Concise route to α-acylamino-β-keto amides: Application to the synthesis of a simplified azinomycin A analogue

Article abstract of DOI:10.1016/S0040-4039(02)02420-6

Condensation of acid chlorides (alkyl, aryl or heteroaryl) with N,N′-dialkyl α-acylamino malonamides in the presence of magnesium ethoxide provides a direct route to α-acylamino-β-keto amides in moderate to good yields (46-95%). Using this method, a conci

Full text of DOI:10.1016/S0040-4039(02)02420-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 9573–9576
Concise route to a-acylamino-b-keto amides: application to the
synthesis of a simplified azinomycin A analogue
Jean-Yves Goujon and Michael Shipman*
School of Chemistry, University of Exeter, Stocker Road, Exeter, EX4 4QD, UK
Received 6 September 2002; revised 12 October 2002; accepted 25 October 2002
Abstract—Condensation of acid chlorides (alkyl, aryl or heteroaryl) with N,N%-dialkyl a-acylamino malonamides in the presence
of magnesium ethoxide provides a direct route to a-acylamino-b-keto amides in moderate to good yields (46–95%). Using this
method, a concise route to an enantiomerically enriched 1-azabicyclo[3.1.0]hexane containing most of the elements of the
‘right-hand’ domain of azinomycin A has been developed. © 2002 Elsevier Science Ltd. All rights reserved.
Azinomycins A 1 and B 2 are naturally occurring
antibiotics which possess potent in vitro cytotoxic activ-
ity, significant in vivo antitumour activity and which
appear to act by disruption of cellular DNA replication
by interstrand cross-link (ISC) formation (Fig. 1).1 The
epoxide and aziridine are known to be responsible for
the cross-linking process which occurs between bases
two residues apart on the complementary DNA
strands, with specificity for 5%–PuPyPy–3% sequences.2
The azinomycins possess a highly unusual molecular
architecture which, combined with their interesting bio-
logical properties, has stimulated considerable interest
from the scientific community.1 The synthetic problems
presented by these natural products are formidable,
because in addition to the density of functional groups,
the azinomycins are unstable in aqueous solution,3 and
prone to opening by nucleophiles at C-10 and C-21.4 In
2001, Coleman overcame these synthetic hurdles and
reported the first total synthesis of azinomycin A.5
ucts.6 To further progress our synthetic efforts towards
azinomycin A itself, and to quantify the role of the
amide side-chain in DNA binding, we required struc-
tures incorporating the azinomycin A side-chain (Cꢀ1
to Nꢀ5). Unfortunately, all our attempts to selectively
hydrolyse the ethyl ester of 5 to facilitate the introduc-
tion of this substituent were entirely unsuccessful. To
circumvent this problem, we sought a way to introduce
Figure 1. Structures of the azinomycins.
As part of our own synthetic efforts, we have developed
a novel reductive cyclisation to construct the dehy-
droamino acid fragment.6 This chemistry involved
homologation of acid chloride 3 to a-amino-b-keto
ester 4 and subsequent catalytic hydrogenation in the
presence of hydrochloric acid to give pyrrolidine 5
(Scheme 1). Furthermore, we were able to convert 5
into epoxy aziridine 6 and demonstrate that it cross-
links DNA in a manner similar to the natural prod-
Keywords: antitumour compounds; azinomycins; aziridines;
carzinophilin; coupling reactions.
* Corresponding author. Tel.: +44-1392-263469; fax: +44-1392-
263434; e-mail: m.shipman@exeter.ac.uk
Scheme 1. Reductive cyclisation approach to azinomycin ana-
logue 6.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02420-6

9574
J.-Y. Goujon, M. Shipman / Tetrahedron Letters 43 (2002) 9573–9576
the amide side-chain prior to implementing the reduc-
tive cyclisation. In this Letter, we describe a new direct
method for the synthesis of a-acylamino-b-keto amides
and demonstrate that it can be used to make a 1-azabi-
cyclo[3.1.0]hexane containing the amide side-chain of
azinomycin A.
Scheme 3. Synthesis of a-acylamino-b-keto amides 9a–h.
stituent (Table 1, entry 1 cf. entry 4). The reactions
were typically performed using a two-fold excess of the
malonamide although it seems that reasonable yields
can be obtained using just 1.1 equivalents of this cou-
pling partner (Table 1, entry 1 cf. entry 9).
