1. Yong Zhang1, 2
Guichen Ping2
Noritada Kaji2, 3
Manabu Tokeshi2, 3
Yoshinobu Baba2, 3, 4, 5
1
Graduate School of
Pharmaceutical Sciences,
University of Tokushima,
Tokushima, Japan
2
Department of Applied Chemistry,
Graduate School of
Engineering, Nagoya University,
Nagoya, Japan
3
MEXT Innovative Research Center
for Preventive Medical
Engineering,
Nagoya University,
Nagoya, Japan
4
Health Technology Research
Center,
National Institute of Advanced
Industrial Science and
Technology (AIST),
Takamatsu, Japan
5
Plasma Nanotechnology
Research Center,
Nagoya University,
Nagoya, Japan
Received February 8, 2007
Revised May 7, 2007
Accepted June 27, 2007
Research Article
Dynamic modification of poly(methyl
methacrylate) chips using poly(vinyl alcohol)
for glycosaminoglycan disaccharide isomer
separation
We describe a microchip electrophoresis (MCE) method for the assay of unsaturated di-
saccharides of chondroitin sulfates, dermatan sulfates, and hyaluronic acid (HA). Poly(vinyl
alcohol) (PVA) could be irreversibly adsorbed onto poly(methyl methacrylate) (PMMA)
substrates and this approach was applicable for dynamic coating. The characteristics of the
PMMA surface with PVA coating were evaluated in terms of the wettability, EOF, and
adsorption of 2-aminoacridone (AMAC)-labeled disaccharide. The water contact angle
decreased from ,737 on a pristine PMMA surface to ,37.57 on a PVA-coated surface,
indicating that the PVA coating increased hydrophilicity. EOF was reduced approximately
twofold and was relatively stable. Scanning electron microscopy and fluorescence micros-
copy images showed that adsorption of AMAC-labeled disaccharides was dramatically sup-
pressed. Using the PVA coating, baseline separation of two pairs of glycosaminoglycan
(GAG) disaccharide isomers, nDi-diSB/nDi-diSD and nDi-0S/nDi-HA, was achieved in
Tris-borate buffer within 130 s by MCE.
Keywords:
Dynamic coating / Glycosaminoglycan disaccharides / Microchip electrophoresis /
Poly(vinyl alcohol) DOI 10.1002/elps.200700088
3308 Electrophoresis 2007, 28, 3308–3314
1 Introduction
Proteoglycans (PGs) consist of glycosaminoglycan (GAG)
chains covalently linked to a “core” protein. Changes in
GAG structure result in different biological functions.
Chondroitin is crucial for cytokinesis, early embryogenesis
and epithelial morphogenesis in Caenorhabditis elegans [1].
Hyaluronan (HA) functions to induce vascular endothelial
growth factor, thereby acting as a negative regulator of
epithelial-to-mesenchymal transformation [2]. Chondroitin
sulfates (CSs) and dermatan sulfates (DSs) have been
implicated in the signaling pathways of various heparin-
binding growth factors and chemokines, and play critical
roles in the development of the central nervous system,
besides acting as receptors for various pathogens. The pre-
cise function in each of these scenarios is closely asso-
ciated with the particular sulfation pattern of the GAG
chains [3]. One recently reported example of CS/DS func-
tion is in binding of low-density lipoprotein (LDL) to artery
wall PGs as a key step in atherosclerosis [4], where the
degree of GAG sulfation determined affinity for LDL [5].
Therefore, profiling of GAGs on the microchip scale has
potential application in monitoring patient response to
therapy or in early diagnosis of disease. In particular,
microfluidic system profiling would be especially beneficial
for GAG profiling [6].
