Unique diastereoselectivity trends in aminyl radical cyclizations onto silyl enol ethers

Article abstract of DOI:10.1021/jo902426z

(Chemical Equation Presented) The cyclization of nitrogen-centered radicals onto silyl enol ethers is an efficient method for the synthesis of polyhydroxylated alkaloids as the 2-hydroxymethylpyrrolidine core can be readily accessed from a linear precursor. During our studies on the synthesis of polyhydroxylated alkaloid CYB-3, we found that the diastereoselectivity of the cyclization was dependent on a complex combination of sterics and olefin geometry. A more thorough understanding of the factors that lead to high diastereoselectivites would greatly expand the utility of this methodology in complex natural product synthesis. We have found that cyclization diastereoselectivities of substrates with alkyl or aryl substitution were excellent regardless of olefin geometry or substitution pattern. When electronegative substituents were introduced adjacent to the silyl enol ether, only Z-silyl enol ethers provide high diastereoselectivites. Temperature, steric size of the silyl group, and sterics and electronics of the metal hydride affected the selectivity to a lesser extent.

Full text of DOI:10.1021/jo902426z

pubs.acs.org/joc
Unique Diastereoselectivity Trends in Aminyl Radical Cyclizations
onto Silyl Enol Ethers
Maria Zlotorzynska, Huimin Zhai, and Glenn M. Sammis*
Department of Chemistry, 2036 Main Mall, University of British Columbia, Vancouver, British Columbia
V6T 1Z1, Canada
gsammis@chem.ubc.ca
Received November 12, 2009
The cyclization of nitrogen-centered radicals onto silyl enol ethers is an efficient method for the
synthesis of polyhydroxylated alkaloids as the 2-hydroxymethylpyrrolidine core can be readily
accessed from a linear precursor. During our studies on the synthesis of polyhydroxylated alkaloid
CYB-3, we found that the diastereoselectivity of the cyclization was dependent on a complex
combination of sterics and olefin geometry. A more thorough understanding of the factors that lead
to high diastereoselectivites would greatly expand the utility of this methodology in complex natural
product synthesis. We have found that cyclization diastereoselectivities of substrates with alkyl or
aryl substitution were excellent regardless of olefin geometry or substitution pattern. When
electronegative substituents were introduced adjacent to the silyl enol ether, only Z-silyl enol ethers
provide high diastereoselectivites. Temperature, steric size of the silyl group, and sterics and
electronics of the metal hydride affected the selectivity to a lesser extent.
Introduction
are synthetically useful radical acceptors as the cyclization
products provide versatile alkoxy or siloxy functionalities.
Cyclizations between carbon-,1 oxygen-,2 and nitrogen-
centered radicals3 onto alkenes have been explored, and their
highly predictable behavior with regard to both chemo- and
diastereoselectivity make these methodologies valuable syn-
thetic transformations. One class of cyclization acceptor that
is relatively unexplored is silyl enol ethers.4,5 Silyl enol ethers
(3) For reviews on nitrogen-centered radicals, see: (a) Neale, R. S.
Synthesis 1971, 1–15. (b) Mackiewicz, P.; Furstoss, R. Tetrahedron 1978,
34, 3241–3260. (c) Stella, L. Angew. Chem., Int. Ed. Engl. 1983, 22, 337–350.
(d) Esker, J. L.; Newcomb, M. Adv. Heterocycl. Chem. 1993, 58, 1–45.
(e) Zard, S. Z. Synlett 1996, 1148–1154. (f) Fallis, A. G.; Brinza, I. M.
Tetrahedron 1997, 53, 17534–17594. (g) Zard, S. Z. Chem. Soc. Rev. 2008, 37,
1603–1618. (h) Curran, D. P. In Comprehensive Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 4, Chapter 4.2, pp
811-812. (i) Stella, L. In Radicals in Organic Synthesis; Renaud, P., Sibi,
M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 2, Chapter 51, pp 407-426.
(j) Minozzi, M.; Nanni, D.; Spagnolo, P. Chem.;Eur. J. 2009, 15, 7830–7840.
(4) For select examples of carbon-radical cyclizations onto silyl enol
ethers, see: (a) Urabe, H.; Kuwajima, I. Tetrahedron Lett. 1986, 27, 1355–
1358. (b) Walkup, R. D.; Kane, R. R.; Obeyesekere, N. U. Tetrahedron Lett.
1990, 31, 1531–1534. (c) Fensterbank, L.; Mainetti, E.; Devin, P.; Malacria,
M. Synlett 2000, 1342–1344. (d) Zhu, H.; Wickenden, J. G.; Campbell, N. E.;
Leung, J. C. T.; Johnson, K. M.; Sammis, G. M. Org. Lett. 2009, 11, 2019–
2022.
(1) For reviews on carbon-radical cyclizations, see: (a) Ramaiah, M.
Tetrahedron 1987, 43, 3541–3676. (b) Curran, D. P. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 4,
€
Chapter 4.1, pp 715-777. (c) Giese, B.; Kopping, B.; Gobel, T.; Dickhaut, J.;
Thoma, G.; Kulicke, K. J.; Trach, F. In Organic Reactions; Paquette, L. A., Ed.;
John Wiley & Sons, Inc.: New York, 1996; Vol. 48, Chapter 2, pp 303-856.
€
(d) Gansauer, A.; Bluhm, H. Chem. Rev. 2000, 100, 2771–2788. and references
therein.
(2) For the first example of oxygen-centered radical cyclizations, see:
(a) Surzur, J. M.; Bertrand, M. P.; Nougier, R. Tetrahedron Lett. 1969, 48,
4197–4200. For recent reviews, see: (b) Hartung, J. Eur. J. Org. Chem. 2001,
619–632. (c) Hartung, J.; Gottwald, T.; Spehar, K. Synthesis 2002, 1469–1498
and references therein.
(5) (a) Kim, S.; Kim, K. H.; Cho, J. R. Tetrahedron Lett. 1997, 38, 3915–
3918. (b) Zlotorzynska, M.; Zhai, H.; Sammis, G. M. Org. Lett. 2008, 10,
5083–5086.
ꢀ
864 J. Org. Chem. 2010, 75, 864–872
Published on Web 12/31/2009
DOI: 10.1021/jo902426z
r
2009 American Chemical Society

Zlotorzynska et al.
