Selective recognition of alkyl pyranosides in protic and aprotic solvents

Article abstract of DOI:10.1021/ja802229f

The design and synthesis of receptors capable of selective, noncovalent recognition of carbohydrates continues to be a signature challenge in bioorganic chemistry. We report a new generation of tripodal receptors incorporating three pyridine (compound 2)

Full text of DOI:10.1021/ja802229f

Published on Web 06/25/2008
Selective Recognition of Alkyl Pyranosides in Protic and
Aprotic Solvents
Prakash B. Palde,† Peter C. Gareiss,† and Benjamin L. Miller*,†,‡
Department of Biochemistry and Biophysics and Department of Dermatology,
UniVersity of Rochester, Rochester, New York 14642
Received March 26, 2008; E-mail: Benjamin_Miller@urmc.rochester.edu
Abstract: The design and synthesis of receptors capable of selective, noncovalent recognition of
carbohydrates continues to be a signature challenge in bioorganic chemistry. We report a new generation
of tripodal receptors incorporating three pyridine (compound 2) or quinoline (compound 3) rings around a
central cyclohexane core for use in molecular recognition of monosaccharides in apolar and polar protic
solvents. These tripodal receptors were investigated using 1H NMR, UV, and fluorescence titrations in
order to determine their binding abilities toward a set of octyl glycosides. Receptor 2 displayed the highest
binding affinity reported to date for noncovalent 1:1 binding of an R-glucopyranoside in chloroform (Ka )
212 000 ( 27 000 M-1) and an approximately 8-fold selectivity for the R anomer over the ꢀ anomer of the
glucopyranoside. Most importantly, 2 retained its micromolar range of affinities toward monosaccharides
in a polar and highly competitive solvent (methanol). The quinoline variant 3 also displayed micromolar
binding affinities for selected monosaccharides in methanol (as measured by fluorescence) that were
generally smaller than those of 2. Compound 3 was found to follow a selectivity pattern similar to that of
2, displaying higher affinities for glucopyranosides than for other monosaccharides. The binding stoichiometry
was estimated to be 1:1 for the complexes formed by both 2 and 3 with glucopyranosides, as determined
by Job plots. Nuclear Overhauser effect spectroscopy allowed for the derivation of a binding model consistent
with the observed selectivities.
Chart 1. Structures of Cyclohexane-Centered Tripodal
Carbohydrate Receptors Studied in This Work
Introduction
Carbohydrates represent a particularly enticing and important
challenge to our understanding of molecular recognition.1 Novel
carbohydrate receptors play an important role in addressing the
fundamental problem of understanding binding and selectivity
and also have significant potential as tools in the rapidly
evolving field of glycomics,2 as components of sensing systems
for isolated sugars3 and pathogens,4 and as leads for the
development of new therapeutic agents.5 However, as others
have noted, the design of organic receptors able to bind simple,
nonionic sugars via noncovalent interations is a daunting task.6
This is especially true when the goal is binding sugars in
† Department of Biochemistry and Biophysics.
‡ Department of Dermatology.
(1) (a) Davis, A. P.; James, T. D. In Functional Synthetic Receptors;
Schrader, T., Hamilton, A. D., Eds.; Wiley-VCH: Weinheim, Germany,
2005; pp 47-109.
(2) (a) Raman, R.; Raguram, S.; Venkataraman, G.; Paulson, J. C.;
Sasisekharan, R. Nature Methods 2005, 2, 817–824. (b) Ratner, D. M.;
Adams, E. W.; Disney, M. D.; Seeberger, P. H. ChemBioChem 2004,
5, 1375–1383.
protonated solvents, since the solute looks much like the ordered
solvent. As recently disclosed by the Davis group, however,
binding carbohydrates in water is not altogether impossible: an
elegant cage structure they synthesized7is able to bind the
disaccharide cellobiose in water with a binding constant (Ka)
of 650 M-1, as measured by isothermal titration calorimetry
(ITC).8
(3) (a) Zhang, T.; Anslyn, E. V. Org. Lett. 2006, 8, 1649–1652. (b) Gao,
X.; Zhang, Y.; Wang, B. Tetrahedron 2005, 61, 9111–9117. (c) Yoon,
J.; Czarnik, A. W. J. Am. Chem. Soc. 1992, 114, 5874–5875. (d)
Shinkai, S.; Takeuchi, M. Biosens. Bioelectron. 2004, 20, 1250–1259.
(4) Chan, S.; Horner, S. R.; Miller, B. L.; Fauchet, P. M. J. Am. Chem.
Soc. 2001, 123, 11797–11798.
(5) (a) Miller, K. A.; Kumar, E. V. K.; Wood, S. J.; Cromer, J. R.; Datta,
A.; David, S. A. J. Med. Chem. 2005, 48, 2589–2599. (b) Ding, B.;
Yin, N.; Liu, Y.; Cardenas-Garcia, J.; Evanson, R.; Orsak, T.; Fan,
M.; Turin, G.; Savage, P. B. J. Am. Chem. Soc. 2004, 126, 13642–
13648.
