New bioorganometallic ferrocene derivatives as efficient mediators for glucose and ethanol biosensors on PQQ-dependent dehydrogenases

Article abstract of DOI:10.1016/S0022-328X(02)02103-4

One known and two new ferrocene-containing mediators incorporating the organometallic moiety and the fragments of natural substrates of oxidative enzymes, viz. 4-ferrocenylphenol (FP), 2-ferrocenyl-4-nitrophenol (FNP), and N-(4-hydroxybenzylidene)-4-ferrocenylaniline (HBFA), were studied as electron transfer mediators between the coenzyme pyrroloquinoline quinone (PQQ) of glucose (GDH) and alcohol (ADH) dehydrogenases and the carbon electrode surface. A screen-printed carbon electrode (SPCE) suitable for ADH and GDH immobilization served as a transducer. The electrodes were integrated into a flow-through amperometric cell. All data were obtained at a flow rate of 1 ml min^-1. The maximal currents (j_max) obtained from the calibration curves for the oxidation of ethanol and D-glucose by ADH and GDH of 2.3 and 3.0 μA, respectively, were obtained when SPCE was modified with HBFA, i.e. with a mediator with a longer arm and a high degree of conjugation. The biosensors were used for ethanol and D-glucose measurements in beverages. There was a good correspondence (r=0.978 for D-glucose and r=0.920 for ethanol) between the data obtained by using the biosensors, on one hand, and by the refractometric or hydrometric methods, on the other. The operational stability of biosensors is determined by the inactivation of the immobilized enzymes rather than by leakage of a mediator from an electrode.

Full text of DOI:10.1016/S0022-328X(02)02103-4

Journal of Organometallic Chemistry 668 (2003) 83ꢂ90
/
www.elsevier.com/locate/jorganchem
New bioorganometallic ferrocene derivatives as efficient mediators
for glucose and ethanol biosensors based on PQQ-dependent
dehydrogenases
J. Razumiene a,*, A. Vilkanauskyte a, V. Gureviciene a, V. Laurinavicius a,
Nataliya V. Roznyatovskaya b, Yulia V. Ageeva b, Marina D. Reshetova b,
Alexander D. Ryabov b,*
a Institute of Biochemistry, Mokslininku 12, LT-2600 Vilnius, Lithuania
b Department of Chemistry, Moscow State University, 119899 Moscow, Russia
Received 20 September 2002; received in revised form 8 November 2002; accepted 12 November 2002
Abstract
One known and two new ferrocene-containing mediators incorporating the organometallic moiety and the fragments of natural
substrates of oxidative enzymes, viz. 4-ferrocenylphenol (FP), 2-ferrocenyl-4-nitrophenol (FNP), and N-(4-hydroxybenzylidene)-4-
ferrocenylaniline (HBFA), were studied as electron transfer mediators between the coenzyme pyrroloquinoline quinone (PQQ) of
glucose (GDH) and alcohol (ADH) dehydrogenases and the carbon electrode surface. A screen-printed carbon electrode (SPCE)
suitable for ADH and GDH immobilization served as a transducer. The electrodes were integrated into a flow-through
amperometric cell. All data were obtained at a flow rate of 1 ml minꢀ1. The maximal currents (jmax) obtained from the calibration
curves for the oxidation of ethanol and
modified with HBFA, i.e. with a mediator with a longer arm and a high degree of conjugation. The biosensors were used for ethanol
and -glucose measurements in beverages. There was a good correspondence (rꢁ0.978 for -glucose and rꢁ0.920 for ethanol)
D
-glucose by ADH and GDH of 2.3 and 3.0 mA, respectively, were obtained when SPCE was
D
/
D
/
between the data obtained by using the biosensors, on one hand, and by the refractometric or hydrometric methods, on the other.
The operational stability of biosensors is determined by the inactivation of the immobilized enzymes rather than by leakage of a
mediator from an electrode.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Ferrocene; Electron transfer; Pyrroloquinoline quinone; Glucose dehydrogenase; Alcohol dehydrogenase; Screen-printed carbon
electrode; Biosensor
1. Introduction
sites of oxidoreductases and electrodes [1ꢂ6]. Ampero-
/
metric biosensors are often comprised of glucose oxidase
and a ferrocene derivative on a carbon paste electrodes
[7]. Ferrocenes are slightly soluble in water and this
reduces applications of such biosensors. Therefore,
biosensors incorporating ferrocene-based redox gels
have been introduced [2]. Hale et al. achieved the
electrical communication between GO and a conven-
tional carbon paste electrode by using covalently
attached ferrocene to insoluble siloxane polymers [5].
