which decompose with chromatographic purification and
storage.1b In this case, the removal of the benzyl groups
through catalytic hydrogenation can also restrict the potential
synthetic applications of these intermediates.1b,6 Rapoport’s
use of the N-PhFl protection largely solved these racemiza-
tion and purification problems.7 However, the installation
of this group employs stoichiometric quantities of the
environmentally unfriendly Pb(NO3)2, and its removal re-
quires harsh conditions (e.g., TFA, 8 h, 80 °C or Li/NH3
(l)). Thus, the appropriate precursors to these aldehydes are
not commercially available, and these groups are rarely used.
For some time, we have had an interest in new chemistry
orchestrated through the triisopropylsilyl (TIPS) substitution.8
The highly compact and effective steric bulk of the triiso-
propyl substitution on silicon retards reactions not only at
the silyl center but also at adjacent centers. We felt that the
TIPS group should combine ease of installation and removal
with effective steric bulk which impedes the racemization
of enolizable amino aldehydes (i.e., 5). Consideration of these
properties and the above limitations for the known R-amino
aldehydes led us to examine the potential of the bulky,
electron-donating TIPS group to provide thermally stable and
isolable N-protected R-amino aldehydes (5) in enantiomeri-
cally and chemically pure form. We envisaged the N,O-bis-
(triisopropylsilyl)-R-amino acids 2 as convenient precursors
to 5 (Scheme 1). While both TIPS amines and esters are
Table 1. N-TIPS-R-Amino Aldehydes 5 from 1
R
series
2 (%)a
3b (%) (c/t)c
5 (%)a
Me
Prd
(CH2)2SMe
i-Bu
Bn
CH2OBn
Ph
a
b
c
d
e
f
74
72
72
82
87
82
85
100 (31/69)
100 (18/82)
100 (13/87)
100 (14/86)
100 (20/80)
100 (20/80)
100 (25/75)
56
56
56
54
52
50
49
g
a Yields of isolated analytically pure material. b The yields for 3 from
this process were quantitative (100 ( 2%). c The cis isomer exhibits J(H4-
H5) ) 5-6 Hz, while the trans isomer has a negligible three-bond H-H
coupling J(H4-H5) ∼ 0 Hz. d The D-amino acid was also silylated to give
D-2b (83%) and reduced to D-3b (100%, c/t ) 15/85) and hydrolyzed to
D-5b (60%).
tocol. More significantly, the inclusion of the bulky Hu¨nig’s
base in the procedure provided the key to avoiding any loss
of optical purity in the product 2 even for challenging
systems such as O-benzylserine (2f) or phenylglycine (2g).
This is a major issue for these examples when DBU is used
as the base.
With the highly soluble and stable silyl derivatives 2 in
hand, we chose to examine their semi-reduction with
borane-dimethyl sulfide complex (DMSB). While DIBAL-H
is commonly employed for related processes with amino
esters,1a,4a to our knowledge, the analogous borane-based
process is unprecedented. TIPS esters specifically are fully
reduced to the corresponding alcohol when heated neat with
1 molar equiv of DMSB. In fact, when a 1:1:1 mixture of
2e, PhCH2COOTIPS, and DMSB is heated, only the PhCH2-
COOTIPS is reduced (to PhCH2CH2OH) with 2e being
unaffected. With the bulky 2°-amine present in 2, after the
initial “hydroboration” of the carbonyl, we view the BH2
moiety as reversibly complexing this amine, ultimately
reacting further to produce hydrogen and the oxazaborolidine
3. This reaction diminishes both the Lewis acidity and the
mobility of the boron atom. Thus, 3 is stable, showing no
tendency to undergo â-elimination of a BOTIPS moiety to
generate an aldehydic carbonyl group which would be further
reduced. The oxazaborolidines 3 are produced quantitatiVely
as cis/trans mixtures (Table 1). These isomers exhibit a
Scheme 1
known,8a the only reported N,O-disilylated R-amino acids
are the hydrolytically unstable trimethylsilyl derivatives.9 We
expected this process to be straightforward. However,
considerable effort was required to find conditions which
provide an efficient route to 2 without its partial racemiza-
tion. Fortunately, through the slow addition of TIPSOTf to
1/(i-Pr)2NEt in refluxing THF, the clean 1 f 2 conversion
was achieved (Table 1). The unwanted formation of poly-
tetrahydrofuran was completely avoided through this pro-
number of clearly resolved signals in both the H and 13C
1
NMR spectra which can be used to evaluate this ratio in
each case (see Supporting Information). Interestingly, while
these BH oxazaborolidines evidently exist as monomers
(ν(B-H) ) 2550-2580 cm-1, 11B NMR δ 31-33), we were
1
unable to observe a doublet in the H-coupled 11B NMR
spectra of 3. This phenomenon has been observed by Corey
in related systems.10 Remarkably, despite containing the
mixed acetal moiety, 3 is stable and can be stored under
nitrogen for months without significant decomposition!
Very mild reaction conditions were developed to effect
the clean hydrolysis of 3 through 4 to provide the pure
R-amino aldehydes 5. This was accomplished through the
(6) Hyun, S. I.; Kim, Y. G. Tetrahedron Lett. 1998, 39, 4299.
(7) (a) Lubell, W. D.; Rapoport, H. J. Am. Chem. Soc. 1987, 109, 236.
(b) Lubell, W. D.; Rapoport, H. J. Org. Chem. 1989, 54, 3824. (c) Jamison,
T. F.; Rapoport, H. Org. Synth. 1993, 71, 226.
(8) (a) Ru¨cker, C. Chem. ReV. 1995, 95, 1009 and references cited therein.
(b) Soderquist, J. A.; Vaquer, J.; D´ıaz, M. J.; Bordwell, F. G.; Zhang, S.
Tetrahedron Lett. 1996, 37, 2561. (c) Justo de Pomar, J. C.; Soderquist, J.
A. Tetrahedron Lett. 1998, 39, 4409. (d) Justo de Pomar, J. C.; Soderquist,
J. A. Tetrahedron Lett. 2000, 41, 3285. (e) Soderquist, J. A.; Justo de Pomar,
J. C. Tetrahedron Lett. 2000, 41, 3537.
(9) (a) Gehrke, C. W.; Leimer, K. J. Chromatogr. 1971, 57, 219. (b)
Venkateswaran, P. S.; Bardos, T. J. J. Org. Chem. 1967, 32, 1256.
(10) For a similar system, see: Corey, E. J.; Bakshi, R. K.; Shibata, S.
J. Am. Chem. Soc. 1987, 109, 5551.
334
Org. Lett., Vol. 10, No. 2, 2008