A novel synthesis of functionalized allylsilanes

Article abstract of DOI:10.1021/ol049311v

Matrix presented. A novel facile synthesis of substituted and functionalized allylsilane has been developed. This essentially one-pot procedure involves (1) (dimethylphenylsilyl)methyl cerium chloride addition to cyclopropyl ketone and (2) MgI₂

Full text of DOI:10.1021/ol049311v

ORGANIC
LETTERS
2004
Vol. 6, No. 11
1849-1852
A Novel Synthesis of Functionalized
Allylsilanes†
,‡
Wei-Dong Z. Li* and Jin-Hui Yang
State Key Laboratory of Applied Organic Chemistry, Lanzhou UniVersity,
Lanzhou 730000, China
liwd@lzu.edu.cn
Received April 14, 2004
ABSTRACT
A novel facile synthesis of substituted and functionalized allylsilane has been developed. This essentially one-pot procedure involves (1)
(dimethylphenylsilyl)methyl cerium chloride addition to cyclopropyl ketone and (2) MgI2 etherate-mediated Julia homoallylic transposition of
the corresponding cyclopropyl carbinol. The bifunctional homoiodo allylsilanes, readily accessible by this method, would be useful and versatile
synthons in organic synthesis.
Allylsilanes are valuable and extremely versatile intermedi-
ates in synthetic organic chemistry for a variety of stereo-
controlled C-C bond formation, carbocyclic ring-forming
reactions.1 Numerous methods have thus been developed for
the synthesis of allylsilanes of various types.2 In connection
with an ongoing project on the total synthesis of Labdane
diterpenoids based on a well-established general biomimetic
polyene cationic cyclization strategy (Figure 1),3 we disclose
here a novel and efficient synthetic method for the synthesis
of highly substituted and functionalized allylsilanes such as
epoxy allylsilane 1, the projected biomimetic cyclization
precursor.
In an attempt to utilize the Julia homoallylic transposition
protocol (Julia olefination synthesis)4 for the synthesis of a
trisubstituted allylsilane model compound 5, we devised a
synthetic plan as shown in Scheme 1, beginning with
cyclopropyl carbinol 3, which was prepared from homoge-
ranyl cyclopropyl ketone 25 by carbonyl addition of a
(trimethylsilyl)methyl cerium reagent derived from the
† Dedicated to Prof. Marc Julia.
‡ Current address: Department of Chemistry and Chemical Biology,
Harvard University, 12 Oxford Street, Cambridge, MA 02138. E-mail:
wdli@fas.harvard.edu.
(1) For a review, see: Fleming, I.; Barbero, A.; Walter, D. Chem. ReV.
1997, 97, 2063.
(2) For a review on allylsilanes, see: Sarkar, T. K. In Science of
Synthesis: Houben-Weyl Methods of Molecular Transformations; Fleming,
I., Ed.; Georg Thieme Verlag: 2002, New York; Vol. 4, pp 837-962. See
also: Sarkar, T. K. Synthesis 1990, 969 and 1101.
(3) For general reviews, see: (a) Johnson, W. S. Bioorg. Chem. 1976,
5, 51. (b) Hughes, L. R.; Schmid, R.; Johnson, W. S. Bioorg. Chem. 1979,
8, 513. (c) Coates, R. M.; Conradi, R. A.; Ley, D. A.; Akeson, A.; Harada,
J.; Lee, S.-C.; West, C. A. J. Am. Chem. Soc. 1976, 98, 4659. For recent
examples, see: (d) Ravn, M. M.; Coates, R. M.; Flory, J. E.; Peters, R. J.;
Croteau, R. Org. Lett. 2000, 2, 573. (e) Aggarwal, V. K.; Bethel, P. A.;
Giles, R. Chem. Commun. 1999, 325. (f) Kashibuchi, Y.; Fukumoto, T.;
Oguchi, M.; Hirukawa, T.; Kato, T. Heterocycles 1998, 49, 255. (g) Coates,
R. M.; Yee, N. K. N. J. Org. Chem. 1992, 57, 4598. (h) Jognson, W. S.;
Lindell, S. D.; Steele, J. J. Am. Chem. Soc. 1987, 109, 5852.
Figure 1. Biomimetic strategy for Labdane diterpenoids.
10.1021/ol049311v CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/07/2004

Scheme 1
Scheme 2
as a Lewis acidic mediator4e,f allowed the Julia ring-opening
homoallylation of 6 to be examined.
After considerable experimentation, we were delighted to
find that treatment of 6 with MgI2 etherate7 in diethyl ether
furnished the desired Julia-type ring-opening product 7 in
78% yield as a mixture of (E,Z)-isomers (ca. 1:1), without
significant formation (<5%) of Peterson elimination product
4 (Scheme 2). Although commercially available powdered
MgI2 could also be used for this reaction, the reaction is
found to be slower and contaminated by a significant amount
of elimination product 4. Other metal halides, i.e., LiX, ZnX2
(X ) Cl, Br, I), or MgBr2, all led to rapid production of
elimination product 4. It is especially worth noting that the
MgI2 etherate is uniquely effective for the homoallylic
transposition of (dimethylphenylsilyl)methyl cyclopropyl
carbinol 6.
corresponding magnesium chloride reagent and anhydrous
CeCl3. However, the labile keto adduct 3 underwent rapid
Peterson-type elimination6 upon exposure to silica gel to give
cyclopropyl alkene 4 exclusively during the attempted
chromatographic purification. Although the crude carbonyl
adduct 3 is relatively stable for structural characterization
(spectroscopic), all efforts4 for the Julia-type olefination by
nucleophilic ring-opening of cyclopropane leading to the
expected trimethylallylsilane 5 resulted solely in the forma-
tion of elimination product 4 instead.
We reasoned that the following factors might account for
the success of this substrate-reagent combination: (1) the
steric bulkiness8 of dimethylphenylsilyl grouping deters the
attack of anionic nucleophile (here an iodo anion) on the
silicon center, thus diminishing the Peterson elimination (path
a); (2) the enhanced â-effect of silicon8 due to the more
electron-rich phenyl substituent, leading to a more stable
cationic intermediate i (Scheme 3), which might facilitate
the nucleophilic ring-opening of the cyclopropane by iodo
anion attack of a pseudo SN2′ type (path b); and (3) the
softness of the more dissociated iodo anion of the mild Lewis
acid MgI2 etherate tends to favor the nucleophilic attack on
the carbon atom (sp2 character) of the cyclopropane ring
(Julia-type ring-opening pathway). The delicate reactivity
alternation demonstrated here again the uniqueness of the
MgI2 etherate as a mild Lewis acid as well as effective source
of iodide ion as a cyclopropane ring-opening nucleophile.4e,f
The twist of the usually facile Peterson elimination pathway
Attempted conversion of cyclopropyl alkene 4 to 5 by the
action of trimethylsilyl halides (TMSX, X ) Cl, I) also failed
by affording homoallylic halides 5′ (X ) Cl, I) as the major
isolable (E)-isomeric product (eq 2), via presumably a protic
acid-mediated Julia-type process. In hoping that the sterically
more bulky silyl substituents might diminish the competitive
Peterson elimination pathway and provide us a chance to
effect the designated Julia homoallylic transposition, the
cyclopropyl keto adduct 6 was prepared analogously from
the (dimethylphenylsilyl)methyl cerium reagent in good
yield, which is indeed more stable than the corresponding
adduct 3, as evidenced by ready chromatographic purification
on silica gel (Scheme 2). Employing various metal halides
(4) For earlier seminal works, see: (a) Julia, M.; Julia, S.; Gue´gan, R.;
Bull. Soc. Chim. Fr. 1960, 1072. (b) Julia, M.; Julia, S.; Tchen, S.-Y.;
Neuville, C. Bull. Soc. Chim. Fr. 1960, 1381. (c) Julia, M.; Julia, S.; Du
Chaffaut, J. M. Bull. Soc. Chim. Fr. 1960, 1735. (d) Julia, M.; Julia, S.;
Tchen, S.-Y. Bull. Soc. Chim. Fr. 1961, 1849. For later modifications, see:
(e) Brady, S. F.; Ilton, M. A.; Johnson, W. S. J. Am. Chem. Soc. 1968, 90,
2882. (f) McCormick, J. P.; Barton, D. L. J. Org. Chem. 1980, 45, 2566.
