Assembling Organic Receptors around Transition Metal Templates: Functionalized Catechols and Dioxomolybdenum(VI) for the Recognition of Dicarboxylic Acids

Article abstract of DOI:10.1021/jo960836d

The synthesis of two receptors for dicarboxylic acids [12]^2- and [13]^2-, based on the self arrangement of two functionalized catechols 1 and 2 around a cis-[MoO₂]²⁺ core, is described. Among the three pairs of e

Full text of DOI:10.1021/jo960836d

2
186
J . Org. Chem. 1997, 62, 2186-2192
Assem blin g Or ga n ic Recep tor s a r ou n d Tr a n sition Meta l
Tem p la tes: F u n ction a lized Ca tech ols a n d Dioxom olybd en u m (VI)
for th e Recogn ition of Dica r boxylic Acid s†
‡
§
,‡
Isabelle Pr e´ vot-Halter, Thomas J . Smith, and J ean Weiss*
Laboratoire d’Electrochimie, Facult e´ de Chimie, URA no. 405 au CNRS, Universit e´ Louis Pasteur,
4
rue Blaise Pascal, F-67000 Strasbourg, France and Department of Chemistry, Kalamazoo College,
Kalamazoo, Michigan 49006
Received May 7, 1996X
The synthesis of two receptors for dicarboxylic acids [12]2- and [13]2-, based on the self arrangement
2
+
of two functionalized catechols 1 and 2 around a cis-[MoO
2
]
core, is described. Among the three
pairs of enantiomers which may be produced during the complexation of two unsymmetrical
catecholates around one molybdenum(VI) ion, only one is observed for each catechol derivative 1
or 2. Depending on the base used during the complexation of catechols to the molybdenum atom,
the dianionic receptors obtained display different solubility properties. These molybdenum-based
receptors are chromogenic and, in methylene chloride, the affinities of the assembled receptors for
dicarboxylic acids ranging from C
4 8
to C have been assessed by UV-visible titrations after
2
-
determining the stoichiometry of the complex formation using J ob’s method. While receptor [12]
4 5 7
and C
acids, the more flexible receptor [13]2 exhibits selectivity for C
-
displays selectivity for C
and C . The binding mode of the diacids to the molybdenum receptor has been determined based
8
1
on H NMR titration. Due to the intrinsic chirality of the receptors, their binding properties versus
chiral dicarboxylic acid have been examined. The enantioselective binding of N-carbobenzyloxy
protected L and D-glutamic acid due to additional π-π interactions of the protecting group with
2
-
the receptor’s framework is reported for [12] in methylene chloride. For comparison, the
association constants of receptor [12]2 with a Boc protected L-glutamic acid and the racemic mixture
of N-carbobenzyloxy protected glutamic acid have been determined.
-
neutral molecules.6
-9
More recently, a combinatorial
In tr od u ction
Architectural control on the directionality of specific
approach based on orienting “half-receptors” around
ruthenium(II) has been developed by Hamilton and
illustrates the versatility of this approach. So far, most
of the coordinating fragments used for the orientation of
fragments around transition metal atoms have been
nitrogen-containing heterocycles, such as bipyridines,
phenanthrolines, and terpyridines. The ability of cat-
interactions such as hydrogen bonds has received con-
siderable attention in the field of molecular recognition.
In particular, adequate spacing of 2-(acylamino)pyridine
10
1
b-3
units
has provided access to numerous receptors for
dicarboxylic acids.1 Since several synthetic steps are
often required to achieve a topographical control on the
orientation of multiple hydrogen bonds, the spontaneous
assembly of smaller constituents to form receptors has
been investigated.4 As transition metals have demon-
strated their ability to gather and orient organic frag-
11
echolates to complex a large variety of transition metals
has prompted us to investigate the potential use of half-
receptors containing this coordinating species. Results
obtained using a cis-dioxomolybdenum(VI) template are
presented.
5
ments, self assembly of half-receptors around metallic
templates has been used to build specific receptors for
Exp er im en ta l Section
†
Gen er a l. All commercial reagents were used without
A preliminary account of this work has been published as a
communication: Prevot-Halter, I.; Smith, T. J .; Weiss, J . Tetrahedron
Lett. 1996, 37, 1201-1204.
2 2
purification except MoO (acac) which was purified by extrac-
tion from commercial material (5-10 g) and placed in a 60
mL glass filtering funnel (medium porosity) with three 30 mL
aliquots of boiling acetylacetone. To the resulting yellow
filtrate was added petroleum ether to yield a yellow precipitate
which was collected by filtration and washed with several
portions of petroleum ether. Spectroscopic characterization
‡
Universit e´ Louis Pasteur.
§
Kalamazoo College.
Abstract published in Advance ACS Abstracts, March 15, 1997.
X
(
1) (a) For a recent review on Molecular Recognition of Organic
Acids, see: Seel, C.; Galan, A.; de Mendoza, J . Top. Curr. Chem. 1995,
5, 101-132 and also (b) Ballester, P.; Costa, A.; Deya, P. M.; Gonzalez,
J . F.; Rotger, M. C.; Deslongchamps, G. Tetrahedron Lett. 1994, 35,
813-3816.
2) (a) Geib, S. J .; Vicent, C.; Fan, E.; Hamilton, A. D. Angew. Chem.,
7
3
(
(5) Chambron, J .-C.; Dietrich-Buchecker, C. O.; Sauvage, J .-P. Top.
Curr. Chem. 1993, 165, 131-162.
Int. Ed. Engl. 1993, 32, 119-121. (b) Owens, L.; Thilgen, C.; Diederich,
F.; Knobler, C. B. Helv. Chim. Acta 1993, 76, 2757-2774. (c) Vicent,
C.; Fan, E.; Hamilton, A. D. Tetrahedron Lett. 1992, 33, 4269-4272.
(6) Schwabacher, A. W.; Lee, J .; Lei, H. J . Am. Chem. Soc. 1992,
114, 7595-7598.
