Mild and expedient asymmetric reductions of α,β-unsaturated alkenyl and alkynyl ketones by TarB-NO2 and mechanistic investigations of ketone reduction

Article abstract of DOI:10.1021/jo101530f

A facile and mild reduction procedure is reported for the preparation of chiral allylic and propargyl alcohols in high enantiomeric purity. Under optimized conditions, alkynyl and alkenyl ketones were reduced by TarB-NO ₂ and NaBH₄ at 25 °C in 1 h to produce chiral propargyl and allylic alcohols with enantiomeric excesses and yields up to 99%. In the case of α,β-unsaturated alkenyl ketones, α-substituted cycloalkenones were reduced with up to 99% ee, while more substituted and acyclic derivatives exhibited lower induction. For α,β-ynones, it was found that highly branched aliphatic ynones were reduced with optimal induction up to 90% ee, while reduction of aromatic and linear aliphatic derivatives resulted in more modest enantioselectivity. Using the (l)-TarB-NO₂ reagent derived from (l)-tartaric acid, we routinely obtained highly enantioenriched chiral allylic and propargyl alcohols with (R) configuration. Since previous models and a reduction of a saturated analogue predicted propargyl products of (S) configuration, a series of new mechanistic studies were conducted to determine the likely orientation of aromatic, alkenyl, and alkynyl ketones in the transition state.

Full text of DOI:10.1021/jo101530f

pubs.acs.org/joc
Mild and Expedient Asymmetric Reductions of r,β-Unsaturated Alkenyl
and Alkynyl Ketones by TarB-NO and Mechanistic Investigations of
Ketone Reduction
2
Scott Eagon, Cassandra DeLieto, William J. McDonald, Dustin Haddenham,
Jaime Saavedra, Jinsoo Kim, and Bakthan Singaram*
Department of Chemistry and Biochemistry, University of California at Santa Cruz, Santa Cruz,
California 94086, United States
singaram@chemistry.ucsc.edu
Received August 17, 2010
A facile and mild reduction procedure is reported for the preparation of chiral allylic and propargyl
alcohols in high enantiomeric purity. Under optimized conditions, alkynyl and alkenyl ketones were
reduced by TarB-NO and NaBH at 25 °C in 1 h to produce chiral propargyl and allylic alcohols
2
4
with enantiomeric excesses and yields up to 99%. In the case of R,β-unsaturated alkenyl ketones,
R-substituted cycloalkenones were reduced with up to 99% ee, while more substituted and acyclic
derivatives exhibited lower induction. For R,β-ynones, it was found that highly branched aliphatic
ynones were reduced with optimal induction up to 90% ee, while reduction of aromatic and linear
aliphatic derivatives resulted in more modest enantioselectivity. Using the (L)-TarB-NO reagent
2
derived from (L)-tartaric acid, we routinely obtained highly enantioenrichedchiral allylic and propargyl
alcohols with (R) configuration. Since previous models and a reduction of a saturated analogue
predicted propargyl products of (S) configuration, a series of new mechanistic studies were conducted
to determine the likely orientation of aromatic, alkenyl, and alkynyl ketones in the transition state.
Introduction
reduction is not enantioselective, it has been used as a diaster-
eoselective reduction agent for several recent compounds with
The asymmetric reduction of prochiral alkenyl and alkynyl
ketones is one of the most expedient methods to synthesize
chiral allylic and propargyl alcohols. These chiral alcohols
are present in a myriad of bioactive compounds and can be
converted into a variety of synthetic intermediates due to
the unique reactivities of the carbon-carbon double and triple
bond. A handful of methods for the regiospecific 1,2-reduction
of R,β-unsaturated alkenyl ketones have been reported, with
one of the most popular being the cerium trichloride/sodium
2
modest success. Modern enantioselective 1,2-reductions of
R,β-unsaturated alkenyl ketones have been reported using
several methods including ruthenium catalysts, enzymes,
3
4
5
and the boron-containing DIP-Cl and CBS reagents. In the
6
(2) (a) Pragani, R.; Roush, W. Org. Lett. 2008, 70, 4613. (b) Gollner, A.;
Mulzer, J. Org. Lett. 2008, 70, 4701. (c) Westermann, J.; Schneider, M.;
Platzek, J.; Petrov, O. Org. Process Res. Dev. 2007, 11, 200. (d) Tsegay, S.;
Huegal, H.; Rizzacasa, M. Aust. J. Chem. 2009, 62, 676. (e) Taber, D.; Gu, P.;
Li, R. J. Org. Chem. 2009, 74, 5516.
1
borohydride mediated Luche reduction. Though the Luche
(3) (a) Van Innis, L.; Plancher, J.; Marko, I. Org. Lett. 2006, 8, 6111.
b) Hannedouche, J.; Kenny, J.; Walsgrove, T.; Wills, M. Synlett 2002, 263.
c) Peach, P.; Cross, D.; Kenny, J.; Mann, I.; Houson, I.; Campbell, L;
(
(
(
1) Luche, J. J. Am. Chem. Soc. 1978, 100, 2226.
Walsgrove, T.; Wills, M. Tetrahedron 2006, 62, 1864.
DOI: 10.1021/jo101530f
r 2010 American Chemical Society
Published on Web 10/29/2010
J. Org. Chem. 2010, 75, 7717–7725 7717

Eagon et al.
JOCArticle
case of R,β-ynones, many of the early reagents used to reduce
these substrates to highly enantioenriched propargyl alcohols
were boron-based compounds, such as M. Midland’s highly
SCHEME 1. Reduction of 2-Cyclohexene-1-one with TarB-
NO and NaBH
2
4
7
enantioselective reagent Alpine-Borane (B-isopinocampheyl-
9-borabicyclo[3.3.1]nonane), which continues to find modern
8
synthetic applications. Another early boron-based reagent
reported to reduce ynones was the saccharide derived
K-Glucoride reagent (potassium 9-O-(1,2:5,6-di-O-isopropy-
lidinene-5-deoxy-R-D-glucofuranosyl)-9-boratabicyclo[3.3.1]-
9
allylic alcohols have recently been prepared with enantiose-
lective reduction agents from R,β-unsaturated ketones in a
variety of recent syntheses including Mycorrhizin A, cyclo-
nonane). More recent asymmetric reduction reagents used
1
with these substrates include ruthenium catalysts, gallium
0
1
catalysts, and enzymes.
