Asymmetric synthesis of pyrrolidinoindolines. Application for the practical total synthesis of (-)-phenserine

Article abstract of DOI:10.1021/ja046690e

A versatile route to enantiopure 3,3-disubstituted oxindoles and 3a-substituted pyrrolidinoindolines is described in which diastereoselective dialkylation of enantiopure ditriflate 10 with oxindole enolates is the central step. These reactions are rare examples of alkylations of prostereogenic enolates with chiral sp3 electrophiles that proceed with high facial selectivity (10-20:1). The scope of this method is explored, and a model to rationalize the sense of stereoselection is advanced. This dialkylation chemistry was used to synthesize (-)-phenserine on a multigram scale in six steps and 43% overall yield from 5-methoxy-1,3-dimethyloxindole (27) and to complete a short formal total synthesis of (-)-physostigmine (2).

Full text of DOI:10.1021/ja046690e

Published on Web 10/12/2004
Asymmetric Synthesis of Pyrrolidinoindolines. Application for
the Practical Total Synthesis of (-)-Phenserine
Audris Huang, Jeremy J. Kodanko, and Larry E. Overman*
Contribution from the Department of Chemistry, 516 Rowland Hall, UniVersity of California,
IrVine, California 92697-2025
Received June 5, 2004; E-mail: leoverma@uci.edu
Abstract: A versatile route to enantiopure 3,3-disubstituted oxindoles and 3a-substituted pyrrolidinoindolines
is described in which diastereoselective dialkylation of enantiopure ditriflate 10 with oxindole enolates is
the central step. These reactions are rare examples of alkylations of prostereogenic enolates with chiral
sp3 electrophiles that proceed with high facial selectivity (10-20:1). The scope of this method is explored,
and a model to rationalize the sense of stereoselection is advanced. This dialkylation chemistry was used
to synthesize (-)-phenserine on a multigram scale in six steps and 43% overall yield from 5-methoxy-1,3-
dimethyloxindole (27) and to complete a short formal total synthesis of (-)-physostigmine (2).
Introduction
Pyrrolidinoindolines bearing a carbon substituent at C-3a of
the hexahydropyrrolo[2,3-b]indole ring system (1), exemplified
by (-)-physostigmine (2), (-)-flustramine B (4), and (-)-
pseudophrynaminol (5), have been obtained from a diverse array
of natural sources (Figure 1).1 For example, physostigmine was
isolated initially from the seeds of the African Calabar bean
Physostigma Venenosum,2 whereas flustramine B was found in
the marine bryozoan Flustra foliacea3 and (-)-pseudophrynami-
nol in the skin of an Australian frog, Pseudophryne coriacea.4
A variety of biological activities are associated with 3a-
substituted pyrrolidinoindolines.1 Physostigmine is a potent
reversible inhibitor of acetyl- and butyrylcholinesterase and is
employed clinically to treat glaucoma. For almost two decades,
physostigmine was evaluated in clinical trials for the symp-
tomatic treatment of Alzheimer’s disease.5 However, its clinical
use was compromised by its short duration of action, low
bioavailability, and narrow therapeutic window.6 (-)-Phenserine,
a synthetic congener of (-)-physostigmine, is physostigmine’s
successor in the clinic5a because of its more-favorable pharma-
cological profile7 and its ability to inhibit both acetylcholinest-
Figure 1. Representative hexahydropyrrolo[2,3-b]indole alkaloids with a
quaternary center at C-3a.
erase8 and â-amyloid plaque deposition in the brain.9 Phenserine
was advanced recently to phase III clinical trials.10
Several asymmetric methods for preparing hexahydropyrrolo-
[2,3-b]indoles have been documented,1b,11 many in the context
of total syntheses of (-)-physostigmine (2).12 One common
route involves the construction of an enantioenriched oxindole
intermediate containing a 2-aminoethyl side chain (or a progeni-
(1) For reviews, see: (a) Anthoni, U.; Christophersen, C.; Nielsen, P. H. In
Alkaloids: Chemical and Biological PerspectiVes; Pelletier, S. W., Ed.;
Wiley: New York, 1999; Vol. 13, pp 163-236. (b) Takano, S.; Ogasawara,
K. In The Alkaloids; Brossi, A., Ed.; Academic: San Diego, CA, 1989;
Vol. 36, pp 225-251.
(2) Jobst, J.; Hesse, O. Justus Liebigs Ann. Chem. 1864, 129, 115-118.
(3) (a) Carle, J. S.; Christophersen, C. J. Org. Chem. 1980, 45, 1586-1589.
(b) Carle, J. S.; Christophersen, C. J. Am. Chem. Soc. 1979, 101, 4012-
4013.
(8) Yu, Q.-S.; Holloway, H. W.; Flippen-Anderson, J. L.; Hoffman, B.; Brossi,
A.; Greig, N. H. J. Med. Chem. 2001, 44, 4062-4071.
(9) Shaw, K. T. Y.; Utsuki, T.; Rogers, J.; Yu, Q.-S.; Sambamurti, K.; Brossi,
A.; Ge, Y.-W.; Lahiri, D. K.; Greig, N. H. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 7605-7610.
(10) See Axonyx Home Page, http://www.axonyx.com/pipeline (accessed
December 2003), or Senior Citizens Information and News, http://
seniorjournal.com/NEWS/Alzheimer’s/04-08-02Phenserine.htm (accessed
December 2003).
(4) Spande, T. F.; Edwards, M. W.; Pannell, L. K.; Daly, J. W.; Erspamer, V.;
Melchiorri, P. J. Org. Chem. 1988, 53, 1222-1226.
(5) (a) Greig, N. H.; Pei, X.-F.; Soncrant, T. T.; Ingram, D. K.; Brossi, A.
Med. Res. ReV. 1995, 15, 3-31. (b) Sano, N.; Bell, K.; Marder, K.; Stricks,
L.; Stern, Y.; Mayeux, R. Clin. Neuropharmacol. 1993, 16, 61-69.
(6) Al-Jafari, A.; Kamal, M. A.; Greig, N. H.; Alhomida, A. S.; Perry, E. R.
Biochem. Biophys. Res. Commun. 1998, 248, 180-185 and references
therein.
(11) For 3a-substituted pyrrolidinoindolines (other than physostigmine), see: (a)
Austin, J. F.; Kim, S.-G.; Sinz, C. J.; Xiao, W.-J.; MacMillan, D. W. C.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5482-5487. (b) Kawasaki, T.;
Ogawa, A.; Takashima, Y.; Sakamoto, M. Tetrahedron Lett. 2003, 44,
1591-1593. (c) Fuji, K.; Kawabata, T.; Ohmori, T.; Shang, M.; Node, M.
Heterocycles 1998, 47, 951-964. (d) Sun, W. Y.; Sun, Y.; Tang, Y. C.;
Hu, J. Q. Synlett 1993, 5, 337-338.
(7) Sorbera, L. A.; Castaner, J. Drugs Future 2003, 28, 18-25.
9
10.1021/ja046690e CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 14043-14053
14043

A R T I C L E S
Huang et al.
Scheme 2
tor), which subsequently is cyclized under reductive conditions
to generate the tricyclic pyrrolidinoindoline ring system.13 One
of the more-notable methods of this type, introduced by
scientists at Hoechst-Roussel, involves alkylation of 5-methoxy-
1,3-dimethyloxindole with chloroacetonitrile in the presence of
a chiral phase-transfer catalyst.12i These conditions deliver the
3-cyanomethyl derivative in 78% ee; this intermediate, or the
derived amine, is then enriched to a high enantiopurity by
classical resolution.12i,14 An optimized version of this sequence
delivers (-)-phenserine in seven steps and 35% overall yield
from 5-methoxy-1,3-dimethyloxindole.12e,15
Some years ago, we described a catalytic asymmetric
construction of (-)- and (+)-physostigmine that utilized an
asymmetric intramolecular Heck reaction to assemble oxindole
8 (Scheme 1).12c,g In this case, the chiral oxindole aldehyde
product is generated in 95% ee and >99% ee after one
recrystallization. However, the preparation of Heck cyclization
precursor 7 was not straightforward and required the use of the
expensive coupling reagent (benzotriazol-1-yloxy)tris(dimethyl-
amino)phosphonium hexafluorophosphate (BOP).
enolate formed from a C3-substituted oxindole 9 (Scheme 2).
