Total asymmetric synthesis of sperabillins B and D

Article abstract of DOI:10.1039/b305740b

A consise route to the core fragment of sperabillins B and D, methyl (3R,5R,6R)-3,6-diamino-5-hydroxyheptanoate, has been developed with a subsequent novel protection strategy allowing the total asymmetric synthesis of sperabillins B and D.

Full text of DOI:10.1039/b305740b

Total asymmetric synthesis of sperabillins B and D†
Stephen G. Davies,* Richard J. Kelly and Anne J. Price Mortimer
Dyson Perrins Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QY.
E-mail: steve.davies@chemistry.ox.ac.uk
Received (in Cambridge, UK) 22nd May 2003, Accepted 8th July 2003
First published as an Advance Article on the web 21st July 2003
A consise route to the core fragment of sperabillins B and D,
methyl (3R,5R,6R)-3,6-diamino-5-hydroxyheptanoate, has
been developed with a subsequent novel protection strategy
allowing the total asymmetric synthesis of sperabillins B and
D.
Conjugate addition of lithium (R)-N-allyl-N-a-methylbenzyla-
mide9 to 6 generated the b-amino acid derivative 7 as a single
diastereomer ( > 96% de).10 Replacement of the allyl protecting
group with benzyloxycarbonyl gave alkene 8, and subsequent
iodocarbamation generated oxazin-2-one 9 in 84% de, which
was improved to > 98% de on recrystallisation. Azide displace-
ment of the iodide with inversion of configuration followed by
reduction installed the C-6 amino group. Global deprotection
using concentrated aqueous acid gave, after re-esterification,
the methyl ester of the core fragment 10 in 19.3% overall yield
on a multigram scale.
The sperabillin family of antibiotics 1–4, isolated from the
culture filtrates of Pseudomonas fluoresens YK-437,1 are active
in vitro and in vivo against Gram-positive and Gram-negative
bacteria including antibiotic resisitant strains.2 Notably, their in
vivo potencies are greater than expected from their in vitro
activities. Sperabillin polymers have also been shown to have
antitumour activity.3 The structures of the sperabillins 1–4,
including the absolute configurations, were elucidiated by the
degradation studies of Hida et al.4 and by the total enantiospe-
cific synthesis of sperabillin D 4 by Natsugari et al.5
Sperabillins A 1 and C 3 have the same core amino acid as the
potent antibiotic negamycin,6 whilst sperabillins B 2 and D 4
(which are more active than A 1 and C 3 respectively)2 bear an
additional C-6 methyl substituent and consequently an addi-
tional stereogenic centre. We have previously reported the
asymmetric synthesis of the highly functionalised (3R,5R,6R)-
3,6-diamino-5-hydroxyheptanoic acid 5,7 the common core
fragment of sperabillins B 2 and D 4 (Fig. 1). Herein we
describe an improved synthesis of the core fragment and its
conversion via a novel protection strategy to both sperabillins B
2 and D 4.
In order to achieve selective amide formation on the C-6
amine group a simple protection strategy was envisaged.
Treatment of the methyl ester 10 with acetone under dehy-
drative conditions was anticipated to generate the correspond-
ing bis-imine which would be expected to cyclise exclusively to
the cis-1,2 disubstituted 5-ring oxazolidine 11 rather than the
alternative trans-1,3-disubstituted oxazinane. Consistent with
this notion, sequential treatment of 10 with acetone/molecular
sieves and sorbyl chloride/Hunig’s base gave, after chromatog-
raphy, the desired C-6 amine coupled derivative 12 (Scheme 2).
N-Boc protection of the C-3 amine of 12 followed by ester
hydroylsis under basic conditions gave the fully protected free
acid 13. Finally the amidine bearing side chain was introduced
by coupling with 3-aminopropionamidine11 using DCC and
Analogously to our previous work,7 a strategy of ster-
eoselectively introducing the C-3 amino group by lithium amide
conjugate addition to a a,b,d,e-unsaturated acceptor, followed
by installation of the other stereogenic centres via an iodocarba-
mation, was envisaged. A more efficient acceptor synthesis than
that previously employed was necessary for a viable synthesis.
Thus, the palladium catalysed co-dimerisation of butadiene and
methyl acrylate8 was optimised to give methyl (2E,5E)-hepta-
2,5-dienoate 6 in very good yield, with only a small amount
(8%) of Diels–Alder product being generated (Scheme 1).
Scheme 1 Reagents and conditions: (i) Pd2(dba)3 (1.55 mmol, 0.2 mol%),
HBF4 etherate (6.20 mmol), PBu3 (3.10 mmol), 80 °C, neat, sealed tube, 5
h; (ii) lithium (R)-N-allyl-N-a-methylbenzylamide (184 mmol, 1.3 eq),
THF (400 mL), 278 °C, 2 h; (iii) Wilkinson’s catalyst (3.14 mmol, 5
mol%), MeCN/H2O (4 : 1, 500 mL), reflux, 2 h; (iv) Cbz2O (214 mmol, 2
eq), vacuum, neat, 4 days; (v) I2 (21.2 mmol, 5 eq), DCM (50 mL), 0 °C, 5
h; (vi) (a) NaN3 (37.7 mmol, 5 eq), DMF/H2O (25 : 1, 62.5 mL), 110 °C, 5
h; (b) H2 (1 atm), Pd/C (0.094 mmol, 0.1 mol%), MeOH (60 mL), 18 h; (vii)
(a) 5 M HCl (50 mL, excess), 100 °C, 24 h; (b) MeOH (50 mL), SOCl2 (21.4
mmol, 4.5 eq), reflux, 18 h.
Fig. 1
† Electronic supplementary information (ESI) available: experimental
details. See http://www.rsc.org/suppdata/cc/b3/b305740b/
2132
CHEM. COMMUN., 2003, 2132–2133
This journal is © The Royal Society of Chemistry 2003

