I
312%.? ' ‘II‘ I
.V I2 *I‘
' I

.‘I‘Il-Z in
,I pnl'flp. .1.
"I; , :I'J‘. "

:‘4‘ ’l;

I I.
I. II. .IIIIII
.....

I "'I
I :. “315k; 4-:
.‘ 'lI'" '.~ I”
." ’: hiilV.‘
I“ I'|'I\l" V
I-" I '7
I " J

II .
II

. I. IIIII -
.\ IIIII I. f I . IVIII In
I II II or. LI I .
III .III I I I .vI I.
III... .. 3. v. I... w. - .
I n I III. I <0 I .F
III. \II I III “IR-III: I .vIIIII
II I. JI I I II<.I MN” -\ (VP: I. Md
- I VnIIv IIIIIQ .lkv I
v I 4. v4 LII-l h-
Osv IL I . o
I I . II. .
o AulI 5‘... It I.
I I III I
a III I II I laI I OIII I. l
I I I I I
.0 I‘ll: r . V0. III, I; I II (War-I II I. .
IV Ofi I: not .I J, I III: III
C I. I. I‘
I I I
‘5'. . IIILII .000!!! In... I. III; III I. 1.! I
.1. .IMI I....IJ IIII..II.I.I: II. I.
. \. III. IrIPI. I 1295. .i. U”
I .IIII. .. I. I {II
I3”. V IIII v .I I I
.II. 0 AI“ .- III. A I
. I. I. .. I I.
I 2"“. I '0.- ‘ I'
?.I.11LII II VIIVI I.‘ .. I I
I. III II IVIII I II
I-.. a. I. I.. . I I .
I P .I... .III .I I... .II I.
II I]. k'I I I I {IQ I II
I}! . III. II I I. . .I
III V I VI II. III I .I
I II -.III ..I I I
I . I I.
I II I
I . I!” III, I3
I II JIMMIIII iq}I I...“ I
II II! I I” III!
..... JIIII I JII II .IJ II
It, .I III,
I0 I. ..... III—II .IIIIIPIHNII. I
I II“ I 1...”!
l- I I I / II I I
I, I I I.
I IIyIP

z I I
a I I.“
III. I
1' I
II x
I. I I Ia}. II
I I I. III.
I I J In
I
I I I~ 01. I I
I II I I to A
I .1 III I
Iul t I»
L.I I I al
A \ II ‘0’
I ’ I I or I,
I 0".» II:
II N I IJIIII
I I .7.- 1|
I I I I I
I I In I I II I I
I? I I-III In I.
It I
‘4 I
J. .
II -I o
I. I. II . I.
I . II I I I. I. . I
III IVII I
I
IIVII. . IIIMII
.. o1
- I K I I
Inxr‘l . .I I .II.
II... I I I III I III.
I I... IIIIIIOT do .7
II- I III. I c INIIII I I
I I (Iflj‘a III
I INIIIHI . IIIIIIIIIIOIuIIJIIIIIII I
III. NIILIIIII III-III III
III: . I I
rIIIrI. I. ”II. «I
I II
94.1? .4! I III 3 II.
A ~ .I D
.. ‘OII I II'
I III. IV'I.I|I I
II I!

II
I
I I. I
I O I I I I. I
' IIIII I II ‘II III
C .
. I I I I
I. l I 0.5.1 .
I. C
III I II
III.
I. I
I I I H.
I III II).
I‘ II. III-
II ’ O I 0‘ |
Q. A g
II I. ..
CIVInII II I
I I
I I \ III. .
II .II II
IIJquhoIII
7 II. IIIIIIII
LIIIIIIIIIIIIIIIIIHUI I
.O‘V’I" On I‘ l I I f-I
I
I. I III
‘3 .VIO. . D
I \ .I v v. I
I PI» I
v I .
I I .I

I.
II
III

II
I I
. II I a
I I I I III I.
I I \
I ’ 0" Ihu' I I I
I II I I III/x.
I III, I’ll- I I. 1"
(I 1 I. I
I. I l I
III! II II ..I
In. I III. II I I I
. IIIII IIII' ,I. I
O l J I! I .I “ III I
III III .I I
IIIII I II I III II
M I I .I I I . I I l I
I II II I
I I.
'I l ' .I
. - lrII IIIIIINIIIIIIIIIIIIIII
I. II IIIIIIIIIIIIIIIIII II II I .

.I
I
I I .
I
I. I
' - '.l | . V
I .I: I.
I I ‘ I V
I II I
I!
I I I
I I II I 0 III .
'. ’
.5. II." o l
' J I n
I I I . 1
. III
I III I
H‘ .II I I I
I '0 I I '1 'I
II IIII III.I .I I .
o \ III I I II o .
I’I I
II -II I I
I I! ‘OIII IOII
'. II 'IIQI
|-III 0'- I '6
I II I II
I I I AIIIIIIf III III III
I II I I. II I'I'IIIIII III I I III
I I"III'|" .I II. I
u I I I, I|\.II I‘
II I III IIIII‘ I.’ I I ’I II I
r I I IIII IV4I‘II'IIfIIIII InfloI‘J m” I“ ”II
I“ ‘II II I7I‘I 'II‘ It'll I I
I .IIIu III IIIIIIII “III! I III,
I II I III III. (III. I III. I III!
. III II I. , III II I]
IWII... I .II" I
I IIIIIIINI .II IIIo»- IIIIIIIIfl II .
‘I’IIOIIIIII’II’I".II- I ,I II 11" ,
t’I‘ II I ‘13- if
I III II III; II I II.I.. IIIIIIIIJLIH.”
IIIIII IIIIIII IL III I .IIIIIIflIIII IIIII
I I? II III III) III“.
.IIIII IIIIIIIIIIIIIIIIIIIII.
III InInIIIII IIIHII I I
III II II I I III III , I .I III
(II |.II. .I HI H V. I
III, III I--. III I’II
. IIICW I I .o I. “I
I? I II I. II I I
..IIII..IU III III! I.
II. I. II‘IIIIv In. .I III
II £1ng IIIIIHI I I
.IIIIIIIIIIIIINTI
I. I I III II
I '

 

         

III.
(III!
II I
(7.,II

II 1 \MI II
‘ If
. II .“II I l l '
I I. II I H I III. I III II .
. . IIWHWI I I III I II I I. I I I I I III
«IV I: - - I - - u III. III... I--I.. -I.
IIIMII IIII I II, . I .I\ I. I. I II- I
I I I I I l I I - J I
.. I u “I, IIIIWIIIIIIIIIIIFII I [mLIIWIIIJHIIIII
IIIII IIIGUIIgnIIIIIIIIIIIHNIIII’III IIIIIIIIHIIIIIIIIIIIIIIIIIIIIIIIII. II I III; I . I
I III I I In“ I’ 'l- {L ”III... I. I
.IIIIIIIdWIIIIIIIIMIIIIH IIIII gfldfijflrhrf III II
“III III .I III III! IIIHI It I IIIIIIII Igfi .IIIIIIII.
if I . .IIIII‘II’JIIIJHIINIIJ‘I’ III’IIIIII III,"- I‘IlIi 'IIIIIIII’IIIII
IIIIIIIIIIIIIIIIIIIIIIIII‘I Ill-IIIIIIIIIIIIIIIIIIIIIIIIIIII . IIIII
IHIUuIIIIIIIIIIIIIIIIIIIIII
I!!! III II.II.I§I III: IAIIIIII I la»? V.

III- II III III.II l
. I”. I‘Llrl
I.
I II .I I II

           

I - .. .4 A -
\V
W -Q.$$£

$339
2%»
W
W l

‘3
(if. ‘n

_‘::‘T' \f'i‘ m2?-

 

 

This is to certify that the

dissertation entitled

THE CHARACTERIZATION OF NYLON OPEN—TUBULAR
IMMOBILIZED ENZYME REACTORS INCORPORATED IN
STOPPED-FLOW AND CONTINUOUS FLOW SYSTEMS

presented by

Robert Q. Thompson

has been accepted towards fulfillment
of the requirements for

PhoDo degree in ChemiStrX

 

 

War/£4

Major fiofessor

nae August 2, 1982

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

MSU J

LIBRARIES

 

“-

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record.~ FINES will
be charged if book is
returned after the date
stamped below.

 

 

 

 

 

 

THE CHARACTERIZATION OF NYLON OPEN-TUBULAR
IMMOBILIZED ENZYME REACTORS INCORPORATED IN
STOPPED-FLOW AND CONTINUOUS FLOW SYSTEMS

By

Robert Q. Thompson

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1982

ABSTRACT

THE CHARACTERIZATION OF NYLON OPEN-TUBULAR
IMMOBILIZED ENZYME REACTORS INCORPORATED IN
STOPPED—FLOW AND CONTINUOUS FLOW SYSTEMS

By

Robert Q. Thompson

Open-tubular immobilized enzyme reactors are becoming
widely useful in the clinical laboratory for specific and
sensitive substrate determinations. However, the dependence
of apparent enzyme activity on reactor design and mass
transport effects has not been well documented. Several
open-tubular immobilized enzyme reactors, prepared from
0.1 cm i.d. O-alkylated nylon tubing, were incorporated
into flow systems and characterized. The apparent activity
of an immobilized glucose oxidase reactor, used in a
miniature, 1.0 mm i.d., continuous flow system to deter-
mine glucose, was found to be independent of coiling
diameter and ionic strength, but was affected by tempera-
ture, segmentation gas composition, and the amount of
co—immobilized catalase. Sodium azide was found to inhibit

catalase, thereby improving the assay sensitivity, and to

inhibit glucose oxidase at higher concentrations.

To investigate the influence of mass transport on
immobilized enzyme kinetics, a novel stopped-flow instru-
ment was developed. The enzyme reactor was employed as a
photometric observation cell to allow direct and continuous
monitoring of the enzyme reaction. The kinetics, as mea-
sured under static conditions, are simpler than in other
flow systems because only molecular diffusion and/or
coulombic forces contribute to mass transport. Kinetics
constants for immobilized lactate dehydrogenase are
shown to be influenced by radial diffusion in the stopped-
flow instrument. The method is also applicable to clinical
substrate determinations. Glucose, in the range of 1—10 mM,
and lactate, in the range of 5-50 uM, were determined by
reaction-rate methods with the instrument.

A new enzyme reactor for use in flow-injection analysis
was also developed. Glucose oxidase was immobilized on
non-porous glass beads and the beads, 0.5 mm in diameter,
were packed into a 0.8 mm i.d. Teflon tube. The resulting
reactor was incorporated into a flow-injection manifold for
the determination of glucose. Calibration curves were very
linear over the range of 0.0—2.5 mM glucose, and at least

60 samples/hour could be processed.

To

Mom and Dad

and Jan

and all my other teachers

ii

ACKNOWLEDGMENTS

I would first like to thank Dr. Stanley R. Crouch
for his helpful advice and help with this and other
manuscripts. He also has good taste in baseball teams.
Special thanks also go to Dr. Gerald Babcock for serving
as my Second Reader, and Dr. Enke andDr. Wood for being
members of my Guidance Committee.

Charlie Patton deserves a lot of credit. The ”Bubble
King" taught me about continuous flow analysis and shared
with me a lot of ideas. He was a friend and helpmate in
the laboratory. Thanks Chas, I couldn't have done it
without you.

Thanks also go to Frank Curran for his friendship in
and out of the chemistry building (inspite of his poor
taste in baseball teams), and Gene Ratzlaff for loaned
equipment and the gift of helpful advice. All the Crouch
group members who lent a helping hand deserve thanks.

The times of play were also very valuable. I thank
all my basketball buddies, my golf partners—- Steve M.,
Steve B., and Pete, my tennis nemesis-— John Leckey, and my
good friends-- Rich and Jane Farmer.

Several support people deserve hearty thanks: Tom
Atkinson for MULPLOT, Dick Menke and Deak Waters, and the
graphics artists-- Katherine, Bev, and Jo. Thank you.

Finally, I thank my wife, Jan, for everything else.

iii

TABLE OF CONTENTS

INTRODUCTION .... ................... .............. 1

CHAPTER 1

Introduction to Immobilized Enzymes .............. 3
A. Immobilized Enzyme Characteristics ...... 3

B. The Uses of Immobilized Enzymes in
Analytical Chemistry ......... ..... ...... 6

CHAPTER 2
Chemical Bonding of Glucose Oxidase to Nylon Tubing 14

A. Nylon Activation ........................ 16

B. Addition of Spacer Molecules ..... ..... .. 25
C. Attachment of Glucose Oxidase ........... 31
CHAPTER 3

Kinetics Of OpGH-TUbUlar Reactors o o o o o o o I o o o o o o o o 37

A. Kinetics Properties of Open-Tubular
Reactors ............. ..... .............. 41

1. Electrostatic and environmental

effects ... ......... ....... .......... 41
2. Mass transport effects .............. 43
B. Experimental Considerations ............. 51

C. The Determination of Michaelis Constants. 57

1. Introduction ........ ..... .... ....... 57
2. Reagent preparation ................. 59
3. Immobilization of lactate dehydro-

genase ........ ...... ..... ..... ...... 60
4. The instrumentation ..... ....... ..... 60
5. Results and discussion .............. 62
D. Analytical Determinations ............... 7O
1. Introduction ........................ 7O
2. Reagent preparation ................. 71
iv

E.

CHAPTER 4

3.
4.
5.

6.

Enzyme immobilization ......... . .....
The instrumentation .... ..... ........
The results of the glucose

determination ............ ..... ......

The results of the lactate
determination ......OOOOOOOOOOOOOOOOO

A New Cell / Reactor Design .............

Colorimetric Determination of Glucose Based on
Immobilized Glucose Oxidase ......................

A.

The

1.
2.
3.
4.

Method and Instrumentation ..........

The glucose oxidase reaction ........
The Trinder/peroxidase reaction .....
The continuous flow manifold ........
The continuous flow instrument ......

Experimental Studies of the Trinder

Reaction ......OOOOOOOOOOO.....OOIOOOOOO.
Reaction Characteristics ... ........ .....
1. Segmentation gas composition ........
2. The effect of ionic strength ........
3. The specificity of glucose oxidase ..
4. Mutarotation kinetics ...............
Reactor Design Characteristics ..... .....
1. Coiling diameter .... ............ ....
2. Coil temperature ....................
The Characterization and Elimination of
the Effect Of catalase ......OOOOOOOOOOO.
1. The nature of the interference ......
2. The reagents and continuous flow
instrument I...OOOOOOOOOOIOOOOOOOOOOO
3. Enzyme immobilization ...............
4. Effect of azide on aqueous peroxidase
5. Effect of azide on immobilized
catalase ......OOOOOOOOOOOOOOOO0.0...
6. Effect of azide on immobilized
glucose oxidase .....................
7. Effect of azide on different
preparations of glucose oxidase .....
8. A method to determine the presence
of catalase ..... ....... ...... .......
9. Conclusions .............. ...........

73
73
77

81

90

97

97

98

99
105
107
114
114
115
118
121
124
124
125
127
127
130
131
131
133
137
140

141
143

CHAPTER 5
Enzyme Immobilization on Non-porous Glass ........ 147

A. Immobilization of Glucose Oxidase
on Glass Tubing ......................... 148

B. Glucose Oxidase Bead String Reactor ..... 150

CHAPTER 6
Future Plans .......... ..................... . ..... 161

APPENDICES

Appendix A. Derivation of Diffusion-control
Rate Equation ..... ............. 164

Appendix B. Derivation of the Absorbance of
a Solution with a Radial
Concentration Gradient ......... 165

LIST OF REFERENCES .......... ........ . ....... ..... 166

vi

LIST OF TABLES

Table 1.1. Recent Examples of Enzyme
Electrodes............................. ........... 8

Table 1.2. Recent Examples of Open-
Tubular Reactors............... ................... 11

Table 1.3. Packed—Bed and Other
ReactorSOOOOOO0.00.00.00.000............OOOOOOOOOO 12

Table 2.1. The Results of the Spacer
Molecule Studies................... ......... . ..... 26

Table 2.2. The Activities of Nylon

Tubular Reactors Resulting From the

Incorporation of Various Spacer
Molecules....................... ....... ........... 29

Table 2.3. Immobilization Procedure ............. . 35
Table 2.4. Various Enzymes Bonded to
Nylon Tubing and Their Approximate

Half-liveSOOOOOOOOO....OOOOOOOOOO ....... o ..... 0.0. 36

Table 3.1. Reviews of Immobilized
Enzyme Kinetics...................... ............. 38

Table 3.2. Kinetics Values for Aqueous
and Immobilized Lactate Dehydrogenase... .......... 68

Table 3.3. Results of Serum Assays
for Glucose....................................... 76

Table 3.4. Results of Serum Assays
for Lactate....................... .......... . ..... 80

Table 4.1. Chromogens for Reaction
with Hydrogen Peroxide ..... ......... .............. 95

Table 4.2. Specificity of Glucose
Oxidase.. ....... . ........... . ..................... 120

vii

Table 4.3. Solutions Used to Prepare
the Immobilized Enzyme Reactors.... ............... 132

Table 4.4. Glucose Concentrations
Found in Control Sera With and Without
Added Azide....................................... 146

Table 5.1. Immobilization Procedure
for Glass Beads................. .......... ........ 152

viii

LIST OF FIGURES

Figure 1.1. Schematic of an instrument
employing an open-tubular reactor..................

Figure 2.1. The immobilization scheme.............

Figure 2.2. Reactor activity versus
TOTFB reaction time.................. ..............

Figure 2.3. Reactor activity versus reactor

length for different TOTFB concentrations.

The concentrations were 0.1 M (A), 0.05 M (B),

0.01 M (C), and 0.005 M (D)............... .........

Figure 2.4. The precision of the
immobilization procedure.......... .............. ...

Figure 2.5. The storage stability of reactors
prepared with different spacer molecules...........

Figure 2.6. Fluorescamine determination of
1’6-diamin0hexane........0...0.0......0. ..... 0.0...

Figure 2.7. The effect of initial glucose
oxidase concentration on the reactor
actiVity.......0...0....................0....00....

Figure 3.1. Immobilized enzyme kinetics...........

Figure 3.2. Influence of slow mass transport
on enzyme kinetics......0 ........ 0.0....0.. ........

Figure 3.3. Concentration gradients in an
Open-tUbUIar reactoroooooooo ooooooooo 000000000 ooooo

Figure 3.4. Bolus flow pattern ....................

ix

15

20

22

24

27

30

33

39

45

47

50

Figure 3.5. Cross-sectional View of the
GCA McPherson stopped-flow mixer/observa-
tion cell unit. The outer aluminum housing
(a), quartz windows (b), and the main KEL-F
body (0) are press fit with three bolts.
Mixing occurs at (e), where the two reagent
flow streams meet at 90° to each other --
one stream is in the plane of the figure
and the other is perpendicular to it.

The immobilized enzyme reactor is placed
inside the observation cell (d). The
dashed arrow represents the light path
inside the cell...................... ......... .... 55

Figure 3.6. Stopped-flow colorimeter... .......... 56

Figure 3.7. Typical progress curves ob-
tained with the instrument. The dotted
lines represent the initial reaction rates
calculated from these curves. The reactions
represented were of 2.00 mM lactate with
0.425 mM (A), 0.850 mM (B), and 2.12 mM (C)

B-NADOC00.00.000.000....0....0...00............... 63

Figure 3.8. Lineweaver-Burk plot of
l/Rate versus 1/[Lactate] at four differ—
ent B—NAD concentrations: 0.26mM (A),
0.43 mM (B), 0.85 mM (C), and 2.12 mM

(D)............................................... 64

Figure 3.9. Lineweaver—Burk plot of

1/Rate versus 1/[B-NAD] at three differ-

ent lactate concentrations: 0.80 mM (A),

2.00 mM (B), and 4.00 mM (C)........... ..... ...... 65

Figure 3.10. Plot of y-intercepts from

Figures 3.8 and 3.9 versus 1/[Substrate]

for different B—NAD and lactate concen-

trations. The y-intercept equals 1/Vm,

and the x-intercepts yield the two Km
values.................................. ...... .... 67

Figure 3.11. Calibration curve for
glucose..................0...... 0000000000 . ..... .0 75

Figure 3.12. Calibration curve for
lactate........ ..... . ........ ........ ............. 78

Figure 3.13. Design of the new cell/

reactor. Two pneumatically actuated

pistons (A) were used to push the slider

(B) back and forth between the fill posi-

tion (P1) and the Viewing position (P2).

Quartz windows were used (C) to contain

the solution in the viewing position...............

Figure 3.14. Reproducibility of cell

positioning. Distilled water was

pumped through the cell at 0.46 mL/min.

Each trial represents movement of the

cell from the viewing position to the

fill position and back again.......................

Figure 3.15. The results of five pro-

gress curve determinations with 0.44 mM

glucose as the sample. The 1 mm i.d.

reactor was used.................... ..... ..........

Figure 3.16. Progress curves obtained

with the 1 mm i.d. reactor. The glucose
concentrations (mg/L) which were used

are written at the end of each curve. ....... .......

Figure 3.17. Progress curves obtained

with the 2 mm i.d. reactor. The glucose
concentrations (mg/L) which were used

are written at the end of each curve...............

Figure 4.1. The three forms of D-glucose..........

Figure 4.2. The reactions involved in
the enzymatic assays for glucose. .............. ....

Figure 4.3. The continuous flow reaction
manifold. The pump speed setting which
gave nominal flow rates was 42......... ..... . ......

Figure 4.4. The pH dependence of the

glucose assay. A 0.1 M phosphate buffer

was used to control the pH. The dotted line
indicates the approximate relationship in the
absence of many data points........................

Figure 4.5. Diagram of the continuous
flow instrument.0 .......... ......0........0..0.....

xi

82

84

86

87

88

91

93

103

106

Figure 4.5. Progress curves for the

reaction of hydrogen peroxide and the

Trinder reagent. The peroxide concen—

trations were 50 uM (A), 25 uM (B),

10 uM (C), 5 uM (D)...................... ..... .... 109

Figure 4.7. The effect of ascorbate
concentration on the assay sensitivity ...... ...... 112

Figure 4.8. First order plots of data

for the hydrogen peroxide-Trinder

reaction in the presence of ascorbate.

