BIOTRANSFORMATION KINETICS OF BENZALDEHYDE TO L-PHENYLACETYLCARBINOL (L-PAC) BY IMMOBILIZED CANDIDA UTIUS AND ITS PYRUVATE DECARBOXYLASE

by

HYOUN SEUNG SHIN

A thesis submitted for the Degree of Doctor of Philosophy at the University of New South Wales

Department of Biotechnology The University of New South Wales September, 1994 Declaration

I hereby declare that this submission is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another person nor material which to a substantial extent has been accepted for the award of any other degree or diploma of the University or other Institute of higher learning, except where due acknowledgment is made in the text.

Hyoun Seung Shin

September, 1994 I dedicate this work to my parents, wife and family. TO THE GLORY OF GOD

We know that in all things God works for the good of those who love him, who have been called according to his purpose. -Romans 8:28. 1

ACKNOWLEDGMENTS

I wish to thank my supervisor, Professor Peter L. Rogers, for his excellent advice, thoughtful guidance and encouragement throughout this study. I am grateful also to Professor Kye Joon Lee, Department of Microbiology, Seoul National University who kindly recommended me to undertake my Ph.D. course in Australia.

I would like to thank Professor Peter. P. Gray, Head of Department, Professor Noel W. Dunn, Dr. R. Cail, all other staff members and postgraduate students of the Dept. of Biotechnology, in particular G17 members including Prof. Y. H. Rhee, Wang, Anka, Yaowaluk, Jang, Michael, Keith, Kanda, Somchit and Yvonne for discussions, suggestions and technical assistance.

It is my pleasure to thank the Department of Employment, Education and Training (DEET) of the Government of Australia for the award of a Korea/ Australia Postgraduate Research Scholarship, and I would like to extend my appreciation also to Mrs. Rosemary Plain, Scholarships Unit of UNSW for her advice and assistance.

I wish to extend my thanks to the good fellows at the Sydney Korean Parish, the Uniting Church in Australia, for their sincere encouragement.

My heartfelt appreciation goes to my family, in particular my wife, Ae Hee and my daughter, Hyo Eun (Sharon) for their encouragement, endurance and understanding during the years. 11

ABSTRACT

Biotransformation of benzaldehyde to L-phenylacetylcarbinol (L-P AC) as a key intermediate for L-ephedrine synthesis has been evaluated using immobilized Candida utilis and its free and immobilized pyruvate decarboxylase (PDC). PDC catalyzes the condensation reaction of benzaldehyde (BZ) and 'active acetaldehyde'.

Enhancement of PDC by induction of fermentative conditions and pulse feeding of glucose was conducted prior to immobilization (entrapment in 3 % calcium alginate). Although the immobilized cells have a higher resistance towards the toxic or inhibitory effects of benzaldehyde, benzyl alcohol (BA) production was higher than for the free cell process. During biotransformation, the BZ level and RQ (respiratory quotient) significantly affected both L-PAC and BA formation. By controlling the BZ level at 2 g/L, maintaining RQ=5-7 and pulse feeding glucose, a concentration of 15.6 g/L L­ p AC was achieved in batch culture. In a single stage continuous process, the steady state L-PAC concentration was reduced, however, 0.33 g/L/h L-PAC productivity could be sustained for more than 110 hat optimal conditions. With purified PDC, several catalytic reactions occurred simultaneously and gave rise to acetaldehyde and acetoin as by-products. Various modifications of the reaction conditions involving pH, temperature, addition of ethanol, optimal PDC activity and substrate ratios, led to an increase in L­ p AC and minimization of by-products. The highest L-P AC concentration of 28.6 g/L was achieved at 7 unit/mL PDC activity and 200 mM benzaldehyde with 2.0 molar ratio of pyruvate to benzaldehyde in 40 mM KH2PO4 reaction buffer (pH 7.0) at 4°C. In an evaluation of the immobilized enzyme process, entrapment of PDC into spherical polyacrylamide beads was successfully performed with 12.5 % activity retention, longer half life and modified kinetic parameters compared to free enzyme. In a batch process, the immobilized PDC generally produced lower L-P AC at the same concentrations of substrates. However, in this latter case, L-P AC formation could occur at higher benzaldehyde concentrations (e.g. 300 mM BZ) with the highest L-PAC being 27.1 g/L. With the continuous process, when 50 mM BZ and 100 mM sodium pyruvate were 111

fed into a packed-bed reactor, 0.56 g/L/h productivity was obtained at an average L-PAC concentration of 4.5 g/L with 32 days enzyme half life at 4°C. In summary, the present study has resulted in L-PAC concentrations as high as 28.6 g/L, which compare well with fed-batch values from our laboratory of 22 g/L, and previously reported maximum values of 12 g/L from the literature. An economic assessment which takes account of substrate (pyruvate) cost, as well as that of producing purified and/ or immobilized PDC, is needed for a realistic evaluation of the commercial potential of the various modes of L-P AC production. lV

Conference Presentations

Hyoun S. Shin and Peter L. Rogers (1993) Kinetics of bioconversion of benzaldehyde to L-phenylacetylcarbinol by purified pyruvate decarboxylase (PDC) from Candida utilis. 11th Australian Biotechnology Conference, 20-24 September, Perth. Australia pp. 240-242.

Wang, Bin, H. S. Shin and P. L. Rogers (1994) Microbial and Enzymatic biotransformation of benzaldehyde to L-phenylacetylcarbinol (L-PAC), an intermediate in L-ephedrine production. 3rd Asia-Pacific Biochemical Engineering Conference, 13-15, June. Singapore. pp. 249-252. V

AIMS OF THE PRESENT STUDY

The present study follows from previous evaluations in our laboratory of L-PAC production using batch, fed-batch and continuous modes of operation with Candida utilis. The specific aims of the investigation are:

(1) evaluation of fermentative enzyme profiles of free and immobilized Candida utilis to find optimal conditions for high PDC activity; (2) kinetic evaluation of immobilized C. utilis for biotransformation of benzaldehyde to L-PAC; (3) characterization and kinetic evaluation of purified pyruvate decarboxylase (PDC) as the key catalytic enzyme for biotransformation. This evaluation includes an investigation of various PDC sources, partial purification of PDC, evaluation of biotransformation kinetics and subsequent improvement of biotransformation efficiency for L-P AC production; (4) immobilization of pyruvate decarboxylase and evaluation of the resultant kinetics of biotransformation. This includes an investigation of immobilization methods, assessment of the stability of the immobilized enzyme, optimization of reaction conditions, and comparison of kinetic parameters with free enzyme; (5) application of the immobilized enzyme system to continuous L-PAC production and comparison with previous batch data.

From these investigations, it is anticipated that a clearer understanding will emerge of L-P AC production from benzaldehyde, as well as the provision of data for a comparative analysis of various reaction modes for the L-PAC biotransformation process. VI

TABLE OF CONTENTS

Acknowledgment i Abstract ii List of Publication iii Aims of the present study V Table of contents vi List of Figures xii List of Tables xv

Chapter 1 Literature Review

1.1 Biotransformations 1 1.1.1 Introduction 1 1.1.2 Characterization of Biotransformations 2 1.1.3 Current Trends of Biotransformation processes 5 1.1.3.1 Introduction 5 1.1.3.2 Biotransformation in homogenous water­ miscible organic solvents 8 1.1.3.3 Biotransformation in two phase system 9 1.1.3.4 Biotransformation in mono-phasic organic solvents 11 1.1.3.5 Biotransformation in reverse micelles 13 1.1.3.6 Biotransformation of supercritical fluids 16 1.2 L-Ephedrine 18 1.2.1 Introduction 18 1.2.2 Medicinal uses of L-ephedrine 20 1.2.3 Production of Ephedrine 22 1.2.3.1 Ephedrine from natural herbs 'Ma Huang' 23 1.2.3.2 Chemical synthesis of ephedrine 27 1.2.3.3 Ephedrine synthesis via biotransformation of benzaldehyde to L-phenylacetylcarbinol 30 1.3 Biotransformation of benzaldehyde to L-phenylacetylcarbinol 34 1.3.1 Reaction mechanism of pyruvate decarboxylase for acyloin formation 34 1.3.2 Screening of microorganisms for L-P AC 36 1.3.3 Factors affecting the biotransformation with yeast cells 38 vu

1.3.3.1 Factors affecting the growth phase 38 1.3.3.2 Factors affecting the biotransformation phase 39 1.3.4 Kinetics of Biotransformation 41 1.4 Pyruvate decarboxylase (PDC) 45 1.4.1 Characteristics of pyruvate decarboxylase 45 1.4.2 Reaction mechanism of PDC 47 1.4.3 Roles of pyruvate decarboxylase 48 1.4.4 Genetic regulation of PDC 52

CHAPTER 2 Materials and Methods 55

2.1 Microorganisms 55 2.2 Media 55 2.2.1 Stock culture medium 55 2.2.2 Seed culture medium 56 2.2.3 Medium for cultivation in Erlenmeyer flasks and LH fermenter 56 2.2.4 Medium for cultivation in 100 L fermenter 56 2.2.5 Media Sterilization 57 2.3 Culture systems 58 2.3.1 Flask culture 58 2.3.2 LH fermenter 58 2.3.3 Porton-type stirred tank reactor 59 2.3.4 Pilot scale 100 L fermenter 62 2.3.4.1 The vessel 62 2.3.4.2 The control console 64 2.3.5 Sterilization of fermenters 65 2.4 Biotransformation system with immobilized cells 66 2.5 Methods of analysis 70 2.5.1 Procedure of sample preparation for analysis 70 2.5.2 Cell disruption and enzyme extraction 72 2.5.2.1 Extraction of enzymes from free cells 72 2.5.2.2 Extraction of enzymes from immobilized cells 72 2.5.2.3 Extraction of enzymes with high pressure 72 homogenizer 2.5.3 Dry cell weight estimation 73 2.5.4 Estimation of glucose concentration 73 Vlll

2.5.5 Estimation of ethanol and acetaldehyde concentrations 74 2.5.6 Estimation of benzaldehyde, benzyl alcohol and L-P AC concentrations 74 2.5.7 Pyruvic acid determination 75 2.5.8 Estimation of acetaldehyde by enzymatic analysis 76 2.5.9 Colorimetric determination of acetoin 76 2.5.10 Analysis of enzyme activities 77 2.5.10.1 Pyruvate decarboxylase (PDC) 77 2.5.10.2 Alcohol dehydrogenase for ethanol (ADHr) 78 2.5.10.3 Aromatic alcohol dehydrogenase (ADHcm) 78 2.5.11 Protein determinations 79 2.6 Preparation of specimens for Scanning Electron Microscope 81 2.7 Determination of respiratory quotient (RQ) values 82 2.8 Evaluation of Kinetic parameters 83 2.8.1 Yield coefficients 83 2.8.2 Biotransformation kinetics 84 2.8.2.1 Kinetic parameters in batch 84 2.8.2.2 Kinetic evaluation in continuous process 85

Chapter 3 Kinetic evaluation of biotransformation of benzaldehyde to L-phenylacetylcarbinol (L-P AC) by Immobilized Candida utilis 87 3.1 Introduction 87 3.2 Enzyme Profiles of pyruvate decarboxylase (POC), total alcohol dehydrogenase (ADHr) and aromatic cinnamyl alcohol dehydrogenase (ADHcm) for C. utilis 88 3.2.1 Time course and enzyme profiles of PDC, ADHr and ADHcin with constant air flow 89 3.2.2 Effects of aeration rate and agitation speed on enzyme profiles 92 3.2.3 Effect of initial glucose concentration on fermentative enzyme profiles 95 3.2.4 Effect of pulse feeding of glucose on enzyme profiles 97 3.3 Immobilization of C. utilis cells and associated enzyme profiles 99 3.4 Biotransformation of benzaldehyde to L-P AC by Immobilized 101 Candida utilis cells IX

3.4.1 Comparison of biotransformation by free cells and immobilized cells with various concentrations of benzaldehyde 102 3.4.2 Effect of benzaldehyde concentration on L-PAC production by immobilized cells under controlled conditions 104 3.4.3 Effect of RQ value on L-P AC production during biotransformation 107 3.4.4 Time course of biotransformation with immobilized C. utilis cells 109 3.5 Evaluation of a continuous immobilized cell process for L-P AC production 114 3.6 Morphological changes of immobilized C. utilis following exposure to benzaldehyde 116 3.7 Discussion and Conclusions 120

CHAPTER 4 Characterization of purified pyruvate decarboxylase (PDC) from Candida utilis and Saccharomyces cerevisiae 123 4.1 Introduction 123 4.2 Evaluation of PDC Enzyme Sources 124 4.3 Comparison of biotransformation kinetics for L-PAC formation by purified PDC enzyme from C. utilis and S. cerevisiae 129 4.4 Purification of PDC from C. utilis 132 4.5 Characterization of PDC 138 4.5.1 Effect of buffer species on the stability of PDC 138 4.5.2 Effect of TPP on PDC stability 140 4.5.3 Determination of Km value for pyruvate 140 4.5.4 Effect of pH on PDC stability 140 4.5.5 The requirement of TPP for reconstruction of PDC apoenzyme to PDC holoenzyme 143 4.5.6 Effect of Mg2+ on PDC activity 146 4.6 Effect of reaction temperature on L-PAC formation 148 4.7 Effect of acetaldehyde on the initial rate of L-P AC formation 151 4.8 Toxic effects of benzaldehyde and L-P AC on PDC activity 153 4.9 Discussion and conclusions 155 X

CHAPTER 5 Kinetic evaluation of biotransformation for L-phenylacetylcarbinol formation by purified pyruvate decarboxylase from Candida utilis 158 5.1 Introduction 158 5.2 Effect of ethanol concentration on L-P AC formation 160 5.2.1 Effect of ethanol on initial reaction rates for L-PAC formation 160 5.2.2 Effect of ethanol on PDC stability 162 5.2.3 Effect of ethanol on the efficiency of L-P AC formation 164 5.3 Effect of pH on the rates of L-P AC and acetaldehyde formation 164 5.4 Effect of PDC activity on the initial rate of L-PAC formation 167 5.5 Effect of molar ratios of pyruvate to benzaldehyde on the initial rates for L-PAC formation 170 5.6 Determination of saturation constant (Ks) for benzaldehyde 172 5.7 Effect of molar ratios of pyruvate to benzaldehyde on final L-PAC concentrations 174 5.8 Effect of PDC activities on L-P AC formation with various molar ratios of pyruvate to benzaldehyde 177 5.9 Typical biotransformation kinetics with purified PDC 177 5.10 Discussion and conclusions 181

CHAPTER 6 Kinetic evaluation of biotransformation of benzaldehyde to L-phenylacetylcarbinol by immobilized pyruvate decarboxylase 182 6.1 Introduction 182 6.2 Immobilization of pyruvate decarboxylase 183 6.2.1 Manometric measurement of pyruvate decarboxylase activity 183 6.2.2 Adsorption of PDC on cationic exchange resins 184 6.2.3 Entrapment of PDC enzyme into gel matrix 187 6.3 Improvement of entrapment of PDC into polyacrylamide gel 189 6.3.1 Effect of acrylamide concentrations on PDC activity of immobilized beads 189 6.3.2 Effect of glutaraldehyde concentration on activity of PDC immobilized in polyacrylamide gel 190 6.3.3 Effect of glutaraldehyde on stability of immobilized PDC 190 Xl

6.3.4 Preparation of spherical polymer beads 193 6.4 Characterization of immobilized PDC 198 6.4.1 Determination of apparent Km for pyruvate 198 6.4.2 Effect of pH on stability of immobilized PDC 198 6.4.3 Effect of organic solvents on stability of immobilized PDC 201 6.4.4 Comparison of toxic effects of benzaldehyde and L-P AC on 201 immobilized PDC 6.5 Biotransformation properties of Immobilized PDC 204 6.5.1 Effect of pH on L-PAC formation with immobilized PDC 204 6.5.2 Effect of organic solvents on L-P AC formation 204 6.5.3 Effect of temperature on L-PAC formation 206 6.5.4 Effect of PDC activity of immobilized beads on L-P AC formation 209 6.5.5 Effect of molar ratio of pyruvate to benzaldehyde on L- P AC formation with immobilized PDC 209 6.5.6 Typical biotransformation kinetics with immobilized PDC 211 6.6 Evaluation of continuous process with immobilized PDC 213 6.6.1 Configuration of continuous immobilized PDC process 213 6.6.2 Effect of space time on acetaldehyde formation of PBR 216 6.6.3 Effect of benzaldehyde concentration on L-P AC formation 216 6.6.4 Effect of pyruvate concentration on L-P AC formation 218 6.6.5 Long term continuous biotransformation of benzaldehyde to L-PAC 222 6.7 Discussion and conclusions 224

CHAPTER 7 General discussion and conclusions 227 7.1 L-PAC formation with immobilized C. utilis 227 7.2 L-P AC formation with purified PDC 229 7.3 L-P AC formation with immobilized PDC 232 7.4 Comparison with other biotransformation process for L-P AC formation 233 7.5 Possible future studies 235 Reference 236 Appendices 260 Appendix 1 Nomenclature 260 Appendix 2 Calculation of the enzyme activity 262 Appendix 3 Manometric measurement 263 Xll

LIST OF FIGURES

Figure 1.1 Schematic presentation of different ways to use enzymes together with organic solvents 7 Figure 1.2 (a) Schematic representation of the interaction of substrate molecules distributed in the reverse micelle system;(b) ratio of initial reaction rate to enzyme concentration in reverse micelles 15 Figure 1.3 The structural formulae for ephedrine and its analogues 19 Figure 1.4 Metabolism of ephedrine 22 Figure 1.5 Picture of Ephedra sinica 24 Figure 1.6 Extraction and purification of ephedrine from Ephedra 26 plant Figure 1.7 Chemical synthesis of Ephedrine 29 Figure 1.8 Synthesis of L-ephedrine via biotransformation 31 Figure 1.9 Summary of L-ephedrine production via semi-biological synthesis 33 Figure 1.10 A proposed mechanism for the pyruvate decarboxylase 49 Figure 1.11 Restriction map YRp7-PDC 1 53 Figure 2.1 (a) Schematic diagram of the Porton type stirred tank reactor; (b) photograph the Porton type stirrer tank 60 reactor Figure 2.2 The design features of the 100 L fermenter 63 Figure 2.3 Schematic diagrams for the biotransformation system with immobilized cells 67 Figure 2.4 Photograph of (a) LH fermenter containing immobilized cells; (b) Exit Gas Analyzer for CO2; (c) Exit Gas Analyzer for 02; (d) Dehumidifier; (e) Data processor; (f) NEC computer 68 Figure 2.5 Procedures for sample preparation for analysis of chemicals and enzymes 70 Figure 3.1 (a) Growth of C. utilis on 60 g/L glucose based medium with 0.6 vvm aeration rate; (b) fermentative enzyme profiles; (c) RQ values computed from analysis of 02 and CO2 concentrations in exit gases 90 Xlll

Figure 3.2 (a) Growth of C. utilis on 60 g/L glucose based medium with changing aeration rate from 0.6 vvm to 0.2 vvm and agitation speed from 1,000 to 500 rpm after 7 h cultivation, (b) fermentative enzyme profiles and (c) RQ values computed from analysis of 0 2 and CO2 concentrations in exit gases as a function of culture time. 93 Figure 3.3 (a) Growth of C. utilis on 90 g/L glucose based medium with changing aeration rate and agitation speed after 7 h cultivation; (b) fermentative enzyme profiles 96 Figure 3.4 Effect of pulse feeding of glucose on (a) kinetics of growth and (b) fermentative enzyme profiles 98 Figure 3.5 (a) Kinetics for immobilized cells; (b) fermentative enzyme profiles with pulse feeding of glucose 100 Figure 3.6 Biotransformation kinetics with immobilized C. utilis on the level of (a) 0.8 g/L; (b) 1.5 g/L; (c) 2.0 g/L; (d) 4.0 g/L BZ after 4 h initial adaptation period. 105 Figure 3.7 Effect of various levels of benzaldehyde on L-P AC formation as a function of time 108 Figure 3.8 Biotransformation time course with immobilized cells 111 Figure 3.9 Phase Contrast Microscope photograph of calcium- alginate beads containing Candida utilis 117 Figure 3.10 Electron Microscope photograph of viable Candida utilis entrapped in calcium-alginate matrix 118 Figure 3.11 Comparison of Electron Microscope photographs of immobilized cells in calcium-alginate beads (a, b) before and (c, d) after biotransformation 119 Figure 4.1 (a) Growth of S. cerevisiae on 90 g/L glucose based medium in 2 L LH fermenter with pulse feeding 30g/L glucose; (b) fermentative enzyme profiles 126 Figure 4.2 (a) Growth of C. utilis on 90 g/L glucose based medium in 100 L fermenter with pulse feeding 30g/L glucose; (b) fermentative enzyme profiles 128 Figure 4.3 Biotransformation kinetics with 3 unit/ mL PDC enzyme in shake flasks at 25°C: (a) PDC from C. utilis ; (b) PDC from S. cerevisiae 130 XIV

Figure 4.4 Photograph of (a) 100 L Fermenter; (b) Westfalia Continuous Separator; (c) Sharples Super Centrifuge; (d) Manton-Gaulin High Pressure Homogenizer 133 Figure 4.5 Chromatography of the pyruvate decarboxylase from C. utilis on Sephacryl S-300 137 Figure 4.6 Effect of buffer species (pH 6.0) on the stability of PDC from C. utilis in presence and absence of 0.03 mM TPP at 25°C 139 Figure 4.7 Effect of TPP concentration on PDC stability in 40mM KH2PO4 buffer (pH 6.0) at 25°C 141 Figure 4.8 Determination of Km value for pyruvate: (a) C. utilis ; (b) S. cerevisiae at pH 6.0 and 25°C 142 Figure 4.9 pH dependence of PDC stability in sodium citrate/phosphate buffer at 4°C 144 Figure 4.10 Effect of TPP concentration on the reactivation of apoenzyme (10 unit/mL) with 40 mM KH2PO4 buffer (pH 6.0) at 4°C 145 Figure 4.11 Effect of Mg2+ on PDC activity at 4°C and pH 6.0 147 Figure 4.12 Biotransformation kinetics at various temperatures (a) 4°C; (b) 10°C; (c) 25°C 149 Figure 4.13 Effect of acetaldehyde on (a) initial reaction rate for L-P AC formation; (b) bioconversion kinetics 152 Figure 4.14 Toxic effect of benzaldehyde and L-PAC on purified PDC from C. utilis 154 Figure 4.15 Proposed two-site reaction mechanism catalysed by pyruvate decarboxylase for L-P AC formation 157 Figure 5.1 Effect of ethanol on initial rate for L-P AC formation at 4°c 161 Figure 5.2 Effect of ethanol on the stability of purified PDC from C. utilis 163 Figure 5.3 Comparison of GC analysis profiles for biotransformat- ion products with and without 2.0 M ethanol 165 Figure 5.4 Comparison of effect of pH on relative reaction rates for acetaldehyde and L-PAC formation at 4°C 166 xv

Figure 5.5 Effect of PDC enzyme activity (unit/mL) on (a) initial reaction rate; (b) specific rate for L-PAC formation in the presence of 3 times higher mole of sodium pyruvate and corresponding benzaldehyde 169 Figure 5.6 Initial reaction rates for L-PAC formation as a function of benzaldehyde concentration and various molar ratios of pyruvate to benzaldehyde 171

Figure 5.7 Determination of saturation constant (K5) for benzaldehyde 173 Figure 5.8 L-PAC yields (mole/mole%) based on benzaldehyde as a function of various molar ratios of pyruvate to benzaldehyde in the presence of various concentrations of benzaldehyde with 7 unit/mL PDC enzyme at pH 7.0 and4°C 175 Figure 5.9 Comparison of the final L-PAC formation with various PDC activities, and equal molar concentration of benzaldehyde and sodium pyruvate in the range of 200 to 300mM 179 Figure 5.10 Biotransformation kinetics: Reaction mixture consists of 40 mM KH2PO4 buffer (pH 7.0) containing 150 mM benzaldehyde, 225 mM sodium pyruvate, 2.0 Methanol and 7 unit/mL PDC at 4°C 180 Figure 6.1 Manometric measurement of PDC activities 185 Figure 6.2 Immobilized PDC activity of various cationic exchangers loaded with various PDC activity solutions 186 Figure 6.3 Immobilized PDC activity with entrapment in 10 % polyacrylamide and 2 % calcium alginate beads compared with various activities of original PDC solution 188 Figure 6.4 Effect of acrylamide monomer concentrations on apparent PDC activity of immobilized bead entrapping from 120 unit/mL PDC activity solution 191 Figure 6.5 Effect of glutaraldehyde concentrations on PDC activity of 10 % polyacrylamide beads entrapping from 120 unit/mL PDC activity 192 Figure 6.6 Effect of glutaraldehyde concentrations on stability of PDC activity of immobilized beads at 4°C 194 XVl

Figure 6.7 Schematic diagram for preparation of bead-form acrylamide polymer 195 Figure 6.8 Photograph of (a) spherical polyacrylamide beads containing PDC enzyme (magnification, lOx) Photograph of active stained immobilized PDC beads (b) on a plate: (c) in suspension 196 Figure 6.9 Determination of apparent Km of immobilized PDC for pyruvate 199 Figure 6.10 Effect of pH on stability of immobilized PDC enzyme after 10 days incubation at 4°C 200 Figure 6.11 Effect of organic solvents on stability of immobilized PDC activity after 10 days incubation at 4°C 202 Figure 6.12 Toxic effects of benzaldehyde and L-P AC on immobilized POC. 203 Figure 6.13 Effect of pH on initial rate of L-P AC formation with immobilized PDC at 4°C 205 Figure 6.14 Effect of organic solvents on initial rate of L-P AC formation with immobilized PDC enzyme at 4°C 207 Figure 6.15 Biotransformation kinetics with immobilized PDC at 4°C 212 Figure 6.16 Schematic diagram of the continuous process with immobilized PDC 214 Figure 6.17 Photograph of the continuous process with a packed bed reactor 215 Figure 6.18 Effect of space time ('t) on acetaldehyde formation in a packed-bed reactor with various concentrations of sodium pyruvate at 4°C 217 Figure 6.19 Effect of benzaldehyde on L-PAC formation in a packed- bed reactor at 4 °C 219 Figure 6.20 Effect of varying pyruvate concentrations with 50 mM benzaldehyde on L-P AC formation in a packed-bed reactor at various space times 220 Figure 6.21 Steady state kinetics of a packed-bed reactor at 4°C with 50 mM benzaldehyde, and (a) 75 mM or (b) 100 mM sodium pyruvate 221 Figure 6.22 Continuous L-PAC formation with 50 mM benzaldehyde and 100 mM sodium pyruvate by immobilized PDC in a packed-bed reactor at 4°C 223 xvn

LIST OF TABLES

Table 1.1 Potential advantages of employing enzymes in organic as opposed to aqueous media 6 Table 1.2 Biotransformation products of water miscible solvents phase 8 Table 1.3 Biotransformation products of aqueous two-phase systems 10 Table 1.4 Biotransformations in water immiscible organic two- phase systems 11 Table 1.5 Biotransformation products with mono-phasic organic solvents 13 Table 1.6 Biotransformation products of reverse micellar phases 14 Table 1.7 Biotransformation products of supercritical phase 17 Table 1.8 Comparison of biotransformation of benzaldehyde to L-PAC by various yeasts 43 Table 3.1 Comparison of biotransformation products by free cells and immobilized cells after 16 h incubation with various initial concentrations of benzaldehyde 102 Table 3.2 Effect of aeration rate on the biotransformation of benzaldehyde to L-P AC 109 Table 3.3 Kinetic parameters of continuous process with immobilized normal cells density 115 Table 3.4 Kinetic parameters of continuous process with immobilized double cells density 116 Table 3.5 Summary of fermentative enzyme activities in C. utilis under various culture conditions 120 Table 3.6 Summary of biotransformation products with immobilized cells with various BZ levels maintained in the bioreactor 121 Table 3.7 Comparison of kinetic parameters of batch and continuous process with immobilized cells 122 Table 4.1 ADHr, ADHcin and PDC activities of wheat germ during germination 125 Table 4.2 Comparison of kinetic parameters of S. cerevisiae and C. utilis grown on glucose based media 129 XVlll

Table 4.3 Disruption of C. utilis by high pressure homogenizer at 9000 psi (640 kPa) 132 Table 4.4 Fractional purification of PDC enzyme with ammonium sulphate 135 Table 4.5 Summary of results of purification of pyruvate decarboxylase from C. utilis 136 Table 4.6 Comparison of initial kinetic parameters and final L-PAC concentration at various temperatures 148 Table 5.1 Comparison of initial rates for biotransformation products over first 30 min incubation at pH 6.0 and pH 7.0 at 4°C 167 Table 5.2 Acetaldehyde formation with various PDC activities in the presence of equi-molar benzaldehyde and sodium pyruvate at 4°C 170 Table 5.3 L-P AC formation with various molar ratios of pyruvate to benzaldehyde at 4°C 176 Table 5.4 Effect of PDC activities on L-P AC formation with various molar ratios of pyruvate to benzaldehyde 178 Table 6.1 Immobilization capacity of various matrices for PDC with 90 unit/mL PDC solution 189 Table 6.2 Effect of temperature on biotransformation kinetics with immobilized PDC 208 Table 6.3 Final products of biotransformation with various PDC activities per unit volume of beads, 70 mM benzaldehyde and 70 mM sodium pyruvate in the reaction buffer at 4°C 209 Table 6.4 Effect of molar ratios of pyruvate to benzaldehyde on L-P AC formation and molar conversion yields by immobilized PDC 210 Table 6.5 Comparison of L-P AC formation and productivities with 50 mM benzaldehyde, 1.5 and 2.0 molar ratios of pyruvate to benzaldehyde for continuous process 222 Table 7.1 Comparison of characteristics of PDC from various sources 232 Table 7.2 Comparison of kinetic parameters from various processes for L-P AC formation 234 Appendices Table 1 The solubility of carbon dioxide in pure water (ex value) 265 Chapter 1 1

CHAPTER 1

LITERATURE REVIEW

1.1 BIOTRANSFORMA TI ONS

1.1.1 Introduction

Bioprocessing has received increasing world wide attention with respect to production of high value products, reduced environmental pollution and economic utilization of natural resources. Biotransforma tion, one of such bioprocesses, can be defined as the transformation or modification of natural products or chemically synthesised substrates, which are catalysed by living organisms or their enzymes. In some cases, it is not possible to clearly distinguish between biotransformation and biodegradation in which multi-stage reaction pathways are involved in the degradation of biopolymers or macro­ molecules of complex structure. In general, biotransformations can be considered as selective and enzymatic modifications of defined pure compounds into defined final products (Kieslich, 1984). Biotransformations have been carried out with cultures of microorganisms, plant cells, animal cells and their purified enzymes. In such processes, both constitutive and inducible enzymes have been employed to synthesize and degrade chemical compounds. Specific reactions can be catalysed by particular enzymes in highly complex and well coordinated metabolic pathways. In addition to their usual substrates, many of these enzymes can interact with other structurally related compounds and thus catalyse 'unnatural' reactions when foreign substrates are added to the medium. Several early processes which fulfilled these requirements were: the oxidation of ethanol to acetic acid by Bacterium xylinum, the oxidation of glucose to gluconic acid by Acetobacter aceti and acyloin formation from benzaldehyde and a fermentable sugar by yeast (Neuberg and Hirsch, 1921). Chapter 1 2

However. biotransformations did not achieve their present significance until microbial transformations to produce pharmaceutical compounds and food additives such as steroids, antibiotics, vitamins, high fructose syrup, amino acids and organic acids were developed. Present! y. biotransformation is one of the fastest growing areas of biotechnology since diverse transformation processes and a variety of types of reaction leading to specific modification of chemical compounds can be employed. The general goals of current biotransformation practice may be considered as the following:

• specific modification of the substrate structures via selective transformation reactions, e.g. steroid modifications (Smith, 1984); • partial degradation of substrate into desirable metabolites by means of controlled microbial reaction pathways, e.g. 6-amino penicillanic acid and corresponding acryl side chains (Okachi and Nara, 1973); • alternation of the substrate structure by the use of biosynthetic reactions to produce artificial structures, e.g. peptide synthesis (Margolin and Klibanov, 1987).

1.1.2 Characterization of Biotransformations

Biotransformations have the following desirable characteristics (Crueger and Crueger, 1982):

• reaction specificity: the catalytic activity is usually restricted to a single reaction type in which side reactions are not likely to occur as long as one enzyme is involved in the biotransformation, e.g.

CHO CH.,OH I I - H-c-oH C=O I I OH-C- H OH-C- H I I H _ C- OH glucose isomeruse H-C-OH I I H-C-OH H-C-OH I I CH OH CH2OH 2

Glucose Fructose Chapter 1 3

• regiospecificity: the substrate molecule is usually attacked at the same site, even if several groups of equivalent or similar reactivity are present (Peterson et al., 1952), e.g.

Rhi;.ophus nixricans

Progesterone 11-u-Hydroxyprogesteron

• stereospecificity: the reactive center of the enzyme provides an asymmetric environment which allows it to distinguish between the enantiomers of a racemic substrate. Therefore, only one (or preferentially one) of the enantiomers is catalysed. The product is, therefore, optically active (Schmidt-Kastner and Egerer, 1984), e.g.

D. L L-specific biocatalyst L D x-cH-COOR,-----~~ x-cH-COOH + x-cH-cooR, I Est erase 1 1 NH1 NH2

D. L L-specific biocatalyst L D X-CH- COOR 1 X- CH - COOH + X- CH- COOR 1 I Esterase I I NH-COR2 NH - COR2 NH- COR1

D D. L L-specific biocatalyst L X- CH- COOH X- CH- COOH + X-CH-COOH I I I NH- COR2 Acylase NH1 NH- COR2

D.L L-specific biocatalyst L D X- CH- CONH, ------X- CH- COOH + X- CH- CONH, I • 1 I • Amidase NH2

D D.L D-specific biocatalyst L x-cH- co x-cH- co + x-cH- COOH I I I I I Hydantoinase NH-CONH2 HN NH HN NH ' / ..... / co co Chapter 1 4

• mild reaction conditions: the activation energy of the reaction is significantly lowered by the interaction of substrate and enzyme, thus biotransformations take place under mild conditions (temperature below 40°C, pH near neutrality and normal pressure). Even labile molecules can be converted without undesired decomposition or isomeriza tion.

On the other hand, biotransformations can have some disadvantages or problems associated with biocatalysts having the following unfavourable characteristics:

• by-products can be synthesised also due to other enzymatic reactions in living cells in addition to the desired reactions. Side reactions in biotransformation processes are observed usually with free and immobilized whole cell processes when a substrate or its product is metabolized by undesirable enzymes, e.g. benzyl alcohol production in L-phenylacetylcarbinol formation (Long and Ward, 1989a). • inhibition by substrates and products can cause the loss of cell viability and biocatalytic activity. To overcome this effect during the catalytic period, addition of inhibitory substrate must be closely matched to the product formation, e.g. biodegradation of chlorophenol mixture (Dapaah and Hill, 1992). • additional nutrients may be required to maintain growth and viability of the cells. The presence of complex media components may cause purification difficulties in addition to low product concentrations in the reaction mixture. • reduction, or loss of activity of the biocatalyst in aqueous and/ or organic solvents can occur, particularly when hydrophobic substrates are used. Most enzymes are inactivated when non-aqueous solvent concentrations exceed 30 % which limits the application of these systems (Reslow et al., 1992). • maintenance of consistent catalytic activity can be difficult. When enzymes require cofactors, or possess a di-, or a tetrameric protein structure, they can be easily dissociated into a monomer, with resulting decreased activity so that conversion efficiency is variable. Physical stress with pressure drop in a packed bed reactor, and shear Chapter 1 5

stress in a stirred tank reactor can also cause some degeneration of activity of a biocatalyst. • in the case of an enzyme or free cell process, it may be difficult or impossible to reuse these biocatalysts.

1.1.3 Current Trends of Biotransformation Processes

1.1.3.1 Introduction

Recently, interest has focused increasingly on the catalytic activity of biocatalysts in the presence of organic solvents such as aqueous organic solvents and/or organic solvents. Originally, addition of an organic solvent to a reaction mixture was aimed at the biotransformations for water­ insoluble steroid formation (Cremonesi et al., 1973). Two phase media, water/ organic solvent, were examined for several biocatalytic reactions (Buckland et al., 1975). The advantages of multiphase systems are: the increased solubilities of less polar species allow higher concentrations of potential reactants and products to be used in the process; cheap and readily available hydrolytic enzymes may be used as efficient catalysts for the reverse, synthetic reactions; the biocatalyst may be protected from non-polar starting materials or products which can inactivate it due to partition into the organic phase; similarly decomposition of substrates and products in the aqueous phase may be reduced by extraction of reactant into the organic phase. Recently changes in regioselectivity, in chemoselectivity and even in stereoselectivity have been reported for enzymes in organic solvents (May, 1992). A theory describing chemical equilibrium shift was suggested by Martinek et al. (1981). These authors have shown quantitatively that the overall product concentration in aqueous/ organic two phase systems can be related to the partition coefficients of substrates as well as products and to the volume ratio of the two phases. Further comprehensive descriptions of how water affects biochemical equilibrium in two phasic reaction mixtures were provided by Halling (1984) and Semenov et al. (1987). There are Chapter 1 6

numerous potential advantages in employing enzymes in organic as opposed to aqueous media as listed in Table 1.1.

Table 1.1 Potential advantages of employing enzymes in organic as opposed to aqueous media (Klibanov, 1986; Halling, 1987)

• Increased solubility of nonpolar substrate. • Shifting of thermodynamic equilibria to favour synthesis over hydrolysis. • Reduction in water-dependent side reaction such as hydrolysis of acid anhydrides or polymerization of quinones. • Immobilization is often unnecessary; enzymes are insoluble in organic solvents. • Recovery of enzyme is possible by simple filtration. • If immobilization is desired, adsorption onto nonporous surface is satisfactory; enzymes are unable to desorb from these surfaces in nonaqueous media. • Ease of product recovery from low boiling solvents. • Enhanced thermal stability of enzymes: water is required to inactivate enzyme at high temperatures. • Elimination of microbial contamination. • Potential of enzymes to be used directly within a chemical process.

However, in organic/water systems, the toxicity of the organic solvent has to be always considered when biotransformation takes place. Thus, interests in 'solvent engineering' is rapidly growing as a means to manipulate the activity and specificity of enzymes in anhydrous media. Organic/water, two-liquid phases have been distinguished as fully dispersed two-liquid phase, non-dispersed two liquid phase and solvent saturated aqueous phase (Bar, 1987). Currently, as shown in Figure 1.1, the application of enzymes in organic/water systems has been schematically presented by Aldercreutz and Mattiasson (1987). Further descriptions and applications of various organic/water systems are summarized as follows: Chapter 1 7

a b C

d e

Figure 1.1 Schematic presentation of different ways to use enzymes together with organic solvents. Water is drav,·n as white area and the organic solvents as shaded areas. In the magnified pictures individual enzyme molecules ( circles marked with E) are shown. (a) Homogeneous system containing water and water-miscible solvent. (b) Two-phase system containing one aqueous phase and one organic phase. The enzyme is located in the aqueous phase. (c) Enzyme immobilized in water-containing, porous particles surrounded by organic solvent. (d) Enzyme adsorbed on porous particle surrounded by organic solvent. Small amounts of water are associated with the enzyme molecules and the support material. (e) Solid enzyme particles suspended in organic solvent. Small amounts of water are associated with the particles. (f) Enzyme in microemulsion (reverse micelles). Inside the reverse micelles the enzyme molecules are surrounded by thin water shells. (g) Covalently modified enzyme soluble in organic solvents. Small amounts of water are associated with the enzyme molecules and in some cases with the modifying residues (Aldercreutz and Mattiasson, 1987). Chapter 1 8

1.1.3.2 Biotransformation in homogenous water-miscible organic solvents

The biotransformation of lipophilic compounds is limited by their solubility in water. Emulsifiers, e.g. tween and sodium dodecyl sulphate (SDS), and low toxicity solvents, e.g. ethanol, acetone, dimethylformamide (DMF), dimethyl sulphoxide (DMSO) and acetonitrile may be used to increase the solubility of poorly soluble compounds. When enzymes are added to aqueous solutions containing high concentration of water-miscible solvents, generally they are unfolded and lose their activity, but some water-miscible compounds enhance the stability of enzyme (Butler, 1979; Freeman, 1984). In continuous processes, there are examples where enzymes have been used in an homogenous mixture of water and organic solvents without any considerable loss of enzyme activity. A column containing immobilized glucose oxidase was operated continuously for 14 days in a 20 % (v /w) acetone solution (Alberti and Klibanov, 1985). Yokozeki et al. (1982) succeeded in transglycosylation reaction using whole cells of Enterobacter aerogenes with 40 % dimethylsulphoxide (DMSO) for 35 days without loss original activity. On the other hand, addition of a water miscible solvent decreases the water activity, thereby changing the equilibrium for hydrolytic reactions as presented in Table 1.2.

Table 1.2 Biotransformation products of water miscible solvents phase

Substrate(s) Product(s) Organic Biocatalyst Reference solvent(s) aromatic amino peptide acetonitrile subtilisin Nagashima acids et al. (1992) neutral, aromatic peptide acetonitrile chymotrypsin Nagashima amino acids etal. (1992) peptides polypeptides dimethylform- a-chymotrypsin Barbas et al. amide (DMF) (1988) peptides penicillin tetrahydrofuran a-chymotrypsin Barbas et al. precursors (THF) subtilisin (1988) tryptophan, formic 1-formyl dioxan trypsin Coletti- hydrochloric acids tryptophan Previero et al. (1969) Chapter 1 9

1.1.3.3 Biotransformation in two phase system

Phase separation generally occurs when two solutions of water­ soluble polymers are mixed, or water-immiscible solvents are added to aqueous solution. In two water-soluble polymer mixtures, the most commonly used polymers are polyethylene glycol and dextran, or instead of using two polymers, a polymer and salt solution can also be used to form a two phase system. Among the polymer/salt systems, PEG/potassium phosphate and PEG/magnesium sulphate are most frequently used (Andersson and Hahn­ Hagerdal, 1990). Although the two phases are no longer miscible, the water content of the two phases exceeds 75 % so that aqueous two-phase systems have several properties that make them suitable for use together with biological material. One of the main advantages of using a water miscible two phase system is that the interfacial tension between both phases is very low (0.0001-0.1 mN m-1) compared with interfacial tension between water and an organic solvent (5-50 mN m-1 ). This leads to dispersion with extremely small droplets of one phase being formed in the other phase upon continuous agitation, Therefore, partition equilibrium of even high molecular weight components, such as proteins, can be reached very quickly under conditions of turbulent mixing. However, some difficulty in the partitioning of the biocatalyst and product in the different phases still remains. Some examples of aqueous two-phase system are listed in Table 1.3.

On the other hand, two liquid phase systems, consisting of water and poorly water miscible solvents, have been currently applied to conduct biotransformation of substrates with low solubility in water (Andersson and Hahn-Hagerdal, 1990; Osborne et al., 1990). In these cases, the system splits into two or more phases, and the biocatalyst remains in a relatively water rich phase that includes only low concentrations of the non-polar organic species. In general, more hydrophobic solvents are preferred to retain enzyme activities (Janssen et al., 1993). Hence, inactivation of catalyst is often less of a problem than with water miscible solvents. Chapter 1 10

Table 1.3 Biotransformation products of aqueous two-phase systems

Substrate(s) Product(s) Phase Biocatalyst Reference components

N-acetyl-D,L- methionine PEG 300, acylase Yamazaki& methionine potassium Suzuki (1979) phosphate penicillin 6-amino- PEG20M, penicillin Andersson penicillanic potassium acylase et al. (1984) acid phosphate hydrocortisone prednisolone PEG 8000 steroid- Kaul& dextran T40 dehydrogenase Mattiasson (1986) fumarate malate PEG 1500, fumarase Yang et al. potassium (1988) phosphate steroids androst-4-ene- PEG 4000, Mycobacterium Flygare (1988) 3,17-dion, dextran T500, androsta-1,4- Brij 35 diene-3,17-dion

In other possible ways to achieve good maintenance of activity, immobilizations of biocatalysts were carried out with various polarity of matrices, in which a greater difference in polarity between the support and the organic solvent led a lower risk of exposure of the enzyme at the interface. Khmelnitski and co-workers (1984) carried out peptide synthesis using a-chymotrypsin in a water/ ethyl acetate two phase system. They found that the free enzyme was inactivated rather quickly while enzyme immobilized terephthalate polymer could be reused successfully. However, high concentration of solvents as well as substrate/products may progressively inhibit the activity of the biocatalyst. Beside the effect of solvents on the activity and stability of the enzyme, there is also a solvent effect on the equilibrium position of reactions in aqueous two phase systems. Substrate and product will partition between both phases in aqueous/organic systems. Martinek et al. (1981) have shown quantitatively that the overall product concentrations in aqueous/ organic two-phase systems can be related to the partition coefficients of substrate as well as products, and to the volume ratio of both phases. To obtain high overall product concentrations, it is essential to use Chapter 1 11

an organic solvent for which the partition coefficient for the product is high. As long as the enzyme is kept in the water phase, good operational stability can be predicted. Antonini et al. (1981) pointed out that the sensitivity of enzymes to organic solvents in two systems varies with solvents. Important parameters in this process are the activity and stability of the enzyme in organic solvents. Several examples of water immiscible organic two-phases are listed in Table 1.4.

Table 1.4 Biotransformations in water immiscible organic two-phase systems

Substrate(s) Product Organic Biocatalyst Reference solvent(s) fatty acid methylketone pentanone, Pseudomonas Cruely et al. neptanone, roquefortii (1992) nonanone cholesterol cholestenone carbon immobilized Atrat et al. tetrachloride, or Nocardia (1980) toluene erythropolis octanoic acid 2-heptanone isoparaffinic Penicillium Larroche et al. solvent roquefortii (1992) dih ydroepiandroster 4-androstene- benzene, Nocardia Fukui et al. -one 1,17-dion heptane rhodocrouds (1980) (R,S) methyl-2-(4- (R)-HPPA toluene - lipase Barton et al. hydroxy-phenoxy) (1990) propionic acids (HPPA-Me)

1.1.3.4 Biotransformation in mono-phasic organic solvents

Biocatalysts can exist in mono-phasic systems which are not predominant aqueous phases. The kinetics and mechanisms of enzyme action in a mono-organic phase can be significantly different from in an aqueous phase. Although variables that affect enzymatic catalysis in organic solvents have yet to be fully elucidated, general trends can be observed in many cases. When exposing an enzyme preparation to organic solvents, it is very important to define: (i) water content, (ii) nature of organic solvent and (iii) purity of enzyme preparation. Chapter 1 12

Although biotransformations can be carried out in mono-organic solvents, the solvents may not be completely anhydrous. Yamane (1987) proposed the technical term 'microaqueous'. It implies that the system is neither aqueous nor anhydrous. Yamane (1988) showed that complete depletion of water from the reaction system results in non-occurrence of a biochemical reaction. Water seems to be essential for a biocatalyst in organic media. One common hypothesis is that the enzyme molecule requires a small hydration layer that acts as the primary component of the enzymic micro-environment (Fukui et al., 1987). This layer acts as a buffer between the enzyme surface and the bulk reaction medium. Furthermore, evidence was provided by Zaks and Klibanov (1988) that an enzyme molecule in aqueous solution becomes surrounded by a few layers of water. In the optimal choice of solvent, it is known that replacing water with an organic solvent in a biotransformation process alters the native conformation of the protein by disrupting hydrogen bonding and hydrophobic interactions, thereby generally leading to reduced activity and stability (Cremonesi et al., 1974). Several factors thus must be taken into account in determining which solvent is most appropriate for a given reaction (Dordick, 1989). First, the compatibility of the solvent with the reaction of interest must be considered. For example, the use of hydrophobic water immiscible solvents is impractical for enzyme catalyzed sugar modifications as there would be no interaction between the insoluble substrate and insoluble enzyme. Similarly, the compatibility of the reaction products with the solvent is also important. Polar products tend to remain in the vicinity of the enzyme and can cause product inhibition or can undergo unwanted side reaction (Stoicheva, 1989). A second factor is that the solvent selected must be inert to the reaction and not react with either substrate or product. Additional factors, which may influence the choice of solvent, include solvent density and viscosity, surface tension, toxicity, flammability, waste disposal and cost. In enzyme preparation, an impure enzyme preparation sometimes may offer good protection for the catalytically active molecules. Duarte and Lilly (1980) suggested that this positive effect could be ascribed to both a protein dilution and creation of a favourable polarity in the vicinity of the catalytic site. Several applications of organic mono-phase are listed in Table 1.5. Chapter 1 13

Table 1.5 Biotransformation products with mono-phasic organic solvents

Substrate(s) Product(s) Organic Biocatalyst Reference Solvent(s) olive oil fatty acids, diisopropylether Rhizopus Bell et al. triglycerides arrhizus (1981) n-benzyloxyl aspartame ethylacetate thermolysin Hiroshi et al. carbony 1-L-aspartic precursor (1985) acid, L-phenylalanine- methylester castanospermine, o-acyl tetrahydrofuran subtilisin Margolin et al. acylating agents derivatives, (1990) castanospermine isoamylalcohol, isoamylacetate n-hexane immobilized Rizzi et al. ethylacetate lipase (1992) racemic fatty acids optical active ether, heptane yeast lipase, · Kirchner et al. esters porcine lipase (1985) racemic alcohols, optical active tetra-butyl subtilisin, Margolin et al. amines (meth)acrylate methyl ester, lipase (1991) ester 3-methyl- 3-pentanol. octanoic acid n-octane n-hexane recombinant Favre-Bulle E. coli (alk+) et al. (1993) amino acid polypeptide toluene porcine Margolin& derivatives pancreatic Klibanov lipase (1987)

1.1.3.5 Biotransformation in reverse micelles

A reverse micelle medium is formed by dissolving amphiphilic molecules in apolar solvents. It consists of tiny water droplets stabilized by surfactant in a bulk water immiscible organic solvent. An enzyme molecule is confined in a reverse micelle, which is solubilized in a water-insoluble aliphatic hydrocarbon. In reverse micellar media, enzymes can interact easily with water-insoluble substrates as schematically presented in Figure l.2(a). In the formation of reverse micelles, the molar ratio of the water solubilized within the reverse micelles to the amount of the surfactant Chapter 1 14

present, which is usually denoted by Wo, is an important parameter. Wo is an indication of the number of water molecules in each reversed micelle. Figure 1.2(b) is a typical profile of the effect of Wo on the reaction rate (Barbaric and Luisi, 1981). The reaction rates as a function of Wo show very similar, sharp peaks at which the reaction rates are maximum. It is interesting that the reaction rates decrease with increase in Wo above the critical Wo, although the reason remains unclear. The use of reverse micelle media has been focussed mostly on lipase catalysed reactions where these reactions can be carried out in a 'pseudo one-phase' system. This specific system has significant potential for the processing of fats and oils. This system results in greater interfacial area, lower mass transfer limitation, enhancement of catalytic activity and higher resistance of thermal deactivation in the micelle (Prazeres et al., 1993). Several examples of biotransformations in reverse micelle phase are summarized in Table 1.6.