Whilst methods for the formation of a-acylamino-b-
keto esters by acylation of EtO2CCH(NHCOR)CO2H
with acid chlorides are known,7 we are unaware of any
published methods for the preparation of a-acylamino-
b-keto amides via the same disconnection. Somewhat
serendipitously, we have discovered that this transfor-
mation can be smoothly accomplished by reacting
N,N%-dialkyl a-acylamino-malonamides with acid chlo-
rides in the presence of magnesium ethoxide.
When benzoyl chloride was reacted with 8a (2 equiv.) in
the presence of excess magnesium ethoxide, in addition
to 9a (92%), PrHNCO2Et (80%) was isolated along
with unreacted 8a (90% of theoretical assuming 100%
conversion). This experiment provides us with a basic
working hypothesis as to the reaction pathway. We
suggest that malonamide 8 is deprotonated by the base
then is C-acylated with the acid chloride. Further reac-
Five representative dialkyl malonamides were used.
Malonamides 8a–d were made directly by treating
amino malonate 7 (R=Boc or Cbz) with an excess of
the amine in refluxing xylene (Scheme 2). Further dou-
ble oxidation of 8d under Swern conditions provided 8e
containing the amide side-chain required for azino-
mycin A.
tion with ethoxide then provides
PrHNCO2Et by subsequent C–C bond fission.
9 along with
We have successfully used this methodology to make
1-azabicyclo[3.1.0]hexane 10 (Scheme 4). Thus, treat-
ment of malonamide 8e with acid chloride 3, made
from 11, provided 12 in 55% yield. Hydrogenation of
Treatment of dialkyl malonamides 8a–e (1.0 equiv.)
with a variety of acid chlorides (2.0 equiv.) in diethyl
ether using magnesium ethoxide (4.0 equiv.) as base
directly provided the corresponding a-acylamino-b-keto
amides 9a-h in moderate to good yields (Table 1).8,9
The reaction works with alkyl, aromatic and heteroaro-
matic acid chlorides (Table 1, entries 1–3) and with
either Cbz and Boc protection of the a-amino sub-
Scheme 2. Preparation of dialkyl malonamides 8a–e.
Reagents and conditions: (i) R1NH2 (8 equiv.), xylene, 140°C,
2 h, 8a (88%), 8b (92%), 8c (89%), 8d (87%); (ii) (COCl)2 (2.2
equiv.), DMSO (4.4 equiv.), Et3N (10 equiv.), CH2Cl2, −78°C,
82%.
Scheme 4. Synthesis of 1-azabicyclo[3.1.0]hexane 10. Reagents
and conditions: (i) (COCl)2, cat. DMF, CH2Cl2; (ii) 8e,
Mg(OEt)2, Et2O, 55% (from 11); (iii) H2, Pd/C, HCl, EtOH,
80%; (vi) MsCl, CH2Cl2, 0°C, 88%; (v) KHMDS, THF,
0°C“rt, 61%.
Table 1. a-acylamino-b-keto amides 9a–h produced via Scheme 3
Entry
Amide
R
R1
R2
Producta (yield%)
1
2
3
4
5
6
7
8
9
8a
8a
8a
8b
8b
8c
8d
8e
8a
Boc
Boc
Boc
Cbz
Cbz
Boc
Boc
Boc
Boc
nPr
Ph
9a (91)
9b (85)
9c (46)
9d (90)
9e (89)
9f (95)
9g (73)
9h (89)
9a (68)
nPr
2-Thiophenyl
nPr
nPr
nPr
Ph
nPr
2-Thiophenyl
Ph
Ph
2-Thiophenyl
Ph
CH2Ph
CH2CH(OH)Me
CH2COMe
nPr
a All reactions were performed using 2 equiv. of amide 8 with respect to the acid chloride except entry 9 where 1.1 equiv. was used.

J.-Y. Goujon, M. Shipman / Tetrahedron Letters 43 (2002) 9573–9576
9575
13 according to our previously developed conditions,6
Acknowledgements
facilitated cyclisation to pyrrolidine 13 in 80% yield.
This pyrrolidine was produced as essentially a single
stereoisomer which was tentatively assigned as possess-
ing the (E)-stereochemistry on the basis of NOE exper-
iments.10 Chiral HPLC analysis indicated that 13 had
been produced in ca. 90% ee,11 revealing that a small
amount of racemisation had occurred in the coupling/
cyclisation sequence. Comparable levels of racemisation
had been noted in our earlier synthesis of ester 5
(Scheme 1).6 Mesylation of alcohol 13 provided 14
which was ring closed using potassium hexamethyldi-
silazide to 10.12 Aziridine 10 was isolated in a reason-
able state of purity (]90% as judged by 13C NMR
spectroscopy) after rapid aqueous work up and precipi-
tation of the impurities using hexane/CH2Cl2 (10:1).
The authors gratefully acknowledge the financial sup-
port provided by EPSRC under GR/M05461. We are
indebted to the EPSRC National Mass Spectrometry
Centre for performing mass spectral measurements, and
the EPSRC Chemical Database Service at Daresbury.14
We thank Professor Robert S. Coleman for providing
us with details of his recent biochemical studies on the
azinomycins prior to publication.
References
1. For a comprehensive review, see Hodgkinson, T. J.;
Shipman, M. Tetrahedron 2001, 57, 4467–4488.
Gratifyingly, it possessed the same (E)-geometry about
the tetrasubstituted double bond as the natural prod-
ucts themselves. This conclusion was reached on the
basis of several pieces of spectroscopic data. Firstly,
irradiation of the amide hydrogen (H-5) produced small
but measurable NOE enhancements of Hꢀ3 (6.2%),
Hꢀ10exo (1.2%) and Hꢀ11 (0.9%); whilst simultaneous
irradiation of Hꢀ13 and Hꢀ13% produced NOE enhance-
ments of Hꢀ10endo (3.3%) and Hꢀ16 (1.4%). Secondly, a
significant downfield shift (l 10.3) of the amide hydro-
gen (Hꢀ5) was observed. A similar chemical shift was
seen for this hydrogen in azinomycin A itself (l 10.09).
Yokoi et al. rationalised this observation by invoking a
hydrogen bond to the nitrogen atom of the aziridine
(N–9) which necessitates the (E)-geometry of the dou-
ble bond.13 Further evidence in support of this assign-
ment came from the fact that the 13C NMR spectrum
of 10 closely agrees with that of azinomycin A (Table
2).
2. For mode of action studies, see: (a) Armstrong, R. W.;
Salvati, M. E.; Nguyen, M. J. Am. Chem. Soc. 1992, 114,
3144–3145; (b) Fujiwara, T.; Saito, I.; Sugiyama, H.
Tetrahedron Lett. 1999, 40, 315–318; (c) Zang, H.; Gates,
K. S. Biochemistry 2000, 39, 14968–14975; (d) Coleman,
R. S.; Perez, R. J.; Burk, C. H.; Navarro, A. J. Am.
Chem. Soc. 2002, 124, 13008–13017.
3. (a) Kamada, H.; Wakaki, S.; Fujimoto, Y.; Tomioka, K.;
Ueyama, S.; Marumo, H.; Uzu, K. J. Antibiot., Ser. A
1955, 8, 187–188; (b) Salvati, M. E.; Moran, E. J.;
Armstrong, R. W. Tetrahedron Lett. 1992, 33, 3711–3714.
4. For example, 4-O-methyl-13-desacetyl-12,13-di-O-benzyl
azinomycin B readily reacts with thiophenol at both the
epoxide and the aziridine, see Hashimoto, M.; Terashima,
S. Heterocycles 1998, 47, 59–64.
5. Coleman, R. S.; Li, J.; Navarro, A. Angew. Chem Int. Ed.
2001, 40, 1736–1739.
6. Hartley, J. A.; Hazrati, A.; Kelland, L. R.; Khanim, R.;
Shipman, M.; Suzenet, F.; Walker, L. F. Angew. Chem.,
Int. Ed. Engl. 2000, 39, 3467–3470.
In summary, we have devised a new simple method for
the synthesis of a-acylamino b-keto amides which we
have demonstrated is of utility in the synthesis of
simplified azinomycin A analogues possessing the cor-
rect (E)-geometry about the tetrasubstitued double
bond. Work to assemble more complex azinomycin
analogues and to evaluate their therapeutic potential is
ongoing in our laboratory.
