Energetics of a low barrier hydrogen bond in nonpolar solvents

Article abstract of DOI:10.1021/ja960288l

A measure of the strength of a low barrier hydrogen bond (LBHB) in apolar organic media was obtained using synthetic molecules derived from Kemp's triacid. The structures feature unusually rigid conformations that enforce intramolecular hydrogen bonds in

Full text of DOI:10.1021/ja960288l

J. Am. Chem. Soc. 1996, 118, 8575-8579
8575
Energetics of a Low Barrier Hydrogen Bond in Nonpolar
Solvents
Yoko Kato, Leticia M. Toledo, and Julius Rebek, Jr.*
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
ReceiVed January 26, 1996X
Abstract: A measure of the strength of a low barrier hydrogen bond (LBHB) in apolar organic media was obtained
using synthetic molecules derived from Kemp’s triacid. The structures feature unusually rigid conformations that
enforce intramolecular hydrogen bonds in a dicarboxylic acid, its corresponding acid-amide and their respective
conjugate bases. Analysis of proton and deuterium NMR spectra established the formation of a LBHB in the conjugate
base of the diacid and a conventional hydrogen bond in the conjugate base of the acid-amide. Through deprotonation
equilibria with organic bases, it was determined that the conjugate base of the diacid was more stable than the
conjugate base of the acid-amide by 2.4 kcal/mol in benzene and 1.4 kcal/mol in dichloromethane. These figures
set the upper limits for the free energy of the additional stabilization arising from the LBHB at 25 °C. This value
is far lower than many estimates but is closer to the recent determinations of Schwartz and Drueckhammer [J. Am.
Chem. Soc. 1995, 117, 11902-11905].
Introduction
or in unnatural amino acid replacement experiments.16 To
complement these recent findings and estimates, we examined
LBHB’s in low dielectric media using synthetic structures.
During the review of this manuscript, a study of LBHB’s in
organic solvents was published.17 This gave a value of 4-5
kcal/mol for phthalate monoanion derivatives in DMSO and
2.7-6 kcal/mol for range of phenol-phenolate derivatives in
THF; no evidence for special stabilization due to pKa-matched
LBHB’s was found. During the revision of this manuscript,
an analysis of hydrogen bonding in solution appeared in which
the relevance of spectroscopic features of LBHB’s to their
strength was seriously questioned.18 The studies described
below are consistent with these recent developments.
Unusually strong hydrogen bonds, known as low barrier
hydrogen bonds (LBHB), have emerged as a new rationale for
the exceptional catalytic abilities of some enzymes.1-5 The
existence of LBHB’s in solution and especially in the solid state
and gas phase has been established for some time6,7 and the
criteria for their detection are well established, but the energetics
involved are harder to assess. Recently, there has been some
debate on the magnitude of the energetics of the LBHB.8-12
For condensed phases, some estimates for the additional
stabilization from a LBHB (vs a conventional hydrogen bond)
exceed 10 kcal/mol. However, a recent study of mesaconic and
citraconic acid derivatives in DMSO solution indicated a
considerably lower value of 4.4 kcal/mol for the extra stabiliza-
tion.13 Another study that compares enzyme-inhibitor com-
plexes comes to an even smaller figure of ∼2 kcal/mol.14 This
value is more in line with those determined for conventional
hydrogen bonds as obtained through site directed mutagenesis15
The system involves m-xylidenediamine-bis(Kemp’s triacid)-
imide (XDK) 1a (Figure 1a), a structure of some utility as a
model in bioorganic and bioinorganic chemistry.19 XDK
features two unconjugated acids, rigidly locked into a conforma-
tion that enforces the formation of two intramolecular hydrogen
bonds.20 Of the four types of hydrogen bonds that could be
formed between two carboxylic acids (Figure 1b), the arrange-
ment in XDK allows only the syn-syn pattern. Upon deproto-
nation of XDK, a monoanion with an intramolecular hydrogen
bond is obtained; the symmetry of this species requires that the
acid and base components have precisely the same strength.
This balance of pKa’s is believed to be an essential feature of
the LBHB.4,5
X Abstract published in AdVance ACS Abstracts, August 15, 1996.
(1) Gerlt, J. A.; Gassman, P. G. Biochemistry 1993, 32, 11943-11952.
(2) Gerlt, J. A.; Gassman, P. G. J. Am. Chem. Soc. 1993, 115, 11552-
11568.
(3) Cleland, W. W.; Kreevoy, M. M. Science 1994, 264, 1887-1890.
(4) Frey, P. A.; Whitt, S. A.; Tobin, J. B. Science 1994, 264, 1927-
1930.
