Surface plasmon resonance study of protein-carbohydrate interactions using biotinylated sialosides

Article abstract of DOI:10.1021/ac702566e

Lectins are carbohydrate binding proteins found in plants, animals, and microorganisms. They serve as important models for understanding protein-carbohydrate interactions at the molecular level. We report here the fabrication of a novel sensing interface

Full text of DOI:10.1021/ac702566e

Anal. Chem. 2008, 80, 4007–4013
Surface Plasmon Resonance Study of
Protein-Carbohydrate Interactions Using
Biotinylated Sialosides
†
†
‡
‡
,†
Matthew J. Linman, Joseph D. Taylor, Hai Yu, Xi Chen, and Quan Cheng*
Department of Chemistry, University of California, Riverside, California 92521, and Department of Chemistry,
University of California, Davis, California 95616
Lectins are carbohydrate binding proteins found in plants,
animals, and microorganisms. They serve as important
models for understanding protein-carbohydrate interac-
tions at the molecular level. We report here the fabrication
of a novel sensing interface of biotinylated sialosides to
probe lectin-carbohydrate interactions using surface
plasmon resonance spectroscopy (SPR). The attachment
of carbohydrates to the surface using biotin-NeutrAvidin
interactions and the implementation of an inert hydro-
philic hexaethylene glycol spacer (HEG) between the
biotin and the carbohydrate result in a well-defined
interface, enabling desired orientational flexibility and
enhanced access of binding partners. The specificity and
sensitivity of lectin binding were characterized using
Sambucus nigra agglutinin (SNA) and other lectins in-
cluding Maackia amurensis lectin (MAL), concanavalin
A (Con A), and wheat germ agglutinin (WGA). The results
indicate that r2,6-linked sialosides exhibit high binding
affinity to SNA, while alteration in sialyl linkage and
terminal sialic acid structure compromises the affinity by
a varied degree. Quantitative analysis yields an equilib-
biologically relevant systems, opening new avenues for
probing carbohydrate-protein interactions in real time.
Lectins are carbohydrate binding proteins of considerable
1
specificity derived from plants, animals, and microorganisms.
Specifically, carbohydrate-protein interactions play key roles in
2
3
modulating intracellular traffic, endocytosis, cell-cell recogni-
4
3
5
tion, signal transduction, inflammation processes, and cancer
cell metastasis. Structurally, lectins have shallow binding pockets,
6
which result in relatively weak, noncovalent binding with affinity
7
–9
constants in the millimolar range.
In addition, lectins are
structurally complex and binding motifs are often not well-
8
understood. To overcome the weak-binding problem of the
carbohydrate-lectin interaction, the use of multivalent interactions
to enhance the binding affinity is often employed. This can be
accomplished by varying the density of the lectin-carbohydrate
6
recognition domains, clustering carbohydrates on the cell sur-
1
0
face, or altering molecular topography by adjusting the length
of linkers to improve their accessibility to ligands on opposing
1
1
binding partners.
Many analytical and bioanalytical techniques have been em-
ployed to study protein–carbohydrate interactions, including NMR
in conjunction with isothermal calorimetry and fluorescence
rium dissociation constant (K
binding to Neu5Acr2,6-LHEB. Transient SPR kinetics
confirms the K value from the equilibrium binding
D
) of 777 ( 93 nM for SNA
1
2
13
spectroscopy, dual polarization interferometry, enzyme-linked
D
1
4
15
lectin assays (ELLAs), quartz crystal microbalance (QCM),
studies. A linear relationship was obtained in the 10-100
µg/mL range with limit of detection of ∼50 nM. Weak
interactions with MAL, Con A, and WGA were also
quantified. The control experiment with bovine serum
albumin indicates that nonspecific interaction on this
surface is insignificant over the concentration range
studied. Multiple experiments can be performed on the
same substrate using a glycine stripping buffer, which
selectively regenerates the surface without damaging the
sialoside or the biotin-NeutrAvidin interface. This surface
design retains a high degree of native affinity for the
carbohydrate motifs, allowing distinction of sialyl linkages
and investigation pertaining to the effect of functional
group on binding efficiency. It could be easily modified
to identify and quantify binding patterns of any low-affinity
(
1) Sharon, N.; Lis, H. Lectins, 2nd ed.; Kluwer Academic: Dordrecht, The
Netherlands, 2003.
(
2) Lis, H.; Sharon, N. Chem. Rev. 1998, 98, 637–674
3) Sacchettini, J. C.; Baum, L. G.; Brewer, C. F. Biochemistry 2001, 40, 3009–
015
(4) Hakomori, S. Glycoconjugate. J. 2004, 21, 125–137
.
(
3
.
.
(
5) Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357–2364
.
(
6) Monsigny, M.; Mayer, R.; Roche, A. C. Carbohydr. Lett. 2000, 4, 35–52
.
