Analytica Chimica Acta 591 (2007) 195–199

An amperometric glucose biosensor constructed by immobilizing

glucose oxidase on titanium-containing mesoporous composite
material of no. 41 modified screen-printed electrodes
Zhihui Dai a,∗ , Min Fang a , Jianchun Bao a , Huaisheng Wang b , Tianhong Lu a,∗
a Department of Chemistry, Laboratory of Materials Science, Nanjing Normal University, Nanjing 210097, PR China
b Department of Chemistry, Liaocheng University, Liaocheng 252059, PR China

Received 20 December 2006; received in revised form 15 March 2007; accepted 27 March 2007
Available online 4 April 2007

Abstract
We have constructed a glucose biosensor by immobilizing glucose oxidase (GOD) on titanium-containing MCM-41 (Ti-MCM-41) modified
screen-printed electrodes. The strategy of the sensing method is to monitor the extent of the decrease of the reduction current of O2 upon adding
glucose at a selected potential. The detection can be done at the applied potential of −0.50 V and can efficiently exclude the interference from
commonly coexisted substances. The constructed sensor has a high sensitivity to glucose (5.4 mAM−1 cm−2 ) and a linear response range of
0.10–10.0 mM. The detection limit is 0.04 mM at a signal-to-noise ratio of 3. The sensor also shows high stability and remains its catalytic activity
up to 60 ◦ C. The biocompatibility of Ti-MCM-41 means that this immobilization matrix not only can be used for immobilizing GOD but also can
be extended to other enzymes and bioactive molecules, thus providing a promising platform for the development of biosensors.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Titanium-containing mesoporous composite material of no. 41; Glucose oxidase; Glucose; Biosensor

1. Introduction H2 O2 is detected at a relatively high anodic potential (>+0.6 V

The determination of glucose concentration is very impor-
tant in biology, chemistry, food processing and fermentation, as
well as in clinic for diagnosing diabetics [1–3]. Amperomet-
ric biosensors based on electron transfer between an electrode
and immobilized glucose oxidase (GOD) are especially promis-
ing because of their practical advantages such as operational
simplicity, low fabrication cost and suitability for real time detec-
tion. GOD consists two identical polypeptide chains. Each chain
contains a flavin adenine dinucleotide (FAD) redox center, which
is deeply embedded in the apoenzyme. During the last 20 years,
numerous attempts have been made to promote electron transfer
between the enzyme and electrode, and some glucose biosensors
were obtained [4–8]. Most of the electrochemical amperomet-
ric biosensors for glucose are based on measuring the increase
of the anodic current due to H2 O2 oxidation [9]. Oxidation of
∗ Corresponding authors. Tel.: +86 25 83598031; fax: +86 25 83243286.
E-mail addresses: daizhihuii@njnu.edu.cn (Z. Dai),
lutianhong@njnu.edu.cn (T. Lu).
versus SCE) [10]. But at this potential, there are interferences
from other oxidizable species such as ascorbic acid (AA), uric
acid (UA) and acetaminophen (AP). It results in difficulty for the
quantitative analysis of the produced H2 O2 and the enzyme sub-
strate. General methods used to suppress the interferences have
involved utilizing electron mediators to reduce the overpotential
of electrochemical oxidation of H2 O2 to prevent the diffusion
of interfering species [7,8]. But the stability and the toxicity of
the mediators limited their in vivo applications. Here, we choose
the method of detecting the decrease of the cathodic current due
to O2 reduction, which can suppress the interferences greatly.
Recently, the unique structure and catalytic properties of
porous materials have attracted considerable attention [11] for
their resistance to biodegradation and ability of building electro-
chemical/electron transfer environments. Porous materials have
large specific surface areas; high mechanical, thermal, and chem-
ical stabilities; as well as good adsorption and penetrability;
and can incorporate with proteins through physical or chemi-
cal method, providing active biomaterials [12]. Porous materials
with the appropriate dimension and functionalization adjacent to
the enzyme redox center can act as current nanocollectors and as

0003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2007.03.063
196 Z. Dai et al. / Analytica Chimica Acta 591 (2007) 195–199
electron relays to electrodes. They provide environments similar
to those of redox proteins in native systems and give the protein
molecules more freedom in orientation. A series of porous mate-
rials such as micromachined silicon [13], titania sol–gel [14],
gold nanoparticles microporous [15] and electrically conductive
polypyrrole composite [16] have been proven to be promising
as the immobilization matrices and applied in constructing glu-
cose biosensors. Here, we use the novel mesoporous material,
Ti-MCM-41 as the immobilized matrix to facilitate the electron
transfer and reduce the insulating property of the proteins.