(6) Wiskur, S. L.; Lavigne, J. J.; Metzger, A.; Tobey, S. L.; Lynch, V.;
Anslyn, E. V. Chem.sEur. J. 2004, 10, 3792–3804.
(7) Velasco, T.; Lecollinet, G.; Ryan, T.; Davis, A. P. Org. Biomol. Chem.
2004, 2, 645–647.
(8) Ferrand, Y.; Crump, M. P.; Davis, A. P. Science 2007, 381, 619–622.
9
9566 J. AM. CHEM. SOC. 2008, 130, 9566–9573
10.1021/ja802229f CCC: $40.75
2008 American Chemical Society

Selective Recognition of Alkyl Pyranosides
A R T I C L E S
Figure 1. (a) (left) Side and (right) top views of the calculated minimum-energy conformation of 2. (b) Dihedral-angle drive calculation (using the MMFFs
force field22 and the GB/SA continuum water solvation model23) showing the 2.5 kcal/mol preference for the conformer shown in (a) as opposed to the one
in which the pyridine ring is flipped 180° across the C-C single bond.
Scheme 1. Synthesis of Receptor 2
Arenes have proven to be successful core elements in the
design of carbohydrate receptors for two primary reasons. First,
they provide a surface variously described as complementary
to the “hydrophobic patch” of target sugars8 or able to provide
stabilizing interactions via C-H/π bonding.16 Second, hexas-
ubstituted benzenes provide rigid control over the positioning
of pendant functionality via a conformational “gearing” effect.
In principle, highly substituted cycloalkanes should provide
some of the same benefits as arenes: a hydrophobic surface and
conformational control17 of pendant functionality via avoidance
of “anti” interactions (though this control is less stringent than
that provided by hexasubstituted benzenes). Indeed, we have
successfully employed cycloalkane oligomers as conformational
control elements in the development of receptors for lipid A.18
This led us to begin considering the possibility that other
cycloalkane-based scaffolds might have utility as receptors for
monosaccharides. In particular, we were intrigued by the
possibility of employing cis-1,3,5-trisubstituted cyclohexane as
a core design element. While highly substituted cyclohexanes,
such as Kemp’s triacid,19c have found considerable utility as
components in molecular recognition systems,19 receptors based
on a simple cis-1,3,5-trisubstituted cyclohexane core are less
well-known.20 We report here that two such receptors, 2 and 3
(Chart 1), display significant affinities and selectivities for simple
alkyl pyranosides in both halogenated and protic solvents.
Tripodal receptors have been a mainstay of the molecular
recognition field9 and have found favor as scaffolds for the
creation of boronic acid-based receptors for covalent carbohy-
drate recognition.10 In the realm of noncovalent carbohydrate
recognition, tripodal receptors based on arenes have been
explored extensively by the Mazik group, who demonstrated
that aromatic compounds incorporating aminopyridine side
chains11 are able to bind alkyl pyranosides as 1:1 and 2:1
complexes in chloroform with substantial affinities and selectivi-
ties.12 Other notable examples of arene-based tripodal carbo-
hydrate receptors have been described by the Roelens,13 Abe,14
and Schmuck15 groups, who studied hexasubstituted benzene-
based tripodal receptors incorporating urea, phenol, and guani-
dinium substituents, respectively.
Results and Discussion
Design and Synthesis. To begin to assess the utility of
cyclohexane-centered tripodal carbohydrate receptors, we syn-
thesized receptor 1 (Chart 1). We observed that 1 showed only
weak binding (Ka ≈ 90 M-1) to n-octyl-ꢀ-D-glucopyranoside
1
in chloroform, as measured by H NMR titration. Conforma-
(9) (a) Lavigne, J. J.; Anslyn, E. V. Angew. Chem., Int. Ed. 1999, 38,
3666–3669. (b) Chin, J.; Walsdorff, C.; Stranix, B.; Oh, J.; Chung,
H. J.; Park, S.-M.; Kim, K. Angew. Chem., Int. Ed. 1999, 38, 2756–
2759. (c) Oshovsky, V. G.; Reinhoudt, D. N.; Verboom, W. Angew.
Chem., Int. Ed. 2007, 46, 2366–2393.
(16) Sujatha, M. S.; Sasidhar, Y. U.; Balaji, P. V. Biochemistry 2005, 44,
8554–8562.
(17) Hoffmann, R. W. Angew. Chem., Int. Ed. 2000, 39, 2054–2070.
(18) (a) Hubbard, R. D.; Horner, S. R.; Miller, B. L. J. Am. Chem. Soc.
2001, 123, 5810–5811. (b) Gareiss, P. C.; Palde, P. B.; Hubbard, R. D.;
Miller, B. L. Eur. J. Org. Chem. 2007, 53–61.
(10) Zhong, Z.; Anslyn, E. V. J. Am. Chem. Soc. 2002, 124, 9014–9015.
(11) Mazik, M.; Bandmann, H.; Sicking, W. Angew. Chem., Int. Ed. 2000,
39, 551–554.
