Biosensors and Bioelectronics 54 (2014) 297305

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

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

Single layer linear array of microbeads for multiplexed analysis

of DNA and proteins
Wanqing Yue a,b, Heng Zou a,b, Qinghui Jin c, Cheuk-Wing Li d, Tao Xu a,b, Huayang Fu a,b,
Lawrence C.H. Tzang e, Hongyan Sun a,b, Jianlong Zhao c, Mengsu Yang a,b,n
a

Department of Biology and Chemistry, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong SAR, People's Republic of China
Key Laboratory of Biochip Technology, Biotech and Health Centre, Shenzhen Research Institutes of City University of Hong Kong, Shenzhen,
People's Republic of China
c
State Key Laboratory of Transducer Technology, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, Shanghai
200050, People's Republic of China
d
State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Macau, People's Republic of China
e
Multigene Diagnostics Limited, 5/F, Biotech Centre 2, 11 Science Park Avenue, Hong Kong Science Park, N. T., Hong Kong SAR, People's Republic of China
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 29 July 2013
Received in revised form
16 October 2013
Accepted 21 October 2013
Available online 1 November 2013

In this study, a microuidic platform was developed to generate single layer, linear array of microbeads
for multiplexed high-throughput analysis of biomolecules. The microuidic device is comprised of eight
microbead-trapping units, where microbeads were immobilized in a linear array format by the exertion
of a negative pressure in the control channel connected to each sieving microstructure. Multiplexed
assays were achieved by using a mixture of different spectrally-encoded microbeads functionalized with
specic probes, followed by on-chip reaction and detection. The microuidic-based microbeads array
platform was employed for multiplexed analysis of DNA and proteins, as demonstrated by the
simultaneous discrimination of four HPV genotypes and the parallel detection of six different proteins.
Compared with the off-chip protocols, the on-chip analysis exhibited better reaction efciency, higher
sensitivity and wider linear detection range. Visual inspection and identication of functionalized
microbeads were facilitated by the single layer arrangement of microbeads so that accurate data
acquisition can be performed during the detection process.
& 2013 Elsevier B.V. All rights reserved.

Keywords:
Microuidics
Microbead array
Multiplexed assay

1. Introduction
Early screening and diagnosis of diseases require the development of sensitive, reliable and inexpensive high-throughput assays
(Braeckmans et al., 2002; Derveaux et al., 2008; Hsu et al., 2009;
Jain, 2005; Mani et al., 2009; Wilson et al., 2006). High-throughput
analysis can be achieved by parallel screening of one analyte for
multiple samples, or by simultaneous detection of multiple analytes from one sample, or by a combination of both (Situma et al.,
2006). Bead-based assays, in which different specic probes are
tethered on encoded particles via chemical or physical means
(Derveaux et al., 2008), have demonstrated feasibility and versatility in numerous multiplexed applications including genotyping
(Zhang et al., 2011), gene expression proling (Lawrie et al., 2006),
enzymatic assays (Holmes et al., 2007) and immunoassays (Puig
et al., 2001). Bioanalytes, including cytokines, cell signaling molecules,
n
Corresponding author at: Department of Biology and Chemistry, City University
of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong SAR, People's Republic of
China. Tel.: 852 3442 7797; fax: 852 2788 7406.
E-mail address: bhmyang@cityu.edu.hk (M. Yang).

0956-5663/$ - see front matter & 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bios.2013.10.034

DNA and proteins from human serum samples, cerebrospinal uid and
synovial uid, have been simultaneously analyzed in the bead-based
multiplexed platform (Diercks et al., 2009; Yu et al., 2010; Zhang et al.,
2012a, 2012b).
Bead-based assays usually use a ow cytometer as the detection system to analyze microparticles rapidly, but the high cost
and lack of portability hinder its wide applications. An alternative
approach is to integrate bead-based assays into microuidic
devices, which has shown several advantages over conventional
ow cytometer (Lion et al., 2004; Tudos et al., 2001; Vo-Dinh and
Cullum, 2000). With functionalized particles being immobilized
within microchannels, continuous uidic ow delivers fresh analyte solution to the reaction site, where a high concentration
gradient of analyte is maintained to enhance mass transport in
short diffusion distance. The miniaturized systems also enable low
reagent consumption, high surface-to-volume ratios, and rapid
diffusion times (Dittrich et al., 2006; Tanaka et al., 2007; Vilkner
et al., 2004; Zhang et al., 2006), thus enhancing the efciency and
sensitivity of analysis. Many microuidic-based platforms have
been developed for bioanalysis, including cell-based assays (Xu
et al., 2010, 2013; Yue et al., 2013), nucleic acid analysis (Huang

298

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

et al., 2006), protein engineering (Joensson and Andersson-Svahn,

2011), mutation detection (Han et al., 2011) and point-of-care
testing (Holland and Kiechle, 2005; Tudos et al., 2001). Recently,
the combinations of bead-based assays with microuidic systems
have attracted increasing attention due to the high-throughput of
the microuidic systems and the multiplexing capability of barcoded microbeads (Dendukuri et al., 2006;Dunbar and Jacobson,
2000; Nicewarner-Pena et al., 2001; Pregibon et al., 2007). Various
bead-based bioassays integrated with microuidic systems have
been reported, including discrimination of single-nucleotide mismatches (Ali et al., 2003; Ng et al., 2007), genotyping of hepatitis B
and HPV viruses (Zhang et al., 2010; Zhang et al., 2011) and
inuenza A virus (Lien et al., 2011). In these experiments,
suspended microbeads were either sequentially detected via ow
cytometer, or immobilized in trapping structures followed by
imaging a cluster, often multilayers of microbeads. In both
approaches, individual microbead could not be visualized and
quantied accurately. The generation of highly ordered single
layer array requires ne microstructures with delicate control by
hydrodynamic (Li et al., 2006; Sochol et al., 2011) and electrical
(Chiou et al., 2005; Taff and Voldman, 2005) means or optical
trapping (Daria et al., 2004; MacDonald et al., 2003). Therefore,
there is a great need for the development of simple microuidic
structures that can be easily integrated with single layer, individually addressable bead-based assays.
In this study, we developed a microuidic device consisting of
sieving microstructures by single step photolithography for the
formation of single layer array of microbeads for multiplexed
bioanalysis. Without the need of specic geometry design of
microchannels and peripheral equipment for liquid manipulation,
microbeads with the diameter of several micrometers could be
immobilized and aligned automatically into a linear array alongside the microchannel simply by exerting negative pressure from a
manually-controlled syringe. Several protocols were established
for developing microuidic bead-based assays which include
functionalization of microbeads with specic probes, formation
of single layer array of microbeads in the microuidic device,
on-chip reaction, image acquisition and data analysis. We have
demonstrated the feasibility using the microuidic device to
develop multiplexed assays for discrimination of viral genotypes
and for parallel analysis of a panel of protein markers. This
microuidic device provides a universal platform for on-chip
bead-based assays by combining the rapid binding kinetics of
homogeneous microbeads assay and the efcient liquid handling
capability of microuidics, and can be used to develop other
multiplexed assays for high-throughput analysis of biomolecules.

2. Materials and methods

2.1. Microuidic chip fabrication
PDMS-based microuidic devices were prepared by molding
PDMS against a printed circuit board (PCB) master (Li et al., 2003;
Yue et al., 2011). Briey, the microchannels were designed in a
CAD program (CorelDRAW 13.0, Corel Corporation, UK) and
printed on a transparent lm which served as a photomask by a
3000 dpi commercial printer. The photomask was placed atop of a
PCB (Kinsten glass epoxy single sided, Chiefskill, Taiwan), followed
by UV exposure for 110 s by a standard PCB exposure unit (KVB-30
exposure unit, Chiefskill, Taiwan). After development in a 1:20 PCB
developing reagent (Kinsten Corp., Taiwan), the developed PCB
was wet etched in a 1:2 ferric chloride solution. The etching time
varied from 50 min to 70 min. The PCB was thoroughly rinsed by
running water and the residual photoresist was removed by acetone,
with the remaining protruding positive copper microstructures to

serve as the master. The surface prole of the PCB master was
evaluated by a step proler (Ambios XP-2 high resolution surface
proler, Ambios Technology, Santa Cruz, CA).
The PCB master was covered by degassed PDMS prepolymer
(10:1 base: curing agent, Sylgard 184, Dow Corning, Midland, MI)
and incubated at 65 1C for 2 h. The cured PDMS replica was
carefully peeled off from the master, and the reservoirs O and I
were punched with 3 mm in diameter for liquid injection while
reservoirs #18 were punched with 1.5 mm in diameter for
pressure regulation (Fig. 1a). The PDMS replica featured with
micropatterns and a piece of cleansed glass slide were oxidized
in a plasma cleaner for 2 min and irreversibly sealed to form a
microdevice (Fig. 1b).
2.2. Microfuidic ow simulation
The pressure and ow velocity prole in the microchannels
were simulated using computational software COMSOL Multiphysics (V4.2.0.105, Burlington, MA, USA). Each model was composed of grid patterns with 583341 nodes. All the microchannels
were simplied in rectangular shape and the microsieves were
assumed as smooth protruding patterns with lower height compared with the regular microchannel intervals, leaving the gap
connection between the main channel and the dead-ended pressure control channel in the sealed device. One trapping unit
(Fig. 1a) was schematic constructed, consisting of multi-height
structures: the main channel and the dead-ended pressure control
channel were both 26 m in height and 150 m in width, and the
connection between the two channels was 2 m in height and
70 m in width. For three-dimensional ow through microchannels, the governing equations are continuity equation and Navier
Stokes equations, respectively (Rawool et al., 2006)
V 0

VV  P 2 V

where V is the velocity of the ow, P is the pressure, is the ow

density and is the ow viscosity. No slip condition was imposed
at all the boundaries and constant inlet and outlet pressure
boundary condition was assumed in the simulation model. Without negative pressure, the ow was driven by the droplet of 5 L
solution in the inlet ( 50 Pa), while a negative pressure
( 20 kPa) was exerted to the dead-ended channel in order to
trap the microbeads.
2.3. Immobilization of the microbeads in the microchip
The microuidic chip was pretreated with 5% bovine serum
albumin (BSA) at 4 1C overnight in order to minimize nonspecic
adsorption (Weibel et al., 2005; Xu et al., 2010). The multiplexed
microbeads mixture was prepared by transferring 1.0  103 each
type of microbeads with Tris-EDNA (TE) buffer to adjust the bead
density to 1.0  105 beads/mL. To immobilize the microbeads in
the chip, 5 L of microbeads mixture was added in the inlet I and
the negative pressure was exerted to the trapping unit from vial
#8 to #1 separately to ensure evenly immobilization of the
microbeads in the eight trapping units.
2.4. Multiplexed microbeads mapping
The spectrally-encoded carboxyl group functionalized microbeads purchased from Bioplex (cat#177-0506, BIO-RAD, USA) were
capsulated with two uorescent dyes at the emission wavelength
of 650 nm and 710 nm, respectively. By varying the ratios of two
internal dyes, different types of the microbeads were identied
based on the differences in the uorescence intensities at the

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

299

multiplexed beads mixture. Among the 23 types of microbeads,

4 types were used to couple oligonucleotides for virus genotyping
and 6 types were selected to immobilize antibodies for protein
detection by immunoassays.
2.5. Functionalization of microbeads
For virus genotyping, four types of specic DNA probes were
designed to discriminate human papilloma virus (HPV) genotypes
6, 11, 16 and 18 as shown in Table S2. The capture probes coupled
on the surface of the microbeads, which were named HPV6P,
HPV11P, HPV16P, and HPV18P, had several base differences and
showed different hybridization dynamics between matched and
mismatched target templates, resulting in successful discrimination of virus genotypes. For each coupling reaction, 5.0  106 of
stock microbeads were re-suspended in 50 L of 0.1 M MES buffer
(pH 4.5) by vortex and sonication for approximately 20 s and
mixed with 2 L of 2 nmol/L capture oligonucleotides in a microfuge tube. 2.5 L of freshly prepared 10 mg/mL EDC was added in
the tube and incubated for 30 min at room temperature, which
was repeated once for complete coupling reaction. For protein
detection, six types of microbeads were conjugated with antibodies of leptin, CEA, CA125, IL-8, HGF and HE4, respectively by using
EDC/NHS cross-linking method (Santra et al., 2001). The mixture
of carboxyl group functionalized microbeads was reacted with
50 L 1 mg/mL antibody solution containing 1.8 mg EDC and
3.5 mg NHS for 2 h at 25 1C. After the coupling reaction, microbeads were washed by vortex and micro-centrifugation at
48000 rpm for 2 min in 1.0 mL of 0.02% Tween-20 and 1.0 mL
of 0.1% SDS, respectively. The microbeads were stored in 100 L of
TE buffer (pH 8.0) at 28 1C in the dark.
2.6. On-chip genotyping of HPV virus

Fig. 1. (a) Schematic design of the microuidic device and enlarged view of one
trapping unit. (b) Optical image of the microuidic device lled with Rose Bengal.
The size of device was 40  25 mm. (c) Optical images of one trapping unit on the
PCB master etched for 70 min. Scale bar: 100 m. (d) Surface prole of four trapping
units on the PCB master etched for 70 min.

distinct emission wavelengths. To establish the microuidic-based

multiplexed analysis, we immobilized various types of the
microbeads in separate chips to form single layer beads array
(Table S1), and analyzed the microbeads individually (Fig. S1a).
By plotting the uorescence intensities at the emission wavelengths of 650 nm and 710 nm, 23 types of microbeads were
identied and mapped as shown in Fig. S1b, which were used as
the standard pool for microbeads recognization. The settings of
the confocal microscope were xed and used in the following
experiments for identication of each type of microbeads in the

The HPV plasmid DNA was amplied by PCR using four pairs of
primers: HPV6F and HPV6R, HPV11F and HPV11R, HPV16F and
HPV16R, HPV18F and HPV18R (Table S2). For each PCR reaction,
the amplication was carried out using a commercial thermal
cycling machine (Rotor-Gene Q, QIAGEN) in 20 L of reaction
mixture containing 10 L of 2  master mix (Premix Taq version
2.0 from Takara, cat#R004A), 4 pmol of each primers, and 2 L of
DNA template. After denaturation step at 95 1C for 5 min, the
amplication was achieved for 40 cycles at 95 1C for 15 s, 52 1C for
30 s, 72 1C for 30 s, and a nal extension at 72 1C for 5 min.
Clinical samples of cervical scraps that were positive for HPV
infection were also analyzed. HPV genomic DNA was extracted
from the samples and amplied targeting the L1 gene of HPV with
biotinylated consensus primers in reaction volume of 40 mL. The
thermal cycling prole was 50 1C for 5 min, 95 1C for 10 min;
followed by 45 cycles of denaturation at 95 1C for 30 s, annealing
at 52 1C for 45 s, and extension at 65 1C for 30 s; and then a nal
extension at 65 1C for 5 min (Yip et al. 2010). The PCR products
were genotyped using the microuidic-based single layer bead
array platform and the genotypes were veried by DNA sequencing.
To discriminate the four types of HPV genotypes, ve microuidic chips were pre-loaded with the microbeads mixture
containing four types of microbeads coupled with HPV6P, HPV11P,
HPV16P, and HPV18P probes, respectively. Four separate microuidic devices were used for identify four HPV genotypes, respectively, and one chip was used as control utilizing deionized water.
After the formation of the single layer microbeads array, the
microuidic chip was lled with 10 L of pre-warmed neutralization buffer.1 L of PCR product of each HPV genotype was mixed
with 5 L of 0.1 mol/L NaOH and kept at room temperature for
10 min, followed by mixing with 10 L of neutralization buffer. The
mixture was transferred into the chip, and incubated at 55 1C for