(7) Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am. Chem.
Soc. 2003, 125, 6140–6148
8) Rini, J. M. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 551–577
9) Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem. Soc.
1998, 120, 10575–10582
10) Zhang, Y.; Luo, S.; Tang, Y.; Yu, L.; Hou, K. Y.; Cheng, J. P.; Zeng, X.;
Wang, P. G. Anal. Chem. 2006, 78, 2001–2008
11) Shinohara, Y.; Hasegawa, Y.; Kaku, H.; Shibuya, N. Glycobiology 1997, 7,
1201–1208
12) Murthy, B. N.; Voelcker, N. H.; Jayaraman, N. Glycobiology 2006, 16, 822–
32
13) Ricard-Blum, S.; Peel, L. L.; Ruggiero, F.; Freeman, N. J. Anal. Biochem.
2006, 352, 252–259
(14) Wu, J. H.; Singh, T.; Herp, A.; Wu, A. M. Biochimie 2006, 88, 201–217
(15) Pei, Y.; Yu, H.; Pei, Z.; Theurer, M.; Ammer, C.; Andre, S.; Gabius, H. J.;
Ramstrom, O. Anal. Chem. 2007, 79, 6897–6902
.
(
(
.
.
(
(
(
(
.
.
8
.
*
To whom correspondence should be addressed. Phone: (951) 827-2702. Fax:
.
(
951) 827-4713. E-mail: quan.cheng@ucr.edu.
.
†
University of California, Riverside.
‡
University of California, Davis.
.
1
0.1021/ac702566e CCC: $40.75  2008 American Chemical Society
Analytical Chemistry, Vol. 80, No. 11, June 1, 2008 4007
Published on Web 05/08/2008

1
6,17
and microarray techniques.
A number of these methods
immobilization of carbohydrates has been further complicated by
the need to conserve the original structure and expose the
functional sites so that the immobilized moiety can mimic the
specific biomolecular interactions occurring on the cell surface.
Because carbohydrates are flexible molecules, it has proven
technically challenging to immobilize carbohydrates in an oriented
fashion while retaining their inherent structures and functional-
require either labeling steps with convoluted chemistry or
expensive materials and optics. To circumvent these drawbacks,
surface plasmon resonance (SPR) has emerged as the method of
choice in recent years to study carbohydrate-lectin interac-
tions. SPR is advantageous for its intrinsic sensitivity that
has been documented to be at least an order of magnitude higher
than that of QCM for a comparable biological system. Also, the
SPR method is fast and suitable for real time measurement. SPR
has been used to study protein-carbohydrate interactions for the
1
2,18–20
2
1
33
ities. Most recently, Karamanska et al. used commercial Neu-
trAvidin-coated gold chips to detect lectin-carbohydrate interac-
3
4
tions in an array manner, seeking more biological interaction
information rather than focusing on method development.
An important family of monosaccharide-containing structures
that can be recognized by certain lectins are sialic acid (SA)-
containing structures. SAs are R-keto acids that widely exist as
components of the sugar chains of glycoconjugates in cells and
tissues, playing important roles in many biological recognition
2
2
determination of affinity constants, equilibrium dissociation
2
3
24
constants, and lectin specificity. However, SPR kinetic/
thermodynamic studies have been limited to the use of proprietary
prefabricated sensor surfaces thus far.
Meanwhile, the use of multifunctional surface chemistry
characterized by SPR for carbohydrate-protein interactions is still
an emerging scientific field. To effectively study carbohydrate-
lectin interactions with SPR, optimal surface chemistry is of most
significance. Previous work has used surface-immobilized lectin
3
5
mechanisms. Specifically, sialic acid is the binding moiety for
numerous plant lectins, including wheat germ agglutinin (WGA),
Sambucus nigra agglutinin (SNA), and Macckia amurensis lectin
2
5
36
due to easy fabrication. However, binding of a low molecular
(MAL). In recent years, derivatives of sialic acid have shown
weight carbohydrate ligand (<1000 Da) is difficult to quantify by
various biological activities, aiding the development and production
of medicines. However, many SA derivatives have yet to be
9,11
37
SPR. In contrast, immobilization of carbohydrates to the surface
has the advantage of straightforward detection of large lectin
molecules with the SPR method. To immobilize the carbohydrate
efficiently, simple and well-understood surface chemistry must be
developed. To date, several techniques have been attempted to
tested on biological systems. In particular, sialyl trisaccharides
containing various naturally existing sialic acid residues have yet
to be studied extensively. These structures have varying levels
of binding affinity to a number of lectins (SNA, MAL, WGA),
making them conducive for structure-function relationship stud-
ies of carbohydrate-lectin interactions. In addition, sialyltrisac-
charides can contain different sialic acid structures, sialyl linkages,
and underlying disaccharide structures, providing enough sophis-
tication for understanding complex oligosaccharide-based biomo-
lecular interactions systems. Furthermore, these sugars are
generally amenable to efficient synthesis and modification by the
one-pot multiple-enzyme chemoenzymatic synthetic system es-
26
immobilize carbohydrates, including copolymerization, reductive
2
7
amination, chemical immobilization of carbohydrates through
2
8
29
diazirine derivatization, self-assembled monolayers, and other
3
0
covalent immobilizations. While these approaches have been
successful for their specific systems, the complexity of most
oligosaccharides can cause surface deformation and inhomoge-
3
1
neity. Therefore, alternative surface modifications have been
2
0,32
attempted.