(19) (a) Kocis, P.; Issakova, O.; Sepetov, N. F.; Lebl, M. Tetrahedron Lett.
1995, 36, 6623–6626. (b) Rebek, J., Jr.; Kolloran, M. K.; Nemeth,
K. D.; Lin, F.-T. J. Am. Chem. Soc. 1987, 109, 2426–2431. (c) Kemp,
D. S.; Petrakis, K. S. J. Org. Chem. 1981, 46, 5140–5143.
(20) (a) Tajc, S. G.; Miller, B. L. J. Am. Chem. Soc. 2006, 128, 2532–
2533. (b) Ruy, J. C.; Shin, H. N.; Kim, D. H.; Lee, S. H. Bull. Korean
Chem. Soc. 2001, 22, 1293–1294.
(12) Mazik, M.; Kuschel, M.; Sicking, W. Org. Lett. 2006, 8, 855–858.
(13) Vacca, A.; Nativi, C.; Cacciarini, M.; Pergoli, R.; Roelens, S. J. Am.
Chem. Soc. 2004, 126, 16456–16465.
(14) Abe, H.; Aoyagi, Y.; Inouye, M. Org. Lett. 2005, 7, 59–61.
(15) Schmuck, C.; Schwegmann, M. Org. Lett. 2005, 7, 3517–3520.
9
J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008 9567

A R T I C L E S
Palde et al.
Figure 2. (a) Summary of the NOE contacts observed in the 1H-1H NOESY spectrum of 2. (b) Expansion of the 1H-1H NOESY spectrum of 2 (500 MHz,
200 ms mixing time, CDCl3, 298 K).
Chart 2. Octyl Pyranosides Used in Binding Studies
tional analysis of 121 led us to hypothesize that a receptor with
fewer torsional degrees of freedom (such as 2) might prove to
be better suited to monosaccharide binding. Indeed, both Monte
Carlo conformational searches and dihedral-angle drives of the
cyclohexane-pyridine Ha-C-C-N torsional angle indicated
that 2 would have a 2.5 kcal/mol preference for a conformation
in which all of the pyridine nitrogen atoms are oriented on the
same face of the molecule (Figure 1) and available for hydrogen
bonding to hydroxyl groups of a target sugar.
reported by us.25 A minor, regioisomeric product (4) was also
obtained. This compound served as a useful control for the
binding abilities of 2 (see below). The structure of 2 was
confirmed by X-ray crystallography and further analyzed in
solution by one- and two-dimensional (1D and 2D) nuclear
Overhauser effect (NOE) spectroscopy. The single-crystal X-ray
structures of 2 in toluene and methanol (see the Supporting
Information) showed that the crystal packing forces and the
recrystallization solvent strongly influence the observed con-
formation in the solid state, causing two of the three pyridine
rings to adopt high-energy torsional angles (as indicated by
molecular mechanics). However, 2D NOE studies on 2 indicated
that the molecule in solution primarily adopts a conformation
similar to that shown in Figure 1. In particular, we observed
moderate NOE cross-peaks between the pyridyl H6 and cyclo-
hexyl Hc and Hc′ protons as well as strong NOEs between the
pyridyl H3 and cyclohexyl Ha, Hc, and Hc′ protons (Figure 2).
Synthesis of 2 was readily achieved via a single-step,
tridirectional Minisci reaction24 between cis-1,3,5-cyclohexan-
etricarboxylic acid and 4-cyanopyridine (Scheme 1), as recently
(21) Maestro/Macromodel, version 7.2; Schro¨dinger, LLC: Portland, OR,
2001.
(22) Halgren, T. A. J. Comput. Chem. 1996, 17, 490–519.
(23) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J. Am.
Chem. Soc. 1990, 112, 6127–6128.
(24) (a) Minisci, F.; Vismara, E.; Fontana, F.; Morini, G.; Serravalle, M.;
Giordano, C. J. Org. Chem. 1987, 52, 730–736. (b) Minisci, F.;
Bernardi, R.; Bertini, F.; Galli, R.; Perchinummo, M. Tetrahedron
1971, 27, 3575–3579.
(25) Palde, P. B.; McNaughton, B. R.; Ross, N. T.; Gareiss, P. C.; Mace,
C. R.; Spitale, R. C.; Miller, B. L. Synthesis 2007, 2287–2290.
9
9568 J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008

Selective Recognition of Alkyl Pyranosides
A R T I C L E S
concentrations greater than 1.2 mM, indicative of aggregation.
No detectable chemical-shift changes were observed upon
further dilution of ꢀ-Glc to concentrations less than 1.2 mM,
consistent with the observations of Vacca et al.13 Therefore,
we followed a reverse-titration format in which a fixed
concentration of ꢀ-Glc (∼1 mM) was titrated with an increasing
concentration of 2. It was also verified that the NMR spectrum
of 2 did not show any concentration-dependent changes.