While conventional CE is now a relatively mature tech-
nology, microchip electrophoresis (MCE) is still in the early
stages of development with significant advantages in terms
of rapid analysis and facile integration of sample preparation
and derivatization steps. Cost and speed are also crucial
concerns, and, as such, MCE is expected to soon overtake
conventional CE as the primary mode of analysis [7]. It has
been shown that, for micro devices using electrophoretic
separation at a constant field, resolution is proportional to
the channel length [8], indicating that resolution becomes a
Correspondence: Yong Zhang, Department of Applied Chemistry,
Graduate School of Engineering, Nagoya University, Furo-cho,
Chikusa-ku, Nagoya 464-8603, Japan
E-mail: yl7173@gmail.com
Fax: 181-52-789-4666
Abbreviations: AMAC, 2-aminoacridone; CS, chondroitin sulfate;
DS, dermatan sulfate; FM, fluorescence microscopy; GAG, glyco-
saminoglycan; HA, hyaluronic acid; MC, methylcellulose; MCE,
microchip electrophoresis; PMMA, poly(methyl methacrylate);
PVA, poly(vinyl alcohol)
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

2. Electrophoresis 2007, 28, 3308–3314 Miniaturization 3309
significant challenge in MCE analysis. Many GAG isomers
differ in sulfate residue positions or different hydroxyl
orientations. Resolving such isomeric GAGs by CE has been
widely reported [9–12]. However, few papers have discussed
resolution of GAGs by MCE, probably due to the restrictions
in resolution due to short separation channels.
Microfluidic devices were initially fabricated in glass or
quartz due to their well-defined surfaces, good EOF prop-
erties, and excellent optical properties. However, for high-
aspect-ratio microstructures and for cheaper mass produc-
tion, alternative polymers or plastics are necessary [13].
Poly(methyl methacrylate) (PMMA) is a polymer that has
significant advantages, including being machinable, opti-
cally transparent and annealable, and having a high
dielectric strength. However, PMMA has a fairly hydro-
phobic surface, which can lead to biomolecule adsorption
issue. Thus, surface modification is critical. Approaches
include grafting of PEG onto the PMMA surface to sup-
press nonspecific adsorption of protein and to control cell
adhesion [14–16]. However, dynamic coating of PEG on
PMMA has been found to be unsatisfactory in certain
applications such as the analysis of fluorescent dye-labeled
oligosaccharides due to the low content of hydroxyl groups
in PEG [17].
Belder et al. [18] confirmed that glass microchips with
poly(vinyl alcohol) (PVA)-coated channels exhibit con-
siderably improved separation performance. This was at-
tributed to reduced analyte wall interaction and reduced
EOF. Chip coating was performed by forcing aqueous PVA
solution into the channel followed by emptying the channel,
then the PVA-layer was thermally immobilized. Wu et al.
[19] modified poly(dimethylsiloxane) (PDMS) with PVA to
suppress protein adsorption and EOF, and observed good
separation of basic proteins. The microchannel and flat
substrate were pretreated with oxygen plasma. PVA aque-
ous solution was introduced into the channels and the
reservoirs, and the PVA coating was thermally immobilized.
Kozlov et al. [20] found that PVA adsorbed irreversibly from
aqueous solutions onto hydrophobic solids in contact with
solutions. Lowering interfacial free energy (hydrophobic
interactions or displacement of water molecules from the
hydrophobic solid-water interface) drives the initial steps of
the adsorption. Crystallization processes drive formation of
continuous thin films of PVA. Since PMMA is also hydro-
phobic, there is a possibility that the PMMA substrate could
attract PVA from the solution. Furthermore, the irreversible
adsorption of PVA to the hydrophobic surface makes dy-
namic coating applicable. Here, we develop a method to
dynamically coat PVA onto a PMMA chip without thermal
immobilization for analyzing GAG disaccharide isomers by
MCE.
Dynamic coating of PVA onto PMMA chips was mon-
itored by measurement by EOF, water contact angles, and
adsorption of 2-aminoacridone (AMAC)-labeled unsaturated
GAGs. Using PVA-coated PMMA chips, the conditions of
resolving two pairs of GAG isomers, nDi-diSB and nDi-diSD,
nDi-0S and nDi-HA, were investigated. To our knowledge,
this is the first reported method to baseline-separate these
two pairs of isomers on a MCE scale.
