Anal.
Chem.
1994,66,
1754-1761
Characterization
of
Glucose Oxidase-Modified
Poly( phenylenediamine)-Coated Electrodes in
Vitro
and in
Vivo:
Homogeneous Interference by Ascorbic Acid in
Hydrogen Peroxide Detection
John
P.
Lowry,
Karl McAteer, Satea
S.
El
Atrash, Adrlenne
Duff,
and Robert
D.
O’Nelll’
Department
of
Chemistry, University College Dublin, Belfield, Dublin
4,
Ireland
The response of glucose oxidase (GOx) modified poly(&
phenylenediamine) coated Pt disk electrodes to glucose
was
well-behaved with
a
rapid response time and displaying
Michaelis-Menten kinetics. However, the glucose response
was
lowered in
a
concentration-dependent manner by ascorbic
acid when the glucose calibrations were carried out in solutions
containing this reducing agent. The possibility of
a
homo-
geneous redox reaction in which the H202 generated by the
enzymatic oxidation of glucose at the GOx/polymer surface
is consumed by ascorbate
was
investigated. Similar “negative”
interference at GOx-modified carbon powder electrodes not
involving membranes and for H202 calibrations at bare Pt
electrodes supported the hypothesis. The observation that this
interference could be blocked by the chelating agent EDTA
suggests that the homogeneous reaction is catalyzed by trace
metal ion impurities in solution.
A
model for the homogeneous
reaction based on these experimental findings is proposed and
tested by comparing quiescent and stirred solutions.
No
homogeneous interference by uric acid
was
observed. The
electrodes were found to be free from lipid fouling in vitro, and
experiments monitoring brain glucose levels in vivo indicate
the absence of the homogeneous reaction in this environment.
The results highlight the need to test each individual assay
procedure involving HzO2 under relevant conditions for both
positive and negative interference by ascorbic acid.
The search for the ideal glucose sensor continues to be one
of the main focuses of biosensor research despite
3
decades
of intense investigation since the development of the first
electrochemical glucose sensor by Clark and Lyons in 1962.’
This is primarily due to the important role of such a glucose
monitoring sensor in industrial2t3 and clinicalw applications,
ranging from the analysis of fermentation media, to the
development of the artificial @-cell or pancreas for the
treatment
of
the metabolic disease diabetes mellitus.’J’
Because of the complexity and serious specificity problems
associated with direct electrooxidation of glucose9J0 the
(1) Clark, L. C.,
Jr.;
Lyons, C.
Ann. N.Y. Acad. Sci.
1962, 102,
29-45.
(2) Grundig,
B.;
Krabisch, C.
Anal. Chim. Acta
1989, 222,
75-81.
(3) Senda, M.
Ann. N.Y. Acad. Sci.
1990,613,
79-94.
(4) Meyerhoff, M. E.
Clin.
Chem.
1990, 36,
1567-1572.
(5)
Arnold, M. A,; Meyerhoff, M. E.
CRC
Crir. Reu. Anal. Chem.
1988,
20,
149-196.
(6) Claremont,
D.
J.;
Pickup,
J.
C.
In
Biosensors: Fundamenralsand Applicationr;
Turner,
A. P. F., Karube,
I.,
Wilson,
G.
S.,
Eds.; Oxford University Press:
New York, 1987; pp 356-376.
(7)
Soeldner,
J.
S.;
Change, K. W.; Aisenberg,
S.;
Hiebert,
J.
M. In
Temporal
Aspecfsof Therapeutics;
Urquhart,
J.,
Yates, F. E., Eds.; Plenum Press: New
York, 1973; pp 181-207.
1754
AnalyticalChemistry,
Vol.
66,
No.
10, May
15,
1994
oxidoreductase enzyme glucose oxidase (GOx) is most often
used in the development of glucose sensors. This enzyme
catalyses the oxidation of glucose according to the reactions
D-glucono-&lactone
+
GOx/FADH, (I)
P-D-glucose
+
GOx/FAD
-*
GOx/FADH,
+
0,
-
GOx/FAD
+
H,O, (11)
where FAD is the oxidized form of the prosthetic group, flavin
adenine dinucleotide.
