Highly enantioselective oxazaborolidine-catalyzed reduction of 1,3-dicarbonyl compounds: Role of the additive diethylaniline

Article abstract of DOI:10.1021/ol900258f

The oxazaborolidine-catalyzed reduction of 2,2-disubstituted cycloalkan-1,3-diones or hindered 2,2-disubstituted cyclic ketones using catecholborane as reductant proceeds with greater enantioselectivity when N,N-diethylaniline is added. It has now been sh

Full text of DOI:10.1021/ol900258f

ORGANIC
LETTERS
2009
Vol. 11, No. 7
1611-1614
Highly Enantioselective
Oxazaborolidine-Catalyzed Reduction of
1,3-Dicarbonyl Compounds: Role of the
Additive Diethylaniline
Rong-Jie Chein, Ying-Yeung Yeung, and E. J. Corey*
Department of Chemistry and Chemical Biology, HarVard UniVersity Cambridge,
Massachusetts 02138
corey@chemistry.harVard.edu
Received February 6, 2009
ABSTRACT
The oxazaborolidine-catalyzed reduction of 2,2-disubstituted cycloalkan-1,3-diones or hindered 2,2-disubstituted cyclic ketones using
catecholborane as reductant proceeds with greater enantioselectivity when N,N-diethylaniline is added. It has now been shown that the effect
of this additive is to catalyze the conversion of a harmful minor impurity in catalyst preparations to the active catalyst.
We recently described a practical and short enantioselective
synthetic route to estrone from the Torgov diketone (1) based
Scheme 1
.
Practical and Short Enantioselective Synthetic Route
to Estrone from Torgov Diketone
on the highly enantioselective and diastereoselective reduc-
tion of 1 to the ꢀ-hydroxy ketone 2 with a catalytic amount
of the chiral oxazaborolidine 3 and stoichiometric cat-
echolborane using N,N-diethylaniline as an additive (Scheme
1).1 The use of this additive had a dramatic and beneficial
effect on the enantioselectivity (96: 4) and also accelerated
the reaction. Remarkably also, in the absence of N,N-
diethylaniline the opposite enantiomer predominated some-
what (by between 60: 40 and 75: 25 in various experiments).1
Our initial surmise was that the additive opened up the
possibility of an alternative mechanistic pathway that is
distinctly different from the normal CBS pathway.2 This
paper reports on a study to ascertain the basis for the salutary
effect of N,N-diethylaniline on the CBS reduction of 1,3-
dicarbonyl compounds, which was found to be quite general
for a whole series of cyclic 1,3-dicarbonyl substrates.1
(1) Yeung, Y.-Y.; Chein, R.-J.; Corey, E. J. J. Am. Chem. Soc. 2007,
The reduction of achiral cyclic 1,3-dicarbonyl compounds
by a chiral reagent is complicated by the fact that the two
carbonyl groups are diastereotopic and not equivalent. As a
129, 10346–10347.
(2) Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986–
2012.
10.1021/ol900258f CCC: $40.75
Published on Web 03/11/2009
2009 American Chemical Society

