Acyl radical cyclisation onto pyrroles

Article abstract of DOI:10.1016/S0040-4039(01)01639-2

Synthetically useful [1,2-a]-fused pyrroles, e.g. 2,3-dihydro-1H-pyrrolizidines substituted in the 1- and 7-positions, have been generated by acyl radical cyclisation onto pyrroles using N-(ω-acyl)-radicals generated from acyl-selenide precursors. The protocol does not require high pressures of CO. Mechanistic studies indicate the key role of azo radical initiators as oxidants of the intermediate π-radicals.

Full text of DOI:10.1016/S0040-4039(01)01639-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 7887–7890
Acyl radical cyclisation onto pyrroles
Steven M. Allin,a,* William R. S. Barton,a W. Russell Bowmana and Tom McInallyb,*
aDepartment of Chemistry, Loughborough University, Loughborough LE11 3TU, UK
bAstraZeneca R&D Charnwood, Bakewell Road, Loughborough LE11 5RH, UK
Received 23 July 2001; accepted 30 August 2001
Abstract—Synthetically useful [1,2-a]-fused pyrroles, e.g. 2,3-dihydro-1H-pyrrolizidines substituted in the 1- and 7-positions, have
been generated by acyl radical cyclisation onto pyrroles using N-(v-acyl)-radicals generated from acyl-selenide precursors. The
protocol does not require high pressures of CO. Mechanistic studies indicate the key role of azo radical initiators as oxidants of
the intermediate p-radicals. © 2001 Elsevier Science Ltd. All rights reserved.
The synthesis of nitrogen heterocycles using radical
intermediates has become a common protocol in modern
organic chemistry.1 One of the protocols that has proved
useful is the Bu3SnH and AIBN [azobisisobutyronitrile,
or by IUPAC nomenclature, 2-(1-cyano-1-methyl-ethyl-
azo)-2-methyl-propio-nitrile] mediated oxidative cyclisa-
tion of alkyl or aryl radicals onto heteroarenes. Our
earlier studies have reported the cyclisation of N-(v-
alkyl)-radicals onto pyrroles and imidazoles with elec-
tron withdrawing groups.2 Similar cyclisation protocols
onto indoles3–5 pyrroles,3,6 pyridinium salts,7,8 pyridines9
and 1,2,3-triazoles10 have also been reported. Normally
in Bu3SnH mediated reactions reductive cyclisation
takes place, whereas in these reactions there is also an
oxidative step resulting in rearomatisation. The mecha-
nism of this oxidative step is not clear but the most
recent evidence indicates hydrogen abstraction from a
p-radical intermediate.11
electron-deficient heteroarenes using nucleophilic radi-
cals as opposed to Friedel–Crafts alkylation of electron
rich heteroarenes using cationic electrophiles. The logical
extension of this work was to study radical cyclisation
onto heteroarenes using acyl radicals, i.e. the umpolung
of Friedel–Crafts acylation. Acyl radicals are also
nucleophilic12 and similar results were predicted. Acyl
radicals have been extensively studied, both mechanisti-
cally and in synthesis.12 During our study, related
research was published; the cyclisation of acyl radicals,
generated by carbon monoxide addition under high
pressure onto intermediate N-(v-alkyl)-radicals, onto
pyrroles and indoles.13 The experimental drawback of
this methodology is the requirement for CO at 80 atm
in the reactions. In this paper we report our initial results
showing that cyclisation of acyl radicals can be carried
out in high yields from acyl selenide precursors without
using high pressures of CO. In this paper we report our
initial studies to develop a facile protocol for the synthe-
sis of [1,2-a]-fused pyrroles, substituted in both rings,
using tributyltin hydride (Bu3SnH) (Scheme 1).
These alkylation reactions can be considered to be the
umpolung of Friedel–Crafts alkylation, i.e. alkylation of
CHO
O
CHO
CHO
H
CHO
Bu3Sn•
•
(– H•)
O
O
N
N
N
N
•
Bu3SnSePh
( )n
( )n
( )n
( )n
O
SePh
1
2
3
4
Scheme 1. Cyclisation of N-(v-acyl) radicals. a, n=1; b, n=2; c, n=3.
