Tandem radical cyclisation and translocation approaches to biologically important mitomycin ring systems

Article abstract of DOI:10.1055/s-2002-33522

New free-radical cyclisation and translocation approaches to the tricyclic mitomycin ring system have been developed. These convergent approaches involve either a tandem 5-endo/5-exo radical cyclisation or alternatively, a 1,6-hydrogen-atom transfer follo

Full text of DOI:10.1055/s-2002-33522

LETTER
1431
Tandem Radical Cyclisation and Translocation Approaches to Biologically
Important Mitomycin Ring Systems
Tandem
Radical
C
i
yclisat
l
ion and
l
Transl
i
ocationan M. Allan, Andrew F. Parsons,* Jean-François Pons
Department of Chemistry, University of York, Heslington, York, UK YO10 5DD
Fax +44(1904)432516; E-mail: afp2@york.ac.uk
Received 17 June 2002
plication of an unusual 5-endo followed by 5-exo tandem
Abstract: New free-radical cyclisation and translocation approach-
es to the tricyclic mitomycin ring system have been developed.
These convergent approaches involve either a tandem 5-endo/5-exo
radical cyclisation or alternatively, a 1,6-hydrogen-atom transfer
followed by 5-exo cyclisation sequence.
radical cyclisation sequence (Scheme 1, Route A). Tri-
cycles of type 5 have previously been elaborated to
mitomycins2 and so an efficient and concise approach to
this ring system, from diene 6, was first investigated. The
proposed tandem cyclisation of halide 6 followed on from
related studies within our group, which has shown that ha-
loenamides can undergo sequential 5-endo/5-exo reac-
tions to form pyrrolizidinones in the presence of
triphenyl- or tributyl-tin hydride.6 This required a radical
stabilising group, namely an ester, on the second acceptor
double bond (i.e. R1 = CO2R in 6) so as to prevent a re-
versible cyclisation leading to the thermodynamically
favoured indolizidinone ring system (derived from a 5-
endo/6-endo cyclisation).
Key words: cyclisations, mitomycin, radical reactions, tandem re-
actions, tin
The mitomycins, which include mitomycin A 1, C 2 and
K 3, are a family of naturally occurring compounds which
were first isolated from Streptomyces verticillatus in the
1960’s.1 These compounds have attracted considerable
medical interest due to their pronounced antibiotic and an-
titumour activities. They are active against both Gram-
positive and -negative bacteria and also several kinds of
tumours. Although numerous synthetic approaches to the
mitomycins have been reported,2,3 the development of a
direct, versatile, diastereo- and enantio-selective approach
to these synthetically challenging compounds is still lack-
ing. As a consequence, we have investigated concise syn-
thetic approaches to the core tricyclic ring system of these
compounds using novel tandem radical cyclisation reac-
tions. These approaches were designed to afford not only
naturally occurring mitomycins but also a range of biolog-
ically important analogues including BMY-25067 4 (Fig-
ure).4
X
R
O
N
O
N
R
Route A
R1
R1
R2
5-endo/5-exo
R2
5
6
Route B
Route C
1,6-H/5-exo
5-exo
R
X
O
N
R
O
N
R1
R1
R2
H
N
H
N
R2
X
7
OMe
OMe
8
N
N
Scheme 1
OCONH2
O
O
O
O
Hence our studies began by the synthesis of a suitable un-
saturated ester (of type 6) starting from 2-nitrobenzalde-
hyde 9 (Scheme 2). Initially, a Wittig reaction followed
by selective reduction of the nitro group afforded aniline
10 in good yield.7 Subsequent condensation of 10 with a
ketone was expected to afford an intermediate imine/
enamine, which on N-acylation would yield a variety of
enamide precursors (with various R groups, Scheme 1).
