Evaluation of amino acids as chiral ligands for the enantiodifferentiation of carbohydrates by TOCSY NMR

Article abstract of DOI:10.1021/jo1004263

Figure presented The use of natural amino acids as chiral ligands on a triethylbenzene scaffold for the binding and enantiodifferentiation of carbohydrates has resulted in moderate affinity and selectivity values for glucose. Selective 1D TOCSY experiments revealed this as a powerful titration technology surpassing the severe overlapping of receptor and carbohydrate signals in ¹H NMR spectra.

Full text of DOI:10.1021/jo1004263

pubs.acs.org/joc
standing of these recognition mechanisms and to the devel-
opment of synthetic carbohydrate receptors.2
Evaluation of Amino Acids as Chiral Ligands for
the Enantiodifferentiation of Carbohydrates by
TOCSY NMR
Two main families of synthetic receptors can be considered
attending to their recognition mechanisms: (i) boronic acid
derivatives with the ability to bind diols covalently have proven
effective carbohydrate receptors under aqueous conditions,
albeit the null resemblance of their recognition mechanism to
that of lectins (natural carbohydrate recognition proteins),2,3
and (ii) receptors interacting with carbohydrates in a biomimetic
way through noncovalent interactions. In this case, the strong
hydrogen bonds between carbohydrates/synthetic receptors and
water molecules render recognition under aqueous conditions
more elusive and of limited efficiency.2,4 Actually, most of these
receptors work exclusively in organic solvents (e.g., CHCl3 and
MeCN),2,5,6 which limits very much their potential bioapplica-
tions, but conversely offers a good playground for the screening
of candidate scaffolds and ligands.
ꢀ
Francisco Fernandez-Trillo, Eduardo Fernandez-Megia,*
and Ricardo Riguera*
Departamento de Quimica Organica, Facultad de Quimica
and Unidad de RMN de Biomoleculas Asociada al CSIC,
Universidad de Santiago de Compostela, 15782 Santiago de
Compostela, Spain
ef.megia@usc.es; ricardo.riguera@usc.es
Received March 6, 2010
In spite of the stereoselective carbohydrate recognition dis-
played by lectins, it is surprising that the enantiodifferentia-
tion capacity of synthetic carbohydrate receptors has been
barely investigated.7-9 Furthermore, the ability of amino acids,
the constituting building blocks of lectins, to bind carbo-
hydrates10 and to induce a chiral recognition9 has been nar-
rowly exploited.
In this Note we describe our results on the noncovalent
recognition and enantiodifferentiation of carbohydrates in solu-
tion (CHCl3) by means of a new family of amino acid-decorated
(3) (a) James, T. D.; Sandanayake, K. R. A. S.; Iguchi, R.; Shinkai, S. J.
Am. Chem. Soc. 1995, 117, 8982. (b) Zhong, Z.; Anslyn, E. V. J. Am. Chem.
Soc. 2002, 124, 9014. (c) Suri, J. T.; Cordes, D. B.; Cappuccio, F. E.;
Wessling, R. A.; Singaram, B. Angew. Chem., Int. Ed. 2003, 42, 5857. (d)
Dowlut, M.; Hall, D. G. J. Am. Chem. Soc. 2006, 128, 4226. For alternative
examples of boronic acid-based carbohydrate receptors in water, see ref 2 and
the Supporting Information.
(4) Selected examples of synthetic carbohydrate receptors in water: (a)
Sugimoto, N.; Miyoshi, D.; Zou, J. Chem. Commun. 2000, 2295. (b) Klein, E.;
Crump, M. P.; Davis, A. P. Angew. Chem., Int. Ed. 2005, 44, 298. (c) Mazik,
M.; Cavga, H. J. Org. Chem. 2006, 71, 2957. (d) Ferrand, Y.; Crump, M. P.;
Davis, A. P. Science 2007, 318, 619.
The use of natural amino acids as chiral ligands on a
triethylbenzene scaffold for the binding and enantiodif-
ferentiation of carbohydrates has resulted in moderate
affinity and selectivity values for glucose. Selective 1D
TOCSY experiments revealed this as a powerful titration
technology surpassing the severe overlapping of receptor
and carbohydrate signals in 1H NMR spectra.