Table 1. Effect of Borane Source on the Oxazaborolidine-Catalyzed Reduction of 2,2-Dibenzylcyclopentanone
d
e
entrya
borane (equiv)
BH3-THF (1.0)
BH3-Me2S (1.0)
BH3-PhNEt2 (1.0)
catecholborane (1.8)
catecholborane (1.8) + PhNEt2 (0.5)
T (°C)
time
conversion (%)b
ee (%)c
[R]D
config
1
2
3
4
5
23
23
23
-50
-50
2 h
2 h
2 h
3 h
3 h
100
100
100
73
95
97
92
-62
92
-35.2
-36.0
-34.6
+23.3
-34.5
R
R
R
S
54
R
a In each case the catalyst was prepared from (S)-1,1-diphenylpyrrolidinemethanol and n-BuB(OH)2 by heating in toluene at reflux in a Dean-Stark
b
apparatus containing 4 Å molecular sieves. Conversion measured by H NMR analysis. c Determined by HPLC analysis using a Chiracel OD column.
1
d Rotation determined in CHCl3 solution (c ) 1.0). e Absolute configuration of 5 was established by X-ray diffraction analysis (see Supporting Information).
result, in the case of Torgov diketone 1 four different
compounds can be produced by reduction, since there are
enantiomeric forms of the hydroxy/methyl cis or trans
diastereomers.
expected enantiomer of the reduction product resulted when
BH3•THF or BH3•Me2S were used as stoichiometric reduc-
tant (without N,N-diethylaniline) (Table 1, entries 1-2).
2,2-Bis-(4-bromobenzyl)cyclopentanone (6) was also stud-
ied in similar experiments with catalyst 3 and catecholborane,
with and without N,N-diethylaniline in toluene at -50 °C,
and displayed the same divergent behavior, as shown by the
data in Table 2 for the reduction product 7.
Scheme 2
Since the effect of N,N-diethylaniline on oxazaborolidine
catalyzed reduction of hindered ketones seemed to be
specified for catecholborane, an interaction between diethy-
laniline and catecholborane was investigated by 1H and 11
B
NMR analysis. However, no evidence of such an interaction
could be found. We further established that catecholborane
is monomeric in benzene by cryoscopic determination of
molecular weight, thus ruling out the complication of more
than one form of the borane in hydrocarbon solution.
To study the question of whether N,N-diethylaniline causes
a change in the face selectivity of reduction catalyzed by
oxazaborolidine 3, we examined the reduction of a simple
monoketone acetophenone. With this substrate in toluene at
-50 °C with catalyst 3 and catecholborane as stoichiometric
reductant, the product was (R)-1-phenylethanol, the expected
product of the regular CBS pathway.2 The same was true
with the somewhat less reactive 4-methoxyacetophenone and
tert-butyl phenyl ketone as substrates under these conditions.
In all of these experiments the catalyst 3 was prepared from
n-butylboronic acid and (S)-1,1-diphenylpyrrolidinemethanol
in toluene at reflux in a Dean-Stark apparatus containing 4
Å molecular sieves. In contrast, when the reduction was
carried out with the more hindered substrate 2,2-dibenzyl-
cyclopentanone (4), quite divergent results were obtained
with catecholborane in toluene at -50 °C in the presence or
absence of N,N-diethylaniline, as summarized in Table 1,
entries 4 and 5. For each of these experiments the reduction
was incomplete after a reaction time of 3 h, consistent with
the relative unreactivity of this hindered ketone. The dramatic
change in the absolute stereochemical course of the reduc-
tion with added N,N-diethylaniline places 2,2-dibenzylcy-
clopentanone in line with the hindered 2,2-disubstituted-1,3-
dicarbonyl class of substrates.1 On the other hand, the
Scheme 3
We then investigated the possibility that a minor amount
of impurity in the oxazaborolidine catalyst might be a factor.
1H NMR and 11B NMR analysis of a typical batch of catalyst,
prepared using the standard method (heating (S)-1,1-diphe-
nylpyrrolidinemethanol and n-butylboronic acid at prolonged
reflux in toluene with a Dean-Stark apparatus containing
heat-activated (250 °C) 4 Å molecular sieves for removal
of water) revealed the presence of a few percent of the
impurity 83 (Scheme 2). In the past catalyst prepared in this
way functioned well in most cases.
1612
Org. Lett., Vol. 11, No. 7, 2009