Koide and Yokoyama designed an enzyme electrode
based on GO and ferrocene-containing polymer made
by cross-linking of polyallylamine [8]. Similarly, a
reusable pseudo-reagentless glucose biosensor incorpor-
Ferrocene derivatives are widely used mediators
(electron shuttles) of the electron transfer between active
Abbreviations: SPCE, screen-printed carbon electrode; GO, glucose
oxidase; HRP, horseradish peroxidase; PQQ, pyrroloquinoline
quinone; ADH, PQQ-dependent alcohol dehydrogenase; GDH,
PQQ-dependent glucose dehydrogenase; CVA, cyclic voltammetry
and cyclic voltammogram.
* Corresponding authors. Present address: Department of
Chemistry, Carnegie Mellon University, 4400 Fifth Avenue,
Pittsburgh, PA 15213, USA. Tel.: ꢃ
2681061
E-mail address: ryabov@andrew.cmu.edu (A.D. Ryabov).
/
1-412-2686177; fax: ꢃ1-412-
/
0022-328X/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 3 2 8 X ( 0 2 ) 0 2 1 0 3 - 4

84
J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ90
/
ating a Nafion polymer and 1,1?-dimethylferrocene has
also been reported [9]. An HRP electrode with electro-
chemically deposited ferrocene-modified polyaniline or
phenylenediamine film has application in monitoring
peroxides in aqueous and non-aqueous media [10,11]. A
non-conducting polymer film of 3-aminophenol with
immobilized HRP was prepared on a carbon paste
electrode by the electrochemical polymerization. This
H2O2 sensor is applicable for fabricating biosensors for
glucose and cholesterol, which are not influenced by
Importantly, the mediators made by this approach react
fast also with other redox enzymes also including the
PQQ-dependent dehydrogenases [23]. This work is a
continuation of our studies in this field. It is demon-
strated here that 4-ferrocenylphenol (FP) and the new
biorganometallic mediators, viz. 2-ferrocenyl-4-nitro-
phenol (FNP) and N-(4-hydroxybenzylidene)-4-ferroce-
nylaniline (HBFA) are excellent electron transfer
mediators between the PQQ coenzyme of glucose and
alcohol dehydrogenases on screen-printed carbon elec-
trodes for D-glucose and ethanol monitoring.
easily oxidizable species such as L-ascorbic or uric acid
[12]. Mediators chemically bound to a flexible polymer
backbone for a better contact with the active centers of
GO have also been investigated. This has involved a use
of methyl(ferrocenylethyl)siloxane homopolymer and
2. Experimental
2.1. Chemicals and materials
methyl(ferrocenylethyl)dimethylsiloxane
copolymer
[13]. Tatsuma et al. reported a glucose electrode
fabrication using poly(N-isopropylacrylamide-co-vinyl-
ferrocene) and analyzed its performance both experi-
mentally and theoretically [14]. Recently, some novel
1,1?-dipeptide-ferrocenes were synthesized which are
being evaluated for a possibility of their binding to
aspartate proteases [15]. Yabuki et al. prepared a
polyionic membrane containing physically entrapped
microperoxidase and ferrocene on a glassy carbon. Thus
made electrode could be used in a matter of 10 days [16].
Some ferrocene derivatives have been tested as
mediators in pyrroloquinoline quinone (PQQ) depen-
dent dehydrogenase-catalyzed reactions. Due to the
capability of PQQ-dehydrogenases to donate electrons
to acceptors other than dioxygen, these enzymes are
unique in the biosensor engineering as such biosensors
Carbon rods (Ultra ‘F’ Purity Ultra† carbon) 3 mm
in diameter were obtained from Ultra Carbon Division
of Carbon USA Corp. MINICO M 7000 G isolating
and Ag MINICO M 4200 pastes were obtained from
Emerson&Cuming Specialty Polymers (Westerio, Bel-
gium). A carbon paste electrode on a RAVEN-M
carbon black, the latter obtained from Columbian
Chemicals Co. (Atlanta, USA), was developed in the
Institute of Biochemistry (Vilnius, Lithuania) [27].
