Noble polymeric surface conjugated with zwitterionic moieties and antibodies for the isolation of exosomes from human serum

Article abstract of DOI:10.1021/bc300339b

New zwitterionic polymer-coated immunoaffinity beads were developed to resist nonspecific protein adsorption from undiluted human serum for diagnostic applications of exosomes. A zwitterionic sulfobetaine monomer with an amine functional group was employed for simple surface chemistry and antifouling properties. An exosomal biomarker protein, epithelial cell adhesion molecule (EpCAM), was selected as a target molecule in this work. The beads were coated with polyacrylic acids (PAA) for increasing biorecognition sites, and protein G was then conjugated with carboxylic acid groups on the surfaces for controlling EpCAM antibody orientation. The remaining free carboxylic acid groups were modified with sulfobetaine moieties, and anti-EpCAM antibody was finally introduced. The amount of anti-EpCAM on the beads was increased by 40% when compared with PAA-uncoated beads. The surfaces of the beads exhibited near-net-zero charge, and nonspecific protein adsorption was effectively suppressed by sulfobetaine moieties. EpCAM was captured from undiluted human serum with almost the same degree of efficiency as from PBS buffer solution using the newly developed immunoaffinity beads.

Full text of DOI:10.1021/bc300339b

Article
pubs.acs.org/bc
Noble Polymeric Surface Conjugated with Zwitterionic Moieties and
Antibodies for the Isolation of Exosomes from Human Serum
Gahee Kim, Chang Eun Yoo, Myoungsoon Kim, Hyun Ju Kang, Donghyun Park, Myoyong Lee,*
and Nam Huh
Bio Research Center, Samsung Advanced Institute of Technology (SAIT), Mt. 14-1, Nongseo-dong, Giheung-gu, Yongin-si, 446-712,
Gyeonggi-do, South Korea
*
S Supporting Information
ABSTRACT: New zwitterionic polymer-coated immunoaffin-
ity beads were developed to resist nonspecific protein
adsorption from undiluted human serum for diagnostic
applications of exosomes. A zwitterionic sulfobetaine mono-
mer with an amine functional group was employed for simple
surface chemistry and antifouling properties. An exosomal
biomarker protein, epithelial cell adhesion molecule
(
EpCAM), was selected as a target molecule in this work.
The beads were coated with polyacrylic acids (PAA) for
increasing biorecognition sites, and protein G was then
conjugated with carboxylic acid groups on the surfaces for
controlling EpCAM antibody orientation. The remaining free
carboxylic acid groups were modified with sulfobetaine
moieties, and anti-EpCAM antibody was finally introduced. The amount of anti-EpCAM on the beads was increased by 40%
when compared with PAA-uncoated beads. The surfaces of the beads exhibited near-net-zero charge, and nonspecific protein
adsorption was effectively suppressed by sulfobetaine moieties. EpCAM was captured from undiluted human serum with almost
the same degree of efficiency as from PBS buffer solution using the newly developed immunoaffinity beads.
INTRODUCTION
nonspecific protein adsorption are poly(ethylene gylcol)
■
9
(
PEG) and oligo(ethylene glycol)s (OEG). However, PEG
Since Norwegian scientist John Ugelstad was first able to create
polystyrene beads, research on bead-based isolations was
extensively conducted for biomedical applications such as
and OEG can decompose in the presence of oxygen or
1
0,11
transition metal ions,
resulting in limited application of
1
2
3
these materials in biological fluids.
clinical diagnosis, drug targeting5, cell isolation and
4
6
Recently, zwitterionic moieties such as phosphorylcholine,
sulfobetaine, and carboxybetaine are being studied intensively
for reducing nonspecific protein adsorption in various
purification, nucleic acid purification, and detection.
Especially, the utility of magnetic beads in biological fluids is
tremendous due to the simple and fast separation process of
beads. In general, it is difficult to isolate target biomolecules
present in complex media with high purity. Therefore, magnetic
bead-based immunoaffinity purification has been envisaged as a
strategy to enhance the effective isolation of a target molecule
from the media. However, an immunoaffinity bead-based
method is not easily applicable to high-protein-containing
body fluids such as serum or plasma due to nonspecific
adsorption of serum proteins. In the case of serum, the primary
serum proteins are albumin, immunoglobulin, and fibrinogen,
12−15
applications.
Zwitterions are excellent candidates as
nonfouling materials because of balanced charge, minimized
dipole moment, and strong hydration capacity through
electrostatic interactions as well as highly hydrophilic character-
7
istics. Target proteins and chemicals in serum or plasma were
effectively detected by sensors of which the surfaces were
16
coated with zwitterionic moieties such as carboxybetaine. The
sensors in undiluted serum showed almost the same sensitivity
as in buffer solutions. Therefore, it is expected that
immunoaffinity magnetic beads modified with zwitterionic
moieties will have a great deal of utility in serum-based
diagnostics.
