A dynamic combinatorial approach for the analysis of weak carbohydrate/aromatic complexes: Dissecting facial selectivity in CH/π stacking interactions

Article abstract of DOI:10.1021/ja3120218

A dynamical combinatorial approach for the study of weak carbohydrate/aromatic interactions is presented. This methodology has been employed to dissect the subtle structure-stability relationships that govern facial selectivity in these supramolecular complexes.

Full text of DOI:10.1021/ja3120218

Communication
pubs.acs.org/JACS
A Dynamic Combinatorial Approach for the Analysis of Weak
Carbohydrate/Aromatic Complexes: Dissecting Facial Selectivity in
CH/π Stacking Interactions
Andres G. Santana,†,# Ester Jimenez-Moreno,†,# Ana M. Gomez,† Francisco Corzana,‡ Carlos Gonzalez,§
́
́
́
́
Gonzalo Jimenez-Oses,∥ Jesus Jimenez-Barbero,⊥ and Juan Luis Asensio*,†
́
́
†Instituto de Química Organ
́
ica (IQOG-CSIC), Juan de la Cierva 3, 28006 Madrid, Spain
‡Departamento de Química, Universidad de La Rioja, 26006 Logrono, La Rioja, Spain
̃
§Instituto de Quimica-Fisica Rocasolano (IQFR-CSIC), E-28006 Madrid, Spain
∥Department of Chemistry & Biochemistry, University of California, Los Angeles, California 90095, United States
⊥Centro de Investigaciones Biolog
́
icas (CIB-CSIC), 28040 Madrid, Spain
S
* Supporting Information
ABSTRACT: A dynamical combinatorial approach for the
study of weak carbohydrate/aromatic interactions is
presented. This methodology has been employed to
dissect the subtle structure−stability relationships that
govern facial selectivity in these supramolecular complexes.
nderstanding how proteins and nucleic acid receptors
Urecognize oligosaccharides represents a fundamental issue
in chemical biology with far-reaching implications in
fundamental biology, biotechnology, and drug design.1 After
two decades of studies, it is now widely accepted that stacking
interactions involving pyranose rings and aromatic systems are
pivotal to these processes.2 While the free energy of these
contacts is dominated by the dispersive component, electro-
static and solvent-dependent effects also must be considered.
Despite extensive efforts in this area, a detailed comprehension
of the existing relationship between these contributions and the
chemical nature of the pyranose and aromatic rings is still
missing. This absence can at least partially be attributed to the
lack of a simple system for the study of stacking complexes that
displays enough sensitivity to reveal minor stability differences.
Employing concepts from dynamic combinatorial chemistry,3
we have developed a simple and versatile model for the analysis
of weak carbohydrate/aromatic complexes.
Figure 1. (a) Schematic representation of the dynamic combinatorial
approach employed for the analysis of carbohydrate/aromatic
interactions. (b) (top) Complexes involving the D-pyranose β/α
faces for both anomers are shown (up and down). Interacting regions
of the D-pyranose are highlighted. (bottom) The charge density at the
ring oxygen is modulated by hyperconjugative delocalization of its lone
pair (a common explanation for the anomeric effect).
Our general strategy is outlined in Figure 1a. First, several
disaccharide model systems were designed and synthesized. All
4
of them included a common aminoaltrose ring locked in a C1
conformation by a benzylidene group and linked to an
equatorial position of a variety of potential “CH/π donor”
sugar units (represented as colored ellipses). Next, equimolar
amounts of the different models were dissolved in 10 mM
phosphate solution [see the experimental section in the
Supporting Information (SI)] and treated with 2-naphthylace-
taldehyde to form a dynamic mixture of hemiaminals/imines
(undetectable under the employed experimental conditions).
Previous studies have shown that these species adopt a
hairpinlike major conformation stabilized by intramolecular
stacking between the aromatic system and the sugar CH/π
donor.4 Since the resulting mixture is formed under
thermodynamic control, the observed composition directly
depends on the energy contribution of the different
carbohydrate/aromatic contacts (the only distinct factor
among the species). Subsequent chemical reduction of the
Received: December 10, 2012
Published: February 18, 2013
© 2013 American Chemical Society
3347
dx.doi.org/10.1021/ja3120218 | J. Am. Chem. Soc. 2013, 135, 3347−3350

