3130131.:L. It. liavtn‘ 33!)..13': \, /I//I/////I////II/lI///l/ll7I7/llll/llfil7illllll This is to certify that the thesis entitled Optimization of a Flow Injection Indicator Reaction with Applications in Determining Sugars in Fruits. presented by Pavlos Aspris has been accepted towards fulfillment of the requirements for Master of-Seienee——d%¥flflfl—Feeé—Seience and Hum. Nutrition gazed (Win/5% Major professor Date 5/7///9‘%€ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution r #1 LIBRARY Michigan State University is - , v—w PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution c:\circ\datedue.pm3-p.1 It» OPTIMIZATION OF A FLOW INJECTION INDICATOR REACTION WITH APPLICATIONS IN DETERMINING SUGARS IN FRUITS BY Pavlos Aspris A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 1990 . x ' / ,. ‘ ) A».."../}’ "5 ,/. , . ABSTRACT OPTIMIZATION OF A FLOW INJECTION INDICATOR REACTION WITH APPLICATIONS IN DETERMINING SUGARS IN FRUITS BY Pavlos Aspris In this research a new dye, leucomalachite green (LMG), has been introduced for the determination of glucose by flow injection analysis (FIA) using immobilized glucose oxidase. By' means of' general univariate and. computerized simplex optimization procedures, nine analytical parameters affecting the indicator reaction and the FIA system. were optimized. Through a specific optimization procedure, four crucial analytical parameters were further optimized. The optimum conditions obtained were: flow rate, 1.47 ml/min; 0.1 M phosphate buffer, pH 6.0, as a carrier; peroxidase activity, 143 units per 10 ml of reagent; concentration of LMG reagent solution, 9.09 x 10—2 mM; concentration of LMG stock solution, 1.5 mM; 0.1 M phosphate buffer, pH 6.0, 2.4 ml per 10 ml reagent solution; 0.1 M acetate buffer, pH 4.0, 7 ml per 10 m1 reagent solution; length of enzyme single bead string reactor (SBSR), 11.6 cm; length of unmodified SBSR reactor, 35 cm; and temperature, 40 oC. The optimized FIA procedure was applied to the analysis of glucose directly, and fructose and sucrose after proper enzyme conversions to glucose. The results of the FIA analysis of 26 fruit samples for these sugars were compared with the results obtained by conventional sugar enzyme analysis and found to be in good agreement. To my parents, and brother, George. iv ACKNOWLEDGMENT I am extremely indebted to Professors Pericles Markakis and Stanley Crouch for their guidance, enthusiasm, encouragement, and understanding throughout the course of my research effort. Their willingness to give me the freedom to do what I want, but offering help when needed, built my confidence and independence in doing research. I would also like to thank the other members of my M.S. committee: Dr. Denise Smith and Dr. William Helferich, who served as my second readers. Sincere appreciation and 'thanks also are extended. to Dr. Amy Iezzoni, Department of Horticulture, MSU, and the Cherry Marketing Institute for their' partial financial support to» me as a graduate research assistant. Then, I wish to thank Dr. Milt and Stella Karayanni, Department of Chemistry, Ioannina, Greece, for their help and guidance in the beginning of my research project and for giving the idea of using Leucomalachite Green as the new dye for the indicator reaction of the Flow Injection Analysis system. Next, I would like to thank Dr. Saha, Michigan Biotechnology Institute, for his advise in the procedure of the glucose isomerase reaction. Also, the Finnsugar Biochemicals, Inc, chemical Company, IL, for providing free sample of glucose isomerase. I would also like to thank all of Dr. Crouch's group members, who have been very cooperative while I worked amongst them; especially Stephen Medlin who always there to help with computer problems. vi I would like to thank Dr. Petros Charalampous , Cypriot Department of Agriculture, and Mr. Tasos Anastasios for providing olive samples from Cyprus and California, respectively. Last but not least I am most grateful to my parents and my brother, George, for their moral and financial support even from so far away, all these years. TABLE OF CONTENTS CHAPTER Page DEDICATION........................................iv ACKNOWLEDGEMENTS...................................v LIST OF PIGURES...................................xi LIST OF TABLES....................................xv I. ImODUCTIONeeeoeeeoeeoeeeoeeoeeeeeeeeooeeeeoeeeeeel II. BACKGROUND INFORMATION.............................3 A. Methods of Sugars Analysis in Fruits.. ........ ..3 l. Non-Specific and Specific Methods............3 2. Comparison of HPLC with FIA..................4 B. Introduction to Continuous Flow Methods.........7 C. Principles and Function of Flow Injection Analysis........................................8 1. Definition..................................10 2. Essential Features..........................10 3. Dispersion..................................ll 4. Types of FIA Manifolds......................12 D. Use of Immobilized Enzymes in FIA..............15 1. Reversible Immunological Immobilization.....16 2. Advantages of Immobilized Enzymes...........18 3. Applications of Immobilized Enzymes.........18 (Continued) vii (Continued) TABLE OF CONTENTS CHAPTER Page III. METHODS AND MATERIALS.............................20 A. Methodology of FIA for Sugars.................20 1. Apparatus...................................20 2. Reagents....................................27 3. Preparation of the samples............ ...... 28 4. Procedure...................................29 a. Glucose Analysis........................29 b. Fructose Analysis.......................30 c. Sucrose Analysis........................30 B. Conventional Enzyme Methods for Sugar Analysis.31 1. Apparatus...................................31 2. Reagents....................................31 3. Preparation of the samples..................32 4. Procedure........................... ........ 32 a. Determination of D-glucose before inversion................................33 b. Determination of D-fructose..............33 c. Determination of sucrose.................33 (Continued) viii CHAPTER (Continued) TABLE OF CONTENTS Page Iv. RESULTSANDDISCUSSION.0.00.00.00.00000.0.0000000035 A. Optimization of FIA............................35 1. Initial Optimization of the Indicator Reaction....................................35 a. Concentration of LMG Stock Solution......40 b. Concentration of LMG Reagent Solution....47 c. Activity of Peroxidase...................47 d. pH of LMG Stock Solution.................49 e. General Simplex Optimization of FIA......51 2. Specific Optimization of Rate-Dependent VariableSOOOOOOOOOOOO...00.00.00.0000000000057 a. Univariate MethOdSOOOOOOOOOOOOOO0.0.0....58 ii. iii. iv. Flow Rate.............................58 Length of Enzymatic SBSR Reactor......59 Length of Unmodified SBSR Reactor.....63 Flow Rate-Length of Unmodified Reactor...............................63 pH of Carrier.........................67 (Continued) ix (Continued) TABLE OF CONTENTS CHAPTER Page vi. Comparison of Different Lengths of SBSRs for Sensitivity.................67 vii. Sensitivity for Different Lengths of Enzymatic SBSR........................70 viii. Effect of Temperature.................70 b. Specific Simplex Optimization of FIA.....77 c. Evaluation of the optimization methods...8l B. Application of FIA in Determining Fruit sugars.........................................83 1. Determination of Sugars in Fruits of Different Maturity....................................83 a. Olives.. ............. . ................... 83 b. Cherries....... ..... .....................86 2. Determination of Sugars in Ripe Fruits......86 C. Determination of Fruit sugars using Conventional Enzyme Methods................................ 87 D. Comparison of the two Methods..................87 E. Test of the Conventional Method Against Sugar StandardBOOOOO..OOOOOOOOOOOOOOOOOO0.00.00.00.0092 V. SUMMARY...0..0.0.0.0...OQOOOOOOOOOOOO000.000.000.100 REFERENCESOOOOOOOOOOOOOOOOO...0.0.0.0000000000000102 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 10. ll. 12. 13. LIST OF FIGURES Page Flow diagrams of Flow Injection Analysis and Liquid Chromatography.........................S (a) The main parts of the simplest flow injection analyzer. (b) Typical output corresponding to analyte......................9 Types of enzymatic reactors that can be used in the Flow Injection Analysis method...........14 Attachment of the enzyme in two ways by using aVidin-biotin interaction. 0 O O O O O O 0 O O O O O O O 0 O O O 17 Parallel multichannel flow injection analyzer for the enzymatic determination of sugars....21 Enzymatic reactions for each sugar...........22 Diagram of the apparatus and the reaction manifold used for the analysis of glucose....23 Enzymatic reaction schemes for glucose, fructose and sucrose..................................24 Oxidation of Leucomalachite Green to Malachite GreenOOOOOO0.0...O.....OOOOOOOOOOOOOOOOOOO0.026 (a) Detection of hydrogen peroxide with the Trinder Reaction. (b) Detection of hydrogen peroxide with the Malachite Green Reaction...36 Comparison of LMG and Trinder reaction using hydrogen peroxide. The lengths of the enzymatic and plain SBSRs were 10 cm and 30 cm, respectively.................................38 Comparison of LMG and Trinder reaction using a glucose oxidase reactor. The lengths of the enzymatic and plain SBSRs were 10 cm and 30 cm, respectively.................................39 Relationship between LMG concentration and absorbance of the H202 reaction product. The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively......................42 (Continued) xi Figure Figure Figure Figure Figure Figure Figure Figure 14. 15. 16. 17. 18. 19. 20. 21. (Continued) LIST OF FIGURES Page Loss of LMG reagent activity during storage up to 270 min. The concentration of the stock LMG solution was 1.5 mM. The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively. X-Axis is time of signal detection..........44 Loss of LMG reagent activity during storage up to 270 min. The concentration of the stock LMG solution was 3.0 mM. The lengths ofenzyme and plain SBSR were 10 cm and 30 cm,respectively X-Axis is time of signal detection...........45 Comparison of two concentrations of LMG stock solution for reaction sensitivity. The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively.................................46 Relationship between LMG reagent concentration and absorbance of reaction product at three concentrations of peroxidase (P0). The lengths of enzyme and plain SBSRs were 10 cm and 30 cm respectively. The standard hydrogen peroxide solution was 0.1 mM..........................48 Movements of standard step size simplex procedure on response surface. The initial simplex is 1, 2, 3 and the optimum region lies close to point 12........................... 52 Relationship of absorbance and flow rate without glucose oxidase. The standard hydrogen peroxide solution was 0.04 mM and the carrier was a 0.1 M phosphate buffer pH 6.0. The lengths of enzyme and plain reactors were 12 cm and 35 cm, respectively.................................60 Relationship of absorbance and flow rate in the presence of glucose oxidase. The standard glucose solution was 0.4 mM and the carrier pH was 6.0. The lengths of enzyme and plain reactors were 12 cm and 35 cm, respectively..61 Relationship of length of the enzyme reactor and absorbance ( MG+ concentration). The standard hydrogen peroxide solution was 0.04 mM. The length of the plain reactor was 35 cm........62 (Continued) xii Figure Figure Figure Figure Figure Figure Figure Figure 22. 