7
,8
and these proteins represent over 90% of total serum protein.
Because these proteins are adsorbed onto the surfaces of
immunoaffinity beads very fast and thus interfere with specific
binding, control of the nonspecific protein adsorption plays a
key role in achieving the successful isolation of target
molecules. Therefore, it is critical to develop immunoaffinity
beads with antifouling surface materials on the surfaces which
are effective enough in undiluted human serum. The most
common materials used for surface coatings to prevent
In general, zwitterionic polymer-based surface platforms are
prepared by surface grafting polymerization through surface-
Received: June 26, 2012
Revised: September 23, 2012
Published: October 1, 2012
©
2012 American Chemical Society
2114
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
initiated atom transfer radical polymerization (ATRP) or self-
assembled monolayer (SAM) formation through thiolating
monomers of zwitterionic monomers on the surface, which
limits the full exploitation of a zwitterionic surface chemistry in
various diagnostic applications. Even though one study
reported that a zwitterionic polymer could be synthesized as
a coatable type of polymer, the preparation process was highly
complicated and polymerization reactions were not efficiently
dimethylammonio)propane-1-sulfonate (B): Sultone was
added to the solution of N-Boc-dimethylaminopropylamine in
anhydrous N,N-dimethylformamide, and this solution was
stirred for 2 days at RT. After evaporating the solvent, the
product was dissolved in ethyl ether. By withdrawing the
solvent, the excess sultone was removed. The remaining solvent
was evaporated under high vacuum. 3-((3-Aminopropyl)-
dimethylammonio)propane-1-sulfonate (sulfobetaine, C): 4
M HCl in 1,4-dioxane was added to the solution of 3-((3-((tert-
butoxycarbonyl)amino)propyl)dimethylammonio)propane-1-
sulfonate in methylene chloride at 0 °C, and this solution was
then warmed to RT. After being stirred for 20 min, the solvent
was evaporated. The final product was obtained as a white solid
by repeated recrystallization in methylene chloride/isopropy-
lalcohol/methanol.
Preparation and Characterization of Immunoaffinity
Beads. The immunoaffinity beads (B-PGSE*) was prepared
through five stepwise reactions. Those are polymer incorpo-
ration, protein G conjugation, zwitterionic moiety modification,
antibody immobilization, and cross-linking of antibody with
protein G. The experimental details on the preparation and
surface characterization are provided in the Supporting
Information.
7
achieved. Therefore, immunoaffinity magnetic beads using a
zwitterionic surface chemistry have not yet been studied to the
best of our knowledge.
In this study, we developed zwitterionic polymer-coated
magnetic beads having a high degree of biorecognition sites.
We designed and synthesized the sulfobetaine monomer with
an amine functional group for simple conjugation with
carboxylic acid groups on bead surfaces, enabling a simple
antibody incorporation process in the bead preparation. The
number of specific binding sites was increased by coating the
surfaces with PAA to maximize the efficiency of target molecule
isolation, and protein G was conjugated with PAA for
controlling antibody orientation. We analyzed the surface
characteristics of the prepared beads and evaluated the
antifouling properties by comparing target detection in PBS
buffer with that in undiluted human serum.
Exosomes are membrane vesicles with a size of 40−100 nm
that are released from many different cell types such as red
blood cells, platelets, lymphocytes, dendritic cells, and tumor
RESULTS AND DISCUSSION
■
Preparation of Immunoaffinity Beads with Multifunc-
tional Groups. The preparation of immunoaffinity beads
comprises five steps (Scheme 2): (i) PAA coating for increasing
antibody reactive site (B-P), (ii) protein G conjugation for
controlling orientation of antibodies (B-PG), (iii) zwitterionic
moiety modification for preventing nonspecific protein
adsorption (B-PGS), (iv) antibody introduction for target
specific binding (B-PGSE), (v) cross-linking of antibodies with
protein G for preventing replacement with human immuno-
globulins (B-PGSE*) in serum. PAA was introduced onto
magnetic beads via EDC and NHS amide coupling for
increasing reactive sites. Additionally, PAA can lead to reduce
steric hindrance caused by bulky proteins such as protein G and
antibodies. PAA is a type of anionic polymer. In a water
solution at neutral pH, many of the side chains of PAA will lose
their protons and acquire a negative charge. Due to these
polyelectrolyte properties, PAA is hydrophilic and resistant to
1
7
cells and are found to occur in body fluids such as breast milk,
1
8,19
serum, plasma, malignant ascites, and urine.
Recently, a
number of studies have demonstrated that exosomes carry
1
8,20
specific proteins, mRNA and miRNA,
which can be utilized
as diagnostic markers of malignancy, including cancer. In this
work, an exosomal tumor marker protein, epithelial cell
adhesion molecule (EpCAM) was selected as a model protein
to demonstrate an antifouling surface chemistry of the
zwitterionic polymer-coated immunoaffinity beads.