Keywords: acyl radicals; radical cyclisation; acyl selenides; heteroarenes; tributyltin hydride.
* Corresponding authors. Tel.: +44 1509 222 569; fax: +44 1509 223 925; e-mail: w.r.bowman@lboro.ac.uk
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01639-2

7888
S. M. Allin et al. / Tetrahedron Letters 42 (2001) 7887–7890
Acyl selenides have proved to be useful precursors of
acyl radicals12,14 and appeared suitable for our putative
protocol. 1H-Pyrrole-2-carbaldehyde was used to
develop the procedure for the synthesis of radical pre-
cursors 7 (a, n=1; b, n=2; c, n=3). Standard alkyla-
tion with v-bromo esters gave high yields and purity
using DMF as a solvent and lower yields when THF
was used. The esters 5 were hydrolysed without purifi-
cation to yield the required carboxylic acids 6. Both
diphenyl diselenide and N-(phenylselenyl)phthalimide
(NPSP)15 were used for the selenation, but the latter
gave cleaner reactions.
nique to ensure that CO was the only gas present. The
atmosphere of CO, together with the use of a two-phase
solvent system of acetonitrile and cyclohexane and
heating at 80°C, gave the best yields of the required
cyclised products 14 with only traces of products from
CO-loss and reduction. The precursors 7a–c and AIBN
are soluble in acetonitrile and not cyclohexane, whereas
the Bu3SnH is soluble in the cyclohexane thereby keep-
ing a low concentration in the acetonitrile which facili-
tates cyclisation over reduction. The yields were as
follows: 14a (55%, 2 h syringe pump addition of
reagents) and 14c [31%, reagents added in two portions
at the beginning and after 4 h and sealed in a Schlenk
tube and heated at 80°C for a total of 8 h). In the latter
reaction the acyl radical intermediate 8c was also
reduced to the uncyclised aldehyde 12c (46%).
The cyclisations were initially studied using ‘normal’
conditions for Bu3SnH mediated oxidative radical reac-
tions [Bu3SnH (2.2 equiv., addition by syringe pump
over 5–6 h), AIBN or AMBN [azobismethylisobutyro-
nitrile, or by IUPAC nomenclature, 2-(1-cyano-1-
methyl-propylazo)-2-methyl-butyronitrile] (2.0 equiv.,
added portion-wise every 30 min), cyclohexane, reflux
under N2] (Scheme 3). The expected cyclised ketones 14
were obtained in each case in moderate yield along with
a mixture of other products. Significant decarbonyla-
tion took place from the acyl radical intermediates 8a,b
to yield the alkyl radical intermediates 9a,b. The pre-
cursor 7b gave cyclisation to the known pyrrolizidine
10b2 and 7a gave reduction to 11a.2 Alternative condi-
tions were determined because these reactions were not
satisfactory for synthesis.
Precursor 7b gave a reduced product, 3-(hydroxy-
methyl)-5,6,7,8-tetrahydroindolizin-8-one 15 (65%, all
reagents added at the beginning of the reaction and
sealed in a Schlenk tube). The formation of a reduced
product 15 from 7b is unusual and suggests a ‘normal’
reductive Bu3SnH reaction as shown in Scheme 4. The
product 15 was not present in the crude reaction mix-
ture and only formed after extraction of products in
dilute hydrochloric acid indicating that a basic product
was formed in the reaction, e.g. 17. Rapid tautomerism
would yield the more stable pyrrole 15. A possible
explanation is that the O-centred canonical form 16 has
an electrophilic centre, which reacts sufficiently fast
with the nucleophilic Bu3SnH to give a reduced product
such as 17. The six-membered ring cyclisation is under
much less strain than the five- and seven-membered
ring cyclisation and possibly facilitates this different
delocalisation of the unpaired electron in the p-radical
(13b/16).
Decarbonylation is an exothermic process and the rate
of CO loss (ca. 2×102 M−1 s−1 at 80°C to yield primary
alkyl radicals) is much slower than CO addition (6.3×
105 M−1 s−1 at 80°C for primary alkyl radicals).12 There-
fore, we reasoned that an atmospheric pressure of CO
may be sufficient to prevent decarbonylation and nitro-
gen was replaced with CO using the free-thaw tech-
(i), (ii)
(iii)
CO2Me
(iv)
OHC
OHC
OHC
CHO
N
H
N
N
N
O
CO2H
( )n
( )n
( )n
SePh
5
6
7
Scheme 2. Synthesis of 1H-pyrrole-2-carbaldehyde radical precursors. a, n=1; b, n=2; c, n=3. (i) NaH, dry DMF, 30 min, room
temperature; (ii) BrCH2(CH2)nCO2Me, 2 h at 0°C then 24 h at room temperature; (iii) LiOH, EtOH/H2O, 4 h, room temperature:
6a, n=1 (73%); 6b, n=2 (95%); 6c, n=3 (100%); (iv) Bu3P (2.5 equiv.), CH2Cl2, room temperature: (PhSe)2 (1.5 equiv.), 24 h: 7a,
n=1 (59%) and 7b, n=2 (57%); NPSP (1.5 equiv.), 6 h: 7c, n=3 (63%).
Bu3SnH
– CO
OHC
OHC
OHC
OHC
7
N
N
N
N
( )n
•
H
•
( )n
( )n
11 a (34%), b (0%)
( )n
O
8
9
10 a (0%), b (13%)
•
H
OHC
(– H•)
N
OHC
O
OHC
O
N
N
( )n
14 a (31%), b (20%)
( )n
12 a,b (0%)
( )n
O
13
Scheme 3. Cyclisation of pyrrole-2-carbaldehyde radical precursors 7 using ‘normal’ Bu3SnH conditions.

S. M. Allin et al. / Tetrahedron Letters 42 (2001) 7887–7890
7889
Bu3SnH
H
H
•O
HO
HO
O
O
O
13b
N
N
N
Bu3Sn•
15
16
17
Scheme 4. Cyclisation of pyrrole-2-carbaldehyde radical precursors 7 using a CO atmosphere.
isolated for the reaction of 1c. The six-membered ring
cyclisation from 1b to yield 4b did not yield any prod-
ucts resulting from decarbonylation showing that the
rate of cyclisation (2b to 3b) is faster than the corre-
sponding rates for the five- and seven-membered ring
cyclisations, i.e. cyclisation of 2b to 3b is faster than
loss of CO under the reaction conditions. Five-mem-
bered ring cyclisation onto heteroarenes is known to be
strained.2 The cyclisations proceeded with complete
regioselectivity to the 2-position on the pyrrole. The
structures of 4b and 4c have been confirmed by X-ray
crystallography.
The logic of our results suggests that the protocol could
be used to add CO to an intermediate alkyl radical 9a
to yield the acyl radical 8a with subsequent cyclisation
to 14a. This reaction has been reported in the literature
but with an 80 atm pressure of CO.13 A reaction was
attempted using our protocol with 1-(2-bromoethyl)-
1H-pyrrole-2-carbaldehyde but gave only 1-ethyl-1H-
pyrrole-2-carbaldehyde 11a (32%) with only traces of
cyclisation. Clearly the high pressure of CO is required
for the carbonylation of 9a even though favoured
thermodynamically.
The analogous 1H-pyrrole-3-carbaldehyde reactions
were also investigated with a view to the synthesis of
target pyrrolizidine natural products. 1H-Pyrrole-3-car-
baldehyde was synthesised from N-(triisopropylsilyl)-
Our synthetic protocol has been used to prepare a
natural product pyrrolizine, which is a pheromone from
the Lepidoptera family, Danainae genus. The synthesis
of other pyrrolizidine natural products is underway. We
tested whether cyclisation would take place without an
electron withdrawing group present on the pyrrole ring
in order to prepare nordanaidone 21.17 The radical
precursor 19 was synthesised using the methods in
Scheme 2 and reacted using our protocol to yield
nordanaidone 21 (23%, 16% from pyrrole) (Scheme 6).
Although the yield was low, cyclisation of the nucle-
ophilic acyl radical 20 onto an electron rich pyrrole ring
is possible.
pyrrole using
a
Vilsmeier reaction with oxalyl
chloride and DMF in 65% yield.16 The synthesis of the
precursors was carried out as for the corresponding
1H-pyrrole-2-carbaldehydes (Scheme 5). The acyl
selenides were formed in high yield but were unstable to
chromatography resulting in lower yields than for the
corresponding pyrrole-2-carbaldehydes. The cyclisa-
tions were carried out using the CO atmosphere and
two-phase solution protocol to give the predicted
cyclised compounds 4a–c in moderate yields. No
reduced uncyclised aldehydes [1-(3-oxoalkyl)-1H-
pyrrole-3-carbaldehydes] were formed in any of the
reactions but 1-ethyl-1H-pyrrole-3-carbaldehyde (17%)
was formed in the reaction of 1a. A cyclised product
18c (n=2) (54%), resulting from decarbonylation, was
The mechanism of cyclisation is shown in Schemes 1
and 3. SH2 Abstraction of the selenyl group with
’
Bu3Sn radicals to form acyl radicals 2 (or 8) and
cyclisation to the p-radical intermediates 3 (or 13) is
CHO
CHO
( )n
CHO
(i), (ii)
CHO
CHO
CHO
(iii)
CO2Me
(iv)
CO2H
(v)
O
N
N
N
H
N
N
N
O
( )n
( )n
16
( )n
17
( )n
SePh
4
1
18
Scheme 5. Synthesis and cyclisation of pyrrole-3-carbaldehyde radical precursors. (i) NaH, DMF; (ii) BrCH2(CH2)nCO2Me; (iii)
LiOH, EtOH/H2O: 17a, n=1 (69%); 17b, n=2 (90%); 17c, n=3 (89%); (iv) Bu3P, CH2Cl2, NPSP: 1a, n=1 (40%); 1b, n=2 (40%);
1c, n=3 (35%); (v) Bu3SnH (1.8 equiv.) in a solution of cyclohexane was added by syringe pump over 7 h and AIBN (for 1a) or
AIBMe (for 1b,c) (1.8 equiv.) added portion-wise over 5 h under an atmosphere of CO to a solution of 1 in acetonitrile under
reflux: 4a, n=1 (32%); 4b, n=2 (50%); 4c, n=3 (38%).
•
δ–
H
O
O
N
N
O
N
N
δ–
•
O
SePh
19
20
21 nordanaidone
Scheme 6. Synthesis of the pyrrolizine insect pheromone, nordanaidone.

7890
S. M. Allin et al. / Tetrahedron Letters 42 (2001) 7887–7890
CHO
H
CHO
3b
H
N
MeO2C
MeO2C
MeO2C
•
•
N
N
O
O
N
N
H
N
H
N
N
CO2Me
CO2Me
CO2Me
4b
22 AIBMe
24
23
3b
4b
Scheme 7. Role of ‘initiator’ in the oxidative step of the oxidative cyclisation. Bu3SnH (1.5 equiv.) and AIBMe (2.0 equiv.) were
added by syringe pump to a solution of 1b under reflux and an atmosphere of CO over 5 h and refluxed for further 3 h.
clear cut. The mechanism of rearomatisation is still
unclear and is discussed in Ref. 11. In the Bu3SnH
mediated oxidative cyclisations reported in the litera-
ture, greater than one equivalent of AIBN is required
suggesting that AIBN is involved with a H-abstraction
mechanism for this step. Azo compounds are known to
act as oxidants and therefore dihydro-AIBN would be
expected as a product, but is unstable to the reaction
conditions and cannot be isolated. Therefore, in initial
studies, AIBMe 22 [azobisisobutyrate methyl ester, or
by IUPAC nomenclature, 2-(1-methoxycarbonyl-1-
methyl-ethylazo)-2-methyl-propionic acid methyl ester]
was used (Scheme 7). Cyclisation of 1b gave the cyclised
product 4b (30%) and dihydro-AIBMe 23 (17%) (analy-
Soc., Perkin Trans. 1 2000, 1; (b) Aldabbagh, F.; Bow-
man, W. R. Contemporary Org. Synth. 1997, 4, 261.
2. (a) Aldabbagh, F.; Bowman, W. R.; Mann, E. Tetra-
hedron Lett. 1997, 38, 7937–7940; (b) Aldabbagh, F.;
Bowman, W. R.; Mann, E.; Slawin, A. M. Z. Tetrahedron
1999, 55, 8111.
3. Antonio, Y.; De La Cruz, E.; Galeazzi, E.; Guzman, A.;
Bray, B. L.; Greenhouse, R.; Kurz, L. J.; Lustig, D. A.;
Maddox, M. L.; Muchowski, J. M. Can. J. Chem. 1994,
72, 15–22.
4. Ziegler, F. E.; Belema, M. J. Org. Chem. 1997, 62,
1083–1094.
5. Moody, C. J.; Norton, C. L. J. Chem. Soc., Perkin Trans.
1 1997, 2639–2643.
1
sis by H NMR spectroscopy using an internal stan-
6. (a) Tim, C. T.; Jones, K.; Wilkinson, J. Tetrahedron Lett.
1995, 36, 6743–6744; (b) Dobbs, P. A.; Jones, K.; Veal,
K. T. Tetrahedron 1997, 53, 8287–8289.
7. Suzuki, F.; Kuroda, K. J. Heterocyclic Chem. 1993, 30,
811–813.
dard) indicating that AIBMe is responsible for the
oxidative step. The relative yields indicate the interme-
diate radical 24 oxidises a second equivalent of the
p-radical intermediate 3b to 4b. Alternatively, 24 is
reduced by Bu3SnH to complete a chain reaction by
8. Murphy, J. A.; Sherburn, M. S. Tetrahedron 1991, 47,
4077–4088.
’
generating Bu3Sn radicals.
9. Harrowen, D.; Nunn, M. I. T. Tetrahedron Lett. 1998,
39, 5875–5876.
10. Marco-Contelles, J.; Rodr´ıquez-Ferna´ndez, M. Tetra-
hedron Lett. 2000, 41, 381–384.
11. Bowman, W. R.; Mann, E.; Parr, J. J. Chem. Soc., Perkin
Trans. 1 2000, 2991–2999.
The products and yields (e.g. cyclisation of 13b to 14b
and 15) differ depending on the conditions used, e.g.
the use of syringe pump addition or sealed tubes, rate
of Bu3SnH addition, presence or absence of CO and the
amount of radical initiator (or oxidant). Further study
is underway to optimise reaction conditions, determine
the mechanism and to develop the protocol for natural
product synthesis.
12. Chatgilialoglu, C.; Crich, D.; Komatsu, M.; Ryu, I.
Chem. Rev. 1999, 99, 1991–2067.
13. Miranda, L. D.; Cruz-Almanza, R.; Paco´n, M.; Alva, E.;
Muchowski, J. M. Tetrahedron Lett. 2000, 40, 7153–7157.
14. (a) Pattenden, G.; Roberts, L.; Blake, A. J. J. Chem.
Soc., Perkin Trans. 1 1998, 863–868; (b) Double, P.;
Pattenden, G. J. Chem. Soc., Perkin Trans. 1 1998,
2005–2007 and references cited therein.
Acknowledgements
15. Nicolaou, K. C.; Petasis, N. A.; Claremon, D. A. Tetra-
hedron 1985, 41, 4835–4841.
16. Bray, B. L.; Mathies, P. H.; Naef, R.; Solas, D. R.;
Tidwell, T. T; Artis, D. R.; Muchowski, J. M. J. Org.
Chem. 1990, 55, 6317–6328.
17. Hartmann, T.; Witte, L. In Alkaloids: Chemical and Bio-
logical Persepectives; Pelletier, S. W., Ed.; Pergamon
Press: Oxford, 1996; Vol. 9, pp. 155–233. Recent synthe-
sis: Miranda, L. D.; Cruz-Almanza, R.; Alvarez-Garcia´,
A; Muchowski, J. M. Tetrahedron Lett. 2000, 40, 3035–
3038.
We thank AstraZeneca (R&D Charnwood) and the
EPSRC for an EPSRC CASE Postgraduate Research
Studentship (WRSB), AstraZeneca (R&D Charnwood)
for generous funding and the EPSRC Mass Spectrome-
try Unit, Swansea University, Wales for mass spectra.
References
1. (a) Bowman, W. R.; Bridge, C. F.; Brookes, P. J. Chem.