However, many of these strategies are multistep
3
8–40
laborious protocols that often lead to inconsistent results. Surface
tablished in the Chen group.
To create a reproducible and efficient biological mimic of
carbohydrate-lectin interaction, we report the use of chemoen-
zymatically synthesized biotinylated sialyldisaccharides to study
carbohydrate-lectin interactions. Figure 1 shows the schematic
illustration of the sensing interface. While a number of surface
chemistries have been employed to study carbohydrate-lectin
interactions as mentioned previously, the use of biotin-avidin
(
(
(
(
(
16) Liang, P. H.; Wang, S. H.; Wong, C. H. J. Am. Chem. Soc. 2007, 129, 11177–
1184
17) Kuno, A.; Uchiyama, N.; Koseki-Kuno, S.; Ebe, Y.; Takashima, S.; Yamada,
M.; Hirabayashi, J. Nat. Methods 2005, 2, 851–856
18) Duverger, E.; Frison, N.; Roche, A. C.; Monsigny, M. Biochimie 2003, 85,
67–179
19) Vornholt, W.; Hartmann, M.; Keusgen, M. Biosens. Bioelectron. 2007, 22,
983–2988
20) Suda, Y.; Arano, A.; Fukui, Y.; Koshida, S.; Wakao, M.; Nishimura, T.;
Kusumoto, S.; Sobel, M. Bioconjugate Chem. 2006, 17, 1125–1135
21) Su, X.; Wu, Y. J.; Knoll, W. Biosens. Bioelectron. 2005, 21, 719–726
22) Kapoor, M.; Thomas, C. J.; Bachhawat-Sikder, K.; Sharma, S.; Surolia, A.
Methods Enzymol. 2003, 362, 312–329
23) Krotkiewska, B.; Pasek, M.; Krotkiewski, H. Acta Biochim. Polon. 2002,
9, 481–490
24) Yamamoto, K.; Ishida, C.; Shinohara, Y.; Hasegawa, Y.; Konami, Y.; Osawa,
T.; Irimura, T. Biochemistry 1994, 33, 8159–8166
25) Kobayashi, Y.; Nakamura, H.; Sekiguchi, T.; Takanami, R.; Murata, T.; Usui,
T.; Kawagishi, H. Anal. Biochem. 2005, 336, 87–93
26) Hatanaka, K.; Takeshige, H.; and Akaike, T. J. Carbohydr. Chem. 1994,
3, 603–610
27) Biskup, M. B.; Muller, J. U.; Weingart, R.; Schmidt, R. R. ChemBioChem
005, 6, 1007–10015
1
.
.
1
.
2
.
1
5
-1
.
chemistry allows for strong binding (K
A
) 10
M ), facilitating
(
(
.
the removal of bound lectin, and eliminating the need for long
incubation steps. The biotinylated sialosides contain a flexible
.
(
(
(
(
(
(
4
.
(32) De Paz, J. L.; Noti, C.; Seeberger, P. H. J. Am. Chem. Soc. 2006, 128,
2766–2767
(33) Fukui, S.; Feizi, T.; Galustian, C.; Lawson, A. M.; Chai, W. Nat. Biotechnol.
2002, 20, 1011–1017
.
.
.
.
(34) Karamanska, R.; Clarke, J.; Blixt, O.; Macrae, J. I.; Zhang, J. Q.; Crocker,
P. R.; Laurent, N.; Wright, A.; Flitsch, S. L.; Russell, D. A.; Field, R. A.
1
.
Glycoconjugate J. 2008, 25, 69–74
(35) Yamamoto, K.; Konami, Y.; Irimura, T. J. Biochem. 1997, 121, 756–761
.
.
2
.
(36) Van Damme, E.; Pusztai, A.; Bardocz, S. Handbook of Plant Lectins: Properties
and Biomedical Applications; John Wiley and Sons: New York, 1998.
28) Chevolot, Y.; Martins, J.; Milosevic, N.; Leonard, D.; Zeng, S.; Malissard,
M.; Berger, E. G.; Maier, P.; Mathieu, H. J.; Crout, D. H. G.; Sigrist, H.
(37) Furuhata, K. Trends Glycosci. Glycotechnol. 2004, 16, 143–169
(38) Yu, H.; Chokhawala, H.; Karpel, R.; Wu, B.; Zhang, J.; Zhang, Y.; Jia, Q.;
Chen, X. J. Am. Chem. Soc. 2005, 127, 17618–17619
(39) Yu, H.; Abba, A.; Huang, S.; Chokhawala, H.; Sun, M.; Zheng, H.; Chen, X.
Angew. Chem., Int. Ed. 2006, 45, 3938–3944
(40) Yu, H.; Abba, A.; Chokhawala; Huang, S.; Chen, X. Nat. Protoc. 2006, 1,
2485–2492
.
Bioorg. Med. Chem. 2001, 9, 2943–2953
29) Zhi, Z.; Powell, A. K.; Turnbull, J. E. Anal. Chem. 2006, 78, 4786–4793
30) (a) Jelinek, R.; Kolusheva, S. Chem. Rev. 2004, 104, 5987–6015 and (b)
Houseman, B. T.; Mrksich, M. Chem. Biol. 2002, 9, 443–454
31) Seo, J. H.; Adachi, K.; Lee, B. K.; Kang, D. G.; Kim, Y. K.; Kim, K. R.; Lee,
.
(
(
.
.
.
.
(
H. Y.; Kawai, T.; Cha, H. J. Bioconjugate Chem. 2007, 18, 2197–2201
.
.
4008 Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

38–40
in the Chen laboratory.
ManNAc, 8) or mannose (9) was converted to Neu5Ac or KDN
by an aldol condensation reaction catalyzed by an Escherichia coli
E. coli) K-12 sialic acid aldolase, and then activated to form CMP-
In this system, N-acetylmannosamine
(
(
sialic acids catalyzed by an Neisseria meningitidis CMP-sialic acid
synthetase (NmCSS). The sialic acid residue in CMP-sialic acid
was then transferred to the galactose moiety of the acceptor 7 by
a multifunctional Pasteurella multocida sialyltransferase (PmST1)
to form an R2,3-linked sialoside 10 (Neu5AcR2,3-LHEB) or 11
(
KDNR2,3-LHEB), or by a Photobacterium damsela R2,6-sialyl-
transferase (Pd2,6ST) to form an R2,6-linked sialoside 12
Neu5AcR2,6-LHEB) or 13 (KDNR2,3-LHEB), in high yields
(
varying from 91 to 94% (Scheme 1). Sialoside products were
purified by Bio-Gel P-2 gel filtration chromatography, and the
1
13
structures were characterized by H and C NMR as well as mass
spectrometry (see Supporting Information).
SPR Analysis of Sambucus nigra Agglutinin on Biotiny-
lated Sialosides Immobilized on the Gold Substrates. Surface
interaction and modification were monitored and characterized
using the tracking mode of SPR angular scanning around the
minimum angle. To prepare the sensing surface, a previous
protocol was modified. Briefly, gold substrates were rinsed with
ethanol and ultrapure water. After drying under a gentle stream
2
of N gas, the gold substrates were clamped down by a flow cell
on a high-refractive index prism for use. Once PBS buffer had
established a smooth baseline across the surface, biotin-BSA (0.5
mg/mL in PBS) was injected across the gold chip. The modified
surface was allowed to incubate for 30 min and then rinsed with
PBS for 15 min to wash away any nonspecifically adsorbed species.
NeutrAvidin (0.5 mg/mL in PBS), a tetramer that forms a very
stable complex with biotin was subsequently injected across the
sensing surface and incubated for 90 min to allow binding to
biotin-BSA, while leaving additional biotin binding sites. Once a
stable signal was observed, the surface was rinsed with PBS for
30 min. The constructed sensor chip was then injected with a
biotinylated sialoside such as Neu5AcR2,6-LHEB (1.0 mg/mL in
PBS) and allowed to incubate for 30 min and rinsed with PBS for
30 min.
Figure 1. Cartoon representation of the carbohydrate-immobilized
surface for lectin sensing.
hexaethylene glycol (HEG) spacer. The HEG linker allows
orientational flexibility to closely mimic the sialoside-lectin
interaction in its native conformation. The binding of SNA to
sialosides was investigated, and the kinetic data are reported and
compared. This surface design is desirable for studies of both
strong and weak biological interactions, enabling rapid determi-
nation of lectin-carbohydrate binding specificity in a highly
sensitive manner.
