BIOINTERFACES BASED ON IMMOBILIZED BORONIC ACID

WITH SPECIFITY TO GLYCATED PROTEINS

Jan Přibyl and Petr Skládal

National Center for Biomolecular Research, Masaryk University, Kotlářská 2, Brno, Czech Republic

Keywords: Glycated hemoglobin, aminophenylboronic acid, biosensors, quartz crystal microbalance, heterogeneous

affinity assay, microtitre plate.

Abstract: Development of bioanalytical assays for determination of glycated hemoglobin content in blood samples is

reported. First, a combined biosensor setup for determination of total and glycated hemoglobin content was

successfully developed and tested. The effect of various operating parameters, such as ionic strength, flow

rate and instrumental set-up was optimized. The total hemoglobin content was analyzed by measuring of

absorbance of the hemoglobin-cyanide derivative at 540 nm. Only one standard (calibrator), diluted in

various proportions, was necessary for the method calibration. The full range of HbA

content (4 to 15 %)

presented in blood can be analyzed. Only 1 μl of blood was required for analysis. The developed method

was successfully evaluated for analysis of blood samples collected from diabetic patients. Next, the

heterogeneous affinity assay performed in a microtitre plate with an immobilized boronic acid is described.

This assay is based on ELISA (Enzyme-Linked Immunosorbent Assay) principle; however stable

chemiselective ligand is used in this case. The content of glycated hemoglobin is determined according to its

peroxidase activity after attachment to immobilized boronic acid derivative; the total hemoglobin

concentration is measured as an absorbance at 405 nm.

1 INTRODUCTION

Diabetes mellitus is a group of diseases

characterised by high levels of blood glucose

resulting from defects in insulin production, insulin

action, or both. Diabetes can be associated with

serious complications and premature death. Diabetes

was the sixth leading cause of death in USA in 2000.

(National Diabetes Information Clearinghouse,

http://diabetes.niddk.nih.gov). The steps to control

the disease and lower the risk of complications

should be taken. In this way, blood and urine

glucose analysis, cholesterol reduction and blood

pressure control should be mentioned. Analysis of

glycated hemoglobin (HbA

) helps to monitor the

long-term progression of diabetes without influence

of the short-term fluctuations of blood glucose. The

fraction of HbA

is usually indicated as percentage

of its presence in the total hemoglobin content. The

content of glycated hemoglobin in blood should

substitute the term “glycemia”; values lying under

7% indicate good health state of patient and

effective practicing of the proposed therapy

(Marshall and Barth, 2000).

Glycated hemoglobin (GHb) refers to a series of

minor hemoglobin components which are stable

adducts formed by reaction of hemoglobin primary

aminogroups with various sugars. Hemoglobin

HbA

is a stable glucose adduct to the N-terminal

group of the β-chain of HbA

. In current opinion,

concentration of the hemoglobin variant HbA

considered to be the only specific and stable

indicator of long-term diabetes progress. Neither the

whole glycated fraction of hemoglobin (HbA

) nor

fructosamine can be any longer used in the disease

diagnosis.

A wide range of methods for analysis of the

glycated hemoglobin (either HbA

or HbA

) has

been reported recently. However, only few of them

were based on biosensor approach and mostly the

chromatographic approaches were employed. From

the area of biosensor development, especially two

concepts should be mentioned. The first one was

based on selection of ligands from the hexapeptide

combinatorial library for binding the glycated

terminus of hemoglobin β-chain; thus found

hexaptides exhibited high specifity and stability

(Chen et al., 1998). However, only the preliminary

study was performed, additional testing of this

107

ribyl J. and Skládal P. (2008).

BIOINTERFACES BASED ON IMMOBILIZED BORONIC ACID WITH SPECIFITY TO GLYCATED PROTEINS.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 107-112

DOI: 10.5220/0001051501070112

 SciTePress

affinity molecules and integration to some analytical

method (e.g. affinity chromatography) is essential

for correct evaluation. A monolayer of boronic acid

conjugate with 11-mercaptoundecaonic acid

immobilised on the surface of gold nanoclusters was

used as recognition element in another study

(Valina-Saba et al., 1999). An easy approach was

reported, when the precipitation reaction between

boronic groups on the particle surface and glycated

protein (horse-radish peroxidase) was visible.