The ascorbate concentrations were

10 HM (A), 20 MM (B), 40 HM (C).

A was the absorbance at equilibrium,

afid A was the absorbance at any

time, t........................... .......... ...... 113

Figure 4.9. Calibration curves for a
series of glucose standards using three
different segmentation gases: NZ, Air,

020.00.00.000000000000......0000000000.000....0... 116

Figure 4.10. The output from the con-

tinuous flow instrument using three

different segmentation gases. The

glucose concentrations were 0.44 mM,

1.44 mM, 2.22 mM, and 4.44 mM. Only

the first three concentrations were

sampled using 02.................................. 117

Figure 4.11. The effect of ionic
strength on the glucose oxidase
reaction000.....00000.00.00.000.....0.....0......0119

Figure 4.12 First order plot of data

for the mutarotation of 4.44 mM

B-D-glucose solution at 22°C. The

results of two independent experiments

are presented. [8] is the concentration

of the B-form of D—glucose at any time,

t, while [B] is the initial B-form
concentratiog....................... ............ .. 123

Figure 4.13. The effect of coiling

diameter on the calibration slope (0)
and wash ([3).......... ............ .. ..... ........ 126

xii

Figure 4.14. The dependence of the assay
sensitivity on temperature....... ..... . ........... 128

Figure 4.15. The effect of azide on the
amount of hydrogen peroxide that was con-
verted to product. A 79 nM H20 solution
was sampled without an enzyme rgactor
([J) and with a 30 cm catalase reactor

(0) in the manif01d00000000.00.00.00.0000000000000 134

Figure 4.16. Lineweaver-Burk plot of the

data which exhibited azide inhibition of

immobilized catalase. The azide concen-

trations ranged from 0.0 (A) to 1.5 mM (B)........ 136

Figure 4.17. The effect of azide on three

glucose oxidase reactors, type B (A),

type C (0), and type D (EJ), containing

various amounts of catalase (refer to

Table 4.3)........................................ 138

Figure 4.18. Dixon plot of data from

Figure 4.17, curve B. Three glucose

concentrations were used: 0.44 mM (1),

0.22 mM (2), and 0.11 mM (3)...................... 139

Figure 4.19. The effect of azide on

three commercial preparations of glucose

oxidase, reactor types G, E, and F

(refer to Table 4.3)............... ......... ...... 142

Figure 4.20. The change in calibration

slope with different lengths of reactors

B, C, and D (refer to Table 4.3). No

azide was added....... ........ ............. ....... 144

Figure 4.21. The change in calibration

slope with different lengths of reactors

B, C, and D (refer to Table 4.3). The

reactions occurred in the presence of

0.8 mM azide...................................... 145

Figure 5.1. View of the immobilized
enzyme single bead string reactor..... ............ 153

Figure 5.2. Schematic diagram of the
flow-injection instrument......................... 155

xiii

Figure 5.3. The degree of incomplete

mixing at various sample volumes. Four
pump speeds were used: 8.3, 12.3, 25.0
and 33.3 uL/sec........................

Figure 5.4. Flow-injection instrument
response. The four glucose concentra-
tions were: 0.44 (A), 0.88 (B), 1.33

(C), and 2.22 (D) mM...................

Figure 5.5. Calibration curve for the

flow-injection determination of glucose.

xiv

..... ...... 159

.......... 160

LIST OF SYMBOLS

AU - Absorbance units

AO - initial absorbance (AU)

Aco - equilibrium absorbance (AU)

- wash value (seconds)

- ratio of mean ionic activity coefficients (unitless)
— a constant (unitless)

reactor diameter (cm)

- substrate diffusion coefficient (cmZ/sec)

(DUO-000‘
I

- charge on the electron (coulombs)

[E]o - total enzyme concentration (Molar) 1

h - mass transport rate constant (second )
H(ss) - height of CF peak at flat portion (AU)
H(t) — height of CF peak at any time, t (AU)

js - Bessel functions (unitless)

k - Boltzmann constant (joules °K-1)

ke - an enzyme rate constant (second-1)
Ki - enzyme interaction constant (Molar)
KI — enzyme inhibition constant (Molar)
Km - Michaelis constant (Molar)

L — reactor length (cm)

Nu - Nusselt number

[P]b - bulk product concentration (Molar)

[P]w - product concentration at the reactor wall (Molar)
r - reactor radius

[S]b - bulk substrate concentration (Molar)

[S]W - substrate concentration at reactor wall (Molar)

t - time (seconds)

T - temperature (°K)

v — initial reaction rate (Molar second-1)

V — volume of a single liquid segment (cm3)

\hn - maximum rate of an enzyme reaction (Molar second-1)

XV

electrovalency of the substrate (unitless)

effective concentration of charges on the carrier (M)
. . —1 —1

molar absorpt1v1ty (AU Molar cm )

ratio of the mass transport rate constants times
the ratio of Michaelis constants of the two substrates
for the same enzyme (unitless)

liquid viscosity (poise)
surface tension of a liquid (dyne cm-l)

electrostatic potential between the carrier
and the substrate (volts)

diffusion layer thickness (centimeters)

ratio of the enzyme rate constant (V /K ) to the
mass transport rate constant (h) (ugitTess)

liquid flow rate (cm sec-1)

xvi

INTRODUCTION

The concept of passing a reagent solution through an
open tube to which an enzyme is attached has led to specific,
sensitive, and fast determinations of clinically important
compounds. The use of immobilized enzymes with continuous
flow instruments is an important and growing area. This
research is a part of the effort to improve such systems.
The goals were to optimize the procedure for binding enzymes
to nylon tubing and then to characterize the resulting
reactors in terms of their physical and chemical properties.

Glucose oxidase was chosen as the major enzyme of
study, because it is relatively inexpensive, highly active,
and bonds well to nylon. Lactate dehydrogenase, catalase,
and other enzymes were used in immobilized form to a
lesser extent.

The enzyme activities were measured by incorporating
the open-tubular reactors into a continuous flow or
stopped-flow system. The continuous flow technique involves
a flowing liquid stream in which the sample is injected,
mixed with reagents, and detected. When gas bubbles are
added to the liquid stream the technique is termed gas—
segmented, continuous flow analysis (CF) [1-3]. Without
{gas segments, the technique is called flow-injection

analysis (FI) [4-6]. The major interest in this research

has been with CF systems. The stopped-flow (SF) technique
[7—9] involves the rapid mixing of reagents and sample,
followed by the rapid cessation of the flow. The reaction
mixture is most often monitored by spectrophotometry.
These flow techniques are not discussed in detail; only
important points pertinent to this research are mentioned.
Informative discussions can be found in the references
given above.

After a brief introduction to immobilized enzymes
(Chapter 1), the optimization of the nylon immobilization
procedure is presented (Chapter 2). The kinetics proper-
ties of enzyme reactors are discussed in Chapter 3. A
novel stopped—flow instrument is described which should
help in learning more about the inherent properties of
immobilized enzymes. In Chapter 4, a glucose oxidase
reactor, incorporated into a CF manifold, is character-
ized. The influence of catalase impurities on the deter-
mination of glucose was studied. In Chapter 5, the
possibility of replacing nylon with non-porous glass is
explored, and finally, suggestions for future projects
are given (Chapter 6).

This research project has provided several surprises
and many interesting results. The work is mostly experi-
mental and practical. Further work should provide a more
mathematical description of the open-tubular immobilized

enzyme reactors (OTIMERS).

CHAPTER 1

Introduction to Immobilized Enzymes

The interest and activity in the field of immobilized
enzymes has been high for the past fifteen years. Hundreds
of papers have appeared in the literature extolling the use
of enzymes attached to an insoluble support. Many excellent
reviews of the theory and practice of immobilized enzyme
technology are available. At least nine books have been
published in this area [10-18], with two of them appearing
in 1980; several other books have chapters dealing with
immobilized enzymes [19,20]. Furthermore, many reviews
on the uses of immobilized enzymes have appeared in the
scientific literature [21-25].

Since the research in this area has been well—docu—
mented, only the fundamental concepts are presented here.
Section A introduces bound enzymes and the factors which
may alter the properties of the enzyme. Section B reviews
the uses of the immobilized enzymes in analytical chemistry
and introduces the concept of open-tubular reactors in flow

systems.

A. Immobilized Enzyme Characteristics

An enzyme is a protein which can catalyze one or a
small set of reactions. Thus, the enzyme is specific and
very sensitive toward a few compounds and is useful as an

analytical probe. Confining an enzyme to an insoluble

3

support provides the further advantages of longer life-
time and easier manipulation. The enzyme reactor can be
used again and again. Thus, immobilized enzymes lower the
time and costs of employing biological catalysts.

The enzyme can be confined or attached to a carrier
in four basic ways. (1) Adsorption [13] occurs to some
extent in all immobilization procedures. This method
includes attachment by coulombic attraction and other short
range forces. The procedure is the mildest and results in
the highest percentage of active bound enzyme. However,
the bond is weak, and the enzyme is gradually lost from the
carrier with use. This method is not frequently used for
analytical applications.

(2) Encapsulation is really an indirect method of
immobilization. A solution or slurry containing the enzyme
is simply contained inside a semipermeable membrane. The
technique is widely used to prepare enzyme electrodes [22]
(See Section B). Substrate molecules must diffuse through
the membrane in order for the reaction to occur. Thus,
the membrane barrier limits the size of molecules that can
react and limits the rate at which the enzyme can function.

(3) Entrapment occurs when a gel is made to polymerize
in the presence of an enzyme [13]. The enzyme molecule is
trapped inside the empty spaces of the gel. Similarly, a
:fiber stretched in a solution of enzyme will retain some

(pf the enzyme in between the layers of fiber. Like

encapsulated enzymes, enzymes trapped in a matrix are limited
by substrate transport.

(4) Covalent bonding of the enzyme to a carrier is the
most widely used immobilization technique in analytical
chemistry [12]. A chemical bond is formed between the
carrier surface and the enzyme molecule. The reactor that
results shows good temperature and time stability. The
enzyme is not easily stripped from the carrier by a flowing
liquid, so the technique is chosen in preparing open—tubular
reactors. However, the percentage of active, bound enzyme
is low. The reason for this is that no technique has yet
been developed to single out a site on the enzyme for
attachment. Thus, enzyme bound at the active site becomes
inactive. Likewise, multiple bonds to the same enzyme
molecule can change its tertiary structure, affecting its
activity. Covalent bonding of enzymes is still rather empir-
ical and not yet based on solid fundamental grounds.

An enzyme covalently bound to a carrier experiences
much of the carrier environment in addition to that of the
bulk solution. The hydrophilicity or hydrophobicity, charge,
and chemical reactivity of the support all influence the
enzyme properties. The substrate affinity for the carrier
and the enzyme is also affected. The enzyme is not always
in intimate contact with the substrate as it is in solution;
'thus, substrate transport from the bulk solution to the

<3arrier surface influences the reaction rate. Changes in

pH optima and enzyme kinetics are discussed in more detail
in Chapter 3. Thus, it is expected that the attached enzyme
would display different properties from those of an enzyme

in aqueous solution.

B. The Use of Immobilized Enzymes in Analytical Chemistry

Enzymes in aqueous solution have been used for
thousands of years to produce beer, cheese, and other
products. Similarly, industrial processes were among the
first to use immobilized enzymes. Several reviews of the
use of enzymes in industry have been written [26-29]. The
lower costs that result are making bound enzymes popular.

In analytical chemistry, enzymes have been transformed
from aqueous reagents to integral parts of chemical instru-
mentation. The immobilized enzyme is usually a part of
the transducer or manifold of the instrument. The chemist
does not need to prepare enzyme solutions each day, but
instead must maintain the activity of the enzyme. Storage
of the enzyme at reduced temperature, in the dark, and
immersed in a buffer at neutral pH is usually all that is
required. The job of the chemist is simplified by using
immobilized enzymes.

Immobilized enzyme instrumentation can be divided into
three major areas: enzyme electrodes, open-tubular reactors,
and miscellaneous methods. Enzyme electrodes are prepared
by confining an enzyme solution or suspension to an area

adjacent to the surface of an ion-selective electrode.

A semipermeable membrane, stretched over the enzyme layer
and the electrode, separates the enzyme from the reagent
solution. Substrate molecules diffuse into the catalytic
layer, the reaction occurs, and the product concentration
is sensed by the electrode. Potentiometric and amperometric
sensing are the usual means of detection. Enzyme electrodes
do not include systems in which the electrochemical detector
is separated from and subsequent to the immobilized enzyme
reaction. The general reviews mentioned previously and
several specific reviews [30-32] show the wide use of
enzyme electrodes. Recent reports, listed in Table 1.1,
are further evidence of the importance of these methods.
Open-tubular reactors are even more widely used. A
general schematic of a typical instrument is displayed in
Figure 1.1. Open tubes allow low pressure pumping and
preserve the integrity of the flow profile which are
required in the analytical flow methods. Delivery of the
reagents is usually performed by a peristaltic pump (CF
and FI) or a pneumatically-driven syringe (SF). The mixed
reagents are pushed into the reactor, where they are
allowed to react for a short period of time. The reaction
time is a compromise between sample throughput and
acceptable instrumental response. The reaction products
are usually measured by absorption spectroscopy, but can
also be monitored electrochemically, thermometrically, or

lay luminescence spectroscopy.

mm

mm

mm

vm

mm

.mmm

cappoEOMoQE<

oflgpmsofipcopoa

ofihpoeouode<

cappmeonmae¢

QHHpoEOAmQE<

AOpoopoQ

canspaaam

mowom ocfiswlq

Hosanna

me<

Hosanna

mpwfidc<

moUOMpomHm

capafimo

mochpooao
opfinmmhw Op
@202 usofia>ov

How encamcdom

How
moHEdazgoahaom

:oflpsaom mzoosv<

 

xflgpmz msmNcm

mszucm mo moHQmem

mmmadpao

mmmoflxOMoQ
mmaofixo mmOUSHw
I®E>Ncmflge .m

U

omapflxo
How ocfls< .v

ommcowonozcoo

HOSOOH¢ .m

mmacwxoxmn
omwoflxo mmoosaw

IoE>Nc®Hm .N

omwsowouomnoo

pcooom

Hosoofi< .H

 

vamEmNcm

.H.H mange

 

 

 

 

 

 

 

 

 

 

 

 

Rquent
Dehvery > Mixer
Systenw
fr,» Enzyme
Reactor
[ DetectorH Woste
Figure 1.1. Schematic of an instrument employing an

open-tubular reactor.

10

In most cases, nylon tubing has been used as the
carrier. It is inert, sturdy, easily molded into a helix,
and readily activated for subsequent enzyme immobilization.
The only disadvantages are that it has a small surface
area (few sites for enzyme attachment), and it tends to
adsorb organic compounds, making sample carryover a problem.
Despite these problems, no other carrier has replaced nylon
for analytical use. Table 1.2 lists some of the more recent
uses of nylon open-tubular reactors.

Several other analytical reactors have been reported
in the literature. A few of these are listed in Table 1.3.
The most common type is a packed bed reactor. The enzyme
is covalently bound to porous glass beads, and the beads
are packed into a tube. Thermometric detection is common
with these systems. More unusual instrumentation includes
a nylon tubular reactor connected to a pipette tip and a
stirring bar coated with an enzyme.

Immobilized enzymes have been termed a "solution in
search of a problem”. Indeed, the potential usefulness of
immobilized enzymes has not been reached. Only a few
enzymes of high inherent activity have been used success-
fully. In order for more and better enzyme reactors to
be developed, the following needs must be met. The
carrier environment needs to be modified so that it is
there conducive to enzyme attachment. Also, the chemist

zaeeds to be able to choose the site of attachment on the

11

 

Ne mocmnpomnw >3
av cosmopomnw >D
ow moonpomao wfisoEE¢
mm mocmnpomnw >3
mm oGOAHQoao cowmxo
mocogommm mammwmwm

mAOHOHom swasnseucoao

H0pmozao

moflgmozawwhe

ocflcflpmopo

eaom can:

Hosanpm

mpwflaa<

coapmfisxfia
mmeoe

eoflpafisxflm
mmeoe

ceapafisxfie
mmeoe

mHm%HOHU>:

 

seawa>apo<

Ho mmHQEMxm pcmomm

omncomoucmnoo
Honmo>ao .m

omwcowOAozsoc
Honookao .v

ommcficflpmmpo .m

.eszme masseuse
ommOHps
I®E>Ncoflm .N

ommoflxo
HOQOUH< .H

 

Amvosmmcm

.N.H manwe

12

man magmafia
a Op conomppa

om OHHpQEOHOQQoupoon omwoua many 20H>z .o
av ofihposoaonQOMHOon omwcom0pc>noo Honooa< Han mafiaaflpm .m
wv QHQHoEoHpcmpom omauozpop mamupflz .H mm oEmm .w

omacmw0hozsmo mpwpomq
be oflppoEOAoQE¢ omwcowoupznmp Hosoofi< .H mm osmm .m

ow ofippoEonoo ommohb .H mm oEmm .m

owmofixo Houmpmoaozo
mmmpflxo mmoosfio :ESHoo m as coxoma

 

mvlmv capmeOEnoze enema: mmwam machom .H
mosopmmom acuoopom oewncm yoflnpwo

myopoemm nmspo new nomucmxomm .m.H manna

13

enzyme. Furthermore, if diffusional limitations and
electrostatic effects could be eliminated, the use of
immobilized enzymes would be simpler since the large

body of soluble enzyme data would be directly applicable.

CHAPTER 2

Chemical Bonding of Glucose Oxidase to Nylon Tubing

The goal of immobilization is to produce an enzyme
reactor with the highest possible activity. This means
that many sites on the carrier must be available for
attachment, and that the enzyme must be bonded in its most
active form. With nylon, four chemical steps are employed
to produce the desired results, as shown in Figure 2.1.
First, the inert, amide polymer is treated so that a reactive
group is formed. Three different activators have been used,
and these are discussed in Section A. Next, a bifunctional
compound is added. This "spacer” molecule increases the
distance between the nylon surface and the enzyme. Therefore,
the enzyme experiences less of the carrier environment and
more of the bulk solution properties. Diffusional and
steric limitations are reduced to some extent by adding the
spacer. In Section B, the use of diamines as spacers is
described.

Third, enzyme attachment is accomplished, usually via
free amino groups (lysine residues or N-terminus) or free
carboxyl groups (aspartic acid, glutamic acid, or C-terminus).
In most cases, bonding through amino groups yields higher
activities than bonding to other sites. This may be due
to participation of carboxyl groups in the active site.

Thus, glutaraldehyde is often used to link the amine group

14

15

.msonom :ofipaNHHHQOEEH one

: Amasmmoq<m<eafioc
_
mom + mzsmzmuzuouxmmHmm<oc + mszzmuzmm + Ammoammv
.
Azoqwzv
m
—,
muwumxmmovuoui
o Ammo<mmv
_ a
mom + Azoqszv + muw-cimmoc-om + Ammo<mmc
_ __
o o Azoqwzc
mmz-:xmmoc-zm mmo<mm
.
C
acmmmo + Azoqszv + mmz- Ammoc-2m: + ammow
Azoqszv
ammoo mmeoe zoqsz

036mm“: + «mm + x% u 8 +

_
m

X
mm+0mimzmov + Az - oc
_ __
I 0

v

.H.m opzwam

“v ampm

"m ampm

"m ampm

"H ampm

16

on the spacer molecule and the enzyme. Such procedures are
examined in Section C. Since glucose oxidase and other
enzymes are glycoproteins, Zaborsky and Ogletree [51] have
suggested that attachment can be performed via the carbo-
hydrate residues. This method may prevent loss of activity
due to bonding at the active site. Another suggestion for
improving the activity of the reactor is to bond the enzyme
in the presence of its substrate or a reversible inhibitor.
Again the active site should be protected from chemical
bonding. In spite of their theoretical merits, these latter
two methods have not gained wide acceptance.

The immobilization of glucose oxidase is examined here.
Activity measurements were made by incorporating the reactor
in a continuous flow manifold and monitoring the amount of
product formed. Glucose standards were reacted with the
enzyme and a calibration curve was prepared. The slope of
the calibration curve was used as the measure of reactor
activity. Details of the instrumental method are given in

Chapter 4.

A. Nylon Activation

A few years ago, hydrolysis of the amide linkage of
nylon was the major means of activation [52]. Free amine
groups were produced for subsequent bonding. First, a
calcium chloride solution was pushed through the tubing to
remove amorphous nylon and to etch the inner wall. Then

a dilute (about 4 M) solution of HCl was allowed to react

17

in the tube at 45 °C for 40 minutes. Washing with water
prepared the tube for subsequent reactions. The advantages
of this method were that a carrier surface free of charges
was prepared and the surface was well pitted to increase the
number of attachment sites. The major disadvantage was that
conditions had to be carefully controlled in order to
prevent the loss of tube wall strength. If the reaction
times were too long or acid concentrations too high, holes
could form in the tubing. The procedure required the use

of strong acids and was messy. The tubing often became
clogged with nylon particles, released by the hydrolysis
treatment. In conclusion, this procedure was expensive

in terms of time and materials.

In 1975, Morris, e£.al. [53] first described the
activation of nylon by O-alkylation. Two compounds,
dimethylsulfate (DMS) and triethyloxonium tetrafluoroborate
(TOTFB), were introduced as reagents that formed secondary
imidate groups with the amide groups on nylon. The
resulting structure was very reactive toward amines. Thus,
spacer molecules were easily added to the nylon.

The procedure employing dimethylsulfate has several
severe disadvantages. First, the reagent gives off poison—
ous vapors and causes burns; it must be used in undiluted
form. Second, the reaction must be carried out at 100 °C.
Third, the amount of enzyme immobilized with this reagent

is small. Despite these disadvantages, this method of

18

activation is still widely used.

The advantage of the TOTFB method is its simplicity.
The reagent is commercially available (Aldrich Chemical Co.),
and although it should be stored under nitrogen, it can be
used safely in the laboratory atmosphere. The diluted
reagent is perfused through the nylon tubing for several
minutes at room temperature, and then the tube is washed
with cold methanol and water. Clogging is rarely a problem.
Excess TOTFB can be neutralized with water, which yields
ethanol and other harmless products. The procedure does
impart to the nylon a positive charge, which may affect the
enzyme properties. Also, only a small surface area is
produced due to the mild conditions. Nevertheless, this
is the method of choice in many cases, because it preserves
the integrity of the tubing and the number of active sites
is adequate to produce analytically useful reactors.

The complete immobilization scheme is shown in Figure 2.1.
The first two steps are of concern here. A 0.1 M solution
of TOTFB in methylene chloride is usually employed. The
amide is converted to an imidate, with the concomitant
development of a positive charge on the amide nitrogen. The
amide bond is not broken, however, and this preserves the
nylon structure. The reaction of an amine group with the
activated nylon (step 2) is rapid, and a C-N bond is formed
between the amide carbon atom and a free amine nitrogen

on the spacer molecule.

19

The amidine formation from the nylon backbone has
been examined closely by Sundaram [54]. He determined that
the stability of the C-N bond is rather poor. Exchange with
proteins as well as small nucleophiles occurs readily.
Also, attempts to improve the stability of the bond by reduc-
tion of the C=N amide bond and concomitant removal of the
positive charge with sodium borohydride was unsuccessful.
Thus, nylon reactors prepared in the above fashion should
not be used as extracorporeal shunts in clinical applications.
The slow loss of enzyme from such reactors could be due to
the lability of the amidine bond. The lifetime of most
reactors, nevertheless, is long.

A study of the TOTFB reaction with nylon was performed
with the help of an undergraduate, Michael A. Schneider.
We studied the precision of the method as well as the effects
of reaction time and reagent concentration on the resulting
activity. The reactors were prepared in identical fashion
in steps two, three, and four, but the treatment in step 1
was varied. The volume of TOTFB solution used was 10 mL,
and the treatment was performed at room temperature. The
nylon tubing (Type '6', 1 mm i.d., Portex Ltd., Hythe, Kent,
U.K.) was all taken from the same lot.

Using a 0.06 M solution of TOTFB in dichloromethane,
the reaction time was varied from five minutes to thirteen
minutes. Figure 2.2 shows the effect of reaction time on

the activity of 25 cm reactors. No trend was apparent,

20

Calibration Slope (x102 AU M‘1)
.p.
I
I

 

Exposure Thne (nflnutes)

Figure 2.2. Reactor activity versus TOTFB reaction time.

21

and the treatments for five and thirteen minutes yielded

the same activities. Since the total amount of TOTFB was
constant, it appears that the reaction occurred rapidly

and was limited by either the amount of TOTFB or the number
of sites on the nylon. With a 0.06 M solution, the duration
of the treatment was not critical. The use of the shortest
possible reaction time, about five minutes, is desirable.

