Synthesis and metalation of a chiral, pyridine-strapped, cyclic bis(benzimidazole) ligand

Article abstract of DOI:10.1021/ol049954a

(Equation presented) Synthesis of a chiral, strapped, cyclic bis(benzimidazole) ligand is reported. This compound forms a stable, Jahn-Teller-distorted, octahedral Cu(II) complex in which the elongated axis is positioned within the main macrocycle. The C<

Full text of DOI:10.1021/ol049954a

ORGANIC
LETTERS
2004
Vol. 6, No. 6
989-992
Synthesis and Metalation of a Chiral,
Pyridine-Strapped, Cyclic
Bis(benzimidazole) Ligand†
Tomasz Fekner, Judith Gallucci, and Michael K. Chan*
Department of Chemistry, The Ohio State UniVersity, 100 W 18th AVenue,
Columbus, Ohio 43210
chan@chemistry.ohio-state.edu
Received January 7, 2004
ABSTRACT
Synthesis of a chiral, strapped, cyclic bis(benzimidazole) ligand is reported. This compound forms a stable, Jahn−Teller-distorted, octahedral
Cu(II) complex in which the elongated axis is positioned within the main macrocycle. The C -symmetric bis(benzimidazole) core adopts a
2
highly ruffled conformation with the ortho-substituted benzene rings located anti relative to the pyridine-bearing strap.
Cyclic, highly ruffled1 bis(benzimidazole)s (e.g., 1 and 2,
Figure 1) represent a new family of chiral C2-symmetric
similarly shaped substrates, the main obstacle in the use of
the unmodified ligands 1 and 2 in transition-metal-mediated
enantioselective catalysis stems from their rapid racemization
via a relatively fast macrocyclic inversion. In previous
studies, we found that introduction of bulky substituents ortho
to the amide NH groups of 2 significantly increases the
barrier to racemization.3 We also demonstrated that spanning
the two external benzimidazole nitrogen atoms with relatively
short straps renders these compounds configurationally stable.
Pursuant to our continued efforts in this arena, we
envisaged that an analogous strapping between the two
nonbenzimidazole nitrogen atoms (N5 and N13, Figure 1) in
macrocyclic compounds based around 1 and 2 could also
potentially lead to extremely ruffled, configurationally stable
Figure 1. Ruffled, cyclic, bis(benzimidazole) ligands.
(1) For classification of various deformation modes in polyazamacrocyclic
ligands, see: Scheidt, W. R.; Lee, Y.-J. Struct. Bonding (Berlin) 1987, 64,
1-70.
ligands capable of creating a well-defined asymmetric
environment around a central metal atom.2 Although we have
previously communicated principles for the design of ligands
capable of enantiofacial discrimination of (E)-alkenes and
(2) (a) Payra, P.; Hung, S.-C.; Kwok, W.-H.; Johnston, D.; Gallucci, J.;
Chan, M. K. Inorg. Chem. 2001, 40, 4036-4039. (b) Kwok, W.-H.; Zhang,
H.; Payra, P.; Duan, M.; Hung, S.-C.; Johnston, D. H.; Gallucci, J.;
Skrzypczak-Jankun, E.; Chan, M. K. Inorg. Chem. 2000, 39, 2367-2376.
(c) Payra, P.; Zhang, H.; Kwok, W.-H.; Duan, M.; Gallucci, J.; Chan, M.
K. Inorg. Chem. 2000, 39, 1076-1080.
† Dedicated to Prof. Sunney I. Chan for his constant support and guidance,
on the occasion of his retirement.
(3) Fekner, T.; Gallucci, J.; Chan, M. K. J. Am. Chem. Soc. 2004, 126,
223-236.
10.1021/ol049954a CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/25/2004

