Enantiospecific, diastereoselective synthesis of Aspidosperma alkaloid analogues from secologanin

Article abstract of DOI:10.1016/S0040-4039(00)00459-7

Pure enantiomers of two diastereomeric Aspidosperma alkaloid analogues antipodal at the spiro and adjacent centres have been prepared, as kinetic and thermodynamic products, respectively, in a biomimetic sequence via a secodine-like intermediate from a secologanin derivative and Kuehne's indolo- azepine. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00459-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 3547–3550
Enantiospecific, diastereoselective synthesis of Aspidosperma
alkaloid analogues from secologanin
∗
Richard T. Brown and Mythily Kandasamy
Department of Chemistry, The University of Manchester, Manchester M13 9PL, UK
Received 17 January 2000; accepted 17 February 2000
Abstract
Pure enantiomers of two diastereomeric Aspidosperma alkaloid analogues antipodal at the spiro and adjacent
centres have been prepared, as kinetic and thermodynamic products, respectively, in a biomimetic sequence via a
secodine-like intermediate from a secologanin derivative and Kuehne’s indolo-azepine. © 2000 Elsevier Science
Ltd. All rights reserved.
Keywords: alkaloids; biomimetic; indole; stereoselection; terpene.
Both enantiomeric series of Aspidosperma alkaloids, exemplified by vincadifformine 1, occur in
nature, even though they are derived from one enantiomer of the universal monoterpenoid indole alkaloid
1
precursor, strictosidine, formed by condensation of tryptamine and secologanin 3. This dichotomy was
plausibly attributed to the intervention of an achiral intermediate, e.g. secodine 2, with cyclisation to
one enantiomer or the other of 1 being controlled by the chiral environment of an enzyme specific to
the individual plant (Scheme 1). Presumably due to its high reactivity, no secodine could be isolated
1
in vivo, although related dimers and dihydrosecodine have been discovered. Eventually, Kuehne and
co-workers substantiated the proposed pathway by elegant in vitro syntheses of a transient secodine 2
2
with biomimetic cyclisation in high yield to racemic vincadifformine. In subsequent work, their indolo-
azepine methodology was used in enantioselective syntheses of (+)- and (−)-vincadifformine in high
3
enantiomeric excess.
Our group has carried out several stereoselective syntheses of Corynanthé alkaloids from secologanin
in which the asymmetric centre at C-7 played a key role in inducing chirality at other centres.4
We considered that the C-7 centre might also direct stereoselectivity in cyclisation of secodine-like
intermediates formed, by analogy with Kuehne’s work, from indolo-azepine 7 and aldehydes derived
from secologanin and hence generate a novel series of chiral Aspidosperma alkaloid analogues (Scheme
2).
∗
Corresponding author. Tel: +44 161 275 4632; fax: +44 161 275 4939; e-mail: r.t.brown@man.ac.uk (R. T. Brown)
0
040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00459-7
tetl 16739

3
548
Scheme 1.
Scheme 2. Reagents and conditions: (i) Ac
MeOH, 1.5 h; (v) β-glucosidase, pH 5 aq. buffer, 37°C, 2 days; (vi) pH 7 aq. Me
viii) silica chromatography; (ix) CHCl , 7 days; (x) MeOH, ∆, 6 days
2
O/py, 12 h; (ii) H
2
/Pd/EtOAc/H
2
O, 48 h; (iii) SOCl
2
/DCM, 15 min; (iv) NaOMe,
2
CO, 37°C, 3 days; (vii) MeOH, ∆, 2–4 days;
(
3
Thus, acetylation and catalytic hydrogenation of secologanin 3 afforded tetrahydrosecologanin tetra-
+
acetate 4 [M 560.2105 (C H O ); λmax(MeOH): 238 nm; νmax (CHCl ): 3500, 1750, 1720, 1630
2
5
36 14
3
−
1
cm )] which was converted with SOCl in DCM to the 5-chloro derivative, mp 125°C, as indicated
2
inter alia by loss of the OH peak in the IR spectrum and elemental analysis for C H O Cl. Zemplen
25
35 13
deacetylation and hydrolysis with β-glucosidase then gave in 42% overall yield from 3 the 5-chloro-
+
35
−1
aglucone 5 [M 248.0817 (C H O Cl); νmax (CHCl ): 3415, 1703, 1630 cm ]. Removal of the
1
1
17
4
3
sugar was confirmed by a shift in the UV maximum from 238 to 272 nm on addition of alkali, due to
opening of the hemi-acetal ring and formation of a conjugated enolate anion, and the H NMR spectrum
1
which also showed 5 to be a pair of C-1 epimers from two H-9 singlets at δ 7.51 and 7.48. After