Table 2. Variation of Enantioselectivity in the Oxazaborolidine-Catalyzed Reduction of 2,2-(4-Bromobenzyl)-cyclopentanone with
Coreactants
d
entrya
borane (equiv)
BH3-Me2S (1.0)
catecholborane (1.8)
catecholborane (1.8)+PhNEt2 (0.5)
T (°C)
time
yield (%)b
ee (%)c
[R]D
1
2
3
23
-50
-50
1 h
8 h
8 h
99
97
98
95
-94
93
+2.9
-2.9
+2.8
a In each case the catalyst was prepared from (S)-1,1-diphenylpyrrolidinemethanol and n-BuB(OH)2 by heating in toluene at reflux in a Dean-Stark
apparatus containing 4 Å molecular sieves. b Isolated yield. c Determined by HPLC analysis using a Chiracel OD column. d Rotation determined in CHCl3
solution (c ) 1.0).
It was possible to eliminate the impurity 8 by conducting
the preparation of catalyst 3 in a Soxhlet apparatus containing
a mixture of potassium hydride and sand for removal of
water. When this very pure catalyst was used for the
reduction of the Torgov diketone 1 with catecholborane in
toluene at -50 °C the reduction product 2 was obtained in
89% yield and 93% ee, a result which is essentially the same
as obtained earlier in the presence of N,N-diethylaniline.
Thus, it is clear that the use of N,N-diethylaniline is
unnecessary with catecholborane as reductant if pure catalyst
3 is used, and that its beneficial effect comes into play only
when small amounts of the impurity 8 are present in the batch
of catalyst 3 which is employed. Further, the deleterious
effect of the impurity 8 is manifested only with hindered
ketones which react more slowly and when catecholborane
is used as the stoichiometric reductant.
analysis. Catalyst 3 so prepared and catecholborane reduced
diketone 1 to the hydroxy ketone 2 (in toluene at -50 °C)
in high yield and 92% ee. In contrast, the reaction of the
hydrate 8 and catecholborane alone with the Torgov diketone
1 gave principally the enantiomer of 2 (36% ee).
There are a number of possible reasons for the harmful
effects of 8 in CBS reductions of hindered ketones and 1,3-
diketones. Perhaps the most obvious possibility is that 8 can
compete with the ketone substrate by binding to the complex
of catecholborane and catalyst 3, thus inhibiting the normal
CBS reduction pathway. Another explanation is that the
presence of 8 results in promotion of undesirable side
reactions such as destruction of catecholborane or catalysis
of ketonic reduction to form the enantiomer of the normal
CBS reduction product.
Given the problems caused by the presence of impurity 8
in certain of the reductions catalyzed by 3 with catecholbo-
rane as reductant, we have investigated an alternative
method4 for the generation of catalyst 3 which avoids the
possibility of forming 8. Reaction of n-BuBCl2 with (S)-
1,1-diphenylpyrrolidinemethanol in benzene with 2 equiv of
triethylamine rapidly formed 3 and a precipitate of Et3NHCl.
Simple filtration with exclusion of air and moisture provided
a solution of pure 3 (Scheme 4). Use of 3 prepared in this
way for the reduction of the Torgov diketone 1 with
catecholborane gave the desired hydroxy ketone 2, in 87%
yield and 92% ee, and without the use of N,N-diethylaniline.
n-BuBCl2 was prepared from dichloroborane and 1-butene.5
Scheme 4
Studies using the hydrated oxazaborolidine 8 revealed a
relatively slow reaction with catecholborane at -50 °C in
toluene which is greatly accelerated upon addition of N,N-
diethylaniline. The catalyzed reaction forms H2 and the
catecholboric anhydride 9, as determined by NMR analysis
(Scheme 3). These results imply that the beneficial effect of
diethylaniline derives from its acceleration of the removal
of the deleterious impurity 8 by conversion to the catalyst
3. Reaction of the oxazaborolidine hydrate with N,N-
diethylaniline and NaH in toluene cleanly converted it to
(4) To a solution of (S)-1,1-diphenylpyrrolidinemethanol (253 mg, 1.0
mmol, from Aldrich) and triethylamine (280 mL, 2.0 mmol) in benzene (1
mL) in a 10-mL round-bottomed flask equipped with a stir bar was added
n-BuBCl2 (1 M in toluene, 1.0 mL, 1.0 mmol) at 23 °C. The resulting
solution was heated to 40 °C for 1 h. After stirring at 23 °C for another
1 h, solvents were removed in Vacuo (ca. 0.1 mmHg). The residue was
diluted with benzene (2 mL) and the suspension was filtered through a
sintered glass funnel under nitrogen. The ammonium salt was washed with
benzene (5mL) and the benzene was removed in Vacuo (ca. 0.1 mmHg).
Toluene (5.0 mL) was added to provide a 0.20 M solution of the pure
oxazaborolidine 3.
the oxazaborolidine 3 as shown by by H and 11B NMR
1
(3) See, Jones, T. K.; Mohan, J. J.; Xavier, L. C.; Blacklock, T. J.;
Mathre, D. J.; Sohar, P.; Jones, E. T. T.; Reamer, R. A.; Roberts, F. E.;
Grabowski, E. J. J. J. Org. Chem. 1991, 56, 763–769.
Org. Lett., Vol. 11, No. 7, 2009
1613

This method of preparation of catalyst 3 may be advanta-
geous for catalyzed reductions using catecholborane as
reductant.
In conclusion, the beneficial effect of N,N-diethylaniline
in the reduction of the Torgov diketone (1) mediated by
catalyst 3 and catecholborane is due to the ability of this
amine to catalyze the destruction of a minor impurity (8) in
preparations of 3 by the process shown in Scheme 3.
Acknowledgment. Rong-Jie Chein is the recipient of a
Taiwan Merit Scholarship. We are grateful to Dr. Nate
Wallock of Sigma-Aldrich for helpful discussions and gifts
of reagents.
Supporting Information Available: Detailed experimen-
tal procedures and characterization for all new compounds
and X-ray crystallographic data. This material is available
free of charge via the Internet at http://pubs.acs.org.
(5) Soundararajan, R.; Matteson, D. S. J. Org. Chem. 1990, 55, 2274–
2275.
OL900258F
1614
Org. Lett., Vol. 11, No. 7, 2009