Investigation of synthetic hosts that model cation-π sites found at protein binding domains

Article abstract of DOI:10.1021/jo011124c

Small cyclophanes containing aromatic groups and dialkyl ammonium ions were created as model systems of the cation-π complexes found at some protein binding domains. The hosts had different shapes in order to investigate the effect the arrangement of ammonium ions to aromatic surfaces has on their reactivity. pK_a values of the hosts were substantially different in DMSO or (95/5) DMSO/D20 solutions, which showed that the ions existed in different environments of the hosts. Electrostatic charges, as determined by density functional calculations, revealed that the magnitude of a cationic charge depends on its position relative to an aromatic ring. Association constants of the hosts bound to the sodium salt of N-acetyl phenylalanine in d₆-DMSO and in d₆-DMSO/D₂0 solutions were inversely proportional to the magnitude of the hosts' acidity constants. These results suggest that the magnitude of the positive charge for cationic groups of cation-π complexes is reduced by being associated with electron-rich faces of aromatic rings. The aromatic rings, however, lessen the desolvation penalty that must be overcome for ligand binding, giving an overall more favorable association.

Full text of DOI:10.1021/jo011124c

2160
J . Org. Chem. 2002, 67, 2160-2167
In vestiga tion of Syn th etic Hosts Th a t Mod el Ca tion -π Sites F ou n d
a t P r otein Bin d in g Dom a in s
Vadims Dvornikovs and David B. Smithrud*
Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 45221-0172
david.smithrud@uc.edu.
Received December 4, 2001
Small cyclophanes containing aromatic groups and dialkyl ammonium ions were created as model
systems of the cation-π complexes found at some protein binding domains. The hosts had different
shapes in order to investigate the effect the arrangement of ammonium ions to aromatic surfaces
has on their reactivity. pKa values of the hosts were substantially different in DMSO or (95/5)
DMSO/D2O solutions, which showed that the ions existed in different environments of the hosts.
Electrostatic charges, as determined by density functional calculations, revealed that the magnitude
of a cationic charge depends on its position relative to an aromatic ring. Association constants of
the hosts bound to the sodium salt of N-acetyl phenylalanine in d6-DMSO and in (95/5) d6-DMSO/
D2O solutions were inversely proportional to the magnitude of the hosts’ acidity constants. These
results suggest that the magnitude of the positive charge for cationic groups of cation-π complexes
is reduced by being associated with electron-rich faces of aromatic rings. The aromatic rings,
however, lessen the desolvation penalty that must be overcome for ligand binding, giving an overall
more favorable association.
In tr od u ction
For example, the interaction between K+ and benzene
in the gas phase produces 19 kcal/mol of energy, which
is actually 1 kcal/mol more favorable than the binding
energy observed for K+ interacting with H2O.5c Artificial
receptors, based on cyclophanes, bind tetraalkylammo-
nium salts tighter than neutral guests in water and
organic solvents,9 which strongly supports the hypothesis
that the cation-π interaction is important source of free
energy for the binding of quaternary ammonium ligands
by proteins.10 Lariet ether based receptors interact with
Na+ and K+ in the solid and solution phases.11 Dough-
erty’s studies, along with theoretical calculations, have
revealed that the free energy is obtained through favor-
able interactions between the cation and the electron-
rich face of an aromatic ring, which is a consequence of
the ring’s permanent quadrupole moment.6f,9 Polarization
and dispersion interactions may also contribute to com-
plex stability.6b,7c,12
Since the time it was first noted that amino groups
were predominantly associated with aromatic rings in
X-ray crystallographic structures of proteins,1 the cat-
ion-π interaction has been postulated to be an important
component of protein stabilization,2 ligand recognition,3
and catalysis.4 Because protein interactions involve a
complex mixture of noncovalent forces, important in-
sights into the cation-π interaction have been obtained
by investigating model systems. Gas-phase studies of
small ions with benzene rings,5 high-level calculations,6,7
and model peptides8 have shown that cations bind
aromatic rings with a substantial amount of free energy.
* To whom correspondence should be addressed. FAX: (513) 556-
9239.
(1) (a) Burley, S. K.; Petsko, G. A. FEBS Lett. 1986, 203, 139-143.
(b) Perutz, M. F.; Fermi, G.; Abraham, D. J .; Poyart, C.; Bursaux, E.
J . Am. Chem. Soc. 1986, 108, 1064-1078.
(2) (a) Fernandez-Recio, J .; Romero, A.; Sancho, J . J . Mol. Biol. 1999,
290, 319-330. (b) Ting, A. Y.; Shin, I.; Lucero, C.; Schultz, P. G. J .
Am. Chem. Soc. 1998, 120, 7135-7136.
(3) (a) Wu, Z.; Demma, M.; Strickland, C. L.; Radisky, E. S.; Poulter,
C. D.; Le, H. V.; Windsor, W. T. Biochemistry 1999, 38, 11239-11249.
(b) Schluckebier, G.; Labahn, J .; Granzin, J .; Saenger, W. Biol. Chem.
1998, 379, 389-400.
(4) (a) Gallivan, J . P.; Dougherty, D. A. P. Natl. Acad. Sci. U.S.A.
1999, 96, 9459-9464. (b) Inoue, Y.; Sugio, S.; Andzelm, J .; Nakamura,
N. J . Phys. Chem. A 1998, 102, 646-648.
(5) (a) Deakyne, C. A.; Meot-Ner (Mautner), M. J . Am. Chem. Soc.
1985, 107, 474-479. (b) Meot-Ner (Mautner), M.; Deakyne, C. A. J .
Am. Chem. Soc. 1985, 107, 469-474. (c) Summer, J .; Nishizawa, K.;
Kebarle, P. J . Phys. Chem. 1981, 85, 1814-1820.
(6) (a) Gapeev, A.; Dunbar, R. C. J . Am. Chem. Soc. 2001, 123,
8360-8365. (b) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M.;
Tanabe, K. J . Am. Chem. Soc. 2000, 122, 11450-11458. (c) Wouters,
J . J . Comput. Chem. 2000, 21, 847-855. (d) Gallivan, J . P.; Dougherty,
D. A. J . Am. Chem. Soc. 2000, 122, 870-874. (e) Dunbar, R. C.;
Klippenstein, S. J .; Hrusak, J .; Stockigt, D.; Schwarz, H. J . Am. Chem.
Soc. 1996, 118, 5277-5283. (f) Basch, H.; Stevens, W. J . J . Mol. Struct.
(THEOCHEM) 1995, 338, 303-315. (g) Lee, J . Y.; Lee, S. J .; Choi, H.
S.; Choi, S. J .; Kim, K. S.; Ha, T. K. Chem. Phys. Lett. 1995, 232, 67-
71.
