Control of structural isomerism in noncovalent hydrogen-bonded assemblies using peripheral chiral information

Article abstract of DOI:10.1021/ja9936262

The results of a systematic study of the structural isomerism in more than 30 noncovalent hydrogen-bonded assemblies are described. These dynamic assemblies, composed of three calix[4]arene dimelamines and six barbiturates/cyanurates, can be present in th

Full text of DOI:10.1021/ja9936262

J. Am. Chem. Soc. 2000, 122, 3617-3627
3617
Control of Structural Isomerism in Noncovalent Hydrogen-Bonded
Assemblies Using Peripheral Chiral Information
Leonard J. Prins, Katrina A. Jolliffe, Ron Hulst, Peter Timmerman,* and
David N. Reinhoudt*
Contribution from the Laboratory of Supramolecular Chemistry and Technology, MESA+ Research
Institute, UniVersity of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands
ReceiVed October 8, 1999
Abstract: The results of a systematic study of the structural isomerism in more than 30 noncovalent hydrogen-
bonded assemblies are described. These dynamic assemblies, composed of three calix[4]arene dimelamines
and six barbiturates/cyanurates, can be present in three isomeric forms with either D3, C3h, or Cs symmetry.
The isomeric distribution can be readily determined via a combination of 1H NMR and 13C NMR spectroscopy.
In one case it is shown that the covalent capture of the dynamic assemblies via a ring-closing metathesis
(RCM) reaction provides a novel analytical tool to distinguish between the D3 and C3h isomeric forms of the
assembly. For the D3 isomer the RCM results in the formation of a cyclic trimer, comprising three dimelamines,
whereas for the C3h isomer a cyclic monomer is formed. Molecular dynamics simulations in chloroform are
qualitatively in agreement with the experimental data and reveal that the isomeric distribution is determined
by a combination of steric, electronic, and solvation effects. A wide range of isomeric distributions covering
all extremes has been found for the studied assemblies. Those with 5,5-disubstituted barbituric acid derivatives
exclusively form the D3 isomer, because steric hindrance between the barbiturate substituents prevents formation
of the C3h and Cs isomers. In contrast, assemblies with isocyanuric acid derivatives exhibit increased stability
of the C3h and Cs isomers upon increasing the size of the isocyanurate substituent. The outcome of the assembly
process can be controlled to a large extent via chiral substituents in the calix[4]arene dimelamines, due to the
preferred orientation of the chiral centers.
Introduction
ognized that there are three different structural motifs in the
isocyanuric acid‚melamine (CA‚M) lattice, i.e., the linear and
crinkled tape motif and the cyclic rosette motif.9,10 Structural
isomerism has also been observed within other well-defined
molecular assemblies as a result of different orientations of the
building blocks within the assembly11 or as a result of
tautomerism.12 These forms of structural isomerism often result
in the formation of a mixture of assemblies. This always
complicates the analysis and will eventually affect the applica-
tion of this type of self-assembled structures.
Noncovalent interactions play an important role in determin-
ing the three-dimensional structure of biological assemblies.1
Although only a limited number of building blocks are used,
the structural variety in these assemblies is enormous. Examples
are the various forms of DNA2 and the R-helices or â-sheets in
proteins.3 Subtle changes in the arrangement of the building
blocks can have a large effect on the secondary structure of
proteins, as was recently demonstrated by Sauer et al.4 The
interchange of just two amino acids in the Arc repressor
completely changes the secondary structure of the protein from
a â-sheet to a right-handed helix. Similar forms of structural
isomerism have been observed in synthetic noncovalent as-
semblies, which are the subject of numerous studies5 because
they offer new perspectives in the fields of catalysis,6 guest
encapsulation,7 and combinatorial chemistry.8 Whitesides rec-
(7) (a) Kusukawa, T.; Fujita, M. Angew. Chem., Int. Ed. 1998, 37, 3142-
3144. (b) MacGillivray, L. R.; Atwood, J. L. Nature 1997, 389, 469-472.
(c) Rivera, J. M.; Mart´ın, T.; Rebek, J., Jr. Science 1998, 279, 1021-1023.
(d) Meissner, R. S.; Rebek, J., Jr.; De Mendoza, J. Science 1995, 270, 1485-
1488. (e) Wyler, R.; De Mendoza, J.; Rebek, J., Jr. J. Am. Chem. Soc. 1993,
32, 1699-1701.
(8) For a short review, see: Timmerman, P.; Reinhoudt, D. N. AdV.
Mater. 1999, 11, 71-74.
(9) Zerkowski, J. A.; Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc.
1992, 114, 5473-5475.
(10) (a) Mascal, M.; Hext, N. M.; Warmuth, R.; Moore, M. H.;
Turkenburg, J. P. Angew. Chem., Int. Ed. Engl. 1996, 35, 2204-2206. (b)
Lehn, J.-M.; Mascal, M.; DeCian, A.; Fischer, J. J. Chem. Soc., Chem.
Commun. 1990, 479-480. (c) Seto, C. T.; Whitesides, G. M. J. Am. Chem.
Soc. 1990, 112, 6409-6411.
(11) (a) Hasenknopf, B.; Lehn, J.-M.; Boumediene, N.; Leize, E.; Van
Dorsselaer, A. Angew. Chem., Int. Ed. 1998, 37, 3265-3268. (b) Castellano,
R. K.; Kim, B. H.; Rebek, J., Jr. J. Am. Chem. Soc. 1997, 119, 12671-
12672. (c) Fujita, M.; Ibukuro, F.; Yamaguchi, K.; Ogura, K. J. Am. Chem.
Soc. 1995, 117, 4175-4176. (d) Mathias, J. P.; Simanek, E. E.; Whitesides,
G. M. J. Am. Chem. Soc. 1994, 116, 4326-4340.
* To whom correspondence should be addressed.
(1) Stryer, L. Biochemistry; W. H. Freeman: New York, 1995.
(2) Dickerson, R. E.; Drew, H. R.; Conner, B. N.; Wing, R. M.; Fratini,
A. V.; Kopka, M. L. Science 1982, 216, 475-485
(3) Fersht, A. R. Enzyme Structure and Mechanism; W. H. Freeman:
New York, 1985.
(4) Cordes, M. H. J.; Walsh, N. P.; McKnight, C. J.; Sauer, R. T. Science
1999, 284, 325-328.
(5) For reviews, see: (a) Conn, M. M.; Rebek, J., Jr. Chem. ReV. 1997,
97, 1647-1668. (b) Linton, B.; Hamilton, A. D. Chem. ReV. 1997, 97,
1669-1680. (c) Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl.
1996, 35, 1154-1196. (d) Lawrence, D. S.; Jiang, T.; Levett, M. Chem.
ReV. 1995, 95, 2229-2260.
(6) For a short review, see: (a) Sanders, J. K. M. Chem. Eur. J. 1998,
4, 1378-1383. (b) Kang, J.; Santamar´ıa, J.; Hilmersson, G.; Rebek, J., Jr.
J. Am. Chem. Soc. 1998, 120, 7389-7390.
(12) (a) Gonza´lez, J. J.; Prados, P.; De Mendoza, J. Angew. Chem., Int.
Ed. 1999, 38, 525-528. (b) Corbin, P. S.; Zimmerman, S. C. J. Am. Chem.
