Angewandte
Chemie
MsCl/DBU protocol proceeded in 90% yield, whereas the
Burgess reagent led to 83% yield. Analogously, carbamate 19
was prepared by application of either the MsCl/DBU proto-
col or the Burgess reagent in good yield. Dienes 18 and 19
were also synthesized in comparable yield by employing one-
pot procedures starting from imine 15 by subsequent addition
of the required reagents. In the next step, the activated
enamine double bond was reduced regio- and stereoselec-
tively under acidic conditions with sodium cyanoborohydride
furnishing the mesyl-protected pentacyclic amine 20 in only
40% yield, but the carbamate 21 was isolated in 78% yield.
1
The H NMR and 13C NMR data of our compound 21
were identical with those reported by Bodwell and Li.[6]
Remarkably, NOE experiments of both 20 and 21 revealed
that the bridgehead protons (11b-H and 13a-H) are cis
positioned as depicted in Scheme 3.[16] This configuration,
which would be wrong for the synthesis of strychnine, can be
rationalized by considering the fairly flat geometry of dien-
amines 18 and 19 and the resulting iminium species. An attack
of the hydride source occurs more likely from the less
hindered convex top face, rather than from the more shielded
bottom side.[17] With these results we had unexpectedly
demonstrated that the sequence as described by Bodwell
and Li leads to the wrong configuration at C-13a, and hence
our first approach to key intermediate 1 had also failed.
In a second attempt we therefore decided to reduce the
tetracyclic building block 2 with Raney nickel to approach
pentacycle 22 directly by in situ formation of imine 15 and its
subsequent reduction (Scheme 4).[17] Gratifyingly, we isolated
pentacyclic amine 22 as a single diastereomer, now bearing
the correct configuration at C-13a, in almost quantitative
yield. Presumably, owing to the steric hindrance because of
the tertiary hydroxy group, the attack of the reducing agent
occurs exclusively from the bottom face of the molecule.[17a]
Without further purification secondary amine 22 was either
mesylated or acylated yielding 23 or 24 in good to very good
yields. NOE and NOESY experiments on amine 22 and
protected amines 23 and 24 indicated that the bridgehead
protons 11b-H and 13a-H are now trans positioned as
required.[16]
Scheme 4. Synthesis of protected strychnine precursors 25 and 26.
Reagents and conditions: a) Raney Ni, H2, 3 d, MeOH, 97%; b) to 23:
MsCl, DMAP, TEA, CH2Cl2, 2 h (92%); to 24: ClCO2Me, DMAP, TEA,
CH2Cl2, 4 h (87%); c) to 25, 27, 29: Burgess reagent, toluene, 708C,
2 h (ratioꢀ2:1:1: ꢀ=77%); d) 1. MsCl, DMAP, TEA, 16 h, 2. DBU,
24 h: to 25, 27, 29: (ratioꢀ5:1:1, ꢀ=89%), to 26, 28, 30:
(ratioꢀ5:1:1, ꢀ=88%); e) one-pot procedure to 25, 27, 29: 1. MsCl,
DMAP, TEA, CH2Cl2, 24 h, 2. DBU, 24 h (ratioꢀ5:1:1, ꢀ=70%); to
26, 28, 30: 1. ClCO2Me, DMAP, TEA, CH2Cl2, 4 h, 2. MsCl, DMAP, TEA,
CH2Cl2, 12 h, 3. DBU, 24 h (ratioꢀ5:1:1, ꢀ=62%).
alkylation of this secondary amine with tosylate 31[5c,10]
furnished the known strychnine precursor 32 in similar yield
to that reported by Rawal[5c] (Scheme 5). Finally, we subjected
Scheme 5. Synthesis of TBS-protected isostrychnine 33. Reagents and
conditions: a) 1. TMSI, CHCl3, 608C, 2 h, 2. MeOH, 608C, 1 h;
b) 1.2 equiv 31, K2CO3, nBu4NI, CH3CN (65% for two steps); c) Pd-
(OAc)2, K2CO3, nBu4NCl, DMF, 708C, 3 h (68%). TMSI=iodotrimethyl-
silane.
Subsequently, we attempted the regioselective conversion
of the tertiary alcohols 23 and 24 into the desired alkenes 25
and 26. In first attempts for elimination we employed the
Burgess reagent which, unfortunately, furnished for both
series mixtures of the three possible alkenes (roughly 2:1:1,
compounds 25 to 30) as depicted in Scheme 4. Fortunately,
treatment of either 23 or 24 with MsCl/DBU at room
temperature afforded the desired key building blocks 25
and 26 as the major products in good selectivity (5:1:1 ratio),
which could easily be isolated by column chromatography. We
also developed one-pot procedures for compounds 25 and 26
starting from amine 22. However, as a result of longer
reaction times and partial decomposition, the overall yields of
25 and 26 were lower than those achieved in the two-step
protocol. 2D NMR experiments (COSY, HMQC, NOESY)
on compounds 25 and 26 clearly indicated the correct relative
configuration as depicted in Scheme 4.[16]
compound 32 to the published Heck reaction which gave the
hexacyclic TBS-protected isostrychnine 33 in 68% yield. The
NMR data for compounds 32 and 33 are in complete
agreement with the previously published data and allowed
further unambiguous proof of our configurational assign-
ments.
In conclusion, with this formal total synthesis of strych-
nine we could demonstrate the power of SmI2-induced
cascade reactions. Starting from simple indole precursors,
the novel process allowed the generation of two new rings and
three stereogenic centers, including a quaternary carbon
atom, in one step. Moreover, the key building block 26 was
obtained from commercially available indolylacetonitrile in
We then converted carbamate 26 by TMSI-induced
deprotection into the pentacyclic amine 1.[18] Subsequent
Angew. Chem. Int. Ed. 2010, 49, 8021 –8025
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim