A highly stable, six-hydrogen-bonded molecular duplex

Article abstract of DOI:10.1021/ja9942742

This paper describes the design, synthesis, and characterization of a hydrogen-bonded molecular duplex (3·4). Two oligoamide molecular strands, 3 and 4, with the complementary hydrogen-bonding sequences ADAADA and DADDAD, respectively, were found to form

Full text of DOI:10.1021/ja9942742

J. Am. Chem. Soc. 2000, 122, 2635-2644
2635
A Highly Stable, Six-Hydrogen-Bonded Molecular Duplex
Huaqiang Zeng, Rebecca S. Miller, Robert A. Flowers, II, and Bing Gong*
Contribution from the Department of Chemistry, The UniVersity of Toledo, Toledo, Ohio 43606
ReceiVed December 6, 1999
Abstract: This paper describes the design, synthesis, and characterization of a hydrogen-bonded molecular
duplex (3‚4). Two oligoamide molecular strands, 3 and 4, with the complementary hydrogen-bonding sequences
ADAADA and DADDAD, respectively, were found to form an extremely stable (Ka ) (1.3 ( 0.7) × 109
1
M-1) molecular duplex (3‚4) in chloroform. Evidence from 1D and 2D H NMR spectroscopy, isothermal
titration calorimetry, and thin-layer chromatography confirmed the formation and the high stability of the
duplex. The exceptional stability is explained by positive cooperativity among the numerous hydrogen-bonding
and van der Waals interactions and the preorganization of the individual strands by intramolecular hydrogen
bonds. This design has opened a new avenue to supramolecular recognition units with programmable specificities
and stabilities.
Introduction
Ghadiri et al.9 designed dimeric structures with eight hydrogen
bonds using modified cyclic peptides. Extremely stable dimers
based on guanine were reported recently by Sessler et al.10
Rebek et al.11 designed concave molecules that assembled into
structures reminiscent of tennis balls.
Currently there is intense interest in constructing supramo-
lecular structures through the self-assembly of both biological
and synthetic systems.1 In Nature, the cooperative action of
numerous noncovalent forces leads to highly specific molecular
recognition events. As a result, nanoscale supramolecular
structures are ubiquitous in Nature. On the other hand, the
preparation of nanostructures with even the best present-day
synthetic methods still presents one of the most daunting
challenges. Based on noncovalent interactions, particularly
hydrogen-bonding interactions, numerous artificial self-assembly
systems have been developed.1b For example, Whitesides et al.1c
described multicomponent structures based on the cyanuric
acid-melamine motif. The groups of Zimmerman2 and Meijer3
reported heterocyclic complexes with arrays of hydrogen-bond
donors (D) and acceptors (A). Hamilton et al.4 developed
complexes based on the 2-aminopyridinecarboxylic acid system.
By combining multiple hydrogen bonds, Zimmerman,2b Lehn,5
and Mascal6 described the creation of cyclic self-assembling
structures. In addition, Reinhoudt et al.7 and de Mendoza et
al.8 have shown the assembly of calix[4]arene derivatives.
A most well-known and intensively studied example of self-
assembly involves the formation of the DNA double helix.12
The pairing of nucleic acid strands is perhaps the most elegant
example of self-assembly in Nature. Nucleobase complemen-
tarity allows the specification of intermolecular association of
nucleic acids, by “sticky-ended” association.13 DNA is the
molecule with the most readily predictable and programmable
intermolecular interactions. The unfavorable entropy price paid
for the association of two strands of DNA is compensated by
numerous noncovalent interactions that act cooperatively. DNA
assembly is characterized by sequence specificity and positive
cooperativity. By taking advantage of the programmable
sequence specificity of duplex DNA, Seeman has described an
elegant strategy for the construction of a variety of DNA-based
nanostructures.14 Although duplex DNA is able to specify
intermolecular associations, the conjugation of DNA molecules
with other types of structural units, such as peptide and unnatural
structures, is not always a trivial task synthetically. In addition,
the requirement for using DNA in aqueous media is incompat-
ible with many nonbiological applications. Therefore, if a diverse
* To whom correspondence should be addressed.
(1) For reviews, see: (a) Lindsey, J. S. New J. Chem. 1991, 15, 153-
180. (b) Lawrence, D. S.; Jiang, T.; Levett, M.; Chem. ReV. 1995, 95, 2229.
(c) Whitesides, G. M.; Simanek, E. E.; Mathias, J. P.; Seto, C. T.; Chin, D.
N.; Mammen, M.; Gordon, D. M. Acc. Chem. Res. 1995, 28, 37-44. (d)
Philip, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1155. (e)
Conn, M. M.; Rebek, J., Jr. Chem. ReV. 1997, 97, 1647. (f) de Mendoza,
J. Chem.sEur. J. 1998, 4, 1373. (g) Sijbesman, R. P.; Meijer, E. W. Curr.
Opin. Colloid Interface Sci. 1999, 4, 24.
(7) Vreekamp, R. H.; van Duynhoven, J. P. M.; Hubert, M.; Verboom,
W.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 1996, 35, 1215-1218.
(8) Gonza´lez, J. J.; Prados, P.; de Mendoza, J. Angew. Chem., Int. Ed.
Engl. 1999, 38, 525.
(2) (a) Corbin, P. S.; Zimmerman, S. C. J. Am. Chem. Soc. 1998, 120,
9710-9711. (b) Kolotuchin, S. V.; Zimmerman, S. C. J. Am. Chem. Soc.
1998, 120, 9092-9093.
(3) (a) Beijer, F. H.; Sijbesma, R. P.; Kooijman, H.; Spek, A. L.; Meijer,
E. W. J. Am. Chem. Soc. 1998, 120, 6761-6769. (b) Beijer, F. H.;
Kooijman, H.; Spek, A. L.; Sijbesma, R. P.; Meijer, E. W. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 75-78. (c) Folmer, B. J. B.; Sijbesma, R. P.;
Kooijman, H.; Spek, A. L.; Meijer, E. W. J. Am. Chem. Soc. 1999, 121,
9001-9007.
(4) Yang, J.; Fan, E.; Geib, S. J.; Hamilton, A. D. J. Am. Chem. Soc.
1993, 115, 5314-5415.
(5) Marsh, A.; Silvestri, M.; Lehn, J.-M. Chem. Commun. 1996, 1527-
1528.
(6) Mascal, M.; Hext, N. M.; Warmuth, R.; Moore, M. H.; Turkenburg,
J. P. Angew. Chem., Int. Ed. Engl. 1996, 35, 2204-2206.
(9) (a) Ghadiri, M. R.; Kobayashi, K.; Granja, J. R.; Chadha, R. K.;
McRee, D. E. Angew. Chem., Int. Ed. Engl. 1995, 34, 93. (b) Kobayashi,
K.; Granja, J. R.; Ghadiri, M. R. Angew. Chem., Int. Ed. Engl. 1995, 34,
95.
(10) Sessler, J. L.; Wang, R. Angew. Chem., Int. Ed. Engl. 1998, 37,
1726.
(11) (a) Szabo, T.; Hilmersson, G.; Rebek, J., Jr. J. Am. Chem. Soc. 1998,
120, 6193-6194. (b) Rebek, J., Jr. Acc. Chem. Res. 1999, 33, 278-286.
(12) (a) Saenger, W. Principles of Nucleic Acid Structure; Springer: New
York, 1984. (b) Blackburn, G. M. In Nucleic Acids in Chemistry and
Biology; Blackburn, G. M., Gait, M. J., Ed.; IRL Press: Oxford, U.K., 1990;
p 17.
