A new approach to indolo[2,3-a]quinolizidines through radical cyclization of 2-acyl-1-phenylthiotetrahydro-β-carbolines bearing pendent α,β-unsaturated esters

Article abstract of DOI:10.1016/j.tetlet.2009.08.100

A new method is reported for the preparation of indolo[2,3-a]quinolizidines based on radical cyclization of a 2-acyl-1-phenylthiotetrahydro-β-carboline bearing a pendent α,β-unsaturated ester. The required radical cyclization precursor is efficiently asse

Full text of DOI:10.1016/j.tetlet.2009.08.100

Tetrahedron Letters 50 (2009) 6342–6346
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A new approach to indolo[2,3-a]quinolizidines through radical cyclization
of 2-acyl-1-phenylthiotetrahydro-b-carbolines bearing pendent
a,b-unsaturated esters
*
Myles W. Smith, Roger Hunter , Devendren J. Patten, Wolfgang Hinz
Department of Chemistry, University of Cape Town, Rondebosch 7701, Cape Town, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 July 2009
Revised 13 August 2009
Accepted 28 August 2009
Available online 2 September 2009
A new method is reported for the preparation of indolo[2,3-a]quinolizidines based on radical cyclization
of a 2-acyl-1-phenylthiotetrahydro-b-carboline bearing a pendent
a,b-unsaturated ester. The required
radical cyclization precursor is efficiently assembled from E-5-ethoxycarbonyl-4-pentenoic acid and
3,4-dihydro-b-carboline through a DCC/HOBt-activation/N-acylation and BF3ꢀEt2O/PhSH iminium-ion
trapping sequence. Tin-mediated radical cyclization of the radical cyclization precursor affords stereose-
lectively a cis-lactam (dr = 7:1) in good yield (81%), bearing the correct D/E ring fusion stereochemistry
for the Tacaman alkaloids. The methodology has been applied to formal syntheses of the indoloquinolizi-
dine alkaloids, ( )-eburnaminol and ( )-larutensine.
Ó 2009 Elsevier Ltd. All rights reserved.
The indolo[2,3-a]quinolizidine framework 1 (Fig. 1) is a struc-
tural motif found within numerous natural products of varying
structure and biological activity, including alkaloids such as the
architecturally interesting (+)-Na-methylvellosimine (2) and the
well-known reserpine (3).1 In addition to its importance within
Nature’s structural palette, this scaffold continues to present
medicinally relevant leads for potential therapies against a range
of diseases.2
Thus, given the significance of the indoloquinolizidine core in
natural and clinical arenas alike, many methods have been
developed for its preparation, including Bischler–Napieralski,3 Pic-
tet–Spengler,4 Fischer indole synthesis5 and vinylogous Mannich6
approaches, as well as strategies based on formal aza-[3+3]-7 and
[3+2]-carbonyl ylide8 cycloadditions. The Pictet–Spengler reaction
has historically been particularly popular, and recently has been
developed into synthetically useful asymmetric variants based on
the use of tryptophan-derived substrates4a–c or asymmetric organ-
ocatalysis.4d–g
which we aimed to instal via a 6-exo-trig cyclization of a benzylic
-acylamino radical11 7, itself generated from 2-acyl-1-phenyl-
thiotetrahydro-b-carboline 8 (Scheme 1).
Herein we describe the development of the aforementioned
method, and apply it to a brief formal synthesis of the Eburn-
amine–Vincamine alkaloids, ( )-eburnaminol and ( )-larutensine.
Our work began with the synthesis of the radical cyclization
precursor 8. As shown retrosynthetically in Scheme 1, we sought
to assemble 8 from acid 9 and 3,4-dihydro-b-carboline12 (10). From
the outset, we envisaged achieving this N-acylation/iminium-ion
a
Despite these advances, indoloquinolizidine synthesis in the con-
text of the pentacyclic alkaloids of the Eburnamine–Vincamine and
Tacaman families (general framework, 4, Scheme 1) has often suf-
fered from low levels of stereocontrol. For this reason, we set out
to develop a stereoselective method that would be particularly use-
ful in accessing indoloquinolizidines of this type, specifically in the
context of a total synthesis of the indole alkaloid tacamonine (5).9
Our planned method involved a novel disconnection10 of the
C3–C14 bond (tacamonine numbering) of the indoloquinolizidine,
* Corresponding author.
