G Model

BIOS 9002 1–7 BIOS : 9002 Model5G pp:17ðcol:fig: : NILÞ

Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

Development of a screen-printed carbon electrode based disposable

enzyme sensor strip for the measurement of glycated albumin
Mika Hatada a, Wakako Tsugawa a,b, Eri Kamio a, Noya Loew b, David C. Klonoff b,c,
Koji Sode a,b,n
a
Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei,
Tokyo, 184-8588 Japan
b
Institute of Global Innovation Research, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo, 184-8588 Japan
c
Diabetes Research Institute, Mills-Peninsula Health Services, 100 South San Mateo Drive, San Mateo, CA, 94401 USA

art ic l e i nf o a b s t r a c t

Article history: Glycated proteins, such as glycated hemoglobin (HbA1c) or glycated albumin (GA) in the blood, are es-
Received 16 June 2016 sential indicators of glycemic control for diabetes mellitus. Since GA, compared to HbA1c, is more sen-
Received in revised form sitive to short term changes in glycemic levels, GA is expected to be used as an alternative or together
1 August 2016
with HbA1c as a surrogate marker indicator for glycemic control. In this paper we report the develop-
Accepted 2 August 2016
ment of a sensing system for measuring GA by combining an enzyme analysis method, which is already
used in clinical practice, with electrochemical principles. We used fructosyl amino acid oxidase, hex-
Keywords: aammineruthenium(III) chloride as the electron mediator, and an inexpensive and economically at-
Diabetes mellitus tractive screen-printed carbon electrode. We used chronoamperometry to measure protease-digested GA
Glycated albumin
samples. The developed sensor strips were able to measure protease-digested samples containing GA in
Fructosyl amino acid oxidase
very small sample volumes (1.3 μL) within about 1 min. We also prepared enzyme sensor strips suitable
Hemoglobin A1c
Point-of-care testing for clinical use in which the enzyme and the mediator were deposited and dried on. This sensor system
Screen-printed carbon electrode showed a clear correlation between the GA concentration and the resulting current. The strips were
stable following 3 months of storage at 37 °C. We conclude that this disposable enzyme sensor strip
system for measuring GA is suitable for point-of-care test (POCT) applications.
& 2016 Elsevier B.V. All rights reserved.

1. Introduction internal lysine residues of circulating albumin in plasma and re-

Diabetes mellitus is a common disease. It is very important to
maintain the blood glucose level of patients with diabetes at a
normal level to avoid long-term complications. Circulating gly-
cated proteins, such as glycated hemoglobin, which is also known
as hemoglobin A1c (HbA1c) and glycated albumin (GA), are im-
portant indicators of mean glycemic control for diabetes mellitus.
Glycated proteins are the product of a non-enzymatic reaction
between glucose and various circulating blood proteins. HbA1c
originates from the glycation of the β subunit N-terminal valine of
hemoglobin (which resides intracellularly within erythrocytes)
and reflects the mean blood glucose level over the prior 2–3
months (Bunn et al., 1978; Goldstein et al., 2004; Franco, 2012). GA
primarily originates from the glycation of the ε-amino group of
n
Corresponding author at: Department of Biotechnology and Life Science,
Graduate School of Engineering, Tokyo University of Agriculture and Technology,
2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan.
E-mail address: sode@cc.tuat.ac.jp (K. Sode).
flects the average blood glucose level over the prior 2–3 weeks
(Guthrow et al., 1979; Andersen et al., 2014). GA reflects a shorter
timeframe than HbA1c because erythrocytes have a lifespan of
over three months whereas albumin has a half-life of three weeks.
Just as HbA1c is expressed as a percentage of the total hemoglobin
molecules that are glycated, GA is usually quantified as the pro-
portion of the GA concentration divided by the total albumin
concentration. GA values for normal individuals range from 11 to
16% (Japan Diabetes Society, 2014).
HbA1c is the standard accepted indicator of glycemic control
because sufficient evidence has been established to support the
relationship of HbA1c and diabetes-associated complications (The
Diabetes Control and Complications Trial Research Group, 1993;
Stratton et al., 2000). Although HbA1c is the world's most widely
used marker of mean glycemia, GA has two advantages over
HbA1c. First, GA reflects glycemic status over a shorter period than
HbA1c which makes it a better metric for monitoring mean gly-
cemia when a patient's condition is rapidly changing. Second,
HbA1c may not accurately reflect glycemic status in the presence

http://dx.doi.org/10.1016/j.bios.2016.08.005
0956-5663/& 2016 Elsevier B.V. All rights reserved.

Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
2 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
of states with altered erythrocyte life span or genetic variations in
the structure of the hemoglobin molecule. Specifically, disease
processes associated with reduced red blood cell formation, such
as iron and vitamin B12 deficiencies, can elevate HbA1c levels,
whereas conditions with increased hemolysis and replacement of
older erythrocytes with young erythrocytes can lower HbA1c le-
vels (Hare and Shaw, 2016). GA levels, however, are unaffected by
all these conditions affecting hemoglobin production (Koga, 2014).
Recent studies have demonstrated that a GA assay can be used in
evaluating mean glycemic status and predicting a diabetic com-
plication similar to an HbA1c assay (Ueda and Matsumoto, 2015).
GA is expected to be increasingly used as an adjunct or alternative
to HbA1c for estimating glycemic control.
Clinical analyses of GA have been conducted using boronate
affinity methods, ion-exchange high-performance liquid-chroma-
tography (HPLC), or immunoassay. However, within the current
principles of GA clinical analyses, only the enzymatic method is
used, which is the enzyme reagent for autoanalyzers commercia-
s
lized by Asahi-Kasei Pharma (Tokyo, Japan) as Lucica GA-L used
during central laboratory testing (Kohzuma and Koga, 2010). The
principle of enzymatic analysis for GA measurement is detailed
below. GA is digested by protease, and ε-fructocyl lysine (ε-FK) is
released. Ketoamine oxidase, also known as fructosyl amino acid
oxidase (FAOx), which catalyzes the oxidation of glycated amino
acid, is then used, and hydrogen peroxide is produced. Subse-
quently, the liberated hydrogen peroxide is measured in the pre-
sence of peroxidase, N, N-Bis(4-sulfobutyl)-3-methylaniline, dis-
odium salt (TODB), and 4-aminoantipyrine (Kouzuma et al., 2002).
This enzymatic assay is optimized to formulate reagents which are
ready to use in conventional autoanalyzers. FAOxs- or fructosyl
peptide oxidase (FPOxs)-based enzymatic assay systems are now
commercially available for both GA and HbA1c clinical analyses
(Ferri et al., 2009; Sakurabayashi et al., 2003; Hirokawa et al.,
2004; Kohzuma and Koga, 2010; Abidin et al., 2013; Yonehara
et al., 2015). There is an expectation that GA measurement will be
increasingly used to assess glycemic control in: 1) hospitalized
patients; 2) outpatients whose control is rapidly changing; or 3)
patients whose hemoglobin formation is altered (Koga, 2014).
Therefore, the development of accurate GA assays, which are sui-
table for point-of-care testing (POCT), is expected.
Electrochemical sensing systems are widely applied to enzyme-
based POCT. A representative example is the multitude of com-
mercial meter/strip systems for self-monitoring of blood glucose
(SMBG) that are currently available in the global market. They are
generally accurate and easy to use. We expect that electrochemical
principles and enzymatic methods can be integrated into well
designed monitoring systems for measuring GA, particularly those
employing disposable sensor strips. Several studies on the devel-
opment of electrochemical enzyme sensors using FAOx focusing
on the measurement of glycated protein have been reported
(Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al.,
2003; Nanjo et al., 2007; Fang et al., 2009; Chawla and Pundir,
2011, 2012; Jain and Chauhan, 2016). However, to the best of our
knowledge, none of these studies have reported on disposable
enzyme sensor strips for measuring GA in a POCT platform.
In this study, we developed a screen-printed carbon electrode-
(SPCE-) based disposable electrochemical enzyme sensor strip for
the measurement of GA using FAOx, and hexaammineruthenium
(III) chloride (Ru complex) as the electron mediator. The mea-
surement principle in the present study is based on the com-
s
mercially available GA measurement kit (Lucica GA-L), as detailed
below. GA in the blood is digested by a protease to release ε-
fructosyl lysine (ε-FK), following which ε-FK is oxidized by FAOx
and the Ru-complex is simultaneously reduced. The amount of
reduced Ru-complex formed is measured by chronoamperometry
with a SPCE, where the enzyme and electron mediator are
Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
Fig. 1. The principle of GA measurement. The principle of GA measurement was
based on the commercially available GA measurement kit: GA in the blood is di-
gested by protease to release ε-fructosyl lysine (ε-FK), following which ε-FK is
oxidized by FAOx and Ru-complex is simultaneously reduced. The amount of re-
duced Ru-complex formed is measured by chronoamperometry with SPCE.
deposited. The ε-FK concentration contained in protease-digested
GA is determined by the concentration of generated reduced
mediator (Fig. 1). We evaluated the performance of this new en-
zymatic electrochemical-based GA sensor, which can be developed
in a POCT platform.
2. Materials and methods
2.1. Materials
FAOx was prepared following the method by Kouzuma (Kou-
zuma et al., 2002). Recombinant human serum albumin was pur-
chased from Merck (Darmstadt, Germany) and its glycated form
was prepared as below. Twenty-five grams of recombinant human
serum albumin and 80 g of D-glucose (Wako Pure Chemical In-
dustries, Ltd., Osaka, Japan) were dissolved in 500 mL of pure
water and incubated for 24 h at 37 °C. After that, the mixture was
dialyzed against pure water for 3 h for 5 times and then ultra-
filtrated (4 °C, 3500 rpm, 20 min). The glycated albumin (GA)
concentration and the albumin concentration of the prepared
s
sample were determined using Lucica GA-L (Asahi-Kasei Pharma,
Tokyo, Japan), as a reference method. Then, two samples of GA
with a ratio of GA to total albumin of either 15% or 30% and ab-
solute GA concentrations of 0.741 g/dL and 1.442 g/dL, respectively,
were prepared. Proteinase (protease type X X VUU) was obtained
α ε
from Sigma-Aldrich (St. Louis, MO). N -carbobenzyloxy-N -fruc-
tosyllysine (Z-FK) was prepared following the method by Hashiba
(Hashiba, 1976). Hexaammineruthenium(III) chloride (Ru com-
plex) was purchased from Sigma-Aldrich. Screen-printed carbon
electrodes (SPCEs, WE: Carbon, 2.4 mm2, RE: Ag/AgCl, CE: Carbon)
were kindly supplied by i-SENS (Seoul, Republic of Korea). All
electrochemical experiments were performed with a HSV-100
potentiostat (Hokuto Denko Co., Tokyo, Japan).
2.2. Electrochemical measurement of fructosyl lysine or glycated
albumin
A solution of 1.3 μL containing various Z-FK concentrations,
60 U/mL of FAOx, and 300 mM of Ru complex in 100 mM po-
tassium phosphate buffer (P.P.B.) (pH 8.0) was absorbed into the
spacer layer (120 mm thick) of the SPCE. After incubation for 1 min
at room temperature, a potential of þ100 mV vs. Ag/AgCl was
applied and the current was monitored. The Ru complex con-
centration of 300 mM was sufficiently high not to limit the se-
quential enzyme reaction of oxidation of the substrate by FAOx,
and subsequent reduction of the mediator (Supplementary mate-
rial S1). We used þ100 mV vs. Ag/AgCl as the suitable potential for
the oxidation of the reduced Ru complex (Supplementary material
S2).
The measurement of GA was conducted as detailed below. The
mixture of prepared GA samples with ratios of GA to total albumin
 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