JOCArticle
SCHEME 1. Cyclization of Azide 1 To Afford Pyrrolidine 4
SCHEME 2. Chairlike (8) and Boatlike (9) Transition States
for Radical 5-Exo Cyclizations
However, not much is known about how the significant
increase in electron density of the olefin affects the diastereo-
selectivity of 5-exo radical cyclizations.
In the course of our investigations into heteroatom-cen-
tered radical cyclizations onto silyl enol ethers,5b,6 we re-
cently reported the first example of a 5-exo aminyl radical
cyclization onto a silyl enol ether (Scheme 1, 1).6 The tin-
bound aminyl radical7 cyclization proceeds in high yield to
afford 2-siloxymethyl pyrrolidines. This new methodology
was utilized as a key step in the syntheses of polyhydroxy-
lated alkaloids8 CYB-3 (Figure 1, 5)9,10 and 1,4-dideoxy-1,4-
imino-D-ribitol (6).11
is that 5-exo cyclizations can be analyzed through both
chairlike (8) and boatlike (9) transition states (Scheme 2).
Regardless of the substitution pattern, the overall cyclization
diastereoselectivity can be predicted through energy mini-
mization of these transition states. Supported by extensive
empirical evidence,1,13 this model for radical cyclization has
emerged as a powerful and predictable tool in organic
synthesis.14
FIGURE 1. Polyhydroxylated alkaloids.
While substituent effects in carbon-centered radical
cyclizations onto alkenes have been thoroughly studied,1
the tin-bound aminyl radical analogue may provide differ-
ent selectivity trends when cyclizing onto an electron-rich
olefin, such as a silyl enol ether. Herein, we report an
investigation into how sterics, temperature, hydride
source, silyl enol ether geometry, and electronics of the
substituent affect the overall diastereoselectivity of this
cyclization.
Despite the significant studies on the diastereoselectivities
of radical cyclizations onto alkenes,1-3 little is known about
the effects substituents have on 5-exo cyclizations onto silyl
enol ethers. Indeed, during our synthetic studies, we ob-
served some atypical substituent effects in regard to sterics
and olefin geometry.6 Typically, the diastereoselectivity of
radical cyclizations is predicted on the basis of computa-
tional studies on radical cyclization transition states by
Beckwith12 and Houk.13 The premise behind these studies
Results and Discussion
(6) Zhai, H.; Zlotorzynska, M.; Sammis, G. M. Chem. Commun. 2009,
5716–5718.
(7) For initial studies on azides as precursors for tin-bound aminyl
radicals, see: (a) Kim, S. G.; Joe, H.; Do, J. Y. J. Am. Chem. Soc. 1993,
115, 3328–3329. (b) Kim, S.; Joe, G. H.; Do, J. Y. J. Am. Chem. Soc. 1994,
116, 5521–5522.
(8) For recent reviews, see: (a) Winchester, B.; Fleet, G. W. J. Glycobiol-
ogy 1992, 2, 199–210. (b) St€utz, A. E. Iminosugars as Glycosidase Inhibitors:
Nojirimycin and Beyond; Wiley-VCH: Weinheim, 1999. (c) Watson, A. A.; Fleet,
G. W. J.; Asano, N.; Molyneux, R. J.; Nash, R. J. Phytochemistry 2001, 56, 265–
295. (d) Wrodnigg, T. M. Monatsh. Chem. 2002, 133, 393. (e) Compain, P.;
Martin, O. R. Curr. Top. Med. Chem. 2003, 3, 541–560.
(9) Nash, R. J.; Bell, E. A.; Fleet, G. W. J.; Jones, R. H.; Williams, J. M.
J. Chem. Soc., Chem. Commun. 1985, 738–740.
(10) CYB-3 has been found to exhibit modest glycosidase inhibition:
Scofield, A. M.; Fellows, L. E.; Nash, R. J.; Fleet, G. W. J. Life Sci. 1986, 39,
645–650.
(11) For early syntheses of 2-hydroxymethyl-3-hydroxypyrrolidine and
protected derivatives, see: (a) Ikota, N.; Hanaki, A. Heterocycles 1988, 27,
2535–2537. (b) Roemmele, R. C.; Rapoport, H. J. Org. Chem. 1989, 54,
1866–1875. For recent, enantioselective syntheses of 2-hydroxymethyl-3-
hydroxypyrrolidine, see: (c) Merino, P.; Delso, I.; Tejero, T.; Cardona, F.;
Marradi, M.; Faggi, E.; Parmeggiani, C.; Goti, A. Eur. J. Org. Chem. 2008,
2929–2947. (d) Toumi, M.; Couty, F.; Evano, G. Tetrahedron Lett. 2008, 49,
1175–1179. (e) Schomaker, J. M.; Bhattacharjee, S.; Yan, J.; Borhan, B.
J. Am. Chem. Soc. 2007, 129, 1996–2003. (f) Toumi, M.; Couty, F.; Evano, G.
Angew. Chem., Int. Ed. 2007, 46, 572–575. (g) Michaelis, D. J.; Ischay, M. A.;
Yoon, T. P. J. Am. Chem. Soc. 2008, 130, 6610–6615.
We started our investigations with siloxy substitution at
C3 as the corresponding substituted pyrrolidine is a common
core in many of the polyhydroxylated alkaloids (Figure 1).8
A general synthesis for the preparation of 3-siloxy-substi-
tuted cyclization precursors is presented in Scheme 3. Synth-
esis of the tert-butyldimethylsiloxy-substituted cyclization
substrates began with known enantioenriched diol 11.15
Conversion of the primary alcohol to the iodide followed
by protection of the secondary alcohol with a siloxy group
afforded ester 12 in excellent yield. A subsequent DIBAL
reduction of the ester followed by silyl enol ether formation
provided 13 as a 40:60 mixture of E- to Z-isomers. Displace-
ment of the iodide with azide afforded the key cyclization
precursor 14.
According to the Beckwith-Houk studies, there are four
transition states that should be analyzed for this cyclization:
chair transition states 15 and 18 and boat transition states 17
(14) For reviews on the use of radicals in natural product synthesis, see:
(a) Curran, D. P. Synthesis 1988, 417–439. (b) Curran, D. P. Synthesis 1988,
440–513. (c) Jasperse, C. P.; Curran, D.; Fevig, T. L. Chem. Rev. 1991, 91,
1237–1286 and references therein.
