A new family of modular chiral ligands for the catalytic enantioselective reduction of prochiral ketones

Article abstract of DOI:10.1021/jo990942q

A family of enantiomerically pure (4R,5R)-2-alkyl-4-phenyl-5-(R- oxymethyl)-1,3,2-oxazaborolidines (5) [boron substituent: H, CH₃, n-C₄H₉; R-oxy group: CH₃O, CH₃OCH₂CH₂0, CH₃(OCH₂CH₂)₂O, PhCH₂0, Ph₂CHO, Ph₃CO] has been prepared from (2S,3S)-2,3-epoxy-3-phenylpropanol (2) through a four-step sequence involving protection of the alcohol, regioselective ring-opening of the epoxide with sodium azide in acetonitrile in the presence of LiClO₄, reduction of the azido group (H₂/Pd-C/MeOH or NaBH₄/THF-MeOH), and formation of the oxazaborolidine ring with the appropriate boron reagent. Both the boron substituent and the R-oxy group have been optimized for maximal enantioselectivity in the reduction of prochiral ketones with borane. The optimal oxazaborolidine (5a-Me) [boron substituent: CH₃; R-oxy group: CH₃O] has been employed (10% molar amount, THF, 0 °C to room temperature) in the reduction of a representative family of 10 substrates comprising alkyl aryl ketones and dialkyl ketones. In these reductions, 5a-Me induces the formation of secondary alcohols of S configuration with high enantioselectivity (93% mean enantiomeric excess). The origin of the enantioselectivity in the reduction has been rationalized by means of semiempirical AM1 calculations.

Full text of DOI:10.1021/jo990942q

7902
J . Org. Chem. 1999, 64, 7902-7911
A New F a m ily of Mod u la r Ch ir a l Liga n d s for th e Ca ta lytic
En a n tioselective Red u ction of P r och ir a l Keton es
Cristina Puigjaner, Anton Vidal-Ferran, Albert Moyano, Miquel A. Perica`s,* and Antoni Riera
Unitat de Recerca en S´ıntesi Asime`trica (URSA), Departament de Quı´mica Orga`nica,
Universitat de Barcelona, Martı´ i Franque`s 1-11, 08028 Barcelona, Spain
Received J une 10, 1999
A family of enantiomerically pure (4R,5R)-2-alkyl-4-phenyl-5-(R-oxymethyl)-1,3,2-oxazaborolidines
(5) [boron substituent: H, CH3, n-C4H9; R-oxy group: CH3O, CH3OCH2CH2O, CH3(OCH2CH2)2O,
PhCH2O, Ph2CHO, Ph3CO] has been prepared from (2S,3S)-2,3-epoxy-3-phenylpropanol (2) through
a four-step sequence involving protection of the alcohol, regioselective ring-opening of the epoxide
with sodium azide in acetonitrile in the presence of LiClO4, reduction of the azido group (H2/Pd-
C/MeOH or NaBH4/THF-MeOH), and formation of the oxazaborolidine ring with the appropriate
boron reagent. Both the boron substituent and the R-oxy group have been optimized for maximal
enantioselectivity in the reduction of prochiral ketones with borane. The optimal oxazaborolidine
(5a -Me) [boron substituent: CH3; R-oxy group: CH3O] has been employed (10% molar amount,
THF, 0 °C to room temperature) in the reduction of a representative family of 10 substrates
comprising alkyl aryl ketones and dialkyl ketones. In these reductions, 5a -Me induces the formation
of secondary alcohols of S configuration with high enantioselectivity (93% mean enantiomeric excess).
The origin of the enantioselectivity in the reduction has been rationalized by means of semiempirical
AM1 calculations.
In tr od u ction
Modularly constructed ligands, which allow the fast
optimization of catalytic properties for a given applica-
tion, are increasingly recognized as a major advancement
at both the academic and the industrial levels. We have
recently applied this philosophy to the preparation of a
family of enantiopure amino alcohol ligands 1 starting
from readily available enantiopure epoxy alcohol 2.
Through an iterative process, the structural parameters
key to high catalytic efficiency in the addition of diethyl-
zinc to aldehydes were established. These parameters
turned out to be the steric bulk of the R-oxy group and
the nature of the dialkylamino substituent as a nitrogen-
containing six-membered ring.1 Further optimization of
the dialkylamino substituent through a mechanism-
guided molecular design2 led to 3a , a most convenient
ligand for the ethylation of aldehydes.
The structural specification of the optimal modules
shows some transferability among related families of
amino alcohols acting in the same reaction. Thus, the
systematic study of perhydroazines as dialkylamino
components in a family of amino alcohols derived from
synthetic, yet enantiomerically pure, triphenylethylene
oxide has led to the development of a superior enantio-
selective catalyst, 3b, for the addition of diethylzinc to
R-substituted aldehydes.3
The catalytic enantioselective reduction of prochiral
ketones stands as an alternative and complementary
method for the synthesis of chiral secondary alcohols. In
particular, the oxazaborolidine-mediated enantioselective
reduction of ketones with borane4 is a well-established
methodology: It has been applied with success to sub-
strates of many different types, and the efficient catalytic
cycle involved in the process has even been compared to
the action mode of enzymes.4b
From a mechanistic point of view, the amino alcohol-
mediated alkylation of aldehydes5 and the oxazaboroli-
dine-mediated reduction of ketones6 bear many common
characteristics (Figure 1): In both reactions, a hetero-
cyclic five-membered ring formed from an amino alcohol
and the corresponding reagent act as a disymmetric
* To whom correspondence should be addressed. E-mail: mpb@
ursa.qo.ub.es.
(1) Vidal-Ferran, A.; Moyano, A.; Pericas, M. A.; Riera, A. J . Org.
Chem. 1997, 62, 4970-4982.
(2) Vidal-Ferran, A.; Moyano, A.; Perica`s, M. A.; Riera, A. Tetrahe-
dron Lett. 1997, 38, 8773-8776.
(3) (a) Sola`, L.; Reddy, K. S.; Vidal-Ferran, A.; Moyano, A.; Perica`s,
M. A.; Riera, A.; Alvarez-Larena, A.; Piniella, J . F. J . Org. Chem. 1998,
63, 7078-7082. (b) For the preparation of the regioisomeric (S)-1,2,2-
triphenyl-2-piperidinoethanol and for its use as ligand in the ethylation
of aldehydes, see: Reddy, K. S.; Sola`, L.; Moyano, A.; Perica`s, M. A.;
Riera, A. J . Org. Chem. 1999, 64, 3969-3974.
(4) (a) Wallbaum, S.; Martens, J . Tetrahedron: Asymmetry 1992,
3, 1475-1504. (b) Singh, V. K. Synthesis 1992, 605-617. (c) Deloux,
L.; Srebnik, M. Chem. Rev. 1993, 93, 763-784. (d) Corey, E. J .; Helal,
C. J . Angew. Chem., Int. Ed. Engl. 1998, 37, 1986-2012.
(5) Yamakawa, M.; Noyori, R. J . Am. Chem. Soc. 1995, 117, 6327-
6335.
(6) Corey, E. J .; Bakshi, R. K.; Shibata, S. J . Am. Chem. Soc. 1987,
109, 5551-5553.
10.1021/jo990942q CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/21/1999

Enantioselective Reduction of Prochiral Ketones
J . Org. Chem., Vol. 64, No. 21, 1999 7903
Sch em e 1
F igu r e 1. Schematic representation of the transition states
in the amino alcohol-mediated reductive alkylation of alde-
hydes and the oxazaborolidine-mediated reduction of ketones.
Sch em e 2
template onto which the reacting molecules coordinate.
These templates present a central heteroatom with Lewis
acid character (zinc or boron) flanked in each case by two
atoms with Lewis base character. It is assumed that in
both reactions coordination of the carbonyl compound
takes place at the Lewis acidic site, and the main
difference between the reacting complexes lies on the
Lewis base type atom where the reagent coordinates:
Diethylzinc at oxygen5 and borane at nitrogen.6
In view of these similarities, it is not surprising that
structurally related amino alcohols have been employed
as ligands with similar results for both types of chemis-
try.7 According to these precedents, we considered that
the free amino alcohols 4 could be convenient precursors
to catalytically active oxazaborolidines. They belong to
a structural type that has not previously been tested in
the enantioselective reduction of ketones and contain in
their structures the additional feature of the R1-oxy-
methyl substituent that allows gradual modification of
catalytic behavior in order to improve enantioselectivity.
The work described here details our efforts in the
preparation of an optimized oxazaborolidine catalyst 5
for the enantioselective reduction of ketones, derived from
amino alcohols 4, through a modular optimization process
involving both the alkoxy group OR1 and the substituent
at boron.
important source of diversity with high relevance for the
purposes of this research, since groups R1 with greatly
different steric and electronic characteristics can be in
principle introduced in the molecule without interference
with the epoxide.