EXPERIMENTAL SECTION
More detailed information on the materials and methods used
can be found in the Supporting Information.
■
Synthesis of Zwitterionic Moiety, Sulfobetaine (SB)
Scheme 1). N-Boc-dimethylaminopropylamine (A): Di-tert-
(
21
nonspecific protein adsorption. Major proteins of human
Scheme 1. Synthesis of the Zwitterionic Moiety, Sulfobetaine
serum are albumin and human IgG. The pI value of albumin is
4
.5−5.0 and that of human IgG is 7.0−7.5. At neutral pH, they
have a slightly negative charge or neutral charge; therefore,
negatively charged PAA is more efficient than positively
charged PAA for nonspecific protein adsorption. Protein G
was conjugated with carboxylic acid groups of PAA via EDC
22
and NHS coupling reactions for orientation of antibodies.
The protein G used in this study was a genetically engineered
truncated protein which retained its affinity for the Fc region of
IgG, but which lacked albumin- and Fab-binding sites and
23
membrane-binding regions. The sulfobetaine (SB) moieties
were also coupled with the remaining carboxylic acid groups of
PAA via EDC and NHS coupling reaction for suppressing
nonspecific protein adsorption. The representative zwitterionic
moieties are phosphorylcholine (PC), carboxybetaine (CB),
and sulfobetaine (SB). PC has been heavily studied because of
butyldicarbonate was added to a solution of N,N-dimethylami-
noproylamine in methanol at 0 °C and the solution was
warmed to RT. The solvent was then evaporated and DI water
was added. After extracting the product with ethyl acetate, the
solvent was evaporated. The white powder product was
obtained and used in the next step without further purification.
7
,12,24
its biomimetic characteristics,
but PC is relatively less
effective at resisting nonspecific protein adsorption in
15
physiological conditions than CB and SB. Most recently,
CB is having been greatly investigated because of its nonfouling
3
- ( ( 3 - ( ( t e r t - B u t o x y c a r b o n y l ) a m i n o ) p r o p y l ) -
2
115
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
Scheme 2. Schematic Representation of the Immunoaffinity
a
Beads Preparation Process
Figure 1. Identification of sulfobetaine (SB) on B-PGSE* by ToF-
SIMS. (A) Relative intensity of molecules S, CH, CN, and O after 0 h
(initial) and 2 h incubation. The relative intensity of sulfur was the
sum of S, SO, SO , and SO . (B) Relative intensity of sulfur molecules
2
3
with incubation time.
sum of S, SO, SO and SO originated from sulfate groups in
2
3
sulfobetaine increased significantly (Figure 1A). The intensities
of sulfur-containing molecules were not changed significantly
after 2 h of reaction time (Figure 1B). Although the amounts of
SB coupled on beads were not assessed directly, it could be
concluded that the beads used in our study were saturated with
SB from the results. The antibody for capturing a specific target
protein was introduced to the protein G via bioaffinity
interaction. We chose the EpCAM (epithelial cell adhesion
molecule) as a target protein, which is a transmembrane
protein. The amount of anti-EpCAM introduced on beads was
assessed by measuring the protein content of the antibody
solution before and after binding, as described in the
Experimental Section. To check the maximum amount of
antibody introduced into beads, solutions with various amounts
of antibody were mixed with beads (B-PGS) and the amount of
antibody in solution before and after mixing was measured. The
result is shown in Figure S1 in the Supporting Information. The
amount of antibody bound to beads did not increase as the
input amount of antibody was over 80 μg. The maximum
amount of antibody bound to B-PGSE* was 80 μg per 100 μL
bead. The same experiment was done with commercial protein
G bead (B′-GE*) and the maximum amount of antibody bound
to B′-GE* was 56 μg per 100 μL bead (Table 1). This amount
corresponds to ca. 65% of that for B-PGSE*. Therefore, the
a
(
A) Sulfobetaine-containing bead. (B) Commercial bead.
characteristics and functionalities. The carboxylate group itself
of CB could be both a reactive site to be coupled with another
molecule and the anionic part of a zwitterionic moiety.
