"L ":III‘III. ’ ‘I1'1 1'" "‘I I: I ' "I 11' III. '.33 II. {34.33.33 33.3% HM ”3333 I- Ixifi'y ""1" . "I! I 2311’ 333333333 I.I...I.I}I{I 33331.; 1393? mm A 333—..313133 ., ,, W W. 333 '1' . 3 3 l '31I'1""1.L31WII, 1.1%":1 .12.. “Q. 33333.3. .33 "A . " I. I.'I.I§I'.I. ‘ «I-w‘II-II. «I... .1 ‘31::"" 1‘. I «313‘ I. 337.. lgI 333 “M .I..I..LI;.3. II (33333 ‘Iw..:|,.133.3.33333333. “.1‘12'1‘1‘1J‘2’I ._ 333133 'LI‘; "I “II . HAIL. ILL 125%.; .33 no»... . .II...I;:II~III "' “VI-'1 III... a.“ . 21.1 .3313 1.33!" {$33 HQLQ‘E‘" .Id ""1193133. "CI". III'III ._ I... IIIII‘. III1 III. I351; :43!“ “ 3.3333131433339333 ' 333.3333. . _ 3 33333333333 """"" "fix" "I. "'2‘: 31: Y L .I. ""' """""'""""'::"""1 "'3' 3"". 2:19. ‘w . "1’5" . ‘L‘I -- I.«I.333II13 333 33 3333333333 "L'L' v 13-: “alkfiufi.” uh. 2:53 L331 _ {EN 1 ., ”.2 1-1.33'113"‘;' 1‘ 'm‘III‘ 53.333 . 3 '1...wa3.33..3, "3:I1 v .'. 'I;I‘I:_ I‘I II I'Ig' 1.,1 "' II"1:.3I "3.13"" ' ' 1 ' I; "I:.:I";I . "I‘1':I‘-I:y 31‘1333 I“: I . I,1 u ., 3333. "3' ‘ ' I‘I“ I 1 I ,3. "'I",'II . 3.':33i31:13333 3:3. " ('IlIYII 13.313" 3 "'31 L'.1§:1,: ‘.3_‘~.6 33'33 3333.333 .. 3 " "" 1"y""'5'"\1"5£11131 I'H‘ : II ":1'133.3I1‘1'1L3333333 I.“‘I1 “‘ 3'1"“ M: 333333 "‘1‘: MW .L. " I.’ Itm "‘31."."1 5:}; . _ '"I I QII‘ "111:1‘1'1 “133333333333“ "'""'.I’.‘3 1.1 III ’ 3 33333 3 3 " ' "" "" """"~ . II “I11.”3 .3. 3 3 33 3 333333 3333 3 1‘1‘."‘31 "‘9 "‘ , 3L 3 .331. ,33333333 I ‘I' 333 I'mdv 3'" 333.3333 333 ' .1333 "'3‘!" '11)“ "1‘1 "‘11:": ‘1"‘I"‘I'1‘I"I . . I 33331111§1 33.3....II II I1 . ' . .1." II ('1I ~ . {“4" 3. [33%1‘." .. I1. " "' "' "1,~"'I ' "" ' "" ' I'I‘I "If I' ”1'1"". 1.1"'I.1I:::I.‘l. """.'""‘1""""1‘. I3, 33 333 333333, I. . J """"""1' I III? , ""'""‘""""""""' ""1‘1"' "1'3" """"' "11‘ "WIéiw" 1. .I ‘I‘ $111 ‘I' I‘iII I .1“ 3333.33; 33 3 33 """ "'M' ""' """"""" 'I "“5 I“ ‘1".IIIIIIIII » .1333 ‘1‘! L113!” I 3:333 “1.3.5.3. .~ ,. 3 II ‘ ‘II III“. III 1.“ I"“1311 313.3335. 333 "3' 1‘1 33 L323“. "f” I 4:715:24} 5.14:. ”‘1'“. 14.1.3 """1.'1'2‘33'33w3333333333 rig ~——w ' “:5"? 752‘. 3-7.] .21:— .I_ _ _.‘Eigcj ’ $.75 III “74?... “J“;TT:L?§-:f:. 4‘13?“ 43:33:33: . . ""':' ‘1: “~45. mi .41... -_ M ' ’J‘ if}. 7 - .131)?- ”wzy- _ “‘M. c.) m. .1. _ :3"; . _3_ "‘4— $713.1: 3—1." .31"- 51:2“: «1-4513, A31?“ 72- as...“ 1:; "1- "‘ :13 , 3 3 W1” . "'1“! '3 3 333 ‘ .‘1: "1.913;: I 3.31:3} . : 1:;3 '2'13‘.‘ 1 "1‘11"""3-13'5231“ 331.1,.“ 11:1" 31,1.3‘3‘33 3 3 L C F'I'H‘I‘g' “$332333 ".81- " ' H 11:33 "" ' , I 1'I'l2713‘v‘ 3 I . 3 L ‘.’1‘.; “L j '1"? II fig; '15:“:- ~Lu’ln. "EVE ‘G‘I‘I'EI ‘39:)th ‘1... 21"};v‘h "‘3 on u, ~ «as. ., 3 33 I3L“I.,‘I,3Y "(-‘ .13.... .3." _ ,3» 33 333332;: ...‘\I.I.3"II , L’IfifiL 33333 333333 3 h .. 01.3.1.1“ 33 333333 333333 I'M'": L . 3. 1-w- §fi§1§5 3333333333333? 33 " '" 13%.: ~ 1.15": a. 1. :73: LI, "1. .IJQI. I I "' "Ln. . 3333 1. ?"""'" "1'3. ""'1.'1.1 I... ... .. " "'" """"'""""' "; "1'5 1.2"". "'K. . V 3 E3133 K "_.'I .(J 11;" 2‘1 3 3 3 333 3333333 "1157‘ , L . .3 I : 33.333.333.33 "" l ‘7: WI I .‘ 3 3 3 3 'th‘fl 1 . I . 3333 3L 3 33 33 .3 I:H 3333:3333 I‘Ifx ‘1, “33""113 “IL"? 333 ‘3 . .. - - 3313:}, E‘I '23:“ 3 3333333 3 "';':"I“‘:I'13I.g ""0"; “'3'": I‘I"I‘I‘ . "1'1" ' 3""? {‘I‘I3"1“I1‘13333-333333' «I "I" II My... """ Q‘II I“ "'1 "fix. ' W1. “VI/«wk... 33 u I ‘1 " " II 3133-5?!” II 3 3 3 L 3 3 "I1 .1”... 33:1 _ 3.3.1231 333.333.131333333353332. '1' l 3 333333 ‘14" "'1.“ ()31“ ”avg." 33:33.33: ""'"""‘ MIL..." I1.3 33";133'1' 33.33333 .1 "RICE"; ”13.3 1 “II ""‘"""""'J' I 3333 .- "'11". '1"'""'§£I3BLW‘W. I'I‘I' LIBRARY Mm SEE Michigan State , University This is to certify that the thesis entitled AN AUTOMATED ANALYTICAL SYSTEM IN WHICH AN IMMOBILIZED-ENZYME REACTION COLUMN IS COUPLED WITH THE STOPPED-FLOW MIXING TECHNIQUE presented by Daniel Joseph Kasprzak has been accepted towards fulfillment of the requirements for Ph . D. degree in _flhemm Wig / 4144 Major professor Date October 12, 1978 0-7639 AN AUTOMATED ANALYTICAL SYSTEM IN WHICH AN IMMOBILIZED-ENZYME REACTION COLUMN IS COUPLED WITH THE STOPPED-FLOW MIXING TECHNIQUE BY Daniel Joseph Kasprzak A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1978 @057 7%” ABSTRACT AN AUTOMATED ANALYTICAL SYSTEM IN WHICH AN IMMOBILIZED-ENZYME REACTION COLUMN IS COUPLED WITH THE STOPPED-FLOW MIXING TECHNIQUE BY Daniel Joseph Kasprzak A microcomputer-controlled system based on the use of an immobilized enzyme for clinical analyses is de- scribed. The immobilized enzyme is applied in the form of a small packed column. A peristaltic pump is used to pass analyte solutions through the column where the en- zyme selectively catalyzes the conversion of the sub- strate of interest. The product of the enzyme reaction is collected in a modified stopped-flow module where it is rapidly and efficiently mixed with indicator reagents. A spectrophotometric measurement is made on the resultant absorbing species and is used to calculate the concentra- tion of the analyte. Daniel Joseph Kasprzak The preparation of the immobilized enzyme involves the use of glutaraldehyde for covalent attachment to amine functionalized controlled-pore glass. Approximately 150 mg of the enzyme-glass product are packed into a jacketed glass column for use in the system. Preliminary work with immobilized glucose oxidase used for the determination of glucose is described. The work was accomplished with a manually operated system and a non-thermostated enzyme reaction column. A linear cali- bration curve was obtained in the concentration range of 5 to 100 ppm glucose with relative precision in the range of 3.0 to 16.4%. An Intel 8080A microprocessor-based microcomputer was interfaced to the system to provide control of the instru- mentation and to automate data acquisition. The hardware and software required to implement this interface are described. The design and assembly of a 12—bit analog-to- digital converter circuit board for use with the microcom- puter is also presented. Characterization studies are described in which potas— sium nitrate solutions and a column packed with untreated Daniel Joseph Kasprzak controlled-pore glass were used to simulate an enzyme— based system. Observation of the column effluent by means of a flow cell showed that the peak heights of bands eluted from the column are directly proportional to sample concen— trations. Furthermore, when the center of a band is col- lected for measurement in the modified stOpped—flow module, a direct proportionality is also observed. The preparation of an immobilized alcohol dehydrogen- ase column for use in the determination of ethanol is de— scribed. The measured value of the Km of the preparation is reported. Results obtained for the determination of glucose with immobilized glucose oxidase and the microcomputer—con- trolled system are presented. A linear calibration curve was obtained with a relative precision of <6%. A control serum sample was analyzed and the value of glucose deter- mined agreed well with the assigned value. Relative pre— cision for three determinations was <2%. Stability studies indicated that there was no detectable decrease in enzyme activity on the column over a six day interval. To my parents and family for their unquestioning support during my stay in graduate school, and to Melody for giving me a future ii ACKNOWLEDGEMENTS I would like to thank Dr. Stan Crouch for his support and helpful discussions. I gratefully acknowledge the fact that he sacrificed his summer salaries to contribute to my support during several summer terms. I would like to thank Dr. Chris Enke for serving as my second reader and Drs. Babcock and Sweeley for serving on my final oral committee. My deepest gratitude is extended to the members of the Crouch group who added an invaluable personal dimension to my life in graduate school. I am especially indebted to Gene, Roy, Gary, Sandy, and Wai for their listening ears, their help and advice, and their friendship. I acknowledge the help of Phil, Floyd, Charlie, Gary, Bob, and Dave who left before me and that of my co—worker, Marty. iii I thank the newer members of the group, Jim, Clay, Rytis, Dave, and Gene, for their friendship, and I wish them success. I am grateful to the Department of Chemistry of Michi- gan State University for financial support in the form of teaching assistantships. Finally, but most important of all, I thank my wife, Melody, for her love and support and for graciously typing the final copy of this dissertation. iv i _Mh—y -—-—. _aa ._ TABLE OF CONTENTS I. INTRODUCTION. . . . . . . . . . . Enzymes in Analytical Chemistry. . . . Immobilized Enzymes. . . . . . . . On the Organization of this Dissertation. II. HISTORICAL . . . . . . . . . Early Work with Immobilized Enzymes . . Immobilized Enzymes for Clinical Analyses . . . . . . . . . . System Approaches in Clinical Analyses . External Probes . . . . . . . . Continuous Flow Systems. . . . . . Stopped-flow Systems. . . . III. AN OVERVIEW OF THE SYSTEM _. . . . . . The Original System Configuration . . . Instrumental Components . . . . . Stopped-flow Mixing Module . . Collection Loop. . . . . . . . Peristaltic Pump . . . . . . . Sample Introduction Mechanism . . . Spectrophotometer . . . . . The Immobilized-Enzyme Reaction Column . . . . . . . . . . Operation of the System. . . . . . The Present System Configuration . . . V WM 10 11 12 14 15 17 17 17 19 19 20 20 21 21 21 23 Instrumental Components . . . . Sample Introduction Mechanism . . Operation of the System. . . . . Key Features of the System . . . . IV. PREPARATION OF THE IMMOBILIZED-ENZYME REACTION COLUMN . . . . . . . . The Choice of Controlled-pore Glass as Support Material . . . . . . . Covalent Attachment of Enzyme to Controlled-pore Glass . . . . . Description of the Procedure . . . Experimental . . . . . . . . V. PRELIMINARY WORK WITH IMMOBILIZED GLUCOSE OXIDASE . . . . . . General Description of the Work. . . A Consideration of Enzyme Kinetics. Experimental and Preliminary Results . Solutions and the Immobilized-Enzyme Column . . . . . . . . . Optimization of Push Volume . . . Calibration Curves and Analytical Results . . . . . . . . . Discussion of Results . . . . . . Conclusions . . . . . . . . . VI. MICROCOMPUTER AUTOMATION OF THE SYSTEM . The Need for Automation . . . . Description of the Microcomputer . . Control of Instrumentation . . . . vi Page 23 25 25 27 29 29 32 32 35 39 40 41 44 45 46 49 55 61 63 64 65 68 ~o—s—m:_A-' ‘_~ — w A _-w*‘ - L-‘ .M-L n-‘Afw'. .. _. <... a. .. —-—.-———- wrv— v—w twaaifl'u Page Hardware . . . . . . . . . . 68 Software . . . . . . . . . . 72 Automated Data Acquisition . . . . . 75 Hardware—-—Genera1 . . . . . . . 76 Hardware---Device Selects. . . . . 79 Software . . . . . . . . . . 79 Serial Communication Link to a Minicomputer . . . . . . . . . 80 Description of Hardware . . . . . 84 Description of Software . . . . . 86 LINK80 - The Transparent Mode. . . 86 LOAD8O - The Binary Loader. . . . 89 DATA8E - The Data Transfer Program . 90 VII. CHARACTERIZATION OF SYSTEM RESPONSE. . . 96 The Non-Enzymatic Study . . . . . . 99 Experimental and Procedure . . . . 100 Flow Cell Experiments . . . . . 100 Stopped-flow Experiments . . . . 101 Results and Discussion. . . . . . 103 Enzymatic Study . . . . . . . . 110 Flow Cell Investigation . . . . . 112 Experimental . . . . . . . . 112 Results and Discussion . . . . . 115 Determination of Km. . . . . . . 117 Experimental . . . . . . . . 117 Results and Discussion . . . . . 119 Conclusions . . . . . VIII. RESULTS WITH THE AUTOMATED SYSTEM Introduction . . . Software . . . . . General Description. The Time of Delay Subroutine. The Analog-to-Digital Conversion Subroutine . . . . Experimental and Procedure Calibration Curves . . Control Serum Standard. Results and Discussion . Conclusions . . . . . IX. FUTURE PROSPECTIVES . . . Sample Handling. . . . Microcomputer Software . CONVERS. . . . . . 7K BASIC . . . . . APPENDIX A . . . . . . . . . APPENDIX B . . . . . . . . . APPENDIX C . . . . . . . . . REFERENCES . . . . . . . . . viii Page 123 126 126 128 128 130 133 135 135 135 137 147 149 149 154 155 157 158 171 175 192 Table 10. 11. 12. 13. 14. LIST OF TABLES Immobilized Enzymes Applied to Clinical Analyses . . . . . . . . . . . . Calibration Curve Results with Immobilized Glucose Oxidase (Day 1) . . . Calibration Curve Results with Immobilized Glucose Oxidase (Day 4) . . . . . Calibration Curve Results with Immobilized Glucose Oxidase (Day 4) . . Printed Circuit Boards Used in the Microcomputer . . . . . . . . . Control Functions on the Stopped—flow Module. Command Words Used with the 8255 I/O Port. ADC Board Device Selects . . KNO3 Peak Height Results . . . . . . . KNO3 Stopped-flow Results . . . . . . Rate Data for Determination of Km of Immobilized Alcohol Dehydrogenase. . . . Glucose Calibration Curve Results . . . . Determination of Glucose Standards Based on the First Day Calibration Curve . . . . Determination of the 50 ppm Standard Based on Daily Two-point Calibration Curves . . . ix Page 50 53 56 66 72 74 81 104 106 120 139 144 146 LIST OF FIGURES Figure 1. 10. 11. 12. 13. 14. 15. Schematic representation of the original system. . . . . . . . . . . . . Schematic representation of the present system. . . . . . . . . . . The immobilization procedure for controlled- pore glass . . . . . . . . . . . The jacketed glass column . . . . . . Glucose calibration plot . . . . . . . Glucose calibration plot . . . . . . . Circuit diagram of the controller interface . Circuit diagram of the analog-to-digital converter board . . . . . . . . . Detail of connections to the ADC. . . . Flow chart for the transparent mode program . Absorbance versus concentration of potassium nitrate O 0 O O O O O O O O O O —1 Hofstee plot of V versus V°[S] Glucose calibration plot . . . . . . . Absorbance versus the day of measurement . . Proposed valve changes . . . . . . . . Page 18 24 34 37 54 57 7O 77 78 88 108 122 140 142 153 Figure 16. 17. 18. 19. Modifications on the 8080A CPU board Modifications on the 8251 board Modifications on the 8255 board Circuit diagram of the RS232 driver board. xi Page 161 162 165 . 168 _. ._.__ 44 _x....____&.. CHAPTER I INTRODUCTION Enzymes in Analytical Chemistry In the past twenty to thirty years, advances in the technology of isolation and purification of enzymes has led to the investigation of their use in analytical chem- istry (1-7). Major emphasis has been placed on clinical analyses where the enzyme is used as a tool to selectively catalyze the conversion of an analyte in body fluids to an easily monitored product. The reason for the interest in the use of enzymes stems primarily from their characteristic specificity to- ward one or a limited number of substrates. This allows the enzyme to be applied to a complex mixture of chemical species without the need for prior separation of those which might otherwise interfere. A second reason for the attractiveness of enzymes for analytical use is the fact that the enzyme is a catalyst and as such is not depleted .-... _ - _ a: .1. in a chemical reaction. Often only small amounts are re- quired for analyses. Further, since enzymes are derived from living organisms they are generally usable under near physiological conditions of pH and temperature, a fact which is advantageous for automated analyses where harsh conditions would be detrimental to equipment and for analy— ses of biological species which would not tolerate the pH extremes required by some non-enzymatic methods. There are, however, some disadvantages to the use of enzymes for analytical purposes. Since enzymes must be obtained from living organisms where they are present at very low levels, they are often quite expensive. Also, unlike other chemicals, enzymes are subject to denatura- tion during storage, both in the crystalline state and especially so in solution where they are most often used. Immobilized Enzymes The disadvantages in the use of the soluble form of enzymes for analytical procedures may eventually be alle- viated by advances in a more recent technology, namely that of the immobilization of enzymes (8-12). An immo- bilized enzyme may be defined as one which is in some way rendered insoluble while still retaining its activity. A.......g_.-_.‘A..._. 1.... _ , Such an enzyme is extremely attractive for analytical use because it has all the advantages of its soluble counter- part along with some additional advantages (12). These include: 1) the possibility for repetative use of the en— zyme, 2) the elimination of sample contamination by the enzyme, 3) the ability to terminate the enzymatic reaction without denaturation, 4) an increase in stability of the enzyme, and 5) the elimination of reagent preparation for analytical precedures employing the enzyme. In short, im— mobilization yields a selective, heterogeneous catalyst which is reusable and more stable. These characteristics tend to overcome the disadvantages in the use of enzymes mentioned earlier, and make the immobilized enzyme very well suited for automated applications. On the Organization of this Dissertation The work presented in this dissertation involves the design and characterization of an automated system which is based on the use of an immobilized enzyme for clinical analyses. The enzyme is attached to controlled-pore glass and is applied in the form of a small packed column which is coupled with the stopped-flow mixing technique in a unique approach to such an enzyme based system. The system is automated by means of a microcomputer for control of in- strumentation and the acquisition of data. In order to place this work into perspective with past work of a similar nature reported in the literature, Chap- ter II of this dissertation presents a brief history of the phenomenon of immobilized enzymes. Emphasis is placed on clinically oriented analytical applications and the dif— ferent approaches taken to analytical systems which incor- porate an immobilized enzyme. Chapter III of this dissertation provides a descrip- tion of the system in its original or first generation form along with modifications which led to its present or second generation form. Included in this chapter are a description of the instrumental components which comprise the system and a generalized description of the operation of the system. Some of the key features of this system are also presented. The immobilized—enzyme reaction column performs the principal function in the system and is described sepa— rately in Chapter IV. The choice of controlled—pore glass as a support material is discussed, and the detailed . 2 41a} experimental procedure for covalent attachment of the en— zyme to this support is presented in this chapter. Some preliminary work done on the original system using an immobilized glucose oxidase column for the deter- mination of glucose is presented in Chapter V. Experimen— tal details and a discussion of results which led to the investigation of dispersion phenomena in the system are included in this chapter. The contents of Chapter VI concern the automation of the system by means of a microcomputer. The microcomputer chosen is described along with the hardware and software used to affect control of the instrumentation. Also in- cluded is a description of an analog-to-digital converter circuit board which was designed to accomplish automated data acquisition. The chapter concludes with a presenta— tion of work done on the establishment of a serial communi- cation link between the controller microcomputer and a more powerful minicomputer. Reasons for this link and the hardware and software required to accomplish it are dis— cussed in this chapter. Automation of the system provided an easy mechanism for characterization of the system. Two studies were performed to evaluate precision, sample dispersion, and the linearity of response of the system. These studies are described in Chapter VII. Immobilized alcohol dehy- drogenase was used in one of these studies, and the measured value of the Km of this enzyme preparation is reported. Chapter VIII presents some final results obtained for the determination of glucose with immobilized glucose oxidase and the microcomputer-controlled system. Included in these results is the value obtained from a glucose determination on a control serum sample. The chapter con- cludes with a discussion of the potential of the system as a clinical analyzer. The final chapter, Chapter IX, contains some ideas for possible future work on the system. Particular empha- sis is placed on improvements in sample handling and micro- computer software. CHAPTER II HISTORICAL This chapter provides an historical perspective for the work to be presented in the following chapters. Pre- sented first is a description of some of the earliest work in the area of immobilized enzymes. Next, work done on the application of immobilized enzymes to clinical analy— ses is described. Finally, the clinical applications are classified in terms of the system into which the immobi— lized enzyme is incorporated. Early Work With Immobilized Enzymes The first recognized immobilized enzyme was reported by Nelson and Griffin in 1916 (13). They showed that the enzyme invertase was removed from solution by adsorption onto charcoal or colloidal aluminum hydroxide and that while adsorbed the enzyme retained its activity. The potential of this discovery was apparently not recognized, for the area received no further investigation until 1953 when Grubhofer and Schlieth reported the successful 7 covalent attachment of several enzymes to diazotized poly- styrene (14). Upon testing the resultant product for ac- tivity they were able to state, "Die Fermente activitat blieb erhalten", and termed the phenomenon, "struktur gebunden Fermentan". Interest in this area began to in- crease thereafter and by the early 1960's numerous workers were investigating different methods for immobilizing en- zymes and characterizing the resultant products (11). The first analytical application of an immobilized enzyme can be attributed to Bauman and coworkers, who in 1965 reported on the immobilization of cholinesterase for the determination of butyrylthiocholine iodide (15). They entrapped the enzyme in a starch matrix on urethane foam pads and proved the analytical utility of the preparation. Immobilized Enzymes for Clinical Analyses Hicks and Updike were the first to report the immo- bilization of a clinically applicable enzyme (16). In 1966 they described the analytical usefulness of a glucose oxidase preparation for the determination of blood glucose. The enzyme was copolymerized in an acrylamide gel which was lyophilized, packed into a 1.0 mL syringe barrel, and hydrated prior to use in a continuous flow system. Table 1 Immobilized Enzymes Applied to Clinical Analyses Enzyme Analyte References Alcohol dehydrogenase Alcohol 28 Alcohol oxidase Alcohol 27 Cholesterol esterase Cholesterol esters 22 Cholesterol oxidase Cholesterol 22 Galactose oxidase Galactose 25 Glucose dehydrogenase Glucose 24 Glucose oxidase Glucose 21 :::::::.::1dase 20 Glycerol kinase Triglycerides 26 Hexokinase G1ucose-6-phosphate Glucose 23 dehydrogenase Hexokinase Glucose 18 Urease Urea l7 Uric acid 19 Uricase 10 Since that time, applications of immobilized enzymes to clinical analyses have been numerous. Enzymes which have been utilized in this manner are listed in Table 1 along with the analyte determined by each. For most of the enzymes listed, there have been different approaches taken in the preparation and method of use of the enzyme. With regard to preparation, differences occur in two major areas. These are: 1) the choice of a support ma- terial for immobilization and 2) the choice of the coup- ling procedure for attachment of the enzyme to a given support material. Support materials which have been used for immobilization of enzymes range from organic polymers such as nylon and polyacrylamide to inorganic materials such as iron oxide and controlled-pore glass (11, 12). Coupling procedures vary as widely as the surface chemis- try and derivatization possibilities of a given support will allow, so that for a single support such as con- trolled pore glass nearly a dozen immobilization proce- dures have been reported (12). With regard to the method of use of the enzyme, differences occur also in two major areas. The first of these is the choice of a detection scheme for monitoring the product of the enzymatic reaction. Detection schemes , -.' u..