For the first part of the sequence, epoxy alcohol 2 was
subjected to a variety of protection schemes. The prepa-
ration of compounds 7a and 7d -f (which are also key
intermediates in the synthesis of ligands 1) has already
been reported.1 In these alkylations, which require the
intermediate formation of a metal alkoxide (NaH/DMF),
low-temperature (-20 to 0 °C) conditions were used to
prevent competing reactions. Under these conditions, the
alkylations took place devoid of any detectable side
reaction.
Along with these derivatives, designed to test the effect
of the steric bulk of R1 on the catalytic properties of the
target oxazaborolidines, compounds 7b,c, containing
ethyleneglycol derived R1 groups, were also prepared to
provide models of oxazaborolidines anchored to poly-
(ethylene glycol) resins with expected increased binding
ability of the ligand in the catalytic process. In these cases
(Scheme 2), an excess of base and alkylating agent was
required to obtain the final epoxy ethers in good yield,
as elimination on the alkyl halide is a competing reaction
even at 0 °C. A 79% yield was recorded for 7b with a
6-fold excess of NaH and alkyl iodide. If the amount of
both the base and alkylating agent is further increased
(10-fold excess), the alkylation reaction takes then place
very selectively (95% yield for 7c).
Resu lts a n d Discu ssion
Syn t h esis of Am in o Alcoh ols 4. The synthetic
strategy envisaged for the conversion of epoxy alcohol 2
into amino alcohols 4, precursors to oxazaborolidines 5,
is presented in retrosynthetic fashion in Scheme 1. In
this strategy, the amino group comes from the reduction
of an azide that, in turn, is introduced on epoxy ethers 7
by means of a well-precedented8 regioselective ring
opening. An initial protection of 2 would be the ultimate
source of these epoxy ethers. This protection step is an
As for the regioselective and stereospecific ring-opening
step of epoxy ethers 7 with an azide-delivering reagent,
Crotti’s method8d (which involves the use of sodium azide
in a 5 M LiClO4 solution in acetonitrile) was used. Under
these conditions, the intermediate azido alcohols 6a -f
were obtained in quantitative yield and were further
converted into the final amino alcohols without purifica-
tion. It is worth mentioning that the ring-opening step
proceeds with complete regio- and stereospecificity, since
NMR data are only in agreement with the diastereomeri-
cally pure anti azido alcohol exclusively arising from
nucleophilic attack at C-3 of the epoxy ether. Two
different procedures were used to reduce the azido to the
(7) For the use of norephedrine-derived ligands in the diethylzinc
addition to benzaldehyde, see: Soai, K.; Yokoyama, S.; Hayasaka, T.
J . Org. Chem. 1991, 56, 4264-4268. For the use of similar ligands in
the reduction of ketones, see: Quallich, G. J .; Woodall, T. M. Synlett
1993, 929-930. For proline-derived ligands in the diethylzinc addition
to benzaldehyde, see: Soai, K.; Ookawa, A.; Kaba, T.; Ogawa, K. J .
Am. Chem. Soc. 1987, 109, 7111-7115. For the use of similar ligands
in the reduction of ketones, see: Corey, E. J .; Bakshi, R. K.; Shibata,
S.; Chen, C. P.; Singh, V. K. J . Am. Chem. Soc. 1987, 109, 7925-7926.
(8) (a) Medina, E.; Vidal-Ferran, A.; Moyano, A.; Perica`s, M. A.;
Riera, A. Tetrahedron: Asymmetry 1997, 8, 1581-1586. (b) Chini, M.;
Crotti, P.; Macchia, F. J . Org. Chem. 1991, 56, 5939-5942. (c) Chini,
M.; Crotti, P.; Flippin, L. A.; Macchia, F. J . Org. Chem. 1991, 56, 7043-
7048. (d) Chini, M.; Crotti, P.; Flippin, L. A.; Gardelli, C.; Giovani, E.;
Macchia, F.; Pineschi, M. J . Org. Chem. 1993, 58, 1221-1227.

7904 J . Org. Chem., Vol. 64, No. 21, 1999
Puigjaner et al.
ess”:12 Over the past years, continuing interest on this
process has led to the development of new efficient borane
reducing agents13 and catalysts based on the oxazaboro-
lidine structure.13a,14 Once the family of amino alcohols
4 had been successfully prepared, we undertook studies
directed to their conversion into oxazaborolidines 5, with
the purpose of studying them as chiral catalysts in this
reaction. The functionalized amino alcohols 4 offer the
possibility of systematic variations of steric, electronic,
and binding characteristics of the ligand in order to
achieve an improved catalytic behavior, while the sub-
stituent at boron represents an additional source of
diversity.
Sch em e 3
Boron unsubstituted oxazaborolidines 5-H, analogous
to the ones initially used by Corey, were generated in
situ from 4 by treatment of the amino alcohol with BH3‚
Me2S (BMS)14h as shown in Scheme 4. B-Alkylated
oxazaborolidines, less air and moisture sensitive than the
B-H analogues,10a, 13a have been described as more robust
catalysts that can be used with no loss in the stereo-
control of the reaction. In the present instance, B-methyl
oxazaborolidines (5-Me) have been easily prepared in
toluene solution from amino alcohols 4 and methyl-
boroxine,14c and B-butyl oxazaborolidine 5-Bu from the
amino alcohol and butylboronic acid,13a also in toluene
(see Scheme 4). The solutions of 5-Me and 5-Bu could
be stored at room temperature under nitrogen atmo-
sphere for extended periods of time without appreciable
decomposition or decrease in activity.
Ta ble 1. Lith iu m P er ch lor a te-In d u ced Regioselective
Rin g Op en in g of Ep oxy Eth er s 7a -f by Sod iu m Azid e
F ollow ed by Red u ction of th e Azid o Gr ou p a
epoxy
overall
R1
ether reduction condns product yield (%)
Me
7a
7b
7c
7d
method A, 36 h, rt
method A, 15 h, rt
method A, 36 h, rt
method B, 17 h,
55-60 °C
4a
4b
4c
4d
91
84
81
92
CH2CH2OMe
(CH2CH2O)2Me
CH2Ph
CHPh2
CPh3
7e
7f
method B, 8 h,
55-60 °C
4e
4f
73
87
method B, 23 h,
55-60 °C
As the initial point in our reactivity study, we used
the unsubstituted oxazaborolidines 5-H derived from
4a -f as the chiral catalysts, and we tested the optimal
protecting group R1 to achieve optimal catalytic proper-
ties in the enantioselective reduction of 1-phenylethan-
1-one (8a) and 2-chloro-1-phenylethan-1-one (8b) (Scheme
5). All reactions were run at room temperature in THF
a
All the ring-opening reactions were performed treating the
starting epoxy ether 7a -f with a 5-fold excess of sodium azide in
a 5 M solution of LiClO4 in acetonitrile at 55 °C for 24 h.
amino group depending on the stability of the protecting
group R1 toward reducing agents (Scheme 3). In those
cases where R1 is stable under hydrogenolytic conditions
(6a -c), H2 with Pd/C in MeOH was chosen as the
reducing agent. Amino alcohols 4a -c were obtained in
this way in high overall yields (81-91%; see Table 1).
Hydrogenolysis-sensitive azido alcohols (6d -f) were re-
duced with sodium borohydride in THF/MeOH9 as these
conditions turned out to be harmless to benzylic ethers.
Overall yields for the preparation of 4d -f were high as
well, ranging from 73 to 92% (Table 1).
(12) Berrisford, D. J .; Bolm, C.; Sharpless, K. B. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1059-1070.
(13) (a) Corey, E. J .; Link, J . O. Tetrahedron Lett. 1989, 30, 6275-
6278. (b) Corey, E. J .; Bakshi, R. K. Tetrahedron Lett. 1990, 31, 611-
614. (c) Corey, E. J .; Link, J . O. J . Am. Chem. Soc. 1992, 114, 1906-
1908. (d) Corey, E. J .; Link, J . O. Tetrahedron Lett. 1992, 33, 4141-
4144. (e) Corey, E. J .; Yi, K. Y.; Matsuda, S. P. T. Tetrahedron Lett.
1992, 33, 2319-2322. (f) Mathre, D. J .; Thompson, A. S.; Douglas, A.
W.; Hoogsteen, K.; Carroll, J . D.; Corley, E. G.; Grabowski, E. J . J . J .
Org. Chem. 1993, 58, 2880-2888. (g) Stone, G. B. Tetrahedron:
Asymmetry 1994, 5, 465-472. (h) Salunkhe, A. M.; Burkhardt, E. R.
Tetrahedron Lett. 1997, 38, 1523-1526.
From a practical perspective, it is important to point
out that the preparation of amino alcohols 4 has been
easily carried out at the multigram scale without yield
decrease. Even in this case, purification of the rather
polar final products can be easily done by a short pad
column chromatography following the reduction step,
since both azido reduction conditions are very clean and
efficient.
(14) (a) Corey, E. J .; Chen, C. P.; Reichard, G. A. Tetrahedron Lett.
1989, 30, 5547-5550. (b) Rao, A. V. R.; Gurjar, M. K.; Sharma, P. A.;
Kaiwar, V. Tetrahedron Lett. 1990, 31, 2341-2344. (c) Mathre, D. J .;
J ones, T. K.; Xavier, L. C.; Blacklock, T. J .; Reamer, R. A.; Mohan, J .