As dialkylation product 11 encodes 2 equiv of oxindole aldehyde
12, an attractive route to enantiopure pyrrolidinoindolines 13
appeared possible. Although three stereoisomeric dialkylation
products could result from the union of oxindole 9 and ditriflate
10, our experience with related reactions of this dielectrophile
with dienolates19 suggested that one C2-symmetric product
should predominate. In this paper, we describe the development
of this new asymmetric synthesis of 3,3-disubstituted oxindoles
and 3a-substituted pyrrolidinoindolines.
Scheme 1
Results
A. Synthesis of Oxindole Substrates. Our investigations into
the asymmetric dialkylation reaction began with the preparation
of a series of oxindoles containing a variety of substituents at
N1, C3, and C5. Several different synthetic sequences were
employed. The 1-alkyl-3-benzyloxindoles 1611c and 1720 were
synthesized by aldol condensation of benzaldehyde and oxindole
to furnish benzylidene oxindole 15 (Scheme 3).21 After N-
alkylation of this intermediate with benzyl bromide or methyl
iodide, reduction of the alkene double bond with zinc under
acidic conditions provided oxindoles 16 and 17 in good yields.
N-Benzyl-3-phenyloxindole (20) was prepared by adding
phenylmagnesium chloride to N-benzylisatin (18)22 and followed
by the Lewis acid-promoted reductive deoxygenation23 of
During an exploratory phase of recent investigations in our
laboratory directed at the total synthesis of the tris(pyrrolidi-
noindoline) alkaloid hodgkinsine,16 a new method for synthesiz-
ing pyrrolidinoindolines of high enantiopurity was conceived.17
The central step in this approach is the dialkylation of
enantiopure, tartrate-derived ditriflate 1018 with 2 equiv of an
Scheme 3
(12) For physostigmine, see: (a) Tanaka, K.; Taniguchi, T.; Ogasawara, K.
Tetrahedron Lett. 2001, 42, 1049-1052. (b) Kawahara, M.; Nishida, A.;
Nakagawa, M. Org. Lett. 2000, 2, 675-678. (c) Matsuura, T.; Overman,
L. E.; Poon, D. J. J. Am. Chem. Soc. 1998, 120, 6500-6503. (d) Node,
M.; Hao, X.; Nishide, K.; Fuji, K. Chem. Pharm. Bull. 1996, 44, 715-
719. (e) Pei, X.-F.; Yu, Q.-S.; Lu, B.-Y.; Greig, N. H.; Brossi, A.
Heterocycles 1996, 42, 229-236. (f) Pallavicini, M.; Valoti, E.; Villa, L.;
Resta, I. Tetrahedron: Asymmetry 1994, 5, 363-370. (g) Ashimori, A.;
Matsuura, T.; Overman, L. E.; Poon, D. J. J. Org. Chem. 1993, 58, 6949-
6951. (h) Marino, J. P.; Bogdan, S.; Kimura, K. J. Am. Chem. Soc. 1992,
114, 5566-5572. (i) Lee, T. B. K.; Wong, G. S. K. J. Org. Chem. 1991,
56, 872-875. (j) Takano, S.; Moriya, M.; Ogasawara, K. J. Org. Chem.
1991, 56, 5982-5984. (k) Node, M.; Itoh, A.; Masaki, Y.; Fuji, K.
Heterocycles 1991, 32, 1705-1707. (l) Takano, S.; Sato, T.; Inomata, K.;
Ogasawara, K. Heterocycles 1990, 31, 411-414. (m) Takano, S.; Moriya,
M.; Iwabuchi, Y.; Ogasawara, K. Chem. Lett. 1990, 109-112.
(13) This strategy traces back to the original Julian synthesis of (()-physos-
tigmine: (a) Julian, P. L.; Pikl, J. J. Am. Chem. Soc. 1935, 57, 563-566.
(b) Julian, P. L.; Pikl, J. J. Am. Chem. Soc. 1935, 57, 755-757.
(14) (a) Pei, X.-F.; Brossi, A. Heterocycles 1995, 41, 2823-2826. (b) Pallavicini,
M.; Valoti, E.; Villa, L.; Lianza, F. Tetrahedron: Asymmetry 1994, 5, 111-
116.
(18) Stang, P. J.; Dueber, T. E. Org. Synth. 1974, 54, 79-84.
(19) (a) Overman, L. E.; Larrow, J. F.; Stearns, B. A.; Vance, J. M. Angew.
Chem., Int. Ed. 2000, 39, 213-215. (b) Hoyt, S. B.; Overman, L. E. Org.
Lett. 2000, 2, 3241-3244.
(15) (a) Brossi, A.; Pei, X.-F.; Greig, N. H. Aust. J. Chem. 1996, 49, 171-181.
(b) Brzostowska, M.; He, X.-S.; Greig, N. H.; Rapoport, S. I.; Brossi, A.
Med. Chem. Res. 1992, 2, 238-246.
(20) Munusamy, R.; Dhathathreyan, K. S.; Balasubramanian, K. K.; Venkatacha-
lam, C. S. J. Chem. Soc., Perkin Trans. 2 2001, 7, 1154-1166.
(21) Villemin, D.; Martin, B. Synth. Commun. 1998, 28, 3201-3208.
(22) Garden, S. J.; Torres, J. C.; Da Silva, L. E.; Pinto, A. C. Synth. Commun.
1998, 28, 1679-1689.
(16) Kodanko, J. J.; Overman, L. E. Angew. Chem., Int. Ed. 2003, 42, 2528-
2531.
(17) Kodanko, J. J. Methods for Synthesis of the High Order Polypyrrolidino-
indolines and Total Synthesis of the Hodgkinsine Trispyrrolidinoindolines;
Ph.D. Dissertation, University of California, Irvine, CA, 2003.
(23) Orfanopoulos, M.; Smonou, I. Synth. Commun. 1988, 18, 833-839.
9
14044 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Asymmetric Synthesis of Pyrrolidinoindolines
A R T I C L E S
Scheme 4
B. Scope of Diastereoselective Dialkylation Reactions.
Initial Survey. Reactions of lithium enolates of oxindoles 16,
17, 20-22, and 24-27 with ditriflate 10 were examined to
explore the scope of the dialkylation reaction. The conditions
chosen for this initial survey involved adding 2.2 equiv of a
1.0 M THF solution of LHMDS to an equimolar amount of the
oxindole at -78 °C in THF containing a small amount of 1,3-
dimethyl-3,4,5,6-tetrahydro-1(1H)-pyrimidinone (DMPU), lead-
ing to a final ratio 98:2 THF/DMPU and an oxindole concen-
tration of 0.15 M. After 15-30 min, ditriflate 10 (1.0 equiv)
was added as a solid in one portion, and the solution was allowed
to warm to room temperature over an 18-24 h period in a -78
°C CO2/i-PrOH bath. The results of these experiments are
summarized in Table 1.
Generally, the dialkylation reactions proceeded cleanly to
generate a mixture of three stereoisomeric dialkylation products.
Stereoisomer ratios were determined by 1H NMR, GC, or HPLC
analysis of the crude product mixtures. In all but one instance,
one C2-symmetric product was formed predominantly; the C1-
symmetric isomer was produced in smaller amounts, and the
second C2-symmetric isomer was present in trace amounts only.
In cases where stereoselection was good, the crude reaction
product was filtered through a short column of silica gel and
the resulting mixture of stereoisomeric dialkylation products was
crystallized to provide the major isomer in pure form. In the
best case (entry 9), diastereoselection was 9:1 (major isomer:
sum of the two minor isomers), with the major dialkylation
product, 36a, being isolated in 70% yield as a high-melting
crystalline solid. To obtain sufficient amounts of the minor C2-
symmetric diastereomers for isolation and characterization, most
dialkylation reactions were repeated under conditions in which
the reaction took place with lower stereoselectivity: 2.2 equiv
of NaH, 2.2 equiv of oxindole in THF, and 1.0 equiv ditriflate
10 at 23 °C.
alcohol 1924 (Scheme 4). Finally, the synthesis of N-benzyl-3-
isopropyloxindole (22) was accomplished in two steps from
N-benzylisatin (18) via a sequential reaction with isopropyl-
triphenylphosphorane and hydrogen over Pd/C (Scheme 5).