Unfortunately, this reaction proved problematic and the
desired enamide was only formed in good to reasonable
yield when using acetophenone followed by chloroacetyl
chloride. This gave enamide 11, which was then reacted
with triphenyltin hydride and AIBN (added over 1 h) in
boiling toluene to initiate free-radical cyclisation. Follow-
Me
R
Me
OMe
1; R = OMe
2; R = NH2
4; R = NH(CH2)2SS-(4-NH2)-Ar
3
Figure Mitomycins 1, 2, 3 and 4
The development of novel cascade (or domino) radical re-
actions is an active area of current research5 and our initial
approach to the mitomycin ring system focused on the ap-
Synlett 2002, No. 9, Print: 02 09 2002.
Art Id.1437-2096,E;2002,0,09,1431,1434,ftx,en;D09402ST.pdf.
© Georg Thieme Verlag Stuttgart · New York
ISSN 0936-5214

1432
G. M. Allan et al.
LETTER
ing work-up and column chromatography, three lactam this case, cyclisation of the intermediate pyrrolidinone
products were isolated from the reaction. The desired tri- radical on to an unsaturated ester double bond (7,
cycle 12 was formed as a single diastereoisomer as indi- R1 = CO2R) was expected to afford the desired 5,5,6-ring
cated by the 1H NMR spectrum8 but only in 12% yield. A system. It should be noted that a competitive 6-endo cycli-
competitive 5-endo-6-endo tandem cyclisation was also sation of 7 is also possible, particularly if a radical stabi-
observed to give tricycle 13 (as a single diastereoisomer)8 lising group (e.g. 7, R = phenyl) is present at the site of
in 16% yield. The 1H NMR spectrum of 13 showed a dis- radical generation. So, to minimise 6-endo cyclisation, re-
tinctive doublet at 9.02 ppm for the aromatic H-1 hydro- action of the parent pyrrolidinone (7, R = H) was investi-
gen, which is characteristic of this type of 5,6,6-ring gated.
system.9 The fact that a 6-endo cyclisation takes place pre-
sumably reflects the stability of the intermediate pyrrolid-
The synthesis of a suitable 5-halopyrrolidinone precursor
started from 2-fluorobenzaldehyde 15 (Scheme 3). Nu-
inone radical; this radical is mesomerically stabilised by
cleophilic aromatic substitution in the presence of pyrro-
both the amide nitrogen atom and the phenyl ring. This
lidine was followed by oxidation of the pyrrolidine ring10
ensures reversible 5-exo ring-closure leading to the for-
to afford the corresponding pyrrolidinone. Olefination of
the aldehyde using a Horner–Wadsworth–Emmons type
reaction then afforded the expected E-alkene 16 in good
overall yield. However, problems arose with the introduc-
tion of a halogen atom at the 5-position of the pyrrolidino-
ne ring. Hence, for example, attempted bromination of 16
to give 17 using NBS in the presence of AIBN gave rise
mation of the thermodynamically favoured 5,6,6-ring sys-
tem. However, the major product from this reaction was
the piperidinone 14. This was derived from a competitive
6-exo cyclisation of the intermediate carbamoyl radical
directly on to the unsaturated ester double bond. Reaction
of the corresponding iodide, under the same reaction con-
ditions, gave similar yields of 12 and 13 (overall 20%),
and once again, the 6-exo product 14 was the major com-
to a number of products in low yield. This may reflect the
instability of the bromopyrrolidinone 1711 and/or compet-
pound isolated (in 36% yield).
ing reactions involving NBS and the alkene double bond
Thus, although the desired 5,5,6-ring system was formed, of 16.
