Investigations of α-siloxy-epoxide ring expansions forming 1-azaspirocyclic ketones

Article abstract of DOI:10.1021/jo049356+

The construction of 1-azaspirocyclic cycloalkanones using a siloxy-epoxide semipinacol ring expansion process was examined. Functionalized l-azaspiro[5.5]undecan-7-ones (1-azaspirocyclic cyclohexanones) proceeded in high chemical yields with complete diastereoselectivity using titanium tetrachloride as the Lewis acid promoter. The formation of functionalized 6-azaspiro[5.4]-decan-1-ones (1-azaspirocyclic cyclopentanones) proceeded in high chemical yield with little diastereoselectivity. Modification of reaction parameters such as the Lewis acid promoter or the nature of the silyl ether allowed for the preferential formation of either ("anti" or "syn" 1,2 alkyl shift) diastereomeric product. An explanation for the different reactivity profiles between the cyclobutanol silyl ethers and cyclopentanol silyl ethers is provided.

Full text of DOI:10.1021/jo049356+

In vestiga tion s of r-Siloxy-Ep oxid e Rin g Exp a n sion s F or m in g
1-Aza sp ir ocyclic Keton es
Gregory R. Dake,* Michae¨l D. B. Fenster, Me´lissa Fleury, and Brian O. Patrick†
Department of Chemistry, 2036 Main Mall, University of British Columbia,
Vancouver, British Columbia, Canada V6T 1Z1
gdake@chem.ubc.ca
Received April 16, 2004
The construction of 1-azaspirocyclic cycloalkanones using a siloxy-epoxide semipinacol ring
expansion process was examined. Functionalized 1-azaspiro[5.5]undecan-7-ones (1-azaspirocyclic
cyclohexanones) proceeded in high chemical yields with complete diastereoselectivity using titanium
tetrachloride as the Lewis acid promoter. The formation of functionalized 6-azaspiro[5.4]-decan-
1-ones (1-azaspirocyclic cyclopentanones) proceeded in high chemical yield with little diastereo-
selectivity. Modification of reaction parameters such as the Lewis acid promoter or the nature of
the silyl ether allowed for the preferential formation of either (“anti” or “syn” 1,2 alkyl shift)
diastereomeric product. An explanation for the different reactivity profiles between the cyclobutanol
silyl ethers and cyclopentanol silyl ethers is provided.
In tr od u ction
synthetic organic chemist.5-7 A useful representative
example of the power of this method is shown in eq 1.
Treatment of epoxide 1 with 1.1 equiv. of titanium
tetrachloride leads to the smooth formation of substituted
cyclohexanone 2 in 89% yield.2a A noteworthy feature of
this process is the facile formation of the spiro-ring
junction with complete diastereocontrol, which is pre-
The stereocontrolled formation of carbon atoms with
four different substituents (quaternary carbons) contin-
ues to be a significant problem in synthetic organic
chemistry.1 Stereoselective 1,2-rearrangement reactions
of epoxides and their derivatives have proven to be a
useful solution for this difficult challenge. Several re-
search groups have contributed to the understanding of
these processes so that a number of procedures for these
rearrangements, including the Tsuchihashi-Suzuki,2 the
Yamamoto,3 or the J ung4 variants, are available to the
(4) (a) J ung, M. E.; D’Amico, D. C. J . Am. Chem. Soc. 1993, 115,
12208. (b) J ung, M. E.; D’Amico, D. C. J . Am. Chem. Soc. 1995, 117,
7379. (c) J ung, M. E.; D’Amico, D. C. J . Am. Chem. Soc. 1997, 119,
12150. (d) J ung, M. E.; Marquez, R. Org. Lett. 2000, 2, 1669. (e) J ung,
M. E.; Lee, W. S.; Sun, D. Org. Lett. 1999, 1, 307. (f) J ung, M. E.;
Hoffmann, B.; Rausch, B.; Contreras, J .-M. Org. Lett. 2003, 5, 3159.
(g) J ung, M. E.; van den Heuvel, A. Org. Lett. 2003, 5, 4705. (h) J ung,
M. E.; van den Heuvel, A.; Leach, A. G.; Houk, K. N. Org. Lett. 2003,
5, 3375.
(5) (a) Marson, C. M.; Walker, A. J .; Pickering, J .; Hobson, A. D.;
Wrigglesworth, R.; Edge, S. J . J . Org. Chem. 1993, 58, 5944. (b)
Marson, C. M.; Harper, S.; Walker, A. J .; Pickering, J .; Campbell, J .;
Wrigglesworth, R.; Edge, S. J . Tetrahedron 1993, 49, 10339. (c) Tu, Y.
Q.; Sun, L. D.; Wang, P. Z. J . Org. Chem. 1999, 64, 629. (d) Fan, C.-
A.; Wang, B.-M.; Tu, Y. Q.; Song, Z.-L. Angew. Chem., Int. Ed. 2001,
40, 3877. (e) Baldwin, S. W.; Chen, P.; Nikolic, N.; Westheimer, D. C.
Org. Lett. 2000, 2, 1193. (f) Tu, Y. Q.; Fan, C. A.; Ren, S. K.; Chan, A.
S. C. J . Chem. Soc., Perkin Trans. 1 2000, 3791. (g) Pettersson, L.;
Frejd, T. J . Chem. Soc., Perkin Trans. 1 2001, 789. (h) Bickley, J . F.;
Hauer, B.; Pena, P. C. A.; Roberts, S. M.; Skidmore, J . J . Chem. Soc.,
Perkin Trans. 1 2001, 1253. (i) Wang, F.; Tu, Y. Q.; Fan, C. A.; Wang,
S. H.; Zhang, F. M. Tetrahedron: Asymmetry 2002, 13, 395. (j) Marson,
C. M.; Khan, A.; Porter, R. A.; Cobb, A. J . A. Tetrahedron Lett. 2002,
43, 6637. (k) Hauer, B.; Bickley, J . F.; Massue, J .; Pena, P. C. A.;
Roberts, S. M.; Skidmore, J . Can. J . Chem. 2002, 80, 546. (l) Marson,
C. M.; Oare, C. A.; McGregor, J .; Walsgrove, T.; Grinter, T. J .; Adams,
H. Tetrahedron Lett. 2003, 44, 141. (m) Fan, C. A.; Hu, X. D.; Tu, Y.
Q.; Wang, B. M.; Song, Z. L. Chem. Eur. J . 2003, 9, 4301. (n) Hu, X.
D.; Fan, C. A.; Zhang, F. M.; Tu, Y. Q. Angew. Chem., Int. Ed. 2004,
43, 1702.
† Manager, UBC X-ray Structural Facility; to whom inquiries
regarding crystal structures should be addressed.
(1) (a) Christoffers, J .; Mann, A. Angew. Chem., Int. Ed. 2001, 40,
4591. (b) Corey, E. J .; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998,
37, 388. (c) Fuji, K. Chem. Rev. 1993, 93, 2037. (d) Martin, S. F.
Tetrahedron 1980, 36, 419.
(2) (a) Maruoka, K.; Hasegawa, M.; Yamamoto, H.; Suzuki, K.;
Shimasaki, M.; Tsuchihashi, G. J . Am. Chem. Soc. 1986, 108, 3827.
(b) Suzuki, K.; Shimasaki, M.; Tsuchihashi, G. Tetrahedron Lett. 1986,
27, 6237. (c) Suzuki, K.; Matsumoto, T.; Tomooka, K.; Matsumoto, K.;
Tsuchihashi, G. Chem. Lett. 1987, 113. (d) Suzuki, K.; Miyazawa, M.;
Tsuchihashi, G. Tetrahedron Lett. 1987, 28, 3515. (e) Shimasaki, M.;
Hara, H.; Suzuki, K.; Tsuchihashi, G. Tetrahedron Lett. 1987, 28, 5891.
(f) Suzuki, K.; Miyazawa, M.; Shimasaki, M.; Tsuchihashi, G. Tetra-
hedron 1988, 44, 4061. (g) Suzuki, K. J . Synth. Org. Chem. J pn. 1988,
46, 49. (h) Shimasaki, M.; Hara, H.; Suzuki, K. Tetrahedron Lett. 1989,
30, 5443. (i) Shimasaki, M.; Morimoto, M.; Suzuki, K. Tetrahedron Lett.
1990, 31, 3335. (j) Nagasawa, T.; Taya, K.; Kitamura, M.; Suzuki, K.
J . Am. Chem. Soc. 1996, 118, 8949. See also: (k) Cheer, C. J .; J ohnson,
C. R. J . Am. Chem. Soc. 1968, 90, 178.
(3) (a) Maruoka, K.; Ooi, T.; Yamamoto, H. J . Am. Chem. Soc. 1989,
111, 6431. (b) Maruoka, K.; Nagahara, S.; Ooi, T.; Yamamoto, H.
Tetrahedron Lett. 1989, 30, 5607. (c) Maruoka, K.; Bureau, R.;
Yamamoto, H. Synlett 1991, 363. (d) Maruoka, K.; Ooi, T.; Nagahara,
S.; Yamamoto, H. Tetrahedron 1991, 47, 6983. (e) Maruoka, K.; Sato,
J .; Yamamoto, H. J . Am. Chem. Soc. 1991, 113, 5449. (f) Maruoka, K.;
Ooi, T.; Yamamoto, H. Tetrahedron 1992, 48, 3303. (g) Maruoka, K.;
Sato, J .; Yamamoto, H.; Tetrahedron 1992, 48, 3749. (h) Ooi, T.;
Maruoka, K.; Yamamoto, H. Org. Synth. 1995, 72, 95. (i) Nemoto, H.;
Ishibashi, H.; Nagamochi, M.; Fukumoto, K. J . Org. Chem. 1992, 57,
1707. (j) Nemoto, H.; Nagamochi, M.; Ishibashi, H.; Fukumoto, K. J .
Org. Chem. 1994, 59, 74.
(6) For reviews of R-hydroxy epoxide reactions, see: (a) Magnusson,
G. Org. Prep. Proc. Int. 1990, 22, 547. (b) Rickborn, B. In Comprehen-
sive Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon:
Oxford, U.K., 1991; Vol. 3, pp 733-775.
(7) R-Hydroxy N-p-toluenesulfonylaziridine rearrangement reactions
have also been performed: (a) Wang, B. M.; Song, Z. L.; Fan, C. A.;
Tu, Y. Q.; Shi, Y. Org. Lett. 2002, 4, 363. (b) Song, Z. L.; Wang, B. M.;
Tu, Y. Q.; Fan, C. A.; Zhang, S. Y. Org. Lett. 2003, 5, 2319.
10.1021/jo049356+ CCC: $27.50 © 2004 American Chemical Society
Published on Web 07/21/2004
5676
J . Org. Chem. 2004, 69, 5676-5683

1-Azaspirocyclics by Siloxy-Epoxide Expansion
SCHEME 1. An a lysis of F a sicu la r in
SCHEME 2
sumably dictated by the stereochemistry of the epoxide.8
As the facility to construct epoxides of one stereochemical
identity has increased substantially over the past twenty-
five years, the potential power of this protocol is signifi-
cant.9 The incorporation of semipinacol reactions in total
synthesis ventures is a testament to their growing
significance in the field.10
SCHEME 3. Gen er a l Syn th esis of Su bstr a tes
dihydropyran with several inductively withdrawing sub-
stituents was sufficiently reactive to induce expansion
of a cyclopentanol to a cyclohexanone.14 In addition, as
exemplified in eq 1, six-membered (and larger) rings are
easily produced using siloxy-epoxide rearrangement
reactions. Thus, it was anticipated that the epoxide
functionality could enable the desired reaction in two
respects. The inductive withdrawing oxygen substituent
resulting from the epoxide was anticipated to render the
intermediate more electrophilic, and thus more reactive
to 1,2-migration reactions. It was also hoped that the
hydroxyl (or siloxyl) functionality in the starting material
would be less prone to acid promoted dehydration, as it
would no longer be tertiary and allylic as in 3.