Soc. 1998, 120, 9710-9711. (c) Beijer, F. H.; Sijbesma, R.; Kooijman, H.;
Spek, A. L.; Meijer, E. W. J. Am. Chem. Soc. 1998, 120, 6761-6769.
10.1021/ja9936262 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/05/2000

3618 J. Am. Chem. Soc., Vol. 122, No. 15, 2000
Prins et al.
Figure 1. Molecular structure of double rosette assemblies such as 13‚(DEB)6 and the three structural isomers of this type of assembly.
In this paper we describe the results of a systematic study
aimed at characterizing the various parameters that govern the
isomeric distribution in a series of nine-component hydrogen-
bonded assemblies. Understanding these parameters is essential
in order to control the assembly process and to generate one
(or more) of the isomers selectively. Initially, we will analyze
the isomeric distribution in more than 30 different assemblies
prising three molecules of calix[4]arene dimelamine 1 and six
molecules of 5,5-diethylbarbituric acid (DEB) are currently
investigated in our group.13 The driving force for the assembly
process is the cooperative formation of 36 hydrogen bonds. In
principle, assemblies such as 13‚(DEB)6 can be present as three
constitutionally different isomers: the D3 isomer (staggered),
the C3h isomer (symmetrically eclipsed), and the Cs isomer
(unsymmetrically eclipsed) (Figure 1).
1
by H and 13C NMR spectroscopy and show that molecular
dynamics simulations provide a qualitative insight into how the
different parameters influence the isomeric distribution. Second,
we show that the products formed after covalent capture of a
dynamic assembly can be used to determine the symmetry of
the assemblies, which renders it a useful analytical tool. Finally,
we show how in certain cases the isomeric distribution of the
assembly can be controlled via the introduction of chiral centers
in the individual building blocks.
In the D3 isomer the two melamine units of all three calix-
[4]arenes are in a staggered orientation with respect to each
other (Figure 2), which causes the assembly to be chiral.14
(13) (a) Jolliffe, K. A.; Crego Calama, M.; Fokkens, R.; Nibbering, M.
M.; Timmerman, P.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 1998,
37, 1247-1251. (b) Timmerman, P.; Vreekamp, R. H.; Hulst, R.; Verboom,
W.; Reinhoudt, D. N.; Rissanen, K.; Udachin, K. A. Chem. Eur. J. 1997,
3, 1823-1832. (c) Vreekamp, R. H.; Van Duynhoven, J. P. M.; Hubert,
M.; Verboom, W.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 1996,
35, 1215-1218.
(14) The orientation of the two melamines can either be clockwise (P)
or counterclockwise (M). In the absence of additional chiral centers present
in the building blocks, the staggered isomer exists as a racemic mixture of
the P and M assembly.
Results and Discussion
1H and 13C NMR Spectroscopic Characterization of the
Isomers. Nine-component assemblies (e.g., 13‚(DEB)6) com-

Structural Isomerism in NoncoValent H-Bonded Assemblies
J. Am. Chem. Soc., Vol. 122, No. 15, 2000 3619
three different positions, i.e., positions X, R1/R2, and R3 (Chart
1). We have investigated the formation of assemblies with
barbituric acid (BA) and a series of derivatives (5,5-dimethyl-
(DMB), 5,5-diethyl- (DEB), and 5-ethyl-5-phenyl- (EPB)) with
substituents that differ mainly in size, and one bisbarbiturate
(BisBAR) in which the two barbituric acid moieties are
connected via a C3-spacer. Furthermore, the formation of
assemblies from a similar series of isocyanuric acid derivatives
with butyl (BuCYA), neohexyl (HexCYA), 1,1,1-triphenylpro-
pyl (TripCYA), and p-tert-butylphenyl (PheCYA) substituents
was investigated.
The influence of the substituents at position R1 on assembly
formation was investigated using calix[4]arene dimelamines 1,
3, and 4, having butyl, benzyl, or (R)-1-phenylethyl substituents,
respectively. Furthermore, the formation of assemblies from
calix[4]arene dimelamine 5, with one (R)-1-phenylethyl and one
(S)-1-phenylethyl moiety, was investigated. Finally, we exam-
ined the effect of the introduction of NO2 substituents at position
R3 of dimelamine 2 on assembly formation.
Figure 2. Staggered and eclipsed orientation of the two melamine
moieties of a calix[4]arene dimelamine. The Ca and Cb carbons are
used as probes to analyze the symmetry of the assemblies by 13C NMR
spectroscopy.
Recently, we have shown that assemblies of one handedness
are formed when chiral centers are present in the building
blocks.15 The C3h and Cs isomers, in which the two melamines
(on each calix[4]arene) are in an eclipsed orientation (Figure
2), have a plane of symmetry and are therefore achiral. The
C3h and Cs isomers differ in the sense that the Cs isomer lacks
a C3-axis due to the 180° rotation of one of the calix[4]arenes.
Dimelamine 4 was prepared using a procedure similar to that
described for dimelamines 1,13b 2,13b and 3.13a For the synthesis
of dimelamine 5, one of the amino groups of 718 was BOC-
protected upon reaction with 1 equiv of BOCON to give 8 in
67% yield. Subsequent reaction of 8 with cyanuric chloride (1
equiv, 0 °C) followed by successive reaction with gaseous NH3
(0 °C) and (R)-1-phenylethylamine (6 equiv, reflux) in THF
afforded compound 9 in 64% yield. Removal of the BOC group
in 9 with TFA gave 10 (93%), which was successively treated
with cyanuric chloride (1.1 equiv, 0 °C), gaseous NH3 (0 °C),
and finally (S)-1-phenylethylamine (10 equiv, reflux) in THF
to give 5 as the final product in an overall yield of 18% (Scheme
1).
1
Whitesides et al. have previously discussed the value of H
NMR spectroscopy as a tool for the analysis of the symmetry
of hydrogen-bonded assemblies.16 Upon the formation of
assemblies, the NH protons of the barbiturate or isocyanurate
component get involved in hydrogen bonding and undergo a
large downfield shift from 8.4 to 13-16 ppm, an area free of
other resonances. The number of resonances reflects the
symmetry of the assembly. For assemblies such as 13‚(DEB)6,
the Cs isomer can easily be distinguished from the other two
isomers. On the basis of symmetry, a total of six proton signals
in the 13-16 ppm region of the 1H NMR spectrum is expected,
since all NHbarb protons within one rosette layer reside in a
different magnetic environment, due to the absence of the C3-
axis. For the D3 and C3h isomers two signals are expected, as
the three different substituents on the melamines render the two
NHbarb protons of each barbiturate chemically inequivalent in
the assembly, thereby giving two anisochronic signals for these
protons. Since the two rosette layers are related either by a C2-
axis (D3 isomer) or a σh-plane (C3h isomer), the protons in one
rosette layer are either homotopic or enantiotopic, respectively,
compared to the protons in the second rosette layer and therefore
give identical resonances.17 The D3 and C3h isomers can be
distinguished using 13C NMR spectroscopy. These isomers differ
in the number of signals for the Ca and Cb carbon atoms of the
unsubstituted aromatic rings of the calix[4]arene (Figure 2). The
C2-axes in the D3 isomer cause the Ca and Cb carbons to be
homotopic (isochronous), resulting in a single resonance. The
Ca and Cb carbons in the C3h isomer, which lacks C2-axes, are
heterotopic (anisochronous) and therefore give different signals.
