Simple reagents for direct halonium-induced polyene cyclizations

Article abstract of DOI:10.1021/ja106813s

Although there are many reagent combinations that can initiate polyene cyclizations, simple electrophilic halogen sources have not yet proven broadly effective as promoters of such processes. Herein is described a readily prepared and stable class of reagents capable of effecting such transformations for a wide range of electron-rich and -deficient terpenes derived from geraniol, farnesol, and nerol, thereby enabling the effective synthesis of a diverse array of complex chlorine-, bromine-, and iodine-containing polycyclic frameworks. Efforts to date have led to the first racemic laboratory total synthesis and structural revision of the anti-HIV natural product peyssonol A as well as an efficient and concise inaugural total synthesis of peyssonoic acid A. They have also permitted formal racemic total syntheses of aplysin-20, loliolide, K-76, and stemodin to be achieved through routes that are typically shorter, higher-yielding, and more environmentally conscious than previous efforts. Preliminary attempts to use chiral forms of the reagent class for enantioselective alkene halogenation are also described.

Full text of DOI:10.1021/ja106813s

Published on Web 09/21/2010
Simple Reagents for Direct Halonium-Induced Polyene
Cyclizations
Scott A. Snyder,* Daniel S. Treitler, and Alexandria P. Brucks
Department of Chemistry, Columbia UniVersity, HaVemeyer Hall, 3000 Broadway, New York,
New York 10027
Received July 30, 2010; E-mail: sas2197@columbia.edu
Abstract: Although there are many reagent combinations that can initiate polyene cyclizations, simple
electrophilic halogen sources have not yet proven broadly effective as promoters of such processes. Herein
is described a readily prepared and stable class of reagents capable of effecting such transformations for
a wide range of electron-rich and -deficient terpenes derived from geraniol, farnesol, and nerol, thereby
enabling the effective synthesis of a diverse array of complex chlorine-, bromine-, and iodine-containing
polycyclic frameworks. Efforts to date have led to the first racemic laboratory total synthesis and structural
revision of the anti-HIV natural product peyssonol A as well as an efficient and concise inaugural total
synthesis of peyssonoic acid A. They have also permitted formal racemic total syntheses of aplysin-20,
loliolide, K-76, and stemodin to be achieved through routes that are typically shorter, higher-yielding, and
more environmentally conscious than previous efforts. Preliminary attempts to use chiral forms of the reagent
class for enantioselective alkene halogenation are also described.
utilized.9 Indeed, the use of simple halogen electrophiles to
achieve such cyclizations, even in racemic form, typically has
Introduction
With little question, the ability to convert polyene starting
materials into far more complex frameworks via stereoselective
cation-π cyclizations constitutes one of the most important
strategies currently available for C-C bond construction.1
Indeed, in the half century since Stork and Eschenmoser first
advanced their hypothesis2 for the existence of such processes,
chemists have devised numerous sets of unique substrates,
reaction conditions, and reagent combinations that enable such
reactions to be conducted with very high levels of stereoselec-
tivity. Specifically, numerous versions of nonmetal-3 and metal-
induced4 [especially Hg(II),4a-i Pd(II),4j-l Pt(II),4l-o and
Au(I)4o-q] cyclizations have been developed and honed to the
point where the efficient synthesis of dozens of molecules of
natural and designed origin can readily be achieved.5
What remains to be accomplished, however, is broadly
initiating such processes with halogen electrophiles. Nature takes
advantage of such reactivity with some frequency, as vanadium-
and heme-based haloperoxidases6 have been shown (or hypoth-
esized) to convert simple polyene precursors into the highlighted
rings of the natural products drawn in Figure 1;7 these molecules
represent a select subset of the nearly 200 chlorine- and bromine-
containing compounds which possess such ring systems that
have been isolated to date from both marine and terrestrial
sources.8 Yet, mirroring such reactivity in the laboratory flask
has proven elusive unless haloperoxidases themselves have been
led to modest product yields and then only for a narrow range
of substrates with certain halogens.10-13
To the best of our knowledge, there have been no examples
of any chemical reagents effecting a chloronium-induced
polyene cyclization in any yield.10 Explorations with bromine-
based systems, by contrast, have been much more extensive.
Nevertheless, no reagent possesses the scope of reactivity needed
to handle the diverse range of CdC double bond nucleophilicity
possible in functionalized terpene precursors.11 Most reagents
convert electron-rich systems into multiple, and often challeng-
ing to separate, products due to issues of olefin chemoselectivity,
with electron-poor substrates typically leading to products where
the cyclizations stall after forming a single ring (i.e., 6f7)11e
or an exogenous species behaves as nucleophile or base prior
to cation-π cyclization (8f9).11g In fact, yields of cyclized
(3) For selected examples, see: (a) Yang, D.; Gu, S.; Yan, Y.-L.; Zhao,
H.-W.; Zhu, N.-Y. Angew. Chem., Int. Ed. 2002, 41, 3014–3017. (b)
Ishibashi, H.; Ishihara, K.; Yamamoto, H. J. Am. Chem. Soc. 2004,
126, 11122–11123. (c) Zhao, Y.-J.; Chng, S.-S.; Loh, T.-P. J. Am.
Chem. Soc. 2007, 129, 492–493. (d) Zhao, Y.-J.; Loh, T.-P. J. Am.
Chem. Soc. 2008, 130, 10024–10029. (e) Surendra, K.; Corey, E. J.
J. Am. Chem. Soc. 2009, 131, 13928–13929. (f) Topczewski, J. J.;
Callahan, M. P.; Neighbors, J. D.; Wiemer, D. F. J. Am. Chem. Soc.
2009, 131, 14630–14631. (g) Rendler, S.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2010, 132, 5027–5029. (h) Knowles, R. N.; Lin, S.;
Jacobsen, E. N. J. Am. Chem. Soc. 2010, 132, 5030–5032. (i) Corey,
E. J.; Lin, S. A. J. Am. Chem. Soc. 1996, 118, 8765–8776. (j) Corey,
E. J.; Lee, J.; Liu, D. S. Tetrahedron Lett. 1994, 35, 9149–9152. (k)
Mi, Y.; Schreiber, J. V.; Corey, E. J. J. Am. Chem. Soc. 2002, 124,
11290–11291. (l) Zhang, J.; Corey, E. J. Org. Lett. 2001, 3, 3215–
3216. (m) Shenvi, R. A.; Corey, E. J. Org. Lett. 2010, 12, 3548–
3551. For selected examples of classic papers in the field, see: (n)
Johnson, W. S.; Gravestock, M. B.; McCarry, B. E. J. Am. Chem.
Soc. 1971, 93, 4332–4334. (o) van Tamelen, E. E.; Willet, J.; Schwartz,
M.; Nadeau, R. J. Am. Chem. Soc. 1966, 88, 5937–5938. (p) van
Tamelen, E. E.; Hwu, J. R. J. Am. Chem. Soc. 1983, 105, 2490–2491.
(1) For reviews, see: (a) Sutherland, J. K. In ComprehensiVe Organic
Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford, 1991; Vol. 5,
pp 341-377. (b) Yoder, R. A.; Johnston, J. N. Chem. ReV. 2005, 105,
4730–4756.
(2) (a) Stork, G.; Burgstahler, A. W. J. Am. Chem. Soc. 1955, 77, 5068–
5077. (b) Eschenmoser, A.; Ruzicka, L.; Jeger, O.; Arigoni, D. HelV.
Chim. Acta 1955, 38, 1890–1904. For a recent perspective on this
work 50 years after its original publication, see: (c) Eschenmoser, A.;
Arigoni, D. HelV. Chim. Acta 2005, 88, 3011–3050.
9
10.1021/ja106813s  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 14303–14314 14303

