AlIII-calix[4]arene catalysts for asymmetric Meerwein-Ponndorf-Verley reduction

Article abstract of DOI:10.1021/cs5001976

Chiral Al^III-calixarene complexes were investigated as catalysts for the asymmetric Meerwein-Ponndorf-Verley (MPV) reduction reaction when using chiral and achiral secondary alcohols as reductants. The most enantioselective catalyst consisted o

Full text of DOI:10.1021/cs5001976

Research Article
pubs.acs.org/acscatalysis
AlIII−Calix[4]arene Catalysts for Asymmetric Meerwein−Ponndorf−
Verley Reduction
Partha Nandi,*,† Andrew Solovyov, Alexander Okrut, and Alexander Katz*,†
†Department of Chemical and Biomolecular Engineering, University of California, Berkeley, California 94720, United States
S
* Supporting Information
ABSTRACT: Chiral AlIII-calixarene complexes were inves-
tigated as catalysts for the asymmetric Meerwein−Ponndorf−
Verley (MPV) reduction reaction when using chiral and achiral
secondary alcohols as reductants. The most enantioselective
catalyst consisted of a new axially chiral vaulted-hemispherical
calix[4]arene phosphite ligand, which attained an enantiose-
lective excess of 99%. This ligand consists of two lower-rim
hydroxyl groups, with the remaining two lower-rim oxygens
directly connected to the phosphorus of the phosphite, which
is derived from a chiral diol. The results emphasize the
importance of the rigid calix[4]arene lower-rim substituents
and point to a possible role of a lower-rim chiral pocket and Lewis-basic phosphorus lone pairs in enhancing asymmetric hydride
transfer.
KEYWORDS: MPV reduction, chiral, asymmetric hydride transfer, Lewis-acid catalysis, calixarene complexes, phosphite ligand
homogeneous as well as in grafted Al(III)-calixarene sites on
silica.7,8 This class of catalyst bridges the homogeneous−
heterogeneous gap in that both homogeneous and grafted
Al(III)-calixarene-on-silica variants of this molecular catalyst
have the same per-site MPV activity. Also, in the case of the
homogeneous catalyst, we demonstrated that the calixarene
enabled synthesis of a molecular pocket, which affected
accessibility and catalytic rate at the Al center.7,8 Here, we
build on the tunability of calixarene-based catalysts, with the
synthesis of chiral 1,3-disubstituted lower-rim calixarene
ligands, including new axially vaulted chiral hemispherical
calixarene catalysts based on phosphite substituents, and
demonstrate their catalytic utility for MPV reduction.
INTRODUCTION
■
The Meerwein−Ponndorf−Verley (MPV) reaction is a mild
reduction method for ketones, which is catalyzed using
nontoxic and earth-abundant main group elementsin this
case, Lewis acidic Al(III)1−4and can be directed to introduce
asymmetric carbons in prochiral ketones. There are several
applications of this reaction, including a stereoselective variant
that has been recently used for the synthesis of pharmaceutical
building blocks for anti-HIV therapeutics.3 In general,
asymmetric MPV reduction can be tuned by using either a
chiral alcohol as a sacrificial reductant or a chiral Lewis acid
complex as a catalyst. Here, in this article, we investigate the
essential catalyst structural features for asymmetric MPV
reduction using Al(III)-calixarene complexes, in which the
metal is placed in a chiral oxo environment. Our results
demonstrate enantioselective Al-based catalysts for MPV
reduction, which are among the few that accomplish this in
the absence of chiral alcohol.5,6
RESULTS AND DISCUSSION
■
Our investigation of asymmetric MPV reduction used
previously reported enantiopure chiral hemispherical calix[4]-
arene ligands 1a−1c, shown in Table 1,9 in which the
asymmetric carbon is directly attached to the calixarene lower
rim. We synthesized Al(III) complexes 2a−2c using these
ligands (Table 1). This was accomplished by treating 1a−2c
with 1 equivalent (with respect to calix[4]arene diol) of
trimethylaluminum in toluene at room temperature for 3 min,
followed by the addition of 4 equivalents (with respect to
ketone substrate) of secondary alcohol as MPV reductant.
Table 1 lists yields and enantioselectivity as measured by chiral
gas chromatography for MPV reduction of 2-chloroacetophe-
Our approach leverages lower-rim-substituted cone Al(III)-
tert-butylcalix[4]arene complexes, which are tunable. We
recently demonstrated these complexes as highly active
homogeneous-catalyst sites for MPV reduction.7,8 The Al-
1
(III)-calixarene complex remained intact as observed using H
NMR spectroscopy during catalysis. The crucial role of the
calixarene is to enforce active-site isolation in these catalysts,
thereby avoiding aggregation of Al-alkoxide-type species,7
which leads to coordinatively saturated hexacoordinate Lewis
acid sites, which are catalytically inactive. This class of catalyst is
2-fold more active per Al site compared with freshly prepared
aluminum isopropoxide, and active-site isolation was charac-
terized previously using 27Al NMR spectroscopy both in
Received: February 13, 2014
Revised: May 6, 2014
© XXXX American Chemical Society
2492
dx.doi.org/10.1021/cs5001976 | ACS Catal. 2014, 4, 2492−2495