1
12
15
16
17
hexenyl nucleosides, Symbiodinolide, Brasilinolides, and
18
Though there are many procedures for the reduction of
R,β-unsaturated ketones to optically active allylic and pro-
pargyl alcohols, each has limitations. Heavy transition metals
are often acutely toxic and their removal, especially in the
preparation of pharmaceutical reagents, is not always trivial.
Enzymatic reductions often require a water-soluble substrate
and/or large excesses of microbial enzymes. Pinene-based
boron reduction reagents such as Alpine-Borane and DIP-Cl
also often require superstoichiometric quantities and extended
reaction times to achieve maximal enantioselectivity. The
advent of the CBS reagent provided a more convenient
boron-mediated reduction procedure, but it often requires
low temperatures (-78 °C), an excess of reagent, and/or
modification of the borane source to achieve the best enanti-
oselectivity and to avoid hydroboration of the alkene or
alkyne. A Ru-BINAP transfer hydrogenation procedure
has been reported by Noyori that can reduce unsaturated
ketones without reduction or migration of the olefinic bond
with good enantioselectivities for some cyclic and acyclic
alkenyl ketones, but these catalysts were unable to reduce
1
9
(þ)-Spongistatin 1. Chiral propargyl alcohols have also been
used as intermediates in several recent syntheses, including
(
2
0
21
22
þ)-Brefeldin, Spicigerolide, Balfilomycin A1, Leiocarpin
23
C, and (þ)-Goniodiol. Some bioactive molecules also contain
chiral propargyl alcohols, such as the potent prostacyclin
analogue Cicaprost and the antibacterial Uncialamycin.
2
4
The great utility of these reagents necessitates the search for
improved methods for the enantioselective reduction of pro-
chiral R,β-unsaturated alkenyl and alkynyl ketones.
Results and Discussion
Reduction of r,β-Alkenyl Ketones. We first sought to
investigate the reduction of cyclic R,β-unsaturated alkenyl
ketones with TarB-NO to see if we could achieve both good
enantioselectivity and regioselectivity for 1,2-reduction to pro-
2
25
duce the chiral allylic alcohol. When we reduced 2-cyclohexen-
1-one with TarB-NO and NaBH , we observed mainly the
2 4
1,4-reduction product cyclohexanone with the slightly enantioen-
riched 1,2-reduction product 2-cyclohexen-1-ol (Scheme 1).
We screened a variety of hydrides and discovered that by
replacing NaBH with NaBH(OAc) we were able to isolate
1
3
R,β-ynones. Noyori, however, developed an alternative
specialized family of ruthenium transfer hydrogenation cat-
alysts that were able to reduce ynones to the corresponding
4
3
only the desired 1,2-reduction product. However, we were
never able to achieve enantioselectivities above 33%. Pre-
vious reduction studies with TarB-NO indicated that a large
1
4
propargyl alcohols with excellent asymmetric induction.
Despite these preparative challenges, chiral allylic and
propargylic alcohols remain extremely useful intermediates
and are widespread throughout the synthetic literature. Chiral
2
steric difference between the substituents flanking the car-
26
bonyl was required to achieve maximal induction. To this
end, we sought more functionalized substrates which could
enhance the enantioselectivity of the reduction. Initially we
(
4) (a) Musa, M.; Ziegelmann-Fjeld, K.; Vieille, C.; Zeikus, J.; Phillips, R.
J. Org. Chem. 2007, 72, 30. (b) Polland, D; Telari, K.; Lane, J.; Humphrey,
G.; McWilliams, C.; Nidositko, S.; Salmon, P.; Moore, J. Biotechnol. Bioeng.
2
2
006, 93, 674. (c) Zagozda, M.; Plenkiewicz, J. Tetrahedron: Asymmetry
006, 17, 1958.
(15) Yu, B.; Jiang, T.; Quan, W.; Li, J.; Pan., X.; She, X. Org. Lett. 2009,
11, 629.
(
(
5) Lovchik, M.; Goeke, A.; Frater, G. J. Org. Chem. 2007, 72, 2427.
6) (a) Corey, E. J.; Bakshi, R.; Shibata, S.; Chen, C.; Singh, V. J. Am.
(16) Ferrer, E.; Alibes, R.; Busque, F.; Figueredo, M.; Font, J.; de March,
P. J. Org. Chem. 2009, 74, 2425.
Chem. Soc. 1987, 109, 7925. (b) Cho, B. T. Tetrahedron 2006, 62, 7621.
c) Holub, N.; Neidhofer, J.; Blechert, S. Org. Lett. 2007, 9, 513. (d) Kita, Y.;
Futamura, J.; Ohba, Y.; Sawama, Y.; Ganesh, J.; Fujioka, H. J. Org. Chem.
003, 68, 5917.
7) Midland, M.; Tramontano, A.; Kazubski, A.; Graham, R.; Tsai, D.;
Cardin, D. Tetrahedron 1984, 40, 1371.
8) (a) Gopalsamuthriam, V.; Predeus, A.; Huang, R.; Wulff, R. J. Am.
Chem. Soc. 2009, 131, 18018. (b) Desrat, S.; Weghe, P. J. Org. Chem. 2009,
4, 6728.
(17) Takamura, H.; Kadonaga, H.; Yamano, Y.; Han, C.; Kadota, I.;
Uemura, D. Tetrahedron 2009, 65, 7449.
(18) Paterson, I.; Burton, P.; Cordier, C.; Housden, M.; Muhlthan, F.;
Loiseleur, O. Org. Lett. 2009, 11, 693.