In the next study, the optimum concentration of TOTFB
was determined. Coiled, 100 cm lengths of nylon tubing
were treated for 5 minutes with TOTFB solutions in the con-
centration range, 0.005 M - 0.1 M. The tubing was then cut
into four 25 cm segments, and each coil was assayed with the
continuous flow instrument. Figure 2.3 displays the results.
The x-axis represents the distance along the nylon tube with
zero being the end in which the TOTFB solution entered and
100 representing the exit end. Each curve contains four
data points, representing the four segments of tubing. The
first segment's data are plotted at 12.5 cm, the second's
at 37.5 cm, and so on, displaying the mean activity over
that 25 cm of reactor.

For the lowest concentration, Figure 2.3 shows that
almost all of the TOTFB was reacted with the first 50 cm of
nylon; the TOTFB was the limiting reagent. With a concen-
tration of 0.01 M, the TOTFB reacted beyond 75 cm, while
tile 0.05 M solution provided similar numbers of active sites

alxong the entire length of tubing. Thus, with reactors up

22

 

 

AST—
‘ .
:2 Z::::::::8::::===~e=:::::::: 1A
B
34——
<
N
O
23——
V
O)
O.
22——
(f)
C
2 C
+4 _—
o1
.0
B D
1 I 1 I 1 J
00 T I I I I l T l
o 25 50 75 100

Figure 2.3.

Length (cm)

Reactor activity versus reactor length for
different TOTFB concentrations. The
concentrations were 0.1 M (A), 0.05 M (B),
0.01 M (C), and 0.005 M (D).

23

to one meter in length, the TOTFB concentration should be

at least 0.05 M. An increase to 0.1 M did not improve the
final reactor activity very much. This could have resulted
from nylon becoming the limiting reagent. A second expla-
nation is that more reactive sites were developed on the
nylon, but the enzyme bonding in active form was limited by
steric or other factors. Since lengths longer than 100 cm
were often used in our studies, a concentration of 0.1 M was
chosen. This required that the reaction time be limited to
3-4 minutes to prevent clogging of the tube. Likewise, it
is recommended that the nylon tubing be kept under 1.5 meters
in length. With higher concentrations of TOTFB, one end
will be over-reacted and clog, while the other end of the
tubing will not be activated enough.

Using a concentration of 0.1 M and a reaction time of
three minutes, the precision of the method was tested. Again,
a 100 cm length of tubing was treated, cut into 25 cm seg-
ments, and tested. The results of two independent experi-
ments are shown in Figure 2.4. The variation in activity
along the reactor length was found to be 2.9% and 4.8%
(relative standard deviation). The variation between
immobilization experiments was 6.4%. Similar results are
also shown in Figure 2.3 for the highest two concentrations
of TOTFB. Thus, the precision of the step 1 procedure is
good. Another interesting fact was that for the experiments

mentioned above, the second 25 cm segment always showed the

24

 

 

 

8.0 --'
£5 __ Gt Or A
I \
2 G//—6\\‘9
:3 6.0 ~--
<
N
o ....
3'.
q, 4 o ——
0.
2
(f) __
c
2
*5 2.0 ——
fl
'5 -..
O

l l I l l l J
0.0 l g I I l I I l
0 25 50 75 100

Length (cm)

Figure 2.4. The precision of the immobilization procedure.

25

highest activity. The reason for this peculiar trait is

not known.

B. Addition of Spacer Molecules

Several diamines of the formula, HzN-(CH2)N-NH were

2.
tested as spacer molecules. A diamine concentration of

0.1 M in a sodium bicarbonate buffer (0.1 M, pH 9.4) was
used in all experiments. The effects of amine concentration
and solution pH on the reactor activity were not examined.
After activation of the nylon, the tubing was cut into

25 cm segments and filled with the appropriate diamine.
After reaction for 4-5 hours at room temperature, the tubes
were treated identically during the remainder of the immo-
bilization. Thus, differences in activity should be due

to differences in the spacer molecule only. The spacers

are abbreviated S-N, where N is the number of carbon atoms
in the molecule.

Three studies were performed, and the results of the
first two are listed in Table 2.1. Bonding of the enzyme
directly to the activated nylon gave relatively low activ-
ities. Ethylenediamine and 1,8-diaminooctane yielded the
best reactors. Study 3, however, gave different results
as shown in Figure 2.5. The S-0 reactor had a much higher
activity than the S-2 reactor, in contrast to the previous
studies. The S—4 reactor showed the best properties. Though
all the reactors lost enzyme activity during storage, the S-4

tube had the highest stability. Thus, 1,4—diaminobutane

26

Table 2.1. The Results of the Spacer Molecule Studies

 

. Calibration _1 Relative
Spacer Slope (100 AU M ) Activity
Study #1
S-0 0.54 0.09
S-2 6.05 1.00
S-6 5.79 0.96
Study #2
S-0 0.27 0.27
S-2 1.01 1.00
S-6 0.72 0.71

3.2
A
I
'5
:3 2.4
<
(\l
c:
3‘.
0 1.6
Q
2
a)
C
2
g 0.8
L.
.0
B
c)
0.0

Figure 2.5.

27

 

 

 

9 .
_._ 2 VA \
2': = S-4
" , °S-6
‘ 7 5-2
° 5-0
I l l L l I
1 I 1 T l I
5 15 25 35

Storage Time (days)

The storage stability of reactors prepared
with different spacer molecules.

28

was used in all subsequent studies.

The very inconsistent results shown here are not
atypical. Table 2.2 shows some of the spacer molecule
experiments which have been described in the literature.
Various spacers have been used, including aminated starch
and pectin, polyethyleneimine (PEI), diamines, and lysine.
When more than one spacer was used, the spacers are ranked
(1=highest) according to the activities that they produced.
One can see many inconsistencies, possibly due to the use
of different enzymes. Only a rough estimate of the best
spacer molecule can be ascertained from the current data.

An attempt to calculate the amount of amine bound to
nylon was made. The diamine was determined by reaction
with fluorescamine [61] and measurement of the product
absorbance at 390 nm. The fluorescamine, 15 mg per 50 mL
of acetone, was mixed with an equal volume of diamine in
pH 8 phosphate buffer in a stopped-flow instrument. The
reaction rate over the first 0.5 second was directly propor-
tional to the diamine concentration. A calibration curve for
1,6-diaminohexane is shown in Figure 2.6. The other
diamines also gave linear plots with similar (i10%) slopes.

The amount of diamine used in the immobilization and
the amount washed form the tubing after the reaction were
determined. The difference between these two values should
be the amount bonded to the tubing. However, because of

imprecision in the method, no significant differences were

29

Table 2.2. The Activities of Nylon Tubular Reactors
Resulting From the Incorporation of Various
Spacer Molecules

dehydrogenase

Sgigeit Enzyme §:2 §:§ P_l Lysine Ref;

DMS NAD kinase 2 1 55

TOTFB Creatininase 1 56

DMS Invertase 57

TOTFB Lactate 58
dehydrogenase

DMS Urease 59

DMS Glucose 60
oxidase

TOTFB Glycerol 41

30

Reaction Rate (AU 3")

 

 

 

0 . p . l . 1 . l . l
0.0 r I i I I I I I I l I
0.0 0.2 0.4 0.6 0.8 1.0

[1,6—Diawfinohexane] OnM)

Figure 2.6. Fluorescamine determination of
1,6-diaminohexane.

31

found. Thus, the amount of amine bound to the tubing was
not determined. Perhaps a direct method in which the bonded
diamine was stripped from the tube and measured by

fluorescence would have been better.

C. Attachment of Glucose Oxidase

In order to bond the enzyme through a free amine group,
glutaraldehyde was reacted first with the spacer molecule.
A Schiff's base formed, and a reactive carbonyl group was
established on the carrier. Most researchers have employed
a dilute solution of the glutaraldehyde in a basic medium.
Our procedure was to fill the nylon tube with a 5% w/v
solution of glutaraldehyde in pH 8.0 phosphate buffer
(0.1 M) for about 1.5 hours at room temperature. The para-
meters involved in this immobilization step were not
investigated.

Free amine groups on the enzyme were coupled to the
glutaraldehyde through a second Schiff's base reaction.
The coupling was performed at the optimal pH for enzyme
activity, so that the enzyme was bonded in its most active
form. Also, the reaction occurred at 4 °C to prevent loss
of the enzyme activity in solution. The duration of the
reaction was 18 hours to allow the bonding to go to completion.

The concentration of the glucose oxidase solution used
in step 4 of the immobilization procedure was varied to
discover the effect on the resulting reactor activity. An

undergraduate, Brenda Kaufmann, performed these studies.

32

With the conditions of the first three steps held constant,
different solutions of glucose oxidase were added to 30 cm
segments of the same tubing and reacted for 18 hours.
Subsequently, the reactors were incorporated into the CF
manifold and their activities (calibration curve slopes)
were determined. The results are presented in Figure 2.7.

The attachment sites on the nylon were not saturated
until the glucose oxidase concentration exceeded 4 mg/mL.
A nearly linear relationship between activity and added
glucose oxidase was found at lower concentrations. The
non-zero intercept may not be significant in light of experi-
mental errors. Further experiments need to be done to
confirm these results. One conclusion that can be drawn,
however, is that a 5 mg/mL solution is optimal. This ensures
the highest activity with the least amount of enzyme.

The value was also used in work with other enzymes, though
the optimal concentration may change.

The amount of glucose oxidase bonded to the nylon was
not successfully determined. Difference methods were used
in which the amount of protein added to the reactor and the
amount washed from the tube after the reaction were deter-
mined. The difference in the two values should be the
amount bonded. First, the biuret method [62] for proteins
was employed. No significant difference between the two
solutions was found. Next, UV absorbance at 280 nm was

monitored. Aromatic amino acid residues absorb strongly

33

 

 

3 ~—
A
l
2 “r
:3
.<
01 2 .1.
c:
I;
O -.
o.
.9
U)
1 ——
C
9
+’ o
0 o
r) -.
'6
L)
L l l l J
0 I I I I I
O 1 2 3 4 5
[Glucose Oxidase] Ung/WnL)

Figure 2.7.

The effect of initial glucose oxidase concen-
tration on the reactor activity.

34

at that wavelength [63]. Very linear calibration curves
were obtained, but the method frequently gave negative
results. Finally, the ability of glucose oxidase to
catalyze the oxidation of glucose was utilized. A linear
relationship between enzyme concentration and initial
reaction rate was established. Positive results for the
amount of enzyme bonded were obtained, but the precision
of the method was poor. No definitive results were found.
In the future, direct methods should be attempted. The
completed reactor could be cut into pieces and treated
with dilute HCl at high temperatures to strip the enzyme
from the tube and to cleave the protein into amino acids.
Then the amount of aromatic amino acids could be measured
by UV absorbance.

The immobilization procedure selected for this work
is listed in Table 2.3. The reactors were washed with
distilled, deionized water between all steps. Cold
methanol was used to wash unreacted triethyloxonium tetra-
fluoroborate from the tubes after step 1. Many enzymes can
be immobilized using this method. Table 2.4 lists the
enzymes with which we have experimented and gives their
approximate half—lives, when the activity drops to 50% of
the activity on day one. The TOTFB procedure for the
covalent attachment of enzymes to nylon tubing is versatile

and easy to perform.

35

Table 2.3. Immobilization Procedure

Step

Reagent Duration
0.1 M TOTFB in CH2C12 3 min.
0.1 M Diaminobutane 4.5 hr.
in pH 9.4 HCOB’ Buffer
5% w/v Glutaraldehyde 1.5 hr.
in pH 8.0 HPO32- Buffer
Enzyme in pH 6.85 18 hr.

Phosphate Buffer

Temp (°C)

 

22

22

22

36

Table 2.4. Various Enzymes Bonded to Nylon Tubing and
Their Approximate Half-lives

E.C.
Enzyme Number

Alcohol dehydrogenase 1.1.1.1

Peroxidase 1.11.1.7
Carbonic anhydrase 4.2.1.1
Catalase 1.11.1.6

Glutamate dehydrogenase 1.4.1.2

Lactate dehydrogenase 1.1.1.27

Glucose oxidase 1.1.3.4

Amino acid oxidase 1.4.3.2

Malate dehydrogenase 1.1.1.38

Days

- Half-life -

Weeks

 

Months

N3

CHAPTER 3

Kinetics of Open—Tubular Reactors

The study of the kinetics properties of immobilized
enzymes is important because these properties are often
very different from those of enzymes in aqueous solution.
While several compilations of soluble enzyme properties
are available [64-66], relatively little is known about the
kinetics of immobilized enzymes. Most of the current infor—
mation can be found in the references listed in Table 3.1
and in many general reviews of immobilized enzymes.

The reason for the lack of data is that immobilized
enzyme properties depend on reactor design and the type of
immobilization used. Figure 3.1 shows the factors that can
change the properties of an enzyme when attached to a
carrier. Environmental and electrostatic effects vary
according to the chemical nature of the carrier and the
linkage to the enzyme. Similarly, the physical structure
of the carrier determines whether internal and/or external
mass transport will be present and the extent that mass
transport influences the overall reaction rate. External
mass transport is the transfer of material from the bulk
solution to the enzyme environment, while internal mass
transport concerns the movement in the enzymic matrix, such
as a gel layer, membrane, or fiber. The kinetics constants

of immobilized enzymes are defined in terms of the

37

38

Table 3.1. Reviews of Immobilized Enzyme Kinetics

1.

10.

K.J. Laidler and P.S. Bunting, Meth. Enzymol. 64,
227 (1980).

J.-M. Engasser and C. Horvath, Applied Biochemistry
and Bioengineering, Volume 1, Academic Press, NY
(1976), pp 127—220.

 

 

G.P. Royer, Immobilized Enzymes, Antigens, Antibodies,
and Peptides, Marcel Dekker, Inc., NY (1975), pp 49-91.

 

 

 

W.R. Vieth and K. Venkatasubramanian, Chem. Technol.
(1974), 303.

T. Kobayashi and K.J. Laidler, Biotech. Bioeng. 16,
99 (1974).

K.J. Laidler and P.S. Bunting, The Chemical Kinetics
pf Enzyme Action, Clarendon Press, Oxford, England
(1973), pp 383-412.

 

 

C. Horvath and B.A. Solomon, Biotech. Bioeng. 14,
885 (1973).

G.J.H. Melrose, Rev. Pure Appl. Chem. 21, 83 (1971).

R. Goldman, L. Goldstein, and E. Katchalski, Biochemical

 

Aspects pf Reactions pp Solid Supports, Academic Press,
NY (1971). Pp 1—78.

 

 

P.V. Sundaram and K.J. Laidler, Chemistry pf the
Cell Interface, Part A, Academic Press, NY (1971),
Chapter V.

 

 

39

 

Soluble
Enzyme

 

 

 

  
   

   

Changes in.
Enzyme Active
Site

V
/\

 

 

 

 

Immobilized
Enzyme Inherent Apparent

lntnnsic Properfies Properfies
Properhes

 

   

 

 

 

 

 

 

 

 

   
   
     
 

 

 

Internal External
Electro- Moss Mass
stafic Transport Transport

Effects

Figure 3.1. Immobilized enzyme kinetics.

40

conditions under which they were determined. Thus,
there are three different Michaelis constants —-
Km(intrinsic), Km(inherent) which accounts for electro-
static and environmental effects, and Km(apparent) which
also accounts for slow mass transfer.

The aim of much of the research on immobilized
enzyme kinetics is to decouple transport effects from
enzyme kinetics in the overall reaction rate. In other
words, it is important to know the properties inherent to
the immobilized enzyme in order to design the carriers and
immobilization procedures better. The inherent properties
are often masked by slow mass transport, and only the appar-
ent enzyme kinetics are measurable. A new instrument has
been developed which should allow the inherent properties
to be seen indirectly. Testing of this instrument is the
focus of this chapter.

Because fine reviews on immobilized enzyme kinetics
are available, Section A serves only to review the major
theoretical findings of researchers in the field. The
kinetic processes which occur inside open-tubular immobil-
ized enzyme reactors (OTIMER) integrated in three different
flow systems are explained in more detail. Section B
describes the experimental requirements in monitoring the
kinetics of immobilized enzymes in open tubes and presents
our novel method. The design of the stopped—flow instru-

ment is detailed. Experimental results show that the

41

instrument is valuable in determining Michaelis constant
(Km) values (Section C), in performing clinical deter-
minations (Section D), and for investigating more fundamen-
tal kinetics properties (Section E). Section E also

describes an improved design for the instrument.

A. Kinetics Properties of Open-Tubular Reactors
1. Electrostatic and environmental effects

The enzyme reaction rate for a single substrate
reaction that obeys Michaelis-Menten kinetics is given in

equation 3.1.
v = (ke[E]O[S]w)/(Km(intrinsic) + [S]w) (3.1)

The rate, v, depends on an enzyme rate constant, ke’ the
enzyme concentration, [E]O, the substrate concentration
at the reactor wall, [S]w, and the intrinsic Michaelis
constant, Km. Differences in the immobilized enzyme rate
from that of the enzyme in aqueous solution are usually
shown by modifying the Km term in equation 3.1. This is
sensible since the Michaelis constant indicates the degree
of enzyme affinity for a substrate, and electrostatic and
other environmental effects can change this affinity.

If both the carrier and substrate are charged, then
the affinity of the substrate toward the enzyme will be

altered. Equation 3.2 shows the change which occurs [67].

Km(inherent) = Km(intrinsic)exp(—zew/kT) (3.2)

42

In equation 3.2, 2 is the electrovalency of the substrate,
e is the charge on the electron, W is the electrostatic
potential between the carrier and substrate (if the carrier
and substrate are of like charge, then w is negative), k is
the Boltzmann constant, and T is temperature in °K. If
like charges are present, the Michaelis constant is larger
than the intrinsic value, while the opposite is true if
unlike charges are present. Wharton, 23.31. considered the
effect of ionic strength on the Km value and derived the

following equation [68].
Km(inherent) = Km(intrinsic)(2cI)/(Z+21) (3.3)

The term, I, is the ionic strength, Z is the effective
concentration of fixed charges on the support, and c is
the ratio of the mean ionic activity coefficients for the
carrier and the bulk solution.

Electrostatics also change the pH optima and substrate
migration rates in carriers which have a surface charge.
The pH optimum changes according to the following mathe—

matical description [67].
pH(inherent) = pH(intrinsic) + 0.43ep/kT (3.4)

The substrates are attracted to the reactor wall if the
carrier is of opposite charge and repelled if of like charge.
This effect on the substrate transport can be described

again by modifying the Michaelis constant term [69].

43

Km(inherent)=Km(intrinsic)/(1—azeC(grad w)/kT) (3.5)

Here a is the diffusion layer thickness, C is a constant,
and the other variables are defined above. In summary,
many factors can change the intrinsic kinetics properties
of an immobilized enzyme. The resulting properties are
termed the inherent properties.
2. Mass transport effects

Inside an OTIMER, the chemical reaction occurs at the
tube wall, catalyzed by the attached enzyme.i Simultaneously,
the substrate molecules migrate under the influence of a
radial concentration gradient or convection from the bulk
solution to the reactor wall. The enzyme reaction rate is
given in equation 3.1, and the mass transport rate is

described below [70].
v = h([S]b-[S]w) (3.6)

The reaction rate, v, is influenced by the mass transport
rate constant, h, the bulk substrate concentration, [S]b,
and the substrate concentration at the reactor wall.
These two rate processes contribute to the overall reactor
performance. In most cases, both processes influence the
reaction rate, although at the extremes of enzyme and/or
substrate concentration one process may dominate.

At steady-state, when d[S]w/dt = 0, the rates of the
enzyme reaction (eqn. 3.1) and mass transport (eqn. 3.6)

are equal. If the two equations are set equal to each

44

other, the equations are solved for [S]w, and then [S]w
is substituted into the reaction rate equation (3.1),

the following equation results.

V [S] [K ’(1+U)+((1+u)2+2[8] (l—u)/K +([s] /K )2) 1/2
bm b m bm

The above equation is equation 3.7. The term, Km is the
inherent Michaelis constant for the enzyme, and u is equal
to ke[E]O/Km(inherent)(h), the ratio of the enzyme rate
constant to the mass transport rate constant. The effect
of slow mass transport on the apparent Michaelis constant
can be seen in Figure 3.2. The plots were produced by
substituting different values for [S]b/Km(inherent) and U
into equation 3.7. The value of Km is increased over the

inherent value as described in the equation shown below [71].
Km(apparent) = Km(inherent)(l + 0.5u) (3.8)

As mass transfer limitations become larger (u increases),
the analytical range for substrate determinations widens.
Thus, the ideal conditions for analytical methods are high
enzyme activity and substrate transport control of the
reaction rate inside the OTIMER.

With two substrate reactions, mass transport has a
very interesting effect on the two Km(inherent) values.
If the inherent affinity of the enzyme toward the two
substrates is very different, then substrate transfer tends

to raise the Km of the higher affinity substrate while

1
E
>'
\\
>
Figure

3.

45

 

 

J“.
‘— 0.0.1
1
__ 2
5
__ 10
J.
I I I I I L, I
I I I I I I I I
O 1 2 3 4 5 6 7
[Substrate]/Km

2. Influence of slow mass transport on enzyme
kinetics.

46

lowering the constant of the other substrate [72,73].

This see-saw effect has been confirmed in several immobil-
ized enzyme studies [74—76]. Under other conditions,

the apparent Km values can both be higher or lower (see
Engasser's paper for details [72]).

The type of mass transport in open-tubular reactors
depends on the system being considered. These reactors
are used almost entirely in three flow system types:
stopped-flow, flow-injection, and continuous flow. The
mass transport processes and kinetics of each of these are
discussed in turn. Only external mass transport needs to
be examined, since the enzyme is bound essentially to a
non-porous wall in a very thin layer.

The reaction in the stopped-flow system occurs under
static conditions. In the simplest mass transfer model,
only molecular diffusion is considered. Once the enzyme
reaction begins, concentration gradients in both substrate
and product are established as shown in Figure 3.3. The
mass transport rate constant under static conditions may be

given by the following equation [70].
h = NuD/(2ra) (3.9)

Here Nu is the Nusselt number, D is the molecular diffusion
coefficient for the substrate, and r is the reactor radius.
Under conditions of high enzyme binding and activity,

molecular diffusion may totally control the overall reaction

47

 

 

Relative Concentration

 

 

 

 

 

 

Distance along Reactor Diameter

Figure 3.3. Concentration gradients in an open—tubular
reactor.

48

rate. A mathematical description has been derived and
is given below. Details of the model are described in
Appendix A.
3
[p1b = [s1b<1 — 4 Eexp(-Dj:t/r2)/j:) (3.10)
s—l
where [P]b is the product concentration in the bulk
solution at time t, r is the reactor radius, jS is a
Bessel function, and t is the time inseconds.

When a flowing liquid passes through an OTIMER,
convection becomes the primary means of transport. Enzyme
reactors used in flow-injection analysis [4-6] are examples
of this system. Laidler and coworkers have been the major
contributors to the theory of mass transport in such

systems [77-80]. Equation 3.11 is a mathematical descrip—

tion of the transport rate constant [77].

3

h = 1.29 0-1(D20/rL)1/ (3.11)

where v is the linear (cm/s) flow rate, and a is the
diffusion layer thickness. The inherent Michaelis constant

is transformed into the following [77].
_ . 2 1/3
Km(apparent)—Km(1nherent)+0.39ake[E]O(rL/D v) (3.12)

Since mass transport is faster in flowing systems than
under static conditions, the enzyme reaction rate should
have a greater influence on the overall kinetics in flowing

systems.