(
1
d) Alcazar, V.; Diederich, F. Angew. Chem., Int. Ed. Engl. 1992, 31,
521-1523. (e) Garcia-Tellado, F.; Albert, J .; Hamilton, A. D. J . Chem.
(7) Fujimoto, K.; Shinkai, S. Tetrahedron Lett. 1994, 35, 1915-2918.
(8) Goodman, M. S.; Hamilton, A. D.; Weiss, J . J . Am. Chem. Soc.
1995, 117, 8447-8455.
Soc., Chem. Commun. 1991, 1761-1763. (f) Alcazar-Montero, V.;
Tomlinson, L.; Houk, K. N.; Diederich, F. Tetrahedron Lett. 1991, 32,
(9) Goodman, M. S.; J ubian, V.; Hamilton, A. D. Tetrahedron Lett.
1995, 36, 2551-2554.
5
309-5312. (g) Garcia-Tellado, F.; Goswami, S.; Chang, S.-K.; Geib,
S. J .; Hamilton, A. D. J . Am. Chem. Soc. 1990, 112, 7393-7394.
3) (a) Tanaka, Y.; Kato, Y.; Aoyama, Y. J . Am. Chem. Soc. 1990,
12, 2807-2808. (b) Rebek, J ., J r.; Nemeth, D.; Ballester, P.; Lin, F.-
T. J . Am. Chem. Soc. 1987, 109, 3474-3475.
4) Schall, O.; Gokel, G. J . Am. Chem. Soc. 1994, 116, 6089-6100,
and references cited therein.
(10) Goodman, M. S.; J ubian, V.; Linton, B.; Hamilton, A. D. J . Am.
Chem. Soc. 1995, 117, 11610-11611.
(
1
(11) Representative examples from (a) Albrecht, M.; Franklin, S.
J .; Raymond, K. N. Inorg. Chem. 1994, 33, 5785-5793. (b) Nielson,
A.; Griffith, W. P. J . Chem. Soc., Dalton Trans. 1978, 1501-1506. (c)
Pierpont, C. G.; Buchanan, R. M. Coord. Chem. Rev. 1981, 38, 45-87.
(
S0022-3263(96)00836-5 CCC: $14.00 © 1997 American Chemical Society

Assembling Receptors for Recognition of Dicarboxylic Acids
of purified material was in excellent agreement with literature
data. When anhydrous conditions were required, all glass-
ware was flame-dried under inert gas flow, experiments were
run under argon, and all solvents were dried as follows.
Toluene, ether, and tetrahydrofuran were distilled from sodium/
benzophenone, and methylene chloride and acetonitrile were
J . Org. Chem., Vol. 62, No. 7, 1997 2187
three times with 2 M aqueous Na
removal of the solvents, the crude mixture was purified by
chromatography (SiO /CH Cl ) to afford 10 or 11. Compound
20 2 3
10: 1.15 g (3.3 mmol, 95%). Anal. Calcd for C21H N O
2 3
CO and water. After
1
2
2
2
2
-
1
(349.60 g mol ): C(72.39), H(5.78), N(8.04). Found: C(72.50),
1
H(5.88), N(7.90). Mp: 96 °C. H NMR CDCl
8.19 (d, J ) 8.1 Hz, 1H), 7.94 (d, J ) 8.1 Hz, 2H), 7.61 (m,
3H), 7.09 (dd, J
3
: 8.81 (s, 1H),
2
distilled from CaH . NMR spectra were recorded on a Bruker
AM 300 spectrometer, and data were processed with Bruker-
WINNMR for Windows. Chemical shifts were determined
1
) 7.3 Hz, J ) 8.1 Hz, 1H), 6.92 (d, J ) 8.1
2
Hz, 2H), 6.88 (d, J ) 7.3 Hz, 1H), 3.88 (s, 3H), 3.57 (s, 3H),
taking the solvent as an internal reference: CHCl
3
(7.26 ppm),
2.39 (s, 3H). Compound 11: 1.09 g (3.1 mmol, 90%). Anal.
-
1
CHDCl (5.32 ppm), DMSO-d (2.49 ppm). UV-visible experi-
2
5
Calcd for C21
20 2 3
H N O (349.60 g mol ): C(72.39), H(5.78),
N(8.04). Found: C(72.60), H(5.99), N(8.07). Mp: 114 °C. H
1
ments were run on a Hewlett-Packard HP8452A diode array
spectrophotometer using a 2 nm resolution in quartz cells with
an optical path length of 1 cm. FTIR spectra were recorded
on a Bruker IFS 28 apparatus with KBr pellets as samples.
Melting points were determined on a Kofler Heating Plate
Type WME and are uncorrected. Elemental analyses were
performed by the Service de Microanalyse of the Institut de
Chimie de Strasbourg. Mass spectra (FAB) were obtained on
a ZAB-HF mass spectrometer. Thin-layer chromatography
NMR CDCl
3
: 8.9 (s, 1H), 8.18 (d, J ) 8.1 Hz, 1H), 8.05 (dd, J
1
) 1.5 Hz, J
2
) 1.1 Hz, 1H), 7.87 (ddd, J
1
) 1.1 Hz, J
2
) 1.5
Hz, J
3
) 7.7 Hz, 1H), 7.71 (ddd, J
1
) 1.5 Hz, J
2
) 1.5 Hz, J
3
) 7.7 Hz, 1H), 7.57 (dd, J
(dd, J ) 7.7 Hz, J
1
) 7.7 Hz, J ) 7.7 Hz, 1H), 7.45
2
1
2
) 7.7 Hz, 1H), 7.05 (dd, J
1
) 8.1 Hz, J )
2
7.7 Hz, 1H), 6.88 (m, 3H), 3.86 (s, 3H), 3.57 (s, 3H), 2.34 (s,
3H).
