Modulating weak interactions for molecular recognition: A dynamic combinatorial analysis for assessing the contribution of electrostatics to the stability of CH-π bonds in water

Article abstract of DOI:10.1002/anie.201411733

Electrostatic and charge-transfer contributions to CH-π complexes can be modulated by attaching electron-withdrawing substituents to the carbon atom. While clearly stabilizing in the gas phase, the outcome of this chemical modification in water is more difficult to predict. Herein we provide a definitive and quantitative answer to this question employing a simple strategy based on dynamic combinatorial chemistry.

Full text of DOI:10.1002/anie.201411733

Angewandte
Chemie
DOI: 10.1002/anie.201411733
Polarizing CH–p bonds
Modulating Weak Interactions for Molecular Recognition: A Dynamic
Combinatorial Analysis for Assessing the Contribution of
Electrostatics to the Stability of CH–p Bonds in Water**
Ester Jimꢀnez-Moreno, Ana M. Gꢁmez, Agatha Bastida, Francisco Corzana, Gonzalo Jimꢀnez-
Oses, Jesffls Jimꢀnez-Barbero, and Juan Luis Asensio*
Abstract: Electrostatic and charge-transfer contributions to
CH–p complexes can be modulated by attaching electron-
withdrawing substituents to the carbon atom. While clearly
stabilizing in the gas phase, the outcome of this chemical
modification in water is more difficult to predict. Herein we
provide a definitive and quantitative answer to this question
employing a simple strategy based on dynamic combinatorial
chemistry.
phase, polarized CH moieties form tighter complexes with
aromatic systems than nonpolarized ones.[3] Similarly, replace-
ment of the CH group by an OH function (which, from an
electrostatic perspective, might resemble an extremely polar-
ized CH) leads to a larger enhancement in stability.[3] Hence,
CH polarization represents a simple chemical strategy for the
stabilization of CH–p complexes in the absence of solvent.
The outcome of this simple chemical modification in
water is, however, more difficult to predict. In fact, exposed
ROH–p contacts[10] are usually highly destabilized in water
due to the competence provided by the stronger and
ubiquitous ROH···H2O hydrogen bonds. Analogously, the
polarization of a particular CH bond could also enhance its
own polar interactions with the solvent, thus leading to
a minor stabilization, a null effect, or even a significant
destabilization of the CH–p contact (Figure S1). However, to
the best of our knowledge, no unambiguous experimental
evidences have been presented so far to support or disprove
the stabilizing role of CH polarization in water. Moreover, the
corresponding effect, if any, has never been quantified.
These questions have profound implications for the
molecular recognition and drug design fields, and providing
the right answers will contribute to the longstanding challenge
of unravelling the complicated influence of water on molec-
ular association.[11]
N
owadays it has become clear that CH–p interactions[1–3]
play a key role in a variety of molecular recognition
processes[4,5] including conformation stabilization,[6] crystal
packing,[7] formation of gas-phase clusters,[8] or chiral dis-
crimination.[9] Nevertheless, there are still aspects of these
interactions that remain poorly understood.