Receptor-dependent changes in the hydroxyl proton chemical
shifts of ꢀ-Glc upon titration with 2 were observed, indicating
the binding event (Figure 3). Although one can use Scatchard
analysis to derive a binding constant for the interaction of 2
with ꢀ-Glc based on the data shown in Figure 3 (and indeed,
such an analysis yielded a Ka value of 1.5 × 105 M-1),
quantitative analysis of this type of NMR experiment is highly
error-prone when the starting concentration of the sugar is more
than 10-fold greater than the dissociation constant.28 Therefore,
as discussed below, we used other analytical techniques to derive
quantitative information. Importantly, no chemical-shift changes
were observed in an analogous titration of 4-cyanopyridine into
ꢀ-Glc. This initial result confirmed that binding requires the
full structure of the receptor and is not due to nonspecific
interactions between the alkyl glucosides and pyridine. Job plot
analysis (as modified for NMR29) verified a 1:1 association
between 2 and ꢀ-Glc.
UV Titrations. In order to verify the binding observed by
NMR and derive quantitative binding information, UV titrations
of the monosaccharides in Chart 2 with 2 were performed. In
each case, the receptor at fixed concentration was first titrated
with the octyl glycoside30 in dry, deacidified chloroform. An
increase in the intensity of absorbance of 2 was observed as
the monosaccharide concentration was increased (Figure 4a).
The data were fit to a 1:1 binding model on the basis of the Job
plot (NMR) results. Receptor 2 showed the highest binding
affinity for R-Glc (Ka ) 212 000 ( 27 000 M-1) among the
set of monosaccharides tested (Figure 4b and Table 1). This is
the highest binding affinity reported to date for any artificial
receptor toward R-Glc in chloroform. There was also an 8-fold
decrease in the affinity observed for ꢀ-Glc, indicating anomeric
selectivity for the R form of the pyranoside. Additionally,
substantial selectivity for glucopyranosides over other sugars
was observed: while ꢀ-Gal was bound with roughly one-third
the affinity of ꢀ-Glc, ꢀ-Man did not reach saturation up to the
limit of the titration (Table 1). No binding was observed between
2 and octanol, indicating that nonspecific interactions between
the receptor and the octyl chain of the glycopyranosides were
at best a minor contributor to binding. Likewise, the regioisomer
4 showed no affinity for R-Glc. Combined with the NMR
titration data described above, these data demonstrate that the
ability of 2 to bind alkyl pyranosides in chloroform is highly
selective and that proper orientation of the three pyridine rings
is an absolute requirement for affinity.
Figure 3. 1H NMR titration (CDCl3, 298 K) of receptor 2 into ꢀ-Glc (1.09
mM). The portion of the 1H NMR spectrum of ꢀ-Glc that includes (left to
right) the 3-, 4-, and 2-OH proton resonances is shown after addition of
(bottom to top) 0, 0.17, 0.35, 0.52, 0.86, 1.20, 1.59, 2.40, 3.20, 3.60, and
4.80 equivalents of receptor 2.
Sensing of saccharides with fluorescence-active receptors is
of particular interest because of the inherent sensitivity and
simplicity of the fluorescence technique.3b,26 We therefore also
synthesized the quinoline variant 3, anticipating that the intrinsic
fluorescence of the quinoline heterocycle would yield a receptor
suitable for binding analysis by fluorescence. Additionally,
extending the aromatic surface from pyridine (2) to quinoline
(3) allowed us to examine the effect of added π surface on the
conformations, binding affinities, and selectivities of cyclohex-
ane-centered tripodal receptors.
Binding Studies. The binding affinities of 2 toward a set of
1
octyl glycosides (Chart 2) at 298 K were measured using H
NMR titrations in CDCl3 and UV titrations in chloroform or
dry methanol. After we confirmed that 3 exhibited intrinsic
fluorescence properties, binding studies of 3 with the selected
monosaccharides were primarily carried out by fluorescence
spectroscopy in dry methanol.
1H NMR Titrations. NMR has been extensively used in
studying carbohydrate recognition by synthetic receptors.27 This
method has been very helpful in understanding the recognition
process because of its convenience for detecting noncovalent
interactions, especially in nonpolar solvents such as CDCl3.
However, as reported by Vacca et al.,13 interpreting the results
of binding studies using this method can be complicated by the
tendency of octyl pyranosides to aggregate and form micelles.
In particular, aggregation of alkylated sugars seriously affects
accurate estimation of binding constants. Therefore, we used
NMR titration experiments primarily as qualitative indicators
of binding. In our binding experiments with ꢀ-Glc in CDCl3
using NMR titration, dilution-mediated changes in the chemical
shifts of sugar hydroxyl protons were observed at ꢀ-Glc
Since 2 showed excellent binding affinities and selectivities
in chloroform, we became interested in investigating its capabil-
ity to function in polar protic solvents, where the competition
from solvent molecules is high. Hence, we next examined
(26) (a) Cao, H.; Heagy, M. D. J. Fluoresc. 2004, 14, 569–584. (b) James,
T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Angew. Chem., Int. Ed.
Engl. 1996, 35, 1910–1922.