Figure 1. The indolo[2,3-a]quinolizidine framework 1 and examples of indolo-
E-mail address: Roger.Hunter@uct.ac.za (R. Hunter).
quinolizidine natural products.
0040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.08.100

M. W. Smith et al. / Tetrahedron Letters 50 (2009) 6342–6346
6343
Scheme 1. Retrosynthesis of ( )-tacamonine (5) and general framework of the Tacaman and Eburnamine–Vincamine alkaloids (4).
trapping sequence under mild conditions, since this would allow
for the potential use of sensitive acids bearing -stereocentres
With 8 in hand, we next investigated the key radical cyclization.
Subjecting 8 to standard radical cyclization conditions [n-Bu3SnH
(3.0 equiv), ACCN (0.1 equiv), degassed PhMe (0.02 M), reflux]
gratifyingly resulted in the consumption of starting material and
formation of more polar products [ACCN = 1,10-azobis(dic-
yclohexylcarbonitrile)]. Careful chromatography of this mixture
revealed the major component to be the desired cis-lactam (cis-6,
51%, Scheme 4),17 accompanied by a smaller amount of the
trans-isomer (trans-6, 15%)17 and radical reduction product (18,
20%). After further optimization, we found, as expected, that con-
ducting the reaction under more dilute conditions (0.01 M), lower-
ing the equivalents of n-Bu3SnH (1.5 equiv) and slowing the
addition of the n-Bu3SnH/ACCN/PhMe solution (over 1.5 h), all
resulted in a decrease in the amount of reduction product (down
to 6%), together with an increase in the yield and diastereoselectiv-
ity (cis:trans = 81%:12%) (Scheme 4).
The structures of cis- and trans-6 were assigned on the basis of 2D
NMR experiments (COSY, HSQC, NOESY and HMBC), with key stereo-
chemical information being afforded by the cross-peak obtained in
the NOESY spectrum of cis-6 between H-3 (d 5.39) and H-14 (d
3.23). Unfortunately, signal overlap in the 1H NMR spectrum of
trans-6 precluded the use of a similar NOESY experiment to show
the absence of this cross-peak in the trans-isomer. Moreover, the
coupling constants between these two protons yielded little infor-
mation, as in both isomers, the J values were small and very similar
[JH3–H14 (cis) = 2.0 Hz; JH3–H14 (trans) = 1.8 Hz], indicating significant
distortion from the chair-like conformers. Thus, with the NMR data
unable to confirm the stereochemistry of the major isomer, we
sought definitive proof in the form of an X-ray crystal structure
determination. While this was not possible for cis-6 (formed as a
gum), its saponification [LiOH (2.5 equiv), THF/H2O (3:1), 76%]
yielded the crystalline acid 19, which could be recrystallized from
MeOH to provide crystals suitable for a single crystal X-ray determi-
nation. The X-ray structure18 (Fig. 2) clearly confirms the desired cis-
arrangement of H-3 and H-14.
a
(cf. structure of tacamonine, 5). Mindful of these requirements,
we began our search by exploring DCC/HOBt-activation of the acid,
followed by Nb-acylation and low temperature BF3ꢀEt2O-mediated
interception of the iminium ion with thiophenol in a one-pot
sequence (DCC = N,N0-dicyclohexylcarbodiimide, HOBt = 1-hydro-
xybenzotriazole).
Our initial studies were carried out on the model acids, acetic
and heptanoic acids. In the event, treatment of these acids
(1.2 equiv) with DCC (1.2 equiv) and HOBt (1.2 equiv) at 0 °C, fol-
lowed by addition of 10 (1.0 equiv) and subsequent exposure to
PhSH (1.3 equiv) and BF3ꢀEt2O (1.3 equiv) at ꢁ78 °C, cleanly affor-
ded the corresponding a-sulfanyl amides as a mixture of rotamers
(acetic, 11: ꢂ1.6:1; heptanoic, 12: ꢂ2.8:1) in good yield following
chromatography (Scheme 2).13 Indoles 11 and 12 could subse-
quently be N-Boc-protected [Boc2O (1.3 equiv), DMAP (0.1 equiv),
THF, RT] in near quantitative yield to give 13 and 14, respectively.