Unfortunately, the application for determination of

HbA

content in whole blood, which is rather

complex mixture, was not tested.

Boronic acid shows ability to bind covalently to

either 1,2- or 1,3-diols and thus forms five- or six-

membered cyclic esters. 3-aminophenylboronic acid

(APBA) binds in this way to the cis-diols of

saccharides, glycated proteins or nucleic acids

(Pickup et al., 2005). The formation of a boronate

ester is usually described as a two step reaction; the

planar boron group initially reacts with hydroxyl

(pH>7.0 is essential) to form tetrahedral boronate

anion, which subsequently binds reversibly to the

positively charged carbon atoms in the diol-

containing structure (Ito et al., 2003). This kind of

ester formation designates boronic acid and its

derivatives to be used as the affinity recognition

elements in variety of applications, such as

construction of sensors for saccharides with

piezoelectric (Lau et al., 2000) and surface plasmon

resonance (Kugimiya and Takeuchi, 2001)

transducers or fluorescent (Kataoka et al., 1995)

detection. Boronic acid derivatives immobilized in

the matrix of columns have formed the basis of a

new field of chromatographic techniques designated

for analysis and separation of sugars and glycated

proteins. This area is commonly known as boronate

affinity chromatography (Bongartz and Hesse,

1995).

The aim of presented study was to continue our

previous attempts with boronate-modified sensors

for sugars (Přibyl and Skládal, 2005) in order to

develop an innovative, easy to handle and cost

effective but reliable biosensor set-up with high

stability and reproducibility for determination of

glycated hemoglobin in blood samples. The

designed system contains two parts, one performs

the analysis of HbA

using a piezoelectric sensor

modified with phenylboronic acid, and the second

one is designed for a photometric determination of

total Hb. The absolute concentration of these blood

components differs in each sample. The percentage

of HbA

presence will be determined as a ratio of

these two concentrations;

(conc

HbA1c

/conc

Total Hb

) x 100%.

Another interesting method for detection of

glycated hemoglobin is reported, too. This

bioanalytical method called AHA (Affinity

Heterogeneous Assay) employs microtitre plates

consisting of the wells covered with

aminophenylboronic acid. The AHA assay allows

determination of total hemoglobin as well as

glycated fraction of hemoglobin in blood samples,

similarly as the biosensor based method.

2 EXPERIMENTAL

2.1 Chemicals and Reagents

Chemicals were obtained from Sigma (St. Luis,

USA) and used as received without any further

purification. Microtitre plates with chemically

reactive surface (NUNC Immobilizer Amino) were

from Nunc (Copenhagen, Denmark).

The special solutions were prepared, stored and

used as officially recommended (International

Committee for Standardization in Haematology,

1978) for analysis of total hemoglobin content in

blood samples.

2.2 Instrumentation

Measurements with the piezoelectric biosensor were

performed using 10 MHz, AT-cut quartz crystals

(ICM, Oklahoma City, OK, USA) with gold-coated

smooth quartz discs (electrode area, 0.8 cm

In the center of the system, there was placed a

PMMA-made flow-trough cell (internal volume

10 μl) from NanoQ (Brno, Czech Rep.) with the

piezoelectric biosensor mounted between two silicon

rubber O-rings. The cell was supplied with flowing

liquid via two stainless steel tubes (i.d., 0.5 mm).

Sensor was connected to MultiLabPlus instrument

(MultiLab) combining oscillator with high resolution

frequency counter.

Handling of liquids and samples was performed

by the FIAlab 3500b instrument (Alitea, Seattle,

WA, USA).

