First total synthesis of ent-gelsedine via a novel iodide-promoted allene N-acyliminium ion cyclization

Article abstract of DOI:10.1021/jo001119t

The first total synthesis of the oxindole alkaloid gelsedine (1) starting from (S)-malic acid is described. The key step is a novel iodide-promoted intramolecular reaction of an allene with an N-acyliminium ion intermediate which provided in a single step the bicyclic vinyl iodide 11. Other important steps are the highly stereoselective Pd-catalyzed Heck cyclization of N-methylanilide 23a which led to the desired spiro-oxindole 24a, the fully regioselective intramolecular oxymercuration of 25a to the desired cyclic ether, and the remarkable oxindole N-demethylation of 29 via a radical mechanism by using dibenzoyl peroxide. The total synthesis was concluded by the stereoselective introduction of the ethyl group from the bis-Boc compound. 41 followed by methoxylation of the oxindole nitrogen. This total synthesis leads to the unnatural (+)-enantiomer of gelsedine in 21 steps and 0.10% overall yield.

Full text of DOI:10.1021/jo001119t

J . Org. Chem. 2000, 65, 8317-8325
8317
F ir st Tota l Syn th esis of en t-Gelsed in e via a Novel Iod id e-P r om oted
Allen e N-Acylim in iu m Ion Cycliza tion
Winfred G. Beyersbergen van Henegouwen, Rutger M. Fieseler, Floris P. J . T. Rutjes, and
Henk Hiemstra*
Laboratory of Organic Chemistry, Institute of Molecular Chemistry, University of Amsterdam,
Nieuwe Achtergracht 129, 1018 WS Amsterdam, The Netherlands
henkh@org.chem.uva.nl
Received J uly 24, 2000
The first total synthesis of the oxindole alkaloid gelsedine (1) starting from (S)-malic acid is
described. The key step is a novel iodide-promoted intramolecular reaction of an allene with an
N-acyliminium ion intermediate which provided in a single step the bicyclic vinyl iodide 11. Other
important steps are the highly stereoselective Pd-catalyzed Heck cyclization of N-methylanilide
23a which led to the desired spiro-oxindole 24a , the fully regioselective intramolecular oxymer-
curation of 25a to the desired cyclic ether, and the remarkable oxindole N-demethylation of 29 via
a radical mechanism by using dibenzoyl peroxide. The total synthesis was concluded by the
stereoselective introduction of the ethyl group from the bis-Boc compound 41 followed by
methoxylation of the oxindole nitrogen. This total synthesis leads to the unnatural (+)-enantiomer
of gelsedine in 21 steps and 0.10% overall yield.
In tr od u ction
present full details of the first total synthesis of enan-
tiopure ent-gelsedine.6
The plant genus Gelsemium (Loganiaceae) consists of
three species Gelsemium elegans Benth., G. sempervirens
Ait., and G. rankinii Small, all of which contain a number
of indole and oxindole alkaloids with remarkably diverse
and complex structures.1 The Gelsemium alkaloids can
be classified in six groups based on their skeletal types,
of which gelsedine (1), gelsemine (2), koumidine (3),
koumine (4), humentenine (5), and gelselegine (6) are
representative members. Although specific biological
activity of most of these alkaloids is poorly documented,
extracts from Gelsemium species have a rich medicinal
history, particularly in China.1 Despite extensive syn-
thetic research efforts for many years, successful de novo
syntheses of Gelsemium alkaloids are only of relatively
recent date. In 1989, Magnus et al. reported the total
synthesis of (+)-koumine (4).2 Well into the 1990s, we3
and others4 have published total syntheses of racemic
gelsemine (2). Very recently, we completed the synthesis
of gelsemine in enantiopure form.5 In this article, we
* Corresponding author. Phone: +31 20 5255941; fax: +31 20
5255670.
(1) For reviews on Gelsemium alkaloids, see: (a) Saxton, J . E. In
The Alkaloids; Manske, R. H. F., Ed.; Academic Press: New York, 1965;
Vol. 8, pp 93-117. (b) Liu, Z.-J .; Lu, R.-R. In The Alkaloids; Brossi,
A., Ed.; Academic Press: New York, 1988; Vol. 33, pp 83-140. (c)
Takayama, H.; Sakai, S. In Studies in Natural Products Chemistry;
Atta-ur-Rahman, Ed.; Elsevier: Amsterdam, 1995; Vol. 15, pp 465-
518.
Gelsedine was first isolated from Carolina jasmine
(Gelsemium sempervirens) in 1953 by Schwarz and
Marion,7 but also occurs in G. elegans.8 One year after
Przybylska and Marion elucidated the structure of
gelsemicine (7) on the basis of its X-ray structure in
1961,9 Wenkert proposed the structure for gelsedine
(2) Magnus, P.; Mugrage, B.; DeLuca, M. R.; Cain, G. A. J . Am.
Chem. Soc. 1990, 112, 5220-5230.
(3) Newcombe, N. J .; Fang, Y.; Vijn, R. J .; Hiemstra, H.; Speckamp,
W. N. J . Chem. Soc., Chem. Commun. 1994, 767-768.
(4) (a) Dutton, J . K.; Steel, R. W.; Tasker, A. S.; Popsavin, V.;
J ohnson, A. P. J . Chem. Soc., Chem. Commun. 1994, 765-766. (b)
Kuzmich, D.; Wu, S. C.; Ha, D.-C.; Lee, C.-S.; Ramesh, S.; Atarashi,
S.; Choi, J .-K.; Hart, D. J . J . Am. Chem. Soc. 1994, 116, 6943-6944.
(c) Fukuyama, T.; Gang, L. J . Am. Chem. Soc. 1996, 118, 7426-7427.
(d) Atarashi, S.; Choi, J .-K.; Ha, D.-C.; Hart, D. J .; Kuzmich, D.; Lee,
C.-S.; Ramesh, S.; Wu, S. C. J . Am. Chem. Soc. 1997, 119, 6226-6241.
(e) Madin, A.; O’Donnell, C. J .; Oh, T.; Old, D. W.; Overman, L. E.;
Sharp, M. J . Angew. Chem., Int. Ed. 1999, 38, 2934-2936.
(5) (a) Dijkink, J .; Cintrat, J .-C.; Speckamp, W. N.; Hiemstra, H.
Tetrahedron Lett. 1999, 40, 5919-5922. (b) Manuscript in preparation.
(6) For a preliminary communication of this work, see: Beyersber-
gen van Henegouwen, W. G.; Fieseler, R. M.; Rutjes, F. P. J . T.;
Hiemstra, H. Angew. Chem., Int. Ed. 1999, 38, 2214-2217.
(7) Schwarz, H.; Marion, L. Can. J . Chem. 1953, 31, 958-975.
(8) J in, H. L.; Xu, R. S. Acta Chim. Sinica 1982, 40, 1129-1135.
10.1021/jo001119t CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/27/2000

8318 J . Org. Chem., Vol. 65, No. 24, 2000
Beyersbergen van Henegouwen et al.
Sch em e 1
lead to the oxindole 8 after hydroboration of the bridge
methylene group. The stereochemical outcome of the
spirocyclization was an open question, but Overman has
shown that this outcome can be drastically influenced
by varying the conditions.24b The anilide 9 was thought
to arise from a Pd-catalyzed aminocarbonylation of a
suitably functionalized vinyl triflate or vinyl iodide (viz.
10 and 11, respectively) after introduction of a one-carbon
fragment onto the bridge. These two Pd-catalyzed steps
have also served extremely well in our gelsemine syn-
thesis.3 The bicyclic lactams 10 and 11 have suitable
skeletons to be accessed via N-acyliminium ion cycliza-
tion chemisty.17 In a previous paper, we showed that 12
is the ideal cyclization precursor.16 Inexpensive (S)-malic
acid was chosen as the starting material, despite the fact
that this will eventually lead to the unnatural isomer of
gelsedine. The early stages of the synthesis bear resem-
blance to our total synthesis of the indole alkaloid
peduncularine.18
based on NMR analogy with gelsemicine.10 Although the
biological activity of gelsedine has not been determined,
gelsemicine was shown to be the most toxic component
in G. sempervirens.1b
Resu lts
In the previous paper, we reported an efficient five-
step synthesis of ethoxylactam 12 from (S)-malic acid and
its N-acyliminium ion cyclization in neat formic acid to
bicyclic ketone 14 in 79% yield (Scheme 2).16 In nine
steps, ketone 14 was further transformed into the ethyl-
substituted ketone 15.16 We regarded 15 a suitable model
system for the parent skeleton to study the introduction
of the oxindole moiety (Scheme 2).
As a minor constituent of Carolina jasmine, the syn-
thesis of gelsedine has not received as much attention
as the parent alkaloid gelsemine. Until sofar, three
groups have reported studies toward a total synthesis of
gelsedine.11,12,13 In 1979, a synthesis of the skeleton of
gelsedine lacking the ethyl and oxindole moieties was
published by Baldwin and Doll,11 and a synthesis of a
comparable structure using a different approach was
reported by Hamer in 1990.12 That same year, Kende et
al. synthesized a more advanced compound, which had
the oxindole moiety in place, albeit in the wrong stere-
ochemical sense.13 In 1994, Takayama et al. reported a
semisynthesis of gelsedine starting from the natural
product koumidine (3), the latter also being an interme-
diate in Magnus’s synthesis of koumine (4).14 This
synthesis was based on the biogenetic pathway proposed
for gelsedine and other Gelsemium alkaloids.1b,c,15
To perform the Heck cyclization in the appropriate
way, vinyl triflate 17 had to be prepared regioselectively.
