Breaking pseudo-symmetry in multiantennary complex N-glycans using lanthanide-binding tags and NMR pseudo-contact shifts

Article abstract of DOI:10.1002/anie.201307845

Controlling NMR shifts by lanthanides tagged to a "symmetrical" N-glycan (see picture) reveals individual resonances for the residues of the otherwise identical A and B arms. This method provides a global perspective of conformational features and interac

Full text of DOI:10.1002/anie.201307845

Angewandte
Chemie
DOI: 10.1002/anie.201307845
NMR Spectroscopy Very Important Paper
Breaking Pseudo-Symmetry in Multiantennary Complex N-Glycans
Using Lanthanide-Binding Tags and NMR Pseudo-Contact Shifts**
Angeles Canales, Alvaro Mallagaray, Javier Pꢀrez-Castells, Irene Boos, Carlo Unverzagt,
Sadine Andrꢀ, Hans-Joachim Gabius, Francisco Javier CaÇada, and Jesffls Jimꢀnez-Barbero*
Molecular recognition is of vital significance for life. Under-
standing the chemical basis of these interactions between
cellular receptors and their specific ligands not only gives
a functional meaning to structures and changes occurring in
diseases but also helps devise innovative therapeutic
approaches. In terms of biological coding and translating
signals into cellular effects, glycans have gained a particular
status, owing to their unsurpassed coding capacity and
widespread presence of receptors (lectins) to read the
encoded information.[1] The glycan sequence and shape and
the intimate interplay between a glycan determinant and its
cognate lectin ensure the flow of information of sugar
coding.[2] Thus, in addition to their peptide scaffold, glyco-
proteins carry a second source of bioinformation in their
glycan chains, realized, for example, by recognition of N-
glycans by lectins.[3]
charide core composed of an N,N’-diacetyl chitobiose (N-
glycosidically linked to the Asn at the protein chain)
substituted at position O-4 of its nonreducing moiety by
a trimannoside, as shown in Scheme 1. Further extensions of
the external Man A and Man B residues generate the differ-
ent N-glycan classes, and internal substituents at the core
further increase the panel size of N-glycans.[4,5]
The conformational features of the different classes of N-
glycans (from complex- to high-mannose type) have been
investigated using different techniques, especially NMR
spectroscopy,[6] since the flexibility of the glycosidic linkages
has usually hampered their study by X-ray crystallography.
However, typical NMR parameters (J couplings and NOEs)
provide only short-range structural information (within a few
ꢀngstrçm).[7] Therefore, it is not easy to define the global
shape of these nonglobular and relatively flexible molecules.
Of course, NOEs and J couplings may define possible
accessible conformations around each glycosidic linkage,
but they cannot provide a global conformational model in
solution. Especially, in our target molecule 1 (Scheme 1), the
existence of eight glycosidic linkages amplifies the level of
uncertainty in the relative positions of the two terminal Gal
(A and B) moieties with respect to the reducing-end GlcNAc
unit. In general, the limited number of available interresidual
restraints in symmetrical complex-type N-glycans and the
recurrent signal overlapping, precludes obtaining nonambig-
uous NMR parameters with conformational information.
In this context, we herein present a novel NMR approach
to individually monitor the behavior of each arm, A and B, of
N-glycans (Scheme 1) and thereby provide a global perspec-
tive of their conformational and interaction features in
solution. Structurally, N-glycans have a common pentasac-
[*] Dr. A. Mallagaray, Prof. Dr. F. J. CaÇada, Prof. Dr. J. Jimꢀnez-Barbero
Department of Chemical and Physical Biology
Centro de Investigaciones Biolꢁgicas, CSIC
Ramiro de Maeztu, 9, CP, 28040 Madrid (Spain)
E-mail: jjbarbero@cib.csic.es
1
In fact, for 1, all the H and 13C chemical shifts of the
Homepage: http://www.cib.csic.es
Dr. A. Canales
terminal Gal and GlcNAc moieties as well as those of the C3,
C4, and C6 atom pairs of Man residues A and B are
isochronous. Therefore, the individual behavior of the two
inner glycosidic linkages of the A and B arm cannot be
estimated by typical NMR analysis. Differing only in their
attachment points, compound 1 displays a pseudo-symmetry
for the two arms, which renders them indistinguishable by
conventional NMR techniques. However, biological recog-
nition of N-glycans tends to be branch-specific in many cases.
