Synthesis of C-Alkylcalix[4] arenes. 4. Design, Synthesis, and Computational Studies of Novel Chiral Amido[4]resorcinarenes

Article abstract of DOI:10.1021/jo962018r

In extending our studies involving BF₃·Et₂O-catalyzed reaction of cinnamic acid analogues, we have shown that amido derivatives also can afford [4]resorcinarene octamethyl ethers. Subsequently, chiral monomeric amides, derived from the mixed anhydride of cinnamic acid and L- or D-valine, upon treatment with BF_a·Et₂O, yielded for the first time chiral amido [4]resorcinarenes in enantiomerically pure forms. Four stereoisomers were isolated, and three of them were attributed the flattened-cone, chair, and 1,2-alternate conformations. The major product was assigned a novel chairlike structure, namely flattened partial cone 1. The flattened-cone stereoisomer, which was indicated by molecular modeling studies to be the most stable, became the major product under more drastic experimental conditions. Chromatographic studies on chiral phases revealed that the above tetramers could be used for the enantiodiscrimination of racemic molecules.

Full text of DOI:10.1021/jo962018r

932
J . Org. Chem. 1997, 62, 932-938
Syn th esis of C-Alk ylca lix[4]a r en es. 4. Design , Syn th esis, a n d
Com p u ta tion a l Stu d ies of Novel Ch ir a l Am id o[4]r esor cin a r en es
Bruno Botta,*,† Giuliano Delle Monache,*,‡ Patrizia Salvatore,† Francesco Gasparrini,†
Claudio Villani,† Maurizio Botta,*,§ Federico Corelli,§ Andrea Tafi,§ Eszter Gacs-Baitz,
Antonello Santini,| Cristopher F. Carvalho,∇ and Domenico Misiti†
Dipartimento di Studi di Chimica e Tecnologia delle Sostanze Biologicamente Attive, Universita` “La
Sapienza”, P. le A. Moro 5, 00185 Roma, Italy, Centro Chimica dei Recettori del C.N.R., Istituto di
Chimica e Chimica Clinica, Universita` Cattolica del S. Cuore, 00168 Roma, Italy, Dipartimento Farmaco
Chimico Tecnologico, Universita` degli Studi, 53100 Siena, Italy, Center Research Institute for Chemistry,
Hungarian Academy of Sciences, H-1525 Budapest, Hungary, and Dipartimento di Scienza degli
Alimenti, Universita` “Federico II”, 80055 Napoli, Italy
Received October 29, 1996X
In extending our studies involving BF3‚Et2O-catalyzed reaction of cinnamic acid analogues, we
have shown that amido derivatives also can afford [4]resorcinarene octamethyl ethers. Subse-
quently, chiral monomeric amides, derived from the mixed anhydride of cinnamic acid and L- or
D-valine, upon treatment with BF3‚Et2O, yielded for the first time chiral amido [4]resorcinarenes
in enantiomerically pure forms. Four stereoisomers were isolated, and three of them were attributed
the flattened-cone, chair, and 1,2-alternate conformations. The major product was assigned a novel
chairlike structure, namely flattened partial cone 1. The flattened-cone stereoisomer, which was
indicated by molecular modeling studies to be the most stable, became the major product under
more drastic experimental conditions. Chromatographic studies on chiral phases revealed that
the above tetramers could be used for the enantiodiscrimination of racemic molecules.
In tr od u ction
[4]arenes by modification of the phenolic groups, were
systematically classified by Shinkai et al.9 Chiral sub-
stituents can be also introduced directly into the calix-
[n]arene framework9-11 either on the phenolic OH group
or at the p-position of the phenolic ring.
Molecular recognition by artificial enzymes is an
important goal in current bioorganic chemistry. The
selectivity of recognition is determined by a number of
featuresssuch as electrostatic or hydrophobic interac-
tions, hydrogen bonds, etc.sthat are incorporated by the
molecular receptor to fit, in a multiple-points binding,
the chemical characteristics of the substrate.1,2
Chiral calix[4]resorcinarenes have been recently pre-
pared by a Mannich reaction of [4]resorcinarene in the
cone conformation with formaldehyde and amino acids
such as L-proline.12
Chiral recognition, one of the most sophisticated func-
tions of enzymes, has been achieved in cyclodextrin-based
artificial enzymes3-6 but remains difficult to realize with
other enzyme mimics.
Calixarenes7 are synthetic macrocycles, endowed with
a cavity-shaped architecture similar to that of cyclodex-
trins. Due to the fact that the basic 14-metacyclophane
skeleton cannot be planar, there are several possibilities
to construct intrinsically chiral calix[4]arenes.8 Molec-
ular asymmetry can be generated not only by different
substituents but also by conformational isomerism. All
possible chiral isomers, which can be derived from calix-
Chiral calix[4]resorcinarene octamethyl ethers are not
known, nor has the synthesis of a chiral [4]resorcinarene
starting from a chiral monomer been reported.
In previous studies, we have shown that ethereal BF3
catalyzes the conversion of (E)-2,4-dimethoxycinnamic
acid esters to the corresponding mixture of [4]resor-
cinarene octamethyl ethers.13 For instance, ethyl ester
1 gave the 1,2-alternate (2a ), the flattened cone (2b), and
the 1,3-alternate (2c) stereoisomers in the ratio 2a :2b:
2c of 2:3:1. The relationship between the nature of the
ester and the stereoisomer distribution in the reaction
mixture has been discussed in a previous publication.14
It was of interest to establish whether a starting mono-
† Universita´ “La Sapienza”.
‡ Universita´ Cattolica del S. Cuore.
§ Universita´ degli Studi di Siena.
(9) Iwamoto, K.; Shimizu, H.; Araki, K.; Shinkai, S. J . Am. Chem.
Soc. 1993, 115, 3997 and references therein.
(10) Muthukrishnan, R.; Gutsche, C. D. J . Org. Chem. 1979, 44,
3962.