Scheme 5
Several oxindoles (Figure 2) were prepared by methods
reported in the literature or by straightforward modifications
After examining the dialkylation reaction with oxindoles
bearing an alkyl or aryl substituent at C3, we explored the
possibility of carrying out the dialkylation of 10 with an oxindole
enolate lacking substitution at C3. However, reaction of the
lithium enolate of N-benzyloxindole (23) with ditriflate 10
yielded only the spiro-annulated product, 37 (eq 1).
Figure 2. Oxindoles 23-27 .
thereof. Wolff-Kishner reduction of N-benzylisatin (18) fur-
nished N-benzyloxindole (23).25 Using a sequence reported by
Fuji, appendage of a prenyl group to 23 was accomplished by
reacting oxindole 23 with 1-bromo-3-methyl-2-butene to afford
isopropylidene derivative 24.26 Oxindoles 2527 and 2628 were
synthesized from 3-thiomethyloxindole by the Gassman oxindole
synthesis.29 Finally, 5-methoxy-1,3-dimethyloxindole (27) was
prepared from commercially available 4-(methylamino)phenol
sulfate using the Brossi group’s modification of the original
Julian synthesis.13a,30
To explore the potential role of oxygens in the dioxolane ring
of dielectrophile 10, the dialkylation reaction of Table 1, entry
9, was repeated using cyclopentyl ditriflate 41 as the chiral
dielectrophile. The synthesis of racemic 41 was accomplished
from â-ketoester 3831 (Scheme 6). Bromination of â-ketoester
38, followed by a Favorski rearrangement,31 led to the formation
of trans-cyclopentanedicarboxylate (39).32 Reduction of this
diester with LiAlH4 gave diol 40, which was triflated to provide
(24) Kafka, S.; Klasek, A.; Kosmrlj, J. J. Org. Chem. 2001, 66, 6394-6399.
(25) Crestini, C.; Saladino, R. Synth. Commun. 1994, 24, 2835-2841.
(26) Lakshmaiah, G.; Kawabata, T.; Shang, M.; Fuji, K. J. Org. Chem. 1999,
64, 1699-1704.
(27) Giannangeli, M.; Baiocchi, L. J. Heterocycl. Chem. 1982, 19, 891-895.
(28) Shaughnessy, K. H.; Hamann, B. C.; Hartwig, J. F. J. Org. Chem. 1998,
63, 6546-6553.
(29) (a) An-naka, M.; Yasuda, K.; Yamada, M.; Kawai, A.; Takamura, N.;
Sugasawa, S.; Matsuoka, Y.; Iwata, H.; Fukushima, T. Heterocycles 1994,
39, 251-270. (b) Gassman, P. G.; van Bergen, T. J. J. Am. Chem. Soc.
1974, 96, 5508-5512.
(31) Paquette, L. A.; Farkas, E.; Galemmo, R. J. Org. Chem. 1981, 46, 5434-
5436.
(32) Reddy, S. H. K.; Chiba, K.; Sun, Y.; Moeller, K. D. Tetrahedron 2001,
57, 5183-5197.
(30) Pei, X.-F.; Greig, N. H.; Brossi, A. Heterocycles 1998, 49, 437-444.
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14045

A R T I C L E S
Huang et al.
Table 1. Dialkylation of Ditriflate 10 and Oxindole Lithium Enolates
entry
substrate
R1 (N1)
R2 (C3)
R3 (C5)
product
a:b:ca
a:b
yield (%)b
a
yield (%)c
a+b+c
1
2
3
4
5
6
7
8
9
16
20
21
22
25
24
17
26
27
Bn
Bn
Bn
Bn
Bn
Bn
Me
Me
Me
Bn
Ph
2-propenyl
CH(CH3)2
Me
H
H
H
H
H
H
H
H
28
29
30
31
32
33
34
35
36
70:22:8d
55:39:6d
47:44:9e
71:24:5e,f
81:16:3d
86:13:1e
87:12:1e
88:11:1e,f
90:9:1g
3.2:1
1.4:1
1.1:1
3.0:1
5.0:1
6.6:1
7.3:1
8.0:1
10.0:1
65
94
81
45
65
41
67
CH2CHdC(CH3)2
88
89
Bn
Me
Me
OMe
70
a Duplicate experiments unless otherwise indicated. Mean product ratio reported ((3%). b Yield of the major C2-symmetric diastereomer isolated by
recrystallization. c Yield of the mixture of diastereomers. d Determined by 1H NMR analysis. e Determined by HPLC analysis. f Result of one run. g Determined
by GC analysis.
Scheme 6
Scheme 7
ditriflate 41. Finally, reaction of 2.2 equiv of the lithium enolate
of 27 with ditriflate 41 furnished a mixture of three diastereo-
meric dialkylation products (42-44) in a ratio of 69:23:8 (major
C2:C1:minor C2; eq 2).
a 98:2 ratio of THF/DMPU (-78 to 23 °C) yielded two
alkylation products, 46 and 47, in a ratio of 3.1:1 (Scheme 7).
The absolute configuration of the quaternary stereocenter of
oxindole acetal 46 was determined by chemical correlation of
this major alkylation product to (S)-oxindole aldehyde 8.12c The
primary alcohol derivative of oxindole 8 was found to be
enantiopure (>99% ee, by HPLC analysis using a chiral
stationary phase).
Relative and Absolute Configurations of Dialkylation
Products. The relative and absolute configurations of the
In an effort to gain insight into individual selectivities of the
first and second alkylation steps, alkylation of the triflate
derivative of glycerol acetonide was examined. Reaction of 1.1
equiv of the lithium enolate of oxindole 27 with triflate 4533 in
(33) Cassel, S.; Debaig, C.; Benvegnu, T.; Chaimbault, P.; Lafosse, M.;
Plusquellec, D.; Rollin, P. Eur. J. Org. Chem. 2001, 5, 875-896.
9
14046 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Asymmetric Synthesis of Pyrrolidinoindolines
A R T I C L E S
Scheme 8
Scheme 10
The relative configuration of diprenyl dialkylation product
33a was established by its chemical correlation to that of
oxindole acid (S)-5826 (Scheme 11). As in previous cases, this
major C2-symmetric dialkylation product was converted in two
steps to oxindole aldehyde 57. Lindgren oxidation36 of this
intermediate then provided the (S)-oxindole acid 5826 in
moderate yield.
dialkylation products reported in Table 1 and eq 2 were
established in the following manner. Because of its lack of
symmetry, the C1-symmetric dialkylation product, whose qua-
ternary stereocenters have opposite absolute configurations, was
1
recognized readily by its diagnostic H and 13C NMR spectra.
The configuration of the newly formed quaternary stereocenters
of the major C2-symmetric dialkylation products was established
by either chemical correlation or single-crystal X-ray analysis
of a heavy atom derivative; the minor C2-symmetric diastere-
omers were assigned the opposite absolute configuration at their
quaternary stereocenters.
Scheme 11
The absolute configurations of the stereogenic quaternary
centers of 29a, 35a, and 36a were determined by chemical
correlation of these products to the configurations of oxindole
alcohols (R)-51,34 (S)-55,35 and (S)-48,12c whose absolute
configurations had been established previously. As summarized
in Schemes 8 and 9, these conversions were accomplished in
good overall yield by routine three-step sequences that were
identical to the sequence employed to form oxindole alcohol
(S)-48 from oxindole acetal 46.
Scheme 9
The relative and absolute configurations of the quaternary
stereocenters of C2-symmetric dialkylation product 28a were
established by X-ray crystallographic analysis of iodopyrroli-
dinoindoline 61 (Scheme 12). This derivative was accessed by
initially transforming 28a to aldehyde 59. Conversion of this
intermediate to the methylimine analogue, followed by reduction
with LiAlH4, provided the corresponding pyrrolidinoindoline.12c
Removal of the N-benzyl substituent of this product by reaction
with Na/NH3 gave pyrrolidinoindoline 60. Installation of an
iodine substituent at the 7-position of this intermediate was
achieved in three steps by ortho-lithiation-iodination of the tert-
butoxycarbonyl derivative.37 Single-crystal X-ray analysis of
product 61 then established the absolute configuration of the
quaternary stereogenic center as S.38
The absolute configuration of dialkylation product 32a was
assigned by chemical correlation to that of the tetramethyl
dialkylation product 35a (Scheme 10). This correlation was
accomplished by cleavage of the benzyl groups of 32a with
Na/NH3, followed by methylation of the oxindole nitrogens, to
yield tetramethyl dioxindole 35a.
The absolute configuration of the quaternary stereocenters
of dialkylation product 34a was established by its chemical
(36) (a) Kraus, G. A.; Taschner, M. J. J. Org. Chem. 1980, 45, 1175-1176. (b)
Lindgren, B. O.; Nilsson, T. Acta Chem. Scand. 1973, 27, 888-890.
(37) (a) Lebsack, A. D.; Link, J. T.; Overman, L. E.; Stearns, B. A. J. Am.
Chem. Soc. 2002, 124, 9008-9009. (b) Iwao, M.; Kuraishi, T. Org. Synth.
1996, 73, 85-93.
(38) Absolute configuration was assigned by analysis of the anomalous dispersion
using Rogers’s η parameter: Flack, H. D. Acta Crystallogr., Sect. A 1983,
39, 876-881.
(34) Dounay, A. B.; Hatanaka, K.; Kodanko, J. J.; Oestreich, M.; Overman, L.
E.; Pfeifer, L. A.; Weiss, M. M.; Ziller, J. W. J. Am. Chem. Soc. 2003,
125, 6261-6271.
(35) Ashimori, A.; Bachand, B.; Calter, M. A.; Govek, S. P.; Overman, L. E.;
Poon, D. J. J. Am. Chem. Soc. 1998, 120, 6488-6499.
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14047

A R T I C L E S
Huang et al.
Scheme 12
correlation with that of tetrabenzyl dialkylation product 28a.
This correlation was accomplished by removing both benzyl
groups, in low yield, by reaction of 28a with potassium tert-
butoxide and O239 and followed by methylation of the oxindole
nitrogens of the resulting product to give 34a (eq 3).40
Finally, the relative configuration of the major C2-symmetric
dialkylation product 42, derived from cyclopentyl ditriflate 41,
was verified by X-ray crystallographic analysis of dibromide
derivative 62. This analogue was accessed, with the tribromi-
nation product 63,41 by reaction of 42 with 3.3 equiv of bromine
in acetic acid at room temperature (eq 4).
Table 2. Effect of Solvent on the Reaction of the Lithium Enolate
of Oxindole 27 with Ditriflate 10a
entry
solvent
36a:36b:36cb
36a:36b
1
2
3
4
5
6
100% THF
89:11:1
90:9:1
90:9:1
86:14:1
84:15:1
72:24:4c
8.1:1
10.0:1
10.0:1
6.1:1
5.6:1
3.0:1
98:2 THF/DMPU
98:2 THF/HMPA
90:10 THF/HMPA
90:10 THF/DMPU
70:30 THF/DMPU
a Conditions: 2.2 equiv of LHMDS, -78 °C to room temperature.
b Determined by GC analysis of duplicate experiments. Mean product ratio
reported ((2%). c Result of one run.
Table 3. Effect of Temperature on the Reaction of the Lithium
Enolate of Oxindole 27 with Ditriflate 10a
entry
temperatureb
36a:36b:36cc
36a:36b
% conversion
1
2
3
4
5
6
-20 °C
76:21:4
82:16:3
86:13:1
88:11:1d
89:10:1
90:9:1
3.6:1
5.1:1
6.6:1
8.0:1
8.9:1
10.0:1
100
100
100
95
90
100
-40 °C
-60 °C
-70 °C
-78 °C
-78 °C to
room temperature
Optimization of the Dialkylation Reaction Conditions.
Although the results in Table 1 establish that diastereoselective
dialkylations of ditriflate 10 with lithium enolates of 1,3-
disubstituted oxindoles have reasonable scope, the reaction
conditions remained to be fully optimized. To this end, studies
were conducted to optimize the solvent, reaction temperature,
and base. The reaction that provided the highest diastereose-
lectivity, that between 5-methoxy-1,3-dimethyloxindole (27) and
ditriflate 10, was chosen for these studies.
First, the effect of the additives DMPU and HMPA on the
reaction was examined (Table 2). Compared to the original
conditions (entry 2), the addition of more than 2% of DMPU
as a cosolvent had a detrimental effect on diastereoselectivity
(entries 5 and 6). The highest diastereoselection was realized
a Conditions: 2.2 equiv of LHMDS, 98:2 THF/DMPU, 18-24 h.
b Unless otherwise indicated, reaction mixtures were maintained at this
temperature throughout the course of the reaction. c Determined by GC
analysis of duplicate experiments. Mean product ratio reported ((2%).
d Result of one run.
using 2% of either HMPA or DMPU (entries 2 and 3).
Diastereoselection was slightly lower in the absence of either
additive (entry 1). Because of the toxicity of HMPA, the original
solvent conditions (98:2 THF/DMPU) were deemed optimal.
The reaction temperature was investigated next (Table 3).
As expected, the dialkylation was found to be more stereose-
lective at lower temperatures, with the highest selectivity being
achieved at -78 °C (entry 5). Unfortunately, the reaction of
the lithium enolate of oxindole 27 with ditriflate 10 at this
temperature was only 90% complete after 24 h.
(39) Haddach, A. A.; Kelleman, A.; Deaton-Rewolinski, M. V. Tetrahedron
Lett. 2002, 43, 399-402.
(40) The quaternary stereocenters of 30a and 31a were assigned the R absolute
configurations by analogy to the relative and absolute configurations
established for compounds 28a, 29a, and 32a-36a.
(41) Tentatively assigned from the diagnostic peaks of the 1H and 13C NMR of
63.
Other bases were explored to determine whether a more
reactive enolate would lead to faster reaction rates without
erosion of the stereoselectivity. As summarized in Table 4,
9
14048 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Asymmetric Synthesis of Pyrrolidinoindolines
A R T I C L E S
Table 4. Effect of Counterion on the Reaction of the Enolates of
Oxindole 27 with Ditriflate 10a
amino)phenol sulfate. The central dialkylation reaction of
oxindole 27 and ditriflate 10 was conducted on a 20 g scale
using KHMDS as the base to furnish the crystalline C2-
symmetric dialkylation product 36a in 70% yield after recrys-
tallization. Deprotection of the acetonide of 36a, followed by
oxidative cleavage of the resultant vicinal diol, provided
aldehyde 8,12c thus completing a formal total synthesis of (-)-
physostigmine. The optical purity of 8 was assayed by reduction
of the aldehyde to give alcohol (S)-48, which was found to be
99% ee by HPLC analysis. Condensation of 8 with methylamine,
followed by in situ reduction of the crude imine with LiAlH4,12g
provided (-)-esermethole (64) in 83% overall yield from
dialkylation product 36a. Finally, using a slight modification
of the conditions reported by Brossi,43 this intermediate was
transformed in two steps and 75% yield to (-)-phenserine (3):
[R]D -73° (c 1.0, CHCl3); (literature)44 [R]D -74° (c 1.0,
CHCl3). This total synthesis of enantiopure (-)-phenserine was
completed in nine steps and 26% overall yield from com-
mercially available 4-(methylamino)phenol sulfate and in six
steps and 43% overall yield from oxindole 27.
entry
base
36a:36b:36cb
36a:36b
% conversion
1
2
3
4
LHMDS
KHMDS
KDA
90:9:1
10.0:1
8.2:1
7.3:1
3.2:1
90
100
100
98
90:11:<1
88:12:<1
75:24:2
NaHMDS
a Conditions: 2.2 equiv of base, 98:2 THF/DMPU, -78 °C. b Determined
by GC analysis of duplicate experiments. Mean product ratio reported
((2%).