(13) Cohen, S. N.; Chang, A. C. Y.; Boyer, H. W.; Helling, R. B. Proc.
Natl. Acad. Sci. U.S.A. 1973, 70, 3240-3244.
(14) Seeman, N. Acc. Chem. Res. 1997, 30, 357-363.
10.1021/ja9942742 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/02/2000

2636 J. Am. Chem. Soc., Vol. 122, No. 11, 2000
Zeng et al.
Figure 1. Oligoamides consisting of the three building blocks
conjugated by proper linkers, leading to molecular strands with all
possible combinations of hydrogen-bond donors and acceptors.
tions that may have led to the formation of polymeric aggregates.
The self-complementary dimers of 1 and 2 showed similar
stabilities (104 M-1 < Kdimer < 105 M-1) in chloroform. The
stabilities of these hydrogen-bonded duplexes are thus sequence-
independent. This is in sharp contrast to the case of previously
reported hydrogen-bonded dimers, whose stability depended on
the particular arrangement of hydrogen-bond donor and acceptor
sites due to secondary electrostatic interactions.18 In fact, the
dimers of 1 and 2 represent the first examples of unnatural
hydrogen bonded duplexes free of secondary interactions. This
system inaugurated the design and systematic study of self-
assembling duplexes presenting cooperativity and adjustable
specificity.
set of readily modifiable structural units leading to highly
specific intermolecular interactions become available, the
development of artificial nanostructures will be greatly facili-
tated.
We are interested in developing unnatural molecular recogni-
tion units with programmable strength and specificity similar
to those demonstrated by DNA. The underlying strategy involves
the combination of the simplest hydrogen-bond donors and
acceptors, i.e., amide O and H atoms. As shown in Figure 1,
linking building blocks derived from 3-aminobenzoic acid, 1,3-
benzenedicarboxylic acid (isophthalic acid), and 1,3-diami-
nobenzene (1,3-phenylenediamine) with proper three- or five-
atom linkers should generate oligoamides with all possible
combinations (sequences) of donors and acceptors. The number
N of duplexes with n intermolecular hydrogen-bonding sites can
be calculated on the basis of the following equations:
To develop the molecular system represented by 1 and 2 into
a more general and predictable platform for designing molecular
recognition units, oligoamides 3 and 4, with the complementary
hydrogen-bonding sequences DADDAD and ADAADA, were
designed and synthesized. The self-assembly of 3 and 4 into a
N ) 2n-2 + 2(n-3)/2
N ) 2n-2 + 2(n-2)/2
when n ) odd number
(1)
when n ) even number (2)
When n is an odd number, only complementary (or hetero)
duplexes are possible. With n being an even number, both
complementary and self-complementary duplexes become pos-
sible. The first term in the second equation is the number of
complementary duplexes and the second term is the number of
self-complementary (or homo) duplexes. The number N in-
creases very rapidly as the number of intermolecular hydrogen-
bonding sites increases. For example, there should be six
different quadruply hydrogen-bonded duplexes. For oligoamides
that associate with six hydrogen bonds, the number of different
duplexes jumps to 20.
We recently described the design and characterization of
oligoamides 1 and 2, with the self-complementary hydrogen-
bonding sequences DADA and DDAA, which formed stable
duplexes via hydrogen-bonding interactions between the back-
bone amide O and H atoms.15 Incorporating alkoxy groups into
1 and 2 led to the formation of the highly favorable S(6) type16
intramolecular hydrogen-bonded rings that have been observed
in numerous structures.17 The S(6) systems preorganized the
benzamide groups in a way that facilitated the dimerization of
these molecules. The incorporation of intramolecular hydrogen
bonds also helped to block unwanted hydrogen-bonding interac-
(15) Gong, B.; Yan, Y.; Zeng, H.; Skrzypczak-Jankunn, E.; Kim, Y. W.;
Zhu, J.; Ickes, H. J. Am. Chem. Soc. 1999, 121, 5607-5608.
(16) Berstein, J.; Davis, R. E.; Shimoni, L.; Chang, N.-L. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 1555-1573.
hydrogen-bonded duplex should address the following ques-
tions: (1) Instead of self-complementary arrays, can duplexes
(17) Etter, M. C. Acc. Chem. Res. 1990, 23, 120-126.

A Six-Hydrogen-Bonded Molecular Duplex
J. Am. Chem. Soc., Vol. 122, No. 11, 2000 2637
Scheme 1. Synthesis of 3 and 4
containing two different strands of complementary hydrogen-
bonding sequences form? (2) Will the backbones of longer
strands stay in register to allow pairing of the duplexes? (3) In
the formation of a longer duplex, is there any cooperativity
among the building blocks and among the hydrogen-bonding
sites? As illustrated, these two strands may form a molecular
duplex held together by six hydrogen bonds. Given its increased
number of hydrogen-bonding sites and the possibility of
cooperative interactions among the hydrogen bonds, such a
duplex is expected to be more stable than the homologous
quadruply hydrogen-bonded duplexes. The long-term objective
of this work is to develop hydrogen-bonded molecular duplexes
whose formation is characterized by programmable sequence
specificity and adjustable stability and which may be used as
(18) (a) Pranata, J.; Wierschke, S. G.; Jorgensen, W. L. J. Am. Chem.
Soc. 1991, 113, 2810-2819. (b) Jorgensen, W. L.; Pranata, J. J. Am. Chem.
Soc. 1990, 112, 2008-2010. (c) Murray, T. J.; Zimmermann, S. C. J. Am.
Chem. Soc. 1992, 114, 4010-4011.

2638 J. Am. Chem. Soc., Vol. 122, No. 11, 2000
Zeng et al.
Figure 2. 1H NMR (400 MHz, CDCl3) spectrum of duplex 3‚4.
specific molecular recognition units (or “molecular glues”) for
constructing supramolecular structures.
Preparation and Characterization of the Hydrogen-
Bonded Duplex. Duplex 3‚4 was prepared by mixing 1 equiv
of 3 with 1 equiv of 4 in chloroform (1 mL). Although the
solubility of 3 (<1 mM) or 4 (e10 mM) in the same solvent
was relatively low, the 1:1 mixture of 3 and 4 was highly soluble
(.100 mM) in chloroform. This observation suggested that
duplex 3‚4 indeed formed, which shielded the polar amide
groups from the solvent, thus preventing the formation of
polymeric aggregates.20 Attempts to obtain X-ray-quality crystals
of 3‚4 have not been successful so far. Slow evaporation of
solvents from solutions of complex 3‚4 in chloroform and in
chloroform/ethanol led to amorphous masses.
Results and Discussion
Synthesis. Oligoamides 3 and 4, which contain six hydrogen-
bonding sites with the sequences DADDAD and ADAADA,
respectively, were synthesized by iterative coupling steps using
either acid chloride or 1-ethyl-3-(3-dimethylaminopropyl)car-
bodiimide (EDC) (Scheme 1). Methyl salicylate was first
alkylated with 1-bromooctane; product I was then nitrated into
ester II, which was converted into 5-nitro-2-(octyloxy)benzoic
acid, III. Acid III was converted into the corresponding acid
chloride, which was then treated with the corresponding amines
to give ester 3a or amide 4a. The nitro compounds 3a and 4a
were then reduced to the corresponding amines 3b and 4b by
catalytic hydrogenation. Compound 3a was acetylated into 3c,
which was hydrolyzed to give acid 3d. Diamine 3f was prepared
from 3,5-dinitrobenzoyl chloride via the octyl benzoate ester
3e. Coupling of 3d and 3f in DMF using EDC gave 3.