25
Therefore, as other molecules are coupled to the carboxylic acid
groups of CB polymer, the zwitterionic characteristics of CB
will be lessened, which limits the use of CB as a surface
modifier in a simple process of surface chemistry exploiting
carboxylic acid functionality. SB moieties are effective at
resisting nonspecific protein adsorption in physiological
conditions but do not have reactive sites to be coupled with
other molecules. Sulfobetaine (SB) was thus chosen as a
zwitterionic moiety in our surface treatment study using PAA-
coated beads. The bead surface modification by SB moieties
was confirmed by ToF-SIMS, as shown in Figure 1. The
intensity of other molecules such as CH, CN, and O was not
changed, but that of sulfur-containing molecules which are the
26
Table 1. Comparative Data on B-PGSE* and B′-GE*
bound antibody (per
100 uL of beads)
zeta potential
nonspecific protein
2
bead ID
(mV, in pH 8.0) adsorption (ng/cm )
B-PGSE*
B′-GE*
80 μg (±2.0)
56 μg (±1.2)
−0.55 (±1.8)
−4.04 (±0.2)
775.0 (±28)
1576.1 (±90)
2
116
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
3
2
amount of antibody bound to B-PGSE* increased by 1.4 times,
compared with that on B′-GE* (According to the manufac-
turer’s manual, both types of beads have the same size (2.8 μm)
from pI values of proteins. Considering the pH of serum
(7.35−7.45) and pI value of main proteins in serum (IgG, 7.4;
albumin, 4.7), it may be expected that nonspecific protein
adsorption by electrostatic interaction in serum can also be
minimized. The charge properties of bead surfaces could be
identified by measuring zeta potential. The zeta potential of B-
PGSE* was determined by streaming potential in sodium
borate buffer (pH 8.0) and shown in Figure 2. The zeta
9
and concentration (2 × 10 /mL). Therefore, direct comparison
of antibody amount is reasonable). It is thought that the
increased amount of antibody on B-PGSE* may have resulted
from more reactive sites and possible reduction of steric
hindrance by introducing PAA on the surfaces of beads as
described above. The cross-linking was conducted between
protein G and anti-EpCAM to prevent an exchange of mouse
EpCAM antibody for human immunoglobulins (IgG) in serum.
The equilibrium constant (K) of protein G and human IgG
interaction is smaller than that of protein G and mouse IgG
11
−1
−10
−1 27
(
6.70 × 10−
M
vs 3.96 × 10 M ). The association
constants are almost the same for these two reactions, but the
dissociation constants differ by 1 order of magnitude.
Considering these values, it may be assumed that anti-
EpCAM which is not cross-linked with protein G will be
exchanged with human IgG if the reaction is conducted in
serum. This assumption was confirmed by incubation of B-
PGSE both in buffer and in serum, followed by Western blot
analysis. The Western blot band of anti-EpCAM when
incubated in undiluted human serum decreased about 40%
due to the exchange of antibody with IgG whereas that in buffer
almost retained (see Figure S2 in the Supporting Information).
Therefore, cross-linking of antibody with protein G must be
needed to prevent exchange of target-specific antibody for
human serum IgG. Dimethyl pimelimidate (DMP) was used as
a cross-linker because the imidoester functional group is one of
the most specific acylating groups available for the modification
of primary amines and has minimal cross reactivity toward
other nucleophilic groups in proteins. In addition, the
imidoamide reaction product does not alter the overall charge
of the protein, potentially retaining the native conformation
Figure 2. Zeta potential of B-P, B-PGE*, and B-PGSE* in sodium
borate buffer (pH 8.0) at RT. Error bars show the standard deviation
of three experiments.
potential value of B-PGSE* was −0.55 mV, almost zero charge.
On the other hand, those of B-P (only PAA coated beads) and
B-PGE* (beads prepared without SB) were −26.07 mV and
−
8.51 mV, respectively. The value of B′-GE* (beads prepared
by using commercial protein G beads) was −4.04 mV (Table
). The highly negative charge of B-P was caused by carboxylic
1
acid groups of PAA. The net negative charge of B-PGE* may
be due to remaining carboxylic acid and conjugated proteins on
the beads when considering pK value of carboxylic acid groups
a
28
and activity of the protein. Completion of cross-linking was
confirmed by disappearance of antibody band in Western blot
assay after reaction (see Figure S2 in the Supporting
Information).
(4.5) and pI value of protein G (4.2) and anti-EpCAM (7.4).
This result indicates that, in the case of B-PGSE*, almost all
carboxylic groups are coupled with protein G and SB, and SB
moieties are highly effective for neutralizing the surface charge.
From these data, we could conclude that the immunoaffinity
beads with almost net zero surface potential were prepared
successfully by incorporating SB moieties on the bead surface.
A measurement of water contact angle on the bead surface
was not technically feasible. However, it may be expected that
the beads with sulfobetaine have a high degree of hydrophilicity
when considering water contact angle value of 15° reported for
Surface Properties of Prepared Immunoaffinity Bead
B-PGSE*). There are four main interactions at the interface
(
between protein and surface. Those are hydrophobic, electro-
static, hydrogen bonding, and van der Waals interactions. In
order to reduce nonspecific protein adsorption onto a surface,
the control of these interactions is very important. Van der
Waals interactions cannot be controlled because they exist
between molecules universally. The control of hydrogen
bonding interaction between surface and nontarget proteins is
also limited, because all biomolecules act in the medium of
water including the human body. As a result, the control of
hydrophobic and electrostatic interactions between surface and
nontarget proteins is needed to resist nonspecific protein
15
the flat surface coated with sulfobetaine. This high degree of
hydrophilicity of SB modified surface was due to the formation
of a thicker hydration layer via charge−charge (electrostatic)
interaction as well as hydrogen bonding, as mentioned
15,16
previously.
