Remote stereochemical control of both reacting centers in ketyl-olefin radical cyclizations: Involvement of a samarium tridentate ligate

Article abstract of DOI:10.1021/ja963195c

High diastereoselection in a samarium(II) iodide-promoted ketyl-olefin cyclization reaction has been achieved using tartramide-derived keto allylic acetals as chiral auxiliaries. The unique features of the reaction include the fact that remote diastereose

Full text of DOI:10.1021/ja963195c

J. Am. Chem. Soc. 1997, 119, 1265-1276
1265
Remote Stereochemical Control of Both Reacting Centers in
Ketyl-Olefin Radical Cyclizations: Involvement of a Samarium
Tridentate Ligate
Gary A. Molander,* J. Christopher McWilliams, and Bruce C. Noll†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Colorado,
Boulder, Colorado 80309-0215
ReceiVed September 11, 1996X
Abstract: High diastereoselection in a samarium(II) iodide-promoted ketyl-olefin cyclization reaction has been achieved
using tartramide-derived keto allylic acetals as chiral auxiliaries. The unique features of the reaction include the
fact that remote diastereoselection is achieved in a radical process and that high levels of stereochemical induction
are observed at both new stereocenters created in the transformation. The source of the asymmetric induction is
postulated to be a highly ordered, tricyclic transition structure made possible by three-point chelation between the
ketyl intermediate and the samarium counterion. As such, this transformation also demonstrates the first example
of the use of a chelating metal to affect high levels of remote asymmetric induction in a radical reaction. Because
of this chelation, the sense of relative stereoselectivity is unusual for a SmI2-mediated cyclization, providing consistently
high ratios of cis/trans isomers. A double-diastereodifferentiating experiment provides additional support for this
mechanistic hypothesis. The preparation of enantiomerically enriched cyclopentanediols and -lactols can be achieved
through this novel asymmetric cyclization protocol.
Introduction
though currently developed methods employing chiral auxiliaries
have demonstrated high stereoselectivity with regard to the
chirality at the carbon atom bearing the radical, or at the alkene
radical acceptor, simultaneous control of facial selectivity in
both species remains an unchallenged field. Therefore, the
currently available methods would necessitate a chiral auxiliary
on each of the two reacting centers.4 In heterolytic bond-
forming reactions, controlling the facial selectivity of two
reacting partners with a single chiral auxiliary can be achieved
by chelation with a Lewis acid. Applications of this approach
to radical reactions are lacking, presumably because of the
absence of appropriate functional groups in close proximity to
both reacting centers.
The asymmetric formation of carbon-carbon bonds utilizing
radical intermediates remains a challenging goal in organic
synthesis.1 In general, successful approaches to inducing
asymmetry in radical reactions have relied on chiral auxiliaries
that utilize steric bulk to block approach to one face of the
radical or radical acceptor. When the reacting center is in close
proximity to the auxiliary, this is a viable option. However, as
the reacting center becomes further removed from the auxiliary
(remote asymmetric induction), developing large auxiliaries with
predictable conformational biases becomes a much more difficult
endeavor. In heterolytic bond-forming processes, examples of
remote asymmetric induction are well precedented.2 These
transformations take advantage of the ability of Lewis acids to
organize substrates through chelation. Although the use of
Lewis acids to control conformational preferences in radical
chemistry has recently provided some impressive successes, few
examples of applications to remote asymmetric induction are
known.1d-f,h-m,3
Herein we describe a samarium-mediated radical cyclization
in which a single chiral auxiliary engenders high levels of
asymmetric induction at both reacting carbon centers. High
(2) For recent examples, see: (a) Molander, G. A.; Haar, Jr., J. P. J.
Am. Chem. Soc. 1993, 115, 40. (b) Molander, G. A.; Bobbitt, K. L. J. Am.
Chem. Soc. 1993, 115, 7517. (c) Teerawutgulrag, A.; Thomas, E. J. J. Chem.
Soc., Perkin Trans. 1 1993, 2863. (d) Carey, J. S.; Coulter, T. S.; Thomas,
E. J. Tetrahedron Lett. 1993, 34, 3933. (e) Carey, J. S.; Thomas, E. J.
Tetrahedron Lett. 1993, 34, 3935. (d) Molander, G. A.; Bobbitt, K. L. J.
Org. Chem. 1994, 59, 2676. (e) Carey, J. S.; Thomas, E. J. J. Chem. Soc.,
Chem. Commun. 1994, 283. (f) Stanway, S. J.; Thomas, E. J. J. Chem.
Soc., Chem. Commun. 1994, 285. (g) Ahzng, H.-C.; Costanzo, M. J.;
Maryanoff, B. E. Tetrahedron Lett. 1994, 35, 4891. (h) Hallett, D. J.;
Thomas, E. J. J. Chem. Soc., Chem. Commun. 1995, 657. (i) Gruttadauria,
M.; Thomas, E. J. J. Chem. Soc., Perkin Trans. 1 1995, 1469. (j) Stanway,
S. J.; Thomas, E. J. Synlett 1995, 214. (k) Stanway, S. J.; Thomas, E. J.
Tetrahedron Lett. 1995, 36, 3417.
(3) For recent examples of the application of Lewis acids to stereose-
lective radical transformations, see: (a) Renaud, P.; Ribezzo, M. J. Am.
Chem. Soc. 1991, 113, 7803. (b) Guindon, Y.; Lavelle´e, J.-F.; Llinas-Brunet,
M.; Horner, G.; Rancourt, J. J. Am. Chem. Soc. 1991, 113, 9701. (c)
Feldman, K. S.; Romanelli, A. L.; Ruckle, Jr., R. E. J. Org. Chem. 1992,
57, 100. (d) Renaud, P.; Moufid, N.; Kuo, L. H.; Curran, D. P. J. Org.
Chem. 1994, 59, 3547. (e) Renaud, P.; Bourquard, T.; Gerster, M.; Moufid,
N. Angew. Chem., Int. Ed. Engl. 1994, 33, 1601. (f) Guindon, Y.; Gue´rin,
B.; Chabot, C.; Mackintosh, N.; Ogilvie, W. W. Synlett 1995, 449. (g) Sibi,
M. P.; Ji, J.; Hongliu, W.; Gu¨rtler, S.; Porter, N. A. J. Am. Chem. Soc.
1996, 118, 9200 and references therein.
Another issue that is unresolved in asymmetric carbon-
carbon bond formation utilizing radical intermediates is that of
controlling the stereochemistry at both reacting centers. Al-
† Author to whom correspondence should be addressed concerning the
X-ray crystal structure.
X Abstract published in AdVance ACS Abstracts, February 1, 1997.
(1) For reviews of diastereoselection in radical reactions, see: (a) Smadja,
W. Synlett 1994, 1. (b) Porter, N. A.; Giese, B.; Curran, D. P. Acc. Chem.
Res. 1991, 24, 296. See also: (c) Snider, B. B.; Zhang, Q. Tetrahedron
Lett. 1992, 33, 5921. (d) Toru, T.; Watanabe, Y.; Tsusaka, M.; Ueno, Y. J.
Am. Chem. Soc. 1993, 115, 10464. (e) Yamamoto, Y.; Onuki, S.; Yumoto,
M.; Asao, N. J. Am. Chem. Soc. 1994, 116, 421. (f) Nishida, M.; Ueyama,
E.; Hayashi, H.; Ohtake, Y.; Yamaura, Y.; Yanaginuma, E.; Yonemitsu,
O.; Nishida, A.; Kawahara, N. J. Am. Chem. Soc. 1994, 116, 6455. (g)
Curran, D. P.; Geib, S. J.; Lin, C.-H. Tetrahedron: Asymmetry 1994, 5,
199. (h) Andrus, M. B.; Argade, A. B.; Chen, X.; Pamment, M. G.
Tetrahedron Lett. 1995, 36, 2945. (i) Urabe, H.; Yamashita, K.; Suzuki,
K.; Kobayashi, K.; Sato, F. J. Org. Chem. 1995, 60, 3576. (j) Murakata,
M.; Tsutsui, H.; Hoshino, O. J. Chem. Soc., Chem. Commun. 1995, 481.
(k) Renaud, P.; Gerster, M. J. Am. Chem. Soc. 1995, 117, 6607. (l) Sibi,
M. P.; Jasperse, C. P.; Ji, J. J. Am. Chem. Soc. 1995, 117, 10779. (m) Wu,
J. H.; Radinov, R.; Porter, N. A. J. Am. Chem. Soc. 1995, 117, 11029.
(4) Presumably, this would also necessitate a matching between the
chirality of the auxiliaries.
S0002-7863(96)03195-2 CCC: $14.00 © 1997 American Chemical Society

