Tandem cyclisations of amidyl radicals derived from O-acyl hydroxamic acid derivatives

Article abstract of DOI:10.1055/s-1999-2641

Amidyl radicals generated from tributylstannane mediated homolysis of O- acyl hydroxamic acid derivatives 5a-b undergo tandem cyclisations to give pyrrolizidinones 6a-c and indolizidinones 7a-b respectively while 5c-d undergo monocyclisation to give β-lac

Full text of DOI:10.1055/s-1999-2641

LETTER
441
Tandem Cyclisations of Amidyl Radicals Derived from O-Acyl Hydroxamic
Acid Derivatives
Andrew J. Clarka*, Robert P. Filika, Joanne L. Peacocka, Gerard H.Thomasb
aDepartment of Chemistry, University of Warwick, Coventry, CV4 7AL, UK
Fax +44 (1203) 524112, E-mail: msrir@csv.warwick.ac.uk
bKnoll Pharmaceuticals, Research and Development Department, Pennyfoot Street, Nottingham, NG1 1GF, UK
Received 13 January 1999
Ph
O
O
Abstract: Amidyl radicals generated from tributylstannane mediat-
ed homolysis of O-acyl hydroxamic acid derivatives 5a-b undergo
tandem cyclisations to give pyrrolizidinones 6a-c and indolizidin-
ones 7a-b respectively while 5c-d undergo monocyclisation to give
b-lactam 10a and g-lactam 11a respectively. On the other hand the
reaction of 5d-f with Cu(OTf)2/DBN furnishes mixtures of reduc-
tion 10b, monocyclisation 10a and tandem cyclisation 9 products
with the ratio dependant upon the nature of the O-acyl group and the
solvent and concentration employed.
N
N
OBz
OBz
5a
5b
O
O
Ph
Ph
N
N
OCOR
OBz
Key words: amidyl radicals, cyclisation, copper complexes, hy-
droxamic acid derivatives, tandem reactions
5d (R = Me)
5e (R = Ph)
5f (R = OMe)
5c
Biologically active nitrogen heterocycles occur widely in
nature. In particular the bicyclic and tricyclic systems
shown below 1-4 are prevalent in a number of naturally
occurring compounds ranging from thienamycin1, the
pyrrolizidine2, indolizidine3 and the lycorane4 family of
alkaloids respectively. We have recently published the use
of O-benzoyl N-alkyl hydroxamic acid derivatives as pre-
cursors for amidyl radicals5 which can undergo 4-exo
cyclisation6, or 5-exo cyclisation7-8 reactions to furnish b-
lactams or pyrrolidinones respectively. We wish to report
in this letter attempts to extend this methodology to fur-
nish bicyclic and tricyclic frameworks related to 1-4.
Ph
O
Ph
Ph
O
H
H
H
(i)
5a
Me
Me
Me
N
N
N
O
6b
6c
6a
6a : 6b : 6c = 3 : 2 :1
H
Me
Me
Me
(i)
5b
N
N
Me
N
Me
O
O
7c
O
7a
7b
7a : 7b : 7c = 5 : 2.5 : 1
(47%)
Scheme 1 (i) syringe pump 6 hrs Bu3SnH, AIBN,1:1 v:v C6H12:MeC6H5
N
N
N
N
O
2
3
1
4
1.4%
2.3%
3.5%
3.3%
Ph
Ph
H
Ph
H
H
N
H
N
H
N
The use of O-benzoyl hydroxamic acid derivatives as pre-
cursors to amidyl radicals was first reported by Zard9. In
order to determine the scope of this methodology for the
synthesis of more functionalised molecules we investigat-
ed the Bu3SnH mediated reactions of the precursors 5a-f.
These were prepared using our previously described
methodology.6-8, 10
1.1%
H
Me
Me
Me 1.7%
H
H
O
O
O
1.4%
Fig 1
7.4%
Fig 2
Fig 3
Reaction of both 5a-b with 1.2 eq of Bu3SnH/AIBN deliv-
ered by a syringe pump over 6 hours at reflux (initial con-
centration of substrate 0.2 mmol/ml) led to mixtures of the
desired tandem cyclised products 6a-c and 7a-b respec-
tively (Scheme 1). Cyclisation of 5a furnished three dias-
tereomeric pyrrolizidinones 6a-c in a ratio of 3 : 2 : 1
respectively11. No reduction or monocyclisation products
were detected. It proved difficult to remove all the last
traces of tin residues from the products to obtain pure
samples (after repetitive chromatography 6a was isolated
in 17% yield). The stereochemistry of each of the diaste-
reomers was tentatively determined by nOe NMR experi-
ments. The diastereoselectivity of the first cyclisation
Synlett 1999, No. 4, 441–443 ISSN 0936-5214 © Thieme Stuttgart · New York

