Application of the 2-azaallyl anion cycloaddition method to an enantioselective total synthesis of (+)-coccinine

Article abstract of DOI:10.1002/(SICI)1521-3773(19980703)37:12<1724::AID-ANIE1724>3.0.CO;2-8

A highly functionalized perhydroindole is formed by the intramolecular [π4s+π2s] cycloaddition of a 2-azaallyl anion with a vinyl sulfide [Eq. (a)]. This is the key step in the total synthesis of (+)-coccinine, the enantiomer of the Amaryllidaceae alkaloi

Full text of DOI:10.1002/(SICI)1521-3773(19980703)37:12<1724::AID-ANIE1724>3.0.CO;2-8

COMMUNICATIONS
Li
+
Application of the 2-Azaallyl Anion Cyclo-
addition Method to an Enantioselective Total
Synthesis of (‡)-Coccinine**
Ar
X
Ar
RO
RO
RO
RO
O
Ar
X
13
X
N
N
H
RO
OP
10
William H. Pearson* and Brian W. Lian
11
12
Figure 2. Synthetic strategy for montanine alkaloids.
The montanine-type alkaloids (e.g., 1 ± 8, Figure 1) form a
small class of Amaryllidaceae alkaloids that share the 5,11-
methanomorphanthridine ring system[1] and differ primarily
anion 11 should be accessible from the epoxide 12 and the
vinyllithium compound 13.[8] The diastereoselectivity of the
intramolecular 2-azaallyl anion cycloaddition is still relatively
unexplored.[9] On the basis of molecular models, we proposed
that anion 11 would produce the isomer of 10 required for a
synthesis of ( )-montanine (1). As revealed below, an
alternate stereochemical outcome which ultimately allowed
access to the alternative enantioseries was observed exper-
imentally.
Our synthesis began with the assembly of the vinyllithium
precursor 17 (Scheme 1). Piperonal was converted into the
Br
PhS
a
b
O
O
Br
Ar
Figure 1. Montanine-type alkaloids.
15
14
in the configurations at the stereocenters and the type of
oxygen substitution at C-2 and C-3. The recent isolation of
(‡)-montabuphine, assigned structure 8,[2] suggests that both
enantiomers of the methanomorphanthridine ring may exist
in nature. While Amaryllidaceae alkaloids have in general
been the subject of much synthetic effort, the 5,11-methano-
morphanthridine class alkaloids has attracted much less
attention. Overman and Shim reported the total syntheses
of ( Æ )- and ( )-pancracine (3),[3] and total syntheses of
racemic montanine (1), coccinine (2), pancracine (3), bruns-
vigine (6), and O-acetylmontanine (7) were published by
Hoshino et al.[4] Most recently, Jin and Weinreb reported the
enantioselective syntheses of ( )-coccinine (2) and ( )-
pancracine (3).[5] We now describe an enantioselective total
synthesis of (‡)-coccinine (9), the nonnatural enantiomer of
( )-coccinine (2), which uses our 2-azaallyl anion cyclo-
addition methodology[6, 7] as the key step. This represents the
first assembly of the 5,11-methanomorphanthridine ring
system in the uncommon enantioseries.
Bu3Sn
O
O
O
O
c
SPh
SPh
16
17
Scheme 1. Synthesis of the vinyl sulfide 17: a) nBuLi (2 equiv), THF,
788C, 1 h, RT, 1.5 h; PhSSPh, THF, 788C, 0.5 h, RT, 0.5 h, 78%;
b) [Pd(PPh3)4] (cat.), Bu3SnH, C6H6, RT, 14 h; c) nBuLi (3 equiv), THF,
788C, 0.5 h, sat. aq. NH4Cl, RT, 89% based on 15.
