Multicomponent Linchpin Couplings. Reaction of Dithiane Anions with Terminal Epoxides, Epichlorohydrin, and Vinyl Epoxides: Efficient, Rapid, and Stereocontrolled Assembly of Advanced Fragments for Complex Molecule Synthesis

Article abstract of DOI:10.1021/ja0376238

The development, application, and advantages of a one-flask multicomponent dithiane linchpin coupling protocol, over the more conventional stepwise addition of dithiane anions to electrophiles leading to the rapid, efficient, and stereocontrolled assembly of highly functionalized intermediates for complex molecule synthesis, are described. Competent electrophiles include terminal epoxides, epichlorohydrin, and vinyl epoxides. High chemoselectivity can be achieved with epichlorohydrin and vinyl epoxides. For vinyl epoxides, the steric nature of the dithiane anion is critical; sterically unencumbered dithiane anions afford S_N2 adducts, whereas encumbered anions lead primarily to S_N2′ adducts. Mechanistic studies demonstrate that the S_N2′ process occurs via syn addition to the vinyl epoxide. Integration of the multicomponent tactic with epichlorohydrin and vinyl epoxides permits the higher-order union of four and five components.

Full text of DOI:10.1021/ja0376238

Published on Web 10/30/2003
Multicomponent Linchpin Couplings. Reaction of Dithiane
Anions with Terminal Epoxides, Epichlorohydrin, and Vinyl
Epoxides: Efficient, Rapid, and Stereocontrolled Assembly of
Advanced Fragments for Complex Molecule Synthesis
Amos B. Smith, III,* Suresh M. Pitram, Armen M. Boldi, Matthew J. Gaunt,
Chris Sfouggatakis, and William H. Moser
Contribution from the Department of Chemistry, Monell Chemical Senses Center and Laboratory
for Research on the Structure of Matter, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104
Received July 29, 2003; E-mail: smithab@sas.upenn.edu
Abstract: The development, application, and advantages of a one-flask multicomponent dithiane linchpin
coupling protocol, over the more conventional stepwise addition of dithiane anions to electrophiles leading
to the rapid, efficient, and stereocontrolled assembly of highly functionalized intermediates for complex
molecule synthesis, are described. Competent electrophiles include terminal epoxides, epichlorohydrin,
and vinyl epoxides. High chemoselectivity can be achieved with epichlorohydrin and vinyl epoxides. For
vinyl epoxides, the steric nature of the dithiane anion is critical; sterically unencumbered dithiane anions
afford SN2 adducts, whereas encumbered anions lead primarily to SN2′ adducts. Mechanistic studies
demonstrate that the SN2′ process occurs via syn addition to the vinyl epoxide. Integration of the
multicomponent tactic with epichlorohydrin and vinyl epoxides permits the higher-order union of four and
five components.
Dithianes have evolved as invaluable tools in organic
synthesis, serving primarily as acyl anion equivalents for
constructing carbon-carbon bonds.1,2 The efficient, convergent
construction of partially or fully protected aldol linkages in a
stereocontrolled manner employing epoxides as the electrophile
is a particularly attractive feature of dithiane chemistry and, as
such, serves as an effective alternative to the classical aldol
reaction. Indeed, an early prominent aspect of several synthetic
ventures undertaken in this laboratory comprised the assembly
of subtargets of varying structural complexity exploiting
dithianes as linchpins.3 In these early examples, two building
blocks were joined, one a substituted lithiated 1,3-dithiane, the
other an electrophile such as an iodide, epoxide, aldehyde, or
ketone.
adducts, exploiting a solvent-controlled Brook rearrangement
(Scheme 1).4 The tactic proved efficient and was subsequently
Scheme 1
employed with considerable success in the construction of the
spiroketal segments of the potent antitumor spongistatins.5
Extension of this tactic to a five-component process achieved
in a single flask enabled the rapid assembly of the extended
1,3-polyol segment of the polyene macrolides, mycoticins A
and B.6
In 1997, we disclosed a three-component, one-flask linchpin
tactic employing 2-tert-(butyldimethylsilyl)-1,3-dithiane with
two different epoxide electrophiles to construct unsymmetrical
(1) (a) Corey, E. J.; Seebach, D. Angew. Chem., Int. Ed. Engl. 1965, 4, 1075.
(b) Seebach, D.; Corey, E. J. J. Org. Chem. 1975, 40, 231.
(2) Reviews: (a) Seebach, D. Synthesis 1969, 1, 17. (b) Grobel, B. T.; Seebach,
D. Synthesis 1977, 357. (c) Page, P. C. B.; Van Niel, M. B.; Prodger, J. C.
Tetrahedron 1989, 45, 7643. (d) Kolb, M. In Encyclopedia of Reagents
for Organic Synthesis; Paquette, L. A., Ed.; John Wiley & Sons: Chichester,
1995; Vol. 5, pp 2983. (e) Yus, M.; Najera, C.; Foubelo, F. Tetrahedron
2003, 59, 6147.
(3) (a) Smith, A. B., III; Condon, S. M.; McCauley, J. A. Acc. Chem. Res.
1998, 31, 35 and references therein. (b) Smith, A. B., III; Lodise, S. A.
Org. Lett. 1999, 1, 1249. (c) Smith, A. B., III; Doughty, V. A.; Lin, Q.;
Zhuang, L.; McBriar, M. D.; Boldi, A. M.; Moser, W. H.; Murase, N.;
Nakayama, K.; Sobukawa, M. Angew. Chem., Int. Ed. 2001, 40, 191. (d)
Smith, A. B., III; Lin, Q.; Doughty, V. A.; Zhuang, L.; McBriar, M. D.;
Kerns, J. K.; Brook, C. S.; Murase, N.; Nakayama, K. Angew. Chem., Int.
Ed. 2001, 40, 196. (e) Smith, A. B., III; Adams, C. M.; Lodise Barbosa, S.
A.; Degnan, A. P. J. Am. Chem. Soc. 2003, 125, 350.