A R T I C L E S
Snyder et al.
reactivity for enantioselective bromonium-induced cyclizations,
however, were not as successful.14 The second advance is the
recent disclosure of Barluenga’s hypervalent iodonium-reagent
(4) (a) Hoye, T. R.; Kurth, M. J. J. Org. Chem. 1979, 44, 3461–3467. (b)
Hoye, T. R.; Caruso, A. J.; Kurth, M. J. J. Org. Chem. 1981, 46,
3550–3552. (c) Nishizawa, M.; Takenaka, H.; Nishide, H.; Hayashi,
Y. Tetrahedron Lett. 1983, 24, 2581–2584. (d) Nishizawa, M.;
Takenaka, H.; Hayashi, Y. J. Org. Chem. 1986, 51, 806–813. (e)
Corey, E. J.; Reid, J. G.; Myers, A. G.; Hahl, R. W. J. Am. Chem.
Soc. 1987, 109, 918–919. (f) Gopalan, A. S.; Prieto, R.; Mueller, B.;
Peters, D. Tetrahedron Lett. 1992, 33, 1679–1682. (g) Nishizawa, M.;
Morikuni, E.; Asoh, K.; Kan, Y.; Uenoyama, K.; Imagawa, H. Synlett
1995, 169–170. (h) Nishizawa, M.; Yadav, A.; Iwamoto, Y.; Imagawa,
H. Tetrahedron 2004, 60, 9223–9234. (i) Takao, H.; Wakabayashi,
A.; Takahashi, K.; Imagawa, H.; Sugihara, T.; Nishizawa, M.
Tetrahedron Lett. 2004, 45, 1079–1082. (j) Kucera, D. J.; O’Connor,
S. J.; Overman, L. E. J. Org. Chem. 1993, 58, 5304–5306. (k)
Overman, L. E.; Abelman, M. M.; Kucera, D. J.; Tran, V. D.; Ricca,
D. J. Pure App. Chem. 1992, 64, 1813–1819. (l) Koh, J. H.; Gagne´,
M. R. Angew. Chem., Int. Ed. 2004, 43, 3459–3461. (m) Mullen, C. A.;
Campbell, A. N.; Gagne´, M. R. Angew. Chem., Int. Ed. 2008, 47,
6011–6014. (n) Feducia, J. A.; Gagne´, M. R. J. Am. Chem. Soc. 2008,
130, 592–599. (o) Fu¨rstner, A.; Davies, P. W. Angew. Chem., Int. Ed.
2007, 46, 3410–3449. (p) Fu¨rstner, A.; Morency, L. Angew. Chem.,
Int. Ed. 2008, 47, 5030–5033. (q) Sethofer, S. G.; Mayer, T.; Toste,
F. D. J. Am. Chem. Soc. 2010, 132, 8276–8277.
(5) Other approaches to such frameworks have been developed as well.
For a recent example featuring a carbonyl-ene reaction, see: Zhao,
Y.-J.; Li, B.; Tan, L.-J. S.; Shen, Z.-L.; Loh, T.-P. J. Am. Chem. Soc.
2010, 132, 10242–10244.
(6) For reviews, see: (a) Vailancourt, F. A.; Yeh, E.; Vosburg, D. A.;
Garneau-Tsodikova, S.; Walsh, C. T. Chem. ReV. 2006, 106, 3364–
3378. (b) Butler, A.; Walker, J. V. Chem. ReV. 1993, 93, 1937–1944,
For work illustrating that the napyradiomycins are prepared via
vanadium-based haloperoxidases, see: (c) Winter, J. M.; Moffitt, M. C.;
Zazopoulos, E.; McAlpine, J. B.; Dorrestein, P. C.; Moore, B. S.
J. Biol. Chem. 2007, 282, 16362–16368.
(7) For the isolation of these natural products, see: (a) Shiomi, K.;
Nakamura, H.; Iinuma, H.; Naganawa, H.; Takeuchi, T.; Umezawa,
H.; Iitaka, Y. J. Antibiot. 1987, 40, 1213–1219, and references therein.
(b) Cho, J. Y.; Kwon, H. C.; Williams, P. G.; Jensen, P. R.; Fenical,
W. Org. Lett. 2006, 8, 2471–2474. (c) Talpir, R.; Rudi, A.; Kashman,
Y.; Loya, Y.; Hizi, A. Tetrahedron 1994, 50, 4179–4184. (d) Loya,
S.; Bakhanaskvili, M.; Kashman, Y.; Hizi, A. Arch. Biochem. Biophys.
1995, 316, 789–796. (e) Howard, B. M.; Fenical, W. Tetrahedron Lett.
1976, 17, 41–44. (f) Wall, M. E.; Wani, M. C.; Manikumar, G.; Taylor,
H.; Hughes, T. J.; Gaetano, K.; Gerwick, W. H.; McPhail, A. T.;
McPhail, D. R. J. Nat. Prod. 1989, 52, 1092–1099.
(8) For a review on halogenated natural products, see: Gribble, G. N. Acc.
Chem. Res. 1998, 31, 141–152.
(9) Carter-Franklin, J. N.; Butler, A. J. Am. Chem. Soc. 2004, 126, 15060–
15066.
(10) Only radical approaches have allowed access to relevant frameworks,
typically through halogenation of every alkene within the substrate: (a)
Yang, D.; Yan, Y.-L.; Zheng, B.-F.; Gao, Q.; Zhu, N.-Y. Org. Lett.
2006, 8, 5757–5760. (b) Helliwell, M.; Fengas, D.; Knight, C. K.;
Parker, J.; Quayle, P.; Raftery, J.; Richards, S. N. Tetrahedron Lett.
2005, 46, 7129–7134.
Figure 1. (A) Selected natural products with rings that arise in Nature via
halonium-induced cation-π cyclizations of appropriate terpene precursors
(cf. ref 7), (B) generalized retrosynthesis for their formation, and (C) selected
examples of attempting such reactions with simple reagents, thereby
revealing challenges in executing such reactions in the absence of enzymes.
(11) For representative examples, see: (a) van Tamelen, E. E.; Hessler, J.
Chem. Commun. 1966, 411–413. (b) Kato, T.; Ichinose, I.; Kumazawa,
S.; Kitahara, Y. Bioorg. Chem. 1975, 4, 188–193. (c) Kato, T.;
Ichinose, I.; Kamoshida, A.; Kitahara, Y. J. Chem. Soc., Chem.
Commun. 1976, 518–519. (d) Wolinsky, L. E.; Faulkner, D. J. J. Org.
Chem. 1976, 41, 597–600. (e) Gonza´lez, A. G.; Mart´ın, J. D.; Pe´rez,
C.; Ram´ırez, M. A. Tetrahedron Lett. 1976, 17, 137–138. (f) Hoye,
T. R.; Kurth, M. J. J. Org. Chem. 1978, 43, 3693–3697. (g) Kato, T.;
Ichinose, I. J. Chem. Soc., Perkin Trans. 1 1980, 1051–1056. (h) Shieh,
H.-M.; Prestwich, G. D. Tetrahedron Lett. 1982, 23, 4643–4646. (i)
Kato, T.; Mochizuki, M.; Hirano, T.; Fujiwara, S.; Uyehara, T.
J. Chem. Soc., Chem. Commun. 1984, 1077–1078. (j) Yamaguchi, Y.;
Uyehara, T.; Kato, T. Tetrahedron Lett. 1985, 26, 343–346. (k)
Fujiwara, S.; Takeda, K.; Uyehara, T.; Kato, T. Chem. Lett. 1986,
1763–1766. (l) Tanaka, A.; Sato, M.; Yamashita, K. Agric. Biol. Chem.
1990, 54, 121–123. (m) Tanaka, A.; Oritani, T. Biosci. Biotech.
Biochem. 1995, 59, 516–517.
material from electron-deficient systems using electrophilic
bromine initiators have been less than 30%.
Iodonium-induced reactions12 are the best developed, largely
due to two recent advances. The first is Ishihara’s use of a
phosphorus-complexed form of N-iodosuccinimide (NIS) to
cyclize three aryl-containing polyenes derived from geraniol
(including 10); when certain chiral phosphoramidites were used
in stoichiometric amounts, the cyclization could be achieved
with high enantioselection (95% e.e.).12a Key was the use of
30 h of controlled cryogenic conditions in the initial halonium-
induced reaction followed by the addition of ClSO3H in a
separate step to convert partially cyclized materials (such as
11) into the final tricycle (i.e., 12). Efforts to deploy such
(12) (a) Sakakura, A.; Ukai, A.; Ishihara, K. Nature 2007, 445, 900–903.
(b) Barluenga, J.; Trincado, M.; Rubio, E.; Gonza´lez, J. M. J. Am.
Chem. Soc. 2004, 126, 3416–3417. (c) Barluenga, J.; Alvarez-Pe´rez,
M.; Rodr´ıguez, F.; Fan˜ana´s, F. J.; Cuesta, J. A.; Garc´ıa-Granda, S. J.
Org. Chem. 2003, 68, 6583–6586.
9
14304 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010