300

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

half an hour. The chip was washed with TE buffer twice, followed
by incubation with 15 L of 0.5 ng/mL streptavidin-phycoerythrin
(SA-PE) for 10 min in the dark. After washing with TE buffer, the
chip was ready for imaging.
In order to compare the sensitivity of microuidic-based
microbeads array to that of commercial DNA hybridization
approach, off-chip hybridization and on-chip hybridization of
probe-bound microbeads with target DNA of different HPV genotypes were carried out. Off-chip hybridization was carried out in a
total volume of 30 L in 0.2 mL microtubes at room temperature
for 10 min followed by 55 1C for half an hour. For each trial, the
reaction mixture consisted of 1 L of 25 pM to 3600 nM target
DNA, 5 L of 0.1 mol/L NaOH, 10 L of the microbeads coupled with
specic probe to the target DNA, and 14 L of the neutralization
buffer. After washing by 100 L TE buffer twice, the microbeads
were incubated with 100 L of 0.5 ng/mL SA-PE for 10 min in the
dark. After washing by 100 L TE buffer, the microbeads were read
by Bio-PlexTM 200 System (BIO-RAD, USA).
The off-chip discrimination of four HPV genotypes was also
carried out in the same experimental condition as the on-chip
detection described above. To eliminate the effect of different
detection systems, after the hybridization and SA-PE labeling in
the tube, part of the microbeads were immobilized on the chip for
visualization, while the rest of the microbeads were read off-chip.
The efciency of on-chip and off-chip hybridization was compared
by the signal-to-noise ratio (SNR) obtained in the two different
detection systems.
2.7. On-chip detection of proteins
The six-plexed microbeads were immobilized in nine microuidic chips: six chips were used for working standard mixtures
containing the above six proteins prepared by serial dilution, and
the other three were used for detection of buffer with no standard
as control, DMEM normal medium, and DMEM conditioned medium
cultured with A549 cells (adenocarcinomic human alveolar basal
epithelial cells, ATCC) for 48 h, respectively. Mixtures containing 24
types of proteins with known concentrations (Table S3) including
target anlyates of HE4, IL8, HGF, CA125, leptin and CEA were serial
diluted as standard solutions (Table S4). To perform the on-chip
protein analysis, 5 L of analyte (control, standards, or samples)
was added into the microuidic chip and incubated overnight at
4 1C. After washing by wash buffer twice, 5 L of detection
antibodies was added and incubated for 30 min at room temperature. The chip was then washed with washing buffer and incubated with 15 L of 0.5 ng/mL SA-PE for 10 min in the dark.
2.8. Data acquisition and analysis
The images of microbeads immobilized in the chip were
captured by a confocal laser scanning microscope (Leica TCS SPE,
Wetzlar, Germany) with excitation wavelength of 635 nm, and
photomultiplier tube 1 (PMT1) for 640660 nm emission and
PMT2 for 700720 nm emission to identify different types of
microbeads. The detection signals for the genotyping assay and
the immunoassays were recorded by laser with excitation wavelength of 488 nm and PMT3 for 560590 nm emission.
The uorescent intensities were retrieved by GenePix Pro (6.0,
Axon Instruments Inc, C.A., USA). The uorescence intensity of
each microbead (F) at different wavelengths was calculated as
F F i  F Blank

where F i is the measured uorescence intensity of one microbead,

F Blank is the mean uorescence intensity (n 50) of the microbeads
mixed with blank solution for the corresponding experiments.
The uorescence intensities of a single microbead at the emission

of 650 nm (F650) and 710 nm (F710) were measured and logarithmically plotted to generate the distribution map for microbeads recognization (n 50 for each type). The uorescence
intensities of each microbead at the emission of 575 nm were
counted only if it was attributed to the corresponding beads region
according to the distribution map, and the averaged uorescence
intensities of each type of microbeads (F575) were calculated as
the detection signal to quality and quantify the analytes of
interests.

3. Results and discussion

3.1. Design and fabrication of the microuidic device
The microuidic device was made by sealing the PDMS slab
with the design shown in Fig. 1a to a clean glass slide. The chip
consisted of one main microchannel (150 m in width, 26 m in
height) with eight trapping units for the entrapment of microbeads, and eight dead-ended microchannels (70 m in width, 26 m
in height) connected to the main channel by the sieving microstructures (70 m in width, 2 m in height) in between. The whole
device was 40  25 mm in size (Fig. 1b).
The sieving microstructure was fabricated by single-step photolithography as described previously (Yue et al., 2011). Briey, the
etching process of copper layer on the PCB was accelerated at the
edges of the features owing to the intrinsic geometry-dependent
etching rates in wet etching, by which phenolic support sheet with
rough features in the PCB was rst revealed and uidic connection
between two closely designed channels could be achieved in the
corresponding PDMS replica. The determining parameters for the
formation of the microsieves are the geometry of the micro-patterns
and the etching time. For the design in Fig. 1a, we varied etching time
from 50 min to 70 min (Fig. S2). It was found under the bright eld
microscope that the sieving microstructures were completely formed
between the protruding microchannels on the PCB master after
etching for 70 min (Fig. 1c). The rough features of the microsieves
were veried by the surface prole results of the PCB master, which
showed different roughness of the channel intervals in the absence
(Fig. S1) and presence (Fig. 1d) of the microsieves.
As the microsieves was generated together with the microchannels, masters with micro-patterns featured with different
roughness and heights could be generated after single-step photolithography and the multi-level structures in the sealed microuidic device could be fabricated without precise alignment. The
advantages of sieving microstructures with a roughness of 2 m in
the sealed microuidic device include: (1) enabling uidic connection between two microchannels connected by the gaps
generated by the microsieves, as shown in Fig. 1b where the uid
was connected in all the microchannels when the inlet I was
added with Rose Bengal, a red dye; (2) size-dependent trapping of
microbeads with the diameter of 5.6 m alongside the microchannels after exertion of negative pressure.
3.2. Formation of single layer microbeads array
The ow velocity and the pressure differences in the trapping
unit were computationally simulated in Fig. 2. Without negative
pressure, the uidic ow passed through the main channel at a
high ow rate driven by the pressure difference between the inlet
and outlet of the main channel. Limited ow passed through the
microsieves because of its relative high uidic resistance with
respect to the main channel (Fig. 2a). When a negative pressure
was exerted into the pressure control channel, a larger amount of
the uid passed through the microseivesas driven by the dramatically decreased pressure in the sieving microstructure (Fig. 2b).

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

301

Fig. 2. Simulation of the pressure and ow velocity eld at the microsieves layer in one trapping unit (a) without negative pressure and (b) with negative pressure exerted to
the pressure control channel. (c) Micrograph of one trapping unit with single layer microbeads array. Scale bar is 100 m. (d) Number of microbeads trapped in eight trapping
units in one chip (n 10).

As the height of the sieving microstructure was about 2 m that

prevented the microbeads with a diameter of 5.6 m from owing
through, the microbeads were trapped alongside the trapezoid
edge of the main channel generated by isotropic etching (Kohler
1999; Yue et al., 2011). When one microbead was trapped on the
trapezoid edge of the main channel connecting the microseives,
other microbeads owing towards the sieving microstructure
would stop at unoccupied positions. Since the microbeads were
spherical in shape that will never completely block the uidic ow
through the microseives, microbeads array with more than onelayer might be created if high negative pressure was persistently
applied to the control channels.
The height of the trapezoid sidewall was optimized by controlling the etching time in order to overcome the collision and
clustering so that single layer of microbeads can be formed
(Fig. 2c). It should be noticed that the trapped microbeads could
not be washed away even when the negative pressure was
removed from the control channels. Stable trapping of immobilized microbeads facilitated the on-going detection process during
which no additional external equipment was required to manipulate the microbeads. The single layer beads array was individually
generated in the eight trapping units in series by separately
controlling the pressure in each unit. The averaged number of
microbeads in each trapping unit was 9578 beads (Fig. 2d).
3.3. Comparison of on-chip and off-chip analysis systems
The sensitivity of on-chip hybridization was determined with
samples containing PCR product of HPV11 ranging from 25 pM to
3.6 M (Fig. 3a). As a comparison, off-chip hybridization was also
carried out in the microtubes and analyzed by the ow cytometer.

Differences in the hybridization efciency and discrimination

sensitivities between on-chip and off-chip hybridization were
compared, indicating that higher signal was obtained in the
microuidic-based assay. The saturation concentration was 700 nM
for the on-chip hybridization, with uorescence intensity increased
slightly when the concentration of target DNA reached 3.6 M
(Fig. 3b). Meanwhile, the microuidic-based assay could discriminate 25 pM of target DNA at a signal-to-noise ratio (SNR)43, while
the off-chip analysis system could only generate a signicant
uorescent signal at a concentration of 360 pM of target DNA
(Fig. 3c).
To further eliminate the systematic differences in signal acquisition, i.e., the confocal microscope system for on-chip detection
and the ow cytometer for off-chip detection, we conducted the
experiments for the discrimination of the four HPV genotypes in
three groups: on-chip hybridization and on-chip detection, offchip hybridization and on-chip detection, and off-chip hybridization and off-chip detection. The values of SNR in each group as
shown in Fig. 3d indicated that different reading systems did not
affect the relative uorescence intensities of the on-chip and offchip hybridization results, where the signal was signicantly
enhanced in the microuidic-based hybridization and detection
system.
Bead-based multiplexed assay exhibited better reproducibility,
wider dynamic range, and shorter time required for assay preparation (Dasso et al., 2002). The microuidics enhanced mass
transport to solid surface of the microbeads during the hybridization process and thus exhibited remarkably higher discrimination
sensitivity compared with the in-tube hybridization process (Park
et al., 2002). The sensitivity of the analysis might be further
increased if other signal amplication methods such as quantum

302

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

Fig. 3. Comparison of the on-chip and off-chip hybridization and detection. (a) Micrographs of microbeads in the bright eld and at the emission of 575 nm after
hybridization with the target DNAs with various concentrations and uorescence labeling of phycoerythrin (PE). Scale bar is 10 m. (b) Fluorescence intensities of the
microbeads after on-chip (red circle) and off-chip (black square) hybridization with the target DNA with concentrations varying from 0 to 3.6 M. (c) Fluorescence intensities
of microbeads hybridized with the target DNA with low concentrations indicated by the dashed square in (b). (d) Comparison of signal-to-noise ratio between on-chip and
off-chip based DNA discrimination methods (n 50 for each type of microbeads). (For interpretation of the references to color in this gure legend, the reader is referred to
the web version of this article.)

dots and/or nanoparticles-based detections (Zhang et al., 2010,

2011) were introduced to the microuidic-based microbeads
system. In addition to the signicant enhancement in detection
range and the sensitivity, other advantages of the microuidicbased single-layer beads array as tabulated in Table 1, include
reduced sample/reagent consumption, and zero bead loss during
the washing steps. The formation of single layer linear beads array
in the microuidic device facilitates visual inspection and analysis
of individual microbead, making it possible to discriminate various
types of microbeads in the mixture for multiplexed analysis.
Moreover, it is convenient to monitor the capture process of the
analyte in real-time and compare the heterogeneity among individual microbead, which is useful for evaluating the performance
of the detection system during the development of new strategy
for high sensitive microbead-based analysis.
3.4. Discrimination analysis of virus genotypes
To demonstrate the feasibility of the multiplexed analysis based
on the microuidic-based single layer beads array platform, we

selected four sets of microbeads from the established pools to

discriminate four types of HPV genotypes. Each type of microbeads
was functionalized with one type of probes for HPV6, HPV11,
HPV16 and HPV18 genotypes, respectively (Table S2). Four types of
microbeads were evenly mixed and arrayed in ve replica chips:
four chips were exposed to one kind of target DNA, respectively,
and one was served as negative control using deionized water. The
images of the immobilized microbeads in each chip were recorded
and analyzed after on-chip hybridization and SA-PE staining
(Fig. 4a). The reporter signal of individual microbead at the
wavelength of 575 nm was counted only if the uorescence
intensities of the microbead at 650 nm and 710 nm were attributed to the specic region as indicated in the distribution map
(Fig. 4b). It was found that only microbeads functionalized with
the probe completely complementary to the target PCR product
could exhibit uorescent signal, as indicated from the averaged
uorescence intensities of 50 microbeads in each type (Fig. 4c).
These results demonstrated that the single layer microbeads array
could be applied for virus genotyping with good specicity in the
microuidic system.