More recently, we developed the chemoselective addition of
lithiated dithianes to various vinyl epoxides. Selectivity between
the SN2 and SN2′ addition manifolds could be easily controlled
(4) Smith, A. B., III; Boldi, A. M. J. Am. Chem. Soc. 1997, 119, 6925.
(5) (a) Smith, A. B., III; Zhuang, L.; Brook, C. S.; Boldi, A. M.; McBriar, M.
D.; Moser, W. H.; Murase, N.; Nakayama, K.; Verhoest, P. R.; Lin, Q.
Tetrahedron Lett. 1997, 38, 8667. (b) Smith, A. B., III; Zhuang, L.; Brook,
C. S.; Lin, Q.; Moser, W. H.; Trout, R. E. L.; Boldi, A. M. Tetrahedron
Lett. 1997, 38, 8671. (c) Smith, A. B., III; Lin, Q.; Nakayama, K.; Boldi,
A. M.; Brook, C. S.; McBriar, M. D.; Moser, W. H.; Sobukawa, M.; Zhuang,
L. Tetrahedron Lett. 1997, 38, 8675.
(6) Smith, A. B., III; Pitram, S. M. Org. Lett. 1999, 1, 2001.
9
10.1021/ja0376238 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 14435-14445
14435

A R T I C L E S
Smith et al.
by appropriately tuning the steric properties of the dithiane
anion.7 Integration of the latter protocol with the earlier
multicomponent tactic has now led to the rapid generation of a
series of complex targets.
this reaction would have a wider impact. Critical to this
proposition would be control of the Brook rearrangement (Vide
infra).
Initial Studies. Our point of departure was 2-tert-(butyldi-
methylsilyl)-1,3-dithiane (8), a substrate also successfully
employed by Tietze.8 The more robust TBS protecting group
was anticipated to hold a significant synthetic advantage over
the TMS group. We also enlisted the lithiation conditions of
Williams and co-workers [e.g., t-BuLi in 10% hexamethylphos-
phoramide (HMPA)/THF at -78 °C],10 a protocol used exten-
sively in our laboratory for the effective deprotonation of simple
dithianes. The initial experiment called for generation of the
lithium anion of 8, a` la Williams, followed by the addition of
2.5 equiv of epoxide (-)-9; bis-adduct (+)-10 was obtained in
86% yield (Scheme 3). The dramatic improvement in reaction
time was quite pleasing.
In this, a full account, we present in turn the development,
application, and advantages of the one-flask multicomponent
dithiane linchpin coupling protocol over the more conventional
stepwise dithiane additions. The study comprises the reactions
of dithiane anions with terminal epoxides, epichlorohydrin, and
vinyl epoxides. Integration of these reactions led to new, higher-
order multicomponent unions capable of coupling four and five
components in a single flask. The overall goal of this program
was, and continues to be, the rapid, efficient, and stereocon-
trolled assembly of highly functionalized intermediates for
complex molecule synthesis.
Results and Discussion.
Scheme 3
Symmetrical Bisalkylation of 2-Silyl-1,3-dithianes. In 1994,
Tietze and co-workers reported a tandem bisalkylation of
2-trimethylsilyl-1,3-dithiane (1) with enantiomerically pure
epoxides.8 The procedure called for the addition of lithiated 1
(Scheme 2) to 2.2 equiv of epoxide (+)-2 in the presence of a
Scheme 2
Early studies to expand the scope of this reaction, employing
two different epoxides to furnish unsymmetrical adducts,
however, were not promising; when epoxides (-)-9 and (-)-
11 were added sequentially to lithiated 8 in THF (Scheme 4),
Scheme 4
the two symmetrical adducts (+)-10 and (+)-12b predominated
over the unsymmetrical adduct (+)-12a. Presumably, the initial
alkylation of 8 was rate-limiting, with the subsequent Brook
rearrangement occurring more rapidly, such that competitive
alkylation could occur with the newly generated dithiane anion
6, thereby leading to a mixture of adducts. Taken together, these
results suggested that the linchpin union of different electrophiles
would only be feasible if the Brook rearrangement could be
suppressed until the first alkylation was complete.
Solvent Effects on Brook Rearrangements. Utimoto and
co-workers, in an elegant study, revealed a dramatic solvent
effect on the Brook rearrangement with adducts derived from
the lithiation of dihalo(trialkylsilyl)methanes (Scheme 5).11
Addition of benzaldehyde to 13 provided a critical observation
crown ether to yield the pseudo-C2-symmetrical 1,5-diol (3) in
65% yield. The reaction was presumed to involve a 1,4-Brook
rearrangement9 after the initial alkylation process, wherein the
silyl group is transferred to the intermediate alkoxide (5),
generating a new dithiane anion (6) for a second alkylation
(Scheme 2). An appealing feature of this transformation was
the differentiation between the two resulting hydroxy groups
permitting subsequent selective manipulation. Long reaction
times (2 days) and use of a single epoxide, however, limited
the general synthetic utility of this reaction sequence. We
reasoned that if two different electrophiles could be employed,
(7) Smith, A. B., III; Pitram, S. M.; Gaunt, M. J.; Kozmin, S. A. J. Am. Chem.
Soc. 2002, 124, 14516.
(8) Tietze, L. F.; Geissler, H.; Gewert, J. A.; Jakobi, U. Synlett 1994, 511.
(9) (a) Brook, A. G. Acc. Chem. Res. 1974, 7, 77. (b) Brook, A. G.; Bassindale,
A. R. In Rearrangements Ground Excited States; de Mayo, P., Ed.;
Academic Press: New York, 1980; Vol. 2, pp 149. (c) Moser, W. H.
Tetrahedron 2001, 57, 2065 and references therein.
(10) (a) Williams, D. R.; Sit, S.-Y. J. Am. Chem. Soc. 1984, 106, 2949. See
also: (b) Merrifield, E.; Steel, P. G.; Thomas, E. J. J. Chem. Soc., Chem.
Commun. 1987, 1826. (c) Toshima, H.; Suzuki, T.; Nishiyama, S.;
Yamamura, S. Tetrahedron Lett. 1989, 30, 6725.
(11) (a) Shinokubo, H.; Miura, K.; Oshima, K.; Utimoto, K. Tetrahedron Lett.
1993, 34, 1951. (b) Shinokubo, H.; Miura, K.; Oshima, K.; Utimoto, K.
Tetrahedron 1996, 52, 503.
9
14436 J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003

Multicomponent Linchpin Couplings
A R T I C L E S
Table 1. Solvent Effects on the Brook Rearrangement
Scheme 5
yield (%)
yield (%)
yield (%)
entry
solvent
additive
(+)-22
(+)-23
(+)-10
1
2
3
4
Et2O
THF
Et2O
Et2O
-
88
6
-
-
-
33
83
91
-
14
-
-
HMPA
DMPU
-
the solvents was critical to achieve optimal yields, presumably
due to the instability of the lithiated 2-silyl-1,3-dithianes.
One-Flask Three-Component Couplings with 2-TBS-1,3-
dithianes. We next conducted a series of linchpin couplings
with the lithium anion of dithiane 8, exploiting the solvent
controlled Brook rearrangement to generate a diverse array of
unsymmetrical bis-adducts; in general, yields were good to
excellent (Table 2). The optimized protocol calls for deproto-
nation of 8 with t-BuLi in Et2O at -78 °C, warming the reaction
mixture to -45 °C over a 1 h period, followed by cooling to
-78 °C and addition of the first epoxide in Et2O. Completion
of the first alkylation is achieved in ca. 1 h at -25 °C. The
Brook rearrangement is then triggered by the addition of HMPA
or DMPU (0.3-0.4 equiv) in Et2O. Addition of the second
electrophile (2 equiv) in Et2O completes the reaction sequence.
Terminal scalemic epoxides comprise excellent electrophiles
for the three-component process, since the stereogenicity of the
resulting carbinol stereocenter is predetermined, thereby cir-
cumventing formation and separation of diastereomers. The
substitution pattern Vis-a`-Vis the silyl ether and free hydroxyl
can be readily achieved simply by the order of epoxide addition.
Thus, access to all possible diastereomers of the 1,3,5-polyol
structural motif is possible by alternating the order of addition
and stereochemistry of the epoxides, followed by stereocon-
trolled reduction of the ketone derived by unmasking the
dithiane. A wide variety of other electrophiles should also prove
viable in the second addition (Vide infra) based upon related
studies, both in our laboratory4,6 and in others.14 Importantly,
the highly functionalized 1,3,5-polyol adducts are poised for
further elaboration.
Vis-a`-Vis our interest in controlling the timing of silyl migration.
With 13, the 1,3-Brook rearrangement did not occur at -78 °C
following initial alkylation in THF but did proceed readily upon
addition of HMPA. Complete consumption of the initial anion
could thus be achieved prior to silyl migration, permitting
alkylation with a second electrophile. Similar effects on the
corresponding 1,4-Brook rearrangement were observed when
13 was reacted with an epoxide. In Et2O and THF, silyl group
migration was completely suppressed at -78 °C, although, in
THF, warming to -40 °C led to rearrangement. Again, HMPA
triggered silyl transfer; addition of a second electrophile to anion
17 derived via Brook rearrangement then furnished the unsym-
metrical adduct 18.12
With these observations as a guide, metalation and alkylation
of the lithium anion derived from 8 with epoxide (-)-9 in Et2O
furnished primarily the unrearranged carbinol (+)-22, whereas
alkylation in THF resulted in a mixture of compounds (Table
1, entries 1 and 2). Thus, Et2O rather than THF appeared to be
the solvent of choice to achieve the initial alkylation and, most
importantly, to suppress premature silyl migration. Addition of
HMPA or DMPU13 after an appropriate reaction time (ca. 1 h;
entries 3 and 4) then triggered the 1,4-Brook rearrangement to
afford predominantly silyl ether (+)-23. Rigorous degassing of
(12) Matsuda and co-workers previously reported an analogous unsymmetrical
bisalkylation of trimethylsilylacetonitrile (19) with two different epoxides
in DME (see scheme below). In this example, how the timing of the Brook
rearrangement was controlled is unclear. One could speculate that the first
alkylation occurs more rapidly than silyl migration. Alternatively, lithiated
19 could be more reactive than the intermediate formed after Brook
rearrangement (21) reducing competitive alkylation. See: Matsuda, I.;
Murata, S.; Ishii, Y. J. Chem. Soc., Perkin Trans. 1 1979, 26.
Epichlorohydrin as the Second Electrophile. The use of
epichlorohydrin as the second electrophile presents the interest-
ing issue of dithiane anion attack at either the carbon bearing
the chloride or the terminal epoxide carbon, followed by
alkoxide displacement of the chloride to generate a new terminal
epoxide. In either case, the coupling process, if successful, would
lead to a new epoxide permitting further elaboration. The
stereochemistry of the resultant epoxide however would differ.
(14) (a) Le Merrer, Y.; Gravier-Pelletier, C.; Maton, W.; Numa, M.; Depezay,
J.-C. Synlett 1999, 1322. (b) Brauer, N.; Dreeben, S.; Schaumann, E.
Tetrahedron Lett. 1999, 40, 2921. (c) Hale, K. J.; Hummersome, H. G.;
Bhatia, G. S. Org. Lett. 2000, 2, 2189. (d) Gravier-Pelletier, C.; Maton,
W.; Le Merrer, Y. Tetrahedron Lett. 2002, 43, 8285. (e) Petri, A. F.;
Kuhnert, S. M.; Scheufler, F.; Maier, M. E. Synthesis 2003, 6, 940.
(13) DMPU is an abbreviation for N,N′-dimethyl- N,N′-propylene urea. For
substituting DMPU in place of HMPA, see: Mukhopadhyay, H.; Seebach,
D. HelV. Chim. Acta 1982, 65, 385.
9
J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003 14437

A R T I C L E S
Smith et al.
Table 2. Three-Component Couplings with 2-TBS-1,3-dithiane in One Flask
+
a Only 1 equiv of E2 was used.