Scheme 13
Using an identical sequence, ditriflate dialkylation products
28a and 32a were converted in good overall yields to the 3a-
substituted pyrrolidinoindolines 68 and 69 (Scheme 14). HPLC
Scheme 14
dialkylation of the sodium enolate with ditriflate 10 was less
selective,42 whereas the potassium enolate reacted with 10 with
a stereoselectivity only slightly less than that realized with the
corresponding lithium enolate (entries 1-4). Using potassium
bis(trimethylsilyl)amide (KHMDS) as the base, the reaction of
oxindole 27 and ditriflate 10 was complete in <3 h. Although
these results indicate that LHMDS produces the most-selective
reaction, the long reaction time (18-24 h) and the need to
slowly ramp the reaction temperature (-78 °C to room
temperature) to realize complete conversion of the starting
material, while retaining high stereoselectivity, make this
procedure cumbersome for large-scale reactions. For these
reasons, KHMDS was chosen as the optimal base for preparative
scale applications.
C. Enantioselective Total Synthesis of (-)-Phenserine and
Other 3a-Substituted Pyrrolidinoindolines. With the optimal
reaction conditions established, a large-scale asymmetric total
synthesis of (-)-phenserine (3) was undertaken (Scheme 13).
The Julian-Brossi procedure for preparing 5-methoxy-1,3-
dimethyloxindole (27)13a,30 was readily scaled and used to
prepare 80 g of this intermediate from inexpensive, 4-(methyl-
analysis of the derived alcohols (S)-66 and (S)-67 confirmed,
as in the case of pyrrolidinoindolines 60 and 64, that these
products were formed in high enantiopurity.
Discussion
Synthetic Implications. The dialkylation of tartrate-derived
ditriflate 10 with lithium or potassium enolates of oxindoles
containing substituents at C3 represents a new method for the
(43) Yu, Q.-S.; Brossi, A. Heterocycles 1988, 27, 745-750.
(44) Brossi, A.; Brzostowska, M.; Rapoport, S.; Greig, N.; He, X. S. Substituted
phenserines and phenylcarbamates of (-)-eseroline, (-)-N1-noreseroline,
and (-)-N1-benzylnoreseroline; as specific inhibitors of acetylcholinest-
erase. WO 9306105 A1, April 1, 1993.
(42) The lower selectivity observed in this reaction using sodium bis(trimeth-
ylsilyl)amide (NaHMDS) as the base is not general; stereoselection in the
reaction of 3-benzyloxindole (16) and ditriflate 10 was 70:22:8 (major C2:
C1:minor C2) using either NaHMDS, LHMDS, or KHMDS as the base.
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14049

A R T I C L E S
Huang et al.
Figure 3. Possible transition structures for the reaction of a 3-methyloxindole enolate with ditriflate 10. For clarity, the counterion of the enolate nucleophile
is omitted.
asymmetric construction of all-carbon quaternary stereocenters.45
One major C2-symmetric dialkylation product typically is
formed, which in most cases can be isolated in pure form by
simple recrystallization. The proclivity to form largely one
stereoisomeric product from these reactions is unusual as high
stereoselection is rarely seen in the union of prostereogenic
enolates with sp3-hybridized chiral electrophiles.19,46 This
method appears limited to the synthesis of oxindoles having
alkyl substituents at C3 since diastereoselection is low when
this substituent is phenyl or 2-propenyl (Table 1). The substituent
on the oxindole nitrogen is of less importance, although
diastereoselectivity is slightly higher when this group is methyl
rather than benzyl. Although it was not examined in a systematic
way, an electron-releasing substituent in the aromatic ring
appears to further increase diastereoselection. For example, the
highest dialkylation diastereoselectivity observed in this study
is seen with enolates derived from 5-methoxy-1,3-dimethylox-
indole (9:1, major isomer:sum of minor isomers). To our
surprise, the enolate counterion appears to play a minor role,
with high diastereoselection being realized with both lithium
and potassium oxindole enolates. As the potassium derivative
reacts more rapidly, this enolate is preferred for preparative scale
applications.
of triflic anhydride is a disadvantage of the method reported
here. For many other applications, this new route to 3a-
substituted pyrrolidinoindolines should be attractive because the
highly crystalline alkylated oxindole intermediate is produced
in enantiopure form, and either enantiomer will be available
depending which enantiomer of tartaric acid is employed.
Alkylation Diastereoselection. Obtaining high diastereose-
lection in the reaction of prostereogenic enolates and chiral sp3
electrophiles is rare;46 thus, it is important to consider the
possible origin of the diastereoselection realized in this study.
Since diastereoselectivity was nearly identical with lithium and
potassium enolates in 98:2 THF/DMPU, we believe open
transition states are involved. Moreover, stereoselection observed
in the dialkylation of the former nucleophile is only slightly
eroded by increasing the amount of HMPA or DMPU to 10%.
Another observation that influences our thinking is the stereo-
selectivity (69:23:8) seen in the dialkylation of cyclopentyl
ditriflate 41 with the lithium enolate of oxindole 27, which is
lower than that realized in the identical reaction of tartrate-
derived ditriflate 10 (90:9:1). We conclude that the oxygens of
the dioxolane ring of 10 are not involved in chelation; however,
they do play a role in organizing the transition state.
Any discussion of possible origins of diastereoselection in
the two alkylation steps must be speculative in the absence of
information regarding the orientation of the conformationally
flexible electrophiles during the alkylation events. One possible
arrangement would involve the enolate nucleophile approaching
the triflatomethyl group in a staggered arrangement in which
the C1-C2 σ-bond of the electrophile is oriented to position
the largest substituent at C2 anti to the approaching nucleophile.
This orientation, depicted in Figure 3 for the first alkylation
step, should be favored by stereoelectronic factors as it places
the electron-releasing C2-C3 σ-bond coplanar to the p orbital
of a trigonal bipyramidal SN2 transition structure.47,48 The six
possible transition structures possessing this arrangement are
Using a diastereoselective dialkylation as the central step,
(-)-esermethole (64) was prepared on a multigram scale from
readily available 5-methoxy-1,3-dimethyloxindole (27) in four
steps and 58% overall yield. This sequence is shorter and is
accomplished in higher yield than that employed in the
asymmetric synthesis of (-)-phenserine reported by Brossi,12e,15
wherein catalytic asymmetric cyanomethylation of oxindole 27
is the key step.12i However, for large-scale synthesis, the cost
(45) For recent reviews of asymmetric synthesis of quaternary carbon centers,
see: (a) Douglas, C. J.; Overman, L. E. Proc. Natl. Acad. Sci. U.S.A. 2004,
101, 5363-5367. (b) Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int.
Ed. 1998, 37, 388-401. (c) Fuji, K. Chem. ReV. 1993, 93, 2037-2066.
9
14050 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Asymmetric Synthesis of Pyrrolidinoindolines
A R T I C L E S
Conclusion
depicted in Figure 3: transition structures A-C arising from
the reaction of the enolate Si face and transition structures D-F
from the reaction of the Re face. Steric considerations would
appear to favor structures C and F by placing the largest
substituent of the nucleophile, the oxindole aryl group, between
the two hydrogens. Of these two transition structures, C should
have lower energy because F would be destabilized by interac-
tions (electrostatic or dipole) between the oxygen of the enolate
and the proximal oxygen of the dioxolane ring of the electro-
phile. Alternatively, if minimizing repulsion between the dipoles
of the enolate and the C2-O σ-bond is of primary importance,
conformations A and D would be the two to consider. Of these,
A should be favored as the two dipoles directly oppose each
other. Both of these models predict alkylation from the Si face
of the enolate, as observed.49
Dialkylation of enantiopure ditriflate 10 with oxindole
enolates, having alkyl substituents at C3, occurs with high
diastereoselection to give largely one C2-symmetric dialkylation
product. This product is readily cleaved to give 2 equiv of the
corresponding enantiopure 3-alkyl-3-(2-oxoethyl)oxindole, which,
in one double reductive amination step, can be converted to
enantiopure 3a-substituted pyrrolidinoindolines. Using this di-
alkylation chemistry, enantiopure (-)-phenserine was prepared
on a multigram scale in six steps and 43% overall yield from
5-methoxy-1,3-dimethyloxindole (27). A model that rationalizes
the observed preference for alkylation to occur preferentially
on the Si face of the oxindole enolate is advanced. The alkylation
reactions that are the subject of this account are atypical
examples of prostereogenic enolates reacting with high facial
selectivity (10-20:1) with chiral sp3 electrophiles.