J .; J ones, E. T. T.; Hoogsteen, K.; Baum, M. W.; Grabowski, E. J . J . J .
Org. Chem. 1991, 56, 751-762. (d) J ones, T. K.; Mohan, J . J .; Xavier,
L. C.; Blacklock, T. J .; Mathre, D. J .; Sohar, P.; J ones, E. T. T.; Reamer,
R. A.; Roberts, F. E.; Grabowski, E. J . J . J . Org. Chem. 1991, 56, 763-
769. (e) Corey, E. J .; Cheng, X. M.; Cimprich, K. A.; Sarshar, S.
Tetrahedron Lett. 1991, 32, 6835-6838. (f) Corey, E. J .; Azimioara,
M.; Sarshar, S. Tetrahedron Lett. 1992, 33, 3429-3430. (g) Corey, E.
J .; Link, J . O.; Bakshi, R. K. Tetrahedron Lett. 1992, 33, 7107-7110.
(h) Quallich, G. J .; Woodall, T. M. Synlett 1993, 929-930. (i) Quallich,
G. J .; Woodall, T. M. Tetrahedron Lett. 1993, 34, 785-788. (j) Quallich,
G. J .; Woodall, T. M. Tetrahedron Lett. 1993, 34, 4145-4148. (k)
Berenguer, R.; Garcia, J .; Gonzalez, M.; Vilarrasa, J . Tetrahedron:
Asymmetry 1993, 4, 13-16. (l) Kim, Y. H.; Park, D. H.; Byun, I. L. J .
Org. Chem. 1993, 58, 4511-4512. (m) Berenguer, R.; Garcia, J .;
Vilarrasa, J . Tetrahedron: Asymmetry 1994, 5, 165-168. (n) Corey,
E. J .; Helal, C. J . Tetrahedron Lett. 1995, 36, 9153-9156. (o) Prasad,
K. R. K.; J oshi, N. N. J . Org. Chem. 1996, 61, 3888-3889. (p) Corey,
E. J .; Helal, C. J . Tetrahedron Lett. 1996, 37, 4837-4840. (q) Corey,
E. J .; Helal, C. J . Tetrahedron Lett. 1996, 37, 5675-5678. (r) Masui,
M.; Shioiri, T. Synlett 1997, 273-274. (s) Bach, J .; Berenguer, R.;
Garcia, J .; Loscertales, T.; Manzanal, J .; Vilarrasa, J . Tetrahedron Lett.
1997, 38, 1091-1094.
Ligan d Optim ization th r ou gh th e Catalytic En an -
tioselective Red u ction of P r och ir a l Keton es. The
oxazaborolidine-mediated catalytic enantioselective re-
duction of ketones by borane, introduced by Corey6,10
following the pioneering work by Itsuno,11 represents a
paradigmatical example of “ligand-accelerated proc-
(9) Soai, K.; Yokoyama, S.; Ookawa, A. Synthesis 1987, 48-49.
(10) (a) Corey, E. J .; Bakshi, R. K.; Shibata, S.; Chen, C. P.; Singh,
V. K. J . Am. Chem. Soc. 1987, 109, 7925-7926. (b) Corey, E. J .;
Shibata, S.; Bakshi, R. K. J . Org. Chem. 1988, 53, 2861-2863.
(11) Itsuno, S.; Ito, K.; Hirao, A.; Nakahama, S. J . Chem. Soc., Chem.
Commun. 1983, 469-470.

Enantioselective Reduction of Prochiral Ketones
J . Org. Chem., Vol. 64, No. 21, 1999 7905
this ketone, oxazaborolidines 5a -c [R1 ) Me, R1)
CH2CH2OMe and R1)(CH2CH2O)2Me, respectively] in-
duced the highest stereoselectivity in the reduction (91%
ee in all three cases).
Sch em e 4
To study the possibility of cross-dependence between
the R1 and the boron substituents with respect to
enantioselectivity, chiral catalysts 5-Me derived from
4a -f were also evaluated in the reduction of prochiral
compounds 8a and 8b. Results are shown in Table 2
(entries 7-12 for 1-phenylethan-1-one 8a and entries 19-
24 for 2-chloro-1-phenylethan-1-one 8b). In full agree-
ment with the results observed for unsubstituted ox-
azaborolidines, steric congestion around the -CH2O-
unit is detrimental to stereoselectivity in the reduction
of prochiral ketones. Again, the most efficient catalyst is
the one incorporating the less bulky substituent as the
protecting group (5a -Me, R1 ) Me). Oxazaborolidine 5a -
Me induced the reduction of ketone 8a with a 96% ee
(entry 7) and in the case of 8b, with a 94% ee (entry 19).
On the contrary, catalyst 5f-Me, which incorporates the
bulkiest R1 group in the series, induced the lowest
stereoselectivity in the reduction of 8a (85% ee, entry 12)
and 8b (82% ee, entry 24). These results tend to indicate
that oxazaborolidines 5 behave as truly modular and that
the variable structural elements (R1 and the boron
substituent) can be optimized separately.
Sch em e 5
When the whole series of results with B-H and B-Me
oxazaborolidines are compared (entries 1-6 and 13-18
for B-H and 7-12 and 19-24 for B-Me), it becomes clear
that the B-Me derivatives are better catalysts. To com-
plete the optimization of this second parameter in
catalysts 5, the B-Bu substituted oxazaborolidine derived
from the amino alcohol containing an optimal R1 group
(4a ) was studied in the reduction of 8a and 8b (see
entries 25 and 26 in Table 2). As it can be seen, the
results obtained with this B-Bu-substituted oxazaboro-
lidine did not represent any improvement over the
corresponding ones with the B-Me catalyst, so that this
possibility was not further investigated.
For the best catalyst in the series (5a -Me), we have
investigated the effect of temperature on the turnover
and stereoselectivity of the reaction (see Scheme 6 and
Table 3). As the temperature was lowered from room
temperature to 0 °C, the turnover of the system was
significantly decreased. Even after extended reaction
times (2 h addition plus 75 min additional stirring),
conversions were 78% for 8a and 71% for 8b. As for the
stereoselectivity, the ee of the reduction products de-
creased from 96 to 65% in the case of 8a and from 94 to
68% in the case of 8b. On the other hand, carrying out
the reaction at 45 °C, led to the final product with
practically no loss in the enantioselectivity for both
studied ketones. According to this observation, the opti-
mal temperature for the use of 5a -Me as a catalyst seems
to be 22 °C. Interestingly, working at temperatures
slightly above this value seems to exert no negative
influence on the enantioselectivity of the process, so that
strict control of the temperature is not necessary. In any
case, as it will be discussed later, adjustment of the
optimal reaction temperature may require further refine-
ment in some instances.
using BMS as the reducing agent and a 10% molar
amount (with respect to the ketone) of oxazaborolidine
catalyst. The starting ketone was slowly added (1 h) onto
the corresponding catalyst solution with the aid of a
syringe pump. Once the addition had been completed,
conversion was very high so that the reactions could be
quenched without delay (a 10 min stirring period was
normally introduced).
The results obtained in the reduction of 8a mediated
by oxazaborolidines 5-H derived from 4a -f are shown
in Table 2 (entries 1-6). An important conclusion can
be drawn from these results. The best enantioselectivity
was recorded for the less bulky protecting group (94% ee
for R1 ) Me, entry 1). Much lower stereoselectivities were
observed with R1 ) trityl (72% ee, entry 6). For the rest
of ligands, ee’s ranged between 81% and 91% as expected
for protecting groups with a size comprised between a
methyl and a trityl group. In this respect, it is interesting
to note the existence of a regularity in the decrease of
enantioselectivity upon increasing the number of phenyl
groups and, hence, the bulkiness, in the protecting group
(91% ee for R1 ) benzyl, 81% ee for R1 ) benzyhydryl
and 72% ee for R1 ) trityl).
The same tendency was observed for 8b under the
same reaction conditions and with the same chiral
catalyst (entries 13-18 in Table 2). The lowest stereo-
induction was recorded for the largest protecting group
(84% ee for R1 ) Tr). For any other specification of R1,
selectivities were higher and ranged from 88 to 91%. With
En a n tioselective Red u ction of a Rep r esen ta tive
Set of Keton es Ca ta lyzed by th e Op tim ized Ox-
a za bor olid in e 5a -Me. As a confirmation of the validity
of the optimization process performed on catalysts 5, the
optimized oxazaborolidine 5a -Me was used in the reduc-

7906 J . Org. Chem., Vol. 64, No. 21, 1999
Puigjaner et al.
