Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Application of optimization for the enzymatic synthesis of sweet aspartyl dipeptide analogs catalyzed… Lee, Grace Ilah 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1992_fall_lee_grace_ilah.pdf [ 2.54MB ]
Metadata
JSON: 831-1.0098875.json
JSON-LD: 831-1.0098875-ld.json
RDF/XML (Pretty): 831-1.0098875-rdf.xml
RDF/JSON: 831-1.0098875-rdf.json
Turtle: 831-1.0098875-turtle.txt
N-Triples: 831-1.0098875-rdf-ntriples.txt
Original Record: 831-1.0098875-source.json
Full Text
831-1.0098875-fulltext.txt
Citation
831-1.0098875.ris

Full Text

APPLICATION OF OPTIMIZATION FOR THE ENZYMATIC SYNTHESIS OF SWEET ASPARTYL DIPEPTIDE ANALOGS CATALYZED BY THERMOLYSIN  by GRACE ILAH LEE B.Sc. (Agr.), The University of British Columbia, 1989  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Food Science)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA September 1992 © Grace Ilah Lee, 1992  In presenting this thesis in partial fulfilment  of  the  requirements for an advanced  degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  Food  Science  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  ABSTRACT  The enzymatic peptide synthesis of two sweet dipeptide analogs (aspartyl derivatives), Z-L-Asp-Tyr-OMe and Z-L-Asp-Met-OMe, was attempted using thermolysin as a catalyst. These two dipeptides are analogs of Z-L-Asp-Phe-OMe, the precursor to aspartame, which has already been successfully synthesized enzymatically using thermolysin. Random Centroid Optimization was employed to obtain optimal conditions for maximum yield. Seven factors were optimized for each peptide and they included: reactant concentration (only the amino component for Z-L-Asp-Tyr-OMe but both carboxyl and amino components for Z-L-Asp-MetOMe), solvent concentration, buffer concentration, enzyme concentration, calcium chloride concentration, pH and temperature. Factor ranges were entered into the Random Centroid Optimization program, which generated the conditions for the experiments. Organic solvents were used in order to shift the equilibrium towards synthesis. A mixture of two solvents, water-miscible dimethylformamide (DMF), and water-immiscible ethyl acetate (EA), were used for both the first and second cycles in the optimization of Z-L-Asp-Tyr-OMe. For Z-LAsp-Met-OMe, glycerol, a polyol, was used in the first cycle but in the second cycle, another polyol, ethanediol was substituted because glycerol appeared to have an inhibitory effect on the enzyme. The maximum yield of Z-L-Asp-Tyr-OMe was 2.13 ± 0.46% (Mean ± S.D., n=3) under the following conditions:  80mM Z-L-Asp, 154mM Tyr-OMe, 2.3/9.7/88  DMF/EA/buffer, 39pM thermolysin, 7mM CaCl2, pH 6.3, at 46°C for 24 hours. The optimum yield of Z-L-Asp-Met-OMe was 9.13 ± 1.22% (Mean ± S.D., n=3) under the following conditions: 53mM Z-L-Asp, 96mM Met-OMe, 2/98 ethanediol/buffer, 35(iM thermolysin, ii  3.8mM CaCl2, pH 5.1, at 49°C for 24 hours. Random Centroid Optimization was shown to be a useful method for attaining the conditions of maximum peptide synthesis, especially with the Z-L-Asp-Met-OMe. However, there were few clear indications of which factors most affected the yield. In general, it was found that the type of solvent used definitely had a great impact on the synthesis of both peptides. Z-L-Asp-Tyr-OMe synthesis was not very successful since the DMF:EA solvents did not sufficiently solubilize the reactants. In addition, high concentrations of solvents which are normally used in other systems could not be used for this synthesis. This indicates that the effect of the solvent on the enzyme must be examined further. It was apparent that not only the solvent had an effect on the synthesis, but that other factors were also important, although the nature of these other factors could not be deduced from this study.  iii  TABLE OF CONTENTS Page  ABSTRACT  ii  TABLE OF CONTENTS  iv  LIST OF TABLES  vii  LIST OF FIGURES  viii  LIST OF APPENDICES  xi  ACKNOWLEDGEMENTS  xii  INTRODUCTION  1  CHAPTER 1. ENZYMATIC SYNTHESIS OF RIBONUCLEASE A  5  CHAPTER 2. ENZYMATIC SYNTHESIS OF SWEET DIPEPTIDES  9  LITERATURE REVIEW  10  A. ENZYMATIC PEPTIDE SYNTHESIS 1. Equilibrium Controlled Synthesis a) Use of Water-Miscible Organic Solvents b) Use of Water-Immiscible Solvents: The Biphasic System . 2. Kinetically Controlled Synthesis 3. Advantages and Disadvantages of Enzymatic Synthesis 4. Enzymatic Synthesis of Polypeptides 5. The Plastein Reaction  10 10 11 15 21 24 25 26  B. ENZYMES 1. Function of Enzymes 2. Enzymatic Synthesis by Thermolysin (Metalloprotease)  28 28 31  C. SWEETENERS 1. Aspartame 2. Enzymatic Synthesis of the Precursor to Aspartame 3. Other Sweet Dipeptides  34 34 35 37  iv  Page  D. RANDOM CENTROID OPTIMIZATION (RCO)  MATERIALS AND METHODS  38  40  MATERIALS  40  METHODS A.  41  B. C. D. E. F.  G. H. I.  2,4,6-TRINTTROBENZENE 1-SULFONIC ACID (TNBS) ASSAY O-PHTHALALDEHYDE (OPA) METHOD THIN-LAYER CHROMATOGRAPHY (TLC) CHEMICAL SYNTHESIS PREPARATIVE HPLC/ PURIFICATION OF SAMPLES ENZYMATIC SYNTHESIS 1. Z-L-Asp-Phe-OMe Synthesis 2. Z-L-Asp-Tyr-OMe Synthesis 3. Z-L-Asp-Met-OMe Synthesis ANALYTICAL HPLC STANDARD CURVES ASSAY FOR THERMOLYSIN  RESULTS AND DISCUSSION  41 41 42 43 44 45 45 46 47 48 49 49  51  A. Z-L-ASP-PHE-OME SYNTHESIS  51  B. Z-L-ASP-TYR-OME SYNTHESIS 1. Random Centroid Optimization: First Cycle 2. Mapping 3. Random Centroid Optimization: Second Cycle 4. Mapping 5. Solvent Suitability for Z-L-Asp-Tyr-OMe Synthesis 6. Synthesis of Z-L-Asp-Tyr-OMe in Different Organic Solvents . . . .  57 57 60 65 66 75 78  C. Z-L-ASP-MET-OME SYNTHESIS 1. Preliminary Work: Solvent Suitability 2. Synthesis of Z-L-Asp-Met-OMe in Different Organic Solvents . . . 3. Random Centroid Optimization: First Cycle  80 80 81 85  v  Page 4. 5. 6. 7.  Mapping Random Centroid Optimization: Second Cycle Mapping Quantification of Z-L-Asp-Met-OMe  88 95 97 102  CONCLUSIONS  106  REFERENCES  110  VI  LIST OF TABLES Page  Table 1. Primary Specificity of Selected Proteases  30  Table 2.  Z-L-Asp-Phe-OMe Synthesis Assayed by TNBS and OPA Methods  52  Table 3.  Synthesis of Z-L-Asp-Phe-OMe as Calculated from HPLC  53  Table 4.  TLC of Amino Acids and Peptides on Silica Gel G plates with fluorescence using 70:29:1 chloroform:methanol:acetic acid  55  Random Centroid Design, Centroid Formulation and Yields for Z-L-Asp-Tyr-OMe Synthesis: First Cycle  58  Random Centroid Design, Centroid Formulation and Yields for Z-L-Asp-Tyr-OMe Synthesis: Second Cycle  67  Table 7.  Solvent Suitability for Z-L-Asp-Tyr-OMe Synthesis  76  Table 8.  Synthesis of Z-L-Asp-Tyr-OMe in the Presence of Organic Solvents . . . .  79  Table 9.  Solvent Suitability for Z-L-Asp-Met-OMe Synthesis  81  Table 5.  Table 6.  Table 10. Synthesis of Z-L-Asp-Met-OMe in the Presence of Organic Solvents . . . .  84  Table 11. Random Design, Centroid Formulation and Yields for Z-L-Asp-Met-OMe Synthesis with Glycerol: First Cycle  86  Table 12. Random Design, Centroid Formulation and Yields for Z-L-Asp-Met-OMe Synthesis in Ethanediol: Second Cycle  96  vii  LIST OF FIGURES Page Figure 1. The dependence of resynthesized ribonuclease A activity in 40% (v/v) dioxane upon the time of incubation with subtilisin BPN'  8  Figure 2. Peptide synthesis catalyzed by serine and thiol proteases  22  Figure 3. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: A) Tyr-OMe concentration and B) DMREA ratio  61  Figure 4. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: C) Buffer and D) Enzyme concentration  62  Figure 5. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: E) Calcium chloride and F) pH . . . .  63  Figure 6. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: G) Temperature  64  Figure 7. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: A) Tyr-OMe concentration and B) DMF/EA ratio  68  Figure 8. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: C) Buffer and D) Enzyme concentration  69  Figure 9. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: E) Calcium chloride and F) pH . . . .  70  viii  Page  Figure 10. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: G) Temperature  71  Figure 11. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: A) Z-L-Asp concentration and B) Met-OMe concentration  89  Figure 12. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: C) Glycerol and D) Enzyme concentration  90  Figure 13. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: E) Calcium chloride and F) pH . .  91  Figure 14. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: G) Temperature  92  Figure 15. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: A) Z-L-Asp concentration and B) Met-OMe concentration  98  Figure 16. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: C) Ethanediol and D) Enzyme concentration  99  Figure 17. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: E) Calcium chloride and F) pH  100  Figure 18. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: G) Temperature  101  IX  Page Figure 19. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Phe-OMe using 80mM Z-L-Asp and 80mM Phe-OMe as the carboxyl and amino components, respectively  121  Figure 20. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Phe-OMe using 80mM Z-L-Asp and 200mM Phe-OMe as the carboxyl and amino components, respectively  122  Figure 21. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Tyr-OMe under optimal conditions from the second cycle of RCO (Vertex 27). . .  123  Figure 22. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe under optimal conditions in glycerol (First cycle of RCO; Vertex 10). .  124  Figure 23. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe of the suspended sample under optimal conditions in ethanediol (Second cycle of RCO; Vertex 32)  125  Figure 24. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe of the freeze dried sample under optimal conditions in ethanediol (Second cycle of RCO; Vertex 32)  126  x  LIST OF APPENDICES  Page  Appendix 1. HPLC Chromatograms of Enzymatically Synthesized Sweet Dipeptides  120  Appendix 2. Glossary  127  xi  ACKNOWLEDGEMENTS  I wish to thank Dr. Shuryo Nakai for his support throughout the course of this thesis. I would like to acknowledge the Biomedical Research Center for the use of the HPLC and laboratory facilities. I also wish to express my sincere appreciation and gratitude to Dr. Ian Clark-Lewis for the use of the HPLC, his assistance in the chemical synthesis and advice. I wish to thank Dr. Brent Skura and Dr. Tim Durance, both of whom served on my committee and offered constructive criticism of my work. Thanks to Phil Owen for his aid in the synthesis of the peptides, Greg Radigan for his technical assistance and especially Peter Borowski, whom I am most grateful to for his help and cooperation.  I wish to thank  Guillermo Arteaga for his computer assistance and helpful comments. Special thanks are extended to Dr. Eunice Li-Chan for her invaluable advice and interest. Finally, thanks to all of my friends, L.F. in particular, for the laughs and their support and encouragement when I needed it.  xii  1  INTRODUCTION Peptide synthesis has been important in the formation and design of biologically active peptides, which in turn has enabled studies of their structure and function. The methodology for peptide synthesis has already been established and the most widely applied method is chemical synthesis. There are more than 140 chemical variations for the formation of peptide bonds (Jakubke, 1987). Chemical synthesis involves the linkage of amino acids or peptide sequences using chemical coupling agents. There are four main steps in chemical synthesis. First, specific side chain groups and also nonparticipating a-amino and a-carboxyl end groups are protected, then chemical coupling agents are used to activate the carboxyl group, allowing peptide bond formation to occur spontaneously. Next, there is deprotection of the groups at certain steps in the synthesis, and a final deprotection is required to obtain the final product. One problem with chemical synthesis is that racemization, the formation of different enantiomers of a product, occurs frequently. In addition, both by-product formation and the solubility of the intermediates affect synthesis. Chemical synthesis is costly on an industrial scale because of its requirement of protected amino acids and chemicals. In the past few decades, there has been interest in the use of proteolytic enzymes as a means to catalyze peptide bond formation because they have potential to be an alternative approach to chemical synthesis. Enzymatic peptide synthesis, also termed peptide ligase activity, is a stereospecific reaction, which means racemization does not occur. Also, side reactions and the formation of by-products is minimized. Enzymatic synthesis involves the selection of a suitable proteolytic enzyme for the formation of the specific peptide bond. The enzyme accelerates the attainment of the equilibrium for formation of the peptide (Jakubke  2 et al., 1985). In order for enzymatic synthesis to occur, conditions must be selected in order to favour the reverse reaction since there is no spontaneous formation of the peptide bond. It is not as dependent on the use of protecting groups as with chemical synthesis, although their use is not entirely eliminated. Thus, the reaction is less costly. Enzymatic peptide synthesis has been demonstrated to be useful not only in the formation of small peptides, but also has potential for synthesis of large peptides, which has been a problem in chemical synthesis. Protein semisynthesis, which is used for large peptides and involves chemical synthesis of limited segments of a polypeptide chain, could use enzymes to assist coupling of these large polypeptide fragments (Chaiken et al., 1982; Rose et al., 1987). Some of the work carried out in genetic engineering could have a large impact on enzymatic peptide synthesis, since there is potential for the engineering of novel enzymes that are able to improve or promote synthesis. Although the use of enzymatic synthesis has been widespread in the area of biochemistry, it has not been applied much in food science. The work that has been done in this area involves the plastein reaction, which uses enzymes to catalyze synthesis between peptide hydrolysates to form complex, insoluble gel-like products.  The application of  enzymes as catalysts for the formation of products is of interest to the food industry since this can lead to the production of novel food ingredients. Since the enzymatic reaction may be carried out under mild conditions as compared to chemical methods, there could be minimization of regulatory hurdles (Gross, 1991). The future goal is to be able to improve synthesis of peptides or to form other products using the enzymatic reaction that would be beneficial to the food industry. One such example is the production of aspartame, a dipeptide  3 sweetener that has been successfully synthesized enzymatically. Much work has been done regarding aspartame production and studies have been completed for its production on a large scale (Nakanishi et al., 1985; Nakanishi et al., 1986; Gross, 1991). The production of other dipeptide sweeteners would be of value to the food industry, especially since aspartame cannot be consumed by individuals who cannot have large amounts of phenylalanine in their diet If these dipeptides could be synthesized in large quantities, there would be some potential application of enzymatic synthesis for the formation of sweeteners. The original objectives of this thesis were to enzymatically synthesize a large peptide, k-casein, using subtilisin as the catalyst for protein semisynthesis of the para-k-casein and glycomacropeptide fragments, to obtain optimal conditions of synthesis and to apply a simple assay method based on the ability of k-casein to stabilize a^-casein from precipitation. Prior to the k-casein system, a well established model of polypeptide synthesis involving the resynthesis of ribonuclease A from ribonuclease S was completed according to Homandberg and Laskowski (1978) (Chapter 1). Since there were difficulties in obtaining resynthesis of ribonuclease S, it was decided that a simpler dipeptide system be used. This was to enable a better understanding of enzymatic peptide synthesis. A simpler dipeptide system using thermolysin to catalyze the formation of the precursor to aspartame, Z-L-Asp-Phe-OMe, was then carried out with the intention of completing the k-casein work later in the project. However, there are difficulties in the resynthesis of k-casein since one of the fragments, para-k-casein, is insoluble. This poses a problem for enzymatic peptide synthesis, since reactants must be soluble. Thus, the  4 k-casein resynthesis plan was dropped. The objectives of this thesis were as follows: two sweet dipeptides, Z-L-Asp-TyrOMe and Z-L-Asp-Met-OMe, which are analogs of Z-L-Asp-Phe-OMe, the precursor of aspartame, were to be enzymatically synthesized using thermolysin and optimal conditions of synthesis were to be developed. Random Centroid Optimization was applied in order to obtain the conditions for synthesis. The reaction was monitored using HPLC.  5  CHAPTER 1. ENZYMATIC SYNTHESIS OF RIBONUCLEASE A  6 ORIGINAL PLAN The original objectives of the thesis were to resynthesize k-casein, a polypeptide which is 169 residues in length, using subtilisin, to obtain optimal conditions for synthesis and to apply a simple assay method based on the ability of k-casein to stabilize oc^-casein from precipitating. Before attempting resynthesis of k-casein, the work by Homandberg and Laskowski (1978) was repeated. They were able to resynthesize ribonuclease A from ribonuclease S using subilisin, and 50% synthesis was attained.  MATERIALS AND METHODS MATERIALS Bovine pancreatic ribonuclease A #R-4875 and S, protease type XXVII (Nagarse-subtilisin BPN'), and cytidine 2',3'-monophosphate (potassium salt) were obtained from Sigma Chemical Co. (St. Louis, MO). Dioxane (ACS assured) and glycerol (ACS assured) were obtained from BDH Canada Ltd. (Missisauga, ON). METHODS Assay Method The reaction was monitored using the cytidine cyclic 2',3'-monophosphate assay which determined ribonuclease activity (Crook et al., 1960). This assay involved measurement of initial velocities of the hydrolysis of cytidine 2',3'-monophosphate, which was proportional to enzyme concentration. The activities of the reaction mixture were compared to a standard solution of ribonuclease A.  7 Enzymatic Synthesis of Ribonuclease A 5 mg of ribonuclease S was dissolved in 0.1 mL water and 0.9 mL glycerol was added slowly. An aliquot (10 |iL) of subtilisin BPN' solution (5 mg/mL stock) was added to the organic cosolvent solution last. Reactions were allowed to proceed at 25°C. The reaction was assayed every two hours until maximum synthesis was obtained. The reaction was monitored using the assay of Crook et al. (1960).  RESULTS AND DISCUSSION The reaction was monitored over 13 days. The yields obtained were not as expected; they were very low compared to those reported by Homandberg and Laskowski (1978). The highest yield obtained was 9.02%, compared to 50% reported yield. Although the yields obtained did not correspond to that of the paper, the general trends in increased synthesis were similar (Figure 1). For instance, the 9.02% yield was obtained on day 9 of the reaction, the same day as the published optimum (50%) occurred. By day 10 of incubation, there was a decrease to 8.75%, which corresponded to the results in the paper since yields dropped off after 9 days. Synthesis of large peptides is always more difficult than of small peptides because a long period is required before the enzyme can recognize the appropriate sites and catalyze the reaction (Rose et al., 1987). Although an organic solvent (glycerol) was used, it did not appear to enhance yields or facilitate equilibrium shift In addition, it is possible that there may have been some negative effect on the enzyme since the yields were low. Some hydrolysis of the product may have occurred, since yields dropped off after 9 days.  8  \  .  /  o <  c <D O O Q.  0  1 3 5 6 8 1011 1315 Time (Days)  Figure 1. The dependence of resynthesized ribonuclease A activity in 40% (v/v) dioxane upon the time of incubation with subtilisin BPN'.  9  CHAPTER 2. ENZYMATIC SYNTHESIS OF SWEET DIPEPTIDES  10 LITERATURE REVIEW  A. ENZYMATIC PEPTIDE SYNTHESIS Enzymatic peptide synthesis is a thermodynamically unfavourable reaction, since it is the reverse of hydrolysis, which is a reaction that releases free energy and is therefore thermodynamically favourable. Under normal physiological conditions, it does not occur by simply reversing the hydrolysis reaction ie. amino acids or peptides in solution will not undergo peptide synthesis in the presence of an enzyme. In order for peptide bond synthesis to occur, there must be ways to reduce or bypass the energetic barriers so that the hydrolysis reaction may be reversed. Three considerations that affect peptide bond formation are: a) the equilibrium constant; b) the ionization constants of selectively protected reactants; and c) the starting concentration of the ionized and nonionized forms of the NHj-terminal protected carboxyl component and the amino component (Jakubke, 1987). Enzymatic peptide synthesis may occur via two methods; as: 1. an equilibrium controlled or thermodynamically controlled process or 2. a kinetically controlled process. It may be further extended to solid phase systems where the enzyme is immobilized onto a support.  