Similarly, the protons Ha and Hb are homotopic in the D3 isomer
and enantiotopic in the C3h isomer, but these protons are
occasionally substituted by other groups, thus preventing a
Cyanurates BuCYA and PheCYA were synthesized in one
step from the corresponding amines via reaction with N-
chlorocarbonylisocyanate in 15% and 25% yield, respectively,
following a literature procedure.19 The remaining compounds
were either commercially available or have been reported
previously.20
Influence of the Barbituric Acid Substituent X on the
Isomeric Distribution. Our initial investigations of the structural
isomerism in hydrogen-bonded assemblies were prompted by
the observation that all DEB-containing assemblies that we
studied form only the D3 isomer (Table 1 summarizes the
isomeric distribution of all assemblies discussed in this paper).
As typical examples, the 1H NMR spectrum and part of the 13
C
NMR spectrum of assembly 23‚(DEB)6 are depicted in Figure
3. That assembly 23‚(DEB)6 exclusively forms the D3 isomer
is confirmed by the presence of only two NHbarb resonances at
1
14.01 and 13.72 ppm in the H NMR spectrum and a single
resonance for the Ca and Cb carbons at 141.08 ppm in the 13C
NMR spectrum. This is in full agreement with previously
reported X-ray crystal diffraction studies and 2D NOESY
experiments of assembly 23‚(DEB)6.13b Assemblies 13‚(DEB)6
and 33‚(DEB)6 give very similar spectra, which indicates that
in these cases the D3 isomer is also formed exclusively.
1
general characterization by H NMR spectroscopy.
(18) Timmerman, P.; Boerrigter, H.; Verboom, W.; Reinhoudt, D. N.
Recl. TraV. Chim. Pays-Bas 1995, 114, 103-111.
(19) Kimizuka, N.; Kawasaki, T.; Hirata, K.; Kunitake, T. J. Am. Chem.
Soc. 1998, 120, 4094-4104.
(20) For the synthesis of BisBAR, see: Lipkowski, P.; Bielejewska, A.;
Kooijman, H.; Spek, A. L.; Timmerman, P.; Reinhoudt, D. N. Chem.
Commun. 1999, 1311-1312. For the synthesis of calix[4]arene dimelamine
6, see ref 22. For the syntheses of HexCYA and TripCYA, see: Mathias,
J. P.; Seto, C. T.; Simanek, E. E.; Whitesides, G. M. J. Am. Chem. Soc.
1994, 116, 1725-1736.
Structural Variations in the Building Blocks. Structural
variations within the components can readily be introduced at
(15) Prins, L. J.; Huskens, J.; De Jong, F.; Timmerman, P.; Reinhoudt,
D. N. Nature 1999, 398, 498-502.
(16) Simanek, E. E.; Wazeer, M. I. M.; Mathias, J. P.; Whitesides, G.
M. J. Org. Chem. 1994, 59, 4904-4909.
(17) This is only the case when the substituents on each melamine are
identical.

3620 J. Am. Chem. Soc., Vol. 122, No. 15, 2000
Prins et al.
Chart 1
Scheme 1. Synthesis of Calix[4]arene Dimelamine 5
Subsequently, the assembly behavior of a series of closely
related barbituric acid derivatives, e.g. DMB, BA, and EPB,
was studied. Substitution of both ethyl for methyl groups at
position X has no effect on the assembly process. Assembly
23‚(DMB)6 shows two signals in the 13-16 ppm region of the
1H NMR spectrum and one resonance for the Ca and Cb carbons
at 141.24 ppm in the 13C NMR spectrum, proving the exclusive
formation of the D3 isomer. For assemblies 13‚(DMB)6 and 33‚
(DMB)6 similar results were obtained.
spectra for the assemblies 13‚(BA)6 and 33‚(BA)6, indicating
the formation of undefined assemblies. This must be due to a
lack of steric bulk at position X, which is essential for the
stability of double rosette assemblies. In contrast to this, the
1H NMR spectrum of assembly 23‚(BA)6 is sharp and very
similar to assemblies comprising DMB or DEB. The 13C NMR
spectrum shows a single resonance for the Ca and Cb carbons,
which confirms the presence of the D3 isomer. The reasons for
the very different assembly behavior of dimelamine 2 compared
to that of dimelamines 1 and 3 are not entirely clear, but they
1
Substitution of DEB by BA gives rise to broad H NMR

Structural Isomerism in NoncoValent H-Bonded Assemblies
Table 1. Summary of the Isomeric Distribution of All Assembliesa
J. Am. Chem. Soc., Vol. 122, No. 15, 2000 3621
barbiturates
cyanurates
calix[4]arene
dimelamine
DEB
DMB
BA
EPB
BisBAR
BuCYA
PheCYA
HexCYA
TripCYA
1
2
3
4
5
6
100/0/0
100/0/0
100/0/0
100/0/0
b
100/0/0
100/0/0
100/0/0
100/0/0
d
b
100/0/0c
100/0/0
100/0/0c
100/0/0
b
b
100/0/0
100/0/0
80/5/15
100/0/0
d
90/5/5
b
80/10/10
100/0/0
0/20/80
-
50/20/30
75/15/10
35/50/15
100/0/0
0/40/60
-
0/90/10
0/10/90
0/100/0
100/0/0
d
100/0/0
0/100/0
b
b
-
-
-
-
-
-
100/0/0
-
-
-
0/65/35
1
a The relative amounts of the D3/C3h/Cs isomers are calculated on the basis of the integrals of the signals in the 13-15 ppm region of the H
NMR spectra. All spectra were recorded at 1 mM concentration in CDCl3. b Broad signals were observed, indicating nondefined aggregation. c Due
to the unsymmetrical orientation of EPB, the two signals for the D3 isomer are accompanied by relatively broad signals. d Sharp signals were
observed, indicating well-defined assemblies, but the number of observed signals was higher than theoretically possible.
Figure 4. (a) 1H NMR spectrum and (b) part of the 13C NMR spectrum
of assembly 23‚(BisBAR)3. The arrows in the 13C NMR spectrum
Figure 3. (a) 1H NMR spectrum and (b) part of the 13C NMR spectrum
indicate the two signals for the Ca and Cb carbons.
of assembly 23‚(DEB)6. The arrow in the 13C NMR spectrum indicates
the signal for the Ca and Cb carbons.
formation of only the eclipsed isomers. The 1H NMR spectrum
of assembly 23‚(BisBAR)3 exhibits sharp signals (Figure 4a),
with two signals at 13.71 and 14.35 ppm. Together with the
two observed signals for the Ca and Cb carbons at 141.21 and
141.10 ppm in the 13C NMR spectrum (Figure 4b), this proves
most likely involve a combination of steric and electronic
factors, due to the presence of the NO2 substituents at position
R3.
Assembly 13‚(EPB)6 exhibits two main signals at 13.70 and
14.38 ppm (80% of the intensity) in the 13-16 ppm region of
the 1H NMR spectrum together with several small broad signals
(20% of intensity). The D3-symmetrical orientation of the EPBs
in which the phenyl groups are pointing away from the assembly
is sterically most likely. We assume that the other 20% of the
signals is attributed to staggered assemblies in which one or
several of the EPBs is oriented upside down, thus destroying
the D3 symmetry. Assembly 33‚(EPB)6 also exhibits two main
signals in the 13-16 ppm region (13.59 and 14.15 ppm), but
these account for only 60% of the intensity. The remaining 40%
is present as relatively broad signals, and the signals are
attributed to assemblies with one or several unsymmetrically
oriented EPBs. As for the assemblies with BA, dimelamine 2
exhibits a different assembly behavior with EPB compared to
dimelamines 1 and 3. Assembly 23‚(EPB)6 exhibits only two
signals (13.59 and 14.21 ppm) in the 13-16 ppm region. No
other signals are observed, showing the identical orientation of
all EPBs in assembly 23‚(EPB)6.