7. For example, see: (a) Schmidt, U.; Griesser, H.;
Lieberknecht, A.; Schmidt, J.; Gra¨ther, T. Synthesis,
1993, 765–766; (b) Krysan, D. J. Tetrahedron Lett. 1996,
37, 3303–3306.
8. Typical procedure: To magnesium (0.16 g, 6.58 mmol)
was added CCl4 (0.15 ml) then ethanol (0.8 ml) dropwise
at room temperature. The reaction was stirred for 20 min,
then diethyl ether (4 ml) was added and the reaction
refluxed for 40 min. On cooling to room temperature,
diamide 8a (1.0 g, 3.32 mmol) was added followed by
ethanol (2 ml) and diethyl ether (2 ml). The mixture was
stirred for 1 h, then benzoyl chloride (190 ml, 1.64 mmol)
added. After stirring for 1 h, saturated sodium hydrogen
carbonate (8 ml) was added and the mixture extracted
with ethyl acetate (3×15 ml). The combined organic
layers were dried (MgSO4), filtered and concentrated in
vacuo. Purification by column chromatography (20%
ethyl acetate in light petroleum) gave 9a as a white solid
(0.48 g, 91%); m.p. 91°C; wmax (thin film) 3292 (NH), 2966
(CH), 1686 (CꢁO), 1643 (CꢁO) cm−1; lH (300 MHz,
CDCl3) 8.07 (2H, d, J=7.3 Hz, Ar), 7.60 (1H, t, J=7.4
Hz, Ar), 7.48 (2H, t, J=7.4 Hz, Ar), 6.63 (1H, bs, NH),
6.21 (1H, bs, NH), 5.66 (1H, bs, CH), 3.18 (2H, m,
Table 2. Selected 13C NMR chemical shifts for azino-
mycin A 1 and 10
Carbona
1b
10
Cꢀ1
Cꢀ2
Cꢀ3
27.2
202.6
50.6
27.2
203.5
50.3
Cꢀ6
Cꢀ7
Cꢀ8
Cꢀ10
Cꢀ11
Cꢀ12
Cꢀ13
Cꢀ17
163.2
120.1
149.6
35.8
45.4
76.9
164.5
117.6
153.3
34.9
44.5
23.3
84.0
163.8
26.6
157.4
t
CH2N), 1.52-1.44 (11H, m, Bu, NCH2CH2), 0.86 (3H, t,
a Spectra recorded in CDCl3 (100 MHz).
b Data from Ref. 13.
J=7.4 Hz, CH3); dC (75 MHz, CDCl3) 193.5 (s), 166.5
(s), 155.4 (s), 134.5 (s), 134.3 (d), 129.3 (d), 128.6 (d), 80.9

9576
J.-Y. Goujon, M. Shipman / Tetrahedron Letters 43 (2002) 9573–9576
(s), 61.1 (d), 41.4 (t), 28.2 (q), 22.5 (t), 11.2 (q); m/z
(0.7 ml min−1; 17.5% IPA in hexane; u 254 nm): 23.5
min (major), 25.1 min (minor).
12. For the first application of this methodology, see:
Hashimoto, M.; Matsumoto, M.; Yamada K.;
Terashima, S. Tetrahedron Lett. 1994, 35, 2207–
2210.
13. Yokoi, K.; Nagaoka, K.; Nakashima, T. Chem. Pharm.
Bull. 1986, 34, 4554–4561.
14. Fletcher, D. A.; McMeeking, R. F.; Parkin, D. J.
(CI+) 321 (MH+), 265, 221; Observed (MH+): 321.1821;
C17H25N2O4 requires 321.1814.
9. All new compounds were fully characterised by stan-
dard spectroscopic and analytical techniques.
10. Partial NOE data for 13 (azinomycin numbering): Irra-
diation of Hꢀ16 enhanced Hꢀ5 (5.8%), Hꢀ13 and Hꢀ13%
(2.5%); simultaneous irradiation of Hꢀ13 and Hꢀ13%
enhanced Hꢀ16 (1.8%), Hꢀ12 (1.1%) and Hꢀ12% (1.2%).
11. HPLC performed using a Daicel chiralpak AD column
Chem. Inf. Comp. Sci. 1996, 36, 746–749.