(27) (a) Kikuchi, Y.; Tanaka, Y.; Sutarto, S.; Kobayashi, K.; Toi, H.;
Aoyama, Y. J. Am. Chem. Soc. 1992, 114, 10302–10306. (b)
Lecollinet, G.; Dominey, A. P.; Velasco, T.; Davis, A. P. Angew.
Chem., Int. Ed. 2002, 41, 4093–4096. (c) Das, G.; Hamilton, A. D.
Tetrahedron Lett. 1997, 38, 3675–3678. (d) Bitta, J.; Kubik, S. Org.
Lett. 2001, 3, 2637–2640. (e) Inouye, M.; Takahashi, K.; Nakazumi,
H. J. Am. Chem. Soc. 1999, 121, 341–345. (f) Mazik, M.; Radunz,
W.; Sicking, W. Org. Lett. 2002, 4, 4579–4582.
(28) Hollenberg, M. D. In Receptor Binding in Drug Research; O’Brien,
R. A., Ed.; CRC Press: Boca Raton, FL, 1986; pp 354-363.
(29) Fokkens, M.; Jasper, C.; Schrader, T.; Koziol, F.; Ochsenfeld, C.;
Polkowska, J.; Lobert, M.; Kahlert, B.; Kla¨rner, F.-G. Chem.sEur. J.
2005, 11, 477–494.
(30) In order to avoid aggregation-induced error in the binding constants,
care was taken to ensure that the concentration of sugar in the cuvette
was never higher than 1.2 mM.
9
J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008 9569

A R T I C L E S
Palde et al.
Figure 4. UV titration of R-Glc into receptor 2. (a) Absorbance spectra of 2 recorded during the course of the titration, showing the sugar-concentration-
dependent increase in the absorbance intensity. (b) Binding isotherms for receptor 2 (25 µM; black squares) and its regioisomer 4 (25 µM; green triangles).
The titration of n-octanol into 2 (red circles) is shown as a control.
Table 1. Results from UV Titrations of Selected Octyl Pyranosides
similar to that displayed by 2: glucosides were favored over
other pyranosides and R-Glc was favored over ꢀ-Glc. The fact
with Receptor 2 in Chloroform or Dry Methanol at 298 Ka
pyranoside
Ka in chloroform (M-1
)
Ka in methanol (M-1
)
that the change in fluorescence was of sufficient intensity to be
visible to the naked eye (Figure 6) suggests that 3 might have
utility as a “turn-on” carbohydrate sensor. A Job plot of the
interaction of 3 with R-Glc confirmed a 1:1 binding stoichi-
ometry (Figure 7). Together, these data suggest a common
binding mode for 2 and 3 with octyl pyranosides.
R-Glc
212000 ( 27000
29000 ( 2500
17700 ( 200
10900 ( 300
6500 ( 700
52000 ( 1600
9700 ( 900
9600 ( 950
5200 ( 400
>1300
ꢀ-Glc
R-Gal
ꢀ-Gal
R-Man
ꢀ-Man
n-octanol
(not saturable)
(no binding)
(not saturable)
(not determined)
Mode of Interaction. A preliminary model for binding was
derived on the basis of 1D difference NOE experiments on 2
in the presence of R- and ꢀ-Glc in CDCl3 or CD3OD. Initial
experiments conducted at a sugar concentration below 1.2 mM
did not provide usable data. Therefore, it was necessary to carry
out the NOE analyses at a higher sugar concentration, where
aggregation was a potential complicating factor. Nevertheless,
these experiments served as a useful starting point for under-
standing the structural basis of sugar recognition by 2 and 3.
Selective irradiation of the frequency corresponding to the C-6
aromatic proton of 2 in the presence of ꢀ-Glc showed enhance-
ment of the 2-, 3-, and 4-hydroxyl protons of the sugar (Figure
8). This suggests the possibility that the sugar hydroxyl protons
are hydrogen-bonded to the pyridine nitrogens of 2 and hence
are located within the NOE distance range of the C-6 aromatic
protons. The key NOE contacts (summarized in Figure 9a)
indicate that R- and ꢀ-Glc bind with similar orientations. A
binding model derived from applying the NOE constraints to
the ꢀ-methylglucoside and receptor 2 is shown in Figure 9b.
This model indicates that crucial hydrogen-bonding interactions
can be lost due to the axial orientation of the 4- and 2-hydroxyl
groups in Gal and Man, respectively, thus reducing the affinity
of 2 toward these sugars; this is consistent with the observed
binding selectivity of 2 for Glc over Gal and Man. The model
also suggests that weak hydrophobic interactions with the octyl
chain may contribute to the observed anomeric selectivity toward
the R forms of the pyranosides.
a n-Octanol was used as a control. Each titration was repeated at least
twice. Each reported Ka value is the average of results from two
independent titrations.
binding in dry methanol. UV titrations again demonstrated that
2 binds alkyl pyranosides with high affinities and selectivities
(Table 1). While the affinities were smaller than those measured
in chloroform (as would be expected because of increased
competition from solvent), their rank order was maintained. This
suggests that the mode of binding is not altered to any great
extent by changing the solvent system from nonpolar to polar.