- '_‘~_ ." <— 11 which have been applied to immobilized—enzyme based analy- ses include those based on potentiometry (l9), fluorometry (21), amperometry (22), coulometry (29), spectrophotometry (30), enthalpimetry (31), and chemiluminescence (32). System Approaches in Clinical Applications The second major area in which differences in the method of use of the enzyme occur is in the choice of the mode of application or the "system" into which the immobi— lized enzyme is incorporated. The "system" refers mainly to the combined mechanism for transporting samples to the enzyme and for transporting the enzymatic product to a detector. The goal of this work was not to investigate new pos- sibilities for support materials, coupling procedures, or detection schemes. Enough viable choices already exist in these areas. Rather, it was felt that investigation of a new possibility with respect to an analytical system which incorporates an immobilized enzyme was warranted and would be of value. The various approaches to analytical systems which incorporate an immobilized enzyme can be classified into three major categories. These are: 1) external probes, 12 2) continuous flow systems, and 3) stopped-flow systems. Each of these in turn will be described below. External Probes As the name suggests, this type of system is based on the insertion of the immobilized enzyme into an analyte solution. The enzymatic reaction is allowed to take place there and the resultant product is measured. Solution transport for both sample and product takes place mainly by diffusion or by mechanical stirring within the solution container. A major contribution to this system category has been made by developments in the area of the enzyme electrode (33). In this approach, the enzyme is immobilized at or near the surface of an electrochemical sensor which selec- tively responds to the activity of a chemical species pro- duced or depleted by the enzymatic reaction. The electrode is inserted into an analyte solution and substrate diffuses through a membrane or gel into the presence of the enzyme. There reaction takes place, and the resultant productdif— fuses to the actual electrode surface. When a steady- state diffusion rate is reached, which is often a very slow process, the measurement of product is completed. 13 Enzyme electrodes based on glucose oxidase for the determination of glucose (34), urease for the determination of urea (35), and uricase for the determination of uric acid (19), among others, have been developed for clinical analyses. Another probe type system has been investigated by Guilbault and coworkers (22). In this approach, an enzyme is immobilized on a particulate support such as cellulose or controlled—pore glass. This product is then sealed into a porous magnetic capsule. This capsule is inserted into an analyte solution and rotated by a magnetic stirring motor. The stirring action aids in the solution transport of sample to the enzyme and in the release of product to the bulk solution. After conversion has taken place, the enzymatic product is measured in the reaction vessel. Im- mobilized-enzyme capsules containing glucose dehydrogenase for the determination of glucose (24), urease for the de- termination of urea (36), glucose oxidase and hydrogen peroxide for the determination of glucose (21), and cho- lesterol oxidase and cholesterol ester hydrolase for the determination of total cholesterol (22) have been prepared and used with success. 14 Continuous Flow Systems Much effort has been put forth in the investigation of systems based on the continuous flow method of analysis. Emphasis has been placed on the adaptation of immobilized enzymes to the Technicon Autoanalyzer system that is prev— alent in clinical laboratories (37), although other con- tinuous flow systems independent of this have also been used (38, 39). All of these systems have in common the use of an immobilized enzyme "reactor". Two major reactor designs have been applied. The first of these is the open tube reactor in which the en— zyme is immobilized on the inside surface of a support tubing (40-43). The second design is the packed column type of reactor in which the enzyme is immobilized on a suitable particulate support and then packed into a small column (16—18, 38). In both cases, the reactor is fitted into the flow system and a pump is used to transport sam- ple solutions to and through the enzyme reactor. As pro- duct is formed, it is carried out of the reactor and on to a detector, if it is directly measurable, or it is mixed with indicator reagents through the use of various mixing or delay coils before reaching the detector. 15 Solution is continuously pumped through the system during operation, hence the term "continuous flow". Stopped-flow Systems A novel approach to an analytical system based on the use of an immobilized enzyme has been investigated in our laboratories. The novelty stems from incorporation of an enzyme reactor into a stopped-flow mixing unit. As with continuous flow systems, two types of reactors have been applied. The application of the open tube type is based on the concept of insertion of a "sample loop" into one of the solution lines of the stopped-flow mixer (44). Instead of an inert and untreated 100p, however, an immo- bilized-enzyme reaction loop is used (45, 46). Sample is introduced rapidly into the reaction loop by suction. There conversion to product takes place during an incuba— tion interval. The product formed is then rapidly mixed with indicator reagents, and the indicator product is carried to the detector by the action of the stopped-flow syringes. The incorporation of the other type of reactor, namely the packed column, with the stopped-flow mixing technique is the focal point of the work to be presented in this 16 dissertation. In this hybrid approach, sample is trans- ported through the reactor in a continuous flow manner, but product is collected within a modified stopped-flow module to allow fast and efficient mixing with indicator reagents and rapid transport of the indicator product to the detector. Research was directed toward characteriza- tion of this type of system for use as a clinical analyzer. , L - A fifi—kg _W -.--.—_ w—H- _- ._..— A-_-_.._.. w . _k_‘-_~__,—_.- _a- CHAPTER III AN OVERVIEW OF THE SYSTEM This chapter presents an overview of the analytical system which is the focal point of this dissertation. As mentioned earlier, the system is based on the use of an immobilized-enzyme reaction column coupled with the stopped-flow mixing technique. The overview begins with a detailed description of the original system configura- tion. Both the instrumental components that comprise the system and a general operational scheme that explains how the system works are presented. This is followed by a description of modifications which led to the present sys- tem configuration. The chapter concludes with remarks on the key features of this system. The Original System Configuration Instrumental Components A schematic diagram of the original system configu- ration is presented in Figure 1. The instrumental com- ponents that comprised the system are each described below. 17 "HMCATOR REAGENTS (ED- ...... . 18 BUFFER Pamsnrnc PUMP ENZYME REACTION PUSH catuuu LJOUN) C OL L ECTION LOOP WASTE SPECTROPHOTOMETER Figure 1. Schematic representation of the original system. —_ _.._..__._ ' .b_—.—'~—=_.—-va-.~_;-,~ .. ~ .— . - g.— s.“ - -- ’ ____3 ‘ .i<._::‘_‘._-_-.-‘_:_-_ _.«—_ .‘g—_~=— .—.—_-',- '~‘~—7?‘- .. - TR W— a nmmfif-fl l9 Stopped-flow Mixing Module A prototype of the GCA McPherson Model 730—S stopped- flow mixing module was obtained from the manufacturer. It is designed to be compatible with the modular spectropho- tometric system that is produced by the same manufacturer (47). The stopped—flow module consists of two drive sy- ringes, A and B, whose volumes are variable from 200 to 500 uL. These syringes are driven by a pneumatic cylinder, D, that is controlled by solenoid Sl. Valves 1 and 2 are fluid check valves that direct solution flow to refill or empty the drive syringes. The mixer, E, and the observa- tion cell, F, are integrated into a single block of Kel-F. The observation cell is a 2 mm by 2 mm cylinder that is fitted with quartz windows and has a volume of 63 uL. Syringe C is the variable volume stop syringe that con- trols the volume of solution driven for each push. Valve 5 is integrated with solenoid S4 and directs the solution flow to the stop syringe or out to the waste. Collection Loop The stopped-flow mixing module was modified by in- sertion of a "collection loop" into one of the solution lines (44). The appropriate connection points were 20 accessed through holes bored in the front of the cabinet. The loop itself was a coiled length of Teflon tubing, 0.8 mm by 40 cm, with a resultant volume of 200 uL. This tubing was fitted with plastic HPLC fittings made by the Altex Co. to allow connection to valves 3 and 4, which are 0.060 inch bore three-way slider valves stacked to- gether in the actuator-return mechanism, Gl. Both valves were placed under control of sOlenoid SZ. Peristaltic Pump The pump used in this system was an early model Tech- nicon Autoanalyzer peristaltic pump. The number of steel rollers on the pump that force solution through flexible tubing were doubled in an effort to improve the regularity of flow rate. A color-coded pump tube, chosen to affect the desired flow rate, was fitted at both ends with plas- tic HPLC fittings for easy connection within the system. Sample Introduction Mechanism The introduction of sample solution into the system was accomplished by means of valve 6, a three-way slider valve with its common port connected to the pump tubing. The position of the valve determined whether sample 21 solution or buffer was drawn into the system by the pump. Valve 6 was fitted into the actuator-return mechanism, G that was controlled by solenoid 83. 2! Spectrophotometer A GCA McPherson EU 700 series modular spectropho- tometer was employed in the system. It consisted of a UV—visible light source module, a monochromator, and a photomultiplier module. A Keithley model 427 current amplifier was used to convert the photomultiplier output current to a measurable voltage. The Immobilized-Enzyme Reaction Column The immobilized-enzyme reaction column performs the principal function in the system, and a separate chapter is set aside for its description. Operation of the System A generalized description of the physical operation of the system pictured in Figure 1 follows. All control functions were performed manually. Under control of solenoids S1 and S4, the stopped— flow module is cycled several times to fill syringes A and B with indicator reagents and with push liquid, 22 respectively, and to fill all lines with solution. Then, under control of solenoid S valves 3 and 4 are simulta— 2, neously positioned to allow solution to travel from the enzyme reaction column into the collection loop. At this same time the peristaltic pump is started. Valve 6, under control of solenoid S3, allows either the sample (substrate) or the buffer to be transported through the enzyme column by the action of the pump. When the sample is selected, a timed interval is used to control the V01- ume introduced. After this interval, the buffer is se— lected for a timed interval to push the sample plug into the enzyme reaction column. On passing through the col— umn, sample is converted to product, and this product is transported to the collection loop, whereupon the pump is stopped. Valves 3 and 4 are repositioned as shown. Activation of the pneumatic cylinder in the stopped— flow module causes the two drive syringes to push their solution contents through the lines and out to the stop syringe until that syringe reaches the end of its travel. This results in the enzymatic product being transported, by the push liquid, from the collection loop to the mixer, where it is rapidly and efficiently combined with 23 indicator reagents from the other syringe. In the same motion, the resultant absorbing species is carried to the observation cell whereupon the flow is stopped, and a spectrophotometric measurement is made. To begin a new cycle, the contents of the stop syringe are expelled to the waste, under control of solenoid S4 and the drive sy- ringes are refilled. After valves 3 and 4 are simultane- ously switched and the pump is turned on, a sufficient volume of buffer is pumped through the enzyme column to avoid carry-over from previous samples prior to the next sample. The Present System Configuration Instrumental Components A schematic diagram of the present system configura- tion is presented in Figure 2. All instrumental compo- nents in this second generation system are the same as those described previously, except for the sample intro- duction mechanism. The modification of this component is described below. In addition to this change, an 8080 microprocessor based microcomputer (not shown) was interfaced to the system to automate control of instru=- mentation and the acquisition of data. A separate ' _, ,' .“ya-fizw‘w— 24 BUFFER ll PERISTALTIC pun? 6 SAMPLE @ ........... i O I .1162 INTRODUCTION _ @ LOOP [T fl 7 VACUUM B ENZYME REACTION COLUMN INDICATOR } I Z PUSH REACENTS LIQUID ET 3 62). ....... I - a. COLLECTION LOOP 4 wAsTE E L SPECTROPHOTOMETER Figure 2. Schematic representation of the present system. 25 chapter describes the manner in which this automation . I was implemented. I I Sample Introduction Mechanism In the second generation system, valve 6 was removed I from its place behind the peristaltic pump. It was placed, along with an identical three-way slider valve, which is labeled 7 in Figure 2, into a single acctuator-return mechanism, G that is controlled by solenoid S3. This 2! dual valve assembly was inserted between the pump and the enzyme reaction column as shown. The common port of each valve was connected to the other by means of a length of 0.8 mm I.D. Teflon tubing to form a sample "introduction loop". Lengths of 40 cm and 80 cm were used to provide volumes of 200 and 400 uL, respectively, for introduction of sample. Operation of the System The operation of the system represented in Figure 2 is identical to that of the original system except for the manner in which a sample is introduced. In this system, after the stOpped—flow module has been cycled several times, and all lines have been filled with solution, 26 valves 6 and 7 are switched simultaneously to allow sample to be drawn into the introduction loop by the action of an aspirator connected to valve 7. When the loop has been purged and filled, both valves are switched to allow solu- tion to flow from the pump to the enzyme reaction column as shown in Figure 2. The peristaltic pump is then turned on and the contents of the sample introduction loop are pushed into the enzyme column by buffer solution from a reservoir. From this point on, the sample undergoes con— version as before, the product is collected and mixed with indicator reagents, and a spectrophotometric measure- ment is made on the resultant absorbing species. A new cycle begins as described previously. The reasons for modification of the sample introduc— tion mechanism are as follows. Unlike the original mech- anism, the present one allows for faster introduction of sample and is based on a discrete volume being filled with solution rather than on a solution being proportioned at a constant flow rate over a timed interval. It is, there- fore, inherently more reproducible, since pump tubes deteriorate with time. Also, the position of the intro- duction loop at the head of the enzyme column reduces if, A ‘A ;___.__.‘L A .. fi_—— $___4._.—— 27 pre-column dispersion of a sample by elimination of the dead volume contribution of the pump tube. This inher- ently increases the response of the system to a given size sample or, from another standpoint, it decreases the size of a sample required for a given response. Key Features of the System The general configuration of the system and the use of the stopped-flow mixing module provide the system with several key features that make it quite attractive. The first of these features is that the enzymatic reaction is completely isolated from the indicator reaction. This avoids possible inhibition or denaturation of the enzyme by indicator reagents and allows the conditions for the two reactions to be Optimized separately. Secondly, the stopped-flow module provides for fast and efficient mixing of the enzymatic product with indi- cator reagents and for the rapid transport of the result- ant absorbing species to the observation cell for a meas- urement. Thus, delays, such as those associated with mixing coils and slow solution transport in continuous flow systems, are avoided. 28 A third key feature of the system is that the speed of mixing afforded by the stopped—flow module makes it ideally suited for use in a rate method of analysis of those indicator reactions which are slow to reach equi- librium. This could result in an added savings in time. For those indicator reactions which reach equilibrium quickly, the system is also well suited for equilibrium measurements. The inclusion of the stopped-flow module thus lends flexibility to this system, and allows its use to be tailored to each individual analysis scheme. CHAPTER IV PREPARATION OF THE IMMOBILIZED-ENZYME REACTION COLUMN This chapter provides a generalized description of the immobiliZed-enzyme reaction column used in the sys— tem. A discussion of the choice of controlled-pore glass as the support material is presented first. This is followed by a description of the method used for co- valent attachment of the enzyme to controlled-pore glass. Experimental details for the precedure are in- cluded in this section. The Choice of Controlled-pore Glass as Support Material Early investigation into the concept of an immobi— lized-enzyme based stopped—flow clinical analyzer made use of nylon tubing as a support for the attachment of enzyme (15). Results were very encouraging. It was felt, however, that for contact times of substrate with 29 3O enzyme on the order of two minutes, a greater signal-to— noise ratio ought to be obtainable. That is, a greater percent conversion of substrate to product ought to be realized in that time or even less. An obvious approach to increase the percent conver- sion of substrate for the same contact time is to provide more enzyme activity. This in turn can be accomplished by an increase in the amount of surface area upon which immobilization occurs and to which substrate solution is exposed. One support material well recognized for its large surface area per unit weight is controlled-pore glass which has typical surface area values from 50 to 100 ng'l. A valid comparison of surface area for an open tube support and a particulate support must be based on a com- mon denominator. If one considers a small packed column containing controlled-pore glass, than a comparison with an open tube can be made based on the ratio of surface area to solution volume exposed to that surface area. For a 1 mm diameter tube, a value typically used (28, 43), the surface area per unit volume contained within can be calculated from the equation .: < gumm— where r is the radius and 1 is the length of the tube, and the value obtained is 0.04 cm7 . For a packed column with controlled-pore glass, it was necessary to measure the solution volume experimen- tally and ratio it to the manufacturer's stated surface area value. To measure the solution volume of a packed column, a 1.4 mm I.D. by 20 cm length of glass tubing was filled with 74-125 um size glass particles with an average pore diameter of 327 A (Sigma). The weight of this glass was determined by difference. The column was then filled with water, capped, and weighed. The volume of the water was The determined from its mass and density to be 225 uL. surface area per unit volume was then determined using a surface area value of 73.5 ng-1 (48). The value ob— tained was 315 cm-1. Conservatively then, there is a factor of at least 5000 increase in surface area for a packed column, as described, over a 1 mm I.D. open tube for the same solu- tion volume exposed to that surface area. Clearly, the 32 use of controlled—pore glass for immobilization of an enzyme, and its subsequent application in the form of a small packed column ought to provide greater enzyme ac- tivity and, therefore, should result in greater percent conversion of substrate for contact times comparable to, or even shorter than, those required in an open tube en— zyme reactor. Controlled-pore glass is attractive as a support for enzyme immobilization for three other reasons (12). First, it is a rigid support material that is not subject to com- pression or swelling in a flowing solution and is, there- fore, well suited for packed column applications. Second, it is not subject to microbial attack such as it might be exposed to in analysis of biological samples. Third, controlled—pore glass lends itself to simple and effective procedures for covalently bonding enzymes to its surface. Covalent Attachment of Enzyme to Controlled-pore Glass Description of the Procedure Several methods are available for the attachment of an enzyme to a solid support (8 - 12). These include covalent bonding, adsorption, and cross-linking or .