An accessible bicyclic architecture for synthetic lectins

Article abstract of DOI:10.1039/c3cc41190g

Bicyclic carbohydrate receptors are easier to synthesise than tri- or tetra-cyclic relatives, and are better adapted to bind monosaccharide residues with bulky appendages. Disaccharides containing β-glucosyl units are preferred substrates.

Full text of DOI:10.1039/c3cc41190g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
An accessible bicyclic architecture for synthetic lectins†
Joshua D. Howgego, Craig P. Butts, Matthew P. Crump and Anthony P. Davis*
Cite this: Chem. Commun., 2013,
49, 3110
Received 13th February 2013,
Accepted 4th March 2013
DOI: 10.1039/c3cc41190g
www.rsc.org/chemcomm
Bicyclic carbohydrate receptors are easier to synthesise than tri- or required, so that unbound portions of the substrate can protrude
tetra-cyclic relatives, and are better adapted to bind monosaccharide from the cavity without steric clashes. Herein we describe a new
residues with bulky appendages. Disaccharides containing b-glucosyl bicyclic architecture for synthetic lectins, possessing larger portals
units are preferred substrates.
than earlier tri- or tetracyclic systems. The system retains the ability
to bind carbohydrates in water despite increased flexibility, and is
Biomimetic carbohydrate recognition remains a significant problem particularly effective for sterically demanding substrates. Moreover,
for supramolecular chemistry.1 There is general interest in binding it is easier to synthesise than more highly-connected relatives, and
polar molecules in water, and saccharides are especially challenging thus has greater promise for variation and optimisation.
due to their hydromimetic2 character. Indeed, the affinities shown
by carbohydrate-binding proteins (lectins) are notably weak, at least
to monosaccharides.3 Even so, carbohydrate recognition is an
important natural phenomenon, playing roles in many biological
processes.4 Synthetic carbohydrate receptors can throw light on the
principles which underlie lectin binding, could potentially comple-
ment these proteins as research tools for glycobiology, and may
eventually find applications in medicine.1c,5
Among the more effective ‘‘synthetic lectins’’ are the macropoly-
cyclic oligoamides developed in our group.6 Targeted at carbo-
hydrates with all-equatorial substitution patterns, these structures
employ parallel biphenyl or terphenyl units to complement the axial
carbohydrate CH groups, and polar spacers (mostly isophthal-
amides) to match polar substituents. The biphenyl-based systems
(e.g. 16a,c) are effective for monosaccharides such as N-acetyl-
glucosamine (GlcNAc) 3 and glucose 5, while the terphenyl-
based structures (e.g. 26f) are complementary to disaccharides
such as N,N0-diacetylchitobiose 4 and cellobise 6.
A feature common to these receptors has been the ability to
surround the mono- or disaccharide target. This has conferred
interesting size-selectivity; for example, 1 binds 3 but not 4, while 2
shows a strong preference for 6 vs. 5 or 7. However, the cost has
been limited versatility. For applications such as glycoprofiling, it
may be useful to bind just part of an oligosaccharide structure, for
example a specific terminal unit. For this, a more open structure is
School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS1 1TS, UK.
E-mail: Anthony.Davis@bristol.ac.uk; Fax: +44 (0)117 9298611;
The new family of receptors is exemplified by prototype 8.
Tel: +44 (0)117 9546334
Molecular modelling7 of 8 predicts generally similar properties
† Electronic supplementary information (ESI) available: Experimental and mod-
to 1, i.e. an open binding site in water appropriately sized for a
monosaccharide and apparently compatible with all-equatorial
elling details, spectra for compound characterisation and binding studies, and
fitting curves for binding analyses. See DOI: 10.1039/c3cc41190g
c
3110 Chem. Commun., 2013, 49, 3110--3112
This journal is The Royal Society of Chemistry 2013

View Article Online
Communication
ChemComm
Receptor 8 was prepared as shown in Scheme 1. Compared
to the synthesis of tricyclic 1, the sequence is considerably more
straightforward. In particular, the biphenyls are initially joined
by a simple linear coupling which proceeds in 90% yield. The
corresponding step in the synthesis of 1 is a high-dilution
macrocyclisation (B50% yield after preparative HPLC).
