Development of a new dimeric cyclophane ligand: Application to enhanced diastereo- and enantioselectivity in the catalytic synthesis of β-lactams

Article abstract of DOI:10.1021/jo049804d

We detail the synthesis of a new C₂-symmetric bis(cyclophane) ligand system that can be thought of as electronically analogous to binol, but which possesses the added third dimension of cyclophane chirality. The ligand synthesis involves a spontaneous (but unexpected) atropisomerization to the desired product. We have employed this ligand to form a metal complex that is an effective cocatalyst for the highly enantio- and diastereoselective catalytic asymmetric synthesis of a β-lactam.

Full text of DOI:10.1021/jo049804d

Develop m en t of a New Dim er ic Cyclop h a n e
Liga n d : Ap p lica tion to En h a n ced
SCHEME 1. Syn th esis of Diol Cou p lin g P r ecu r sor 4
Dia ster eo- a n d En a n tioselectivity in th e
Ca ta lytic Syn th esis of â-La cta m s
Harald Wack, Stefan France, Ahmed M. Hafez,
William J . Drury III, Anthony Weatherwax, and
Thomas Lectka*
Department of Chemistry, New Chemistry Building,
J ohns Hopkins University, 3400 North Charles Street,
Baltimore, Maryland 21218
We envisioned the construction of ligand 1 from dimer-
ization of a suitably functionalized cyclophane monomer,
such as 4 (Scheme 1). The synthesis of 4 can be ac-
complished straightforwardly starting with the known
lectka@jhu.edu
Received February 3, 2004
2
carbamate 2, conveniently available in either racemic
Abstr a ct: We detail the synthesis of a new C2-symmetric
bis(cyclophane) ligand system that can be thought of as
electronically analogous to binol, but which possesses the
added “third dimension” of cyclophane chirality. The ligand
synthesis involves a spontaneous (but unexpected) atrop-
isomerization to the desired product. We have employed this
ligand to form a metal complex that is an effective cocatalyst
for the highly enantio- and diastereoselective catalytic
asymmetric synthesis of a â-lactam.
or optically pure forms as described by Pamperin et al.
Iodination of either racemic or (R)-2, followed by depro-
tection by ethereal hydrazine and reprotection as the
methyl ether, afforded the coupling precursor 4 in 87%
overall yield from 2.3
Formation of a di-(()-copper(I)-ate complex from
iodide (()-4 and its oxidation with molecular oxygen gave
4
two new major products. These were shown to be both
5
the “quasi-meso” (a racemic mixture of (S,R
R,R
dimer 5 ((S,R
p
,R) and
(
p
,S) in 12% yield) and the “quasi-dl” form of the
,S) and (R,S ,R), 21% yield) (eq 1). To our
We thought it would be of interest to devise a new
sterically bulky biaryl ligand system wherein the chiral
bulk not only projects “horizontally” back from the metal
center but also “vertically,” up and down from the site of
catalysis, taking advantage of the added “third dimen-
p
p
satisfaction, repeating the reaction with optically pure 4
produces one major dimeric product 5. A single-crystal
X-ray structure of 5 obtained from the coupling of (R)-4
6
established its absolute configuration as (S,R
p
,S), the
1
sion” of cyclophane chirality, in contrast to binol, which
opposite of what we had desired.
projects much less steric bulk in the vicinity of the metal.
The result is a ligand system that possesses both planar
and axial chirality and which may be appropriate for
applications in which the use of binol is unsatisfactory.
To demonstrate, in a preliminary fashion, the utility of
this new ligand system, we apply an Al(III)-based com-
plex of ligand 1 in the bifunctional, catalytic, asymmetric
synthesis of â-lactams. The use of the ligand-metal
complex in conjunction with a cinchona alkaloid cocata-
lyst provides enhanced yields of product in high diaste-
reo- and enantioselectivity.
Surprisingly, treatment of (R,S
C not only resulted in the removal of one methyl ether
but also initiated a concomitant atropisomerization to the
desired (R,R ,R)-6 diastereomer (eq 2). Its configuration
p 3
,R)-5 with BBr at -40
°
p
was also implied by the presence of an intramolecular
hydrogen bond in the IR spectrum, necessitating proxim-
ity between the methoxy and hydroxyl groups. Depro-
(2) Hopf, H.; Grahn, W.; Barrett, D. G.; Gerdes, A.; Hillmer, J .;
Hucker, J .; Y. O.; Kaida, Y. Chem. Ber. 1990, 123, 841-845.
(3) Pamperin, D.; Schulz, C.; Hopf, H.; Syldatk, C.; Pietzsch, M. Eur.
J . Org. Chem. 1998, 7, 1441-1446.
(
1) For some examples of the use of cyclophanes as ligands, see: (a)
Worsdorfer, U.; V o¨ gtle, F.; Nieger, M.; Waletzke, M.; Grimme, S.;
Glorius, F.; Pfaltz, A. Synthesis 1999, 4, 597-602. (b) Hou, X.-L.; Wu,
X.-W.; Dai, L.-X.; Cao, B.-X.; Sun, J . Chem. Commun. 2000, 1195-
(4) Whitesides, G. M.; SanFilippo, J . J .; Casy, C. P.; Panek, E. J . J .
Am. Chem. Soc. 1967, 89, 5302-5303.
1
196. (c) Dahmen, S.; Brase, S. Chem. Commun. 2002, 26-27. (d)
(5) The stereochemical assignment for the coupled products adhere
Inoue, M. B.; Velazquez, E. F.; Medrano, F.; Ochoa, K. L.; Galvez, J .
C.; Inoue, M.; Fernando, Q. Inorg. Chem. 1998, 37, 4070-4075. (e) Pye,
P. J .; Rossen, K.; Reamer, R, A.; Tsou, N. N.; Volante, R. P.; Reider, P.
J . J . Am. Chem. Soc. 1997, 119, 6207-6208.
to the following denotation: (A, B
stereochemistry of the first cyclophane moiety; B
p
, C): where A represents the absolute
represents the
p
stereochemistry about the pivot bond (pivot or axial isomers); and C
represents the configuration of the second cyclophane moiety.
1
0.1021/jo049804d CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/25/2004
J . Org. Chem. 2004, 69, 4531-4533
4531