These hydrophilic and electrically neutral
characteristics are thought to contribute to the resistance of
nonspecific protein adsorption on the bead surface.
7
,29,30
adsorption.
This could be done by incorporation of
zwitterionic groups on a surface. Zwitterionic surfaces induce
more water molecules to bind to the surface via ionic
interactions in addition to hydrogen bonding, resulting in the
thicker hydration layer and large amount of free water in the
Assessment of Nonspecific Protein Adsorption with
B-PGSE*. We used B′-GE* (beads prepared by using
commercial protein G beads) as references to evaluate isolation
capacity of our beads system. We evaluated the nonspecific
resistance properties of B-PGSE* by reacting in undiluted
human serum, and the level of serum protein adsorption was
compared with that of B′-GE*. The result was shown in Table
1. The amount of bound protein onto B-PGSE* was 775.0 ng/
1
6,31
hydration layer.
Therefore, zwitterionic surfaces are very
hydrophilic, and nonspecific protein binding can be mini-
1
3
mized. Zwitterionic surfaces have also balanced (net-zero)
charge and minimized dipole, and these characteristics are not
changed by biological pH range. In general, electrostatic
interactions between balanced charge surfaces and proteins can
be minimized if the pH of media is not significantly different
2
cm and there was only about 50% of B′-GE* nonspecific
2
binding (1576.1 ng/cm ). Nonspecific protein adsorption was
highly suppressed in the zwitterionic moiety modified beads (B-
2
117
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
PGSE*) as compared to control beads (B′-GE*). As previously
mentioned, charge balanced and highly hydrophilic surfaces are
the most important factors to reduce nonspecific protein
adsorption. These two factors were embodied on bead surfaces
by employing sulfobetaine, and consequently, nonspecific
adsorption of protein was reduced by about 50% when
compared with the control beads which have hydrophilic but
negatively charged surfaces.
In our results, the absolute amount of nonspecific protein
adsorption was rather higher than those reported in other
7
,14
references.
This discrepancy results from the different
detection systems used in the studies. In the most studies,
the nonspecific protein adsorption was measured by SPR
spectroscopy. The SPR spectroscopic study of protein
adsorption is very restricted. The SPR spectroscopy is a flow
cell detection system in which a contact time between surfaces
and analytes is limited. Even though, the flow rate can be
controlled, the contact time is within a very restricted range,
several tens of minutes because of systematic limitation. As
described in the Experimental Section, our beads were
incubated in human serum overnight and the contact condition
between bead surfaces and human serum proteins is not flow
system but mixed system. It allows us to adjust contact time
between surfaces and analytes as much as we want. Considering
these factors mentioned above, it is a matter of course that
nonspecific protein adsorption on bead surface is much higher
than that on flat surface, even if the same functional group such
as sulfobetaine are on both surface.
Figure 3. Detection of captured of EpCAM by B-PGSE* and B′-GE*.
(A) Relative band intensity of captured EpCAM by B-PGSE* after
incubation with various amounts of EpCAM positive exosomes in
buffer and in undiluted human serum. (B) Relative band intensity of
captured EpCAM by B′-GE* after incubation with various amount of
EpCAM positive exosomes in buffer and in undiluted human serum.
The relative band intensity is calculated by dividing intensity of input
exosome amounts band. Error bars show the standard deviation of
three experiments. (C) Western blot image of (A). (D) Western blot
image of (B).
PGSE* captures EpCAM more efficiently than B′-GE* due
to a higher degree of antibody on the surface, the amount of
antibody conjugated on beads does not explain the difference
entirely. We compared the capture efficiencies of B-PGE* and
B-PGSE* with the same amount of antibody to validate the
zwitterionic effect of sulfobetaine moieties. When reacted with
ca. 20 ng of exosomal EpCAM, B-PGE* showed 30% lower
capture efficiency than B-PGSE* in undiluted serum. There-
fore, the relatively lower decrease of captured amount of
EpCAM in serum by B-PGSE* may be largely due to the
decrease of nonspecific protein adsorption by sulfobetaine. As
mentioned in the Introduction, these results are parallel to
results in ref 16 in which the sensors coating the surface with
zwitterionic moieties in undiluted serum showed almost the
same sensitivity as in buffer solutions.
The comparative data of B-PGSE* and B′-GE* are
summarized in Table 1.