(7) (a) Hay, B. P.; Nicholas, J . B.; Feller, D. J . Am. Chem. Soc. 2000,
122, 10083-10089. (b) Sulpizi, M.; Carloni, P. J . Phys. Chem. B 2000,
104, 10087-10091. (c) Zhu, W. L.; Tan, X. J .; Puah, C. M.; Gu, J . D.;
J iang, H. L.; Chen, K. X.; Felder, C. E.; Silman, I.; Sussman, J . L. J .
Phys. Chem. A 2000, 104, 9573-9580. (d) Dunbar, R. C. J . Phys. Chem.
A 2000, 104, 8067-8074. (e) Zhu, W. L.; J iang, H. L.; Puah, C. M.;
Tan, X. J .; Chen, K. X.; Cao, Y.; J i, R. Y. J . Chem. Soc., Perkin Trans.
2 1999, 2615-2622.
(8) (a) Pletneva, E. V.; Laederach, A. T.; Fulton, D. B.; Kostic, N.
M. J . Am. Chem. Soc. 2001, 123, 6232-6245. (b) Olson, C. A.; Shi, Z.
S.; Kallenbach, N. R. J . Am. Chem. Soc. 2001, 123, 6451-6452.
(9) Ma, J . C.; Dougherty, D. A. Chem. Rev. 1997, 97, 1303-1324.
(10) (a) Sussman, J . L.; Harel, M.; Frolow, F.; Oefner, C.; Goldman,
A.; Toker, L.; Silman, I. Science 1991, 253, 872-879 (b) Shafferman,
A.; Velan, B.; Ordentlich, A.; Kronman, C.; Grosfeld, H.; Leitner, M.;
Flashner, Y.; Cohen, S.; Barak, D.; Ariel, N. EMBO J . 1992, 11, 3561-
3568.
(11) (a) Meadows, E. S.; De Wall, S. L.; Barbour, L. J .; Gokel, G. W.
J . Am. Chem. Soc. 2001, 123, 3092-3107. (b) De Wal, S. L.; Meadows,
E. S.; Barbour, L. J .; Gokel, G. W. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 6271-6276.
(12) Cubero, E.; Luque, F. J .; Orozco, M. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 5976-5980.
10.1021/jo011124c CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/13/2002

Synthetic Hosts That Model Cation-π Sites
J . Org. Chem., Vol. 67, No. 7, 2002 2161
Ch a r t 1
The studies mentioned above examine the interaction
energies between a cation and a π-system. They do not
explore the interaction energies that occur between a
cation-π site, displayed at some protein binding domains,
and polar or negatively charged ligands. For example,
Arg43 of the human growth hormone receptor (hGHR) is
nestled between Trp104 and Trp169, and this cation-π
site contacts Thr175 and Asp171 of the human growth
hormone binding protein.13 An even more pronounced
cation-π stacked domain is seen in the extracellular
domain of hGHR.14 Lys-Tyr-Arg-Phe-Arg-Trp-Lys exists
in an almost continuous stacked arrangement, but the
biological function, if any, of this domain has not been
determined.9 At first glance, it seems reasonable that a
cation-π site would interact favorably with negatively
charged or polar ligands. Ostensively, the role of aromatic
rings is to partially desolvate the cation, which lowers
the desolvation penalty for salt bridge formation or to
align the positive charge, giving more structure to the
binding domain. On the other hand, favorable interac-
tions between the positively charged side chain and the
aromatic surface of a cation-π site must detract from the
free energy available for salt bridge formation.
To investigate the structure-function relationship of
cation-π sites exposed at binding domains, we have
created a series of artificial receptors that contain
hydrophobic patches and dialkyl ammonium ions. The
shapes of the hosts were altered to determine what effect
the spatial arrangement of ammonium ions and aromatic
rings has on a host’s activity. Activity of the cation-π
domain was ascertained by comparing the acidity con-
stants of the hosts and their association constants for
complexes of the sodium salt of N-acetyl phenylalanine
(Phe(I)). Phe(I) was chosen as a representative ligand
because it contains a carboxylate and a phenyl ring that
are likely to interact favorably with a cation-π site. We
specifically wanted to determine whether the spatial
arrangement of ammonium ions to aromatic rings affects
the magnitude of the cationic charge, whether the shape
of cation-π sites affects their ability to form a salt bridge,
and what affect solvent has on these interactions.
erties of a half host were investigated as well to deter-
mine whether the more rigid alignment of cationic
charges and aromatic rings in preformed pockets en-
hances Phe(I) association.
Syn th esis of th e Hosts. The hosts were designed to
have a rigid aromatic domain, defined by a diphenyl-
methane unit,15 and cationic domain through a trans-
cyclohexanediamine ring in which the amino groups are
in a fixed gauche orientation. Alkylation of bisphenol 1
with para, meta, ortho bromides 2a -c in the presence of
sodium hydride gave the corresponding diesters 3a -c in
good yields (Scheme 1). Reduction of the esters to alcohols
using either DIBALH or LAH also resulted in the
undesirable reduction of the tertiary amide to the amine.
Selective ester reduction was obtained by using lithium
borohydride generated in situ from sodium borohydride
and lithium chloride. Alcohols 4a -c were readily con-
verted into aldehydes 5a -c via Swern oxidation. Host
formation was completed upon the addition of 1,2-
diaminocyclohexane 6 through a reductive amination
reaction. Disappearance of the aldehyde signal in 1H
NMR spectra and the concomitant emergence of the
imine proton provided an easy means to monitor the
reaction progression. Once the imine was formed, sodium
borohydride in methanol was added to give essentially
pure hosts. Half host 7 was prepared via a condensation
reaction between trans-cyclohexanediamine 6 and para-
tolualdehyde followed by reduction with sodium borohy-
dride. The hosts were exposed to HCl(g) and studied as
the dichloride salts.
Resu lts a n d Discu ssion
Host Design . Three hosts, referred to as para, meta,
and ortho, were designed to bind Phe(I) in DMSO and
DMSO-water mixtures through mainly electrostatic
interactions. The pockets of the hosts were made too
small to fully encapsulate an aromatic guest. We wanted
to mimic the narrow grooves found at protein surfaces
and not their deep binding pockets. Thus, van der Waals
interactions and the hydrophobic effect, for complexes
formed in a DMSO-water mixture, should be minor
contributors to the overall binding energy. Changes in
the position of the hydrophobic pocket to the dialkyl
ammonium ions were accomplished by changing the
substitution pattern of the central aromatic ring, which
we refer to as the hinge. The cyclohexyl ring also gives
the hosts complex shapes similar to the clefts and bulges
observed at protein binding domains. The physical prop-
Hosts’ geometries were obtained by analyzing two-
dimensional H NMR spectra and performing molecular
1
modeling. Proton resonances were readily assigned from
one-dimensional 1H NMR, COSY, and NOESY spectra.16
There were no long range NOEs observed in the para
host, which is consistent with it having a flat shape. The
meta host appears to be slightly bent because a weak
NOE was observed between a â-H of the cyclohexyl
(15) Odashima K., Koga K. In Cyclophanes; Keehn, P. M., Rosenfeld,
S. M., Eds.; Academic Press: New York, 1983; Vol. II.
(16) (a) Kessler, H.; Bermel, W.; Muller, A.; Pook, K. H. In The
Peptides; Underfriend, S., Meienhofer, J ., Eds.; Academic Press:
London, 1985; Vol. 7, pp 438-473. (b) Biomolecular NMR Spectroscopy;
Evans, J . N. S.; Oxford University Press: Oxford, 1995; Chapter 4. (c)
Dyson, H. J .; Wright, P. E. Annu. Rev. Biophys. Biophys. Chem. 1991,
20, 519-538. (d) Deber, C. M.; Madison, V.; Blout, E. R. Acc. Chem.