(19) Smith, A.; Sfouggatakis, C.; Risatti, C.; Sperry, J.; Zhu, W.; Doughty,
V.; Tomoika, T.; Gotchev, D.; Bennett, C.; Sakamoto, S.; Atasoylu, O.; Shohei,
S.; Bauer, D.; Takeuchi, M.; Makoto, J.; Sakamoto, Y. Tetrahedron 2009, 65,
6489.
(20) Kim, M.; Kim, H.; Tae, J. Synlett 2009, 8, 1303.
(21) Georges, Y.; Ariza, X.; Garcia, J. J. Org. Chem. 2009, 74, 2008.
(22) Kleinbeck, F.; Carriera, E. Angew. Chem., Int. Ed. 2009, 48, 578.
(23) Yadav, J.; Premalatha, K.; Harshavardhan, S.; Subba Reddy, B.
Tetrahedron Lett. 2008, 49, 6765.
(24) (a) Gais, H.; Kramp, G.; Wolters, D.; Reddy, L. Chem.;Eur. J. 2006,
12, 5610. (b) Skuballa, W.; Schillinger, E.; Sturzebecher, C.; Vorbruggen, H.
J. Med. Chem. 1986, 29, 313. (c) Davies, J.; Wang, H.; Taylor, T.; Warabi, K.;
Huang, X. Org. Lett. 2005, 7, 5233.
(25) Kim, J.; Brunigh, J.; Park, K.; Lee, D.; Singaram, B. Org. Lett. 2009,
11, 4358.
(26) (a) Eagon, S. C.; Kim, J.; Yan, K.; Haddenham, D.; Singaram, B.
Tetrahedron Lett. 2007, 48, 9025. (b) Eagon, S. C.; Kim, J.; Singaram, B.
Synthesis 2008, 23, 3874.
(
2
(
(
7
(
(
9) Cho, B.; Park, W. Bull. Korean Chem. Soc. 1987, 8, 257.
10) (a) Morris, D.; Hayes, A.; Wills, M. J. Org. Chem. 2006, 71, 7035.
(
b) Tanaka, K.; Shoji, T. Org. Lett. 2005, 7, 3561.
11) (a) Ford, A.; Woodward, S. Angew. Chem., Int. Ed. 1999, 38, 335.
b) Blake, A.; Cunningham, A.; Ford, A.; Teat, S.; Woodward, S. Chem.;
(
(
Eur. J. 2000, 6, 3586.
12) Schubert, T.; Hummel, W.; Kula, M.; Muller, M. Eur. J. Org. Chem.
001, 4181.
13) (a) Ohkuma, T.; Ikehira, H.; Ikaria, T.; Noyori, R. Synlett 1997, 467.
b) Ohkuma, T.; Ooka, H.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1995,
17, 10417.
14) Matsumura, K.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1997, 119, 8738.
(
2
(
(
1
(
7
718 J. Org. Chem. Vol. 75, No. 22, 2010

Eagon et al.
JOCArticle
a
considered adding an alkyl or aryl substituent at the
R-position of a cyclic alkenyl ketone, but such substrates
are not commercially available and their synthesis is not
straightforward. We did, however, find a simple procedure
for the synthesis of 2-bromo-2-cyclohexen-1-one by reacting
TABLE 1. Reduction of r,β-Unsaturated Ketones with (L)-TarB-NO
2
2
-cyclohexen-1-one with molecular bromine followed by
7
2
triethylamine. When we reduced this brominated com-
pound with TarB-NO and NaBH(OAc) , we were pleased
to obtain the regioselective 1,2-reduction product in 92% ee.
2
3
When we reduced the same compound with NaBH , we were
4
pleased to again obtain only the 1,2-reduction product with
9% enantiomeric excess. Given the excellent reduction
9
results we obtained using 2-bromo-2-cyclohexen-1-one, we
prepared a variety of analogous substrates including
R-substituted iodine derivatives under Baylis-Hillman con-
2
8
ditions and an R-substituted phenyl derivative using a
2
Suzuki coupling. The results of our reduction studies are
9
summarized in Table 1.
We were greatly pleased to find that all of our substrates were
reduced with good to excellent enantioselectivity. All of the
simple R-substituted cyclohexenones (Table 1, entries 1, 2,
and 3) were reduced with a superb 99% enantiomeric
excess. More highly substituted R,β-unsaturated cyclohex-
enones were also reduced with good enantioselectivity by
TarB-NO (Table 1, entries 4 and 8). In addition to cyclo-
2
hexenone derivatives, we also investigated several R-sub-
stituted R,β-unsaturated cyclopentenones and were pleased
to find that all were reduced with greater than 90% en-
antioselectivity (Table 1, entries 5-7). We investigated the
reduction of some acyclic R,β-unsaturated ketones, but
unfortunately achieved very poor induction for these sub-
strates. However, an R-alkyl cyclic enone was reduced by
TarB-NO with good enantioselectivity (Table 1, entry 9).
2
It should also be noted that we routinely recovered the
R)-allylic alcohol when using (L)-TarB-NO . Given our
2
(
promising results with cyclic R,β-unsaturated alkenyl
ketones, we wished to next extend our investigation to
alkynyl derivatives.
Reduction of r,β-Unsaturated Alkynyl Ketones. Several
R,β-ynones were prepared by adding the appropriate alde-
hyde to a 1 M solution of the lithiated terminal alkyne in
THF at 0 °C under a dry and inert atmosphere. The racemic
product was then oxidized with pyridinium chlorochromate
(PCC) in DCM to give the desired substrate. The ynone
5-phenyl-4-butyn-3-one, however, was obtained commer-
cially and used without further purification. Optimization
studies indicated that 1 equiv of TarB-NO at room tem-
2
perature gave optimal results when reduced with sodium
1
1
borohydride. B NMR analysis of the reaction mixture also
confirmed that there was no hydroboration of the alkyne
during the reaction. Using this optimized procedure, we
reduced a variety of R,β-ynones with (L)-TarB-NO . The
2
results are summarized in Table 2.