23. 24. 25. 26. 27. 28. 29. (Continued) LIST OF FIGURES Page Relationship of length of the plain reactor and absorbance (MG+ concentration) at three concentrations of H202. The length of the enzyme reactor was 12 cm....................64 Relationship of absorbance (MG+ concentration) and flow rate with three different lengths of plain reactors. The carrier was pH 6.0 and the length of the enzymatic reactor was 12 cm. The standard solution of hydrogen peroxide was 0.04 Meeeeeeeeeooooeeeoooeoooeoeeeeooooeeoeooooo066 The effect of flow rate on dispersion. The lengths of enzyme and plain reactors were 12 cm and 35 cm, respectively. The carrier pH was 6.0 and the standard solution of hydrogen peroxide was 0.04 mM.... ....... .... ..... ..............68. Relationship between the carrier pH and absorbance (MG+ concentration) with different lengths of the plain reactor. The length of the enzyme reactor was 12 cm and the standard solution of hydrogen peroxide was 0.04 mM....69. Comparison of the sensitivity of LMG reaction with different lengths of plain SBSR. The enzyme reactor was 12 cm and the pH of the carrier was 6.0........ ........ ..............71 Sensitivity of the LMG reaction for different lengths of enzymatic SBSR. The plain SBSR was 35 cm and the pH of the carrier was 6.0......72 Apparatus for the controlled temperature determination of glucose using immobilized glucose oxidase ....................... .. ..... 74 Diagram of the glass part used for the temperature experiments with glucose oxidase. Water with specific temperature was passing around the glass tee.........................75 (Continued) xiii Figure 30. Figure 31. (Continued) LIST OF FIGURES Page Effect of temperature on the activity of the glucose oxidase reactor. The enzyme and plain SBSRs were 12 cm and 35 cm, respectively; the pH of the carrier was 6.0; the flow rate was 1.47 ml/min and the standard glucose solution was 0.2 mM...................................76 Comparison of specific Simplex and Univariate methods of optimization for sensitivity......82 Table Table Table Table Table Table Table Table Table Table Table Table Table 10. 11. 12. 13. LIST OF TABLES Page Comparison of attributes between FIA and HPLC...6 Comparison of buffer conditions for LMG stock solution.OOOOOQOOOOOOOOOOO..OOOOOOOOOO0.0.00.0050 Range and precision of variables for the general Simplex optimization.OO...0.0.00.0000000000000054 Initial and optimal experimental conditions for the flow injection analysis system.............56 Range and precision of variables for the specific simplex optimization...........................78 Optimum information given by simplex optimization ProgramooooO00.0..0......00.00.000.00000000000079 Initial and optimal experimental conditions for the specific simplex optimization of FIA systemOOOOOOOO0.00.0000...0.00.00.00.000000000080 Determination of sugars in fruits using FIA systemOOOOO0.0.00000000000000000.0000000000000084 Determination of fruit sugars using Conventional Enzymatic method.....OOOOOOOOOOOOO0.0.00.0000088 Analysis of variance for sugars with treatment and replication numbers equal to 26 and 6, respectiveIY0.00000000000000000000000.0.000000091 Statistical comparison of glucose measurements for FIA and conventional methods by using the Least Significant Difference (LSD) procedure...93 Statistical comparison of fructose measurements for FIA and conventional methods by using the Least Significant Difference (LSD) procedure...95 Statistical comparison of sucrose measurements for FIA and conventional methods by using the Least Significant Difference (LSD) procedure...97 XV I. INTRODUCTION The significance of chemical analysis in Food Science and Human Nutrition is steadily increasing. New techniques are sought which are rapid, do not require highly skilled personnel and are not very costly. Dramatic advances in electronics, computer technology and biotechnology over the past decades have led to the emergence of a number of techniques to fulfill the needs of the modern food analysis laboratory. The availability of fast, low-cost computers allows automation in instrument control and data handling so that the classical methods themselves can be automated. In addition, the data handling capability allows the development of sophisticated techniques based on measurement of physical properties of the sample; these include spectroscopy, chromatography and electrochemistry. Flow Injection Analysis (FIA) is a fairly recent technique (1) and features major methodological innovations such as simplicity, relatively inexpensive equipment, handy operation and great capacity for achieving results that are excellent in rapidity, accuracy and precision. The extreme versatility of this methodology makes it stand out from most new analytical techniques. For example, FIA can be adapted to meet many types of requirement without major technical changes. FIA differs from traditional analytical techniques in that it is not necessary for measurements to be made at a state of equilibrium with respect either to the course of the chemical reaction or flow dynamics. FIA is a microchemical technique in which beakers, pipettes and volumetric flasks are replaced by small (0.5 mm i.d.) open—ended tubes through which the solutions are pumped. Since FIA has proven to be very effective by making the handling of liquid samples an easy task, it has great potential in all areas that require chemical analysis. This research contributes to the development of a rapid, parallel continuous flow analyzer, for the simultaneous enzymatic determination of six nutritionally important sugars present in food samples with complex matrices, without prior separation. Work on this project has been going on for several years already under the supervision of Dr. Crouch, with various researchers focusing on different aspects, such as enzyme immobilization procedures, sample preparation, and construction of the FIA manifold. Specifically this research deals with the optimization of a new flow-injection indicator reaction using Leucomalachite Green (LMG) and the application of this novel technique to the direct and indirect determination of three major free sugars (glucose, fructose and sucrose) in fruits using immobilized glucose oxidase. I I . BACKGROUND INFORMATION A. Methods of Sugars Analysis in Fruits Fresh fruits and certain vegetables are major sources of unprocessed sugars in the human diet. In addition, fruits contain a higher proportion of free sugars than vegetables and a lower proportion of unavailable carbohydrates than most vegetables (2). Recently, high fiber natural health foods, consumption of unprocessed vs. processed foods, and metabolic differences in fructose, glucose and sucrose have all received wide spread attention. The free sugars in fruits are usually mixtures of glucose, fructose and sucrose. Occasionally maltose and other oligosaccarides are present. The proportions of the different sugars are characteristic of the fruit, although different varieties of the same fruit show some variations. 1. Non-Specific and Specific Methods High performance liquid chromatography (HPLC) and methods of enzymatic analysis have rapidly become the techniques of choice for the quantitative analysis of sugars and other carbohydrates in most fruits. Prior to the development of such analytical techniques, sugars were determined quantitativeLy in many fruits as total reducing sugars and total non-reducing sugars. These non-specific methods had serious limitations (3). In those measurements fructose and glucose were usually assumed to be the reducing sugars and sucrose the only nonreducing sugar. These assumptions are generally true for most fruits. Some fruits, however, contain significant amounts of sorbitol, which is not accounted for by those earlier measurements, or maltose, which would be included in the total reducing sugar value. As the role of individual sugars, such as fructose, in health and nutrition became more well defined (4), the need for rapid and simple quantitative methods for determining individual sugars (specific methods) in foods became more important. Individual sugars have been determined by GC and by enzymatic methods (using soluble enzymes). HPLC was shown to be generally faster than either of those two methods. A relatively new analytical technique, FIA, based on continuous flow and immobilized enzymes has been applied for the determination of individual sugars in fruits with greater advantages than the HPLC technique. 2. Comparison of HPLC with FIA Figure 1 shows two flow diagrams corresponding to FIA and liquid chromatography. Table 1 lists the common and differential features of FIA and HPLC. The following similarities should be emphasized: miniaturization capability, injection, unsegmented flow, small sample volume, and signal profile. On the other hand, there are substantial differences between them, the most important of which is probably their principle, since in HPLC there is always an interface which affects the separation of a mixture of substances passing through the column, and this is not so common in FIA. The similarity between both techniques becomes more apparent when a column (packed, open or single bead sting reactor) with an ion exchange resin or an immobilized enzyme is used in the manifold behind the injection point in a FIA system. The working pressure is a FLOW INJECTION ANALYSIS SAMPLE _l_ LIQUID PROPELLING ' INJECTION nmcron _> _’ _. REACTOR _> RESERVOIR SYSTEM SYSTEM LIQUID CHROMATOGRAPHY SAMPLE IIQUID PROPEIJING INJECTION | CO DETECTOR RESERVOIR . SYSTEM SYSTEM . . Figure 1. Flow diagrams of Flow Injection Analysis and Liquid Chromatography. Table 1. Comparison of attributes between FIA and HPLC. Characteristics HPLC FIA Pressure high low Column essential possible Interface always occasionally Data produced Cost Versatility Main analytical purpose Tubing diameter Flow rate Sample volume Sample introduction Unsegmented flow peak height/area high limited several components in a single sample small variable small injection yes peak height/area/ width/peak-to-peak distance low great a single component in many samples small variable small injection yes major factor responsible for significant differences between the two techniques. FIA uses low pressures, whereas in HPLC the pump (usually dual-piston) must exert a high pressure to overcome the hydrodynamic resistance of columns packed with material that is finely divided to improve the efficiency of the separation process. Despite the fact that some FIA methods have been developed with the aid of HPLC components, typical FIA systems are much simpler, since they are designed to at low pressures. Therefore, HPLC instruments are much more expensive. The scope of application of the two techniques is very different. The basic aim of an HPLC instrument is to separate and analyze a complex mixture of substances, whereas FIA. is