Resorcinol was converted into 4,6-dihydroxy-1,3-benzene-
dicarboxylic acid, 4c, by modifying a reported procedure.19 The
acid 4c was esterified and alkylated into dimethyl 4,6-bis-
(octyloxy)-1,3-benzenedicarboxylate, 4e. This was followed by
hydrolysis to give the acid 4f. The coupling between 4f and
glycine ethyl ester using EDC in DMF gave 4g. Diester 4g was
then hydrolyzed into the dicaid 4h. Coupling between 4h and
4b using EDC in DMF led to 4.
(1) 1H NMR Spectroscopy. Figure 2 shows the 1D 1H NMR
spectrum of duplex 3‚4. The assignment of this spectrum to
the structure of 3‚4 was assisted by using COSY and NOESY
spectra.
1H NMR spectroscopy in chloroform-d (CDCl3) revealed
downfield shifts of the aniline NH signals of the 1:1 mixture of
3 and 4 (at 1 mM: Hb 9.93 ppm, Hi 10.09 ppm, and Hr 10.19
ppm) compared to the corresponding signals of either 3 (Hb
<7.5 ppm and Hi <7.5 ppm; the signals overlap with other
peaks) or 4 (Hr 10.04 ppm), suggesting the formation of duplex
3‚4.
Mixing 3 and 4 in stoichiometries other than 1:1 revealed
separate sets of signals corresponding to both duplex 3‚4 and
the uncomplexed strands. Figure 3 shows the NMR spectrum
of a 1:4 mixture of 3 and 4. This spectrum is consistent with
(20) Whitesides et al. observed similar phenomena for the solubilities
of hydrogen-bonded aggregates: Mathias, J. P.; Simanek, E. E.; Whitesides,
G. M. J. Am. Chem. Soc. 1994, 116, 4326-4340.
(19) Gore, T. S.; Inamdar, P. K.; Nagarkar, P. M. Indian J. Chem. 1974,
946-947.

A Six-Hydrogen-Bonded Molecular Duplex
J. Am. Chem. Soc., Vol. 122, No. 11, 2000 2639
Figure 4. Aniline resonances of duplex 3‚4 at (a) 100 µM, (b) 10
µM, and (c) 1 µM.
Figure 3. Aniline NH resonances of (a) a 1:4 mixture of 3 and 4, (b)
amide 4 at 1 mM, and (c) duplex 3‚4 at 1 mM.
In an attempt to determine the association constant, 1H NMR
binding studies were carried out in CDCl3 by diluting a solution
of the 1:1 mixture of 3 and 4. Within the sensitivity of detection
of the NMR spectrometer (400 MHz), no significant upfield
shifts of the aniline NH signals were observed across a broad
concentration range (100 mM to 1 µM). Figure 4 shows the
regions of the three aniline NH signals at 100, 10, and 1 µM.
Detection of the NH signals at 1 µM was made possible by
first determining the spin-lattice relaxation time T1 of duplex
the existence of the three aniline signals corresponding to duplex
3‚4 and one aniline signal corresponding to uncomplexed 4
(Figure 3). The observation of separate sets of resonances for
duplex 3‚4 and the uncomplexed 4 at this stoichiometry for 3:4
indicated that exchange between the assembled and the uncom-
plexed states was very slow on the time scale of NMR
spectroscopy, which provided another piece of evidence for the
very high stability of the duplex. Similar phenomena were
observed by Whitesides et al.20 for highly stable hydrogen-
bonded aggregates with multiple components.
1
3‚4. The H NMR spectrum (400 MHz) of 3‚4 at 1 µM was
recorded for a total of ∼29 h (acquisition time ) 0.251 s; delay
time ) 0.05 s; excitation pulse width ) 25 ms; number

2640 J. Am. Chem. Soc., Vol. 122, No. 11, 2000
Zeng et al.
of acquisitions ) 332 168). The short acquisition time was based
on the spin-lattice relaxation time of 3‚4 that was experimen-
tally determined to be 0.6 s. On the basis of this short relaxation
time, a short repetition time was adopted, which led to a large
acquisition number within a reasonable period of time. Because
the signal-to-noise ratio is proportional to the square root of
the acquisition number, the large acquisition number allowed a
spectrum with much improved sensitivity. The three aniline NH
signals, which were involved in intermolecular hydrogen-
bonding interactions, remained sharp and well above the
background under these NMR conditions. Instead of moving
upfield, one NH signal corresponding to Hr moved slightly
(0.078 ppm) downfield upon dilution from 100 to 10 µM. This
was probably due to the weakening of stacking interactions
among the dimers. Since there was no detectable change in the
chemical shifts and number of the aniline NH signals, assuming
conservatively a 10% dissociation at 1 µM led to a lower limit
of 9 × 107 M-1 for the association constant of duplex 3‚4.3a
Variable-temperature measurements of the amide NH signals
of 4 were carried out to probe the possibility for this molecule
to adopt folded conformations with intramolecular hydrogen
bonds.21 At 1 mM 4 and from 25 to 65 °C in CDCl3, the
temperature-dependent change of the aniline NH signal (Hr) of
4 was 1.4 × 10-2 ppm/K. The two glycine NH signals, which
were involved in forming the S(6) intramolecular hydrogen
bonds, showed much smaller changes in the same temperature
range: 3.9 × 10-3 ppm/K (Hw) and 3.9 × 10-3 ppm/K (Hp).
These results clearly indicated that the aniline NH group in 4
was not involved in an intramolecular hydrogen-bonding
interaction, and 4 was thus very likely to adopt an extended,
rather than folded, conformation in its uncomplexed form.
Similar variable-temperature experiments could not be carried
out on 3 due to overlap of its amide NH signals with other
aromatic resonances in its 1H NMR spectrum. However, given
its structural similarity to 4, particularly in its backbone rigidity,
3 is also very likely to adopt an extended conformation in
solution.
Figure 5. (a) Amide 4 possibly associating into dimer 4‚4 with four
attractive hydrogen bonds (dotted lines) and one repulsive interaction
between two amide O atoms (double-headed arrow). (b) Amide 3
possibly forming dimer 3‚3 with a stability similar to that 4‚4.
is reasonable to imagine that the self-association of 3 may also
involve four intermolecular hydrogen bonds (Figure 5b).
Considering the equilibrium
The self-association of 4 was investigated by diluting its
solution in CDCl3 from 10 to 0.375 mM. Upfield shifts of its
aniline NH signal from 10.11 to 9.72 ppm was observed. A
dimerization constant of (4.4 ( 2.0) × 104 M-1 was obtained
for 4.22 This value is surprisingly similar to those for the self-
complementary quadruply hydrogen-bonded duplexes we re-
ported recently.15 One possibility is that the self-association of
4 involves four intermolecular hydrogen bonds. Examining the
sequence of hydrogen-bond donors and acceptors in 4 leads to
only one possible arrangement of a self-complementary dimer
with four attractive hydrogen-bonding interactions and one
repulsive interaction involving two amide O atoms (Figure 5a).