Optical part for determination of total

hemoglobin content was located in front of the

biosensor cell. The detector consisted of a Z-type

flow-trough absorption cell (optical path, 10 mm)

supplied with flowing liquids trough the Teflon

tubing and standard visible light source and optical

fibre spectrophotometer from Ocean Optics

(Dunedin, FL, USA).

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

108

2.3 Immobilization Procedure

2.3.1 Affinity Biosensor Fabrication

Matrix based layer was prepared when 2% solution

of polyethylene imine (PEI) in methanol (3 μl) was

used to activate the gold surface. The APBA layer

was attached through the glutaraldehyde linker (8%,

8 hours, 4

C) In the last step, the recognition layer

was stabilised by the reduction of Schiff bonds with

10 mg/ml solution of sodium borohydride (2 hours).

The thiocompound-APBA conjugates were

prepared in order to modify the gold biosensor

surface with a monolayer of boronate groups. In the

first step, the carboxygroup of mercapto-terminated

(on the opposite side of chain) acids was activated

with carbodiimide (3 hours, 99

C). Conjugation of

aminophenylboronic acid to bellow mwntioned

thiocompounds was performed during the next step:

DTSP, 11-MUA, 16-mercaptohexadecanoic acid and

lipoic acid (3 hours, 99

C). Final products exhibited

light-yellow color and were stored at -20

C prior

use.

A monolayer of boronic groups was prepared,

when the freshly cleaned gold electrodes were

incubated with 15 μl of the conjugate for 24 hours at

laboratory temperature in a closed chamber.

Gold surface modified with

11-mercaptoundecanoic acid and the freshly cleaned

gold electrode were used as reference surfaces. For

comparison of binding specifity to the matrix-

modified surfaces, the polyethylene imine layer was

attached to the piezosensor.

2.3.2 Specific Modification of Microtitre

Plate

The ‘Amino Immobilizer’ microtitre plate from

Nunc shows ability to bind covalently any molecule

containing primary aminogroup. 10 mg/mL solution

of aminophenylboronic acid (APBA) in 50 mM

carbonate buffer pH=9.5 was used to cover the

microtitre plate with boronic groups. The solution of

APBA was kept overnight under ambient

temperature in order to cover the wells of plate with

boronic groups. After 4-times repeated washing

(PBS pH=7.4), the unreacted surface group were

saturated with glycine (25 mg/mL in PBS buffer

pH=7.4) during 2 hours reaction performed under

ambient temperature. After thorough washing with

PBS, the plates were dried in the air stream of

ambient temperature (4 hours). Such modified plates

can be long-term stored in a well sealed box (4

without any significant loosing of their binding

activity.

2.3.3 Biosensor Setup - Measuring

Procedure

A similar protocol was used for all experiments:

Figure 1: I – Immobilization of the APBA molecule to the

activated surface of microtitre plate (A, activated surface;

B, APBA solution added; C, APBA-modified surface. II –

Procedure of total and glycated hemoglobin determination

in boronic acid modified plate (A, well filled with diluted

blood sample – total hemoglobin determination during

binding of GHb to the surface; B, peroxidase activity of

bound GHb is used to oxidase a substrate; C, reaction

stopped and the overall activity measured).

after 5 min of the base-line signal stabilisation with

the carrier buffer, the flow of sample - glycated

hemoglobin dissolved in the carrier buffer

(alternatively supplied by sorbitol solution) followed

for 7 min. For the next 7 min, the flow cell was

washed with the carrier buffer in order to equilibrate

the signal (non-specifically adsorbed molecules

dissociated from the biosensor surface). Injection of

200 mM aquatic solution of HCl for 120 seconds

disintegrated the complex formed between glycated

hemoglobin and monolayer of boronic acid groups;

complex [GHb-matrix immobilized APBA]

dissociated spontaneously. Washing with working

buffer for a few minutes followed (new base-line

stabilization) prior performing the next measuring

cycle.