spectroscopic data were entirely consistent with the natural
product, in 56% overall yield from 10.
This straightforward procedure for the selective sequential C-
6 NH2 and C-1 CO2H amide formation should be applicable to
the rapid generation of libraries of analogues of sperabillins D
4 and B 2 and hence allow construction of detailed structure–
activity relationships, as evidenced by its application to the first
asymmetric synthesis of sperabillin B 214 (Scheme 3), whose
spectroscopic data were entirely consistent with those of the
natural product.
In conclusion, a concise route to the core b,e-diamino-g-
hydroxy fragment 10 has been developed (in 19.3% yield from
methyl acrylate) and a novel protection strategy utilised to
complete the syntheses of the highly functionalised antibiotics
sperabillins B 2 and D 4. The first asymmetric total synthesis of
sperabillin B 2 was achieved in 5.8% overall yield from methyl
acrylate (30% from 10). An efficient total synthesis of
sperabillin D 4 has been achieved in 10.8% overall yield (56%
from 10).
Rhoˆne-Poulenc Roˆrer (R. J. K.) and EPSRC (A. J. P. M.) are
gratefully acknowledged for funding.
Notes and references
1 (a) S. Harada and H. Ono, Eur. Pat. Appl., 1986, EP 206068; (b) N.
Katayama, Y. Nozaki, S. Tsubotani, M. Kondo, S. Harada and H. Ono,
J. Antibiot., 1992, 45, 10.
2 N. Katayama, Y. Nozaki, S. Tsubotani, M. Kondo, S. Harada and H.
Ono, J. Antibiot., 1992, 45, 10.
3 T. Hida, S. Tsubotani, A. Hori, M. Murakami, H. Natsugari, Y. Kozai
and S. Harada, Chem. Pharm. Bull., 1993, 41, 889.
4 T. Hida, S. Tsubotani, Y. Funabashi, H. Ono and S. Harada, Bull. Chem.
Soc. Jpn., 1993, 66, 863.
Scheme 2 Reagents and conditions: (i) acetone (15 mL), 3 Å molecular
sieves, Hunig’s base (1.52 mmol, 2 eq), reflux, 1 h then sorbyl chloride
(0.863 mmol), Hunig’s base (0.863 mmol), 0 °C to room temperature, 18 h;
(ii) Boc2O (0.74 mmol, 1.5 eq), NaHCO3 (1.47 mmol), MeOH (15 mL), 72
h; (iii) NaOH (aq) (1.65 mmol, 4 eq), THF/MeOH (2 : 1, 15 mL), 18 h; (iv)
DCC (0.08 mmol, 1.35 eq), HOBt (0.071 mmol), THF (2 mL), 3 Å
molecular sieves, 3 h, then 3-aminopropionamidine dihydrobromide (0.059
mmol), NaHCO3 (0.12 mmol), THF/H2O (10 : 1, 20 mL), 48 h; (v) TFA (1
mL, excess), DCM (1 mL), 30 min then Amberlite IRA-402 (Cl2 form).
5 S. Hashiguchi, A. Kawada and H. Natsugari, J. Chem. Soc., Perkin
Trans. 1, 1991, 2435.
6 M. Hamada, T. Takeuchi, S. Kondo, Y. Ikeda, H. Naganawa, K. Maeda,
Y. Okami and H. Umezawa, J. Antibiot., 1970, 23, 171.
7 S. G. Davies and O. Ichihara, Tetrahedron Lett., 1999, 40, 9313.
8 P. Grenouillet, D. Neibecker and I. Tkatchenko, Fr. Demande, 1987, FR
86-4644 19860327.
HOBt. The purity of the highly polar amidine product was
ensured by work-up with MP-carbonate scavenger resin.12
Acidic global deprotection afforded sperabillin D 4,13 whose