The development over the last 30 years of enzyme-based
amperometric devices for glucose determination can essentially
be divided into three categories. The first involves “classical”
devices which monitored either the consumption of oxygen’
or the formation of hydrogen peroxide.12 Such devices were
originally affected by the ambient concentration
of
oxygen in
the sample and required a large overpotential. In order to
avoid these problems “second generation” systems were
developed in which the natural dioxygen in reaction I1 is
replaced by a mediat0r3-13.1~ resulting in electrodes which are
relatively insensitive to changes in dioxygen tension and which,
depending on the choice of mediator, can be operated at lower
applied potentials. The successful application of these
mediated systems relies on the appropriate choice of mediator,
based on good kinetic interaction with the reduced enzyme
(GOx/FADH2), and the effective entrapment of the mediator
at the electrode surface. The latter has conventionally been
achieved by the use of precast films (e.g., dialysis mem-
branes13J5J6 and more recently using electropolymerization
techniques to entrap both the mediator and enzyme in a
molecular film.17-19 However, problems associated with these
devices include leeching of the mediator from the membrane
(8) Velho,
G.
D.;
Reach,
G.;
Thhenot,
D.
R.
In
Biosensors: Fundamentals
and
Applicafionr;Turner,
A. P. F.,Karube,
I.,
Wilson,G.S., Eds.;OxfordUniversity
Press: New York, 1987; pp 390-408.
(9) Vassilyev,
Yu.
B.;
Khazova,
0.
A.;
Nikolaeva, N. N.
J.
Electroanal. Chem.
Interfacial Electrochem.
1985, 196,
105-125.
(10) Vassilyev,
Yu.
8.;
Khazova,
0.
A.;
Nikolaeva, N. N.
J.
Electroanal. Chem.
Interfacial Electrochem.
1985. 196,
127-144.
(1 1) Updike,
S.
J.;
Hicks,
G.
P.
Nafure
(London)
1967, 214,
986-988.
(12) Guilbault,
G. G.
Analyfical Uses
of
Immobilised Enzymes;
Marcel Dekker:
New York, 1984.
(13) Cass. A. E. G.;Davis,G.; Francis,G.
D.;
Hill, H. A. O.;Aston, W.
J.;
Higgins,
I.
J.;
Plotkin,
E.
V.; Scott, L.
D.
L.;
Turner,
A. P. F.
Anal. Chem.
1984, 56,
667-671.
(14) Crumbliss, A. L.; Hill, H. A.
0.;
Page,
D.
J.
J.
Elecrroanol. Chem. Interfacial
Electrochem.
1986, 206,
327-331.
(15) Gunasingham, H.; Tan, C. H.
Electroanalysis
1989,
I,
423-429.
(16) Clark, L. C.,
Jr.;
Noyes, L. K.; Spokane, R.
B.;
Sudan, R.; Miller,
M.
L.
(17) Foulds, N. C.; Lowe, C. R.
Anal. Chem.
1988.60,
2473-2478.
Methods Enzymol.
1988, 137,
68-89.
0003-2700/94/0366-1754$04.50/0
0
1994
American Chemical Society
compartment of precast films and sensitivity to dioxygen
tension at physiological concentrations of glucose with elec-
tropolymerized films.20
In an attempt to eliminate mediators from the reaction
scheme “third generation” electrodes made from conducting
organic salts have been reported.21*22 The most common
example is the TTF+TCNQ- (tetrathiafulvalenium tetracy-
anoquinodimethanide) electrode where it has been suggested
that reduced GOx is oxidized directly at the electrodesurface.21
However, the mechanism of electron transfer is contr~versial~~
and there are also interferen~e~~ and problems. Thus,
notwithstanding this prolific research into mediated and
nonmediated enzyme electrodes, the peroxide detecting system
still remains the most common form of enzyme electrode.26
As we are interested in developing a glucose sensor for
neurochemical applications in the living brain, typical problems
associated with analysis of such a complex biological environ-
ment include fouling by electrode poisons (e.g., proteins) and
surfactants (e.g., lipids) and direct faradaic interference by
endogenous electroactive species. Two of the most common
interferences are the reducing agents ascorbic acid27 (AA)
and uric acid,28 both of which exist in anionic form at
physiological pH,29y30 and whose concentrations in body fluids
are continuously
hanging.^^.^^.^'
Recently, H202-detecting
sensors based on the electropolymerization of poly(pheny1ene-
diamine) (PPD) films on Pt electrodes in which the enzyme
is incorporated in the polymer film have been reported as
interference-free sensors having a high enzyme activity and
a low response time, making them suitable for detecting
substrates in biological fl~ids.~~-~~
We recently reported preliminary results suggesting that
the H202 formed by the catalytic oxidation of glucose at these
electrodes is depleted by an homogeneous redox reaction
involving AA in vitro.35 Because of the important implications
for the reliability of a wide variety of assays involving H202
detection, including clinical analyses where negative interfer-
(18)
Kajiya, Y.; Sugai,
H.;
Iwakura, C.; Yoneyama, H.