Direct Halonium-Induced Polyene Cyclizations
A R T I C L E S
Ipy2BF4. When coupled with an additional equivalent of HBF4,
this reagent was able to convert several terpenes into cyclized
products.12b However, neither of these two reagent combinations
has been reported to successfully cyclize an electron-deficient
polyene substrate; a complete listing of these, as well as other
previous examples of terpene-based halonium-induced cycliza-
tions, can be found in the Supporting Information section.
Thus, given this global range of present capabilities for all
direct halonium-based cyclizations, especially that for bromine
and chlorine, most natural product structures of the types
represented by 1-5 (cf. Figure 1) have been targeted through
strategies that feature indirect, multistep alternatives. These
variants have included the formation and cyclization of halo-
hydrin intermediates,15 stoichiometric Hg(II)-induced cycliza-
tions followed by stereoselective replacement with chlorine,
bromine, or iodine16,4a (a reaction sequence we recently rendered
enantioselective),17 or two-step inversion and replacement
sequences from oxygen-cyclized materials.18
Figure 2. Structures of previously synthesized materials derived from
molecular Br2, a sulfide, and a Lewis acid; in all cases, their chemical
reactivity with double bonds was unexplored.
substrates in yields that are often multifold improvements over
previously available alternatives.
Results and Discussion
Herein, we describe the development of the first class of
reagents that can render possible the direct synthesis of a diverse
range of chlorine-, bromine-, and iodine-containing polycycles
via cation-π cyclizations. Each reagent is a readily prepared
crystalline solid that reacts with olefins highly chemoselectively
and rapidly, with reactions normally complete within 5 min at
low temperature. Moreover, added acids are not typically
required to drive cyclizations to completion. To date, these
reagents have allowed us to accomplish racemic total and formal
syntheses of 7 different natural products, 6 of which are
disclosed for the first time in this article (including a substantial
structural revision), as well as to cyclize nearly 20 additional
1. Development of BDSB. Having carefully studied the
precedents cited above, we chose to begin our investigations
into the problem of achieving direct halonium-induced reactions
by attempting to solve the bromonium-based challenge. Our goal
was to identify a novel reagent with higher alkene chemose-
lectivity and less proclivity for side-product formation. We also
added as a requirement for that search that such a reagent would
need to be a stable and easily handled material rather than one
that would have to be prepared in situ. In addition, we hoped
that its molecular structure could ultimately prove applicable
to generating the corresponding iodine- and chlorine-based
variants and, eventually, chiral versions for asymmetric
applications.
(13) For selected examples of other halonium-induced cyclization reactions
not using terpene-based materials, as well as approaches that form
related structures through nucleophilic attack of halogen onto carbon
electrophiles, see: (a) Inoue, T.; Kitagawa, O.; Ochiai, O.; Shiro, M.;
Taguchi, T. Tetrahedron Lett. 1995, 36, 9333–9336. (b) Cui, X.-L.;
Brown, R. S. J. Org. Chem. 2000, 65, 5653–5658. (c) Appelbe, R.;
Casey, M.; Dunne, A.; Pascarella, E. Tetrahedron Lett. 2003, 44, 7641–
7644. (d) Hajra, S.; Maji, B.; Karmakar, A. Tetrahedron Lett. 2005,
46, 8599–8603. (e) Hanessian, S.; Tremblay, M.; Marzi, M.; Del Valle,
J. R. J. Org. Chem. 2005, 70, 5070–5085. (f) Barluenga, J.; Trincado,
M.; Rubio, E.; Gonza´lez, J. Angew. Chem., Int. Ed. 2003, 42, 2406–
2409. (g) Barluenga, J.; Va´zquez-Villa, H.; Ballesteros, A.; Gonza´lez,
J. M. J. Am. Chem. Soc. 2003, 125, 9028–9029. (h) Barluenga, J.;
Gonza´lez, J. M.; Campos, P. C.; Asensio, G. Angew. Chem., Int. Ed.
1988, 27, 1546–1547. (i) Kang, S. H.; Lee, S. B.; Park, C. M. J. Am.
Chem. Soc. 2003, 125, 15748–15749. (j) Haas, J.; Piguel, S.; Wirth,
T. Org. Lett. 2002, 4, 297–300. (k) Grossman, R. B.; Trupp, R. J.
Can. J. Chem. 1998, 76, 1233–1237. (l) Zhang, W.; Zheng, S.; Liu,
N.; Werness, J. B.; Guzei, I. A.; Tang, W. J. Am. Chem. Soc. 2010,
132, 3664–3665.
Our initial idea to achieve the desired reactivity was
predicated on enhancing the electrophilicity of a typical bromine
source (like Br2 or NBS) while concurrently removing the
potential for any other species (such as a counterion) to serve
as either nucleophile or base.19 An extensive search of the
literature revealed the existence of several reagents that formally
met this criteria; 6 of these compounds are presented in Figure
2, all of which are complexes of Br2 with Me2S and a Lewis
acid.20
Interestingly, although these materials have been known for
some time (one was reported over 60 years ago), no report
describes their chemical reactivity. As such, we sought to fill
this gap, and began by preparing several members of the class.
Preliminary screens quickly revealed that the use of SbCl5 as
the Lewis acid component most consistently afforded solid
materials relative to boron or aluminum alternatives.21a Ad-
ditionally, of the various simple dialkyl sulfides that could be
(14) Part of the challenge may lie in the rapid transfer of bromonium to
unreacted alkene, thereby eroding enatioselectivity: (a) Brown, R. S.
Acc. Chem. Res. 1997, 30, 131–137. For similar challenges in
achieving the enantioselective addition of chalcogens such as sulfur
onto alkenes, see: (b) Denmark, S. E.; Collins, W. R.; Cullen, M. D.
J. Am. Chem. Soc. 2009, 131, 3490–3492. The same problem may
not exist for iodine, and has recently been indicated not to be as
profound an issue for chlorine: (c) Denmark, S. E.; Burk, M. T.;
Hoover, A. J. J. Am. Chem. Soc. 2010, 132, 1232–1233.
(19) Conceptually, BDSB is an example where a Lewis base has activated
a Lewis acid. For a review on this important mode of promoting
reactivity, see: Denmark, S. E.; Beutner, G. L. Angew. Chem., Int.
Ed. 2008, 47, 1560–1638.
(20) (a) Bohme, H.; Boll, E. Z. Anorg. Alleg. Chem. 1957, 290, 17–23. (b)
Goetz-Grandmont, G. J.; Leroy, M. J. F. J. Chem. Res. 1982, 160–
161. (c) Minkwitz, R.; Gerhard, V.; Werner, A. Z. Anorg. Alleg. Chem.
1989, 575, 137–144. (d) Askew, H. F.; Gates, P. N.; Muir, A. S. J.
Raman Spectrosc. 1991, 22, 265–274. (e) Minkwitz, R.; Baeck, B. Z.
Naturforschung B: Chem. Sci. 1993, 48, 694–696. (f) Regelmann, B.;
Klinkhammer, K. W.; Schmidt, A. Z. Anorg. Alleg. Chem. 1997, 623,
1633–1638.
(15) (a) Couladouros, E. A.; Vidali, V. P. Chem.sEur. J. 2004, 10, 3822–
3835. (b) Murai, A.; Abiko, A.; Masamune, T. Tetrahedron Lett. 1984,
25, 4955–4958.
(16) Replacement with iodine is often not completely stereoselective due
to competing electrophilic and radical substitution pathways. (a) Jensen,
F. R.; Gale, L. H. J. Am. Chem. Soc. 1960, 82, 148–151. (b) DePuy,
C. H.; McGirk, R. H. J. Am. Chem. Soc. 1974, 96, 1121–1132.
(17) Snyder, S. A.; Treitler, D. S.; Schall, A. Tetrahedron 2010, 66, 4796–
4804.
(21) (a) Arsenate anions were not investigated due to their toxicity. (b)
The di-t-butyl variant decomposed rapidly, likely due to the weakness
of the tertiary carbon-sulfur bond, while both the methyl and isopropyl
variants were solid, crystalline materials. In terms of stability,
solubility, and cost, however, 13 was superior.
(18) (a) Su, W.-C.; Fang, J.-M.; Cheng, Y.-S. Tetrahedron Lett. 1995, 36,
5367–5370. (b) Pulicani, J.-P.; Bouchard, H.; Bourzat, J.-V.; Com-
merc¸on, A. Tetrahedron Lett. 1994, 35, 9709–9712.
9
J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010 14305