Furthermore, screen-print electrodes (SPE) have been used
to produce disposable electrochemical sensors. The technology
for the preparation of SPE is simple, inexpensive, versatile, and
mass-produced [17]. The screen-printed biosensor challenges
the conventional electrochemical biosensors for disposablil-
ity and portability. Therefore, it has been widely used in the
determination of biomolecules [18], pesticides [19], antigens
[20], and nucleic acid [21]. Herein, we report the applica-
tion of Ti-MCM-41 to GOD immobilization, and biosensing
on SPE. The constructed sensor has a high sensitivity to glu-
cose (5.4 mAM−1 cm−2 ) with the linear response range of
0.10–10.0 mM.
2. Experimental
2.1. Materials
Glucose oxidase (GOD, EC 1.1.3.4, 35.3 units mg−1 . Type II
from Aspergillus niger) and ␤-d-(+)-glucose were purchased
from Sigma and used as received. Polyvinyl alcohol (PVA,
average degree of polymerization, 1800 ± 100) was purchased
from Shanghai Laize Factory of Fine Chemicals. Tetrabutyl
titanate was purchased from Shanghai Reagent Factory. Phos-
phate buffer solutions (PBS, 0.1 M) with various pH values were
prepared by mixing stock standard solutions of K2 HPO4 and
KH2 PO4 and adjusting the pH with H3 PO4 or NaOH. All other
chemicals were of analytical grade and were used without fur-
ther purification. All solutions were made up with twice distilled
water.
2.2. Instrumentation
Cyclic voltammetry were performed on CHI 660 electro-
chemical workstation (CH Instruments, USA). All electrochem-
ical experiments were carried out in a cell containing 5.0 mL
0.1 M, pH 6.2 PBS at room temperature (25 ± 2 ◦ C) and using the
modified SPE as working electrodes. The experiments for glu-
cose determinations and electrocatalytical studies were carried
out in air-saturated solutions. Energy dispersive spectroscopy
(EDS) was performed on Thermo Noran analyzer.
2.3. Procedures
The SPE was fabricated according to the steps reported pre-
viously [20]. Firstly, the silver ink was screen-printed on a
nylon sheet as conductive bands. Then the graphite ink was
imprinted to cover the area that served as the working and aux-
Table 1
Physical properties of Ti-MCM-41
ABET (m2 g−1 ) d100 (nm) a0 (nm) D (nm) L (nm) Si:Ti
1050 3.84 4.43 2.81 1.62 1:0.01
ABET , total specific surface area; d100, XRDd100 space; a0 , lattice parameter; D,
mesopore diameter; L, wall thickness; Si:Ti, the atom ratio of Si to Ti.
iliary electrode. After that the silver chloride ink was applied
onto the area of reference electrode. Finally, the conductive
bands were insulated by overlaying an insulating dielectric
material except for the area of electric connectors and the three-
electrodes.
Ti-MCM-41 was prepared following a recipe similar to that
reported by Tatsumi [22]. The specific surface area was cal-
culated based on N2 adsorption data according to the BET
method of Barrett et al. [23]. The diameter of the Ti-MCM-
41 was derived from the adsorption branch using the BJH
method. The physical properties of the prepared Ti-MCM-41
were listed in Table 1. 30 mg Ti-MCM-41 was dispersed into
10 mL water to obtain a suspension of 3 mg mL−1 Ti-MCM-41.
Three microliters of 3 mg mL−1 Ti-MCM-41, 3 ␮L of 0.1 mM
GOD (prepared in pH 6.2 PBS) and 3 ␮L of 3% PVA solu-
tion of ethanol/water (1:1, v:v) were successively dispersed on
SPE surface, and allowed to dry under ambient condition for
3 h. After the modified electrode was rinsed with doubly dis-
tilled water twice or thrice, the GOD/Ti-MCM-41 modified SPE
was obtained. Ti-MCM-41, GOD/MCM-41 and GOD modi-
fied SPEs were prepared following the recipe similar to that of
GOD/Ti-MCM-41 modified SPE. When not in use the electrodes
were stored in 0.1 M pH 6.2 PBS at 4 ◦ C.