ACS Catalysis
Research Article
observed for the reaction and catalysts in Table 1, when using
isopropanol as a hydride donor instead of the chiral alcohol.
Guided by the observation above of increasing MPV
reduction enantioselectivty upon incorporation of noncovalent
contacts when using catalyst 2c, we aimed to further increase
MPV reduction enantioselectivity using even more bulky
phosphite lower-rim substituents, which are derived from
axially chiral diol ligands. Our approach involved the design and
synthesis of catalysts 3a−3e in Table 2. The phosphite in these
catalysts is directly attached to the calixarene lower rim and
provides a Lewis base consisting of the phosphorus lone pair in
proximity to the Lewis acid Al(III) center. 31P NMR
spectroscopy of the ligands before and after complexation
with Al does not show evidence of a P···Al interaction (i.e.,
treatment with Me3Al in toluene failed to show characteristic
shifts in the 31P NMR resonance of the phosphite that would be
indicative of direct communication of a phosphorus lone pair
with aluminum). This may suggest that the phosphorus and Al
comprise a frustrated Lewis pair. Similar frustrated Lewis pairs
have been previously shown to bind and activate hydrogen as
well as transfer hydrogenation catalysts.10−13 We hypothesized
that the frustrated Lewis pair in our system may also act to
facilitate asymmetric hydride transfer in the MPV reduction,
when using catalysts 3a−3e. Related benzyl- and fluorenyl-
capped lower-rim disubstituted calix[4]arene phosphite ligands
have been previously reported and have been used as ligands
for asymmetric catalysis, in which transition metals have been
complexed within a chiral hemispherical cavity on the lower rim
of a calix[4]arene macrocycle.14−17 Crucially, in contrast to
these previously reported calixarene ligands, those used for
synthesis of complexes 3a−3e consist of calix[4]arene lower-
rim OH groups, which are available for direct covalent
attachment to the Al(III) metal center.
Table 1. Asymmetric MPV Reduction with Chiral
Calix[4]arene Ligands
a
Conversion measured by GC.
none at room temperature, when using a chiral hydride donor
consisting of (S)-2-butanol. Data show that when increasing the
steric bulk of the catalyst, lower-rim substituents from α-phenyl
methyl in 1a to α-naphthyl methyl in 1b, the enantioselectivity
of reduction modestly increased from barely detectable levels
up to 20%, as shown by entries 1 and 2, respectively, in Table 1.
Such a result is consistent with our previous correlation of
greater degree of chirality transfer throughout the calix[4]arene
scaffold, as quantified by the geminal coupling constant
corresponding to the splitting of calix[4]arene diastereotopic
bridging hydrogens,14−17 when using more bulky and
conformationally rigid lower-rim substituents. On the basis of
this correlation, we attempted to increase enantioselectivity
further by increasing the rigidity of the calix[4]arene lower-rim
substituents. We hypothesized that specific dative interactions
between the Al(III) site and lower-rim substituents might be
beneficial in this regard and therefore incorporated a Cl
substituent in ligand 1c, which was hypothesized to facilitate a
noncovalent Cl···Al interaction in complex 2c. This hypothesis
was based on a previous demonstration of the catalytic
significance of Cl···Al interactions in MPV reduction reactions,
which were hypothesized to account for nearly an 8-fold
increase in MPV reduction catalytic rate and were supported by
single-crystal X-ray diffraction data as well as density-functional
theory-based molecular modeling.7,8 Entry 3 of Table 1 using
catalyst 2c shows an increase in the ee to 40% for the same
model reaction. This result in conjunction with entries 1 and 2
in Table 1 clearly demonstrates the catalyst structural features
that control enhancement of MPV reduction enantioselectivity
and specifically highlights the crucial role of lower-rim
calixarene substituent rigidification in this process. No ee was
The design and synthesis of 3a−3e is modular insofar as it
can be expanded to include a variety of chiral metal complexes,
when using other enantiopure phosphites derived from
commercially available chiral diols, as well as achiral diols.
The synthesis of catalysts 3a−3e is accomplished according to
Table 2, by treating unfunctionalized tert-butylcalix[4]arene
with a toluene solution consisting of two equivalents of
enantiopure chlorophosphite. This mixture was stirred over-
night at room temperature, in the presence of excess
triethylamine, and the reaction was followed using 31P NMR
spectroscopy. The chlorophosphite was synthesized in a
previous step without isolation, according to literature
precedent,18 by treating the corresponding enantiopure
BINOL with PCl3 in the presence of excess triethylamine in
toluene (monitored via 31P NMR spectroscopy). The synthesis
of 3a−3e crucially proceeds with the calix[4]arene adopting a
1
cone conformation according to H and 13C NMR spectros-
copies. This conformation is preferred because it is the one that
leads to a hemispherical cavity on the lower rim, which can
serve as a chiral pocket during catalysis.
Table 2 shows the results of the asymmetric MPV reduction
when using catalysts 3a−3e, isopropanol as a secondary alcohol
hydride donor, and various substituted acetophenones. To the
best of our knowledge, the 99% ee achieved with catalyst 3a
and ortho-fluorobenzophenone as a ketone reactant represents
the highest ee reported for Al-catalyzed asymmetric MPV
reduction catalysis, when using an achiral alcohol as a hydride
donor. At higher fractional conversions approaching 0.8, the ee
of this reaction decreases to 80% ee when using 3a as a catalyst.
Changing the ketone reactant to ortho-chloroacetophenone
2493
dx.doi.org/10.1021/cs5001976 | ACS Catal. 2014, 4, 2492−2495