Citrateꢂphosphate buffer (0.1 M, pH 6.0) was used as
/
a default buffer. All inorganic and organic reagents used
were of the highest purity available. Doubly distilled
water was obtained using a ‘Purator Bi’ device (Glas
Keramic, Germany). 4-Ferrocenylphenol (FP) was ob-
tained as previously described [26].
are insensitive to the dioxygen level in samples [17ꢂ20].
/
2.1.1. 2-Ferrocenyl-4-nitrophenol
Promising results have also been obtained using PQQ-
dependent membrane-bound aldose dehydrogenase and
1,1?-dimethylferrocene or ferrocene carboxylic acid as
mediators [21]. This aldose sensor has been improved by
using some new polymer-bound ferrocene derivatives
[22]. We have previously shown that 4-ferrocenylphenol
adsorbed on graphite electrodes mediates the ADH- and
GDH-catalyzed reactions [23]. The biosensors were then
improved by using SPCE [24].
2-Amino-4-nitrophenol (3.85 g, 0.025 mol), which was
obtained from Aldrich, was placed in a 500 ml glass
beaker equipped with a stirrer. Water (35 ml) and concd.
HCl (6.4 ml) were added. The mixture was cooled to ꢀ
/
2 8C using an ice bath and a solution of NaNO2 (2.1 g,
0.03 mmol) in 5 ml H2O was added in a matter of 40 min
to avoid heating the mixture above 0 8C. Excess of
nitrite was neutralized by urea and the solution pH was
adjusted to 5ꢂ6 by NaOAc. After addition of 50 ml
/
The fixation of mediators on various supports is an
important component in construction of modern bio-
sensors. Other research is focused on substantial im-
provement of the mediator performance in terms of
stability and reactivity [25]. We have recently introduced
the ‘biorganometallic’ strategy of the mediator design. It
is based on combining a redox-active organometallic
fragment and an organic moiety reminiscent of the
substrate of HRP in one molecule (Scheme 1) [26].
water, this solution was mixed using an addition funnel
with a stirred solution of ferrocene (3.64 g, 0.025 mol) in
100 ml CH2Cl2 placed in a 500 ml round-bottom flask.
The mixing continued for 1 h and the temperature of the
flask was kept at ꢀ3 8C using an ice bath. The
/
temperature was raised to 22 8C; the mixture was stirred
for 2 h, and then heated to 60 8C. The color changed
from greenishꢂ
separated, SnSO4 was added to reduce the ferricenium
ion, and the products were extracted with ether (3ꢄ30
/
yellow to dark-red. An aq. layer was
/
ml). An organic layer was washed twice with water,
combined with the ether solution, and dried with
Na2SO4. The solvent was removed using a rotary
Scheme 1.

J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ
/90
85
evaporator, and the residue was column-chromato-
graphed (17ꢄ3 cm) on Al2O3. Unreacted ferrocene
(1.9 g) was eluted with petroleum ether and FNP with
electric contact. Finally, the working electrode was
coated with 3 mL FP, FNP or HBFA (1 mg mlꢀ1 in
C3H6O), 2 ml ADH or GDH, held for 30 min over 25%
aq. solution of glutaric aldehyde and covered with semi-
permeable terylene film [31]. The enzyme electrode was
installed into a home-made flow-through three-electrode
amperometric cell made from polymethylmethacrylate
[24]. The working electrode (0.04 cm2) is a bottom of the
flow cell. A Ag wire (1 mm diameter and 2 cm length) in
saturated KCl was used a reference electrode. A Ti rod
(2 mm diameter and 2 cm length) was used as a counter
electrode.
/
EtOAc. FNP (1.1 g) was isolated. Crystallization from a
1:1 EtOHꢂwater mixture gave 0.97 g of analytically
/
pure material (33% with respect to consumed ferrocene).
M.p. 114 8C. 1H-NMR (d, CDCl3): 4.32s (5H, Cp),
4.55s and 4.61s (2H, a and b C5H4), 7.00d (H5, J 8 Hz),
7.96bs (OH), 8.09dd (H4, J 8 and 2 Hz), 8.12d (H3, J 2
Hz). IR (KBr): 1290, 1336, 3300ꢂ
3550 cmꢀ1. Anal.
Found: C, 59.21; H, 4.17; Fe, 16.93. Calc. for
C16H13FeNO3: C, 59.17; H, 4.05; Fe, 17.28%.