Res. 1976, 9, 106-109. (e) Bystrov, V. F.; Portnova, S. L.; Balashova,
T. A.; Koz’min, S. A.; Gavrilov, Yu. D.; Afanas’ev, V. A. Pure Appl.
Chem. 1973, 36, 19-25. (f) Bax, A.; Davis, D. G. J . Magn. Reson. 1985,
63, 207-213.
(13) Clackson, T.; Ultsch, M. H.; Wells, J . A.; de Vos, A. M. J . Mol.
Biol. 1998, 277, 1111-1128. (b) de Vos, A. M.; Ultsch, M.; Kossiakoff,
A. A. Science 1992, 255, 306-312.
(14) Livnah, O.; Stura, E. A.; J ohnson, D. L.; Middleton, S. A.;
Mulchay, L. S.; Wrighton, N. C.; Dower, W. J .; J olliffe, L. K.; Wilson,
I. A. Science 1996, 273, 464-471.

2162 J . Org. Chem., Vol. 67, No. 7, 2002
Dvornikovs and Smithrud
Sch em e 1
Ta ble 1. Acid ity Con sta n ts for Hosts a t 298 K
diammonium ring and a meta proton of a nonhinge-
aromatic ring. A more intense NOE was observed be-
tween these two protons in the ortho host, showing that
it has a very compacted shape. Energy minimized struc-
tures (Monte Carlo search procedure, MMFF94 force field
as presented by SpartanPro,17 Figure 1) were consistent
with the observed NOEs. The calculated structures
showed that the size of the hydrophobic pockets decrease
from the para to the ortho host and the N+-H’s are held
in different environments. The four N+-H’s (italics are
used to indicate that the H-atom is being discussed) of
the para host are held within an aromatic pocket,
pointing inside. For the meta and ortho hosts, two N+-
H’s are held within and two are held without their hosts’
pockets. All N+-H’s are held at the face of an aromatic
ring, except for the two N+-H’s of the meta host that are
positioned outside its cleft nearer to an aromatic edge
(H3 and H4, Figure 1).
Whether an N+-H is positioned at the face or edge of
an aromatic ring should be reflected in its pKa. Increasing
the electronic density around a N+-H should stabilize its
positive charge and raise its pKa, whereas decreasing the
electronic density should have the opposite effect. Lund-
Katz observed that the pKa’s of Lys residues are raised
when they reside near aromatic surfaces and lowered
when they are held close together because of electrostatic
repulsion.18 The electron density of Thr13 OH proton is
increased by being placed in the center of the aromatic
ring of Tyr6, according to density functional theory (DFT)
calculations performed on this complex found in µ-glu-
tathione S-transferase (µ-GST).7b
pKa1
(DMSO/H2O,
95/5)
pKa2
(DMSO/H2O,
95/5)
a
pKa1
pKa2
(DMSO)
host
(DMSO)
para
meta
ortho
half
6.2
4.5
3.6
5.4
9.8
9.2
8.2
9.5
5.8
3.8
3.3
5.1
9.6
8.4
8.4
9.9
a
Uncertainties for the pKa’s < 4%.
free hosts (Figure 1) and the cation-π theory. Ammonium
ions held at the electron-rich face of aromatic rings, such
as the ones buried within the para host, are stabilized
through the cation-π interaction, and thus their pKa’s are
raised. Similarly large pKa’s are observed for the half
host, which has two N+-H’s (H1 and H2, Figure 1D) held
in the face of aromatic rings. Although the pKa’s of the
meta and ortho hosts are more similar, their energy
minimized structures are dissimilar. H1 and H2 of the
meta host are held within an aromatic cleft, which is
smaller than the pocket of the para host. The other two
N+-H’s of the meta host are positioned outside the pocket
(H3 and H4). The N+-H’s of the ortho host are not held
within an aromatic cleft, but instead they are arranged
in a more elongated groove. Each N+-H group is associ-
ated with a sole aromatic ring. The calculated partial
charges (vide infra) of the N+-H’s positioned outside the
pocket of the meta host are substantially larger (espe-
cially H3) than for the ortho hosts. Because the pKa’s are
similar, the charges of the dialkyl ammonium ions of the
meta and ortho hosts should be more similar. To account
for this inconsistency, the exposed N+-H’s of the meta
host could interact strongly with solvent molecules, which
reduces the magnitude of their charge. The meta host
has a slightly greater drop in pKa1 and pKa2 (∆pKa’s =
0.7) than the other hosts (∆pKa’s = 0.3) when comparing
the values obtained in DMSO to DMSO/water solutions.
This larger solvent effect on pKa1 and pKa2 is consistent
with the meta host’s N+-H proton being more exposed to
solvent. A smaller drop is observed across the host series
for pKa2. This smaller change may be caused by the
stabilization of the remaining single N+-H through
interactions with the electron pair of its neighboring
amine.
In our studies, a large drop in pKa1 is observed when
one compares the para to the meta and to the ortho host
in DMSO (Table 1). This lowering across the series is
consistent with the energy minimized structures of the
(17) (a) SpartanPro; Hehre, J . W.; Yu, J .; Klunzinger, P. E.; Lou, L.
Brief Guide to Molecular Mechanics and Quantum Chemical
A
Calculations; Wavefunction, Inc.: Irvine, CA, 1998. The Becke-Perdew
model: Becke, A. D. Phys. Rev. A 1988, 38, 3089-3100. (b) Perdew, J .
P. Phys, Rev. B 1986, 33, 8822-8824.
(18) (a) Lund-Katz, S.; Laplaud, P. M.; Phillips, M. C.; Chapman,
M. J . Biochemistry 1998, 37, 12867-12874. (b) Lund-Katz, S.; Phillips,
M. C.; Mishra, V. K.; Segrest, J . P.; Anantharamaiah, G. M. Biochem-
istry 1995, 34, 9219-9226.