Hungarian Academy of Sciences.
| Universita´ “Federico II”.
∇ Present address: School of Physical Sciences, University of
Southern Queensland, Toowoomba, Australia.
X Abstract published in Advance ACS Abstracts, J anuary 1, 1997.
(1) Relevant examples have been provided to illustrate this con-
cept: (a) Diedrich, F. Angew. Chem., Int. Ed. Engl. 1988, 27, 362. (b)
Lehn, J . M. Ibid. 1988, 27, 89. (c) Cram, D. J . Ibid. 1988, 27, 1009.
(2) (a) J eong, K. S.; Rebek, J . J . Am. Chem. Soc. 1988, 110, 3327
and references therein. (b) Pant, N.; Hamilton, A. D. Ibid. 1988, 110,
2002.
(11) (a) Shinkai, S.; Arimura, T.; Sath, H.; Manabe, O. J . Chem.
Soc., Chem. Commun. 1987, 1495. (b) Arimura, T.; Edamitsu, S.;
Shinkai, S.; Manabe, O.; Muramatsu, T.; Tashiro, M. Chem. Lett. 1987,
2269. (c) Arimura, T.; Kawabata, H.; Matsuda, T.; Muramatsu, T.;
Satoh, H.; Fujio, K.; Manabe, O.; Shinkai, S. J . Org. Chem. 1991, 56,
301. (d) Ikeda, A.; Nagasaki, T.; Shinkai, S. J . Phys. Org. Chem. 1992,
5, 699.
(12) (a) Matsushita, Y.; Matsui, T. Tetrahedron Lett. 1993, 34, 7433.
(b) Yanagihara, R.; Tominaga, A. M.; Aoyama, Y. J . Org. Chem. 1994,
59, 6865. (c) Schneider, U.; Schneider, H. J . Chem. Ber. 1994, 127,
2455.
(3) Breslow, R. Acc. Chem. Res. 1980, 13, 170.
(4) Tabushi, I. Acc. Chem. Res. 1982, 15, 66.
(5) Bender, M. L.; Komiyama, M. In Cyclodextrin Chemistry;
Springer-Verlag: New York, 1977.
(6) D’Sonza, V. T.; Bender, M. L. Acc. Chem. Res. 1987, 20, 146.
(7) (a) Gutsche, C. D. Calixarenes; Royal Society of Chemistry:
(13) Botta, B.; Iacomacci, P.; Di Giovanni, M. C.; Delle Monache,
G.; Ga´cs-Baitz, E.; Botta, M.; Tafi, A.; Corelli, F.; Misiti, D. J . Org.
Chem. 1992, 57, 3259.
Cambridge, 1989. (b) Vicens, J ., Bo¨hmer, V., Eds. Calixarenes:
Versatile Class of Macrocyclic Compounds; Kluwer: Dordrecht, 1991.
(8) Bo¨hmer, V. Angew. Chem., Int. Ed. Engl. 1995, 34, 713.
A
(14) Botta, B.; Di Giovanni, M. C.; Delle Monache, G.; De Rosa, M.
C.; Ga´cs-Baitz, E.; Botta, M.; Corelli, F.; Tafi, A.; Santini, A.; Benedetti,
E.; Pedone, C.; Misiti, D. J . Org. Chem. 1994, 59, 1532.
S0022-3263(96)02018-X CCC: $14.00 © 1997 American Chemical Society

Synthesis of C-Alkylcalix[4]arenes
J . Org. Chem., Vol. 62, No. 4, 1997 933
obtained by coupling of the mixed anhydride of 2,4-
dimethoxycinnamic acid with L-valine, the latter pro-
tected as the ethyl ester. The enantiomeric monomers
D-5 and the racemic (D,L-5) were also prepared by the
same procedure starting from D-valine and D,L-valine,
respectively. Both D- and L-5 showed an ee >98% [1H
NMR after addition of Eu(fod)3] or >99% by HPLC on
CSP1 (vide infra).
The monomer L-5 was converted into the expected
mixture of macrocycles in a 75% overall yield by treat-
ment with BF3‚Et2O (see Experimental). Extensive chro-
matography of the reaction mixture afforded three opti-
cally active stereoisomeric tetramers (MH + at m/ z 1341
+
and M - chain at m/ z 1154 in the FAB MS spectra)
with a gross [14] metacyclophane structure 6 (Figure 1).
1
Compound L-6b showed the simplest H NMR spectra
(only one signal for each type of proton) and was at-
tributed a flattened cone conformation. Notably, the
methoxyl groups and the quaternary nonoxygenated
aromatic carbons gave two signals in the 13C NMR
spectrum, thus showing the nonequivalence, e.g., of C-1
and C-3 carbons due to the presence of a chiral center in
the C-2 substituent. The influence of the chiral center
on the C-25/C-28 or C-4/C-24 pair was not observable as
two distinct lines.
F igu r e 1. Tetramerization reaction of 2,4-dimethoxycinnamic
acid amido derivatives.
mer incorporating an amide group could similarly afford
the calixarene system nucleus. Successful results would
then allow extension of this route to chiral end products
since amides derived from chiral amino acids could be
utilized.
With these objectives in mind, we investigated the
BF3‚Et2O-catalyzed cyclization of amide 3 and found that
the corresponding calixarenes bearing the 1,2-alternate,
flattened-cone, and 1,3-alternate conformations were
indeed obtained. It was therefore appropriate to extend
our studies into the chiral series with amides derived
from D- and L-valine. The present discussion describes
our results directed to the first synthesis of chiral C-alkyl-
[4]resorcinarenes, starting from a chiral unit.
The substituents in L-6b were all cis and pseudoaxial
since DIF NOE experiments revealed the proximity of
the methine and the methoxyl groups as well as that of
the methylene group and the aromatic Hi protons.14
In the 1H- and 13C-NMR spectra of compound L-6a
every carbon and every proton gave its own signal, thus
not revealing any diagnostic symmetry.
