Rapid assembly of the 1-azabicyclo[3.1.0]hexane skeleton of ficellomycin

Article abstract of DOI:10.1055/s-2004-831313

The 1-azabicyclo[3.1.0]hexane core of ficellomycin can be assembled in a concise manner by way of a double cyclisation reaction involving in situ generation of an NH aziridine and further intramolecular conjugate addition onto a dehydroamino ester.

Full text of DOI:10.1055/s-2004-831313

2212
LETTER
Rapid Assembly of the 1-Azabicyclo[3.1.0]hexane Skeleton of Ficellomycin
1-Azabicyclo[3.1.
.
Pleton of Ficel
a
a
b
c
Department of Chemistry, University of Exeter, Stocker Road, Exeter, EX4 4QD, UK
Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, UK
AstraZeneca, Process Research and Development, Charter Way, Macclesfield, SK10 2NA, UK
Fax +44(2476)524429; E-mail: m.shipman@warwick.ac.uk
Received 3 June 2004
NH2
Abstract: The 1-azabicyclo[3.1.0]hexane core of ficellomycin can
be assembled in a concise manner by way of a double cyclisation re-
action involving in situ generation of an NH aziridine and further in-
tramolecular conjugate addition onto a dehydroamino ester.
H
NH2
CO2H
3
N
2
H
3
N
2
N
O
5
H
6
CO2H
O
N
N
HN
6
5
Key words: antibiotics, aziridines, 1,2-azido alcohols, conjugate
addition reactions, polymer supported triphenylphosphine
HN
NH
1a
1b
H2N
NH2
Kuo et al. (1989)3
Armstrong and Zhao (1993)5
Figure 1 Structural proposals for ficellomycin
In 1976, Argoudelis et al. described the isolation of ficel-
lomycin (1) from the culture broths of Streptomyces ficel-
lus.1 This natural product was shown to inhibit the in vitro
growth of Gram positive bacteria and Staphylococcus au-
reus infections in mice. Little is known about its mode of
action although there is evidence to suggest that it inhibits
semiconservative DNA replication in bacteria.2 In 1989,
structure 1a was proposed for this antibiotic by Kuo et al.
based on a combination of NMR, mass spectrometry and
derivatisation studies (Figure 1).3 Most intriguingly ficel-
strong have undertaken exploratory studies to test the fea-
sibility of this route to ficellomycin.5 In seeking the most
expedient route to this natural product, we were drawn to
the idea of generating the NH aziridine in situ from the
corresponding 1,2-azido alcohol using a phosphine in-
duced Staudinger-type reaction, which might undergo
spontaneous closure onto a suitably functionalised dehy-
droamino acid. The basic concept is illustrated in
Scheme 1 with the C-5 amino substituent omitted. In this
way, both rings of the natural product might be assembled
in a single chemical transformation. This reaction estab-
lishes the stereochemistry at C-2 and C-3 in the resulting
bicycles. Since the stereochemistry of ficellomycin at
these stereocentres is uncertain (vide supra), the possibil-
ity of fine-tuning the cyclisation to provide access to any
or all of the possible diastereomers is an additional attrac-
tive feature.
lomycin contains
a
highly strained 1-azabicyc-
lo[3.1.0]hexane ring system which has only been
identified once before, in the anti-tumour agents azinomy-
cin A and B.4 Although it is known from degradation stud-
ies that the valine residue of ficellomycin possesses the S-
configuration,1 questions remained concerning the stereo-
chemistry at the other asymmetric centres. The relative
stereochemistry at C-3, C-5 and C-6 within the rigid bicy-
cle was assigned by Kuo using 1H-1H coupling constants
assuming a chair conformation. Armstrong and Zhao have
revised the ring conformation to a boat and the relative
stereochemistry at C-6 to 1b based upon additional NMR
experiments including NOE difference measurements.5
On the basis of its unique chemical architecture, stereo-
chemical ambiguities, and interesting antibiotic activity,
ficellomycin is an attractive and challenging synthetic tar-
get. In this Letter, we describe preliminary studies aimed
at accomplishing its total synthesis.
