Redox hydrogel-based amperometric bienzyme electrodes for fish freshness monitoring

Article abstract of DOI:10.1021/ac990848+

This work presents the design and optimization of amperometric biosensors for the determination of biogenic amines (e.g., histamine, putrescine, cadaverine, tyramine, cystamine, agmatine, spermidine), commonly present in food products, and their application for monitoring of freshness in fish samples. The biosensors were used as the working electrodes of a three-electrode electrochemical cell of wall-jet type, operated at -50 mV vs Ag/AgCl, in a flow injection system. Two different bienzyme electrode designs were considered, one based on the two enzymes [a newly isolated and purified amine oxidase (AO) and horseradish peroxidase (HRP)] simply adsorbed onto graphite electrodes, and one when they were cross-linked to an Os-based redox polymer. The redox hydrogel-based biosensors showed better biosensors characteristics, i.e., sensitivity of 0.194 A M^-1 cm^-2 for putrescine and 0.073 A M^-1 cm^-2 for histamine, and detection limits (calculated as three times the signal-to-noise ratio) of 0.17 μM for putrescine and 0.33 μM for histamine. The optimized redox hydrogel-based biosensors were evaluated in terms of stability and selectivity, and were used for the determination of total amine content in fish samples kept for 10 days in different conditions.

Full text of DOI:10.1021/ac990848+

Anal. Chem. 2000, 72, 1591-1597
Redox Hydrogel-Based Amperometric Bienzyme
Electrodes for Fish Freshness Monitoring
Mihaela Niculescu, Catalin Nistor,† Ivo Fre´ bort,‡ Pavel Pecˇ,‡ Bo Mattiasson, and Elisabeth Cso1regi*
Department of Biotechnology, Lund University, P.O. Box 124, S-22100, Lund, Sweden
erine,8,10 produced by microbial decarboxylation of the amino acids
histidine, ornithine, and lysine, respectively, have been proposed.
The biogenic amine content of various foodstuffs has been
intensively studied due to their potential toxicity.11 Histamine is
the most biologically active compound from that class, affecting
the normal functions of heart, smooth muscle, motor neurons,
and gastric acid secretion.12 Other biogenic amines, such as
putrescine and cadaverine, may amplify the effects caused by
histamine intoxication, inhibiting the enzymes involved in hista-
mine biodegradation: diamine oxidase and histamine-N-methyl
transferase.13 Numerous countries adopted maximum levels for
histamine in food, especially in fish products; e.g. Italian laws fixed
this level at 100 mg kg-1 food,3 and similar limits have been
adopted by EEC regulations.6
Classical methods for the analysis of biogenic amines generally
involve chromatographic techniques, such as gas chromatogra-
phy,14 thin-layer chromatography,15 reversed-phase liquid chro-
matography,15,16 and liquid-chromatography with derivatization
techniques.17-19 However, they often require sample pretreatment
steps and skilled operators, and the relatively long analysis time
and high costs make these methods unsuitable for routine use.
Enzymatic determination of biogenic amines was previously
carried out and represents an alternative that can solve the above-
mentioned problems. In this context, amperometric,9-20 spectropho-
tometric,21-24 fluorimetric,25 or chemiluminometric detection meth-
ods14,26 have been used. Amperometric electrodes using AO as
the biological recognition element were also previously reported,
This work presents the design and optimization of am-
perometric biosensors for the determination of biogenic
amines (e.g., histamine, putrescine, cadaverine, tyramine,
cystamine, agmatine, spermidine), commonly present in
food products, and their application for monitoring of
freshness in fish samples. The biosensors were used as
the working electrodes of a three-electrode electrochemi-
cal cell of wall-jet type, operated at -5 0 mV vs Ag/ AgCl,
in a flow injection system. Two different bienzyme elec-
trode designs were considered, one based on the two
enzymes [a newly isolated and purified amine oxidase
(AO) and horseradish peroxidase (HRP )] simply adsorbed
onto graphite electrodes, and one when they were cross-
linked to an Os-based redox polymer. The redox hydrogel-
based biosensors showed better biosensors characteris-
tics, i.e., sensitivity of 0 .1 9 4 A M-1 cm-2 for putrescine
and 0 .0 7 3 A M-1 cm-2 for histamine, and detection limits
(calculated as three times the signal-to-noise ratio) of 0.17
µM for putrescine and 0 .3 3 µM for histamine. The
optimized redox hydrogel-based biosensors were evalu-
ated in terms of stability and selectivity, and were used
for the determination of total amine content in fish
samples kept for 1 0 days in different conditions.
Rapid evaluation of fish and meat quality is permanently
required in food industry, motivating a continuous search for
freshness biomarkers and efforts to develop simple and inexpen-
sive methods for their determination. Among these biomarkers,
inositol monophosphate, hypoxanthine, and xanthine, which are
intermediate degradation products of nucleic acids,1,2 and some
biogenic amines, such as histamine,3-7 putrescine,8,9 and cadav-
(10) Gasparini, R.; Scarpa, M.; Di Paolo, M. L.; Stevanato, R.; Rigo, A. Bioelec-
trochem. Bioenerg. 1 9 9 1 , 25, 307-315.