(5) Prepared from geranyl iodide by alkylation with cyclopropyl methyl
ketone (LDA, -78 °C, THF) in 52% yield.
(7) Magnesium(II) iodide and its etherate have emerged as unique Lewis
acidic catalysts in a number of synthetic transformations and organic reaction
catalysis; for recent examples, see: (a) Lautens, M.; Han, W.; Liu, J. H.-C.
J. Am. Chem. Soc. 2003, 125, 4028. (b) Lautens, M.; Han, W. J. Am. Chem.
Soc. 2002, 124, 6312. (c) Li, W.-D. Z.; Zhang, X.-X. Org. Lett. 2002, 4,
3485 and references therein.
(6) For a review on â-silyl alcohols and the Peterson reaction, see: Ager,
D. J. In Science of Synthesis: Houben-Weyl Methods of Molecular
Transformations; Fleming, I., Ed.; Georg Thieme Verlag: 2002, New York;
Vol. 4, pp 789-809.
(8) We are currently investigating these factors by incorporating other
trisubstituted silyl groupings in â-silyl cyclopropyl carbinol adducts, for
example, triisopropylsilyl, triethylsilyl (as suggested by a reviewer), and
(para-methoxyphenyl)dimethylsilyl, etc.
1850
Org. Lett., Vol. 6, No. 11, 2004

derived cyclopropyl ketone 10 was similarly converted into
cyclopentenyl allylsilane 11 and ring-expended cyclohexenyl
allylsilane 12 in a ratio of 1.2:1 (Scheme 5).10 Presumably,
Scheme 3
Scheme 5
of â-silyl cyclopropyl carbinol of type 6 toward the Julia
homoallylic transposition is a result of the synergetic effect
of bulky and electron-rich silyl substituent and the strained,
yet cation-stablizing cyclopropane ring system. The lower
stereoselectivity (E vs Z) may be attributed to the relatively
even steric interaction of two conformers (A and B, eq 2) of
the cationic intermediate i.
cyclic allylsilanes 11 and 12 might have resulted from the
corresponding diastereomeric keto adducts 11a and 12a,
respectively,11 where ring-opening of cyclopropane by an
iodo anion is assumed to follow a trans anti-periplanar
conformational alignment,12 or from the major adduct 11a
by iodo anion attack on two cyclopropane carbons4f as shown
(Scheme 5). It is interesting to compare with the Magnus
synthesis13 of endo-cyclic allylsilanes14 by ring-opening
olefin-transposition of silylcyclopropane carbinol derivatives,
in which the seven-membered cyclic allylsilanes were formed
exclusively (eq 3) regardless of the stereochemistry of the
initial silylcyclopropane carbinol. These bifunctional
homoiodo allylsilanes (i.e., 9 and 11) would be potentially
useful synthons in natural product synthesis via tandem C-C
bond formation (formally an annulation).15
To extend the scope of this approach, some bicyclic
cyclopropyl ketone substrates were tested next. As shown
in Scheme 4, the (+)-carvone-derived cyclopropyl ketone
To demonstrate the current allylsilane synthesis method
in a more functionalized substrate for the aforementioned
general strategic plan (Figure 1), the synthesis of an epoxy
allylsilane cyclization precursor of type 1 was outlined in
Scheme 6. Allylic alcohol 13, available from 2 by allylic
oxidation (SeO2-TBHP), was epoxidized (Sharpless epoxi-
dation with L-(+)-DET as a chiral ligand) and followed by
O-benzylation to afford the epoxy ketone 14 (ca. 50% from
13). Upon addition of (dimethylphenylsilyl)methyl cerium
chloride to epoxy ketone 14 at 0 °C in THF, the desired
carbonyl adduct 15 was produced after neutral extractive
workup, which without further purification was taken in
anhydrous diethyl ether and treated with a freshly prepared
solution of MgI2 etherate (ca. 0.25 M, 1:1 diethyl ether-
benzene) at 0 °C for 10 min to give the iodohydrin allylsilane
Scheme 4
89 was transformed into the optically pure bifunctional cyclic
allylsilane 9 in a convenient one-pot procedure in good yield
as the sole isolable product. Although the corresponding
cyclopropyl carbinol intermediate was not isolated, it is safe
to suggest its structure as 9a. 2-Methyl cyclopentenone-
(10) Although inseparable on chromatographic silica gel, they were
1
distinctly identified by H NMR.