1. Equilibrium Controlled Synthesis The equilibrium controlled process, also termed thermodynamically controlled process by Jakubke et al. (1985), refers to the reversal of hydrolysis and is a condensation reaction. There are two equilibria of importance: RCOO + H3N+R' -H. RCOOH + H2NR' **• RCO-NHR' + H 2 0  (Eq. 1)  The first equilibrium is an ionization equilibrium, Kion, and the second is the conversion  11 equilibrium, Kcon. The synthesis equilibrium may be summarized by: K„ = Kim • Kcon = [RCO-NHRl/[RCOO][H3N+Rl  (Eq. 2)  In order for the reaction to favour product formation, the substrates must be uncharged or not ionized because they are not reactive in the ionized form (Jakubke et al., 1985; Jakubke, 1987). Thus, the pK values of the substrates are also important. pH is an important factor in the reaction, since ionization of the substrates is affected by pH. The concentration of uncharged substrates must be sufficiently high enough for product to be formed. Amino acids which are partially blocked/protected at the terminal sites provides reactive, uncharged forms for the ionization equilibrium (Jakubke, 1987). Since the ionization of the reactants is related to the pH of the system, the optimal pH conditions is normally between pH 6 and 7 (Jakubke et al., 1985). However, the pH optimum for synthesis may differ from the pH optimum of the enzyme stability (Martinek and Semenov, 1981a; Fruton, 1982). The two methods by which the equilibrium controlled peptide formation may be affected are: by altering the pK values of the substrates such that K^ increases and using the law of mass action. As already mentioned, ionization equilibria play a major role in dictating peptide bond synthesis. K^n will be increased if the difference in pK values between the amino and carboxyl substrates is decreased. To improve synthesis, either an increase in pKt (the carboxyl group of the carboxyl component) and/or a decrease of pKj (the amino group of the amino component) is required (Kullmann, 1985). There are a number of strategies employed to affect the equilibrium and they are as follows:  a) Use of Water-Miscible Organic Solvents The addition of organic solvents to the reaction mixture causes a shift in ionic  12 equilibria. The equilibrium is shifted towards synthesis rather than the utilization of water since there is reduced water activity (Ingalls et al., 1975; Brink et al., 1988). By lowering the water concentration in the system, the product yield may be increased according to Eq. 1 and hydrolytic side reactions may be suppressed (Martinek and Semenov, 1981a; Clapes and Valencia, 1992). Water-miscible organic solvents and polyhydroxy alcohols cause the dielectric constant to decrease. The major effect that the solvent has is to decrease the acidity of the a-carboxy group of the carboxyl component but it only has a marginal effect on the amino group of the nucleophile (Homandberg et al., 1978). In other words, the pK value of the carboxyl group of the carboxyl component is increased (Jakubke, 1987). Water-miscible organic solvents have been used in systems ranging from 10 to nearly 100% (Oka and Morihara, 1977; Reslow et al., 1988). Malak et al. (1992) used 18-20% dimethylformamide (DMF) for pepsin-catalyzed synthesis, with no loss in pepsin activity. Unfortunately, water-miscible organic solvents generally cannot be used for peptide synthesis at levels higher than 50-70% because of problems with the enzyme (Khmelnitksy et al., 1988). The enzyme may lose its catalytic activity, specificity, and stability, resulting in decreased synthesis or no synthesis at all.  Jakubke et al. (1985) found that ethanol,  dimethylformamide (DMF), dimethylsulfoxide (DMSO), dioxane, acetone and acetonitrile at 50% caused inactivation of a-chymotrypsin. The addition of organic solvents in increasing concentrations may cause activation of enzyme (at the lower levels such as 10-30%), but eventually, the enzymatic activity will decrease (Butler, 1979; Martinek and Semenov, 1981a). Recently, Fernandez et al. (1991) found that a mixture of N-N-dimethylformamide and dimethylsulfoxide caused an initial increase ofKcat (up to 15%) for papain, but there was  13 a decrease when solvent concentration was increased. In another study with papain as the catalyst, Stevenson and Storer (1991) found that methanol promoted product formation for a few minutes but the enzyme quickly became inactivated. There are cases where a watermiscible solvent can be used at high levels, although only when there is a favourable combination of the properties of the enzyme and the solvent (Khmelnitsky et al., 1988). Acetonitrile, ethanol, tetrahydrofuran, acetone and butanediol were used at levels higher than 90% for a-chymotrypsin catalyzed synthesis (Reslow et al., 1988; Clapes et al., 1990a). There are several considerations that must be taken into account when a particular solvent is chosen for an enzymatic reaction, one of which is its compatibility with the reaction (Dordick, 1991). For example, hydrophobic solvents cannot be used for reactions involving sugars because sugars are soluble in hydrophilic solvents (Dordick, 1991). In addition, the products should be compatible with the solvent. A good solvent is one which can reduce the water content in the reaction without loss in enzyme activity (Mozhaev et al., 1989). Another important criterion in choosing the solvent is that it must be able to solubilize the reactants (Carrea, 1984). Reactants that are not soluble in the solvents are unable to react with the enzyme, so synthesis will not occur or will be hindered. If a solvent is able to maintain catalytic stability but is not able to solubilize the reactants, it is less useful than one which can solubilize the reactants although there is a partial effect on the enzyme (Antonini et al., 1981; Stevenson and Storer, 1991). A lot of work has been done to study the effect of organic solvents on enzyme function (Zaks and Russell, 1988; Dordick, 1989). Proteins in their native state have a surrounding hydration shell composed of water molecules attached to the protein surface by  14 hydrogen bonds. It is thought that solvents act to displace the water molecules from the hydration shell, resulting in denaturation of the enzyme since there is an alteration of the protein structure (Arakawa and Goddette, 1985; Khmelnitsky et al., 1988; van Erp et al., 1991). This effect occurs particularly with the highly polar organic solvents, which are able to solubilize water and strip away water from the enzyme that is required for its catalytic structure and function (Dordick, 1991). Inouye et al. (1981) performed the semisynthesis of insulin analogs using dimethylformamide as the solvent and trypsin and found that K^ decreased at DMF concentrations that were greater than 50%, and that there was no synthesis at 70% since there was loss of activity due to denaturation. Zaks and Klibanov (1988a) found that a-chymotrypsin exhibited different activities in all of the hydrophilic solvents tested eg. acetone, dioxane, dimethylformamide, etc. because the essential water was being stripped from the enzyme. Zaks and Klibanov (1988b) demonstrated that enzymatic activity depended on the amount of water bound to the enzyme (the essential layer of water) in the presence of organic solvents. Recently, Gorman and Dordick (1992) showed that waterstripping from an enzyme in the presence of solvents occurs nearly immediately and can be significant in polar solvents. In addition to the water-stripping effect, solvents can alter substrate binding to the enzyme, alter protein structure and flexibility, and affect the enzyme kinetics (Butler, 1979; Dordick, 1991). Since a-chymotrypsin and subtilisin remain rigid in organic solvents and retain their native conformation, they remain active in organic solvents (Zaks and Klibanov, 1988a). The organic solvents that may be used in high concentrations with the least effect on the enzyme are the polyols such as glycerol. They may be used at concentrations exceeding  15 80-90% without affecting the enzyme's catalytic ability. Polyols are able to maintain solvophobic interactions; that is, there is a hydration shell separating the enzyme from the organic phase and so it remains in its native conformation (Khmelnitsky et al., 1988; Gupta, 1992). Polyols also enhance thermostabilization, so reactions can occur at high temperatures and the enzymes will have better stability towards other denaturing conditions (Gupta, 1992). However, polyols are not always the ideal solvent because of their high viscosity and their poor solubilizing ability of substrates that have low polarity. In addition, they are not always successful in promoting synthesis since they can still have an effect on the enzyme (Durrant et al., 1986). Mozhaev et al. (1989) studied the effect of mixed aqueous media on the stability of enzymes and found that polyols could be used at higher concentrations (eg. 65% for 1,4-butanediol) but there was an abrupt decrease in enzyme activity when the concentration was increased only by 3% (up to 68%). Therefore, there is a threshold concentration for organic solvents which causes abrupt changes in the structure and catalytic properties of the enzyme. Polyols may be used with other solvents in order to enhance enzymatic synthesis; for example, Ingalls et al. (1975) used a 1:1 (v/v) ethanolrglycerol mixture and had maximum rates when the water concentration was 5-15%. Inouye et al. (1981) used 70% dimemylformamide/l,4-butanediol (1:1) to synthesize insulin analogs. Kitaguchi and Klibanov (1989) used a combination of tert-amyl alcohol with 9% ethylene glycol to successfully form oligopeptides (tripeptides and longer chains) using thermolysin. b) Use of Water-Immiscible Solvents: The Biphasic System Use of water-immiscible solvents in the reaction system is the other method by which the pK values may be affected and cause Kion to increase. This involves a 2-phase reaction,  16 since there is an aqueous phase and the immiscible organic phase, thus the term "biphasic" system. Extensive work using water-immiscible solvents for enzymatic peptide synthesis has been conducted because of their advantages over the water-miscible ones. The main advantage is that there is minimum contact of the enzyme with the solvent Hydrophobic solvents do not strip the essential water from the enzyme, thus the enzyme is able to maintain its catalytic ability since their native conformation is not affected (Zaks and Klibanov, 1988b; Dordick, 1989). In the biphasic system, the reactants are dissolved in the water-immiscible organic phase and the enzyme in the aqueous phase. The reaction undergoes moderate stirring or shaking to ensure transfer and the reactants diffuse from the organic phase into the aqueous phase where enzymatic peptide synthesis occurs, while the end products diffuse from the aqueous phase back into the organic phase (Carrea, 1984; Morihara, 1987). The reaction rate is dependent on the concentration of the substrate in the aqueous phase rather than on its total concentration in the mixture, which indicates that the catalysis occurs in the aqueous phase (Antonini et al., 1981; Carrea, 1984). Product inhibition, which can be a problem in some systems, is avoided since the product is transferred to the organic phase. Since the enzyme is in the aqueous phase, it is possible to use the biphasic reaction for repeated reactions by removing the organic phase and adding substrate again (Nakanishi and Matsuno, 1988). Kuhl et al. (1980) re-used cc-chymotrypsin for synthesis of Ac-Leu-Phe-Leu-NHj with only a 5% decrease in yield after two re-utilization experiments. Klibanov et al. (1977) proposed the use of the 2-phase system since the free energies of transfer provide a source for equilibrium shift. The amount of water required in a biphasic system is much less than that of a water-miscible system; only the amount necessary to  17 dissolve the enzyme is required. Morihara (1987) stated that the water content is usually 25%. Since water partitions onto the enzyme, water contents as low as 1% in the solvent yields up to 30% water on the protein (Zaks and Klibanov, 1988b). A low water content results in a large equilibrium shift in favour of synthesis. In addition, the use of the waterimmiscible solvents causes an increase in the pK values of the carboxyl components and a decrease in the pK values of the amino components (Jakubke, 1987). Martinek and Semenov (1981b) studied the shift of ionic equilibria in biphasic aqueous-organic systems and concluded that the pK values of the reactants became closer when the water content decreased and synthesis occurred. That is, the ionic equilibria of the reactants were shifted towards their nonionic forms when water content decreased. Martinek et al. (1981) and Eggers et al. (1989) noted that the magnitude of the shift depends on the partition coefficients of the reactants and the volume ratio of the two phases. The resulting product should partition entirely into the organic phase to avoid any effect of product inhibition on the reaction (Lilly, 1982). Martinek et al. (1981) showed that by varying the ratio of the aqueous and organic phases, the equilibrium constant could be shifted and that the yield for N-benzoyl-Lphenylalanine methyl ester was 80% in a biphasic system using chloroform and <xchymotrypsin. The type of solvent used will affect the partition coefficients of the reactants and products. Thus, proper selection of the solvent is required and factors such as compatibility with the reactants and products must be considered. Just as with the water-miscible solvents, the solvent of choice must be able to solubilize the reactants and should not affect the enzyme's catalytic ability (Antonini et al., 1981). In general, the solvents of low polarity (ie.  18 more hydrophobic) have a smaller effect on enzyme stability and thus the enzyme has higher activity (Dordick, 1991). Different proposals have been made to enable selection of organic solvents for use in biocatalytic systems. Laane et al. (1987) correlated enzyme activity with solvent hydrophobicity and had used the log P model as a quantitative measure of solvent polarity. There are some problems with this model since it does not provide a kinetic basis of solvent compatibility to enzymatic catalysis (Dordick, 1991). Hailing (1990) made predictions for solvents based on the equilibrium position of enzyme catalyzedreactionsbut they were only reliable for dilute systems and are based on the partition coefficients. Khmelnitsky et al. (1991) used denaturation capacity as a quantitative criterion for selection of solvents. Gorman and Dordick (1992) suggested that both solvent dielectric constants and a measure of the saturated molar solubility of water in a given solvent provide accurate correlations between the properties of the organic solvents and the extent of T 2 0 (tritiated water) desorption rather than the log P model (log P = the solvent hydrophobicity). Although the use of the biphasic system results in an equilibrium shift, less contact with the enzyme and product separation, there are some disadvantages. It may not be applied to all enzymatic systems since not all substrates are soluble in hydrophobic solvents. The time required for enzymatic synthesis may be longer because there is a low rate of mass transfer across the interface (Jakubke et al., 1985). Kuhl et al. (1987) found that dipeptide formation was greatly enhanced in the biphasic system of buffer and ethyl acetate as compared to a pure aqueous medium, which had insignificant synthesis. However, it took seven days before the reaction equilibrium was attained. The rate of reaction could be altered by varying the enzyme concentration and using higher stirring rates, but excessive stirring can  19 cause denaturation of the enzyme (Carrea, 1984; Khmelnitsky et al., 1988). In addition, even though the solvents are nonpolar, they are still soluble in water to a certain extent and thus affect the enzyme's catalytic properties (Lilly, 1982; Khmelnitsky et al., 1988). The second method to affect equilibrium controlled peptide synthesis is to increase product formation by manipulations based on the law of mass action (Jakubke et al., 1985; Kullmann, 1985). The basis for these actions is to drive the reaction towards peptide synthesis. Factors include:  a) Concentration Dependence The amount of product formed is dependent on the concentration of the substrates, according to the law of mass action. As the concentration of the initial reactants increases, equilibrium yield also increases. Khmelnitsky et al. (1984) found that the yield of N-acetylL-tryptophenyl-L-leucine amide was a function of the concentration of the starting material; the yield increased when the reactant concentration was increased. Yield may be improved with respect to one reactant if the other reactant is added at a higher concentration (Kullmann, 1985; Nakanishi and Matsuno, 1988). In addition, increasing temperature is favourable for synthesis, based on the principle of Le Chatelier, which states the extent of a reaction increases in the direction in which heat is absorbed by the equilibrium reaction system (Kullmann, 1985; Jakubke, 1987).  b) Precipitation or Solubility-Controlled Process The precipitation reaction relies on the use of protected carboxyl and amine  20 components that are used at sufficiendy high concentration so that the solubility of the product is decreased. The reaction requires a relatively high concentration of enzyme. The concentration of the reactants is very important because it determines the concentration of the product (Morihara, 1987). If the peptide concentration exceeds the maximal saturation concentration, the product will precipitate and be removed from solution. Thus, as the solubility decreases, product yield increases. Product formation will continue until the concentration of one of the reactants is so reduced that it equals the solubility of the product (Kullmann, 1985). Oka and Morihara (1978) found that product yield of CBZ-Phe-Leu-NH2 was small when the starting material concentration was less than lOmM but once lOmM was used with the other component at 0.1M, there was considerable yield. Use of one of the reactants in excess is common since the product is usually of lower solubility than the reactants (Jakubke et al., 1985). The precipitation reaction may be enhanced by the addition of salts to the reaction. Isowa and Ichikawa (1979) and Miranda and Tominaga (1991) added ammonium sulfate and sodium chloride salts to increase dipeptide yields since only small amounts of product were obtained initially. Salts act to increase ionic strength, reduce the solubility of hydrophobic products and induce precipitation in the reaction media (Fruton, 1982; Miranda and Tominaga, 1991). The other solubility-controlled reaction was already mentioned, which is the use of the biphasic system where the products are formed in the aqueous layer and transferred to the organic phase. This depends on the volume ratio of the phases and the partition coefficient of the reactants between the two phases (Martinek and Semenov, 1981a). accumulates in the organic phase and can be easily separated.  The product  21 Although the equilibrium controlled process does rely on insoluble product formation to shift the equilibrium it does not mean that it is a requirement for peptide synthesis (Petkov, 1982). Water-soluble peptide formation can occur by using the law of mass action; that is, by using a high concentration of the reactant (Tsuzuki et al., 1980; Morihara, 1987). Often in enzymatic peptide synthesis, not just one of the strategies that have been described is applied; in fact, usually a combination of them are used. In this way, the best conditions are provided for favouring product synthesis. For example, organic solvents are used in the enzyme mixture to shift the ionic equilibrium while a precipitation reaction might occur due to the formation of an insoluble product (Kullmann, 1985; Morihara, 1987).  2. Kinetically Controlled Synthesis The alternate method to the equilibrium controlled peptide synthesis is the kinetically controlled process, also known as the nonequilibrium approach (See Fig.2). The kinetically controlled synthesis is limited to the serine and thiol (cysteine) proteases such as trypsin, achymotrypsin, carboxypeptidase Y, subtilisin BPN' (serine proteases) and papain (thiol protease) and use of acyl amino acid esters as the carboxyl component (Jakubke et al., 1985). The acyl donor to the peptide bond is an amino acid ester, which has the amino group Nacylated, and theresultingacyl-enzyme intermediate is deacylated either by water or an added nucleophile (Gaertner and Puigserver, 1989). The ratio between aminolysis and hydrolysis of the acyl-enzyme determines the acyl transfer efficiency of the enzyme (Schellenberger et al., 1991). If the amino component competes with water for the acyl enzyme, there will be a kinetically controlled transient accumulation of the product which will occur independent  22 of solubility ie. a soluble product can form (Jakubke et al., 1982). The distinguishing feature of the kinetically controlled process is that the products that appear at the highest rate and disappear at the lowest rate will accumulate (Petkov, 1982; Jakubke, 1987). R-CO-NH-R'  +  H-E  H 2 NR'  RCO-X  +  H-E  5 t  ' [ R - C O - X • H-E] — ^ — * R - C O - E HX H20  RCOOH + H-E  Figure 2. Peptide synthesis catalyzed by serine and thiol proteases (H-E) with R-CO-X = carboxyl component, H2N-R'= amino component (Jakubke, 1987).  Unlike equilibrium controlled synthesis, this method allows peptide synthesis to occur in a short time and does not require large amounts of enzyme. The reaction must be stopped before the equilibrium is reached because the concentration of product is a function of time and if the ester substrate is completely used up, there is the possibility of product hydrolysis. The reaction is usually stopped at the "kinetic optimum" (Kullmann, 1985; Jakubke, 1987). The ester donors (X=OR~) are favoured because they form the required acyl-enzyme complex rapidly. There are a number of factors that will affect the rate of formation of product since  23 hydrolysis and aminolysis both involve the same acyl-enzyme complex (Refer to Figure 2). They include the concentration of R"NH2 and the binding constant for the interaction of R'NHj with the acyl-enzyme complex. The ratio of kjk^ must favour transamidation (Fruton, 1982). In order for accumulation of the product to occur, k2 > k3 + k4, and the acylamino acid alkyl esters usually fulfil this condition if they match the substrate specificity of the enzyme (Jakubke, 1987). The nucleophile specificity of the enzyme used in kinetic synthesis is important since this influences 1^, which determines the product. Some substrates are more suitable for the substrate binding sites in the active centres of enzymes. Much work has been done to study the nucleophile specificity for synthesis.  Fastrez and Fersht (1973)  demonstrated the acyl-enzyme mechanism for peptide synthesis using a-chymotrypsin. Morihara and Oka (1977) used Ac-Phe-OEt as the donor and found that the type of nucleophile used was important; amide or hydrazide derivatives had to be used, not free amino acids or ester derivatives. For example, Leu and Leu-OEt as nucleophiles resulted in 0% yield while Leu-NH2 and Phe-NH-NH2 produced yields of 78% and 79%, respectively. Just as with the equilibrium controlled synthesis, a number of factors will affect synthesis using the kinetically controlled method.  