Synthetic Hosts That Model Cation-π Sites
J . Org. Chem., Vol. 67, No. 7, 2002 2163
F igu r e 1. Structures of the (A) para, (B) meta, (C) ortho, and (D) half hosts without (Free) and with (Bound) Phe(I). Free
conformations were obtained through Monte Carlo simulations, whereas the bound structures were solved using a combination
of Monte Carlo docking simulations combined with conjugate gradient energy minimizations. Phe(I), C-H, N-acetylpiperidine,
and cyclohexyl rings are removed for clarity. These structures were used to calculate the electrostatic charges of the N-H (H1-H4)
hydrogens (DFT calculations), and the values are given in Table 2.
The combination of experimental and calculated results
that interact favorably with the electron-rich face of
show that larger pKa’s are obtained for ammonium ions
aromatic rings. Stronger cation-π interactions occur for

2164 J . Org. Chem., Vol. 67, No. 7, 2002
Dvornikovs and Smithrud
Ta ble 2. Electr osta tic Ch a r gesa for Hosts a n d Th eir
Dia m in oeth yl P or tion s Held in En er gy Min im ized
Str u ctu r es in F r ee a n d Bou n d to P h e(I) Con for m a tion s
Ta ble 3. Associa tion Con sta n ts for Host-P h e(I)
Com p lexes a t 298 Ka
KA (M-1
(d6-DMSO/D2O)
)
host
para
structureb
H1c
H2
H3
H4
KA (M-1
)
host
(d6-DMSO)
(95/5)
free
0.171
0.283
0.160
0.187
0.338
0.203
0.121d
0.285
0.205
0.156
0.290
0.238
0.210
0.317
0.186
0.239
0.321
0.167
0.258
0.301
0.181
0.156
0.290
0.198
0.256
0.336
0.289
0.330
0.310
0.252
0.205
0.315
0.213
0.273
0.318
0.270
0.257
0.342
0.267
0.298
0.326
0.246
0.229
0.313
0.301
0.273
0.318
0.287
diamine
bound
free
diamine
bound
free
diamine
bound
free
diamine
bound
para
meta
ortho
half
81 ( 5
59 ( 6
51 ( 7
47 ( 5
81 ( 7
65 ( 13
18 ( 5
<5
meta
ortho
half
Obtained from 1H NMR titrations; Phe(I) was the sodium salt.
a
throughout the host series were not substantially differ-
ent. Both results support our conclusion that the mag-
nitude of a host’s pKa is controlled by the juxtaposition
of its N+-H to an aromatic surface.
a
Calculated from density functional theory pBP/DN*.17 b Struc-
tures are defined as free means hosts without guest, diamine
means only the diamine portion of the hosts, and bound means
that the host were bound to Phe(I). c See Figure 1 for atom labeling
To more accurately represent the environment of
protein binding domains, we added 5% H2O to the
DMSO-host solutions. A higher percentage of water
precipitated the hosts. The addition of water lowered the
hosts’ pKa1’s about 10% (Table 1), which can be explained
by the higher basicity of water as compared to DMSO.
The similar drop in pKa1’s throughout the host-series
suggests that their structures are not altered in the
presence of 5% H2O. Furthermore, similar 2D NMR
spectra were observed for each host in both solvent
systems. More variable changes in pKa2’s are observed
with the addition of water except for the meta hosts,
which is once again consistent with its N+-H being more
exposed to solvent.
d
and host geometries. H1 of the ortho host makes contact with
an ether-oxygen atom and is unavailable for binding.
ammonium ions placed within narrow aromatic pockets
than in broader clefts or near a single aromatic ring.
Stronger interactions also occur for cations positioned at
an aromatic face than their edge, which is consistent with
previous mp2 calculations on ammonia-benzene com-
plexes.6b
Differences in the hosts’ pKa’s may result from the
juxtaposition of their dialkyl ammonium ions to each
other and not from the positions of N+-H’s to the aromatic
rings. To address this possibility, electrostatic charges
of the ammonium ions were derived for the energy
minimized structures of the hosts and half host using
DFT calculations (Table 2; pBP/DN* model as presented
by SpartanPro).17 Semiempirical calculations using the
AM1 force field did not produce significant differences
in the magnitude of the charges. DFT calculations have
been used successfully to predict and understand the
properties of cation-π complexes.7 The N-acetylpiperidine
and cyclohexyl portions of the hosts were replaced with
H-atoms after the hosts were minimized to lessen the
computational time (Figure 1). The electrostatic charges
of the hosts’ N+-H’s were compared to ones calculated
for only the diaminoethyl domain of the hosts, which had
the pockets, cyclohexanecarboxamidyl, and cyclohexyl
rings of the minimized structures replaced by H-atoms.
Electrostatic charges of the diaminoethyl domains are
similar throughout the host series (Table 2). Only in the
presence of the host’s pockets are the charges of the N+-
H’s substantially different. In these latter calculations,
smaller partial charges were obtained for N+-H’s posi-
tioned within electron-rich pockets (H1 and H2, Table
2) and larger charges occurred for N+-H’s held at aro-
matic edges or at exposed regions (H3 and H4, Table 2).
These results are consistent with the measured pKa’s, and
they are in accordance with the cation-π theory.
Having different cation-π properties makes these hosts
a good model system of protein receptors, allowing us to
test whether cation-π sites bind negatively charged
ligands and determine what factors promote this associa-
tion. All hosts bound Phe(I) in d6-DMSO (Table 3), which
shows that the dialkyl ammonium ions are accessible for
1
guest association. 2-D H NMR analysis of host-guest
complexes did not reveal NOEs between the binding
components. Thus, it appears that association forms
mainly through salt bridges. We were surprised to find
that the para host binds Phe(I) more favorably in
d6-DMSO than the ortho and meta hosts, which had
similar binding energies. According to their large pKa’s,
N+-H’s of the para host interact the strongest with the
aromatic rings, and thus they should interact the weakest
with the guest. One possible advantage afforded the para
host is that its ammonium ions are both positioned within
an aromatic cleft and partially desolvated, which should
lower the desolvation penalty for salt bridge formation.
A greater solvent effect for association is observed with
the addition of 5% D2O to the solutions (Table 3). The
affinities of the para and meta hosts for Phe(I) are
approximately unchanged. A new NOE was observed
between an aromatic proton of the phenyl ring of Phe(I)
and the â-H of the cyclohexyl diammonium ring of the
para host. This interaction is consistent with the energy
minimized structure of the para host-Phe(I) complex. A
large drop in binding energy is observed for the ortho
and half hosts. Most likely, the more exposed N+-H’s of
the ortho host and the half host, after it undergoes a
rearrangement, vide infra, interact favorably with water
molecules, increasing the desolvation penalty for salt
bridge formation. Another interesting feature is that with
the addition of water, the smallest drops in pKa1’s are
observed for the ortho and half hosts, but their ability to
bind Phe(I) is greatly reduced. The converse phenomenon
is observed for the para and meta hosts. It is possible
Another possibility is that the properties of the con-
jugate bases of the hosts are responsible for the differ-
ences observed in the pKa’s. Molecular modeling results
of the various conjugate bases showed that the structure
of these host were not significantly different than the
fully protonated hosts except there was a slight flattening
of the aromatic rings of the para host in its doubly
deprotonated state. Preference for this conformation
appears to be caused by an H-bond existing between the
amines. Electrostatic charges of the deprotonated amines

Synthetic Hosts That Model Cation-π Sites
J . Org. Chem., Vol. 67, No. 7, 2002 2165
that N+-H’s held within rigid pockets or clefts are
protected from solvent (H1 and H2 for the para and meta
hosts), and thus they are not deprotonated with added
base. These protons instead interact with a ligand such
as Phe(I) because their weaker association with solvent
molecules gives them a smaller desolvation penalty for
complex formation. The lack of a preformed cleft in the
ortho host or a rigid structure of the half host keeps their
ammonium ions from being discriminated into binding
and acidic N+-H’s.
results show that protein binding domains need to be
finely tuned; the arrangement of side chains in cation-π
sites needs to provide both structural stability, desolva-
tion, and potential energy for ligand binding.