2.3.4 AHA Analysis

Another way of determination of total hemoglobin in

blood samples is, comparing to the previously

employed conversion to a cyanomethemoglobin, its

conversion to alcalic hematin. The carbonate buffer

pH=9.0 was used for this purpose, when the blood

samples were diluted 400-times in this medium, the

most of hemoglobin molecules was converted to the

hematin. This can be quantified by measuring of

absorbance at 405 nm. Moreover, the alcalic pH is

an optimal value to support the affinity interaction

between the boronic group and GHb in a solution.

O=C

5 minutes

O=C

NHS

O=C

NHS

OH OH

OHOH

A B

BIOINTERFACES BASED ON IMMOBILIZED BORONIC ACID WITH SPECIFITY TO GLYCATED PROTEINS

109

The blood solution (in a carbonate buffer) was left to

interact with the surface immobilized boronic groups

for 60 minutes, at ambient temperature, in a closed

box. The total hemoglobin content was quantified as

change of A

405

during 20 minutes following the

reagent addition. Afterwards, the wells were washed

4-times with PBS buffer pH=7.4 and the peroxidase

substrate solution, containing 0.075% hydrogen

peroxide and 105 ug/mL of tetramethylbenzidine in

50 mM acetate buffer pH=4.5 (solution freshly

prepared before each experiment). The enzymatic

reaction releasing intensive blue color was left to

proceed for 5 minutes in a dark box heated to 37

The reaction was stopped by addition of 50 uL of 1

M H

solution to each well. The color of solution

in wells turns yellow immediately. The amount of

the enzymatic reaction product was measured as

absorbance at 405 nm in a microtitre plate reader.

The absorbance of the whole blood solution

corresponds to a total hemoglobin in a sample, the

enzymatic activity of bound hemoglobin (measured

as A

405

in a next step) is proportional to a glycated

hemoglobin content. The percentage of GHb

presence in the total hemoglobin can be easily

calculated by simple dividing of those two values.

3 RESULTS AND DISSCUSSION

3.1 Biosensor based Experiments

The amount of boronic groups deposited on the

surface of piezoelectric sensors was first monitored

during the immobilisation procedure. The deposited

mass was calculated according to Saurbray equation

from the difference of resonant frequency during

deposition. These results indicate that the highest

amount of boronic groups was coupled to the

biosensor surface via 3,3′-Dithiodipropionic acid

di(N-hydroxysuccinimide ester (DTSP),

11-mercaptoundecanoic acid (11-MUA) and mainly

inside the polyethylene imine structure.

However, the evidence of optimal affinity for

samples containing glycated hemoglobin provided

the comparative experiments. Within those, the eight

types of prepared biosensors with either specific or

reference surfaces were consequently mounted into

the flow-trough cell and the response to GHb sample

(410 μg/ml, constant concentration) was monitored.

The experiments were done in duplicate with each

sensor; the equal scheme of experiment was used in

all cases. The lowest response provided the Gold-

PEI-GA-APBA sensor (together with the appropriate

reference one). Therefore these were tested for their

ability to bind sorbitol (low molecular compound

containing vicinal diol group) in concentration of

10 mg/ml (in phosphate buffer pH=9.0). Response of

the specific sensor (229.6 Hz) together with the

reference one (19.8 Hz) showed correctness of

theoretical predictions. Low density of boronate

groups presented on the top of PEI-matrix (low

affinity to glycated hemoglobin) allowed only low

binding of glycated hemoglobin. Moreover, the

difference between specific and non-specific

response (66.5 vs. 44.6 Hz, respectively) was the

next, and probably main, reason to exclude the PEI-

GA-APBA recognition layer from further use in

GHb analysis.

As it was commonly considered the boronic

acid-diol interaction is not substantially affected by

ionic strength of environment. However, most recent

publications (Zhong et al., 2004) indicated a

substantial increase of boronate affinity to diol group

in low ionic strength solutions (co-solute

concentrations up to 0.25 M). Determination of the

ionic strength effect on glycated hemoglobin

interaction with immobilised boronic groups was not

the principal aim of our study. However, the

examination of influence of the used various

reagents on interaction were carried out prior to the

final analysis. The ionic strength of tested reagents

proceeded from 0.9 to 84.3 mM (Modified Drabkin

reagent and 50 mM phosphate buffer, respectively),

pH was in range 7.4 - 9.6. Low ionic strength, i.e.

use of Modified Drabkin Reagent, supports the

affinity interaction. This result well correlates with

findings of other authors.