If both alkylation steps took place with a Si:Re facial
preference of n:1, the three diastereomeric dialkylation products
would be produced in a ratio of n2:2n:1 (major C2-symmetric
isomer:C1-symmetric isomer:minor C2-symmetric isomer). In
this scenario, the ratio of the major C2-symmetric isomer:C1-
symmetric isomer observed in the most-selective dialkylation
reactions (entries 7-9 of Table 1) would arise from Si:Re
selectivities of 15, 16, and 20, respectively. This analysis predicts
that the minor C2-symmetric isomer would have been produced
in somewhat lower amounts (<0.3%) than observed (∼1.0%
in the reactions of 7-9 of Table 1), although this difference is
close to the error in measuring the amount of this minor product.
The spectator trifloxymethyl substituent of electrophile 10 must
contribute to the high facial diastereoselection observed in
reactions of this electrophile with oxindole enolates since low
diastereoselection was observed in the alkylation of glycidol
acetonide monotriflate 45 with the lithium enolate of oxindole
27 (3.1:1, Scheme 7).50,51
Experimental Section52,53
General Procedure for Dialkylation of Lithium Oxindole Eno-
lates with Ditriflate 10. Preparation of 32a-c. A solution of 25 (1.50
g, 6.33 mmol), DMPU (0.85 mL), and THF (41.7 mL) was cooled to
-78 °C in a dry ice/i-PrOH bath and was deoxygenated by vigorously
sparging with argon for 30 min. A 1 M solution of LHMDS in THF
(6.3 mL, 6.3 mmol) was added dropwise. After 20 min, ditriflate 10
(1.23 g, 2.88 mmol) was added as a solid. The reaction mixture was
covered with aluminum foil and allowed to warm slowly to room
temperature overnight. The reaction mixture was quenched with
saturated aqueous NH4Cl (23 mL) and diluted with 20 mL of 1:1
benzene/EtOAc. The layers were separated, and the aqueous phase was
extracted with EtOAc (2 × 20 mL). The combined organic layers were
dried over Na2SO4, filtered, and concentrated to afford an orange
residue. Purification of this residue by silica gel chromatography (eluant,
10% EtOAc/toluene-100% EtOAc) yielded a residue consisting of a
mixture of three diastereomers. Recrystallization from hot EtOH
afforded the major C2-symmetric product 32a as a colorless solid (1.12
g, 65%). The mother liquor was concentrated to yield a colorless solid.
A small portion of this solid was purified further by HPLC [Phenom-
enex, Luna C-18 (2), 5 µm, 250 × 21.2 mm, column temperature 23
°C, 75% MeOH in H2O, flow rate 16 mL/min, UV detection at 254
nm, tr ) 46 min (major C2), 62 min (C1), and 71 min (minor C2)] to
afford pure analytical samples of the C1-symmetric product 32b (7.0
mg) and the minor C2-symmetric product 32c (1.3 mg).
(46) (a) A recent survey of stereoselective C-C bond-formation cites no
examples of alkylation reactions with chiral electrophiles: Methoden Org.
Chem. (Houben-Weyl), 4th Ed., 1996; Vol. E21-E22, pp 1077-1118
(Workbook Edition). (b) For examples of selectivity in the range of 7-10:1
in alkylations of lactate triflates, see: Vedejs, E.; Daugulis, O. J. Am. Chem.
Soc. 1999, 121, 5813-5814.
(47) (a) The rate of SN2 reactions is typically diminished by the incorporation
of electron-withdrawing substituents at the â-carbon;48 in the case at hand,
the electron-withdrawing substituent would be oxygen. This deactivation
should have the stereoelectronic component we propose, although we are
unaware of specific discussions of this issue as it pertains to SN2 reactions.
(b) Conformations that place the electron-releasing C-H σ-bond coplanar
to the p orbital of the trigonal bipyramidal SN2 transition structure would
be favored also by this stereoelectronic feature.
Major C2-symmetric product, 32a: [R]27 -55°, [R]27 -57°,
589
577
[R]27 -65°, [R]27 -103°, [R]27 -117°; mp 147-149 °C; H
1
546
435
405
NMR (500 MHz, CDCl3) δ 7.33-7.23 (m, 10H), 7.15-7.12 (m, 4H),
7.02 (t, 2H, J ) 7.6 Hz), 6.65 (d, 2H, J ) 7.7 Hz), 4.99 (d, 2H, J )
15.9 Hz), 4.74 (d, 2H, J ) 15.8 Hz), 3.31 (m, 2H), 2.09 (dd, 2H, J )
14.4, 9.6 Hz), 1.75 (d, 2H, J ) 12.5 Hz), 1.37 (s, 6H), 1.09 (s, 6H);
13C NMR (125 MHz, CDCl3) δ 180.3, 142.7, 136.2, 133.0, 128.6, 127.8,
127.4, 127.2, 122.8, 122.2, 109.2, 108.9, 77.7, 46.6, 43.9, 40.5, 26.9,
25.2; IR (thin film) 3056, 2968, 2929, 1711, 1611, 1490, 1372, 754
cm-1; HRMS (ESI) m/z calcd for C39H40N2O4 (M + Na)+ 623.2886,
found 623.2885. Anal. Calcd for C39H40N2O4: C, 77.97; H, 6.71; N,
4.66. Found: C, 78.02; H, 6.71; N, 4.78.
(48) Streitwieser, A. SolVolytic Displacement Reactions; McGraw-Hill: New
York, 1962; pp 16-20.
(49) The erosion of stereoselection observed in the reaction of the lithium enolate
of oxindole 27 with ditriflate 10 as the DMPU content of the solvent is
increased from 2 to 30% (Table 2) would be consistent with the importance
of electrostatic or dipole interactions in organizing the alkylation transition
state.
(50) If Si:Re selectivity in the first step was indeed 3.1:1 and the second
alkylation occurred exclusively from the Si face, the major C2-symmetric
isomer and the C1-symmetric isomer would be produced in a 3.1:1 ratio;
if selectivity in the second step was also 3.1:1, this ratio would be only
1.6:1. To realize the observed 9:1 ratio of the major C2-symmetric isomer:
C1-symmetric isomer, the major monoalkylation product would have to
react with high Si selectivity and the minor monoalkylation product would
have to partition in the alternate sense (i.e., react preferentially from the
Re face). However, this scenario is not consistent with the observed results
as it would lead to the formation of significant amounts of the second
(minor) C2-symmetric isomer.
C1-Symmetric product, 32b: [R]27 +10°, [R]27 +13°, [R]27
589
1
577
546
+16°, [R]27 +41°, [R]27 +56°; H NMR (500 MHz, CDCl3) δ
435
405
7.34-7.24 (m, 10H), 7.16 (t, 2H, J ) 6.5 Hz), 7.12 (app t, 2H, J )
7.7 Hz), 7.01 (ddd, 1H, J ) 7.6, 7.6, 1.0 Hz), 6.96 (ddd, 1H, J ) 7.6,
7.6, 1.0 Hz), 6.71 (d, 1H, J ) 7.8 Hz), 6.63 (d, 1H, J ) 7.8 Hz), 5.03
(51) It is not surprising that this substituent would effect diastereoselection in
the initial alkylation step and the oxindolemethyl substituent in the second.
Interactions of these groups with the triflate leaving group could modulate
the alignment of the C1-C2 σ-bond of the electrophile and/or the
conformation of the dioxolane ring, with the latter orienting the dimethyl
group in a way that influences the approach of the nucleophile.
