3.2;}

q
z.

.
.17

- 3

5w .

. 2.

. 3.“!
.:‘y

. 4 ﬁle...
3%.. . . Rpm“? .
2.. 7.4..” . .. . . ..4. , .... a
.... Earn. _, , . . . . . . summ‘ﬂwwm ..
. , . , 4m
. . . . . . gunman;

A. “:34
I.

4 ',o-\.
”gun...
" H‘.‘ " \

 

n,

 

.Ir 4:

4 v13...» U. 4

. I. . .. I.I..vl-l
.94 ....c\r.

r... 21.4 924.2! .
. z. .. ~ . . u: .
. a4 . . ,
. , , . . . . 4...........l..ﬁ , .. 4 :l.

 

YHESlS

44444444444444444 44 444444 4444 44444444444444

3 12930

1

This is to certify that the

thesis entitled

Immobilized Glucose Oxidase in Capillary
Flow Injection for the Determination
of Glucose

presented by
Yi Shi

has been accepted towards fulfillment
of the requirements for

M.S . degree in Chemistry

étanley R. Crouch

Major professor

 

 

 

Date July 17, 1998

 

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1“ WM“

IMMOBILIZED GLUCOSE OXIDASE IN
CAPILLARY FLOW INJECTION FOR THE _
DETERMINATION OF GLUCOSE

By

Yi Shi

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE
Department of chemistry

1998

ABSTRACT

IMMOBILIZED GLUCOSE OXIDASE IN CAPILLARY FLOW
INJECTION FOR DETERMINATION OF GLUCOSE

By

Yi Shi

The principles of capillary ﬂow injection (CFI) analysis and immobilized
enzymes were used to develop methodologies for the determination of glucose. Indirect
detection of glucose was accomplished by means of enzymatic reactions that yield
hydrogen peroxide. The enzyme reaction was coupled with the Trinder reaction, in which
hydrogen peroxide reacts to form a quinoneimine dye. Then, the dye product was
monitored spectrophotometn'cally. The basic components of the CPI system include a
peristaltic pump, a 4-port, two-position conical rotary valve, a glucose oxidase
immobilized capillary reactor, a mixing tee, and a ﬁlter colorimeter. Data acquisition was
controlled by a computer.

The initial immobilization procedure was modiﬁed in order to be applied in
capillaries. The optimized procedure was evaluated by its reproducibility, amount of
enzyme immobilized, lifetime, and conversion percentage of the reaction. Enzyme kinetic
studies were performed. System optimizations, such as capillary dimensions, ﬂow rates
and reaction time were investigated. The calibration curve of glucose was obtained and
its deviation from linearity compared with that of hydrogen peroxide. The linear range,
signal-to-noise ratio, detection limit, and other factors were calculated. Finally, CFI with

immobilized enzyme reactor was applied to real fruit juice and soft drink samples.

ACKNOWLEDGEMENTS

I am glad for the opportunity to ofﬁcially thank those who have contributed to this
achievement in my life. I would ﬁrst and foremost like to express my sincerest gratitude
to Dr. Stanley R. Crouch. I am thankful for his guidance, enthusiasm, encouragement,
and understanding throughout the course of my research effort. He gives me the freedom
to do what I want, but offers help when need, and built my conﬁdence and independence
in doing research.

I would also like to thank the other members of my committee: Dr. Simon Garrett,
Dr. John McCracken, and Dr. Chi K. Chang. A special thanks goes to Dr. Victoria
McGufﬁn for her excellent teaching in my ﬁrst year and advice in my literature seminar.

There have also been Crouch group members, who have helped along the way. I
owe thanks to Dana Spence, who introduced me to capillary ﬂow injection (CF 1)
analysis. I would like to thank Tom Cullen for his great help in using computer software;
Chunhong Peng for his good technical suggestions; and also Dan Severs for his
friendship in the laboratory. In addition, I have also made numerous friends, who have
made my stay all the more enjoyable. Qing Li, Jaycoda Major, Charles Ngowe, Lee
Kelepouris, Rui Lu, Lili Duan, Jian Li will always be remembered. There are so many
others: support staff Lisa D., my ﬁ'iend in China Jinghong Wei, my college friend Liping
Zhou, who have made my life abroad like home.

Last but not least, I am most grateful to my parents and my sister for always
believing in me, even from so far away. I am also deeply grateful to my best friend, Bing

Ren, for providing me a brand new and colorful world.

iii

TABLE OF CONTENTS

Chapter 1 INTRODUCTION .............................................................................................. l
1.1 Miniaturization of Analytical Systems ...................................................................... 1
1.2 Glucose Analysis in Clinical and Food Chemistry .................................................... 2

1.2.1 Glucose Analysis in Clinical Chemistry .............................................................. 2
1.2.2 Glucose Analysis in Food Chemistry .................................................................. 2
1.3 Project Goals ............................................................................................................ 2

Chapter 2 HISTORICAL BACKGROUND ....................................................................... 2

2.1 Flow Injection Analysis ............................................................................................. 2
2.1.1 Introduction of F low Injection Analysis .............................................................. 2
2.1.2 Principles of Flow Injection Analysis ................................................................. 2
2.1.3 Application of Flow Injection Analysis ............................................................... 2
2.1.4 Miniaturization of Flow Injection Analysis ......................................................... 2

2.2 Enzyme Immobilization ............................................................................................. 2
2.2.1 Introduction to Immobilized Enzymes ................................................................ 2
2.2.2 Methods of Enzyme Immobilization ................................................................... 2
2.2.3 Applications of Immobilized Enzymes ............................................................... 2

2.3 Glucose Determination .............................................................................................. 2
2.3.1 Non-speciﬁc Methods and Separation Methods .................................................. 2
2.3.2 Enzymatic Methods ............................................................................................. 2

2.4 Previous Work in Our Laboratory ............................................................................. 2
2.4.1 Capillary Flow Injection Development ............................................................... 2

iv

 

2.4.2 Glucose Oxidase Immobilization ......................................................................... 2

2.4.3 Glucose Determination ........................................................................................ 2
2.4.4 System Optimizations .......................................................................................... 2
Chapter 3 INSTRUMENTATION AND EXPERIMENTAL REAGENTS ....................... 2
3.1 F low Injection System ............................................................................................... 2
3.1.1 Capillary Flow Injection (CFI) System ............................................................... 2
3.1.2 Conventional Flow Injection System .................................................................. 2

3.2 Experimental Reagents .............................................................................................. 2
3.2.1 Reagents Used in Optimized Immobilization Procedure ..................................... 2
3.2.2 Reagents Used in the Analysis of Glucose and System Studies .......................... 2
Chapter 4 ENZYME IMMOBILIZATION ......................................................................... 2
4.1 General Enzyme Immobilization Procedure .............................................................. 2
4.2 Initial Immobilization Procedure ............................................................................... 2
4.3 Immobilization Optimization ..................................................................................... 2
4.3.1 Cleaning ............................................................................................................... 2
4.3.2 Etching ................................................................................................................. 2
4.3.3 Silylation .............................................................................................................. 2
4.3.4 Glutaraldehyde Activation ................................................................................... 2
4.3.5 Glucose Oxidase Immobilization ........................................................................ 2

4.4 Optimized Immobilization Procedure ........................................................................ 2
Chapter 5 ANALYZER MANIFOLD STUDIES AND SYSTEM OPTIMIZATION ....... 2
5.1 Immobilization Reproducibility ................................................................................. 2

5.2 Determination of the Amount of Enzyme Immobilized .............................................. 2

5.3 Lifetime of the Capillary Enzyme Reactor ................................................................ 2
5.4 Percentage of the Glucose Oxidase Reaction Completed .......................................... 2
5.5 The Lineweaver-Burk Plot and Michaelis Constant .................................................. 2
5.6 Dimensions of the Glucose Oxidase Reactor and the Trinder Reaction Reactor ...... 2
5.6.1 Dimensions of the Glucose Oxidase Reactor ...................................................... 2
5.6.2 Length of the Trinder Reactor ............................................................................. 2
5.7 Flow Rate Study ......................................................................................................... 2
5.8 Pass Number Study ................................................. 4 ................................................... 2
Chapter 6 CALIBRATION CURVES AND APPLICATIONS ......................................... 2
6.1 Glucose Calibration Curves ....................................................................................... 2
6.2 Deviation from Linearity ........................................................................................... 2

6.3 Linear Range, Signal-To-Noise Ratio, Detection Limit, Sample Consumption, and

Sample Throughput .................................................................................................... 2

6.4 An Alternative Method .............................................................................................. 2
6.5 Applications to Real Samples .................................................................................... 2
6.6 Concluding Remarks .................................................................................................. 2
Chapter 7 FUTURE PROJECTS ......................................................................................... 2
7.1 Increase of Enzyme Immobilization Efﬁciency on Capillary Walls ......................... 2
7.2 Alternative Enzyme Immobilization Methods and Reactor Conﬁgurations .............. 2
7.3 Determination of the Fraction of the Amount of Active Enzyme Immobilized ........ 2

7.4 Multichannel Parallel Sugar Determinations by Capillary Flow Injection Analysis. 2

7.5 The Use of Immobilized Enzyme Reactors with Electroosmotic F low ..................... 2

vi

 

LIST OF REFERENCES .................................................................................................... 2

vii

LIST OF TABLES

Table 4.1 Initial glucose immobilization procedure ............................................ 38
Table 4.2 Optimized glucose immobilization procedure ....................................... 47

Table 6.1 Comparison of capillary enzyme immobilized reactor with
SBSR ...................................................................................... 75

Table 6.2 Results of real samples ................................................................. 78

viii

LIST OF FIGURES

Figure 2.1 Effect of convection and diffusion on concentration proﬁles of analytes at the
detector: (a) no dispersion, (b) dispersion by convection, (c) dispersion by
convection and radial diffusion, and (d) dispersion by

diffusion .................................................................................. 8
Figure 2.2 Principal methods of immobilization ................................................ 12
Figure 2.3 Enzymatic pathways for glucose determination .................................... 18
Figure 2.4 Immobilized enzyme capillary reactor conﬁgurations ............................ 21
Figure 2.5 Glucose oxidase / Trinder reaction ................................................... 23

Figure 3.1 Schematic of CFI manifold for the determination of glucose with immobilized
enzyme reactor. The sample loading mode is shown
here ....................................................................................... 26

Figure 3.2 Schematic of CF] manifold for the determination of glucose without
immobilized enzyme reactor. The sample loading mode is shown
here ....................................................................................... 27

Figure 3.3 Conventional flow injection manifold for the determination of glucose with
immobilized enzyme reactor. The sample loading mode is shown

here ...................................................................................... 30

Figure 3.4 SBSR ..................................................................................... 31
Figure 4.1 General procedure for 3-APTS / glutaraldehyde immobilization of enzymes
onto glass supports ..................................................................... 37

Figure 4.2 Glass surface hydroxylation / dehydroxylation equilibrium ..................... 40
Figure 4.3 Common binding models for 3-APPTS ............................................. 43
Figure 4.4 Response signals obtained with the CPI system by injection 2 mM solutions of
glucose twice .......................................................................... 48

Figure 5.1 Enzyme immobilization reproducibility for glucose oxidase ..................... 51

Figure 5.2 Loss of glucose oxidase activity for routinely used capillary reactor and stored
reactor ................................................................................... 54

ix

Figure 5.3 Stop-ﬂow ﬂow injection analysis (FIA) for capillary immobilized enzyme

reactor ................................................................................... 56
Figure 5.4 Stop-ﬂow FIA for SBSR .............................................................. 57
Figure 5.5 Lineweaver—Burk plot ................................................................. 59
Figure 5.6 Effect of glucose oxidase reactor length ............................................ 61
Figure 5.7 Effect of glucose oxidase reactor diameter ......................................... 63
Figure 5 .8 Effect of Trinder reactor length ...................................................... 64
Figure 5.9 Effect of ﬂow rate ..................................................................... 66
Figure 5.10 Effect of pass number ............................................................... 67
Figure 6.1 Calibration curve for glucose ......................................................... 70
Figure 6.2 Calibration curve for hydrogen peroxide ........................................... 72
Figure 6.3 Measured baseline noise level ........................................................ 74
Figure 6.4 Calibration curve for glucose (alternative method) ................................ 77
Figure 7.1 Malachite green reaction .............................................................. 83
Figure 7.2 Enzymatic reaction schemes ........................................................... 85
Figure 7.3 Parallel, multichannel capillary ﬂow injection analyzer .......................... 86

 

Chapter 1

Introduction

1.1 Miniaturization of Analytical Systems

Miniaturization of analytical systems is becoming increasingly important as
demand continually rises for small-scale rapid determinations. In the past several years,
numerous systems have been developed based on this idea. In chromatography and
electrophoresis, capillary electrochromatographyl, capillary zone electrophoresisz, and
capillary gel electropphoresis3 are powerful techniques that have been utilized for many
different sample types, including a wide range of biological samples. Separation speed
and column efﬁciency can be signiﬁcantly improved in capillaries compared to
conventional media“. When probing dynamic chemical changes within ultra small
volume samples, a major technique today involves using miniaturized electrochemical
sensorss. Microelectrodes have been used to make in vivo electrochemical measurements
or to obtain spatially resolved electrochemical responses. More recently, the concept of a

miniaturized Total chemical Analysis System (uTAS) was proposed by Manz, et a16.

Microfabricated devices, for example, microchips, have been developed. The application
of photolithographic techniques7, using a standard double-sided single mask procedure, to
design and fabricate ﬂow manifolds on planar substrates, such as silicon and glass, allows
structures with micrometer dimensions to be fabricated. They can perform many sample
processing tasks, such as injection, separation, dilution, reagent mixing, and a variety of

chemical measurements.

In addition to these speciﬁc advantages of micro separators, micro sensors, and
micro reactors, general practical beneﬁts to be derived from such miniaturized systems
include the ability to analyze small sample volumess. The use of small sample quantities
can be advantageous in situations, where initial sample amounts are limited (e. g., infant
blood samples, biological micro-samples), where interlaboratory validation studies must
be performed (e. g., criminal cases), and where reagent costs are high (e. g., enzymatic
methods or immunoassays). The use of small reagent volumes not only lowers chemical
costs, but also results in increased speed of analysis, and a consequent reduction in waste
disposal. The reduction in device size also allows for increased ﬂexibility and can result
in instruments being located in the ﬁeld near the process or clinical site. This allows for
continuous analyte monitoring, which is of great beneﬁt in biomedical applications,
where elimination of dependence on external laboratory analyzers should have an
enormous impact on the way in which chemical and biochemical processes are
controlled, as the relevant chemical information may beprovided to the user in the form
of a continuous electronic readout, over the monitoring periodg.

Flow injection analysis (FIA) is an area in which miniaturization is especially
attractive. Capillary ﬂow injection (CF I) has been under study in our research group since
the early 1990510. The CF] method has been shown to have low dispersion, which can
lead to increased sensitivity in some sample determinations. In addition, it has
signiﬁcantly reduced sample consumption, and at the same time, keeps the simplicity and

efﬁciency associated with FIA.

1.2 Glucose Analysis in Clinical and Food Chemistry

1.2.1 Glucose Analysis in Clinical Chemistry

Continuous on—line monitoring of biochemical analytes, such as glucose in
critically ill patients, would be of great beneﬁt for therapeutic decisions and disease
prognosis. No goal has been more important than the realization of continuous blood
glucose monitoring as an aid to diabetes therapyl 1. The determination of oral (e. g. saliva,
surface of teeth) glucose concentration, after ingestion of foods, would also be of

beneﬁcial to teeth health 12’ '3.

1.2.2 Glucose Analysis in Food Chemistry

Food analysis is an extremely diverse area and presents a unique challenge to
analytical chemists due to the complex chemical composition and dynamic nature
inherent to various foodstuffs. As a result of society’s growing awareness of the
importance of health and nutrition, food industries and government ofﬁcials are
concerned with the nutrient content of goods and their subsequent labeling. An important
area within food analysis is that of carbohydrate determinations, such as the
determination of glucose. Determining carbohydrates is essential to many areas of food
science including nutrition, agriculture, formulation, preparation, and industrial
processing, from the raw materials used to the ﬁnal processed products. There is a
continuing demand for rapid, accurate, and precise analytical methods for carbohydrate

determination.

1.3 Project Goals

The primary goals of this research were to develop methodologies for the
determination of glucose with the combination of CFI and immobilized enzyme reactors.
The reusability of immobilized enzyme reactors leads to a signiﬁcant economic
advantage in many cases. And, the capillary immobilized enzyme reactors should be
advantageous over conventional open tubular reactors because of the small capillary
radii”.

In achieving these goals, the ﬁrst consideration was the immobilization of glucose
oxidase. The appropriate attachment procedure for immobilizing glucose oxidase on
fused silica surface of capillary wall was studied. Investigations were also conducted to
improve enzyme immobilization efficiencies over those initially obtained. These studies
are discussed in Chapter 4.

The optimized immobilization procedure was evaluated by studying the
reproducibility of the procedure and the amount of enzyme immobilized. The reactor’s
lifetime and the reaction conversion percentage were also investigated to test the enzyme
immobilization efficiency. Enzyme kinetics, which include the Lineweaver-Burk plot and
Michaelis constant, are presented in Chapter 5. System factors are extremely important to
consider in developing the sensitivity, stability, and sample throughput. Studies of the
effect of reactor lengths and diameters, ﬂow rate, and reaction time are also covered in
Chapter 5.

A glucose calibration curve is presented in Chapter 6. Its deviation from linearity
is compared with that of the hydrogen peroxide calibration curve. The analytical ﬁgures

of merit such as linear range, signal-to-noise ratio, detection limit, sample consumption,

and sample throughput are calculated and compared with those of the conventional
system. The applications of the CPI, immobilized enzyme system to fruit juices and soft
drinks are detailed in Chapter 6.

Background information pertinent to the miniaturization of ﬂow injection and the
development of the sugar determination methodologies is given in Chapter 2, along with
the results of previous studies in our laboratory relevant to the work described in this
Thesis. A general introduction to the instrumentation and experimental reagents involved
is presented in Chapter 3. Finally, conclusions and future perspectives of this work are

discussed in Chapter 7.

Chapter 2

HISTORICAL BACKGROUND

Background information pertaining to three unique yet related ﬁelds is important
in developing the methodologies for glucose determination described in this thesis. These
ﬁelds are ﬂow injection analysis (F IA), enzyme immobilization, and glucose
determination. This chapter provides a condensed review of each of the topics. A section

describing previous relevant studies in our laboratory is also included.

2.1 Flow Injection Analysis

2.1.1 Introduction of Flow Injection Analysis

One of the major developments in analytical chemistry during the last four
decades has been the appearance of commercial automatic analytical systems, which
provide analytical data with a minimum of operator intervention. FIA, in its present form,
was ﬁrst reported by Kent Stewart, Jaromir Ruzicka, and E10 H. Hansen in the mid
19708”’ '6. It is an outgrowth of segmented ﬂow procedures, which were widely used in
clinical laboratories in the 19603 and 19705 for automatic routine determination of a
variety of species in blood and urine samples for medical diagnostic purposes. F IA is
based on the injection of a liquid sample into a ﬂowing, nonsegrnented continuous carrier
stream of a suitable liquid. The injected sample forms a zone, which is then transported

toward a detector that continuously records a physical parameter, such as absorbance or

electrode potential, as it continuously changes due to the passage of the sample material

through a ﬂow cell.

2.1.2 Principles of Flow Injection Analysis

F IA is widely accepted in analytical and process control laboratories due to its
simplicity, high degree of automation, high sample throughput, and reproducible sample
handling capabilities. As commonly practiced, FIA depends on precise timing,
reproducible injection of samples, and a controlled amount of dispersion of the sample
and reagent zones involved.

The shape of the sample zone, after injection with a sampling valve, is determined
by convection and diffusion, which can be seen in Figure 2.1. The convection arising
from laminar ﬂow creates the parabolic shaped front of the FIA stream, which dominates
the dispersive processes in conventional F IA tubing. Diffusion (both radial and axial)
also contributes to the overall dispersive process. The radial diffusion is always important
in narrow tubing, while the axial process is not signiﬁcant under this circumstance.

The formula for dispersion D is D = C0 / c, where c0 is the analyte concentration of
the injected sample and c is the peak concentration at the detector. Although dispersion
can be controlled by appropriate choices of three interrelated variables: sample volume,
tube length, and pump rate, the dispersion that occurs in F IA methods can lead to a lack
of sensitivity or an inability to detect a product at all. Dispersion and the resulting
broadening can also severely limit the sampling frequency of FIA methods involving
complicated or slow reactions. To make FIA an ideal analytical tool, low dispersion

needs to be achieved and combined with the simplicity and efﬁciency of FIA.

deﬁant? .3 286.8qu A3 can .commahmv BEE 28 8:93:80 .3 EVER—mm“. A8 £28268
.3 :ommbmmmv 3v .zommcoammw on A3 ”.8883 05 an morn—mam mo 8an cease—3:8 co coast? 93 2880250 .«o 60km mm oSmE

A3 3 3v 3

can. 25 25 2::

 

 

 

 

 

 

 

 

 

 

nowadays—00:00
332...

 

 

 

 

 

 

 

 

 

\
A

Ea co Essa

2.1.3 Application of Flow Injection Analysis

F IA has evolved into an extremely versatile technique, capable of carrying out
many on-line operations for sample pretreatment, manipulation, and analysis. This

technique is also powerful due to its compatibility with many types of detection, such as

18

spectrophotometry“ , ﬂuorimetry'g, chemiluminescence” 21, FTIRZZ’ 23, mass

24,25 26,27

spectrometry (MS) , atomic absorption , amperometryzs’ 29, conductivity”, and etc.

The area of FIA application includes: reactions / derivatizationsn, kinetic analysis”,

33, 34 8 39, 40

immunoassays , preconcentrations35' 36, dilutions”, ﬁltrations3 , extractions ,

45,46

dial sis“ 42, as diffusion“ 44, ion exchan e , etc. FIA has recentl been used as a
Y g g Y

source of renewable surfaces in chemical microscopy techniques“.

2.1.4 Miniaturization of Flow Injection Analysis

As prescribed in 2.1.2, the efﬁciency of a FIA system can be signiﬁcantly
enhanced by reducing the amount of dispersion in the F IA streams. Solid particles such
as beads or coiled reactors can be employed to disrupt the laminar ﬂow proﬁles.
However, since the signal-to-noise ratio and the peak width are not signiﬁcantly
improved, to promote FIA to a higher level, other technical changes are required.

According to the Aris Taylor theory of dispersion“,

 

(02) = Trz
[novv 24I)na

the dispersion (zone variance) is directly proportional to the square of the tubing radius,

the residence time T, and the reciprocal of diffusion coefﬁcient Dm of the sample

molecules or ions. Thus, the amount of dispersion arising from convective forces should

be substantially decreased by using tubing of smaller diameter. The dispersion in fused

silica capillaries with an inside diameter of 75 pm is only 1/117 of that in 0.8lmm inside
diameter TEF LON tubing.

The lowering of dispersion in capillary ﬂow injection (CF 1) allows complex
chemistries or long reaction periods. In addition, since the zones from one injection to the
next should be less broadened inside the reactor, the sample throughput can be improved.
The CFI system developed by Spence and Crouch has shown improved efﬁciency over
conventional FIA; furthermore, reagent and sample consumption are reduced to micro

scales”.

2.2 Enzyme Immobilization

2.2.1 Introduction to Immobilized Enzymes

Enzymes are catalysts of biological processes”. They have a number of distinct
advantages over conventional chemical catalysts. Foremost amongst these are their
speciﬁcity and selectivity, not only for particular reactions, but also in their
discrimination between similar parts of molecules or between optical isomers. Enzymes
work under generally mild processing conditions of temperature, pressure and pH. In
addition, they bring high catalytic efﬁciencies and straightforward catalytic regulation.
There are some disadvantages in the use of enzymes, which cannot be ignored. In
particular, the high cost of enzyme isolation and puriﬁcation still discourages their use.
The generally unstable nature of enzymes, when removed from their natural environment,
is also a major drawback to their more extensive use.

Immobilized enzymes are deﬁned as “enzymes physically conﬁned or localized in

a certain deﬁned region of space with retention of their catalytic activities, which can be

10

”5'. Many of the disadvantages of the enzyme are

used repeatedly and continuously
eliminated and some additional advantages gained when they are immobilized. The
enzymes are reusable and so provide an economical alternative to other methods. The

enzymes retain their activity over a broader range of temperature and pH. They are easily

manipulated and have fewer interferences.

2.2.2 Methods of Enzyme Immobilization

Research and development work has provided a bewildering array of support
materials and methods for immobilization of enzymes. Selection of these should be made
by weighing the various characteristics and required features of the enzyme application
against the properties, limitations, and characteristics of the combined immobilization
method and support. Many practical aspects needed to be considered including: physical
strength, chemical hydrophilicity, available functional groups, stability, resistance to pH
changes, temperature, organic solvent, economic availability, safety, etc. There are ﬁve
principal methods for immobilization of enzymes”: adsorption, entrapment,
encapsulation, crosslinking, and covalent binding (see Figure 2.2).
2.2.2.1 Adsorption

Immobilization by adsorption is the simplest method and involves reversible
surface interactions between the enzyme and a support. The forces involved are mostly
electrostatic, such as van der Waals forces, ionic and hydrogen bonding interactionsﬁ.
These forces are very weak, but sufﬁciently large in number to enable reasonable
binding. However, leakage of enzymes from the support and contamination of product,

nonspeciﬁc binding, overloading of the support, and steric hindrance by the support have

limited the use of this method.

11

 

 

 

 

I -------------------------- 1 Adsorption
Entrapment
O ~~~~~
O 0 .' Encapsulation
‘‘‘‘‘‘ F, ,
Crosslinking

 

 

l Covalent Binding

Figure 2.2 Principal methods of immobilization.

12

2.2.2.2 Entrapment

Immobilization by entrapment differs from adsorption in that enzyme molecules
are free in solution, but restricted in movement by the lattice structure of a gel”. The
porosity of the gel lattice is controlled to ensure that the structure is tight enough to
prevent leakage of enzyme. At the same time, however, free movement of the substrate
and product is allowed. Inevitably, the support acts as a barrier to mass transfer, and this
can have serious implications for reaction kinetics.
2.2.2.3 Encapsulation

Encapsulation of enzymes can be achieved by enveloping the biological
components within various forms of semipermeable membranes”. Encapsulation is
similar to entrapment in that the enzymes are free in solution, but restricted in space.
Large proteins or enzymes cannot pass out of or into the capsule, but small substrates and
products can pass freely across the semipermeable membrane. This method suffers from
problems associated with acute diffusion, which may result in rupture of the membrane if
products from a reaction accumulate rapidly. A further problem is that the immobilized
enzyme particle may have a density fairly similar to that of the bulk solution with
consequent problems in reactor conﬁguration, ﬂow dynamics, and so on.
2.2.2.4 Crosslinking

This type of immobilization is support-free and involves joining the enzymes to
each other to form a large, three-dimensional complex structure. Crosslinking can be
achieved by chemical or physical methods“. But crosslinking is rarely used as the only
means of immobilization because of its poor mechanical properties and poor stability,

which are severe limitations.

13

2.2.2.5 Covalent Binding

Covalent binding is the most extensively used technique, and involves the
formation of a covalent bond between the enzyme and a support material”. The bond is
normally formed between functional groups present on the surface of the support and
functional groups belonging to amino acid residues on the surface of the enzyme. A
number of amino acid functional groups are suitable for participation in covalent bond
formation. Those that are most often involved are the amino group (NHz) of lysine or
arginine, the carboxyl group (COzH) of aspartic acid or glutarrric acid, the hydroxyl group
(OH) of serine or threonine, and the sulfydryl group (SH) of cysteine. Popular supports
include porous silica, porous glass, polyacrylamide, nylon, and agarose. Basically, two
steps are involved in covalent binding of enzymes to support materials. First, functional
groups on the support material are activated by a speciﬁc reagent, and second, the
enzyme is added in a coupling reaction to form a covalent bond with the support material.
To compensate for the somewhat rigid environment encountered with covalent binding,
spacer molecules are often used to increase separation of the enzyme from the support.
This allows the enzyme more freedom for conformational changes, which may be needed
for substrate interaction. The covalent binding method eliminates or reduces many of the
limitations associated with the other methods. The enzyme immobilization is not easily
reversed by pH, ionic strength, substrate, solvents, or temperature. Also, activity is

generally high and stability with time is good.

2.2.3 Applications of Immobilized Enzymes

The beneﬁts of an increased understanding of enzymes, and especially

immobilized enzymes, should allow many novel solutions to analytical problems

14

involving substrates, activators or inhibitors of these enzymes. In addition, the potential

for using immobilized enzymes as catalysts in areas such as food and clinical analysisss’

59 60, 61

, medicinal and chemical synthesis , and biotechnological and environmental
analysis62’ 63, has been widely promoted.
Immobilized enzymes can be used in combination with many analytical

64, 65

techniques, such as high performance liquid chromatography (HPLC) , laser-

6 69
S66, 67 8,

desorption ionization M , capillary zone electrophoresis (CZE) , sequential
injection analysis”, and FIA71 " 73. The enzymes are usually used as components of
bioreactors, bioseparators, and biosensors. In analytical applications, immobilized
enzymes are usually employed as enzyme reactors, enzyme probes, and enzyme
membranes.

Immobilized enzyme reactors (IMERs) are ﬁnding increased use due to their
compatibility with ﬂow systems and many types of detectors“. A wide range of reaction
types can be catalyzed by IMERs, including oxidation / reduction, inter and
intramolecular transfer of a variety of chemical groups, hydrolysis, cleavage of covalent
bonds, isomerization and addition of chemical groups across double bonds. 80, both
organic and inorganic reactions and complex biotechnological processes can be catalyzed
by one or more IMERs.

Enzyme probes are devices in which the enzyme is directly attached to the
sensing system in form of enzyme electrodes or sensors”. They have been used in both
potentiometric and amperometric applications.

Enzyme membranes are commonly used as the transducer in the enzyme electrode

probes. Enzyme membranes can also be used in conjunction with optical detection in

15

methods such as solid surface ﬂuorescence”. Immobilized enzymes have also been used

in various surfaces to form enzyme stirrers and separators.

2.3 Glucose Determination

2.3.1 Non-speciﬁc Methods and Separation Methods

Polarimetry, refractometry, and hydrometry can be employed in glucose
determination, if glucose is the only carbohydrate presented in the sample77 ' 79. These
techniques are non-speciﬁc, because they are generally incapable of distinguishing
between individual sugars. However, usually there are more than one carbohydrates
present, especially in food samples. Also, non-speciﬁc methods tend to be very time
consuming and suffer considerably from various interferences, which can result from
sugars other than glucose or from non-sugar components exist in the sample. Other
optically active sample components, such as amino acids and glycosides affect
polarimetric measurements. Acids and inorganic ions interfere with the speciﬁc rotation
of sugar.

Glucose has to be isolated from the sample mixture if an non-speciﬁc detection
method is employed”. Separation techniques such as gas chromatography (GC), HPLC,
and CZE have been used81 '83. These techniques suffer from complicated pre-separation

procedures and insensitive detection methods“.

2.3.2 Enzymatic Methods

Enzymatic methods are speciﬁc because of the high selectivity of enzymes.

Generally, enzymatic methods are sensitive, rapid, and reproducible. Glucose

l6

determination with enzymatic methods can be accomplished by measuring the reaction
product directly or indirectly after equilibrium is reached. Or, one can determine the
initial reaction rate, which is proportional to substrate concentration. Two popular
enzymatic pathways for glucose determination are illustrated in Figure 2.3.

In a direct method, glucose can be oxidized by oxygen catalyzed by glucose
oxidase to yield gluconic acid or gluconolactone and hydrogen peroxide.
Electrochemically monitoring the disappearance of oxygen or the production of hydrogen
peroxide can determine the corresponding glucose amount. Alternatively, the hydrogen
peroxide from the glucose oxidization reaction can be used in a second enzymatic
reaction in the presence of peroxidase and a reduced dye (leuco-dye) to produce a colored
product that can be determined with UV-visible absorption spectroscopy. O-dianisidine,
benzidine, leucomalachite green, 2,4-dichlorophenol, and 4-methoxy-l-napthol are some
examples of popular leuco-dyes. The coupling reaction of 4-aminoantipyrine (AAP) with
dichlorobenzenesulfonic acid, called the Trinder reaction”, is also widely used,
especially in clinical chemistry. This reaction is discussed in detail later in this chapter.

The other pathway is also shown in Figure 2.3. Glucose and adenosine
triphosphate (ATP) can be catalyzed by hexokinase and yield glucose-6-phosphate.
Glucose-6-phosphate can react with nicotinamide adenine dinucleotide phosphate
(NADP), which can be subsequently reduced to NADPH. The NADPH can be monitored
by direct absorption at 340 nm, or by ﬂuorescence at 460 nm with 340 nm excitation.

In recent years, enzyme sensors, enzyme electrodes, and enzyme membranes are
commonly used in the determination of glucose in various types of samples (food,

serum). These enzyme sensors and reactors are used in conjunction with such analytical

l7

dogma—count 883w 8m £8553 oumgwcm m.m oSmE

Ema/«Z + Bazooﬁwosmmogmé

 

x4 AHA—<2 + owmsmmoﬁé- 3830
035363an 03:98:90- 3830

 

 

 

o m mmo a- .88: 4
H a a o 5 / 33:20on AER. + 8820
of + $583.... A 25-853 + NON:
omeEonm
NONE + Eon 2:820 4 NO + 3820

33me 8820

18

techniques as HPLC, CZE, FTIR, and MS.

2.4 Previous Work in Our Laboratory

‘4 were focused on the

Previous studies in our laboratory conducted by Spence
design and development of the CPI system. His investigations include pressure effects,
zone variance, and mixing considerations. Work done by Thompson“, Stults87 and
Kurtz88 focused on methodologies for covalent immobilization of glucose oxidase andrthe

enzymatic determination of glucose with subsequent optimizations in a conventional FIA

system.

2.4.1 Capillary Flow Injection Development

Spence89 designed the ﬂow injection employing capillary tubing with inside
diameter of 75 um or less. He studied the pumping mechanism, designed a photometric
ﬂow cell, and chose appropriate injector and mixing aids, so that the miniaturized
instrumentation could be used in conjunction with conventional detectors. His studies
indicated that there was still a need for proper mixing techniques within a CFI system to
get efﬁcient reagent and sample zone overlap; otherwise, since the length of the sample
zone increases as the sample volume increases, the reagent and sample zones eventually
will reach a point where there is no overlap, and hence no mixing or reaction. The
stopped-ﬂow methods conducted by Spence90 show some of the beneﬁts of performing
F IA in a miniaturized format. He reported that CFI maintains the simplicity, speed of

analysis, and the ease of automation of conventional ﬂow injection. At the same time,

19

CF I also reduces the amount of sample volume needed, the reagent consumption, and the

waste generated by two orders of magnitude in selected cases.

2.4.2 Glucose Oxidase Immobilization

Thompson ﬁrst immobilized glucose oxidase onto nylon tubing in the form of an
open tubular reactor (OTR, see Figure 2.4), which was employed in a continuous ﬂow
system for glucose determinations". He investigated various immobilization methods for
covalent attachment of glucose to nylon, and linkage via glutaraldehyde provided
favorable results. His research showed that glutaraldehyde, a bifunctional molecule, not
only acts as the binding reagent, but also functions as a spacer between the enzyme and
support, thereby allowing more conformational ﬁ'eedom for enzyme activity. He also
carried out studies in the immobilization of glucose oxidase onto non-porous glass beads.
Enzyme attachment with glutaraldehyde following surface modiﬁcation with 3-
aminopropyltriethosysilane (3-APTS) was proved very successful for glucose oxidase.
This procedure is also advantageous due to the mild reaction conditions required when
compared to other common immobilization methods such as cyanogen bromide and
diazotization. Both the cyanogen bromide and diazotization methods involve hazardous
chemicals and harsh conditions, and can result in poor enzyme activity. Enzyme reactor
lifetimes have also been good with the 3-APTS / glutaraldehyde immobilization method,
on the order of several months in many cases.

Stults92 and Kurtz93 continued Thompson’s work with the immobilization of
glucose oxidase on non-porous glass beads in the form of a single bead string reactor
(SBSR) and controlled-pore glass (CPG) in the form of a packed bed reactor. These are

shown in Figure 2.4, for ﬂow injection determination of glucose. The immobilization

20

 

 

 

OTR

 

 

 

SBS(C)R

 

 

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
...............................................
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
ooooooooooooooooooooooooooooooooooooooooooooooo
...............................................
ooooooooooooooooooooooooooooooooooooooooooooooo
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
...............................................
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
00000000000000000000000000000000000000000000000
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I L I I

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIII I I I I I I I I I I I I I. I I I I .I I I I I .I I I I I I I I I I I I I I I

 

 

 

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
I I I I I I I I I I I I I_ I I I I I I I I I I I I I I I I I I I I I I I I I I I III. I LI I I

 

 

I I I I I I I I I I I I I I I I I I I I I I O I I I I I I I I I I I I I I I I I I If .....
I I I I I I I I I I I I I I I I—I I I I I I I I I I I I I I I I I III I I I I I I I I IAI I

 

 

 

. n o a o o l o n n u e c n I o o n I I ow oooooooooooooooooooooooooo
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn

 

CPG Embedded Reactor / SBS(C)R

Figure 2.4 Immobilized enzyme (capillary) reactor conﬁgurations.

21

 

 

 

 

 

procedure developed by Thompson was reevaluated and optimized by Stults94 and
Kurtz“. They also investigated the conditions for optimum surface area and glass
derivatization with 3-APTS. These are the factors that eventually limit the number of

sites available for enzyme binding and subsequent activity.

2.4.3 Glucose Determination

Thompson, Stults and Kurtz developed the enzymatic determination of glucose
via glucose oxidase. Thompson employed an OTR in an air-segmented continuous ﬂow
system. Stults designed a SBSR in a ﬂow injection system. Kurtz compared the SBSR, a
CPG packed bed reactor, a CPG embedded reactor, and a CPG embedded reactor / SBSR
(see Figure 2.4) by using immobilized glucose oxidase in a flow injection system.

All three previous workers chose the same method of detection (see Figure 2.5).
Hydrogen peroxide produced by the glucose oxidation reaction is colorimetrically
determined in the presence of peroxidase by the Trinder reaction”. In this reaction, 4-
aminoantipyrine (AAP) and 3,5-dichloro-2-hydroxy-benzenesulfonic acid (DCPS) are
oxidized and coupled to form a quinoneimine dye. The high molar absorptivity, stability,
and non-toxic nature of the Trinder reaction species make it a favorable choice over the
other oxygen acceptors such as benzidine, o-toluidine, and o-dianisidine. The reaction
performed very well under the instrumental conditions described by Thompson, Stults,
and Kurtz. However, limitations were encountered in Kurtz’s research, and she
investigated an alternative detection reaction, the malachite green reaction, which is more
sensitive and has a wider linear range than the Trinder reaction. This reaction was studied

in detail by Aspris95 in our laboratory, for determining glucose in olives.

22

.5382 SEE. \ 83me 88:5 Wm 2sz

 

 

 

 

m<< mmUQ
MID
_ m £0 £2 a
Z :0 5
N I. + N N
+
vOmNm + ONE? @1! Z _ A £0 '2 0 mm
032x80; , Z O
MOm /5
o / @ mo
0 U
ocowoﬂlméaooﬁwd omooﬂmdé
IO $0
or 0:
N N O “R A N
O I + :0 83me 8829 + O + 30

mofo mof

23

2.4.4 System Optimizations

Stults and Kurtz” 88

performed optimizations of the glucose F IA manifold in an
effort to attain maximum sensitivity and sample throughput. They employed a composite
modiﬁed simplex routine to optimize seven system parameters, which include carrier
ﬂow rate, carrier pH, temperature, peroxidase concentration, AAP concentration, DCPS

concentration, and reagent pH. Their optimization resulted in an improvement in

response 22.5 times over the initial conditions.

24

Chapter 3

INSTRUMENTATION AND EXPERIMENTAL REAGENTS

A brief discussion of the general instrumentation and experimental reagents is

presented in this chapter.