mainly’ devoted to the rapid determination of a single species in a large number of samples. B. Introduction to Continuous Flow Methods The analytical procedures in which the analyte concentration is measured without stopping the flow of a gas or liquid are referred to as continuous flow methods (CFA) (5). There are two general types of continuous flow methods: segmented and unsegmented. In the segmented flow methods, the samples are introduced onto the manifold. which made of interconnected tubing, by aspiration for a defined period of time, and air bubbles separate (segment) the flow. In each segment complete mixing takes place so that the signal obtained at the output has a rectangular shape similar to what would be expected in the ideal case of a plug-shape sample. The air bubbles are usually removed before they reach the detector cell. Unsegmented flow methods, are the Flow Injection Analysis (FIA) methods and differ from segmented flow methods in that the flow is not segmented by air bubbles, the sample is injected instead of aspirated and neither flow homogenization nor chemical equilibrium has been accomplished by the time the signal is recorded (5). In addition, FIA methods require less sophisticated and expensive equipment. C. Principles and Function of Flow Injection Analysis The simplest flow injection analyzer (Figure 2a) consists of a pump, which is used to propel the carrier stream through a narrow tube; an injection port, for injection of a well defined volume of a sample solution into the carrier stream in a reproducible manner; and a microreactor in which the sample zone disperses and reacts with the components of the carrier stream, forming a species that is sensed by a flow through detector and recorded. A typical recorder output has the form of a peak (Figure 2b), the height H and width W, or area A, of which is related to the concentration of the analyte. The time span between the sample injection and the peak maximum, is the residence time T during which the chemical reaction takes place. A.well designed FIA system has an extremely rapid response, because T is in the range of 5-20 sec. Therefore, a sampling cycle is less than. 30 sec (T+tb), and ‘thus, typically, two samples can be analyzed per minute. The injected sample volumes may be between 1 and 200 ”L (typically 25 pL). This makes FIA a simple, automated microchemical technique, capable of having a high sampling rate and a minimum sample and reagent consumption. (a) Recorder Sample Pump 1 Carrier I—_ - . Detector _, Waste l ‘—J Injection Valve Reactor I I l | l l l I (b) (I I.— T —. l l H A W—b <— S l tb I Figure 2. (a) The main parts of the simplest flow injection analyzer. (b) Typical output corresponding to analyte. 10 Flow Injection Analysis is based on a combination of three principles: sam le in'ection, controlled dispersion of the injection sample zone, and reproducible timing of its movement from the injection point toward and into the detector. Thus, in contrast to all other methods of instrumental analysis, the chemical reactions are taking place while the sample material is dispersing within the reagent prior to the detection point. This is why the concept of dispersion, controlled within space and time, is the central issue of FIA. 1. Definition Flow Injection Analysis (FIA) is based on the injection of a liquid sample into a moving, nonsegmented continuous carrier stream of a suitable liquid. ‘The injected sample forms a zone, which. is then transported. toward. a detector that continuously records a physical parameter, such as absorbance or electrode potential, as it continuously changes due to the passage of the sample material through the flow cell (1,5). 2. Essential Features In principle, Flow Injection Analysis is an automatic method of analysis. The essential features of the FIA used here are the following: (a) The flow is not segmented by air bubbles, which is the fundamental difference from classical CFA methods. (b) The sample is injected or inserted directly into the flow stream instead of being aspirated into it. (c) The injected plug is carried along the system. A physicochemical process (chemical reaction, dialysis, liquid-liquid extraction, etc.) may occur in addition to transport. 11 (d) The partial dispersion or dilution of the analyte throughout this transport operation can be manipulated by controlling the geometric and hydrodynamic characteristics of the system(tubing i.d., length, flow rate ect). (e) A continuous sensing system yields a transient signal which is recorded. (f) Neither physical equilibrium (which would involve the homogenization of a portion of the flow) nor chemical equilibrium (completeness of reaction) has been attained when the signal is detected. (9) The operational timing must be highly reproducible because measurements are made under non-equilibrium conditions and small variations may result in serious alterations in the results. 3. Dispersion The flow injection technique involves the injection of a sample into a nonsegmented carrier stream. Since the conditions are usually such that laminar flow is predominant, the development of a parabolic velocity profile is responsible for the dispersion of the sample along the axis of the tube. This dispersion, although much greater than that found in CFA, can be controlled by appropriate choices of tubing length and inner diameter, flow rate, sample size, and other components such as valves and flow cells which determine the overall volume of the reactor in the FIA system. ' Ruzicka and Hansen have proposed an empirical method by which dispersion can be measured (5). The sample has an initial concentration (R) as it enters the carrier stream. As the plug travels through the manifold, axial and radial mixing take place. This result is a predominantly Gaussian-shaped signal profile. The maximum concentration 12 sensed by the detector, Cmax' is only a fraction of C0. The formula for the dispersion is: D = CO/Cmax = Ho/H x Const'/Const" The height of the peak obtained with the undiluted sample is H0. After the sample has traversed the manifold a lower peak height, H, will be obtained due to dispersion. If the two constants are equal, as in the case of photometric detection for a system that obeys Beer’s law, the peak heights of the signals can be used to determine the dispersion of the FIA system. The amount of dispersion that can be tolerated in an FIA system depends on the application for which it is applied. For mere transportation of a sample, limited dispersion is ideal. On the other hand, for a chemical reaction requiring reagent additional dispersion must take place. In conclusion, an FIA peak is a result of two kinetic processes that occur simultaneously: the physical process of zone dispersion and the chemical processes resulting from reactions between sample and reagent species. 4. Types of FIA Manifolds The physical foundations of FIA are related to dispersion, which is caused by injecting a sample volume into the flowing stream. The dispersion is characterized by the concentration profile adopted by a zone or plug inserted at a given point in the system without stopping the flow. The recorder output from the detector (the analytical signal) is representative of the dispersion at such a point, and can be used to access the extent of dispersion. .m- 13 One way of manipulating dispersion is by selection of the appropriate type of manifold. A comparison between various types of FIA manifolds was carried out by Ruzicka and Hansen (6). Three general types of reactors have been used in FIA: the open tubular (OTR), packed bed (PBR), pnd single bead string reactors (SBSR) ( Figure 3). Since straight open tubes yield relatively large amounts of dispersion (7), little attention has been paid to the preparation of reactors with enzymes immobilized on the inner wall of the tubes (8-10). However, such. reactors have advantages. over' packed Ibeds in certain applications because they permit an unobstructed flow of the substrate solution. Coiled tubes show less dispersion due to the presence of a secondary flow (11). This type of flow is a result of the centrifugal forces which affect the flow perpendicular to the axis of the tube. A relatively small amount of dispersion has been found in the case of packed tubes (12-13). This is due to the fact that the parabolic profile is broken up as the sample passes through the packed material. The high pressure drop associated with tubes that have been packed with very small diameter particles makes them difficult to use with the peristaltic low pressure pumps normally present in an FIA system. The single bead string reactor (14-15) has gained acceptance as a viable alternative to open tubes and those packed with small particles. The SBSR consists of ordinary Teflon tubes packed with glass beads having diameters that are 60-80 % of that of the tube. This type of reactor in FIA provides the following advantages: 14 OoOoOoOoQoOo Single Bead String Reactor Open Tubular Reactor Figure 3. Types of enzymatic reactors that can be used in the Flow Injection Analysis method. 15 (a) Decreased dispersion due to the break up of the velocity profile. (b) The sampling rate is significantly high, with up to 500 samples assayed per hour. (c) It allows easy merging of streams, which is especially desirable when there is a chemical reaction between species dissolved in the streams. (d) The pressure drop is small. Therefore, the SBSR can be used to provide longer residence time without an increase in dispersion. (e) For use with immobilized enzymes an added advantage of the SBSR is the additional surface area available compared to that of an open tube. D. Use of Immobilized Enzymes in FIA Enzymes have been insolubilized by irreversible covalent attachment 'to ‘various rorganic 'polymers (16-22), and. cellulose derivatives (23-25). Immobilization has also been accomplished by entrapment in starch (26) and acrylamide gel (27—28). These types of derivatives have been studied in detail by Silman and Katchatski (29). Also, enzymes can be irreversibly covalently coupled to inorganic carriers (30-33). In general, inorganic carriers are run: subject to microbial attack. The carrier does not change configuration over an extensive pH range or under various solvent conditions, and is therefore, easier to use in continuous flow systems. In addition, the inorganic carriers have greater rigidity and they immobilize enzymes to a greater degree than do organic polymers. In general, immobilization by covalent attachment has proved to be the most suitable for continuous flow analysis and can be applied with 16 the three different types of immobilized enzyme reactors, such as packed columns, open tubular wall reactors, and single bead string reactors. 