ACS Catalysis
Research Article
Table 2. Asymmetric MPV Reduction through Distally Positioned Chiral Substituents in Calix[4]arene Ligands
entry
catalyst
chiral diol
R1
Ph
X
% yield
% ee
1
2
3
4
5
6
7
8
3a
3a
3a
3a
3b
3c
3d
3e
(S)-BINOL
(S)-BINOL
(S)-BINOL
(S)-BINOL
(S)-DINOL
(S)-TADDOL
(S)-VANOL
(R)-VAPOL
F
1
73
83
75
76
80
60
20
99
80
Ph
F
a
CH3
CH3
Ph
Cl
Cl
F
60
40
22
12
95
05
Ph
F
Ph
F
Ph
F
a
Reaction conducted at 0 °C.
resulted in a lower ee of 40% under otherwise identical
conditions in entry 4 of Table 2. This was increased up to 60%
ee by decreasing the temperature from room temperature to 0
°C in entry 3 of Table 2. Catalyst 3a shows poor
enantioselectivity for slightly smaller ketones that lack halogen
substituents, as shown by acetophenone reduction, which has
2494
dx.doi.org/10.1021/cs5001976 | ACS Catal. 2014, 4, 2492−2495

ACS Catalysis
Research Article
almost no ee. This result is unlikely to be due to any possible
two-point ketone binding to the active site, since we previously
demonstrated that such more rigid two-point versus one-point
binding is not an attribute that leads to higher enantioselec-
tivity.5−7,19,20 It is interesting to note that F ketone entries in
Table 2 show high enantioselectivity despite what we previously
hypothesized to be much weaker F···Al interactions relative to
Cl···Al interactions, and Cl ketones in Table 2 show lower
enantioselectivity.7 The latter hypothesis was based on a
comparative analysis of substituted ketones and their catalytic
activity for MPV reduction, which demonstrated enhanced
catalytic rates for Cl-substituted versus either F- or H-
substituted ketones. The latter two types of ketones showed
similar MPV reduction activity and were inferred to bind
weaker as a result of this.7 Here, when using MPV
enantioselectivity rather than rate as a probe, it is likely that
other factors besides halogen···Al interactions play a role for
causing high enantioselectivity. These could include the
presence of a chiral pocket that is formed by lower-rim
calixarene substituents for catalysts in Table 2. Previously, we
demonstrated that the presence of such a pocket acted to
decrease the MPV reduction rate for bulkier substrates as well
as when using bulky lower-rim substituents.7 If cavity effects are
important, the data in Table 2 suggest that the right sterically
tight fit, which is correlated with lower activity and yield, results
in higher enantioselectivity. Such a result is consistent with the
relatively poor observed MPV reduction activity when using
catalyst 3a and enantioselective recognition in other supra-
molecular host pockets, in which guests binding more tightly
into hosts exhibit a greater degree of stereoselective
discrimination.21
The degree of π-delocalization of the P lone pair was
decreased in DINOL- and TADDOL-derived phosphite
catalysts in entries 5 and 6 of Table 2, respectively, which
both lack direct aryl oxygen-P connectivity in catalysts 3b and
3c, respectively. This delocalization may be important for
electronic transmission of chiral information to the metal
center. Consistent with this and in contrast, VANOL-derived
phosphite catalyst 3d possesses extended π-delocalization, and
this catalyst exhibits both good yield and enantioselectivity in
entry 7 of Table 2. Yet if delocalization is too great, as in (R)-
VAPOL-based catalyst 3e, this may hamper the electronic