3.1 Flow Injection System

A single channel ﬂow injection system was used in all studies reported here. A
peristaltic pump propels the carrier and reagent streams. For glucose determinations, the
sugar sample is injected into the carrier by means of a valve. As the sample zone is
transported downstream, the enzymatic reactions take place to yield hydrogen peroxide.
Following the formation of hydrogen peroxide, the detection reaction takes place. Aﬁer
the dye formation, the stream enters a ﬂow cell. Detection was performed using an in-
house designed colorimeter with an interference ﬁlter at 540 um. All data acquisition was
accomplished using a program written with the LabWindows/CVI soﬁware package and

a LabPC+board (National Instruments)96.

3.1.1 Capillary Flow Injection (CFI) System

Figures 3.1 and 3.2 show the CPI systems used in the determination of glucose
with and without the immobilized enzyme reactor.
3.1.1.1 Manifold Tubing and Immobilized Enzyme Reactor Tubing

Polyimide coated fused silica capillaries (Polymicro Technologies) with an inside

diameter of 75 or 100 um and an outside diameter of 365 pm were used as the

25

.20: 56%
mm 258 9:32 0388 25. #382 08,35 poi—5255 53> 36on mo aouaﬁgoaov 05 .5.“ 29:58 EU mo ozaﬁonom _.m onE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:oEmEcom 03m.»
8% 839280 2280 _.£ wEnB v88 32m
N338 SEEP ESQ

Noam woman _ Eowmom

Busch

V7 L 038mm

A ..... DIFAHR . BEND

Spooﬁn—
08 @632 ESE oﬁﬁmtom

 

N888 ommeo 382w
coNEnoEEm 85m woman

26

.80: 8527. mm 258
@532 oEEam 2C. #882 2835 vogEoEEm 5923 382w no 5:83:83 05 .5.“ 29:58 EU mo oumﬁonom mac. oSmE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

536332“ 035$
3% Newsmﬁou 538.9:
ESQ
maﬁa @822 32m
-_ macaw
A ....... ﬂ _ Bow—Nod
860on
955 on 830
N888 x3598 . ﬂ . m

826 woman

27

immobilized enzyme reactor tubing for the studies. The inside diameter of the manifold
tubing used was 75 um.
3.1.1.2 Pumping Mechanism

A 12-channel peristaltic pump (Ismatec, model IP-12) was used to induce ﬂow. In
order to reduce pump pulsation, the pump had been previously modiﬁed in our laboratory
by doubling the number of rollers from 8 to 16. The reagents were introduced into the
capillaries using typical ﬂow-rate tubes (Cole Parmer) with inside diameters of 190 um.
The capillaries are simply inserted into the ﬂow rated tubing. No leaking was
experienced.
3.1.1.3 Injector

The injector used was a 4-port, two-position conical rotary valve (V alco
Instruments). The sample size is determined by a passage engraved on the valve rotor,
allowing precise, repeatable injections. The internal sample (500 nL) ﬂowpath was
chosen for the studies.
3.1.1.4 Mixing Aid

The mixing aid used in the CPI system was a 290 nL mixing tee (Upchurch
Scientiﬁc).
3.1.1.5 Photometric Flow Cell

The ﬂow cell for the photometric detector used with the CFI system was designed
by Spence and Crouch”. It is mainly comprised of two black Delrin units, a couple of
ball lenses and a PEEK sleeve as the optical path. With this ﬂow cell, the internal volume

is 320 nL with a path length of 1 cm.

28

3.1.2 Conventional Flow Injection System

The conventional ﬂow injection system with immobilized enzyme single bead
string reactor (SBSR) for glucose determination is shown in Figure 3.3. The deference
between this system and the capillary systems is listed below.

3. l .2.1 Manifold Tubing And Single Bead String Reactor Tubing

Teﬂon tubing (Cole Parmer) with an inside diameter of 0.81 mm and an outside
diameter of 1/16 " was employed as the manifold tubing and the SBSR tubing for the
studies.
3.1.2.2 Injector

A 6-port injection valve (Upchurch Scientiﬁc) was used. The sample size is
determined by the volume of a injection loop made of 15 cm long Teﬂon tubing with
inside diameter of 0.5 mm and outside diameter of 1/16 ” (Cole Farmer), allowing
precise, repeatable injections. An injection volume of 30 p.L was chosen for the studies.
3.1.2.3 Single Bead String Reactor

The dimensions of this reactor assembled with non-porous glass beads are
illustrated in Figure 3.4. The SBSR was constructed by aspirating glucose oxidase
immobilized beads with 0.6 mm diameter into a 10.0 cm length of Teﬂon tubing with an
inside diameter of 0.81 mm. The tube ends were then crimped to contain the beads.
3.1.2.4 Mixing Aid

The mixing aid used in the conventional ﬂow injection system was a 2.9 uL Peek

tee (Upchurch Scientiﬁc).

29

.20: :32? mm «.on 9552

0183 2:. .8882 08,35 3N=_N_oEE_ 5:» 883m mo coumﬁcbowou 05 8.“ Bahama: 20:8? Boa 3:283:00 m.m ouswi

 

 

 

 

 

 

    
 

 

 

 

 

 

:oEmEcom
8.8 .63980
63S, couoo ME
maﬁa. @822 32m

«f “nowmom

_ BEEP

_ ~3me

888.09 _ 038mm

03 @832

5.5 . a
ESE omzﬁmcom

 

N888 wagm ween ﬂwﬁm

30

 

.mmmm am 055

 

 

2

31

3.1.2.5 Photometric Flow Cell
The same photometric ﬂow cell as described in 3.1.1.5 was employed with a

internal volume of 2.2 “L.

3.2 Experimental Reagents

All solutions for glucose oxidase immobilizations and glucose determinations

were prepared with distilled water. No special ﬁltering of the solutions was needed.

3.2.1 Reagents Used in Optimized Immobilization Procedure

3.2.1.1 Saturated Ammonium Biﬂuoride Solution

A saturated (NI-I4F)HF (J. T. Baker Chemical) solution was prepared by
transferring 0.5 g of (N H4F )HF to 10 mL of distilled water.
3.2.1.2 3-aminopropyltriethoxysilane (3-APTS) Solution

A 10 % (v / v) 3-APTS (Sigma) solution was made by dissolving 1 mL 3-APTS in
10 mL of distilled water.
3.2.1.3 Phosphate Buffer Solutions

A 0.05 M phosphate buﬂ‘er solution was prepared by mixing 6.81 g of KHzPO4
(Baker) and 7.10 g of NazHPO4 (Baker) in 2 L of distilled water. HCl and NaOH were
used to adjust the solutions to pH 6.85 and pH 8.00 respectively.
3.2.1.4 Glutaraldehyde Solution

A 1 % (v / v) glutaraldehyde (Sigma) solution was made by dissolving 0.4 mL 25

% glutaraldehyde in 10 mL of 0.05 M phosphate buffer solution (pH ~ 8.00).

32

3.2.1.5 Glucose Oxidase Solution
Approximately 89 units (3.9 mg) of glucose oxidase (Sigma, type 2) was

dissolved in 1.0 mL of 0.05 M phosphate buffer solution (pH ~ 6.85).

3.2.2 Reagents Used in the Analysis of Glucose and System Studies

3.2.2.1 Glucose Standard Solutions

A 10 mM glucose (Sigma) standard solution was made by transferring 1.802 g
glucose to a l-L volumetric ﬂask and diluting to the mark with distilled water. Glucose
standards with concentrations ranging from 0.50 mM up to 10 mM were prepared by
appropriate dilution of the stock solution with distilled water.
3.2.2.2 Phosphate Buffer Solution

Preparation of the 0.05 M (pH ~ 6.85) phosphate buffer is described in 3.2.1.3.
3.2.2.3 4-Aminoantipyrine (AAP) Solution

A 10 mM solution was prepared by dissolving 0.508 g of AAP (Sigma) in 250 mL
of distilled water.
3.2.2.4 3,5-Dichloro-2-Hydroxy-Benzenesulfonic Acid (DCPS) Solution

A 10 mM solution was made by transferring 0.663 g DCPS (Sigma) to a 250 mL
volumetric ﬂask and diluting to the mark with distilled water.
3.2.2.5 Trinder Reagent Solution

The Trinder reagent consisted of 5 mL of 10 mM AAP solution, 5 mL of 10 mM
DCPS solution, 12 mg of horseradish peroxidase (Sigma, type 2), 10 mL of 0.05 M
phosphate buffer @H ~ 6.85) and 10 mL of distilled water mixed together in an amber

bottle before use.

33

3.2.2.6 Hydrogen Peroxide Standard Solution

A 100 mM standard solution was made by diluting 30 % (w / w) hydrogen
peroxide solution (Columbus Chemical Industries) in a l-L volumetric. Hydrogen
peroxide standards with concentrations ranging from 0.5 mM up to 5 mM were prepared
by appropriate dilution of the stock solution with distilled water.
3.2.2.7 Composite Reagent Solution

The composite reagent can be prepared in an amber bottle by mixing 20 mL of
glucose oxidase solution (dissolve 0.06 g of glucose oxidase (Sigma, type 2) in 50 mL of
distilled water), 10 mL of the 10 mM AAP solution, 10 mL of 10 mM DCPS solution, 24
mg of horseradish peroxidase (Sigma, type 2), 20 mL of 0.05 M phosphate buffer (pH ~

6.85) and 20 mL of distilled water.

34

Chapter 4

ENZYME IMMOBILIZATION

An immobilization procedure for glucose oxidase is required in developing
analytical methods for glucose determination. As previously stated, the 3-
aminopropyltriethoxysilane (3-APTS) / glutaraldehyde method of enzyme
immobilization has been successful with several enzymes, particularly on controlled-pore
glass and non-porous glass supports. However, this method of glucose oxidase
immobilization has never been used in fused silica capillaries.

As was discussed in Chapter 2, there is a smaller amount of dispersion associated
with fused silica capillaries than with conventional manifolds. This allows longer
residence times, which are desirable in enzyme applications. In an effort to apply the 3-
APTS / glutaraldehyde method of enzyme immobilization with fused silica capillaries
and to improve immobilization efﬁciencies for glucose oxidase, the various steps of the
immobilization process were investigated further and modiﬁcations were made to
improve enzyme loadings. The aim of this work was to minimize the amount of time
required to perform the immobilization and to establish a procedure that gave

reproducible results.

4.1 General Enzyme Immobilization Procedure

Covalent attachment of enzymes to an inert matrix provides a sturdy environment

and yet allows the enzyme to be readily accessible for reaction in continuous ﬂow

35

reactors. Glass provides an ideal inert support because of its low reactivity toward most
compounds and mechanical stability.

The general steps involved in the immobilization of enzymes onto glass supports
via 3-APTS and glutaraldehyde are illustrated in Figure 4.1. There are three main steps in
the procedure: pretreatment, activation of surface, and enzyme attachment. The ﬁrst step,
prior to actual immobilization steps, may include procedures to clean the support surface
for removal of impurities. Also, support pretreatments can include methods to increase
the surface area and initial reactivity”. For glass supports, the surface reactivity is
increased as the concentration of silanol groups increases relative to siloxane groups.
Following any pretreatment steps, the support is derivatized via 3-APTS, yielding
reactive amino groups on the glass surface. Glutaraldehyde, a bifunctional aldehyde, is
then used as a link group between the alkylamine modiﬁed surface and the enzyme,
according to well-known Schiff base chemistry”. Glutaraldehyde attachment is simple
and can be done under moderate conditions (e. g., pH and temperature). One aldehyde
group reacts with the amino group on the support surface while the other reacts with
various amino acid functionalities within the protein structure of the enzyme, such as the
amine groups of lysine and histidine and the phenol group of tyrosine. This procedure
covalently binds the enzyme to the glass support, and the glutaraldehyde carbon skeleton

provides conformational freedom for substrate interaction.

4.2 Initial Immobilization Procedure

The immobilization procedure developed by Kurtz88 for attaching glucose oxidase

onto non-porous glass beads is showed in Table 4.1. This procedure was initially used to

36

1. Pretreatment:

0 Cleaning
0 Increase surface area

0 Increase surface reactivity (silanol groups)

2. Activate Surface: convert surface silanol groups to active groups for
enzyme attachment

0 Silylation: (a-APT S)

OCHzCH3

l
I-OH + (CH3CH20)3SI(CH2)3NH2 —) I-O-Si(CH2)3NH2

I
OCH2CH3

0 Glutaraldehyde:

O O O

I II II I II
|-O-Si(CH2)3NH2 + HC(CH2)3CH —+ |-O-Si(CH2)3N=CH(CH2)3CH

3. Enzyme Attachment:

0

I II I
|-O-Si(CH2)3N=CH(CH2)3CH + HzN-E —> |-O-Si(CH2)3N=CH(CH2)3CH=N-E

Figure 4.1. General procedure for 3-APTS / glutaraldehyde immobilization of enzymes
onto glass supports.

37

Table 4.1 Initial glucose immobilization procedure.

 

 

 

 

Step Reagent Time Temperature (°C)
Cleaning concentrated HNO3 30 min 23
rinse HZO/acetone 23
dry N2 23
Elm saturated (5 % w / v) (NH4F)HF/ 1 hr 23
MeOH
dry N2 23
heat (open tube) 3 hr 450
SurfaLce Actiiition concentrated HCl 1 hr 23
Silylation 2 % (v /v) 3-APTS / dry acetone 5 hr 23
cure 15 hr 70
Glutaraldehyde l3 % (v / v) glutaraldehyde / 0.05 3 hr 23
M phosphate buffer pH 8.00 15 hr 4
rinse 0.05 M phosphate buffer pH 6.85 23
445 U / 5.0 mL 0.05 M phosphate 24 hr 4

Glucose Oxidase

 

buffer pH 6.85

 

38

immobilize glucose oxidase onto the inner surface of polyimide coated fused silica
capillaries with inside diameters of 100 or 75 pm. A peristaltic pump circulated the same
solutions through the capillary at the rate of 2 cm / sec or 5.3 uL / min for the same
amount of time at the same temperature as described in the initial procedure.