Anal. Chem.
1991, 63,
(19)
Hale,P.D.;Boguslavsky,L.I.;Inagaki,T.;Karan,H.I.;SuiLee,H.;Skotheim,
(20)
Cenas, N. K.; Kulys,
J. J.
J.
Electroanal. Chem. Interfacial Electrochem.
(21)
Albery,
W.
J.;
Bartlett, P. N.; Craston, D.
H.
J.
Electroanal. Chem. Interfacial
(22)
Bartlett, P. N. In
Biosensors: A Practical Approach;
Oxford University Press:
(23)
Bartlett, P. N.
J.
Electroanal. Chem. Interfacial Electrochem.
1991, 300,
(24)
Lowry,
J.
P.; O’Neill,
R.
D.
J.
Electroanal. Chem. Interfacial Electrochem.
(25)
Heller, A,; Degani, Y. In
Redox Chemistry and Interfacial Behavior
of
Biological Molecules;
Dryhurst, G., Niki,
K.,
Eds.; Plenum Press: New York,
(26)
Methods in Enzymology;
Mosbach, K., Ed.; Academic Press: New York,
(27)
Grunewald,
R.
A,; ONeill,
R.
D.; Fillenz, M.; Albery,
W.
J.
Neurochem. Int.
(28)
Cespuglio,
R.;
Sarda, N.; Gharib, A,; Faradji,
H.;
Chastrette, N.
Exp. Brain
(29) Youngblood, M. P.
J.
Am. Chem.
Soc.
1989,
111,
1843-1849.
(30)
Dryhurst, G.; Kadish, K. M.; Schellcr, F.; Renneberg,
R.
Biological
(31)
O’Neill,
R.
D.
Brain Res.
1990,
507,
267-272.
(32)
Sasso,
S.
V.;
Pierce,
R.
J.;
Walla,
R.;
Yacynych, A.
M.
Anal. Chem.
1990,
(33)
Malitesta, C.; Palmisano, F.; Torsi, L.; Zambonin,
P.
G.
Anal. Chem.
1990,
(34)
Lowry,
J.
P.; ONeill,
R.
D.
Electroanalysis,
in press.
(35)
Lowry,
J.
P.; ONeill,
R.
D.
Anal. Chem.
1992, 64,
453-456.
49-54.
T.
A.; Okamoto, Y.
Anal. Chem.
1991,63,
677-682.
1981,
128,
103-113.
Electrochem.
1985, 194,
223-235.
New York,
1990;
pp
47-95.
175-189.
1992, 334,
183-194.
1988;
pp
151-171.
1988;
Vol.
137.
1983,
5,
773-778.
Res.
1986,
64,
589-595.
Electrochemistry;
Academic Press: New York,
1982;
Vol.
1,
Chapter
5.
62,
1111-1117.
62,
2735-2740.
ence by AA in glucose monitoring has been reported recently,36
we have investigated this phenomenon further. The results
support the original hypothesis and highlight the need to test
each individual assay procedure involving Hz02 under relevant
conditions for both positive and negative interference by
AA.
Since we are developing enzyme-based sensors for monitoring
neurochemical dynamics, we have investigated the relevance
of AA interference at Pt/PPD/GOx electrodes in vivo.
EXPERIMENTAL SECTION
Chemicals and Solutions.
The enzyme glucose oxidase
(GOx) fromhpergillus niger (EC 1.1.3.4, type VII-S), insulin
from bovine pancreas, and the lipid L-a-phosphatidyletha-
nolamine (type 114) were obtained from Sigma Chemical
Co.
Ethylenediaminetetraacetic
acid (EDTA, disodium salt)
and hydrogen peroxide (30% (w/v) solution) were from BDH
Chemicals Ltd., GPR grade. The exact concentration of H202
was determined by titrating against a standardized potassium
permanganate solution. The o-phenylenediamine (1,2-di-
aminobenzene, Merck Chemical Co.,
99+%),
a-D-(+)-glucose
(Sigma), L-ascorbic acid (BDH, Biochemical grade), and uric
acid (Sigma, sodium salt), were used as supplied. Carbon
paste was prepared37 by thoroughly mixing 2.83 g of carbon
powder (UCP-1-M, Ultra Carbon Corp., Bay City, MI) and
1.0 mL of silicone oil (Aldrich, Catalog No. 17 563-3).