The only repulsive interaction in this dimer could be alleviated
by slightly twisting the backbone of the two molecules, which
leads to a dimer with a stability similar to those of quadruply
hydrogen-bonded 1 and 2. Given the considerable degree of
rotational freedom in 4, such a twist of the backbone in 3 should
be easily realized. The self-association of 3 could not be
investigated because of the limited solubility of this compound
in CDCl3. However, given its hydrogen-bonding sequences, it
the dimerization of 4 (and very likely that of 3) should not
interfere with the pairing of 3 and 4. Despite its relatively large
value, K4 is still 5 orders of magnitude (see the ITC data below)
smaller than K3‚4. On the basis of the above equilibrium, if
K3 ≈ K4 and the initial (added) concentrations of 3 and 4 are
the same, a simple calculation shows that at equilibrium only
0.001% of 4 (or 3) will be the dimer 3‚3 or 4‚4, while 99.999%
of 4 and 3 will exist as 3‚4. It needs to be pointed out that this
conclusion is consistent with the 1H NMR observations, which
indicated that diluting a solution of 3‚4 to as low as 1 µM did
not result in the appearance of any signals corresponding to
uncomplexed 3 (Figure 4).
Unequivocal evidence for the formation of 3‚4 in solution
was provided by two-dimensional (2D) NMR (NOESY, CDCl3,
400 MHz) studies (Figure 6). Numerous interstrand contacts
were observed. For example, contacts between protons l and
m, i and m, and h and q, as well as a and x, were particularly
diagnostic of duplex formation. In addition, contacts between
protons f ang g, o and p, and w and v were indicative of the
intramolecular hydrogen bonds that act to preorganize both 3
and 4. These 2D NMR data indicated that, in solution, 3 and 4
indeed aligned and associated in accord with the original design.
Once again, the extraordinary stability of 3‚4 is demonstrated
by the fact that NOESY spectra recorded at 1 or 10 mM 3‚4
showed the same intra- and interstrand contacts.
(21) (a) Kessler, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 512-523.
(b) Hamuro, Y.; Geib, J.; Hamilton, A. D. J. Am. Chem. Soc. 1996, 118,
7529-7541.
(22) Wilcox, C. S. In Frontiers in Supramolecular Organic Chemistry
and Photochemistry; Schneider, H.-J., Durr, H., Eds.; VCH: New York,
1991. The dimerization constant was obtained by fitting the NMR data to
a modified dimerization equation with the program Kaleidagraph on a
Macintosh computer.
(2) Isothermal Titration Calorimetry (ITC). Dilution
experiments showed that the stability constant of duplex 3‚4

A Six-Hydrogen-Bonded Molecular Duplex
J. Am. Chem. Soc., Vol. 122, No. 11, 2000 2641
Figure 7. Calorimetry binding isotherm for the titration of 3 with 4
in 5% DMSO/chloroform.
Since the experiment was carried out at micromolar concentra-
tions, the measurements were carried out near the detection limit
of the calorimeter. When even lower concentrations of 3 and 4
were tested, the resulting data led to association constants with
greater (>10%) error. In all cases, analysis of the resulting
binding isotherm in CHCl3 showed that the binding constant
was in the range of 109 M-1
.
More accurate data were obtained by carrying out the ITC
titration in 5% DMSO/CHCl3. Titration of 4 into 3 produced
the thermogram shown in Figure 7. Integration of the heat data
plotted vs the mole ratio of host and guest produced the iso-
therm also shown in Figure 7. An association constant of
(3.5 ( 1.3) × 106 M-1 was obtained, along with ∆G )
-8.9 ( 0.2 kcal/mol and ∆H ) -9.7 ( 0.2 kcal/mol. The
excellent fit obtained for the thermogram, in combination with
the NMR data, supports our belief that the values reported here
accurately describe the thermodynamic equilibrium.
Given the relatively small entropic change (-T∆S ≈ 0.8 kcal/
mol), the ITC data showed that the formation of 3‚4 was driven
almost entirely by enthalpy. This result implies that 3 and 4
adopt preorganized, extended conformations, which may have
provided optimal orientations for pairing the two strands. This
is consistent with the results of the variable-temperature
measurements, which indicated that 4 was more likely to adopt
an extended, rather than folded, conformation in solution. To
account for the thermodynamic data from the ITC experiment,
a very likely, and yet simplified, scenario can be proposed:
dimer 4‚4 (and perhaps 3‚3) dissociates into single-stranded 4,
which then combines with single-stranded 3 to form the highly
stable 3‚4. This process involves mainly enthalpic, rather than
entropic, changes, during which the existence of any uncom-
plexed, folded molecules is very unlikely.
(3) Thin-Layer Chromatography (TLC). The stability of
3‚4 was further confirmed by straight-phase thin-layer chro-
matography (TLC) analysis (silica gel plate, 10% DMF in
chloroform). As shown in Figure 8, the presence of duplex 3‚4
is clearly demonstrated by the different Rf values of 3 (Rf )
0.00), 4 (Rf ) 0.10), and 3‚4 (Rf ) 0.96). The fact that 3‚4
shows a tailing on the TLC plate indicates its partial dissociation
under the rather polar analytical conditions. That 3‚4 could be
detected under the rather harsh straight-phase TLC conditions
Figure 6. NOESY spectra of 3‚4 (8 mM) in CDCl3 (mixing time 0.5
s): (a) cross-strand contacts between protons l and m and protons i and
m; (b) cross-strand contacts between protons a and x and protons h
and q. NOESY spectra of 3‚4 recorded at concentrations of 1 and 100
mM are similar to the ones shown here.
was far above the detection limit of the NMR method. Titrations
using isothermal titration microcalorimetry can directly measure
∆H° and the association constant over a large range of Ka values,
so this method was employed to acquire thermodynamic profiles
for complexation. We first carried out titration of 4 into 3 in
pure CHCl3. By titration of a solution of 0.1 mM 3 with 0.8
mM 4 in chloroform, an association constant of (1.3 ( 0.7) ×
109 M-1 was obtained. Due to the very high stability of 3‚4,
the concentrations of 3 and 4 had to be decreased significantly
to the micromolar range to obtain a reasonable binding isotherm.

2642 J. Am. Chem. Soc., Vol. 122, No. 11, 2000
Zeng et al.
stable duplex.24 These results, along with those for 1 and 2,
have paved the way for the design of longer duplexes. The
development of a wide variety of extremely tight-binding
duplexes with high sequence specificities can be envisioned.
In addition, this class of molecules offers several advantages,
including ease of synthesis based on highly efficient amide
(peptide) chemistry and readily available building blocks.
Structural modification can be conveniently carried out on this
system. As a result, duplexes compatible with a variety of
environmental conditions should be easily generated. Highly
specific, tight-binding duplexes should find numerous applica-
tions in chemistry, biochemistry, and materials science.
Experimental Section
General Methods. All chemicals were purchased from Aldrich,
Fluka, and Sigma and were used as received unless otherwise noted.
The organic phases from all liquid extractions were dried over Na2-
SO4 unless specified otherwise. All products were detected as single
spots by thin-layer chromatography (precoated 0.25 mm silica plates
from Aldrich). All samples were purified either by recrystallization or
by flash column chromatography and dried completely under high
vacuum before being characterized by 1H NMR (400 MHz), 13C NMR
(100 MHz), and elemental analysis. All 1H NMR and 13C NMR spectra
were recorded on a Varian VXR 400 spectrometer (400 MHz). NMR
chemical shifts are reported in ppm relative to TMS. For the 1H NMR
dilution experiments, CDCl3 (99.8% D) and DMSO-d6 (99.8% D) were
purchased from Carmbridge Isotope Laboratory and used without further
purification. T1 measurements were carried out with degassed samples
(freeze-thaw cycles with nitrogen). NOE measurements were per-
formed with the steady-state NOEDIF protocol on degassed samples.