1
that the assembly is present as the C3h isomer. The H NMR
spectra of a 1:1 mixture of BisBAR with either dimelamine 1
or 3 show broad resonances, indicating nonspecific hydrogen
bonding.
The above results suggest that 5,5-disubstituted barbituric acid
derivatives preferentially form the D3 isomer (DEB, DMB,
EPB). Only when formation of the D3 isomer is sterically
impossible, e.g. with BisBAR, is formation of the C3h isomer
observed. These results can be rationalized by comparing the
position of the barbituric acid substituents in the D3 isomer and
in the C3h and Cs isomers. In the C3h and Cs isomers the two
barbituric acid moieties in the upper and lower rosette layer
are perfectly aligned, causing steric hindrance between the
5-alkyl substituents of both barbiturate fragments. In the D3
isomer this steric hindrance is relieved as a result of the
staggered orientation of the barbiturate components leaving
ample space for the 5-alkyl substituents.
Influence of the Isocyanuric Acid Substituents on the
Isomeric Distribution. Isocyanurates have a DAD‚DAD hy-
drogen bonding pattern similar to that of barbiturates, but both
the number and the orientation of their substituents is different
because they are attached to an sp2-hybridized nitrogen atom
To obtain information about the intrinsic stability of the C3h
and Cs isomers, the assembly behavior of compound BisBAR
was studied. The C3-spacer severely limits the relative freedom
of the two barbituric acid moieties and consequently allows the

3622 J. Am. Chem. Soc., Vol. 122, No. 15, 2000
Prins et al.
1
Figure 5. Part of the H NMR spectrum of assembly 33‚(BuCYA)6.
instead of an sp3-hybridized carbon as in the barbiturates. In
general, the stability of isocyanurate assemblies is higher since
the NH groups are stronger H-bond donors.21 The increase in
hydrogen bonding strength is reflected in the 1H NMR spectra
by a stronger downfield shift for the NHcya protons (typically
to 14-16 ppm) compared to the NHbarb protons (typically to
13-14 ppm) upon assembly.
Figure 6. Part of the 1H NMR spectrum of assemblies (a) 13‚
(HexCYA)6, (b) 23‚(HexCYA)6, and (c) 33‚(HexCYA)6. The signals
marked with an asterisk correspond to two protons indicated by the
integration. One of the signals in the spectrum of 23‚(HexCYA)6 is
visible as a shoulder of the signal at 14.17 ppm and is indicated by an
arrow.
The isomeric distribution was studied for assemblies compris-
ing isocyanurates BuCYA, PheCYA, HexCYA, and TripCYA.
These isocyanurates differ only in the nitrogen substituents,
which gradually increase in size.
1
The H NMR spectra of assemblies 13‚(BuCYA)6 and 23‚
(BuCYA)6 both show two signals in the 13-16 ppm region.
Furthermore, the presence of one single resonance for the Ca
and Cb carbons in the 13C NMR spectra of both assemblies
(120.55 and 141.10 ppm, respectively) leads to the conclusion
that in both cases exclusively the D3 isomer is formed. Replacing
the melamine butyl substituent with a benzyl group, as in
assembly 33‚(BuCYA)6, causes a small but significant change.
The 1H NMR spectrum exhibits two main signals at 14.72 and
14.19 ppm (80% of the intensity) attributed to the D3 isomer
(Figure 5), accompanied by a total of eight (two + six) sharp
signals (20% of the intensity) arising from the corresponding
C3h and Cs isomers, respectively. Assembly 33‚(BuCYA)6 is the
first example of a double rosette in which all three isomers are
observed. Remarkably, the larger size of the benzyl substituent
seems to favor the formation of eclipsed isomers.
Figure 7. Part of the 1H NMR spectrum of assembly 13‚(TripCYA)6.
the C3h isomer, whereas introduction of a nitro group at the Ca
and Cb carbons, as in assembly 23‚(HexCYA)6, stabilizes the
D3 isomer.
Assembly 13‚(PheCYA)6 forms almost quantitatively the D3
The increase in stability of the eclipsed isomers upon
increasing the size of the isocyanuric acid substituent becomes
even more pronounced for assemblies containing TripCYA,
which form mainly as the C3h isomer. The 1H NMR spectra of
13‚(TripCYA)6, 33‚(TripCYA)6, and 63‚(TripCYA)6 all exhibit
two main signals at 14.55/14.53, 14.43/14.34, and 14.53/14.44
ppm, respectively. The two signals observed for assemblies 13‚
(TripCYA)6 and 63‚(TripCYA)6 are accompanied by six ad-
ditional signals (10% and 35% of the intensity, respectively)
attributed to the Cs isomer (Figure 7). This once more shows
how sensitive the assembly process is to changes in the building
blocks, since even a seemingly small and negligible change such
as the elongation of the alkyl side chains already has a significant
effect on the isomeric distribution.
1
isomer according to the H and 13C NMR spectra. Assembly
33‚(PheCYA)6 is also formed mainly as the D3 isomer (80%),
but again the eclipsed isomers are observed (20%). Mixing of
dimelamine 2 and PheCYA in a 1:2 ratio leads to unspecific
hydrogen bonding. CPK models clearly show that, due to the
nitro substituents, the periphery of the assembly is sterically
too congested to form a double rosette assembly.
A further increase in the size of the isocyanuric acid
substituent leads to increased amounts of the C3h and Cs isomers.
Assemblies 13‚(HexCYA)6, 23‚(HexCYA)6, and 33‚(HexCYA)6
all show a total of 10 (two + two + six) signals, indicating
that all three isomers are present (Figure 6). The ratio between
the D3, C3h, and Cs isomers is 50:20:30, 75:15:10, and 35:50:
15, respectively, in the three assemblies. These values confirm
our earlier observations that replacement of a butyl for a benzyl
group, as in assembly 33‚(HexCYA)6, causes a preference for
Conclusive evidence for the fact that TripCYA exclusively
forms the C3h and Cs isomers was obtained from a covalent
capture experiment with assembly 63‚(TripCYA)6 via a ring-
closing metathesis (RCM) reaction using the Grubbs catalyst.22
Recently, we have demonstrated that assembly 63‚(DEB)6, which
(21) Mascal, M.; Fallon, P. S.; Batsanov, A. S.; Heywood, B. R.; Champ,
S.; Colclough, M. J. Chem. Soc., Chem. Commun. 1995, 805-806. Shieh,
H. S.; Voet, D. Acta Crystallogr., Sect. B 1976, 32, 3254-3260
(22) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413

Structural Isomerism in NoncoValent H-Bonded Assemblies
J. Am. Chem. Soc., Vol. 122, No. 15, 2000 3623
Scheme 2. Products of the Covalent Capture of Assembly 63‚(DEB)6 (Trimer 1123) and Assembly 63‚(TripCYA)6 (Monomer
12) by a RCM Reaction
is exclusively present as the D3 isomer, forms the cyclic trimer
11 in 96% yield, as proven by the signal at m/z ) 3100 (calcd
for C180H240N36O12 ) 3100) in the MALDI-TOF spectrum
(Scheme 2).23 The reason for this is that the alkene moieties
within one dimelamine fragment in assembly 63‚(DEB)6 are too
remote to react, while the alkene moieties of adjacent dimelamines
are in close proximity to each other.