A R T I C L E S
Snyder et al.
Scheme 1. Synthesis and Properties of BDSB (13)
Table 1. Exploration of the Generality of Direct,
Bromonium-Induced Cation-π Cyclizations using BDSB (1.1 equiv)
and 0.1 mmol of Substrate in Nitromethane
used (such as methyl, ethyl, isopropyl, or t-butyl), the ethyl
variant was the most easily prepared.21b As indicated in Scheme
1, addition of a slight excess of Et2S and SbCl5 to Br2 in 1,2-
dichloroethane at -30 °C immediately produced a yellow solid
that could be recrystallized from the reaction solution to give
the material shown in the inset photo in 87% yield. This odorless
crystalline solid, which we have named BDSB (for bromodi-
ethylsulfonium bromopentachloroantimonate, 13), can be pre-
pared smoothly on hundred-gram scale,22 is stable at ambient
temperature in an enclosed vial for at least 1 week (and for a
year or more at -20 °C), and possesses good solubility in
several organic solvents.23
Our optimism that BDSB would possess the desired reactivity
for cation-π cyclizations was enhanced upon obtaining its X-ray
crystal structure. As indicated, the relatively short bromine-sulfur
bond and effective sequestration of bromide to the antimonate
counterion constitutes a significant departure from typical
bromosulfonium complexes, an example of which is given in
Scheme 1 (and which is ineffective for bromonium-induced
cation-π cyclizations).24 Indeed, as we recently reported in a
preliminary communication,25 BDSB is very effective at induc-
ing cation-π cyclizations for a variety of substrates, including
those that possess electron-deficient alkenes, as well as polyenes
containing Z-alkenes.25
Table 1 provides a subset of the examples that were
previously disclosed. It is worth noting that some of these
reactions have been conducted on scales as large as 5.0 mmol
in equivalent yields and that the nitromethane utilized can be
recovered and reused in these large scale processes. Additionally,
reactions are generally very fast (usually complete in less than
5 min), and in all cases product yields are superior to those
obtained by other available methods reported in the literature.
As with most cation-π cyclizations, reaction concentrations need
to be kept dilute (0.05 M on small scale; 0.01 M on larger scale)
for optimal yields. In terms of chemoselectivity, BDSB will
typically react cleanly with olefins prior to aromatic systems,
even those that are electron-rich (such as those in 17 and 19,
a Produced as a 6.5:2.5:1.0 mixture of tri:tetra:disubstituted alkene
isomers. b Produced as a 3.8:1.0 mixture of separable diastereomers at
the highlighted carbon favoring the drawn product. c Generated
alongside some very minor diastereomers. d MeSO3H (15 equiv) added
with 1 h of additional stirring to promote the final cyclization.
Entries 3 and 4), though it can perform electrophilic aromatic
bromination if no CdC bonds are present.
In its reactions with olefins, BDSB possesses typical elec-
trophilic reactivity patterns: more substituted and more elec-
tronically activated double bonds will react faster, and usually
selectively, over their less-substituted and/or electron-deficient
counterparts.26 Fortunately, steric considerations appear to be
more important than electronic considerations given that in
polyenes such as 19, the more accessible, yet less electron-rich
distal double bond consistently reacts preferentially to the
hindered, more electron-rich central double bond. A minor side-
product formed in many reactions is the proton-cyclized
homologue; an acidic byproduct, likely protonated Et2S, is
formed as the reaction progresses and is responsible for the
observed yield (usually <5%) of this undesired compound.27
This acid cannot be neutralized in situ with added base, but is
of value in that it may help to drive many cyclizations to
(22) Snyder, S. A. Treitler, D. S. Org. Synth. 2010, submitted.
(23) 13 is fully soluble at ambient temperature in MeNO2, EtNO2, MeCN,
DMSO, DMF, and EtOAc, moderately to slightly soluble in CH2Cl2,
1,2-dichloroethane, chloroform, and toluene, and insoluble in benzene,
hexanes, and pentane. We have observed that 13 is soluble in acetone,
methanol, ethanol, and THF, but reacts with these solvents.
(24) Allegra, G.; Wilson, G. E.; Benedetti, E.; Pedone, C.; Albert, R. J. Am.
Chem. Soc. 1970, 92, 4002–4007.
(26) These assessments were made with various terpene-like polyenes
possessing different substitution patterns in the terminal alkene position
(the typical site of initiation for a cation-π cyclization).
(27) Generally formed in higher amounts on large scale, this side-product
could be suppressed by using dilute reaction concentrations (0.01 M)
and adding a nitromethane solution of BDSB rapidly to the substrate.
(25) Snyder, S. A.; Treitler, D. S. Angew. Chem., Int. Ed. 2009, 48, 7899–
7903.
9
14306 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010

Direct Halonium-Induced Polyene Cyclizations
A R T I C L E S
Scheme 2. Retrosynthetic Analysis of Peyssonol A (3)
Scheme 3. Racemic Total Synthesis of the Proposed Structure of
Peyssonol A (3) from Acyclic Terpene Precursors 30 and 31a
completion by promoting formation of multiple rings, especially
when synthesizing tricyclic or tetracyclic materials. We note
as well that while catalytic versions of this reagent design are
conceivable (in terms of the sulfide), we have not pursued such
explorations since we anticipate that they would not be ideal in
many cases. For instance, the conversion of 8 into 14 (Entry 1,
Table 1) provides products with reactive olefins, which we
believe we are able to obtain in good yield only because there
is rapid consumption of the starting material prior to the
formation of significant amounts of product (which would likely
react with BDSB if it were formed via a slower, catalytic
process).
2. Further Explorations into the Power of BDSB: Total
and Formal Syntheses of Peyssonol A, Peyssonoic Acid A,
and Aplysin-20. Recent investigations with BDSB have centered
on exploring its reactivity with progressively longer polyenes,
especially trienes possessing unique (i.e., Z) stereochemistry in
hopes of accessing the frameworks of several complex and
structurally intriguing natural products. In particular, our atten-
tion was drawn to the structure of the secondary metabolite
peyssonol A (3, Scheme 2), a material that was obtained from
the Red Sea marine alga Peyssonnelia sp., that has been shown
to act as an allosteric inhibitor of the reverse transcriptases of
the Human Immunodeficiency Virus.7c,d To the best of our
knowledge, this compound is the only known natural product
possessing a cis-decalin framework likely arising from a
halonium-induced cation-π cyclization. As such, we felt it would
be an ideal proving ground to evaluate the power of BDSB to
effect a direct and highly challenging cation-π cyclization to
access a framework distinct from those we had previously
prepared.28
a Reagents and conditions: (a) BDSB (1.1 equiv), MeNO2, 0 °C, 30 s,
34% for substrate 30; BDSB (1.1 equiv), MeNO2, -25f25 °C, 15 min,
26% for substrate 31; (b) K2CO3 (5.0 equiv), MeOH, 40 °C, 30 min; (c)
(COCl)2 (2.0 equiv), DMSO (5.0 equiv), CH2Cl2, -78 °C, then substrate,
then Et3N (10 equiv), -78f-50 °C, 1 h; (d) SOCl2 (2.0 equiv), Et3N (6.0
equiv), CH2Cl2, -97 °C, 1 h, 91% over three steps; (e) 34 (1.2 equiv),
n-BuLi (1.4 M in hexanes, 1.1 equiv), THF, -78 °C, 20 min, then 26, -78
°C, 20 min, 63%; (f) TFA (5.0 equiv), Et3SiH (10 equiv), CH2Cl2, 0 °C,
30-90 min, 64%; (g) n-BuLi (1.4 M in hexanes, 1.2 equiv), THF, -78
°C, 20 min, then DMF (5.0 equiv), -78 °C, 20 min, 62%, (h) p-TsOH·H2O
(0.2 M in t-BuOH), 65 °C, 2 h, 91%.
projecting a nucleophilic addition onto the aldehyde within
compound 26 to effect its incorporation, might afford the most
efficient means to reach a suitable polyene cyclization precursor.
Compound 26 could potentially arise from cis-decalin 27, which
could in turn directly result from a bromonium-induced cation-π
cyclization of the (2E,6Z)-farnesol derivative 28. Either an
acetate or carbonate as group R within 28 would hopefully give
rise to the desired functionality within 27, assuming that the
cation-π cyclization could indeed be induced to proceed despite
the higher degree of strain anticipated in the transition state to
reach the requisite cis-fused ring system.
As indicated in Scheme 2, our retrosynthetic analysis sug-
gested that a late-stage disconnection of the pendant aryl ring,
(28) Examples of polyene cyclizations involving Z-alkene geometries to
prepare cis-fused decalin systems are in fact quite rare. For the seminal
example, see: (a) Smit, W. A.; Semenovzky, A. V.; Kucherov, V. P.
Tetrahedron Lett. 1964, 5, 2299–2306. For a more recent example,
see: (b) Snowden, R. L.; Eichenberger, J.-C.; Linder, S. M.; Sonnay,
C. V.; Schulte-Elte, K. H. J. Org. Chem. 1992, 57, 955–960. In general,
such systems are prepared by other methods, including the cyclization
of partially cyclized materials, formation of the ring junction using a
Friedel-Crafts approach, or post-cyclization modification of a trans-
fused system. (c) Saito, A.; Matsushita, H.; Kaneko, H. Chem. Lett.
1984, 591–594. (d) Ishihara, K.; Ishibashi, H.; Yamamoto, H. J. Am.
Chem. Soc. 2002, 124, 3647–3655. (e) Bhar, S. S.; Ramana, M. M. V.
Tetrahedron Lett. 2006, 47, 7805–7807. (f) Von Schlatter, H.-R.;
Lu¨thy, C.; Graf, W. HelV. Chim. Acta 1974, 57, 1044–1055. (g)
Raeppel, F.; Heissler, D. Tetrahedron Lett. 2003, 44, 3487–3488.
The translation of this general plan into a synthesis of the
proposed structure for peyssonol A (3) proceeded largely without
incident as shown in Scheme 3. Thus, commercially available
nerol (29) was advanced into polyene cyclization precursors 30
and 31 in six steps each through a series of previously disclosed
transformations,29 details of which can be found in the Sup-
porting Information section. Subsequent exposure of these
materials separately to 1.1 equiv of BDSB in nitromethane
(29) (a) Kato, T.; Suzuki, M.; Toyohiko, K.; Moore, B. P. J. Org. Chem.
1980, 45, 1126–1130. (b) Yu, J. S.; Kleckley, T. S.; Wiemer, D. F.
Org. Lett. 2005, 7, 4803–4806.
9
J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010 14307

