Using a Temporary Silicon Connection in Stereoselective Allylation with Allylsilanes: Application to the Synthesis of Stereodefined 1,2,4-Triols

Article abstract of DOI:10.1021/ol035762o

(Equation presented) Treatment of aldehyde 6 with TMSOTf, in the presence of a Bronsted acid scavenger, effects an intramolecular allylation to provide the oxasilacycle 7 as the major diastereoisomer. Tamao oxidation of the C-Si bond in 7 affords the corresponding 1,2,4-triol 9.

Full text of DOI:10.1021/ol035762o

ORGANIC
LETTERS
2003
Vol. 5, No. 22
4231-4234
Using a Temporary Silicon Connection
in Stereoselective Allylation with
Allylsilanes: Application to the
Synthesis of Stereodefined 1,2,4-Triols
Julien Beignet and Liam R. Cox*
School of Chemistry, UniVersity of Birmingham, Edgbaston,
Birmingham B15 2TT, U.K.
l.r.cox@bham.ac.uk
Received September 12, 2003
ABSTRACT
Treatment of aldehyde 6 with TMSOTf, in the presence of a Brønsted acid scavenger, effects an intramolecular allylation to provide the
oxasilacycle 7 as the major diastereoisomer. Tamao oxidation of the C−Si bond in 7 affords the corresponding 1,2,4-triol 9.
The reaction between an aldehyde and an allylmetal to
provide a homoallylic alcohol remains one of the most
powerful and widely used methods for the stereoselective
formation of C-C bonds.1 Nucleophilic allyl reagents can
be grouped into three classes according to the mechanism
by which they react with aldehydes.1,2 This classification is
not only a function of the metal; it is also dependent on the
coordinating ligands, the presence of additives, and the
reaction conditions. We are interested in employing so-called
Type II allylmetals, specifically allyltrialkylsilanes, in stereo-
selective synthesis.1c,3 These are potentially very attractive
reagents for synthesis, owing to their ease of preparation,
relatively high stability, and low toxicity. Unfortunately, the
inherently low nucleophilicity of allyltrialkylsilanes offsets
some of these advantages. Thus, to use these reagents as
nucleophiles in synthesis, it is usually necessary to either
increase the nucleophilicity of the allylsilane (e.g. with
fluoride4), or improve the reactivity of the participating
electrophile (usually with a Lewis acid). This requirement,
in combination with the weak Lewis acidity of the metal
center, has important ramifications for the reaction mecha-
nism: allyltrialkylsilanes invariably5 react with aldehydes
through an open, acyclic T. S.1c This type of T. S. is poorly
defined when compared with the closed Zimmerman-Traxler
T. S.s through which reactions between Type I allylmetals
(e.g. allylboranes) and aldehydes proceed.1b Nevertheless the
reaction between a Type II allylmetal and an aldehyde can
(1) (a) Denmark, S. E.; Fu, J. Chem. ReV. 2003, 103, 2763-2794. (b)
Roush, W. R. in ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Pergamon: Oxford, 1991; Volume 2, Chapter 1.1, pp 1-53. (c)
Fleming, I. in ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Pergamon: Oxford, 1991; Volume 2, Chapter 2.2, pp 563-593.
(d) Yamamoto, Y.; Asao, N. Chem. ReV. 1993, 93, 2207-2293.
(2) Denmark, S. E.; Weber, E. J. HelV. Chim. Acta 1983, 66, 1655-
1660.
(4) Hosomi, A.; Shirahata, A.; Sakurai, H. Tetrahedron Lett. 1978, 3043-
3046.
(5) Silacyclobutane systems would be an important exception: (a)
Denmark, S. E.; Griedel, B. D.; Coe, D. M.; Schnute, M. E. J. Am. Chem.
Soc. 1994, 116, 7026-7043. (b) Myers, A. G.; Kephart, S. E.; Chen, H. J.
Am. Chem. Soc. 1992, 114, 7922-7923.
(3) Fleming, I.; Dunogue`s, J.; Smithers, R. Org. React. 1989, 37, 57-
575.
10.1021/ol035762o CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/10/2003

sometimes be highly diastereoselective.1,3 Moreover, chiral
Lewis acids can render the reaction highly enantioselective
which further increases the synthetic utility of these reagents.1a,6
A number of approaches have been devised for over-
coming some of the shortfalls associated with allylsilanes.
One of the most successful is to modify the ligands on the
allylic silicon center such that reaction proceeds through a
closed T. S. and behaves more like a Type I allyl reagent.7
We are interested in retaining the alkyl ligands around the
silicon and investigating an alternative strategy, namely to
use a temporary connection, for controlling the stereoselec-
tivity of an allylation reaction.8 By employing a silyl ether
connection to temporarily tether the two reacting partners, a
sequence of tether formation, followed by reaction, and
finally tether cleavage, provides a product that is the result
of a net intermolecular reaction, yet one that has benefited
from all the advantages associated with an intramolecular
process. Thus, by using this approach we will have an
allylsilane that retains the attractive properties of Type II
reagents, but one that should exhibit increased reactivity and
react through a better defined - and hopefully more
predictable - T. S., leading to improved - or different -
levels of stereocontrol to the analogous intermolecular
process.
clearly demonstrate that temporarily tethering an allylsilane
in this fashion leads to consistently higher (or opposite in
the Reetz example) levels of stereocontrol compared to the
analogous intermolecular reactions.
We envisaged that relocating the silyl ether tether from
the allylic silicon group to the γ-position of the allylsilane
would confer a number of potential advantages over the
system employed by Reetz and others (Scheme 2).
Scheme 2. Relocating the Silyl Ether Tether to the γ-Position
of the Allylsilane Provides an Oxasilacyclic Product Rather than
an Acyclic 1,3-Diol as is Obtained in the Reetz System
Reetz has shown that tethering an allylsilane through the
carbinol center of a â-hydroxy aldehyde led to the 1,3-syn
diol allylation product on activation with TiCl4 (Scheme 1).9
Scheme 1. Tethering an Allylsilane to an Aldehyde
Electrophile Leads to a Different Sense of Stereoinduction
(Ref 9)
In our modified system, the reacting allyl group would
now be exocyclic in the T. S. and therefore more closely
resemble the corresponding intermolecular reaction. Reaction
would also proceed through a tighter, and better-defined, T.
S. to provide a cyclic product in which the tether remains
intact. Furthermore, by preserving the tether in the reaction
we would generate two - as opposed to one - new
stereogenic centers and therefore obtain improved stereo-
chemical transcription. Finally, the oxasilacycle product is
a very versatile intermediate ripe for further elaboration
(Scheme 2).
The γ-(amino)silyl-substituted allylsilane 1 was readily
synthesized according to the procedure of Tamao et al. as
outlined in Scheme 3.12 Aminosilanes are attractive silylating
agents: treatment of â-hydroxy ester 2 with aminosilane 1
provided the desired silyl ether 3 in excellent yield after
purification by chromatography (Scheme 3). Silylation was
neither particularly rapid nor exothermic, and was best
achieved by simply mixing equimolar quantities of the two
reagents in the absence of solvent and with slight warming
of the reaction mixture to aid mixing of the two reactants.
Since the only byproduct in the reaction is a volatile
secondary amine, acid scavengers are not required which
facilitates workup. DIBALH reduction of the ester in 3 then
This result was significant in that the sense of induction in
this intramolecular process9 was opposite that found in the
analogous intermolecular reaction.10 This, and other studies,11
(6) For an enantioselective allylation reaction employing allyltrialkyl-
silanes: Gauthier, D. R.; Carreira, E. M. Angew. Chem., Int. Ed. Engl. 1996,
35, 2363-2365.
(7) For some very recent examples: (a) Denmark, S. E.; Fu, J. Chem.
Commun. 2003, 167-170. (b) Kubota, K.; Leighton, J. L. Angew. Chem.,
Int. Ed. 2003, 42, 946-948. (c) Yamasaki, S.; Fujii, K.; Wada, R.; Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2002, 124, 6536-6537.
(8) Cox, L. R.; Ley, S. V. in Templated Organic Synthesis; Diederich,
F., Stang, P. J., Eds.; Wiley-VCH: Weinheim, 1999; Chapter 10, pp 275-
395 and references therein.
(9) Reetz, M. T.; Jung, A.; Bolm, C. Tetrahedron 1988, 44, 3889-3898.
The choice of TiCl4 as a Lewis acid was important here for ensuring an
intramolecular allylation. A T. S. in which the Ti Lewis acid coordinates
to both the aldehyde and silyl ether oxygen was invoked for rationalizing
the stereochemical outcome of the reaction.
(11) See for example: (a) Hioki, H.; Okuda, M.; Miyagi, W.; Itoˆ, S.
Tetrahedron Lett. 1993, 34, 6131-6134. (b) Martin, O. R.; Rao, S. P.; Kurz,
K. G.; El-Shenawy, H. A. J. Am. Chem. Soc. 1988, 110, 8698-8700.
(12) Tamao, K.; Nakajo, E.; Ito, Y. Tetrahedron 1988, 44, 3997-4007.
(10) Reetz, M. T.; Jung, A. J. Am. Chem. Soc. 1983, 105, 4833-4835.
4232
Org. Lett., Vol. 5, No. 22, 2003