(5) (a) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. 1998, 37,
2270. (b) Ishi-i, T.; Mateos-Timoneda, M. A.; Timmerman, P.; Crego-
Calama, M.; Reinhoudt, D. N.; Shinkai, S. Angew. Chem., Int. Ed. 2003,
42, 2300. (c) Mazik, M.; Cavga, H.; Jones, P. G. J. Am. Chem. Soc. 2005, 127,
9045. (d) Waki, M.; Abe, H.; Inouye, M. Angew. Chem., Int. Ed. 2007, 46,
Carbohydrate recognition events mediate a myriad of
biological processes such as cell-cell interactions, infection
by pathogens, tumor metastasis, and certain pathways of the
immune response.1 As this recognition takes place in Nature
under aqueous conditions, it represents an important chal-
lenge for supramolecular chemistry. Thus, water molecules
compete with the carbohydrate hydroxyl groups for the
receptor binding site, and only the precise arrangement of
polar and apolar domains (hydrogen bonds and van der
Waals forces) enables the saccharide recognition to occur.
Consequently, a great effort has been devoted to the under-
ꢀ
3059. (e) Arda, A.; Venturi, C.; Nativi, C.; Francesconi, O.; Gabrielli, G.;
~
ꢀ
Canada, F. J.; Jimenez-Barbero, J.; Roelens, S. Chem.;Eur. J. 2010, 16, 414.
For alternative examples of non-covalent carbohydrate receptors in organic
solvents, see ref 2 and the Supporting Information.
(6) Vacca, A.; Nativi, C.; Cacciarini, M.; Pergoli, R.; Roelens, S. J. Am.
Chem. Soc. 2004, 126, 16456.
(7) (a) Bhattarai, K. M.; Bonar-Law, R. P.; Davis, A. P.; Murray, B. A. J.
Chem. Soc., Chem. Commun. 1992, 752. (b) Kim, H.-J.; Kim, Y.-H.; Hong,
J.-I. Tetrahedron Lett. 2001, 42, 5049. (c) Anderson, S.; Neidlein, U.;
Gramlich, V.; Diederich, F. Angew. Chem., Int. Ed. 1995, 34, 1596. (d)
Das, G.; Hamilton, A. D. Tetrahedron Lett. 1997, 38, 3675.
ꢀ
~
(8) Arda, A.; Venturi, C.; Nativi, C.; Francesconi, O.; Canada, F. J.;
ꢀ
Jimenez-Barbero, J.; Roelens, S. Eur. J. Org. Chem. 2010, 64.
(9) (a) Welti, R.; Diederich, F. Helv. Chim. Acta 2003, 86, 494. (b) Welti,
R.; Abel, Y.; Gramlich, V.; Diederich, F. Helv. Chim. Acta 2003, 86, 548. (c)
Li, C.; Wang, G. T.; Yi, H. P.; Jiang, X. K.; Li, Z. T.; Wang, R. X. Org. Lett.
2007, 9, 1797.
(10) (a) Hubbard, R. D.; Horner, S. R.; Miller, B. L. J. Am. Chem. Soc.
2001, 123, 5810. (b) Bitta, J.; Kubik, S. Org. Lett. 2001, 3, 2637. (c) Billing, J.;
Grundberg, H.; Nilsson, U. J. Supramolecular Chem. 2002, 14, 367. (d)
(1) (a) Dwek, R. A. Chem. Rev. 1996, 96, 683. (b) Bertozzi, C. R.;
Kiessling, L. L. Science 2001, 291, 2357. (c) Carbohydrate-based Drug
Discovery; Wong, C.-H., Ed.; Wiley-VCH: Weinheim, Germany, 2003.