A R T I C L E S
Snyder et al.
Scheme 4. Racemic Total Syntheses of Various Diastereomers of Peyssonol A (40, 45, and 50), Ultimately Revealing that 50 is the Correct
Structure
afforded access to cis-decalin 32 in 34% yield from 30 and its
homologue 33 in 26% yield from 31. Although the efficiency
of these transformations is not as high as it was for many of
the substrates we explored previously, the strain within the cis-
fused transition states leading to 32 and 33 is significantly higher
than that for the corresponding trans-fused system.30 In fact, to
the best of our knowledge, these cyclizations constitute the first
examples of halonium-induced cation-π cyclizations leading to
cis-decalin frameworks, with an X-ray crystal structure of 33
(see Supporting Information section) confirming the stereo-
chemical assignment.
In any event, both 32 and 33 could be funneled into 26
through ester or carbonate hydrolysis as achieved with K2CO3
in MeOH, oxidation of the resultant primary alcohol, and
regioselective elimination of the remaining tertiary alcohol as
achieved with SOCl2 and Et3N in CH2Cl2 at -97 °C. Use of
warmer temperatures or a less hindered base (such as pyridine)
in this final step led to the formation of significant amounts of
the regioisomeric trisubstituted alkene.31 The remainder of the
sequence proceeded smoothly as designed, with only 4 ad-
ditional steps needed to complete a total synthesis of structure
3. To our surprise, however, comparison of the spectral
properties of synthetic 3 to those reported for natural peyssonol
A revealed stark differences; the inset table within Scheme 3
highlights several key, and readily identifiable, peaks from their
respective 1H NMR spectra. As such, assuming that our
stereochemical assignment for synthetic 3 was accurate, the
reported structure for peyssonol A (3) would have to be
incorrect.
To confirm this hypothesis, especially given the potential for
epimerization during the formation or subsequent arylation of
aldehyde 26, we synthesized cis-decalin 40 (see Scheme 4) with
altered stereochemistry at C-9 (the highlighted center). This
compound was readily prepared utilizing the same 8 step
sequence, with reduced cation-π cyclization efficiencies noted
for conversion of (Z,Z)-isomers 36 and 37 into polycycles 38
and 39 (20 and 28% yield, respectively). More important,
however, was that the homologue of aldehyde 26 (cf. Scheme
3) obtained through this sequence had a unique 1H NMR
spectrum, thus suggesting that neither material had been
epimerized; all other intermediates were distinct as well. As a
result, we concluded that the stereochemical integrity of our
assignments had not been compromised. Unfortunately, the
spectral data of 40 also did not match those reported for natural
peyssonol A.
Thus, based on these results, coupled with the fact that no
other cis-decalin natural products of this type are known, we
hypothesized that the correct orientation for these rings must
include a trans-decalin framework, despite the arguments
counter to this analysis presented in the original isolation paper.7c
Consequently, we prepared the two C-9 diastereomers of such
a trans-ring fusion (i.e., 45 and 50), and found that compound
50 had nearly identical 1H and 13C spectral data to those
published for the natural isolate;32 a crystal structure of this
final product was obtained as well, thereby removing any
(30) The rate determining transition state for the formation of products 32
and 44 is associated with the first C-C bond formation. The activation
barrier leading to 32 was estimated to be approximately 4.3 kcal/mol
higher in energy than that leading to 44. Geometries and energetics
were obtained from semi-emperical (PM3) calculations using the
GAMESS suite of programs. Schmidt, M. W.; Baldridge, K. K.; Boatz,
J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.;
Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.;
Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347–1363.
(31) We note that the use of a bulkier base such as i-Pr2NEt afforded inferior
regioselectivity in this dehydration step.
9
14308 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010

Direct Halonium-Induced Polyene Cyclizations
A R T I C L E S
Scheme 5. Racemic Total Synthesis of Peyssonoic Acid A (51) via
a BDSB-Induced Cation-π Cyclization and a Terminating
Regioselective Ring-Opening/Elimination Reactiona
potential ambiguity concerning the stereochemical integrity of
our sequence.33 As such, we believe that 50 reflects the true
configuration of peyssonol A, a reassignment strengthened by
the fact that it matches the carbon framework of peyssonoic
acid A (51), a compound which was recently obtained from a
related marine alga along with the rearranged framework
peyssonoic acid B (52).34 These materials all possess an
uncommon stereochemical configuration at C-9, one which
places the large substituent axial; to the best of our knowledge,
this synthesis of 50 constitutes a rare example of forming any
such framework through an electrophilic-induced polyene
cyclization.35 It is also worth noting that the BDSB-induced
cation-π cyclization leading to this final structure was the highest
yielding of all four diastereomers of t-butyl farnesyl carbonate,
with an optimized yield of 56% obtained for tricycle 49.
Intriguingly, the (E,E)-farnesol-derived substrates 41 and 42 also
provided a fair amount (26 and 17% yield, respectively) of the
cation-π cyclization products possessing the axial C-9 orienta-
tion of revised peyssonol A (i.e., 48 and 49) in addition to the
expected materials (i.e., 43 and 44), thus reflecting a shift in
reaction trajectory away from an all-chair conformation.36
A
similar switch in selectivity was recently observed by Shenvi
and Corey using a differentially protected oxygen-linked
termination group in the same position along the carbon
framework as 41 and 42.3m
Our next efforts sought to achieve additional refinement in
the route to 50 to determine whether a sequence could be
developed in which the aromatic ring was incorporated prior to
cation-π cyclization, since much of the overall step count
derived from the postcyclization incorporation of this unit. We
hoped that such an approach would also enable a total synthesis
of peyssonoic acid A (51) to be achieved, assuming that its
alternate double bond location relative to peyssonol A could
be formed readily and selectively. Scheme 5 presents those
endeavors, efforts which were able ultimately to achieve the
total synthesis of peyssonoic acid A (51), but not an enhanced
preparation of peyssonol A (50).
(32) The Supporting Information section contains a complete comparison
table. Compound 50 possesses 1H NMR data that are in excellent
agreement with the natural isolate; their 13C spectra possess small
discrepancies, most of which are 0.5 ppm or less. Efforts based on
altering water content as well as adding base or acid never afforded
spectra that were perfectly identical in terms of reported values, though
such factors can affect the 13C NMR spectra of compounds of this
type. None of the other isomers synthesized (3, 40, and 45) had 1H or
13C NMR data that even remotely resembled those of the natural
product. Prof. Kashman (Tel Aviv University) was contacted in hopes
of obtaining either a natural sample or copies of the original physical
spectra of peyssonol A, but unfortunately neither could be located,
making direct comparison impossible.
a Reagents and conditions: (a) 34 (1.0 equiv), n-BuLi (2.5 M in hexanes,
1.0 equiv), THF, -78 °C, 20 min, then Cul (0.5 equiv), 0 °C, 10 min, then
allyl bromide (3.0 equiv), 0 °C, 20 min, 75%; (b) PBr3 (0.5 equiv), Et2O,
-20f0 °C, 1 h; (c) 54 (1.7 equiv), n-BuLi (2.5 M in hexanes, 1.7 equiv),
THF, -78 °C, 20 min, then s.m., -40f5 °C, 2 h, 74% over two steps; (d)
BDSB (1.1 equiv), MeNO2, -25 °C, 5 min, 31%; (e) OsO4 (0.2 equiv),
NalO4 (5.0 equiv), pyridine (3.0 equiv), THF/t-BuOH/H2O (4/1/3), 0f25
°C, 2 h, 89%; (f) NaClO2 (5.0 equiv), NaH2PO4 (10 equiv), 2-methyl-2-
butene (10 equiv), THF/t-BuOH/H2O (5/2/3), 0 °C, 20 min, 81%; (g) BCl3
(1.0 M in CH2Cl2, 6.0 equiv), CH2Cl2, -78 °C, 1 h, 72%; (h) BDSB (1.1
equiv), MeNO2, -25 °C, 5 min, 42%.
(33) The Supporting Information section contains the complete synthetic
route employed to access the revised structure of peyssonol A (50),
for which further route improvement allowed for the shortening of
the sequence described in Scheme 3 by one step to ultimately achieve
an overall yield of 14% from (2Z,6E)-farnesol.
(34) (a) Lane, A. L.; Mular, L.; Drenkard, E. J.; Shearer, T. L.; Engel, S.;
Fredericq, S.; Fairchild, C. R.; Prudhomme, J.; Le Roch, K.; Hay,
M. E.; Aalbersberg, W.; Kubanek, J. Tetrahedron 2010, 66, 455–461.
For a review on misassigned natural product structures, see: (b)
Nicolaou, K. C.; Snyder, S. A. Angew. Chem., Int. Ed. 2005, 44, 1012–
1044. For recent examples of reassignment involving a major
architectural change within two fused 6-membered carbon rings, see:
(c) Maugel, N.; Mann, F. M.; Hillwig, M. L.; Peters, R. J.; Snider,
B. B. Org. Lett. 2010, 12, 2626–2629. (d) Spangler, J. E.; Carson,
C. A.; Sorensen, E. J. Chem. Sci. 2010, 1, 202–205.
Our sequence began by adding an allylated form of building
block 34 (i.e., 54) onto a (2Z,6E)-farnesyl backbone to forge
cation-π cyclization precursor 55. The allyl group was incor-
porated onto the aromatic ring to enable the eventual generation
of the aryl acetic acid moiety of peyssonoic acid A (51) through
oxidative cleavage. In addition, however, this monosubstituted
double bond would provide a critical test for olefin chemose-
(35) For an example of an epoxide-induced cyclization leading to such a
framework, see: van Tamelen, E. E.; Coates, R. M. Bioorg. Chem.
1982, 11, 171–196.
(36) Specifically, these products should arise if the second ring forms as a
boat.
9
J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010 14309