the yield of the product was modest because of competing
free-radical cyclisations. In order to avoid the formation
An alternative cyclisation approach was then devised us-
ing a related vinyl halide precursor of type 8 (Scheme 1,
of a piperidinone our attention turned to the cyclisation of
Route C). In this case, the required pyrrolidinone radical
a 5-halopyrrolidinone of type 7 (Scheme 1, Route B). In
was expected to be formed from a 1,6-hydrogen atom
Cl
(i) PhCOCH3,
toluene, TsOH,
heat
NO2
(i) Ph3P=CHCO2Et,
CH2Cl2
NH2
O
N
Ph
CHO
CO2Et
75%
CO2Et
79%
(ii) Fe, EtOH, AcOH
(ii) ClCOCH2Cl,
CH2Cl2, PhNEt2
9
10
11
Ph3SnH, AIBN
Toluene, heat
O
Ph
Ph
CO2Et
O
O
N
Ph
N
N
+
+
CO2Et
CO2Et
12 (12%)
13 (16%)
14 (63%)
Scheme 2
F
(i) pyrrolidine, K2CO3, DMF,
heat (82%)
O
N
O
Br
N
CHO
CO2Et
CO2Et
(ii) Hg(OAc)2, EDTA, H2O,
heat (54%)
15
(iii) (EtO)2P(O)CH2CO2Et,
NaH, THF, 0 °C to r.t. (82%)
16
17
Scheme 3
Synlett 2002, No. 9, 1431–1434 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Tandem Radical Cyclisation and Translocation
1433
transfer reaction.12 Hence, reaction of 8 (R = H) with the benzaldehyde gave 23 in 76% yield, which was then elab-
tributyltin radical should produce reactive vinyl radical orated to vinyl bromide 24 in four steps.16 On reaction of
18, which could undergo a rearrangement reaction so as to 24 with tributyltin hydride (added over 2 h) the 5,5,6-tri-
form the more stable pyrrolidinone radical 19 (Scheme 4). cycle 25 was isolated as the major product in 50% yield
(as a 7.3:1 mixture of isomers) together with the 6-endo
product 26 (as a single isomer) in 20% yield. Interesting-
ly, no products derived from simple reduction or 1,6-hy-
drogen atom transfer from the methoxy group were
isolated.
1,6-hydrogen
atom transfer
H
O
O
N
N
R1
R1
R2
R2
This work has shown that the 5,5,6-ring system present in
mitomycins can be prepared, for the first time, via tandem
radical cyclisation sequences. The mild cyclisation condi-
tions and convergent approach offers a potentially flexible
approach to these types of biologically important com-
pounds. The ability to control the ratio of 5-exo/6-endo
radical cyclisation pathways by appropriate substitution
of the precursor is of particular mechanistic interest as is
the novel 1,6-hydrogen atom transfer reaction. This is
shown to afford an elegant approach to an intermediate
pyrrolidinone radical, which proved impossible to access
from a classical halogen-atom transfer route because of
the difficulty in preparing the requisite 5-halopyrrolidi-
none precursor.
18
Scheme 4
19
The required Z-vinyl bromide precursor 20 was prepared
in a similar manner to pyrrolidinone 16, but using the bro-
mophosphonate (EtO)2P(O)CHBrCO2Et13 in the Horner–
Wadsworth–Emmons step (Scheme 5). Following reac-
tion of 20 with tributyltin hydride and AIBN (added over
2 h), the desired 5,5,6-tricycle 21 was isolated in 64%
yield as a 1.4:1 mixture of diastereoisomers. It was pleas-
ing to see that the 6-endo product 22 was formed in only
6% yield (as a single diastereoisomer) and that no simple
reduced product was isolated. Whereas the 1H NMR spec-
tra for the two diastereoisomers of 21 showed signals at
4.81–4.75 and 4.34–4.30 ppm for the NCH hydrogen,14
the corresponding NCH hydrogen in 22 was shifted to
4.07–3.97 ppm.15
Acknowledgement
We thank the Leverhulme Trust for funding.
A similar cyclisation reaction was observed when using
the dimethoxyaryl analogue 24 (Scheme 6). Nitration
(HNO3/SiO2) of commercially available 2,5-dimethoxy-
CO2Et
O
O
O
N
N
N
Bu3SnH, AIBN
Toluene, heat
CO2Et
CO2Et
+
Br
20
21 (64%; d.r. = 1.4:1)
22 (6%)
Scheme 5
NO2
NH2
(i) (EtO)2P(O)CHBrCO2Et,
NaH, 0 °C to r.t. (63%)
CO2Et
Br
MeO
CHO
OMe
MeO
(ii) Fe, EtOH, AcOH,
heat, (58%)
OMe
23
(i) ClCH2CH2COCl,
Na2HPO4, CHCl3
(ii) NaOEt, EtOH, heat
O
N
CO2Et
O
O
N
MeO
N
Bu3SnH, AIBN
Toluene, heat
CO2Et
OMe
CO2Et
Br
+
MeO
MeO
OMe
26 (20%)
OMe
24
30%
25 (44% + 6%)
Scheme 6
Synlett 2002, No. 9, 1431–1434 ISSN 0936-5214 © Thieme Stuttgart · New York

1434
G. M. Allan et al.
LETTER
(10) Moehrle, H.; Mehrens, J. Z. Naturforsch., B: Chem. Soc.
References
1998, 53, 1369.
(1) (a) Webb, J. S.; Cosulich, D. B.; Mowat, J. H.; Patrick, J. B.;
Broschard, R. W.; Meyer, W. E.; Williams, R. P.; Wolf, C.
F.; Fulmor, W.; Pidacks, C.; Lancaster, J. E. J. Am. Chem.
Soc. 1962, 84, 3185. (b) Webb, J. S.; Cosulich, D. B.;
Mowat, J. H.; Patrick, J. B.; Broschard, R. W.; Meyer, W. E.;
Williams, R. P.; Wolf, C. F.; Fulmor, W.; Pidacks, C.;
Lancaster, J. E. J. Am. Chem. Soc. 1962, 84, 3187.
(2) (a) Nakatsubo, K.; Fukuyama, T.; Cocuzza, A. J.; Kishi, Y.
J. Am. Chem. Soc. 1977, 99, 8115. (b) Fukuyama, T.;
Nakatsubo, K.; Cocuzza, A. J.; Kishi, Y. Tetrahedron Lett.
1977, 18, 4295. (c) Kishi, Y. J. Nat. Prod. 1979, 42, 549.
(d) Fukuyama, T.; Yang, L. J. Am. Chem. Soc. 1987, 109,
7881. (e) Fukuyama, T.; Yang, L. J. Am. Chem. Soc. 1989,
111, 8303. (f) Benbow, J. W.; Schulte, G. K.; Danishefsky,
S. J. Angew. Chem., Int. Ed. Engl. 1992, 31, 915.
(g) Benbow, J. W.; McClure, K. F.; Danishefsky, S. J. J. Am.
Chem. Soc. 1993, 115, 12305. (h) Kasai, M.; Kono, M.
Synlett 1992, 778.
(3) For recent approaches see for example: (a) Coleman, R. S.;
Chen, W. Org. Lett. 2001, 3, 1141. (b) Jones, G. B.; Guzel,
M.; Mathews, J. E. Tetrahedron Lett. 2000, 41, 1123.
(c) Ziegler, F. E.; Berlin, M. Y. Tetrahedron Lett. 1998, 39,
2455. (d) Ziegler, F. E.; Berlin, M. Y.; Lee, K.; Looker, A.
R. Org. Lett. 2000, 2, 3619. (e) Dobbs, A. P.; Jones, K.;
Veal, K. T. Tetrahedron 1998, 54, 2149. (f) Brunton, S. A.;
Jones, K. J. Chem. Soc., Perkin Trans. 1 2000, 763.
(g) Jones, K.; Storey, J. M. D. J. Chem. Soc., Perkin Trans.
1 2000, 769.
(4) (a) He, Q.-Y.; Maryendu, H.; Tomasz, M. J. Am. Chem. Soc.
1994, 116, 9349. (b) Kohn, H.; Wang, S. Tetrahedron Lett.
1996, 37, 2337. (c) Edstrom, E. D.; Yu, T. Tetrahedron
1997, 53, 4549.
(5) McCarroll, A. J.; Walton, J. C. J. Chem. Soc., Perkin Trans.
1 2001, 3215.