p
SCHEME 2. Syn th esis of (R,R ,R)-Bis(cyclop h a n e)
Diol 1
p
F IGURE 1. DFT structure of (R,R ,R)-1 (B3LYP/6-31G*).
tection of either the (R,S
situ generated TMSI (83% yield) or the (R,R
of monoether 6 with BBr resulted in the desired (R,R
diol ligand 1 (Scheme 2).
p
,R)-isomer of diether 5 with in
,R)-isomer
,R)-
p
3
p
F IGURE 2. MMFF-calculated comparison of isomer stability
for (R,R ,R)-14 and (R,R ,R)-1.
p
p
SCHEME 3. Syn th esis of th e P u ta tive
Bis-(Cyclop h a n yl d iol)AlOTf Com p lex
Although diffractable crystals of the product 1 were
not obtained, confirmation of its structure was achieved
through the observation of a hydrogen bond in the IR
p p
spectrum analogous to that in (R,R ,R)-6 (the (R,S ,R)
configuration of diol 1 would be unable to achieve this
intramolecular H-bond for steric reasons), as well as the
standard assortment of analytical data (NMR, MS). Also
of note are the molecular mechanics calculations we
performed on the cyclophane diol 1 and relevant inter-
mediates. Monte Carlo calculations employing the MMFF
force field predict that the (R,R
kcal/mol more stable than the (R,S
R,R ,R)-6 is more stable than (R,S ,R)-6 by 8.6 kcal/mol,
whereas (R,S ,R)-5 is more stable than (R,R ,R)-5 by 2.0
p
,R)-diol 1 should be 9.1
p
,R). Similarly,
(
p
p
p
p
7
kcal/mol. These predictions are also mirrored in the DFT
calculations (B3LYP/6-31G*) we performed that indicate
a strong preference for (R,R
by 5.2 kcal/mol (see Figure 1 for DFT structure).
To quantify the stabilization energy provided by the
hydrogen bond, as a control, we modeled the difluorinated
isostere 14, in which the hydroxyl groups are replaced
by fluorines (Figure 2). MMFF calculations predict that
p p
,R)-diol 1 over (R,S ,R)-diol
calculations, it is probably not solely responsible for the
observed atropisomer equilibrium. Given the hydrocarbon
based nature of the ligand system, both MM and DFT
calculations should prove very reliable. In summary,
these calculations explain the experimental data very
well from a thermodynamic standpointsnamely, why
p
8
1
the (R,R
its (R,S ,R) counterpart. Thus, we can conclude that while
hydrogen bonding accounts for almost 3 kcal/mol of
stabilization for the (R,R ,R) isomer according to MM
p
,R)-14 should be 6.7 kcal/mol more stable than
(R,S ,R)-5 may be produced in the coupling reaction and
p
why the atropisomerizations to (R,R
p p
,R)-6 and (R,R ,R)-1
occur spontaneously upon deprotection. We can also
conclude that on the basis of these calculations (R,R ,R)-1
p
p
must be formed in the deprotection.
Mixing (R,R
initiates immediate evolution of 2 equiv of CH
loss of the phenolic protons (evidenced by H NMR)
Scheme 3). Likewise, stirring (R,R ,R)-1 with TiCl
( PrO) and molecular sieves causes the solution to turn
2
p
,R)-1 with either Zn(CH
3
)
2
or Al(CH
3 3
)
(
6) Employing (S)-4 in the coupling reaction provided the (R,S
enantiomer in identical yield.
7) We used the program Macromodel (version 7.0) for the calcula-
tions (Schrodinger, Inc.).
p
,R)
4
1
gas and
(
(
p
2
-
(
.0.
8) Calculations were performed using the Titan Program, version
i
2
4
532 J . Org. Chem., Vol. 69, No. 13, 2004