A R T I C L E S
Snyder et al.
Scheme 6. Racemic Formal Total Synthesis of Aplysin-20 (64)
lectivity in the key BDSB-induced cyclization. Pleasingly,
exposure of 55 to BDSB in nitromethane for 5 min at -25 °C
afforded materials in which the allyl group remained intact; the
isolated yield of 56 was 31%, thereby reflecting a cyclization
efficiency of 68% per ring. From 56, the remainder of the
sequence occurred smoothly, with the key operation being a
terminating exposure to excess BCl3 in CH2Cl2 at -78 °C for
1 h which served to remove the protecting group and cleave
the C-O bond at C-8, regioselectively affording the trisubsti-
tuted alkene of the target molecule (51).37,38 Peyssonoic acid
A (51) could also be accessed from polycycle 59 (prepared from
58 in 42% yield with BDSB) through a sequence involving
initial lithiation and addition of CO2 to afford a carboxylic acid
that was then homologated via an Ardnt-Eistert sequence; this
route, unfortunately, proceeded in significantly reduced yield
relative to that of Scheme 5. In no case, however, were we ever
able to convert tetracyclic materials like 56 or 59 into exocyclic
alkenes, despite numerous attempts. As such, the route described
earlier for peyssonol A (cf. Scheme 4) proved to be the only
one capable of delivering the desired functionality chemose-
lectively.
Patterned from the Work of Murai (ref 15b) in which a BDSB-Based
Cyclization Was Utilized to Access 65 Directly from 60a
As a final investigation into the power of BDSB to cyclize
trienes, we then targeted a formal total synthesis of the natural
product aplysin-20 (64, Scheme 6).39 This unique bicycle was
synthesized by Murai and co-workers in 198415b through a route
which employed a Lewis acid-catalyzed polyene cyclization of
protected bromohydrin derivative 61, a compound formed in 2
steps from the known nitrile 60.40 When the key cyclization
reaction was conducted with BF3•OEt2 in CH2Cl2 at reflux for
40 min, polycycles 62 and 63 were obtained in 53 and 14%
yield, respectively, from 61. Of these 4 cyclized diastereo- and
regio-isomers, only 2 (39% combined yield) had the proper
configuration (both -Br and -CH2CN in equatorial, i.e. ꢀ-posi-
tions) to be advanced to the natural product.
a Reagents and conditions: (a) BDSB (1.1 equiv), nitromethane, 25 °C,
5 min, 72%, 5.3:1.3:1.0 tri:di:tetrasubstituted alkenes.
In an effort to render this sequence far more direct, we found
that BDSB could convert 60 directly into 65 (as a 5.3/1.3/1.0
mixture of all alkene regioisomers) in 72% isolated yield. When
using BDSB, in contrast to Murai’s bromoacetate cyclization,
stereochemical control was observed at the highlighted center
for the di- and trisubstituted alkene forms of 65, indicative of
the strong preference for a chair-chair transition state as well
as the synchronous nature of this cyclization.2 Thus, all of the
cyclized products (65) could formally be advanced to the natural
product.
As a concluding comment on the uniqueness of BDSB as a
reagent to effect polyene cyclizations, we note that many
variants are not as effective overall, either due to challenges in
their preparation or their global reactivity. For instance, attempts
to prepare aryl variant 66 (Figure 3) have failed, due entirely
Figure 3. Structures of selected BDSB derivatives (66-69) that have been
investigated; of these, 67-69 have all been prepared successfully.
to the reagent brominating itself; this problem can be avoided
by prehalogenating the rings to form reagents such as 67, but
these materials are not readily solidified or handled. By contrast,
carbonyl variants such as 68 and 69 are easily prepared and
crystallized but, interestingly, afford reduced stereocontrol in
cation-π cyclizations, suggesting that they may react through a
different mechanism.
3. Synthesis and Reactivity of IDSI: Application to the
Formal Total Syntheses of Loliolide, K-76, and Stemodin. We
next sought to determine whether or not a related iodine variant
of BDSB could be prepared. After several failed attempts, we
were able to synthesize a crystalline form of such a material by
combining molecular I2, Et2S, and SbCl5 in 1,2-dichloroethane
followed by the addition of hexanes to a saturated solution of
the reagent prior to cooling.41 We have termed this material
(70, Scheme 7) IDSI on the basis of what we hoped would be
(37) The reaction temperature was essential to preventing lactone formation
between the pendant carboxylic acid and the adjacent phenol.
(38) The synthetic sequence produced the fully protonated version of
peyssonoic acid A. Initial NMR spectroscopic analysis of synthetic
51, however, did not match that reported for the natural isolate (ref
34), primarily around the carboxylic acid residue, aromatic ring, and
adjoining methylene group. Subsequent exposure of our material to
NaHCO3 provided the sodium salt of the carboxylate form; spectra
obtained from this material fully matched the reported data. See
Supporting Information for all relevant spectra and NMR data tables.
(39) (a) Matsuda, H.; Tomiie, Y.; Yamamura, S.; Hirata, Y. J. Chem. Soc.,
Chem. Commun. 1967, 898–899. (b) Yamamura, S.; Hirata, Y. Bull.
Chem. Soc. Jpn. 1971, 44, 2560–2562.
(40) Kato, T.; Kumazawa, S.; Kabuto, C.; Honda, T.; Kitahara, Y.
Tetrahedron Lett. 1975, 16, 2319–2322. A total synthesis of aplysin-
20 was also achieved via a TBCO-mediated cation-π cyclization of a
related starting material, albeit in low yield (cf. ref 11i).
(41) IDSI (70) is fully soluble at 25 °C in MeNO2, EtNO2, MeCN, CH2Cl2,
DMSO, DMF, acetone, and EtOAc. It is moderately to slightly soluble
in dioxane and CHCl3, and insoluble in benzene, toluene, and hexanes.
9
14310 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010