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

The capability to detect clinical samples is a major requirement

during the development of diagnostic platform. In our platform,
on-chip virus genotyping was carried out using PCR products from
HPV positive cervical scrapes samples. The genomic DNAs of
cervical scrapes from four clinical samples with positive HPV

Table 1
Comparison of the microuidic device (on-chip) and ow cytometer-based system
(off-chip) for bead-based assays including the reagents/sample consumption,
manipulation of the microbeads, and detection limits of oligonucleotides and
proteins, respectively.
Consumption for each
trial

Microuidic device
(on-chip)

Flow-cytometer-based
system (off-chip)

Number of mixture
beads
Volume of Analyte (L)
Volume of SA-PE (L)
Volume of wash
solution (L)
Beads loss in the wash
steps
Visualization of each
microbead
Detection limit of
HPV (pM)
Leptin (pg/mL)
CEA (pg/mL)
CA125 (U/mL)
IL-8 (pg/mL)
HGF (pg/mL)
HE4 (pg/mL)

5  102

1  103

5
5
20

25
25
200

No

Yes

Yes

No

25
4.5
3.6
0.2
0.5
6.7
29.6

360
42.8
5.2
0.2
0.3
6.8
193.5

303

genotypes conrmed by DNA sequencing (Table S5) were extracted

and amplied by PCR, followed by analysis in the microuidic
device preloaded with a single layer array of four-plexed microbeads, each of which was functionalized with the probes for selected
four genotypes. The HPV genotypes were correctly discriminated by
the microuidic-based platform (Fig. 4d).

3.5. Microuidic-based multiplexed protein analysis

Protein analysis was also performed in the microuidic-based
microbeads platform. Similar to the DNA analysis, the procedure of
protein detection includes functionalization of the microbeads, onchip beads immobilization, on-chip reaction, microscopic imaging
and data analysis. We chose six proteins as the target analytes,
including leptin, carcinoembryonic antigen (CEA), carbohydrate
antigen 125 (CA125), interleukin 8 (IL-8), hepatocyte growth factor
(HGF), and human epididymis protein (HE4). The conditioned
medium cultured with A549 cells, one type of non-small-cell lung
cancer (NSCLC), was used as the sample which contains a mixture
of proteins including the six target molecules.
Of these six proteins, CEA and CA125 have been widely used for
early screening of cancer, and CEA has been used as a tumor
marker in NSCLC (Grunnet and Sorensen, 2012). Expression of HE4
protein has been validated in lung cancer recently (Galgano et al.,
2006) and shown as a potential diagnostic and prognostic marker
for lung cancer (Iwahori et al., 2012). HGF has the ability to
stimulate mitogenesis and matrix invasion (Yang et al., 2009),
and the interaction between HGF and its receptor mescenchymalepithelial transition factor gene (MET) and activate downstream

Fig. 4. Discrimination of four HPV genotypes in the microbead-based microuidic platform. (a) Micrographs of four-plexed microbeads mixture in the bright eld, at the
emission of 650 nm, 710 nm, and 575 nm, respectively. Different types of microbeads were identied according to the uorescent intensities at 650 nm and 710 nm and
marked with 1, 2, 3, and 4 respectively in the images. Scale bar is 20 m. (b) Identication of four-plexed microbeads mixture (dark triangle spot) according to the
distribution map (grey zone). (c) The uorescence intensity of individual microbead at 575 nm was analyzed (n 50 for each type of microbeads). Only the microbeads
completely complementary to the targets could show the uorescence signal. (d) Discrimination of the HPV genotypes in the clinical samples using the microuidic-based
microbead array platform.

304

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

Fig. 5. (a) Micrographs of the six-plexed microbeads mixtures for protein analysis. (a) Serial diluted standard mixture solutions, fresh medium, and conditioned medium
cultured with A549 cells for 48 h were analyzed in the identical microuidic chips. Scale bar is 20 m. (b) Standard curves for the analysis of leptin, CEA, CA125, IL-8, HGF,
and HE4 obtained by measuring the uorescence intensities at 575 nm of the six-plexed microbeads mixture reacted with serial diluted standard solutions in separate
microuidic chips. (c) Concentrations of six proteins in the fresh medium and in the conditioned medium cultured with A549 cells for 48 h. The concentration unit of leptin,
CEA, IL-8, HGF and HE4 was pg/mL and the unit of CA125 was U/mL.

signaling pathways promotes carcinogenesis in various tumor

types including lung cancer (Lesko and Majka, 2008). IL-8 is a
transcription factor essential for the induction of immune
responses and a tumor suppressor gene in different types of
cancers (Abrams, 2010; Schiavoni et al., 2004). It has been
demonstrated that IL-8 production from NSCLC could be initiated
by their own produced factors, leading to the formation of
inammatory microenvironment and recruitment of inammatory
cells (Zhang et al., 2012a, 2012b).
Six types of spectrum-barcoded microbeads encapsulated with
different dosages of two uorescent dyes were used (Table S1 and
Fig. S3), and the surface of each type of microbeads was conjugated with a capture antibody targeting one of the six mentioned proteins. Following the same microbeads immobilization
step, the analyte was added and incubated on-chip with the
microbeads overnight. Biotinylated secondary antibody was then
added into the chip followed by the addition of excessive SA-PE.
The feasibility and specicity of the functionalized microbead was
veried by reacting each type of beads with the standard solutions
containing 24 types of proteins and there was no crosstalk among
the target proteins in the mixture that would affect the result
precision. Multiplexed analysis was performed by mixing six types
of functionalized microbeads and analyzing the images of the
multiplexed microbeads reacted with (1) serial diluted standards,
(2) fresh medium, and (3) conditioned medium cultured with
A549 cells for 48 h (Fig. 5a). Standard curves of individual protein
were generated simultaneously using the protein mixture with
serial dilution (Fig. 5b), and the detection limits of each protein
were determined as shown in Table 1. The concentrations of six
types of proteins in the fresh medium and conditioned medium
cultured with A549 cells were quantied (Fig. 5c). Up to 20-folds
elevated protein concentration of CEA, IL-8 and HE4 was found in cell
treated medium when compared with the fresh medium, indicating
that A549 cells secreted those proteins during their growth. These

results demonstrated the feasibility of the microuidic bead-based

immunoassay for multiplexed analysis of protein markers.