Exp er im en ta l Section
Gen er a l Meth od s. All syntheses were carried out under
positive argon pressure, solvents were freshly distilled prior
to use, and other reagents were used as purchased. NMR
spectra used for compound identification and deriving associa-
tion constants were measured at 400.14 or 249.99 MHz for
proton nuclei and 62.9 MHz for carbon nuclei. Chemical shifts
are referenced using an internal TMS standard for proton
spectra and CDCl3 for carbon spectra. COSY and NOESY
spectra were obtained at 298 K, and NOESY experiments had
mixing times of 150, 300, and 500 ms.
Meth yl m -Br om om eth ylben zoa te (2b). Methyl m-toluate
(5.00 g, 33.3 mmol) was dissolved in carbon tetrachloride (85
mL), and N-bromosuccinimide (5.93 g, 33.3 mmol) was added,
followed by benzoyl peroxide (0.16 g, 0.67 mmol). The mixture
was gently refluxed for 4 h, and the white succinimide residue
was filtered off. The solvent was evaporated under reduced
pressure to give a yellow liquid that consisted of the product
and starting methyl m-toluate in a 90:10 ratio (1H NMR). After
the mixture was dissolved in ether and cooled to -78 °C (CO2/
acetone), the product precipitated and liquid was decanted.
Repeating this procedure two more times gave product of
sufficient purity (> 95%) as a yellow liquid at room temper-
ature (6.48 g, 85% yield). The product’s 1H NMR spectrum
matched literature data.19
Methyl o-bromomethylbenzoate (2c) was prepared by the
same procedure from methyl o-methyltoluate in a 63% yield.
The product’s 1H NMR spectrum matched literature data.20
1-{4,4-Bis[4-(4-m et h oxyca r b on ylb en zyloxy)-p h en yl]-
p ip er id in -1-yl}-eth a n on e (3a ). Bisphenol 121 (4.00 g, 12.9
mmol) was dissolved in 50 mL of DMF, and NaH suspension
in mineral oil (0.68 g, 28 mmol) was added in small portions.
The mixture was stirred for 30 min until evolution of hydrogen
ceased, and then methyl p-bromomethylbenzoate 2a (6.48 g,
28.3 mmol) in 30 mL of DMF was added dropwise. The reaction
mixture was stirred for 2 h. After removal of the solvent under
reduced pressure, the residue was taken up in CH2Cl2 and the
inorganic salts were filtered off. Filtrate was concentrated
under reduced pressure and purified by column chromatog-
raphy in CH2Cl2/EtOH (95:5) to give a white foam (7.33 g, 94%
yield): mp 80-82 °C. 1H NMR (CDCl3): 2.08 (3H, s), 2.33 (4H,
m), 3.47 (2H, m), 3.64 (2H, m), 3.92 (6H, s), 5.08 (4H, s), 6.89
(4H, d, J ) 9.0 Hz), 7.14 (4H, d, J ) 9.0 Hz), 7.48 (4H, d, J )
8.0 Hz), 8.05 (4H, d, J ) 7.8 Hz). 13C NMR: 21.5, 36.0, 36.9,
38.6, 52.2, 69.3, 114.9, 127.0, 128.1, 129.7, 129.9, 139.2, 142.3,
156.7, 166.9, 168.9. HR-MS(ES) calcd for C37H37NO7 [M+]
607.2570, found 607.2538.
To relate the measured pKa’s with KA’s, energy mini-
mized host-guest complexes (molecular mechanics cal-
culations using the MM+ force field as presented by
HyperChem) were compared to the conformations of the
free hosts. The structure of the para host free and bound
to Phe(I) is basically unchanged (Figure 1). Minimized
structures of the other hosts bound to Phe(I), however,
were different than their unbound states. These com-
pounds had both sets of N+-H’s aligned for carboxylate
binding instead of having two N+-H’s inside and two
outside the host’s pocket. Similar binding-conformers for
the meta and ortho hosts were observed in the set of
conformers obtained from the Monte Carlo simulations
of the free hosts. These conformers were approximately
2 and 3 kcal/mol higher in energy, respectively, than their
global minima. If these are the binding conformers, the
smaller association constants for the ortho and meta
hosts in d6-DMSO, as compared to the para host, could
be caused by their conversion to less stable conformers
during association. Additional loss in entropic energy
upon association will occur for flexible hosts that bind
through a single conformer. If the hosts do not rearrange
upon guest binding, the larger association constants
obtained for the para host could be due to it having a
greater number of N+-H’s held in the binding domain
than the meta and ortho hosts. One consistent feature
for the computed structures of the free and bound hosts
is that the alignment of the aromatic rings provides for
a pocket, cleft, and groove for the para, meta, and ortho
hosts, respectively. More importantly electrostatic charges
calculated for N+-H’s of the hosts in the binding-conform-
ers have values similar to the ones obtained for free hosts
(Table 2).
Con clu sion
These studies demonstrated that dialkyl ammonium
ions interact with aromatic rings in a manner consistent
with the cation-π theory. According to the theoretical
results and experimental pKa’s, the cationic charge of N+-
H’s positioned at electron-rich aromatic domains are
reduced compared to ones residing at the edges of
aromatic rings or exposed to solvent. Domains displaying
cation-π sites can form stable salt bridges in DMSO and
in DMSO-water mixtures. The binding energy depends
strongly on the desolvation penalty that occurs for
ammonium ions exposed to solvent, especially water.
Therefore, it appears that the activity of positively
charged side chains involved in cation-π sites is reduced,
but this reduction in potential energy for ligand binding
is compensated for by a decrease in the desolvation
penalty for complex formation. Slight differences in the
position of the cation with the aromatic surface produced
large differences in the properties of the hosts. DFT
calculations provided electrostatic charges that were
consistent with the observed properties of the hosts. Our
Meta and ortho diesters 3b,c were prepared by this proce-
dure from methyl m-bromomethylbenzoate 2b and methyl
o-bromomethylbenzoate 2c respectively.