Our interest in siloxy-epoxide rearrangement pro-
cesses stemmed from our investigations of semipinacol
reaction-based approaches to 1-azaspirocycle ring sys-
tems present in alkaloids.11 Specifically, a proposed route
to fasicularin required a ring expansion of a cyclopentanol
to a cyclohexanone (Scheme 1).12 Unfortunately, the
reaction of cyclopentanol 3 with Bronsted acids did not
provide the desired ring expansion product but rather
enone 4 (eq 2).13 Through NMR studies, it was estab-
lished that 4 is produced via initial acid promoted
dehydration of the tertiary alcohol followed by hydrolytic
ring opening of the enesulfonamide moiety.
We were also aware of some points of concern (Scheme
2). The first was whether compounds such as a could be
made and isolated. A second equally important matter
was the predictability of the stereochemical outcome of
the reaction. Within carbocyclic systems at least, it
appeared that the stereochemistry of the epoxide function
effectively controlled this process (anti migration of the
alkyl group to the epoxide). In the proposed heterocyclic
system, the result was less obvious. Would a Lewis acid
induce a synchronous epoxide opening-1,2-alkyl migra-
tion reaction (path a) or would an azacarbenium ion b
be a significant intermediate (path b). In short, what
would be the relative stereochemistry between the alcohol
function and the spiro ring junction in the products? This
report presents our investigations into this matter.
Given this setback, we considered a siloxy-epoxide
variant of this proposed semipinacol reaction. It had been
demonstrated that an oxacarbenium ion derived from a
(8) A review on the formation of spirocyclic ring systems: Sanni-
grahi, M. Tetrahedron 1999, 55, 9007.
(9) (a) Katsuki, T.; Martin, V. S. Org. React. (N. Y.) 1996, 48, 1. (b)
Pfenninger, A. Synthesis 1986, 89. (c) Larrow, J . F., J acobsen, E. N.
Org. Synth. 1998, 75, 1. (d) Schaus, S. E,; Brandes, B. D.; Larrow, J .
F.; Tokunaga, M.; Hansen, K. B.; Gould, A. E.; Furrow, M. E.; J acobsen,
E. N. J . Am. Chem. Soc. 2002, 124, 1307. (e) Frohn, M.; Shi, Y.
Synthesis 2000 1979.
(10) (a) Saito, T.; Suzuki, T.; Morimoto, M.; Akiyama, C.; Ochiai,
T.; Takeuchi, K.; Matsumoto, T.; Suzuki, K. J . Am. Chem. Soc. 1998,
120, 11633. (b) Li, J .; J eong, S.; Esser, L.; Harran, P. G. Angew. Chem.,
Int. Ed. 2001, 40, 4765. (c) Eom, K. D.; Raman, J . V.; Kim, H.; Cha, J .
K. J . Am. Chem. Soc. 2003, 125, 5415. (d) Kita, Y.; Kitagaki, S.;
Yoshida, Y.; Mihara, S.; Fang, D.-F.; Kondo, M.; Okamoto, S.; Imai,
R.; Akai, S.; Fujioka, H. J . Org. Chem. 1997, 62, 4991.
(11) Fenster, M. D. B.; Patrick, B. O.; Dake, G. R. Org. Lett. 2001,
3, 2109.
Resu lts a n d Discu ssion
Syn th esis of Su bstr a tes. A general outline of the
synthetic route used to build the R-epoxy silyl ethers used
in this study is shown in Scheme 3. Functionalized
vinylstannanes of general structure 5 were reacted with
ketones to produce allylic alcohols of type 6.13,15 The
(12) (a) Fenster, M. D. B.; Dake, G. R. Org. Lett. 2003, 5, 4314. For
other syntheses of fasicularin, see: (b) Abe, H.; Aoyagi, S.; Kibayashi,
C. J . Am. Chem. Soc. 2000, 122, 4583. (c) Maeng, J .-H.; Funk, R. L.
Org. Lett. 2002, 4, 331.
(13) Dake, G. R.; Fenster, M. D. B.; Hurley, P. B.; Patrick, B. O. J .
Org. Chem., preceding paper in this issue.
(14) Paquette, L. A.; Kinney, M. A.; Dullweber, U. J . Org. Chem.
1997, 62, 1713.
J . Org. Chem, Vol. 69, No. 17, 2004 5677

Dake et al.
11 in 95% yield as one detectable diastereomer. The
connectivity and relative stereochemistry of 11 was
established using the X-ray crystallography of its p-
nitrobenzoate derivative.