—._-4._~ .~..._ ._ ‘g- 4 . A4.-. «_.—.__-_—_ 33 copolymerization. Of these, covalent attachment was chosen for this work because it is believed to yield the most stable product. There are a number of procedures for the covalent attachment of an enzyme to controlled-pore glass. Reagents used in these procedures include the following (12): N,N'-dicyclohexylcarbodiimide, thiophos— gene, phosgene, cyanuric chloride, cyanogen bromide, p- nitro benzoyl chloride, and glutaraldehyde. Most of these chemicals are relatively hazardous or noxious to work with, except for the last two. The glutaraldehyde procedure was used with success in our laboratory on nylon tubing and was found to be a relatively simple procedure. The adap- tation of the procedure to controlled-pore glass has been reported (17) and is equally simple. Thus, for reasons of simplicity, ease of reagent handling, and previous success on another material, the glutaraldehyde method of covalent attachment was selected for this work. A schematic representation of the attachment proce- dure is shown in Figure 3. In this procedure, the glass is first functionalized with a silylating agent that con- tains a primary amine. The glutaraldehyde is then al- lowed to react with the amine glass to form imine 34 H CH S I ?C 2 3 a ——— '———o + -—- '——— glass:3 f1 H CH3CH20 Sl (CH2)3NH2 + o H H o C 2c 3 y—aminopropyltriethoxysilane a o O O o 8 . -... _I-.... . glass: 31 O 1 (CH2)3NH2 HC(CH2)3CH + O 0 glutaraldehyde a o o o e . I. II g1ass§-——Si-O-—Si——(CH ) N=CH(CH ) CH + NH -(Enzyme) a I 2 3 2 3 2 o o 8 o o 8 g1ass§-—Si—-O-—Si-—%CH2)3N=CH(CH2)3CH=N-——(Enzyme) O 0 Figure 3. The immobilization procedure for controlled- pore glass. 4 41.1 —-—‘_-_AA . .—_=._.——_fi._ 35 linkages and provide an active aldehyde group on the glass surface. The enzyme is attached to the aldehyde glass presumably via the imine linkages with lysine residues on the enzyme. The procedure is essentially a three step series, where the first step can be performed on larger amounts of glass and stored for later use. Experimental The amine glass was prepared using the method re— ported by Weetal (49). A 5 g quantity of 74-125 um size, 327 3 pore diameter, controlled-pore glass (Sigma) was placed into a 200 mL round bottom flask. To this were added 100 mL of a toluene solution containing 10 mL of y-aminopropyltriethoxysilane (Aldrich). The flask was fitted with a reflux condenser and a heating mantle, and the solution was refluxed overnight. Periodic bumping of the solution caused some glass to be splashed up into the lower part of the condenser column and this could not be avoided even with the use of boiling chips. After it was cooled, the amine functionalized glass was washed with acetone and air dried. The method of Bowers et a1. (17) was used with slight modifications to attach the enzyme to the amine glass with 36 glutaraldehyde. A 150 mg quantity of the amine glass was placed into a 50 mL round bottom flask. To this were added 10 mL of a 0.1 M phosphate buffer solution, pH 7.2, containing 2 mL of a 2% aqueous solution of glutaraldehyde. The glass became yellow upon contact with the glutaralde— hyde. The flask was attached to a rotary evaporator, and a vacuum was applied with rotation for one hour in an at— tempt to remove air trapped in the pores of the glass. The flask was rotated an additional 2 hrs. at atmospheric pressure. The aldehyde glass, now red brown in color, was then washed in a 30 mL fritted glass filter of medium porosity with 200 mL 1 M NaCl, 200 mL distilled water, and 60 mL 0.1 M phosphate buffer, pH 7.2. The aldehyde glass was then washed into a 10 mL round bottom flask, and excess water was drawn off by suction. To this were added 2 mL of a 0.1 M phosphate buffer solu— tion, pH 7.2, that contained 15—25 mg of enzyme. The flask was immersed in an ice bath and a vacuum was applied for 15 min., after which the flask was capped and refrig— erated for approximately 18 hours. The enzyme—glass product was washed as described pre- viously. The washed product was packed into the jacketed glass column pictured in Figure 4. The column was clamped 37 I EPOXIED HPLC FITTING ——|.7 cm I.D. I8 cm L45 mm I.D. 1 FILTER PAPER DISC Figure 4. The jacketed glass column. 38 vertically in place and fitted with a glass funnel to which an appropriate HPLC fitting has been attached with epoxy cement. The column was filled with buffer by a vacuum applied at the downstream end, and the enzyme-glass was then washed into the funnel and allowed to fill the column by gravity. The column was occassionally tapped on the outside with a stirring rod to help assure uniform packing. The glass was held in place in the column by means of a small disc of filter paper (Whatman #1) sand- wiched between two HPLC connector fittings at the down- stream end of the column. CHAPTER V PRELIMINARY WORK WITH IMMOBILIZED GLUCOSE OXIDASE Before automation of the immobilized-enzyme based system was implemented, some preliminary work was done on the original system to test its potential as a practical clinical analyzer. This was accomplished with the use of immobilized glucose oxidase for the determination of glucose. This chapter begins with a general description of this preliminary work which leads into a discussion of kinetic aspects of the determination of substrates with enzymes. Then, the experimental details of the glucose oxidase work and the results obtained are presented. The chapter concludes with a discussion of these results with emphasis on those which led to an investigation of sample dispersion phenomena in the system. 39 40 General Description of the Work Glucose oxidase was immobilized on controlled-pore glass and packed into a small glass column. The column was used at room temperature, but not thermostated. Buf- fered glucose solutions were passed through the glucose oxidase column by means of a peristaltic pump to allow conversion of the glucose to products by the following reaction, lucose oxidase Glucose + O + H20 (g::::::::::::::9 2 Gluconic acid + H202. The hydrogen peroxide produced was transported to the col— lection loop by the action of the pump. After the pump was stopped, the action of the drive syringes in the stopped—flow module carried the hydrogen peroxide to the mixer where it was mixed with indicator reagents to form triiodide by the following reaction, — + MO(VI) ' O + I . H202 + 31 + 2H é:::::::9 2H2 3 After transportation to the observation cell, the absorb- ance of triiodide was determined spectrophotometrically at 365 nm where the difference spectrum between triiodide and 41 the molybdenum catalyst is near a maximum. Transmittance was measured relative to the indicator reagent as a blank by monitoring the output voltage from a current-to-voltage converter with a digital voltmeter. Absorbance values were calculated from the measured transmittance and used to calculate the concentration of glucose in the sample. A Consideration of Enzyme Kinetics It is important to consider the kinetics of the enzy- matic determination of substrates, since, in this system, the product concentration measured does not represent an equilibrium value after total conversion of substrate. Rather, it represents only partial conversion of substrate that takes place during the time the substrate is in con- tact with the enzyme. As such, the product concentration reflects a measurement of the rate of the enzyme reaction over a fixed time interval. A direct proportionality be- tween the concentration of product and the initial concen- tration of substrate will hold only if conditions are con- strained to attain the desired reaction kinetics. If a Michaelis and Menten model for an enzyme reaction is assumed, the mechanism can be written as, 42 k1 k2 E+S,“—"__ ES 5‘— E+P kl k2 where E represents the enzyme, S the substrate, ES an in- termediate complex, and P the product. Rate constants for elementary steps are given by k1' k k and k . The 2 initial rate of the reaction, according to a steady state -1’ 2’ treatment, is given by, where E is the initial enzyme concentration and Km = o Ingle and Crouch (50) have shown, by rigorous manipu— lation of integrated rate equations, that for such a Nfichaelis and Menten formulation, the rate equation reduces to the form of a pseudo—first—order rate equation to within % when initial substrate concentrations are 5 0.01 Km. Furthermore, they point out that, given a choice of the fixed-time and variable—time approaches to measurement of the rate of a pseudo—first—order reaction, the fixed—time _. viale? 43 approach is by far the better choice. The reason for this is that, in a pseudo-first-order reaction, the relative change in concentration over a fixed time interval is con- stant, and, therefore, the absolute change is directly proportional to the initial concentration of reactant. This direct proportionality is true regardless of whether the rate is constant over the measurement interval. Therefore, rate measurements are not restricted to the initial part of the reaction where pseudo-zero—order reaction kinetics hold. In terms of the enzymatic reac- tion in this system, a substantial percent conversion can take place during the fixed-time interval, and, as long as the initial substrate concentration is S 0.01 Km, a linear relationship will hold between the measured con- centration change and the initial substrate concentration. Workers have reported apparent changes in the Mi- chaelis constant of enzymes upon immobilization. In gen- eral, the observed change is an increase in Km' which is believed to arise from charge effects, diffusion effects, or changes in the conformation of the enzyme (8). Thus, as a guideline for the limitation of initial substrate concentrations, the Km of the soluble form of the enzyme 44 can most often be used, and, if the Km of the immobilized enzyme is increased, linearity may extend beyond the ex- pected concentration limit. A value which has been reported for the Michaelis constant of soluble glucose oxidase with respect to glu- cose is 9.6 x 10-3 M at 20 OC, pH 5.6, and a 20% oxygen atmosphere (51). This value was used as a guideline in the preparation of glucose standards for this work. The con- centration of the highest standard was 5.55 x 10“4 M (100 ppm) which is 20.06 Km, and that of the lowest was 2.77 x 10‘ M which is 20.003 Km. According to kinetic con- siderations, a direct proportionality between the rate of the glucose oxidase reaction and the initial glucose con- centration should exist for these solutions to within 6% with a greater deviation expected at the higher concentra— tions. Furthermore, if the apparent Michaelis constant is increased, then direct proportionality could hold beyond the highest glucose concentration. Experimental and Preliminary Results In this section the experimental conditions that were used in these preliminary studies, and the results that were obtained are described. 45 Solutions and the Immobilized—Enzyme Column A 0.6 M phosphate buffer solution of pH 6.0 was prepared by dissolving 71.61 g of potassium dihydrogen phosphate and 12.85 g of dipotassium hydrogen phosphate in deionized water and diluting to 1 liter. A 1000 ppm stock glucose solution was prepared by dissolving 1.000 g of anhydrous D—glucose in deionized water and diluting to 1 liter. This solution was left unused for 24 hours to allow mutarotation to equilibrate. A 10 mL buret was used to dispense this stock solution to prepare 50 mL each of the glucose standard solutions listed in Tables 2-4. These solutions were diluted to the mark with the 0.6 M phosphate buffer solution. Indicator reagent solution was prepared fresh daily by dissolving 0.65 g of ammonium molybdate and 10 g of potassium iodide in a 0.1 M phosphate buffer solution of pH 6.0, and diluting to 100 mL with the buffer. The rather high iodide concentration (0.6 M) was used to assure that greater than 99% of all iodine in solution existed as triiodide at equilibrium. The 0.6 M phosphate buffer, used as push liquid and diluent in the preparation of glucose standards, promoted better mixing in the 46 stopped-flow module. It minimized concentration gradients between solutions and thereby reduced fluctuations in transmittance, otherwise observed, when solutions entered the observation cell incompletely mixed. Glucose oxidase (D-glucose:oxidoreductase; E.C.1.l.3.4; Sigma type II) was immobilized on 74-125 um, 327 A pore diameter, controlled-pore glass by the procedure described in Chapter III. Approximately 100 mg of the product were packed into a 1.4 mm x 18 cm glass tube onto which plastic HPLC fittings had been epoxied. This column was placed in the position shown in Figure 2. An effective flow rate of 1.6 mL min.1 was obtained with a green color-coded pump tube. Optimization of Push Volume If one considers the process of transportation of the enzymatic product into the stopped-flow module, in the ideal situation, the solution in the collection loop would reach the mixer undiluted by the push liquid. The solution that entered the observation cell would then accurately represent the original product solution. Ideally, the next push of the syringes would completely clear the ob- servation cell of product so that it would contain only a 4.. 47 1:1 mixture of the indicator reagent and the push liquid. This would allow the 100% transmittance to be measured periodically to compensate for source fluctuations in the single beam spectrophotometer. The push volume of the mod- ified stOpped-flow module was optimized to correspond, within experimental uncertainty, to the ideal situation described. The procedure used to accomplish this is de- scribed below. The variable-volume stop syringe in the stopped-flow module allows control of the volume of solution driven with each push. This volume can be set by means of a thumbscrew that limits the retraction of the syringe plunger. A calibrated scale alongside this thumbscrew in- dicates the volume selected. Likewise, the volumes of the drive syringes can be set to correspond to the volume of the stop syringe by means of a thumbscrew. Connections to the collection loop were temporarily modified to allow the loop to be filled with solution by suction. A 10 g L-1 potassium nitrate solution was pre- pared by dissolving the solid in distilled water. This solution was used to represent an enzymatic product solu— tion. Distilled water was used as both the push liquid 48 and thewindicator reagent solution- Transmittance meas- urements were made at 302 nm versus distilled water. The volumes of the stop syringe and the two drive syringes were set to 200 uL and the collection loop was filled with the potassium nitrate solution. The stopped- flow module was then cycled and the transmittance of the potassium nitrate solution in the observation cell was measured. Without introduction of a new solution, the stopped-flow module was again cycled and a transmittance measurement was made. This procedure was performed in triplicate and average absorbance values were calculated. from the measured transmittances. The volume of all three syringes was then incremented by 25 uL and the above procedure was repeated. This process was continued through a push volume of 400 uL. The results obtained indicated that 300 uL was the optimum value for the push volume in the modified stopped- flow module. At this volume, the absorbance value ob- tained in the potassium nitrate solution was a maximum, and the "residual" absorbance obtained after the second push was negligible. At smaller volumes, the residual absorbance was significant which indicates that some IL .. ___.___ _ ‘_ 49 product solution was carried into the observation cell with the second push. At larger volumes, the absorbance of the potassium nitrate solution was significantly reduced which indicates that dilution of the product solu- tion by the push liquid had occurred. Therefore, a push volume of 300 uL was used for all work with the modified stopped—flow module. Calibration Curves and Analytical Results On the first day of preparation of the immobilized glucose oxidase column, the results presented in Table 2 were obtained with the original system under manual con- trol. Each absorbance value listed is the average of five replicate determinations. The procedure for obtaining these results is described below. The system was initialized by pumping a 0.6 M phos- phate buffer of pH 6.0 through the enzyme column to remove trapped air pockets and to fill all connecting tubing with solution. The stopped-flow module was then cycled several times to fill the drive syringes with the iodide—molybdate indicator reagent and with the 0.6 M phosphate buffer used as push liquid. The 100% transmittance was set with a 1:1 50 Table 2 Calibration Curve Results with Immobilized Glucose Oxidase (Day 1) Glucose -1 Average Concentration (mg L ) Absorbance %RSD 20 0.327 1.2 40 0.541 2.6 60 0.719 1.3 80 0.842 3.2 0.889 2.7 100 51 mixture of the indicator reagent and the push liquid in the observation cell. For a single determination, 0.8 mL (30 s) of a buf- fered glucose sample were pumped through the immobilized enzyme reaction column at a flow rate of 1.6 mL min- , followed by 0.8 mL (30 s) of a 0.6 M phosphate buffer of pH 6.0. As the glucose passed through the enzyme column, the hydrogen peroxide produced was eluted and transported to the collection loop. After the total 60 s pump inter- val, the pump was stopped, and the collection loop valves were switched. By cycling the stopped-flow module, the hydrogen peroxide in the collection loop was transported to the mixer where it combined with indicator reagents to form triiodide. The transmittance of the triiodide solu- tion was measured as described previously. A second cycle of the stopped-flow module cleared the observation cell of triiodide, and allowed the 100% transmittance setting to be checked. After the first glucose sample, all following samples were preceeded by 0.8 mL of phosphate buffer. Thus, if one includes the volume of buffer pumped after each sam- ple, a total of 1.6 mL (60 s) of buffer separated 52 replicate samples. An additional 0.8 mL of buffer were pumped in between different glucose standards in an effort to reduce carry-over. Thus a total of 2.4 mL (90 s) of buffer separated different glucose standards. On day four after preparation of the immobilized glu- cose oxidase column, glucose standards were again deter- mined by the method just described. The results of these determinations are presented in Table 3. The absorbance values listed are the average of three replicate determi- nations. The results listed in Tables 2 and 3 are presented graphically in Figure 5 which is a plot of absorbance versus concentration of glucose. Curve A represents the results from day one and curve B represents those from day four. After the determination of the highest glucose stan- dard on day four, an attempt was made to assess the depen- dence of the absorbance value obtained as a function of the volume of buffer pumped to elute the hydrogen peroxide from the enzyme column. To accomplish this, the volume of buffer pumped after the introduction of the 100 ppm stan- dard was increased from 0.8 mL (30 s) to 1.2 mL (45 s). 