Most theoretical analyses have revealed that, in the gas
phase, the stability of CH–p complexes largely arises from
dispersion forces.[2,3] However, although relatively small,
electrostatic interactions cannot be neglected, because they
contribute to both the strength and the directionality of the
contact. Actually, this dual dispersive/electrostatic nature
represents an essential feature of CH–p bonds and explains
their ubiquitous presence under different environments.[2] It is
important to note that electrostatic forces can be enhanced by
the presence of electron-withdrawing substituents attached to
the interacting carbon atom (Figure S1). Indeed, in the gas
To this aim, we have employed a modified version of the
simple strategy recently developed by us,[12] based on the
principles of dynamic combinatorial chemistry.[13] This meth-
odology is schematically represented in Figure 1 (see also
Figures S2–S10). First, several model systems, including
[*] E. Jimꢀnez-Moreno, Dr. A. M. Gꢁmez, Dr. A. Bastida,
Dr. J. L. Asensio
Departamento de Quꢂmica Bio-orgꢃnica
Instituto de Quꢂmica Orgꢃnica General (IQOG-CSIC)
Juan de la Cierva 3, E-28006 Madrid (Spain)
E-mail: juanluis.asensio@csic.es
a
reactive 3-amino-a-d-allopyranose scaffold (unit I in
Figure 1) and alternative CH–p donor units (dubbed II and
IIa in Figure 1) were designed and synthesized. As a next step,
buffered water solutions containing equimolecular amounts
of two or more specific model compounds were treated with
an arylacetaldehyde to form a dynamic mixture of hemi-
aminals/imines in exchange. It should be noted that, in these
transient species, the aromatic system is optimally positioned
to establish intramolecular interactions with a single face of
the alternative CH–p donor units (a conclusion based on our
previous experience with related altrose disaccharides and
also supported by molecular dynamics calculations). The
resulting energy contributions depend on the chemical nature
of the CH–p donor and acceptor units, and render the
hemiaminal/imine species nonequivalent in terms of stability.
Consequently, they exhibit a distinct equilibrium population.
Subsequent chemical reduction of the transient species with
Dr. F. Corzana
Departamento de Quꢂmica
Universidad de la Rioja (Spain)
Dr. G. Jimꢀnez-Oses
Department of Chemistry and Biochemistry
University of California
Los Angeles (USA)
Dr. J. Jimꢀnez-Barbero
Centro de Investigaciones Biolꢁgicas (CSIC), Madrid (Spain), CIC-
bioGUNE, Derio-Bizkaia (Spain), and Ikerbasque, Bilbao (Spain)
[**] This investigation was supported by research grants of the Spanish
“Plan Nacional” (MINECO) CTQ2010-19073, CTQ2013-45538-P,
CTQ2012-32025, and CTQ2012-32114.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201411733.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
with a substoichiometric amount of the corresponding
aldehyde to establish an effective competition (see the
Experimental Section). The solution was then allowed to
equilibrate and then, sodium cyanoborohydride was added.
After completion of the reduction step, the relative fraction of
the two secondary amine products was evaluated from the
well-resolved NMR signals. The obtained results are repre-
sented in Figure 2 and Table 1. With aldehyde 5, the product
ratio (denoted as R in Figure 2) A-5/C-5 amounts to 1.8–2.0,
implying that the interaction energy between the CH–p-
Figure 1. Schematic representation of the dynamic combinatorial
approach employed to test the influence of CH polarization on the
stability of CH–p complexes (see the main text). Our library of CH–p
acceptor systems (1–5) is represented in the yellow square. The library
of CH–p donor units (A–D) is shown at the bottom together with the
calculated electrostatic potential surfaces. Partial charges at the
interacting axial hydrogens are also represented.
Figure 2. a) Example of competition experiments performed with pair
A/C and aldehyde 5 (see the main text). 1D-NMR spectra (anomeric
proton region) collected at t=0 and t=5 h (reaction completion) are
shown at the bottom. The ratio between products C-5 and A-5
(represented above) is highlighted (grey-filled square). The outcome of
the experiment performed with aldehyde 1 is also represented for
comparison (dotted black square). b) Comparison of the product
ratios (R, shown in black) obtained with different aldehydes. Interac-
tion free energies (kcalmolÀ1) for the CH–p contacts established in
the respective products are represented in grey.
an externally added reagent irreversibly converts them into
a nonequimolecular mixture of secondary amines, whose
relative populations can be easily evaluated by NMR
methods. These data were then employed to determine the
relative interaction free energies (DG) of each complex.
The employed library of the novel 3-amino-a-d-allopyr-
anose derivatives (A–D) and arylacetaldehydes (1–4) is
represented in Figure 1. The former compounds include
CH–p donor units with zero (B), one (C), and three (D)
polarized CH groups exposed at the interacting surface.