Ha lf-Recep tor s 1 a n d 2. To a deoxygenated solution of
CH Cl (200 mL) containing the appropriate veratrole deriva-
(TLC) was performed using Macherey-Nagel Polygram Sil
2
2
G/UV254 (0.25 mm) and Polygram Alox N/UV254 analytical
polyethylene-coated plates. E. Merck silica gel 60 (70-230
mesh) and aluminum oxide 90 (70-230 mesh) were used for
column chromatography.
Bor on ic Acid 4. To a solution of veratrole (3) (10.0 g, 72
mmol) in 50 mL of THF at 0 °C was added dropwise 40.5 mL
of n-BuLi (1.6 M in hexane). After 2 h at 0 °C, the solution
was transferred via cannula into 100 mL of THF containing
tive 10 or 11 (2.0 g, 5.7 mmol) was added 1.5 mL (16 mmol) of
BBr . After refluxing 2 h, the solution was carefully quenched
3
with 200 mL of MeOH, and the solvents were removed under
reduced pressure. The crude catechol derivatives were suc-
cessively dissolved and evaporated three times with 100 mL
of MeOH to remove the boronic acid as its methyl ester. The
half-receptors were isolated in a pure form as white solids.
Com p ou n d 1: 1.60 g (4.8 mmol, 85%). Anal. Calcd for
3
-1
8
.2 mL (72 mmol) of B(OMe)
3
.
The resulting solution was
C
19
H
16
N
2
O
3
, (CH
2
Cl
2
,
2 2
/ H O) (431.45 g mol ): C(55.52),
stirred for an additional 1 h at room temperature, and the
solvents were evaporated. The crude boronate was taken into
ether (250 mL), washed with 10% aqueous HCl, and then
extracted with 10% aqueous NaOH. The aqueous phase was
washed twice with ether and then acidified with concentrated
HCl. The acidic (pH ) 1) aqueous phase was extracted three
H(4.85), N(6.47). Found: C(55.25), H(4.45), N(6.90). Mp: 295
-
1
-1
°C dec. UV-visible λmax nm (ꢀ in M cm ) in MeOH: 310
-
1
(22900); FTIR ν in cm (bands): 3076 (νOsH), 1650 (νCdO),
1
1577 (δNsH); H NMR in (CD
) 8.5 Hz, 1H), 8.65 (s, 1H), 8.48 (d, J ) 8.1 Hz, 2H), 8.42 (dd,
) 8.5 Hz, J ) 8.1 Hz, 1H), 7.85 (d, J ) 8.1 Hz, 2H), 7.58
(s, 1H), 7.51 (d, J ) 8.1 Hz, 1H), 6.93 (dd, J ) 8.1 Hz, J
7.3 Hz, 2H), 6.80 (dd, J ) 7.3 Hz, J ) 8.1 Hz, 1H), 2.81 (s,
3H). Com p ou n d 2: 1.68 g (5.1 mmol, 90%). Anal. Calcd
3 2
) CO: 12.10 (s, 1H), 8.77 (d, J
J
1
2
times with ether. The organic layer was dried over MgSO
4
,
1
2
)
filtered, and evaporated to dryness. The crude boronic acid
was then purified by chromatography over silica gel (7/3
hexane/ether) to afford an oil which crystallizes on standing,
1
2
3
-1
for C19
16 2 3 2 2 2 2
H N O , (CH Cl , / H O) (431.45 g mol ): C(55.52),
7
.0 g (44 mmol, 60%) of which was used without further
H(4.85), N(6.47). Found: C(55.21), H(4.81), N(6.44). Mp: 253
1
-1
-1
purification. Mp: 72-73 °C. H NMR (CDCl
1.5 Hz, J ) 7.5 Hz, 1H), 7.11 (dd, J ) 7.5 Hz, J
H), 7.04 (dd, J ) 1.5 Hz, J ) 8.0 Hz, 1H), 3.94 (s, 3H), 3.47
s, 3H).
Am id es 8 a n d 9. The appropriate acid chloride, 5 or 6 (4.8
3
): 7.42 (dd, J
1
°C; UV-visible λmax nm (ꢀ in M cm ) in MeOH: 218 (19,-
1
-
)
1
(
2
1
2
) 8.0 Hz,
300), 250 (16400), 304 (13100); FTIR ν in cm (bands): 3070
1
1
2
(νOsH), 1641 (νCdO), 1565 (δNsH); H NMR in (CD
11.34 (s, 1H, NH), 8.47 (s, 2H), 8.20 (s, 1H), 8.12 (dd, J
Hz, J ) 7.7 Hz, 1H), 7.98 (d, J ) 8.1 Hz, J ) 8.1 Hz, 2H)
7.83 (d, J ) 7.7 Hz, 1H, H ), 7.58 (dd, J ) 7.5 Hz, J ) 7.5
) 7.5 Hz, J
) 8.1 Hz, 1H), 2.62 (s,
3
)
2
SO:
1
) 8.1
2
1
2
g, 20 mmol), was dissolved in 20 mL of THF, and 2.15 mL of
DBU (0.03 mmol) was added. Dropwise addition of 2-amino-
9
1
2
Hz, 1H), 7.34 (d, J ) 7.5 Hz, 1H), 6.84 (d, J
1
2
)
6
-methylpyridine (7) (2.16 g, 20 mmol) with further stirring
overnight and evaporation of the solvent afforded the crude
amide which was purified over a silica gel column (CH Cl
hexane: 1/1) to yield the corresponding amide as a white solid.
Compound 8: 5.3 g (18 mmol, 90%). Anal. Calcd for C13
7.5 Hz, 2H) 6.74 (dd, J
1
) 7.5 Hz, J
2
3H).