Mitin et al. (1984) found that the  kinetically controlled process could be carried out at alkaline pH for papain, which meant that it was possible to synthesize peptides without hydrolysis since the peptidase activity of papain is negligible at pH > 8. Since the nonprotonated form of the amino component (nucleophile) is required in the deacylation process of the acyl-enzyme, the pH for the synthesis should be 8 or higher since the pK values of the a-amino group and peptide derivatives are around 8 (Jakukbe, 1987). Oka and Morihara (1977) synthesized Bz-Arg-Leu-NH2 using trypsin and  24 found the reaction was completed in 5 minutes and that it was better if the pH was higher since pH affects the solubility of the product by making it less soluble. They found that secondary hydrolysis did not occur when there was product precipitation.  Nucleophile  concentration is important because it affects the inhibition of secondary hydrolysis of the products. Thus, a high concentration of nucleophiles is desirable (Oka and Morihara, 1977). Reaction conditions that needed to be optimized for peptide synthesis were the pH, which was dependent on the side chains of the N-terminal amino acid; the ionic strength should be as low as possible, temperature had minimal effect on the reaction and the type of C-terminal protection group used was dependent on the enzyme (Schwarz et al., 1990). As with the equilibrium controlled synthesis, the addition of organic solvents will decrease the amount of water in the system and thus increase yields (Kasche, 1986). Again, the solvents must not have any effect on the enzyme in terms of affecting its structure and function.  3. Advantages and Disadvantages of Enzymatic Synthesis The application of proteases for peptide synthesis has several advantages over chemical methods since the reaction is stereospecific and there is minimization of racemization. Unlike chemical synthesis, protection of side chain functions is unnecessary because of the specificity of the reaction, although protecting groups for the carboxyl and amino groups are still required (Kullmann, 1985; Jakubke, 1987; Nakanishi and Matsuno, 1988). Costs for the enzymatic method are lower than chemical synthesis since protection groups and chemicals are not needed. The reaction can be carried out under mild conditions (eg. ordinary temperatures and pressure) and does not produce unsafe by-products during the formation of peptide bonds.  Also, the catalyst may be reused by employing the biphasic  25 reaction or immobilizing the protease, thus making the enzymatic process economical. Enzymatic peptide synthesis also has several disadvantages. Peptide yields are lower than those obtained by chemical synthesis, since losses occur as a result of hydrolysis and because the reaction is a thermodynamically unfavourable process. Also, the enzymatic process has not yet achieved general applicability Because the reaction is very specific, the proper enzyme must be chosen whereas with chemical synthesis the reagents are applicable to any amino acid residues (Nakanishi and Matsuno, 1988). An enzyme with a broader specificity allows more applications, but there is the problem of nonspecific cleavages in the starting materials and the end product (Jakubke, 1987). This can result in the formation of unwanted by-products. In addition, the reaction usually requires a large amount of enzyme, which may be expensive and also promotes autolysis. Coprecipitation of the enzyme with the end product could cause the reaction to cease (Fruton, 1982).  4. Enzymatic Synthesis of Polypeptides The application of enzymatic peptide synthesis for the formation of large polypeptide fragments is of interest to some researchers. The enzymatic method is advantageous because of the specificity of the reaction and also because it is often difficult to chemically synthesize large polypeptides. A few biologically active peptides have been resynthesized using the enzymatic process, including insulin analogs, cytochrome C, and leucine enkephalin (Inouye et al., 1981; Juillerat and Homandberg, 1981; Kimura et al., 1990). The conditions for synthesis must be chosen in order to favour equilibrium shift and promote the reverse reaction. Organic solvents are often added to the reaction mixture in order to decrease the water content and also to affect K  (Jakubke, 1987).  26 There are problems associated with enzymatic synthesis of the large polypeptides; there may be steric constraints since the peptides are large and have to be oriented so that there is enzyme contact It is sometimes difficult to use the substrates at high enough concentrations, therefore there are difficulties in promoting equilibrium shift. In addition, the time of the reaction is much longer and synthesis can take days to occur (Rose et al., 1987). In order to facilitate enzymatic reaction with large framents, Rose et al. (1987) suggested enzyme-assisted specific carboxyl-terminal activation. This would involve the use of a high concentration of a nucleophile agent in arapidenzyme coupling to a carboxyl component and a subsequent chemical coupling to an amino component This differs from direct enzymatic coupling since activation is rapid, steric constraints are reduced and the enzyme is not present when the fragments are brought together in the subsequent coupling step.  5. The Plastein Reaction The term "plastein" was used by Sawyalow in 1901 to describe the insoluble material that resulted from the addition of proteolytic enzymes to partially digested protein hydrolysates.  The plastein reaction is a protease-catalyzed process that uses protein  hydrolysates or peptide oligomers as the starting materials. It results in the formation of complex, irregular polypeptides or insoluble proteins that may be of high molecular weight (Jakubke, 1987). If intact proteins are used, they are subjected to enzymatic hydrolysis so that there is formation of hydrolysates. These hydrolysates are then used to form plastein, by incubating them with an enzyme (which may be different than the one used for the formation of the hydrolysates). This differs from enzymatic peptide synthesis since the latter produces definite end-products. The plastein reaction is of interest to the food industry  27 because of applications such as the improvement of amino acid composition of food proteins, debittering of hydrolyzed protein and deodorization of soybeans (Fujimaki et al., 1970; Yamashita et al., 1976; Watanabe and Arai, 1988). The mechanism of the plastein reaction is still under investigation because there have been conflicting reports challenging its occurrence.  The disagreement is between two  theories: a) plastein reaction results in an increased molecular weight due to peptide bond reformation; b) plastein synthesis results because of an increased number of hydrophobic interactions. Some authors have proposed that the plastein reaction occurs as a result of transpeptidation and condensation reactions (Arai and Fujimaki, 1978; Gololobov et al., 1986; Watanabe and Arai, 1988) whereas others have rejected the idea of peptide bond synthesis and have suggested that it is only an aggregation process (Sukan and Andrews, 1982; Andrews and Alichanidis, 1990). Methods of examining the plastein products are based on molecular weight including: gel permeation chromatography; trichloroacetic acid (TCA) precipitation; and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Yamashita et al., 1976; Andrews and Alichanidis, 1990).  Arai and Fujimaki (1978)  suggested that when certain enzymes are used, the plastein reaction could form acyl-enzyme intermediates, much like that in the kinetically controlled enzymatic synthesis. After the acylintermediate was formed, the acyl group was transferred to a nucleophile and resulted in peptide bond synthesis. This observation was also supported by Gololobov et al. (1986). However, von Hofsten and Lalasidis (1976) and Edwards and Snipe (1978) both suggested that the plastein products were due to a hydrophobic association, in which non-covalent bonds were important since no material of high molecular weight was formed.  28 Recent papers still have not elucidated the mechanism and the conflict is unresolved. Hartnett and Satterlee (1990) studied plastein gels formed from pepsin-hydrolyzed soy protein isolate and concluded that the product formed was due to an increased number of hydrophobic interactions since there was no peptide bond formation. Lozano and Combes (1991) found that the plastein reaction was an enzymatic process and that substrate concentration was important in driving the reaction towards synthesis. They used gel permeation chromatography to examine the fractions and attributed that enzymatic activity was occurring. Further investigations of the plastein reaction will be required in order to determine its mechanism.  B. ENZYMES 1. Function of Enzymes Enzymes have several functions in enzyme-catalyzed peptide synthesis. Enzymes are required to increase the rate of the reaction. More importantly, there is specific binding of the substrate molecules at the active site so that they are in a close distance to the catalytic groups and in a favourable orientation for reaction (Fruton, 1982; Jakubke, 1987). Thus, the substrate specificity of the enzyme is very important since it will dictate the type of substrates that will bind and ultimately form the peptide bond. This criterion is the reason why enzymatic peptide synthesis is specific and why the proper selection of the enzyme is so vital. The information about enzyme substrate specificity is essential because the same structural and electronic features apply to the synthesis reaction as the hydrolysis reaction (Jakubke,  29 1987). The enzymes used for peptide bond formation are proteases or proteinases (endopeptidases, which hydrolyze peptide bonds in the interior of a polypeptide chain) but peptidases (exopeptidases, which catalyze hydrolysis of terminal amino acid residues from the peptide) can not be used (Morihara, 1987). The proteases may be classified into four groups; the serine, thiol (cysteine), acid (or aspartic), and metalloproteases.  This  classification shows the differences in the participation of the active-site groups (Jakubke, 1987). In addition, most of the proteases within a classification share similar properties such as optimum pH, inhibitors, substrates, species of enzyme producer and also effect of secondary interaction (Morihara, 1974). The specificity of proteases may be described as "primary" and "secondary". Primary specificity is characterized by the amino acid in the Pj position while secondary specificity includes the additional structural elements near the sensitive bond (Jakubke, 1987). Schechter and Berger (1967) used a notation to characterize substrate specificity, where H2N —P3-P2Pj- Pi'-P2'-P3'— COOH denotes the amino acid residues flanking the sensitive bond, and —S3S2-S1-S1/-S2/-S3'— denote the complimentary subsites on the enzyme. The peptide substrate (P=amino acid residue) is split between subsite amino acid residues Sx and S/ of the active site, which results in hydrolysis of the peptide bond between Px and P,' (Jakubke, 1987). Table 1 shows the primary specificity for proteases from all four groups. Papain differs from the other proteases in that it requires hydrophobic residues such as Phe, Val, Leu in the P2 position, and small residues are favoured in the Pj position (Jakubke, 1987). Fruton (1982) noted that knowledge about the secondary specificity of proteases would be helpful in their  30  Table 1. Primary Specificity of Selected Proteases" Enzyme  Preferred Cleavage Site  Serine Proteases Trypsin  -Arg(Lys)i  Achromobacter protease  -Lys^  Chymotrypsin, subtilisins  -Trp(TyrJ,he,Leu,Met)i  Elastases, a-lytic protease  -Ala(Ser,Met,Phe)i'  Staphylococcus aureus V8 protease  -Glu(Asp)i  Thiol (cysteine) Proteases Papainb, ficin, streptococcal protease Clostripain, cathepsin B Cathepsin C  -Phe(Val,Leu)-xi  -Argf H-X-Phe(Tyr,Arg)i  Metalloproteases Thermolysin, B. subtilis neutral protease  -Phe(Gly^euKLeu(Phe)-  Acid (carboxyl) Proteases Pepsin, penicillopepsin, chymosin  -Phe(TyrJJeu)iTrp(Phe,Tyr)-  •References: Fruton (1982), Jakubke, (1987) and Morihara, (1987) •Tor papain, X may be a small uncharged residue (Gly, Ala, etc.)  31 use for peptide synthesis; however, little information exists. More studies on enzyme structure would be required in the future. Usually, proteases catalyze peptide bond synthesis of L-amino acids, but not D-amino acids because of their L-specificity (Margolin et al., 1987). Only in certain cases can Damino acids be incorporated into a peptide via enzymatic synthesis. There is interest in applying enzymatic peptide synthesis for the formation of peptides with D-amino acids, as many of the biologically important peptides contain D-amino acids eg. antibiotic peptides. The enzymes which are able to catalyze these reactions are the serine and cysteine proteases such as a-chymotrypsin and trypsin.  Organic solvents could be used to alter the  stereospecificity of enzymes such as subtilisin and a-chymotrypsin so that they could catalyze reactions involving D-amino acids (Zaks and Klibanov, 1986; Margolin et al., 1987; Sakurai et al., 1988).  2. Enzymatic Synthesis by Thermolysin (Metalloprotease) Thermolysin, isolated from Bacillus thermoproteolyticus, is a thermostable enzyme of molecular weight 37,500, and is classified as a metalloprotease. It requires the participation of a zinc atom at the active site for its catalytic function and it also requires 4 atoms of calcium for its stability (Latt et al., 1969; Reddy, 1991). It is a neutral protease which is most active at neutral pH and is sensitive to metal chelating agents such as ethylenediaminetetraacetic acid (EDTA) and 1,10-phenanthroline (Morihara, 1974). Other compounds that exhibit inhibitory effects are oxalate, citrate and phosphate (Matsubara, 1970). The optimum pH for thermolysin hydrolytic activity is between 5 and 9 (Oyama and Kihara, 1984). Thermolysin cleaves peptide bonds involving amino groups of hydrophobic  32 amino acids with bulky side chains such as isoleucine, leucine, valine, phenylalanine, methionine and alanine (Matsubara, 1970). Other amino acids found to be susceptible only to a minor extent were tyrosine, glycine, threonine and serine. The rate of cleavage may be enhanced if there is also the presence of a hydrophobic amino acid in the donor of the carbonyl group of the sensitive bond (Matsubara, 1970). Although the application of thermolysin for enzymatic peptide synthesis has not been as extensive as that for a-chymotrypsin, it has been proven as a useful protease for peptide synthesis. The studies done showed that synthesis by thermolysin reflected the specificity requirements for hydrolysis. There was a strong preference for a hydrophobic P / residue (eg. Phe, Leu, lie) with a relative preference for a hydrophobic residue in the P! position (Fruton, 1982).  Isowa and Ichikawa (1979) noted that hydrophobic amino acids as the amine  component were most suitable and also found the unsuitable carboxyl components to be Val, lieu, Pro and Trp, which is in agreement with the hydrolysis studies. Suitable carboxyl components were Z-derivatives of all hydrophobic amino acids and other Z-amino acids. Oka and Morihara (1980) showed that the coupling efficiencies for the amino acid residues on the amino side of the coupling point for synthesis with Cbz-Phe-OH (carboxyl component) to be in the following order: Leu-NH2, Phe-NH2, De-NH2> Val-NH2 > Ala-NH2 > Tyr-NH2, GlyNH2, Pro-NH2, D-Leu-NH2. In addition, the carboxyl component was also of importance and reflected the specificity of hydrolysis.  Cbz-Gly-OH was a poor carboxyl component  compared to Cbz-Ala-OH and Cbz-Phe-OH. The use of different protecting groups for the carboxyl component affected results. Miranda et al. (1986) found that peptides protected with Z and Moz groups for L-asparagine gave higher yields for the formation than those protected  33 with Boc groups eg. Boc-Asn-Ile-OBzl, Z-Asn-Ile-OBzl (70 and 87% yield, respectively). The enzymatic synthesis conditions usually used were 0.1M reactants, lmg/mL enzyme, pH 6.5-8 and incubation varying from 3 and 24 hours at 30-50°C (Reddy, 1991). Sakina et al. (1988) used 50mM carboxyl component, 50fiM thermolysin and 250mM amino component since the reaction was better when the amino component was in excess. They also used organic solvents to shift the equilibrium towards peptide bond formation but found that 70% aqueous 1,4-butanediol and 70% aqueous methanol were unsatisfactory. Yields were better when 90% DMF-ethanol (1:1) were used. The addition of salts to the reaction medium has been useful in enhancing yield when there was difficulty in precipitating the product. Salts, by reducing the solubility of the products and inducing their precipitation, help to increase yield (Miranda and Tominaga, 1991). Isowa and Ichikawa (1979) used 20% ammonium sulfate or sodium chloride to increase yields of dipeptides. Similarly, Cheng et al. (1988) found that 20-25% ammonium sulfate was required to increase the yield of Z-ProLeu-OEt since it was partially soluble in the reaction medium. Recently, thermolysin has been modified by attaching polyethylene glycol to the amino groups, which resulted in a change in the substrate specificity to hydrophilic as well as acidic amino acids (Ferjancic et al., 1988). Information obtained from the mechanism of the hydrolysis reaction is relevant in understanding the mechanism of peptide synthesis. Thermolysin resembles pepsin and related aspartyl proteases in its general mechanism and primary specificity (Fruton, 1984). However, its secondary specificity is much different from pepsin.  Initially, it was thought that  thermolysin-catalyzed reactions proceed via the formation of covalent acyl and amino  34 intermediates (Morihara et al., 1978). This proposal has since been rejected because other researchers have found it to be a base-catalyzed pathway (Reddy, 1991). Oyama et al. (1981) studied the mechanism of thermdlysin and showed that it involved the reverse of a general base-catalyzed hydrolysis of a peptide bond, which was proposed by Kester and Matthews (1977). Oyama et al. (1981) suggested that there was an ordered binding mechanism where the donor binds first.  Wayne and Fruton (1983) proposed a rapid-equilibrium random  bireactant mechanism where both substrates (donor and acceptor) bind in random order before the condensation. Fruton (1984) suggested that thermolysin's mechanism was similar to that of pepsin, where there would be coupled interaction of products with the active site and peptide synthesis via condensation reactions. Riechmann and Kasche (1986) also rejected the idea of the covalent enzyme-substrate intermediate. Their studies were similar to those of Wayne and Fruton (1983), which supported the rapid-equilibrium random bireactant mechanism. All of the studies required the analysis of the kinetics of the reaction, and used indices which defined the specificity of proteases in hydrolysis, as well as synthesis. The indices were: km (which is substrate concentration at half-maximal velocity) and kcat (maximal initial velocity per unit enzyme concentration) for different substrates (Fruton, 1982).  C. SWEETENERS 1. Aspartame Much work has been done in the area of sweeteners in order to find sweet-tasting compounds other than sucrose. The studies completed in the area of peptide sweeteners have mainly focused on the dipeptide aspartame, which is L-aspartyl-L-phenylalanine methyl ester.  35 Aspartame is 180 times sweeter than sucrose (Janusz, 1989). It has achieved commercial significance because its taste is very similar to that of sucrose and it has many food applications. Aspartame has found wide usage in the chewing gum, breath mint and softdrink industries and has been used in products such as Jello® (gelatin and pudding mixes), and in dry beverage mixes such as instant coffees, teas and cocoa mixes. However, it lacks hydrolytic stability and cannot be used in acidic food and beverage products (Kim and Dubois, 1991). Intensive studies have been conducted to ensure aspartame's stability, and it has been approved by the Food and Drug Administration (FDA), which allows an acceptable daily intake (ADI) of 50mg/kg body weight (Kim and Dubois, 1991). Aspartame is metabolized by the body as a protein, but because of its intense sweetness, the amounts used in foods are very small. Therefore, it is suitable for people on low-calorie diets. Since it gets broken down to the two amino acids, people suffering from phenylketonuria, who cannot metabolize phenylalanine, should not use aspartame (Kim and Dubois, 1991).  2. Enzymatic Synthesis of the Precursor to Aspartame Aspartame has been synthesized chemically but there have been problems such as the production of a mixture of a and P-L-Asp-L-Phe-OMe, which required separation (Yang and Su, 1986). Thus, enzymatic peptide synthesis methods have been suggested as an alternative to chemical synthesis by a number of researchers. Enzymatic synthesis of the precursor to aspartame, Z-L-Asp-Phe-OMe, was achieved by a number of researchers. Isowa et al. (1979) successfully synthesized Z-L-Asp-Phe-OMe using thermolysin and the yield was 96%. A series of other papers followed.  Oyama et al. (1981) used the thermolysin-catalyzed  condensation reaction of Z-L-Asp with L-Phe-OMe to elucidate the mechanism of the action  36 of thermolysin and concluded that it was the reverse of base-catalyzed hydrolysis. Petkov and Stoineva (1984) suggested the use of a nucleophile pool for synthesis of Z-L-Asp-PheOMe, where the nucleophile (Phe-OMe) was dissolved with thermolysin and then Z-L-Asp was added. They obtained 93% yield by separating the product and re-using the nucleophile pool as an alternate method to immobilization.  Nakanishi et al. (1985) successfully  synthesized Z-L-Asp-Phe-OMe by a continuous method, using immobilized thermolysin in an organic solvent, ethyl acetate. They used a continuous stir tank reactor (CSTR) and a plug flow type reactor (PFR) and found that the CSTR was more suitable for long-term stability and ease of operation. Ooshima et al. (1985) were able to synthesize Z-L-Asp-Phe-OMe in 98% organic medium using a mixture of ethyl acetate:benzene:methanol:water (50:29:19:2), as a homogeneous solvent, not a 2-phase reaction. Nakanishi et al. (1986) synthesized Z-LAsp-Phe-OMe in both an aqueous and biphasic system and found that the reaction was more efficient when there was an excess of Phe-OMe. Oyama et al. (1987) used immobilized thennoase, a crude preparation of thermolysin, to synthesize Z-L-Asp-Phe-OMe using a continuous column operation and also a batchwise operation. They found that the continuous column operation was more advantageous than the batchwise operation since deactivation occurred more rapidly with the batchwise method. Chen and Wang (1988) used papain in a biphasic reaction to synthesize the precursor to aspartame and found the yield to be 75% for the reaction involving Moz-L-Asp and L-Phe-OMe. Although all of these studies have shown that enzymatic synthesis has potential for the production of aspartame precursor, the method has not yet been applied industrially because the enzyme is very expensive (Yang and Su, 1985; Gross, 1991).  37  3. Other Sweet Dipeptides Aso (1989) synthesized N-acetyl-L-phenylalanine-L-lysine, a dipeptide 20 times sweeter than sucrose, using a-chymotrypsin and kinetically controlled synthesis and obtained a yield of 75%. Another enzymatic synthesis completed by Aso et al. (1988) was not a sweet peptide but a L-glutamic acid oligomer, which has a brothy (unami) taste. A yield of 80% was obtained for the oligo-L-glutamic acid using papain. There are two analogs of aspartame that are also similar in sweetness potency which involve the replacement of the L-phenylalanine. They are L-aspartyl-L-tyrosine methyl ester and L-aspartyl-L-methionine methyl ester (Mazur et al., 1969). They both are 180 times sweeter than sucrose (Janusz, 1989). Little information is available on these dipeptides and there has been no work on their enzymatic synthesis. These two dipeptides would be of interest if they could be enzymatically synthesized, since there is the possibility that they could serve as alternate dipeptide sweeteners to aspartame. The methionine methyl ester sweetener would be of particular interest since methionine is an essential amino acid, so it would be a nutritive sweetener. Both L-tyrosine methyl ester and L-methionine methyl ester have the potential to be enzymatically synthesized with Z-L-Asp using the same enzyme, thermolysin, as with the L-phenylalanine.  This would result in the formation of the  precursors to L-Asp-L-Tyr-OMe and L-Asp-L-Met-OMe, since thermolysin requires the presence of Z-blocked carboxyl components.  L-tyrosine is structurally similar to L-  phenylalanine, except for the presence of an -OH group and both L-tyrosine and L-methionine have been noted to be suitable substrates for thermolysin, according to the hydrolysis studies (Matsubara, 1970; Fruton, 1982).  38 D. RANDOM CENTROID OPTIMIZATION (RCO) Random centroid optimization is a computer-aided optimization technique which may be applied as a tool to enable prediction of conditions for experiments instead of using trial and error. It is used in order to decrease the amount of time to carry out experiments and decrease the number of experiments required, therefore improving efficiency. The method is flexible, and allows the use of up to 20 different factors, without requiring a significant increase in the number of experiments. The advantages are: a) that there is high search efficiency; b) both its theory and computer operation are simple; c) boundary constraints that may cause search stalling are not needed; d) there is less chance of merely reaching a local optimum instead of the global optimum; e) unexpected conditions may be discovered by chance (Nakai, 1990). The random centroid optimization method is based on the use of random search and centroid search. The limits or constraints for each factor are set and entered into the program. The program then completes a random search, in which the conditions for the experiments are chosen randomly according to the limits for the factors (Nakai, 1990). Once the first cycle of experiments are completed, the conditions for other cycles are again selected randomly, but within a narrower range. The main advantage of the program is that it allows for much flexibility. The limits for the conditions may be narrowed in order to improve the optimization. A separate feature in the program is used when the number of factors is larger than 8. The centroid search is used since it helps avoid inefficiency of random search (Nakai, 1990). Once the centroid search is completed and results are reported to the computer, the  39 mapping process may be executed. Mapping may be defined as an approximation of the response surface (Nakai, 1990). A series of curves are fit to the data points and are plotted against the factors. Based on these maps, new limits may be chosen or narrowed for the second cycle in order to focus the direction towards the optimum. After the second cycle is complete, the optimization may be terminated depending on the results or may continue to the step called simultaneous shift. Simultaneous shift is carried out when the directions of all factors are well defined from mapping, otherwise the cycle is repeated (Nakai, 1984). It involves the continuation of experiments until the response value becomes worse than the preceding one. Usually, the optimization is terminated when the response value has only slight fluctuations or when the same response value is repeated 3 times (Nakai, 1984).  40  MATERIALS AND METHODS  MATERIALS L-aspartic acid, N-CBZ-L-aspartic acid (Z-L-Asp), L-tyrosine, L-leucine, L-phenylalanine methyl ester hydrochloride (Phe-OMe), L-tyrosine methyl ester (Tyr-OMe), L-methionine methyl ester hydrochloride (Met-OMe), L-aspartyl phenylalanine methyl ester, thermolysin (Protease Type X), o-phthaldialdehyde (OPA), N-CBZ-L-asp-P-t-butyl ester, N-acetyl-Lcysteine, calcium chloride, triethyleneglycoldimethylether (triglyme), 5% picrylsulfonic acid solution (2,4,6 trinitrobenzenesulfonic acid) and MES buffer (2-[N-morpholino]-ethanesulfonic acid) were obtained from Sigma Chemical Co. (St. Louis, MO). Silica Gel G60 plates, 0.2mm layer thick with fluorescent indicator F-254 and aluminum backed (20 x 20 cm) and Silica Gel 60 plates without fluorescence indicator (20 x 20 cm) were from BDH Canada (Toronto, ON). Trichloroacetic acid (TCA), sodium acetate, Hammarsten casein, and sodium sulfite were obtained from BDH Canada (Toronto, ON). Sodium borate and ethylene glycol (ethanediol) were obtained from Fisher Scientific (Vancouver, BC).  0.5M N-dicyclo-  hexylcarbodiimide dichloromethane (DCC) was from Applied Biosystems (Foster City, CA). Dichloromethane (DCM) and trifluoroacetic acid (TFA) were obtained from Baxter Canlab (Missisauga, ON). Acetonitrile (OmniSolv), methanol (OmniSolv), chloroform (OmniSolv), acetic acid, butan-1-ol (Assured, ACS), 1,2-dichloroethane  (Assured, ACS), N,N-  dimethylformamide (DMF) (Assured, ACS), glycerol (Assured, ACS) and toluene (OmniSolv) were from BDH Canada (Toronto, ON). Ethyl acetate (HPLC grade) was obtained from Caledon Labs Ltd. (Georgetown, ON). N-methyl-pyrrolidone was obtained from Baxter  41 Canlab (Missisauga, ON). AG1-X8 analytical grade, styrene type, quaternary ammonium ion exchange resin was from Biorad Laboratories (Canada) Ltd. (Missisauga, ON).  METHODS A. 2,4,6-TRINITROBENZENE 1-SULFONIC ACID (TNBS) ASSAY The method of Kwan et al. (1983) was used with a few modifications. The solution to be assayed was first diluted 100X, then 0.1 mL was mixed with 2.0 mL of 0.2M sodium borate buffer, pH 9.2 and 1.0 mL of 4.0mM TNBS. The solution was incubated for 30 minutes at room temperature and then 1.0 mL of 2.0M NaHjPO,, containing 18mM NajSOg was added. A calibration curve was made from a 2mM stock solution of L-leucine, with saturated ethyl acetate as the blank. Absorbance at 420 nm was measured with a Shimadzu UV-Vis Recording Spectrophotometer, UV-160 (Kyoto, Japan). The % synthesis was calculated based on the loss of amino groups: % synthesis =  Conc^ - Conc^  B. O-PHTHALALDEHYDE (OPA) METHOD The method by Hernandez et al. (1990) was used with the following modifications: 5 x 10~2M solutions of o-phthalaldehyde in ethanol and N-acetyl-L-cysteine (NAC) in water were prepared and stored in the refrigerator. The boric acid-borate buffer, pH 9.5 was prepared by dissolving 6.18 g of orthoboric acid and 2.8 g of NaOH in 1 L of water. The  42 OPA-NAC-buffer reagent was prepared fresh before every reaction by mixing 25 mL of OPA and 25 mL of NAC with 200 mL of the boric acid-borate buffer. NAC-buffer reagent (3 mL) was prepared as directed but the sample was diluted 20X first, then a 10 fxL aliquot of the sample was mixed with the reagent. The mixture was vortexed and the absorbance at 335 nm was measured against buffer/solvent (as used in the resynthesis reaction) using a Shimadzu UV-Vis Recording Spectrophotometer, UV-160 (Kyoto, Japan).  A calibration curve was made from a 7 x 10~3M stock solution of  phenylalanine methyl ester hydrochloride using a range of 5-50 ^L. The % synthesis was calculated based on the loss of amino groups: % synthesis =  Conc^ - C o n c ^  C. THIN-LAYER CHROMATOGRAPHY (TLC) Thin-layer chromatography (TLC) was carried out on 0.2 mm-thick silica gel G-60 plates without fluorescent indicator using the following solvent systems for development: a) chloroform:methanol (95:5) and b) butanol:acetic acid:water (4:1:1). Silica gel G-60 plates, 0.2 mm-thick, with fluorescent indicator were also used with the following solvent systems: a) ethyl acetate:heptane:acetic acid (6:3:1); b) ethyl acetate:heptane:acetic acid (6:3:3); c) ethyl acetate:heptane:acetic acid (6:3:6); d) methanol:water (99:1); e) chloroform: methanohacetic acid (85:10:5); f) chloroform:methanol:acetie acid (70:20:10); g) chloroform: methanohwater (70:20:10); h) cMoroform:methanol:triethylamine (70:29:1); i) chloroform: methanohacetic acid (70:29:1).  43 Standard compounds of Z-L-Asp, Phe-OMe-HCl, Z-L-Asp-Phe-OMe were dissolved in the same solvents used in the mobile phase. 3 to 5 uL of each compound were spotted onto the plates. Plates were visualized using a UV lamp. Rj values were calculated for Z-LAsp, Phe-OMe-HCl and Z-L-Asp-Phe-OMe.  D. CHEMICAL SYNTHESIS Z-L-Asp-Phe-OMe, Z-L-Asp-Tyr-OMe and Z-L-Asp-Met-OMe standards were chemically synthesized using the dicyclohexylcarixxliimide (DCC) mediated coupling reaction to form symmetrical anhydrides (Stewart and Young, 1984). Phe-OMe-HCl and MetOMe-HCl, were treated with AG1-X8 Styrene Type, Quaternary Ammonium Resin (BioRad Laboratories Ltd., Missisauga, ON) prior to reaction to remove the hydrochloride group and replace it with ammonium ion. 3M Phe-OMe-HCl (or Met-OMe-HCl) was dissolved in a few mLs of dimethylformamide (DMF). 8 g of AG1-X8 resin was washed several times with DMF, then mixed with the Phe-OMe-HCl (or Met-OMe-HCl) and left to react for 15 minutes. 2M CBZ-L-Asp-P-t-butyl ester was dissolved in approximately 5 mL dichloromethane (DCM), then 2 mL of 0.5M N-dicyclohexylcarbodiimide dichloromethane (DCC) was added, resulting in a white precipitate, which was dicyclohexylurea (DCU). This mixture was filtered using a Buchner filtering funnel with coarse fritted disk (Fisher Scientific, Vancouver, BC) into a flask to obtain the filtrate (the symmetrical anhydride), and the Phe-OMe or MetOMe (after being treated as above) was filtered into the same flask. For the Z-L-Asp-TyrOMe synthesis, the Tyr-OMe was dissolved in approximately 5 mL of dimethylformamide (DMF) and added to the filtered CBZ-L-Asp-pVt-butyl ester solution. The flask was heated  44 at 50°C and allowed to react for 30 minutes. The reaction was monitored using a ninhydrin reaction (Sarin et al., 1981). The sample was roto-evaporated to remove excess solvent, then 50-70% trifluoroacetic acid (TFA) was added, and allowed to react for 45 minutes at 50°C. Roto-evaporation was repeated to remove excess TFA. Samples from each step (eg. before TFA addition and after TFA addition) were taken and analyzed on an analytical column, a 4.6 x 250 mm C18 Vydac column with 5 fim particle, 300-A pore size packing (Separations Group, Hesperia, CA) with a precolumn guard using a 0-60% acetonitrile/water gradient in 0.1% TFA, and flow rate of 0.7 mL/min. Absorbance was detected by a Beckman System Gold with Diode Array Detector Module 168 at 214 nm. The sample was then frozen with liquid nitrogen and lyophilized overnight. After lyophilization, water (approximately 20 mL) was added until there was separation of two phases. The sample was then purified using HPLC.  E. PREPARATIVE HPLC/ PURIFICATION OF SAMPLES All three dipeptides were purified in the same manner. Samples were filtered using a 0.8 micron Nalgene sterilization filter unit (Nalge Co., Rochester, NY) prior to application to the column. A 10 mm x 25 cm Cjg Vydac column, 5 Jim particle size, 300 A pore size packing (Separations Group, Hesperia, CA) was used. 10 mL of 6M guanidine hydrochloride in 0.1 M acetic acid, pH 3, was applied to the column first, followed by the filtered sample and another 10 mL of 6M guanidine hydrochloride and acetic acid to ensure that no precipitation of the sample would occur on the column. Peptides were eluted with a 0-60% acetonitrile/water gradient in 0.1% TFA at a flow rate of 1.1 mL/min over 2 hours. The  45 absorbance was detected by a Spectroflow 757 detector at 225 nm. The peaks obtained were compared to the crude peak from the chemical synthesis using the analytical column and then fractions with the major peaks were collected, pooled and lyophilized.  F. ENZYMATIC SYNTHESIS  1. Z-L-Asp-Phe-OMe Synthesis Z-L-Asp-Phe-OMe was synthesized according to the method of Nakanishi et al. (1986) with the following changes: Ethyl acetate was saturated with water instead of buffer. Saturated ethyl acetate was prepared as follows: water was added to a separatory funnel, then ethyl acetate was added until two layers were formed. The mixture was shaken, and the bottom layer (the saturated layer) was decanted and used. 80mM Z-L-Asp was dissolved first in the saturated organic solvent, then dry Phe-OMeHCl (80 or 160mM) was added along with the 5mM CaClz The mixture became cloudy but once the pH was adjusted to 6 using 0.2M NaOH, it became clear. A 10 mg/mL stock of thermolysin in water was prepared and a 0.1 mL aliquot was used for each reaction. 0.1 mL thermolysin was pipetted into a tube, then 0.9 mL of the reactants in saturated ethyl acetate was slowly pipetted in. The tubes were then incubated at 40°C for 5 hours in a shaking water bath (Julabo SW-20C reciprocal water bath at 200 r.p.m., Baxter Canlab, Missisauga, ON). The enzymatic reaction was stopped by the addition of 5-10 |iL of 100% trifluoroacetic acid (TFA). The tubes with enzyme were centrifuged (12 x 1000 min"1 for 10 minutes using an Eppendorf Centrifuge 5415C, Brinkmann Instruments Canada Ltd., Rexdale, ON) in order to obtain the precipitate. The  46 supernatent was decanted and the precipitate was redissolved in a mixture of water and methanol (80%). The blank was diluted 10X in water before being analyzed on the HPLC.  2. Z-L-Asp-Tyr-OMe Synthesis The total volume for each reaction was 1.0 mL, as with the Z-L-Asp-Phe-OMe synthesis. Large stock solutions were not made up for the reactants due to the different conditions required for each individual vertex (experiment). Instead, the Z-L-Asp and TyrOMe were weighed out and made to a volume of 2 mL at the appropriate concentration. This "stock" was used for both the blank and the enzyme. Z-L-Asp was first dissolved in the solvents, along with the CaCl2, and then the Tyr-OMe was added. This was because there was better dissolution of the Z-L-Asp when it was dissolved in the solvents first The amount of enzyme required was weighed on a microbalance and dissolved separately in the buffer solution (either 0.2M sodium acetate or 0.05M Mes buffer was added depending on the pH used: sodium acetate was used for pH 5 - 5.5 while Mes was used for pH 5.6 to 7.0) at the appropriate pH and volume. The pH of the stock solution was checked before it was used in the reaction. Both of the buffers maintained the appropriate pH. The required amount of reactants was pipetted slowly into the tube containing the dissolved enzyme. This method was modified for those experiments involving a small volume of solvent eg. 0.06 mL solvent and 0.94 mL buffer. Instead, the reactants were dissolved in the solvents as described above but were also dissolved in the appropriate amount of buffer. The dry enzyme was dissolved in 0.1 mL buffer and then a 0.9 mL aliquot of the reactants was pipetted in slowly. The tubes were incubated for 24 hours (or 5 hours as in the preliminary work) in a shaking water  47 bath (Julabo SW-20C reciprocal water bath at 200 r.p.m., Baxter Canlab, Missisauga, ON) at the appropriate temperature. After incubation, 150 |iL aliquots were removed from the tubes with the enzyme, and 3 - 4 drops of 100% TFA were added to stop the reaction. Samples were also removed from the blanks (tubes with reactants and no enzyme) for analysis. Both the blank and sample were diluted 10X in water before being analyzed on the HPLC. All of the vertices for the optimization were completed once, but the optimum vertex was repeated in duplicate.  3. Z-L-Asp-Met-OMe Synthesis Z-L-Asp-Met-OMe synthesis was similar to that of Z-L-Asp-Tyr-OMe. However, the reactants were measured together in the same tube, along with the CaCl2, and were dissolved with the appropriate volume of solvent. Once again, large stock solutions were not made up due to the different conditions required for each individual vertex. Instead, the Z-L-Asp and Met-OMe were weighed out and made to a volume of 2 mL at the appropriate concentration. This "stock" was used for both the blank and the enzyme tube. The enzyme was weighed out on the microbalance and dissolved in the appropriate volume of buffer (either 0.2M sodium acetate or 0.05M Mes buffer was added depending on the pH used: sodium acetate was used for pH 5 - 5.5 while Mes was used for pH 5.6 to 7.0) at the appropriate pH and volume. The pH of the stock solution was checked before it was used in the reaction. The appropriate volume of reactants was slowly pipetted into the enzyme solution. Again, if the volume of solvent was not sufficient enough to dissolve the reactants, buffer was added to dissolve them and the enzyme was dissolved in 0.1 mL of buffer while a 0.9 mL aliquot of  48 the reactants was pipetted in. After the 24 hour incubation in a shaking water bath (Julabo SW-20C reciprocal water bath at 200 r.p.m., Baxter Canlab, Missisauga, ON), 200 (XL aliquots were removed from the enzyme tubes and 4 - 5 drops of 100% TFA were added to stop the reaction. The samples were analyzed three ways: a) samples were vortexed and diluted 10X; b) samples that appeared to have particulate matter were centrifuged at 12 x 1000 min"1 for 5 minutes using an Eppendorf Centrifuge 5415C (Brinkmann Instruments Canada Ltd., Rexdale, ON) to obtain the precipitate, which was redissolved in 150 \iL of 100% methanol and analyzed; c) a 200 \iL aliquot taken directly from the enzyme tube was lyophilized as is (no separation of supernatant or precipitate), resuspended in 150 nL of 100% methanol and analyzed. Samples were removed from the blanks and diluted 10X in water. All of the vertices for the optimization were completed once, but the optimum vertex was repeated in duplicate.  G. ANALYTICAL HPLC All samples were analyzed using a Beckman HPLC with a 4.6 mm x 25 cm C18 Vydac peptide column, 5 fim particle size, 300 A pore size packing (Separations Group, Hesperia, CA) with a precolumn guard using a 0-60% acetonitrile/water gradient in 0.1% TFA, and flow rate 0.7 mL/min. The absorbance was detected at 214 nm using a Beckman System Gold with Diode Array Detector Module 168 (Beckman Instruments Inc., Missisauga, ON). The software used was System Gold Version 510.  49 H. STANDARD CURVES Standard curves were constructed from stock solutions of the reactants, Z-L-Asp, TyrOMe, Met-OMe and also the pure products. Z-L-Asp-Phe-OMe, Z-L-Asp-Tyr-OMe, and ZL-Asp-Met-OMe were dissolved in 100% methanol or methanol/water mixtures (80% methanol). Constant volumes (eg. 50 ^L) were injected into the HPLC in a series of concentrations and peak areas were measured. Graphs were plotted using Systat 5.01 and were curve fitted using the quadratic function. The equations obtained from the plots were used along with the quadratic equation to calculate concentration of the material from the peak area. % Yield = mmole of product x 100% mmole of limiting reactant where mmole concentration is obtained from the standard curves, mmole of product is obtained from the enzyme tube after incubation and mmole of limiting reactant is from the blank (initial reactant concentration) Although the % yield could theoretically be based on the change in reactant concentration alone, it was more appropriate to use the concentration of the product itself for the calculation, since there could be a decrease in reactants due to formation of other byproducts or complexing of the reactants with the enzyme.  I. ASSAY FOR THERMOLYSIN The method by Matsubara (1970) was used but was slightly modified. A 1.2% casein (Hammarsten) solution made up in 0.03M Tris-HCl buffer, pH 8. A thermolysin stock solution containing 0.5 mg/mL was diluted appropriately with 0.01M Tris-HCl buffer, pH 8.  50 1 mL of thermolysin was mixed at 30°C with 5 mL of 1.2% casein at pH 8, and reacted for 10 minutes. 4 mL of a 1.2M TCA (trichloroacetic acid) solution was added, and after shaking, the tubes were kept in the water bath at 30°C for 30 minutes. The precipitate was filtered off through Whatman No. 1 filter paper, 7 cm (Fisher Scientific, Vancouver, BC) and the absorbance of the filtrate was measured at 275 nm using a Shimadzu UV-Visible Recording Spectrophotometer, UV-160 (Kyoto, Japan). A standard curve was constructed from a tyrosine stock solution, at appropriate dilutions.  The E 1 ^ for thermolysin is 17.85 (Sober and Harte, 1970). The Ajgo was determined by dissolving 3 mg of thermolysin in 3 mL of 0.05M Mes buffer, pH 7, (which had to be diluted 2X since the solution was very turbid). The absorbance of the thermolysin solution was measured by a Shimadzu UV-Visible Recording Spectophotometer, UV-160 (Kyoto, Japan).  51  RESULTS AND DISCUSSION A. Z-L-ASP-PHE-OME SYNTHESIS Z-L-Asp-Phe-OMe was synthesized according to the method by Nakanishi et al. (1986), where 80mM Z-L-Asp and 80mM or 160mM Phe-OMe was dissolved in saturated ethyl acetate with 5mM CaCl2, in a total volume of 1.0 mL, catalyzed by 1 mg thermolysin and incubated for 5 hours at 40°C in a shaking water bath. In order to assay the synthesis, several methods were applied including two indirect free amino group assays, the 2,4,6 trinitrobenzene (TNBS) and o-phthalaldehyde (OPA) methods, and thin-layer chromatography (TLC). Didziapetris et al. (1991) used the OPA method to assay the extent of the enzymatic synthesis of leucine-enkephalin, which was determined as the change in concentration of amino groups. The results for the TNBS and OPA assays are in Table 2. The TNBS results were lower than those obtained by OPA, which was expected since the OPA method is more sensitive (Church et al., 1983). The detection limit for the OPA method is 9 x 10"7M (Hernandez et al., 1990). The results from the indirect methods were quite different from those calculated from the high-performance liquid chromatography (HPLC) data (Table 3). For example, the results from the 80mM/200mM Z-L-Asp/Phe-OMe were lower while the 80mM/80mM values were higher. These discrepancies were probably due to the nature of the assays themselves since the indirect methods only measure the free amino groups available.  