1. Reversible Immunological Immobilization The use of immobilized enzymes packed into reactors and coupled to flow systems in analytical applications has been well demonstrated (34- 35). However, the immobilized enzymes in these configurations suffer from several limitations. Some of these are a) limited lifetime, and b) susceptibility to inhibitory and steric problems created by immobilization, which limit the transfer of substrate to the enzyme layer and block access to the active site. By ‘using immobilized antibodies which are specific to the enzyme or using indirect immunochemical reactions, enzymes can be immobilized with high efficiency while retaining maximum enzymic activity. The use of antibodies in the immobilization of enzymes allows the operator to replace the bound enzyme reproducibly in a few minutes in the event of a loss of enzyme activity without removing or replacing the packing material (Figure 4). The flow injection analysis method has the advantage of rapid sample throughput and minimal sample handling. Another coupling technique is based on the fact that avidin binds to biotin with a binding constant of 1015 (36) and that the resulting binding is therefore irreversible under conditions where the antibody- antigen interactions can be reversed (37—40). This situation provides a method for immobilizing the primary antibody with high efficiency (41- 46). Furthermore, the use of avidin-biotin interactions for the reversible or irreversible immobilization of enzymes is very critical (Figure 4). These methods of immobilization of enzymes in reactors use 17 I. Irreversible Attachment : + B ___.,/WE/ 1. NW (Activated support) (Biotin) (Immobilized Biotin) -A‘ 2 AAJBV + A ——" AA/E/ (Immobilized Biotin) (Avidin) (Biotin-Avidin) + B-E .___.> (Biotin-Avidin) (Biotin-Enzyme) (Immobilized Enzyme) E. II. Reversible Attachment : t -A-B- - + B-Ab + E pH6.8 /\/\/E/ Ab E ——> : (Biotin-Avidin) (Biotin-Antibody) (Immobilized Enzyme) -A-B-Ab’ M sz.o E + 2. ______.. (Immobilized Enzyme) (Free Enzyme) Figure 4. Attachment of the enzyme in two ways by using avidin-biotin interaction. ‘- 18 two biospecific reactions where one of the reactions is irreversible and the other reversible (Figure 4). 2. Advantages of Immobilized Enzymes In recent years there has been an increasing use of immobilized enzyme preparations in industrial, analytical and medical procedures (47). The most obvious advantages are products free from enzymes, continuous run, greater efficiency of substrate conversion, higher yields and good product uniformity. However, these advantages must be balanced against the additional costs of enzyme immobilization, and the relatively poor stability of purified soluble enzymes. Additionally, some enzyme solutions cannot be used in organic solvents, and all are very sensitive to elevated temperatures. These particular drawbacks have slowed the advancement of enzyme applications, and much research effort has been. expended. to overcome these jproblems. ILarge scale procedures of enzyme immobilization have helped to reduce enzyme immobilization costs. New procedures of enzyme immobilization like pre- treatment (for maximizing the surface area on support) or new immobilization methods such as affinity chromatography and immuno— techniques have pmovided useful preparations, with greater stability, that are also suitable for reuse (48). 3. Applications of Immobilized Enzymes The benefits of an increased understanding of enzymes, and especially immobilized enzymes, should allow many novel solutions to analytical problems involving substrates, activators (n: inhibitors of these enzymes. In addition, the potential for using immobilized enzymes as catalysts in areas such as food and clinical analysis, medicine, chemical synthesis and conversions, has been widely promoted. The scope 19 for using enzymes as industrial catalysts is indicated by the wide range of reaction types that can be catalyzed by enzymes. These include oxidation/reduction, inter and intramolecular transfer of a variety of chemical groups, hydrolysis, cleavage of covalent bonds, isomerization and addition of chemical groups across double bonds; so some organic and many inorganic reactions can be catalyzed by one or more enzymes. The best known and widely used enzymic assay is for glucose. Glucose Oxidase, oxidizes B-D-glucose to gluconic acid and hydrogen peroxide. The latter is used to oxidise an appropriate dye using horseradish peroxidase, giving ‘a useful colorimetric assay (20,49 and 50). Some mutarotase activity should be present to convert a-D-glucose to B—D—glucose, as the latter is oxidized 150 times faster at 20 0 C by the glucose oxidase used. Furthermore, many other substrates can be assayed using immobilized enzymes and FIA, including: Vitamin C by Ascorbate Oxidase (51), Ethanol by Alcohol Dehydrogenase (52), Amino Acids by L-Amino Acid Decarboxylases (53), L-Lactic acid by Lactate Oxidase (54), Oxalate by Oxalate Oxidase (55), Penicillin by Penicillinase (56), Urea by Urease (57), Malate by Malate Dehydrogenase (58), Cholesterol by Cholesterol Oxidase (59-61) etc. III. METHODS AND MATERIALS Unlike many other theses, in this work the development and optimization of an analytical method was the major objective. For this reason in this section only the general interactions on which the method is based are described, and the details of the proposed method are given in the Results and Discussion section. The conventional method for the determination of sugars is described here. A. Methodology of FIA for Sugars 1. Apparatus For the determination of the six nutritionally important sugars, the proposed novel parallel continuous flow analyzer and the appropriate enzyme reaction schemes are shown in Figures, 5 and 6. In this work, direct determinations of glucose and indirect determination of fructose and sucrose were done with the flow injection apparatus shown in Figure 7 and the enzymatic reaction schemes shown in Figure 8. The flow injection analysis apparatus consisted of a 12— channel peristaltic pump (Ismatec, Glattbrugg, Switzerland) with flow- rated pump tubing (Technicon Instruments, Tarrytown, NY), a pneumatically activated injection valve with a 30 nL sample loop (Rheodyne Inc., Cotati, CA), and a miniaturized flow through filter colorimeter designed and constructed by Patton and Crouch (62). A light source of variable intensity was connected to the channel of the detector via a fiber optic. The wavelength of the operation was 620 nm and was accomplished by a filter. An IBM PC compatible microcomputer, equipped with an RTI-815 (Analog Devices, Norwood, MA) interface board, 20 ,... a.-_..._ "—' " 21 l to 6 Selector km S '3 mil) Carrier ll Enzyme Plain Reactors Reactors D Cunfl I ud E Reagent C Shams O R PUMP. Figure 5. Parallel multichannel flow injection analyzer for the enzymatic determination of sugars. 22 Glucose Oxidase _ D‘EIUCOSC + 02 + H20 D-glucono—d-lactone + I1202 Galactose Oxidaoe_ D-galactose + 02 + H20 7 D—galacto-hexodialdose + H202 Sucrose + H20 1mm." : mutant": ; D-glucose + D—fructose Lactose + H20 fi-"hcmida" ; D-glucose + fi-D-galactose a - glucosidue L Maltose + H20 D-glucose + a—D-glucose 01 I D-fructose “w" ”men" E D-glucose + D—fructose Figure 6. Enzymatic reactions for each sugar. 23 PC/XT I I ‘ k Sam Glucose Oxidase Carrier fijBZJ SBSR V23/r EEEEQE}-— lkggm Figure 7. Glucose Plain SBSR #J/ j ___l Detector Diagram of the apparatus and the reaction manifold used for the analysis of glucose. 24 D—glucose + 02 + 320 mum” and.” : D-glucfino—d -lactone + H202 D—fructose mum" I'°m°“'° > D—glucose + D—fructose Sucrose + H20 Immue ; mum-owe = D—glucose + D-fructose Figure 8. Enzymatic reaction schemes for glucose, fructose and sucrose. 25 controlled the pump speed, sample injection and data acquisition. Software was written in Quick BASIC (Microsoft Corp., Rendmond, WA). More details about this apparatus are given by Stults (63). This system was based on immobilized enzyme SBSRs for high selectivity and sensitivity. The sample containing glucose passed first through a SBSR which had glucose oxidase covalently bonded to 0.6 mm diameter non-porous glass beads (Propper MFG. CO., L. I. City, NY) (64). The beads were packed into a reactor of 0.86 mm i.d. poly(tetrafluoroethylene) tubing (Benton-Dickinson, Parsipanny, NJ). The sample was then mixed with a reagent stream that contained horseradish peroxidase, and Leucomalachite Green (LMG) (Figure 9). A colored product (MG+) was formed in a plain SBSR and was detected photometrically. The pump tubing for the carrier, the sample and the reagent stream were, 0.42 cc/min, 0.32 cc/min, and 0.06 cc/min respectively. Different pieces of tubing were connected to one another as well as to the different components of the system by means of connectors. Particular care was taken, in making connections, to avoid dead volumes, leakage or the introduction of air bubbles. The parameters and their ranges of the values studied for the optimization effort were the following: concentration of LMG stock solution, 0.303-15.151 mM; concentration of LMG reagent solution, 2 6 x 10-2 - 12 x 10- mM; activity of peroxidase, 72-178 units; pH of LMG stock solution, 1.65-3.0; flow rate, 0.2-1.8 ml/min; length of enzymatic SBSR, 8-16 cm; length of unmodified SBSR, 14—40 cm; pH of carrier, 5.0- 0 6.5; and effect of temperature, 20-50 C. 26 i": N\ . on, i on . O C -N/ 3 + PO 6 \\ /N‘\ N+-CH3 I cu3 cu, C”! Lee Figure 9. Oxidation of Le ucomalachite Gr ' Green. een to Malachite 27 For the optimization of the FIA method and for the applications in determining fruit sugars, six replications were performed for each sample. 2. Reagents All stock solutions were prepared with distilled water and filtered before use. All stock solutions were diluted with 0.05 M phosphate buffer, pH 6.85. All chemicals (reagent grade) were used without further purification. Anhydrous fi—D(+)-glucose grade III, sucrose grade II and B-D(—)- fructose crystalline (all from Sigma Chemical Co., St.Louis, MI USA) were used to prepare the standard solutions (all 0.01 M) that contained 0.5 g L'-1 benzoic acid as preservative. The reagent for the Leucomalachite Green (p,p'-Benzylidene-bis-N,N-dimethylaniline) (LMG) indicator reaction was prepared immediately before use and contained peroxidase, dissolved in 0.1 M phosphate buffer, pH 6.0, LMG, and 0.1 M acetate buffer, pH 4.0. For comparison purposes the Trinder indicator reaction (66) was also employed. The reagent for the Trinder indicator reaction was prepared immediately before use and contained 143 units peroxidase, 1 mM 4-aminoantipyrine (from Sigma), 1 mM 3,5-dichloro-2- hydroxyphenyl sulfonic acid mixed together and then diluted with 0.05 M phosphate buffer, pH 6.85, to 10 ml in a volumetric flask (63). A stock solution of 0.06 M of H202 (30 % W/W from Sigma Chemical Company) was used for the preparation of the standard solutions. Stock magnesium chloride (MgClz.6H20) solution 0.1 M and cobalt chloride (COC12.6H20) solution 0.01 M were prepared for use as activators for the glucose isomerase conversion reaction in 0.05 M phosphate buffer and pH 7.5 environment. 28 The enzymes used were horseradish peroxidase (Sigma, Type II, from Aspergillus niger, activity approximately 17800 units 9-1 ), invertase (Sigma, grade VII, from Baker's Yeast, activity approximately 400 units mg-l), glucose isomerase (Spezyme GI-M600) (Finnsugar Biochemicals, Inc., activity approximately 3290 units g-l), and ascorbate oxidase (Boehringer Mannheim, from Cucurbita species, activity approximately 170 units mg-l). Glucose oxidase was immobilized on non-porous glass beads by the procedure described by Stults (64). 