Despite the lack of molecular rationales for these findings,
sialylation, galactosylation, and lectin recognition of N-
glycans were found to show branch specificity.[8]
Departamento de Quꢂmica Orgꢃnica I, Fac. C. C. Quꢂmicas
Universidad Complutense de Madrid, 28040 Madrid (Spain)
I. Boos, Prof. Dr. C. Unverzagt
Lehrstuhl fꢄr Bioorganische Chemie, Gebꢅude NWI
Universitꢅt Bayreuth, 95444 Bayreuth (Germany)
Prof. Dr. J. Pꢀrez-Castells
Departamento de Quꢂmica, Facultad de Farmacia
Universidad San Pablo-CEU, Boadilla del Monte
28668 Madrid (Spain)
Dr. S. Andrꢀ, Prof. Dr. H.-J. Gabius
Institute of Physiological Chemistry, Faculty of Veterinary Medicine
Ludwig-Maximilians-University, 80539 Munich (Germany)
[**] We acknowledge financial support by grants from Spain CTQ2012-
32025 (JJB) and CTQ2012-31063 (JPC), from CAM S2010/BMD-
2353 (J.J.B. and A.C.), from Deutsche Forschungsgemeinschaft
(CU) the GlycoHit and Glycopharm EU projects, BM1003 and
CM1102 COST actions. We thank CESGA Supercomputing Center
for computational resources and UCM CAI-NMR. We also thank Dr.
G. Domꢂnguez (USP-CEU) and Dr. A. Vꢃzquez (U. Vigo) for
discussions.
Our method is based on exploiting pseudo-contact shifts
(PCS) induced by paramagnetic lanthanide ions.[9,10] Recent
studies have demonstrated their applicability in carbohydrate
conformational analysis of small oligosaccharides.[11–15] It is
well-known that paramagnetic metal ions (such as most
lanthanides) cause PCS, enhanced nuclear relaxation (PRE),
and molecular alignment in the presence of magnetic fields.[16]
Lanthanides offer a wide range of paramagnetic properties,
and thus allow access to detailed structural information,
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201307845.
Angew. Chem. Int. Ed. 2013, 52, 13789 –13793
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
13789

.
Angewandte
Communications
Mana1!6Man linkage. According to
the MD data, all the F glycosidic
torsions were defined by the exo-
anomeric effect, while the rest of the
y angles displayed moderate flexibility
around the global minimum. The differ-
ent conformers of the Mana1!6Man
linkage were populated between 11%
(“extended gt conformer”, with y 1808,
w 1808) and 41% (“back-folded” con-
former, with y 908, w 608). Additional
contributions of the “extended gg”
(15%, with y 1808, w 608), “half back-
folded” (15%, with y 608, w 608), and
“tight back-fold” (17%, y 608, w 1808)
were also predicted by the simula-
tions.[19]
Thus, the observed PCS were inter-
preted considering the presence of the
postulated low-energy conformers[19]
for the oligosaccharide of 1 (Figure 1).
Inspection of the geometries of the five
different conformers indicated that the
Scheme 1. a) Common pentasaccharide core of N-glycans. The torsional angles that define the
conformation around the Mana1-6Man linkage are defined. b) Nonasaccharide derivatives
studied in this work. The 1–3 and 1–6 arms attached to the b-mannose unit are labeled as A and
B arms, respectively.
because of the spatial anisotropy of the induced paramagnetic
effects.[17]
In the context of oligosaccharides, lanthanide-binding tags
have been designed to decrease the effect of the paramagnetic
relaxation enhancement (PRE), by separating the metal from
the biomolecule by a rigid structure.[11–13] The synthesis of
1 was based on the corresponding anomeric azide[18] and
planned following our previously tested strategy (see the
Supporting Information).
As first step, 1H, 1H–13C HSQC and HSQC-TOCSY NMR
spectra were acquired for 1 in the presence of one molar
equivalent of either La3+ (1a) as diamagnetic reference or
Dy3+ (1b) as paramagnetic ion (Figure 2). The 1H/13C signals
of the terminal Gal, GlcNAc, and Man (3, 4, and 6 positions)
have isochronous chemical shifts in the HSQC spectra of both
1 and 1a. In the case of 1b, the aromatic signals as well as
those of the GlcNAc-1 residue (and most of GlcNAc-2)
disappeared since they are very close to the metal, while the
other sugar signals were clearly visible in the HSQC spectra.
Significant shifts were observed because of the presence of
Figure 1. Minimum-energy conformations of free 1 according to MD
calculations.[19] Structures were built and energy minimized using
MacroModel/Batchmin package (version 9.6) and the MM3* force
field.[20] The B arm trisaccharide is shown in black.
the paramagnetic ion and, strikingly, signal separation was
seen in the Dy3+-containing sample 1b for the identical
residues of the A and B arms (see Figure 2). Therefore, in the
1
presence of the paramagnetic metal, the H and 13C reso-
nances for each a-Man, GlcNAc, and Gal residues at either A
1
or B arms can be easily distinguished. In fact, 34 H NMR
nuclei of the B arm are always closer to the paramagnetic
metal than their analogs at the A arm, independently of the
chosen conformation. Therefore, as first approximation, the
peaks arising from the larger PCS were assigned to the nuclei
of the B-arm. This initial assumption was backed by
comparison of the back-calculated PCS values with the
experimental ones.
PCS (see Table S1 in the Supporting Information) could be
measured in a nonambiguous manner, ranging from 0.03 (H4
Gal A) to 0.94 ppm (H3 GlcNAc-2). The estimated distance
between H4 Gal A and the metal is larger than 30 ꢀ.
Recently, a detailed theoretical MD analysis of the
biantennary N-glycan part of 1 has been presented.[19] This
work describes the existence of five conformers (Figure 1) in
equilibrium, especially differing in the y and w angles of the
The individual fittings for the five independent con-
formers of 1[19] using MSpin software[21] are given in Table S1.
13790 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13789 –13793

Angewandte
Chemie
Figure 2. Pseudo-contact shifts arise from dipolar interactions between the unpaired electrons of a metal ion and the nuclei in the vicinity. PCS
give rise to changes in chemical shifts that decrease with the distance between the metal ion and the nuclear spin with a 1/r3 dependence. PCS
are measured by comparison of the spectrum of a complex with a diamagnetic ion with that of a paramagnetic ion. a) PCS values measured for
the two arms of 1 are shown as bar graphs. b) 1H-13C HSQC spectrum of 1a, (diamagnetic reference, Ln3+ =La3+). c) 1H–13C HSQC spectrum of
1b, (paramagnetic sample, Ln3+ =Dy3+). The signals of Gal, GlcNAc and aMan of arms A and B (inside the box in 2a) display identical chemical
shifts in the reference spectrum. Fittingly, they can be individually observed in the presence of Dy3+. d) Superimposition of the 1H–13C HSQC
spectra (600 MHz) acquired for 1a and 1b. The PCS for some of the signals are assigned.
Angew. Chem. Int. Ed. 2013, 52, 13789 –13793
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13791

.
Angewandte
Communications
When individually taken, none of them provided a satisfactory
fit to the experimental PCS, especially for the B arm. The
comparison of the experimental versus calculated PCS values
for the extended and half backfolded conformations (see
Figure 1) are shown in Figure 3. Indeed, the global-minimum
geometry for all glycosidic torsions constituting the A-arm
gave a very good fit between expected and observed PCS
(Figure 3a), independently of the conformation at the B arm.
Regarding the B arm, the best Q factor (0.145) and SVD
condition (28.75) was obtained for the “extended gg”
conformer (green dots in Figure 3b). The predicted global
minimum by MD [9] gave a poorer fitting (Q factor of 0.451).
Overall, a good correlation between experimental and back-
calculated PCSs could not be obtained considering just one
conformer.
We thus applied the differentiation of the NMR signals of
the A and B arms by PCS to biorecognition. As proof-of-
principle, we tested the carbohydrate recognition domain
(CRD) of human galectin-3 (hGal3), a galactose-binding
protein with affinity to N-glycans.[22] Addition of hGal3 to
a sample containing paramagnetic 1b (protein:1b = 1:8)
induced further line broadening of the carbohydrate signals
(Figure S1a). Clear changes were detected for the H1 and H4
protons of the Gal and GlcNAc units at the termini of both
arms (positions 5, 6, 5’, and 6’). In addition, the CH3 groups of
the GlcNAc 5 and 5’ also displayed clear line broadening.
HSQC spectra of both samples were then acquired for
further analysis of those signals that overlapped in the
1D NMR spectra. The analysis of the HSQC peaks allowed
confirming the specific interaction of hGal3 with the Gal and
GlcNAc residues at both ends of the A and B arms. This result
is in line with data from precipitation analysis of N-glycans
with a related receptor from the galectin family.[23] The signals
of the b-Man unit of the core were not affected by the
presence of the protein (Figure S1b).
In conclusion, the use of the tagged nonasaccharide
derivative 1 loaded with a paramagnetic ion (1b Dy3+) has
allowed to unequivocally defining the conformational behav-
ior of this flexible molecule in solution for the first time. The
use of the lanthanide tag has permitted to break the inherent
pseudo-symmetry of the NMR spectra of the identical
branches, revealing that the T-shaped gg rotamer at the
Mana1-6Man junction is the major one in solution, with
minor contributions of other backfolded geometries. This
approach has permitted the characterization of the binding
epitopes of the symmetrical N-glycan 1, showing that both
arms are involved in the recognition of human galectin-3. Our
approach provides new tools to further pinpoint the relevance
of multiantennary N-glycans as biomedically important
receptors.
Received: September 6, 2013
Published online: November 4, 2013
Keywords: lanthanides · molecular recognition · N-glycans ·
NMR spectroscopy · pseudo-contact shifts
.
Figure 3. a,b) Correlation between experimental and computed PCSs
for the protons at the “A” and “B” arms. PCSs calculated from the
individual fitting of the structures of the extended conformations
(green, gg (y 180); red, gt (y 180); blue, gg (y 60), half backfolded
conformation). c) Correlation between experimental and computed
PCSs for 1 calculated considering the presence of four conformations
(gg (y 180), gt (y 180), gg (y 60), and gg (y 90)).
[1] H.-J. Gabius, S. Andrꢁ, J. Jimꢁnez-Barbero, A. Romero, D. Solꢂs,
Trends Biochem. Sci. 2011, 36, 298 – 313.
[2] a) The Sugar Code. Fundamentals of glycosciences (Ed.: H.-J.
Gabius), Wiley-VCH, Weinheim, 2009; b) V. Roldꢃs, F. J.
CaÇada, J. Jimꢁnez-Barbero, ChemBioChem 2011, 12, 990 –
1005.
[3] a) S. Andrꢁ, T. Kozꢄr, S. Kojima, C. Unverzagt, H.-J. Gabius,
Biol. Chem. 2009, 390, 557 – 565; b) C. Zuber, J. Roth in The
Sugar Code. Fundamentals of glycosciences (Ed.: H.-J. Gabius),
Wiley-VCH, Weinheim, 2009, pp. 87 – 100; c) M. Amano, H.
Eriksson, J. C. Manning, K. M. Detjen, S. Andrꢁ, S.-I. Nishimura,
J. Lethiç, H.-J. Gabius, FEBS J. 2012, 279, 4062 – 4080; d) A.
Ardꢄ, D. V. Silva, V. Schubert, S. Andrꢁ, M. Bruix, F. J. CaÇada,
H.-J. Gabius, C. Unverzagt, J. Jimꢁnez-Barbero, J. Am. Chem.
Soc. 2013, 135, 2667 – 2675.
Then, combinations of the different geometries were
considered and their corresponding fittings were calculated
by MSpin. Avery good correlation with the experimental data
(Q factor of 0.10, Figure 3c) was obtained when four con-
formers were included, with the “extended gg” and “extended
gt” as the major contributors (about 45% and 35%,
respectively). The presence of the half-backfolded and back-
folded conformations (10% each) gave the best fit. The
employed protocol enables to experimentally define the
conformational behavior in solution of the A and B arms of
the complex-type N-glycan (1) in an independent manner.
[4] a) M. Aebi, R. Bernasconi, C. Clerc, M. Molinari, Trends
Biochem. Sci. 2010, 35, 74 – 82; b) A. Antonopoulos, S. J.
13792 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13789 –13793

Angewandte
Chemie
North, S. M. Haslam, A. Dell, Biochem. Soc. Trans. 2011, 39,
1334 – 1340.
[12] A. Mallagaray, A. Canales, G. Dominguez, J. Jimenez-Barbero,
J. Perez-Castells, Chem. Commun. 2011, 47, 7179 – 7181.
[13] S. Yamamoto, T. Yamaguchi, M. Erdelyi, C. Griesinger, K. Kato,
Chem. Eur. J. 2011, 17, 9280 – 9282.
[14] S. Yamamoto, Y. Zhang, T. Yamaguchi, T. Kameda, K. Kato,
Chem. Commun. 2012, 48, 4752 – 4754.
[15] Y. Zhang, S. Yamamoto, T. Yamaguchi, K. Kato, Molecules 2012,
17, 6658 – 6671.
[16] G. Otting, J. Biomol. NMR 2008, 42, 1 – 9.
[5] a) S. Andrꢁ, C. Unverzagt, S. Kojma, X. Dong, C. Fink, K.
Kayser, H.-J. Gabius, Bioconjugate Chem. 1997, 8, 845 – 855;
b) C. Unverzagt, S. Andrꢁ, J. Seifert, S. Kojima, C. Fink, G.