A stock 1 mol/L solution of glucose was prepared, left for
24 h at room temperature to allow equilibration of the anomers,
and then stored at 4 OC. Hydrogen peroxide solutions (100
mmol/L) were prepared just before use by dilution of the
commercial 30% stock solution. AA solutions (100 mmol/L)
for experiments in vitro were also prepared just before use
because of their gradual decomposition, while EDTA solutions
(100 mmol/L) were used within 48 h of preparation. AA
solutions (0.34 g/mL) used for experiments in vivo were
prepared just before use, and the pH was adjusted to between
6.5 and 7.0 using NaOH. The insulin solution was prepared
in PBS (4.0 mg/25 mL) and remained cloudy even after
sonication for ca. 15 min. Experiments in vitro were carried
out in a phosphate buffer saline (PBS) solution, pH 7.4: NaCl
(Merck, BP USP grade, 0.15 mol/L), NaH2P04 (BDH,
AnalaR grade, 0.04 mol/L), and NaOH (BDH, GPR grade,
0.04 mol/L). All solutions were prepared using distilled water
and were kept refrigerated when not in use.
Atomic absorption spectroscopy of the PBS indicated that
it contained ca. 1 pmol/L copper and 3 pmol/L iron impurities.
The level of these metals in the distilled water and AA and
glucose reagents was negligible.
Instrumentation and Software.
All
,
experiments were
microcomputer controlled with data collection accomplished
using either a Biodata Microlink interface or a National
Instruments (NI, Austin, TX) AT-MIO-16 data acquisition
board and a specially designed low-noise, low-damping
potentiostat. In-house software was written in QuickBASIC
(version 4.0) andNI Labwindows (version 2.1) QuickBASIC
environments to perform all voltammetric experiments and to
collect, plot, and analyze the data. The enzyme parameters
V,,,,
Km,
and
(Y
(see Data Analysis below) were obtained by
(36) Strijdom,
J.
G.; Marais, B.
J.;
Kocslag,
J.
H.S.
Afr. Med.
J.
1993.84.64-65.
(37)
O’Neill,
R.
D.; Grunewald,
R.
A.; Fillcnz, M.; Albery,
W.
J.
Neuroscience
1982,
7,
1945-1954.
Analytical
Chemistty,
Vol.
66,
No.
10,
May
15,
1994
1755
using the nonlinear regression data analysis computer package
Enzfitter (Elsevier-Biosoft, version 1.05).
Working Electrode Preparation.
Glucose oxidase was
immobilized in
poly(o-phenylenediamine)
films by potentio-
metric electropolymerization of the monomer
(300
mmol/L)
on the bare disk end of a freshly cut Teflon-insulated Pt wire
(Plastics One Inc., Virginia). The codeposition procedure
was similar to that reported by Malitesta et al.33 and Almeida
et Briefly, a deoxygenated solution of the monomer
(300
mmol/L) was prepared in PBS. An aliquot of
5
mL of this
solution was then added to a weighed quantity
of
GOx
(5
mg/mL) in a 10-mL beaker and the solution sonicated for
approximately 2 min to ensure complete dissolution of the
enzyme. The reference (SCE) and auxiliary (large Pt wire)
electrodes were then immediately connected and the elec-
tropolymerization procedure was begun. The working elec-
trode potential was maintained at
+0.65
V
versus SCE for 15
min, although 10 min was sufficient to reach the small steady-
state current indicative of this self-sealing process.32 After
the electropolymerization, the Pt/PPD/GOx electrodes were
rinsed by immersion in PBS for
30
min to remove loosely
bound enzyme and unreacted monomer.
The carbon paste working electrodes (CPEs) were made
from Teflon-coated silver wire3' (Clark Electromedical
Instruments, Reading). The Teflon insulation was slid along
the wire to create an approximately 2 mm deep cavity. Further
slipping of the Teflon was prevented by bending the wire at
the opposite end to the cavity. The cavity was packed with
carbon paste using a silver wire as plunger. A small gold
electrical contact was attached to the end of the wire. The
carbon powder electrode (CPWEs) were prepared in the same
way except that the Teflon cavity was filled with the unpasted
carbon powder. Glucose oxidase adsorption onto the CPWEs
was accomplished by dipping the electrode tip in the enzyme
powder until the surface was completely coated and then
dropping a sufficient volume of PBS (0.1-0.2 FL) onto the
electrode surface to apparently dissolve the enzyme. The active
diameter of the disk electrodes was 250 pm with a
320-pm
external diameter. When not in use all electrodes were stored
in PBS at
4
"C.
Voltammetry in Vitro.