Therefore, ketone 15 was subjected to kinetic deproto-
nation conditions (LiHMDS, -78 °C) and then treated
1
with N-phenyltriflimide.19 The H NMR spectrum of the
crude product showed two signals in the double bond
region in a ratio of 4:1. The major component was isolated
in 15% yield and was identified by a COSY experiment
as being the undesired regioisomer 16. The minor com-
ponent could not be isolated and characterized due to
(17) (a) Klaver, W. J .; Hiemstra, H.; Speckamp, W. N. Tetrahedron
Lett. 1987, 28, 1581-1584. (b) Klaver, W. J .; Moolenaar, M. J .;
Hiemstra, H.; Speckamp, W. N. Tetrahedron 1988, 44, 3805-3818.
(18) Klaver, W. J .; Hiemstra, H.; Speckamp, W. N. J . Am. Chem.
Soc. 1989, 111, 2588-2595.
(19) For a review on vinyl and aryl triflates, see: Ritter, K. Synthesis
1993, 735-762.
(20) (a) Overman, L. E.; Sharp, M. J . J . Am. Chem. Soc. 1988, 110,
612-614. (b) Lin, N.-H.; Overman, L. E.; Rabinowitz, M. H.; Robinson,
L. A.; Sharp, M. J .; Zablocki, J . J . Am. Chem. Soc. 1996, 118, 9062-
9072. (c) Caderas, C.; Lett, R.; Overman, L. E.; Rabinowitz, M. H.;
Robinson, L. A.; Sharp, M. J .; Zablocki, J . J . Am. Chem. Soc. 1996,
118, 9073-9082.
(21) (a) Nossin, P. M. M.; Speckamp, W. N. Tetrahedron Lett. 1981,
22, 3289-3292. (b) Nossin, P. M. M.; Hamersma, J . A. M.; Speckamp,
W. N. Tetrahedron Lett. 1982, 23, 3807-3810. (c) Nossin, P. M. M.,
Ph.D. Thesis, University of Amsterdam (The Netherlands), 1983. For
intermolecular reactions of N-acyliminium ions with allenes, see: (d)
Danheiser, R. L.; Kwasigroch, C. A.; Tsai, Y.-M. J . Am. Chem. Soc.
1985, 107, 7233-7235. (e) Prasad, J . S.; Liebeskind, L. S. Tetrahedron
Lett. 1988, 29, 4257-4260. (f) Karstens, W. F. J .; Rutjes, F. P. J . T.;
Hiemstra, H. Tetrahedron Lett. 1997, 38, 6275-6278.
(22) Brosius, A. D.; Overman, L. E. J . Org. Chem. 1997, 62, 440-
441.
Syn th etic P la n
Our retrosynthetic approach is depicted in Scheme 1.
We envisioned that the introduction of the ethyl sub-
stituent into 8 should be possible via methodology
developed earlier by our group on a model system.16 The
tetrahydropyran ring was expected to arise via mercuric
ion-induced ring closure of the hydroxyl group onto the
double bond for which we had collected precedence in our
gelsemine synthesis.3 Suitable protection of anilide 9 and
subsequent Pd-catalyzed Heck spirocyclization should
(9) Przybylska, M.; Marion, L. Can. J . Chem. 1961, 39, 2124-2127.
(10) Wenkert, E.; Orr, J . C.; Garratt, S.; Hansen, J . H.; Wickberg,
B.; Leicht, C. L. J . Org. Chem. 1962, 27, 4123-4126.
(11) Baldwin, S. W.; Doll, R. J . Tetrahedron Lett. 1979, 20, 3275-
3278.
(12) Hamer, N. K. J . Chem. Soc., Chem. Commun. 1990, 102-103.
(13) (a) Kende, A. S.; Luzzio, M. J .; Mendoza, J . S. J . Org. Chem.
1990, 55, 918-924. (b) Luzzio; M. J ., Ph.D. Thesis, University of
Rochester, 1987.
(14) Takayama, H.; Tominaga, Y.; Kitajama, M.; Aimi, N.; Sakai,
S. J . Org. Chem. 1994, 59, 4381-4385.
(15) Takayama, H.; Sakai, S. In The Alkaloids; Cordell, G. A., Ed.;
Academic Press: New York, 1998; Vol. 50, pp 415-452.
(16) Beyersbergen van Henegouwen, W. G.; Hiemstra, H. J . Org.
Chem. 1997, 62, 8862-8867.
(23) Cacchi, S.; Morera, E.; Ortar, G. Tetrahedron Lett. 1985, 26,
1109-1112.
(24) (a) Abelman, M. M.; Oh, T.; Overman, L. E. J . Org. Chem. 1987,
52, 4130-4133. (b) Madin, A.; Overman, L. E. Tetrahedron Lett. 1992,
33, 4859-4862. (c) Ashimori, A.; Bachand, B.; Overman, L. E.; Poon,
D. J . J . Am. Chem. Soc. 1998, 120, 6477-6487. (d) Ashimori, A.;
Bachand, B.; Calter, M. A.; Govek, S. P.; Overman, L. E.; Poon, D. J .
J . Am. Chem. Soc. 1998, 120, 6488-6499. For a review, see: (e) Link,
J . T.; Overman, L. E. Chemtech 1998 (August), 19-26.

Total Synthesis of ent-Gelsedine
J . Org. Chem., Vol. 65, No. 24, 2000 8319
Sch em e 2
Sch em e 4
Sch em e 3
tively, when allene 12 was dissolved in formic acid in the
presence of a large excess of sodium iodide and heated
to 85 °C for 18 h we were delighted to find vinyl iodide
11 as the main product in 42% yield (Scheme 4). We were
unable to prevent competitive attack of formic acid
providing ketone 13 as a byproduct in 34% yield. Lower-
ing the temperature to as low as 40 °C or temperatures
between 40 °C and 80 °C resulted in longer reaction times
without improvement of the ratio. In fact, at these
temperatures the reaction did not go to completion within
18 h. Varying the concentration and nature of the iodide
source had no beneficial effect either. Probably formic
acid-mediated formation and solvolysis of the intermedi-
ate N-acyliminium ion 20 is crucial for the cyclization to
occur. We were nevertheless pleased with this unique
reaction, which produced vinyl iodide 11 in only one step,
suitably functionalized for elaboration to the oxindole
moiety. Vinyl iodide 11 was readily separated from 13
by flash chromatography so that sufficient quantities of
11 could be assembled for the further study of the total
synthesis of gelsedine.
purification problems. Thermodynamic conditions (KH,
rt, and treatment with N-phenyltriflimide or treatment
with sterically hindered 2,6-di-tert-butyl-4-methylpyri-
dine and Tf2O, rt) did not effectively yield any vinyl
triflate.
As it appeared difficult to create the desired regioiso-
mer at this stage of the synthesis, the formation of the
vinyl triflate at an earlier stage of the synthesis was
studied. Ketone 14 (which does not yet contain the ethyl
group) was protected as its tert-butyldimethylsilyl ether
18 and subjected to kinetic deprotonation conditions as
well. The 1H NMR spectrum of the crude product showed
only one product, which after isolation was identified by
a COSY experiment to be the undesired regioisomer 19.
In addition, with similar but differently functionalized
bicyclic ketones we were also unable to form the desired
regioisomeric vinyl triflate in acceptable yield.
Because regioselective functionalization of the seven-
membered ring after N-acyliminium cyclization appeared
unpractical, we considered methods to regioselectively
introduce a vinyl triflate (or its synthetic equivalent) in
the cyclization step. However, few methods exist to
introduce such functionality via an N-acyliminium ion
reaction. There is one example by Overman where an
iodide-terminated iminium ion cyclization of an alkyne
leads to an exocyclic vinyl iodide.20 Because a similar
reaction in our case would lead to the undesired regio-
isomer, we decided to investigate a similar cyclization of
the allene 12. N-Acyliminium ion cyclizations of allenes
are scarce in the literature,21 and iodide-promoted vari-
ants are unknown to the best of our knowledge. Unfor-
tunately, subjection of 12 to the conditions used by
Overman, i.e. CSA, NaI, H2O, 100 °C or TfOH, n-Bu4NI,
CHCl3, produced no vinyl iodide or ketone.20,22 Alterna-
The construction of the spiro-oxindole moiety onto 11
commenced by applying a Pd-catalyzed aminocarbony-
lation process23 to give anilide 21 after subsequent
deformylation. Because it is known that the stereochem-
ical course of the Heck spiro-cyclization is dependent
upon steric influences, we envisioned that a sterically
nondemanding sp2-hybridized bridge carbon would be
ideal. The approach of the arylpalladium complex from
the exo direction would then be sterically the least
encumbered. Furthermore, by using an inert methylene
substituent at the bridge, the use of protecting groups
in this region of the molecule was avoided. Therefore,
alcohol 21 was first oxidized with PCC to ketone 22 in
66% yield (Scheme 5). A Swern oxidation, which was
successfully applied on our model system,16 did not
proceed satisfactorily in this case. The exocyclic double
bond was then introduced by a Wittig reaction to provide
9 in 83% yield.
Overman discovered that Heck cyclizations of N-
unsubstituted anilides are ineffective suggesting N-
substitution of the anilide 9 is necessary.24a Initially,
anilide 9 was N-functionalized with several groups, i.e.
methyl, benzyl, methoxymethyl, and acetyl in good yields.