2 Materials and methods
2.1 Chemicals and solutions
Unsaturated GAG disaccharides, 2-acetamido-2-deoxy-3-O-
(2-O-sulfo-b-D-gluco-4-enepyranosyluronic acid)-4-O-sulfo-D-
galactose (DDi-diSB), 2-acetamido-2-deoxy-3-O-(2-O-sulfo-b-
D-gluco-4-enepyranosyluronic acid)-6-O-sulfo-D-galactose
(DDi-diSD), 2-acetamido-2-deoxy-3-O-(b-D-gluco-4-enepyr-
anosyluronic acid)-D-galactose (DDi-0S), and 2-acetamido-2-
deoxy-3-O-(b-D-gluco-4-enepyranosyluronic acid)-D-glucose
(DDi-HA), were obtained from Seikagaku Corporation
(Chuo-ku, Tokyo, Japan) (Fig. 1). AMAC was purchased
from Molecular Probes (Eugene, OR), Tris from ICN Bio-
medicals Inc. (Aurora, OH), boric acid from Katayama
Chemical (Osaka, Japan), and acetic acid from Nacalai Tes-
que Inc. (Nakagyo-ku, Kyoto, Japan). DMSO, sodium cya-
noborohydride, and PVA were from Sigma (St. Louis, MO,
USA). Buffer pH values were adjusted by mixing appropri-
ate concentrations of Tris and boric acid solutions. Running
buffer with PVA was prepared by adding PVA to Tris-boric
acid and stirring until the solution appeared homogeneous
and transparent.
Figure 1. The GAG disaccharide structures and derivatization
procedure.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

3. 3310 Y. Zhang et al. Electrophoresis 2007, 28, 3308–3314
2.2 Dynamic coating of PMMA chips with PVA
For MCE, a buffer solution with PVA polymer was gradually
forced through four wells of the Hitachi PMMA i-chips using
a syringe. Except for the sample well, all other wells were
filled with the solution. The chips were then kept at room
temperature for about 15 min to ensure a thorough coating
with PVA. The step was followed by MCE analysis.
2.3 Measurement of contact angles
The static contact angles of the surfaces were measured with
a Contact-angle Meter (Kyowa Interface Science Co. Ltd, Sai-
tama, Japan). For contact angle measurement, the buffer so-
lution with PVA polymer was filled into the reservoirs of a
PMMA chip prepared for measurement. The chip was then
dried by evaporation at room temperature. The PMMA chip
was appropriately cut along the reservoirs to fit the size of the
platform and the surface was either modified with PVA or
was pristine. Approximately 2.0 mL water was added to the
surface of the PMMA chip. Contact angles were measured as
a mean of both left and right side of the water drop and
reported values are the average of eight separate droplets on
each substrate.
2.4 Measurement of EOF and viscosity
The measurements of EOF were carried out using the cur-
rent-monitoring method [21]. In our studies we used an
electrolyte consisting of 0.4 M Tris-sodium phosphate buffer
as the higher concentration buffer; the lower concentration
buffer was prepared by diluting the electrolyte with water to a
concentration of 0.36 M. The results are the average of three
time measurements. The viscosity of the polymer water so-
lution was measured with Viscolite 700 (Hydramotion, Eng-
land).
2.5 Detection by fluorescence microscopy and
scanning electron microscopy
Fluorescence microscopy (FM) detection was carried out
using an Axiovert 135T (Carl Zeiss, Tokyo, Japan), illumi-
nated by a 100-W mercury arc lamp. Images were captured
by a CCD camera (EB-CCD, Hamamatsu Photonics, Hama-
matsu, Japan). The sample labeled by AMAC was introduced
to the microchannel of both pristine and PVA-modified
PMMA chip. The solution was kept in the channel for 5 min
and then pumped out by a syringe. The microchip was
placed on the FM stage. The same region of the micro-
channel was monitored.