communication and results in lower enantioselectivity in entry
8 of Table 2 when using this catalyst relative to 3d.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: partha.nandi@gmail.com.
*E-mail: askatz@berkeley.edu.
■
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) Corma, A.; Domine, M. E.; Nemeth, L.; Valencia, S. J. Am. Chem.
Soc. 2002, 124, 3194−95.
(2) Boronat, S.; Corma, A.; Renz, M. J. Phys. Chem. B 2006, 110,
21168−74.
(3) Flack, K.; Kitagawa, K.; Pollet, P.; Eckert, C.; Richman, K.;
Stringer, J.; Dubay, W.; Liotta, C. L. Org. Process Res. Dev. 2012, 16,
1301−06.
(4) Kiyoharu, N.; Node, M. Chirality 2002, 14, 759−67.
(5) Campbell, E. J.; Nguyen, H. Angew. Chem., Int. Ed. 2002, 41,
1020−22.
(6) Graves, C. R.; Zhou, H.; Stern, C. L.; Nguyen, S. T. J. Org. Chem.
2007, 72, 9121−4.
(7) Nandi, P.; Matvieiev, Y. I.; Boyko, V. I.; Durkin, K. A.; Kalchenko,
V. I.; Katz, A. J. Catal. 2011, 284, 42−49.
(8) Nandi, P.; Tang, W.; Okrut, A.; Kong, X.; Hwang, S.-J.; Neurock,
M.; Katz, A. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 2484−89.
(9) Solovyov, A.; Notestein, J. M.; Durkin, K. A.; Katz, A. New J.
Chem. 2008, 32, 1314−1325.
(10) Miller, A. J. M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc.
2010, 132, 3301−3.
(11) Karkamkar, A. J.; Parab, K.; Camaioni, D. M.; Neiner, D.; Cho,
H. M.; Nielsen, T. K.; Autrey, T. Dalton Trans. 2013, 42, 615−19.
(12) Chase, P. A.; Welch, G. C.; Jurca, T.; Stephan, D. W. Angew.
Chem., Int. Ed. 2007, 46, 8050−8053.
(13) Spies, P.; Schwendemann, S.; Lange, S.; Kehr, G.; Frohlich, R.;
Erker, G. Angew. Chem., Int. Ed. 2008, 47, 7543−7546.
(14) Dieleman, C.; Steyer, S.; Jeunesse, C.; Matt, D. J. Chem. Soc.,
Dalton Trans. 2001, 2508−2517.
(15) Jeunesse, C.; Dieleman, C.; Steyer, S.; Matt, D. J. Chem. Soc.,
Dalton Trans. 2001, 881−886.
(16) Semeril, D.; Matt, D.; Toupet, L. Chem.Eur. J. 2008, 14,
7144−55.
(17) Paciello, R.; Siggel, L.; Roper, M. Angew. Chem. Int. Ed. 1999,
̈
38, 1920−1923.
(18) Ini, S.; Oliver, A. G.; Tilley, T. D.; Bergman, R. G.
Organometallics 2001, 20, 3839−3841.
(19) Ooi, T.; Miura, T.; Maruoka, K. Angew. Chem., Int. Ed. 1998, 37,
2347−2349.
(20) Ooi, T.; Ichikawa, H.; Maruoka, K. Angew. Chem., Int. Ed. 2001,
40, 3610−3612.
In summary, a chiral cavity alone is insufficient for
enantioselective hydride delivery in MPV reduction, when
using chiral Al(III)-calix[4]arene complexes, since catalysts 2a−
2c possessing only a chiral pocket, even a rigidified one as in 2c,
are unable to perform asymmetric MPV reduction, when using
achiral 2-propanol as a hydride donor. The directed lone pair of
the Lewis basic P on the calixarene-phosphite substituent
clearly plays a significant role in direct asymmetric hydride
delivery to a ketone in the absence of a chiral alcohol hydride
donor, when using catalysts 3a−3e. We hypothesize that
directed chirality of the phosphite P works synergistically with
the presence of a chiral hemispherical pocket, as defined by
calix[4]arene lower-rim substituents for directing asymmetric
MPV reduction catalysis.
(21) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. J.
Am. Chem. Soc. 2004, 126, 3674−75.
ASSOCIATED CONTENT
* Supporting Information
This material is available free of charge via the Internet at
http://pubs.acs.org.
■
S
2495
dx.doi.org/10.1021/cs5001976 | ACS Catal. 2014, 4, 2492−2495