Exp a n sion s of Cyclobu ta n ol Silyl Eth er s. In at-
tempting to define the scope of the method, ring expan-
sions of 12 and 13 were attempted. Cyclobutane rings
are known to undergo ring expansions,18 and thus
compounds 12 and 13 appeared to be minor modifications
of 8 and 10. Although 12 and 13 each undergo facile ring
expansions to cyclopentanones, it was somewhat surpris-
ing to discover that these reactions occurred with low
stereoselectivity (eq 4). Epoxide 12 produced ketones 14
and 15 in 95% yield as a 1.1:1 mixture of diastereomers.19
Cyclopentanone 14 is the apparent product of an anti-
periplanar 1,2-alkyl migration toward the epoxide, while
15 seemingly results via a synperiplanar 1,2-alkyl migra-
tion process. Similar results were obtained with 13, as
cyclopentanones 16 and 17 were produced as a 2.6:1
mixture in 96% yield. As 17 could have derived from 16
through a retro-aldol-aldol sequence, each of 16 and 17
were individually resubjected to titanium tetrachloride
in dichloromethane. Even at higher reaction tempera-
tures or prolonged reaction times, compounds 16 and 17
did not interconvert, suggesting that this ratio repre-
sented a “kinetic” mixture of the products of ring expan-
sion. X-ray analysis of crystals of 16 as well as a
derivative of 17 established each of their structures.
F IGURE 1.
optimal ordering and method of subsequent transforma-
tionssthe silylation of the tertiary alcohol and the
epoxidation16 of the enesulfonamidesto produce R-epoxy
silyl ethers of general structure 7 were best defined
through experimentation for each particular compound.
The synthetic details for each of the substrates described
in the article are given in the Supporting Information.
Exp a n sion s of Cyclop en ta n ol Silyl Eth er s. The
first R-siloxy epoxide that we examined was derived from
3, in part because of our keen interest in a synthetic route
toward fasicularin. Interestingly, the reaction of 3 (or its
corresponding trimethylsilyl ether) with dimethyldi-
oxirane (DMDO) produced a single epoxide. The diaste-
reoselectivity of this process can be rationalized by the
“axial” addition of the electrophilic oxygen on the most
stable half-chair conformer of 3 (Figure 1).
We were delighted to find that treatment of R-siloxy
epoxide 8 with 1.1 equiv of titanium tetrachloride in
dichloromethane at -78 °C for 30 min furnished a new
compound 9 in 96% isolated yield (eq 3). The trimethyl-
silyl ether in 8 was required for good efficiency as the
ring expansion reaction of the analogous 3° alcohol
produced 9 in only 30% yield.17 Signals in the IR and 13C
NMR spectra (ν ) 1718 cm-1 13C NMR: δ 209.9 ppm) of
;
9 clearly established the presence of a cyclohexanone
function. The structural connectivity and relative ster-
eochemistry of 9 was ultimately established using X-ray
analysis.
Superficially it appears that the cyclobutane-derived
substrates 12 and 13 reacted through an azacarbenium
ion intermediate while the cyclopentane-derived com-
pounds 8 and 10 underwent a synchronous epoxide
opening-ring expansion process. These counterintuitive
results prompted further investigation. Three mechanis-
tic options were formulated in an attempt to interpret
this phenomenon (Scheme 4).20 It was hoped that further
experiments could be used to distinguish between these
options.
In the case of the cyclobutane-containing substrates,
the first explanation invokes the formation of an azac-
arbenium ion intermediate before an unselective ring
expansion (option 1). This explanation requires that 8
and 12 (or 10 and 13) undergo ring expansions through
different mechanistic paths. Put another way, why is the
reaction of 8 (or 10) selective while the reaction of 12 (or
That 9 was formed as a single diastereomer suggested
that the titanium(IV) induced epoxide opening and 1,2-
alkyl migration occurred as a synchronous step. The
stereochemistry of the spiro center appeared to be
dictated by the epoxide. To test if the tert-butyldimeth-
ylsiloxy function in 9 conformationally anchored this
substrate as it reacted through an azacarbenium ion
intermediate, this reaction sequence was repeated on 10.
This compound lacks the tert-butyldimethylsiloxy func-
tion and so might be expected to form the spiro compound
with lower diastereoselectivity if an azacarbenium ion
was an important intermediate. Epoxide 10, when sub-
jected to the identical reaction conditions as those used
for 8, similarly generated azaspirocyclic cyclohexanone
(18) For examples, see: (a) Creary, X.; Inocencio, P. A.; Underiner,
T. L.; Kostromin, R. J . Org. Chem. 1985, 50, 1932. (b) Hirst, G. C.;
J ohnson, T. O., J r.; Overman, L. E. J . Am. Chem. Soc. 1993, 115, 2992.
(c) Trost, B. M.; Chen, D. W. C. J . Am. Chem. Soc. 1996, 118, 12541.
(d) Larock, R. C.; Reddy, C. K. J . Org. Chem. 2002, 67, 2027. For
reviews of the chemistry of cyclobutane ring containing compounds:
(e) Lee-Ruff, E.; Mladenova, G. Chem. Rev. 2003, 103, 1449. (f)
Namyslo, J . C.; Kaufmann, D. E. Chem. Rev. 2003, 103, 1485.
(19) Interestingly, ring expansions promoted by N-bromosuccinimide
apparently produce only the “anti” 1,2-migration product. See ref 13.
(20) It is important to note that these options represent mechanistic
extremes. Both subtle variations from these extremes and the consid-
eration that the reactions could occur simultaneously through more
than one option complicate matters significantly.
(15) Luker, T.; Hiemstra, H.; Speckamp, W. N. J . Org. Chem. 1997,
62, 8131.
(16) For related epoxidations using DMDO: (a) Burgess, L. E.;
Gross, E. K. M.; J urka, J . Tetrahedron Lett. 1996, 37, 3255. (b)
Sugisaki, C. H.; Carroll, P. J .; Correia, C. R. D. Tetrahedron Lett. 1998,
39, 3413.
(17) At least three other uncharacterizable byproducts were also
obtained.