Heck cyclizations were carried out (Pd(PPh3)4, Et3N,
MeCN, 100-120 °C, sealed tube) with each of these
precursors to give products 23a -d , respectively.24 The
yields in the case of the methyl, benzyl, and methoxy-
methyl substituents were high, while with the acetyl

8320 J . Org. Chem., Vol. 65, No. 24, 2000
Beyersbergen van Henegouwen et al.
Sch em e 5
Sch em e 6
with Hg(OTf)2 in MeNO2 for 2 days followed by addition
of saturated aqueous NaCl gave chloromercurial 26b in
58% yield together with 30% of the starting material.
The low yield, long reaction time, and the in situ
preparation of Hg(OTf)2 led us to adopt the procedure
employing commercially available Hg(O2CCF3)2. Treat-
ment of 25a with Hg(O2CCF3)2 in THF followed by
addition of saturated aqueous NaCl gave the chloromer-
curial 26a (Scheme 6).27 In the case of 25d , application
of an identical procedure provided the required product
26d . However, the yield could not be determined, because
26d was too polar to be purified by column chromatog-
raphy. In all cases the six-membered ring ether was
produced as a single product. MM2 force field calculations
(CSChem3D Pro version 4.0) on model systems 27 and
28 showed that the steric energy of the six-membered
ring ether is significantly lower (41.85 kcal/mol) than the
undesired five-membered ring ether (45.48 kcal/mol,
Figure 1). This difference is probably reflected in the
transition states of the cyclic ether formation.
Reduction of the chloromercurial proved to be prob-
lematic. For the reduction of 26a we first examined a
procedure utilizing a two-phase system of aqueous NaOH
and CH2Cl2.3,28 After addition of the phase-transfer
catalyst benzyltriethylammonium chloride and an excess
of NaBH4, metallic mercury precipitated, and the desired
product 29 along with alcohol 30 was obtained in 35%
and 37% yield, respectively. The stereochemistry of
alcohol 30 was tentatively assigned based on the absence
of a NOE effect between H-2 and H-10. This product
probably arose from reaction of the intermediate alkyl
radical with molecular oxygen present in the solution.29
A similar problem was encountered in our total synthesis
group it was somewhat lower, potentially due to in situ
deacetylation.
Clearly, the nature of the N-substituent does not
influence the outcome of the reaction. In principle all of
these groups can be used. The stereochemistry in the case
of R ) Me was unambiguously proven in a later stage of
the synthesis. By comparison of the 1H NMR and 13C
NMR spectra, we concluded that 24b-d possess the same
stereochemistry. In all cases less than 5% of the spiro-
isomer was observed in the 1H NMR spectrum of the
crude product.
The bridge carbon was now further elaborated by
hydroboration with sterically hindered dicyclohexylbo-
rane25 to discriminate between the exo- and endocyclic
double bond. In the case of 24a -c, hydroxymethylene
products 25a -c were obtained in comparably high yields.
Initially, commercially available 9-BBN was used for
hydroboration of 24a , but the resulting byproduct after
workup (1,5-cyclooctanediol) could not be separated by
column chromatography from product 25a due to com-
parable Rf-values. In all cases, the desired stereoisomer
was formed exclusively, as a result of attack of the borane
from the less sterically encumbered exo face. Hydrobo-
ration of 24d with dicyclohexylborane afforded deacety-
lated alcohol 25d , which was probably formed by attack
of hydroxide on the acetyl carbonyl during workup. The
synthesis was eventually continued with the methyl as
N-substituent, because in the course of our work we found
out that it can be selectively removed, as will be detailed
later.
To close the tetrahydropyran ring we initially tried to
activate the double bond with iodine, but no reaction was
observed. We then relied upon a procedure from our total
synthesis of gelsemine.3,26 For example, treatment of 25b
(26) Nishisawa, M.; Takenaka, H.; Hayashi, Y. J . Org. Chem. 1986,
51, 806-813.
(27) Sutherlin, D. P.; Armstrong, R. W. J . Org. Chem. 1997, 62,
5267-5283.
(28) Benhamou, M. C.; Etemad-Moghadam, G.; Speziale, V.; Lattes,
A. Synthesis 1979, 891-892.
(25) Brown, H. C.; Mandal, A. K.; Kulkarni, S. U. J . Org. Chem.
1977, 42, 1392-1398.
(29) Hill, C. L.; Whitesides, G. M. J . Am. Chem. Soc. 1974, 96, 870-
876.

Total Synthesis of ent-Gelsedine
J . Org. Chem., Vol. 65, No. 24, 2000 8321
Sch em e 7
Sch em e 8
F igu r e 1. Optimized geometries (CSChem3D Pro) of the
theoretical six- and five-membered ring ethers 27 and 28,
indicating that the former is more stable than the latter by
ca. 3.5 kcal/mol (see text).
of gelsemine when the substrate concentration in the
organic phase of the reaction mixture was too low. How-
ever, at a concentration of 0.4 M the desired product was
selectively obtained in 88% yield.3 Unfortunately, the use
of higher concentration was not successful in the present
case. Product 30 could in principle be elaborated toward
14-hydroxygelsedine (31), another Gelsemium alkaloid.30
Finally, we relied upon a procedure initially developed
by Whitesides and optimized by Evans for reduction of a
tetrahydropyran chloromercurial in the total synthesis
of the polyether antibiotic X-206.31 Thus, treatment of
26a with n-Bu3SnH and AIBN at 55 °C in toluene
provided exclusively the desired product 29 in 80% yield.
The concentration of chloromercurial 26a is crucial and
should not be lower than 0.5 M. In this way, the
concentration of n-Bu3SnH remains sufficiently high to
react directly with the intermediate alkyl radical.
Sch em e 9
methyl group in 33 (Scheme 8). The resulting intermedi-
ate radical can then react with benzoyl peroxide to form
36. This extraordinary reaction is probably viable due
to the stability of the intermediate radical 35.35 Subse-
quent treatment with NH3 in methanol was necessary,
because after 18 h still a certain amount of intermediate
37 was present. In the presence of ammonia, 37 reacted
smoothly to provide the deprotected oxindole 34.
Application of these conditions to the more complex
system 29 proceeded remarkably well providing the
deprotected oxindole 38 in 74% yield after treatment with
ammonia to cleave the intermediate benzoate (Scheme
9). It is a remarkable observation that the methylene
hydrogens of the benzyl group do not react at all,
although the benzyl radical might be even better stabi-
lized. The benzyl group was then removed by treatment
with lithium in liquid ammonia. The reaction was best
Having installed the ether bridge, the next stage
consisted of introduction of the ethyl moiety. This re-
quires reductive removal of the benzyl group from the
saturated lactam nitrogen. However, to achieve this using
a dissolving metal reduction, the oxindole ring needs
protection as the anion. Therefore, the oxindole methyl
group should be removed first. We hoped that this might
be accomplished by using a procedure which is known
for removal of an N-methyl group from an indole.32
Initially, we attempted this reaction on model system 33,
which was synthesized via exhaustive methylation of
oxindole 32 (Scheme 7).33 Treatment of 33 with dibenzoyl
peroxide (CH2Cl2, 80 °C, sealed tube) followed by NaOH
in methanol and subsequently NH3 in methanol provided
34 in a remarkable 74% yield.
(33) Other syntheses of compound 34: (a) J ulian, P. L.; Pikl, J .;
Boggess, D. J . Am. Chem. Soc. 1934, 56, 1797-1801. (b) Goehring, R.
R.; Sachdeva, Y. P.; Pisipati, J . S.; Sleevi, M. C.; Wolfe, J . F. J . Am.
Chem. Soc. 1985, 107, 435-443. (c) Clark, A. J .; Davies, D. I.; J ones,
K.; Millbanks, C. J . Chem. Soc., Chem. Commun. 1994, 41-42. (d)
Beckwith, A. L. J .; Storey, J . M. D. J . Chem. Soc., Chem. Commun.
1995, 977-978.
(34) Ghosez, A.; Giese, B.; Zipse, H. In Methoden Der Organischen
Chemie (Houben-Weyl); Regitz, M.; Giese, B., Eds.; Thieme: Stuttgart/
New York, 1989; Vol. E19a, 533-716.
(35) For R-acylamino radicals, see: Hart, D. J .; Tsai, Y.-M. J . Am.
Chem. Soc. 1982, 104, 1430-1432 and ref 4d.
This reaction presumably starts with the formation of
a phenyl radical,34 which abstracts a hydrogen from the
(30) Schun, Y.; Cordell, G. A. J . Nat. Prod. 1985, 48, 788-791.
(31) (a) Whitesides, G. M.; San Filippo, J ., J r. J . Am. Chem. Soc.
1970, 92, 6611-6624. (b) Evans, D. A.; Bender, S. L.; Morris, J . J .
Am. Chem. Soc. 1988, 110, 2506-2526.
(32) Nakatsuka, S.; Asano, O.; Goto, T. Heterocycles 1986, 24, 2791-
2792.

8322 J . Org. Chem., Vol. 65, No. 24, 2000
Beyersbergen van Henegouwen et al.