Journal of the American Chemical Society
Communication
transient hemiaminals/imines (normally performed under
slightly acidic conditions for an optimum reaction rate; see
the SI) with an externally added reagent irreversibly converts
them into a set of secondary amines, whose relative populations
must reflect free energy differences among the alternative stacking
modes. For practical reasons, we employed NMR spectroscopy
to follow the chemical reactions. The populations of the
different species, which were roughly invariant during the time
course of the experiment, were evaluated by integration of the
different sets of resolved signals. Finally, the chemical shift
perturbations promoted by the naphthyl ring at the interacting
pyranose in the final species were measured and taken as an
indication of the subtle geometrical differences among the
alternative stacking modes.
In particular, we focused our attention on the subtle factors
that govern (α/β) “facial” selectivity in carbohydrate/aromatic
complexes (Figure 1b). It should be noted that stacking
through a D-pyranose β face implies the participation of the ring
oxygen in n/π bonds with the aromatic system. This particular
contact, would be expected to be modulated by the anomeric
effect5 and consequently should be sensitive to structural factors
such as the anomeric configuration and the presence of
electron-withdrawing substituents at the anomeric center.5d,e
The employed disaccharide library is shown in Figure 2 (also
see Figures S1−S7 in the SI). Its design permitted comparison
expected for the alternative stacking modes. In contrast, smaller
aldehydes (e.g., benzaldehyde) were much less selective, in
agreement with their limited ability to contact the CH/π donor
unit in the disaccharide products.
As a first step in our analysis, we performed the reductive
amination reactions with these disaccharides and characterized
the corresponding products by NMR spectroscopy (Figure 3).
Next, we carried out pairwise competition experiments with all
of the library members (Figures 4, S9, and S10). Figure 4 shows
the ratio of products resulting from each pair of primary
amines. This value was taken as an indicator of the free energy
differences among the alternative carbohydrate/aromatic
stacking modes. Experiments performed with more complex
mixtures (see Figure 4b) or alternative buffer conditions
(Figure S11) rendered similar results. Several conclusions can
be obtained from these data (Figures 4 and 5).
(a) First, in regard to the α complexes (formed by 2−4),
derivative 2 leads to a favored interaction relative to that
established by 3. The observed preference probably reflects the
contribution of the carbohydrate/aromatic contacts mediated
by the hydroxymethyl moiety in 2, which are hardly feasible in
3 and 4. This hypothesis was backed up by inspection of the
geometries of the corresponding complexes and by the Δδ
values observed in the reaction products (Figure 5).
(b) It is known that axial OH groups exposed on the
interacting face of the pyranose are highly disruptive for
carbohydrate/aromatic stacking.1,2 Our results highlight the
unfavorable influence that equatorial OH functions can also
have on these complexes. In particular, inversion of position 4
from the D-gluco (in 3) to the D-galacto configuration (in 4)
leads to preferred CH/π contacts, together with a pronounced
adjustment of the stacking geometry (as revealed by the Δδ
values). The destabilizing role of this particular OH in 3 most
likely results from a combination of repulsive electrostatic
interactions and nonfavorable desolvation effects.4,6
(c) In the case of an identical β-OMe glucose unit,
complexation through the D-pyranose α face (compound 3)
was preferred over the β-face (compound 5). In conclusion, n/π
interactions mediated by the ring oxygen are less stable than the
CH/π analogues. The observed effect amounts to 0.2−0.4 kcal/
mol.
(d) This nonfavorable influence would be expected to be
partially relieved upon inversion of the anomeric position (from
β in 5 to α in 6; see Figures 4 and 5) as a result of
hyperconjugative delocalization of the ring oxygen lone pair (a
common explanation for the anomeric effect). In fact, our
experiments indicate that derivative 6 allows the formation of a
tighter carbohydrate/aromatic complex with a slightly different
geometry (see the Δδ values in Figure 5). However, the
stabilization promoted by inversion of the anomeric position in
5 (to give 6) was identical to that promoted by inversion of the
equivalent position 4 in 3 (to give 4) within the experimental
error (Figure 5). In conclusion, the observed effect is not specific
for the anomeric position and seems simply to reflect the
unfavorable influence of particular equatorial OH moieties of
the D-pyranose. According to our data, the modulation οf the β-
type carbohydrate/aromatic stacking by the anomeric effect is
below the 0.2 kcal/mol limit.
Figure 2. Representation of the synthesized disaccharide library. The
configuration of the interacting pyranoses is indicated. Carbon
positions involved in the glycosidic bonds are labeled in gray. Finally,
carbohydrate/aromatic contacts formed upon reaction with 2-
naphthylacetaldehyde are also shown.
of the stacking of the naphthyl ring against either the α
(compounds 2−4) or β (compounds 5−8) face of the CH/π
donor D-pyranose unit. These unique stacking modes are to a
large extent determined by the strong conformational
preferences of the α(1−4)/α(1−2) and α(1−3) linkages,
respectively (see Figure S8). The library also included
disaccharides with different configurations for the key OH
groups and punctual OH/F substitutions. Monosaccharide 1,
with no interacting additional sugar unit, was also prepared as a
negative reference.
(e) Further enhancements of the hyperconjugative inter-
actions promoted by the incorporation of electron-withdrawing
substituents at the anomeric center (in 7 and 8) were barely
reflected in either the stability or the geometry of the resulting
complexes. The obtained results are consistent with a minor
Control reductive amination assays confirmed that 2-
naphthylacetaldehyde exhibits a strong preference for dis-
accharides 2−8 with respect to the reference compound 1 (3−
10-fold increased reactivity). Moreover, the observed differ-
ences in reactivity were in agreement with the stabilities
3348
dx.doi.org/10.1021/ja3120218 | J. Am. Chem. Soc. 2013, 135, 3347−3350

