B-ally-10-Ph-9-borabicyclo[3.3.2]decanes: Strategically designed for the asymmetric allylboration of ketones

Article abstract of DOI:10.1021/ja053865r

The simple and efficient syntheses of B-allyl-10-(phenyl)-9-borabicyclo[3.3.2]decane (1) in both enantiomeric forms are reported. The remarkable enantioselectivity (81-99% ee) of these reagents in the allylboration process at -78 °C is only modestly diminished when the process is conducted at 0 °C, a phenomenon attributable to its rigid bicyclic structure. In addition to providing the homoallylic alcohols 6 efficiently (70-92%), the procedure also permits the efficient recovery of the chiral boron moiety (67-82%) as an air-stable crystalline N-methylpseudoephedrine complex 4 for the direct regeneration of 1 with allylmagnesium bromide in ether (98%). The reagent gives predictable stereochemistry, providing a strategically designed "chiral pocket" which is particularly receptive to leading methyl groups (e.g., methyl ethyl ketone, 87% ee). Copyright

Full text of DOI:10.1021/ja053865r

Published on Web 07/30/2005
B-Allyl-10-Ph-9-borabicyclo[3.3.2]decanes: Strategically Designed for the
Asymmetric Allylboration of Ketones
Eda Canales, K. Ganeshwar Prasad, and John A. Soderquist*
Department of Chemistry, UniVersity of Puerto Rico, Rio Piedras, PR 00931-3346
Received June 12, 2005; E-mail: jas@janice.uprr.pr
Scheme 1
The asymmetric allylation of aldehydes provides a highly
effective method for the preparation of nonracemic homoallylic
alcohols.1 However, a general procedure for preparation of 3°-
homoallylic alcohols through the asymmetric allylation of ketones
has remained conspicuously wanting. These substrates can be
selectively allylated with allylsilanes through their chiral ketals,
but obtaining the free alcohols is nontrivial and inconvenient.2
Recently, significant progress has been made with stereoelectroni-
cally biased ketones (e.g., PhCOMe) through the use of asymmetric
catalysts and either allyltins or achiral allylboronic ester reagents
or, alternatively, through the stoichiometric reactions of chiral
allylboronic esters derived from 3,3′-disubstituted-1,1′-bi-2-
naphthol.3 Unfortunately, these methods are significantly less
selective for dialkyl ketones containing pendent groups, which are
more similar in size (e.g., PhCH2CH2COMe). Careful examination
of these and other systems employed for asymmetric allylation
suggested that a fully elaborated “chiral pocket” may better address
the challenges posed by these ketones. We envisioned a process in
which the smaller of the two groups could better fit into this pocket.
A key feature in this new design paradigm was the inclusion of a
“chiral floor”, which we felt could be achieved through the
incorporation of the reactive site into a bicyclic system.
Recently, we reported the synthesis of the B-allyl- and B-crotyl-
10-trimethylsilyl-9-borabicyclo[3.3.2]decanes (BBDs) and their
additions (3 h, -78 °C) to aldehydes achieving excellent selectivities
(94 f 99% ee, >98:2 dr) with a wide range of substrates.4
Unfortunately, with these reagents, the asymmetric allylboration
of acetophenone is both very slow (2 days, 25 °C) and far less
Scheme 2
selective (62% ee). This disappointing result led us to view the
10-TMS-9-BBD system as providing a chiral pocket which is
receptive to aldehydic hydrogens, but is too small to easily
accommodate even a methyl group. With model studies suggesting
that this pocket may be incrementally enlarged with a 10-Ph rather
than 10-TMS BBD system, we chose to prepare the reagents 1 for
the asymmetric allylboration of methyl ketones.
The preparation of the thermally stable (()-3 from 2 was
accomplished through the simple reaction of the stable PhCHN2,
which cleanly inserts the CHPh group into a ring B-C bond
(hexanes, 10 h, 0 °C, 90%) (Scheme 1). In contrast to its essentially
air-stable 10-TMS counterpart, 3 is readily oxidized in the open
atmosphere, but is indefinitely stable when handled and stored under
nitrogen. Employing a modified version of the Masamune resolution
protocol,5 (()-3 was added to 0.5 equiv of (1S,2S)-N-methylpseu-
doephedrine (NMPE) in hexanes which provides (+)-4S as a pure
crystalline compound (39%). After concentration of the supernatant
to remove the liberated MeOH, 0.5 equiv of (1R,2R)-NMPE was
added to a fresh solution of the residue in hexanes, ultimately giving
a 28% yield of the pure crystalline (-)-4R. Reversing this order
gives first, (-)-4R and second, (+)-4S, also in a 67% combined
total yield of enantiomerically pure forms of 4 from (()-3! These
complexes are air-stable and can be stored indefinitely. The complex
(+)-4R is wholly chelated in the solid state, while in solution, 4
exists in both the open and closed forms (11B NMR (C6D6) δ 55.5,
10.0). The preparation of the B-allyl reagents (+)-1S or (-)-1R
through the reaction of either complex 4 with allylmagnesium
bromide in Et2O is both simple and efficient (98%). The byproduct
NMPE Mg2+ salt is easily converted back to NMPE for reuse
(86%).