3
549
some experimentation, it was found that on standing in aq. acetone at pH 7 and 37°C for three days,
an intramolecular displacement of the chlorine at C-5 by the C-9 enol could be achieved to give, in 80%
yield, the C-1 aldehyde 6. The molecular ion at m/z 212.103 (C H O ) indicated loss of HCl, and the
11
16
4
UV maximum at 237 nm was attributable to a β-alkoxyacrylate chromophore, but the lack of a shift with
alkali indicated that the product was not a hemi-acetal. Accordingly, no OH signal was present in the IR
−
1
spectrum, but there were now two C_O bands at 1750 and 1710 cm , confirmation of the former as the
1
7
C-1 aldehyde and hence structure 6 coming from its H NMR spectrum.
The indolo-azepine 7 and aldehyde 6 were then heated under reflux in MeOH for four days, and
subsequent chromatography afforded two isomeric products, characterised as lactams A, mp 238–240°C
5
[
α] +150°, and B, mp 205–207°C [α] −131°, in ca. 10 and 20% unoptimised yield, respectively. For
D
D
both products the UV absorption with λ
226, 296, 326 nm was no longer indolic but corresponded to
max
a combination of β-anilinoacrylate and β-alkoxyacrylamide chromophores, as did two C_O IR bands at
−1
ca. 1710 and 1650 cm ; molecular ions at m/z 406 analysing for C H O N and the presence of only
24
26
4
2
one methyl ester in the NMR spectra were consistent with formation of a lactam. In both mass spectra
the base peak at m/z 214 plus an ion at m/z 192 corresponded to a fragmentation by retro Diels–Alder
and allylic cleavage characteristic of an Aspidosperma skeleton.
1
Lactam A was assigned structure 10 and relative stereochemistry from a complete analysis of the H
7
NMR spectrum with decoupling and NOE experiments. The ethyl group was shown to be cis to H-1 and
trans to H-7 when irradiation of both C-3 protons gave enhancement of the former (3.1/2.5%) but not
0
the latter, whereas irradiation of H-7 did enhance one of the C-7 protons, showing that the ethanamine
bridge and H-7 were on the same side. Finally, the absolute configuration of the spiro centre in lactam 10
was established as S and C-1 as R from correlation of the positive optical rotation and, more specifically,
the associated positive Cotton effect at 326 nm in the CD spectrum with those of known alkaloids of the
6
1
7
vincadifformine group. For lactam B structure 11 was deduced in a similar way from H NMR and
CD spectra. Salient points were that a negative Cotton effect established spiro R and C-1 S chirality, and
that substantial, reciprocal NOE enhancements observed between H-1, H-7 and the protons of the ethyl
group showed that they were all on the same side of the molecule. Since on heating lactam 10 in refluxing
MeOH it was slowly converted into 11, but the latter gave no indication of reverting to 10 under the
same conditions, lactam 10 must be the kinetic and 11 the thermodynamic product. Molecular modelling
indicated that the difference in stability could be attributed essentially to an eclipsing interaction along
the 2–7 bond in 10 that was relieved on conversion to 11.
After condensing 7 and 6 in MeOH for a shorter time of two days, chromatography afforded the
0
intermediate enamine 8 as a ca. 9:1 mixture of isomers, probably 3 -epimers, in ca. 15% unoptimised
+
1
yield [M 438.2149 (C H O N ); λ
224, 282, 291 nm]. The structure was established from the H
max
NMR spectrum with signals for both indolo-azepine and monoterpene moieties, and characterised in
2
5
30
5
2
7
particular by a singlet for the enamine H-1 at δ 6.81. Furthermore, NOE studies not only showed that the
major enamine had E 1,2-geometry, but also that it was the E N,1-rotamer: irradiation of H-1 enhanced
0
H-7 (4.5%), 9 (2.6%) and 2a (9.7%) but not any protons of the ethyl group, whereas irradiation of H -3
a
0
0
gave enhanced peaks for protons on C-7 (5.2%) but not C-2 .
Significantly, on standing in CHCl for seven days the enamine 8 was quantitatively converted into
3
a single product, the kinetic lactam 10. These results can be rationalised by invoking a secodine-type
intermediate 9 formed by retro-Michael fragmentation and lactamisation of 8, after E/Z isomerisation
of the enamine. At room temperature, cyclisation is then directed by the C-7 chiral centre to the less
hindered face of the enamide ring to afford the kinetic lactam 10 exclusively. However, at 65°C reversal
to 9 must occur and thus permit a slow accumulation of the thermodynamic lactam 11 by attack from the
more hindered face, a process that is apparently irreversible under these conditions.