The selectivity of the reaction was found to be highly depen-
dent upon the steric bulk of the nonalkynyl group attached
directly to the ketone. Methyl and ethyl alkynyl ketones were
reduced with little or no enantioselectivity, but isopropyl and
cyclohexyl alkynyl ketones were reduced with enantioselectivities
a
Standard reaction carried out under argon on a 4 mmol scale with
b
1
alcohols on a Supelco Beta-Dex 120 column. Assigned by comparison
4
.2 equiv of NaBH . Determined by GC analysis of the acetylated
c
d
to the literature value. Assigned by analogy.
from 75% to 83% (Table 2, entries 5-7). tert-Butyl alkynyl
ketones were reduced with the best enantioselectivity up to a
maximum of 90% ee for 2,2-dimethyl-non-4-yn-3-one (Table 2,
entry 10). This preference for highly branched aliphatic alkynyl
(
088.
27) Kowalski, C. J.; Weber, A. E.; Fields, K. W. J. Org. Chem. 1982, 47,
5
(
(
28) Krafft, M. E.; Cran, J. W. Synlett 2005, 1263.
29) Felpin, F. X. J. Org. Chem. 2005, 70, 8575.
J. Org. Chem. Vol. 75, No. 22, 2010 7719

Eagon et al.
JOCArticle
TABLE 2. Reductions of Ynones with (L)-TarB-NO
2
2
FIGURE 1. Chiral alcohols obtained with (L)-TarB-NO .
to branched aliphatic substrates (Table 2, entries 3 and 4).
Varying the size of the group attached on the alkyne distal to
the ketone seemed to have a limited effect, with larger groups
slightly reducing the enantioselectivity of the reaction. Also, we
were able to compare the optical rotation of several of our
products to those previously reported in the literature and
found that all were of the (R) configuration. Though TarB-
NO achieved poor reduction with unhindered n-alkyl alkynyl
2
ketones, these substrates have been successfully reduced with
31
high enantioselectivity by a modified CBS reagent. The NB-
Enantrane and Alpine-Borane reagents have also demon-
strated excellent stereoselectivity when reducing unhindered
7,32
n-alkyl alkynyl ketones.
This makes TarB-NO mediated
2
reductions of hindered alkyl alkynyl ketones complementary to
the reduction of unhindered analogues by these other popular
boron-based reagents.
Mechanistic Implications. We were pleased to find that like
previously investigated aromatic and aliphatic substrates,
(
L)-TarB-NO was able to reduce a variety of R,β-unsatu-
2
rated alkenyl and alkynyl ketones with excellent induction
Figure 1).
L)-TarB-NO2 mediated reductions of these substrates
(
(
consistently resulted in products of (R) configuration as
determined by the observed optical rotation in comparison
2
to previously published literature values. Given the sym-
6
metry of the TarB-NO reagent, these products are assumed
2
to arise from one of the four most likely transition states
determined by the coordinating lone pair of the substrate
ketone and the face of the substrate carbonyl presented to
the active hydride. These transition states (TS), with arbi-
trary conformations of the aromatic groups, are shown in
Figure 2.
We previously published ab initio calculations modeling
the transition state of TarB-NO based on early results with
the reduction of prochiral ketones and the results suggested
2
3
3
an orientation consistent with TS(I). The optical rotation
values corresponding to products of the (R) configuration
for previously investigated aromatic and aliphatic substrates
had always agreed with the configuration predicted from this
model (Figure 3).
In the case of acetophenone, the absolute configuration of
the product alcohol arises from a si-face attack correspond-
ing to either TS(I) or TS(II). This initial model, however, was
greatly simplified to reduce computation time and as such
a
Determined by GC analysis of the acetylated alcohols on a Supelco
Beta-Dex 120 column. Observed rotation value too small to accurately
b
c
d
measure. Assigned by analogy. Assigned by comparison to the litera-
ture value. See the Supporting Information.
ketones by a boron-based reduction reagent was previously
30
observed with the popular DIP-Cl reagent. Interestingly,
aromatic and heteroaromatic substituents bound to the ketone
gave markedly depressed enantioselectivities when compared
(31) Helal, C.; Magriotis, P.; Corey, E. J. Am. Chem. Soc. 1996, 118,
1
0938.
(
32) Brown, H. C.; Pai, G. J. Org. Chem. 1985, 50, 1384.
33) Cordes, D.;Nguyen, T.;Kwong, T.;Suri, J.;Luibrand, R.;Singaram, B.
(
Eur. J. Org. Chem. 2005, 5289.
(
30) Brown, H. C.; Ramachandran, V. Acc. Chem. Res. 1992, 25, 16.
7
720 J. Org. Chem. Vol. 75, No. 22, 2010

Eagon et al.
JOCArticle
FIGURE 4. Transition state models for the reduction of 2-phenyl-
cyclohex-2-enone. Given the fact that the absolute configuration of
the product dictates that the carbonyl must orientate the si-face of
the carbonyl toward the acyloxyborohydride, it is highly unlikely
that the phenyl ring of the ketone substrate is positioned immedi-
2
ately adjacent to the TarB-NO aromatic ring as per TS(I).
is obtained with a 99% ee and the phenyl ring is a convenient
fluorophore. The absolute configuration of the product
dictates that the hydride must be delivered to the si-face of
the ketone consistent with either TS(I) or TS(II). If TS(I) was
indeed the preferred orientation for this substrate, the close
proximity of the two aromatic rings of TarB-NO2 and
acetophenone suggested the possibility of a π-π stacking
interaction. To elucidate whether or not such stacking could
be observed, a UV spectrum was obtained for a THF
2
FIGURE 2. Likely transition states for reduction with (L)-TarB-NO .
solution of (L)-TarB-NO , acetophenone, and a mixture of
2
both. The resultant spectrum of the (L)-TarB-NO2 and
acetophenone mixture displayed neither a shift in the ob-
served λmax values, the appearance of new absorption peaks,
nor any temperature dependence;the standard hallmarks
3
4
of π-stacking interactions. Additionally, we examined the
1
H NMR spectrum of a 1:1 (L)-TarB-NO /acetophenone
mixture for the characteristic upfield shift of aromatic pro-
2
tons in π-π systems, but we observed no change in the
observed ppm values for any of the aromatic signals. These
35
results all indicate that no significant π-π interactions are
FIGURE 3. Previously published transition state model with acet-
ophenone corresponding to TS(I).
present between the TarB-NO reagent and acetophenone.