49

Similar interplay between the enzyme and mass trans-
port is expected in air-segmented continuous flow systems
[21-23], but the mass transport rate has not been described
mathematically. A thin film of liquid which moves only
slowly along the reactor wall and secondary, bolus, flow
which is established in each liquid segment complicate the
fluid dynamics. Therefore, primary flow, secondary flow,
and mass transport in the thin film all contribute to the
mass transfer of the system. Snyder has described the
effect of incomplete mixing in the liquid segments on the
sample dispersion [82,83]. One should be able to relate
the extent of dispersion to the rate of radial mass trans-
port, since it is known that mixing along the direction
of flow is fast while mixing across adjacent streamlines
is the limiting factor. Figure 3.4 shows the pattern of
bolus flow.

One can assume that the rate of mass transport is
inversely proportional to the sample dispersion due to
incomplete mixing. From Snyder's theory, the following
equation may be a good estimate of mixing in OTIMERs in

segmented—flow systems.

2/3 1/2

h = (1-25Y VD )/(v7/6n2/3L1/2d3

) (3.13)
Here Y is the surface tension of the liquid, V is the

volume of a single liquid segment, n is the solution vis—

cosity, v is the linear flow rate, L is the reactor

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/
‘\
1’
‘\
1’
\\
Air Liquid Air
Segment Segment Segment
1’
‘\
1’
‘\
,/
l \
Liquid
I

 

 

Figure 3.4. Bolus flow pattern.

51

length, and d is the reactor diameter. Further theoretical

and experimental work on this system is needed.

B. Experimental Considerations

The kinetics of open—tubular immobilized enzyme
reactors are difficult to monitor. In all systems, the
reagents are pushed into the reactor, allowed to react for
a time, and then the substrates or products are monitored
as they exit the tube. Thus, the reaction rate obtained
is the result of a fixed-time measurement, with the initial
time (t1) being zero and the final time (t2) being the
moment that the substrate solution exits the reactor.
While t2 can be varied, tl is usually restricted to a
value of zero. Enzyme lag phases, therefore, are included
in the rate measurements, which introduces error.

Continuous flow methods are used widely in conjunc—
tion with immobilized enzymes. Yet few kinetics experi-
ments have been performed because data at early reaction
times are impossible to obtain and entire progress curves
would require many experiments under varying conditions.
For example, a change in pump speed to change the reaction
time might also alter the mass transport rate. Bi-direc—
tional flow, continuous flow methods [84] and other designs
are possible ways of reducing the number of required
experiments, but the instrumentation and procedures are
complicated. A stopped—flow method using OTIMERs has been

described [85], but again many experiments are required in

52

order to draw a complete progress curve of the enzyme
reaction.

Ideally, one would like to monitor the reaction
directly and continuously. We have developed an instrument
that is ideal in its concept, and only a little less so
in its function. A tubular reactor, containing an immobi-
lized enzyme, is fitted inside the observation cell of a
stopped-flow instrument. With the reagents mixed and
pushed into the observation cell, the reaction begins. The
light path of the photometer passes through the reactor and
reacting solution; thus, the formation of product or de-
pletion of substrate can be monitored directly and contin-
uously by absorption spectroscopy.

The advantages of the instrument are numerous. Early
reaction times are obtainable since the injection of the
mixed reagents is performed quickly by the stopped-flow.
The time window can be varied so as to exclude lag periods
and other error-prone regions. Also, initial reaction
rates can be calculated more accurately by the linear
regression slope method rather than by the fixed-time
method. Only diffusion and/or electrostatic effects con-
tribute to mass transport in the reactor so that funda-
mental studies of the kinetics can be done more easily.

The only disadvantage is that one of the substrates or
products must have a unique absorbance maximum in the range

of the monochromator. If not, a follow-up reaction can be

53

used, but only if its reaction rate is much faster than
that of the analytical reaction rate.

The requirements for the instrument described above
are somewhat different from those of a conventional stopped-
flow instrument. First, fast mixing is not essential since
the reaction does not begin until the reagents enter the
observation cell. Even manual mixing is possible. However,
the injection time, the time taken to push the solution
through the cell, must be small so that the reaction extent
prior to the first data point is negligible. These re-
quirements can be provided by simply employing a T-mixer
and a strong syringe push. Second, downstream stopping
is not required since typical enzyme reactions are moni—
tored on the order of seconds rather than milliseconds.
Third, mixing between reagents and the reaction volume
inside the observation cell should be minimized. Unlike
standard stopped—flow experiments, products are only formed
in the observation cell/reactor; diffusion of product into
the reagent solution at the entrance and exit of the cell
will cause errors in the rate measurement. Any siphoning
or solution leakage will lead to drastic errors. The
design of an instrument which minimizes this mixing is
a difficult problem. A prototype instrument which elimi-
nates the reaction volume/reagent solution interface was
built, and preliminary test results are given in Section E.

A modified GCA McPherson stopped-flow module was used

54

in most of the kinetics studies. A 1.75 cm immobilized
enzyme reactor, 1 mm i.d., was fit tightly inside the
observation cell of the module. Figure 3.5 shows the
location of the reactor in the cell housing. A T-mixer
was employed, along with upstream stopping. The remainder
of the instrument was constructed from commercial com-
ponents as shown in Figure 3.6. Data acquisition was
performed by a microcomputer [86], linked to a PDP-8/e
minicomputer (Digital Equipment Corporation). Further
details of the instrument and its use are given in Section C.

One other important consideration in the use of this
method is that the absorbance measurement is atypical.
Because a product (and substrate) concentration gradient
exists radially across the tube, the relationship between
absorbance and concentration does not depend only on the
molar absorptivity, c, and the reactor length, but also
upon reactor radius and the ”steepness” of the concentra-
tion gradient. A mathematical model of the time-indepen-
dent absorbance versus reactor design relationship is
described in Appendix B. The time dependent case is com-
plicated by the rate of product formation and mass transport
and has not been considered. The final result of the

mathematical treatment is given in equation 3.14.
Absorbance = eL([P]b+a([P]w—[P]b)/2r) (3.14)

Here [P]b is the concentration of the product in the bulk

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TEL—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.5.

 

 

l l

Solufion Flow

Cross-sectional View of the GCA McPherson
stopped—flow mixer/observation cell unit.

The outer aluminum housing (a), quartz win-
dows (b), and the main KEL-F body (0) are
press fit with three bolts. Mixing occurs at
(e), where the two reagent flow streams meet
at 90° to each other-- one stream is in the
plane of the figure and the other is perpen-
dicular to it. The immobilized enzyme reactor
is placed inside the observation cell (d).
The dashed arrow represents the light path
inside the cell.

56

 

GCA McPherson

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

GCA McPherson :UOb-se-rotn'onfi O , I GCA McPherson
Model EU—701-5o -C°,,/;"'c,°ra FP‘M Model EU-701-30
Light Source Ilter Photomultiplier
Stopped-flow
Module
Keithley
Microcom uter Model 427
[3 Current to Voltage
Amplifier
PDP B/e Strip Chart
Minicomputer Recorder

 

 

 

 

 

 

Figure 3.6. Stopped-flow calorimeter.

57

solution, and [P]w is the concentration of product at the
reactor wall. The absorbance can vary with reactor diameter.
However, during a single experiment, the reactor diameter
is a constant and the absorbance depends on the diffusion
layer thickness, 0. Changes in this factor during an
experiment are expected to occur until steady-state between
the enzyme reaction rate and mass transport is established.
If steady—state conditions are established quickly, then
most of the data will be free of mass transport influence
on the absorbance measurement. More definitive experiments
are needed to establish when steady-state is attained and

to verify equation 3.12.

C. The Determination of Michaelis Constants
1. Introduction

We report here a study of the kinetics of L(+)-lactate
dehydrogenase, LDH (E.C. 1.1.1.27). This enzyme converts
L-lactic acid and B-nicotinamide adenine dinucleotide
(B-NAD) to pyruvic acid and B-NADH, the reduced form of
B-NAD. The reaction goes by a compulsory order bi-bi

mechanism [87]:

+
E1 + B-NAD + 2

E
+
E2 + Lactate + E3

E3 + E1 + Pyruvate + B-NADH + H+

Here El represents the free enzyme, and E2 and E3 represent

58

the bimolecular and termolecular enzyme—substrate complexes
respectively. The initial reaction rate for such a mecha—

nism is given in equation 3.15.
v=k [E] /(1+KN/[NAD]+KL/[Lac]+K KL/([NAD][Lac])) <3 15)
e o m m i m °

The term, Ki’ represents the degree of interaction between
the two substrates on the enzyme; the value is zero for a
ping-pong type mechanism.

Lactate dehydrogenase is a tetrameric molecule, com—
posed of heart (H) or skeletal muscle (M) type monomers [87].

MH M H M H,

4’ 3’ 2 2’ 3

and their relative proportion varies among different

Thus, five different isozymes exist—— H

and M4,

species and tissues. The preparation used in this study
was obtained from beef muscle, and so it contained mostly

the M4 and M3H tetramers. The molecular weight of the

enzyme is about 140,000. Each subunit contains a B-NAD
binding site, but these sites are independent of one another.
High concentrations of pyruvate and lactate inhibit the

enzyme. The KI value, the inhibition constant, for lactate

varies from 26 mM for the H isozyme to 130 mM for the

4
M4 tetramer.

In this section, the three kinetics constants and Vm’
which equals ke[E]O, are reported for aqueous and nylon-

immobilized LDH, and the values are compared. Partial

diffusion control of the reaction rates is evident.

59

2. Reagent preparation

Tris Buffer. The following quantities of materials
were dissolved in distilled, deionized water (DDW): 10.0 g
trishydroxyaminomethane, 13.0 g hydrazine sulfate, and 2.0 g
ethylenediaminetetracetic acid. The solution was diluted
to one liter with DDW and then the pH of the solution was
adjusted to 9.6 with NaOH. The buffer was stable for one
week.

B-NAD Solutions. The desired amount of B-NAD (Grade
III, molecular weight=706, Sigma Chemical Company) was
dissolved in buffer and diluted to 50.0 mL with buffer to
give a stock solution. The working solutions were prepared
by appropriate dilution of this stock with buffer. The
stock solution and working standards were prepared daily.

Lactate Solutions. Purified L—lactic acid (0.960 g,
Grade L-X, Sigma) and 0.125 mL of concentrated sulfuric
acid were added to 500 mL of DDW. The resulting solution
was diluted to 1.00 L with DDW. This stock was diluted
appropriately with DDW to give the desired standard solu-
tions. The stock solution was stable for two weeks, and
the standards were prepared daily.

Soluble Enzyme Reagent. Eighty to one hundred micro-
liters of a suspension of beef muscle LDH (Type X, 8800
Units/mL suspension, Sigma) were dissolved in Tris buffer,
and then diluted to 0.200 L with buffer. This reagent was

then substituted for the Tris buffer in the B—NAD solution

60

preparations described above. The B-NAD/enzyme solutions
were used within two hours.
3. Immobilization of lactate dehydrogenase

Lactate dehydrogenase was covalently bonded to
straight, 20 cm lengths of nylon tubing (Portex Ltd.,
1 mm i.d.). The immobilization procedure is given in
Chapter 2, Table 2.3. The lactate dehydrogenase suspension
obtained from Sigma was used directly in the immobilization.
4. The instrumentation

The stopped-flow system was assembled from commercially
available components. The stopped—flow module (GCA Mc-
Pherson) was modified for upstream stopping by removal of
the stop syringe. The two push syringes had volumes of
370 uL. A filter photometer was employed for absorbance
measurements. The radiation source was a deuterium lamp
housed in a commercial light source module (GCA McPherson).
A 340 nm interference filter (bandwidth = 12 nm) was mounted
between the stopped-flow unit and a photomultiplier module
(GCA McPherson). The photocurrents from the photomul-
tiplier were converted to voltages (Kiethley model 427
current amplifier), digitized, and the digital voltages
stored in RAM with a microcomputer data acquisition system
[86]. Data were acquired at 1.0 Hz. At the end of each
trial, the data were shipped serially to a PDP-8/e mini-
computer (Digital Equipment Corporation) for processing and

analysis. The initial reaction rates were computed as

61

the linear regression slopes over the time period from
10 to 20 seconds after the final reagent push.

The instrument was operated in the following manner.
The two syringes were filled with DDW, the water was pushed
through the observation cell, and the flow was stopped.
This procedure was repeated six times to wash out the reac-
tor completely. Next, the syringes were filled with the
reagents, the reagents were pushed through the mixer and
cell, and the flow was stopped. This second procedure was
repeated five times in order to ensure quantitative 1+1
mixing of the B-NAD and lactate solutions. After the final
reagent push, the microcomputer data—taking routine was
started. These steps were repeated for each experiment.
The blank consisted of the B—NAD working solution mixed
with distilled deionized water.

For the soluble LDH studies, the instrument was used
as described above. After the soluble enzyme studies were
completed, the observation cell of the stopped-flow module
was disassembled, and a 1.75 cm section of the immobilized
LDH tube was fitted tightly inside the cell. The cell was
reassembled, and the study was repeated. Between trials,
the IME tube was filled with pH 6.85 phosphate buffer,
.0.05 M. A room temperature water bath kept the reagents
at a constant temperature, between 18 and 20 °C, during

all experiments.

 

(.3‘

H:

(‘11

62

5. Results and discussion

The precision of the rate measurement was determined
by repetitively montoring the immobilized LDH catalyzed
reaction of 0.50 mM lactate and 1.35 mM B-NAD. The rela-
tive standard deviation was found to be 4%. The enzyme
activity declined rapidly during the first two days of use
and, thereafter, only slowly. Even after seven days at
room temperature and about 175 experiments, 60% of the
original enzyme activity remained.

Typical progress curves obtained with our instrument
are shown in Figure 3.7. The immobilized LDH displays
a sigmoid-shaped curve, that is similar to the one produced
by the aqueous LDH. No odd reaction kinetics are apparent.
Also, a short lag phase is noticeable. These facts are
quickly and easily shown, in contrast to other indirect
kinetics methods.

The Km values were determined by the method of
Florini and Vestling [88]. The mean values of the initial
reaction rates for each pair of substrate concentrations
were fit to equation 3.15 by a computer regression program.
The experimental and regression results are shown as
Lineweaver-Burk plots in Figures 3.8 and 3.9. Because the
lines in the double reciprocal plots intersect at a common
point below the x-axis, the formation of a ternary complex
is confirmed.

A small concave curvature exists in the experimental

63

 

Absorbance

 

 

 

Time (seconds)

Figure 3.7. Typical progress curves obtained with the
instrument. The dotted lines represent the
initial reaction rates calculated from these
curves. The reactions represented were of
2.00 mM lactate with 0.425 mM (A), 0.850 mM
(B), and 2.12 mM (C) B-NAD.

64

6.0-m-

1/R0te (x102 sec)

 

 

 

'-2.0 T ! ' g ; t . I ' i

-45.0 —30.0 -15.0 030 15.0 30.0

1/{Lc0‘tate} (x302 M-i)

Figure 3.8. Lineweaver—Burk plot of 1/Rate versus
1/[Lactate] at four different B-NAD
concentrations: 0.26 mM (A), 0.43 mM (B),
0.85 mM (C), and 2.12 mM (D).

65

 

 

6.0'—r
A
’5‘ 4.0——
o 0 B
m .
j_ C
a:
0 ° . °
3'. 20——
v o e
o __ . '
o 0
I:
'\~ 0 O..-
l I . I . I I I
-2‘0 ' I ‘ I l 1
-10.0 '-5.0 0.0 5.0 ‘10.0
1/[NA03 (x103 20“)

Figure 3.9. Lineweaver-Burk plot of 1/Rate versus
1/[B—NAD] at three different lactate
concentrations: 0.80 mM (A), 2.00 mM (B),
and 4.00 mM (C).

66

values. Such curvature is not apparent in the aqueous

enzyme data. Sundaram and Igloi also reported such curved
plots in a multi-enzyme system bound inside nylon tubing [56].
Diffusional resistance or the presence of several iso-

zymes may be the cause of this effect, but no realistic
mathematical model has been found.

The y-intercepts of the Lineweaver-Burk plots
represent the reaction rate at infinite lactate or B-NAD
concentrations. A second Lineweaver-Burk plot of these
rates yields the desired Km and Vm values, as shown in
Figure 3.10. The results of five independent trials were
weighted and averaged to give final values for the kinetics
constants. The values found for soluble and immobilized
LDH by this method are given in Table 3.2.

The Km of B-NAD increases and the Km of lactate
decreases upon immobilization. Intuitively, Km values of
bound enzymes should not be lower than those of soluble
enzymes, unless a conformational change occurs upon immo-
bilization. However, Engasser, £3.31. [72] recently ex—
plained that if the Km values and diffusion rate constants
(hs) of the enzyme substrates differ greatly, then the
tendency is to increase the Km of the higher affinity
substrate and decrease the Km of the lower affinity sub—
strate. This is due to diffusional limitations on the
reaction rate.

Engasser, §£.al. restricted the model to ping-pong

'1.
o

 

67

 

 

I
I
l
I.
L /
a. I!
(D 1’
I
N , 1’
o 2.0—— I, 1’
I /
I ’l
X . l’ ,
V l p
...-I T. a, ,’
Q l- [I
Q I I
o I I
a; " ”
I
*0 100 —_ I /d
C ’ I
.- a ,
' ’sz
>‘ D,
l/
4'
”l
I” i
/ I’ i
’ I
..’ ’ I I l l l l
0.0 I I i j
-2 0 2 4 6 3 10

Figure 3.10.

Plot of y-intercepts from Figures 3.8 and
3.9 versus 1/[Substrate] for different B-NAD
and lactate concentrations. The y-intercept
equals 1/Vm, and the x-intercepts yield the
two Km values.

68

Table 3.2. Kinetics Values for Aqueous and Immobilized
Lactate Dehydrogenase

 

Agueous Immobilized
vm (10'6 Ms‘l) 2.0 i 0.1 2.1 i 0.1
KS (10‘4 M) 2 5 1 0.3 4.3 i 0 5
K; (10'4 M) 19 i 4 9 i 1
K (10'4 M) 5 8 i 1.6 1.9 1 0.5

69

type reactions, but we believe that the same effect should
occur with compulsory order mechanisms. In the case of
LDH, the important dimensionless factor, 5 =(hNK:)/(hLKg),
equals approximately 0.07. Since this is very much
different from 1, it may be expected that the see-saw
effect on the Km values will occur. Using Figures 1 a,b,
e, and f of reference 72 and interpolating to approximate
the conditions of this experiment, weestimated that the
enzyme reaction rate was about five times faster than the
diffusion rate.

A second explanation of the KIn changes is based an
electrostatic effects. The immobilization procedure im-
parts a positive charge to the nylon surface. Since lactate
has a negative charge and B-NAD has a positive charge, the
apparent Km values could decrease and increase respectively,
because the surface concentrations of the two substrates
differ from those in the bulk solution. The kinetics values
found here are relevant to other methods in which a reagent
solution is allowed to react inside an OTIMER under static
conditions. Sundaram's impette method [50] and stopped-
flow, flow—injection methods employing OTIMERs should
benefit directly. The method is very useful in determining
the apparent Michaelis constant values for enzymes immobil-
ized inside open tubes. A variety of enzymes can be tested

in this manner.

70

D. Analytical Determinations
1. Introduction

Direct reaction rate methods for two clinically
important substrates, glucose and lactate, are reported
here. The glucose determination was performed under fixed-
time conditions, while the lactate results were obtained
from rates calculated by linear regression analysis of the
absorbance—time data. The instrument allows calculation
of the initial reaction rate by the fixed—time, variable-
time, or the slope approach [89].

The Trinder method [90,91] for the determination of
glucose with glucose oxidase was adopted here. The two

reactions involved are given below.

B-D—glucose + 02 + H20 + 0—gluconolactone + H202

2 H202 + DCPS + AAP + Dye + 3 H20 + st04

Glucose oxidase (GO) catalyzes the first reaction, while
peroxidase (P0) does the same for the second reaction.
DCPS is 2,4-dichlorophenolsulfonate and AAP is 4-amino—
antipyrine. The absorbance of the dye product at 505 nm
was monitored.

Lactate was determined by measuring the absorbance of

B-NADH produced in the following reaction.
Lactate + B-NAD + Pyruvate + B—NADH

Lactate dehydrogenase (LDH) catalyzes this reaction.

71

The spectroscopic measurement was done at 340 nm. A basic
solution and hydrazine were used to drive the reaction
toward completion. The results of several control serum
assays for glucose and lactate are also reported.

2. Reagent preparation

Glucose Standards. The glucose stock solution was
prepared by dissolving 2.000 g of anhydrous a-D-glucose

and 0.50 g of benzoic acid in about 0.75 L of distilled
deionized water (DDW). The solution was then diluted to
exactly 1.00 L with DDW. The standard solutions were pre-
pared by appropriate dilution of the stock with DDW. The
stock was stable for a month at 4 °C.

Trinder/Peroxidase Reagent. The method of Barham and
Trinder [91] was used to prepare a solution of 2,4-dichloro-
phenolsulfonate. Two mL of this solution (about 0.2 mmoles),
110 mg (1,670 Units) of peroxidase (Type II, Sigma), and
ILO mg (about 0.05 mmoles) of 4-aminoantipyrine were dis-
ssolved in 75 mL of pH 6.4 phosphate buffer, 0.1 M. The
ssolution was then diluted to exactly 0.100 L with the pH

€5.4 buffer. This reagent was prepared daily.

Lactate Standards. Purified L-lactic acid (0.960 g,
Cirade L—X, Sigma) and 0.125 mL of concentrated sulfuric
aJZid were added to 500 mL of DDW. The resulting solution
‘VEts diluted to exactly 1.00 L with DDW. This stock was
(ii.luted appropriately with DDW to give the desired standard

SOlutions. The stock was stable for two weeks; standards

72

were prepared daily.

B-NAD Reagent. The following quantities of materials
were dissolved in DDW: 10.0 g of trishydroxyaminomethane,
13.0 g hydrazine sulfate, and 2.0 g ethylenediaminetetra-
acetic acid. The pH of the solution was adjusted to 9.6
with NaOH after dilution to exactly 1.0 L with DDW. The
buffer was stable for one week. In 40 mL of this buffer,
0.300 g of B-NAD (Grade III, Sigma) were dissolved. The
resulting solution was diluted to exactly 50.0 mL to give
a 8.5 mM B-NAD solution. This reagent was prepared daily.

Serum Samples. The control sera (Monitrol and Patho-
trol, Dade Division, American Hospital Supply) were re—
constituted according to the manufacturer's instructions.
For the glucose determination, 1.00 mL of the serum, 1.50
mL of a 20 g/L solution of Ba(OH)2, and 1.40 mL of a

20 g/L solution of ZnSO in that order, were added to a

4,
centrifuge tube. The solution was thoroughly mixed and
then centrifuged for three minutes. The supernatant fluid
was used directly for the determination. No protein pre-
cipitation was involved in sample preparation for the
lactate determination. One millilter of the control serum
was diluted to exactly 0.100 L with DDW, and the resulting
solution was used in the analysis.