known vinylidene dibromide 14[10] with CBr4/PPh3/Zn.[11]
Conversion of 14 into the alkynyllithium compound followed
by quenching with diphenyl disulfide produced the alkynyl
sulfide 15. Regioselective hydrostannylation[12] of 15 provided
the vinyl stannane 16, which was best used without purifica-
tion. Tin ± lithium exchange followed by protonation gave the
(Z)-vinyl sulfide 17 in good overall yield.[13]
The synthesis of the (2-azaallyl)stannane 26 required for
the 2-azaallyl anion cycloaddition is shown in Scheme 2. The
epoxide 22 needed for coupling with the vinyl sulfide 17 was
prepared by the Sharpless asymmetric dihydroxylation meth-
od.[14] The known p-methoxybenzyl-protected diol 18[15] was
subjected to Swern oxidation, Horner± Wadsworth ± Emmons
olefination, reduction with diisobutylaluminum hydride, me-
sylation, and in situ displacement of the chloride substituent[16]
to provide the allylic chloride 19. Asymmetric dihydroxyla-
tion of 19 gave the diol 20,[17] which was converted into the
epoxyalcohol 21 with sodium hydride. Mosherꢁs ester analysis
of 21 indicated an enantiomeric excess of 88% for this
material. Protection of 21 as a benzyl ether provided the
desired epoxide 22. Deprotonation of the vinyl sulfide 17 with
tert-butyllithium followed by reaction with epoxide 22 in the
presence of BF3 ´ OEt2[18] afforded the alcohol 23 in good yield
with complete retention of the alkene geometry.[19] Methyl-
ation of 23 gave 24, which was selectively deprotected to
For the retrosynthesis (Figure 2) we envisioned the assem-
bly of montanine-type alkaloids from the perhydroindole 10,
which should be available by the intramolecular cycloaddition
of the 2-azaallyl anion 11.[6] Pictet ± Spengler cyclization,
previously applied by Overman and Shim,[3] would be used to
incorporate the methano bridge, and elimination of HX from
10 would form the requisite alkene. The precursor to the
[*] Prof. W. H. Pearson, B. W. Lian
Department of Chemistry
University of Michigan
930 N. University, Ann Arbor, MI 48109 ± 1055 (USA)
Fax: (‡1)734-763-2307
E-mail: wpearson@umich.edu
[**] This work was supported by the National Institutes of Health (GM-
52491).
1724
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3712-1724 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 12

COMMUNICATIONS
Ar
Cl
Cl
PhS
MeO
BnO
OH
MeO
a
Ar
OH
a
b
SPh
N
OPMB
N
BnO
OPMB
HO
OPMB
H
H
18
19
20
27
26
SnBu3
O
RO
O
Ar
c
e
O
SPh
PhS
BnO
f
MeO
HO
RO
d
OPMB
21 (R=H)
c
b
OPMB
N
23 (R=H)
H
22 (R=Bn)
28
24 (R=Me)
O
O
O
O
MeO
BnO
MeO
BnO
Ar
Ar
MeO
HO
MeO
HO
d
g
h
SPh
SPh
N
N
N
H
H
OH
26
25
29
9
SnBu3
(+)-coccinine
Scheme 2. Synthesis of (2-azaallyl)stannane 26: a) 1. (COCl)2, DMSO,
NEt3, CH2Cl2; 2. (MeO)2P(O)CH2CO2Me, NaH, C6H6, RT, 14 h;
3. iBu2AlH, toluene/THF, 08C, 45 min, RT, 30 min; 4. MsCl, iPr2NEt,
CH2Cl2, 238C, 45 min, dilute with DMF, add LiCl (2 equiv), RT, 5 h, 41%
based on 18; b) AD-mix-a, THF, tBuOH, H2O, MeSO2NH2, 08C, 48 h,
100%; c) NaH, THF, DMSO, RT, 12 h, 81%; d) NaH, BnBr, THF, DMSO,
08C, 2 h, RT, 24 h, additional portion of NaH/BnBr, RT, 8 h, 82%; e) 17
(2.8 equiv) ‡ tBuLi (2.8 equiv), 788C, THF, 15 min, add 22, add BF3 ´
OEt2 (2.8 equiv), 788C, 45 min, 80%; f) NaH, THF, 08C, MeI, RT,
14 h, 100%; g) EtSH (27 equiv), CH2Cl2, BF3 ´ OEt2, 788C, then 258C,
24 h, 92%; h) 1. (COCl)2, DMSO, NEt3, CH2Cl2, 94%, or tetrapropylam-
monium perruthenate (cat.), N-methylmorpholine N-oxide, 4-Š molecular
sieves, CH2Cl2, RT, 14 h, 78%; 2. Bu3SnCH2NH2, Et2O, 4-Š molecular
sieves, RT, 16h, 86%. Bn ˆ benzyl, Ms ˆ methanesulfonyl, PMB ˆ p-
methoxybenzyl.
Scheme 3. Synthesis of (‡)-coccinine (9): a) nBuLi (4 equiv), THF,
788C, 2 h, 45%; b) 37% aq. CH2O, MeOH, RT, 5 min; 6n HCl, 808C,
12 h, 53%; c) 1. Et2O, anhydrous HCl, 08C, 20 min, concentrate in vacuo;
m-CPBA, CH2Cl2, 08C; 2. C6H6, K2CO3, 808C, 1.5 h, 76% based on 28;
d) 1. Ms2O, pyridine, CH2Cl2, 08C, 1 h; 2. CsOAc, DMF, [18]crown-6,
1258C, 48 h; 3. K2CO3, MeOH, RT, 3 h, 52% based on 29. m-CPBA ˆ m-
chloroperbenzoic acid.
CHCl3); natural ( )-coccinine: [a]D25
EtOH), [a]2D5
218.2 (c ˆ 0.055, CHCl3)].
ˆ
104 (c ˆ 0.07,
ˆ
In our initial synthetic planning, examination of molecular
models suggested that two conformations, chairlike 30 and
twist-boat-like 32, had good orbital overlap between the 2-
azaallyl anion and the anionophile, leading to the cis-fused[21]
perhydroindoles 33 and 34, respectively (Figure 3). The
provide the alcohol 25. Oxidation and condensation with
(aminomethyl)tri-n-butylstannane[6c] gave the desired (2-
azaallyl)stannane 26.
N
Ar
X
RO
RO
Ar
Without purification, 26 was subjected to tin ± lithium
exchange at low temperature with n-butyllithium (Scheme 3).
Quenching the reaction with water provided a single stereo-
isomer of the perhydroindole 27 in 45% yield. At this point,
the configuration of the three new stereocenters was not
known. Pictet ± Spengler cyclization of 27 with concomitant
removal of the benzyl protecting group gave the 5,11-meth-
anomorphanthridine 28. Oxidation of the HCl salt of 28 to
generate the sulfoxide followed by thermolysis gave 29, which
differed from an authentic sample of ( )-montanine (1).[20]
The optical rotation of 28 (and all subsequent compounds)
was dextrorotatory rather than levorotatory, which provides
evidence that the cycloaddition had produced the enantio-
meric 5,11-methanomorphanthridine skeleton. Thus, we invert-
ed the alcohol of 29 to produce the nonnatural enantiomer of
coccinine. Mesylation of 29 followed by displacement with
acetate and saponification provided (‡)-coccinine (9), which
was identical to natural ( )-coccinine (2)[20] in all respects (Rf
value in thin-layer chromatography (TLC), 500-MHz 1H
NMR and 90-MHz 13C NMR spectra, and mass spectrum)
with the exception of optical rotation [synthetic (‡)-cocci-
nine: [a]2D5 ˆ ‡96 (c ˆ 0.05, EtOH), [a]2D5 ˆ ‡182.2 (c ˆ 0.045,
RO
(–)-montanine 1
(+)-coccinine 9
N
H
RO
X
H
30
31
X
Ar
X
RO
RO
RO
Ar
RO
N
N
H
H
32
33
Figure 3. Possible conformations of the 2-azaallyl anion in the cyclo-
addition.
chairlike conformation 30 appeared to be less congested than
31. Thus, we predicted that the perhydroindole 31 rather than
33 would result from the cycloaddition.[21] An explanation of
this surprising result is under investigation.