1266 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
a
Table 1. Cyclization of 1 with SmI2
Scheme 1
additive
ratio of 5:6:7:8b
none
t-BuOH
61:13:11:15
07:00:56:37
a Reactions were run by adding a solution of 1 to a preformed
solution of SmI2 with HMPA in THF at room temperature. b Ratios
were determined by GLC, uncorrected for response factors.
to reverse the inherent diastereoselectivity associated with SmI2
cyclizations. Furthermore, if this ligating functionality was
present in a chiral environment, it would be possible to control
the absolute stereochemistry of the final product as well.
Our initial results with the cyclization of 1 implied that a
chiral allylic alcohol or possibly an ether might provide the
desired stereocontrolling element. However, the use of such
functionalities has several disadvantages associated with it,
including the difficulty of incorporating these moieties into the
substrate, and dissociating them from the final product. With
this in mind, we began our investigation with a series of chiral
acetals. Acetals offer the advantages of ease of incorporation
into the starting material, ease of removal from the products
under conditions of protonation or elimination, and of being
readily available in diverse forms. Furthermore, chiral acetals
have been successful in directing asymmetric transformations
with a variety of organometallics.9 To our knowledge, no
previous application of chiral acetals to organolanthanide
chemistry has been demonstrated.
levels of remote asymmetric induction are observed at the radical
center furthest removed from the auxiliary. The source of the
observed stereoselectivity is postulated to be a highly ordered,
tricyclic transition structure made possible by three-point chela-
tion between the ketyl intermediate and the samarium cation.
Background
The SmI2-mediated cyclization of keto olefins has been
established as an excellent method for the synthesis of trans-
1,2-dialkylcyclopentanols.5 Initially, we were interested in
extending this method to the cyclization of keto allylic alcohols,
providing products with functionality readily amenable to further
manipulation. Utilizing keto allylic alcohol 1 as an example
(Scheme 1), four possible products could arise from cyclization
with SmI2. Thus, initial reduction with SmI2 generates the ketyl
radical 2. Cyclization of 2 to 3, followed by further reduction
generates an intermediate organosamarium 4. The intermediate
4 can either protonate to form the diols 5 and 6 or undergo
elimination to the alkenes 7 and 8.
Ideally, it would be desirable to develop conditions to produce
each of the four possible products selectively. We had expected
that formation of the trans isomers (5 and 7) would be preferred
to the cis isomers (6 and 8) based upon substantial precedent.
Earlier results from these laboratories on similar substrates
indicated the presence or absence of proton sources (i.e.,
alcohols) had a distinct effect on the outcome of reactions
conducted with intermediates analogous to 4.6 Indeed, cycliza-
tion of 1 in the presence of tert-butyl alcohol demonstrated
selectivity for alkene formation (Table 1). In the absence of
added alcohol, the diols were formed preferentially. However,
the stereoselectivity of the cyclizations was very low, in stark
contrast with the usual high selectivity observed with SmI2
cyclizations of keto olefins.5 The implication of these results
was that the hydroxyl group was somehow involved in the
stereodifferentiating, bond formation step.7,8 This raised the
possibility of employing a suitably disposed ligating auxiliary
Results and Discussion
A representative series of keto allylic acetals was prepared
from commercially available keto ester 9 (Scheme 2). Dithiane
protection of the ketone functionality of 9 gave 10.10 Reduction
of the ester, followed by oxidation,11 provided the aldehyde 11.
A two-step homologation procedure yielded the enal 12.12
Direct acetalization, or transacetalization, from 12, followed by
deprotection to the ketone, yielded the dimethyl- (13), diester-
(14), and diamide-substituted (15a) keto allylic acetals.13
Cyclization of the dimethyl-substituted acetal 13 with SmI2/
HMPA yielded mixtures of cyclopentanol products (Scheme 3).
At low temperature (-78 °C), cyclization proceeded without
elimination, producing the cyclopentanols with the dioxolane
intact (16 and 17). At room temperature, elimination products
(18 and 19) were observed along with 16 and 17. Both 18 and
19 were mixtures of E- and Z-enol ethers. Although the relative
stereoselectivity was typical for SmI2-THF/HMPA cyclizations,
no asymmetric induction from the chiral auxiliary was ob-
served.14 In the absence of HMPA, attempts to initiate
cyclization of 13 with SmI2 was met with low reactivity and
(8) Other examples in which a pendant hydroxyl group influences the
stereoselectivity of SmI2-mediated reactions are known. For examples of
ketyl-olefin coupling, see ref 5e,h. (a) Kawatsura, M.; Matsuda, F.;
Shirahama, H. J. Org. Chem. 1994, 59, 6900. For a related example, see:
(b) Inanaga, J.; Ujikawa, O.; Handa, Y.; Otsubo, K.; Yamaguchi, M. J.
Alloys Compounds 1993, 192, 197.
(9) For a review, see: (a) Alexakis, A.; Mangeney, P. Tetrahedron:
Asymmetry 1990, 1, 477. For more recent work in this field, see: (b)
Sammakia, T.; Smith, R. S. J. Am. Chem. Soc. 1992, 114, 10998, and
references cited therein. (c) Aube´, J.; Heppert, J. A.; Milligan, M. L.; Smith,
M. J.; Zenk, P. J. Org. Chem. 1992, 57, 3563. (d) Parrain, J.-L.; Cintrat,
J.-C.; Qintard, J.-P. J. Organomet. Chem. 1992, 437, C19. (e) Longobardo,
L.; Mobbili, G.; Tagliavini, T.; Trombini, C.; Umani-Ronchi, A. Tetrahedron
1992, 48, 1299. (f) For a recent example of the application of chiral acetals
to asymmetric cyclopropanations, see: Armstrong, R. W.; Maurer, K. W.
Tetrahedron Lett. 1995, 36, 357.
(5) (a) Molander, G. A.; Kenny, C. Tetrahedron Lett. 1987, 28, 4367.
(b) Molander, G. A.; Kenny, C. J. Am. Chem. Soc. 1989, 111, 8236. (c)
Molander, G. A.; Kenny, C. J. Org. Chem. 1991, 56, 1439. (d) Molander,
G. A.; McKie, J. A. J. Org. Chem. 1992, 57, 3132. (e) Kito, M.; Sakai, T.;
Yamada, K.; Matsuda, F.; Shirahama, H. Synlett 1993, 158. (f) Molander,
G. A.; McKie, J. A. J. Org. Chem. 1994, 59, 3186. (g) Kan, R.; Nara, S.;
Ito, S.; Matsuda, F.; Shirahama, H. J. Org. Chem. 1994, 59, 5111. (h) Kan,
T.; Hosokawa, S.; Nara, S.; Oikawa, M.; Ito, S.; Matsuda, F.; Shirahama,
H. J. Org. Chem. 1994, 59, 5532. (i) Molander, G. A.; McKie, J. A. J.
Org. Chem. 1995, 60, 872.
(10) Seebach, D.; Corey, E. J. J. Org. Chem. 1975, 40, 231.
(11) Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651.
(12) Cresp, T. M.; Sargent, M. V.; Vogel, P. J. Chem. Soc., Perkin Trans.
1 1974, 37.
(6) Unpublished results.
(7) Alternative explanations, such as an equilibrating cyclization process,
although less probable, cannot be excluded.
(13) The diamide acetal 15 was produced as a 9:1 mixture of E:Z
stereoisomers. The mixture could be enriched by silica gel chromatography.

Ketyl-Olefin Radical Cyclizations
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1267
Scheme 2a
Scheme 4
a (a) 1,2-Ethanediol, 0.4 equiv of ZnCl2, CH2Cl2, 93%; (b) LiAlH4,
THF, 100%; (c) (COCl)2, DMSO, Et3N, CH2Cl2, 68%; (d) (1,3-
dioxolan-2-yl-methyl)triphenylphosphonium bromide, KH, THF/
HMPA, 75% (85:15 Z:E); (e) Amberlyst-15, H2O, acetone, 96%; (f)
(2R,3R)-2,3-butanediol, p-TsOH, benzene reflux, 60%; (g) diethyl-L-
tartrate, p-TsOH, benzene, reflux, 66%; (h) HC(OMe)3, dry Amberlyst,
82%; (i) N,N,N′,N′-tetramethyl-L-tartaramide, PPTS, toluene, 4Å mol.
sieves, reflux, 86%; (j) NCS, AgNO3, collidine, 9:1 CH3CN/H2O, 13
(65%), 14 (53%), 15a (73%).
Scheme 3
the cyclization of 14 is excellent, providing the cis isomers (23
+ 24) with g98:2 selectivity.16 This result represents a
complete reversal of the usually observed trans stereoselectivity
associated with SmI2-mediated ketyl-olefin cyclizations. In
terms of asymmetric induction from the chiral auxiliary, the
cyclization of 14 provided an 85:15 ratio of 23 and 24,
respectively. Although this level of stereoinduction is modest,
this proVides the first example in which the stereochemistry at
both reacting centers of a radical cyclization is controlled by a
single chiral auxiliary. Finally, the high stereoselectivity for
the E-enol ether in both 23 and 24 is intriguing. When
compared to the cyclization of the acetal 13, this indicates
another level of stereoselectivity associated with the diester-
substituted acetal 14 that is not present in the simple dimethyl-
substituted acetal system.
The diamide-substituted acetal 15a was envisioned to be an
improvement over the diester-substituted acetal 14 for two
reasons. First, the amide functionalities should be better ligands
for samarium ions, forming closer contacts with the samarium
metal center. Shorter bond lengths in the transition structure
should accentuate energetic differences in competing cyclization
pathways. Secondly, an amide functionality would be less prone
to R-C-O bond cleavage in the product enol ether.15 In line
with these predictions, cyclization of 15 with SmI2 in the
absence of HMPA proceeded with excellent relative stereo-
chemistry and stereoinduction from the chiral auxiliary (Scheme
4). The cis (27a and 28a) and trans (29a and 30a) isomers
were formed with g99:1 stereoselectivity, respectively. The
ratio of the two cis isomers (27a and 28a) was 97:3. The sense
of relative and absolute stereochemistry was confirmed by
chemical correlation (Vide infra) and a single crystal X-ray
structure of 27a (Figure 1). No R-C-O bond cleavage was
observed, and the overall combined yield of cyclization products
was 79%. Not surprisingly, the diamide-substituted acetal
the observation of additional, uncharacterized products. It was
quite apparent from these results that a simple acetal functional-
ity, combined with modest steric biases far removed from the
reacting center, would have an insignificant influence on the
stereochemistry of the cyclization.
We then turned to the diester-substituted acetal 14, which
offered additional sites for metal ligation (i.e., the ester
functionalities). Submitting 14 to typical cyclization conditions
with SmI2 in the presence of HMPA yielded intractable
mixtures. However, we were pleased to discover that in the
absence of HMPA, 14 cyclized readily, yielding products 23-
26 (Scheme 4). Indeed, the enhanced reactivity of 14 toward
SmI2 was the first observation that distinguished this substrate
from the acetal 13. Although 13 displayed incomplete conver-
sion after 24 h at room temperature, the cyclization of 14 was
complete within 1 h at room temperature. A small amount of
R-C-O bond cleavage of the enol ether was observed in the
latter reaction. In spite of this, an overall combined yield of
70% could still be obtained in the cyclization of 14.15
Several points concerning the stereoselectivity in the cycliza-
tion of 14 are worthy of note. The relative stereoselectivity in
(14) Stereoselectivity was determined by submitting crude reaction
mixtures to p-TsOH in toluene heated at reflux. These conditions yielded
quantifiable mixtures of 16 and 17 from the cyclization of enol ethers, 18
and 19. The relative stereoselectivity was assumed to be trans based upon
the cyclization of the achiral keto allylic acetal 20, which gave a 7:3 mixture
of E- and Z-enol ethers rac-21. Submitting mixtures of 16-19 or rac-21 to
hydrolysis conditions yielded the hydroxy aldehyde rac-22. The corre-
sponding cis isomers would be expected to give a cyclic hemiacetal (Vide
infra).
(15) Evidence for R-C-O cleavage of the enol ether bond comes from
the observation of diethyl malate in crude reaction mixtures. We believe
this process to be the source of several side reactions leading to intractable
mixtures when 14 is submitted to stronger reducing conditions (i.e., SmI2/
THF-HMPA). (a) Molander, G. A.; Hahn, G. J. Org. Chem. 1986, 51, 1135.
(b) Molander, G. A. Org. React. 1994, 46, 211.
(16) The trans isomers, 25 and 26, made up less than 2% of the product
mixture. The minor isomers were not isolated or characterized, and the ratio
of 25:26 is unknown.