SEM detection was performed using a JSM-5600 (JOEL,
Tokyo, Japan). The surface of the reservoirs on the PMMA
chips modified with PVA was examined by SEM in the case
of no adsorption or adsorption. Uncoated PMMA reservoirs
were detected as controlled experiments.
2.6 Preparation and labeling of unsaturated GAG
disaccharides
Derivatization of DDi-diSB, DDi-diSD, DDi-0S, and DDi-HA
with AMAC was performed as described previously [22]
(Fig. 1). Briefly, 10 nmol of each standard D-disaccharide in
water was completely evaporated in a microcentrifuge tube at
12 0006g at 07C. Then 5 mL 0.1 M AMAC in glacial acetic
acid–DMSO (15:85) and 5 mL freshly prepared 1 M
NaBH3CN in water were added to each sample and the mix-
tures incubated at 457C for 3 h. The AMAC-labeled dis-
accharides were kept at –207C until use and then diluted to
the desired concentrations with buffer solution prior to
analysis.
2.7 MCE
All experiments were carried out on a HITACHI SV1100
with a light-emitting diode (LED) confocal fluorescence
detector that provided a median excitation source of 470 nm.
The PMMA microchips, i-chip 3 (Hitachi Chemical, Tokyo,
Japan), consisted of a simple cross channel 100 mm wide and
30 mm deep. Distances from the channel intersection to the
sample, sample waste, buffer, and buffer waste wells were
5.25, 5.25, 5.75, and 37.5 mm, respectively.
Buffer solutions were loaded into the microchannels
using a syringe. All microchip reservoirs were filled with
either buffer solution or sample solution prior to analysis.
High voltages were applied using an additional power supply
(Hitachi). For sample injection, around 600 V was applied to
the sample waste reservoir, while the other three reservoirs
were grounded. During separation, the same voltage was
applied to both sample and sample waste reservoirs. High
voltages were then applied to the buffer waste reservoir and
the buffer reservoir was kept grounded.
3 Results and discussion
3.1 Modification of microchannel surface
Adsorption of the analytes to the microchannel surface
always leads to poor resolutions and reproducibilities. The
water contact angle of the pristine PMMA microchip was
measured to be 71.37, close to that measured by H. Bi and co-
workers [16]. This indicated that the surface of PMMA is
highly hydrophobic and the surface energy of PMMA is low.
It was found that the analytes derivatized by the fluorescence
dye AMAC, which also has a hydrophobic character, strongly
adsorbed to the PMMA surface through hydrophobic inter-
actions (Fig. 2c). Thus, modification of the microchannel
surface before analysis was a prerequisite for further separa-
tion.
To investigate the wettability of the surfaces before and
after PVA coating, water contact angles for pristine and
modified PMMA substrates were measured (Table 1). The
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

4. Electrophoresis 2007, 28, 3308–3314 Miniaturization 3311
Figure 2. SEM images of pris-
tine PMMA substrate (a) and
PVA-modified PMMA substrates
(b), and adsorption (c and d) of
AMAC-labeled GAG disac-
charide samples. Fluorescence
microscopy (FM) images of
adsorption of AMAC-labeled
GAG disaccharide samples onto
the PMMA microchannel (e) and
PVA-modified PMMA micro-
channel (f).
Table 1. Contact angles for PVA-coated and untreated PMMA
substrates
Untreated PMMAa)
71.3 6 87
PVA-coated PMMAb)
37.5 6 87
a) The value was averaged from contact angle measurements of
eight separate water drops on the same PMMA chip.
b) The value was averaged from contact angle measurements of
ten separate water drops on the same PMMA chip.
average static contact angle for PVA-coated PMMA surface
was approximately 37.57. The decrease in contact angle indi-
cated that the surface of the PVA-coated PMMA substrate
was wettable. Decreased contact angles also demonstrated
that PVA successfully covered the PMMA surface. Hydro-
phobic substrates, such as PMMA and PDMS, always display
a negative zeta potential at pH 7. Beattie [23] hypothesized
that surface charge arises from enhanced autolysis of water
at hydrophobic surfaces, with the preferential adsorption of a
hydroxide ion. Thus, PVA was attracted to the PMMA surface
by hydrogen bonding.
