Dual binding mode of methylmethanetriacetic acid to tripodal amidopyridine receptors

Article abstract of DOI:10.1021/jo025787l

A series of tripodal amidopyridine receptors capable of selective recognition of methylmethanetriacetic acid (MMTA) in organic solvents is described. Intramolecular hydrogen-bonding groups, built into some of the receptors, were designed as preorganization devices. Binding was studied by NMR titration, variable temperature NMR experiments, 2D-NMR, isothermal titration calorimetry, and single-crystal X-ray crystallography. The results reveal that a balancing act between inter- and intramolecular hydrogen-bonding interactions in the complexes governs both the dynamics and the geometry of binding. Receptor 1b (without intramolecular hydrogen-bonding groups) features a simple symmetric MMTA binding geometry with optimal enthalpic interactions. In sharp contrast, receptor 1a (with intramolecular hydrogen-bonding groups) reveals a temperature-dependent dual binding mode where MMTA can bind in two completely different geometries. The two solution binding geometries of 1a·MMTA were unraveled by NMR experiments and correlated to the X-ray structures.

Full text of DOI:10.1021/jo025787l

Du a l Bin d in g Mod e of Meth ylm eth a n etr ia cetic Acid to Tr ip od a l
Am id op yr id in e Recep tor s
Pablo Ballester,*,† Magdalena Capo´,† Antoni Costa,† Pere M. Deya`,† Rosa Gomila,†
Andreas Decken,‡ and Ghislain Deslongchamps‡
Departament de Quı´mica, Universitat de les Illes Balears, 07071 Palma de Mallorca, Spain, and
Department of Chemistry, University of New Brunswick, Fredericton, New Brunswick, E3B 6E2, Canada
pablo.ballester@uib.es
Received April 2, 2002
A series of tripodal amidopyridine receptors capable of selective recognition of methylmethanetri-
acetic acid (MMTA) in organic solvents is described. Intramolecular hydrogen-bonding groups, built
into some of the receptors, were designed as preorganization devices. Binding was studied by NMR
titration, variable temperature NMR experiments, 2D-NMR, isothermal titration calorimetry, and
single-crystal X-ray crystallography. The results reveal that a balancing act between inter- and
intramolecular hydrogen-bonding interactions in the complexes governs both the dynamics and
the geometry of binding. Receptor 1b (without intramolecular hydrogen-bonding groups) features
a simple symmetric MMTA binding geometry with optimal enthalpic interactions. In sharp contrast,
receptor 1a (with intramolecular hydrogen-bonding groups) reveals a temperature-dependent dual
binding mode where MMTA can bind in two completely different geometries. The two solution
binding geometries of 1a ‚MMTA were unraveled by NMR experiments and correlated to the X-ray
structures.
In tr od u ction
accounting of the interplay between inter- and intra-
molecular forces.4
Intermolecular interactions are the fundamental basis
for the formation of supermolecules.1 They occur between
atoms that are not directly linked by a covalent frame-
work. In nonpolar organic solvents, the so-called “strong”
or “conventional” hydrogen bond2 (O,N)-H‚‚‚(O,N), an
electrostatic force with some covalent character, is prob-
ably the most important directional intermolecular force
used for supramolecular construction. The design of
abiotic receptors possessing high affinity and selectivity
and based primarily on complementary intermolecular
hydrogen-bonding arrays depends wholly on the optimi-
zation of these intermolecular interactions. This is typi-
cally achieved by tailoring the host molecular structure
to that of the guest.3 While the broad features of the
supermolecule structure may be appreciated by simple
consideration of optimal intermolecular hydrogen-bond-
ing interactions and their anisotropic properties, an in-
depth understanding of the design may require a detailed
In particular, we have been interested in the effect of
remote intramolecular hydrogen bonding, used as a
receptor preorganization device, on the detailed thermo-
dynamics of binding and on the geometry of the complex.
Initial studies focused on the molecular recognition of cis-
1,3,5-cyclohexanetricarboxylic acid (CTA) 2 by triamido-
pyridine receptors 1a and 1b, in which the three inter-
action units are separated by a 1,3,5-triarylbenzene
molecular scaffold. Receptor 1a features one intra-
molecular hydrogen bond between the phenolic hydrogen
and the amide carbonyl on each arm, restricting the
receptor conformation and flexibility. These interactions
were eliminated in receptor 1b by removing the hydroxyl
groups.
The structural modifications are remote with respect
to the binding site and do not seem likely to affect the
binding properties of the receptor due to enthalpic,5
steric, or solvation effects.
On one hand, our binding studies on receptors 1a and
1b with CTA indicate that only one of two possible
binding geometries is observed in solution6,7 and in the
crystalline state.7,8 The preferred geometry for the bound
* To whom correspondence should be addressed. Current address:
The Skaggs Institute for Chemical Biology, The Scripps Research
Institute, MB-26, 10550 North Torrey Pines Road, La J olla, CA 92037,
on sabbatical leave from UIB. E-mail: pauba@scripps.edu.
† Universitat de les Illes Balears.
‡ University of New Brunswick.
(1) Atwood, J . L.; Davies, J . E. D.; Macnicol, D. D.; Vo¨tgle, F., Eds.
Comprehensive Supramolecular Chemistry; Pergamon: 1996.
(2) J effrey, G. A. An Introduction to Hydrogen Bonding; Oxford
University Press: New York, 1997.
(3) Examples of abiotic receptors based on hydrogen bonding: (a)
Hamilton, A. D.; Springer-Verlag: Berlin, 1991; Vol. 2, pp 115-174.
(b) Rebek, J ., J r. Angew. Chem., Int. Ed. Engl. 1990, 29, 245-255. (c)
Webb, T. H.; Wilcox, C. S. Chem. Soc. Rev. 1993, 22, 383-395. (d) Bell,
T. W.; Hext, N. M.; Khasanov, A. B. Pure Appl. Chem. 1998, 70, 2371-
2377.
(4) (a) Adams, H.; Carver, F. J .; Hunter, C. A.; Osborne, N. J . Chem.
Commun. 1996, 2529-2530. (b) Rebek, J ., J r.; Williams, K.; Parris,
K.; Ballester, P.; J eong, K.-S. Angew. Chem., Int. Ed. Engl. 1987, 26,
1244-1245. (c) Muehldorf, A. V.; Vanengen, D.; Warner, J . C.;
Hamilton, A. D. J . Am. Chem. Soc. 1988, 110, 6561-6562. (d) Low, J .
N.; Storey, E. J .; McCarron, M.; Wardell, J . L.; Ferguson, G.; Glidewell,
C. Acta Crystallogr. 2000, B56, 58-67.
(5) Rebek, J ., J r. Acc. Chem. Res. 1990, 23, 399-404.
(6) Ballester, P.; Costa, A.; Deya`, P. M.; Gonzalez, J . F.; Rotger, M.
C. Tetrahedron Lett. 1994, 35, 3813-3816.
10.1021/jo025787l CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/16/2002
8832
J . Org. Chem. 2002, 67, 8832-8841

Dual Binding Mode of Methylmethanetriacetic Acid
different binding behaviors. Indeed, while receptor 1a
binds MMTA in a single geometry, receptor 1b features
a dual binding mode. The dynamic process in which
complex 1b‚3 is involved was studied by variable tem-
perature (VT) 1H NMR spectroscopy. As will be apparent
from the following section, at low temperature, the
favored binding geometry results from a maximization
of intermolecular interactions, while at room temperature
it is the maximization of intramolecular interactions that
appears to govern the preferred geometry. Thus, the
relative distribution of the geometric complexes for 1b‚3
is controlled by the interplay of inter- and intramolecular
interactions, which in turn is regulated by temperature.
Resu lts a n d Discu ssion
Design a n d Syn th esis. Some time ago, some of us
reported10 the use of a tetrachlorosilane-ethanol-induced
self-condensation of ketones11 for preparing 1,3,5-tris(3-
bromo-4-hydroxy-5-carboxyphenyl)benzene. With this fac-
ile synthetic methodology, one can obtain 1,3,5-triaryl-
substituted benzenes in multigram quantities. Molecular
modeling studies12 on the corresponding debrominated
2-amidopyridine derivative 1 suggested that the syn
conformation of the molecule was very well suited for the
complexation of CTA 2 (Figure 1) and MMTA 3 (Figure
4). In this conformation, the hydrogen-bonding groups
converge and act simultaneously in “grasping” the guest
molecule in the induced cleft of the receptor. It is worth
noting that, in the original design of the host, intra-
molecular hydrogen bonding between the amide carbo-
nyls and the phenol protons was intended simply as a
preorganization tool.13 The other possible conformer, in
which the amide protons are hydrogen-bonded to the
phenolic oxygens, was dismissed on thermodynamic
grounds.14 Thus, the favored intramolecular hydrogen
bonds restrict the rotation of the Caryl-Camide single bonds,
forcing the amide hydrogens to point toward the interior
of the cavity, locking each binding subunit in a conforma-
tion that is quasiplanar with each aryl ring. The evalu-
ation of the goodness of the designed intramolecular
hydrogen bonds as a preorganization tool required the
synthesis of the analogous deoxygenated receptor 1b.