Table 1.6 Biotransformation products of reverse micellar phases

Substrate(s) Product(s) Phase Biocatalyst Reference solvent(s)

peptide amino acids toluene lipase Margolin& Klibanov (1987)

olive oil fatty acids dioctylsodium lipase B Prazeres et al. sulfosuccinate (1993) /isooctane

cholesterol cholestenone phosphatidyl- Rhodococcus Goetschel et al. choline erythropolis (1992)

propionic acid, oleyl isooctane lipase Schlatmann oleyl alcohol propionate et al. (1991)

cholesterol, oleate cholesterol- bis(ethylhexyl) lipase Hedstrom et al. oleate (ester) sulfosuccinate (1992) /isooctane Chapter 1 15

ORGANIC SOLVENT • (a)

(b) 1.5

N 0

0 UJ ...... E > 0.5

5 10 15

Figure 1.2 (a) Schematic representation of the interaction of substrate molecules (S) distributed in the reverse micelle system with entrapped hydrophilic (El), substrate-active (E2) and hydrophobic (E3) enzymes (Martinek et al., 1981); (b) Ratio of initial reaction rate to enzyme concentration in reverse micelles (for the a-chymotrypsin-ca tal ysed hydrolysis of N-glutaryl-L-phenyl alanine-p-nitroanilide) (Barbaric and Luisi, 1981). Chapter 1 16

1.1.3.6 Biotransformation in supercritical fluids

Supercritical fluids which are substances existing above their critical temperatures and pressures, possess physical properties intermediate between a gas and a liquid. Physical properties such as dielectric constant and hydrophobicity of supercritical fluids can be regulated by changing temperature and/ or pressure. The use of enzymes in supercritical fluids has been proposed as a means of improving the activity and utility of enzymes in anhydrous environments. The attractive properties of this system are: (i) relatively high diffusivities which facilitate heterogeneous reactions between enzyme and substrate, (ii) high dependence of solubility on pressure which permits good substrate solubility, (iii) enhanced product recovery, (iv) low toxicity and environmental impact and (v) use of relatively mild temperatures. Previous studies on enzymes in supercritical fluids have been limited to carbon dioxide as a solvent because of a lack of other fluids with critical temperatures and pressures low enough to support enzyme catalysis (Nakamura et al., 1986). Recently, Kamat et al. (1992) demonstrated that supercritical carbon dioxide is not a suitable solvent for the esterase­ catalyzed transesterification of acetylates. The supercritical carbon dioxide may either alter the pH of the microaqueous environment associated with the protein, or form a reversibly covalent complex with a free amine group on the surface of the enzyme. Thus, many other supercritical fluids (ethane, ethylene, sulphur hexafluoride and fluoroform) have been examined and showed better characteristics than carbon dioxide. These authors found that with supercritical ethane, it was possible to control the activity of the enzyme by changing pressure, and the enzyme appeared to follow Michaelis-Menten kinetics. Furthermore, sulphur hexafluoride, the first anhydrous inorganic solvent, was proposed also as a good candidate for supercritical organic solvent. Current results of biotransformations in the supercritical phase are showed in Table 1.7. Chapter 1 17

Table 1.7 Biotransformation products of supercritical phase

Substrates Product Solvent(s) Biocatalyst Reference oleic acid, ethanol ethyloleate supercritical C02, immobilized Marty et al. n-hexane lipase (1992)

2-ethylhexanol, 2-ethylhexyl supercritical ethane lipase Kamat et al. methylmethacrylate methacrylate or (1992) sulphurhexafluorde Chapter 1 18

1.2 L-EPHEDRINE

1.2.1 Introduction

L-ephedrine is a natural plant alkaloid isolated from the dried young branches of Ephedra, a plant with an interesting pharmacological activity. Extracts of Ephedra sp., particularly Ephedra sinica, E. equisetina and E. distachya commonly called wild herbs 'Ma Huang' in China, have been used as folk remedies for inducing sweat, soothing breath and easing excretion of urine in the Orient and Russia for more than several thousand years. The active ingredient of the these extracts, L-ephedrine, was first isolated by Yamanaski in 1885. Since international interest in this drug was stimulated by the classical investigations of Chen and Schmidt in 1930, who reported its cardiovascular effects and its similarity to epinephrine, pharmacological research on L-ephedrine has significantly extended its medicinal activities (Hu, 1969). L-ephedrine, known chemically as 1-methylamino-ethyl-benzyl alcohol, or 2-methylamino-1-phenyl-1-propanol, contains two asymmetric carbon atoms, so that there are four optically active forms and two racemic mixtures, of which the (-) isomers are usually used clinically. These four isomers shown in Figure 1.3 occur naturally in the Ephedra plants, and commonly can be extracted with alcohol and benzene. Purified ephedrine is obtained as odourless, colourless crystals or as a white crystalline power with bitter taste (S0resen and Spenser, 1988). Ephedrine and pseudoephedrine are very stable. A solution of ephedrine hydrochloride sealed for six years showed neither oxidation nor loss of activity. Currently, Germany and China are the major suppliers to the world markets for the commercial production of L-ephedrine. China produces this chemical from the herbal plants, while Germany uses chemical and biological synthesis. ICI Australia P /L in 1988 planned to establish a semi­ biological process based on the biotransformation of benzaldehyde to L­ phenylacetylcarbinol (L-PAC) which is a key intermediate in the synthesis of L-ephedrine, using yeast Candida utilis. The ICI fed-batch process for the production of L-PAC involved a two stage approach: growth of yeast (Candida utilis) on a fermentable sugar (growth phase), and feeding of Chapter 1 19

benzaldehyde into culture medium according to a pre-programmed flow rate (biotransformation phase). A project designed to optimise the production of L-P AC was initiated in Department of Biotechnology, the University of New South Wales and ICI Australia P /Lin 1989.

OH

..•,,. ~ ~NH2

(- )norephedrine ( +)norpseudoephedri ne

OH H OH

..,,. ~ NHCH3

( -)ephedrine ( +)pseudoephedri ne

Figure 1.3 The structural formulae for ephedrine and its analogues (S0rensen and Spenser, 1988) Chapter 1 20

1.2.2 Medicinal uses of L-ephedrine

Ephedra plants containing L-ephedrine have been used by the Chinese for over 5000 years. They made a tea of twigs to allay fever and coughs, and to increase blood pressure. Since the mydriatic activity of the isolated L-ephedrine was first demonstrated in 1887 by Miura, the intrinsic alkaloid, L-ephedrine, has been a valuable multipurpose drug and an ingredient in many preparations (Mark et al., 1963). L-ephedrine is a sympathomimetic agent with direct and/ or indirect effects on adrenergic receptors like adrenalin. It acts indirectly by releasing neuro-transmitter from storage sites in the sympathetic nerves to the effector organ. L-ephedrine has been used for treatment of allergic conditions (hay fever), prophylaxis, hypotension and can cause mild stimulation of the central nervous system (Dollery, 1991). L-ephedrine increases arterial blood pressure in man both by peripheral vasoconstriction and by cardiac stimulation. Heart rates and cardiac output are increased. It prevents the fall in blood pressure which occurs after spinal anaesthesia in surgery (Bhagwanjee et al., 1990), and has been used for initial treatment of hypotension which could later cause haemorrhagic shock due to bleeding (Eldor, 1991). L-ephedrine has been shown to stimulate oxygen uptake and thermogenesis in man (Astrup et al., 1984), and has been used as an ergogenic agent in athletic competition (Smith and Perry, 1992). Some reports indicate that L-ephedrine may be a potential drug of abuse, and can produce symptoms of schizophrenia similar to those found in amphetamine psychosis. Chronic ephedrine over-dosage may result in either severe cardiac toxicity or psychosis (American Pharmaceutical Association, 1986). The most generally appreciated virtue of L-ephedrine is in causing a significant increase in specific airway conductance so that as a bronchodila tor, it can provide effective relief of nasal congestion and is commonly used for the treatment of mild to moderate seasonal or chronic asthma with obstructed airways. In view of the current interest in the pharmacological stimulation of metabolic rate, L-ephedrine has been used to assist the management of obesity with low calorie regimens. Walker (1990) reported that L-ephedrine Chapter 1 21

caused anorexia in rats so that food-intake was significantly decreased. This finding agreed well with other investigations on anorexic activity of a mixture of L-tyrosine and L-ephedrine (Hull and Maher, 1991). Jonderko and Kucio (1991) also demonstrated that L-ephedrine could be an anti­ obesity drug by promoting energy expenditure. They observed a significant delay in gastric emptying after administration of L-ephedrine, and concluded that the inhibitory influence of L-ephedrine on gastric emptying could make it a candidate for trial as pharmacological support for a low energy diet treatment of obesity. In clinical reports, Pasquali et al. (1987, 1992) reported that L­ ephedrine contributed to weight loss in overweight and obese women who had found difficulty in losing weight with conventional hypocaloric treatment, and they found that L-ephedrine partially prevented the fall in resting metabolic rate and significantly improved the nitrogen balance. These effects may be of importance in the treatment of patients in whom a reduced capacity for energy expenditure may be involved in their obese state. In the study of synergistic effects on anti-obesity, Dulloo and Miller (1987) reported the effects of both aspirin and L-ephedrine on obesity. L­ ephedrine increased energy expenditure, and reduced body weight and body fat significantly without reversing the obesity. In the presence of both ephedrine and aspirin, these effects were much greater than ephedrine alone. Furthermore, they reported that the thermogenetic effect of ephedrine could involve minor adenosine antagonism, but was mainly explained by the inhibition of phosphodiesterase activity (Dulloo et al., 1992). Similarly, Lowell et al. (1990) reported that a mixture of L-ephedrine and caffeine in murine greatly diminished adipsin gene expression in certain forms of genetic and acquired obesity. Further, Astrup et al. (1992a) also investigated whether a weight reducing synergism between ephedrine and caffeine was present in obese patients. The ephedrine/ caffeine combination was found to be strongly effective, while caffeine and ephedrine separately were either not effective or less effective for the treatment of human obesity. Additionally, these authors reported that treatment with caffeine/ephedrine promoted fat loss by stimulation of lipolysis, and preserved the fat free mass during weight reduction (Astrup et al., 1992b). Chapter 1 22

L-ephedrine is metabolised in the liver with both the parent drug and metabolites being excreted in the urine (half life 3-11 h). Up to 95 % of the dose can be recovered in 24 hours, with 55-75 % as the unchanged drug (Dollery, 1991). L-ephedrine is metabolised to norephedrine as a major metabolite by N-demethylation, which has pharmacological activity, producing a central stimulant effect. L-ephedrine is also deaminated yielding benzoic acid, hippuric acid and L-phenylpropan-1,2-diol. Metabolism of ephedrine is summarised in Figure 1.4.

cu-cu-r{ 8 O I I CH3 OH CH3 Ephedrine

Unchang~ i ~ Deaminated drug in urine 0 metabolite (4-13 % ) (55%-75%) \ .i -qi-yH-NH2 OH CH3

Norephedrine(8-20%)

Figure 1.4 Metabolism of Ephedrine

1.2.3 Production of Ephedrine

Currently, three main ways of preparing ephedrine are used: traditional extraction from plant species of Ephedra, chemical synthesis and semi-biological synthesis which combined with biotransformation for a key intermediate is followed by chemical synthesis for the final products. Chapter 1 23

1.2.3.1 Ephedrine from natural herbs 'Ma Huang'

1. Sources of ephedrine

Ephedra plants called 'Ma Huang' are classified as the division Gnetophyta, family Ephedraceae, genus Ephedra. The seemingly leafless Ephedra species grow in tundra-like regions, and are widespread in China, India, Mongolia, Pakistan, Tibet and the northwest Himalayas. The pale green, grooved stems are erect and 30-90 cm in height. They are articulated, the length of the internode being 3-5 cm. The scale-like leaves growing at the nodes are 2-4 cm long and toothed at the top as shown in Figure 1.5 (Thomson, 1978). The inconspicuous yellow­ green flowers appear in terminal catkins; male and female flowers are separated. Five species of Ephedra, E. major, E. gerardiana, E. intermedia, E. sinica and E. equisetina are considered to contain significant amount of L­ ephedrine. These species have been experimentally cultivated in Australia, Kenya, England and United States. Most efforts were successful, but the crop proved to be economically unfeasible because of high labour costs. Commercial cultivation has been contemplated in India (Morton, 1977). The woody basal stems of the Ephedra plant are very low in alkaloid content; roots and fruits are nearly alkaloid-free. The green branches and twigs of the cited Mediterranean and Asiatic species contain the alkaloids ephedrine and pseudoephedrine in varying amounts depending on the plant species, the altitude at which it grows, the harvest time and weather. E. major may contain over 2.5 % total alkaloids, nearly 75 % of total being ephedrine. E. intermedia is low in ephedrine but fairly high in pseudoephedrine. Harvesting which begins when the plants reach four years of age takes place in autumn, during the blooming season, when the alkaloid content is highest. Throughout the summer rains, alkaloid content declines, then it gradually increases until it is double that of spring time (Morton, 1977). In plants of the ephedra, ephedrine is usually found together with norephedrine, pseudoephedrine, norpseudoephedrine (Hu, 1969). Ephedra sp. have a strong pine odour and very astringent taste. Chapter 1 24

, 1978) sinica (Thomson ture of Ephedra Figure 1.5 Pic Chapter 1 25

For medical preparations, all stems are cut to less than 1/2 inch in diameter. These are dried in the sun for 15 days to reduce the volume by 50- 60 %. After drying, the stems are beaten to break off the joints and then screened to separate the unwanted joints and stored in a dry atmosphere. Exposure to humidity during storage can result in complete loss of alkaloids. Traditionally, the Chinese used an aqueous reflux method of extraction to obtain ephedrine from plants for medical preparations (Hu, 1969). The procedure for extraction and isolation of ephedrine developed by Yamasaki et al. (1973) is summarised in Figure 1.6. More recently, S0rensen and Spenser (1988) successfully carried out the purification and isolation of ephedrine, pseudoephedrine and their isomers from crushed dried plant materials. This purification involved Soxhlet extraction with methanol followed by acid/base partition, treatment with acetone/silica gel, liquid chromatography and recrystallization.

2. Biosynthesis of ephedrine in plants

Investigation of the biosynthesis of ephedrine in plants was initially carried out. It was originally thought that the amino-phenyl propanoid system of ephedrine was derived directly from phenylalanine or by reaction of phenylalanine-derived phenethylamine with a formate-derived one carbon unit. This view had to be abandoned when it was found that label from [14C2J phenylalanine did not enter ephedrine (Yamasaki, et al. 1973). It was shown that ephedrine was synthesised by the condensation of a C6-Ct portion which is derived from phenylalanine via cinnamate. Feeding of benzoic acid, benzaldehyde to Ephedra distschya resulted in efficient incorporation of 14C into the a-carbon atom of the side chain of ephedrine. Thus, the relatively low incorporation of phenylalanine and cinnamic acid suggested that an alternative route produced benzoic acid or benzaldehyde from shikimate, and the remaining part of C2-N of ephedrine was derived from aspartate or related compounds, formate and methionine. However, S0rensen and Spenser (1989) found through NMR investigation that the carbon skeleton of ephedrine bases did not appear to be derived directly from an amino acid. Chapter 1 26

Dried Ephedra plant at 60°C ! Extracted three times with methanol

Removed m!.hanol through the evaporation

Dissolved JI% H2S04 ! Acidic and neutral substances discarded by filtration and extration with diethylether. ! Alkalinized to pH 11 with K2C03 ! Extracted with diethylether i Washed with H20 i Dried with MgS04 i Oily basic compound obtained after evaporating residual dietrlelher

Dissolved in H20 and neutralized with diluted HCI (crude ephedrine HCI) I Recrystallized with ethanol/diethylether l Pure ephedrine HCI

Figure 1.6 Extraction and purification of ephedrine from Ephedra plant (Yamasaki et al., 1973) Chapter 1 27

They concluded that the carbon skeletons of the Ephedra alkaloids were generated from two fragments, a C6-C1 unit related to benzaldehyde or benzoic acid, and a CH3CO-moiety derived from pyruvate based decarboxylation by pyruvate carboxylase. Further, the stereospecific transfer of the amino group of L-glutamic acid into (+ )-1-hydroxy-1-phenyl-2-propanone (i.e. phenylacetylcarbinol) to yield a mixture of (-)-norephedrine as a major and ( + )-norephedrine as a minor product, was catalysed by transaminase at a late stage. Plant tissue cultures using callus tissue also showed a significant potential to produce ephedrine. Studies with Ephedra gerardiana callus tissues showed that indole butyric acid (IBA) was the best auxin amongst growth regulators for ephedrine production (Ramawat and Arya, 1979a). The appearance of ephedrine in the late phase of growth suggested that synthesis of these compounds took place after the exponential phase was over. A synergistic effect of IBA and some amino acids on alkaloid production was observed with E. gerardiana tissue (Ramawat and Arya, 1979b). They found that phenylalanine, methionine, and glycine (0.1-0.4 g/L) increased the alkaloid yield even though these amino acids were not considered to be direct precursors of ephedrine. The highest yields, 0.6 % ephedrine (based on dry callus) was obtained with synthetic medium containing phenylalanine, IBA and kinetin after 8 weeks growth with 1000 lux of light intensity.

1.2.3.2 Chemical synthesis of ephedrine

Chemical synthesis of ephedrine occurs in three steps: preparation of primary substrate, condensation of methylamine and resolution of racemic mixture. These procedures are summarised in Figure 1.7.

1. Preparation of 1-phenyl propane-1,2-dione (Coles et al., 1929)

The synthesis of ephedrine necessitated the preparation of considerable quantities of 1-phenylpropane-1,2-dione. This substance was prepared by the decomposition of isonitrosoethylphenyl ketone by means of Chapter 1 28

dilute sulphuric acid or amyl nitrate. In the first case, decomposition was slow, and in the second case, by-products were formed in large amounts so that the yield of diketone in either case was small. Therefore, it was proposed that the oxidation of the methylene group of ethyl phenyl ketone be accomplished by nitrogen tetroxide. The reactivity conferred on the methyl atom by the phenyl and carbonyl groups provides sufficient stability to the resulting diketone to prevent complete disintegration by the further action of the oxidant, although it was not possible to prevent the formation of benzoic acid in considerable quantities. The yield of pure diketone was greater than that by any other feasible method.

2. Condensation of l-phenyl-1,2-propanedion with methylamine (Manske and Johnson, 1929a)

It was observed that a solution of 1-phenyl-1,2-propanedion (C6HsCOCOCH3) in petroleum ether reacts exothermally with dried gaseous methylamine to form the intermediate C6HsCOC=(NCH3)CH3, with the elimination of water. This substance was sequentially reduced to ephedrine, a colourless crystalline compound. Practically, a mixture of absolute ethanol, methyl phenyl diketone and an alcoholic solution of methylamine was reduced catalytically with hydrogen in the presence of platinum oxide. When reduction no longer proceeded, the catalyst was removed by filtration and about half the alcohol was removed under reduced pressure. The solution was made acidic with alcoholic hydrogen chloride and evaporated to dryness. The solid hydrochloride was washed with cold acetone and dried. The D,L-ephedrine hydrochloride was purified by recrystallising once with alcohol-acetone and melted at 189°C.

3. Resolution of racemic mixtures (Manske and Johnson, 1929b)

Inactive ephedrine readily forms crystalline salts with many optically active acids, among which are tartaric acid, D-camphor sulphonic acid and mandelic acids. No resolution can be effected with the first two although the use of D-camphoric acid yielded a small amount of crystals which contained only D,L-ephedrine. 29 Chapter 1

0 0 II II c- u-CH3 0 C- CH2-CH3 0 1 N02 Ethylphenyl ketone lsonitroso-ethylphenyl ketone

Oxidation of {{-methyl group ~ Lomp<~itioo by di!"'od solph"

1-phenyl-1.2-propanedion Rcadion with dried gaseous mcthylammc m absolute ! ethanol 0 II 0 C- YI- CH3 N I I CH3 f RcJuclJon by H~. Pt OH

Q-bH-rH-CH3 NH I with D.L-EphedrineCHJ Rccr)slalli1ation Rccnslallual1on\\1lh L . ~ _- ... -mandchcaciJin ~ -mandclu.: ,tuJ in 9,;,7, h· 1 Rcsolu11on by - ' cl ano L-ac1d-L-basc,and 95'7, ethanol D-ac1d-D-basc OH combinal1on H

L-ephedrine Pseudoephed ri ne

and Figure 1.7 Chemical synthesis of Ephedrine (Coles et al., 1929; Manske Johnson, 1929a, b) Chapter 1 30

However, the mandelic acids are very well suited to the purpose, the D-acid-O-base and L-acid-L-base compounds being far less soluble in 95 % ethanol than the other possible combinations. Three recrystalizations sufficed to obtain optically pure products and the yields were high.

1.2.3.3 Ephedrine synthesis via biotransformation of benzaldehyde to L-phenylacetylcarbinol (L-PAC)

L-phenylacetylcarbinol (L-P AC) is a key intermediate in the synthesis of L-ephedrine. The first L-P AC was produced microbiologically by Neuberg and co-workers in 1921. It was found that benzaldehyde added to yeast cells in the presence of fermentable sugar could be transformed to L-P AC. The production of L-PAC by biotransformation of benzaldehyde is generally ascribed to pyruvate decarboxylase (PDC). The properties of the yeast pyruvate decarboxylase have been studied with regard to the condensation reaction via the enzyme-bound intermediate 2-a.-hydroxylethylthiamine pyrophosphate (HETPP). In the presence of pyruvate and an aldehyde, the enzyme catalyses the condensation of the C-C bond via an acyloin reaction in which free aldehyde competes with a proton for bond formation with a carbanion of HETPP. This results in the formation of acetoin when acetaldehyde accumulates or is added to the reaction mixture. With benzaldehyde as the aldehyde component, L-PAC is synthesised enantioselectively. However, this process did not gain industrial significance until the competing chemical synthesis of L-ephedrine from L­ p AC was patented (Hilldebrandt and Klavehn, 1934). This former process proceeds in two steps: the first step is to transform benzaldehyde to L-P AC by Saccharomyces cerevisiae and the second step is to convert L-P AC to ephedrine by reductive amination with methylamine as summarized in Figure 1.8. Most of the research effort towards improvement of L-PAC production has concentrated on the microbiological transformation of benzaldehyde using various yeasts of which the most important is S. cerevisiae. Chapter 1 31

oo Pyruvate decarboxylase 11 II (EC 4. l. l. I) CHrC-C-OH Oc;:+ bi otransformati on Benzaldehyde Pyruvate

;=\_OH chemical synthesis ~i-iuCH3 I CH3 L-phenylacetylcarbinol D.L-ephedrine

Figure 1.8 Synthesis of L-ephedrine via biotransformation.

The importance of S. cerevisiae is largely due to its capacity to both withstand relatively high concentrations of benzaldehyde, and produce higher concentration of L-PAC. Considerable success has been achieved in the use of S. cerevisiae in the production of L-PAC. However, a relatively low conversion yield (viz. 60-70 % theoretical) of benzaldehyde to L-PAC has been serious drawback in the commercial application of this system. The reasons for this low conversion yield have been attributed to the side reaction of benzaldehyde to benzyl alcohol and the high toxicity of benzaldehyde (substrate) on the free cells. The chemical reaction step to catalyse the reductive amination of L­ PAC to ephedrine was significantly improved by Nebesky et al. (1978). L­ ephedrine hydrochloride was prepared from L-PAC by the following Chapter 1 32

procedure: active carbon was mixed with fermentation broth and suspended solids were separated without any heating by centrifugation. The supernatant solution was filtered through a layer of 'Supercell'. The clear filtrate was extracted with butyl acetate. The extracted L-PAC was concentrated under vacuum in an evaporator with circulation at the temperature not exceeding 60°C. L-P AC was then subjected to the reductive amination. L-P AC can be hydrogenated at 50-55°C under a pressure of 202.6 kPa of hydrogen gas atmosphere, in the presence of a platinum catalyst. Aqueous 35-40 % methylamine solution is added portion by portion in fed­ batch fashion to maintain the reaction. The product is a racemic mixture of D, L-ephedrine, which can be separated by differential crystallisation in 88 % aqueous hydrochloric acid. For optimization of this process, surfactant 'Slovefol 909', and p-methylbenzene sulphonate as an initiator can be added to decrease the surface tension and initiate the hydrogenation reaction at such a rate that a yield of 69.5 kg of ephedrine-HCl can be obtained from 77.8 kg of phenylacetykarbinol within 5 h. This sequential procedure is summarised in Figure 1.9. Chapter 1 33

Biotransformation broth I Mixed with active carbon I Suspended solid separated by centrifugation I Filtered through 'Supercell' membrane I Extracted with butylacetate j Concentrated by vacuum evaporation at 60°C I Hydrogenated by 202.6 kPa H2 gas atmosphere at 50-55°C in presence of Pt, Slovefol 909, and p-methyl benzene sulphonate l Reacted with 35-45% methylamine l Obtained racemic mixture of D,L-ephedrine ! Isomers separated by differential crystallisation in 88% aqueous HCl l L-ephedrine

Figure 1.9 Summary of L-ephedrine production via semi-biological synthesis (Nebesky et al. 1978) Chapter 1 34

1.3. BIOTRANSFORMATION OF BENZALDEHYDE TO L-PHENYLACETYLCARBINOL (L-PAC)

1.3.1 Reaction mechanism of pyruvate decarboxylase for acyloin formation

Since Neuberg and Hirsch (1921) suggested that a special enzyme, 'carboligase' could bring about the formation of ketols in yeast, many studies on enzymatic reactions for acetoin and other ketol synthesis, followed by decarboxylation of keto acids, have been conducted. Neuberg and Hirsch (1921) proposed the condensation of acetaldehyde or benzaldehyde with a 'nascent' aldehyde to yield the corresponding ketol. Isotopic carbon studies by Gross and Werkman (1947) and by Juni (1952) substantiated that a condensation of this general type occurs. It was suggested that yeast preparations were capable of condensing aldehyde with pyruvic acid to yield ketols and C02 according to a following equation:

CH3COCOOH + RCHO

It was recognized that two possible ketols, CH3COCHOHR and CH3CHOHCOR, which would give the same osazone, might be formed. Gross and Werkman (1947) found that the addition of 13C labelled acetaldehyde to a cell free yeast extract in presence of pyruvate resulted in the labelling of all carbon atoms of the acetoin formed. The heaviest labelling occurred in the carbinol end of molecule. Juni (1952) showed that the addition of pyruvic acid-14C2 and unlabelled acetaldehyde to a cell free enzyme preparation of brewers' yeast yielded acetoin with the heaviest labelling at the carbonyl portion of acetoin. In both cases, the presence of labelled carbon at all carbon atoms could be accounted for by the oxidation or reduction of some of the acetoin to a symmetrical molecule. Kluyver and Donker (1924) reported that the acetoin was not formed from glucose alone due to the rapid removal of acetaldehyde by reduction using endogenous NADH. Although acetylmethylcarbinol (acetoin) is rarely formed during the normal fermentation of glucose by yeast, addition of acetaldehyde (or a hydrogen acceptor) causes its formation. Chapter 1 35

The comparison of a bacterial system with non-bacterial systems (yeasts and other eucaryotes) indicated that there are several different mechanisms for the formation of acetoin. Bacteria, such as Aerobacter aerogenes (formerly Klebsiella aerogenes), do not utilize acetaldehyde for the formation of acetoin (Juni, 1952). In this case, it was demonstrated that a-acetolactic acid [CH3COH(COCH3)COOH] was an intermediate in the formation of acetoin from pyruvic acid. However, a-acetolactic acid is not an intermediate in the formation of acetoin by yeast or animal cells. While the yeast enzyme always requires at least 'nascent' aldehyde resulting from decarboxylation of pyruvate, a mechanism which can form acetoin from acetaldehyde alone is present in the animal tissue. With yeast pyruvate decarboxylase, decarboxylation proceeds by the action of the TPP C-2 carbanion on the a-keto group of the pyruvate molecule at the active site. The initial product is the stabilised anion of hydroxyethyl-TPP (HETPP), which then dissociates from the initially added coenzyme carbanion.

Juni (1961) proposed a two-site mechanism to explain the formation of free aldehyde and acetoin from a-keto acids by decarboxylase. At the first site, pyruvate is decarboxylated to an aldehyde-thiamine pyrophosphate complex. The aldehyde moiety is then irreversibly transferred to the second site, where reversible dissociation to free aldehyde take place. It is postulated that acetoin synthesis involves condensation of free acetaldehyde with acetaldehyde at the first site, independent of whether the second site is or not blocked with acetaldehyde. The mechanism for biotransformation of benzaldehyde to L-PAC by yeast was elucidated by Smith and Hendlin (1953; 1954). They explained that the synthesis of L-P AC, accompanied with a concomitant reduction of a portion of benzaldehyde to benzyl alcohol, was closely associated with NADH and a coenzyme A dependent system. Thus, acetyl-CoA from pyruvate might be involved in the acetylation of benzaldehyde to L-P AC. But later, Hane and Kakac (1956) clearly showed that the formation of L­ P AC from benzaldehyde required the presence of pyruvic acid and yeast carboxylase. The enzyme catalysed not only the decarboxylation of pyruvic acid, but also the carboligase reaction, by which the acetyl residue is bound to benzaldehyde for the formation of L-P AC. Chapter 1 36

1.3.2 Screening of microorganisms for L-PAC

The efficiencies of L-PAC formation by various microorganisms were found to be dependent on the level of pyruvate decarboxylase, its affinity for the substrate (e.g. benzaldehyde), tolerance to inhibitory substrate and products and by-product formation (e.g. benzyl alcohol). Most efforts in screening of microorganisms to improve levels of L-P AC have been focussed on high ethanol production capacity attributed to high PDC activity.

Comparison of PDC activity and affinity for substrates, pyruvate, and benzaldehyde with Zymomonas mobilis and Saccharomyces carlsbergensis has been conducted by Bringer-Meyer and Sahm (1988). Although Z. mobilis showed five times higher PDC activity than S. carlsbergensis, Z. mobilis produced 4-5 times less L-PAC than yeast. Studies with 2-p-toluidino naphthalene-6-sulphonate and other substrate analogues showed that the catalytic sites of PDC of Z. mobilis were less lipophilic than that of PDC of yeast (Ullrich and Donner, 1970). This difference could explain the lower affinity of the Z. mobilis PDC toward benzaldehyde, leading to lower L-P AC formation. Because of the limited solubility of benzaldehyde (0.3 g/100 mL water), PDC is never likely to be saturated with benzaldehyde for L-PAC formation.

The tolerance to benzaldehyde of 6 species of yeasts, Hansenula an o ma la, Brettanomyces vini, Saccharomyces carlsbergensis, Saccharomyces cerevisiae, Saccharomyces ellipsoides and Candida utilis were measured after a single addition (2 g/L) and after 4 additions of benzaldehyde (Becvarova and Hane, 1963). The inhibition effects were different for the different yeasts. For Hansenula anomala, benzaldehyde caused less inhibition of fermentation (16 %), while for other cultures, the inhibition was more significant (35.5-63.2 %). After single addition of benzaldehyde, 46-52 % benzaldehyde was converted to L-PAC by H. anomala, S. cerevisiae and S. carlsbergensis. During biotransformation with 4 additions of benzaldehyde, 70 % benzaldehyde was converted to L-P AC by S. carlsbergensis, while the rest in large part was reduced to benzyl alcohol. Chapter 1 37

In terms of adaptation to benzaldehyde and L-P AC formation, eight yeast strains, viz. S. cerevisiae (CBS 1171, NCYC 324), S. carlsbergensis, S. fragilis, S. rouxi, S. lactis, S. veronae and S. microellipsoides were tested for their ability of growth and L-P AC formation in the presence of various concentrations of benzaldehyde (Gupta et al., 1979). Most of the yeasts which showed growth in the presence of benzaldehyde (0.5 g/L) were further adapted to high concentrations. Although growth of yeasts were decreased with increasing benzaldehyde concentration, after an adaptation period, growth of S. cerevisiae (CBS 1171) was gradually restored to more than 50 % of that in the absence of benzaldehyde. Under these concentrations, 5.2 g/L of L-PAC was obtained in 8 h incubation with S. cerevisiae CBS 1171.

The relation between initial reaction rate and final production of L­ p AC was determined for 38 species of yeasts- most from the genera Saccharomyces, Candida and Pichia (Netrval and Vojtisek, 1982). In many strains, the biotransformation, having started quickly, suddenly ceased when the L-PAC reached 3-4 g/L. The results showed that a relatively high initial rate of L-P AC did not necessarily indicate that a given strain would produce a large amount of L-PAC. The highest L-PAC production (6.3 g/L) was reached with a strain of S. carlsbergensis, while slightly lower concentrations were found for Hansenula sp. (5.9 g/L) and Candida sp. (2.8- 5.6 g/L).

Twelve different microorganisms including yeasts, fungi and bacteria were examined for their ability to perform the acyloin formation and the reduction to corresponding alcohols (Cardillo et al., 1991). Saccharomyces fermentati and S. delbrueckii were found to promote the acyloin condensation with higher yields (56 % and 60 %, respectively, based on benzaldehyde) compared with other previous reports. For all other microorganisms tested, while the condensation activity was low or absent, the reductive activities were relatively high. Aspergillus niger and Hansenula anomala for example produced distinctively higher benzyl alcohol (74 % and 85 %, respectively) and lower L-PAC (14 % and 11 %, respectively). Among bacteria, besides the earlier study on Z. mobilis, L­ p AC formation has been obtained with Bacillus subtilis in limited amounts (9 % of benzaldehyde converted to L-P AC). Chapter 1 38

1.3.3 Factors affecting the biotransformation with yeast cells

The biotransformation of benzaldehyde to L-P AC using yeast cells essentially involves a two stage process; in the first stage, the microorganism is cultivated under optimum growth and fermentation conditions, and in the second stage, biotransformation is performed sequentially with sugar and benzaldehyde feeding. Thus, the physicochemical factors influencing the growth phase, and bioconversion phase can be considered separately.

1.3.3.1 Factors affecting the growth phase

During the growth phase prior to biotransformation, it is necessary to obtain higher concentrations of biomass to achieve better rates of biotransformation, high levels of pyruvate decarboxylase as the key enzyme and significant pyruvate accumulation as a co-substrate. Metabolic rates of yeasts depend on yeast species, carbon sources, physicochemical factors and the culture media. Among these, glucose and oxygen have a significant effect on sugar metabolism. Two primary modes of glucose metabolism in yeast can be defined by reference to the end products. In respiration, biomass and carbon dioxide are the main products, while in oxidoreductive respiration (fermentation), ethanol and C02 are produced in association with cell growth. A shift from aerobic to fermentative conditions has been shown to accelerate the specific rate of glycolysis and cause a fall in biomass yield (Yx/s) from 0.5 to as low as 0.1 g/g (Kappeli, 1986). Yeasts can be grouped according to their response to carbon and oxygen limitation as follows: glucose sensitive yeasts exhibit aerobic ethanol production in the presence of excess glucose, while oxygen sensitive yeasts produce ethanol when oxygen is limiting (Fraleigh et al., 1989). For example, when a Crabtree positive strain, such as S. cerevisiae is grown in aerobic culture containing a high concentration of glucose, the sugar is fermented to ethanol no matter whether or not sufficient oxygen is available. During this phase, aerobic respiration is repressed by the high concentration of glucose. Chapter 1 39

In contrast, Crabtree negative yeasts, such as Candida sp. and Hansenula sp. do not produce ethanol when grown on high glucose media in fully aerobic conditions. They respire under aerobic conditions and give a high yield of biomass. Although higher cell densities have greater potential to achieve higher biotransformation rates in both type yeasts, ethanol production was also found to be necessary as it was associated with increased activity of cytoplasmic pyruvate decarboxylase. Moreover, levels of pyruvate and other glycolytic intermediates are also increased during fermentative growth.

1.3.3.2 Factors affecting the biotransformation phase

(1) Benzaldehyde concentration

Studies on the effect of benzaldehyde have been undertaken with a view to find out its effect on the capacity of microorganisms for growth and L-PAC formation. Although benzaldehyde is used as a substrate with pyruvate, the toxicity of benzaldehyde has a significant effect on cell growth and enzyme activities. Long and Ward (1989b) observed reduction in lipid and protein contents in cells during biotransformation, and they postulated that benzaldehyde alters the cell permeability, and directly and/ or indirectly inactivates enzyme protein. Gupta et al. (1979) measured growth and L-PAC formation with different yeasts species at various concentrations of benzaldehyde. L-P AC formation and tolerance toward benzaldehyde were variable, and found to be dependent on yeast species. All of the yeasts studied were found to grow in the presence of 0.5 g/L benzaldehyde although the amount of growth was variable. S. cerevisiae and S. carlsbergensis showed positive formation of L­ P AC, S. fragilis and S. microellipsoides were doubtful while S. rouxi, S. lactis and S. veronae did not produce L-P AC. None of these organisms showed growth above this concentration of benzaldehyde. Studies on determining the optimum benzaldehyde level for biotransformation demonstrated that while cell viability and enzyme activities were maintained at a low level of benzaldehyde (0.4-0.5 g/L), benzyl alcohol was formed preferentially, and L-P AC production was relatively low. However, at higher levels of benzaldehyde (2-6 g/L Chapter 1 40

dependent on yeasts), L-PAC production was increased although cell viabilities were decreased significantly and finally biotransformation ceased due to pyruvate depletion and PDC deactivation (Long and Ward, 1989b).

After adaptation to low concentrations of benzaldehyde, S. cerevisiae continued to grow and carry out biotransformation at higher concentrations (Gupta et al., 1979). To maximize the L-PAC production, the level of benzaldehyde might be maintained empirically within the range 0.5-5 g/L in order to maximize L-P AC production while maintaining cell viability and PDC activity.

(2) Additional nutrient supplements and available oxygen

For the biotransformation, as pyruvate is required as a substrate, sugar metabolism is an important factor which is affected by available oxygen and residual glucose concentrations (as previously mentioned in section 1.3.3). Thus, supplementary sugar feeding and subsequent pyruvate accumulation are essentially required for L-PAC formation and maintenance of PDC activity. Furthermore in a thiamine deficient in medium, C. utilis significantly accumulated pyruvate in culture broth. If this was followed by addition of thiamine, it significantly enhanced PDC activity, and led to much higher L-PAC formation (Wang, 1993). Additionally with respect to other nutrients, Wiimpelmann and co­ workers (1984) found a correlation between specific ethanol productivity and potassium concentration, which also influenced intracellular pyruvate concentration. The specific ethanol productivity was found to be highly dependent on the ratio of internal to external potassium concentration at constant pH. The increased concentration gradient of potassium was accompanied by a substantial increase in the glucose assimilation rate and in the ethanol production rate. During oxidoreductive metabolism, pyruvate dehydrogenase and pyruvate decarboxylase both compete for pyruvate from glycolysis: the dehydrogenase converts pyruvate to acetyl­ CoA for the TCA cycle, and decarboxylase converts pyruvate to acetaldehyde which is further reduced to ethanol. Based on a limitation of oxidative glucose catabolism, they suggested that the increased pyruvate pool resulting from K+ stress leads to changes in the ratio of activities of these Chapter 1 41

two enzymes. Thus, when pyruvate accumulates, the activity of pyruvate decarboxylase increases accordingly to catalyse the further reduction to ethanol via acetaldehyde. At the same time, alcohol dehydrogenase also increases with the probability of enhanced benzyl alcohol production using endogenous NADH.

(3) Others factors

Smith and Hendlin (1954) reported that the addition of nicotinic acid analogues, 3-acetyl pyridine, nicotine amide and pyrazin amide helped to decrease benzyl alcohol production by 18-22 %. It seems that competition of these analogues with NAD+ at ADH catalytic sites inhibited the reduction of benzaldehyde to benzyl alcohol. Addition of acetaldehyde to the biotransformation medium of S. cerevisiae was found to retard the conversion of benzaldehyde to L-PAC, and to lower the formation of benzyl alcohol. Presumably, the acetaldehyde competitively blocks the reducing system of yeast which causes reduction of benzaldehyde to benzyl alcohol, and increases the yield of L-PAC (Becvarova et al., 1963) The effect of external addition of sodium pyruvate on initial reaction rates and the final production of L-PAC was observed by Vojtisek and Netrval (1982). Initial reaction rates were not influenced by addition of pyruvate, but overall L-PAC production was significantly increased. When pyruvate and benzaldehyde were present at sufficient concentrations (or in excess), the rate of L-PAC production depended on the pyruvate decarboxylase activity. However, when pyruvate concentration was not in excess, the reaction was not zero order with respect to pyruvate, but its rate depended directly on pyruvate concentration.

1.3.4 Kinetics of Biotransformation

Most biotransformations have been carried out in batch culture, with external feeding of benzaldehyde, although there are a few reports of semi­ continuous and continuous processes using immobilized cells. Chapter 1 42

Batch processes have been studied by a number of authors since the discovery of L-P AC production. A major problem encountered in the biotransformation of benzaldehyde to L-P AC is possible toxic effects of substrate and/ or products on cells and their constituents including cell membranes and intracellular enzymes (Long and Ward, 1989b). Although relatively high initial rates of L-PAC formation were observed in biotransformations containing a high starting benzaldehyde level, a significant reduction in viable cells was observed resulting in early cessation of L-P AC formation. Thus, pulse feeding to maintain lower benzaldehyde concentrations resulted in lower initial reaction rates, but it prolonged yeast viability and extended the biotransformation. This resulted in improving overall production of L-PAC. Finally, yeast growth and viability were reduced, and PDC was significantly inhibited by continuous contact with toxic substrate and products. This caused accumulation of benzaldehyde in the culture broth with further inhibition and final cessation of the biotransformation.

Two step continuous feeding of benzaldehyde in a batch process with Candida utilis has been conducted by Wijono (1991). In the early biotransformation stage, C. utilis was exposed to a low concentration of benzaldehyde (viz. 0.8 g/L) prior to an increased rate of biotansformation in which benzaldehyde was added continuously with slow feeding (0.45-0.65 g/L/h) for about 4-5 h. Finally, fast feeding with benzaldehyde increased to about 1.3-1.6 g/L/h was implemented until the end of biotransformation. L-PAC concentrations as high as 16.5 g/L were achieved with 8.3 g/L benzyl alcohol within 17 h. In further work, Wang (1993) successfully performed a biotransformation by recognizing the fact that C. utilis can synthesis its own growth factors in vitamin-free fully-defined synthetic medium. The use of a vitamin-free synthetic medium for C. utilis in the growth phase resulted in the pyruvate decarboxylase activity being maintained at a low level and gave rise to increased accumulation of pyruvate under slightly fermentative conditions (respiratory quotient, RQ=2-4). When pyruvic acid reached a relatively high concentration (about 15-20 g/L), the addition of vitamins, particularly thiamine, activated the PDC. The increased levels of both pyruvic acid and PDC activity resulted in increasing final L-PAC concentrations over 22.2 g/L. The literature results of previous L-PAC production are summarised in Table 1.8. Chapter 1 43

Table 1.8 Comparison of biotransformation of benzaldehyde to L-PAC by various yeasts

Biotransfonnation Microorganism Carbon Max. L-PAC Reference process source(s) (g/L) Fed-batch with C. utilis glucose 22.2 Wang (1993) free cells 16.5 Wijono (1991) S. cerevisiae sucrose 6.0 Becvarova et al. (1963) S. carlsbergensis sucrose 8.0 Vojtisek and Netrval (1982) sucrose+ S. 10.5 Vojtisek and carlsbergensis sodium- pyruvate Netrval (1982) S. coreanus sucrose 12.4 Culic et al. (1984) S. cerevisiae sodium- 5.2 Long and Ward. pyruvate (1989a) S. cerevisiae sodium- 10.0 Seely et al. pyruvate (1989b) C. flareri sodium- 9.3 Seely et al. pyruvate (1989b) Fed batch with S. cerevisine glucose 9.9 Seely et al. immobilized cells (1989a) S. cerevisine glucose 10.0 Mahmoud et al. (1990a)

In the development of a continuous process for L-P AC production, single and multi-stage chemostat systems with C. utilis were tested by Wijono (1991) and Wang (1993). They concluded that a single stage continuous process was not suitable for biotransformation because non­ steady state reduction of biomass and wash-out of cells would occur due to the significant growth inhibition caused by continuous feeding of benzaldehyde, thus a multi-stage chemostat system would need to be employed as different conditions could be established then in each of the separate stages. Chapter 1 44

A system was designed to allow cell growth in stage 1, partially fermentative conditions for accumulation of pyruvic acid and PDC activity enhancement in stage 2, and finally biotransformation in stage 3 with separate benzaldehyde feeding. A maximum L-PAC concentration of 11.7 g/L was achieved with productivity of 0.35 g/L/h in complex medium containing yeast extract (Wijono 1991), and a maximum of 10.6 g/L was obtained with productivity of 0.46 g/L/h with a defined synthetic medium (Wang, 1993).

Immobilized cell processes have been found to have several advantages for biotransformation. They have been reported to provide a concentration gradient of toxic substrate within calcium alginate beads which resulted in decreasing the toxicity of benzaldehyde. As a result of increased cell density and substrate concentration, an increased rate of L­ PAC production and higher concentration of 10 g/L was achieved in a batch process (Mahmoud et al., 1990a). The cyclic, semicontinuous production of L-PAC was undertaken in an air bubble column reactor containing S. cerevisiae ATCC 834 immobilized in calcium alginate beads (Mahmoud et al., 1990b). While in the first and second cycles (length of each production period, 24 h) with the addition of 6.0 g/L benzaldehyde, L-PAC production reached to 4.5 g/L, a significant decrease was observed with further cycles in the production capacity of the immobilized cells. Over the seven cycles, the total amount of L-PAC was about five fold higher than that obtained in a single batch biotransformation. Chapter 1 45

1.4 PYRUVATE DECARBOXYLASE (PDC)

In developing a full understanding of the L-P AC production process, it is desirable to review the characteristics of the key enzyme, pyruvate decarboxylase, as fully as possible.

1.4.1 Characteristics of pyruvate decarboxylase

Thiamine pyrophosphate dependent pyruvate decarboxylase (2- oxoacid carboxylase: EC 4.1.1.1) catalyses the non-oxidative decarboxylation of pyruvic acid producing acetaldehyde and C02. Since PDC was detected in yeast extracts by Neuberg and Hirsch (1921), PDC has been found widely distributed in yeasts, fungi, many plant materials (Vennesland and Felsher, 1946) and a few bacteria (Scrutton, 1971; Neale et al., 1987). In conjunction with alcohol dehydrogenase, PDC in yeast and Zymomonas sp. enables them to produce ethanol from pyruvate via acetaldehyde.

The PDC holoenzyme is a tetramer structure with molecular weight of 230-250 kilodalton (kd) for yeasts and Zymomonas mobilis (Bringer­ Meyer et al., 1986) or 275 kd for wheat germ (Zehender et al., 1987). In both yeast and wheat germ, it consists of two dimeric s"ubunits (a.2f32) of slightly different chain lengths while PDC of Z. mobilis consists of four identical subunits (a.4). The holoenzyme contains 2-4 molecules of thiamine pyrophosphate (TPP) and magnesium ions which are obligatory cofactors. The binding strength of thiamine pyrophosphate is pH dependent and differs among the binding sites in the various enzyme molecules (Ullrich and Donner, 1970). When increasing pH to alkaline conditions (pH 7.5-8.0), the native tetramer dissociates into dimeric halves with concomitant release of TPP and Mg2+ accompanied by loss of enzyme activity.