On consideration of the presence in the 1H NMR
spectrum of two high-field signals (δ 2.24 and 2.02)
attributable to the protons of the C-14 methylene group,14
a 1,2-alternate structure can be proposed for L-6a . As
confirmation of this, irradiation of the signal at 2.24 ppm
in a DIF NOE experiment showed the proximity of the
methylene with aromatic (Hi) protons at 6.80 and 6.43
(H-26/H-27), that is, cis-trans-cis configuration relative
to C-2.
Resu lts a n d Discu ssion
The most abundant isomer (45%) 6d was assigned a
novel conformation on the basis of the spectral data
(Table 1). The compound revealed in the 1H NMR
spectrum three signals with an integrated intensity of
1:2:1 for the external (He) and internal (Hi) aromatic
protons and two signals (2H:2H distribution) for the
bridged methine protons. Accordingly, the distribution
pattern of signals in the 13C NMR spectrum suggested
the presence of a plane of symmetry passing through two
opposite nonequivalent aromatic rings (e.g., A and C in
Figure 1).
This statement is only a rationale since with the chiral
side chains an “actual” plane of symmetry is not possible.
The splitting patterns of the NMR signals of the aromatic
framework are not significantly influenced by the distant
chiral centers of the side chains and thus approximate
the result that would be expected if a plane of symmetry
were present.
Syn th esis a n d Ch a r a cter iza tion . The monomer 3
was synthesized by treatment of (E)-2,4-dimethoxycin-
namic acid in anydrous THF with diethyl chlorophos-
phate and benzylamine in the presence of K2CO3 under
reflux for 2 h. Treatment of 3 with BF3‚Et2O (1:6 molar
ratio) under reflux for 2 h gave a mixture of tetramers
in 60% overall yield. Three products (4a , 4b, 4c in the
ratio 2:3:1) were purified by crystallization and column
chromatography. On the basis of the similarity of their
1H and 13C NMR spectra with those of esters 2a , 2b, and
2c, the three stereoisomers (Figure 1) were attributed
the structures 4a (1,2-alternate conformation with a cis-
trans-cissrelative to C(2)sarrangement of the pseudo-
axial side chains), 4b (flattened cone conformation with
an all-cis relative configuration of the pseudoaxial side
chains), and 4c (1,3-alternate conformation with the
pseudoequatorial side chains in an all-cis position). The
relative orientation of the side chains was confirmed by
additional DIF NOE experiments.14
The symmetry consistent with the distribution pattern
of protonated carbon signals in 6d can be associated with
a flattened partial cone 1 geometry (Cs type), obtained
by our previous molecular-modeling studies of C-alkyl-
[4]resorcinarene.14 Irradiation of the 2H signal at 6.92
ppm in a DIF NOE experiment showed the proximity of
the equivalent H-25 and H-27 protons with both the C-2/
As expected from C-alkyl[4]resorcinarenes,14 FAB MS
spectra of 4a -c showed a base peak corresponding to the
loss of one side chain, whereas EIMS spectra were
characterized by further losses of neutral benzylamine.
The above findings prompted us to synthesize chiral
amido[4]resorcinarenes starting from the chiral unit L-5,

934 J . Org. Chem., Vol. 62, No. 4, 1997
Botta et al.
Ta ble 1. 1H- a n d 13C-NMR Sp ectr a l Da ta of Com p ou n d s
L-6d /D-6d in th e F la tten ed -P a r tia l Con e 1 Con for m a tion a
position
δC
δH
1,21; 3,19
7,15; 9,13
2,20
124.27, 124.07
122.74, 122.48
33.19
5.04 (2H, dd, 9, 6)
4.99 (2H, t, 7)
8,14
33.16
4,18; 6,16
10,12; 22,24
5,17
156.34, 156.31
156.14, 156.02
96.17 (×2)
95.69
6.38 (2H, s)
6.44 (2H, s)
11
23
25
26,27
28
OMe
95.57
127.62
F igu r e 2. Amido[4]resorcinarene with flattened-partial cone
2 (chair) conformation.
6.45 (1H, brs)
6.92 (2H, brs)
6.40 (1H, brs)
3.76 (3H, s)
3.90 (6H, s)
3.74 (3H, s)
2.75 (2H, dd, 15, 7)
2.66 (2H, dd, 15, 9)
2.77 (4H, d, 7)
126.95 (×2)
127.21
ent brush-type CSPs. CSPs 1-3 are known to afford
facile resolutions of many racemates having different
types and combinations of stereogenic elements and
functionalities, the highest levels of enantioselective
separations usually being observed for compounds that
can simultaneously establish H-bond (or dipole-dipole)
and aromatic-aromatic interactions with the stationary
phase. Our chiral calixarenes, featuring the electron-
rich aromatic nucleus and a series of H-bond donor/
acceptor sites surrounding the macrocycle core, possess
the required complementary functionalities and are thus
expected to have high and different affinities for the
above phases. For instance, when the product mixture
from the reaction of L-5 with BF3‚Et2O was submitted to
separation on chiral columns CSP 1, CSP 2, and CSP 3,
it was possible to separate the chiral calixarenes L-6a ,b,d .
The analysis revealed the presence of another minor
product (L-6f), which was in turn obtained by preparative
TLC of column fractions selected on the basis of CSP
chromatograms.
55.89
55.83 (×2)
55.77
CH2
41.69
41.63
172.03
172.01
CO
NH
CHN
CH
6.31 (2H, d, 8.5)
6.11 (2H, d, 8.5)
57.04
56.97
31.32
31.02
18.59, 18.55
17.76, 17.71
171.82, 171.75
60.77
14.19
14.12
4.35 (2H, dd, 8.5, 6)
4.37 (2H, dd, 8.5, 6)
1.84 (2H, dh, 7 × 6, 6)
1.94 (2H, dh, 7 × 6, 5)
0.70 (9H, d, 7)
Me
0.59 (3H, d, 7)
CO
OCH2
Me
4.09 (8H, m)
1.23 (6H, t, 7)
1.19 (6H, t, 7)
a
Proton and carbon signals were correlated by a HETCOR
experiment.