42
EXPERIMENTAL SECTION
Materials and Instrumentation. Biotinylated bovine serum
albumin (b-BSA) and NeutrAvidin were purchased from Pierce
Biotechnology (Rockford, IL). Sambucus nigra agglutinin was
purchased from Vector Laboratories (Burlingame, CA). Bovine
serum albumin (BSA), wheat germ agglutinin, and Concanavalin
A (Con A) were from Sigma-Aldrich (St. Louis, MO). An SPR
spectrometer Biosuplar-2 (Analytical µ-Systems, Germany) with
a light-emitting diode light source (λ ) 670 nm), high-refractive
index prism (n ) 1.61), and 30 µL flow cell was used for all SPR
measurements. SPR gold chips were fabricated with a 46 nm thick
gold layer deposited by an e-beam evaporator onto cleaned BK-7
glass slides. All SPR experiments used 20 mM phosphate buffered
saline (pH 7.4 with 150 mM NaCl) as a running buffer.
SNA solutions diluted in PBS with concentrations varying from
10 to 1000 µg/mL were injected onto the carbohydrate-function-
alized surface. Each aliquot was incubated for 60 min and then
rinsed with PBS for 30 min. Upon rinsing the lectin-immobilized
surface, the minimum angle decreased slightly over time, char-
1
8
Synthesis of Biotinylated Sialosides. Monotosylation of
hexaethylene glyco (HEG, 1) was achieved in the presence of
acteristic of lectin dissociation kinetics. The free biotinylated-
carbohydrate-functionalized surface was regenerated with 100 mM
glycine (pH 1.7) stripping buffer to remove bound lectin. Once
the carbohydrate baseline had been established, a second aliquot
of the same concentration of lectin was injected and the entire
process was repeated. Each calibration standard was obtained in
triplicate on the same substrate.
4
1
silver(I) oxide and potassium iodide with good yield. The HEG
azide 3 was then prepared in 98% yield from the mesylate 2 by
using sodium azide in N,N-dimethylformamide (DMF) at 60 °C.
Coupling of lactose trichloroacetimidate 4 with acceptor 3 in the
presence of TMSOTf afforded compound 5 in 87% yield. After
deacetylation of 5 using NaOMe in methanol followed by
hydrogenation of the azido group to the primary amino group,
the reduced product coupling with N-hydroxy succinamide ester
of biotin (Biotin-NHS) in DMF afforded biotintylated lactoside
For the experiments with surfaces immobilized with different
carbohydrates, the same surface modification procedure was
followed as above except that a different biotinylated carbohydrate
was used in place of Neu5AcR2,6-LHEB. Similarly, for experiments
with various lectins, the same surface modification procedure was
followed as above except that a different lectin was used in place
of SNA. For each binding process, carbohydrates and lectins
(LHEB, 7) in 78% yield (Scheme 1).
Chemoenzymatic sialylation of 7 was achieved using a highly
efficient and convenient one-pot three-enzyme system established
(41) Bouzide, A.; Sauve, G. Org. Lett. 2002, 4, 2329–2332
.
(42) Whelan, R. J.; Zare, R. N. Anal. Chem. 2003, 75, 1542–1547.
Analytical Chemistry, Vol. 80, No. 11, June 1, 2008 4009

Scheme 1. Chemoenzymatic Synthesis of Biotinylated Sialosides 10-13
diluted in PBS with a final concentration of 1.0 mg/mL were used
as standards.
Data Analysis and Determination of Equilibrium and
D
Kinetic Constants. The equilibrium dissociation constant (K )
where kassoc is the association constant. By plotting the initial rate
against lectin concentration the result is a straight line with a slope
equal to ABmax assoc. After kaassoc is determined, kdiss, the dissociation
constant, can be determined by eq 3:
k
value for the interaction between lectin SNA and Neu5AcR2,6-
LHEB was determined from saturation binding experiments.
Increasing concentrations of SNA (10 µg/mL to 1 mg/mL) were
injected over the Neu5AcR2,6-LHEB-functionalized surface, and
AB ) (AB - AB )[exp(-kdisst)] + AB∞
(3)
t
0
∞
In this equation, AB
the dissociation curve), AB
dissociated. The value for the equilibrium dissociation constant,
0
is the initial response (i.e., the beginning of
the minimum angle shift was recorded. The value for K
D
was
∞
is the final response once completely
obtained using a nonlinear least-squares analysis outlined previ-
4
3
ously with eq 1:
K
D
, is thus determined by the following:
ABeq ) ABmax(1/(1 + K /[A]))
(1)
D
kdiss
K )
(4)
D
where ABeq is the average of the response signal at equilibrium
in defined intervals for each concentration of SNA ([A]) and ABmax
is the maximum response that can be obtained for lectin binding
depending on the number of binding sites available on the
kassoc
RESULTS AND DISCUSSION
Surface Chemistry, Detection, and Regeneration. The
surface chemistry employed in this work reflects our desire of
usingSPRtostudytherelativelyweakaffinityincarbohydrate-lectin
interactions with high accuracy. The choice of Sambucus nigra
agglutinin as the target was aimed to understand the general
binding characteristics of lectins to carbohydrates with a wide
affinity distribution in the micromolar-millimolar range. The
scheme in Figure 1 offers simplicity in surface chemistry to
provide a unique interface for detecting the generally weak
interactions. The structures for the synthesized biotinylated SA
are depicted in Scheme 1 (labeled 10-13). We have focused
initially on Neu5AcR2,6-LHEB due to the high SNA affinity for
4
4
Neu5AcR2,6-LHEB-functionalized surface.
association constant, can then be determined by the reciprocal
value of the K
In addition, to confirm the K
A
K , the equilibrium
D
.