Ta ble 2. Effect of th e R1 a n R2 Gr ou p s in th e Ca ta lytic En a n tioselective Red u ction of P r och ir a l Keton esa 8a ,b Lea d in g
to Alcoh ols 9a ,bb Med ia ted by Ch ir a l Oxa za bor olid in es 5
d
chiral
catalyst
starting
ketone
resulting
alcohol
conversionc
(%)
selectivity
(%)
eee
(%)
entry
R1
R2
1
2
3
4
5
6
7
8
9
5a -H
5b-H
5c-H
5d -H
5e-H
5f-H
5a -Me
5b-Me
5c-Me
5d -Me
5e-Me
5f-Me
Me
H
H
H
H
H
H
Me
Me
Me
Me
Me
Me
8a
8a
8a
8a
8a
8a
8a
8a
8a
8a
8a
8a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
(S)-9a
>99
>99
88
>99
87
>99
>99
>99
98
99
99
>99
>99
>99
>99
>99
>99
94
86
81
91
81
72
96
94
95
92
95
85
CH2CH2OMe
(CH2CH2O)2Me
CH2Ph
CHPh2
CPh3
83
Me
>99
>99
>99
>99
>99
86
CH2CH2OMe
(CH2CH2O)2Me
CH2Ph
CHPh2
CPh3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
5a -H
5b-H
5c-H
5d -H
5e-H
5f-H
5a -Me
5b-Me
5c-Me
5d -Me
5e-Me
5f-Me
Me
H
H
H
H
H
H
Me
Me
Me
Me
Me
Me
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
(R)-9b
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
92
94
91
91
91
88
89
84
94
92
89
92
92
82
CH2CH2OMe
(CH2CH2O)2Me
CH2Ph
CHPh2
CPh3
98
>99
>99
>99
90
>99
>99
>99
>99
Me
CH2CH2OMe
(CH2CH2O)2Me
CH2Ph
CHPh2
CPh3
25
26
5a -Bu
5a -Bu
Me
Me
Bu
Bu
8a
8b
(S)-9a
(R)-9b
>99
96
>99
>99
95
85
a
All reactions were performed in THF under N2 at room temperature, using a BH3‚SMe2/ketone/catalyst molar ratio of 1.2/1/0.1. A
solution of the ketone in THF was added onto a solution of the catalyst over a period of 1 h with the aid of a syringe pump. Acidic
b
quenching was made 10 min after. The absolute configuration of the final alcohols has been established by comparing the sign of the
optical rotation with reported values for (S)-1-phenylethanol and (R)-2-chloro-1-phenylethanol. c Determined by integration of residual
d
starting material in front of all new products in the gas chromatogram of the reaction crude. Determined by integration of the reduction
products (both enantiomers) in front of other reaction products. e By GC using a â-DEX 120 column.
Ta ble 4. Ca ta lytic En a n tioselective Red u ction of a
F a m ily of P r och ir a l Keton esa 8a -j Lea d in g to Alcoh ols
9a -jb Med ia ted by Ch ir a l Oxa za bor olid in e 5a -Me
Sch em e 6
conver- selec-
resulting
alcohol
sionc
(%)
tivityd eee
starting ketone
(%)
(%)
1-phenylethan-1-one (8a )
2-chloro-1-phenylethan-
1-one (8b)
(S)-9a
(R)-9b
>99
>99
>99
>99
96
94
1-(3-methoxyphenyl)ethan-
1-one (8c)
1-(3-chlorophenyl)ethan-
1-one (8d )
1-(4-chlorophenyl)ethan-
1-one (8e)
3,3-dimethylbutan-
2-one (8f)
1-phenylpropan-1-one (8g)
1,2,3,4-tetrahydronaphthalen-
1-one (8h )
1-(2-methoxyphenyl)ethan-
1-one (8i)
(S)-9c
(S)-9d
(S)-9e
(S)-9f
>99
>99
>99
>99
>99
>99
>99
>99
97
94
93
97
Ta ble 3. In flu en ce of th e Tem p er a tu r e on th e Ca ta lytic
En a n tioselective Red u ction of Keton esa 8a ,b Lea d in g to
Alcoh ols 9a ,bb Med ia ted by Ch ir a l Oxa zobor olid in e
5a -Me
conver- selec-
chiral
T
starting resulting
sionc tivityd eee
(S)-9g
(S)-9h
94
>99
>99
>99
90f
98g
entry catalyst (°C) ketone
alcohol
(%)
(%)
(%)
1
2
3
4
5
6
5a -Me
5a -Me
5a -Me
5a -Me
5a -Me
5a -Me
0
8a
8a
8a
8b
8b
8b
(S)-9a
(S)-9a
(S)-9a
(R)-9b
(R)-9b
(R)-9b
78
>99
>99
71
>99
>99
>99
>99
>99
>99
>99
>99
65
96
95
68
94
93
(S)-9i
(S)-9j
>99
>99
98
92h
79i
rt
45
0
rt
45
1-cyclohexylethan-1-one (8j)
>99
As described in Table 2. e By GC using a â-DEX 120 column
for 8a -i and a R-DEX 120 column for 8j. f The reaction was
performed at 0 °C. Enantioselectivity at other temperatures, ee
a-d
a-e As described in Table 2 except for the reaction time in
experiments at 0 °C (2 h addition plus 75 min stirring before
quenching).
g
(%)/°C: 78/-10; 87/10; 85/22; 82/33. The reaction was performed
at 0 °C. Enantioselectivity at other temperatures, ee (%)/°C: 97/
5; 87/22. The reaction was performed at 10 °C. Enantioselectiv-
ity at other temperatures, ee (%)/°C: 85/0; 86/22; 83/33. Enan-
tioselectivity at other temperatures, ee (%)/°C: 59/0; 79/10; 77/
33.
h
i
tion of a family of structurally diverse ketones including
several alkyl aryl and dialkyl ketones (Scheme 6). Results
on catalyst efficiency (conversion and selectivity) and
enantiomeric excess of the final secondary alcohols have
been collected in Table 4. Working at room temperature
and using a 10% molar amount of oxazaborolidine
catalyst, conversions were complete after short reaction
times (1 h addition plus 10 min stirring), and the
reductions took place in a very clean manner (selectivity
>99%). With respect to enantioselectivity, alcohols 9a -f

Enantioselective Reduction of Prochiral Ketones
J . Org. Chem., Vol. 64, No. 21, 1999 7907
were obtained with >93% ee, while enantiomeric purity
was slightly lower in the other cases (9g-j). By progres-
sive lowering of the reaction temperature, a point with
maximal enantioselectivity could be located in the reduc-
tion of 8g-i. Maximal enantioselectivities for these
reductions are provided in the corresponding entries in
Table 4, while enantioselectivities at other studied tem-
peratures are given as footnotes. Up to 11 points in ee
(from 87 to 98%) can be gained in the reduction of 1,2,3,4-
tetrahydronaphthalen-1-one (8h ) by simply lowering the
reaction temperature from 22 to 0 °C. No increase of
enantiomeric purity of the resulting alcohol (79% ee)
could be recorded through variation of the reaction
temperature in the case of 1-cyclohexylethan-1-one (8j).
Much in the same way, the use of higher amounts of 5a -
Me in the reduction (up to a stoichiometric amount) did
not provoke any improvement in the enantioselectivity
of the reaction. In this context, it is worth noting that
the same reduction, when performed with the CBS
reagent,10a affords alcohol 9j of unusually low enantio-
meric purity (84% ee).
MNDO analysis of the origin of the enantioselectivity
with prolinol-based oxazaborolidines.
In any case, the calculated energy differences between
the different transition states are so small that all of
them should be considered when analyzing the observed
enantioselectivity with a new type of oxazaborolidine.
In the present instance, we studied at the AM1 level
of theory the possible modes of complexation of 1-phen-
ylethan-1-one and 3,3-dimethylbutan-2-one with the bo-
rane adduct of 5a -Me, as well as the derived transition
states, in agreement with the commonly accepted mech-
anism of the reaction. Whereas for all the studied
complexes a complete geometry optimization was per-
formed, in the corresponding transition states the critical
B-H-C distances were fixed to the values reported by
Quallich for a model system at the ab initio MP2/6-
31G(d)//MP2/6-31G(d) level of theory. This finds justifica-
tion in the fact that geometries optimized at these two
levels of theory tend to be similar, since electron correla-
tion is considered (explicitly in the MP2 ab initio calcula-
tions and implicitly in the semiempirical ones) in both
cases.
When the overall performance of 5a -Me is considered,
the mean enantiomeric excess of alcohols 9a -j obtained
with a 10% molar amount of the chiral agent is found to
be 93%. When 5a -Me is compared with the CBS catalyst,
the reference reagent for this chemistry, over a set of 8
ketones (8a -c,f-j) that have been reduced with both
oxazaborolidines, the difference in favor of the CBS
reagent between the mean enantiomeric excesses of the
resulting alcohols is only 2% (95 vs 93%).