(11) Taylor, S. L. Crit. Rev. Toxicol. 1 9 8 6 , 17, 91-128.
(12) Sttraton, J. E.; Hutkins, R. W.; Taylor, S. J. Food Protect. 1 9 9 1 , 54, 460-
470.
(13) Hui, J. Y.; Taylor, S. L. Toxicol. Appl. Pharmacol. 1 9 8 5 , 8, 241-249.
(14) Bachrach, U.; Plesser, Y. M. Anal. Biochem. 1 9 8 6 , 152, 423-431.
(15) Yang, X.; Rechnitz, G. A. Electroanalysis 1 9 9 5 , 7, 105-108.
(16) Hauschild, M. Z. J. Chromatogr. 1 9 9 3 , 630, 397-401.
(17) Hui, J. Y.; Taylor, S. L. J. AOAC Int. 1 9 8 3 , 66, 853-857.
(18) Yen, G. C.; Hsieh, C. L. J. Food Sci. 1 9 9 1 , 56, 158-160.
(19) Veciana-Nogues, M. T.; Hernandez-Jover, T.; Marine-Font, A. J. AOAC Int.
1 9 9 5 , 78, 1045-1050.
† Department of Analytical Chemistry, Lund University, P.O. Box 124, S-22100,
Lund, Sweden.
‡ Department of Biochemistry, Faculty of Science, Palacky´ University, Sˇlech-
titelu˚ 11, 78371 Olomouc, Czech Republic.
(1) Volpe, G.; Mascini, M. Talanta 1 9 9 6 , 43, 283-289.
(2) Gruger, E. J. Agric. Food Chem. 1 9 7 2 , 20, 781-784.
(3) Tombelli, S.; Mascini, M. Anal. Chim. Acta 1 9 9 8 , 358, 277-284.
(4) Male, K. B.; Bouvrette, P.; Luong, J. H. T.; Gibbs, B. F. J. Food Sci. 1 9 9 6 ,
61, 1012-1016.
(20) Xu, C. X.; Marzouk, S. A. M.; Cosofret, V. V.; Buck, R. P.; Neuman, M. R.;
Sprinkle, R. H. Talanta 1 9 9 7 , 44, 1625-1632.
(5) Bouvrette, P.; Male, K. B.; Luoung, J. H. T.; Gibbs, B. F. Enzymol. Microb.
(21) Matsumoto, T.; Suzuki, O.; Katsumata, Y.; Oya, M.; Suzuki, T.; Nimura, Y.;
Hattori, T. J. Cancer Res. Clin. Oncol. 1 9 8 1 , 100, 73-84.
(22) Stevanato, R.; Mondovi, B.; Sabatini, S.; Rigo, A. Anal. Chim. Acta 1 9 9 0 ,
273, 391-397.
Technol. 1 9 9 7 , 20, 32-38.
(6) Chemnitius, G. C.; Bilitewski, U. Sens. Actuators B 32, 1 9 9 6 , 107-113.
(7) Ohashi, M.; Nomura, F.; Suzuki, M.; Otsuka, M.; Adachi, O.; Arakawa, N.
J. Food. Sci. 1 9 9 4 , 59, 519-522.
(23) Rinaldi, A. C.; Sanjust, E.; Rescigno, A.; Finazziagro, A.; Rinaldi, A. Biochem.
(8) Yano, Y.; Yokoyama, K.; Tamiya, E.; Karube, I. Anal. Chim. Acta 1 9 9 6 ,
Mol. Biol. Int. 1 9 9 4 , 34, 699-704.
320, 269-276.
(24) Toul, Z.; Machola´n, L. Collect. Czech. Chem. Commun. 1 9 7 5 , 40, 2208-
2217.
(25) Hungerford, J. M.; Arefyev, A. A. Anal. Chim. Acta 1 9 9 2 , 261, 351-359.
(9) Chemnitius, G. C.; Suzuki, M.; Isobe, K. Anal. Chim. Acta 1 9 9 2 , 263, 93-
100.
10.1021/ac990848+ CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/04/2000
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1591

both in single10,27 and coupled enzyme-based designs.15,28 However,
most of the AO biosensors required a high applied potential (>500
mV vs Ag/ AgCl),5,27 which can lead to high background currents
and bias signals caused by interferants present in complex
matrixes to be analyzed.
Therefore, in this work, a bienzymatic approach based on a
recently isolated and purified amine oxidase from grass pea (AO)
and horseradish peroxidase (HRP) immobilized on solid graphite
has been considered, focusing on electrode designs which operate
at a low potential where biases from interferants are minimal.