(15) For a synthesis of (S)-methyl-3,5-dihydroxypentanoate, see: Loubinaux,
B.; Sinnes, J. -L.; O’Sullivan, A. C.; Winkler, T. Tetrahedron 1995, 51, 3549–
3558.
(12) (a) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron Lett. 1985, 26,
373–376. (b) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985, 41, 3925–
3941.
(13) Spellmeyer, D. C.; Houk, K. N. J. Org. Chem. 1987, 52, 959–974.
J. Org. Chem. Vol. 75, No. 3, 2010 865

Zlotorzynska et al.
JOCArticle
SCHEME 3. Synthesis of Cyclization Precursor 14
SCHEME 5. Cyclization of Azide 14 To Afford Pyrrolidine 16
and Acyclic Amine 21
SCHEME 6. Cyclization of Azide 14 To Afford Pyrrolidine 16
and Acyclic Amine 21
SCHEME 4. Beckwith-Houk Transition States for the Cycli-
zation of Azide 14
SCHEME 7. Cyclization of Azide 14 To Afford Pyrrolidine 16
and Acyclic Amine 21
observed for both carbon and oxygen-centered radical cycliza-
tions. In addition, pyrrolidine 16 was formed in a virtually
equimolar ratio with acyclic amine 21 presumably arising from
trapping of the nitrogen radical prior to cyclization. This yield
was also significantly lower than expected compared to cycliza-
tion of simple azide 1.
In an effort to increase the diastereoselectivity of the
cyclization, we next examined the effects of lowering the
reaction temperature. We focused on a different initiation
method (Scheme 6) as azobisisobutyronitrile (AIBN) re-
quires elevated temperatures for radical formation. Treat-
ment of azide 14 with triethylborane and oxygen at ambient
temperature17 in the presence of tributyltin hydride formed
pyrrolidine 16 with increased diastereoselectivity. However,
there was also a significant increase in the amount of acyclic
amine 21.
We hypothesized that we may be able to increase the ratio
of pyrrolidine 16 to amine 21 by decreasing the electron
density of the tin-bound nitrogen-centered radical. A more
electrophilic nitrogen-centered radical is expected to have
faster rates of cyclization with the electron-rich silyl enol
ether. Cyclization of azide 14 with triphenyltin hydride led to
a slight increase in the ratio of cyclized pyrrolidine 16 and
amine 21 (Scheme 7). It also afforded an increase in the
diastereoselectivity, presumably due to the increased steric
bulk around the nitrogen-centered radical.
and 20 (Scheme 4). Of the two chair transition states, 15
should be lower in energy than 18 through simple minimiza-
tion of steric interactions of the siloxy substitutent. Like-
wise, boat transition state 20 should be lower in energy
than 17 through simple steric minimization. Furthermore,
boat-transition state 20 does not have more significant
steric interactions than are present in chair transition
state 18. On the basis of these models, and on analogous
oxygen-centered radical5b,16 cyclizations onto simple al-
kenes, it was expected that cyclization of the nitrogen-
centered radical should provide the trans-pyrrolidine 16 as
the major diastereomer.
Cyclization of azide 14 afforded pyrrolidines 16 and 19 as a
60:40 ratio of trans- to cis-isomers (Scheme 5), a diastereos-
electivity significantly lower than what has been previously
Formation of acyclic amine 21 may be due to steric conges-
tion in the transition state 15, thus lowering cyclization rates.
(16) For representative diastereoselectivity studies, see: (a) Hartung, J.;
Hiller, M.; Schmidt, P. Chem.;Eur. J. 1996, 2, 1014–1023. (b) Hartung, J.;
ꢀ
Kneuer, R.; Laug, S.; Schmidt, P.; Spehar, K.; Svoboda, I.; Fuess, H. Eur.
J. Org. Chem. 2003, 4033–4052. (c) Hartung, J.; Stowasser, R.; Vin, D.;
Bringmann, G. Angew. Chem., Int. Ed. Engl. 1996, 35, 2820–2823.
(17) (a) Nozaki, K.; Oshima, K.; Utimoto, K. J. Am. Chem. Soc. 1987,
109, 2547–2549. (b) Yorimitsu, H.; Oshima, K. In Radicals in Organic Synthesis;
Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, pp 11-27.
866 J. Org. Chem. Vol. 75, No. 3, 2010

Zlotorzynska et al.
JOCArticle
SCHEME 8. Cyclization of Azide 22 To Afford Pyrrolidine 23
and Acyclic Amine 24
SCHEME 10. Beckwith-Houk Transition States for the Cycliza-
tion of Z-Silyl Enol Ethers
SCHEME 9. Cyclization of Azides 25 and 22 To Afford Pyr-
rolidines 26 and 23
To test this hypothesis, we synthesized azide 22 using a similar
protocol as azide 14.18 Indeed, cyclization of azide 22 afforded
an increased ratio of pyrrolidine 23 to amine 24 (Scheme 8)
compared to cyclization of the TBS analog under similar
conditions (Scheme 5). Despite the improved yields using
cyclization precursor 22,19 the diastereoselectivity was still
significantly lower than comparable cyclizations with oxygen-
centered radicals.5b,16
ether is the dominant factor in the diastereoselectivity of this
reaction.
The high selectivity for the cyclization of Z-silyl enol ether
22 can be understood using the standard Beckwith-Houk
transition states (Scheme 10). Chair transition state 29,
which provides the observed trans-substituted pyrrolidine,
has few major steric interactions. The siloxy substituent at
C3 is in the equatorial position and the silyl enol ether is
oriented to minimize A1,3-strain. Transition states 32 and 34,
which lead to the cis-substituted pyrrolidine, should both
be significantly higher in energy than 29. Chair transition
state 32 has significant A1,3 -strain between the siloxy
substitutents.