R1HN
CO2R
H
R1HN
CO2R
NH
R1HN
CO2R
N
2
PPh3
3
OH
N3
where R = alkyl or H; R1= –COCH(iPr)NH2
Scheme 1 Proposed double cyclisation to the ficellomycin skeleton
In considering synthetic approaches to ficellomycin, we
felt that existing routes to 1-azabicyclo[3.1.0]hexanes
were not well suited for the purpose.6 Instead, we were
drawn to the idea of using an intramolecular conjugate ad-
dition of an aziridine onto a dehydroamino acid deriva-
tive,7 i.e. disconnection across the aziridine nitrogen/C-3
bond (ficellomycin numbering). Indeed, Zhao and Arm-
To explore these ideas, we embarked on the synthesis of a
model substrate, namely 1,2-azido alcohol 2. Reaction of
commercially available phosphonate 3 with 4-pentenal8 in
the presence of DBN provided dehydroamino ester 4 in
79% yield as a single geometric isomer after column chro-
matography. Chemoselective epoxidation of diene 4 was
achieved using excess m-chloroperoxybenzoic acid
(mCPBA) providing epoxide 5 in 88% yield. Competitive
oxidation of the alkene functionality of the dehydroamino
SYNLETT 2004, No. 12, pp 2212–2214
0
1
.1
0
.2
0
0
4
Advanced online publication: 26.08.2004
DOI: 10.1055/s-2004-831313; Art ID: D14704ST
© Georg Thieme Verlag Stuttgart · New York

LETTER
1-Azabicyclo[3.1.0]hexane Skeleton of Ficellomycin
2213
7
ester was not observed. At this juncture, it was possible to
confirm the anticipated Z-stereochemistry9 by NOE dif-
ference experiments, which produced strong reciprocal
enhancements between the NH and allylic hydrogens (i.e.
-CH2CH=CNHCbzCO2Me). Regioselective opening of
epoxide 5 was achieved using sodium azide and ammoni-
um chloride in ethanol. This provided azido alcohol 2 in
three steps in 45% overall yield (Scheme 2).
CO2Me
7
CbzHN
PPh2,
2
3
N
PhMe, 70 °C
+
3
2
N
2
CbzHN
CO2Me
6
6a and 6b (21%)
7a (20%);
7b (4%)
Scheme 3
syn-diastereomers 7a and 7b showed NOE enhancements
between H-2 and H-7endo. At this juncture, we are unable
to deduce the relative stereochemistry at C-2 in either the
syn- or anti-diastereomers.
4-pentenal,
DBN, CH2Cl2,
0 °C → r.t.
CbzHN
CO2Me
NHCbz
PO(OMe)2
MeO2C
79%
4
3
Whilst little stereoselectivity is observed in this model
system, the suitability of this strategy for the assembly of
‘ficellomycin-like’ molecules has been demonstrated.
The modest yield obtained for the key step (45% com-
bined) is largely offset by the brevity of the approach. On-
going studies are focused on examining whether greater
stereocontrol can be achieved by the incorporation of the
C-5 amino substituent and the (S)-valine residue into azi-
do alcohol precursors, and to completing the total synthe-
sis of ficellomycin.
mCPBA, CH2Cl2
88%
NHCbz
OH
NHCbz
NaN3, NH4Cl,
EtOH, 3 d
O
MeO2C
MeO2C
65%
N3
2
5
Scheme 2
Several Staudinger-type conditions were examined for the
cyclisation of azide 2 including variations in the nature of
the phosphine (PPh3 or PBu3), solvent (toluene, acetoni-
trile or THF) and reaction temperature (r.t. or heating). It
was rapidly established that heating azide 2 in acetonitrile
at reflux with triphenylphosphine (1.6 equiv) for 24 hours
led to the production of all four expected 1-azabicyc-
lo[3.1.0]hexanes, namely 6a,b and 7a,b. However, the pu-
rification of these aziridines was hampered by the
presence of excess triphenylphosphine and triphenylphos-
phine oxide in the reaction mixture. To overcome this
problem, we resorted to the use of a polymer-supported
variant of triphenyl phosphine, namely diphenylphosphi-
nopolystyrene (ca 3 mmol/g, Fluka). As far as we are
aware, supported phosphines have not been used previ-
ously for NH aziridine formation from 1,2-azido alcohols.