/
2.1.2. N-(4-Hydroxybenzylidene)-4-ferrocenylaniline
4-Ferrocenylaniline (0.4 g, 0.0014 mol) obtained as
described elsewhere [28] was dissolved in 10 ml of EtOH
on heating and mixed with 4-hydroxybenzaldehyde (0.1
g, 0.0014 mol) in 5 ml of EtOH and the mixture was
refluxed for 60 min. The precipitated dark cherry-red
material was filtered off, washed with EtOH, and dried
using an Abderhalden drying vacuum apparatus. Yield:
2.4. Instrumentation
The flow cell was supplied with a ZALIMP 315
peristaltic pump (Poland). The pumping rate was 1 ml
minꢀ1. Steady-state currents were recorded using a
Radelkis OH-105 polarograph (Hungary). The reference
data on
obtained by using a
D
-glucose concentrations in beverages were
-glucose analyzer EKSAN-G
D
66% (0.35 g). M.p. 189ꢂ
/
191 8C. IR (Nujol): 3100br
(Analita, Lithuania) and a RE 40 refractometer (Met-
tler-Toledo, Switzerland). Ethanol was quantified by a
hydrometer graduated to 0.1 proof. CVA experiments
were carried out using an Ag/AgCl reference electrode.
All potentials are against this electrode throughout. A
Pt coil and a carbon rod sealed in a Teflon tube were
counter and working electrodes, respectively. The sys-
tem was operated by a PA-2 polarographic analyzer
(Laboratorni pristroje, Czechoslovakia).
1
(OH), 1635 (Cꢀ
/
N), 1600, 1530, 1120, 1000 cmꢀ1. H-
NMR (d, d6-Me2SO): 4.02s (5H, Cp), 4.34s and 4.78s
(2H, a and b C5H4), 6.88 and 7.55 (4H, AA?BB? system
from C6H4O), 7.17 and 7.78 (4H, AA?BB? system from
C6H4N), 8.52s (1H, CHꢀ
/
N), 10.1bs (1H, OH). Anal.
Found: C, 72.25; H, 5.18; Fe, 14.46. Calc. for
C23H19NO: C, 72.46; H, 5.02; Fe, 14.65%.
2.2. Enzymes
2.5. Determination of D-glucose in beverages
ADH was purified from Gluconobacter sp.33 [29]
40 U mgꢀ1) and used as a solution
(specific activity 25ꢂ
/
For determining a total amount of glucose and
sucrose in soft drinks, sucrose was inverted into
glucose and -fructose by Maxinvert L 1000 invertase.
A 10 ml sample was diluted 100-fold, incubated for 20
min with 0.1 ml of invertase at 22 8C, and -glucose was
in 0.02 M potassium phosphate buffer (pH 7.0) contain-
ing 0.02% Triton X-100 and 0.5% sucrose. GDH was
purified from Erwinia sp. 34-1 [30] (specific activity 12 U
mgꢀ1) and used as a solution in 0.02 M the same buffer
containing 10% glycerol. Invertase (Maxinvert L 1000)
of specific activity 72 U mlꢀ1 at 20 8C and pH 4.5 was
purchased from Gist-Brocades (Germany).
D-
D
D
then measured in the flow-through biosensor system.
3. Results and discussion
2.3. Construction of screen-printed biosensor and
electrochemical flow-through cell
3.1. Synthesis of ‘bioorganometallic’ mediators
Screen-printed carbon biosensors were constructed by
using the following procedure [24]. A Ag track of
MINICO M 4200 paste (10 mm thick) on the poly-
ethylene terephthalate film (180 mm) was produced by
the screen-printing technology. Printing on the Ag track
by the conducting carbon ink [27] produced the surface
of the working electrode. Water-insoluble insulating
MINICO M 7000 G paste layer was deposited on the
electrode surface. Two openings were made, viz. a
carbon ink window with an area of 0.04 cm2 for a
working electrode and a 0.5 cm2 Ag track for an external
Synthetic routes to new organometallic molecules are
shown in Schemes 2 and 3. Preparation of FNP (Scheme
Scheme 2.

86
J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ
/90
Scheme 3.