of 15% or 30% along with 100 kU/mL of protease in 100 mM P.P.B.
(pH 8.0) was incubated for 1 h at 37 °C to digest GA and release ε-
FK. The mixture was then ultrafiltrated at 4 °C at 14,000  g for
20 min to remove the protease. Subsequently, 1.3 μL of mixture
containing the sample of protease-digested GA, 60 U/mL of FAOx,
and 300 mM of Ru complex in 100 mM P.P.B. (pH 8.0) was ab-
sorbed into the spacer layer (120 mm thick) of the SPCE. After in-
cubation for 1 min at room temperature, a potential of þ100 mV
vs. Ag/AgCl was applied, and the resulting current was observed.
2.3. Preparation and electrochemical measurement of fructosyl ly-
sine or glycated albumin using enzyme sensor strips
The mixture containing 60 U/mL of FAOx, 300 mM of Ru com-
plex, and 0.25% of sucrose in 100 mM P.P.B. (pH 8.0) was deposited
onto the SPCE and dried at 25 °C for more than 30 min. A spacer
(120 mm thick) and a cover were attached afterwards.
For the measurement, a volume of 1.3 μL of each sample (Z-FK
or protease-digested GA) in 100 mM P.P.B. (pH 8.0) was absorbed
into the spacer layer of multiple enzyme sensor strips and in-
cubated at room temperature for 1 min. Subsequently, a potential
of þ100 mV vs. Ag/AgCl was applied, and the current was
observed.
2.4. Evaluation of the storage stability of the enzyme sensor strips
Enzyme sensor strips were prepared as above. To evaluate the
storage stability under different temperatures, the sensors were
placed in a plastic container containing silica gel. They were stored
at 25 °C, 37 °C, or 50 °C in a shaded location. After 91 days, various
Z-FK concentrations were measured using the enzyme sensor
strips. The stability was evaluated by comparing the 91-day
readings with the results obtained from measurements with sen-
sors immediately after preparation.
3. Results
3.1. Electrochemical measurement of fructosyl lysine or glycated
albumin
To evaluate the electrochemical measurement principle used in
the present study, we first attempted to measure a synthetic
analog of ε-FK, namely Z-FK. After mixing FAOx, mediator, and
Z-FK, the mixture was applied on SPCE and incubated at room
temperature for 1 min to facilitate the enzyme reaction coupled
with reduction of the mediator. Afterwards, þ100 mV vs. Ag/AgCl
was applied to the electrode to reoxidize the reduced mediator.
Figs. 2a and b show the response curve and calibration curve,
respectively, of the Z-FK measurements. Good reproducibility was
observed. The resulting currents decreased gradually to reach a
steady state current (Cottrell current), and the current values
correlated positively with Z-FK concentration. The calibration
curve shows the plot of currents obtained at 20 s after applying
potential versus the Z-FK concentrations. From the calibration
curve, a linear response was obtained over the entire Z-FK con-
centration range (0–500 μM), with a linear regression coefficient
of R2 ¼0.999.
The protease-digested products of GA were then measured
using the same electrochemical method. Samples containing ratios
of either 15% or 30% GA to total albumin were used in the present
study. The lower value represents approximately the standard
value of GA, whereas the higher value represents the GA propor-
tion observed during serious diabetes (Japan Diabetes Society,
2014; Tahara, 2009). The ε-FK concentrations of prepared GA were
estimated as 220 μM and 432 μM for 15% GA and 30% GA re-
spectively. Fig. 3a shows the response curve obtained by measur-
ing the mixture of the protease digestion product of the GA, FAOx,
and Ru complex. There was a clear difference between the re-
sulting currents obtained with samples of high and low GA con-
centrations. Protease-digested non-glycated albumin (non-GA)
samples containing known Z-FK concentrations were also mea-
sured, and a calibration curve was plotted (Fig. 3b) to evaluate the
impact of the presence of protease-digested materials on the
sensor response. A slight decrease in the sensitivity (slope) and
linear regression coefficient (R2 ¼ 0.993) was observed (Fig. 3b) in
the presence of protease-digested materials compared with those
in the absence of protease-digested materials (Fig. 2b). The ε-FK
concentrations contained in protease-digested GA were calculated
by applying the current value to the calibration curve formula. The
calculated ε-FK concentrations were 51 and 86 μM for 15% and
30% GA, respectively. However, ε-FK concentrations for these
samples calculated from GA concentrations, that were spectro-
s
scopically determined using the reference method Lucica GA-L,
were 55 and 108 μM, respectively (assuming that 2 molecules of ε-
FK were released per one GA molecule). Because the calibration
curve was based on Z-FK, whereas the protease digested GA
sample contained ε-FK, this slight difference might have resulted
from the difference of reactivity between Z-FK and ε-FK with
FAOx. Since the calibration curve would ultimately be based on GA

a b
1200 300
y=0.54 x + 14
1000 250
R2=0.999
Current (nA)