(2) (a) James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 1910. (b) Davis, A. P.; Wareham, R. S. Angew. Chem.,
Int. Ed. 1999, 38, 2978. (c) Davis, A. P.; James, T. D. In Functional synthetic
receptors; Schrader, T., Hamilton, A. D., Eds.; Wiley-VCH: Weinheim, Germany,
2005; p 45. (d) Mazik, M. ChemBioChem 2008, 9, 1015. (e) Kubik, S. Angew.
Chem., Int. Ed. 2009, 48, 1722.
~
Segura, M.; Bricoli, B.; Casnati, A.; Munoz, E. M.; Sansone, F.; Ungaro, R.;
Vicent, C. J. Org. Chem. 2003, 68, 6296. (e) Benito, J. M.; Meldal, M. QSAR
Comb. Sci. 2004, 23, 117. (f) Reenberg, T.; Nyberg, N.; Duus, J. Ø.; van
Dongen, J. L. J.; Meldal, M. Eur. J. Org. Chem. 2007, 5003.
3878 J. Org. Chem. 2010, 75, 3878–3881
Published on Web 05/13/2010
DOI: 10.1021/jo1004263
r
2010 American Chemical Society

ꢀ
Fernandez-Trillo et al.
JOCNote
SCHEME 1
the only clearly distinguishable signal of the carbohydrate in
the whole set of spectra.
Since the analysis of a single carbohydrate signal notice-
ably limits the quality of the pursued affinity data, 1D
DPFGSE-TOCSY (Total Correlation Spectroscopy)16 NMR
experiments were attempted with the aim of filtering off the
receptor signals. Indeed, by selective irradiation at the anomeric
proton at 4.3 ppm, the transfer of magnetization through the
whole spin system of 4a (via vicinal-coupled protons, mixing
time 120 ms) resulted in the selective and clean visualization of
the carbohydrate ring protons with no significant increase in
titration time (Figure 1b). In this way, it was possible to analyze
the chemical shift variations of all the carbohydrate resonances
independently on the carbohydrate/receptor ratio, and so a
more precise measurement of the binding affinity was per-
formed. It was revealed that no saturation of the sugar signals
occurred even after the addition of 11 equiv of 3a, in agreement
with a moderate affinity (Figure 1c). Attemps to determine the
binding stoichiometry by using mole ratio and job plots led to
inconsistent results, suggesting a complex binding mode. Actu-
ally, an accurate fitting resulted when a 1:1 and 1:2 mix model
(sugar to receptor ratio) was used for the calculation of the
binding affinity (EQNMR software).17,18 Average values for
receptors based on 1,3,5-triethylbenzene (1) (Scheme 1). Com-
pound 1 is a common scaffold in the construction of artificial
receptors due to its marked preference for an alternate confor-
mation when fully substituted.11
the cumulative association constants β11 (119.7 ( 4.12 M-1
)
and β12 (9.10ꢀ 103 (263 M-2) were obtained for the binding of
3a and 4a, with averaged standard deviations below 10%
(Table 1).
With this aim, the bromomethylated derivative 26 was
treated with various methyl L-R-amino esters (AA-OMe)
derived from tryptophan (Trp), alaline (Ala), leucine (Leu),
valine (Val), and aspartate (Asp), leading to the desired
receptors 3a-e in very good to excellent yields (86-96%).
In a similar fashion, receptors 3f and 3g, incorporating
L-Ala-NHMe and the dipeptide Gly-Ala-OMe, were also
synthesized (Scheme 1 and Supporting Information).12
The capacity of 3 to establish hydrogen bonds (amino ester
side arms with donor-acceptor ability) and hydrophobic
interactions (benzene core) with carbohydrates was assessed
in the binding and enantiodifferentiation of the octyl-β-
glucopyranosides 4a and 4b,13 convenient organosoluble
derivatives of D- and L-glucose (selected as a model mono-
saccharide based on the presence of the D enantiomer as
terminal sugar in glycans that participate in saccharide
recognition processes, Scheme 1). As Trp is present at the
binding site of many lectins, and the indol group resembles
the pyridine/pyrrol present in some of the most successful
carbohydrate receptors based on 1,14 we decided to analyze
first the behavior of the Trp-based receptor 3a. The interac-
At this point the enantiomeric differentiation capacity of
receptor 3a was evaluated by titration with octyl-β-L-glucopyr-
anoside 4b.13 Association constants β11 and β12 in the range of
those shown for the enantiomer 4a evidenced an almost negli-
gible chiral differentiation (Table 1). To elucidate the reasons
behind the moderate affinity and null selectivity values for 3a,
semiempirical (AM1) energy calculations of the binding pocket
geometry were performed with Gaussian 03.19 An interesting
feature emerged: the most stable geometry for 3a places the
indol residues slightly below the binding cavity for glucose,
imparting a rather rigid conformation with limited adaptability,
and hence reduced recognition capacity (Figure 2). In this
situation, the role of the indol group as a binding motif is
limited.