Literature precedent suggests that carbon nucleophiles attack
Scheme 6
preferentially at C(1) of the epoxide versus C(3) bearing the
halide.15 In the event, execution of the three-component protocol
with the two enantiomers of epichlorohydrin (-)-30 and (+)-
30 furnished (+)-31 and (+)-32, respectively, as the sole product
(Table 2, entries 4 and 5). The stereochemistry of epoxides (+)-
31 and (+)-32 were readily established by the conversion of
mixed ketal (+)-27, of known stereochemistry, to epoxide (+)-
31, which proved identical in all respects to (+)-31 derived from
(-)-30 (Scheme 6). Thus attack occurred selectively at the ter-
minal epoxide carbon followed by displacement of the chloride.
(15) (a) Kasai, N.; Suzuki, T.; Furukawa, Y. J. Mol. Cat. B. 1998, 4, 237. (b)
Hanson, R. M. Chem. ReV. 1991, 91, 437.
9
14438 J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003

Multicomponent Linchpin Couplings
A R T I C L E S
Scheme 8
Intra- versus Intermolecular Silyl Migration. To establish
the mode of silyl migration in the multicomponent linchpin tactic
(i.e., intra- or intermolecularity), we carried out a crossover
experiment, wherein an equimolar mixture of the lithium anions
derived from dithianes (+)-22 and (+)-41 was generated in the
presence of HMPA at -78 °C, followed by warming the mixture
to room temperature (Scheme 7). No crossover products were
observed. Thus, as expected silyl migration occurs intramo-
lecularly.
Scheme 7
metalation conditions led only to decomposition. The successful
generation and reaction of the corresponding lithium anion of
(+)-44, however, could be achieved via the lithium alkoxide.
That is, addition of epoxide (-)-45 in a mixture of HMPA/
THF to the anion of (+)-44 furnished (+)-46 in 80% yield by
way of a Brook rearrangement. In another case, all attempts to
effect lithiation of (+)-47, as demonstrated by deuterium
incorporation, did not lead to the desired dithiane (+)-51
(Scheme 9). Instead, quantitative deuterium incorporation oc-
Reversible Nature of Silyl Migration. To determine the
potential reversibility of the silyl migration process, dithiane
(+)-23 was deprotonated with t-BuLi in THF at -78 °C
followed by warming to 0 °C; a mixture of (+)-22 and (+)-23
(1.0:1.3) was observed after 3 h. Repeating the same reaction
in Et2O also yielded (+)-22 and (+)-23 (1.0:1.5). Alternatively,
deprotonation of alcohol (+)-22 in Et2O, employing the identical
conditions, provided (+)-22 and (+)-23 (1.4:1.0). Taken
together, these results demonstrate that the Brook rearrangement
is indeed reversible under these conditions; however, the
equilibrium state could not be achieved due to the instability
of the dithiane anions over the time and/or temperature required
to achieve complete equilibration.
Scheme 9
Lithiation of 2-Substituted Dithianes: A More Effective
Tactic. In contrast to the parent 1,3-dithiane, which is easily
metalated with n-BuLi, a myriad of strong bases, solvent
additives, and time/temperature regimens have been employed
to generate the lithium anion of 2-substituted derivatives,
sometimes to no avail.16 Indeed, dithianes possessing a high
level of oxygenation and unsaturation are often difficult to
deprotonate and, in some cases, result in capricious behavior.17
We have not been spared from such difficulties. In our second
generation synthesis of the CD-spiroketal fragment of the potent
antitumor agents, the spongistatins,18 attempts to generate the
lithium anion of (+)-43 (Scheme 8) employing a variety of
curred at the benzylic position of the naphthyl protecting group
(48).19 The requisite dithiane anion could however be generated,
as illustrated by formation of the 2-d-dithiane (+)-51, by the
reaction of epoxide (-)-50 with the lithium anion obtained from
2-TES-1,3-dithiane (49), followed by a Brook rearrangement
triggered by HMPA.
(16) (a) Lipshutz, B. H.; Garcia, E. Tetrahedron Lett. 1990, 31, 7261. (b) Ide,
M.; Nakata, M. Bull. Chem. Soc. Jpn. 1999, 72, 2491 and references therein.
(17) (a) Hanessian, S.; Pougny, J.-R.; Boessenkool, I. K. Tetrahedron 1984,
40, 1289. (b) Oppong, I.; Pauls, H. W.; Liang, D.; Fraser-Reed, B. J. Chem.
Soc., Chem. Commun. 1986, 1241. (c) Kinoshita, M.; Taniguchi, M.;
Morioka, M.; Takami, H.; Mizusawa, Y. Bull. Chem. Soc. Jpn. 1986, 61,
2147. (d) Jones, T. K.; Reamer, R. A.; Desmond, R.; Mills, S. G. J. Am.
Chem. Soc. 1990, 112, 2998. (e) Okamura, H.; Kuroda, S.; Ikegami, S.;
Tomita, K.; Sugimoto, Y.; Sakaguchi, S.; Ito, Y.; Katsuki, T.; Yamaguchi,
M. Tetrahedron 1993, 49, 10531. (f) De Brabander, J.; Vandewalle, M.
Synthesis 1994, 855. (g) Nicolaou, K. C.; Patron, A. P.; Ajito, K.; Richter,
P. K.; Khatuya, H.; Bertinato, P.; Miller, R. A.; Tomaszewski, M. J. Chem.
Eur. J. 1996, 2, 847. (h) Smith, A. B., III; Friestad, G. K.; Barbosa, J.;
Bertounesque, E.; Hull, K. G.; Iwashima, M.; Qiu, Y.; Salvatore, B. A.;
Spoors, P. G.; Duan, J. J. W. J. Am. Chem. Soc. 1999, 121, 10468. (i)
Armstrong, A.; Barsanti, P. A.; Jones, L. H.; Ahmed, G. J. Org. Chem.
2000, 65, 7020 and references therein.
To understand the steric and stereoelectronic issues associated
with the metalation of 2-substituted-1,3-dithianes, we turned to
(18) Smith, A. B., III; Doughty, V. A.; Sfouggatakis, C.; Bennett, C. S.;
Koyanagi, J.; Takeuchi, M. Org. Lett. 2002, 4, 783.