1H NMR spectra of 8 (r0.5 mM) in D2O were well-resolved
and invariant with concentration, suggesting that the system is
monomeric under these conditions.
Binding of 8 to carbohydrates in D2O was studied using
1H NMR titrations.8 In many cases, addition of the substrates
caused significant changes to the spectrum of the receptor,
including signal movements of up to 0.25 ppm and loss of
some equivalences.6a Where feasible, the movements were
analyzed according to a 1 : 1 binding model, giving mostly good
fits with internal consistency. The resulting binding constants
Ka are listed in Table 1, along with the corresponding data for 1.
For the hexose monosaccharides such as glucose and GlcNAc,
the change from tri- to bicyclic architecture results in a low-
ering of affinities. This is perhaps unsurprising given the
reduction in preorganisation (for example, the increased rota-
tional freedom within the biphenyl units9). Interestingly, pen-
toses such as xylose are bound slightly more strongly, perhaps
reflecting a decrease in the average size of the cavity. However,
more significant is the increased affinity for disaccharide
substrates. Bicyclic receptor 8 shows moderate-good affinities
for cellobiose 6, lactose 12, maltose 13 and gentiobiose 14. By
contrast the tricyclic system 1 fails to bind lactose and maltose,
and is three times less effective for cellobiose. The obvious
interpretation is that, unlike 1, receptor 8 can bind b-glucosyl
residues in a wide variety of structural environments. Although
the cavity cannot accept more than one monosaccharide
residue (Fig. 1), interactions to parts of a second residue can
presumably enhance binding. Although more forgiving than 1,
receptor 8 still discriminates between disaccharides, failing to
bind sucrose 15 (which lacks b-glucosyl). Lastly the small but
Fig. 1 Model of 8 binding b-D-cellobiose 6 (magenta), viewed from above the
plane of the biphenyl units. Side chains are omitted, and the biphenyls are shown
with transparent CPK surfaces.
substrates. However clear differences emerge with disaccharide
substrates. As illustrated in Fig. 1, receptor 8 can bind a
glucosyl residue in cellobiose 6 while leaving ample space
for the unbound monosaccharide unit. In corresponding struc-
tures of 1 + 6, the unbound glucosyl is hindered and twisted
from its preferred orientation.7 Similar disparities are observed
for 8 and 1 binding maltose 13.
Table 1 Association constants (Ka) for receptors 8 and 1 with carbohydratesa
Ka [Mꢀ1
]
Carbohydrate
8a
1b
D-Glucose 5
Me b-D-glucoside
D-Mannose
D-Galactose
D-Xylose
4
5
B0
B0
8
9
27
B0
2
5
D-Lyxose
2-Deoxy-D-glucose
D-Ribose
B0
6
7
B0
7
3
N-Acetyl-D-glucosamine 3 (GlcNAc)
D-Cellobiose 6
D-Lactose 12
D-Maltose 13
D-Gentiobiose 14
D-Sucrose 15
19
52
37
14
22
B0
7
56
17
B0
B0
n.m.
n.m.
B0
N-Acetylneuraminic acid (Neu5Ac) 16
Scheme 1 (i) Pd(PPh3)3, K3PO3, 80 1C, DMF, 4 h, 57%; (ii) TFA, CH2Cl2, 0 1C, 3 h,
76%; (iii) 11, DIPEA, THF, 3 h, 90%; (iv) PPh3, THF-H2O, 60 1C, 12 h, 86%; (v) 11,
DIPEA, THF (high dilution), RT, 5 days, 19%; (vi) TFA, CH2Cl2, 0 1C, 3 h; (vii) NaOH
(aq), B100%.
a
Determined by 1H NMR titration. Values denoted B0 were too small
for analysis. Errors were estimated at rꢁ10% for most cases where
Ka Z 10 Mꢀ1
.