Thus, if other products were formed (eg. other peptide linkages) or even  complexing of products that were not specifically the product, there would be higher values since the loss in free amino groups would be detected. In addition, there may have been interference from the reaction reagents with the enzymatic solutions that were assayed. In  52  Table 2. Z-L-Asp-Phe-OMe Synthesis Assayed by TNBS and OPA Methods  Yield (%) Reactant Concentration  TNBS  OPA  80mM Z-L-Asp, 80mM Phe-OMe  78.9 ± 0.03*  > 97  80mM Z-L-Asp, 200mM Phe-OMe  86.2 ± 0.1  94.7 ± 0.01  "Results are Mean values ± S.D., n=3  Table 3. Synthesis of Z-L-Asp-Phe-OMe as Calculated from HPLC  Carboxyl and Amino Components  Yield (%)  80mM Z-L-Asp and 80mM Phe-OMe  54.9 ± 6.9%a  80mM Z-L-Asp and 200mM Phe-OMe  89.7 ± 3.5%b  "Values are: Mean ± S.D., n=2 (Duplicates were 59.7% and 50%) "Values are: Mean ± S.D., n=3 (Triplicates were 90%, 86%, 93%)  54 contrast, the HPLC analysis relies on the measurement of the peak area of the product only, therefore, results would be more quantitative than the indirect methods. The results indicate that it would not be very reliable to use only the indirect methods for assaying enzymatic peptide synthesis since there is no specific assay for the product. The indirect methods may be useful as indicators of peptide formation. The thin-layer chromatography results for reactants and Z-L-Asp-Phe-OMe are in Table 4. Many different solvent conditions were evaluated (See Materials and Methods) but there was difficulty in obtaining separation of the spots. The solvent systems evaluated were based on papers involving peptide synthesis with thermolysin (Miranda and Tominaga, 1991) and from Brenner et al. (1969). It was found that the Silica Gel G plates with fluorescent indicator were the most useful. The best condition was 70:29:1 chloroform:methanol:acetic acid. TLC is not a very quick method for determining the presence of a resynthesized product because the procedure requires the repeated spotting and drying of the plates, and development in the chosen solvent system. Also, rinding the proper solvent conditions is time consuming.  Moreover, trial and error must be used in order to spot the proper  concentration of the peptide if the amount of peptide present is unknown (Touchstone and Dobbins, 1983). TLC is only useful as an assay for peptide synthesis if the product is produced in a high concentration, otherwise detection on the plate may be a problem. The concentration of the solutions used for TLC are usually 0.1 to 1 |ig/mL and the volume applied to the plate is 1-10 \iL (Fried and Sherma, 1982). Touchstone and Dobbins (1983) stated that samples may be applied as a 0.01-1.00% solution. Thus, the TLC method was less sensitive than the OPA method.  Table 4. TLC of Amino Acids and Peptides on Silica Gel G plates with fluorescence using 70:29:1 chloroform:methanol:acetic acid.  Amino Acid/ Peptide  Rf  Aspartame  0.16  Phe-OMeHCl  0.73  Z-L-Asp  0.45  Z-L-Asp-Phe-OMe  0.85  56 The third method of assaying Z-L-Asp-Phe-OMe synthesis was by analytical HPLC. The yields were calculated based on the peak areas and compared to standard curves of the reactants and products. The precipitates from the reactions were redissolved in methanol for analysis by HPLC. HPLC was more advantageous than the indirect methods because the calculation of the yield was based on the product formation itself, not just by the loss of free amino groups. Thus, it would be able to detect even very small amounts of product which would not be detected by the other methods.  Chromatograms of Z-L-Asp-Phe-OMe  synthesized using 80mM/80mM and 80mM/200mM Z-L-Asp/Phe-OMe are presented in Appendix 1. Z-L-Asp-Phe-OMe was synthesized successfully and the yield (calculated from the HPLC data) of the 80mM/80mM Z-L-Asp/Phe-OMe experiment was lower than the 80mM/200mM reaction, as expected and reported by Isowa et al. (1979) and Nakanishi et al. (1986). The repeatability of the 80mM/80mM Z-L-Asp/Phe-OMe reaction was not very good (Table 3). This was probably due to difficulties in obtaining precipitation of the product. Cassells and Hailing (1990) used thermolysin to catalyze synthesis of CBZ-Phe-Phe-OMe and found that they had difficulties with the precipitation of the product. In contrast, the results for the 80/200mM reaction were quite consistent. As Isowa et al. (1979) and Nakanishi et al. (1986) noted, the product precipitated in the reaction mixture, especially when Phe-OMe was in excess.  57 B. Z-L-ASP-TYR-OME SYNTHESIS 1. Random Centroid Optimization: First Cycle The first dipeptide that was optimized was Z-L-Asp-Tyr-OMe, because of its similarity in structure to Phe, except Tyr has an -OH group. The conditions for the optimization were modelled closely on a previous Z-L-Asp-Phe-OMe synthesis (Nakanishi et al., 1986). Since it was uncertain which factors would be important for this particular peptide, all of the factors considered to be of importance were included. Seven factors were used for the Random Centroid Optimization and they were: 1. 2. 3. 4. 5. 6.  Reactant concentration (Tyr-OMe): 80 - 160mM Solvent Ratio-DMF/EA: 0 - 1 (0-100%) Water (Buffer): 0.05 - 0.95 (5-95%) Enzyme: 13 - 40\iM CaCl2: 0 - 5mM pH: 5 - 7  7. Temperature: 35 - 50°C A total of 17 experiments were obtained for the first cycle (Table 5). Since it was expected that the synthesis would be similar to that of Z-L-Asp-Phe-OMe, the Z-L-Asp concentration was kept constant at 80mM and the Tyr-OMe (nucleophile) was varied up to 160mM so that it would be in excess ie. 1:2 ratio of Z-L-Asp to Tyr-OMe. A mixture of two solvents, dimethylformamide (DMF) and ethyl acetate (EA), was used as a single factor since this enabled optimization of two solvents at the same time. Ethyl acetate was used because was successful for Z-L-Asp-Phe-OMe synthesis and also because it had small impairing effects on thermolysin as compared to other solvents such as the chlorinated hydrocarbons (Oyama et al., 1987). DMF was chosen as a more hydrophilic solvent and because previous work using DMF as a solvent with thermolysin was successful (Sakina et al., 1988). A wide  58  Table 5. Random Centroid Design, Centroid Formulation and Yields for Z-L-Asp-Tyr-OMe Synthesis: First Cycle Tyr-OMe (mM)  DMF:EA: H2Ob  Enz (uM)  Ca (mM)  PH  1  118  1.2/57.8/41  40  2.3  6.6  36  0  2  145  28.1/8.9/63  40  4.7  5.8  36  0.10  3  102  25.5/41.5/33  27  4.3  6.8  39  0  4  148  1.5/49.5/49  16  3.4  7.0  44  0  5  157  1.6/24.4/74  30  3.9  5.7  48  1.40  6  135  25.8/42.2/32  15  2.4  5.6  37  0  7  110  42.3/30.7/27  14  3.9  6.0  36  0  8  135  26.7/31.3/42  35  .88  5.2  43  0.04  9  125  20.1/17.9/62  22  1.0  5.7  49  0.01  10  81  16/39/45  26  1.5  6.5  45  0.02  11  84  54.9/6.1/39  38  2.6  5.4  37  0  12  98  4/29/67  40  1.8  6.5  37  0.28  13  89  13.5/11.5/75  15  3.5  5.0  46  0  14  119  4.1/9.9/86  35  4.1  6.6  48  0.37  15  80  24.4/12.6/63  16  3.2  6.5  35  0  16  115  26.4/6.6/67  26  .50  5.5  39  0.001  17  158  42.8/50.2/7  29  1.5  5.4  35  0  18  125  14.1/23.9/62  30  2.4  6.0  43  0.07  19  122  15.2/21.8/63  32  2.3  6.0  43  0.09  Vertex*  Temp  •c  Yield (%)  "Vertices 1 to 17 represent the random design; Vertices 18 and 19 were obtained from the centroid formulation b Refers to ratio of solvents used: total volume of solvents and buffer = l.OmL  59 range was chosen for the solvent ratio and for the buffer since it was unknown whether or not the reaction would work better under conditions of high solvent/little buffer or high buffer/low solvent In general, the ranges set for the optimization were not very narrow in order that the search area be wide and not miss any possible areas that would lead to optimization of synthesis. The calcium concentration chosen was based on the Nakanishi et al. (1986) work, which used 5mM. Calcium was required since it is essential in mamtaining the stability of thermolysin (Matsubara, 1970). Although the pH optimum for thermolysin is 6.5-8.0 (Matsubara, 1970; Reddy, 1991), the range for optimization was 5-7 since synthesis did occur at pH 5 (Nakanishi et al., 1986), and pH 7 was the maximum since the Tyr-OMe precipitated out at pH 7 or higher. The temperature range that may be used for thermolysin is 35-55°C (Reddy, 1991) but the limits used were 35-50°C. The incubation time used for the reaction was 24 hours. A preliminary experiment was carried out with two incubation times, 5 and 24 hours, and it was found that product formation was very small at 5 hours, while yields were increased at 24 hours. Once the experiments from the first cycle were completed, yields were calculated and expressed as % synthesis. The results were entered into the random centroid program, which resulted in the centroid formation, consisting of two experiments.  After these were  completed, responses were also entered into the program (Table 5). Although not all of the vertices had a measurable yield, they all had a peak area for the products. However, some were too small to be calculated and thus were expressed as zero % synthesis. Nine out of the nineteen experiments had zero % synthesis. In general, synthesis occurred at solvent levels up to 43%, with Vertex 8 and 10 being the exceptions with solvent levels of 58% and  60 55%, respectively. There was definitely no synthesis at >58% solvent Although Vertices 13 and 15 had low solvent levels, there was no synthesis, which indicated that other factors were influencing the reaction. In this case, it was probably because of the low enzyme level. All of the vertices that had <22\iM enzyme had 0% synthesis. The two vertices that had the highest yields had lower levels of solvent; Vertex 5 had 26% while Vertex 14 had 14%. The other vertices all had solvent concentrations higher than 30% eg. 33%, 37%, etc. These all had lower yields than Vertex 5 and 14. The highest yield from the first cycle was Vertex 5, which had 1.40%.  2. Mapping All of the information from the 19 experiments was then used for the mapping step of the optimization. The responses were plotted against each individual factor, and data points which were in clusters or groups were linked together. The maps were then used in order to narrow the limits for the second cycle of optimization.  However, for this  optimization, the information obtained from the maps was limited since there was no linkage of the data points and therefore, no observable pattern that indicated the direction of the optimum. Because the original plots that were obtained were very small, all of the maps were replotted for better clarity. Thus, these plots do not have any of the lines (if any) drawn in but the observations were based on the original maps (Figures 3, 4, 5, and 6). The limits for the second cycle were still changed based on the mapping information. Since a maximization of yield was desired, the conditions were changed according to the area where the highest point was located. The limits for Tyr-OMe were increased since the reaction seemed to favour an excess of Tyr-OMe. However, the Tyr-OMe concentration  61  X  s >»  o -  -1  60  80  100  120  140  160  Tyr-OMe Concentration (mM)  B  0.2  0.4  0.6  0.8  1.0  DMF/EA Ratio (mL)  Figure 3.  Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: A) Tyr-OMe concentration and B) DMF:EA ratio  62  1  -  0  -  1 c (0  -1 0.0  0.4  0.6  1.0  Buffer (mL)  D  I 20  30  40  50  Enzyme tyiM)  Figure 4.  Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: C) Buffer and D) Enzyme concentration  63 E  1 h  I to  o f-  2  3  Calcium (uM)  1 h  1 CO  I  0 h  ••«••! «...  6  8  PH  Figure 5.  Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: E) Calcium chloride and F) pH  64  I  G  I  •  x •  •  c w  .1.  -1  30  •  • • • •  1  1  35  40  45  •-  50  Temperatura  Figure 6.  Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: G) Temperature  65 could only be increased to 180mM (not 200mM as with the Phe-OMe) since there was difficulty in dissolving the reactants. The DMF/EA range was decreased considerably since yields were quite low at the higher solvent concentrations ie. higher than 30%. It was expected that the optimization would give some indication as to which solvent would be the better of the two, but unfortunately, there was not much information from the maps. Based on the synthesis, only three out of the ten vertices that had a measurable yield had DMF as the major solvent Thus, for the second cycle, ethyl acetate made up the major portion of the DMF/EA ratio. The limits were increased for buffer concentration because the maps indicated an upward trend. There was no particular trend for the enzyme concentration as indicated from the maps, but the limits was increased slightly since no yields were obtained when the enzyme concentration was <20nM. There was a definite upward trend from the maps for the calcium chloride, so its concentration was increased. The limits for pH were narrowed since the points were within the boundary limits. The maps for temperature indicated an upward trend and since there did not seem to be favourable synthesis at levels lower than 40°C, the temperatures were increased. 3. Random Centroid Optimization: Second Cycle The ranges for the second cycle were: 1. 2. 3. 4. 5. 6. 7.  Reactant concentration (Tyr-OMe): 150 - 180mM Solvent Ratio-DMF/EA: 0 - 0.3 (0-30%) Water (Buffer): 0.75 - 0.95 (75-95%) Enzyme: 25 - 40pM CaCl2: 4 - 8mM pH: 5.4-6.5 Temperature: 40 - 50°C Ten experiments were chosen for the second cycle (Table 6). Once the second cycle  66 and centroid experiments were completed, yields were calculated and the mapping process was repeated. Table 6 shows the yields for the second cycle, which were more successful than the first cycle because all of the vertices had measurable yields and the yields were increased. A maximum synthesis of 2.65% was obtained in Vertex 27. The reaction was also repeated twice but it was found that yields were lower in both of the duplicates; yields were 1.79% and 1.96%. The overall % synthesis for Vertex 27: n=3, Mean ± S.D. was 2.13 ± 0.46%. A chromatogram of Z-L-Asp-Tyr-OMe, which was obtained for Vertex 27, is in Appendix 1.  4. Mapping By changing and narrowing the limits in the second cycle, synthesis of Z-L-Asp-TyrOMe was increased (Figures 7, 8, 9 and 10). The mapping was again not very useful since there was no joining of data points and no particular direction or pattern in the maps. Yields were still low because there was precipitation of the reactants and this was probably due to the two solvents, DMF and ethyl acetate, rather than the pH. Changing the Tyr-OMe concentration was effective because yields increased slightly, but there was no indication that higher Tyr-OMe concentrations gave better yields. The overall effect of decreasing the DMF/EA ratio was successful. However, it was difficult to conclude whether a reduction in the DMF content had an impact (as in the first cycle) since some vertices had good yields with higher DMF. For example, Vertex 21 had a higher yield than Vertex 28 even though the DMF ratio was smaller for Vertex 21. It was apparent that the solvent:buffer ratio themselves did not affect yields. The results for Vertex 30 and 31 were not as expected since the conditions were very similar, but the yield for Vertex 31 was about half that of Vertex  67 Table 6. Random Centroid Design, Centroid Formulation and Yields for Z-L-Asp-Tyr-OMe Synthesis: Second Cycle Vertex*  Tyr-OMe (mM)  DMF:EA: H2Ob  Enz (uM)  Ca (mM)  PH  Temp °C  Yield (%)  20  167  3.7/19.3/77  36  4.2  6.3  49  0.15  21  154  2.9/13.1/84  31  6.4  6.4  42  1.08  22  169  1.5/12.5/86  25  7.7  5.9  45  0.82  23  180  5.5/13.5/81  30  5.7  5.7  47  0.77  24  175  0.7/23.3/76  36  7.9  5.9  45  1.04  25  159  6.1/14.9/79  29  4.9  5.9  47  0.77  26  176  7.2/16.8/76  28  6.6  6.0  41  0.81  27  154  2.3/9.7/88  39  7.0  6.3  46  2.65c  28  162  1.1/13.9/85  35  5.4  5.8  41  0.76  29  162  0.9/5.1/94  26  4.8  6.3  40  0.85  30  165  3.2/14.8/82  31  6.1  6.0  45  0.82  31  166  3.1/14.9/82  31  6.2  6.0  45  0.43  'Vertices 20 to 29 represent the second cycle; Vertices 30 and 31 were obtained from the centroid formulation b Refers to ratio of solvent used: total volume of solvents and buffer = l.OmL c Actual value for % synthesis calculated using the mean values of this value and duplicates: 2.13 ± 0.46% (Mean + S.D., n=3; duplicates were 1.96% and 1.79%)  68  3  I  (0  1 -  150  • I • • 160  • •  170  180  190  Tyr-OMe Concentration (mM)  ! (0  0.0  0.1  0.2  DMF/EA Ratio (mL)  Figure 7. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: A) Tyr-OMe concentration and B) DMF/EA ratio  69  ~  2  1 >•  (0  1 -  0.7  0.8  0.9 Buffer (mL)  1 £  Enzyme (uM)  Figure 8.  Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: C) Buffer and D) Enzyme concentration  70  E  5  1 >»  5  6  7  8  Calcium tyiM)  3  2 -  5 ! #  1 •  5.6  •  5.8  •  •  •  6.0  6.2  6.4  6.6  PH  Figure 9.  Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: E) Calcium chloride and F) pH  71  X  40  45  Temperature  Figure 10. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Tyr-OMe: G) Temperature  72 30. In general, the reaction favoured higher enzyme concentrations, >25 |iM. Synthesis was better when the CaCl2 concentration was >4.2mM. There was not any particular trend for the pH. Synthesis was best for those vertices which had a temperature of >40°C, but only up to 46°C since yields dropped at temperatures above 46°. There are a number of reasons that may account for the low yield of Z-L-Asp-TyrOMe. The main reason was the solubility of the reactants. Even though reactants appeared to be soluble initially since they formed a slightly turbid suspension, precipitation occurred by the end of the 24 hour incubation. This occurrence was evident in both cycles, and even when ethyl acetate was the major solvent in the second cycle. It was noted that this phenomenon would occur after 5 hours of incubation; up to 5 hours, the reactants would appear to be dissolved. Precipitation had a negative effect on the reaction since the availability of the reactants that were in contact with the enzyme was decreased. Antonini et al. (1981) and Carrea (1984) stated that a suitable solvent for synthesis must be able to solubilize the substrates. A solvent which is able to solubilize large amounts of substrates that partially affects enzyme activity or stability is more useful than one which is not able to solubilize the reactants and does not affect enzyme activity (Antonini et al., 1981). Thus, an alternate solvent system or mixture of solvents would be better for this reaction. Not only are the solvents important in solubilizing the reactants, but also are required to shift the equilibrium toward peptide formation. Thus, the conditions were set in order to favour equilibrium shift. The presence of organic solvents is required in order to limit the water concentration. However, use of organic solvents at high concentrations can affect the enzyme by causing it to lose its catalytic ability, or cause a decrease in solubility and stability  73 (Jakubke, 1987).  Water-miscible solvents cannot be used at levels higher than 50%  (Khmelnitsky et al., 1988). This was evident in the first cycle of the Z-L-Asp-Tyr-OMe synthesis, where higher concentrations of solvents eg. higher than 43% resulted in either no synthesis or poor synthesis. DMF, a water-miscible solvent, has been noted as being capable of solubilizing enzymes (Dordick, 1991). DMF did not appear to be very favourable for promoting synthesis when it was the major solvent This is probably because DMF stripped the essential water required by the enzyme and caused a loss in catalytic ability (Zaks and Russell, 1988). Although DMF has been used in enzymatic synthesis using thermolysin, perhaps its combination with ethyl acetate was not very favourable and could have affected the enzyme, especially at higher concentrations. It was expected that ethyl acetate would be a suitable solvent since it is water-immiscible and has been used for thermolysin-catalyzed synthesis, but not for this peptide. Use of higher concentrations of solvents would be most favourable for the reaction, since the presence of organic solvents means that there is less water available and therefore, less hydrolysis. In this reaction, the concentration of solvents in the second cycle was decreased, which may explain why yields could not be increased. In both cycles, there was no clear indication of which factors influenced the synthesis, because the mapping from both cycles showed no definite trends for any of the factors. Moreover, attempts to provide favourable conditions for precipitation of the product were not successful. A method often used to promote peptide synthesis involves precipitation of the product by using the appropriate concentration of reactants with blocking groups on either the carboxyl or amino end to cause an equilibrium shift (Morihara, 1987; Nakanishi and Matsuno, 1988). Although it was found that the reaction required an excess of Tyr-OMe,  74 and that yields were slightly increased, most of the product remained in solution and very little was precipitated out. After the chemical synthesis, the pure Z-L-Asp-Tyr-OMe was found to be water-insoluble, therefore it was expected to precipitate. The conditions used were supposed to enhance yield since one of the reactants (Tyr-OMe) was used in excess, and it was expected that the reaction would follow the law of mass action and form the product (Kullmann, 1985). In this case, only the clear portion of the enzyme reaction mixture was removed and diluted for HPLC analysis. This was because it was known that the reactants were precipitating out in the blanks and that the precipitate would mostly be the reactants. However, to check, the precipitate was separated from the supernatant by centrifugation and redissolved in 100% methanol for analysis by HPLC. It was found that the majority of the precipitate was Z-L-Asp and Tyr-OMe. Cassells and Hailing (1990) found precipitation of reactant complexes resulted in low product yields. They noted that reduced yields may have been due to a decreased reaction rate. It was not expected that a lot of product would be in the precipitate because of the difficulty in achieving soluble reactants. Even when conditions were changed in order to maximize synthesis, the changes did not make a major difference in increasing yields. A larger ratio of ethyl acetate:DMF still did not solubilize the reactants well. This showed that the solvent was the major cause of the poor yields. It would not be expected that the yields could be increased any more after the second cycle, so the optimization was terminated. Although Tyr-OMe has a similar structure to Phe-OMe, with the -OH group being the difference, tyrosine is not the most favourable amino acid for thermolysin. As evident from studies of the hydrolysis of peptides and proteins using thermolysin, the most favoured amino  75 acids at the amino side of the splitting point in the peptide are those with hydrophobic residues with bulky side chains, such as isoleucine, leucine, valine, phenylalanine, methionine and alanine (Fruton, 1982). The amino site of tyrosine is susceptible to hydrolysis by thermolysin, but only to a minor extent (Matsubara, 1970). The specificity of the enzyme is greatly influenced by the types of amino acids at the region of the sensitive peptide bond for hydrolysis, and this also applies for the reverse reaction. Oka and Morihara (1980) noted that enzymatic peptide synthesis was best when the amine components used were hydrophobic and that the order of efficiency was: Leu, Phe, lie, Val, Ala, and then Tyr. Miranda and Tominga (1991) also found that hydrophobic amino acids were required as the amino component This means that the tyrosine probably had less efficient binding capacity to the Sj subsite of thermolysin, which would account for the poor levels of synthesis of the Z-LAsp-Tyr-OMe. The yields for Z-L-Asp-Tyr-OMe were not expected to be very high for this reason, thus the results obtained are reasonable even though they are very much lower than those obtained with the Z-L-Asp-Phe-OMe. They might have been higher if a different solvent was used and if there was better solubility of the reactants.  