The synthesis of host 1 was straightforward and is
outlined in Scheme 1. Receptor 1 was prepared from
triester 5, which was obtained in 78% yield through
tetrachlorosilane-ethanol-induced self-condensation of
4-acetylmethylsalicylate15 4. Next, the phenol hydroxyls
of triester 5 were protected as benzyl ethers to afford the
tribenzylated triester 6,14 a step that was found necessary
in order to cleanly obtain the triacyl chloride in a later
step. The tribenzylated triester was hydrolyzed under
F IGURE 1. Preferred binding geometry of 1a or 1b and 2.
guest features the three axial hydrogens R to the carboxyl
groups pointing outside the receptor cavity (Figure 1, HR).
On the other hand, receptor 1b, which lacks the intramo-
lecular hydrogen bonds, exhibits an enthalpy for binding
CTA that is ∼ 3 kcal/mol higher than that measured for
intramolecularly hydrogen-bonded receptor 1a . On the
basis of Koshland’s induced-fit hypothesis,9 the optimum
host conformation for binding CTA could be inferred from
the X-ray complex structure for 1b‚2,7 which reveals an
average value of 23° (0.97) for the dihedral angle between
the carboxamide and aryl planes. In contrast, that same
dihedral angle is only 10° (2.3) for the 1a ‚28 complex.
Thus, host 1a cannot achieve the optimum binding
conformation due to conformational restrictions imposed
by the intramolecular phenolic hydrogen bonds. We
concluded that the preferred binding geometry was
derived from a maximization of the intermolecular
interactions, while the difference in the enthalpy of
binding was a consequence of a balance between the
inter- and intramolecular hydrogen-bonding interactions.
Herein we report a complementary study focused on
the molecular recognition of methylmethanetriacetic acid
(MMTA, 3-carboxymethyl-3-methylpentanedioic acid, 3)
by receptors 1a and 1b, complete with full experimental
details of the receptor syntheses.
(10) Rotger, M. C.; Costa, A.; Saa´, J . M. J . Org. Chem. 1993, 58,
4083-4087.
(11) Elmorsy, S. S.; Pelter, A.; Smith, K.; Hursthouse, M. B.; Ando,
D. Tetrahedron Lett. 1992, 33, 821-824.
In the case of MMTA recognition, the balancing of
inter- and intramolecular hydrogen-bonding interactions
was not only reflected in a difference of binding enthal-
pies toward receptors 1a and 1b but also in totally
(12) Macromodel V6.0: Mohamadi, F.; Richards, N. G. J .; Guida,
W. C.; Liskamp, R.; Lipton, M.; Caulfield, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J . Comput. Chem. 1990, 11, 440.
(13) (a) Cram, D. J . Angew. Chem., Int. Ed. Engl. 1986, 25, 1039-
1058. Examples of conformational restriction by intramolecular hy-
drogen bonding: (b) Lo´pez de la Paz, M.; Ellis, G.; Penade´s, S.; Vicent,
C. Tetrahedron Lett. 1997, 38, 1659-1662. (c) Kohmoto, S.; Koyano,
I.; Kawatsuji, T.; Kasimura, H.; Kishikawa, K.; Yamamoto, M.;
Yamada, K. Bull. Chem. Soc. J pn. 1996, 69, 3261-3265.
(14) Garrett, T. M.; Cass, M. E.; Raymond, K. N. J . Coord. Chem.
1992, 25, 241-253.
(7) Ballester, P.; Capo´, M.; Costa, A.; Deya`, P. M.; Gomila, R.;
Decken, A.; Deslongchamps, G. Org. Lett. 2001, 3, 267-270.
(8) Ballester, P.; Costa, A.; Deya`, P. M.; Deslongchamps, G.; Mink,
D.; Decken, A.; Prohens, R.; Toma`s, S.; Vega, M. J . Chem. Soc., Chem.
Commun. 1997, 357-358.
(9) Koshland, D. E. Proc. Natl. Acad. Sci. 1958, 44, 98.
J . Org. Chem, Vol. 67, No. 25, 2002 8833

Ballester et al.
F IGURE 2. 1H NMR spectra of complexes (a) 1a ‚3 and (b) 1b‚3 in CDCl3 at 298 K. See Figure 4 for protons assignments.
F IGURE 3. Expanded regions of the 2D-ROESY spectra of (a) complex 1a ‚3 and (b) complex 1b‚3 in CDCl3 at 298 K.
basic conditions to afford, after acidification, the tribenz-
ylated triacid 7 in 96% yield. Triacid 7 was converted
almost quantitatively to the corresponding triacid
chloride 8 by treatment with thionyl chloride. Next, the
triacid chloride 8 was coupled with 2-amino-6-methyl-
pyridine, yielding triamide 9. Finally, the benzyl groups
were removed using TMSI or HBr, affording receptor 1
(38% overall yield from 4, Scheme 1) as a white solid,
insoluble in CHCl3 or other organic solvents that cannot
disrupt hydrogen bonding.
To increase the chloroform solubility of the receptor, a
more lipophilic version of the triaryl spacer was devised.
The synthesis involved the etherification of the phenols
of 5 with allyl bromide16 to obtain compound 10, followed
by a triple Claisen17 rearrangement, yielding 11, and
finally catalytic hydrogenation of the terminal double
bonds to afford the propylated triaryl benzene triester
5a in 61% overall yield (Scheme 2).
trimethylaluminum complex of 2-aminopicoline for 48 h,18
resulting in a 26% yield of host 1a after column chro-
matography (Scheme 2). Chloroform solutions of receptor
1a of up to 10-2 M could be easily prepared.
Host 1b was also prepared from triester 5a through a
more elaborated route. First, the phenols of 5a were
removed by converting triester 5a to the corresponding
tris aryl triflate 12, followed by ring deoxygenation
19
mediated by Cl2Pd(PPh3)2 to afford triester 5b in 63%
overall yield. Hydrolysis of triester 5b afforded triacid
14, which was converted to the triacyl chloride 15 on
treatment with thionyl chloride. Finally, 15 was coupled
with 6-methyl-2-aminopyridine to produce tripodal recep-
tor 1b (38% overall yield, from 5a , Scheme 2). Receptor
1b was also quite soluble in chloroform.
Bin d in g Stu d ies. Addition of solid MMTA 3 to a
room-temperature CDCl3 solution of either receptor 1a
or 1b led to rapid dissolution of the otherwise insoluble
substrate. In both cases, large downfield shifts (∼3 ppm)
were observed for the NH resonances, indicating the
formation of hydrogen-bonded complexes, which integra-
tion established as having 1:1 stoichiometry. Dilution of
a 5 mM CDCl3 solution of each pair of components to 0.5
mM shows negligible changes in the chemical shifts,
suggesting a Ka > 1 × 104 M-1 for hosts 1a and 1b with
triacid 3 in chloroform.
Receptors 1a and 1b were prepared uneventfully using
lipophilic spacer 5a . For the preparation of 1a , a benzene
solution of 5a (previously treated with 1.5 equiv of
trimethylaluminum) was refluxed with 1.5 equiv of the
(15) Originally the triester 1b was prepared through the tetrachloro-
silane-ethanol-induced self-condensation of 4-acetylsalicylic acid (see
ref 10), followed by esterification with diazomethane. In the course of
this work, we found that when the self-condensation reaction is carried
out on 4-acetylmethylsalicylate, the triester 1b is obtained in just one
synthetic step in a higher yield and with a simpler workup.
(16) Winters, R. T.; Sercel, A. D.; Showalter, H. D. H. Synthesis 1988,
712-714.
(18) Basha, A.; Lipton, M.; Weinreb, S. M. Tetrahedron Lett. 1977,
4171-4173.
(19) (a) Saa´, J . M.; Dopico, M.; Martorell, G.; Garcı´a-Raso, A. J . Org.
Chem. 1990, 55, 991-995. (b) Cacchi, S.; Ciattini, P. G.; Morera, E.;
Ortar, G. Tetrahedron Lett. 1986, 27, 5541-5544.
(17) Danishefsky, S.; Berman, E. M.; Ciufolini, M.; Etheredge, S.