Again, the receptor showed the highest binding affinity for R-Glc
(Ka ) 51 000 ( 1600 M-1) among the tested monosaccharides.
The anomeric selectivity for the R form over the ꢀ form for
each glucopyranoside was found to be similar (nearly within
experimental error) in chloroform and methanol.
Fluorescence Titrations. The exceptionally strong binding
constants measured for 2 led us to seek other analytical methods
to confirm our observations. Unfortunately, 2 is not fluorescent,
precluding the use of direct fluorescence methods. Therefore,
we conceived receptor 3 as a fluorescent analogue of 2. Receptor
3 was found to exhibit a fluorescence maximum at 422 nm when
excited with a wavelength of 342 nm (Figure 5a). Initial
fluorescence titrations of 3 in CHCl3 showed a high affinity for
R-Glc in CHCl3 (Ka ) 42 000 M-1) and established a cor-
respondence with the UV-vis titrations of 2 described above.
Next, we measured octyl pyranoside affinities of 3 in dry
methanol, again by fluorescence. As for 2, we found that
switching to methanol caused only modest decreases in the
affinities. While 3 showed an overall decrease in affinity for
octyl glucosides relative to 2 (Table 2), the measured binding
constants were nevertheless quite large (Ka ) 15 300 ( 150
M-1 for R-Glc). Compound 3 showed a selectivity pattern
Conclusion
In the present work, we have prepared and analyzed three
novel tripodal receptors for monosaccharides using a cis-1,3,5-
trisubstituted cyclohexane as a core structural element. While
compound 1 showed only weak binding ability, our findings
with respect to compounds 2 and 3 confirmed that cyclohexane
9
9570 J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008

Selective Recognition of Alkyl Pyranosides
A R T I C L E S
Figure 5. (a) Fluorescence spectrum of 3 in dry methanol. (b) Binding isotherm for the interaction between 3 (12 µM) and R-Glc.
Table 2. Results of Fluorescence Titrations of the
Glycopyranosides in Chart 2 with Receptor 3 in Dry Methanol at
298 Ka
pyranoside
Ka in methanol (M-1
)
R-Glc
ꢀ-Glc
R-Gal
ꢀ-Gal
R-Man
ꢀ-Man
octanol
15300 ( 150
9200 ( 600
6600 ( 200
3200 ( 80
2600 ( 200
>400
(no binding)
a n-Octanol was used as a control. Each reported Ka value is the
average of results from two independent titrations.
Figure 7. Job plot for the complex formed between 3 and R-Glc, showing
1:1 binding stoichiometry in dry methanol by fluorescence.
for simple derivatization via several methods, such as hydrolysis
of the nitrile groups of 2 and oxidation of the methyl groups of
3. These experiments are currently underway. Thus, we
anticipate that incorporation of analogues of 2 and 3 into more-
complex structures that target complex carbohydrates and
carbohydrate derivatives will enable the synthesis of novel
structures targeting a broad range of carbohydrates and provide
fundamental insight into the recognition of simple sugars by
designed receptors.
Figure 6. Visible fluorescence detection of octyl R-glucopyranoside: 400
µM 3 in CH3OH in the (left) absence and (right) presence of 1 mM R-Glc,
illuminated with a hand-held long-wave (365 nm) UV lamp.
Experimental Section
General. Reagents were purchased from commercial suppliers
can serve as an effective scaffold for the production of receptors
capable of efficient carbohydrate recognition. Importantly,
compound 2 showed the highest affinity recorded to date for
noncovalent recognition of monosaccharides in a protic solvent
(methanol) and also demonstrated significant selectivities in both
solvents tested. Compound 3, while displaying somewhat lower
carbohydrate affinities than 2, is intriguing because of its ability
to function as a turn-on fluorescence sensor for monosaccharides
in protic solvents. Compounds 2 and 3 should both be suitable
1
and used without purification unless otherwise stated. H and 13C
NMR spectra were recorded on a Bruker Avance 400 instrument.
Chemical shifts (δ) are reported in parts per million relative to
tetramethylsilane. Melting points are uncorrected. High-resolution
mass spectrometry (HRMS) was performed by the mass spectrom-
etry facility at the University of California, Riverside, using
electrospray ionization/atmospheric pressure chemical ionization.
Preparation of Octyl Monosaccharides. Octyl R-Glc, ꢀ-Glc,
ꢀ-Gal, and ꢀ-Man were obtained commercially. Octyl R-Gal and
9
J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008 9571

A R T I C L E S
Palde et al.
Figure 8. (bottom) 1H NMR spectrum of 2 (1 mM) complexed with ꢀ-Glc (3 mM) in CDCl3 at 298 K. The irradiation frequency, corresponding to C-6 ArH
on the pyridines of 2 (δ 8.785), is marked, as are (left to right) the 3-, 2-, and 4-OH groups of the sugar. (top) The 1D difference NOE spectrum shows a
strong positive intramolecular NOE from the irradiated C-6 ArH to C-5 ArH and weak positive intermolecular NOEs (labeled with *) from the irradiated
proton to the 3-, 2-, and 4- OH groups of the sugar.