Scheme 3. Synthesis of Allylsilane 1, Tethering, and Initial
Table 1. An Intramolecular Allylation/Tamao Oxidation
Sequence Provides a Route to Stereodefined 1,2,4-Triolsa
Cyclization Studiesa
combined
yield
7 + 8b (%)
isolated
ratioc
yield
9d (%)
entry aldehyde
R
7:8
a Reaction conditions: (a) LiNEt2 (1.0 equiv), THF, 0 °C; (b)
(allyl)MgBr (1.0 equiv), Et2O, reflux; (c) nBuLi (1.1 equiv),
TMEDA (1.0 equiv), Et2O, -5 °C, then TMSCl (1.0 equiv), -5
°C to rt (19%, three steps); (d) NaBH4 (1.1 equiv), EtOH, 0 °C
(81%); (e) 1 (1.0 equiv), 2 (1.0 equiv), 40 °C (90%); (f) DIBALH
(1.1 equiv), CH2Cl2, -78 °C (93%); (g) Lewis acid (ref 13), CH2Cl2,
-78 °C or MeCN, -40 °C, 5 (>95% crude).
1
2
3
4
5
6
7
8
9
6a
6b
6c
6d
6e
6f
6g
6h
6i
Ph
2-furyl
88
86
82
92
83
76
55
75
tracee
3.9:1
4.0:1
3.5:1
6.7:1
4.9:1
8.5:1
2.7:1
9.7:1
73
61
85
34
51
65
60
63
cinnamyl
TIPS-CtC-
cyclohexyl
nBu
sBu
BnOCH2CH2-
Me
provided the corresponding aldehyde 4 ready for investigat-
ing the intramolecular allylation. Unfortunately however, all
attempts to effect cyclization using a range of Lewis and
Brønsted acids13 in CH2Cl2 and MeCN at a range of
temperatures only ever provided the desired oxasilacyclic
products in trace amounts; in every case dienes 5 were
isolated as the major products and as single (E)-stereoisomers
(Scheme 3).
a Reaction conditions: (a) TMSOTf (1.0 equiv), 2,6-DTBMP (1.2 equiv),
CH2Cl2, -78 °C; (b) H2O2 (60% aq) (20 equiv), KF (5 equiv), KHCO3 (3
equiv), THF/MeOH (1:1). b Diene products accounted for the remaining
material in the crude reaction mixture. c Ratio calculated from analysis of
the crude reaction mixture by 1H NMR. d Isolated yields following column
chromatography. e (E)-Dienes were the major products from this reaction.
It is not clear at the present time why this substrate does not provide the
desired oxasilacycles.
We reasoned that our problem was associated with
preferential collapse of the silyl ether in the carbocationic
intermediate14 and that, by increasing the steric bulk of the
ligands around the silyl ether tether, we would redirect the
course of the reaction to the desired products. We were
therefore delighted to observe that allylsilane 6a15 containing
Et substituents at the silyl tether efficiently suppressed the
formation of diene products and provided the desired
cyclization products 7a and 8a in very good yield (Table
1). Similar reactions were carried out on a range of
aldehydes. The results are summarized in Table 1.
As already discussed, relocating the silyl ether to a position
where it is retained in the initial allylation product, generates
a product ripe for further elaboration. In a first investigation
we chose to oxidatively cleave the Si-C bond in our
allylation products using a Tamao oxidation16 to provide a
route into stereodefined 1,2,4-triols. Treatment of the crude
reaction mixture from the allylation to the standard Tamao
oxidation conditions effected a relatively slow cleavage of
the tether to give the desired triol products 9 in moderate to
good yield (Table 1). Interestingly, under the mildly basic
reaction conditions of the oxidation process, the minor
diastereoisomer 8 was found to be particularly susceptible
to elimination (to provide (E)-dienes)17 and only small
quantities of the minor triol product 10 were identified. It is
unclear at the present time why the two diastereoisomeric
allylation products behave so differently; however the result
fortuitously allowed the isolation of essentially diastereo-
isomerically pure triol product 9.
In all cases, only two of the four possible diastereoisomers
were observed by analysis of the crude reaction mixture by
HPLC and 1H NMR. A series of nOe experiments was used
to elucidate the structure of the minor diastereoisomer 8
(Figure 1). However, we were unable to use similar
procedures for confirming the relative stereochemistry in the
major diastereoisomer 7. Fortunately since the Tamao
oxidation is a stereospecific process that proceeds with
retention of configuration, we were able to identify the
relative stereochemistry of the major diastereoisomer indi-
rectly using the triol oxidation products. Exposure of triol
9a to acetone in the presence of pTSA and Na2SO4 led to
the formation of the two acetonides 11a and 12a with the
(13) The following activators were tried: TiCl4, SnCl4, BF3‚OEt2,
Yb(OTf)3, MeAlCl2, TMSOTf with/without 2,6-DTBMP, TBDMSOTf with/
without 2,6-DTBMP, Cu(OTf)2, ZrCl4, MgBr2‚OEt2, TfOH.
(14) Diene 5 may be formed in a variety of ways. We currently favor a
mechanism involving a vinylogous silicon-mediated olefination as this best
accounts for the excellent (E)-stereoselectivity observed. (a) Stragies, R.;
Blechert, S. Tetrahedron 1999, 55, 8179-8188. (b) Bradley, G. W.; Thomas,
E. J. Synlett 1997, 629-631. (c) Angoh, A. G.; Clive, D. L. J. J. Chem.
Soc., Chem. Commun. 1984, 534-536.
(15) 6a was prepared according to Scheme 3 starting from Et2SiCl2 in
place of Me2SiCl2.
(16) For a detailed review on this reaction: Fleming, I. Chemtracts: Org.
Chem. 1996, 9, 1-64.
(17) Ager, D. J. Org. React. 1990, 38, 1-223 and references therein.
Org. Lett., Vol. 5, No. 22, 2003
4233