To examine the adsorption of AMAC-GAG on PMMA
substrates, we compared the surface of PMMA chips with or
without PVA coating using SEM detection (Figs. 2a and b).
Figure 2c shows that the hydrophobic sample labeled by
AMAC strongly adsorbed onto the pristine PMMA surface.
However, the adsorption was dramatically suppressed with a
PVA coating (Fig. 2d). Similar results were observed for the
microchannel using FM detection. As shown in Fig. 3e, the
sample adsorbed strongly on the surface of the untreated
PMMA microchannel, showing bright fluorescence. In con-
trast, no obvious fluorescence was observed for the PVA-
coated microchannel (Fig. 2f), indicating significantly less
adsorption.
EOFs were measured in the range of pH 7.0–11.0
(Fig. 3). The EOF mobility of the pristine PMMA micro-
channel was unstable with increasing pH. The EOF mobility
of the PVA-treated PMMA microchannel was lower and
more stable. This relatively stable surface of the PMMA
microchannel is critical for reproducible analysis.
3.2 High-performance analysis of GAG disaccharide
isomers
Although the separation efficiency obtained by MCE is gen-
erally as high or slightly higher than that in conventional CE,
the obtainable resolution is often lower than conventional CE
due to the shorter channel length for separation. Therefore,
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

5. 3312 Y. Zhang et al. Electrophoresis 2007, 28, 3308–3314
Figure 3. Influence of pH on the mEOF values of the untreated
PMMA (o), and PVA-modified (D) channel.
analysis of sugar isomers by MCE is still under development.
Here, we have developed a method of MCE to analyze two
pairs of GAG disaccharide isomers, nDi-diSB/nDi-diSD and
nDi-0S/nDi-HA.
In our previous study, three monosulfate GAG positional
isomers nDi-UA2S, nDi-4S, and nDi-6S, were baseline-
separated by MCE. Increased resolutions were achieved
when 1,4-dioxane was added into the buffer [19]. The three
analytes were presumably surrounded by different solvation
layers due to different steric confirmations. The difference in
thickness of the layers gave rise to different masses for each
isomer, which was manifested as different electrophoretic
mobilities. Here, we first tested the ability of MCE to resolve
nDi-diSB/nDi-diSD and nDi-0S/nDi-HA under conditions
found to be successful in resolving mono-sulfate GAG posi-
tional isomers. The results showed that nDi-diSB/nDi-diSD
were partly resolved and nDi-0S/nDi-HA co-eluted (data not
shown). It was deduced that steric conformation of the ana-
lyte was critical to the forming of solvation layers. Two sulfate
groups in nDi-diSB/nDi-diSD led to a tendency towards an
identical steric conformation with the same solvation layer.
An opposite orientation of the hydroxyl group in nDi-0S/
nDi-HA led to an increased difference in steric conforma-
tion. To summarize, the solvation effects caused by 1,4-diox-
ane were unable to provide enough resolution for analyzing
nDi-diSB/nDi-diSD and nDi-0S/nDi-HA by MCE.
3.2.1 Enhancing the resolution of GAG disaccharide
isomers by PVA coating
Not only did the dynamic coating with polymer PVA led to a
hydrophilic character, but also improved the resolution of
AMAC-labeled GAG disaccharide isomers on the PMMA
chips. As shown in Table 2, when methylcellulose (MC) in
the buffer was substituted with PVA, the resolutions of GAG
disaccharide isomers were enhanced to different extents.