All experiments in vitro were
performed in a standard three-electrode glass electrochemical
cell which was thermostated at 25.0
f
0.5
OC. The working
electrodes used were either bare Pt disks, PPDIGOx-modified
Pt, or GOx-modified carbon powder electrodes. A saturated
calomel electrode (SCE) was used as the reference electrode,
and a large silver wire, isolated in a compartment containing
PBS, served as the auxiliary electrode. The applied potentials
for these amperometric studies were
+0.7
V
versus SCE for
the Pt/PPD/GOx electrodes, which is the value generally
used for H202 dete~tion~~,~~J* and
+0.78
V
versus SCE for
the CPWEs. The current was measured when the solution
was quiescent unless otherwise stated. To facilitate mixing
in calibration experiments, solutions were bubbled with air
using a RENA 101 air pump for ca. 10
s
following the addition
of each aliquot; the current was then measured when the
solution was quiescent. In hydrodynamic experiments solu-
tions were stirred at ca.
5
Hz throughout the current sampling.
(38)
Almeida, N.
F.;
Wingard,
L.
B.
Jr.; Malmros, M.
K.
Ann. N.Y. Acad. Sci.
1990,
613,
448-451.
Voltammetry in Vivo.
Male SpragueDawley rats (ca.
300
g initial weight) were stereotaxically implanted, under chloral
hydrate anaesthesia
(3.5
mL/kg i.p.,
10%
solution in normal
saline), with working electrodes in specific brain structures
as described in detail.37 The working electrode combination
was either a PPD/GOx-modified Pt electrode and a CPE or
bilateral CPEs placed in the right and left dorsal striatum,
respectively. The coordinates used, with the skull leveled
between bregma and lambda, were: AP
-0.5
mm (from
bregma), L
f
3.0
mm, DV
-4.8
mm (from skull). Silver
wires (250 pm) placed in the cortex served as reference and
auxiliary electrodes. The reference potential provided by the
Ag wire in brain tissue is very similar to that of the SCE.39
The animals were placed in the recording cage immediately
after surgery, connected to the recording equipment, and left
to recover for at least 24 h. The light/dark cycle was the
same as that of the presurgery housing room: 12/12 h light/
dark with lights on at
8
a.m. Food and water were available
ad libitum. The protocol for these experiments was approved
by the college President's committee on animal welfare.
Staircase voltammograms were recorded with CPEs at 12-
min intervals at a scan rate of either 10 or
50
mV/s. Current-
voltage plots show a distinct peak between
0
and
300
mV. The
peak, centered at about 200 mV, corresponds to the oxidation
of AA present in the extracellular fluid (ECF) around the
electrode tip.39s40 The height of the peak,
h,
was measured
by constructing a baseline between the two minima and
measuring the distance from this to the current maximum.
Thevalue of potential for the amperometric experiments with
CPEs was selected from these linear sweep recordings. A
potential of +0.2 V was chosen as this was just above the peak
potential for AA oxidation.40 Amperometry was carried out
at the Pt/PPD/GOx electrode in vivo at a potential of
+0.7
V.
Data Analysis.
Because of the classical nonlinearity of
enzyme/substrate calibrations both in solution and for
immobilized enzymes,41 the apparent Michaelis-Menten
constants,
Vma,
(nA) and
Km
(mmol/L), were determined
using the following Hill-type equation
I=*
(111)
[glucl
where the current
I
is a measure of the rate of reaction and
CY
is used as a measure of the deviation from ideal Michaelis-
Menten behavior.
All data are reported as mean
f
SEM. The significance
of differences observed was estimated using Student's two-
tailed t-tests. Paired tests were used for comparing signals
recorded with the same electrode (e.g., before and after drug
administration); unpaired tests were used for comparing data
from different electrodes.
RESULTS
AND
DISCUSSION
Glucose Calibrations.
Figure 1 (top) illustrates the effect
on the amperometric current of adding an aliquot of glucose
(39)
ONeill,
R.
D.
Analyst
1993,
118,
433438.
(40) Boutelle,
M.
G.;
Svensson,
L.;
Fillenz,
M.
Neuroscience
1989,
30,
11-17.
(41)
Inoue,
Y.;
Kato,
Y.;
Sato,
K.
J.
Chem.
Sw.,
Faraday
Trans.
1992.88.449-
454.
1758
AnalyticalChemistry,
Vol.
66,
No.
10,
May
15, 1994
5
1
50
mM Gluc
..
............
-
PBS
-
1
mMM
_.___.
30
mM
Gluc
1-
-1
0
80
160
240
320
400
480
time/s
I
/"
20
10
0
0
20
40
60
80
100
[Gluc]/mM
Flgure
1.