(19) This observation was not surprising since the pKa’s of both the dithiane
and benzylic protons are quite similar. See: Bordwell, F. G. Acc. Chem.
Res. 1988, 21, 456.
9
J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003 14439

A R T I C L E S
Smith et al.
the elegant work of Eliel and co-workers.20 Their pioneering
studies established (a) that deprotonation of conformationally
constrained 4,6-disubstituted-1,3-dithianes, such as 52 and 54,
occurs with high stereoselectivity at the equatorial position, (b)
that the equatorial anion is thermodynamically more stable, and
(c) that, upon exposure to an electrophile, the new group is
introduced in the equatorial position (Scheme 10). In addition,
electrophile, higher-order multicomponent couplings might be
possible. Such a tactic would provide exceptionally concise
routes to advanced subtargets required for the 1,3-polyol
backbone segments of the polyene macrolides.22 The first
example, leading to the formation of four carbon-carbon
σ-bonds in one-flask, involved alkylation of the anion derived
from dithiane 8 with epoxide (-)-9 to generate ca. 2 equiv of
the lithium alkoxide 60 (Scheme 11). Sequential addition of
Scheme 10
Scheme 11
abstraction of an axial methine hydrogen (e.g., 54) was found
to be quite sluggish, requiring long reaction times and higher
temperatures, compared to the equatorial isomer. Thus, the
difficulties experienced with direct metalation of structurally
complex dithianes can, in some cases, be attributed to the
preference for removal of equatorial hydrogens. This analysis,
however, does not completely explain why some 2-substituted
dithianes are readily deprotonated and others are not. From a
pragmatic point of view, however, it is often possible to
circumvent such difficulties by taking advantage of the solvent
controlled Brook rearrangement as illustrated in Schemes 8 and
9. These observations suggest that migration of the 2-silyl group
may not suffer the same conformational limitations as direct
metalation. Support here comes from the recent study by
Degl’Innocenti and Pollicino,21 demonstrating that both axial
and equatorial silyl groups on conformationally constrained 1,3-
dithianes couple to benzaldehyde with essentially the same rate
upon treatment with fluoride ion.
HMPA and epichlorohydrin (-)-30 (1 equiv) furnished the bis-
(silyloxy dithiane) carbinol (+)-61 in 66% yield, accompanied
by a minor amount (ca. 2%) of epoxide (+)-31.4
A second example6 of a five-component coupling tactic was
designed to permit ready access to the pseudo-C2-symmetric
trisacetonide (+)-64 employed by Schreiber et al. in a synthesis
of (+)-mycoticin A.23 Lithiation of dithiane 8 (2.5 equiv)
followed in turn by alkylation with (-)-9 (Scheme 12), addition
Scheme 12
Application of the Multicomponent Tactic in Natural
Product Synthesis. We, along with other research groups,14
have demonstrated the considerable utility of the one-step,
multicomponent coupling protocol as strategic reactions in
several ongoing synthetic programs.3b,c,18 Importantly, these
three-component coupling reactions can be conducted routinely
on a multigram scale.18
Higher-Order Multicomponent Couplings: Union of Five
Components. We reasoned that, by careful choice of the second
(20) (a) Hartmann, A. A.; Eliel, E. L. J. Am. Chem. Soc. 1971, 93, 2572. (b)
Eliel, E. L.; Abatjoglou, A.; Hartmann, A. A. J. Am. Chem. Soc. 1972, 94,
4786. (c) Eliel, E. L.; Hartmann, A. A.; Abatjoglou, A. J. Am. Chem. Soc.
1974, 96, 1807. (d) Eliel, E. L. Tetrahedron 1974, 30, 1503. (e) Abatjoglou,
A.; Eliel, E. L.; Kuyper, L. F. J. Am. Chem. Soc. 1977, 99, 8262.
(21) The following scheme illustrates that the conformationally constrained
2-silyl-1,3-dithianes 56 and 58 furnish 57 and 59 upon treatment
with benzaldehyde and a fluoride source. See: Capperucci, A.; Cere,
V.; Degl’Innocenti, A.; Nocentini, T.; Pollicino, S. Synlett 2002, 9,
1447.
of HMPA to trigger the 1,4-Brook rearrangement, and diep-
oxypentane (+)-6224 (1.0 eq) furnished diol (+)-63 in 59% yield,
9
14440 J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003

Multicomponent Linchpin Couplings
A R T I C L E S
Table 3. Reactions of Dithiane Lithium Anions with Vinyl Epoxides
accompanied again by a minor amount (ca. 13%) of the epoxide,
arising from an incomplete second alkylation. Notably, use of
THF as the cosolvent with HMPA in the second alkylation
resulted in higher yields as compared to Et2O (ca. 33%).
Alternatively, when DMPU was employed to trigger the Brook
rearrangement, diol (+)-63 was obtained in a somewhat lower
yield (55%) compared with HMPA. Enhanced efficiency with
HMPA versus DMPU is a general observation in our laboratory.
Due caution with HMPA must, of course, be observed. In both
examples, four new carbon-carbon σ-bonds were generated
efficiently in one flask.
Vinyl Epoxides as Electrophiles: The SN2 versus the SN2′
Manifold. As with epichlorohydrin, the use of vinyl epoxides
as electrophiles holds the interesting promise of reactivity at
two electrophilic sites. Surprisingly, relatively few reports exist
on the addition of carbon nucleophiles, particularly dithiane
anions, to vinyl epoxides (e.g., SN2 versus SN2′),25 although
nucleophilic additions to alkenes possessing a leaving group at
the allylic position have been examined extensively over the
past 40 years.26 Controlled access to the SN2 and SN2′ reaction
manifolds would significantly augment the utility of dithiane
linchpins and would permit expansion of the multicomponent
tactic (Scheme 13).
