Expeditious route towards (±)-desethyleburnamonine, a precursor of (±)-vindeburnol

Article abstract of DOI:10.1002/ejoc.201101058

An expeditious route to (±)-desethyleburnamonine (1), a direct precursor of vindeburnol, is reported. Microwave irradiation of an indolo-tetrahydropyridine ethyl ester 2 in CH₃CN in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) gave

Full text of DOI:10.1002/ejoc.201101058

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201101058
Expeditious Route Towards (؎)-Desethyleburnamonine, a Precursor of
(؎)-Vindeburnol
Laurence Jung-Deyon,[a] Bruno Giethlen,[a] and André Mann*[b]
Keywords: Lactams / Synthesis design / Isomerisation / Condensation reactions / Microwave chemistry
An expeditious route to (Ϯ)-desethyleburnamonine (1), a di-
rect precursor of vindeburnol, is reported. Microwave irradia-
tion of an indolo-tetrahydropyridine ethyl ester 2 in CH3CN
in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)
gave the pentacyclic (Ϯ)-desethyleburnamonine in a one-pot
process. Experimental evidence support an allylamine–en-
amine isomerization, followed by a rare Pictet–Spengler con-
densation in basic media. From commercially available 3-(2-
bromo-ethyl)indole and ethyl pyridin-3-yl acetate, (Ϯ)-des-
ethyleburnamonine was obtained in 52% yield over three
steps.
investigated to achieve our goal. At first, the opportunity
to isomerize cyclic allylamine of type A to the correspond-
ing cyclic enamine of type B was examined. Indeed, the
cyclic six-membered allylamines are generally available by
partial reduction of pyridinium salts, and the corresponding
cyclic enamines – if accessible – would provide valuable in-
termediates towards the synthesis of alkaloids by C–C
bonding with nucleophiles at C2 and with electrophiles at
C3 [reaction (a), Scheme 2].
Introduction
(Ϯ)-Vindeburnol, a synthetic derivative of eburnamine, is
known as a central vasodilator with some benefits for severe
depression.[1–3] (Ϯ)-Vindeburnol features a pentacyclic
structure and has the D/E ring junction in a trans relation-
ship.[4] For extensive pharmacological studies, we need
multigram quantities of vindeburnol,[5] which is accessible
from (Ϯ)-desethyleburnamonine (1) by diastereoselective re-
duction of the carbonyl function in 1[1] (Scheme 1).
Scheme 2. (a) Elusive isomerization of cyclic allylamine; (b) Joule’s
synthesis of indoloquinolizidine.[16]
Scheme 1. Synthesis of (Ϯ)-vindeburnol via (Ϯ)-desethyleburn-
amonine (1).
Decades ago, Wenkert was tackling the subject without
success.[15] In 1978, Joule’s group reported a simple way to
perform the synthesis of indoloquinolizine D from the
indolyl-ethyl-piperideine C through a two-step procedure:
(i) double-bond isomerization and (ii) Pictet–Spengler con-
densation in AcOH [reaction (b), Scheme 2]. The double-
bond isomerization was operative with tBuOK in
DMSO.[16,17] As we needed a practical method to isomerize
allylamine 2 into cyclic enamine 3, and to convert enamine
3 to desethyleburnamonine 1, we built on Joule’s results and
submitted substrate 2 to a chemical trial under basic condi-
tions.
Of course, it was essential to identify other bases than
tBuOK to avoid hydrolysis or trans-esterification in sub-
strate 2. Our strategy is depicted in Scheme 3: first double-
bond isomerization under basic conditions, then a Pictet–
Spengler condensation followed by a spontaneous lactam-
ization.