2) involved transformation of 2-amino-4-nitrophenol
into the corresponding diazonium cation which was
used for arylation of ferrocene to afford FNP. The
Schiff base HBFA (Scheme 3) was made from 4-
ferrocenylaniline, the preparation of which was de-
scribed in detail previously [26], and 4-hydroxybenzal-
dehyde. Both compounds were thoroughly characterized
by spectral and analytical data.
segment of HBFA is oxidized at a more positive
potential (0.68 V) as compared with Ea of 0.6 V for
FP (cf. peaks 2 and 5). The oxidation potentials of the
both moieties of FNP are anodically shifted and seen at
0.47 and 0.75 V (8 and 9). The Ec of 0.35 and 0.5 V
characterize the reduction of FP and FNP at the carbon
electrode, respectively (3 and 10), whereas two reductive
peaks are found for HBFA, viz. at 0.3 and 0.45 V (6 and
7 in Fig. 1). It should be noted that the peak current for
the oxidation of phenolic moieties of ferrocene deriva-
tives decreases with increasing the number of scans. This
effect was previously studied by the example of FP and
accounted for electrochemical polymerization of the 4-
hydroxyphenyl fragments [23]. This process also de-
creases slightly the peak current from the ferrocenyl
groups. At lower potentials, when no polymerization
occurs, all investigated mediators display steady redox
behavior of the ferrocene moiety (Fig. 2). In addition,
the values of Ea and Ec are practically independent, but
the peak currents increase proportionally to the sweep
rate. This emphasizes good electrochemical reversibility
of the ferrocenes and suggests that they could be
promising mediators in the ADH and GDH-catalyzed
3.2. Electrochemical characterization
FP adsorbed on a carbon electrode was previously
studied by CVA [23]. The CVAs exhibited a pair of
anodic peaks (Ea) at 0.35 and 0.65 V and poorly
resolved cathodic peak (Ec) at 0.2 V. The first anodic
peak was ascribed to the oxidation of the ferrocene
group and the second referred to the oxidation of the 4-
hydroxyphenyl moiety. CVAs of FP, HBFA, and FNP
obtained on SPCE are shown in Fig. 1. There is a pair of
oxidation peaks corresponding to the oxidation of
ferrocene and 4-hydroxyphenyl fragments similar to
FP [23]. Slightly lower compared to FP (peak 1 in Fig.
1) value of Ea (0.35 V) was obtained for the oxidation of
ferrocenyl moiety of HBFA (peak 4). The phenolic
Fig. 1. CVAs of FP, FNP, and HBFA adsorbed on SPCE: 0.1 M,
KCl; scan rate, 10 mV sꢀ1. See text for details.
Fig. 2. CVAs of HBFA adsorbed on SPCE at different potential
sweep rates (in mV sꢀ1).

J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ
/90
87
ethanol and D-glucose oxidations at potentials, when the
polymerization of the 4-hydroxyphenyl moieties is
eliminated.
3.3. Evaluation of mediating efficiency of FP, HBFA, and
FNP in glucose biosensor
The main objective of this work was evaluating
ferrocene compounds as effective electron transfer
mediators in GDH-catalyzed oxidation of D-glucose,
since GDH show no electrocatalytic behavior on carbon
electrodes in the absence of mediating material. The
current responses of the carbon electrode modified with
GDH, on one hand, and FP, HBFA, or FNP, on the
other, were investigated as a function of the working
potential. It was always below 0.5 V, thus eliminating
the electropolymerization. Measurements were per-
formed in a flow system in the buffer using 0.5 mM
glucose solutions. CVAs obtained are shown in Fig. 3.
The highest catalytic current, close to the FP-modified
electrode, was observed for HBFA as a mediator. The
maximal responses for FP and FNP electrodes are
obtained at potentials around 0.4 V, i.e. slightly higher
than required for the highest response of the HBFA-
modified electrode. Obviously, a highly conjugated arm
of HBFA facilitates the electron exchange rate. This
leads to increased sensitivity of the biosensor even at low
potentials (Fig. 3). The negative mesomeric effect of the
nitro group in FNP shifts the peak anodically and
decreases the rate of the electron exchange between the
active center of enzyme and the mediator.
Fig. 4. Calibration curves for D-glucose obtained using the carbon
electrode modified with GDH and HBFA, FP, and FNP. Applied
potential, 0.4 V; carrier buffer, 50 mM acetate (pH 6) with 1 mM
Ca2ꢃ
.
lated from data in Fig. 4 equal 2.4, 0.83, and 2.7 mM for
HBFA, FP, and FNP, respectively. Clearly, structurally
different ferrocenes display different mediating abilities.
The best performance for HBFA, viz. a compound with
a 4-iminomethylphenolic arm, is likely due to the fact
that the latter is longer than the phenolic arms of FP
and FNP. An extra reason for the better response could
be higher hydrophobicity of HBFA that enables a better
interaction between ferrocene and the cofactor buried in
the protein globule of the enzyme. A clear dependence of
the current response on the polymer structure was also
observed for PQQ-dependent aldose dehydrogenase by
Smolander et al. [22].