J .; Segmuller, B. E. J . Am. Chem. Soc. 1985, 107, 3891-3898.
8834 J . Org. Chem., Vol. 67, No. 25, 2002

Dual Binding Mode of Methylmethanetriacetic Acid
SCHEME 1a Syn th esis of Recep tor 1
a
Reagents and conditions: (a) SiCl4/ EtOH, rt 14 h; (b) Cs2CO3,
DMF/acetone, BrCH2Ph, reflux, 24 h; (c) NaOH/THF, reflux, 20
h; HCl, 5%; (d) SOCl2/CH2Cl2, reflux, 2 h; (e) 2-amino-6-meth-
ylpyridine, rt, 4 h; (f) Me3SiI/CH2Cl2, rt, 72 h or HBr/CHCl3, rt,
10 min.
F IGURE 4. Possible binding geometries for 1:1 complexes 1a ‚
3 and 1b‚3. The terms endo and exo refer to the orientation of
the methyl group of 3 with respect to the receptor cavity.
1
The H NMR spectra for the complexes 1a ‚3 and 1b‚3
the receptor and the methyl and methylene triacid
signals was 0.79 (cross-peak volume CH3(3)-Ha/cross-
peak volume CH2(3)-Ha). The corresponding ratio for the
Hb proton was 0.53. Similarly, for complex 1b‚3, the
calculated ratios yielded values of 1.27 and 0.93, respec-
tively. In other words, host protons Ha and Hb in complex
1a ‚3 have more intense intermolecular cross-peaks with
the methylene than with the methyl group of 3. For
complex 1b‚3, the opposite is observed for the host Ha
proton, while for the Hb proton, similar volumes are
observed.
Molecular modeling studies suggest two possible C3-
symmetric geometries for a complex between 3 and 1a
or 1b in which six intermolecular hydrogen bonds be-
tween the three COOH residues of the triacid and the
three CONH(py)N groups of the receptor can be estab-
lished (Figure 4). In geometry A, the methyl group of 2
is endo to the receptor cavity, while in geometry B, it is
exo. Both geometries were optimized with MacroModel
6.012 using the AMBER* force field and the GB/SA CHCl3
solvation model, resulting in an energy difference of ∆E
) ∼ 3 kcal/mol in favor of binding geometry A for the
two complexes.
are shown in Figure 2. Similar chemical shifts are
observed for several of the equivalent protons in the two
different complexes, such as, the propyl chain protons and
some of the aromatic and aliphatic resonances. However,
the 1H NMR spectrum of complex 1b‚3 (Figure 2b) shows
a significant downfield shift for the singlet assigned to
the central aromatic ring protons (Ha), one of the pyridine
doublets (He), and the amide NH signal when compared
to complex 1a ‚3. Also, a moderate upfield shift is ob-
served for the methyl singlet of triacid 3 in complex 1b‚
3 compared to the same signal for complex 1a ‚3. As a
whole, these observations lead one to surmise that a
different binding mode for each of the complexes was
operative. To explore this hypothesis, 2D-ROESY20 spec-
tra (mixing time ) 500 ms, spinlock ) 2 kHz) were
acquired at room temperature for each of the complexes
(Figure 3).
Figure 3a shows an expansion of the 2D-ROESY
spectrum obtained for complex 1a ‚3. Intermolecular
contacts between the protons of the receptor and the
triacid were observed. Indeed, correlation peaks between
the methyl and methylene signals of MMTA and the Ha
and Hb protons of the receptor are clearly evident. The
2D-ROESY for complex 1b‚3 (Figure 3b) is identical in
terms of observed cross-peaks. A difference arises when
the ratio of integral volumes for the above-mentioned
cross-peaks is calculated. For complex 1a ‚3, the ratio of
integrals for the cross-peaks between the Ha proton of
On the basis of the complexes obtained by molecular
modeling, the A geometry (methyl group endo) should
yield intermolecular contacts between the Ha and Hb
protons on the receptor and the methyl and methylene
protons of the triacid. Conversely, for the B geometry
(methyl group exo), one should expect only intermolecular
contacts between the above-mentioned receptor protons
and the methylene proton of triacid 3.
(20) Bothner-By, A. A.; Stephens, R. L.; Lee, J .-M. J . Am. Chem.
Soc. 1984, 106, 811.
J . Org. Chem, Vol. 67, No. 25, 2002 8835

Ballester et al.
SCHEME 2a Syn th esis of Recep tor s 1a a n d 1b
1
F IGURE 5. Downfield region of the VT H NMR spectra for
(a) complex 1a ‚3 and (b) complex 1b‚3.
F IGURE 6. Upfield region of the VT 1H NMR spectra for
complex 1a ‚3.
downfield region corresponding to the acidic protons (NH
and phenolic OH of complex 1a ‚3, NH of complex 1b‚3).
For complex 1a ‚3, two sets of signals for each NH and
OH proton are clearly seen at low temperature (230 K).
These sets of signals had arisen through a decoalescence
process from the two signals observed at room temper-
ature and were assigned to the amide NH (∼10.98 ppm)
and phenolic OH (∼11.79 ppm). Both sets of signals have
chemical shifts that can be assigned to hydrogen-bonded
protons. The observed temperature dependence of the
system provides clear evidence for an equilibrium that
becomes slow on the NMR time scale at around 240 K,
allowing for the simultaneous observation of both of its
components below that temperature. The relative popu-
lation of the equilibrium components below decoalescence
can be measured by integration of the signals of the
different sets and was determined to be a 3:1 ratio at
220 K.
The remaining task was to determine the geometries
of the major and minor components of the observed
equilibrium. Close inspection of the region corresponding
to the methyl group of triacid 3 in the VT 1H NMR
spectra revealed some important clues (Figure 6).
During the cooling process, the methyl group of 3,
which appeared initially as a singlet at 0.86 ppm,
decoalescences into two signals clearly observable at 220
K. The major signal appears at 0.6 ppm, while the minor
is at 1.2 ppm. In the endo geometry (A), the methyl group
of 3 is directed toward the central aromatic ring of the
receptor and is likely to experience an upfield shift under
the influence of the magnetic anisotropy of the aromatic
a
Reagents and conditions: (a) Cs2CO3, DMF/acetone, Br-CH2-
CHdCH2, reflux, 24 h; (b) N,N-dimethylaniline, reflux, 12 h; (c)
H2, 10% Pd/C, ethyl acetate, 10 psig, 5 h; (d) 4.5 equiv Me3Al, then
Me3Al/2-amino-6-methylpyridine complex, benzene, reflux, 48 h;
(e) Tf2O, CH2Cl2; (f) Cl2Pd(PPh3)2, Bu3N, HCO2H/DMF, 90 °C; (g)
NaOH, H2O/THF, then H+; (h) SOCl2/CH2Cl2; (i) 6-methyl-2-
aminopyridine, CH2Cl2.
First, the observed ROEs between the methyl group
of triacid 3 and the intracavity protons of the receptors
provide strong experimental evidence for solution geom-
etry A, but they do not rule out the presence of geometry
B as well. Second, due to the considerably different Ha/
Hb-methyl/methylene ROESY cross-peak volume ratios
and the unequal chemical shift for the triacid methyl
group for each complex, one can speculate on the exist-
ence of a solution equilibrium that involves different
populations of geometries A and B. The relative amounts
of the solution populations would be dictated by receptor
structures that differ only on the phenolic OH. The
interplay of intra- and intermolecular forces could be a
possible cause for the different geometry populations.
1
VT H NMR experiments were carried out in order to
detect the existence of different complex geometries in
solution. Lowering the temperature of the 1H NMR
experiment should slow the equilibration process, if one
exists. If the temperature passes below the decoalescence
point, the simultaneous observation of different H NMR
signals for the two complex geometries is possible. Figure
1
1
5 shows two VT H NMR experiments, centered on the
8836 J . Org. Chem., Vol. 67, No. 25, 2002

Dual Binding Mode of Methylmethanetriacetic Acid
TABLE 1. Isoth er m a l Titr a tion Ca lor im etr y for Hosts
1a a n d 1b w ith 3 (MMTA) in 20% THF /CHCl3, a t 294 K
ring, compared to the same methyl group in the exo
geometry. Also, 1D NOE difference experiments carried
out at 220 K reveal intense negative NOEs for the
intracavity protons of the receptor upon irradiation at
0.6 ppm (upfield methyl signal of 3). Thus, the major
component of the equilibrium at 220 K can be assigned
to the endo geometry (A) of complex 1a ‚3, while the minor
component can be assigned the exo geometry (B).
host
Ka (M-1
)
∆H (kcal/mol)
1a
1b
10185 ( 350
6690 ( 200
-5.8 ( 0.2
-6.5 ( 0.4
promising the solubility of the host and guest. Accord-
ingly, hosts 1a and 1b were titrated by ITC as ≈1.5-2
mM solutions with a solution of 3 (12-17 mM), and the
results are reported in Table 1. The enthalpy results
clearly show that host 1b, lacking the intramolecular
hydrogen bond (phenolic hydroxyl to amide carbonyl),
promotes more optimal H-bonding interactions with 3
than host 1a . The conformational changes induced upon
complexation in host 1a reduce the exothermicity of
association by ≈0.7 kcal/mol compared to 1b, correspond-
ing to the gain in energy due to the distortion of the intra-
and intermolecular hydrogen-bonding networks of com-
plex 1a ‚3. The more flexible host 1b can achieve better
complementarity with 3 by maximizing the H-bonding
interactions at the expense of a higher entropy loss.23 The
result is a lower value for its association constant.