These values have been previously reported (see ref. 6c).
b
n.m. = not measured.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3110--3112 3111

View Article Online
ChemComm
Communication
measurable affinity for Neu5Ac 16, another bulky substrate, Unfortunately it seems that polar interactions to the pyridyl
provides further evidence for receptor flexibility.
nitrogen are not significant, at least for the substrates tested.
In conclusion, we have shown that bicyclic designs are viable for
biphenyl-based synthetic lectins, and can provide complementary
binding properties. The systems discussed in this paper retain the
ability to bind carbohydrates in water and (in the case of 8) show
improved performance with bulky substrates. The new receptors
are more accessible than their predecessors, and therefore more
Following our normal practice we attempted to find alternative tuneable. Their activity towards disaccharides suggests potential for
methods which could be used to support the data in Table 1. this architecture in the binding of oligosaccharides, the key long-
Unfortunately UV-vis/fluorescence titrations or isothermal titra- term objective for work on synthetic lectins.
tion calorimetry (ITC) could not be applied. However, addition of
This work was supported by the Engineering and Physical
cellobiose and maltose to 8 caused measurable induced circular Sciences Research Council (grant number EP/D060192/1 and
dichroism (ICD), which could be analysed to give Ka values DTA studentship to JH). JH is grateful to J.-M. Casas-Solvas and
consistent with those from NMR. In the case of Neu5Ac 16 we C. Ke for useful discussions.
also employed a less usual method, that of 13C NMR titration. It is
Notes and references
well established that changes in 13C chemical shifts can be used to
detect binding, and there is precedent for the determination of
binding constants in biochemical systems using signals from
13C-enriched substrates.10 On the other hand, as far as we can
determine, the method has not been applied to unenriched
synthetic supramolecular systems. In the present case we
observed the 13C NMR signals of 8 (2 mM in D2O) on addition
of NeuAc (6 aliquots) up to 0.22 M. Movements of up to 0.3 ppm
were observed for several aromatic carbons. These were readily
followed and analysed to yield Ka = 5.5 Mꢀ1, close to the value
achieved using 1H NMR. The titration took several days, as it was
performed with a proton-sensitive probe that was not ideal for the
task.11 However, the use of a 13C-observe cryoprobe should reduce
this time to a few hours. As such probes become more widely
available, 13C NMR titration could become a routine method for
quantitative binding studies in supramolecular chemistry.
Finally, as mentioned earlier, an advantage of the bicyclic
architecture is the ease of synthesis and thus the potential for
variation. To illustrate the possibilities we prepared a second
receptor 9 in which the p-phenylene unit was replaced by a
2,5-linked pyridylene. An interesting feature of this structure
was the ability of the pyridyl nitrogen to reach inwards towards
the substrate, potentially assisting binding through polar inter-
actions. Receptor 9 was synthesized in B7 steps from commer-
cially available starting materials 17 and 18.7 Binding of 9 to
carbohydrates was studied by 1H NMR titrations as for 8.
Disappointingly the pyridyl nitrogen produced a uniformly
negative effect on binding, resulting in a general lowering of
1 For reviews, see: (a) A. P. Davis and R. S. Wareham, Angew. Chem., Int.
Ed., 1999, 38, 2978–2996; (b) A. P. Davis, Org. Biomol. Chem., 2009, 7,
3629–3638; (c) S. Jin, Y. F. Cheng, S. Reid, M. Y. Li and B. H. Wang,
Med. Res. Rev., 2010, 30, 171–257; (d) M. Mazik, Chem. Soc. Rev., 2009,
38, 935–956; (e) Y. Nakagawa and Y. Ito, Trends Glycosci. Glycotechnol.,
2012, 24, 1–12. For recent examples, see: ( f ) C. Nativi, O. Francesconi,
G. Gabrielli, I. De Simone, B. Turchetti, T. Mello, L. D. C. Mannelli,
C. Ghelardini, P. Buzzini and S. Roelens, Chem.–Eur. J., 2012, 18,
5064–5072; (g) G. Fukuhara and Y. Inoue, J. Am. Chem. Soc., 2011, 133,
768–770; (h) A. Pal, M. Berube and D. G. Hall, Angew. Chem., Int. Ed.,
2010, 49, 1492–1495; (i) M. Rauschenberg, S. Bomke, U. Karst and
B. J. Ravoo, Angew. Chem., Int. Ed., 2010, 49, 7340–7435.
2 C. Ke, H. Destecroix, M. P. Crump and A. P. Davis, Nat. Chem., 2012,
4, 718–723.