1268 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
Scheme 6a
Figure 1. X-ray structure of 27a.
a (a) n-BuLi, 1-chloro-3-iodopropane, DMSO/THF, 49%; (b) H2,
Lindlar catalyst, EtOAc, 97%; (c) N,N,N′,N′-tetramethyl-L-tartaramide,
PPTS, DMF, 89%; (d) NaI, NaHCO3, Na2SO3, acetone, 94%; (e) Zn,
TMSCl (catalyst), 1,2-dibromoethane (catalyst), THF, 90%; (f)
CuCN‚LiCl, then isobutyryl chloride or 2,4-dimethoxybenzoyl chloride,
27% (15c), 21% (15g).
Scheme 5a
of isopropylmagnesium chloride or DIBAH (the latter was used
in an attempt to generate the aldehyde).21 Syntheses of these
derivatives are described below. As previously noted, transac-
etalization of allylic methyl acetals with N,N,N′,N′-tetramethyl-
L-tartaramide in toluene heated at reflux provides products with
only modest E/Z-stereoselectivity. Thus, the desired allylic
diamide acetals (15a,b,d,f) were prepared by the method of
Noyori20 or by our newly developed transacetalization proto-
col.22 This latter method was found to give the best yields,
with little loss in stereoselectivity when compared to the Noyori
method.
Scheme 6 outlines our syntheses of the isopropyl- and 2,4-
dimethoxyphenyl-substituted derivatives, 15c and 15g, respec-
tively. This route takes advantage of organozinc chemistry
developed by Knochel.23 The commercially available propi-
olaldehyde diethyl acetal 41 was monoalkylated with 1-chloro-
3-iodopropane, yielding 42.24 After partial reduction under
Lindlar conditions, the olefin 43 was obtained. Transacetal-
ization to 44, followed by a Finkelstein reaction, provided the
pivotal iodide 45. Conversion to the organozinc reagent 46 by
the method of Knochel proceeded to 90% conversion.23,25 After
transmetalation to the copper species, the appropriate acid
chlorides were added to yield the isopropyl and 2,4-dimethox-
yphenyl ketones, 15c and 15g, respectively.26
a (a) AlMe3, Me(MeO)NH2Cl, CH2Cl2, 91%; (b) (COCl)2, DMSO,
Et3N, CH2Cl2, 76%; (c) LDA (2 equiv), acetaldehyde tert-butylimine,
THF, then diethyl chlorophosphate (1 equiv), then 33, then oxalic acid,
38%; (d) TMSOMe, TMSOTf (catalyst), CH2Cl2, 96%; (e) RMgBr or
RLi, THF, 63-80%; (f) N,N,N′,N′-tetramethylbis(O-trimethylsilyl)-L-
tartaramide, TMSOTf, CH2Cl2, -20 °C, 41-57%, or N,N,N′,N′-
tetramethyl-L-tartaramide, PPTS (catalyst), DMF, g79%.
15a reacted faster than the corresponding diester-substituted
acetal 14.
Scope and Limitations
Cyclization of unactivated keto olefins with SmI2-THF/
HMPA provides trans cyclopentanols as the major products.5
The relative stereoselectivity in these transformations is excellent
when the steric bulk around the ketone is small but decreases
with increasing alkyl substitution R to the carbonyl.5i We set
out to determine what the effect of changing the steric and
electronic properties around the carbonyl functionality might
be on the reactivity and stereoselectivity of this related, but
distinctly different, transformation in which the olefin was
substituted with acetals.
Syntheses of representative unsaturated carbonyl substrates
are outlined in Schemes 5, 6, and 8. Scheme 5 depicts a route
in which diversity is incorporated through Weinreb amide
intermediate 35.17 Starting with commercially available δ-vale-
rolactone 31, Weinreb amidation provided an acid- and base-
sensitive hydroxyamide 32.18 Swern oxidation of 32 yielded
the aldehyde 33.11 Using the procedure developed by Meyers,
33 was homologated to the enal 34.19 Acetalization of 34 by
the method of Noyori provided the pivotal intermediate 35.20
Conversion of 35 to a variety of ketones (36-40) was generally
successful, although low yields were obtained with the addition
The final substrate we wished to examine was an aldehyde
derivative. We had hoped to acquire the aldehyde precursor
48 from a dithiane alkylation of the iodide 45 (Scheme 7).
However, treating the lithium salt of 47 with 45 afforded
products that displayed NMR resonances consistent with mono
and bis-addition (49 and 50, respectively) of 47 to the amide
functionalities of 45. No evidence of iodide displacement from
45 was observed. Presumably, the amides direct the addition
by prior ligation to the lithium cation.27
(21) These comments should be taken in the context of single run,
unoptimized experiments.
(22) Molander, G. A.; McWilliams, J. C., Tetrahedron Lett. 1996, 40,
7197.
(23) (a) Knochel, P.; Yeh, M. C. P.; Berk, S. C.; Talbert, J. J. Org. Chem.
1988, 28, 351. (b) For a recent review, see: Knochel, P.; Singer, R. D.
Chem. ReV. 1993, 93, 2117.
(24) Chong, J. M.; Wong, S. Tetrahedron Lett. 1986, 27, 5445.
(25) Determined by GC after hydrolysis of the crude mixture (uncorrected
for response factors).
(26) The isolated yields of 15c and 15g were low overall, but this is a
likely reflection of the isolation process. The relatively polar nature of the
diamide acetals tends to swamp out subtleties in polarity differences between
the impurities of this transformation (e.g., the chloride 44, iodide 45, and
protonated material are all present in the crude mixture). For 15c, isolation
from the crude mixture required a reduction to an isolable mixture of
epimeric alcohols, followed by oxidation back to 15c. It should be noted
that these yields represent single run, unoptimized reactions.
(27) We have found that lithium anions derived from hydrazones will
alkylate 45 in good yield at low temperatures (-78 °C).
(17) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.
(18) (a) Lipton, M. F.; Basha, A.; Weinreb, S. M. In Organic Syntheses;
Noland W. E., Ed.; John Wiley & Sons: New York, 1988; Collect. Vol. 6,
pp 492-495. (b) Basha, A.; Lipton, M.; Weinreb, S. M. Tetrahedron Lett.
1977, 4171.
(19) Meyers, A. I.; Tomioka, K.; Fleming, M. P. J. Org. Chem. 1978,
43, 3788.
(20) Tsunoda, T.; Suzuki, M.; Noyori, R. Tetrahedron Lett. 1980, 21,
1357.

Ketyl-Olefin Radical Cyclizations
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1269
Scheme 7
levels relative to the Sm(III)-ketyl radicals derived from aliphatic
ketones.29 Because the alkene LUMO remains unchanged for
all the substrates, the net result is that conjugated Sm(III)-ketyl
radicals have less orbital overlap in the transition structure (i.e.,
less stabilization).30
We expected that additional alkoxy functionalities at the ortho
and para positions of the aromatic ring in 15g would increase
the SOMO energy of the resultant Sm(III)-ketyl radical relative
to 15f and, therefore, enhance reactivity. To test our hypothesis,
the 2,4-dimethoxyphenyl ketone 15g was submitted to cycliza-
tion with SmI2. In this event, 15g did display improved
reactivity toward cyclization (Table 2), albeit with lower yield
and stereoselectivity when compared to the aliphatic substrates.
A small increase in the temperature of cyclization (40 °C instead
of 23 °C) yielded cleaner reaction mixtures. Interestingly, we
found that utilizing a large excess of SmI2 resulted in an increase
in the diastereoselectivity.31 We believe these observations are
consistent with the mechanism outlined in Scheme 10. An
initial reduction of 15g with SmI2 yields the Sm(III)-ketyl radical
53. The cyclization of 53 to the intermediate carbon radicals
54 or 55 is a reversible process. It is this reversible step that
differentiates 15g from the aliphatic substrates 15a-d.32 The
intermediates 54 and 55 are removed from the equilibrium by
irreversible reduction to the organosamarium(III) species, 56
and 57. At lower temperatures, the intermediate Sm(III)-ketyl
radical 53 has a longer lifetime and can undergo side reactions
leading to other products. When excess SmI2 is present, the
rates of the irreversible reductions leading to 56 and 57 increase
relative to the unimolecular equilibration of 54 and 55. Thus,
at higher temperatures and higher concentrations of SmI2, the
cyclization of 15g proceeds more cleanly and with greater
diastereoselectivity.
We examined the potential for aldehyde substrate 15h to
undergo stereoselective cyclization with SmI2. In this event,
good relative stereoinduction (98:2) and stereoinduction from
the chiral auxiliary (95:5) was observed (Table 2). The
combined yield for cyclization products was, however, only
modest. Side reactions giving pinacol and reduction products
appeared to be the source of diminished yields.33 Optimized
conditions consisted of slow addition of 15h to an excess of
SmI2, thereby minimizing intermolecular side reactions.
The distinct features of this cyclization process have provided
the basis for a reasonable mechanistic rationale. Although the
overall diastereoselectivity decreases with increasing steric bulk
around the ketyl center, the high relative cis-stereoselectivity
of these cyclizations appears to be independent of the size of
the ketone substituent. This is in stark contrast to the trans-
selective SmI2 cyclizations of unactivated keto olefins.5 There
is a large acceleration in the rate of cyclization when the diester-
substituted acetal 14 is compared with the nonselective dimethyl-
substituted acetal 13. An additional, less dramatic, increase in
Scheme 8a
a (a) NaI, NaHCO3, Na2SO3, acetone, 93%; (b) 1,3-dithiane, n-BuLi,
then 41, THF, 65%; (c) N,N,N′,N′-tetramethyl-L-tartaramide, PPTS,
DMF, 71%; (d) bis(trifluoroacetoxy)iodobenzene, 9:1 CH3CN/H2O,
76%.
To circumvent the problems associated with dithiane alky-
lation of 45, we turned to alkylating the iodide 51, derived from
a precursor lacking the diamide acetal functionality 43 (Scheme
8). After Finkelstein conversion of 43 to the iodide 51,
alkylation proceeded without incident, providing 52. Transac-
etalization of 52 to the diamide acetal 48, followed by oxidative
removal of the dithiane functionality,28 yielded the aldehyde
substrate 15h.
With the desired substrates in hand, an investigation of the
ketyl olefin cyclization reactions was begun. Addition of the
keto allylic acetals (15a-h) to a THF solution of SmI2 (Scheme
9) provided the results tabulated in Table 2. All of the aliphatic
ketones cyclized to cis-cyclopentanols in good yields. Notably,
even the bulky tert-butyl derivative 15d cyclized readily without
the need for HMPA as an additive. The relative stereoselectivity
was excellent, providing high cis/trans ratios for the entire series
of aliphatic ketone substrates. Again, the tert-butyl derivative
15d showed little deviation from the smaller homologues. In
terms of stereoinduction from the chiral auxiliary, some decrease
was observed as the steric bulk around the ketone increased.
The unsaturated ketones 15e-g did not show the same
reactivity and stereoselectivity as their aliphatic counterparts.
The propargylic ketone substrate 15e appeared to give only
pinacol coupling products. The unsubstituted phenyl substrate
15f also indicated little propensity for cyclization, although some
cyclopentanol products could be discerned from the crude
reaction mixtures.
(29) The stability of Sm(III)-ketyls derived from aromatic ketones is
exemplified by the isolation and X-ray determination of the Sm(III)-ketyl
derived from fluorenone: Hou, Z.; Miyano, T.; Yamazaki, H.; Wakatsuki,
Y. J. Am. Chem. Soc. 1995, 117, 4421.
(30) An analogous argument could be stated as a decrease in the electron
density coefficient at the ketyl carbon orbital because of delocalization into
the π-system.
(31) For example, the use of 2.7 and 8 equiv of SmI2 yielded 60:40 and
75:25 mixtures of 59/60, respectively. Also, the reactions appeared cleaner
overall under these conditions.
(32) The cyclization of phenyl-substituted ketyl radicals has been shown
to be a reversible process. Furthermore, the Sm(III)-ketyl derived from
fluorenone was found to undergo reversible pinacol coupling: ref 29.
(33) When comparing the aldehyde substrate 15h to the aliphatic ketone
substrates 15a-d, the rate of intermolecular pinacol coupling relative to
cyclization is probably increased by diminished steric hindrance around
the Sm(III)-ketyl, and this effect is compounded by the diminished
nucleophilicity of the Sm(III)-ketyl.
It seems unlikely that the low reactivity of substrates 15e-g
can be attributed to a high energy barrier in the initial reduction
to the Sm(III)-ketyl radical, especially when compared to the
aliphatic ketones which should be more resistant to reduction.
The poor reactivity of substrates 15e-f may be a consequence
of relatively large energy gaps between the SOMO of the Sm(III)
-
ketyl radical and the LUMO of the alkene. The conjugated
Sm(III)-ketyl radicals in 15e-f likely have lower SOMO energy
(28) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 287.