442
A. J. Clark et al.
LETTER
(6a+6c:6b = 2:1) is similar to that reported for recent DBN in either acetone, acetonitrile or THF was heated at
amidyl radical cyclisations7 while that of the second 120°C in a sealed tube for 48 hours varying ratios of 11a-
would be expected to be similar to a 2-substituted hexenyl b and 9 were produced (Scheme 3, Table 1).
radical cyclisation. The Beckwith model12 therefore pre-
dicts the observed formation of the all cis isomer 6a as the
major product.
see
table
H
Ph
N
Cyclisation of 5b under identical conditions produced a
mixture of the indolizidinone 7a and pyrrolizidinone 7b in
a 2:1 ratio (42%). In addition a small amount of the mono-
cyclisation product 7c (5%) was also isolated. The pres-
ence of the monocyclisation product is in line with a
slower second cyclisation relative to 5a caused by the
steric hindrance now presented by the methyl alkene
substituent13.
5d-f
N
Ph
N
O
O
O
11b
11a
9
Scheme 3
The results clearly indicate that the nature of the O-acyl
substituent is crucial for successful cyclisation reactions
with 5d-e giving a mixture of all three products (entries 1-
2) while 5f gave the reduction product 11b only (entry 3).
The nature of the solvent was also important in controlling
the ratio of products (entries 4-5) but by far the most cru-
cial factor in optimising the yield of the tandem product 9
was found to be concentration with best results at high di-
lution (entry 8).
Attempts to facilitate a tandem cyclisation of 5c to give
the bicyclo [2,3,0] system 8 found in thienamycin under
identical conditions failed and the only cyclised product
isolated was the monocyclised b-lactam 10a. In addition
to 10a an equal amount of the reduction product 10b and
the rearranged product 10c were isolated in a 50% overall
yield. Attempts to prepare the tetrahydro-2H-pyrrolo-iso-
quinolinone skeleton 9 from 5d via a tandem 5-exo amidyl
radical cyclisation followed by a 6-exo carbon radical ad-
dition into the aromatic ring of the N-benzyl substituent
also failed and only reduced 11b and monocyclised 11a
products were isolated in 58% combined yield.
Table 1
Ratiob
(11a:11b:9)
Conc
Yielda
Entry
5
Solvent
0.19M
0.19M
0.19M
0.19M
0.19M
0.05M
0.02M
0.01M
84% (17%)
91% (7%)
77% (0%)
85% (30%)
85% (23%)
81% (28%)
93% (38%)
93% (53%)
1
2
3
4
5
6
7
8
5d
5e
5f
MeCN
MeCN
MeCN
THF
4 : 76 : 20
6 : 86 : 8
N
N
0 : 100 : 0
18 : 44 : 38
15 : 58 : 27
3 : 62 : 35
3 : 56 : 41
0 : 43 : 57
O
O
8
9
5d
5d
5d
5d
5d
acetone
MeCN
MeCN
MeCN
O
Ph
(i)
Ph
10b R = H
10c R = OBz
5c
N
N
H
O
R
10a
10a : 10b : 10c = 1 : 1 : 1
a Combined yield, with yield of 9 in brackets,
b Determined by 250 MHz 1H NMR
(50%)
O
O
(i)
N
5d
Cu(OTf)2
Ph
N
Ph
5d
H
N
DBN, MeCN
120ºC, 0.01M
N
11b
11a
12
O
O
11a : 11b = 1 : 1
(58%)
CuI
CuII