Isolation of Exosome with B-PGSE*. To isolate exosome
from serum, we selected EpCAM as a specific protein
biomarker of exosome. EpCAM is the cell surface molecule
that is known to be highly expressed in many primary and
33
metastatic epithelial cancers. Recently, some reports review
that the cancer-derived exosomes contain a higher level of
33
EpCAM than normal exosomes. In this research, serum of
cancer patients was mimicked by spiking EpCAM expressed
exosomes derived from a cancer cell line into normal human
serum. The normal human serum was purchased as sterile-
filtered solutions, and EpCAM was undetectable our assay. The
spiked samples were reacted with B-PGSE* and B′-GE*, and
exosomal EpCAM was then detected (Figure 3). The EpCAM
capture by the beads increased depending on the concen-
trations of spiked exosomes. The exosomes were also spiked
into PBS buffer, and the binding was compared with that in
undiluted human serum. Figure 3A showed the captured
amount of EpCAM by B-PGSE* from the buffer and undiluted
human serum. The relative intensities of detected EpCAM were
confirmed by Western blot band (Figure 3C,D) and were
calculated by using ImageJ.
Exosomes used in this experiment were prepared by
ultracentrifugation of cell culture media. Pellets obtained by
ultracentrifugation contain not only exosomes, but also various
34
proteins. EpCAM might also exist as a soluble form or
membrane debris. Therefore, it is useful to identify that the
signal of EpCAM is largely from exosomes for the possible
utilization of exosomes in diagnostic applications. We
determined that the captured protein fractions would have
another exosomal marker protein, integrin β, as well as
35
EpCAM, in order to determine that EpCAM was not
recovered in a soluble form. The result was shown in Figure 4A.
Both proteins could be detected in the same sample, and the
captured amounts of both proteins were proportional to the
total protein amounts of exosome fractions used for spiking.
Additionally, the same experiments were carried out with
another two kinds of beads which had antimouse IgG (B-
PGSI*) and no antibody (B-PS) instead of anti-EpCAM
(Figure 4B,C). Although amount input of total protein in
exosomes increased, EpCAM and integrin β were barely
detected in both beads, excluding the possibility of nonspecific
interaction or adsorption of integrin β.
The amounts of EpCAM bound to B-PGSE* in serum were
the same as that in PBS within the experimental errors over a
concentration range of 4.8 ng to 32 ng of exosomal EpCAM
was used. In contrast, the amount of EpCAM bound to B′-GE*
in serum were decreased by about 25−57% compared to that in
PBS over the same range of concentrations (Figure 3B). It was
reported that, in immunoaffinity reactions, a decrease in
nonspecific adsorption could reduce nonspecific interaction of
26
antibody with nontarget proteins. In our study, nonspecific
protein adsorption on beads decreased by incorporation of
sulfobetaine, and thus, anti-EpCAM might interact with
EpCAM more effectively on B-PGSE*. Even though B-
Last, the visualization of exosomes bound on beads was
performed by using field emission scanning electron micros-
copy. Figure 5A is an image of exosomes itself, Figure 5B is an
image of exosomes captured by beads in buffer, and Figure 5C
2
118
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
CONCLUSIONS
■
We have successfully prepared immunoaffinity beads with
zwitterionic moieties for controlling nonspecific protein
adsorption. The surface characteristics of beads were
determined by ToF-SIMS, BCA assay, and zeta potential
measurements. We also demonstrated that our novel beads
could successively suppress nonspecific protein adsorption and
effectively capture a target protein, EpCAM expressed on
exosomes in undiluted human serum. This study described that
our novel beads were very useful for isolating specific protein
molecules from body fluids such as human blood and could be
readily applicable for in vitro diagnostics use.
ASSOCIATED CONTENT
Supporting Information
Additional experimental section and figures. This material is
■
*
S
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
Phone: +82 31 2806943. Fax: +82 31 2806816. E-mail:
myoyong.lee@samsung.com.
■
*
Figure 4. Detection of captured EpCAM and integrin β expressed on
exosome surface by B-PGSE*, B-PGSI*, and B-PS in undiluted human
serum. (A) Relative band intensity of captured EpCAM and integrin β
by B-PGSE*, B-PGSI*, and B-PS after incubation of each bead with
various amount of exosomes. Relative band intensity is calculated by
dividing intensity of 2 μg exosome band into that of each band. Error
bars show the standard deviation of three experiments. (B) Western
blot image of captured integrin β by B-PGSE*, B-PGSI*, and B-PS.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The first two authors (G. Kim and C. E. Yoo) contributed
equally to this work. We would like to acknowledge Kyunghee
Park, Sungwoo Hong, Dr. Yeryoung Yong, and Jongmyeon
Park for helpful discussion and technical assistance.
(
C) Western blot image of captured EpCAM by B-PGSE*, B-PGSI*,
and B-PS.
REFERENCES
■
(
1) Olsvik, O., Popoic, T., Skjerve, E., Cudjoe, K. S., Hornes, E.,
Uglstad, J., and Uhlen, M. (1994) Magnetic separation techniques in
diagnostic microbiology. Clin. Microbiol. Rev. 7, 43−54.