1270 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
Scheme 9
Table 2. Asymmetric Cyclization of Tartrate-Derived Keto Allylic
Of the chair conformers in Chart 1, only 58 and 61 lead to
the observed major diastereomer 27. A distinct difference
between conformers 58 and 61 is the orientation of the acetal-
bearing carbon with respect to the alkene. In 58, the acetal-
bearing carbon is poised in an eclipsed conformation, but a
bisected conformation is adopted in 61. For propene, the
eclipsed conformer has been calculated to be 2 kcal/mol lower
in energy than the bisected conformer.35 It is also interesting
to note that elimination (following further reduction of the
subsequently formed carbon radical) from any of the eclipsed
conformations would provide the trans enol ether directly,
whereas formation of the trans enol ether from the bisected
conformer 61 would require a breakdown of the ligated structure
in order to allow for bond rotation prior to the elimination
event.36 Furthermore, the samarium center in 61 is located on
what may be considered the concave side of the acetal unit,
whereas the samarium center occupies the convex side in 58.
We have reason to believe that the concave side of the acetal is
a less stable environment for the samarium ion and its attached
ligands (Vide infra). Therefore, the chair conformer 58, with
an eclipsed conformation about the alkene-acetal carbon and
with the samarium center located on the convex face of the
acetal, likely represents the conformation of the transition
structure leading to the major product 27.
Conformer 60 represents the only conformation in Chart 1
that could lead to the minor diastereomer 28. A bisected
conformer analogous to 60 is not likely, for it would place the
samarium center too far from the amide located on the convex
face of the acetal for any significant bonding interaction. It is
not obvious why conformer 58 is favored over that of conformer
60. The distinct difference between 58 and 60 resides in the
spatial placement of the samarium center. Although the metal
center is located on the convex side of the acetal in 58, it is on
the concave side of the acetal in 60. This suggests that locating
the samarium center on the concave side of the acetal is
relatively destabilizing. It is possible that there are more
destabilizing van der Waals interactions between the ligands
attached to samarium and the substrate on the concave side.37
Conformers 62 and 63 represent those that would lead to trans
isomers 29 and 30, respectively. Although some bonding
between the amide oxygen and samarium may be possible in
62, it would be very weak at best. In conformer 63, no
interaction between the amide and samarium is possible.
Because the carbonyl functionalities serve to increase the
oxidation potential of samarium, and it is likely this activation
leads to enhanced reactivity, very little of the trans isomers
Acetalsa
substrate
R
E:Zb
% yieldc
cis/transd
27:28
14
Me
Me
Et
98:2
98:2
98:2
97:3
98:2
98:2
98:2
98:2
98:2
70
79
83
82
86
0
98:2e
99:1
97:3
97:3
97:3
N/A
N/A
g
85:15f
97:03
97:03
93:07
91:09
N/A
15a
15b
15c
15d
15e
15f
15g
15h
i-Pr
t-Bu
1-hexynyl
Ph
Arh
H
g
N/A
43
53
80:20j
95:05
98:2
a Ratios of isomers (E/Z, cis/trans, 27-30) were determined by GC
(uncorrected for response factors). N/A ) not applicable. b Ratio of
alkene isomers in 14, 15a-h. c Combined isolated yield. d Ratio of (27
+ 28)/(29 + 30). e Ratio of (23 + 24)/(25 + 26). f Ratio of 23/24.
g Not determined. h Ar ) 2,4-dimethoxyphenyl. i Determined by 13C
NMR.
the rate of cyclization is observed when comparing the diamide-
substituted acetal 15a with the less selective diester-substituted
acetal 14. Finally, this cyclization process yields the E-enol
ethers exclusively. We believe these observations are consistent
with a tridentate-ligating transition structure in which the ketyl
oxygen, an ether oxygen, and one of the carbonyl functionalities
of the acetal are ligated to the samarium center. Several possible
transition structure conformers that fit this criteria are shown
in Chart 1.
Of the transition structures in Chart 1, conformers 58, 59,
and 61 would yield the observed major cis diastereomer.
Conformer 60 is the only transition structure in Chart 1 that
would yield the observed minor cis diastereomer. The last two
conformers in Chart 1, 62 and 63, are those that would lead to
the trans diastereomers. Only chair and boat conformers are
considered for the cyclopentanol ring formation. The only boat
conformer represented in Chart 1 is 59. With regard to ligation
around the samarium metal center, and to the resultant diaste-
reoselectivity, conformer 59 is essentially equivalent to its chair
counterpart 58. For clarity, we have omitted analogous boat
conformations that would correspond to conformers 60-63. We
believe this is justified based on the following argument.
Although there are no flagpole interactions in 59, two destabiliz-
ing torsional interactions (one consisting of eclipsing carbon-
hydrogen bonds and the other a carbon-hydrogen bond
eclipsing a carbon-carbon bond) more than compensate for any
relief gained from the loss of the two 1,3-diaxial interactions
in 58. Similar arguments can be made for equivalent boat
conformations of 60-63. Therefore, for the consideration of
only those low energy conformers having a distinctly different
ligand sphere around samarium and those leading to different
diastereomers, conformers 58 and 60-63 should suffice.34
(35) (a) Wiberg, K. B.; Martin, E. J. Am. Chem. Soc. 1985, 107, 5035.
(b) Dorigo, A. E.; Pratt, D. W.; Houk, K. N. J. Am. Chem. Soc. 1987, 109,
6591.
(36) This is assuming an E2 elimination pathway.
(34) In reality, a Boltzmann distribution of all reactive conformers would
be more accurate, but with only semiquantitative approximations (i.e., energy
approximations based upon correlation to related interactions), this is
unworkable.
(37) Although greater interaction between the R group of the ketyl with
the ligands around samarium is suggested, it is difficult to rationalize a
switch from conformer 58 to 60 as the size of R gets larger (i.e., the
diastereoselectivity of the cyclization decreases as the size of R increases).

Ketyl-Olefin Radical Cyclizations
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1271
Table 3. Crystallographic Data for (2R,3R)-O-((E)-2-((1R,2R)-2-Hydroxy-2-methylcyclopentyl)vinyl)-N,N,N′,N′-tetramethylsuccinamide, 27a
compd no.
empirical formula
formula mass
crystal system
space group
a, Å
27a
C16H28N2O5
328.40
triclinic
P1
6.0950(10)
7.659(2)
10.042(2)
101.89(3)
93.06(3)
94.15(3)
456.4(2)
1
transmission coeffs.
T (K)
λ, Å
reflcns collected
unique reflcns
0.96 to 0.98
293(2)
0.71069 (MoKR)
1801
1667 (R(int) ) 0.0175)
1324
reflections observed
R indicesa [I > 2σ(I)]
R indicesa (all data)
weighting coeffsb
goodness-of-fitc on F2
b, Å
c, Å
R1 ) 0.0390, wR2 ) 0.0759
R1 ) 0.0549, wR2 ) 0.0818
R, deg
a ) 0.0414, b ) 0
1.020
â, deg
γ, deg
F
calc, Mg/m3
µ, mm-1
1.195
0.088
vol., Å3
Z
2
2
2
a R1 ) ∑||F | - |F ||/∑|F |; wR2 ) ∑[w(F -F 2)2]/∑[w(F )2]. b w-1 ) [σ2(Fo ) + (aP)2 + bP], where P ) (Fo2 + 2Fc2)/3. c Goodness-of-fit
x
o
c
o
o
c
o
2
)
∑[w(F -F 2)2]/(M-N) where M is the number of reflections and N is the number of parameters refined.
x
o
c
Scheme 10
Scheme 11a
a (a) Propiolaldehyde diethyl acetal, n-BuLi, DMSO/THF, 72%; (b)
n-Bu4NF, THF/H2O, 84%; (c) H2, Lindlar catalyst, quinoline EtOAc,
then MeOH, p-TsOH, 74%; (d) I2, Ph3P, imidazole, CH3CN, 38%; (e)
2-methyl-1,3-dithiane, n-BuLi, 65%; (f) N,N,N′,N′-tetramethyl-L-tart-
aramide, PPTS (catalyst), DMF, 79%; (g) N,N,N′,N′-tetramethyl-D-
tartaramide, PPTS (catalyst), DMF, 79%; (h) Chloramine-T, MeOH,
H2O, 55% (71), 50% (72).
Chart 1
yielded 64.38 Desilylation of 64 afforded 65. Partial reduction
of the alkyne 65, followed by transacetalization produced the
unsaturated methyl acetal 66. After formation of the iodide 67,
alkylation with 2-methyl-1,3-dithiane yielded the protected
ketone 68.39 From 68, the synthesis diverged by transacetal-
ization with both isomers of N,N,N′,N′-tetramethyltartaramide,
yielding the L- and D-isomers, 69 and 70, respectively. Depro-
tection of 69 and 70 provided the desired diastereomeric L- and
D-substrates 71 and 72, respectively.40
According to our mechanistic hypothesis regarding the source
of stereoselectivity in these cyclizations, diastereomeric sub-
strates 71 and 72 should represent mismatched and matched
chirality, respectively. Cyclization of 71 through conformer 73,
which is analogous to 58, places the methyl group in an
unfavorable axial orientation (Scheme 12). However, conformer
74, which is analogous to 60, places the methyl group in a
favorable equatorial orientation. Thus, we would expect to see
resulting from transition structures that proceed through con-
formers 62 and 63 are observed.
The methyl-substituted, diastereomeric keto allylic acetals 71
and 72 were prepared to test the hypothesis that 58 and 60
represent the transition structures leading to the major cis
isomers, 27 and 28, respectively (Scheme 11). Alkylation of
propiolaldehyde diethyl acetal with the known bromide 63
(38) McWilliams, J. C. Ph.D. Thesis, Cornell University, May, 1994.
(39) Seebach, D.; Corey, E. J. J. Org. Chem. 1975, 40, 231.
(40) Huurdeman, W. F. J.; Wynberg, H.; Emerson, D. W. Tetrahedron
Lett. 1971, 3449.

1272 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
Scheme 12
Scheme 13a
a decrease in the diastereoselectivity for the cyclization of 71
(i.e., the consequence of mismatched chirality). In contrast,
cyclization of 72 through conformer 77, which is analogous to
58, places the methyl group in a favorable equatorial orientation.
Conformer 78, which is analogous to 60, should be destabilized,
because this conformer would place the methyl group in an
unfavorable axial orientation. Thus, we would expect to see
enhanced diastereoselectivity in the cyclization of 72. It should
be noted that if these cyclizations proceeded through conformers
analogous to 61, rather than 58, we would expect no change in
stereoselectivity for 71 (i.e., the methyl group is equatorial in
both transition structures leading to major and minor cis
diastereomers) and a mismatched effect for 72.
The diastereomeric substrates 71 and 72 were submitted to
cyclization under identical conditions (Scheme 12). In line with
our predictions, cyclization of 71 yielded the diastereomers 75
and 76 with 88:12 diastereoselectivity, respectively (77%
isolated yield). Cyclization of 72, on the other hand, yielded
the diastereomers 79 and 80 with g99:1 diastereoselectivity
(88% isolated yield). Thus, the reactions of diastereomeric
substrates 71 and 72 do indeed represent examples of mis-
matched and matched chirality, respectively, and reinforce the
validity of the chelating model proposed.41
Another issue to be addressed was how the stereochemistry
of the alkene in the starting material affected the diastereose-
lectivity of the cyclization. With this in mind, we prepared the
Z-isomer, 85 (Scheme 13). The chloroalkyne 42 was converted
to the iodide 81 under Finkelstein conditions. Alkylating
2-lithio-2-methyl-1,3-dithiane with 81 provided 82.40 Transac-
etalization with N,N,N′,N′-tetramethyl-L-tartaramide provided the
chiral acetal 83, albeit in modest yield.42 Following deprotection
to the methyl ketone 84,43 Lindlar reduction then provided 85
as a 98:2 mixture of Z- and E-isomers, respectively.
a
(a) NaI, NaHCO3, Na2SO3, acetone, 86%; (b) n-BuLi, 2-methyl-
1,3-dithiane, THF, 92%; (c) N,N,N′,N′-tetramethyl-L-tartaramide, CSA
(cat.), 4A mol. sieves, toluene, reflux, 42%; (d) NCS, AgNO3, CH3CN/
H2O, 56%; (e) H2, Lindlar catalyst, quinoline, EtOAc, 90%.
Submitting the Z-isomer 85 to cyclization with SmI2 yielded
a 14:86:<1 ratio of 27a:28a:(29a + 30a), respectively. Thus,
cyclization of 85 proceeded with the same high degree of relative
stereoselectivity (g99:1). Although stereoinduction from the
chiral auxiliary was only modest (14:86), the major isomer 28a
was diastereomeric to the one formed upon cyclization of the
E-isomers. Accordingly, the 2% impurity of 85 in the cycliza-
tion of the E-isomer 15a should contribute 0.3% and 1.7% to
the observed formation of 27a and 28a from 15a. Thus, the
diastereoselectivity resulting from cyclization of the pure
E-isomer 15a is actually 99:1.44
Although the stereochemistry of the major isomer 27a had
been established by an X-ray determination (Vide supra), the
conclusive proof for the stereochemical assignment of the minor
isomer 28a came from chemical correlation (Scheme 14). The
enriched isomers of 27a and 28a (derived from the cyclization
of 15a and 85, respectively) were hydrolyzed to the lactols 86
(42) Low yields were generally obtained for all propargylic acetals
examined (i.e., transacetalization with N,N,N′,N′-tetramethyl-L-tartaramide
proceeded only at high temperatures and gave low yields of chiral products).
(43) Corey, E. J.; Erickson, B. W. J. Org. Chem. 1971, 36, 3553.
(44) This ratio of 27a/28a was calculated as follows: 27a/28a ) (97 -
0.3)/(3 - 1.7) ) 99/1.
(41) Interestingly, the diastereoselectivity in the cyclization of 72 (g99:
1), exceeded the isomeric purity of the starting alkene geometry (96:4).
Because transition structures for the E-isomers cannot be applied to the
Z-isomer, this implies a matched case for the transition state with which
Z-72 cyclizes to yield 79 as well.