Direct Halonium-Induced Polyene Cyclizations
A R T I C L E S
Similar results were obtained with NIS/Ph3P.43 Of course,
materials like 73 can be converted into 72 in a subsequent step
through the addition of acid; however, IDSI (like BDSB),
typically avoids the need for this additional step as an acidic
byproduct is produced during the course of the cyclization which
can complete the sequence effectively in most cases, thereby
enabling a more direct and efficient synthetic protocol.44
Table 2 provides our preliminary survey of IDSI reactivity
with various electron-rich and electron-poor substrates derived
from geraniol, farnesol, and nerol, each of which was performed
with 0.1 mmol of substrate at a reaction concentration of 0.05
M. In the electron-rich cases (Entries 1-4), cyclization yields
were commensurate with those observed previously with BDSB
with equally fast reaction times, and only in the case of the
conversion of 21 into 78 was an added acid needed at the end
of the sequence to achieve complete cyclization. For electron-
deficient systems (Entries 5-8; the final entry includes a nerol
derivative), IDSI also worked well, though the use of various
oxygen species to terminate those processes were not as efficient
as BSDB. The main side-product in all of these cases was an
uncyclized vicinal chloroiodide such as acetate 86 (formed from
attempted IDSI cyclization of 15, Entry 6), revealing that IDSI
may have potential as an effective ICl source outside of polyene
cyclizations. In any event, it is important to note that Entries
5-8 represent, to the best of our knowledge, the first examples
of successful iodonium-based cyclizations of electron-deficient
polyenes. All product stereochemistries were established based
on comparison to previously synthesized materials and/or
literature data.
Scheme 7. Synthesis of IDSI (70)
Scheme 8. Evaluation of Different Electrophilic Iodine Sources in
Achieving Iodonium-Based Cyclizations of Substrate 71a
a Reagents and conditions: (a) IDSI (1.2 equiv), MeNO2, -25 °C, 5 min,
93%; (b) Ipy2BF4 (1.0 equiv), HBF4 (1.0 equiv), CH2Cl2, -40 °C, 3 h,
41%; (c) NIS (1.0 equiv), Ph3P (0.3 equiv), CH2Cl2, -78 °C, 24 h; -40
°C, 6 h, 1.8:1.0 of 73:71 (<5% 72).
On a global level, however, the true value in a direct and
high yielding iodine-based cyclization lies not in forming an
iodinated material (as there are no natural products isolated to
date resulting from iodonium-induced cation-π cyclizations) but
rather in the ability to couple, displace, or easily eliminate the
alkyl iodide within the product. For instance, we were able to
readily form an alkene (i.e., 87) in 86% yield from 72 with
DBU in refluxing pyridine; the corresponding bromide is far
more robust and does not participate in such chemistry.45 As
such, it seemed reasonable, given the established cyclization
scope and capability for further iodine functionalization, to
attempt to utilize IDSI to render more efficient and/or expedi-
tious several previous total syntheses of various nonhalogenated
natural product polycycles, particularly those cases where
stoichiometric amounts of metals were required for success.
For instance, in 1983, Rouessac and co-workers46 synthesized
the natural product loliolide (92, Scheme 9)47 through a Hg(II)-
based polyene cyclization of 88, which, following replacement
of the organomercurial with iodine under radical conditions16
and subsequent elimination, afforded key alkene 91 in 25%
overall yield. In our hands, exposure of 88 to 1.2 equivalents
of IDSI afforded cation-π cyclization product 93 in 79% yield
reactivity equivalent to BDSB in polyene cyclizations, given
that the reagent itself does not possess a structure or level of
stability commensurate to BDSB. Indeed, X-ray diffraction
revealed that IDSI is actually a chlorine-linked dimer, one whose
crystalline form requires a maximum of -20 °C for effective
storage; in addition, though the reagent can be weighed normally
in air, it will decompose relatively quickly (within 30 min at
25 °C) if not properly attended, losing ICl in the process.42 The
inset picture of some needles within Scheme 7 shows this
process through the discoloration of the paper on which the solid
has been placed. Despite these differences, however, IDSI is
quite effective and just as chemoselective as BDSB for initiating
polyene cyclizations.
For instance, as shown in Scheme 8, exposure of polyene 71
to 1.2 equiv of IDSI in nitromethane at -25 °C for 5 min at a
reaction concentration of 0.05 M afforded polycycle 72 as a
single diastereomer in 93% yield. By contrast, neither
Ishihara’s12a nor Barluenga’s reagent combinations12b were
nearly as effective. For instance, in the case of the latter species,
we obtained (after multiple attempts using various solvents and
differential amounts of added HBF4) an optimized 41% yield
of 72, with other major products being the partially cyclized
73, unidentified diastereomers of 72, and proton cyclized 74.
(43) See Supporting Information section for complete details.
(44) We note, however, that IDSI should be viewed as having comple-
mentary reactivity to Barluenga’s reagent. For instance, electron-rich
aromatic rings undergo electrophilic aromatic substitution when
pendant monosubstituted double bonds are activated with Ipy2BF4;
IDSI will not cleanly perform such reactions.
(45) For selected examples of the elimination of alkyl iodides to form
alkenes, see: (a) Furber, M.; Kraft-Klaunzer, P.; Mander, L. N.; Pour,
M.; Yamauchi, T.; Murofushi, N.; Yamane, H.; Schraudolf, H. Aust.
J. Chem. 1995, 48, 427–444. (b) Jin, L.; Nemoto, T.; Nakamura, H.;
Hamada, Y. Tetrahedron: Asymmetry 2008, 19, 1106–1113. (c)
Eidman, K. F.; MacDougall, B. S. J. Org. Chem. 2006, 71, 9513–
9516. (d) Soorukram, D.; Qu, T.; Barrett, A. G. M. Org. Lett. 2008,
10, 3833–3835.
(42) The lengths of the I-Cl bonds within this material are 2.814 and 2.714
Å. These values compare favorably to related compounds such as a
chlorine-linked NIS-dimer which has very similar bond lengths (2.845
and 2.910 Å) and is a source of ICl as well: Ghassenzadeh, M.;
Dehnicke, K.; Goesmann, H.; Fenske, D. Zeit. Naturforschung B:
Chem. Sci. 1994, 49, 602–608. We note that the average I-Cl bond
length is 2.553 Å according to the Cambridge Structural Database,
version 5.31, 2009. The dimethysulfonium variant of IDSI was reported
as a monomeric species: Minkwitz, R.; Prenzel, H. Z. Anorg. Alleg.
Chem. 1987, 548, 97–102.
9
J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010 14311

A R T I C L E S
Snyder et al.
Scheme 9. Racemic Formal Total Synthesis of Loliolide (92)
Patterned from the Work of Rouessac (ref 46) in which an
IDSI-Based Cyclization and Subsequent LiCl-Induced Elimination
Was Utilized to Access 91a
Table 2. Exploration of the Generality of Direct, Iodonium-Induced
Cation-π Cyclizations using IDSI (1.2 equiv) and 0.1 mmol of
Substrate in Nitromethane
a Reagents and conditions: (a) TFAA (5.0 equiv), t-BuOH, 0 to 25 °C,
20 min, 68%; (b) IDSI (1.2 equiv), MeNO2, 5 min, 79% and 19:1 d.r. from
88 at -25 °C, 88% from 89 at 0 °C; (c) LiCl (50 equiv), DMF, 80 °C,
12 h, 97%.
in fewer steps) from 88, without the use of stoichiometric Hg(II).
Similarly, IDSI proved quite effective in our efforts to prepare
96 (Scheme 10), a key intermediate in the McMurry and Erion
total synthesis48 of K-76 (97)49 reported in 1985. In this case,
bicycle 98 was prepared in 77% yield using IDSI, illustrating
its utility as a powerful cation-π initiator as even the very
electron-deficient olefin within 94 participated in this cyclization
reaction. Typically, such non-nucleophilic olefins (in this case
an R,ꢀ-unsaturated ester) do not participate in cation-π cycliza-
tions unless Hg(II) is utilized.1,4 A subsequent elimination using
DBU at elevated temperatures provided the requisite alkene 96
in 66% overall yield for the two-step sequence. This outcome
compares favorably to the 53% overall yield obtained over the
4 steps of the McMurry and Erion route in which stoichiometric
Hg(II) and Se were employed.50 It should be noted that in our
hands neither NIS/Ph3P nor Ipy2BF4/HBF4 was able to fully
cyclize the same substrate of Scheme 10 (i.e., 94) in any yield.43
It must be mentioned, however, that Hg(II)-based cyclizations
certainly do have merit. For instance, in the Corey total
synthesis51 of stemodin (101, Scheme 11),52 polyene 99 was
(46) (a) Rouessac, F.; Zamarlik, H.; Gnonlonfoun, N. Tetrahedron Lett.
1983, 24, 2247–2250. For other papers in the series, see: (b) Rouessac,
A.; Rouessac, F.; Zamarlik, H. Bull. Chim. Soc. Fr. 1981, 199–203.
(c) Zamarlik, H.; Gnonlonfoun, N.; Rouessac, F. Can. J. Chem. 1984,
62, 2326–2329.
a Isolated as a 2:1 mixture of inseparable stereoisomers about the
highlighted carbon atom favoring the drawn diastereomer. b MeSO3H
(15 equiv) added with 1 h of additional stirring to promote the final
cyclization. c Produced as a 8.5:1.4:1.0 mixture of tri:tetra:disubstituted
alkene isomers.
(47) (a) White, E. P. New Zealand J. Agric. Res. 1958, 1, 859–865. (b)
Hodges, R.; Porte, A. L. Tetrahedron 1964, 20, 1463–1467.
(48) (a) McMurry, J. E.; Erion, M. D. J. Am. Chem. Soc. 1985, 107, 2712–
2720. (b) Erion, M. D.; McMurry, J. E. Tetrahedron Lett. 1985, 26,
559–562.
(49) Kaise, H.; Shinohara, M.; Miyazaki, W.; Izawa, T.; Nakano, Y.;
Sugawara, M.; Sugawara, K. J. Chem. Soc., Chem. Commun. 1979,
726–727.
with 19:1 diastereoselectivity at the bridgehead methyl position,
while the use of the t-butyl ester-protected variant (89, formed
in 68% yield from 88) enabled an IDSI-based synthesis of 93
as a single diastereomer in 88% yield. Subsequent LiCl-induced
elimination afforded 91 in 97% yield, thereby accounting for
an overall yield of 73% of alkene 91 (an ∼3 fold improvement
(50) For an alternate total synthesis of K-76, see: Mori, K.; Komatsu, M.
Liebigs Ann. Chem. 1988, 107–119. This route did not utilize a
cation-π cyclization, but used the same key intermediate, compound
96, to achieve their synthesis. Compound 96 was accessed in 9 steps
and 2.5% overall yield from commercial materials.
(51) Corey, E. J.; Tius, M. A.; Das, J. J. Am. Chem. Soc. 1980, 102, 7612–
7613.
9
14312 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010