3. Enzyme immobilization

The enzymes, glucose oxidase (Type II, Sigma) and

beef muscle lactate dehydrogenase (Type X, Sigma), were

73

immobilized on Type '6' nylon tubing (0.1 cm i.d., Portex
Ltd.) by a procedure described in Chapter 2, Table 2.3.
A 10 mg/mL solution of glucose oxidase was prepared for
use in step 4 of the procedure, while the lactate dehy-
drogenase suspension was used directly as sold by Sigma.
The completed reactor was filled with buffer (pH 6.4,
phosphate) and stored at 4 °C when not in use. Short,
1.75 cm, segments of the tubing were cut for use in the
stopped—flow observation cell when required.
4. The instrumentation

The instrument has been described in detail in Section C.
For the fixed-time measurements, a monostable timer circuit
was built to cause a sample-and—hold device to hold light
intensity values obtained thirty seconds (11) and sixty
seconds (12) after the reaction began. The change in ab-
sorbance was taken to be log(11/I2). This assumes no drift
in the 100% T setting over the 30 second time interval.
The light source was very stable, so this was probably a
good (< 2% error) assumption. The time range was chosen
as an Optimum between sample throughput and errors due to
mixing effects and enzyme lag periods. The lactate data
were taken and processed by computer as described in Section C.
5. The results of the glucose determination

Glucose standards and the Trinder/peroxidase reagent
were mixed rapidly and pushed into the observation cell/

immobilized enzyme reactor by the stopped-flow instrument.

74

The decrease in light intensity reaching the detector

was monitored and the change in absorbance was calculated
as described above. A blank was tested between each sam-
ple, because the Trinder reagents and product adsorbed
strongly to the nylon, producing some carryover. Thus,
the sample throughput was limited to a maximum of thirty
samples per hour.

The results of five separate experiments over a
four day period using the same glucose oxidase reactor
are shown in Figure 3.11. A very linear calibration
plot with a slope of 8.3 i 0.2 x10"2 and a y—intercept
of -3.4 i 0.1 x10_3 was obtained. The error in the
absorbance measurements was mostly due to a decline in
the enzyme activity with time. The reactor, stored at
room temperatrue for five days, dropped in activity by
only 8%. The linear range of the method, 1-10 mM,
covered the normal levels found in human blood serum
(3.8—5.9 mM [92]).

Three control serum samples were assayed with the
instrument. During the protein precipitation treatment,
the samples were diluted 1:3.9; the resulting concentrations
were still within the linear range. Table 3.3 compares
the concentrations found in our laboratory with those given
on the data sheets accompanying the samples. The accuracy
of the method was good, with the greatest deviations

occurring with samples of low glucose content. This

75

 

 

0.6 —-
0.5-1—

‘1’ _

g) 0.4 —

C

a

.o

L.

O 03*

U)

.a

<3:

<1 0.2 —r
0.1 ~—
00. I l I l I

[Glucose] UnM)

Figure 3.11. Calibration curve for glucose.

76

Table 3.3. Results of Serum Assays for Glucose

 

 

 

Serum No. Detns. Expected(mM) Found(mM)
Monitrol I 5 5.67 4 i 1
Pathotrol 6 12.8 13.3 i 0.3

Monitrol II 9 13.4 13.2 i 0.5

77

deviation was probably due to the presence of ascorbic
acid, gentisic acid, and other reducing agents, which can
interfere with peroxidase-catalyzed reactions [93,94].
More details of this problem are given in Chapter 4,
Section B.
6. The results of the lactate determination

Initial reaction rates for the reaction of lactate
and B-NAD were determined between 10 and 20 seconds after
the final reagent push, with a computer linear regression
program. Sets of standards were tested with a blank run
at the end of each set. Sample carryover was small with
this system, apparently because little adsorption of
reagents or products on the nylon occurred. Thus, the
processing of 90 samples per hour was possible.

Reaction rates obtained from five experiments over
a four day period are plotted versus lactate concentrations
in Figure 3.12. The same lactate dehydrogenase reactor was
used in all of the experiments. The linearity is excellent,
but a small positive intercept is apparent. This inter-
cept may have been caused by a slow non—enzymatic reaction
between the reagents. The blank, which consisted of DDW
and the B-NAD reagent, would not correct for this. The
slope of the calibration curve was 4.3 i 0.3 x10-3, and
the y-intercept was 1.0 i 0.1 xio'lo.

The errors in the reaction rate data were mostly due

to fluctuations in the reactor activity. Sixty percent

/'\
,—
I
(n
.2
|\
I
O
X
v 1
(1)
+1
0
01 1
C
O
.4:
0
O
0)
0:

Figure

78

 

J1 1 l 1 l 1 l
O J T I l I I r
0 2O 4O 60

[Loctote] (MND

3.12. Calibration curve for lactate.

 

79

of the original enzyme activity remained after seven days
of heavy use at room temperature. The linear range of
the method, 5—50 uM, is limited at the high end by the
following requirement: [Lactate] < 0.05 Kg. The K; of
immobilized LDH was determined to be 9 mM in Section C of
this chapter. Thus, the calibration curve should bend off
near 45 pM and it does.

Serum assays were performed on 1:100 dilutions of
the reconstituted controls with DDW in order to adjust
the sample concentrations to the linear range of the
method. This was quite advantageous, since the sample
preparation was simplified and interfering substances should
have a much smaller effect. Table 3.4 shows that the lac-
tate concentrations in all three samples were accurately
determined. However, the average values were slightly
higher than those expected. The reason for this was that
a small amount of aqueous lactate dehydrogenase was present
in the sera, and the soluble enzyme reaction also occurred
inside the reactor. Nevertheless, the method is very
sensitive and the accuracy is within the limits of error.

Since the sensitivity and accuracy are good, but the
sample throughput is low, this stopped-flow instrument
should be suitable for method development and substrate
determinations in small clinical laboratories. Many other
enzymes can be immobilized inside nylon tubing and used in

the instrument to determine a wide variety of compounds.

80

Table 3.4. Results of Serum Assays for Lactate

 

 

 

Serum No. Detns. Expected(mM) Found(mM)
Monitrol II 6 1.8 2.1 i 0.2
Monitrol I 5 2.2 2.2 i 0.4

H-

Monitrol I 6 3.1 3.2 0.6

81

The calculation of reaction rates is direct, and lag periods
can be avoided. With greater automation and the use of

more inert supports, such as glass or Teflon, higher through-
put is likely, and the method should become more advanta-
geous. The instrument is versatile in that it can function
as a device which determines kinetics values for assay

development and also performs the actual determination.

E. A New Cell / Reactor Design

In order to eliminate any error due to mixing between
reacting and unreacted solutions in the stopped-flow
instrument, the cell/reactor was redesigned. Figure 3.13
is a diagram of the new instrument, built around two Altex
slider valves. The cell was moved between positions 1 and
2 with a pneumatic valve and applied gas pressure.
Typically, the cell was filled while in position 1 and
then was moved quickly into position 2 for spectrophoto-
metric monitoring. In this way, the interface between the
reaction volume and the mixed reagents was eliminated.
Errors due to undesirable mixing were prevented.

The slider, shown in Figure 3.13, was constructed of
nylon to allow direct immobilization of enzymes to the
walls of the observation cell. In the previous design,
slight movements of the nylon tube at early reaction times
often caused large fluctuations in the instrumental out-

put. This problem was eliminated. Another advantage of

82

P1 P2

Fhfid

Li ht

 

Figure 3.13. Design of the new cell/reactor. Two
pneumatically actuated pistons (A) were
used to push the slider (B) back and forth
between the fill position (P1) and the
viewing position (P2). Quartz windows were
used (C) to contain the solution in the
viewing position.

83

using nylon sliders was that various reactor diameters
could easily be implemented by simply changing sliders.

One obvious disadvantage of this new method was the
increase in cost of materials and time. While nylon
tubing is fairly cheap, it is expensive to machine a
new nylon block for each experiment. Also, more diffi-
culty and time is involved in immobilizing enzymes on
these blocks with high precision. The greatest disadvan—
tage was found to be that the movement of the slider some-
times caused bubbles to form in the observation cell,
because the solution was lost from the ends of the cell
during changes in position. The development of bubbles was
infrequent and was detected by a large drop in instrumental
output. Several stopped-flow pushes usually restored the
original output. The results of a series of experiments
with distilled water are displayed in Figure 3.14, and
they show fairly good reproducibility in the cell
positioning.

The use of the instrument was demonstrated by
determining glucose with immobilized glucose oxidase.
The enzyme was bonded on two sliders-— one with a 1 mm
i.d. cell and one with a 2 mm i.d. cell. The immobiliza—
tion procedure was identical to that described above
(Chapter 2). A stopped-flow unit employing a T-mixer
and upstream stopping was used with the cell assembly.

Each syringe held 0.16 mL. Four reagent pushes and

84

 

 

o o o o o o
4" oo oo
3 —- o o 0
o
<1)
cm 6 ‘—
C
+1
0 —11—
> o
5—1...
_ oo 0
o
4 I I I I I I I I I I
O 5 1O 15 20 25

Figure 3.14.

Triol Number

Reproducibility of cell positioning.
Distilled water was pumped through the

cell at 0.46 mL/min. Each trial represents
movement of the cell from the viewing
position to the fill position and back again.

85

four wash pushes were employed to reduce sample carry-
over. One reagent was the glucose standards, while the
Trinder/peroxidase reagent filled the other syringe. The
Trinder reagent was composed of the following: 4 mg
peroxidase, 2 mM AAP, and 2 mM DCPS made up to 25.0 mL
with a pH 6.85 phosphate buffer, 0.05 M.

The precision of the progress curve determination is
shown in Figure 3.15. Both reactors displayed very good
precision. The determination of initial reaction rates
from repetitive experiments with three different glucose
concentrations gave an average relative standard deviation
of 7%.

A preliminary study of the effect of reactor diameter
on overall reaction progress displayed some interesting
results. Figure 3.16 displays the data for the 1 mm i.d.
reactor and Figure 3.17 for the 2 mm i.d. reactor. The
enzyme reactors were not prepared at the same time, nor
were the absorbances corrected for the effect of reactor
diameter (equation 3.14). Despite these possible errors,
it appears that the reaction was much slower in the larger
diameter reactor. The reaction in the 2 mm i.d. reactor
is about five times slower than that in the other reactor.
Equation 3.8 predicts about a three-fold change in the
reaction rate with this change in reactor diameter.
However, the diffusion—control model does not predict the

absolute reaction rate very well.

86

 

 

 

0.6 ——
—I—-
0) 0.4
U
C
o /
n 11
L.
O
(n
.O
< 0.2 ——
o.o — ~ I I I I I I
o 1 2 :5

Time (min)

Figure 3.15. The results of five progress curve
determinations with 0.44 mM glucose as
the sample. The 1 mm i.d. reactor was used.

1.2

Absorbance

 

87

(.40 320 leo 8'0

 

Figure 3.16.

 

Progress curves obtained with the 1 mm
i.d. reactor. The glucose concentrations
(mg/L) which were used are written at the
end of each curve.

88

um

1.0 -T-

Absorbance

 

 

 

 

Tinie (swim)

Figure 3.17. Progress curves obtained with the 2 mm
i.d. reactor. The glucose concentrations
(mg/L) which were used are written at the
end of each curve.

89

The enzyme concentration also changes with the
reactor radius. The concentration is directly dependent
on the surface-to—volume ratio or 1/r. The increase in
reactor radius from 0.5 mm to 1.0 mm could decrease the
reaction rate and also change the ratio of the enzyme
rate constant to the diffusion rate constant. In these
ways, variation in reactor radius can change the kinetics
indirectly. To keep the enzyme concentration constant
with changes in reactor radius, additional enzyme should
be bound to tubes of larger diameter.

The stopped-flow instrument should provide a fairly
direct route to the determination of inherent enzyme
properties. By varying the reactor diameter and enzyme
concentration, information about the enzyme reaction and
diffusional influences should be obtainable. Reactors
from these studies under static conditions can then be
incorporated into a continuous flow manifold. Differences
in kinetics between the two systems can be attributed to
the change in mass transport, from diffusion to primary
and secondary flow. Thus, definitive information con-
cerning continuous flow systems with integrated open-
tubular immobilized enzyme reactors should be forthcoming.
Much more research on the kinetics of such reactors is

needed.

CHAPTER 4

Colorimetric Determination of Glucose

Based on Immobilized Glucose Oxidase

The determination of blood glucose is the most
frequently performed test in clinical laboratories.
Because glucose is the major energy source for living
cells, any variation in the concentration of this sugar
has a large effect on cell function. The most common
disease related to carbohydrate metabolism is diabetes
mellitus; nearly two percent of the U.S. population has
this disease [95]. Glucose assays can reveal such abnor-
malities, and then treatment of the illness can begin.

Glucose is a monosaccharide, containing six carbon
atoms. The structure of D—glucose, the form that is found
in nature, is shown in Figure 4.1. The D—form has the
hydroxyl group on carbon #5 on the right side of the carbon
chain. The aldehyde group and the hydroxyl group on carbon
#5 react to form a hemiacetal and a six-membered ring.
The carbon #1 atom becomes asymmetric and, thus, can exist
in two anomeric forms, a and B, shown in Figure 4.1.
Mutarotation of glucose occurs slowly in aqueous solution
[96,97]. At equilibrium, the D—glucose exists as 64%
a—form, 36% B-form, and traces of the free aldehyde form
[95,98,99]. The normal range of blood glucose concentrations

is 70—105 mg/dL or 3.9-5.8 mM [95].

90

D-glucose

CHO
H-C-OH
Ina—EH.
1.42-01.
HJC-OH
£11,011

Aldehyde Form

   

a-D-glucose B—D—glucose

Figure 4.1. The three forms of D—glucose.

92

Because the determination of glucose has been so
important for so many years, a large number of methods
have been used. Non—enzymatic methods are based on the
reducing nature of D—glucose and include reactions with
ferricyanide ion, cupric ion, and o—toluidine [95]. These
methods are sensitive but relatively non-specific. The
enzymatic methods are much more selective. A list of the
three common enzymatic systems is given in Figure 4.2.
These systems have been used in immobilized forms. The
hexokinase/glucose-6-phosphate dehydrogenase method has
been adopted for use in the Technicon SMAC instrument, the
major clinical analyzer in hospital laboratories.

In this study, the glucose oxidase method was used
for several reasons. First, the enzyme has a high Km value
for glucose. This property ensures a large analytical range
of glucose concentrations. Second, the enzyme and other
reagents necessary for the assay are relatively inexpen—
sive. The hexokinase and glucose dehydrogenase methods
require B-NAD, which is expensive. Furthermore, commercial
preparations of the enzymes are more than fifty times as
expensive as glucose oxidase [103]. The reason that these
other methods are used is that they are somewhat less
susceptible to chemical interferences. Another reason
glucose oxidase was chosen is its high specific activity
(up to 150,000 Units/g) and the fact that it bonds well to

nylon tubing. Therefore, glucose oxidase is well—suited

93

GLDH

l) B-D-glucose + B-NAD ----- 9~ é—D—gluconolactone + B-NADH

pH optimum = 7.6 [100]
Measure B-NADH absorbance

at 340 nm
3- HK
2) D-glucose + P04 ----- 9. D—glucose-6-phosphate
D-glucose-6-phosphate+B-NAD 993§329g9>6-phospho-D-glu-

conate + B-NADH

pH optimum = 8.5 [101]
Measure B—NADH absorbance
at 340 nm

3) B-D-glucose + H20 + 02 __§Q_9. 6-D-gluconolactone + H202

pH optimum = 6.0 [102]
Measure hydrogen peroxide conc.

GLDH: glucose dehydrogenase (E.C. 1.1.1.47)
HK: hexokinase (E.C. 2.7.1.1)
GL-6-PDH: glucose-6-phosphate dehydrogenase (E.C. 1.1.1.49)

GO: glucose oxidase (E.C. 1.1.3.4)

Figure 4.2. The reactions involved in the enzymatic
assays for glucose.

94

for use in immobilized form.

Most assays using glucose oxidase are based on the
increase in concentration of peroxide or the depletion
of dissolved oxygen. Electrochemical methods have involved
an alternate oxidizing agent, p—benzoquinone [102,104],
the measurement of the decrease in oxygen concentration
with an oxygen electrode [105,106], and amperometric sen-
sing of the production of hydrogen peroxide [107-109].
Luminescence methods include the reaction of peroxide to
give chemiluminescence [110], the reaction of peroxide and
homovanillic acid to yield a fluorescent compound [111],
and the measurement of oxygen via its ability to quench the
fluorescence of pyrene [112]. A novel method based on the
absorbance or fluorescence of B—NADH using additional en-
zymes, catalase and aldehyde dehydrogenase, has also been
described [113].

Most often, colorimetric methods are used. Hydrogen
peroxide, in the presence of peroxidase, can oxidize many
aromatic compounds to give colored products. Under the
proper conditions, the absorbance of these molecules is
proportional to the initial D-glucose concentration. A
variety of color-forming reactions have been used; some are
listed in Table 4.1. Note that two of the methods do not
require peroxidase, making then less costly and less suscep—
tible to chemical interferences. A recent study presented

several more water—soluble reagents for reaction with

95

Table 4.1. Chromogens for Reaction with Hydrogen Peroxide

1

Reference

 

 

 

Chemical System :flfii €(AU M- cm_ )
AAP—DCPS 510 5 x 104
AAP-HE 500 -

DPAS 465 ~3 x 101
MBTH-CTA 572 -

MBTH-DMA 590 -
o-Dianisidine 410 -
o—Toluidine 630 —

KI/Moo42'* 365 2 x 104
Xylenol orange* 582 2.5 x 104

94,114,115

116

117

118

101,119,120

121

93

85

122

 

AAP: 4-aminoantipyrine

CTA: Chromotropic acid

DCPS: 2,4-dichlorophenol sulfonate
DMA: N-N-dimethylaniline

DPAS: p-diphenylamine sulfonate
HB: p-hydroxybenzoate

MBTH: 3-methyl-2—benzothiazolinone hydrazone

* - peroxidase is not required

96

hydrogen peroxide [123]. The multitude of available
chromogenic agents reveals the fact that no one method is
significantly better than the others.

The 4-aminoantipyrine (AAP) - 2,4-dichlorophenol
sulfonate (DCPS) system was chosen for this study. The
product has a high molar absorptivity, the reaction occurs
rapidly, and the reagents are readily available and water
soluble. A comparative study of eight chromogens showed
that systems involving AAP were most suitable [124].

Also, a variety of phenolic compounds are acceptable
coupling agents for AAP [125]. Trinder and coworkers first
reported the use of AAP-DCPS in the determination of glucose
[90,91]. Since that time, the reagent was been termed the
Trinder reagent and has enjoyed wide use.

This chapter describes the characterization of an
immobilized glucose oxidase/Trinder determination of glu—
cose. The chemical reactions and continuous flow instru-
mentation involved in this study are reported in Section A.
The Trinder color—forming reaction is explored in more
detail in Section B. Next, experiments concerning the
effect of reactor design on the assay sensitivity (Section
D) and the characterization of the glucose oxidase reaction
(Section C) are discussed. Finally, the sensitivity of
the determination was improved by adding sodium azide to
the system. The reasons for doing this are given in

Section E.

97

A. The Method and Instrumentation
1. The glucose oxidase reaction
The reaction which is catalyzed by immobilized

glucose oxidase, the analytical reaction, is given below.

B-D-glucose + 02 + H20 + 6-D-gluconolactone + H202 (4.1)

The enzyme is specific for the B-form of the substrate.
Glucose oxidase reacts much more slowly with the a-form
and other sugars [126]. The enzyme considered here is

extracted from Aspergillus niger and contains two flavine

 

adenine dinucleotide (FAD) groups per molecule [127]. The
reaction involves only the fully oxidized and fully reduced

forms of the enzyme in the following manner [128].

E-FAD + B-D-glucose I E-FAD—glucose
E-FAD-glucose + H20 2 E-FADH2 + 6-D-gluconolactone

E-FADH2 + 02 + E-FAD + H202

The mechanism is a ping-pong bi-bi type, and, consequently,

the rate equation is of the following form.
v = k [E] /(1 + KG/[Glucose] + KO/[Oxygen]) (4.2)
e o m m

The reported kinetics constants for glucose oxidase
have recently been reviewed [129]. The mean values for
Kg, K3, and ke are 9 xlo’2 M, 6 1110'4 M, and 1.2 x102 s"1

respectively. Several compounds have been found to be

inhibitors: 8-hydroxyquinoline, sodium nitrate,

98

semicarbazide [128], a-D—glucose [130], and hydrogen
peroxide [131], among others.

The glucose oxidase preparations used here were
purchased from Sigma Chemical Company. Type II enzyme,
obtained several years ago, was used for the studies
reported in Sections A-D of this chapter. The enzymes
used in the experiments described in Section E were all
purchased in 1982. This distinction is important because
the two lots of enzymes displayed different properties.

2. The Trinder/peroxidase reaction

Because neither the substrates or the products of the
analytical reaction absorb light in the visible region,

a second, color-forming reaction is required. Hydrogen
peroxide is a good oxidizing agent (H202 + 2 H+ + 2 e- +

2 H20, E° = 1.77 V), and is usually chosen for further
reaction. The reaction of hydrogen peroxide and a hydrogen
donor is slow, however, so a soluble enzyme, horseradish

peroxidase, is used as a catalyst. The reaction mechanism

is given below [132,133].

E + H O I Complex I

2 2
Complex I + AH2 1 Complex II + AH
Complex II + AH + E + A + 2 H20
Here AH2 represents the hydrogen donor. Excess peroxide

can combine with Complex II to produce an inactive enzyme

species. The substrate specificity of the free enzyme is

99

high; only peroxide, formic acid, and acetic acid yield
complexes that are active [133]. However, the speci-
ficity of the enzyme-substrate complexes is poor. Many
hydrogen donors can participate, including phenols,
aromatic amines, ascorbic acid, and leucodyes.

The overall reaction in the case of the Trinder

reagents is given below.

Cl (C) c' 0 CH3
2 H202 + c.@.o.. + N\ :99 GN£N@+ 3 H20 + H250,
503" (IQ-c”: CI "3° \CH3 (4 ' 3)
H
DCPS AAP

2H CH:

The exact coupling mechanism of the two organic compounds
is not known. A possible mechanism is that each reagent
first loses two electrons and two hydrogen ions to the
hydrogen peroxide, forming water and a reactive intermediate.
Then the intermediates instantaneously react to form the
dye product. Therefore, two hydrogen peroxide molecules
and, consequently, two glucose molecules are required for
the production of one molecule of dye. This fact is impor-
tant when the molar absorptivity of the product or the
percent glucose conversion is calculated.

The peroxidase used in the following experiments was
purchased in 1982 from Sigma (Type II) and contained about
200 Units/mg of solid.

3. The continuous flow manifold
The analytical and color-forming reactions occur in

the manifold of a continuous flow instrument. The reagents

100

are added, mixed, and reacted in a flowing, gas-segmented,
liquid stream. Figure 4.3 is a diagram of the reaction
manifold, showing the order and rate in which the reagents
are added. A peristaltic pump meters the reagents. The
composition of the reagents is considered below in the order
in which they are added to the flow stream.

Sample and Wash Reagents. In continuous flow methods,
the samples and another liquid are alternately injected
into the flowing liquid stream. This other liquid is called
the "wash” solution, and in the following studies was
distilled, deionized water (DDW). The wash solution mini-
mizes sample carryover by rinsing the manifold tubing and
flow cell between consecutive samples. The sample:wash
timing was usually 40 seconds:20 seconds. Thus, the sample
throughput was 60 samples/hour.