Ketyl-Olefin Radical Cyclizations
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1273
Scheme 14a
layer was separated, and the aqueous phase was extracted with EtOAc
(2×). The combined organic extracts were washed with brine and then
concentrated. The resultant oil was diluted with THF (15 mL) and
saturated aqueous K2CO3 (15 mL) and stirred for 60 min (until residual
silyl ethers were completely hydrolyzed as judged by GC analysis).
The organic layer was separated, and the aqueous phase was extracted
with EtOAc (2×). The combined organic extracts were washed with
brine and dried over MgSO4. The resultant oil was submitted to
chromatography on silica gel (92:8 EtOAc/MeOH), yielding 15a (1.09
g, 57%): 1H NMR (400 MHz, CDCl3) δ 5.88 (dt, J ) 14.5, 6.7 Hz,
1H), 5.48 (ddt, J ) 14.5, 6.9, 1.4 Hz, 1H), 5.44 (d, J ) 6.9 Hz, 1H),
5.27 (d, J ) 5.6 Hz, 1H), 5.18 (d, J ) 5.6 Hz, 1H), 3.15 (s, 3H), 3.13
(s, 3H), 2.94 (s, 6H), 2.40 (t, J ) 7.4 Hz, 2H), 2.09 (s, 3H), 2.08-2.03
(m, 2H), 1.69-1.62 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 208.4,
168.7, 166.9, 138.1, 125.9, 106.0, 76.5, 74.4, 42.6, 37.07, 36.92, 35.71,
a (a) 2 N HCl, THF; (b) LiAlH4, THF.
and 88. Reduction of each of these mixtures gave diols (+)-
87 and (-)-87, which differed only in their optical rotations.
Thus, the two diastereomers 27a and 28a were of the same
relative configuration about the cyclopentanol but differed in
absolute stereochemistry.
The transformations depicted in Scheme 14 demonstrate the
ease with which the chiral auxiliary can be removed. Although
(-)-87 is not available in high enantiomeric purity from the
Z-isomer 85, both enantiomers of N,N,N′,N′-tetramethyltartara-
mide are commercially available. Thus, both diols can be
procured in high enantiomeric excess from the E-isomeric keto
allylic acetal (15 or its enantiomer) by selecting the correspond-
ing enantiomer of the auxiliary.
35.65, 31.1, 28.8, 22.2; IR (neat) 1713, 1650, 1503, 1149, 1056 cm-1
;
20
[R]D -19.0° (c 1.89, EtOAc).
(4R,5R)-2-((E)-6-Oxo-1-hexenyl)-N,N,N′,N′-tetramethyl-1,3-diox-
olane-4,5-dicarboxamide (15h). To a solution of 48 (2.15 g, 5.34
mmol) in 9:1 CH3CN/H2O (75 mL) was slowly added bis(trifluoroac-
etoxy)iodobenzene (3.45 g, 8.03 mmol) at 0 °C. After stirring the
mixture for 10 min, NaHCO3 (2.10 g, 25.0 mmol), followed by brine
(75 mL) were added. The aqueous phase was extracted with EtOAc
(3×). The combined organic extracts were washed with brine and dried
over MgSO4. The resultant oil was chromatographed on silica gel
(EtOAc f 93:7 EtOAc/MeOH), yielding 15h (1.27 g, 76%): 1H NMR
(400 MHz, CDCl3) δ 9.73 (broad t, J ) 1.4 Hz, 1H), 5.89 (dt, J )
15.1, 6.5 Hz, 1H), 5.50 (broad dd, J ) 15.1, 6.5 Hz, 1H), 5.45 (d, J )
6.9 Hz, 1H), 5.28 (d, J ) 5.8 Hz, 1H), 5.19 (d, J ) 5.8 Hz, 1H), 5.15
(s, 3H), 3.14 (s, 3H), 2.94 (s, 6H), 2.42 (td, J ) 7.4, 1.4 Hz, 2H),
2.13-2.07 (m, 2H), 1.75-1.61 (m, 2H); 13C NMR (400 MHz, CDCl3)
δ 202.0, 168.8, 167.0, 137.7, 126.2, 105.9, 76.5, 74.5, 42.9, 37.12,
36.98, 35.76, 35.72, 31.1, 20.6; IR (neat) 2727, 1722, 1652, 1505, 1152,
Conclusions
In conclusion, the SmI2-mediated cyclization of tartrate-
derived keto allylic acetals provides the first example of an
asymmetric radical cyclization in which high levels of stereo-
chemical induction are achieved at both reacting centers. This
transformation also demonstrates the first example of the use
of a chelating metal to effect high levels of remote asymmetric
induction in a radical reaction. The sense of relative stereose-
lectivity is also unusual for a SmI2-mediated cyclization,
providing consistently high ratios of cis/trans isomers. The
stereoselectivity observed in this cyclization appears to arise
from an intermediate tridentate ligate involving the ketyl oxygen
and the acetal auxiliary. A double-diastereodifferentiating
experiment provides additional support for this mechanistic
hypothesis.
The preparation of enantiomerically enriched cyclopen-
tanediols and -lactols can be achieved through this novel
asymmetric cyclization protocol. We are currently exploring
applications of this method to the synthesis of more complex
structures as well as the extension of this method to acyclic
substrates.
20
1059 cm-1; [R]D -21.5° (c 1.98, EtOAc).
Diethyl (2R,3R)-O-((E)-2-((1R,2R)-2-Hydroxy-2-methylcyclopen-
tyl)vinyl)succinate (23). Samarium (104 mg, 0.69 mmol) was weighed
into a flask in a glovebox. After sealing the flask with a septum, the
flask was transferred to an argon manifold. To the solid samarium
was added THF (3 mL), followed by diiodomethane (49 µL, 0.61
mmol). The mixture was stirred vigorously at room temperature for 2
h. A solution of 14 (95 mg, 0.29 mmol) in THF (3 mL) was added to
the SmI2 mixture by syringe pump over a period of 20 min. After
stirring at room temperature for an additional 40 min, a minimum of
saturated aqueous NaHCO3 was added. The mixture was filtered
through a bed of Celite and dried over K2CO3. The resultant crude oil
was chromatographed on silica gel (1:1 hexanes/EtOAc), yielding an
85:15 ratio of 23 and 24, respectively (67 mg, 70% combined yield):
1H NMR (400 MHz, CDCl3) δ 6.16 (d, J ) 12.4 Hz, 1H), 4.87 (dd, J
) 12.4, 8.5 Hz, 1H), 4.64 (d, J ) 2.2 Hz, 1H), 4.62 (d, J ) 2.2 Hz,
1H), 4.37-4.21 (m, 4H), 2.06-1.99 (m, 1H), 1.78-1.68 (m, 3H), 1.67-
1.53 (m, 3H), 1.28 (q, J ) 6.9 Hz, 6H), 1.17 (s, 3H); 13C NMR (100
MHz, CDCl3) δ 170.68, 168.0, 146.0, 106.0, 79.85, 78.62, 71.6, 62.3,
61.8, 48.8, 39.8, 30.04, 26.13, 21.22, 14.2, 14.1; IR (neat) 3500, 1747,
1732, 1668, 1652, 1264, 1200, 1156, cm-1; MS (EI+) calcd for C16H25O7
(M - H): m/e 329.1601, found 329.1586; 329 (<5 (M - H)), 190
Experimental Section
Reagents. Unless otherwise noted, all reagents were purchased from
Aldrich. Samarium metal was purchased from Cerac Inc., Milwaukee,
WI, weighed, and stored under an inert atmosphere. Diiodomethane
was purchased from Aldrich but was distilled and stored with copper
over a nitrogen atmosphere prior to use. Propiolaldehyde diethyl acetal
and δ-valerolactone were purchased from Lancaster. Tetrahydrofuran
(THF) and diethyl ether (Et2O) were distilled from benzophenone ketyl
under argon prior to use. Dimethyl sulfoxide (DMSO) and dimethyl-
formamide (DMF) were stored over 4 Å molecular sieves prior to use.
All reactions were performed under a dry argon or nitrogen atmosphere.
(4R,5R)-N,N,N′,N′-Tetramethyl-2-((E)-6-oxo-1-heptenyl)-1,3-di-
oxolane-4,5-dicarboxamide (15a). Trimethylsilyl trifluoromethane-
sulfonate (0.11 mL, 0.57 mmol) was added to 36 (1.09 g, 5.84 mmol)
and (2R,3R)-N,N,N′,N′-tetramethyl-2,3-bis(trimethylsilyloxy)succin-
bisamide (2.24 g, 6.43 mmol) in CH2Cl2 (15 mL) at -78 °C. The
solution was allowed to warm to -20 °C and stirred for 48 h at that
temperature. Pyridine (0.23 mL, 2.8 mmol) was added to the solution.
The solution was cannulated into ice-cold, saturated aqueous K2CO3,
where it was stirred at room temperature for several hours. The organic
20
(10), 117 (100), 89 (15), 71 (13), 43 (40), 29 (32); [R]D +18.6° (c
2.02, EtOAc). Diethyl (2R,3R)-O-((E)-2-((1S,2S)-2-hydroxy-2-me-
thylcyclopentyl)vinyl)succinate (24): 13C NMR (100 MHz, CDCl3)
δ 170.65, 145.88, 105.78, 78.62, 30.14, 26.23, 21.31.
General Method for the Cyclization of Keto Allylic Acetals with
SmI2. Samarium metal (Cerac, 2.6-2.8 equiv) was weighed into a
flask in a glovebox. After sealing the flask with a septum, the flask
was transferred to an argon manifold. To the solid was added THF (9
mL/mmol diiodomethane), followed by diiodomethane (2.2-2.4 equiv).
The mixture was stirred vigorously at room temperature for 2 h, during
which time a deep blue color developed. A solution of the keto allylic
acetals (0.3-0.4 mmol) in THF (6 mL/mmol) was added by syringe
pump over a period of 30-90 min at 0 °C to 35 °C. After stirring at
0 °C to 35 °C for an additional 30-60 min, Celite was added until a
viscous slurry formed. To this mixture was added saturated aqueous
NaHCO3 (2 mL), and stirring was continued for 10 min. After dilution
with CH2Cl2, the mixture was filtered through a bed of Celite. The