As a preliminary step, a conformational analysis of the
N-BH3 (I) and the N-BH3/B-THF (II) adducts of ox-
azaborolidine 5a -Me was performed. In both cases, the
Ra tion a liza tion of th e Obser ved En a n tioselectiv-
ity in th e Red u ction of P r och ir a l Keton es Med ia ted
b y 5a -Me. As we have already discussed, it is widely
accepted that oxazaborolidines, through their adjacent
boron and nitrogen atoms with complementary Lewis
acid and Lewis base character, can give rise to complexes
incorporating one borane and one ketone molecule in an
arrangement very favorable for reaction. In this process,
the role of the substituents at the carbon atoms of the
oxazaborolidine ring (those of the starting amino alcohol)
is that of directing the complexation of the reactants
toward the less hindered diastereotopic face of the
oxazaborolidine ring. When a prochiral ketone (RLCORS)
is involved in the process, four possible geometries can
exist for these complexes (Figure 2) depending on the
oxygen lone pair involved in the complexation and the
enantiotopic face of the carbonyl compound offered to the
BH3 moiety. In their names, the exo/endo term refers to
the value of the dihedral angle CH3-B-OdC (nearly 0°
in exo complexes and nearly 180° in endo ones), whereas
Re/Si refers to the topicity of the face of the carbonyl
group opposite to the BH3 moiety.
preferred conformation of the CH3OCH2CHCH- frag-
ment turned out to be of the g+a type, other conformers
being at least 2 kcal‚mol-1 less stable. According to this,
g+a conformers were considered as the starting point for
all geometry optimizations in this study.
We have summarized in Table 5 the AM1-calculated
heats of formation of the different complexes and transi-
tion-state models involving the N-BH3 adduct of ox-
azaborolidine 5a -Me and 1-phenylethan-1-one (8a ) and
3,3-dimethylbutan-2-one (8f) as substrate ketones. Rela-
tive energies within the different families of complexes
and transition-state models are also given, as well as the
energy barrier connecting every complex/transition state
pair.
It is interesting to note that, according to the least
motion principle, exo complexes must evolve into chair
type transition states, whereas endo complexes should
correspondingly evolve into boat type ones. For conven-
ience, we have retained for the transition states the same
nomenclature employed for the complexes. According to
it, Re transition states lead to R alcohols and Si transition
states correspondingly lead to S alcohols.
If interconversion between complexes is much faster
than reaction, the enantioselectivity of the reaction is
simply determined, according to the Curtin-Hammett
principle,16 by the energy difference between the most
favorable Re- and Si-type transition states. If this inter-
Several theoretical studies, involving different levels
of theory, have been devoted to this process on both
model15a and real systems.15 Whereas Quallich and Blake
have reported the highest level ab initio calculations for
a model system, Liotta has performed an exhaustive
(15) (a) Quallich, G. J .; Blake, J . F.; Woodall, T. M. J . Am. Chem.
Soc. 1994, 116, 8516-8525. (b) Nevalainen, V. Tetrahedron: Asym-
metry 1992, 3, 1563-1572. (c) J ones, D. K.; Liotta, D. C.; Shinkai, I.;
Mathre, D. J . J . Org. Chem. 1993, 58, 799-801. (d) Nevalainen, V.
Tetrahedron: Asymmetry 1993, 4, 1597-1602. (e) Linney, L. P.; Self,
C. R.; Williams, I. H. J . Chem. Soc., Chem. Commun. 1994, 1651-2.

7908 J . Org. Chem., Vol. 64, No. 21, 1999
Puigjaner et al.
Sch em e 7
Ta ble 6. Solven t-Assisted Dissocia tion of th e Rea ctive
Com p lexes of 5a -Me w ith Bor a n e a n d
1-P h en yleth a n -1-on e (8a ) or 3,3-Dim eth ylbu ta n -2-on e (8f)
1-phenylethan-1-one
3,3-dimethylbutan-2-one
∆H° (kcal mol-1
)
∆H° (kcal mol-1
+3.4
)
exo-Si
+2.7
+0.5
-0.5
+0.6
exo-Re
endo-Si
endo-Re
+1.9
sponding reduction step. Therefore, the direct comparison
between heats of formation of the most favorable transi-
tion states of Re- and Si-type for a given ketone repre-
sents a valid approximation to the analysis of the
enantioselectivity of the reduction.
F igu r e 2. Schematic representation of the four possible
reactive complexes and derived transition states in the oxaza-
borolidine-mediated reduction of a prochiral ketone with
borane.
Ta ble 5. AM1-Ca lcu la ted Hea ts of F or m a tion a of th e
Rea ctive Com p lexes of Oxa za bor olid in e 5a -Me w ith
Bor a n e a n d 1-P h en yleth a n -1-on e or
3,3-Dim eth ylbu ta n -2-on e a n d of th e Cor r esp on d in g
Tr a n sition Sta tes Lea d in g to Secon d a r y Alcoh ols
According to this, the enantioselectivity in the reduc-
tion of 1-phenylethan-1-one arises from the energy dif-
ference between the exo-Si and the exo-Re transition
states, the former being preferred by 1.6 kcal‚mol-1. In
the reduction of 3,3-dimethylbutan-2-one, in turn, the
very important steric bulk of the tert-butyl substituent
precludes the achievement of some of the transition
geometries (the exo-Re and the exo-Si ones, which would
present unavoidable steric interactions between the
ketone and the oxazaborolidine moieties). In this case,
enantioselectivity in the reaction seems to obey to the
energy difference between the exo-Si and the endo-Re
transition states, the former being preferred by 2.3
kcal‚mol-1. A representation of the relevant transition
states in the reduction of the ketones discussed above
can be found in Figure 3.
1-phenylethan-1-one
T.S.
3,3-dimethylbutan-2-one
T.S.
complex model barrier complex model barrier
exo-Si
-140.6 -134.8
(0.0) (0.0)
-138.4 -133.2
(2.2) (1.6)
endo-Si -137.4 -132.3
(3.2) (2.5)
endo-Re -138.4 -132.2
(2.2) (2.6)
5.8
5.2
5.1
6.2
-187.5 -181.9
(0.0) (0.0)
5.6
exo-Re
-186.0 -179.6
(1.5) (2.3)
6.4
a
All values are in kcal mol-1. Values in parentheses are relative
energies within a column.
The theoretical calculations reported here predict that
the reduction with borane of prochiral ketones mediated
by oxazaborolidine 5a -Me will preferentially lead to
alcohols of S configuration, as is experimentally ob-
served.17 Again, in agreement with experimental obser-
vation, it is predicted that the reduction of 3,3-dimethyl-
butan-2-one at room temperature will take place with
higher enantioselectivity than that of 1-phenylethan-1-
one. Moreover, the present calculations could also provide
a clue on the capricious enantioselectivity/temperature
dependence observed in the reduction of the ketones in
this paper. Thus, it is well-known that the occurrence of
a nonlinear dependence between enantioselectivity and
temperature obeys to some kind of mechanism change
in the process.18 In processes where every component of
a product stereoisomeric mixture can be formed through
two or more transition states, a simple variation with
temperature of the preference for the different reaction
modes (as a consequence of, for instance, their different
activation entropies) would be a sufficient condition for
the observation of a nonlinear behavior. The possibility
of attaining all four possible transition states in the
conversion is to take place through a dissociative process,
an estimation of the corresponding barriers could be done
through the enthalpy change of the solvent-assisted
dissociation represented in Scheme 7. We have sum-
marized in Table 6 the AM1-calculated values for the
dissociation of the different complexes of 1-phenylethan-
1-one (8a ) and 3,3-dimethylbutan-2-one (8f). Taking into
account the nature of the considered processes and the
small enthalpy changes predicted by the calculations, it
can be deduced that barriers for complex interconversion
should also be small.
Barriers for reaction of the different complexes, in turn,
are directly available from their heats of formation and
those of the corresponding transition states (5.1-6.4
kcal‚mol-1, Table 5). Interestingly, they are only slightly
higher than those found at the MP2/6-31G(d)//MP2/6-
31G(d) level of theory (3.1-4.1 kcal‚mol-1) for a reduction
of the same type.15a A comparison between the data in
Tables 5 and 6 indicates as discussed above that complex
interconversion should be much faster than the corre-
(16) (a) Carey, F. A.; Sundberg, R. J . Advanced Organic Chemistry,
3rd ed.; Plenum Press: New York, 1990; pp 215-216. (b) Eliel, E. L.;
Wilen, S. H. Stereochemistry of Organic Compounds; J ohn Wiley &
Sons: New York, 1994; pp 647-655.
(17) The priority order of the substituents in the reduction product
of 8b accounts for the R configuration of the final compound.
(18) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.; Esser, P. Angew.
Chem., Int. Ed. Engl. 1991, 30, 477-515.

Enantioselective Reduction of Prochiral Ketones
J . Org. Chem., Vol. 64, No. 21, 1999 7909
F igu r e 3. Representation of the relevant transition states (AM1) in the reduction of 1-phenylethan-1-one (8a ) and 3,3-
dimethylbutan-2-one (8f).
d’Ana`lisi Elementals del CSIC de Barcelona. High-resolution
mass spectra (CI) were measured by the Servicio de Espec-
trometr´ıa de Masas de la Universidad de Co´rdoba. Chromato-
graphic separations were carried out using NEt3-pretreated
(2.5% v/v) SiO2 (70-230 mesh). THF was freshly distilled from
benzophenone ketyl under N2. 2-Iodo-1-methoxyethane and
2-iodo-1-(2-methoxyethoxy)ethane19 were prepared as de-
scribed in the literature. (2S,3S)-2,3-Epoxy-3-phenylpropanol,
2, was prepared according to the procedure described by
Sharpless et al.20
1-[((2S,3S)-3-P h en yloxir a n -2-yl)m eth oxy]-2-m eth oxy-
eth a n e, 7b. A solution of 2 (3.00 g, 20 mmol) in DMF (24 mL)
was added via cannula to a suspension of sodium hydride (3.60
g, ca. 120 mmol) in DMF (24 mL) at 0 °C under N2. The
mixture was stirred for 20 min, and 2-iodo-1-methoxyethane
(22.30 g, 120 mmol) was added via syringe into the mixture.
