Stereocontrolled reduction of an oxazepinohexahydroindolo[2,3-a]quinolizine derivative: Asymmetric total synthesis of (+)-tacamonine

Article abstract of DOI:10.1016/S0040-4039(01)01488-5

Efficient, stereocontrolled total synthesis of the title compound is described, starting from enantiopure intermediates. A key step was the diastereoselective catalytic hydrogenation of a pentacyclic oxazepinohexahydroindolo[2,3-a]quinolizine derivative.

Full text of DOI:10.1016/S0040-4039(01)01488-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 7237–7240
Stereocontrolled reduction of an
oxazepinohexahydroindolo[2,3-a]quinolizine derivative:
asymmetric total synthesis of (+)-tacamonine
Bruno Danieli, Giordano Lesma,* Daniele Passarella, Alessandro Sacchetti and Alessandra Silvani
Dipartimento di Chimica Organica e Industriale, Universita` degli Studi di Milano,
Centro CNR di Studio per le Sostanze Organiche Naturali, via Venezian 21-20133 Milan, Italy
Received 19 July 2001; revised 9 August 2001; accepted 13 August 2001
Abstract—Efficient, stereocontrolled total synthesis of the title compound is described, starting from enantiopure intermediates.
A key step was the diastereoselective catalytic hydrogenation of a pentacyclic oxazepinohexahydroindolo[2,3-a]quinolizine
derivative. © 2001 Elsevier Science Ltd. All rights reserved.
Tacamonine (1) is an indole alkaloid isolated1 in 1984
from the constituents of Tabernaemontana eglandulosa
Stapf, a plant which is widely distributed in Central
Africa. The structural similarity between 1 and Hunte-
ria alkaloids (e.g. (−)-eburnamonine (2)), which possess
valuable vasodilator and hypotensive activities, pro-
motes several efforts towards the synthesis of 1.2,3 The
attachment of the ethyl group away from the D/E ring
junction of the pentacyclic framework engenders an
additional stereocenter, setting out three methine
hydrogen atoms 3bH, 14bH and 20bH cis related as a
point significant in pursuing a synthetic study of taca-
monine (Fig. 1).
Among the reported syntheses, an asymmetric
approach4 to 1 featured free radical cyclization to form
the D ring, whereas another synthesis5 relied on the
development of the D ring component from a substi-
tuted pyridine. Just recently, an original route6 was
designed on the basis of symmetric considerations
which emphasized the establishment of the relative cis
configuration at C14 and C20 stereocenters. In a previ-
ous work,7 we reported an efficient lipase-mediated
preparation of the possible precursor 3 (Scheme 1), in
an optically pure form, starting from asymmetrized
2-substituted
propane-1,3-diols.
Following
this
approach, we secured a formal enantioselective entry to
tacamonine according to literature precedents,8 which
Figure 1.
Scheme 1.
Keywords: alkaloids; asymmetric synthesis; hydrogenation; diastereoselection.
* Corresponding author. Fax: 39 2 2367606; e-mail: giordano.lesma@unimi.it
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01488-5

7238
B. Danieli et al. / Tetrahedron Letters 42 (2001) 7237–7240
describe the conversion of racemic 3 into 1, via 4, using
standard chemistry.
indolylethyl piperidinone 9, easily preparable in opti-
cally pure form, as previously reported by us.5
However, in that work, no attempt was made to control
the relative configuration of the newly created stereo-
genic centers at C3 and C14. Moreover, to our know-
ledge, none of the above-mentioned stereoselective
routes to 1 have been able to install efficiently the
required cis D/E ring junction, avoiding in this way the
final step of tedious diastereoisomeric separation. As a
consequence, our studies were on going in order to
furnish an efficient answer to this challenge. We report
here an original approach that allowed us to reach a
completely enantio- and diastereoselective synthesis of
(+)-tacamonine 1.
As will be seen, we assume that the stereofacial selectiv-
ity of hydrogen attack on enamine 7 would be influ-
enced by the presence of the bulky acetoxymethyl
substituent at C20. The required b-face diastereoselec-
tivity would be favoured allowing the set-up of the
correct absolute stereochemistry at the C3 and C14
carbon atoms.
With the hydroxymethyl derivative 6 at hand, we would
be able to carry out the homologation of the C20 side
chain and the construction of ring E, according to a
standard reaction sequence, and thus to complete the
totally stereoselective synthesis of tacamonine.