1274 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
filter bed was washed several times with CH2Cl2, and the combined
filtrates were dried over MgSO4. The resultant crude oil was chro-
matographed on silica gel.
(2R,3R)-O-((E)-2-((1R,2S)-2-(1,1-Dimethylethyl)-2-hydroxycyclo-
pentyl)vinyl)-N,N,N′,N′-tetramethylsuccinamide (27d). Using the
general procedure above, substrate 15d was cyclized, and the products
were isolated after chromatography (8:1 EtOAc/MeOH) to afford an
88:9:3 mixture of 27d, 28d, and (29d + 30d), respectively (86%
combined yield): 1H NMR (400 MHz, CDCl3) δ 6.17 (d, J ) 12.7
Hz, 1H), 5.01 (dd, J ) 12.7, 9.0 Hz, 1H), 4.80 (broad dd, J ) 6.7, 4.0
Hz, 1H), 4.67 (d, J ) 4.0 Hz, 1H), 3.85 (broad d, J ) 6.7 Hz, 1H
(OH)), 3.14 (s, 3H), 3.09 (s, 3H), 2.97 (s, 3H), 2.92 (s, 3H), 2.48 (dt,
J ) 7.7, 9.0 Hz, 1H), 1.95-1.87 (m, 1H), 1.73-1.38 (m, 5H), 0.90 (s,
9H); 13C NMR (100 MHz, CDCl3) δ 170.3, 168.3, 144.7, 108.6, 85.6,
79.2, 69.4, 43.3, 37.50, 36.86, 36.79, 36.27, 36.20, 35.90, 33.2, 35.9,
22.2; IR (neat) 3416, 1644, 1504, 1146, 1058 cm-1; MS (CI+ (NH3))
calcd for C19H35N2O5 (M + H): m/e 371.2511, found 371.2546; 388
(24 (M + NH4)), 371 (100 (M + H)), 353 (57), 343 (50), 327 (33);
(EI) 116 (100), 72 (53), 57 (29); [R]D20 +2.5° (c 1.79, EtOAc). Anal.
Calcd for C19H34N2O5: C: 61.60; H: 9.25; N: 7.56; found: C: 61.62;
H: 9.16; N: 7.39.
(2R,3R)-O-((E)-2-((1R,2R)-2-Hydroxy-2-methylcyclopentyl)vinyl)-
N,N,N′,N′-tetramethylsuccinamide (27a). Using the general proce-
dure above, substrate 15a was cyclized, and the products were isolated
after chromatography (6:1 EtOAc/MeOH) to afford a 96:3:1 mixture
of 27a, 28a, and (29a + 30a), respectively (81% combined yield): 1H
NMR (400 MHz, CDCl3) δ 6.18 (d, J ) 12.7 Hz, 1H), 4.92 (dd, J )
12.7, 8.4 Hz, 1H), 4.80 (broad dd, J ) 7.1, 4.7 Hz, 1H), 4.72 (d, J )
4.7 Hz, 1H), 4.10 (broad d, J ) 7.1 Hz, 1H (OH)), 3.13 (s, 3H), 3.09
(s, 3H), 2.94 (s, 3H), 2.91 (s, 3H), 2.23-1.98 (m, 1H), 1.76-1.48 (m,
6H), 1.15 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 170.4, 168.3, 145.5,
106.1, 79.73, 79.14, 69.31, 48.9, 39.9, 36.92, 36.86, 36.19, 35.92, 30.0,
26.0, 21.1; IR (neat) 3408, 1644, 1504, 1149, 1058 cm-1; MS (EI+)
calcd for C16H29N2O5 (M + H): m/e 329.2076, found 329.2070; 329
20
(<5 (M + H)), 188 (38), 116 (100), 98 (34), 72 (97), 43 (73); [R]D
+4.9° (c 2.47, EtOAc); Anal. Calcd for C16H28N2O5: C: 58.51; H:
8.59; N: 8.53. Found: C: 58.58; H: 8.83; N: 8.30.
(2R,3R)-O-((E)-2-((1R,2S)-2-(2,4-Dimethoxyphenyl)-2-hydroxy-
cyclopentyl)vinyl)-N,N,N′,N′-tetramethylsuccinamide (27g). Using
the general procedure above, substrate 15g was cyclized, and the
products were isolated after chromatography (9:1 EtOAc/MeOH) to
afford a 4:1 mixture of 27g and 28g, respectively (43% combined
yield): 1H NMR (400 MHz, CDCl3) δ 7.23 (d, J ) 8.7 Hz, 1H), 6.43
(d, J ) 2.0 Hz, 1H), 6.40 (dd, J ) 8.7, 2.0 Hz, 1H), 6.10 (d, J ) 12.8
Hz, 1H), 4.87 (dd, J ) 12.8, 7.5 Hz, 1H), 4.74 (broad m, 1H), 4.62 (d,
J ) 4.3 Hz, 1H), 3.79 (s, 3H), 3.76 (s, 3H), 3.04 (s, 3H), 3.00 (s, 3H),
2.93 (s, 3H), 2.86 (s, 3H), 2.18-2.09 (m, 1H), 1.98-1.65 (m, 6H);
13C NMR (100 MHz, CDCl3) δ 170.2, 168.2, 159.5, 157.7, 145.4, 127.1,
125.6, 106.3, 103.6, 99.1, 82.5, 79.2, 69.2, 55.2 (2 carbons), 46.5, 39.7,
36.84, 36.67, 36.21, 35.90, 29.9, 21.7; IR (neat) 3418, 1651, 1644,
1583, 1504, 1207, 1158, 1048 cm-1; MS (EI+) calcd for C23H34N2O7
(M+): m/e 450.2366, found 450.2363; 432 (7), 188 (30), 116 (100),
X-ray Crystallography for 27a. A suitable crystal was selected
and mounted on a Nicolet P3 4-circle diffractometer. Unit cell
dimensions were determined after carefully centering 24 reflections
chosen such that 20° < 2θ < 35°. Peak profiles examined from this
group indicated a large mosaic spread, and an ω scan width of 1.2°
was chosen to ensure that the entire intensity peak was recorded. Data
were collected in two shells, the first, 0° < 2θ < 40° was measured at
3.91°/min and the second shell, to 45°, was measured at 2.02°/min.
No decay was observed in the intensities of two standard reflections
monitored every 198 reflections.
Structure solution via direct methods in the noncentrosymmetric
space group P1 revealed all non-hydrogen atoms. Hydrogens were
placed at calculated positions which were allowed to ride on the position
of the parent atom in subsequent cycles of least-squares refinement.
Absolute configuration was assigned from known stereochemistry of
the precursor. All non-hydrogen atoms were refined with anisotropic
thermal parameters. Hydrogen thermal parameters were set at 1.2 times
the equivalent isotropic value of the parent atom. Full details of the
crystallographic results are included in the Supporting Information.
(2R,3R)-O-((E)-2-((1R,2R)-2-Ethyl-2-hydroxycyclopentyl)vinyl)-
N,N,N′,N′-tetramethylsuccinamide (27b). Using the general proce-
dure above, substrate 15b was cyclized, and the products were isolated
after chromatography (6:1 EtOAc/MeOH) to afford a 94:3:3 mixture
of 27b, 28b, and (29b + 30b), respectively (83% combined yield):
1H NMR (400 MHz, CDCl3) δ 6.17 (d, J ) 13.0 Hz, 1H), 4.90 (dd, J
) 13.0, 8.4 Hz, 1H), 4.81 (broad dd, J ) 6.8, 4.1 Hz, 1H), 4.70 (d, J
) 4.1 Hz, 1H), 3.90 (broad d, 6.8H, 1 (OH)), 3.15 (s, 3H), 3.10 (s,
3H), 2.97 (s, 3H), 2.93 (s, 3H), 2.13-2.04 (m, 1H), 1.68-1.48 (m,
7H), 1.37-1.26 (m, 1H), 0.89 (t, J ) 7.5 Hz, 3H); 13C NMR (100
MHz, CDCl3) δ 170.4, 168.3, 145.4, 106.3, 82.3, 79.3, 69.4, 47.5, 36.91,
36.85, 36.56, 36.22, 35.95, 32.0, 30.0, 21.1, 8.7; IR (neat) 3416, 1651,
1644, 1504, 1150, 1058 cm-1; MS (CI+, NH3) calcd for C17H31N2O5
(M + H): m/e 343.2255, found 343.2233; 343 (100), 325 (50); (EI+)
20
72 (60); [R]D +24.8° (c 2.07, EtOAc).
(2R,3R)-O-((E)-2-((1R,2R)-2-Hydroxycyclopentyl)vinyl)-N,N,N′,N′-
tetramethylsuccinamide (27h). Samarium (465 mg, 3.09 mmol) was
weighed into a flask in a glovebox. After sealing the flask with a
septum, the flask was transferred to an argon manifold. To the solid
samarium was added THF (25 mL), followed by diiodomethane (686
mg, 2.56 mmol). The mixture was stirred vigorously at room
temperature for 2 h. A solution of 15h (200 mg, 0.64 mmol) in THF
(5 mL) was added by syringe pump over a period of 7 h at room
temperature. After stirring at room temperature for an additional 12
h, air was bubbled into the mixture until the blue-green color had faded
to brown. To this mixture was added saturated aqueous Na2SO3 (0.5
mL), saturated aqueous NaHCO3 (6 mL), and enough Celite to produce
a viscous slurry. After dilution with CH2Cl2, the mixture was filtered
through a bed of Celite. The filter bed was washed several times with
CH2Cl2, and the combined filtrates were dried over MgSO4. The
resultant crude oil was chromatographed on silica gel (6:1 EtOAc/
MeOH), yielding a 93:5:2 mixture of 27h, 28h, and (29h + 30h),
respectively (107 mg, 53% combined yield): 1H NMR (400 MHz,
CDCl3) δ 6.22 (d, J ) 12.6 Hz, 1H), 5.00 (dd, J ) 12.6, 7.9 Hz, 1H),
4.81 (broad dd, J ) 6.3, 4.6 Hz, 1H), 4.57 (d, J ) 4.6 Hz, 1H), 4.13
(broad s, 1H), 3.99 (broad s, 1 (OH)), 3.12 (s, 3H), 3.10 (s, 3H), 2.94
(s, 3H), 2.91 (s, 3H), 2.31-2.24 (m, 1H), 2.00 (broad s, 1 (OH)), 1.85-
1.73 (m, 2H), 1.73-1.58 (m, 2H), 1.56-1.46 (m, 2H); 13C NMR (100
MHz, CDCl3) δ 170.3, 168.4, 145.2, 106.5, 78.8, 75.4, 69.2, 44.4, 36.92,
36.89, 36.11, 35.85, 33.6, 28.6, 21.9; IR (neat) 3406, 1644, 1504, 1147,
1058 cm-1; MS (EI+) calcd for C15H27N2O5 (M + H): m/e 315.1920,
found 315.1897; 315 (<5 (M + H)), 188 (15), 116 (100), 72 (89), 44
20
116 (100), 72 (48); [R]D +5.2° (c 2.04, EtOAc). Anal. Calcd for
C17H30N2O5: C: 59.63; H: 8.83; N: 8.18. Found: C: 59.69; H: 8.82;
N: 8.16.
(2R,3R)-O-((E)-2-((1R,2S)-2-Hydroxy-2-(1-methylethyl)cyclopen-
tyl)vinyl)-N,N,N′,N′-tetramethylsuccinamide (27c). Using the general
procedure above, substrate 15c was cyclized, and the products were
isolated after chromatography (9:1 EtOAc/MeOH) to afford a 90:7:3
mixture of 27c, 28c, and (29c + 30c), respectively (82% combined
yield): 1H NMR (400 MHz, CDCl3) δ 6.18 (d, J ) 13.0 Hz, 1H), 4.92
(dd, J ) 13.0, 8.1 Hz, 1H), 4.81 (broad dd, J ) 7.0, 4.2 Hz, 1H), 4.69
(d, J ) 4.2 Hz, 1H), 3.88 (broad d, J ) 7.0 Hz, 1H (OH)), 3.14 (s,
3H), 3.10 (s, 3H), 2.97 (s, 3H), 2.92 (s, 3H), 2.31 (broad dt, J ) 10.3,
7.7 Hz, 1H), 1.79-1.57 (m, 5H), 1.53-1.44 (m, 2H), 0.98 (d, J ) 6.9
Hz, 3H), 0.83 (d, J ) 6.9 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ
170.3, 168.3, 145.4, 106.2, 84.5, 78.9, 69.2, 44.5, 36.81, 36.78, 35.99,
35.74, 34.6, 32.6, 30.2, 20.9, 17.74, 17.63; IR (neat) 3417, 1644, 1503,
1148, 1058 cm-1; MS (EI+) calcd for C18H33N2O5 (M + H): m/e
357.2389, found 357.2406; 357 (M + H (<5)), 188 (15), 116 (100),
72 (95); [R]D20 +6.5° (c 2.00, EtOAc). Anal. Calcd for C18H32N2O5:
C: 61.65; H: 9.05; N: 7.86. Found: C: 60.63; H: 9.29; N: 7.33.
20
(21); [R]D +2.8° (c 3.02, EtOAc).
(2R,3R)-O-((E)-2-((1S,2S)-2-Hydroxy-2-methylcyclopentyl)vinyl)-
N,N,N′,N′-tetramethylsuccinamide (28a). Using the general proce-
dure above, substrate 85 was cyclized, and the products were isolated
after chromatography to afford a 13:83:3 mixture of 27a, 28a, and (29a
+ 30a), respectively (83% combined yield): 1H NMR (400 MHz,
CDCl3) δ 6.21 (d, J ) 12.6 Hz, 1H), 4.86 (dd, J ) 12.6, 8.7 Hz, 1H),
4.82 (dd, J ) 6.9, 4.3 Hz, 1H), 4.71 (d, J ) 4.3 Hz, 1H), 3.92 (d, J )
6.9Hz, 1H (OH)), 3.15 (s, 3H), 3.10 (s, 3H), 2.97 (s, 3H), 2.93 (s, 3H),
2.05-1.99 (m, 1H), 1.81-1.68 (m, 3H), 1.67-1.50 (m, 3H), 1.159 (s,
3H); 13C NMR (100 MHz, CDCl3) δ 170.22, 168.22, 145.55, 106.07,