(11) A pair of diastereomers was spotted on TLC and by 1H NMR
analysis of the crude adducts.
(12) Cf.: Sakaguchi, K.; Fujita, M.; Ohfune, Y. Tetrahedron Lett. 1998,
39, 4313.
(13) (a) Cooke, F.; Magnus, P. J. Chem. Soc., Chem. Commun. 1978,
714. (b) Magnus, P.; Cooke, F.; Sarkar, T. Organometallics 1982, 1, 562.
(14) For an alternative method, see: Kende, A. S.; Hebeisen, P.;
Newbold, R. C. J. Am. Chem. Soc. 1988, 110, 3315.
(15) Studies along this line are ongoing in this laboratory; cf. eq 4.
(9) Cf.: (a) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87,
1353. (b) Chandrasekhar, S.; Narasihmulu, C.; Jagadeshwar, V.; Reddy,
K. V. Tetrahedron Lett. 2003, 44, 3629.
Org. Lett., Vol. 6, No. 11, 2004
1851

of Labdane diterpenoids in general, and studies on the
synthesis of andrographolide,18 a typical bioactive natural
product of this class, are in progress in our laboratory.
Scheme 6a
In summary, we have developed a novel approach for the
synthesis of highly substituted functionalized allylsilanes
based on the Julia homoallylic transposition protocol, which
features the combination of the incorporation of sterically
more hindered and more electron-rich dimethylphenylsilyl
grouping and the use of MgI2 etherate as a unique Lewis
acidic reagent. The method is operationally simple (es-
sentially a one-pot procedure),19 effective, and applicable to
bicyclic cyclopropyl ketones and substrates with sensitive
functionality (i.e., epoxide). Iodo allylsilanes such as 9 and
11 would be useful bifunctional synthons for the synthesis
of terpenoid natural products, as illustrated in eq 4 as an
example.20
Acknowledgment. We thank the National Natural Sci-
ence Foundation (Distinguished Youth Fund 29925204 and
Fund 20021001) and Ministry of National Education (Funds
99114 and 2000-66) for financial support. The Cheung Kong
Scholars program is gratefully acknowledged.
a Conditions: (a) SeO2, TBHP, CH2Cl2, rt, 2 h, 56%. (b) (i)
Ti(OiPr)4, L-(+)-DET, TBHP, CH2Cl2, -20 °C, 5 h, 94%; (ii) NaH,
BnBr, n-Bu4NI, THF, rt, 1.5 h, 53%. (c) ClMgCH2SiMe2Ph, CeCl3,
THF, 0 °C, 1 h. (d) MgI2‚(OEt2)n, Et2O, 0 °C, 15 min. (e) K2CO3,
MeOH, 0 °C, 30 min, 58% from 14. (f) BF3‚OEt2, CH2Cl2, -78
°C, 15 min, 61%.
Supporting Information Available: Experimental pro-
cedures and spectral data for compounds 2-4, 6, 7, 9, 11-
14, and 17-19. This material is available free of charge via
the Internet at http://pubs.acs.org.