Direct Halonium-Induced Polyene Cyclizations
A R T I C L E S
Scheme 10. Racemic Formal Total Synthesis of K-76 (97)
Patterned from the Work of McMurry (ref 48) in which an
IDSI-Based Cyclization and Subsequent DBU-Induced Elimination
Was Utilized to Access 96a
Scheme 12. Synthesis of CDSC (103)
our hands, only a portion of these could only be successfully
converted into 102 through the use of an acid-promoted
cyclization (concentrated H2SO4 in toluene) in a separate step;
extensive efforts to differentially functionalize the enol ether
in 99 (including groups such as a methyl-, methoxymethyl-,
and various silyl-enol ethers) afforded no improvement above
the 40% yield indicated within Scheme 11. Thus, in this case,
the overall yield of the polycycle was not superior through the
use of IDSI, though the toxic metal species used for the polyene
cyclization could still be avoided.
4. Synthesis and Reactivity of CDSC. We next sought to
determine if direct, chloronium-induced cyclizations could be
achieved with a reagent of the general design of BDSB and
IDSI. The synthesis of our test reagent, a derivative of a
previously reported Me2S variant54 which we name CDSC
(chloro diethylsulfonium hexachloroantimonate, 103), is shown
in Scheme 12.
a Reagents and conditions: (a) IDSI (1.2 equiv), MeNO2, -25 °C, 5 min,
77%; (b) DBU (20 equiv), pyridine, 120 °C, 12 h, 86%.
Similar to BDSB and IDSI, this compound is a crystalline
solid that is stable at -20 °C for at least several months and
can be handled and weighed in air. As indicated in Table 3,
polyene cyclizations of various materials possessing differential
electron wealth were undertaken with CDSC, all at a reaction
concentration of 0.05 M. Though the resultant product yields
are not nearly as high as those observed with BDSB and IDSI
for the same substrates, these entries represent, to the best of
our knowledge, the first examples of effecting chloronium-
induced polyene cyclizations in any yield via an ionic pathway.10
Of note, these cyclizations do not, for the most part, possess
diastereocontrol, perhaps reflecting a global challenge in reactiv-
ity due to greater tertiary carbocation rather than bridged
chloronium-character in the initial reactive intermediate (as
indicated by the structures at the bottom of Table 3).55
5. Efforts toward Asymmetric Induction. Finally, we desired
to prepare chiral versions of CDSC, BDSB, and IDSI in a
preliminary investigation of their potential to achieve asym-
metric versions of the reactions described above. Although
Scheme 11. Racemic Formal Total Synthesis of Stemodin (101)
Patterned from the Work of Corey (ref 51) in which an IDSI-Based
Cyclization Was Utilized to Access 102a
(55) Several experimental studies (pioneered by the Olah group) have
postulated that some halonium ions, especially chloronium ions, exist
not as traditional 3-membered rings but as “open” carbocations, or
perhaps as an equilibrium between the two species. This appears
especially true when one or more of the carbons of the potential
chloronium ion is tertiary: (a) Olah, G. A.; Westerman, P. W.; Melby,
E. G.; Mo, Y. K. J. Am. Chem. Soc. 1974, 96, 3565–3573. (b) Olah,
G. A.; Bollinger, J. M.; Mo, Y. K.; Brinich, J. M. J. Am. Chem. Soc.
1972, 94, 1164–1168. (c) Olah, G. A.; Bollinger, J. M. J. Am. Chem.
Soc. 1968, 90, 947–953. (d) Berman, D. W.; Anicich, V.; Beauchamp,
J. L. J. Am. Chem. Soc. 1979, 101, 1239–1248. A more recent NMR
analysis of tetrasubstituted chloronium ions indicates that open
structures predominate. (e) Ohta, B. K.; Hough, R. E.; Schubert, J. W.
Org. Lett. 2007, 9, 2317–2320. Theoretical studies also corroborate
that the open carbocation is favored in the specific case of the
trisubstituted chloronium ion. (f) Yamabe, S.; Tsuji, T.; Hirao, K.
Chem. Phys. Lett. 1988, 146, 236–242. As noted by a referee, a
plausible alternative explanation for the lack of stereocontrol in
chloronium ion cyclizations is reaction through both chair/chair and
boat/chair conformations, perhaps owing to the higher reactivity and
electrophilicity of the chloronium ion electrophile compared to the
bromonium and iodonium counterparts.
a Reagents and conditions: (a) IDSI (1.2 equiv), MeNO2, 25 °C, 5 min;
(b) conc. H2SO4 (15 equiv), toluene, 0 °C, 30 min, 40% over two steps.
smoothly converted into 100 in 60% yield via treatment with
Hg(OCOCF3)2 to effect the cyclization followed by replacement
of the intermediate organomercurial with iodine.53 IDSI was
able to form similar materials from 99, but in reduced yield as
the predominant products obtained were partially cyclized. In
(52) Manchand, P. S.; White, J. D.; Wright, H.; Clardy, J. J. Am. Chem.
Soc. 1973, 95, 2705–2706.
(53) The Corey group also used a similar route to prepare the natural product
K-76: Corey, E. J.; Das, J. J. Am. Chem. Soc. 1982, 104, 5551–5553.
(54) Meerwein, H.; Zenner, K.-F.; Gipp, R. Justus Liebigs Ann. Chem. 1965,
67–77.
9
J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010 14313

A R T I C L E S
Snyder et al.
Table 3. Exploration of the Generality of Direct,
Chloronium-Induced Cation-π Cyclizations using CDSC (1.1 equiv)
and 0.1 mmol of Substrate in Nitromethane
all putative structures). Unfortunately, all endeavors with these,
and related compounds, afforded no asymmetry in the cation-π
cyclization of substrate 71, though they all led to the formation
of the expected racemic products. Interestingly, however, with
reagent 107 we were able to add Cl2 across the double bond of
111 with some enantioselection (up to 14% e.e.);58 initial screens
have shown that solvent is a critical factor in the efficiency of
this process, suggesting that further refinement may enable
improvement on this preliminary finding. It is important to note
that the reagent formed with the omission of SbCl5 did not afford
any 112, indicative of the importance of the normally inert SbCl6
counterion. Explorations seeking to build upon these initial
results are the subject of current endeavors.
Conclusion
Although direct halonium-induced polyene cyclizations have
long been achievable only through the use of enzymes in ViVo,
this article describes the first class of simple reagents that can
effect such reactions in a reaction flask for a broad range of
substrates. Consequently, the direct total and formal syntheses
of 6 different natural products were achieved, one of which
required a substantive structural revision following the synthesis
and cyclization of 4 stereochemically distinct polyene systems.
In addition, chiral forms of the reagent class have shown some
promise in enantioselective halogenations, indicating that they
have potential for a number of significant applications outside
of cation-π cyclizations.
a Isolated as a 1.0:1.0 mixture of inseparable stereoisomers. b Isolated
as a 2.2:1.0 mixture of separable diastereoisomers at the highlighted
carbon favoring the drawn product. c Produced as a 4.0:1.0 mixture of
separable diastereomers at the highlighted carbon favoring the drawn
product.
Acknowledgment. We thank Drs. J. Decatur and Y. Itagaki
for NMR spectroscopic and mass spectrometric assistance, respec-
tively. We acknowledge Prof. Gilbert Stork for several helpful
discussions and Mr. Adel ElSohly for performing our calculations.
We also thank the National Science Foundation (CHE-0619638)
for an X-ray diffractometer and Prof. Gerard Parkin, Mr. Wesley
Sattler, and Mr. Aaron Sattler for analyzing our crystal structures.
Financial support for this research was provided by Columbia
University, an NSF CAREER Award (CHE-0844593), the ACS
Petroleum Research Fund (47481-G), the Camille and Henry
Dreyfus Foundation (New Faculty Award to S.A.S.), Eli Lilly
(Grantee Award to S.A.S.), and the NSF (Predoctoral Fellowships
to D.S.T. and A.P.B.). S.A.S. is a fellow of the Alfred P. Sloan
Foundation and a Cottrell Scholar of the Research Corporation for
the Advancement of Science.
Scheme 13. Initial Explorations into Asymmetric Versions of the
Developed Reagent Class
Supporting Information Available: Experimental procedures,
compound characterization, copies of spectral data, and X-ray
crystallographic structures. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA106813S
(56) C2-symmetric ligands were chosen due to our previous success in
utilizing them for asymmetric Hg(II)-induced cyclizations (cf. ref
17).
(57) (a) Julienne, K.; Metzner, P. J. Org. Chem. 1998, 63, 4532–4534. (b)
Braun, W.; Calmuschi, B.; Haberland, J.; Hummel, W.; Liese, A.;
Nickel, T.; Steizer, O.; Salzer, A. Eur. J. Inorg. Chem. 2004, 11, 2235–
2243.
several chiral sulfides are known, we focused our attention on
materials with C2-symmetry,56 using a sequence involving an
enzymatically controlled step to synthesize (2R,5R)-(+)-2,5-
dimethylthiolane57 for reagents 107, 108, and 109 (Scheme 13,
(58) For a recent total synthesis featuring an enantioselective chlorination
of an isolated alkene, see: Snyder, S. A.; Tang, Z.-Y.; Gupta, R. J. Am.
Chem. Soc. 2009, 131, 5744–5745.
9
14314 J. AM. CHEM. SOC. VOL. 132, NO. 40, 2010