Journal of the American Chemical Society
Communication
Figure 3. Reductive amination reaction performed with disaccharide 2 and 2-naphthylacetaldehyde in D2O (10 mM phosphate, pH 6.2). The time
evolution of the reaction mixture is represented at the right. Key signals of the glucose unit that were shifted upfield (by up to 2.12 ppm) because of
the proximity of the naphthyl ring are highlighted.
Figure 4. (a) Pairwise competition experiments performed with the library described herein. Equimolecular mixtures of two disaccharides in D2O
(10 mM phosphate, pH 6.2) were treated with substoichiometric amounts of 2-naphthylacetaldehyde. After a short equilibration period, the
reductive amination reactions were carried out by adding sodium cyanoborohydride. The ratio of secondary amine products obtained for each
disaccharide pair (Ix/Iy) is indicated. These values were taken as indicators of the free energy differences between the alternative stacking modes (Ex
− Ey, in kcal/mol). Dotted boxes highlight the formation of the reaction products for pairs 2/3, 2/4, and 2/6 as revealed by NMR spectroscopy. For
each signal, the compound number (in bold), proton assignment (in brackets), and chemical shifts in ppm are indicated. (b) Competition
experiment performed with a 1:1:1:1 mixture of 2, 3, 5, and 6. Representative NMR signals for the disaccharide models and reaction products are
shown. The compound numbers (in bold) and proton assignments (in brackets) are indicated.
3349
dx.doi.org/10.1021/ja3120218 | J. Am. Chem. Soc. 2013, 135, 3347−3350