The poor diastereoselectivity when cyclizing E-silyl enol
ether 25 is much more difficult to understand using the
Beckwith-Houk transition states. Chair transition state
35, leading to trans-substituted pyrrolidine 30, has few major
steric interactions (Scheme 11). The alternative chair transi-
tion state (37) should be slightly higher in energy as the A
value for a tert-butyldimethylsiloxy group is 1.06 kcal/mol.20
While transition state 36 should be high in energy, boat
transition state 38 also has reduced steric interactions be-
tween the silyl enol ether substituents and the rest of the
substrate compared to the chair transition state 37. Though
transition states 37 and 38 are closer in energy to 35 than for
Z-silyl enol ether 22, the cyclization should still favor the
trans-pyrrolidine as cyclization of alkoxy radicals with sub-
stitution at C3 provide excellent selectivity to the correspond-
ing tetrahydrofuran. The lack of selectivity in the cyclization
of azide 25 indicates steric minimization is likely not the sole
factor in reaction diastereoselectivity.
The decreased diastereoselectivities in all of these cycliza-
tions could arise from the difference in selectivities for the
E- and Z-silyl enol ethers. To test this hypothesis, we
synthesized both the E- and Z-silyl enol ethers 22 and 25
(Scheme 9). These substrates were synthesized according to
Scheme 3 with the exception of the silyl enol ether formation
step. Synthesis of the E-silyl enol ether was accomplished
using TBSCl and DBU in dichloromethane at 35 °C to
form the thermodynamic product. The Z-silyl enol ether
was prepared using TESCl and DBU at 0 °C to form the
Z-enriched product which was subsequently purified using
flash chromatography. Cyclization of E-silyl enol ether 25
afforded pyrrolidine 26 in moderate yield with no diaster-
eoselectivity while cyclization of Z-silyl enol ether 22 af-
forded pyrrolidine 23 in excellent yield exclusively as the
trans-isomer (Scheme 9). While there is a slight contribution
to diastereoselectivity from the difference in steric size
between the silyl enol ether substituents (Schemes 5 and 8)
the dramatic difference in selectivities in the cyclizations of
azides 25 and 22 suggests that the geometry of the silyl enol
(18) The enolization was accomplished by treatment with TESCl and
DBU at ambient temperature. See the Supporting Information for further
details.
(19) The Z- to E-ratios of amines 21 and 24 were slightly higher than the
Z- to E-ratios of azide-containing silyl enol ethers 14 and 22. This suggests
that the rate of cyclization of the Z-silyl enol ether is slightly slower than the
cyclization rate of E-silyl enol ether.
(20) Eliel, E. L.; Satici, H. J. Org. Chem. 1994, 59, 688–689.
J. Org. Chem. Vol. 75, No. 3, 2010 867

Zlotorzynska et al.
JOCArticle
SCHEME 11. Beckwith-Houk Transition States for the Cycliza-
tion of E-Silyl Enol Ethers
SCHEME 13. Synthesis of Cyclization Precursors 42a and 42b
SCHEME 14. Synthesis of Cyclization Precursors 43a and 43b
interaction is not present in transition state 37 as the σ*C-O
of the siloxy group and the π-system of the silyl enol ether
are orthogonal. Similarly, there is either poor or no overlap
between the σ*C-O and the olefin in boat transition states 36
and 38. Since the nitrogen-centered radical is slightly elec-
trophilic, the cyclization rate of transition state 35 should be
slowed in comparison with transition states 37 and 38 due to
the decreased electron density in the olefin. Presumably, the
stereoelectronic interactions are comparable to the higher
steric demand in transtion states 37 and 38, leading to similar
rates of cyclization. This interaction is also present in transi-
tion state 29, but the steric interactions are apparently
sufficiently severe to outweigh the stereoelectronic effects.
To test this hypothesis, we next investigated substrates
with alkyl and aryl substitution at C3. The σ* of these substi-
tuents are less accessible for interaction with the π-system
as they are higher in energy. Synthesis of these substrates
began from readily accessible diols (Scheme 13, 39a and 39b)
Monotosylation and Swern oxidation afforded aldehydes
40a and 40b, which were subjected to TBSOTf and diisopro-
pylethylamine at 0 °C to afford Z-enriched silyl enol ethers
41a and 41b. The cyclization precursors (42a and 42b) were
synthesized by a final displacement of the tosylate with
sodium azide.
SCHEME 12. Stereoelectronic Interaction in Beckwith-houk
Transition States for the Cyclization of E-Silyl Enol Ethers
The E-silyl enol ethers (43a and 43b) were synthesized
using an analogous procedure except the final two steps
had to be reversed as the enolization conditions necessary for
E-selectivity led to elimination of the tosylate (Scheme 14).
Displacement of tosylate 40 followed by treatment with
TBSCl and DBU at 35 °C afforded E-enriched silyl enol
ether 43a and 43b in good yields.
Cyclization of Z-silyl enol ethers (42a and 42b) proceeded
with high diastereoselectivity to yield pyrrolidines 44a and
44b (Table 1, entries 1 and 3). Methyl substitution (42a)
provided the cyclized product in high yield and excellent
diastereoselectivity. Increasing the steric size to a phenyl
group afforded pyrrolidine 44b in excellent yields. Only the
trans-isomer could be detected by 1H NMR spectroscopy.
A possible explanation for the low diastereoselectivity
during the cyclization of azide 25 is a stereoelectronic inter-
action in one of the transition states (Scheme 12). In chair
transition state 35, the σ*C-O of the siloxy group is in
alignment with the π-system of the silyl enol ether, which
will lead to decreased electron density of the olefin. This
868 J. Org. Chem. Vol. 75, No. 3, 2010

Zlotorzynska et al.
JOCArticle
TABLE 1. Cyclization of Methyl- and Phenyl-Substituted Substrates
SCHEME 15. Cyclization of Azide 45 To Afford Pyrrolidine 46
To further investigate whether the steric difference between
methyl and siloxy groups is sufficient to explain the differ-
ence in selectivities in the cyclization of azides 25 and 43a, we
cyclized azide 45 (Scheme 15). Since the siloxy substituent
is homoallylic to the silyl enol ether, there should be no
stereoelectronic interaction between the π-system and the
σ*C-O of the silyl ether. The cyclization diastereoselectivity
should therefore be governed exclusively by steric minimiza-
tion. If indeed the siloxy substituent does not have a suffi-
ciently high A-value to impart any diastereoselectivity, there
should be an equimolar amount of the cis and trans diaster-
eomers formed. However, the cyclization provided a 75:25
ratio of cis to trans isomers. Furthermore, substrates with
substitution at C3 are known to cyclize in higher diastereo-
selectivity than those with substitution at the C4.12,13
Thus, the 50:50 mixture obtained for the cyclization of azide
25 cannot be explained by a simple steric minimization
argument.