2
Though we found no direct evidence for π-stacking inter-
actions that could arise from TS(I), this could be due to the
fact that the ketone substrate is known to be only weakly
did not account for steric or solvent effects, nor did it contain
any aromatic rings. The limitations of the model soon
became apparent given our results with R,β-unsaturated
alkenyl and alkynyl ketones. When the R-substituted cyclic
R,β-unsaturated ketone 2-phenylcyclohex-2-enone (Table 1,
bound to the TarB-NO substrate in THF as evidenced by
2
1
1
B NMR studies. This could also be due to the fact that an
aromatic ring bearing a nitro group and another ring bearing
a carbonyl make poor partners for π-π interactions. To
make a more tightly bound model, we prepared the product-
like analogue boronate 1 in situ as per Scheme 2.
entry 3) is reduced with (L)-TarB-NO , the corresponding
2
(
R) alcohol is obtained with a 99% enantiomeric excess. The
absolute configuration of the product, as determined from
the optical rotation value, results from a si-face attack of the
hydride consistent with TS(I) and TS(II). However, when
modeling these two options, it is readily apparent that TS(I)
would have significant steric repulsion (Figure 4).
In addition to the questions raised by the reduction of
R,β-unsaturated alkenyl ketones, our model of TS(I) pre-
dicted the absolution configuration of our propargyl alco-
hols to be (S). In fact, when we compared the optical rotation
of our products to those previously reported in the literature,
we found that all were of the (R) configuration (Table 2).
Again, it should be noted that our original computational
model was greatly simplified to reduce calculation time.
Given these contradictory results and the limitations of our
initial computational model, we decided to conduct a more
Formation of 1 was confirmed by the shift observed in the
1
B NMR spectra, as well as the H NMR downfield shift of
1
1
the methine proton on the 1-phenylethanol moiety. More
notably, the aromatic signals for both the 1-phenylethanol
and TarB-NO phenyl ring in 1 were identical to the signals
2
observed for the free compounds, again suggesting that no
significant π-stacking interactions are present. It should be
noted that while boronate 1 may approximate the stable
configuration of the final product, it does not necessarily
approximate the active transition state that leads to the
experimentally observed products since the most stable
complex does not necessarily arise from the most reactive
one as per the Curtin-Hammett principle. However, these
solution-phase analogues can provide valuable insight
into constructing a better mechanistic model for TarB-NO2
detailed mechanistic investigation of TarB-NO mediated
2
reductions.
The reduction of acetophenone by (L)-TarB-NO2 was
selected first for investigation since the product (R)-alcohol
(34) Swager, T. M.; Song, C. Org. Lett. 2008, 10, 3575.
(35) Martin, R. B. Chem. Rev. 1996, 96, 3043.
J. Org. Chem. Vol. 75, No. 22, 2010 7721

Eagon et al.
JOCArticle
FIGURE 5. Selected NOESY interactions for 1.
SCHEME 2. Preparation of Boronate 1
FIGURE 6. Comparative reduction of an alkyl and alkynyl ketone
with TarB-NO .
2
mediated reductions. Specifically, the lack of any evidence for
π-π interactions in 1 suggests that other geometrical and/or
steric effects are responsible for the induction achieved with
TarB-NO2.
We next decided to investigate whether correlation
spectroscopy could provide additional evidence for either
TS(I) or TS(II) in the reduction of acetophenone. When we
attempted to acquire NOESY spectra of a (L)-TarB-NO2/
acetophenone mixture, we did not observe any through-
space interactions, again likely due to the weak association
FIGURE 7. Likely transition states for the reduction of alkynyl
ketones with TarB-NO . From a sterics standpoint, it would be more
reasonable to expect the tert-butyl group to be positioned above and
2
away from the TarB-NO reagent corresponding to TS(III) ratherthan
2
directly above the dioxoborolane ring of TarB-NO as in TS(IV).
2
Nevertheless, we decided to look for additional experimental evidence
supporting either TS(III) or TS(IV).
of the substrate to TarB-NO in a THF solution. As an
2
alternative, we decided to substitute boronate 1 prepared
in situ and perform a series of NOESY experiments
SCHEME 3. Preparation of Boronate 2
(Figure 5).
Irradiation of the methine and ortho protons of TarB-
NO all displayed through-space coupling to the methyl
2
protons on the phenylethanol moiety of 1. Irradiation of
phenylethanol methyl protons also resulted in coupling to
the methine protons on the tartrate backbone of TarB-NO2.
Additionally, when the aromatic protons of the phenyletha-
nol moiety are irradiated, no NOE effects are observed. Since
˚
presence of the triple bond was responsible for the observed
6
inversion of configuration.
NOE interactions are generally not observed beyond 4 A, the
observed interactions suggest that, in terms of 1, the aro-
3
The absolute configuration of the product propargyl
alcohol indicated that the alkyne was presenting the opposite
face of the carbonyl to the active acyloxyborohydride of
matic ring of the phenylethanol is positioned away from the
TarB-NO while the methyl group is close to the dioxobor-
olane ring of TarB-NO . Such a configuration would be
2
2
TarB-NO . Such a configuration would require an orienta-
2
consistent with either TS(II) or TS(III). Given that the
observed absolute configuration of the product can only
arise from a si-face attack represented by TS(I) and TS(II),
these results suggest that TS(II) is the most likely orientation
for reduction.
Having examined the evidence for the reduction of
acetophenone, we next turned our attention to alkynyl
ketones. Since the optical rotation values of six of our chiral
propargyl alcohols had been previously reported and all
were of the (R) configuration, we were confident in our
assignment of the absolute configuration of the products.
We suspected that the alkyne triple bond was responsible in
some manner for the apparent inversion of configuration.