In sharp contrast to this, a RCM reaction with assembly 63‚
(TripCYA)6 did not show any trace of the cyclic trimer 11.
Instead, the two alkene moieties of dimelamines react intramo-
lecularly to give cyclic monomer 12 as the only product, proven
by the signal at m/z ) 1034.1 (calcd for C60H80N12O4 ) 1033.4)
in the FAB-MS spectrum (Scheme 2). This experiment un-
equivocally proves that the two main signals at 14.53 and 14.44
Figure 8. Part of the 1H NMR spectrum of assembly 23‚(TripCYA)6.
ppm belong to the C3h instead of the D3 isomer. Unfortunately,
formation of the Cs isomer for assembly 23‚(TripCYA)6. On
in this case 13C NMR spectroscopy is not a suitable technique
the other hand, assembly 13‚(HexCYA)6 is formed as a mixture
to distinguish between the D3 and the C3h isomers, as it fails to
of all three isomers. Steric effects play an important role, as
illustrated by the effects of replacing a butyl for a benzyl
substituent on the dimelamine or the increase in size of the
isocyanurate substituents, but other factors including electronic
interactions and solvation effects most likely contribute to the
isomeric distribution as well. This has been clearly illustrated
by the peculiar assembly behavior of dimelamine 2 and the
bulky cyanurate TripCYA.
show two signals for the Ca and Cb carbons of assembly 13‚
(TripCYA)6, most probably due to a coincidental magnetic
equivalence.
Surprisingly, the introduction of a nitro substituent at position
R3 completely interchanges the stabilities of the C3h and Cs
isomers. Assembly 23‚(TripCYA)6 almost quantitatively (90%)
forms as the Cs isomer, as illustrated by the presence of six
1
signals in the 13-16 ppm region of the H NMR spectrum
Molecular Dynamics Studies. In an attempt to gain visual
(Figure 8).
insight into the assembly process, we performed a series of
The above results clearly show that the outcome of the
24
molecular dynamics (MD) simulations in OPLS CHCl3 with
assembly process is extremely diverse and mostly unpredictable.
assemblies 13‚(DEB)6, 13‚(HexCYA)6, and 13‚(TripCYA)6.
The isomeric distributions of the assemblies cover the whole
These assemblies were observed to form with very different
spectrum of possible isomeric distributions, ranging from the
isomeric distributions (D3 ) 100% for 13‚(DEB)6, D3/C3h/Cs )
quantitative formation of the D3 isomer for assemblies compris-
50/20/30% for 13‚(HexCYA)6, and C3h ) 95% for 13‚(Trip-
ing DEB or DMB to the exclusive assembly (90%) of the C3h
CYA)6, respectively), although they differ only in the barbiturate/
isomer for assembly 13‚(TripCYA)6, and a nearly quantitative
isocyanurate component. For each individual assembly, all three
(23) Cardullo, F.; Crego Calama, M.; Snellink-Rue¨l, B. H. M.; Weidmann,
J.-L.; Bielejewska, A.; Fokkens, R.; Nibbering, N. M. M.; Timmerman, P.;
Reinhoudt, D. N. Chem. Commun. 2000, 367-368.
(24) Jorgensen, W. L.; Briggs, J. M.; Contreras, M. L. J. Phys. Chem.
1990, 94, 1683.

3624 J. Am. Chem. Soc., Vol. 122, No. 15, 2000
Prins et al.
isomers were generated starting from the X-ray crystal structure
of assembly 23‚(DEB)6 using Quanta 97.25 The MD calculations
were carried out with CHARMm version 2426 for a period of
250 ps, after which the structural integrity of the assemblies
was examined. The flatness of the rosette motif, rotation of one
or more of the melamine/barbiturate fragments out of the rosette
motif, and the number of intact hydrogen bonds are regarded
as qualitative tools to judge the stability of each of the
isomers.27,28
After 250 ps, only the D3 isomer of assembly 13‚(DEB)6 is
structurally intact (Figure 9a). In both the C3h and Cs isomers,
barbiturate or melamine fragments have rotated away from the
assembly, which is clearly visible (Figure 9b,c). This is most
likely due to repulsion of the ethyl groups. These results are in
agreement with our experimental observation that 5,5-disubsti-
tuted barbituric acid derivatives exclusively form the D3 isomer.
A similar analysis of the three isomers of assembly 13‚
(TripCYA)6 after 250 ps shows that only the C3h isomer has
34-36 intact hydrogen bonds and does not show any sign of
destruction. In the Cs isomer, steric interactions between the
two TripCYAs in the most crowded part of the assembly force
an almost 90° deviation of planarity in the rosette motifs. The
rosette motifs in the D3 isomer deviate strongly from planarity
as well, but in this case the deviation is due to destructive
interactions with solvent molecules. Early in the simulation, two
chloroform molecules penetrate between the rosette motifs,
breaking the assembly apart (Figure 10). Apparently, the
solvation of assembly 13‚(TripCYA)6 in CHCl3 plays an
important role in determining the isomeric distribution, with
the interior of the D3 isomer being more accessible to solvent
molecules than the eclipsed isomers.
For assembly 13‚(HexCYA)6, all three isomers show nearly
flat rosette motifs with 34-36 intact hydrogen bonds after 250
ps. No signs of disruption are observed, while in general the
disruptive processes began early in the simulation (0-50ps).
The MD simulations presented here contribute significantly
to the visualization of how different parameters affect the
isomeric distribution. These studies show that the origin of the
destabilization of specific isomers either may be present within
the assembly, e.g. steric hindrance, or can arise from direct
solvation effects. A quantitative analysis of the MD simulations
as well as the use of MD to predict the stability of hydrogen-
bonded assemblies in general is currently under investigation.
From the assembly experiments described in this paper, it is
evident that small structural variations in the molecular com-
ponents have an unpredictable effect on the outcome of the
assembly process. Although the outcome of the assembly in
most cases can be rationalized via MD simulations, it is far
more complicated to predict the outcome of the assembly
process. This is most clearly illustrated by the opposite effect
of increasing steric bulk on the barbiturate and isocyanurate
assemblies. Whereas an increase in size of the isocyanurate
substituent leads to preferential formation of the eclipsed
isomers, the same steric parameter is used to explain induction
Figure 9. Structures of the (a) D3, (b) C3h, and (c) Cs isomers of
assembly 13‚(DEB)6 after MD runs of 250 ps (top views). Nonpolar
hydrogens have been omitted for clarity. The arrows depict places in
the assemblies where the structural integrity is lost, due to the expulsion
of cyanurate (b) and melamine fragments (c).
(25) Quanta 97, Molecular Simulations, Waltham, MA, 1997.
(26) (a) Brooks, B. R.; Bruccoleri, R. E.; Olafsen, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983, 4, 187-217. (b)
Momany, F. A.; Klimkowski, V. J.; Scha¨fer, L. J. Comput. Chem. 1990,
11, 654-662. (c) Momany, F. A.; Rone, R.; Kunz, H.; Frey, R. F.; Newton,
S. Q.; Scha¨fer, L. J. Mol. Struct. 1993, 286, 1.
(27) Chin, D.; Gordon, D. M.; Whitesides, G. M. J. Am. Chem. Soc.