200
Current (nA)

800
600 150

400 100

200 50

0 0
0 20 40 60 0 200 400 600
Time (s) Z-FK concentration (µM)
Fig. 2. a) The response curve of the Z-FK measurement. A solution of 1.3 μL containing various Z-FK concentrations, 60 U/mL of FAOx, and 300 mM of Ru complex in
100 mM P.P.B. (pH 8.0) was measured. The concentration range of Z-FK was 0, 50, 100, 300, 500 μM (down to up). (n¼ 3), b) The calibration curve of the Z-FK measurement.
Currents obtained at 20 s after applying potential were plotted versus Z-FK concentration.

Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
4 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎

a b
250
200 30% GA y = 0.45 x + 18
200 R2=0.993
15% GA
150

Current (nA)
Current (nA)

150

100 5% GA 100
100
50
50
50
0
30% GA
0 50 100 150
0 0
0 10 20 30 40 50 60 70 0 100 200 300 400 500
Time (s) Z-FK concentration (µM)

Fig. 3. a) The response curve obtained by measuring the mixture of the protease digestion product of GA, FAOx, and Ru complex. Dark lines were obtained from 30% GA and
light lines were obtained from 15% GA respectively. (n ¼ 3), b) Calibration curve of protease digested non-GA samples containing known Z-FK concentrations measurement
(n¼ 3). The current values obtained with protease digested 15% GA (circle) and 30% GA (square) were plotted on the calibration curve. The inset shows enlarged curve.

samples of known concentrations, this difference is not significant
for practical use. Furthermore, GA might not be completely di-
gested and less ε-FK might have been released before the elec-
trochemical measurement with the current conditions for pro-
tease digestion. The protease digestion condition of the commer-
s
cial assay kit Lucica GA-L, which was used as the reference
spectroscopic measurement, has been optimized, so that maximal
ε-FK was released. This might lead to less ε-FK in the sample used
for the electrochemical method compared to the sample used for
the spectroscopic method. Therefore, further optimization of the
condition for protease digestion of GA, suitable for electrochemical
measurement should be necessary to develop a POCT device em-
ploying this disposable enzyme sensor strip.
These results suggested that the current electrochemical
method is principally applicable for the measurement of GA.
3.2. Electrochemical measurement of fructosyl lysine or glycated
albumin using enzyme sensor strips
We then investigated the performance of enzyme sensor strips,
which were prepared initially by depositing the enzyme and
mediators on SPCEs, followed by a drying process. Various Z-FK
concentrations and two different GA samples digested by protease
were measured by absorbing 1.3 μL of sample solution to the
spacer layer of enzyme sensor strips and applying þ100 mV vs.
Ag/AgCl after incubation for 1 min at room temperature. Fig. 4a
shows the response curve of Z-FK measurements. The resulting
current was gradually decreased until it reached a steady state
current, as observed in the measurement of the mixture of Z-FK,
FAOx, and Ru complex (solution condition) (Fig. 2a), and correlated
positively with Z-FK concentration. Once again, good reproduci-
bility was observed. Fig. 4b shows the calibration curve. The linear
response between Z-FK concentrations and current values was
obtained over the entire concentration range of Z-FK (0–500 μM).
Slight decreases in the sensitivity (slope) and linear regression
coefficient (R2 ¼0.993) were observed compared with the solution
condition (Fig. 2b). This might be because of the time and process
required to dissolve of enzyme and mediator on the sensor strips
by the sample solution, which might affect the homogeneity of the
enzyme reaction, and consequently affect the kinetics of the
electrochemical reactions.
Fig. 5a shows the response curve for the measurement of
protease-digested GA samples using the enzyme sensor strips.
There was a clear difference between the resulting currents of high
and low concentration GA samples. Once again, protease-digested
non-GA samples containing known Z-FK concentrations were
measured, and the calibration curve was plotted (Fig. 5b) to
evaluate the impact of the presence of protease-digested materials

300
1000
250
y=0.49 x + 31
800 R2=0.990
Current (nA)

Current (nA)

200
600
150
400
100
200
50
0
0
0 20 40 60 0 200 400 600
Time (s) Z-FK concentration (µM)
Fig. 4. a) The response curve of Z-FK measurement with enzyme sensor strips which were prepared by depositing the enzyme and mediators on SPCE, and were dried. A
solution of 1.3 μL containing various Z-FK concentrations in 100 mM P.P.B. (pH 8.0) was measured. The concentration range of Z-FK was 0, 50, 100, 300, 500 μM (down to up).
(n¼ 3), b) The calibration curve of Z-FK measurement with enzyme sensor strips. The currents obtained at 20 s after applying potential are plotted versus Z-FK concentration.

Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

a b
250
140 y = 0.41 x + 23
30% GA
120 200 R2=0.990
15% GA
100
Current (nA)

150

Current (nA)
80
5% GA 100
60 100
50
40
50
20 0
30% GA 0 50 100 150
0 0
0 20 40 60 0 100 200 300 400 500
Time (s) Z-FK concentration (µM)

Fig. 5. a) The response curve obtained by measuring the solution of the protease digestion product of GA with enzyme sensor strips. Dark lines were obtained from 30% GA
and light lines were obtained from 15% GA, respectively. (n ¼3), b) Calibration curve of protease digested non-GA samples containing known Z-FK concentrations mea-
surement with enzyme sensor strips (n¼ 3). The current values obtained with protease digested 15% GA (circle) and 30% GA (square) were plotted on the calibration curve.
The inset shows enlarged curve.

on the sensor response. A slight decrease in the sensitivity (slope) 250

was observed; however, the same linear regression coefficient 25
(R2 ¼0.990) was observed compared with those in the absence of
protease digested materials (Fig. 4b). Fig. 5b also shows the current 200 37 50 : y=0.57 x + 46
R2=0.988
values obtained with the GA sample indicated on the plot. The 50
calculated ε-FK concentrations from protease-digested GA samples
Current (nA)
150
were 29 and 81 μM for 15% and 30% GA, respectively. However, ε- 0 day
FK concentrations calculated from GA concentrations that were
s
spectroscopically determined using Lucica GA-L were 110 and 100 25 : y=0.59 x + 12
216 μM for 15% and 30% GA, respectively. The gaps between R2=1.000
electrochemically determined values with enzyme sensor strips
50 37 : y=0.56 x + 12
and spectroscopically determined values were larger than the gaps R2=0.999
between electrochemically determined values with solution con- 0 day : y=0.62 x + 4.4
dition and spectroscopically determined values. This might also be 0
because of the time and process required to dissolve enzyme and 0 100 200 300
mediator on the sensor strips by sample solution, which might Z-FK concentration (µM)
have led to the low sensitivity observed. As was stated in Section
3.1, the protease digestion in this study might have been in- Fig. 6. Evaluation of storage stability. The calibration curves of Z-FK measurement
with the stored enzyme sensor strips. The sensor strips were stored at 25 °C
complete and, as a consequence, ε-FK might not have been re- (square), 37 °C (triangle), and 50 °C (circle) for 3 months. Dotted line (0 day) shows
leased completely. Therefore, the response currents obtained for the current with enzyme sensor strips soon after preparation.
the GA samples were lower than what would be expected for the
spectroscopically determined ε-FK concentrations of these sam- 4. Discussion
ples. Furthermore, the expected difference of ε-FK concentration
between 15% and 30% GA samples based on the spectroscopically A number of studies have developed glycated protein biosen-
determined concentrations was about 100 μM. Because the sen- sing methods, including the use of FAOx/FPOx (Tsugawa et al.,
sitivity of the calibration curve (Fig. 5b) was sufficiently high to 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003; Nanjo et al.,
distinguish the expected difference of 100 μM, the sensitivity of 2007; Fang et al., 2009; Chawla and Pundir, 2011, 2012; Jain and
the developed enzyme sensor strips was considered to be sa- Chauhan, 2016), binding proteins (Sakaguchi et al., 2007; Saka-
tisfactory for practical application. guchi-Mikami et al., 2013; Kameya et al., 2015), antibodies (Qu
et al., 2009; Liu et al., 2012; Chopra et al., 2013), molecularly im-
3.3. Evaluation of the storage stability of the enzyme sensor strips printed polymers (Yamazaki et al., 2003; Sode et al., 2003; Raj-
kumar et al., 2008), and aptamer (Apiwat et al., 2016). We recently
The storage stability of prepared enzyme sensor strips was reported an electrochemical sensor for GA using fructosamine
evaluated. After incubation at 25 °C, 37 °C, and 50 °C for 3 months 6-kinase (FN6K) which recognized GA as the substrate (Kameya
(91 days), the various Z-FK concentrations were measured with et al., 2016). This sensing system did not require the protease di-
the stored enzyme sensor strips, and the calibration curves were gestion of GA, but the measurement principle was relatively
constructed (Fig. 6). The slope and background values of the cali- complicated and the sensitivity of the sensor was limited. Fur-
bration curves were little-changed after storage at 25 °C and 37 °C thermore, FN6K has not been used in commercially available en-
for 3 months. In contrast, the background value was increased zymatic assay system for any clinical analyses, unlike FAOx. Al-
while the slope was not changed when the sensor was stored at though these reports demonstrated novelty of the principles, none
50 °C. of them employed clinically relevant principles intended for use in
These results indicated that the sensor was stable after a future POCT system.
3 months of storage under 25 °C or 37 °C, thereby demonstrating In the present study, we developed a disposable electro-
the viability of these strips for practical clinical application. chemical enzyme sensor strip for the measurement of GA based on

Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
6 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
an enzyme analysis method which is already used in clinical
practice, with the combination of FAOx, Ru complex as the elec-
tron mediator, and an inexpensive and economically attractive
SPCE, which will be utilized for POCT of GA. We developed this
sensor strip for GA measurement by referring to the formulation of
the sensor strip utilized in commercial SMBG. In this type of
measurement, the enzyme, mediator and substrate react in a mL-
scale volume, and the potential is applied after starting the reac-
tion. In these cases, the enzyme reaction reaches its equilibrium
relatively fast, usually before the potential is applied. The response
current then is an indicator for the concentration of the reduced
mediator at equilibrium. Therefore, the mediator must be present
enough not to limit the enzyme reaction. Consequently, in SMBG,
several tens to hundreds mM of mediator is often used. For ex-
ample, in one of our group's contribution reported previously,
60 mM Ru complex was used as mediator (Yamaoka and Sode,
2007a) and in another, 60, 120 and 240 nmol/sensor (120, 240,
480 mM) Ru complex were compared and 240 nmol/sensor was
chosen as the optimal concentration (Yamaoka and Sode, 2007b).
s s
In the commercially available OneTouch Ultra Test Strips (for
SMBG), more than 22 μg (67 mM) ferricyanide is used as mediator
(LifeScan, Inc., 2008). We investigated the impact of the Ru com-
plex concentration on the Z-FK measurements (S1), and used
300 mM as the minimal concentration of Ru complex not limiting
the reaction. The FAOx and protease digestion procedures we
employed in the present study are identical to those utilized in
clinical practice. Therefore, the process presented here is a realistic
method to measure GA in blood samples. Furthermore, electro-
chemical methods are often used for POCT systems because of
their suitability for miniaturized systems, good sensitivity, re-
producibility, and low cost, and, moreover, their compatibility with
enzymatic methods. These advantages of the electrochemical
systems are evidenced in the fact that electrochemical sensors are
dominating commercialized SMBG systems, although SMBG sys-
tems were first developed based on reflectometer technology
(Newman and Turner, 2005; Yoo and Lee, 2010; Clarke and Foster,
2012).
In general, electrochemical measurements are subject to in-
terferences, such as ascorbic acid and other electroactive sub-
stances in blood. However, Ru-complex electron transfer allowed
us to conduct electrochemical measurements with a relatively low
operational oxidation potential (þ 100 mV vs Ag/AgCl), which
should minimize the impact of such electroactive ingredients in a
sample solution. Other electroactive interferences might be pro-
tein-derived interfering substances including oxidizable amino
acids, such as cysteine. Because about 60–70% of plasma protein is
albumin, the majority of amino acids would be derived from al-
bumin. Albumin consists of 585 amino acid residues and has 34
cysteine residues. Therefore, any interferences due to a release of
cysteine or other oxidizable amino acids during the protease di-
gestion should occur when measuring protease digested albumin
or GA. The present results indicated that the background current,
namely the y-intercept of the calibration curve of Z-FK measure-
ments, were nearly the same whether they contained (Fig. 2b) or
did not contain (Fig. 3b) protease-digested albumin. Therefore, it
appeared that this measurement system was not affected by in-
terfering substances that could be contained in protease-digested
albumin.
The development of a cost-effective SPCE will facilitate mass
production of GA sensor strips, which can be fabricated using the
same type of procedure that is used for BG monitor systems. The
sensor strips described in this article could measure a protease-
digested GA sample with a very small sample volume (1.3 μL)
within about 1 min after sample preparation by protease diges-
tion. These volume and measurement time conditions are ap-
proximately the same as those required by currently marketed
Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
POCT devices used for monitoring lipid panels (Méndez-González
et al., 2010). We also prepared enzyme sensor strips that are
practical and can be easily used, which showed a clear correlation
between measured current and GA concentrations. The storage
stability of these enzyme sensor strips was 3 months at 37 °C. This
storage capability was achieved with chemical components used
to develop this electrochemical principle only on a laboratory
scale. Further steps to stabilize the system for longer shelf life
storage of the sensor strips could include adding chemical stabi-
lizers to the enzymes, optimizing the drying procedure, and de-
signing an envelope or cartridge for storage.