The mixture was stirred for 15 h at 0 °C. MeOH (250 mL) and
brine (250 mL) were added. The aqueous solution was ex-
tracted with Et2O. The combined organic extracts were dried
and concentrated in vacuo. The residual oil was chromato-
reduction of prochiral ketones not containing a too bulky
substituent, as seen for 1-phenylethan-1-one (see also ref
15c) opens the way for such behavior.
Con clu sion s
In summary, the synthesis of chiral oxazaborolidines
from enantiomerically pure (2S,3S)-2,3-epoxy-3-phenyl-
propanol has led to the development of a new family of
catalysts for the enantioselective reduction of ketones.
Thanks to the modular design of the target molecules,
some of their structural elements have been optimized
with great ease for improved enantioselectivity. As a
result, (4R,5R)-2-methyl-4-phenyl-5-methoxymethyl-1,3,2-
oxazaborolidine (5a -Me) has been developed as a new
mediator for the enantioselective reduction of prochiral
ketones with interesting characteristics: It can be used
in most cases at room temperature and the enantiomeric
excess of the resulting alcohols approaches that obtained
with the CBS reagent (2% difference in mean ee over a
family of 8 ketones). As indicated by semiempirical AM1
calculations the observed enantioselectivity arises, within
the framework of the commonly accepted mechanism for
these reactions, from the preference for a chairlike
transition state with the hydride being transferred from
boron to the Re face of the reacting ketone.
graphed using hexane/EtOAc (95:5) as eluent to give 3.27 g
1
(79%) of 7b as an oil: [R]23 ) -36.1 (c ) 1.1 in CHCl3); H
D
NMR δ 7.19-7.38 (m, 5H), 3.89 (dd, J ) 11.7, 2.9 Hz, 1H),
3.79 (d, J ) 2.2 Hz, 1H), 3.55-3.75 (m, 5H), 3.40 (s, 3H), 3.24
(ddd, J ) 5.3, 3.1, 2.2 Hz, 1H); 13C NMR δ 136.7 (C), 128.3
(CH), 128.1 (CH), 125.6 (CH), 71.8 (CH2), 71.0 (CH2), 70.7
(CH2), 61.0 (CH), 58.9 (CH3), 55.7 (CH); IR (film) 2881, 1200,
1109 cm-1; MS (CI, CH4) m/z 209 (C12H16O3‚H+, 100); HRMS
(CI, CH4) for C12H16O3 (M+) 208.1099, found 208.1088.
1-[((2S,3S)-3-P h en yloxir a n -2-yl)m et h oxy]-2-(2-m et h -
oxyeth oxy)eth a n e, 7c. Compound 2 (0.52 g, 3.4 mmol) in
Exp er im en ta l Section
Gen er a l Met h od s. 1H NMR and 13C NMR spectra in
solution were recorded in CDCl3 at 200 and 50 MHz, respec-
tively, unless otherwise cited. 1H chemical shifts are quoted
relative to TMS and 13C chemical shifts relative to solvent
signals. Elemental analyses were carried out by the Servei
(19) Prugh, J . D.; Hartman, G. D.; Mallorga, P. J .; McKeever, B.
M.; Michelson, S. R.; Murcko, M. A.; Schwam, H.; Smith, R. L.; Sondey,
J . M.; Springer, J . P.; Sugrue, M. F. J . Med. Chem. 1991, 34, 1805-
1818.
(20) Gao, Y.; Hanson, R. M.; Klunder, J . M.; Ko, S. Y.; Masamune,
H.; Sharpless, K. B. J . Am. Chem. Soc. 1987, 109, 5765-5780.

7910 J . Org. Chem., Vol. 64, No. 21, 1999
Puigjaner et al.
DMF (7 mL), sodium hydride (1.19 g, ca. 40 mmol) in DMF (9
mL), and 2-iodo-1-(2-methoxyethoxy)ethane (7.93 g, 34 mmol)
were treated as described for 7b with stirring 24 h at 0 °C to
give 820 mg (95%) of 7c as an oil after chromatography using
hexane/EtOAc (100:0/90:10): [R]23D ) -30.5 (c ) 1.0 in CHCl3);
1H NMR (500 MHz) δ 7.27-7.32 (m, 5H), 3.87 (dd, J ) 11.6,
3.5 Hz, 1H), 3.78 (d, J ) 2.0 Hz, 1H), 3.63 (dd, J ) 11.6, 5.0
Hz, 1H), 3.54-3.77 (m, 8H), 3.37 (s, 3H), 3.20 (ddd, J ) 5.0,
3.5, 2.0 Hz, 1H); 13C NMR δ 136.8 (C), 128.4 (CH), 128.1 (CH),
125.6 (CH), 71.8 (CH2), 70.9 (CH2), 70.6 (CH2), 70.52 (CH2),
70.47 (CH2), 61.0 (CH), 58.9 (CH3), 55.7 (CH); IR (film) 2877,
1104 cm-1; MS(CI, NH3) m/z 270 (C14H20O4‚NH4+, 100) and
253 (C14H20O4‚H+, 3); HRMS(CI, CH4) for C14H21O4 (M + H+)
253.1440, found 253.1417.
Gen er a l P r oced u r e for th e Regioselective Oxir a n e
Rin g Op en in g F ollow ed by Hyd r ogen a tion of th e Azid o
Gr ou p : (1R,2R)-1-Am in o-3-m eth oxy-1-p h en ylp r op a n -2-
ol, 4a . A solution of 7a (4.90 g, 30 mmol), LiClO4 (78.40 g,
737 mmol), and NaN3 (9.70 g, 149 mmol) in acetonitrile (147
mL) was stirred at 55 °C for 24 h under N2. H2O (1.8 L) was
added, and the aqueous layer was extracted with Et2O. The
combined organic extracts were dried and concentrated in
vacuo to give 6.17 g of 6a as an oil that was used in the next
step without further purification. A solution of 6a (6.17 g, 29.8
mmol) in MeOH (75 mL) was added via syringe to a suspension
of Pd/C 10% (0.62 g) in MeOH (108 mL) at room temperature
under H2. After 36 h of stirring at this temperature, the
mixture was filtered through Celite. The residual oil was
chromatographed through a short SiO2 column using hexane/
(CH), 73.5 (CH), 72.1 (CH2), 71.8 (CH2), 70.5 (CH2), 70.4 (CH2),
59.0 (CH3), 57.8 (CH); IR (film) 3367, 1109 cm-1; MS (CI, NH3)
m/z 270 (C14H23NO4‚H+, 100); HRMS (CI) calcd for C14H23NO4‚
H+ 270.1705, found 270.1710.
Gen er a l P r oced u r e for th e Regioselective Oxir a n e
Rin g Op en in g F ollow ed by Red u ction of th e Azid o
Gr ou p w ith Na BH4: (1R,2R)-1-Am in o-1-p h en yl-3-(p h en -
ylm eth oxy)p r op a n -2-ol, 4d . Compound 7d (3.20 g, 13 mmol),
LiClO4 (34.90 g, 328 mmol), and NaN3 (4.30 g, 66 mmol) in
acetonitrile (66 mL) were treated as described for 7a during
24 h. The workup was identical to the one described for 7a to
give 3.76 g of 6d as an oil that was used in the next step
without further purification. A solution of 6d (3.76 g, 13 mmol)
and sodium borohydride (1.60 g, 42 mmol) in THF (30.4 mL)
was heated at 55-60 °C under N2. MeOH (6.6 mL) was added
during 1 h. After the mixture was heated at this temperature
for 16 h, H2O (560 mL) was added, and the aqueous layer was
extracted with CH2Cl2. The combined organic extracts were
dried and concentrated in vacuo. The residual oil was chro-
matographed through a short SiO2 column using hexane/
EtOAc (80:20-60:40) as eluent to give 3.14 g (92%) of 4d as a
colorless oil: [R]23 ) -21.8 (c ) 0.9 in CHCl3); 1H NMR δ
D
7.38-7.25 (m, 10H), 4.48 (s, 2H), 4.16 (d, J ) 5.2 Hz, 1H),
3.97 (ddd, J ) 5.2, 5.2, 5.2 Hz, 1H), 3.44-3.46 (m, 2H), 2.07
(br s, 3H, NH2 + OH);13C NMR δ 141.8 (C), 137.8 (C), 128.3
(CH), 127.9 (CH), 127.7 (CH), 127.6 (CH), 127.2 (CH), 127.1
(CH), 73.3 (CH), 73.3 (CH2), 71.4 (CH2), 57.7 (CH); IR (film)
3363, 1102, 1073 cm-1; MS (CI, NH3) m/z 275 (C16H19NO2‚
NH4+, 4) and 258 (C16H19NO2‚H+, 100); HRMS (CI) calcd for
EtOAc (50:50) as eluent to give 4.9 g (91%) of 4a as an oil:
C
16H19NO2‚H+ 258.1494, found 258.1504.