1-{4,4-Bis[4-(3-m et h oxyca r b on ylb en zyloxy)-p h en yl]-
p ip er id in -1-yl}-eth a n on e (3b). White foam, yield 65%, mp
1
62-64 °C. H NMR (CDCl3): 2.08 (3H, s), 2.33 (4H, m), 3.48
(m, 2H), 3.65 (2H, m), 3.92 (6H, s), 5.06 (4H, s), 6.90 (4H, d,
J ) 8.8 Hz), 7.45 (2H, t, J ) 7.5 Hz), 7.62 (2H, d, J ) 7.5 Hz),
8.00 (2H, d, J ) 8.0 Hz), 8.09 (2H, s), 13C NMR: 21.1, 35.7,
36.6, 38.3, 51.9, 69.1, 114.5, 127.7, 128.2, 128.4, 128.9, 130.2,
131.6, 137.2, 138.8, 156.4, 166.5, 168.6. HR-MS(ES) calcd for
C
37H38NO7 [M + 1] 608.2648, found 608.2697.
1-{4,4-Bis[4-(2-m et h oxyca r b on ylb en zyloxy)-p h en yl]-
p ip er id in -1-yl}-eth a n on e (3c). White foam, yield 64%, mp
1
68-70 °C. H NMR (CDCl3): 2.09 (3H, s), 2.34 (4H, m), 3.48
(19) Budesinsky, M.; Exner, O. Magn. Reson. Chem. 1989, 27, 585-
591.
(20) Norman, M. H.; Minick, D. J .; Rigdon, G. C.; J . Med. Chem.
1996, 39, 149-157.
(21) Diederich, F.; Dick, K.; Griebel, D. Chem. Ber. 1985, 118, 3588-
3619.

2166 J . Org. Chem., Vol. 67, No. 7, 2002
Dvornikovs and Smithrud
(2H, m), 3.65 (2H, m), 3.89 (6H, s), 5.46 (4H, s), 6.92 (4H, d,
J ) 8.5 Hz), 7.14 (4H, d, J ) 9.0 Hz), 7.37 (2H, t, J ) 7.8 Hz),
7.55 (2H, t, J ) 7.5 Hz), 7.74 (2H, d, J ) 7.5 Hz), 8.02 (2H, d,
J ) 8.3 Hz). 13C NMR: 21.2, 35.5, 36.7, 38.6, 51.9, 68.0, 114.8,
127.1, 127.3, 127.5, 127.8, 130.5, 132.4, 138.7, 139.4, 156.7,
167.2, 169.0. HR-MS(ES) calcd for C37H38NO7 [M + 1] 608.2648,
found 608.2625.
38.3, 68.9, 114.6, 127.6, 127.8, 128.0, 129.0, 132.9, 136.4, 138.1,
139.0, 156.3, 168.6, 191.8. HR-MS(ES) calcd for C35H34NO5
[M + 1] 548.2437, found 548.2432.
1-{4,4-Bis[4-(2-for m ylben zyloxy)-p h en yl]-p ip er id in -1-
yl}-eth a n on e (5c). Off-white solid, yield 94%, mp 83-85 °C.
1H NMR (CDCl3): 2.09 (3H, s), 2.35 (4H, m), 3.49 (2H, m),
3.65 (2H, m), 5.48 (4H, s), 6.94 (4H, d, J ) 8.5 Hz), 7.16 (4H,
d, J ) 8.5 Hz), 7.52 (2H, t, J ) 7.5 Hz), 7.64 (2H, t, J ) 7.5
Hz), 7.78 (2H, d, J ) 7.5 Hz), 7.90 (2H, d, J ) 7.5 Hz), 10.18
(2H, s). 13C NMR: 21.3, 35.9, 36.8, 38.6, 67.4, 114.8, 127.7,
127.9, 127.9, 133.0, 133.7, 134.0, 139.1, 139.4, 156.5, 168.9,
193.0. HR-MS(ES) calcd for C35H34NO5 [M + 1] 548.2437,
found 548.2432.
1-{4,4-Bis[4-(2-h yd r oxym et h ylb en zyloxy)-p h en yl]-p i-
p er id in -1-yl}-eth a n on e (4c). Diester 3c (0.43 g, 0.71 mmol)
was dissolved in 10 mL of EtOH, and NaBH4 (0.43 g, 11 mmol),
dissolved in 10 mL THF, was added in portions. Anhydrous
LiCl (0.48 g, 11 mmol) was added very slowly over a period of
15 min (exothermic!). The thick mixture was stirred for 24 h,
and the solvent was evaporated under reduced pressure. After
partitioning the residue between methylene chloride and
water, the water layer was extracted with methylene chloride
(2 × 20 mL). The combined extracts were dried over Na2SO4,
and the solvent was evaporated. The residue was purified by
column chromatography in CH2Cl2/EtOH (95:5) to give a white
solid (0.24 g, 61% yield), mp 68-70 °C. 1H NMR (CDCl3): 2.08
(3H, s), 2.33 (4H, m), 3.48 (2H, m), 3.65 (2H, m), 4.73 (4H, s),
5.02 (4H, s), 6.92 (4H, d, J ) 8.5 Hz), 7.15 (4H, d, J ) 8.5 Hz),
7.3-7.5 (8H, m). 13C NMR: 21.1, 35.7, 36.6, 38.5, 60.3, 68.2,
114.7, 127.3, 127.8, 128.4, 128.6, 128.9, 134.5, 138.9, 139.4,
156.5, 169.0. HR-MS(ES) calcd for C35H37NO5Na+ [M + Na+]
574.2569, found 574.2529.
8,11-Dia za -4,14-d ioxa -2(4,4)-N-a cetylp ip er id in a -1,3,6,-
12(1,4)-t et r a b en zen a -9(1,2)-tr a n s-cycloh exa n a cyclot et -
r a d eca p h a n e (P a r a Host). Cyclohexane-trans-1,2-diamine
6 (0.044 g, 0.38 mmol) was dissolved in 20 mL of acetonitrile.
Dialdehyde 5a (0.21 g, 0.38 mmol) in 20 mL of acetonitrile
was added slowly, and the reaction mixture was stirred for
24 h. After removing the solvent under reduced pressure, the
residue was dissolved in 10 mL of methanol, and 7 mL of
CHCl3 and NaBH4 (0.10 g, 2.7 mmol) were added. The reaction
mixture was stirred for 2 h. Solvents were evaporated under
reduced pressure, and the residue was dissolved in CH2Cl2 and
extracted with water (2 × 10 mL). Organic extracts were
collected and dried over Na2SO4, and the solvents were
removed in vacuo. The resulting product (0.24 g, 99%) was of
sufficient purity to be used in the studies (>95%, by 1H NMR),
mp 118-120 °C. 1H NMR (CDCl3): 0.88 (2H, m), 1.04 (2H,
m), 1.24 (2H, m), 1.6-1.9 (4H, m), 2.06 (3H, s), 2.2-2.4 (6H,
m), 3.45 (2H, m), 3.58 (2H, m), 3.55 (2H, d, 13.4 Hz), 3.97 (2H,
d, 13.4 Hz), 5.17 (4H, s), 6.74 (4H, d, 9.2 Hz), 6.99 (4H, d, 7.5
Hz), 7.18 (8H, m). 13C NMR: 21.3, 25.0, 31.3, 35.4, 36.3, 38.5,
50.6, 61.6, 68.8, 115.1, 126.2, 127.5, 127.8, 135.9, 138.6, 139.9,
Para and meta dialcohols 4b,c were prepared by the same
procedure from diesters 3a and 3b, respectively.