Methyl 2-(Octyloxy)benzoate (I). Sodium metal (2.76 g, 120 mmol;
washed with clean petroleum ether prior to use) was added in small
pieces to anhydrous MeOH (120 mL) in a well-ventilated hood. The
resulting solution was heated under reflux. Methyl salicylate (15.2 g,
100 mmol) was then introduced, followed by 1-bromooctane (23.2 g,
120 mmol). The reaction was allowed to proceed for 2 days under
reflux. The solid was then filtered off, and the solvent was removed
under reduced pressure. Excess 1-bromooctane was removed in vacuo,
and the residue was dissolved in ethyl acetate. The solution was washed
with diluted NaOH solution and was dried over Na2SO4. Evaporation
of the solvent gave pure liquid product I (25.1 g, 79%). 1H NMR
(CDCl3): δ 7.77 (q, J ) 1.2, 7.6, 1H), 7.43 (m, 1H), 6.95 (m, 2H),
4.02 (t, J ) 6.6, 2H), 3.88 (s, 3H), 1.83 (m, 2H), 1.48 (m, 2H), 1.34-
1.28 (m, 8H), 0.88 (t, J ) 6.6, 2H). (All J values are reported in hertz.)
13C NMR (CDCl3): δ 167.01, 158.66, 133.17, 131.51, 120.79, 120.01,
113.40, 69.10, 51.75, 31.80, 29.29, 29.22, 25.95, 22.62, 14.01. Anal.
Calcd for C16H24O3: C, 72.69; H, 9.15. Found: C, 72.63; H, 9.25.
Methyl 2-(Octyloxy)-5-nitrobenzoate (II). Compound I (26.4 g,
100 mmol) was added dropwise to concentrated H2SO4 (60 mL) cooled
in an ice-water bath, to which a mixture of 70% HNO3 (7.7 mL, d
1.41 g/mL, 120 mmol) and concentrated H2SO4 (32.3 mL) was added
dropwise over a period of 20 min at 0 °C. After being stirred for 1.5
h, the reaction mixture was poured into cracked ice (1 kg) and the
mixture was left overnight while the temperature was kept at about 0
°C. The precipitate was dissolved in ether (600 mL), and the solution
was washed with 1 N NaOH (2 × 30 mL). Drying over Na2SO4 and
evaporation of the solvent gave the crude product as a solid, which
was then recrystallized from MeOH, the give the pure product as a
Figure 8. Thin-layer chromatography of 3 (lane A), 3 + 4 (1:1, lane
B), and 4 (lane C). TLC condition: silica gel plate/10% DMF in CHCl3.
The spot corresponding to 3 (lane A) is weak due to the low solubility
of 3. The developed TLC plate was detected under short-wavelength
UV light.
is very surprising and is a direct confirmation of the extremely
high stability shown by this duplex.23
Positive Cooperativity in the Self-Assembling Process.
Comparing the stabilities of doubly (Ka ) 25 M-1, -0.9 kcal/
mol for each H bond) and quadruply (dimers of 1 and 2, 104
M-1 < Ka < 105 M-1, -1.4 to -1.7 kcal/mol-1 for each H
bond) hydrogen-bonded duplexes15 with that of 3‚4 (1 × 109
M-1, 2.0 kcal/mol for each H bond) indicates that the increase
in stabilities is not just due to the additive effect from increased
numbers of hydrogen bonds. Instead, these data clearly dem-
onstrate that the self-assembly of 3‚4 is highly cooperative: after
the initial association of the two strands, which may involve
one or, at most, two hydrogen bonds and is entropically
unfavorable, the subsequent hydrogen-bond formation during
the growth of the duplex is enhanced by multiple, enthalpically
favorable interactions. This type of cooperativity is one of the
most prominent features of the self-assembly of DNA and lies
at the heart of many self-assembling and self-organizing
systems.1a In addition to hydrogen-bonding interactions, van der
Waals contacts between the methylene hydrogens of the glycine
linkers may also contribute to the stability of the duplexes. Such
contacts, as clearly shown by the NOESY spectra of 3‚4, may
play a dominant role in the future design of duplexes that will
be stable in aqueous media.
Conclusions
1
white solid (21.66 g, 70%). H NMR (CDCl3): δ 8.70 (d, J ) 3.2,
1H), 8.33 (q, J ) 3.2, 9.2, 1H), 7.03 (d, J ) 9.2, 1H), 4.14 (t, J ) 6.4,
2H), 3.92 (s, 3H), 1.88 (m, 2H), 1.50 (m, 2H), 1.38-1.29 (m, 8H),
0.88 (t, J ) 6.8, 3H). 13C NMR (CDCl3): δ 164.69, 163.22, 140.54,
128.69, 127.79, 120.84, 112.72, 69.921, 52.35, 31.77, 29.19, 28.88,
25.81, 22.63, 14.03. Anal. Calcd for C16H23NO5: C, 62.12; H, 7.49;
N, 4.53. Found: C, 62.20; H, 7.75; N, 4.78.
The assembly of 3‚4 indicates that (1) in addition to self-
complementary duplexes, highly stable duplexes containing two
different strands can also be designed, (2) the backbones of 3‚
4 do stay in register to allow the formation of all the hydrogen
bonds and thus the pairing of the duplexes, and (3) the self-
assembly of 3‚4 is highly cooperative, which results in a very
(24) There were reports on hydrogen-bonded complexes with very high
stabilities: (a) Bell, T. W.; Hou, Z.; Zimmerman, S. C.; Thiessen, P. A.
Angew. Chem., Int. Ed. Engl. 1995, 34, 2163. (b) Sessler, J. L.; Wang, R.
J. Org. Chem. 1998, 63, 4079.
(23) Previously only reverse-phase TLC had been used for the charac-
terization of hydrogen-bonded complexes: Seto, C. T.; Whitesides, G. M.
J. Am. Chem. Soc. 1990, 112, 6409.

A Six-Hydrogen-Bonded Molecular Duplex
J. Am. Chem. Soc., Vol. 122, No. 11, 2000 2643
2-(Octyloxy)-5-nitrobenzoic Acid (III). The ester II (3.09 g, 10
mmol) was dissolved in hot MeOH (40 mL), to which 1 N NaOH (20
mL, 20 mmol) was added. The mixture was heated under relux for
30-40 min, and more water (100 mL) was added. The aqueous layer
was neutralized by addition of concentrated HCl to pH 3.0. The
precipitated crude product was collected, and recrystallization from
acetyl chloride (0.79 g, 10 mmol) over 5 min at 0 °C, after which the
ice-water bath was removed. After 6 h, the solvent was evaporated,
the residue was dissolved in ethyl acetate, and the solution was washed
with diluted HCl and NaOH solutions alternatively. Drying and
evaporation of the solvent gave the crude product, which was
recrystallized from MeOH, giving the pure product as a white solid
1
1
MeOH gave a white solid (2.66 g, 90%). H NMR (CDCl3): δ 10.62
(3.56 g, 90%). H NMR (CDCl3): δ 8.70 (s, 1H), 8.13 (q, J ) 2.4,
8.8, 1H), 7.83 (d, J ) 2.4, 1H), 7.38 (s, 1H), 6.96 (d, J ) 8.8, 1H),
4.25 (q, J ) 8.8, 14.0, 4H), 4.13 (t, J ) 6.6, 2H), 2.17 (s, 3H), 1.93
(br, 1H), 9.04 (d, J ) 2.8, 1H), 8.43 (q, J ) 2.8, 9.2, 1H), 4.34 (t, J )
6.6, 2H), 1.97 (m, 2H), 1.51 (m, 2H), 1.41-1.29 (m, 8H), 0.89 (t, J )
6.8, 3H). 13C NMR (CDCl3): δ 129.87, 129.63, 129.55, 118.79, 113.09,
71.41, 71.33, 31.67, 29.09, 20.04, 28.75, 25.75, 22.58, 14.01. Anal.