Attempts to obtain X-ray crystal structures for com-
plexes 1a ‚3 and 1b‚3 were successful. Slow diffusion of
cyclohexane into 1:1 chloroform solutions of either 1a ‚3
or 1b‚3 at room temperature produced colorless plates
that were relatively unstable once removed from the
mother liquor. For this reason, they were mounted on
Lindeman capillaries prior to the dataset acquisition.
Interestingly, the X-ray structure of complex 1a ‚3 reveals
an unique host-guest geometry (Figure 7).
As expected, triacid 3 is docked into the receptor via
six intermolecular hydrogen bonds. However, the struc-
ture shows a complex geometry that is related to the exo
geometry proposed previously, based on the NMR data
and modeling studies, except that it does not possess the
idealized C3-symmetry. Two of the guest carboxylic acids
are almost coplanar with their respective amidopyridine
binding unit, while the third acid is almost perpendicular
to its amidopyridine partner. The O‚‚‚N distances for the
former NH‚‚‚O interactions are shorter (2.8 Å) than for
the latter (3.1 Å), while the O‚‚‚N distances for the N‚‚‚
HO hydrogen bonds are the same for both types of
hydrogen-bonded pairs (3.1 Å). The methyl group of the
triacid does point away from the receptor cavity (exo).
The dihedral angle about the amide-aryl planes is
different for each arm (5°, 11°, and 13°). These values
represent a distortion from the planarity required for the
intramolecular phenolic hydrogen bond. The crystal data
are in line with the hypothesis that, at room temperature,
the exo/endo equilibrium for the geometries of complex
1a ‚3 is biased toward the exo geometry, since the latter
is the adopted geometry for crystallization of the complex
at room temperature. The lack of C3-symmetry of the (1a ‚
3)A structure may be attributed to packing preferences
As for the room-temperature populations, the chemical
shifts for the NH and the OH of the receptor and the
methyl signal of 3 are much closer to the values ascribed
to the minor geometry (i.e. exo) at 220 K. Furthermore,
as discussed above, the room temperature 2D-ROESY
experiment reveals more intense correlation peaks be-
tween the intracavity protons of the receptor (Ha, Hb) and
the methylene signal of 3 than with the methyl signal of
3. These observations suggest a change of the relative
populations of the complex geometries from 220 K to
room temperature, that is, an increase of the relative
amount of the exo geometry upon warming. A likely
explanation is that the temperature regulates the inter-
play between the inter- and intramolecular forces, which
in turn controls the binding mode. The maximization of
the intermolecular hydrogen bonds established in the
endo geometry requires some conformational distortion
and, therefore, a weakening of the intramolecular hy-
drogen bonds (vide infra). This maximization can only
be achieved at low temperatures, where intermolecular
hydrogen bonds become stronger due to the loss of
vibrational motion in the complex. Analysis of the 1H
NMR signals of the hydrogen-bonded protons at 220 K
(Figure 5a) provides further experimental support for this
hypothesis. Indeed, the phenol proton for the major
isomer (endo) is shifted upfield by 0.3 ppm relative to
that of the minor isomer, a sign of weaker intramolecular
hydrogen bonding. Furthermore, the amide proton for the
major isomer is shifted downfield by 0.3 ppm compared
to the same signal for the minor isomer, another sign of
strengthened intermolecular hydrogen bonding.
Similar VT 1H NMR experiments performed with
complex 1b‚3 gave a completely different picture. First
of all, the amide NHs of 1b do not show any sign of
decoalescence in the same range of temperatures used
for the study of complex 1a ‚3 (Figure 5b). The only
temperature-dependent phenomenon observed for the
amide NH protons of 1b and the methyl group protons
of 3 is a slight chemical shift change (downfield shift for
the NH and upfield shift for the methyl), which can be
simply ascribed to a tightening of the complex at lower
temperature. These observations provide clear evidence
for the nonexistence of equilibrating geometries for this
host-guest system. Thus, for complex 1b‚3, one geometry
is highly predominant in solution within a wide range of
temperatures. Again, through 1D NOE difference experi-
ments at 220 K and by comparing the chemical shift of
the methyl group of 3 at that temperature to that of the
endo geometry of complex 1a ‚3, the geometry of complex
1b‚3 was assigned to be almost exclusively endo.
(21) (a) Wadso¨, I. Chem. Soc. Rev. 1997, 26, 79-86. (b) Freire, E.;
Mayorga, O. L.; Straume, M. Anal. Chem. 1990, 62, 950-959. (c)
J elesarov, I.; Bosshard, H. R. J . Mol. Recognit. 1999, 12, 3-18.
(22) The “c” parameter (c ) KaMtotn, Mtot ) total host concentration,
n ) number of binding sites) should be between 5 and 500 for the
accurate simultaneous determination of ∆H, Ka, and n. Micro Calo-
rimetry System User’s Manual, MicroCal ITC manual, pp 46-47.
(23) Searle, M. K.; Westwell, M. S.; Williams, D. H. J . Chem. Soc.,
Perkin Trans. 2 1995, 141-151.
The thermodynamic binding data for systems 1a ‚3 and
1b‚3 were determined using isothermal titration calo-
rimetry (ITC)21 in a solvent mixture of 20% THF/CHCl3.
The host concentrations were adjusted to return a “c”22
value approximately equal to 9, high enough for a reliable
calorimetric determination of Ka and ∆H without com-
J . Org. Chem, Vol. 67, No. 25, 2002 8837

Ballester et al.
F IGURE 7. X-ray structures of (a) complex 1a ‚3 and (b) complex 1b‚3. Only polar hydrogens and MTA methyl hydrogens in
calculated positions are shown. Dotted lines indicate inter- and intramolecular hydrogen bonds.
of the crystal and to the presence of disordered solvent
molecules, as was previously observed in the crystal
structure of complex 1a ‚2.8
for binding C3-symmetric triacids, such as 1,3,5-cyclo-
hexanetricarboxylic acid (2) and methylmethanetriacetic
acid (3). We have studied the effect of remote intra-
molecular hydrogen bonding, used to restrict receptor
conformation, on the binding mode of complex formation
with triacid 3 and on the thermodynamic parameters of
the binding process. Molecular modeling studies suggest
two possible C3-symmetric geometries for a complex
between 3 and 1a or 1b in which six intermolecular
hydrogen bonds between the triacid guest and the recep-
Gratifyingly, the X-ray structure of complex 1b‚3 is in
complete agreement with the endo structure proposed
from solution analyses. It features a C3-symmetry axis,
and the methyl group of the triacid 3 points directly
toward the interior of the receptor cavity, perpendicular
to the central aromatic ring. Triacid 3 is docked into the
receptor, establishing an array of six hydrogen bonds.
Only two types of O‚‚‚N distances for the six hydrogen-
bonded interactions are observed, namely 2.6 Å and 2.9
Å for the NH‚‚‚O and N‚‚‚HO hydrogen bonds, respec-
tively. The dihedral angle about the amide-aryl planes
is 30° for all three arms. Since it is known that aromatic
carboxamides prefer to remain coplanar to their aryl ring
(conjugation), the observed tilt of the amides in complex
1b‚3 provides a nice example of Koshland’s induced-fit
hypothesis in an abiotic setting.9
The observed amide-aryl dihedral angle in complex 1b‚
3 has the optimum value required for maximizing the
intermolecular interactions in the endo geometry. This
conformational condition cannot be achieved with recep-
tor 1a on binding triacid 3. In receptor 1a , the intramo-
lecular hydrogen bonds between phenols and amide
carbonyls restrict the rotation of the Caryl-Camide bond and
lock each amidopyridine subunit in an almost coplanar
configuration with respect to their respective aryl ring.