Bienzyme electrodes were prepared by simply adsorbing the two
enzymes on the electrode surface using a direct electron-transfer
approach and by applying the principle of a mediated electron-
transfer cross-linking the enzymes and a redox polymer formed
of poly(1-vinylimidazole) complexed with [Os(4,4′-dimethylbipyri-
dine)2Cl]+/ 2+ (PVI13-dmeOs) using poly(ethylene glycol) diglycidyl
ether (PEGDGE) as the cross-linking agent. The optimal biosensor
design was evaluated in terms of sensitivity, lifetime, and selectiv-
ity, and it was used for the analysis of fish samples stored under
different conditions.
EXPERIMENTAL SECTION
Materials. Amine oxidase from grass pea (EC 1.4.3.6, AO)
was isolated and purified according to a previously published
protocol.29 Peroxidase from horseradish (EC 1.11.1.7, HRP) was
purchased from Sigma Chem. Co., St. Louis, MO (cat. no. P-6782)
as a lyophilized powder with a declared activity of 1100 U mg-1
solid. Histamine dihydrochloride (cat. no. 100340), putrescine
dihydrochloride (cat. no. 100450), tyramine hydrochloride (cat.
no. 103173), cystamine dihydrochloride (cat. no. 100492), agmatine
sulfate (cat. no. 100274), spermidine phosphate salt (cat. no.
102943) were from ICN Biochemicals Inc., Aurora, OH. Ethyl-
enediamine (cat. no. 800947) was from Merck, Darmstadt,
Germany. Cadaverine dihydrochloride (cat. no. C-8561) was
purchased from Sigma. Z- And E-2-butene-1,4-diamino dihydro-
chloride were synthesized according to previously published
protocols.30,31 PVI13-dmeOs was prepared by complexing poly(1-
vinylimidazole) with [osmium(4,4′-dimethylbipyridine)2Cl]+/ 2+, as
described elsewhere.32 Poly(ethylene glycol) (400) diglycidyl ether
(PEGDGE, Polysciences, Warrington, PA, cat. no. 08210) was used
for cross-linking AO and HRP to the osmium-complexed polycat-
ion. Disodium hydrogen phosphate dihydrate and potassium
dihydrogenphosphate, purchased from Merck (cat. no. 1.06580
and no. 1.04873, respectively), and water purified in a Milli-Q
system (Millipore, Bedford, MA) were utilized to prepare the
phosphate buffer 0.1 M, pH 7.2 (PB), used as the carrier solution
and supporting electrolyte in the flow injection experiments, if
not otherwise stated. Hydrogen peroxide, 35 wt % water solution
was from Across Organics, Geel, Belgium.
Figure 1. Preparation of AO-HRP bienzyme electrodes: type I
electrodes, AO + HRP (mixture); type II electrodes, AO + HRP +
PVI13-dmeOs + PEGDGE (mixture); Type III electrodes, (a) first
coating, HRP + PVI13-dmeOs + PEGDGE (mixture); second coating,
AO; (b) first coating, AO; second coating, HRP + PVI13-dmeOs +
PEGDGE (mixture); (c) first coating, HRP; second coating, AO
+ PVI13-dmeOs + PEGDGE (mixture); (d) first coating, AO +
PVI13-dmeOs + PEGDGE (mixture); second coating, HRP.
All experiments were performed at room temperature. All
solutions were daily prepared using PB as solvent, if not otherwise
stated, and were filtered through 0.45 µm (Millipore, Molshem,
France, type HA) filters and degassed before use.
Biosensor P reparation. Bienzyme graphite electrodes were
prepared as follows: first, rods of spectroscopic graphite (Rings-
dorff Werke GmbH., Bonn, Germany, type RW001, 3.05 mm
diameter) were cut and polished on a wet fine emery paper
(Tufback, Durite P1200, Allar, Sterling Heights, MI), followed by
rinsing the electrode surface with water and drying at room
temperature before coating with enzymes. Three different elec-
trode types were prepared (see Figure 1):
Type I electrodes were prepared by placing 6 µL of a premixed
solution containing various amounts of AO (stock 20 mg mL-1
)
and HRP (stock 10 mg mL-1) on the graphite electrode (direct
electron-transfer approach).
Type II electrodes were prepared using 6 µL of a mixture formed
of AO (same as above), HRP (same as above), PVI13-dmeOs (stock
10 mg mL-1), and PEGDGE (5 mg mL-1 water solution freshly
prepared and used within 15 min) in different w/ w (%) ratios
placed on the top of the graphite electrode (one-layer electrodes).
Type III electrodes were prepared using a sequential coating
procedure. The four components of the above-described mixture
were separated in two groups; 6 µL of the first one were initially
added on the graphite, and after drying, the second layer formed
by 6 µL of the other group of components was placed over the
first one (two layer electrodes), as follows:
(26) Alam, M. K.; Sasaki, M.; Watanabe, T.; Maeyama, K. Anal. Biochem. 1 9 9 5 ,
229, 26-34.