D
value obtained for the
Neu5AcR2,6-LHEB surface, an initial rate kinetic method was
employed to extract kinetic information with transient SPR
response. At t ) 0, the equation for initial rate analysis is
4
5
dAB
dt
)
ABmax[A]kassoc
(2)
(
(
43) Harris, D. C. J. Chem. Educ. 1998, 75, 119–121
44) Navratilova, I.; Myszka, D. G. Surface Plasmon Resonance Based Sensors;
.
46
R2,6-linked sialosides. The hydrophilic HEG spacer allows
flexible movement of the carbohydrate immobilized to the surface
Springer-Verlag: Weinheim, Germany, 2006; Vol. 4, pp 159-161.
(45) Edwards, P. R.; Leatherbarrow, R. J. Anal. Biochem. 1997, 246, 1–6
.
via biotin-NeutrAvidin binding, thus enhancing the affinity of
4010 Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

Figure 2. Characteristic sensorgram for carbohydrate functionalized
sensing surface in response to 400 µg/mL SNA.
Figure 3. Sensorgram of Neu5AcR2,6-LHEB-functionalized surface
upon addition of 250 µg/mL SNA. Inset: Identical sensorgram except
that a 250 mM glycine stripping buffer of pH 1.7 is added before the
addition of SNA.
interaction of the carbohydrates and carbohydrate-binding lec-
tins. Biotinylated sialosides that varied either in the sialyl linkage
or the substituent at C-5 of the Neu5Ac were used for comparison.
Biotin occupies the opposite end of the PEG spacer, providing
highly specific binding to an avidin protein sublayer. The strong
including self-assembled monolayers (SAMs), are time-consuming
(requiring overnight incubation), require chemical modification,
4
7
and are less sensitive to lectin binding than the aforementioned
surface.
interaction of biotin-avidin was used in hope to remove the
bound lectin for repeat use without damaging the sialoside-
immobilized surface. After testing a class of avidin proteins, we
found that NeutrAvidin has the lowest nonspecific binding in the
avidin family most likely due to its deglycosylated structure.
Together with the better performance in biotin binding, it is
selected for use in all the experiments reported here.
To ensure that the surface regeneration did not damage the
carbohydrate-functionalized surface, 250 mM glycine, pH 1.7, was
injected over the Neu5AcR2,6-LHEB-derivatized surface before the
addition of 250 µg/mL SNA (Figure 3, inset). The data obtained
were then compared to a surface in which 250 µg/mL SNA was
injected without prior addition of glycine (Figure 3). A comparison
of the two experiments shows a similar equilibrium binding level
of lectin, with corresponding angular changes of 0.145 and 0.149°,
the latter representing the glycine-free assay. This demonstrates
the unique ability of this low-pH buffer to provide a fresh sensor
surface after each injection of lectin without compromising surface
functionality. A higher concentration of glycine (250 mM) was
used in this case to prove that, under even harsher conditions,
the lectin binding response would not be diminished by the
stripping buffer. For all other experiments, the concentration of
glycine was optimized at 100 mM, providing complete lectin
removal to the carbohydrate baseline and establishing a fully
functional surface capable of subsequent binding analyses, as
depicted in Figure 3.
The SPR sensorgram for the interaction of Neu5AcR2,6-LHEB
with SNA is shown in Figure 2. Upon addition of 400 µg/mL SNA
to the Neu5AcR2,6-LHEB -functionalized surface, slow binding was
observed, which is common for carbohydrate-lectin interactions.
The curve shape is consistent, even for long incubation times.
Upon rinsing with buffer, the lectin dissociated rapidly due to weak
11,18,25
carbohydrate-lectininteractions,similartopreviousSPRstudies.
It is highly desirable if such a surface can be regenerated with a
simple reagent that largely preserves the native binding structures.