1994, 116, 12033-12044.
(28) The number of intact hydrogen bonds for stable structures ranges
from 34 to 36, due to the fact that occasionally one or several of the NHmels
in the outer ring of the rosette motif temporarily rotate away.
of the D3 isomer for barbiturates. In addition to this, electronic
interactions and solvation effects are even more difficult to
predict.
Controlling the Isomeric Distribution via Chiral Substit-
uents. The final part of this paper describes a series of
experiments demonstrating that the introduction of chiral centers

Structural Isomerism in NoncoValent H-Bonded Assemblies
J. Am. Chem. Soc., Vol. 122, No. 15, 2000 3625
Figure 10. Structure of the D3 isomer of assembly 13‚(TripCYA)6 after
a MD simulation of 250 ps (top view). Nonpolar hydrogens have been
omitted for clarity. Two solvent molecules (van der Waals representa-
tion) have penetrated between the rosette layers, causing a disruption
of the assembly.
in the building blocks can be used to control the assembly
process.15 As an example we use assembly 33‚(HexCYA)6,
which is present as a mixture of the D3, C3h, and Cs isomers in
a 50:20:30 ratio (vide supra, Figure 6a). Substitution of the
achiral benzyl substituents in 3 for (R)-1-phenylethyl groups,
as in 4, completely changes the outcome of the assembly
process. Only the D3 isomer is formed, as concluded from the
Figure 11. (a) Orientation of the chiral substituents of dimelamine 4
with respect to the rosette core as elucidated from 2D NOESY
measurements (NOESY enhancements are indicated by arrows). (b)
Corresponding orientation of the chiral substituents of dimelamine 5.
1
presence of the two signals at 13.92 and 14.61 ppm in the H
NMR spectrum of assembly 43‚(HexCYA)6 (Figure 12a).29 The
reason for this behavior is that the chiral substituents have a
strongly preferred orientation with respect to the rosette core
forms the C3h and Cs isomers, whereas DEB and EPB are
observed to form only the D3 isomer, as a result of steric effects
(vide supra). This results in obstruction of the double rosette
assembly process, leading to the formation of nondefined
assemblies.
1
(Figure 11a) and are unable to rotate, as revealed by 2D H
NMR ROESY measurements of assembly 43‚(DEB)6.14
Only when the assembly is formed as the D3 isomer can both
chiral substituents adopt the preferred orientation. Similarly,
assemblies 43‚(DEB)6 (Figure 12b), 43‚(DMB)6, 43‚(EPB)6, 43‚
(BuCYA)6, 43‚(TripCYA)6 and 43‚(PheCYA)6 (Figure 12c) also
selectively form the D3 isomer, even though in some cases the
corresponding assembly with achiral dimelamine 3 favors the
eclipsed isomers.
The assembly studies with dimelamine 5 and either DMB,
BuCYA, or TripCYA show that the inducing effects of the chiral
substituents may be occasionally overruled by other factors. All
three assemblies exhibit sharp signals in the 13-16 ppm region
1
of the H NMR spectrum, indicating the formation of well-
defined assemblies. However, the number of observed signals
is much larger (15, 14, and 16, respectively) than theoretically
possible (12) when all isomers are present. The reason for this
assembly behavior remains unclear.
In summary, we can state that chiral substituents can have a
strong directing effect on the isomeric distribution of hydrogen-
bonded assemblies. The handedness of chiral groups (i.e. (R)
or (S)) can generate an energetic difference between the
staggered and eclipsed isomers, resulting in preferential forma-
tion of the D3 isomer in the case of dimelamine 4 and
preferential formation of the C3h and Cs isomer in the case of
dimelamine 5.
Quantitative formation of the C3h and Cs isomers can be
achieved with meso dimelamine 5, in which the two melamines
carry substituents of opposite chirality. The 1H NMR spectrum
of assembly 53‚(HexCYA)6 exhibits eight signals in the 13-16
ppm region, which proves the exclusive formation of the
eclipsed isomers (Figure 12d). No trace of the staggered isomer
is observed. This result can similarly be rationalized by
examination of the orientation of the chiral substituents. Only
for the eclipsed isomers of this assembly can both chiral
substituents adopt the energetically most favorable orientation
(Figure 11b). Similarly, the selective formation of the C3h and
Cs isomers was also observed for assembly 53‚(PheCYA)6
(Figure 12e).
Conclusions
The 1H NMR spectra for mixtures of dimelamine 5 and DEB
or EPB show only broad signals (Figure 12f), indicating the
formation of undefined assemblies. In these cases the inducing
effects of the chiral substituents in 5 oppose the inducing effect
of the DEB and EPB components. Dimelamine 5 preferentially
In this paper we show that the isomeric distribution of
noncovalent hydrogen-bonded assemblies can be controlled in
a predictable manner via the introduction of chiral substituents
in the assembly components. We describe the results of a
systematic study on the assembly of a series of calix[4]arene
dimelamines with barbituric acid and isocyanuric acid deriva-
(29) As a result of the presence of the (R)-R-methylbenzyl substituents,
the C3h isomer no longer contains a σh-plane, and therefore four signals are
expected for the C3h isomer instead of two.
1
tives. Analysis of the results using H and 13C NMR spectros-

3626 J. Am. Chem. Soc., Vol. 122, No. 15, 2000
Prins et al.
Figure 12. Part of the 1H NMR spectra of assemblies (a) 43‚(HexCYA)6, (b) 43‚(DEB)6, (c) 43‚(PheCYA)6, (d) 53‚(HexCYA)6, (e) 53‚(PheCYA)6,
and (f) 53‚(DEB)6.
was removed under reduced pressure to give a yellow oil which was
copy together with MD simulation studies clearly shows that
the isomeric distribution between D3, C3h, and Cs isomers is
determined by a combination of different factors, i.e., steric,
electronic, and solvation effects. For most assemblies this makes
an a priori prediction of the outcome of the assembly process
virtually impossible. However, this changes when chiral building
blocks are used. It has been shown for the first time that chiral
substituents present in the molecular components determine the
isomeric distribution of the assembly in a rational manner. We
believe that the role of chirality in controlling self-assembly
processes will become increasingly important.
purified by column chromatography (SiO2; EtOAc/hexane ) 1:3).
1
Compound 8 was obtained as a white solid (0.70 g, 67%). H NMR
(250 MHz, CDCl3): δ 7.4 (br s, 1H, NH), 6.9-6.8 (m, 4H, ArH), 6.7-
6.6 (m, 2H, ArH), 6.10 (s, 2H, ArH), 5.68 (s, 2H, ArH), 4.43 and 3.15
2
2
(ABq, 4H, J ) 16.5 Hz, ArCH2Ar), 4.38 and 3.05 (ABq, 4H, J )
16.1 Hz, ArCH2Ar), 4.0-3.8 (m, 4H, OCH2), 3.7-3.5 (m, 4H, OCH2),
2.0-1.7 (m, 8H, OCH2CH2), 1.43 (s, 9H, C(CH3)3), 1.05 and 0.90 (2t,
12H, J ) 9.0 Hz). MS (FAB): m/z 722.9 ([M + H+], calcd 723.4).
3
Anal. Calcd for C45H58N2O6‚0.5H2O: C, 73.8; H, 8.1; N, 3.8. Found:
C, 74.0; H, 8.2; N, 3.9.