4. Conclusions
We have developed a sensitive platform for multiplexed
bioanalysis based on the integration of microuidics and
microbead-based assays. The PDMS replica featured with eight
microbead-trapping units in series was fabricated by single-step
photolithography and replica molding. Single layer microbead
array was formed by the exertion of negative pressure in the
control channel connecting the bead loading channels by sieving
microstructures. The single layer array of beads were securely
immobilized alongside the channels after the removal of the
negative pressure. Multiplexed assays were achieved by using a
mixture of spectrally-encoded microbeads with specic probes,
which formed a linear array in the microuidic device, followed by
on-chip reaction and detection. The on-chip assay could correctly
detect 25 pM of target DNA with sequence-specic discrimination
ability, in contrast to the off-chip test that required  360 pM of
target DNA to reach a signal-to-noise ratio larger than 3. Parallel
detection and discrimination of four HPV genotypes and simultaneous identication of six proteins markers were demonstrated on
this microuidic-based microbeads array platform. Compared
with conventional ow cytometry, the advantages of this
microuidic-based platform include small consumption of
reagents, simple sample injection, and minimal sample contamination. In addition, as the microbeads were arranged in a single
layer array such that individual microbead could be visualized, the
platform would be useful in providing more accurate information
of the detection process when designing new strategies for novel
microbead-based bioanalysis in microuidic systems.

W. Yue et al. / Biosensors and Bioelectronics 54 (2014) 297305

Acknowledgment
This work was supported by grants from the National Key
Scientic Research Program (973 Program no. 2012CB933302),
Shenzhen Key Laboratory Funding Scheme of Shenzhen Municipal
Government, and the Research Grants Council of the Hong Kong
Special Administrative Region, China (Project no. CityU_103312).
Appendix A. Supporting information
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.bios.2013.10.034.
References
Abrams, S.I., 2010. Immunol. Res. 46 (13), 5971.
Ali, M.F., Kirby, R., Goodey, A.P., Rodriguez, M.D., Ellington, A.D., Neikirk, D.P.,
McDevitt, J.T., 2003. Anal. Chem. 75 (18), 47324739.
Braeckmans, K., De Smedt, S.C., Leblans, M., Pauwels, R., Demeester, J., 2002. Nat.
Rev. Drug Discovery 1 (6), 447456.
Chiou, P.Y., Ohta, A.T., Wu, M.C., 2005. Nature 436 (7049), 370372.
Daria, V.R., Rodrigo, P.J., Gluckstad, J., 2004. Biosens. Bioelectron. 19 (11),
14391444.
Dasso, J., Lee, J., Bach, H., Mage, R.G., 2002. J. Immunol. Methods 263 (12), 2333.
Dendukuri, D., Pregibon, D.C., Collins, J., Hatton, T.A., Doyle, P.S., 2006. Nat. Mater. 5
(5), 365369.
Derveaux, S., Stubbe, B.G., Braeckmans, K., Roelant, C., Sato, K., Demeester, J.,
De Smedt, S.C., 2008. Anal. Bioanal. Chem. 391 (7), 24532467.
Diercks, A.H., Ozinsky, A., Hansen, C.L., Spotts, J.M., Rodriguez, D.J., Aderem, A.,
2009. Anal. Biochem. 386 (1), 3035.
Dittrich, P.S., Tachikawa, K., Manz, A., 2006. Anal. Chem. 78 (12), 38873907.
Dunbar, S.A., Jacobson, J.W., 2000. Clin. Chem. 46 (9), 14981500.
Galgano, M.T., Hampton, G.M., Frierson, H.F., 2006. Mod. Pathol. 19 (6), 847853.
Grunnet, M., Sorensen, J.B., 2012. Lung Cancer 76 (2), 138143.
Han, K., Lee, T.Y., Nikitopoulos, D.E., Soper, S.A., Murphy, M.C., 2011. Anal. Biochem.
417 (2), 211219.
Holland, C.A., Kiechle, F.L., 2005. Curr. Opinion Microbiol. 8 (5), 504509.
Holmes, D., She, J.K., Roach, P.L., Morgan, H., 2007. Lab Chip 7 (8), 10481056.
Hsu, H.Y., Joos, T.O., Koga, H., 2009. Electrophoresis 30 (23), 40084019.
Huang, F.C., Liao, C.S., Lee, G.B., 2006. Electrophoresis 27 (16), 32973305.
Iwahori, K., Suzuki, H., Kishi, Y., Fujii, Y., Uehara, R., Okamoto, N., Kobayashi, M.,
Hirashima, T., Kawase, I., Naka, T., 2012. Tumor Biol. 33 (4), 11411149.
Jain, K.K., 2005. Nanotechnology-based lab-on-a-chip devices. Encyclopedia of Diagnostic Genomics and Proteomics, 2. Marcel Dekkar Inc., New York, pp. 891895.
Joensson, H.N., Andersson-Svahn, H., 2011. Lab Chip 11 (24), 41444147.
Kohler M., 1999. Etching in Microsystem Technology. Wiley-VCH Verlag GmbH,
pp. 1012, 6268.
Lawrie, G.A., Robinson, J., Corrie, S., Ford, K., Battersby, B.J., Trau, M., 2006. Int. J.
Nanomed. 1 (2), 195201.