One critical issue of the current status of the method is the so-
called pretreatment time required to prepare a sample of protease-
digested GA to be measured by the enzyme sensor strips. The
s
current commercially available enzyme analysis kit, Lucica GA-L
also uses a protease digestion procedure. With the Lucia system,
the required time for digestion is 5 min and protease is not re-
moved before adding FAOx contained reagent. Therefore, by op-
timizing the reaction conditions such as pH or temperature, the
period for GA pretreatment can be shortened. Furthermore, the
step for protease removal also can be omitted before electro-
chemical measurement. So the entire analysis procedure will be-
come simpler and ideally, an on-chip reaction for protease diges-
tion will be developed.
The measurement units of GA are specified to be the proportion
of the GA concentration relative to the total albumin concentra-
tion, and therefore, the albumin concentration has to be de-
termined to achieve clinically relevant values. Currently, a spec-
s
troscopic method is used in Lucica GA-L for measuring total al-
bumin. Combining this spectroscopic method with the electro-
chemical GA measurement system in this study can be applicable
for quick practical application. Ideally, an electrochemical mea-
surement system of albumin should be combined. Electrochemical
methods for albumin measurement were also reported (Barisci
et al., 1998; Chuang et al., 2006; Luque et al., 2007; Cieplak et al.,
2015; Yasukawa et al., 2015), which may lead to the development
of a principle suitable for disposable sensing system. Therefore, the
future combination of an electrochemical principle for total albu-
min concentration and our electrochemical principle for GA sen-
sing will realize an electrochemical POCT device to measure GA
units specified as the proportion of the concentration relative to
the total albumin concentration.
Our research group is also engaged in the development of
various biomolecules dedicated to sensing glycated proteins. In
particular, we have reported the construction of engineered FAOx/
FPOx (Kim et al., 2010, 2012; Ferri et al., 2004, 2005, 2013). In
contrast to the wild-type enzymes, the engineered enzymes
strongly preferred artificial molecules to oxygen as electron ac-
ceptor. By employing these types of engineered FAOx enzymes,
then an increase in sensitivity will also likely be achieved.
To elevate the research presented in this article into a com-
mercial product, then further study will be required to use human
serum or plasma as the sample measured by these enzyme sensor
strips under development. Furthermore, a monitor will need to be
created with an algorithm to convert current to a GA concentra-
tion. These developments will result in the development of an
electrochemical POCT sensor for GA.
5. Conclusion
In our study, we developed a sensing system for measurement
of GA by combining an enzyme analysis method, which is already
used in clinical practice, with electrochemical principles. FAOx, Ru
complex as the electron mediator, and an inexpensive and eco-
nomically attractive SPCE were used in this study. The prepared
 M. Hatada et al. / Biosensors and Bioelectronics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
enzyme sensor strips were able to measure a protease-digested GA
sample with a very small sample volume (1.3 μL) within about
1 min after sample preparation by protease digestion. The system
had a storage stability of 3 months at 37 °C. These results indicate
that the developed sensor has a legitimate potential for a POCT
application. Further studies, such as optimization of protease di-
gestion procedure will advance the progression of this sensor to-
ward clinical use.
Acknowledgments
We acknowledge i-SENS (Seoul, Republic of Korea) for kindly
supplying screen printed carbon electrodes. We also acknowledge
Asahi-Kasei Pharma (Tokyo, Japan) for kindly supplying samples
including FAOx, Z-FK, GA, recombinant albumin, and protease, and
furthermore, for helping with the determination of GA con-
s
centration by Lucica GA-L.
Appendix A. Supplementary material
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.bios.2016.08.005.
References
Andersen, J.T., Dalhus, B., Viuff, D., Ravn, B.T., Gunnarsen, K.S., Plumridge, A.,
Bunting, K., Antunes, F., Williamson, R., Athwal, S., Allan, E., Evans, L., Bjørås, M.,
Kjærulff, S., Sleep, D., Sandlie, I., Cameron, J., 2014. J. Biol. Chem. 289,
13492–13502.
Abidin, D., Liu, L., Dou, C., Datta, A., Yuan, C., 2013. Anal. Methods 5, 2461–2469.
Apiwat, C., Luksirikul, P., Kankla, P., Pongprayoon, P., Treerattrakoon, K., Pai-
boonsukwong, K., Fucharoen, S., Dharakul, T., Japrung, D., 2016. Biosens. Bioe-
lectron. 82, 140–145.
Bunn, H.F., Gabbay, K.H., Gallop, P.M., 1978. Science 200, 21–27.
Barisci, J.N., Hughes, D., Minett, A., Wallace, G.G., 1998. Anal. Chim. Acta 371, 39–48.