(6) (a) Baker, S. R.; Parsons, A. F.; Pons, J.-F.; Wilson, M.
Tetrahedron Lett. 1998, 39, 7197. (b) Baker, S. R.; Burton,
K. I.; Parsons, A. F.; Pons, J.-F.; Wilson, M. J. Chem. Soc.,
Perkin Trans. 1 1999, 427.
(7) All new compounds exhibited satisfactory spectral and
analytical (high-resolution mass) data.
(11) Easton, C. J.; Pitt, M. J.; Ward, C. M. Tetrahedron 1995, 51,
12781.
(12) (a) Robertson, J.; Pillai, J.; Lush, R. K. Chem. Soc. Rev.
2001, 30, 94. (b) Bogen, S.; Fensterbank, L.; Malacria, M. J.
Org. Chem. 1999, 64, 819. (c) Wessig, P.; Schwarz, J.;
Lindermann, U.; Holthausen, M. C. Synthesis 2001, 1258.
(d) Leardini, R.; McNab, H.; Minozzi, M.; Nanni, D.; Reed,
D.; Wright, A. G. J. Chem. Soc., Perkin Trans. 1 2001, 2704.
(13) Wadsworth, W. S.; Emmons, W. D. J. Am. Chem. Soc. 1961,
83, 1733.
(14) For a related 5,5,6-tricycle see: Wee, A. G. H.; Liu, B.;
Zhang, L. J. Org. Chem. 1992, 57, 4404.
(15) 5,5,6-Tricycle 21. Diastereoisomer 1: 1H NMR (500 MHz,
CDCl3): = 7.63 (1 H, d, J = 7.8 Hz, H-1 aromatic), 7.27–
7.18 (2 H, m, aromatic), 7.05–7.01 (1 H, m, aromatic), 4.81–
4.75 (1 H, m, NCH), 4.17 (2 H, q, J = 7.1 Hz, CO2CH2),
3.74–3.69 (1 H, m, CHCH2CO2), 2.90–2.82 (1 H, m,
NCOCH), 2.63–2.57 (2 H, m, CHCO2 and NCOCH), 2.43 (1
H, dd, J = 16.8 and 5.7 Hz, CHCO2), 2.19–2.14 (1 H, m,
NCOCH2CH), 2.00–1.91 (1 H, m, NCOCH2CH) and 1.26
(3 H, t, J = 7 Hz, CO2CH2CH3). 13C NMR (75 MHz, CDCl3):
= 171.6, 170.7 (NCO and CO2), 138.0, 137.0 (2 C=CH
aromatic), 128.4, 125.2, 124.3, 114.5 (4 C=CH aromatic),
65.0 (NCH), 60.9 (CO2CH2), 38.3 (CHCH2CO2), 36.4, 36.2
(CHCH2CO2 and NCOCH2), 23.3 (CH2CH2CH2), 14.2
(CO2CH2CH3). MS (CI, NH3): m/z (%) = 260 (100) [M +
H+]. Found (CI, NH3): [M + H+] 260.1290. C15H17NO3
requires for [M + H+] 260.1287. Diastereoisomer 2: 1H NMR
(500 MHz, CDCl3): = 7.61 (1 H, d, J = 7.8 Hz, H-1
aromatic), 7.41–7.04 (3 H, m, aromatic), 4.34–4.30 (1 H, m,
NCH), 4.25–4.16 (2 H, m, CO2CH2), 3.60–3.55 (1 H, m,
CHCH2CO2), 3.03 (1 H, dd, J = 16.4 and 4.4 Hz, CHCO2),
2.86–2.78 (1 H, m, NCOCH), 2.61–2.51 (3 H, m, CHCO2,
NCOCH and NCOCH2CH), 2.15–2.06 (1 H, m,
NCOCH2CH) and 1.30 (3 H, t, J = 7 Hz, CO2CH2CH3). 13
C
NMR (75 MHz, CDCl3): = 171.8, 171.6 (NCO and CO2),