To test this hypothesis, we prepared an alkane analogue of
one of our alkynyl ketones and reduced it with (L)-TarB-
NO2 (Figure 6). The striking result confirmed that the
tion consistent with either TS(III) or TS(IV). The remaining
question was whether the sterically demanding tert-butyl
group was positioned above the TarB-NO reagent as with
2
the phenyl ring of acetophenone, or if the alkyne was
positioned above the TarB-NO reagent (Figure 7).
2
To gain a better understanding of the orientation of the
transition state for the reduction of alkynyl ketones, we
decided to create the new boronate 2. Formation of the
1
11
B
new compound in situ was confirmed by both H and
NMR (Scheme 3).
To test for any interactions between the alkyne and
aromatic π systems, the UV spectra of both 2 and an alkane
(36) Note that the alkyne group has a higher priority than the tert-butyl
group under the Cahn-Ingold-Prelog convention.
7
722 J. Org. Chem. Vol. 75, No. 22, 2010

Eagon et al.
JOCArticle
FIGURE 8. Selected NOESY and ROESY correlations for 2.
analogue were obtained. Both spectra were identical, sug-
gesting no such interactions were present. We next took 2
and performed several NOESY and ROESY experiments,
the results of which are summarized in Figure 8.
A strong correlation was observed between the methine
proton of the propargyl alcohol and the methine proton on
the backbone of the dioxoborolane ring of the TarB-NO2
reagent, suggesting the conformer depicted above. More
importantly, however, was an interaction observed between
the methylene protons of the aliphatic carbon bonded to
FIGURE 9. Calculated transition state for the reduction of acet-
ophenone.
the alkyne and the ortho protons on TarB-NO . This sug-
2
gested that in terms of boronate 2, the alkyne was relatively
close to the aromatic ring of TarB-NO . If this orientation
2
holds true for the coordination complex between R,β-ynones
and TarB-NO , it would suggest a conformation consistent
with TS(III). Such a conformation would imply that R,β-ynones
2
coordinate to TarB-NO with the sterically demanding
2
group in the same position as with other aromatic and
aliphatic substrates, but presents the opposite face of the
carbonyl for reduction by the active hydride when compared
to TS(II). Since we have already shown that the alkyne alone
is responsible for this change in coordination configuration,
we believe that this is due either to the electronic properties or
the hybridization peculiar to the carbon-carbon triple bond.
Spectroscopic studies of the reduction of aromatic and
alkynyl ketones demonstrated that TS(I), indicated as the
active reducing species by earlier computational studies, was
not sufficiently accurate to predict the observed configuration
of the product alcohols. NMR studies of product-like boronate
analogues suggested that TS(II) could explain the absolute
configuration of the product alcohols obtained from the reduc-
tion of aromatic, aliphatic, and R,β-unsaturated cyclic ketones.
R,β-Ynones, however, are thought to assume an orientation
consistent with TS(III).
FIGURE 10. Calculated transition state for an R,β-ynone.
analogue was utilized with the nonparticipating carboxylic
acid and nitro group removed.
When the originally published transition state was recal-
culated with the above parameters, it was found that the
previous model was highly unstable and strongly favored
dissasociation. In fact, in the absence of the sodium atom the
previously published transition state completely disassoci-
ates. Under our computational parameters, we also found
To further develop a more accurate model, a series of new
ab initio computational studies were performed to model
the previously published transition state as well as the
reduction of acetophenone and an R,β-ynone. All calcula-
that substrate and TarB-NO favor disassociation, in agree-
2
11
ment with B NMR studies. However, when the reduction
3
7
tions were performed with the GAMESS software suite
of acetophenone was calculated, it was found that the attack
of the hydride from the si-face, resulting in the experimen-
tally observed (R) product, was energetically perferred by
1.62 kcal/mol (Figure 9).
The newly calculated transition state indicated an early
transition state wherein the hydride delivery occurs before a
formal coordination is established between the substrate
utilizing unconstrained geometry optimizations, double-
differenced seminumerical nuclear Hessian calculations,
and intrinsic-reaction coordinate calculations preformed
3
8
with the (R)PBE0 hybrid density functional and the
6
3
-31G(d) basis. In addition, s,p-diffuse functions were
9
40
added to all boron and oxygen atoms, as well as the
carboxylate carbon of TarB-X and the carbonyl carbon of
the ketone substrate. To aid in computation, a TarB-NO2
ketone and TarB-NO . Since the calculations were per-
2
formed in the absence of molecular motion (0 K) in the gas
phase and the hydride delivery occurs before coordination,
the orientation of the substrate ketone does not conform
directly to previously modeled transition states for TarB-
NO reductions performed in THF at room temperature.
2
However, these studies confirm on an ab initio basis that a
(37) Schmidt, M.; Baldridge, K.; Boatz, A.; Elbert, S.; Gordon, M.;
Jensen, J.; Koseki, S.; Matsunaga, N.; Nguyen, K.; Windus, T.; Dupuis,
M.; Montgomery, J., Jr. J. Comput. Chem. 1993, 14, 1347.
(
(
(
38) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158.
39) Hehre, W.; Ditchfield, R.; Pople, J. J. Chem. Phys. 1972, 56, 2257.
40) Frish, M.; Pople, J.; Binkley, J. J. Chem. Phys. 1984, 80, 3265.
J. Org. Chem. Vol. 75, No. 22, 2010 7723

Eagon et al.
JOCArticle
2
FIGURE 11. Proposed transition state for ketone reduction with (L)-TarB-NO .
2
FIGURE 12. Proposed transition state for R,β-ynone reduction with (L)-TarB-NO .
si-face hydride delivery is energetically favored in the case
of acetophenone, resulting in the experimentally observed
product.
Following the calculation of the transition state for acet-
ophenone, the same parameters were used to calculate the
reduction of an R,β-ynone. It was found that the ynone
presents the opposite face of the substrate carbonyl to the
approaching hydride in the energetically preferred transition
state, in agreement with spectroscopic and experimental
results (Figure 10).
substrate ketone with TarB-NO . Regardless of the ultimate
2
accuracy of this proposed transition state, the new model for
reduction correctly predicts the observed absolute config-
uration of alcohols obtained without assuming sterically
crowded, high-energy configurations as predicted by the
previous model. Reductions with R,β-ynones is an exception,
however. Ab initio, experimental, and spectroscopic studies
suggest that these substrates are likely reduced via a different
transition state, shown in Figure 12.
The ability of TarB-NO to prepare highly enantioen-
2
The calculated transition state for the alkynyl ketone also
indicated an early transition state where hydride delivery occurs
before a full coordination between the substrate and TarB-NO2
and thus does not directly correspond to previously modeled
transition states. Most importantly, however, this result con-
firms on an ab initio basis that an R,β-alkynyl ketone energe-
tically prefers a different orientation than that of acetophenone,
which leads to the experimentally observed product.
riched secondary alcohols from R,β-unsaturated ketones in 1
h at room temperature with NaBH , coupled with more
4
accurate mechanistic models for the prediction of product
configuration make it a highly effective tool for the asym-
metric reduction of prochiral ketones.
Experimental Section
2
Preparation of (L)-TarB-NO . A 500-mL round-bottomed
flask with a side arm and a stir bar was oven-dried, cooled under
argon, and charged with 3-nitrophenylboronic acid (20.04 g,
Conclusion
1
(
20 mmol), (L)-tartaric acid (18.00 g, 120 mmol), and CaH
10.10 g, 240 mmol), sealed with a septum, and fitted with a
2
We have reported a facile room temperature boron-mediated
reduction procedure for producing highly enantioenriched allyl
and propargyl alcohols using NaBH as the hydride source.
reflux condenser. Dry THF (240 mL) was added slowly through
the side arm and the system was refluxed for 1 h followed by
cooling. The suspension was cannulated with argon from the
sealed reaction vessel to a sealed medium frit attached to a dry
flask to remove calcium salts. The clear, yellow-brown solution
was then transferred via cannula to a dry ampule as a 0.5 M
solution in THF. H NMR indicates the yield was 99%. TarB-
NO solutions were stored in sealed ampules at room tempera-
ture and kept away from light. These stock solutions are stable
for as long as one year. (Note: TarB-X reagents are moisture-
sensitive. Air-sensitive techniques and dry glassware should be
used in their preparation and application.)
4
Reductions of R-substituted R,β-unsaturated cyclic alkenyl ke-
tones and R,β-ynones with TarB-NO proceed smoothly and
2
quickly under mild conditions at room temperature, producing
chiral alcohols of up to 99% ee with isolated yields up to
1
9
9%. For R,β-unsaturated alkenyl ketones, R-substituted
cycloalkenones exhibited the highest induction, while R,β- and
R,γ-substituted derivatives exhibited slightly lower enantioselec-
tivity. In the case of R,β-ynones, it was found that highly
branched aliphatic ynones were reduced with optimal induction,
while reduction of aromatic and linear aliphatic derivatives
resulted in more modest induction. Results obtained from spec-
troscopic and ab initio experiments indicated that most substrates
are likely reduced via the transition state depicted in Figure 11.
Though this transition state was not directly observed in
solution; spectroscopic, calculational, and experimental evi-
dence suggest that this is the most likely orientation of the
2
Procedure for the Reduction of r,β-Unsaturated Ketones with
2
TarB-NO . An oven-dried 100-mL round-bottomed flask
equipped with a stir bar was sealed with a septa and cooled under
argon. The flask was charged with TarB-NO (8 mL of a 0.5 M
solution in THF, 4 mmol) and the ketone (4 mmol). The solution
4
was stirred for 15 min after which solid NaBH (0.18 g, 4.8 mmol)
was added directly causing vigorous evolution of H gas. The
2
2
7
724 J. Org. Chem. Vol. 75, No. 22, 2010

Eagon et al.
JOCArticle
solution was allowed to stir for 1 h. The reaction was quenched
dropwise by the addition of H O (10 mL) and the mixture was
2
(R)-1-phenylethanol (0.242 mL, 2 mmol). The solution was then
cooled over an ice bath under argon. n-BuLi (0.83 mL of a 2.4 M
solution in hexanes, 2 mmol) was then added dropwise to the
solution. Once the addition was complete, the ice bath was
removed and the solution was allowed to warm to room
brought to pH 12 with 3 M NaOH (5 mL). The solution was
extracted with Et O (3 ꢀ 20 mL) and the combined extracts were
2
4
dried over MgSO , filtered, and concentrated to isolate the product
alcohol.
2
temperature (15 min.). A solution of TarB-NO (4 mL of a
Preparation of r,β-Ynones:. The Preparation of 2-Methynon-
-yn-3-one Is Representative An oven-dried 50-mL round-bot-
tomed flask equipped with a stir bar was sealed with a septa and
cooled under argon. A 20 mL sample of dry THF was trans-
ferred to the flask. The flask was then cooled over an ice bath.
0.5 M solution in THF, 2 mmol) was then added dropwise to the
flask. After 30 min of stirring, an aliquot (700 μL) was trans-
ferred to a sealed NMR tube for B NMR analysis to confirm
4
1
1
1
formation of the product (160.4 MHz, þ8.5 ppm). For H NMR
analysis, an aliquot (150 μL) of the solution was transferred to a
sealed NMR tube and evaporated under argon, leaving a white
1
-Hexyne (2.30 mL, 20 mmol) was added to the flask, followed
by the dropwise addition of n-BuLi (8.0 mL of a 2.5 M solution
in hexanes, 20 mmol). The mixture was stirred for 15 min, then
butyraldehyde (1.83 mL, 20 mmol) was added slowly to the
reaction mixture. The reaction was stirred an additional 30 min.
The ice bath was then removed and the reaction was quenched
with 2 mL of MeOH. The mixture was then transferred to a
separatory funnel and diluted with 45 mL of a 2:1 ethyl ether/
pentanes mixture. The organic layer was washed with 1 M HCl
chalky residue on the walls of the tube. CDCl
added and used as such for NMR analysis.