3. Results and discussion
3.1. The immobilization of GOD on Ti-MCM-41 modified
SPE
EDS shows that there is no N element in the Ti-MCM-41 film
while N element appears on GOD/Ti-MCM-41 film which indi-
cates GOD is indeed immobilized in Ti-MCM-41. The big pore
size of the Ti-MCM-41 film makes the immobilized GOD easy
to be accessed by its substrate and brings out a good performance
of the modified electrode.
The interaction between GOD and Ti-MCM-41 might be
electrostatic through hydrogen bonding and hydrophilic attrac-
tion [24,25] and it is much stronger than that between GOD and
GCE due to the presence of SiOH groups on the external surface
of Ti-MCM-41 [26].
3.2. The determination of glucose
Upon addition of ␤-d(+)-glucose to the air-saturated PBS,
the reduction current of O2 at GOD/Ti-MCM-41 modified SPE
decreased (Fig. 1B), while no obvious change was observed
at Ti-MCM-41 modified SPE (Fig. 1A). The cyclic voltam-
mograms of GOD/Ti-MCM-41 modified SPE with successive
addition of ␤-d(+)-glucose to air-saturated 0.1 M, pH 6.2 PBS
 Z. Dai et al. / Analytica Chimica Acta 591 (2007) 195–199
Fig. 1. Cyclic voltammograms of the Ti-MCM-41 (A) and GOD/Ti-MCM-41 (B) modified SPEs in air-saturated 0.1M pH 6.2 PBS in the absence (a and c) and
presence of 2.0 mM glucose (b and d) at a scan rate of 50 mV s−1 .
were shown in Fig. 2. With increasing ␤-d(+)-glucose con-
centration the reduction current decreased. The amperometric
responses of the GOD/Ti-MCM-41 modified SPE upon succes-
sive additions of 1.0 mM glucose to 0.1 M air-saturated pH 6.2
PBS at an applied potential of −0.5 V were shown in Fig. 3.
Upon addition of an aliquot of glucose to the buffer solution,
the reduced current decreased steeply to reach a stable value.
The enzyme electrode achieved 95% of the steady-state-current
in less than 10 s. The results demonstrated clearly that the elec-
trocatalytic response was very fast, which could be used as an
efficient sensor for glucose detection.
Fig. 2. Cyclic voltammograms of the GOD/Ti-MCM-41 modified SPE in air-
saturated 0.1 M pH 6.2 PBS containing 0,1.0 and 3.0 mM glucose (from top to
bottom) at a scan rate of 50 mV s−1 .
Fig. 3. Amperometric responses of GOD/Ti-MCM-41 modified SPE at −0.50 V
upon successive additions of 5 ␮L 1.0 M glucose to 5.0 mL air-saturated 0.1 M
pH 6.2 PBS. Inset: plot of the reducing peak current of O2 vs. glucose concen-
tration at −0.50 V.
197
The calibration curve of the sensor under the optimized
experimental conditions was shown as inset in Fig. 3. The cal-
ibration range of ␤-d(+)-glucose concentration is from 0.10 to
15.6 mM. The linear response range of the sensor to ␤-D(+)-
glucose concentration is from 0.10 to 10.0 mM (correlation
coefficient 0.9992) with a sensitivity of 5.4 mAM−1 cm−2 glu-
cose and a detection limit is 0.04 mM at a signal-to-noise
ratio of 3. It is well known that the diabetic glucose con-
centration is above 7.0 mM. Compared with the linear range
of 0.04–0.28 mM for GOD immobilized on a colloidal gold
modified carbon paste electrode [27], 0–7.8 mM for GOD
immobilized on a carbon nanotubes/chitosan matrix [8] and
0–4.75 mM for immobilizing GOD through silica sol–gel film
onto Prussian Blue modified electrode [28], the linear range of
our prepared sensor are much wider and more suitable for diag-
nosing diabetics. The detection limit is 0.04 mM which is similar
to 0.05 mM for GOD immobilized on carbon nanotube materials
[29].