Figure 9. (a) Summary of NOE interactions observed between 2 and (left) ꢀ- and (right) R-Glc in CDCl3. (b) Binding model derived by applying the NOE
constraints from 1D difference NOE experiments to models of ꢀ-alkylglucoside and 2.
R-Man are known compounds31 and were prepared according to a
literature method.32 Spectral assignments were in agreement with
the literature data.31
cis-Cyclohexane-1,3,5-tricarboxylic Acid Tris[(6-methylpyridin-
2-yl)amide] (1). The cis-1,3,5-cyclohexanetricarboxylic acid chloride
was prepared by previously described methodology.33 Briefly,
anhydrous N,N-dimethylformamide (0.12 g, 1.65 mmol) was added
dropwise to a slurry of cis-1,3,5-cyclohexanetricarboxylic acid (0.60
g, 2.72 mmol) and thionyl chloride (1.60 g, 13.60 mmol) in
anhydrous CH2Cl2. The slurry was allowed to reflux for 4 h and
dried in vacuo to yield a yellow oil. An aliquot of the triacid chloride
(110 mg, 0.40 mmol) was diluted to 0.1 mM in CH2Cl2. Next,
2-amino-6-methylpyridine (0.44 g, 4.05 mmol), 4-(dimethylami-
no)pyridine (0.24 g, 0.20 mmol) and triethylamine (0.41 g, 4.05
mmole) were dissolved in 10 mL of anhydrous CH2Cl2 and added
dropwise to the acid chloride. After 24 h, the resulting yellow
solution with precipitate was treated with 0.5 mL of H2O and dried
in vacuo to yield a whitish solid. The compound was purified by
silica gel flash chromatography (49:1 CH2Cl2/CH3OH) followed
by preparative reversed-phase high-performance liquid chromatog-
raphy (RP-HPLC) (85:15 H2O/CH3CN + 0.1% trifluoroacetic acid)
to obtain a yellow oil in 51% overall yield. FTIR (thin film from
CDCl3) ν (cm-1): 722, 798, 1153, 1196, 1431, 1636, 2867. 1H
NMR (400 MHz, CDCl3) for the trifluoroacetic acid salt from
preparative RP-HPLC: δ 8.39 (d, J ) 8 Hz, 3H), 8.06 (t, J ) 8
Hz, 3H), 7.13 (d, J ) 8 Hz, 3H), 2.97 (t, J ) 14 Hz, 3H), 2.71 (s,
9H), 2.39 (d, J ) 12 Hz, 3H) 1.80 (q, J ) 12 Hz, 3H). 13C NMR
(100.7 MHz, CDCl3): δ 19.4, 30.1, 43.2, 114.4, 119.6, 145.8, 148.9,
(31) Mizutani, T.; Kurahashi, T.; Murakami, T.; Matsumi, N.; Ogoshi, H.
J. Am. Chem. Soc. 1997, 119, 8991–9001.
(32) Adascha, V.; Hoffmanna, B.; Miliusc, W.; Platza, G.; Voss, G.
Carbohydr. Res. 1998, 314, 177–187.
(33) Pryor, K. E.; Shipps, G. W.; Skylera, D. A.; Rebek, J. Tetrahedron
1998, 54, 107–4124.
9
9572 J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008

Selective Recognition of Alkyl Pyranosides
A R T I C L E S
150.7, 174.9. HRMS (m/z): calculated for C29H34O4 ([M + H]+),
487.2458; found, 487.2457.
UV Titrations. UV titrations were performed on a Shimadzu
UV-160IPC UV-vis spectrophotometer using 1 cm path-length
cuvettes. The chloroform solvent was passed through a short basic
alumina column to remove traces of acid and stored over 4 Å
molecular sieves. Methanol used in titrations was freshly distilled
over magnesium turnings and iodine (a procedure for making
superdry methanol). Titrations were performed by placing the
receptor (17-40 µM, 0.6 mL) into the cuvette and adding increasing
amounts of the octyl pyranoside. Hamilton microsyringes were used
for the addition of monosaccharides. After each addition, an
equilibration time of 10-15 min was allowed before the absorbance
was recorded. Statistical analysis of the data was carried out using
Origin 7 (OriginLab Corporation). Binding constants were deter-
mined by fitting the data to a one-site binding model equation.
Fluorescence Titrations. Fluorescence titrations were performed
on an Aminco-Bowman Series 2 luminescence spectrometer fluo-
rometer using a 1 cm path-length quartz fluorescence cell. The
titrations with R- and ꢀ-Glc were also repeated on a Horiba Jobin
Yvon Fluoromax spectrofluorometer. Chloroform was passed
through a short basic alumina column to remove traces of acid and
stored over molecular sieves. Superdry methanol used in the
titrations was freshly distilled over magnesium turnings and iodine.