53 Table 3 Calibration Curve Results with Immobilized Glucose Oxidase (Day 4) Glucose _ Average Concentration (mg L ) Absorbance %RSD 10 0.102 3.9 20 0.167 0.6 40 0.299 7.4 60 0.410 5.4 80 0.536 7.6 100 0.593 4.2 1.999: ABSORBANCE 54 GLUCOSE CALIBRATION PLOT l .J- I+I f W71 Ill l_l iITfiIUT'I1III‘TTTIWII Illl4lJLJ14J1JJllllJll l IIT—rTTIIITTTTIl’I Curve A - Results from Day 1 Curve B — Results from Day 4 8°amyM%%H+HfifiHHfimfifi%Hmfi%%%HH 0.000 1.000 CONC. GLUCOEE (MG/L) X10 Figure 5. Glucose calibration plot. 55 When this was done, a higher absorbance value was obtained for that standard. All of the glucose standards were then determined with this increased volume of eluting buffer. The results for these determinations are listed in Table 4 and plotted in Figure 6. Each absorbance value repre— sents the average of four replicate trials. The line drawn through the points in Figure 6 is the result of a linear least squares analysis of the data (52). Discussion of Results There are three features of the curves plotted in Fig- ure 5 that merit discussion. The first of these is the lack of coincidence of the two curves. The lower absor- bance values that were obtained for all glucose standards in curve B indicate a loss of activity of the immobilized glucose oxidase during storage under refrigeration. Such a loss of activity is not uncommon (53, 54), and even though not desirable, it is acceptable for analytical ap— plications provided that adequate sensitivity can still be obtained. The second noteworthy feature of the curves in Figure 5 is the non—zero intercept obtained for both series of determinations. This is attributed to a virtually constant l‘ 56 Table 4 Calibration Curve Results with Immobilized Glucose Oxidase (Day 4)* Glucose _ Average Concentration (mg L ) Absorbance %RSD 5 0.0605 16.4 10 0.100 3.0 20 0.170 3.5 40 0.329 7.3 60 0.434 3.7 80 0.551 6.7 100 0.678 4.3 * Volume of eluting buffer increased to 1.2 mL 57 GLUCOSE CALIBRATION PLOT l 1)— I LL ‘I" 2; -+ l LLJ I U -L z I < I m I I T o -- m n: m l < 4 if L. 3: I _.L. 0. eeettt'I-I-I-‘I-H-H-I-I-H-I-I-HIIIIIIIIIIIIIIII-I-HIH-j-HH-I-H‘I-I 0.000 1.000 CONC. GLUCOEE (MG/L) X10 Figure 6. Glucose calibration plot. 58 low-level background of hydrogen peroxide eluting from the enzyme column after the first glucose sample passed through. Since the 100% transmittance was set with the indicator reagent as blank, no adjustment was made for this background, and it appears as a constant added ab- sorbance. Complete elution of the hydrogen peroxide be- tween samples could not be realized in a reasonable time. The third important feature of the two curves pre- sented in Figure 5 is the curvature at high glucose con— centrations. For curve A, which represents data obtained of the first day of preparation of the column, this curvature is attributed mainly to a depletion of dissolved oxygen in the solution that momentarily occupies the en- zyme column. Enzyme activity was significantly higher on this first day so that, at high glucose concentrations, the amount of oxygen consumed is believed to have limited the production of hydrogen peroxide. Since the system is closed, oxygen is replenished only by transport of fresh solution that contains dissolved oxygen. Within a small volume element of glucose solution, all of the available oxygen could be consumed before that element leaves the enzyme column. The concentration of hydrogen peroxide 59 that results from a summation of such volume elements would not correctly represent the velocity of the enzyme reaction over a fixed-time interval. The hydrogen peroxide concentration would, in fact, reach a limit independent of the initial glucose concentration. This limit is ap- parent in curve A where the absorbance due to hydrogen per— oxide rapidly approaches a constant value at high glucose concentrations. The depletion of dissolved oxygen also helps explain the long times required to elute all of the hydrogen per- oxide from the enzyme column. The release of hydrogen peroxide from the enzyme is reported to be a separate mechanistic step that requires oxygen (55). If the oxygen is depleted or greatly reduced, the hydrogen peroxide will not be released until fresh buffer solution replenishes the oxygen, and this would be a gradual process. The curvature at high glucose concentrations in curve B is not as pronounced, and, therefore, is not as satisfac— torily explained by the same oxygen depletion process alone. The results obtained on the fourth day (Table 4), with the increased volume of eluting buffer, indicate that a major contribution to the curvature in curve B was due 60 to an error in “collecting" the hydrogen peroxide eluted from the column. This concept bears further explanation, but this is deferred until Chapter VII. Suffice it to men- tion here that the volume half-width of the eluted hydrogen peroxide "band" was suspected to be large relative to the volume of the collection loop. Therefore, the quantity of hydrogen peroxide collected represented only a discrete portion of the entire eluted product band. The volume of buffer pumped after the introduction of a sample influ— enced the position of the product band center relative to the center of the collection loop. Thus, it determined which portion of the product band was collected. The re- sults plotted in Figure 6 show that the curvature at high glucose concentrations was removed by an increase in the volume of buffer pumped after each sample. It was felt that an investigation of dispersion phe- nomena in the system would provide valuable insight into these results and would serve as a guideline for further work. But, before this investigation was attempted, the system was automated to allow microcomputer control of its operation. It was felt that automation should improve the 61 accuracy and precision of dispersion studies and enhance all further work with the system. Conclusions Several important conclusions were drawn from the preliminary work done with immobilized glucose oxidase. First, the use of controlled-pore glass as the support material for immobilization of the enzyme proved to be advantageous inthis system. For analysis times of less than two minutes, the signal obtained for a given glucose sample was much higher than those obtained with the Open tube reactor (46). The anticipated advantage of increased surface area was thus realized. Second, a linear calibration curve was obtained over a clinically useful range which indicates the potential of the system as a clinical analyzer. The glucose stan- dards used represent approximately a 1:20 dilution of normal serum to allow for deproteinization and buffering. The standards listed in Table 4 thus cover the range of 10 to 200 mg dL“l of glucose. Abnormally high glucose samples could be further diluted to bring them within the workable range. Illlllllll 62 The third conclusion drawn from the preliminary work was that improvements could be made onthe system to in- crease precision and accuracy. Since control of tempera— ture is critical when rate methods of analysis are used, the enzyme reaction column needed to be thermostated. This was implemented by construction of a water-jacketed glass column. Also, the timing of control events and the ac- quisition of data were subject to human error. This error could best be reduced by automation of the system. The implementation of this automation is described in Chapter VI. CHAPTER VI MICROCOMPUTER AUTOMATION OF THE SYSTEM This chapter describes the automation of the immobi- lized enzyme-based system by means of a microcomputer for control of operation and acquisition of data. A discus- sion of the need for automation of the system and the de- cision to use a microcomputer for this purpose is presented first. Next, a description of the Intel Corp. 8080A mi- croprocessor-based microcomputer used in this work is presented. This is followed by a detailed description of the hardware and software used to accomplish automated data acquisition. Included in this latter section is a descrip— tion of an analog-to-digital (ADC) circuit board designed and built for the microcomputer system. Finally, the establishment of a serial communication link between the 8080 microcomputer and a minicomputer is described. 63 64 The Need for Automation The original system described in Chapter 3 was used for preliminary work with immobilized glucose oxidase. In the course of that work, it became obvious that manual control of the system was both tedious and slow. It required the user's complete attention to perform menial switching operations in the correct sequence and a stop watch was required to time various intervals in the opera— tion sequence. Also, data collection required the presence of the user to read the digital voltmeter display. All of this human intervention contributed to an increase in the analysis time for a series of determinations and decreased the precision and accuracy of results. It became apparent that some form of automation of the system was needed whereby control functions, the timing of events, and the acquisition of data would be accom- plished electronically. This would minimize the tedium associated with operation of the system and would increase the speed, precision, and accuracy of the system. The recent advent of rather inexpensive microprocessor-based microcomputers provided an attractive and relatively easy avenue for accomplishing this automation. 65 It was envisioned that a small microcomputer could be "dedicated" to the immobilized-enzyme analyzer to provide control of instrumentation and to acquire data automati- cally. As an "intelligent" controller that could be pro- grammed by the user, the microcomputer would provide a great deal of flexibility for different configurations of the system. This flexibility would be extremely valuable in a developmental system such as the immobilized-enzyme analyzer. Description of the Microcomputer The microcomputer selected to automate the system was an 8080A microprocessor-based system designed and developed at the University of Illinois, called the "ADD-8080" (56). The system is composed of a number of highly functional printed circuit (PC) boards that are connected by a clip-on ribbon cable bus line. These PC boards are designed to plug into the Analog Digital Designer (E&L), which provides a power bus and allows easy patch wire interfacing of digital circuits to the microcomputer. The plug-in nature of the components of this microcomputer system allows it to be tailored to the user's individual application and provides for easy expansion of the system. Table 5 66 Table 5 Printed Circuit Boards Used in the Microcomputer 8080A 2102 AM9130 1702A 2708 8255 8251 ADCHZ12BC Microprocessor CPU Board (Revision lA)* 1K Read-Write Memory Board (Revision 1B) 4K Read-Write Memory Board 1K Programmable Read-Only Memory Board (Revision 2) 4K Programmable Read-Only Memory Board (Revision 2) Parallel I/O Board (Revision 1A) Asynchronous Communications Board (Revision 1B, 1C) Analog-to-Digital Converter Board** *All revisions refer to University of Illinois design **Designed and built at Michigan State University 67 provides a listing of the PC boards used in this appli- cation. The 1702A PROM board contains 1K of programmable read— only memory (PROM) in which a monitor program resides. This monitor is a slightly modified version of the Intel SDK-80 Monitor (ADD—8080 Monitor; University of Illinois). The monitor provides for minimal communication with the microcomputer through an interactive console device. With this monitor resident in PROM, the user issues commands to perform the following: 1) display selected areas of mem- ory, 2) initiate execution of user programs, 3) set "break points" in user programs, 4) modify the contents of memory and the microprocessor registers, and 5) input hexadecimal code from the console device into memory (57). User pro— grams must be entered in hexadecimal code that represents data, addresses, and individual assembly language instruc- tions. Commands are issued in the form of a single alpha- betic character that specifies the command, followed by a hexadecimal parameter that specifies a memory address or a processor register. Resident in 3K of read—only memory on the 2708 PROM board is a modification of the Intel version of Tiny BASIC 68 (Version 2.1; University of Illinois). Tiny BASIC is a higher level interpretive computer language that is a sub- set of the BASIC programming language and is designed for small computer systems with limited memory. This version is written for use with 4K of read-write memory. Tiny BASIC provides the user with the four simple arithmetic operators, the logical AND and OR functions, relational operators, and several useful mathematical functions. All numbers in Tiny BASIC must be integers in the range between -32,768 and +32,767. Variables may be either simple or single-dimensioned arrays. Version 2.1 simplifies I/O routines with the use of the FIN and OUT commands. User communication during operation of a pro- gram is simplified by use of the INPUT and PRINT commands. User programs are entered from the console as a series of numbered lines of text and arithmetic expressions. A limited editing mode in Tiny BASIC allows modification of user programs after they are written (58). Control of Instrumentation Hardware Control of the instrumentation in the system repre- sented in Figure 2 was implemented through the 8255 69 Parallel I/O Board on the ADD-8080. It required only a simple interface between one output port on the 8255 and the solenoids, peristaltic pump, and the stopped—flow module to affect complete control of the system. The cir- cuit used for this interface is pictured schematically in Figure 7. Since the logic portion of the controller interface consisted of only two integrated circuits, it was unnec- essary to build a full—size ADD compatible circuit board. Rather, a board which could be directly plugged into the SK—10 socket on the top of the 8255 board was deemed more practical. This was accomplished on a small piece of Vectorbord with the use of wire wrap pins as the plug-in connectors. Cables with Cinch-Jones connectors extended from the interface board to allow easy and reliable con— nection to the stopped-flow module and the solid state relays. Control of the peristaltic pump and all solenoids ex- ternal to the stopped-flow module was accomplished with the use of two zero-crossing, TTL controlled, solid state relays capable of switching ac line voltage. The pump and the solenoid that controls the collection loop valves were 70 oomwumucH Hwaaouucoo may mo Emnmmflo pflsonflo coau8pcossuumcH m+ oommumch omomlaom U " mo>ao> u Q one " mooq .--- _ m _ COHHUSUOHDSH . _ u _ H NH 3 . “ m+ have " mmmm _ mm _ mm>Hm> “ y N m _ hm mooq -- mmMII ” “mm coflpowaaou w _ _ .IHW one _ _ .3 mm _ H _ m mean " m+ “mm m oauampmanom _ _ . . _ . «m _ V: .mm _ J» _ m u . 3582 _ a m H gofluemaaonm my " 39$] 8: _ QOmuonmoz a _ _ moo m _ m a. u _ VHH _ _ . _ . .n ousmflm Ln + 71 connected to one relay, since their action is always co- ordinated, and the solenoid that controls the introduction loop valves was connected to the other relay. The two re- lays were mounted in a small metal cabinet to which ac line voltage was brought. The solenoids were connected to the ac side of the relays at the back of the cabinet, and the pump was connected to the relay by means of a line cord receptacle. Cables with Cinch—Jones connectors extended from the dc side of the relays to allow reliable connection to the controller interface. The interface between the relays and the 8255 I/O port consisted of a single 7417 open collector buffer/driver. The GCA McPherson stopped-flow module is set up for digital control by means of a small plug-in socket at the back of the cabinet. Two pins on this plug allow complete control of the module. Since documentation for this proto- type is lacking, the exact circuit function is not known. However, Table 6 lists the control functions that corre- spond to the possible logic states at the two pins labelled B and F on the module. 72 Table 6 Control Functions on the Stopped-flow Module Active Pin Logic Level Function F 0 expel waste F 1 fill syringes B 0 set delay (shortly after F 1+0) B 1 drive syringes Logic level "0" in each case corresponds to a con- nection to ground (Pin D) on the module. In order to buffer the 8255, and to provide the current-sinking abil- ity needed, a single 7403 open collector NAND gate was used to interface the stopped-flow module to the I/O port. Software Command of the controller interface was accomplished in software through use of the Tiny BASIC "OUT A,B" com- mand, where A is the decimal device select and B is the decimal command work for an output port on the microcom- puter. The device select for the 8255 control word 73 register is 6310. The command, "OUT 63,0", initializes the device to operate in a basic input/output mode with all three ports A, B, and C selected for output mode. Port B is used exclusively for generation of all control signals. The device select for port B of the 8255 I/O port is 6110. Thus, all control commands are of the form, "OUT 61,B". The binary value of B determines the state of bits B0-B7 of which bits B4—B7 are active in the system. Table 7 lists the decimal values of B, the I/O bits affected, and the corresponding control function for all control com- mands used with the system. I I The B value of "O" is placed in parentheses in Table 7 since that value is seldom used as such. Obviously, it resets all of bits B0-B7 to zero and could therefore affect all the control functions listed. In actual practice the B value is chosen so that it resets only the desired bit or bits and leaves all others set. For example, to turn on the pump and fill the drive syringes before turning the pump off, the following command sequence must be used: OUT 61,64 (turn pump on), OUT 61,80 (pump on, expel waste), OUT 61,64_(pump on, fill syringes), OUT 61,0 (turn pump 74 Table 7 Command Words Used with the 8255 I/O Port Control Function Value of B I/O Bit Condition 16 B4 Set (0) B4 Reset 32 B5 Set (0) B5 Reset 64 B6 Set (0) B6 Reset 128 B7 Set (0) B7 Reset Expel waste Fill syringes Set delay Drive syringes Start pump and open collection loop Stop pump and close collection loop Open introduction loop to aspirator Close introduction loop to aspirator 75 off). Note that the B value of "0" appears only once even though bit B4 also is reset in the third command. Automated Data Acquisition As mentioned earlier, it was found desirable to auto- mate the acquisition of data from the system, both to re- move the tedium of manual methods and to improve the speed, precision, and accuracy of the process. In terms of the spectrophotometric measurement system employed, this automation entailed the acquisition of a voltage measure- ment that was proportional to the transmittance of an ana- lyte solution in the observation cell. The ADD-8080 was employed for this automated voltage measurement by in- corporation of an analog-to-digital converter (ADC) into the microcomputer system. Since at the time of this work, an ADC circuit board compatible with the ADD-8080 was not available, it was necessary to design, lay out, and assem- ble one. This was done using the basic design approach described by Titus et a1. (59). The attractiveness of this design is that it allows for the interface of an ADC with more than 8 bits of resolution. In this case, a 12—bit converter was employed. A circuit diagram of the ADC 76 board is shown in Figure 8. Details of the connections to the ADC itself are shown in Figure 9. Hardware-—-General There are several key components in the ADC circuit diagram. The most important of these is the ADC itself. The one used for this work was the ADC H212 BC (Datel) which is a 12-bit successive approximation converter with a conversion time of 8.4 us. The ADC was wired for uni- polar operation in the range of 0 to +10 volts. The analog input to the ADC is preceeded by an SHA 5 sample-and-hold (Analog Devices), which has a gain of unity and an input range of -10 to +10 volts. This device allows the voltage from the current amplifier to be sampled on command and maintained at a constant value while the A-to-D conversion takes place on that voltage value. Address decoding and the generation of device selects is achieved through the use of the 74LSl38s and associated logic gates. This will be described in greater detail below. Finally, the output of the ADC is tri-stated onto the ADD-8080 data bus by the two 82123 wired as gated drivers. These gated drivers allow the 8080 CPU to input the 4 most 77 .8 .pumon mo. On Hmuum>coo HMUHmeIODImOHmcm can mo Ecummep DHDOHHU :<.I m4 nZ‘I“§< \ run r Ms 3 2 d anal—‘1. mnfiqch § «s 3 Dom-NI 004 . oboe. :35 > 0. com on 0-41m .m musmwm 0+ :8... 22:. .Uad can Op mcofluooccoo mo Henson 9-9 - “saw “3%. I .w H .. .116 H “.17? i I q I]; 8 3.00. 9 33.3 Eco 2m. Elan” 7 $ H 1 $0.73 :65 0296 9 A l .. an. > $50. 8. .8 23 wt? I 9+? I :99 c. co_w..o>coo to; I mm on mm ON .vN NN ON fiwd . T. omN_NI oo< m m h m m. w. III. I l I I I I l I . m OHDmHm III 332.3 23 I J +.H.U n+ 1 uo_H W 2 8+: So co_u._o>coo Lo 9:. T I _H 79 significant bits (MSB's) and the 8 least significant bits (LSB's) separately in time and only at the command of the CPU. Hardware---Device Selects The 74LSl38 is a three-to-eight line decoder. Two of these were used to decode only the 8 MSB's of the ADD-8080 address bus since these are identical to the 8 LSB's during an I/O instruction. Device select pulses were ob- tained by logical combination of the IN and OUT pulses from the bus with the outputs of the 74LSl38 decoders. These device selects were then used to control the function of various components in the ADC circuit. The decimal value of the device selects used in this application, along with the device selected and the control function performed, are listed in Table 8. Software Software control of the operation of the ADC board was accomplished through use of two Tiny BASIC commands. The "OUT A,B" command described earlier was used for the sam— ple, hold, and convert commands. For these, it was only necessary to generate device selects to affect the desired 80 control function on the board. Thus, the value of B was always zero. All control commands were, therefore, of the form, "OUT A,0". For the input of data from an ADC result, the Tiny BASIC "Y=FIN(X)" command was used, where X is the device select of the input device and Y is the variable name under which the input value is stored in memory. Thus the com- mand, "L(I)=FIN(221) caused the 8080 CPU to input the 8 LSBs of an ADC result from IC G and store that value in memory under the variable name, "L(I)". A similar command was used to acquire the 4 MSBs of a given ADC result. Serial Communication Link to a Minicomputer The ADD—8080 microcomputer with Tiny BASIC (Version 2.1) and the SDK—80 monitor resident in PROM is a powerful system that lends itself well to control functions and to user interactions in the form of terminal output and key- board input. However, the computational power of the system leaves much to be desired. The restriction to in- teger arithmetic imposed by Tiny BASIC, and the lack of logarithmic and exponential functions severely limits the amount of useful calculations that can be performed on the system. In terms of data acquired from the 81 Table 8 ADC Board Device Selects Device Select Device Control Function 217 ADC start conversion 218 S&H hold command 219 S&H sample command 220 IC H (8212) input 4 MSB's of ADC 221 IC G (8212) input 8 LSB's of ADC 82 spectrophotometric measurement system, it was not even possible to calculate absorbance values from measured transmittances. It would of course be possible to equip the system with a higher level language to increase its computational power, but at the time of this work such a language was not readily available. The need for higher level computational abilities thus led to another approach, namely that of a communication link to a minicomputer. It was felt that it would be extremely advantageous to set up a communication link between the ADD-8080 micro— computer and a more powerful minicomputer. The minicom— puter system would support a full operating system and would be equipped with useful peripherals such as a floppy disc drive, and a printer. The ADD-8080 would be able to pass raw data to the mini for higher level calculations and would have indirect access to mass storage. Such a communication link, if attempted in a parallel fashiOn, would involve complex hardware requirements that would be amplified if the bus structures of the two com— puter systems were not identical. Concerns over bus con- flicts and of the coordination of communication would need to be resolved, for the most part, in hardware. There is, 83 however, another approach to a communication link that relies heavily on software and is, therefore, relatively simple