Heating azide 2 with this supported phosphine (1.1 equiv)
for 16 hours in toluene at 70 °C provided 6a,b and 7a,b
devoid of phosphine contamination after filtration and
evaporation (Scheme 3). The distinctive pattern of the
aziridine hydrogens in the 1H NMR spectrum made their
identification straightforward. Integration of these signals
revealed that only modest diastereoselectivity was ob-
served (crude dr 6a:6b:7a:7b = ca 3:1:5:1). These bicy-
cles were quite stable and could be purified by
chromatography on silica gel pretreated with triethyl-
amine. The anti-diastereomers 6a and 6b (which are
epimeric at C-2) were isolated in 21% yield as an insepa-
rable 75:25 mixture. The more polar syn-diastereomers 7a
(20%) and 7b (4%) were isolated as separated compo-
nents. The reaction shows very modest preference for the
production of the syn-diastereomers. All the bicycles have
been fully characterised and the relative stereochemistry
at the ring junction (C-3) solved by NMR methods.10 Both
anti-diastereomers 6a and 6b showed strong reciprocal
NOE enhancements between H-3 and H-7endo. In contrast,
Acknowledgment
We thank the EPSRC (GR/N22489/01) and AstraZeneca for finan-
cial support of this work. We thank the EPSRC National Mass
Spectrometry Centre for performing some of the mass measure-
ments.
References
(1) Argoudelis, A. D.; Reusser, F.; Whaley, H. A.; Baczynskyj,
L.; Mizsak, S. A.; Wnuk, R. J. J. Antibiot. 1976, 29, 1001.
(2) Reusser, F. Biochemistry 1977, 16, 3406.
(3) Kuo, M.-S.; Yurek, D. A.; Mizsak, S. A. J. Antibiot. 1989,
42, 357.
(4) For a review, see Hodgkinson, T. J.; Shipman, M.
Tetrahedron 2001, 57, 4467.
(5) Zhao, Z. PhD Thesis; University of California: Los Angeles,
1993.
(6) For some alternative approaches to 1-azabicyclo[3.1.0]-
hexanes, see: (a) Buyle, R. Chem. Ind. (London) 1966, 195.
(b) Laurent, D. A. Chimia 1976, 30, 274. (c) Hudlicky, T.;
Frazier, J. O.; Seoane, G.; Tiedje, M.; Seoane, A.; Kwart, L.
D.; Beal, C. J. J. Am. Chem. Soc. 1986, 108, 3755.
(d) Mulzer, J.; Becker, R.; Brunner, E. J. Am. Chem. Soc.
1989, 111, 7500. (e) Sasaki, M.; Yudin, A. K. J. Am. Chem.
Soc. 2003, 125, 14242.
(7) Intramolecular conjugate additions involving aziridine
nucleophiles are known. For example, see: (a) Moran, E. J.;
Tellew, J. E.; Zhao, Z.; Armstrong, R. W. J. Org. Chem.
1993, 58, 7848. (b) Coleman, R. S.; Li, J.; Navarro, A.
Angew. Chem. Int. Ed. 2001, 40, 1736. (c) ref. 5.
(8) Prepared by oxidation of of 4-penten-1-ol (PCC, CH2Cl2, 3
h, r.t.), which was used directly in the olefination reaction
without purification.
(9) Hiebl, J.; Kollmann, H.; Rovenszky, F.; Winkler, K. J. Org.
Chem. 1999, 64, 1947.
(10) Selected Spectroscopic Data. Compound 6a/b: IR (neat):
n
max = 3334, 2951, 1713, 1498 cm–1. 1H NMR (400 MHz,
C6D6): d = 7.22–7.15 (2 H, m, ArH), 7.11–7.00 (3 H, m,
Synlett 2004, No. 12, 2212–2214 © Thieme Stuttgart · New York

2214
D. Paumier et al.
LETTER
ArH), 5.87 (0.75 H, d, J = 8.5 Hz, NH), 5.73 (0.25 H, d, J =
8.0 Hz, NH), 5.14–4.98 (2 H, m, CH2Ph), 4.48 (0.25 H, dd,
J = 8.5, 3.8 Hz, H-2), 4.40–4.33 (0.75 H, m, H-2), 3.41 (0.25
H, dd, J = 8.5, 3.5 Hz, H-3), 3.33 (0.75 H, s, CH3O), 3.29
(2.25 H, s, CH3O), 2.95 (0.75 H, t, J = 7.5 Hz, H-3), 2.02–
1.98 (0.75 H, m, H-6), 1.91–1.85 (0.25 H, m, H-6), 1.61–
1.43 (3 H, m, 2 × H-5, H-4), 1.31 (0.75 H, d, J = 5.2 Hz, H-
(1 H, d, J = 4.8 Hz, H-7exo), 1.20–1.05 (2 H, m, H-4), 0.99 (1
H, d, J = 2.0 Hz, H-7endo). 13C NMR (100 MHz, C6D6): d =
172.2 (CH3OCO), 156.2 (NHCOO), 137.2 (ArC), 128.4–
128.3 (ArCH), 67.0 (CH2Ph), 65.4 (C-3), 56.8 (C-2), 51.7
(CH3O), 38.7 (C-6), 26.0 (C-5), 23.1 (C-7), 22.2 (C-4). MS
(ES+): m/z = 305 [MH+], 91. HRMS (ES+): m/z calcd for
C16H21N2O4: 305.1496; found: 305.1492 [MH+]. Anal.