In light of these considerations, we decided to investigate the
steric factors at the amino acid side chain influencing the
recognition of 4a and 4b by this family of receptors. The binding
affinities of a series of Ala-, Leu-, and Val-decorated receptors
3b-d (Scheme 1), characterized by an increasing degree of
substitution, were determined by TOCSY NMR titration (1:1
and 1:2 mix model, Table 1 and Supporting Information).18 In
addition to β, median binding concentrations (BC50), an affinity
descriptor independent of the binding nature of the association,
were also determined.20 As shown in Table 1, there seems to be
an influence of the bulkiness of the amino acid side chain in the
1
tion of 3a with both enantiomers of 4 was studied by H
NMR in CDCl3 (Figure 1). Titrations were performed with a
solution of 4a at a constant concentration (0.5 mM)15 and
increasing concentrations of 3a. The chemical shifts of the
time-averaged signals for the free and complexed 4a were
monitored. Unfortunately, as seen in Figure 1a, severe over-
lapping of carbohydrate and receptor resonances was over-
whelming even at 750 MHz, with the anomeric proton being
(16) Xu, G.; Evans, J. S. J. Magn. Reson., Ser. B 1996, 111, 183.
(17) Hynes, M. J. J. Chem. Soc., Dalton Trans. 1993, 311.
(18) Control NMR experiments confirmed that no dimerization of
receptors 3 occurred in the range of concentrations used for the titrations.
(19) Frisch, M. J. et al. Gaussian03, revision E.01; Gaussian, Inc.:
Wallingford, CT, 2004.
(11) Hennrich, G.; Anslyn, E. V. Chem.;Eur. J. 2002, 8, 2218.
(12) The enantiomeric integrity of receptors 3 was established by chiral
HPLC (see the Supporting Information).
(13) Adasch, V.; Hoffmann, B.; Milius, W.; Platz, G.; Voss, G. Carbo-
hydr. Res. 1998, 314, 177.
(14) (a) Mazik, M.; Bandmann, H.; Sicking, W. Angew. Chem., Int. Ed.
2000, 39, 551. (b) Nativi, C.; Cacciarini, M.; Francesconi, O.; Vacca, A.;
Moneti, G.; Ienco, A.; Roelens, S. J. Am. Chem. Soc. 2007, 129, 4377.
(15) Aggregation of 4 in CDCl3 at concentrations higher than 1 mM has
been reported (see ref 6).
(20) BC50 has been defined by S. Roelens as the total concentration (in
mol/L) of titrating agent (in this case receptors 3) that causes a 50% reduction
in the concentration of unbound analyte (in this case saccharide 4) (for
calculating BC50 values see ref 6). BC50 is usually expressed in mM, and the
lower it is the better the affinity.
J. Org. Chem. Vol. 75, No. 11, 2010 3879

ꢀ
Fernandez-Trillo et al.
JOCNote
FIGURE 1. Titration of 4a (0.52 mM in CDCl3, 25 °C, 750 MHz) with Trp-based receptor 3a: (a) 1H NMR and (b) TOCSY NMR spectra. (c)
Plot of the observed (symbol) and calculated (line) upfield chemical shifts of 4a as a function of the concentration of added 3a.