The reagents 1 were examined in the asymmetric allylboration
of representative ketones (Scheme 2). In contrast to the slow and
moderately selective reactions of its 10-TMS counterpart, 1 reacts
rapidly with methyl ketones (e3 h, -78 °C). The resulting 3°-
homoallylic alcohols (6) are obtained efficiently (70-92%) in high
enantiomeric excess (81-99%) (Table 1). These selectivities equal
or exceed the selectivities of any known process for sterically biased
methyl ketones (6a,c-e,h,i). With the more demanding cases, where
the groups on the ketone are similar in size, the levels of
enantioselectivity observed with 1 are particularly noteworthy (i.e.,
f (MEK, 87% ee), g (MVK, 81% ee). With the ethyl ketone,
9
11572
J. AM. CHEM. SOC. 2005, 127, 11572-11573
10.1021/ja053865r CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Table 1. Allylboration of RLRsCO with 1
ArCOMe), anti-complexation is clearly dominant (cf., B vs D). This
leads to B being preferred over its “trans” counterpart C on steric
grounds with the smaller oxygen atom versus the methylene of the
allyl group being best positioned “cis” to the 10-R substituent. This
is a major effect, and allylation occurs nearly exclusively with (-)-
1R on the re face of the stereoelectronically biased ketones (see
Table 1, (+)-1S 6 si face) as it does in the analogous aldehyde
process with the 10-TMS reagent (96 f 99% ee).4,7 However, for
even methyl groups, this pocket is too small with the 10-TMS
system, and slow allylation at 25 °C is less selective with respect
to the cis versus trans reaction pathways. Fortunately, with 1, the
combination of smaller flat Ph group and chair form of the ring on
the cis side works in concert to provide an ideal pocket for methyl
groups, and the allylation is both rapid and highly selective at low
temperatures (B). With a leading Et versus Me group, computational
analysis reveals that the Et group must rotate away from the ring
to form a viable pretransition state complex resulting in a slower
reaction for propio- versus acetophenone. However, both do react
at low temperature with high enantioselectivity. As a consequence,
for ketones such as MEK and MVK (Table 1, series f, g), in addition
to C, the isomeric syn complex, D, must also be considered as a
potential source of diminished enantioselectivity.
RL
RS
1
series
6 (%)a
%eeb (abs config)
Ph
Ph
Me
Et
R
S
S
S
R
S
S
S
R
R
a
b
c
d
e
f
92
70
96
89
90
80
77
74
70
82
96 (R)
94 (S)
98 (S)
94 (S)
>98 (R)
87 (R)
81 (S)
92 (S)
99 (R)
90 (R)
4-BrC6H4
4-MeOC6H4
4-O2NC6H4
Et
CH2dCH
i-Pr
Me
Me
Me
Me
Me
Me
Me
H
g
h
i
t-Bu
Ph
j
a All runs were made in duplicate (at least), and the a and e series were
performed with both (-)-1R and (+)-1S. The intermediate 5 was isolated
and converted to 6, and 4 was recovered (67-82%) via the NMPE workup
procedure. b Product enantiomeric excesses were determined by conversion
to the Alexakis esters and analysis by 31P NMR. For 6a,d, this enantiomeric
excess value was confirmed by HPLC (DAICEL CHIRACEL OD, hexane/
2-propanol 99:1).
propiophenone, the allylation is significantly slower (∼8 h, -78
°C) than that with methyl ketones, but 1 still exhibits high selectivity
(6b (94% ee)). In a competitive experiment employing a 1:1.33:
1.33 mixture of 1, PhCOMe, and PhCOEt, the acetophenone is
selectively allylated, leaving the propiophenone completely unre-
acted. With PhCHO, 1 is less selective (90% ee) than with its 10-
TMS counterpart (g98% ee).4 Additionally, when the allylation of
PhCOMe is conducted at 0 °C, the homoallylic alcohol 6a is
obtained in 90% ee. This modest diminution in enantioselectivity
at significantly higher allylation temperatures is a signature feature
of the rigid bicyclic BBD reagents.
The reagents 1 are easily prepared in either enantiomerically pure
form from air-stable precursors and are highly reactive, environ-
mentally friendly, and fully recyclable. In combination with their
high selectivities for a wide range of unsymmetrical ketones, these
features of the new BBD reagents make them highly attractive alter-
natives to existing methods for the asymmetric allylation of ketones.