(1R,2R)-1-Am in o-1-ph en yl-3-(diph en ylm eth oxy)pr opan -
1
[R]23 ) -31.2 (c ) 1.0 in CHCl3); H NMR δ 7.37-7.26 (m,
D
5H), 4.14 (d, 1H, J ) 5.2 Hz), 3.93 (ddd, J ) 5.2, 5.2, 5.2 Hz,
2-ol, 4e. Compound 7e (0.20 g, 0.6 mmol), LiClO4 (1.65 g, 15
mmol), and NaN3 (0.20 g, 3 mmol) in acetonitrile (3.1 mL) were
treated as described for 7a during 24 h. The workup was
identical to the one described for 7a to give 0.23 g of 6e as an
oil that was used in the next step without further purification.
A solution of 6e (0.23 g, 0.6 mmol) and sodium borohydride
(77 mg, 2 mmol) in THF (1.5 mL) was heated at 55-60 °C
under N2. MeOH (0.32 mL) was added during 1 h, and the
mixture was heated at this temperature for 7 h. A workup
identical to the one described for 4d followed by chromatog-
raphy through a short SiO2 column using hexane/EtOAc (60:
40) as eluent yielded 0.15 g (73%) of 4e as a white solid that
1H), 3.32-3.34 (m, 2H), 3.33 (s, 3H), 2.09 (br s, 3H, NH2
+
OH);13C NMR δ 142.0 (C), 128.4 (CH), 127.4 (CH), 127.1 (CH),
73.6 (CH2), 73.2 (CH), 59.1 (CH3), 57.8 (CH); IR (film) 3361,
3298 cm-1; MS (CI, NH3) m/z 199 (C10H15NO2‚NH4+, 20) and
182 (C10H15NO2‚H+, 100). Anal. Calcd for C10H15NO2: C, 66.27;
H, 8.34; N, 7.73. Found: C, 65.93; H, 8.19; N, 7.37.
(1R,2R)-1-Am in o-3-(2-m eth oxyeth oxy)-1-ph en ylpr opan -
2-ol, 4b. Compound 7b (1.10 g, 5 mmol), LiClO4 (13.90 g, 130
mmol), and NaN3 (1.76 g, 27 mmol) in acetonitrile (27.1 mL)
were treated as described for 7a during 24 h. The workup was
identical to the one described for 7a to give 1.30 g of 6b as an
oil that was used in the next step without further purification.
A solution of 6b (1.30 g, 5 mmol) in MeOH (13 mL) was added
via syringe to a suspension of Pd/C 10% (0.13 g) in MeOH (18.7
mL) at room temperature under H2. After 15 h of stirring at
this temperature, the mixture was filtered through Celite. The
residual oil was chromatographed through a short SiO2 column
using hexane/EtOAc (30:70) as eluent to give 0.98 g (84%) of
was recrystallized from MeOH to obtain an analytically pure
1
sample: mp 125-7 °C; [R]23 ) -21.5 (c ) 1.0 in CHCl3); H
D
NMR (300 MHz) δ 7.31-7.19 (m, 15H), 5.27 (s, 1H), 4.13 (d, J
) 4.8 Hz, 1H), 4.00 (ddd, J ) 4.8, 4.8, 4.8 Hz, 1H), 3.39-3.41
(m, 2H), 2.57 (br s, 3H, NH2 + OH); 13C NMR (75 MHz) δ 141.9
(C), 128.5 (CH), 128.4 (CH), 127.6 (CH), 127.5 (CH), 127.4
(CH), 127.2 (CH), 127.1 (CH), 127.0 (CH), 126.9 (CH), 84.3
(CH), 73.6 (CH), 70.3 (CH2), 58.0 (CH); IR (film) 3357, 1075
cm-1; MS (CI, NH3) m/z 351 (C22H23NO2‚NH4+, 5) and 334
(C22H23NO2‚H+, 100). Anal. Calcd for C22H23NO2: C, 79.25; H,
6.95; N, 4.20. Found: C, 79.10; H, 7.02; N, 4.16.
4b as an oil: [R]23 ) -31.7 (c ) 1.3 in CHCl3); 1H NMR δ
D
7.36-7.27 (m, 5H), 4.13 (d, J ) 5.0 Hz, 1H), 3.97 (ddd, J )
5.2, 5.2, 5.2 Hz, 1H), 3.60-3.48 (m, 4H), 3.43-3.40 (m, 2H),
3.36 (s, 3H), 2.33 (br s, 3H, NH2 + OH);13C NMR δ 142.1 (C),
128.3 (CH), 127.2 (CH), 127.1 (CH), 73.5 (CH), 72.2 (CH2), 71.7
(CH2), 70.4 (CH2), 58.9 (CH3), 57.8 (CH); IR (film) 3365, 3298,
1104 cm-1; MS (CI, CH4) m/z 226 (C12H19NO3‚H+, 100); HRMS
(CI) calcd for C12H19NO3‚H+ 226.1443, found 226.1434.
(1R,2R)-1-Am in o-3-[2-(2-m eth oxyeth oxy)eth oxy]-1-ph en -
ylp r op a n -2-ol, 4c. Compound 7c (1.00 g, 4 mmol), LiClO4
(10.40 g, 98 mmol), and NaN3 (1.30 g, 20 mmol) in acetonitrile
(19.8 mL) were treated as described for 7a during 24 h. The
workup was identical to the one described for 7a to give 1.17
g of 6c as an oil that was used in the next step without further
purification. A solution of 6c (1.17 g, 4 mmol) in MeOH (10
mL) was added via syringe to a suspension of Pd/C 10% (0.12
g) in MeOH (20 mL) at room temperature under H2. After 36
h of stirring at this temperature, the mixture was filtered
through Celite. The residual oil was chromatographed through
a short SiO2 column using hexane/EtOAc (30:70) as eluent to
(1R,2R)-1-Am in o-1-ph en yl-3-(tr iph en ylm eth oxy)pr opan -
2-ol, 4f. Compound 7f (2.30 g, 6 mmol), LiClO4 (15.40 g, 145
mmol), and NaN3 (1.90 g, 29 mmol) in acetonitrile (29.3 mL)
were treated as described for 7a during 24 h. The workup was
identical to the one described for 7a to give 2.60 g of 6f as an
oil that was used in the next step without further purification.
A solution of 6f (2.60 g, 6 mmol) and sodium borohydride (0.72
g, 19 mmol) in THF (13.7 mL) was heated at 55-60 °C under
N2. MeOH (3 mL) was added during 1 h, and the mixture was
heated at this temperature for 22 h. A workup identical to the
one described for 4d followed by chromatography through a
short SiO2 column using hexane/EtOAc (60:40) as the eluent
yielded 2.13 g (87%) of 4f as a white solid that was recrystal-
lized from Et2O to obtain an analytically pure sample: mp 57-
1
58 °C; [R]23 ) -22.6 (c ) 1.0 in CHCl3); H NMR (300 MHz)
D
δ 7.43-7.18 (m, 20H), 4.17 (d, J ) 5.1 Hz, 1H), 3.91 (ddd, J )
5.1, 5.1, 5.1 Hz, 1H), 3.15-3.06 (m, 2H), 2.24 (br s, 3H, NH2 +
OH);13C NMR (50 MHz) δ 143.7 (C), 141.8 (C), 128.6 (CH),
128.3 (CH), 127.8 (CH), 127.6 (CH), 127.2 (CH), 127.0 (CH),
86.8 (C), 73.9 (CH), 64.4 (CH2), 58.1 (CH); IR (film) 3363, 1067
give 0.86 g (81%) of 4c as an oil: [R]23 ) -28.5 (c ) 1.0 in
D
CHCl3); 1H NMR δ 7.36-7.23 (m, 5H), 4.15 (d, J ) 5.2 Hz,
1H), 3.97 (ddd, J ) 5.2, 5.2, 5.2 Hz, 1H), 3.64-3.53 (m, 8H),
3.43-3.41 (m, 2H), 3.37 (s, 3H), 2.61 (br s, 3H, NH2 + OH);13C
NMR δ 142.0 (C), 128.3 (CH), 127.9 (CH), 127.3 (CH), 127.1
cm-1; MS (CI) m/z 427 (C28H27NO2‚NH4+, 2) and 410 (C28H27
-

Enantioselective Reduction of Prochiral Ketones
J . Org. Chem., Vol. 64, No. 21, 1999 7911
NO2‚H+, 50). Anal. Calcd for C28H27NO2: C, 82.12; H, 6.65; N,
3.42. Found: C, 81.75; H, 6.79; N, 3.29.
â-DEX 120, 112 °C, tR R isomer 47.4 min, tR S isomer 49.1
min. For 1,2,3,4-tetrahydronaphthol: â-DEX 120, 125 °C, tR
S isomer 89.2 min, tR R isomer 93.3 min. For 3,3-dimethyl-
Gen er a l P r oced u r e for th e Red u ction of P r och ir a l
Keton es 8a ,b Usin g (4R,5R)-4-P h en yl-5-(R1-oxym eth yl)-
1,3,2-oxa za bor olid in es (5a -H, 5b-H, 5c-H, 5d -H, 5e-H, a n d
5f-H) a s Ch ir a l Ca ta lysts. Borane-methyl sulfide complex
(66 µL of a 10 M BH3 solution in Me2S, 0.7 mmol) was added
via syringe to a solution of the amino alcohol (0.055 mmol, 10
mol %) in THF (1.5 mL) at room temperature under N2. The
mixture was stirred at this temperature for 16 h, and a
solution of the ketone (0.55 mmol) in THF (0.7 mL) was added
dropwise with the aid of a syringe pump over a 1 h period.