139.1, 136.3, 128.4, 124.4, 123.9, 115.0 (C=CH aromatic),
69.8 (NCH), 60.9 (CO2CH2), 44.8 (CHCH2CO2), 37.8, 36.1
(CHCH2CO2 and NCOCH2), 29.2 (CH2CH2CH2), 14.4
(CO2CH2CH3). MS (CI, NH3): m/z (%) = 260 (100) [M +
H+]. Found (CI, NH3): [M + H+] 260.1288. C15H17NO3
requires for [M + H+] 260.1287. 5,6,6-Tricycle 22. 1H NMR
(270 MHz, CDCl3): = 8.70 (1 H, d, J = 9.1 Hz, H-1
aromatic), 7.27–7.04 (3 H, m, aromatic), 4.26 (2 H, q, J = 7.2
Hz, CO2CH2), 4.07–3.97 (1 H, m, NCH), 3.14–3.09 (2 H, m,
CHCHCO2 and CH2CHCO2), 2.74–2.33 (4 H, m, NCOCH2,
CHCHCO2 and NCOCH2CH), 1.93–1.77 (1 H, m,
NCOCH2CH) and 1.32 (3 H, t, J = 7.2 Hz, CO2CH2CH3).
13C NMR (75 MHz, CDCl3): = 173.6, 172.6 (NCO and
CO2), 136.0, 128.9, 127.4, 124.0, 123.9, 119.1 (C=CH
aromatic), 61.2 (CO2CH2), 58.9 (NCH), 45.4 (CHCO2),
31.9, 31.5 (NCOCH2 and CH2CHCO2), 24.0 (CH2CH2CH2),
14.3 (CH2CH3). Found (CI, NH3): [M + H+] 260.1285.
C15H17NO3 requires for [M + H+] 260.1287.
(8) Tricycle 12: 1H NMR (270 MHz, CDCl3): = 7.80 (1 H, d,
J = 8 Hz, H-1 aromatic), 7.53–6.97 (8 H, m, aromatic), 4.26
(2 H, q, J = 7 Hz, CO2CH2), 3.78 (1 H, dd, J = 11 and 4.5 Hz,
CHCH2CO2), 2.75 (1 H, dd, J = 17.5 and 11 Hz, CHCO2),
2.62–2.15 (5 H, m, CHCO2, NCOCH2 and NCOCH2CH2)
and 1.32 (3 H, t, J = 7 Hz, CO2CH2CH3). MS (CI, NH3):
m/z (%) = 336 (100) [M + H+]. Found (CI, NH3): 336.1598
[M + H+]. C21H21NO3 requires for [M + H+], 336.1600.
Tricycle 13: 1H NMR (270 MHz, CDCl3): = 9.02 (1 H, d,
J = 8.2 Hz, H-1 aromatic), 7.46–6.99 (8 H, m, aromatic),
4.36–4.26 (2 H, m, CO2CH2), 3.03 (1 H, dd, J = 13.3 and 4.1
Hz, CHCHCO2), 2.89–2.30 (6 H, m, CHCHCO2, NCOCH2
and NCOCH2CH2) and 1.36 (3 H, t, J = 7 Hz, CO2CH2CH3).
MS (CI, NH3): m/z (%) = 336 (100) [M + H+]. Found (CI,
NH3): [M + H+] 336.1599. C21H21NO3 requires for [M + H+]
336.1600.
(16) The low yield (30%) for the N-acylation/cyclisation
reactions (to form the pyrrolidinone ring) was due to the
formation of an alkyne in 55% yield, which resulted from
dehydrobromination of vinyl bromide 24 by ethoxide.
(9) (a) Kocián, O.; Ferles, M. Collect. Czech. Chem. Commun.
1978, 43, 1413. (b) Speckamp, W. N.; de Boer, J. J. J. Recl.
Trav. Chim. Pays-Bas 1983, 102, 410.
Synlett 2002, No. 9, 1431–1434 ISSN 0936-5214 © Thieme Stuttgart · New York