nDi-di4S/nDi-di6S(nDi-UA2S), nDi-diSB/nDi-diSD were
Table 2. Effect of the polymers in the buffer on the Rs of AMAC-
labeled GAG disaccharide isomersa)
Rs DDi-4S/
DDi-6S
DDi-4S/
DDi-UA2S
DDi-6S/
DDi-UA2S
DDi-diSB/
DDi-diSD
DDi-0S/
DDi-HA
MCb)
0 0 0 0.74 0
PVAc)
1.52 1.52 0 2.19 0.89
a) n = 3.
b) Conditions: 0.1 M Tris-phosphate buffer pH 8.5, Esep =161 V/
cm, 0.5% MC.
c) Conditions: 0.1 M Tris-phosphate buffer pH 8.5, Esep = 439 V/
cm, 2% PVA.
baseline-separated and nDi-0S/nDi-HA were partly sepa-
rated with PVA coating. Previously, the observed resolution
of these analytes was relatively low using an MC coating,
which was often performed on PMMA chips [17, 19]. With
the same concentration (2% water solution), the viscosity of
MC solution was 136.4 cp. However, the viscosity of PVA so-
lution was only 2.1 cp. The lower viscosity of PVA solution
was more applicable in the operation of MCE. The relatively
lower number of hydrogen groups in PVA molecule com-
pared to MC results in a longer dynamic coating process
(15 min), while MC coating is effectively immediate.
We further investigated the effect of PVA concentration
on resolution. As shown in Table 3, the resolution of nDi-
diSB/nDi-diSD decreased with PVA concentration, while that
of nDi-0S/nDi-HA increased with PVA concentration. The
reason is not clear, although it may be possible that PVA is
involved in complex formation with the analytes. A 2% PVA
solution was optimal for further analysis. However, it is
worth noting that baseline separation for all pairs of isomers
could not be obtained in Tris-phosphate buffer just by
adjusting polymer coating (data not shown).
Table 3. Effect of the PVA concentration in the buffer on the Rs of
AMAC-labeled GAG disaccharide isomersa)
Rs DDi-diSB/DDi-diSD DDi-0S/DDi-HA
0.5% PVA b)
0.74 4.28
1.0% PVA b)
1.19 3.91
2.0% PVA b)
1.63 2.51
3.0% PVA b)
1.73 1.52
a) n = 3.
b) Conditions: 0.4 M Tris-borate buffer, pH 8.5, Esep = 650 V/cm.
3.2.2 Forming borate ion complexes to further
improve the resolution
To obtain a higher resolution, especially for nDi-0S/nDi-
HA, Tris-borate buffer was investigated. Borate buffers are
often added to carbohydrate solutions to form borate com-
plexes. The complex formation can be described by the fol-
lowing equations:
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

6. Electrophoresis 2007, 28, 3308–3314 Miniaturization 3313
BÀ
þ L , BLÀ
þ H2O (1)
BLÀ
þ L , BLÀ
2 þ H2O (2)
whereL isthepolyolligandandBtetrahydroxyborate,B[OH]4
–
.
For any particular sugar molecule, complex stability
depends strongly on the configuration, the number of hy-
droxyl groups, and the presence of substituents [24]. In
theory, the two pairs of GAG isomers, DDi-0S/DDi-HA and
DDi-diSB/ DDi-diSD, have the same mass/charge ratio.
Accordingly, the same mass/charge ratio leads to the same
mobility of GAG isomers in CZE and makes separation
impossible. However, the mass/charge ratios of GAG iso-
mers change when forming complexes with borate,
depending on the stability of the borate complex: higher
stabilities result in smaller mass/charge ratios. The
reduced ring of GAG disaccharides opened after derivati-
zation. Fluorescence dye (AMAC) was then attached on C1
and a hydroxyl group was attached on C5 of the N-acetyl
galactosamine residue. The hydroxyl groups on C4/C6 and
C5 of N-acetyl galactosamine residue were involved in
borate complexation. The stability of complexes was deter-
mined for different positions of the hydroxyl group. In
particular, DDi-0S and DDi-HA differ in the cis or trans
position of the hydroxyl group on C4 of the N-acetyl galac-
tosamine residue (Fig. 1). We speculated that the hydroxyl
group on C5 was orientated upward such that cis 1,2-diols
were more stable than trans 1,2-diols. Furthermore, lower
mass/charge ratios gave rise to higher mobilities of DDi-0S
than DDi-HA (Fig. 4a). As the degree of dissociation of
boric acid is a function of pH, the maximum difference in
mobilities of DDi-0S and DDi-HA was obtained at pH 8.5.