5. Solvent Suitability for Z-L-Asp-Tyr-OMe Synthesis After the second cycle was completed and problems with solubility were found, additional work was completed to determine whether an alternative solvent could be used for the synthesis. Ten solvents, tested for suitability in terms of their ability to solubilize the reactants, are listed in Table 7. The reactant concentrations used were 80mM Z-L-Asp and 160mM Tyr-OMe, and they were dissolved with 0.5 mL of the solvent first, then 0.5 mL of 0.05M Mes buffer, pH 6, was added. These conditions were used to model the work done  Table 7. Solvent Suitability for Z-L-Asp-'  -OMe Synthesis*  Solvent  Observations  Ethanediol  Poor dissolution; suspended particles  N-Methyl-Pyrrolidone  Complete dissolution  Triethyleneglycoldimethylether (Triglyme)  Poor dissolution; large fluffy crystals  Dichloroethane  Poor dissolution; 2 phases with clump of reactants at bottom  Acetonitrile  Complete dissolution  Methanol  Dissolution except for a few crystals  Toluene  Opaque solution; 2 phase separation  Butanol  Clump of reactants at bottom and 2 phase separation  Dimethylformamide  Complete dissolution  Ethyl Acetate  Dissolution but forms emulsion  "Used combination of 80mM Z-L-Asp and 160mM Tyr-OMe with 0.5mL solvent and 0.5mL buffer  77 by Nakanishi et al. (1986). A different range of solvents was used, including water-miscible (DMF, N-methyl-pyrrolidone, acetonitrile, and methanol), water-immiscible (dichloroethane, ethyl acetate, toluene and butanol) and two polyols, ethanediol (ethylene glycol) and triglyme=triethyleneglycoldimethylether, which are both water soluble. This range of solvents was important in order to determine which type of solvent system was most suitable for synthesis. In order to keep the systems simple, mixtures of solvents were not attempted. Water-miscible solvents have been widely used in enzymatic peptide synthesis; however, at high concentrations, eg. 50-70% they can affect the catalytic ability of the enzyme (Khmelnitsky et al., 1988). The importance of the water-immiscible solvents is that there may be a possibility that the reaction could be run as a biphasic system, if there is good solubilization of the reactants. The biphasic system is advantageous in that the enzyme has limited contact with the solvent and is less likely to be damaged by it since the enzyme is dissolved in the aqueous layer (buffer) (Kuhl et al., 1980; Carrea, 1984; Eggers et al., 1989). There is also easier product recovery if the product is more soluble in the solvent phase and hydrophobic reactants may have increased solubility. However, not all of the substrates can be solubilized in the non-polar solvents and the reaction is much slower (Jakubke et al., 1985). In addition, solvents that are too hydrophobic can cause enzyme inactivation due to large amounts of the substrate partitioning into the aqueous phase (Stevenson and Storer, 1991). The polyols have good potential for enzymatic synthesis since they are the only solvents that can be used in high concentrations (eg. water contents of 5-10%) and catalytic activity of the enzyme is maintained (Khmelnitsky et al., 1988). It was found that the water-miscible solvents were the most suitable for solubilizing  78 Z-L-Asp and Tyr-OMe. All of the water-immiscible solvents, with ethyl acetate being the exception, were poor and thus the synthesis could not be run as a biphasic system, unless there was a suitable water-miscible solvent that could be combined with it Both polyols were very poor at solubilizing the reactants, which was unfortunate in that polyols are the only solvents that may be used at high concentration with less effects on the stability of the enzyme (Khmelnitsky et al., 1988). A peak area of the product was obtained for each solvent but was too small to be calculated so was expressed as 0% synthesis. 6. Synthesis of Z-L-Asp-Tyr-OMe in Different Organic Solvents Six of the ten solvents that were tested for solubilizing ability were used in the synthesis of Z-L-Asp-Tyr-OMe using 80mM Z-L-Asp, 160mM Tyr-OMe, 5mM CaCl2, 1 mg thermolysin, at pH 6, 40°C, in a total volume of 1 mL for 5 hours. A shorter reaction time was used because it was thought that the reaction might have occurred faster with other solvents. The solvents chosen were the water-miscible ones (methanol, acetonitrile, DMF, N-methyl-pyrrolidone) and two water-immiscible ones (ethyl acetate and dichloroethane). Five of the solvents chosen had good solubilizing ability, except for dichloroethane, which was purposely chosen to see the effect the solvent had on peptide synthesis. The polyols were excluded since they were such poor solubilizers. At first, the reactions were carried out with 40% solvent and 60% buffer, which resulted in very low synthesis (Table 8). As expected, the water-miscible solvents gave betterresultsthan the water-immiscible ones since they were better at dissolving the reactants. The best solvent from these experiments was methanol, followed by acetonitrile, and lastly, ethyl acetate. Thus, ethyl acetate may not have been the best choice for the optimization. DMF, N-methyl-pyrrolidone and dichloroethane  79 Table 8. Synthesis of Z-L-Asp-Tyr-OMe in the Presence of Organic Solvents Organic Solvent  Amount of Solvent* (mL)  Yield  Methanol  0.12  0.66  Methanol  0.40  0.09  Acetonitrile  0.12  0.35  Acetonitrile  0.40  0.07  DMF  0.12  0.18  DMF  0.40  0  N-Methyl-Pyrrolidone  0.12  0.21  N-Methyl-Pyrrolidone  0.40  0  Ethyl Acetate  0.12  0.07  Ethyl Acetate  0.40  0.01  Dichloroethane  0.12  0.06  Dichloroethane  0.40  0  (%)  The reactions were made up to l.OmL with 0.05M Mes Buffer, pH 6, with lmg thermolysin, 5mM CaCl2, 80mM Z-L-Asp and 160mM TyrOMe. The reaction mixtures were shaken for 5 hours at 40°C.  80 had 0% synthesis at 40% levels, indicating that there may have been some effect on the enzyme. Based on these results, it was felt that Z-L-Asp-Tyr-OMe optimization should not be continued since none of the solvents seemed to have a good potential for synthesis. Later, it was decided that this experiment should also be carried out at a low level of solvent, so 12% solvent and 88% buffer with the above conditions was also completed. The 12% level was chosen based on the optimization of Z-L-Asp-Tyr-OMe. The highest concentration was 40%, since denaturation of the enzyme often occurs at levels above 50% (Khmelnitsky et al., 1988; Clapes et al., 1990b) and it was decided that the level should be less than 50%. At the 12% level, surprisingly, all of the solvents had some synthesis of Z-L-Asp-Tyr-OMe, including dichloroethane. The order of solvents, starting from the best, was: methanol, acetonitrile, N-methyl-pyrrolidone, DMF, ethyl acetate and dichloroethane. Once again, the two best solvents were methanol and acetonitrile. If more work were to be completed for the synthesis, these two solvents could be considered.  C. Z-L-ASP-MET-OME SYNTHESIS 1. Preliminary Work: Solvent Suitability It was decided that the second sweet dipeptide, Z-L-Asp-Met-OMe, should also be synthesized enzymatically using thermolysin. In order to favour equilibrium shift, organic solvents were used, and also an excess of one of the reactants. For this dipeptide, preliminary experiments were completed before optimization was applied. Fifteen solvents were tested for solvent suitability, using 80mM Z-L-Asp, 160mM Met-OMe, 5mM CaCl2, 1 mg thermolysin, with 0.5mL solvent followed by 0.5mL 0.05M Mes buffer, pH 6 (See Table 9).  81 Table 9. Solvent Suitability for Z-L-Asp-Met-OMe Synthesis* Solvent  Observations  Ethanol  Complete dissolution only when buffer added  Butanol  Forms 2 layers; top layer is slightly opaque, bottom layer is clear  Methanol  Complete dissolution  Propan-2-ol (Iso-propyl alcohol)  Complete dissolution only when buffer added  Acetonitrile  Complete dissolution only when buffer added  Dioxane  Complete dissolution after buffer added  Dimethylformamide  Complete dissolution only when buffer added  N-Methyl-Pyrrolidone  Dissolution but had small clump of viscous material on bottom; dissolved when buffer added  Ethyl Acetate  Had white precipitate but dissolved when buffer added  Dichloroethane  Poor dissolution; 2 phase separation with clump of white particles on bottom, clear layer on top  Toluene  Poor dissolution; 2 phase separation with clump of white particles on bottom, clear layer on top  Chloroform  Poor dissolution; 2 phase separation with opaque layer on top, clear layer on bottom  Triglyme  Poor dissolution until buffer added; solution was translucent  Glycerol  Slightly cloudy but had complete dissolution once buffer added  Ethanediol  Complete dissolution  "Used combination of 80mM Z-L-Asp and 160mM Met-OMe with 0.5mL solvent and 0.5mL buffer  82 The range of solvents used was the same as that used for Z-L-Asp-Tyr-OMe. A larger number of solvents were used, some of which were used previously for enzymatic reactions (Auriol et al., 1990; Kitaguchi and Kilbanov, 1989; Mozhaev et al., 1989). The watermiscible solvents used were methanol, propan-2-ol, acetonitrile, dioxane, ethanol, DMF, and N-methyl-pyrrolidone; water-immiscible solvents were ethyl acetate, dichloroethane, toluene, and chloroform; and three polyols used were triglyme, glycerol and ethanediol (ethylene glycol). There were a lot more suitable solvents for solubilizing Z-L-Asp and Met-OMe as compared to the Tyr-OMe. All of the water-miscible solvents were good at dissolving the reactants, but most required the addition of buffer before there was complete dissolution. Most of the water-immiscible solvents were poor solubilizers of the reactants, with butanol and ethyl acetate being the exceptions. Therefore, there was the possibility of applying the biphasic system for enzymatic synthesis, which could result in better yields. Unlike the situation with Tyr-OMe, the polyols were suitable solvents, although triglyme did not completely dissolve the reactants since they appeared slightly translucent. Thus, solvents from all three categories had potential to be used for enzymatic synthesis.  2. Synthesis of Z-L-Asp-Met-OMe in Different Organic Solvents Nine of the fifteen solvents tested for solvent suitability were used for synthesis of ZL-Asp-Met-OMe using 80mM Z-L-Asp, 160mM Met-OMe, 5mM CaCl2, 1 mg thermolysin, pH 6, in a total volume of 1 mL at 40°C, in a shaking water bath for 5 hours. The same conditions used for the Z-L-Asp-Tyr-OMe synthesis were used. Six of the solvents used for the Z-L-Asp-Tyr-OMe synthesis were repeated for this peptide, and three different solvents,  83 glycerol, ethanediol, and ethanol were also used. Once again, the majority of the solvents had good dissolving properties except for dichloroethane. The same levels of solvent used for the Z-L-Asp-Tyr-OMe synthesis were used again, 12 and 40%. The results are in Table 10. The product peak appeared in the chromatograms for all of the solvent-reaction mixtures, but some were too small to be measurable and thus were expressed as 0% synthesis. Although solubility of the reactants in the solvents was much better than for Z-L-Asp-TyrOMe, synthesis using the same conditions was not. This was unexpected since Z-L-Asp-TyrOMe synthesized under the same conditions had better results even though Met is supposed to be a more suitable substrate for thermolysin.  Thus, the conditions used for these  experiments were probably not the best for synthesis with Met and shows that enzymatic synthesis is very different for every peptide system. Six solvents which were used in the trial experiments first were methanol, acetonitrile, DMF, N-methyl-pyrrolidine, ethyl acetate and dichloroethane. At the 12% level, only methanol, DMF, N-methyl-pyrrolidone supported some synthesis, while at the 40% level, there was 0% synthesis in all solvents except for methanol. Based on these results, it would be expected that methanol should be used for the optimization. However, since yields with the water-miscible and water-immiscible solvents were so poor, it was presumed that a polyol might be better, especially since it may be used at high concentrations without damaging the enzyme (Khmelnitsky et al., 1988). The other water-miscible solvent of particular interest was ethanol, since it is a food grade solvent and would be favoured over other solvents such as methanol, which is toxic. Glycerol, ethanediol and ethanol were used in the synthesis trials last. For these three solvents, aliquots were removed at 5 hours, and also left longer until 24 hours (upper and lower parts of Table 10).  84 Table 10. Synthesis of Z-L-Asp-Met-OMe in the Presence of Organic Solvents Organic Solvent  Amount of Solvent* (mL)  Yield (%)  Methanol  0.12  0.03  Methanol  0.40  0.009  Acetonitrile  0.12  0  Acetonitrile  0.40  0  DMF  0.12  0.02  DMF  0.40  0  N-Methyl-Pynolidone  0.12  0.02  N-Methyl-Pyrrolidone  0.40  0  Ethyl Acetate  0.12  0  Ethyl Acetate  0.40  0  Dichloroethane  0.12  0  Dichloroethane  0.40  0  Glycerol  0.12  0.008  Glycerol  0.40  0  Ethanediol  0.12  0  Ethanediol  0.40  0  Ethanol  0.12  0  Ethanol  0.40  0  Glycerolb  0.12  0.47  Glycerol  0.40  0.12  Ethanediol  0.12  0.20  Ethanediol  0.40  0.14  Ethanol  0.12  0.13  Ethanol  0.40  0.05  •The reactions were made up to l.OmL with 0.05M Mes buffer, pH 6, with lmg thermolysin, 5mM CaC12, 80mM Z-L-Asp and 160mM MetOMe. The reaction mixtures were shaken for 5 hours, at 40°C. b Conditions were exactly the same as above but reaction time was 24 hours.  85 The 5 hour incubation was equally poor as compared to the other solvents. However, 24 hours was a more suitable incubation time since yields increased for all three solvents. Just as with the Z-L-Asp-Tyr-OMe, peptide synthesis was higher at the 12% level as compared to the 40% level, even for the polyols, which are supposed to be able to be used at very high concentrations. Since synthesis was highest with glycerol, it was the solvent of choice for the optimization. Rather than using a mixture of solvents, it was decided that only one solvent be used in the optimization. Kitaguchi and Klibanov (1989) had found that enzymatic synthesis could be improved when water mimics such as glycerol, ethylene glycol or formamide were added to tert-amyl alcohol because they acted to partially replace water, and therefore limited the water concentration to 1%.  3. Random Centroid Optimization: First Cycle Since it was uncertain which factors were of importance, seven factors were used for the Z-L-Asp-Met-OMe synthesis, with wider ranges than for the Z-L-Asp-Tyr-OMe. They were the following: 1. 2. 3. 4. 5. 6.  Z-L-Asp: 20-200mM Met-OMe: 20 - 200mM Solvent-Glycerol: 0-1.0 (0-100%) Enzyme: 5 - 40\iM CaCl2: 2 - lOmM pH: 5-7  7. Temperature: 35 - 50°C A total of 17 experiments were completed for the first cycle (Table 11). Both Z-LAsp and Met-OMe were varied for the optimization in order to see whether or not synthesis could occur at the lower concentrations and to find the reactant which was the limiting one, even though in the previous two syntheses the nucleophile (Tyr-OMe and Phe-OMe) was  86 Table 11. Random Design, Centroid Formulation and Yields for Z-L-Asp-Met-OMe Synthesis with Glycerol: First Cycle Vertex*  Z-L-Asp (mM)  Met-OMe (mM)  Glycb (mL)  Enz (UM)  Ca (mM)  pH  Temp °C  Yield (%)  1  151  68  .65  14  7.7  5.1  39  0  2  31  22  .89  14  3.5  6.2  40  0  3  113  134  .56  32  2.9  5.2  43  0  4  69  86  .08  38  2.7  6.3  38  0  5  33  158  .68  21  8.9  6.3  42  0  6  157  193  .63  27  5.0  5.2  45  0.09  7  32  76  .58  13  6.5  5.6  41  0  8  176  135  .59  22  8.6  6.5  38  0  9  155  23  .22  27  9.0  6.3  41  0  10  41  117  .01  35  8.3  5.0  40  0.72  11  192  57  .19  23  6.6  6.2  42  0  12  118  32  .23  22  2.1  5.8  50  0  13  104  182  .59  33  9.3  6.7  36  0  14  71  128  .12  8.5  5.3  5.4  39  0  15  196  188  .59  18  8.1  6.2  48  0  16  76  33  .03  37  4.1  5.4  45  0.18  17  142  105  .08  40  2.0  6.9  36  0  18  122  115  .27  27  5.6  5.9  41  0  19  114  122  .28  27  5.5  5.8  41  0  'Vertices 1 to 17 represent the random design; Vertices 18 and 19 are the centroid formulation 'Total volume for reaction = l.OmL; Buffer content is l.OmL - glycerol content eg, 1.0 - 0.65 = 0.35mL buffer  87 limiting. Previous investigations that used thermolysin as a catalyst found that increasing the concentration of either the carboxyl or amino component resulted in an increase in yield (Oka and Morihara, 1980). Cheng et al! (1988) obtained quantitative yields of Z-Pro-Leu-Gly-NH2 when the carboxyhamine component ratios were 2:1 and 1:2. Although the preliminary work in this study has indicated that glycerol does not improve yields at higher concentrations (40%), the limits were still set to vary up to 100% in case there is an optimal combination which requires high concentrations. Enzyme concentration was varied since some researchers have been able to obtain resynthesis at lower concentrations eg. 1.3 x 10"7 mol enzyme to 5.1 x 10"4 mol carboxyl and amine component (Miranda and Tominaga, 1991). Based on the work with Z-L-Asp-Tyr-OMe, the calcium concentration range was widened since it was expected that the calcium requirements would be important. The pH and temperature ranges were kept the same as in the Z-L-Asp-Tyr-OMe optimization experiments, as well as the incubation time (24 hours). Once the 17 experiments were completed and yields were calculated, the results were entered into the random centroid program, which resulted in the 2 centroid experiments (Table 11). The peptide yields for the first cycle are also shown in Table 11. The results from the first cycle of optimization with glycerol were poor; in thirteen of the vertices there was no peak area identified at all. Since Vertices 3, 12 and 14 had peak areas that were too small to be calculated, their yields were expressed as 0%. The only vertices that had measurable yield were Vertex 6, 10 and 16, where Vertex 10 had the highest yield. Vertex 10 and 16 had a very small amount of glycerol vs. buffer while Vertex 6 had a high concentration but the least yield. None of the other vertices with high concentrations of  88 glycerol had any yield. A chromatogram of Z-L-Asp-Met-OMe synthesized in glycerol for the optimum (Vertex 10) is in Appendix 1.  4. Mapping After the centroid experiments were completed, all of the information from the first cycle was entered into the program and the responses were plotted for each individual factor (Figures 11, 12, 13, and 14). Just as with the Z-L-Asp-Tyr-OMe, the maps were not very helpful since there were so few positive results. Although there was some linkage of points for some factors (eg. Z-L-Asp, pH, temperature) it was difficult to obtain information suitable for narrowing the selection of conditions.  There were two main reasons why synthesis of Z-L-Asp-Met-OMe was not successful. The solvent definitely was important in affecting synthesis. Although the preliminary work showed that glycerol was the solvent of choice it was apparent that it was affecting synthesis. Glycerol appeared to have an inhibitory effect on peptide synthesis since there were some vertices that did not have any product appear at all and the two vertices that did have some product had very low concentrations of glycerol (with Vertex 6 being the exception). Durrant et al. (1986) studied the effect of glycerol on thermolysin-catalyzed peptide bond synthesis and found that it inhibited the synthesis by affecting the structural properties of thermolysin. Synthesis of the peptide, which happened to be Z-L-Asp-Phe-OMe, was possible in glycerol but the rate of synthesis was affected. This probably also occurred with the Z-L-Asp-MetOMe, except that in this case it did not appear that the amount of product would increase even if the incubation time was longer since there were so many vertices with no yield at all.  89 0.8  -0.2  50  100  150  200  Z-L-Asp Concentration (mM)  B  l  v.o  • 0.6 -  -  6  0.4 -  -  i  0.2 -  r?  • •  0.0  -n 4  •  •  •  •  • • •  —•&  •  l  i  i  50  100  150  200  Met-OMe Concentration (mM)  Figure 11. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: A) Z-L-Asp concentration and B) MetOMe concentration  90 0.8  0.6  £  0.4  £  0.2  CO  0.0 -  -0.2 0.0  0.2  0.4  0.6  0.8  1.0  Glycerol (mL)  VI.O  I  1  1  D  1  •  t;  S !  0.6 -  -  0.4 -  -  0.2 -  -  • •  0.0  m  •  I  10  1  20  ••  i  i  30  40  50  Enzyme tyiM)  Figure 12. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: C) Glycerol and D) Enzyme concentration  91  u.o  i  I  • 0.6 -  -  s  0.4 -  -  t  0.2 -  t?  -  • •  0.0  -n o  —• • +  • • • • • •  i  I  5  6 PH  Figure 13. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: E) Calcium chloride and F) pH  92  G  i  v.o  i  •  s  0.6  -  0.4  -  ^J  1  0.2 -  -  • •  0.0  •  30  •  4  •  i  I  40  50  60  Temperature  Figure 14. Mapping results from the first cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: G) Temperature  93 The second observation of the Z-L-Asp-Met-OMe synthesis was that some vertices appeared to have precipitation occurring in the tubes with enzyme after the 24 hour incubation. Tubes appeared to be slightly cloudy and a few had suspended particulate matter. It was thought that there may have been precipitation of the product, since it was found after the Z-L-Asp-Met-OMe was chemically synthesized that it is not a water soluble peptide. In order to determine the nature of the particulate matter, the tubes were centrifuged, the supernatant was decanted and the precipitate was redissolved in 100% methanol. Both supernatant and precipitate were analyzed for presence of product. Because there was great difficulty in dissolving the precipitates even though 100% methanol was used, a small amount of water was also added to see if the precipitate would dissolve. After running the samples on the HPLC, it was apparent that the precipitate was mainly Z-L-Asp and that there was no product present.  This meant that the precipitate could have been due to the enzyme  complexing with some of the reactants, producing the difficult-to-dissolve mass. This also meant that the Z-L-Asp, although appearing to be soluble initially, was precipitating out as the reaction progressed and was not available for reaction with the Met-OMe. As mentioned earlier, solubility of the reactants is essential for peptide synthesis, otherwise the reactants are not able to have contact with the enzyme (Carrea, 1984). Glycerol was probably causing the Z-L-Asp to precipitate out, which would account for the low yield. Based on the above observations, it was decided that the second cycle of optimization should not be completed using glycerol as the solvent. It was not expected that yields would improve if glycerol were to be used because of its inhibitory effect on thermolysin. Since ethanediol (ethylene glycol) was the second most promising solvent based on the preliminary  94 work, and is another polyol, it was the next solvent of choice. However, before the second cycle of optimization was started, a few preliminary experiments with ethanediol were carried out in order to ensure that it was a better solvent than glycerol. Thus, three of the vertices from the first cycle that were carried out with glycerol were repeated using ethanediol. Vertex 10, which had the highest yield was repeated, along with Vertex 19 which had a midrange amount of solvent (0.28mL), and Vertex 5 (0.68mL) were all found to have higher yields with ethanediol as compared to glycerol. It was decided that ethanediol was indeed more suitable since there was synthesis ie. a peak area for the product was detected. The limits for the second cycle were only slightly modified since it was uncertain whether or not the information obtained was sufficient since only three out of the nineteen experiments had positive results. Some of the maps were used in order to change the limits. The Z-L-Asp concentration was made the limiting reactant and the range was made very small for it since it was expected that the Met-OMe concentration was of more importance, based on the work by Nakanishi et al. (1986). Thus, the Met-OMe concentration was changed in order for it to be in excess because the reaction would be able to promote precipitation or product formation via the law of mass action (Kullmann, 1985). The ethanediol range was small since it was found from the preliminary work that yields were not very high when the concentration was increased; therefore, the range was kept the same as for a water-miscible solvent The enzyme concentration was increased very slightly since it did not appear that the reaction worked better at the low level. Ranges for the calcium content, and pH were kept the same. The temperatures were increased very slightly.  95 5. R a n d o m Centroid Optimization: Second Cycle The limits for the second cycle were as follows: 1. 2. 3. 4. 5. 6.  Z-L-Asp: 40 - 80mM Met-OMe: 80 - 200mM Solvent-Ethanediol: 0 - 0.4 (0-40%) Enzyme: 10 - 40 fiM Calcium: 2 - 1 0 mM pH: 5 - 7  7. Temperature: 40 - 50°C The number of experiments chosen for the second cycle was 14 (2 x 7 factors). Once these and the centroid experiments were completed, the yields were calculated and results were used for mapping. Z-L-Asp-Met-OMe synthesis was much more successful using ethanediol (Table 12). All of the vertices had a measurable peak area and the yields were much higher than those obtained with glycerol. The highest yield for Z-L-Asp-MetOMe synthesis was 9.13 ± 1.22% (Mean ± S.D., n=3) for Vertex 32 using 58mM Z-L-Asp, 96mM Met-OMe, 2/98 ethanediol/buffer, 35{iM thermolysin, 3.8mM CaCl2, pH 5.1 and at 49°C. This was significantly higher than any of the other vertices. The second highest, Vertex 21, was only 2.59%. This showed that the Random Centroid Optimization can be very useful in obtaining the optimum conditions. However, it should be noted that only one cycle was completed for the synthesis in ethanediol and that the true optimum conditions were not obtained because the Random Centroid Optimization usually requires completion of two cycles. The "optimum" conditions discussed here refer to the best conditions for this cycle. The second cycle would have to be completed in order to obtain the optimum conditions.  96 Table 12. Random Design, Centroid Formulation and Yields for Z-L-Asp-Met-OMe Synthesis in Ethanediol: Second Cycle Met-OMe (mM)  Ethanediol" (mL)  Enz (uM)  Ca (mM)  pH  (mM)  Temp °C  Yield (%)  20  48  178  .40  10  9.8  5.1  48  2.26  21  66  174  .15  28  6.8  6.8  50  2.59  22  76  115  .34  38  3.6  5.8  48  1.20  23  45  136  .08  18  8.2  5.6  45  0.49  24  71  200  .13  28  2.2  6.6  50  2.46  25  48  199  .24  33  8.3  6.5  49  1.91  26  49  115  .05  28  9.2  6.5  45  1.39  27  44  98  .22  13  3.5  5.9  44  0.76  28  69  93  .28  15  6.1  5.0  45  1.11  29  44  82  .10  27  6.8  7.0  40  1.50  30  47  178  .24  40  8.4  6.2  49  1.97  31  76  134  .33  23  9.1  5.5  45  2.89  32  53  96  .02  35  3.8  5.1  49  8.28c  33  66  154  .28  23  4.3  5.9  40  0.42  34  56  151  .18  28  7.4  6.2  47  1.88  35  57  155  .20  28  7.2  6.1  47  1.45  Vertex*  Z-L-Asp  "Vertices 20 to 33 represent the random design; Vertices 34 and 35 were obtained from the centroid formulation "Total volume of reaction = l.OmL; Buffer content is l.OmL - ethanediol content (eg. 1.0 0.28 = 0.72mL) c Actual value for % synthesis calculated using this value and duplicates: 9.13 ± 1.22% (Mean ± S.D., n=3; duplicates were: 8.59% and 10.5%)  97  6. Mapping The results from the mapping were not particularly useful even though there was some joining of points because the lines appeared to be horizontal (Figure 15, 16, 17 and 18). Results were still summarized as follows: The greatest effect on the synthesis of Z-L-Asp-Met-OMe synthesis was the use of ethanediol instead of glycerol, since ethanediol did not inhibit the reaction. In addition, the slight changes in the ranges for the factors also helped. By making Z-L-Asp the limiting reactant and Met-OMe in excess, there was increased yield. Therefore, the equilibrium was successfully shifted towards synthesis, to some extent by manipulating the concentration of the reactants (Kullmann, 1985; Jakubke et al., 1985). However, it did not appear that more Met-OMe resulted in higher synthesis; there was no particular trend, which suggests that the other factors were affecting the synthesis. The best ratio of Z-L-Asp to Met-OMe was 1:1.8, for Vertex 32 ("optimum") and Vertex 31, which had the second highest yield (2.89%). Although it was expected that there would be some trend for the ethanediol concentration, there was no indication whether or not synthesis was better at the low or high concentrations since the highest yield was with a very low concentration (0.02mL) but there was still synthesis at the high level (0.40mL) for Vertex 20. In addition, this was apparent for others such as Vertex 23, which had 0.08mL ethanediol but only 0.49% yield, while Vertex 33 had only 0.42% yield with 0.28mL ethanediol. There was a trend for the enzyme concentration, since most of the vertices that had low enzyme concentration had low yields. In general, enzyme concentrations above 20|iM were more favourable, with Vertex 20 being the exception where there was high yield with low enzyme concentration. Synthesis was not  98 I U  i  •  8  s  i  I  *•••»  6 -  -  4 -  -  J5 "5 Q jg +rf  ^  • 2  -  •  4  •  -  • •  •  •  ••  n  40  • •  • 50  60  70  80  Z-L-Asp Concentration (mM)  100  200  Met-OMe Concentration (mM)  Figure 15. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: A) Z-L-Asp concentration and B) MetOMe concentration  99 10  _•  8 -  S  6h  c w  4 -  w "5  2 -  0.0  -  0.1  .  .  *  0.2  •  0.3  •  0.4  0.5  Ethanediol (mL)  1U  I  I  D  ...... !  •  8 *•%  6 -  -  4  -  9>  "5 £• *• C  >»  (0  2  *  •  •  •  •  • •  A  1  1  10  20  • I  30  40  50  Enzyme tyiM)  Figure 16. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: C) Ethanediol and D) Enzyme concentration  100 IV  i  i  r  i  i  _  -?•  i  --  •  8 #"*  s  6 -  -  4 -  -  0 40  ©  £  +*  ^  w •  •  •  2  • •  o 2  I  3  •  •  •  i  I  5  6  •  0  •  7  8  i  9  10  Calcium QiM)  1-  ... , 8 -  E  1  -  •  6  _  4 -  "  1  • •  2 -  •  • •  •  ""  • A  1  6  •  8  PH Figure 17. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: E) Calcium chloride and F) pH  101  30  40  50  60  Tamperature  Figure 18. Mapping results from the second cycle of experiments of Random Centroid Optimization for Z-L-Asp-Met-OMe: G) Temperature  102 improved by increasing the calcium concentration since the highest yield was at 3.8mM while the second highest yield, Vertex 31 had 9.1mM of calcium. No trends were indicated for the pH, meaning that the reaction was influenced by the combination of factors. The optimum pH for synthesis of Z-L-Asp-Met-OMe was lower (pH 5.1) than for Z-L-Asp-Tyr-OMe (pH 6.3), which shows that it depends on the conditions and substrates used. Both of the pHs for this study are slightly lower than those reported in the literature, although they are still within the range of hydrolysis, which is pH 5-9 (Oyama and Kihara, 1984). Other researchers have found the optimum pH to vary from 6 (Oyama et al., 1981; Nakanishi et al., 1986) up to 7.5 (Riechmann and Kasche, 1986). Cheng et al. (1988) found the optimum pH to be 6.5. Z-LAsp-Met-OMe synthesis was favoured at higher temperatures, because the vertices with the higher yields all had been incubated at temperatures at 45°C or more. However, since there were only three vertices that were less than 45°C, it was difficult to see whether or not there was a difference at the higher temperatures.  7. Quantification of Z-L-Asp-Met-OMe The method by which the Z-L-Asp-Met-OMe yield was calculated differed from that of Z-L-Asp-Tyr-OMe, due to the nature of the formation of product. At first, the sampling of the tube with enzyme for the Z-L-Asp-Met-OMe was the same as was done for the Z-LAsp-Tyr-OMe; the tube was vortexed to ensure a uniform sample, then an aliquot was removed, diluted 10X and subjected to HPLC analysis. This was repeated for all of the vertices. However, since there were a few samples that were opaque and had suspended particulate matter, they were centrifuged down to obtain a pellet, the supernatant was decanted and the precipitate was redissolved in 100% methanol and analyzed. It was found  103 that a larger peak area for the product resulted as compared to the peak areas from the samples that were taken directly from the samples with enzyme. This meant that most of the product was appearing in the precipitate, although there was also some that was dissolved. In order to obtain a representative yield for the product, the sample should be freeze dried so that there would be the total amount of product (the amount suspended and the amount that was precipitated). Thus, the entire second cycle had to be repeated in order to obtain the samples, and aliquots were removed from the reaction and freeze dried. The freeze dried samples were then redissolved in 100% methanol and analyzed by HPLC. Yields calculated from the lyophilized samples were higher than those taken directly from the enzyme containing tube. This procedure was not necessary for the Z-L-Asp-Tyr-OMe synthesis since the product was not present in large quantities in the precipitate nor did it apply for the Z-LAsp-Met-OMe synthesis in glycerol. See Appendix 1 for chromatograms of Z-L-Asp-MetOMe that were obtained from a sample taken directly from the tube and for a freeze dried sample. Although the best yield for Z-L-Asp-Met-OMe was higher than for Z-L-Asp-Tyr-OMe, the rest of the Z-L-Asp-Met-OMe yields were only slightly higher. Based on the optimum yield, it did appear that enzymatic synthesis with Met-OMe was better than that of the TyrOMe, but further studies would have to be done due to the problems with the insolubility of the reactants. According to the specificity of thermolysin, it is favourable to use hydrophobic amino acids such as He, Phe, Val, Leu and Met as the amine component (nucleophile) so it was expected that the enzymatic reaction would be better with Met-OMe (Reddy, 1991). There were a number of reasons why yields for Z-L-Asp-Met-OMe were not as high  104 as those obtained for other thermolysin-catalyzed reactions. It was found on examination of the chromatograms of the products from the freeze dried fractions that there were a number of large peaks other than the product peak. (Refer to Appendix 1). This suggests that a number of side reactions or by-products or other peptide/complex forming reactions were occurring, which would affect product yield. Due to the lack of time, identification of the peaks was not possible. It would be very time consuming to either chemically synthesize possible combinations of peptides or to apply other determination methods such as mass spectrometry. There were 5 distinct unknown peaks in the HPLC chromatograms, two of which occurred before the product peak that were quite large and often were larger than the product peak. These peaks were thought to be a result of methionine oxidation (Clark-Lewis, 1992). Interference from the other peaks would definitely cause lower yields of Z-L-AspMet-OMe. In addition, the ratio of the concentration of the reactants was probably not optimal since the Met-OMe was much higher than the Z-L-Asp and it was unnecessary to have it in such excess. Morihara (1987) stated that lower concentrations of reactants are required in the precipitation reaction as product solubility decreases. This explains why Vertex 32 was the best condition: there was 53mM Z-L-Asp and 96mM Met-OMe in the presence of 35|iM enzyme. Morihara (1987) noted that a high concentration of enzyme is required for peptide synthesis. It was expected that there would be some trend for the amount of solvent required for the synthesis but after optimization (1 cycle) was completed there was no indication of whether or not synthesis was better at the lower or higher levels. A better indication might  105 be obtained if the second cycle of optimization were completed.  Based on those  observations, it may have been possible to use the ethanediol at concentrations >40%, since yields did not really drop off when there was higher solvent levels. It did appear that there may have been little effect of ethanediol on the synthesis since only 2% was required under optimal conditions. However, it is possible that there is an increase in the catalytic activity but this is not known since kinetic studies have not been performed. Fernandez et al. (1991) noted that there was a slight increase in kcat for esterase activity of papain when solvents were increased from 0-15%. This suggests that there is some activation of the enzyme, even at low solvent concentrations. More work has to be done to find out if ethanediol causes any inhibitory effects on thermolysin. This study showed that synthesis is not entirely dependent on one factor alone and that all other contributing factors are important (eg. concentration of reactants, enzyme concentration, pH, temperature). Perhaps the precipitation of the product would have been facilitated if a mixture of solvents was used instead of ethanediol alone. The reaction in ethanediol under the particular conditions did cause precipitation of the product, but not in high amounts except for Vertex 32, which was significantly higher than any other vertex. Product precipitation may have been enhanced by the inclusion of salts such as ammonium sulfate. Isowa and Ichikawa (1979), Cheng et al. (1988) and Miranda and Tominaga (1991) all used ammonium sulfate in thermolysin-catalyzed systems to decrease the solubility of the product and increase yields.  106  CONCLUSIONS Enzymatic peptide synthesis has been suggested as an alternative approach to chemical synthesis because the reaction is stereospecific and racemization is minimized. Application of enzymes for the formation of novel food products is of interest to the food industry. Enzymatic synthesis is a potential tool for the food industry to produce products. One such example is the production of the dipeptide sweetener, aspartame. Enzymatic peptide synthesis may also be applied to catalyze the formation of other dipeptide sweeteners. Two analogs of the precursor to aspartame, Z-L-Asp-Tyr-OMe and Z-L-Asp-Met-OMe were enzymatically synthesized by thermolysin.  Optimal conditions for synthesis were  obtained by using Random Centroid Optimization (RCO), and yields were calculated from the peak areas obtained by HPLC. The optimum yield of Z-L-Asp-Tyr-OMe was 2.13 ± 0.46% with the following conditions:  80mM Z-L-Asp, 154mM Tyr-OMe, 2.3/9.7/88  DMF/Ethyl acetate/buffer, 7.0mM CaCl2, at pH 6.3, 46°C, for 24 hours and catalyzed by 39|iM thermolysin. The best yield for Z-L-Asp-Met-OMe after the completion of 1 cycle of optimization was 9.13 ± 1.22% with the following conditions: 53mM Z-L-Asp, 96mM MetOMe, 2/98 ethanediol/buffer, 3.8mM CaCl2, at pH 5.1, 49°C, for 24 hours and catalyzed by 35\iM thermolysin. Synthesis for Z-L-Asp-Met-OMe was much better than for Z-L-Asp-Tyr-OMe, which was expected since Met-OMe is a more suitable substrate for thermolysin than Tyr-OMe, in terms of specificity. However, there were difficulties with the Z-L-Asp-Tyr-OMe synthesis since the solvent mixture used (dimethylformamide:ethyl acetate) did not solubilize the reactants well, which accounted for the low yields. It is therefore not certain how the Z-L-  107 Asp-Tyr-OMe yields would have been if a different solvent or solvent mixture were used. By applying Random Centroid Optimization, it was possible to increase yields and obtain the conditions for synthesis of both peptides. After the optimization was completed for the Z-L-Asp-Tyr-OMe synthesis, yields were increased from 1.40% in the first cycle to 2.13%, which was not substantially higher but some of the yields in the first cycle had 0% yield. In the case of the Z-L-Asp-Met-OMe, there was a substantial increase in yields due to the change of the solvent in the second cycle from glycerol to ethanediol.  Synthesis with  glycerol was very poor and only three vertices had a measurable yield. The best yield for synthesis with ethanediol was significantly higher than that of any of the other vertices, but another cycle would have to be completed to obtain the optimum. Although it was evident that the organic solvents used affected synthesis, there was no indication of which other factors were important because the mapping results from RCO were not very conclusive. However, it was evident that there were influences other than the solvent effect since some results could not be explained by the solvent alone.  Some  promotion of synthesis in the presence of organic solvents was expected, and did occur but unfortunately in both cases, increasing the solvent level did not improve synthesis. Usually, in enzymatic peptide synthesis, organic solvents are used to limit water and to cause an equilibrium shift in favour of formation of the peptide. Additional experiments showed that even if a solvent may be suitable in solubilizing the reactants, other studies must be done that show its effect on the enzyme. Enzyme kinetic studies would provide the most information about the effect of solvents on enzymes and enable better selection of solvents for enzymatic peptide synthesis.  108 FUTURE DIRECTIONS Some suggestions for future studies are: A.  To do more extensive studies on the effect of organic solvents on thermolysin. At present, there is little information available in the literature for the effect of organic solvents on thermolysin, although there is much more for a-chymotrypsin.  Not  enough information can be obtained from examining solvent suitability alone. This would require enzyme kinetic studies and could be very useful for future applications of thermolysin in enzymatic synthesis. B.  To include ammonium sulfate salts (or sodium chloride) in order to increase the yield of the product In some cases, it is difficult to obtain precipitation of the product since some of it remains in solution.  The addition of salts would reduce the  solubility of the product and facilitate the precipitation reaction. However, the amount of salts to be added to the system should be optimized since there is a critical point at which the solubility of the substrates will be affected. In addition, the enzyme activity may also be affected by high concentrations of salts. C.  To test other enzymes such as papain which may be suitable for synthesis of Z-L-AspTyr-OMe or Z-L-Asp-Met-OMe. Papain has already been used for Z-L-Asp-Phe-OMe synthesis, and since it has a similar specificity to thermolysin, it is probably a suitable alternative.  Other protecting groups might be attempted for papain-catalyzed  synthesis, such as Boc (tert-butyloxycarbonyl) or Moz [(p-methoxybenzyl)oxy]carbonyl (Chen and Wang, 1988).  109 D.  To use mixtures of solvents at higher concentrations for synthesis. The main problem with this study was that the amount of organic solvent used was very small and hydrolysis of the product probably occurred. For the Z-L-Asp-Met-OMe synthesis, perhaps a mixture of ethanediol with another solvent would be better since ethanediol would act as a water mimic and partially substitute the water in the system, as Kitaguchi and Klibanov (1989) found for oligopeptide synthesis.  