For comparison, we also prepared the GOD/MCM-41 modi-
fied SPE and GOD modified SPE and studied their responses to
glucose. Results are given in Fig. 4. Upon addition of ␤-d(+)-
glucose to air-saturated PBS, the reduction current of O2 also
decreases, but the decrease is much smaller than that of GOD/Ti-
MCM-41 modified SPE (Fig. 4B). The linear response range of
GOD/MCM-41 modified SPE to ␤-d(+)-glucose concentration
is from 0.21 to 2.40 mM with a correlation of 0.9983 which indi-
cates the narrower linear range. GOD immobilized on PVA has
virtually no response towards glucose (Fig. 4A).
Based on the above results, the sensing mechanism is
proposed as follows. When ␤-d(+)-glucose is added to the
air-saturated solution, it reacts with the oxidized form of
GOD, GOD(FAD), immobilized on Ti-MCM-41 nearly quanti-
tatively according to Eq. (1), giving the reduced form of GOD,
GOD(FADH2 ). GOD(FADH2 ) will then reduce the dissolved
oxygen quantitatively, giving H2 O2 according to Eq. (2), caus-
ing the decrease of the concentration of O2 . Therefore, the added
glucose concentration is proportional to the decreasing extent
of the reducing peak current of O2 . Similar results have been
reported previously [30,31].
Glucose + GOD(FAD) → gluconolactone + GOD(FADH2 )
(1)
GOD(FADH2 ) + O2 → GOD(FAD) + H2 O2 (2)
198 Z. Dai et al. / Analytica Chimica Acta 591 (2007) 195–199
Fig. 4. Cyclic voltammograms of the GOD (A) and the GOD/MCM-41 (B) modified SPEs in air-saturated 0.1 M pH 6.2 PBS in the absence (a, c) and presence of
2.0 mM glucose (b, d) at a scan rate of 50 mV s−1 .
GOD immobilized on PVA has no reactivity with glucose.
GOD immobilized on MCM-41 has less reactivity towards glu-
cose, thus the decrease of the reduction current of O2 is less
dramatic. These results indicate that mesoporous structure of
MCM-41 provides a microenvironment similar to that of redox
proteins in native systems and causes the immobilized GOD to
react with glucose while Ti-MCM-41 has a better response than
that of MCM-41. TiO2 is semiconductor whose band gap of
3.2 eV [32] is much smaller than that of SiO2 of 8.0 eV [33] and
has a better conductivity than that of SiO2 . Ti-MCM-41 which
contains Ti in tetrahedral coordination with oxygen is likely also
has a smaller band gap than that of MCM-41, and can promote
the electron transfer between the immobilized GOD and the
electrode. Another possibility is probably due to the different
adsorption content of GOD. The isoelectric point (PI) of GOD
is 4.2 [34]. Thus, GOD is negative at pH 6.2. Ti-MCM-41 might
be positive in pH < 7 and adsorb more GOD molecules than
MCM-41, resulting in its better response to glucose. The third
explanation was attributed to acceleration of the proton-transfer
step by the metal oxide component of the composite [35].
3.3. Factors influencing the sensor response to glucose
Fig. 5 illustrates the effect of pH value of PBS solution on
the reduction current of 2.0 mM glucose at the GOD/Ti-MCM-
41 modified SPE. With an increasing solution pH from 5.1 to
Fig. 5. Effect of pH on the reducing peak current of O2 in 0.1 M PBS containing
2.0 mM glucose at −0.50 V.
8.3 the reduction current first increases to the maximum at pH
6.2, then decreases. At pH values less than 5.1, the reduction
current decreases steeply which might result from the protein
denaturation. Therefore, we select pH 6.2 PBS as the supporting
electrolyte for the determination of glucose.
The effect of applied potential on the responses of the glu-
cose sensor showed the electrocatalytic reduction of O2 could be
observed at around −0.45 V. Upon a decreasing applied potential
from −0.45 to −0.70 V, the steady-state-current increased due
to the increased driving force for the fast reduction of glucose
at the lower potentials. Here, we select −0.50 V as the working
potential for determinations.
The effect of the scan rate was also evaluated. We selected
different scan rates of 50, 100, 150, 200 mV s−1 to demonstrate
the effect of the scan rate on the response of the sensor. The peak
current was found to be proportional to the square root of the
scan rates, indicating a diffusion-controlled process. That meant
that glucose diffused from solution to the electrode surface.