ligands. Herein, we wish to disclose the results of our
preliminary synthetic studies of this hypothesis.
As the previously reported preparation of ligand 1 was
tedious and inefficient,2a we initially sought to develop an
improved procedure that would be amenable to multigram
preparations. Our first synthesis (Scheme 1) commenced with
As the first step in the construction of the primary
macrocyclic framework, ligand 1 was prepared according
to the known,2a high-yielding procedure involving treatment
of amine 6 with HClO4 in MeCN, followed by neutralization
with Et3N (Scheme 3). Conversion of ligand 1 to the cyclic
amine 10 was achieved by reduction with NaBH3(CN) in
glacial AcOH in excellent yield.
Scheme 1. First Synthesis of Amine 6
Scheme 3. Synthesis of the Cyclic Amine 10
conversion of the known acid 4,3 obtained in eight steps and
30% overall yield from the commercially available diacid
3, to aldehyde 5 via BH3-mediated chemoselective reduction,
followed by oxidation with active MnO2. Protection of
aldehyde 5 as its ethylene glycol acetal and subsequent
catalytic hydrogenation (H2, Pd/C) proceeded smoothly
furnishing the requisite amine 6 in 70% overall yield. This
compound served as a convergence point for the two novel
syntheses of ligand 1.
Crystals of amine 10 suitable for X-ray analysis were
grown by slow evaporation of its EtOAc/CH2Cl2 solution.4
Unlike ligands 12a and 2,3 for which the macrocyclic core
adopts a well-defined ruffled conformation, amine 10 exhibits
a mode of macrocyclic deformation in the solid state that is
strikingly different (Figure 2). The two phenylbenzimidazole
The known benzimidazole 73 was selected as the starting
material in an alternative synthesis of the key amine 6
(Scheme 2). Conversion of 7 to benzylic alcohol 8 via a
Scheme 2. Alternative Synthesis of Amine 6
Figure 2. ORTEP view (50% probability thermal ellipsoids) of
the cyclic amine 10. The oxygen and nitrogen atoms are hatched.
Hydrogen atoms are omitted for clarity.
components are parallel to each other but shifted so that there
is no intramolecular π-stacking. When the molecule is viewed
along the aryl-heteroaryl axis, the macrocycle creates a
three-step protocol involving catalytic hydrogenation (H2, Pd/
C) of the nitro group, subsequent protection of the resulting
amine as its benzyl carbamate, and reduction of the ester
function with LAH in THF proceeded uneventfully in 61%
overall yield. Sequential oxidation of alcohol 8 with active
MnO2, followed by acetalization (ethylene glycol, pTSA),
produced benzimidazole 9 in 59% overall yield. Finally,
hydrogenolysis of the amino protecting group freed the
desired amine 6 in excellent yield.
(4) Crystal data for 10: C30H26N6; MW ) 470.57; colorless, rhombus-
shaped chunk, 0.31 × 0.35 × 0.38 mm; monoclinic; space group P21/n; T
) 200 K; λ ) 0.71073 Å; a ) 9.2514(1), b ) 7.3897(1), c ) 17.7530(3)
Å; â ) 100.5505(5)°; V ) 1193.17(3) Å3; Z ) 2; Dcalc ) 1.310 Mg/m3;
F(000) ) 496; µ(Mo KR) ) 0.080 mm-1; 30 188 reflections collected with
2.71 < θ < 27.47°, 2723 of which were independent (Rint ) 0.030); 173
parameters; R1 ) 0.0460, wR2 ) 0.1193 [for reflections with I > 2σ(I)];
R1 ) 0.061, wR2 ) 0.1580 (all data).
990
Org. Lett., Vol. 6, No. 6, 2004