3. Preparation of the Samples All solutions of H202 and standard sugars, for the optimization studies , were prepared with 0.05 M phosphate buffer pH 6.85. The fruit samples tested were olives from California (Manzanillo and Ascolano), from Greece (Coroneiki and Amphisis) and from Cyprus (Cypriot, Manzanillo and Ascolano); cherries from Michigan (Wolynska, Montmorency and I 20(36)); and citrus fruits (oranges "NAVEL" from California, lemons from California, and grapefruit from Florida). The olive and cherry samples were in different maturity stage, but the citrus fruit samples were in the ripe stage. Fifty milliliters of the citrus juice samples, which weighed 52.5 g, 52.3 g and 52.4 g for the oranges, lemons and grapefruit juice, respectively, were diluted 1:10 prior to use for preparation of the FIA working solution . Ten grams of olive fruit without seeds were blended for 5 min with 30 ml water. The slurry was centrifuged at 2100 rpm for 10-15 min and then filtered under vacuum twice. The filtrate was diluted to 50 ml total volume with water and used for the preparation of the FIA working solution without any further dilution. Cherry selections were harvested at the 29 Clarksville Horticultural Experiment Station and frozen at -20 0C under nitrogen, a few hours after harvesting. Fifty grams of frozen cherries without seeds were blended at high speed with 50 ml water for 10-15 min. The slurry was centrifuge at 2100 rpm for 10-15 min and then filtered under vacuum, twice. The solution was diluted to 100 ml by using distilled water. A.1:10 dilution was done prior to use for preparing the FIA working solution. All working solutions were prepared with 0.05 M phosphate buffer, pH 6.85, and used 0.2 ml of the sample solution. 4. Procedure Six replicate measurements were done for each sample. The conditions for all measurements were the optimum for the new indicator reaction of LMG (see below). a. Glucose Analysis For the determination of glucose, each sample was transferred into a separate 10 ml volumetric flask; the amounts taken were 0.2 ml of the citrus juice (after dilution), 0.2 ml of cherry juice, and 0.2 ml of olive juice. 1 m1 stock ascorbate oxidase solution was prepared using 0.1 M phosphate buffer pH 5.5 and 10 mg ascorbate oxidase (1700 units mg-l). 0.02 ml of this stock solution was transferred into all juice samples to destroy the undesirable ascorbic acid. The volumetric flasks were filled to volume with 0.05 M phosphate buffer, pH 6.85, to make the working solutions. The FIA determinations begun after a delay of 15 min to allow the a and 3 forms of D-glucose to reach equilibrium in the phosphate buffer (65). After injection, approximately 85 sec passed under the conditions used before the FIA signal was obtained with the 30 computer data acquisition system. Peak absorbance values were used for the calculations. b. Fructose Analysis For the determination of fructose, the same procedure was used as for glucose except that the samples were treated with glucose isomerase (GI) prior to the glucose determination. The conversion reaction of fructose to glucose proceeded in the presence of the enzyme activators, M9012, CoClz, at pH 7.5 and at 60 °C. A stock activator mixture was prepared, which contained 2 ml of 0.1 M MgClz, 2 ml of 0.01 M 00012, 2 ml of 1M phosphate buffer, pH 7.5 and 6 ml of H20. A 0.3 ml volume of the activator stock solution and 0.1 ml glucose isomerase enzyme were transferred to all 10 ml volumetric flasks. The reaction was run at pH 7.5 at 60 °C for 25 min. Then the volumetric flasks containing the treated samples for fructose analysis were filled to volume with the buffer solution. The FIA determinations of glucose were again begun 15 min after the dilution. A blank sample without glucose isomerase was run and the appropriate corrections to the samples with enzyme were done. A relatively high concentration of glucose isomerase was used in order to speed up the conversion reaction. c. Sucrose Analysis For the determination of sucrose, the same procedure was used except that the samples were treated with invertase prior to the glucose determination. Ten milliliters of stock solution of invertase was prepared using 0.1 M acetate buffer pH 4.5 and 5 mg invertase. A 0.1 ml volume of this stock solution was transferred to all volumetric flasks to convert the sucrose to the invert sugar. The conversion reaction proceeded under optimum conditions: one unit of invertase hydrolyzed 31 1.0 pmole of sucrose to invert sugar per min at pH 4.5 and 55 CT: for 15 min. 8. Conventional Enzyme Methods for Sugar Analysis 1. Apparatus The apparatus used for the conventional enzyme method of sugar analysis consisted of a Perkin-Elmer Lambda 4B, UV/VIS spectrophotometer with a Perkin-Elmer Lambda Accessory Interface and an Epson printer. The ‘wavelength. of the determination. was 340 nm and .Absorbance ‘was measured. 2. Reagents All stock solutions were prepared with distilled water. For these experiments a kit for sucrose, D—glucose and D-fructose was used (Methods of Food Analysis using Test-Combinations, Boehringer Mannheim). This kit consisted of five bottles. Bottle 1 with approximately 0.5 g lyophilisate, contained: citrate buffer, pH 4.6; b-fructosidase, 720 units; stabilizers. Bottle 2 with approximately 7.2 9 powder mixture contained: triethanolamine buffer, pH 7.6; NADP, 110 mg ; ATP, 260 mg; magnesium sulfate; stabilizers. Bottle 3 with 1.1 ml enzyme suspension, consisting of: hexokinase, 320 units; glucose-6-phosphate dehydrogenase, 160 units. Bottle 4 contained approximately 0.6 ml phosphoglucose isomerase suspension, 420 units. And bottle 5 contained standard sucrose. Anhydrous fi-D-glucose grade III and B-D—fructose crystalline (all from Sigma Chemical Co.) and sucrose (Boehringer Mannheim) were used to prepare ‘the standard solutions (2 g/L) for the evaluation. of this method. 32 3. Preparation of the Samples The samples tested were olives (from California, Greece and Cyprus), cherries (from Michigan) of different maturity, and citrus fruits (oranges from California, lemons from California, and grapefruit from Florida) in the ripe stage. Fifty milliliters of the citrus juice samples weighing 52.5 g for orange, 52.3 g for lemon and 52.4 g for grapefruit juice were diluted 1:25 prior to use for the soluble enzyme analysis. Ten grams of olive samples without seeds were used for the sugar extraction. Fifty milliliters distilled water were used as extraction solvent and the sample was diluted to 1:10 prior to analysis. Fifty grams of cherry without seeds were blended with water for 10 min and the slurry was centrifuged and filtered. The final volume of the filtrate was made to 100 ml and diluted 1:50 prior to use for the sugar analysis. 4. Procedure The D-glucose concentration was determined before and after the enzymatic hydrolysis of sucrose; D-fructose was determined after the determination of D-glucose. In all determinations the volume of the fruit sample solution used was 0.1 ml. For each measurement of D- glucose or D—fructose the whole procedure took 50-60 min plus 50-60 min for the D-glucose\D-fructose blank sample. For the sucrose measurement the procedure took 35-40 min plus 35-40 min for the sucrose blank sample. Three replicate measurements were done for each standard sample of glucose, fructose and sucrose, but only one measurement was made for the fruit samples. All the appropriate corrections in the calculations were applied. 33 a. Determination of D-glucose before inversion The enzyme hexokinase (HR) catalyzes the phosphorylation of D- glucose by adenosine-S—triphosphate (ATP) with the simultaneous formation of adenosine-S-diphosphate (ADP). D-Glucose + ATP -------- > G-6-P + ADP In the presence of glucose-6-phosphate dehydrogenase (G6P-DH) the glucose-6-phosphate (G-6-P) formed is specifically oxidized by nicotinamide—adenine dinucleotide phosphate (NADP) to gluconate-6- phosphate with the formation of reduced nicotinamide-adenine dinucleotide phosphate (NADPH). G-6-P + NADP+ -------- > gluconate-6-phosphate + NADPH + H+ The NADPH formed in this reaction is stoichiometric with the amount of D-glucose and is measured by means of its absorbance at 340 nm. b. Determination of D-fructose Hexokinase also catalyzes the phosphorylation of D-fructose to fructose-6-phosphate (F-6-P) with the aid of ATP. D-fructose + ATP -------- > F-6-P + ADP On completion of this reaction F-6-P is converted by phosphoglucose isomerase (PGI) to G-6-P. F-6-P -------- > G-6-P G-6-P reacts again with NADP+ with formation of gluconate-G-phosphate and NADPH. The amount of NADPH formed now is stoichiometric with the amount of D-fructose. c. Determination of sucrose Sucrose is hydrolyzed by the enzyme invertase to D-glucose and D- fructose. 34 Sucrose + H20 -------- > D-glucose + D-fructose The determination of D-glucose after inversion (total D-glucose) is carried out according to the reaction above. The sucrose content is calculated from the difference of the D-glucose concentrations before and after enzymatic inversion. IV. RESULTS AND DISCUSSION The main objective of this work was to optimize the conditions for the FIA determination of glucose, using a new indicator reaction. Two more sugars, fructose and sucrose, were subjected to this analysis, after their conversion to glucose, and a comparison was made with a conventional method of sugar analysis. A. Optimization of the FIA System 1. Initial Optimization of the Indicator Reaction One of the most commonly used method, for the colorimetric determination of glucose is the Glucose Oxidase / Trinder reaction (66- 68). In this reaction, glucose and molecular oxygen in the presence of glucose oxidase (GO) produce hydrogen peroxide (Figure 10a). The H202 then reacts with 4-aminoantipyrine (AAP) and 3,5 dichloro-Z- hydroxyphenyl sulfonic acid (DCPS) in the presence of peroxidase (PO) to produce a colored compound with an absorbance maximum at 510 nm (69). The Trinder reaction has been used by many researchers for glucose determinations in clinical applications using immobilized glucose oxidase and flow injection analysis (FIA) systems (70-72). Also Stults (64) optimized the Trinder reaction for the enzymatic determination of glucose with a flow injection analysis system. However, the Trinder reaction has certain shortcomings such as limited sensitivity and a small linear dynamic range. Also, for applications with food samples with high concentration of sugars, the Trinder reaction appeared to be not 'the reaction. of choice, because ‘the samples had. to be diluted several times. 35 36 (a) Trinder Reaction: . DCPS + AAP + H202 m» Quinonimine Dye + H20 DCPS = 3.5-dichlorp-2-hydroxyphenyl sulfonic ac: AA? = 4—eminoantipyrine lb) Malachite Green Reaction: LMG + 1120,: ”mm” Malachite Green + H20 LMG = leucomalachite green Figure 10. (a) Detection of hydrogen peroxide with the Trinder Reaction. (b) Detection of hydrogen peroxide with the Malachite Green Reaction. 