Temperature is another important parameter affecting the
electrocatalytic activity of protein. Fig. 6 shows the effect of
temperature on sensor response. With an increasing temperature
from 15 to 40 ◦ C the electrocatalytic activity of the immobi-
lized GOD increases. The immobilized GOD has activity even
at 60 ◦ C. It is evident that the immobilized GOD has a good ther-
mal stability. These results indicate that this sensor can handle
in a wide range of temperature.
Fig. 6. Effect of temperature on the reducing peak current of O2 in 0.1 M pH
6.2 PBS containing 2.0 mM glucose at −0.50 V.
 Z. Dai et al. / Analytica Chimica Acta 591 (2007) 195–199 199

3.4. Interference
The influences of foreign species were investigated by
analyzing a standard solution of 2.0 mM glucose to which
interfering species were added. 0.3 mM uric acid and 0.3 mM
p-acetaminophenol did not cause any observable interference to
the response to glucose, and only ascorbic acid at the concentra-
tion of 0.3 mM reduced the reduction current by 2.1%. This was
probably due to the smaller size of ascorbic acid, which could
diffuse through the porous film to the electrode surface and be
oxidized. Thus, the response was not interfered by cooxidizable
substances such as ascorbic acid, uric and acetaminophen.
implies that it not only can be used to immobilize GOD but also
can be used as immobilizing matrices for other enzymes and
bioactive molecules, thus providing a promising platform for
the development of biosensors.
Acknowledgments
We gratefully acknowledge the financial support of the
National Natural Science Foundation of China (Nos. 20505010,
20473038, 20475024), and by the Natural Science Founda-
tion of the Education Committee of Jiangsu province (No.
04KJB150066).

References
3.5. Stability of the GOD/Ti-MCM-41 modified SPE
[1]
When the enzyme electrode was stored at 4 ◦ C, it retained [2]
96% of its initial current response for glucose after intermitted [3]
[4]
use over a 10-day period. Thus, the presence of Ti-MCM-41
was very efficient for retaining the enzyme activity of GOD. [5]
The sensor could keep a constant current when successively
swept for 100 cycles in the presence of glucose. The fabrication [6]
reproducibility of six electrodes, made independently, showed
an acceptable reproducibility with a relative S.D. of 5.2% for the [7]
current determined at 2.0 mM ␤-d(+)-glucose concentration. [8]
3.6. Determination of glucose in serum sample [9]
[10]
The sample was diluted to its half concentration by mixing [11]
it with 0.1 M, pH 6.2 PBS. Five parallel determinations were [12]
carried out. The glucose level was determined to be 7.35 mM
close to the 7.54 mM determined by spectrophotometry, showing [13]
a good accuracy. The recoveries for the assays of 1.0–8.0 mM [14]
glucose were between 98% and 101% for 10 measurements. [15]
[16]
4. Conclusions [17]
In this work, we have constructed a glucose biosensor which [18]
can operate under air by immobilizing GOD on Ti-MCM-41 on
[19]
the SPE. The immobilized GOD can catalyze the reduction of O2
and resulted in a high reduction current in the cyclic voltammo- [20]
gram. Upon adding glucose, the reducing peak current decreases [21]
proportionally. The strategy of the sensing method is to mon-
[22]
itor the extent of the decrease of the reduction current upon
[23]
adding glucose at a selected potential. The proposed biosensor [24]
can decrease the applied potential and efficiently exclude the [25]
interference of commonly coexisted substances. [26]
Ti-MCM-41 is demonstrated as a suitable kind of mate- [27]
[28]
rial for GOD loading and can promote the electron transfer
[29]
between GOD and electrodes. The immobilized GOD main- [30]
tains its bioactivity up to 60 ◦ C. The resulting sensor displays
low detection limit, low interference and long-term stability [31]
for glucose determination and the wide linear response range
[32]
from 0.10 to 10.0 mM. Because of its convenient preparation,
good properties, and disposability, this biosensor can be fur- [33]
ther developed for practical clinic detection of glucose. Due to
good immobilization on the Ti-MCM-41 and a good catalytic [34]
response for glucose, the good biocompatibility of Ti-MCM-41 [35]
P.W. Barone, R.S. Parker, M.S. Strano, Anal. Chem. 77 (2005) 7556.
H.T. Zhao, H.X. Ju, Anal. Biochem. 350 (2006) 138.