Although various stripping buffers have been used for SPR studies,
we found that very few of them performed well on a carbohydrate
surface. After testing multiple stripping buffers, the best results
were obtained with 100 mM glycine, pH 1.7. This stripping buffer
was modified from a previous protocol in which the buffer was
4
8
Lectin Quantitation and Kinetic Analysis. The Neu5AcR2,6-
LHEB-functionalized surface was further tested over a wide range
of physiologically relevant concentrations of SNA to evaluate its
used to strip antibodies from a binding surface. The capability
of reusing the sialoside-immobilized surface was tested by
repeated injection of aliquots of 400 µg/mL SNA, followed by
addition of the glycine buffer. The sudden drop of the baseline
upon glycine injection is due to the difference in the solution
refractive index relative to the buffer. Rinsing with the buffer
quickly established the original carbohydrate baseline. This
greatly reduces material consumption required by the SPR
apparatus and significantly decreases assembly and experimental
time. In addition, the new stripping buffer allows facile quantitation
and direct comparison on the shape of kinetic curves by providing
the same baseline of the carbohydrate-coated surface. In com-
parison, other surface chemistries that have been attempted,
1
5,19
analytical performance (Figure 4).
At higher concentrations
of SNA, more binding sites become occupied, resulting in
saturation binding kinetics as expected. Using the linear portion
of the response curve to increasing concentrations of SNA (Figure
4
, inset), the limit of detection (LOD) was determined to be 8.01
µg/mL or 53.3 nM. Our results are comparable to the most recent
1
9
work using a similar approach. It should be noted that the
literature reported system used a higher concentration of im-
mobilized carbohydrate and more expensive instrumentation, and
19
each assay took more than a day to complete. Our work yielded
a dynamic range over several orders of magnitude (10-400 µg/
mL SNA or high nanomolar range), which is considerably better
(
(
(
46) Shibuya, N.; Goldstein, I. J.; Broekaert, W. F.; Nsimba-Lubaki, M.; Peeters,
B.; Peumans, W. J. J. Biol. Chem. 1987, 262, 1596–1601
47) Grun, C. H.; Van Vliet, S. J.; Schiphorst, W.; Bank, C.; Meyer, S.; Van Die,
I.; Van Kooyk, Y. Anal. Biochem. 2006, 354, 54–63
48) Phillips, K. S.; Han, J. H.; Cheng, Q. Anal. Chem. 2007, 79, 899–907
.
15
than the range observed for a similar QCM system and on par
19
with another SPR study. In comparison to recent carbohydrate-
.
.
lectin studies, our design offers distinctive advantages in that the
Analytical Chemistry, Vol. 80, No. 11, June 1, 2008 4011

5
0,51
previous reports
and emphasizes the importance of sialyl
linkage in lectin binding specificity. KDNR2,3-LHEB displayed
slightly weaker binding than Neu5AcR2,3-LHEB for reasons
indicated above with regard to replacing the C5 N-acetyl group
in Neu5Ac to the C5 hydroxyl group in KDN. This information
would be useful for future work to design other carbohydrate-
functionalized surfaces to study weak biomolecular interactions.
The Neu5AcR2,6-LHEB-functionalized surface was also used
to test four additional proteins or lectins at saturation concentra-
tions (Supporting Information). The strongest binding protein was
SNA, followed by Con A, which is known to specifically bind to
5
2
mannosylated surfaces. Mannose is a stereoisomer of both
galactose and glucose, which are the monosaccharides comprising
the lactose that is present on all of the four biotinylated sialosides
used here. In fact, Con A exhibits a primary specificity for
Figure 4. Sensing response obtained by addition of increasing
concentrations of Sambucus nigra lectin to a Neu5AcR2,6-LHEB-
immobilized surface. Inset: Linear portion of response curve before
saturation binding occurs.
5
3
mannose, but will also bind glucose. Therefore, a relatively
strong binding signal to Con A is expected, yet weaker than that
of SNA. Additionally, the binding to wheat germ agglutinin was
slightly weaker than that of Con A. While WGA does have affinity
experiments were performed with inexpensive instrumentation,
along with simple and well-understood chemistry, allowing re-
generation of the sensing surface to yield reproducible results.
54
for sialic acid, it is a secondary interaction with affinity 4-fold
less than its affinity to N-acetylglucosamine (GlcNAc) according
SPR sensorgrams provided values for the determination of K
D
53
to a hapten-inhibition study of the methyl ester of sialic acid.
6
-1
and K
A
. The values for both K
A
(1.29 × 10 M ) and K
D
(777 ±
Thus the lack of strong binding exhibited by WGA to the
Neu5AcR2,6-LHEB-immobilized surface was expected. The bind-
9
3 nM) were quite consistent with previously reported constants
10,18,20
for similar systems.
Specifically, two separate reports studied
ing of Maackia amurensis lectin (a lectin known to bind to
the interaction of SNA with sialic acid-terminated R2,6 conjugates
by immobilizing SNA. Haseley et al. employed a similar oligosac-
charide and determined an affinity constant that was very similar
35,51
Neu5AcR2,3Lac(NAc) trisaccharide containing structures)
to
the Neu5AcR2,6-LHEB-immobilized surface was even weaker.
