A copper-catalyzed domino radical cyclization route to benzospiro-indolizidinepyrrolidinones

Article abstract of DOI:10.1016/j.tetlet.2007.07.211

In this Letter the synthesis of benzospiro-indolizidinepyrrolidinones is described by a domino atom transfer radical cyclization reaction using a copper catalyst. The structure of one of the products was established by single crystal X-ray diffraction. Th

Full text of DOI:10.1016/j.tetlet.2007.07.211

Tetrahedron Letters 48 (2007) 7108–7111
A copper-catalyzed domino radical cyclization route
to benzospiro-indolizidinepyrrolidinones
Christian V. Stevens,a,* Ellen Van Meenen,a Kurt G. R. Masschelein,a Yves Eeckhout,a
Wim Hooghe,a Bart D’hondt,a Victor N. Nemykinb and Viktor V. Zhdankinb
aDepartment of Organic Chemistry, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium
bDepartment of Chemistry and Biochemistry, University of Minnesota Duluth, Duluth, MN 55812, USA
Received 2 July 2007; revised 27 July 2007; accepted 31 July 2007
Available online 3 August 2007
Abstract—In this Letter the synthesis of benzospiro-indolizidinepyrrolidinones is described by a domino atom transfer radical cycli-
zation reaction using a copper catalyst. The structure of one of the products was established by single crystal X-ray diffraction. The
investigated precursors, bearing a homo allyl substituent on the N-indole, result in a 5-exo-trig, followed by a 6-endo-trig cyclization.
When the N-indole is substituted with an allyl group, only the spiro-cyclization occurs.
Ó 2007 Elsevier Ltd. All rights reserved.
Indoles form extremely important and basic skeletons in
many biologically active natural products. Next to the
need for new effective methods for the construction of
the indole moiety, there is a continuing search for new
functionalization strategies. In our ongoing research
using the copper-catalyzed atom transfer radical cycliza-
tion (ATRC)1 for the formation of lactams,2 the indole-
skeleton was already evaluated as a substrate to synthe-
size spiro-indoles.3 With this report we would like to
extend the scope of this reaction using a domino radical
cyclization reaction for the synthesis of benzospiro-
indolizidinepyrrolidinones.
ring skeletons. Both stages of the reaction are intramo-
lecular processes, which result in the formation of two
extra ring systems (Scheme 1).
This strategy can be useful for the synthesis of natural
products related to strychnine 3 (Fig. 1) comprising a
1,2-annulated piperidine and a spiro ring as well.6
To the best of our knowledge, only one paper described
in the literature reports an analogous skeleton as
a side product in their reaction.7 Nevertheless the
R
R
O
N
Domino and tandem catalysis reactions are defined
involving ‘two or more bond-forming transformations,
which take place under the same reaction conditions,
without adding additional reagents and catalysts’. When
the multiple transformations are effected via a single cat-
alytic mechanism, the reaction is called a domino reac-
tion, while otherwise the term tandem reaction is
used.4 Within the domino and tandem reactions, the
radical domino reaction occupies an important posi-
tion.5 In the present Letter, the utility of domino radical
cyclizations is illustrated to quickly build up multiple
N
CCl3
Cl
Cl
O
Cu(I)Cl
TMEDA
N
N
Cl
1
2
R
R
N
O
O
N
Cl
Cl
Cl
Cl
N
N
Keywords: Cu(I)Cl; Atom transfer radical cyclization; ATRC; Domino
reaction.
*
Corresponding author. Tel.: +32 9 264 59 57; fax: +32 9 264 62 43;
e-mail: Chris.Stevens@UGent.be
B
A
Scheme 1. Domino radical cyclization.
URL: http://www.synbioc.ugent.be.
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.07.211

C. V. Stevens et al. / Tetrahedron Letters 48 (2007) 7108–7111
7109
Table 1. Conversion of aldehydes 6 to alkylated amines 9
Alkylation Imination Reduction
Product Y (%) Product Y (%)
NHCOR
N
R1
Product
Y (%)
H
H
7a
92
8a
8b
8c
8d
8e
98
94
99
99
98
9a
9b
9c
9d
9e
94
96
99
92
95
N
N
O
O
4 R1 = H
5 R1 = OMe
R = CH3, n-C3H7
3
H
7d
7e
90
90a
a For the more expensive 5-bromo-2-methyl-pent-2-ene only 1.02 equiv
was used.