Compound A, with no CH–p donor unit, was also synthesized
as a reference. Regarding the aldehydes, we prepared
derivatives equipped with both, electron-rich and electron-
poor CH–p acceptor aromatic systems. Compound 5 (2-
phenylacetaldehyde) was employed as control, because it
incorporates a significantly smaller aromatic ring and is
consequently unable to establish optimum contacts with the
CH–p donor units upon reaction with the allose scaffold.
We then performed competition experiments in NMR
tubes, employing the A/C pair and the different arylacetalde-
hydes. Thus, 1:1 A/C mixtures in buffered water were treated
Table 1: Net interaction free energies (kcalmolÀ1) for the CH–p com-
plexes established between donors present in B–C and the aromatic
rings of 1, 3, and 4. Stability differences (DDG, kcalmolÀ1) between
complexes involving three and zero polarized CH groups are also
represented. These values represent a substantial fraction (DDG (%),
right column) of the net interaction free energies estimated for D.
B
C
D
DDG[c]
DDG
zero polarized one polarized three polarized (D–B)
[%]
CH–p[b]
CH–p[a]
CH–p[b]
1
3
4
0.97Æ0.11
0.62Æ0.07
0.57Æ0.11
1.25Æ0.10
0.88Æ0.06
0.64Æ0.10
1.94Æ0.11
1.28Æ0.13
0.84Æ0.11
0.91Æ0.09 47
0.65Æ0.06 51
0.23Æ0.08 27
[a] Measured from pairwise competition experiments performed with
mixtures A/C. [b] Estimated from column (C), considering the stability
differences represented in Figure 4 (see the Experimental Section).
[c] Estimated from pairwise competition experiments performed with
mixtures D/B.
2
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
donor unit present in C and the aromatic system is in the
range of 0.35–0.41 kcalmolÀ1. This relatively small value
reflects the reduced size of the phenyl ring. In contrast, with
aldehyde 1 (with a larger aromatic ring) the A-1/C-1 product
ratio increases to 7.0–10.0, which corresponds to a complex
stability of 1.15–1.36 kcalmolÀ1. It is noteworthy to mention
that this interaction energy is similar to that reported for
comparable CH–p contacts observed in natural systems. For
example, carbohydrate/aromatic interactions (ubiquitous in
lectin/sugar complexes) typically involve the formation of two
to three simultaneous CH–p bonds and contribute to
carbohydrate recognition by 1–2 kcalmolÀ1.[4]
Thus, the B-1/C-1/D-1 product ratio (with zero, one, and
three polarized CH groups involved in contacts with the
aromatic ring, respectively) is 1:2:4. Moreover, product A-
1 (with no interaction) is almost undetectable. Analogous
experiments performed with aldehyde 2 (Figure 3, bottom)
yielded similar results (product ratio B-2/C-2/D-2 1:2:3.7).
Although the HSQC provided valuable qualitative informa-
tion, more precise conclusions were gathered from integra-
tion of NMR signals in 1D spectra. Thus, smaller sets of allose
derivatives were used for additional competition experiments.
The outcome of these experiments is shown in Figure 4. The
Finally, replacement of the p-rich hydroxynaphthyl
system with electron-deficient aromatic rings, as those present
in 3 and 4, leads to a progressive destabilization of the
interactions, as reflected by the obtained A3/C3 and A4/C4
ratios (see Figure 2). These experiments are in agreement
with previous knowledge on CH–p interactions and support
the validity of our strategy.[4b] Similar tests carried out with
pairs A/B and A/D were inconclusive due to signal over-
lapping or poor signal-to-noise ratio.
Dynamic combinatorial experiments performed with
more complex mixtures provided key evidences for the role
played by CH polarization on the stability of CH–p com-
plexes. More specifically, we carried out simultaneous com-
petition experiments with our four allose derivatives (A/B/C/
D 1:1:1:1), and one single aldehyde. The fraction of the
corresponding reaction products in the final mixture was
evaluated by 2D-HSQC experiments to alleviate the signifi-
cant signal-overlapping present in the NMR spectra. The
obtained results provide strong evidence for a highly stabiliz-
ing influence of CH polarization (Figure 3).