to implement and is, in fact, quite elegant. This alternative approach is a serial link from one computer to another. Its elegance stems from the fact that serial communication is virtually universal among mini— and microcomputers and is independent of the bus structure of a particular computer. This makes the serial approach applicable to a variety of quite different computer sys— tems. As will be shown, it is possible to set up a serial communication link between a minicomputer with a 12—bit address and data bus and a microcomputer with an 8-bit data bus and a l6-bit address bus. The implementation of this serial communication link is described by Denker (60) and was brought to the atten— tion of our research group by E. H. Pals. In this method, the microcomputer is equipped with two asynchronous serial ports. One of these is connected to a user terminal while the other is connected to a serial port on the minicom- puter. Software on the microcomputer monitors both of the serial ports. In one mode, the micro is essentially "transparent" in that it merely passes characters between '9‘ v 84 the user terminal and the minicomputer. In this manner, the user has access to the full operating system on the mini, along with all the peripherals. While in the trans- parent mode, the microcomputer checks for "special" characters from both ports. These special characters divert the micro to other software routines. These rou- tines may ready the microcomputer to download a file from the mini into its memory, or merely to return to a monitor program. The number of these special functions is easily expanded in software. The only hardware requirement for this communication link is an extra asynchronous serial port on the microcomputer. Description of Hardware A serial communication link was established between the ADD-8080 microcomputer and a PCM-12 minicomputer (Pacific Cyber Metrics). The PCM—12 is a 12-bit computer based on the Intersil IM6100 microprocessor, and it exe— cutes the instruction set of the DEC PDP 8/e minicom— puter (Digital Equipment Corporation). The POM—12 is equipped with 16K of core memory, a dual floppy drive (Data Systems), and a printer (Digital Equipment Corporation), and runs under the 08/8 operating system (DEC Version 85 3D). Technically, the PCM-12 is a microcomputer, but be— cause it supports a full operating system, it is indis- tinguishable from a PDP 8/e minicomputer. The ADD—8080 was equipped with two 8251 Asynchronous Communication Boards. The device select of the extra 8251 I/O port was changed to E8l6 to differentiate it from the other port. A teletype served as the user terminal that was con- nected to the ADD-8080. It was chosen for the availability of a paper tape reader to load initializing programs for the communication link. Use of the teletype necessitated operation of the communication link at 110 baud. The link between the extra serial port on the ADD-8080 and the serial port on the PCM-12 was run under RS—232 to avoid conflict with the 20 mA current loop. As alluded to earlier, the main purpose of the serial communication link was to pass raw data from ADD-8080 mem- ory to the PCM-12 for higher level calculations. Another goal was to be able to download Intel "binary" files from mass storage on the PCM-12 to core memory on the ADD-8080 in order to bypass the need for keyboard or paper tape 86 entry of programs in hexadecimal code. The software writ- ten for these tasks is described next. Description of Software The software written for use on the ADD—8080 to estab- lish a communication link consisted of three major parts. These are: l) the "transparent" mode program to monitor the two serial ports, 2) the loader program to accept files from the minicomputer and place them in core memory, and 3) the data transfer program to pass raw data from core mem- ory to the minicomputer. Each of these programs was written in 8080 assembly language (61) on the PDP 8/e minicomputer. They were cross assembled on the 8/e using a program called CA8080 (Version 4.1; University of Illi— nois) to obtain hexadecimal code that could be entered through the ADD-8080 monitor program. Source listings of these programs can be found in Appendix C. The function of each of them is described below. LINK80 - The Transparent Mode The "transparent" mode program is the most important part of the serial communication link because it actually controls the exchange of information. A simplified flow 87 chart of this program, entitled "LINK80", is presented in Figure 10. The program begins with the initialization of the extra 8251 serial port. The 8080 CPU then checks the user terminal port for a character. If none has been sent, the PCM-12 port is checked for a character. Again, if none is received the other port is checked. The program remains in this loop until a character is received at one of the ports. When a character is received, it is checked to de- termine if it is a special character. If it is not, the character is merely passed out to the other port. Thus, from the terminal, the user is able to obtain directories, create programs, list files, and perform all other OS/8 operations as if he were communicating directly with the PCM-12 minicomputer. When a special character is received, it causes the 8080 CPU to execute a special function, i.e., it causes a jump to another program. At the user terminal, "CTRL B" is the special character that causes an escape from the LINK80 program to the ADD-8080 monitor. At the PCM-12 port, the special character, "CTRL A", indicates that a 88 C START ) PASS CHARACTER TO PCM-12 EXECUTE CHAR.\. YES SPECIAL FROM FUNCTION TTY? (ESCAPE) NO EXECUTE SPECIAL FUNCTION (DOWNLOAD) PAS S CHARACTER TO TTY . Figure 10. Flow chart for the transparent mode program. 89 "binary" file is to be downloaded. It causes program control to transfer to LOAD80, a program that loads the incoming file into ADD-8080 memory. LOAD80 - The Binary Loader LOAD80 is a program written for the ADD-8080 that allows it to accept and load Intel format "binary" files downloaded from the PCM—12. The word "binary" has been used with quotes because the .BM files resulting from the cross assembler, CA8080, are not true binary files. Rather, they are ASCII files for hexadecimal numbers used to represent instructions, data, and addresses on the 8080 microcomputer. These hexadecimal numbers are arranged in a coded format. The LOAD80 program decodes this format, strips away the ASCII to obtain true binary numbers, and loads these numbers into the proper memory locations on the ADD-8080. Included in the coded format of a .BM file are starting addresses for program storage, byte counts, and intermediate checksums to validate correct transmission of all files. An Intel binary file contains no distinguishing char- acter that could be used as a special character to signal that an incoming file is to be downloaded. Thus, it was 90 necessary to create a file that contained a special charac- ter to preceed all binary (.BM) files. The OS/8 editor was used to create the file, LEAD, that contains a series of CTRL A characters. As mentioned earlier, a CTRL A re- ceived at the PCM-12 port signaled that a binary file was to be downloaded. The procedure for downloading a binary file makes use of the paper tape punch (PTP) handler on the PCM-12. The user calls OS/8 PIP and, in response to the prompt, enters the following command string: PTP: msmuo> > m0 uoHQ mmummom .NH ousmflh .1553: 72 .. 73$ 3. cm. 0.... ed. 0.8 cm on. o.» o“... o... o.» o.~ o._ oo _ _ — - C! o Ex- . 28...... 5.8 123 Conclusions This chapter described the characterization of the analytical system described in Chapter III. Two different studies were performed for this purpose. The first of these was a non-enzymatic study in which an absorbing species was pumped through a simulated enzyme column. Both the flow cell results and the results with the automated stopped-flow system showed that, for this simple case, a linear response was obtained with a relative precision of about 1%. The second study, in which an actual immobilized en— zyme column was used, provided some valuable insight into to system even though satisfactory results were not ob- tained. First of all, volume half-widths calculated for the observed NADH peaks agreed well with those obtained in the potassium nitrate studies as did the time required for elution of the band center. This indicates that, at least for ADH, within the resolution provided by the flow cell, there is no additional contribution to dispersion from the interaction of substrate with enzyme at the glass surface. Therefore a simple model can be used for dispersion in the 124 enzyme column. With this model, an "instantaneous" con- version of substrate to product can be pictured, such that the dispersion apparent in the eluted product band is the same as would be observed if the substrate were monitored after passage through the column. Or similarly, in the case of alcohol dehydrogenase, it is as if NADH were pumped from the introduction loop through an untreated glass column. Measurement of the Michaelis constant points out a practical limitation to the analytical usefulness of im— mobilized enzymes and of enzymes in general. If the Km is relatively small, the reduction of initial substrate con— centrations to levels necessary for linear results may make the enzyme unusable. As the initial substrate con— centration is lowered, the concentration of measurable product produced in the enzymatic reaction may be below the limit of detection. Or, as in the case of alcohol de- hydrogenase, as the enzyme begins to loose activity this limit may be reached. Also, the preparation of standards at very low levels may not be accurate. In the case of ADH, the low molecular weight of ethanol and its volatility make the preparation of ppm level standards undesirable. 125 Because of the difficulty of working with the im- mobilized alcohol dehydrogenase and the rather rapid loss in activity of the preparation, an alternative approach to an enzymatic study was taken. It was decided that a hardier enzyme would be more practical for this work and thus glucose oxidase was again employed for characteriza- tion of the system. This work is described in the next chapter. CHAPTER VIII RESULTS WITH THE AUTOMATED SYSTEM This chapter describes the results obtained for the determination of glucose with immobilized glucose oxidase and the microcomputer controlled system. Included is an overview of the software used for interaction with and control of the system. An evaluation of the stability of the enzyme column and the precision of day to day calibra— tion was performed and is described. Glucose was also determined in a human control serum sample, and the value obtained is compared to the assigned value. Introduction The previous chapter described the use of an immobi- lized alcohol dehydrogenase column to characterize the re- sponse of the system. That enzyme was chosen because a product of the conversion of ethanol can be monitored 126 127 directly. However, the stability of the enzyme was unsat— isfactory and accurate results could not be obtained. Preliminary work with immobilized glucose oxidase in the original system yielded very satisfactory results and indicated that the stability of the preparation was much better than that of ADH. The conversion of glucose by glucose oxidase does not yield a product that can be di- rectly monitored spectrOphotometrically. An indirect method was tried, whereby a composite iodide-molybdate- glucose solution was passed through the enzyme column. This solution appeared to denature or dislodge the enzyme from the column, however, and further flow cell studies were deemed impractical. Nevertheless, it was felt that stopped-flow studies would suffice for glucose oxidase since inferences about peak heights and dispersion could be drawn based on data collected in the potassium nitrate and ADH studies. Furthermore, the use of the same enzyme with the automated system would allow comparisons to be made with the original, manually operated system. Expected improvements in precision and accuracy due to automation of the system could easily be assessed. 128 Software In chapters IV and VI of this dissertation, portions of the software used for operation of the system were pre- sented. Those portions dealt mainly with generation of device selects for control functions and the operation of the serial communication link with the minicomputer. This section presents an overview of the software used for opera- tion of the system for the case in which the user obtains a calibration curve. Emphasis is placed on two subroutines which form a major part of the control program. These sub— routines are the time of delay subroutine and the analog- to-digital conversion subroutine. Source listings of these subroutines can be found in Appendix C. General Description The control program for operation of the system is written in Tiny BASIC, which has been described previously. The INPUT and PRINT commands were used for all interactive communication between the microcomputer and the user. Ini— tialization of the system, which involves filling all lines and syringes with solution and setting the 100% transmit— tance, is assumed to be complete before the program is started. The program begins with initialization of the 8255 129 I/O port. Then the user is instructed to close the shutter on the photomultiplier tube (PMT) module so that a dark current measurement can be made and stored. An A-to-D conversion takes place after the user responds at the key— board. This zero percent transmittance value is printed on the teletype, and the program asks the user for two pa- rameters. These are the number of replicate determinations for each sample and the volume of buffer to be pumped after introduction of a sample. It is assumed that the user has determined the optimum eluting volume by the procedure de— scribed in the previous chapter. After these parameters are entered, the user opens the PMT shutter and signals the microcomputer through the keyboard. The system then goes into full operation starting with a cycle of the stopped-flow module to allow a 100% transmittance measure- ment to take place. The stored dark current value is sub- tracted from this result and the number is stored and printed on the teletype. Next, buffer is pumped to cleanse the column, a sample is introduced and is pumped through the enzyme column, and the optimum volume of eluting buffer follows. The stopped—flow module is cycled, and a trans- mittance measurement is made. The zero value is subtracted 130 as before and this data point is stored and printed. The above process is repeated for the requested number of repli— cate determinations. After the completion of a sample, the microcomputer presents the user with the option of continuation or termi- nation of the program. Continuation allows the user to change samples and to enter new parameters for a fresh series of determinations. Termination of the program causes the number of stored data points to be printed out along with the hexadecimal starting address of data storage in microcomputer memory. This provides the user with the necessary information to transfer data to the minicomputer as described in Chapter VI. As can be judged from this brief description, the software required for operation of the system is relatively simple. Two key subroutines bear further consideration, however, because they point out a special use of Tiny BASIC commands to circumvent limita- tions of that language. The Time of Delay Subroutine Much of the operation of this system involves the use of timed intervals. For example, all volumes proportioned by the peristaltic pump are determined by the time interval 131 that the pump is turned on, since the pump operates at a constant flow rate. Thus, the ability to control timed in— tervals accurately is necessary in the system. Tiny BASIC provides a command that appeared to be the answer to this need. The DELAY N command, where N is the time in milliseconds, provides a variable—length software timing loop that can be used in programming time intervals to the nearest millisecond. However, the accuracy and linearity of this timing loop were found to be unacceptable for direct use. The inability to make CALLS to assembly language sub— routines in Version 2.1 of Tiny BASIC made the incorpora— tion of an external timing loop impossible. Thus it was necessary to use the DELAY command and to compensate for its limitations. The DELAY command can be used in the form DELAY N*Q, where Q is a variable defined elsewhere in a program and * signifies multiplication. The approach taken was to create empirical weighting factors (N values) that would result in accurate times over a short range of Q values. The accuracy of the intervals was determined with a digital counting circuit. Two Counter-Timer-Frequency (CTF) cards made by E and L Co. were cascaded and used to 132 count an accurate 1 KHz clock signal. A bit on the 8255 I/O board in the microcomputer system was used as the start-stop signal. It was possible to insert a DELAY com- mand between two I/O commands and accurately time the in- terval between them. With this approach it was possible to create time intervals accurate to better than 1% over the range of 10 — 120 sec. The limitation on the size of numbers in Tiny BASIC requires that the product N*Q be less than or equal to 32,767. Thus, delays greater than about 30 sec required Special handling in software. For longer delays, several DELAY commands were used in succession. The weighting fac- tor N was scaled down in each case so that the size of N*Q was within the limit. Values of Q were integers that equalled the delay time interval in seconds. Thus, time resolution with the subroutine was limited to the nearest second. Therefore, all volumes delivered by the pump were pre- portionally limited to a resolution of about 20 uL. This was found to be sufficient for all of the work done on the system. 133 The Analog—to—Digital Conversion Subroutine Generation of device selects for operation of the ADC board is described in Chapter VI. These commands were in- corporated into a subroutine that performs 100 A-to—D con- versions in succession, and averages them into a single value. Raw data from the converter are rapidly acquired and then later manipulated before storage. Because the available core memory is limited, the use of large arrays for data storage had to be restricted. The use of a 12-bit converter required that two 100 point arrays be used, each of which used two locations for each value. Therefore, 400 locations were required for the raw data from 100 A-to-D conversions. Thus, when manipulation of the data occurred, they were placed back into the same array. A complete 12—bit conversion is stored in two differ- ent arrays. The 8 LSB's are stored in the first array while the 4 MSB's are stored in the second. The first manipulation required was the combination of these two numbers into a single value. The logical AND and the FSH commands were used to mask and rotate the 4 MSB's so that they could be added to the 8 LSB's in the correct sense. 134 The output of the ADC-HZ12BC converter is complementary binary, so that the final manipulation of a single con- version was to subtract it from 4095 to obtain its comple- ment. In this way, the magnitude of the number corre- sponds directly to the magnitude of the analog input volt- age. To obtain an average of 100 A-to-D conversions re- quired special handling again because of the limit imposed on the size of numbers. The averaging process was divided into two smaller ones. First, groups of 5 numbers were averaged. Then these twenty smaller averages were combined in groups of five. Finally, the resultant 4 values were averaged together. To avoid truncation errors in the pro— cess of averaging, the FMOD (X,Y) function was used. This returns the first digit of any fractional remainder re- sulting from the division of X by Y. This remainder was checked and standard rounding was used to increase the quotient by one when called for. The main program stores the result from the subroutine under the proper variable name for later retrieval. Data were transferred to the PCM-12 minicomputer for calcula— tion and storage as described in Chapter VI. ws'wa » 135 Results and Discussion Calibration Curves All solutions and the immobilized glucose oxidase column were prepared according to the procedures described in Chapter V. The basic procedure for determination of glucose stan- dards was the same as that described in Chapters III and V. The 400 uL introduction loop was used for this work. The effective flow rate with the jacketed glass column was measured to be 1.3 mL min-l. The column was thermostated at 25 r 0.05 0C with the water bath described in the pre— vious chapter. Each glucose sample was separated from the next by 1.5 mL (68 s) of buffer in a series of determina— tions. The enzyme column was stored in a 0.6 M phosphate buffer of pH 6.0 under refrigeration when not in use. The glucose standards were also refrigerated when not in use and were brought to room temperature before use. Control Serum Standard ______________________ An 18 g Lq'solution of zinc sulfate heptahydrate was prepared by dissolving 9.0 g of the solid in distilled water and diluting to 500 mL. 136 A 20 g L.1 solution of barium hydroxide octahydrate was prepared by dissolving 10.0 g of the solid in 500 mL of distilled water. This solution was left unused for 48 hours after which it was filtered and stored. These two solutions were used in combination to de- proteinize a control serum sample prior to determination of glucose in the sample. The deproteinization scheme is as follows. The zinc ion precipitates protein as zinc pro- teinate and sulfhydryl compounds are removed as zinc salts. Excess zinc and barium ions precipitate as zinc hydroxides and barium sulfate. Also, uric acid and creatinine are precipitated and absorbed on barium sulfate. The final centrifuged sample is thus virtually free of protein and deproteinization reagents and is nearly a neutral solution (67). Before use, a 5 mL portion of the zinc sulfate solu- tion was titrated to a phenolphthalein endpoint with the barium hydroxide solution. The more concentrated solution was then diluted so that equal volumes of each solution would contain chemically equivalent amounts of the two species. 