As shown retrosynthetically in Scheme 2, we envisaged
the chiral, all cis-indoloquinolizidine 5 as a key
advanced intermediate, whose conversion into taca-
monine should be easily achievable.9 Our strategy was
based on the condition that an efficient stereocontrolled
installation of C3 and C14 stereocenters could be
accomplished via the catalytic hydrogenation of the
chiral enamine 7 to give 6. Access to 7 was envisioned
through Bischler–Napieralsky and C14 hydroxymethyl-
ation reactions directly from the appropriate
Enamine 8 was prepared by treatment of 9 with POCl3
and then further reacted with paraformaldehyde in
acetonitrile at room temperature (Scheme 3). The reac-
tion gave cleanly the unstable enamine 7, which was
immediately hydrogenated over a Pt catalyst. Unfortu-
nately, the only isolable product was the indoloquino-
lizidine 10, whose formation can be explained by the
kinetically favoured reduction of the less hindered
Scheme 2.
Scheme 3. (a) POCl3, CH2Cl2, reflux, 78%; (b) (CH2O)3, THF, rt; (c) H2/PtO2, dioxane, 62%.

B. Danieli et al. / Tetrahedron Letters 42 (2001) 7237–7240
7239
product. Molecular mechanics calculations suggested a
rigid framework for 13, with a minimum energy confor-
mation in which approach from the concave a-face of
the molecule would be severely hindered by the pseudo-
axial oriented acetoxymethyl group at C20.
trisubstituted enamine double bond of 8, formed trans-
iently from 7, by retroaldol condensation.
Looking for the possibility of making compound 7
more stable, discouraging in this way the observed
reaction pathway, we attempted to protect the C14-
hydroxymethyl group. A variety of classical methods
(Ac2O/Py; MOMCl, Et3N; TBDMSCl, DIPEA) proved
to be unsuccessful.
Indeed, Pt-catalyzed hydrogenation of 13 afforded 1413
in 75% yield and 95% d.e., in which the 3bH, 14bH
configuration was set. The assignment of the stereo-
chemistry was performed by comparison of the CD
spectrum of 14 with those of pentacyclic tacamane
alkaloids.14 Analysis of NMR spectra (1H and 13C)
allowed the establishment of the diastereoisomeric ratio
and confirmed the stereochemistry, by observation of
diagnostic NOE contacts between 3bH (l 4.57) and
14bH (l 2.25) and between 14bH and 20bH (l 2.05).
Having attained the complete control of absolute
configuration at all three stereogenic centers of taca-
monine framework, we proceeded to complete our syn-
thetic plan. The homologation of the C20
acetoxymethyl side chain was attained by a three-step
sequence involving hydrolysis of acetate, tosylation and
displacement of the tosylate with lithium dimethyl-
cuprate (Scheme 6).
On the basis of our earlier experience on the chemistry
of eburnamine-vincamine indole alkaloids, we were
aware that enamines such as 11 (Scheme 4) react with
formaldehyde and then sodium borohydride to give
pentacyclic
compounds
(12)
embodying
an
oxazepinoindoloquinolizine skeleton,10,11 which pos-
sesses a seven-membered heterocyclic ring joined by a
trans-junction with ring D.
We felt that if we were able to convert 8 into the
pentacycle 13, still containing the C3ꢀC14 double bond,
we would obtain an ideal substrate for achieving
diastereoselection in the addition of hydrogen to the
enamine double bond. Following these ideas, we
allowed the N-Boc-protected enamine 8 to react with
excess aqueous formaldehyde and catalytic formic acid
in refluxing acetonitrile. After 1 h, we could isolate the
expected compound 1312 (Scheme 5) as the unique
Deprotection of 1515 with BF3·Et2O, according to
Sundberg protocol,16 gave 1617 in good yield. At this
point, completion of the total enantiosynthesis of 1 was
performed according to literature precedents.9 The
hydroxy group of 16 was converted into the corre-
sponding tosylate and then homologated with KCN.
Eventually, base treatment with MeONa in MeOH at
reflux, followed by acid hydrolysis allowed the closure
of the fifth ring to give enantiopure (+)-tacamonine 1,
which was recrystallized from acetone to provide color-
less needles {[h]D=122 (c=0.2, CHCl3), mp 180°C (lit.1
1
180–181°C)}, whose UV, CD, H NMR and MS spec-
tral data were identical in all respects to the reported
Scheme 4.
ones.1,8
Scheme 5. (a) (CH2O)3, HCO2H, THF, reflux, 90%; (b) H2/PtO2, dioxane, 75%.
Scheme 6. (a) NaOH, THF/H2O, 98%; (b) TsCl, Et3N, CH2Cl2, 76%; (c) Me2CuLi, Et2O, −10°C, 72%; (d) Ac2O, BF3·Et2O, LiBr,
rt, 61%; (e) see Ref. 9, 48%.