(5) Tobin, J. B.; Whitt, S. A.; Cassidy, C.; Frey, P. A. Biochemistry 1995,
34, 6919-6924.
Diacid 1a shows peculiar pKa’s in aqueous ethanol.21,22 The
first pKa (4.8) is only slightly lower than that of other analogous
(6) Hibbert, F.; Emsley, J. AdV. Phys. Org. Chem. 1990, 26, 255-379.
(7) Gilli, P.; Bertolasi, V.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 1994,
116, 909-915.
(8) Guthrie, J. P.; Kluger, R. J. Am. Chem. Soc. 1993, 115, 11569-
11572.
(16) Thorson, J. S.; Chapman, E.; Schultz, P. G. J. Am. Chem. Soc. 1995,
117, 9361-9362.
(17) Shan, S.; Loh, S.; Herschlag, D. Science 1996, 272, 97-101.
(18) Guthrie, J. P. Chem. Biol. 1996, 3, 163-170.
(19) Rebek, J.; Ballester, A. P.; Buhr, C.; Jones, S.; Nemeth, D.; Williams,
K. J. Am. Chem. Soc. 1987, 109, 5033-5035. Rebek, J., Jr.; Marshall, L.;
Wolak, R.; J. McManis, J. J. Am. Chem. Soc. 1984, 106, 1170-1171.
Watton, S. P.; Masschelein, A.; Rebek, J., Jr.; Lippard, S. J. J. Am. Chem.
Soc. 1994, 116, 5196-5205.
(20) Marshall, L.; Parris, K.; Rebek, J., Jr.; Luis, S. V. Burguete, M. I.
J. Am. Chem. Soc. 1988, 110, 5192-5193.
(21) Parris, K. Ph. D Thesis, University of Pittsburgh, 1988.
(22) Rebek, J., Jr.; Duff, R. J.; Gordon, W. E.; Parris, K. J. Am. Chem.
Soc. 1986, 108, 6068-6069.
(9) Warshel, A.; Papazyan, A.; Kollman, P. A.; Cleland, W. W.; Kreevoy,
M. M.; Frey, P. A. Science 1995, 269, 102-106.
(10) Scheiner, S.; Kar, T. J. Am. Chem. Soc. 1995, 117, 6970-6975.
(11) Alagona, G.; Ghio, C.; Kollman, P. A. J. Am. Chem. Soc. 1995,
117, 9855-9862.
(12) Mock, W. L.; Chua, D. C. Y. J. Chem. Soc., Perkin Trans. 2 1995,
2069-2074.
(13) Schwartz, B.; Drueckhammer, D. G. J. Am. Chem. Soc. 1995, 117,
11902-11905.
(14) Usher, K. C.; Remington, S. J.; Martin, D. P.; Drueckhammer, D.
G. Biochemistry 1994, 33, 7753-7759.
(15) Fersht, A. R. Trends Biochem. Sci. 1987, 12, 301-304.
S0002-7863(96)00288-0 CCC: $12.00 © 1996 American Chemical Society

8576 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Kato et al.
Figure 1. (a) Previously studied XDK derivatives. (b) Four types of
hydrogen bonds which can be formed between two carboxylic acids.
diacids under comparable conditions, but the second pKa of 11.1
is remarkably high. The large second pKa has been attributed
to the repulsion of two negative charges at a small distance,
especially since the more basic syn lone pairs of the carboxylates
are directed at one another in the dianion. Large ∆pKa’s in
aqueous solution are also characteristic properties of hydrogen
maleate and hydrogen phthalate, which form short and (presum-
ably strong) hydrogen bonds. A strong LBHB in the XDK
monoanion might also explain the large ∆pKa’s in 1a, although
it is generally believed that such hydrogen bonds do not persist
in polar protic media.23
The magnesium complex 1b in DMSO does indeed show a
downfield NMR signal at 19 ppm, a clear indication of a
strongly hydrogen bonded proton.24 We had also observed a
short hydrogen bond in the crystal structure of calcium complex
1c. The relevant O-O distance was 2.43 Å in 1c, while in 1a
the O-O distance of the two conventional hydrogen bonds were
2.68 and 2.69 Å.21
A highly lipophilic hexabenzyloxymethyl-XDK diacid 2a was
synthesized for study in apolar media. The corresponding
monoamide-monoacid 3, which offers a control was also
prepared, as was the diamide 4. The crystalline-calcium
complex 2b enabled analysis of the geometry of the LBHB-ed
complex. The syntheses were carried out as outlined in Figure
2. Conversion of benzyloxymethyl-Kemp’s anhydride-acid25
to acid chloride 5 proceeded smoothly with oxalyl chloride. This
material was used to acylate diaminoxylene in refluxing
pyridine to yield XDK diacid 2a in 70% yield. The corre-
sponding diacid chloride 6 was obtained by treatment of 2a with
oxalyl chloride. When 3 equiv of ammonia was placed with
diacid chloride 6 in the presence of triethylamine (TEA) in a
warm sealed tube, diamide 4 was obtained in 78% yield. When
compound 6 was reacted with only 1 equiv of ammonia under
the same conditions, both XDK acid-amide 3 (6% yield) and
diamide 4 (16% yield) could be isolated. The 2:1 diacid/calcium
complex (2b) was prepared by treating 2a with 0.5 equiv of
CaH2.