The samples were either hydrogen peroxide or glucose
standards. A hydrogen peroxide stock solution was pre-
pared by diluting 0.50 mL of 30% peroxide solution to 0.500
liters with DDW. This solution was standardized by reaction
with acidic KI and subsequent titration with standard thio-
sulfate [134]. Working standards were made daily by appro-
priate dilution of the stock solution with DDW. The primary
glucose standard was prepared by dissolving 2.000 g of
anhydrous, granular a-D-glucose and 0.5 g of benzoic acid
in about 750 mL of DDW contained in a 1.000 L volumetric

flask. The resulting soultion was then diluted to the

101

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NOMINAL
REAGENT FLOW RATE
(mL/nfin)
0.10
0.05
G“ s T 0mm: 20 1 was“
Sample 0-23 0909 9000 0090 f
Buffer A I
Trinder/P0 0-10
510 nm

Figure 4.3. The continuous flow reaction manifold.
The pump speed setting which gave nominal
flow rates was 42.

102

mark with DDW and allowed to stand at room temperature
for 24 hours to ensure complete mutarotation. Working
glucose standards were prepared by appropriate dilution
of the primary standard with DDW. The hydrogen peroxide
and glucose stock solutions were stored in a refrigerator
at 4 °C.

Gas Segments. The sample and wash slugs were seg-
mented with gas bubbles using dual pump tubes. The
"helper" tube at a nominal flow rate of 0.10 mL/min builds
up pressure in the second, smaller i.d. tube, and the
bubbles are forced rapidly into the liquid stream. This
procedure helps phase the segmentation with the pump roller
lift-off. Phasing of the gas bubbles improves the system
characteristics. Room air was used unless otherwise noted.
The segmentation rate was 1.67 bubbles per second at a
pump speed setting of 42.

Buffer Reagent. The next reagent added to the mani-
fold is a buffer which provides the optimal hydrogen ion
concentration for the enzymatic reactions. Potassium
phosphate buffers (0.1 M) were used in these studies. The
pH optimum of the chemical system was found to be 6.85.
The experimental evidence is presented in Figure 4.4.
Several drops of Brij—35 wetting agent were always added
to the pH 6.85 buffer. This agent promoted pulseless flow
through the manifold and preserved the integrity of the

gas bubbles.

103

Figure 4.4. The pH dependence of the glucose assay.
A 0.1 M phosphate buffer was used to
control the pH. The dotted line indicates
the approximate relationship in the
absence of many data points.

104

Glucose Oxidase Reactor. After mixing in a five-turn
glass coil, the sample and buffer were directed into the
immobilized enzyme reactor. The preparation of the glucose
oxidase reactor was detailed in Chapter 2, Table 2.3. The
analytical reaction occurred in the presence of the bound
glucose oxidase. The percentage of glucose that was reacted
in the enzyme loop was never more than 10%. The product,
hydrogen peroxide, was swept out of the reactor by the
flowing liquid stream.

Trinder/Peroxidase Reagent. Finally, the Trinder/
peroxidase reagent was added. The reagent was prepared
as follows. A 10 mM stock solution of 4-aminoantipyrine
(AAP) was pepared by dissolving 1.02 g of the solid (Sigma)
in 0.500 L of DDW. A 10 mM stock solution of 2-hydroxy-
3,5-dichlorobenzenesu1fonate was prepared by dissolving
1.33 g of the solid (Research Organics) in 0.500 L of DDW.
The working reagent was made by dispensing 5.00 mL of each
of these stock solutions and 5 mg of peroxidase into a
50.0 mL volumetric flask and then diluting to the mark with
the pH 6.85 phosphate buffer. This composite reagent was
prepared daily. A 20-turn mixing coil (about 100 seconds
residence time at nominal flow rates) was used at the exit
end of the manifold to ensure a complete reaction between
the peroxide and the Trinder reagents. The pink dye was

then passed into the flow cell.

105

4. The continuous flow instrument

A modular single-channel continuous flow instrument,
diagrammed in Figure 4.5, was used in all of the experiments.
A Brinkmann IP-12 peristaltic pump, modified to accomodate
sixteen rollers, was used to proportion the reagents. The
pump speed setting that provided nominal flow rates was 42.
This speed was used unless otherwise mentioned. The mani-
fold was constructed from 1.0 mm i.d. glass mixing coils,
injector fittings, and a 1.0 cm flow cell (0.5 mm i.d.,
Technicon). Sampling was performed manually.

The colorimeter was designed and built in this labor-
atory. The light source was a miniature tungsten—halogen
lamp (Model 03000, Welch-Allyn), and the light was trans-
mitted to the flow cell via a glass fiber optic bundle
(Model EK15-12, Dolan-Jenner). Wavelength selection was
accomplished with a 510 nm interference filter (Ditric
Optics, bandpass = 8 nm) mounted between the exit window
of the flow cell and the detector. The detection system
consisted of a photodiode (Princeton Applied Research) and
an operational amplifier current-to-voltage converter.

A bubble-gate removed the air segments electronically [135].

The data were collected and stored on a strip-chart
recorder or in RAM by a microcomputer [86]. The data
stored in the microcomputer were shipped to a PDP-8/e
(Digital Equipment Corporation) minicomputer following each

experiment for long-term storage and computation of the

106

 

 

Peristaltic

 

 

—

Reacfion

Calorimeter

 

 

 

 

_ Manifold

 

UWW

Reagents

Sampling

Pulse

 

 

>0 O——>Waste

 

 

 

 

 

 

Bubble Gate

 

 

 

 

 

Microcomputer

 

 

Data
Lines

 

 

Chart Recorder]

 

Figure 4.5. Diagram of the continuous flow instrument.

107

steady-state absorbance values. The data were initial
sample concentrations and their associated absorbance
values. In most cases, a calibration curve of absorbance
versus analyte concentration was constructed for a series
of standards. The slope of the calibration curve was used
as a measure of the immobilized enzyme activity. A larger

slope meant that the assay conditions had improved.

B. Experimental Studies of the Trinder Reaction

With coupled enzyme reactions, the rate of one of the
reactions can be measured accurately (< 2% error) if the
two reaction rates differ by a factor of 50 or more.
Consequently, the Trinder reaction must occur at a high
velocity so that the measured rate depends strictly on the
glucose oxidase reaction. Another way to prevent errors is
to allow the indicator reaction to go to completion. This
is only possible, however, when the two reactions are
separated in time, as they are in this determination. Thus,
conditions were adjusted so that the color-forming reac-
tion went to completion.

The concentrations of the Trinder reagents and the
peroxidase were set as high as possible without unduly
increasing the cost of the assay. An indicator reagent
containing 6 mM AAP, 5 mM DCPS, and 80 Units/mL of per-
oxidase was prepared and used in the continuous flow

manifold. Likewise, 1:5 and 1:10 dilutions of this reagent

108

were prepared and used. The least concentrated reagent
only produced a 3% smaller instrumental response than did
the most concentrated. Thus, concentrations of 1 mM AAP,
1 mM DCPS, and 20 Units/mL of peroxidase were chosen as
optimal; the reaction approached completion without
wasting reagents.

The optimal reaction time for the Trinder reaction
was determined. With peroxide standards as the sample and
the glucose oxidase reactor removed, the effect of reaction
time on the instrumental response was small but noticeable.
Assuming that the reaction was complete after 200 seconds,
96% of the 50 uM H202 sample was reacted within the first
100 seconds. Even at a pump speed setting of 56 and a 20-
turn glass mixing coil in place (about a 70 second reaction
time), the reaction was more than 90% complete.

Results of stopped-flow studies of the Trinder reaction,
shown in Figure 4.6, revealed that reactions with peroxide
solutions of 50 uM or less were more than 90% complete in
about thirty seconds. The Trinder reagent and peroxidase
concentrations were the same as described in Section A.

The hydrogen peroxide, 0 - 50 uM, reacted according to
pseudo-first order reaction kinetics, with a rate constant
of 0.09 sec—1. This confirmed that the Trinder reagents
were in large excess and that the mechanism involving a sin-
gle peroxide molecule, given above, is correct. In all the

studies presented here, the Trinder reaction was made to

0.9 -
<1)
0
C.
O
.C) 0.6-
x.
O
(I)
.0
<1
0.3 -
0.0

Figure 4.6.

 

109

 

 

 

 

A

B

C

D

. l 1 I . L, 1 l . l
' I ' I ' I ' I I I
7 14 21 28 35

Thne (seconds)

Progress curves for the reaction of hydrogen
peroxide and the Trinder reagent. The
peroxide concentrations were 50 uM (A),

25 uM (B), 10 uM (C), and 5 uM (D).

110

approach completion with the use of excess reagents and
long reaction times.

The molar absorptivity of the Trinder product was
determined. A series of hydrogen peroxide solutions of
known concentration were tested, and the resulting steady-
state absorbance values were recorded. To account for
dilution in the manifold, the actual flow rates were cal-
culated by weighing the amount of water delivered by the
various pump tubes over a known period of time. The molar

absorptivity at 510 nm was calculated to be 5 i 1 x104

AU M_1cm-1. This does not compare well with the value of

1.2 x104 AU M-lcm"1 reported in the literature [136]. The
reason for this discrepancy is not known.

In almost all reports concerning the GO/Trinder assay,
several substances are mentioned as interferences. Uric
acid, ascorbic acid, gentisic acid, and L-DOPA are most
often cited [94,115,136,137]. These interferences occur
primarily because the peroxidase-peroxide complex does not
react specifically with a single reducing agent. The
severity of the problem can be lessened by performing a
protein precipitation on the serum sample. A common
precipitating agent for glucose assays is the Somogyi
reagent [138]. A stepwise treatment with Ba(0H)2 and then
ZnSO4 precipitates the proteins along with uric acid and

other metabolites, but ascorbic acid remains in the super-

natant.

 

111

Ascorbic acid, added to the CF manifold, lowered the
instrumental response. Figure 4.7 reveals the effect of
ascorbate on the apparent peroxidase activity. The glucose
oxidase reactor was not used, and a series of peroxide
standards were sampled. The loss in assay sensitivity may
be due to inhibition of peroxidase, loss of peroxide, or
other factors. Stopped—flow studies were performed to
determine the exact nature of the ascorbate interference.
Ascorbic acid absorbs light at 261 nm; its oxidized form
does not [93]. By monitoring the decrease in the ascor—
bate concentration, it was found that hydrogen peroxide and
ascorbic acid react rapidly, but only in the presence of
peroxidase. Thus, ascorbate lowers the concentration of
peroxide in the glucose assay. Further evidence was pro-
vided by studying the formation of the Trinder product in
the presence of ascorbate. A lag phase which increases with
increasing ascorbate concentration is evident in Figure 4.8.
The rate constants, measured after the lag period, were
nearly independent of the ascorbate concentration. These
facts suggest that ascorbate acts as an alternate substrate
for peroxidase and, in this way, inhibits the enzyme.

The problem of removing the ascorbate interference is
a difficult one. Prior oxidation to dehydroascorbic acid
is possible, but this risks oxidation of the glucose. Also,
remaining oxidizing agents may react with the Trinder solu—

tion. One researcher has used ascorbate oxidase to remove

112

 

 

 

29-—
A
I
E
D
<
.0 6
O
:2
v
a) u—lI-
O.
O
(f)
c: 3——
O
1.:
C
L.
.0 —I—
B
O
I l I I
O I I I I

O 10 20 30 4O
[Ascorbote] (uM)

Figure 4.7. The effect of ascorbate concentration
on the assay sensitivity.

4.0 —

113

00

 

 

3.0 -
(‘4-
\\
.4
L_ 2.0 —
:5
C
1.0 —
0.0
Figure 4.8.

 

-II-
I-
1-
—I—
-l.
.3
-I-
b

Time (seconds)

First order plots of data for the hydrogen
peroxide-Trinder reaction in the presence
of ascorbate. The ascorbate concentrations
were 10 uM (A), 20 uM (B), and 40 uM (C).

A was the absorbance at equilibrium, and A
was the absorbance at any time, t.

114

ascorbic acid [136], but this approach is very expensive.

A satisfactory solution has not been found.

C. Reaction Characteristics
1. Segmentation gas composition

In analytical reaction rate determinations, a linear
relationship between the rate and the analyte concentration
must be established. In the case of glucose, the Trinder
reaction has a negligible effect, and, thus, the reaction
rate is described by equation 4.2. In order to attain a
rate that is only proportional to the glucose concentration,
the oxygen concentration must be larger than 20 K: or
1 x10.2 M, and the glucose concentration must be smaller
than Kg/ZO or 4.5 x10.3 M.

The concentration of glucose after dilution or dialysis
of the sample is usually well below 4.5 mM. The concentra—
tion of oxygen in 0.1 M phosphate buffer saturated with air
is 0.26 mM at 22 °C [139]. This value is less than the
K3 value, and, therefore, linear calibration curves should
not be possible. This is not the case, however. A large
source of oxygen is provided by the air bubbles which
segment the liquid stream. To prove this, the effect of
segmentation gas composition on the determination was studied.

The manifold and reagents were used as described in

Section A. The gases, N and 0 were bled from pressur-

2 2’
ized tanks and allowed to fill a large bottle. The

115

appropriate pump tube was placed in the bottle. Room air
was the source of air. A series of glucose standards was
sampled, and the resulting calibration curves are shown in
Figure 4.9. When the assay is entirely dependent on dis-
solved oxygen, the analytical range of glucose concentra-
tions is limited to 0.5 mM or less. The linearity improves
with increasing oxygen concentration. With air segmentation,
samples containing less than about 3 mM glucose can be
tested; this is well within the range for blood glucose
analysis.

The gas composition also affected the continuous flow
peak shapes. When nitrogen gas was used, large Spikes
occurred at the falling edges of the peaks. See Figure 4.10.
These spikes were due to air bubbles which entered the sam-
pling system as the probe was moved between the sample and
wash solutions. With the extra oxygen, the reaction could
occur to a greater extent. Small spikes also occur with
the use of air segments.

2. The effect of ionic strength

The ionic strength of the solution passing through the
glucose oxidase reactor had little effect on the enzyme
activity. The usual continuous flow manifold was used.

The buffer concentration was lowered to 1.0 mM, so that the
ionic strength was controlled by adding KCl to the sample
and wash solutions. Chloride ion is not known to affect

either of the enzymatic reactions. The ionic strength

116

 

 

1.250 —-
0 2
O O
1.000 —— AI r
0)
U ...—
: 0.750 N
C
D 2
L
O
m __
D 0.500
<
0.250 -—-—
0.000 I I I I I I I I a I
0 1 2 3 4 5

[Glucose] (mM)

Figure 4.9. Calibration curves for a series of glucose
standards using three different segmentation
gases: NZ, Air, 02.

ABSORBANCE

 

117

N2 AIR 02

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 4.10.

TlME

The output from the continuous flow
instrument using three different seg-
mentation gases. The glucose concentra-
tions were 0.44 mM, 1.44 mM, 2.22 mM,
and 4.44 mM. Only the first three
concentrations were sampled using 02.

 

118

through the final mixing coil, in which the color-
forming reaction occurred, was held nearly constant by
preparing the Trinder reagent in 0.1 M buffer. Conse-
quently, the final ionic strength only varied from
0.010 M to 0.015 M.

Figure 4.11 shows an initial increase and then a
slow decline in instrumental response as the ionic strength
is increased. Only a small change was expected since the
substrates are uncharged. However, since the carrier was
positively charged, enzyme-carrier interactions may have
been altered. According to equation 3.3, an increase in
ionic strength should increase the Km value and lower the
reaction rate. This study shows the same general trend.
3. The specificity of glucose oxidase

The oxidation of most sugars, except B-D-glucose, is
catalyzed very slowly by glucose oxidase. Nine different
sugars, obtained commercially in the highest purity forms,
were sampled, first with the immobilized glucose oxidase
reactor in place, and then with soluble glucose oxidase
added to the buffer reagent (0.2 mg/mL) and the enzyme
reactor removed. The continuous flow manifold and the
other reagents were the same in both experiments. Table
4.2 is a list of the results. No sugar other than B—D—
glucose showed high reactivity toward glucose oxidase. The
aqueous and immobilized enzyme results were remarkably

similar. This indicates that the conformation of the enzyme

119

 

 

 

0.3 m—-
l __
E
I)
< M
r!) 0.2 "" *3
O
3?
a) -r-
0.
_°
(.0
c; 0.1 ""_
.2
‘6
L.
:9 l_
B
U
l l l I 1 4| 1 l I J
0.0 I l T I l l I l I 1
O 10 20 30 4O 50

Figure 4.11.

Ionic Strength (an)

The effect of ionic strength on the
glucose oxidase reaction.

120

Table 4.2. Specificity of Glucose Oxidase

Relative Reactivity

 

Sugar Immobilized Agueous Aqueous*
B-D-glucose 100 100 100
2—deoxy—D-glucose 2.5 2.4 25
Maltose 1.1 0.95 0.19
a—D-glucose 0.6 - 0.64
Xylose 0.31 0.26 0.98
Mannose 0.27 0.23 0.98
Galactose 0.03 0.05 0.14
Fructose 0.03 0.02 -
Glucosamine 0.03 0.02 <0.05
Sucrose 0.00 0.00 -

* - Values reported in the literature [126]

 

121

was not significantly altered during immobilization.
Correlation of the results with an aqueous enzyme study,
performed in 1952, was fair. The literature values may
be inflated by impure reagents; the amount of B—D-glucose
in various sugar preparations is probably much lower today.
4. Mutarotation kinetics

To replace the B-D—glucose which is converted to
hydrogen peroxide, the a-form of D—glucose is converted
spontaneously to the B—form. This process is termed muta-
rotation. The chemical equation and associated rate

equation are shown below.

B-D-glucose I a-D—glucose (4.4)

-d[B]/dt = k1[B] — k2[d] (4.5)

If [B]O-[B] is substituted for [a] ([B]O is the initial
concentration of the B—form) and the equation integrated,

the rate equation appears in the following form.

-1n(1.565[B]/[B]O - 1) = (k +k2)t (4.6)

1

The term, k +k is identified as the ”mutarotation

1 2’
constant" [96,97].
A 4.44 mM B-D-glucose solution was prepared, and
immediately the solution was sampled with the continuous
flow instrument. Steady-state absorbance values were

obtained for the glucose solution at intervals over a two

hour period. Assuming that the a-form had a negligible

122

reaction rate with the immobilized glucose oxidase, first
order data were calculated and plotted in Figure 4.12.
The mean rate constant was determined to be 0.017 r 0.002

min , in good agreement with the literature value of
0.012 min‘1 [96,97].

The specificity value for d-D-glucose was also
obtained. Initially, only the B-form existed in the
sample solution, and, thus, the sensitivity toward the
B-form was calculated. Knowing this and that 64% of the
glucose is of the B-form at equilibrium, reaction rate

measurements at equilibrium allowed calculation of the

a—form reactivity. The following equation was derived.

S /S = 2.78(A /A ) - 1.77 (4.7)
a B m 0

Here 80 and S are the sensitivities of glucose oxidase

8
toward the a and 8 forms of glucose, and Am and A0 are
the equilibrium and initial absorbances. The reactivity
of the d-form was only 0.6% of that for the B—form.
Glucose oxidase shows very high specificity indeed.

The effect of glucose mutarotation on the overall
reaction kinetics is small because the mutarotation rate
constant is about ten times smaller than the glucose
oxidase reaction rate constant. If one considers a 30
second residence time in the enzyme reactor, up to 10%

of the B-D—glucose is consumed by the enzyme, while only

about 0.5% of the a—form can be converted to the B-form in

 

-|n(1.565 [Bl/[1810 -1)

 

Figure 4.12.

123

 

l 1 l
I ' l

1 l
' I
30 60 90

‘I’

Thne (nfinutes)

First order plot of data for the muta-
rotation of a 4.44 mM B-D-glucose solution
at 22 °C. The results of two independent
experiments are presented. [8] is the
concentration of the B-form of D-glucose
at any time, t, while [B]O is the initial
B-form concentration.

124

the same time period. Consequently, once the standard
glucose solutions have come to equilibrium, the muta-

roatation process can be neglected.

D. Reactor Design Characteristics
1. Coiling diameter

Immobilized enzyme reactors are usually molded into
helical coils to make them easier to incorporate into a
continuous flow manifold. The coiling of the reactor
should also improve the radial mixing in each liquid seg-
ment [83]. The reaction rate should increase with better
radial mass transport, and, thus, with more tightly wound
reactors. Horvath and coworkers found that the apparent
activity of trypsin immobilized inside a nylon tube improved
as the coiling diameter decreased [140]. However, this
effect only occurred at flow rates greater than 3 mL/min.

By improving the radial mixing, coil formation de-
creases sample dispersion, or the ”wash" of the system.
Ideally, the rising portion of a continuous flow peak can

be described by the following equation [1].

H(t) = H(ss)(1 - exp(-t/b)) (4.8)

Here H(t) is the absorbance at any time along the peak,
H(ss) is the steady-state absorbance (flat portion of the
peak), t is the time, and b is a measure of the magnitude

of the sample dispersion. The term, b, is called the wash

125

value of the system and is defined as the time required
for the absorbance to change from H(t) to 0.37 H(t).
Smaller b values are desirable and can result from tighter
coiling of the mixing loops and enzyme reactor.

A 25 cm glucose oxidase reactor, coiled to various
diameters, was inserted into the manifold, and a series of
glucose standards, 0.4 to 4.4 mM, was sampled. The manifold
and reagents were the same as those described in Section A.
The total flow rate through the reactor was only 0.38 mL/min.

The results are presented in Figure 4.13. The reaction
rate was unaffected by the changes in the reactor design.
This indicates that either the reaction rate is limited by
the enzyme kinetics or that radial substrate transport was
not improved significantly at the low flow rate. The latter
explanation is in agreement with Horvath's work [140].

The wash, however, was improved by about 30%. Thus, small
coiling diameters are desirable, but the coiling process
does not have to be very precise to obtain good analytical
results.

2. Coil temperature

An experiment was performed to determine the effect
of temperature on the apparent activity of glucose oxidase.
The manifold and reagents were the same as those described
in Section A. The enzyme reactor was heated by gluing the
reactor into a plastic drying tube, sealing the drying tube,

and then passing heated water through the tube. The water,

126

 

 

 

1.2 Er- WA 7— 9.0
U —6
l __ _r_
‘2
2:)
4
N 0.8 -h -- 6.0
O
32
a) " 1'
o.
2
(f)
c 0.4 m—' ‘—'3.0
2
B
L...
:9 .. _-
B
O
L l l I l l
0.0 I l 1 I 1 l 0.0
O 20 4O 60

Figure 4.13.

 

CoiHng [Morneter (nun)

The effect of coiling diameter on the
calibration slope (O) and wash ([3).

(spuoaas) usoM

 

127

contained in a 2 liter beaker, was thermostatted using a
Haake Model E52 heating bath/circulator unit. Since the
color-forming reaction went nearly to completion, tempera-
ture had little effect on the Trinder reaction.

The temperature dependence of the analytical reaction
is shown in Figure 4.14. The results do not conform to the
Arrhenius equation. It is also interesting to note that
at room temperature (22 °C) a small temperature shift
can change the reaction rate by several percent. This fact
could explain the fluctuations of the activity of enzyme
reactors in long—term stability studies (e.g. Figure 2.5).
The stability of the enzyme reactor at high temperatures
was tested by heating the reactor to 41 °C, and sampling
glucose standards over a one hour interval. The initial
and final results were identical, which indicates that
immobilized glucose oxidase is very stable with respect
to temperature.

E. The Characterization and Elimination of the Effect of
Catalase
1. The nature of the interference

In the glucose determination described here, the
hydrogen peroxide produced in the immobilized enzyme tube
is reacted downstream with the Trinder/peroxidase reagent.
The analytical and color-forming reactions are separated
in time for two reasons. First, separation allows the

reactions to be optimized individually in terms of reaction

 

128

4.00 -—

3.00

2.00

 

1.00 ——

CaHbrofion Slope (x102 AU hi—1)

Mb
I...