(52) General experimental details have been described: Ando, S.; Minor, K.
P.; Overman, L. E. J. Org. Chem. 1997, 62, 6379-6387.
(53) Crystallographic data for the structures reported in this paper have been
deposited with the Cambridge Crystallographic Data Centre as a supple-
mentary publication: CCDC-241957 (61) and CCDC-241958 (62).
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14051

A R T I C L E S
Huang et al.
(d, 1H, J ) 15.8 Hz), 4.95 (d, 1H, J ) 15.8 Hz), 4.91 (m, 1H), 4.70
(d, 1H, J ) 15.9 Hz), 3.45 (ddd, 1H, J ) 8.3, 8.3, 3.0 Hz), 3.26 (ddd,
1H, J ) 10.1, 7.9, 2.2 Hz), 2.11 (dd, 1H, J ) 14.2, 10.1 Hz), 1.95 (dd,
1H, J ) 14.3, 8.6 Hz), 1.85 (ddd, 2H, J ) 14.0, 9.7, 3.0 Hz), 1.39 (d,
6H, J ) 16.5 Hz), 1.10 (s, 3H), 0.97 (s, 3H); 13C NMR (125 MHz,
CDCl3) δ 180.6, 180.3, 142.7, 142.0, 136.2, 136.0, 133.6, 133.0, 128.8,
128.6, 127.7, 127.6, 127.5, 127.3, 127.2, 123.7, 122.9, 122.3, 122.2,
109.1, 108.9 (2), 78.1, 78.0, 47.2, 46.7, 43.9, 43.7, 40.5, 39.7, 26.9,
26.8, 25.4, 24.8; IR (thin film) 3056, 2968, 2927, 1710, 1611, 1488,
1372, 1179, 753 cm-1; HRMS (ESI) m/z calcd for C39H40N2O4 (M +
Na)+ 623.2886, found 623.2876.
Minor C2-symmetric product, 32c: 1H NMR (500 MHz, CDCl3) δ
7.31-7.24 (m, 10H), 7.19 (app d, 2H, J ) 7.4 Hz), 7.10 (ddd, 2H, J
) 7.8, 7.8, 1.3 Hz), 6.97 (ddd, 2H, J ) 7.6, 7.6, 1.0 Hz), 6.67 (d, 2H,
J ) 7.4 Hz), 4.88 (d, 2H, J ) 15.7 Hz), 4.81 (d, 2H, J ) 16.0 Hz),
3.53 (m, 2H), 1.95 (dd, 2H, J ) 14.2, 8.7 Hz), 1.85 (dd, 2H, J ) 14.4,
2.4 Hz), 1.42 (s, 6H), 1.02 (s, 6H); 13C NMR (125 MHz, CDCl3) δ
180.5, 142.0, 136.1, 133.8, 128.7, 127.5, 127.2, 123.7, 122.2, 108.9
(2), 78.4, 47.1, 43.6, 39.7, 29.7, 27.0, 24.6; IR (thin film) 3047, 2917,
1711, 1613, 1490, 1378, 1181, 741 cm-1; HRMS (ESI) m/z calcd for
C39H40N2O4 (M + Na)+ 623.2886, found 623.2882.
mixture was purified further by HPLC [Alltech, Altima silica, 5 mm,
250 × 21.2 mm, column temperature 23 °C, 5% IPA/hexanes, flow
rate 10-12 mL/min, UV detection at 254 nm, tr ) 39 min (C1), 45
min (major C2), and 53 min (minor C2)] to afford pure analytical
samples of the C1-symmetric diastereomer 36b (1 mg) and the minor
C2-symmetric diastereomer 36c (1 mg).
C1-Symmetric product, 36b: 1H NMR (500 MHz, CDCl3) δ 6.84
(dd, 2H, J ) 10.8, 2.4 Hz), 6.78 (ddd, 2H, J ) 8.5, 6.1, 2.5 Hz), 6.73
(d, 1H, J ) 8.4 Hz), 6.69 (d, 1H, J ) 8.4 Hz), 3.79 (d, 6H, J ) 12.5
Hz), 3.34 (ddd, 1H, J ) 8.0, 8.0, 3.4 Hz), 3.27 (m, 1H), 3.20 (s, 3H),
3.10 (s, 3H), 1.91 (d, 1H, J ) 14.2 Hz), 1.89 (dd, 1H, J ) 14.2, 1.5
Hz), 1.71 (m, 2H), 1.35 (s, 3H), 1.30 (s, 3H), 1.14 (s, 3H), 1.04 (s,
3H); 13C NMR (125 MHz, CDCl3) δ 180.2 (2), 155.9, 155.7, 137.1,
136.4, 134.6, 134.5, 112.1 (2), 111.3, 110.6, 108.3, 108.2, 108.0, 78.1,
77.8, 55.8, 55.8, 47.5, 47.0, 40.3, 39.5, 26.9 (2), 26.3, 24.3, 23.9; IR
(thin film) 3054, 2933, 1702, 1600, 1495, 1291, 1036, 803 cm-1; HRMS
(ESI) m/z calcd for C29H36N2O6 (M + Na)+ 531.2471, found 531.2465.
Minor C2-symmetric product, 36c: 1H NMR (500 MHz, CDCl3) δ
6.83 (d, 2H, J ) 2.5 Hz), 6.77 (dd, 2H, J ) 8.4, 2.5 Hz), 6.69 (d, 2H,
J ) 8.4 Hz), 3.78 (s, 6H), 3.49 (m, 2H), 3.12 (s, 6H), 1.88 (dd, 2H, J
) 14.3, 8.8 Hz), 1.66 (dd, 2H, J ) 14.2, 2.4 Hz), 1.35 (s, 6H), 1.16 (s,
6H); 13C NMR (125 MHz, CDCl3) δ 180.2, 155.9, 136.5, 134.7, 112.1,
111.5, 108.7, 108.1, 78.3, 55.8, 47.4, 39.3, 27.1, 26.2, 24.0; IR (thin
film) 2933, 1706, 1497, 1291, 1119, 1040, 805 cm-1; HRMS (ESI)
m/z calcd for C29H36N2O6 (M + Na)+ 531.2471, found 531.2487.
(S)-3-(2-Oxoethyl)-1,2-dihydro-5-methoxy-1,3-dimethyl-2-oxo-
[3H]indole ((S)-8). p-Toluenesulfonic acid monohydrate (25.1 g, 132
mmol) and H2O (50 mL) were added to a stirred solution of 36a (17.5
g, 34.4 mmol) and MeOH (403 mL). The reaction mixture was heated
at reflux for 21 h and then allowed to cool to room temperature.
Saturated aqueous NaHCO3 (130 mL) was added; the mixture was
concentrated to remove the organic solvent, and EtOAc (300 mL) and
saturated aqueous NaHCO3 (130 mL) were added. The layers were
separated, and the aqueous phase was extracted with EtOAc (8 × 200
mL). The combined organic layers were dried over Na2SO4, filtered,
and concentrated to afford a yellow solid. Recrystallization of this solid
from hot EtOH (13 mL/1 g) gave the corresponding diol as colorless
crystals (9.62 g). An additional crop (6.49 g) was obtained by
concentrating the mother liquor and purifying the residual solid by silica
gel chromatography (2-10% MeOH in CH2Cl2). The two crops were
combined and carried forward.