5678 J . Org. Chem., Vol. 69, No. 17, 2004

1-Azaspirocyclics by Siloxy-Epoxide Expansion
SCHEME 4
13) is nonselective? A (simplistic) consideration of ring
strain arguments would suggest that a cyclobutane ring
would expand more readily (or more likely through a
synchronous manner) than a cyclopentane ring. An
assumption for option 2 in Scheme 4 is that the migrating
σ bond of the cycloalkane undergoing expansion is aligned
with the breaking C-O bond (the σ*) of the epoxide.
Generally, antiperiplanar orientations are preferred for
1,2-migration reactions.21 Because the cyclopentane ring
is less reactive, the lowest energy transition state reflects
this preference for an antiperiplanar orientation between
the migrating and rupturing bonds. The complexation of
12 or 13 with titanium(IV) would generate an electron-
deficient site adjacent to the cyclobutane. The more
reactive cyclobutane ring could undergo expansion through
either the antiperiplanar or the synperiplanar orienta-
tions. A third option is, under Lewis acid promotion, in
addition to the aforementioned antiperiplanar 1,2-migra-
tion reaction, the oxygen atom of the silyl ether could
participate by opening the epoxide ring (an anti process).
If 1,2-alkyl migration of the cyclobutane ring took place
in an anti fashion through this intermediate, the net
result would be an overall syn 1,2-alkyl migration reac-
tion. Our efforts to sort through these options began.
Ca r b ocyclic Ver sion . The first scenario suggested
the intermediacy of an azacarbenium ion. Did the adja-
cent cyclobutane ring stabilize this intermediate such
that it became a factor in the reaction expansion reac-
tion?22 Because we could not find carbocyclic versions of
a R-epoxy silyl ether ring expansion involving a cyclo-
butane ring, the expansion of epoxide 18 was attempted
(eq 5).2a Interestingly, unlike 12 or 13, the ring expansion
of 18 (1.1 equiv of TiCl4, -78 °C, 30 min) produced a
single diastereomeric cyclopentanone 19 in 62% yield.
The lowered yield of this reaction probably reflects the
difficulty in recovering the volatile product and not the
efficiency of the semipinacol reaction (see below). The
relative stereochemistry of 19 was established using
X-ray crystallography. This reaction was repeated at
progressively higher temperatures (0 °C, 25 °C) in order
to examine if alternative diastereomers could be ob-
served. Other diastereomeric products were not observed
(TLC, HPLC, NMR) during these experiments. In the
carbocyclic system, the cyclobutane ring was not suf-
ficient to induce the formation of a carbenium ion
intermediate. Clearly, the sterics and electronic proper-
ties of the p-toluenesulfonylamido function in 12 or 13
affects the reactivity profile of the expansion process
considerably, and consequently option 1 in Scheme 4
cannot be ruled out.
Effect of Differ en t Lew is Acid s. At this point the
effect of the Lewis acid on the outcome of the reaction of
13 was examined (eq 6 and Table 1). In general, 1.1-1.4
equiv of a Lewis acid was used to promote a given
reaction. The temperatures presented in the table rep-
resent the minimum at which a reaction was observed
to take place. Warming of the reaction temperature was
occasionally required in order to force reactions to
completion. Reactions with triethylaluminum or titanium
tetraisopropoxide were quite sluggish even at room
temperature (entries 1 and 2). The lowered isolated yields
(21) (a) Deslongchamps, P. In Stereoelectronic Effects in Organic
Chemistry; Pergamon: New York, 1983; pp 182-190. (b) Goodman,
R. M.; Kishi, Y. J . Am. Chem. Soc. 1998, 120, 9392. (c) Crudden, C.
M.; Chen, A. C.; Calhoun, L. A. Angew. Chem., Int. Ed. 2000, 39, 2851.
(d) Sahasrabudhe, K.; Gracias, V.; Furness, K.; Smith, B. T.; Katz, C.
E.; Reddy, D. S.; Aube´, J . J . Am. Chem. Soc. 2003, 125, 7914.
(22) (a) Wilcox, C. F., J r.; Mesirov, M. J . Am. Chem. Soc. 1962, 84,
2757. (b) Hahn, R. C.; Corbin, T. F.; Shecter, H. J . Am. Chem. Soc.
1968, 90, 3404.
J . Org. Chem, Vol. 69, No. 17, 2004 5679

Dake et al.
migration in which the trimethylsilyl group has also been
transferred to the newly formed 2° alcohol.26 Although
magnesium bromide-diethyl etherate appeared to result
in the formation of 16 in a highly selective manner (10:
1) relative to 17, R,â-unsaturated ketone 21 also was
obtained in 24% yield. These byproducts were not ob-
served, even in trace amounts, using other Lewis acids.
The observation that “mild” Lewis acid promotion
results in a greater amount of “anti”-1,2-migration sug-
gests that an antiperiplanar arrangement is energetically
preferred. In fact, the most selective ring expansion that
was uncovered to form an azaspirocyclic cyclopentanone
was one that did not involve a Lewis acid promoter (eq
7). While the optimal ordering of epoxidation and tri-
alkylsilylation steps for substrate construction was being
defined, the allyl alcohol 22 was epoxidized using DMDO.
After workup the R-hydroxy epoxide was placed under
vacuum for 10 h. Examination of the synthetic material
at that point clearly demonstrated it to be a 13.2:1
mixture of 16 and 17.