Titrations were performed by placing the receptor (12-30 µM, 1
mL) into the 1 mL cuvette and adding increasing amounts of the
octyl pyranoside using a Hamilton microsyringe. The receptor
solution was excited at 342 nm, and the emission spectrum was
recorded from 365 to 500 nm. After each addition, an equilibration
time of 10-15 min was allowed before the fluorescence was
recorded. A PMT voltage of 955 V was used for collecting the
fluorescence data. Statistical analysis of the data was carried out
using Origin 7 (OriginLab Corporation). Binding constants were
determined by fitting the data to a one-site binding model equation.
6-[(3R,5S)-3,5-Bis(4-cyanopyridin-2-yl)cyclohexyl]nicotinoni-
trile (2). To a solution of 1,3,5-cyclohexanetricarboxylic acid (1.50
g, 6.93 mmol) in 10% H2SO4 (40 mL) were added silver nitrate
(0.43 g, 2.49 mmol) and 4-cyanopyridine (8.64 g, 83.16 mmol).
The reaction mixture was heated to 80 °C. A saturated solution of
ammonium persulfate (19.00 g, 83.16 mmol) in water (10 mL) was
added to the reaction mixture dropwise over 10 min with evolution
of carbon dioxide (indicated by bubbling in the solution). After
emission of carbon dioxide ceased, the reaction mixture was allowed
to stir for an additional 20 min at 80 °C. The mixture was then
poured into ice and neutralized using a sufficient quantity of
saturated ammonium hydroxide solution. The resulting suspension
was then extracted with chloroform (3 × 30 mL), and the organic
extract was washed with brine solution (3 × 20 mL). Finally, the
organic layer was dried over Na2SO4 and concentrated to yield a
yellow oil. The crude product was then purified using flash
chromatography (0-3% CH3OH in CHCl3) to obtain compound 2
as the major product in 24% yield (0.65 g) and its regioisomer 4
as a minor product in 4% yield (95 mg). Mp of 2: 163-165 °C. IR
(thin film from CHCl3) ν (cm-1): 2925, 2237, 1594, 1550, 1472,
1397, 1260, 842, 754. 1H NMR (400 MHz, CDCl3): δ 8.73 (d, J )
4.8 Hz, 3H), 7.48 (s, 3H), 7.38 (dd, J ) 5 Hz, 1.4 Hz, 3H),
3.27-3.21 (m, 3H), 2.33 (d, J ) 12 Hz, 3H), 2.02 (q, J ) 12.4
Hz, 3H). 13C NMR (100.7 MHz, CDCl3): δ 165.8, 150.3, 123.2,
121.1, 116.7, 45.7, 37.4. HRMS (m/z): calculated for C24H18N6
(M+), 390.1600; found, 390.1601.
2,2′-[(1R,3S)-5-(4-Cyanopyridin-3-yl)cyclohexane-1,3-diyl]di-
isonicotinonitrile (4). Mp 140-142 °C. IR (thin film from CHCl3)
ν (cm-1): 2926, 2856, 2236, 1694, 1594, 1550, 1473, 1412, 840,
754. 1H NMR (400 MHz, CHCl3): δ 8.85 (s, 1H), 8.74 (d, J ) 4.8
Hz, 2H), 8.65 (d, J ) 4.4 Hz, 1H), 7.50 (d, J ) 4.4 Hz, 1H), 7.48
(s, 2H), 7.39 (dd, J ) 4.8 Hz, 1.4 Hz, 2H), 3.43 (t, J ) 12.2 Hz,
1H), 3.31-3.24 (m, 2H), 2.34-2.28 (m, 3H), 2.17-2.01 (m, 3H).
13C NMR (100.7 MHz, CDCl3): δ 164.5, 149.7, 148.5, 147.7, 122.4,
120.3, 119.1, 115.8, 114.9, 44.9, 39.8, 36.9, 36.5. HRMS (m/z):
calculated for C24H19N6 ([M + H]+), 391.1671; found, 391.1671.
NMR Studies. The solvent used in NMR studies (CDCl3) was
passed through a basic alumina column and stored on 4 Å molecular
sieves. The monosaccharides were dried thoroughly for 12 h under
high vacuum at 61 °C in a drying pistol using P2O5 as the desiccant.
NMR titrations were performed in 5 mm NMR tubes using a Bruker
Avance 400 instrument. For the NOE experiments, the samples were
degassed using nitrogen, and the experiments were carried out on
a Bruker Avance 500 instrument. NMR spectral data were processed
and analyzed using MestReC version 4.4.1.0 (Mestrelab Research).
Acknowledgment. We gratefully acknowledge financial support
from the NIH-NIGMS (5RO1-GM62825). X-ray crystallographic
analysis was conducted by Dr. William W. Brennessel and the X-ray
Crystallographic Facility of the Department of Chemistry at the
University of Rochester.
1
Supporting Information Available: Detailed H NMR, UV,
and fluorescence titration data; details of 1D and 2D NOE
experiments and spectra; and single-crystal X-ray data in CIF
format. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA802229F
9
J. AM. CHEM. SOC. VOL. 130, NO. 29, 2008 9573