0.00

 

‘—
"ll"
_L

15.0 25.0 35.0 45.0

Tenwperature (°C)

Figure 4.14. The dependence of the assay sensitivity
on temperature.

 

 

129

time, reagent concentrations, pH, etc. In soluble enzyme
methods, a compromise between the optimal values for each
reaction is necessary. Second, the separation improves the
wash characteristics of the system. The Trinder compounds
adsorb strongly to the nylon tubing, which degrades the
wash. These two factors prohibit the simultaneous reaction
of glucose and peroxide inside the immobilized enzyme
reactor.

Because of the separation of the two reactions, the
sensitivity of the method was found to be reduced due to
contamination of the glucose oxidase by catalase (E.C.
1.11.1.6) [141]. Catalase destroys hydrogen peroxide via

the following mechanism [142]:

+
E + H202 + Compound I

Compound I + H202 + E + 02 + 2 H20

where E is free catalase. Some of the hydrogen peroxide
that otherwise would react to form product is destroyed
inside the enzyme reactor by co-immobilized catalase. Since
catalase has a specific activity of forty million units

per gram [103], the small amounts found in most glucose
oxidase preparations will have a large effect. If the
color-forming reaction also occurred inside the reactor,
then the effect of the catalase would be small since the
peroxide would quickly be converted to the colored product

by aqueous peroxidase. This is not the case in CF systems.

130

In a method proposed recently, the catalase was
selectively inhibited by adding sodium azide to the
system [141]. Hydrazoic acid is well known as an inhibi-
tor of catalase [142-144]. Although the exact nature of
the inhibition is still in question, it is known that
hydrazoic acid reacts with Compound I to yield peroxide
or other products. Thus, the addition of azide to the
reaction system inhibits catalase by forming a compound
with the enzyme-substrate complex.

A study of the effect of sodium azide on the co-
immobilized catalase is presented here. The improvement
in the assay sensitivity with the addition of azide and
the variation in catalase content of several commercial
glucose oxidase preparations are described in detail.
Finally, suggestions for the detection and elimination of
the catalase interference are made.

2. The reagents and continuous flow instrument

The reagents were prepared as outlined in Section A.
Sodium azide solutions were prepared in the following
manner. A 250 mM stock solution was made by dissolving

3.25 g of NaN in 0.200 L of pH 6.85 phosphate buffer (0.1 M).

3
This solution was stable for several weeks at room tempera—
ture. Working azide solutions were prepared by appropriate
dilution of the stock with buffer. Five drops of Brij-35

surfactant were added to these solutions prior to use.

The azide solutions were substituted for the buffer reagent

131

in the continuous flow manifold. The term, "azide” is
used to mean sodium azide.

The concentrations of azide, hydrogen peroxide, and
glucose reported here are the concentrations actually
present inside the immobilized enzyme reactor, after dilu-
tion in the continuous flow manifold. The instrument was
the same as outlined in Section A. The data were acquired
by the computer in all the experiments, and the pump was
operated at a speed setting of 56.

3. Enzyme immobilization

The catalase (Type C-10) and glucose oxidase (Types II,
IX, and X) were purchased in 1982 from Sigma Chemical Co.
and used without further purification. The enzymes were
immobilized as outlined in Chapter 2, Table 2.3. The seven
enzyme reactors used in this study were prepared with the
aqueous enzyme solutions listed in Table 4.3. The listed
catalase concentrations were derived from the catalase found
in the glucose oxidase and from the amount of pure catalase
added to the solutions -- reactors A, C, and D. The
different reactor types will be referred to by letter names
as listed in Table 4.3. When not in use, the reactors were
filled with pH 6.85 phosphate buffer and stored at 4 °C.

4. Effect of azide on aqueous peroxidase

A series of experiments was performed to investigate

the effect of azide on the color—forming reaction. In the

first experiment, the immobilized enzyme reactor was removed

132

Table 4.3. Solutions Used to Prepare the Immobilized
Enzyme Reactors

 

 

 

 

 

Reactor Type GO (U/mL) CAT (U/mL) Sigma Type GO
A 0 3765 -
B 89 0.4 II
C 89 380 II
D 89 1880 II
E 125 0.58 II
F 125 1200 IX

G 125 1.7 X

133

from the manifold, and peak heights were obtained for a
79 0M peroxide standard. Then the experiment was re-
peated with increasing amounts of azide added to the
buffer reagent. The results, shown in the upper curve of
Figure 4.15, revealed that for azide concentrations of
up to 30.3 mM the sensitivity of the assay remained con-
stant within the limits of experimental error (about 3%).
Thus, azide does not appear to inhibit peroxidase activity
or interfere with the color-forming reaction.
5. Effect of azide on immobilized catalase

In the next series of experiments, the ability of
azide to inhibit catalase activity was investigated. A-
30 cm immobilized catalase reactor (reactor A in Table
4.3) was inserted into the manifold, and peak heights were
obatined for a 79 0M peroxide standard with increasing
concentrations of azide in the buffer. The results of this
experiment are also shown in Figure 4.15. Without azide
in the buffer, more than 99% of the peroxide was destroyed
as it passed through the catalase reactor, but as the azide
concentration increased, peroxide recovery improved drama-
tically. Even at an azide concentration of 15 mM, however,
20% of the peroxide was destroyed.

The manner in which azide inhibited catalase was
investigated in the next set of experiments. Shorter
catalase reactors (type A) were used in these studies

so that the mean residence time of peroxide in the enzyme

134

 

 

 

 

100M 1']

,0 80—— U

(I)

L.

a

>

o

o __

a) 60

a:

o

E

x 40-

E

Q)

a.

N 20—

1 1 1 l 1 I

or”1 T I I l I l
O 10 20 30

Afide Concentnfljon OnM)

Figure 4.15. The effect of azide on the amount of
hydrogen peroxide that was converted to
product. A 79 uM H 0 solution was sampled
without an enzyme reactor ([3) and with a
30 cm catalase reactor (0) in the manifold.

 

135

reactor was about 3.5 seconds. Under these conditions,
Michaelis-Menten kinetics were applicable. A new 5 cm
segment was cut from a fresh 50 cm catalase reactor and
inserted into the manifold for each experiment. The dif-
ferences between peak heights obtained with and without
the catalase reactor in the manifold were determined for

a series of hydrogen peroxide standards in the range of

20 0M — 80 0M. The decrease in steady-state absorbance
with the catalase reactor in place was proportional to the
catalase-induced rate of peroxide decomposition.

Figure 4.16 shows Lineweaver-Burk (L-B) plots of the
results of three separate experiments. All lines inter-
sected in a small region just below the x-axis. Dixon
plots [145] of the data were hyperbolic, and replots of the
L-B intercepts and slopes were non-linear. These results
indicate a mixed-type noncompetitive inhibition [146]; in
addition to interaction with Compound I, azide may also
bind at two sites on the free enzyme. The kinetics were
complicated, and more experiments would be required to
completely elucidate the action of azide on immobilized
catalase.

We next investigated the extent of catalase interfer-
ence under non-separated reaction conditions. The Trinder/
peroxidase reagent was passed through the 5 cm catalase
reactor (type A). The results revealed that peroxidase

successfully captured 80% of the available peroxide in the

 

136

\\ g‘

l
I

G
‘\

\\

A
l
I

K‘

‘0

o
>

O
l 1
I I
\
1x1. >0 o\\~ a a
0‘ Q.

l/Rate (x105 M—Is)

 

 

-4 I I l
—4 0 4

1/[H202] (x104 M“)

.0-

m_ll_

Figure 4.16. Lineweaver-Burk plot of the data which
exhibited azide inhibition of immobilized
catalase. The azide concentrations ranged
from 0.0 (A) to 1.5 mM (B).

 

137

presence of catalase. Thus, for the small amount of
catalase impurity normally found in commercial glucose
oxidase preparations, the loss of peroxide from the
catalase reaction in the presence of peroxidase should
be negligible.

It should be noted that azide did not have an imme-
diate effect on the immobilized catalase. The azide/buffer
solution had to be passed through the reactor for about
three minutes before its full effect was attained. To re-
move the inhibitory effect of azide, the reactor was
flushed with pure buffer for five minutes. This process,
however, only returned the catalase to 90% of its original
activity. Thus, azide inhibition of catalase appears to
be only weakly reversible, in agreement with a previous
study [142].

6. Effect of azide on immobilized glucose oxidase

A preparation of glucose oxidase containing a very
small amount of catalase was chosen for this study. A
50 cm glucose oxidase (GO) reactor (type B) was inserted
into the CF manifold, and glucose standards in the range
of 0.1 mM - 0.5 mM were used as the sample. Curve B of
Figure 4.17 shows the effect of azide on the GO reactor.
Azide solutions of more than 0.5 mM had a large, negative
effect on glucose oxidase activity. Dixon plots of these
data, Figure 4.18, show that the inhibition was purely non-

competitive and that the KI value was 18 mM.

138

1.2 5—

 

 

Calibration Slope (x103 AU M“)
O
O)

 

-II-

I
I
0 4 8 12

“JE

Azide Concentrohon (nnM)

Figure 4.17. The effect of azide on three glucose
oxidase reactors, type B (A), type C (0),
and type D (EJ), containing various
amounts of catalase (refer to Table 4.3).

 

139

 

 

16n—
-I-- '1 3
m
I cu—
IE '6
L0 8“ '
o 9 2
x -
V
4 1. '95
m ' e 1
+1
O -11— ‘9:
m .
\\
r- O_.._.
—4 I I I I I I I I I
—20 —1O 0 10 20

Azide Concentration (mM)

Figure 4.18. Dixon plot of data from Figure 4.17,
curve B. Three glucose concentrations
were used: 0.44 mM (1), 0.22 mM (2),
and 0.11 mM (3).

140

Two other reactors were prepared from Type II
glucose oxidase spiked with small (type C) and large
(type D) amounts of catalase. The different concentration
ranges at which azide first inhibits catalase and then
glucose oxidase in these reactors can be seen in Figure
4.17. The optimal azide concentration increased as the
amount of catalase in the reactor increased. The rapid
rise in assay sensitivity observed as the concentration
of azide in the buffer approached 1 mM is due to catalase
inhibition. The more gradual decrease in assay sensitivity
at higher azide concentrations is due to glucose oxidase
inhibition.

We were surprised that the addition of a second
enzyme to the immobilization solution had only a small
effect on the bonding of the primary enzyme. At an azide
concentration of 15.2 mM, at which the action of catalase
is very small, the calibration slopes of all three reac—
tors (types B,C, and D) were similar (relative standard
deviation between reactors = 4%), which indicates that the
amount of active glucose oxidase bonded in all three
reactors was nearly the same.

7. Effect of azide on different preparations of glucose
oxidase

Reactor types E, F, and G (40 cm in length) repre—
sented immobilized forms of Sigma Types II, IX, and X

glucose oxidase respectively. The catalase activities of

 

141

of these preparations (Table 4.3) were very different,

and the differences can be seen if one compares the points
at zero azide concentration for the three curves in Figure
4.19. The calibration curves were prepared by sampling

peroxide standards in the range of 10 uM - 40 uM. Reac-

 

tor F exhibited very little activity which reflects the

higher catalase activity.

 

Addition of azide improved the sensitivity only a
small amount for reactor types G and E, while the charac-
teristics of reactor F changed noticeably. At an azide
concentration of 15.2 mM, the assay sensitivity with
reactor type F in place approached that of the other two
reactors. Again, the presence of a second enzyme had only
a small effect on the bonding of the primary enzyme.

8. A method to determine the presence of catalase

 

The properties of the same enzyme preparation can
vary from lot to lot. For example, some Sigma Type II
glucose oxidase purchased five years ago exhibited high
catalase activity [141], while a batch purchased several
months ago showed negligible catalase activity. Thus, it
is important to know the catalase activity in the glucose
oxidase preparations to be used.

To determine whether catalase activity in a given lot
of glucose oxidase is sufficient to cause problems in an
immobilized enzyme glucose assay, a simple test is suggested.

First, immobilize the glucose oxidase in a 50 cm or longer

142

 

 

 

 

 

 

 

A
IE 0 6’0
1.0 C G E
I) F
<( 1
d-
x 1-11—
v
CL
2 Hi-
(I)
0.4 ‘I'
C
O -U-
1;
S 0.2 ~—
.0
ES -
C) 0.0 I I I I I I I I
O 4 8 12 16

Figure 4.19.

AJide Concentrofion (nflw)

The effect of azide on three commercial
preparations of glucose oxidase, reactor
types 0, E, and F (refer to Table 4.3).

143

reactor, and then cut the tube into 10 and/or 20 cm
segments. Determine, with glucose as the sample, the
calibration slopes with various lengths of reactor in the
CF manifold. If the catalase content of the reactor is
significant, a curved plot of the calibration slope
versus reactor length will result. Figure 4.20 shows the
results of such an experiment. As the catalase activity
in the reactors increases (types B+D), the plot becomes
more hyperbolic. To verify the presence of catalase,
azide is added to the system so that the concentration
inside the enzyme reactor is about 1 mM. If straight lines
are produced, as shown in Figure 4.21, then catalase inter-
ference is confirmed. In this way, one can determine if
the use of sodium azide is warranted.
9. Conclusions

The results presented here should be applicable to
systems that use chromogenic agents other than the Trinder
reagent. The analytical and color-forming reactions, how-
ever, must be separated in time for the effect of azide to
be evident. Azide does not affect the accuracy of control
serum assays. Table 4.4 shows that no significant differ—
ence was found between the results obtained with and with-
out added azide. The sensitivity of the glucose determina-
tion is improved by the addition of sodium azide when
glucose oxidase immobilized enzyme reactors, containing

appreciable amounts of catalase, are used.

 

144

1.0 *-

 

 

 

l
22
:3 0.8 -— a 8
<1
[(7
O
; CL6-—
v
e C
a
0. O
O —I—
5; 0.4 . 9
c
o o
+) —n-—
L.
.o e '
B
0 0.0 I I I I I I
O 20 40 60

Reactor Length (cnfl

Figure 4.20. The change in calibration slope with
different lengths of reactors B, C, and
D (refer to Table 4.3). No azide was added.

145

1.0 fl—

O.8 ~—

 

Calibration Slope (x103 AU M“)

0.0 {I I I I

O 20 4O 60

«I-
«L-
-

Reactor Length (cm)

Figure 4.21. The change in calibration slope with
different lengths of reactors B, C, and
D (refer to Table 4.3). The reactions
occurred in the presence of 0.8 mM azide.

146

Table 4.4. Glucose Concentrations Found in Control Sera
With and Without Added Azide

Glucose Concentrations (mg/dL)

 

 

 

 

Found Found
Control Serum Expected (no azide) (2 mM azide)
Calibrate 165 143 153
Monitrol 95 91 95
Sigma 92 74 75

The 95% confidence limits for the experimentally determined
values are i 5 mg/dL.

CHAPTER 5

Enzyme Immobilization on Non—porous Glass

As shown in previous chapters, nylon serves well
as a carrier for immobilized enzymes. Yet, nylon is far
from perfect in many cases. Aromatic compounds tend to
adsorb to the nylon, and nylon has a small surface area-to-
volume ratio. Also, a charge develops an the surface during
immobilization; this charge can alter the enzyme properties.
Since these problems are not easily solved, the use of other
carriers has been an active area of research.

Glass offers the advantages of being inert, sturdy,
and readily available in a variety of forms; glass shows
good wash characteristics in continuous flow manifolds.
0n the other hand, non-porous glass has a small surface
area-to—volume ratio and has no reactive groups for direct
enzyme attachment. These problems can be overcome, however,
by growing “whiskers” on the glass to increase the surface
area and reacting the glass with a silynating reagent to
activate the surface. Coiled glass tubing for use in
continuous flow instruments is somewhat expensive, but the
use of straight tubing in the stopped-flow instrument
presented in Chapter 3 or the use of beads packed into a
tube is ideal.

The preparation and use of immobilized enzyme glass
reactors are discussed in this chapter. Glucose oxidase was

147

148

again chosen as the enzyme; the reactions were discussed
in Chapter 4. First, glass tubing was used in a seg—
mented CF system. The results are presented in Section A.
Then glucose oxidase was attached to non-porous glass
beads for use in a flow-injection instrument. Section B

includes analytical results from this novel system.

A. Immobilization of Glucose Oxidase on Glass Tubing

Borosilicate glass tubing of 1.25 mm i.d. and 30 cm
in length was bent in a U-shape for use in the segmented
continuous flow instrument. The glucose oxidase was attached
to the glass by procedures adapted from those described by
Iob and Mottola [147]. Four major steps are involved:
tube cleaning, ”whisker” formation, activation, and enzyme
bonding. The details are given below.

The glass was cleaned first by filling the tube with
concentrated HCl and heating it to 60 °C for 24 hours.
Next, the tube was washed with 100 mL each of distilled
water and acetone. The glass was dried by passing nitrogen
gas over the surface.

In the second step, the glass was etched and silica
needles were formed on the inner surface of the tube. The
procedure was taken from that reported by Onuska, §£.al.
[148]. The tube was filled with a 5% w/v (saturated)
methanol solution of ammonium bifluoride, NH F-HF for about

4

one hour at room temperature. The excess solution was

149

removed by passing a stream (1 mL/min) of nitrogen gas
through the tube. The tube became milky white in colora—
tion as the methanol was removed. A constant, but slow,
stream of gas was crucial for uniform coverage of the tube
with the reagent. This was the most difficult portion of
the treatment. The ends of the coated tube were then
sealed, and the tube was placed in a muffle furnace at
425 °C for three hours. Under these conditions, the bi-
fluoride is decomposed to ammonia and hydrogen fluoride gas.
The HF removes glass at one point and then deposits silica
in the form of ”whiskers” at other points on the tube.
The surface area of the reactor should be increased greatly.
After cooling, the tube was opened and flushed with
nitrogen gas. Activation was effected by reaction with
3-aminopropyltriethoxysilane (Sigma). The original pro-
cedure was reported for activation of porous glass beads
[149,150]. A neat solution was reacted at 60 °C for 1.5
hours or a 5% v/v solution in acetone was reacted at 60 °C
for 24 hours. Both procedures were hampered by the fact
that the solution slowly evaporated, leaving behind a
yellow gum, which often restricted solution flow through
the tube. The formation of the residue needs to be reme—
died before this procedure can be used reliably for the
activation of glass tubing. Perhaps lower reaction tem-
peratures and times will have to be used.

The activated tube was then washed with acetone and

150

water and filled with a 5% v/v solution of glutaraldehyde
in pH 8.0 phosphate buffer. The reaction occurred for
1.5 hours at room temperature. No change in the colora-

tion of the tube was evident. The glucose oxidase, 10 mg/mL

 

in pH 6.85 phosphate buffer, was next added to the tube and I
allowed to react for 18 hours at 4 °C. After these pro-

cedures were performed, the tube was washed with distilled,

 

deionized water and filled with the phosphate buffer when
not in use.

The results of the experiment were somewhat encoura-
ging. The activity of the reactor was tested as described
in Chapter 4. The best glass reactors had activities that
were about 25% of those prepared from nylon tubing. The

wash characteristics were somewhat better. Considering the

 

small amount of time spent on this venture, the results

were very good. Further testing is warranted.

B. Glucose Oxidase Bead String Reactor

In flow—injection (FI) analysis, sample dispersion and
mixing between the reagents and sample are coupled. In
order for complete mixing to occur, the sample dispersion
must be fairly large. Consequently, sample throughput
and sensitivity are sacrificed as longer residence times
are employed. In 1981, van der Linden, et.al. [151,152]
suggested the use of tubing packed with impervious glass

beads, which have a diameter that is 60—80% of the inner

151

diameter of the tube; the single bead string reactor (SBSR)
was born. The liquid must travel a tortuous path through
the reactor. Mixing is enhanced, while dispersion is

about ten times smaller than in open tubing of the same
dimensions. Wide use of such packed tubes is expected in
future FI research.

Since tiny, porous glass beads had been used as
carriers for enzymes in the past, it was expected that
larger, non-porous glass beads would also prove to be
good carriers. Once immobilized on the beads, the enzyme
could be packed into a tube and used in a flow-injection
instrument. The advantages of immobilized enzymes and
SBSRs would be combined to yield a powerful technique.

Jim Litch tested this method during an undergraduate re-
search project.

The immobilization procedure is listed in Table 5.1
and follows the same concepts as described for glass tubing.
The treatment was easier to perform, since most of the
reactions occurred in a beaker, and the beads were easier
to handle than the tubing. The beads acquired a red color
due to the reaction between the glutaraldehyde and the
amine group on the glass; this is a sign of a successful
activation step. A diagram of the completed reactor is
shown in Figure 5.1. A large overall reaction rate is
promoted by the rapid mixing and the intimate contact

between the enzyme and the bulk solution.

Table 5.1. Immobilization Procedure for Glass Beads

Step

1. Cleaning

2. Whisker
formation

3. Silanate

4. Enzyme

152

Reagent

Chromic Acid

6 M HCl
Distilled water
Acetone

Dry with N2

5% w/v solution
of NHuF-HF

in methanol

Dry with N2
Apply heat

2% acetone
solution of
3-aminopropyl-
triethoxysilane

-Decant excess

Apply heat

1% glutaraldehyde
in pH 8.0 buffer

5 mg/mL solution
of glucose oxidase
in pH 6.85 buffer

Time

15 min.
10 min.
10 min.
10 min.

1 hour

3 hours

30 min.

24 hours

30 min.

24 hours

Temp. (°C)

 

23
23
23
23

23

450

23

90

23

153

 

 

 

 

0.8 mm i.d. 0.5 mm diam.
Teflon Enzyme
Tubwng Coated Bead

Figure 5.1. View of the immobilized enzyme single
bead string reactor.

 

154

The continuous flow instrument described in Chapter 4
was modified to include a different manifold and to omit
the bubble—gate. The manifold that was used is shown in
Figure 5.2. A four-way slider valve (Altex), pneumatically
actuated, and controlled by a ZX81 microcomputer (Sinclair)
interfaced with a relay board (Byte-Back), was at the heart
of the system. In position 1, the carrier stream (Trinder/
peroxidase reagent) was pumped through the valve and SBSR.
In position 2, the sample (glucose) was pumped into the
system. The volume of glucose injected was determined by

the pump speed and the time the valve spent in position 2.

Sample volume = (Pump speed)(Valve time) (5.1)

Each experiment included one cycle of the valve.

The valve times were controlled by ZX81 software so
that the sample volume could be varied reproducibly over a
wide range. In such systems, mixing only occurs at the
leading edge of the sample zone until the valve is returned
to position 1. Thus, mixing between reagents and sample
in this design may differ from more common FI instruments
which employ a 6-way valve and a sample 100p.

The analytical and color-forming reactions occurred
simultaneously inside the reactor. Results indicated that
the indicator reaction was much faster than the glucose
oxidase reaction. This is necessary for a useful deter—

mination. All the reagents were prepared as described in

155

Glucose 4

 

 

 

 

 

 

  

 

1,Waste IIME- Waste
\ SBS fl
THnder/PO

 

 

 

 

 

 

 

Colorhneter
ZX81 510 nm

 

 

 

 

Figure 5.2. Schematic diagram of the flow-injection
instrument.

 

 

156

Chapter 4. The reactor was 21 cm in length and had an
internal diameter of 0.8 mm. If the beads take up 30%
of the total reactor volume, the void volume of the
reactor was about 75 uL.