Ketyl-Olefin Radical Cyclizations
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1275
79.76, 78.94, 69.1, 48.8, 39.82, 36.79 (2 carbons), 36.00, 35.69, 30.37,
25.96, 21.15; IR (neat) 3416, 1644, 1504, 1151, 1058 cm-1; MS (EI+)
calcd for C16H29O5N2 (M + H): m/e 329.2076, found 329.2085; 329
The organic phase was separated, and the aqueous layer was extracted
with CH2Cl2 (5×). The combined organic extracts were washed with
brine, back-extracting with CH2Cl2 (2×). The organic phase was dried
(MgSO4) and concentrated, yielding 35 (8.45 g, 96%): 1H NMR (400
MHz, CDCl3) δ 5.80 (dt, J ) 15.8, 6.6 Hz, 1H), 5.46 (ddt, J ) 15.8,
5.4, 1.2 Hz, 1H), 4.69 (d, J ) 5.4 Hz, 1H), 3.64 (s, 3H), 3.28 (s, 6H),
3.15 (s, 3H), 2.41 (broad t, J ) 7.4 Hz, 2H), 2.15-2.09 (m, 2H), 1.77-
1.69 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 173.6 (broad), 134.1,
126.8, 102.7, 60.6, 52.0, 31.5 (broad), 31.3, 30.5 (broad), 23.1; IR (neat)
20
(<5 (M + H)), 116 (84), 72 (75), 18 (100); [R]D +43.3° (c 2.24,
EtOAc).
5-Hydroxy-N,O-dimethylpentanohydroxamic Acid (32). Trim-
ethylaluminum (450 mL, 0.90 mol) was slowly added over 2 h to N,O-
dimethylhydroxylamine hydrochloride (88.0 g, 0.90 mol) in CH2Cl2
(500 mL) at -78 °C. The bath was removed, and the solution was
allowed to stir overnight. Upon cooling the solution to 0 °C,
δ-valerolactone (30.15 g, 0.298 mmol) was cannulated in over a 30
min period. The cooling bath was removed, and the solution was stirred
for several hours. The solution was cooled to 0 °C, whereupon
Rochelle’s salt (100 g) in H2O (150 mL) was added (very slowly at
first). The mixture was filtered through a bed of Celite, washing several
times with CH2Cl2. Drying the solution over MgSO4 and concentration
yielded 32 (43.62 g, 91%): 1H NMR (400 MHz, CDCl3) δ 3.65 (s,
3H), 3.60 (t, J ) 6.3 Hz, 2H), 3.15 (s, 3H), 2.44 (broad t, J ) 6.9 Hz,
2H), 1.74-1.67 (m, 2H), 1.61-1.54 (m, 2H); 13C NMR (100 MHz,
CDCl3) δ 174.4 (broad), 61.5, 60.9, 31.9, 31.8 (broad), 31.0 (broad),
1667, 1132, 1050, 993 cm-1
.
General Method for the Preparation of Ketones from Weinreb’s
Amide 35. A solution of the organolithium or Grignard reagent was
added to 35 in THF (2.5 mL/mmol 35) at -60 °C to 0 °C. After stirring
20-120 min at 0 °C to 23 °C, H2O (for organolithium reagents) or a
10% solution of Rochelle’s salt (for Grignard reagents) was added.
The aqueous layer was extracted with Et2O (2×). The combined
organic extracts were washed with brine and dried over MgSO4. The
resultant oil was subjected to chromatography on silica gel.
(E)-8,8-Dimethoxy-6-octen-2-one (36). Utilizing the general pro-
cedure above, methylmagnesium bromide in Et
2O provided an oil that
20.3; IR (neat) 3424, 1644, 1179, 1060, 999 cm-1
.
was subjected to chromatography on silica gel (3:1 hexanes/EtOAc),
yielding 36 (80%): 1H NMR (400 MHz, CDCl3) δ 5.75 (dt, J ) 15.6,
6.2 Hz, 1H), 5.43 (ddt, J ) 15.6, 5.3, 1.3 Hz, 1H), 4.69 (d, J ) 5.3
Hz, 1H), 3.27 (s, 6H), 2.40 (t, J ) 7.4 Hz, 2H), 2.09 (s, 3H), 2.08-
2.02 (m, 2H), 1.69-1.62 (m, 2H); 13C NMR (100 MHz, CDCl3) δ
208.6, 134.4, 127.2, 103.1, 52.6, 42.8, 31.3, 29.9, 22.7; IR (neat) 1716,
N,O-Dimethyl-5-oxopentanohydroxamic acid (33). Oxalyl chlo-
ride (2.4 mL, 28 mmol) was slowly added to DMSO (4.0 mL, 56 mmol)
in CH2Cl2 (20 mL) at -78 °C. After stirring 5 min, 5-hydroxy-N,O-
dimethylpentanohydroxamic acid (3.00 g, 18.6 mmol) in CH2Cl2 (5
mL) was cannulated into the solution over a 5 min period. The
heterogeneous solution was stirred for 60 min, whereupon triethylamine
(13 mL, 93 mmol) was added. The mixture was warmed to room
temperature and stirred for several hours. The mixture was poured
into Et2O (100 mL) and filtered through Celite (washing with Et2O).
After concentration, the oil was diluted with EtOAc and Et2O and then
filtered through Celite. This solution was concentrated again, diluted
with Et2O, and filtered through Celite. After a final concentration, the
homogenous oil was submitted to vacuum distillation, collecting 33
(4.4 g, 76%) at 70-74 °C (0.2 torr): 1H NMR (400 MHz, CDCl3) δ
9.75 (t, J ) 1.5 Hz, 1H), 3.65 (s, 3H), 3.15 (s, 3H), 2.52 (td, J ) 7.0,
1.5 Hz, 2H), 2.46 (t, J ) 7.1 Hz, 2H), 1.98-1.91 (m, 2H); 13C NMR
(100 MHz, CDCl3) δ 202.0, 173.5 (broad), 61.0, 42.9, 31.9 (broad),
1674, 1132, 1052 cm-1
.
(E)-9,9-Dimethoxy-7-nonen-3-one (37). Utilizing the general
procedure above, ethylmagnesium bromide in Et2O provided an oil that
was subjected to silica gel chromatography (3:1 hexanes/EtOAc),
yielding 37 (65%): 1H NMR (400 MHz, CDCl3) δ 5.76 (dt, J ) 15.6,
6.8 Hz, 1H), 5.44 (ddt, J ) 15.6, 5.3, 1.3 Hz, 1H), 4.69 (d, J ) 5.3
Hz, 1H), 3.29 (s, 6H), 2.42-2.36 (m, 4H), 2.09-2.03 (m, 2H), 1.71-
1.63 (m, 2H), 1.04 (t, J ) 7.3 Hz, 3H); 13C NMR (100 MHz, CDCl3)
δ 210.7, 134.2, 127.0, 102.8, 52.3, 41.1, 35.5, 31.1, 22.5, 7.4; IR (neat)
1715, 1674, 1131, 1052 cm-1
.
(E)-9,9-Dimethoxy-2,2-dimethyl-7-nonen-3-one (38). Utilizing the
general procedure above, tert-butyllithium in pentane provided an oil
that was subjected to silica gel chromatography (7:2 hexanes/EtOAc),
yielding 38 (169 mg, 63%): 1H NMR (400 MHz, CDCl3) δ 5.78 (dt,
J ) 15.5, 6.9 Hz, 1H), 5.44 (ddt, J ) 15.5, 5.3, 1.4 Hz, 1H), 4.69 (d,
J ) 5.3 Hz, 1H), 3.29 (s, 6H), 2.46 (t, J ) 7.3 Hz, 2H), 2.07-2.02 (m,
2H), 1.68-1.61 (m, 2H), 1.10 (s, 9H); 13C NMR (100 MHz, CDCl3) δ
215.7, 134.9, 127.1, 103.2, 52.6, 44.1, 35.6, 31.5, 26.4, 22.9; IR (neat)
30.5 (broad), 16.8; IR (neat) 2726, 1722, 1661, 1180, 998 cm-1
.
(E)-N,O-Dimethyl-7-oxo-5-heptenohydroxamic acid (34). To a
solution of diisopropylamine (28 mL, 200 mmol) in THF (450 mL)
was added n-BuLi (80 mL, 203 mmol) at -78 °C. After stirring the
solution for 2 h, acetaldehyde N-tert-butylimine (13.7 mL, 101 mmol)
was added over 10 min. The solution was stirred at -78 °C for 90
min, and then diethyl chlorophosphate (14.6 mL, 101 mmol) was added
over 20 min. The solution was stirred at -78 °C for 2 h and then
allowed to warm to -10 °C over 3 h. After stirring for an additional
30 min at -10 °C, the solution was cooled again to -78 °C, whereupon
33 (14.7 g, 92 mmol) was added gradually. The solution was warmed
to 0 °C overnight and then poured into an ice cold solution of oxalic
acid (35 g, 277 mmol) in H2O (400 mL). Benzene (400 mL) was added,
and the resulting mixture was stirred for 36 h. The organic layer was
separated, and to the aqueous phase was added solid NaCl until the
solution was saturated. The aqueous phase was extracted with EtOAc
until no more product was detected in the extract. The combined
organic extracts were washed with saturated aqueous NaHCO3 (2×)
and brine, back-extracting with EtOAc (2×). After drying (MgSO4)
and concentration, the crude oil was submitted to chromatography on
silica gel (1:2 hexanes/EtOAc), yielding 34 of 92% purity by GC (7.05
g, 38%): 1H NMR (400 MHz, CDCl3) δ 9.46 (d, J ) 8.0 Hz, 1H),
6.80 (dt, J ) 15.7, 6.6 Hz, 1H), 6.08 (ddt, J ) 15.7, 8.0, 1.4 Hz, 1H),
3.62 (s, 3H), 3.12 (s, 3H), 2.42 (broad t, J ) 7.2 Hz, 2H), 2.39-2.33
(m, 2H), 1.85-1.78 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 193.9,
173.5 (broad), 157.8, 133.2, 61.2, 32.1 (two carbons), 30.8, 22.5; IR
1705, 1131, 1052 cm-1
.
(E)-1,1-Dimethoxy-2-tridecen-8-yn-7-one (39). Utilizing the gen-
eral procedure above, n-butyllithium in hexanes provided an oil that
was subjected to silica gel chromatography (6:1 hexanes/EtOAc),
yielding 39 (80%): 1H NMR (400 MHz, CDCl3) δ 5.76 (dt, J ) 15.8,
6.3 Hz, 1H), 5.44 (ddt, J ) 15.8, 5.3, 1.2 Hz, 1H), 4.69 (d, J ) 5.3
Hz, 1H), 3.28 (s, 6H), 2.50 (t, J ) 7.4 Hz, 2H), 2.32 (t, J ) 7.0 Hz,
2H), 2.10-2.05 (m, 2H), 1.78-1.70 (m, 2H), 1.56-1.49 (m, 2H), 1.44-
1.35 (m, 2H), 0.89 (t, J ) 7.2 Hz, 3H); 13C NMR (100 MHz, CDCl3)
δ 187.2, 133.8, 127.2, 102.6, 93.8, 80.5, 52.1, 44.3, 30.8, 29.3, 22.8,
21.5, 18.2, 13.1; IR (neat) 2212, 1674, 1131, 1052 cm-1
.
(E)-7,7-Dimethoxy-1-phenyl-5-hepten-1-one (40). Utilizing the
general procedure above, phenyllithium in 7:3 cyclohexane: Et2O
provided an oil that was subjected to silica gel chromatography (5:2
hexanes/EtOAc), yielding 40 (76%): 1H NMR (400 MHz, CDCl3) δ
7.94-7.91 (m, 2H), 7.55-7.51 (m, 1H), 7.45-7.41 (m, 2H), 5.83 (dt,
J ) 15.6, 6.6 Hz, 1H), 5.48 (ddt, J ) 15.6, 5.2, 1.5 Hz, 1H), 4.70 (d,
J ) 5.2 Hz, 1H), 3.29 (s, 6H), 2.96 (t, J ) 7.3 Hz, 2H), 2.19-2.14 (m,
2H), 1.88-1.81 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 199.4, 136.6,
134.2, 132.6, 128.2, 127.6, 127.1, 102.8, 52.3, 37.3, 31.2, 23.0; IR (neat)
(neat) 1689, 1661, 1130, 992 cm-1
.
1686, 1598, 1131, 1050 cm-1
.
(E)-7,7-Dimethoxy-N,O-dimethyl-5-heptenohydroxamic acid (35).
Trimethylsilyl trifluoromethanesulfonate (0.8 mL, 4.13 mmol) was
added to a solution of methoxytrimethylsilane (21 mL, 142 mmol) and
34 (7.0 g, 38 mmol) in CH2Cl2 (50 mL) at -78 °C. The reaction was
stirred overnight, whereupon pyridine (1.5 mL, 18 mmol) was added
to the now heterogeneous mixture. After stirring for 20 min, the
mixture was poured into ice-cold saturated aqueous NaHCO3 (100 mL).
(2R,3R)-O-((E)-2-((1R,2R,4S)-2,4-Dimethyl-2-hydroxycyclopen-
tyl)vinyl)-N,N,N′,N′-tetramethylsuccinamide (75). To a saturated (0.1
M) solution of SmI2 (6.2 mL, 0.62 mmol) in THF was added a solution
of 71 (81 mg, 0.24 mmol) in THF (2 mL) by syringe pump over a
period of 40 min at 30 °C. After stirring at 30 °C for an additional 60
min, air was bubbled into the mixture until the blue-green color had
faded to brown. To this mixture was added saturated aqueous Na2SO3