Sch em e 10
Sch em e 11
molecules was followed.14,36 This involves protection of
the amine in 44 with 2,2,2-trichloroethyl chloroformate
to give 45 and subsequent reduction of the oxindole
carbonyl with borane to afford spiro-indoline 46 (Scheme
11). The oxidation of this indoline (urea‚H2O2, Na2WO4‚
2H2O, 10% aq MeOH) proceeded rather slowly, but did
produce 47 after subsequent treatment with diaz-
omethane.
quenched with isoprene, and in this fashion bis-amide
39 was obtained in 79% yield.
With 39 in hand, the introduction of the ethyl moiety
was studied (Scheme 9). In initial model studies, a methyl
carbamate was successfully used to activate the lactam
carbonyl for attack of EtMgBr leading to the desired
stereoisomer.16 Therefore, we anticipated that a similar
sequence of reactions could also be applied to lactam 39.
Although we were aware that it would be impossible to
selectively react the N-7 lactam NH with methyl chlo-
roformate, we felt confident that in the Grignard step
we would be able to discriminate between the carbamate
and the oxindole carbonyl for steric reasons. Thus, by
deprotonation with NaH, followed by addition of methyl
chloroformate, the polar bis-amide 39 was converted to
the less polar bis-methyl carbamate 40. Subsequent
treatment with excess of EtMgBr, however, gave exclu-
sively the bis-amide 39 resulting from attack of EtMgBr
onto the carbonyls of both carbamates. Variation of the
temperature did not change the outcome of this reaction.
The sterically more hindered Boc protecting group was
then used to suppress this undesired reaction (Scheme
10). Thus, treatment of 39 with Boc2O and DMAP in THF
gave bis-Boc carbamate 41 in 78% yield. On reaction with
excess of EtMgBr the N,O-hemiacetal 42 was produced
as a single product. The Boc group was removed from
the oxindole in this process, presumably due to attack of
EtMgBr on the Boc carbonyl as the oxindole carbonyl was
too sterically hindered. As expected, the ethyl group was
introduced selectively from the top face. Although 42 was
stable at room temperature, it was treated in crude form
with TFA (trifluoroacetic acid) generating the N-acylimin-
ium ion 43, which was reduced in situ with triethylsilane.
Again, reduction took place selectively from the least
hindered side to form a single product. Additional TFA
removed the remaining Boc group, and the TFA salt of
desmethoxygelsedine 44 was obtained in 61% yield over
two steps. At this stage the stereochemistry of the ethyl
and oxindole moieties was unambiguously proven by
NOE measurements (a NOE of 5.0% was observed
between H-8 and H-9 and a NOE of 3.9% between H-10
and H-15). The presence of the trifluoroacetate counterion
was indicated with 19F NMR and 13C NMR. The 19F NMR
spectrum of 44 in CD3OD showed one peak at δ -77.4
(19F NMR spectrum of trifluoroacetic acid in CD3OD, one
peak at δ -78.2). In the 13C NMR spectrum two quartets
were observed at δ 163.4 (J CF ) 35.4 Hz) and δ 118.3
(J CF ) 290.6 Hz), respectively.
The last step was the deprotection of the Troc group
with zinc in acetic acid. After workup, the 1H NMR
spectrum of the crude product showed less than 5% of a
byproduct, which was only different in the aromatic
region. Presumably the N-O bond was partially reduced
during the deprotection. After purification with prepara-
tive TLC, pure (+)-gelsedine was obtained in 84% yield,
which was identical to natural (-)-gelsedine14 (400 MHz
1H NMR, 100 MHz 13C NMR spectra and HRMS (FAB)),
except for the optical rotation (synthetic (+)-gelsedine:
[R]22 ) +120 (c ) 0.25, CHCl3); natural (-)-gelsedine:
D
[R]25 ) -159 (c ) 1.35, CHCl3)).7 The spectral compari-
D
sons were made by using NMR spectra and a sample of
(-)-gelsedine kindly made available to us by Professor
H. Takayama of Chiba University in J apan.
Con clu sion
We have successfully completed the first de novo total
synthesis of enantiopure ent-gelsedine in 21 steps from
(S)-malic acid in an overall yield of 0.10%. The unique
combination of iodide-promoted cationic allene cyclization
producing a vinyl iodide and palladium-catalyzed elabo-
ration of the latter in carbonylation and subsequent Heck
chemistry has been the key to success. The Heck cycliza-
tion and the introduction of the ethyl moiety via N-
acyliminium ion chemistry both proceeded with complete
stereoselectivity. In addition, a methyl group has been
shown to be a useful oxindole N-protective group. We
anticipate that the novel transformations developed in
this project will find further application in synthesis.
Exp er im en ta l Section
Gen er a l. All reactions involving oxygen or moisture sensi-
tive compounds were carried out under a dry nitrogen atmo-
sphere. Unless otherwise noted, reagents were added by
syringe. IR spectra were measured as thin films on NaCl plates
unless otherwise noted and wavelengths (ν) are reported in
cm-1. NMR spectra were measured as solutions in CDCl3
(unless otherwise noted). Chemical shifts (δ) are expressed in
ppm relative to an internal standard of CHCl3 (7.26 ppm for
1H NMR and 77.0 ppm for 13C NMR). Mass spectra and
accurate mass determinations were performed using electron
(36) (a) Murahashi, S.-I.; Oda, T.; Sugahara, T.; Masui, Y. J . Org.
Chem. 1990, 55, 1744-1749. (b) Kitajima, M.; Takayama, H.; Sakai,
S. J . Chem. Soc., Perkin Trans. 1 1994, 1573-1578.
For the final stage introduction of the N-methoxy
group, a procedure precedented to work on similar

Total Synthesis of ent-Gelsedine
J . Org. Chem., Vol. 65, No. 24, 2000 8323
impact (EI) mass spectrometry, unless otherwise noted. Rf
values were obtained using thin-layer chromatography (TLC)
on Merck F-254 silica gel plastic sheets with the indicated
solvent (mixture). Chromatographic purification refers to flash
chromatography using the solvent (mixture) mentioned above
(unless otherwise noted) and Acros silica gel (0.030-0.075
mm).37 Melting and boiling points are uncorrected. All starting
materials were obtained from commercial suppliers and used
without further purification. THF was distilled from sodium/
benzophenone prior to use. DMF, CH2Cl2, and MeCN were
distilled from CaH2 and stored over 4 Å molecular sieves under
dry nitrogen atmosphere. Triethylamine and diisopropylamine
were distilled from and stored over KOH pellets.
solution of 21 (2.50 g, 5.66 mmol) in CH2Cl2 (28 mL) was added
PCC (4.88 g, 22.64 mmol), and the reaction mixture was heated
at 40 °C for 17 h. Then extra PCC (1.22 g, 5.66 mmol) was
added, and the mixture was stirred at 40 °C for another 5 h,
cooled to room temperature, and poured onto a short column
of silica (4.0 × 6.0 cm). This was eluted with acetone. The top
layer of the short column was removed and thoroughly stirred
with acetone, and then poured back onto the column and eluted
further with acetone until no more product came off. The
acetone was evaporated and the residue was chromatographed
(CH2Cl2/acetone 8:1) to give 22 (1.64 g, 66%) as a colorless oil;
Rf 0.39 (CH2Cl2/acetone 8:1); [R]22 -30.8 (c ) 1.0, CHCl3); IR
D
3279, 2921, 1723, 1696, 1679, 1577, 1517, 1434, 1299; 1H NMR
(400 MHz) δ 8.21 (d, J ) 8.2 Hz, 1H), 7.80 (s, 1H, NH), 7.53
(d, J ) 8.0 Hz, 1H), 7.33-7.24 (m, 5H), 7.22-7.19 (m, 1H),
6.99 (t, J ) 7.9 Hz, 1H), 6.54 (s, 1H), 4.91 (d, J ) 14.7 Hz,
1H), 4.27 (d, J ) 14.7 Hz, 1H), 3.99 (s, 1H), 3.10-3.06 (m,
3H), 2.68 (d, J ) 16.9 Hz, 1H), 2.46 (d, J ) 17.8 Hz, 1H); 13C
NMR (100 MHz) δ 207.8, 170.7, 167.0, 135.2, 135.0, 133.7,
(1S,6R,9S)-F or m ic
Acid
7-Ben zyl-4-iod o-8-oxo-7-
a za bicyclo[4.2.1]n on -3-en -9-yl Ester (11). To a solution of
1216 (10.62 g, 0.0369 mol) in formic acid (185 mL) was added
anhydrous sodium iodide (110.77 g, 0.739 mol). After stirring
the reaction mixture at 85 °C for 18 h, the formic acid was
evaporated, and the residue was taken up in CH2Cl2 (250 mL)
and water (250 mL). The water layer was extracted with CH2-
Cl2 (2 × 75 mL), and the combined organic layers were washed
with water (75 mL), dried (MgSO4), and concentrated. The
residue was chromatographed (first with EtOAc/petroleum
ether 1:1, then with EtOAc/petroleum ether 4:1) to give 11
(6.16 g, 42%) and 13 (3.60 g, 34%) both as colorless oils; 11:
132.2, 129.7, 128.9, 128.4, 128.3, 128.1, 125.6, 121.9, 113.9,
79
62.9, 49.4, 43.9, 30.9, 29.6; HRMS calculated for C22H19
BrN2O3 438.0579, found 438.0605.
-
(1S,6R)-7-Ben zyl-9-m eth ylen e-8-oxo-7-a za bicyclo[4.2.1]-
n on -3-en e-4-ca r boxylic Acid (2-Br om op h en yl)a m id e (9).