Calcd for C15H21NO5: C, 61.00; H, 7.17; N, 4.74. Found: C, 60.89;
H, 7.21; N, 4.79.
(m, 2H), 1.48 (m, 2H), 1.37-1.29 (m, 12H), 0.88 (t, J ) 6.4, 3H). 13
C
NMR (CDCl3): δ 169.76, 168.49, 164.96, 153.89, 131.80, 125.59,
123.24, 120.64, 113.06, 69.74, 61.41, 42.16, 31.78, 29.27, 29.20, 29.07,
26.14, 24.35, 22.63, 14.19, 14.06. Anal. Calcd for C21H32N2O5: C,
64.26; H, 8.22; N, 7.14. Found: C, 64.51; H, 8.45; N, 7.16.
N-((Ethoxycarbonyl)methyl)-2-(octyloxy)-5-nitrobenzamide (3a).
The acid III (2.95 g, 10 mmol) was dissolved in 10 mL of thionyl
chloride, and the mixture was heated under reflux with a calcium
chloride drying tube for 2 h. The solvent was removed in vacuo, and
anhydrous ether (2 × 30 mL) was added and then removed. The crude
acid chloride was directly used for the next step without further
purification.
N-(Carboxymethyl)-5-acetamido-2-(octyloxy)benzamide (3d). Hy-
drolysis of 3c based on the procedures described for 1c gave the crude
product, which was crystallized from MeOH, to give the pure product
1
as a white solid (97%). H NMR (DMSO): δ 12.80 (br, 1H), 9.95 (s,
1H), 8.52 (t, J ) 5.0, 1H), 8.0 (s, 1H), 7.78 (d, J ) 8.8, 1H), 7.10 (d,
J ) 8.4, 1H), 4.08 (t, J ) 6.2, 2H), 4.01 (d, J ) 3.6, 2H), 2.0 (s, 3),
1.80 (m, 2H), 1.40-1.24 (m, 10H), 0.843 (t, J ) 6.2, 3H). 13C NMR
(DMSO): δ 171.07, 167.99, 164.23, 152.49, 132.64, 123.59, 121.71,
121.19, 113.57, 69.16, 41.55, 31.20, 28.70, 28.62, 28.43, 25.53, 23.78,
22.07, 13.96. Anal. Calcd for C19H28NO5: C, 62.62; H, 7.74; N, 7.69.
Found: C, 62.62; H, 7.97; N, 7.68.
Octyl 3,5-Dinitrobenzoate (3e). To a solution of 3,5-dinitrobenzoyl
chloride (18.5 g, 80 mmol) and triethylamine (8.1 g, 80 mmol) in 250
mL of methylene chloride was added dropwise 1-octanol in 40 mL of
methylene chloride over 5 min at 0 °C. The ice-water bath was
removed. After 6 h, the solvent was evaporated and the residue was
dissolved in ethyl acetate and washed with diluted HCl and NaOH
solutions alternatively. The ethyl acetate solution was dried over Na2-
SO4. Evaporation of the solvent and recrystallization of the crude
product from MeOH gave the product as a white solid (25.9 g, 84%).1H
NMR (CDCl3): δ 9.22 (t, J ) 2.0, 1H), 9.16 (d, J ) 2.0, 2H), 4.48 (t,
J ) 2.0, 2H), 1.83 (m, 2H), 1.47-1.29 (m, 10H), 0.88 (t, J ) 6.6,
3H). 13C NMR (CDCl3): δ 162.23, 148.43, 133.93, 129.04, 121.92,
66.83, 31.43, 28.84, 28.25, 25.58, 22.29, 13.71. Anal. Calcd for
C15H20N2O6: C, 55.55; H, 6.22; N, 8.64. Found: C, 55.45; H, 6.34; N,
8.70.
Octyl 3,5-Diaminobenzoate (3f). Hydrogenation of 3e based on
procedures similar to those described for 1e gave crude 3f (90%), which
was used directly for the next step without further purification. 1H NMR
(CDCl3): δ 6.78 (d, J ) 1.6, 2H), 6.18 (s, 1H), 4.25 (t, J ) 6.6, 2H),
3.67 (s, 4H), 1.73 (m, 2H), 1.42-1.28 (m, 10H), 0.88 (t, J ) 6.0, 3H).
13C NMR (CDCl3): δ 166.91, 147.47, 132.56, 106.94, 105.55, 64.97,
31.80, 29.25, 29.19, 28.76, 26.05, 22.63, 14.05. Anal. Calcd for
C15H24N2O2: C, 68.15; H, 9.15; N, 10.60. Found: C, 68.14; H, 9.14;
N, 10.67.
To a solution of glycine ethyl ester hydrochloride (10 mmol) and
2.02 g (20 mmol) of triethylamine in CH2Cl2 was added a solution of
the above prepared acid chloride in 40 mL of methylene chloride over
a period of 5 min at 0 °C. The ice-water bath was removed, and the
reaction mixture was allowed to warm to room temperature. After 6 h,
the solvent was evaporated, the residue was dissolved in ethyl acetate,
and the solution was washed alternatively with diluted HCl and NaOH
solutions. Drying over Na2SO4 and evaporation of solvent gave the
crude product, which was recrystallized from MeOH, to give the product
as a white solid (3.61 g, 95%). 1H NMR (CDCl3): δ 9.10 (d, J ) 2.8,
1H), 8.44 (s, 1H), 8.32 (q, J ) 3.4, 9.2, 1H), 7.08 (d, J ) 9.2, 1H),
2.26 (m, 6H), 2.01 (m, 2H), 1.508 (m, 2H), 1.41-1.29 (m, 13H), 0.88
(t, J ) 6.8, 3H). 13C NMR (CDCl3): δ 169.70, 162.91, 161.55, 141.72,
128.56, 128.21, 121.70, 112.54, 70.70, 61.57, 42.26, 31.75, 29.19, 29.14,
28.81, 26.07, 22.61, 14.18, 14.05. Anal. Calcd for C19H28N2O6: C,
59.98; H, 7.42; N, 7.37. Found: C, 59.93; H, 7.34; N, 7.47.
N-Hexyl-2-(octyloxy)-5-nitrobenzamide (4a). This compound was
synthesized by the same procedure used for preparing 3a. Yield: 3.10
g, 82%. 1H NMR (CDCl3): δ 9.10 (d, J ) 2.8, 1H), 9.29 (q, J ) 2.8,
8.8, 1H), 7.78 (s, 1H), 7.05 (d, J ) 9.2, 1H), 4.23 (t, J ) 6.6, 2H),
3.47 (m, 2H), 1.93 (m, 2H), 1.61 (m, 2H), 1.51 (m, 2H), 1.41-1.30
(m, 14H), 0.90 (m, 6H). 13C NMR (CDCl3): δ 162.90, 161.13, 141.91,
128.52, 127.70, 122.84, 112.41, 70.32, 40.11, 31.75, 31.54, 29.49, 29.27,
29.19, 29.10, 26.83, 26.17, 22.60, 22.59, 14.01, 13.99. Anal. Calcd for
C21H34N2O4: C, 66.65; H, 9.06; N, 7.40. Found: C, 66.53; H, 9.22; N,
7.48.