The sample volume was varied at four different flow
rates. When large sample volumes were used, two peaks
often resulted from a single sample injection. The reason
for this was incomplete mixing; the reaction occurred at
both ends of the sample slug, but the reagents never
reached the center of the slug. The two peaks varied as
to which was the higher, but in all cases the heights
were within 25% of each other. Large volumes were de-
sirable because increased instrumental output resulted.
However, if the sample volume was too large double peaks
occurred. The optimal sample volume for the system was
found by plotting the volume between the double peaks
versus the sample volume. The x-intercept, the point
where only a single peak is produced, should be the op-
timal volume. Figure 5.3 shows the interesting results.
The curves for the four pump speed settings all follow
nearly the same line. This is evidence that mixing in a
SBSR is independent of flow rate and residence time.

The Optimal volume was about 45 uL. In practice, however,
sample volumes up to 65 uL did not give perceptible double
peaks.

Four glucose standards were each sampled for five

157

 

 

,\ 600 --

..l

3

m a.

x

O

0)

a 400 ~—

C

Q)

m 0

3 __

Q)

a)

g 200 -~

3

B

>' _-

4 I l | L l I
O T l I l I I I
0 200 400 600
Soaufle Vothe (0L)
Figure 5.3. The degree of incomplete mixing at various

sample volumes. Four pump speeds were used:
8.3, 12.3, 25.0, and 33.3 uL/sec.

 

 

158

seconds at a flow rate of 12.3 uL/sec. The residence time
inthe enzyme reactor was about 6 seconds. The instrumental
output is shown in Figure 5.4. The average relative
standard deviation of the results was less than 2%. A
calibration curve is shown in Figure 5.5. The results

were very good; the correlation coefficient was 0.99994.

The use of this method for the determination of glucose
looks promising. With further optimization, the sample
throughput should be increased from the current value of

60 samples/hour. Many other substrates could be determined

in this way, using other enzymes.

159

0.3 -T'

Absorbance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fi-

Ill

Thne UnM)

Figure 5.4. Flow-injection instrumental response.
The four glucose concentrations were:
0.44 (A), 0.88 (B), 1.33 (C), and 2.22 (D) mM.

160

Mean Absorbance

 

 

 

[Glucose] OnM)

Figure 5.5. Calibration curve for the flow—injection
determination of glucose.

CHAPTER 6

Future Plans

The use of open—tubular immobilized enzyme reactors
in flowing systems continues to grow, and more research
is needed to characterize such systems. A few of my ideas
for future research in this area are presented below. I
plan to continue some portions of my doctoral research,
and, hopefully, others at Michigan State will spend their

time working with immobilized enzymes.

1) Fast immobilization procedure for nylon. The current
procedure for bonding enzymes to nylon takes about 1.5 days.
The duration of the procedure might be shortened without
adversely affecting the enzyme atachment. A three-hour or
less immobilization treatment would be advantageous for
undergraduate research and teaching laboratory experiments.
It is known that most of the enzyme bonds to the activiated
nylon within the first ten minutes of reaction. Thus, the

reduction of the procedure in terms of time is feasible.

2) Properties of bi-enzyme systems. Many determinations
involve two or more enzyme-catalyzed reactions. It would
be helpful to know how each enzyme affects the attachment
of the other. The studies presented in Chapter 4, Section
E, suggested that the secondary enzyme, catalase, did not

significantly disturb the bonding of the primary enzyme,

161

162

glucose oxidase. The properties of bi-enzyme systems

should be studied in more detail.

3) Properties of immobilized enzymes in continuous flow
systems. Immobilized glucose oxidase reactors have been
fully characterized for use in a continuous flow instru-
ment. Other enzymes could be incorporated into the CF

manifold, and their properties could be studied. After

several enzyme systems have been tested in this fashion,
general conclusions about the characteristics of OTIMERs

in continuous flow systems can be drawn.

4) Fundamental studies of the influence of molecular
diffusion on immobilized enzyme kinetics. The stopped—
flow instrument described in Chapter 3 may prove to be

a very powerful tool in probing the interaction of
diffusion and immobilized enzyme kinetics. By changing
the reactor parameters and measuring the reaction rates,
a mathematical description of such systems should be
possible to derive. Subsequently, the inherent enzyme
properties could be calculated from the measured apparent

properties. This project deserves a lot of attention.

5) Immobilized enzyme bead string reactor. Very fast and
sensitive clinical determinations may be possible with the
use of enzyme-coated glass beads packed into a Teflon tube.

Flow-injection determinations using such reactors show

163

great promise and can be expanded for use with highly
active enzymes, other than glucose oxidase. The experi-

mental design would be similar to that described in

Chapter 5.

APPENDICES

164

APPENDIX A

Derivation of Diffusion-control Rate Equation

Model: 1) Circular tube of radius, r

2) Substrate reacts at the reactor wall
very fast to form product.

3) A radial concentration gradient in
substrate and product is established.

4) Substrate diffuses toward reactor wall.

5) Product diffusing away from reactor wall.

Initial condition:

Boundary condition:

Diffusion equation:

Final result:

[S] = [S] for 0<x<r at t=0,
where x ig the distance from the
center of the reactor.

[S] = 0 for x=r at t>0.

3[S]/3t = D(32[S]/3x2 + (3[S]/8x)/x)

3
[P] = [S]b(1-4Z (eXp(-Dtj:/r2))/j:)
S=1

165

APPENDIX B

Derivation of the Absorbance of a Solution with a Radial
Concentration Gradient

Model: 1) Reaction solution is axially uniform.
2) The concentration gradient is radially symmetric.

3) The concentration gradient does not change
with time (steady-state conditions).

4) The gradient is linear (d[P]/dx = constant).
5) The product is the absorbing species.

6) Light traverses the reactor parallel to the axis.

Considering a slice of solution from the center of
the tube to the reactor wall:

Absorbance = eL[P]

o r

= - — +

[P] (I ([P]w ([P]W [P]b)x/o)dx é [P]bdx)/r

where x is the distance from the center of

the reactor, 0 is the diffusion layer thickness,

and where r is the reactor radius.

Integrating the above equation and combing terms:

[P] = o/2r([P]w-[P]b) + [P]b

Thus, Absorbance = eL([P]b + o/2r([P]w-[P]b))

Case 1: No enzyme reaction (common absorbance measurement)
[P]w=[P]b and Absorbance = eL[P]b
Case 2: Start of enzyme reaction

[P]b = 0 and Absorbance = eL[P]w(o/2r)

L I ST OF REFERENCES

166

LIST OF REFERENCES

1. W.B. Furman, editor, Continuous Flow Analysis: Theory
and Practice, Marcel Dekker, NY (1976).

 

 

2. L.R. Snyder, J. Levine, R. Stay, and A. Conetta,
Anal. Chem. 48, 942A (1976).

3. L.R. Snyder, J. Chromatog. 125, 287 (1976).
4. J. Ruzicka and E.H. Hansen, Chemtech 9, 756 (1979).

5. J. Ruzicka and E.H. Hansen, Flow Injection Analysis,
John Wiley and Sons, NY (1981).

 

 

6. D. Betteridge, Anal. Chem. 59, 832A (1978).
7. B. Chance, J. Franklin Inst. 229, 455 (1940).

8. S.R. Crouch, F.J. Holler, P.K. Notz, and P.M. Beck-
with, Appl. Spec. Rev. 13, 165 (1977).

9. E.T. Gray and H.J. Workman, J. Chem. Ed. 51, 752
(1980).

10. P.W. Carr and L.D. Bowers, Immobilized Enzymes in
Analytical and Clinical Chemistry, John Wiley and Sons, NY
(1980).

 

 

 

11. M.D. Trevan, Immobilized Enzymes, John Wiley and Sons,
NY (1980).

 

12. I. Chibata, editor, Immobilized Enzymes, Research and
Development, John Wiley and Sons, NY (1980).

 

 

13. T.M.S. Chang, Immobilized Enzymes and Proteins,
Volumes 1 and 2, Plenum Press, NY (1977)

 

14. L.B. Wingard, E. Katchalski-Katzir, and L. Goldstein,
editors, Applied Biochemistry and Bioengineering, Volume 1,
Academic Press, NY (1976).

 

15. H.H. Weetall, editor, Immobilized Enzymes, Antigens,
Antibodies, and Peptides, Marcel Dekker, NY (1975).

 

 

 

 

167

16. R.A. Messing, editor, Immobilized Enzymes for Indus-
trial Reactors, Academic Press, NY (1975).

 

 

17. M. Salmona, C. Saronio, and S. Garattini, Insolu-
bilized Enzymes, Raven Press, NY (1974).

 

18. O. Zaborsky, Immobilized Enzymes, CRC Press, Cleveland
(1973).

 

19. S. Aiba, A.E. Humphrey, and N.F. Millis, Biochemical
Engineering, Academic Press, NY (1973), pp 393-416.

 

 

20. L.B. Wingard, Enzyme Engineering, John Wiley and Sons,
NY (1972). Pp 177-400.

 

21. L.D. Bowers, Trends Anal. Chem. 1, 191 (1982).

22. L.D. Bowers and P.W. Carr, Adv. Biochem. Eng. 15,
89 (1980).

23. L.D. Bowers and P.W. Carr, Anal. Chem. 45, 544A
(1976).

24. H.H. Weetall, Anal. Chem. 45,602A (1974).
25. L.B. Wingard, Adv. Biochem. Eng. g, 1 (1972).
26. G.G. Birch, N. Blakebrough, and K.J. Parker, Enzymes

and Food Processing, Applied Science Publishers, London
(1981).

 

27. L.B. Wingard, E. Katchalski—Katzir, and L. Goldstein,
Applied Biochemistry and Bioengineering, Volume 2, Academic
Press, NY (1979).

 

28. S. Kindel, Technology 1, 62 (1981).
29. K.J. Skinner, Chem. Eng. News, August 18,1975, 22.
30. C.R. Martin, Trends Anal. Chem. 1, 175 (1982).

31. G.G. Guilbault and M.H. Sadar, Acc. Chem. Res. 15,
344 (1979).

32. G.G. Guilbault and M.H. Sadar, Proc. Analyt. Div.
Chem. Soc., October 1977, 302.

33. W.J. Blaedel and R.C. Engstrom, Anal. Chem. 55,
1691 (1980).

168

34. F. Scheller and D. Pfeiffer, Anal. Chim. Acta 117,
383 (1980).

35. W.J. Blaedel and J. Wang, Anal. Chem. 55, 1426
(1980).

36. R.M. Ianniello and A.M. Yacynych, Anal. Chim. Acta
131, 123 (1981).

37. R. Renneberg, D. Pfeiffer, F. Scheller, and M.
Janchen, Anal. Chim. Acta 134, 359 (1982).

38. E.L. Gulberg and G.D. Christian, Anal. Chim. Acta
123, 125 (1981).

39. W. Hinsch and P.V. Sundaram, Fres. Z. Anal. Chem.
309, 25 (1981).

40. M. Mascini and G. Palleschi, Anal. Chim. Acta 136,
69 (1982).

41. W. Hinsch, W.—D. Edersbach, and P.V. Sundaram,
Clin. Chim. Acta 104, 95 (1980).

42. W. Hinsch and P.V. Sundaram, Clin. Chim. Acta
104, 87 (1980).

43. B. Danielsson, B. Mattiasson, and K. Mosbach, Pure
Appl. Chem. 54, 1443 (1979).

44. C. Spink, CRC Critical Rev. Anal. Chem. 5, 1 (1980).

45. R.S. Schifreen, D.A. Hanna, L.D. Bowers, and P.W.
Carr, Anal. Chem. 45, 1929 (1977).

46. A.O. Lindberg, Anal. Chim. Acta 131, 133 (1981).
47. A.S. Attiyat and G.D. Christian 105, 154 (1980).

48. C.-H. Kiang, S.S. Kuan, and G.G. Guilbault, Anal.
Chem. 55, 1319 (1978).

49. J.-C. Kaun, S.S. Kuan, and G.G. Guilbault, Anal.
Chim. Acta 100, 229 (1978).

50. P.V. Sundaram and S. Jarayaman, Clin. Chim. Acta
54, 309 (1979).

51. O.R. Zaborsky and J. Ogletree, Biochem. Biophys.
Res. Comm. 5;, 210 (1974).

169

52. D.J. Inmann and W.E. Hornby, Biochem. J. 129, 255
(1972).

53. D.L. Morris, J. Campbell, and W.E. Hornby, Biochem.
J. 147, 539 (1975).

54. P.V. Sundaram, Biochem. J. 183, 445 (1979).

55. P.V. Sundaram and D.K. Apps, Biochem. J. 161, 441
(1977).

56. P.V. Sundaram and M.P. Igloi, Clin. Chim. Acta 54,
295 (1979).

57. F.N. Onyezili and A.C. Onitiri, Anal. Biochem. 113,
203 (1981).

58. G.A. Noy, A. Buckle, and K. Alberti, Clin. Chim.
Acta 55, 135 (1978).

59. F.N. Onyezili and A.C. Onitiri, Anal. Biochem. 117,
121 (1981).

60. J. Campbell, W.E. Hornby, and D.L. Morris, Biochim.
Biophys. Acta 384, 307 (1975).

61. V. Toome, S. DeBernardo, K. Manhart, and M. Weigele,
Anal. Lett. 1, 437 (1974).

62. W.—T. Law and S.R. Crouch, Anal. Lett. 45, 1115
(1980).

63. O. Warburg and W. Christian, Biochem. Z. 310, 384
(1941).

64. P.D. Boyer, editor, The Enzymes, Academic Press, NY.

 

65. M. Dixon and E.C. Webb, Enzymes, Academic Press, NY
(1964).

66. J.B. Sumner and G.F. Somers, Chemistyy and Methods
55 Enzymes, Academic Press, NY (1953).

67. K.J. Laidler and P.S. Bunting, The Chemical Kinetics
5: Enzyme Action, Clarendon Press, Oxford, England (1973),
p 398.

 

 

68. E.W. Wharton, E.M. Crook, and K. Brocklehurst, Eur.
J. Biochem. 5, 572 (1968).

170

69. W.E. Hornby, M.D. Lilly, and E.M. Crook, Biochem.
J. 55, 420 (1966).

70. J.-M. Engasser and C. Horvath, Applied Biochemistry
and Bioengineering, Volume 1, Academic Press, NY (1976),
pp 146-149.

 

 

71. G.P. Royer, Immobilized Enzymesy Antigens, Antibodies,

 

 

and Peptides, Marcel Dekker, NY (1975), pp 66-68.

 

72. J.-M. Engasser and P.Hisland, Biochem. J. 173, 314
(1978).

73. J.-M. Engasser, P.R. Coulet, and D.C. Gautheron,
J. Biol. Chem. 252, 7919 (1977).

74. I.C. Cho and H. Swaisgood, Biochim. Biophys. Acta
334, 243 (1974).

75. P. Nitisewojo and H.O. Hultin, Eur. J. Biochem.
51, 87 (1976).

76. N.L. Smith and H.M. Lenhoff, Anal. Biochem. 54,
392 (1974).

77. K.J. Laidler and P.S. Bunting, Meth. Enzymol. 54,
227 (1980).

78. T. Kobayashi and K.J. Laidler, Biotech. Bioeng.
45, 99 (1974).

79. N.J. Daka and K.J. Laidler, Can. J. Biochem. 55,
774 (1978).

80. N.J. Daka and K.J. Laidler, Biochim. Biophys. Acta
612, 305 (1980).

81. Advances 45 Automated Analysis, Technicon International
Congress 1976, Volume 1, Mediad Inc., Tarrytown, NY (1977).

 

82. L.R. Snyder and H.J. Adler, Anal. Chem. 45, 1017
(1976).

83. L.R. Snyder and H.J. Adler, Anal. Chem. 45, 1022
(1976).

84. S. Morgenstern, Advances 45 Automated Analysis,
Volume 1, Therman Assoc., Miami (1971), p 85.

 

171

85. M.D. Joseph, D.J. Kasprazak, and S.R. Crouch, Clin.
Chem. 55, 1033 (1977).

86. D.R. Christmann, S.R. Crouch, and A. Timnick, Anal.
Chem. 55, 276 (1981).

87. P.D. Boyer, editor, The Enzymes, Volume XI, Academic

 

Press, NY (1975), pp 289- 292.

88. J.R. Florini and C.S. Vestling, Biochim. Biophys.
Acta 251 575 (1957).

89. H.L. Pardue, Advances 55 Analytical Chemistry and
Instrumentation, Volume 7, John Wiley and Sons, NY (1968),
pp 141-207.

 

 

90. P. Trinder, Ann. Clin. Biochem. 5, 24 (1969).
91. D. Barham and P. Trinder, Analyst 55, 142 (1972).

92. N. W. Tietz, Fundamentals of Clinical Chemistry, W. B.
Saunders, Philadelphia (1976),— p 1213.

 

 

93. R.H. White-Stevens, Clin. Chem. 55, 578 (1982).

94. J.A. Lott and K. Turner, Clin. Chem. ‘55, 1754
(1975).

95. N. W. Tietz, editor, Fundamentals of Clinical Chemistry,
W.B. Saunders, Philadelphia (1976), pp 234- 263.

 

 

96. F. Gram, J.A. Hveding, and A. Reine, Acta Chem.
Scand. 55, 616 (1973).

97. E.N. Drake and C.E. Brown, J. Chem. Ed. 54, 124
(1977).

98. J. Okuda and I. Miwa, Anal. Biochem. 55, 387 (1971).

99. A.S. Kaston and R. Brandt, Anal. Biochem. 5, 461
(1963).

100. P.V. Sundaram, B. Blumemberg, and W. Hinsch, Clin.
Chem. 55 1436 (1979).

101. L.P. Leon, M. Sansur, L.R. Snyder, and C. Horvath,
Clin. Chem. 55, 1556 (1977).

102. L. Gorton and K.A. Bhatti, Anal. Chim. Acta 105,
43 (1979).

172

103. Sigma Chemical Company Catalog, 1982 edition.

104. D.L. Williams, A.R. Doig, and A. Korosi, Anal.
Chem. 45, 118 (1970).

105. M. Koyama, Y. Sato, M. Aizawa, and S. Suzuki,
Anal. Chim. Acta 116, 307 (1980).

106. T. Osaki, Biotech. Bioeng. 55, 1287 (1980).

107. G.J. Lubrano and G.G. Guilbault, Anal. Chim. Acta
55, 229 (1978).

108. B. Watson, D.N. Stifel, and F.E. Semersky, Anal.
Chim. Acta 106, 233 (1979).

109. G. Sittampalam and G.S. Wilson, J. Chem. Ed. 55,
70 (1982).

110. D.C. Williams, G.F. Huff, and W.R. Seitz, Clin.
Chem. gg, 372 (1976).

111. J.N. Miller, B.F. Rocks, and D.T. Burns, Anal.
Chim. Acta 55, 353 (1977).

112. L.B. McGowan, B.J. Kirst, and G.L. LaRowe, Anal.
Chim. Acta 117, 363 (1980). -

113. F. Heinz and T.W. Beuhausen, J. Clin. Chem.
Clin. Biochem. ‘55, 977 (1981).

114. R. Chirillo, G. Caenaro, B. Pavan, and A. Pin,
Clin. Chem. 55, 1744 (1979).

115. J. Ziegenhorn, U. Neumann, A. Hagen, W. Bablok,
and K. Stinshoff, J. Clin. Chem. Clin. Biochem. 45,
13 (1977).

116. American Research Products Company, bulletin.

117. L.G. Morin and J. Prox, Clin. Chem. 55, 959
(1973).

118. R.H. White-Stevens and L.R. Stover, Clin. Chem.
55, 589 (1982).

119. N. Gochman and J.M. Schmitz, Clin. Chem. 45,
943 (1972).

173

120. L.P. Leon, S. Narayanan, R. Dellenbach, and C.
Horvath, Clin. Chem. 55, 1017 (1976).

121. A.T. Romano, Clin. Chem. 45, 1152 (1973).

122. C. Matsubura, K. Ishii, and K. Takamura, Microchem.
J. 55, 242 (1981).

123. K. Tamaoku, Y. Murao, K. Akiura, and Y. Ohkura,
Anal. Chim. Acta 136, 121 (1982).

124. I. Seigo, Eisei Kensa 55, 760 (1979).

125. J.D. Artiss, R.J. Thibert, J.M. McIntosh, and B.
Zak, Microchem. J. 55, 487 (1981).

126. D. Keilin and E.F. Hartree, Biochem. 55, 331 (1952).
127. D. Keilin and E.F. Hartree, Biochem. 45, 221 (1948).
128. R. Bentley, Meth. Enzymol. l, 340 (1955).

129. V. Linek, P. Benes, J. Sinkule, O. Holecek, and
V. Maly, Biotech. Bioeng. ‘55, 2515 (1980).

130. S. Gondo, T. Osaki, and M. Morishita, J. Mol.
Cat. 45, 365 (1981).

131. P.F. Greenfield, J.R. Kittrell, and R.L. Laurence,
Anal. Biochem. 55, 109 (1975).

132. B.C. Saunders, A.G. Holmes-Siedle, and B.P. Stark,
Peroxidase, Butterworths, Washington, D.C. (1964).

 

133. A.C. Maehly, Meth. Enzymol. 5, 801 (1955).

134. I.M. Kolthoff, E.B. Sandell, E.J. Meehan, and S.
Bruchenstein, Quantitative Chemical Analysis, The MacMillan
Co., NY (1969), p 854.

 

135. C.J. Patton, M. Rabb, and S.R. Crouch, Anal. Chem.
54, 1113 (1982).

136. S. Klose, M. Sholtz, E. Munz, and R. Portenhauser,
Clin. Chem. 54, 250 (1978).

137. F. Zoppi and D. Fenili, Clin. Chem. 55, 1229
(1980).

174

138. N.W. Teitz, editor, Fundamentals 54 Clinical Chemistry,
W.B. Saunders, Philadelphia (1976) p 254.

 

 

139. Y. Malpiece, M. Sharan, J.-N. Barbotin, P. Personne,
and D. Thomas, J. Biol. Chem. 225, 6883 (1980).

140. C. Horvath, B.A. Solomon, and J.-M. Engasser, Ind.
Eng. Chem. Fundam. 45, 431 (1973).

141. R.Q. Thompson, C.J. Patton, and S.R. Crouch, Clin.
Chem. 55, 556 (1982).

142. G.R. Schonbaum and B. Chance, The Enzymes, Volume 13,
Academic Press, NY (1976), pp 363-408.

 

143. D. Keilin and P. Nicholls, Biochim. Biophys. Acta
55, 302 (1958).

144. P. Nicholls and G.R. Schonbaum, The Enzymes,
Volume 8, Academic Press, NY (1963), pp 147-225.

145. M. Dixon, Biochem. J. 55, 170 (1953).

146. I.H. Segel, Enzyme Kinetics, John Wiley and Sons, NY
(1975). PP 170-203.

 

147. A. lab and H.A. Motolla, Clin. Chem. 55, 195
(1981).

148. F.I. Onuska, M.E. Comba, T. Bistricki, and R.J.
Wilkinson, J. Chrom. 142, 117 (1977).

149. H.H. Weetall, Science 166, 616 (1969).

150. P.J. Robinson, P. Dunnill, and M.D. Lilly, Biochim.
Biophys. Acta 242, 659 (1971).

151. W.E. van der Linden, Trends Anal. Chem. 4, 188
(1982).

152. J.M. Reijn, W.E. van der Linden, and H. Poppe,
Anal. Chim. Acta 123, 229 (1981).