Figure 1. Some interesting analogues.
tetrahydro-pyrido[1,2-a]indole core is often described in
the literature.8 For example, compounds 4 and 5,
analogues of melatonin, show to be potent antagonists
(4) or agonists (5) of this vertebrate hormone.9
side products and alkylated indoles 7 could be isolated
in high yields (P90%) and purities (>95%) (Table 1).
Reaction of 7 with an amine and magnesium sulfate
(3 equiv) in dichloromethane as a solvent under reflux
resulted, after laborious filtration and evaporation, in
the corresponding imines 8 in almost quantitative yields
(94–99%) (Table 1). When non-volatile amines were
used, the use of an excess of amine was impossible
because of purification problems, which therefore
resulted in longer reaction times. Previous research
shows that the reduction of these imines was not
straightforward.3 Only the use of sodium borohydride
in combination with ethanol as a solvent led to satisfy-
ing results. Using these optimized reductive reaction
conditions the secondary amines were recovered in 92–
99% yield (Table 1). The next step in the preparation
of the ATRC-precursors was the acylation with trichlo-
roacetyl chloride. Pyridine was chosen as a base and
tetrahydrofuran as a suitable solvent. This reaction pro-
ceeded straightforwardly (3–24 h at room temperature),
The synthetic strategy (Scheme 2) starts from the com-
mercially available indole-3-carbaldehyde (or deriva-
tives) 6. After the alkylation of the N-indole with an
appropriate alkenyl bromide, imine 8 was formed and
subsequently reduced to secondary amine 9. The latter
was acylated with trichloroacetyl chloride to form amide
1 as a suitable precursor for the ATRC reaction.
N-Alkylation of indole-3-carbaldehydes 6 was first
attempted using sodium hydride in THF. However, no
complete conversion took place. Literature search
revealed a promising alternative using a 3–5 M excess
of potassium carbonate in acetone and a similar excess
of alkenyl bromide.10 No remaining starting material
was present, but some side products appeared in the
reaction mixture. Reflux (24–29 h) in acetonitrile as an
aprotic solvent for this reaction shows no formation of
R3
O
O
N
H
H
H
R1
i
R1
ii
R1
94-99%
90-92 %
N
N
N
H
R2
R2
8a-e
7a,d,e
6a,d
R2
R2
a R1=H; R2=H; R3=Bn
b R1=H; R2=H; R3=
Hex
c R1=H; R2=H; R3=
Pr
c
i
92-99%
iii
d R1=OMe; R2=H; R3=
iPr
e R1=H; R2=Me; R3=Bn
R3
N
R3
NH
R3
N
O
O
R1
R1
a,c,d
R1
Cl
v
iv
Cl
CCl3
20-39%
Cl
50-65%
N
N
N
R2
R2
2a-d
1a,c,d
R2
R2
9a-e
b,e
R2 = H
vi
40%
Scheme 2. Synthesis of benzospiro-indolizidinepyrrolidinones 2. Reagents and conditions: (i) 5 equiv K2CO3—1.02–3 equiv alkenyl bromide—24–
29 h reflux in acetonitrile; (ii) 1.03–3 equiv amine—0.6–3 equiv MgSO4—21–47 h reflux in CH2Cl2; (iii) 1 equiv NaBH4—17–28 h in EtOH at rt; (iv)
2 equiv CCl3COCl—2 equiv pyridine—2–24 h in THF (N2) (rt or reflux); (v) 0.4 equiv Cu(I)Cl—0.8 equiv TMEDA—4–5 h reflux in CH2Cl2 or
CH3CN (Ar); (vi) (a) 1 equiv CCl3COCl—1.1 equiv pyridine—1 h at rt in Et2O (N2); (b) filtration–evaporation Et2O—adding CH3CN—0.4 equiv
Cu(I)Cl—0.8 equiv TMEDA—21 h reflux (Ar).

7110
C. V. Stevens et al. / Tetrahedron Letters 48 (2007) 7108–7111
Table 2. Conversion of alkylated amines
lizidinepyrrolidinones 2
9
to benzospiro-indo-
Cl126
Cl127
Cl125
C111
Acylation
ATRC
O124
C101
C105
C104
C107
N106
C119
Product
Y (%)
Product
Time (h)
Solvent
Y (%)
C120
C121
C108
1a
1b
1c
1d
1e
65
—
50
51
—
2a
2b
2c
2d
2e
4
21
4
CH3CN
CH3CN
CH2Cl2
CH2Cl2
39
C110
a
40b
27
N102
C117
C109
C103
C112
C118
C123
5
20
C113
C114
a
Complex reaction mixture
C122
a Product could not be isolated in a pure form.
b Yield of two steps: ATRC was performed without isolating the
substrate 1b.
C115
C214
C213
C215
O224
C219
C220
Cl225
C216
C212
C201
C211
but problems arose during work-up. Previous research
already illustrated that these compounds are light-sensi-
tive.3 In this case however, breakdown already started
during evaporation of the solvent after aqueous extrac-
tion. As an alternative work-up procedure, the majority
of the pyridinium hydrochloride salts were filtered and
most of the tetrahydrofuran was evaporated. Subse-
quently dry methanol was added and the product was
allowed to crystallize in the freezer. Unfortunately, this
was only an appropriate alternative for a selected num-
ber of derivatives (see Table 2). In the other cases (b,e)
the acylated product could not be isolated in a pure
form to perform the atom transfer radical cyclization
(vide infra).
C217
N206
C221
C222
C218
N202
C203
C204
C205
C207
C208
C222
Cl226
C210
C209
Cl227
Figure 2. CAMERON representation of X-ray structure of compound
2b. The triclinic unit cell consist of four molecules of 2b, from which
two are unique. The unique molecule 1 labelled using 1xx numbers,
while unique molecule 2 labelled using 2xx numbers. All three
stereocenters in molecule 1 are R, while those in molecule 2 are S.
In the final step, the obtained precursors are evaluated
in the atom transfer radical cyclization reaction
(ATRC) using the cheap and environmental friendly
catalytic system Cu(I)Cl/TMEDA (tetramethyl ethylene
diamine). The optimized reaction conditions are listed in
Table 2. When reactions were performed at room tem-
perature, a lot of side products were formed because
of the long reaction times. The latter were shortened
using reflux temperatures resulting in cleaner reaction
mixtures. The end products 2 were isolated via crystalli-
zation using methanol or diethyl ether after aqueous
work-up. Although the 5-exo-cyclization is kinetically
favored,11,12 only 5-exo-6-endo domino products could
be isolated. The structure of product 2b was established
by single crystal X-ray analysis and is shown in Figure
2.13 Interestingly, 2b crystallized as a racemic mixture
of two enantiomers in the unit cell. Indeed, the absolute
configurations for all three stereocenters in unique mol-
ecule 1 is R, while that for all three stereocenters in mol-
order to obtain the ring closed products. In this case
equimolar amounts of acid chloride and 1.1 equiv of
pyridine were used in Et2O for the acylation reaction.
In this manner no excess of reagents is present, which
could interfere during the ATRC reaction. The solvent
was altered from THF to diethyl ether because of the
higher volatility. After the reaction, the pyridine hydro-
chloride salts were filtered and three quarters of the
solvent was evaporated. After addition of acetonitrile,
the remaining Et2O was removed under reduced pres-
sure. In this way the acylated products 1 were not con-
centrated, which prevented breakdown. Subsequently,
Cu(I)Cl and N-ligand were added. With this strategy a
40% yield was obtained for the cyclohexyl substituted
derivative 2b for the one pot acylation–ATRC reaction.
Unfortunately, only a complex reaction mixture was
obtained starting from amine 9e. Due to the steric
hindrance and the more stabilized tertiary radical, a 5-
exo-5-exo domino cyclization could be expected, how-
ever more research is needed to substantiate this
hypothesis.
ecule
2 is S. As expected, both five-membered
heterocyclic fragments in 2b adopt an envelope confor-
mation, while both six-membered substituents are
locked in the most energetically favored chair conforma-
tion. In both unique molecules, chlorine atoms attached
to the six-membered piperidine fragments were found in
the axial position. The crude 1H NMR spectra also
shows the presence of a considerable amount of an inter-
mediate product where only the 5-exo-spirocyclization
occurred. Longer reaction times however resulted in
complex reaction mixtures.
To investigate the regioselectivity of the second cycliza-
tion, an allyl substituent was introduced on the N-indole
instead of a homo allyl group. The ATRC precursor 10
was prepared following the above mentioned proce-
dures. However, evaluating solvents (CH3CN, CH2Cl2),
N-ligands (tetramethyl ethylene diamine, pentamethyl
diethylene triamine) and different time-temperature
combinations, no domino cyclization was observed
(Scheme 3). The lack of the domino cyclization might
As mentioned before, precursors 1b and 1e could not be
isolated. Therefore, another procedure was developed in

C. V. Stevens et al. / Tetrahedron Letters 48 (2007) 7108–7111
7111
V.; Grandi, R.; Libertini, E.; Pagnoni, U. M.; Schenetti, L.
Tetrahedron 2003, 59, 1147.
3. Stevens, C. V.; Van Meenen, E.; Eeckhout, Y.; Vander-
hoydonck, B.; Hooghe, W. Chem. Commun. 2005, 4827.
4. Fogg, D. E.; dos Santos, E. N. Coord. Chem. Rev. 2004,
248, 2365.
O
N
N
N
0.4 eq. Cu(I)Cl
0.8 eq. N-ligand
O
Cl
Cl
X
CCl3
CH2Cl2 or CH3CN
N
5. Albert, M.; Fensterbank, L.; Lacoˆte, E.; Malacria, M.
Top. Curr. Chem. 2006, 264, 1–62.
11 X = Cl
12 X = OH
10
Scheme 3. ATRC of N-allyl-precursor 10.
6. Bodwell, G. J.; Li, J. Org. Lett. 2002, 4, 127.
7. Veenstra, S. J.; Speckamp, W. N. J. Am. Chem. Soc. 1981,
103, 4645.
8. Some examples describing compounds containing the 1,2-
annulated indole core structure: (a) Dobbs, A. P.; Jones,
K.; Veal, K. T. Tetrahedron 1998, 54, 2149; (b) Moody, C.
J.; Norton, C. L. J. Chem. Soc., Perkin Trans. 1 1997,
be explained in perspective to the Baldwin rules stating
that a 5-endo-trig cyclization is disfavored. The other
option, a 4-exo-trig, is favored, but needs to overcome
a considerable ring strain for the four-membered ring
formation.
´
2639; (c) Miranda, L. D.; Cruz-Almanza, R.; Pavon, M.;
Romero, Y.; Muchowski, J. M. Tetrahedron Lett. 2000,
41, 10181; (d) Wang, S.-F.; Chuang, C.-P.; Lee, W.-H.
Tetrahedron 1999, 55, 6109.
In all cases, a mixture of spirocompounds 11 and 12 was
formed. 2-Hydroxydihydroindole 12 was formed during
aqueous work-up.
9. Tsotinis, A.; Panoussopoulou, M.; Hough, K.; Sugden, D.
Eur. J. Pharm. Sci. 2003, 18, 297.
10. Dobbs, A. P.; Jones, K.; Veal, K. T. Tetrahedron Lett.
1995, 36, 4857.
In the current reactivity study of the obtained tetracyclic
compounds 2 the possibility to open the lactam ring is
investigated in order to produce a c-aminobutyric acid
analogue. This pathway is now under investigation,
however the first step, that is, the reduction of the two
chloro atoms was already successful. Refluxing of benz-
ospiro-indolizidinepyrrolidinone 2b in THF in the pres-
ence of 10 equiv of glacial acetic acid and 5 equiv of zinc
resulted in a 56% yield of the reduced compound.
11. Beckwith, A. L. J.; O’Shea, D. M. Tetrahedron Lett. 1986,
27, 4525.
12. Crich, D.; Mook, R., Jr. Tetrahedron Lett. 1996, 37, 3105.
0
13. Spectroscopic data for one selected example: N1 -cyclo-
hexyl-3,40,40-trichloro-50-oxo-10,20,40,50-tetrahydrospiro-
benzo[6,7-a]indolizidine-5,30-pyrrole 2b:
1' N
O
5'
In summary, the synthesis of benzospiro-indo-
lizidinepyrrolidinones 2 is described by a domino atom
transfer radical cyclization reaction using a copper cata-
lyst. The investigated precursors, bearing a homo allyl
substituent on N-indole, result in a 5-exo-trig, followed
by a 6-endo-trig cyclization. The structure of a key prod-
uct was established by single crystal X-ray diffraction.
When N-indole is substituted with an allyl group, only
the spiro-cyclization occurs. Further work on the reac-
tivity of these tetracyclic compounds 2, is currently
under investigation.
2'
4' Cl
3'
6
5
Cl
4
4a
7
N
Cl
3
1
2
1H NMR (300 MHz, CDCl3, ppm): 1.01–1.84 (10H, m, 5x
CH2(cHex)); 1.94–2.07 (3H, m, NC1H2C2H2 and C4HH);
2.44 (1H, 0 d, J = 13.5 Hz, C4HH); 3.29 (1H, d, JAB
=
9.6 Hz, C2 HH); 3.33–3.40 (1H, m, C1HH); 3.49–3.60 (1H,
0
m, C1HH); 3.63 (1H, d, JAB = 9.6 Hz, C2 HH); 4.07 (1H, t,
J = 11.5 Hz, CH(cHex)); 4.30 (1H, dxd, J = 11.5 Hz,
J2 = 2.5 Hz, C4aH); 4.67 (1H, t, J = 3.0 Hz, C3HCl); 6.53
(1H, d, J = 7.7 Hz, CH(Ph)); 6.64 (1H, dxd, J1 =
J2 = 7.7 Hz, CH(Ph)); 6.94 (1H, d, J = 7.7 Hz, CH(Ph));
7.18 (1H, dxd, J1 = J2 = 7.7 Hz, CH(Ph)); 13C NMR
(75.6 MHz, CDCl3, ppm): 25.09 (CH2(cHex)); 25.19
(CH2(cHex)); 25.25 (CH2(cHex)); 29.60 (CH2(cHex));
Acknowledgments
We are grateful to the Fund of Scientific Research-Flan-
ders (FWO-Vlaanderen) and Ghent University for
financial assistance of this research.
29.79 (CH2(cHex)); 31.94 (C2); 35.54 (C4); 39.28 (C1);
0
47.47 (C2 ); 52.13 (CH(cHex)); 56.79 (C3); 59.73 (C5);
0
60.06 (C4a); 89.14 (C4 ); 106.83 (C(Ph)); 118.22 (C(Ph));
123.99 (C(Ph)); 129.32 (C6 or C7); 129.81 (C(Ph)); 149.43
References and notes
0
(C6 of C7); 165.52 (C5 ); IR (KBr, cmꢀ1): 2932, 2856, 1724,
1478; MS: m/z = 427, 429, 431, 433 (MH+); melting point:
207–208 °C. Single crystals of compound 2b were grown
from chloroform. CCDC-652072 contains the supplemen-
tary crystallographic data for this Letter. These data can
be obtained free of charge from The Cambridge Crystal-
lographic Data Centre via www.ccdc.cam.ac.uk/datare-
quest/cif.
1. For an excellent review on copper-catalysed ATRC-
reaction: Clark, A. J. Chem. Soc. Rev. 2002, 31, 1.
2. (a) Bellesia, F.; De Buyck, L.; Colucci, M. V.; Ghelfi, F.;
Laureyn, I.; Libertini, E.; Mucci, A.; Pagnoni, U. M.;
Pinetti, A.; Rogge, T. M.; Stevens, C. V. Tetrahedron Lett.
2001, 43, 4573; (b) Ghelfi, F.; Stevens, C. V.; Laureyn, I.;
Van Meenen, E.; Rogge, T. M.; De Buyck, L.; Nikitin, K.