5-N-[4-Amino-6-(R)-1-phenylethylamino-1,3,5-triazin-2-yl]-amino-
17-N-(tert-butoxycarbamino)-25,26,27,28-tetrapropoxycalix[4]-
arene (9). To an ice-cold solution of 8 (0.50 g, 0.69 mmol) and DIPEA
(0.21 mL, 1.21 mmol) in THF (40 mL) was added cyanuric chloride
(0.13 g, 0.70 mmol), and the solution was stirred for 3 h at 0 °C. A
stream of ammonia gas was then bubbled through the solution for a
period of 2 h, while the temperature was maintained at 0 °C. The solvent
was removed under reduced pressure, and the residue was partioned
between CH2Cl2 (50 mL) and water (50 mL). The organic layer was
washed with water (20 mL) and brine (20 mL) and then dried (MgSO4),
and the solvent was removed under reduced pressure. Part of the white
solid (0.45 g, 0.53 mmol) was redissolved in THF (10 mL), and
subsequently DIPEA (0.83 mL, 4.75 mmol) and (R)-R-methylbenzyl-
amine (0.70 mL, 5.43 mmol) were added. The mixture was heated at
reflux for 2 days, after which the solvent was removed under reduced
pressure. The residue was redissolved in CH2Cl2 (50 mL), washed with
water (50 mL) and brine (50 mL), and subsequently dried (Na2SO4).
The solvent was removed under reduced pressure, and the resulting
yellow solid was purified by column chromatography (silica, EtOAc)
to yield 9 as a white solid (0.81 g, 64%). 1H NMR (250 MHz, DMSO-
d6): δ 9.1 (br s, 1H, NH), 8.6 (br s, 1H, NH), 7.4-7.2 (m, 9H, ArH +
Experimental Section
THF was freshly distilled from Na/benzophenone, EtOAc, and
hexane (referring to petroleum ether fraction with bp 60-80 °C) from
K2CO3 and CH2Cl2 from CaCl2. All chemicals were of reagent grade
and used without further purification. NMR spectra were recorded on
a Bruker AC 250 (250 MHz) or a Varian Unity 300 (1H NMR 300
MHz) spectrometer at room temperature. Residual solvent protons were
used as internal standard, and chemical shifts are given relative to
tetramethylsilane (TMS). FAB-MS spectra were recorded with a
Finnigan MAT 90 spectrometer with m-nitrobenzyl alcohol (NBA) as
a matrix. EI mass spectra were recorded on a Finnigan MAT 90
spectrometer with an ionizing voltage of 70 eV. Elemental analyses
were performed using a Carlo Erba EA1106. The presence of solvents
1
in the analytical samples was confirmed by H NMR spectroscopy.
Flash column chromatography was performed using silica gel (SiO2,
E. Merck, 0.040-0.063 mm, 230-240 mesh).
5,17-N,N′-Bis[4-amino-6-(R)-1-phenylethylamino-1,3,5-triazin-2-
yl]diamino-25,26,27,28-tetrapropoxycalix[4]arene (4). This com-
pound was obtained as a white solid (0.58 g, 65%) in a manner similar
to the previously reported dimelamine calix[4]arene 113b starting from
5,17-N,N′-bis[4-amino-6-chloro-1,3,5-triazin-2-yl]diamino-25,26,27,28-
tetrapropoxycalix[4]arene13b (0.75 g, 0.85 mmol), (R)-R-methylbenzyl-
amine (4.0 mL, 31 mmol), and DIPEA (1.8 mL, 10.3 mmol) in THF
(10 mL). 1H NMR (250 MHz, DMSO-d6): δ 8.5 (br s, 1H, NH), 7.4-
7.2 (m, 9H, ArH + NH), 6.2-6.1 (m, 10H, ArH + NH2), 5.26 (br s,
3
NH), 6.3-6.2 (m, 8H, ArH + NH2), 5.27 (t, 1H, J(H,H) ) 6.13 Hz,
CHCH3), 4.33 and 3.06 (ABq, 8H, 2J(H,H) ) 12.8 Hz, ArCH2Ar), 3.90
and 3.63 (2t, 8H, 3J(H,H) ) 7.9 Hz, OCH2), 2.0-1.7 (m, 8H,
OCH2CH2), 1.49 (s, 9H, C(CH3)3), 1.22 (d, 3H, 2J ) 6.13 Hz,CHCH3),
3
1.09 and 0.89 (t, 12H, J ) 7.4 Hz, OCH2CH2CH3). MS (FAB): m/z
936.6 ([M + H+], calcd 936.5). Anal. Calcd for C56H69O6N7: C, 71.77;
2
H, 7.53; N, 10.46. Found: C, 72.15; H, 7.32; N, 10.68.
1H, CHCH3), 4.32 and 3.02 (ABq, 8H, J(H,H) ) 12.8 Hz, ArCH2-
3
5-N-[4-Amino-6-(R)-1-phenylethylamino-1,3,5-triazin-2-yl]amino-
17-amino-25,26,27,28-tetrapropoxycalix[4]arene (10). Trifluoroacetic
acid (3 mL) was added dropwise to a solution of 9 (0.41 g, 0.44 mmol)
in CH2Cl2 (30 mL) at 0 °C. The mixture was warmed to room
temperature and stirred overnight. Water (20 mL) was then added,
followed by portions of solid NaHCO3 until the solution reached pH
7. Subsequently, the mixture was separated, and the organic layer was
washed with aqueous NaOH (1 M, 2 × 10 mL) and brine (10 mL) and
dried on Na2SO4. The solvent was removed under reduced pressure to
yield 10 as a white solid (0.34 g, 93%). 1H NMR (250 MHz, DMSO-
d6): δ 8.5 (br s, 1H, NH), 7.4-7.2 (m, 10H, ArH), 6.3-6.2 (m, 10H,
ArH + NH + NH2), 5.26 (br s, 1H, CHCH3), 4.54 (br s, 2H, NH2),
4.29 and 3.00 (ABq, 8H, 2J(H,H) ) 13.2 Hz, ArCH2Ar), 3.9-3.6 (m,
Ar), 3.89 and 3.63 (2t, 8H, J(H,H) ) 8.3 Hz, OCH2), 2.1-1.8 (m,
2
8H, OCH2CH2), 1.42 (d, 3H, J ) 6.1 Hz,CHCH3), 1.09 and 0.89 (t,
3
12H, J ) 7.4 Hz, OCH2CH2CH3). MS (FAB): m/z 1049.6 ([M +
H+], calcd 1049.6). Anal. Calcd for C62H72O4N12‚0.2CH3OH: C, 70.76;
H, 6.95; N, 15.92. Found: C, 70.56; H, 6.83; N, 15.85.