(27) Draisci, R.; Volpe, G.; Lucentini, L.; Cecilia, A.; Frederico, R.; Palleschi, G.
Food Chem. 1 9 9 8 , 62, 225-232.
(28) Gasparini, R.; Scarpa, M.; Vianello, F.; Mondovi, B.; Rigo, A. Anal. Chim.
Acta 1 9 9 4 , 294, 299-304.
(29) Sˇebela, M.; Luhova´, L.; Fre´bort, I.; Faulhammer, H. G.; Hirota, S.; Zajoncova´,
L.; Stuzˇka, V.; Pecˇ, P. Phytochem. Anal. 1 9 9 8 , 9, 211-222.
(30) Pecˇ, P.; Chudyˇ, J.; Machola´n, L. Biolo´gia (Bratislava) 1 9 9 1 , 46, 665-672.
(31) Machola´n, L.; Huba´lek, F.; Sˇubova´, H. Collect Czech. Chem. Commun. 1 9 7 5 ,
40, 1247-1265.
(32) Ohara, T. J.; Rajagopalan, R.; Heller, A. Anal. Chem. 1 9 9 4 , 66, 2451-2457.
1592 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

Type IIIa. First, 6 µL of a premixed solution of HRP, PVI13-
dmeOs, and PEGDGE was placed on the top of the electrode.
Next, the electrodes were dried for 1 h before coating with 6 µL
of AO.
Type IIIb. First 6 µL of AO was placed on the top of the
electrode. After drying for 1 h, 6 µL of a premixed solution of
HRP, PVI13-dmeOs, and PEGDGE was added.
Type IIIc. In the first step, 6 µL of HRP was placed on the top
of the electrode, and after drying, a second layer containing 6 µL
of a premixed solution of AO, PVI13-dmeOs, and PEGDGE was
added.
ever, the structure, selectivity, and biological functions are very
different, depending on the isolation source. The grass pea amine
oxidase,29 used during this work, is a copper-containing AO, which
besides the metal ions contains also an organic cofactor with a
quinoide structure (topa quinone) in its catalytic site,36 most of
the proposed reaction mechanisms being related to the structure
of this cofactor. However, in most of the developed methods for
monitoring of biomarkers using an amine oxidase, the enzyme is
converting the analyte to a corresponding aldehyde with NH3 and
H2O2 release, according to reaction 1:
RCH2NH2 + H2O + O2 f RCHO + H2O2 + NH3 (1)
Type IIId. First, a premixed solution formed of 6 µL of AO,
PVI13-dmeOs, and PEGDGE was placed on the top of the
electrode. Next, electrodes were dried for 1 h before coating with
6 µL of HRP.
The amount of all the components added on type III electrodes
was constant and the same as the optimum found for the type II
electrodes.
If not otherwise stated, all modified electrodes were stored at
4 °C for 14 h in a glass beaker and were rinsed with PB before
use. All results presented in this paper are means of at least three
equally prepared electrodes.
Both, the oxygen consumption7,37 and hydrogen peroxide
formation3,4 have been used as detection principles of biogenic
amines based on the above-mentioned reaction, usually requiring
high overvoltages. We recently presented a biosensor design that
is based on the same enzyme as the one used in this work38
displaying very good characteristics but still requiring a quite high
applied potential (+200 mV vs Ag/ AgCl). This aspect represents
a serious drawback when considering monitoring in complex
matrixes (e.g., food samples).
Instrumentation. The bienzyme-modified graphite electrodes
were inserted as the working electrode in a single channel flow
injection system33 containing a manual injection valve (Valco
Instruments Co. Inc., Houston, TX) with a 50 µL injection loop
and a three-electrode flow-through wall-jet type electrochemical
cell. A peristaltic pump (Alitea AB, Stockholm, Sweden) was used
to pump the carrier solution at desired flow rates through Teflon
tubings (0.5 mm i.d.) to the flow cell. A potentiostat (Z¨ata-
Electronik, Ho¨ o¨ r, Sweden) maintained the constant potential
between the working and the Ag/ AgCl (0.1 M KCl) reference
electrode. A platinum wire was used as the counter electrode. The
response current was monitored with a single channel strip-chart
recorder (Model BD 111, Kipp & Zonen, Delft, The Netherlands).
Operational stability experiments were made using an auto-
mated sample injection analyzer (Ismatec, Glattburg-Zu¨ rich,
Switzerland) by injecting samples of 100 µM histamine and 50
µM putrescine, respectively, with a sample throughput of 30
injections h-1, using PB as the carrier solution at a flow rate of
The combination of peroxidases with hydrogen peroxide-
producing oxidases for the development of amperometric biosen-
sors has been extensively used during the past years.39,40 The
possibility of using biosensors based on coupled enzymes using
a direct electron transfer between a peroxidase and an appropriate
electrode at low applied potentials (around 0 mV vs SCE)41 makes
these types of biosensors theoretically more suitable for applica-
tions in real matrixes. AO also has been previously coupled to
peroxidases,15,28 but the obtained biosensors worked at high
potential (+200 mV),28 being prone for biases when used in
complex matrixes. All these electrodes used, however, AOs from
different sources and, hence, with different characteristics.