305

Lesko, E., Majka, M., 2008. Frontiers in BioscienceLandmark, 13; , pp. 12711280.
Li, C.W., Cheung, C.N., Yang, J., Tzang, C.H., Yang, M.S., 2003. Analyst 128 (9),
11371142.
Li, C.W., Yang, J., Yang, M.S., 2006. Lab on a Chip 6 (7), 921929.
Lien, K.Y., Hung, L.Y., Huang, T.B., Tsai, Y.C., Lei, H.Y., Lee, G.B., 2011. Biosens.
Bioelectron. 26 (9), 39003907.
Lion, N., Reymond, F., Girault, H.H., Rossier, J.S., 2004. Curr. Opinion Biotechnol. 15
(1), 3137.
MacDonald, M.P., Spalding, G.C., Dholakia, K., 2003. Nature 426 (6965), 421424.
Mani, V., Chikkaveeraiah, B.V., Patel, V., Gutkind, J.S., Rusling, J.F., 2009. Acs Nano 3
(3), 585594.
Ng, J.K.K., Feng, H.H., Liu, W.T., 2007. Anal. Chim. Acta 582 (2), 295303.
Nicewarner-Pena, S.R., Freeman, R.G., Reiss, B.D., He, L., Pena, D.J., Walton, I.D.,
Cromer, R., Keating, C.D., Natan, M.J., 2001. Science 294 (5540), 137141.
Park, S.J., Taton, T.A., Mirkin, C.A., 2002. Science 295 (5559), 15031506.
Pregibon, D.C., Toner, M., Doyle, P.S., 2007. Science 315 (5817), 13931396.
Puig, O., Caspary, F., Rigaut, G., Rutz, B., Bouveret, E., Bragado-Nilsson, E., Wilm, M.,
Seraphin, B., 2001. Methods 24 (3), 218229.
Rawool, A.S., Mitra, S.K., Kandlikar, S.G., 2006. Microuid. Nanouid. 2 (3), 215221.
Santra, S., Zhang, P., Wang, K.M., Tapec, R., Tan, W.H., 2001. Anal. Chem. 73 (20),
49884993.
Schiavoni, G., Mattei, F., Borghi, P., Sestili, P., Venditti, M., Morse, H.C., Belardelli, F.,
Gabriele, L., 2004. Blood 103 (6), 22212228.
Situma, C., Hashimoto, M., Soper, S.A., 2006. Biomol. Eng. 23 (5), 213231.
Sochol, R.D., Casavant, B.P., Dueck, M.E., Lee, L.P., Lin, L., 2011. J. Micromech.
Microeng. 21, 5.
Taff, B.M., Voldman, J., 2005. Anal. Chem. 77 (24), 79767983.
Tanaka, Y., Sato, K., Shimizu, T., Yamato, M., Okano, T., Kitamori, T., 2007. Lab Chip 7
(2), 207212.
Tudos, A.J., Besselink, G.A.J., Schasfoort, R.B.M., 2001. Lab Chip 1 (2), 8395.
Vilkner, T., Janasek, D., Manz, A., 2004. Anal. Chem. 76 (12), 33733385.
Vo-Dinh, T., Cullum, B., 2000. Fresenius J. Anal. Chem. 366 (67), 540551.
Weibel, D.B., Garstecki, P., Ryan, D., Diluzio, W.R., Mayer, M., Seto, J.E., Whitesides, G.M.,
2005. Proc. Natl. Acad. Sci. USA 102 (34), 1196311967.
Wilson, R., Cossins, A.R., Spiller, D.G., 2006. Angew. Chem. Int. Ed. 45 (37),
61046117.
Xu, T., Li, C.W., Yao, X.S., Cai, G.P., Yang, M.S., 2010. Anal. Biochem. 396 (2), 173179.
Xu, T., Yue, W.Q., Li, C.W., Yao, X.S., Yang, M.S., 2013. Lab Chip 13 (6), 10601069.
Yang, Z.J., Zhang, Y.R., Chen, B., Zhang, S.L., Jia, E.Z., Wang, L.S., Zhu, T.B., Li, C.J.,
Wang, H., Huang, J., Cao, K.J., Ma, W.Z., Wu, B., Wang, L.S., Wu, C.T., 2009. Mol.
Biol. Rep. 36 (6), 13231329.
Yip, Y.C., Ngai, K.L.K., Vong, H.T., Tzang, L.C.H., Ji, S.L., Yang, M.S., Chan, P.K.S., 2010.
J. Med. Virol. 82 (10), 17241729.
Yu, X.B., Hartmann, M., Wang, Q.A., Poetz, O., Schneiderhan-Marra, N., Stoll, D.,
Kazmaier, C., Joos, T.O., 2010. Plos One 5, 10.
Yue, W.Q., Li, C.W., Xu, T., Yang, M.S., 2011. Lab Chip 11 (19), 33523355.
Yue, W.Q., Li, C.W., Xu, T., Yang, M.S., 2013. Biosens. Bioelectron. 41, 675683.
Zhang, H., Xu, T., Li, C.W., Yang, M.S., 2010. Biosens. Bioelectron. 25 (11), 24022407.
Zhang, H., Liu, L., Li, C.W., Fu, H.Y., Chen, Y., Yang, M.S., 2011. Biosens. Bioelectron. 29
(1), 8996.
Zhang, H., Fu, X., Liu, L., Zhu, Z.J., Yang, K., 2012a. Anal. Biochem. 426 (1), 3039.
Zhang, L., Yin, X.F., Fang, Z.L., 2006. Lab Chip 6 (2), 258264.
Zhang, Y., Wang, L.Y., Zhang, M.M., Jin, M.L., Bai, C.X., Wang, X.D., 2012b. J. Cell.
Physiol. 227 (1), 3543.