137 A vial of Moni-Trol I Chemistry Control (DADE Div. Amer. Hospital Supply Corp.) was reconstituted with 5 mL of distilled water as recommended. Moni-Trol is a human-serum based control material designed for evaluation of the ac- curacy and precision of clinical laboratory procedures. For determination of glucose in the sample, 1.0 mL of the reconstituted material was pipeted into a small test tube, and to this was added 2.0 mL of the barium hydroxide solu- tion. After 30 sec, 2.0 mL of the zinc sulfate solution was added, the tube was stoppered and shaken, and the con- tents were allowed to stand for 2 minutes. The solution was then centrifuged for 2 minutes, after which a 2.5 mL portion was pipeted into a 10 mL volumetric flask. This solution was diluted to the mark with 0.6 M phosphate buf- fer solution of pH 6.0. The final buffered sample repre- sented a 1:20 dilution of the original control serum sam- ple. Glucose in this sample was determined in a manner identical to that used for the glucose standards. Results and Discussion The results of a calibration curve obtained on the first day after preparation of an immobilized glucose 138 oxidase column are presented in Table 12 and plotted in Figure 13. The absorbance values listed are the average of three replicate determinations. The line drawn through the points in Figure 13 is the results of a linear least squares analysis of the data. Linearity was excellent and the relative precision ranged from 1.4 to 5.6%. These results indicate by inference that there is a direct pro— portionality between eluted product peak heights and ini— tial substrate concentrations in an enzyme based system. Also, dispersion of substrate and product must be compara- ble to that observed in previoUs studies. On this same day, the clinical control sample was analyzed and the concentration of glucose was determined to be 81.4 r 0.9 mg dL-l. This compares very well with the reported value of 79.0 i 3.3 mg dL-l determined by a "well established method" based on the use of hexokinase (68). Other reported assay values that are based on the use of glucose oxidase for determination of glucose range from 72 to 86 mg dL”l with relative precision ranging from 0.95% to 6.22% for a minimum of 20 replicate determina- \ tions. 139 Table 12 Glucose Calibration Curve Results Glucose Concentration (ppm) Average Absorbance %RSD 5 0.0937 1.4 25 0.308 5.6 50 0.569 3.8 100 1.09 2.5 ABSORBANCE 0.0 Figure 13. 140 1 I I I l 25 so 75 IOO GLUCOSE CONCENTRATION (mg-L") -h Glucose calibration plot. 141 On the five consecutive days after the above results were obtained, three point calibration curves were ob- tained using glucose standards of 25, 50, and 100 ppm. The absorbance values obtained for these three standards are plotted in Figure 14, which is a plot of absorbance versus the day of measurement. The day-to-day relative precision of the determination of each of the standards was less than 4.1%. However, the deviations in the values were not random. As can be seen in Figure 14, there was a signifi- cant increase in the absorbance values obtained for all three standards on the second and third days. It is possi- ble that several sources of error contributed to the ob- served increase. However, the likelihood of a unidirec— tional contribution is small. The one single factor which seems most likely to have caused a change of the magnitude observed is temperature. The glucose standards were not thermostated at 25 oC. The residence time of a glucose solution in the enzyme column was approximately 10 seconds. This may not have been enough time for thermal equilibra- tion to occur. Therefore, the temperature of the substrate solutions was largely dependent on room temperature. Since a rate method of analysis is applied to the enzyme reaction, 142 Laj- I.I Tl- LO‘P 0.9“” 0.8 “t u 0.7 at 0 i m o s-II- o 0 g: C) ID ID (D g 50 m 4 0.5“ pp 0.47“ O o 3.. 0 ° 0 o o 25 ppm 0.2? O.I 4r 0.0 l J_ l _L l L O I ‘4 5 A— 15 Is DAY OF MEASUREMENT Figure 14. Absorbance versus the day of measurement. 143 an increase in room temperature would have resulted in an increase in the measured absorbance. It is believed that if the glucose standards, and possibly the introduction loop, had been thermostated at 25 OC, better day-to—day precision would have resulted. Based on the calibration curve obtained on the first day, the "error“ in determining the three standards on the next five days was calculated. The results of these calcu— lations are listed in Table 13. It can be seen that large positive errors are associated with the determination of all three standards on the second and third days, as ex- pected. However, on the other days, acceptable errors were obtained for the 25 and 50 ppm standards. A fairly large positive error was associated with the determination of the 100 ppm standard on all five days. This is attribu- ted to an error in the preparation of the standard, since the expected deviation at this relatively high concentra— tion, based on kinetic considerations, would be negative. Next, the "error" in the determination of the 50 ppm standard on those same five days was calculated, based on daily two-point calibration curves constructed from the 25 and 100 ppm standards. The results of these 144 Table 13 Determination of Glucose Standards Based on the First Day Calibration Curve Actual Value Day Average Value % Relative Error %RSD (ppm) Obtained (PPm) 25.0 2 27.5 +10.0 3.5 3 26 5 + 6 0 4 7 4 24 9 + 0 4 4 4 5 24 5 + 2 0 1 1 6 25.0 0.0 4 7 50.0 2 54.3 + 8.6 2.5 3 52.9 + 5.8 1.5 4 50 2 + 0.4 0 24 5 49 9 - 0.2 0.67 6 49.2 - 1.4 1.2 100.0 2 111.2 +11.2 1.4 3 110.8 +10.8 0.75 4 104.1 + 4.1 0.45 5 104.0 + 4.0 1.1 6 105.7 + 5.7 0.47 145 calculations are presented in Table 14. It can be seen that better precision can be obtained with daily two-point calibration curves. The constant negative error observed is consistent with the positive error associated with the 100 ppm standard. The improved stability of the glucose oxidase column compared to the one described in Chapter V is attributed to the storage buffer. The first column was stored in a 0.1 M phosphate buffer of pH 7.0 while the later one was stored in a 0.6 M phosphate buffer of pH 6.0. The storage buffer is an important factor in determining the lifetime of an enzyme column, since, unless it is used continuously, it may spend greater than half its lifetime in the storage buffer. In this case, the difference in pH appears to have been quite favorable to the immobilized enzyme. When glucose standards of 150 ppm and above were de- termined with the present column, results similar to those obtained previously were obtained. That is, a dramatic negative deviation from linearity was observed. At times, absorbance values lower than those for the 100 ppm stan— dard were obtained. This indicates that product inhibi- tion (55) may contribute to the observed deviation in 146 Table 14 Determination of the 50 ppm Standard Based on Daily Two-point Calibration Curves Day Average Value Obtained (ppm) Relative Error %RSD 2 49.0 —2.0% 2.5 3 48.5 —3.0% 1.5 4 48.9 -2.2% 0.24 5 49.1 -l.8% 0.67 6 47.5 -5.0% 1.2 147 addition to the oxygen depletion mechanism proposed in Chapter V. Conclusion The potential of the analytical system described in this dissertation as a practical clinical analyzer has been verified by the work described in this chapter. It has been shown that automation of the system improved the accuracy and precision of results as compared to those presented in Chapter V. The determination of glucose in a clinical control sample yielded a value in excellent agree- ment with established clinical methods. The relative pre- cision of that analysis was about 1.1%, which is a highly acceptable value. The enzyme column was shown to be usable over a six day period with a single calibration curve. Relative errors ranged from -1.4% to +11.2%, and much of the error is attributed to a temperature effect. More con- sistent results were obtained with daily calibration curves. It is difficult to predict the useful lifetime of the glucose oxidase column. In a clinical laboratory, it would receive heavier use than it did in this work. Daily use consisted of 15-20 determinations at 25 0C over a 1-2 hour 148 interval. The remainder of the time, the column was re— frigerated. Based on the observed stability, however, it is estimated that a useful lifetime of several weeks can be expected, and it is possible that such a column could last for several months. Although it has a great deal of potential, the system in its present form is less than practical for routine clinical analyses. Suggested improvements on the system toward that end are presented in the next chapter. CHAPTER IX FUTURE PROSPECTIVES This chapter presents some ideas and suggestions for possible improvements of the system described in this dis— sertation. These improvements range from the very practi- cal to the very idealistic, but all are nonetheless included. Emphasis is placed on two major areas which ex— perience has shown to be weak or lacking in the system. These areas are sample handling and microcomputer software. Each of these is discussed in turn below. Sample Handling An automated system, in the broadest sense of that phrase, is one which requires almost no human intervention for operation, except perhaps to start it. By this stan— dard, the system described in the preceeding chapters is not fully automated. One particular step in its operation requires frequent user intervention. This step is the sample changing step. In its present form, the system 149 150 requires the presence of the user to change samples after the requested number of replicate determinations on a sample is completed. A possible future addition to the system, to alleviate the need for frequent intervention, would be an automated sample changing mechanism. This would consist of a ro- tating sample turret driven by a stepper motor. This turret would have multiple sample capacity and would allow all samples to be partially immersed in a thermostated water bath. The thermostating of samples is highly recom— mended to enhance the day—to-day precision of determina- tions. A retractable probe, coordinated with the rotating turret, would dip into a sample or a buffer solution at the appropriate time. This sample probe would be connected to the inlet of the sample introduction loop with flexible tubing. In this way, since the turret could be loaded with multiple samples at one time, the user would not need to change samples during operation of the system. The micro- computer would control the "change" of samples by rotation of the turret to a new position when necessary. Calibra— tion standards could be included in the turret to check or prepare calibration curves before the determination of 151 samples. Such an automated sample changing mechanism would not only free the user from sample changing duties, but would also increase system throughput, since the time required for the change of a sample would be greatly re- duced by automation. Other steps could be taken to increase system through- put. The flow rate of sample solutions through the enzyme reaction column could be increased. This would result in a decrease in the system's response since less measurable product would be produced. However, results presented in the previous chapters indicate that some reduction in sig- nal could easily be tolerated. To implement an increased flow rate would necessitate the use of a different pump in the system. A practical limit on the flow rate that can be obtained with the peristaltic pump is about 1.6 mL min- . Also, irregularity of flow is still a problem with the present pump. It is suggested that a stepper-motor driven syringe may serve well as a pump mechanism in the system. It would almost certainly provide a more regular flow pattern and may be able to provide higher flow rates. A second step toward an increase in system throughput would involve the replacement of the present introduction 152 loop valves with four—way HPLC slider valves and the addi- tion of a three-way valve preceeding the collection loop. A schematic diagram of the resultant flow configuration is shown in Figure 15. The use of the four-way slider valves and the extra three-way valve would allow continuous Opera- tion Of the pump without any additional waste of solution. With the introduction loop valves in the position shown in Figure 15, a sample could be introduced while the pump is operating. The three-way valve preceeding the collection loop would allow solution from the enzyme column to be di- verted to the waste while the contents of that loop are pushed into the stopped-flow module for a transmittance measurement. Thus, a new sample could be introduced into the system, while a second was eluted from the enzyme col— umn, and a third was determined in the stopped-flow module. The total time that was formerly needed for introduction, elution, and measurement of a sample would be reduced be- cause the first and last Operations could take place during the elution Of a sample from the enzyme column. Since elu- tion is the slowest step, the inclusion of the other two steps during this interval should not pose a problem. INTRODUCTION LOOP Figure 15. 153 SAMPLE __ VACUUM .___ _ WASTE COLLECTION LOOP VALVE Proposed valve changes. (— PUMP ENZYME REACTION COLUMN 154 The alternate position of the introduction loop valves would allow a sample to be pushed into the enzyme column as described previously, and the alternate position of the extra three-way valve would allow solution to enter the collection loop for collection of the product band center. Microcomputer Software As.described in Chapter VI, Tiny BASIC Version 2.1 is a useful, but limited computer language. It is limited both with respect to its computational power and with re- spect to the speed of its I/O capabilities. The computa- tional limitations have already been discussed. The limitation on the speed of I/O transfers stems from the fact that Tiny BASIC is an interpretive language. Thus, I/O commands are greatly slowed down in the interpretation process. As an example, the execution time Of the OUT A,B command was measured to be 4.1 msec. This time was found to be tolerable in this system, but it would be unaccep- table is faster I/O capabilities were needed. Earlier versions of Tiny-BASIC allowed CALLS to assembly language subroutines. This would have served to speed up I/O 155 capabilities, since I/O subroutines could have been stored in directly executable machine code. However, this option was removed from Version 2.1. The establishment of a serial communication link to a minicomputer helped to alleviate the computational problem. However, to program in Tiny BASIC was still an exercise in ingenuity. That is, one had to constantly work within rather strict limitations to Obtain the desired results. It is felt that higher computational abilities are needed on the microcomputer itself. Microcomputers are still in a relative state of in- fancy, as is the software available for use with them. It is assumed that, as time progresses, software Of the cali- ber now available for minicomputers will become available for microcomputers. At present, two languages that should be considered for ADD-8080 software are CONVERS (69) and a 7K version of BASIC. Each of these is described briefly below. CONVERS CONVERS was developed at the Department of Chemistry Of the University of Arizona for 8080—based microcomputers. 156 It is described by its authors as an "interpretive com- piler" because it is interactive during entry of a program and yet it reduces source code to machine code for later execution. The most attractive feature of this language is that user defined subroutines become a permanent and integral part of CONVERS. This includes subroutines built from previously defined subroutines. Thus, programming becomes progressively easier once the user has created a suitable "library" of subroutines. This author has had no practical experience with CONVERS and so it is difficult to assess a feature such as the computational power of the language. However, its authors describe a floating point package for the four Sim- ple arithmetic Operations. It is assumed that algorithms for other Operations could be added to the language. In any case, floating point arithmetic is already a step up from the integer arithmetic of Tiny BASIC. Since subroutines are compiled to machine code, the execution Of I/O routines should not be slowed down in CONVERS as they are in Tiny BASIC. Thus, the limitation on the Speed Of I/O transfers would be overcome with the use of CONVERS. 157 7K BASIC A 7K version of BASIC has been developed at the Uni- versity of Illinois for use of the ADD-8080. It is a modification and extension of a 5K version (70). The con- tents of this new language make it an excellent choice for microcomputer software. The 7K version of BASIC contains the high-level arithmetic Operators such as exponents and logarithms that are absent in Tiny BASIC. Also, the ability to CALL assembly language subroutines is included. Thus, the major limitations of Tiny BASIC would be over— come with the use of this new 7K version. The choice of either CONVERS or BASIC Should be made in the context Of its intended use. One may prove more suitable than the other for a particular application. APPENDIX A ADDITIONS AND MODIFICATIONS TO THE ADD-8080 APPENDIX A ADDITIONS AND MODIFICATIONS TO THE ADD-8080 This appendix describes additions and modifications to the ADD-8080 microcomputer system implemented by this au- thor after the system was received from the University of Illinois. All notation used to describe changes on the PC boards follows exactly the notation used in "The ADD-8080 Microprocessor Manual" (71). The reader is referred to this manual for complete documentation of the system. Cir— cuit diagrams are included in this appendix to clarify the changes made. Dotted lines in these diagrams indicate original connections and asterisks designate new connec- tions. The RESET Modification In the ADD-8080 microcomputer system, the RESET sig- nal generated by the 8224 System Controller on the Micro- processor CPU Board is bussed to the 8251 Asynchronous 158 .- ". '.T'. 11% n}? " ' ‘ 159 Communications Board and to the 8255 Parallel I/O Board so that when the 8080A CPU is reset, both Of these periph- erals are also reset. The RESET signal is active high on the bus, however, and this leads to a serious problem in the system. Noise Spikes picked up on the bus can cause a reset of both of the peripherals without a reset of the CPU. Thus, serial and parallel I/O capabilities can be lost during operation of a program. To circumvent this problem, the CPU board was modi- fied to generate an active low RESET signal so that the RESET line on the bus in normally high. This makes the peripherals virtually immune to noise-spike resets. The two peripheral boards were also modified to accept the in- verted RESET. The changes required to implement this modi- fication on each of the three boards are described in detail below. 8080A Microprocessor CPU Board (Revision 1A) A small piece of Vectorbord was epoxied to the top edge of the CPU board over the connector area for the secondary bus. This was used to mount a 7404 hex inverter. Ground and +5 V were brought to this chip with insulated jumper wires directly from the power pins. A 0.1 uF W m --L_x~_- L-J‘TL‘K“? .3v ”1... .. ’_."L.'. "-1 I 1.: 160 decoupling capacitor was connected between power and ground at the chip. The foil pattern between the RESET main bus connection (#35) and pin 5 of IC G was severed. Insulated jumper wires were used to connect pin 5 of IC G to pin 1 of the 7404, and to connect pin 2 Of the 7404 to the RESET main bus connection (#35). This modification resulted in inversion of the RESET line on the bus. All further refer- ences to this line will thus use the notation RESET. The changes described above are represented in the circuit dia— gram of Figure 16. 8251 Asynchronous Communications Board Revision 1B The foil pattern between the RESET main bus connection (#35) and pin 21 of the 8251 was severed. Insulated jumper wires were used to connect the RESET main bus connection to pin 5 of IC I and to connect pin 6 of IC I to pin 21 of the 8251. This resulted in inversion of the RESET line prior to reception at the 8251. Thus, the 8251 receives the re- quired active-high RESET signal. The change described above is represented in the circuit diagram of Figure 17. 161 G I Reset 1 12 . I (35) Reset : 8224 L----_-__ Reset 8080A Figure 16. Modifications on the 8080A CPU Board 162 pumom Hmmm map so mQOHpoOAMHpoz .BH ossmwm m+ WI IV I WED R C III m+ 4 m A HA HA 0H ea c m ADO use oil R .N.H s ma MMAmw A ma o A IIIIIIIIII venom m+ m+ ” 4 mm powmm . mom a a m mm 9 a TIIIIIIII I. owe _ N. _ EUR v a - m "E. a we m _ m Hmmm Hm RHI. (H R m 2H mm OII I mmm e .Hu HA AHI. a mH Rh Nb mm.,flm m a 163 Revision 1C The design of this board differs from Revision 1B in that IC J is replaced by a 7428, and two additional gates on IC I are used to invert signals from pins 1 and 10 of IC J before they are sent to the card top connectors labeled Z4 and Z6, respectively. This author replaced IC J with a 7433 which is the Open collector analog of 7428. Two 680 ohm pull-up resis- tors were added to the board between pins 4 and 13 of IC J and the +5 V power pin. Since the Z connector strip is not used in this sys- tem, the foil pattern between pin 5 of IC I and pin 1 of IC J was severed, as was the pattern between pin 6 of IC I and connector Z4. This freed the inverter on IC I for use with the RESET line as described previously for Re- vision 1B. 8255 Parallel I/O Board A 7404 hex inverter was mounted in the lower left hand corner of the 8255 PC board. Ground and +5 V were jumper wired to this chip from nearby power bus lines and a 0.1 uF decoupling capacitor was connected between power and ground at the chip. The foil pattern between the 164 RESET main bus connection (#35) and pin 35 of the 8255 was severed. Insulated jumper wires were used to connect the RESET line to pin 1 of the 7404, and to connect pin 2 of the 7404 to pin 35 of the 8255. This resulted in inversion of the RESET line from the bus before reception at pin 35 of the 8255, as required. The change on the 8255 board is represented in the circuit diagram of Figure 18. The RS232 Modification The ADD-8080 microcomputer system was modified to allow the use Of an RSZ32-based CRT terminal for serial communication with the system. This modification required some additional wiring on the 8251 Asynchronous Communica- tions Board, and the addition of an RS232 driver circuit to the system. Also, special cables were needed for con- nections within the system. Each of these changes is de- scribed in detail below. Wiring Changes on the 8251 Board Both revisions of the 8251 Asynchronous Communications Board were changed in the same manner. Pin 7 of IC K was connected to ground and pins 5 and 6 were connected to +5 V via insulated jumper wires. This causes the output at 165 pumom mmmm on» so mcoHPOOHMHpoS mm .wa onsmflm wommm mmmm 166 pins 11 and 12 of IC K to remain at a logic "1" so that the 8251 clear-to-send (CTS) line at pin 17 receives a constant logic "0". This enables the 8251 to transmit when the 75116 is in place without the need for an externally gen- erated CTS signal. The 75116 is capable of accepting RS232 generated at the CRT terminal, since it has differential line receivers that accept input voltages of i25 V. These receivers gen- erate TTL level signals that can be input directly to the 8251. Thus, for RS232 input from the terminal at IC L, T15 V are sent to pins 5 and 6, and pin 7 is connected to ground. A +15 V signal at the inputs results in a logic "0" at imput pin 3 Of the 8251. Thus, the correct sense Of the signal is received at the 8251. The wiring changes described above are represented in Figure 17. The R8232 Driver Circuit Board The line driver portion of the 75116 generates only TTL level output signals. Thus, for communication back to a CRT terminal, it is necessary to generate RS232 level signals externally. This was implemented on a small piece Of Vectorbord that was equipped with ADD compatible power pin connectors. The circuit used is based on an LM339 “huh...