Molecular analysis of carbohydrate-antibody interactions: Case study using a bacillus anthracis tetrasaccharide

Article abstract of DOI:10.1021/ja104027w

The process for selecting potent and effective carbohydrate antigens is not well-established. A combination of synthetic glycan microarray screening, surface plasmon resonance analysis, and saturation transfer difference NMR spectroscopy was used to dissect the antibody-binding surface of a carbohydrate antigen, revealing crucial binding elements with atomic-level detail. This analysis takes the first step toward uncovering the rules for structure-based design of carbohydrate antigens.

Full text of DOI:10.1021/ja104027w

Published on Web 07/08/2010
Molecular Analysis of Carbohydrate-Antibody Interactions: Case Study
Using a Bacillus anthracis Tetrasaccharide
Matthias A. Oberli,†,∇ Marco Tamborrini,§ Yu-Hsuan Tsai,† Daniel B. Werz,†,⊥ Tim Horlacher,†
Alexander Adibekian,†,# Dominik Gauss,| Heiko M. Mo¨ller,| Gerd Pluschke,§ and
Peter H. Seeberger†,‡,
*
Department of Biomolecular Systems, Max-Planck Institute for Colloids and Interfaces, 14476 Potsdam, Germany,
Freie UniVersita¨t Berlin, 14195 Berlin, Germany, Swiss Tropical and Public Health Institute, UniVersity of Basel,
4002 Basel, Switzerland, and Department of Chemistry, UniVersity of Konstanz, 78457 Konstanz, Germany
Received May 11, 2010; E-mail: seeberger@mpikg.mpg.de
highly immunogenic glycoprotein found on the surface of the spores
Abstract: The process for selecting potent and effective carbo-
of Bacillus anthracis, the agent that causes the acute zoonotic
hydrate antigens is not well-established. A combination of
synthetic glycan microarray screening, surface plasmon reso-
nance analysis, and saturation transfer difference NMR spec-
troscopy was used to dissect the antibody-binding surface of a
carbohydrate antigen, revealing crucial binding elements with
atomic-level detail. This analysis takes the first step toward
uncovering the rules for structure-based design of carbohydrate
antigens.
disease anthrax. This tetrasaccharide was a particularly attractive
target because glycans found on the B. anthracis spore surface are
excellent candidate antigens for developing anthrax vaccines and
diagnostic tools.11,12
The BclA tetrasaccharide contains three rhamnose residues and
an unusual terminal sugar, 2-O-methyl-4-(3-hydroxy-3-methylbu-
tanamido)-4,6-dideoxy-D-glucopyranose, which is named anthrose.13
This tetrasaccharide and related structures were synthesized chemi-
cally and successfully used as both antigens and molecular probes
to identify B. anthracis.14-18 Furthermore, synthetic tetrasaccharide
Carbohydrate-based vaccines, which induce an immune response
against cell-surface oligosaccharides found on disease-causing
bacteria, are now used to protect young children from meningitis
and other severe bacterial diseases.1-4 However, selecting an oligo-
or polysaccharide antigen that will generate highly selective
antibodies is not a straightforward process, and the rules have not
yet been defined. Typical oligosaccharides contain few, if any,
charged residues and primarily display only hydroxyl and amine
groups on their sugar rings. This starkly contrasts with the rich
structural and chemical diversity of peptides, which are better-
understood antigens. Understanding how complex carbohydrates
interact with antibodies is an important first step toward establishing
rules for guided carbohydrate antigen design.5 After the pioneering
work by Lemieux,6 the antibody response to immune challenge
with antigens from Salmonella,7 Shigella flexneri,8 Vibrio cholerae,9
and Candida albicans10 has been investigated. Since crystallization
is rarely possible, enzyme-linked immunosorbent assay (ELISA)-
based formats, immunoblotting, and surface plasmon resonance
(SPR) analysis have been utilized to define key epitopes. Here we
show that crucial antibody-binding positions on the sugar antigen
can be identified using a combination of three established tech-
niques: synthetic glycan microarray screening, SPR analysis, and
saturation transfer difference (STD) NMR spectroscopy. It is the
combination of these complementary techniques that reveals the
molecular interactions between antibodies and carbohydrate anti-
gens. We have examined the tetrasaccharide component of the
glycoprotein Bacillus collagen-like protein of anthracis (BclA), the
1 (Figure 1), was immunogenic in mice and generated the
monoclonal antibodies (mAbs) MTA1-3.19,20 The selectivity of
the anti-tetrasaccharide and anti-disaccharide mAbs was shown by
immunofluorescence analyses, where the antibodies distinguished
B. anthracis spores even from those of closely related strains such
as Bacillus thuringiensis.19-21 For the purposes of this study, anti-
disaccharide mAbs (designated MTD1-6) were obtained by B-cell
hybridoma technology using spleen cells of mice immunized with
the BclA-related disaccharide 15 (Figure 1).