3
550
We have thus achieved the synthesis from a single chiral precursor of two diastereomeric Aspi-
dosperma alkaloid analogues that are antipodal at every centre except one, the original C-7 centre
of secologanin. If this asymmetric centre were eliminated (e.g. by dehydrogenation) then we would
have the prospect of readily controlled selective syntheses for both enantiomeric series. A further
possibility is the use of secologanin itself with the indolo-azepine and hence a route to novel hybrid
Aspidosperma–Corynanthé structures.
Acknowledgements
We thank the CVCP for an ORS Award and the University of Manchester for a Bursary (M.K.).
References
1. Herbert, R. B. In Monoterpenoid Indole Alkaloids; Saxton, J. E., Ed.; John Wiley: New York, 1983; Chapter 1 and
references cited therein.
2
3
4
. Kuehne, M. E.; Bohnert, J. C. J. Org. Chem. 1981, 46, 3443–3447 and references cited therein.
. Kuehne, M. E.; Podhorez, D. E. J. Org. Chem. 1985, 50, 924–929.
. (a) Brown, R. T. In Monoterpenoid Indole Alkaloids; Saxton, J. E., Ed.; John Wiley: New York, 1983; pp. 135–9, 192–4.
(
b) Brown, R. T.; Dauda, B. E. N.; Santos, C. A. M. Chem. Commun. 1991, 825–826.
. Kuehne, M. E.; Bohnert, J. C.; Bornmann, W. G.; Kirkemo, C. L.; Kuehne, S. E.; Seato, P. J.; Zebvitz, T. C. J. Org. Chem.
985, 50, 919–924.
. Hesse, M. Indolalkaloide in Tabellen; Springer-Verlag: Berlin, 1968; pp. 58–63.
5
1
6
7
1
. H NMR spectra (300 MHz, CDCl
3
) (numbering based on secologanin): 6 δ 9.68 (d, J=3 Hz, H-1), 7.62 (s, H-9), 4.14 (m,
J=12, 4.5, 1.5 Hz, H-5eq), 3.92 (m, J=12, 10.5, 4 Hz, H-5ax), 3.70 (s, OMe), 3.00 (m, J=1.5 Hz, H-7), 2.47 (m, J=3 Hz,
H-2), 1.9–1.7 (m, H
m, ar-H
2
-6, H-3
b
), 1.35 (m, H-3
a
), 0.93 (t, J=7 Hz, H
3
-4). 8 (major isomer) δ 8.63 (NH), 7.48 (s, H-9), 7.1–7.6
0
(
4
), 6.81 (s, H-1), 4.25 (dd, J=5, 2.5 Hz, H-3 ), 4.07 (m, H-5
a
), 3.79 (m, H-5
b
), 3.72 (s, OMe), 3.69 (s, OMe),
0
0
0
3
H
.5–3.7 (m, H
2
-7 ), 3.16 (m, J=10.5, 2.5 Hz, H-2
-6 ), 2.15 (m, H-3 ), 1.97 (m, H-6 ), 1.7–1.5 (m, H-6
), 4.28 (td, J=12, 4 Hz, H-5
a
), 2.97 (bd, J=2.5 Hz, H-7), 2.64 (m, J=10.5, 5 Hz, H-2
b
), 2.5–2.3 (m,
0
2
a
b
a
, H-3
b
), 1.00 (t, J=7 Hz, H
3
-4). 10 δ 8.95 (NH), 7.38 (d, J=2.5 Hz,
0
H-9), 6.9–7.2 (m, ar-H
4
a
), 4.00 (td, J=12, 12, 4 Hz, H-5
b
), 3.96 (ddd, J=12, 8, 2.5 Hz, H-7
a
),
0
0
a0
3
6
0
.81 (s, OMe), 3.59 (d, J=2.5 Hz, H-1), 3.48 (ddd, J=12, 11, 6 Hz, H-7
, 2.5 Hz, H-7), 2.16 (d, J=16 Hz, H-2 ), 1.9–2.1 (m, H-6 , H -6), 1.90 (ddd, J=12, 6, 2.5 Hz, H-6 ), 1.22 (m, H-3
b
), 2.70 (dd, J=16, 2.5 Hz, H-2
a
), 2.47 (ddd, J=10,
0
0
b
b
2
b
),
.97 (m, H-3
a
), 0.68 (t, J=7 Hz, H
3
-4). 11 δ 8.93 (NH), 7.70 (d, J=2.5 Hz, H-9), 6.9–7.2 (m, ar-H
4
), 4.67 (td, J=10.5, 2.5
0
Hz, H-7
H-7
a
), 4.36 (td, J=10.5, 2.5 Hz, H-5
), 2.85 (dt, J=9, 2.5 Hz, H-7), 2.46 (dd, J=16, 2.5 Hz, H-2
a
), 4.04 (d, J=1.5 Hz, H-1), 3.95 (m, H-5
b
), 3.75 (s, OMe), 3.16 (dm, J=10.5 Hz,
0
0
0
0
b
), 1.9–2.0 (m, H-6 , H
b
b
), 2.04 (d, J=16 Hz, H-2
a
2
-6), 1.90
0
(ddd, J=12, 6, 2.5 Hz, H-6
a
), 1.22 (m, H-3
b
), 0.97 (m, H-3
a
), 0.68 (t, J=7 Hz, H
3
-4).