The bienzyme approaches outlined in this work were based
on the direct coupling of AO and HRP either simply immobilized
on the electrode surface (type I) or cross-linked into a redox
hydrogel (types II and III). In all configurations, the amine oxidase
first converts the amine substrate (e.g., histamine) to an aldehyde
product, the active form of the enzyme being recovered by
oxidation of the organic cofactor in the presence of molecular
oxygen (see Figure 2). The hydrogen peroxide formed during
the first reaction is subsequently reduced to water by peroxidase;
its native form being regenerated either by direct reduction of its
heme cofactor on the electrode surface (see Figure 2a, type I
biosensors) or by receiving electrons from a mediator, maintained
in its reduced form by the potential applied on the graphite
electrode (see Figure 2b, type II biosensors).
0.5 mL min-1
.
Sample P reparation. The frozen fish-muscle samples (turbot,
Psetta maxima) were kindly provided by Dr. Gunilla O¨ nning,
Deparment of Applied Nutrition and Food Chemistry, Lund
University, Sweden. Triplets of 1.0-g fish samples kept under
different conditions were homogenized in 10 mL of PB. The
homogenate was centrifuged at 13000g for 60 min. at 4 °C. The
supernatant was separated and immediately analyzed by direct
injection into the flow system.
(36) Janes, S. M.; Mu, D.; Wemmer, D.; Smith, A. J.; Kaur, S.; Maltby, D.;
Burlingame, A. L.; Klinman, J. P. Science 1 9 9 0 , 248, 981-984.
(37) Karube, I.; Satoh, I.; Araki, Y.; Suzuki, S. Enzymol. Microb. Technol. 1 9 8 0 ,
2, 117-120.
RESULTS AND DISCUSSION
Amine oxidases represent a class of enzymes with a ubiquitous
distribution in mammals, plants, and microorganisms.34,35 How-
(38) Niculescu, M.; Fre´bort, I.; Pecˇ, P.; Galuszka, P.; Mattiasson, B.; Cso¨ regi, E.
In press.
(33) Appelqvist, R.; Marko-Varga, G.; Gorton, L.; Torstensson, A.; Johansson, G.
Anal. Chim. Acta 1 9 8 5 , 169, 237-242.
(39) Vijayakumar, A. R.; Cso¨ regi, E.; Heller, A.; Gorton, L. Anal. Chim. Acta 1996,
327, 223-234.
(34) Rinaldi, A.; Floris, G.; Giartosio, A. Structure and functions of amine oxidases;
(40) Ghobadi, S.; Cso¨ regi, E.; Marko-Varga, G.; Gorton, L. Curr. Separations
CRC Press: Boca Raton, FL, 1 9 8 5 .
1 9 9 6 , 14, 94-102.
(35) Knowles, P. F.; Dooley, D. M. Metal ions in biological systems; Marcel
Dekker: New York, 1994.
(41) Marko-Varga, G.; Emne´us, J.; Gorton, L.; Ruzgas, T. Trends Anal. Chem.
1 9 9 5 , 14, 319-328.
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1593

Figure 2. Scheme of the electron-transfer pathways using histamine as a model substrate for (a) direct electron transfer and (b) mediated
electron transfer. Symbols: AOox, AOred, HRPox, HRPred are the oxidized and reduced forms of the mentioned enzymes, respectively; TOPA-
inactive and TOPA-native represent trihydroxyphenylalanine and its quinonic form, being the reduced and oxidized states of the AO cofactor.
Figure 3. Hydrodynamic voltammogram recorded for type I elec-
trodes using 100 µM histamine (b), the background current (O), and
their ratio (‚), respectively. Conditions: electrode structure, AO:HRP
Figure 4. Effect of flow rate on current signals (0) and sample
1:1 (w/w); applied potenitial, -50 mV vs Ag/AgCl; flow rate, 0.5 mL
throughput (9) of type I electrodes. Conditions: 100 µM histamine;
min-1
.
other parameters as mentioned in Figure 3.
Optimization Steps. The bienzyme electrodes were optimized
with regard to several parameters, namely, working potential, flow
rate, influence of various enzyme, polymer, cross-linker ratios, and
electrode coating procedure.
chosen for all further experiments as a compromise between the
observed response and the background currents.