Ta ble 2. P h ysica l Da ta of Ch ir a l
C-Alk yl[4]r esor cin a r en es a n d Sta r tin g Mon om er s
1
An interesting feature in the H NMR spectra of the
new compound was the “absence” of signals for the
methylene protons at room temperature. Heating to 60
°C led to the appearance of a broad signal centered at
ca. 2.56 ppm, which by irradiation of the unique methine
signal (t at 5.15 ppm) gave two broadened doublets at
2.61 and 2.51 ppm with a geminal coupling of ca. 15 Hz.
These findings suggest a very rigid structure and a
hindered rotation of the substituents.
Moreover, the distribution pattern of the signals for
the methoxyl groups (2 signals) and aromatic protons
and carbons (two signals for both CHe and CHi) in 1H-
and 13C-NMR spectra is compatible with either a C2h
(flattened partial cone 2)14 or D2d (1,3-alternate) sym-
metry.
conformation
[R]D
compd amino acid of the tetramers mp (°C) (CHCl3)
c
6a
6a
6b
6b
6d
D-valine
L-valine
D-valine
L-valine
D-valine
1.2-alternate
1.2-alternate
flattened cone
flattened cone
flattened partial 218-219 +23.0 0.30
cone 1
flattened partial 219-220 -22.9 0.20
cone 1
chair
a
a
+77.4 0.16
-77.4 0.23
192-193 +41.0 0.12
194-195 -41.6 0.12
6d
L-valine
6f
5
5
L-valine
D-valine
L-valine
a
-22.0 0.18
112-113 -59.6 0.4
112-113 +59.7 1.2
a
Vitreous solid.
Irradiation of the signal at 2.56 ppm (at 60 °C) in a
DIF NOE experiment revealed the proximity of the
methylene group with both Hi protons. Conversely, the
irradiation of the methine signal showed no effect on the
aromatic protons.
On the basis of these results, a flattened-partial cone
2 structure, with all axial substituents having a trans-
trans-cis configuration relative C-2, was attributed to
compound L-6f (Figure 2). A similar structure, named
chair, was attributed to the minor [4]resorcinarenes
obtained from the reaction of resorcinol with heptanal
and dodecanal.17
C-20 and C-8/C-14 methylene pairs, thus requiring the
side chains to be all-cis and axial.
When the tetramerization reaction was repeated with
the D-valine derivative 5 as the starting monomer,
1
compounds D-6a ,b,d were obtained and showed H and
13C NMR spectra identical with those of L-6a ,b,d . Fur-
thermore, the same absolute values, but of course op-
posite sign (Table 2), were observed in optical rotations
for each couple of enantiomers.
Finally, treatment of the racemic monomer (D,L-5) with
BF3‚Et2O gave a very complex mixture of products that
resisted chromatographic separation. Studies with these
racemic products were discontinued.
(15) (a) Iwamoto, K.; Yanagi, A.; Arimura,T.; Matsuda, T.; Shinkai,
S. Chem. Lett. 1990, 1901. (b) Pappalardo, S.; Caccamese, S.; Giunta,
L. Tetrahedron Lett. 1991, 32, 7747. (c) Shinkai, S.; Arimura, A.;
Kawabata, H.; Murakami, H.; Iwamoto, K. J . Chem. Soc., Perkin Trans.
1 1991, 2429. (d) Araki, K.; Inada, K; Shinkai, S. Angew. Chem., Int.
Ed. Engl. 1996, 35, 72.
Ch r om a togr a p h ic Beh a vior of Ch ir a l Ca lixa r e-
n es on HP LC Ch ir a l P h a ses. Optical resolutions of
inherently chiral calixarenes have been recently obtained
by HPLC using both polymeric15 and brush-type16 chiral
stationary phases. Here, we report the separation of the
enantiomeric pairs of compounds 6a ,b,d on three differ-
(16) Pickard, S. T.; Pirkle, W. H.; Tabatabai, M.; Vogt, W.; Bo¨hmer,
V. Chirality 1993, 5, 310.

Synthesis of C-Alkylcalix[4]arenes
J . Org. Chem., Vol. 62, No. 4, 1997 935
The mixtures comprising the D-series (D-6a ,b,d ,f) were
also separated on chiral columns CSP 1, CSP 2, and CSP
3. The best results in both the cases were obtained with
CSP 3.
mechanics, less reliable due to the difficulty of accurate
energetical evaluation of the H-bond contribution to the
total conformational energy.
The conformational searching of compound 6 was
performed by adding the amino acidic side chains at the
methylene bridges of every minimum energy conforma-
tion of the [4]resorcinarene nucleus13 and executing five
separate random studies of the preferred spatial orienta-
tions of these moieties. The reliability of such a proce-
dure of conformational analysis of resorcinarenes has
been discussed previously.13,14 According to that protocol
the program BKM (generalized MM2 force field)20-22 was
chosen in a first attempt to carry out statistical searches
on the side chains of 6a -e. In order to reduce the
number of rotatable bonds of the side chains, the four
carbethoxy residues of the valine moieties were replaced
with four carbomethoxyl groups. We were confident that
the computational results were not dependent on these
minor structural modifications.
A further Monte Carlo conformational search was
performed on the side chains of 6a -e with the program
MacroModel22 with the aim to take into account the
contribution of intramolecular H-bond formation to the
total steric energy in a more appropriate manner. The
AMBER*23 force field, as implemented in the version 4.0
of this program, was selected for this purpose for two
main reasons: first, the possibility to compare the results
obtained in the search carried out with the generalized
MM2 force field implemented in BKM with those ob-
tained using AMBER*, a force field expecially suited to
model molecular systems possessing H-bond donor and
acceptor groups, and second, the finding that, among the
different force fields implemented in MacroModel, only
AMBER* allowed reliable parameters to be applied in
this study. It is well known that conformational energy
differences and even more geometries may not be ac-
curate if low-quality generalized parameters are in use.