Figure 2. Synthesis of XDK derivatives.
at 12.60 ppm, characteristic of a neutral, normal hydrogen bond,
was replaced by a signal at 18.0 ppm when 1 equiv of
triethylamine was added. This downfield signal is consistent
with that expected for a LBHB. When the system was cooled
to -70 °C, this resonance moved even further downfield to 19.2
ppm. Excess triethylamine (4 equiv) was added, but the peak
positions did not change beyond those observed at 1 equiv.
Accordingly, the second deprotonation of the diacid does not
occur under these conditions.
For acid-amide 3 in CDCl3 the chemical shift was at 13.80
ppm, and amide protons appeared at 8.50 and 6.03 ppm. When
approximately 1 equiv of triethylamine was added, the acid
signal disappeared, and the downfield (HZ) amide signal moved
to 11.6 ppm. This signal, while unusually downfield for an
amide, is far upfield from the region (16-20 ppm) associated
with strong hydrogen bonds. The amide HE signal at 6.0 ppm
moved only very slightly. Accordingly, the hydrogen bonds
formed appear to be of a conventional type in both the neutral
and monodeprotonated forms of 3. A recent study of an amide-
acid pair in peptide helices reports a much more modest shift
(from 7.6 to 7.8 ppm) when a neutral glutamine-aspartate pair
was deprotonated. The hydrogen bond in that case was with
the HE proton of the amide; the HZ proton shift (∼7.0 ppm)
remained unchanged.26 In diamide 4, the chemical shifts of the
amide protons were at 8.90 and 5.86 ppm; these signals also
remained unchanged in the presence of added organic bases.
A deuterium isotope shift experiment27 gave evidence that
the hydrogen bond in the monoanion of 2a could be further
characterized as a double-well LBHB. The acid protons in 2a
were exchanged with deuterium by treatment with DCl in D2O.
Characterization of the Hydrogen Bonding in Solution
The titration of diacid 2a with triethylamine in anhydrous
CD2Cl2 was monitored by 1H NMR at 25 °C. The diacid signal
2
The signal for the diacid in the H NMR of 2c (Figure 2) at
(23) Perrin, C. L. Science 1994, 266, 1665-1668.
(24) Yun, J.; Lippard, S. J. Unpublished results.
(25) Kato, Y.; Conn, M. M.; Rebek, J., Jr. J. Am. Chem. Soc. 1994, 114,
3279-3284.
(26) Huyghues-Despointes, B. M. P.; Klingler, T. M.; Baldwin, R. L.
Biochemistry 1995, 34, 13267-13271.
(27) Altman, L. J.; Laungani, D.; Gunnarsson, G.; Wennerstro¨m, H.;
Forse´n, S. J. Am. Chem. Soc. 1978, 100, 8264-8266.

Energetics of a Low Barrier Bond in Apolar SolVents
J. Am. Chem. Soc., Vol. 118, No. 36, 1996 8577
Figure 3. Front view of the 2b structure showing the water molecule
coordinated to the calcium and hydrogen-bonded to the carbonyl
oxygens of the imides.
Figure 4. Side view of the 2b complex showing the geometry of the
calcium ion (benzyloxy groups and CH2Cl2 molecules were omitted
for clarity).
room temperature in dry CH2Cl2 moved to 11.86 ppm, a ∆δ-
(H, D) of 0.74 ppm. When 1 equiv of triethylamine was added,
and the monoanion form was cooled to -70 °C, a broad peak
at 17.8 ppm for the acid deuteron was observed. An energy
barrier for the deuterium transfer between the two oxygens was
confirmed from the sizable positive ∆δ(H, D) value of 1.4 ppm.
From the chemical shifts and the isotope shifts, the diacid
appears to exhibit conventional hydrogen bonds, while the
monoanion forms a LBHB rather than a single-well hydrogen
bond.
Table 1. Summary of Deprotonation Studies at 25 °C
solvent
base
substrate
Keq (sd)
∆G (kcal/mol)
benzene DPG
2a
3
2a
3
2.2 (( 0.2) × 106
2.1 (( 0.7) × 104
6.6 (( 3.1) × 104
3.5 (( 1.5) × 103
-8.3 ( 0.1
-5.9 ( 0.2
-6.2 ( 0.4
-4.8 ( 0.3
CH2Cl2
TEA
molecule filling the void cavity on the face of the xylene
opposite the calcium had the two chlorine atoms in close
proximity (∼3.5 Å) to the acid and imide oxygens.
There exists, nonetheless, some uncertainty about the solution
structure of the hydrogen bonds featured in the monoanions.
There appears to be enough flexibility in the skeleton of XDK
to permit independent motions of the two ends of the system.28
The syn-syn arrangement is forced on the molecules and
adjustment of the structure to give the best hydrogen bonds with
guest species has been observed.29 It is likely that the monoanion
of 2a can also attain the best geometry for intramolecular
hydrogen bonds. In the absence of a metal ion, the arrangement
with the spectator oxygens well-separated (as shown in Figure
1b) should be the best arrangement. However, a bifurcated
hydrogen bond with the acid placed between the oxygens cannot
be excluded by the evidence outlined above. Similar consid-
erations apply to the structure of the monoanion of 3. As we
shall see, the presence of a metal ion stabilizes the spectator
oxygens to give rise to another array, that observed in the solid
state.
Equilibria and Energetics
Deprotonations of 2a and 3 with diphenyl guanidine (DPG)
were examined in benzene, using procedures developed to
measure acid-base equilibria and pKa’s in nonpolar organic
solvents.30 The indicator dye bromophthalein magenta E was
applied with triethylamine as the base in CH2Cl2 at 25 °C. The
equilibrium constant (Keq) determined for the acid-base reac-
tions 1 and 2 are summarized in Table 1.
Crystallography
Recrystallization of 2b from CH2Cl2 and isooctane yielded
single crystals suitable for X-ray structure determination. A
single crystal X-ray diffraction analysis unambiguously defined
the structure of 2b (Figures 3 and 4). The overall geometry of
2b was analogous to that obtained earlier for 1c.21 The O-O
distance of the short hydrogen bond was 2.42 Å, within the
range for a LBHB. This short bond was in the same plane as
the xylene atoms of 2b, causing the cyclohexyl rings of the
Kemp’s imide-acid to be tilted out of the plane of xylene . The
calcium had octahedral coordination with two molecules of 2b
positioned in the equatorial plane and two water molecules
coordinated in the axial positions. The oxygen of the water
molecule was coplanar with the carbonyl oxygens of the two
imides, and each hydrogen of the water formed a hydrogen bond
to the carbonyl oxygens of the two imides. The CH2Cl2
The deprotonation of the diacid was easier than the depro-
tonation of acid-amide 3, a result that anticipates the generation
of a stronger hydrogen bond in reaction 1 than in reaction 2.
Control experiments using various amounts of diamide 4 and
the indicator-base solution, showed that the nonacidic portions
of the molecule neither affect the equilibrium nor alter the
absorption of the indicator-base salt.
From the Keq for eqs 1 and 2, the free energy of a single
deprotonation to form a charged hydrogen bond was calculated,
using a statistical correction for the two acid sites for the
(28) Rebek, J., Jr.; Marshall, L.;.McManis, J.; Wolak, R. J. Org. Chem.
1986, 51, 1649-1653.
(29) Pascal, R. A.; Ho, D. M. J. Am. Chem. Soc. 1993 115, 8507-8508.
Duffy, E. M.; Jorgensen, W. L. J. Am. Chem. Soc. 1994, 116, 6337-6343.
(30) Davis, M. M. Acid-Base BehaVior in Aprotic Organic SolVents; U.S.
National Bureau of Standards: Washington, D.C., 1968; Vol. Monograph
105. Davis, M. M.; Hetzer, H. B. J. Res. Natl. Bureau Standards 1958, 60,
569-592.

8578 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Kato et al.
The reaction mixture was filtered through Celite and then concentrated
in vacuo to give a faintly purple crystalline solid. This was immediately
treated with acid chloride 5 (1.00 g, 1.74 mmol) and DMAP (2 mg,
0.016 mmol) dissolved in pyridine (30 mL). The mixture was heated
at reflux for 18 h under Ar. The reaction mixture was concentrated,
and the brownish yellow residue was taken up in EtOAc. The organic
phase was washed with 1 N HCl(aq) (2 × 50 mL) and brine (50 mL),
then dried over Na2SO4, filtered, and concentrated in Vacuo. The
resulting residue was purified using flash chromatography through a
silica gel column (20, 30, 40% EtOAc/CH2Cl2) to give crude diacid
2a (778 mg, 70%). This was recrystallized from CH2Cl2/Et2O to give
608 mg (56%) of pure 2a as white solid: mp 187.6-188.4 °C; IR
(KBr) 3488, 2862, 1735, 1686, 1648, 1453, 1364, 1192, 1097, 737,
697 cm-1; 1H NMR (500 MHz, DMSO-d6) δ 18.70 (br s, 1 H), 7.32-
7.23 (m, 31 H), 6.982 (s, 1 H), 4.486 (s, 8 H), 4.437 (s, 4 H), 3.802 (d,
4 H, J ) 8.8 Hz), 3.348 (d, 4 H, J ) 8.8 Hz), 3.197 (s, 4 H), 2.588 (d,
2 H, J ) 12.7 Hz), 2.343 (d, 4 H, J ) 13.2 Hz), 1.792 (s, 6 H), 1.591
(d, 4 H, J ) 12.2 Hz), 1.536 (d, 2 H, J ) 12.7 Hz); HRMS (FAB in
3-nitrobenzyl alcohol) calcd for C74H77N2O14 (M + H), 1217.5375;
found, 1217.5382.
deprotonation of 2a (Table 1). The difference in the free
energies for the diacid and acid-amide can be partially attributed
to the extra stabilization energy arising from a charged LBHB.
The values in benzene and CH2Cl2 were determined to be -2.4
and -1.4 kcal/mol, respectively.
Discussion
This values obtained above are in general agreement with
the results of Schwartz and Drueckhammer, who recently
showed that the anti-anti LBHB of a carboxylate-acid pair was
4.4 kcal/mol more stable in DMSO than the hydrogen bond of
the corresponding carboxylate-amide pair.14 Our results are
somewhat lower, and even then are likely to be an overestimate
of the stabilization afforded by a LBHB vs its conventional
counterpart. The reason is that, as in the Schwartz and
Drueckhammer study, the hydrogen bonds being compared are
to two acids of different aciditysthe carboxylic acid and the
primary amide. To be sure, the acid is unusually weak (pKa )
11.1 in aqueous ethanol) but still several orders of magnitude
stronger than the value estimated for a primary amide (pKa ∼
15). In addition, the primary amide features different electro-
statics from the carboxylic acid as presented to the carboxylate.
According to Gilli et al.,7 this situation leads to inherently
weaker strength of a heteronuclear (N-H‚‚‚O) hydrogen bond
vs a homonuclear (N-H‚‚‚N or O-H‚‚‚O) one. A primary
alcohol opposite the carboxylate would be a better reference
compound from these perspectives and we are hopeful that it
can be prepared for parallel studies.
Deuterated Diacid 2b. Diacid 2a was dissolved in CH2Cl2 and
washed with 1 N DCl in D2O (3 × 8 mL). The organic phase was
separated, concentrated, and dried in a vacuum desiccator.
Diacid Calcium Complex 2c. Diacid 2a (30 mg, 0.025 mmol) and
CaH2 (0.5 mg, 0.5 equiv) were placed in dry benzene (1 mL) and stirred
at room temperature for 1 h. The solution was filtered through glass
wool and concentrated in vacuo: mp 200.9-203.2 °C; IR (KBr) 3488,
1
2862, 1735, 1686, 1648, 1453, 1364, 1192, 1097, 737, 697 cm-1; H
NMR (300 MHz, CD2Cl2) δ 19.4 (br s, 2 H), 7.36-7.12 (m, 62 H),
7.028 (s, 2 H), 4.58-4.47 (m, 16 H), 4.394 (d, 4 H, J ) 12.2 Hz),
4.281 (d, 4 H, J ) 11.7 Hz), 3.964 (d, 4 H, J ) 8.8 Hz), 3.912 (d, 4
H, J ) 9.3 Hz), 3.51-3.44 (m, 12 H), 3.349 (d, 4 H, J ) 8.3 Hz),
2.769 (d, 4 H, J ) 13.2 Hz), 2.676 (d, 4 H, J ) 13.2 Hz), 2.590 (d, 4
H, J ) 14.2 Hz), 1.91-1.84 (m, 16 H), 1.702 (d, 4 H, J ) 12.7 Hz),
1.629 (d, 4 H, J ) 13.2 Hz).
The recent correlation study5 of ∆pKa to hydrogen bond
strength also showed no special energetics for the hydrogen bond
forming at ∆pKa ) 0. Even though these studies involved
different geometries, solvents, and methods, it appears that the
contributions of pKa matched LBHB’s vs pKa mismatched
hydrogen bonds are quite modest in solution. The implications
of these findings for biological catalysis will be left for others
to define.17,18
Single Crystal X-ray Diffraction Analysis of Diacid Calcium
Complex 2b. Compound 2b crystallized as clear rods from slow
diffusion of isooctane into CH2Cl2 containing 2b. A specimen with
approximate dimensions 0.4 × 0.3 × 0.3 × mm was selected for
analysis and was mounted on a fiber embedded in a matrix of Paraton
N. Data were collected at -66 °C on a Siemens CCD diffractometer
(equipped with an automated three circle goniometer and a solid state
generator) using graphite-monochromatized Mo, KR (0.710690 Å) by
the ω scan method operating under the program SMART.32 A total of
15 frames at 20 s measured at 0.3° increments of ω at three different
values of 2θ and φ were collected, and, after least squares, a preliminary
unit cell was obtained. For data collection three sets of frames of 20
s exposure were collected. Data were collected in three distinct shells.
For the first shell, 606 frames were collected with values of φ ) 0°
and ω ) -26°. For the second shell, 435 frames were collected with
φ ) 88° and ω ) -21°, and, for the third shell, values of φ ) 180°
and ω ) -23° were used to collect 230 frames. At the end of the
data collection, the first 50 frames of the first shell were recollected to
correct for any crystal decay, but no anomalies were observed. The
data were integrated using the program SAINT.33 The integrated
intensities of the three shells were merged into one reflection file. The
data were filtered to reject outliers based on the agreement of the
intensity of the reflection and the average of the symmetry equivalents
to which the reflection belongs. A total of 26 381 reflections were
measured, (2θmax ) 46.6°) of which 9955 (Rint for averaged reflections
) 0.06).
Experimental Section
General Methods. Air-/water-sensitive reactions were performed
in flame-dried glassware under argon. Tetrahydrofuran and dioxane
were distilled from Na/benzophenone ketyl. Methylene chloride was
distilled from P2O5. Unless otherwise stated, all other commercially
available reagents were used without further purification.
NMR spectra were obtained on Bruker AC-250, Varian XL-300,
Varian UN-300, and Varian VXR-500 spectrometers. All chemical
shift values are reported in parts per million. Spectra taken in CDCl3
were referenced to residual CHCl3 (7.26). Spectra taken in DMSO-d6
were referenced to residual solvent (2.49). Fourier transform infrared
spectra were taken on a Perkin-Elmer infrared spectrometer. UV
measurements were taken on a Perkin-Elmer Lambda2 spectropho-
tometer. High-resolution mass spectra were obtained with a Finnegan
Mat 8200 instrument. Melting points were taken on a Elctrothermal
9100 melting-point apparatus. Flash chromatography31 was performed
using Silica Gel 60 (ICN, 230-400 mesh).
Synthesis. Anhydride Acid Chloride 5. Oxalyl chloride (120 µL,
8 equiv) was added to a cooled (ice bath) solution of Kemp’s anhydride
acid (100 mg, 0.174 mmol) in CH2Cl2 (5 mL) and DMF (5 µL) and
was stirred for 5 min. The solution was warmed to room temperature
and stirred overnight. The solvent was then removed in Vacuo, and
the resulting white solid was used immediately without further
purification.
2,4-Bis{((cis,cis-2,4-dioxo-1,5,7-tris((benzyloxy)methyl)-3-azabicyclo-
[3.3.1]non-7-carboxylic Acid m-Xylene (2a). 2,4-Dinitro-m-xylene
(176 mg, 0.897 mmol) and 10% Pd-C (18 mg) in tetrahydrofuran (4
mL) was stirred under a H2 atmosphere at room temperature for 24 h.
The systematic extinctions and crystal density were consistent with
the monoclinic space group P21/n with half the molecule, one CH2Cl2
and one water of composition C75H80N2O15Cl2Ca0.5 forming the
asymmetric unit. The unit cell dimensions were a ) 16.631(1), b )
23.990(2), and c ) 17.267(1) Å, â ) 95.9645(1)°, V ) 6852.0(6) Å3
with Z ) 4. The structure was solved using the direct methods program
(32) SMART, V. 4.0, 4.0th ed.; Siemens Industrial Automation, Inc.:
Madison, WI, 1994.
(31) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923-
2925.
(33) SAINT, V. 4.0, 4.0 ed.; Siemens Industrial Automation, Inc.:
Madison, WI, 1995.

Energetics of a Low Barrier Bond in Apolar SolVents
J. Am. Chem. Soc., Vol. 118, No. 36, 1996 8579
Sir92 of the TeXsan34 crystallographic package of the Molecular
Structure Corporation. The final full-matrix least-squares refinements
with 4477 reflections (2.7° < 2θ < 40.9°, I > 3σ) with anisotropic
thermal parameters for all non-hydrogen atoms and located and riding
697 cm-1; 1H NMR (300 MHz, CDCl3) δ 8.903 (br s, 2 H), 7.35-7.24
(m, 30 H), 7.079 (s, 1 H), 7.015 (s, 1 H), 5.856 (br s, 2 H), 4.572 (d,
4 H, J ) 12.32 Hz), 4.490 (d, 4 H, J ) 11.2 Hz), 4.471 (s, 4 H), 3.889
(d, 4 H, J ) 9.4 Hz), 3.475 (d, 4H, J ) 8.5 Hz), 3.323 (s, 4 H), 2.767
(d, 2 H, J ) 13.0 Hz), 2.404 (d, 4 H, J ) 14.3 Hz), 1.917 (s, 6 H),
1.689 (d, 2 H, J ) 13.0 Hz), 1.618 (d, 4 H, J ) 14.9 Hz); HRMS
(FAB in 3-nitrobenzyl alcohol) calcd for C74H79N4O12 (M + H),
1215.5694; found, 1215.5704.
Determination of Equilibrium Constants by UV. Reagent grade
benzene, dried over 4Å molecular sieves and subsequently stored over
fresh 4Å molecular sieves, and anhydrous methylene chloride stored
over 4A molecular sieves, were used as solvent. Teflon stoppered
quartz cuvettes (10 mm cell length) were dried in a vacuum desiccator
prior to use. The 3′,3′′,5′,5′′-tetrabromophenolphthalein ethyl ester
(Bromophthalein magenta E), diphenyl guanidine, and substrates 2a,
3, and 4 were stored in a desiccator prior to use. TEA was dried over
4Å molecular sieves.
Stock solutions of bromophthalein magenta E (BME) indicator (2.5
× 10-4 M, 5 × 10-4 M) and substrates 2a (2.5 × 10-4 M, 5 × 10-4
M) and 3 (2.5 × 10-4 M, 1.25 × 10-3 M, 1.0 × 10-3 M) were prepared
in both benzene and CH2Cl2. Solutions of diphenyl guanidine in
benzene (5 × 10-4 M), triethylamine in CH2Cl2 (5 × 10-4 M), and
diamide 4 (5 × 10-4 M) in benzene were prepared. Aliquots of
indicator, base, and substrate (2a, 3, or 4) were added to the cuvette
and diluted with either benzene or CH2Cl2 to 1.00 mL such that the
indicator was 2.5 × 10-5 M, the base concentrations were 0.14-4.0
equiv, and the substrate concentration was in the range of 1-15 times
that of the indicator. The solutions were capped immediately and were
mixed well using a vortex mixer. The cuvettes were placed in a
temperature-controlled cell holder at 25 °C and were equilibrated for
15-20 min before making absorbance measurements. A cuvette
containing 1 mL of solvent was used as reference. The changes in
absorbance that accompanied the conversion of BME to its diphe-
nylguanidinium salt (540 nm in benzene) or triethylammonium salt (577
nm in CH2Cl2) were recorded.
A literature value for the equilibrium constant for the reaction of
BME and diphenyl guanine forming the salt complex in benzene at 25
°C was used.30 The equilibrium constant of BME and TEA in CH2Cl2
at 25 °C was determined from a constant indicator UV titration: aliquots
of a solution of 2.5 × 10-4, 1.0 × 10-3, 2.5 × 10-3 M TEA in 2.5 ×
10-5 M BME were added to a 2.5 × 10-5 M BME solution. The
resulting curve (18 point) obtained by following the absorbance of the
indicator-TEA salt was fitted to the 1:1 binding isotherm.35 Nonlinear
least-squares regression was used to curve-fit the experimental data
with the Simplex algorithm as implemented in Systat 5.2.36 Equilibrium
constants for the deprotonation of the substrate were calculated as
described by Davis et al.30 At least 10 reliable measurements were
taken and averaged to obtain the value for each equilibrium constant.
isotropic hydrogens converged smoothly to a final R ) 0.072, Rw
)
0.085. A computer-generated perspective model of the final model is
given in Figure 1.
Diacid Chloride 6. Oxalyl chloride (100 µL, 14 equiv) was added
to diacid 2a (100 mg, 0.082 mmol) dissolved in CH2Cl2 (15 mL) and
DMF (5 µL) and cooled in an ice bath. After stirring for 10 min, the
bath was removed and stirred for 3 h at room temperature. The clear
solution gradually became cloudy. Concentration in Vacuo yielded the
diacid chloride as a white solid, and this material was used without
further purification: 1H NMR (250 MHz, CDCl3) δ 7.4-7.2 (m, 31
H), 7.047 (d, 1 H, J ) 6.1 Hz), 4.572 (d, 4 H, J ) 12.0 Hz), 4.494 (s,
4 H), 4.483 (d, 4 H, J ) 12.0 Hz), 3.889 (d, 4 H, J ) 9.0 Hz), 3.518
(s, 4 H), 3.502 (d, 4 H, J ) 7.7 Hz), 2.724 (d, 6 H, J ) 14.2 Hz),
1.907 (s, 6 H), 1.717 (d, 2 H, J ) 12.7 Hz), 1.662 (d, 4 H, J ) 14.4
Hz).
Acid-Amide 3. Triethylamine (250 µL) was added to diacid chloride
6 (0.103 mmol) dissolved in dioxane (20 mL), and the flask was sealed
with a septum. Ammonia (0.5 M in dioxane, 205 µL, 1 equiv) was
added through the septum, and the slightly cloudy mixture was stirred
at room temperature for 20 h. An aqueous solution of 1 N HCl (1
mL) was added to quench the reaction. The solvent was removed under
reduced pressure, and the residue was taken up in CH2Cl2 (20 mL)
and washed with 1 N HCl(aq) (2 × 20 mL) and brine (20 mL). The
organic layer was dried over Na2SO4, filtered, and concentrated in
Vacuo. The crude material was purified by flash chromatography
through a silica gel column (using a solvent system gradient of 5, 7,
20% acetone/CH2Cl2) to give a mixture of acid-amide 3 and diamide
4 (30 mg). Diacid was recovered (26.6 mg, 22%). Rechromatography
of the mixture (4, 8% acetone/CH2Cl2) yielded pure acid-amide 3 (7
mg, 6%) and diamide 4 (20 mg, 16%) as white solids: for 3 mp 86.1-
87.5 °C; IR (KBr) 3421, 3179, 2860, 1700, 1453, 1400, 1365, 1193,
1101, 737, 698 cm-1; 1H NMR (300 MHz, CDCl3) δ 13.80 (br s, 1 H),
8.495 (br s, 1 H), 7.35-7.19 (m, 30 H), 7.095 (s, 1 H), 6.828 (2, 1 H),
6.030 (br s, 1 H), 4.59-4.43 (m, 12 H), 3.892 (t, 4 H, J ) 9.9 Hz),
3.467 (dd, 4 H, J ) 2.7, 8.8 Hz), 3.397 (s, 2 H), 3.311 (s, 2 H), 2.747
(t, 2 H, J ) 12.2 Hz), 2.570 (d, 2 H, J ) 13.7 Hz), 2.434 (d, 2 H, J )
14.4 Hz), 1.935 (s, 3 H), 1.927 (s, 3 H), 1.73-1.59 (m, 6 H); HRMS
(FAB in 3-nitrobenzyl alcohol) calcd for C74H78N3O13 (M + H),
1216.5535; found, 1216.55215.
Diamide 4. Acid chloride 6 (0.0821 mmol) dioxane (3 mL), and
TEA (200 µL, 17 equiv) were combined in a round-bottomed flask
and then sealed with a septum. Ammonia solution (0.5 M in dioxane,
500 µL, 3 equiv) was added through the septum, and the suspension
was stirred for 24 h. The reaction mixture was concentrated, taken up
in CH2Cl2 (20 mL), and washed with 1 N HCl (20 mL). The layers
were separated, and the organic layer was washed with 1 N HCl (2 ×
10 mL) and with brine (10 mL), then dried over Na2SO4, filtered, and
concentrated. The crude white solid was purified by flash chroma-
tography through a silica gel column (5, 10% acetone/CHCl3) to give
diamide 4 (78.2 mg) in 78% yield: mp 88.4-89.7 °C; IR (KBr) 3477,
3168, 2859, 1735, 1694, 1496, 1453, 1400, 1365, 1194, 1100, 736,
Acknowledgment. We thank Joanne Yun for helpful dis-
cussions and Prof. Perry Frey for stimulating correspondence.
This work was supported by the National Institutes of Health.
JA960288L
(35) Connors, K. A. Binding Constants; John Wiley & Sons, Inc.: New
York, 1987.
(36) Systat 5.2, 5.2 ed.; Systat Inc.: Evanston, IL, 1992.
(34) TeXsan Single Crystal Analysis Package, V. 1.7-1; 1.7-1 ed.;
Molecular Structure Corporation: The Woodlands, TX, 1995.