TABLE 1. Cumulative Association Constants β11 (M-1) and β12 (M-2),
Standard Free Energies of Binding -ΔG° (kJ mol-1), Median Binding
Concentration BC50 (mM), and Relative Affinity (RA) for 1:1 and 2:1
Complexes of Receptors 3 and Glucopyranosides 4 (0.5 mM)
β11
-ΔG°11
β
12/103
-ΔG°12
BC50
RAa
3a 4a
3
119.7
113.7
269.0
465.0
84.3
11.9
11.8
13.9
15.3
11.1
10.1
11.4
12.2
11.7
10.9
11.3
12.1
9.1
9.6
18.4
44.4
5.2
22.6
22.8
24.4
26.6
21.3
20.2
23.8
23.8
23.1
22.1
21.1
23.5
6.14
6.23
3.37
2.12
8.35
11.4
5.96
5.27
6.12
7.83
7.99
5.07
1.01
3a 4b
3
3b 4a
3
1.58
1.37
1.13
1.28
1.58
3b 4bb
3c 4a
3
3
3c 4b
3e 4a
58.5
3.4
3
FIGURE 2. Energy minimized geometry (AM1) of receptor 3a: (a)
axial and (b) equatorial views.
100.2
133.4
110.2
80.2
93.4
132.6
14.3
14.5
10.9
7.2
4.8
12.8
3
3e 4b
3f 4a
3
3
3f 4b
3g 4a
3
Finally, we analyzed receptor 3g incorporating the dipep-
tide Gly-Ala-OMe. The Gly spacer introduces an extra
binding motif (amide) and places the Ala residue further
away from the triethylbenzene scaffold, when compared to
receptor 3b (Scheme 1). Although binding affinity was not
improved,21 probably again as a result of the enhanced
substitution around the benzene ring, 3g was able to induce
a RA similar to 3b.
3
3
3g 4b
aDefined as (BC50)max/BC50. bA receptor with D-Ala was prepared
[(D)-3b] and its binding affinity to 4a evaluated, showing identical binding
constants.
binding strength of these receptors. Thus, receptor 3b, having a
methyl group at the side chain, offers the lowest steric hindrance
and the highest binding affinity. As steric hindrance increases,
affinity is reduced (Me in 3b > i-Bu in 3c > i-Pr in 3d), with 3d
being a limiting example where the affinity constants could not
be determined as a result of its poor binding ability.
In addition, a correlation between binding affinity and chiral
discrimination was observed within this family of receptors, as
3b also showed the highest relative affinity (RA 1.58, based on
BC50) for the L-glucopyranoside 4b relative to the natural
saccharide configuration 4a (Table 1).
Having explored the steric effects over the binding con-
stants and RA, we introduced variations at the number and
nature of binding motifs at the amino acid side chain
(Scheme 1). Comparison of receptor 3c with the Asp
derivative 3e (similar steric hindrance) revealed the ester
group in 3e as positive for the binding, with reductions up
to 50% in BC50 values (Table 1 and Supporting Informa-
tion). Conversely, the increase in number of binding sites
led to reduced RA, probably as a result of the higher
adaptability of the receptor cavity to both carbohydrate
enantiomers. The nature of the terminal carboxylic acid
was also analyzed. Substitution of the methyl ester in the
Ala derivative 3b for a methyl amide (receptor 3f) led to
reduced binding strength and chiral discrimination (Table 1
and Supporting Information).18
In addition to the quantitative affinity data offered by β
and BC50, structural information on the interaction can
1
be obtained from the H chemical shift differences (Δδ =
δmax - δ0) for the carbohydrates during titration thanks to
the TOCSY filtering (Figure 1, as well as Figures S9-S16 in
the Supporting Information). Thus, significant Δδ resulted
when comparing the NMR spectra for the free and bound
glucopyranosides, with the largest shifts (up to 1.8 ppm)
observed for H2 and H4 protons (Figure S16, Supporting
Information). This feature is in agreement with the β face of 4
being in close proximity to the benzene ring of the receptors
in the complex, as recently reported by the groups of Roelens
ꢀ
and Jimenez-Barbero for alternative receptors based on 1
(NMR and modeling calculations).8
In conclusion, natural amino acids, in spite of being the
constituting building blocks of lectins, have demonstrated
moderate affinity and selectivity for the binding and enantio-
differentiation of carbohydrates, when combined with a
triethylbenzene scaffold. We have also proven TOCSY
experiments to be a useful filtering strategy widening the
scope of NMR titrations to receptors with complex structures
(21) Fitting the data for 3g proved to be more difficult than for the other
receptors, suggesting a different binding mode. Nevertheless, the 1:1 and 1:2
mix model afforded again the most accurate correlation.