Through a trial and error process, it was discovered that the method, which was
well suited for controlled-pore glass and non-porous glass beads, did not work for fused
silica capillaries. First of all, the strong acid (concentrated HNO3, concentrated HCl) and
the organic solvent (acetone and methanol) used in the initial procedure reacted with the
TYGON ﬂow-rate tubes. Secondly, the polyimide coating of the fused silica capillary
could not survive the high temperature of 450 °C in the etching step, initially developed
by Onuska98 for etching the inner surface of glass open-tubular columns. Third, with this
procedure, the capillary was easily clogged at the step of glutaraldehyde linkage, even
though the length of the capillary was only 30 cm. Further more, the entire
immobilization procedure was too time consuming, taking four days to complete.

Several details of the initial procedure are noteworthy. One is that, HNO3 was
used in the initial support cleaning due to its superiority over alcoholic KOH. The other is
that, as shown in Figure 4.2, under acidic condition, surface siloxane groups are
converted Oiydroxylation) to silanol groups; this process is reversed (dehydroxylation)
via dehydration at elevated temperatures (above 200 °C). In an effort to increase the
concentration of surface silanol groups, the glass surface was washed by acid for an hour

aﬁer the etching step. Both of these procedures were not altered for the capillaries.

39

.Estﬂzauo coca—axegnou \ coca—$96.3 08.35 mam—O .Né 2:me

3:35

:0 IO

_ _ 028
la Io law I llllv

_ _ E

02.32%
\ o /
cam + Ia Io la I

40

4.3 Immobilization Optimization

4.3.1 Cleaning

In order to immobilize an enzyme efﬁciently, it is necessary to eliminate any
contaminants from the surface of the glass before the procedure is begun. The initial
support cleaning was performed at 23 °C with HNO3. It is known that in order for an acid
treatment to be effective, it must be carried out at elevated temperatures”. So this
procedure was modiﬁed, and the temperature increased to 80 °C. Both ends of the
capillary were sealed. Higher temperature and a sealed capillary resulted in improved

surface activity.

4.3.2 Etching

The initial conditions for etching involved soaking the glass beads in methanol
saturated with ammonium biﬂuoride for an hour, decanting the solution and drying the
beads, followed by heating them in an open glass tube at 450 °C for three hours.

Because of reaction with the TYGON ﬂow-rate tubes, methanol was replaced by
water to make the saturated ammonium biﬂuoride solution. The solution was not
circulated through the capillary for an hour. Instead, it was pumped into the capillary and
allowed to stand at room temperature for an hour before the capillary was put into the
oven. Also the temperature was decreased to 200 °C to save the polyimide coating of the
fused capillary. The reaction happened slowly in an open capillary, and the time was

increased to allow reaction overnight. The activity exhibited was poor. Then both ends of

41

the capillary were sealed under the above conditions. The activity turned out not to be
signiﬁcantly improved.

These observations can be explained by considering the reaction equilibrium:

(NH4F)HF (S) H ZHF (g) + NH3 (3)

The gaseous products will be quickly released from an open tube, resulting in little time
for the etching step. On the other hand, the gaseous products will be contained within the
sealed tube, shifting the equilibrium back to the reactant. This results in less silica etching
of the glass surface. Another consideration is that support surface water will be trapped in
the sealed tube. The presence of surface water may inhibit the etching process.

Based on these observations, one end of the capillary was sealed in the etching
step, the activity was greatly increased in this procedure over a completely sealed or

completely open capillary.

4.3.3 Silylation

The Silylation step in the immobilization procedure is extremely important with
respect to eventual enzyme loadings. The extent of Silylation or surface activation
ultimately controls the number of sites available for glutaraldehyde linkage and
subsequent enzyme attachment.

A complex sequence of reactions is possible when glass surfaces react with 3-
APTS. Which reactions predominate depends on the availability of protons in the
Silylation solvent (essentially H2O). The most common bonding models for
trialkoxysilanes with surface silanol groups are shown in Figure 4.3. For conditions
where the solvent is free of protic impurities, the top two bonding models are possible. In

these cases, no hydrolysis takes place between neighboring ethoxy groups, preventing

42

Aprotic Solvents:

' | OCH2CH3
— Sli-OH + 3-APTS —'9 — 8‘1—0- S|i(CH2)3NH2

OCH2CH3
— Si-OH _ Sli
‘ I \O\ / OCHzCH3
O + 3-APTS % O Si
| / \
— Si-OH _ SIi /o (CH2)3NH2
I
Protic Solvents:
OH
_ Stan ' l
I: —— Si- O- |i(CH2)3NH2
(I) + 3-APTS —> O o
_ ._ I I
Sll OH — S|i— O- S|i(CH2)3NH2
‘?
— S|i(CH2)3NH2
OH
OH l'
81- S|1\O / OCH2CH3
o + 3-APTS —-> 0 S‘
l. o/ \ o
— Sl-OH - Sll/

‘ S‘i(CH2)3NH2

Figure 4.3. Common binding models for 3-APTS.

43

polymerization of the silane. This result is very desirable in the preparation of bonded-
phase chromatography supports such as octadecylsilane (ODS) supports. When protic
impurities are present, the bottom two bonding models predominate. Here the ethoxy
groups are hydrolyzed and extensive polymerization of the 3-APTS can occur. The
polymerization should result in higher yields of active amino groups when compared to
conditions where polymerization is prevented. Thus, protic impurities may be beneﬁcial
in surface activations for enzyme immobilization applications. However, there is a limit
to the extent of desirable polymerization, which can be attained by controlling the
experimental conditions'oo.

An aqueous solution of 10% (v / v) 3-APTS was applied in the Silylation.
Although the original immobilization procedure allowed this reaction to proceed for 5
hours at room temperature, and 15 hours at 70 °C, it was found in the immobilization

optimization, that the period after the ﬁrst hour did not improve the activities. Thus, the

3-APTS solution was circulated in the capillary at 95 °C for only 1 hour.

4.3.4 Glutaraldehyde Activation

The glutaraldehyde was attached to the silane group as speciﬁed in Figure 4.1.
Instead of 13 % (v / v) in the initial procedure, a solution of 1% (v / v) glutaraldehyde in
0.05 M phosphate buffer (pH 8.00) was found to be sufficient for linking the enzyme to
the support. A speciﬁc investigation of glutaraldehyde reaction yield versus reaction time
was not conducted. Since the activities did not improve signiﬁcantly after 1 hour of
reaction at room temperature, the 15 hours in the initial procedure was deemed

unnecessary for complete reaction.

4.3.5 Glucose Oxidase Immobilization

The covalent binding of the enzyme was carried out with an aqueous solution of
89 units / 1.0 mL glucose oxidase as in the initial procedure. It was circulated through the
capillary at 0 °C overnight. The use of 0.05 M phosphate buffer (pH 6.85) made sure that
the enzyme was immobilized in its most active state. Finally, buffer solution was
circulated through the capillary reactor to remove the temporarily adsorbed enzyme

molecules.

4.4 Optimized Immobilization Procedure

The optimized immobilization procedure is summarized in Table 4.2. The entire
modiﬁed immobilization procedure takes two days to complete, whereas the initial
procedure required four days. The reaction conditions are more moderate. High
temperature (450 °C) is no longer needed. This procedure provides reproducible
immobilization and satisfactory reactor lifetime. The immobilization eﬂiciency was also
highly improved. These results and the comparison of the capillary enzyme reactor with a
conventional single head string reactor (SBSR) are discussed in detail in Chapter 5.

A 100 cm long fused silica capillary glucose oxidase reactor with inside diameter
of 100 um was prepared according to the optimized immobilization procedure. It was
evaluated by injecting 500 nL 0f 2 mM standard glucose solution twice. The carrier was
0.05 M phosphate buffer (pH ~ 6.85) with a carrier ﬂow rate of 5.3 11L / min. The Trinder
reagent was used with the same ﬂow rate. Figure 4.4 shows the two reproducible glucose

peaks obtained.

45

In all cases, the capillary reactors were stored at 4 °C.

46

Table 4.2 Optimized glucose immobilization procedure.

 

 

 

 

 

 

Step Reagent Time Temperature (°C)
Cleaning (sealed tube) concentrate HNO3 30 min 80
rinse H20 23
Mg saturated (5 % w / v) 1 hr 23
(N H4F)HF / H2O
heat (in a one-end-sealed overnight 200
tube)
rinse H20 23
Surface Activation concentrated HCl 1 hr 23
Silylation (circulated) 10 % (v /v) 3-APTS / H20 1 hr 95
Glutaraldehyde l % (v / v) glutaraldehyde 1 hr 23
(circulated) / 0.05 M phosphate buffer
pH 8.00
rinse 0.05 M phosphate buffer 23
pH 6.85
Glucose Oxidase 89 U / 1.0 mL 0.05 M overnight 0
(circulated) phosphate buffer pH 6.85
rinse 0.05 M phosphate buffer 23
pH 6.85

 

47

.0225 08on mo muons—cm 28 N maven? .3 88m? EU 05 5E» 358cc A5095. Soﬁrsv magma omaoqmom .vé oSmE

 

 

:36

L6

GOUEQJOSQV

..m~.o

 

..N.o

 

 

mmd

 

48

Chapter 5

ANALYZER MANIFOLD STUDIES AND SYSTEM
OPTIMIZATION

Optimal conditions for immobilization were evaluated based on the
immobilization reproducibility. In some cases it is desirable to know the amount of
enzyme actually immobilized. A method for such a determination is presented. Since it is
important to work with a reactor that retains its activity for a long period of time, the
useful lifetime of the immobilized glucose oxidase capillary reactor was investigated.
Ideally the reactor should provide high substrate conversion. The percentage of the
glucose oxidation reaction with an immobilized capillary reactor was calculated through
the stopped-ﬂow ﬂow injection technique and compared with the results obtained with a
conventional single bead string reactor (SBSR).

The enzyme kinetics were also studied. A Lineweaver-Burk plot was made, and
the Michaelis constant was calculated.

Achieving the glucose determination methodology goals, such as high sensitivity,
and sample throughput, not only requires highly effective immobilization, it also requires
consideration of such aspects as the ﬂow injection manifold design, and the system
optimization. The effects of the dimensions of the glucose oxidase reactor and the Trinder
reaction reactor were studied. In addition, investigations of ﬂow rate and reaction time
were carried out to obtain the optimized reaction condition for the analysis of glucose

with capillary ﬂow injection (CFI) techniques.

49

5.1 Immobilization Reproducibility

In order for enzyme activity to be reproducible from one reactor to another, where
the reactors are capillaries from separate immobilizations, the immobilization procedure
must be reproducible with respect to reaction yields at each step. The Silylation step in
both the initial and modiﬁed immobilization procedures is the least reproducible from a
reaction yield standpoint. The Silylation reaction yields were particularly inconsistent
when it was required to release the beads from the glass container before the step of
curing. This step was eliminated in the modiﬁed procedure and therefore, the Silylation
reaction yield should be more reproducible.

An investigation of the reproducibility of the optimized immobilization procedure
was conducted by comparing three 100 cm long capillaries with 100 um inside diameter
prepared simultaneously. Standard glucose solutions ranging in concentration from 1.0
mM to 4.0 mM were injected (500 nL) into each of the three capillary enzyme reactors
with a ﬂow rate of 5.3 uL/ min. The conﬁguration of the system is shown in Figure 3.1
with a 30 cm long Trinder reactor. All the ﬂow injection parameters were held constant.

Calibration curves for the three capillary enzyme reactors are shown in Figure 5.1.
The three reactors exhibited nearly identical activities. The relative standard deviation
(RSD) for the three calibration slopes was 1.1 %, compared with a RSD of 8.8 % using
the initial immobilization procedure in a SBSR reported by Kurtz“. The slope of another
batch of capillary reactor is also shown in Figure 5.1. The activity for this reactor was
much less than the three reactors from the same batch. The optimized procedure appears
to be reproducible for the same batch of capillaries. However, the reproducibility is poor

among different batches.

50

0.40

 

 

 

 

 

O Column 1
0,35 —_ I Column2
A Column3
0 30 _ V Column4 (another batch)
0.25 -
d)
“é 0.20 -—
.D
:93,
:2 0.15 —
0.10 -
0.05 -
0.00 -
-0-05 I r l l I

0 1 2 3 4
Glucose Concentration (mM)

Figure 5.] Enzyme immobilization reproducibility for glucose oxidase.

51

 

5.2 Determination of the Amount of Enzyme Immobilized

An assay by difference was used to measure the glucose oxidase loading on a 100
cm long fused silica capillary with 100 um inside diameter. Two milliliters of an 89 units
/ mL glucose oxidase solution were prepared, half of which was used for the covalent
binding step in the immobilization procedure and the other half as a control. The control
solution was diluted with the Trinder reagent to 10 mL. This solution was used as the
carrier solution at a ﬂow rate of 5.3 11L / min in the CPI analysis of glucose. A plain 100
cm long reactor was used in the place of the glucose oxidase reactor, and the Trinder
reactor was 30 cm long. Figure 3.2 shows the basic conﬁguration for this determination.
No mixing tee was employed in this study. Aﬂer glucose oxidase was attached to the
capillary, the other half of the reacted glucose oxidase solution was diluted and used as
the carrier solution in a similar manner. For each glucose oxidase solution, a 500 nL, 0.5
mM glucose standard was injected with a ﬂow rate of 5.3 1.1L / min, and the absorbance
was measured.