Figure 4. Dynamic combinatorial experiments for the B/C/D triad and
C/D, B/D, B/C pairs (from right to left), employing aldehydes 1 (top),
3 (middle panel), and 4 (bottom). NMR signals for the allose anomeric
proton in the final products are shown. Chemical shifts (ppm) are
indicated above the signals. The measured product ratios (grey),
together with the estimated free energy differences between the
alternative CH–p complexes (black) are also represented. Overlapping
signals were deconvoluted before integration (see the Supporting
Information).
net interaction free energies for the alternative CH–p donors,
extracted from the results shown in Figures 2 and 4 are
collected in Table 1. Interestingly, for electron-rich aromatic
rings (such as 1), the extra stability provided by CH polar-
ization amounts, on average, to 0.3 kcalmolÀ1 per interacting
CH group. Accordingly, the complex formed by the 1,3-
dioxane unit (with three polarized interacting CH groups) is
ca. 0.6 kcalmolÀ1 and 0.9 kcalmolÀ1 more stable than those
formed by cyclohexanediol (with one polarized interacting
CH group) and cyclohexanol (with no polarized CH groups)
rings, respectively (results further confirmed by directly
Figure 3. Dynamic combinatorial experiments performed with equimo-
lecular solutions of the four allose derivatives (yellow square) and
aldehydes 1 (top) and 2 (bottom). After reaction completion, the final
mixtures were analyzed by means of 2D-HSQC experiments (anomeric
regions shown). NMR signals of the final products together with the
corresponding cross-sections are represented. Their intensities reflect
the relative stabilities of the CH–p contacts established in the final
products (also shown). NMR signal corresponding to unreacted
derivative A is also labeled.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
systems by also measuring these values from the relative populations
of the intermediate imines/enamines prior to reduction (these species
are intrinsically unstable in water but could be detected employing
concentrated NMR samples and long acquisition times; Figure S8).
Net interaction free energies for complexes involving the CH–p
donor unit present in C were measured from pairwise competition
experiments with equimolecular A/C mixtures. Analogous experi-
ments with mixtures A/B and A/D did not render accurate values for
the corresponding interaction energies (because of signal overlapping
problems or extremely high product ratios). Therefore, the stability of
complexes involving B and D were measured with respect to those
formed by C employing pairwise competition experiments with
equimolecular B/C and D/C mixtures.
inspecting the relative population of the intermediate imine/
enamines before the reduction step, from concentrated NMR
samples Figure S8). Of note, the later value corresponds to
a significant fraction (around 50%) of the net interaction
energy estimated for the 1,3-dioxane unit. This contribution,
together with the global stability of the respective complexes
decreases for very electron-poor aromatic rings such as that
present in 4. However, in all cases the stabilization promoted
by the attachment of electron-withdrawing groups to the
interacting CH functions remains significant (Figure 4 and
Table 1).
Finally, addition of organic cosolvents, such as acetonitrile
(see Figure S11), leads to a progressive destabilization of all
the CH–p complexes, reflecting the attenuation of the
hydrophobic effect (i.e., less polar medium) and the appear-
ance of competitive interactions with the more affine alkyl
groups of the cosolvent.
In conclusion, CH polarization in water largely stabilizes
the CH–p contacts exacerbated by the hydrophobic “solvent
cage” effect, proving that electrostatics and charge transfer
forces are remarkably relevant. It could be anticipated that
more polar CH bonds would also experience stronger
interactions with the solvent, and, therefore, larger desolva-
tion penalties. According to our data, even for fully exposed
CH moieties as those considered herein, this unfavorable
influence does not counteract the improvement in the direct
CH–p contacts established upon complex formation.
Finally, from a drug design perspective, we have shown
that the attachment of electron-withdrawing atoms to the
interacting CH groups is a valid strategy to stabilize CH–p
bonds in physiological environments. Larger increases in
stability could probably be achieved by employing more
electronegative polarizing atoms, such as fluorine.
Received: December 5, 2014
Published online: && &&, &&&&
Keywords: CH–p interactions · drug design ·
.
dynamic combinatorial chemistry · molecular recognition
[1] E. A. Meyer, R. K. Castellano, F. Diederich, Angew. Chem. Int.
Ed. 2003, 42, 1210 – 1250; Angew. Chem. 2003, 115, 1244 – 1287.
[2] a) M. Nishio, M. Hirota, Y. Umezawa, The CH/p interaction.
Evidence, Nature and Consequences, Wiley-VCH, New York,
1998; b) M. Nishio, ChemPhysChem 2011, 13, 13873 – 13900;
c) M. Nishio, Y. Umezawa, J. Fantini, M. S. Weiss, P. Chakrabarti,
Phys. Chem. Chem. Phys. 2014, 16, 12648 – 12683.
[3] a) S. Tsuzuki, A. Fujii, Phys. Chem. Chem. Phys. 2008, 10, 2584 –
2594; b) A. Fujii, K. Shibasaki, T. Kazama, R. Itaya, N. Mikami,
S. Tsuzuki, Phys. Chem. Chem. Phys. 2008, 10, 2836 – 2843.
[4] a) N. K. Vyas, M. N. Vyas, F. A. Quiocho, Science 1988, 242,
1290 – 1295; b) J. L. Asensio, A. Ardꢀ, F. J. CaÇada, J. Jimꢁnez-
Barbero, Acc. Chem. Res. 2013, 46, 946 – 954; c) A. P. Davis, R. S.
Wareham, Angew. Chem. Int. Ed. 1999, 38, 2978 – 2996; Angew.
Chem. 1999, 111, 3160 – 3179; d) A. P. Davis, R. S. Wareham,
Science 2007, 318, 619 – 622; e) Z. R. Laughrey, S. E. Kiehna,
A. J. Riemen, M. L. Waters, J. Am. Chem. Soc. 2008, 130, 14625 –
14633.
[5] a) A. Scarso, L. Trembleau, J. Rebek, J. Am. Chem. Soc. 2004,
126, 13512 – 13518; b) P. S. Lakshminarayanan, D. K. Kumar, P.
Ghosh, J. Am. Chem. Soc. 2006, 128, 9600 – 9601; c) A. Frontera,
C. Garau, D. Quinonero, P. Ballester, A. Costa, P. M. Deya, Org.
Lett. 2003, 5, 1135 – 1138.
In our opinion, this work provides an unambiguous and
quantitative answer to a simple question that is of general
relevance in different fields of chemistry and biology.
Experimental Section
[6] a) M. Brandl, M. S. Weiss, A. Jabs, J. Suhnel, R. Hilgenfeld, J.
Mol. Biol. 2001, 307, 357 – 377; b) Y. Umezawa, M. Nishio,
Nucleic Acids Res. 2002, 30, 2183 – 2192; c) Y. Umezawa, S.
Tsuboyama, H. Takahashi, J. Uzawa, M. Nishio, Tetrahedron
1999, 55, 10047 – 100056.
[7] a) A. Matsumoto, T. Tanaka, T. Tsubouchi, K. Tashiro, S.
Saragai, S. Nakamoto, J. Am. Chem. Soc. 2002, 124, 8891 –
8902; b) S. Nagahama, K. Inoue, K. Sada, M. Miyata, A.
Matsumoto, Cryst. Growth Des. 2003, 3, 247 – 256.
[8] a) K. Shibasaki, A. Fujii, N. Mikami, S. Tsuzuki, J. Phys. Chem.
A 2006, 110, 4397 – 4404; b) K. Shibasaki, A. Fujii, N. Mikami, S.
Tsuzuki, J. Phys. Chem. A 2007, 111, 753 – 758.
[9] R. Carrillo, M. Lopez-Rodrꢂguez, V. S. Martꢂn, T. Martin,
Angew. Chem. Int. Ed. 2009, 48, 7803 – 7808; Angew. Chem.
2009, 121, 7943 – 7948.