TABLE 1. Lew is Acid Scr een for Rea ction s of Ep oxid e
13
entry Lewis acida T (°C) t (h)
yieldb (%)
ratio 16:17c
1
2
3
4
5
6
7
8
9
10
11
12
13
Al(Me)3
Ti(OiPr)4
TiCl4
ZnI2
ZnCl2
EtAlCl2
Et2AlCl
TMSOTf
ZrCl4
BF3Et2O
Y(OTf)3
MgBr2Et2O
Yb(OTf)3
-78-rt
-78-rt
-78
sluggish reacn
sluggish reacn
0.5 96
3.5 89
93
0.5 95
51
0.5 89
89
0.5 89
54
2.5 68
99
2.6:1
2.5:1
4.2:1
4.6:1
5.9:1
6.0:1
6.1:1
6.8:1
6.9:1
10:1
-78-0
0-rt
17
-78
-78-rt 12
-78
-78
1
-78
-15-0
7
0
-45-0
7
7.4:1
a
b
1.1-1.4 equiv of Lewis acid was used. Isolated yields (after
purification using column chromatography). c Ratios were deter-
mined by HPLC analysis of the product mixture.
SCHEME 5
Effect of Silicon P r otectin g Gr ou p . The steric bulk
of the trialkylsilyl protecting group in these substrates
was then modified to examine its effect on the reaction.
If the silyl ether actively participates as a “neighboring
group” as delineated in the third mechanistic suggestion
(Scheme 4, option 3), then modifying this group should
have a substantial impact on the diastereoselectivity of
the reaction. Specifically, because bulkier groups should
inhibit the “bridging” ability of the silyl ether, it was
anticipated that an increase in steric bulk of the trialkyl-
silyl unit would lead to a corresponding increase of anti
1,2-alkyl migration product.
In fact, the opposite effect was observed (eq 8 and Table
2). As the steric bulk of the trialkylsilyl ether increases
from trimethylsilyl (13) to triethylsilyl (23) to tert-
butyldimethylsilyl (24), 17 is formed in increasingly
larger proportions. The most dramatic example is in
entry 4. Surprisingly, treatment of the triisopropylsilyl
ether 25 with titanium tetrachloride led to the formation
of 17 as the exclusive diastereomer.27
of 16 and 17 in entries 7 and 11 were the result of
byproduct formation that was not observed using other
Lewis acid promoters (see below). Although a rationaliza-
tion of the stereochemical result for each entry is not
possible, “milder” Lewis acids such as magnesium bro-
mide etherate or ytterbium(III) trifluorosulfonate gave
greater selectivity compared to titanium(IV) tetrachlo-
ride.23,24 The most selective, synthetically useful result
is that in entry 13. The use of ytterbium(III) trifluoro-
sulfonate smoothly promoted the ring expansion of 13 to
produce 16 and 17 in 99% isolated yield as a 7.4:1 ratio
of diastereomers.25
Certain Lewis acids promoted the formation of byprod-
ucts as well as the desired ring expansion reaction
(Scheme 5). When diethylaluminum chloride was used,
a substantial amount (37%) of 20 was also isolated.
Compound 20 was the formal result of “syn” 1,2-alkyl
TABLE 2. Effect of Bu lk of Tr ia lk ylsilyl P r otectin g
Gr ou p
(23) The organization (and even the definition) of “strong” and “mild”
Lewis acids is still unclear at the present time. (a) Childs, R. F.;
Mulholland, D. L.; Nixon, A. Can. J . Chem. 1982, 60, 801. (b) Laszlo,
P.; Teston, M. J . Am. Chem. Soc. 1990, 112, 8750. (c) Kobayashi, S.;
Busujima, T.; Nagayama, S. Chem. Eur. J . 2000, 6, 3491, and
references therein. For a recent review discussing Lewis acids, see:
(d) Corma, A.; Garcia, H. Chem. Rev. 2003, 103, 4307. For earlier,
useful reviews, see: (e) J ensen, W. B. Chem. Rev. 1978, 78, 1. and (f)
Satchell, D. P. N.; Satchell, R. S. Chem. Rev. 1969, 69, 251.
(24) For a review of rare earth salts as catalysts in synthetic organic
chemistry, see: Kobayashi, S.; Sugiura, M.; Kitagawa, H.; Lam, W.
W. L. Chem. Rev. 2002, 102, 2227.
entry
compd
SiR3
SiMe3
% yielda
ratio 16:17b
1
2
3
4
13
23
24
25
96
99
89
88
2.6:1
2.5:1
1.3:1
0:1
SiEt3
Si(tBu)Me2
Si(iPr)3
a
Isolated yields (after purification using column chromatogra-
b
phy). Ratios were determined by HPLC analysis of the product
mixture.
5680 J . Org. Chem., Vol. 69, No. 17, 2004

1-Azaspirocyclics by Siloxy-Epoxide Expansion
SCHEME 6
To examine whether this “TIPS” effect was more
general, the triisopropyl silyl ether 26 was treated with
titanium tetrachloride (eq 9). This initiated a smooth ring
expansion process from which only 19 (anti migration)
was obtained in 93% yield. Unfortunately, attempts to
form the triisopropylsilyl ether analogues of either 8 or
10 have to this point been unsuccessful.
89% yields, respectively (Scheme 6).13 This result was
significant in our laboratories, as substitution at the
3-position was not tolerated in the established synthetic
sequence (cf. Scheme 3) used to build precursors for
semipinacol reactions. It was surmised that if the ene-
sulfonamide moiety could react chemoselectively with an
epoxidizing agent, the possibility of an alternative aza-
spirocycle-building epoxide-based rearrangement reaction
existed.
In the event, the reaction of 29 or 30 with purified
m-CPBA took place uneventfully to produce 31 and 32
in 60 and 82% yields, respectively. In each case a single
diastereomeric product was observed (GC and NMR
analysis). Fortunately, X-ray quality crystals of 32 were
obtained and established the relative stereochemistry of
the epoxide as depicted. The stereochemistry of 31 is
assigned by analogy. This stereochemical result is ex-
pected as the allyl group in 29 or 30 would be expected
to be pseudoaxial in order to minimize A1,3 strain with
the cycloalkylidene function on the heterocyclic ring,
causing the epoxidizing agent to approach the opposite
alkene face.28 Because of the structural similarity be-
tween 31 or 32 and the “bridging siloxy” intermediate in
option 3 shown in Scheme 4, we envisioned that the
reactions of this substrate would be a good test of this
mechanistic premise.