7240
B. Danieli et al. / Tetrahedron Letters 42 (2001) 7237–7240
d, 1H, J=12.9 Hz), 3.91 (dd, 1H, J=11.0, 5.8 Hz), 3.74
Acknowledgements
(dd, 1H, J=11.0, 7.5 Hz), 3.34 (br dd, 1H, J=12.8, 6.6
Hz), 3.23 (ddd, 1H, J=12.8, 11.2, 5.0 Hz), 3.05 (m, 1H),
2.78 (dd, 1H, J=11.5, 3.5 Hz), 2.59 (dd, 1H, J=15.8, 4.5
Hz), 2.45 (t, 1H, J=11.5 Hz), 2.25 (m, 1H), 2.05 (m, 1H),
2.00 (s, 3H), 1.61 (m, 1H), 1.24 (m, 1H); 13C NMR
(CDCl3, 75.4 MHz) l 170.7, 135.1, 133.7, 126.9, 121.7,
119.3, 118.1, 109.9, 108.9, 79.0, 77.9, 66.8, 58.7, 51.2,
47.9, 42.5, 36.4, 27.0, 20.7, 16.4; FAB+MS m/z: 341
[MH+]; EI-HRMS for C20H24N2O3: calcd 340.1787;
found 340.1789. Selected data for 3-epi,14-epi 14: CD
(MeOH): u (Dm) 236 (+4.2), 252 (+1.2), 263 (+1.8), 271
(+1.4); 1H NMR (300 MHz, CDCl3): l 7.46 (d, 1H,
J=8.0 Hz), 7.26 (d, 1H, J=8.0 Hz), 7.18 (t, 1H, J=8.0
Hz), 7.10 (t, 1H, J=8.0 Hz), 5.92 and 5.01 (AB system,
2H, J=11.0 Hz), 4.49 (br s, 1H), 4.37 (dd, 1H, J=14.5,
7.5 Hz), 4.33 (dd, 1H, J=14.5, 7.0 Hz), 4.20 (dd, 1H,
J=12.8, 4.0 Hz), 3.90 (dd, 1H, J=12.8, 2.0 Hz), 3.21 (m,
2H), 3.03 (m, 1H), 2.88 (dd, 1H, J=12.0, 4.2 Hz), 2.58
(br d, 1H, J=12.0 Hz), 2.53 (m, 1H), 2.33 (m, 1H), 2.05
(s, 3H), 2.03 (m, 1H), 1.71 (dt, 1H, J=5.0, 13.8 Hz), 1.51
(br d, 1H, J=13.8 Hz); 13C NMR (CDCl3, 75.4 MHz) l
171.0, 135.0, 133.8, 127.0, 121.6, 119.4, 119.0, 110.1,
109.0, 79.0, 78.1, 65.2, 59.1, 51.5, 45.3, 37.3, 32.6, 24.8,
20.9, 16.5; FAB+MS m/z: 341 [MH+].
This work was supported by Ministero dell’Universita` e
della Ricerca Scientifica e Tecnologica (MURST).
References
1. van Beek, T. A.; Verpoorte, R.; Svendsen, A. B. Tetra-
hedron 1984, 40, 737.
2. Lounasmaa, M.; Tolvanen, A. In The Alkaloids; Cordell,
G. A., Ed. The Eburnamine-Vincamine Alkaloids; Aca-
demic Press: New York, 1992; Vol. 42, pp. 1–116.
3. Lounasmaa, M.; Din Belle, D.; Tolvanen, A. Tetrahedron
Lett. 1995, 36, 7141.
4. Ihara, M.; Setsu, F.; Shohda, M.; Taniguchi, N.; Toku-
naga, Y.; Fukumoto, K. J. Org. Chem. 1994, 59, 5317.
5. Din Belle, D.; Tolvanen, A.; Lounasmaa, M. Tetrahedron
1996, 52, 11361.
6. Ho, T.-L.; Su, C.-Y. Tetrahedron 2001, 57, 507.
7. Danieli, B.; Lesma, G.; Macecchini, S.; Passarella, D.;
Silvani, A. Tetrahedron: Asymmetry 1999, 10, 4057.
8. Massiot, G.; Sousa, Oliveira, F.; Le´vy, J. Bull. Soc. Chim.
Fr. II 1982, 185.
9. The N-deprotected analogue of all cis-indoloquino-
lizidine 5, namely 16, was recently synthetized as a race-
mate and then converted into ( )-tacamonine (Ref.: Din
Belle, D.; Tolvanen, A.; Karinen, K.; Lounasmaa, M.
Tetrahedron 1998, 54, 157).