Further study on circular dichroism changes (Hopmann, 1980) indicated that the dissociation process of the holoenzyme into its subunits occurs in two distinct steps. The first step is fast, but remains time unresolved, the second step is slow and rate determined by the subunit dissociation step. Both reactions which liberate two protons in each step are Chapter 1 46

dependent on pH and temperature. The subunit dissociation appears to be controlled by the enthalpy of activation indicating that a number of bonds (i.e. ionic, hydrogen and hydrophilic bridges) are broken. When shifted to pH 6-6.5, the dimeric halves of yeast PDC were reassociated to the tetramer form in the presence of cofactors (Schellenberger and Hubner, 1967), or even in the absence of cofactors (Hubner et al., 1990). In Z. mobilis, inactive PDC could be restored to 99.5 % of original activity at pH 6.0 in the presence of cofactors (Bringer-Meyer et al., 1986). Although the reconstituted enzyme appeared to be identical with the original enzyme, the reconstitution process resulted in some rather unexpected characteristics. When the same amount (as originally present) of thiamine pyrophosphate was added to the reconstitution mixture, a small amount of activity was recovered. A large excess of TPP was required to recover total activity. Schellenberger and Hubner (1967) proposed the following mechanism for the reconstitution process:

Holoenzyme ------> Apoenzyme + TPP + Mg2+ (1)

/TPP Apoenzyme + TPP + Mg2+ ------> Enz (2) '---Mg

TPP TPP / Enz...... __ ------> Enz~ I (3) Mg Mg

In this mechanism at pH 8.0, TPP and Mg2+ were released forming apoenzyme as shown in reaction (1). Reaction (2) represents a fast non­ ordered addition of TPP and Mg2+ to form an inactive ternary complex. Reaction (3) represents a slow cyclization of the enzyme-bound TPP and Mg2+ to form active enzyme. The above authors have suggested that the 'quasi-irreversible'. coenzyme binding is explained by a kinetically controlled equilibrium reaction (3), since the concentration of the inactive ternary complex is determined by the equilibrium constants for the binding of TPP and Mg2+ to apoenzyme as indicated by reaction (2). Chapter 1 47

The effects of thiamine monophosphate and thiamine thiazole pyrophosphate on the protein-association step were compared with that of thiamine pyrophosphate (Gounaris et al, 1975). The complete reassociation of subunits to form active oligomer was obtained only when TPP was present. It was concluded that both the pyrimidine ring and the pyrophosphate group were required for productive coenzyme binding, and it was proposed that this interaction effects a conformational changes which promotes an aggregation to form the enzymatically active holoenzyme.

The Michaelis-Menten curve of PDC was found to be sigmoidal (Boiteux and Hess, 1970). The measured Hill coefficient of 1.7-1.8 indicates a high positive cooperativity between the active sites (Ullrich and Donner, 1970). With addition of substrate to the resting enzyme, a lag phase was observed until the enzymatic conversion started. This was explained by the existence of allosteric activation sites, which must bind substrate in order to activate the inactive catalytic sites (Hubner et al. 1978).

PDC has allosteric properties influenced by phosphate (Boiteux and Hess, 1970) and acetaldehyde (Gruber and Wassenaar, 1960). PDC enzyme activity can be modified by phosphate, with the phosphate increasing the substrate cooperativity accompanied by a decrease in the apparent affinity for pyruvate. The inhibition is restricted to an increase in Km while the maximum activity is not affected. In product· inhibition, Gruber and Wassenaar (1960) reported that PDC in fresh bakers' yeast was inhibited in an apparent noncompetitive or uncompetitive manner by the acetaldehyde formed at a saturating pyruvate concentration.

1.4.2 Reaction mechanism of PDC

Thiamine pyrophosphate (TPP) is essential to the action of enzymes catalysing the decarboxylation of cx.-keto acids, the formation of corresponding aldehydes and their condensation products (cx.-keto aldehydes). In a study on decarboxylation reactions with TPP without enzyme (Yatco-Manzo et al, 1959), it was found that thiamine catalyses the decarboxylation of pyruvate to cx.-acetolactate and acetoin as the major Chapter 1 48

products, together with small amounts of acetaldehyde and unidentified products. While carbon dioxide evolution began immediately, the appearance of acetolactate and acetoin exhibited a characteristic lag due to possible formation of one or more unstable intermediates. When acetaldehyde was added, some of the intermediates reacted with it to produce acetoin directly.

The natural occurrence and enzymatic formation of hydroxyethyl thiamine pyrophosphate (HETPP) has been described by Carlson and Brown (1961), and evidence was presented also on the role of HETPP in 'active acetaldehyde' production, i.e. the intermediate formed by enzymatic decarboxylation of pyruvate, which serves as an acetaldehyde donor in enzymatic reactions. Crosby et al. (1970) used an analogue, viz. 2-(l-carboxy- 1-hydroxyethyl)-3,4-dimethyl thiazoliumchloride to determine the kinetics of non-enzymatic decarboxylation. Figure 1. 10 presents a proposed mechanism for the entire reaction that is consistent with previous model and enzymatic studies. The first step is written as a direct proton transfer between the carboxylate anion and carbon 2 of the thiazolium ring, rather than proton transfer via other acid-base groups of the enzyme. In the second step the ylid attack pyruvic acid rather than the pyruvate anion, because the greater electron-withdrawing effect of the carboxylic group should make the keto function of the acid more reactive, and the condensation reaction with other carbonyl compounds may occur. The final step is elimination of the ylid from the alcoholate anion of HETPP. The release of acetaldehyde from this compound is a reaction that would be strongly catalysed by PDC. This reaction would occur more rapidly in the hydrophobic medium due to higher concentrations of dipolar ion and more rapid elimination of dipolar ions in an aqueous solvent system.

1.4.3 Roles of pyruvate decarboxylase (PDC)

In the metabolism of sugar by fermentative yeast and bacteria, various pathways diverge at pyruvate. Pyruvate may be oxidized via the TCA cycle, or converted to fermentation products, or utilized as a key intermediate in anabolic metabolism. Chapter 1 49

+ :8 carhanion of Pyruvate initial HETPP tetrahedral adduct

HC 3 ~/R' R-~I \~s e

Figure 1.10 A proposed mechanism for the pyruvate decarboxylase (Crosby et al., 1970) Chapter 1 50

The diversion of pyruvate through the various pathways depends on its intracellular concentration, and the activities and kinetic properties of the pyruvate metabolising enzymes. These enzymes are influenced by the rate of substrate consumption, type of metabolism and the stage of budding cycle (for yeasts). In yeasts, the difference in metabolism is mainly ascribed to partitioning between respiration and fermentation, that is, between the use of oxygen or organic compounds as terminal electron acceptors. Both types of metabolism are significantly affected by oxygen and glucose. The induction of oxidoreductive metabolism incorporates a variety of individual biochemical control processes including repression of TCA cycle enzymes, and alteration of components in the electron transport chain. Different types of yeasts, e.g. S. cerevisiae and C. utilis, show different behaviour on glucose-based medium. The difference between S. cerevisiae and C. utilis is mainly in the aerobic ethanol fermentation in S. cerevisiae upon transition from glucose limitation to excess, and is accompanied by differences in the level of PDC (Alexander and Jeffries, 1990). The 'Crabtree positive yeast', S. cerevisiae immediately starts fermentation with associated initiation of ethanol formation, while the 'Crabtree negative yeast' C. utilis does not produce similar fermentation products under aerobic respiration. In S. cerevisiae, the additional glucose increases the rate of carbon dissimilation via the glycolytic pathway even under aerobic conditions. The phenomenon, called 'aerobic fermentation' is largely explained by a regulatory mechanism. This metabolic change to fermentation following the addition of glucose initiates a high rate of energy production in which glucose is rapidly metabolised to ethanol and C02. During oxidoreductive metabolism, pyruvate dehydrogenase (PDH) and PDC compete for pyruvate. Previously, it was considered that fermentation under aerobic conditions resulted from repression of respiration by fermentation, however later studies have shown that aerobic fermentation results from an inherently limited respiratory capacity (which is too small to utilize oxidatively all the sugar taken up) rather than from specific repression of respiration (Harford and Hall 1981). Rieger and co-workers (1983) reported that the onset of ethanol production was attained at a constant or critical growth rate of S. cerevisiae regardless of the inlet substrate concentration. So far, the explanation of aerobic ethanol formation (Crabtree effect) is still based on these events. For Chapter 1 51

aerobic fermentation yeasts, respiration seems to be optional; they can grow in the absence of oxygen, and fermentation is favoured when grown in the presence of large amounts of fermentable sugar. Other yeasts, which are insensitive to free glucose, are the aerobic respirating yeasts such as Candida, Rhodotorula, Pichia, Torulopsis, Trichoderma and Hansenula. They appear to have a fundamental requirement for respiration, and can not be grown in the absence of oxygen. These yeasts show relatively fast growth and high yields of biomass, and ethanol is not produced under unrestricted oxygen supply. The regulation of PDC in C. utilis differs from that in S. cerevisiae, and has been described by Schmitt et al. (1983), Butler and McConnell (1988), and Kellermann and Hollenberg (1988). The PDC of C. utilis may not be synthesised in fully aerobic conditions. When conditions are changed from aerobic respiration to fermentative growth, PDC is synthesised 'de novo'. At high cell density in batch culture, there is usually sufficient depletion of oxygen to allow C. utilis to synthesis PDC. Indeed, the activity of PDC is not likely to be constant during growth and fermentation. The PDC activity can be subject to large and rapidly reversible changes, which are influenced by the addition or deprivation of oxygen or glucose. For example, PDC was reported to be about 50 % deactivated on the depletion of glucose, but reactivation occurred immediately when glucose was added under fermentative conditions (Sims and Barnett, 1991) The rate of ethanol production and PDC activity are closely associated. Sharma and Tauro (1986) reported a high yield of ethanol production in mutant strains of yeast which overexpressed PDC and alcohol dehydrogenase. When PDC was deactivated, ethanol formation ceased. Under conditions of low PDC activity, the rate of ethanol production was limited by the PDC activity rather than by sugar transport in C. utilis (Schaaff et al., 1989; Sims et al., 1991). Bacteria such as Z. mobilis also contain relatively high amounts of PDC, which can contribute up to 4-6 % of soluble cell protein (Neale et al., 1987). In Z. mobilis, up to 95 % of carbon substrate is converted through PDC activity via acetaldehyde to ethanol and C02 (Rogers et al., 1982). Although an immunological cross reaction between PDC of yeast and of Z. mobilis has not been demonstrated (Bringer-Meyer et al., 1986), the enzymes are similar in molecular weight and subunit number (Gounaris et Chapter 1 52

al., 1975). The high PDC activity of Z. mobilis reflects to some extent the efficiencies of ethanol production. Many yeasts can utilize particular disaccharides aerobically, but not anaerobically, although these yeasts can use one more of the component monosaccharides anaerobically (called 'Kluyver effect'). For a given yeast, this Kluyver effect may be found with one disaccharide but not another. The effect is widespread among the fermentative yeasts. When Candida viswanathii, Debarymyces polymorphus, Kluyveromyces dobzhanskii and K. wickerhamii grown on glycosides which might demonstrate the Kluyver effect, the yeasts showed significantly less PDC activity than when grown on glucose or other glycosides (Sims and Barnett, 1991). In general, the Kluyver effect probably depends on a significantly reduced glycolytic metabolic flux associated with a low substrate transport rate, low substrate affinity of relevant glycosidases and deactivation of PDC under anaerobic conditions.

1.4.4 Genetic regulation of PDC

Fermentable sugars can bring about various changes in the metabolism of yeasts: they activate enzymes of the glycolytic pathway, increase the glycolytic flux and induce ethanol fermentation. The extent of this activation varies with different enzymes, species, strains, culture and testing conditions and sugar substrates (Maitra and Lobo, 1971). The loss of PDC activity 'in vitro' is reflected in a lower fermentative rate, increased pyruvate excretion and simultaneous accumulation of endogenous NADH (Schmitt and Zimmermann, 1982). Further, Schmitt and co-workers (1983) reported that the synthesis of yeast PDC was regulated by corresponding mRNA levels. The yeast structural gene PDC 1 coding for PDC has been isolated from S. cerevisiae, and cloned into the YRp7 plasmid as shown in Figure 1.11. Purified plasmid YRp7-PDC 1 showed homologous recombination between the plasmid and a chromosomal sequence. The labelled YRp7-PDC 1 sequence was used to identify the corresponding mRNA by hybridisation. The synthesis of PDC was efficiently regulated by various amounts of PDC 1 mRNA. While very low levels of PDC 1 mRNA were found in cells growing on a medium containing ethanol, glucose addition to these cells Chapter 1 53

'%~. /, ~­ ~ Pstl ~~ YRp7-PDC1 (1) ~~ 11,2kbp EcoRI

Sall EcoRI

Figure 1.11 Restriction map YRp7-PDC 1. The open box indicates sequences of the PDC 1 gene, while the filled box represent sequences not linked to PDC 1 in the genome. The area where two independent yeast DNA fragments were joined to the same vector molecule could be assigned to a short sequence (hatched box) (Schmitt et al., 1983). Chapter 1 54

led to rapid accumulation of PDC 1 mRNA.

A PDC 1 (pyruvate decarboxylase structural gene) deleted mutant of S. cerevisiae was isolated by Hohmann and Cederberg (1990). This mutant was found to have PDC activity also due to the presence of a second structural gene (called PDC 5 ). The amino acid sequences of PDC 1 and PDC 5 were 88 % identical. Deletion of PDC 5 did not cause any decrease in the specific PDC activity while pdc 1 deletion mutant had 80 % of wild type activity. It seem that PDC 5 can compensate for the loss of PDC 1 because PDC 5 is expressed only when pdc 1 is deleted. Deletion of both PDC 1 and PDC 5 resulted in no detectable PDC activity 'in vitro'. Beside regulation by glucose induction, PDC synthesis seems to be subject to autoregulation which may function at the mRNA transcriptional level.

PDC from Z. mobilis has been reported to consist of four_ identical subunits of 55 kb each (Hoppner and Doelle, 1983) indicating the presence of one structural pdc gene. The complete nucleotide sequence of the structural gene encoding PDC from Z. mobilis has been determined by Neale et al. (1987). The coding region is 1704 nucleotides long and encodes a polypeptide 567 amino acid sequence with a calculated subunit mass of 60.8 kd. Brau and Sahm (1986) investigated cloning and expression of the structural gene for PDC from Z. mobilis into E. coli. Cloning of the pdc gene from Z. mobilis into E. coli using the cosmid vector ·pHC 79 was successfully conducted (cloned plasmid called pZM 11) and revealed the following characteristics: the pdc gene was well expressed in E. coli during both respiratory and fermentative growth; the ethanol yield of the recombinant E. coli clone was considerably enhanced compared with plasmid free strains while the amount of acids formed was reduced; ethanol production appeared to restrict further increase in the substrate concentration. For example, while recombinant E. coli ZM 11 containing pZM 11 completely converted 25 mM glucose to 41.5 mM ethanol with almost no acids, in the presence of 50 mM glucose, substrate consumption and ethanol production were significantly limited compared to Z. mobilis due to the higher inhibitory effect of ethanol at the increased concentration of substrate (Ingram and Buttke, 1984). Chapter 2 55

CHAPTER 2

MATERIALS AND METHODS

2.1 MICROORGANISMS

Candida utilis used in the present study was provided by ICI Australia Pty. Ltd. Biospecialities, Mayfield, N.S.W. Australia.

Saccharomyces cerevisiae was obtained from the Dept. of Biotechnology, the University of New South Wales.

Both C. utilis and S. cerevisiae were maintained by transferring to a fresh stock culture medium every two weeks at 25°C. To maintain strains when growing on a stock culture medium (except agar, see Section 2.2), for exponential phase cultures, an equal volume of sterile glycerol was mixed with the culture broth, and stored in a freezer at -20°C. When needed for growth, it was transferred to fresh media and incubated at 25°C.

2.2 MEDIA

2.2.1 Stock culture medium

Glucose 20g Yeast extract 3.0 g (NH4)2SO4 2.0 g KH2PO4 1.0 g MgSO4.7H2O 1.0 g Agar 15 g Reverse Osmosis (RO) water up to 1 L

The pH of the medium was 6.0 before autoclaving. Chapter 2 56

2.2.2 Seed culture medium

Glucose 30g Yeast extract 5.0 g (NH4)2SO4 5.0 g KH2PO4 2.0 g MgSO4.7H2O 1.0 g RO water up to 1 L

The pH of the medium was 6.0 before autoclaving.

2.2.3 Medium for cultivation in Erlenmeyer flasks and LH fermenter

Glucose 60g Yeast extract 5.0 g (NH4)2SO4 10 g KH2PO4 2.0 g MgSO4.7H2O l.0g Na2HPO4.l2H2O l.0g CaC!i.2H2O 0.05g FeSO4.7H2O 0.05g MnSO4.4H2O 0.05g RO water up to 1 L

The pH of the medium was 6.0 before autoclaving.

2.2.4 Medium for cultivation in 100 L Fermenter

Glucose 90g Yeast extract 10g (NH4)2SO4 10g KH2PO4 3.0 g MgSO4.7H2O l.0g Na2HPO4.l2H2O 2.0 g CaCl2.2H2O 0.05 g Chapter 2 57

FeSO4.7H2O 0.05g MnSO4.4H2O 0.05g RO water up to 1 L

The pH of the medium was 6.0 before autoclaving.

2.2.5 Media Sterilization

Media for flasks and LH fermenter in batch cultures were sterilized by autoclaving at 121 °C for 20 min. Yeast extract and KH2PO4 were separately sterilized and added aseptically to the autoclaved medium.

Large amounts of continuous culture media were prepared by membrane filtration. All ingredients of the media were dissolved in RO water and filtered through asbestos prefilters. The pre-filtered media were finally passed through a 0.22 µm pore size cellulose acetate/ cellulose-nitrate membrane filter. The sterilized media were collected in previously sterilized 20 L bottles and kept at 4°C not longer than 4 weeks until use.

For 100 L fermenter medium, 50 L RO water was added into the fermenter. With starting agitation at 80 rpm, medium components were gradually added into fermenter except yeast extract and KH2PO4 which were separately sterilized and added aseptically to the fermenter. When components were dissolved in the reactor, more RO water was added up to 80 L, and then the sterilization procedure was begun. The air was vented from the reactor head space and replaced by steam, then all ports were closed and the pressure allowed to increase to 15 psi (121 °C). This temperature and pressure were held for 20 min making sure that the steam reached all pipes and ports in contact with the medium, then the steam sterilization cycle was turned off and the reactor was allowed to cool for at least 1 hour. Chapter 2 58

2.3 CULTURE SYSTEMS

2.3.1 Flask culture

Flask cultures were conducted using 500 mL baffled Erlenmeyer flasks. Culture media (100 mL) were added into flasks, which were stoppered with non-absorbent cotton wool. The prepared flasks were placed on a rotary shaker, and incubated at 25°C and 180 rpm.

2.3.2 LH fermenter

The LH fermenter (2 L total volume) consisted of three major components: fermenter vessel, agitation system, and measurement and control cabinet. The fermenter vessel was made of a tempered glass cylinder. The top lid was fitted with ports for inoculum, media, acid, alkali, antifoam, air inlet, exit gas outlet and sampling device. Various measuring probes were fitted, e.g. an autoclavable pH probe (Ingold Type number 465- 50-57), dissolved oxygen probe (a modified silver-lead electrode: Lee and Tsao, 1979), a heating element, Pt 100 temperature sensor, an antifoam sensor, a cooling finger, and agitator which was mounted. An overflow outlet was attached to the lower part of the vessel. Inside the vessel there was a removable baffle device with four stainless steel baffles. An agitator shaft was positioned in the center of the vessel with two open turbine impellers, with an opening to release sterile air into the medium. The sterile air after passing through a sterile Millipore filter (pore size 0.25 µm ), entered the top part of the agitation shaft. The control panel was equipped with a temperature controller (Shinko, Model No. 586284) connected to a temperature sensor, a dissolved oxygen tension controller (Kent, Model No. 96 M), pH controller (Dynaco Epson, Surrey England) and an antifoam controller connected with foam sensor. A sterile antifoam solution was added to the fermenter by an on/off mode open-loop sequence controller when necessary. The sequence time could be varied within wide limits. The overall system design includes peristaltic pumps to transfer alkali (3.0 N NaOH), acid (5 % HCl) and Chapter 2 59

antifoam (polyethyleneglycol 1025 BDH). The fermenter has 2 L capacity with a working volume of 1.6 L. The gas outlet was connected to exit gas analyzers, e.g. CO2. and 02 (See Section 2.4).

2.3.3 Porton-type Stirred Tank Reactor

The Porton-type stirred tank reactor consisted of a standard glass cylinder (15 cm ID x 30 cm L) which was sealed at the top and bottom by stainless steel plates (280 mm ID x 10 mm T). Two stainless steel plates were tightened against both ends of a standard QCV glass cylinder with neoprene gaskets as seals. The top plate was fitted with 7 process ports which were used for the connection of the measuring probes, e.g. pH probe, DOT probe and condenser equipped with a filter for gas out as well as connectors for medium, alkali, antifoam agent and inoculum. The agitator shaft attached to the center of base plate was equipped with two six-bladed flat turbine impellers. The bottom plate was equipped with an overflow pipe, air inlet line, temperature sensor, steam lines, a cooling finger and sampling port. A schematic diagram of the Porton type fermenter and photograph are shown in Figure 2.l(a) and (b). A thermistor and a temperature controller (Pye Ether, Type 19-90/1, England) were used in measuring and controlling the cultivation temperature. A 250 watt infrared reflection lamp (Osram, England) connected to the temperature controller was used for heating and the stainless steel cold finger immersed into the culture broth was used for cooling. The pH was controlled by means of a Dynaco pH meter-controller (Epson, England) in association with a steam sterilisable pH probe (Probion, England) and a Delta pump (Watson-Marlow Ltd, England) for alkali/acid addition. DOT level was measured by a modified Johnson type silver-lead electrode (Lee and Tsao, 1979) connected to a Kent P 96M microprocessor with digital display. A pH probe and an oxygen electrode were inserted into the fermenter vessel through the ports on the top plate, and sealed with silicone. The air supply was regulated using a manostat and a rotameter. After passing through a sterile Gamma 12 filter unit (Whatman grade 12- 03), the air was passed into the culture broth through holes at the base of I. Fermenter 141(\\ 2. Oxygen probe 0 3. Cold finger () 4. pH probe ::r 5. Thermistor Q) --0 6. Temperature controller ~ 7. Infrared Lamp .... N 8. Air filter 9. Rotameter 18 13 10. Manostat 11. DOT meter 12. Timer 13. Antifoam reservoir 19 14. Motor 15. Gear box 20 16. Condenser 9,. 17. Air filter 4 II 112 18. pH meter and controller 1 19. Alkali reservoir 20. Inoculum port 3 21. Sampling port 10

21

Air in

Figure 2.l(a) Schematic diagram of the Porton type stirred tank reactor.

0\ 0 Chapter 2 61

Figure 2.l(b) Photograph of the Porton type stirred tank reactor. Chapter 2 62

the agitator shift. Volatile components escaping from the vessel were condensed in the condenser and the residual gas passed out through a miniature line filter.

2.3.4 Pilot scale 100 L Fermenter

2.3.4.1 The vessel

The pilot scale 100 L fermenter was a conventional stirred tank reactor. A schematic diagram of the fermenter system is shown in Figure 2.2. The fermenter vessel (1) was obtained from L H Engineering, U.K. It was constructed from stainless steel (316), mirror polished internally, and had a maximum working volume of 100 L. The vessel had a bottom driven impeller shaft (2) which was isolated from the outside via a pressurized crane seal (3). The crane seal was lubricated by pumping sterile condensate with a SMC pump (30 W, 0.16 A bronze), through the seal. The sterile condensate was stored in a stainless steel reservoir (19), and was maintained at 1 atm above the internal vessel pressure. A rotameter (29) controlled the air flow sparged into the bottom of the fermenter (4), just below one of the paddle turbines (5) attached to the impeller shaft, and separated from each other by two turbine diameters. The air was sterilized by the passage through two filters, one a 9 inch glass wool packed prefilter (6), the other, a cellulose/acetate Millepore filter (7) (pore size 0.2 µm). Agitation was provided by a 2 HP motor (NECO type 1 NS. B. Shunt) (8) attached by a belt drive to the impeller shaft, and controlled by an electronic tacho-feedback speed control unit on the console panel. The vessel has six ports. One was located on the under side of the base, for removal of the product liquor (9), and five were located on the sides of the vessel. Three of these were grouped together in a recessed housing and were used for the placement of the pH, temperature and dissolved oxygen probes (10).

N N

~ ~

~ ~

n n

(t) (t)

~ ~

::r ::r

......

'"O '"O

pump pump

sterilization sterilization

for for

sensor sensor

controller controller

reservoir reservoir

circulation circulation

pump pump

jacket jacket

port port

reservoir reservoir

probe probe

controller controller

gauge gauge

reservoir reservoir

port port

controller controller

in in

reservoir reservoir

port port exit exit

in in

in in

exit exit

probe probe

controller controller

probe probe

outlet outlet

filter filter

controller controller

pH pH

pH pH

Pressure Pressure Antifoam Antifoam

Rotameter Rotameter

Alkali Alkali

Inlet Inlet

Condensate Condensate

Antifoam Antifoam Fermenter Fermenter

Condensate Condensate Circulation Circulation DO DO

Acid Acid Sampling Sampling Antifoam Antifoam

Motor Motor

DO DO Air Air Temperature Temperature

Temperature Temperature

Temp Temp

Outlet Outlet Prefilter Prefilter

Air Air

water water

steam steam

water water

water water

water water

<2 <2

ci ci

16. 16.

18. 18.

15. 15.

17. 17.

19. 19.

13. 13.

t> t> 12. 12.

14. 14.

11. 11.

10. 10.

30. 30.

29. 29. 22. 22.

24. 24.

26. 26.

28. 28. 21. 21.

23. 23.

25. 25. 20. 20. 27. 27.

8. 8.

9. 9.

7. 7.

6. 6.

shaft shaft

seal seal

sparger sparger

• •

'-J_ '-J_

r---

: :

,, ,,

. .

,:: ,::

• •

Vessel Vessel

Crane Crane

Air Air Impeller Impeller

Impeller Impeller

......

., .,

-

l. l.

.. ..

3. 3.

4. 4.

5. 5. 2. 2.

-

------<1 ------<1

exit exit

......

~------·-·1 ~------·-·1

V V

~ ~

~ ~

______

water water

'------·------.,: '------·------.,:

: :

......

! ! ; ;

~ ~

~ ~

r r

; ;

I I I I

. . . .

in in

: :

• •

., .,

~ ~

'' ''

: : .. ..

iooJ--: iooJ--:

wL.. wL..

steam steam

----

/l,. /l,.

2,~·H1 2,~·H1

. .

27 27

?!:!. ?!:!.

--

------·4 ------·4

......

;:,------..--4------1> ;:,------..--4------1>

fermenter. fermenter.

---

-

......

-······------~ -······------~

. .

L L

-

-

~ ~

~ ~

1

-

i i

: :

......

-

i: i:

.. ..

. .

-: -:

14 14

100 100

-

-

--

2 2

-

......

the the

------1 ------1

---

of of

. .

r·------~ r·------~ ' '

-

......

t. t.

.-

6 6

V V

111 111

features features

exit exit

V V

==::;;:::; ==::;;:::;

i.. i..

11 11

......

, ,

steam steam

7 7

design design

19 19

The The

I> I>

2.2 2.2

18 18

in in

I I

t,, t,,

air air

2811 2811

Figure Figure

17 17

exit exit

16 16 water water Chapter 2 64

One of the other two remaining ports, near the top of the vessel, was connected to the steam line and is used for the cleaning of a viewing port on the lid of the fermenter. The other was a steam sterilizable sample port (11). The lid of the fermenter has 12 ports, and was attached to the vessel by 18 x 1/2 inch bolts and sealed by a large O-ring. Eight of these ports were used as inlets, the remaining 4 included an illuminated viewing port, an air outlet (12), the antifoam probe (14), and a pressure gauge with a pressure release valve (15). Of the eight inlet ports, two were connected to the steam line and so are resterilizable (13). One was a 3 inch diameter port for addition of semi-solid materials, and for cleaning purposes. This port has a lid which was fastened by 4 wing nuts to the top of the fermenter and sealed by an O-ring. Three of the remaining 5 inlets were connected to one of three separate stainless steel (316) reservoirs, each reservoir capable of being steam sterilized or pressurised by compressed air to 1 atm. The three reservoirs contained antifoam (17), acid (18) and alkali (19). Additions from these reservoirs were controlled from the control console via solenoid valves.

2.3.4.2 The Control Console

Sterilization of the vessel was by autoclaving. This was done by the passage of steam, via a solenoid valve into the jacket, and controlled by a Leeds & Northrup 2022 Star 1, on-off controller (20), mounted on the control console. During sterilization, steam was also introduced into the vessel via the product outlet port, and the sampling port, all valves connected to the various ports being closed. Temperature regulation was achieved by the circulation of water, using an SMC pump (31), inside the jacket (21). The temperature was altered by introduction of steam or cooling water, via solenoid valves, as required. The solenoid valves were in turn regulated by signals produced by the temperature sensor (resistance thermometer, 100 ohm platinum 3-lead type) (22), and transmitted to a Leeds & Northrup Electromax Controller (23). The pH was monitored and controlled by a steam sterilizable pH probe (Ingold, type 764-31) (24), capable of being pressurized to 6 atm, whose signal was fed to a Kent pH meter/ controller (Model No. 9150 EIL) (25) which then controlled a solenoid valve via one of two 3 minute, percentage Chapter 2 65

timers (Crouzet, Model 88 282). Dissolved oxygen was measured by a galvanic type Pb/ Ag probe (27) and the signal fed to a dissolved oxygen meter (Fermentation Design) (28). The signals from the various probes were transmitted to and recorded on a cyclic 6-channel chart recorder (Kent Clearspan P 120L).

2.3.5 Sterilization of Fermenters

The LH fermenter was autoclaved at 121 °C for 20 min. Before autoclaving, the vessel was fitted with autoclavable pH and DO probes. All stainless steel connectors at the ends of each line were plugged with cotton wool and covered by aluminium foil. The condenser, the reservoirs for alkali or acid and the reservoir with antifoam were separately sterilized. Having been cooled, all of the necessary connections were made aseptically, and then sterilized medium was added into the fermenter. The Porton-type stirred tank reactor was sterilized by steam injection for at least 24 h. The air filter and lines used for sterile aeration were separately autoclaved. After calibration, the pH and DO probes were sterilized in 10 % formaldehyde for a few hours, and then washed with sterile water and inserted aseptically into the fermenter. For the 100 L fermenter, prior to autoclaving the fermenter contents, it was necessary to pass live steam through all ports on the fermenter for at least twenty minutes on two successive occasions. These ports then remained closed when autoclaving the fermenter contents. During sterilization, steam was injected directly into the fermenter vessel and jacket. Time taken to reach 100°C was approximately 20 minutes. This was followed by 10 minutes at that temperature, to ensure escape of all oxygen from the medium and fermenter head space. The pH probe was pressurised to 22 psi using a handy pump, and the pressure on the coolant/lubricant reservoir was increased to 20.5 psi. All valves on the lid were then closed and the pressure was allowed to rise to 15 psi. At this pressure the temperature inside the fermenter should be 121 °C. This temperature was maintained constant by adjustment of the temperature controller so that the green LED on to display was illuminated. The controller regulated the steam entering the jacket by the opening and closing of a solenoid valve located in the steam line. The holding time for sterilization was normally Chapter 2 66

around 20 minutes, after which the temperature setting on the controller was readjusted following turn off. The air outlet valve was then opened and the pressure allowed to decrease to around 4 psi. Air at a slow flow rate was then turned on. The flow rate was gradually increased to around 50-70 L/min with the air outlet valve fully opened. If the internal pressure started to rise again the air flow rate was reduced accordingly. The temperature normally dropped rapidly and reached operating temperature in about 30-40 minutes, depending on the ambient temperature.

2.4 BIOTRANSFORMATION SYSTEM WITH IMMOBILIZED CELLS

The biotransformation system consisted of several components as follows: fermenter; benzaldehyde feeding pump; exit gas analyzers; computer linked on-line RQ measurement. The schematic diagrams are shown in Figure 2.3 and their photographs in Figure 2.4. In this system, the 2 L LH fermenter was used usually as described in Section 2.3.2. For the continuous process, an over-flow outlet was covered with a stainless steel sieve (mesh size 1.0 mm) to maintain the immobilized cell beads in the fermenter. The culture media were fed into the fermenter by means of a peristaltic pump (Gilson Minipul, French). To feed benzaldehyde into the fermenter, a syringe pump (Perfusor®VII, B. Braun, Germany) was used with variable feed rates in the range of 0.1-99.0 mL/h, by means of 50 cc disposable syringe. Prior to analysis for 02 and CO2, the exit gas was dehumidified by a cold dehumidifier (Komatsu Electronics Inc. Model OH 1052G, Japan) to meet the requirements of the gas analyzers. The requirement of gas for the analyzer was for a clean, dry sample with dew point 5°C below ambient temperature. The basic principles of these analyzers are that the content of oxygen is measured by the paramagnetic susceptibility of the sample (Servomix type 1400A), while the content of carbon dioxide was measured by an infrared gas analyzer with single beam dual wavelength (Servomix-R type 1410). Output signals (4-20 mA) from both gas analyzers were fed into a data interface (Data system FC-4, Real Time Engineering, Australia) linked to an NEC Powermate 286 Plus computer. RQ values were calculated instantaneously by this computer based on Equation 2.1 (see Section 2.8) and the results were saved in a reserved memory space.

N N

n n

(I) (I)

'"I '"I

Ill Ill

,..,. ,..,.

::r ::r

-....::i -....::i

0\ 0\

"O "O

~ ~

h h

controller controller

pump pump

reservoir reservoir

controller controller

reservoir reservoir

box box

15 15

pump pump

-

computer computer

I I

processor processor

analyzer analyzer

controller controller

L:. L:.

16 16

analyzer analyzer

controller controller

exit exit

NEC NEC

Data Data

Overflow Overflow

Peristaltic Peristaltic

02 02

Dehumidifier Dehumidifier

C02 C02

Syringe Syringe

Fermenter Fermenter

pH pH

Medium Medium

Agitator Agitator

Antifoam Antifoam

DOT DOT

DOTmeter DOTmeter

Temperature Temperature

l. l.

l. l. l l

16. 16.

14. 14.

13. 13.

12. 12.

15. 15. 10. 10.

8. 8.

•air •air

6. 6.

5. 5.

9. 9.

2. 2.

3. 3.

7. 7.

4. 4.

_ _

-

1-1· 1-1·

.11 .11

g g

1----1= 1----1=

;----~-1-·--;=1~~~ ;----~-1-·--;=1~~~

12 12

.. ..

Cil Cil

13 13

with with

D D

a a

[I] [I]

(@) (@)

system system

9 9

1 1 1 1

D D

in in

biotransformation biotransformation

air air

• •

the the

(I (I

for for

I, I,

I I

jl jl

,, ,,

11 11

10 10

,~, ,~,

,o ,o

I I

I I

! !

n n

2 2

-, -,

::c=:: ::c=::

-

CD CD

-

-

diagrams diagrams

______

-

-

-

-=------=---=-

c::! c::!

= =

= =

-

= =

- --

------~ ------~

: :

cells. cells.

Schematic Schematic

~------

----

--

2.3 2.3

4 4

,~0~1 ,~0~1

BJ BJ

-----

31@@ 31@@

Figure Figure

immobilized immobilized

--

7 7 6

5 5

E] E]

@] @] ---- Chapter 2 68

(a)

Figure 2.4 Photograph of (a) LH fermenter containing immobilized cells; 0\

..c::

\()

u

N

...

(13

I-<

Figure

Analyzer

2.4

for

Photograph

02,

(d)

Dehumidifier

of

(b)

Exit

,

(e)

Data

Gas

processor

Analyzer

and

for

(f)

C02

NEC

,

(c)

computer

Exit

gas

. Chapter 2 70

2.5 METHODS OF ANALYSIS

2.5.1 Procedure of Sample Preparation for Analysis

The schematic diagram in Figure 2.5 illustrates the various procedures for sample pretreatment. Ten millilitre of fresh samples were taken from the culture system. Immediately after sampling, cells were separated by centrifugation (5,000 rpm, 10 min) and the supernatants were kept in sealed test tubes at 4°C prior to analysis. If the analyses were not carried out on the day of sampling, samples were stored at -20°C. The packed cells were resuspended in 10 mL of saline (0.85 % NaCl) and used for estimation of cell mass, or further enzyme assays.

(1) Sample of culture broth

Centrifugation (Sigma 203, B. Braun) at 5,000 rpm for 10 min

I I I Supernatant Cells Washing \ I Glucose Cell cake Pyruvate I Ethanol Dry cell Weight Disrubtion

Centrifugation (Eppendorf Centrifuge at 12,000 rpm for 3 min)

POC IADHr ADHcin Chapter 2 71

(2) Sample from biotransformation with immobilized cells

Centrifugation (Sigma 203, B. Braun) at 5,000 rpm for 10 min I SupernatantI Calciul alginate beads I I PyruvateI Extraction by Disruption Ethanol dichloromethane Glucose CentrifugationI L-PACl (Eppendorf Centrifuge Benzaldehyde at 12,000 rpm for Benzyl alcohol 3 mim)

PDC ADHrI ADHcin

(3) Sample from biotransformation with PDC enzyme

Centrifugation (B. Braun) at 5,000 rpm for 10 min

SupernatantI

I I Pyruvate Extraction1by dichloromethane I Acetaldehyde L-PAC Acetoin Benzaldehyde

Figure 2.5 Procedures for sample preparation for analysis of chemicals and enzymes. Chapter 2 72

2.5.2 Cell disruption and enzyme extraction

2.5.2.1 Extraction of enzymes from free cells

Cells from one milliliter of culture broth were harvested by Eppendorf Centrifuge at 12,000 rpm for 1 min and washed twice with 30 mM Tris buffer (pH 6.5). Cells were resuspended in the same buffer and adjusted to approximately 0.4 mL of cell suspension. Approximately 1 g of glass beads (size 0.5 mm, B. Braun Cat No. 854 170/1) was mixed with 0.4 mL of cell suspension, and then vortexed at maximum speed for 2 min. For every 30 sec vortexing, the sample was given a 1 min cooling in an ice bath. Cell debris was removed by Eppendorf Centrifuge at 12,000 rpm for 3 min. The supernatant was collected for subsequent enzyme assays and protein determination.

2.5.2.2 Extraction of enzymes from immobilized cells

To extract enzymes from immobilized cells, 3 mL beads containing the immobilized cells were put into a ceramic hammer mill and gently extracted with 3 g of pre-treated fine sand (which was washed 2-3 times with 3 N HCl and RO water until neutralized). Crushed immobilized beads were suspended into 20 mL water, and centrifuged at 1,000 rpm. From the resultant supernatant, yeast cells were harvested and washed with 30 mM Tris buffer (pH 6.0) by B. Braun Centrifuge at 5,000 rpm for 10 min. The harvested cells were resuspended into 3.0 mL 30 mM Tris buffer, and then 1 mL of yeast suspension was centrifuged at 12,000 rpm for 1 min. Cells were resuspended in the same buffer to give approximately 0.4 mL of cell suspension, and then enzymes were extracted by ball milling.

2.5.2.3 Extraction of enzymes with high pressure homogenizer

Large quantities of cells were ruptured using a Manton-Gaulin APV high pressure homogenizer. The degree of cell disruption estimated by measuring the release of protein using the Bradford test (1970). The cells Chapter 2 73

were suspended in cooled 40 mM KH2PO4 buffer (pH 6.0) and added to the reservoir of the homogenizer. The pressure was set to 500-700 kg/cm2. The cells were disrupted by several discrete, single passes with cooling in between. The cell debris was removed by centrifugation and the protein concentration of the supernatant was determined by the Bradford method (Bradford, 1970).

2.5.3 Dry cell weight estimation

The dry cell weight of biomass was estimated using pre-weighed glass tubes (10 mm ID x 70 mm L), and the average values of 3 measurements of each sample were determined. The procedure was as follows: after centrifuging and resuspending in saline, 4 mL of the cell suspension was transferred to the pre-weighed glass tubes and centrifuged at 5,000 rpm for 10 min. Then the glass tubes were dried in an oven at 105°C for 24 h, cooled in a desiccator and weighed.

2.5.4 Estimation of glucose concentration

Glucose concentrations were determined by a YSI glucose analyzer (Yellow Springs Instruments Co., Model 27). This estimation was based on the use of immobilized glucose oxidase. After the sample was injected, glucose diffused through an outer polycarbonate membrane and was oxidized to H2O2 by immobilized glucose oxidase. The hydrogen peroxide produced diffused through a cellulose acetate membrane on the inner side of the enzyme membrane sandwich and reacts with a platinum anode. The instrument measured the current flow in the anode circuit which was proportional to the local concentration of hydrogen peroxide, and the result was a signal current proportional to the glucose content in the sample. Glucose concentrations were determined after calibration of the analyzer with 1.0 g/L standard solution.

glucose oxidase Glucose+ 02 ------> Glucose-6-lactone + H2O2 Chapter 2 74

platinum anode ,----->

silver cathode 2 AgCl +2 e------> 2 Ag + 2 Cl-

2.5.5. Estimation of ethanol and acetaldehyde concentrations

Ethanol and acetaldehyde concentrations were estimated using a gas chromatograph (Packard, U.S.A. Series 427). The relevant column and its operation were as follows:

Column material 1 / 4 inch glass Column length 1.5 meter Packing material PorapakQ Column mesh range 100-200 µm Carrier gas Nitrogen (30 cc/ min) Oven temperature 180°C Injector temperature 220°c Detector temperature 220°c Detector type Flame ionization detector (FID) with hydrogen (20 cc/min) and air (300 cc/ min) Operation mode Isotherm Injection sample 3 µL

The ethanol and acetaldehyde concentrations of the sample were measured by comparison with standard samples.

2.5.6 Estimation of benzaldehyde, benzyl alcohol and L-PAC concentrations

Concentration of both substrate, benzaldehyde, and products, benzyl alcohol and L-PAC, were determined by gas chromatography. Samples were prepared by extracting in volatile solvent (sample:solvent=l:5), The Chapter 2 75

biotransformation sample (0.2 mL) was mixed with 1 mL of dichloromethane in an Eppendorf tube and vortexed for 2 min. A sample from the bottom organic layer was injected into the Packard Series 427 gas chromatograph with the column and its operating conditions as follows:

Column material 1 / 4 inch glass Column length 1 meter Packing material Chromosorb W.Hr/SE 30WTX 10 Column mesh range 80-100 µm Carrier gas Nitrogen (30 cc/min) Oven temperature 115°C Injector temperature 180°C Detector temperature 180°C Detector type FID with hydrogen (20 cc/min) and air (300 cc/min) Operation mode Isotherm Injection sample 3 µL

The concentrations of benzaldehyde, benzyl alcohol and L-P AC were determined by comparison with standard samples of benzaldehyde and benzyl alcohol provided from BDH, and L-PAC from ICI Australia Pty. Ltd.

2.5.7 Pyruvic acid determination

Determination of pyruvic acid was carried out by an enzymatic analysis (Boehringer-Mannheim Analytical Kit No. 718 882). This analytical kit consisted of lactate dehydrogenase, and NADH+H+ and standard pyruvic acid (100 mg/L). In the presence of NADH+H+, lactate dehydrogenase reduced pyruvic acid to lactic acid.

lactate dehydrogenase Pyruvic acid+ NADH+H+ ------> L-lactic acid+ NAD+

The amount of NADH+H+ oxidized to NAD+ corresponded stoichiometrically to the amount of pyruvic acid. The decrease in Chapter 2 76

NADH+H+ was determined by difference of its absorbency at 340 nm, measured with a Shimadzu UV-120-02 Spectrometer.

2.5.8 Estimation of acetaldehyde by enzymatic analysis

The determination of acetaldehyde in small amounts (below 100 mg/L) was carried out by enzymatic analysis (Boehringer-Mannheim Analytical Kit, Cat. No. 668 613). Acetaldehyde was oxidized quantitatively at pH 9.0 in the presence of aldehyde dehydrogenase by nicotinamide­ adenine dinucleotide (NAD+) to acetic acid.

aldehyde dehydrogenase Acetaldehyde + NAD+ + H2O ------> Acetic acid + NADH + H+

The amount of NADH+H+ formed was stoichiometrically related to the amount of acetaldehyde. NADH+H+ was determined by means of an increase of absorbency at 340 nm. The amount of acetaldehyde in the cuvette should range between 2 µg-20 µg (in 0.2-0.5 mL sample volume).

2.5.9 Colorimetric determination of acetoin

Determination of acetoin was based on the colour reaction with creatine and a-naphthol (Westerfeld, 1945). Reagents were prepared as follows: 0.5 % creatine: 1 g of creatine dissolved in 200 mL of water 5 % a-naphthol: 1 g of powdered colourless a-naphthol (redistilled under nitrogen) dissolved in 20 mL of 2.5 N NaOH. The solution was prepared immediately before using. Details of the method for determining acetoin are as follows: To 5 mL of solution containing between 1-12 µg of acetoin was added consecutively 1 mL of 0.5 % creatine and 1 mL of the 5 % a-naphthol solution. The latter reagent should not be prepared until after the creatine has been added to the test solution, and it should be used as soon as the a- Chapter 2 77

naphthol dissolved. The colour was allowed to develop at room temperature for 1 hour. The amount of colour was then read at 540 nm with a Shimadzu spectrometer UV-120-02. The colours obtained by this method were reproducible but somewhat unstable. The maximum intensity was reached in 60 min after the addition of the alkaline a-naphthol; fading could not be detected within 75 min. The acetoin concentration of the sample was measured by comparison with a standard sample (Aldrich Cat No. 10897-9).

2.5.10 Analysis of Enzyme Activities

2.5.10.1 Pyruvate decarboxylase (PDC)

The activity of POC was assayed by coupling the decarboxylation reaction with the alcohol dehydrogenase mediated reaction and monitoring the oxidation of NAOH+H+ to NAO+ at 340 nm. The basic principle of the POC activity assay is given below (Bergmeyer, 1974):

COi pyruvate decarbo~ alcohol dehydrogenase Pyruvate - Acetaldehyde .. ~ --...... • Ethanol TPP, Mg2+ /' ~ NAOH+H+ NAO The reaction mixture consisted of the following:

Solution Volume (µL) 200 mM sodium citrate buffer (pH 6.0) 950 10 mg/mL NAOH (sodium salt) 10 10 mg/mL Alcohol dehydrogenase 3 (Sigma Chem. Co, Prod. No. A-3263) 100 mg/mL sodium pyruvate 32 Enzyme sample 5

One unit of the enzyme activity is defined as that activity which converts 1.0 micromole of pyruvate to acetaldehyde per minute at pH 6.0 and 25 °C. Chapter 2 78

2.5.10.2 Alcohol dehydrogenase for ethanol (ADHT)

The basic reaction for determination of alcohol dehydrogenase activity for ethanol is as follows (modified from Bergmeyer, 1974).

alcohol dehydrogenase Ethanol ..,,7 ~ > Acetaldehyde NAD+ NADH + H+ The reaction mixture consisted of the following:

Solution Volume {µL) 35 mM Trisma base (pH 8.5) 935 20 mg/mL NAD+ 30 Absolute ethanol 30 Enzyme sam le 5

The activity of the enzyme was monitored by changes in absorbency at 340 nm for NADH formation using a Shimadzu UV-120-02 spectro­ photometer. The results were recorded on an Omniscribe Chart Recorder Series D-1000 (Houston Instruments. USA), operating at 1 inch/min. One unit of enzyme activity is defined as that activity which converts 1.0 micromole ethanol to acetaldehyde per minute at pH 8.5 and 25°C.

2.5.10.3 Aromatic alcohol dehydrogenase (ADHcin)

The basic reaction of an aromatic alcohol dehydrogenase assay is shown below. Cinnamyl alcohol (Aldrich Cat No.10819-7) was selected as representative of aromatic alcohols as this alcohol has been used previously as a typical test of alcohol dehydrogenase type II (Ebisuzaki and Barron, 1957)

aromatic alcohol dehydrogenase Cinnamyl alcohol ---~ > Cinnamaldehyde

NAD+ NADH + H+ Chapter 2 79

The reaction mixture consisted of the following:

Solution Volume (µL) 35 mM Tris-base (pH 8.5) 865 20 mg/mL NAD 30 10 mg/mL cinnamyl alcohol in 12.5% methanol 100 Enz me sample 5

The activity of the enzyme was monitored by changes in absorbency at 340 nm for NADH formation as outlined for estimation of alcohol dehydrogenase for ethanol. One unit of enzyme activity is defined as that activity which converts 1.0 micromole cinnamyl alcohol to cinnamaldehyde per minute at pH 8.5 and 25°C.

2.5.11 Protein determinations

Protein determinations were carried out by two different methods according to the protein concentration. When crude enzyme (PDC) was extracted from yeast cells, the Bradford method (Bradford, 1970) was used, while following enzyme purification and for measurement of specific activity, the Lowry method (Lowry et al., 1951) was used.

(1) Bradford method:

The Bradford reagent was prepared according to the following procedure: 1. Dissolve 100 mg of Coomassie Brilliant Blue G 250 in 50 mL of 95 % ethanol. The Coomassie Blue is difficult to dissolve and requires a minimum of one hour using a magnetic stirrer for adequate dissolution. 2. Add to 500 mL of distilled water in a 2 L beaker. 3. Slowly add 100 mL of 85 % orthophosphoric acid while stirring; do this over about ten minutes to make the filtration step easy. Chapter 2 80

4. Make the final volume up to 1 L with distilled water. 5. When the solution has cooled, filter through a sintered glass filter using vacuum. Store the solution in an amber bottle covered with aluminium foil.

Protein determination

1. Transfer 200 µL sample into a clean dry test tube. 2. Add 3 mL of reagent and vortex the sample to ensure adequate mixing. Allow the sample to stand at room temperature for at least five minutes. The OD of the sample should be read on a spectrophotometer at 595 nm against a blank (200 µL RO water + 3 mL reagent). Samples should be read within half an hour of their preparation. 3. The protein concentration is measured by comparison with standard samples (Bovine serum albumin). Turbid samples need to be centrifuged to remove suspended cells and debris. For sample preparation, centrifuge approximately 1 mL of sample in an Eppendorf tube for 2 minutes and use the supernatant for protein analysis as above.

(2) Lowry method for protein determination

Reagents were prepared as follows: Reagent A: dissolve 100 g Na2CO3 in 1 L (final volume) 0.5 N NaOH. Reagent B: dissolve 1 g CuSO4.5H2O in 100 mL distilled water. Reagent C: dissolve 2 g potassium tartrate in 100 mL distilled water.

Protein determination

1. Bring total volume of sample to 1.0 mL by adding an appropriate amount of glass-distilled water. 2. Mix thoroughly 15 mL Reagent A, 0.75 mL Reagent Band 0.75 mL Reagent Cina 50 mL Erlenmeyer flask. 3. Add 1 mL of above reagents mixture to a prepared sample. Vortex the tube to mix it thoroughly. 4. Incubate the tube for 15 minutes at room temperature. Chapter 2 81

5. While the tube is incubated, add 5.0 ml 2 N Folin-phenol reagent to 50 ml distilled water. Mix the solution thoroughly. 6. At the conclusion of the incubation period, pipette 3.0 ml of the solution into the tube. Vortex the resulting solution immediately. 7. Incubate the sample at room temperature for 45 minutes, and determine the absorbency of the sample at 540 nm. 8. The protein concentration is determined by comparison with standard samples (Bovine serum albumin).

2.6 PREPARATION OF SPECIMENS FOR SCANNING ELECTRON MICROSCOPE

Specimens to be viewed by the Scanning Electron Microscope (SEM) must first be dried. Otherwise, the low pressure in the electron microscope will cause water (and other volatile liquids) to boil violently out of the specimen, disrupting its structure and increasing electron scattering by raising the pressure in the microscope. Yeast cells with a high water content will collapse if air dried, since the water/air interface tension forces associated with the cell walls will cause structural damage. To minimize the effect of surface tension changes when drying specimens, preparation of specimens was carried out as follows (Hayat, 1978; Cohen, 1979): 1. Fix with buffered aldehyde After appropriate washing, the specimen was fixed in glutaraldehyde in buffer for up to 4 hours. After this, or at some other convenient stage, the specimen was enclosed in a suitable solvent proof basket to give protection and enable different specimens that were to be dried together to be segregated. Gentle agitation was recommended during this stage and the subsequent stage of dehydration. The specimens must remain immersed throughout processing.

2. Wash with buffer Chapter 2 82

Unless osmium fixation was to follow, prolonged washing was unnecessary. 3. Dehydrate specimens in a series of alcohol solution Treatment Duration (min) 50 % ethanol + 0.1% NaCl 5 50 % ethanol + 0.1% NaCl 5 70 % ethanol + 0.1% NaCl 10 95 % ethanol + 0.1% NaCl 10 100 % ethanol+ 0.1% NaCl 15 100 % ethanol+ 0.1 % NaCl 15 4. Substitute acetone for ethanol Treatment Duration (min) 100 % acetone (dry) 10 100 % acetone(dry) 15 5. Load samples into the Critical Point Drying Chamber (CPD) 6. Substitute acetone with CO2 7. Heat chamber to produce phase transition 8. Release pressurized CO2 from the heater chamber 9. Remove specimens from the chamber 10. Mount the specimen 11. Examine the sample using the Cambridge Stereoscan S-360.

2.7 DETERMINATION OF RESPIRATORY QUOTIENT (RQ) VALUES

RQ values were estimated based on a nitrogen balance, i.e, N2(inlet) = N2(exit), since nitrogen is inert (Brix et al., 1988)

Carbon dioxide evolution rate (CER) Respiratory Quotient (RQ) = ------(Eq. 2.1) Oxygen uptake rate (OUR)

Flow rate of the exit gas : F(e) = F(i) x ---- (L/h) (Eq. 2.2) % N2(e) CER = [F(e) x % CO2] /22.4 (mole/h) (Eq. 2.3) Chapter 2 83

OUR= [(F(i) x %Ow)) - (F(e) x %02(e))]/22.4 (mole /h) (Eq. 2.4)

Definitions of the terms are as follows: F(e): exit air/ gas flow rate (L/h) F(i): inlet air/ gas flow rate (L/h) N2(e): exit nitrogen content (%) N2(i): inlet nitrogen content (%)

2.8 EVALUATION OF KINETIC PARAMETERS

2.8.1 Yield coefficients

The observed yield coefficients for growth and biotransformation were calculated from the following general equations (Wang et al., 1979). The yield coefficient calculation was based on the particular substrate selected, either glucose or benzaldehyde.

(1) Cell growth in batch

Cell yield (based on glucose) Ax Yx/s1 = --(g/g) (Eq. 2.5) As1 Ethanol yield (based on glucose) Ap1 Yp1/s1 = --(g/g) (Eq. 2.6) As1 (2) Molar conversion yield for biotransformation

Product yield (based on benzaldehyde) Ap Yp/s2= --- (mole/mole) (Eq. 2.7) As2 p=p2, p3 Product yield (based on pyruvate) Ap Yp/S3 = ---- (mole/mole) (Eq. 2.8) Chapter 2 84

p=p2, p4

Acetoin yield (based on pyruvate) Aps (mole/mole)* (Eq. 2.9) As3 *The theoretical value was calculated based on following reaction:

2 2 pyruvate ---- ,2 > acetoin Definitions of the terms are as followings: s1: glucose s2: benzaldehyde s3 :pyruvate p: product p1: ethanol p2: L-PAC p3: benzyl alcohol p4: acetaldehyde

p5: acetoin x: biomass

2.8.2 Biotransformation Kinetics

2.8.2.1 Kinetic parameters in batch

The kinetic parameters in biotransformation were evaluated using following equations:

(1) Volumetric substrate consumption rate -ds -As Qs= = (g/L/h) (Eq. 2.10) dt At S=Sz, S3

(2) Volumetric production rate dp Ap Qp= = (g/L/h) (Eq. 2.11) dt At p=p2, P3. p4 and Ps Chapter 2 85

(3) Specific substrate consumption rate

1 ds qs = - (g/g/h) (Eq. 2.12) X dt S=S2 (4) Specific production rate

1 dp qp=-·- (g/g/h) (Eq. 2.13) x dt p=p2, p3 Definitions of the terms are as follows: s: substrate s2: benzaldehyde s3 :pyruvate p; product p2: L-PAC p3: benzyl alcohol

p4: acetaldehyde ps: acetoin x: biomass

2.8.2.2 Kinetic evaluation in continuous process

(1) Kinetic parameters with immobilized cells Dilution rate: F D=-- (h-1) (Eq. 2.14) V Specific substrate consumption rate: D(so- s) (g/g/h) (Eq. 2.15) X

s0 = inlet of Si, s2 s= outlet of s1, Sz Specific production rate: D.p 9p= --- (g/g/h) (Eq. 2.16) x p=p1, p2 and p3 Chapter 2 86

(2) Kinetic parameters with immobilized PDC enzyme

Space time: V 't= (h) (Eq. 2.17) F Substrate consumption rate -ds (so - s) - = (g/L/h) (Eq. 2.18) dt 't So= inlet of s2, S3 s=outlet of s2 S3 Productivity p p = (g/L/h) (Eq. 2.19) 't p=p2, p4 and Ps Definitions of the terms are as follows: D: dilution rate F: flow rate 't: space time V: working volume s: substrate s1: glucose s2: benzaldehyde s3: pyruvate p: product p1: ethanol p2: L-PAC p3: benzyl alcohol p4: acetaldehyde p5: acetoin Chapter 3 87

CHAPTER 3

KINETIC EVALUATION OF BIOTRANSFORMATION OF BENZALDEHYDE TO L-PHENYLACETYLCARBINOL (L-PAC) BY IMMOBILIZED CANDIDA UTlLIS.

3.1 INTRODUCTION

The study of fermentative enzymes profiles of Candida utilis and evaluation of its biotransformation of benzaldehyde to L-phenylacetyl­ carbinol (L-P AC) has been conducted to establish its potential as a possible alternative to Saccharomyces cerevisiae. Most of the previous research on the biotransformation of benzaldehyde to L-P AC has been concentrated on various yeasts and also on Zymomonas mobilis which contains relatively high levels of cytoplasmic pyruvate decarboxylase associated with high rates of ethanol production. The significance of S. cerevisiae is largely due to its capacity to withstand relatively high concentrations of benzaldehyde and produce high levels of L-PAC. However, Z. mobilis with even higher PDC levels produced significantly less L-PAC due to the low affinity of its PDC toward benzaldehyde. Considerable success has been achieved in the use of S. cerevisiae for the production of L-P AC. However the non-stoichiometric conversion of benzaldehyde to L-PAC has been a significant drawback in the commercial application of this process. The main reasons have been attributed to high toxicity of benzaldehyde (substrate) and the side reaction of benzaldehyde reduced to benzyl alcohol (Mahmoud et al., 1990a). In the present study, an immobilized cell system rather than free cells was employed in order to reduce the toxic effects of benzaldehyde, based on a report by Mahmoud et al. (1990a). These authors achieved 65 % molar conversion of benzaldehyde to L-PAC with immobilized S. cerevisiae in contrast to about 15.3 % with free cells. The higher conversion associated with immobilization was attributed to a reduction in the toxic or inhibitory effect of benzaldehyde resulting from diffusional gradients inside the beads. The choice of C. utilis in the present study is based on the fact that it is a Crabtree negative yeast which is likely to be more amenable to physiological control than S. cerevisiae. C. utilis is an oxygen sensitive Chapter 3 88

yeast so that the generation of biomass, induction of fermentation enzymes and production of pyruvate are predominantly affected by available oxygen which can be readily controlled by aeration rate and agitation speed.

The aims of this initial study are the following: (1) to study the effect of culture conditions on fermentative enzyme profiles in C. utilis; (2) to compare the key enzyme profiles both in free and immobilized cells under various conditions; (3) to evaluate of the kinetic parameters for L-PAC production with immobilized C. utilis in both fed-batch and continuous processes.

3.2 ENZYME PROFILES OF PYRUV A TE DECARBOXYLASE (PDC), TOTAL ALCOHOL DEHYDROGENASE(ADHT)AND AROMATIC CINNAMYL ALCOHOL DEHYDROGENASE (ADHc1N) FOR C. UTILIS

An investigation of the key enzyme profiles of pyruvate decarboxylase (PDC), total alcohol dehydrogenase (ADHr) and aromatic cinnamyl alcohol dehydrogenase (ADHcin) which are enzymes associated with ethanol production, was carried out in order to determine the optimal conditions for biotransformation. PDC is the key enzyme responsible for L­ PAC formation, while ADHr and ADHcin may be involved in by-product (benzyl alcohol) formation. These enzyme activities are likely to be dependent on the rate of substrate consumption, source of carbon and the extent of fermentative metabolism. In this study, the effects were examined of aeration rate, initial concentration of glucose and pattern of supplementary feeding of glucose on fermentative enzymes profiles in free and immobilized cells. Chapter 3 89

3.2.1 Time course and enzyme profiles of PDC, ADHT and ADHcin with constant air flow.

C. utilis was cultivated on 60 g/L glucose based medium in a 2 L LH fermenter with 0.6 vvm constant air flow rate. As shown in Figure 3.l(a, b), C. utilis gave a high yield of biomass under respiratory conditions within the first 6 h. In this period, biomass was predominantly produced with a 0.45g/ g yield (Yx/ s glucose). Ethanol production was maintained at a low level. It was found that when C. utilis was grown on glucose, this yeast possessed certain activities of PDC, ADHr and ADHcin no matter whether ethanol was produced or not. These findings are in good agreement with those from a previous report (Franzblau and Sinclair, 1983). These authors observed that C. utilis, when grown aerobically on glucose medium, always possessed some PDC activity irrespective of culture age. However, when the yeast was grown on non-fermentable carbon compounds (e.g. citrate, ethanol and acetate), PDC activity was observed only after anaerobic induction in glucose containing medium. As the cell density increased with culture age, the constant air flow rate of 0.6 vvm could not supply sufficient oxygen for the requirements of the increased cell population. As a result of the limited available oxygen, fermentation associated with ethanol and pyruvate production was initiated. Once ethanol production commenced, the growth rate gradually decreased. Generally, a 1.5 to 2 times higher level of ADHr compared to PDC was observed over the whole time course. It seems that higher ADHr activity can immediately convert acetaldehyde to ethanol coupled with endogenous NADH+H+ to NAD+. These enzymatic reactions may prevent accumulation of free acetaldehyde (via decarboxylation of pyruvate by PDC), and diminish the possible toxic effect of acetaldehyde. Most of enzyme activities involved in fermentation reached maximum values prior to complete glucose utilization, and then declined following glucose depletion. The activities of PDC, ADHr and ADHcin reached peak levels of 0.27, 0.44 and 0.17 unit/mg protein, respectively after 15 h cultivation. After glucose exhaustion, PDC and ADHr activities decreased with time in a period when C. utilis was likely to adapt to the utilization of ethanol as an alternative carbon and energy source. The estimation of RQ (respiratory quotient) values with on-line computer­ linked measurement of 02 and CO2 concentration provided an insight into the changes in the physiological state of the cells (Figure 3.1 (c)). Chapter 3 90

(a) 70.0 C Glucose • Biomass 60.0 ...... 0 Ethanol c~ • Pyruvate 50.0 -..,J -bO a"'a -c:: 40.0 ~ ....0 a\ ·-~ ....i., c:: (IJ 30.0 V c:: ~ 0 u 20.0 ..-.-• o--0 10.0 .....,.,,,-. 0 __.-• ~ --• _o--o

2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h) (b) 1.0 ------, o ADHT • ADHcin c POC c=' 0.8· ....(IJ 0 i., 0.. bO 0.6 - E :::i o-0 -o-,,~ 0.4- ....>. 0.,,,...-0--- .> ,,,,,0/ .... / _.--o---°-ccJ-c--c ·­V _-o ___,..c < 0.2 · ...--- ..,.. -o-" _...... c ,_./d__ ._. ------•••--· 1,- 0 -c ---· -- 1 on -----I I I I I I I I I 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h)

Figure 3.1 (a) Growth of C. utilis on 60 g/L glucose based medium with 0.6 vvm aeration rate; (b) fermentative enzyme profiles. Chapter 3 91

(c) 25.0 ------, QJ ~ !U -> OI 20.0 - ~ ... . ~ -Q . -fll 15.0 - !U bO . ·-->< QJ C 10.0 - COi ·-C 0 __ .-...... ,,· ·-!U .,· -i., C 5.o- .,.,.,·" -QJ _., V C - .-----·-· RQ u0 .,,,,:.-·::P"'------­ --.---·-·-·--·------~-=--~-· 0.0 ...... -::.a··;;;.;·..;;.-·,...,--·- ....- ...... ,---.--r--,....-"T,---r--r--,....---r, ...... --,,,--r--....-,~--,,r--r----1 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h)

Figure 3.1 (c) RQ values computed from analysis of 02 and C02 concentrations in exit gases as a function of culture time. Chapter 3 92

The RQ values higher than 1.0 indicate the initiation of fermentation, with ethanol production corresponding well with increase in C02 evolution and RQ values. With extension of culture time, the high oxygen requirements induced the more fermentative conditions with an increase in RQ values up to 2.2. A volumetric rate of 1.5 g/L/h ethanol after 18 h cultivation was much higher compared to 0.9 g/L/h of growth rate. At the end of cultivation, a carbon balance based on 60 g/L glucose was estimated and showed that 54 % glucose was used for biomass (16.2 g/L), 40 % of glucose for ethanol (12 g/L), and 6 % on pyruvate and other minor products.

3.2.2 Effects of aeration rate and agitation speed on enzyme profiles.

The transition from fully respiratory to respiro-fermentative conditions was accomplished by reducing the aeration rate from 0.6 vvm to 0.2 vvm and agitation speed from 1000 to 500 rpm after 8 h cultivation. As shown in Figure 3.2, there were significant increases in both ethanol production and the activities of fermentative enzymes. Changes in the aeration rate and agitation speed gave rise to an increase in PDC activity to 0.31, ADHr to 0.52, while ADHcin was nearly the same at 0.2 unit/mg protein as for a constant aeration rate. These peak levels of activity were reached after 12 h incubation. When the available oxygen falls to levels below that required to metabolize pyruvate via the TCA cycle, pyruvate cannot be metabolized purely oxidatively, so conversion via PDC catalytic activity occurs. Pyruvate, therefore, can be metabolized via two metabolic routes. Two enzymes, pyruvate dehydrogenase (PDH) and PDC, function at the beginning of these two pathways. The Km value of PDH for pyruvate is approximately ten times lower than that of PDC (Holzer and Goedde, 1957), which implies that the oxidative pathway is preferentially saturated with pyruvate. The high glycolytic rate compared with respiratory activity results in significant fermentative activity when oxygen limitation occurs. This respiratory capacity with limited oxygen represents a metabolic bottle-neck. The metabolic change to respiro-fermentation give rise to increased glucose metabolism and enhanced overall energy production. Chapter 3 93

(a) 70.0 ------, C Glucose • Biomass 60.0 o Ethanol • Pyruvate 3 50.0 eh -C o 40.0 .... ·-le ....i-, ~ 30.0 V C 0 U 20.0

10.0

2.0 4.0 6.0 8.0 10.0 12.0 14.0 Time (h) (b) 1.0------, o ADHT 0.9 • ADHcin a PDC 0.8 -C ·-....Q,/ 0 0.7 i-, 0.. bO 0.6 E 0 .....--0-0 ~ 0.5 C ::s 0/ 0.4 0/ ...... >. > 0.3 /. ---a-a .... 0 a ·-V o-- 0---- c,C/"' < ---- 0.2 a-C~_,,.., e-• ------·-- 0.1 ====:.:.:-__a-----~ • 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 Time (h)

Figure 3.2 (a) Growth of C. utilis on 60 g/L glucose based medium with changing aeration rate from 0.6 vvm to 0.2 vvm and agitation speed from 1,000 to 500 rpm after 7 h cultivation; (b) fermentative enzyme profiles; Chapter 3 94

(c) 25.0 ------, QJ ~ COi (U ,._ - , '• > ,· '· OI 20.0 • 0:: ...... i',...... ,· -~0 ··············-... ,· -~ 15.0 · .~::,_·.:: ...... ?.~.... . b0 , .... / ·x ,· QJ ' .,. C ,· ·- 10.0 · , C I 0 I RQ .... I ·­(U ,., __ ---.... .,, ~ ----- .... I - ,....,_ _-,, C ;- - QJ s.o · ~ V /.· C ' / //. u0 ~.,·' ..... :. =:.-::. ---.--· ---· --.- --._-· __,_,_, 0.0 -----.-,-----,...... -- .....,--.--..-, -.....----.,....-...... ---r, -.....----1 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 Time (h)

Figure 3.2 (c) RQ values computed from analysis of 02 and C02 concentrations in exit gases as a function of culture time. Chapter 3 95

In the present experiment with a change to partial fermentation conditions, 40 g/L glucose was consumed within 4 h compared to 20 g/L for 8 h for fully aerobic respiration. It is clear that PDC activity is thus linked to metabolic regulation in yeast during glycolysis, and increased PDC activity is associated with increased rate of catabolic flux in the anaerobic utilization of glucose.

3.2.3 Effect of initial glucose concentration on fermentative enzyme profiles.

Considerable research has been focussed on the effect of glucose on both respiratory and fermentative enzyme activities. For S. cerevisiae, the induction of respiratory enzymes by oxygen and repression of these enzymes by glucose is a well known mode of catabolic regulation. The Crabtree negative yeast, C. utilis is known to be not sensitive to high concentrations of glucose, but can be affected by available oxygen. Thus, it does not produce significant amount of ethanol when growing on high glucose media in fully respiratory conditions. The PDC enzyme activity in C. utilis has been shown to be dependent on the growth conditions. Although C. utilis is insensitive to high concentrations of glucose, changes in PDC activity with increased glucose have been reported for other Candida yeasts. Hommes (1966) demonstrated that an increase in glucose from 0.6 to 20 % led to a remarkable increase in PDC activity of up to 100 fold in Candida parapsilosis. These studies prompted similar investigations with C. utilis in the present study. Comparison of the results at 60 g/L and 90 g/L glucose showed that at 90 g/L, the lag phase of growth was prolonged, and the levels of fermentative enzyme were significantly increased as shown in Figure 3.3. The activities of PDC, ADHr and ADHcin were enhanced to 0.45, 0.68 and 0.26 unit/mg protein respectively, after 16 h cultivation. The rate of carbon dissimilation via glycolysis is increased at the higher glucose concentrations although some substrate inhibition occurs. Ethanol production is due primarily to an overflow reaction at the pyruvate level when the oxidative pathway is limited by available oxygen. Ethanol production rate is determined by the difference between glucose uptake rate and its respiratory metabolism. Chapter 3 96

(a) 100.0 ~------. C Glucose 90.0 • Biomass o Ethanol 80.0 • Pyruvate -,_J 70.0 -~ -C 60.0 ...0 ·-...~ 50.0 C -QJ V 40.0 C 0 u 30.0

20.0

10.0

2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h) (b) 1.0 ...------, o ADHT 0.9 - • ADffcin· a PDC c 0.8- 0.1 -0... c.. ~ E -·-- -.....>. ·;;: ......

0.0 -----.-----,--.-,------.-.---,--.-,----.,---..-,~-,,---,.---i 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h)

Figure 3.3 (a) Growth of C. utilis on 90 g/L glucose based medium with changing aeration rate and agitation speed after 7 h cultivation; (b) fermentative enzyme profiles. Chapter 3 97

Fiechter et al. (1981) reported that fermentation in Candida tropicalis was observed only with high concentrations of glucose over 75 g/L under aerobic conditions. Induction of fermentation resulted in enhanced PDC activity and ethanol production. In 90 g/L glucose medium, it was found that a higher glucose consumption rate resulted in higher ethanol production rate than at 60 g/L. For 90 g/L glucose medium, ethanol production rate increased to nearly double that at 60 g/L while biomass was similar to that at 60 g/L. As a result, 25.1 g/L of ethanol and 16.7 g/L of biomass were obtained from 90 g/L glucose, and the activities of PDC, ADHr and ADHcin were increased about 1.5 times.

3.2.4 Effect of pulse feeding of glucose on enzyme profiles

For the results shown in Figure 3.4, respiratory metabolism (RQ=l) was maintained for 8-9 h to obtain a high biomass for biotransformation, and then a switch was made from aerobic respiration to fermentative growth by reducing agitation and aeration rate as in previous experiment. This procedure resulted in a decline in the oxygen concentration and an increase in the RQ value. During the aerobic growth, an overall glucose consumption rate (ds/dt) of 2.8 g/L/h and biomass yield 0.45 g/g glucose were obtained, while the levels of PDC, ADHr and ADHcin increased gradually even though ethanol production was negligible. After the shift to respiro­ fermentative conditions, glucose consumption rate significantly increased. Before the initial glucose was completely exhausted, pulse feeding of a supplement containing glucose and yeast extract (approximate concentrations 30g/L and 5g/L, respectively) was initiated. This resulted in enhanced PDC and ADHr activities, and increased production of ethanol and pyruvate. The activities of PDC, ADHr and ADHcin reached peaks of 0.59, 0.83 and 0.19 unit/mg protein, respectively. However, the increase in glucose did not lead to an increase ADHcin activity (an aromatic alcohol dehydrogenase). After entering the late stationary phase, there was a steady decline in the PDC and ADHr activity even in the presence of excess glucose. Pulse feeding of glucose proved to be successful with enhanced induction of fermentative enzymes in C. utilis and associated products. Chapter 3 98

(a) 70 C Glucose Biomass 60 • 'C 0 Ethanol Pyruvate 50 • -.,.J bO ~ -c:: ·-....0 40 ....a..~ c:: QJ 30 c::V 0 u 20

10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Time (h) (b) 1.0 o ADHT 0.9- • ADffcin c PDC 0.8- ~-0-=------/o° 0 0 0 ' 0.7 · /co 0.6- o o c ... a-~ 0/ .,....D""' o.s- / c_...C c:: ·-::s o/o ~cc 0.4· / a ->.. /4a .... /40/c,c cc ·-> .... 0.3- /. -/1 ·-V < 0.2 _)_o __o , /c -.-.-e-.-e-e-e-e-•-.---. /C ...-• e / C ..... 0.1· ...,...... c...... - ...... : ·-· J g-•--- •

0.0 I I I I I I " I • I I I I I I 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Time (h)

Figure 3.4 Effect of pulse feeding of glucose on (a) kinetics of growth; (b) fermentative enzyme profiles. Chapter 3 99

3.3 IMMOBILIZATION OF C. UTILIS CELLS AND ASSOCIATED ENZYME PROFILES

It has been shown that additional feeding of glucose prior to its depletion extended cell growth and enhanced PDC activity. When the level of PDC reached a maximum, cells were harvested, and resuspended into sodium alginate (3 % w /v). 150 g wet cells (approximate 30 g dry weight)/300 mL solution were prepared for immobilized beads. This preparation was extruded into 2% CaCl2 solution through a peristaltic pump. In order to stabilize the immobilized beads, they were maintained for 1-2 h in fresh 2 % CaCl2 solution containing 3 % glucose. Then, the immobilized beads were reintroduced into their own former supernatant with further supplementation of glucose and yeast extract.

The kinetics of fermentation for the immobilized cells are shown in Figure 3.S(a) with most of the added glucose being converted to ethanol instead of biomass. This resulted in higher initial ethanol production rates than with free cells which gradually decreased as more ethanol accumulated. Measurement of the fermentative enzymes in immobilized C. utilis as described in Figure 3.S(b) indicated that the levels of PDC and ADH1 were increased with glucose feeding, while ADHcin remained at a similar level. The highest activities of PDC, ADHr and ADHcin were 0.57, 0.91 and 0.24 unit/mg protein, respectively after 12 h incubation. Comparison of enzyme profiles for free and immobilized cells indicated that ADH1 activity was significantly higher in immobilized cells than in free cells. This may be due to the oxygen diffusional barrier imposed as a result of immobilization, which would tend to favour more fermentative conditions. Additionally, the reduced agitation speed (200 rpm) designed to avoid excessive mechanical stress would also contribute to enhanced ADH1 activity and ethanol production. Chapter 3 100

(a) 70 ------, c Glucose o Ethanol 60 • Pyruvate

50 ~ o o--O C: /0 ....0 40 ·­!U ....lo< 0/ C: QJ c'-a V 30 \ • ..------v C: u0 X \ 20 ~ c"-.

10 \. ~c~ "' c--c....._ -•-•-•-•-•--•--. I i-a-• 0 -+--.-"""'T"---.---r----.r----r---r-~""T""'"""'T"'"""'T"~----,,--,.-----r--r-""T""'"""T"""""'T"'---1 0 2 4 6 8 10 12 14 16 18 20 Time (h) (b) 1.0------, o ADHT 0.9 • ADffcin c PDC 'c 0.8 ·-....QJ 0 0.7 lo< c.. 00 0.6 a--a-0 -c-a-c- E .,,,.,- 0 -::i 0.5 ·-c: ~ ::s ~a ....>. 0.4 / ·-....> 0.3 ·­V < _.!!--~------·-·-·- . 0.2 --- • -·-·-· 0.1

0 2 4 6 8 10 12 14 16 18 20 Time (h)

Figure 3.5 (a) Kinetics for immobilized cells; (b) fermentative enzyme profiles with pulse feeding of glucose. Chapter 3 101

3.4 BIOTRANSFORMATION OF BENZALDEHYDE TO L-PAC BY IMMOBILIZED CANDIDA UTILIS CELLS

Immobilized living and growing cells of C. utilis have some self­ regenerating properties as a catalytic system, and can also catalyse efficient biotransformation involving coenzyme regeneration. While the immobilized cells can be growing, resting or dead, the enzymes are maintained in their native state (in a cellular matrix) which can provide favourable conditions for their stability. With the use of whole-cell preparations, time-consuming and costly enzyme isolation and purification are not necessary, but the specific activity is generally lower than that for immobilized purified enzyme, except when enzyme activity is increased 'in situ' in immobilized living cells. Whole cell immobilization frequently improves operational stability in a continuous process due to a reduction in toxic effect, however overall productivity is relatively low. The potential disadvantages are that cells contains numerous catalytically active enzymes, which may catalyze side reactions. In addition, there are diffusional barriers in immobilized beads which may hinder benzaldehyde access to the active sites of the enzyme. Major difficulties may also be encountered with oxygen transport, within the beads to meet the aerobic requirements of C. utilis. In the present study, the PDC activity in C. utilis was enhanced by controlling available oxygen, and then harvested cells were immobilized in 3 % calcium alginate prior to biotransformation. Comparisons can then be made between L-P AC production with free cells and immobilized cells. The effects of benzaldehyde levels, RQ values on L-P AC formation and the characteristics of operation of a continuous L-P AC process were evaluated. An assessment can be made also whether or not a higher productivity can be achieved in continuous operation than in a batch process, when a relatively high cell concentration can be maintained in the continuous bioreactor. Chapter 3 102

3.4.1 Comparison of biotransformation by free cells and immobilized cells with various concentrations of benzaldehyde

Biotransformations were carried out with free and immobilized cells in 500 mL flasks at 20°C and 180 rpm after PDC enzyme was induced with pulse feeding of 30 g/L glucose. Results for various concentrations of benzaldehyde (BZ) on L-P AC formation are summarized in Table 3.1.

Table 3.1 Comparison of biotransformation products by free cells and immobilized cells after 16 h incubation with various initial concentrations of benzaldehyde.

Free cells Immobilized cells BZ Residual BA L-PAC Residual BA L-PAC (mM) BZ(mM) (mM) (mM) BZ (mM) (mM) (mM) 5 nil 4.8 0.1 nil 4.8 0.05 10 nil 9.1 0.6 nil 9.3 0.4 15 nil 11.7 2.9 nil 12.3 2.5 20 nil 13.8 5.2 nil 15.2 4.3 25 0.5 15.2 8.2 nil 16.5 7.6 30 1.2 16.0 12.1 nil 17.7 11.9 40 1.3 17.0 21.5 nil 19.3 20.0 50 2.1 18.2 28.5 0.5 22.5 26.5 60 11.2 19.2 27.3 2.1 23.8 35.5 70 30.5 16.3 20.1 3.8 26.5 37.5 80 56.1 10.7 10.8 22.1 22.1 32.5 90 73.3 8.2 6.3 40.1 20.0 25.1 100 92.1 1.5 1.6 55.2 15.6 23.2

With both free and immobilized cells, below 30 mM BZ, benzyl alcohol (BA) was preferentially produced instead of L-PAC. However, L­ p AC was preferentially formed above 40 mM BZ. Higher concentrations of BZ were accompanied by higher L-P AC formation until inhibition occurred. Increased BZ was accompanied by an enhanced molar fraction of L-P AC formation to BA. Chapter 3 103

For free cells, the highest biotransformation efficiency was 57 % yield based on BZ. This was achieved at 50 mM benzaldehyde at which 28.5 mM L-PAC and 18.2 mM BA were produced. Above 50 mM BZ, L-PAC and BA production were significantly inhibited by BZ. An increased resistance of immobilized cells toward BZ is illustrated clearly in this Table 3.1. Above 60 mM BZ, immobilized cells exhibited their potential catalytic ability. While the highest conversion yield (Yp/s) of 58.8 % (mole/mole) was achieved at 60 mM BZ, L-PAC reached peak value of 37.5 mM at 70 mM BZ. In the range of 60 to 80 mM of BZ, L-P AC and BA production showed only small variations, however residual BZ increased in accordance with the initially the higher BZ. Above 80 mM, L-P AC production with free cells was almost completely inhibited, while with immobilized cells L-PAC production still proceeded. Even at 100 mM BZ, 23 % BZ was converted to L-P AC with immobilized cells. In contrast to L-PAC, BA formation gradually decreased with increasing BZ. This result was found with both free and immobilized cells. These results support those from a previous report (Long and Ward, 1989b). The authors found that PDC in S. cerevisiae was more resistant to denaturation by BZ than alcohol dehydrogenase. In their report, while only 33 % residual ADH activity remained at 3.0 g/L BZ (viz. 28.3 mM), the PDC enzyme maintained almost its complete activity. Moreover, 87 % PDC activity was still maintained at 7 g/L BZ (viz. 66 mM) although ADH activity was completely lost. As previously shown in this Chapter, C. utilis has generally 1.5-2.5 times higher levels of ADHr than PDC under fermentation conditions. Thus, although ADHr is significantly inhibited by BZ, at low concentrations of BZ, partial ADHr activity indicates that the enzyme could still be involved in benzyl alcohol production. In the range of 5-40 mM BZ, the reduction of BZ to BA was a predominant reaction. However, at higher concentrations of BZ, ADHr or possible other oxido-reductases involved in BA formation, were more significantly inhibited by BZ. Therefore, L-P AC production via PDC became predominant. It was apparent also that BA production with immobilized cells was higher than with free cells over the range of BZ concentrations. These results were supported by previous data in which relatively high levels of ADHr occurred in immobilized C. utilis possibly due to diffusional oxygen transfer limitation which could lead to high BA production. From these studies, it was suggested that maintenance Chapter 3 104

of an optimum level of BZ is necessary to promote L-PAC formation as well as reducing BA formation.

3.4.2 Effect of benzaldehyde concentration on L-PAC production by immobilized cells under controlled conditions

Investigations of the effect of BZ (molecular weight, 106) levels on L­ PAC (molecular weight, 150) formation were carried out with four different concentrations in a 2 L LH fermenter following a period of adaptation for 2- 4 h. A short acclimatisation was used with addition of 0.8 g/L/h of BZ for this period. During this time, it appeared that the cells adapted to the toxic substrate. This acclimatisation was provided to minimize viability and/ or enzyme activity loss following extended exposure to highly toxic substrate (Wang 1993). After this period, biotransformation kinetics were evaluated at various feeding rates of BZ until L-P AC concentrations reached their peak values. To maintain relatively constant BZ levels, samples were taken every hour, and then BZ concentrations were measured rapidly by gas chromatography (GC). Through this analysis, BZ feed rates were adjusted to maintain various levels of BZ in the range of 0.8-4 g/L through changing the set point on the syringe pump. At 0.8 g /L BZ as shown in Figure 3.6(a), 7.1 g/L L-PAC and 6.4 g/L BA were achieved with relatively low molar fraction of BZ (46.6 mM) converted to L-PAC compared with BA (59.3 mM). However, at 1.5 g/L BZ L-P AC was significantly increased up to 9.5 g/L while BA formation was nearly similar level at 6.1 g/L (Figure 3.6(b)). The highest level of L-PAC (viz. 10.8 g/L) was achieved at 2 g/L BZ level as shown in Figure 3.6(c) at which the production of L-PAC corresponded to a significant improvement over previous levels. The production of BA was similar to that obtained at 0.8 and 1.5 g/L BZ levels. At 4 g/L BZ as shown in Figure 3.6(d), while BA production was drastically depressed to 4.6 g/L, 7.2 g/L L-P AC was achieved within the relatively shorter period of 16 h. Chapter 3 105

(a) 12.0 0 BA 11.0 • L-PAC 10.0 9.0 ,_J bO- 8.0 -C: 7.0 0 .... ./o--o·-· ·-ru 6.0 ....'"' C: A/ Q,I 5.0 u C: 0 4.0 u ...... / 3.0 ,.....~.,o fY° .,...... ,, 2.0 .~3::..--0 ~~ 1.0 a~ 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h) (b) 12.0 0 BA 11.0 • L-PAC 10.0 .__. 9.0 ,_J -bO 8.0 -C: 0 7.0 .... /.O-o /e ·-i,: 6.0 ....'"' .~ 0.,,,..- C: Q,I 5.0 e/ ./'6' u ./ 0 C: 0 4.0 ,( 0/o u / o-' 3.0 . . / .7/.0 2.0 *oo 1.0 0~ 0.0 a.::::::• 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h)

Figure 3.6 Biotransformation kinetics with immobilized C. utilis on the level of (a) 0.8 g/L BZ; (b) 1.5 g/L BZ; Chapter 3 106

(c) 12.0 0 BA 11.0 ...... - ·---. • L-PAC 10.0 ...... / 9.0 /. ..,J • -bO 8.0 -c:: 7.0 0 ...o--0 0 0 .... • o,-9.,o ·-~ 6.0 I ....lo< I/(;,- c:: QJ 5.0 c::V u0 4.0 ·) 3.0 2.0 1/o~ 1.0 /" 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h) (d) 12 o BA 11 • PAC 10

9 -..,J bO 8 -C 7 .---. ....0 •~ ·-II: 6 ....lo< ./ c:: QJ 5 -o-o-O-o V / C u0 4 3 //..r·-· 2 al 1

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h)

Figure 3.6 Biotransformation kinetics with immobilized C. utilis on the level of (c) 2.0 g/L BZ; (d) 4.0 g/L BZ after 4 h initial adaptation period. Chapter 3 107

Comparison of results for all of the four levels of BZ (Figure 3.7) indicated that a 2 g/L BZ concentration appeared to be optimal for L-PAC production since at this level, L-P AC production was maximized, and the production of BA reduced. From these results, it established also that although high levels of BZ promoted L-P AC production rate and depressed BA formation, they caused significant loss in catalytic activity of immobilized cells in the long term. At 2 g/L BZ, maximum L-P AC production rates could be maintained for up to 9-10 h before the rate began to decrease. Besides the inhibition effect of BZ, it seems that BA accumulation could also strongly inhibit L-PAC formation. L-PAC formation did not continue when BA was above 6-7 g/L (as was evident also in the other levels of BZ). This inhibition might be due to that degeneration of catalytic activity resulted from continuous contact with both toxic substrate and products.

3.4.3 Effect of RQ value on L-PAC production during biotransformation

As mentioned earlier, the respiratory /fermentative growth characteristics of C. utilis are influenced more by available oxygen than by available sugar. The metabolism of C. utilis was shown previously to be significantly affected by different aeration rates. The RQ value is an indicator of pyruvate flux through the TCA cycle, or its further reduction to ethanol via acetaldehyde. The biotransformation of BZ to L-P AC can be related also to the production of C02 from decarboxylation of pyruvate to acetaldehyde, thus the effect of RQ value on L-PAC formation was conducted with 2 g/L BZ level. Table 3.2 shows that a low aeration rate of 0.3 vvm resulted in a high RQ value. Even though this was accompanied by adequate pyruvate accumulation, a high level of BA was formed presumably due to relatively high levels of ADHr (and/or other oxido-reductases) as well as sufficient NADH under more fermentative conditions. At the high aeration rate 1.0 vvm, a low RQ value resulted, with consequent lower accumulation of pyruvate and low level of PDC, in lower L-PAC production. Chapter 3 108

12.0 BZ level (g/L) 11.0 ...... -o..c-a 0 0.8 .P 10.0 ~ ...... • 1.5 9.0 ,o C 2.0 ..J L -eh 8.0 • 4.0 -_.!""..I:8 -C 7.0 j" ·- 0 ~--~. ·-IU 6.0 -....i.. /a,1' ./

Figure 3.7 Effect of various levels of BZ on L-PAC formation as a function of time. Chapter 3 109

Table 3.2 Effect of aeration rate on the biotransformation of benzaldehyde to L-PAC.

Aeration Max cone. of L-PAC Benzyl alcohol Molar ratio rate RQrange pyruvate (vvm) (g/L) (g/L) (g/L) (L-PAC/BA)

0.3 12-100 4.2 10.7 6.7 1.15 0.6 7-20 3.8 12.4 4.9 1.82 0.75 5-7 3.5 12.5 4.7 1.91 1.0 2-4 2.2 10.1 5.2 1.40

From Table 3.2, it can be concluded that respiro-fermentative conditions for which the RQ value was maintained between 5 to 20, were the most favourable to maxmize L-P AC and minimize BA formation.

3.4.4 Time course of biotransformation with immobilized C. utilis cells

As mentioned in Section 3.2.3, when C. utilis was grown on 90 g/L glucose medium, the PDC and ADHr activities were significantly increased compared with growth on 60 g/L glucose medium. In the present study, to achieve enhanced PDC levels, free cells were cultivated on 90 g/L initial glucose medium, and this was followed by 30 g/L glucose pulse feeding before glucose depletion and cell harvesting. Immobilization was carried out according to previous procedures. Biotransformation was maintained at 2 g/L BZ level with 30 g/L glucose pulse feeding and in the optimum range of RQ values. As shown in Figure 3.8 (b ), there was a significant increase in L-P AC production up to 15.2 g/L. This compares with 10.8 g/L achieved previously with 60 g/L initial glucose (See Section 3.4.2), while maintaining BZ at similar levels. Presumably, this increase can be ascribed to an association of various factors viz. pulse feeding of glucose during cell growth to facilitate pyruvate production, with 6.2 g/L pyruvate accumulation before the biotransformation phase (Figure 3.8(a)) compared with 3-4 g/L pyruvate Chapter 3 110

with initial 60 g/L glucose medium. The more favourable RQ with optimum aeration rate was likely also to contribute to increased L-PAC formation and the slightly higher PDC was likely also to have a positive effect. Furthermore, a high L-PAC formation rate also resulted in reduced BA formation. Profiles of enzyme activities with increasing reaction time showed that the PDC activity seems to be not more stable than ADHr or ADHcin (viz. final activities of PDC, ADHr and ADHcin were 0.12, 0.55 and 0.1 unit/mg protein, respectively) compared with their activities at initial biotransformation (viz. initial activities of PDC, ADHr and ADHcin were 0.7, 1.0 and 0.35 unit/mg protein, respectively). However, some residual PDC activity could still be detected in the cells with residual glucose left in reaction mixture, although L-PAC production had ceased (Figure 3.8(c)). This cessation of L-PAC production may be ascribed to depletion of pyruvate rather than loss of PDC activity. C. utilis cells at the end of biotransformation appeared to have completely lost metabolic activity (viz. monitored by CO2 evolution as shown Figure 3.8(e)) due to accumulating BZ and biotransformation products. Chapter 3 111

(a) 50.0 ------D Glucose 45.0 o Ethanol • Pyruvate 40.0

-...J 35.0 be -t::: 30.0 0 ·-~ 25.0 ....~ t::: ~ 20.0 t::: 0 U 15.0

10.0

5.0 ---;--~...... ---.---. D 0.0 +-...... r--T--.---~.,...... ,,...... --,-._~T""""""ir--r~...... :~;::::;=--~------,-~ -· 0 2 4 6 8 10 12 14··-- 16 18 20 22 24 Time (h) (b) 18.0 ------, • BZ 16.0 o BA • L-PAC 14.0 -...J be 12.0

-t::: .s.... 10.0 I<: ....""' t::: 8.0 ~ V t::: 0 6.0 u ,.-(/'__.o-o-o

4.0 / ,.,.,.,0.,,,,,..0 .,.. .L'. 0~ __,,,.. • 2.0 /~----g~•--•-•-·---·

o.o .....~·::::::---~._:.,.__..-..__,,---r---r---r--.--r--..--.--.__,.---r- ...... --r--r--..--,.....,.--.--r---r--1 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 3.8 Biotransformation time course with immobilized cells: (a) glucose consumption, ethanol and pyruvate production; (b) biotransformation kinetics; Chapter 3 112

(c) 1.2 0 ADHT 1.1 • ADHcin 1.0 -o-0-Cb CJ PDC C .,,0 ~ - 0.9 ·-Cl .... 0..___0 0 ~ 0.8 '-0 0 0.. ---.:::.g...._0 00 0.7 -...... __0 E __.,....o--"-'ll~ '- :.::i 0.6 0 C ::, 0.5 -j>,...... 0.4 c'-.....1:1.._CJ > ·-.... ~ ·-V 0.3 -.-•-. ~ CJ...... __ < • •--• • CJ...... _ 0.2 -- ~CJ 0.1 ··~• • 0.0 ______...... ,...... ,,...... ______....----- ...... ----.---.----1

0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h) (d) 3.0 ------,

2.5 -

.c- -...J -...J -E :::: /1

1.0 -._____ J·11 1: ..... :r-- -ti::::••••• 1····1 ..... r····1 .... ~.i:==,: ••••• : 0.5 -

0.0 -+-....-....,--.---,.-, ...... -....,-....--.-,-.---, ...... ,-....-....,--.---,.--,,.---.-....,-....--1 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 3.8 Biotransformation time course with immobilized cells: (c) enzyme profile; (d) controlled feeding profile for BZ; Chapter 3 113

16.0

14.0

5.0

2.0 2.5

0 -ll-()~=--"-l2.~~-4_C-14!J-=~L-~_,_.__,___;J.CW.lil!J.UfU0===~~~~=~-='1--"--"--'--_.__~+ 0 0 5 10 15 20 25 30 Time (h)

Figure 3.8 Biotransformation time course with immobilized cells: (e) C02 and 02 gas analysis during cell growth and biotransformation. Chapter 3 114

3.5 EVALUATION OF A CONTINUOUS IMMOBILIZED CELL PROCESS FOR L-PAC PRODUCTION

A continuous process for L-P AC production has been proposed as having some advantages when compared with batch processes. These advantages includes operational stability, quality control of product and possibly increased productivity. In the present investigation, a continuous process with immobilized C. utilis was evaluated in a continuous stirred tank reactor (CSTR) which is likely to be more favourable than a plug flow reactor due to substrate inhibition. The kinetics of the continuous process were evaluated with normal cell density and double density of immobilized cells. For the normal cell density experiment, immobilized cells from 1.5 L of culture broth were reintroduced into the same volume of reaction mixture, and for double cell density, immobilized cells from 3 L culture broth was resuspended into 1.5 L reaction mixture. Prior to biotransformation, a continuous culture with immobilized cells was conducted with 60 g/L glucose based medium at D=0.15 h-1, which was maintained for a sufficient period to reach a steady state.

As summarized in Table 3.3, with normal cell density, almost 60 g/L of glucose was fermented to ethanol with the relevant enzyme activities achieving normal levels for fermentation. Furthermore, 2.1 g/L pyruvate accumulated in the medium. When benzaldehyde was fed into medium, ethanol production was significantly reduced with increasing BZ feed rate, and the various fermentative enzymes were strongly inhibited by continuous contact with BZ. While higher BZ feed rates at 1.5 mL/h (BZ specific gravity, 1.04) resulted in higher L-PAC production, the steady state could be maintained only for 48-50 h. With 1.0 mL/h feed rate, operation stability was prolonged for 110-120 h, however, BA production exceeded L­ p AC production. Therefore, this continuous process with immobilized cells has a major problem for long term operation due to deactivation of catalytic activity caused by continuous exposure to the toxic substrate. Chapter 3 115

Table 3.3 Kinetic parameters of continuous process with immobilized normal cells density.

BZ feed rate So(BZ) S(BZ) P(L-PAC) P(BA) -q(BZ)'° q(L-PAC)• q(BA)'° (mL/h) {g/L) (g/L) (g/g/h)

0.5 2.2 0.1 0.7 1.6 0.021 0.007 0.016

1.0 4.4 0.36 2.2 2.3 0.040 0.022 0.023

1.5 6.7 0.45 4.0 3.6 0.063 0.040 0.036

BZ feed rate So(glucose) S(glucose) P(ethanol) P(pyruvate) PDC ADHr ADHcin (mL/h) (g/L) (g/L) (unit/mg protein)

0 60 1.35 27.1 2.1 0.47 0.94 0.42

0.5 60 10.5 23.5 0.9 0.40 0.87 0.40

1.0 60 16.9 19.5 0.6 0.30 0.85 0.31

1.5 60 35.2 10.1 0.4 0.26 0.71 0.25

*The cell density for the immobilized normal cell density was 15 g/L.

The higher density of immobilized cells would have a higher total catalytic activity. As summarized in Table 3.4, the results showed that a two fold increase in cell density caused higher fermentative activity which produced more ethanol from glucose (Yp/s=0.48 g/ g), close to the theoretical yield (Yp/s=0.51 g/ g).It is likely that with increased cell density, the oxygen requirements were increased resulting in more fermentative conditions within the beads with associated higher activities of fermentation enzymes, particularly ADHr. This resulted in more ethanol production than previously occurred. Under these circumstances, BA was preferentially produced as the major product, and L-P AC was produced at a significantly lower level. Chapter 3 116

Table 3.4 Kinetic parameters of continuous process with immobilized double cells density.

BZ feed rate SO(BZ) S(BZ) P(L-PACG P(BA) --* %,-PAC)* 'l(BA)* (mL/h) (g/L) g/L) (g/g/h)

1.0 4.4 0.1 0.7 3.6 0.022 0.004 0.018

1.5 6.7 0.2 1.3 5.6 0.034 0.007 0.028

BZ feed rate SO(glucose) S(fucose) P(ethanol) P(pyruvate) PDC ADH1 ADHcin (mL/h) (g/L (g/L) (unit/mg protein)

0 60 0.9 28.5 0.8 0.52 1.4 0.43

1.0 60 11.1 22.4 0.4 0.35 1.2 0.31

1.5 60 25.2 16.4 0.2 0.29 1.0 0.29

*The cell density for the immobilized double cell density was 30 g/L.

From these results, it is evident that a single stage system with immobilized C. utilis is strictly limited in its application to continuous L­ PAC production. Particularly, when toxic substrates such as benzaldehyde are used in a continuous process, partial replacement or regeneration of biocatalysts are essential to maintain sustained catalytic activity.

3.6 MORPHOLOGICAL CHANGES OF IMMOBILIZED C. UTIUS FOLLOWING EXPOSURE TO BENZALDEHYDE

Pictures of immobilized C. utilis in calcium alginate beads were taken by Phase Contrast Microscopy as shown (Figure 3.9) and Electron Microscopy (Figure 3.10). The size range of immobilized beads was 2-3 mm diameter. A relatively homogenous cell population in the beads was clearly exhibited in EM photographs as shown in Figure 3.11, in which form the morphology of the cells looked healthy and undamaged prior to exposure to benzaldehyde. During biotransformation, it has been reported that BZ altered the cell permeability due to a reduction in the content of lipids in the cell membrane (Long and Ward, 1989b). Chapter 3 117

(a)

(b)

Figure 3.9 Phase Contrast Microscope photograph of calcium-alginate beads containing Candida utilis (magnification of (a) xlO; (b) x5 ). Chapter 3 118

Figure 3.10 Electron Microscope photograph of viable Candida utilis entrapped in calcium-alginate matrix. Chapter 3 119

Figure 3.11 Comparison of Electron Microscope photographs of immobilized cells in calcium-alginate beads (a, b) before; (c, d) after biotransforma tion. Chapter 3 120

After biotransformation, most of cells had collapsed, and the cell population had decreased drastically due to cell lysis .

3.7 DISCUSSION AND CONCLUSIONS

Comparison of the enzyme activities likely to be involved in the biotransformation {Table 3.5) showed that induction of PDC, ADHr and ADHcin in C. utilis were affected by culture conditions, particularly available oxygen, culture age and glucose concentration. Under fermentative conditions, induction of PDC occurred in parallel with ADHr and AD Hein· However, high PDC activity could not guarantee high L-P AC production due to increased possibility of by-product formation. Although it is not clearly understood, there is evidence that BA production may be associated with ADHr activity. Additionally, it is possible that aromatic alcohol dehydrogenase (e.g. ADHcin) and possibly NADH dependent oxido­ reductases might be involved also in BA formation.

Table 3.5 Summary of fermentative enzyme activities in C. utilis under various culture conditions.

Initial Supplemental Maximum enzyme activity Aeration rate glucose glucose feeding (unit/mg protein) (g/L) (g/L) PDC ADHr ADHcin 0.6 vvm 60 nil 0.23 0.41 0.17 0.6 vvm changed to 60 nil 0.31 0.52 0.18 0.3 vvm 0.6 vvm changed to 90 nil 0.45 0.68 0.26 0.3 vvm 0.6 vvm changed to 60 three times of 0.59 0.83 0.20 0.3 vvm 30 ~/L

Comparison of L-PAC formation with free and immobilized cells clearly showed that immobilized cells had an enhanced resistance toward BZ (Section 3.2, Table 3.1). Thus, immobilized C. utilis might have Chapter 3 121

advantages as a biocatalyst for toxic benzaldehyde biotransformation. However, a significant drawback with the immobilized cell process was demonstrated when it was found that BA production was higher than for free cells. This is most likely due to changed respiration/fermentation conditions within the beads (favouring BA production) or to reduced BZ levels due to diffusional effects (which also favours BA production). The present results as summarized in Table 3.6, support this conclusion as higher levels of BZ contributed to lower BA formation, although both enzymes were strongly inhibited above 4 g/L BZ.

Table 3.6 Summary of biotransformation products with immobilized cells with various BZ levels maintained in the bioreactor.

BZ level dp L-PAC dp BA Reaction Molar L-PAC BA (g/L) dt dt time conversion (g/L/h) (g/L/h) (g/L) (g/L) (h) yield (%)

0.8 0.65 0.57 7.0 6.4 20 37.4 1.5 0.93 0.64 9.5 6.1 22 43.1 2 1.20 0.75 10.8 6.6 20 44.9 4 0.85 0.83 7.3 4.8 16 34.8

Since the fate of pyruvate produced via glycolysis, which could be metabolized via either the respiratory or fermentative pathway, was governed by aeration rate, control of the aeration rate during biotransformation indicated that an aeration rate which resulted in an RQ in a 5-7 range, enhanced final L-P AC production. Optimization of various factors such as PDC activity, glucose level (via pulse feeding), BZ level and aeration rate are required for the effective biotransformation of BZ to L­ p AC. Optimization of these factors gave the highest L-P AC of 15.2 g/L in a fed-batch process with immobilized cells.

Comparison of a continuous process with a fed-batch process (both using immobilized cells) indicated that the final L-PAC concentration was markedly higher in the fed-batch process (Table 3.7). Although the productivities are similar, the conversion efficiency is higher in the fed- Chapter 3 122

batch process. From the data, a single stage continuous process with immobilized cells would not be recommended for long term operation, without PDC regeneration in fresh medium or replacement with fresh biocatalyst, as significant activity loss was evident after 100-120 h with 1.0 g/h BZ feed rate to a 1.5 L reactor medium.

Table 3.7 Comparison of kinetic parameters of batch and continuous process with immobilized cells.

Type of Cone. of Cone. of Productivity Operation Molar process L-PAC BA of L-PAC time conversion (g/L) (g/L) (g/L/h) yield(%)

Fed-Batch 15.2 4.8 0.72 22h 58.0

Continuous 4.1 3.6 0.60 more than 48 h 37.9 Chapter 4 123

CHAPTER 4

CHARACTERIZATION OF PURIFIED PYRUVATE DECARBOXYLASE (PDC) FROM CANDIDA UTILIS AND SACCHAROMYCES CEREVISIAE

4.1 INTRODUCTION

As demonstrated by using whole cells for L-PAC production, by­ product (benzyl alcohol) formation significantly reduced the biotransformation yield, and the high toxicity of benzaldehyde on the cells also imposed limitations on the level of benzaldehyde that can be used for the biotransformation process. Various attempts have been made in addressing the problems of benzaldehyde inhibition and benzyl alcohol production. These include the pulse feeding of BZ to the culture broth following a period of growth and enhancement of pyruvate under oxidoreductive conditions (Long and Ward, 1989a), and the use of nicotinic acid analogues for inhibition of alcohol dehydrogenase (Smith and Hendlin, 1954). It was found that in a batch process, cell growth was inhibited at a BZ level between 1 and 2 g/L, and cell viability was drastically reduced above 2 g/L (Agarwal et al., 1987). The BZ in the media appeared to destroy the gel structure of cell cytoplasm, and had a non-specific inhibitory effect on the respiratory metabolism. Thus, high efficiency biotransformations were accomplished for only a few hours. It was found that nicotinic acid analogues slightly reduced BA formation, but did not significantly increase L-P AC. Cell immobilization appeared to offer the prospect for overcoming, or at least minimizing, the toxicity of benzaldehyde. However, as described in Chapter 3, because biotransformation by immobilized C. utilis cells was associated with a significant level of benzyl alcohol formation, an immobilized cell process would not be competitive as an alternative process. Therefore, it was considered as purified PDC had been reported to have a higher resistance to denaturation by BZ than free cells (Bringer­ Meyer and Sahm, 1988), and as the formation of BA would be eliminated in Chapter 4 124

an enzyme process, that it would be worthwhile to evaluate the biotransformation of BZ to L-PAC using a purified PDC enzyme process. Purification of enzymes is often tedious, time consuming and expensive, however complete enzyme purification is not always necessary. Boiteux and Hess (1970) have shown for example that the kinetic properties determined for high purity PDC enzyme and a crude enzyme extract from yeast cells were identical. In the present case, it seemed that a low purity or even crude cell extract would carry out biotransformation efficiently. Another possible advantage of the crude enzyme extract is that reaction rates may be much faster than with whole cells, and this may result in higher productivities due to direct contact with substrates. However, for the purified PDC process, pyruvate as substrate (which in the whole cell process is produced from glucose via glycolysis) would need to be provided to the reaction mixture.

The study reported in this chapter includes: (1) evaluation of PDC enzyme sources; (2) characterization of purified PDC from yeast; (3) kinetic evaluation for L-P AC formation using purified PDC.

4.2. EVALUATION OF PDC ENZYME SOURCES

PDC is considered to be a key enzyme for rapid alcohol production in yeasts as well as Z. mobilis. This enzyme is widely distributed in plants (Vennesland, 1951) with relatively high levels reported also in wheat germ. This enzyme converts pyruvate to acetaldehyde and carbon dioxide in a non-oxidative reaction using TPP and Mg2+ as cofactors (Green et al., 1941; Carlson and Brown 1961). The production of L-PAC is also ascribed to a PDC mediated reaction. As stated by others (Bringer-Meyer and Sahm, 1988), PDC from Z. mobilis has a lower affinity toward BZ compared to S. cerevisiae, and this resulted in low L-PAC production. This result was confirmed in our laboratories, and for this reason, the PDC from Z. mobilis was not evaluated further. Chapter 4 125

In present study, three different sources, wheat germ, S. cerevisiae and C. utilis were selected for preparation of purified PDC enzyme.

The extraction of PDC enzyme from dried wheat germ was conducted as described in a previous report (Singer and Pensky, 1952a). Wheat germ was defatted with 10 volumes of acetone at -10°C for 24 h, and homogenized in a Waring blender. Defatted wheat germ was placed in a vacuum dryer overnight. Acetone defatted wheat germ was extracted with 40 mM KH2PO4 buffer (pH 6.2) using a Blend mixer with vigorous stirring for 15 min, followed by centrifugation for 20 min at 5,000 g. PDC, ADHr and ADHcin activities from the resulting supernatant were determined as 0.05, 0.2 and 0.03 unit/mg protein, respectively. When the wheat germ was germinated under dark and high humidity conditions, the PDC activity showed a slight increase up to 0.07 unit/mg protein with germination (Table 4.1).

Table 4.1 ADHr, ADHcin and PDC activities of wheat germ during germination.

Activity (unit/mg protein) Time (h) Enzyme 0 12 24 36

PDC 0.05 0.05 0.06 0.07

ADHr 0.21 0.48 0.54 0.64

ADHcin 0.03 0.04 0.04 0.05

Evaluation of other sources of PDC was conducted with S. cerevisiae and C. utilis grown on glucose based media. When S. cerevisiae was grown on 90 g/L glucose in a 2 L LH fermenter at 28°C and pH 5.5 with 0.6 vvm aeration rate, ethanol production occurred at an early stage of growth while biomass gradually increased with culture time. In the exponential phase, maximum specific growth rate (µmax) and maximum specific ethanol production rate (qp max) were calculated to be 0.12 g/ g/h and 0.91 g/ g/h, respectively. With increasing culture time the ethanol production rate increased faster than biomass as shown in Figure 4.1. Chapter 4 126

1()(}"'1""'"(a) ______a Glucose • Biomass 0 Ethanol 80

70

C 60 ....0 ·-!U 50 ....1-, C QJ V 40 C 0 u 30 20

10

2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h) (b) 2.0,------o ADHT 1.8 • ADHcin 0 a PDC 1.6 /a 000_0-o-o -C ·-....QJ 0 1.4 /0 1-, 0.. /0 o 00 1.2 a~a'°t,:J_.-c---a-c E c/ .... a/ - 1.0 C ::s o/o/ .,,a - 0.8 / // ·-....> 0.6 /C ·­V < 0.4 . 0.2 _..._. __ _.-.-·- ·-· -----·- -· 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (h)

Figure 4.1 (a) Growth of S. cerevisiae on 90 g/L glucose based medium in 2 L LH fermenter with pulse feeding 30g/L glucose; (b) fermentative enzyme profiles. Chapter 4 127

The enzyme pattern for S. cerevisiae can be broadly correlated with the growth conditions. The changes in the enzyme patterns of S. cerevisiae indicate that PDC activity and ethanol production were closely parallel. For S. cerevisiae, PDC, ADHr and ADHcin activities reached peaks of 1.3, 1.7 and 0.4 unit/mg protein, respectively at 16-17 h. It seems that high ethanol production was associated with high PDC rather than alcohol dehydrogenase (ADHr). Pulse feeding of glucose prior to exhaustion of initial 90 g/L glucose did not further increase PDC and ADHr activities while ethanol production steadily increased. These levels of enzymes appeared to be sufficient to accomplish full fermentative metabolism.

C. utilis was grown on 90 g/L glucose medium in a 100 L fermenter with 0.6 vvm aeration rate at 25°C, pH 5.5 for 8 h, and then its metabolism was shifted from respiration to fermentation by the reduction of aeration rate to 0.3 vvm and agitation speed from 300 to 100 rpm to give rise to fast ethanol production. The increased profiles of PDC and fermentative enzymes corresponding to the metabolic change to fermentation (Figure 4.2) were similar to those in the 2 L LH fermenter as described in Chapter 3. A comparison of the kinetics for the 100 L and 2 L fermenters indicated that ethanol production in the 100 L fermenter occurred earlier at 4-5 h, presumably due to limited oxygen transfer in the bulk medium. The pulse feeding of glucose slightly enhanced the fermentative enzymes, whereas ethanol production continuously increased up to 35 g/L. The highest PDC activity of 0.9 unit/mg protein was achieved after 18 h cultivation.

A comparison of kinetic parameters for S. cerevisiae and C. utilis is summarized in Table 4.2, and indicates that the PDC activity in S. cerevisiae is generally higher than in C. utilis although the levels were obviously dependent on culture conditions. While S. cerevisiae had a higher specific activity, the total activity of S. cerevisiae on a volumetric basis was lower compared with C. utilis due to lower biomass yield of the former. Chapter 4 128

(a) 100.0------, D Glucose 90.0 • Biomass o Ethanol 80.0

-..J 70.0 bb -t: 60.0 0 ·-~... 50.0 -t: ~ 40.0 t: 0 U 30.0

20.0

10.0

2 4 6 8 10 12 14 16 18 20 22 24 26 28 Time (h) (b) 2.0------, o ADHT 1.8 • ADficin a PDC 1.6 -t: ·-....~ ...0 1.4 ~ 00 1.2 E .... 1.0 ·--t: ::I 0.8 -....>. ....:> 0.6 ·-V

0.2

0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Time (h)

Figure 4.2 (a) Growth of C. utilis on 90 g/L glucose based medium in 100 L fermenter with pulse feeding 30g/L glucose; (b) fermentative enzyme profiles. Chapter 4 129

Table 4.2 Comparison of kinetic parameters of S. cerevisiae and C. utilis grown on glucose based media.

Kinetic values Total Yeast µmax qs biomass ethanol PDC ADHr ADHcin PDC (unit/ml) (g/g/h) (g/L) ( unit/mg protein) s. cerevisiae 0.09 1.86 0.73 7.8 46.1 1.3 1.7 0.4 4.4

C. utilis 0.15 1.05 0.27 14.7 35.2 0.9 1.4 0.55 5.9

4.3 COMPARISON OF BIOTRANSFORMA TION KINETICS FOR L-PAC FORMATION BY PURIFIED PDC ENZYME FROM C. UTILIS AND S. CEREVISIAE

Biotransformation of benzaldehyde and pyruvate by PDC enzymes was carried with 3 unit/mL PDC, 50 mM benzaldehyde and 150 mM sodium pyruvate in 200 mM sodium citrate buffer (pH 6.0) at 25°C. In this experiment, PDC enzymes which had been extracted from C. utilis and S. cerevisiae, and partially purified by (~)2SO4 fractionation in the range of 45-55% saturation were used (see Section 4.4). Although a PDC enzyme process will eliminate the significant loss of benzaldehyde to benzyl alcohol, it was evident that all pyruvate added was not involved in biotransformation of benzaldehyde to L-P AC because some acetaldehyde (and possibly acetoin) were produced concomitantly. Thus, an excess of pyruvate (3 fold) was used initially for the PDC enzyme process.

The results are shown in Figure 4.3, In the reaction with PDC from C. utilis, a high initial pyruvate consumption rate occurred within 1 h, with 70 mM pyruvate being converted to acetaldehyde, 10 mM pyruvate to L­ PAC and 9 mM pyruvate to acetoin. Even with further incubation, L-PAC did not reach more than 20 mM, and the conversion yield was less than 40 % (mole/mole) based on an initial 50 mM BZ. Chapter 4 130

(a) 15.0 ....,;.._;,,..------, 0 BZ 14.0 • L-PAC 13.0 D Pyruvate 12.0 • Acetaldehyde A Acetoin :J 11.0 bh - 10.0 C: ....0 9.0 ·­i,: ...... 8.0 C: 7.0 QJ V C: 6.0 0 u 5.0 4.0 3.0 2.0 1.0 0.0 ...... --..--...... ,,...... ,.--,---,-""T-""T--r---r---r---r---r-"'T""""T""" ...... --T""""S,...... ,.---t 0 1 2 3 4 5 6 7 8 9 10 11 12 Time (h) (b) 15.0

14.0 o BZ 13.0 • L-PAC 12.0 a Pyruvate • Acetaldehyde -..:l 11.0 A Acetoin 310.0-

=0 9.0 ·-....c,: 8.0 ....~ 7.0 =i:.i V 6.0 =0 u 5.0 4.0 ~--:a- 3.0 -~--'~a~..]"arw-.-·-·- -0-0-0- •-.__._. -a- 0-0-0- __ 2.0 _ _C:i--g-.o.:~---e-e-iJ-i-i 1.0 t~•-·-·,..-.-.-.-•-.:.:.~-A-~-~~ .. 0.0 0 1 2 3 4 5 6 7 8 9 10 11 12 Time (h)

Figure 4.3 Biotransformation kinetics with 3 unit/ml PDC enzyme in shake flasks at 25°C: (a) PDC from C. utilis ; (b) PDC from S. cerevisiae. Reaction mixture consisted of 200 mM sodium citrate buffer (pH 6.0) containing 50 mM BZ and 150 mM sodium pyruvate. Chapter 4 131

In the reaction with PDC from S. cerevisiae, the acetaldehyde formation rate was even faster with 78 mM pyruvate converted to acetaldehyde, 9 mM pyruvate incorporated into L-PAC formation and 10 mM pyruvate to acetoin. Final L-P AC production was even lower than with PDC from C. utilis. In both cases, biotransformation with purified PDC accomplished only 30-40 % conversion efficiency based on benzaldehyde.

As described in previous investigations (Green et al., 1941; Singer and Pensky, 1952), the PDC enzyme can catalyse 4 types of reaction as follows:

1. Decarboxylation RCOCOOH ,----> RCOH + C02 2. Decarboxylation & Condensation (2 species) RCOCOOH + R'CHO --> RCHOHCOR'+ C02 3. Condensation & Decarboxylation (same species) 2 RCOCOOH ----> RCHOHCOR + 2 C02 4. Condensation (2 species) RCHO +R'CHO ---> RCHOHCOR'

R, R' = CH3, aliphatic or aromatic species

In all microorganisms tested, reaction 1 was usually predominant. Condensation reactions 2 and 3 were less likely or absent, with reaction 4 normally found with plant or animal PDC enzymes. Reaction 2 is required as the major reaction for L-PAC formation, however reactions 1, 2 and 3 are all likely to occur simultaneously in the reaction mixture, producing both acetaldehyde and acetoin as well as L-P AC. For further L-P AC formation, the present study is focussed on selecting optimal conditions for reaction 2 to become predominant in the biotransformation process with purified PDC from C. utilis. Chapter 4 132

4.4. PURIFICATION OF PDC FROM C. UTILIS

As described in section 4.2, C. utilis cells were cultivated on 90 g/L glucose medium with pulse feeding of 30 g/L glucose in a 100 L fermentor (Figure 4.4 (a)). In the stationary phase in which the PDC level was 0.9 unit/mg protein, cells were harvested with a continuous Sharples Super Centrifuge (Sharples Centrifuge Co., Ltd. U.S.A., Figure 4.4 (c)) at 20,000 rpm. The cell cake was washed in water twice through a Westfalia continuous separator SA 1-01 (Westfalia Separator AG., Germany, Figure 4.4 (b)) at 9,700 rpm. Washed cells were resuspended in 40 mM KH2PO4 (pH 6.0) containing 0.05 mM TPP and 0.3 mM MgSO4. The disruption was effected with a Manton­ Gaulin-APV high pressure homogenizer at 640 kPa {The APV Co., Ltd. England, Figure 4.4(d)), The degree of disruption in the multiple process was examined microscopically, and the released protein was determined by the Bradford method (Bradford, 1970). Since the temperature 'at each pass rose about 10-12°C due to adiabatic compression, the samples were pre­ cooled and cooled again between each step of the multiple process. PDC activities at each cycle were measured, and the results are summarized in Table 4.3. During this process some slight enzyme inactivation might be caused by heat, proteolysis, loss of cofactor and mechanical stress.

Table 4.3 Disruption of C. utilis by high pressure homogenizer at 9000 psi (640 kPa).

Enzyme activity Number of Released protein (unit/mg protein) cycle (mg/mL) PDC ADHr ADHcin

1 27.2 0.51 1.12 0.36

2 28.5 0.65 1.24 0.48

3 30.3 0.85 1.38 0.51

4 33.5 0.87 1.45 0.58 Chapter 4 133

(a)

Figure 4.4 Photograph of (a) 100 L Fermentor;

~ ~

(.,.) (.,.)

,.... ,....

~ ~

'"I '"I

(1) (1)

...... PJ PJ

n n

::r­

'"O '"O

(d) (d)

. .

Super Super

;,-, ;,-,

Sharples Sharples

.: .:

Homogenizer

(c) (c)

, ,

Pressure Pressure

Separator

High High

(c) (c)

Continuous Continuous

Manton-Gaulin Manton-Gaulin

Westfalia Westfalia

(d) (d)

(b) (b)

and and

4.4 4.4

Figure Figure

Centrifuge Centrifuge (b) (b) Chapter 4 135

Although higher pressure and multiple passes gave rise to more efficient disruption, the further passage of already broken cells resulted in fine debris which would be excessively difficult to remove further downstream. Thus, a 3 cycle homogenization process was selected to disrupt C. utilis cells, with 0.85 unit/mg protein of PDC being obtained. The resulting cell homogenate was diluted with 3 times volume of 40 mM KH2PO4 buffer, and the cell debris and other solid particles were removed by centrifugation at 3,800 g for 30 min. For further purification, fractionation of the protein mixture by stepwise increase in ammonium sulphate concentration was used. The results shown in Table 4.4 demonstrate that 50 % saturation of (NH4)2SO4 provided 2.1 fold purification and 63.5 % recovery. On the basis of these results, PDC enzyme precipitated in the range of 45-55 % (NH4)2SO4 saturation was harvested and desalted by dialysis in purification buffer. Finally, the active PDC enzyme solution was loaded on a Sephacryl S-300 (Pharmacia Chem. Co.) gel chromatography column equilibrated with 40 mM KH2PO4 (pH 6.0) containing 30 µM TPP and 0.3 mM MgSO4.

Table 4.4 Fractional purification of PDC enzyme with ammonium sulphate

% of saturated Total PDC activity Specific activity Recovery (NH4)zS04 (unit) (unit/mg protein) (%)

Crude extract 5572.5 0.75

0-20 % 5.8 0.34 0.1

-30% 2.9 0.05 0.05

-40 % 400.2 0.27 7.2

-50 % 3514.1 1.61 63.5

-60 % 508.4 0.29 9.2

-70 % 24.8 0.03 0.8 Chapter 4 136

As shown in Figure 4.5, a reasonable separation of PDC from other proteins including ADHr and ADHcin was achieved. As a result of collection of 6 active fractions (No. 22-27), purification of PDC via chromatography was achieved with 5.6 fold increase in purity and an overall 54.3 % recovery. A summary of the purification results is presented in Table 4.5.

Table 4.5 Summary of results of purification of pyruvate decarboxylase from C. utilis.

Total Total Process protein activity Specific activity Purification Recovery {mgl {unitl (unit/mg eroteinl fold :rield {%) 1. Cell homogenate 928.6 789.3 0.85 100 (640 kPa, 3 cycle) 2. (NH4)2S04 360 543.6 1.51 1.8 72.5 fractionation (45-55 % saturation) 3. Gel filtration 85.3 407.6 4.76 5.6 54.3 on Seehacr~l S-300 Chapter 4 137

24.0 ...... 0 ...... A 280nm ...... _ ADHT e 22.0 c:: ···"' .. o-·- ADH.:.in 0 20.0 00 ...... •-·- PDC N < 18.0 ... 16.0 ..,J -e 14.0 -:.:i.... c:: 12.0 ::s 10.0 ->...... w > 8.0 .w ·-V < 6.0 4.0 2.0 0.0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Elution volume (x 4.2mL)

Figure 4.5 Chromatography of the pyruvate decarboxylase from C. utilis on Sephacryl S-300. A column (3 cm ID x 45 cm L) was equilibrated with 40 mM KH2PO4 containing 0.3 mM MgSO4, 0.03 mM TPP at 4°C. A flow 8 mL/h was maintained with 12 cm hydrostatic pressure head. Chapter 4 138

4.5 CHARACTERIZATION OF PDC

The properties of purified PDC enzyme from C. utilis were determined with regard to the effect of buffer species, pH, temperature and cofactor requirements on the relevant kinetic constants.

4.5.1 Effect of buffer species on the stability of PDC

The stability of the purified PDC enzyme was evaluated by measuring a time profile of residual activity in various buffers with and without 0.03 mM TPP at 25°C, pH 6.0. Although some loss of PDC activity occurred in all buffers, it was found that PDC was considerably more stable in potassium phosphate and sodium citrate buffers compared with Tris-Cl buffer (Figure 4.6). Furthermore, the stability of PDC was significantly improved by addition of TPP, this trend being revealed in all buffers under test. The loss in activity with time is consistent with a previous report (Gounaris et al., 1975) which established that the dimeric tetramer structure (cx2f3z) of PDC dissociated into dimer and monomer subunits with concomitant release of TPP and magnesium ions in vitro. This dissociation, predominantly a function of pH (Hopmann, 1980), is affected also by the buffer species, and appears to be greater in Tris-Cl compared with phosphate and citrate buffers (Gounaris et al., 1971). The loss of activity in phosphate buffer containing TPP was similar to that in sodium citrate buffer with half lives of PDC being about 28 h in both cases, while in Tris-Cl buffer, the half life of PDC was 17 h. It was emphasised that half lives without TPP were significantly less compared to those with TPP. In distilled water for example PDC retained less than 50 % activity after 2 h. This result confirms that TPP is involved in the maintenance of active conformation of PDC holoenzyme. However, it is possible also that the purified enzyme from yeast may contain some contaminating protease which could enhance decay of PDC enzyme. Chapter 4 139

110 With TPP a KH2P04 100 6 Tris-Cl o Sodium citrate 90 Without TPP 80 • KH2P04 -0~ A Tris-Cl ->.. 70 .... • Sodium citrate ·-> .... t:I Water ·-V 60 ns ns 50 -:s "0 ·-fll 40 ~ ~ 30 8 20

10

0 0 5 10 15 20 25 30 35 40 Time (h)

Figure 4.6 Effect of buffer species (pH 6.0) on the stability of PDC from C. utilis in presence and absence of 0.03 mM TPP at 25°C. Chapter 4 140

4.5.2 Effect of TPP on PDC stability

To determine the optimum concentration of TPP in the buffer system, 10 unit/mL PDC enzyme was incubated at 25°C in 40 mM KH2PO4 (pH 6.0) with various concentrations of TPP. As shown in Figure 4.7, it was observed that the half life of PDC was gradually extended with increasing TPP concentration. However, above 5 µM TPP the stability of PDC was independent of TPP concentration. It is likely therefore that PDC enzyme was saturated with TPP above 5 µM. It was evident that TPP in addition to its catalytic role also functioned in maintaining the optimal conformational structure for the holoenzyme.

4.5.3 Determination of Km value for pyruvate

As estimated from data shown in Figure 4.8 (a), the Km values of PDC from C. utilis for pyruvate has been calculated as 2.4 mM at 25°C and pH 6.0. This enzyme was saturated at a concentration of about 15 mM pyruvate. As a comparison, the Km value of PDC from S. cerevisiae shown in Figure 4.8 (b) was determined to be 3.2 mM under the same conditions. The Km values are in agreement with other reported values (e.g. 3.0-3.6 mM, Urk et al., 1989; 1.2 mM, Lehmann et al., 1973; 1.3 mM, Boiteux and Hess, 1970). When compared with Z. mobilis for which Km value was 0.4 mM (Bringer­ Meyer et al., 1986), yeast PDC enzyme possesses a low affinity toward pyruvate.

4.5.4 Effect of pH on PDC stability

The lower temperature of 4°C was selected for a more complete evaluation of PDC characteristics as previous studies reported that enzyme stability was enhanced at 4°C (Cooper, 1977). Furthermore, other data (see section 4.5) indicated that at low temperatures the reactions of decarboxylaytion and condensation of pyruvate with benzaldehyde to L­ PAC Chapter 4 141

110 Cone. of TPP (µM) • 0 11 0.05 a 0.1 o 0.3 • 0.5 11 1.0 C 3.0 A 5.0 A 10.0 o 30.0

5 225 3o 35 L--.~.--~~~--10 15 20 Time (h)

Effect o f TPP concentra ho·non Poe stab11·1·t y in 4 0 m M KH2P04 bufferFigure (p 4~ 6 ·0) at 2soc. Chapter 4 142

(a) 2.0

1.8 -c:: ·-e 1.6 -QJ -0 1.4 e ----- o;.,..---=-o 0 0 :i 1.2 -QJ !IS 1.0 -i.. 0I c:: 0 0.8 ·-V -!IS QJ 0.6 c:i:: 0.4

0.2 p

0.0 -¥-r""'T'-.--,...,.-,-'l'""'T...... T"""'l"'-r-'l'""'T-.--'l'""'T-.--r--r-r-.,....,.-,-,....,...,....,....,.~,..... 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Pyruvate (mM) (b) 2.0 ------,

1.8

-c:: 1.6 ·-e -QJ 1.4 -0e -- -- 0 :i 1.2 0------0 -QJ .... 1.0 ~ i.. 0 / c:: 0 0.8 I ·-V -!IS 0.6 QJ c:i:: 0.4 -7 0 I 0.2 / I I 0.0 -¥-,...,....-,:...... ,....,.....T"""T"-,-r-T"""T"""r-T"-r-T"""'l"'"'T"""r-r-r-r-r,--,r-,-r-r--,--,--r; 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Pyruvate (mM)

Figure 4.8 Determination of Km value for pyruvate: (a) C. utilis ; (b) S. cerevisiae at pH 6.0 and 25°C. Chapter 4 143

were preferred to the direct decarboxylation of pyruvate to acetaldehyde. The pH dependence of PDC stability was investigated in the pH range 4.5-8.0 with Universal citrate-phosphate buffer containing 0.03 mM TPP at 4°C. As shown in Figure 4.9, the PDC stability was significantly affected by pH. High stability of PDC was maintained in weak acidic conditions between pH 5.5-6.5. Previous investigations (Hopmann 1980; Hilbner et al., 1990) had shown that the binding strength of TPP differed for the various binding sites, and was pH dependent. It has been found that PDC can dissociate easily into apoenzyme and TPP under alkaline conditions, and recombination of the apoenzyme to the holoenzyme takes place below pH 7.0 in the presence of TPP. However, the original activity of the a holoenzyme was not completely recovered, and the degree of reconstitution of the holoenzyme was found to be inversely related to incubation time (Gounaris et al., 1971 ). At 4°C, the half life of PDC in pH 6.0 buffer was 110 h compared to 28 h at 25°C (see Figure 4.6). This emphasises the need to keep the enzyme preparation cold to maintain the active conformation at the molecular level, and to slow any protease action that may occur. A protease inhibitor could have been added, however it was evident that at 4°C the rate of deactivation was slow.

4.5.5 The requirement of TPP for reconstitution of PDC apoenzyme to PDC holoenzyme.

The requirement of TPP for reconstitution of the holoenzyme from the apoenzyme was investigated. This study was performed using a dialysis method as follows: 10 unit/mL PDC enzyme was dialyzed at 4°C against an excess amount of 40 mM KH2PO4 buffer (pH 7.8) changed every 6 h until the PDC enzyme was almost inactive when assayed directly without prior incubation. Reactivation of PDC enzyme was carried out by the dialyzing prepared apoenzyme with various concentrations of TPP in 40 mM KH2PO4 (pH 6.0) containing 5 mM MgSO4 for 48 h. Reconstituted PDC activity was measured and plotted against TPP concentration (Figure 4.10). The degree of PDC reactivation was increased with increase of TPP up to 10 µM, and above 10 µM TPP, 82 % of PDC activity was recovered. Chapter 4 144

100 Buffer pH Ill 4.5 90 6 5.0 o 5.5 ~ 80 • 6.0 ~ 70 • 6.5 ·;: D 7.0 ·-ti 60 • 7.5 CU O 8.0 -CU 50 "O= -~ 40 ~ 01:: 30

20

10

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 Time (h)

Figure 4.9 pH dependence of PDC stability in sodium citrate/phosphate buffer at 4°C. Chapter 4 145

100

90

0 80 0 -#- 70 ->- ...... 60 ...... > V !U 50 QI 0 > ...... 40 !U -QI ~ 30

20

10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 TPP (uM)

Figure 4.10 Effect of TPP concentration on the reactivation of apoenzyme (10 unit/mL) with 40 mM KH2P04 buffer (pH 6.0) at 4°C. Chapter 4 146

Ullrich (1982) have reported that added TPP appeared to occupy specific binding sites in the protein subunits reversibly and independently, and the activation of the apoenzyme by TPP was relatively slow. If there is sufficient saturation with both of TPP and Mg2+, the reconstitution of ternary complex of the holoenzyme is a rate limiting step. While the holoenzyme contains 2-4 molecules of TPP and Mg2+, the requirement for cofactor is dependent on making up this loss during purification. Usually the reconstitution process requires excess amounts of TPP compared to the amount of TPP required solely for activation (Gounaris et al., 1971).

4.5.6 Effect of Mg2+ on PDC activity

The optimum requirement for Mg2+ was determined using the dialysis method outlined previously. 10 unit/ml PDC enzyme was dialyzed with an excess amount of 40 mM KH2PO4 (pH 6.0) containing 50 µM TPP for 36 h. Exhaustive dialysis within the pH stability range failed to reduced the activity to zero in the absence of added Mg2+. The results shown in Figure 4.11 indicate that added Mg2+ did not affect PDC activity. It seems that the Mg2+ was strongly bound to the protein, or that the range examined exceeded the requirement for PDC saturation. From the literature (Singer and Pensky 1952a; Bringer et al., 1986), apparent Km values for Mg2+ for the apoenzyme were reported in the range of 100-200 µM. It was reported also that Mg2+ could be replaced by Mn2+ without loss of binding stability. Chapter 4 147

llO ...... • 100 • • • • 90 • PDC

-~ 80 -->. 70 ·:::... 60 ·-...(J CQ

4,1 50 > 40 ·-...CQ 1j c:z:: 30 20· 10 0 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 MgS04 (µM)

Figure 4.11 Effect of Mg2+ on PDC activity at 4°C and pH 6.0. Chapter 4 148

4.6 EFFECT OF REACTION TEMPERATURE ON L-PAC FORMATION

The dependence of the PDC reaction on temperature was evaluated at 4°C, 10°C and 25°C. The biotransformations were carried out with the reaction mixture consisting of 40 mM KH2PO4 (pH 6.0) containing 7 unit/mL PDC from C. utilis, 70 mM BZ, 70 mM sodium pyruvate, 30 µM TPP and 0.3 mM MgSO4. The biotransformation kinetics are shown in Figure 4.12 and their data analysis summarised in Table 4.6. As shown in Figure 4.12, in the first 3-4 h reaction L-PAC formation increased with time. Then due to depletion of pyruvate and possible product inhibition by acetaldehyde and L-P AC, the rate of L-PAC formation decreased significantly towards the end of reaction. A comparison of the kinetic parameters at different temperatures showed the interesting result that while the formation of acetaldehyde was the major reaction at 25°C (on a molar basis), the formation of L-PAC was predominant at 4°C.

Table 4.6. Comparison of initial kinetic parameters (initial rates determined over first 30 min) and final L-P AC concentration at various temperatures. Reactions were carried out with 70 mM benzaldehyde, 70 mM pyruvate and 7 unit/mL PDC enzyme in 40 mM KH2PO4 buffer (pH 6.0).

Kinetic values Temperature (mM h·1) 4°C 10°C 25°C -ds pyr 46.2 63.0 58.1 dt -ds BZ 21.0 27.0 9.5 -dt dpL-PAC at 20.9 26.6 8.3 ~ acetoin 8.5 11.4 10.9 dt dp acetaldehyde 6.1 13.3 27.1 dt L-PAC max (mM) 37.7 34.1 22.0 Chapter 4 149

__.!.:(a::!)______-, 8.0 o BZ • L-PAC 7.0 • Acetaldehyde • Acetoin 6.0 c Pyruvate ..,J -bb 0 • 5.0 -C: ·-·-· ....0 ·-fU ...... 4.0 c~- ~ C: QJ c • 0 V \ e o--- --O-Y..0--o--o C: 3.0 u0 2.0 .\ C 1.0 ·-·-·-~~=•-•~•-•==-• 0.0 _::;;4-::::.!!•C-:::•:.-...... •_T-""~c~:----C~..::;::..=;=~c!=-=-:;::-~"=:;=:.:;"~=:;:-=-:ljl..~-I 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 Time (h) (b) 8.0 ------, o BZ • L-PAC • Acetaldehyde • Acetoin c Pyruvate ..,J bO- -C: ·-....0 ...... fU C: QJ V C: u0

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 Time (h)

Figure 4.12 Biotransformation kinetics at various temperatures: (a) 4°C ; (b) 10°C. Reaction mixture contains 70 mM benzaldehyde, 70 mM sodium pyruvate and 7 unit/mL PDC of C. utilis in 40 mM KH2PO4 buffer (pH 6.0); Chapter 4 150

(c) 8.0-.------, o Benzaldehyde • L-PAC 7.0 °"-o • Acetaldehyde ' A Acetoin c Pyruvate 6.0 -..,J ~ a.___ bb 0---o--______,.,.. O-o -c:: 5.0 ....0 ·-11: ....i.. 4.0 c:: [] ~ V c:: 3.0 u0 \ C • ...... -. --·-·--· 2.0 \. ./

c

Figure 4.12 Biotransformation kinetics at various temperatures: (c) 25°C. Reaction mixture contains 70 mM benzaldehyde, 70 mM sodium pyruvate, and 7 unit/ mL PDC of C. utilis in 40 mM KH2P04 buffer (pH 6.0). Chapter 4 151

4.7 EFFECT OF ACETALDEHYDE ON THE INITIAL RATE OF L-PAC FORMATION

As mentioned previously (section 4.5), the effect of acetaldehyde on L-P AC formation needs to be clearly understood for the PDC enzyme process at different temperatures. The results shown in Figure 4.13 (a) indicate that the initial reaction rates for L-P AC formation were drastically decreased with increasing acetaldehyde concentration. This means that the condensation reaction of 'active acetaldehyde' with benzaldehyde may be greatly affected by free acetaldehyde. Gruber and Wassenaar (1960) have already designated acetaldehyde as a non-competitive inhibitor of pyruvate decarboxylase. As acetaldehyde increased, the velocity of the reaction catalyzed by PDC quickly decreased even in the presence of an excess of pyruvate. The extent of inhibition by a given concentration of acetaldehyde was found to be nearly the same whether it was added at the beginning of the experiment, or accumulated as a result of decarboxylation of pyruvate. Juni (1961) suggested a two site reaction in which pyruvate may be bound and decarboxylated at the first catalytic site and irreversibly transferred to the second site. Acetaldehyde at the second active site can be reversibly released into the medium as free acetaldehyde. In a biotransformation, it seems that free acetaldehyde binding at the releasing site of PDC can induce an allosteric inhibition effect, so that the substrate binding site may be reversibly inactivated with increase of acetaldehyde. A high concentration of acetaldehyde may be responsible also for reducing acetaldehyde transfer from the decarboxylation site to the releasing site of the PDC. On the other hand biotransformation for L-P AC production with various initial concentrations of acetaldehyde showed that although initial rates of L-P AC production decreased with increase in acetaldehyde concentrations, L-PAC production gradually increased with time even in the presence of initial 300 mM acetaldehyde as shown in Figure 4.13(b ). Chapter 4 152

(a) 0.10 -C 0.09 ·-E ~ 0.08 E u 0.07 < ~ I 0.06 ...,J -....QJ 0.05 i,: i.. C 0.04 0\ ....0 ·-y i,: 0.03 °'\ QJ i.. o, ,; 0.02 ·-.... 0~ ·-C 0.01 o-___o - 0 0 0.00 0 50 100 150 200 250 300 Acetaldehyde (mM) (b) 7.0 Cone. of acetaldehyde

1:1 OmM 6.0 • lOmM C 30mM 5.0 • SOmM o lOOmM -...,J & lSOmM .._,bb 4.0 6 200mM u o 300mM < ~ j 3.0

2.0

1.0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time (h)

Figure 4.13 Effect of acetaldehyde on (a) initial reaction rate for L-PAC formation; (b) bioconversion kinetics. Reaction mixture contains 70 mM sodium pyruvate, 70 mM BZ and 7 unit/mL PDC in the presence of various concentrations of acetaldehyde at 4°C. Chapter 4 153

4.8 TOXIC EFFECT OF BENZALDEHYDE AND L-PAC ON PDC ACTIVITY

As reported previously, purified PDC enzyme from S. carlsbergenesis has a higher resistance to benzaldehyde toxicity compared with whole yeast cells (Bringer-Meyer and Sahm, 1988). In the present study, the inhibition or inactivation effect of benzaldehyde and L-PAC toward PDC enzyme has been examined. After the PDC enzyme was exposed to various concentrations of benzaldehyde and L-P AC in 40 mM KH2PO4 buffer (pH 6.0) containing 30 µM TPP, 0.3 mM MgSO4 and 2.0 M ethanol (which used for enhancing solubilities of both substances, see Section 5.2) for 1 h at 4°C, the residual activities were measured. The results showed that the PDC activity was drastically decreased with increasing benzaldehyde concentration, while PDC activity was diminished more gradually with increasing L-P AC concentration (Figure 4.14). At 200 mM benzaldehyde, there was no detectable PDC activity after 1 h incubation. Furthermore, extended incubation time brought about a greater toxic effect at a given concentration of substrate. The apparent Ki in the present study was revealed at 80 mM benzaldehyde which is considerably higher than a Ki of 14.2 mM for free cells of C. utilis (Dissara et al., 1993). Generally, it appeared that the toxicity of benzaldehyde was about 2.5 times higher than that of L-P AC (on a molar basis) in the range of 50-200 mM. This result implied that for a given concentration of benzaldehyde, a higher conversion rate of benzaldehyde to L-PAC would result in reduced toxic effect of benzaldehyde due to the more highly toxic substrate being converted to the less toxic product. Chapter 4 154

l()()"'T""" ______

90 '--~s---...... _ 80 .---_ L-PAC -'#- 70 ·------>. -.... 0 ·------·s: 60 .... "o • ·-~ 50

40 Benzaldehyde 0 30

20

10

0 20 40 60 80 100 120 140 160 180 200 220 240 Concentration (mM)

Figure 4.14 Toxic effects of benzaldehyde and L-P AC on purified PDC from C. utilis. After 1 h exposure to various concentrations of benzaldehyde or L­ P AC at 4 °C, residual activities were measured. Chapter 4 155

4.9 DISCUSSION AND CONCLUSIONS

The levels of PDC activity has been reported previously to vary with different sources, induction conditions and purification procedures. In the present study, wheat germ, S. cerevisiae and C. utilis were selected as source for PDC enzymes. Although the levels of enzyme were significantly affected by environmental conditions, PDC activity from yeasts was found to be 15- 20 times higher than that from wheat germ. Furthermore, S. cerevisiae had a higher specific activity compared to C. utilis. However, as a potential enzyme source of PDC, C. utilis has several advantages. These include: a higher biomass yield of C. utilis compared to S. cerevisiae, which resulted in higher total PDC activity per unit volume, the PDC activity could be controlled by available oxygen with C. utilis, its purification is generally simpler and cheaper than for wheat germ enzyme. In the study of PDC stability in buffer, it was found that three factors significantly influenced PDC stability, viz. buffer species, pH and TPP concentration. Activity was decreased significantly at alkaline pH, and this decay was greater in Tris-Cl buffer. Stability was enhanced by addition of TPP. Thus, it was confirmed that the stability of the PDC could be closely correlated with the binding affinities of the tetramer protein subunits and both cofactors TPP and Mg2+, which in turn were influenced by pH, buffer species and temperature. It was evident that during the purification of the enzyme, TPP plays an important role as a stabilizer in addition to being an essential cofactor. Comparison of the kinetics for L-P AC production and the Km values for pyruvate with PDC from both C. utilis and S. cerevisiae showed similar biotransformation characteristics and values. In a PDC mediated-reaction, it was considered that loss of carbon dioxide and transfer of acetaldehyde from catalytic site to releasing site are likely to be rate limiting steps (Crosby et al., 1970), and thus active acetaldehyde could be involved in 3 types of catalytic reactions. The production of acetaldehyde as a major product is accelerated with increasing temperature. Thus, at higher temperatures, more acetaldehyde (which is not reduced at the expense of NADH) will be accumulated in the reaction mixture, while at lower temperatures, the a-carbanion at the PDC active site may have an extended life. As a result, L-PAC production is likely to be predominate at lower temperatures (e.g. 4°C). Chapter 4 156

Furthermore, it was found that the biotransformation rate was significantly inhibited by an increasing residual acetaldehyde. This result is in agreement with previous studies that acetaldehyde could be designated as a noncompetitive inhibitor (Gruber and Wassenar, 1960). It seems that acetaldehyde binds at a site on the enzyme other than substrate binding site, thereby altering the conformation of the enzyme so that reversible inactivation of the catalytic site results. Noncompetitive inhibitors bind reversibly to both free enzyme and enzyme-substrate complex to form inactive complexes EI and ESI (Lehninger, 1982). On the basis of previous reports Guni, 1961; Carlson and Brown, 1961; Bringer-Meyer and Sahm, 1988) and the present results, a 2-site reaction mechanism for L-PAC formation can be described as follows (Figure 4.15): Initially, substrate, pyruvate attaches to the PDC at the substrate binding site in which the a-carbon of pyruvate binds with the thiazole ring of TPP (step 1). This is followed sequentially by decarboxylation (step 2). The 'active acetaldehyde' which is bound with TPP, could follow one of three irreversible pathways. First, as a major reaction, acetaldehyde could transfer to the second active site (step 3), and then reversibly dissociate to yield free acetaldehyde (step 4). Second, 'active acetaldehyde' could condense with another aldehyde such as benzaldehyde or acetaldehyde to produce L-PAC or acetoin, respectively (step 5). Acyloin production will depend on the potential binding affinity of substrate and HETPP (2-hydroxyethyl thiamine pyrophosphate), with benzaldehyde likely to have higher hydrophobic interaction with 'active acetaldehyde' than with free acetaldehyde (Lehmann et al., 1973). Lastly, when free acetaldehyde is available (from the either reaction mixture or first active site via decarboxylation) at the second active site (step 6), condensation with 'free acetaldehyde' could occur. This reaction however may be strongly inhibited due to a conformational change of the PDC resulting in reversible inactivation of the catalytic site. This reversible inactivation (step 3, 6) may be magnified with increased free acetaldehyde as shown in the results in Figure 4.13. Comparison of the toxic effects of benzaldehyde and L-PAC suggest that since increasing irreversible inactivation with time can be caused by benzaldehyde and L-PAC, high initial reaction rate should be employed in order to enhance the conversion efficiencies before significant PDC inhibition occurs. Chapter 4 157

1 2 3 4

s I

(or CH3CHO)

CHOHCOCH3 0 CTl30fOHCDCH3 L-PAC Acetoin l 4

Table 4.15 Proposed two-site reaction mechanism catalysed by pyruvate decarboxylase for L-PAC formation. Chapter 5 158

CHAPTER 5

KINETIC EVALUATION OF BIOTRANSFORMATION FOR L-PHENYLACETYLCARBINOL FORMATION BY PURIFIED PYRUVATE DECARBOXYLASE FROM CANDIDA UTILIS

5.1 INTRODUCTION

One of the main limitations of biotransformations lies in the relatively low product concentrations from the reaction when compared with chemical processes. This limitation is partially due to the hydrophobic .character of the compounds used which leads to a low solubility of the substrate in an aqueous system. Other reasons are that biocatalysts are often very sensitive to changes in pH, temperature, buffer species and organic solvents. However, the most crucial consideration in favour of a biotransformation process is its ability to achieve a rapid and efficient reaction for desired products with their unique properties of regio- and stereo selectivity. Furthermore, such processes often have the capacity to operate in non-extreme conditions. One of the ways to improve the efficiency of a biotransformation is that since some enzymes can function in non-aqueous solvents which contain relatively low concentrations of water, various such solvents can be used for the enzyme reaction with higher substrate concentrations (Rose and Schneider, 1991). However in such an organic solvent/water system, the toxicity of the organic solvents towards the biocatalyst needs to be considered. Beside the effect of solvents on the activity and stability of the enzyme, there is also a solvent/water effect on the equilibrium of the reaction. It was found that since only a small amount of water may be required to maintain enzyme catalytic activity (Nikolova and Ward, 1992), the relative rates of the forward and the reverse reactions are often controlled by the water content of the reaction mixture (Malcata et al., 1992). The major advantage of carrying out an enzymatic reaction in organic solvents is the significant enhancement in solubilities of most organic compounds. Hydrophobic substrates, particularly aromatic substrates, can be more efficiently converted (Zaks and Klibanov, 1984). Chapter 5 159

In the production of L-PAC, PDC catalyzes the condensation of benzaldehyde and active acetaldehyde. However, the reaction rate of L-PAC formation can be significantly reduced by partial deactivation of the enzyme by the toxic benzaldehyde as well as product inhibition by acetaldehyde and L-P AC. Moreover, the limited solubility of benzaldehyde (0.3 g/ 100 mL water) can reduce the extent of the saturation of PDC with substrate. As described in Chapter 4, the benzaldehyde was not totally converted to L-P AC by purified PDC. Because of including the limited solubility of benzaldehyde, it is unlikely that the enzyme will be fully saturated with substrate for L-P AC synthesis. It is known that the synthesis of L-PAC is greatly influenced by concentration of active acetaldehyde, which is not only condensed with benzaldehyde, but also released into the medium as free acetaldehyde. As a result, an excess concentration of pyruvate compared to benzaldehyde is desirable to achieve high L-P AC formation.

The objectives of the research in this Chapter are: (1) to examine various reaction conditions in order to increase the reaction rate of L-P AC production and minimize by-products; (2) to evaluate the biotransformation kinetics for L-P AC formation with various PDC activities and molar ratios of substrates.

All experiments were carried at 4°C as it had been established previously in Chapter 4 that L-P AC formation was favoured at this lower temperature. Chapter 5 160

5.2 EFFECT OF ETHANOL CONCENTRATION ON L-PAC FORMATION

As described above, it is desirable to use an organic solvent in order to improve the rate of 1-PAC production. In selecting an organic solvent, it was considered that PDC should retain its catalytic capability in the aqueous organic solvent, and that this system should favour the stability of substrates and products, and enhance the desired reaction. Since Jencks (1975) showed that PDC possesses highly lipophilic substrate binding sites, the hydrophobic interaction of substrate molecules at apolar active sites of PDC is important for 1-PAC formation. The selection of ethanol as a water miscible organic solvent in these experiments was based on the following results: PDC has significant resistance to denaturation by ethanol up to 150 g/1 (Scopes, 1989), ethanol is a less hydrophilic solvent than water (Godfrey, 1972), and benzaldehyde is reported to have infinite solubility in ethanol (Perry, 1975).

5.2.1 Effect of ethanol on initial reaction rates for L-PAC formation

Initial reaction rates were determined within the first 30 min after addition of substrates and PDC to the reaction mixture in the presence of various concentrations of ethanol. The biotransformations were conducted with 30 ml of reaction mixture containing 70 mM benzaldehyde, 70 mM sodium pyruvate and 7 unit/ml PDC with ethanol in the range of 0-6.0 M in 50 ml Erlenmeyer flasks. These were incubated on a reciprocal shaker at 100 rpm and 4°C. As shown in Figure 5.1, the enhancement in initial rate of 1-PAC formation due to the presence of ethanol was a significant observation. These results demonstrated that the rate of 1-PAC formation gradually increased with increasing ethanol concentration up to 3.0 M. At 3.0 M ethanol, the initial reaction rate was 1.4 times higher than that in the absence of ethanol. A further increase in ethanol concentration resulted in drastically decreased initial reaction rates. Chapter 5 161

160

~ 140 _,,--o -0 -u /0 120 <~ I ..J ... ~ / ....0 100 QJ -...IU 80 IU ·-- 0 ·--c:: 60 ·-QJ > ·- 40 -IU -=:QJ 20 o"--o

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 Ethanol (M)

Figure 5.1 Effect of ethanol on initial rate for L-PAC formation at 4°C. Reaction mixture contained 70 mM BZ, 70 mM sodium pyruvate, 7 unit/mL PDC and various concentrations of ethanol in 40 mM KH2PO4 buffer (pH 6.0). Chapter 5 162

Possible reasons for this acceleration of L-PAC formation are (1) that benzaldehyde in the less hydrophilic ethanol solution may have more easy access to the hydrophobic active sites, and (2) that the enhanced solubility of benzaldehyde in the ethanol solution is likely to increase the substrate saturation of PDC. Thus, the condensation reaction with enzyme bound acetaldehyde at the lipophilic site is likely to be preferred to that releasing free acetaldehyde from the decarboxylation site of PDC (Schellenberger and Hubner, 1967). These results are in agreement with a previous report (Crosby et al., 1970). These authors showed in an investigation of non-enzymatic decarboxylation with TPP, that the rate of decarboxylation was significantly increased in solvents less polar than water, in which the rates of decarboxylation in ethanol were 104 to 105 fold faster compared with the rates in water. In apolar aprotic solvent, the rate was even faster.

5.2.2 Effect of ethanol on PDC stability

PDC stability with various concentrations of ethanol was measured as a function of time. The results shown in Figure 5.2 illustrate that PDC stability was enhanced in the presence of ethanol in the range of 0.5 to 3.0 M. At 2.0 Methanol, 70-75 % of PDC activity was maintained after 100 h incubation at 4°C, while only 40 % initial activity was evident in the absence of ethanol. It is likely that the configuration of the hydrophobic active sites of the PDC are more stable in the less hydrophilic ethanol solution. These results are in agreement with other previous findings (Butler, 1979; Freeman, 1984). These authors found that modest concentrations of some water-miscible compounds (e.g. glycerol and other polyakohols) enhanced the stability of the enzyme. However, PDC stability was significantly decreased with increasing ethanol concentration over 4.0 M. At 6.0 M ethanol, PDC activity was decreased drastically in a few hours. It seems that the PDC when added to an aqueous solution containing a high concentration of water-miscible solvents may be unfolded, and lose its activity. From this result, it is evident that ethanol in the range of 0.5-3.0 M is likely to significantly improve the PDC enzyme stability. Chapter 5 163

120 Cone. of ethanol 110 • OM 100 0 O.SM • 1.0M 90 • 1.SM -~0 2.0M 80 • -...... >, C 3.0M ....> 70 El 4.0M V A S.OM -ns 60 A 6.0M ns El -:::s ...."'0 50 (I) QJ 40 c=::: A 30 20 10 -. A A 0 0 20 40 60 80 100 120 140 160 Time (h)

Figure 5.2 Effect of ethanol on the stability of purified PDC from C. utilis. After exposure to various concentrations of ethanol in KH2P04 buffer (pH 6.0) at 4 °C, the residual activities were measured at different time intervals. Chapter 5 164

5.2.3 Effect of ethanol on the yield for L-PAC formation

The biotransformation yields with and without 2.0 M ethanol were compared with 140 mM sodium pyruvate and 70 mM benzaldehyde. As shown in Figure 5.3 for an analysis of GC profiles, 95 % benzaldehyde was converted to L-P AC in the presence of ethanol, while 78 % benzaldehyde was converted to L-P AC in the absence of ethanol. It was shown clearly that the biotransformation yield (based on initial benzaldehyde) was significantly enhanced by the addition of ethanol. This enhancement can possibly be attributed to the presence of a strongly lipophilic zone around the active site of the PDC which acts to stabilize the cx.-carbanion of HETPP for suitable conversion to L-P AC.

5.3 EFFECT OF pH ON THE RA TES OF L-PAC AND ACETALDEHYDE FORMATION

The pH dependence of L-P AC and acetaldehyde formation at various pH values in the range of 3.5 to 8.0 was examined with and without 2.0 M ethanol at 4°C. The reaction rates could be measured with citrate/sodium phosphate buffer in the pH range 4.0 to 8.0, however the enzyme appeared to be denatured at pH 3.5. The results in Figure 5.4(a) shows that the effect of pH on the rate of L-P AC formation did not relate directly to the pH dependence of free acetaldehyde formation. These results indicated that while the optimum pH for acetaldehyde formation was 6.0, optimum for L-PAC formation was at pH 7.0. Quantitative comparison of reaction rates demonstrated that the rate of L-PAC at pH 7.0 was 3 times as high as at pH 6.0, while the rate of acetaldehyde formation at pH 7.0 was half that at pH 6.0. Further comparison of kinetic parameters at pH 6.0 and pH 7.0 is summarized in Table 5.1 illustrating that at pH 7.0, high L-PAC formation was accompanied by relatively low acetaldehyde formation, which is in general agreement with previous results in Chapter 4. Chapter 5 165

(a)

.22 with 2.0 M Ethanol 1 .49 2.05 'L-PAC

3 .43 PEAKU AREA% RT AREA BC 1 0.008 0. 11 1192 02 2 89.684 0.22 14194207 08 3 0.024 t .49 3782 05 4 9.561 2.05 1513209 05 5 0.723 3.43 114454 05 TOTAL 100. 15826844

(b) .22 . 63 Benzaldehyde without Ethanol 2.04 -.. L-PAC

PEAKU AREA% RT AREA BC

1 0.008 0.08 1280 02 2 86.571 0.22 14425000 02 3 3.53 0.63 588164 02 4 0.516 t .49 86041 02 5 8.726 2.04 1453997 08 6 0.649 3.43 108101 05 TOTAL 100. 16662583

Figure 5.3 Comparison of GC analysis profiles for biotransformation products (a) with; (b) without 2.0 Methanol. Reaction mixtures contained 70 mM BZ, 140 mM sodium pyruvate and in the presence, or absence of 2.0 M ethanol at 4°C. Chapter 5 166

120_:...__:__(a) ______-,

- 100 ~ -QJ -~ 80 C: 0

60

QJ > ·­-ie 40 -QJ c::::: 20

3.0 4.0 5.0 6.0 7.0 8.0 9.0 pH (b) 120------,

100 -2ft. - Acetaldehyde r• "'o • C: L-PAC~ 0 • ~ - 40 ie • -QJ c::::: \. 20

.J/ \ 0 0 -+--~___,;::-=--,,----r---r---r---i.-----,-----r---,--T--,.L•..,./ 3.0 4.0 5.0 6.0 7.0 8.0 9.0 pH

Figure 5.4 Comparison of effect of pH on relative reaction rate for acetaldehyde and L-PAC formation at 4°C: (a) with 2.0 M ethanol; (b) without ethanol. Reaction mixtures contained 70 mM benzaldehyde, 70 mM sodium pyruvate and 7 unit/ml PDC in citrate/sodium phosphate buffer at various pH. Chapter 5 167

Table 5. 1 Comparison of initial rates for biotransformation products over first 30 min incubation at pH 6.0 and pH 7.0 at 4°C. The reaction mixture contains 70 mM benzaldehyde, 70 mM sodium pyruvate, 2.0 Methanol and 7 unit/ml PDC.

Kinetic values (mM/min) pH -ds pyruvate d pacetaldehyde dp L-PAC dp acetoin dt dt dt dt

6.0 0.80 0.17 0.35 0.14 7.0 1.26 0.08 1.10 0.09

The finding of a relatively high l-PAC formation rate at pH 7.0 is in agreement with a previous investigation (Yatco-Manzo et al., 1959) which reported that when sodium pyruvate was incubated with TPP in slight alkaline solution, pyruvate was gradually decarboxylated even in the absence of PDC enzyme. The rate of decarboxylation was maximum at pH 8.9, and was insignificant at acidic pH. Thus, it seems that slight alkaline or neutral conditions may be more favourable for l-PAC formation. Even under favourable conditions for acyloin formation at pH 7.0, low acetoin formation was revealed. The reason is likely to be due to the relatively low concentration of free acetaldehyde at this pH value. Although almost the same types of curves were obtained in the absence of ethanol for the various pH ranges (Figure 5.4 (b )), addition of ethanol slightly extended the catalytic activity of PDC enzyme for both acetaldehyde and l-PAC formation towards the alkaline pH region.

5.4 EFFECT OF PDC ACTIVITY ON THE INITIAL RA TE OF L-PAC FORMATION

Initial reaction rates for l-PAC formation at various PDC activities from 4.0 to 10.0 unit/ ml were determined over first 30 min incubation. The reactions were carried out with various concentrations of benzaldehyde, 3 Chapter 5 168

times higher molar ratio of sodium pyruvate to benzaldehyde and 2.0 M ethanol in 40 mM KH2PO4 buffer (pH 7.0) at 4°C. As shown in Figure 5.5(a), it was observed that the initial L-PAC formation rates were greatly dependent on PDC activity and benzaldehyde concentration. An approximately linear increase in initial rates for L-PAC formation was evident only in the range of from 30 to 70 mM benzaldehyde. The highest L-PAC formation rates with 4.0, 5.0 and 7.0 unit/ml PDC occurred at 70, 100 and 120 mM benzaldehyde, respectively. Above 7.0 unit/ml PDC, the L-PAC formation rates were significantly increased with higher levels of benzaldehyde. Generally, it appeared that the higher initial rates for L-PAC formation occurred at higher benzaldehyde concentration and increasing PDC activies. However, after a peak value of L-PAC formation, the toxic effect of further increasing benzaldehyde was evident and dependent on PDC activity.

Specific L-P AC production for each PDC activity was computed in order to estimate the enzyme efficiency. As shown in Figure 5.5(b), 3 different trends can be identified over the range of benzaldehyde concentrations tested. Firstly, in the range of 30 to 70 mM benzaldehyde, specific L-PAC formation rates were the higher with the lower PDC activities. It seems that the catalytic capacity of PDC was not completely saturated with benzaldehyde in this range. Secondly, in the range of 100 mM to 150 mM benzaldehyde, the highest specific L-P AC formation was found at 7 unit/ml PDC activity rather than 8.2 and 10.0 unit/ml PDC. It is evident that further increasing PDC activity increased acetaldehyde formation, and therefore competed with L-PAC formation (Table 5.2). Lastly, for 150 to 200 mM benzaldehyde, the capacity for L-PAC formation was significantly inhibited. However, the extent of inhibition by benzaldehyde was more evident for the lower levels of PDC activity.

It is evident that the higher activities of PDC resulting in the higher reaction rates were readily able to carry out biotransformations even in the high range of benzaldehyde concentrations (i.e. from 120 to 200 mM). This was presumably due to the high initial conversion rates with high PDC activity which resulted in decreasing the concentration of benzaldehyde relatively quickly with a consequent reduction in its toxic effects. Chapter 5 169

(a)

14.0 • PDC activity (unit/ml) • 4.0 :§- 12.0 - o 5.0 eo • 6.0 o 7.0 -Q.I 10.0 - -...re • 8.2 § 8.0- C 10.0

·­V -i,: ...Q,I 6.o- ·­-....i,: ·--c 4.0· 2.0-

40 60 80 100 120 140 160 180 200 220 Benzaldehyde (mM) (b) 2.0 ------, PDC activity (unit/ml)

1.8 • 4.0 -..c :.:i o 5.0 C 1.6 :, • 6.0 eo O 7.0 E 1.4 • 8.2 -Q,I .... 1.2 ...re C 10.0 C 1.0 0 ·­....u re 0.8 ...Q,I V 0.6 -=·c Q,I c.. 0.4 rJl 0.2

0.0 -+------...... ,...... "'T"""....,...... "'T""""'T"""""r-"'T""""'T"""""r-""r-,...... ,...... ~~~~-r---i 0 20 40 60 80 100 120 140 160 180 200 220 Benzaldehyde (mM)

Figure 5.5 Effect of PDC enzyme activity (unit/ml) on (a) initial reaction rate; (b) specific rate for L-PAC formation in the presence of 3 times higher mole of sodium pyruvate and corresponding benzaldehyde. The initial reactions were determined within 30 min at pH 7.0 and 4°C. Chapter 5 170

Table 5.2 Acetaldehyde formation with various PDC activities in the presence of equi-molar benzaldehyde and sodium pyruvate at 4°C.

PDC Acetaldehyde (mM) activity substrates (mM) (unit/ml) 30 50 100 150 4.0 6.1 9.5 "'2.3 "'n.d 5.0 6.5 10.2 7.5 "'n.d 6.0 7.1 10.8 10.5 8.3 7.0 7.5 11.0 11.4 4.8 8.2 8.2 12.0 16.0 9.2 10.0 9.1 13.5 17.8 10.1

* Reaction did not proceeded fully due to toxicity of benzaldehyde.

5.5 EFFECT OF MOLAR RATIOS OF PYRUV A TE TO BENZALDEHYDE ON THE INITIAL RA TES FOR L-PAC FORMATION

Since both pyruvate and benzaldehyde are required for L-P AC formation, it is possible that a high concentration of one substrate in the PDC reaction mixture may bring about significant formation of unwanted acetaldehyde or enzyme denaturation. As mentioned previously, since PDC catalyses the conversion of pyruvate to acetaldehyde during biotransformation, higher concentrations of pyruvate are required to achieve higher conversion yields of L-PAC. However, excess pyruvate may cause substrate inhibition. Thus, in order to determine optimum pyruvate concentrations in relation to benzaldehyde, initial reaction rates were evaluated with various molar ratios of pyruvate to benzaldehyde. Initial reaction rates for L-PAC were determined with 7 unit/ml PDC and 2.0 M ethanol in 40 mM KH2PO4 buffer (pH 7.0) at 4°C. It was found that initial reaction rates for L-PAC formation increased rapidly with increasing ratios of pyruvate to benzaldehyde in the range of 0.3 to 0.5 mole (Figure 5.6). At higher benzaldehyde levels, initial L-PAC Chapter 5 171

16.0 Cone. of benzaldehyde

14.0 • SOmM • 70mM -.c o lOOmM -...J 12.0 • 120mM eh Cl 150mM -QJ • 180mM ..ns 10.0 ... 6 200mM C: ..0 8.0 ·-V ns QJ ... 6.0 • • -..ns ·--C: 4.0 2.0

0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Molar ratio of pyruvate to BZ

Figure 5.6 Initial reaction rates for L-PAC formation as a function of benzaldehyde concentration and various molar ratios of pyruvate to benzaldehyde. Reaction mixture contained 7 unit/ml PDC, 2.0 Methanol and indicated substrates in 40 mM KH2PO4 buffer (pH 7.0) at 4°C. Chapter 5 172

formation rates were almost constant in the range of 150 mM to 200 mM benzaldehyde. With 200 mM benzaldehyde and 3.0 times higher ratios (i.e. 600 mM pyruvate), the initial rate of L-PAC formation decreased significantly to a half of that with equal moles of pyruvate. Substrate inhibition by pyruvate was more evident at relatively high benzaldehyde concentrations. Excess pyruvate is likely to cause increased acetaldehyde formation resulting in either non-competitive inhibition or competitive inhibition with pyruvate/benzaldehyde, or acetaldehyde/benzaldehyde. In general, it was observed that the inhibition by pyruvate occurred above 180- 200 mM pyruvate, and the initial reaction rates were maximal between 1.0 to 1.5 molar ratios of pyruvate to benzaldehyde.

5.6 DETERMINATION OF SATURATION CONSTANT (Ks) FOR BENZALDEHYDE

As described previously in this study, the rate of L-PAC formation was significantly affected by various factors, viz. temperature, pH, addition of organic solvent, PDC activity and substrate concentrations. Thus, it is difficult to represent the saturation constant (K5) for benzaldehyde in the general PDC reaction, but it may be determined restrictively under the given reaction conditions. On the basis of present results, a determination of K5 value was conducted by measuring the initial rate for L-PAC formation at 4°C over the first 30 min. The reaction mixtures consisted of 40 mM KH2PO4 reaction buffer (pH 7.0) containing 7 unit/mL PDC, 200 mM sodium pyruvate, 2.0 M ethanol and various initial concentration of benzaldehyde. As the results show in Figure 5.7, the initial reaction rate increased with increased benzaldehyde concentration up to 180 mM. An apparent Ks was determined as 42 mM benzaldehyde which is similar when compared with other results in the literature (viz. 50 mM. Bringer-Meyer and Sahm, 1988). It is possible that the slightly enhanced affinity of PDC toward benzaldehyde in the present study may be due to co-operative effects, viz. presence of ethanol, optimization of pH, PDC from C. utilis and lower reaction temperature. Chapter 5 173

1s.o------

16.0 ~ ::J- eo 14.0 ------0-0 -u 0-- ---0 < 12.0 ~ I ..,J i.. 10.0 0 -....QJ 8.0 ~ i.. C 0 6.0

4.0

2.0 0.0 +-...... _.....-_...____ .,.....,...-,, ______...... ,...-,, ______..

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 Benzaldehyde (mM)

Figure 5.7 Determination of K5 for benzaldehyde. Reaction mixture consisted of 40 mM KH2P04 reaction buffer (pH 7.0) containing 7 unit/ml P DC, 200 mM sodium pyruvate, 2.0 M ethanol and various initial concentration of benzaldehyde. Chapter 5 174

5.7 EFFECT OF MOLAR RATIOS OF PYRUV A TE TO BENZALDEHYDE ON FINAL L-PAC CONCENTRATIONS

The final conversion yields of L-P AC as a function of the molar ratios of pyruvate to benzaldehyde were evaluated with 7 unit/mL PDC, 2.0 M ethanol in 40 mM KH2PO4 buffer (pH 7.0) at 4°C. Biotransformation was considered complete when L-PAC production stopped. The results presented in Figure 5.8 indicate that an increase in molar ratio was accompanied by an increased final L-PAC yield, up to a 2.0 molar ratio. Excess pyruvate did not result in enhancement of L-P AC yield. The requirement for pyruvate to achieve higher conversion yields was dependent on the benzaldehyde concentration used. It was found that higher benzaldehyde levels required a lower molar ratio to achieve the maximum conversion. For instance, the molar ratio requirements with 50, 70, 100 and 150 mM benzaldehyde to achieve over 90 % yield (Yp/s: mole

Conversion yields above 90 % (mole/mole) based on pyruvate were achieved only in the range of 150 mM to 200 mM benzaldehyde when an equal or less molar ratio were used. In other cases, more fraction of pyruvate was presumably converted to by-products (free acetaldehyde and acetoin) or remained in excess concentrations in the reaction mixture. Chapter 5 175

110

Cone. of benzaldehyde 100 a ; ==•- t • SOmM 90 • 70mM 0 lOOmM

-0~ 80 • a 120mM -0 0 l""'4 lSOmM >< 70 • -QJ l 6 180mM -0 E A 60 200mM -QJ -0 E -Cl) 50 Q.. ->-

40

30

20

10

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Pyruvate/BZ (mole/mole)

Figure 5.8 L-PAC yields (mole/mole %) based on benzaldehyde as a function of various molar ratios of pyruvate to benzaldehyde in the presence of various concentrations of benzaldehyde with 7 unit/mL PDC enzyme at pH 7.0 and 4°C. Chapter 5 176

Table 5.3 L-PAC formation with various molar ratios of pyruvate to benzaldehyde at 4°C. The reaction mixture consisted of 40 mM KH2PO4 (pH 7.0) containing 7 unit/mL PDC, 2.0 Methanol and indicated concentrations of sodium pyruvate and benzaldehyde.

Final concentration of L-PAC (g/L) and molar conversion yields (%) Molar ratio of pyruvate to benzaldehyde BZ!mM) 0.3 0.5 1.0 1.2 1.5 2.0 2.5 3.0 1.6 2.3 5.0 5.9 6.8 7.2 7.2 7.2 *a 50 69.3 60.5 67.1 65.9 60.6 47.8 38.4 32.o*b 20.8 30.2 67.1 78.1 90.9 95.7 96.0 96.o*c 2.3 3.8 7.9 9.5 9.9 10.0 10.0 9.9 70 72.8 72.0 75.2 75.1 62.8 47.6 38.1 31.4 21.9 36.0 75.2 90.2 94.3 95.3 95.3 95.3 3.5 5.8 11.7 14.2 14.5 14.5 14.5 14.4 100 78.4 77.0 78.2 78.3 64.5 48.4 38.6 32.0 23.3 38.3 78.2 94.6 96.7 96.7 96.7 96.6 4.5 7.0 14.6 16.8 17.6 17.6 17.6 17.4 120 83.3 77.2 81.1 7.60 65.2 48.9 39.0 32.0 25.0 38.6 81.8 93.2 97.5 97.5 97.5 96.6 6.3 10.9 20.5 21.5 22.0 22.0 22.1 22.0 150 92.8 97.0 90.6 79.6 64.9 48.9 39.30 32.5 27.8 48.6 90.6 95.3 97.8 97.8 98.2 97.8 7.4 12.1 24.2 25.1 25.5 25.5 25.0 25.1 180 91.8 89.5 90.0 77.5 63.0 47.2 37.0 31.0 27.2 44.4 90.0 93.0 94.4 94.4 92.6 93.0 8.3 13.3 22.7 24.9 27.9 28.6 28.0 28.0 200 91.6 88.5 75.6 69.1 62.0 47.6 37.3 31.0 27.5 4.2 75.6 82.9 93.0 95.3 93.0 93.0

*a L-P AC (g/L) *b Yp/s (pyr) xl00 (%) *c Yp/s (benzaldehyde) x 100 (%) Chapter 5 177

5.8 EFFECT OF PDC ACTIVITIES ON L-PAC FORMATION WITH VARIOUS MOLAR RATIOS OF PYRUV A TE TO BENZALDEHYDE

On the basis of the previous results, an attempt was made to achieve higher L-PAC concentrations with higher PDC activities and substrate concentrations. The biotransformations were carried out with in the range of PDC activities of 7.0 to 20.1 unit/mL, 200 mM to 300 mM BZ and 1.0 to 1.5 molar ratios of pyruvate to benzaldehyde. The results shown in Table 5.4 indicate that although the higher levels of PDC activities can function at the higher substrate concentrations, the highest L-PAC formation did not exceed 27.9 g/L. It was evident that higher benzaldehyde concentrations were required for higher L-P AC formation with increasing PDC activity (Figure 5.9). It seems that the higher PDC activity resulted in relatively high free acetaldehyde formation which caused 'active acetaldehyde' depletion and product inhibition.

5.9 TYPICAL BIOTRANSFORMA TION KINETICS WITH PURIFIED PIX:

The kinetics of typical time course of a biotransformation with purified PDC were evaluated with 150 mM benzaldehyde, 225 mM sodium pyruvate, 7 unit/mL PDC and 2.0 Methanol in 40 mM KH2P04 buffer (pH 7.0). The results shown in Figure 5.10 demonstrated the time dependence of simultaneous L-P AC, acetaldehyde and acetoin formation together with biotransformation of pyruvate and benzaldehyde with PDC. In the first 2-3 h, L-PAC formation increased rapidly, further L-PAC formation occurred more slowly and was influenced by substrate depletion and possible product inhibition. After 6 h incubation with maximum L-PAC (22 g/L), a mass balance on pyruvate indicated that of the original 225 mM pyruvate, 147 mM pyruvate contributed to L-PAC formation, 25 mM pyruvate was converted to acetoin formation, 22 mM pyruvate was converted to free acetaldehyde. The residual pyruvate was 30 mM indicating a closing mass balance based on pyruvate. Furthermore after 6 h incubation, 20-30 % of initial PDC activity still remained. Chapter 5 178

Table 5.4 Effect of PDC activities on L-P AC formation with various molar ratios of pyruvate to benzaldehyde.

Final L-P AC concentration (g/L)

BZ Pyruvate to BZ PDC activity (unit/mL)

(mM) (mole/mole) 7.0 10.2 15.0 20.1 200 1.0 22.7 18.2 17.6 15.5 1.2 24.9 20.7 15.9 15.3 1.5 27.9 23.9 16.2 15.0 230 1.0 20.4 25.3 24.3 22.5 1.2 25.6 24.8 25.6 23.2 1.5 21.0 20.3 24.8 19.4 250 1.0 13.7 22.8 23.5 24.8 1.2 10.7 20.3 20.8 25.6 1.5 9.1 9.2 20.5 21.2 280 1.0 5.1 6.5 7.9 23.3 1.2 5.0 6.3 8.1 22.1 1.5 4.1 6.7 10.5 12.7 300 1.0 1.1 1.2 3.0 4.7 1.2 1.2 1.3 2.8 4.7 1.5 1.0 1.7 2.7 4.5 Chapter 5 179

30 PDC activity (unit/ml) 28 26 o 7.0 24 • 10.2 22 D 15.0 20 • 20.1 -,..;i 18 ~ -u 16 14

Figure 5.9 Comparison of the final L-P AC formation with various PDC activities, and equal molar concentration of benzaldehyde and sodium pyruvate in the range of 200 to 300 mM. Results shown are for a molar ratio of 1.0. Chapter 5 180

24 o Benzaldehyde • L-PAC 22 / ~--·-·-·. • Acetaldehyde & Acetoin : ~ /" c Pyruvatc C ....0 ·-l'C ....""' C :: Q,I ~ / '\~ V 10 C 0 u 8 0 1'\ 6 4 \:"a'- ~ c--c-c 2 o~~~~;...... ,,...... ;~_Jjl:::::;::=q.;=;::=9==;==9=::;::::9=;::==-.;,-,,--•=t=• e~_._.~e 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Time (h)

Figure 5.10 Biotransformation kinetics: Reaction mixture consists of 40 mM KH2P04 buffer (pH 7.0) containing 150 mM benzaldehyde, 225 mM sodium pyruvate, 2.0 Methanol and 7 unit/mL PDC at 4°C. Chapter 5 181

5.10 DISCUSSION AND CONCLUSIONS

Efficient L-P AC formation using purified PDC has been successfully demonstrated in the present study. The introduction of a water-miscible organic solvent, viz. ethanol, resulted in improved L-P AC production. It is likely that addition of ethanol provides a range of positive effects: increased solubility of hydrophobic substrate, enhancement of reaction rates and improved conversion efficiencies and prolongation of enzyme stability in reaction buffers. The use of an organic solvent such as ethanol appeared to enhance lipophilic interaction between substrates and enzyme. The optimal pH for L-PAC formation did not correspond to optimal pH for acetaldehyde formation. It was evident that the PDC exhibited slightly different characteristics for its minor reaction, thereby, allowing a pH to be selected (viz. pH 7.0) which optimized L-P AC production, and minimized free acetaldehyde release at 4 °C. The influence of PDC activity on L-P AC formation was evaluated. It was found that L-PAC formation was significantly influenced by PDC activity and substrate concentrations. During biotransformation, L-P AC, acetaldehyde and acetoin were all produced as a result of the catalytic activity of the purified PDC. The molar ratios of substrates also influenced initial reaction rates, final L-P AC yields and by-product formation. From the present results, the optimum concentrations of substrates to achieve maximum biotransformation, can be determined. On the basis of the present results, the highest L-P AC concentration (viz. 28.6 g/L) was achieved with 200 mM benzaldehyde, 2.0 pyruvate molar ratio to benzaldehyde (i.e. 400 mM pyruvate) and 7 unit/mL of purified PDC in 40 mM KH2PO4 buffer (pH 7.0) containing 2.0 M ethanol at 4°C. After biotransformation was completed, 20-30 % initial PDC activity was still apparent, indicating the possibility for further product formation. Although the nature of PDC/substrate interactions are not fully elucidated, it was clear that L-PAC formation is associated with a number of complex interactions with variables such as addition of ethanol, optimized pH, the level of PDC activity, substrate concentrations and inhibiting product concentrations all being significant. Chapter 6 182

CHAPTER 6

KINETIC EVALUATION OF BIOTRANSFORMATION OF BENZALDEHYDE TO L-PHENYLACETYLCARBINOL BY IMMOBILIZED PYRUVATE DECARBOXYLASE

6.1 INTRODUCTION

It is now increasingly recognized that immobilized enzyme processes may offer significant advantages over free enzyme systems. The main advantages with immobilized enzymes are the possibility of using a continuous process for long term operation, maintenance of enzyme activity, high catalytic density, clean product streams and in some cases modification of the properties of the enzyme. It has been claimed often that immobilization of an enzyme on an insoluble support will result in enhanced stability against the denaturing effects of heat, pH, toxic substances etc. (Lenders et al., 1985). An enhanced resistance to generally unfavourable conditions can be of particular importance for immobilized enzymes in a continuous process, and result in minimal need for renewal of the biocatalyst. It has been reported although enzyme activity is generally lowered through immobilization, that much of the original activity can be retained by using physical or chemical treatment with insoluble supporting materials. A large number of supports including natural and synthetic polymers as matrices has been used for immobilizing enzymes. The choice of the support is largely governed by the properties required in the enzyme application. The physicochemical characteristics of such a matrix can have an effect on the reaction and can be critical parameters in selecting a matrix for a particular process. Despite the fact that much research has been devoted to the immobilization of biocatalysts, their technological exploitation so far has been rather limited. The advantages of using immobilized biocatalysts are often overshadowed by drawbacks such as diffusion limitations, problems with steric hindrance and limited stability of the biocatalysts, as well as the costs of immobilization and the supporting materials. Chapter 6 183

One of the potential advantages of an immobilized enzyme process, viz. an enhanced resistance to toxic substrates and/or products, strongly suggests the evaluation of an immobilized PDC process for L-PAC formation. As discussed previously in Chapters 4 and 5, the substrate, benzaldehyde, as well as products, acetaldehyde and L-PAC, can cause significant inhibition or inactivation of the PDC enzyme. A continuous process with immobilized PDC may be therefore a possible means to improve long term L-P AC production. In evaluating a continuous process with immobilized PDC, it is necessary to carefully consider various influencing factors as follows: the characteristics of the immobilized enzyme, concentration of substrate, operational requirements of the reactor, reaction kinetics, enzyme loading capacity, mass transfer characteristics and ease of enzyme replacement etc.

The aims of this Chapter are as follows: (1) to immobilize the PDC from C. utilis; (2) to characterize the immobilized PDC; (3) to evaluate the kinetic parameters for L-P AC formation with immobilized PDC in batch and continuous processes.

6.2 IMMOBILIZATION OF PYRUV A TE DECARBOXYLASE (PDC)

6.2.1 Manometric measurement of PDC activity

It was essential to develop an assay for immobilized PDC activity under well defined conditions to obtain useful information on the overall characteristics of the bound enzyme. Indirect determination of POC activity was conducted by measuring the liberation of C02 from pyruvate substrate using a Warburg manometer (B. Braun) with shaking at 100 rpm and 25°C. The Warburg manometer contained 1 mL 200 mM potassium citrate buffer (pH 6.0) with 50 µM TPP, 0.3 mM MgS04• and 1 mL of free or immobilized PDC enzyme. When the reaction was started by the addition of 1 mL of 200 mM sodium pyruvate from the side arm of a manometer, the amount of Chapter 6 184

CO2 released, as indicative of PDC activity, was plotted as a function of time (see Appendix 3). The results shown in Figure 6.1 indicate that CO2 evolution could be correlated linearly with PDC activity over the first 4 min incubation. With low PDC activities from 1.0 to 2.5 unit/mL, CO2 evolution was linearly proportional over a longer period of time. Particularly for high PDC activities, reaction rates decreased with increasing incubation times. As a consequence, PDC activities corresponding to CO2 evolution could be reasonably determined in the range of 0-25 µmole of CO2. This was measured over the first 2-3 min incubation. For measuring PDC activities above 10 unit/mL, the dilution of sample volumes was required.

6.2.2 Adsorption of PDC on cationic exchange resins

As previously described in Sections 4.2 and 5.3, PDC stability could be well maintained in 40 mM potassium phosphate buffer with the optimum pH for L-P AC formation at 7.0. On the basis of these results, it was decided that 40 mM KH2PO4 buffer (pH 7.0), containing 2.0 M ethanol, 50 µM TPP and 0.3 mM MgSO4 would be consistently used for washing and equilibrating the ionic exchange resins. The choice of buffers followed the general rule that anionic buffers (i.e. phosphate) should be used with cationic exchangers. The cationic exchangers Amberlite IRC-50, Amberlite IR-200, CM-Sephadex and Dowex 50W-X8, were activated with 0.5 N NaOH and 0.5 N HCl, and then equilibrated with the reaction buffer. After each resin was equilibrated with 40 mM KH2PO4 buffer (pH 7.0), it was gently mixed with enzyme solutions containing various activities of PDC (2.6 unit/mg protein of specific activity) for 24 h at 4°C. Unbound PDC was washed away with the reaction buffer. The immobilization capacity of each cationic exchanger was determined by measuring CO2 evolution with a Warburg manometer at 25 °C. The computed immobilized PDC activities, determined from the CO2 evolution of 1 mL of prepared cationic exchanger, were plotted as functions of the free enzyme activities (Figure 6.2). Chapter 6 185

40.0 PDC activity (unit)

El 1.0 35.0 • 2.0 o 2.5 30.0 • 3.5 -QJ • 5.0 -0 25.0 e a 7.0 ::i. 6 9.0 - 20.0 8 6 10.0 15.0

10.0

5.0

0 2 4 6 8 10 12 14 16 18 20 22 Time (min)

Figure 6.1 Manometric measurement of PDC activities. The total C02 evolution was measured from reaction mixtures containing different PDC activities. with 200 mM sodium citrate buffer (pH 6.0) made to 2.0 ml and 1 ml 200 mM sodium pyruvate at 100 rpm and 25°C. Chapter 6 186

6.0 o Amerlite IRC-SOH 5.5 • Amberlite-lR 200 5.0 c CM-Sephadex ~o -c:: 4.5 • Dowex SOW-XS rll ·-~ ... 4.0 /0 ..,J e 3.5 ";;:i c:: 3.0 :s -....>. 2.5 0 I ·-....> 2.0 .,,,--.----·----· ·-V . < 1.5 1.0 0.5 f _-a---c::===== I 0.0 ·----=~--- . 0 10 20 30 40 so 60 70 80 90 100 110 120 130 140 PDC activity (unit/ml)

Figure 6.2 Immobilized PDC activity of various cationic exchangers loaded with various PDC activity solutions. Chapter 6 187

It was revealed that generally the PDC activity of immobilized beads increased with increasing PDC activity in solution. However, for Amberlite IR-200 and CM-Sephadex, the activity remained constant with further increments of PDC activity from 45 to 60 unit/mL PDC activity. It seems that the binding capacities of both cationic resins are saturated with PDC under these conditions. The highest activity of 5.0 unit/mL beads was achieved with Amberlite IRC-50 H adsorbed from a 120 unit/mL PDC enzyme solution.

6.2.3 Entrapment of PDC enzyme into gel matrix

Evaluation of entrapment of PDC into two inert materials, calcium alginate and polyacrylamide, was also conducted. Various PDC activity solutions were mixed with 2 % sodium alginate and extruded into 1.0 % calcium chloride solution, and then stabilized for 4 hat 4°C. For entrapment into polyacrylamide gel, the enzyme solutions were mixed with 10 % acrylamide (containing 5 % bis-acrylamide) and polymerized with 0.004 % riboflavin and N ,N ,N' ,N' -tetramethylethylenediamine (TEMED) (Hames, 1981). The polymerized matrix was cut into 1-2 mm diafueter particles and regularly sized particles were collected through experimental sieves (size 1000 and 2000 µm). As illustrated in Figure 6.3, the PDC activities in the polyacrylamide gels were almost proportional to increasing PDC activities and were much higher than with calcium alginate. The highest activity of 8.8 unit/mL beads was achieved with entrapment by polyacrylamide with 90 unit/mL PDC enzyme solution, while 4.5 unit/mL was the peak PDC activity with 2 % calcium alginate.

A comparison of the capacities of various matrices for immobilization of PDC from a 90 unit/mL PDC solution is summarized in Table 6.1. It was shown that entrapment methods may provide higher immobilization capacities for PDC compared with adsorption on cationic exchanger resins. It is possible that the high molecular weight and tetrameric structure of PDC may provide more stability in an inert gel matrix than attachment on the surface of ion exchanger resins. Chapter 6 188

10.0 9.0 Ac/• 8.0 ~ -"O tU • QI 7.0 ,.Q ...J e 6.0 Alginate ....~ C 5.0 ::s 0 ....-....>. 4.0 ....> 3.0 V <- 2.0

1.0

0.0 0 10 20 30 40 so 60 70 80 90 100 PDC activity (unit/ml)

Figure 6.3 Immobilized PDC activity with entrapment in 10 % polyacrylamide and 2 % calcium alginate beads compared with various activities of original PDC solution. Chapter 6 189

Table 6.1 Immobilization capacity of various matrices for PDC with 90 unit/mL PDC solution.

Matrix Activity "'Activity yield (unit/mL beads) (%) Ion-exchanger Amberlite IRC-50H 4.4 4.9 Amberlite IR-200 1.8 2.0 CM-Sephadex 0.6 0.7 Dowex 50W-X80 0.4 0.45 Entrapment Calcium alginate (2 % ) 4.6 5.1 Polyacrylamide (10 %) 8.9 9.9 "'Activity yields were calculated by reference to the original equilibrium PDC solution (i.e. 90 unit/mL PDC).

6.3 IMPROVEMENT OF ENTRAPMENT OF PDC INTO POLYACRYL­ AMIDE GEL

6.3.1 Effect of acrylamide concentrations on PDC activity of immobilized beads

The average pore size of a polyacrylamide gel can be manipulated by varying the total content of acrylamide. It was considered that the effective pore size would decrease as acrylamide concentration increases, and the diffusional mass transfer would be inversely related to acrylamide concentration. The apparent activity and retention ability of immobilized beads may be greatly influenced also by the concentration of acrylamide monomer. Various concentrations from 6.0 to 16.0 % of acrylamide monomer (containing 5 % bis-acrylamide) were examined for immobilization of PDC. In fact, immobilizations were conducted with 8.0 % to 16.0 % acrylamide Chapter 6 190

due to immobilization not being successful with 6 %. After immobilization, PDC activities were measured at 25°C. As the results show in Figure 6.4, the highest activity was achieved with 10 % acrylamide. It was considered that although 12 % and 16 % of acrylamide may retain PDC more tightly, which resulted in higher mechanical stability and longer life, limitation of mass transfer in high concentrations of the gel matrix could occur more readily.

6.3.2 Effect of glutaraldehyde concentration on activity of PDC immobilized in polyacrylamide gel

As a bifunctional coupling agent which may increase the binding between matrix and enzyme, glutaraldehyde was evaluated also for PDC immobilization. Various concentrations of glutaraldehyde were added to immobilization mixtures containing 10 % acrylamide solution and 120 unit/mL PDC (2.6 unit/mg protein of specific activity). The immobilization procedure was as described previously. The PDC activity of immobilized beads (Figure 6.5) indicated that a gradual increase in PDC activity was observed with increasing glutaraldehyde up to 0.3 % (v /w). Further increasing glutaraldehyde has a negative effect possibly due to inactivation or deterioration of the enzyme. The PDC activity yields were significantly enhanced up to 14 % of free PDC enzyme solution with 0.3 % glutaraldehyde, compared with 9.9 % in the absence of glutaraldehyde.

6.3.3 Effect of glutaraldehyde on stability of immobilized PDC

When PDC was entrapped into polyacrylamide in the presence of glutaraldehyde, other factors to be considered included the binding capacity of the supporting matrix as well as possible denaturation of the enzyme by the coupling agent. The optimum concentration of glutaraldehyde to maintain PDC activity was determined by measuring the residual PDC activity of polyacrylamide beads containing various Chapter 6 191

12.0

0~ 10.0

0~ -'#- 8.0 -....>, 0 .... 0 ...... > V 6.0 QJ"' ...... > "' 4.0 -=QJ

2.0

0.0 +---"T"""--"T"""--,--.....---,--...----.--...... -...------.1 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 % Concentration (w/v)

Figure 6.4 Effect of acrylamide monomer concentrations on apparent PDC activity of immobilized bead entrapping from 120 unit/ml PDC activity solution. The relative activity (%) was determined by reference to the activity of the original PDC solution. Chapter 6 192

20.0

18.0

16.0

-0~ 14.0 ->. 12.0 ·-....-> ·-V 10.0

"'~ > ·-.... 8.0 0 "'~ -c=:: 6.0

4.0

2.0 0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Glutaraldehyde % (w/v)

Figure 6.5 Effect of glutaraldehyde concentrations on PDC activity of 10 % polyacrylamide beads entrapping from 120 unit/ml PDC activity. The relative activities (%) were determined by reference to the activity of the original PDC solution. Chapter 6 193

concentrations of glutaraldehyde after incubation for 10 days at 4°C. Figure 6.6 illustrates the relative residual activities of immobilized beads as a function of incubation time compared to initial PDC activities. Generally, it appeared that higher glutaraldehyde concentrations prolonged the life of immobilized PDC although a negative effect after 14 days was apparent at 0.5 % glutaraldehyde. Over 60 % of initial PDC activity was maintained with 0.3 % glutaraldehyde even after 20 days incubation at 4°C. From the results in Figures 6.5 and 6.6, 0.2 % glutaraldehyde was selected as a reasonable concentration to retain PDC activity in the gel matrix while sustaining a longer life for the immobilized PDC.

6.3.4 Preparation of spherical polymer beads

On the basis of previous investigations that spherical beads of polyacrylamide have excellent mechanical stabilities and facilitate high flow rates in a column process (Nilsson et al., 1972), preparation of spherical beads was conducted with PDC and 10 % acrylamide. The bead polymerization process was carried out in a 1 L beaker with magnetic stirrer, and an inlet and outlet for N2 (Figure 6.7). The beaker was cooled in ice water, and contained 400 mL of hydrophobic phase, toluene/chloroform (290/110, v/v), in order to achieve the same density as the aqueous phase. This former phase contained 1-3 mL of suspension stabilizing agent, sorbitan sesquioleate. To the cold monomer solution, consisting of 5.7 g acrylamide, 0.3 g. bis-acrylamide, 0.2 % glutaraldehyde and 120 unit/mL PDC activity (specific activity, 2.6 unit/mg protein) in 200 mM potassium citrate buffer (pH 6.0) in a total volume 58.5 mL, 0.5 mL TEMED and 1.0 mL ammonium persulphate solution (0.6 g/mL) were added. The polymerization was completed within 40 min. The beads were washed on a glass filter, first with toluene to remove the chloroform, and then with several liters of 40 mM KH2PO4 buffer (pH 7.0). As demonstrated in Figure 6.8 (a), bead diameters were distributed in the range of 0.4-1.2 mm. The activity the spherical beads was 12.S % of original PDC activity, and active immobilized PDC beads stained with fuchsin/sulphurous acid (Feigl and Anger, 1966) are shown in Figure 6.8 (b) and (c). Chapter 6 194

110 Cone. of glutaraldehyde 100 a 0%

-0~ 90 • 0.05 % >. -.... 80 o 0.1 % ....-> a 0.2% V 70 -IU • 0.3 % IU -::s 60 A 0.5% ...."0 ell Q,I 50 i. Q,I ...... > 40 a IU 30 -c:::Q,I 20

10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (day)

Figure 6.6 Effect of glutaraldehyde concentrations on stability of PDC activity of immobilized beads at 4°C. Chapter 6 195

N2 Gas - =:::;i

----=- N2 Gas

Magnetic stirrer

Reagents

Hydrophilic phase Acrylamide monomer 5.7 g N ,N-methylenebisacrylamide 0.3g TEMED 70-150 µl Ammonium persulphate 0.6g 120 unit/ ml PDC enzyme in 25 % glutaraldehyde 0.48ml 200 mM potassium citrate buffer containing 50 µM TPP (pH 6.0) up to 60 ml

Hydrophobic phase Chloroform 290 ml Toluene 110 ml Sorbitan sesquioleate 1-3 ml

Figure 6.7 Schematic diagram for preparation of bead-form acrylamide polymer. Chapter 6 196

Figure 6.8 (a) Photograph of spherical polyacrylamide beads containing PDC enzyme (10 x, magnification) Chapter 6 197

(b) (c)

Figure 6.8 Photograph of active stained immobilized PDC beads on (b) a plate; (c) in suspension. Chapter 6 198

6.4 CHARACTERIZATION OF IMMOBILIZED PDC

6.4.1 Determination of apparent Km for pyruvate

Determination of apparent Km for pyruvate was conducted by measuring CO2 evolution from reaction mixtures at 25°C. The reaction mixture consisted of 1 ml immobilized PDC beads (14 unit/ml) and 2 ml 200 mM potassium citrate buffer (pH 6.0) containing 50 µM TPP, 0.3 mM MgSO4 and various concentrations of sodium pyruvate. Rates of CO2 evolution were plotted against pyruvate concentration, and Lineweaver­ Burk analysis carried out as presented in Figure 6.9. The apparent Km value for pyruvate was 3.2 mM which is higher than 2.4 mM for the free enzyme at 25°C (see Section 4.4). This higher affinity constant might be due to diffusional mass transfer limitation.