The influence of carrier flow rate on the biosensor response
for histamine was also considered for type I electrodes, the results
being presented in Figure 4. The decrease in peak height with
increasing flow rates demonstrates a limitation either due to the
bioconversion of the amine substrate by AO or to the reduction
of H2O2 by the direct electron transfer between HRP and graphite
electrode. The phenomenon is not clarified yet, since the later
reaction is known to be a kinetically slow process,42 but also our
recent study has indicated that the reaction catalyzed by AO is
Hydrodynamic voltammograms were recorded using 100 µM
histamine as substrate and using AO-HRP-modified type I
electrodes in order to establish the optimal working potential. The
voltammogram, together with the ratio between the response and
the background current obtained on the same condition, respec-
tively, is shown in Figure 3. Although the response of the
biosensor drastically increased when the applied potential was
below -100 mV, so did the background current, demonstrating
a possible oxygen electroreduction interference with the biosens-
ing process. A potential of -50 mV vs Ag/ AgCl was therefore
(42) Ruzgas, T.; Cso¨ regi, E.; Emne´us, J.; Gorton, L.; Marko-Varga, G. Anal. Chim.
Acta 1 9 9 6 , 330, 123-138.
1594 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

Table 1. Biosensor Characteristics of Type I Electrodes Obtained for Different AO:HRP Ratios (w/w%)a
enzyme ratio
analyte
Kmapp (µM)
Imax (µA)
S (mA/ M cm2)
C (%)
DL (µM)
AO 87%
HRP 13%
histamine
putrescine
H2O2
histamine
putrescine
H2O2
histamine
putrescine
H2O2
histamine
putrescine
H2O2
histamine
putrescine
H2O2
histamine
putrescine
H2O2
279 ( 16
153 ( 15
93 ( 3
332 ( 17
228 ( 15
112 ( 8
370 ( 22
240 ( 15
153 ( 6
437 ( 43
268 ( 23
175 ( 8
441 ( 23
276 ( 22
206 ( 3
479 ( 41
287 ( 12
211 ( 18
1.03 ( 0.02
1.96 ( 0.06
1.80 ( 0.21
1.34 ( 0.03
3.01 ( 0.07
2.70 ( 0.06
1.30 ( 0.03
3.10 ( 0.01
3.64 ( 0.04
1.22 ( 0.04
3.05 ( 0.10
3.83 ( 0.05
1.16 ( 0.02
3.69 ( 0.06
4.94 ( 0.03
1.37 ( 0.10
3.84 ( 0.06
4.95 ( 0.15
50.6 ( 0.8
175.5 ( 1.4
265.1 ( 1.6
55.3 ( 0.7
180.8 ( 0.9
330.2 ( 1.0
48.1 ( 0.1
176.9 ( 0.9
325.9 ( 0.5
38.2 ( 1.4
155.9 ( 1.3
299.8 ( 0.6
36.0 ( 0.7
183.1 ( 1.1
328.5 ( 0.2
39.2 ( 1.5
183.3 ( 0.6
321.4 ( 1.2
19.0
66.2
0.16
0.06
AO 80%
HRP 20%
16.6
54.7
0.20
0.07
AO 67%
HRP 33%
14.7
54.2
0.25
0.07
AO 50%
HRP 50%
12.7
52.0
0.33
0.08
AO 40%
HRP 60%
10.9
55.7
0.34
0.13
AO 33%
HRP 67%
12.2
57.0
0.41
0.08
Imax and Km values were estimated from the Michaelis-Menten equation: I ) (Imax×[A])/ (Kmapp+[A]). S is the sensitivity (calculated as
a
app
Imax/ Kmapps, where s is the active area of the electrode), C is the conversion efficiency (calculated as Sanalyte/ SH O ), and DL is the detection limit
(calculated as 3S/ N). The linear range for all studied configurations was 1-100 µM.
2
2
Table 2. Bioelectrochemical Characteristics Obtained for Bienzyme Electrodes Based on Direct and Mediated ETa
electrode type
type I
analyte
Kmapp (µM)
Imax (µA)
S (mA/ M cm2)
C (%)
DL (µM)
LR (µM)
histamine
putrescine
H2O2
histamine
putrescine
H2O2
332 ( 17
227 ( 16
112 ( 8
901 ( 85
512 ( 40
977 ( 92
1.34 ( 0.02
3.01 ( 0.07
2.70 ( 0.06
4.85 ( 0.41
7.26 ( 0.53
22.8 ( 1.68
55.3 ( 0.7
181.6 ( 1.0
330.1 ( 1.0
73.74 ( 1.7
194.1 ( 1.4
319.6 ( 1.6
16.74
55.01
0.16
0.06
1-100
1-100
1-100
1-150
1-400
1-250
type II
23.07
60.73
0.33
0.17
a
LR is linear range and the other symbols are as in Table 1.
the rate-limiting one.38 According to the obtained results an optimal
working flow rate was chosen to be 0.5 mL min-1, as a compromise
between the biosensor kinetics and its sample throughput.