(top) Examples
of
the amperometric current increment
(A
r)
recorded
wkh
a Pt/PPD/GOx electrode at
700
mV versus
SCE
for
the addition
of
20
mmol/L glucose Into a PBS solution containing either
30
mmol/L glucose or
30
mmol/L glucose plus
1
mmol/L ascorbate
(AA).
(bottom) Glucose calibrations of PtIPPDIGOx electrodes
(n
=
3)
in the range
0- 100
mmol/L glucose carried out either In PBS or
a
solution
of
1
mmol/L
AA
in PBS.
The
ordinate
is
A
Zsince
the
current
recorded at zero glucose concentration
has
been subtracted.
See
Glucose Calibrations.
toa quiescent solution recorded with a Pt/PPD/GOx electrode
at
0.7
V vs SCE. The response time was satisfactory, with
the
95%
steady-state response being achieved in less than the
mixing time (ca. 10
s),
as reported by Malitesta et al.33 A
similar response time was observed in hydrodynamic studies.
This behavior is not unexpected as there is evidence that PPD
films formed under similar conditions have a thickness of the
order of
5-10
nm33342
so
that GOx molecules (diameter ca.
9
nm43) would be located at the polymer/electrolyte interface32
(see Figure
4,
later). The increase in current, however, caused
by adding glucose was clearly less in the presence of
1
mmol/L
AA compared with its absence. There are at least three
possible mechanisms that would lead to reduced sensitivity to
glucose in the presence of AA: (i) fouling of the Pt surface
by AA; (ii) "clogging" of the membrane by AA; (iii) a
homogeneous reaction between AA and H202 (reaction IV).
This reaction has recently been observed at bare Pt electrodes,35
and homogeneous interference by H202 in AA detection44
and by AA in a spectrochemical assay for oxalate45 has also
(42)
Ohnuki,
Y.;
Matsuda, H.; Ohsaka, T.; Oyama, N.
J.
Electroanal. Chem.
(43)
Nakamura,
S.;
Hayashi,
S.;
Koga,
K.
Eiochim. Eiophys. Acta
1976,
445,
(44)
Matuszewski,
W.;
Trojanowicz, M.
Electrounalysis
1990,
2, 147-153.
(45)
Inamdar,
K.
V.;
Raghavan,
K.
G.;
F'radhan, D.
S.
Clin. Chem.
1991,
37,
Interfacial Electrochem.
1983,
158,
55-67.
294-308.
864-868.
been reported. To investigate more fully the processes involved,
we next quantified the glucose and AA concentration
dependence of this negative interference.
M'+
H202
+
ascorbic acid
-
2H20
+
dehydroascorbic acid
(IV)
Calibration plots for 0-100 mmol/L glucose for Pt/PPD/
GOx electrodes in the presence and absence of 1 mmol/L AA
are shown in Figure 1 (bottom). The calibrations followed
Michaelis-Menten kinetics in both cases, since
a
(see eq
111)
was not significantly different from unity in the absence (1.03
f
0.02)
or presence
(0.96
f
0.04,
n
=
3,
p
>
0.2)
of AA.
Although
Km
was similar under the twoconditions (AA absent,
29
f
3
mmol/L; AA present,
29
f
4
mmol/L,
n
=
3,p
>
0.9),
the
Vma,
value calculated for PBS solutions,
44
f
8 nA
(n
=
3),
was reduced by
24
f
6%
(p
C
0.05)
by 1 mmol/L AA.
Similar percentage decreases were observed over the range of
glucose concentrations studied (Figure
1,
bottom) ranging
from
22
f
9%
(n
=
3) at 1 mmol/L glucose to
25
f
5%
(n
=
3) at
100
mmol/L. This reduction was less than that
reported in our recent corresponden~e~~ because the decrease
in electrode sensitivity to glucose over time was not taken into
account in the earlier work. By use of stability data,34 the
recalculated reduction in
Vma,
for
500
pmol/L AA reported
previously was 30
f
13%,
n
=
3.
The GOx activity of the PPD electrodes used in this study
compares well with previous reports.33 The current density
corresponding to the
Vma,
value calculated from calibrations
in PBS was 88
f
15
pA/cm2
(n
=
10).
Ascorbate Amperometric Titrations.
To ensure saturation
of the enzyme by glucose in these experiments, the maximum
flux of H202 was generated
(66
f
7
nA,
n
=
3) using a
concentration of glucose (100 mmol/L) well in excess of the
Km.