A mixture of this diol, NaIO4 (110 g, 513 mmol), THF (370 mL),
and H2O (193 mL) was stirred at room temperature for 22 h and then
concentrated to remove the THF. The resulting mixture was filtered,
and the filter cake was washed with Et2O (1 L). Water (500 mL) was
added to the filtrate, and the layers were separated. The aqueous phase
was extracted with Et2O (14 × 450 mL). The combined organic layers
were dried over Na2SO4, filtered, and concentrated to afford (S)-8 as
an orange solid (14.7 g, 92% over two steps). Spectral data for this
product were consistent with those previously reported.12c
(-)-Phenserine (3). A slight modification of Brossi’s procedure was
employed.43 Boron tribromide (3.0 g, 12.9 mmol) was added dropwise
over 15 min to a stirred solution of (-)-esermethole (64) (3.0 g, 12.9
mmol) and CH2Cl2 (57 mL). The reaction mixture was maintained at
room temperature for 3 h, and then the reaction mixture was
concentrated to afford a foam. This foam was cooled to 0 °C and
dissolved in dry MeOH (50 mL); the resulting solution was stirred for
10 min, and the solution was concentrated. This step was repeated one
time. The resulting residue was dissolved in H2O (60 mL), basified
with saturated aqueous NaHCO3 (40 mL), and extracted with Et2O (15
× 200 mL). The organic layers were combined, dried over Na2SO4,
filtered, and concentrated to afford (3aS,8aR)-1,3a,8-trimethyl-1,2,3,-
3a,8,8a-hexahydropyrrolo[2,3-b]indol-5-ol as a light orange solid (2.56
g, 91%), which was carried forward without further purification.
A 60% dispersion of NaH (46.8 mg, 1.17 mmol) was added to a
solution of (3aS,8aR)-1,3a,8-trimethyl-1,2,3,3a,8,8a-hexahydropyrrolo-
(3S)-5-Methoxy-3-[((4S,5S)-5-{[(3S)-5-methoxy-1,3-dimethyl-2-
oxo-1,3-dihydro-2H-indol-3-yl]methyl}-2,2-dimethyl-1,3-dioxolan-
4-yl)methyl]-1,3-dimethyl-1,3-dihydro-2H-indol-2-one (36a). A so-
lution of oxindole 27 (20.0 g, 105 mmol), DMPU (14.2 mL), and THF
(700 mL) was cooled to -78 °C in a cooling bath and deoxygenated
by vigorously sparging with argon for 60 min. Solid KHMDS (22.0 g,
105 mmol) was added in one portion. After 70 min, a solution of
ditriflate 10 (22.3 g, 52.3 mmol) and THF (50 mL) was added dropwise
over 100 min with a syringe pump, being careful to keep the reaction
mixture temperature below -74 °C. The reaction mixture was
maintained at -78 °C for 100 min and then allowed to warm to room
temperature. A solution of 3% acetic acid in THF (100 mL) was added.
After 20 min, EtOAc (800 mL) and H2O (800 mL) were added, and
the layers were separated. The aqueous phase was extracted with EtOAc
(1 × 800 mL and 1 × 300 mL). The combined organic layers were
washed with a solution of saturated aqueous NaHCO3 (1 × 500 mL)
and a solution of saturated aqueous NH4Cl (1 × 500 mL), dried over
Na2SO4, filtered, and concentrated to afford a yellow solid. Recrystal-
lization of this solid from hot ethanol (8 mL/1 g) yielded 36a as
colorless crystals. The mother liquor was concentrated, and the resulting
solid was taken up in Et2O and filtered to remove residual DMPU.
The filtered solid was recrystallized from hot ethanol to afford two
additional crops (total yield of 18.7 g, 70%): [R]27 -33°, [R]27
589
577
-35°, [R]27 -38°, [R]27 -51°, [R]27 -50°; mp 209-211 °C;
546
435
405
1H NMR (500 MHz, CDCl3) δ 6.81, 6.77, 6.69, 3.80, 3.22, 3.10, 1.92,
1.64, 1.31, 1.04; 13C NMR (125 MHz, CDCl3) δ 180.1, 155.7, 137.2,
134.7, 111.8, 110.9, 108.1, 108.0, 77.6, 55.8, 47.0, 40.3, 26.8, 26.3,
23.8; IR (thin film) 3058, 2933, 1708, 1495, 1293, 1040, 803 cm-1
;
HRMS (ESI) m/z calcd for C29H36N2O6 (M + Na)+ 531.2471, found
531.2478. Anal. Calcd for C29H36N2O6: C, 68.48; H, 7.13; N, 5.51.
Found: C, 68.59; H, 7.12; N, 5.58.
Isolation of Minor Dialkylation Products 36b and 36c. A solution
of oxindole 27 (300 mg, 1.57 mmol) and THF (10.4 mL) was cooled
to 0 °C and deoxygenated by vigorously sparging with argon for 50
min. A 60% dispersion of NaH (62.8 mg, 1.57 mmol) was added to
this solution. After 15 min, ditriflate 10 (304 mg, 0.714 mmol) was
added as a solid, and the reaction mixture was allowed to warm to
room temperature and stirred overnight. The reaction mixture was
cooled to 0 °C and quenched with saturated aqueous NH4Cl (12 mL)
and diluted with EtOAc (16 mL). The layers were separated, and the
aqueous phase was extracted with EtOAc (2 × 16 mL). The combined
organic layers were dried over Na2SO4, filtered, and concentrated to
yield a yellow residue. Purification of this residue by silica gel
chromatography (eluant, 60-80% EtOAc/hexanes) afforded a solid
consisting of a mixture of three diastereomers. A small amount of this
9
14052 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Asymmetric Synthesis of Pyrrolidinoindolines
A R T I C L E S
[2,3-b]indol-5-ol (2.56 g, 11.7 mmol) and THF (229 mL). This mixture
was stirred for 10 min at room temperature, and then phenyl isocyanate
(1.53 mL, 14.0 mmol) was added. After 4.5 h, the solution was
concentrated. EtOAc (170 mL) and saturated aqueous NaHCO3 (170
mL) were added; the layers were separated, and the aqueous phase
was extracted with EtOAc (2 × 170 mL). The organic layers were
combined, dried over Na2SO4, filtered, and concentrated. Purification
of the crude product by silica gel chromatography (eluant, 1-10%
MeOH in CH2Cl2) furnished (-)-phenserine (3) as a pink foam (3.22
g, 82%): [R]27 -73.4°, [R]27 -76.4°, [R]27 -85.2°, [R]27
Acknowledgment. This research was supported by the
National Institute of General Medical Sciences (GM-30859).
A.H. was supported, in part, by an Allergan Graduate Fellow-
ship. NMR, mass spectra, and X-ray diffraction data were
determined at the University of California, Irvine, with instru-
mentation purchased with the assistance of the NSF and NIH
shared instruments programs.
Supporting Information Available: Experimental procedures
for the preparation of 15-17, 19-22, 24-26, 28a-c, 29a-c,
30a-c, 31a-c, 33a-c, 34a-c, 35a-c, 37, 39, 41-44, 46-
63, and 65-69; copies of 1H and 13C NMR spectra for 21, 22,
28b,c, 29a-c, 30a-c, 31a-c, 32b,c, 33a-c, 34a-c, 35a-c,
36b,c, 37, 41-44, 46, 47, 49-57, 59, 62, 63, 65-67, and 69;
HPLC traces used to determine the enantiopurity of 48, 66, and
67; and CIF files for single-crystal X-ray analyses of 61 and
62. This material is available free of charge via the Internet at
http://pubs.acs.org.
589
577
546
435
-141°, [R]27405 -162°; mp 151-152 °C; 1H NMR (500 MHz, CDCl3)
δ 7.44 (d, 2H, J ) 7.8 Hz), 7.34-7.31 (m, 2H), 7.09 (t, 1H, J ) 7.4
Hz), 6.87 (dd, 2H, J ) 8.4, 2.4 Hz), 6.82 (d, 1H, J ) 2.4 Hz), 6.36 (d,
1H, J ) 8.4 Hz), 4.13 (s, 1H), 2.93 (s, 3H), 2.74-2.71 (m, 1H), 2.68-
2.63 (m, 1H), 2.55 (s, 3H), 1.98-1.94 (m, 2H), 1.44 (s, 3H); 13C NMR
(125 MHz, CDCl3) δ 152.5, 149.8, 142.5, 137.6 (2), 129.1, 123.7, 120.4,
118.6, 116.1, 106.5, 98.0, 53.2, 52.6, 40.8, 38.4, 36.9, 27.2; IR (thin
film) 2960, 1721, 1602, 1495, 1198, 1009 cm-1. Anal. Calcd for
C20H23N3O2: C, 71.19; H, 6.87; N, 12.45. Found: C, 71.02; H, 6.95;
N, 12.36.
JA046690E
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14053