entry
R
epoxide yield(%)
SN2:SN2′a
product
76
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
65 (R ) H)
71
72
73
74
75
71
72
73
74
75
71
72
73
74
75
71
72
73
74
75
71
72
73
74
75
71
72
73
74
75
83
74
85
89
69
85
82
88
86
74
83
85
88
76
78
81
78
84
82
73
81
84
74
63
74
85
86
81
84
80
100 (anti):0
100 (anti):0
100 (anti):0
100 (syn):0
100 (syn):0
100 (anti):0
100 (anti):0
100 (anti):0
100 (syn):0
100 (syn):0
100 (anti):0
100 (anti):0
100 (anti):0
100 (syn):0
100 (syn):0
0:100 91
0:100 92
1 (anti):3.5
1 (syn):3
1 (syn):5
0:100 96
0:100 97
1 (anti):5
1 (syn):4
1 (syn):3
65
65
65
65
77
78
79
80
81
82
83
84
85
86
87
88
89
90
66 (R d Ph)
66
66
66
66
Scheme 13
67 (R ) TMS)b
67
67
67
67
68 (R ) Et)
68
68
68
68
93a/b
94a/b
95a/93b
From the outset of this strategy, we speculated that selectivity
over the SN2 and SN2′ reaction manifolds might be achievable
by varying the steric properties of the lithiated 2-substituted-
1,3-dithianes. That is, a small nucleophile would be expected
to attack preferentially at the allylic position via an SN2 process,
whereas a more sterically demanding nucleophile would add
to the more accessible alkene terminus in an SN2′ fashion.
Toward this end, a series of dithianes (65-70) were metalated
(t-BuLi in 10%HMPA/THF) and reacted with a series of vinyl
epoxides (71-75)27 in THF (Table 3). Lithium anions derived
from dithianes with small substituents (R ) H, Ph, and TMS)
afforded solely the SN2 adduct in good yield (entries 1-15),
suggesting that the allylic position of the vinyl epoxide is indeed
69 (R ) i-Pr)
69
69
69
69
98a/b
99a/b
100a/98b
70 (R ) TIPS)
0:100 101
70
70
70
70
0:100 102
0:100 103
0:100 104
0:100 103
a Ratio of isolated compounds. b Products from entries 11-15 underwent
1,4-Brook rearrangement to a 1:1 mixture of non-Brook:Brook products.
activated.28 Not surprisingly, in the presence of HMPA, reactions
with 2-TMS-1,3-dithiane (67) led to a mixture of homoallylic
alcohols and the corresponding TMS-ethers resulting from 1,4-
Brook rearrangement. The stereochemical outcome of the SN2
reaction with cis vinyl epoxides 71, 72 and 73 provided, as
anticipated, exclusively the anti vicinal adducts, whereas the
trans epoxides 74 and 75 afforded only the syn vicinal adducts.29
The yields were good in both series.
A complete reversal of selectivity to the SN2′ manifold was
achieved with the lithium anions derived from alkyl-substituted
dithianes (R ) Et and i-Pr, entries 16-25); again yields were
good. The chemoselectivity was however compromised with
(22) For reviews on polyol syntheses, see: (a) Schneider, C. Angew. Chem.,
Int. Ed. 1998, 37, 1375. (b) Oishi, T.; Nakata, T. Synthesis 1990, 635. (c)
Rychnovsky, S. D. Chem. ReV. 1995, 95, 2021.
(23) Poss, C. S.; Rychnovsky, S. D.; Schreiber, S. L. J. Am. Chem. Soc. 1993,
115, 3360.
(24) Rychnovsky, S. D.; Griesgraber, G.; Zeller, S.; Skalitzky, D. J. J. Org.
Chem. 1991, 56, 5161.
(25) For examples, see: (a) Semeniuk, F.; Jenkins, G. L. J. Am. Pharm. Assoc.,
Sci. Ed. 1948, 37, 118. (b) Herr, R. W.; Johnson, C. R. J. Am. Chem. Soc.
1970, 92, 4979. For Lewis acid promoted SN2 additions of alkyllithiums,
see: (c) Alexakis, A.; Vrancken, E.; Mangeney, P.; Chemla, F. J. Chem.
Soc., Perkin Trans. 1 2000, 3352. For transition metal catalyzed reactions
with vinyl epoxides, see: (d) Fagnou, K.; Lautens, M. Org. Lett. 2000, 2,
2319. (e) Trost, B. M.; Bunt, R. C.; Lemoine, R. C.; Calkins, T. L. J. Am.
Chem. Soc. 2000, 122, 5968 and references therein. For a review on SN2′
additions of organocuprates to vinyl epoxides, see: Marshall, J. A. Chem.
ReV. 1989, 89, 1503.
(26) (a) Stork, G.; White, W. N. J. Am. Chem. Soc. 1953, 75, 4119. (b) Stork,
G.; White, W. N. J. Am. Chem. Soc. 1956, 78, 4609. (c) Magid, R. M. J.
Org. Chem. 1971, 36, 2105. (d) Stork, G.; Kreft, A. F., III. J. Am. Chem.
Soc. 1977, 75, 3850. (e) Magid, R. M. Tetrahedron 1980, 36, 1901. (f)
Bordwell, F. G.; Clemens, A. H.; Cheng, J.-P. J. Am. Chem. Soc. 1987,
109, 1773. (g) Paquette, L. A.; Stirling, C. J. M. Tetrahedron 1992, 48,
7383.
(27) For syntheses of vinyl epoxides 71-73, see: (a) Hu, S.; Jayaraman, S.;
Oehleschlager, A. C. J. Org. Chem. 1996, 61, 7513. For synthesis of vinyl
epoxides 74-75, see: (b) Diez-Martin, D.; Kotecha, N. R.; Ley, S. V.;
Mantegani, S.; Menedez, J. C.; Osbourn, H. M.; White, A. D.; Banks, J.
B. Tetrahedron 1992, 48, 7899.
(28) Nicolaou, K. C.; Prasad, C. V. C.; Somers, P. K.; Hwang, C.-K. J. Am.
Chem. Soc. 1989, 111, 5330.
9
J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003 14441

A R T I C L E S
Smith et al.
epoxides 73, 74, and 75. Addition of the more sterically
demanding 2-triisopropylsilyl-1,3-dithiane (70) (entries 26-30)
also proceeded exclusively via SN2′ attack in good yield with
all vinyl epoxides. Thus, at the two extremes of dithiane anion
steric encumberance (R ) H or R ) TIPS), access to either the
SN2 or the SN2′ manifold can be achieved with excellent
selectivity.30
yield (Scheme 14); the corresponding diene elimination product
109 was also observed (25% and 19%, respectively). Impor-
tantly, of the two reaction manifolds, only the SN2′ product was
observed in each case.
Scheme 14
Effect of Lewis Acids on the SN2 and SN2′ Reaction
Manifolds. In an attempt to provide additional control over the
SN2 and SN2′ reaction manifolds, we explored the effect of
Lewis acids.25c We selected the reactions of 2-Et- and 2-i-Pr-
1,3-dithianes (68 and 69) with vinyl epoxide 71 (Table 3, entries
16 and 21); the reaction proceeds exclusively via the SN2′
manifold. No reaction was observed in the presence of 1.0 equiv
of either zinc chloride or magnesium bromide (Table 4, entries
Table 4. Effect of Additives on the Addition of Dithiane Anions to
Vinyl Epoxides
Attempts to determine the relative stereochemistry of 108 and
111 via X-ray crystallography proved unsuccessful; a chemical
correlation with compounds of known stereochemistry was
therefore established, demonstrating that this SN2′ addition of
lithiated dithiane anions occurs via a syn process.34
Multicomponent Tactic Employing Vinyl Epoxides as
Electrophiles. Recognizing the potential of the SN2 versus SN2′
addition manifolds of vinyl epoxides, we explored the possibility
of higher-order multicomponent couplings, first utilizing vinyl
epoxide (-)-75 as the second electrophile (Scheme 15). Toward
entry
R
additive
ZnCl2
yield
SN2:SN2′
1
2
3
4
Et
i-Pr
Et
i-Pr
Et
i-Pr
Et
-
-
-
ZnCl2
-
MgBr2
MgBr2
LiCl
LiCl
BF3‚OEt2
BF3‚OEt2
-
-
-
-
5
65
63
35
40
4 (105):1 (91)
3 (106):1 (96)
3 (105):1 (91)
2 (106):1 (96)
6
7a
8a
i-Pr
Scheme 15
a Entries 7 and 8 were conducted in Et2O.
1-4). However, dithiane addition did occur in the presence of
1.0 equiv of lithium chloride or boron trifluoride etherate (entries
5-8), with reversal of selectivity to favor the SN2 product. With
boron trifluoride etherate, addition of the dithiane anions to THF
was solely observed; thus Et2O was the solvent of choice in
these examples.31 Yields, while not optimized, were modest to
good.
Stereochemistry of the SN2′ Addition Process. To define
the stereochemical outcome of the SN2′ addition process, we
explored the effect of substitution at the terminal position of
the alkene in the vinyl epoxide.32 Toward this end, vinyl
epoxides 10733 and 11033 were subjected to the lithium anion
of 2-triisopropylsilyl-1,3-dithiane (70). Adducts 108 and 111,
resulting from SN2′ addition, were isolated, albeit in modest
(29) The relative configurations of the SN2 adducts derived from the cis and
trans vinyl epoxides were determined by chemical correlation (see
Supporting Information).
(30) We recently employed this tactic toward the synthesis of rimocidin: Smith,
A. B.; Pitram, S. M.; Fuertes, M. J. Org. Lett. 2003, 5, 2751.
(31) Fang, J.-M.; Chen, M.-Y. Tetrahedron Lett. 1988, 29, 5939.
(32) For references on the mechanism of the SN2′ reaction, see: (a) Magid, R.
M.; Fruchey, O. S. J. Am. Chem. Soc. 1970, 92, 2107. (b) Denmark, S. E.
PhD. Thesis, Swiss Federal Institute of Technology, Zurich, 1982. (c)
Stohrer, W.-D. Angew. Chem., Int. Ed. Engl. 1983, 22, 613. (d) Park, Y.
S.; Kim, C. K.; Lee, B.-S.; Lee, I. J. Phys. Chem. 1995, 99, 13103.
this end, deprotonation of dithiane 8, followed in turn by
addition of (+)-9 and vinyl epoxide (-)-75 in Et2O/HMPA,
9
14442 J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003

Multicomponent Linchpin Couplings
A R T I C L E S
furnished a mixture (2:1) of 112 and 113 in a modest yield
(50%), favoring the expected SN2′ pathway in the second
alkylation.
That a change in the solvent system (THF/HMPA to Et2O)
affects the observed selectivity between the SN2 and SN2′
manifolds did not come as a complete surprise. Studies by Reich
and co-workers35 demonstrated that in THF/HMPA dithiane
anions are monomeric entities; therefore selectivity between the
SN2 and SN2′ manifolds can be attributed to the steric nature of
the 2-substituent of the dithiane. However, in Et2O at low
temperature (-78 to -25 °C) the dithiane anions form ag-
gregates.36 In this case, we would expect the aggregate to favor
SN2′ addition. Indeed, in THF/HMPA, the lithium anion derived
from 2-TMS-1,3-dithiane (67) afforded exclusively the SN2
adduct with epoxides 71 and 73 (Table 3, entries 11 and 13),
whereas in Et2O the SN2′ adduct was the sole product observed
with both epoxides (Table 5, entries 2 and 4). Similarly, 2-Et-
1,3-dithiane (68) resulted in a 3.5:1 mixture of SN2′:SN2 adducts
in THF/HMPA (Table 3, entry 18), versus exclusive SN2′
addition in Et2O, albeit in modest yield (Table 5, entry 3).
One-Flask, Three- and Four-Component Coupling Reac-
tions with a Halomethyl Vinyl Epoxide Hybrid. Having
demonstrated the utility of both epichlorohydrin and vinyl
epoxides in multicomponent couplings, we turned next to
halomethyl vinyl epoxides (117 and 118; Scheme 16), hybrids
of the two electrophiles readily available from the corresponding
alcohol. Addition of lithiated 2-Ph-1,3-dithiane (66, 1.2 equiv)
to the chloro congener (+)-11737 furnished a mixture of (+)-
119 and (+)-120 (Scheme 16), the outcome, respectively, of
We next examined vinyl epoxides as the first electrophile.