We next investigated the scope of alkyl and aryl substitu-
tion at C4 and C5 to determine if these substrates were
consistent with the Beckwith-Houk studies on radical cycli-
zation diastereoselectivity (Table 2).23 We began by examin-
ing phenyl substitution on the silyl enol ether (entry 1).
Cyclization of silyl enol ether 47a provided desired pyrroli-
dine 48a in good yield and in a 90:10 ratio of erythro to threo
isomers, which is substantially higher than was observed
with analogous oxygen-centered radical cyclizations.5b
Substrates 47b and 47c (entries 2-3) were used to examine
how substituents at C5 would influence the yield and diastereo-
selectivity. Methyl substituent 47b cyclized in excellent yield,
but only in a 65:35 ratio of trans- to cis-isomers. Increasing
the steric size of the substituent from methyl (47b) to phenyl
(47c) provided comparable yields and a remarkable increase
in diastereoselectivity (entry 3). The diastereoselectivity
again was higher than what was observed for analogous
oxygen-centered radical cyclizations.5b The diastereomers
obtained were in accordance with the Beckwith-Houk
studies.
aReactions were carried out on >0.25 mmol scale. bThe relative
stereochemistry was determined by derivatization of the product and
comparison to known compounds. See the Supporting Information for
experimental details. cIsolated yields of the mixture of diastereomers
after flash chromatography. dThe diastereomeric ratio was determined
by 1H NMR spectroscopy of purified products.
We next examined the effect of the geometry of the silyl enol
ether acceptor on the diastereoselectivity of the cyclization for
substrates with the same substitution pattern (Table 1). As
was observed in substrates with siloxy substituents, E-silyl
enol ether 43a cyclized in much lower diastereoselectivity
(entry 2) than did the corresponding Z-diastereomer 42a
(entry 1). However, the difference in diastereoselectivity be-
tween the Z- and E-silyl enol ethers with a methyl substituent
was not as pronounced as what was observed for the siloxy
substituted substrates. Furthermore, this drop in selectivity
was not observed in the cyclization of the phenyl substituted
E-silyl enol ether 43b (entry 4),21 which cyclized to give
exclusively the trans-isomer. As the A-value for a methyl
substituent (1.74 kcal/mol)22 is only moderately higher than
a siloxy substitutent (1.06 kcal/mol), sterics alone do not
explain the increase in selectivity between the cyclizations of
azides 25 and 43a. However, it is consistent with the stereo-
electronic interaction shown in transition state 35 (Scheme 12)
as the siloxy substituent has a significantly lower energy σ*
orbital.
We next examined whether possible β-fragmentation
pathways would provide lower yields of cyclized product.
Substrate 47d (entry 4) cyclized to desired pyrrolidine 48d in
1
high yield with no fragmentation product observed by H
NMR spectroscopy. Products resulting from 1,5-hydrogen
abstraction were also not observed. The cyclization also
proceeded in a 72:28 ratio of cis- to trans-isomers. Phenyl
substituted substrate 47e cyclized in high yield. The increase
in steric size also provided an increase in diastereoselectivity
to an 89:11 ratio of cis- to trans-isomers. As in previous
entries, the selectivity observed was consistent with the
Beckwith-Houk studies. Despite the added stabilization of
the β-fragmentation product, no other products were ob-
served by 1H NMR spectroscopy. Thus, in addition to high
(21) The σ*C-C of phenyl substituents have been demonstrated to be
acceptors in nucleophilic additions to acyclic carbonyls: Heathcock, C. H.;
Flippin, L. A. J. Am. Chem. Soc. 1983, 105, 1667–1668. In 6-membered rings,
the sterics of the phenyl ring override the stereoelectronic component:
Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019–5087.
(22) Booth, H.; Everett, J. R. J. Chem. Soc., Chem. Commun. 1976, 278–
280.
(23) Analogous oxygen-centered radical cyclizations onto silyl enol
ethers (ref 5b) were consistent with the Beckwith-Houk model for diastereo-
selectivity.
J. Org. Chem. Vol. 75, No. 3, 2010 869

Zlotorzynska et al.
JOCArticle
TABLE 2. Cyclization of Methyl- and Phenyl-Substituted Substrates
diastereoselectivities of substrates with alkyl or aryl substi-
tution were excellent regardless of olefin geometry or sub-
stitution pattern. Efforts to further explore this cyclization
methodology in the context of natural product synthesis are
currently underway.
Experimental Section
General Methods. All reactions were performed under a
nitrogen atmosphere in flame-dried glassware. Tetrahydrofur-
an, diethyl ether, dichloromethane and benzene were purified
by a solvent purification system. Thin-layer chromatography
(TLC) was performed on UV254-precoated TLC plates. Chro-
matographic separations were effected over silica gel. Triethy-
lamine washed silica gel has been stirred with triethylamine prior
to packing. All chemicals were purchased from commercial
sources and used as received. Azide-containing silyl enol ethers,
such as azides 1, 14, 25, 42a,b, 43a,b, 45, and 47a-e, are bench-
stable for at least 2 weeks.
General Cyclization Procedure. A solution of Bu3SnH (1.2
equiv), AIBN (0.1 equiv), and silyl enol ether in degassed
benzene24 (0.05 M) was heated to 80 °C and stirred for 12 h,
the solution was allowed to cool to room temperature, and the
solvent was removed by rotary evaporation. Purification by
flash chromatography afforded the cyclized product.
General Tosylation followed by Deprotection Procedure for
Sterechemistry Determination. To a solution of pyrrolidine 48c
(25 mg) and p-toluenesulfonyl chloride (20 mg, 0.1 mmol)
in CH2Cl2 (3 mL) at 0 °C was added triethylamine (20 mg,
0.2 mmol). The resulting solution was stirred for 2 h, quenched
with water (1 mL), and extracted with CH2Cl2 (2 ꢀ 3 mL). The
combined organic extracts were dried over Na2SO4 and filtered,
and the solvent was removed by rotary evaporation to provide
a pale yellow oil. Purification by flash chromatography
(5% EtOAc in hexanes) provided the tosylate as a colorless
oil, which was dissolved in MeOH (1 mL), and p-toluenesulfonic
acid monohydrate (10 mg, 0.053 mmol) was added. After being
stirred for 4 h, the solution was diluted with Et2O (2 mL) and
washed with water (2 mL). The solvent was removed by rotary
evaporation to provide a crude oil. The NMR spectra of the
crude alcohols were compared with literature compounds.