Fig. 4 shows the glucose calibration curves for GDH
electrodes modified with FP, HBFA, and FNP at the
working potential of 0.4 V, which was selected for
comparing ferrocenes used. The highest current density
(28 mA (cm2 mM)ꢀ1) was observed for the HBFA-
modified carbon electrode. For FP and FNP this equals
3.4. Evaluation of mediating efficiency of FP, HBFA, and
FNP in ethanol biosensor
In contrast to GDH, ADH adsorbed on carbon
electrodes is involved in non-mediated electron ex-
change with an electrode in the presence of ethanol
[32,33]. This is due to the structure of the enzyme. ADH
contains four heme moieties, which act as bridges for the
electron transfer from PQQ to the electrode surface.
Therefore, the catalytic efficacy of the mediated oxida-
tion of ethanol was compared with the non-mediated
process. The current responses of the carbon electrode
modified with ADH and FP, HBFA, FNP or ADH only
were obtained as a function of the working potential.
Measurements were performed in a flow system in the
buffer using 1 mM ethanol solutions. CVAs obtained
are shown in Fig. 5. The highest and the lowest catalytic
currents were observed for HBFA as a mediator and at a
bare carbon electrode, respectively. The relative med-
10.5 and 1.8 mA (cm2 mM)ꢀ1, respectively. The
app
apparent Michaelis constants KM
for glucose calcu-
iator efficacy (HBFAꢀ
/
FPꢀFNP) was as in the GDH
/
case, whereas a difference in the peak positions was
smaller. This could be accounted for in terms of the
cooperative effect of the PQQ and heme groups. As a
Fig. 3. Hydrodynamic voltammograms for the GDH and FP, HBFA,
FNP-modified carbon electrodes. Conditions: flow rate, 1 ml minꢀ1
229
2 8C, 50 mM acetate (pH 6), 1 mM Ca2ꢃ
;
/
.

88
J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ90
/
3.5. Kinetic evaluation of stability of biosensors
The operational stability of the GDH- or ADH-
modified carbon electrode with adsorbed HBFA was
studied in a flow system. The performance of the both
biosensors was investigated in a matter of 3 weeks at
room temperature at pH 6. Measurements were made
using standard ethanol or D-glucose solutions (0.5 mM)
at 0.4 V. Different kinetics of functioning of HBFA and
ADH- or GDH-modified carbon electrodes (Fig. 7)
indicates that the stability of the immobilized enzyme
plays a dominant role. This was proven by adding a
soluble mediator, i.e. phenazine methosulfate (data not
shown). The response of the ethanol biosensor decreases
by ca. 50% in 2 days presumably due to dissociation of
poorly bound ADH molecules from the carbon surface.
However, a decrease in the sensor response is indistinct
after 2 weeks. In contrast, the response of the glucose
biosensor increases during first working day. The
protein molecules are likely tightly bound to the
electrode surface and both GDH and the mediator
adopt an arrangement most favorable for the mutual
contact. Studies of time-dependencies of the perfor-
mance of both biosensors suggest that it is unaffected by
the mediator leaking, but the immobilization procedure
based on the cross-linking with glutaric aldehyde is more
appropriate for GDH than for ADH.
Fig. 5. Current response of the bare ADH-modified carbon electrode
and same electrode additionally modified with ferrocene derivatives
HBFA, FP, and FNP. Data is obtained in a flow system; 0.5 mM
EtOH; carrier buffer, 50 mM acetate (pH 6) with 1 mM Ca2ꢃ
.
result, the catalytic current is comparable for all three
mediators investigated.