1276 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Molander et al.
(0.3 mL), saturated aqueous NaHCO3 (3 mL), and enough Celite to
produce a viscous slurry. After dilution with CH2Cl2, the mixture was
filtered through a bed of Celite. The filter bed was washed several
times with CH2Cl2, and the combined filtrates were dried over MgSO4.
The resultant crude oil was chromatographed on silica gel (6:1 EtOAc/
MeOH), yielding an 88:12 mixture of 75 and 76, respectively (63 mg,
77% combined yield): 1H NMR (400 MHz, CDCl3) δ 6.18 (d, J )
12.3 Hz, 1H), 4.90 (dd, J ) 12.3, 8.4 Hz, 1H), 4.82 (dd, J ) 7.0, 4.0
Hz, 1H), 4.71 (d, J ) 4.0 Hz, 1H), 3.88 (d, J ) 7.0 Hz, 1H), 3.15 (s,
3H), 3.11 (s, 3H), 2.98 (s, 3H), 2.93 (s, 3H), 2.26-2.17 (m, 1H), 1.96
(dd, J ) 13.5, 7.8 Hz, 1H), 1.88-1.80 (m, 2H), 1.37-1.31 (m, 1H),
1.18 (dd, J ) 13.5, 8.7 Hz, 1H), 1.15 (s, 3H), 0.95 (d, J ) 6.9 Hz,
3H); 13C NMR (100 MHz, CDCl3) δ 170.4, 168.3, 145.3, 106.2, 80.7,
78.9, 69.3, 49.2, 47.4, 38.4, 36.90, 36.85, 36.14, 35.87, 29.2, 26.0, 22.3;
IR (neat) 3417, 1644, 1504, 1145, 1058 cm-1; MS (EI+) calcd for
C17H31N2O5 (M + H): m/e 343.2233, found 343.2220; 343 (9 (M +
H)), 188 (58), 116 (99), 72 (100), 46 (58); [R]D20 +6.8° (c 2.10, EtOAc).
(2R,3R)-O-((E)-2-((1S,2S,4S)-2,4-Dimethyl-2-hydroxycyclopentyl)-
vinyl)-N,N,N′,N′-tetramethylsuccinamide (76): 1H NMR (400 MHz,
CDCl3) δ 6.20 (d, J ) 12.5 Hz, 1H), 1.01 (d, J ) 6.2 Hz, 3H); 13C
NMR (400 MHz, CDCl3) δ 170.3, 168.2, 145.5, 105.9, 80.0, 79.1, 49.9,
48.6, 40.2, 31.0, 27.5, 21.5.
EtOAc), yielding an epimeric mixture of lactols 86 (33 mg, 66%): 1H
NMR (400 MHz, CDCl3) δ major: 5.52-5.51 (m, 1H), 3.91 (s, 1H
(OH)), 2.47-2.43 (m, 1H), 2.16 (ddd, J ) 13.3, 9.1, 1.1 Hz, 1H), 1.83-
1.77 (m, 1H), 1.73-1.66 (m, 2H), 1.65-1.52 (m, 3H), 1.45 (s, 3H),
1.32 (ddd, J ) 13.3, 11.2, 6.9 Hz, 1H), minor: 5.50-5.49 (m, 1H),
4.20 (broad s, 1H (OH)), 2.28-2.18 (m, 1H), 2.00-1.91 (m, 1H); 13
C
NMR (100 MHz, CDCl3) δ major: 99.27, 93.7, 46.7, 41.7, 40.9, 33.1,
28.6, 24.0, minor: 99.31, 94.3, 47.3, 40.4, 33.6, 26.7, 24.6; IR (neat)
3405, 1037, 1005 cm-1
.
A solution of epimeric lactols 86 (33 mg, 0.23 mmol) in THF (2
mL) was cannulated into a mixture of LiAlH4 (8 mg, 0.21 mmol) in
THF (2.5 mL) at room temperature. After stirring for 30 min, the
reaction was quenched with 2.5 N NaOH (1 mL). To the mixture was
added Celite, followed by CH2Cl2. The slurry was filtered, washing
with CH2Cl2. After drying over MgSO4, the resultant oil was
chromatographed on silica gel (100% EtOAc), yielding (+)-87 (31 mg,
93%): 1H NMR (400 MHz, CDCl3) δ 3.75 (dt, J ) 10.5, 6.0 Hz, 1H),
3.61 (ddd, J ) 10.5, 7.8, 5.7 Hz, 1H), 1.85-1.43 (m, 9H), 1.28 (s,
3H); 13C NMR (100 MHz, CDCl3) δ 79.8, 61.7, 46.7, 41.6, 31.7, 30.0,
26.6, 21.2; IR (neat) 3362, 1056 cm-1; MS (EI+) calcd for C8H15O2
(M - H): m/e 143.1072, found 143.1074; 126 (10 (M - H2O)), 111
20
(23), 97 (38), 71 (87), 43 (100), 31 (36), 17 (15); [R]D +14.4° (c
(2S,3S)-O-((E)-2-((1S,2S,4S)-2,4-Dimethyl-2-hydroxycyclopentyl)-
vinyl)-N,N,N′,N′-tetramethylsuccinamide (79). To a saturated (0.1
M) solution of SmI2 (6.2 mL, 0.62 mmol) was added a solution of 72
(78 mg, 0.23 mmol) in THF (2 mL) by syringe pump over a period of
40 min at 30 °C. After stirring at 30 °C for an additional 60 min, air
was bubbled into the mixture until the blue-green color had faded to
brown. To this mixture was added saturated aqueous Na2SO3 (0.3 mL),
saturated aqueous NaHCO3 (3 mL), and enough Celite to produce a
viscous slurry. After dilution with CH2Cl2, the mixture was filtered
through a bed of Celite. The filter bed was washed several times with
CH2Cl2, and the combined filtrates were dried over MgSO4. The
resultant crude oil was chromatographed on silica gel (6:1 EtOAc/
MeOH), yielding 79 as the major isomer of a g 99:1 mixture of
diastereomers (69 mg, 88% combined yield): 1H NMR (400 MHz,
CDCl3) δ 6.18 (d, J ) 12.4 Hz, 1H), 4.93 (dd, J ) 12.4, 8.0 Hz, 1H),
4.82 (dd, J ) 6.9, 4.0 Hz, 1H), 4.71 (d, J ) 4.0 Hz, 1H), 3.85 (d, J )
6.9 Hz, 1H (OH)), 3.15 (s, 3H), 3.11 (s, 3H), 2.98 (s, 3H), 2.93 (s,
3H), 2.11-2.05 (m, 1H), 2.00-1.90 (m, 2H), 1.82-1.75 (m, 1H), 1.41-
1.28 (m, 2H), 1.15 (m, 3H), 1.01 (d, J ) 6.3 Hz, 3H); 13C NMR (100
MHz, CDCl3) δ 170.4, 168.3, 145.5, 105.9, 79.9, 79.5, 69.4, 49.9, 48.5,
39.6, 36.98, 36.89, 36.35, 36.07, 30.9, 27.4, 21.8; IR (neat) 3421, 1644,
1503, 1148, 1057 cm-1; MS (EI+) calcd for C17H31N2O5 (M + H):
m/e 343.2233, found 343.2230; 188 (50), 142 (20), 116 (100), 98 (37),
2.05, CHCl3).
(1S,2S)-2-(2-Hydroxyethyl)-1-methylcyclopentanol ((-)-87). To
a 14:86 mixture of 27a and 28a, respectively (87 mg, 0.27 mmol), in
THF (4 mL) at room temperature was added 6 N HCl (2 mL). The
solution was stirred for 30 min, and then solid Na2CO3 and saturated
aqueous NaHCO3 were added until no more effervescence was
observed. The aqueous layer was extracted with EtOAc (3×). The
combined organic extracts were washed with brine (2×) and dried over
K2CO3. The resultant oil was chromatographed on silica gel (1:1
hexanes/EtOAc), yielding a mixture of epimeric lactols 88 (27 mg,
0.19 mmol, 70%). This mixture was diluted in THF (2 mL) and
cannulated into a mixture of LiAlH4 (7 mg, 0.18 mmol) in THF (2.5
mL) at room temperature. After stirring 20 min, the reaction was
quenched with 2.5 N NaOH (1 mL). To the mixture was added Celite,
followed by CH2Cl2. The slurry was filtered, washing with CH2Cl2.
After drying over MgSO4, the resultant oil was chromatographed on
silica gel (100% EtOAc), yielding (-)-87 (24 mg, 86%): [R]D20 -11.7°
(c 0.99, CHCl3).
Acknowledgment. We thank the National Institutes of
Health for their generous financial support of this program. We
would also like to thank Dr. Todd Nelson for helpful comments
regarding the preparation of the manuscript.
20
72 (90), 46 (50); [R]D -12.9° (c 2.30, EtOAc).
Supporting Information Available: Crystallographic data
and tables for 27a. Full experimental details for the preparation
of 10-12, 14, 15b-g, 42-45, 48, 51, 52, 64-72, and 81-85.
Copies of 1H and 13C spectra of compounds for which no
elemental analysis was obtained (134 pages). See any current
masthead page for ordering and Internet access instructions.
(1R,2R)-2-(2-Hydroxyethyl)-1-methylcyclopentanol [(+)-87]. To
a 97:3 mixture of 27a and 28a (116 mg, 0.35 mmol) in THF (4 mL)
at room temperature was added 6 N HCl (2 mL). The solution was
stirred for 30 min, and then solid Na2CO3 and saturated aqueous
NaHCO3 were added until no more effervescence was observed. The
aqueous solution was extracted with EtOAc (3×). The combined
organic extracts were washed with brine (2×) and dried over K2CO3.
The resultant oil was chromatographed on silica gel (1:1 hexanes/
JA963195C