N-((Ethoxycarbonyl)methyl)-5-amino-2-(octyloxy)benzamide (3b).
Compound 3a was reduced by catalytic hydrogenation in methanol at
room temperature, using Pd-C (10%) as the catalyst. Removal of
catalyst and solvent gave the crude product, which was used for the
next step without isolation and further purification. Typical yield: 90-
Octyl 3,5-Bis[(((((5-acetamido-2-(octyloxy)phenyl)carbonyl)ami-
no)methyl)carbonyl)amino]benzoate (3). To a solution of acid 3d
(3.64 g, 10 mmol), 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide
hydrochloride (1.92 g, 10 mmol), and 1-hydoxybenzotriazole (1.35 g,
10 mmol) in 30 mL of dimethylformamide (DMF) was added the
diamine 3f (1.32 g, 5 mmol) in 10 mL of DMF. The reaction was
allowed to proceed overnight. Distilled water (50 mL) was then added
to the reaction mixture. The precipitated solid was filtered off and stirred
with anhydrous MeOH under reflux for 2 h. The solution was cooled
to room temperature and filtered, giving the crude product. Recrystal-
lization from DMF/CHCl3 gave the pure product as a white solid (2.6
1
95%. H NMR (CDCl3): δ 8.74 (s, 1H), 7.56 (d, J ) 2.8, 1H), 6.79
(m, 2H), 4.24 (q, J ) 6.6, 13.2, 4H), 4.05 (t, J ) 6.6, 2H), 3.54 (s,
2H), 1.90 (m, 2H), 1.46 (m, 2H), 1.35-1.28 (m, 13H), 0.88 (t, J )
6.8, 3H). 13C NMR (CDCl3): δ 169.97, 165.36, 150.66, 140.04, 121.40,
119.70, 118.48, 114.02, 69.94, 61.26, 42.07, 31.78, 29.29, 29.21, 26.17,
22.63, 14.19, 14.06. Anal. Calcd for C19H30N2O4: C, 65.11; H, 8.63;
N, 8.00. Found: C, 64.82; H, 8.71; N, 7.97.
N-Hexyl-5-amino-2-(octyloxy)benzamide (4b). This compound was
prepared by the same procedure for used preparing 3b. 1H NMR
(CDCl3): δ 8.154 (s, 1H), 7.57 (d, J ) 3.2, 1H), 6.76 (m, 2H), 4.01 (t,
J ) 6.4, 2H), 3.53 (s, 2H), 3.44 (q, J ) 6.8, 12.8, 2H), 1.82 (m, 2H),
1.59 (m, 2H), 1.46 (m, 2H), 1.43-1.29 (m, 14H), 0.89 (t, J ) 6.4,
6H). 13C NMR (CDCl3): δ 165.21, 150.17, 140.34, 122.42, 118.97,
118.51, 113.95, 69.72, 39.77, 31.80, 31.60, 29.56, 29.51, 29.38, 29.26,
26.89, 26.30, 22.63, 22.61, 14.06. Anal. Calcd for C21H36N2O2: C,
72.38; H, 10.41; N, 8.04. Found: C, 72.38; H, 10.40; N, 8.21.
1
g, 56%). H NMR (DMSO): δ 10.36 (s, 2H), 9.92 (s, 2H), 8.66 (t,
J ) 4.8, 2H), 8.10 (s, 1H), 8.05 (d, J ) 2.8, 2H), 8.02 (d, J ) 1.2,
2H), 7.79 (q, J ) 2.8 Hz, 9.2, 2H), 7.12 (d, J ) 9.2, 2H), 4.24 (t, J )
6.6, 2H), 4.19 (d, J ) 2.4, 4H), 4.11 (t, J ) 6.4, 4H), 2.00 (s, 6H),
1.83 (m, 4H), 1.69 (m, 2H), 1.41-1.14 (m, 26H), 0.82 (t, J ) 6.6,
3H), 0.741 (t, J ) 6.6, 6H). 13C NMR (DMSO): δ 167.76, 167.35,
165.29, 164.14, 152.46, 139.25, 132.57, 130.66, 123.50, 121.72, 121.27,
114.75, 113.93, 113.50, 69.24, 64.57, 43.44, 31.01, 28.57, 28.48, 28.44,
28.38, 28.05, 25.58, 25.25, 23.59, 21.84, 21.82, 13.66, 13.62. Anal.
Calcd for C53H76N6O10: C, 66.50; H, 8.00; N, 8.78. Found: C, 66.12;
H, 8.12; N, 8.79.
N-((Ethoxycarbonyl)methyl)-5-acetamido-2-(octyloxy)benzamide
(3c). To a solution of 3b (3.52 g, 10 mmol) and triethylamine (1.01 g,
10 mmol) in 60 mL of methylene chloride was added dropwise

2644 J. Am. Chem. Soc., Vol. 122, No. 11, 2000
Zeng et al.
4,6-Dihydroxy-1,3-benzenedicarboxylic Acid (4c). Resorcinol (2.2
g, 20 mmol) was mixed with anhydrous potassium bicarbonate (4.0 g,
40 mmol) in a Carius tube at 210 °C for 3 h. The mixture was then
treated with water (100 mL), and the aqueous mixture was extracted
with ether. The crude product was obtained upon acidification of the
aqueous layer. Recrystallization from 2000 mL of boiling water gave
described for 1c gave 4h. Recrystallization of the crude product from
MeOH gave pure 4h as a white solid (90%). H NMR (DMSO): δ
1
12.76 (br, 2H), 8.53 (s, 1H), 8.29 (t, J ) 5.2, 2H), 6.79 (s, 1H), 4.25
(t, J ) 6.4, 4H), 4.02 (d, J ) 5.2, 4H), 1.85 (m, 4H), 1.43-1.20 (m,
20H), 0.85 (t, J ) 6.8, 6H). 13C NMR (DMSO): δ 170.88, 163.52,
160.14, 134.87, 113.87, 97.88, 69.44, 41.39, 31.02, 28.49, 28.41, 28.11,
25.33, 21.88, 13.73. Anal. Calcd for C28H44N2O8: C, 62.66; H, 8.26;
N, 5.22. Found: C, 62.79; H, 8.56; N, 5.09.
1
pure 4c (1.78 g, 45%). H NMR (DMSO): δ 11.87 (br, 2H), 8.28 (s,
1H), 6.40 (s, 1H). 13C NMR (DMSO): δ 171.02, 166.46, 134.31,
105.96, 103.16. Anal. Calcd for C8H6O6: C, 48.50; H, 3.05. Found:
C, 48.46; H, 3.09.
N,N′-Bis[(((3′-((hexylamino)carbonyl)-4-(octyloxy)phenyl)amino)-
carbonyl)methyl]-4,6-bis(octyloxy)-1,3-benzenedicarboxamide (4).
To a solution of 4b (2.48 g, 3.56 mmol) in 30 mL of DMF was added
a solution of the diacid 4h (1.91 g, 3.56 mmol), 1-ethyl-3-(3-
(dimethylamino)propyl)carbodiimide hydrochloride (1.36 g, 3.5 mmol),
and 1-hydroxybenzotriazole (0.96 g, 3.56 mmol) in 40 mL of DMF.