To achieve an effective electron transfer (ET) all electrode
types were optimized with regard to the ratio of the used enzymes
(type I), composition of the redox hydrogel (type II), and influence
of electrode coating procedures (types IIIa-d). Table 1 shows
the kinetic parameters and the main biosensor characteristics
obtained for type I electrodes at different ratios of AO:HRP. The
increasing tendency of the apparent Michaelis constant ob-
served for type I electrodes with the increasing amount of
immobilized HRP was attributed to an increase in the thickness
of the total protein loading on the electrode surface, decreasing
the diffusion rate of the analytes into the film. The conversion
efficiency of type I biosensors (defined as the ratio between the
sensitivities for amines and H2O2, respectively) decreased with
decreasing AO content, showing a catalytic limitation of the AO-
catalyzed reaction and a first-order kinetics for the amine
substrates within the dynamic range. As a compromise between
response time, sensitivity, and detection limit for the two sub-
strates, the electrodes containing 20% HRP and 80% AO were
chosen as optimal for further experiments. The dynamic range
for all type I biosensors was 1-100 µM for both histamine and
putrescine.
electrode design was considered that integrated the two enzymes
into an Os-based redox polymer. Redox hydrogels have been
previously shown to represent an effective matrix for enzyme
immobilization, resulting in increased stability and enhanced rate
of the ET.32,43,44 It is known, however, that the rate of the ET is
highly influenced by the composition of the redox hydrogel, as
well as by the kinetics of the used enzymes. Therefore various
biosensor preparations (types II and III) were considered in order
to find the optimal electrode structure displaying the most efficient
ET pathway.
The redox hydrogel based biosensors were first examined in
order to determine the influence of the amount of redox polycation
and the cross-linking agent in the biosensor’s structure. Table 2
shows the obtained results. Increasing the number of components
on the electrode surface resulted in an increased diffusional
barrier, a tendency reflected in the apparent Michaelis constants
(3-fold increase for histamine and 2-fold increase for putrescine).
However, the introduction of the electrochemical mediator
caused a considerable improvement in bioelectrocatalytic ef-
ficiency, indicated by an increased Imax (3-fold increase for
histamine and 1,5-fold for putrescine) and sensitivity (30% and 7%
increase for histamine and putrescine, respectively). The hydrogen
(43) Cso¨ regi, E.; Schmidtke, D. W.; Heller, A. Anal. Chem. 1 9 9 5 , 66, 2451-
2457.
(44) Larsson, N.; Ruzgas, T.; Gorton, L.; Kokaia, M.; Kissinger, P. T.; Cso¨ regi,
E. Electrochim. Acta 1 9 9 8 , 43, 3541-3554.
To improve the electron transfer kinetics between HRP and
the graphite electrode, as well as the biosensor stability, a new
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1595

Table 3. Influence of the Electrode Coating Procedure
on the Biosensor Characteristicsa
type of
electrode
analyte
Kmapp (µM)
Imax (µA)
S (mA/ M cm2)
type II
histamine
putrescine
901 ( 85
512 ( 40
789 ( 35
449 ( 34
687 ( 47
473 ( 28
689 ( 33
422 ( 35
649 ( 19
425 ( 24
4.85 ( 0.41
7.26 ( 0.53
3.56 ( 0.08
7.72 ( 0.69
2.66 ( 0.24
2.04 ( 0.13
2.17 ( 0.06
7.83 ( 0.82
1.90 ( 0.02
2.14 ( 0.20
73.7 ( 1.7
194.1 ( 1.4
61.8 ( 0.7
235.5 ( 1.6
53.0 ( 1.5
59.1 ( 1.2
43.1 ( 0.7
254.2 ( 1.8
40.1 ( 0.4
69.0 ( 1.5
type IIIa histamine
putrescine
type IIIb histamine
putrescine
type IIIc histamine
putrescine
type IIId histamine
putrescine
a
Symbols are as in Table I.
Figure 5. Relative selectivity recorded for type I (white) and type II
(black) electrodes for different substrates using histamine signals
obtained for type I electrodes as reference. Symbols: His, histamine;
Cys, cystamine; Tyr, tyramine; Spr, spermidine; EDA, ethylenedi-
amine; Agm, agmatine; Put, putrescine; Cad, cadaverine; Z,E-Ab,
Z,E-1,4-diamino-2-butene.
peroxide sensitivity remained practically unchanged, confirming
previously reported results for other bienzyme hydrogel elec-
trodes.32 Type II electrodes displayed improved detection limit
and linear range for the studied analytes. Therefore, this design
(type II electrodes) was chosen for further experiments also due
to improved electrode stability (results not shown).
The effect of the coating procedure was also studied. Besides
coating with a premixed solution of all four components (type II),
different possibilities of sequential coating of the electrode surface
(type III) were considered (see Table 3). It was previously
demonstrated that both HRP and AO can be electrically wired to
the redox polymer and, thus, might cause a partial short circuit32,38
when all components are mixed together. This assumption was
confirmed for the main substrate (putrescine) for which an
increase in sensitivity of about 30% was observed for the two-layer
electrodes (type III), compared to the single-layer ones (type II).