The mixture was further stirred for 10 min under N2. The
reaction was quenched by the addition of a 2 M HCl solution
(2 mL) and the mixture extracted with Et2O. The combined
organic extracts were dried and concentrated in vacuo. The
enantiomeric excesses were determined from the crude mix-
ture by GC analyses. For the conditions of the GC analyses
and the determination of the absolute configuration of the final
compounds, see below.
butan-2-ol: â-DEX 120, 70 °C, tR R isomer 14.7 min, tR
S
isomer 15.0 min. For 1-cyclohexylethan-1-ol: R-DEX 120, 70
°C, tR R isomer 67.6 min, tR S isomer 69.5 min.
To stablish the absolute configuration of the final com-
pounds, the alcohols were purified by bulb-to-bulb distillation
of the crude mixtures. The optical rotation was measured in
each case and its sign was compared with the reported value
((1S)-1-phenylethan-1-ol,21a (1R)-2-chloro-1-phenylethan-1-
ol,21b (1S)-1-(3-methoxyphenyl)ethan-1-ol,21c (1S)-1-(4-chloro-
phenyl)ethan-1-ol,21d (1S)-1-(2-methoxyphenyl)ethan-1-ol,21e (1S)-
1-phenylpropan-1-ol,21f (1S)-1,2,3,4-tetrahydronaphthol,21a (1S)-
1-cyclohexylethan-1-ol,21g and 3,3-dimethylbutan-2-ol21g).
Gen er a l P r oced u r e for th e Red u ction of P r och ir a l
Keton es Usin g (4R,5R)-2-Bu tyl-4-p h en yl(1,3,2-oxa za bor o-
lid in -5-yl)m eth oxym eth a n e (5a -Bu ) a s th e Ch ir a l Ca ta -
lyst. (a ) P r ep a r a tion of th e B-Bu Oxa za bor olid in e 5a -
Bu . Butylboronic acid (62 mg, 0.605 mmol) was added to a
solution of the amino alcohol 4a (0.1 g, 0.55 mmol) in toluene
(7 mL) in a round-bottom flask fitted with a Dean-Stark
apparatus and a reflux condenser. The reaction mixture was
stirred under N2 at room temperature for 0.5 h and then
refluxed for 16 h. The solvent was distilled off until ca. 1 mL
was left. Toluene (0.1 mL) was then added to provide a 0.5 M
solution of catalyst. (b) Ca ta lytic En a n tioselective Red u c-
tion of P r och ir a l Keton es 8a ,b. The chiral catalyst 5a -Bu
(78 µL of a 0.5 M solution in toluene, 0.039 mmol, 10%)
together with 1.1 mL of THF were introduced via syringe into
a flame-dried flask under N2. Borane-methyl sulfide complex
(47 µL of a 10 M BH3 solution in Me2S, 0.47 mmol) was added
via syringe to the previous solution. A solution of the ketone
(0.39 mmol) in THF (0.5 mL) was added dropwise with the
aid of a syringe pump over a 1 h period. The mixture was
further stirred for 10 min under N2. The reaction was
quenched by the addition of a 2 N HCl solution (2 mL) and
the mixture extracted with Et2O (3 × 2 mL). The combined
organic extracts were dried and concentrated in vacuo. The
enantiomeric excesses were determined from the crude mix-
ture by GC analyses. For the conditions of the GC analyses
and the determination of the absolute configuration of the final
compounds see above.
Gen er a l P r oced u r e for th e Red u ction of P r och ir a l
Ket on es Usin g (4R,5R)-2-Met h yl-4-p h en yl-5-(R1-oxy-
m eth yl)-1,3,2-oxa za bor olid in es (5a -Me, 5b-Me, 5c-Me, 5d -
Me, 5e-Me, a n d 5f-Me) a s Ch ir a l Ca ta lysts. (a ) P r ep a r a -
tion of th e B-Me Oxa za bor olid in es. Trimethylboroxine
(0.21 mL, 1.5 mmol) was added via syringe to a solution of
the amino alcohol 4a -f (2.2 mmol) in toluene (14.4 mL), and
the mixture was stirred under N2 at room temperature for 1
h. Toluene, excess trimethylboroxine, and water were distilled
off at atmospheric pressure until ca. 4 mL remained. The
mixture was diluted with toluene (3 × 7 mL), each time
distilling until ca. 4 mL was left. Toluene (0.4 mL) was then
added to provide a 0.5 M solution of the B-methyl oxazaboro-
lidines in toluene. (b) Ca ta lytic En a n tioselective Red u c-
tion of P r och ir a l Keton es 8a -j. The corresponding B-Me
oxazaborolidine 5-Me (78 µL of a 0.5 M solution in toluene,
0.039 mmol, 10%) together with 1.1 mL of THF was introduced
via syringe into a flame-dried flask under N2. Borane-methyl
sulfide complex (47 µL of a 10 M BH3 solution in Me2S, 0.47
mmol) was added via syringe to the previous solution. The
mixture was heated or cooled to the desired temperature if
necessary. A solution of the ketone (0.39 mmol) in THF (0.5
mL) was added dropwise with the aid of a syringe pump over
a 1 h period. The mixture was further stirred for 10 min at
the appropriate temperature under N2. The reaction was
quenched by the addition of a 2 M HCl solution (2 mL) and
the mixture extracted with Et2O. The combined organic
extracts were dried and concentrated in vacuo. The enantio-
meric excesses were determined from the crude mixture by
GC analyses. Conditions of GC analyses: â-DEX 120 or R-DEX
120 column, 30 m length, 0.25 mm internal diameter, isotherm
temperature program, He as carrier gas (2.4 mL/min). For
1-phenylethan-1-ol: â-DEX 120, 100 °C, tR R isomer 52.1 min,
Ack n ow led gm en t. Financial support from DGICYT
(PB95-0265) and CIRIT (1998SGR-00005) is gratefully
acknowledged. C.P. thanks CIRIT for a fellowship, and
A.V.-F. thanks MEC for a contract under the program
Acciones para la Incorporacio´n a Espan˜a de Doctores y
Tecno´logos.
J O990942Q
t
R S isomer 55.7 min. For 2-chloro-1-phenylethan-1-ol: â-DEX
120, 125 °C, tR S isomer 70.1 min, tR R isomer 73.8 min. For
1-(3-methoxyphenyl)ethan-1-ol: â-DEX 120, 122 °C, tR
isomer 78.4 min, tR S isomer 82.3 min. For 1-(3-chlorophenyl)-
ethan-1-ol: â-DEX 120, 125 °C, tR R isomer 57.7 min, tR
(21) (a) Kim, Y. H.; Park, D. H.; Byun, I. S.; Yoon, I. K.; Park, C. S.
J . Org. Chem. 1993, 58, 4511-4512. (b) Hartgerink, J . W.; van der
Laan, L. C. J .; Engberts, J . B. F. N.; Boer, T. H. J . Tetrahedron 1971,
27, 4323-4334. (c) Chen, C. P.; Prasad, K.; Repic, O. Tetrahedron Lett.
1991, 32, 7175-7178. (d) Polavarapu, P. L.; Fontana, L. P.; Smith, H.
E. J . Am. Chem. Soc. 1986, 108, 94-99. (e) Meier, C.; Laux, W. H. G.
Tetrahedron 1996, 52, 589-598. (f) Kitamura, M.; Suga, S.; Kawai,
K.; Noyori, R. J . Am. Chem. Soc. 1986, 108, 6071-6072. (g) Brown, H.
C.; Cho, B. T.; Park, W. S. J . Org. Chem. 1988, 53, 1231-1238.
R
S
isomer 61.1 min. For 1-(4-chlorophenyl)ethan-1-ol: â-DEX 120,
125 °C, tR R isomer 60.6 min, tR S isomer 65.8 min. For 1-(2-
methoxyphenyl)ethan-1-ol: â-DEX 120, 122 °C, tR S isomer
58.6 min, tR R isomer 64.0 min. For 1-phenylpropan-1-ol: