Sterically biased 3,3-sigmatropic rearrangement of chiral allylic azides: Application to the total syntheses of alkaloids

Article abstract of DOI:10.1021/jo800817p

(Chemical Equation Presented) We describe a tandem Mitsunobu/3,3- sigmatropic rearrangement of allylic azides on a chiral auxiliary system that favors one regioisomer thanks to its exceptional steric bias. The sequence may be completed by the oxidative cleavage of the auxiliary or by a ring-closing metathesis reaction that produces a carbo-or heterocycle directly and a recyclable form of the chiral auxiliary. Applications of the methodology to the total synthesis of (+)-coniine, (+)-lentiginosin, and (+)-pumiliotoxin C are reported.

Full text of DOI:10.1021/jo800817p

Sterically Biased 3,3-Sigmatropic Rearrangement of Chiral Allylic
Azides: Application to the Total Syntheses of Alkaloids
Sophie Lauzon, Franc¸ois Tremblay, David Gagnon, Ce´drickx Godbout, Christine Chabot,
Catherine Mercier-Shanks, Ste´phane Perreault, He´le`ne DeSe`ve, and Claude Spino*
UniVersite´ de Sherbrooke, De´partement de Chimie, 2500 BouleVard UniVersite´, Sherbrooke,
Quebec, Canada, J1K 2R1
Claude.Spino@USherbrooke.ca
ReceiVed April 15, 2008
We describe a tandem Mitsunobu/3,3-sigmatropic rearrangement of allylic azides on a chiral auxiliary
system that favors one regioisomer thanks to its exceptional steric bias. The sequence may be completed
by the oxidative cleavage of the auxiliary or by a ring-closing metathesis reaction that produces a carbo-
or heterocycle directly and a recyclable form of the chiral auxiliary. Applications of the methodology to
the total synthesis of (+)-coniine, (+)-lentiginosin, and (+)-pumiliotoxin C are reported.
Introduction
Indeed, most allylic azides exist as mixtures of regioisomers
that interconvert rapidly at ambiant temperature, a drawback
that has hampered their use in synthesis. VanderWerf and
Heasley have noted that, in general, tertiary and secondary allylic
azides rearrange more rapidly than primary allylic azides.4 In
addition, the regioisomer with the more substituted alkene was
usually thermodynamically favored. A high degree of regiose-
lectivity could be obtained in cases where the double bond was
conjugated with an insaturation5 while some degree of regio-
chemical control was achieved using competitive reactivity of
either the azide or alkene moiety.6 A chiral scaffold capable of
biasing the thermodynamic ratio of regioisomeric allylic azides
and lead to useful structural motifs is highly desirable.
The 3,3-sigmatropic rearrangement of azides were first ob-
served in 1960 by Winstein and co-workers.1 They prepared a
70-30 mixture of 2-azido-2-methyl-3-butene and 1-azido-3-
methyl-2-butene that were interconverting fast at room-temper-
ature but were stable for days at -80 °C. Others reported
obtaining thermodynamic mixtures of acyclic allylic azides at
or lower than room temperature.2 Cyclic allylic azides, it seems,
require higher temperatures to rearrange.3
(1) (a) Gagneux, A.; Winstein, S.; Young, W. G. J. Am. Chem. Soc. 1960,
82, 5956–5957. See also: (b) VanderWerf, C. A.; Heilser, R. Y.; McEwen, W. E.
J. Am. Chem. Soc. 1954, 76, 1231–1235.
(2) (a) Murahashi, S.-I.; Tanigushi, Y.; Imada, Y.; Tanigawa, Y. J. Org.
Chem. 1989, 54, 3292–3303. (b) Safi, M.; Fahrang, R.; Sinou, D. Tetrahedron
Lett. 1990, 31, 527–530. (c) Arimoto, M.; Yamaguchi, H.; Fujita, E. Tetrahedron
Lett. 1987, 28, 6289–6292.
We recently reported a novel approach to construct chiral
nonracemic carbo- and heterocycles using a tandem Mitsunobu
reaction/3,3-sigmatropic rearrangement of allylic azide.7 We
(3) (a) Trost, B. M.; Pulley, S. R. Tetrahedron Lett. 1995, 36, 8737–8740.
(b) Fava, C.; Galeazzi, R.; Mobbili, G.; Orena, M. Tetrahedron: Asymm. 2001,
12, 2731–2741. (c) Capaccio, C. A. I.; Varela, O. Tetrahedron: Asymmetry 2000,
11, 4945–4954. (d) Cardillo, G.; Fabbroni, S.; Gentillucci, L.; Perciaccante, R.;
Piccinelli, F.; Tolornelli, A. Org. Lett. 2005, 7, 533–536. (e) Chida, N.; Tobe,
T.; Murai, K.; Yamazaki, K.; Ogawa, S. Heterocycles 1994, 38, 2383–2388. (f)
Magg, H.; Rydzewski, R. M. J. Org. Chem. 1992, 57, 5823–5831. (g) Askin,
D.; Angst, C.; Danishefsky, S. J. Org. Chem. 1985, 50, 5005–5007.
(4) VanderWerf, C. A.; Heasley, V. L. J. Org. Chem. 1966, 31, 3534–3537.
(5) Panek, J. S.; Yang, M.; Muler, I. J. Org. Chem. 1992, 57, 4063–4064.
(6) Feldman, A. K.; Colasson, B.; Sharpless, K. B.; Fokin, V. V. J. Am.
Chem. Soc. 2005, 127, 13444–13445.
(7) Gagnon, D.; Lauzon, S.; Godbout, C.; Spino, C. Org. Lett. 2005, 7, 4769–
4771.
10.1021/jo800817p CCC: $40.75
Published on Web 07/22/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 6239–6250 6239

Lauzon et al.
TABLE 1. AlMe3-Promoted Vinyllithium Addition of Vinyllithium
to 1
SCHEME 1. General Strategy for the Synthesis of Amino
Acids, and Carbo- or Heterocycles
herein report a more complete investigation of this methodology
as well as applications to the synthesis of natural alkaloids.
The general synthetic strategy is depicted in Scheme 1. The
auxiliary p-menthane-3-carboxaldehyde (1) is easily available
from menthone in two steps in either enantiomeric form.8 The
first two steps of the sequence are key steps that set up the
stereochemistry of the nitrogen-bearing chiral carbon. Then,
different methods of cleavage of the chiral auxiliary either lead
to N-heterocycles 8, carbocycles 9, or amino acids 11. Com-
pounds 8 and 9 are particularly well suited for the synthesis of
mono- or polycyclic alkaloids.
a Isolated yield of pure 3. b Measured by HPLC with authentic
mixtures of 3 and 4. c Without AlMe3. d ratio measured by 1H NMR.
e yield and ratio correspond to the addition of alkynyllithium to 1: pure
propargyl alcohol was then reduced to 3k with Red-Al in 77% yield.
TABLE 2. Tandem Mitsunobu/3,3-Sigmatropic Rearrangement of
ꢀ-Alcohols 3a-l
Results and Discussion
Synthesis of the Allylic Azides. The required vinyliodides
2 were prepared from the corresponding alkynes by hydrozir-
conation or hydroboration and a halogen quench. Metal-halogen
exchange with common lithium bases generated the vinyllithium
species corresponding to 2 that were added to p-menthane-3-
carboxaldehyde (1) in the presence of AlMe3 to give, stereo-
selectively, separable mixtures of diastereomeric alcohols 3 and
4 (Table 1, entries 1, 2, 4-10, and 12). The Felkin-Anh
selectivity of the addition reaction decreases considerably in
the absence of AlMe3 (c.f. entries 1, 3 and 11).9
Submitting alcohols 3a-l to the Mitsunobu reaction condi-
tions in the presence of hydrazoic acid led to their stereoselective
conversion to the allylic azides 5a-l (Table 2). With the
exception of azides 12j, 12k, and 12l (entries 10, 11, and 12),
all other azides were obtained as a single regioisomer 5 in good
to excellent diastereomeric ratios. The diastereomeric ratio of
the azide product 5 was generally higher when the size of the
substituent R was larger (compare entries 2-3 with 4-5). The
minor alcohols 4 were also submitted to the Mitsunobu/3,3-
sigmatropic rearrangement reaction (Table 3, vide infra).
Two mechanistic pathways were initially considered to
explain our results: an anti SN2′ displacement of the phospho-
nyloxy leaving group by the azide or a SN2 displacement
a Isolated yield of 5 + 6. b Measured by HPLC with authentic
mixtures of 5 and 6. c measured by NMR. d measured by NMR after
reduction and conversion to the Mosher amide. e d.r. of 12.
(8) Spino, C.; Godbout, C.; Beaulieu, C.; Harter, M.; Mwene-Mbeja, T. M.;
Boisvert, L. J. Am. Chem. Soc. 2004, 126, 13312–13319.
(9) Spino, C.; Granger, M.-C.; Tremblay, M.-C. Org. Lett. 2002, 4, 4735–
4737.
followed by a 3,3-sigmatropic rearrangement to the thermody-
namically more stable regioisomer. Mitsunobu reactions on
allylic alcohols rarely proceed by a SN2′ pathway, although some
6240 J. Org. Chem. Vol. 73, No. 16, 2008