A. Salimi, R.G. Compton, R. Hallaj, Anal. Biochem. 333 (2004) 49.
P. Kotzian, P. Brazdilova, S. Rezkova, K. Kalcher, K. Vytras, Electroanal-
ysis 18 (2006) 1499.
M.S.P. Lopez, D. Mecerreyes, E. Lopez-Cabarcos, B. Lopez-Ruiz, Biosens.
Bioelectron. 21 (2006) 1013.
B. Sljukic, C.E. Banks, C. Salter, A. Crossley, R.G. Compton, Analyst 131
(2006) 670.
Z.H. Dai, J. Ni, X.H. Huang, G.F. Lu, J.C. Bao, Bioelectrochemistry 70
(2007) 250.
Y. Liu, M. Wang, F. Zhao, Z. Xu, S. Dong, Biosens. Bioelectron. 21 (2005)
984.
S.J. Updike, G.P. Hicks, Nature 214 (1967) 986.
F. Tian, G. Zhu, Anal. Chim. Acta 45 (2002) 251.
T.J. Pisklak, M. Macias, D.H. Coutinho, R.S. Huang, K.J. Balkus Jr., Top.
Catal. 38 (2006) 269.
Y. Xiao, F. Patolsky, E. Katz, J.F. Hainfeld, I. Willner, Science 299 (2003)
1877.
G. Piechotta, J. Albers, R. Hintsche, Biosens. Bioelectron. 21 (2005) 802.
J.H. Yu, S.Q. Liu, H.X. Ju, Biosens. Bioelectron. 19 (2003) 401.
F.H. Zhang, S.S. Cho, S.H. Yang, S.S. Seo, G.S. Cha, H. Nam, Electro-
analysis 18 (2006) 217.
Y.L. Li, K.G. Neoh, L. Gen, E.T. Kang, Langmuir 21 (2005) 10702.
M.J.B. Alvarez, M.T.F. Abedul, A.C. Garcia, Anal. Chim. Acta 462 (2002)
31.
J.P. Hart, A. Crew, E. Crouch, K.C. Honeychurch, R.M. Pemberton, Anal.
Lett. 37 (2004) 789.
A. Ivanov, G. Evtugyn, H. Budnikov, F. Ricci, D. Moscone, G. Palleschi,
Anal. Bioanal. Chem. 377 (2003) 624.
H. Yu, F. Yan, Z. Dai, H.X. Ju, Anal. Biochem. 331 (2004) 98.
D. Hernandez-Santos, M. Diaz-Gonzalez, M.B. Gonzalez-Garcia, A.
Costa-Garcia, Anal. Chem. 76 (2004) 6887.
K.A. Koyano, T. Tatsumi, Microporous Mater. 10 (1997) 259.
E.P. Barrett, L.G. Joyner, P.H. Halenda, J. Am. Chem. Soc. 73 (1951) 373.
W.B. Stockton, M.F. Rubner, Macromolecules 30 (1997) 2717.
P.L. He, N.F. Hu, J.F. Rusling, Langmuir 20 (2004) 722.
P.T. Tanev, T.J. Pinnavaia, Access Nanoporous Mater. (1995) 13.
S. Liu, H. Ju, Biosens. Bioelectron. 19 (2003) 177.
T. Li, Z.H. Yao, L. Ding, Sens. Actuators B 101 (2004) 155.
Y.L. Yao, K.K. Shiu, Anal. Bioanal. Chem. 387 (2007) 303.
T. Seisuke, T. Shinji, H. Satoshi, O. Akiyoshi, Bioceram. Proc. Int. Symp.
Ceram. Med. 12 (1999) 551.
D.P. De Taxis, S. Miyamoto, T. Murakami, J. Kimura, I. Karube, Anal.
Chim. Acta 235 (1990) 255.
K.M. Reddy, S.V. Manorama, A.R. Reddy, Mater. Chem. Phys. 78 (2002)
239.
S.C. Cheng, S.L. Schiefelbein, L. Moore, M. Pierson-Stull, S. Sen, C.
Smith, Microsc. Microanal. Microstruct. (2005) 730.
B.E.P. Swoboda, V. Massey, J. Biol. Chem. 240 (1965) 2209.
J. Wang, P.V.A. Pamidi, M. Jiang, Anal. Chim. Acta 360 (1998) 171.