14. To´th, G.; Clauder, O.; Gesztes, K.; Yemul, S. S.;
Snatzke, G. J. Chem. Soc., Perkin Trans. 2 1980, 701.
15. Selected data for 15: [h]D=−9.12 (c=0.8, CHCl3). 1H
NMR (300 MHz, CDCl3): l 7.45 (d, 1H, J=8.0 Hz),
7.25–7.05 (m, 3H), 5.91 and 5.00 (AB system, 2H, J=
10.8 Hz), 4.55 (br s, 1H), 4.29 (dd, 1H, J=12.0, 3.6 Hz),
3.93 (br d, 1H, J=12.0 Hz), 3.36 (m, 2H), 3.05 (m, 1H),
2.82 (dd, 1H, J=11.5, 3.5 Hz), 2.56 (dd, 1H, J=16.5, 4.7
Hz), 2.43 (t, 1H, J=11.5 Hz), 2.22 (m, 1H), 2.00 (m, 1H),
1.51 (br d, 1H, J=13.5 Hz), 1.15 (m, 3H), 0.88 (t, 3H,
J=7.5 Hz); 13C NMR (CDCl3, 75.4 MHz) l 133.5, 122.8,
121.8, 119.4, 109.9, 109.4, 79.0, 77.5, 65.9, 59.9, 51.1,
42.3, 39.4, 29.6, 26.9, 26.6, 16.5, 11.4; FAB+MS m/z: 297
[MH+]; EI-HRMS for C19H24N2O: calcd 296.1889; found
296.1892.
10. Din Belle, D.; Tolvanen, A.; Lounasmaa, M. Rec. Trav.
Pays-Bas 1995, 114, 37.
11. Kalaus, G.; Malkieh, N.; Katona, I.; Kajta´r-Peredy, M.;
Koritsa´nszky, T.; Ka´lma´n, A.; Szabo´, L.; Szanta´y, C. J.
Org. Chem. 1985, 50, 3760.
12. Selected data for 13: [h]D=−7.20 (c=1, CHCl3). 1H
NMR (300 MHz, CDCl3): l 7.45 (d, 1H, J=8.1 Hz),
7.20–7.05 (m, 3H), 5.75 and 5.52 (AB system, 2H, J=
10.0 Hz), 4.48 (m, 2H), 4.14 (dd, 1H, J=11.5, 5.5 Hz),
4.05 (dd, 1H, J=11.5, 6.2 Hz), 3.22 (br t, 1H, J=10.2),
3.17 (dd, 1H, J=10.9, 5.6 Hz), 3.08 (ddd, 1H, J=10.9,
7.2, 5.0 Hz), 2.94 (m, 2H), 2.82 (dd, 1H, J=10.9, 9.2 Hz),
2.39 (m, 1H), 2.10 (m, 1H), 2.08 (s, 3H), 1.93 (dd, 1H,
J=16.0, 9.1 Hz); FAB+MS m/z: 339 [MH+]; EI-HRMS
for C20H22N2O3: calcd 338.1630; found 338.1622.
13. Selected data for 14: [h]D=−5.38 (c=1, CHCl3); CD
(MeOH): u (Dm) 236 (−5.2), 248 (−1.4), 268 (−2.6); 1H
NMR (300 MHz, CDCl3): l 7.48 (d, 1H, J=8.0 Hz), 7.27
(d, 1H, J=8.0 Hz), 7.19 (t, 1H, J=8.0 Hz), 7.10 (t, 1H,
J=8.0 Hz), 5.93 and 5.01 (AB system, 2H, J=10.8 Hz),
4.57 (br s, 1H), 4.29 (br dd, 1H, J=12.9, 2.9 Hz), 3.93 (br
16. Sundberg, R. J.; Russell, H. F. J. Org. Chem. 1973, 38,
3324.
17. Selected data for 16: Amorphous solid; 1H NMR (300
MHz, CDCl3): l 10.22 (br s, 1H), 7.50–7.05 (m, 4H), 4.78
(t, 1H, J=1.9 Hz), 4.05 (d, 2H, J=5.0 Hz), 3.18–2.88 (m,
3H), 2.85–2.45 (m, 4H), 1.15 (m, 2H), 0.85 (t, 3H, J=7.5
Hz); 13C NMR (CDCl3, 75.4 MHz) l 135.5, 130.0, 127.0,
121.3, 119.2, 118.0, 111.1, 106.9, 65.1, 59.6, 51.5, 41.3,
37.4, 28.6, 27.0, 16.5, 11.4; FAB+MS m/z: 285 [MH+];
EI-HRMS for C18H24N2O: calcd 284.1889; found
284.1880.