Rearrangement of Chiral Allylic Azides
TABLE 3. Effect of Alcohol Stereochemistry and Double Bond
Geometry on the Selectivity of the Mitsunobu Reactions of Alcohols
3 or 4
FIGURE 1. Proposed transition states for the SN2 and SN2′ displace-
ments of azide on the allylic phosphonyloxy intermediate from alcohols
3 and 4.
varied from 73% in THF to 79% in benzene and 82% in
dichloromethane while its %de was 87, 88, and 82% respectively.
We can safely conclude from these results that the reaction
does not involve an allylic carbocation. Indeed, alcohols 3 and
4 would then lead to identical carbocations and should thus have
given indentical ratios of azides 5 and 6. In addition, it would
be surprising if the stereoselectivity of a purported SN2′ pathway
would be so sensitive to alkene geometry or carbinol stereo-
chemistry (Figure 1). Rather, we think that the stereoselectivity
depends on a competition between the SN2 and SN2′ pathways,
the former being stereospecific, the latter being poorly stereo-
selective. The displacement is followed by a stereospecific
rearrangement that favors isomer 5 because of steric repulsion
between the azide and the bulky menthyl nucleus. The alkene
geometry and the carbinol stereochemistry could influence the
amount of SN2 vs SN2′ attack by the azide ion: for example, a
small change in conformation around the carbinol methine could
raise or lower the energy of the SN2 transition state while the
energy of the SN2′ transitions state would likely remain the same
(Figure 1).
On the other hand, SN2′ displacements are thought to occur
principally syn to the leaving group, especially with small
nucleophiles like an azide ion.12 In our case, all stereoselec-
tivities were in favor of the “anti” displacement product. Thus,
we have carried out a series of reactions varying the phosphine
for its electronic and steric size (Table 4): if indeed a SN2′
displacement were the prevalent pathway, we would expect to
see variations in the stereoselectivity of the reaction. Electron-
deficient phosphines gave better yields of products, tris(pen-
tafluorophenyl)phospine being an exception (entry 9), but more
importantly, the regioselectivity was again complete for the
γ-product 5d. In addition, the diastereoselectivity seemed
independent of the size of the phosphine group. In fact, many
bulkier phosphines gave a slightly lower ratio of diastereomers
(entries 2, 10, and 11). Such results are difficult to reconcile
with a direct SN2′ displacement of the leaving group.
cases involving electronically biased alkenes are known.10 Large
nucleophiles, such as imides, can also have a preference for
SN2′ attack, and Mulzer and co-workers established that the
substitution pattern on the allylic alcohol may influence the
regioselectivity of Mitsunobu reactions involving phthalimide
as nucleophile.11 Since the phthalimide products cannot rear-
range, the study provides a regioselectivity trend (SN2 Vs SN2′)
for such substrates and nucleophile. These authors suggested
that steric effects and positive charge densities both influence
the regiochemical outcome of the reaction. Predictably, the SN2
reactions were completely stereospecific while the stereoselec-
tivities of the SN2′ reactions ranged from 4:1 to 9:1. The
stereochemistries of the major product from the SN2′ attack
corresponded to an anti attack of the phthalimide on an
A1,3-strain minimized conformation.12
As a first mechanistic investigation into our particular system,
we looked at what effect the stereochemistry of the carbinol
carbon and the geometry of the double bond would have on
the stereoselectivity of the reaction (Table 3). It is clear from
the results that both are affecting considerably the stereoselec-
tivity of the reaction (compare entries 1 vs 3 and 2 vs 4) but
not its regiochemical outcome as only regioisomers 5 were
observed in all cases studied. The Z geometry was particularly
detrimental to the stereoselectivity of the reaction (entries 3 and
4). Also, the R-alcohols generally gave slightly lower stereo-
selectivities than their corresponding ꢀ-alcohols (compare entries
1, 2, 4, 7-12 in Table 2 with entries 1, 5, 6, 7-12 in Table 3,
respectively). Conversely, the solvent seemed to have only a
marginal effect overall: for compound 3a, yields of azide 5a
Now, lets consider the fact that alcohols 3j-k led to mixtures
of regioisomers 5j-k and 12j-k, favoring the latter, probably
because conjugation renders isomer 12 more stable (Table 2,
entries 10 and 11). Moreover, the reactions of 3j and 3k were
less stereoselective than the other ones. Yet, their diastereomers
4j and 4k underwent the same reaction to give the opposite
mixture of diastereomers 13j:12j (and 13k:12k) in similar yields
and ratios (Table 3, entries 10 and 11). Also, alcohol 3l, having
a silyl group on the alkene, led to a 1:1 mixture of regioisomers
(10) (a) Charrette, A. B.; Coˆte´, B.; Monroc, S.; Prescott, S. J. Org. Chem.
1995, 60, 6888–6894. (b) Sobti, A.; Sulikowski, G. A. Tetrahedron Lett. 1994,
35, 3661–3664. (c) Ramesh, N. G.; Balasubramanian, K. K. Tetrahedron 1995,
51, 255–272.
(11) Mulzer, J.; Funk, G. Synthesis 1995, 101–112.
(12) (a) Magid, R. M. Tetrahedron 1980, 36, 1901–1904. (b) Stohrer, W.-
D. Angew. Chem., Int. Ed. Engl. 1983, 22, 613–614.
J. Org. Chem. Vol. 73, No. 16, 2008 6241

Lauzon et al.
SCHEME 2. Tandem Mitsunobu/3,3-Sigmatropic Rearrangement of Alcohol 3m
SCHEME 3. Mitsunobu Reaction of Dienol 3n
TABLE 4. Tandem Mitsunobu/3,3-Sigmatropic Rearrangement of
Alcohols 3d with Different Phosphines
geometry effectively keeps the stereochemical integrity of the
allylic azide intact despite a fast equilibrium between the two
regioisomers even at room temperature (Figure 2, II and III)
because allylic azides Z-5 possessing a Z double bond were
never observed.
The above results are thus best explained in the following
way: displacement of the phosphonyloxy group by the azide
ion takes place mainly by way of the SN2 mechanism. That
part is stereospecific. Then a concerted 3,3-sigmatropic
rearrangement generates azides 5, stereoselectively. However,
as the positive charge density becomes higher on the distal
alkene carbon of the alcohols 3, an increased amount of
displacement by the azide ion occurs via the SN2′ mechanism,
which is poorly stereoselective. Thus, aryl-substituted allylic
alcohols 3j and 3k as well as trisubstituted allylic alcohol
3m give rise to more SN2′ displacement then the alkyl
monosubstituted allylic alcohols 3a-i. That causes the poorer
stereoselectivity observed in the reactions of 3j,k,m. This is
certainly in line with the findings of Mulzer and Funk on
the Mitsunobu of imides.11 The silyl-substituted allylic
alcohol 3l probably undergoes solely a SN2 displacement
because of the steric effect of the bulky silyl group and a
decreased positive charge at the distal alkene carbon (the
one bearing the silyl group). The size (and perhaps the
electronics) of the silyl group leads to an equal mixture of
regioisomers 5l and 12l, each diastereomerically pure because
the rearrangement is concerted.
entry
PR3
yield (%)
5d/6da
5d/12d
1
2
3
Ph3P
94
15
0
0
85
0
50
92
90
0
50
99
94:6
91:9
s
s
92:8
s
92:8
92:8
92:8
s
91:9
91:9
<99:1
<99:1
s
s
<99:1
s
<99:1
<99:1
<99:1
s
<99:1
<99:1
(o-Tol)3P
Cy3P
(p-MeO-Ph)3P
(p-Cl-Ph)3P
(EtO)3P
4
5
6
7
8
9
10
11
(thp)3P
(pyr)Ph2P
(m-Cl-Ph)3P
(C6F5)3P
DPPA
(Napht)Ph2P
a Measured by GC (pure 6d was prepared separately starting from
4d).
5l and 12l but, this time, each isomer was diastereomerically
pure (Table 2, entry 12). Similarly, starting with diastereomeric
alcohol 4l, diastereomerically pure 6l and 13l were obtained as
a 1:1 mixture. Lastly, trisubstituted allylic alcohol 3m was
treated under the same reaction conditions to afford a single
regioisomers 5m with a modest 4:1 ratio of diastereomers
(Scheme 2).
Mechanistic studies by Padwa and co-workers on the 3,3-
sigmatropic rearrangement of some allylic azides led them to
believe that the reaction is a concerted one and likely does not
involve ion pairs.13 This is certainly the case here as regioiso-
mers 5 would likely be isolated as mixtures of diastereomers
were ionic intermediates involved, unless tight ion pairs were
involved with no time for bond rotation (Figure 2, ion pair I).
A stereoselective rearrangement following a stereospecific SN2
displacement of the azide ion, on the other hand, would account
for the overall stereoselectivity of the reaction. The double bond
Interestingly, we submitted dienol 3n to the Mitsunobu
reaction conditions and observed the formation of azide 14n
exclusively, isolated as a 3:1 mixture of diastereomers
(Scheme 3). The major diastereomer was unambiguously
identified as 14n by reduction to the mixture of amines 16n/
17n and their conversion to the separable Mosher esters 18n
and 19n (Scheme 4). Compound 18n crystallized and a X-ray
diffraction analysis on a single crystal confirmed the structure.
It is likely that the SN2 pathway is again the major one and
that neither regioisomer 12n nor 5n is thermodynamically
favored, the former because of steric reasons, the latter
because the diene is not conjugated. Stereoisomeric alcohol
4n gave the corresponding azide 15n as a 1:2.3 mixture of
diastereomers (Scheme 3). Conversion to the separable
(13) Padwa, A.; Sa, M. M. Tetrahedron Lett. 1997, 38, 5087–5090.
6242 J. Org. Chem. Vol. 73, No. 16, 2008

Rearrangement of Chiral Allylic Azides
SCHEME 4. Proof of Stereochemistry of Azide 14n and 15n
SCHEME 5. Synthesis of Non-Racemic r-Amino-Acids 20
Mosher esters 19n and 18n confirmed that 15n was the major
compound. The low observed diastereomeric ratios indicate
that the vinylic substituent, like the phenyl substituent in 3j,
favors an increased amount of SN2′ displacement probably
because of an increase in positive charge density at that
carbon. There could have been some attack at the carbon δ
to the alcohol as well.
Thus, only groups capable of stabilizing the double bond,
by conjugation or otherwise, will disrupt the perfect ther-
modynamic regioselectivity for this reaction. Notably, even
the bulky t-butyl group does not appreciably destabilize the
desired regioisomer 5d. To the best of our knowledge, this
is the first chiral auxiliary system capable of biasing to such
a degree the thermodynamic equilibrium of allylic azides.
Regioisomers 5 are quite useful compounds that can lead to
amino acids or N-heterocycles depending on how we cleave
the auxiliary, that is to say by oxidative cleavage or by RCM
(vide infra), respectively.
Synthesis of Amino Acids. Reduction of azides 5b,d,e,g
procured the corresponding amines 16b,d,e,g with no trace of
undesired regioisomers (similarly, the corresponding azides
6b,d,e,g were converted to amines 17b,d,e,g). This is perhaps
a reflection of the very low concentration of azides 12 (or 13)
and/or of the higher reactivity of azides 5 (or 6). We have then
made a number of Fmoc-protected R-amino acids 20 by
protection of the amine and ozonolysis of the double bond
(Scheme 5). The measured enantiomeric excesses of the amino
acids matched the diastereomeric excesses of the corresponding
azide.
FIGURE 2. Mechanism and stereospecificity of the 3,3-sigmatropic
rearrangement.
Synthesis of N-Heterocycles. Perhaps one of the most
interesting feature of our chiral auxiliary system is its capability
to undergo RCM cleavage to give directly carbo- or heterocyclic
compounds and the chiral auxiliary in a recoverable form (Figure
3).14 The latter (compound 21) is highly volatile and very easy
to remove from the mixture and recover. It can be transformed
FIGURE 3. Carbo- and heterocycles from RCM cleavage of the chiral
auxiliary.
J. Org. Chem. Vol. 73, No. 16, 2008 6243

Lauzon et al.
SCHEME 6. RCM Cleavage of the Auxiliary on N-Boc Derivatives 22
back to 1 by ozonolysis. Because the chain bearing the alkene
for the RCM cleavage can be at one of three places, one can
prepare a number of interesting structures (Figure 3).
biological activities.19 To demonstrate the usefulness of our
overall strategy, we carried out the synthesis of some simple
alkaloids (Figure 4).
Schemes 6 and 7 depict examples of RCM cleavages of
certain derivatives of some of the azides listed in Table 2. The
RCM reactions are divided in two categories: those that were
carried out on protected amines 22 (Scheme 6) and those that
were carried out on amides 24, 25, or 27 (Scheme 7). The first
category requires very mild conditions using 1 mol% of the
Grubbs-Nolan catalyst or Grubbs’ first generation catalyst, in
refluxing dichloromethane. Protection of the amine is necessary
otherwise no reaction takes place. Recovery of the auxiliary 21
was not done systematically but it was isolated several times
in excess of 80% yield.
The amides 24, 25, and 27 required higher catalyst loading
at 5 mol% but proceeded well at the refluxing temperature of
dichloromethane. Normally, R,ꢀ-unsaturated carbonyls are not
good alkenes with which to initiate the RCM.15,16 That is why
we first performed the RCM via the relay RCM strategy starting
with amide 24o (Scheme 7).17 In 10 min, the reaction was
complete and gave 85% yield of pyrrolidone 26o. Surprisingly,
the RCM was slower but just as efficient when carried out
directly from amide 25o. In this case, the RCM must be initiated
at the R,ꢀ-unsaturated amide since initiation at the internal
double bond is not possible.15b
Double bond reduction in 23b gave the known N-protected
alkaloid (+)-coniine 32 for a total of eight steps and 19% yield
starting from 1-pentyne (Scheme 8).20 Coniine is often used as
prototypical target for piperidine synthesis and due to its high
volatility, it is either converted to its N-Boc derivative or isolated
as it hydrochloride salt.18 It is a major toxic principle of poison
hemlock (Conium maculatum) and was apparently the first
alkaloid ever synthesized by Ladenburg in 1886.21
Lentiginosine is part of a small family of polyhydroxylated
indolizidine alkaloids, some of which are potent antiviral
agents.22 Lentiginosine was isolated from Astralagus lentigi-
nosus and its structure was elucidated in 1990 by Elbein and
co-workers.20b However, its absolute configuration was revised
after Brandi and co-workers achieved the total synthesis of (+)-
and (-)-lentiginosine.23 It was found to be a reasonably good
inhibitor of fungal R-glucosidase, amyloglucosidase (Ki ) 1 ×
10-5 M), but not of other R-glucosidases. It has been synthesized
by Isobe and his team using a rearrangement of cyanates to
isocyanates.24 We prepared (+)-lengitinosine from 23h by
stereoselective epoxidation of the alkene,25 cyclization, and
regioselective opening of the epoxyde (Scheme 9). Starting from
(19) (a) Hesse, M. In The Alkaloids, Nature’s Curse or Blessing?; Wiley-
VCH: Zu¨rich, 2002; p 413. (b) Yamamura, S.; Hirata, Y. In The Alkaloids;
Manske, R. H. F. Ed.; Academic Press: New York, 1975; Vol. 15, p 41. (c)
Yamamura, S. In The Alkaloids; Brossi, A. Ed.; Academic Press: New York,
1986; Vol. 29, p 265. (d) Weinreb, S. Chem. ReV. 2006, 106, 2531–2549. (e)
Kintzios, S. E. Critical ReV. Plant Sci. 2006, 25, 79–113.
RCM cleavage of amides 27c and 27h required the use of an
external Lewis acid.18 The cause of failure when there is no
Lewis acid is the strong coordination of the amide oxygen to
the ruthenium carbene intermediate.16 We surveyed several
Lewis acids but PhBCl2 gave the best results.
Synthesis of Natural Alkaloids. N-Heterocycles have a huge
potential as therapeutic agents and they are valuable synthetic
intermediates to natural alkaloids that possess large spectra of
(20) For selected syntheses of coniine or N-Boc-coniine, see: (a) Reding,
M. T.; Buchwald, S. L. J. Org. Chem. 1998, 63, 6344–6347. (b) Hunt, J. C. A.;
Laurent, P.; Moody, C. J. Chem. Comm. 2000, 1771–1772. (c) Charette, A. B.;
Grenon, M.; Lemire, A.; Pourashraf, M.; Martel, J. J. Am. Chem. Soc. 2001,
123, 11829–11830. (d) Wilkinson, T. J.; Stehle, N. W.; Beak, P. Org. Lett. 2000,
2, 155–158. (e) Bois, F.; Gardette, D.; Gramain, J.-C. Tetrahedron Lett. 2000,
41, 8769–8772.
(14) (a) Boisvert, L.; Beaumier, F.; Spino, C. Can. J. Chem. 2006, 84, 1290–
1293. (b) Spino, C.; Boisvert, L.; Douville, J.; Roy, J.; Lauzon, S.; Minville, J.;
Gagnon, D.; Beaumier, F.; Chabot, C. J. Organomet. Chem. 2006, 691, 5336–
5355, and references cited therein.
(21) (a) Lo´pez, T. A.; Cid, M. S.; Bianchini, M. L. Toxicon 1999, 37, 841–
865, and references cited therein. (b) Ladenburg, A. Ber. Chem. Gesel. 1886,
19, 439–441.
(22) (a) Pastuszak, I.; Molyneux, R. J.; James, L. F.; Elbein, A. D.
Biochemistry 1990, 29, 1886–1891. (b) Alkaloids: Chemical and Biological
PerspectiVes; Elbein, A. D., Molyneux, R. J., Pelletier, S. W. Eds.; Pergamon
Press: New York, 1996; Vol. 5, Chapter 1.
(15) (a) Handbook of Metathesis; Grubbs, R. H. Ed.; Wiley-VCH: Weinheim,
Germany, 2003. (b) Huang, J.; Stevens, E. D.; Nolan, S. P.; Peterson, J. L. J. Am.
Chem. Soc. 1999, 121, 2674–2678. (c) Scholl, M.; Trnka, T. M.; Morgan, J. P.;
Grubbs, R. H. Tetrahedron Lett. 1999, 40, 2247–2250.
(23) Brandi, A.; Cicchi, S.; Corderon, F. M.; Frignoli, R.; Goti, A.; Picasso,
S.; Vogel, P. J. Org. Chem. 1995, 60, 6806–6812.
(16) For examples, see: Rodr´ıguez, S.; Castillo, E.; Carda, M.; Marco, J. A.
Tetrahedron 2002, 58, 1185–1192.
(24) (a) Ichikawa, Y.; Ito, T.; Isobe, M. Chem. Eur. J. 2005, 11, 1949–1957.
For other selected syntheses see: (b) Cardona, F.; Moreno, G.; Guarna, F.; Vogel,
P.; Schuetz, C.; Merino, P.; Goti, A. J. Org. Chem. 2005, 70, 6552–6555. (c)
Raghavan, S.; Sreekanth, T. Tetrahedron: Asymmetry 2004, 15, 565–570. (d)
El-Nezhawy, A. O. H.; El-Diwani, H. I.; Schmidt, R. R. Eur. J. Org. Chem.
2002, 4137–4142. For a review on synthetic approaches to related indolizidines,
see: (e) Burgess, K.; Henderson, I. Tetrahedron 1992, 48, 4045–4066.
(25) Yang, D.; Wong, M. K.; Yip, Y. C. J. Org. Chem. 1995, 60, 3887–
3889.
(17) (a) Hansen, E. C.; Lee, D. Org. Lett. 2004, 6, 2035–2038. (b) Hoye,
T. R.; Jeffrey, C. S.; Tennakoon, M. A.; Wang, J.; Zhao, H. J. Am. Chem. Soc.
2004, 126, 10210–10211. (c) Wallace, D. Angew. Chem., Int. Ed. 2005, 44, 1912–
1915.
(18) (a) Vedrenne, E.; Dupont, H.; Oualef, S.; Elka¨ım, L.; Grimaud, L. Syn.
Lett 2005, 4, 670–672. (b) Binot, G.; Quiclet-Sire, B.; Zard, S. Z. Syn. Lett.
2003, 3, 382–384. (c) Giese, B.; Kopping, B.; Dickhaut, J.; Thoma, G.; Kuliecke,
K. J.; Trach, F. Org. React. 1996, 48, 301–315.
6244 J. Org. Chem. Vol. 73, No. 16, 2008

Rearrangement of Chiral Allylic Azides
SCHEME 7. RCM of Amides 24a, 25o, 27c, and 27h
SCHEME 8. Synthesis of (+)-N-Boc-coniine (+)-32
SCHEME 9. Synthesis of (+)-Lentiginosine 30
5-hexynol, the synthesis of (+)-lentiginosine was achieved in
13 steps in 7% overall yield for an average of 82% yield per
step.
FIGURE 4. Alkaloids synthesized.
South American poison dart frogs secrete toxic alkaloids in
their skin. Pumiliotoxin C was the first member to be isolated
from Dendrobates pumilio, also called the strawberry poison
frog in Costa Rica.26 Frogs held in captivity do not secrete
alkaloids. Scheloribatid mites have been identified as an
important source of pumiliotoxins for the frogs.27 The structure
of pumiliotoxin C was revealed by synthesis and X-ray analysis
in 197528 but the absoluted configuration of natural (-)-
pumiliotoxin C was confirmed as that shown in Scheme 10 after
Oppolzer and co-workers completed the synthesis of both
enantiomers.29
Iodide 2i was prepared in four steps from racemic 2-meth-
ylfuran (Scheme 10). Tables 1 and 2 indicate the yields and
selectivities of its converstion to azide 5i. Then, reduction of
azide 5i with LiAlH4 followed by acylation of the resulting
amine afforded amide 27i. RCM in refluxing dichloromethane
in the presence of dichlorophenylborane was efficient and led
directly to the N-heterocycle. It was found that an aqueous
workup of this reaction caused partial deprotection of the
alcohol, probably from hydrochloric acid generated by the
reaction of water and the dichloroborane. Indeed, it was possible
to optimize this procedure to obtain directly 28i in quantitative
yield using an acidic workup. Conversion of the alcohol to
iodide 38 went uneventfully. Radical cyclization formed the
second ring of pumiliotoxin C stereoselectively (39) in modest
yield. We tried a vast number of variations on this cyclization,
including replacing the iodide by bromide, thiocarbonate, or
xanthate, changing the solvents, temperature, radical initiation
procedures, using hydridre free conditions and atom transfert
techniques, but we were unable to augment the yield of 39.30
Zard pointed out that 6-exotrig cyclizations to give decaline-
type systems are particularly difficult.31 In our particular case,
facile hydrogen abstraction compounds the problem and leads
to reduced (40) or aromatized products (41) as well as
dimerization products (42 and 43) (Figure 5). Nonetheless, the
radical cyclization created two chiral centers stereoselectively
via transition state 38A. The synthesis was completed by imino
ether formation, Grignard addition and reduction of the unstable
imine in good yield for the three steps.32 Overall, the synthesis
was carried out in 15 steps from 2-methylfuran with an overall
(26) (a) Daly, J. W.; Tokuyama, T.; Habermehl, G.; Karle, I. L.; Witkop, B.
Liebigs Ann Chem 1969, 729, 198–212. Close to 200 other alkaloids have been
isolated from poisonous frogs. See: (b) Tokuyama, T.; Tsujita, T.; Shimada, A.;
Girrafo, H. M.; Spande, T. F.; Daly, J. W. Tetrahedron 1991, 47, 5401–5414,
and references cited therein.
(27) Takada, W.; Sakata, T.; Shimano, S.; Enami, Y.; Mori, N.; Nishida, R.;
Kuwahara, Y. J. Chem Ecol. 2005, 31, 2403–2415.
(30) Maruoka, K.; Miyazaki, T.; Ando, M.; Matsumura, Y.; Sakane, S.;
Hattori, K.; Yamamoto, H. J. Am. Chem. Soc. 1983, 105, 2831–2843.
(31) Zard, S. Z. In Radical Reactions in Organic Synthesis; Oxford University
Press: New York, 2003; p 224.
(28) Ibuka, T.; Inubushi, Y.; Saji; Tankana, K.; Masaki, N. Tetrahedron.
Lett 1975, 16, 323–326.
(29) Oppolzer, W.; Flaskamp, E. HelV. Chem. Acta 1977, 60, 204–207.
(32) Mehta, G.; Praveen, M. J. Org. Chem. 1995, 60, 279–280.
J. Org. Chem. Vol. 73, No. 16, 2008 6245

Lauzon et al.
SCHEME 10. Synthesis of (+)-Pumiliotoxin C 31
combined organic extracts were washed with 1 portion of saturated
solution of Na2S2O3, one portion of water, and once with brine,
dried over anhydrous MgSO4, filtered, and concentrated under
reduced pressure to give a colorless oil. The crude product was
purified on silica gel with 5:95/EtOAc:Hexanes as eluant to afford
1
iodide 2h as a colorless oil (10.67 g, 77%). H NMR (300 MHz,
CDCl3): δ (ppm) 7.67-7.64 (dd, 4H, J ) 7.7 Hz, 1.7 Hz),
7.43-7.35 (m, 6H), 6.53-6.44 (m, 1H), 5.94 (d, 1H, J ) 14.3
Hz), 3.64 (t, 2H, J ) 5.8 Hz), 2.03 (q, 2H, J ) 5.1 Hz), 1.59-1.41
(m, 4H), 1.05 (s, 9H). 13C NMR (75.5 MHz, CDCl3) δ (ppm) 146.5
(d), 135.7 (d), 134.1 (s), 139.7 (d), 127.9 (d), 74.9 (s), 63.6 (t),
35.9 (t), 31.9 (t), 27.1 (q), 24.8 (t), 19.4 (s). IR (film) • (cm-1):
3070, 3049, 2931, 2858, 1472, 1427, 1111, 943. LRMS (m/z,
relative intensity): 407 [M - C4H9]+, 64), 309 (100), 249 (20),
199 (44), 181 (19), 81 (27). HRMS calcd for C18H20IOSi: 407.0328,
found 407.0325.
(E)-5-t-Butyldimethylsiloxy-1-iodo-1-heptene 2i. 2-Methyltet-
rahydrofuran 36 (20 mL, 200 mmol) was added to a solution of
aluminum chloride (150 g, 1.00 mol) and sodium iodide (133 g,
1.00 mol) in acetonitrile (220 mL) at 0 °C. The mixture was stirred
2 h at 0 °C, and a 1 N HCl aqueous solution was added until
complete dissolution of salts. The aqueous layer was extracted with
pentane (3 × 300 mL), and the combined organic layers were dried
over anhydrous MgSO4 and then concentrated under reduced
pressure. The iodoalcohol (41.6 g, 97%) was obtained as a colorless
oil and used without further purification.
Imidazole (33.0 g, 485 mmol) and t-butylchlorodimethylsilane
(32.0 g, 212 mmol) were added to the iodoalcohol (41.6 g, 194
mmol) in dichloromethane (800 mL) at RT. The mixure was stirred
at RT for 16 h and then quenched with water (500 mL). The
aqueous layer was extracted with dichloromethane (3 × 400 mL)
and the combined organic layers were washed once with brine, dried
over anhydrous MgSO4 and then concentrated under reduced
pressure. The crude product was purified by flash chromatography
on silica gel (from 0 to 2% diethyl ether in hexanes) to yield pure
1
iodide 37 (50.3 g, 83%) as a colorless oil. H NMR (300 MHz,
CDCl3) δ (ppm) 3.82 (sext, 1H, J ) 6.0 Hz), 3.19 (t, 2H, J ) 7.2
FIGURE 5. Transitions state and byproduct of the radical cyclization
of 36.
Hz), 1.98-1.75 (m, 2H), 1.54-1.46 (m, 2H), 1.13 (d, 3H, J ) 6.0
Hz), 0.88 (s, 9H), 0.05 (s, 3H), 0.04 (s, 3H). IR (neat) • (cm-1
)
2955, 2927, 2885, 2856, 1465, 1317, 1248, 1129, 1082, 832, 770.
yield of 4% and an average yield of 81% per step. So despite
the modest yield of the radical cyclization, the overall synthesis
is short and efficient.
LRMS (m/z, relative intensity) 313 ([M+ - CH3], 2), 271 ([M+
-
C4H9], 45), 229 (100). HRMS calcd for C7H16IOSi [M+ - C4H9]:
271.0015, found: 271.0006.
In conclusion, an unusual tandem Mitsunobu/3,3-sigmatropic
rearrangement of allylic azides combined with RCM cleavage
of our chiral auxiliary allows for synthesis nonracemic of
N-heterocycles. The role of the chiral auxiliary is to induce
stereochemistry but also to sterically bias the equilibrium
mixture of the allylic azides. Chiral quaternary azides are also
available by this method, albeit in lower diastereomeric excess.
We have successfully applied the strategy to the synthesis of
amino acids and natural alkaloids. The synthesis of complex
alkaloids are underway and will be presented in due course.
A solution of the iodide 37 (46.4 g, 141 mmol) in diethyl ether
(700 mL) was slowly added to lithium acetylide complexed with
ethylene diamine (85%, 31.1 g, 289 mmol) in a pentane:DMSO
solvent system (5:2, 1.23 L) at 0 °C. The mixture was stirred,
allowed to slowly warm up to RT for 16 h, and then quenched
with a half-saturated NH4Cl aqueous solution (1.0 L). The aqueous
layer was extracted with diethyl ether (3 × 800 mL), and the
combined organic layers were washed once with water, once with
brine, dried over anhydrous MgSO4, and then concentrated under
reduced pressure. The crude product was purified by flash chro-
matography on silica gel (from 0 to 5% ethyl ether in hexanes) to
1
yield pure alkyne (29.6 g, 93%) as a colorless oil. H NMR (300
MHz, CDCl3) δ (ppm) 3.81 (sext, 1H, J ) 6.0 Hz), 2.19 (td, 2H,
J ) 5.9, 2.5 Hz), 1.94 (t, 1H, J ) 2.5 Hz), 1.66-1.46 (m, 4H),
1.13 (d, 3H, J ) 6.0 Hz), 0.88 (s, 9H), 0.05 (s, 6H). 13C NMR
(75.5 MHz, CDCl3) δ (ppm) 84.3 (s), 68.2 (d), 68.0 (d), 38.6 (t),
25.8 (q), 24.7 (t), 23.8 (q), 18.4 (t), 18.0 (s), -4.5 (q), -4.8 (q). IR
(neat) • (cm-1) 3317, 2957, 2935, 2858, 2120, 1471, 1377, 1255,
1138, 1093, 1030, 837, 778. LRMS (m/z, relative intensity) 211
([M+ - CH3], 2), 169 ([M+ - C4H9], 5), 75 (100). HRMS calcd
for C12H23OSi [M+ - CH3]: 211.1518, found: 211.1514.
Shielded from light, a solution of lithium triethylborohydride in
THF (1.0 M, 33.1 mL, 33.1 mmol) was added dropwise to bis-
(cyclopentadienyl)zirconium chloride (9.68 g, 33.1 mmol) in THF
(190 mL) at 0 °C. The mixture was stirred 1 h at RT before a
solution of the 5-t-butyldimethylsiloxy-1-heptyne (5.00 g, 22.1
Experimental Section
6-t-Butyldiphenylsilyloxy-1-iodo-1-hexene 2 h. In a 500 mL
r.b. flask was dissolved 1-t-butyldiphenylsilyloxy-5-hexyne (10.00
g, 29.71 mmol) in anhydrous CH2Cl2 (100 mL). The solution was
cooled to 0 °C, and a solution of Br2BH·SMe2 1.0 M in CH2Cl2
(35.7 mL, 35.66 mmol) was added dropwise over a period of 10
min. The reaction mixture was was stirred at RT overnight (18 h),
and a solution of NaOH 3.0 M (49.5 mL, 148.57 mmol) was added
dropwise after cooling to 0 °C. A white solid was precipated. This
slurry was then stirred at RT for 3 h and then cooled again to 0 °C.
Iodine (11.30 g, 44.57 mmol) was added in small portions. The
resulting reaction mixture was stirred for 2 h at RT. Then, the
aqueous phase was extracted with 3 portions of Et2O and the
6246 J. Org. Chem. Vol. 73, No. 16, 2008