perfectly rhomboidal cavity in the center (the dihedral angle
between the joined aryl and heteroaryl components, which
corresponds to the smaller angle in the rhomboidal cavity,
is equal to 65°).
These results suggest that the dramatic ruffling exhibited
by bis(benzimidazole) ligands (e.g., 1 and 2) is linked, at
least in part, to the restricted rotation about the nonbenz-
imidazole carbon-nitrogen bonds (C6-N5 and C14-N13,
Figure 1). Introduction of a double-bond (e.g., 1) or double
bond-like character (e.g., 2) at this position likely induces
the macrocycle to adopt the more typical and, arguably, more
useful conformation.
to their inherent nonplanarity, a distortion of the core
macrocycle in ligand 11 is not a prerequisite for the
handedness of the entire molecular system.
When a blue solution of ligand 11 and Cu(BF4)2 in MeOH
was kept at reflux (Scheme 5), its color gradually changed
Scheme 5. Metalation of the Strapped, Cyclic Amine 11
Modeling studies suggested, however, that an alternative
way to promote the desired ruffling might be to join the two
nonbenzimidazole nitrogen atoms in 10 with a short, rigid
strap. As such structural elements restrict axial binding of
external ligands, we chose a strap possessing a nucleophilic
center that could coordinate the metal directly. We were also
cognizant of the fact that the incorporation of this additional
axial ligand would block the face it occupies, leaving only
the trans face open for chemistry, a potentially desirable
element for future catalytic studies.
to green, and the Cu(II) complex 12 was formed in high
yield. Crystals of complex 12 suitable for X-ray analysis were
obtained by slow evaporation of its MeOH/CH2Cl2 solution.8
The structure reveals a six-coordinate species with a pseudo-
octahedral geometry (Figure 3). The four internal nitrogen
Double alkylation of the cyclic amine 10 with 2,6-bis-
(bromomethyl)pyridine (Scheme 4) was accomplished by
Scheme 4. Synthesis of the Pyridine-Strapped Amine 11
heating the mixture in boiling MeCN in the presence of
Na2CO3 according to the methodology of Guilard et al.5 The
desired strapped ligand 11 was formed in good yield.6,7
Importantly, unlike ligands 1 and 2 that are chiral solely due
(5) (a) Achmatowicz, M.; Hegedus, L. S.; David, S. J. Org. Chem. 2003,
68, 7661-7666. (b) Wynn, T.; Hegedus, L. S. J. Am. Chem. Soc. 2000,
122, 5034-5042. (c) Brande`s, S.; Lacour, S.; Denat, F.; Pullumbi, P.;
Guilard, R. J. Chem. Soc., Perkin Trans. 1 1998, 639-641. (d) Denat, F.;
Lacour, S.; Brande`s, S.; Guilard, R. Tetrahedron Lett. 1997, 38, 4417-
4420.
Figure 3. ORTEP view (50% probability thermal ellipsoids) of
the Cu(II) complex 12. The oxygen and nitrogen atoms are hatched.
Hydrogen atoms are omitted for clarity.
(6) Although in the literature, the terms “capping” and “strapping” are
very often regarded as synonymous, with the former frequently covering
all of the possible topologies, we intentionally differentiate between them,
in particular when aromatic rings or similar components are used to block
one of the macrocyclic faces. When the aromatic ring is, more or less,
parallel to the averaged macrocyclic plane, the former serves as a cap. In
contrast, when the aromatic ring is approximately perpendicular to the
averaged macrocyclic plane, it acts as a strap.
atoms of the main macrocycle bind in an equatorial fashion
to the copper ion, while the two remaining ligands are
provided by the nucleophilic centers of methanol and the
(8) Crystal data for 12: C39H39B2CuF8N7O2; MW ) 874.93; green,
rectangular plate, 0.04 × 0.27 × 0.35 mm; monoclinic; space group P21/c;
T ) 150 K; λ ) 0.71073 Å; a ) 1.595(1), b ) 16.922(2), c ) 19.919(3)
Å; â ) 106.12(1)°; V ) 3754.7(7) Å3; Z ) 4; Dcalc ) 1.548 Mg/m3; F(000)
) 1796; µ(Mo KR) ) 0.669 mm-1; 55 721 reflections collected with 2.13
< θ < 24.99°, 6608 of which were independent (Rint ) 0.038); 538
parameters; R1 ) 0.0515, wR2 ) 0.1252 [for reflections with I > 2σ(I)];
R1 ) 0.0667, wR2 ) 0.1367 (all data).
(7) As the crude reaction product is spectroscopically different (1H NMR)
from the purified amine 11, it is postulated that the crude reaction mixture
contains a sodium complex of the capped amine 11. The cation may serve
as a template on which strapping takes place. However, as a reviewer pointed
out, other phenomena (partial protonation, hydration, kinetic alkylation, etc.)
might be responsible for the observed spectroscopic discrepancies.
Org. Lett., Vol. 6, No. 6, 2004
991