To a solution of methyltriphenylphosphonium bromide (3.33
g, 9.33 mmol) in THF (40 mL) was added at 0 °C a 1.6 M
solution of n-BuLi in hexane (5.72 mL, 9.14 mmol). After
stirring for 25 min a solution of 22 (1.64 g, 3.73 mmol) in THF
(10 mL) was added. The reaction mixture was stirred at 70
°C for 18 h and then poured onto saturated aqueous NH4Cl
(100 mL). The water layer was extracted with CH2Cl2 (3 × 60
mL). The combined organic layers were dried (MgSO4) and
concentrated. The residue was chromatographed to give 9 (1.37
g, 83%) as a colorless oil; Rf 0.34 (EtOAc/petroleum ether 2:1);
[R]22D -31.0 (c ) 1.0, CHCl3); IR 3279, 2907, 1699, 1673, 1564,
Rf 0.31 (EtOAc/petroleum ether 1:1); [R]22 +10.3 (c ) 1.0,
D
CHCl3); IR 2920, 1735, 1682, 1495, 1446, 1165; 1H NMR (400
MHz) δ 7.97 (s, 1H), 7.37-7.23 (m, 5H), 6.34 (s, 1H), 5.02 (s,
1H), 4.86 (d, J ) 15.3 Hz, 1H), 4.12 (d, J ) 15.3 Hz, 1H), 3.44
(t, J ) 3.3 Hz, 1H), 2.99 (d, J ) 18.7 Hz, 1H), 2.83 (d, J ) 18.9
Hz, 1H), 2.77 (s, 1H), 2.70 (d, J ) 18.5 Hz, 1H), 2.57 (d, J )
18.5 Hz, 1H); 13C NMR (100 MHz) δ 173.0, 159.8, 137.3, 135.4,
128.7, 127.6, 127.5, 93.1, 75.4, 61.9, 45.8, 45.2, 44.2, 32.5;
HRMS calculated for C16H16INO3 397.0174, found 397.0146.
13: Rf 0.40 (EtOAc/petroleum ether 4:1); [R]22 +6.4 (c ) 0.5,
D
1
CHCl3); IR 2932, 1721, 1698, 1694, 1451, 1337, 1224, 1164;
1H NMR (400 MHz) δ 7.94 (s, 1H), 7.37-7.27 (m, 5H), 5.00 (d,
J ) 15.1 Hz, 1H), 4.76 (s, 1H), 4.20 (d, J ) 15.1 Hz, 1H), 3.70
(dd, J ) 2.6, 4.3 Hz, 1H), 2.75 (t, J ) 4.4 Hz, 1H), 2.60 (dd, J
) 2.6, 16.2 Hz, 1H), 2.55 (dd, J ) 4.4, 16.3 Hz, 1H), 2.48-2.38
(m, 2H), 2.30-2.23 (m, 1H), 1.93-1.84 (m, 1H); 13C NMR (100
MHz) δ 207.9, 172.9, 159.7, 135.3, 129.0, 128.1, 128.0, 75.8,
57.5, 47.7, 44.3, 43.6, 39.9, 24.4; HRMS (FAB+) calculated for
1513, 1434, 1298; H NMR (400 MHz) δ 8.28 (d, J ) 8.2 Hz,
1H), 7.81 (s, 1H, NH), 7.54 (d, J ) 7.9 Hz, 1H), 7.34-7.19 (m,
6H), 6.99 (t, J ) 7.6 Hz, 1H), 6.52 (s, 1H), 5.05 (d, J ) 7.9 Hz,
2H), 4.88 (d, J ) 15.1 Hz, 1H), 4.20 (s, 1H), 4.10 (d, J ) 15.2
Hz, 1H), 3.22 (s, 1H), 3.11 (d, J ) 17.9 Hz, 2H), 2.63 (d, J )
19.2 Hz, 1H), 2.49 (d, J ) 18.6 Hz, 1H); 13C NMR (100 MHz)
δ 174.4, 168.1, 147.1, 136.0, 135.2, 133.4, 132.2, 131.2, 128.7,
128.3, 128.1, 127.6, 125.3, 121.9, 113.8, 106.8, 59.4, 46.1, 43.9,
35.5, 34.7; HRMS calculated for C23H2179BrN2O2 436.0786,
found 436.0767.
C
15H18NO3 (M + H - CO) 260.1287, found 260.1294.
(1S,6R,9S)-7-Ben zyl-9-h ydr oxy-8-oxo-7-azabicyclo[4.2.1]-
n on -3-en e-4-ca r boxylic Acid (2-Br om op h en yl)a m id e (21).
A solution of 11 (1.67 g, 4.20 mmol), Pd(OAc)2 (0.141 g, 0.630
mmol), triphenylphospine (0.330 g, 1.26 mmol), o-bromoaniline
(2.16 g, 12.6 mmol), and triethylamine (2.93 mL, 21.0 mmol)
in DMF (14 mL) was stirred under carbon monoxide atmo-
sphere (balloon) for 20 h at 40 °C. Then the solvents were
evaporated under reduced pressure using a vacuum pump
(<50 mbar, heat slightly with a heat gun to evaporate DMF),
and the residue was chromatographed (EtOAc/petroleum ether
4:1) to give a reddish oil, which was dissolved in a saturated
methanolic NH3 solution (12 mL) and stirred for 15 min at
room temperature. The solvent was evaporated and the residue
chromatographed to give 21 (1.19 g, 64%) as a colorless oil; Rf
(1S,6R)-7-Ben zyl-9-m eth ylen e-8-oxo-7-a za bicyclo[4.2.1]-
n on -3-en e-4-ca r boxylic Acid Meth yl-(2-br om op h en yl)-
a m id e (23a ). To a solution of 9 (1.36 g, 3.11 mmol) in THF
(15 mL) was added 60% sodium hydride in mineral oil (0.373
g, 9.33 mmol). Methyl iodide (0.484 mL, 7.78 mmol) was added
at 0 °C. The reaction mixture was stirred at room temperature
for 4.5 h (not longer) and then poured onto saturated aqueous
NH4Cl (25 mL). The water layer was extracted with CH2Cl2
(3 × 25 mL). The combined organic layers were dried (MgSO4)
and concentrated. The residue was chromatographed to give
23a (1.10 g, 78%) as a colorless oil; Rf 0.36 (EtOAc/petroleum
ether 10:1); [R]22D +56.6 (c ) 0.5, CHCl3); IR 3059, 2919, 1697,
1673, 1584, 1478, 1444, 1420, 1364; 1H NMR (400 MHz,
benzene-d6, T ) 340 K) δ 7.25 (d, J ) 7.5 Hz, 1H), 7.19-7.03
(m, 5H), 6.82 (t, J ) 7.4 Hz, 1H), 6.72 (d, J ) 6.6 Hz, 1H),
6.61 (t, J ) 6.6 Hz, 1H), 5.65 (s, 1H), 5.08 (d, J ) 15.1 Hz,
1H), 4.45 (s, 2H), 3.79 (s, 1H), 3.68 (d, J ) 14.1 Hz, 1H), 3.10
(s, 3H), 3.04 (d, J ) 19.2 Hz, 1H), 2.76 (s, 1H), 2.60 (d, J )
17.9 Hz, 1H), 2.23 (d, J ) 11.8 Hz, 1H), 1.87 (d, J ) 18.1 Hz,
1H); 13C NMR (100 MHz, benzene-d6, T ) 340 K) δ 173.9,
172.7, 149.2, 145.0, 137.9, 134.4, 132.7, 130.9, 129.3, 128.9,
128.7, 128.5, 128.2, 127.9, 123.5, 105.8, 59.8, 47.0, 44.1, 37.1,
36.5, 35.9; HRMS (FAB+) calculated for C24H2479BrN2O2 (M
+ H) 451.1021, found 451.0993.
0.45 (CH2Cl2/acetone 1:1); [R]22 -16.3 (c ) 1.0, CHCl3); IR
D
3405, 2918, 1685, 1676, 1590, 1518, 1433, 1298, 1249; 1H NMR
(400 MHz) δ 8.23 (d, J ) 8.3 Hz, 1H), 7.80 (s, 1H, NH), 7.54
(d, J ) 8.0 Hz, 1H), 7.34-7.25 (m, 5H), 7.20-7.16 (m, 1H),
6.99 (t, J ) 7.8 Hz, 1H), 6.47 (s, 1H), 4.78 (d, J ) 15.2 Hz,
1H), 4.20 (d, J ) 15.2 Hz, 1H), 4.04 (s, 1H), 3.76 (s, 1H), 2.96
(d, J ) 19.4 Hz, 1H), 2.87 (d, J ) 18.5 Hz), 2.74-2.63 (m, 2H),
2.52 (d, J ) 18.4 Hz, 1H), 2.35 (s, 1H, OH); 13C NMR (100
MHz) δ 175.2, 167.8, 136.0, 135.4, 133.0, 132.2, 131.3, 128.6,
128.4, 128.2, 127.8, 127.5, 125.5, 122.4, 114.4, 75.0, 63.2, 48.7,
44.5, 30.6, 30.1; HRMS (FAB+) calculated for C22H2279BrN2O3
(M + H) 441.0813, found 441.0815.