Although the solvent can play a crucial role when, as
in calixarenes, there is a cavity in the molecular struc-
ture, two observations convinced us to neglect the study
of the solvent effects on the conformational preferences
of 6: first, the lack of high-quality reliable parameters
in the MacroModel force field OPLS*,24 usually used to
calculate the steric energy of molecules in a continuum
solution model,25 when it was applied to 6, and second,
the molecular dimensions of 6 (20 rotatable bonds) that
prevented us from applying an explicit solvatation model,
because very time-consuming searches were required
even without explicit solvent molecules added to the
system.
In order to fully establish the utility of this HPLC
method in future studies relating to the separation of
racemate calixarene mixtures, the D- and L-derived
calixarenes were mixed to provide authentic samples of
D,L-6a ,b,d ,f. The three phases shown in Scheme 2s
(Supporting Information) were tested for their ability to
separate the above racemic mixtures into the pure
enantiomers. Pertinent chromatographic data are pro-
vided in the Supporting Information. All of the examined
chiral phases show some degree of selectivity for the
enantiomers; retention and enantioselectivity are largely
dependent (although at present in unpredictable manner)
on both the stationary-phase structure and analyte
conformation, a finding that was expected considering the
large differences in the spatial arrangement of the
interaction sites among the four calixarenes derivatives.
Thus, racemic compound D,L-6a is the only stereoisomer
separated on the three phases, with the highest separa-
tion factor (R ) 3.31, corresponding to a ∆∆G ∼ 0.7 kcal/
mol) found on CSP 3. Interestingly, while CSP 2 and 3
always select for the same amino acid configuration of
the monomer and of the cyclic tetramers, CSP 1 prefer-
entially retains the D enantiomer of the former and the
L enantiomer of the latter.
Although the chiral tetramers are obtained in enan-
tiomerically pure form, and therefore chromatography on
chiral phases is of little interest from the analytical point
of view in the present study,18 examination of our
chromatographic data in a reciprocal sense19 suggests
that the roles played by hosts and guests could be
conveniently exchanged. For example, by covalent at-
tachment of a homochiral calixarene to silica particles it
should be possible to generate new, easily accessible
enantioselective stationary phases with recognition abil-
ity toward aromatic substrates. Such HPLC phases
should facilitate the screening of larger sets of racemic
candidates for enantioselective binding. Work in this
area is underway in our laboratories.
Molecu la r Mod elin g Stu d ies
The conformational study of the [4]resorcinarene with
general formula 6 (Figure 1) bearing four valine moieties,
one for each side chain, has been performed.
The presence of amide groups in the side chains of
compound 6 allows intramolecular H-bond formation,
thereby distinguishing this compound from the [4]resor-
cinarenes with ester side chains previously studied by
us.13,14 For this reason, a complete conformational study
of compound 6 becomes necessary, since the sequence of
the steric energies of the different geometries of the
calixarene nucleus can be substantially modified by the
presence of nonbonded intramolecular interactions among
the amide side chains. Nevertheless, it is necessary to
point out that the presence in 6 of amide, instead of ester,
functionalities makes in principle the conformational
steric energies of this compound, calculated by molecular
The results of the conformational study of compound
6 are summarized in Tables 3 and 4. In Table 3 are
reported the global steric energies (kcal/mol) of the lowest
energy conformations found in the searches performed
(20) The BKM Program is available from Prof. Kosta Steliou, Boston
University, MA.
(21) In this conformational study the BAKMDL Program has been
substituted by the BKM Program. The latter one is the Unix version
of BAKMDL and maintains the force field and the capabilities of that
program.
(22) MacroModel V4.0: Mohamadi, F.; Richards, N. G. J .; Guida,
W. C.; Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J . Comput. Chem. 1990, 11, 440.
(23) Ferguson, D. M.; Kollman, P. A. J . Comput. Chem. 1991, 12,
620.
(17) Abis, L.; Dalcanale, E.; Du Vosel, A.; Spera, S. J . Org. Chem.
1988, 53, 5475.
(18) However, we do use HPLC on CSP 3 to rapidly determine
stereoisomer ratios in crude mixtures, taking advantage of the large
diastereoselectivity often displayed by brush-type phases.
(19) Pirkle, W. H.; Da¨ppen, R. J . Chromatogr. 1987, 404, 107.
(24) J orgensen, W. L.; Severance, D. L. J . Am. Chem. Soc. 1990,
112, 4768.
(25) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J .
Am. Chem. Soc. 1990, 112, 6127.

936 J . Org. Chem., Vol. 62, No. 4, 1997
Botta et al.
Ta ble 3. Ster ic En er gies of MM2-Min im ized a n d
AMBER*-Min im ized (All-Atom s a n d Un ited -Atom s
Ch a r ge Sets in Ma cr oMod el) Con for m a tion s of
C-Alk yl[4]r esor cin a r en es
in Table 5. Both AMBER* all-atoms and AMBER*
united-atoms calculations found a very low steric energy
for 6f justifying its obtainment as one of the products of
the synthesis.
steric energy (kcal/mol)
MacroModel MacroModel
Con for m a tion a l Stu d ies
AMBER*
all-atoms
AMBER*
united-atoms BKM
conformation
According to the relative orientation of the substituents
at the methylene bridges of the [4]resorcinarene skeleton
6 (Figure 3), the isomers 6b and 6d (cone and flattened
partial cone 1 conformations), possessing an all-cis
orientation of the side chains, are conformationally
isomeric, whereas the isomer 6a (1,2-alternate conforma-
tion with cis-trans-cis relative configuration of the side
chains) and the isomer 6f (chair conformation with trans-
trans-cis relative configuration of the side chains) have
a diastereomeric relationship with 6b and 6d .