After correcting for the dilution factors, the amount of glucose oxidase
immobilized was determined to be 2.1 i- 053 ug / mmz. The total amount of glucose
oxidase immobilized in the capillary reactor was around 0.66 mg. The activity of the
glucose oxidase on a 100 cm long 100 um inside diameter capillary reactor was found to
be equivalent to a solution containing 0.3 mg / mL of enzyme in the composite reagent. It
would take 6.9 hours under continuous operation (a ﬂow rate of 5.3 uL / min) to use an
equivalent amount of glucose oxidase in solution as that immobilized on the capillary

reactor.

52

5.3 Lifetime of the Capillary Enzyme Reactor

The lifetime study of the capillary enzyme reactor was done in two 100 cm long
fused silica capillary reactors of the same batches with inside diameters of 100 um, on
which the enzyme was immobilized with the optimized procedure. One of them was
stored at 4 °C, while the other one was routinely tested. The system shown in Figure 3.1
with a 30 cm long Trinder reactor was employed. Glucose solutions (500 nL of 1 mM to
4 mM) were injected at a ﬂow rate of 5.3 [IL / min.

The loss in activities for routinely used capillary and stored capillary is shown in
Figure 5.2, which indicates the trend of the calibration curve slope of glucose versus
time. It can be seen that the activity in the capillary reactor, which was routinely used,
dropped to approximately 74 % of the original, and became fairly stable after 10 days.
The activity of the stored capillary reactor matched that of the capillary reactor, which
had been used for periodic testing, after around 30 days. Compared with the results
obtained by Stults”, which reported that the activity dropped to 56 %, it can be
concluded that the glucose oxidase immobilized by the optimized procedure for capillary

substantiates the cost advantage of enzyme immobilization.

5.4 Percentage of the Glucose Oxidase Reaction Completed

This approach focused on the conversion percentage of the total amount of
glucose. The system used is shown in Figure 3.1 with a 100 cm long glucose oxidase
reactor and a 30 cm long Trinder reactor. A glucose sample (500 nL of 2 mM) was

injected at a ﬂow rate of 5.3 uL / min. Next, the ﬂow was stopped after the entire sample

53

 

0.090

0.085 -

0.080 -*

0.075 -

0.070 —

Calibration Curve Slope (x 103 AU / M)

0.065 —l

 

O Routinely Used Capillary
I Stored Capillary

 

 

0.060

10

I I I I
1 5 20 25 30 35

Time (day)

Figure 5.2 Loss of glucose oxidase activity for routinely used capillary reactor and stored

reactor.

54

plug occupied the glucose oxidase capillary reactor. Then, various delay times from 0
min to 20 min were applied. Finally the pump was restarted.

As Figure 5.3 shows, the delay time allowed more reaction to occur before the
ﬂow was resumed. The absorbance increased linearly till 4 min of delay and then reached
a plateau, which tells us when the reaction reaches equilibrium. The conversion
percentage of the glucose was determined to be 47 % by the ratio of the absorbance with
no delay and the absorbance at the plateau. Almost half the total amount of glucose
reacted with the glucose oxidase immobilized on the 100 cm long, 100 um inside
diameter fused silica capillary with the optimized immobilization procedure.

This result was compared with that of a similar study on a conventional SBSR.
The dimensions of this reactor are illustrated in Figure 3.4. The initial immobilization
procedure developed by Kurtz was used to immobilized glucose oxidase onto non-porous
glass bead. 30 [AL of 5 mM glucose was injected. The ﬂow rate was 0.38 mL /min. The
construction of the system employed is given in Figure 3.3 with a 20 cm long Trinder
reactor. The result was given in Figure 5.4.

The conversion percentage of glucose in the SBSR was calculated to be 22 %
compared with 47 % in the capillary enzyme reactor. The conversion percentage is
directly related with the enzyme efﬁciency, which was doubled with the optimized
immobilization procedure for capillaries. In order to make a more realistic comparison,
one must take into account the enzyme loading area. The enzyme loading area of a 100

cm long, 100 um inside diameter capillary reactor is the same as that of a 16 cm long

SBSR.

55

om

588.. 0850 9323255 >3:an N8 ASE max—SE canoe? Boa 33?.on 0m oSwE

2:5 as: 85
B om

— _ L

 

 

eoueqrosqv

 

 

 

 

Yo

 

56

om

.Mmmm é Sm 38.85 an 2%;

2.5 95 85
2 cm m

 

 

 

 

 

 

eoueqrosqv

57

5.5 Lineweaver—Burk Plot and Michaelis Constant

The enzyme kinetic study was done with the system illustrated in Figure 3.1 using
a 30 cm long, 100 um inside diameter immobilized glucose oxidase capillary reactor and
a 30 cm long Trinder reactor. Various concentrations of glucose were injected (500 nL)
with a ﬂow rate of 5.3 [IL / min, allowing the glucose oxidation reaction time to be less
than 15 sec. Since the reaction time is short, the reaction rate, represented by the peak
height of absorbance of the dye, can be considered as the initial reaction rate.

The Lineweaver-Burk plot was made by plotting one over absorbance versus one
over the initial concentration of glucose. A straight line was obtained, which is shown in
Figure 5.5. The Michaelis constant was calculated from the intercept on the x axis to be
28 mM. That is to say, the reaction rate reaches half of its maximum value when the
concentration of glucose is 28 mM. This value agrees favorably with those found in the

101-

literature ‘06 for soluble or immobilized glucose oxidase, which range from 10 to 110

mM.

5.6 Dimensions of the Glucose Oxidase Reactor and the Trinder
Reaction Reactor

5.6.1 Dimensions of the Glucose Oxidase Reactor
5.6.1.1 Length of the Glucose Oxidase Reactor

Glucose oxidase reactors with lengths from 20 cm to 100 cm were made by
cutting a 100 cm long reactor. A glucose solution (500 nL of 2 mM) was injected with a
ﬂow rate of 5.3 11L / min using the system shown in Figure 3.1 with a 30 cm long Trinder

reactor. The relationship between peak absorbance and reactor length is given in Figure

58

SE 3 $820 6 8:928:00 .35 F

mLNF

L

.83 0339.850305]— m.m 0.53m

mm...

-

mmd

mNd

 

 

 

 

eoueqrosqv/ L

59

5.6. As expected, longer reactor lengths resulted in higher absorbance. The linear
relationship of absorbance and reactor length can be explained by the following two
reasons.

First of all, the reactor internal surface area is proportional to the reactor length.
Increasing the reactor surface increases the enzyme loading area. Thus, the glucose
oxidase immobilized in a 100 cm capillary is 5 times that of a 20 cm capillary. With a
longer reactor, more enzyme molecules react with glucose, resulting in a higher
conversion percentage and a higher absorbance of the dye product. Secondly, the reaction
time also increases in a longer reactor. With a ﬂow rate of 2 cm / sec or 5.3 uL / min, it
takes 40 sec longer for the glucose plug to travel through the 100 cm enzyme reactor than
through the reactor of 20 cm length. Because glucose has more time to react in a longer
reactor, the absorbance increases.

Obviously, reactors with lengths longer than 100 cm will provide higher
absorbance. In the following studies, 100 cm was chosen since its activity gave
acceptable absorbance values.
5.6.1.2 Diameter of the Glucose Oxidase Reactor

The surface area to volume ratio of the capillary increases as its inside diameter
decreases. With the same volume of the glucose sample, smaller diameter capillaries can
provide more surface area, so that glucose reacts with more enzyme molecules and gives
higher absorbance of the dye product. Also, smaller diameter capillaries produce less
diffusion and band broadening. So the absorbance peaks become higher and sharper.

Capillary reactors (100 cm long) with 100 um and 75 um inside diameters were

made with the optimized immobilization procedure. Glucose solutions with

60

Absorbance

0.20

 

0.18 -J

0.16 n

0.14 -

0.12 n

0.10 —

0.08 —

0.06 —

0.04 -

0.02 -+

0.00 —

 

 

I I I I I
0 20 40 60 80

Glucose Oxidase Reactor Length (cm)

Figure 5.6 Effect of glucose oxidase reactor length

61

I
100

 

120

concentrations ranging from 1.0 mM to 4.0 mM were injected (500 nL) with a ﬂow rate
of 5 .3 ILL / min using the system shown in Figure 3.1 with a 30 cm long Trinder reactor.
Calibration curves of glucose with both reactors were plotted in Figure 5.7. The curves

show that, as expected, the slope of glucose oxidase reactor with 75 um inside diameter is

higher than that of the reactor with 100 um inside diameter.

5.6.2 Length of the Trinder Reactor

The length of the reactor determines the time of the reaction, and this can
inﬂuence the absorbance of the product. The Trinder reactor length was varied from 10
cm to 120 cm. Glucose solution (500 nL of 2 mM) was injected into a 100 cm long, 100
um inside diameter glucose oxidase reactor using the system illustrated in Figure 3.1.

The result is shown in Figure 5.8. The absorbance only increased 0.04 AU when
the Trinder reactor length was changed from 10 cm to 120 cm. The curve shows that
peroxidase-catalyzed reaction that converts hydrogen peroxide to a quinoneimine dye is a
fast reaction. It takes a short period of time for the reaction to reach equilibrium. The
absorbance dominating reaction is the glucose oxidation reaction instead of the Trinder

reaction. Because of this, a 30 cm long Trinder reactor was used for the following studies.

62

0.5

 

—O— 100 um Capillary Reactor i
+ 75 um Capillary Reactor /

0.4 — /

0.3 —

0.2 -1

Absorbance

0.0 -

 

0.0 —

 

 

 

I I I I I
0 1 2 3 4

Glucose Concentration (mM)

Figure 5.7 Effect of glucose oxidase reactor diameter.

63

mNF

00—.

p

4.0ch .883.— ..ovcth .8 80.0m— mﬁ 0.5me

0:3 505.. uoaommm 2005....
mm on mm

 

 

 

 

0N00
.000

050.0
.. _..0
.rmmwd
..m_..0
.0230

 

N0

 

eoueqrosqv

5.7 Flow Rate Study

The system shown in Figure 3.1 was employed in the ﬂow rate study using
various pump settings ranging from 01 to 60. This represents ﬂow rates of 0.1 cm / sec to
6 cm / sec. Glucose solution (500 nL of 1 mM) was injected with different ﬂow rates.

The results are presented in Figure 5.9. The absorbance decreases as the ﬂow rate
increases and almost reaches a plateau when the ﬂow rate reaches 4 cm / sec. The
absorbance difference between the highest and lowest ﬂow rate is 0.2 AU. The change of
ﬂow rate signiﬁcantly changes the absorbance. This can be explained by the reaction
time. Small ﬂow rates allow long reaction times, which results in high absorbances.
Considering the dispersion, the resulting peak broadening, as well as the low sample
throughput, which can be caused by slow ﬂow rates, a ﬂow rate of 2 cm / sec was chosen

for most of the studies.

5.8 Pass Number Study

The reaction time was directly studied by reversing the direction of the peristaltic
pump. The ﬂow direction was changed several times, while the glucose sample plug was
in the immobilized enzyme reactor of the system shown in Figure 3.1. The absorbance of
500 nL of 1.0 mM glucose with different number of passes is illustrated in Figure 5.10.

The curve shows that 4 passes resulted in an increase of 0.1 AU, while 12 passes
resulted in an increase of 0.2 AU. Results show that the glucose oxidation reaction is a
slow reaction, which requires around 10 min to reach equilibrium. Hence, the reaction

time is one of the most important experimental factors that inﬂuence the slopes of the

65

00

p

.88 >50 .8 80.0.5 06 2:3”—

NBE: Z wﬁtom 08:0
0v 0N

 

 

..m0.0

:00

:20

 

mmd

 

aoueqrosqv

66

0..

 

.8083: 83 .8 80.0w 0 _ .m onE

89:32 9.80
0 _.

_ _

HO

..mN.0

 

0.0

 

eoueqrosqv

67

calibration curves. Because reversing the ﬂow causes more diffusion and band
broadening than slowing down the ﬂow rate, the glucose oxidase reaction time was
increased by using pump setting 05 ( 0.5 cm /sec) while the glucose sample plug was in
the immobilized enzyme reactor and then changing it to 20 (2 cm /sec) while it was in the

Trinder reactor. The calibration curve obtained using this method is shown in Chapter 6.

68

Chapter 6

CALIBRATION CURVES AND APPLICATIONS

The calibration curves were generated under the optimized experimental
conditions. Deviations from linearity were observed for the calibration curves. These
were compared with the deviations of the hydrogen peroxide calibration curve. The linear
range, the signal-to-noise ratio, the detection limit, the sample consumption, and the
sample throughput for glucose determinations were calculated and compared with those
using the conventional single bead string reactor (SBSR). An alternative method to
decrease the detection limit is described.

Another aim of this work was to develop a suitable and reproducible method for the
quantitative determination of glucose in real samples, such as fruit juices and soﬁ drinks.
The glucose concentrations of four real samples were determined using both of the
capillary ﬂow injection (CFI) systems with and without the immobilized enzyme

capillary reactor, and the results were compared.