The ethanol calibration curves obtained using the
ADH carbon electrode and ADH electrodes modified
with FP, HBFA, and FNP are shown Fig. 6. The highest
current density (28 mA (cm2 mM)ꢀ1) was reached for
the HBFA-modified electrode. It is at least 10 times
higher compared to the current obtained at the ADH-
modified bare carbon electrode. For the FP- or FNP-
modified electrodes, the values are 17.5 and 12 mA (cm2
mM)ꢀ1, respectively. Thus, HBFA provides the highest
electrocatalytic activity for the both PQQ-dependent
enzymes. One and the same catalytic mechanism is
3.6. Determination of D-glucose and ethanol in beverages
by carbon electrodes modified with FP and ADH or GDH
obviously operative in both cases (Section 3.3). The
app
The biosensors inserted in a flow system were
evaluated for measuring D-glucose and ethanol in
various non-alcoholic and alcoholic beverages. Soft
drinks produced in Lithuania (with or without CO2)
are made of oranges, peaches, pears, grapes or other
values of KM calculated from data in Fig. 6 for
HBFA, FP, and FNP equal 1.0, 1.5 and 2.5 mM,
respectively. The numbers are slightly higher than the
MichaelisꢂMenten constants obtained for the ADH
/
reactions in solution [29].
fruits. They contain 0.1ꢂ200 mM of D-glucose, which
/
Fig. 6. Calibration curves for ethanol obtained on the bare ADH-
modified carbon electrode and same electrode additionally modified
with ferrocene derivatives HBFA, FP, and FNP. Data is obtained in a
flow system; 0.5 mM EtOH; carrier buffer, 50 mM acetate (pH 6) with
Fig. 7. Stability of ethanol and glucose biosensors assembled on the
ADH- and GDH-modified carbon electrodes using FP as a mediator.
Experimental details: working potential, 0.4 V; carrier buffer, 50 mM
acetate (pH 6) with 1 mM Ca2ꢃ; response to 0.5 mM substrate, at least
15 assays per day.
1 mM Ca2ꢃ
.

J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ
/90
89
was quantified by using the carbon electrode modified
with GDH and FP. Samples were analyzed after dilution
and the invertase-catalyzed conversion of sucrose into
D-glucose. The data obtained by using the biosensors,
an EKSAN-G analyzer, and refractometrically are
compared in Fig. 8. The correlation coefficients of
0.978 and 0.975 as well as insignificant intercepts are
indicative of low background interfering currents.
Therefore, impurities and electrochemically active com-
pounds in beverages have a low influence on the
biosensors response.
The ADH/FP-based sensor was evaluated by measur-
ing ethanol in red and white wines that contain 9ꢂ18
/
wt.% ethanol. Samples were analyzed in a flow-through
system after ca. 10 000-fold dilution with the buffer.
Comparison of the bioanalytical and the standard
hydrometer-based methods is presented in Fig. 9. The
correlation coefficient equals 0.9204 indicating a good
applicability of the biosensor for measuring
and ethanol levels in foodstuffs.
D-glucose
Fig. 9. Comparison of the data obtained using the ethanol biosensor
and hydrometer. See text for details.
ability of the biosensors for quantifying ethanol and D-
glucose in beverages.
4. Conclusions
The carbon electrode modified with ‘bioorganometal-
lic’ ferrocene derivatives permits the detection of ethanol
Acknowledgements
and
maximal sensitivity of biosensors for ethanol and
D
-glucose in the potential range 0.25ꢂ0.45 V. The
/
Financial support from the Lithuania state Science
and Studies Foundation is acknowledged. The authors
thank L. Marcinkeviciene and I. Bachmatova for the
preparation of the enzymes.
D-
glucose were reached by using a carbon electrode
modified with HBFA and the corresponding PQQ-
dependent enzymes. The biosensors thus prepared
were used for measuring ethanol and
D-glucose in
beverages. Good agreement between the data obtained
bioanalytically, on one hand, and refractometrically or
hydrometrically, on the other, demonstrates the applic-
References
[1] A.E.G. Cass, G. Davis, G.D. Francis, H.A.O. Hill, W.J. Aston,
I.J. Higgins, E.O. Plotkin, L.D.L. Scott, A.P.F. Turner, Anal.
Chem. 56 (1984) 667.
[2] A.E.G. Cass, G. Davis, M.J. Green, H.A.O. Hill, J. Electroanal.
Chem. 190 (1985) 117.
[3] A. Heller, Acc. Chem. Res. 23 (1990) 128.
[4] A.D. Ryabov, Angew. Chem. Int. Ed. Engl. 30 (1991) 931.
[5] P.D. Hale, L.I. Boguslavsky, T. Inagaki, H.I. Karan, H.S. Lee,
T.A. Skotheim, Y. Okamoto, Anal. Chem. 63 (1991) 677.
[6] H.-z. Bu, S.R. Mikkelsen, A.M. English, Anal. Chem. 67 (1995)
4071.
[7] L. Gorton, Electroanalysis 7 (1995) 23.