37 Several other dyes have been used by many workers for the coloremetric assays of H202 in the presence of peroxidase (PO) such as benzidine (73), leucomalachite green (74-82), and o-dianisidine (73). In this work the LMG reaction has been optimized and used as the new indicator reaction in order to overcome the limitations of the Trinder reaction for practical applications in the food science area (Figure 10b). There are several advantages of using LMG in the indicator reaction. First, LMG is more sensitive in its response over the desirable absorbance range than the Trinder reaction. Figures 11, and 12 show a comparison of the LMG and Trinder reactions, using H202 and glucose standard solutions in 620 nm and 510 nm, respectively. The experimental conditions for this comparison were as follows: 1M acetate buffer, pH 2.25, for LMG stock solution; 3 mM of stock LMG solution; 143 units of peroxidase per 10 ml of reagent; 12 x 10-2 mM LMG of reagent solution; 5.0 ml, 0.1 M phosphate buffer, pH 6.0, per 10 ml of reagent; 4.2 ml, 0.1 M acetate buffer, pH 4.0, per 10 ml of reagent; 10 cm of enzymatic SBSR; 30 cm of unmodified SBSR; pH 6.5, 0.1 M phosphate buffer as a carrier; and pump setting 80. The pump setting of 80 was not the optimum but gave higher absorbance values than the pump setting of 45, used with the Trinder reaction (63). For the LMG and Trinder reaction, the slopes in Figure 11 were 0.1147 and 0.0406 A/mM, and the standard errors of the estimate (relative to the mean of absorbance), were 1.55 % and 1.78 %, respectively. In addition, absorption measurements at 620 nm, as used for LMG, are often an obvious advantage, because fewer potentially interfering materials absorb significantly at Absorbonce 1.6-4 LMG 1.4—4 / Tflnder T 0.0 r T I I r I—_ r r 0.00 0.04 0.08 0.12 0.16 0.20 Hydrogen Penmdde Onbfl Figure 11. Comparison of LMG and Trinder reaction using hydrogen peroxide. The lengths of the enzymatic and plain SBSRs were 10 cm and 30 cm, respectively. Absorbonce Figure 12. [ LMG I 1 I Tflnder #1 I l . 0.6 0.7 I l I l I l 1 l 1 I 0.1 0.2 0.3 0.4 0.5 Ghncose concentroflon, wan Comparison of LMG and Trinder reaction using a glucose oxidase reactor. The lengths of the enzymatic and plain SBSRs were 10 cm and 30 cm, respectively. 40 this wavelength than at 510, 520 or 395-460 nm, the wavelength used in analyses with the Trinder, benzidine or o-dianisidine, reactions, respectively (73). For the initial optimization of the LMG indicator reaction univariate methods were carried out in order to optimize primarily the conditions of the reaction. The simplex optimization methods were used in order to optimize the indicator reaction adapted by the FIA system. For the univariate experiments all the FIA parameters such as flow rate, tubing size, injected sample size, and pH of the carrier stream were kept constant. The variables that optimized. were the following: concentrjition of the LMG stock solution; concentration of the LMG rea ent solution; ectivitv of peroxidase; end pH of the LMG stock solution. For the general simplex optimization, nine variables were employed in the procedure. Those nine variables were: the um settin , the carrier pH, the carrier concentration, the length of enzype SBSR, and the length of plain SBSR as the instrumental variables; and the peroxidase activity, the volume of the LMG stock solution, the volume of 0.1 M phosphate buffer‘ pH 6.0, and the volume of 0.1 M acetate buffer, pH 4.0, used for the preparation of 10 ml of reagent. a. Concentration of LMG stock solution According to Ahlquist (73), on a molar basis, the optimum concentration of LMG was 50 % less than all the other dyes tested (benzidine, o—dianisidine), while the optimum concentration of H202 for LMG was 25 % less than all the other dyes tested. This difference in H202 requirement seems reasonable, because twice as many hydrogen atoms are lost from LMG in the oxidation reaction than for any of the other dyes tested by the same author (Figure 9). This fact makes LMG very 41 sensitive even at very low concentrations. Also, according to the same author (73), the sensitivity is dependent on the ratio of the LMG to hydrogen peroxide. In excessive amounts of either LMG or hydrogen peroxide, the peroxidase activity is inhibited. This could have been a disadvantage of the LMG, but working with low concentrations of LMG stock solutions the absorbance values were in a desirable range (up to 1.6) and the hydrogen peroxide concentration low. For concentrations up to 0.3 mM H202 the peroxidase enzyme was not inhibited (73). The optimum concentration for the LMG stock solution was selected to be approximately' 1.515 mM for 0.06 mM. H202 solution (Figure 13). The experimental conditions applied here were as follows: 1M acetate buffer, pH 2.25, for the LMG stock solution; 143 units of peroxidase per 10 ml of reagent; 12 x 10'2 mM LMG in the reagent solution; 5.0 ml, 0.1 M phosphate buffer, pH 6.0, per 10 ml of reagent; 4.2 ml, 0.1 M acetate buffer, pH 4.0, per 10 ml of reagent; 10 cm of enzymatic SBSR; 30 cm of unmodified SBSR; pH 6.5, 0.1 M phosphate buffer as a carrier; and pump setting 80. For the preparation of 10 ml of reagent solution, 143 units of peroxidase were dissolved in 5.0 ml of 0.1 M phosphate buffer, pH 6.0. To this 0.8 ml of 1.515 mM LMG stock solution and 4.2 m1 of 0.1 M acetate buffer, pH 4.0, were added. Always, the same sequence was remained for the preparation of the reagent solution. For such low H202 concentrations (0.06 mM) the absorbance values were in the range of 0.3-0.64 for LMG concentrations in the range of 0.303-15.151 mM (Figure 13). In order to ascertain the stability of the LMG/peroxidase reagent, the activity of this solution was examined at room temperature every 30 min. Comparing 1.515 mM and 3.030 mM LMG, at higher concentrations of Absorbonce, 620 nm 42 0.8 0f7-J J 0.6—J amt 0.5-— .—J 0.4-e —( 0.3—1 l 0.2—4 and 0.1-4 -4 l . ~ I . C10 013 Figure 13. I l l I r T I V T j! I 2.0 r“ I 4u0 i 6u0 T 813 l 1430 I“ I I . 101) 1213 16x3 LMG concentration, mivi Relationship between LMG concentration and absorbance of the H202 reaction product. The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively. 43 H202, such as 0.12 mM, the two mixed solutions showed no considerable differences (Figures 14, and 15). A small loss of activity can be seen over the 270 min periods for both LMG concentrations. The experimental conditions for this comparison were as follows: 1M acetate buffer, pH 2.25, for the LMG stock solution; 143 units of peroxidase per 10 ml of reagent; 12 x 10.2 mM LMG of reagent solution; 5.0 ml, 0.1 M phosphate buffer, pH 6.0, per 10 ml of reagent; 4.2 ml, 0.1 M acetate buffer, pH 4.0, per 10 ml of reagent; 10 cm of enzymatic SBSR; 30 cm of unmodified SBSR; pH 6.5, 0.1 M phosphate buffer as a carrier; and pump setting 80. As can be seen in Figure 14 and 15, the solutions prepared from 1.515 mM and 3.030 mM had a similar profile of losing activity over time ( up to 270 min), and the same maximum absorbance of 1.27 at time zero, which is sufficient and accurate for making calibration curves and measuring real food samples. For low concentrations of H202, the mixed solution which was prepared with the 1.515 mM stock LMG solution was slightly more sensitive than that prepared with 3.030 mM (Figure 16). When the concentration of H202 was increased, the mixed solution, which was prepared with 3.030 mM LMG stock solution, had the disadvantage of losing its sensitivity. For the LMG stock solutions of 0.303, 1.515 and 3.030 mM, the slopes of the Figure 16 were 0.1225, 0,1147, 0.1048 A/mM, and 'the standard errors of the estimate (relative to the: mean. of absorbance), were 1.16 %, 1.32 %, 1.45 %,, respectively; Also, disadvantages such as precipitation in the plain reactor and flow cell occurred with the 3.030 mM stock LMG solution. As a result, the instrument was noisy and sluggish in returning to the base line for a new injection. In this case more reagent and more solution were consumed Absorbonce, 620 nm 44 1.4 _( ‘1 the so min 12—1 t2= 90 min I t3= 180 nfin t4= 210 nfin l-O—J t5= 270 min 0.8—l ..T 0.6— 0.4—a 0.2— ..l 0.0 I f T r r F I 0.0 20.0 40.0 60.0 80.0 Time (min) Figure 14. Loss of LMG reagent activity during storage up to 270 min. solution was 1.5 mM. x-Axis is time of signal detection. The concentration of the stock LMG The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively. 100.0 0nrn C L Absorbonce, 6 45 1.4 d “ t1= so min *2 t2= 90 min 12.4 *3 t3= 180 min 4 ’4 {4: 210 min 1 O __ t5 t5= 270 min ; 0.8— 0.6— i l _ I i 0.4—4 ‘ g 0.2— i 0.0 I I I l I I 1 I I 0.0 20.0 40.0 60.0 80.0 100.0 Time (min) Figure 15. Loss of LMG reagent activity during storage up to 270 min. The concentration of the stock LMG solution was 3.0 mM. The lengths ofenzyme and plain SBSR were 10 cm and 30 cm,respectively X—Axis is time of signal detection. Absorbonce, 620 nm Figure 16. 2.0 46 1.8—~ 1.6~ 1.4-4 1.2—4 1.0-— 0.8-4 0.64 0.4—4 d 0.24 J . 3 mMLMG 1.5 mM LMG 0.3 mM LV / 0.0 0.00 r *— I I 0.16 0.20 f r I I I I 0.04 0.08 0.12 Hydrogen Peroxide (rnIvI) Comparison of two concentrations of LMG stock solution for reaction sensitivity. The lengths of enzyme and plain SBSRs were 10 cm and 30 cm, respectively. 47 The biggest disadvantage was the insolubility of the LMG at higher concentrations. b. Concentration of LMG Reagent Solution Since the concentration of stock solution of LMG of 1.515 mM was sufficient for linearity and sensitivity in food applications, the concentration of the LMG reagent solution was optimized. This experiment was performed only to find an optimum volume (ml) of the stock LMG solution needed to prepare 10 ml of reagent. The experimental conditions applied here were as follows: 1M acetate buffer, pH 2.25, for the LMG stock solution; 143 units of peroxidase per 10 ml of reagent; 5.0 ml, 0.1 M phosphate buffer, pH 6.0, per 10 ml of reagent; 4.2 ml, 0.1 M acetate buffer, pH 4.0, per 10 ml of reagent; 10 cm of enzymatic SBSR; 30 cm of unmodified SBSR; pH 6.5, 0.1 M phosphate buffer as a carrier; and pump setting 80. Figure 17 shows the optimum concentration of LMG reagent solution to be 9.09 x 10'-2 mM (equivalent to 0.6 ml of 1.515 mM LMG stock solution), when the peroxidase activity was 143 units per 10 m1 of reagent. The standard solution of H202 was 0.1 mM and the absorbance found under the optimum conditions was 1.42. c. Activity of Peroxidase Once the concentrations of stock and reagent solutions of LMG were optimized, the testing of the activity of peroxidase was carried out using 72, 143, and 178 units of peroxidase per 10 m1 reagent. This experiment was performed only to find an estimate of the optimum activity of the peroxidase enzyme. The experimental conditions applied here were the following: 1M acetate buffer, pH 2.25, for LMG stock 6201mm L. N I 178 Unfis PO 143 unfis PO 72 unfis P0 Absorbdncc .0 0‘) I C) 3 | I I I I .(D—I W 5.0 A 7.0 910 win 1.3.0 5 LMG 8e gen V 0 nt ConcenUcUon (.0 0.01 an) Figure 17. Relationship between LMG reagent concentration and absorbance of reaction product at three concentrations of peroxidase (P0). The lengths of enzyme and plain SBSRs were 10 cm and 30 cm respectively. The standard hydrogen peroxide solution was 0.1 mM. 49 solution; 5.0 ml, 0.1 M phosphate buffer, pH 6.0, per 10 ml of reagent; 0.6 ml of 1.515 mM LMG stock solution; 4.4 ml, 0.1 M acetate buffer, pH 4.0, per 10 ml of reagent; 10 cm of enzymatic SBSR; 30 cm of unmodified SBSR; pH 6.5, 0.1 M phosphate buffer as a carrier; and pump setting 80. Figure 17 shows the optimum LMG reagent concentration to be 9.09 x 10‘2 the at all these different activities of peroxidase. The concentration of the standard H202 used was 0.1 mM. Since the results for 143, and 178 peroxidase (PO) units at the optimum LMG reagent concentration are very close, the value of 143 units peroxidase per 10 ml of reagent solution was selected as optimum for technical and economic reasons . d. pH of LMG Stock Solution According to Ahlquist (73) LMG is sparingly soluble in water, but is quite soluble in low concentrations, in organic liquids, such as acetic acid. It is nearly ideal in producing a intense and stabile color with low reagent blank. LMG was tested for solubility and stability in solution with four preparations of acetate buffer with different molarities and acidities. In phosphate buffer LMG was oxidized to MG+, as noted by the stock solution turning dark green upon preparation. Also, the absorbance values obtained with the phosphate buffer preparation of LMG were low. The optimum conditions of LMG stock reagent concentrations and peroxidase activity were used for the preparation of the reagent solution. The concentration of the standard H202 was 0.1 mM. Table 2 shows that the absorbance values were higher at a carrier pH of 6.0 with all LMG stock buffers. Also, the absorbance values increased when the Table 2. Comparison of Buffer Conditions for LMG Stock Solution Absorbance (620 nm) Carrier pH LMG Stock Buffer 6.85 6.00 Acetate 1.0 M, pH 3.00 0.45 0.66 Acetate 1.0 M, pH 2.25 1.40 1.60 Phosphate 1.0, M pH 1.65 0.20 0.30 Acetate 2.0 M, pH 1.65 2.00* 2.20* 30% Acetic Acid, pH 1.65 2.20* 2.40* * Values at upper limit of detection 51 pH values of the LMG stock buffers were lower. LMG gave the best results in 30 % acetic acid adjusted to pH 1.65. In general the solubility of LMG increased in higher acidity. e. General Simplex Optimization of FIA The major assumption during the univariate experiments was that there was no interaction between variables. The nine variables mentioned previously were studied in order to find the optimum indicator reaction activity and the optimum FIA system. Under the conditions of optimization of the nine different variables, which may interact with each other, the univariate method of optimization has the disadvantage of requiring several thousand experiments. Therefore, a better alternative is the simplex optimization (83), which allows simultaneous variation of all parameters. The simplex method is widely applied and accepted in many research areas. A simple two dimensional surface, as pictured in Figure 18, can be used to illustrate the principles employed. The x and y axes represent the two parameters to be varied and each concentric circle represents combinations of those two which have the same response. This surface can be thought of as a topographical map; as the circles get smaller the response increases in magnitude. The simplex is generated initially by choosing a set of experimental conditions which are known to be suboptimal. After the response from that experiment is obtained another set of conditions is specified. This process is repeated until the geometric shape made of n+1 vertices is obtained, where n is the number of parameters. In the two parameter case, three experiments are required and the simplex takes the shape of a triangle. 52 .1". “- \AVL/ X Figure 18. Movements of standard step size simplex procedure on response surface. The initial simplex is 1, 2, 3 and the optimum region lies close to point 12. 53 The response at each vertex is ranked based on its magnitude: the largest is taken as best, the next largest as next best, etc. In the triangular case, the point that gives the worst response (point 1 in the Figure 18) is reflected an equal distance through the line between the other two points. A new set of conditions is specified, the response obtained, and the responses are again ranked. In this example, the response surface is shown so that the movement of the simplex can be understood. Normally, the response surface is unknown and becomes defined by the movement of the simplex. As the optimum set of conditions is reached the simplex may begin to oscillate. Such behavior can be thwarted by reflecting the next best point rather than the worst. The response function chosen was based only on the maximum absorbance obtained. It is also possible (64) to optimize a more complex function that includes response time and peak width as well as peak absorbance. The whole simplex optimization was carried out only with standard hydrogen peroxide solutions, in order to avoid the problems arising from loss of activity of the immobilized enzymes with time and the consequent enprecision. Basically the reason for doing a preliminary optimization of all nine variables was to find an estimate of the optimum conditions of the FIA system using the optimum experimental conditions of the univariate methods for the indicator reaction and adapting that into the FIA system. The initial simplex was obtained by entering the information listed in Table 3 into the modified simplex program which was run on an IBM PC compatible microcomputer. From the univariate experiments that were done prior to this optimization, the acceptable range for each of the parameters were identified. For the parameters that were not tested 54 Table 3. Range and Precision of Variables for Simplex Optimization Experimental Forward Reverse Variables Boundary Boundary Precision Pump Setting 99 10 5 Carrier pH 8 5 0.5 Carrier Concentration. 0.3 0.05 0.05 Peroxidase Activity 322 72 36 LMG Stock Vol.(ml/10ml) 1 0.1 0.1 Phosphate Buffer(ml/10ml) 5 0 0.1 Acetate Buffer(ml/10ml) 4 0 0.1 Length Enzyme SBSR(cm) 14 8 2 Length Plain SBSR(cm) 4O 20 5 Response to optimize: Absorbance Precision of Absorbance: 0.05 55 in the univariate experiments, we used reasonable values to start with. The settings listed under "Reverse" are those which gave low absorbance values and those under "Forward" gave high absorbance values (Table 3). Since there were nine variables, ten experiments were needed to form the initial simplex. These were performed and the results entered. The first point of the simplex was the response obtain from the baseline experiment with the initial conditions. Therefore, each time the simplex program specified a set of experimental conditions, this experiment was performed and the absorbance value was entered. Sixty- two times the program specified a set of experimental conditions and sixty—two absorbance values entered. Table 4 shows the current optimum values which are the current top 5 values after sixty-two experiments. The concentration of standard hydrogen peroxide was 0.012 mM. Six replications were done for each set of experimental conditions, and the 56 mnan eon m amcpm a. stnwmw use OanBmH mxpmnwamsnmw OODQHnHODm mow nzm wwos Hammonwo: >swwomnmnm mwmm twee: 20 made. m: 00:. 00:. mammmfi mammmm masked mmmw H no m.mm o.Ho Ho o.m w.o a.» Ho we we we m.o o.~o m o.m H.m q.m pm we wp mo m.m o.Hm Ho o.n ~.q m.m Ho mo we em m.m o.Ho m o.o ~.w m.o HM um mm mo m.m o.Ho m o.m N.» u.o Hm um mo om m.o o.Ho m o.m ~.w a.w Ho wo zoos" um m.m o.Hw q.m o.m N.» q.o HH.~ ww mo" ~.qn o.m~ o.oq H.mo o.oq o.u o.ww H.Ho N.ca >cmonomsom mwo :3 o.~wm o.~wa o.won o.wHo o.wNm o.waw o.me o.ow mo" mnmomwma Um ‘ ‘ 9 J | / 99 The t test for glucose samples gave a calculated t value (2.139) smaller than that of the table for a two tailed test with a=0.01 (ta/2,6=3'7°7)° Therefore, for the level of a=0.01 the null hypothesis was accepted as reasonable (pi = 2.0). On the other hand, the t test for fructose and sucrose was significant (14.97, 15.591 respectively), and the null hypothesis was rejected for these two sugars. Hence the conventional method appears to give slightly low values for fructose and sucrose. If we take into account the fact that six out of nine values obtained by the FIA method for the food sugars were higher than those obtained by the conventional method, it may be concluded that the FIA method may be more accurate for those nine food samples showing statistical differences by the two methods. In addition, statistical comparison by the least significant difference procedure between standard pure sugar solutions with the FIA method (with three replications) and the Conventional method (with three replications) shows that the two methods are not significantly different at the 99 % confidence level. Hence we may conclude that the FIA method is at least as accurate as the conventional method and may, in fact, be more reliable. Further comparisons need to be done to prove or disprove this last.hypothesis. Q’TT’ P E; ijj—h f-i \ c". 17“); x ’“if‘; ‘< , §?:T:?2JC//) ,7 l " W1 5 #11 y SUMMARY In this work a new dye, leucomalachite green (LMG), was introduced for the indicator reaction of the Flow Injection Analysis (FIA) determination of glucose, based on immobilized glucose oxidase. First, an initial optimization of the indicator reaction was carried out using univariate methods and optimizing the concentration of the LMG stock solution, the concentration of the LMG reagent solution, the activity of peroxidase, and the pH of the LMG stock solution. Then, a general simplex optimization of the FIA system was performed for the following nine variables: pump setting, pH of the carrier, concentration of the carrier, activity of peroxidase, volume of LMG stock solution (ml/10ml), volume of 0.1 M phosphate buffer, pH 6.0 (ml/10ml), volume of 0.05 M acetate buffer, pH 4.0 (ml/10ml), length of enzyme Single Bead String Reactor (SBSR) (cm) and length of plain SBSR (cm). The experimental conditions used as the initial points for the general simplex optimization were the optimal of the initial optimization. The general simplex optimization after sixty-two runs gave 33 % improvement in absorbance values. A more specific optimization of rate dependent variables, such as flow rate, length of enzymatic and unmodified SBSRs and pH of the carrier were carried out using univariate and simplex optimization methods. A new set of optimum values was obtain by those two methods. An additional improvement of 7 % in absorbance values was achieved. A comparison of the specific simplex and univariate methods of optimization was done by comparing the slopes of two calibration curves which were obtained by applying the optimum conditions for each method. The calibration curve obtained, using the optimum values given by the 100 Simplex method gave a slightly higher sensitivity than that of the univariate method. The optimum values which were obtained after the optimization of the FIA system for all variables were the following: flow rate, 1.47 ml/min; 0.1 M phosphate buffer, pH 6.0, as a carrier; peroxidase activity 143 units per 10 ml of reagent; concentration of 2 mM; concentration of LMG stock LMG reagent solution, 9.09 x 10- solution, 1.5 mM; 0.1 M phosphate buffer, pH 6.0, 2.4 ml per 10 m1 reagent solution; 0.1 M acetate buffer, pH 4.0, 7 ml per 10 ml reagent solution; length of enzyme SBSR, 11.6 cm; length of unmodified SBSR reactor, 35 cm; and temperature 40 oC. The injected volume of the sample was 30 pL. The above optimum conditions were applied to the determination of glucose, fructose and sucrose in citrus fruits (oranges, lemons, and grapefruit), thirteen olive samples and ten cherry samples. Conventional enzyme methods for the analysis of these sugars were also used. When pure sugar solutions were analyzed by the FIA and conventional methods, the agreement was satisfactory. When the sugars present in the fruit samples were analyzed by the two methods, the results in nine out seventy—eight samples were different at the 0.01 probability level. This discrepancy may be due to the fact that only one value for each food samples was available by the conventional method, compared to six such values by the FIA method. The FIA method appears to be as accurate as the conventional method. 