In complex 1a ‚3, competition between intramolecular vs
intermolecular hydrogen bonding results in only a partial
tilt of the amides, such that the exo geometry arises as
the preferred binding conformation instead of the endo
geometry seen in complex 1b‚3.
1
tor can be established. VT H NMR experiments, as well
as 1D and 2D NOE experiments, were carried out to
ascertain the existence of an equilibrium between com-
plex geometries and to map out their structural features.
While the conformationally restricted receptor 1a shows
that its preferred binding geometry is temperature
dependent, receptor 1b binds 3 in a fixed geometry at
any temperature. This difference in binding behavior was
explained in terms of an interplay between intermolecu-
lar and intramolecular interactions. The binding pro-
cesses have been studied by ITC, and again, an important
influence of the remote intramolecular hydrogen bonding
on the thermodynamics of binding was observed. Crystal
structures of complexes 1a ‚3 and 1b‚3 were obtained by
X-ray crystallography. Each complex has a different
binding geometry in the solid state that is in agreement
with its predominant solution structure at room temper-
ature (complex 1a ‚3 has exo geometry, while complex 1b‚
3 has preference for the endo mode). The geometrical
parameters derived from the X-ray data clearly support
a balance between intermolecular and intramolecular
hydrogen-bonding interactions that is responsible for the
different binding modes. On one hand, the remote in-
tramolecular hydrogen bonds present in 1a and ring
conjugation of the carboxamide promote coplanarity of
the amide and aryl planes. On the other hand, maximi-
zation of the intermolecular interactions with the more
Con clu sion
We have reported the synthesis of several tripodal
amidopyridine abiotic receptors that display high affinity
8838 J . Org. Chem., Vol. 67, No. 25, 2002

Dual Binding Mode of Methylmethanetriacetic Acid
4 was prepared by recrystallization from cyclohexane. Anal.
Calcd for C10H10O4: C, 61.85, H, 5.19. Found: C, 61.85, H,
5.23.
flexible receptor 1b forces a tilting of the amides away
from coplanarity. The geometrical requirements of the
intramolecular hydrogen bonds in 1a force a change in
binding mode in order to maximize the intermolecular
interactions. Only at low temperatures is the energy gain
of the strengthened intramolecular interactions capable
of overcoming the loss of stability due to the distorted
intramolecular interactions required to achieve optimal
endo geometry. This model system shows how subtle
effects, acting far away from the binding site, and
temperature changes can switch the preferred binding
geometry of a host-guest complex. Indeed, the interplay
of inter- and intramolecular interactions should always
be taken into consideration whenever predictions of
binding modes are made.
1,3,5-(4-H yd r oxy-3-m et h oxyca r b on ylp h en yl)b en zen e
(5). A 500-mL three-neck round-bottom flask equipped with
an argon inlet, a reflux condenser, a rubber septum and a
magnetic stir bar was charged with 13.5 g (69.6 mmol) of ester
4 and 275 mL of dry ethanol. To the resulting solution, 48 mL
(419 mmol) of SiCl4 were slowly transferred via cannula.
During the transfer, the reaction temperature rose to solvent
reflux and a strong evolution of HCl was observed. The
resulting wine red mixture was allowed to stir at room
temperature for 15 h. then poured into 200 mL of ice water. A
pink solid precipitated and was filtered on a fritted filter
funnel. The collected pink solid was dissolved in CH2Cl2 and
the resulting solution was dried over Na2SO4, filtered and
concentrated by rotary evaporation to afford a solid residue
that was washed with a small amount of diethyl ether to yield
10.7 g (87%) of pure triester 5 as a white solid: mp 222-225
Exp er im en ta l Section
1
°C; IR (KBr) 3300-3600 (broad), 1670, 1210, 1290 cm-1; H
NMR (CDCl3) δ 10.83 (s, 3H), 8.14 (d, J ) 2.4 Hz, 3H), 7.81
(dd, J ) 8.4, 2.4 Hz, 3H), 7.63 (s, 3H), 7.12 (d, J ) 8.4 Hz,
3H), 4.00 (s, 9H); 13C NMR (CDCl3) δ 171.13, 161.9, 141.9,
135.3, 132.8, 129.0, 124.7, 118.8, 113.2, 53.1. An analytical
sample of 5 was prepared by recrystallization from benzene.
Anal. Calcd for C30H24O9: C, 68.18, H, 4.58. Found: C, 68.15,
H, 4.67.
Gen er a l Meth od s a n d Ma ter ia ls. In str u m en ta tion . All
reactions were run under argon in oven-dried glassware. 1H
(300 MHz) and 13C NMR (75 MHz) spectra were recorded in
CDCl3 solutions. Chemical shifts (δ) are reported in ppm and
referenced to the solvent peak.24 Melting points were taken
on a capillary melting point apparatus and are uncorrected.
High-resolution mass spectra were obtained by electron impact
ionization with an energy of 70 eV. Elemental analyses were
performed by Servei de Microa`nalisi Elemental del Centre
d'Investigacio´ i Desenvolupament C. S. I. C. (Barcelona, Spain).
Ma ter ia ls. Preparative flash chromatography was per-
formed on Merck silica gel 60 (40-63 mM). Analytical TLC
was performed on precoated Merck silica gel 60 F-254 plates.
Preparative TLC was performed on 0.5 mm precoated Scharlau
silica gel SiF-254 Glasschrom plates.
Com p ou n d s. All commercially available compounds (Ald-
rich) were used without further purification unless otherwise
indicated. Tetrahydrofuran (THF), toluene, benzene, and
diethyl ether were distilled from sodium benzophenone ketyl.
Ethyl alcohol and dichloromethane were distilled from calcium
hydride. Acetone was distilled from K2CO3. N,N-Dimethyl-
formamide (DMF) was dried over activated 4 Å molecular
sieves.
Rou te to 1,3,5-Tr is[3-(N-(6-m eth yl-2-p yr id yl)ca r ba m -
oyl-4-h yd r oxyp h en yl]ben zen e (1). 1,3,5-(4-Ben zyloxy-3-
m eth oxyca r bon ylp h en yl)ben zen e (6). A 500-mL three-
neck round-bottom flask equipped with a magnetic stir bar, a
rubber septum, and a reflux condenser fitted with an argon
inlet adaptor was charged with 3.91 g (7.4 mmol) of triester
5, 120 mL of dry acetone, and 75 mL of DMF. Benzyl bromide
(3.6 mL, 33.3 mmol) was added by syringe, followed by 14.8 g
(45.4 mmol) of Cs2CO3. The dark brown solution was refluxed
for 14 h, allowed to cool to room temperature, filtered over
Celite, and concentrated by rotary evaporation to afford a
brown thick paste.25 The paste was dissolved in 50 mL of
diethyl ether and extracted with water (3 × 40 mL). The
organic phase was dried over Na2SO4, filtered, and concen-
trated to yield the benzylated triester (5.3 g, 90%) as a light
yellow solid: mp 176-178 °C; IR (KBr) 1730, 1690, 1610, 1500,
1
1450, 1260, 1090, 900 cm-1; H NMR (CDCl3) δ 8.14 (d, J )
Syn th esis. Meth yl 5-Acetyl-2-h yd r oxyben zoa te (4). An
oven-dried 500-mL, three-neck round-bottom flask, equipped
2.4 Hz, 3H), 7.74 (dd, J ) 8.7, 2.4 Hz, 3H), 7.67 (s, 3H), 7.42
(m, 15H), 7.12 (d, J ) 8.7 Hz, 3H), 5.26 (s, 6H), 3.95 (s, 9H);
13C NMR (CDCl3) δ 167.3, 158.4, 141.7, 137.3, 134.0, 132.7,
131.1, 129.2, 128.5, 127.5, 124.8, 121.7, 114.2, 71.3, 52.8. An
analytical sample was prepared by recrystallization from ethyl
acetate. Anal. Calcd for C51H42O9: C, 76.68, H, 5.30. Found:
C, 76.49, H, 5.40.
with
a mechanical stirrer, a thermometer, and a reflux
condenser fitted with an argon inlet, was charged with 200 g
of polyphosphoric acid (PPA). The flask was immersed in an
oil bath heated at 80 °C and stirred. When the PPA temper-
ature reached 80 °C, 12.2 mL (213 mmol) of glacial acetic acid
was added at once. After 15 min, 25 mL of methyl salicylate
was added at once and the reaction mixture stirred at this
temperature for 2 h. The reaction mixture had turned dark
and was poured over 1500 g of an ice-water mixture to
hydrolyze the PPA. The resulting aqueous mixture was
extracted with CH2Cl2 (5 × 100 mL). The organic extracts were
combined, washed with 5% NaHCO3 aqueous solution and
brine, dried over Na2SO4, and concentrated by rotary evapora-
tion to yield a dark red oil. The oil was vacuum distilled (0.1
mmHg), yielding a fraction of unreacted methyl salicylate
collected between 60 and 70 °C and 7.7 g of 4 (20%) collected
between 115 and 120 °C as a colorless thick liquid which
turned to a white solid on standing: mp 59-61 °C; IR (KBr)
1,3,5-Tr is(4-ben zyloxy-3-ca r boxyp h en yl)ben zen e (7).