Rearrangement of Chiral Allylic Azides
mmol) in THF (30 mL) was added. The mixture was stirred 2 h at
room temprature, cooled down to 0 °C. and iodine (9.30 g, 36.6
mmol) was added in three portions. The mixture was stirred and
allowed to slowly warm up to RT for 16 h. The resulting red
solution was quenched with a saturated NaHCO3 aqueous solution
(150 mL) until the solution became yellow. The aqueous layer was
extracted with diethyl ether (3 × 150 mL). and the combined
organic layers were washed once with a saturated Na2S2O3 aqueous
solution, dried over anhydrous MgSO4. and then concentrated under
reduced pressure. The crude product was purified by flash chro-
matography on silica gel (from 0 to 1% diethyl ether in hexanes)
to yield pure vinyliodide 2i (6.1 g, 78%) as a colorless oil. 1H NMR
(300 MHz, CDCl3) δ (ppm) 6.50 (dt, 1H, J ) 14.3, 7.0 Hz), 5.97
(d, 1H, J ) 14.3 Hz), 3.77 (sext, 1H, J ) 6.0 Hz), 2.05 (q, 2H, J
) 7.0 Hz), 1.52-1.31 (m, 4H), 1.11 (d, 3H, J ) 6.0 Hz), 0.88 (s,
9H), 0.04 (s, 6H). IR (neat) • (cm-1) 2953, 2930, 2856, 1464, 1372,
1252, 1137, 1095, 1031, 833, 768. LRMS (m/z, relative intensity)
339 ([M+ - CH3], 1), 297 ([M+ - C4H9], 65), 75 (100). HRMS
calcd for C9H18IOSi [M+ - C4H9]: 297.0172, found: 297.0180.
Allylic Alcohol 3h. Prepared according to the general procedure
for allylic alcohols 3 starting with vinyliodide 2h (10.60 g, 22.84
mmol) and p-menthyl-3-carboxaldehyde 1 (3.2 g, 19.03 mmol) to
63.5 (t), 46.7 (d), 44.7 (d), 43.1 (t), 35.1 (t), 34.4 (t), 32.4 (d), 32.1
(t), 31.6 (t), 28.0 (d), 26.9 (q), 23.8 (t), 22.5 (q), 22.3 (t), 21.4 (q),
19.2 (s), 15.1 (q), 14.1 (q). IR (film) • (cm-1): 3072, 3049, 2962,
2868, 2095, 1471, 1428, 1237, 1111, 973. LRMS (m/z, relative
intensity): 549 ([M - NH4]+, 12), 508 (15), 274(45), 248 (30),
182 (50), 112 (99), 98 (100), 78 (93). HRMS calcd for
C33H53N4OSi: 549.3988, found 549.3984 (for [MNH4]). [R]20
-25.1 (c 3.67, CHCl3).
)
D
Allylic Azide 5i. Prepared according to the general procedure
for allylic azides 5 starting with allylic alcohol 3i (4.06 g, 10.2
mmol, diethyl azodicarboxylate (3.2 mL, 20.4 mmol), hydrazoic
acid solution in benzene (1.44 M, 14.2 mL, 20.4 mmol), and
triphenylphosphine (5.37 g, 20.5 mmol) in benzene (50 mL) to yield
pure azide 5i (4.16 g, 97%) as a colorless oil. 1H NMR (300 MHz,
CDCl3) δ (ppm) 5.48 (dd, 1H, J ) 15.2, 9.1 Hz), 5.32 (dd, 1H, J
) 15.2, 8.2 Hz), 3.82-3.72 (m, 2H), 2.00-1.87 (m, 2H), 1.75-1.71
(m, 1H), 1.65-1.55 (m, 2H), 1.51-1.26 (m, 7H), 1.11 (d, 3H, J )
6.0 Hz), 1.03-0.67 (m, 10H), 0.88 (s, 9H), 0.71 (d, 3H J ) 6.6
Hz), 0.04 (s, 6H). 13C NMR (75.5 MHz, CDCl3) δ (ppm) 140.3
(d), 126.9 (d), 68.4 (d), 65.0 (d), 46.7 (d), 44.7 (d), 43.2 (t), 39.3
(t), 39.2 (t), 35.1 (t), 34.8 (t), 32.4 (d), 28.0 (d), 25.9 (q), 23.8 (q),
22.5 (q), 22.2 (t), 21.4 (q), 18.1 (s), 15.1 (q), -4.4 (q), -4.8 (q).
IR (neat) • (cm-1) 2958, 2934, 2859, 2098, 1468, 1374, 1251, 1054,
979, 843. LRMS (m/z, relative intensity) 393 ([M+ - N2], 5), 378
(M+ - CH3N2, 10), 364 (M+-C4H9, 5), 74 (100). HRMS calcd for
1
give allylic alcohol 3h as a colorless oil (4.31 g, 45%). H NMR
(300 MHz, CDCl3): δ (ppm) 7.66 (dd, 4H, J ) 7.7 Hz, 1.7 Hz),
7.45-7.34 (m, 6H), 5.65-5.47 (m, 2H), 4.37 (d, 1H, J ) 4.4 Hz),
3.67 (t, 2H, J ) 6.1 Hz), 2.13 (dquint, 1H, J ) 6.6 Hz, 2.2 Hz),
2.04 (q, 2H, J ) 6.8 Hz), 1.72-1.46 (m, 9H), 1.44-1.24 (m, 3H),
1.04 (s, 9H), 1.05-0.79 (m, 2H), 0.93 (d, 3H, J ) 7.2 Hz), 0.87
(d, 3H, J ) 6.6 Hz), 0.76 (d, 3H, J ) 7.2 Hz). 13C NMR (75.5
MHz, CDCl3) δ (ppm) 135.6 (d), 134.1 (s), 132.6 (d), 130.4 (d),
129.5 (d), 127.6 (d), 71.3 (d), 63.7 (t), 44.8 (d), 43.0 (d), 35.2 (t),
33.9 (t), 32.8 (d), 32.1 (t), 26.9 (q), 26.3 (d), 25.6 (t), 24.3 (t), 22.9
(q), 21.6 (q), 19.2 (s), 15.5 (q). IR (film) • (cm-1): 3447, 3069,
2956, 2931, 2859, 1472, 1428, 1111, 973. LRMS (m/z, relative
intensity): 449 ([M - C4H9]+, 10), 293 (14), 233 (21), 199 (100),
177 (22), 137 (49), 109 (29), 95 (64), 81 (39). HRMS calcd for
C20H38N3OSi [M+ - C4H9]: 364.2784, found: 364.2786. [R]20
-35.7 (c ) 1.63, CHCl3).
)
D
Allylic Amine 16h. As per amine 16i starting from azide 5h
(1.00 g, 1.88 mmol), LiAlH4 (powder 95%, 107 mg, 2.82 mmol)
1
to afford 16h as a colorless oil (820 mg, 86%). H NMR (CDCl3,
300 MHz): δ 7.67-7.65 (m, 4H), 7.44-7.34 (m, 6H), 5.39-5.22
(m, 2H), 3.64 (t, 2H, J ) 6.6 Hz), 3.12-3.22 (m, 1H), 1.89-1.24
(m, 12H), 1.04 (s, 9H), 1.01-0.65 (m, 6H), 0.85 (d, 3H, J ) 7.2
Hz), 0.84 (d, 3H, J ) 6.6 Hz), 0.69 (d, 3H, J ) 6.6 Hz). 13C NMR
(75.5 MHz, C6D6) δ (ppm) 135.6 (s), 135.1 (s), 134.0, (s), 129.5
(s), 127.7 (s), 63.7 (t), 53.9 (s), 47.2 (s), 44.4 (s), 43.5 (t), 37.9 (t),
35.2 (t), 32.7 (t), 32.5 (s), 28.0 (s), 26.9 (t), 24.2 (q), 21.4 (t), 19.1
(q), 15.4 (q). IR (neat, cm-1): 3891, 3366, 3070, 2955, 2929, 2859,
1472, 1428, 1111, 971. LRMS (m/z, (relative intensity)): 506
([MH]+, 84), 489 (49), 448 (100), 293 (30), 194 (100), 137 (13),
95 (8). HRMS calcd for C33H51NOSi: 505.3740, found: 505.3731.
C29H41O2Si: 449.2876, found 449.2870. [R]20 ) -7.7 (c 2.57,
D
CHCl3).
Allylic Alcohol 3i. Prepared according to the general procedure
for allylic alcohols 3 starting with vinyliodide 2i (320 mg, 0.90
mmol) in diethyl ether (3 mL) at -78 °C, (-)-p-menthane-3-
carboxaldehyde 1 (130 mg, 0.77 mmol) to yield pure alcohol 3i
[R]20 ) -20.5 (c ) 2.34, CHCl3).
D
1
(173 mg, 56%) as a colorless oil. H NMR (300 MHz, CDCl3) δ
Allylic Amine 16i. Lithium aluminum hydride (595 mg, 14.9
mmol) was added portion wise to the azide 5i (4.16 g, 9.86 mmol)
in diethyl ether (100 mL) at 0 °C. The mixture was stirred, allowed
to slowly warm up to RT for 13 h and then quenched with water
(0.6 mL), a 15% NaOH aqueous solution (0.6 mL) and water (1.8
mL). The mixture was stirred 20 min at RT and then stirred with
anhydrous Na2SO4 for 30 min. The mixture was filtered and the
solvent was removed under reduced pressure. The crude product
was purified by flash chromatography on silica gel (20% ethyl
acetate in hexanes saturated with NH4OH) to yield pure amine 16i
(ppm) 5.59 (dt, 1H, J ) 15.4, 6.3 Hz), 5.50 (dd, 1H, J ) 15.4, 4.7
Hz), 4.34 (d, 1H, J ) 4.7 Hz), 3.76 (sext, 1H, J ) 6.0 Hz), 2.12
(septd, 1H, J ) 7.0, 1.8 Hz), 2.02 (q, 2H, J ) 6.3 Hz), 1.69-1.60
(m, 3H), 1.49-1.21 (m, 9H), 1.09 (d, 3H, J ) 6.0 Hz), 0.99-0.78
(m, 5H), 0.90 (d, 3H, J ) 7.0 Hz), 0.86 (s, 9H), 0.74 (d, 3H, 7.0
Hz), 0.02 (s, 6H). 13C NMR (75.5 MHz, CDCl3) δ (ppm) 132.4
(d), 130.4 (d), 71.2 (d), 68.4 (d), 44.7 (d), 42.9 (d), 39.2 (t), 35.1
(t), 33.8 (t), 32.7 (d), 32.2 (t), 26.2 (d), 25.9 (q), 25.4 (t), 24.2 (t),
23.9 (q), 22.8 (q), 21.5 (q), 18.1 (s), 15.4 (q), -4.4 (q), -4.8 (q).
IR (neat) • (cm-1) 3385 (large), 2957, 2929, 2864, 1466, 1378,
1257, 838, 777. LRMS (m/z, relative intensity) 339 ([M+ - C4H9],
5), 74 (100). HRMS calcd for C20H39O2Si [M+ - C4H9]: 339.2719,
1
(3.50 g, 90%) as a colorless oil. H NMR (300 MHz, CDCl3) δ
(ppm) 5.30-5.17 (m, 2H), 3.75-3.66 (m, 1H), 3.17-3.15 (m, 1H),
1.87-1.76 (m, 2H), 1.68-1.63 (m, 1H), 1.56-1.42 (m, 4H),
1.41-1.19 (m, 7H), 1.06 (d, 3H, J ) 6.0 Hz), 1.01-0.70 (m, 7H),
0.83 (s, 9H), 0.81 (d, 3H, J ) 6.6 Hz), 0.65 (d, 3H, J ) 6.6 Hz),
-0.01 (s, 6H). IR (neat) • (cm-1) 2958, 2934, 2864, 1468, 1378,
1256, 1139, 838, 777. LRMS (m/z, relative intensity) 395 ([M+],
1), 380 ([M+ - CH3], 2), 338 ([M+ - C4H9] 5), 194 (100). HRMS
calcd for C24H49NOSi: 395.3583, found: 395.3591. [R]20D ) -30.5
(c ) 1.13, CHCl3).
found: 339.2723. [R]20 ) -7.9 (c ) 2.06, CHCl3).
D
Allylic Azide 5h. Prepared according to the general procedure
for allylic azides 5 starting with allylic alcohol 3h (1.50 g, 2.960
mmol) and triphenylphosphine (1.55 g, 5.919 mmol) in benzene
(30.0 mL), hydrazoic acid (4.2 mL, 1.4 M in benzene, 5.92 mmol),
and diethylazodicarboxylate (DEAD) (1.030 g, 5.919 mmol) to
1
afford azide 5h as a colorless oil (1.55 g, 99%). H NMR (300
MHz, CDCl3): δ (ppm) 7.66 (dd, 4H, J ) 7.7 Hz, 1.7 Hz),
7.45-7.34 (m, 6H), 5.51-5.43 (m, 1H), 5.36-5.27 (m,1H), 3.76
(q, 1H, J ) 7.1 Hz), 3.64 (t, 2H, J ) 6.3 Hz), 1.96-1.85 (m, 2H),
1.75-1.70 (m, 1H), 1.65-1.25 (m, 10H), 1.04 (s, 9H), 1.05-0.83
(m, 3H), 0.86 (d, 3H, J ) 7.1 Hz), 0.85 (d, 3H, J ) 6.6 Hz), 0.71
(d, 3H, J ) 6.6 Hz). 13C NMR (75.5 MHz, CDCl3): δ (ppm) 140.2
(d), 135.5 (d), 133.9 (s), 129.5 (d), 127.5 (d), 126.9 (d), 65.0 (d),
t-Butyl Carbamate 22h. In a 100 mL r.b. flask was dissolved
the amine 16h (1.575 g, 3.114 mmol) in acetonitrile (31.5 mL).
Anhydrous K2CO3 (452 mg, 3.270 mmol) was added into the
solution, and this solution was stirred 5 min at RT. Then, the allyl
bromide (414 mg, 3.425 mmol) was added very slowly (20 µL/
min) and the resulting suspension was stirred at RT for 1.5 h and
an excess of allyl bromide (0.2 eq.) was added. The resulting
J. Org. Chem. Vol. 73, No. 16, 2008 6247