strap-incorporated pyridine. The Cu-O(methanol) and Cu-
N(pyridine) distances are equal 1.98 and 1.99 Å, respectively,
and the central metal ion exhibits some doming (0.31 Å
relative to the mean 4N-plane) in the direction of the axially
bound MeOH.
The two sets of Cu-N bonds within the primary macro-
cyclic ring are markedly different. The average length of
the Cu-N(benzimidazole) bond is 2.00 Å, which is common
for a Cu-N bond in other complexes of polyazamacrocyclic
ligands. In contrast, the average length of the Cu-N(aniline)
bond is 2.45 Å, presumably reflecting a Jahn-Teller (JT)
distortion.9,10 While such bond elongations are fairly common
in octahedral Cu(II) complexes, their orientation within the
macrocyclic 4N-plane is unusual.2c,5b,11-12 For example, in
the structure of Cu(II)‚1‚(ClO4)2, the JT distortion is per-
pendicular to the 4N-plane, with the perchlorate ions having
long, 2.6 Å, interactions with the open axial sites of the
copper ion.2c
In addition to the JT distortions at the metal center, as
anticipated, the macrocyclic core of 12 is highly nonplanar
with the two benzimidazole subunits located syn to the strap
relative to the average macrocyclic plane.12 The level of
distortion for the core 16 atoms in complex 12, however, is
significantly lower than that observed for highly ruffled
ligands 1 and 2, arguably the most distorted polyazamacro-
cyclic ligands reported to date2a,3 (Figure 4; see Figure 1 for
numbering system). The maximum distortion for 12 reaches
Figure 4. Deviation from planarity (Å) relative to the mean 4N-
plane for the core 16 atoms in ligands 1‚(HClO4)2 and 2 and the
Cu(II) complex 12. For the atom numbering system, see Figure 1.
The values of distortion for the meso carbons are shown in bold
font.
2a
0.84 Å (1.15 and 1.29 Å for ligands 1‚(HClO4)2 and 2,3
respectively). The average distortion for the core 16 atoms
in 12 is 0.40 Å, compared to 0.58 and 0.54 Å for 1‚(HClO4)2
and 2, respectively. Analogously, the average distortion for
the four meso carbons (i.e., atoms 3, 7, 11, and 15) in 12 is
0.73 Å, compared to 0.93 and 0.94 Å for 1‚(HClO4)2 and 2,
respectively.
As there is no spectroscopic evidence (1H and 13C NMR)
that ligand 11 consists of a mixture of diastereomers due to
macrocyclic inversion of the bis(benzimidazole) core, it is
possible that its geometry is well-defined and analogous to
that of complex 12 (with the two benzimidazole subunits
located syn to the strap relative to the averaged macrocyclic
plane). However, at this stage we cannot rule out very fast
interconversion (relative to the NMR time-scale) of the two
possible diastereomeric species that could also account for
the simplicity of the NMR spectra.
In conclusion, a synthesis of the pyridine-strapped, benz-
imidazole-based cyclic ligand 11 and its Cu(II) complex 12
has been developed. The core macrocycle in 12 has been
shown to adopt a significantly ruffled, well-defined confor-
mation. Work is currently underway to prepare enantio-
enriched samples of 11 and to develop asymmetric syntheses
of its analogues in order to use them as ligands in transition-
metal-catalyzed enantioselective transformations.13
(9) Jahn, H. A.; Teller, E. Proc. R. Soc. London 1937, 161, 220-235.
(10) For reviews on the Jahn-Teller effect in Cu(II) complexes, see:
(a) Bersuker, I. B. Chem. ReV. 2001, 101, 1067-1114. (b) The Jahn-
Teller Effect. A Bibliographic ReView; Bersuker, I. B., Ed.; IFI/Plenum:
New York, 1984; pp 1-590. (c) Hathaway, B. J. Struct. Bonding (Berlin)
1984, 57, 55-118. (d) Bersuker, I. B. Coord. Chem. ReV. 1975, 14, 357-
412.
Acknowledgment. This work was supported by the
National Science Foundation (CAREER Award No. 9984071).
Supporting Information Available: Synthetic procedures
and copies of 1H and 13C NMR spectra for all new
compounds and crystallographic details for 10 and 12. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(11) For representative structures of Cu(II) bound to 4N-macrocyclic
amine ligands, see: (a) Curtis, N. F.; Gladkikh, O. P. Aust. J. Chem. 2000,
53, 727-741. (b) Goeta, A. E.; Howard, J. A. K.; Maffeo, D.; Puschmann,
H.; Williams, J. A. G.; Yufit, D. S. J. Chem. Soc., Dalton Trans. 2000,
1873-1880. (c) Kim, J. C.; Fettinger, J. C.; Kim, Y. I. Inorg. Chim. Acta
1999, 286, 67-73. (d) Kang, S.-G.; Kim, S.-J.; Jeong, J. H. Polyhedron
1998, 17, 3227-3234. (e) Lancashire, R. J.; Newman, P. D.; Stephens, F.
S.; Vagg, R. S.; Williams, P. A. J. Coord. Chem. 1995, 34, 345-350. (f)
Bu, X. H.; An, D. L.; Chen, Y. T.; Shionoza, M.; Kimura, E. J. Chem.
Soc., Dalton Trans. 1995, 2289-2295. (g) Pattrick, G.; Hancock, R. D.
Inorg. Chem. 1991, 30, 1419-1422. (h) Comba, P.; Curtis, N. F.; Lawrance,
G. A.; O’Leary, M. A. J. Chem. Soc., Dalton Trans. 1988, 2145-2152.
(12) For reviews on nonplanar distortions in other classes of polyaza-
macrocyclic ligands, see: (a) Senge, M. O. In The Porphyrin Handbook;
Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: San Diego,
2000; Vol. 1, Chapter 6, pp 239-347. (b) Shelnutt, J. A.; Song, X.-Z.; Ma,
J.-G.; Jia, S.-L.; Jentzen, W.; Medforth, C. J. Chem. Soc. ReV. 1998, 27,
31-41.
OL049954A
(13) Preliminary studies indicate that the molecule of methanol in the
Cu(II) complex 12 is not tightly bound to the central metal atom. As a
result, 12 can potentially serve as a Lewis acid and activator in a host of
important transformations. For instance, it catalyzes the hydrolytic epoxide
ring opening in 1,2-epoxypropane. Therefore, the hydrolytic kinetic
resolution of racemic oxiranes catalyzed by enantioenriched samples of 12
and its analogues should be possible. For a recent report on the Cu(BF4)2-
catalyzed epoxide-ring opening, see: Barluenga, J.; Vazquez-Villa, H.;
Ballesteros, A.; Gonzalez, J. M. Org. Lett. 2002, 4, 2817-2819.
992
Org. Lett., Vol. 6, No. 6, 2004