N-Meth yl Sp ir o-oxin d ole 24a . A solution of 23a (0.365
g, 0.809 mmol), Pd(PPh3)4 (0.187 g, 0.162 mmol) and triethy-
lamine (1.12 mL, 8.08 mmol) in MeCN (8 mL) in a sealed tube
was heated slowly to 120 °C. After stirring for 40 h, the
reaction mixture was cooled to room temperature, and the
solvent was evaporated. The residue was chromatographed
(1S,6R)-7-Ben zyl-8,9-d ioxo-7-a za b icyclo[4.2.1]n on -3-
en e-4-ca r boxylic Acid (2-Br om op h en yl)a m id e (22). To a
(37) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.

8324 J . Org. Chem., Vol. 65, No. 24, 2000
Beyersbergen van Henegouwen et al.
(load with CH2Cl2) to give 24a (0.270 g, 90%) as a colorless
oil; Rf 0.42 (EtOAc/petroleum ether 10:1); [R]22D +14.9 (c ) 1.0,
CHCl3); IR 3027, 2930, 1712, 1673, 1608, 1492, 1343, 1248;
1H NMR (400 MHz) δ 7.33-7.21 (m, 6H), 7.07 (dd, J ) 7.3,
16.4 Hz, 1H), 7.00 (t, J ) 7.5 Hz, 1H), 6.77 (d, J ) 7.7 Hz,
1H), 6.30 (dd, J ) 8.7, 11.4 Hz, 1H), 5.32 (d, J ) 15.7 Hz, 1H),
5.04 (s, 1H), 4.95 (s, 1H), 4.93 (d, J ) 11.1 Hz, 1H), 4.17 (s,
1H), 3.98 (d, J ) 15.7 Hz, 1H), 3.62 (d, J ) 8.7 Hz, 1H), 3.17
(s, 3H), 2.56 (dd, J ) 3.5, 15.5 Hz, 1H), 2.06 (dd, J ) 2.8, 15.5
Hz, 1H); 13C NMR (100 MHz) δ 177.4, 172.9, 147.8, 142.2,
136.5, 132.2, 131.6, 128.8, 128.4, 128.2, 127.7, 127.1, 123.7,
122.9, 107.9, 105.9, 61.0, 54.7, 49.9, 44.4, 38.2, 26.4; HRMS
calculated for C24H22N2O2 370.1681, found 370.1692.
J ) 7.6 Hz, 1H), 5.29 (d, J ) 15.4 Hz, 1H), 4.23 (dd, J ) 3.0,
11.3 Hz, 1H), 4.18 (d, J ) 1.5, 11.2 Hz, 1H), 3.81-3.77 (m,
1H), 3.74-3.72 (m, 1H), 3.70 (d, J ) 15.4 Hz, 1H), 3.19 (s,
3H), 2.80 (dd, J ) 8.6, 9.9 Hz, 1H), 2.66 (t, J ) 8.0 Hz, 1H),
2.57 (d, J ) 15.0 Hz, 1H), 2.29-2.23 (m, 1H), 2.15-2.14 (m,
2H); 13C NMR (100 MHz) δ 176.4, 175.5, 141.9, 136.8, 134.9,
128.7, 128.5, 128.0, 127.3, 123.9, 122.6, 107.5, 75.2, 61.8, 58.6,
56.9, 43.1, 36.5, 35.0, 31.7, 26.8, 26.2; HRMS (FAB+) calcu-
lated for C24H25N2O3 (M + H) 389.1865, found 389.1871.
Sp ir o-oxin d ole 38. A solution of 29 (0.646 g, 1.66 mmol)
and benzoyl peroxide (0.806 g, 3.32 mmol) in CH2Cl2 (3.5 mL)
in a sealed tube was heated slowly to 80 °C. After stirring for
18 h, the reaction mixture was cooled to room temperature,
and the solvent was evaporated. The residue was dissolved in
a saturated methanolic NH3 solution (15 mL) and stirred at
room temperature for 20 h. The solvent was evaporated and
the residue chromatographed (pulverize the residue and load
on column with the eluens) to give 38 (0.463, 74%) as a
N-Meth yl Sp ir o-oxin d ole 25a . To a solution of 24a (0.803
g, 2.16 mmol) in THF (7 mL) was added at 0 °C a freshly
prepared 0.5 M solution of dicyclohexylborane in THF (8.67
mL, 4.33 mmol).25 After stirring the reaction mixture for 2 h
at 0 °C, a 3 M solution of NaOH in water (4.33 mL, 13.0 mmol)
and a 35% solution of H2O2 in water (1.68 mL, 17.3 mmol)
were added slowly. The mixture was stirred for 1 h at room
temperature and then poored onto saturated aqueous NH4Cl
(10 mL). The aqueous layer was extracted with CH2Cl2 (3 ×
20 mL). The combined organic layers were dried (MgSO4) and
concentrated. The residue was chromatographed to give 25a
(0.665 g, 79%) as a white solid; Rf 0.37 (CH2Cl2/acetone 1:1);
colorless oil; Rf 0.33 (CH2Cl2/acetone 1:3); [R]22 +141.8 (c )
D
1.0, CHCl3); IR 3169, 2923, 1713, 1677, 1619, 1472, 1451, 1113;
1H NMR (400 MHz) δ 8.67 (s, 1H, NH), 7.42 (d, J ) 7.4 Hz,
1H), 7.31-7.17 (m, 6H), 7.01 (t, J ) 7.6 Hz, 1H), 6.90 (d, J )
7.7 Hz, 1H), 5.27 (d, J ) 15.4 Hz, 1H), 4.21 (dd, J ) 2.8, 11.4
Hz, 1H), 4.18 (dd, J ) 1.1, 11.4 Hz, 1H), 3.82-3.80 (m, 2H),
3.74 (d, J ) 15.4 Hz, 1H), 2.84 (t, J ) 8.8 Hz, 1H), 2.67 (t, J
) 8.2 Hz, 1H), 2.61 (d, J ) 16.9 Hz, 1H), 2.32-2.24 (m, 1H),
mp 87-89 °C; [R]22 -13.4 (c ) 1.0, CHCl3); IR 3394, 2932,
D
1699, 1682, 1610, 1495, 1471, 1450, 1342, 1252. 1H NMR (400
MHz) δ 7.31-7.24 (m, 7H), 7.07 (t, J ) 7.5 Hz, 1H), 6.81 (d, J
) 7.8 Hz, 1H), 6.00 (dd, J ) 9.1, 11.6 Hz, 1H), 5.33 (d, J )
15.7 Hz, 1H), 5.20 (d, J ) 11.7 Hz, 1H), 4.12-4.02 (m, 2H),
3.86 (d, J ) 15.7 Hz, 1H), 3.78 (t, J ) 4.5 Hz, 1H), 3.26 (dd, J
) 6.8, 8.7 Hz, 1H), 3.19 (s, 3H), 2.79-2.75 (m, 1H), 2.30 (dd,
J ) 4.3, 16.3 Hz, 1H), 2.21 (dd, J ) 1.6, 16.1 Hz, 1H); 13C NMR
(100 MHz) δ 176.6, 174.9, 142.2, 137.2, 135.1, 132.4, 128.5,
128.4, 128.2, 127.7, 127.2, 123.2, 123.0, 107.9, 59.3, 56.9, 55.3,
45.3, 45.1, 45.0, 32.7, 26.5; HRMS calculated for C24H24N2O3
370.1787, found 370.1787.
2.26 (d, J ) 16.9 Hz, 1H). 2.16 (dd, J ) 3.5, 15.9 Hz, 1H); 13
C
NMR (100 MHz) δ 178.6, 176.7, 139.4, 136.7, 135.6, 132.9,
130.0, 128.7, 128.3, 128.1, 127.4, 124.3, 122.6, 109.6, 75.1, 61.9,
58.7, 57.6, 44.0, 36.7, 35.1, 31.9, 26.8; HRMS calculated for
C
23H22N2O3 374.1631, found 374.1641.
Sp ir o-oxin d ole 39. To condensed NH3 (10 mL) was added
lithium (0.036 g, 5.18 mmol) at -78 °C. After stirring the deep
blue solution for 5 min, 38 (0.200 g, 0.531 mmol) in THF (4
mL) was added slowly. The reaction mixture was stirred for
30 min at -78 °C, and then isoprene (1.32 mL, 13.3 mmol)
was added carefully. The ammonia was allowed to evaporate
slowly at room temperature by passing a stream of nitrogen
over the solution, and the THF was evaporated. The residue
was chromatographed (pulverize the residue and load on
column with the eluens) to give 39 (0.119 g, 79%) as a white
Ch lor om er cu r ia l 26a . To a solution of 25a (0.540 g, 1.39
mmol) in THF (14 mL) was added mercury(II) trifluoroacetate
(1.19 g, 2.78 mmol). After stirring the reaction mixture for 5
h at room temperature, saturated aqueous NaCl (7 mL) was
added, and the mixture was stirred vigorously for an additional
18 h. The aqueous layer was extracted with CH2Cl2 (3 × 20
mL). The combined organic layers were dried (MgSO4) and
concentrated. The residue was chromatographed (load with
CH2Cl2) to give 26a (0.762 g, 88%) as a white solid; Rf 0.33
solid; Rf 0.40 (CH2Cl2/methanol 4:1); mp 191-193 °C; [R]22
D
+143.4 (c ) 1.1, CHCl3); IR 3407, 1698, 1682, 1621, 1472, 1109.