[8] S. Koide, K. Yokoyama, J. Electroanal. Chem. 468 (1999) 193.
[9] R.S. Brown, J.H.T. Luong, Anal. Chim. Acta 310 (1995) 419.
[10] C.-L. Wang, A. Mulchandani, Anal. Chem. 67 (1995) 1109.
[11] A. Mulchandani, S. Pan, Anal. Biochem. 267 (1999) 141.
[12] Y. Nakabayashi, H. Yoshikawa, Anal. Sci. 16 (2000) 609.
[13] P.D. Hale, T. Inagaki, H.I. Karan, Y. Okamoto, T.A. Skotheim,
J. Am. Chem. Soc. 111 (1989) 3482.
[14] T. Tatsuma, K. Saito, N. Oyama, Anal. Chem. 66 (1994) 1002.
[15] Y. Xu, H.B. Kraatz, Tetrahedron Lett. 42 (2001) 2601.
[16] S. Yabuki, F. Mizutani, Y. Hirata, J. Electroanal. Chem. 468
(1999) 117.
Fig. 8. Comparison of the data obtained using the glucose biosensor
and refractometer (k) or analyzer ‘EKSAN-G’ (m). See text for
details.

90
J. Razumiene et al. / Journal of Organometallic Chemistry 668 (2003) 83ꢂ90
/
[17] V.L. Davidson (Ed.), Principles and Applications of Quinopro-
teins, Marcel Dekker, Inc, New York, Basel, Hong Kong,
1993.
[26] A.D. Ryabov, V.S. Kurova, V.N. Goral, M.D. Reshetova, J.
ˇ
Razumiene, R. .Simkus, V. Laurinavicius, Chem. Mater. 11
(1999) 600.
[27] V.-S. Laurinavie`ius, B. Kurtinaitiene, K. Bernotas, R. Meskys, R.
[18] B. Schmidt, Clin. Chim. Acta 266 (1997) 33.
[19] J.A. Duine, J. Biosci. Bioeng. 88 (1999) 231.
[20] V. Laurinavicius, J. Razumiene, B. Kurtinaitiene, I. Lapenaite, I.
Bachmatova, L. Marcinkeviciene, R. Meskys, A. Ramanavicius,
Bioelectrochemistry 55 (2002) 29.
ˇ
Simkus, V. Liauksminas, T. Skotheim, L. Boguslavsky, Patent Lt
4605 2000.01.25 OB, 2000/1.
[28] A.N. Nesmeyanov, E.G. Perevalova, R.V. Golovnya, L.S.
Shilovtseva, Dokl. Akad. Nauk SSSR 102 (1955) 535.
[29] L. Marcinkeviciene, I. Bachmatova, R. Semenaite, R. Rudomans-
kis, G. Brazenas, R. Meskiene, R. Meskys, Biologija (1999) 24.
[30] L. Marcinkeviciene, I. Bachmatova, R. Semenaite, R. Rudomans-
kis, G. Brazenas, R. Meskiene, R. Meskys, Biotechnol. Lett. 21
(1999) 187.
[21] M. Smolander, H.L. Livio, L. Rasanen, Biosens. Bioelectron. 7
(1992) 637.
[22] M. Smolander, L. Gorton, H.S. Lee, T. Skotheim, H.-L. Lan,
Electroanalysis 7 (1995) 941.
[23] J. Razumiene, R. Meskys, V. Gureviciene, V. Laurinavicius, M.D.
[31] B.V. Mchedlishvili, G.N. Flerov, Zh. Vses. Khim. O-va. im. D.I.
Mendeleeva 32 (1987) 641.
Reshetova, A.D. Ryabov, Electrochem. Commun.
2 (2000)
307.
[24] J. Razumiene, V. Gureviciene, V. Laurinavicius, J.V. Grazulevi-
cius, Sens. Actuators B: Chem. B78 (2001) 243.
[32] T. Ikeda, in: F.W. Scheller, F. Schubert, J. Fedrowitz (Eds.),
Frontiers in Biosensorics. Fundamental Aspects, vol. 1, Birkhau-
ser Verlag, Basel, 1997, p. 243.
[25] A.D. Ryabov, V.S. Sukharev, L. Alexandrova, R. Le Lagadec, M.
Pfeffer, Inorg. Chem. 30 (2001) 6529.
[33] J. Razumiene, M. Niculescu, A. Ramanavicius, V. Laurinavicius,
E. Csoregi, Electroanalysis 14 (2002) 43.