In a 250-mL round-bottom flask equipped with a magnetic stir
bar and a reflux condenser fitted with an argon inlet were
placed 4.7 g (5.9 mmol) of the benzylated triester prepared
above and 150 mL of THF. To the resulting solution was added
150 mL of 3 N NaOH, and the mixture was refluxed under
argon atmosphere for 20 h. A yellowish solid was collected after
having removed most of the THF under reduced pressure. The
solid was suspended in 50 mL of 5% HCl, stirred for 15 min,
collected again, and dissolved in CHCl3. The organic solution
was dried over Na2SO4, filtered, and evaporated to afford
triacid 6 (4.3 g, 6%) as a white solid: mp 130-132 °C. IR (KBr)
3300-3500 (broad), 1720, 1680, 1600, 1490, 1380, 1230 cm-1
;
1
3300-3600 (broad), 1680, 1200, 1360 cm-1; H NMR (CDCl3)
1H NMR (CDCl3) δ 8.41 (d, J ) 2.1 Hz, 3H), 7.85 (dd, J ) 8.7,
2.1 Hz, 3H), 7.65 (s, 3H), 7.45 (m, 15H), 7.24 (d, J ) 8.7 Hz),
5.35 (s, 6H); 13C NMR (CDCl3) δ 166.8, 157.8, 140.7, 135.2,
134.8, 134.1, 132.3, 129.7, 129.6, 128.6, 124.6, 118.8, 114.4,
δ 11.24 (s, 1H), 8.48 (d, J ) 2.1 Hz, 1H), 8.09 (dd, J ) 8.7, 2.1
Hz, 1H), 7.04 (d, J ) 8.7 Hz, 1H), 4.00 (s, 3H), 2.58 (s, 3H);
13C NMR (CDCl3) δ 196.33, 170.62, 165.74, 135.96, 131.91,
129.49, 118.46, 112.50, 53.24, 26.80. An analytical sample of
(25) 1H NMR spectroscopy of an aliquot of the crude mixture reveals
the presence of a significant amount of DMF. The following water
extraction greatly reduces the amount of DMF.
(24) Reference values were taken from the Bruker Almanac 1992,
as indicated in the tables of Properties of Some Important Solvents.
J . Org. Chem, Vol. 67, No. 25, 2002 8839

Ballester et al.
72.8. Anal. Calcd for C49H36O9: C, 76.18, H, 4.79. Found: C,
76.15, H, 4.76.
mmol) of the triester 5 with 300 mL of freshly distilled dry
acetone and 275 mL of dry DMF. Next, 7.9 mL (91.3 mmol, d
) 1.398 g/mL) of allyl bromide was added to the solution by
syringe. After several minutes, 39.7 g (121.8 mmol) of Cs2CO3
was added, producing a change of the solution’s color from pink
to brown dark. The mixture was stirred and refluxed under
argon during 14 h. The reaction mixture was allowed to reach
room temperature and filtered through Celite. The filtered
solution was concentrated by rotary evaporation to yield a
brown sticky solid. The solid was dissolved in 50 mL of diethyl
ether and extracted with water (3 × 40 mL). The organic phase
was dried over Na2SO4, filtered, and concentrated to yield
1,3,5-tris(4-allyoxy-3-methoxycarbonylphenyl)benzene (11.2 g,
85%) as a brown solid: mp 56-60 °C; IR (KBr) 1720, 1600,
1500, 1260, 850 cm-1; 1H NMR (CDCl3) δ 8.12 (d, J ) 2.5 Hz,
3H), 7.76 (dd, J ) 8.7, 2.5 Hz, 3H), 7.67 (s, 3H), 7.07 (d, J )
8 Hz, 3H), 6.10 (ddt, J ) 17.2, 10.6, 4.7 Hz, 3H) 5.55 (ddt, J )
17.2, 1.5, 1.5 Hz, 3H), 5.34 (ddt, J ) 10.6, 1.5, 1.3 Hz, 3H),
4.70 (ddd, J ) 4.7, 1.5, 1.3 Hz, 6H); 13C NMR (CDCl3) δ 167.35,
158.39, 141.68, 133.81, 133.27, 132.70, 131.09, 124.79, 121.47,
118.22, 114.68, 70.22, 52.84.
1,3,5-Tr is(4-b e n zyloxy-3-ch lor ofor m ylp h e n yl)b e n -
zen e (8). 1,3,5-Tris(4-benzyloxy-3-carboxyphenyl)benzene (3.95
g, 5 mmol) was dissolved in a mixture of 30 mL of CH2Cl2 and
11.4 mL (156.7 mmol) of SOCl2. The resulting clear solution
was refluxed for 2 h. The solvent was evaporated in vacuo,
affording a light brown solid of the triacid chloride (3.9 g, 87%),
which was used in the next step without further purification:
IR (KBr) 1770, 1730, 1600, 1490, 1290, 1260, 900 cm-1 1H
;
NMR (CDCl3) δ 8.36 (d, J ) 2.3 Hz, 3H), 7.85 (dd, J ) 8.7, 2.3
Hz, 3H), 7.66 (s, 3H), 7.42 (m, 15H), 7.17 (d, J ) 8.7 Hz, 3H),
5.29 (s, 6H); 13C NMR (CDCl3) δ 164.3, 158.9, 141.3, 136.3,
135.2, 133.8, 133.6, 129.3, 128.7, 127.4, 124.9, 123.5, 114.7,
71.3.
1,3,5-Tr is[4-b en zyloxy-3-(N-(6-m et h yl-2-p yr id yl)ca r -
ba m oyl)p h en yl]ben zen e (9). In an oven-dried 250-mL,
three-neck round-bottom flask equipped with a magnetic stir
bar, a stopper, a rubber septum, and an argon inlet adaptor
were placed 3.9 g (4.8 mmol) of the triacyl chloride prepared
above and 100 mL of dry CH2Cl2. 2-Amino-6-methylpyridine
(3.12 g, 28.8 mmol) was added in one portion and the resulting
solution was stirred at room temperature for 4 h. The solution
was diluted with 50 mL of CH2Cl2 and washed with 5% HCl
(4 × 50 mL), 5% NaHCO3, and brine, dried over Na2SO4,
filtered, and concentrated in vacuo to afford the title compound
(3.9 g, 78%) as a white solid: mp 223-225 °C; IR (KBr) 3450,
An analytical sample was prepared by recrystallization from
CH2Cl2/MeOH. Anal. Calcd for C39H36O9: C, 72.21, H, 5.59.
Found: C, 72.17, H, 5.60.
1,3,5-Tr is(5-allyl-4-h ydr oxy-3-m eth oxycar bon ylph en yl)-
ben zen e (11). In a dry 100-mL round-bottom flask equipped
with magnetic stir bar and reflux condenser was dissolved 10
g (15.4 mmol) of the above prepared triester in 20 mL of N,N-
dimethylaniline freshly distilled under vacuum. The reaction
mixture was refluxed under argon atmosphere during 13 h.
The reaction mixture was allowed to cool and was diluted with
100 mL of diethyl ether. The organic solution was extracted
with 5% HCl (3 × 50 mL) and brine (50 mL), dried over Na2-
SO4, filtered, and evaporated to dryness, affording 9 g (90%)
of the title compound as a brown solid: mp 199-202 °C; IR
1660, 1530, 1450, 1300, 800, 740 cm-1 1H NMR (CDCl3) δ
;
10.44 (s, 3H), 8.62 (s, J ) 2.4 Hz, 3H), 8.22 (d, J ) 8.4 Hz,
3H), 7.82 (dd, J ) 8.7, 2.4 Hz, 3H), 7.80 (s, 3H), 7.60 (m, 9H),
7.40 (m, 9H), 7.20 (d, J ) 8.7 Hz, 3H), 6.86 (d, J ) 7.5 Hz,
3H), 5.36 (s, 6H), 2.40 (s, 9H); 13C NMR (CDCl3) δ 163.9, 157.5,
156.9, 151.9, 141.4, 138.9, 135.9, 134.9, 132.7, 131.7, 129.4,
129.1, 128.5, 124.8, 122.8, 119.5, 114.2, 111.9, 72.3, 24.7. Anal.