6b (flattened cone)
6a (1,2-alternate)
-4.8
-0.4
-11.2
-8.4
62.2
68.5
77.4
75.1
76.7
6d (flattened partial cone 1) not founda
6e (flattened partial cone 2) 2.8
-4.2
not foundb
0.9
6c (1,3-alternate)
7.0
a
The exhaustive reminimization of the geometries collected in
the SPMC AMBER* all-atoms searches on 6d gave the flattened
partial cone 2 (6e) minimum energy conformation. The exhaus-
tive reminimization of the geometries collected in the SPMC
AMBER* united-atoms searches on 6e gave the flattened partial
cone 1 (6d ) minimum energy conformation.
b
The sequence of stabilities (in terms of steric energy)
derived from the molecular modeling studies, that is, 6b
> 6a ∼ 6f > 6d , was largely different from the order of
decreasing yields, that is, 6d > 6a > 6b > 6f. The
discrepancy could be explained if we assume that 6d was
the kinetic product, whereas 6b was the thermodynamic
one. It could thus be expected that 6d may be converted
in the more stable conformational isomer 6b by heating.
However, when L-6d was heated in refluxing CHCl3 for
8 h, no conversion was detected and L-6d remained
stable. Analogous results were obtained with all the
other stereoisomers. On the other hand, inspection of
CPK molecular models revealed that in each stereoiso-
mers all the side chains are located in a crowded region
of the [4]resorcinarene and would suffer in any confor-
mational change a severe unfavorable steric interaction
with the methoxyl groups of two adiacent aromatic rings.
Conversely, an interconversion could take place when at
least two covalent bonds are broken.14,26 In order to
provide additional information, we performed a series of
experiments, in which each stereoisomer was treated
with an excess of BF3‚Et2O. The reaction mixtures were
maintained under reflux for 6 h and, after workup, were
partially purified by silica gel column chromatography.
on 6a -e with both BKM and MacroModel (all-atoms and
united-atoms AMBER*). Table 4 shows the contributions
of bonded and nonbonded intramolecular interactions to
the steric energies found in the AMBER* all-atoms
search.
The BKM analysis yielded energy values of 6a -e
distributed in a large range but did not find a substantial
energy difference between the flattened partial-cones and
the 1,3-alternate conformations. The MacroModel analy-
ses, on the contrary, clearly indicate that the 1,3-
alternate geometry of the calixarene nucleus corresponds
to the less stable conformation of compound 6.
Regardless of which of the two input geometries (6d
or 6e) the search was initiated with, the Monte Carlo
conformational analysis always found both of the flat-
tened partial-cone geometries among the output struc-
tures collected within 25 kJ /mol energy window. The
following exhaustive reminimization of these structures
with AMBER* all-atoms and AMBER* united-atoms
force fields gave 6d and 6e, respectively, as the minimum
energy conformation of the flattened partial-cone geom-
etry (Table 3). This result is due to different evaluations
of the relative steric energies performed by the two force
fields.
As shown in Table 4, the steric energy differences
among the studied geometries of 6 can be ascribed to a
balance between bonded and nonbonded interactions. The
disposition in 3D space of the side chains in the flattened
cone and 1,2-alternate geometries of the calixarene
nucleus appears to be favorable for the optimization of
the intramolecular H-bonds (see Figure 3). By contrast,
in the flattened partial cone and in the 1,3-alternate
geometries the optimization of the H-bonds causes a
distortion of the calixarene nucleus, as indicated by the
bonded terms of the steric energy (Table 4), which
disfavors these conformations with respect to the previ-
ous ones. This effect is even stronger for the 1,3-alternate
geometry.
1
The H NMR spectra of the total mixture of tetramers
were analyzed in comparison with those of the single
starting calix[4]resorcinarenes to establish in a semi-
quantitative way (by integration of selected signals)
which ones were present. The results, summarized in
Table 5, showed that stereoisomers L-6d and L-6f gave
the thermodynamic isomer (L-6b) as the major product
(experiments 3 and 4). Under similar conditions, the
isomer L-6a reacted only partially to give a mixture of
L-6d and L-6b in a 2:1 ratio (experiment 1). Finally, L-6b
was recovered almost unchanged (experiment 2).
In summary, these findings indicate the flattened-cone
conformation to be the thermodynamically most stable
product in agreement with the results of molecular
modeling calculations. As a confirmation, treatment of
monomer L-5 with a large excess of BF3‚Et2O for 12 h
(Table 5, experiment 5) gave a mixture in which L-6b was
predominant. Notably, the pure product L-6b showed no
racemization ([R]D ) +40.5, see Table 2).
Due to the new experimental findings about the
stereochemistry of the side chains of compound 6f (trans-
trans-cis instead of all-cis as in the flattened partial-cone
2 structure 6e), a further MacroModel-AMBER* confor-
mational analysis was performed on this compound.
The steric energies of the lowest-energy conformation
found in this search are comparable to the values
reported in Table 4 (-0.1 and -9.1 kcal/mol, respec-
tively). The contributions of the bonded and nonbonded
intramolecular interactions to the steric energy are listed
Our experiments also revealed that [4]resorcinarenes
L-6a and L-6d are in equilibrium, since they are converted
into each other, but by more drastic conditions afford
finally L-6b. Also, the minor stereoisomer L-6f was
(26) Timmerman, P.; Verboom, W.; Reinhoudt, D. N. Tetrahedron
1996, 52, 2663.

Synthesis of C-Alkylcalix[4]arenes
J . Org. Chem., Vol. 62, No. 4, 1997 937
F igu r e 3. Minimized conformation of 6 via MacroModel AMBER* all-atoms calculations (dotted lines represent hydrogen bonds).