(2) Chertok, B., Moffat, B. A., David, A. E., Yu, F., Bergemann, C.,
Ross, B. D., and Yang, V. C. (2008) Iron oxide nanoparticles as a drug
delivery vehicle for MRI monitored magnetic targeting of brain
tumors. Biomaterials 29, 487−496.
(3) Lauer, J. L., Furcht, L. T., and Fields, G. B. (1997) Inhibition of
melanoma cell binding to type IV collagen by analogs of cell adhesion
regulator. J. Med. Chem. 40, 3077−3084.
(4) Rabouille, C., Kondo, H., Newman, R., Hui, N., Freemont, P., and
Warren, G. (1998) Syntaxin 5 is a common component of the NSF-
and p97-mediated reassembly pathways of golgi cisternae from mitotic
golgi fragments in vitro. Cell 92, 603−610.
(5) McKay, S. J., and Cooke, H. (1992) hnRNP A2/B1 binds
specifically to single stranded vertebrate telomeric repeat TTAGGGn.
Nucleic Acids Res. 20, 6461−6464.
Figure 5. SEM images of the exosome as free and bead bound form:
(
A) Free exosome. (B) Exosome captured by B-PGSE* in buffer. (C)
(6) Dubus, S., Gravel, J. F., Drogoff, B. L., Nobert, P., Veres, T., and
Boudreau, D. (2006) PCR-free DNA detection using a magnetic bead-
supported polymeric transducer and microelectromagnetic Traps.
Anal. Chem. 78, 4457−4464.
Exosome captured by B-PGSE* in undiluted human serum.
(7) Kim, G., Park, S., Jung, J., Heo, K., Yoon, J., Kim, H., Kim, I. J.,
Kim, J. R., Lee, J. I., and Ree, M. (2009) Novel brush polymers with
phosphorylcholine bristle ends: synthesis, structure, properties, and
biocompatibility. Adv. Funct. Mater. 19, 1631−1644.
is an image of exosomes captured by beads in serum. The
exosomes were rather aggregated in both free and bound forms.
However, in this study we could not confirm that this
aggregation already existed in the solution or arose from the
sample preparation for SEM measurement. From these two
experiments, we could conclude that the captured EpCAM on
beads originated from exosomes.
(
8) Eskin, S. G., Horbett, T. A., McIntire, L. V., Mitchell, R. N.,
Ratner, B. D., Schoen, F. J., and Yee, A. (2004) Some background
concepts: Biomaterials Science, Elsevier Academic Press, Amsterdam.
(9) Harris, J. M. (1992) Poly(ethylene glycol) Chemistry: Biotechnical
and Biomedical Applications, Plenum Press, New York.
(10) Ostuni, E., Chapman, R. G., Holmlin, R. E., Takayama, S., and
Whitesides, G. M. (2001) A survey of structure-property relationships
2
119
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120

Bioconjugate Chemistry
Article
of surfaces that resist the adsorption of protein. Langmuir 17, 5605−
(29) Chen, S., Li, L., Zhao, C., and Zheng, J. (2010) Surface
hydration: Principles and applications toward low-fouling/nonfouling
biomaterials. Polymer 51, 5283−5293.
5620.
(
11) Harder, P., Grunze, M., Dahint, R., Whitesides, G. M., and
Laibinis, P. E. (1998) Molecular conformation in oligo(ethylene
glycol)-terminated self-assembled monolayers on gold and silver
surfaces determines their ability to resist protein adsorption. J. Phys.
Chem. B 102, 426−436.
(30) Chapman, R. G., Ostuni, E., Takayama, S., Holmlin, R. E., Yan,
L., and Whitesides, G. M. (2000) Surveying for surfaces that resist the
adsorption of proteins. J. Am. Chem. Soc. 122, 8303−8304.
(
31) Chen, S., Yu, F., Yu, Q., He, Y., and Jiang, S. (2006) Strong
resistance of a thin crystalline layer of balanced charged groups to
(
12) Iwasaki, Y., and Ishihara, K. (2005) Phosphorylcholine-
protein adsorption. Langmuir 22, 8186−8191.
containing polymers for biomedical applications. Anal. Bioanal.
Chem. 381, 534−546.
(32) Mike, K., and Williem, N. (1995) The adsorption of proteins
from aqueous solution on solid surfaces. Heterogeneous Chemistry
(
13) Colak, S., and Tew, G. N. (2012) Dual-functional ROMP-based
Reviews 2, 157−172.
betaines: effect of hydrophilicity and backbone structure on nonfouling
(33) Osta, W. A., Chen, Y., Mikhitarian, K., Mitas, M., Salem, M.,
properties. Langmuir 28, 666−675.
Hannun, Y. A., Cole, D. J., and Gillanders, W. E. (2004) Target for
breast cancer gene therapy EpCAM is overexpressed in breast cancer
and is a potential. Cancer Res. 64, 5818−5824.