Scheme 2 (i) syringe pump 6 hrs Bu3SnH, AIBN,1:1 v:v C6H12:MeC6H5
- H +
10
N
N
Hypothesising that the failure of the latter tandem reaction
was due to the reversibility of the second cyclisation into
the aromatic ring (12-13) we investigated an alternative
approach where the intermediate cyclohexadienyl radical
13 would be trapped irreversibly as a carbocation using
added copper (II) salts14, (Scheme 4). Successful condi-
tions to facilitate the desired transformations involved us-
ing modified conditions to those reported for asymmetric
Kharash reactions15. Hence when 5d-f and Cu(OTf)2/
O
14
13
O
Scheme 4
The use of copper salts in oxidising radicals to cations is
well established as is their use in catalysing the decompo-
sition of the weak O-O bond of diacyl peroxides to give
acyloxy radicals16. Due to the similarity in structure of N-
Synlett 1999, No. 4, 441–443 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Tandem Cyclisations of Amidyl Radicals Derived from O-Acyl Hydroxamic Acid Derivatives
443
(5) For a review on nitrogen centered radicals see Esker, J.L.;
Newcomb, M.; Advances in Heterocyclic Chemistry. 1993,
58, 1
(6) Clark, A J.; Peacock, J.L.; Tetrahedron Lett. 1998, 39,1265
(7) Clark, A J.; Peacock, J.L.; Tetrahedron Lett. 1998, 39, 6029.
(8) Clark, A.J.; Deeth, R.J.; Samuel, C.J.; Wongtap, H.; Synlett
1999, 444.
(9) Boivin, J.; Callier-Dublanchet, A-C.; Quiclet-Sine, B.; Schia-
no, A.M.; Zard, S.Z.; Tetrahedron 1995, 51, 6517.
(10) Al-Faiyz, Y.S.S.; Clark, A J.; Filik, R.P.; Peacock, J.L.;
Thomas, G.H.; Tetrahedron Lett. 1998, 39,1269.
(11) Ratio determined from crude 1H 250MHz NMR (CDCl3).
(12) Beckwith, A.L.J.; Schiesser, C.H. Tetrahedron 1985, 45,
3925.
alkyl-O-acyl hydroxamic acid derivatives 5d-f to diacyl
peroxides a possible mechanistic rationale for the reaction
sequence involves homolytic cleavage of the weak N-O
bond in 5d-f to furnish the desired amidyl radicals which
then undergo tandem cyclisation to 13 followed by oxida-
tion to 14 followed by elimination of a proton, (Scheme
4).
References and Notes
(1) Albers-Schonberg, G.; Arison, B.H.; Hensen, O.D.;
Horshfeld, J.; Hoogsten, J.; Kaczka, E.A.; Rhodes, R.E.; Ka-
han, J.H.; Kahan, F.M.; Ratcliffe, R.W.; Walton, E.; Ruswink-
le, L.J.; Morin, R.B.; Christensen, B.G.; J. Am. Chem. Soc.
1978, 100, 6491.
(2) Michael. J.P. Nat. Prod. Rep. 1997, 14, 619.
(3) Lidell, J. R. Nat. Prod. Rep. 1996, 13, 187.
(4) Martin, S. F. In “The Alkaloids”, Brossi, A.; Ed. Academic
Press: New York, Vol. 30, 1987, 251-376.
(13) Keusenkothen, P.F.; Smith, M.B.; Tetrahedron 1992, 15,
2977
(14) Kochi, J.K.; Acc. Chem. Res., 1974, 351.
(15) Sekar, G.; DattaGupta, A.; Singh, V.K.; Tetrahedron Lett.
1996, 8435.
(16) Minisci, F.; Acc. Chem Res. 1975, 165.
Synlett 1999, No. 4, 441–443 ISSN 0936-5214 © Thieme Stuttgart · New York