1H NMR (400 MHz) δ 9.65 (s, 1H, NH), 7.37 (d, J ) 7.1 Hz,
1H), 7.08 (t, J ) 6.6 Hz, 1H), 7.06 (s, 1H, NH), 6.95 (t, J ) 7.4
Hz, 1H), 6.82 (d, J ) 7.6 Hz, 1H), 4.19-4.16 (m, 2H), 4.00 (d,
J ) 7.6 Hz, 1H), 3.77 (d, J ) 3.3 Hz, 1H), 2.76 (t, J ) 7.9 Hz,
1H), 2.57 (t, J ) 8.6 Hz, 1H), 2.46 (d, J ) 14.7 Hz, 1H), 2.26-
2.17 (m, 2H), 1.92 (d, J ) 14.7 Hz, 1H); 13C NMR (100 MHz)
δ 180.8, 179.0, 139.7, 135.9, 127.9, 124.2, 122.4, 109.7, 74.7,
61.9, 57.5, 56.9, 36.7, 36.6, 36.2, 26.7; HRMS calculated for
(EtOAc/petroleum ether 4:1); mp 158-160 °C; [R]22 +14.5 (c
D
) 1.0, CHCl3); IR 2919, 1687, 1680, 1608, 1494, 1470, 1449,
1
1349, 1258; H NMR (400 MHz) δ 7.49 (d, J ) 7.5 Hz, 1H),
7.33-7.22 (m, 6H), 7.10 (t, J ) 7.6 Hz, 1H), 6.83 (d, J ) 7.7
Hz, 1H), 5.29 (d, J ) 15.4 Hz, 1H), 4.28 (dd, J ) 3.1, 11.4 Hz,
1H), 4.19 (dd, J ) 1.1, 11.3 Hz, 1H), 3.92-3.89 (m, 2H), 3.74
(d, J ) 15.4 Hz, 1H), 3.27 (s, 3H), 3.22 (dd, J ) 3.9, 9.8 Hz,
1H), 2.97 (t, J ) 9.4 Hz, 1H), 2.75 (t, J ) 7.4 Hz, 1H), 2.26
C
16H16N2O3 284.1161, found 284.1167.
Bis-ter t-bu tyl Ca r ba m a te 41. To a solution of 39 (0.285
g, 1.00 mmol) in THF (8 mL) were added N,N-(dimethylami-
no)pyridine (0.735 g, 6.02 mmol) and Boc2O (1.75 g, 8.03
mmol). The reaction mixture was stirred for 5 h and then
poured onto saturated aqueous NH4Cl (15 mL). The waterlayer
was extracted with CH2Cl2 (3 × 15 mL). The combined organic
layers were dried (MgSO4) and concentrated. The residue was
chromatographed to give 41 (0.379 g, 78%) as a colorless oil;
(dd, J ) 3.4, 15.4 Hz, 1H), 2.17 (dd, J ) 1.7, 15.2 Hz, 1H); 13
C
NMR (100 MHz) δ 179.6, 177.1, 141.4, 136.1, 134.9, 128.5,
128.2, 127.9, 127.4, 124.2, 123.1, 108.0, 79.5, 61.2, 59.1, 57.8,
45.1, 44.4, 43.3, 36.7, 32.6, 26.7; HRMS (FAB+) calculated for
C
24H24ClHgN2O3 (M + H) 625.1182, found 625.1215.
Tetr a h yd r op yr a n 29. To a solution of 26a (0.600 g, 0.962
mmol) in toluene (1.9 mL, 0.5 M!), which was stirred vigor-
ously, were added AIBN (0.0158 g, 0.0962 mmol) and n-Bu3-
SnH (0.647 mL, 2.40 mmol) at one time. The reaction mixture
was stirred at room temperature for 1 h and then at 55 °C for
1 h. After cooling to room temperature, CCl4 (0.5 mL) was
added, and the mixture was stirred for an additional 1 h and
then poured onto a 5% aqueous solution of KF (15 mL). The
aqueous layer was extracted with CH2Cl2 (3 × 25 mL). The
combined organic layers were dried (MgSO4) and concentrated.
The residue was chromatographed (first load and eluate with
CH2Cl2, then with CH2Cl2/acetone 2:1) to give 29 (0.299 g, 80%)
as a white solid; Rf 0.34 (CH2Cl2/acetone 2:1); mp 68-70 °C;
[R]22D +139.0 (c ) 1.0, CHCl3); IR 2917, 1698, 1693, 1609, 1493,
1470, 1447, 1345, 1256; 1H NMR (400 MHz) δ 7.45 (d, J ) 7.4
Hz, 1H), 7.31-7.21 (m, 6H), 7.05 (t, J ) 7.6 Hz, 1H), 6.77 (d,
Rf 0.41 (EtOAc/petroleum ether 6:1); [R]22 +60.4 (c ) 1.0,
D
CHCl3); IR 2979, 1789, 1759, 1730, 1711, 1605, 1369, 1306,
1
1251; H NMR (400 MHz) δ 7.74 (d, J ) 8.1 Hz, 1H), 7.52 (d,
J ) 7.6 Hz, 1H), 7.27 (t, J ) 7.8 Hz, 1H), 7.15 (t, J ) 7.6 Hz,
1H), 4.57-4.52 (m, 1H), 4.25 (dd, J ) 1.3, 11.4 Hz, 1H), 4.20
(dd, J ) 2.9, 14.4 Hz, 1H), 3.88 (d, J ) 3.6 Hz, 1H), 2.81 (t, J
) 8.7 Hz, 1H), 2.72 (t, J ) 8.2 Hz, 1H), 2.66 (d, J ) 15.8 Hz,
1H), 2.53 (d, J ) 15.3 Hz, 1H), 2.35-2.27 (m, 1H), 2.22 (dd, J
) 3.8, 15.9 Hz, 1H), 1.58 (s, 9H), 1.52 (s, 9H); 13C NMR (100
MHz) δ 175.2, 173.8, 150.2, 149.3, 138.0, 133.7, 128.4, 124.6,
124.1, 114.2, 84.2, 82.7, 74.8, 61.6, 60.3, 56.7, 38.2, 34.2, 33.3,
28.1, 28.0, 26.6; HRMS calculated for C26H32N2O7 484.2210,
found 484.2197.
TF A Sa lt of Desm eth oxygelsed in e (44). To a solution of
41 (0.168 g, 0.346 mmol) in THF (2 mL) at -78 °C was added

Total Synthesis of ent-Gelsedine
J . Org. Chem., Vol. 65, No. 24, 2000 8325
a 1 M solution of ethylmagnesium bromide in THF (2.08 mL,
2.08 mmol). The reaction mixture was stirred at -78 °C for
15 min at room temperature for 15 min and then quenched
with saturated aqueous NH4Cl (5 mL). The aqueous layer was
extracted with CH2Cl2 (3 × 15 mL, water (a few mL) was added
to get a clear separation). The combined organic layers were
dried (MgSO4) and concentrated. The residue was not purified,
but directly used in the following reaction. To a solution of
the residue in CH2Cl2 (1.5 mL) was added trifluoroacetic acid
(0.320 mL, 4.15 mmol) at 0 °C, and the mixture was stirred at
0 °C for 10 min. Then triethylsilane (0.416 mL, 2.77 mmol)
was added. After stirring the solution for 1.5 h at 0 °C and 2.5
h at room temperature, additional trifluoroacetic acid (0.4 mL)
was added, and the reaction mixture was stirred for an
additional 18 h. The solvents were evaporated, and the residue
was chromatographed to give 44 (0.087 g, 61%) as a white
Tr oc-P r ot ect ed Gelsed in e (47). To a solution of 46
(0.0147 g, 0.0319 mmol) in 10% aqueous MeOH (0.6 mL) were
added Na2WO4‚2H2O (6.3 mg, 0.019 mmol) and urea‚H2O2
(0.030 g, 0.32 mmol). After stirring the reaction mixture for 2
h at room temperature, more urea‚H2O2 (0.030 g, 0.32 mmol)
was added, and the mixture was stirred for an additional 2 h.
Then it was poured onto saturated aqueous NH4Cl (5 mL) and
extracted with CH2Cl2 (3 × 10 mL). The combined organic
layers were dried (MgSO4) and concentrated. The residue was
dissolved in MeOH (4 mL), and a freshly prepared solution of
diazomethane in ether (0.5 mL) was added at 0 °C. After
stirring the reaction mixture for 1 h at room temperature, the
solvent was evaporated and the residue chromatographed to
give 47 (5.0 mg, 31%) as a colorless oil; Rf 0.23 (EtOAc/
petroleum ether 1:2); [R]22 +95.5 (c ) 0.26, CHCl3); IR 2925,
D
1716, 1621, 1463, 1402, 1291, 1119; 1H NMR (400 MHz,
benzene-d6, T ) 340 K) δ 7.38 (d, J ) 7.5 Hz, 1H), 7.04 (t, J )
7.6 Hz, 1H), 6.92 (t, J ) 7.6 Hz, 1H), 6.71 (d, J ) 7.3 Hz, 1H),
4.66 (s, 2H), 4.18 (s, 1H), 3.83 (dd, J ) 3.4, 10.9 Hz, 1H), 3.81
(d, J ) 11.0 Hz, 1H), 3.60 (s, 3H), 3.54-3.51 (m, 1H), 3.43 (d,
J ) 6.7 Hz, 1H), 3.04-2.98 (m, 1H), 2.72 (d, J ) 15.9 Hz, 1H),
2.55 (d, J ) 15.2 Hz, 1H), 2.52-2.46 (m, 1H), 2.03-1.85 (m,
3H), 1.67-1.58 (m, 1H), 0.83 (t, J ) 7.6 Hz, 3H); HRMS
(FAB+) calculated for C22H26Cl3N2O5 (M + H) 503.0908, found
503.0904. A useful 13C NMR spectrum could not be obtained
due to rotamers.