6.4.2 Effect of pH on stability of immobilized PDC

The pH dependence of immobilized PDC stability was determined at various pH in the range of 4.5 to 8.5. After 10 days incubation at 4°C and various pH values of citrate/sodium phosphate buffer containing cofactors, residual activities were measured under a saturated concentration of sodium pyruvate (70 mM). The relative PDC activities (%) of immobilized beads compared to initial activity were plotted against exposed pH (Figure 6.10). The highest residual activity was revealed at pH 6.0, in which 78 % of initial activity was detected. Comparatively, it is evident that immobilized PDC has a much higher stability under the same buffer conditions after 72 h when compared with free enzyme (see Section 4.5.4). Chapter 6 199

(a) 20.0

18.0

16.0 0/o-o 0 0 14.0 -C E 12.0 -~ 7 -e0 10.0 :::1. 8.0 - 0l > 6.0 4.0 I 2.0 i

0.0 0 5 10 15 20 25 30 35 40 45 so 55 60 Pyruvate (mM)

0.6

0.5

0.4

-0.5 0.0 0.5 1.0 1.5 2.0 2.5 [1/S]

Figure 6.9 Determination of apparent Km of immobilized PDC for pyruvate. Chapter 6 200

100

90

-~0 80 ->. .. 70 ·-..> (0~ ·-V (IS 60 (IS 0 0\ -:::s 50 "'O rn ·-QJ i.. 40 / 0 QJ > .. 30 0 ·-(IS QJ -c::: 20 0 I 0~ 10 0

0 4.0 5.0 6.0 7.0 8.0 9.0 pH

Figure 6.10 Effect of pH on stability of immobilized PDC enzyme after 10 days incubation at 4°C. Chapter 6 201

6.4.3 Effect of organic solvents on stability of immobilized PDC

The stability of immobilized PDC in the presence of various concentrations of ethanol and acetone was examined. These solvents which are water miscible organic solvents, may have a possible function in the enhancement of hydrophobic interactions and increased solubilities of benzaldehyde and L-P AC. Immobilized PDC beads were exposed to various concentrations of solvents in the range from O to 6.5 M. After 10 days incubation at 4°C, residual activities were measured and plotted against both solvent concentrations. As shown in Figure 6.11, the stability of immobilized PDC increased with addition of both ethanol and acetone up to 5.0 M. Eighty eight percent of initial PDC activity was maintained in the range of 1.5-4.0 M ethanol while about 80 % initial activity was sustained with 0.5 to 5.5 M acetone. Generally, higher residual PDC activities were found in the presence of ethanol compared with acetone. It was clearly shown that both solvents provided a positive effect in maintaining activity of immobilized PDC, which is similar to the results for free enzyme (see Section 5.2).

6.4.4 Comparison of toxic effects of benzaldehyde and L-PAC on immobilized PDC

Toxic effects of benzaldehyde and L-P AC on activity of immobilized PDC were determined in the presence of various concentrations of both substances. After 5 h exposure to various concentrations of benzaldehyde and L-P AC in 40 mM KH2PO4 buffer (pH 7.0) containing 2.0 M of ethanol, 50 µM TPP and 0.5 mM MgSO4, residual activities were measured. As shown in Figure 6.12, inhibition or inactivation gradually increased with increasing concentrations of both substances. In this experiment, the Ki values for benzaldehyde and L-P AC were determined to be 160 mM and 240 mM, respectively. The determination of Ki was carried out at 4°C rather than 25 °C, for direct comparison with results for free enzyme. Chapter 6 202

100 o EtOH 90 • Acetone ~-- 0 0~ 80 o-i ·--·-·-. 0 -?fl. /·----- . ->. 70 ·--> ·-V 60 <- ~o u 50 0 • ~ -ns 40 :I -0 fll 30 ·-Q,I c=:: 20

10

0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Concentration (M)

Figure 6.11 Effect of organic solvents on stability of immobilized PDC activity after 10 days incubation at 4°C. The relative residual activities (%) were measured by reference to the initial activity of immobilized PDC. Chapter 6 203

110

100

90

80 -0~ ->. 70 -> ·- 60 ·-f,j -!U 50 -!U:, 'O (lj 40 ·-QI c=:: 30

20

10

0 'T""r--r"'r'jj""l"'"T"'"",....,..,..,....,..."T""T""'T""T'""T-r",...... ,....T""T"...,...... ,...... ,....,...__,~ 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 Concentration (mM)

Figure 6.12 Toxic effects of benzaldehyde and L-P AC on immobilized PDC. After 5 h exposure to various concentrations of benzaldehyde or L-P AC at 4°C, residual activities were measured. Chapter 6 204

In particular, the Ki for immobilized PDC with benzaldehyde even after 5 h longer incubation was found to be higher compared with 80 mM for free enzyme after 1 h incubation, indicating the protective effect of immobilization on the enzyme.

6.5 BIOTRANSFORMA TION PROPERTIES OF IMMOBILIZED PDC

6.5.1 Effect of pH on L-PAC formation with immobilized PDC

The optimum pH for L-P AC formation was determined by measuring initial reaction rates over the first 30 min. The reaction mixture consisted of 10 mL immobilized beads (14.2 unit/mL PDC activity), 10 mL citrate/ sodium phosphate buffer (pH adjusted in the range from 4.5 to 8.5) containing 70 mM benzaldehyde, 70 mM sodium pyruvate, 2.0 M ethanol, 50 µM TPP and 0.5 mM MgSO4. Biotransformations were performed at 4°C in a shake flask on a reciprocal shaker at 100 rpm. Aliquot samples were taken every 5 minutes and the concentration of L-P AC measured. Initial L­ p AC formation rates over the first 30 min were computed and relative reaction rates, compared to maximum rate, were plotted against pH. As demonstrated in Figure 6.13, the results showed that the highest initial rate for L-PAC formation was at pH 6.5. When compared with optimum pH of 7.0 for free enzyme (see Section 5.3), the relative initial rate at pH 7.0 was 88 % that at pH 6.5. The optimum pH with immobilized PDC was shifted slightly to more acidic conditions.

6.5.2 Effect of organic solvents on L-PAC formation

As described in more detail in Section 6.4.3, biotransformations were carried out with various concentrations of ethanol and acetone at 4°C. For a comparison, acetone with a similar hydrophobicity to ethanol (Polarity Chapter 6 205

110 100 -~ (0\ -u 90

Figure 6.13 Effect of pH on initial rate of L-PAC formation with immobilized PDC at 4°C. Rates are relative to maximum rate at pH 6.5. Chapter 6 206

Index, ethanol, 5.2; acetone, 5.4. Godfrey, 1972) was also examined. The reaction mixture contained 10 mL immobilized beads (14.0 unit/ml PDC) and 10 mL 40 mM KH2PO4 buffer (pH 6.5) containing 70 mM benzaldehyde, 70 mM sodium pyruvate, 50 µM TPP, 0.5 mM MgSO4 and indicated amounts of organic solvents. Initial rates for L-PAC formation were measured as described Section 6.5.1. As shown in Figure 6.14, initial rates of L-PAC formation were gradually increased with increasing concentrations of both organic solvents up to 4.0 M. The extent of increase with ethanol was considerably higher than with acetone. With 4.0 M ethanol it was 1.4 times higher, while with 4.0 M acetone, it was 1.25 times that in the absence of solvent.

With the immobilized PDC enzyme, a higher concentrations of ethanol was required to increase the L-PAC formation rate to a similar extent compared to the free enzyme (see Section 5.2). From the results in Figures 6.11 and 6.14, it was proposed that addition of up to 4.0 Methanol can be recommended to facilitate the immobilized PDC process. However, the addition of high concentrations of ethanol may impose purification difficulties for L-P AC production in commercial practice.

6.5.3 Effect of temperature on L-PAC formation

The temperature dependence of immobilized PDC for L-P AC formation was examined at 4°C, 10°C and 25°C. Biotransformations were conducted in flasks on a reciprocal shaker with 100 rpm for 20 h, The reaction mixture consisted of 10 ml of immobilized beads (14.0 unit/ml PDC) and 10 mL of reaction buffer containing 4.0 M ethanol, 70 mM benzaldehyde, 70 mM sodium pyruvate, 50 µM TPP and 0.5 mM MgS04 in 50 mL Erlenmeyer flasks. Initial kinetic parameters were determined over the first 30 min, and final L-PAC concentrations were estimated after 20 h incubation. Kinetic parameters at various temperatures are summarized in Table 6.2. Chapter 6 207

160------, o EtOH

-~0 • Acetone -C 0 ·­IU -s ....0 u C: I ...J ....0""" ~ . -t? ~ 40 - > ·--IU ~ 20 - -c:::

0 -+---,.---.,--,---"Tl-...,....--,,.---,--T,-~---r,--r--,r---..-----1 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Concentration (M)

Figure 6.14 Effect of organic solvents on initial rate of L-PAC formation with immobilized PDC enzyme at 4°C. Chapter 6 208

Table 6.2 Effect of temperature on biotransformation kinetics with immobilized PDC. Biotransformations were carried out in a 20 mL reaction mixture (pH 6.5) containing 70 mM benzaldehyde, 70 mM sodium pyruvate, 10 mL immobilized PDC (14.0 unit/mL), 4.0 M ethanol, 50 µM TPP, 0.5 mM MgSO4 and 40 mM KH2PO4 buffer (pH 6.5).

Temperature Parameters 4°c 10°c 25°C

dp L-PAC• 6.0 10.4 14.8 dt dp acetaldehyde 2.4 5.7 15.5 dt dp acetoin 2.0 2.8 5.4 dt L-PACmax" 41.6 36.4 28.1 Yp/s (mole%) 59.4 52.0 40.1 "'mM/h and"'"' mM

The results indicate that all initial rates at 25°C were higher than at lower temperatures. In particular, L-PAC formation rate at 25°C was 2.5 times higher than at 4°C, and 1.4 times of that at 10°C. Furthermore, the acetaldehyde formation rates was even higher at 25°C, being 6.4 times of that at 4°C and 2.7 times that at 10°C. This high rate of acetaldehyde formation resulted in a reduced final L-PAC formation, and the results agreed with previous findings with free enzyme. In this case, it is likely that the influence of temperature on the rate of diffusional mass transfer may be more significant compared to the free enzyme process.

Although the higher temperature gave rise to increased productivity for L-P AC, it was accompanied by a lower conversion yield. In practice biotransformation conditions may be varied depending on whether productivity or conversion yield has priority. In the present study, it was decided to continue the evaluation at 4°C due to the improved conversion yield. Chapter 6 209

6.5.4 Effect of PDC activity of immobilized beads on L-PAC formation

To examine the influence of PDC activity of immobilized beads on l­ PAC formation, various PDC activities per unit bead volume were employed for biotransformation with 70 mM benzaldehyde and 70 mM sodium pyruvate in the reaction buffer at 4°C. As summarized in Table 6.3. final product concentrations and ratios were significantly influenced by PDC activity.

Table 6.3 Final products of biotransformation with various PDC activities per unit volume of beads and 70 mM benzaldehyde and 70 mM sodium pyruvate in the reaction buffer at 4°C.

PDC activity (unit/mL) Product (mM) 3.8 8.0 10.2 14.2 18.0 L-PAC 11.6 29.0 36.2 36.0 32.1 Acetaldehyde 5.2 10.2 15.1 16.4 18.2 Acetoin 6.2 5.2 7.50 8.0 9.2

The results showed that while the highest l-PAC formation was achieved with 10.2 unit/ml PDC activity, the lowest production of l-PAC as well as acetaldehyde were found at 3.8 unit/ml PDC activity. Partial inactivation or inhibition by benzaldehyde appears to be more significant at lower activity. At 18.0 unit/ml PDC, l-PAC formation was lower than the maximum. It is evident that higher PDC activity resulted in higher acetaldehyde formation and might reduce l-PAC formation.

6.5.5 Effect of molar ratio of pyruvate to benzaldehyde on L-PAC formation with immobilized PDC

Various molar ratios of substrates were employed for l-PAC formation with immobilized PDC. Reactions were conducted with 20 ml reaction mixture in 50 ml Erlenmeyer flasks consisting of 10 ml of beads (13.5 unit/ml PDC) and indicated ratios of substrates in 10 ml reaction Chapter 6 210

buffer (pH 6.5) with reciprocal shaking at 100 rpm and 4°C. Final L-PAC and molar conversion yields based on initial benzaldehyde are summarized in Table 6.4.

Table 6.4 Effect of molar ratios of pyruvate to benzaldehyde on L-P AC formation and molar conversion yields by immobilized PDC.

Final concentration of L-PAC (g/L) and Molar conversion yield (%) Benzaldehyde Pyruvate molar ratio ~mM) 0.5 1.0 1.5 2.0 50 2.2 4.1 5.1 5.3"'a 29.4 54.7 68.0 70.6·b 100 4.2 8.9 10.3 10.8 28.0 59.3 68.7 72.0 150 7.5 13.7 17.0 18.1 33.4 60.8 75.5 80.4 200 11.3 18.2 20.2 19.2 37.7 60.7 67.3 64.0 250 13.3 21.0 24.6 21.9 35.5 56.0 65.6 58.4 300 15.0 25.8 27.1 25.5 33.4 57.3 60.2 56.7 *a L-P AC {g/L): unit of (g/L) selected for comparison with results from free enzyme process. *b Yp/s (L-PAC/benzaldehyde: mole/mole%)

The higher ratios of pyruvate to benzaldehyde resulted in higher L­ P AC formation until excessive pyruvate caused inhibition which was observed above 200 mM benzaldehyde with 2.0 molar ratio. For similar molar ratios of pyruvate to benzaldehyde within the range of 50-150 mM benzaldehyde, Yp/s increased with increasing benzaldehyde. Further increase in benzaldehyde was accompanied by inhibition. The highest conversion yield, 80.4 % (mole/mole) based on benzaldehyde, was achieved with 150 mM benzaldehyde and 2.0 molar ratio (i.e. 300 mM pyruvate) for which 18.1 g/L L-PAC was obtained. Although conversion yield was 60.2 %, the highest L-P AC formation, viz. 27.1 g/L was Chapter 6 211

accomplished with 300 mM benzaldehyde and 1.5 molar ratio of pyruvate to benzaldehyde (i.e. 450 mM pyruvate). When compared with free PDC reactions (see Table 5.4), the immobilized PDC process resulted in lower conversion efficiencies which did not exceed 85 % based on initial benzaldehyde. However, at the concentration of 300 mM benzaldehyde, the immobilized PDC could still carry out the biotransformation, whereas significant inhibition was evident for the free enzyme process.

6.5.6 Typical biotransformation kinetics with immobilized PDC

Biotransformation kinetics for L-PAC formation were evaluated with 50 mL reaction mixture consisting of 25 mL immobilized PDC enzyme (13.5 unit/mL), 150 mM benzaldehyde, 225 mM sodium pyruvate, 4.0 Methanol, 50 µM TPP, 0.5 mM Mg5O4 and 40 mM KH2PO4 buffer (pH 6.5) in a 250 mL Erlenmeyer flask loaded on a reciprocal shaker at 4°C and 100 rpm. As shown Figure 6.15, L-PAC formation increased rapidly in the first 1 h, and then gradually declined as products, L-P AC, acetaldehyde and acetoin accumulated. After 16 h incubation, 17.1 g/L L-PAC was produced and the biotransformation was complete even though 2.8 g/L benzaldehyde and 1.2 g/L pyruvate still remained in the reaction mixture. Final molar conversion yield of L-PAC was 76 % (mole/mole) based on initial benzaldehyde. An estimated mass balance for pyruvate accounted for 114.0 mM pyruvate incorporated into L-PAC formation, 42.2 mM pyruvate converted into acetaldehyde, 52.3 mM into acetoin and 13.5 mM pyruvate was still left in reaction mixture. Only 5 mM pyruvate (less than 2.3 %) was not accounted for, and some evaporation losses might have occurred after conversion to acetaldehyde. When compared with free enzyme processes (see Section 5.8), the immobilized PDC processes have shown lower productivities and lower final L-P AC concentration under the same substrate concentrations although capable of tolerating higher benzaldehyde concentrations. This probably resulted from diffusional mass transfer limitations imposed by the immobilizing matrix. Chapter 6 212

24.0 o Benzaldehyde 22.0 • L-PAC Acetaldehyde 20.0 • • Acetoin 18.0 D Pynavate -..J -co 16.0 a ~ -~----·-·---- -C: 0 14.0 ~ /.,. ·-...(U ~ / ..."' 12.0 . C: ~ .,,;e QJ y 10.0 ~/ C u0 8.0 .?~ 6.0 •/ ~O....o . •' ,o-..._o__ 4.0 / D"-o 0~ • --- -o-o 2.0 ..._--.--•-•~-·-·-·-·-·-·_. -• i: s--Ji 6=1fl-=c o.o.-='-r----.---...-...... -...... -..---.-----.----.---.,...... --.-.--..--..--..--....-~ 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Time (h)

Figure 6.15 Biotransformation kinetics with immobilized PDC at 4°C. Reaction mixture consisted of 50 mL 40 mM KH2P04 buffer (pH 6.5) containing 150 mM benzaldehyde, 225 mM sodium pyruvate, 4.0 M ethanol, 50 µM TPP, 0.5 mM MgS04 and 25 mL immobilized PDC beads (13.5 unit/ml). Chapter 6 213

6.6 EVALUATION OF CONTINUOUS PROCESS WITH IMMOBILIZED PDC

6.6.1 Configuration of continuous immobilized PDC process

The selection of a suitable reactor can be approached according to the following considerations. For the case of substrate inhibited enzyme systems, CSTR performance is superior to that of a packed column (Kennedy and Cabral, 1983) due to minimizing the effect of toxicity of substrate, with the enzyme in the CSTR always operating at a substrate concentration equal to that in the product stream. In a packed bed reactor (PBR), however, the substrate concentration varies along the length of the reactor, thus enzyme is inactivated to a different extent in various sections of the reactor. From the previous results that products as well as substrates caused considerable inhibition of the L-PAC formation (see Section 4.7 and Section 6.4.4), it can be deduced that the adverse effects of both substrates and products in PBR may be paralleled by negative effects due to products in the CSTR. Additionally, substrate concentration has an enhancement effect with respect to rates and final conversions at every point in a PBR, while it is minimal at every point in a CSTR. The average reaction rate in a PBR is higher than it would be in a CSTR. Thus, for a continuous process for L­ P AC formation, it is likely that plug flow characteristics would be an advantage over CSTR operation. A continuous process with immobilized PDC was designed to consist of double jacket glass column reactor (30 mm ID x 300 mm L), substrate reservoir, syringe pump for benzaldehyde feeding, mixing reservoir with magnetic stirrer, peristaltic pump, fraction collector and cooling unit (Figures 6.16 and 6.17). 120 mL immobilized PDC beads (14.0 unit/mL) were packed into a glass column, and its operation temperature was maintained with circulation of constant temperature water (4°C) from a cooling unit. The substrate reservoir contained sodium pyruvate solution which was varied in concentration to give different ratios to benzaldehyde. Chapter 6 214

Substrate Reservoir Packed-Bed Reactor (sodium pyruvate)

Mixing Ves.sel

Cooling water Cooling water 11----r11c.w - 11-----,i,-- --

Magnetic Stirrer

Substrate Feeding Pump

Fraction Collector ITJ()

Benzaldehyde Feeding Pump

Figure 6.16 Schematic diagram of the continuous process with immobilized PDC. Chapter 6 215

Figure 6.17 Photograph of the continuous process with a packed bed reactor. The inset photograph shows a glass column packed with immobilized POC beads. Chapter 6 216

Sodium pyruvate solution was well mixed in a m1xmg reservoir with benzaldehyde fed by a syringe pump at various flow rates. The substrates were introduced into the bottom of the PBR. Overall substrate feeding rates were controlled by a peristaltic pump with variable set point. When the need arose to change substrate formulations, the concentration of sodium pyruvate and feed rate of benzaldehyde were adjusted.

6.6.2 Effect of space time on acetaldehyde formation of PBR

To determine the operating characteristics of the PBR at 4°C, various concentrations of sodium pyruvate in 40 mM KH2P04 buffer (pH 6.5) containing 4.0 M ethanol, 50 µM TPP and 0.5 mM MgS04 were fed into the column reactor at various flow rates. Acetaldehyde concentration was plotted as a function of space time ('t=VR/F: volume of reactor/flow rate through the reactor). The results shown in Figure 6.18 demonstrate that the catalytic activity of the PBR was dependent on the flow rate. Increasing space time resulted in increasing product formation, while productivities were decreased. Furthermore, higher substrate concentration gave rise to higher product formation, but was accompanied by lower fractional conversion yields (mole/mole). For instance, at 8 h space time, molar conversion yields for 25 mM and 100 mM pyruvate were 75 % and 45 % respectively, with the reaction with 25 mM sodium pyruvate completed in 8 h. The lower molar yields could be due to the inhibitory effect of increased acetaldehyde corresponding to higher substrate concentrations.

6.6.3 Effect of benzaldehyde concentration on L-PAC formation

To determine optimum feeding rates for benzaldehyde, several concentrations of benzaldehyde, viz. 25 mM, 50 mM and 75 mM with equal molar sodium pyruvate, were fed into the PBR at various flow rates. PBR operation was sustained ex~ept when 75 mM benzaldehyde was fed into the reactor, the continuous process did not reach steady state Chapter 6 217

(a) 60.,------Cone. of Pyruvate 55 • o 25mM 50 • SOmM a 45 a 75mM -~ • lOOmM E 40 -QJ 35 "'O >. .c: 30 QJ "'O 25 -...~ QJ V 20 < 15

10 5 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Space time (h) (b) 1.0 ------. Cone. of pyruvate 0.9 - o 25mM

QJ • 50mM - 0.8 - -0 a 75mM E QJ ------0 • lOOmM - 0.7 - -0 E 0.6 - /::----: -"'O QJ 0.5 - -">. C: 0.4- ;;~~- 0 C: 0 1~~ u 0.2 - ft 0.1 -

0.0 I I I I I I I I I 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Space time (h)

Figure 6.18 Effect of space time (-c) on acetaldehyde formation in a packed-bed reactor at 4°C and various concentrations of sodium pyruvate: (a) acetaldehyde formation; (b) molar conversion yields. Chapter 6 218

presumably due to toxicity of benzaldehyde. After 3-4 generations of space time, L-PAC in outlet was measured and plotted against space time. The results shown in Figure 6.19 demonstrate that with 50 mM benzaldehyde and 50 mM pyruvate, the L­ p AC concentration was increased with increasing space time. At 20 h space time, L-P AC formation reached a peak level of 3.9 g/L with 52 % molar conversion yield. Whereas with 25 mM benzaldehyde and 25 mM pyruvate, L-PAC formation reached 0.9 g/L with 34 % molar conversion yield in 10 h space time. Further increase in space time did not enhance L-PAC formation. It is likely that the biotransformation was completed within this reaction time. Comparison of molar conversion yields indicated that 50 mM benzaldehyde gave higher yields for L-P AC formation compared with 25 mM benzaldehyde, and are in good agreement with results for the batch process where the higher concentrations of benzaldehyde more strongly inhibited acetaldehyde formation. Higher productivities of L-PAC formation with both concentrations of benzaldehyde were achieved in shorter space times, which is similar to the results for acetaldehyde formation (see Section 6.6.2). For example, productivities with 25 mM and 50 mM benzaldehyde at 1 h space time were 0.41 and 0.9 g/L/h respectively, while at 20 h space time they were 0.07 and 0.2 g/L/h, respectively.

6.6.4 Effect of pyruvate concentration on L-PAC formation

To examine the effect of changing the molar ratio of pyruvate to benzaldehyde, continuous biotransformations were conducted with 50 mM benzaldehyde and 1.0, 1.5 and 2.0 molar ratios (i.e. 50 mM, 75 mM and 100 mM pyruvate, respectively) at various flow rates. L-PAC formation, productivities and molar conversion yields based on benzaldehyde are presented in Figure 6.20. Steady state biotransformation kinetics with 1.5, 2.0 ratios are presented in Figure 6.21 and the data analysis is summarized in Table 6.5. These results showed similar trends to those for a batch process, with increasing molar ratios resulting in increased L-PAC formation as well as by-products, acetaldehyde and acetoin. Chapter 6 219

(a)

5.o------Cone. of benzaldehyde 45 o 25mM 4.0 ----· • SOmM 3.5

-..J 3.0 bO -u 2.5 /"

0.0 ------....--....--...--...--....-....-.....------__. 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 Space time (h) (b) 1.0....------0 25mM 0.9 • • SOmM c Yp/s (25 mM) 0.8 • Yp/s (50 mM) -..c 0.7 -..J -bO 0.6 .\ ·-;;.. 0.5 .... • -· ·-u ::s 0.4 "Cl ' .--- 0 i.. Col 0.3 \,.~:~o---0 0.2 ~11'-...SL • 0.1 .~ ------0 -----0 o.o-+------...... -...... -...--...------' 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 Space time (h)

Figure 6.19 Effect of benzaldehyde on L-P AC formation in a packed-bed reactor at 4°C. Substrate feed contained 25 or 50 mM benzaldehyde, and equal molar sodium pyruvate at various space times: (a) L-PAC formation; (b) productivities and molar conversion yields. Chapter 6 220

(a) 6.0 Cone. of pyruvate 5.5 • 0 SOmM 5.0 • 75mM 4.5 • • lOOmM 4.0 0 ...J -bb 3.5 -u 3.0 ~: < ~ I 2.5 ...J t1/ 2.0 f/o 1.5 ~/·/il O

1.0 •/'0 0.5 0.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 Space time (h) (b) 1.6 1.0 Cone. of pyruvate 0.9 • SOmM 1.4 • 70mM 0.8 • lOOmM 1.2 • 0 SOmM -..c:: 0 -- 0.7 -...J -QJ C 75mM -ClO 1.0 =-\ a- o.6 - 6 0 lOOmM ->-. .... o- -QJ > 0.8 ~~ ~- 0.5 0 .... ·-V e ~ \~y.½: 0.4 - "0 0.6 Cll 0 i.. 0.3 -o.. ~ )~i~ 0.4 1/tt~a~. >- i/0 ----====: 0.2 0 0 C/ I 0.2 0 • 0.1

0.0 0.0 0 2 4 6 8 10 12 14 16 18 20 22 Space time

Figure 6.20 Effect of varying pyruvate concentrations with 50 mM benzaldehyde on L-P AC formation in a packed-bed reactor at various space times: (a) L-PAC formation; (b) productivities and molar conversion yields. Chapter 6 221

(a) 10 0 Benzaldehyde 9 • L-PAC • Acetaldehyde 8 A Acetoin D pyruvate ,..J 7 -bb -C 6 ·-.....0 ~ .....1-< 5 C • QJ V 4 C 0 u 3

0 2 / D~ 1 Iii=- 0::::::::::: • .=:::a=• a 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 Space time (h) (b) 10.0 0 Benzaldehyde 9.0 • L-PAC • Acetaldehyde 8.0 A Acetoin a pyruvate -,..J bb 7.0 a -c:: 0 6.0 ·-..... i,: .....i.. 5.0 \a • c:: QJ V c:: 4.0 u0 3.0 '\~ 2.0 /~o / • 1.0 • I .,...,-:.~ a 0.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 Space time (h)

Figure 6.21 Steady state kinetics of a packed-bed reactor at 4°C with 50 mM benzaldehyde and (a) 75 rnM; (b) 100 rnM sodium pyruvate. Chapter 6 222

Table 6.5 Comparison of L-P AC formation and productivities with 50 mM benzaldehyde, 1.5 and 2.0 molar ratios of pyruvate to benzaldehyde for continuous process.

L-PAC (g/L) Productivity (g/L/h) Flow rate Space time Molar ratio Molar ratio (mL/h) (h) 1.5 2.0 1.5 2.0 120 1.0 1.1 1.4 1.1 1.4 60 2.0 1.6 1.9 0.8 0.95 30 4.0 2.2 2.6 0.55 0.65 15 8.0 3.5 4.3 0.44 0.54 6 20 4.7 5.3 0.23 0.27

6.6.5 Long term continuous biotransformation of benzaldehyde to L-PAC

An evaluation of a continuous process for long term operation with immobilized PDC for L-PAC formation was conducted with the PBR. The PBR reactor was set up as described previously (see Section 6.8.1). Operating conditions were the following:

Total loading (beads) 120mL PDC activity 13.8 unit/mL Inlet substrates 50 mM benzaldehyde, mixture 100 mM sodium pyruvate, S0µM TPP, 0.5 mM MgSO4, 4.0 M ethanol, 40mM KH2PO4 pH 6.5 Overall flow rate 15 mL/h (space time t= 8 h) Temperature 4°c

Concentrations of L-P AC in the outlet were determined and plotted as a function of operation time as shown in Figure 6.22. Chapter 6 223

6.0

5.5 5.0

4.5 4.0 -.,.J -~ 3.5 u 3.0

1.0 0.5

0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Time (day)

Figure 6.22 Continuous L-PAC formation with 50 mM benzaldehyde and 100 mM sodium pyruvate by immobilized PDC in a packed-bed reactor at 40c. Chapter 6 224

The results indicate that continuous production of L-PAC was stably maintained for 5 days and then gradually decreased. When supplied with fresh substrate (indicated arrow), L-PAC production slightly increased. It is possible that partial decomposition of chemicals may have occurred in the substrate reservoir even though it was maintained at 4°C. Over 5 days, the average productivity was 0.53 g/L/h with 4.5 g/L L-PAC, and then productivity gradually decreased with further operation. Eventually, it maintained only 55 % of the initial productivity after 27 days operation.

6.7 DISCUSSION AND CONCLUSIONS

An immobilized PDC process was examined with respect to its potential application for a biotransformation process to produce L-P AC. Entrapment of PDC enzyme into spherical polyacrylamide beads was successfully performed with 12.5 % activity retained compared to original enzyme solution. Furthermore, addition of the bifunctional reagent, glutaraldehyde, provided an enhancement of PDC activity as well as increasing storage stability at 4°C. Under these conditions, activity yields in the range of 14.5-15.0 % were achieved, and the empirical half life of immobilized PDC was calculated to be 32 days in 40 mM KH2PO4 buffer at 4oe, Several different characteristics of the immobilized PDC compared to free enzyme were found. In particular, immobilized PDC has a higher Km for pyruvate and higher resistance to the toxic effects of benzaldehyde and L-PAC due to a concentration gradient of these substances in the supporting matrix. The concentrations of organic solvent (e.g. ethanol) to accelerate biotransformation rate and to enhance the stability were higher than for free enzyme. For a comparison, acetone as a similar hydrophilic solvent was evaluated as well as ethanol. It was clearly shown that addition of either acetone or ethanol enhanced stability and accelerated L-PAC formation rates. Both solvents showed similar enhancement although the positive effect with ethanol was slightly higher than with acetone. Chapter 6 225

Immobilization resulted also in a change of optimum pH for biotransformation to slightly more acidic conditions possibly due to changes of microenvironment at the reaction sites. The influence of temperature was more significant for immobilized PDC and appeared to affect both reaction mechanisms and reaction rates. With immobilized PDC, L-PAC formation was significantly greater at higher temperatures compared with free enzyme under similar conditions (See section 4.5). For example with immobilized PDC at 25°C, L-PAC formation rate was significantly higher than at 4 and 10°C, although still with lower conversion yields compared to lower temperatures. In this case, the temperature dependence of the mass transfer rate is likely to be a considerable influencing factor. The effect of enzyme density on biotransformation was evaluated, with high PDC activity resulting in high L-P AC formation accompanied by relatively higher acetaldehyde formation. However, low PDC activity resulted in incomplete biotransformation due to toxic effects of residual benzaldehyde. Under these conditions the reaction was accompanied by lower acetaldehyde formation. The results that immobilized PDC has a higher resistance to toxic substrates could allow a possible extension of the range of benzaldehyde concentrations compared with free enzyme. While with 300 mM benzaldehyde and 450 mM pyruvate at which the free enzyme was significantly inhibited and/or inactivated, the highest L-PAC 27.1 g/L was accomplished with immobilized PDC. This was close to the value accomplished with free enzyme, although at lower benzaldehyde levels. In a comparison with free enzyme under similar conditions, biotransformations with immobilized PDC showed that the biotransformation took longer time to complete reaction in batch, and resulted in final L-P AC formation which was considerably lower than for free enzyme, with lower conversion efficiency. In continuous process, the potential of the packed bed reactor was evaluated by increasing flow rate and substrate (sodium pyruvate) concentrations. Increased space time (i.e. decreased flow rate) resulted in higher conversion yields as expected. For continuous biotransformation, L-PAC production was quite stable with 25 mM and 50 mM benzaldehyde for 12-15 generations (4-5 days). When 75 mM benzaldehyde was introduced into the column reactor, L-PAC formation was continuously decreased and operational stability of the Chapter 6 226

immobilized PDC appeared to be affected by irreversible denaturation of enzyme and leakage of enzyme from immobilized beads. Although higher ratios of pyruvate to benzaldehyde promoted L-P AC formation, there was not a proportional increase in conversion yields based on benzaldehyde. The main advantages of a continuous process based on immobilized PDC are that a relatively high productivity per enzyme unit can be achieved with long term operation compared with batch process. However, the low L-PAC concentration on a long term basis is a significant disadvantage. Chapter 7 227

CHAPTER 7

GENERAL DISCUSSION AND CONCLUSIONS

The potential use of biocatalysts is increasing particularly in the area of organic chemical synthesis. One of the most important areas is that of biotransformation involving asymmetric catalysis, notably in the pharmaceutical industry where there is a growing need to produce only one enantiomeric form of a compound. For the case of L-ephedrine, there is increasing worldwide demand for its use in the medicinal treatment of allergic conditions, hypotension, chronic asthma, management of obesity etc. In the biotransformation process for L-ephedrine, a stereoselective carbon-carbon condensation reaction is catalysed by pyruvate decarboxylase (PDC) to produce L-phenylacetykarbinol (L-P AC) as the key intermediate for L-ephedrine. In the present study, biotransformation processes for L-PAC formation have been evaluated using immobilized Candida utilis, purified pyruvate decarboxylase from C. utilis and immobilized PDC. The study complements previous research in our laboratory on fed-batch and continuous processes with free cells for L-PAC production. The conclusions are summarized as follows:

7.1 L-PAC FORMATION WITH IMMOBILIZED C. UTILIS

Determination of PDC levels, the enzyme responsible for L-P AC formation, indicated that the PDC activity in C. utilis was significantly affected by the available oxygen in the culture. When fermentative conditions were induced by reducing aeration rate or agitation speed, PDC activity was immediately enhanced. Although the PDC activity of C. utilis was less sensitive to glucose concentration, pulse feeding of glucose (3 sequences) under fermentative conditions enhanced PDC activity to 0.59 unit/mg protein compared with 0.31 unit/mg protein in the absence of additional glucose feeding. When PDC activity was at peak levels, Chapter 7 228

immobilization of C. utilis was effected through entrapping cells in 3 % calcium alginate. A comparison of L-PAC formation between free and immobilized cells indicated that immobilized cells had resistance to higher concentrations of benzaldehyde. For example, at initial 70 mM benzaldehyde, L-P AC formation with immobilized cells reached 37.5 mM (viz. 5.6 g/L), while L-PAC formation with free cells was significantly inhibited at 20.1 mM (viz. 3.0 g/L) in flasks. Furthermore, L-PAC formation with free cells was almost totally inhibited at initial 100 mM benzaldehyde (i.e. 1.6 mM L-PAC produced), while immobilized cells continued to produce L-PAC up to 23.2 mM (viz. 3.5 g/L) at the same concentration of benzaldehyde. However, immobilized cells exhibited a significant drawback: generally higher levels of by-product, benzyl alcohol, were formed compared with free cells, presumably due to more reductive conditions in the calcium alginate gel matrix imposed by diffusional oxygen limitations. Other results indicated that benzyl alcohol formation was affected by initial benzaldehyde concentrations in flask cultures. Under these conditions, benzyl alcohol formation decreased significantly with increasing initial benzaldehyde concentrations. These results indicated the importance of controlling benzaldehyde feeding levels. Studies on various levels of benzaldehyde in a 2 L fermenter demonstrated that at a controlled level of 0.8 g/L benzaldehyde, the proportion of benzyl alcohol formation was maximal, while at 4 g/L benzaldehyde, benzyl alcohol formation was minimized. Control of benzaldehyde at 2 g/L resulted in the highest L-P AC formation up to 10.8 g/L. As a comparison, Mahmoud et al. (1990a) reported that with pulse feeding of benzaldehyde (1.2 g/L x 4 times at 30 min intervals in the early stages), 5.5 g/L L-PAC was achieved with immobilized S. cerevisiae ATCC 834, while further increase in benzaldehyde (2 g/L x 4 times at 30 min interval) drastically decreased L-PAC formation. From a comparison of the present results with literature values, it was confirmed that continuous control of benzaldehyde levels in the reaction mixture to about 2 g/L gave rise to a significant improvement in L­ p AC formation. During biotransformation, the RQ was also found to be important in L-PAC formation. In the range of RQ=5-7, increasing L-PAC formation occurred. With pulse feeding of glucose, maintaining 2 g/L benzaldehyde Chapter 7 229

level and with the optimum range of RQ values, L-P AC of 15.2 g/L was achieved after 22 h biotransformation.

In a continuous process with immobilized cells, although the higher benzaldehyde feed rate resulted in a higher productivity of L-PAC in the short-term, the stability of the process was drastically decreased. With a feed rate 1.0 g/L/h (1.5 mL benzaldehyde into 1.5 L working volume), a L-PAC productivity of 0.6 g/L/h (1.0 mL benzaldehyde into 1.5 L working volume) was maintained for 50 h, while with 0.67 g/L/h benzaldehyde, 0.33 g/L/h of L-P AC was sustained for more than 110 h. A surprising result was that higher cell densities gave rise to a higher ratio of benzyl alcohol to L-PAC, presumably due to higher oxygen requirements at high cell density resulting in more reductive conditions under limited oxygen in the immobilizing matrix. Overall for the immobilized cell system, it can be concluded that final L-P AC formation was significantly influenced by a pulse feeding of glucose, PDC activity, benzaldehyde level, the RQ value for the biotransformation phase and the cell density.

7.2 L-PAC FORMATION WITH PURIFIED PDC

To eliminate production of the by-product, benzyl alcohol, in the biotransformation process, purified pyruvate decarboxylase was employed in producing L-P AC. Following a comparison of various PDC sources such as C. utilis, S. cerevisiae and wheat germ, C. utilis was selected as it produced the highest total PDC activity per unit culture volume (viz. 5.9 unit/mL). The PDC was purified (5.6 fold purification) through homogenization, (NH4)2S04 fractionation and gel-filtration. A comparison of PDC stability under various conditions indicated that PDC stability was significantly influenced by buffer species, pH, temperature and TPP concentration. A 40 mM KH2P04 buffer (pH 6.0) also containing 30 µM TPP and 0.5 mM MgS04, was selected as a suitable buffer. Chapter 7 230

The application of purified PDC in the biotransformation process resulted in formation of other by-products viz. acetaldehyde and acetoin, instead of benzyl alcohol. Under normal conditions (25°C and pH 6.0), PDC from both C. utilis and S. cerevisiae predominantly formed acetaldehyde rather than L-PAC. However, an investigation of the temperature dependence of the reactions catalysed by PDC established that while acetaldehyde formation was promoted by increasing the temperature, at low temperatures (e.g. 4°C) L-PAC formation was predominant. It is possible that a longer half life of the PDC-acetaldehyde complex ('active acetaldehyde') at the lower temperature may promote the condensation reaction with benzaldehyde for L-P AC formation. Further investigation confirmed that free acetaldehyde formed in the reaction mixture was a strong inhibitor of L-P AC formation, as well as being associated with reduced substrate conversion efficiency. Purified PDC had a higher resistance to the toxic effects of benzaldehyde and L-P AC in the short term compared with immobilized cells. However, after 1 h exposure to 200 mM benzaldehyde the PDC was completely inactivated. Generally, the toxic effect of benzaldehyde on PDC was 2.5 times higher than that of L-P AC. It is suggested that the higher initial conversion rates at higher benzaldehyde levels achieved the higher conversion yield through converting highly toxic benzaldehyde to less toxic L-P AC before complete inactivation of PDC occurred.

Various strategies were followed to increase the conversion efficiency of L-P AC formation. Addition of ethanol to the reaction mixture resulted in improvement in biotransformation with enhanced reaction rates, better conversion yields and improved enzyme stability. With 2.0 M ethanol, the initial reaction rates for L-PAC formation were increased up to 1.4 times, conversion yields increased up to 98 % (mole/mole) based on initial benzaldehyde, and 70-75 % initial PDC activity was maintained after 140 h incubation at 4°C compared with 40 % activity in the absence of ethanol. The finding that the optimal pH for L-PAC formation was different from that for acetaldehyde brought a significant improvement in L-P AC formation. At neutral or slightly alkaline pH, acetaldehyde formation was decreased while L-P AC formation rate was increased up to 3 times compared to that at pH 6.0. Chapter 7 231

Increasing PDC activity per unit volume did not correlate directly with increasing L-PAC formation. In fact, higher PDC activity was accompanied by more acetaldehyde formation, while low PDC activity resulted in incomplete biotransformation at high benzaldehyde concentrations. PDC activities in the range 7-10 unit/mL were optimal for efficient biotransformation. As discussed in Chapter 4, due to the complexity of the PDC reaction, the concentration of both substrates, benzaldehyde and pyruvate, significantly influenced the reaction rates as well as final L-PAC values. Usually the requirement for pyruvate was higher than for benzaldehyde due to partial conversion of pyruvate to acetaldehyde and acetoin. Although higher molar ratios of pyruvate to benzaldehyde resulted in higher L-P AC formation, the effect of increasing the molar ratio became less significant with increasing benzaldehyde concentration. The reason seems to be that at high concentrations of benzaldehyde, there is preferential condensation with active acetaldehyde rather than release of free acetaldehyde. This resulted in increased L-P AC formation accompanied by reduced acetaldehyde. In batch biotransformations with purified PDC, L-P AC formation significantly increased in the first few hours. Further L-P AC formation occurred more slowly as a result of substrate depletion and product accumulation. The highest L-P AC formation (viz. 28.6 g/L) was achieved with purified PDC (7.0 unit/mL), initial 200 mM benzaldehyde and 400 mM pyruvate, with 95.3% molar conversion yield based on benzaldehyde. In previous investigations (Bringer-Meyer and Sahm, 1988), a comparison of L-PAC formation by PDC from Z. mobilis and S. carlsbergensis has shown that L-P AC formation with the yeast enzyme was significantly higher than with PDC from Z. mobilis. L-P AC formation rate reached a peak value with 8 unit/mL of PDC from S carlsbergensis at 46.2 mM benzaldehyde (viz. 4.9 g/L), while further increase of benzaldehyde resulted in drastically decreased L-P AC values. In the present study, it was clearly shown that biotransformation efficiency for L-PAC formation was significantly improved by optimizing physicochemical factors, viz. temperature, pH, addition of organic solvent (e.g. ethanol) as well as PDC activity and substrate concentrations. Chapter 7 232

7.3 L-PAC FORMATION WITH IMMOBILIZED PDC

Entrapment of PDC into spherical polyacrylamide beads has been successfully performed with 12.5 % activity yield based on activity of original enzyme in solution. Investigation of characteristics and kinetic constants of immobilized PDC showed different values from those of the free enzyme. Presumably, the reduced rates of diffusion of substrate to entrapped enzyme and product released from the enzyme can bring about considerable changes in the kinetic behaviour of immobilized enzyme. In particular as illustrated in Table 7.1, immobilized PDC has a higher Km for pyruvate when compared with free enzyme, and more than 2 times higher resistance to toxicity of benzaldehyde. The optimum pH for L-PAC production was also slightly shifted to acidic conditions for free cells of C. utilis.

Table 7.1 Comparison of characteristics of PDC from various sources

PDC source Km Ks Ki pH Reference(s) (Prr) (BZ) (BZ) (L-PAC) Z. mobilis 0.6mM 125mM NA* 6.0 Bringer-Meyer et al. (1986); Bringer-Meyer & Sahm (1988) S. carlsbergensis l.2mM S0mM NA 6.0 Lehmann et al. (1973); Bringer-Meyer & Sahm (1988) C. utilis 2.4mM 42mM 80mM 7.0 This work Immobilized PDC 3.2mM NA 161mM 6.5 This work * Not available

Similar to the results with free PDC, the addition of organic solvents such as ethanol and acetone, enhanced the reaction rates of L-PAC formation as well as increasing the stability of the immobilized PDC. Ethanol had a greater influence on reaction rates and stability. With immobilized PDC, formation of L-P AC, acetaldehyde and acetoin were all influenced significantly by temperature. The catalytic activity and Chapter 7 233

diffusional mass transfer in immobilized beads were accelerated with increased temperature as expected. Although higher conversion yields were achieved at lower temperatures, L-PAC formation rates were lower. For example at 4°C, the L-PAC formation rate was 2.5 times lower than at 25°C. Thus, selection of the optimum temperature for L-P AC formation depends on whether the productivity or the conversion yield has a higher priority. In the batch process, L-PAC formation was generally lower than for free enzyme with the same concentrations of substrates. However, biotransformations could be carried out at higher benzaldehyde concentrations (e.g. at 300 mM benzaldehyde) with immobilized cells, and the L-PAC concentration reached a peak of 27.1 g/L, at which L-PAC formation with free enzyme was almost completely inhibited. In a continuous process with a packed bed reactor containing immobilized PDC, the molar conversion yield with a feed of 50 mM benzaldehyde was higher than with 25 mM benzaldehyde with the same molar ratio of pyruvate to benzaldehyde. L-P AC productivity increased with increasing pyruvate concentration and flow rates, while the steady state concentration of L-P AC decreased. Biotransformation with 50 mM benzaldehyde (viz. 5.3 g/L) and 2.0 times molar ratio of pyruvate (viz. 8.8 g/L) to benzaldehyde achieved 30 mM L-PAC (4.5 g/L) at 0.51 g/L/h productivity and 60 % molar conversion yield based on inlet benzaldehyde. However, continuous contact with benzaldehyde, leakage of PDC and enzyme decay caused a steady reduction of catalytic activity in the packed bed reactor. The half life of the immobilized PDC under these conditions was estimated to be 32 days.

7.4 COMPARISON WITH OTHER BIOTRANSFORMA TION PROCESSES FOR L-PAC FORMATION

A comparison of L-P AC formation and productivities with the results of other authors is summarized in Table 7.2, As can be seen from the results, higher yields and productivities have been achieved in the present study. This is particularly true in batch using purified free and immobilized PDC from C. utilis. However, the costs of producing the purified PDC in Chapter 7 234

larger quantities as well as substrate cost (pyruvate) may preclude the development of large scale enzyme based processes for L-PAC production.

Table 7.2 Comparison of kinetic parameters from various processes for L­ PAC formation

(1) Batch and fed-batch processes

*Molar Process L-PAC Reaction Productivity conversion Reference (g/L) time (h) (g/L/h) yield(%) Free cells 12.4 17 0.73 57.5 Culic et al. (1984) 22.2 12 1.85 60.4 Wang (1993) 9.9 3 3.3 30.0 Seely et al. (1989b) Immobilized cells 10 24 Mahmoud et al. 0.4 61.1 (1990a) 15.2 22 0.7 58.0 This work Purified PDC 28.6 8 3.6 95.6 This work Immobilized PDC 27.1 12 2.3 60.2 This work *Molar conversion yield (% theoretical) is based on benzaldehyde added.

(2) Continuous processes

Average Molar Process L-PAC productivity conversion Half life Reference (g/L) (g/L/h) yield(%) Immobilized cells 3.75 1.3-1.8 22.7 11 days Seely et al. (1989a) 4.0 0.6 37.9 120 h This work 3 stages 11.7 0.35 70 NA• Wijono (199I: continuous culture 10.6 0.46 52 more than Wang (1993) 300h Immobilized PDC 4.5 0.56 60.2 32 dai:s This work ,. Not available Chapter 7 235

7.5 POSSIBLE FUTURE STUDIES

In the present study, a detailed evaluation has been made of the use of immobilized C. utilis, as well as purified and immobilized pyruvate decarboxylase for L-PAC production. Possible future studies could be directed at:

(1) development of strains of C. utilis with higher resistance to benzaldehyde and increased PDC levels; (2) further investigation of the biotransformation process in aqueous/organic or non-aqueous phase to increase solubility of substrate/ product; (3) optimal design of the continuous process, including development of a multi-stage packed bed reactor to enhance the conversion yields and facilitate replacement of degenerated immobilized PDC; (4) mathematical modelling of the kinetics of L-PAC formation and use of the model as a basis for optimal process control and benzaldehyde feeding.