(
14) Chen, S. F., Zheng, J., Li, L. Y., and Jiang, S. (2005) Strong
resistance of phosphorylcholine self-assembled monolayers to protein
adsorption: insights into nonfouling properties of zwitterionic
materials. J. Am. Chem. Soc. 127, 14473−14478.
(34) Mativanan, S., Ji, H., and Simpson, R. (2010) Exosomes:
Extracellular organelles important in intercellular communication. J.
Proteomics 73, 1907−1920.
(
15) Holmlin, R. E., Chen, X. X., Chapman, R. G., Takayama, S., and
Whitesides, G. M. (2001) Zwitterionic SAMs that resist nonspecific
(35) Welton, J. L., Khanna, S., Giles, P. J., Brennan, P., Brewis, I. A.,
adsorption of protein from aqueous buffer. Langmuir 17, 2841−2850.
Staffurth, J., Mason, M. D., and Clayton, A. (2010) Proteomics analysis
of bladder cancer exosomes. Mol. Cell. Proteomics 9, 1324−1338.
(
16) Yang, W., Xue, H., Carr, L., Wang, J., and Jiang, S. (2011)
Zwitterionic poly(carboxybetaine) hydrogels for glucose biosensors in
complex media. Biosens. Bioelectron. 26, 2454−2459.
(
17) Hendix, A., Westbroek, W., Bracke, M., and Wever, O. D.
(
2010) An ex(o)citing machinery for invasive tumor growth. Cancer
Res. 70, 9533−9537.
18) Skog, J., Wurdinger, T., van Rijn, S., Meijer, D. H., Gainche, L.,
(
Sena-Esteves, M., Curry, W. T., Jr., Carter, B. S., Krichevsky, A. M., and
Breakefield, X. O. (2008) Glioblastoma microvesicles transport RNA
and proteins that promote tumour growth and provide diagnostic
biomarkers. Nat. Cell Biol. 10, 1470−1476.
(
19) Pisikun, T., Shen, R. F., and Knepper, M. A. (2004)
Identification and proteomic profiling of exosomes in human urine.
Proc. Natl. Acad. Sci. U.S.A. 101, 13368−13373.
(
20) Lotvall, J., and Valadi, H. (2007) Cell to cell signaling via
exosomes through esRNA. Cell Adhes. Migr. 1, 156−158.
(
21) Shiratory, S. S., and Rubner, M. F. (2000) pH-Dependent
thickness behavior of sequentially adsorbed layers of weak
polyelectrolytes. Macromolecules 33, 4213−4219.
(
22) Dutta, P., Sawoo, S., Ray, N., Bouloussa, O., and Sarkar, A.
(
2011) Engineering bioactive surfaces with fischer carbene complex:
protein a on self-assembled monolayer for antibody sensing.
Bioconjugate Chem. 22, 1202−1209.
(
23) Goward, C. R., Murphy, J. P., Atkinson, T., and Barstow, D. A.
(
1990) Expression and purification of a truncated recombinant
streptococcal Protein G. Biochem. J. 267, 171−177.
24) Kyomoto, M., Iwasaki, Y., Moro, T., Konno, T., Miyaji, F.,
(
Hawaguchi, H., Takatori, Y., Nakamura, K., and Ishihara, K. (2007)
High lubricious surface of cobalt−chromium−molybdenum alloy
prepared by grafting poly(2-methacryloyloxyethyl phosphorylcholine).
Biomaterials 28, 3121−3130.
(
́
25) Vaisocherova, H., Yang, W., Zhang, Z., Cao, Z., Cheng, G.,
Piliarik, M., Homola, J., and Jiang, S. (2008) Ultralow fouling and
functionalizable surface chemistry based on a zwitterionic polymer
enabling sensitive and specific protein detection in undiluted blood
plasma. Anal. Chem. 80, 7894−7901.
(
26) Chang, Y., Shu, S. H., Shih, Y. J., Chu, C. W., Ruaan, R. C., and
Chen, W. Y. (2010) Hemocompatible mixed-charge copolymer
brushes of pseudozwitterionic surfaces resistant to nonspecific plasma
protein fouling. Langmuir 26, 3522−3530.
(
27) Nahshola, O., Bronnera, V., Notcovicha, A., Rubrechtb, L.,
Launeb, D., and Bravmana, T. (2008) Parallel kinetic analysis and
affinity determination of hundreds of monoclonal antibodies using the
ProteOn XPR36. Anal. Biochem. 383, 52−60.
(
28) O’Keeffe, E. T., Mordick, T., and Bell, J. E. (1980) Bovine
galactosyltransferase: interaction with a-Lactalbumin and the role of a-
lactalbumin in lactose synthaset. Biochemistry 19, 4962−4966.
2
120
dx.doi.org/10.1021/bc300339b | Bioconjugate Chem. 2012, 23, 2114−2120