Ta ble 4. Con tr ibu tion s of Bon d in g a n d Non bon d in g In tr a m olecu la r In ter a ction s to th e Ster ic En er gies of th e
AMBER*-Min im ized (All-Atom s Ch a r ge Sets in Ma cr oMod el) Con for m a tion s of th e C-Alk yl[4]r esor cin a r en es 6a -d (Ta ble
3) a n d 6f
steric energy (kcal/mol)
conformation
stretching
bending
torsion
impr tors
electrost
van der Waals
6b (flattened cone)
6a (1,2-alternate)
6d (flattened partial cone 1)
6c (1,3-alternate)
6f (chair)
4.4
4.9
4.4
4.4
4.6
20.1
22.1
22.7
26.9
21.8
25.4
20.6
29.7
32.8
20.0
0.3
0.2
0.5
0.3
0.1
-53.9
-48.2
-56.7
-52.8
-49.5
-1.1
0.0
2.2
-4.6
2.9
Ta ble 5. Su m m a r y of Rea ction s w ith a n Excess of
BF 3‚Et2O (1:400 Mola r Ra tio) of L-Va lin a m id o
[4]Resor cin a r en es
and indicate that the flattened-cone stereoisomer 6b is
the thermodynamic sink. Compound 6b, which is formed
as minor product in kinetic condiction, becomes the main
component of the reaction mixture under more drastic
conditions (higher amount of BF3‚Et2O).
exp substrate time (h) major product(s) minor product(s)
1
2
3
4
5
L-6a
L-6b
L-6d
L-6f
L-5
6
6
6
6
12
6b, 6d a
6b
6b
6b
6b
6d
6a
6d
6a , 6d
These results, in agreement with those previously
obtained by us13,14 and by other authors,17,26 suggest that
the length and the steric hindrance of the side chains
play an important role in the stereoselectivity of the
reaction, when it is carried out under kinetic conditions.
Moreover, the above tetramerization reaction also re-
vealed a high reproducibility since each pair of enanti-
omers showed the same absolute value of optical rotation.
Although the chiral [4]resorcinarenes were obtained
in enantiomerically pure form, chromatographic studies
on chiral phases revealed that the above tetramers could
be used, in future studies, for the enantiodiscrimination
of racemic molecules bearing 3,5-dinitrobenzoylamido
substituents.
a
40-45% of 6a unreacted.
shown to be in equilibrium with L-6d and to give L-6b as
final product. These equilibria require that at least two
C-C bonds are cleaved before recombination can give the
other isomer.
Con clu sion s
The experimental results obtained in the Lewis acid-
catalyzed cyclization of chiral cinnamic acid amide
analogues afford a novel and direct synthetic route to
chiral [4]resorcinarenes octamethyl ethers.
A [4]resorcinarene with the novel flattened-partial-cone
1 conformation was isolated for the first time from the
reaction mixture together with three other stereoisomers
(1,2-alternate, flattened cone, and chair17).
Exp er im en ta l Section
Melting points are uncorrected. NMR spectra, in CDCl3,
are referred to TMS (0.00 ppm). FAB-MS were obtained using
a thioglycerol matrix.
The molecular modeling studies suggest that hydrogen
bond formation could be the driving force of the reaction
(E)-N-Ben zyl-2,4-d im eth oxycin n a m a m id e (3). To a so-
lution of (E)-2,4-dimethoxycinnamic acid (2.0 g, 9.6 mmol) and

938 J . Org. Chem., Vol. 62, No. 4, 1997
Botta et al.
distilled Et3N (3.0 mL, 21.4 mmol) in anhydrous THF (60 mL)
was added a solution of diethyl chlorophosphate (2.0 g, 11.7
mmol) in anhydrous THF (15 mL) slowly under N2 atmosphere.
The mixture was stirred for 2 h at room temperature, filtered,
and added dropwise to a solution of benzylamine (1.4 g, 13.3
mmol) and Et3N (3.0 mL, 21.4 mmol) in anhydrous CH2Cl2
(107 mL) under N2. The reaction mixture was stirred for 2 h
at room temperature, concentrated, and extracted with an
aqueous solution of 10% Na2CO3. The organic phase was
washed with H2O, dried over Na2SO4, and evaporated. The
residue gave by crystallization from n-hexane/ EtOAc, 3:2,
compound 3 (1.7 g, yield 60%), as white needles, mp 139-40
°C. 1H- and 13C-NMR spectra are provided in Supporting
Information.
2,8,14,20-Tetr akis(ben zylam ido)[4]r esor cin ar en es (4a-
c). To a solution of the (E)-N-benzyl-2,4-dimethoxycinnama-
mide (0.3 g, 1 mmol) in CHCl3 (6 mL) was added BF3‚Et2O
(0.7 mL, 5 mmol), and the mixture was left under reflux and
stirring for 2.5 h. The reaction mixture was poured into ice-
water and left under stirring overnight at room temperature.
Column chromatography (on silica gel eluting with CH2Cl2-
EtOAc-MeOH, 90:7:3) and crystallization (from acetone) gave
[4]resorcinarenes 4a (64 mg, 21%), 4b (97 mg, 32%), and 4c
(30 mg, yield 10%).
with CHCl3 and then acidified with concentrated HCl and
reextracted with CHCl3. The pooled organic layers were dried
over Na2SO4 and evaporated. Silica gel chromatography
(CH2Cl2:EtOAc, 95:5) of the residue yielded 1.2 g (75%) of the
amide 5. 1H and 13C-NMR and MS spectral data are provided
in the Supporting Information.
2,8,14,20-Tetr a k is(va lin a m id o)[4]r esor cin a r en es. To a
solution of L-5 (1 mmol, 335 mg) in CHCl3 (5 mL) was added
BF3‚Et2O (6 mmol, 0.72 mL). The mixture was stirred at
reflux for 1.50 h. The reaction mixture was quenched with
ice and water and stirred at room temperature overnight. The
solution was extracted with CH2Cl2, and the combined organic
layers were washed with H2O and dried with Na2SO4. Evapo-
ration of the solvent and purification by column chromatog-
raphy eluting with CH2Cl2/EtOAc/MeOH mixtures afforded
L-6a (100 mg, yield 30%. Anal. Calcd for C72H100N4O20: C,
64.44; H, 7.52; N, 4.18. Found: C, 64.38; H, 7.58; N, 4.20);
L-6b (27 mg, 8%. Anal. Calcd for C72H100N4O20: C, 64.44; H,
7.52; N, 4.18. Found: C, 64.26; H, 7.65; N, 4.12); L-6f (10 mg,
3%. Anal. Calcd for C72H100N4O20: C, 64.44; H, 7.52; N, 4.18.