The amperometric glucose current at
0.7
V was recorded
as a function of AA concentration
(0-10
mmol/L) in order
to determine the AA concentration dependence of the negative
interference. Injection of AA aliquots into the cell containing
100 mmol/L glucose produced a rapid
decrease
in the current
that reached a new relatively stable value (inset Figure
2);
the
increases associated with direct oxidation of AA were negligible
due to the blocking effect of the polymer fi1m.32*33~35~42~46 The
half-life of this response to AA was similar for each AA
addition at ca. 30
s.
The low-frequency fluctuations in the
glucose signal (see inset Figure
2)
which were often observed
in the presence of AA may be due to oscillation reactions
between the H202 and AA, as has been reported in other
H202-containing system~.~~,~* A plot of normalized steady-
state currents for three electrodes versus AA concentration
is shown in Figure
2.
The glucose current was very sensitive
to AA at low concentrations giving a half-maximum decrease
of
22
f
6%
(n
=
3) at ca. 300 pmol/L AA. The effective
maximum decrease (38
f
8%,
n
=
3) was observed at
2
mmol/L
since this was not significantly different from the value at 10
mmol/L
(45
f
7%,
n
=
3). The fit of these data to an
exponential decay curve was good
(eq
V,
r
=
0.959)
considering
~ ~~
(46)
Cheek,
0.;
Wales,
C.
P.;
Nowak,
R.
J.
Anal.
Chem.
1983,55,
380-381.
(47)
Mori,
Y.;
Hanazaki,
I.
J.
Phys.
Chem.
1992,96,9083-9087.
(48) Kopcr, M.
T.
M.; Meulenkamp,
E. A.;
Vanmaekelbergh, D.
J.
Phys.
Chem.
1993,
97,
7337-7341.
Analytical Chemistry.
Vol.
66,
No.
10,
May
15,
1994
1757
50
WM
PA
+
50
pM
AA
64
62
90
-
80
-
012345
time
/
min
50
0
2
4
6
8
10
[AAI/mM
Flgure
2.
Normallzed steady-state
100
mmoi/L glucose current at
Pt/PPD/GOx
electrodes
(n
=
3)
as a function of ascorbate (AA)
concentration between
0
and
10
mmol/L. The half-maxlmum decrease
was observed at
ca.
300
pmoi/L AA and the maximum decrease was
45
f
7%.
The nonilnear regresslon curve is given by
eq
V.
Inset:
Example
of
the effect of an AA injection on
the
absolute amperometric
current.
See
Ascorbate Amperometric Titrations.
the experimental error. A model for this behavior is proposed
below.
I(%)
=
(58
f
2%)
+
(34
f
4%) exp{(-1.7
f
0.5
mM-')[AA]] (V)
To test the possibility that AA wasinterfering with transport
through the PPD membrane, the effect of AA on the glucose
response obtained at membrane-free GOx-modified CPWEs
was investigated. The current density corresponding to the
100
mmol/L glucose response at these membrane-free
electrodes was 215
f
17 pA/cm2
(n
=
3), indicating both
good loading and activity of the adsorbed enzyme and good
reproducibility of the adsorption method. When 100 mmol/L
glucose was added to PBS containing
1
mmol/L AA, the
increase in current recorded with the GOx-modified CPWEs
was 37
f
11%
(n
=
3,
p
C
0.04)
smaller than that obtained
with the same electrodes for similar additions of glucose into
PBS alone. The finding that this value was not significantly
different from thedecrease (33
f
7%,
n
=
3,p
C
0.03) observed
for corresponding concentrations at Pt/PPD/GOx electrodes
indicates that interference with transport through the PPD
membrane is
not
responsible for the depression by AA of the
glucose sensitivity at Pt/PPD/GOx electrodes.
Titrations of urate
(0-1
00
pmol/L) had
no
significant effect
on the glucose current recorded in experiments similar in design
to those for AA. We thus conclude that urate does not interact
appreciably with H202 in solution, as suggested by Tyagi and
Dryh~rst.~~
EDTA
Studies.
Previous reports in the literature of the
presence of a homogeneous redox reaction between AA and
(49)
Tyagi,
S.
K.;
Dryhurst, G.
J.
Electroanal. Chem. Interfacial Electrochem.
1984,
174,
343-364.
1758
Analytical
Chemistry,
Vol.
66,
No.