solid; Rf 0.42 (CH2Cl2/methanol 4:1); mp 118-120 °C; [R]22
D
+83.0 (c ) 1.0, MeOH); IR 2971, 1679, 1620, 1472, 1435, 1201,
1124; 1H NMR (400 MHz) δ 11.5 (s, 1H, NH), 11.3 (s, 1H, NH),
9.32 (s, 1H, NH), 7.33 (d, J ) 7.4 Hz, 1H), 7.24 (t, J ) 7.7 Hz,
1H), 7.06 (t, J ) 7.5 Hz, 1H), 6.95 (d, J ) 7.7 Hz, 1H), 4.51 (s,
1H), 4.28-4.24 (m, 2H), 3.80 (s, 1H), 3.72 (d, J ) 6.5 Hz, 1H),
2.88 (s, 1H), 2.56 (s, 1H), 2.35 (dd, J ) 3.4, 16.4 Hz, 1H), 2.26
(d, J ) 15.5 Hz, 1H), 2.21-2.06 (m, 2H), 2.03-1.95 (m, 2H),
1.08 (t, J ) 7.5 Hz, 3H); 13C NMR (100 MHz, CD3OD) δ 184.2,
163.4 (q, J CF ) 35.4 Hz), 141.5, 136.5, 129.9, 126.0, 124.4, 118.3
(q, J CF ) 290.6 Hz), 111.5, 76.6, 66.4, 63.7, 61.3, 60.9, 40.9,
35.2, 32.0, 22.1, 20.6, 11.4; 19F NMR (282 MHz, CD3OD) δ )
-77.4; HRMS calculated for C18H22N2O2 298.1682, found
298.1679.
en t-Gelsed in e (1). To a solution of 47 (4.7 mg, 0.093 mmol)
in acetic acid (0.6 mL) was added zinc dust (0.012 g, 0.18
mmol). After stirring the reaction mixture at room temperature
for 2 h, more zinc dust (0.012 g, 0.18 mmol) was added, and
the mixture was stirred for an additional 18 h. The suspension
was filtered and diluted with ice-water (2 mL). The mixture
was basified with a 25% aqueous solution of NH4OH (pH >
10) and was extracted with MeOH/CHCl3 (1:9, 3 × 10 mL).
The organic layers were washed with brine (4 mL), dried
(MgSO4), and concentrated, and the residue was purified by
preparative TLC (plate 10 × 10 cm, after eluting with MeOH/
CHCl3 (1:3), the plate was turned around (180°) and eluted
with EtOAc) to give 1 (2.6 mg, 84%) as a colorless oil; Rf 0.81
Tr oc-P r otected Desm eth oxygelsed in e (45). To a solu-
tion of 44 (0.144 g, 0.349 mmol) in CH2Cl2 (0.50 mL) and dry
pyridine (1.00 mL) was added TrocCl (2,2,2-trichloroethyl
chloroformate) (0.057 mL, 0.419 mmol). After stirring for 4.5
h at room temperature, extra TrocCl (0.048 mL, 0.349 mmol)
was added, and the reaction mixture was stirred for another
18 h. Then the solvents were evaporated, and the residue was
chromatographed (EtOAc/petroleum ether 2:3, load with CH2-
Cl2) to give 45 (0.111 g, 67%) as a colorless oil; Rf 0.31 (EtOAc/
petroleum ether 2:3); [R]22 +126.5 (c ) 1.0, CHCl3); IR 2934,
(MeOH/CHCl3 1:3), [R]22 +120.4 (c ) 0.25, CHCl3); IR 3246,
D
D
2874, 1715, 1699, 1618, 1472, 1404, 1290, 1118; 1H NMR (400
MHz, benzene-d6, T ) 340 K) δ 7.37 (d, J ) 7.4 Hz, 1H), 6.97
(t, J ) 7.6 Hz, 1H), 6.88 (t, J ) 7.4 Hz, 1H), 6.35 (d, J ) 7.3
Hz, 1H), 4.71 (d, J ) 11.8 Hz, 1H), 4.60 (s, 1H), 4.21 (s, 1H),
3.86-3.82 (m, 2H), 3.56-3.50 (m, 2H), 3.08-3.00 (m, 1H), 2.74
(d, J ) 16.0 Hz, 1H), 2.62 (d, J ) 15.3 Hz, 1H), 2.52-2.48 (m,
1H), 2.00-1.87 (m, 2H), 1.66-1.58 (m, 1H), 0.82 (t, J ) 7.6
Hz, 3H); HRMS (FAB+) calculated for C21H24Cl3N2O4 (M +
H) 473.0801, found 473.0794. A useful 13C NMR spectrum
could not be obtained due to rotamers.
2926, 1707, 1616, 1465, 1435, 1327, 1221, 1044; 1H NMR (400
MHz) δ 7.39 (d, J ) 7.4 Hz, 1H), 7.29 (t, J ) 7.7 Hz, 1H), 7.12
(t, J ) 7.6 Hz, 1H), 6.94 (d, J ) 7.7 Hz, 1H), 4.33 (dd, J ) 4.1,
10.9 Hz, 1H), 4.24 (d, J ) 10.8 Hz, 1H), 4.00 (s, 3H), 3.74 (s,
1H, NH), 3.70 (s, 1H), 3.50 (d, J ) 6.9 Hz, 1H), 2.97 (s, 1H),
2.49-2.47 (m, 1H), 2.19 (d, J ) 6.9 Hz, 1H), 2.13-2.09 (m,
1H), 2.12 (d, J ) 3.6, 16.0 Hz, 1H), 2.02 (d, J ) 16.0 Hz, 1H),
1.96-1.87 (m, 1H), 1.85-1.78 (m, 1H), 1.76-1.69 (m, 1H), 1.00
(t, J ) 7.4 Hz, 3H); 13C NMR (100 MHz) δ 174.5, 137.9, 131.7,
128.2, 125.3, 123.7, 107.1, 74.6, 65.4, 63.8, 63.4, 59.6, 57.3, 41.6,
34.6, 33.8, 21.4, 21.2, 11.9; HRMS (FAB+) calculated for
C19H25N2O3 (M + H) 329.1865, found 329.1867.
In d olin e 46. To a solution of 45 (0.110 g, 0.232 mmol) in
THF (2 mL) was added at 0 °C BH3‚SMe2 (0.440 mL, 4.64
mmol) and the reaction mixture was heated at 70 °C for 18 h.
Then it was poured onto cold 10% aqueous NaHCO3 (10 mL)
and extracted with CH2Cl2 (3 × 10 mL). The organic layers
were dried (MgSO4) and concentrated. The residue was dis-
solved in MeOH (2 mL), and trimethylamine N-oxide dihydrate
(0.128 g, 1.16 mmol) was added. The reaction mixture was
heated at reflux for 2 h and then poured onto water (10 mL)
and extracted with CH2Cl2 (3 × 10 mL). The organic layers
were dried (MgSO4) and concentrated, and the residue was
chromatographed to give 46 (0.078 g, 73%) as a colorless oil;
Ack n ow led gm en t. We are very grateful to Profes-
sor H. Takayama of Chiba University (J apan) for
providing us with NMR spectra and a sample of (-)-
gelsedine. We thank Professor A. S. Kende of the
University of Rochester (USA) for sending us a copy of
the thesis of M. J . Luzzio. We thank Professor C.
Sza´ntay of the Technical University of Budapest (Hun-
gary) for drawing our attention to the article on the
N-demethylation of an indole. R. H. Balk is kindly
acknowledged for synthesizing considerable amounts of
starting materials.
Rf 0.29 (EtOAc/petroleum ether 2:3); [R]22 +110.9 (c ) 1.0,
D
benzene); IR 3363, 2933, 2861, 1713, 1604, 1398, 1292, 1121;
1H NMR (400 MHz, benzene-d6, T ) 340 K) δ 7.63 (d, J ) 7.4
Hz, 1H), 7.00 (t, J ) 7.4 Hz, 1H), 6.81 (d, J ) 7.4 Hz, 1H),
6.43 (d, J ) 7.7 Hz, 1H), 4.70 (d, J ) 12.0 Hz, 1H), 4.58 (d, J
) 12.0 Hz, 1H), 4.12 (s, 1H), 3.89 (d, J ) 5.3 Hz, 1H), 3.83 (s,
2H), 3.41 (d, J ) 8.9 Hz, 1H), 3.41 (s, 1H), 2.96 (br s, 1H), 2.83
(d, J ) 8.7 Hz, 1H), 2.84-2.76 (m, 1H), 2.68 (d, J ) 15.2 Hz,
1H), 1.94-1.92 (m, 1H), 1.81-1.76 (m, 2H), 1.69-1.57 (m, 2H),
1.45-1.37 (m, 1H), 0.74 (t, J ) 7.5 Hz, 3H); HRMS (FAB+)
calculated for C21H26Cl3N2O3 (M + H) 459.1009, found 459.1005.
A useful 13C NMR spectrum could not be obtained due to
rotamers.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and spectral data for compounds 18, 19, 23b-d , 24b-
1
d , 25b-d , 26b, 30, 33, 34, and 40 and H NMR and 13C NMR
spectra of many compounds including ent-gelsedine. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O001119T