Comparing type III electrodes, it seems confirmed that IIIb type
electrodes represent the worse structure for the diffusion of the
substrates, resulting in the highest Kmapp, a value similar to that
obtained for type II electrodes. Signals recorded for type IIIb and
IIId electrodes were about 30% less than those recorded for types
IIIa and IIIc, confirming partly that the ET is hindered when the
HRP layer (wired or not) is not in direct contact with the electrode
surface. Some intermixing of the layers was observed, as previ-
ously demonstrated for other enzymes such as glucose oxidase
and lactate oxidase.45 No considerable change was observed for
the other substrate, histamine, the slight decrease in sensitivity
being unrepresentative considering the differences of about 10-
15% in similar electrode preparation. However, a certain interaction
between histamine and horseradish peroxidase could be observed,
causing a decreased sensitivity (see electrode type IIId) compared
to type II electrodes. More studies have to be done to fully
elucidate the mechanism of the ET observed for the two
substrates. Clearly, the less sensitive electrode configuration is
represented by type IIId electrodes, for which the bias currents
due to the wiring of AO are the most explicit. Considering the
simplicity of electrode preparation and the small differences in
the electrode characteristics between type II and type III elec-
trodes, type II was chosen as the optimal electrode design.
Figure 6. Monitoring of fish freshness using type II electrodes. Total
amine content recorded for fish kept at 4 °C (O) and at 25 °C (9),
respectively. The total amine concentration was expressed in hista-
mine equivalent units.
Characteristics and Applications. Type II biosensors were
further characterized with regard to selectivity, response time,
and operational and storage stability. Figure 5 shows the relative
selectivity for different AO substrates, using histamine as the
reference compound, since it is considered to be a biomarker of
major interest. As seen, the response for aliphatic amines is
generally higher than those observed for the aromatic ones. Also,
type II biosensors yielded higher sensitivities than type I ones,
probably caused by better electron-transfer kinetics.
The response time of the sensor, calculated as the time elapsed
between 5% and 95% of response height, was fast (less than 1 min).
The operational stability of the biosensor was studied both for
histamine and putrescine as substrates. The response current of
the bienzymatic enzyme electrodes decreased with about 30% and
50% for histamine and putrescine, respectively, after 10 h of
continuous operation with a sample throughput of 30 injections
h-1. The storage stability of the electrodes was good, a decrease
of only about 10% and 15% being observed for histamine and
putrescine, respectively, after 10 days of storage.
(45) Cso¨ regi, E.; Quinn, C. P.; Schmidtke, D. W.; Lindquist, S. E.; Pishko, M. V.;
Ye, L.; Katakis, I.; Hubbell, J. A.; Heller, A. Anal. Chem. 1 9 9 4 , 66, 3131-
3138.
1596 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

The optimized biosensor was considered for monitoring
biogenic amines in real samples. The differentiation between the
signals given by different amines is, however, not possible, only
the total amine content of the sample was determinable. Fish-
muscle samples, kept for 10 days at 4 and 25 °C, respectively,
were analyzed after extraction in PB by direct injection in the flow
system. The total amine content expressed in histamine equiva-
lents is presented in Figure 6. The maximum accepted limit for
total amine concentration in food products is 100-200 mg kg-1
sample, and a concentration of 1000 mg kg-1 is considered to be
toxic.6 After 3 days of storage at room temperature, the fish
became improper for consumption, while even after 10 days of
storage at 4 °C no major changes could be observed in the total
amine concentration.
electrodes were characterized by high sensitivity, good operational
and storage stability, fast response time, and low detection limit,
making them very promising for food quality assessing. As
previously demonstrated, a biosensor with only AO immobilized
on a graphite electrode is selective for histamine, cystamine, and
tyramine but not for cadaverine and putrescine, while the bien-
zyme sensors described in this work are selective for all of them.
Thus, using a combination of the two developed biosensors the
separation of the signals given by the mentioned biogenic key
amines is possible.
ACKNOWLEDGMENT
This work was financially supported by the European Comis-
sion (INCO-COPERNICUS, ERBIC 15CT96 1008), the Swedish
Institute (Mihaela Niculescu), the Swedish Council for Forestry
and Agricultural Research (SJSF), the Swedish National Board
for Industrial and Technical Development (NUTEK), and the
Grant Agency (203/ 97/ 0097) Czech Republic.
CONCLUSIONS
The present work shows the development, optimizations, and
possible application of bienzyme electrodes for the determination
of biogenic amines. The developed electrodes are based on an
amine oxidase, which is readily producible with high yield, and
were based either on a direct or on a mediated electron-transfer
approach. The optimized redox hydrogel incorporated bienzyme
Received for review July 29, 1999. Accepted January 3,
2000.
AC990848+
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1597