/
Assem bled Recep tor s: [MoO
2
]2 Com p lexes. The same
-
2
2
typical procedure, under argon, was followed for the synthesis
of each complex. A solution of half-receptor 1 or 2 (0.10 g,
0.31 mmol) in 100 mL of EtOH containing N(n-Bu) OH (0.65
4
mL 0.65 mmol) or KOH (0.035 g, 0.65 mmol) was added
dropwise via cannula transfer into an ethanolic solution (300
2 2
mL) of MoO (acac) (0.051 g, 0.15 mmol). After 2 h at room
temperature, the solvents were removed under vacuum. The
purification methods for the receptors were set up depend-
ing on their respective solubility properties. Com p ou n d
H
11
N
2
-
-
1
BrO (291.14 g mol ): C(53.63), H(3.90), N(9.62). Found:
C(53.40), H(3.85), N(9.42). Mp: 203 °C; H NMR CDCl
1
3
: 7.96
) 6.5 Hz, 2H), 7.37 (dd, J
) 7.3 Hz, 1H), 6.46 (d, J ) 7.3 Hz, 1H), 6.35 (d, J
8.3 Hz, 1H), 5.81 (br s, 1H), 2.43 (s, 3H). Compound 9: 5.3
BrO (291.14g
(d, J
1
) 6.5 Hz, 2H), 7.52 (d, J
.3 Hz, J
2
1
)
1
8
2
1
)
g (18 mmol, 90%); Anal. Calcd for C13
11 2
H N
-1
mol ): C(53.63), H(3.90), N(9.62). Found: C(53.68), H(4.05),
N(9.53); Mp: 99 °C; H NMR CDCl
J ) 8.5 Hz, 1H), 8.05 (dd, J
1
3
: 8.63 (br s, 1H), 8.14 (d,
{12B}[N(n -Bu )
(200 mL) and washed twice with water. Drying over MgSO
and evaporation afforded {12B}[N(n -Bu )
0.190 g (0.15 mmol, 98%). Anal. Calcd for C70
4
]
2
: The crude mixture was taken in CH
2
Cl
2
1
) 1.8 Hz, J
2
) 1.8 Hz, 1H), 7.81
4
(
)
1
dm, J ) 7.7 Hz, 1H), 7.65 (dm, J ) 8.5 Hz, 2H), 7.62 (dd, J
1
4
]
2
as an orange solid,
Mo
(1248.79 g mol ): C(67.32), H(8.06), N(6.72). Found: C(67.46),
7.7 Hz, J
2
) 8.5 Hz, 1H), 7.32 (dd, J
1
) 7.7 Hz, J
2
) 8.5 Hz,
100 6 8
H N O
-1
H), 6.92 (d, J ) 7.7 Hz, 1H), 2.42 (s, 3H).
P r otected Ca tech ols 10 a n d 11. To a solution of the
appropriate amide 8 or 9 (1.0 g, 3.5 mmol) and Pd(PPh
0.12g, 0.1 mmol) in 20 mL of toluene were added 3.5 mL of
aqueous Na CO (2 M) and 4 (0.63 g, 3.5 mmol) dissolved in 4
H(8.29), N(6.14). Mp: 115 °C. UV-visible in CH
2
Cl
2
λ
max nm
-
1
-1
-1
)
4
(ꢀ in M cm ): 304 (40400), 394 (10000). FTIR ν in cm
3
(
(bands): 1669 (νCO), 1602 (δNH), 892 and 843 (cis-νModO).
-
Negative FAB Mass Spectrometry: Calculated for {M}
)
2
3
-
mL of EtOH. All solvents were carefully deoxygenated prior
to use. After refluxing overnight, this mixture was washed
[C38
H
28
N
4
O
8
Mo] : 764.56; measured for I ) 6.9 V: 765.10
(20%). H NMR in CDCl : 8.38 (s, 2H), 8.26 (d, J ) 8.8 Hz,
H), 7.81 (broad m, 8H), 8.11 (d, J ) 1.5 Hz, 2H), 7.60 (dd, J
1
3
2
)
(
1
2
8.8 Hz, J ) 1.5 Hz, 2H), 6.59 (m, 2H), 6.40 (m, 4H), 3.07
m, 16H), 2.29 (s, 6H), 1.59 (m, 16H), 1.34 (m, 16H), 0.96 (t, J
(12) (a) For IR see: Chakravorty, M. C.; Bandyoppadhyay, D.;
Chisolm, M. H.; Hammond, C. E. Inorg. Synth. 1992, 29, 130-131. (b)
) 7.5 Hz, 24H). Com p ou n d {12B}K
2
2
‚9H O: The crude
for UV-vis see: Moore, F. W.; Rice, R. E. Inorg. Chem. 1968, 7, 2510-
2
514.
mixture was taken in MeOH and filtered over Sephadex

2
188 J . Org. Chem., Vol. 62, No. 7, 1997
Prevot-Halter
F igu r e 1. Synthesis of the half-receptors 1 and 2.
(
LH20) to afford {12B}K
2
‚9H
2
O as an orange solid, 0.125 g,
Mo (1004.87
2 in dilute conditions and in presence of a base: KOH or
(0.15 mmol, 99%). Anal. Calcd for C38
46 4 17 2
H N O K
4
(n-Bu) NOH. Depending on the base used at this stage
-1
g mol ): C(45.42), H(4.61), N(5.57). Found: C(45.28), H(4.30),
N(5.58). Mp: 123 °C. UV-visible in CH
cm ): 298 (40400), 376 (10800). FTIR ν in cm (bands): 1660
(
of the synthesis, the solubility of the targeted molybde-
num complex in organic solvents can be controlled, and
-1
3
CN λmax nm (ꢀ in M
-1
-1
+
+
4
the use of (n-Bu) N cation (vs K ) considerably enhances
νCO), 1598 (δNH), 892 and 841 (cis-νModO). Positive FAB
+
the solubility of the receptor in halogenated organic
solvents.
Mass Spectrometry: Calculated for {M} ) [C38
28 4 8
H N O -
] : 842.74; measured for I ) 9.9 V: 843.90 (75%). 1H
+
2
MoK
NMR in CD
H), 7.59 (m, 2H), 6.63 (m, 2H), 6.45 (m, 4H), 2.28 (s, 6H).
Com p ou n d {13B}[N(n -Bu ) ‚H O: The crude mixture was
taken in CH Cl (200 mL) and washed twice with water.
Drying over MgSO and evaporation afforded {13B}[N(n -
Bu ) ‚H O as an orange solid, 0.190 g (0.15 mmol, 98%).
Anal. Calcd for C70
H(8.15), N(6.53). Found: C(65.25), H(8.06), N(6.26). Mp: 113
3
CN: 8.82 (s, 2H), 8.16 (m, 4H), 7.75 (broad m,
Regardless of the base used, only one pair of enanti-
omers out of the three possible was obtained. The
reasons for ruling out the two other structures are briefly
8
4
]
2
2
2
2
discussed hereafter. The receptors [12B]2 and [13B]
-
2-
4
have been characterized by standard techniques. FTIR
clearly shows that the cis-MoO2 geometry has been
retained as two bands corresponding to the ν(ModO) are
4
]
2
2
-
1
H
102 6 9
N O Mo (1266.79 g mol ): C(65.44),
-1
-1
13
°
2
1
C. UV-visible in CH
2
Cl
86 (41000) 366 (6500). FTIR ν in cm (bands): 1671 (νCO),
577 (δNH), 892 and 849 (cis-νModO); Negative FAB Mass
2
λmax nm (ꢀ in M cm ): 252 (41700),
observed. In addition to the correct molecular mass,
-
1
mass spectrometry measurements confirmed in each
+
+
4
case, the presence of the counter cations K or (n-Bu) N .
-
-
Spectrometry: Calculated for {M} ) [C38
measured for I ) 6.0 V: 765.10 (10%). H NMR in CDCl
8
(
(
6
3
H
28
1
N
4
O
8
Mo] : 764.56;
The use of potassium as counter cation induces the
presence of nine water molecules detected in the elemen-
tal analyses of the receptors which should decrease its
H-bonding ability toward dicarboxylic acids.
Due to the two possible orientations of each catecholate
around the molybdenum atom, three pairs of enantiomers
3
:
.69 (s, 2H), 8.13 (d, J ) 8.2 Hz, 2H), 8.10 (broad m, 4H), 7.60
dd, J ) 8.2 Hz, J ) 7.35 Hz, 2H), 7.53 (broad m, 2H), 7.11
m, 2H), 6.88 (d, J ) 7.35 Hz, 2H), 6.67 (d, J ) 9.0 Hz, 2H),
1
2
.42 (d, J ) 6.0 Hz, 2H), 6.37 (dd, J
.03 (m, 16H), 2.46 (s, 6H), 1.33 (m, 16H), 1.21 (m, 16H), 0.84
t, J ) 7.5 Hz, 24H).
1
) 9 Hz, J ) 6.0 Hz, 2H),
2
(
A, B, and C could theoretically be obtained. In Scheme
-
1
, for receptor [12]2 , are represented the possible
Discu ssion
The syntheses of the half-receptors 1 and 2 were
complexes exhibiting different symmetries, which are a
function of the position of the substituents of the catechol
bidentates within the coordination sphere of the molyb-
denum. Form B is obtained from A by rotation (180°) of
the front catecholate while form C is obtained from A by
rotation (180°) of the back catecholate.
performed as depicted in Figure 1, starting from veratrole
and either p- or m-bromobenzoic acid. Low tempera-
3
ture lithiation of 3 followed by quenching with trimethyl
borate and acidic workup afforded the boronic acid 4 (60%
after chromatography over silica gel). The acid chlorides
No symmetry axis is present in isomer A while both B
5
and 6, prepared from the appropriate bromobenzoic
2
and C exhibit a C symmetry axis, and thus, will exist
acid, were readily condensed with 2-amino-6-methylpy-
as pairs of enantiomers. Consequently, if A is formed
ridine (7) in the presence of DBU to afford 8 or 9 in good
during complexation, it should be easily identified by a
more complicated H NMR spectrum. Only signals
1
yields after chromatography (SiO
2
, CH
2
Cl
2
/traces of
MeOH). The boronic acid 4 was coupled to the phenyl
spacers bearing the amide 8 or 9 by “Suzuki” coupling
corresponding to symmetrical species have been detected,
and thus, data collected in the Experimental Section
provide no evidence for the presence of form A. H NMR
1
reaction in excellent yields, using Pd(PPh
heterogeneous medium (MeOH/2 M aqueous Na
3
)
4
in alkaline
CO
2
3
/
characterization of the complexes obtained and NMR
titration performed with chiral reagents (Pirkle’s reagent)
suggested that only one racemate had been isolated.
Although isomers B and C afford the same spacing
between the H-bonding sites, C is sterically disfavored
due to the presence of the cis-dioxo ligands on the same
side of the complex as the two (acylamino)pyridine
functions. Complexation studies with glutaric acid af-
toluene, 1/1/5). Finally, deprotection of the catecholate
chelating site was achieved using stoichiometric amounts
of boron tribromide in CH
2 2
Cl , affording the half-receptor
1
in 85% yield. Syntheses of the fully assembled recep-
2
-
2-
tors [12] and [13] were realized starting from cis-
bis(acetylacetonato)dioxomolybdenum(VI) and by reac-
tion in stoichiometric amounts of the half-receptor 1 or

Assembling Receptors for Recognition of Dicarboxylic Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2189
Sch em e 1. Th r ee P ossible Ar r a n gem en ts of Su bstitu ted Ca tech ols a r ou n d MoO
2
2
Sch em e 2. P r oba ble Bin d in g Mod e for [12B] - w ith Glu ta r ic Acid
forded sufficient information to rule out structure C. If
form C would be present, the proximity of the dioxo
moiety should interfere with the binding occurring at the
recognition site. However, monitoring glutaric acid com-
1
plexation by H NMR does not show significant displace-
ment of chemical shifts for the protons located on the
carbon chain of glutaric acid, which affords good evidence
that the diacid binding occurs far from the cis-dioxo
moiety.