Received: December 9, 1997 [Z11245IE]
German version: Angew. Chem. 1998, 110, 1782 ± 1784
Keywords: alkaloids ´ cycloadditions ´ nitrogen heterocycles
´ synthetic methods ´ total synthesis
Angew. Chem. Int. Ed. 1998, 37, No. 12
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3712-1725 $ 17.50+.50/0
1725

COMMUNICATIONS
[1] S. F. Martin in The Alkaloids, Vol. 30 (Ed.: A. Brossi), Academic
Press, New York, 1987, pp. 251 ± 376.
[2] F. Viladomat, J. Bastida, C. Codina, W. E. Campbell, S. Mathee,
Phytochemistry 1995, 40, 307 ± 311.
[3] a) L. E. Overman, J. Shim, J. Org. Chem. 1991, 56, 5005 ± 5007; b) ibid.
1993, 58, 4662 ± 4672.
A New Base-Pairing Motif Based on Modified
Guanosines**
Jonathan L. Sessler* and Ruizheng Wang
Watson ± Crick base pairing involving purine and pyrimi-
dine subunits plays a crucial role in regulating the structures
and properties of, for example, duplex DNA and hairpin
RNA. Studying synthetic systems with unconventional bind-
ing modes could serve to extend the genetic alphabet of DNA
and RNA, and produce systems of greater structural diversity,
functionality, and catalytic potential.[1] In this context, modi-
fied systems derived from guanine are of considerable interest
because of their potential antiviral activity and their possibly
unique binding ability.[2] However, the number of such
systems that have been analyzed in terms of their self-
association properties remains limited. One example is 7,9-
dimethylguanine, a species that dimerizes in aqueous solution
with the formation of three hydrogen bonds.[3] A second
example is 5'-(tert-butyldimethylsilyl)-2',3'-O-isopropylidene
isoguanosine; this forms a tetramer[4] in organic media that is
more stable than the corresponding guanosine tetramer.[5]
Here we report the new guanine derivative 1, which, when
constrained within a rigid framework, self-associates in
organic solution to form an unprecedented tetrameric gua-
nine-containing array (dimer I). What is unique about this
system is that it is held together by a pair of four-point
hydrogen bonds.[6±8]
[4] a) M. Ishizaki, O. Hoshino, Y. Iitaka, Tetrahedron Lett. 1991, 32,
7079 ± 7082; b) M. Ishizaki, O. Hoshino, Y. Iitaka, J. Org. Chem. 1992,
57, 7285 ± 7295; c) M. Ishizaki, K. Kurihara, E. Tanazawa, O. Hoshino,
J. Chem. Soc. Perkin Trans. 1 1993, 101 ± 110.
[5] a) J. Jin, S. M. Weinreb, J. Am. Chem. Soc. 1997, 119, 2050 ± 2051; b) J.
Jin, S. M. Weinreb, J. Am. Chem. Soc. 1997, 119, 5773 ± 5784.
[6] a) W. H. Pearson, D. P. Szura, W. G. Harter, Tetrahedron Lett. 1988,
29, 761 ± 764; b) W. H. Pearson, D. P. Szura, M. J. Postich, J. Am.
Chem. Soc. 1992, 114, 1329 ± 1345; c) W. H. Pearson, M. J. Postich, J.
Org. Chem. 1992, 57, 6354 ± 6357.
[7] For a review of early work in this area, see T. Kauffmann, Angew.
Chem. 1974, 86, 715 ± 727; Angew. Chem. Int. Ed. Engl. 1974, 13, 627 ±
639.
[8] For related reactions of a-metalated vinyl sulfides with epoxides, see
a) K. Oshima, K. Shimoji, H. Takahishi, H. Yamamoto, H. Nozaki, J.
Am. Chem. Soc. 1973, 95, 2694 ± 2695; b) I. Vlattas, L. D. Veccia, A. O.
Lee, J. Am. Chem. Soc. 1976, 98, 2008 ± 2010.