In this case, if the addition of a 2-silyl-1,3-dithiane anion were
to occur via the SN2 manifold, 1,4-Brook rearrangement would
lead to a new dithiane anion available for a second alkylation.
Alternatively, if the SN2′ manifold were to pertain, an analogous
1,6-silyl migration across the E-alkene would not be possible.
However, since our initial studies on the reactions of dithiane
anions with vinyl epoxides had been performed in THF with
HMPA, control of the Brook rearrangement under these
conditions would not be feasible. We therefore examined the
addition of lithiated 2-silyl-1,3-dithianes to vinyl epoxides in
Et2O, in the absence of HMPA (Table 5). The addition occurred
smoothly to furnish exclusively the SN2′ congeners. Thus, vinyl
epoxides are not viable first electrophiles in a three-component
process. Nonetheless, we were pleased with the high selectivity
observed.
Table 5. Addition of Dithiane Anions to Vinyl Epoxides in the
Absence of HMPA
Scheme 16
direct chloride displacement and epoxide opening at the allylic
position, followed in the latter case by ring closure to furnish a
(33) Vinyl epoxide (()-107 was prepared from commercially available trans,
trans-2,4-hexadien-1-ol as shown in the following scheme:
(35) (a) Reich, H. J.; Borst, J. P.; Dykstra, R. R. Tetrahedron 1994, 50, 5869.
(b) Sikorski, W. H.; Reich, H. J. J. Am. Chem. Soc. 2001, 123, 6527. (c)
Reich, H. J.; Sanders, A. W.; Fiedler, A. T.; Bevan, M. J. J. Am. Chem.
Soc. 2002, 124, 13386.
(36) Visual evidence for aggregation was observed where upon treatment of
2-Et-1,3-dithiane with t-BuLi in Et2O (ca. 0.2 M) at -78 °C, a cloudy
suspension formed and gradually dissipated upon warming.
(37) Vinyl epoxides (+)-117 and (+)-118 were prepared from the preceding
alcohol as shown in the following scheme:
Vinyl epoxide (()-110 was prepared from propargyl alcohol and Z-1-
bromopropene as shown in the following scheme:
For preparation of epoxy alcohol, see: Romero, A.; Wong, C. H. J. Org.
Chem. 2000, 65, 8264.
(34) See Supporting Information.
9
J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003 14443

A R T I C L E S
Smith et al.
new epoxide. Interestingly, the bromo congener (+)-11837
furnished only (+)-119. These results are in complete agreement
with the reactions observed with epichloro- and epibromohydrin,
Table 6. Multicomponent Addition of Dithiane Anions to
Bromomethyl Vinyl Epoxide (+)-118
although chloride displacement is only rarely observed.15
A
similar lack of selectivity with the chloro congener (117) was
observed with other dithiane anions. We therefore employed
the bromo congener (+)-118 in all future studies to limit side
products arising via attack at the epoxide.
A three-component coupling employing vinyl epoxide (+)-
118 as the initial electrophile appeared quite feasible given our
earlier results with epichlorohydrin. Toward this end, addition
of the anion derived from 2-Ph-1,3-dithiane (66) to (+)-118,
followed by addition of a second sterically encumbered dithiane
anion (70) furnished adduct (-)-121 in 61% yield (Scheme 17).
Scheme 17
Importantly, the second addition occurred exclusively via the
SN2′ manifold, a result anticipated from our earlier studies. In
a similar fashion, reversing the order of additions resulted in
the adduct (+)-122. Again, as expected we observed exclusive
SN2 addition of the second dithiane anion. Additional examples
of this class of three-component couplings, providing advanced
intermediates suitable for further elaboration are illustrated in
Table 6.
advantage of solvent effects. Since THF/HMPA would not
permit control of the Brook rearrangement, we employed Et2O
in the absence of HMPA. Under these conditions, treatment of
(+)-118 with 2 equiv of the lithium anion of 8 delivered the
bis-adduct (+)-128 in 72% yield (Scheme 18). We could also
induce a Brook rearrangement in this reaction after addition of
(+)-118 by introducing HMPA. Alkylation with epoxide
To expand the scope of the latter synthetic tactic with the
intention to incorporate a Brook rearrangement, we again took
Scheme 18
9
14444 J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003

Multicomponent Linchpin Couplings
A R T I C L E S
(+)-9, as the second electrophile, furnished the four-component
adduct (-)-129 in 52% yield (Scheme 18). Importantly, no SN2
adducts were observed, again highlighting the preference for
the SN2′ manifold with sterically encumbered anions. An
expanded version of this example leading to the union of four
different building blocks employs vinyl epoxide (+)-118 as the
second electrophile, followed by addition of the lithium anion
derived from 2-TIPS-1,3-dithiane (70); the resultant adduct (-)-
130 was produced in 50% yield (Scheme 18).
anions afford SN2 adducts, whereas encumbered anions lead
primarily to SN2′ adducts. Mechanistic studies demonstrate that
the SN2′ process occurs via syn addition to the vinyl epoxide.
Finally, integration of the multicomponent tactic with epichlo-
rohydrin and vinyl epoxides permits the ready higher-order
union of four and five components in a single flask.
Acknowledgment. This paper is dedicated to Professors
Corey, Seebach and Eliel, pioneers in the area of dithiane
chemistry. Financial support was provided by the National
Institutes of Health (National Cancer Institute) through Grant
CA-70329.
Summary
The development of a one-flask multicomponent dithiane
linchpin coupling protocol exploiting a solvent controlled Brook
rearrangement has been achieved. Competent electrophiles
include terminal epoxides, epichlorohydrin, and vinyl epoxides.
High chemoselectivity can be achieved with epichlorohydrin
and vinyl epoxides. For vinyl epoxides, the steric nature of the
dithiane anion is critical; sterically unencumbered dithiane
Supporting Information Available: Spectroscopic and ana-
lytical data and selected experimental procedures. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA0376238
9
J. AM. CHEM. SOC. VOL. 125, NO. 47, 2003 14445