bulkier chiral diol ligand system could provide enhance-
ments in selectivity and yield and thus afford the desired
results.
We screened complex 8 (10 mol % BQ 9, stoichiometric
proton sponge as base) in the formation of â-lactam 13
1
1
(
a potent inhibitor of human leukocyte elastase that we
need in large quantities for screening), under standard
conditions (acid chloride 11, imino ester 12, toluene
solvent, -78 °C), and found the diastereoselectivity of the
reaction improved to 99:1. The enantioselectivity also
improved to 99%, and the chemical yield increased to
85%, combining for a nearly ideal result. When the
nonchelating “quasi-meso” diol 1 was examined as a
control, the diastereoselectivity dropped to 6:1 cis/trans;
the yield also dropped (70%). These results confirmed our
original hypothesis that a bulky chiral diol ligand with
appropriate stereochemistry attached to a metal could
have positive effects on reaction selectivity and yield
F IGURE 3. B3LYP/6-31G* structure of the putative bis-
cyclophanyl diol)AlCH complex 7.
(
3
deep red, similar in color and appearance to the analo-
gous complex obtained from binol. In contrast, the same
experiments repeated with “quasi-meso” diol 1 resulted
in reduced (or sporadic) gas evolution and with no
apparent formation of a deep-red Ti(IV)-based complex.
(eq 3).
p
A DFT (B3LYP/6-31G*) structure of complex (R,R ,R)-7
is shown in Figure 3, revealing a metal surrounded by
the considerable chiral bulk of the ligand. The methyl
group of complex 7 could be burned off to form the
corresponding triflate with 1 equiv of triflic acid to form
8
. Alternatively, in situ generation of complex 8 could
be accomplished through treatment of diol 1 with Al-
OTf) and proton sponge base.
We immediately sought to establish the utility of the
(
3
new ligand system in a Lewis acid-catalyzed asymmetric
reaction as a test. An ideal candidate from our point of
view is the bifunctional, Lewis acid/nucleophile-catalyzed
synthesis of â-lactams recently developed in our labora-
In conclusion, further studies on ligand system 1 and
9
tories. We discovered that Lewis acids (In(III), Zn(II),
its use as a ligand for chiral Lewis acid-catalyzed
reactions are underway and will be reported in due
course.
and Al(III), among others) work in tandem with cinchona
alkaloid derivatives such as benzoylquinine (BQ, 9) to
produce â-lactams in high chemical yield and enantiose-
lectivity in most cases. However, a side effect of some of
our metal-cocatalyzed reactions is reduced diastereo-
selectivity (dr) and, in a few cases, enantioselectivity. We
reasoned that the reactions could be brought to high
levels of selectivity by the choice of an appropriate metal
complex, perhaps one that was significantly bulkier than
a bare metal salt. In earlier efforts, we screened a number
of ligand-metal complexes, with limited success. For
example, the team of Al-binol(OTf) 10 and BQ 9 (each
Ack n ow led gm en t . T.L. thanks the NIH (GM
64559), the Sloan and Dreyfus Foundations, and Merck
Co. for generous support. S.F. thanks the UNCF,
Merck, and Pfizer for graduate fellowships. H.W. thanks
J ohns Hopkins for Whittaker Chambers and Sonneborn
Fellowships. We also thank Drs. Victor G. Young, J r.,
and Maren Pink (Minnesota) and Dr. Glenn P. A. Yap
0
&
(
Delaware) for obtaining crystal structures of 5 and 6,
respectively.
1
8
0 mol %) afforded product in moderate ee (80%), but only
Su p p or tin g In for m a tion Ava ila ble: General procedures
for the synthesis of ligand 1, its precursors, its complexes, and
â-lactam 13. X-ray crystallographic data. This material is
available free of charge via the Internet at http://pubs.acs.org.
:1 dr.10 We were intrigued by the possibility that a
(
9) France, S.; Wack, H.; Hafez, A. M.; Taggi, A. E.; Witsil, D. R.;
Lectka, T. Org. Lett. 2002, 4, 1603-1605 and references therein.
10) When Al(OTf) was used as the cocatalyst without BINOL
J O049804D
(
3
ligand, lactam product was formed in 75% yield with 11:1 dr and 99%
ee.
(11) Firestone, R. A.; Barker, P. L.; Pisano, J . M.; Ashe, B. M.;
Dahlgren, M. E. Tetrahedron 1990, 46, 2255-2262.
J . Org. Chem, Vol. 69, No. 13, 2004 4533