To uncover which structural and chemical elements of the
carbohydrate influenced this selectivity, microarray screening was
performed using a family of synthetic oligosaccharides related to
the original BclA tetrasaccharide. The synthetic oligosaccharide
analogues ranged from mono- to tetrasaccharides and were equipped
with different side chain appendages as well as a thiol-modified
linker at the reducing end for attachment to a maleimide-
functionalized microarray.22 These synthetic glycans were screened
for their abilities to bind the anti-disaccharide mAbs (MTD1-6)
and the anti-tetrasaccharide mAbs (MTA1-3) (Figure 1). The anti-
tetrasaccharide and the anti-disaccharide mAbs exhibited profoundly
different binding patterns. The anti-disaccharide mAbs recognized
all of the synthetic structures with intact anthrose, including anthrose
monosaccharides (1-6 and 12-16). Similarly, the anti-tetrasac-
charide mAbs strongly bound tetrasaccharide analogues 1 and 2
and trisaccharide 16. However, the anti-tetrasaccharide antibodies
only weakly bound tetrasaccharide analogues 3, 4 and 6, and
tetrasaccharide analogues 5 and 7 were not bound at all. Notably,
each of these structures contained a modified terminal anthrose.
None of the antibodies (anti-disaccharide or anti-tetrasaccharide)
recognized mono-, di-, or trirhamnose structures (8-11). Altogether,
these results demonstrate that anthrose is the minimal unit required
for binding anti-disaccharide mAbs. Interestingly, while a terminal
anthrose is absolutely required for oligosaccharide recognition by
the anti-tetrasaccharide mAbs, these mAbs failed to bind the
anthrose-containing truncated mono- and disaccharide structures
† Max-Planck Institute of Colloids and Interfaces.
‡ Freie Universita¨t Berlin.
§ Swiss Tropical and Public Health Institute, University of Basel.
| University of Konstanz.
⊥ Current address: Institut fu¨r Organische und Biomolekulare Chemie, Georg-
August-Universita¨t Go¨ttingen, 37077 Go¨ttingen, Germany.
# Current address: The Scripps Research Institute, 10550 North Torrey Pines
Road, La Jolla, CA 92037.
∇ Current address: Department of Chemistry, Massachusetts Institute of Technol-
ogy, Cambridge, MA 02139.
9
10.1021/ja104027w  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 10239–10241 10239

C O M M U N I C A T I O N S
Figure 1. Synthetic glycans related to the B. anthracis cell-surface tetrasaccharide BclA. The synthetic glycans were used for antibody mapping by microarray
screening, SPR, and STD NMR analyses. Microarray analysis demonstrates the cross-reactivity of monoclonal antibodies generated against anthrose-rhamnose
disaccharide 15 (MTD1-6) and tetrasaccharide 1 (MTA1-3).
12-15. Therefore, the anti-tetrasaccharide mAbs require at least
two rhamnose units as well as the terminal anthrose for tight
oligosaccharide binding.
Unfortunately, the extremely slow dissociation of the MTD6-
disaccharide complex prevented further analysis of this antibody-
oligosaccharide pair using STD NMR spectroscopy.23,24 This
method is particularly suited for characterizing binding differences
within ligands (discriminating tightly bound domains from weakly
bound ones) without having to assign the resonances of the
macromolecular receptor. However, slow kinetics results in very
limited transfer of ligands from the antibody-bound state to the
free state, strongly affecting the signal-to-noise ratio in STD NMR
experiments. In addition, an increased antibody-ligand complex
lifetime results in intraligand spin diffusion that decreases the
discrimination between individual positions of the ligand and
prevents detailed epitope mapping. The complex of MTA1 and the
tetrasaccharide, however, was a good candidate for further analysis
using STD NMR spectroscopy. Assessment of antibody binding at
a 30:1 ratio of carbohydrate ligand to protein confirmed that MTA1
tightly bound all four sugars of the tetrasaccharide (Figure 2) but
had little effect on the unnatural linker at the reducing end of
rhamnose D. Strong STD effects indicated that tight-binding sites
were located throughout the entire tetrasaccharide on all four sugars,
with a cluster of tight-binding sites found within the anthrose-
(ꢀ1-3)-rhamnose substructure (Figure 2). Binding was relatively
weaker at the opposite end of the molecule, but STD effects showed
that binding in this region was still significant (Figure 2). Looking
at STD effects throughout the structure, we observed that one face
of the trirhamnose chain (protons H1-H2-H3) was bound more
tightly by the antibody than the opposite side (protons H4-H5-H6).
The H1-H2-H3 face of the trirhamnose chain is likely to be
oriented closer to the antibody within the tetra-
saccharide-antibody complex. STD analysis had indicated that
within the anthrose unit there is a cluster of sites tightly bound by
MTA1. Specifically, on the anthrose sugar ring, three protons
showed strong STD effects, but the 2′′-O-methyl group was bound
less strongly. This observation agrees with our microarray data,
which indicated that this side chain appendage was of minor
importance for recognition by MTA1 (1 and 2 in Figure 1). The
microarray data also demonstrated that the anthrose C4 side chain
was crucial for binding MTA1, and this finding is supported by
the STD NMR results, which showed that within the C4 chain, a
methylene group as well as two methyl groups appear to have
Many glycans, particularly in mammalian systems, do not contain
side chain appendages, but anthrose does. Since the anthrose sugar
was essential for antibody binding, its distinctive side chain was
investigated in greater detail. A drastic truncation of the chain
produced by reducing 3-hydroxy-3-methylbutyrate to acetate (7 in
Figure 1), resulted in a structure that was not recognized by any
mAb. However, deleting a methyl group within the side chain by
replacing 3-hydroxy-3-methylbutyrate with 3-hydroxybutyrate (3
and 4 in Figure 1) reduced binding of the anti-tetrasaccharide mAbs
dramatically but had little effect on binding anti-disaccharide mAbs.
Similarly, placement of a trimethylacetyl moiety (5 in Figure 1) or
deletion of a 3-hydroxyl group (6 in Figure 1) significantly affected
only anti-tetrasaccharide mAb binding. Therefore, while the an-
throse side chain must be present for the glycan to bind both classes
of mAbs, only the anti-tetrasaccharide mAbs are affected by altering
the specific chemical composition of the side chain, such as
removing the C3 methyl group.
Quantification of the carbohydrate-antibody binding interactions
commenced using SPR analysis, which reconfirmed the tight
interaction (KD ) 9.1 µM) between the anti-tetrasaccharide mAb
MTA1 and its original tetrasaccharide antigen [Table 1 and Figure 2
in the Supporting Information (SI)]. SPR analysis also demonstrated
that this interaction is fast, indeed much faster than binding of anti-
disaccharide antibodies with any ligand (SI Table 1 and SI Figure 2).
Consistent with the microarray results, MTA1 did not bind with
significant affinity any of the other synthetic glycans tested (SI Table
1). The anti-disaccharide mAb MTD6 showed unusually high
affinity for its original disaccharide antigen (KD ) 0.51 µM; SI
Table 1). Few carbohydrate-antibody interactions with KD values
less than 1 µM have been reported, which makes this discovery
particularly significant. Interestingly, the KD values were comparable
for interactions between MTD6 and two structurally diverse
oligosaccharides, the tetrasaccharide (KD ) 3.7 µM) and the
anthrose monosaccharide (KD ) 7.2 µM) (SI Table 1 and SI
Figure 2). On the basis of this small difference in KD, we can
conclude that the rhamnose units in the tetrasaccharide contribute
little to binding MTD6.