1-{4,4-Bis[4-(4-h yd r oxym et h ylb en zyloxy)-p h en yl]-p i-
p er id in -1-yl}-eth a n on e (4a ). White solid, yield 40%, mp 75-
77 °C. 1H NMR (CDCl3): 1.76 (1H, br s), 2.08 (3H, s), 2.33
(4H, m), 3.48 (2H, m), 3.63 (2H, m), 4.71 (4H, d, J ) 3.8 Hz),
5.02 (4H, s), 6.89 (4H, d, J ) 9.1 Hz), 7.13 (4H, d, J ) 8.4 Hz),
7.40 (8H, m). 13C NMR: 21.2, 35.8, 36.7, 38.6, 43.6, 64.6, 69.6,
114.7, 127.0, 127.5, 127.8, 136.0, 138.7, 140.9, 156.7, 169.0.
HR-MS(ES) calcd for C35H37NO5 [M+] 551.2672, found 551.2658.
1-{4,4-Bis[4-(3-h yd r oxym et h ylb en zyloxy)-p h en yl]-p i-
p er id in -1-yl}-eth a n on e (4b). White solid, yield 30%, mp 58-
60 °C. 1H NMR (CDCl3): 1.79 (1H, br s), 2.07 (3H, s), 2.33
(4H, m), 3.48 (2H, m), 3.64 (2H, m), 4.71 (4H, s), 5.02 (4H, s),
6.90 (4H, d, J ) 9.0 Hz), 7.13 (4H, d, J ) 8.5 Hz), 7.3-7.5
(8H, m). 13C NMR: 20.8, 35.3, 36.1, 38.2, 63.8, 69.4, 114.3,
125.4, 125.8, 125.9, 127.4, 128.1, 136.5, 138.4, 141.5, 156.4,
168.9. HR-MS(ES) calcd for C35H37NO5Na+ [M + Na+] 574.2569,
found 574.2529.
1-{4,4-Bis[4-(4-for m ylben zyloxy)-p h en yl]-p ip er id in -1-
yl}-eth a n on e (5a ). A 1.30 mL (2.59 mmol) portion of a 2 M
solution of oxalyl chloride was diluted in 10 mL of CH2Cl2 and
cooled to -65 °C. DMSO (0.40 mL, 5.59 mmol) in 2.4 mL of
CH2Cl2 was added dropwise during a period of 5 min. Stirring
was continued for 10 min, and alcohol 4a (0.24 g, 0.44 mmol)
in 4.8 mL of methylene chloride was added during a period of
5 min. Stirring was continued for an additional 15 min, and
triethylamine (1.65 mL, 11.9 mmol) was added over 5 min.
The reaction mixture was brought to room temperature, and
3 mL of water were added. The organic layer was extracted
with 1 M HCl, 1 M Na2CO3, and then dried over Na2SO4.
Solvent was evaporated under reduced pressure, and the
residue was purified by column chromatography in CH2Cl2/
EtOH (95:5) to give a yellow amorphous solid (0.21 g, 87%
yield), mp 135-137 °C. 1H NMR (CDCl3): 2.08 (3H, s), 2.33
(4H, m), 3.48 (2H, m), 3.64 (2H, m), 5.11 (4H, s), 6.89 (4H, d,
J ) 8.5 Hz), 7.15 (4H, d, J ) 8.5 Hz), 7.58 (4H, d, J ) 7.8 Hz),
7.90 (4H, d, J ) 7.6 Hz), 10.02 (2H, s). 13C NMR: 21.1, 35.7,
36.6, 38.3, 43.6, 68.9, 114.5, 127.2, 127.7, 129.7, 135.7, 139.0,
143.7, 156.2, 168.5, 191.5. HR-MS(ES) calcd for C35H33NO5
[M+] 547.2359, found 547.2332.
+
156.0, 168.7. HR-MS(ES) calcd for C41H48N3O3 [M + 1]
630.3696, found 630.3719.
Meta and ortho cyclophanes were prepared by the same
procedure from dialdehydes 5b and 5c, respectively.
8,11-Diaza-4,14-dioxa-2(4,4)-N-acetylpiper idin a-1,3(1,4),-
6,12(1,3)-tetr a ben zen a -9(1,2)-tr a n s-cycloh exa n a cyclotet-
r a d eca p h a n e (Meta Host). White foam, yield 79%, mp 95-
97 °C. 1H NMR (CDCl3): 0.7-1.0 (4H, m), 1.0-1.3 (2H, m),
1.4-1.9 (4H, m), 2.06 (3H, s), 2.14 (2H, m), 2.25 (4H, m), 3.45
(2H, m), 3.5-3.7 (4H, m), 3.83 (2H, d, J ) 13.3 Hz), 5.13 (4H,
m), 6.72 (4H, d, J ) 8.5 Hz), 6.93 (4H, d, J ) 8.8 Hz), 7.15-
7.30 (8H, m). 13C NMR: 21.3, 24.8, 31.4, 36.8, 38.6, 51.0, 60.9,
69.8, 114.7, 125.3, 126.6, 127.4, 127.8, 128.6, 137.5, 138.8,
+
141.2, 156.1, 168.8. HR-MS(ES) calcd for C41H48N3O3 [M +
1] 630.3696, found 630.3691.
8,11-Diaza-4,14-dioxa-2(4,4)-N-acetylpiper idin a-1,3(1,4),-
6,12(1,2)-tetr a ben zen a -9(1,2)-tr a n s-cycloh exa n a cyclotet-
r a d eca p h a n e (Or th o Host). White foam, yield 88%, mp 103-
1
105 °C. H NMR (CDCl3): 0.7-1.0 (4H, m), 1.0-1.3 (2H, m),
1.5-1.9 (4H, m), 2.0-2.4 (9H, m), 3.42 (2H, m), 3.5-3.7 (4H,
m), 3.92 (2H, d, J ) 13.3 Hz), 5.02 (4H, s), 6.80 (4H, d, J ) 8.5
Hz), 7.06 (4H, d, J ) 8.8 Hz), 7.2-7.5 (8H, m). 13C NMR: 21.4,
24.9, 31.5, 36.0, 36.9, 38.6, 53.4, 61.4, 67.8, 114.6, 127.2, 127.4,
127.9, 128.2, 129.3, 135.0, 138.8, 139.1, 156.9, 168.8. HR-MS-
(ES) calcd for C41H48N3O3+ [M + 1] 630.3696, found 630.3732.