9 S. G. Davies and D. R. Fenwick, J. Chem. Soc., Chem. Commun., 1995,
1109.
10 S. G. Davies, K. Iwamoto, C. A. P. Smethurst, A. D. Smith and H.
Rodriguez-Solla, Synlett, 2002, 7, 1146.
11 G. Hilgetag, H. Paul, J. Günther and M. Witt, Chem. Ber., 1964, 704.
12 J. J. Weidner, J. J. Parlow and D. L. Flynn, Tetrahedron Lett., 1999, 40,
239.
13 Sperabillin D 4: [a]D25 +27.4 (c 0.22, H2O), lit.4 [a]2D5 +30.4 (c 0.50,
H2O); nmax/cm21 (KBr disc) 3387 (s), 1686 (m), 1654 (s), 1627 (m),
1612 (m), 1550 (m); dH (500 MHz, D2O) 0.96 (3H, d, J 7.0, CH3CHN),
1.47–1.62 (2H, m, CHCH2CH), 1.61 (3H, d, J 6.0, CH3CHNCH), 2.43
(2H, t, J 6.9, CH2CNN), 2.49 (2H, d, J 6.6, CH2CNO), 3.27–3.37 (2H, m,
NCH2), 3.60–3.65 (2H, m, CHOH and CH2CHCH2), 3.77 (1H, dq, J 7.0,
3.5, CH3CHN), 5.75 (1H, d, J 15.2, CHNCHCNO), 5.99–6.10 (2H, m,
CH3CHNCH), 6.90 (1H, dd, J 15.2, 9.9, CHNCHCNO); dC (125 MHz,
D2O) 18.4 (CH3CHN), 20.3 (CH3CHNCH), 34.9 (CH2CNN), 37.0
(CHCH2CH), 38.8 (CH2N), 39.5 (CH2CNO), 48.8 (CHNH2), 51.8
(CHNH), 72.0 (CHOH), 122.6 (CHNCHCNO), 131.6 (CH3CHNCH),
142.7 (CH3CHNCH), 144.9 (CHNCHCNO), 171.1, 171.3 (CNN and
CHNCHCNO), 174.2 (CH2CNO).
14 Sperabillin B 2: [a]2D2 +48.3 (c 0.24, H2O), lit.4 [a]D +56.0 (c 1.0, H2O);
nmax/cm21 (KBr disc) 3269 (m), 3068 (m), 1692 (s), 1654 (s), 1618 (s),
1609 (s), 1546 (s); dH (500 MHz, D2O) 1.10 (3H, d, J 6.7, CH3CHN),
1.61–1.74 (2H, m, CHCH2CH), 1.77 (3H, d, J 7.3, CH3CHNCH),
2.53–2.60 (2H, m, CH2CNN), 2.62 (2H, d, J 6.7, CH2CNO) 3.41–3.49
(2H, m, CH2N) 3.72–3.78 (2H, m, CH2CHCH2 and CHO), 3.92 (1H, qd,
J 6.7, 3.8, CH3CHNH), 5.91 (1H, dq, J 10.7, 7.3, CH3CHNCH), 5.97
(1H, d, J 15.2, CHNCHCNO), 6.12 (1H, app t, J 11.2, CH3CHNCH), 7.45
(1H, dd, J 15.2, 11.7, CHNCHCNO); dC (125 MHz, D2O) 13.6
(CH3CHNCH), 16.4 (CH3CHN), 35.0 (CH2CNN), 36.6 (CH2N and
CHCH2CH), 37.5 (CH2CNO), 46.7 (CH2CHCH2), 49.8 (CH3CHN),
70.0 (CHO), 122.8 (CHNCHCNO), 127.1 (CH3CHNCH), 136.9
(CH3CHNCH), 137.0 (C HNCHCNO), 169.2, 172.2 (CNNH, CH2CNO
and CHNCHCNO).
Scheme 3 Reagents and conditions: (i) acetone (15 mL), 3 Å molecular
sieves, Hunig’s base (1.70 mmol, 2 eq), reflux, 1 h then (2E,4Z)-
hexadienoyl chloride15 (0.933 mmol), Hunig’s base (0.933 mmol), 0 °C to
room temperature, 18 h; (ii) Boc2O (0.681 mmol, 1.5 eq), NaHCO3 (1.36
mmol), MeOH (15 mL), 72 h; (iii) NaOH (aq) (0.509 mmol, 4 eq), THF/
MeOH (2 : 1, 4.5 mL), 5 h; (iv) DCC (0.115 mmol, 1.35 eq), HOBt (0.102
mmol), THF (2 mL), 3 Å molecular sieves, 3 h then 3-aminopropionamidine
dihydrobromide (0.085 mmol), NaHCO3 (0.170 mmol), THF/H2O (10 : 1,
20 mL), 48 h; (v) TFA (1 mL, excess), DCM (3 mL), 30 min then
preparative reverse-phase HPLC then Amberlite IRA-402 (Cl2 form).
15 (2E,4Z)-Hexadienyl chloride was synthesised as > 95 : 5 (2E,4Z) :
(2E,4E) via Heck type coupling of tert-butyl acrylate and (Z)-1-bromo-
1-propene.
CHEM. COMMUN., 2003, 2132–2133
2133