- 77.7 .35... _ . .. .— ., _._- ,: .a' 7:.-—_- - ,3- ~ 167 quad comparator, and its function is described in Kilobaud Magazine (72). A circuit diagram of the dual R8232 driver circuit is presented in Figure 19. The TTL output at pins 1 and 2 of IC K is sent to either of pins 4 and 10 on the comparator to obtain the correct sense of the output sig- nal. Connecting Cable A cable was designed for connection of the 8251 board to the CRT terminal. The necessary connectors were at- tached to each end of a multi-stranded cable. The use of the R8232 driver circuit required a special modification on one line of the cable. A short wire from the TTL out— put connector pin (#15) on the 8251 board was fitted with a small spring clip connector. This spring Clip attached to a metal post at the top of the R8232 driver board to bring the TTL signal to the circuit. The appropriate line of the cable was also fitted with a spring clip that at- tached to a second post on the R8232 board. This line carried the R8232 signal to the terminal. The clip-on de- sign of the cable provided flexibility in the use of the 8251 board by allowing it to be disconnected from the driver board when a teletype was used with the system. 168 BDO BDO pumom um>HuQ mmmmm may mo EMHWMHQ HHSUHHU m omm ma mmmmm oii¢<<< Mm.m mH+ Mm.m H mm Nmm u )2)», - omm N mmm SQ ullllé mH+ .ma musmum m illllo mHI NH 0 2H muse OH _ _ a o IIAltfiII 2H Hues Ha m Mos Mma 169 The Additional 8251 Board As described in Chapter VI, an extra serial port was added to the system to establish a serial communication link between the ADD-8080 microcomputer and the PCM-lZ minicomputer. The use of this extra port required a change in its device select code from the normal value. Also, the cable used for connection of the two computers re- quired a slight modification of the cable normally used with a CRT terminal. Each of these changes is described below. Change of the Device Select The device select code of the extra 8251 serial port (Revision 1C) was changed from a hexadecimal value of F8 to a value of E8. This was accomplished by moving the jumper wire labeled Jl from pin 7 of IC A to pin 9 of IC A. Note that IC A and IC B, J1 and J2, and 81 and 82 are incorrectly labeled in Figure A-28 of the ADD-8080 Manual (71). Each pair of labels should be exchanged for correct labeling. The change in position of the jumper wire and the correct labeling are shown in Figure 17. 170 Connecting Cable The cable used to connect the ADD-8080 serial port to the serial port of the PCM-lZ is not identical to the cable used for connection of the ADD-8080 to a CRT terminal. When the serial link is established, the ADD-8080 mimics the function of a terminal and so the position of its send and receive lines relative to those of the PCM—lZ must mimic those of a terminal. Otherwise, normal connection of the two ports results in conjunction of two send lines and two receive lines. Thus, the normal position of the send and receive lines was reversed at the PCM-12 cable connec- tor for correct operation of the link. l APPENDIX B A STOPPED-FLOW CLINICAL ANALYZER IN WHICH IMMOBILIZED-ENZYME REACTION LOOPS ARE USED Reprinted by permission of the American Association of Clinical Chemists 171 [Reprinted from CLINICAL CHEMISTRY. 23. 1033 l 1977).] . Copyright 1%? by the American Association of Clinical Chemists and reprinted by permission of the OOPYfl‘ht owner. A Stopped-Flow Clinical Analyzer in Which Immobilized-Enzyme Reaction Loops Are Used Marlin 0. Joseph. Daniel J. KaSprzak, and S. R. Crouch‘ A stapped-llow clinical analyzer is described that makes use of a reaction loop containing immobilized enzyme(s) for the determination of the analyte/substrate. The analyzer has been evaluated by determining glucose with immov bilized glucose oxidase. The stooped-flow mixing system was constructed at a current cost of less than $500. The analyzer separates the enzymatic reaction from a followup. spectrOphotometric indicator reaction. This separation allows the enzymatic reaction to be used in either a fixed-time. kinetic mode or in an equilibrium mode. Like- wise. the indicator reaction can be used in either mode. Results tor glucose in blood serum indicate that good precision and accuracy can be obtained. In recent years immobilized enzymes have become increasingly useful as specific catalysts for determining clinically important substrates (1-4 ). Immobilized en- zymes are attractive as reagents in the clinical labora- tory because they possess the usual specificity and sensitivity of soluble enzymes and can be reused. often in hundreds or thousands of determinations, thus greatly reducing the cost of routine use of an enzymatic kinetic or equilibrium method. In addition, immobilized enzymes are often more stable than they are in solution: calibration is thus simpler and less frequently re- quired. In this report, a stopped-flow clinical analyzer is de- scribed that combines the advantages of immobilized enzymes with the speed. mixing efficiency, and ease of automation of the stopped-flow mixing technique. The system utilizes a simple. low-cost. stoppedilow sam- pling and mixing unit together with a reaction loop that contains the immobilized enzyme. The reaction loop is similar in principle to the sample loop described by Pardue e‘. al. (5, 61 except that the enzyme-catalyzed reaction occurs in the loop. After a suitable incubation period in the loop. the sample solution is rapidly mixed by the stopped-flow unit with any desired reagent(sl and sent to an observation cell for spectrophotometric monitoring of the appropriate indicator reaction. The separation of the enzyme-catalyzed reaction from the . Department of Chemistry. Michigan State University. East Lan. sing. Mich. 48824 ‘ To whom reprint requests should he addressed. Received Feb. 23. 1977; accepted Mar. 30. 1977. indicator reaction provides extreme versatility. The enzyme-catalyzed reaction can be allowed to go to completion (equilibrium method) or it can be incubated for only a short. fixed time interval (kinetic method). Likewise. either the reaction rate or the equilibrium absorbance of the spectrophotometric indicator reaction can be used in the final measurement. Because of stopped-flow mixing. measurements can he made on time scales ranging from milliseconds to hours. The semi-automated stopped-flow clinical analyzer described below illustrates the measurement principle and chemical flexibility of the system. We evaluated the analyzer by determining glucose in blood serum with a reaction loop containing immobilized glucose oxidase (EC 1.1.3.4). Possible extensions of the system to mula tiple determinations of the same substrate or to near- simultaneous determinations of multiple substrates are discussed. Materials and Methods instrumentation The instrumentation we used to determine glucose is schematically represented in Figure 1. This system. except for the enzyme reaction loop. was constructed completely from commercially available items. Two l-ml gas syringes with threaded ends. A and B in Figure 1 (from Glenco Scientific. Houston. Tex. 7700?). are used to deliver indicator reagent and push liquid. re- spectively. Valves I through 4 are 0.060-inch hore. threewvay slider valves complete with accuator-return mechanisms C and D and solenoid controllers $2 and 83 (from Altex, Berkeley. Calif. 94710). The syringes are driven by a 2-inch stroke pneumatic cylinder. PC. with solenoid controller 51 (Scovill. Wake Forest. N. C. 275871. Solutions are mixed by forcing them through E. a KEL-F tee (Altex). The indicator reaction is moni- tored in a 250411 micro flow cell (F) with IO-mm light path (Thomas. Philadelphia. Pa. 19105). The flow cell is fitted into a singleoheam xiv-visible spectrophotomo eter (GCA McPherson EU-TOO. Acton. Mass. 01720). All tubing except for the enzyme reaction loop is 1.5 mm id, 3 mm o.d. Teflon with appropriate connector fit~ tings (Altex). CLINICAL CFEMlSTRY. Vol. 23. No. 6. 1977 1033 172 “CYRWHOYOKTER Fig. 1. Diagram of stepped-flow clinical analyzer The heart of the system is the enzyme reaction loop. prepared as described below and fitted with plastic high-pressure liquid chromatography connectors. The enzymatic reaction that occurs in the loop for the de termination of glucose is: glucose Glucose + 02 + [—120 om gluconic acid + H202 The extent of this reaction is determined outside the loop by allowing the H202 produced to mix with and rapidly oxidize iodide to triiodide in the presence of a molybdate catalyst (7"). The indicator reaction is: M VI) H202 + 31- + no —°‘—» 2H20 + 13- Triiodide. measured spectrophotometrically at 365 nm. indicates the concentration of glucose in the sample. The indicator reaction is followed by an equilibrium method in this case. because the absorbance of triiodide is measured at equilibrium. The enzymatic reaction, however, is of the fixed-time reaction-rate type. The enzymatic reaction time is fixed by the incubation time of the sample in the loop. This time is kept short (2 min) to ensure that the enzymatic reaction is pseudo-zero order. This combination of reaction-monitoring meth- ods exemplifies the flexibility of the system. The actual function of the enzyme reaction loop in the system is best described in terms ofa typical sequence of events for the determination of glucose in a sam- p e. First, syringes A and B are filled with KI/molybdate 1034 CLINICAL CtEMlSTRY. Vol. 23. No. 6. 1977 reagent and a push liquid (phosphate buffer). respec- tively, by simultaneous switching of valves I and 2 to the fill position and retraction of PC. Valves I and 2 are then switched to the “push-ready" position. as are valves 3 and 4 (positions shown in Figure 1). Next. PC is caused to push the contents of A and B through the system lines and out to the waste. This is repeated several times to remove air bubbles and fill all lines with solution. With A and B filled and valves I and 2 ready. valves 3 and 4 are simultaneously switched to the fill position, and a buffered glucose sample is aSpirated into the enzyme reaction loop by a gentle negative pressure. Then valves 3 and 4 are switched to the “push-ready" position, which terminates sample introduction. The sample is incubated for a controlled time until PC is caused to push the contents of the enzyme loop through the mixer. Here the H202 produced by the enzymatic reaction is rapidly mixed with the Kl/molybdate re- agent, and the mixture is transported to the flow cell for observation and detection of 13‘. The flow is stopped upstream when the drive syringes reach the ends of their travel. The system described above was constructed at a cost of less than $500, not including the specific flow cell and Spectrophotometer mentioned here. for which there are suitable substitutes. The instrument can be assembled within a day or two due to the ease with which the valves and tubing can be connected via plastic fittings rated at 3450 kPa (500 psi). The simplicity, low cost. and versatility of this system make it an attractive analyzer for clinical determinations. Reagents Buffer. A 0.2 mol/liter solution of dibasic potassium phosphate in de-ionized water was adjusted to pH 6.30 :t 0.02 by use of a Heath pH meter (Model EU-BOZA) by adding concentrated hydrochloric acid. This buffer was then used to prepare all other solutions except where noted. Indicator reagents. A sodium molybdate stock so- lution, 90 g/liter. was prepared and allowed to equili- brate for 48 h. The iodide/molybdate indicator reagent was then prepared daily by adding a fresh solution of potassium iodide to molybdate stock. resulting in a so- lution containing. per liter, 18 g of sodium molybdate and 0.5 mol of iodide. Glucose standards. A glucose stock solution was prepared by dissolving 1.00 g of anhydrous granular D-glucose and diluting to 1.0 liter with de-ionized water. This solution was allowed to stand at room temperature for 24 h to ensure complete mutarotation, and was then stored at 5 °C. Aqueous glucose standards were then prepared by appropriately diluting this stock with the buffer. Deproteinizotion reagents. We deproteinized the serum samples by adding chemically equivalent amounts of barium hydroxide and zinc sulfate. A bari~ um hydroxide solution (20 g/liter) was titrated with zinc sulfate (20 g/liter) to a pH of 7 to determine the required volume ratio of the two reagents. 173 Preparation 0/ serum samples. A control serum (“Montrol"; Dade Division, American Hospital Supply Corp., Miami. Fla. 33152) was reconstituted according to the manufacturer’s instructions, and 0.5 ml of the solution was pipetted into a centrifuge tube. The chemically equivalent amounts of barium hydroxide and zinc sulfate were added. and the contents of the tube were thoroughly mixed. centrifuged for 3 min. and 1.0 ml of the supernatant fluid was pipetted into 3 ml of the phosphate buffer. This solution was then drawn into the enzyme reaction loop for analysis. Immobilization of glucose oxidase. A procedure re- ported by lnman and Hornby (8) was modified for im- mobilizing glucose oxidase on the inner surface of nylon tubing, 100 cm long and 0.86 mm id Most of the imo mobilization steps were accomplished by slowly pumping the required reagent through the tube with the use of a Technicon AutoAnalyzer pump to provide a constant stream of fresh reagent. Amorphous nylon was first removed by filling the tube with a 200 g/kg solution of CaCl2 in methanol and incubating at 50 °C. The nylon was then mildly hydrolyzed by pumping 1.0 mol/liter HCI through the tube at room temperature, followed by rinsing with water. Glutaraldehyde. 1‘25 ml/liter in 0.1 mol/liter trislhydroxymethyllmethyl- amine buffer, pH 9.2. was then attached to the hydro- lyzed nylon at 0 °C. followed by rinsing with the buffer. An 8 g/liter-solution of glucose oxidase (Sigma, Type II) in 0.1 mol/liter phOSphate buffer, pH 6.3, was then pumped through the tube at 0 cC for 4 h in a closed loop. during which time covalent linkages were formed be- tween the glutaraldehyde and free amino groups on the enzyme (8). Finally. the tube was rinsed with 0.1 mol/ liter sodium chloride to remove any physically adsorbed enzyme. The enzyme reaction loop was filled with buffer and stored at 5 °C when not in use. Results and Discussion The analytical capabilities of the analyzer system were evaluated by determining glucose in aqueous samples and in blood serum. with use of a reaction loop containing immobilized glucose oxidase. The first generation analyzer we used in these studies was man- ually sequenced. That is, the solenoid controllers for each of the valves in Figure 1 were actuated in the ap- propriate sequence by a series of manual toggle switches, which applied or disconnected the ac line voltage from the controllers. The spectrophotometric readout was obtained from a strip-chart recorder. linear in transmittance. Thus. the final absorbance values recorded for 13' were obtained by converting chart- recorder readings from transmittance to absorbance. The temperature of the enzyme reaction loop was not controlled. The optimum push volume of the stopped-flow sys- tem was determined to be 0.78 ml per syringe by filling a dummy reaction loop with dye (p-nitrophenol). This volume ensured thorough purging and refilling of the observation cell, but prevented significant dilution of the observed solution by the push liquid. The incuba- Tablo 1. Results tor Aqueous Glucose Standards W" CV * m’ 20 0.5549 2'0 40 0.0510 3.0 so 0.0960 2.1 30 0.127 4.7 100 0.162 1.3 150 0.240 0.8 200 0.314 0.1 250 0.333 1.3 Slope = 1.60 X 10‘3 A per mg/llter (CV. 0.24%) Intercept = *6.39 X 10‘3 A Corral- coett. (r) = 0.99990 tion time for the enzymatic reaction was fixed at .‘2 min. Shorter incubation times did not provide adequate precision because of decreased response and also be. cause of imprecision in manual sequencing: longer in cubation times only added to analysis time without greatly enhancing the signalato-noise ratio. Hence. transmittance was directly read out slightly over ‘2 min after the sample was introduced into the reaction loop. A calibration curve for glucose was prepared in the range 20—250 mg/liter by running aqueous standards in triplicate. Table I shows the results. The observed negative intercept is expected in such a fixed-time method. because the enzymatic reaction is known to exhibit an induction period that is inversely propor- tional to glucose concentration 19), and hence a two- point calibration should be done periodically. The re- sults in Table 1 indicate the excellent linearity and precision which may be obtained with this simple ana- lyzer. Each determination requires only 1 ml of sample (about 100 pl of blood serum). The major sources of imprecision in this simple system appear to be varia- tions in incubation time due to manual control. fluctu- ations in reaction rate due to variations in room tem- perature, and imprecision in reading the transmittance from the chart recorder. For determining glucose in blood-serum samples. we prepared a two-point calibration curve. using the ‘20 and 200 mg/liter aqueous glucose standards. Two reconsti- tuted Monitrol samples were analyzed. and the results (Table 2) agree well with the reported values. Note that the first two values reported for sample 1 were obtained on a different day from the second two values. although the same two-point calibration was used. C learly. be- cause of the high stability of the immobilized enzyme loop. the same calibration curve may be useful for sev- eral days. A reaction loop prepared earlier than the one used in these determinations showed excellent activity for six months before the enzyme was de-activated in- advertently by incubation with a highly acidic solution. Research is underway to determine more precisely the activity loss of immobilized enzyme loops with time under various conditions of storage and use. CLINICAL CPEMlSTRY. Vol. 23. No. 6. 1977 1035 174 Table 2. Results of Determination or Glucose In Serum aueeee canon. Witter Detennhed Reported Werenoe‘ Sample 1 690 2.6 725 2.1 703 690.” 730° 0.9 673 5.2 Sample 2 2006 2090" 4.0 2078 0.6 ' 96 difference from the av reported value ' Beckman “Glucose Anaryzer" ‘ American Monitor "Program" In this study we have demonstrated the potential of the analyzer for the determination of glucose in serum samples. Although the system was evaluated by using one particular enzyme/substrate couple. many other clinically useful enzyme/substrate couples are obviously applicable. Versatility is inherent. not only in the en- zyme chosen for the reaction loop, but also in the re- agent chosen for the indicator reaction. Furthermore, any of several combinations of methods (kinetic or equilibrium) may be used to follow the separate enzy- matic and indicator reactions. We are now undertaking investigations on the capabilities of this new system in the clinical laboratory. New immobilization techniques are being studied for bonding other. more expensive enzymes to various inert supports. The multiplexing potential inherent in the system will be used, to allow several reaction loops containing the same or different 10” CLINDCAL CFEMISTRY. Vol. 23. No. 6. 1977 enzymes to be used nearly simultaneously, thereby minimizing the time lost in incubation and maximizing the use of one detection system. The entire system will then be automated by interfacing it to a minicomputer or microprocessor for sample handling, instrument control, and data acquisition and subsequent process- ing. A more extensive paper describing this work will be submitted at a later date. We thank Mr. Robert Lantz for his initial work on immobilization procedures and Mr. F. J. Holler for his assistance in constructing the stopped-flow apparatus. This work was supported in part by National Science Foundation grant No. CH 876-81203. References l. Weetall, H. H.. Immobilized enzymes: Analytical applications. Anal. Chem. 46. 602A (1974). 2. Bowers, L. D.. and Carr. P. W.. Applications of immobilized en- zymes in analytical chemistry. Anal. Chem. 48. 544A (1976). 3. Leon. L. P.. Narayanan. S., Dellenbach. R., and Horvath. C.. [m- mobilized glucose oxidue used in the continuous-flow determination ot'scrum glucose. Clin. Chem. 22. 1017 (1976). l. Bowers. L. 0.. and Carr. P. W., An immobilized-enzyme flow- enthalpimetric analyzer: Application to glucose determination by direct phosphorylation catalyzed by hexokinase. Clin. Chem. 2. 1427 ”976). 5. Sanderson. D.. Bittikofer. J. A.. and Pardue. H. I... Computer controlled stopped-flow studies—application to simultaneous kinetic analyses. Anal. Chem. 44. 1934 (1972). 