2-(tert-Butyldimethylsilanyloxymethyl)pyrrolidine (4). Silyl
enol ether 16 (301 mg, 1.27 mmol) was subjected to the general
cyclization procedure. Purification by flash chromatography
(1% methanol in EtOAc) afforded 206 mg (75%) of pyrrolidine
4 as a yellow oil: IR (neat) 3354, 2954, 2857, 1652, 1418 cm-1; 1H
NMR (400 MHz, CDCl3) δ 3.56 (dd, J = 10.0, 4.8 Hz, 1 H), 3.49
(dd, J = 10.0, 5.6 Hz, 1 H), 3.19-3.08 (m, 1 H), 2.96 (dt, J =
10.0, 6.4 Hz, 1 H), 2.81 (dt, J = 10.0, 7.2 Hz, 1 H), 2.57 (s, br,
1 H), 1.79-1.49 (m, 3 H), 1.56-1.46 (m, 1 H), 0.86 (s, 9 H), 0.02
(s, 6 H); 13C NMR (100 MHz, CDCl3) δ 65.6, 59.9, 46.4, 27.4,
25.8, 25.3, 18.2, -5.5; HRMS-ESI (m/z) [M þ Na]þ calcd for
C11H26NOSiNa 216.1784, found 216.1785.
aReactions were carried out on >0.25 mmol scale. bThe relative
stereochemistry was determined by derivatization of the product and
comparison to known compounds. See the Supporting Information for
experimental details. cIsolated yields of the mixture of diastereomers
after flash chromatography. dThe diastereomeric ratio was determined
by 1H NMR spectroscopy of crude reaction mixtures. (e)Isolated yield of
the trans-stereoisomer. The cis-isomer was also isolated in 12% yield.
diastereoselectivity, these cyclizations also exhibit a high
level of chemoselectivity.
Conclusion
We have found conditions to achieve high diastereoselec-
tivities for cyclizations between tin-bound aminyl radicals
and silyl enol ethers regardless of the substituent or substitu-
tion pattern. While temperature, source of hydride, and
sterics of the substituent all influenced the cyclization selec-
tivity, the geometry of the silyl enol ether had the most
dramatic effect on selectivity. Cyclizations of Z-silyl enol
ethers provide high diastereoselectivities for all substrates
examined. However, the diastereoselectivity using E-silyl
enol ethers was significantly lower when an electronegative
substituent was adjacent to the silyl enol ether. The low
diastereoselectivities in cyclizations of E-silyl enol ethers
with electron-withdrawing substituents are likely due to a
combination of small steric differentiation between the
transition states and a stereoelectronic interaction that de-
creases the electron density from the olefin. Cyclization
2-(tert-Butyldimethylsilanyloxymethyl)-3-methylpyrrolidine
(44a). Silyl enol ether 42a (375 mg, 1.0 mmol) was subjected to
the general cyclization procedure. Purification by flash chro-
matography (5% methanol in EtOAc) gave 163 mg (71%) of
pyrrolidine 44a (trans/cis = 93:7) as a light yellow oil. The
relative stereochemistry was determined by analogy to pyrro-
1
lidine 44b: IR (neat) 3361, 2957, 1645, 1456, 1412 cm-1; H
NMR (400 MHz, CDCl3) δ 5.74 (s, br, 1H), 3.78 (dd, J = 10.4,
4.8 Hz, 1H), 3.66 (dd, J = 10.4, 4.8 Hz, 1H), 3.17-3.04 (m,
2H), 2.91-2.81 (m, 1H), 2.08-1.95 (m, 2H), 1.03 (d, J = 6.4
Hz, 3H), 0.86 (s, 9H), 0.054 (s, 3H), 0.045 (s, 3H); 13C
NMR (100 MHz, CDCl3) δ 66.7, 62.5, 44.9, 34.9, 33.5, 25.8,
(24) Benzene is a listed carcinogen within the EEC. Appropriate ventila-
tion and safety precautions should be taken when working with this solvent.
870 J. Org. Chem. Vol. 75, No. 3, 2010

Zlotorzynska et al.
JOCArticle
18.2, 17.8, -5.5, -5.6; HRMS-ESI (m/z) [M þ H]þ calcd for
(m, 0.64H), 3.20-3.07 (m, 0.66H), 2.04 (s, br, 1H), 1.95-1.69
(m, 2H), 1.50-1.19 (m, 2H), 1.16 (d, J = 6.4 Hz, 3H, cis), 1.12
(d, J = 6.4 Hz, 3H, trans), 0.90 (s, 9H), 0.05 (s, 6H); 13C NMR (100
MHz, CDCl3) δ 66.0, 65.7, 60.5, 59.2, 54.9, 53.1, 33.8, 33.6, 27.9,
25.9, 21.5, 21.1, 18.3, -5.3, -5.4; HRMS-ESI (m/z) [M þ H]þ calcd
for C12H28NOSi 230.1940, found 230.1934.
C
12H28NOSi 230.1940, found 230.1946.