The biosensor Gold-MUA-APBA and the

previously optimized conditions (peristaltic pump;

flow rate of 100 μl/min; Modified Drabkin reagent

as the working medium and the 2 min regeneration

of sensing surface with 200 mM HCl) were used in

all calibration experiments. The presented method

shows the advantage of calibration using only one

standard solution – blood sample with defined

content of glycated hemoglobin. After dilution in

various proportions, thus obtained standards were

used for calibration. The response of the

piezoelectric biosensor as well as photometric sensor

was increasing with increasing concentration of

glycated and total Hb, respectively. The percentage

of glycated hemoglobin was calculated as the

glycated hemoglobin / total Hb ratio (x 100%). Both

values (glycated and total hemoglobin concentration,

respectively) were taken from the calibration curves,

constructed as the response of biosensor and

photometric sensor to concentration of glycated and

total hemoglobin, respectively. The biosensor can

not be calibrated only using samples containing

various percentage of glycated hemoglobin, the

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

110

10 20 30 40 50

0.1

0.2

0.3

0.4

0.5

0.6

405

GHb [μg/ml]

APBA

Glycine

( )

amount of total hemoglobin should be considered,

too.

A blood sample of diabetic patient with high

content of glycated hemoglobin (14.3%; determined

by the ion-exchange HPLC) was used for calibration

of our setup. The set of six samples for calibration of

analyser was prepared by dilution of blood with the

Modified Drabkin reagent in the following

sequence: 300, 375, 500, 600, 1000 and 2875-times.

Thus prepared samples were placed to the

autosampler and after 15 min of preincubation

(including the base-line stabilisation) were

consequently measured. The combined calibration

graph was constructed using the responses of

photometric and piezoelectric sensors in the time

5 min (Fig. 2).

Figure 2: Calibration of a combined setup, upper curves

show absorbance changes due to the different hemoglobin

content in samples, the curves below were recorded as a

result of binding of various concentration of glycated

hemoglobin.

3.2 AHA Experiments

The AHA method was optimized according its

ability to bind glycated hemoglobin. Buffers of

various pH (7 and 9, respectively) were used to

maximize the surface affinity to GHb. Although

there was found higher adsorption of glycated

hemoglobin at pH=7, the next experiment showed

the low specifity of binding at this pH. On the other

hand, use of buffer of pH=9.0 provides quite a lower

capacity of the surface, however, the binding is

highly specific (Fig. 3). The reference surface,

covered only with glycine, was employed to

compare specifity of binding.

In the further experiments the assay was calibrated

by use of hemoglobin standard solution (total Hb

calibration) and blood sample (with known GHb

content, determined by a standard method), both

diluted in various ratio in order to get at least five

points in a calibration graph. The total hemoglobin

assay provided a linear response in range 10 – 1000

µg/mL; the GHb analysis can be performed in the

concentration range varying between 10 and 40 µg

GHb/mL.

Figure 3: Measured enzymatic activity as result of binding

of glycated hemoglobin to reference (modified with

glycine) and specific surface (modified with

aminophenylboronic acid, APBA). Experiments

performed at pH=9.0.

4 CONCLUSIONS

Two methods for analysis of hemoglobin A

are

presented. First, a combined biosensor for

determination of glycated and total hemoglobin in

blood is reported. The amount of total hemoglobin is

measured in flow-through photometric sensor,

concentration of the glycated fraction is

subsequently monitored by its binding to the

APBA-modified piezoelectric biosensor (higher

content cause higher damping of resonant

frequency).