3 E. J. Toone, Curr. Opin. Struct. Biol., 1994, 4, 719–728.
4 (a) The Sugar Code – Fundamentals of Glycoscience, ed. H.-J. Gabius,
Wiley-Blackwell, Weinheim, 2009; (b) B. Wang and G.-J. Boons, Carbo-
hydrate Recognition: Biological Problems, Methods and Applications,
´
Wiley, Hoboken, 2011; (c) H. J. Gabius, S. Andre, J. Jimenez-Barbero,
A. Romero and D. Solis, Trends Biochem. Sci., 2011, 36, 298–313.
5 J. Balzarini, Nat. Rev. Microbiol., 2007, 5, 583–597.
6 (a) E. Klein, M. P. Crump and A. P. Davis, Angew. Chem., Int. Ed.,
2005, 44, 298–302; (b) Y. Ferrand, M. P. Crump and A. P. Davis,
Science, 2007, 318, 619–622; (c) Y. Ferrand, E. Klein, N. P. Barwell,
´
M. P. Crump, J. Jimenez-Barbero, C. Vicent, G.-J. Boons, S. Ingale
and A. P. Davis, Angew. Chem., Int. Ed., 2009, 49, 1775–1779;
(d) N. P. Barwell, M. P. Crump and A. P. Davis, Angew. Chem., Int.
Ed., 2009, 48, 7673–7676; (e) N. P. Barwell and A. P. Davis, J. Org.
Chem., 2011, 76, 6548–6557; ( f ) B. Sookcharoenpinyo, E. Klein,
Y. Ferrand, D. B. Walker, P. R. Brotherhood, C. Ke, M. P. Crump
and A. P. Davis, Angew. Chem., Int. Ed., 2012, 51, 4586–4590.
7 For further details see ESI†.
8 P. Thordarson, Chem. Soc. Rev., 2011, 40, 1305–1323.
9 In models of unbound 8, the internal dihedral angles of the biphenyl
groups are B50–551 (see Fig. S39, ESI†). On binding to cellobiose these
are reduced by B151, representing an energetic cost of B4 kJ molꢀ1
.
affinities. Binding constants to cellobiose and maltose were 14 10 K. J. Neurohr, N. M. Young, I. C. P. Smith and H. H. Mantsch,
and 10 Mꢀ1 respectively, while the affinity for glucose was too
Biochemistry, 1981, 20, 3499–3504; E. Berman, J. H. Brown, H. Lis
and N. Sharon, Eur. J. Biochem., 1985, 152, 447–451; M. A. Kenyon
weak to be measured.7 Binding to galactose and mannose, with
and J. A. Hamilton, J. Lipid Res., 1994, 35, 458–467.
axial OH groups, was also negligible. The lower affinities 11 Varian VNMRS 600 with H{C,N} coldprobe. For 13C measurements
this instrument requires ꢂ25 longer acquisition times than current
may result from weaker CH–p interactions due to the electron-
generation direct-observe 13C{H} cryogenically cooled probes.
deficient nature of the pyridyl ring,4c,6f or reduced hydro-
12 E. Klein, Y. Ferrand, N. P. Barwell and A. P. Davis, Angew. Chem.,
phobic interactions12 due to hydration of the pyridyl nitrogen.
Int. Ed., 2008, 47, 2693–2696.
c
This journal is The Royal Society of Chemistry 2013
3112 Chem. Commun., 2013, 49, 3110--3112