Calcd for C66H54N6O6: C, 77.17, H, 5.30, N, 8.18. Found: C,
77.15, H, 5.36, N, 8.20.
(KBr) 3300-3600 (broad), 1670, 1440, 800 cm-1 1H NMR
;
(CDCl3) δ 11.12 (s, 3H), 8.02 (d, J ) 2.4 Hz, 3H), 7.64 (d, J )
2.4 Hz, 3H), 7.59 (s, 3H), 6.07 (ddt, J ) 2.4, 10.2, 6.5 Hz, 3H),
5.13 (m, 6H), 3.98 (s, 9H), 3.53 (d, J ) 6.5 Hz, 6H); 13C NMR
(CDCl3) δ 171.53, 159.98, 142.25, 136.67, 135.46, 132.54,
129.93, 127.18, 124.90, 116.88, 112.65, 53.11, 34.58.
1,3,5-Tr is[3-(N-(6-m et h yl-2-p yr id yl)ca r b a m oyl-4-h y-
d r oxyp h en yl]ben zen e (1). Meth od A. In a 10-mL round-
bottom flask equipped with a magnetic stir bar and a rubber
septum was dissolved under argon atmosphere 100 mg (0.097
mmol) of the benzylated triamide prepared above in 2.5 mL
of dry CH2Cl2. Iodotrimethylsilane (206 mL, 1.455 mmol) was
added to the above solution by syringe and the reaction
mixture was stirred for 72 h. MeOH (2 mL) was added and
the reaction stirred further for 5 min. The solvent was
evaporated in vacuo to afford a dark brown solid. Trituration
of the solid with several 5 mL portions of diethyl ether yielded
62 mg (85%) of the triamide 1 as a brown solid.
1,3,5-Tr is(4-h ydr oxy-3-m eth oxycar bon yl-5-pr opylph en -
yl)ben zen e (5a ). In a 250-mL Parr shaker hydrogenation
flask, 900 mg of Pd/C 10% was suspended in a solution
prepared by mixing 9 g (13.9 mmol) of the triester above in
the minimum amount of ethyl acetate. The suspension was
shaken for 2 h at room temperature with 9 psig of H2. The
catalyst was filtered through Celite and the solvent was
evaporated in vacuo, yielding 8.4 g (92%) of 5a as a white
solid: mp 212-215 °C; IR (KBr) 3400 (broad), 3150 (broad),
Meth od B. HBr gas was bubbled for 5-10 min into a
solution of 250 mg (0.243 mmol) of the benzylated triamide
prepared above in 15 mL of CHCl3. A yellowish precipitate
was formed, then 10 mL of 5% NaHSO4 followed by 20 mL of
THF was added, and the mixture was stirred for 20 min. When
all the solids had completely dissolved, the organic layer was
separated, washed with brine, dried over Na2SO4, filtered, and
evaporated in vacuo to afford a solid. Trituration of the solid
with several portions of hexanes yielded 184 mg of pure
triamide 1 (58%) which was dried on a drying pistol under
vacuum over P2O5 using refluxing ethyl acetate: mp >300 °C;
1
1676, 1440, 1240, 1200, 1140, 800 cm-1; H NMR (CDCl3) δ
11.10 (s, 3H), 8.02 (d, J ) 2.1 Hz, 3H), 7.66 (d, J ) 2.1 Hz,
3H), 7.62 (s, 3H), 3.98 (s, 9H), 2.75 (t, J ) 7.5 Hz, 6H), 1.73
(m, 6H), 1.02 (t, J ) 7.3 Hz, 9H); 13C NMR (CDCl3) δ 171.70,
160.26, 142.36, 135.52, 132.40, 132.33, 126.71, 124.82, 112.68,
53.09, 32.70, 23.45, 14.75. Anal. Calcd for C39H42O2: C, 71.54,
H, 6.47. Found: C, 71.60, H, 6.46.
1,3,5-Tr is(3-(N-6-m eth yl-2-pyr idyl)car bam oyl-4-h ydr oxy-
5-p r op ylp h en yl)ben zen e (1a ). An oven-dried 250-mL three-
neck round-bottom flask equipped with a magnetic stir bar, a
glass stopper, and a rubber septum was charged with 4 g (6.1
mmol) of triester 5a and 50 mL of freshly distilled dry benzene
under argon. The resulting suspension was cooled at 0 °C with
an ice bath, and 13.8 mL (27.5 mmol) of 2 M AlMe3 was slowly
transferred via cannula, maintaining the reaction mixture
temperature at 0 °C throughout the addition. The reaction
mixture was stirred at room temperature for 20 min. Using a
similar procedure, 2.98 g (27.5 mmol) of 2-amino-6-methyl-
pyridine dissolved in 50 mL of benzene was treated with 13.8
mL (27.5 mmol) of 2 M AlMe3. The amine solution was added
via cannula to the ester solution and the reaction mixture was
heated to reflux for 48 h. During this time, a brownish
IR (KBr) 3400-3500 (broad), 1640, 1570, 1300, 1170, 800 cm-1
;
1H NMR (DMSO-d6) δ 12.3 (s, 3H), 11.07 (s, 3H), 8.58 (s, 3H),
8.18 (d, J ) 8.1 Hz, 3H), 8.10 (d, J ) 8.1 Hz, 3H), 8.00 (s, 3H),
7.85 (t, J ) 7.8 Hz, 3H), 7.30 (d, J ) 8.1 Hz, 3H), 7.14 (d, J )
7.5 Hz, 3H), 2.54 (s, 9H); 13C NMR (DMSO-d6) δ 170.5, 162.9,
162.7, 156.6, 146.6, 144.5, 140.3, 138.5, 137.5, 134.5, 128.9,
125.1, 123.8, 117.2, 29.5.
Rou te to 1,3,5-Tr is(3-(N-6-m eth yl-2-p yr id yl)ca r ba m -
oyl-4-h yd r oxy-5-p r op ylp h en yl)ben zen e (1a ). 1,3,5-Tr is-
(4-a llyloxy-3-m eth oxyca r bon ylp h en yl)ben zen e (10). In
an oven-dried 1-L three-neck round-bottom flask equipped
with an argon inlet adaptor, a magnetic stirring bar, a reflux
condenser, and a rubber septum were dissolved 10.7 g (20.3
8840 J . Org. Chem., Vol. 67, No. 25, 2002

Dual Binding Mode of Methylmethanetriacetic Acid
precipitate appeared in the mixture. The reaction mixture was
poured over 50 g of ice and 2 M HCl was added until the pH
of the aqueous layer was approximately 4-5. The aqueous
layer was extracted with CHCl3 (4 × 40 mL). The combined
organic extracts were washed with 5% NaHSO4 (2 × 60 mL)
and brine (50 mL) and dried over Na2SO4, and the solvent was
removed by rotary evaporation, yielding a solid. 1H NMR
analysis of the solid revealed a complex mixture. Flash
chromatography of the solid using CH2Cl2/EtOAc (98:2) af-
forded three different fractions: starting material, monoamide
diester, and diamide monoester. The triamide was eluted using
CH2Cl2/EtOAc (90:10), and rotary evaporation of the solvent
yielded 1.4 g (25.9%) of 1a as a pale yellow solid: mp 281-
285 °C. IR (KBr) 3200-3500 (broad), 1640, 1540, 1450, 1310,
) 7.62 Hz, 6H), 1.739 (m, 6H), 0.992 (t, J ) 7.2 Hz, 9H); 13C
NMR (CDCl3) δ 167.21, 143.72, 141.83, 141.15, 132.14, 130.71,
128.81, 125.98, 125.58, 52.17, 37.86, 24.54, 13.78.
An analytical sample was prepared by recrystallization from
ethanol. Anal. Calcd for C39H42O6: C, 77.20, H, 6.98. Found:
C, 77.25, H, 7.00.
1,3,5-Tr is(3-ca r boxy-5-p r op ylp h en yl)ben zen e (14). A
250-mL one-neck round-bottom flask was charged with 0.8 g
(1.32 mmol) of the triester prepared above, 57 mL of THF, and
57 mL of 3 N NaOH. The reaction mixture was refluxed
overnight. Then the THF was eliminated by rotary evapora-
tion. The aqueous layer was washed three times with 20 mL
of diethyl ether and acidified by slow addition of concentrated
HCl, affording a white solid. The solid was taken up in a
solvent mixture THF/Et2O (3:7). The organic phase was dried
over Na2SO4, filtered, and evaporated to dryness, yielding 0.62
g (83.3%) of the title product as a white solid: mp 260 °C dec;
1
790 cm-1; H NMR (CDCl3) δ 12.19 (s, 3H), 8.93 (s, 3H), 8.09
(d, J ) 8.1 Hz, 3H), 7.78 (s, 3H), 7.62 (m, 9H), 6.90 (d, J ) 7.2
Hz, 3H), 2.77 (t, J ) 7.6 Hz, 6H), 2.34 (s, 9H), 1.76 (m, 6H),
1.04 (t, J ) 7.2 Hz, 9H); 13C NMR (CDCl3) δ 169.88, 160.77,
157.46, 150.92, 141.72, 139.29, 134.20, 133.39, 131.56, 123.85,
123.03, 120.28, 114.50, 112.18, 32.84, 24.21, 23.42, 14.89.