The reaction was allowed to proceed for 6 h at room temperature, and
the precipitated product was collected by filtration. Recrystallization
Dimethyl 4,6-Dihydroxy-1,3-benzenedicarboxylate (4d). A solu-
tion of 4c (5.94 g, 30 mmol) and concentrated H2SO4 (∼7.5 mL) in
120 mL of MeOH was heated under reflux for 2 days. The product
was obtained upon cooling and was filtered off and washed with a
small amount of MeOH, to give pure 4d (6.11 g, 90%). 1H NMR
(CDCl3): δ 11.25 (s, 2H), 8.40 (s, 1H), 6.48 (s, 1H), 3.94 (s, 6H). 13
C
1
NMR (CDCl3): δ 169.73, 167.15, 134.16, 105.78, 104.20, 52.24. Anal.
Calcd for C10H10O6: C, 53.10; H, 4.46. Found: C, 52.93; H, 4.64.
Dimethyl 4,6-Bis(octyloxy)-1,3-benzenedicarboxylate (4e). A mix-
ture of 4d (6.79 g, 30 mmol), K2CO3 (24.86 g, 180 mmol), and
1-bromooctane (23.2 g, 120 mmol) in 150 mL of DMF containing 30
mL of methanol was heated at 100 °C for 2 days. The solid was filtered
off, and the solvent was removed in vacuo at 130 °C. The residue was
dissoloved in ethyl acetate, and the solution was washed with diluted
HCl and NaOH solutions alternatively. Evaporation of the solvent gave
from DMF/CHCl3 gave pure 4 (3.37 g, 79%). H NMR (CDCl3): δ
10.12 (s, 2H), 8.96 (s, 1H), 8.57 (s, 2H), 8.29 (d, J ) 8.8, 2H), 8.18-
8.14 (m, 4H), 6.94 (d, J ) 9.2, 2H), 6.47 (s, 1H), 4.62 (s, 4H), 4.159
(t, J ) 6.6, 4H), 4.10 (t, J ) 6.6, 4H), 3.49 (q, J ) 7.2, 12.8, 4H), 2.03
(m, 4H), 1.86 (m, 4H), 1.61-1.25 (m, 56H), 0.91-0.83 (m, 18H). 13
C
NMR (CDCl3): δ 167.83, 165.01, 164.52, 160.54, 153.22, 136.44,
132.89, 124.95, 123,51, 121.85, 115.33, 112.97, 96.51, 69.91, 69.58,
44.69, 40.13, 31.92, 31.81, 31.62, 29.52, 29.47, 29.41, 29.38, 29.26,
29.24, 29.06, 26.86, 26.30, 22.65, 22.61, 22.60, 14.07, 13.99, 13.97.
Anal. Calcd for C70H112N6O10: C, 70.20; H, 9.43; N, 7.02. Found: C,
70.10; H, 9.62; N, 7.24.
Binding Studies. The binding parameters were determined by
titrating a solution of 3 with that of 4 in an Omega isothermal titration
calorimeter (MicroCal, Northampton, MA). In chloroform, 0.1 mM 3
and 0.8 mM 4 were used. Lower concentrations of host and guest were
used in pure chloroform due to the very high binding affinities. In 5%
DMSO/CHCl3, a solution of 0.5 mM 3 was titrated with a 4 mM stock
solution of 4. The cell was thermostated to (0.1 °C using a circulating
bath. All of the experiments were performed at 25 °C. The enthalpy of
binding between 3 and 4 was determined from heats of multiple single
injections. Injection volumes were 5 mL, with 3 min of equilibration
time allowed between injections. The heat of dilution of both host and
guest into solvent was determined, and the host-guest titration heat
was adjusted by this contribution.
1
pure 4e (11.9 g, 88%). H NMR (CDCl3): δ 8.458 (s, 1H), 6.423 (s,
H), 4.057 (t, J ) 6.4, 4H), 3.85 (s, 6H), 1.87 (m, 2H), 1.51 (m, 4H),
1.35-1.28 (m, 16H), 0.88 (t, J ) 6.8, 3H). 13C NMR (CDCl3): δ
165.43, 163.63, 137.01, 112.05, 97.70, 69.28, 51.62, 31.80, 29.28, 29.21,
29.06, 25.53, 22.64, 14.02. Anal. Calcd for C26H42O6: C, 69.30; H,
9.40. Found: C, 69.46; H, 9.68.
4,6-Bis(octyloxy)-1,3-benzenedicarboxylic Acid (4f). Hydrolysis
of 4e based on procedures similar to those described for 1c gave the
product 4f, which was recrystallized from MeOH, to give pure 4f (93%).
1H NMR (DMSO): δ 12.34 (br, 2H), 8.16 (s, 1H), 6.64 (s, 1H), 4.09
(t, J ) 6.4, 4H), 1.70 (m, 4H), 1.42-1.24 (m, 20H), 0.84 (t, J ) 6.4,
4H). 13C NMR (DMSO): δ 165.84, 162.63, 135.92, 111.86, 98.24,
68.51, 31.16, 28.62, 28.57, 29.40, 25.28, 22.04, 13.90. Anal. Calcd for
C24H38O6: C, 68.22; H, 9.06. Found: C, 68.12; H, 9.21.
N,N′-Bis((ethoxycarbonyl)methyl)-4,6-bis(octyloxy)-1,3-benzene-
dicarboxamide (4g). To a solution of 4f (1.27 g, 3 mmol), 1-ethyl-
3-(3-(dimethylamino)propyl)carbodiimide hydrochloride (1.15 g, 6
mmol), and 1-hydroxybenzotriazole (0.81 g, 6 mmol) in 40 mL of DMF
were added glycine ether ester hydrochloride (0.98 g, 7 mmol) and
triethylamine (0.7 g, 7 mmol) in 20 mL of DMF. The reaction mixture
was stirred for 6 h, and the solvent was removed in vacuo. The residue
was dissolved in ethyl acetate, and the solution was washed with diluted
HCl and NaOH solutions alternatively. The ethyl acetate solution was
dried over Na2SO4. Evaporation of the solvent gave pure 4g as a white
The binding constants, K, and the number of binding sites, n, were
extracted from the calorimetric data by employing the Origin data
analysis software supplied with the Omega titration calorimeter. A
complete description of the data analysis has been published by Brandts
and co-workers.25
NMR dilution experiments were performed on a Varian VXR 400
spectrometer at 25 °C in CHCl3. The solution of a sample was diluted,
1
and the chemical shifts of the NH protons were followed in the H
NMR spectra and analyzed with an equation described previously.22
1
solid (1.07 g, 69%). H NMR (CDCl3): δ 9.01 (s, 1H), 8.26 (t, J )
Acknowledgment. This work was supported by the NIH
(Grant CA74402), NASA (Grant NAG5-8785), and the donors
of the Petroleum Research Fund, administered by the American
Chemical Society (Grant 34456-AC4).
4.4, 2H), 6.44 (s, 1H), 4.23 (m, 8H), 4.14 (t, J ) 6.6, 4H), 1.96 (m,
4H), 1.49 (m, 4H), 1.39-1.27 (m, 22H), 0.87 (t, J ) 6.8, 6H). 13C
NMR (CDCl3): δ 170.13, 164.18, 160.68, 137.33, 114.58, 96.44, 69.86,
61.20, 42.07, 31.75, 29.24, 29.14, 28.98, 26.14, 22.58, 14.15, 13.97.
Anal. Calcd for C32H52N2O8: C, 64.84; H, 8.84; N, 4.73. Found: C,
64.85; H, 9.12; N, 4.81.
JA9942742
N,N′-Bis(carboxymethyl)-4,6-bis(octyloxy)-1,3-benzenedicarboxa-
mide (4h). Hydrolysis of 4g based on procedures similar to those
(25) Wiseman, T.; Williston, S.; Brandts, J. F.; Lin, L.-N. Anal. Biochem.
1989, 179, 131-137.