The other method is based on ELISA principle

(AHA, Affinity Heterogeneous Assay) in either

direct or indirect arrangement. Boronic acid

derivative, capturing the glycated fraction of

hemoglobin, is immobilized on the surface of the

microtitre plate. Amount of glycated hemoglobin is

visualized by measuring of its peroxidase activity.

Total hemoglobin concentration is measured

photometrically at 405 nm.

104

50 H

1 min

219

367

444

587

730

2 min

0.1 A.U.

(540 nm)

sample

buffer

104

50 H

1 min

104

50 H

1 min

219

367

444

587

730

2 min

0.1 A.U.

(540 nm)

sample

buffer

BIOINTERFACES BASED ON IMMOBILIZED BORONIC ACID WITH SPECIFITY TO GLYCATED PROTEINS

111

Both methods present promising approach in

diagnosis of glycohemoglobin. The first one presents

fully automatic, low-cost instrument, the other one

offers the possibility to monitor simultaneously

HbA

content in 96 blood samples within a

relatively short time (2 hours).

Possible use of those methods in routine analysis

of blood samples and their comparison with

conventional methods (HPLC) was shown, too.

ACKNOWLEDGEMENTS

The work was supported by the grant no.

KJB401630701 of the Grant Agency of the Czech

Academy of Science.

REFERENCES

Bongartz, D., Hesse, A., 1995. Selective extraction of

quercetrin in vegetable drugs and urine by off-line

coupling of boronic acid affinity chromatography and

high-performance liquid chromatography. J.

Chromatogr., B. 673, 223-230.

Chen, B., Bestetti, G., Day, R.M., Turner, A.P.F., 1998.

The synthesis and screening of a combinatorial peptide

library for affinity ligands for glycosylated

hemoglobin. Biosens. Bioelectron. 13, 779-785.

International Committee for Standardization in

Haematology, 1978. Recommendations for reference

method for hemoglobinometry in human blood (ICSH

Standard EP 6/2: 1977) and specifications for

international hemoglobincyanide reference preparation

(ICSH Standard EP 6/3: 1977). J. Clin. Pathol. 31,

139-143.

Ito, H., Kono, Y., Machida, A., Mitsumoto, Y., Omori, K.,

Nakamura, N., Kondo, Y., Ishihara, K., 2003. Kinetic

study of the complex formation of boric and boronic

acids with mono- and diprotonated ligands. Inorg.

Chim. Acta 344, 28-36.

Kataoka, K., Hisamitsu, I., Sayama, N., Okano, T.,

Sakurai, Y., 1995. Novel sensing system for glucose

based on the complex formation between phenylborate

and fluorescent diol compounds. J. Biochem. 117,

1145-1147.

Kugimiya, A., Takeuchi, T., 2001. Surface plasmon

resonance sensor using molecularly imprinted polymer

for detection of sialic acid. Biosens. Bioelectron. 16,

1059-1062.

Lau, O.W., Shao, B., Lee, M.T.W., 2000. Affinity mass

sensors: determination of fructose. Anal. Chim. Acta

403, 49-56.

Marshall, S.M., Barth, J.H., 2000. Standardization of

HbA

measurements: a consensus. Ann. Clin.

Biochem. 37, 45-46.

Pickup, J.C., Hussain, F., Evans, N.D., Rolinski, O.J.,

Birch, D.J.S., 2005. Fluorescence-based glucose

sensors. Biosens. Bioelectron. 20, 2555–2565.

Přibyl, J., Skládal, P., 2005. Quartz crystal biosensor for

detection of sugars and glycated hemoglobin, Anal.

Chim. Acta 530, 75-84.

Valina-Saba , M., Bauer , G., Stich , N., Pittner , F.,

Schalkhammer T., 1999. A self assembled shell of 11-

mercaptoundecanoic aminophenylboronic acids on

gold nanoclusters. Mat. Sci. Eng. C 9, 205-209.

Zhong, H., Li, N., Zhao, F., Li, K., 2004. Determination of

proteins with Alizarin Red S by Rayleigh light

scattering technique. Talanta 62, 37–42.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

112