5-Amino-17-N-(tert-butoxycarbamino)-25,26,27,28-tetrapropoxy-
calix[4]arene (8). A mixture of 718 (0.90 g, 1.45 mmol), BOC-ON
(0.36 g, 1.45 mmol), and triethylamine (0.40 mL, 2.87 mmol) in DMF
(45 mL) was heated at 50 °C under an argon atmosphere for 3 h and
then cooled to room temperature, after which the solvent was removed
under reduced pressure. The residue was partitioned between CH2Cl2
(100 mL) and aqueous NaOH (1 M, 50 mL), and the aqueous portion
was extracted with CH2Cl2 (2 × 25 mL). The combined organic
solutions were washed with aqueous NaOH (1 M, 2 × 50 mL), water
(50 mL), and brine (50 mL) and then dried (MgSO4), and the solvent
2
8H, OCH2), 2.0-1.8 (m, 8H, OCH2CH2), 1.43 (d, 3H, J ) 6.1 Hz,
CHCH3), 1.05 and 0.91 (t, 12H, 3J ) 7.4 Hz, OCH2CH2CH3). MS

Structural Isomerism in NoncoValent H-Bonded Assemblies
J. Am. Chem. Soc., Vol. 122, No. 15, 2000 3627
(FAB): m/z 837.2 ([M + H+], calcd 837.1). Anal. Calcd for
C51H61O4N7‚0.2H2O: C, 72.95; H, 7.37; N, 11.68. Found: C, 72.87;
H, 7.30; N, 11.36.
athesis (RCM) reactions were performed in toluene-d8 and were
monitored by H NMR spectroscopy. Conditions in toluene-d8: room
1
temperature; 48 h; [6]0 ) 6.0 mM; [TripCYA]/[6]t)0 ) 2.5; [catalyst]t)0
) 0.6 mol %; reactions were quenched by bubbling with oxygen.
Molecular Dynamics Calculations. Initial structures, created by
manual modification of the X-ray crystal structure of assembly
23‚(DEB)6,13b as well as visualizations were carried out with Quanta
97.25 The MD calculations were run with CHARMm, version 24.0.26
Parameters were taken from Quanta 97, and point charges were assigned
with the charge template option in Quanta/CHARMm; excess charge
was smoothened, rendering overall neutral residues. The structures were
placed in a cubic box of approximately 66.0 Å dimensions, initially
filled with 2136 OPLS chloroform.25 Solvent molecules that overlap
with the complexes were removed (based on heavy atom interatomic
distances of 2.3 Å). Full periodic boundary conditions were imposed.
Before the MD simulations were run, the system was minimized by
steepest descent, to remove the worst contacts, until the rms on the
energy gradient was e1.0 kcal‚mol-1‚Å-2 or a maximum of 1000 steps
was reached.
During the simulation the nonbonded list was updated every 20 time
steps with a cutoff of 12 Å. The van der Waals interactions were treated
with a switch function between 10 and 11 Å, whereas the shift function
was applied to the electrostatic interactions (cutoff 11 Å). A constant
dielectric constant and an ꢀ value of 1 were applied. The system was
heated to 300 K in 5 ps, followed by 10 ps for equilibration with scaling
of the velocities within a temperature window of 10 deg. After
equilibration, no scaling of the velocities was applied. The production
phase lasted 250 ps, and coordinates were saved regularly for subsequent
analysis. The verlet/leapfrog algorithm was used for the numerical
integration. The SHAKE algorithm30 on bonds involving hydrogen was
applied, allowing a time step of 1 fs.
5-N-[4-Amino-6-(R)-1-phenylethylamino-1,3,5-triazin-2-yl]-
amino,17-N-[4-amino-6-(S)-1-phenylethylamino-1,3,5-triazin-2-yl]-
amino-25,26,27,28-tetrapropoxycalix[4]arene (5). A procedure similar
to that described for the synthesis of 9 was used, starting from 10 (0.34
g, 0.41 mmol), cyanuric chloride (82 mg, 0.44 mmol), and DIPEA (0.2
mL, 1.14 mmol) in THF (25 mL). The intermediate component (0.5 g,
0.5 mmol), (S)-R-methylbenzylamine (0.67 mL, 5.2 mmol), and DIPEA
(0.92 mL, 5.3 mmol) were refluxed in THF (5 mL) to give 5 as a
white solid (0.26 g, 43%) after column chromatography (SiO2, CH2-
1
Cl2:MeOH:(25% aqueous NH3) ) 99:9.5:0.5). H NMR (250 MHz,
DMSO-d6): δ 8.5 (br s, 1H, NH), 7.4-7.2 (m, 9H, ArH + NH), 6.2-
6.1 (m, 10H, ArH + NH2), 5.26 (br s, 1H, CHCH3), 4.32 and 3.02
(ABq, 8H, 2J(H,H) ) 12.8 Hz, ArCH2Ar), 3.89 and 3.63 (2t, 8H,
3J(H,H) ) 8.3 Hz, OCH2), 2.1-1.8 (m, 8H, OCH2CH2), 1.42 (d, 3H,
3
2J ) 6.1 Hz, CHCH3), 1.09 and 0.89 (t, 12H, J ) 7.4 Hz, OCH2-
CH2CH3). MS (FAB): m/z 1049.7 ([M + H+], calcd 1049.6). Anal.
Calcd for C62H72O4N12‚CH3OH: C, 69.98; H, 7.08; N, 15.54. Found:
C, 69.72; H, 7.20; N, 15.27.
N-Butyl-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (BuCYA). BuCYA
was prepared by following a literature procedure to give 5-substituted
cyanuric acid derivatives.19 1H NMR (300 MHz): δ 11.36 (s, 2H, NH),
3.61 (m, 2H, NCH2), 1.44 (m, 2H, NCH2CH2), 1.26 (m, 2H, NCH2-
CH2CH2), 0.86 (t, 3H, 3J ) 7.3 Hz, CH3). MS (FAB): m/z 186.1 ([M
+ H+], calcd 186.1). Anal. Calcd for C7H11O3N3‚0.1CH3OH: C, 45.27;
H, 6.10; N, 22.31. Found: C, 44.92; H, 6.02; N, 22.31.
N-(p-tert-Butylphenyl)-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (Phe-
CYA). PheCYA was prepared by following a literature procedure to
give 5-substituted cyanuric acid derivatives.19 1H NMR (300 MHz,
DMSO-d6) δ 11.51 (s, 2H, NH), 7.45 (d, 2H, 2J ) 8.4 Hz, ArH), 7.22
(d, 2H, 2J ) 8.4 Hz, ArH) 1.30 (s, 9H, C(CH3)3). MS (EI): m/z 261.9
([M + H+], calcd 262.1). Anal. Calcd for C13H15N3O3: C, 59.8; H,
5.8; N, 15.8. Found: C, 59.7; H, 5.8; N, 15.8.
Acknowledgment. Prof. R. Glaser is gratefully acknowl-
edged for helpful discussions. Dr. F. Cardullo is acknowledged
for providing compound 6 and for assistance with the RCM
experiments. Dr. P. Lipkowski is acknowledged for providing
compound BisBAR. This research has been financially sup-
ported by the Council for Chemical Sciences of The Netherlands
Organization for Scientific Research (CW-NWO) (L.J.P) and
the Japan Science and Technology Corp. (Chemotransfiguration
Project) (K.A.J.).
Assembly Preparation. Typically, 1 mM solutions of assemblies
were prepared either by stirring the calix[4]arene dimelamine and
barbiturate/cyanurate components overnight in 1.0 mL of CDCl3 or by
dissolving both components in THF, evaporating THF, drying under
high vacuum, and subsequently resolvating in CDCl3.
RCM Reaction on Assembly 63‚(TripCYA)6. Ring-closing meth-
(30) Berendsen, H. J. C.; Postma, J. P. M.; Dinola, A.; van Gunsteren,
W. F.; Haak, J. R. J. Chem. Phys. 1984, 81, 3684-3690.
JA9936262