Chawla, S., Pundir, C.S., 2011. Biosens. Bioelectron. 26, 3438–3443.
Chawla, S., Pundir, C.S., 2012. Anal. Biochem. 430, 156–162.
Chopra, A., Tuteja, S., Sachdeva, N., Bhasin, K.K., Bhalla, V., Suri, C.R., 2013. Biosens.
Bioelectron. 44, 132–135.
Chuang, M.C., Liu, C.C., Yang, M.C., 2006. Sens. Actuators B 114, 357–363.
Clarke, S.F., Foster, J.R., 2012. Br. J. Biomed. Sci. 69 (2), 83–93.
Cieplak, M., Szwabinska, K., Sosnowska, M., K.C., C.B, Borowicz, P., Noworyta, K.,
D'Souza, F., Kutner, W., 2015. Biosens. Bioelectron. 74, 960–966.
Franco, R.S., 2012. Transfus. Med Hemother. 39, 302–307.
Ferri, S., Miura, S., Sakaguchi, A., Ishimura, F., Tsugawa, W., Sode, K., 2004. Mar.
Biotechnol. 6, 625–632.
Ferri, S., Sakaguchi, A., Goto, H., Tsugawa, W., Sode, K., 2005. Biotechnol. Lett. 27,
27–32.
Ferri, S., Kim, S., Tsugawa, W., Sode, K., 2009. J. Diabetes Sci. Technol. 3, 585–592.
Ferri, S., Miyamoto, Y., Sakaguchi-Mikami, A., Tsugawa, W., Sode, K., 2013. Mol.
Biotechnol. 54, 939–943.
Please cite this article as: Hatada, M., et al., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.08.005i
7
Fang, L., Li, W., Zhou, Y., Liu, C.-C., 2009. Sens. Actuators B 137, 235–238.
Goldstein, D.E., Little, R.R., Lorenz, R.A., Malone, J.I., Nathan, D., Peterson, C.M.,
Sacks, D.B., 2004. Diabetes Care 27, 1761–1773.
Guthrow, C.E., Morris, M.A., Day, J.F., Thorpe, S.R., Baynes, J.W., 1979. Biochemistry
76, 4258–4261.
Hare, M.J., Shaw, J.E., 2016. Curr. Diabetes Rev. 12, 51–57.
Hirokawa, K., Nakamura, K., Kajiyama, N., 2004. FEMS Microbiol. Lett. 235, 157–162.
Hashiba, H.,J., 1976. Agric. Food Chem. 24, 70–73.
Japan Diabetes Society, Treatment Guide for Diabetes 2014–2015. 〈http://www.fa.
kyorin.co.jp/jds/uploads/Treatment_Guide_for_Diabetes_2014-2015.pdf〉.
Jain, U., Chauhan, N., 2016. Biosens. Bioelectron., doi:10.1016/ j.bios.2016.02.033 [e-
pub ahead on line]
Kameya, M., Sakaguchi-Mikami, A., Ferri, S., Tsugawa, W., Sode, K., 2015. J. Diabetes
Sci. Technol. 9, 183–191.
Kameya, M., Tsugawa, W., Yamada-Tajima, M., Hatada, M., Suzuki, K., Sakaguchi-
Mikami, A., Ferri, S., Klonoff, D.C., Sode, K., 2016. Biotechnol. J. 2016, 1–8.
Kim, S., Nibe, E., Ferri, S., Tsugawa, W., Sode, K., 2010. Biotechnol. Lett. 32,
1123–1129.
Kim, S., Nibe, E., Tsugawa, W., Kojima, K., Sode, K., 2012. Biotechnol. Lett. 34,
491–497.
Koga, M., 2014. Clin. Chim. Acta 43, 96–104.
Kohzuma, T., Koga, M., 2010. Mol. Diagn. Ther. 14, 49–51.
Kouzuma, T., Usami, T., Yamakoshi, M., Takahashi, M., Imamura, S., 2002. Clin. Chim.
Acta 324, 61–71.
Liu, G., Iyengar, S.G., Gooding, J.J., 2012. Electroanalysis 24, 1509–1516.
LifeScan, Inc., 2008. OneTouchs Ultras Test Strips, 〈https://www.onetouch.com/
sites/d7admin-onetouch.jnj.com.professional/files/docfile/1391164268C_UBlue
TestStrip_Instructions_for_use.pdf〉.
Luque, G.L., Ferreyra, N.F., Rivas, G.A., 2007. Talanta 71, 1282–1287.
Méndez-González, J., Bonet-Marqués, R., Ordóñez-Llanos, J., 2010. Point Care: J.
Near-Patient Test. Technol. 9, 102–107.
Nanjo, Y., Hayashi, R., Yao, T., 2007. Anal. Chim. Acta 583, 45–54.
Newman, J.D., Turner, A.P.F., 2005. Biosens. Bioelectron. 20, 2435–2453.
Ogawa, K., Stoöllner, D., Scheller, F., Warsinke, A., Ishimura, F., Tsugawa, W., Ferri, S.,
Sode, K., 2002. Anal. Bioanal. Chem. 373, 211–214.
Qu, L., Xia, S., Bian, C., Sun, J., Han, J., 2009. Biosens. Bioelectron. 24, 3419–3424.
Rajkumar, R., Katterle, M., Warsinke, A., Möhwald, H., Scheller, F.W., 2008. Biosens.
Bioelectron. 23, 1195–1199.
Sakaguchi, A., Tsugawa, W., Ferri, S., Sode, K., 2003. Electrochemistry 71, 442–445.
Sakurabayashi, I., Watano, T., Yonehara, S., Ishimaru, K., Hirai, K., Komori, T., Tagi, M.,
2003. Clin. Chem. 49, 269–274.
Sakaguchi, A., Ferri, S., Tsugawa, W., Sode, K., 2007. Biosens. Bioelectron. 22,
1933–1938.
Sakaguchi-Mikami, A., Ferri, S., Katayama, S., Tsugawa, W., Sode, K., 2013. J. Appl.
Microbiol. 114, 1449–1456.
Sode, K., Ohta, S., Yanai, Y., Yamazaki, T., 2003. Biosens. Bioelectron. 18, 1485–1490.
Stratton, I.M., Adler, A.I., Neil, H.A.W., Matthews, D.R., Manley, S.E., Cull, C.A.,
Hadden, D., Turner, R.C., Holman, R.R., 2000. BMJ 321, 405–412.
The Diabetes Control and Complications Trial Research Group, 1993. N. Engl. J. Med.
329, 977–986.
Tahara, Y., 2009. Diabetes Res. Clin. Pract. 84, 224–229.
Tsugawa, W., Ishimura, F., Ogawa, K., Sode, K., 2000. Electrochemistry 68, 869–871.
Tsugawa, W., Ogawa, K., Ishimura, F., Sode, K., 2001. Electrochemistry 69, 973–975.
Ueda, Y., Matsumoto, H., 2015. J. Diabetes Sci. Technol. 9, 177–182.
Yamazaki, T., Ohta, S., Yanai, Y., Sode, K., 2003. Anal. Lett. 36, 75–89.
Yamaoka, H., Sode, K., 2007a. J. Diabetes Sci. Technol. 1, 28–35.
Yamaoka, H., Sode, K., 2007b. Open Biotechnol. J. 1, 26–30.
Yoo, E.H., Lee, S.Y., 2010. Sensors 10, 4558–4576.
Yonehara, S., Inamura, N., Fukuda, M., Sugiyama, K., 2015. J. Diabetes Sci. Technol. 9,
200–205.
Yasukawa, T., Kiba, Y., Mizutani, F., 2015. Anal. Sci. 31, 583–589.