Thus, with a reliable method in place, the preparation of 8 could
be attempted. Acid 9 was prepared from 4-pentenoic acid (15)
through a high-yielding ozonolysis/Horner–Wadsworth–Emmons
sequence (80%, E:Z = 13:1; Scheme 3).14 Thereafter, the coupling
of 9 (1.0 equiv) with 3,4-dihydro-b-carboline (10) (1.2 equiv) pro-
ceeded smoothly, affording after treatment with benzenethiol the
sulfanyl amide 17 in 88% yield (3 mmol scale) in a manner that
was easily scalable (14 mmol: 83%). As before, indole N-Boc protec-
tion15 proceeded in essentially quantitative yield to furnish 8 as a
mixture of rotamers (ꢂ3.1:1) about the amide bond.16
The observed diastereoselectivity in the radical cyclization can
be rationalized by examining the possible transition states leading
to each isomer (Fig. 3). Assuming reasonably that the larger indole
substituent at the reacting radical centre adopts a pseudoequatori-
al position in the transition states, cis-6 would arise from one in
which the enoate ester group adopts a pseudoaxial orientation
(TS-1), while in the transition state leading to trans-6 this group
Scheme 2. Model a-sulfanyl amide synthesis.

6344
M. W. Smith et al. / Tetrahedron Letters 50 (2009) 6342–6346
Scheme 3. Preparation of radical cyclization precursor 8.
Scheme 4. Radical cyclization of 8.
Having established the requisite stereochemistry to be in place
for ( )-tacamonine, we sought to demonstrate the utility of cis-6 in
a short formal synthesis of the natural products, eburnaminol (20)
and larutensine (21) (Scheme 5). These two Eburnamine–Vinca-
mine alkaloids were isolated20 in 1991 from Kopsia larutensis and
had previously been synthesized by Lounasmaa and Karvinen from
indoloquinolizidine ester 22 in six and seven steps, respectively.21
Thus, lactam 6 was converted through a three-step sequence (final
two steps unoptimized) into 22. First, the N-Boc group of 6 was re-
moved by treatment with TFA in the presence of thioanisole as a
cation scavenger to give free indole 23 in excellent yield (93%)
(Scheme 5). Thereafter, conversion of 23 into thiolactam 24 was
achieved with Lawesson’s reagent (64%), and which was then re-
duced with excess Raney Nickel (77%) to deliver 22. The spectral
data (1H, 13C, IR and HRMS) of 22 were in good agreement with
those reported by Lounasmaa, except that the melting point of
our material differed significantly from that reported by Husson
et al.,22 although the solvent used in their case was different from
ours [120–124 °C (EtOAc/hexanes); lit. mp: 160 °C (benzene/
hexane)22].
Figure 2. X-ray crystal structure of acid 19 showing the cis-arrangement of H-3 and
H-14.
would be pseudoequatorial (TS-2). In the latter orientation, the es-
ter suffers significant steric interaction with the bulky N-t-butoxy-
carbonyl group, which destabilizes TS-2 and thereby disfavours the
formation of trans-6; TS-1, on the other hand, avoids such clashing
and for this reason cis-6 predominates.19
In summary, we have developed an efficient method for the
preparation of indolo[2,3-a]quinolizidines that is particularly

M. W. Smith et al. / Tetrahedron Letters 50 (2009) 6342–6346
6345
Figure 3. Proposed transition state model for the observed stereoselectivity (X = CO2Et).
Scheme 5. Formal syntheses of ( )-eburnaminol (20) and ( )-larutensine (21) via 22.
3. (a) Dieters, A.; Petterson, M.; Martin, S. F. J. Org. Chem. 2006, 71, 6547–6561; (b)
Takasu, K.; Nishida, N.; Tomimura, A.; Ihara, M. J. Org. Chem. 2005, 70, 3957–
3962; for a related approach via a (benzylthio)iminium salt intermediate, see:
(c) Amat, M.; Gómez-Esqué, A.; Escolano, C.; Santos, M. M. M.; Molins, E.;
Bosch, J. J. Org. Chem. 2009, 74, 1205–1211.
4. (a) Cox, E. D.; Cook, J. M. Chem. Rev. 1995, 95, 1797–1842; (b) Ma, J.; Yin,
W.; Zhou, H.; Cook, J. M. Org. Lett. 2007, 9, 3491–3494; (c) Allin, S. M.;
Thomas, C. I.; Doyle, K.; Elsegood, M. R. J. J. Org. Chem. 2005, 70, 357–359;
(d) Wanner, M. J.; Boots, R. N. A.; Eradus, B.; de Gelder, R.; van
Maarseveen, J. H.; Hiemstra, H. Org. Lett. 2009, 11, 2579–2581; (e)
Klausen, R. S.; Jacobsen, E. N. Org. Lett. 2009, 11, 887–890; (f) Mergott,
D. J.; Zuend, S. J.; Jacobsen, E. N. Org. Lett. 2008, 10, 745–748; (g) Seayad,
J.; Seayad, M. J.; List, B. J. Am. Chem. Soc. 2006, 128, 1086–1087.
5. Fornicola, R. S.; Subburaj, K.; Montgomery, J. Org. Lett. 2002, 4, 615–617.
6. Martin, S. F. Acc. Chem. Res. 2002, 35, 895–904.
applicable to the synthesis of pentacyclic alkaloids of the Tacaman
family. Our method involves a high-yielding assembly of the
required radical cyclization precursor using a mild DCC/HOBt-acti-
vation and BF3ꢀEt2O/HSPh iminium-ion trapping sequence, as well
as an n-Bu3SnH-mediated radical cyclization that is selective for
the desired cis-isomer. From this cis-product, the formal syntheses
of ( )-eburnaminol and ( )-larutensine were completed through
the preparation of a common indoloquinolizidine intermediate.
Efforts to expand the scope of this method and apply it to a racemic
total synthesis of tacamonine are underway in our laboratory, and
will be reported in due course.
7. (a) Sydorenko, N.; Zificsak, C. A.; Gerasyuto, A. I.; Hsung, R. P. Org. Biomol. Chem.
2005, 3, 2140–2144; (b) Luo, S.; Zificsak, C. A.; Hsung, R. P. Org. Lett. 2003, 5,
4709–4712.
Acknowledgements
8. (a) England, D. B.; Padwa, A. J. Org. Chem. 2008, 73, 2792–2802; (b) England, D.
B.; Padwa, A. Org Lett. 2007, 9, 3249–3252.
9. For model studies directed towards this natural product from this laboratory,
see: (a) Hunter, R.; Richards, P. Org. Biomol. Chem. 2003, 1, 2348–2356; (b)
Hunter, R.; Richards, P. Tetrahedron Lett. 2000, 41, 3755–3758; (c) Hunter, R.;
Clauss, R. J. Chem. Soc., Perkin Trans. 1 1997, 41, 71–76.
10. To the best of our knowledge, this is the first time this bond has been installed
by radical means. Generally, indoloquinolizidine synthesis via the C3–C14
bond involves annulation or cycloaddition onto 3,4-dihydro-b-carboline or its
derivatives, which simultaneously forms the N4–C21 bond. For examples, see:
(a) Nagata, K.; Sekishiro, Y.; Itoh, T. Heterocycles 2007, 72, 175–179; (b) Itoh, T.;
Yokoya, M.; Miyauchi, K.; Nagata, K.; Ohsawa, A. Org. Lett. 2006, 8, 1533–1536;
(c) Oppolozer, W.; Hauth, H.; Pfaffli, P.; Wenger, R. Helv. Chim. Acta 1977, 60,
1801–1810.
We thank the University of Cape Town and the South African
National Research Foundation for funding.
References and notes
1. (a) Saxton, J. E. Nat. Prod. Rep. 1997, 14, 559–590; (b) Herbert, R. B.; Brown, R. T.;
Joule, J. A.; Saxton, J. E.; Cordell, G. A.; Creasey, W. A. In Indoles. The
Monoterpenoid Indole Alkaloids; Saxon, J. E., Ed. In The Chemistry of Heterocyclic
Compounds; Weissberger, A., Taylor, E. C., Eds.; John Wiley & Sons: New York,
1983; Vol. 25, Part 4; Chapters 1, 3, 4, 5, 9, 11 and 14.
2. For recent examples, see: (a) Wehner, F.; Nören-Müller, A.; Müller, O.; Reis-
Corrêa, I., Jr.; Giannis, A.; Waldmann, H. ChemBioChem 2008, 9, 401–405; (b)
Reis-Corrêa, I., Jr.; Nören-Müller, A.; Ambrosi, H.; Jakupovic, S.; Saxena, K.;
Schwalbe, H.; Prinz, H.; Kaiser, M.; Waldmann, H. Chem. Asian J. 2007, 2, 1109–
1126; (c) Takayama, H.; Ishikawa, H.; Kurihara, M.; Kitajima, M.; Aimi, N.;
Ponglux, D.; Koyama, F.; Matsumoto, K.; Moriyama, T.; Yamamoto, L. T.;
Watanabe, K.; Murayama, T.; Horie, S. J. Med. Chem. 2002, 45, 1949–1956.
11. For initial studies on the cyclization of
phenylthio precursors, see: (a) Hart, D. J.; Tsai, Y.-M. J. Am. Chem. Soc. 1982, 104,
1430–1432; for more recent examples of -acylamino radical cyclizations, see:
a-acylamino radicals generated from a-
a
(b) Chen, M.-J.; Tsai, Y.-M. Tetrahedron Lett. 2007, 48, 6271–6274; (c) Shizaki,
M.; Kai, Y.; Hoshino, O. Heterocycles 2002, 57, 2279–2297.
12. Whittaker, N. J. Chem. Soc. (C) 1969, 85–89.

6346
M. W. Smith et al. / Tetrahedron Letters 50 (2009) 6342–6346
13. All new compounds delivered 1H and 13C NMR and IR spectral data consistent
with the assigned structures, as well as satisfactory microanalyses or accurate
HRMS masses.
14. The small amount of Z-isomer could be separated by careful chromatography
after conversion into 8 but was more conveniently carried through the radical
cyclization, as both isomers were converted into product. Moreover, alkene
isomerization is known to occur under tributyltin radical conditions, see, for
example: Lowinger, T. B.; Weiler, L. J. Org. Chem. 1992, 57, 6099–6101.
15. Indole protection was necessary as radical cyclization of free indole 17 resulted
in a complex mixture (>6 products), which only contained a small amount of
the desired lactam 23 (<10%).
C24H31N2O5 requires 427.2233. trans-6 (minor): IR (thin film) 2979, 2927,
1727, 1648, 1456, 1419, 1370, 1357, 1307, 1156, 1139, 755 cmꢁ1 1H NMR
;
(400 MHz, CDCl3) d 7.96 (d, J = 8.0 Hz, 1H, H-11), 7.45 (dd, J = 7.5, 0.7 Hz, 1H, H-
8), 7.37–7.24 (m, 2H, H-9 + H-10), 5.11 (d, J = 1.8 Hz, 1H, H-12b), 5.02 (m, 1H,
1 ꢃ H-6), 4.11 (q, J = 7.1 Hz, 2H, OCH2CH3), 2.98–2.61 (m, 6H, CH2CO + 1 ꢃ H-
6 + H-1 + H-7), 2.55 (m, 1H, 1 ꢃ H-3), 2.40 (dt, J = 8.1, 4.3 Hz, 1H, 1 ꢃ H-3),
1.78–1.65 (m, 11H, OC(CH3)3 + H-2), 1.25 (t, J = 7.1 Hz, 3H, OCH2CH3); 13C NMR
(100 MHz, CDCl3) d 171.8 (CH2CO), 170.6 (C-4), 150.9 (NCO2t-Bu), 136.6 (C-
11a), 135.0 (C-12a), 129.0 (C-7b), 124.7 (C-10), 123.1 (C-9), 120.7 (C-7a), 118.3
(C-8), 115.5 (C-11), 84.7 (OC(CH3)3), 60.5 (OCH2CH3), 60.3 (C-12b), 41.8 (C-6),
36.8 (C-1), 35.1 (CH2CO), 28.3 (C-3), 28.2 (OC(CH3)3), 22.2 (C-2), 21.5 (C-7), 14.2
(OCH2CH3); HRMS (ESI): [M+H]+, 427.2236; C24H31N2O5 requires 427.2233.
18. The molecule of 19 selected from the racemate by the X-ray analyst had
absolute stereochemistry opposite to that depicted in the preceding schemes.
Crystallographic data (excluding structure factors) for the structure in this
paper have been deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC 737199. Copies of the data can be
obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge
CB2 1EZ, UK, (fax: +44 (0)1223 336033 or e-mail: deposit@ccdc.cam.ac.uk).
19. Attempts to test this model with the less sterically demanding methoxy-
carbonyl-protected analogue of 8 were complicated by the instability of this
group to the cyclization conditions. Nevertheless, the methoxycarbonyl-
protected analogue of cis-6 could be isolated as the major product in 53% yield.
20. Awang, K.; Pa, M.; Sévenet, T.; Schaller, H.; Nasir, A. M.; Hadi, A. H. A.
Phytochemistry 1991, 30, 3164–3167.
16. Although only one rotamer of 8 has the correct configuration to undergo
radical cyclization, equilibration to the reactive rotamer seemingly occurs
under the reaction conditions, as all starting material is consumed to form
product.
17. 12-(tert-Butoxycarbonyl)-1-(ethoxycarbonylmethyl)-4-oxo-1,2,3,4,6,7,12,12b-
octahydroindolo[2,3-a]quinolizine (6). [Numbering for NMR assignments
follows the IUPAC name] cis-6 (major): IR (thin film) 2931, 1732, 1645, 1456,
1410, 1369, 1309, 1141, 752 cmꢁ1
;
1H NMR (400 MHz, CDCl3) d 8.08 (d,
J = 8.3 Hz, 1H, H-11), 7.44 (d, J = 7.7 Hz, 1H, H-8), 7.31 (ddd, J = 8.3, 7.3, 1.4 Hz,
1H, H-10), 7.25 (m, 1H, H-9), 5.39 (d, J = 2.0 Hz, 1H, H-12b), 5.14 (ddd, J = 12.0,
4.1, 1.2 Hz, 1H, 1 ꢃ H-6), 3.82 (dq, J = 10.9, 7.1 Hz, 1H, 1 ꢃ OCH2CH3), 3.68 (dq,
J = 10.9, 7.1 Hz, 1H, 1 ꢃ OCH2CH3), 3.23 (m, 1H, H-1), 2.84-2.65 (m, 3H, H-
7 + 1 ꢃ H-6), 2.55 (m, 2H, H-3), 2.20–2.03 (m, 2H, 1 ꢃ H-2 + 1 ꢃ CH2CO), 1.96-
1.82 (m, 2H, 1 ꢃ CH2CO + 1 ꢃ H-2), 1.72 (s, 9H, OC(CH3)3), 0.99 (t, J = 7.1 Hz, 3H,
OCH2CH3); 13C NMR (100 MHz, CDCl3) d 171.9 (CH2CO), 169.6 (C-4), 150.0
(NCO2t-Bu), 136.7 (C-11a), 132.0 (C-12a), 128.1 (C-7b), 124.8 (C-10), 122.9 (C-
9), 120.2 (C-7a), 118.2 (C-8), 115.8 (C-11), 84.4 (OC(CH3)3), 60.3 (OCH2CH3),
58.2 (C-12b), 38.3 (C-6), 33.0 (C-1), 31.8 (CH2CO), 28.2 (OC(CH3)3), 28.0 (C-3),
24.1 (C-2), 21.3 (C-7), 13.8 (OCH2CH3); HRMS (ESI): [M+H]+, 427.2212;
21. Lounasmaa, M.; Karvinen, E. Heterocycles 1993, 36, 751–760.
22. Husson, H.-P.; Imbert, T.; Thal, C.; Potier, P. Bull. Soc. Chim. Fr. 1973, 2013–2016.
Lounasmaa et al. prepared 22 according to this procedure but do not provide a
melting point.