Lauzon et al.
mixture in stirred again for a another 2 h at RT, diluted in H2O
and extracted with 3 portions of EtOAc. The combined organic
extracts were washed once with brine, dried over anhydrous MgSO4,
filtered, and concentrated under reduced pressure to give a yellow
oil. The crude product was purified on silica gel with 1:1/Et2O:
Hexanes as eluant to afford alkylated amine as a colorless oil (986
mg, 59%). 1H NMR (CDCl3, 300 MHz): δ 7.65 (dd, 4H, J ) 7.7,
1.7 Hz), 7.44-7.34 (m, 6H), 5.91 (ddt, 1H, J ) 17.0, 16.5, 6.1
Hz), 5.27 (dd, 1H, J ) 15.4, 8.8 Hz), 5.17-5.07 (m, 3H), 3.63 (t,
2H, J ) 6.6 Hz) 3.28 (dd, 1H, J ) 8.3-5.5 Hz)), 3.12 (dd, 1H, J
) 13.8, 6.6 Hz), 3.02-2.90 (m, 1H), 1.98-1.83 (m, 2H), 1.74-1.67
(m, 1H), 1.59-1.24 (m, 10H), 1.04 (s, 9H), 1.02-0.64 (m, 3H),
0.87 (d, 6H, J ) 6.6 Hz), 0.83 (d, 3H, J ) 6.6 Hz), 0.70 (d, 3H,
J ) 6.6 Hz). IR (neat, cm-1): 3071, 2945, 2930, 2859, 1471, 1455,
1427, 1111, 909. LRMS (m/z, (relative intensity)): 545 ([M+], 2),
504 (23), 488 (31), 234 (100), 199 (18), 183 (10), 96 (43). HRMS
calcd for C36H55NOSi: 545.4053, found 545.4038. [R]20D ) -21.7
(c 1.09, CHCl3).
Hz), 5.56-5.54 (m, 1H), 5.35-5.14 (m, 4H), 4.34 (quint, 1H, J )
6.9 Hz), 3.71 (sext, 1H, J ) 6.0 Hz), 2.94 (d, 2H, J ) 7.2 Hz),
1.87-1.65 (m, 3H), 1.58-1.19 (m, 9H), 1.05 (d, 3H, J ) 6.0 Hz),
0.97-0.72 (m, 10H), 0.84 (s, 9H), 0.63 (d, 3H, J ) 7.2 Hz), -0.01
(s, 6H). 13C NMR (75.5 MHz, CDCl3) δ (ppm) 169.3 (s), 136.6
(d), 131.6 (d), 128.9 (d), 119.4 (t), 68.5 (d), 50.8 (d), 47.0 (d),
44.4 (d), 43.0 (t), 41.8 (t), 39.4 (t), 35.3 (t), 35.1 (t), 32.4 (d), 27.9
(d), 25.9 (q), 24.0 (t), 23.8 (q), 22.5 (q), 21.9 (t), 21.4 (q), 18.1 (s),
15.3 (q), -4.5 (q), -4.7 (q). IR (neat) • (cm-1) 3273 (large), 2957,
2928, 2853, 1641, 1547, 1457, 1367, 1250, 830, 769. LRMS (m/z,
relative intensity) 463 ([M+], 20), 448 ([M+ - CH3], 5), 406 ([M+
- C4H9], 100). HRMS calcd for C28H53NO2Si: 463.3845, found:
463.3835. [R]20 ) -41.9 (c ) 0.48, CHCl3).
D
Dihydropyridinone 28i. A refluxing solution of amide 27i (702
mg, 1.51 mmol) in dichloromethane (700 mL) was bubbled with
argon for 30 min. Dichlorophenylborane (0.19 mL, 1.45 mmol)
and Grubbs second generation catalyst (65 mg, 0.077 mmol) were
respectively added at RT and the mixture was refluxed for 2 h.
The mixture was cooled down to RT and a solution of concentrated
HCl aqueous solution and methanol (1:99, 8 mL) was added. The
mixture was stirred for 16 h and concentrated under reduced
pressure. The crude product was purified by flash chromatography
on silica gel (from 5 to 10% methanol in ethyl acetate) followed
by an activated carbon treatment to yield pure alcohol 28i (277
In a 100 mL r.b. flask was dissolved the amine (1.31 g, 2.40
mmol) in DMF (20 mL) and this solution was stirred at RT for 5
min. Then, triethylamine (368 mg, 3.60 mmol) was added and this
solution was stirred for another 5 min at RT and the di-t-butyl-
dicarbonate (786 mg, 3.60 mmol) was added and this resulting
mixture was stirred at RT during 23 h. The mixture was then treated
with a saturated solution of NH4Cl and the aqueous phase was
extracted with 3 portions of EtOAc and the combined organic
extracts were washed once with water and once with brine, dried
over anhydrous MgSO4, filtered and concentrated under reduced
pressure to give a yellow oil. The crude product was purified on
silica gel with 3:97/AcOEt:Hexanes as eluant to afford the butoxy
1
mg, 100%) as a colorless oil. Diastereomer A (R or ꢀ OH): H
NMR (300 MHz, CDCl3) δ (ppm) 7.93 (s, 1H), 5.66-5.55 (m,
2H), 4.02-3.92 (m, 1H), 3.73-3.44 (m, 2H), 2.82-2.76 (m, 2H),
1.58-1.29 (m, 6H), 1.07 (d, 3H, J ) 6.0 Hz). 13C NMR (75.5 MHz,
CDCl3) δ (ppm) 170.8 (s), 125.8 (d), 121.4 (d), 66.8 (d), 53.4 (d),
38.4 (t), 36.6 (t), 30.9 (t), 23.2 (q), 19.9 (t). Diastereomer B (R or
ꢀ OH): 1H NMR (300 MHz, CDCl3) δ (ppm) 8.10 (s, 1H),
5.66-5.55 (m, 2H), 4.02-3.92 (m, 1H), 3.73-3.44 (m, 2H),
2.82-2.76 (m, 2H), 1.58-1.29 (m, 6H), 1.07 (d, 3H, J ) 5.0 Hz).
13C NMR (75.5 MHz, CDCl3) δ (ppm) 170.8 (s), 125.6 (d), 121.2
(d), 67.2 (d), 53.0 (d), 38.4 (t), 36.3 (t), 30.9 (t), 23.8 (q), 20.9 (t).
IR (neat) • (cm-1) 3541-3122, 3046, 2969, 2928, 2863, 1682, 1658,
1410, 1346, 1139. LRMS (m/z, relative intensity) 183 ([M+], 5),
182 ([M+ - H], 10), 96 (100). HRMS calcd for C10H17NO2:
1
carbamate 22h a colorless oil (1.39 g, 92%). H NMR (CDCl3,
300 MHz): δ 7.65 (dd, 4H, J ) 7.7, 1.7 Hz), 7.44-7.34 (m, 6H),
5.82-5.70 (m, 1H), 5.40-5.23 (m, 2H), 5.10-5.00 (m, 2H), 3.63
(t, 2H, J ) 6.6 Hz), 1.90-1.76 (m, 2H), 1.72-1.67 (m, 1H),
1.61-1.51 (m, 4H), 1.44 (s, 9H), 1.35-1.25 (m, 6H), 1.04 (s, 9H),
0.97-0.79 (m, 6 H), 0.84 (d, 6H, J ) 6.6 Hz), 0.67 (d, 3H, J )
6.6 Hz). LRMS (m/z, (relative intensity)): 589 ([M - C4H9]+, 1),
532 (44), 455 (61), 282 (51), 278 (100), 235 (70), 199 (41), 140
(49). HRMS calcd for C37H55NO3Si: 589.3951, found 589.3942.
Pyrrolidine 23h. In a 500 mL r.b. flask was dissolved the butyl
carbamate 22h (1.00 g, 1.55 mmol) in CH2Cl2 (310 mL, 0.005 M)
and the reaction was refluxed for 10 min. The reflux was stopped
and the Grubbs catalyst (63.7 mg, 0.077 mmol) was added in small
portions and the reaction mixture was refluxed for 18 h. The
resulting solution was concentrated under reduced pressure and the
crude product was purified on silica gel with 1:20/EtOAc:Hexanes
as eluant to afford pyrrolidine 23h as a colorless oil (743 mg,
100%). 1H NMR (CDCl3, 300 MHz): δ 7.65 (dd, 4H, J ) 7.7, 2.2
Hz), 7.44-7.33 (m, 6H), 5.75-5.68 (bd, 2H, J ) 6.6 Hz), 4.50
(m, 1H), 4.17(bd, 1H, J ) 13.2 Hz), 3.98 (dd, 4H, J ) 9.9, 5.5
Hz), 3.63 (t, 2H, J ) 6.6 Hz), 1.69-1.51 (m, 4H), 1.46 (s, 9H),
1.35-1.24 (m, 2H), 1.04 (s, 9H). IR (neat, cm-1): 3071, 2861, 1704,
1697, 1427, 1392, 1174, 1111, 910. LRMS (m/z, (relative inten-
sity)): 480 ([MH]+, 39), 380 (62), 366 (58), 288 (49), 244 (16), 83
(100). HRMS calcd for C29H42NO3Si: 480.2934, found 480.2942
183.1259, found: 183.1252. [R]20 ) +40.8 (c ) 1.22, CHCl3).
D
Epoxide 34. In a 50 mL r.b. flask was dissolved the pyrrolidine
adduct 23h (500 mg, 1.04 mmol) in acetonitrile (10.0 mL) and a
0.0004 M solution of Na2•EDTA (5.20 mL, 0.002 mmol) was
added. The reaction mixture was cooled to 0 °C and trifluoacetone
(2.00 mL, 2.51 g, 22.4 mmol) was added. Anhydrous NaHCO3 (1.36
g, 16.2 mmol) and Oxone (6.41 g, 10.4 mmol) were mixed together
and added by small portions over 1.4 h keeping the temperature at
0 °C. The reaction mixture was then stirred at 0 °C for 1.5 h, at
RT for 17 h and then diluted with H2O. The aqueous phase was
extracted with 3 portions of DCM and the combined organic extracts
were washed once with brine, dried over anhydrous MgSO4, filtered
and concentrated under reduced pressure to give a yellow oil. The
crude product was purified on silica gel with 1:4/AcOEt:Hexanes
1
as eluant to afford a colorless oil (460 mg, 89%). Rotamer 1: H
NMR (CDCl3, 300 MHz): δ 7.65 (dd, 4H, J ) 7.4, 1.9 Hz),
7.45-7.35 (m, 6H), 4.05 (t, 1H, J ) 5.8 Hz), 3.88 (d, 1H, J )
13.2 Hz), 3.66 (dt, 2H, J ) 6.1, 2.2 Hz), 3.58 (d, 1H, J ) 3.3 Hz),
3.43 (d, 1H, J ) 3.3 Hz), 3.22 (d, 1H, J ) 3.3 Hz), 1.60-1.43 (m,
6H), 1.42 (s, 9H), 1.04 (s, 9H). 13C NMR (CDCl3, 75.5 MHz): δ
(ppm) 155.1 (s), 135.6 (d), 133.9 (s), 129.6 (d), 127.6 (d), 63.6 (t),
58.5 (d), 57.8 (d), 54.9, 47.1 (t), 33.1 (s), 32.6 (t), 30.9 (t), 28.4
(for [MH]+). [R]20 ) + 3.5 (c ) 1.25, CHCl3).
D
Amide 27i. 3-Butenoic acid (0.83 mL, 9.74 mmol) was added
to a solution of the amine 16i (3.49 g, 8.83 mmol), N,N′-
dicyclohexylcarbodiimide (2.00 g, 9.69 mmol) and 4-(dimethy-
lamino)pyridine (184 mg, 1.51 mmol) in dichloromethane (52 mL)
at 0 °C. The mixture was stirred 2 h at 0 °C before the addition of
an additional quantity of 3-butenoic acid (0.3 mL, 3.58 mmol). The
mixture was stirred at 0 °C for 45 min and then concentrated under
reduce pressure. The resulting solid was triturated once with
hexanes. The solid was filtered and the filtrate was concentrated
under reduced pressure. The crude product was purified by flash
chromatography on silica gel (20% ethyl acetate in hexanes) to
1
(q), 26.9 (q), 21.9 (t), 19.2 (s). Rotamer 2: H NMR (CDCl3, 300
MHz): δ 7.65 (dd, 4H, J ) 7.4, 1.9 Hz), 7.45-7.35 (m, 6H), 3.95
(t, 1H, J ) 5.8 Hz), 3.78 (d, 1H, J ) 12.7 Hz), 3.66 (dt, 2H, J )
6.1, 2.2 Hz), 3.56 (d, 1H, J ) 3.3 Hz), 3.41 (d, 1H, J ) 3.3 Hz),
3.20 (d, 1H, J ) 3.3 Hz), 1.60-1.43 (m, 6H), 1.42 (s, 9H), 1.04
(s, 9H). 13C NMR (CDCl3, 75.5 MHz): δ (ppm) 155.1 (s), 135.6
(d), 133.9 (s), 129.6 (d), 127.6 (d), 63.6 (t), 58.1 (d), 57.5 (d), 54.3
(d), 46.5 (t), 33.1 (s), 32.6 (t), 30.2 (t), 28.4 (q), 26.9 (q), 21.9 (t),
19.2 (s). IR (neat, cm-1): 3069, 3049, 2933, 2860, 1697, 1427, 1389,
1
yield pure amide 27i (4.03 g, 99%) as a colorless oil. H NMR
(300 MHz, CDCl3) δ (ppm) 5.88 (ddt, 1H, J ) 16.9, 10.1, 7.2
6248 J. Org. Chem. Vol. 73, No. 16, 2008