These epoxides were unstable to acid and promptly
underwent ring expansion reactions when treated with
1 M HCl. In particular, epoxide 31 reacted within 10 min
of stirring with hydrochloric acid to form azaspirocyclic
cyclopentanone 33 in 90% yield. An important feature of
the reaction is that the 1,2-migration reaction generates
a single diastereomeric product (established by X-ray
crystallography) in which the migrated C-C bond is on
the same face as the epoxide C-O bond. When treated
with acid, epoxide 32 formed an azaspirocyclic cyclohex-
anone. As in the case with 31, a single diastereomeric
product was obtained. Unfortunately, 34 exists as a low-
melting solid, and X-ray quality crystals could not be
obtained. At this time, based on similar spectroscopic
data, the stereochemical identity of 34 is assigned in
analogy to that of 33.
Because it appeared possible that one could manipulate
the diastereoselective outcome of the cyclobutanol silyl
ether to cyclopentanone semipinacol reaction by adjusting
the nature of the Lewis acid and the trialkylsilyl protect-
ing group, we examined this effect on 12 (eq 10). As
observed previously, the reaction of 12 with titanium
tetrachloride produces 14 and 15 in a 1.1:1 ratio, albeit
in 95% yield. Using ytterbium(III) triflate to promote the
semipinacol process led to a much higher ratio of 14
versus 15 (4.4:1) in 87% yield. This enables the isolation
of 14 in 71% yield. Conversely, reaction of 27 with
titanium tetrachloride generates 14 and 15 in 94% as a
1:6.2 ratio of diastereomers. Cyclopentanone 15 can be
isolated in 81% yield from this reaction. This may provide
a useful method to generate either diastereomeric cyclo-
pentanone in the context of total synthesis ventures.
3-Su bstitu ted Heter ocyclic Su bstr a tes. An epoxide
resembling the “bridging” siloxy intermediate in option
3 allowed us to consider the reactivity and selectivity of
this putative intermediate. Quite by accident, it was
discovered that the allyl alcohol functionality in 22 or
28 could be reacted smoothly with allyltrimethylsilane
in the presence of boron trifluoride-diethyl etherate to
produce the allylated compounds 29 and 30 in 84 and
The results of the epoxide rearrangement of 31 tend
to disfavor the third option in Scheme 4. It appears that
the stereocontrolling element in the rearrangement of 31
is not the epoxide, but rather the adjacent allyl substitu-
ent. The alkyl group undergoes migration on the face
opposite the allyl group, perhaps to minimize steric
interactions during the transition state. To the extent
that these results and option 3 in Scheme 4 can be
compared (e.g. an allyl group is not the same as an
(25) The nature of the displaced anion after Lewis acid complexation
is an additional consideration that could affect the diastereoselectivity
of the expansion, although a clear trend is still not observed. We thank
a reviewer for this suggestion.
(26) The stereochemistry was initially assigned by comparison of
the 13C chemical shifts of the spiro carbon and the methylene carbon
adjacent to the nitrogen atom to that of 16 and 17. This assignment
was confirmed by conversion of 20 to 17 (TBAF, THF, 82%).
(27) For a review of the chemistry of the triisopropylsilyl group:
Ru¨cker, C. Chem. Rev. 1995, 95, 1009.
(28) The allyl group is pseudoaxial in the solid-state structure of
29. For reviews on 1,3-allylic strain and its impact on acyclic confor-
mational space: (a) Hoffmann, R. W. Chem. Rev. 1989, 89, 1841. (b)
Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1992, 31, 1124. (c)
Hoffmann, R. W. Angew. Chem., Int. Ed. 2000, 39, 2054.
J . Org. Chem, Vol. 69, No. 17, 2004 5681

Dake et al.
situation could be occurring in these systems as well
(C-O breaking prior to C-C migration), but under those
circumstances the selectivity differences in the expan-
sions of the cyclobutane and cyclopentane rings still are
difficult to rationalize.
Ch ela tion ? Another conceivable explanation for the
anomalous selectivity profiles between the TMS ether 13
and the TIPS ether 25 is a chelation effect. Specifically,
if chelation of a Lewis acid between the silyl ether and
epoxide oxygen atoms takes place, the conformation of
the reactive intermediate would most resemble that of
A (Figure 2). Provided that a large amount of structural
reorganization does not occur prior to rearrangement,
semipinacol rearrangements proceeding through A should
provide anti 1,2-migration products.30 It is possible that
the trimethylsilyl ether could significantly chelate, but
the triisopropylsilyl ether, being substantially more
hindered, does not chelate.31 If chelation of the Lewis acid
between the epoxide and silyl ether oxygen plays a
significant role in these reactions, the selectivity differ-
ence in the expansions of the cyclobutanol silyl ethers
and the cyclopentanol silyl ethers is still not clear. It is
not obvious how a chelation effect could be invoked for
one set of substrates and not its homologues. Although
it cannot be stringently ruled out, an explanation of the
“TIPS effect” invoking chelation is disfavored.
La r ger Rin gs. Unfortunately, our efforts to extend
this system beyond the formation of cyclopentanones and
cyclohexanones have as yet been unsuccessful. Typical
results are shown in eq 11. Treatment of the silyl ether
35 with a Lewis acid promoter generates a mixture of
36-38. “Optimized” conditions for the formation of each
byproduct are given in eq 11. Other Lewis acids such as
titanium tetrachloride give only these three products; no
products of ring expansion are observed. As larger rings
have a reduced tendency to ring-expand, reactions that
involve the intermediacy of an azacarbenium ion (such
as in option 1; Scheme 4) become favored.