16 after flash silica gel chromatography.16 Brief treatment
of the allylsilane 16 with K2CO3 in methanol at 0 °C gave
the epoxy allylsilane 17 as an inseparable mixture of
geometric isomers (Z/E ) 1.8:1) in 58% overall yield from
epoxy ketone 14. Exposure of the epoxy allylsilane 17
(mixture of isomers) to BF3 etherate in CH2Cl2 at -78 °C
for 15 min and quenching with saturated aqueous sodium
bicarbonate at -78 °C furnished the bicyclic product 18 and
19 (61%) as a mixture of two diastereomers (ca. 1:1),17 which
were separable by HPLC (hexane-2-propanol 40:1). The
structures of cyclized products 18 and 19 were fully
characterized by spectroscopic analysis. We believe this
approach would be applicable to the biomimetic synthesis
OL049311V
(18) For an earlier synthetic study, see: Pelletier, S. W.; Chappell, R.
L.; Prabhakar, S. J. Am. Chem. Soc. 1968, 90, 2889.
(19) Typical Procedure for the Preparation of Homoiodo Allysilane
from Cyclopropyl Ketone. Freshly dried (1 mmHg at 150 °C for 7 h from
the heptahydrate) CeCl3 (296 mg, 1.2 mmol) was suspended in 10 mL of
anhydrous THF, stirred vigorously for 2 h at ambient temperature, and
cooled to 0 °C by an ice bath, to which a stock solution of (dimethyl-
phenylsilyl)methylmagnesium chloride in diethyl ether (1.0 M, 1.2 mL, 1.2
mmol) was added dropwise at the same temperature. After the mixture was
stirred for 1 h at 0 °C, cyclopropyl ketone (1.0 mmol) in 1 mL of THF was
added dropwise and the resulting mixture was warmed gradually to room
temperature. When the consumption of starting ketone was complete
(monitored by TLC), the reaction mixture was cooled to 0 °C and quenched
with water (1 mL). The organic layer was separated, and the aqueous phase
was extracted with diethyl ether. The organic layers were washed with water
and brine and dried (MgSO4). The solvent was evaporated carefully in vacuo
at 0 °C to give the cyclopropyl carbinol adduct as a light yellowish oil.
The crude cyclopropyl carbinol was dried azeotropically with benzene, taken
in anhydrous diethyl ether (10 mL) under a nitrogen atmosphere, and stirred
at 0 °C, to which a freshly prepared7c MgI2 etherate (1.1 mmol, 0.25 M)
solution mixture in Et2O-benzene (1:1) was added dropwise. The resulting
mixture was stirred for 10-15 min at 0 °C, quenched with saturated
NaHCO3, and extracted with diethyl ether. The organic extracts were washed
with 10% sodiun thiosulfate solution, water, and brine and dried (MgSO4).
The solvent was evaporated in vacuo to give an oily residue, which was
purified by flash silica gel chromatography eluting with a mixture of diethyl
ether-petroleum ether (bp 30-60 °C) to afford the homoiodo allysilane
product as a colorless oil.
(16) This product was characterized by 1H NMR and IR spectral analysis.
Although MgI2 was reported as the reagent of choice for iodohydrin
synthesis via epoxide ring-opening (cf.: (a) Bonini, C.; Righi, G.; Sotgiu,
G. J. Org. Chem. 1991, 56, 6206. (b) Coutrot, P.; Legris, C. Synthesis 1975,
118) and deoxygenation of epoxide (cf.: Chowdhury, P. K. J. Chem. Res.,
Synop. 1990, 192), we assumed that the species, i.e., Mg(OH)I, generated
from the Julia homoallylic transposition process may serve as the iodo anion
source for the production of iodohydrin 16, in view of the fact that 1 equiv
of MgI2 etherate was used for this transformation.
(17) It is noteworthy that the diastereomeric ratio of the cyclization
product 18 and 19 is scrambled in regards to the geometric ratio of epoxy
allylsilane precursor 17, which that implies one of the cyclization conforma-
tions (chair-chair vs chair-boat) predominates slightly over the other. Further
work will be needed to address this interesting mechanistic aspect through
the preparation (or separation) of pure isomeric precursor 17.
(20) Cf.: (a) Majetich, G.; Defauw, J. Tetrahedron 1988, 44, 3833. (b)
Wang, J.-C.; Krische, M. J. Angew. Chem., Int. Ed. 2003, 42, 5855.
1852
Org. Lett., Vol. 6, No. 11, 2004