Rearrangement of Chiral Allylic Azides
1173, 1113. LRMS (m/z, (relative intensity)): 513 ([MNH4]+, 14),
2.91-2.85 (m, 1H), 2.80 (dd, 1H, J ) 12.1, 2.8 Hz), 1.83-1.77
(m, 2H), 1.53-1.40 (m, 3H), 1.38-1.29 (m, 2H), 1.16 (dt, 1H, J
) 12.5, 3.9 Hz). IR (neat) V (cm-1): 3625-3105, 3032, 2926, 2851,
2805, 1443, 1175, 1139. LRMS (m/z, relative intensity): 139 ([M]+,
39), 120 (20), 83 (88), 55 (100), 41 (41). HRMS calcd for C8H13NO:
496 ((MH)+, 100), 440 (90), 396 (70), 362 (48), 304 (33). HRMS
calcd for C29H42NO4Si [MH]+: 496.2883, found 496.2873. [R]20
) +27.6 (c ) 0.99, CHCl3).
D
In a 25 mL r.b. flask was dissolved the above epoxide (440 mg,
0.888 mmol) in anhydrous THF (8.8 mL). This solution was cooled
to 0 °C and TBAF (1.0 M in THF, 0.976 mL, 0.976 mmol) was
added. The reaction mixture was stirred at RT for 4.5 h and
quenched with H2O and 1N aqueous solution of HCl. The aqueous
phase was extracted with 3 portions of Et2O and the combined
organic extracts were washed once with brine, dried over anhydrous
MgSO4, filtered and concentrated under reduced pressure to give a
yellow oil. The crude product was purified on silica gel with 2:3
to 7:3/AcOEt:Hexanes as eluant to afford a colorless oil (198 mg,
139.0997, found 139.1002. [R]20 ) +16.3 (c ) 0.95, MeOH).
D
Lentiginosine (+)-30. Indolozidine 35 (130 mg, 0.934 mmol)
was dissolved in a mixture of dioxane (1.0 mL) and a 10% mixture
of H2SO4 (1.0 mL). The reaction mixture was refluxed for 6 h, and
then cooled to RT and directly purified by silical gel column flash
chromatography eluting with a mixture of methanol, dichlo-
romethane, and 30% NH4OH (15:84: 1). A white solid (104 mg,
71%) was obtained. m.p. 98-100 °C.1H NMR (300 MHz, D2O):
δ (ppm) 4.06 (ddd, 1H, J ) 7.7, 3.9, 1.7 Hz), 3.64 (dd, 1H, J )
8.8, 3.9 Hz), 2.94 (br d, 1H, J ) 11.0 Hz), 2.83 (dd, 1H, J ) 11.0,
1.7 Hz), 2.62 (dd, 1H, J ) 11.0, 7.7 Hz), 2.05 (td, 1H, J ) 11.6,
3.3 Hz), 1.99-1.91 (m, 2H), 1.82-1.79 (m, 1H), 1.63 (br d, 1H,
J ) 13.2 Hz), 1.52-1.37 (m, 1H), 1.32-1.17 (m, 2H). IR (NaCl/
CHCl3) V (cm-1): 3529-3514, 3021, 3012, 2942, 2857, 2806, 1443,
1212, 1140. LRMS (m/z, relative intensity): 158 ([MH]+, 100), 140
(10), 97 (44). HRMS calcd for C8H16NO2 [MH]+: 158.1181, found
1
87%). Rotamer 1: H NMR (CDCl3, 300 MHz): δ (ppm) 4.08 (t,
1H, J ) 6.1 Hz), 3.88 (d, 1H, J ) 13.2 Hz), 3.65 (t, 2H, J ) 6.1
Hz), 3.61 (d, 1H, J ) 2.7 Hz), 3.46 (s, 1H), 3.24 (d, 1H, J ) 13.2
Hz), 1.72 (br s, 1H), 1.68-1.47 (m, 6H), 1.43 (s, 9H). Rotamer 2:
1H NMR (CDCl3, 300 MHz): δ (ppm) 3.98 (t, 1H, J ) 6.1 Hz),
3.79 (d, 1H, J ) 13.2 Hz), 3.65 (t, 2H, J ) 6.1 Hz), 3.58 (d, 1H,
J ) 2.7 Hz), 3.46 (s, 1H), 3.24 (d, 1H, J ) 13.2 Hz), 1.72 (br s,
1H), 1.68-1.47 (m, 6H), 1.43 (s, 9H). IR (neat, cm-1): 3728-3026,
2976, 2944, 2930, 2870, 1682, 1424, 1392, 1174, 1119. LRMS
(m/z, (relative intensity)): 258 ([MH]+, 78), 202 (100), 158 (100),
140 (37), 122 (15), 112 (30), 84 (50). HRMS calcd for C13H23NO4:
158.1177. [R]20 ) +2.4 (c ) 0.41, MeOH).
D
Iodide 38. Triethylamine (0.55 mL, 3.95 mmol) and methane-
sulfonyl chloride (0.25 mL, 3.23 mmol) were added to a solution
of the alcohol 28i (360 mg, 1.97 mmol) in THF (16 mL) at 0 °C.
The mixture was stirred 16 h at RT and quenched with water (10
mL). The aqueous layer was extracted with ethyl acetate (3 × 15
mL), and the combined organic layers were washed once with brine,
dried over anhydrous Na2SO4 and then concentrated under reduced
pressure. The crude product was purified by flash chromatography
on silica gel (from 0 to 5% methanol in ethyl acetate) to yield pure
257.1627, found 257.1617. [R]20 ) +40.6 (c ) 0.70, CHCl3).
D
The alcohol (500 mg, 1.94 mmol) and pyridine (204 µL, 200
mg, 2.53 mmol) were dissolved in dichloromethane (10.0 mL). The
solution was stirred for 10 min, upon which time tosyl chloride
(445 mg, 2.33 mmol) was added. The reaction mixture was stirred
26 h at RT, upon which time water and a saturated aqueous solution
of NaHCO3 were added. The organic phase was separated and
extracted with dichloromethane (3 × 25 mL). The combined organic
layers were washed with brine, dried over anhydrous magnesium
sulfate, and evaporated under reduced pressure. The crude product
was purified on silica gel with 7:3/AcOEt:Hexanes as eluant to
afford epoxide 34 as a colorless oil (683 mg, 85%). Rotamer 1: 1H
NMR (300 MHz, CDCl3): δ (ppm) 7.78 (d, 2H, J ) 7.7 Hz), 7.35
(d, 2H, J ) 7.7 Hz), 4.02 (t, 2H, J ) 6.1 Hz), 4.01 (t, 1H, J ) 6.1
Hz), 3.89 (d, 1H, J ) 13.2 Hz), 3.59 (d, 1H, J ) 3.6 Hz), 3.40 (d,
1H, J ) 3.6 Hz), 3.22 (d, 1H, J ) 9.6 Hz), 2.45 (s, 3H), 1.73-1.64
1
mesylate (496 mg, 97%) as a colorless oil. H NMR (300 MHz,
CDCl3) δ (ppm) 7.57 (s, 1H), 5.72 (dtd, 1H, J ) 9.9, 3.2, 1.5 Hz),
5.63-5.58 (m, 1H), 4.74 (sext, 1H, J ) 6.0 Hz), 4.07-4.02 (m,
1H), 2.96 (s, 3H), 2.85-2.81 (m, 2H), 1.74-1.39 (m, 6H), 1.36
(d, 3H, J ) 6.0 Hz). IR (neat) • (cm-1) 3217, 3046, 2987, 2934,
2863, 1676, 1665, 1464, 1334, 1169, 915. LRMS (m/z, relative
intensity) 261 ([M+], 2), 182 ([M+ - CH3SO2], 5), 166 ([M+
-
CH3SO3], 30), 96 (100). HRMS calcd for C11H19NO4S: 261.1035,
found: 261.1040. [R]20 ) +48.9 (c ) 2.06, CHCl3).
D
The above mesylate (490 mg, 01.88 mmol), sodium iodide (2.86
g, 19.1 mmol), and NaHCO3 (1.63 g, 19.4 mmol) were stirred 18 h
in acetone (8.5 mL), and the mixture was quenched with water (25
mL). The aqueous layer was extracted with ethyl acetate (4 × 25
mL), and the combined organic layers were washed once with brine,
dried over anhydrous Na2SO4, and then concentrated under reduced
pressure. The crude product was purified by flash chromatography
on silica gel (from 0 to 5% methanol in ethyl acetate) to yield pure
iodide 38 (450 mg, 82%) as a colorless oil. 1H NMR Diastereomer
A (300 MHz, CDCl3) δ (ppm) 7.05-7.01 (m, 1H), 5.77 (dtd, 1H,
J ) 9.9, 3.2, 1.5 Hz), 5.69-5.64 (m, 1H), 4.19-4.07 (m, 2H),
2.92-2.88 (m, 2H), 1.90 (d, 3H, J ) 7.2 Hz), 1.87-1.77 (m, 2H),
1
(m, 2H), 1.57-1.36 (m, 6H), 1.42 (s, 9H). Rotamer 2: H NMR
(300 MHz, CDCl3): δ (ppm) 7.78 (d, 2H, J ) 7.7 Hz), 7.35 (d,
2H, J ) 7.7 Hz), 4.02 (t, 2H, J ) 6.1 Hz), 3.95 (t, 1H, J ) 6.1
Hz), 3.78 (d, 1H, J ) 13.2 Hz), 3.58 (d, 1H, J ) 3.6 Hz), 3.38 (d,
1H, J ) 3.6 Hz), 3.18 (d, 1H, J ) 9.6 Hz), 2.45 (s, 3H), 1.73-1.64
(m, 2H), 1.57-1.36 (m, 6H), 1.42 (s, 9H). IR (neat) V (cm-1): 3040,
2972, 2939, 2873, 1699, 1599, 1420, 1389, 1172, 1118. LRMS
(m/z, relative intensity): 429 ([MNH4]+, 4), 412 ([MH]+, 11), 373
(86), 258 (60), 237 (90), 202 (88), 140 (100), 84 (85). HRMS calcd
for C20.5HNO6S (MH)+: 412.1794, found 412.1785. [R]20D ) +34.9
(c ) 0.55, CHCl3).
1
Indolizidine 35. Tosylate 34 (680 mg, 1.65 mmol) was dissovled
in dichloromethane (4.0 mL), and trifluoroacetic acid (0.80 mL)
was added dropwise. The reaction mixture was stirred at RT for
1 h and was then concentrated under reduced pressure. The residue
was coazeotroped with benzene (2 × 5 mL) to remove most of the
water. It was then taken up in dichloromethane (10.2 mL) and
triethylamine (1.40 mL, 1.01 g, 9.91 mmol) was added. The mixture
was stirred at RT for 18 h and was then concentrated under reduced
pressure. The resulting oil was taken up in dioxane (2.0 mL), and
a 5N aqueous solution of NaOH (0.5 mL, 2.45 mmol) was added
to remove the tosyl salts. This mixture was stirred for 1 h at RT
and was then concentrated under reduced pressure. The crude
product was purified on silica gel eluting with a mixture of
methanol, dichloromethane, and a 30% solution of NH4OH (15:
84: 1). Compound 35 was obtained as colorless oil (145 mg, 63%).
1H NMR (300 MHz, CD3OD): δ (ppm) 3.62 (d, 1H, J ) 2.8 Hz),
3.47 (d, 1H, J ) 2.8 Hz), 3.12 (dd, 1H, J ) 12.1, 3.9 Hz),
1.66-1.38 (m, 4H). H NMR Diastereomer B (300 MHz, CDCl3)
δ (ppm) 7.05-7.01 (m, 1H), 5.77 (dtd, 1H, J ) 9.9, 3.2, 1.5 Hz),
5.69-5.64 (m, 1H), 4.19-4.07 (m, 2H), 2.92-2.88 (m, 2H), 1.90
(d, 3H, J ) 6.0 Hz), 1.87-1.77 (m, 2H), 1.66-1.38 (m, 4H). IR
(neat) V (cm-1) 3209, 3049, 2918, 2870, 1677, 1663, 1338. LRMS
(m/z, relative intensity) 293 ([M+], 1), 197 ([M+ - C5H6NO], 5),
166 ([M+ - I], 20), 96 (100). HRMS calcd for C10H16INO:
293.0277, found: 293.0269. [R]20 ) +66.4 (c ) 1.40, CHCl3).
D
Perhydroquinolinone 39. A refluxing solution of iodide 38 (450
mg, 1.54 mmol) in benzene (980 mL) was bubbled with argon for
30 min. Tributyltin hydride (0.54 mL, 2.0 mmol) in benzene (20
mL) and AIBN (77 mg, 0.47 mmol) were added and the mixture
was refluxed 12 h. The tributyltin hydride solution was added with
a seringe pump in 4 h after an initial addition of 3 mL. AIBN was
added in seven portions at each hour. The solution was concentrated
under reduced pressure and the crude product was purified by flash
chromatography on silica gel (from 50% hexanes in ethyl acetate
J. Org. Chem. Vol. 73, No. 16, 2008 6249

Lauzon et al.
to 100% ethyl acetate to 10% methanol in ethyl acetate) to yield
the known perhydroquinolinone 39 (64 mg, 25%) as a white solid.30
m.p. 108-110 °C. 1H NMR (300 MHz, CDCl3) δ (ppm) 5.40-5.30
(m, 1H), 3.63 (q, 1H, J ) 3.3 Hz), 2.32-2.28 (m, 2H), 2.10-2.01
(m, 1H), 1.73-1.41 (m, 9H), 0.93 (d, 3H, J ) 6.6 Hz). IR (neat)
V (cm-1) 3390, 2940, 2869, 1646, 1461, 1319. LRMS (m/z, relative
intensity) 167 (M+, 20), 152 (M+-CH3, 5), 124 (M+-CHNO, 100).
The imine (30 mg, 0.16 mmol) and 5% palladium on carbon
were stirred at RT in methanol (5 mL) under hydrogen atmosphere
for 40 h at RT. The mixture was filtered on Celite and gaseous
HCl was bubbled in the resulting solution. The mixture was
concentrated under reduced pressure and the crude solid was
recrystallized in isopropyl alcohol-diethyl ether to yield pure HCl
salt of pumiliotoxin C ((+)-31·HCl) (22 mg, 59% for 3 steps) as
1
HRMS calcd for C10H17NO: 167.1310, found: 167.1314. [R]20
+33.8 (c ) 0.08, CHCl3).
)
white crystals. m.p. 228-230 °C. H NMR (300 MHz, CDCl3) δ
D
(ppm) 9.57 (s, 1H), 8.37 (s, 1H), 3.33-3.29 (m, 1H), 2.99-2.95
(m, 1H), 2.51-2.35 (m, 2H), 2.17-2.07 (m, 4H), 1.86 (d, 1H, J )
12.7 Hz), 1.78 (d, 1H, J ) 13.2 Hz), 1.63-1.27 (m, 6H), 1.25-1.20
(m, 1H), 1.01-0.98 (m, 1H), 0.92 (t, 3H, J ) 6.3 Hz), 0.89 (d,
3H, J ) 5.5 Hz). LRMS (m/z, relative intensity) 195 ([M+], 3),
194 ([M+ - H], 3), 152 ([M+ - C3H7], 100). HRMS calcd for
Pumiliotoxin C, Hydrochloride Salt ((+)-31·HCl). The per-
hydroquinolinone 39 (31 mg, 0.19 mmol) in d2-dichloromethane
(1 mL) was added to a solution of trimethyloxonium tetrafluorobo-
rate (53 mg, 0.36 mmol) and diisopropylethylamine (1 drop) in
d2-dichloromethane (1 mL). The mixture was stirred 1 h at RT and
quenched with a saturated NaHCO3 aqueous solution (2 mL) at 0
°C. The organic layer was dried over anhydrous Na2SO4 and then
concentrated under reduced pressure to yield the crude imino ether
(49 mg) used quickly to avoid degradation.
A solution of propylmagnesium bromide in benzene (0.9 M, 1
mL, 0.9 mmol) was added to a solution of imino ether (49 mg,
0.36 mmol) in benzene (3 mL). The mixture was stirred 4 h in a
sealed vial at 85 °C. The reaction mixture was cooled down to 0
°C, dropped in diethyl ether (5 mL), and quenched with a saturated
NaHCO3 aqueous solution (5 mL). The aqueous layer was extracted
with diethyl ether (2 × 5 mL), and the combined organic layers
were dried over anhydrous Na2SO4 and then concentrated under
reduced pressure. The imine (30 mg) was used quickly to avoid
degradation.
C13H25N: 195.1987, found: 195.1982. [R]20 ) +12.9 (c ) 0.41,
D
CHCl3).
Acknowledgment. We thank the Natural Sciences and
Engineering Council of Canada, AstraZeneca (Canada) Inc.,
Merck-Frosst Canada Ltd, and the Universite´ de Sherbrooke
for their financial support.
Supporting Information Available: Experimental proce-
dures and 1H NMR spectra for all new compounds not included
in the Experimental Section of the article. Crystallographic data
for compound 18n. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO800817P
6250 J. Org. Chem. Vol. 73, No. 16, 2008