10. ll. 12. 13. REFERENCES J. Ruzicka and E. H. Hansen, 1975, Flow Injection Analyses; PART I. A New Concept of Fast Continuous Flow Analysis, Anal. T. E. Doty, E. Vanninen, 1975, Crystalline fructose: Use as a food ingredient expected to increase, Food Technol., 29(11), 34- 38. M. A. JOslyn, 1970, Methods in Food Analysis, Physical, Chemical, and Instrumental Methods of Analysis,Academic Press, N.Y. Y. M. Wang and J. V. Eys, 1981, Nutritional Significance of Fructose and Sugar Alcohols., Ann. Rev. Nutr., 1, 437-475. J. Ruzicka and E. H. Hansen, 1981, Flow injection Analysis, John Wiley and Son, New York. J. Ruzicka and E. H. Hansen, 1988, Flow Injection Analysis, John Wiley and Son, New York Anal, Second Edition. J. T. Vanderslice, K. K. Stewart, A. G. Rosenfeld and D. Higgs, 1981, Laminar Dispersion if Flow Injection Analysis, Talanta, 28, 11. W. E. Hornby and H. H. Filippusson, 1970, The Preparation of Trypsin. Chemically attached to Nylon Tubes, Biochim. Biophys. Acta, 220, 343. M. D. Lilly, and P. Dunnill, 1971, Biochemical reactors, Process Biochem., 6, 29. T. Kobayashi and M. Moo-Young, 1971, Backmixing and mass transfer in the design of immobilized-enzyme reactors, Biotechnol. Bioeng., 13, 893. R. Tijssen, 1980, Axial Dispersion and FLow Phenomena in Helically Coiled Tubular Reactors for Flow Analysis and Chromatography, J. H. Van Den Berg, R. S. Deelder and H. G. M. Egberink, 1980, Dispersion Phenomena in Reactors For Flow Analysis, Anal. Chim. Acta, 114, 91-104. B. Olsson, and L. Ogren, 1983, Optimization of peroxidase immobilization and of the design of packed -bed enzyme reactors for Flow Injection Analysis, Anal. Chim. Acta, 145,87. 102 103 14. J. M. Reijn, W. E. Van Der Linden and H. Poppe, 1981, Dispersion in Open Tubes Packed With Large Glass Beads, Anal. Chim. Acta, 123, 229—237. 15. J. M. Reijn, H. Poppe, W. E. Van der Linder, 1984, Kinetics in a single bead string reactor for flow injection analysis, Anal. Chem. 56, 943. 16. B. Alexander, A. Rimon, and E. Katchalski, 1965, Effects of Water- Insoluble Trypsin Derivatives on Fibrinogen, Fed. Proc., 24, 804. 17. A. B. Eli and E. Katchalski, 1960, A Water—Insoluble Trypsin Derivative and its Use as a Trypsin Column, Nature, 188, 856-857. 18. A. B. Eli and E. Katchalski, 1963, Preparation and Properties of Water-Insoluble Derivatives of Trypsin, J. Biol. Chem., 238,1960. 19. R. Goldman, I. H. Silman, S. Caplan, O. Kedem and E. Katchalski, 1965, Papain Membrane on a Collodion Matrix: Preparation and Enzymic Behavior,Sience, 150, 758. 20. H. H. Weetall and N. Weliky, 1966, A New Technique for the Enzymic Detection of Hydrogen Peroxide, Anal. Biochem., 14,160. 21. L. Goldstein, Y. Levin and E. Katchalski, 1964, A Water Insoluble Plyamionic Derivative of Trypsin. II. Effect of the Plyelectrolyte Carrier on the Kinetic Behavior of the Bound Trypsin, Biochemistry, 3, 1913. 22. L. B. Wingard, C. c. Liu and N. L. Nagda,1971, Electro chemical measurements with glucose oxidase immobilized in polyacrylamide gel: constant current voltametry, Biotechnol. Bioeng., 13, 629. 23. E. M. Grook, K. Brocklehurst and C. W. Wharton, 1970, Cellulose- insolubilized enzymes, in Methods Enzymol., Vol. 19, Academic Press, New York. 24. M. A. Mitz and L. J. Summaria, 1961, Synthesis of Biologically Active Cellulose Derivatives of Enzymes, Nature, 189, 576-577. 25. W. E. Hornby, M. D. Lilly and E. M. Crook, 1968, Some Changes in the reactivity of enzymes resulting from their chemical attachment to water-insoluble derivatives of cellulose, Biochem. J., 107,669. 26. E. K. Bauman, L. H. Goodson, G. G. Guilbault and D. N. Kramer, 1963, Preparation of Immobilized Cholinesterase for Use in Analytical Chemistry, Anal. Chem. 37, 137- 1381. 27. S. J. Updike and G. P. Hicks, 1967, The Enzyme Electrode, Nature, 214, 986-988. 28. S. Updike and G. Hicks, 1967, Reagentless substrate analysis with immobilized enzymes, Science, 158, 270. 29. 30. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. I. H. Silman and E. Katchalski, 1966, Water—insoluble derivatives of enzymes, antigens, and antibodies, Ann. Rev. Biochem., 35, 873. H. H. Weetall, 1969, Trypsin and papain covalently coupled to porous glass: preparation and characterization, Science, 166, 615. H. H. Weetall, L. S. Hersh, 1969, Urease covalently coupled to porous glass, Biochim. Biophys.Acta, 185, 464. H. H. Weetal, 1971, Enzymes immobilized on inorganic carriers, Biochem. J., 117, 257. H. H. Weetal, 1970, Storage stability of water—insoluble enzymes covalently coupled to organic and inorganic carriers, Biochim. Biophys. Acta, 212, 1. P. V. Carr, and L. D. Bowers, (1980), Immobilized enzymes in analytical and clinical chemistry, John Wiley and Sons, New York. R. A. Messing, (1975), Immobilized enzymes for industrial reactors, Academic Press, New York. N. M. Green, 1963, Avidin: The nature of the biotin —binding site, Biochem. J., 89, 599-620. U. D. Alwis and G. S. Wilson, 1989, Strategies for the reversible immobilization of enzymes by use of biotin -bound Anti-Enzyme Antibodies, Talanta, 36, 249—253. W. U. De Alwis, B. S. Hill, B. I. Meiklejohn and G. S. Wilson, 1987, Reversibly Immobilized Glucose Oxidase in the Amperometric Flow—Injection Determination of Glucose, Anal. Chem., 59, 2688- 2691. E. A. Bayer and M. Wilchek, 1980, Methods of Biochemical Analysis, Vol. 26, D. Glick, ed Wiley. New York. S. Luo and D. R. Walt, 1989, Avidin—Biotin coupling as a general method for preparing enzyme—based fiber-optic sensors, Anal. Chem., 1069-1072. T.A. Kelly and G. D. Christian, 1982, Homogeneous Enzymatic Fluorescence Immunoassay of serum IgG by continuous flow-injection analysis, Talanta, 29, 1109 -1112. P. J. Worsfold, A. Hughes, 1984, A model immunoassay using automated flow injection analysis, Analyst, 109, 339-341. A. Hughes, P. J. Worsfold, 1985, Monitoring of Immunoprecipitin Reactions using flow injection analysis, Analyst, 22, 16-17. W. U. Alwis, B. S. Hill, I. B. Meiklejohn, and G. S. Wilson, 1987, Reversibly immobilized glucose oxidase in the amperometric flow injection determination of glucose, Anal. Chem., 59, 2698-2691. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 105 W. U. Alwis, and G. S. Wilson, 1987, Rapid heterogeneous competitive electrochemical immunoassay for IgG in the picomole range, Anal. Chem., 59, 2789-2794. W. U. Alwis, and G. S. Wilson, 1985, Rapid sub-picomole electrochemical Enzyme immunoassay for immunoglobulin G, Anal. Chem., 57’ 2756-27580 G. Johansson, L. Ogree, B. Olsson, 1983, Enzyme reactors in usegmented Flow Injection Analysis, Anal. Chem. Acta, 145, 71. G. F. Bickerstaff, 1978, Enzyme in Industry and Medicine, Edward Arnold, 8. N. Gochman and J. M. Schmitz, 1972, Application of a new Peroxide Indicator Reaction to the Specific, Automated Determination of Glucose with Glucose Oxidase, Clin. Chem. 18, 943-950. L. G. Morin and J. Prox, 1973, Single glucose oxidase -peroxidase reagent for two-minute determination of serum glucose, Clin. Chem., 19, 959-962. C. W. Bradberry and R. N. Adams, 1983, Flow Injection Analysis with an Enzyme Reactor Bed for Determination of Ascorbic Acid in Brain Tissue, Anal. Chem. 55, 2439- 2440. P. J. Worsford, 1983, Enzyme Assays Using Flow Injection Analysis, Anal. Proc., 20,486. K. Nobutoshi, S. Hori and M. Furusawa, 1989, Flow Injection Determination of Branched Chain L-Amino Acids with Immobilized Leucine Dehydrogenase, Anal. Chim. Acta, 218,161-166. B. A. Petersson, 1988, Amperomertic Assay of Glucose and Lactic Acid By Flow Injection Analysis, Anal. Chim. Acta, 209, 231-237. P. Solich, M. Polasek, R. Karkicek, O. Valentova and M. Marek, 1989, Spectophotometric Flow-Injection Determination of Urea in Body Fluids By using an Immobilized urease Reactor, Anal. Chim. Acta, 218, 151 -155. Bo Olsson, 1988, A Flow Injection System Based on Immobilized Penicillinase and a Linear pH Buffer for Potentiometric Determination of Penicillin, Anal. Chim. Acta, 209,123-133. N. Kiba, S. Hori and M. Furusawa, 1989, Flow-Injection Determination of Branched-Chain L-Amino Acids With Immobilized Leucine Dehydrogenase, Anal. Chim. Acta, 218, 161-166. Ala'Dini M. Almuiabed and .Alan, Townshend, 1989, Flow-injection determination of malate with immobilized malate dehydrogenase Anal. Chim. Acta, 221, 337-340. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. M. Masoom and A. Townshend, 1985, Determination of Cholesterol by Flow Injection Analysis With Immobilized Cholesterol Oxidase, Anal. Chim. Acta, 174, 293-297. J. M. Fernandez—Romero, M. D. Luque de Castro and M. Valcarcel, 1987, Enzymatic Determination of Total Cholesterol in Serum by Flow Injection Analysis, J. Pharm. and Biom. Anal., 5, 333-340. J. M. Fernandez-Romero, M. D. Luque de Castro and M. Valcarcel, 1987, Determination of Total Cholesterol in Serum by Flow Injection Analysis with Immobilized Enzymes Clin. Chem. Acta, 167,97-104. C. L. Patton and S. R. Crouch, 1986, Experimental Comparison of Flow Injection Analysis and Air Segmented Continuous Flow Analysis, Anal. Chim. Acta, 179,189—201. C. L. M. Stults, Practical considerations for the use of immobilized enzymes in flow injection analysis. PhD Thesis, Michigan State University, East Lansing, MI, USA. C. L. M. Stults, A. P. Wade, and S. R. Crouch, 1987, Computer Assisted Optimization of an Immobilized Enzyme Flow Injection System For the Determination of Glucose, Anal. Chim. Acta, 192, 155-163. C. L. M. Stults, A. P. Wade, and S. R. Crouch, 1987, Immobilized enzyme Kinetic study of D-glucose mutarotation by flow injection analysis, Anal. Chem., 59, 2245—2247. P. Trinder, 1969, Determination of glucose in blood using glucose oxidase with an alternative oxygen acceptor, Ann. Clin. Biochem. 6, 24. M. J. Medina, J. Bartroli, J. Alonso, M. Blanco, J. Fuentes, 1984, Direct determination of glucose in blood serum using Trinder reaction, Anal. Lett., 17, 385. J. A. Lott and K. Turner, 1975, Evaluation of Trinder’s glucose oxidase method for measuring glucose in serum and urine, Clin. Chem., 21, 1754-1760. D. Barham and P. Trinder, 1972, An Improved Colour Reagent for the Determination of Blood Glucose by the Oxidase System, Analyst, 97, 142-145. H. A. Mottola, 1983, Enzymatic Preparations in Analytical Continuous Flow Systems, Anal. Chim. Acta, 145, 27-39. P. J. Worsfold, 1983, The Bio—Analytical Potential of Flow Injection Analysis, Anal. Chim. Acta, 145, 117-124. lb" ‘. n i-"""“ " .I 1 .~. .rfi 'I v ' ‘ l J r , . ; J . 4 I I “,1 - 4" ~ 1 \ u l --" b ‘ ” I’m I" ’1 (4 fl] ’ Ll) .‘J ’/’\' ) L/J ‘l‘ ‘4' I r I f . 1...-¢z‘~“"¥i on-s' 0 l l "\._ l / I l I ‘ d I .‘ l l ‘ / I I 'v '4 .4- . ‘1 i ‘ ‘ ~~ ,-.. a -“-~7.- . \ \