were ever observed. We believe this difference is a conse-
quence of the C-O bond and its associated dipole affecting
the relative energies of the reacting T. S.s, a controlling factor
that is absent in the substrates employed by Keck. Thus,
levels of 1,3-stereoinduction are excellent - there is a clear
preference for the aldehyde to adopt a pseudoaxial orientation
in the reacting chairlike T. S. (Figure 2). Levels of 1,4-
Figure 1. Selected nOes and δC resonances used to elucidate the
relative stereochemistry of the two diastereoisomers.
Figure 2. Transition state leading to the major diastereoisomer 7.
The transition state has the aldehyde and allylsilane adopting
pseudoaxial orientations.
dioxolane predominating as expected (Figure 1). NMR
spectroscopy (13C NMR18 and nOe experiments) on this
mixture readily allowed the identification of the relative
stereochemistry in these products and therefore in the major
allylation diastereoisomer 7 (Figure 1).
It is worthwhile to compare the diastereoselectivity of this
set of reactions with those from a similar study carried out
by Keck.19 The results are rather different. In Keck’s very
detailed study employing an allylstannane linked through an
all-carbon chain to the reacting aldehyde, the diastereo-
selectivity of the reaction with the analogous (E)-allyl-
stannane was heavily dependent on the Lewis acid employed,
although generally, all four possible diastereoisomers were
observed in varying amounts.19 In our system, only two out
of the possible four diastereoisomeric allylation products
stereoinduction, however, are more modest with the major
diastereoisomer resulting from the allylsilane also preferen-
tially adopting a pseudoaxial orientation in the reacting
T. S. (Figure 2). We believe that this is a consequence of
the allylsilane minimizing steric interactions with the ethyl
substituents at the silyl tether. We expect that the (Z)-
stereoisomeric allylsilane will better differentiate the two T.
S.s leading to 7 and 8 and improve the levels of 1,4-
induction. Efforts are currently underway to synthesize this
novel allylsilane.
Acknowledgment. We thank the Engineering and Physi-
cal Sciences Research Council (Grant No. GR/N32556) for
a studentship (to J.B.).
Supporting Information Available: Experimental pro-
cedures and full characterization data for allylation products
7 and 8, triols 9, and acetonides 11 and 12. This material is
available free of charge via the Internet at http://pubs.acs.org.
(18) For elucidating the 1,3-relative stereochemistry: Rychnovsky, S.
D.; Rogers, B.; Yang, G. J. Org. Chem. 1993, 58, 3511-3515; and the
1,2-relative stereochemistry: Dana, G.; Danechpajouh, H. Bull. Soc. Chim.
Fr. 1980, 2, 395-399.
(19) Keck, G. E.; Dougherty, S. M.; Savin, K. A. J. Am. Chem. Soc.
1995, 117, 6210-6223.
OL035762O
4234
Org. Lett., Vol. 5, No. 22, 2003