3880 J. Org. Chem. Vol. 75, No. 11, 2010

ꢀ
Fernandez-Trillo et al.
JOCNote
where severe overlapping hampers the direct analysis of the
carbohydrate resonances.
CDCl3) δ 175.4, 142.6, 136.3, 133.5, 127.1, 123.5, 121.8, 119.1,
118.5, 111.4, 110.6, 62.2, 51.8, 46.4, 29.1, 21.6, 16.4; HR-MS calcd
for C51H61N6O6 (MHþ) 852.4574, found 852.4598.
NMR Titrations and Data Analysis. Titration experiments
were monitored by 1H NMR spectroscopy (750 MHz) in CDCl3
at 298 K. Five millimeter NMR tubes and Gilson micropipets
were used. To avoid the interference of traces of acid in solution,
Experimental Section
General Procedure for the Preparation of Receptors 3a-g. A
solution of 1,3,5-tris(bromomethyl)-2,4,6-triethylbenzene (2)
(1 equiv) in CH3CN was added to a suspension of the corres-
ponding aminoacid hydrochloride (4 equiv) and K2CO3 (6 equiv)
in CH3CN. Reaction was stirred at 80 °C. Then, it was filtered
(Celite) and evaporated. The resulting crude product was dis-
solved in CH2Cl2 and 4-benzyloxybenzaldehyde resin was added.
The mixture was gently shaken for 5 h, after which the resin was
filtered off, and pure 3 was isolated after evaporation. In those
cases where it was needed, flash chromatography was employed
to further purify the receptors.
Receptor 3a. From 51 mg (0.12 mmol) of starting 2, 94mg(95%)
of receptor 3a were obtained after purification by flash chromato-
graphy (silica, CH2Cl2/MeOH 90:10). 1H NMR (300 MHz, CDCl3)
δ 8.07 (br s, 3H), 7.62 (d, J=8.2 Hz, 3H), 7.30 (d, J=7.6 Hz, 3H),
7.18 (dd, J=8.2, 7.6 Hz, 3H), 7.12 (dd, J=8.2, 7.6 Hz, 3H), 6.92 (d,
J=2.1 Hz, 3H), 3.73 (s, 9H), 3.71-3.67 (m, 3H), 3.67 (d, J=11.3
Hz, 3H), 3.38 (d, J=11.3 Hz, 3H), 3.23 (dd, J=14.5, 6.7 Hz, 3H),
3.06 (dd, J=14.5, 6.7 Hz, 3H), 2.54-2.42 (m, 3H), 2.36-2.26 (m,
3H), 1.52 (br s, 3H), 0.89 (t, J=7.4 Hz, 9H); 13C NMR (100.6 MHz,
˚
CDCl3 was filtered over basic alumina, and stored over 4 A
molecular sieves. Mathematical analysis of data was done with
EQNMR software,17 using a mix model of 1:1 and 1:2 sugar:
receptor binding. BC50 values were calculated as reported by
Roelens and co-workers (Supporting Information).6
Acknowledgment. This work was financially supported by
the Spanish MICINN and the Xunta de Galicia. The authors
wish to thank Dr. Manuel Martin-Pastor for his collabora-
tion with the NMR experiments, and Mr. Enrique Lallana
for helpful discussions.
Supporting Information Available: Experimental details for
the synthesis and characterization of receptors 3a-g, NMR
titration and data analysis, and complete refs 3, 5, and 19. This
material is available free of charge via the Internet at http://
pubs.acs.org.
J. Org. Chem. Vol. 75, No. 11, 2010 3881