An analytical sample was prepared by recrystallization from
benzene. Anal. Calcd for C54H54N6O6: C, 73.45, H, 6.16, N,
9.52. Found: C, 73.50, H, 6.18, N, 9.54.
1
IR (KBr) 3000 (broad), 1692, 1599, 1450, 1300, 731 cm-1; H
NMR (DMSO-d6) δ 13.105 (s, 3H), 8.233 (s, 3H) 8.044 (s, 3H),
8.018 (s, 3H), 7.920 (s, 3H). 2.839 (t, J ) 7.2 Hz, 6H), 1.784
(m, 6H), 1.038 (t, J ) 7.2 Hz, 9H); 13C NMR (DMSO-d6) δ
171.35, 147.47, 145.25, 144.34, 135.91, 135.43, 132.43, 129.42,
129.06, 40.92, 28.06, 17.60.
Rou te to 1,3,5-Tr is[3-(N-6-m eth yl-2-p yr id yl)ca r ba m -
oyl-5-p r op ylp h en yl]ben zen e (1b). 1,3,5-Tr is(3-m eth oxy-
ca r b on yl-5-p r op yl-4-t r iflu or om et h a n esu lfon yp h en yl)-
ben zen e (12). A 100-mL round-bottom flask equipped with a
magnetic stir bar, a rubber septum, and an argon inlet was
charged with 1 g (1.53 mmol) of triester 5a and 20 mL of
anhydrous pyridine. The resulting suspension was slightly
heated to induce complete solution of the triester. The solution
was cooled in an ice bath and 1.5 mL (d ) 1.719 g/mL, 9.14
mmol) of triflic anhydride was added by syringe. During the
addition, the temperature rose and the color of the solution
turned intense violet. The reaction mixture was stirred during
3 days at room temperature. The reaction mixture was diluted
with 50 mL of CH2Cl2 and washed with 5% HCl (5 × 20 mL),
an aqueous solution of CuSO4 (until persistence of the blue
color), and water. The organic phase was dried over Na2SO4,
filtered, and evaporated to dryness, affording an amber solid
that was purified by percolation on silica gel using 100 mL of
CH2Cl2. Rotary evaporation of the solvent yields 1.25 g (78%)
of the title compound as a white solid: mp 153-155 °C; 1H
NMR (CDCl3) δ 8.097 (d, J ) 2.43 Hz, 3H), 7.762 (s, 3H), 7.750
(s, 3H), 3.983 (s, 3H), 2.837 (t, J ) 7.71 Hz, 6H), 1.768 (m,
6H), 1.033 (t, J ) 7.26 Hz, 9H); 13C NMR (CDCl3) δ 165.13,
145.12, 140.56, 140.16, 137.51, 133.44, 128.79, 126.21, 126.09,
118.55 (q), 52.81, 32.06, 23.42, 13.81.
Anal. Calcd for C36H36O6: C, 76.57, H, 6.43. Found: C, 76.50,
H, 6.49.
1,3,5-Tr is(3-ch lor ofor m yl-5-pr opylph en yl)ben zen e (15).
The triacid (0.620 g, 1.1 mmol) prepared above was suspended
in 15 mL of SOCl2. The suspension was refluxed for 2 h, and
the solvent was evaporated to dryness, affording a light brown
solid. Trituration of the solid with hexane gave 0.654 g (99.8%)
of the triacyl chloride, which was used in the next step without
further purification: 1H NMR (CDCl3) δ 8.36 (d, J ) 2.3 Hz,
3H), 7.85 (dd, J ) 8.7, 2.3 Hz, 3H), 7.66 (s, 3H), 7.42 (m, 15H),
7.17 (d, J ) 8.7 Hz, 3H), 5.29 (s, 6H); 13C NMR (CDCl3) δ 164.3,
158.9, 141.3, 136.3, 135.2, 133.8, 133.6, 129.3, 128.7, 127.4,
124.9, 123.5, 114.7, 71.3.
1,3,5-Tr is[3-(N-6-m eth yl-2-p yr id yl)ca r ba m oyl-5-p r op -
ylp h en yl]ben zen e (1b). A 50-mL, one-neck round-bottom
flask equipped with a magnetic stir bar and an argon inlet
was charged with 654 mg (1.1 mmol) of the triacyl chloride
prepared above and 20 mL of dry CH2Cl2. 2-Amino-6-meth-
ylpyridine (1.1 g, 10.2 mmol) was added in one portion and
the resulting solution was stirred at room temperature for 4
h. The solution was diluted with 20 mL of CH2Cl2, washed
with 5% HCl (4 × 50 mL), 5% NaHCO3, and brine, dried over
Na2SO4, filtered, and evaporated in vacuo. The residue was
purified by column chromatography (CH2Cl2/hexane/THF, 78:
20:2) to furnish 664 mg (72.6%) of 1b as a white solid: mp
An analytical sample was prepared by recrystallization from
ethanol. HRMS calcd for C42H39F9O15S3 1050.1307, found
1050.1309.
1
178-180 °C; MS: m/z 835 (24.5%, M+), 739 (100%); H NMR
(CDCl3) δ 8.67 (s, 3H), 8.22 (d, J ) 8.3 Hz, 3H), 8.08 (t, J )
1.5 Hz), 7.86 (s, 3H), 7.77 (t, J ) 1.5 Hz, 3H), 7.72 (t, J ) 1.5
Hz, 3H), 7.65 (t, J ) 7.7, 3H), 6.93 (d, J ) 7.5, 3H), 2.773 (t,
J ) 7.2 Hz, 6H), 2.452 (s, 9H), 1.76 (m, 6H), 1.01 (t, J ) 7.2
Hz, 9H); 13C NMR (CDCl3) δ 166.1, 157.3, 151.3, 144.7, 142.2,
141.9, 139.2, 135.5, 131.8, 126.9, 126.2, 124.2, 119.8, 111.4,
38.4, 25.0, 24.4, 14.3. Anal. Calcd for C54H54N6O3: C, 77.67,
H, 6.52, N, 10.06. Found: C, 77.65, H, 6.46, N, 10.10.
1,3,5-Tr is(3-m e t h oxyca r b on yl-5-p r op ylp h e n yl)b e n -
zen e (5b). A 50-mL three-neck round-bottom flask equipped
with stir bar, reflux condenser, and argon inlet adaptor was
charged with 1 g (0.95 mmol) of the triflate prepared above
and 19 mL of anhydrous DMF. To the resulting solution were
successively added 0.163 g (0.23 mmol) of PdCl2(PPh3)3, 0.22
g (0.55 mmol) of 1,3-bisdiphenylphosphinopropane, 3.75 mL
(15.74 mmol) of triethylamine, and 0.38 mL (10.07 mmol) of
formic acid. The color of the solution turned brown dark. The
reaction mixture was stirred for 9 h at 110 °C under argon.
The reaction mixture was diluted with 50 mL of CH2Cl2 and
washed with 5% HCl (3 × 15 mL), 5% NaHCO3 (3 × 15 mL),
and brine (3 × 15 mL). The organic layer was dried over Na2-
SO4, filtered, and concentrated under vacuum. The crude
product was purified by column chromatography on silica gel
(CH2Cl2) to give 0.810 g (81%) of the title compound as a white
solid: mp 144-147 °C; 1H NMR (CDCl3) δ 8.186 (s, 3H), 7.908
(s, 3H), 7.796 (s, 3H), 7.694 (s, 3H), 3.957 (s, 9H), 2.744 (t, J
Ack n ow led gm en t. We thank the DGESIC (Projects
PB 98-0129 and SAB 1998-0109) for financial support
and the sabbatical leave of Prof. Deslongchamps, re-
spectively.
Su p p or tin g In for m a tion Ava ila ble: Crystal data (CIF)
for complexes 1a ‚3 and 1b‚3. This material is available free
of charge via the Internet at http://pubs.acs.org.
J O025787L
J . Org. Chem, Vol. 67, No. 25, 2002 8841