6. Mieling. C. E.. Taylor. R W" Hargis. L. D.. et al.. Fully automated stopped~flow studies with a hierarchical computer-controlled system. Anal. Chem. 48. 1686 (1976). 7. hialmstadt. H. V.. and Pardue. H. L.. Quantitative analysis by an automatic potentiometric reaction rate method. Anal. Chem. 33. 1040 (1961). 8. Inman. D. J.. and Hornby. W. E.. The immobilization of enzymes on nylon Structures and their use in automated analysis. Biochem. J. 129. 255 (1972). 9. Malmstadt. H. V.. and Crouch. S. R. Systems for automatic direct readout of rate data. J. Chem. Educ. 43. 340 (1966). APPENDIX C SOURCE LISTINGS OF MICROCOMPUTER PROGRAMS 175 Tiny BASIC Subroutines The Time of Delay Subroutine 200 IF Q>=12l GO TO 220 202 IF Q>=115 GO TO 222 204 IF Q>=100 GO TO 224 206 IF Q>=9l GO TO 226 208 IF Q>=78 GO TO 228 210 IF Q>=64 GO TO 230 212 IF Q>=48 GO TO 232 214 IF Q>=32 GO TO 234 216 IF Q>=18 GO TO 236 218 DELAY 1040*QzRETURN 220 DELAY 205*Q:DELAY 205*Q:DELAY 205*Q:DELAY 205*Q: DELAY 210*Q:RETURN 222 DELAY 257*Q:DELAY 257*QzDELAY 257*Q:DELAY 261*Q:RETURN 224 DELAY 257*Q:DELAY 257*Q:DELAY 257*Q:DELAY 262*Q:RETURN 226 DELAY 257*Q:DELAY 257*QzDELAY 257*Q:DELAY 263*Q:RETURN 228 DELAY 340*QzDELAY 340*Q:DELAY 352*Q:RETURN 230 DELAY 340*Q:DELAY 340*Q:DELAY 354*Q:RETURN 232 DELAY 520*Q:DELAY 512*QzRETURN 243 DELAY 520*QzDELAY 515*Q:RETURN 236 DELAY 1032*Q:RETURN The Analog-to-Digital Conversion Subroutine 150 OUT 219,0:OUT 218,0:OUT 217,0 151 L(I)=FIN(221):M(I)=FIN(220) 153 IF I+I+1<=100 GO TO 150 156 1:1 159 M(I)=M(I)&15:M(I)=FSH(M(I),8):L(I)=4095-(L(I)+M(I)) 162 IF I+I+1<=100 GO TO 159 165 J=l:K=1 168 s(K)=(L(J)+L(J+l)+L(J+2)+L(J+3)+L(J+4))/5 171 IF FMOD(L(J)+L(J+1)+L(J+2)+L(J+3)+L(J+4),5)>=3 G(K)=G(K)+1 174 J=J+5:IF K+K+l<=20 GO TO 168 177 L=le=l 180 A(L)=(G(M)+G(M+1)+G(M+2)+G(M+3)+G(M+4)l/5 183 IF FMOD(G(M)+G(M+1)+G(M+2)+G(M+3)+G(M+4),5)>=3 A(L)=A(L)+l 186 M=M+5zIF L+L+l<=4 GO TO 180 189 I=lzRETURN ___ __,-, . .“._,_ . . :10 3 L a 8 Z A .3.”_m.:.._HZOZ >33 maze: 3H._...,.H_G_.C :3. .m. G .n. ._. H.”— O .m— ..3. ”H. O a. . D .._H Z 3 m .p. Gm. a 2 Ga H I a M.” Z 176 S... omom m2... #5.; H”. m G .1 5 Z .3“ . am: Ge-.. 233i m5 3% ma” ....m...._n_ru.3 2.3.; 2 m; H.” mom .1 a...” ”E._.UGEG_..§ m .1 m._ .Z. U G .1 G I U m “a r m G ...._ tr 2. m. .1 m_ E o m O m 3.....H.._..m_3._.w.....~. Ext... .3 ..:H._H_..x..w:w G mwG amw._...,.q_G_.. 53—. I... .._H 3 Emir”;35m. m.“ ._. .1 G 3 ._. 3 G .._. 3 CK 5.312;” m 3 m. G .Z. m. .._. a on 5. 3L... 3 Q\ h. 3.”. Z H a 3 ._. G ._. m a . .._. ...n_ m. i .3 G H... a m a. a .._. H”. G q 0 Z . O m a G C a. Q .h ._ >EGZH m ._ m. I .Z. Mo. .h. u G .b. H 20503:: G ZG mm m. N H s. G H .h. N.” Z w.” a. .TdGD FZGD mmxomom: .._. 1G3 .G. E ._. m M.“ Z m I m. G .3 ”_.. m. Z... P in ._.3....._.~.33 .. m N .Z. 1G: :35 a ._H ._H ._H ._H co ._H H a G IOOUN. “_U D Z 6 v.— O m_ m \ O Mw O m m I mw m— 0 I A» ”m 0 I a L 0 I a... .u. 0 3 8 m. 3 a m. 3 8 m. 3 a m. .... z o \ o m o a a. .. ._. a n. \. o a O m s m~u_m._.MZGzGL >4mzmmmG .ZGNZGmGz . .r ”Eta: min.“ O._.ZH_.. m_...;.n_ a m a m a. m Z G .m. .Z. Z m 2 .r m H.“ JumEZH.“ ZG FGI... EZHFGQHQZH m 3;. _HZG mm .>G..._n~mH_H\.HH.,.._GQm>mz tum G _HZG ._. G I .h. I u 3 m m_ a O E _. ._. 2 m. z G .u_ m z G .m. .3. _. _QG-._.H_ EU H I3 ZGHHEFEL G m.“ ”_H oz . omzzafi. Z G _H > m Z m_ .h. ._. H .13 m I m. i. O .3.._ .P. 7. O a m m L 2 .._. m I .r 2 .m. m_ 3 .~. m_ m oz. omzzH- £9. ‘9. (9. 3s N ._. w. G 3 m m... P .m— G 3 3H ._. a. G 3 m .n .p. .m. G 3 a 4 4’. 19. 9 "U 45 43s 49» 49. _H G D .3 (9'. «Is 4’. ‘3- (D. ‘3- 4:. (9. m. m 2 G 4:. 49. G Z ”_H 43. d <9. <3. ‘3'- ‘3? mi. _H .n_ U m. m” NOON. .fioem ooom. comm. mwmoo 0 m0 0 MWLOC «VLOO 177 A H o m a s_- 1 1.M..G1 11-11111 .. .1..3.1GZH3...1.. 11.1....1 .1 1.1_ ....1G1a._11m1 _H 3.11.111 11:11.”... .11 H): .3 ...1 >112 ......1..._r..__..1 Z 1 ..H 11.12 H E. 1....“ H Z G B 1131.1111G mG .1..1G1.1 112.1111 1.1-1.111. mm.13.1G_..1 2H 111.11 111.1 .11..1.M.1G.1G1.._M.1 . 11 Z 1 1.111.111.1111 .121. 111.1...1.1311..11.1 1.1.1. .3111va 1M.....1> 1131111.”1.111.141.1111 11.16.111.111 1.1.. M31. Im1¢w1N. N3 1.1 .120 <3 1% I-I.V ‘ 1-1.1 1G...1G1..1..1 1.1Gn1Mwm 3H\-._.1.1.m1 111.1 ._.H...:..G.3 amcm .31 H311 1 11.1“: .1. 1.1.3.1 ...1 ZH .1. 1.. ...13.._1.1.1M..1.. G .311 3111.1 .1. H 1.1.1....1ZH 3%...1GM11 .111- ...1G.3..1G1.._...11 .1.. 1.7.1613 ZH .1..1G1.1 1111.11.11 11111.10 1.1.111». 1111.11G1G11n1 112”. 'v‘ 11111121“ 1.11.. 1. H H1 > 111G 11.1 1-1) H 21.11131 111.1 .1sz 1 .... HZG a; 2.11.13 M11111. 1.1.1.1 .1..1G1.1 1.1.11.1 1.11-1.1. m1 1.1.1G3 «3 M1 H .1..1G1.1 Z H 11.1113 _HZGZZOU .H _..1...1..1.1.1_111 M11... .1..1G1.1 .1.—111 31.1. 1.. H H 00 H 00 .3 G 1.51.1 . {111.1 Z _ M .1. .1 “.1 Z H .1.. {.1N: ...11.n.......1\... ”V... A... GHU\. a... .H ..........1N.. A”... m... M.1\ .. ... .. ...? \. enfi \ In .eOe omun ”Hum 11.1% QHUA uHon mHom GHom oHum mHUA AHQA QHUA mHoh VHOA MHQN NHUA HHUA oHun 100\ 101A moon 000A 100% Goon moon moon noun coon L3o\ coon m 0 M1 .1.. 178 11>Q£G ”111.1. G..11.11.11.1M1M1G 1.1.1. 1.11.1.1. 1.. Z111>1 1.1.1. .1. .11 G .1. M1 ..1 .3.1 .1 va .H11n1 n11. .r11511.a1m11 FZLZH ED; 211101 11m1 111. .r11nrr11m11 a Z1..GM.1G 11111.11 1 1.11» 1. .1M1H1M1G 1.1.1. >111G11~1 111 >2G1: 11H; >1GZHm F11 :GO1 EH71 C» 1111.1.111u.~11u1:1m1 1.111.? 11G.3.1G1..1M1 13111.1 .1 >1H1Gm1.3..1 FGZM :MFFHszGzF :31 szzm 1:21 1.11.11.11.11 1.11.5.1 1M111>1 92G11 +H 1H1 11Z1ZH an E HbzG: +30 11G DD: .HZU1H:1 N? H HZG mezGD ZH %zqu 121 mu.1.~11111 .1.1111 11G >8: m1M1H1.1.s111 111; H HZG mmbz13 ZH bzu1az1.127 119¢mrn N1 1 121 1.1.1111-1.a111 ”umzkzo .1 U SR 30 MU M11 .11 NE N9 N: Uh mm GU HO OH 01 mi UR $0 MU QM MQ 9R 0% ' J4 ... 9s GU Ho 01 91 mm mm mm mo 1A a? G0 mm m¢um NGQA e¢om mcum ¢¢UN mvom N¢M1A H¢om ovum amok 1&15 anon anon mmom Gmom anon anon hmom QMOA mmom ¢MUA mmum m... ”an Hmum omok anon mmom anon ”mon amen Gwen amok mom ;mom 179 maMFMZGZGi >szammG .DHZD MFGFm ZGeHIQHZ Gfibim: >2Fmazmzu .ZGNEGmGz 2G: >m ZMF+HEB .QZGZZOQ a EIF SZHm: omom MZF 20 2:: mm 2&2? ZGU ZGaeczi MIF .EQGZ mH KOFHZCZ Ob ZZDFHK G 22G awkzfizi ma ** G agjummmuusm mH EGO; G 1H .OQZZHJ Db wzzzbmz :2G a? meHxi ZGZGOEG MI? Azzmzouzn OEMN:ZGZV misuse aozzm FDGZH 2G LH .mm MI? 2822 aubamomz ma moaGmJ Eczuv G 41km zen HaomG MZF ZMIB OGZZHJ 20:; QMKMHZM ma EMQGOJ MIF .mZMFmHSMK Kc >zozm2 omom 2H MGGKOHW as; 5233 me m 3.... MJGZHW G OFZH mezbzm HHumG 03F mqumzou QZG HHQmG Nib >G3G fiinkm EMQGOJ Elk .Qmom MIF 2H mmmfimziaG .DZG .mZOHFUDKFmZH “GFG: bzmmmzimz Ob im$3 wise JGZHQMQGXMI Kan mmgau HHQmG >44sz MiG mMJHL =>ZG2Hm= MI? Azm.m;Hn m.fiGm4»>m: vuibiv .GHG m\ms osz: 060% MI? 5% 233: >44GHzmm amaiHIm sz mMJHL 2m. NIP .Aoz.v mMJHL muzsom MBGDGZGJ >4mzwmmG omom 20 AomomGev qumzmmmG amaze G asz: ma MI? 20 akasz MEG mMJHG 2m. m2+ .EMFDGZOUHZHZ um um: G GZHmD omomQQG 2G asz 22m.u MAM; =>2G2Hm= FGZESG JMbZH 2G BZHQGQJ is; ZGESCKG G mH 92.0mQGO; Q2.0m:ng ‘9 (A g}, 4:. 4a €9- 9. t. 4:. ‘,~ (3. ¢ 4:. ‘3, c). <9. ‘D- 4:. 43 4:. 4% 4:. 4:. 4,- 45, 1.5 4" (9. ‘55 ‘9. <3. 49- {3. ‘3 4,. :0 It 180 222222.22 22.22.2222 .....2 22” 22222.22 ...2 22:. ”.2 G .2 G I “.2 (3. G32 2.22 «H.222 moon ..K. .2.” «can. 2.22.2 2.. 22 UK. 22:22:22.2”. 2.2.2. Om. .22....2._.>2...2 22.2222 2222 222222.22 N... GU 02.32% 1’. 22222322232... 3.2. 2.22.2 22.22:“.2 H2._.>2.2.2H_.. ..2..2G..2 2.22 92.12% .8; H2 .\.. H. .2:.. “.2 ...2. 2G m_wux :GD22H N21 NU Gfiom. 2 220 mm $2JCA. Gm Swan 2 H32 22 “non mm 070\ ZMFUGKGIG HDGZH 2>Dszm N*iG: 2H ma 732% .H— .n._ a an vnom .2 2 I .2. 222G 2.22.2 2 PM; .22 2 2.2222322 G 2:2“ mm 2.. (By 4’. 222.250 2.222.... ..F. .2. ..2 22G. Baum... 9'. .22.. when 2.2. GoG 2......2In2 2.22m; 22.2.2.2.2G22GIU 2.222 :GO..22.2H N... GO .....21.K tho.— v we .nm 0\ FHm >:G22 ZMDHMUMZ 28.2 2m G22 m HZG c2 omen $2 GGQA mNPEG: 2H “:GO2ZH ma 2¢om .2223 222 72227:22 2m G22..»2mm .2I.2m an. .HGUA 2h UGUN I22022 H32 mm mvom 2A GGUA I252: H32 22QG32 Gm a¢un 32.032222 22F2G >22FGH2222H :2QGC2 um 3H2 IGGQA Gas a¢oh (3. '29. 4’. <9. 2222.228: 2.2.. 22.2. 2222222322.. .2. 22. 2.2.2222 .2.. 22.22.2222. ...22 _2G.22_m .282 222.2 GLMO .2220. ..2 2.3.2.»2.252.173.2222“ LEG: 22222220220222 222222.20 2.22.22 2.222.225: 22200 .2. .2 2.2 .222 .u 2. 2 .2. G .2. 2w .2. .2 G 3 m2 2.2 \ O 222 o m ..2 I 2» 2 o 22 a2 .... 2.2 2.... 2. ..2 G 2 2 222 .2 O o a It. 181 2.2.2.2 22.2.2. 2H 22222 G 222.222.222H2 2b>m mmMZDQG 382m mmmzaaG QZHFZGFm LO 2b>m wmmzfiaG IGH22 22222.22 .2.. 22222.22 0 22.. >...2...2G..2. 2.222622%; 2P>22H 22G0 WFZG: 2H GEa N1 N HZG mNFzG: 2H NFKG: 2H ¢i§ N1 m HZG mNFZG: 2H > -2 ..2 G .2.. -.2 ..2 G 0 G22 be: 2F>EZH 22G0 >sfitG» -fi2Gm2 G272 Dm:2 .2 .2. 2. m2 2 .2“ ...2 ..2 G 0 G20 :32 “ZMXZH 0h 0$ D0 mm mi 0 N. 0% G0 NO Q2 02 ma mm ED 05 NA G0 NO 62 0m mi AG Q0 20 0% 0% D0 0R AG Q0 A0 0% 09 Q0 u¢ A$0A @m0h @005 €00N mm0fi N003 Hw0fi 000A 250% 250A QAON 050A mNOK GNOA $AOR 0N0 m. hh0h QRON EBUA ¢A0fi 050R NRON HAOA 030A 200% 200% scum 000% mfl0h €fl0fi 060m QOUA 560% 000% @003 V605 182 zzmzomzo G we mbzmm 5+ 23400 Iocm zmuBFMmm FDizH Hmec 44¢ aszflq mimmz MZHbsozmzm mHIFm zwkmfiema : MI?“ 2H %H matchm :2€ QEOB FHm m MZO Zia; OFQ amMJmmazg FHm ESQ; 03H mafia mszzaimsm mHIba mmvgamb.s_m_~&qugmw ~U.P .umu .mu‘fiuugummm b s.n&74H m QQOJZH be ZHGGG ZHemm .za Esmzsmzom Z...mv_u.....m_.+u Fame 5 Gym—mu.“ my” .rzzaom ”my—CZ ._.m_e .. OKEN #32 m; .rzzcu ...: a ‘ F2330 aubuqaczu Fzmzmzomam EMFZHG; >Kszmz Fzmzuzqum A4,: FQV >zczmz 2H Eczeczi as maqubm Zfifimmm Fug Qau >0: o :aq km: a a a c HHomq 44cc MFPGFOW_.%4€Q Haumc 44cm aszam it? a¢eng N7 0 QQG MF>mZH 44cc ZMZZH N21 maq DD: m is: I XZH > 2. 4 c .h. ..._ ;_ a. 0 Ga: >02 ”>JJGH “MF>QZH 00 g¢ Hm 00 mm on ¢€ :0 on 00 so an «G am an co me on m¢ co gm as 00 :0 GA am me an 3 mm on AG :0 AA OGUA aqua AGUR QGQA QGUA ¢¢UA M€UA NQUN HGUA 0¢QR ...._ m. U A. maofi Q90m QQQA mach caun oaoh moon Baum coon mask ¢0UA moon m,“ my C \ flack 090% AQUA mach :mom 0 m U m. mmun Gmum QQUR mmun 183 :I03 FIG 28 @& MZHHIGI HIGHIO :DHHZGI mwmzw unu__3..._.._..G.n_ Ismzumfizfiu as; MSGSM; ...:L iszgm: IIHmHemz : ZH HH EIGHmM HHm a ZG EKG: OH IJIQHZ amaze 333 G 5% QMQQGM mm ZGU HH HGIH om H.G:3 IIIH O.I. mmIHM. EIGGM ....w3mMHHzg .MHm ”233.“. G M....;.:.GHOI .HIZHIMDOIMHIMw MwH_..:.M :IIm HZH.G fiMTG LOG >IGZHm 3. .M.3:mm.w_m In. so XII A >G3G iHIHm LEG mGB HHUmGM imImHzHu aaso mammzzz ECG >IGZHm AuH HIGVV HHm >IIGU m.Hmm M.,_M.Mw. IMG M ..33MZWHE «Mic 3G3 HHM..M .,G GHM IMHmHem: m ZH G XMIW XII 0m >G3G .IHIG.3 ¢ H3fiwwG PH: \ IHGwIU .m me szzM IIHQGIGIU G HDLZH .>QGIIM ZHGGG ZUMIH 3Hm> Hozm HHm >QGII ZMDHIUII ECG ZQIIQM v_UMIm3 3:..G:_m HIGIM MWHZIQ%GM \&\.u H): Hfiu.33GU \axao H): G a _H be: A HHO m+é $3 a £20 IGo‘m H32 Ion Hsm IGA HZG NHVG: ZH HHU.MwG NM. N HZG mmHsz 2H Mazuszm €3.2MGHM.:I “HHUmG LN m0 MO £1 :0 1m MIHV 90 m5 KO AC NO 50 96 A0 6Q UR LE GE mm GO MH OM QQ LR Gm mm m& UR €€ GU fiO Gm am Mi 000% m U Q N. Goon auum much A. u U x, 003% moon Goon much moon HOUR 000A LEON NEON. DECK Baum amok Gmoh amH3N amok Amok emo\ mmox vmun mmofi. NmUA .HMH .3m omen LGDR MGUA :Gom QGUN; mGom GGUA 184 no 0105 GG Gmou so 33Go no mass Go amen :MMG mzH3M IGOMQ H): we each no mmom GG GMQA so 33cc so amofi :0 amok ZGSHm: IEGHGGGUM 120.3 H3: mo muck no omen GG muon so 33cc :9 ¢MUA Gm mmcm \*\33 H3: we muck mo Hmon Gm omum Ham zo ** mHszim so 33cc mo Gian -m mach :Gs3 33Gmmmu03m as; meGmmmz MmHzHGM \*\mu H3: “ImHzHG mo anon on anon mo mnom szHGSG >sz was: szmGGQZGGH 0H mzzszzm :woua G21 mo Gaom mo amen Gm @505 so 33Go :u AQUA Go 020m :mmu mzH3M IGo.o H3: mo mack no vaum GG maom so 33Gu :9 much no Hawk zzszL IGGHGGGQM 220.3 H3: mo 023% no Loom GG moon :m3.33Gm3 _Mo Hfiuum 185 Z G a m. o I n. m I .H. _H 2 G 2 I I I, m .3“. I m. G H 3.... I G u G G, I _H m— 3 O -3 .3 O .3. m H. I Q G m. . H 2 H I m. I .H M 26 IGKGQGG HH ZGIHIOG QZHHHGB G 5% :mmwGG ZMIH MIG mimmzzz mmIIHW .mHio mammzzz GOG mice HHumG IIF DHZH w3mmHz HHm ¢ IQGI mHIIDZOUm QZG %GOIMI omomiaiG IOIG GHG: mIH i3 mzuHi MIGHGQ .mmGHGQ OHM mGHGG HZHOG GHG: 3GDHQG.LO H2300 MIH mmGL DH mum: mH :ZGIIOU :x MIHW .IQHHZDZ ZH :ZGIZOU Ix MIH GHD mmGHG: OH :ImmGG mH mmmGQQG mHIHW GO m23G> XII HIP .m>GGIG 03H mmMIH no HGGHW NIP GO mmmzanG mzHIUGZ. I. g JGXHUMEGXMI MIH ZHGHmO OH amm: II» @3533G UHme >ZHH 2H :ZGZIOU :GLNA MIH .>Iozmz ZH >3wIGBGHHZGQ QIHGUOJ ‘03 HG :monmzmzH: IQGM m>GGIG (a 45 4. 33% no mthmzco >33G3HUG GHGQ MIH ¢>GOZII IZOD OHZH PH mIHHGB 49- IUHII omomiaaG MIH zo ZGIGOIG UHme >ZHH G ZOGL mmHGZHeHGO GHGQ IIH .zzH3 3GHGmm G GHD Mum owiv GIHDGIOQHZHI G OH >zozmz omomsafiG IOGG GHG: @ZszmeZGzH GOG ZGzeomG MSGDSZG3 >3mzmmmG omom ZG mH mmGHGQ <3. ‘9. £9. (a 1,. 4:. <5. 49. QI.MQGHGQH sz I _H a I .M. 2 HI I G G MW 2 G I .H o .H z .3”. 3 .M. I I M I MW 0 u A a. Z M; on we no MLUR NLQK H .u_ 0 m. 186 am I..IT..I ,IG. .IIIZGIII. Z I.. II ...:IIAI ..IIIILI G III GFI I ..IIGI I. ...IGwU II mtInIII II III I I I ”IS 3 ...I Z G I I I “...I I I m. I. I G _.I IAN...” I G _HZIIII IQCII LIAISTIG: ..MIII..I,.. ..IIIILIG .I.~IG.I.mI ...I...:IU -. mHIG: INH...I.IHG.I:IHZH.I IzIno.G I o o ..I N. IGGmmII IIHUGIGI..IMI mm. IZ HFIIIIIIEIm I.D.NIIIIO III 12.3 . .TIOI O\H FIG: .CIMI\omIomm FISI IWIJIIGILW .I..,IGII HHIU\OIOI.II III .I. .I.. 28 I Z I” m... III. I I G”... >I3I..m~I....I>HZII IFGIII ZGmIIHIxIUIII ...I..NII: >~II.m.IIIII.I...I .vIGNyInIvaI 2G: .ZIIH.w..II>ZO.I _HI\G I...:.I NIH G IGII I...I_.IIH.II~I III—I: IGD GIIGQ ...IIMWII IIG III m... I” >420 . Ik>m 332. “III I £722» A m . I..I.>mI 3II...I I.I.>m IIIHII “III I mix “I “HHIIIIIQ III. 2H 1P III II G IIIIIIZ IIIIIDMI I.I.>IH.I 3III..I .....I_I.,.II.I3 III I III. I. mJIImIIIZ A..I,.I:IA.. qIHIHmIGm I.Z.I...I. . a 7. H I Z H » Z O ”I I ..II II .I I I I I G I I I .I. .I II ...I I ”I I Z G I a ...I ”I I I. ”I G I G I ”I I ...I G .m I“ DI I. I II I“ DI I. 3 ...I I” D I I .«I III I m. a I a .u ImIIo IIGI IaIO DQI ‘3» a ( «I m I ...I I G ..I m... o I I I a QI >..ImZmLm¢ >m ...I ...I G I I .I. I I Z .I.. «I .I I .I. ...I G ...I ... I G ..I ....I II mewIVCH ...: MU ZII. I. I” ”I3 IZII ..I. .I.. G 3 ‘ ‘35 4:, 4:. 49, 13. u N .r .I G I 3m .I. ...IGI I I £9. ‘9. «as 4’- 43» ‘9. 4:. 49. ‘3 ‘9. ’. D. IOIA GOIIA acim moan KOQR GOQA ao:m Gonfi moan moan HOQA OOGA Conn «Loo GuImo mmIoo aIOO 187 IIICH 30:. IO ,.:mw..I¢ :..II. .I..:z: IF>I 334 OF %ZHIII IHEH I»... H I I0 I ~. I M3? .I.I..:.II._:: Ik>m IDI.I C.I. FZHOI i IIIAZCII III: ZHIIII kc IMIIZI.D.I IUCI.L. IIFZI D» HIGmeII 5.0 5 Cad QZGIZOUIX IFHB wIII:IZ 23¢ Dc: I.I.;Iwc :fiIGQ I .I. a. .I. G I ...I ..._ G I IOIO qu ZaG DO: I XUQ HH I II :figch IIO HZ¢ III >02 I XZH I0. I H): Im cam H3: IOIIA.II Hx; H III IZHII N21 I II: I XZH II IIGU Emu >32 ”FIGII MIHZH “IZHII .1 .1 A. s a... DO IO OI IR m% 00 m NO 00 In OI Hm IU SK :0 .dq \; n.~ I. 3 MN MO :I I? 0H Iain INJN :3 Q\ I.wQ\ IOQK QNQN Imam Imih mush GWEN {IN—AN QNEA .,ND\ .HN:\ .HNQN ONDA IHQR m— H _H N. GHQR UHGN mHDA GHQI OHQA IHGK RHDR @HDN D.HQA VHQN ”HER WHEN HHDA CHEN IOQA IOQK Donn 002m 188 $202 muzs .rzszgu Hmwzuhmflwflm auszei 3&2 : 4;: Ch iazmfiwacfi: fifiém mZQHF¢QGJ >Hizm bQMimzm ifizmfimm MZHZEMFMQW fi+ onbcusqgc ZOKL H2200 ub>m Foczbmsmm U 2H Aom * ma+>m my ..m_ a a Z 1.. 2 m z a. i m 2 G .2 .2. O .2. Z z 3 ._. m .m. >€ific Gbci Fxmz bum #2280 wbcmm oofi as 2.233”: szmru ,figcm mazes“ ago: Dam EMFZDCQ FZMZMEQMQQ HmF>m 30s_ 3&2 3+ %Zaeim 3 m a. _. mm 3 .m— FEE EGE ZGK KG: Efii FHZH it? was be: : :ca Gam )5: m mzw m ism $20 30: I¢Qa€ H): I m 0 U m. N1 22¢ 30: m a u _H kzc+m N21 25¢ >3: 2 as: w zzH I sz : sz Hfiumfi 44cc :20 Hzc ”HHQmG ”m+€+:£ «....— ca— cII not as m, w :5 ...... ...I A; h. cm 20 cm mama fimam omam AVQR Evin :¢:m mean mvmm cqnn O¢QN qum nvnh cqzm mcmm ¢¢QA m¢2m mcsm ficak ocam mmaa mmam swam owns mmam ammn 0min QMQA mmzm emaa mmam ¢mcm mmah mmzm fimax Omflfi 189 iiG mu.g.Av ESL H a o zzzhmz mefifizzqum wifmZa:2 ggfgbszum £32 Haom¢ MFGMEQQ GZHQ€MJ 2wzfibuw_ «. Fsr¢#s:uQ me€szcom EMEZ_._Z #31228“ is; ”_H H.” sac 1:. cmzu m 3 F300 JJQU 2:029 HDZ .PDDMH.€;€Q Qmu >82 26% HQQ 3m; i3; #380 AJGU Ifimmu H32 FEE .~.a~aunu 22:2flgnu ID 0 a U H 32 .P_4~UMU zdgfinyng GaU )8: Ion HQG 3mxm .mnr; :4i2Mk U21 lazy Hie ”mJiZMP cm cm H; an AA :0 fin mo ow an AA :0 :0 mo in cman mama ¢mzm mans mmcm fixzm cans Loam mean scam wean mean coma acnm mean mean comm mean vein mesa mean Honk ocam 2min mmam ammm 0min main «man amfifi Mann mack emzm mama emam mama 190 \Q\ m: x \ ma \x\ m: \ s. M“ _H \m\ m: Wax m: .... x m : _2va mm~ 2:0 m: was: 0: 223 323mm Fm: ZHGGG zumzu §mzqw MIH F02“ Koala N21 Fzmm EMFQGEQIU wI% IkHB macizoum a £20 HHnmq HHm A Mbcmzom ILA H26 w. x_. ”NE 3 2 H 2Um:fiu N? I N O ...H 2 G GZHZKDFME mzoumm 820m :8; zumzum mmkzqs 2H NFZQD #30 omfi DB: #353 N7 EHO H26 m m; .16.: 2 H 2.242UHM “3m: ”zomzu d M? ON mm 0% 5? MB BE GO :0 98 DA aw HO 0% LR 0m mm mi 5% fim GU NO Gm 0m m: mm Mi an 2% NR GO fiO QM Qaim mfifim Nfiflm fi$§h mgfih %QQN m9fim NGQA HQQR O$QA .nzwfik. mmfih Dmih Umfifi mmfik Gmfim 0min QQDA AQQA owns mean ¢mfih M$ifi mach HQDA OQQA LKQA whim QAQR DADA ENDS ¢AQA mu «MGR. GNQR RAGE 191 a? UN ¢¢ om mm A? W 1? ; AR M...— 0m '9‘ ?m L? €¢ ON ¢¢ 3 aqua mqam scam ecam mean ¢¢an mqam mcam fiqnm ocmn Lafifi moan imam mama maam qoam REFERENCES lo. 11. 12. 13. 192 REFERENCES G. G. Guilbault, Anal. Chem. 8, 527R (1966). ..._— —— G. G. Guilbault, Anal. Chem. 40, 459R (1968). G. G. Guilbault, Anal. Chem. 42, 335R (1970). _- fl M. M. Fishman and H. F. Schiff, Anal. Chem. 44, 543R (1972). —_’ M. M. Fishman and H. F. Schiff, Anal. Chem.-4§, 367R (1974) . ‘— M. M. Fishman and H. F. Schiff, Anal. Chem. 48, 332R (1976). M. M. Fishman, Anal. Chem. ég, 261R (1978). H. H. Weetal, Anal. Chem. fig, 602A (1974). L. D. Bowers and P. W. Carr, Anal. Chem. 4g, 544A (1976). _— D. N. Gray, M. H. Keyes, and B. Watson, Anal. Chem. 49, 1067A (1977). O. Zaborsky, Immobilized Enzymes, CRC Press, Cleve- land, OH, 1973. R. A. Messing, Immobilized Enzymes for Industrial Reactors, Academic Press, New York, NY, 1975. J. M. Nelson and E. G. Griffin, J. Amer. Chem. Soc. 3g, 1109 (1916). 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 193 N. Grubhofer and L. Schleith, Naturwissenschaften 40, 508 (1953). E. K. Bauman, L. H. Goodson, G. G. Guilbault, and D. N. Kramer, Anal. Chem. 3;, 1378 (1965). G. P. Hicks and S. J. Updike, Anal. Chem. 38. 726 (1966). L. D. Bowers, L. M. Canning, C. N. Sayers, P. W. Carr, Clin. Chem. 32, 1314 (1976). L. D. Bowers and P. W. Carr, Clin. Chem. ii, 1427 (1976). T. Kawashima and G. A. RechnitZ, Anal. Chim. Acta 3;, 9 (1976). J. N. Miller, B. F. Rocks, and D. T. Burns, Anal. Chim. Acta 2;, 353 (1977). S. W. Kiang, J. W. Kuan, S. S. Kuan, and G. G. Guil— bault, Clin. Chem. 22, 1378 (1976). . Huang, S. S. Kaun, and G. G. Guilbault, Clin. Chem. 3, 671 (1977). H 2 . Morris, J. Campbell, and W. E. Hornby, Biochem. 4 , 593 (1975). J. W. Kuan, S. S. Kaun, and G. G. Guilbault, Clin. Chem. ii, 1058 (1977). P. J. Taylor, E. Kmetec, and J. M. Johnson, Anal. Chem. :2, 789 (1977). L. P. Leon, M. Samsur, L. R. Snyder, C. Horvath, Clin. Chem. 23, 1556 (1977). . G. Guilbault and G. T. Lubrano, Anal. Chim. Acta 9, 189 (1974). W. E. Hornby, D. J. Inman, and A. McDonald, FEBS Letters 23, 114 (1972). 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 194 R. E. Adams and P. W. Carr, Anal. Chem. 99, 944 (1978). _— H. Filippusson, W. E. Hornby, and A. McDonald, FEBS Letters 99, 291 (1972). R. S. Schifreen, D. A. Hanna, L. D. Bowers, and P. W. Carr, Anal. Chem. 49, 1929 (1977). D. T. Bostick and D. M. Hercules, Anal. Chem. 49, 447 (1975). _- G. A. Rechnitz, Chem. Eng. News, p. 27, Jan. 27 (1975). S. J. Updike and G. P. Hicks, Nature 214, 986 (1967). G. G. Guilbault and J. G. Montalvo, J. Amer. Chem. Soc. 99, 2533 (1970). G. G. Guilbault and W. Stockbro, Anal. Chim. Acta 19, 237 (1975). Technicon Instruments Corporation, Tarrytown, NY. M. K. Weibel, W. Dritschilo, H. J. Bright, and A. E. Humphrey, Anal. Biochem. 99, 402 (1973). C. Horvath and B. A. Solomon, Biotech. Bioeng. 14, 885 (1972). _— W. E. Hornby, H. Filippusson, and A. McDonald, FEBS Letters 9, 8 (1970). P. V. Sundaram and W. E. Hornby, FEBS Letters 49, 325 (1970). _— L. P. Leon, M. Sansur, L. R. Snyder, and C. Horvath, Clin. Chem. 99, 1556 (1977). P. V. Sundaram, M. P. Igloi, R. Wasserman, W. Hinsch, and K. J. Knoke, Clin. Chem. 94, 234 (1978). 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 195 D. Sanderson, J. A. Bittikofer, and H. L. Pardue, Anal. Chem. 44, 1934 (1972). M. D. Joseph, D. J. Kasprzak, and S. R. Crouch, Clin. Chem. 99, 1033 (1977). M. D. Joseph, Ph.D. Dissertation, Michigan State University, 1978. GCA McPherson Corporation, Acton, MA. Product # PG-350-200 specification label, Lot 76C- 0013, Sigma Chemical Co., St. Louis, MO. R. A. Messing and H. H. Weetal, United States Patent No. 3,519,538 (1970). J. D. Ingle and S. R. Crouch, Anal. Chem. 49, 697 (1971). _— D. Klielen and E. F. Hartree, Biochem. J. 99, 331 (1952). 7‘ P. R. Bevington, Data Reduction and Error Analysis for the Physical Sciences, McGraw-Hill, New York, NY, 1969, p. 104. J. N. Miller, B. F. Rocks, and D. T. Burns, Anal. Chim. Acta 99, 93 (1976). L. D. Bowers and P. W. Carr, Biotech. Bioeng. 49 1331 (1976). _— Q. H. Gibson, B. E. Swoboda, and V. Massey, J. Biol. Chem. 230, 3927 (1964). J. Avery, Ph.D. Dissertation, University of Illinois, 1978. SDK-80 Monitor, Intel Corporation, Santa Clara, CA, (1975). Tiny BASIC Technical Manual, HAL Communications Corp., Urbana, IL, 1976. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 196 J. A. Titus, C. A. Titus, D. G. Larsen, and P. R. Rony, Amer. Lab., p. 101, June, 1977. J. Denker, Electronics, p. 110, March 17, 1977. Intel 8080 Assemblnyanguage Programming Manual, Intel Corporation, Santa Clara, CA, 1975. H. V. Malmstadt and T. P. Hadjiioannou, Anal. Chem. 94, 455 (1962). Worthington Enzyme Manual, Worthington Biochemical Corp., Freehold, NJ, 1978, p. 1. R. C. Weast, ed., Handbook of Chemistry and Physics, Chemical Rubber Co. Press, Cleveland, OH, 1969. Assay Data, Product No. N—7004, Lot No. 116C-7050, Sigma Chemical Co., St. Louis, MO, 1977. H. N. Christensen and G. A. Palmer, Enzyme Kinetics, W. B. Saunders Co., Philadelphia, PA, 1974. N. W. Tietz, ed., Fundamentals of Clinical Chemistry, W. B. Saunders Co., Philadelphia, PA, 1976, p. 245. Moni-Trol I Chemistry Control Comprehensive Assay List, Lot No. LTD—50, DADE Div., Amer. Hospital Supply Corp., Miami, FL, 1977. S. B. Tilden and M. B. Denton, CONVERS An Interpretive Compiler, Dept. of Chemistry, University of Arizona, Tucson, AR, 1978. BASIC/5, Processor Technology Corporation, Emeryville, CA, 1976. J. Avery and D. Lovse, The ADD-8080 Microprocessor Manual, Dept. of Chemistry, University of Illinois, Urbana, IL, 1977. G. Pickles, Kilobaud 9, p. 50, May, 1977. "WWW 83 H H N” I II All H H 0 3 0 3 9 2 1 3 IHI ”NIH