N,N′-Bis(4-m et h ylb en zyl)-cycloh exa n e-1,2-tr a n s-d ia -
m in e (7). trans-1,2-Cyclohexanediamine 6 (0.50 g, 4.4 mmol)
was dissolved in 5 mL of MeOH, and p-tolualdehyde (1.1 mL,
9.0 mmol), also dissolved in 5 mL of methanol, was added. A
white precipitate was produced, and the mixture was stirred
for 1 h. After the liquid phase was decanted off, the precipitate
was dissolved in 10 mL of an EtOH/CHCl3 (50:50) solution.
NaBH4 (0.32 g, 9.0 mmol) was added, and the reaction mixture
was stirred for 1 h. After the solvents were evaporated under
reduced pressure, the residue was taken up in 50 mL of ether
and extracted with 1 M NaOH (2 × 50 mL). The ether layer
was dried over pellets of NaOH and evaporated to give a yellow
solid (1.16 g, 82% yield), mp 62-64 °C. 1H NMR (CDCl3): 1.04
(2H, m), 1.22 (2H, m), 1.71 (2H, m), 1.80 (2H, m), 2.1-2.3 (4H,
m), 2.33 (6H, s), 3.60 (2H, d, J ) 13.0 Hz), 3.86 (2H, d, J )
13.3 Hz), 7.10 (4H, d, J ) 7.8 Hz), 7.20 (4H, d, J ) 8.0 Hz).
The meta and ortho dialdehydes 5b,c were prepared by the
same procedure from dialcohols 4b and 4c, respectively.
1-{4,4-Bis[4-(3-for m ylben zyloxy)-p h en yl]-p ip er id in -1-
yl}-eth a n on e (5b). Off-white solid, yield 99%, mp 137-139
°C. 1H NMR (CDCl3): 2.08 (3H, s), 2.35 (4H, m), 3.49 (2H, m),
3.65 (2H, m), 5.10 (4H, s), 6.91 (4H, d, J ) 8.8 Hz), 7.15 (4H,
d, J ) 8.5 Hz), 7.56 (2H, t, J ) 7.5 Hz), 7.70 (2H, d, J ) 7.5
Hz), 7.94 (2H, s), 10.04 (2H, s). 13C NMR: 21.1, 35.7, 36.6,

Synthetic Hosts That Model Cation-π Sites
J . Org. Chem., Vol. 67, No. 7, 2002 2167
13C NMR: 20.9, 24.9, 31.4, 50.4, 60.7, 127.9, 128.8, 136.0,
137.9. HR-MS(ES) calcd for C22H31N2 [M + 1] 323.2487, found
323.2497.
P r ep a r a tion of Host’s Dia m in e Dih yd r och lor id es. Host
diamines (1 mmol) were dissolved in 10 mL of chloroform, and
1 mL of 4 M HCl in dioxane was added. The solvents were
evaporated under reduced pressure to give the hosts as white
solids.
Phe(I) was combined with hosts in their energy minimized
structures, obtained from the Monte Carlo simulations, at
various positions and distances from the host’s combining site.
The energy of the complexes was minimized using a conjugate
gradient method (Polak-Ribiere) to a rms gradient of 0.01 kcal/
(mol‚Å) (molecular mechanics calculations, MM+ force field).
The structure of the para host did not change significantly,
whereas new binding conformations were obtained for the
meta and ortho hosts. Monte Carlo docking simulations were
performed starting from these energy minimized complexes
(molecular mechanics calculations, MM+ force field, 1000
structures were analyzed). Low energy structures were chosen
and a final round of energy minimized calculations were
performed to give the final energy minimized complexes.
P oten tiom etr ic Titr a tion s. For titrations in DMSO solu-
tions, the electrode was calibrated using standard solutions
of 2,6-dinitrophenol and 4-chloro-2,6-dinitrophenol, according
to literature procedures.23 Solutions (5 mM) of the templates
were prepared in freshly distilled DMSO (over CaH2) and
titrated under Ar with a 0.10 M solution of Bu4NOH in DMSO.
Similar titrations were performed for solutions containing
templates, DMSO, and boiled water (95:5 v/v). The pKa values
of the templates were calculated using the BEST program.24
Mon itor in g Associa tion . Association of hosts with the
sodium salt of N-acetyl L-phenylalanine was detected by
monitoring changes in the chemical shifts of the guest’s R-H
that occurred with the increasing concentration of a host.
Association constants were derived by performing a nonlinear
least-squares fitting procedure of plots that correlated the
changes in the R-H of Phe(I) against the changes in host
concentrations.22
Molecu la r Mod elin g. Global minimum structures of the
hosts were determined through a Monte Carlo search method
using the MMFF94 force field (package procedure of Spartan-
Pro, Wavefunction, Inc. Irvine, CA). Generally, over 1500
structures were created and analyzed for potential energy
through a simulated annealing process starting at 5000 K and
ending at 300 K. The lowest energy conformers were greater
than 1 kcal/mol lower in energy than the next stable structure.
To obtain electrostatic charges of the hosts in a reasonable
amount of time, the cyclohexyl and N-acetylpiperidine rings
were replaced by hydrogen atoms (see Figure 1 for the struc-
tures used in these calculations). These structures were not
altered before being subjected to single point energy calcula-
tions using the density functional method (pBP/DN* as
presented by SpartanPro). Similar calculations were performed
on the diaminoethyl portion of the hosts.
Ack n ow led gm en t. The authors would like to thank
the University of Cincinnati for their generous funding
of this project. V.D. was supported by a University of
Cincinnati Research Council Fellowship.
Su p p or tin g In for m a tion Ava ila ble: 2D COSY and NOE-
SY spectra of the hosts, energy minimized host-Phe(I) com-
plexes, and potentiometric plots. This material is available free
of charge via the Internet at http://pubs.acs.org.
Docking interactions were determined by first obtaining the
energy minimized structure of Phe(I) using a conjugate gradi-
ent method (Polak-Ribiere) to a rms gradient of 0.001 kcal/
(mol‚Å) (molecular mechanics calculations, MM+ force field
as presented by HyperChem, Hypercube, Inc. Gainesville, FA).
J O011124C
(23) Kolthoff, I. M.; Chantooni, M. K., J r.; Bhowmik, S. J . Am. Chem.
Soc. 1968, 90, 23-28.
(24) Martell, A. E.; Motekaitis, R. J . The Determination and Use of
Stability Constants; VCH: New York, 1988.
(22) Conners, K. A. Binding Constants, The Measurement of Molec-
ular Complex Stability; Wiley: New York, 1987.