Similar mobilities of DDi-0S and DDi-HA were observed at
pH 11.0. This may have been due to trihydroxylated Tris
buffering most of the borate, resulting in lower GAG di-
saccharides complex formation (Fig. 4a).
DDi-diSB and DDi-diSD have two sulfated groups on
each monosaccharide ring. Both DDi-diSB and DDi-diSD
have one of these sulfated groups on C2 of the uronic acid
residue. DDi-diSB has the second sulfated group on C4 of
the N-acetyl galactosamine residue, while DDi-diSD has the
second sulfated groups on C6 of the N-acetyl galactosamine
residue (Fig. 1). As shown in Fig. 4b, DDi-diSB had a faster
mobility than DDi-diSD, indicating that steric hindrance
caused by sulfation may be responsible for the stability of
the complex. Similar mobilities of DDi-diSB and DDi-diSD
were obtained at pH 10.0 because of surplus Tris in the
buffer.
The resolutions of these two pairs of GAG disaccharide
isomers were calculated at different pH (Fig. 4). At pH 8.5,
the resolution of nDi-diSB/nDi-diSD was 1.60 and the reso-
lution of nDi-0S/nDi-HA was 2.12. All of them were base-
line separated (Rs .1.50). Thus, pH 8.5 was chosen for fur-
ther analysis. Under optimal conditions, these two pairs of
GAG isomers were well resolved within 130 s by MCE
(Fig. 5).
Figure 4. Effect of buffer pH on the mobility of DDi-0S/DDi-HA
derivatives (a) and DDi-diSB/ DDi-diSD (b). Experimental condi-
tions: 0.4 M Tris-borate, 2% PVA, Esep = 650 V/cm.
Figure 5. Electropherograms of AMAC-labeled DDi-diSB/ DDi-
diSD and DDi-0S/DDi-HA. Experimental conditions: pH 8.5, other
conditions as in Fig. 4.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

7. 3314 Y. Zhang et al. Electrophoresis 2007, 28, 3308–3314
The reproducibility of separation was evaluated by calcu-
lating the RSD of migration time for each GAG disaccharide.
As shown in Table 4, RSDs of less than 1.9% proved this
method was fairly reliable. The detection limit for each GAG
disaccharide was in the range of 1.58610–7
–3.16610–7
mol/
L, which was around ten times lower than that reported for
UV detection [11].
Table 4. Reproducibility of migration time and LOD of GAG di-
saccharide derivativesa)
DDi-diSB DDi-diSD DDi-0S DDi-HA
Migration time
(RSD%)b)
1.87 1.72 1.63 1.36
LOD (10–7
M)c)
2.25 2.25 3.16 1.58
a) Conditions: 2% PVA, other conditions as in Table 3.
b) RSD, n = 10.
c) LOD, S/N of 3.
4 Concluding remarks
A dynamic coating of PVA on the PMMA surface has been
demonstrated. Decreased water contact angles indicated that
the pristine PMMA surface was hydrophobic and the PMMA
surface with PVA coating was relatively hydrophilic. The
EOF of the PVA-modified PMMA microchannel was reduced
approximately twofold and was relatively stable within the
pH range 7.0–11.0. The increased wettability led to a sup-
pression of adsorption of AMAC-labeled GAG disaccharide
samples on the PMMA surface. This dynamic coating meth-
od also contributed to an improved resolution of GAG dis-
accharide isomers. Forming complexes in Tris-borate buffer
also further improved the resolution for separation of two
pairs of GAG disaccharide isomers, nDi-diSB/nDi-diSD and
nDi-0S/nDi-HA by MCE. Under optimized conditions,
these disaccharide isomers were baseline-separated within
130 s. To our knowledge, this is the first report of baseline
separation of these two pairs of isomers by MCE.