9
10240 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

C O M M U N I C A T I O N S
Acknowledgment. We thank the Max-Planck Society, the Swiss
National Science Foundation (SNF Grant 1201260), the Swiss
Tropical and Public Health Institute, and ETH Zu¨rich for generous
support. A Feodor Lynen Fellowship of the Humboldt Foundation
and a DFG Emmy Noether Fellowship (to D.B.W.) are gratefully
acknowledged. A.A. thanks the FCI for a Kekule´ Fellowship. We
thank J. Sobek for assistance with array printing.
Supporting Information Available: Experimental procedures,
characterization data for new compounds, supporting figures and tables,
and complete ref 2 (as SI ref 7). This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Ada, G.; Isaacs, D. Clin. Microbiol. Infect. 2003, 9, 79–85.
(2) Verez-Bencomo, V.; et al. Science 2004, 305, 522–525.
(3) Hecht, M.-L.; Stallforth, P.; Varo´n, D.; Adibekian, A.; Seeberger, P. H.
Curr. Opin. Chem. Biol. 2009, 13, 354–359.
(4) Vliegenthart, J. F. G. FEBS Lett. 2006, 580, 2945–2950.
(5) Serruto, D.; Rappuoli, R. FEBS Lett. 2006, 580, 2985–2992.
(6) (a) Young, W. W., Jr.; Johnson, H. S.; Tamura, Y.; Karlsson, K.-A.; Larson,
G.; Parker, J. M. R.; Khare, D. P.; Spohr, U.; Baker, D. A.; Hindsgaul, O.;
Lemieux, R. U. J. Biol. Chem. 1983, 258, 4890–4894. (b) Lemieux, R. U.
Chem. Soc. ReV. 1989, 18, 347–374. (c) Delbare, L. T. J.; Vandonselaar,
M.; Prasad, L.; Quail, J. W.; Pearlstone, J. R.; Carpenter, M. R.; Smillie,
L. B.; Nikrad, P. V.; Spohr, U.; Lemieux, R. U. Can. J. Chem. 1990, 68,
1116–1121. (d) Lemieux, R. U.; Szweda, R.; Paszkiewicz-Hnatiw, E.;
Spohr, U. Carbohydr. Res. 1990, 205, c12–c17.
(7) Sigurskjold, B. W.; Bundle, D. R. J. Biol. Chem. 1992, 267, 8371–8376.
(8) (a) Vyas, N. K.; Vyas, M. N.; Chervenak, M. C.; Johnson, M. A.; Pinto,
B. M.; Bundle, D. R.; Quiocho, F. A. Biochemistry 2002, 41, 13575–13586.
(b) Johnson, M. A.; Pinto, B. M. Bioorg. Med. Chem. 2003, 12, 295–300.
(c) Mu¨ller-Loennies, S.; Brade, L.; MacKenzie, C. R.; Di Padova, F. E.;
Brade, H. J. Biol. Chem. 2003, 278, 25618–25627. (d) Vulliez-Le, B.; Saul,
F. A.; Phalipon, A.; Be´lot, F.; Guerreiro, C.; Mulard, L. A.; Bentley, G. A.
Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 9976–9981.
(9) (a) Liao, X.; Poirot, E.; Chang, A. H. C.; Zhang, X.; Zhang, J.; Nato, F.;
Fournier, J.-M.; Kova´cˇ, P.; Glaudemans, C. P. J. Carbohydr. Res. 2002,
337, 2437–2442. (b) Provenzano, D.; Kova´cˇ, P.; Wade, W. F. Microbiol.
Immunol. 2006, 50, 899–927.
(10) Nitz, M.; Ling, C.-C.; Otter, A.; Cutler, J. E.; Bundle, D. R. J. Biol. Chem.
2002, 277, 3440–3446.
(11) Steichen, C. T.; Chen, P.; Kearney, J. F.; Turnbough, C. L., Jr. J. Bacteriol.
2003, 185, 1903–1910.
Figure 2. Epitope mapping of the BclA tetrasaccharide 1-MTA1 interac-
tion by STD NMR spectroscopy. Percent STD effects are shown for
individual protons of tetrasaccharide 1. In addition, strong (>10%), medium
(5-10%), and weak (<5%) STD effects are indicated by red, orange, and
yellow spheres of decreasing size. Positions marked with an asterisk could
not be determined with high accuracy because of resonance overlap. The
methyl groups marked with & superscripts were not assigned stereospe-
cifically.
significant STD values (Figure 2). Interestingly, the two methyl
groups have different STD values, indicating that one methyl group
is oriented closer to MTA1 and is bound more tightly. It is
remarkable that this difference in binding affinity was also detected
by microarray screening, where tetrasaccharide 4 containing a
4-(3S)-3-hydroxy-3-methylbutyrate side chain showed stronger
affinity toward an anti-tetrasaccharide mAb, MTA2, than did 3,
which is decorated with a 4-(3R)-3-hydroxy-3-methylbutyrate side
chain (SI Figure 1). We therefore conclude that the methyl group
presented in the S enantiomer, 4, is proximal to the antibody and
thus makes a greater contribution to binding.
(12) Daubenspeck, J. M.; Zeng, H.; Chen, P.; Dong, S.; Steichen, C. T.; Krichna,
N. R.; Pritchard, D. G.; Turnbough, C. L., Jr. J. Biol. Chem. 2004, 279,
30946–30953.
(13) Werz, D. B.; Seeberger, P. H. Angew. Chem. 2005, 117, 6474-6476;
Angew. Chem., Int. Ed. 2005, 44, 6315-6318.
(14) Adamo, R.; Sakesna, R.; Kova´cˇ, P. Carbohydr. Res. 2005, 340, 2579–
2582.
(15) Mehta, A. S.; Saite, E.; Zhong, W.; Buskas, T.; Carlson, R.; Kannenberg,
E.; Reed, Y.; Quinn, C. P.; Boons, G.-J. Chem.sEur. J. 2006, 12, 9136–
9149.
(16) Guo, H.; O’Doherty, G. A. Angew. Chem. 2007, 119, 5298-5300; Angew.
Chem., Int. Ed. 2007, 46, 5206-5208.
The power of systematically combining microarray profiling, SPR
analysis, and STD NMR spectroscopy has been revealed in this
study, where we precisely mapped the molecular elements of the
BclA tetrasaccharide that participate in tight antibody binding.
Understanding which structural features of the oligosaccharide are
most important for this interaction will enable the design of better
carbohydrate-based anthrax vaccines. Furthermore, this study has
illuminated the binding requirements for mAbs that are currently
under development as a highly sensitive spore detection system.25
The approach, however, is a more general tool that we hope could
ultimately help to elucidate the general principles of carbohydrate-
antibody interactions, enabling guided structure-based design of a
broad spectrum of carbohydrate-based antigens and therapeutics.
(17) Crich, D.; Vinogradova, O. J. Org. Chem. 2007, 72, 6513–6520.
(18) Werz, D. B.; Adibekian, A.; Seeberger, P. H. Eur. J. Org. Chem. 2007,
1976–1982.
(19) Tamborrini, M.; Werz, D. B.; Frey, J.; Pluschke, G.; Seeberger, P. H. Angew.
Chem. 2006, 118, 6731-6732; Angew. Chem., Int. Ed. 2006, 45, 6581-
6582.
(20) Tamborrini, M.; Oberli, M. A.; Werz, D. B.; Schu¨rch, N.; Frey, J.;
Seeberger, P. H.; Pluschke, G. J. Appl. Microbiol. 2009, 106, 1618–1628.
(21) Kuehn, A.; Kova´cˇ, P.; Saksena, R.; Bannert, N.; Klee, S. R.; Ranisch, H.;
Grunow, R. Clin. Vaccine Immunol. 2009, 16, 1728–1737.
(22) Adams, E. W.; Ratner, D. M.; Bokesch, H. R.; McMahon, J. B.; O’Keefe,
B. R.; Seeberger, P. H. Chem. Biol. 2004, 11, 875–881.
(23) Mayer, M.; Meyer, B. Angew. Chem. 1999, 111, 1902-1906; Angew.
Chem., Int. Ed. 1999, 38, 1784-1788.
(24) Meyer, B.; Peters, T. Angew. Chem. 2003, 115, 890-918; Angew. Chem.,
Int. Ed. 2003, 42, 864-890.
(25) Balillie, L. J. Appl. Microbiol. 2001, 91, 609–613.
JA104027W
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10241