110  REFERENCES Andrews, A.T. and Alichanidis, E. 1990. The plastein reaction revisited: evidence for a purely aggregation reaction mechanism. Food Chem. 35:243. Antonini, E., Carrea, G., and Cremonesi, P. 1981. Enzyme catalysed reactions in waterorganic solvent two-phase systems. Enz. Microb. Technol. 3:291. Arai, S. and Fujimaki, M. 1978. The plastein reaction: theoretical basis. Ann. Nutr. Alim. 32:701. Arakawa, T., and Goddette, D. 1985. The mechanism of helical transition of protein by organic solvents. Arch. Biochem. Biophys. 240:21. Aso, K., Uemura, T., and Shiokawa, Y. 1988. Protease-catalyzed synthesis of oligo-Lglutamic acid from L-glutamic acid diethyl ester. Agric. Biol. Chem. 52:2443. Aso, K. 1989. Enzymatic approach to the synthesis of a lysine-containing sweet peptide, Nacetyl-L-phenylalanyl-L-lysine. Agric. Biol. Chem. 53:729. Auriol, D., Paul, F., and Monsan, P. 1990. A new peptidase enzyme useful for peptide and amino acid derivative synthesis by reverse hydrolysis reactions. Ann. N.Y. Acad. Sci. 613:201. Brenner, M., Niederwieser, A., and Pataki, G. 1969. Thin-Layer Chromatography Laboratory Handbook, pp. 740-755. RStafl (Ed.), Springer Verlag, NY. Brink, L.E.S., Tramper, J., Luyben, K.Ch.A.M., and Van't Riet, K. 1988. Biocatalysis in organic media. Enz. Microb. Technol. 10:736. Butler, L.G. 1979. Enzymes in non-aqueous solvents. Enz. Microb. Technol. 1:253. Carrea, G. 1984. Biocatalysis in water-organic solvent two-phase systems. Trends Biotechnol. 2:102. Cassells, J.M. and Hailing, P.J. 1990. Protease-catalysed peptide synthesis in aqueous-organic two-phase systems: reactant precipitation and interfacial inactivation. Enz. Microb. Technol. 12:755. Chaiken, I.M., Komoriya, A., Ohno, M., and Widmer, F. 1982. Use of enzymes in peptide synthesis. Appl. Biochem. Biotechnol. 7:385.  Ill Chen, S.T. and Wang, K.T. 1988. Papain-catalyzed synthesis of aspartame precursor in biphasic system. J. Org. Chem. 53:4589. Cheng, E., Miranda, M.T.M. and Tominaga, M. 1988. Thermolysin and a-chymotrypsin mediated synthesis of tripeptides containing proline. Int. J. Peptide Prot. Res. 31:116. Church, F.C., Swaisgood, H.E., Porter, D.H. and Catignani, G.L. 1983. Spectrophotometric assay using o-phthaldialdehyde for determination of proteolysis in milk and isolated milk proteins. J. Dairy Sei. 66:1219. Clapes, P., Adlercreutz, P. and Mattiasson, B. 1990a. Enzymatic peptide synthesis in organic media: a comparative study of water-miscible and water-immiscible solvent systems. J. Biotechnol. 15:323. Clapes, P., Adlercreutz, P. and Mattiasson, B. 1990b. Enzymatic peptide synthesis in organic media: nucleophile specificity and medium engineering in a-chymotrypsin-catalyzed reactions. Biotechnol. Appl. Biochem. 12:376. Clapes, P. and Valencia, G. 1992. Influence of solvent and water activity on kinetically controlled peptide synthesis. Enz. Microb. Technol. 14:575. Clark-Lewis, I. 1992. Personal communication. Crook, E.M., Mathias, A.P., and Rabin, B.R. 1960. Spectrophotometric assay of bovine pancreatic ribonuclease by the use of cytidine 2',3'-phosphate. Biochem J. 74:234. Didziapetris, R., Drabnig, B., Schellenberger, V., Jakubke, H-D., Svedas, V. 1991. Penicillin acylase-catalyzed protection and deprotection of amino groups as a promising approach in enzymatic peptide synthesis. FEBS Lett. 287:31. Dordick, J.S. 1989. Enzymatic catalysis in monophasic organic solvents. Enz. Microb. Technol. 11:194. Dordick, J.S. 1991. Principles and Applications of Nonaqueous Enzymology. Chpt. 1. In Applied Biocatalvsis Volume 1, pp. 1-51. H.W. Blanch and D.S. Clark (Eds.) Marcel Dekker Inc., NY Durrant, I., Beynon, R.J., and Rodgers, P.B. 1986. Effect of glycerol on thermolysincatalyzed peptide bond synthesis. Arch. Biochem. Biophys. 250:280. Edwards, J.H. and Shipe, W.F. 1978. Characterization of plastein reaction products formed by pepsin, a-chymotrypsin, and papain treatment of egg albumin hydrolysates. J. Food Sci. 43:1215.  112 Eggers, D.K., Blanch, H.W., and Prausnitz, J.M. 1989. Extractive catalysis: solvent effects on equilibria of enzymatic reactions in two-phase systems. Enz. Microb. Technol. 11:84. Fastrez, J. and Fersht, A.R. 1973. Demonstration of the acyl-enzyme mechanism for the hydrolysis of peptides and anilides by chymotrypsin. Biochemistry 12:2025. Ferjancic, A., Puigserver, A. and Gaertner, H. 1988. Unusual specificity of polyethylene glycol-modified thermolysin in peptide synthesis catalyzed in organic solvents. Biotechnol. Lett. 10:101. Fernandez, M.M., Clark, D.S. and Blanch, H.W. 1991. Papain kinetics in the presence of a water-miscible organic solvent. Biotechnol. Bioeng. 37:967. Fried, B. and Sherma, J. 1982. Application of samples and basic TLC design. Chpt 5. In Thin-Layer Chromatography, pp. 67-76. Marcel Dekker, Inc., NY. Fruton, J.S. 1982. Proteinase-catalyzed synthesis of peptide bonds. Adv. Enzymol. 53:239. Fruton, J.S. 1984. Pepsin and thermolysin as catalysts of peptide bond synthesis. Carlsberg Res. Commun. 49:231. Fujimaki, M., Yamashita, M., Arai, S., and Kato, H. 1970. Enzymatic modification of proteins in foodstuffs. Part I. Agr. Biol. Chem. 34:1325. Gaertner, H., and Puigserver, A. 1989. Kinetics and specificity of serine protease in peptide synthesis catalyzed in organic solvents. Eur. J. Biochem. 181:217. Gololobov, M.Y., Antonova, T.V. and Belikov, V.M. 1986. Transpeptidation as the main chemical reaction causing the formation of plasteins. Nahrung. 30:289. Gorman, L.A.S. and Dordick, J.S. 1992. Organic solvents strip water off enzymes. Biotechnol. Bioeng. 39:392. Gross, A. 1991. Enzymatic catalysis in the production of novel food ingredients. Food Technol. 45(1):96. Gross, E., and Meienhoffer, J. (Eds.) 1979. In The Peptides: Analysis, Synthesis, Biology. Volume 1, pp. xiii-xvii. Academic Press, NY. Gupta, M.N. 1992. Enzyme function in organic solvents. Eur. J. Biochem. 203:25. Hailing, P.J. 1990. Solvent selection for biocatalysis in mainly organic solvents: predictions of effects on equilibrium position. Biotechnol. Bioeng. 35:691.  113 Hartnett, E.K. and Satterlee, L.D. 1990. The formation of heat and enzyme induced (plastein) gels from pepsin-hydrolyzed soy protein isolate. J. Food Biochem. 14:1. Hernandez, M.J.M., Camanas, R.M.V. and Alvarez-Coque, M.C.G. 1990. Determination of total free amino acids with o-phthalaldehyde and N-acetyl-L-cysteine. Microchem. J. 42:288. Homandberg, G.A., Mattis, J.A., and Laskowski, M.(Jr.) 1978. Synthesis of peptide bonds by proteinases. Addition of organic cosolvents shifts peptide bond equilibria toward synthesis. Biochemistry 17:5220. Homandberg, G.A. and Laskowski, M.(Jr.) 1978. Enzymatic resynthesis of the hydrolyzed peptide bond(s) in ribonuclease S. Biochemistry 18:586. Ingalls, R.G., Squires, R.G., and Butler, L.G. 1975. Reversal of enzymatic hydrolysis: rate and extent of ester synthesis as catalyzed by chymotrypsin and subtilisin Carlsberg at low water concentrations. Biotechnol. Bioeng. 17:1627. Inouye, K., Watanabe, K. and Tochino, Y. 1981. Semisynthesis and properties of some insulin analogs. Biopolymers. 20:1845. Isowa, Y., Ohmori, M., Ichikawa, T. and Mori, K. 1979. The thermolysin-catalyzed condensation reaction of N-substituted aspartic and glutamic acids with phenylalanine alkyl esters. Tetrahedron Lett. 28:2611. Isowa, Y. and Ichikawa, T. 1979. Syntheses of N-acyl dipeptide derivatives by metalloproteinases. Bull. Chem. Soc. Jpn. 52:796. Jakubke, H-D., Kuhl, P., Konnecke, A., Doring, G., Walpuski, J., Wiledorf, A., and Zapevalova, N.P. 1982. Further application of the aqueous-organic two-phase approach to enzyme-catalyzed peptide synthesis. In Peptides 1982. Proceedings of the 17th European Peptide Symposium, p.43-54. K.Blaha and P. Malon (Eds.), Walter de Gruyter, NY. Jakubke, H-D., Kuhl, P. and Konnecke, A. 1985. Basic principles of protease-catalyzed peptide bond formation. Angew. Chem. Int. Ed. Eng. 24:85. Jakubke, H-D. 1987. Enzymatic Peptide Synthesis. Chpt. 3. In The Peptides, Volume 9 pp. 103-165. S. Udenfriend and J. Meinenhofer (Eds.), Academic Press, Inc., NY. Janusz, J.M. 1989. Peptide Sweeteners Beyond Aspartame. Chpt.1. In Progress in Sweeteners, pp. 1-46. T.H. Grenby (Ed.), Elsevier Applied Science, London, UK. Juillerat, M. and Homandberg, G.A. 1981. Clostripain-catalyzed re-formation of a peptide bond in a Cytochrome C fragment complex. Int. J. Peptide Prot. Res. 18:335.  114 Kasche, V. 1986. Mechanism and yields in enzyme catalysed equilibrium and kinetically controlled synthesis of pMactam antibiotics, peptides and other condensation products. Enz. Microb. Technol. 8:4. Kester, W.R. and Matthews, B.W. 1977. Crystallographic study of the binding of dipeptide inhibitors to thermolysin: implications for the mechanism of catalysis. Biochemistry 16:2506. Khmelnitsky, Yu L., Dien, P.K., Semenov, A.N. and Martinek, K. 1984. Optimization of enzyme catalyzed peptide synthesis in a "water-water-immiscible organic solvent" biphasic system. Tetrahedron 40:4425. Khmelnitsky, Yu L., Levashov, A.V., Klyachko, N.L., and Martinek, K. 1988. Engineering biocatalytic systems in organic media with low water content. Enz. Microb. Technol. 10:710. Khmelnitsky, Yu L., Mozhaev, V.V., Belova, A.B., Sergeeva, M.V. and Martinek, K. 1991. Denaturation capacity: a new quantitative criterion for selection of organic solvents as reaction media in biocatalysis. Eur. J. Biochem. 198:31. Kim, S.H. and Dubois, G.E. 1991. Natural High Potency Sweeteners. Chpt.6. In Handbook of Sweeteners, pp.116-178. S. Marie and J.R. Piggott (Eds.), Blackie and Son Ltd., Bishopbriggs, Glasgow. Kimura, Y., Nakanishi, K. and Matsuno, R. 1990. Enzymatic synthesis of the precursor of Leu-enkephalin in water-immiscible organic solvent systems. Enz. Microb. Technol. 12:272. Kitaguchi, H. and Klibanov, A.M. 1989. Enzymatic peptide synthesis via segment condensation in the presence of water mimics. J. Am. Chem. Soc. 111:9272. Klibanov, A.M., Samokhin, G.P., Martinek, K., and Berezin, I.V. 1977. A new approach to preparative enzymatic synthesis. Biotechnol. Bioeng. 19:1351. Kuhl, P., Konnecke, A., Doling, G., Daumer and Jakubke, H.D. 1980. Enzyme-catalyzed peptide synthesis in biphasic aqueous-organic systems. Tetrahedron Lett. 21:893. Kuhl, P., Schaaf, R. and Jakubke, H-D. 1987. Studies on enzymatic peptide synthesis in biphasic aqueous-organic systems with product extraction. Monatsh. Chem. 118:1279. Kullmann, W. 1985. Proteases as catalytic agents in peptide synthetic chemistry: shifting the extent of peptide bond synthesis from a "Quantity N6gligeable" to a "Quantite Considerable". J. Prot Chem. 4:1.  115 Kwan, K.K.H., Nakai, S., and Skura, B J . 1983. Comparison of four methods for determining protease activity in milk. J. Food. Sci. 48:1418. Laane, C , Boeren, S., Vos, K. and Veeger, C. 1987. Rules for optimization of biocatalysis in organic solvents. Biotechnol. Bioeng. 30:81. Latt, S., Holmquist, B. and Vallee, B.L. 1969. Thermolysin: a zinc metalloenzyme. Biochem. Biophys. Res. Commun. 37:333. Lilly, M.D. 1982. Two-liquid-phase biocatalytic reactions. 32:162.  J. Chem. Tech. Biotechnol.  Lozano, P. and Combes, D. 1991. a-Chymotrypsin in plastein synthesis: influence of substrate concentration on enzyme activity. Biotechnol. Appl. Biochem. 14:212. Malak, C.A.A., Filippova, I.Y., Lysogorskaya, E.N., Anisimova, V.V., Lavrenova, G.I. and Stepanov, V.M. 1992. Pepsin as a catalyst of peptide synthesis. Int. J. Peptide ProL Res. 39:443. Margolin, A.L., Tai, D-F, and Klibanov, A.M. 1987. Incorporation of D-amino acids into peptides via enzymatic condensation in organic solvents. J. Am. Chem. Soc. 109:7885. Martinek, K. and Semenov, A.N. 1981a. Enzymes in organic synthesis: physicochemical means of increasing the yield of end product in biocatalysis. J. Appl. Biochem. 3:93. Martinek, K. and Semenov, A.N. 1981b. Enzymatic synthesis in biphasic aqueous-organic systems. II. Shift of ionic equilibria. Biochim. Biophys. Acta. 658:90. Martinek, K., Semenov, A.N., and Berezin, I.V. 1981. Enzymatic synthesis in biphasic aqueous-organic systems. I. Chemical equilibrium shift. Biochim. Biophys. Acta. 658:76. Matsubara, H. 1970. Purification and assay of thermolysin. Methods Enzymol. 19:642. Mazur, R.H., Schlatter, J.M., and Goldkamp, A.H. 1969. Structure-taste relationships of some dipeptides. J. Amer. Chem. Soc. 91:2684. Miranda, M.T.M., Cheng, E., Muradian, J., Seidel, W.F., and Tominaga, M. 1986. Thermolysin as a catalyst in enzymatic synthesis of asparagine-containing peptides. Bioorg. Chem. 14:182. Miranda, M.T.M, and Tominaga, M. 1991. Thermolysin as a catalyst in enzymatic synthesis of asparagine-containing peptides II. Int. J. Peptide Prot. Res. 37:128.  116 Mitin, Y.V., Zapevalova, N.P. and Gorbunova, E.Y. 1984. Peptide synthesis catalyzed by papain at alkaline pH values. Int. J. Peptide Prot. Res. 23:528. Morihara, K. 1974. Comparative specificity of microbial proteinases. Adv. Enzymol. 41:179. Morihara, K. 1987. Using proteases in peptide synthesis. TEBTECH. 5:164. Morihara, K. and Oka, T. 1977. a-Chymotrypsin as the catalyst for peptide synthesis. Biochem T 163:531. Motihara, K., Tsuzuki, H. and Oka, T. 1978. Acyl and amino intermediates in reactions catalyzed by thermolysin. Biochem. Biophys. Res. Commun. 84:95. Mozhaev, V.V., Khmelnitsky, Y.L., Sergeeva, M.V., Belova, A.B., and Klyachko, N.L. 1989. Catalytic activity and denaturation of enzymes in water/organic cosolvent mixtures, achymotrypsin and laccase in mixed water/alcohol, water/glycol and water/formamide solvents. Eur. J. Biochem. 184:597. Nakai, S. 1984. A new mapping super-simplex optimization for food product and process development. J. Food. Sci. 49:1143. Nakai, S. 1990. Computer-aided optimization with potential application in biorheology. J. Jpn. Soc. Biorheology 4:143. Nakanishi, K., Kamikubo, T. and Matsuno, R. 1985. Continuous synthesis of N-(benzyloxycarbonyl)-L-asparryl-L-phenylalanine methyl ester with immobilized thermolysin in an organic solvent. Bio/Technol. 3:459. Nakanishi, K., Kimura, Y., and Matsuno, K. 1986. Kinetics and equilibrium of enzymatic synthesis of peptides in aqueous/organic biphasic systems: thermolysin-catalyzed synthesis of N-(benzyloxycarbonyl)-L-aspartyl-L-phenylalanine methyl ester. Eur. J. Biochem. 161:541. Nakanishi, K. and Matsuno, R. 1988. Recent developments in enzymatic synthesis of peptide. Adv. Biotechnol. Proc. 10:173. Oka, T. and Morihara, K. 1977. Trypsin as a catalyst for peptide synthesis. J. Biochem. 82:1055. Oka, T. and Morihara, K. 1978. Peptide bond synthesis catalyzed by a-chymotrypsin. J. Biochem. 84:1277. Oka, T. and Morihara, K. 1980. Peptide bond synthesis catalyzed by thermolysin. J. Biochem. 88:807.  117 Ooshima, H., Mori, H. and Harano, Y. 1985. Synthesis of aspartame precursor by solid thermolysin in organic solvent. Biotechnol. Lett. 7:789. Oyama, K., Kihara, K., and Nonaka, Y. 1981. On the mechanism of the action of thermolysin: kinetic study of the thermolysin-catalyzed condensation reaction of N-Benzyloxylcarbonyl-L-aspartic acid with L-phenylalanine methyl ester. J. Chem. Soc., Perkin Trans. H. 2:356. Oyama, K. and Kihara, K. 1984. A new horizon for enzyme technology. Chemtech. 2:100. Oyama, K., Irino, S. and Hagi, N. 1987. Production of aspartame by immobilized thermoase. Methods Enzymol. 136:503. Petkov, D.D. 1982. Enzyme peptide synthesis and semisynthesis: kinetic and thermodynamic aspects. J. Theor. Biol. 98:419. Petkov, D.D. and Stoineva, LB. 1984. Enzyme peptide synthesis by an iterative procedure in a nucleophile pool. Tetrahedron Lett. 25:3751. Reddy, A.V. 1991. Thermolysin: a peptide forming enzyme. Indian J. Biochem. Biophys. 28:10! Reslow, M., Adlercreutz, P., and Mattiasson, B. 1988. The influence of water on proteasecatalyzed peptide synthesis in acetonitrile/water mixtures. Eur. J. Biochem. 177:313. Riechmann, L. and Kasche, V. 1986. Reaction mechanism, specificity and pH-dependence of peptide synthesis catalyzed by the metalloproteinase thermolysin. Biochim. Biophys. Acta. 872:269. Rose, K., Herrero, C , Proudfoot, A.E.I., Wallace, C.J.A., and Offord, R.E. 1987. Enzymeassisted semisynthesis of polypeptide active esters for subsequent spontaneous coupling. In Peptides 1986 pp.219-222. D. Theodoropoulos (Ed.) Walter de Gruyter and Co. NY. Sakina, K., Kawazura, K., Morihara, K., and Yajima, H. 1988. Thermolysin-catalyzed synthesis of peptide amides. Chem. Pharm. Bull. 36:4345. Sakurai, T., Margolin, A.L., Russell, AJ. and Klibanov, A.M. 1988. Control of enzyme enantioselectivity by the reaction medium. J. Am. Chem. Soc. 110:7236. Sarin, V.K., Kent, S.B.H., Tam, J.P. and Menifield, R.B. 1981. Quantitative monitoring of solid-phase peptide synthesis by the ninhydrin reaction. Anal. Biochem. 117:147. Sawyalow, W.W. 1901. Pfluger's Arch. Ges. Physiol. 85:171. (In Jakubke, H-D., 1987. Enzymatic Peptide Synthesis. Chpt3. In The Peptides, Vol.9, p. 106. Academic Press, NY.)  118 Schechter, I. and Berger, A. 1967. On the size of the active site in proteinases. I. Papain. Biochem. Biophys. Res. Commun. 27:157. Schellenberger, V., Jakubke, H-D., Zapevalova, N.P. and Mitin, Y.V. 1991. Protease-catalyzed peptide synthesis using inverse substrates: the influence of reaction conditions on the trypsin acyl transfer efficiency. Biotechnol. Bioeng. 38:104. Schwarz, A., Steinke, D., Kula, M-R., and Wandrey, C. 1990. Optimization of enzymemediated peptide bond formation. Biotechnol. Appl. Biochem. 12:188. Sober, H.A. and Harte, R.A.(Eds.) 1970. CRC Handbook of Biochemistry. 2nd Edition, p.C87. The Chemical Rubber Co., Cleveland, Ohio. Stevenson, D.E. and Storer, A.C. 1991. Papain in organic solvents: determination of conditions suitable for biotcatalysis and the effect on substrate specificity and inhibition Biotechnol. Bioeng. 37:519. Stewart, J.M. and Young, J.D. 1984. The Chemistry of Solid Phase Peptide Synthesis. Chpt 1. pp. 1-51. In Solid Phase Peptide Synthesis, 2nd Edition, Pierce Chemical Co., Rockford, IL. Sukan, G. and Andrews, A.T. 1982. Application of the plastein reaction to caseins and to skim-milk powder, n. Chemical and physical properties of the plasteins and the mechanism of the plastein reaction. J. Dairy Res. 49:279. Touchstone, J.C. and Dobbins, M.F. 1983. Preparation and application of the sample. Chpt. 4. pp.73-100. In Practice of Thin Layer Chromatography, John Wiley and Sons Inc., NY. Tsuzuki, H., Oka, T., and Morihara, K. 1980. Coupling between Cbz-Arg-OH and Leu-X catalyzed by trypsin and papain. J. Biochem. 88:669. van Erp, S.H.M., Kamenskaya, E.O., and Khmelnitsky, Y.L. 1991. The effect of water content and nature of organic solvent on enzyme activity in low water media. Eur. J. Biochem. 202:379. von Hofsten, B. and Lalasidis, G. 1976. Protease-catalyzed formation of plastein products and some of their properties. J. Agric. Food Chem. 24:466. Watanabe, M. and Arai, S. 1988. The Plastein Reaction and Its Applications. Chpt 5. In Developments in Food Proteins - 6, pp. 179-217. BJ.F. Hudson (Ed.), Elsevier Applied Science, UK. Wayne, S.I. and Fruton, J.S. 1983. Thermolysin-catalyzed peptide bond synthesis. Proc. Natl. Acad. Sci. USA. 80:3241.  119 Yamashita, M , Arai, S., and Fujimaki, M. 1976. Plastein reaction for food protein improvement. J. Agric. Food Chem. 24:1100. Yang, C-P., Su, C-S. 1986. Effects of solvents and additives on the reaction of Nbenzyloxycarbonyl-L-aspartic anhydride with L-phenylalanine methyl ester (synthesis of aspartame). J. Org. Chem. 51:5186. Zaks, A. and Klibanov, A.M. 1986. Substrate specificity of enzymes in organic solvents vs. water is reversed. J. Am. Chem. Soc. 108:2767. Zaks, A. and Klibanov, A.M. 1988a. Enzymatic catalysis in nonaqueous solvents. J. Biol. Chem. 263:3194. Zaks, A. and Klibanov, A.M. 1988b. The effect of water on enzyme action in organic media. J. Biol. Chem. 263:8017. Zaks, A. and Russell, AJ. 1988. Enzymes in organic solvents: properties and applications. J. Biotechnol. 8:259.  120  APPENDIX 1. HPLC CHROMATOGRAMS OF ENZYMATICALLY SYNTHESIZED SWEET DIPEPTIDES  121 Absorbance J  0.5000 I  I  L.  20.00  40.00 ~  49.25  60.00 -  65.52  ^ 0.0000  0.5000  Absorbance  1.0000  Figure 19. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Phe-OMe using 80mM Z-L-Asp and 80mM Phe-OMe as the carboxyl and amino components, respectively.  122  Absorbance 0.0000 .00  I  I  '  0.5000 I  l  I  1.0000 I  L  •  L  J  I  20.00  22.79  Phe-OMe  30.84 Z - L - A s p  40.00 ~  48.46; 49.09  Z-L-Asp-Phe-OMe  60.00 ~  65.52  ~1 0.0000  "I  '  '  '  1 0.5000  '  '  '  '  1 1.0000  '  '  '  '  Absorbance  Figure 20. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Phe-OMe using 80mM Z-L-Asp and 200mM Phe-OMe as the carboxyl and amino components, respectively.  123  Absorbance 0.00  i  0.5000 I  i  I  I  I  1.0000 1  1  1  1  1.5000 1  1  wm*  18.04; 18.15; 18.24  Tyr-OMe  20.00 ~  30.70  Z-L-Asp  40.00 -  40.97  Z-L-Asp-Tyr-OMe  60.00 -  65.52  ~| 0.0000  '  "  '  "  1 0.5000  <  «  '  Absorbance  '  I 1.0000  '  r  ""  1 1.5000  Figure 21. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Tyr-OMe under optimal conditions from the second cycle of RCO (Vertex 27).  Absorbance 0.0000 0.00 I  i  ^  i  i  •  0.5000 I  i  i  i  1  1.0000 1  J  L  I  16.13; 16.43; 16.67  Met-OMe 28.13; 30.78  Z-L-Asp 43.64  #  65.52  *T 0.0000  l  1  Z-L-Asp-Met-OMe  1  1  1 0.5000  1  1  1  1  1 1.0000  1  1  r  Absorbance  Figure 22. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe under optimal conditions in glycerol (First cycle of RCO; Vertex 10).  125 Absorbance .0000 0.00  j  0.5000 L_  1.0000 i  i  i  J  -  I  I  I  H.45 Met-OMe  20.00 _  30.69  Z-L-Asp  40.00  43.58 Z-L-Asp-Met-OMe  60.00 -  65.52  T 0.0000  T  r  ">  [—  0.5000  ""  1  «•  -j  1  1.0000  Absorbance  Figure 23. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe of the suspended sample under optimal conditions in ethanediol (Second cycle of RCO; Vertex 32).  126 Absorbance 0.0000 i  i  I  I  0.5000 I  I  I  l_  1.0000 I  '  •  •  I  1.5000 I  Met-OMe  20.00  Z-L-Asp  40.00 ~  42.92; 43.52 Z-L-Asp-Met-OMe  T 60.00 -  65.52  H 0.0000  *  — I —  1 — .5000  1.0000  T  r  -1 1.5000  Absorbance  Figure 24. HPLC chromatogram of enzymatically synthesized Z-L-Asp-Met-OMe of the freeze dried sample under optimal conditions in ethanediol (Second cycle of RCO; Vertex 32).  127  APPENDIX 2. GLOSSARY  128  GLOSSARY This is a list of the abbreviations used in the thesis (Gross and Meinenhoffer, 1979). ABBREVIATION Ac  acetyl  Boc  tert-butyloxycarbonyl  CBZ (which is the same as Z )  carboxybenzyl benzyloxycarbonyl  DCC  dicyclohexylcarbodiimide (method)  DMF  dimethylformamide  EA  ethyl acetate  Moz  4-methyloxybenzyloxycarbonyl  OBzl  benzyl ester  OEt  ethyl ester  OMe  methyl ester  TFA  trifluoroacetic acid  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098875/manifest

Comment

Related Items