Journal of the American Chemical Society
Communication
REFERENCES
■
(1) (a) Nishio, M. Phys. Chem. Chem. Phys. 2011, 13, 13873.
(b) Asensio, J. L.; Arda, A.; Canada, F. J.; Jimenez-Barbero, J. Acc.
Chem. Res. 2012, DOI: 10.1021/ar300024d.
(2) (a) Terraneo, G.; Potenza, D.; Canales, A.; Jimen
Baldridge, K. K.; Bernardi, A. J. Am. Chem. Soc. 2007, 129, 2890.
(b) Wimmerova, M.; Kozmon, S.; Necasova, I.; Mishra, S. K.;
Komarek, J.; Koca, J. PLoS One 2012, 7, No. e46032. (c) Ramírez-
Gualito, K.; Alonso-Ríos, R.; Quiroz-García, B.; Rojas-Aguilar, A.; Díaz,
D.; Jimenez-Barbero, J.; Cuevas, G. J. Am. Chem. Soc. 2009, 131,
́
́
̃
́
ez-Barbero, J.;
́
̌
́
́
̌
́
18129. (d) Ke, C.; Destecroix, H.; Crump, M. P.; Davis, A. P. Nat.
Chem. 2012, 4, 718.
(3) (a) Ramstrom, O.; Lehn, J.-M. Nat. Rev. Drug Discovery 2002, 1,
̈
26. (b) Corbett, P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.;
Sanders, J. K. M.; Otto, S. Chem. Rev. 2006, 106, 3652.
Figure 5. Stability differences for stacking complexes. Chemical shift
perturbations promoted by the aromatic ring at the interacting D-
pyranose unit in the reaction products are shown. For derivatives 3−8,
the key positions 1 and 4 in the β and α complexes, respectively, are
highlighted in purple.
́ ́ ́
(4) Vacas, T.; Corzana, F.; Jimenez-Oses, G.; Gonzalez, C.; Gomez,
A. M.; Bastida, A.; Revuelta, J.; Asensio, J. L. J. Am. Chem. Soc. 2010,
132, 12074.
(5) (a) Tvaroska, I.; Bleha, T. Adv. Carbohydr. Chem. Biochem. 1989,
47, 45. (b) Lemieux, R. U.; Koto, S.; Voisin, D. ACS Symp. Ser. 1979,
87, 17. (c) Thatcher, G. R. J. The Anomeric Effect and Associated
Stereoelectronic Effects; American Chemical Society: Washington, DC,
1993. (d) Cocinero, E. J.; Carcabal, P.; Vaden, T. D.; Simons, J. P.;
Davis, B. G. Nature 2011, 469, 76. (e) Wang, C. W.; Ying, F. M.; Wu,
W.; Mo, Y. R. J. Am. Chem. Soc. 2011, 133, 13731.
strengthening of the carbohydrate/aromatic forces in 7 and 8
(as also reflected in the Δδ values). However, according to our
data, this effect amounts to less than 0.1 kcal/mol (Figures 4, 5,
and S10). This conclusion is in agreement with a recent
theoretical study showing that the hyperconjugative electron
delocalization at the anomeric center in pyranose rings is fairly
weak.7
(6) Laughrey, Z. R.; Kiehna, S. E.; Riemen, A. J.; Waters, M. L. J. Am.
Chem. Soc. 2008, 130, 14625.
(7) Mo, Y. Nat. Chem. 2010, 2, 666.
In summary, the dynamic combinatorial approach herein
presented provides a simple and versatile method for the study
of weak carbohydrate/aromatic complexes. It can be observed
that the maximum energy differences revealed by our data
amount to 0.7 kcal/mol, which represents a very significant
fraction of the whole carbohydrate/aromatic interaction free
energy (estimated to be 1−2 kcal/mol).1 On the other hand,
the employed strategy allows the detection of minor stability
differences as small as 0.1 kcal/mol, a limit hardly detectable by
other means. Efforts to extend these studies to alternative
carbohydrate models including a variety of chemical
modifications and configurations and different aromatic systems
are currently underway.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis, NMR competition experiments, and Figures S1−
S11. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
juanluis.asensio@csic.es
Author Contributions
#A.G.S. and E.J.-M. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This investigation was supported by research grants from the
Spanish “Plan Nacional” (MCYT) (CTQ2010-19073,
CTQ2009-08536, and CTQ2009-10343) and the Comunidad
de Madrid (S2009/PPQ-1752).
3350
dx.doi.org/10.1021/ja3120218 | J. Am. Chem. Soc. 2013, 135, 3347−3350