F IGURE 2.
oxygen atom-Lewis acid complex, or that an epoxide is
not identical to a “bridging” siloxy group) it could be
extrapolated that the putative “bridged” siloxy interme-
diate in Scheme 4 may react through similar pathways.
Thus, the putative “bridged” siloxy intermediate in
Scheme 4 would not dictate the stereochemical outcome
of the ring expansion process. That diastereoselectivity,
assuming that the reaction proceeded in a fashion similar
to that of 31, would in part be dictated by the adjacent
C-O-Lewis acid complex, forming the “anti” migration
product.
Taken together, these experimental results appear to
be most consistent with the second proposal in Scheme
4. The expansion in the absence of Lewis acid favors the
stereoelectronically preferred antiperiplanar orientation
of the migrating C-C σ bond and the epoxide C-O bond.
However, as progressively “stronger” Lewis acids are
used to promote the ring expansion reaction, the amount
of “syn” migration increases. This observation seems to
corroborate well with the notion that the cyclobutane
C-C bond will migrate readily as an electron-deficient
atom is formed at an adjacent position. The unusual
selectivity in the reactions of the triisopropylsilyl pro-
tected substrates are also consistent with this interpreta-
tion (Figure 2). Conformations that have a coplanar
relationship between a cyclobutane C-C bond that could
migrate and the C-O bond of the epoxide are shown. Two
of these conformations (A and B) would migrate in an
antiperiplanar fashion, while in conformations C and D
these two bonds are in a syn orientation. The 20 lowest
energy conformations of 13 (no Lewis acid is templated;
using the MM2* force field) resemble C or D.29 No
conformations resembling A or B were found. These
conformations are believed to be destabilized through
steric interactions between either the cyclobutane ring
(in A) or the trialkylsilyl ether (in B) and the aromatic
ring of the p-toluenesulfonyl group. The results of this
ring expansion reflect the conformational constraints
of the starting materials (in a non-Curtin-Hammett
scenario). A similar explanation was used to explain the
“syn” migration of alkyl groups in a J ung non-aldol-aldol
process.4h In the case examined by J ung and Houk, the
syn 1,2-migration was believed to derive through specific
conformations via a carbenium ion intermediate. This
Con clu sion s
The predictability and stereochemical outcome of R-si-
loxy-epoxide rearrangements within carbocyclic frame-
works do not appear to be significantly affected by
structural variations (size of “expanding” ring; nature
(cyclic or acyclic) of epoxide).2-5 In contrast, the reactions
that form 1-azaspirocyclic ketones through a R-siloxy-
epoxide rearrangement reaction in this present work can
be dramatically affected by structural changes. Although
(30) A reviewer points out that the use of the “nonchelating” Lewis
acid, BF3‚OEt2, provides the “anti” migration product as the major
product.
(31) Lewis acid promoted nucleophilic additions to R- or â-siloxy
aldehydes are not believed to proceed in a chelate-controlled manner.
(a) Keck, G. E.; Castellino, S. Tetrahedron Lett. 1987, 28, 281. (b) Chen,
X.; Hortelano, E. R.; Eliel, E. L. Frye, S. V. J . Am. Chem. Soc. 1992,
114, 1778. An exception: (c) Evans, D. A.; Allison, B. D.; Yang, M. G.;
Masse, C. E. J . Am. Chem. Soc. 2001, 123, 10840.
(29) Calculations were performed as a Monte Carlo conformational
search over the MM2* surface (see: (a) Burkert, U.; Allinger, N. L.
Molecular Mechanics; ACS Monograph 177; American Chemical
Society: Washington, DC, 1982) using MacroModel v. 4.5 (see: Mo-
hamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J . Comput. Chem.
1990, 11, 440) on a Silicon Graphics R5000 Indy workstation.
5682 J . Org. Chem., Vol. 69, No. 17, 2004

1-Azaspirocyclics by Siloxy-Epoxide Expansion
cyclopentanol silyl ethers expand to form cyclohexanones
in a 1,2-anti fashion, cyclobutanol silyl ethers expand via
either a 1,2-anti or 1,2-syn manner, depending on condi-
tions and substituents. Although the mechanistic under-
pinnings of this selectivity difference are still not fully
understood, the lower selectivity in the process may be
prescribed to the greater propensity of the cyclobutane
ring to expand compared to the cyclopentane ring. With
appropriate conditions, diastereomeric 1-azaspiro[5.4]-
decanones can be constructed in a complementary fash-
ion. A new method to form 1-azaspiro[5.4]decanones
substituted at the 3-position of the heterocycle was also
uncovered.
in these systems that has not been addressed to date is
the nature of the substituent on the nitrogen atom. In
addition, enabling larger rings to undergo ring expansion
would be a useful advance as well. Efforts to resolve these
issues as well as to apply these reactions in the total
synthesis of alkaloids are ongoing.
Ack n ow led gm en t. The authors thank the Univer-
sity of British Columbia and the Natural Sciences and
Engineering Research Council (NSERC) of Canada for
the financial support of our programs. M.D.B.F. thanks
the Gladys Estella Laird Foundation for a Laird Fel-
lowship.
Although not as general as the standard “carbocyclic”
versions of the R-siloxy-epoxide process, this method can
generate complex spiro-fused heterocyclic ring systems
containing two contiguous stereogenic carbon centers in
a single operation. Because of the relative difficulties in
accessing the necessary substrates, an important factor
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures, characterization data, copies of spectra, and CIF files
for previously unreported compounds. This material is avail-
able free of charge via the Internet at http://www.pubs.acs.org.
J O049356+
J . Org. Chem, Vol. 69, No. 17, 2004 5683