[9] a) W. H. Pearson, M. J. Postich, J. Org. Chem. 1994, 59, 5662 ± 5671;
b) W. H. Pearson, F. E. Lovering, Tetrahedron Lett. 1994, 35, 9173 ±
9176; c) W. H. Pearson, F. E. Lovering, J. Am. Chem. Soc. 1995, 117,
12336 ± 12337.
[10] M. J. Postich, PhD thesis, University of Michigan (USA), 1994.
[11] E. J. Corey, P. L. Fuchs, Tetrahedron Lett. 1972, 3769.
[12] P. A. Magriotis, J. T. Brown, M. E. Scott, Tetrahedron Lett. 1991, 32,
5047 ± 5050.
[13] While the tin ± lithium exchange and protonation sequence proceeds
via the vinyllithium compound required for the epoxide-opening step
in Scheme 2, it is best generated by deprotonation of purified 17.
[14] K. B. Sharpless, W. Amberg, Y. L. Bennani, G. A. Crispino, J. Hartung,
K.-S. Jeong, H.-L. Kwong, K. Morikawa, Z.-M. Wang, D. Xu, X.-L.
Zhang, J. Org. Chem. 1992, 57, 2768 ± 2771.
[15] a) N. Adje, P. Breuilles, D. Uguen, Tetrahedron Lett. 1993, 34, 4631 ±
4634; the method of Takano et al. was used: b) S. Takano, M.
Akiyama, S. Sato, K. Ogasawara, Chem. Lett. 1983, 1593 ± 1596.
[16] Meyers method, as modified by Overman et al., was used: a) F. W.
Collington, A. I. Meyers, J. Org. Chem. 1971, 36, 3044 ± 3045; b) S. D.
Knight, L. E. Overman, G. Pairaudeau, J. Am. Chem. Soc. 1995, 117,
5776 ± 5788.
H
R
H
C
N
N
O
N
N
N
N
H
H
O
N
H
H
N
N
N
N
N
C
N
H
R
H
R
H
H
C
O
N
N
H
N
N
N
N
H
O
C
[17] For the asymmetric dihydroxylation of allylic bromides in with 72% ee
or lower, see H. C. Kolb, Y. L. Bennani, K. B. Sharpless, Tetrahedron:
Asymmetry 1993, 4, 133 ± 141.
[18] M. J. Eis, J. E. Wrobel, B. Ganem, J. Am. Chem. Soc. 1984, 106, 3693 ±
3694.
[19] The success of the epoxide opening was very sensitive to the reaction
conditions and the nature of the protecting groups. A discussion will
appear in the full account of this work.
H
H
N
N
N
N
1
N
H
H
R
I
1
O
t-BDMSO
t-BDMSO
R =
O-t-BDMS
[20] We thank Dr. Henry Fales of the National Institutes of Health for
samples of ( )-montanine and ( )-coccinine, and Professor Osamu
Hoshino of the Science University of Tokyo for spectra of montanine.
[21] Transition states leading to trans ring junctures are not considered
based on previous experience.[6b, 9] The alternate chairlike conforma-
tion that may lead to 33 is not shown, since it displays severe 1,3-
diaxiallike interactions.
The synthesis of 1 (Scheme 1) involves initially a Pd-
catalyzed cross-coupling between N2-(N,N-dimethylformami-
dine)-protected 8-bromoguanosine (5) and organostannyl
derivative 4[9] produced in situ from 1,8-diethynylanthracene
(3). This sequence gave the bis(guanosine) derivative 6.
Treatment of 6 with methanolic ammonia at room temper-
ature did not give the expected deprotected bis(guanine)
derivative 2, but rather 1, in which the NMe2 group of 6 is
[*] Prof. J. L. Sessler, R. Wang
Department of Chemisty and Biochemistry
University of Texas
Austin, Texas 78712 (USA)
Fax: (‡1)512-471-7550
E-mail: sessler@mail.utexas.edu
[**] This research was supported by the Robert A. Welch foundation. We
thank Prof. Horst Kessler and Dr. Bing Wang for helpful discussions.
1726
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3712-1726 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 12