2-(tert-Butyldimethylsilanyloxymethyl)-3-phenylpyrrolidine
(44b). Silyl enol ether 42b (206 mg, 0.65 mmol) was subjected
to the general cyclization procedure. Purification by flash
chromatography (67% EtOAc in hexanes) gave 155 mg
(82%) of pyrrolidine 44b (trans/cis = >95:5) as a light yellow
oil. The pyrrolidine was subjected to the general tosylation
followed by deprotection procedure and the relative stereo-
chemistry was confirmed by comparison to the desilylated
2-(tert-Butyldimethylsilanyloxymethyl)-5-phenylpyrrolidine
(48c). Silyl enol ether 47c (280 mg, 0.88 mmol) was subjected
to the general cyclization procedure. Purification by flash
chromatography (20% EtOAc in hexanes) gave 169 mg
(66%) of trans-pyrrolidine 48c as a light yellow oil. The
pyrrolidine was subjected to the general deprotection proce-
dure, and the relative stereochemistry was confirmed by
comparison to the desilylated alcohol:26 IR (neat) 3330,
alcohol:25 IR (neat) 3317, 2928, 1651, 1456, 1254 cm-1 1H
;
NMR (400 MHz, CDCl3) δ 7.33-7.18 (m, 5H), 3.71 (dd, J =
10.0, 3.2 Hz, 1H), 3.52 (dd, J = 10.0, 4.8 Hz, 1H), 3.21-3.12
(m, 2H), 3.11-3.04 (m, 1H), 2.97 (q, J = 8.8 Hz, 1H), 2.32-
2.16 (m, 2H), 2.04-1.92 (m, 1H), 0.90 (s, 9H), 0.05 (s, 3H), 0.04
(s, 3H); 13C NMR (100 MHz, CDCl3) δ 143.9, 128.4, 127.6,
126.1, 68.5, 62.9, 46.9, 46.5, 35.7, 25.9, 18.2, -5.5; HRMS-ESI
(m/z) [M þ Na]þ calcd for C17H29NONaSi 314.1916, found
314.1911.
2928, 2856, 1673, 1462, 1255, 1112 cm-1 1H NMR (400
;
MHz, CDCl3) δ 7.41 (d, J = 7.2 Hz, 2H), 7.36 (t, J = 7.2
Hz, 2H), 7.30-7.25 (m, 1H), 4.31 (t, J = 7.2 Hz, 2H), 3.63-
3.53 (m, 3H), 2.29-2.20 (m, 1H), 2.15-2.01 (m, 2H), 1.78
(dq, J=12.4, 8.8 Hz, 1H), 1.66-1.54 (m, 1H), 0.95 (s, 9H),
0.12 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 145.4, 128.3,
126.7, 126.4, 66.2, 61.0, 59.6, 34.9, 27.8, 25.9, 18.3, -5.3;
HRMS-ESI (m/z) [M þ H]þ calcd for C17H30NOSi 292.2097,
found 292.2092.
Further elution afforded 31 mg (12%) of cis-pyrrolidine 48c
as a light yellow oil: IR (neat) 2954, 2856, 1255, 1097 cm-1; 1H
NMR (400 MHz, CDCl3) δ 7.40 (d, J =7.2 Hz, 2H), 7.32 (t, J =
7.6 Hz, 2H), 7.23 (t, J =7.2 Hz, 1H), 4.17 (t, J =7.2 Hz, 2H),
3.71 (dd, J =10.0, 4.8 Hz, 1H), 3.65 (dd, J = 10.0, 5.6 Hz, 1H),
3.53 (quin, J = 6.0 Hz, 1H), 2.20-2.10 (m, 1H), 1.94-1.81 (m,
2H), 1.75-1.65 (m, 2H), 0.92 (s, 9H), 0.09 (s, 6H); 13C NMR
(100 MHz, CDCl3) δ 144.8, 128.3, 126.8, 126.6, 66.7, 62.8, 60.3,
34.3, 27.8, 26.0, 18.4, -5.3; HRMS-ESI (m/z) [M þ H]þ calcd
for C17H30NOSi 292.2097, found 292.2090.
2-(tert-Butyldimethylsiloxymethyl)-3-methylpyrrolidine (46).
Silyl enol ether 45 (372 mg, 1.0 mmol) was subjected to the
general cyclization procedure. Purification by flash chromatog-
raphy (5% methanol in CH2Cl2) gave 300 mg (86%) of pyrro-
lidine 46 (cis/trans = 75:25) as a light yellow oil. The relative
stereochemistry was determined by analogy to pyrrolidine 48d:
IR (neat) 2928, 2856, 1472, 1463, 1361, 1252 cm-1; H NMR
1
(400 MHz, CDCl3) δ 4.36-4.31 (m, 1H), 3.64 (d, J = 5.4 Hz,
1.5H, cis), 3.57-3.35 (m, 0.25H, trans), 3.43-3.34 (m, 0.25H,
trans), 3.17-3.04 (m, 1H), 2.90-2.83 (m, 1.5H, cis), 2.76-2.58
(m, 0.40H, trans), 2.06-1.99 (m, 1H), 1.86 (br, 1H), 1.46 (ddd,
J = 10.5, 6.7, 3.8 Hz, 1H), 0.90 (s, 9H), 0.88 (s, 9H), 0.06-0.04
(m, 12H); 13C NMR (100 MHz, CDCl3) δ 73.4, 65.6, 59.8, 55.8,
38.1, 25.9, 25.8, 25.2, 18.4, 18.1, -4.8, -5.4; HRMS-ESI (m/z)
[M þ H]þ calcd for C17H40NO2Si2 346.2598, found 346.2590.
2-[(tert-Butyldimethylsilanyloxy)phenylmethyl]pyrrolidine (48a).
Silyl enol ether 47a (309 mg, 0.97 mmol) was subjected to the
general cyclization procedure. Purification by flash chromatogra-
phy (3% methanol in EtOAc) gave 131 mg (50%) of erythro
pyrrolidine 48a as a colorless oil: IR (neat) 2955, 2856, 1471, 1253,
2-(tert-Butyldimethylsilanyloxymethyl)-4-ethylpyrrolidine (48d).