The present work was partially supported by MEXT Innova-
tive Research Center for Preventive Medical Engineering, Nagoya
University, supported by Ministry for Education, Culture, Sports,
Science, and Technology, Japan (MEXT), a Grant from New
Energy and Industrial Technology Development Organization
(NEDO) of the Ministry of Economy, Trade and Industry, Japan,
and a Grant-in-Aid for Scientific Research from MEXT.
5 References
[1] Hwang, H. Y., Olson, S. K., Esko, J. D., Horvitz, R., Nature
2003, 423, 439–443.
[2] Rodgers, L. S., Lalani, S., Hardy, K. M., Xiang, X. et al., Circ.
Res. 2006, 99, 583–589.
[3] Sugahara, K., Mikami, T., Uyama, T., Mizuguchi, S. et al.,
Curr. Opin. Struct. Biol. 2003, 13, 612–620.
[4] Skalen, K., Gustafsson, M, Rydberg, E. K., Hulten, L. M. et al.,
Nature 2002, 417, 750–754.
[5] Sambandam, T., Baker, J. R., Christner, J. E., Ekborg, S. L.,
Arterioscler. Thromb. 1991, 11, 561–568.
[6] Whitesides, G. M., Nature 2006, 442, 368–373.
[7] Li, S. F., Kricka, L. J., Clin. Chem. 2006, 52, 37–45.
[8] Schmalzing, D., Adourian, A., Koutny, L., Ziaugra, L. et al.,
Anal. Chem. 1998, 70, 2303–2310.
[9] Carney, S. L., Osborne, D. J., Anal. Chem. 1991, 195, 132–
140.
[10] Scapol, L., Marchi, E., Viscomi, G. C., J. Chromatogr. A 1996,
735, 367–374.
[11] El Rassi, Z., Postlewait, J., Mechref, Y., Ostrander, G. K.,
Anal. Biochem. 1997, 244, 283–290.
[12] El Rassi, Z., Electrophoresis 1997, 18, 2400–2407.
[13] Soper, S. A., Ford, S. M., Qi, S., McCarley, R. L. et al., Anal.
Chem. 2000, 72, 642A–651A.
[14] Liu, J., Pan, T., Woolley, A. T., Lee, M. L., Anal. Chem. 2004,
76, 6948–6955.
[15] Patel, S., Thakar, R. G., Wong, J., McLeod, S. D., Li, S., Bio-
materials 2006, 27, 2890–2897.
[16] Bi, H., Meng, S., Li, Y., Guo, K. et al., Lab Chip 2006, 6, 769–
775.
[17] Dang, F., Zhang, L., Hagiwara, H., Mishina, Y., Baba, Y., Elec-
trophoresis 2003, 24, 714–721.
[18] Belder, D., Husmann, H., Warnke, J., Electrophoresis 2001,
22, 666–672.
[19] Wu, D., Luo, Y., Zhou, X., Dai, Z., Lin, B., Electrophoresis
2005, 26, 211–218.
[20] Kozlov, M., Quarmyne, M., Chen, W., McCarthy, T. J., Mac-
romolecules 2003, 36, 6054–6059.
[21] Huang, X. H., Gordon, M. J., Zare, R. N., Anal. Chem. 1988,
60, 1837–1838.
[22] Zhang, Y., Ping, G., Zhu, B., Kaji, N. et al., Electrophoresis
2007, 28, 414–421.
[23] Beattie, J. K., Lab Chip 2006, 6, 1409–1411.
[24] Hoffstetter-Kuhn, S., Paulus, A., Gassmann, E., Widmer, H.
M., Anal. Chem. 1991, 63, 1541–1547.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com