[10] a) S. Scheiner, T. Kar, J. Pattanayak, J. Am. Chem. Soc. 2002,
124, 13257 – 13264; b) M. S. Weiss, M. Brandl, J. Suhnel, D. Pal,
R. Higenfeld, Trends Biochem. Sci. 2001, 26, 521 – 523.
[11] a) P. W. Snyder, M. R. Lockett, D. T. Moustakas, G. M. White-
sides, Eur. Phys. J. Special Topics 2014, 223, 853 – 891; b) S. K.
Pal, A. H. Zewail, Chem. Rev. 2004, 104, 2099 – 2123; c) F.
Biedermann, W. M. Nau, H.-J. Schneider, Angew. Chem. Int. Ed.
2014, 53, 11158 – 11171; Angew. Chem. 2014, 126, 11338 – 11352;
Experimental procedures for the synthesis of derivatives A–D and 1–
4 are described in the Supporting Information (SI). Dynamic
combinatorial competition experiments were performed as follows.
Equimolecular mixtures of two to four derivatives ( ꢀ 1 mm each) in
D2O (10 mm phosphate, pH 6.2) were treated with a substoichiometric
amount of the corresponding arylacetaldehyde. The amount of the
later was carefully adjusted to achieve conversions of the allose
derivatives not superior to 40% (usually in the 300–600 mm range
depending on the number of components present in the mixture).
After a 2 h equilibration period, we added sodium cyanoborohydride
(5 mm). Formation of the reaction products was followed by 1D-
NMR. At reaction completion, qualitative analysis of the final
mixtures was performed by looking at the intensity of key signals of
the products in 2D-HSQC NMR experiments. We derived the ratio
between the alternative secondary amine products by integrating
NMR signals in 1D experiments acquired with long relaxation delays.
Overlapping signals were deconvoluted before integration employing
the line-fitting module implemented in the program MestreNova (see
SI). The obtained values were assumed to reflect differences in the
population of the corresponding imine/enamine intermediates and
therefore, employed to derive the relative stability (DDG) of the
alternative stacking modes. Our protocol was validated in selected
4
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
d) H.-J. Schneider, Angew. Chem. Int. Ed. 2009, 48, 3924 – 3977;
Angew. Chem. 2009, 121, 3982 – 4036; e) G. V. Oshovsky, D. N.
Reinhoudt, W. Verboom, Angew. Chem. Int. Ed. 2007, 46, 2366 –
2393; Angew. Chem. 2007, 119, 2418 – 2445.
Corzana, G. Jimꢁnez-Oses, C. Gonzꢀlez, A. M. Gꢃmez, A.
Bastida, J. Revuelta, J. L. Asensio, J. Am. Chem. Soc. 2010, 132,
12074 – 12090.
[13] a) O. Ramstrom, J. M. Lehn, Nature 2002, 415, 26 – 36; b) P. T.
Corbett, J. Leclaire, L. Vial, K. R. West, J. Wietor, J. K. M.
Sanders, S. Otto, Chem. Rev. 2006, 106, 3652 – 3711.
[12] a) A. G. Santana, E. Jimꢁnez-Moreno, A. M. Gꢃmez, F. Cor-
zana, C. Gonzꢀlez, G. Jimꢁnez-Oses, J. Jimꢁnez-Barbero, J. L.
Asensio, J. Am. Chem. Soc. 2013, 135, 3347 – 3350; b) T. Vacas, F.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Polarizing CH–p bonds
Electrostatic and charge-transfer contri-
butions to CH–p complexes can be
modulated by attaching electron-with-
drawing substituents to the carbon atom.
Although they have a clearly stabilizing
effect in the gas phase, the influence of
this chemical modification for the
E. Jimꢁnez-Moreno, A. M. Gꢂmez,
A. Bastida, F. Corzana, G. Jimꢁnez-Oses,
J. Jimꢁnez-Barbero,
J. L. Asensio*
&&&& — &&&&
Modulating Weak Interactions for
Molecular Recognition: A Dynamic
Combinatorial Analysis for Assessing the
Contribution of Electrostatics to the
Stability of CH–p Bonds in Water
behavior in water is more difficult to
predict. Dynamic combinatorial chemis-
try is used to provide a definitive and
quantitative answer to this question.
6
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!