The predictability and consistently high levels of enantioselec-
tivity in the allylation of ketones with 1 warrant further discussion.
As previously discovered,4 the 10-TMS system is ideally suited
for the asymmetric allylation of aldehydes, but is too hindered to
effectively allylate ketones. The greater reactivity and selectivity
of 1 are a direct result of the changes engendered in the chiral pocket
with the 10-Ph versus 10-TMS substitution. A comparison of the
single-crystal X-ray structure of (+)-4S versus the analogous
pseudoephedrine complex of the 10-TMS-9-BBD system4 suggests
that this reactivity can be attributed both to (1) the lesser steric
bulk of the Ph versus TMS groups, and (2) the shorter C-Ph (1.53
Å) versus C-TMS (1.87 Å) bond, which forces the more open
chair component of the favored boat/chair conformation of the ring
to be cis to the Ph group. This is reversed for the 10-TMS system,
where the boat component is favored on this side (cf., B vs A).
Acknowledgment. The support of the NSF (CHE0217550),
NIH (S06GM8102), Merck, Sharpe & Dohme Quimica de Puerto
Rico, Inc., and the Department of Education GAANN Program
(P200A030197-04) is gratefully acknowledged. We thank Dr.
Charles L. Barnes, University of Missouri, Columbia for the X-ray
structure of (+)-4S. This work is dedicated to the memory of
Professor Satoru Masamune.
Note Added after ASAP Publication. In the Supporting
Information published on the Internet July 30, 2005, there was an
error in Figure 2 on page 2. The SI published August 4, 2005, is
correct.
Supporting Information Available: Experimental procedures,
analytical data and selected spectra for 1, 4, 6, and derivatives, and
X-ray data for (+)-4S. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) (a) Ramachandran, P. V. Aldrichimica Acta 2002, 35, 23. (b) Jadhav, P.
K.; Bhat, K. S.; Perumal, P. T.; Brown, H. C. J. Org. Chem. 1986, 51,
432. (c) Denmark, S. E.; Fu, J. Chem. ReV. 2003, 103, 2763. (d) Roush,
W. R.; Grover, P. T. J. Org. Chem. 1995, 60, 3806.
(2) Tietze, L. F.; Schiemann, K.; Wegner, C. J. Am. Chem. Soc. 1995, 117,
5851.
(3) (a) Casolari, S.; D’Addario, D.; Tagliavini, E. Org. Lett. 1999, 1, 1061.
(b) Waltz, K.; Gavenonis, J.; Walsh, P. J. L. Angew. Chem., Int. Ed. 2001,
41, 3697. (c) Kim, J. G.; Waltz, K. M.; Garcia, I. F.; Kwiatkowski, D.;
Walsh, P. J. J. Am. Chem. Soc. 2004, 126, 12580. (d) Wada, R.; Oisaki,
K.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2004, 126, 8910. (e)
Wu, T. R.; Shen, L.; Chong, J. M. Org. Lett. 2004, 6, 2701. (f) Teo,
Y.-C.; Goh, J.-D.; Loh, T.-P. Org. Lett. 2005, 7, 2743.
(4) Burgos, C. H.; Canales, E.; Matos, K.; Soderquist, J. A. J. Am. Chem.
Soc. 2005, 127, 8044.
(5) Short, R. P.; Masamune, S. J. Am. Chem. Soc. 1989, 111, 1892.
(6) (a) Yamamoto, Y.; Asoa, N. Chem. ReV. 1993, 93, 2207-2293. (b) Omoto,
K.; Fujimoto, H. J. Org. Chem. 1998, 63, 8331. (c) Gung, B. W.; Xue,
X.; Roush, W. R. J. Am. Chem. Soc. 2002, 124, 10692. (d) Li, Y.; Houk,
K. J. Am. Chem. Soc. 1989, 111, 1236. (e) Performed using the Spartan
4.0.4a GL MM program.
(7) Allenylation is slightly less selective (93-95% ee) for aldehydes because
the smaller sp2 versus sp3 carbon in an allenyl versus allyl group leads to
a lesser preference for cis versus trans selection: Lai, C.; Soderquist, J.
A. Org. Lett. 2005, 7, 799.
The 10-substituted-9-BBD ring clearly defines a chiral pocket,
as is illustrated in the energetically favored pretransition state
complexes A and B for aldehydes and ketones, respectively.6 These
complexes are anti with respect to the borane/carbonyl compound,
cis with respect to position of carbonyl relative to the 10-R group,
and “down” with respect to the orientation of the aldehyde or ketone
with respect to the ring system. Eight diastereomeric complexes
can result by varying these three parameters. All four of the “up”
complexes are prevented from reaching energetically competitive
transition states by adverse interactions with the 10-R groups. For
substrates with groups which differ greatly in size (e.g., RCHO,
JA053865R
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11573