Calcd for C16H20NO4: C, 63.14; H, 6.62; N, 9.20%. Found:
C, 63.10; H, 6.65; N, 8.80%. Compound 7b: IR (neat):
7exo), 1.29 (0.25 H, d, J = 5.5 Hz, H-7exo), 1.18–0.88 (1 H, m,
H-4), 0.56 (0.75 H, d, J = 3.3 Hz, H-7endo), 0.53 (0.25 H, d,
J = 3.5 Hz, H-7endo). 13C NMR (100 MHz, C6D6): d = 172.0
(CH3OCO), 171.4 (CH3OCO), 157.0 (NHCOO), 156.2
(NHCOO), 137.2 (ArC), 128.5–128.4 (ArCH), 67.1 (CH2),
67.0 (CH2), 66.9 (C-3), 66.0 (C-3), 59.9 (C-2), 58.1 (C-2),
51.9 (CH3O), 51.7 (CH3O), 41.0 (C-6), 40.0 (C-6), 28.2 (C-
7), 28.0 (C-7), 26.1 (C-5), 25.0 (C-5), 24.3 (C-4), 22.6 (C-4).
MS (ES+): m/z = 305 [MH+], 261, 91. HRMS (ES+): m/z
calcd for C16H21N2O4: 305.1496; found: 305.1493 [MH+].
Compound 7a: IR (neat): nmax = 3319, 3172, 2947, 1712,
1527, 1435 cm–1. 1H NMR (400 MHz, C6D6): d = 7.22–7.18
(2 H, m, ArH), 7.11–7.01 (3 H, m, ArH), 5.99 (1 H, d, J = 9.0
Hz, NH), 5.09 (1 H, d, J = 12.3 Hz, CHHPh), 5.01 (1 H, d,
J = 12.3 Hz, CHHPh), 4.54 (1 H, t, J = 8.4 Hz, H-2), 3.33 (3
H, s, CH3O), 3.21–3.14 (1 H, m, H-3), 2.07–2.02 (1 H, m, H-
6), 1.59–1.52 (1 H, m, H-5), 1.43–1.33 (1 H, m, H-5¢), 1.26
n
max = 3368, 3181, 2948, 1710, 1499 cm–1. 1H NMR (400
MHz, C6D6): d = 7.22–7.16 (3 H, m, ArH and NH), 7.11–
6.99 (3 H, m, ArH), 5.08 (1 H, d, J = 12.0 Hz, CHHPh), 5.03
(1 H, d, J = 12.0 Hz, CHHPh), 4.33 (1 H, dd, J = 8.8, 6.6 Hz,
H-2), 3.38–3.29 (4 H, m, CH3O and H-3), 2.15–2.09 (1 H, m,
H-6), 1.51 (1 H, dd, J = 13.0, 8.0 Hz, H-5), 1.44–1.34 (1 H,
m, H-5¢), 1.22 (1 H, d, J = 5.0 Hz, H-7exo), 1.18–1.09 (1 H,
m, H-4), 1.05–0.96 (1 H, m, H-4¢), 0.93 (1 H, m, H-7endo). 13
NMR (100 MHz, C6D6): d = 172.0 (CH3OCO), 156.6
(NHCOO), 137.3 (ArC), 128.5–128.2 (ArCH), 70.0
C
(CH2Ph), 63.7 (C-3), 57.7 (C-2), 51.8 (CH3O), 38.8 (C-6),
26.2 (C-5), 22.8 (C-7), 22.5 (C-4). MS (ES+): m/z = 305
[MH+], 261, 91. HRMS (ES+): m/z calcd for C16H21N2O4:
305.1496; found: 305.1496 [MH+].
Synlett 2004, No. 12, 2212–2214 © Thieme Stuttgart · New York