6.1 Glucose Calibration Curve

The optimal system performance conditions were described in Chapter 5. A
calibration curve, obtained under such conditions, is shown in Figure 6.1. It was obtained
by using the system shown in Figure 3.1. The 100 cm long, 100 um inside diameter
glucose oxidase capillary reactor was immobilized with the optimized procedure. The

Trinder reactor was 30 cm long with an inside diameter of 75 um. 10 cm long capillaries

69

Absorbance

 

0.40

0.35 -

0.30 —

0.25 -

0.20 -

0.15 r

0.10 n

0.05 -

0.00 -

 

EIIIIH

 

I I I I
2 4 6 8

Glucose Conentration (mM)

Figure 6.1 Calibration curve for glucose.

70

10

 

12

with inside diameters of 75 um were used from the peristaltic pump to the injection
valve. The injection volume was 500 nL. The ﬂow rate employed was 5.3 mL / min.
Concentrations of glucose from 1 mM to 10 mM were chosen. Over the glucose
concentrations employed, relative standard deviations (RSD) of absorbance

measurements were approximately 0 - 4 %.

6.2 Deviation from Linearity

A signiﬁcant deviation from linearity was observed in the glucose calibration
curve illustrated in Figure 6.1, when the glucose concentration was more than 4 mM.
This non-linearity was also observed in the Trinder reaction. A calibration curve using
hydrogen peroxide as the analyte was constructed by injecting 500 nL of standard
hydrogen peroxide solutions (0.4 mM to 3.0 mM) into the same system described above.
All the experimental conditions were the same as those for glucose determinations. The
results are shown in Figure 6.2. The curve indicates that the linear range for hydrogen
peroxide is from 0 mM to around 2 mM.

Kinetic investigations of the Trinder reaction show that the initial rate becomes a
nonlinear function of hydrogen peroxide concentration at moderate concentrations of
hydrogen peroxide (approximately 2 mM) for a peroxidase concentration of 0.8 mg / mL
(1760 units)'°7. There is also a strong dependence of the reaction rate on pH for the
Trinder reaction. The slope is quite steep for the initial rate versus pH in the pH range of
6.00 to 7.50 with a maximum at approximately pH 7.75”. Unfortunately, this strong pH
dependence occurs within the typical pH values used for the glucose determination (pH

6.85). Another reason could be that the solubility of the Trinder reagent restricts the

71

Absorbance

 

0.40

0.35 -

0.30 —

0.25 —

0.20 n

0.15 —

0.10 -

0.05 -

0.00 -

 

 

F I I I
0 1 2 3

Hydrogen Peroxide Concentration (mM)

Figure 6.2 Calibration curve for hydrogen peroxide.

72

 

sensitivity of this detection reaction. The limited sensitivity and small dynamic range of

the Trinder reaction cause the deviation from linearity of the glucose calibration curve.

6.3 Linear Range, Signal-To-Noise Ratio, Detection Limit, Sample
Consumption, and Sample Throughput

The limitations of the Trinder reaction dictate the linear range for the
determination of glucose, which is from 0 mM to 4 mM. The signal-to-noise ratios were
calculated to be from 6.8 to 28. The lower limit of detection is largely determined by the
noise level of the baseline. The baseline of the system employed is shown in Figure 6.3,
which indicates the noise level to be 0.0125 AU. Based on the slope of the calibration
curve, the detection limit was calculated to be 0.15 mM. The ﬂow rate and variance of
the peak determine the sample throughput. At a ﬂow rate of 5.3 mL / min, the sample
consumption is 0.32 mL / hour, which is only 1.4 % of that in conventional ﬂow injection
analysis (FIA). The sample throughput was 28 samples / hour. A comparison of the above
values between the immobilized enzyme capillary reactor and a SBSR is given in Table
6.1. As can be seen, although the slopes of the calibration curves and the sample
throughputs of these two methods are similar, the linear range, detection limit, and

sample consumption of CFI are signiﬁcantly improved compared with those of F IA.

6.4 An Alternative Method

In an effort to increase the slope of the calibration curve, an alternative method

was developed. As discussed in Chapter 5, an increase in the reaction time signiﬁcantly

73

00m

._o>o_ 3.8: 05880 008802 m0 083.0

600 2::

mmmﬂmmﬁ 500.000

0

 

 

 

 

00.0
smbw000

030.0

 

.3000 .0

..mN~.0

05.3 .0

..mm0~.0

.3320
N0

 

 

aoueqrosqv

74

Table 6.1 Comparison of immobilized enzyme capillary reactor with SBSR.

 

 

Difference Capillary Reactor SBSR
Length (cm) 100 10
Inside Diameter (mm) 0.100 0.81
Enzyme Loading Area (cmz) 0.314 0.188
Slope Of The Calibration Curve ( x 0.085 0.08
103 AU / M)
Linear Range (mM) 0 - 4 0 — 2.5
Signal-To-Noise Ratio 6.8 - 28
Detection Limit (mM) 0.15 0.5
Sample Consumption (mL / hour) 0.32 23
Sample Throughput (samples / hour) 28 36

 

75

increases the absorbance of the quinoneimine dye. The reaction time was increased by
using a ﬂow rate of 0.5 cm / sec for the glucose oxidation reaction. This allowed the
reaction time to be 200 sec. After the glucose plug went into the Trinder reactor, the ﬂow
rate was changed back to 2 cm / sec. The sample throughput was decreased by this
method tol7 samples / hour. However, this limitation can be overcome by a higher slope
of the calibration curve and a lower detection limit. As shown in Figure 6.4, the slope of
the calibration curve of glucose was signiﬁcantly increased by a factor of 1.6 compared

with that using a uniform ﬂow rate. The detection limit was also decreased to 0.09 mM.

6.5 Applications to Real Samples

Rapid quantitative analysis of sugar in solution is important in fermentation and
brewing processes, and in the manufacture of fruit juice and soft drinks. The applicability
of the proposed method to the analysis of real samples was tested. The CF] analysis
system with the immobilized glucose oxidase reactor (100 cm long, 100 um inside
diameter) shown in Figure 3.1 was employed. Usually, in fruit juice and non-diet soft
drinks, the glucose is present in relative high quantities and a dilution by a factor of a
hundred is required to ﬁt for the relatively low linear range of the Trinder reagent. The
concentrations found were compared with those obtained using the CFI analysis system
without the immobilized enzyme reactor, which is shown in Figure 3.2. The four real
fruit juice and soft drink samples chosen are apple juice, icetea, Coca-Cola, and Sprite.
The concentration values obtained are listed in Table 6.2. By comparison of the results, it
can be seen that the difference of real samples concentrations obtained from both

approaches is small. And, the RSD of CFI with immobilized enzyme is slightly lower

76

Absorbance

0.35

 

0.30 -

0.25 0

0.20 -

0.15 -

0.10 —

0.05 -

0.00 ~

 

 

Glucose Concentration (mM)

Figure 6.4 Calibration curve for glucose (alternative method).

77

 

Table 6.2 Results of real samples.

 

Sample Name Concentration (M) (with Concentration (M) Difference (M)
immobilized enzyme (without immobilized

reactor) (RSD %, n=3) enzyme reactor) (RSD %,

 

n=3)
Apple Juice 0.115 (3.1) 0.129 (3.2) 0.014
Icetea 0.049 (3.8) 0.056 (4.1) 0.007
Coca-Cola 0.102 (3.2) 0.110 (3.3) 0.008
Sprite 0.095 (3.0) 0.090 (3.0) 0.005

 

78

 

than that of CF I without immobilized enzyme.

6.6 Concluding Remarks

CF I analysis with an immobilized enzyme reactor has been shown to be a fast and
suitable method for the determination of glucose. The method provides satisfactory
reproducibility in terms of RSD. This is of interest, in the quantitative analysis of glucose
in various fruit juices and soft drinks. This technique signiﬁcantly enhances the

performance in term of speed, robustness, simplicity, and cost efﬁciency.

79

Chapter 7

FUTURE PROJECTS

The original goals of this research were to develop methodologies for the
determination of glucose in a scaled-down capillary ﬂow injection (CFI) system.
Desirable characteristics such as high sensitivity, and sample throughput were pursued in
the manifold developments. Through consideration and investigation of various system
factors, including strategies of enzyme immobilization on capillaries, and optimization of
the system, these goals have been attained. Miniaturized ﬂow injection techniques have
many advantages over their conventional counterparts, such as the reduced amount of
sample needed and waste generated, and the increased efﬁciency and overall
performance.

However, the work accomplished still leaves many avenues open for further
investigation. These include increasing the efﬁciency of enzyme immobilization on
capillary walls, pursuing the alternative enzyme immobilization methods and reactor
conﬁgurations, determining the fraction of the total amount of active enzyme
immobilized, developing automated multichannel parallel sugar determinations, and

using immobilized enzyme reactors with electroosmotic ﬂow.

7.1 Increase of Enzyme Immobilization Efﬁciency on Capillary Walls

Although the optimized immobilization procedure gave good activity, higher
enzyme activity is always desirable to get better sensitivity, higher conversion efﬁciency,

and lower detection limits. This can be obtained through more thorough and systematic

80

studies of the immobilization procedure (e.g., the temperature control, the reaction time,
the pH and ionic strength of the buffer solution, and the reaction conditions). The
composite modiﬁed simplex procedure can be utilized as a mean of multivariate

optimization of the system parameters.

7.2 Alternative Enzyme Immobilization Methods and Reactor
Configurations

Alternative attachment procedures should be considered. Sol-gel technology has
been usedms, for example, to prepare porous silica materials that can modify the inner
walls of fused silica capillaries. In some cases, proteins can be directly entrapped in the
sol-gel matrix with retention of activity'og. It may also be possible to silanize the porous
silica surface followed by bifunctional group attachment. In yet a third approach, avidin-
biotin interactions can be studied as a means of immobilization. Here a biotinylated
enzyme is attached to a surface prepared with bound or coated avidinl 10.

To increase the enzyme loading area, four reactor design conﬁgurations (single
bead string capillary reactor (SBSCR), controlled-pore-glass (CPG) packed bed capillary
reactor, CPG embedded capillary reactor, CPG embedded / SBSCR) can be applied. They

are illustrated in Figure 2.4. Their kinetic and ﬂow characteristics need to be investigated

before they are applied in the determination of glucose.

7.3 Determination of the Fraction of the Amount of Active Enzyme
Immobilized

The fraction of the amount of active enzyme immobilized was not determined in

this study. The reason is that the linear range of the calibration curve is restricted by the

81

 

Trinder reaction instead of the glucose oxidation reaction. The absorbance of the glucose
solution leveled off before all active enzyme molecules reacted with the substrate. The
100 cm long immobilized enzyme reactor was replaced by a shorter one (10 cm) in order
to let the enzyme reaction dominate the linear range of the calibration curve of glucose.
However, it did not work due to the low sensitivity of the short reactor.

This step can be taken after the immobilization efﬁciency is further improved, or
by using another detection reaction. The malachite green detection reaction shown in
Figure 7.1 can be substitute for the Trinder reaction. In this reaction, leucomalachite
green (LMG) is oxidized by hydrogen peroxide in the presence of peroxidase to
malachite green, which has an absorption maximum at 620 nm. In contrast to the Trinder
reaction (Figure 2.5), the reaction stoichiometry is 1:1 for LMG and hydrogen peroxide,
where the Trinder reaction stoichiometry is 1:1:2 for 4-aminoantipyrine (AAP), 3,5-
dichloro-Z-hydroxy-benzenesulfonic acid (DCPS), and hydrogen peroxide respectively.
The malachite green reaction was shown by Kurtz88 to be approximately eight times more
sensitive for hydrogen peroxide than the Trinder reaction. Also, the reaction exhibits a
wider linear range. Alternatively, the hydrogen peroxide can be detected by

chemiluminescence with the luminol reaction.

7.4 Multichannel Parallel Sugar Determinations by Capillary Flow
Injection Analysis

A multichannel capillary sugar analyzer for the determination of glucose,
galactose, sucrose, maltose, lactose and fructose can be developed. All of these sugars are

important to determine in foods and in other systems. The disaccharides (sucrose, maltose

82

 

880 880232

88%.

ONEN + @I U

.8082 :0on 23382 2. 0.5me

580 8808888304

8:82

28:00 A E + ~on + @ 0 @ 28:8
0800880 _
:

 

83

and lactose) can be enzymatically hydrolyzed to yield one or two molecules of glucose by
the enzymes invertase (for sucrose), a-glucosidase (for maltose) and lactase (for lactose).
Fructose can be converted to an equilibrium mixture of fructose and glucose with glucose
isomerase. The enzymatic reaction schemes are given in Figure 7.2. In each case the
glucose produced can be enzymatically converted to gluconic acid or gluconolactone and
hydrogen peroxide. The six sugars are anticipated to be determined simultaneously with
parallel immobilized enzyme capillary reactors. Hence, the original sample can be split
into six streams with a splitter and sent to the various reactors. Either one detector could
be used with the products from the reactors entering at different times, or parallel
detection could be employed. In this system, CFI will be of advantage in minimizing
sample volumes and in avoiding excessive dispersion in the serial reactors. The
determinations will essentially be ﬁxed-time kinetic determinations, since the enzyme
reactions are slow and incomplete during the residence times in the reactors. A

preliminary design of the multichannel analyzer is illustrated in Figure 7.3.

7.5 The Use of Immobilized Enzyme Reactors with Electroosmotic Flow

The use of immobilized enzyme reactors with electroosmotic ﬂow (EOF) can be a
possible choice for future applications. EOF is capable of producing steady ﬂow in the
sub ILL / min range and should also minimize dispersion due to convective forces. EOF
has the major advantage of a nearly ﬂat ﬂow proﬁle since ﬂow originates near the walls.

A major difﬁculty is having the correct pH and ionic strength to sustain EOF.

84

 

880200 0000000 0008.380 NH 082.0

 

0802w..Q A v
0008802 08020 08800-9

0880890 + 0802w-Q A 08804 ONE + 0882

 

080: w- -8 + 080: w- A N
_ Q _ D 08280208 0 I + 08208

A

08800-9 + 080205 A ONE + 08800

088882 080.880

NONI + 082080x02-080_0w-m \
M8205 088080

 

ONE + NO + 08808w..Q

 

N N - A N +N +080: wd-
O I + 200 080020 G 08080 08020 O E O _ a

85

889000

saws“ 50 0505035 .5300 2 050E

8808C 088000 205

088000 088m

08%

955

 

80005
20w000

0:0
8:000

0EE0m

86

LIST OF REFERENCES

1. Altria K. D.; Kelly M. A.; Clark B. J. T rac-T rend Anal. Chem. 1998, 17, 204-226.

2. Shelton C. M.; Koch J. T.; Desai N.; et a]. J Chromatogr. A. 1997, 792, 455-462.

3. Banks J. F. Electrophor 1997, 18, 2255-2266.

4. Pesek J. J .; Matyska M. T. Electrophor 1997, 18, 2228-2238.

5. Wang J; Bhada R. K.; Lu J. M.; et a1. Anal. Chim. Acta. 1998, 361, 85-91.

6. Manz A. Chimia. 1996, 50, 140-143.

7. Dale G; Ewen P. J. S. Iee P-Optoelectron. 1997, 144, 426-432.

8. Matsuda. R.; Hayashi Y.; Suzuki T.; et a1. Fresen. J. Anal. Chem. 1993, 347, 225-229.
9. Dempsey 13.; Diamond D.; Smyth M. R.; et a1. Anal. Chim. Acta. 1997 346, 341-349.
10. Spence D.M.; Sekelsky A.M.; Crouch S. R. Instrum. Sci. Technol. 1996, 24, 103-113.
11. Reach G.; Wilson G. S. Anal. Chem. 1992, 64, 381A-386A.

12. Jonas R.; Farah L. F. Polym Degrad. Stabil. 1998, 59, 101-106.

13. Brooks H. W.; Hall G. A.; Wagstaff A. J .; et a1. Vet. J. 1998, 155, 263-274.

14. Spence D. M. Ph. D. Thesis, Michigan State University, East Lansing, MI, 1997.
15. Stewart, K. K.; Beecher, G. R.; Hare P. E. Anal. Biochem. 1976, 70, 167.

16. Hansen, E. H.; Ruzicka, J. Anal. Chim. Acta. 1975, 78, 145.

17. Ye Y. Z.; Mao H. Y.; Chen Y. H. Talanta. 1998, 45, 1123-1129.

18. Safavi A.; Haghighi B. Talanta, 1997, 44, 1009-1016.

19. Tran C. D.; Lu J. Appl. Spectrosc. 1996, 50, 1578-1584.

20. Barnett N. W.; Hindson B. J.; Lewis S. W. Anal. Chim. Acta. 1998, 362, 131-139.

21. Emteborg M.; Irgum K.; Gooijer C.; et a1. Anal. Chim. Acta. 1997, 357, 111-118.

87

 

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38

39.

40.

41

42

43

Ramos M. L.; Tyson J. F .; Curran D. J. Anal. Chim. Acta. 1998, 364,107-116 .
Abderrazak H.; Lendl B.; Kellner R. Quim. Anal. 1997, 16, 27-30.

Schroder H. F. J. Chromatogr. A. 1997, 777,127-139.

Kishimoto K.; Nakamura H. Anal. Sci. 1997, 13, 815-816.

Haghighi B.; Maleki N.; Safavi A. Anal Chim. Acta. 1997, 35 7, 151-156.
Jimenez M. S.; Castillo J. R. J. Anal. Atom. Spectrom. 1997, 12, 1397-1402.
Schachl K.; Alemu H.; Kalcher K.; et a1. Analyst. 1997, 122, 985-989.

Trojanowicz M.; Krawczyk T. K. V.; Zmorzynska M.; et a1. Electroanal. 1997, 9,
1062-1066.

Kuban V. Collect. Czech. Chem. C. 1997, 62, 609-619.

Papaefstathiou 1.; Bilitewski U.; deCastro M. D. L. Fresen. J. Anal. Chem. 1997, 35 7,
1168-1173.

Hashemi P.; Olin A. T alanta. 1997, 44, 1037-1053.

Stiene M.; Bilitewski U. Analyst. 1997, 122, 155-159.

DelCarlo M.; Mascini M. Anal. Chim. Acta. 1996, 336, 167-174.

Lee G. H.; Park S. A.; Song K.; et a1. Anal. Sci. 1997, 13, 27-30.

Song W. L.; Wang L. S.; Zhi Z. L. Chem. J. Chinese.U. 1997, 18, 1621-1623.

Elazouzi H.; PerezJordan M. Y.; Salvador A.; et al. Spectrochim. Acta. B, 1996, 51,
1747-1752.

Bechmann I. E. Talanta. 1997, 44, 585-591.

Grudpan K.; Jakmunee J .; Sooksamiti P. Lab. Robotics. Automat. 1998, 10, 25-31.
Ensaﬁ A A.; Rezaei B. Anal. Lett. 1998, 31, 167-177.

Hasebe T.; Nagao J .; Kawashima T. Anal. Sci. 1997, 13, 93-98.

Hernandez 0.; Jimenez F.; Jimenez A. I.; et a1. Anal. Chim. Acta. 1996, 320, 177-183.

Su X. L.; Yu B. S.; Tan H. W.; et al. J. Pharmaceut. Biomed. 1998, 16, 759-769.

45.
46.
47.
48.
49.
50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

_ 60.

61.
62.

63.

. Su X. L.; Nie L. H.; Yao 8. Z. T alanta. 1997, 44, 2121-2128.

Mifune M.; Mukuno T.; Tani M.; et a1. Anal. Sci. 1998, 14, 519-522.
Vela M.; Saurina J .; Hernandez-Cassou S. Anal. Lett. 1998, 31, 313-331.
Pollema C. H.; Ruzicka. J. Anal. Chem. 1994, 66, 1825-1831.

Aris R. Pro. R. Soc. Ser. A. 1956, 245, 268.

Spence D. M.; Crouch S. R. Anal. Chem. 1997, 69, 165-169.

Gerhartz W. Enzymes in Industry; VCH Publishers: New York, 1990.

Chibata I. Immobilized Enzymes, Research and Development; Kodansha: Tokyo,
1978.

Poppe L.; Novak L. Selective Biocatalysis; VCH Publishers: New York, 1992.

Bickerstaff G. F. Immobilization of Enzymes and Cells; Humana Press: New Jersey,
1997.

Trevan M. D. Immobilized Enzymes: John Wiley & Sons Ltd.: Chechester, 1980.

Guilbault G. G.; Mascini M. Uses of Immobilized Biological Compounds; Kluwer
Academic Publishers: Dordrecht: 1993.

Wiseman A. Handbook of Enzyme Biotechnology; Ellis Horwood Limited: London,
1995.

Lam S.; Malikin G. Analytical Applications of Immobilized Enzyme Reactors; Blackie
Academic & Professional: London, 1994.

Stein K.; Shi R. B.; Schwedt G. Anal Chim. Acta. 1996, 336, 113-122.

Krawczyk T. K. V.; Trojanowicz M.; Lewenstam A.; et al. Biosens. Bioelectron.
1996, 11,1155-1165.

Hemandez-Justiz 0.; Fernandez-Lafuente R.; Terreni M.; et a1. Biotechnol. Bioeng.
1998, 59, 73-79.

Wainer I. W.; Johnson D. V.; Wahnon D.; et al. Chim. Oggi. 1997, 15, 45-48.
Mita D. G.; Portaccio M.; Russo P.; et a1. Biotechnol. Appl. Bioc. 1995, 22, 281-294.

Dezotti M.; InnocentiniMei L. H.; Duran N. J. Biotechnol. 1995, 43, 161-167.

89

64. Osborne P. G.; Yamamoto K. J. Chromatogr. B. 1998, 707, 3-8.

65. Wada M.; Kuroda N.; Ikenaga T.; et a1. Anal. Sci. 1996, 12, 807-810.

66. Marquez C. D.; Lee M. L.; Weintraub S. T.; et al. J. Chromatogr. B. 1997, 700, 9-21.
67. Righetti P. G.; Bossi A.; Wenisch E.; et al. J. Chromatogr. B. 1997, 7699, 105-115.
68. Huang W.; Wang J. Q.; Bhattacharyya D.; et a1. Anal. Chem. 1997, 69, 4601-4607.
69. Nashabeh W.; Rassi Z. E. J. Chromatogr. B. 1992, 596, 251-264.

70. Schindler R.; Watkins M.; Vonach R.; Lendl B.;Kellner R. Anal. Chem. 1998, 70,
226-231.

71. Wada M.; Kuroda N.; Akiyama S.; et a1. Anal. Sci. 1997, 13, 945-950.

72. Chen R. L. C.; Chang B. W.; Chang H. C. J. Chin. Chem. Soc.-Taip. 1997, 44, 423-
426.

73. Yaqoob M.; Nabi A.; MasoomYasinzai M.; J. Biolum.Chemilum. 1997, 12, 1-5.

74. Christian G.; Krull 1. Tyson J. Anal. Chem. 1990, 62, A455-A458.

75. Scheller F.; Pfeiffer D.; Hintsche R.; Dransfeld 1.;Schubert F .;Wollenberger U.
Applications in Medicine, Environmental Protection, and Process Control; VCH
Publishers: New York, 1989.

76. Bowers L. D.; Carr P. W. Anal. Chem. 1976, 48, 545A.

77. Paton N. H.; Player M. R.; Urquhart R. M. RM, et al. Zuckerindustrie. 1993, 118,
705-709.

78. Hughes B; Jelks V.; Hughes D. L. J. Chem. Educ. 1988, 65, 1007-1008.
79. Pomper K. W.; Breen P. J. J. Exp. Bot. 1995, 46, 743-752.

80. Churms S. C.; CRC Handbook of Chromatography: Carbohydrates, Vol. II; CRC
Press: Boca Raton, 1990.

81. Boldizsar I.; Horvath K.; Szedlay G.; et a1. Chromatographia. 1998, 47, 413-419.
82. Magne V.; Mathlouthi M.; Robilland B. Food. Chem. 1998, 61, 449-453.

83. Honda S.; Taga A.; Suzuki K.; et al. J. Chromatogr. 1992, 597, 377-382.

90

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

Vratny P.; Brinkman U.; Frei R. Anal Chem. 1985, 5 7, 224.

Trinder P. Ann. Clin. Biochem. 1969, 6, 24.

Thompson R. Q. Ph. D. Thesis, Michigan State University, East Lansing, MI, 1982.
Stults C. L. M. Ph. D. Thesis, Michigan State University, East Lansing, MI, 1987
Kurtz K. S. Ph. D. Thesis, Michigan State University, East Lansing, MI, 1992
Spence D.M.; Crouch S. R. Quim. Anal. 1996, 15, 296-300.

Spence D.M.; Crouch S. R. Anal. Chim. Acta. 1998, 358, 95-101.

Thompson R. Q.; Kim H. Biochim. Biophys. Acta. 1988, 964, 116-120.

Stults C. L. M.; Wade A. P.; Crouch S. R. Anal. Chim. Acta. 1987, 192, 155-163.
Kurtz K. S.; Crouch S. R. Anal. Chim. Acta. 1991, 254, 201-208.

Stults C. L. M.; Wade A. P.; Crouch S. R. Anal. Chem. 1987, 59, 2245-2247.
Aspris P. Master Thesis, Michigan State University, East Lansing, MI, 1990.

Spence D. M.; Cullen T. F.; Hybl J. D.; et a1. Abstr. Pap. Am. Chem. S. 1996, 212,
132-ANYL Part 1.

Carr P. W.; Bowers L. D. Immobilized Enzymes in Analytical and Clinical Chemistry;
John Wiley & Son: New York, 1980.

Onuska F. I.; Comba M. E.; Bistricki T.; Wilkinson R. J. J. Chromatogr. 1977, 142,
117.

Holland L. The Properties of Glass Surfaces; John Wiley & Sons: New York, 1964.

100. Caravajal G. S.; Leyden D. 13.; Quinting G. R.; Maciel G. E. Anal. Chem. 1988, 60,

1776-1786.

101. Cha C. S.; Chen J .; Liu P. F. Biosens. Bioelectron. 1998, 13, 87-94.

102. Zook C. M.; LaCourse W. R. Anal. Chem. 1998, 70, 801-806.

103. Kondrateva E. G.; Reshetilov A. N.; Yaropolov A. 1. Appl. Biochem. Micro+. 1997,

33, 530-534.

104. Chang H. Y.; Bae Z. U. Anal. Lett. 1997, 30, 1981-1992.

91

 

105. Schlereth D. D. J. Electroanal. Chem. 1997, 425, 77-85.

106. Khan G. F.; Ohwa M.; Wemet W. Anal. Chem. 1996, 68, 2939-2945.

107. Karayannis M. I. Unpublished work, Michigan State University, 1987.

108. Brinder C. J .; Scherer G. W. Sol — Gel Science; Academic Press: New York, 1990.
109. Barreau S.; Miller J. N. Anal. Comm. 1996, 33, H5-H6.

110. Guo Y.; Colon L. A. J. Microcolumn. Sep. 1995, 7, 485-491.

92

 

"‘11111111181011ES