Found: C, 64.28; H, 7.55; N, 4.21); and L-6d (114 mg, 34%.
Anal. Calcd for C72H100N4O20
: C, 64.44; H, 7.52; N, 4.18.
Found: C, 64.30; H, 7.64; N, 4.16). The minor product 6f was
obtained by preparative TLC (CH2Cl2:EtOAc:MeOH, 90:7:3)
of enriched fractions selected on the basis of CSPs chromato-
grams. The reaction of D-5 with BF3‚Et2O gave comparable
results, but compound D-6f was not isolated
Compounds 6a ,b,d ,f gave very similar FAB MS spectra. A
typical fragmentation (6d ) is reported: 1341 [MH]+ (78), 1340
[M]+ (100), 1454 [M - chain]+ (96), 981 (19), 808 (9).
Rea ction of L-Va lin e [4]r esor cin a r en e w ith a n Excess
of BF 3‚Et2O. In a typical experiment, to a solution of L-valine
[4]resorcinarene (27 mg, 2 × 10-2 mmol) in CHCl3 (5 mL) was
added BF3‚Et2O (0.96 mL, 8 mmol). Standard workup gave a
residue that was partially purified by silica gel column
chromatography (CH2Cl2:EtOAc:MeOH, 90:7:3) to give the
mixture of tetramers.
Compound 4a (1,2-alternate conformation): vitreous solid;
1H- and 13C-NMR spectra are provided in the Supporting
Information; EIMS m/ z (rel int) 1188 [M]+ (100), 1040 [M -
CH2CONHCH2Ph]+ (73), 933 [1040 - H2NCH2Ph]+ (21), 824
[933-H2NCH2Ph]+ (38). Anal. Calcd for C72H76N4O12
: C,
72.69; H, 6.44; N, 4.71. Found: C, 72.67; H, 6.45; N, 4.70.
1
Compound 4b (cone conformation): mp 241-2 °C; H- and
13C-NMR spectra are provided in the Supporting Information;
EIMS m/z (rel int) 1188 [M]+ (48), 1040 [M - CH2CONHCH2-
Ph]+ (41), 933 [1040 - H2NCH2Ph]+ (16), 824 [933 - H2-
NHCH2Ph]+ (100). Anal. Calcd for C72H76N4O12: C, 72.69;
H, 6.44; N, 4.71. Found: C, 72.58; H, 6.51; N, 4.65.
Compound 4c (1,3-alternate conformation): mp 286-7°C;
1H- and 13C-NMR spectra are provided in the Supporting
Information; EIMS m/ z (rel int) 1188 [M]+ (96), 1040 [M -
CH2CONHCH2Ph]+ (100), 933 [1040 - H2NCH2Ph]+ (17), 824
Ack n ow led gm en t . We thank Professor J ames P.
Kutney, University of British Columbia, Vancouver, for
a discussion of the paper. This work was supported by
grants from C.N.R., from MURST, as well as from the
Agreement between the C.N.R. of Italy and the Hungar-
ian Academy of Sciences.
[933-H2NCH2Ph]+ (33). Anal. Calcd for C72H76N4O12
: C,
72.69; H, 6.44; N, 4.71. Found: C, 72.64; H, 6.50; N, 4.62.
D-Va lin e Eth yl Ester Hyd r och lor id e. To absolute EtOH
(20 mL), cooled to -5 °C, was added SOCl2 (1.0 mL, 1.62 g,
13.6 mmol) over 5 min. This mixture was then cooled to -5
°C, and D-valine hydrochloride 1.90 g (12.4 mmol) was added.
The solution was stirred for 2 h at room temperature, heated
under reflux for 2 h, and then cooled to rt. Removal of the
solvent in vacuo afforded the title compound (yield 80%): [R]D
) -6.7 (c ) 2, H2O); mp 102-105 °C.
(E)-N-[1-(Ca r b oxyet h yl)-2-m et h ylp r op yl]-2,4-d im et h -
oxycin n a m a m id e (L-5, D-5). To a solution of (E)-2,4-
dimethoxycinnamic acid (1 g, 4.8 mmol) in dry THF (40 mL)
under N2 were added distilled Et3N (1.5 mL, 10.7 mmol) and
diethyl chlorophosphate (0.85 mL) in dry THF (1 mL). The
resultant solution was maintained at 30 °C for 2 h, after which
time was added, dropwise, a solution of the appropriate valine
ethyl ester hydrochloride (1.6 g, 4.8 mmol) in dry THF (5 mL),
and the mixture was stirred at 30 °C for a further 3 h. The
reaction mixture was concentrated in vacuo and diluted with
aquous NaHCO3 solution. The aqueous layer was extracted
Su p p or tin g In for m a tion Ava ila ble: Tetramerization
reactions of 2,4-dimethoxycinnamic acid amido derivatives
(Scheme 1s); 1H- and 13C-NMR and MS spectral data of
monomers 3 and D-5/L-5; 1H- and 13C-NMR spectra of C-alkyl-
[4]resorcinarenes 2a , 4a , D-6a /L-6a (1,2-alternate, Tables 1s
and 2s), 2b, 4b, D-6b/L-6b (flattened-cone, Table 3s), and 2c,
4c, D-6f/L-6f (1,3-alternate and chair, Table 4s). Chiral phases
used for enantiodiscrimination studies (Scheme 2s) and sepa-
ration of the racemic mixture of valinamido calix [4]resor-
cinarenes (Table 5s). Details on molecular modeling studies
(11 pages). This material is contained in libraries on micro-
fiche, immediately follows this article in the microfilm version
of the journal, and can be ordered from the ACS; see any
current masthead page for ordering information.
J O962018R