As indicated in Table 7.2, L-P AC concentration of 22.2 g/L in a fed­ batch process, and 28.6 g/L with purified PDC have now been achieved. This compares with maximum L-P AC concentrations in the previous literature of 12.4 g/L (Culic et al, 1984). The improvements indicated above should lead to ever greater concentrations and productivity of L-P AC. References 236

References

Agarwal, S. C., Basu, S. K., Vora, V. C., Mason, J. R. and Pirt, S. J. {1987) Studies on the production of 1-acetylphenylcarbinol by yeast employing benzaldehyde as precursor. Biotech. Bioeng., 29: 783-785. Alberti, B. N. and Klibanov, A. M. {1985) The mechanism of irreversible enzyme inactivation at 100°C. Science, 228: 1280-1284. Aldercreutz, P. and Mattiasson, B. (1987) Aspects of biocatalyst stability in organic solvents. Biocatalysis, 1 (2): 99-108. Alexander, M. A. and Jeffries, T. W. (1990) Respiratory efficiency and metabolite partitioning as regulatory phenomena in yeasts. Enz. Microb. Technol., 12: 2-19. American Pharmaceutical Association {1986) In 'Handbook of Non-Prescription Drugs'. 8th edn. American Pharmaceutical Association, The National Professional Society of Pharmacists. Washington D. C. p. 142. Andersson, E. and Hahn-Hagerdal B. (1990) Bioconversion in aqueous two-phase systems. Enz. Microb. Technol., 12: 242-254. Andersson, E., Mattiasson, B. and Hahn-Hagerdel, B. (1984) Enzymatic conversion in aqueous two phase systems: decarboxylation of benzyl penicillin to amino penicillanic acid with penicillin acylase. Enz. Microb. Technol., 6: 301-306. Antonini, E., Carrea, G. and Cremonesi, P. (1981) Enzyme catalyzed reactions in water-organic solvent two-phase systems. Enz. Microb. Technol., 3: 291-296. Astrup, A., Bulow, J., Christensen, N. J. and Madson, J. (1984) Ephedrine-induced thermogenesis in man: No role for interscapular brown adipose tissue. Clinical Sci., 55: 179-186. References 237

Astrup, A., Breum, L., Toubro, S., Hein, P. and Quaade, F. (1992a) The effect and safety of an ephedrine-caffeine compound compared to ephedrine, caffeine and placebo in obese subjects on an energy restricted diet: A double blind trial. Intl. J. Obesity, 16 (4): 269-277. Astrup, A., Buemann, B., Christensen, N. J., Toubro, S., Thoebek, G., Victor, 0. J. and Quaade, F. (1992b) The effect of ephedrine/ caffein mixture in energy expenditure and body composition in obese women. Metabolism Clinic. & Experiment., 41 (7): 686-688. Atrat, P., Hiiller, E. and Horhold, C. (1980) Steroid-transformationen mit immobilisierten microorganismen. 1. Transformation von cholesterin zu cholestenon in organschen Losungsmittelm. Z. Allg. Mikrobiol., 20: 79-84. Bar, R. (1987) In 'Biocatalysis in organic media'. ed by Laane, C., Tramper, J. and Lilly, M. D., Elsevier. Amsterdam, vol. 29, pp. 147-153. Barbaric, S. and Luisi, L. (1981) Micellar solubilization of biopolymers in organic solvents: 5. Activity and conformation of cx-chymotrypsin in isooctane-AOT reverse micelle. J. Am. Chem. Soc., 103: 4239-4244. Barbas, C., Matos, J. R., West, J. B. and Wong, C. H. (1988) A search for peptide ligase: cosolvent-mediated conversion of protease to esterase for irreversible synthesis of peptides. J. Am. Chem. Soc., 110: 5162-5166. Barford, J. P. and Hall, R. J. (1981) A mathematical model for the aerobic growth of Saccharomyces cerevisiae with a saturated respiratory capacity. Biotech. Bioeng., 23: 1735-1767. Barford, J. P. (1987) The technology of aerobic yeast growth. In 'Yeast Biotechnology'. ed. by Berry R. I. and Stewart, G. C., Allen and Unwin. London. pp. 220-227. References 238

Barton, M. J., Hamman, J.P., Fichler, K. C. and Calton, G. J. (1990) Enzymatic resolution of (R, S)-2-(4-hydroxylphenoxy) propionic acid. Enz. Microb. Technol., 12: 577-583 Becvarova, H. and Hane, 0. (1963) Production of phenylacetykarbinol by various yeast species. Folia Microbiol., 8: 42-47. Becvarova, H., Hane, 0. and Macek, K. (1963) Course of transformation of benzaldehyde by Saccharomyces cerevisiae. Folia Microbiol., 8: 165-169. Bell, G., Todd, J. R., Blain, J. A., Patterson. J. D. E. and Shaw, C. E. (1981) Hydrolysis triglyceride by solid phase lipolytic enzymes of Rhizopus arrhizus in continuous reactor system. Biotech. Bioeng., 23: 1703-1719. Bergmeyer, U. H. (1974) In 'Methods of Enzymatic Analysis'. vol. 1, Academic Press Inc., New York, London, pp. 429-509. Berry, D. R (1989) Growth of yeast. In 'Fermentation process development of industrial microorganisms' ed. by Neway, J. 0. Dekker Pub. vol. 25, pp. 277-311. Bhagwanjee, S., Rocke, D. A., Rout, C. C., Koovarijee, R. V. and Brijaball, R. (1990) Prevention of hypotension following spinal anaesthesia for elective cesarean section by wrapping of the legs. British J. Anaesthesia, 65 (6): 819-822. Boiteux, A. and Hess, B. (1970) Allosteric properties of yeast pyruvate decarboxylase. FEBS Letters, 9 (5): 293-296. Bradford, M. M. (1970) A rapid and sensitive method for the quantification of microgram quantities of protein utilization, the principal of protein dye binding. Anal. Biochem., 72: 248-254. Brau, B. and Sahm, H. (1986) Cloning and expression of the structural gene for pyruvate decarboxylase of Zymomonas mobilis in E. coli. Arch. Microbiol,. 144: 296-301 References 239

Bringer-Meyer, S., Schimz, K. L. and Sahm, H. (1986) Pyruvate decarboxylase from Zymomonas mobilis. Isolation and partial characterization. Arch. Microbiol., 146: 105-110. Bringer-Meyer, S. and Sahm, H. (1988) Acetoin and phenylacetylcarbinol formation by the pyruvate decarboxylase of Zymomonas mobilis and Saccharomyces carlsbergensis. Biocatalysis, 1: 321-331. Brix, T., Drew, S. W. and Buckland, B. C. (1988) Fermentation process development within a computer controlled pilot plant. In 'Progress in industrial microbiology' ed. by Bushell, M. E. vol. 25, pp. 151-193. Bruinenberg, P. M., van Dijken, J. P. and Scheffers, W. A. (1983) An enzyme analysis of NADPH production and consumption in Candida utilis. J. Gen. Microbiol., 129: 965-971. Buckland, B. C., Dunnill, P. and Lilly, M. D. (1975) The enzymatic transformation of water-insoluble reactions in non­ aqueous solvents. Biotech. Bioeng., 17: 815-816. Butler, L. G. (1979) Enzymes in non-aqueous solvents. Enz. Microb. Technol., 1: 253-259. Butler, L. G and Mc Connell, D. J. (1988) Identification of an upstream activation site in the pyruvate decarboxylase structural gene (PDCl) of Saccharomyces cerevisiae. Curr. Genet., 14: 405-412. Cardillo, R., Servi, S. and Tinti, C. (1991) Biotransformation of unsaturated aldehydes by microorganisms with pyruvate decarboxylase activity. Appl. Microbiol. Biotech., 36: 300-303. Carlson, G. L. and Brown, G. M. (1961) The natural occurrence, enzymatic formation and biochemical significance of a hydroxyethyl derivative of the thiamine pyrophospha te. J. Biol. Chem., 236 (7): 2099-2108. References 240

Chandra, A. H. (1979) Effects of amino acids on ephedrine production in Ephedra gerardiana callus cultures. Phytochem., 18 (3): 484-485. Cohen, A. L. (1979) Critical point drying-principles and procedures. In 'Scanning electron microscopy', ed. by SEM Ind., AMF O'Hane, pp. 303-324. Coles, H. W., Manske, R. H. F. and Johnson, T. B. (1929) Synthesis of ephedrine and structurally similar compounds. III. A new synthesis of orthodiketones. J. Chem. Soc., 51: 2269-2272. Coletti-Previero, M., Previero, A. and Zuckerkland, E. (1969) Separation of the proteolytic and esterasic activities of trypsin by reversible structure modification. J. Mol. Biol., 39: 493-501. Cooper, T. G. (1977) Protein Purification. In 'The Tools of Biochemistry' A Wiley Interscience Pub. New York, London, Ch 10. pp. 355-405. Cremonesi, P., Carrea, G., Sportoletti, G. and Autonini, E. (1973) Enzymatic dehydrogenation of steroid in two phase systems. Arch. Biochem. Biophys., 159: 7-10. Cremonesi, P., Carrea, G., Ferrara, L. and Antonini, E. (1974) Enzymatic dehydration of testosterone coupled to pyruvate reduction in a two-phase system. Eur. J. Biochem., 44: 401-405. Crosby, J., Stone, R. and Lienhard, G. E. (1970) Mechanisms of thiamine-catalysed reactions, decarboxylation of 2-(1)­ carboxyl-1-hydroxyethyl-3,4-dimethylthisodium chloride. J. Am. Chem. Soc., 92 (6): 2891-2900. Crueger, W. and Crueger, A. (1982) Microbial transformation. In 'Biotechnology. A Textbook of Industrial Microbiology'. translated by Brock, T. D. Science Tech. Ind., Madison, Ch. 15, pp. 249-266. Cruely, C., Larroche C. and Gros, J.-B. (1992) Bioconversion of fatty acids into methylketones by spores of Penicillium roquefortii in a water-organic solvent, two-phase system. Enz. Microb. Technol., 14: 669-678. References 241

Culic, J., Netrval, J., Souhrada, J., Ulbrecht, S., Vojtisek, V. and Vodnansky, M. {1984) Method of the cost reduction in the production of D-(-)-1-phenyl-1-hydroxy-2-propanone for the preparation of L-{-)ephedrine. Czech Patent 222,941(CI. C07C 49/24). Dapaah, S. Y. and Hill, G. A. (1992) Biodegradation of chlorophenol mixtures by Pseudomonas putida. Biotech. Bioeng., 40: 1353-1358. Dissara, Y. , Moradi-Tafreshi, Z. and Regers, P. L. (1993) Kinetic modelling and simulation of bioconversion of benzaldehyde to L-phenylacetylcarbinol by Candida utilis. 11th Australian Biotechnology Conference, Perth. pp. 91-93. Dollery, C. (1991) Ephedrine (hydrochloride). In 'Therapeutic Drugs' ed. by Dollery, C., Churchill Pub., London, Livingstone, E26-29. Dordick, J. S. (1989) Enzymatic catalysis in monophasic organic solvents. Enz. Microb. Technol., 11: 194-211. Duarte, J.M. C. and Lilly, M. D. (1980) The use of free and immobilized cells in the presence of organic solvents: the oxidation of cholesterol by Nocardia rhodochrous. In 'Enzyme Engineering', vol. 5 . Plenum Press, New York, pp. 363-367. Dulloo, A. G. and Miller, D.S. (1987) Aspirin as a promoter of ephedrine-induced thermogenesis, potential use in the treatment of obesity. Am. J. Clinical Nutrition, 45 (3): 564-569. Dulloo, A. G., Seydoux, J. and Giravidier, L. (1992) Potentiation of the thermogenetic antiobesity effects of ephedrine by dietary methylxanthines; adenosine antagonism or phosphodiesterase inhibition ? Metabolism Clinic. & Experiment., 41 (11): 1233-1241. Ebisuzaki, K. and Barron, E. G. S. (1957) Yeast Alcohol Dehydrogenase II. Arch. Biochem. Biophys, 69: 555-564. Eldor, J. (1991) Ephedrine in the initial treatment of haemorrhagic shock. Medical Hypotheses, 35 (3): 250-252. References 242

Favre-Bulle, 0., Weenink, E., Vos, T. Preusting, H. and Witholt, B. (1993) Continuous bioconversion of n-octane to octanoic acid by recombinant E. coli (alk+) growing in a two-liquid phase chemostat. Biotech. Bioeng., 41: 263-272. Feigl, F. and Anger, V. (1966) Aldehydes (-CHO group). In 'Spot Tests in Organic Analysis'. translated by Oesper, R., Elsevier Pub. Co., Amsterdam, London, NewYork. pp. 195-197. Fiechter, A., Fuhrmann, G. F. and Kappeli, 0. (1981) Regulation of glucose metabolism in growing yeast cells. Adv. Micro. Physiol., 22: 123-183. Flygare, S. (1988) Biotransformation of steroids by Mycobacterium sp. Ph. D. Thesis, Pure and Appl. Biochemistry, Lund Uni., Sweden. Fraleigh, S., Bungay, H. and Fiechter, A. (1989) Regulation of oxidoreductive yeast metabolism by extracellular factors. J. Bacteriol., 12: 185-198. Franzblau, S. G. and Sinclair, N. A. (1983) Induction of pyruvate decarboxylase in Candida utilis. Mycopathologia, 83: 29-33. Freeman, A (1984) Understanding enzyme stabilization. Trends in Biotechnol., 2: 147-148. Fukui, S., Ahmed, S. A., Omata, T. and Tanaka, A. (1980) Bioconversion of lipophilic compounds in non-aqueous solvent. Effect of gel hydrophobicity on diverse conversion of testosterone by gel entrapped Nocardia rhodochrous cells. Eur. J. Appl. Microbiol. & Biotechnol., 10: 289-301. Fukui, S., Tanaka, A. and Iida, T. (1987) Immobilization of biocatalysts for bioprocesses in organic media. In 'Biocatalysis in organic media' ed. by Laane, A., Tramper, J. and Lilly, L. M. Elsevier Pub., Amsterdam, pp. 21-24. Gerald K. G., Scollar, M. P. and Klibanov, A. M. {1985) Resolution of racemic mixtures via lipase catalysis in organic solvents. J. Am. Chem. Soc., 107: 7072-7076. References 243

Godfrey, N. B. (1972) Solvent selection via miscibility number. Chemtech, 2: 359-363. Goetschel, K., Barenholz, Y. and Bar, R. {1992) Microbiol conversions in a liposomal medium. Part 2: Cholesterol oxidation by Rhodococcus erthropolis. Enz. Microb. Technol., 14: 390-395. Gounaris, A. D., Turkenkopf, I., Buckwald, S. and Young, A. (1971) Pyruvate decarboxylase: 1. Protein dissociation into subunits under conditions in which thiaminepyrophosphate is released. J. Biol. Chem., 246 (5): 1302-1309. Gounaris, A. D., Turkenkopf, I., Civerchia, L. D. and Greenlie, J. (1975) Specificity restrictions for thiamine pyrophosphate in the protein dissociation step, subunit structure. Biochim. Biophys. Acta, 405: 492-498. Green, D. E., Herbert, D., and Subrahmanyan, V. (1941) Carboxylase. J. Biol. Chem., 138 327-339. Gross, N. H. and Werkman, C. H. {1947) Isotopic composition of acetylmethylcarbinol formed by yeast juice. Arch. Biochem., 15: 125-131. Gruber, M. and Wassenaar, J. S. (1960) Inhibition of yeast carboxylase by acetaldehyde. Biochim. et Biophys. Acta, 38: 355-357. Gupta, K. G., Singh, J., Sahni, G. and Dhawan, S. (1979) Production of Phenylacetylcarbinol by yeast. Biotech. Bioeng., 21: 1085-1089. Halling, P. J. (1984) Effects of water on equilibria catalysed by hydrolytic enzymes in biophasic reaction systems. Enz. Microb. Technol., 6: 513-516. Halling, P. J. (1987) Biocatalysis in low water, predominantly organic reaction mixture. In 'Bioreactors and Biotransformations', ed. by Moody, G. W. and Baker, P. B., Elsevier Applied Science, New York. p. 189. References 244

Hames, B. D. (1981) Introduction to polyacrylamide gel electrophoresis. In 'Gelelectrophoresis of proteins, a practical approach', ed. by Rickwood. D. and Hames, B. D, IRI Press. Oxford, England, Ch. 1, pp. 1-28. Hane, 0. and Kakac, B. (1956) Hefe carboxylase und Bildung von phenylacetylcarbinol. Naturwissenschaften 43: 498. Haukeli, A. D. and Lie, S. (1972) Production of diacetyl, 2-acetolactate and acetoin during fermentation. J. Inst. Brew., 78: 229-232. Hayat, M. A. (1978) In 'Introduction to biological scanning electron microscopy', University Park Press, Baltimore, pp. 131-169. Hedstrom, G., Slotte, J. P, Backlund, M., Molander, 0. and Rosenholm, J. B. (1992) Lipase-catalyzed synthesis and hydrolysis of cholesterol oleate in AOT / isooctane microemulsions. Biocatalysis, 6: 281-290. Hilldebrandt, G. and Klavehn, W. (1930) Verfahren zur Herstellung von 1-1-Phenyl-2-methylaminopropan-1 -ol. German Patent (DRP) 548,459; 1930 (77)'. Hilldebrandt, G. and Klavehn, W. (1934) In 'US Patent 1,956,950; 1934'. Hiroshi, 0., Hideki, M. and Yoshio, H. (1985) Synthesis of aspartame precursor by solid thermolysin in organic solvent. Biotech. Letters, 7 (11): 789-792. Hohmann, S. and Cederberg, H. (1990) Autoregulation may control the expression of yeast pyruvate decarboxylase structural genes PDC 1 and PDC 5. Eur. J. Biochem., 188: 615-621. Holzer, H. and Goedde, H. W. (1957) Zwei wege von pyruvat zu Acetyl-coenzyme A in Hefe. Biochem. Z. 329: 157-191. References 245

Holzer, H. and Beaucamp, K. (1961) Detection and characterization of cx.-lactyl-TPP and cx.-hydroxy-ethyl­ TPP as intermediary products of pyruvate decarboxylation. Biochim. Biophys. Acta, 46: 225-243. Hommes, F. A. (1966) Effect of glucose on the level of glycolytic enzymes in yeast. Arch. Biochem. Biophys, 114: 231-233. Hopmann, R. F. W. (1980) Hydroxy-ion-induced subunit dissociation of yeast cytoplasmic pyruvate decarboxylase. A circular dichroism study. Eur. J. Biochem., 110: 311-318. Hoppner, C. and Doelle, H. W. (1983) Purification and kinetic characteristics of pyruvate decarboxylase and ethanol dehydrogenase from Zymomonas mobilis in relation to ethanol production. Eur. J. Appl. Microbiol. & Biotechnol., 17: 152-157. Hu, Shui-Ying (1969) Ephedra (Ma-Huang) in 'The New Chinese Materia Medica'. Economic Botany, 23: 346-351. Hull, K. M. and Maher, T. J. (1991) L-Tyrosine fails to potentiate several peripheral actions of the sympathomimetics. Pharmacology Biochem. & Behaviour, 39 (3): 755-760. Hubner, G. and Weidhase and Schellenberger, A. (1978) The mechanism of substrate activation of pyruvate decarboxylase: A first approach. Eur. J. Biochem., 92: 175-181. Hubner, G., Konig, S., Schellenberger, A. and Koch, M. H.J. (1990) An X-ray solution scattering study of the cofactor and activator induced structural changers in yeast pyruvate decarboxylase. FEBS Letters, 266 (1, 2): 17-20. Hubner, G., Konig, S., and Schnackerz, K. D. (1992) Correlation of cofactor binding and the quaternary structure of pyruvate decarboxylase as revealed by 3lp NMR spectroscopy. FEBS, 314 (1): 101-103. References 246

Ingram, L. 0. and Buttke, T. M. (1984) Effects of alcohols on microorganisms. Adv. Microb. Physiol., 25: 253-300. Janssen, A. E. M., van Padt. A., van Sonesbeek, H. M. and van Riet, K. (1993) The effect of organic solvents on the equilibrium position of enzymatic acylglycerol synthesis. Biotech. Bioeng., 41: 95-103. Jencks, W. P. (1975) In 'Advances in Enzymology and Related Areas of Molecular Biology', vol. 43, ed. by Meister, A., Wiley Pub., New York, p 219. Jonderko, K., and Kudo, C. (1991) Effect of anti-obesity drugs promoting energy expenditure, yohimbine and ephedrine, on gastric emptying in obese patients. Alimentary Pharmacology & Therapeutics, 5 (4): 413-418. Juni, E. (1952) Mechanisms of the formation of acetoin by yeast and mammalian tissue. J. Biol. Chem., 195: 727-734. Juni, E. (1961) Evidence for a two site mechanism for decarboxylation of a.-keto acid by a.-carboxylase. J. Biol. Chem., 236: 2302-2308. Kamat, 5., Barrera, J., Beckman, E. J. and Russell, A. (1992) Biocatalytic synthesis of acrylates in organic solvent and supercritical fluids: 1. Optimization of enzyme environment. Biotech. Bioeng., 40: 158-166. Kaul, R. and Mattiasson, B. (1986) Extractive bioconversion in aqueous two-phase systems: Production of prednisolone from hydrocortisone. Appl. Microbiol. Biotechnol., 24: 259-265. Kappeli, o. (1986) Regulation of carbon metabolism in Saccharomyces cerevisiae and related yeasts. Adv. Micro. Physiol., 28: 183-209. References 247

Kellermann, E. and Hollenberg, C. P. (1988) The glucose and ethanol-dependent regulation of PDC 1 from Saccharomyces cerevisiae is controlled by two distinctive promoter regions. Curr. Genet., 14: 337-344. Kennedy, J. F. and Cabral, J.M. S. (1983) Immobilized living cells and their application. In 'Applied Biochemistry and Bioengineering; vol. 4, Immobilized Microbial Cells'. ed. by Chibata, I. and Wingard, L. B., Academic Press. New York. pp. 198-280. Khmelnitski, Y. L., Fam K. D., Semenov, A. V. Martinek, K., Veruovic, B., and Kubanek, V. (1984) Optimization of enzyme catalyzed peptide synthesis in a 'water-water-immiscible organic solvent' biphasic system. Tetrahedron, 40: 190-194. Kieslich, K (1984) What are Biotransformations?. In 'Biotechnology. A Comprehensive Treatise in 8 volumes; Biotransformation' vol. 6a. ed by. Rehm, H. -J. and Reed G., Verlag Chemie GMbH., Weinheim, pp. 1-4. Kirchner, G., Scollar, M. P, Klibanov, A. M. (1985) Resolution of racemic mixtures via lipase catalysis in organic solvents. J. Am. Chem. Soc., 107: 7072-7076. Klibanov, A. M. (1986) Enzymes that work in organic solvents. Chemtech, 16: 354-357. Kluyver, A. J. and Donker, J. L. (1924) Proc. Acad. Sci Amsteradam, 28: 314. Krampitz, L. 0. (1959) An active acetaldehyde-thiamine intermediate. Fed. Proc. Czechoslovak, 18 (1): 266. Larroche, C., Creuly. C. and Gros, J.-B (1992) 2-Heptanone production by spores of Penicillium roquefortii in a water-organic solvent two-phase system. Biocatalysis 5: 163-173. Lee, Y. Hand Tsao, G. T. (1979) Dissolved oxygen electrodes. Adv. Biochem. Eng,. 13: 36-83. References 248

Lehmann, H., Fisher, G., Hilbner, G., Kohnert, K. D. and Schellenberger, A. (1973) The influence of steric and electronic parameters on the substrate behaviour of a-oxo acids to yeast pyruvate decarboxylase. Eur. J. Biochem., 32: 83-87. Lehninger, A. L. (1982) Enzymes. In 'Principles of Biochemistry', ed. by Anderson, S. and Fox J., Worth Pub, Inc., New York, Ch. 9, pp. 207-247. Lenders, J.P., Germain, P. and Chrichton, R.R. (1985) Immobilization of a soluble chemically thermostabilized enzyme. Biotech. Bioeng., 27: 572-578. Leuenberger, H. G. W. (1984) Methodology. In' Biotechnology. A Comprehensive Treatise in 8 volumes; Biotransformation' vol. 6a, ed. by Rehm, H. -J. and Reed G., Verlag Chemie GMbH., Weinheim, pp. 6-29. Long, A. and Ward, 0. P. (1989a) Biotransformation of aromatic aldehydes by Saccharomyces cerevisiae: investigation of reaction rates. J. Ind. Microbiol., 4: 49-53. Long, A. and Ward, 0. P. (1989b) Biotransformation-of benzaldehyde by Saccharomyces cerevisiae: characterization of the fermentation and toxicity effects of substrates and products. Biotech. Bioeng., 33: 933-941. Lowell, B. B., Napolitano, A., Usher, P., Dulloo, A. G., Rosen, B. N., Spiegelman, B. M. and Flier, J. S. (1990) Reduced adipsion expression in murine obesity: effect of age and treatment with the sympathomimetic-thermogenic drug mixture ephedrine and caffeine. Endocrinology, 126 (3): 1514-1520. Lowry, 0. H., Roseburough, N. J., Farr, A. L. and Randall, R. J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem., 193: 265-275. Mahmoud, W. M., El-Sayed, AdH. M. M. and Robert, W. (1990a) Production of L-phenylacetylcarbinol by immobilized yeast cells: I. Batch Fermentation. Biotech. Bioeng., 36: 47-54. References 249

Mahmoud, W. M., El-Sayed, AdH. M. M. and Robert, W. (1990b) Production of L-phenylacetylcarbinol by immobilized yeast cells: II. Semicontinuous Fermentation. Biotech. Bioeng., 36: 55-63. Maitra, P. K. and Lobo, Z. (1971) A kinetic study of glycolytic enzyme synthesis in yeast. J. Biol. Chem., 246: 475-488. Malcata, F. X., Reyer, H. R. Garcia, H. S., Hill Jr, C. G. and Amundson, C. H. (1992) Kinetics and mechanisms of reactions catalysed by immobilized lipases. Enz. Micro. Technol., 14: 426-446. Manske, R. H. F. and Johnson T. B. (1929a) Synthesis of ephedrine and structurally similar compounds. I. A new synthesis of ephedrine. J. Chem. Soc., 51: 580-582. Manske, R.H. F. and Johnson T. B. (1929b) Synthesis of ephedrine and structurally similar compounds. II. The synthesis of some ephedrine homologs and resolution of ephedrine. J. Chem. Soc., 51: 1906-1909. Margolin, A. L., Delinck, D. L and Whalon, M. R. (1990) Enzyme-catalyzed regioselective acrylation of castnospermine. J. Am. Chem. Soc., 112 (8): 2849-2854. Margolin, A. L., Fitzpatrick, P.A., Dubin, P. L. and Klibanov, A. M. (1991) Chemoenzymatic synthesis of optically active (meth)acetyl polymer. J. Am. Chem. Soc., 113: 4693-4694. Margolin, A. L. and Klibanov, A. M (1987) Peptide synthesis catalyzed by lipases in anhydrous organic solvents. J. Am. Chem. Soc., 109: 3802-3804. Mark, H. F., McKetta, J. and Othmer, D. F. (1963) Alkaloids. In 'Kirk-Othmer Encyclopedia of Chemical Technology', 2nd edn. vol. 1, ed. by Standen, A., Interscience, a Div. of John Wiley & Sons Inc., New York, London, pp. 788-789. Martinek, K. Semenov, A. N. and Berezin, I. V. (1981) Enzymatic synthesis in biphasic aqueous-organic systems. I. Chemical equilibrium shift. Biochim. et Biophys. Acta, 658: 76-89. References 250

Marty, A., Chulalakasananukul, W., Willemot, R. M. and Condoret, J. S. (1992) Kinetics of lipase-catalyzed esterification in supercritical CO2. Biotech. Bioeng., 39: 273-280. May, S. W (1992) Biocatalysis in the 1990s: A prospective. Enz. Microb. Technol., 14: 80-84. Morton, J. F. (1977) Ephedra. In 'Major Medical Plants. Botany, Culture and Uses'. Charles G. Thomas Pub. Spring field, Illinois., pp. 33-37. Nagashima, T., Akira, W. and Hideo, K. (1992) Peptide synthesis by proteases in organic solvents: Medium effect on substrate specificity. Enz. Microb. Technol., 14: 842-847. Nakamura, K., Chi, Y. M. Yamada, Y. and Yano, T. (1986) Lipase activity and stability in supercritical fluids. Chem. Eng. Commun., 45: 207-212. Neale, A. D., Scopes, R. K., Wettenhall, R. E. H. and Hoogenrad, N. J. (1987) Nucleotide sequence of the pyruvate decarboxylase gene from Zymomonas mobilis. Nucleic Acids Research, 15 (4): 1753-1761. Nebesky, F., Souhrada, J. and Jaki, V. (1978) L-ephedrine from phenylacetylcarbinol. Czeck Patent 186,027 (Cl C07C 19 /16) Netrval, J. and Vojtisek, V. (1982) Production of Phenylacetylcarbinol in Various Yeast Species. Eur. J. Appl. Microbial. & Biotechnol., 16: 35-38. Neuberg, C. and Liebermann, L. (1921) Zur Kenntnis der carboligase Mitteilung II. Biochem. Z., 115: 311-325. Neuberg, C. (1946) The Biochemistry of Yeast Annual Review of Biochemistry, 15: 435-437. Neuberg, C. and Hirsch, J. (1921) Uber ein kohlenstoffketten knupfendes ferment (Carboligase). Biochem. Z., 115: 282-310. References 251

Nikolova, P. and Ward, 0. P. (1991) Production of L-phenylacetylcarbinol by biotransformation: product and by-product formation and activation of the key enzymes in wild­ type and ADH isoenzyme mutants of Saccharomyces cerevisiae. Biotech. Bioeng., 38: 493-498. Nikolova, P. and Ward, 0. P. (1992) Reductive biotransformation by wild type and mutant strains of Saccharomyces cerevisiae in aqueous-organic solvent biphasic system. Biotech. Bioeng., 39: 870-876. Nilsson, H., Mosbach, R. and Mosbach, K. (1972) The use of bead polymerization of acrylic monomers for immobilization of enzymes. Biochim. et Biophys. Acta, 268: 253-256. Oba, K. and Uritani, I. (1992) Pyruvate decarboxylase from sweet potato roots. In 'Methods in Enzymology'. vol. 90, Academic Press, New York, San Francisco, London, pp. 528-532. Okachi, R. and Nara,_T. (1973) Penicillin acylase of Pseudomonas melanogenum KY 3987 Agri. Biol. Chem., 37 (12): 2797-2804. Osborne, S. J., Leaver. J., Turner, M. K. and Dunnill, P. (1990) Correction of biocatalytic activity in an organic-aqueous two liquid phase system with solvent concentration in the cell membrane. Enzyme. Microb. Technol., 12: 281-291. Ose, S. and Hironaka, J. (1957) In Intl. Symp. Enzyme Chem; Tokyo, Koyto, pp. 457-460. Pasquali, R., Cesari, M. P., Mekhionda, N., Stefanini, C., Raitano. A, and Labo, G. (1987) Does ephedrine promote weight loss in low energy adapted obese women? Internal J. Obesity (London), 11 (2): 163-168. Pasquali, R., Casimirri, F., Mikhionda, N., Grossi, G., Bortoluzzi, L., Morselli-Labate, A. M., Stefanini, C., and Raitano, A. (1992) Effects of chronic administration of ephedrine during very low calories diets on energy expenditure, protein metabolism and hormone levels in obese subjects. Clinical Science (London), 82 (1): 85-92. References 252

Perry, R.H. (1975) In 'Perry's Chemical Engineering Handbook'. 6th edn, McGraw-Hill International Edition, Chemical Engineering Series, pp. 3-27. Peterson, D. H., Murray, H. C., Eppstein, H. S., Reineke, L. M, Weintraub, A., Meister, P. D. and Leigh, H. M. (1952) Microbiological transformations of Steroids. 1. Introduction of oxygen at carbon-11 of progesterone. J. Am. Chem. Soc., 74: 5933-5936. Prazeres, D. M. F., Garcia, F. -A. P. and Cabral, J. M .S. (1993) An ultrafiltration membrane bioreactor for the lipolysis of olive oil in reverse micelle media. Biotech. Bioeng., 41: 761-770. Ramawat, K. G. and Arya, H. C. (1979a) Alkaloid content of Ephedra in vivo and in vitro. Indian J. Exp. Biol., 17 (1): 106-107. Ramawat, K. G. and Arya, H. C. (1979b) Effect of some growth regulators on ephedrine production in Ephedra gerardiana callus cultures. Phytochem., 18: 484-485. Reslow, M., Adlercreutz, P. and Mattiasson, B. (1992) Modification of the microenvirorunent of enzymes in organic solvents. substitution of water by polar solvents. Biocatalysis, 6: 307-318. Rieger, M., Kappeli, 0 and Fiechter, A. (1983) The role of limited respiration in the incomplete oxidation of glucose by Saccharomyces cerevisiae. J. Gen. M~crobiol., 129: 653-661. Rizzi, M., Stylos, P. Riek, A. and Reuss, M. (1992) A kinetic study of immobilized lipase catalysing the synthesis of isoamyl acetate by transesterification in n-hexane. Enz. Microb. Technol., 14: 709-714. Rogers, P. L., Lee, K. J., Skotnicki, M. L. and Tribe, D. E. (1982) Ethanol production by Zymomonas mobilis. Adv. Biochem. Eng., 23: 37-84. References 253

Rose, N. W. and Schneider, H. (1991) Activities of Candida rugosa lipase and other esterolytic enzymes coated on glass beads and suspended in substrate and water vapor: Enzymes in their liquid film. Enzyme Microb. Technol., 13: 370-377. Rozano, L. S., Zamyshlysera, M. D., Kondratlev, S. A. and Lzotova, G. S. (1978) Study on the Extraction of a L-ephedrine base in apparatus of continuous action. Khim-Farm. Zh.• 12 (11): 92-94. Ruiz-Amil, M., Aparicio, M. L. and Canovas, J. L.(1969) Regulation of the synthesis of glyceraldehyde dehydrogenase in Pseudomonas putida. FEBS Letters, 3: 65-67. Schaaff, I., Heinisch, J. and Zimmermann, F. K. (1989) Overproduction of glycolytic enzymes in yeast. Yeast, 5: 285-290. Schellenberger, A. and Hilbner (1967) Mechanismus und kinetik der rekombination und darausabgeleitete bindungsvergaltnisse im aktiven zentrum der Hefe-pyruvat decarboxylase. Hoppe-Seyler's Z. Physiol Chem., 348: 491-500. Schlatmann, J., Aires-Borros , M. R. and Cabral, J. M. (1991) Esterification of short chain organic acids with alcohols by a lipase microencapsulated in reverse micelles. Biocatalysis, 5: 137-144. Schmidt-Kastner, G and Egerer, P. (1984) Amino acids and peptides. In 'Biotechnology. A Comprehensive Treatise in 8 volumes; Biotransformation' vol. 6a. ed by. Rehm, H. -J. and Reed G., Verlag Chemie GMbH., Weinheim, p389. Schmitt. H. D. and Zimmermann, F. K. (1982) Genetic analysis of the pyruvate decarboxylase reaction in yeast glycol ysis. J. Bacteriol.. 151 (3): 1146-1152. Schmitt. H. D .. Ciriacy, M. and Zimmermann, F. K. (1983) The synthesis of yeast pyruvate decarboxylase is regulated by large variation in the messenger RNA level. Mol. Gen. Genet., 192: 247-252. References 254

Scopes, R. K. (1989) Effect of ethanol on glycolytic enzymes. In 'Alcohol toxicity in Yeast and bacteria'. ed. by van Unden, N. CRC Press Inc., Ch. 3, pp. 89-100. Scrutton, M. C. (1971) Assay of enzymes of CO2 metabolism. In 'Methods in Enzymology' vol. 6A. ed. by Morris, J. R, and Ribbons, D. W. New York, San Francisco, London, pp-479-514. Seely, R. J., Heefner, D. L., Hageman, R. V., Yarus, M. J. and Sullivan, S. A. (1989a) Process for producing L-phenylacetykarbinol (PAC); An immobilized cell mass for use in the process and method for preparing the cell mass. US Patent PCT /US 89 /04421. Seely, R. J., Heefner, D. L., Hageman, R. V., Yarus, M. J. and Sullivan, S. A. (1989b) Process for making phenylacetykarbinol (PAC). Microorganisms for use in the process, and method for preparing the microorganism. US Patent PCT /US 89 / 04423. Semenov, A. N., Khmelnitski, Y. L., Berezin, I. V. and Martinek, K. (1987) Water-organic solvent two phase systems as media for biocatalytic reactions: potential for shifting chemical equilibria toward the higher yield of end products. Biocatalysis, 1: 3-8. Sharma, S. and Tauro, P. (1986) Control of ethanol production by yeast: role of pyruvate decarboxylase and alcohol dehydrogenase. Biotech. Letts., 8 (10): 735-738. Sharma, S. and Tauro, P. (1987) Control of ethanol production by yeast: pyruvate accumulation in slow ethanol producing Saccharomyces cerevisiae. Biotech. Letts., 9 (8): 585-586. Sims, A. P. and Barnett, J. A. (1991) Levels of activity of enzymes involved in anaerobic utilization of sugar by six yeast species: observations towards understanding the Kluyver effect. FEMS Microbiol. Letters, 77: 295-298. Sims. A. P .. Stalbrand. H. and Barnett, J. A. (1991) The role of pyruvate decarboxylase in the Kluyver Effect in the food yeast. Candida utilis. Yeast, 7: 479-487. References 255

Singer, T. P and Pensky, J. (1952a) Isolation and properties of the a-carboxylase of wheat germ. J. Biol. Chem., 196: 375-388. Singer, T. P. and Pensky, J. (1952b) Mechanism of acetoin synthesis by a-carboxylase. Biochim. Biophys. Acta, 9: 316-327. Smith, P. F. and Hendlin, D. (1953) Mechanism of phenylacetykarbinol synthesis by yeast. J. Bacteriol., 65: 440-445. Smith, P. F. and Hendlin, D. (1954) Further studies on phenylacetykarbinol synthesis by yeast. J. Appl. Microbiol., 2: 294-296. Smith, L. L. (1984) Steroids. In 'Biotechnology. A Comprehensive Treatise in 8 vols; vol 6a Biotransformation'. ed. by Rhem, H. -J. and Reed G., Verlag Chemie GMbH. Weinheim., pp 33-36. Smith, D. A and Perry, P. J. (1992) The efficacy of ergogenic agents in athletic competition: Part II. Other performance enhancing agents. Annals of Pharmacotherapy, 26 (5): 653-659. S0rensen, G. G. and Spenser, I. D. (1988) Biosynthesis of ephedrine. J. Am. Chem. Soc., 110: 3714-3715. S0rensen, G. G. and Spenser, I. D. (1989) The biosynthesis of ephedrine. Can. J. Chem. 67: 998-1009. Stoicheva, N. G., Davey, C. L., Mark, G. H. and Kell, D. B. (1989) Dielectric spectroscopy: A rapid method for the determination of solvent biocompatibility during biotransformation. Biocatalysis, 2(4): 245-255. Thomson, William A. R. (1978) In 'Healing Plant. A Modern Herbal'. Mc Gaw-Hill Book Company, p.62 Tilston, J. (1988) 'ICI to build plant for exports' in Sydney Morning Herald: 22nd Sep. p.25. References 256

Tripathi, C. K. M., Basu, S. K., Vora, V. C., Mason, J. R. and Pirt, S. J. (1988) Continuous cultivation of a Yeast Strain for Biotransformation of L-acetylphenykarbinol from benzaldehyde. Biotech. Letters, 10 (9): 635-636. Ullrich, J, and Donner, I. (1970) Fluorimetric study of 2-p-toluidinonaphthalene-6-sulphonate binding to cytoplasmic yeast pyruvate decarboxylase. Hoppe-Seylers's Z. Physiol. Chem., 351: 1030-1034. Ullrich, J. (1982) Structure-function relationship in pyruvate decarboxylase of yeast and wheat germ. Annals New York Academy of Science, 378: 278-305. Umbreit, W. W. (1972) Manometric and Biochemical Techniques . ed. by Umbreit, W.W., Burris, R. H. and Stauffer, J. F., Burgess Pub. Minneapolis, pp. 1-69. Urk, H. V., Mak, P. R., Scheffers, W. A., and van Dijken, J. P. (1988) Metabolic responses of Saccharomyces cerevisiae CBS 8066 and Candida utilis CBS 621 upon transition from glucose limitation to glucose excess. Yeast, 4: 283-291. Urk, H., Schipper, 0., Breedveld G. J., Paul, R. M. Scheffers, W. A. and van Dijken J. P. (1989) Localization and kinetics of pyruvate­ metabolizing enzymes in relation to aerobic alcohol fermentation in Saccharomyces cerevisiae CBS 8066 and Candida utilis CBS 621. Biochim. Biophys. Acta, 992: 78-86. van Dijken, J. P., Bosch, E. V. 0., Hermans, J. J., Miranda, R. D. and Scheffers, W. A. (1986) Alcohol fermentation by 'Non-fermentative' yeasts. Yeast, 2: 123-127. Venneland, B. and Felsher, R. Z. (1946) Oxalacetic and pyruvic carboxylase in some dicotyledonoud plants. Arch. Biochem., 11: 279-306. Vennesland, B. (1951) In 'The Enzymes' ed. by Summer, J.B. and Myrback, K., Academic Press: New York, vol. 2, pp. 183-215. References 257

Voets, J.P., van Damme, E. J., and Vlerick, C. (1973) Some aspects of the phenylacetykarbinol biosynthesis by Saccharomyces cerevisiae. Z. Allg. Mikrobiol., 13 (4): 355-366. Vojtisek, V. and Netrval, J. (1982) Effect of pyruvate decarboxylase activity and pyruvate concentration on the production of L-hydroxy-1-phenylpropanone in Saccharomyces carlbergensis. Folia Microbiol., 27: 173-177. Walker, R. B., Wood, D. M. and Akmal, M. M. (1990) Hyperthermic and anorexic effects of oxazolidines derived from L­ ephedrine in rats. Life Science, 47 (7): 595-600. Walker, R. B., Wood, D. M., Akmal, M. M. and Sharks, E. (1992) Effects of oxazolidines derived from neoephedrine in the rat. Gen. Pharmacol., 23 (4): 729-732. Walsh, C. (1979) Enzyme catalyzed decarboxylase. In 'Enzyme Reaction Mechanisms' W. H. Freeman and Company, San Francisco. Ch. 21, pp. 682-688. Wang, D. I. C., Cooney, C. L., Demain, A. L. Dunnill, P., Humphry, A. E. and Lilly, M. D. (1979) In 'Fermentation and Enzyme Technology'. John Wiley & Sons, New York, Ch 6 &7, pp. 57-137. Wang, B. (1993) Kinetic study of fed-batch and continuous bioconversion process for L-phenylacetykarbinol (L-P AC) production by Candida utilis. Ph D. Thesis, Dept. of Biotechnology, The University of New South Wales, Australia. Westerfeld, W. W. (1945) A colorimetric determination of blood acetoin. J. Biol. Chem., 161: 495-502. Wijono, D. (1991) Kinetics of phenylacetykarbinol (PAC) production by yeast Candida utilis. Ph D. Thesis. Dept. of Biotechnology, The University of New South Wales, Australia. References 258

Wiimpelmann, M., Kjaeargaard, L. and Joergenson, B. (1984) Ethanol production with Saccharomyces cerevisiae under aerobic conditions at different potassium concentrations. Biotech. Bioeng., 26: 301-307. Yamane, T. (1987) Microaqueous biocatalysis system (in Japanese). Japan Oil Chem. Soc., 36: 761-768. Yamane, T. (1988) Importance of moisture content control for enzymatic reactions in organic solvents: a novel concept of 'microaqueous'. Biocatalysis, 2 (1): 1-9. Yamasaki, K. Tamaki, T, Uzawa S., Sankawa, U. and Shibata S. (1973) Participation of C6-Cl unit in the biosynthesis of ephedrine in Ephedra. Phytochem., 12: 2877-2882. Yamazaki, Y., and Suzuki. H. (1979) Report Ferment. Res. Institute, Japan, 52: 33-40. Yang, L. W., Hustedt, H. and Kula, M.-R. (1988) Partition of Brevibacterium ammoniageneses cells in an aqueous two-phase system and their use for the conversion of fumarate. Biotechnol. Appl. Biochem., 10: 173-182. Yatco-Manzo, E., Roddy, F., Yount, R. G. and Metzler, D. E. (1959) The catalytic decarboxylation of pyruvate by thiamine. J. Biol. Chem., 234 (4): 733-737. Yokozeki, K., Yamanaka, S., Utagawa, T. and Fukui, S. (1982) Production of adenine arabinoside by gel-entrapped cells of Enterobacter aerogenes in water-organic cosolvent system. Eur. J. Appl. Microbiol. & Biotechnol., 14: 225-231. Zaks, A., and Klibanov, A. M. (1984) Enzymatic catalysis in organic media at 100°C. Science, 224: 1249-1251. Zaks, A., and Klibanov, A. M. (1986) Substrate specificity of enzymes in organic solvents vs. water is reversed. J. Am. Chem. Soc., 108: 2767-2768. Zaks, A. and Klibanov, A. M. (1988) The effect of water on enzyme action in organic media. J. Biol. Chem., 263: 8017-8021. References 259

Zehender, H., Trescher, D. and Ullrich, J. (1987) Improved purification of pyruvate decarboxylase from wheat germ. Its partial characterization and comparison with yeast enzyme. Eur. J. Biochem., 167: 149-154. Appendices 260

APPENDICES

Appendix 1. Nomenclature

1.1 Terms/Abbreviations

A absorbance ADH alcohol dehydrogenase ADHcin cinnamyl alcohol dehydrogenase ADHr total alcohol dehydrogenase BA benzy1 alcohol BZ benzaldehyde CSTR continuous stirred tank reactor D dilution rate DO dissolved oxygen DOT dissolved oxygen tension g gram GC gas chromatography h hour HETPP Hydroxyl ethyl thiamine pyrophosphate ID inside diameter Km Michaelis Menten constant Ks substrate saturated constant L length L liter L-PAC L-phenylacetylcarbinol min minute mg milligram mM millimole NAD+ nicotinamide adenine dinucleotide (oxidized form) NADH+H+ nicotinamide adenine dinucleotide (reduced form) OD optical density PBR packed bed reactor POC pyruvate decarboxylase Appendices 261

pyruvate dehydrogenase pyruvate specific benzyl alcohol formation rate specific benzaldehyde consumption rate qL-PAC specific L-P AC formation rate RO reverse osmosis rpm revolution per minute RQ respiratory quotient 't space time T temperature TCA tricarboxylic acid TEMED N,N,N' ,N'- tetramethy lethylenediamine TPP thiamine pyrophosphate u enzyme unit v/v volume/volume vvm volume per volume per minute w/v weight/volume w/w weight/ weight Appendices 262

Appendix 2. Calculation of the enzyme activity

The enzyme activity is most conveniently expressed in terms of units (U) such that one unit is the amount of enzyme that catalyses the conversion of one micromole of substrate per minute (µmole/min) under defined conditions. The activity of an enzyme is expressed in terms of the specific activity which is the number of enzyme units per milligram of protein (unit/mg protein). The progress of any enzyme catalysed reaction involving an NAD­ NADH coenzyme such as ADHr, ADHcin and PDC can be observed by following the rate of appearance or disappearance of NADH from the absorbance (A) at a wavelength of 340 nm. The absorbance coefficient or molar extinction coefficient (£) of NADH at 340 nm is 6.2 cm2/µmole. This means that the conversion of one µmole substrate/mL is indicated by a change in absorbance of 6.2 for a 1 cm cuvette. For certain volume of cell extract (mL), the enzyme activity per volume can be expressed as: (d A/min). Vt Enzyme activity (unit/ml) = E. b. Ve (d A/ min). Vt Specific enzyme activity (unit/mg protein) = ------­ E. b.Vc. C

where terms are defined as follows: Vt: the total reaction volume (mL); b: the length of cuvette (cm); Ve: the sample volume (mL); C: protein concentration (mg/mL). Appendices 263

Appendix 3. Manometric measurement of enzyme activity (Umbreit, 1971)

Manometric techniques are based upon the simple gas low: PV=nRT To measure a gas-consuming or production reaction occurring in the flask. the gas present at the start was that in the gas plus that in the fluid phase or: 273 (P-R) (P-R) Gas at start = Vg ----- + Vf a------T.Po Po (gas phase) (fluid phase) At the end of the observation period this gas has been changed by the amount of which has resulted in a pressure change of hmm. If gas is taken up. h is negative; if gas is given off, h is positive.

273 (P-R+h) Gas phase is thus: Vg T Po (P-R+h) Liquid phase: V f a Po 273 (P-R+h) (P-R+h) Gas at end: Vg--- + Vfa ---­ T Po Po Evaluated gas (x) was presented as: x = final gas -initial gas

+ VI a 273 (P-R) (P-R)j ~~::~:'.'.~L lg + Vf a ~ l T Po Po 273 h h + Vfa- T Po Po

273 X= h Vg + Vf a = hk T Po Appendices 264

Summary x = amount of gas exchanged = h x k alteration in flask constant reading on open arm of manometer

k=flask constant 273 k= Vg +Vta T Po

Definition of symbols are as follows:

h : the observed change in the manometer reading in mm Vg: volume of gas phase in flask including connecting tubes down to the reference point V f: volume of fluid in vessel P: initial pressure in vessel of the gas involved in the determination

P0 : standard pressure, which is 760 mm Hg or 10,000 mm Krebs' fluid T : temperature of bath in absolute degrees (=273 + temp. in °C) a : solubility in reaction liquid of gas involved R: vapor pressure of water (or other fluid) at temperature (T)

Table 1. The solubility of carbon dioxide in pure water (a value)

Tern erature (°C) a 0 1.713 10 1.194 15 1.019 20 0.878 25 0.759 30 0.665 35 0.592 40 0.530 Appendices 265

Krebs' Manometer Fluid

Anhydrous NaBr 44 g Triton X-100 0.3 g Evans blue (or acid fuchsin) 0.3 g Water 1000 ml Density 1.033 at 20°C