However, from a literature survey, it appears that the
routes towards desethyleburnamonine 1 are rather cumber-
some: (i) Potier reported a short access to 1, using the re-
ductive cyclization of a pyridinium salt in the presence of
hydrogen;[6–8] (ii) Lounasmaa disclosed several routes
towards 1 and analogues based on the Potier–Polonovski
reaction;[9–13] (iii) later Massiot reported on the isomeriza-
tion–cyclization of indolyl-tetrahydropiperidine 2, assisted
by RhI.[14] However, all these sequences are incompatible
for scale-up, mainly because of tedious purification steps
and low overall yields. Thus, other approaches had to be
[a] Prestwick Chemical,
Boulevard Gonthier d’Andernach, 67400 Illkirch, France
[b] Laboratoire d’Innovation Thérapeutique, UMR 7200 CNRS –
Université de Strasbourg Route du Rhin Faculté de Pharmacie,
67401 Illkirch, France
E-mail: andre.mann@unistra.fr
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201101058.
Eur. J. Org. Chem. 2011, 6409–6412
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6409

L. Jung-Deyon, B. Giethlen, A. Mann
SHORT COMMUNICATION
as 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene (MTBD) or
1,1,3,3-tetramethylguandidine (TMG). The conversion of 2
was roughly similar to that with DBU; with organic bases
having lower pKas, such as DIPEA, 1,4-diazabicyclo[2.2.2]-
octane (DABCO), or TMEDA, no conversion was observed
at all. As a trend, it seems that strong organic bases (with
a pKa value of about 12) are the best choice to achieve good
conversion, to realize the capture and shift of the two pairs
of the acidic protons present in substrate 2 (HaHaЈ and
HbHbЈ, see Scheme 5). In order to examine the scope of
this unprecedented reaction, several other substrates 4–8
were prepared as well as an aryl isoster 9. When submitted
to the above experimental conditions, substrates 4 and 5
were converted to 1 and 5Ј in a one-pot reaction with good
yields [reactions (a) and (b), Scheme 4]. Substrate 6 bearing
a Na–Me group underwent solely the isomerization to en-
Scheme 3. Three-step designed route from 2 to 1: double-bound
isomerization to 3, Pictet–Spengler condensation (P–S) and lactam-
ization to 1.
Here, we report a new route towards 1 using a one-pot
transformation of allylamine 2 to 1 by a rare example of a
Pictet–Spengler condensation performed under basic condi-
tions.[18–20]
Results and Discussion
The starting compound 2 was prepared by a two-step amine 6Ј [reaction (c), Scheme 4], whereas amide 8 or ester
sequence from commercially available bromoethylindole 7 (the methylene homologue of 2) remain unaltered even
and ethylhomonicotinate (for experimental details see the after extensive heating, [reactions (d) and (e), Scheme 4].
Supporting Information). The direct condensation of the Interestingly, with the arylethyl-piperideine 9, enamine 9Ј
two partners resulted in the formation a pyridinium salt, was obtained as the unique adduct, which resulted from the
which was reduced with excess NaBH4 to yield the known double-bond shift [reaction (f), Scheme 4].
tetrahydropiperidine 2.[8,15]
At the onset, we deliberately investigated organic bases
and solvents compatible with scale-up development. In a
preliminary experiment, a solution of 2 in the presence of
excess 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) was
heated at reflux for 20 h in toluene. From the reaction mix-
ture, two compounds could be isolated and identified: 3 in
25% yield, originating from the expected isomerization (al-
lylamine–enamine) pathway, and surprisingly, lactam 1 in
5% yield. The presence of lactam 1 was unexpected and
was obviously the result of a further transformation of 3:
(i) a C–C bound formation and (ii) a subsequent lactam-
ization. The trans relationship of ring D/E in 1 was proven
by NMR analysis (J = 12 Hz in [D5]pyridine), the presence
of Bohlman’s bands in the IR spectrum in the range 2793–
2823 cm–1, and comparison with reported physical
data.[8,14,21] However, the chemical conversion was modest,
as the starting compound 1 was recovered in 60% yield. At
the time we suspected that a base-induced Pictet–Spengler
condensation was responsible for the transformation of 3 to
1. In order to improve the yield of desethyleburnamonine 1
and get insight into the mechanism, a screening of experi-
mental conditions was performed. At first, conventional
heating was replaced by microwave irradiation (sealed
flask), and toluene by CH3CN. As a result, the reaction
time was shortened to 2 h, and, more importantly, trans-
(Ϯ)-desethyleburnamonine (1) was isolated with a 51%
yield with traces of enamine 3 [reaction (a), Scheme 4]. No
traces of the corresponding cis epimer of 1 was observed. It
is assumed that 1 is the kinetic and thermodynamic product
under the reaction conditions studied. Indeed, the indole
lone pair participates in the amide bond, prohibiting pos-
sible epimerization.[22] Several other organic bases similar
to DBU were examined, amidines such as 1,5-diazabicy-
clo[4.3.0]non-5-ene (DBN) or 6-(dibutylamino)-1,8-diazabi-
cyclo[5.4.0]undecene-7-ene (DBADBU), or guanidines such crowave experiments were performed in sealed flasks.
Scheme 4. Tentative transformations of analogues of 2. All the mi-
6410
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6409–6412

Expeditious Route Towards (Ϯ)-Desethyleburnamonine
assistance of DBU, the final lactamization concludes the
sequence. One argument in favor of the proposed conver-
sion of 2 to 2Ј follows. When 2Ј, prepared as the E stereo-
isomer,[28] was subjected to the standard conditions (DBU,
CH3CN, μW), lactam 1 was obtained with a similar yield
as that for the direct transformation of 2 to 1, [reaction (g),
Scheme 4]. Therefore, 2Ј is certainly a probable intermediate
of this two-step transformation. Except for the loss of
1 equiv. ethanol in the chemical transformation from 2 to
1, the proton balance is unchanged, therefore DBU oper-
ates as a catalyst according to the proposed mechanism.
However, an excess of DBU has to be used. Reduction of
the amount of DBU did result in a very low conversion
rate, 8 equiv. DBU were found to be ideal.
As enamine 6Ј did not undergo Pictet–Spengler conden-
sation, it may be suggested that DBU produces the pyrrole
amide by proton abstraction. Nucleophilic attack from
either C2 or C3 then proceeds intramolecularly from the
electron rich indole towards the iminium fragment en route
to 1. As this reaction is not observed with enamine 9Ј, C2
attack is probably involved in the observed Pictet–Spengler
condensation from 3 to 1. Finally, to establish a preparative
value for the sequence, an experiment performed on the
gram-scale gave 1 with an optimized yield of 64% after pu-
rification by column chromatography on silica gel. For the
trans selectivity observed for adduct 1, the following tenta-
Scheme 5. Tentative mechanism for the three-step conversion of 2
to 1.
The above results deserve the following comments: (i) the tive mechanism can be proposed.[29] Transition state TS A
two-step sequence is substrate dependent, a complementar- is favored over TS B, which is destabilized by the A[1,2] in-
ity between the proton acidity of the substrate and base teraction between the pseudo equatorial bulky acetate
strength is mandatory; (ii) for the isomerization to take group and the iminium C–H unit[30,31] (Scheme 6). Ap-
place, the acidity of the proton next to the carboxylate proach of the indole moiety would occur from the less hin-
group must be high; (iii) the Pictet–Spengler step requires dered Si-face of the iminium moiety, which leads to the for-
the presence of a Na–H group. Therefore, the following mation of 1 in either mechanism [pathway (a) or (b)] by C-
chemical scenario can be established to explain the one-pot 2 or C-3 attack of the indole unit. To reach (Ϯ)-vindeburn-
transformation of 2 to 1.
ol, it remains to reduce the ketone to carbinol, a known
At first, one of the HaHaЈ protons is abstracted by DBU procedure.[1]
from substrate 2, then a sequence of enolization and conju-
gation occurs with the assistance of DBU/DBU-H+, which
may operate as a proton shuttle, ending with adduct
2Ј.[23–26] One of the Hb protons next to the nitrogen atom
is next in line, with regard to its acidity, to be abstracted,
and the DBU-assisted proton transfer operates again. This
results in the formation of adduct 3, which is the expected
enamine, arising from the desired allylamine–enamine isom-
erization. Then, as 3 continues its journey, the presence of
DBU-H+ may serve as a proton source to convert the en-
amine to the iminium, which then enters into the Pictet–
Spengler condensation. The mechanism of the Pictet–
Spengler reaction is still matter of discussion.[24] The C2
and C3 atoms of the pyrrole fragment are nucleophilic.
Therefore, after the formation of the iminium, the reaction
can proceed either by attack at C2 to directly yield the six-
membered ring or by attack at C3 to yield a spiroindolenine
intermediate that would undergo a 1,2-alkyl shift to form
the same adduct. Evidence for both mechanisms in solution
exist, and the predominant mechanism is not entirely
clear.[27] The spiro intermediate is obtained if the proton
Scheme 6. Transition state for the Pictet–Spengler reaction.
Conclusions
By combining microwave irradiation and the use of an
from Na of the pyrrole ring is abstracted by DBU. With the organic base (DBU), pentacyclic (Ϯ)-desethyleburnamonine
Eur. J. Org. Chem. 2011, 6409–6412
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6411

L. Jung-Deyon, B. Giethlen, A. Mann
SHORT COMMUNICATION
[3] J. F. Pujol, D. Weissmann, Patent FR 2865649, 2005.
[4] M. Cesario, C. Pascard-Billy, Acta Crystallogr., Sect. B 1973,
29, 529–537.
[5] P. Ciapetti, L. Deyon, C. G. Wermuth, J. F. Pujol, D.
Weissmann, Patent FR 2869034, 2005.
[6] C. Thal, T. Imbert, H. P. Husson, P. Potier, Bull. Soc. Chim. Fr.
1973, 2010–2013.
(1), a direct precursor of vindeburnol, was prepared from
indolethyl-piperideine (2) on a gram-scale with a yield of
64%. The presented three-step sequence features allyl-
amine–enamine isomerization, Pictet–Spengler condensa-
tion in basic conditions, and lactamization.
[7] H. P. Husson, T. Imbert, C. Thal, P. Potier, Bull. Soc. Chim. Fr.
1973, 6, 2013–2016.
[8] E. Wenkert, B. Wickberg, J. Am. Chem. Soc. 1965, 87, 1580–
Experimental Section
1589.
General Procedure for the Isomerization, Pictet–Spengler Condensa-
tion and Lactamization on a One-Gram Scale
[9] M. Lounasmaa, D. D. Belle, A. Tolvanen, Heterocycles 1999,
51, 1125–1130.
[10] K. Rönkkö, M. Berner, R. Jokela, A. Tolvanen, J. Heterocycl.
Chem. 2000, 37, 1145–1149.
[11] M. Lounasmaa, L. Miikki, A. Tolmvanen, Tetrahedron 1996,
52, 9925–9930.
[12] M. Lounasmaa, R. Jokela, E. Karvinen, A. Tolvanen, Tetrahe-
dron 1988, 44, 2367–2375.
[13] A. R. J. Castaigne, Patent DE 2302821, 1973.
[14] G. Massiot, A. Cherif, Bull. Soc. Chim. Fr. 1990, 648–655.
[15] E. Wenkert, K. G. Dave, F. Haglid, R. G. Lewis, T. Oishi, R. V.
Stevens, M. Terashima, J. Org. Chem. 1968, 33, 747–753.
[16] S. J. Martinez, J. A. Joule, Tetrahedron 1978, 34, 3027–3036.
[17] M. Rivière, A. Lattes, Bull. Soc. Chim. Fr. 1972, 730–736.
[18] E. D. Cox, J. M. Cook, Chem. Rev. 1995, 95, 1797–1842.
[19] K. Yamada, S. Yamaguchi, N. Hatae, T. Abe, T. Iwamura, M.
Ishikura, Heterocycles 2011, 83, 815–826.
(؎)-(3α,16β)-Desethyleburnamonine (1): To a suspension of allyl-
amine 2 (1 g, 3.2 mmol) in dry CH3CN (12 mL) was added DBU
(3.9 mL, 25.6 mmol). The mixture was irradiated with microwave
irradiation at 160 °C for 6 h, and the solvents were evaporated un-
der vacuum. The resulting residue was purified by silica gel column
chromatography, eluting with a gradient of cyclohexane 100% to
AcOEt 100% as eluent to yield lactam 1 as a white solid [550 mg,
m.p. 136–139 °C (m.p. 144 °C from ref.[14]), 64%]. IR: ν = 2793,
˜
1
2823 cm–1 (Bohlmann’s bands). H NMR (300 MHz, CDCl3): δ =
1.26 (m, 1 H); 1.81–1.87 (m, 3 H); 1.94 (m, 1 H); 2.31 (m, 1 H);
2.35–3.02 (m, 6 H); 3.12 (m, 1 H), 3.14 (m, 1 H); 7.23–7.29 (m, 2
H); 7.42 (m, 1 H); 8.34 (m, 1 H) ppm. 13C NMR (75 MHz, CDCl3):
δ = 21.5, 25.6, 30.2, 38.0, 39.8, 52.4, 54.8, 62.1, 112.0, 116.3, 118.4,
124.0, 124.3, 130.1, 134.3, 135.3, 167.9 ppm. C17H18N2O (266.34):
calcd. C 76.66, H 6.81, N 10.52; found C 76.64, H 6.87, N 10.43.
MS-ESI: calcd. for [M + H]+ 269; found 267.
[20] N. N. Smolyar, M. G. Abramyants, T. L. Zavyazkina, D. I.
Matveeva, Y. S. Borodkin, I. A. Voloskii, Russ. J. Org. Chem.
2009, 45, 1219–1223.
[21] M. Lounasmaa, M. Berner, M. Brunner, H. Suomalainen, A.
Tolvanen, Tetrahedron 1998, 54, 10205–10216.
[22] T. M. Lowry, J. Chem. Soc. 1927, 2554–2265.
[23] L. Fedor, P. H. Gray, J. Am. Chem. Soc. 1976, 98, 783–787.
[24] I. Kaljurand, A. Kütt, L. Sooväli, T. Rodima, V. Mäemets, I.
Leito, I. A. Koppel, J. Org. Chem. 2005, 70, 1019–1028.
[25] M. M. Changalov, G. D. Ivanova, M. A. Rangelov, P. Acharya,
S. Acharya, N. Minakawa, A. Földesi, I. B. Stoineva, V. M.
Yomtova, C. D. Roussev, A. Matsuda, J. Chattopadhyaya, D.
Petkov, ChemBioChem 2005, 6, 992–996.
[26] J. J. Maresh, L. A. Giddings, A. Friedrich, E. A. Loris, S. Panji-
kar, B. L. Trout, J. Stoeckigt, B. Peters, S. E. O’Connor, J. Am.
Chem. Soc. 2008, 130, 710–723.
[27] F. Ungemach, J. M. Cook, Heterocycles 1978, 9, 1089–1119.
[28] A. Da Silva Goes, C. Ferroud, J. Santamaria, Tetrahedron Lett.
1995, 36, 2235–2238.
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedures and characterization data for the
compounds are presented.
Acknowledgments
The authors thank Biocortech (Paris) for a gift of vindeburnol.
L. J. expresses her gratitude to J. F. Pujol and D. Weissmann
(Biocortech) who entitled her to incorporate part of the work done
on vindeburnol in her PhD. Thanks are expressed to Professor Ca-
mille-Georges Wermuth, the founder of Prestwick Chemical, for
support.
[29] A. G. Wee, Q. J. Yu, J. Org. Chem. 2001, 66, 8935–8943.
[30] R. W. Hoffmann, Chem. Rev. 1989, 89, 1841–1860.
[31] F. Johnson, Chem. Rev. 1968, 68, 375–413.
Received: July 20, 2011
[1] A. Farcilli, I. Medici, R. Fournex, F. Barzaghi, Patent
DE 2807643, 1978.
[2] F. Barzaghi, M. Dragonetti, J. R. Boissier, Drug Dev. Res. 1982,
2, 533–537.
Published Online: October 10, 2011
6412
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6409–6412