Srikrishna, H. H. Freeze, K. Kayser, H.-J. Gabius, J. Med. Chem.
2002, 45, 478 – 491; c) S. Andrꢁ, C. Unverzagt, S. Kojima, M.
Frank, J. Seifert, C. Fink, K. Kayser, C.-W. von der Lieth, H.-J.
Gabius, Eur. J. Biochem. 2004, 271, 118 – 134; d) S. Andrꢁ, T.
Kozꢄr, R. Schuberth, C. Unverzagt, S. Kojima, H.-J. Gabius,
Biochemistry 2007, 46, 6984 – 6995.
[17] I. Bertini, C. Luchinat, G. Parigi, R. Pierattelli, Dalton Trans.
2008, 3782 – 3790.
[6] S. W. Homans, R. A. Dwek, T. W. Rademacher, Biochemistry
1987, 26, 6571 – 6578.
[7] M. R. Wormald, A. J. Petrescu, Y. L. Pao, A. Glithero, T. Elliott,
R. A. Dwek, Chem. Rev. 2002, 102, 371 – 386.
[18] V. Ullmann, M. Rꢆdisch, I. Boos, J. Freund, C. Pçhner, S.
Schwarzinger, C. Unverzagt, Angew. Chem. 2012, 124, 11734 –
11738; Angew. Chem. Int. Ed. 2012, 51, 11566 – 11570.
[19] a) W. Nishima, N. Miyashita, Y. Yamaguchi, Y. Sugita, S. Re, J.
Phys. Chem. B 2012, 116, 8504 – 8512; b) S. Re, N. Miyashita, Y.
Yamaguchi, Y. Sugita, Biophys. J. 2011, 101, L44 – 46.
[20] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J.
Comput. Chem. 1990, 11, 440 – 467.
[8] a) D. H. Jozaisse, W. E. C. M. Schiphorst, D. H. van den Eijn-
den, J. A. van Kuik, H. van Halbeek, J. F. G. Vliegenthart, J.
Biol. Chem. 1985, 260, 714 – 719; b) A. M. Barb, E. K. Brady,
J. H. Prestegard, Biochemistry 2009, 48, 9705 – 9707; c) V.
Ramasamy, B. Ramakrishnan, E. Boeggeman, D. M. R, P. H.
Seeberger, P. K. Qasba, J. Mol. Biol. 2005, 353, 53 – 67; d) S. Guo,
T. Sato, K. Shirane, K. Furukawa, Glycobiology 2001, 11, 813 –
820; e) Y. Kaneda, R. F. Whittier, H. Yamanaka, E. Carredano,
M. Gotoh, H. Sota, Y. Hasegawa, Y. Shinohara, J. Biol. Chem.
2002, 277, 16928 – 16935.
[21] http://www.mestrelab.com Program MSpin.
[22] a) H. Sanchez-Ruderisch, C. Fischer, K. M. Detjen, M. Welzel,
A. Wimmel, J. C. Manning, S. Andrꢁ, H. J. Gabius, FEBS J. 2010,
277, 3552 – 3563; b) M. Krzeminski, T. Singh, S. Andrꢁ, M.
Lensch, A. M. Wu, A. M. J. J. Bonvin, H.-J. Gabius, Biochim.
Biophys. Acta Gen. Subj. 2011, 1810, 150 – 161; c) H. Dawson, S.
Andrꢁ, E. Karamitopoulou, I. Zlobec, H.-J. Gabius, Anticancer
Res. 2013, 33, 3053 – 3060.
[9] T. H. Nguyen, K. Ozawa, M. Stanton-Cook, R. Barrow, T.
Huber, G. Otting, Angew. Chem. 2011, 123, 718 – 720; Angew.
Chem. Int. Ed. 2011, 50, 692 – 694.
[10] G. Pintacuda, M. John, X. C. Su, G. Otting, Acc. Chem. Res.
2007, 40, 206 – 212.
[11] M. Erdꢁlyi, E. dꢅAuvergne, A. Navarro-Vꢄzquez, A. Leonov, C.
Griesinger, Chem. Eur. J. 2011, 17, 9368 – 9376.
[23] a) D. Gupta, H. Kaltner, X. Dong, H.-J. Gabius, C. F. Brewer,
Glycobiology 1996, 6, 843 – 849; b) H. Kaltner, H.-J. Gabius,
Histol. Histopathol. 2012, 27, 397 – 416.
Angew. Chem. Int. Ed. 2013, 52, 13789 –13793
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13793