1728
P. D. Bailey et al. / Tetrahedron Letters 54 (2013) 1726–1729
OMe
aldehyde proximal
to indole 3-position
diaxial boat
conformation
CHO
OHC
OMe
N
N
Note 16
R
R
N
Me
N
Me
16
15
Scheme 4. Chair-boat equilibrium in the cyclization to the indolic 3-position.
Table 1
Reducing agents for the conversion of 9 or 10 into the pentacyclic compounds 11/12 (see Scheme 3 and Ref. 13); yields are quoted for isolated products that were fully
characterized, whilst the natural/epi ratio was determined from the NMR spectra of the crude products
Compound
R (in 10)
Reduction
Yield (%)
Natural/epi (11:12)
9
10a
—
Ph
Various conditions
For example, NaBH4; H2/Pd-C;
BH3ÁTHF; N-selectride;
H2/PtO2, 6 M HCl17
H2/PtO2, 6 M HCl
H2/Pd-C, 6 M HCl
NaBH4
No cyclization
10–60
—
0:100
10b
10b
10c
10c
10c
10d
10d
10d
cy-C6H11
cy-C6H11
But
70
68
32
47
37
100
69
48
22:78
59:41
0:100
45:5518
58:42
0:100
42:58
72:28
But
H2/Pd-C
But
LiAlH(OBut)3
Adamantyl
Adamantyl
Adamantyl
NaBH4
H2/Pd-C
LiAlH(OBut)3
For related substrates, Masamune3 reported a natural/epi ratio of 2:1, whilst Cook6 achieved a 40:60 ratio for the reductive cyclization step in his total synthesis of ajmaline.
by Cook6—this essentially involved the use of AcOH/Ac2O/HCl to
trap the alcohol as the acetate 13 (which could be isolated as the
hydroxy-amine 14), followed by reduction of this iminium inter-
mediate. However, no hint of cyclization was observed until the
conditions became so forcing (>60 °C) that the starting aldehyde
was degrading faster than cyclization was occurring, and no signif-
icant amounts of ring-closed product could be isolated. Of course,
pre-closure of the SE ring facilitates this ring-closure, but leads to
stereochemical problems at the reduction step (loc. cit.). It oc-
curred to us, however, that replacing the benzyl by benzoyl might
aid the cyclization in two ways associated with the nitrogen
becoming sp2 hybridized (see Scheme 4):
group occurred, and the resulting cyclohexyl derivative clearly
showed the presence of the singlet (ca. d 2.6) characteristic of
the ajmaline series.17 We reasoned that the bulkier protecting
group was partially blocking the si face of the imine, allowing
reduction to the ajmaline series. We therefore tried to optimize
this with bulkier protecting groups and reducing agents, eventually
leading to a 72:28 preference for the ajmaline series—the best
selectivity ever reported for this cyclization.
We believe these results are important for three reasons:
(a) We now have a clearer insight into the subtle conforma-
tional factors that affect ring-closure to the indole 3-position
(Scheme 3), for which the functionalization of the remote
piperidine nitrogen is a crucial factor (see Scheme 4 and
Ref. 16).
(b) The stereoselectivity that we observed for the reductive
cyclization of the benzoyl derivative 10a (exclusively epi)
is in contrast to that reported by Masamune.3
(c) These results clearly take us very close to a synthesis of
ajmaline itself, with seven of the nine chiral centres
installed; introduction of the ethyl group and closure of
the SE ring are the steps still required.
(a) increasing the conformational flexibility of the boat con-
former 16, allowing a closer approach to the indole 3-
position;
(b) lowering the energy of the boat conformer 16 relative to the
chair 15, thereby increasing the concentration of 16 in the
appropriate conformation for ring-closure.
Removal of the benzyl group was accomplished by catalytic
hydrogenation in trifluoroethanol,12 followed by O-methylation
and N-benzoylation; Weinreb amide formation and LiAlH4 reduc-
tion yielded the aldehyde 10a. To our delight, relatively mild con-
ditions led to cyclization,13 for which reduction by NaBH4 gave a
single diastereomeric product14 with three new chiral centres.
The stereochemistry at the 2-indolyl position is characterized by
the coupling constant—ca. 5 Hz in the epi-ajmaline series, but a
broad singlet (J <1 Hz) in the ajmaline series.15 To our dismay, it
was the epi-ajmaline derivative 12 that we had obtained, and this
remained the case under a wide range of hydrogenation and hy-
dride reducing conditions; this is in contrast to results obtained
by Masamune,3 in which preferential reduction to the ajmaline ste-
reochemistry (natural/epi, 2:1) was reported.
Acknowledgment
We thank the EPSRC and the Mass Spectrometry Service at
Swansea, and Keele University for ACORN funding (M.C.).
References and notes
1. (a) Siddiqui, S.; Siddiqui, R. J. Indian Chem. Soc. 1931, 8, 677–680; (b)
Woodward, R. B. Angew. Chem. 1956, 68, 13–20.
2. Brugada, J.; Brugada, P. Am. J. Cardiol. 1996, 78, 69–75.
3. Masamune, S.; Ang, S. K.; Egli, C.; Nakatsuka, N.; Sarkar, S. K.; Yasunari, Y. J. Am.
Chem. Soc. 1967, 89, 2506–2507.
One result did however give us hope. Using platinum oxide as
the catalyst, some reduction of the phenyl ring of the benzoyl
4. Mashimo, K.; Sato, Y. Chem. Pharm. Bull. 1970, 18, 353–355.