Silyl enol ether 47d (69 mg, 0.26 mmol) was subjected to the general
cyclization procedure. Purification by flash chromatography
(5% methanol in EtOAc) gave 49 mg (79%) of pyrrolidine 48d
(cis/trans = 72:28) as a light yellow oil. The relative stereochem-
istry was determined by analogy to pyrrolidine 48e: IR (neat) 3300,
1
1062 cm-1; H NMR (400 MHz, CDCl3) δ 7.35-7.20 (m, 5H),
2957, 2857, 1463, 1254 cm-1 1H NMR (400 MHz, CDCl3)
;
4.45 (d, J = 6.8 Hz, 1H), 3.18 (q, J = 7.2 Hz, 2H), 3.08-2.97 (m,
1H), 2.90-2.79 (m, 1H), 2.26 (s, br 1H), 1.80-1.61 (m, 2H),
3.64-3.46 (m, 2H), 3.27-3.18 (m, 1H), 3.16 (dd, J = 10.0, 7.2
Hz, 1H, trans), 3.03 (dd, J = 10.0, 7.2 Hz, 1H, cis), 2.56 (dd, J =
10.0, 7.2 Hz, 1H, trans), 2.45 (dd, J = 10.0, 8.4 Hz, 1H, cis), 2.08
(s, br, 1H), 2.07-1.92 (m, 2H), 1.74-1.63 (m, 1H), 1.43-1.32
(m, 1H), 1.08-0.98 (m, 1H), 0.90 (d, J = 7.2 Hz, 3H), 0.09 (s, 9H),
0.06, (s, 6H); 13C NMR (100 MHz, CDCl3) δ 66.2, 65.9, 60.4, 59.6,
52.9, 52.0, 41.8, 41.0, 34.6, 34.0, 27.6, 27.4, 25.9, 18.3, 12.9, -5.4;
HRMS-ESI (m/z) [M þ H]þ calcd for C13H30NOSi 244.2097,
found 244.2092.
2-(tert-Butyldimethylsilanyloxymethyl)-4-phenylpyrrolidine
(48e). Silyl enol ether 47e (203 mg, 0.64 mmol) was subjected to
the general cyclization procedure. Purification by flash chro-
matography (5% methanol in EtOAc) gave 143 mg (77%)
of pyrrolidine 48e (cis/trans = 89:11) as a light yellow oil. The
pyrrolidine was subjected to the general tosylation followed by
deprotection procedure and the relative stereochemistry and
dr were confirmed by comparison to the desilyled alcohol:26
IR (neat) 3312, 2953, 2855, 1654, 1454, 1255 cm-1; 1H NMR
(400 MHz, CDCl3) δ 7.35-7.16 (m, 5H), 3.71 (dd, J = 10.0, 4.8
Hz, cis, 1H), 3.66 (dd, J = 10.0, 5.6 Hz, cis, 1H), 3.65 (dd, J =
10.0, 5.2 Hz, trans, 1H), 3.61 (dd, J = 10.0, 5.6 Hz, trans, 1H),
3.50-3.20 (m, 3H), 3.00-2.84 (m, 1H), 2.33-2.24 (m, 1H),
2.10-1.88 (m, 2H), 1.71-1.61 (m, 1H), 0.93 (s, 9H), 0.09
(s, 6H); 13C NMR (100 MHz, CDCl3) δ 144.2, 144.1, 128.4,
128.0, 127.3, 127.2, 126.2, 66.0, 65.8, 60.7, 60.3, 55.4, 54.2,
1.49-1.32 (m, 2H), 0.88 (s, 9H), 0.04 (s, 3H), -0.21 (s, 3H); 13
C
NMR (100 MHz, CDCl3) δ 143.3, 128.0, 127.3, 126.8, 78.5, 66.0,
45.9, 27.2, 25.8, 24.5, 18.1, -4.5, -5.0; HRMS-ESI (m/z) [M þ
H]þ calcd for C17H30NOSi 292.2097, found 292.2091.
Further elution afforded 33 mg (12%) of pyrrolidine 48a
(erythro/threo = 50:50) as a light yellow oil: 1H NMR (400 MHz,
CDCl3) δ 7.35-7.20 (m, 5H), 4.45 (d, J = 6.8 Hz, 0.5H), 4.45 (d,
J = 6.8 Hz, 0.5H), 3.18 (q, J = 7.2 Hz, 1H), 3.21-3.19 (m, 1H),
3.08-2.97 (m, 2H), 2.91-2.73 (m, 1H), 2.03 (s, br 1H), 1.80-1.61
(m, 2H), 1.52-1.32 (m, 2H), 0.90 (s, 4.5H), 0.88 (s, 4.5H), 0.05
(s, 1.5H), 0.04 (s, 1.5H), -0.19 (s, 1.5H), -0.21 (s, 1.5H); 13C NMR
(100 MHz, CDCl3) δ 143.6, 143.3, 128.0, 127.4, 127.3, 126.8, 126.5,
78.5, 77.2, 66.2, 66.0, 46.7, 45.9, 27.2, 27.1, 25.8, 25.0, 24.5, 18.1,
-4.5, -5.0, -5.1.
2-(tert-Butyldimethylsilanyloxymethyl)-5-methylpyrrolidine (48b).
Silyl enol ether 47b (142 mg, 0.56 mmol) was subjected to the
general cyclization procedure. Purification by flash chromatography
(5% methanol in EtOAc) gave 99 mg (78%) of pyrrolidine 48b
(trans/cis = 65:35) as a light yellow oil. The relative stereochemistry
was determined by analogy to pyrrolidine 48c: IR (neat) 3364, 2927,
1645, 1462, 1379 cm-1; 1H NMR (400 MHz, CDCl3) δ 3.64 (dd,
J = 10.0, 4.8 Hz, 0.34H, cis), 3.56 (dd, J = 10.0, 4.8 Hz, 0.34H, cis),
3.51-3.45 (m, 1.3H, trans), 3.38-3.29 (m, 0.64H), 3.27-3.20
(25) Miura, K.; Hondo, T.; Nakagawa, T.; Takahashi, T.; Hosomi, A.
Org. Lett. 2000, 2, 385–388.
(26) Momotake, A.; Togo, H.; Yokoyama, M. J. Chem. Soc., Perkin
Trans. 1 1999, 1193–1200.
J. Org. Chem. Vol. 75, No. 3, 2010 871

Zlotorzynska et al.
JOCArticle
45.9, 45.2, 36.8, 36.2, 25.9, 25.7, 18.3, -5.3; HRMS-ESI (m/z)
(NSERC), and a doctoral fellowship from NSERC to
M.Z. We also thank Prof. Ciufolini for useful discussions.
[M þ H]þ calcd for C17H30NOSi 292.2097, found 292.2092.
Acknowledgment. This work was supported by the Uni-
versity of British Columbia, Merck-Frosst, the Natural
Sciences and Engineering Research Council of Canada
Supporting Information Available: Complete experimental
procedures and characterization data. This material is available
free of charge via the Internet at http://pubs.acs.org.
872 J. Org. Chem. Vol. 75, No. 3, 2010