3
(700 μL) was then
Preparation of 2,2-Dimethylnonan-3-one. To a clean, dry, 100-mL
two-necked round-bottomed flask equipped with a stir bar
was added magnesium turnings (3.504 g, 120 mmol) and a
crystal of iodine. A condenser and addition funnel (50 mL)
were attached and the iodine was sublimed onto the turnings
with a heat gun. Once the flask had cooled, dry THF (30 mL)
was added to the flask. In a separate 50-mL round-bottomed
flask, a solution of 1-bromohexane (4.21 mL, 30 mmol) in dry
THF (30 mL) was prepared and then added to the addition
funnel. A small amount (∼5 mL) of the halide solution was
added to the THF over the magnesium. Once the reaction had
begun, the rest of the solution was added dropwise from the
addition funnel over the course of 30 min. Once addition was
complete, the solution was refluxed for 2 h. The Grignard
solution was then cooled over an ice bath and pivalaldehyde
(1.09 mL, 10 mmol) was added dropwise. After the solution
was stirred for 15 min, the ice bath was removed and the
solution was stirred an additional 2 h. The reaction mixture
was quenched with water, and then decanted into a separatory
funnel containing pentane (30 mL), leaving the unreacted
magnesium. The organic layer was washed with 1 M HCl
(3 ꢀ 20 mL), 1 M NaOH (2 ꢀ 20 mL), and then D.I. water (1 ꢀ
(
The organic layer was then dried over MgSO , filtered, and
2 ꢀ 20 mL), 1 M NaOH (2 ꢀ 20 mL), then D.I. water (1 ꢀ 20 mL).
4
evaporated under reduced pressure to yield the product as a colorless
or slightly yellow oil (2.74 g, 17.8 mmol, 89% yield). The crude
product was then transferred to an oven-dried 100-mL round-
bottomed flask equipped with a stir bar and diluted in enough
DCM to obtain approximately a 0.25 M solution (70 mL). To this
solution was added 1.5 equiv of PCC (5.76 g, 26.7 mmol) and the
solution was allowed to stir for 16 h. The DCM was removed under
reduced pressure, then pentanes (50 mL) and Celite (∼2 g) were
added to the flask and the solution was stirred vigorously for 30 min,
or until the chromium salts were free-flowing in the solution. The
suspension was then filtered through a Buchner funnel to remove the
salts. The remaining pentane layer was evaporated under reduced
pressure to yield the product 2-methylnon-4-yn-3-one (1 g) as a
colorless oil (1.96 g, 12.8 mmol, 72% yield).
Procedure for the Reduction of Ynones with TarB-NO :. The
2
20 mL). The organic layer was dried over MgSO , filtered, and
4
Reduction of 2-Methylnon-4-yn-3-one Is Representative A 25-mL
oven-dried round-bottomed flask was equipped with a stir bar
evaporated under reduced pressure to yield the intermediate
alcohol 2,2-dimethylnonan-3-ol as a colorless oil (1.502 g,
8.7 mmol, 87%). A 1.10 g sample of the alcohol (6.4 mmol) was
transferred to a 50-mL round-bottomed flask equipped with a stir
bar. Glacial acetic acid was added to the flask (4.27 mL, 6.4 mmol)
then an addition funnel was connected to the top of the flask.
Bleach (14.3 mL of a 5% solution, 9.6 mmol) was added to the top
of the addition and allowed to slowly drip into the acid with stirring
at a rate of approximately 1 drop per 5 s. The solution was stirred
for 1 h following the complete addition of the bleach solution.
Starch-iodide paper was used to confirmed the presence of excess
hypochlorite, which was subsequently quenched by adding an
excess of a saturated sodium bisulfite solution. The destruction
of the hypochlorite is confirmed by starch-iodide test strips, then
the solution is extracted with pentantes (3 ꢀ 15 mL). The com-
bined organic layers are washed with 1 M NaOH (2 ꢀ 10 mL), then
and cooled under argon. The flask was charged with TarB-NO
(
2
6 mL of a 0.5 M solution in THF, 3 mmol) and 2-methylnon-
4
-yn-3-one (1 g) (0.454 g, 3 mmol). The ketone and TarB-NO
(0.227 g, 6 mmol) was
added directly to the solution causing vigorous evolution of H
2
were stirred for 15 min after which NaBH
4
2
gas. The mixture was allowed to stir for 30 min. The solution was
quenched dropwise with 1 M HCl until gas evolution ceased.
The mixture was brought to pH 12 with 3 M NaOH and stirred
for 30 min. The solution was extracted with a 2:1 ether/pentane
mixture (3 ꢀ 10 mL). The combined organic layers were then
washed with 3 M NaOH (2 ꢀ 10 mL), 1 M HCl (2 ꢀ 10 mL), and
D.I. water (1 ꢀ 10 mL). The organic layer was then dried over
MgSO , filtered, and concentrated under reduced pressure to
4
obtain the product 2-methylnon-4-yn-3-ol (2 g) as a light amber
oil (0.347 g, 2.25 mmol, 75% yield). The enantiomeric excess of
the acetylated alcohol was determined by GC on a Supelco
β-cyclodextrin 120 column (30 m ꢀ 0.25 mm). To recover the
arylboronic acid, the combined aqueous layers were acidified
with concd HCl to pH 1 and extracted with diethyl ether (3 ꢀ
4
dried over MgSO , filtered, and evaporated under reduced pres-
sure to obtain the product ketone as a colorless oil (1.09 g,
6.3 mmol, 97%).
Acknowledgment. The authors would like to thank
Mr. Jim Loo for his assistance with the NMR spectroscopy
experiments.
1
0 mL). The combined organic layers are washed quickly with a
small amount of sat. NaCl solution (5 mL) and dried over
MgSO4. Evaporation under reduced pressure yields the phen-
ylboronic acid as an off-white flakey solid, which was stored in a
desiccator.
Preparation of Boronate (1). A 25-mL oven-dried round-
bottomed flask was equipped with a stir bar and cooled under
argon. Dry THF (3.08 mL) was added to the flask, followed by
Supporting Information Available: Spectra for alkynyl and
transition state analogue substrates along with computational
parameters. This material is available free of charge via the
Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 22, 2010 7725