In addition, previously reported crystal structures of
MoO
2
species with substituted catecholates13 have shown
1
structures similar to B. Low temperature H NMR
experiments resulted in the broadening of the resonances
corresponding to the phenyl spacers; no other signals
were affected. Thus, no exchange occurs between forms
B and C, which was expected considering the nonlability
of the Mo-O bond. Thus, the complexation of dicarboxy-
lic acids with the receptor B is likely to occur as depicted
in Scheme 2.
UV-visible titrations were performed to assay the
2-
6
F igu r e 2. UV-vis titration of [12B] with adipic acid (C ).
[12B]2 ) 5.0 × 10 M and [C ] ) 6.5 × 10 M.
tions more accurate than NMR for the determination of
binding constants.
Titration curves, an example of which is given in
Figure 2 for adipic acid with [12B]2 , were analyzed
using the Letagrop program (ver 2.4.93) and a 1/1 model.
For each titration, the 1/1 stoichiometry of the complex
-
-5
-4
4 8
receptors’ affinities for diacids ranging from C to C . As
6
8
it has been reported previously, significant changes in
the electronic absorption spectrum of receptors assembled
around transition metal centers render UV-visible titra-
-
(13) See for example: Griffith, W. P.; Nogueira, H. S.; Parkin, B.
C.; Sheppard, R. N.; White, A. J . P.; Williams, D. J . J . Chem. Soc.,
Dalton Trans. 1995, 1775-1781.

2
190 J . Org. Chem., Vol. 62, No. 7, 1997
Prevot-Halter
Ta ble 1.a ,c log Ka Va lu es a s a F u n ction of Dica r boxylic Acid s Len gth a
receptor
counter cation
K+
solvent
C4
C5
C6
C7
C8
2
2
2
2
-
-
-
-
[
[
[
[
12B]
12B]
12B]
13B]
CH3CN
CH3CN
CH2Cl2
CH2Cl2
6.1 ( 0.6
5.9 ( 0.6
4.3 ( 0.4
4.5 ( 0.5
7.1 ( 0.7
6.0 ( 0.6
2.9 ( 0.3
3.5 ( 0.4
5.2 ( 0.5
7.2 ( 0.7
3.4 ( 0.3
3.9 ( 0.4
6.1 ( 0.6
8.9 ( 0.9
3.1 ( 0.3
2.6 ( 0.3
6.2 ( 0.6
6.5 ( 0.7
+
(n-Bu)4N
(n-Bu)4N
+
b
b
+
(n-Bu)4N
a
[12B]2-] ) [[13B]2-] ) 5.0 × 10 M. Data used for Ka calculation were collected every 2 nm from 240 to 260 nm, every 4 nm from
-5
[
b
2
70 to 306 nm, and every 2 nm from 308 to 314 nm. Thus, 9 solutions were analyzed for 25 wavelengths. Insoluble. c Italic corresponds
to higher association constants. For comparison, the value of log Ka for propionic acid with the half-receptor 10 has been measured in
CH2Cl2 to be 2.1 ( 0.2.
formation was confirmed by J ob’s method and the binding
1
8
+
isotherm method. Binding constants in acetonitrile (K
or (n-Bu)
+
4 2 2
N as counter cation) and also in CH Cl are
collected in Table 1.
Subunits 10 and 11 by themselves exhibit poor affinity
for mono- or dicarboxylic acids; however, binding of
diacids is effective once the two H-bonding sites are
2
maintained preorganized around the MoO template. For
comparison, association constants with propionic acid and
the half-receptor 10 has been measured and is given in
Table 1 caption. Alternative electrostatic binding occur-
ring after proton transfers from the carboxylic acid to the
pyridine has been ruled out for two reasons. As demon-
strated by earlier studies of molecular recognition through
F igu r e 3. Topographical changes induced in [12B]2- and
[13B] by free rotation around the axis of the spacer.
H-bonds,1
4-16
a proton transfer to the pyridine should give
2-
rise to a band at 332 nm in the UV region of the
spectrum, which is not observed, neither with the half-
receptors nor with the fully assembled receptors. Ad-
ditionally, the release of two protons by the dicarboxylic
acid would form a neutral receptor, and the electrostatic
interaction would occur between the dicarboxylate and
a small variation is obtained for receptor [12B]2-. Thus,
the adjustment to the length of the diacid for binding has
a smaller effect on the electronic distribution of the
2-
2-
chromophore in receptor [13B] than in receptor [12B] .
2-
As a result of this enhanced flexibility of receptor [13B]
in comparison to [12B] , the difference in the binding
+
+
the counter cations (K or NBu
For the higher values of log K
4
) present in solution.
observed (from ca. 7 to
2-
a
constants as well as the selectivity regarding the length
of the complexed diacid are significant. As observed in
1
7
8
), the corresponding ∆Gs are comprised between -8.5
and -9.5 kcal/mol, which is consistent with the formation
of 4 H-bonds between the dianionic receptors and a
neutral substrate. Despite the relative rigidity of the
2-
Table 1, the less chromogenic receptor [13B] exhibits
a different selectivity, with a peak observed for C . The
case of N-CBz-glutamic acid will be treated separately
7
1
6
2
-
[
MoO
and C
however, consistent with previous observations on [Cu-
2
(ca t)
2
]
framework, only a slight preference for
(vide infra).
C
4
5
dicarboxylic acids is noticeable, which is,
The relative selectivity of the receptor [12B]2- for
4 5 3 5 2 2
diacids C and C in CH CN and C in CH Cl may be
+
8
(phenanthroline)
2
] based receptors.
expected from molecular models (CPK) and from litera-
ture reports8 on copper(I) phenanthroline-based recep-
tors. However, in the case of copper(I) receptors, the
tetrahedral arrangement of substituted phenanthrolines
around copper(I) seems to restrict the free rotation of the
C-C bond between the phenyl spacer and the metal
coordinating moiety, thus enhancing the stability of
5
,9
This could be explained by the free rotation of the CO-
phenyl (spacer) bond enabling the receptor to adapt to
the length of the diacid carbon chain. This free rotation
is also responsible for the color changes observed during
titrations. Adjustment of the distance between the two
aminopyridine units to the length of each diacid leads to
partial weakening of the conjugation between the amide
bonds and the spacer, thus inducing a redistribution of
selectively formed C complexes. The steric compression
around the MoO template in our case is apparently not
2
the electronic density within the (catecholate)-MoO
moiety which is responsible for the visible absorption
2
sufficient for such restricted mobility which may explain
the weak difference in C and C binding.
4
5
2
-
bands. Due to its less flexible structure, [12B] is the
most chromogenic receptor. On the basis of the same
number of free rotations, [13B]2 is more flexible. As
depicted in Figure 3, free rotation of the same C-C bond
between the catecholate moiety and the phenyl spacer
allows large variations of the distance between the two
aminopyridine units in the receptor [13B]2 , while only
2 2
Considering the marked preference, in CH Cl , of
receptor [12B]2 for C5 dicarboxylic acids, and the
intrinsic chirality of these receptors, the binding of chiral
N-protected glutamic acid derivatives has been investi-
gated. In a first analysis, for the binding of pure
-
-
enantiomers, two log K have been calculated and are
a
-
listed in Table 2. The first binding constant was obtained
using a 1/1 model which corresponds to the lower limit
(
14) Hirst, S. C.; Tecilla, P.; Geib, S. J .; Fan, E.; Hamilton, A. D.
Isr. J . Chem. 1992, 32, 105-111.
15) Geib, S. J .; Hirst, S. C.; Vicent, C.; Hamilton, A. D. J . Chem.
a
of log K , and a second using a 2/1 model corresponding
to the higher limit of the corresponding association
constant.
(
Soc., Chem. Commun. 1991, 1283-1285.
(
(
(
16) Hamilton, A. D. Adv. Supramol. Chem. 1990, 1, 1-64.
From data collected in Table 2, the binding activities
17) Cram, D. J . Science 1988, 240, 760-767.
2-
of [12B] regarding the optically pure D- and L-N-CBz-
18) K. A. Connors in Binding Constants: The Measurements of
glutamic acid are clearly enhanced, which has prompted
us to examine this binding by the combination of UV-vis
Molecular Complex Stability; J ohn Wiley and Sons Eds.: New York,
987.
1

Assembling Receptors for Recognition of Dicarboxylic Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2191
Ta ble 2.a ,c log Ka Va lu es for Recep tor s [12B]2- a n d [13B]2-
L-N-Cbz-glutamic D-N-Cbz-glutamic D,L-N-Cbz-glutamic
counter
cation
L-N-Boc-
glutamic
receptor
solvent
1/1
2/1
1/1
2/1
1/1
2/1
2
2
2
-
-
-
K+
(n-Bu)4N
(n-Bu)4N
[
[
[
12B]
12B]
12B]
CH3CN
CH3CN
CH2Cl2
5.5 ( 0.6
6.6 ( 0.7
8.0 ( 0.8
9.6 ( 1.0
9.6 ( 1.0
10.6 ( 1.0
6.0 ( 0.6
7.0 ( 0.7
8.0 ( 0.8
9.8 ( 1.0
10.2 ( 1.0
11.3 ( 1.1
6.3 ( 0.6
6.7 ( 0.7
b
10.3 ( 1.0
10.6 ( 1.1
11.0 ( 1.1
5.8 ( 0.6
6.1 ( 0.6
7.0 ( 0.7
+
+
a
[12B]2-] ) 5.0 × 10 M. Data used for Ka calculation were collected every 2 nm from 240 to 260 nm, every 4 nm from 270 to 306
-5
[
b
nm, and every 2 nm from 308 to 314 nm. Thus, 9 solutions were analyzed for 25 wavelengths. Calculations not converging toward a
single K4 value. Italic corresponds to higher association constants.
c
F igu r e 4. J ob plot of the titration of [12B]2- with optically
pure L-N-CBz-glutamic acid and with L-N-BOC-glutamic acid.
F igu r e 5. Evolution of
1
H NMR spectra during titration of
[12B]2
-
with L-N-CBz-glutamic acid: ([) first complex formed,
O) second complex formed.
(
1
methods (J ob plot and K
a
), and H NMR. Attempts to
limit for log K using a 2/1 model is adequate. No
a
determine the stoichiometry of the receptor-substrate
complex formation using J ob plots have resulted in
graphs (Figure 4) in which, for both L and D N-CBz
additional information could be obtained from UV-visible
methods and attempts were made to confirm the stepwise
1
formation of the diastereomeric complexes by H NMR
species, Y ) Aexp - A
A: absorbance at a given wavelength) reaches a mini-
mum value for X) η ) ) 0.66 (η: molar
R R
X - AacidXacid (X: molar fraction;
titrations. Due to the nature and the structure of the
receptor and the substrate, only part of the proton NMR
spectrum has been precisely assigned and considered as
significant. The rest of the spectrum corresponded to the
overlay of aromatic signals from the receptor and the
N-CBz group of the substrate. In particular, the area
between 1.5 and 2.5 ppm is useful and corresponds to
the region in which the methyl groups of the aminopi-
coline function in both the free receptors and the result-
ing complexes are observed. The formation of one or
several species could be distinguished in this region as
well as at fields higher than 8.5 ppm. The evolution of
the proton NMR spectra during titration in the areas
defined above is represented in Figure 5.
This evolution indicates that upon addition of L-N-CBz-
glutamic acid an initial type of complex forms im-
mediately (peaks indexed [). This complex is stable and
continues forming until 0.5 equiv of acid is added. When
more than 0.5 equiv of acid is added, a second type of
complex is formed (peaks labeled O). As the first complex
formed ([) exhibits sharper signals than the second
complex (O), to a first approximation, this demonstrates
that the less favored complex is in fast exchange with
its dissociated form, while the first complex produced
preferentially is in slow exchange with its dissociated
form. In presence of an excess of L-N-CBz-glutamic acid
the uncompleted formation of the less favored complex
still occurs as shown by the continuous shift of the peaks
(O) corresponding to the complex. It is remarkable that
R
/(ηacid + η
R
amount; R: receptor; acid: dicarboxylic acid). This
corresponds to a 2/1 receptor/substrate stoichiometry. The
first explanation, which is less realistic, would be to
consider the specific formation of a complex in which two
receptors [12B]2 are bound to one substrate.
-
Such a possibility has been discarded on the basis that
it has never been observed in any other case for regular
dicarboxylic acids. The second option, which has been
seriously considered, is based on the asymmetric nature
of these receptors which have been considered as race-
mates. Compounds [12B]2- or [13B] are, respectively,
2-
2
-
mixtures of equivalent amounts of (λ)-[12B] and (δ)-
12B]2 , or (λ)-[13B] and (δ)-[13B] . Thus, during the
-
2-
2-
[
complexation of pure L-N-CBz-glutamic acid, pairs of
diastereomeric complexes will be formed. If the differ-
ence of stability between these two diastereomeric com-
plexes is large enough, then the stoichiometry determined
with the help of a J ob plot will be different from 1/1. The
2
/1 stoichiometry observed strongly supports the hypoth-
esis that one diastereomeric complex is very stable and
rapidly formed, while the other is not observed in the
conditions of J ob’s method. In other words, one of the
two enantiomers of the receptor is not involved in the
complexation process until the most stable diastereomeric
complex is completely formed. In these conditions, the
first approximation made when calculating an upper

2
192 J . Org. Chem., Vol. 62, No. 7, 1997
Prevot-Halter
2
2
Sch em e 3. Rep r esen ta tion of (λ)-[12B] - a n d (δ)-[12B] - a n d P ossible In ter a ction s w ith L-N-CBz-Glu ta m ic
Acid
the peaks corresponding to the most favored complex are
only slightly affected by the excess of chiral diacid added,
most of the peaks assigned being only slightly broadened
for 2.08 equiv added. The first complex formed is more
stable than the second and corresponds to the preferen-
length of the diacid chains and illustrates the fact that
strong interactions are not a prerequisite for selective
recognition and that several weak interactions appropri-
ately preorganized may be responsible for distinction
between optical isomers. Such titrations have not been
performed with receptor [13B]2 as this receptor exhibits
-
tial binding of one component of the racemic mixture of
-
[
12B]2 and thus, the preferential formation of a dia-
7
a marked selectivity for C dicarboxylic acids. The same
stereomeric complex. When a large excess of acid is
added, the equilibria leading to the formation of both
enantiomeric complexes are displaced, no free ligand
being observed and both complexes being formed in
comparable amounts. A qualitative analysis confirms
that in the presence of chiral substrate, two different
complexes are formed, with different rates of formation
and different stability constants. This is in agreement
with the hypothesis of two forms (λ and δ) of the receptor
complexing one form of a chiral substrate.
type of experiments carried out with a racemic mixture
of D,L-N-CBz-glutamic acid has given the same type of
J ob plot, binding isotherms, and the same value for the
upper limit of log K
titration data for the racemic mixture of N-CBz glutamic
acid did not converge toward a single log K value. This
a
. Using a 1/1 model for treating the
a
would suggest that the titration of the racemic substrate
should be considered as the simultaneous titration of two
enantiomers of the substrate with the appropriate enan-
tiomers of the receptors and that only one pair of
diastereomers are formed with identical association
constants. The specificity of this recognition is currently
Such behavior may result from an additional interac-
tion in the binding of one enantiomer. After duplication
of the titration and repeating the exact same experiment
with the pure D-enantiomer of the protected acid with
the racemic receptor [12B]2 , the examination of molec-
ular models led to the schematic representation depicted
in Scheme 3 and afforded a reasonable interpretation of
the observed enantioselective binding.
1
studied, mostly by 2D H NMR and titrations with other
chiral substrates. The detailed results and proofs of this
interpretation will be published separately.
-
Con clu sion
The concept of preorganizing multipoint H-bonds around
transition metal complexes developed on copper and
ruthenium with nitrogen-containing ligands may be
extended to functionalized catechols and molybdenum.
This approach offers the advantage of producing, under
mild conditions, nonlabile transition metal complexes,
As illustrated by the differences between log K
Cbz and BOC protected glutamic acids (even when the
lower limit of log K is considered), the nature of the
a
s for
a
protecting group on the glutamic acid is essential in the
selection of one enantiomer compared to the other, and
the only additional interaction which may be developed
between the receptor and the substrate is a π-π stacking
between the aromatic spacers of the receptor and the
protecting group of the acid. Scheme 3 summarizes the
favored and disfavored formation of the diastereomeric
such as receptors [12B]2 and [13B] , which, respec-
tively, display strong binding for glutaric and pimelic
acids. In addition, the receptor racemic [12B]2 displays
enhanced enantioselective recognition of optically pure
L- or D-N-(carbobenzyloxy)glutamic acid, via directed
multipoint hydrogen bonding and lipophillic interactions
with a topographical control achieved via a cis-dioxo-
molybdenum core.
-
2-
-
2
-
complexes of the L-protected acid with [12B] . This
hypothesis is in agreement, for receptor [12B]2-, with the
results of binding experiments with the optically pure
BOC protected L-glutamic acid. As expected, the J ob plot
from UV-visible data confirmed a 1/1 stoichiometry
Ack n ow led gm en t. The Centre National de la Re-
cherche Scientifique is gratefully acknowledged for
financial support. We are indebted to Pr. A. D. Hamil-
ton and Dr. M. S. Goodman for stimulating discussions
and helpful advice.
(
Figure 4), and the formation of the receptor-substrate
complex was less efficient. UV-visible titrations yielded
a log K value of 6.1 in acetonitrile and 7.0 in methylene
a
chloride similar to those obtained for glutaric acid (7.1
in methylene chloride). This selectivity is not in contra-
diction with the weak selectivity observed regarding the
J O960836D