Finally, the weakest binding protein was bovine serum albumin.
BSA has no known affinity toward any specific carbohydrate, so
binding should be minimal. The result also confirms that the
nonspecific interaction on this surface is insignificant over this
concentration range. Overall, we believe that these results show
the applicability of this surface to be used to distinguish many
different types of lectins and carbohydrates in a quantitative and
real-time manner.
6
-1 49
to ours (1.9 × 10 M ), while Duverger et al. showed a slightly
6
-1 50
higher binding affinity (4.5 × 10 M ), likely due to the use of
11
a neoglycoprotein which is known to enhance affinity constants.
For comparison, we also used the transient SPR response
method to determine the binding constant on the carbohydrate-
functionalized surface. The calculation of the dissociation constant
-
3 -1
(
1
k
0
diss ) 6.49 × 10
M
s
) and association constant (kassoc ) 8.60 ×
4
-1 -1
s ) enables the determination of the equilibrium
dissociation constant to be 754 nM. This value agrees very well
with the number determined through equilibrium binding analysis
and thus verifies the results and method employed in this work.
Screening of Carbohydrate-Lectin Interactions. The sur-
face system was further tested using three additional structurally
similar sialosides (10, 11, and 13 in Scheme 1) at the saturation
concentration of SNA. Results demonstrate structure-dependent
variation on binding characteristics, with the Neu5AcR2,6-LHEB-
immobilized surface having the highest affinity to SNA (Support-
ing Information). The KDNR2,6-LHEB-immobilized surface also
showed significant binding. By using two biotinylated sialosides
containing a Neu5Ac or a KDN, our surface design was able to
clearly distinguish these compounds altering only one functional
group on the sialic acid (Neu5Ac has an N-acetyl group at C5,
while KDN has a hydroxyl group at C5). SNA showed a significant
decrease in binding to the Neu5AcR2,3-LHEB surface compared
to Neu5AcR2,6-LHEB, demonstrating the importance of sialyl
linkages in the SNA binding. This result is consistent with several
CONCLUSIONS
With the use of simple and highly effective surface chemistry
based on biotinylated sialosides, we have shown the study of
carbohydrate-lectin interactions in a detailed, quantitative fashion.
Structure-dependent variation of sialoside-lectin binding was
observed for different lectins and different carbohydrate-im-
mobilized surfaces. These results indicate the sensing surface
described here could be broadly applied to a wide variety of
biologically relevant low-affinity systems beyond the carbohydrate-
lectin interaction studies shown here. The bare carbohydrate
surface could be reproduced multiple times without degradation
of the binding signal. In addition, this surface design has high
sensitivity for weak carbohydrate-lectin interactions due to unique
surface chemistry employed. Its ability to distinguish subtle
changes on the ligand of carbohydrate-binding protein is remark-
(51) Knibbs, R. N.; Goldstein, I. J.; Ratcliffe, R. M.; Shibuya, N. J. Biol. Chem.
1
991, 266, 83–88
.
(
(
52) Mangold, S. L.; Cloninger, M. J. Org. Biomol. Chem. 2006, 4, 2458–2465
.
53) The Lectins: Properties, Functions, and Applications in Biology and Medicine;
(
49) Haseley, S. R.; Talaga, P.; Kamerling, J. P.; Vliegenthart, J. F. Anal. Biochem.
999, 274, 203–210
50) Duverger, E.; Coppin, A.; Strecker, G.; Monsigny, M. Glycoconjugate J.
999, 16, 793–800
Liener, I. E., Sharon, N., Goldstein, I. J., Eds.; Academic Press: Orlando,
1
.
FL, 1986; pp 103-115
(54) Peters, B. P.; Ebisu, S.; Goldstein, I. J.; Flashner, M. Biochemistry 1979,
18, 5505–5511
.
(
1
.
.
4012 Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

able. One could envision transforming the current system to a
microarray format for high-throughput screening of weak biologi-
cal interactions. Future work will focus on optimizing the im-
mobilization density of the oligosaccharides and determining the
ideal biotin linker for the sialosides for maximum sensitivity in
monitoring carbohydrate-lectin interactions.
SUPPORTING INFORMATION AVAILABLE
Experimental details for synthesis, NMR, MS data, and
carbohydrate/protein structure dependent binding results (pdf).
This material is available free of charge via the Internet at
http://pubs.acs.org.
ACKNOWLEDGMENT
Received for review December 18, 2007. Accepted March
Q.C. would like to acknowledge the support from NSF Grant
CHE-0719224. X.C. acknowledges the support from NIH Grant
R01 GM076360.
10, 2008.
AC702566E
Analytical Chemistry, Vol. 80, No. 11, June 1, 2008 4013