10,
May
15, 1994
H202 (reaction IV) suggest that it only occurs in the presence
of a catalyst, e.g., small amounts of heavy metal ions (Fe3+
and Cu2+,50-52 or peroxidase enzymes (ascorbate peroxidase53
and horseradish peroxidases4). Even when the reaction was
observed in the absence of added catalysts, it was assumed to
be promoted by impurities in the buffer.50 To test the
possibility that the homogeneous reaction(1V) was occurring
at Pt/PPD/GOxelectrodes, catalyzed by metal ion impurities
in the PBS, we investigated the effect of the chelator EDTA,
which has been used to inhibit the metal catalyzed oxidation
of AA.55356
In
the absence of EDTA the
100
mmol/L glucose
current was reduced by the addition of
1
mmol/L AA from
38
f
3 nA to 27
f
3 nA
(n
=
9), representing an average
decrease of 29
f
5%
(p
C
0,001).
In the presence of EDTA
(1 mmol/L), however, addition of AA did not affect the glucose
response: 52
f
7 nA
(n
=
5)
before AA to 52
f
7 nA
5
min
later
(p
>
0.9). This one particular finding provides strong
evidence that the reduced sensitivity of Pt/PPD/GOx elec-
trodes to glucose in the presence of AA (Figures 1 and
2)
is
not due to either fouling of the Pt surface by AA as suggested
recentlys7 or "clogging" of the membrane by AA, but to an
homogeneous reaction between AA and H202 (reaction IV).
It was noted that addition of 1 mmol/L EDTA caused a
small, marginally significant, increase in the
100
mmol/L
glucose current of 2.2
f
0.8%,
n
=
5,
p
<
0.07, but no increase
when glucose was absent from the solution. This may be due
to a blocking by EDTA of either heavy metal catalyzed
degradation of H20258,59 or heavy metal inhibition of GOx.60
Since the proposed homogeneous redox reaction would take
place near the Pt/PPD/GOx surface, and might be affected
by it, we examined the extent of this negative interference
reaction in the bulk solution to determine whether the problem
might have a more general significance. Amperometric
peroxide calibrations
(0-1
mmol/L) for unmodified bare Pt
disk electrodes at +0.7
V
vs SCE were performed in PBS
(containing low micromolar background levels of heavy metal
ions, see Chemicals and Solutions) in the presence of either
AA (1 mmol/L) or AA (1 mmol/L) and EDTA (1 mmol/L).
When EDTA and AA were both present, the concentration
steps were well-defined for each addition of H202 (Figure 3,
top). The calibration slope was 40
f
13 nA/mM,
(n
=
4),
which is in the range expected-between fresh Pt disks (1 12
f
8 nA/mM,
n
=
11) and surfaces several weeks old
(28
f
3 nA/mM,
n
=
5).
When EDTA was absent from the AA
solution, the injection of H202 resulted in an initial increase
in current which decayed with an average half-life of 33
f
10
s
(n
=
8)
and a pseudo-first-order rate constant of (3.2
f
0.6)
X
s-*
(n
=
8),
for the metal ion catalyzed homogeneous
reaction (Figure
3,
bottom). The H202 calibration slope in
the absence of EDTA was significantly reduced to
I
f
3 nA/
mM
(n
=
4),p
C
0.001 compared with EDTA present. Thus,
(50)
Hand, D.
B.;
Greisen,
E.
C.
J.
Am. Chem. SOC.
1942,
64,
358-361.
(51)
DeChatelet,
L.
R.;
Cooper, M.
R.;
McCall,
C.
E.
Anfimicrob. Agenfs
(52)
Timberlake, C.
F.
J.
Sci.
Food
Agric.
1960,
11,
268-273.
(53)
Kelly, G.
J.;
Latzko,
E.
J.
Agric.
Food
Chem.
1980,
28,
1320-1321.
(54)
Maidan,
R.;
Heller, A.
Anal.
Chem.
1992,
64,
2889-2896.
(55)
Butt,
V.
S.;
Hallaway, M.
Arch. Biochem. Biophys.
1961,
92,
24-32.
(56)
Deutsch,
J.
C.;
Kolhouse,
J.
F.
Anal. Chem.
1993,
65,
321-326.
(57)
Palmisano,
F.;
Zambonin, P. G.
Anal. Chem.
1993,
65,
2690-2692.
(58) Aruoma,
0.
I.
Chem. Br.
1993,
29,
210-214.
(59)
Dekker, A.
0.;
Dickinson,
R.
G.
J.
Am. Chem. SOC.
1940,
62,
2165-2171.
(60)
Nakamura,
S.;
Ogura,
Y.
J.
Eiochem.
(Tokyo)
1968,
64,
439447.
Chemother.
1972,
I,
12-16.