Angewandte
Chemie
Figure 2. 1H NMR spectra (400 MHz, D2O): a) 1:3 mol ratio of syn-
thetic (ꢀ)-10 and natural (+)-1; b) synthetic (ꢀ)-10; c) synthetic (ꢀ)-1;
d) 1:2 mol ratio of synthetic (ꢀ)-1 and natural (+)-1. Concentrations
of approximately 10 mm, no solvent suppression.
Scheme 1. Synthesis of (ꢀ)-10. Reagents and conditions: a) TBDPSCl,
imidazole, DMF, 08C!RT, 4 h, 91%; b) MeOCH2Cl, Hünig’s base,
CH2Cl2, 08C!RT, 56 h, 98%; c) TBAF, THF, ꢀ108C, 4 h, 95%;
d) Lindlar cat., H2, (1 atm), EtOH, 14 h, 98%; e) BnBr, K2CO3, CH3CN,
31 h, 91%; f) 1. (COCl)2, DMSO, CH2Cl2, ꢀ788C; 2. Et3N, 94%;
g) 1. 5, nBu2BOTf, Et3N, CH2Cl2, ꢀ78!08C, 3 h; 2. 4, ꢀ78!08C,
2.5 h, 77%, d.r. 24:1; h) H2O2, LiOH, 08C, 30 min, 96%; i) TFA, 08C,
1 h, 98%; j) 1. 6, EDCI, HOBt, DMF, 08C, 10 min; 2. (ꢀ)-8, Et3N,
08C!RT, 1 h, 81%; k) 1. HCl, MeOH, H2 (5 atm), Pd/C, 1 h; 2. HCl,
H2O, H2 (5 atm), Pd/C, 1 h, 76%. Bn=benzyl, Boc=tert-butoxycar-
bonyl, DMF=N,N-dimethylformamide, DMSO=dimethyl sulfoxide,
EDCI=1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride,
HOBt=1-hydroxybenzotriazole, MOM=methoxymethyl, TBAF=tetra-
n-butylammonium fluoride, TBDPS=tert-butyldiphenylsilyl, TFA=tri-
fluoroacetic acid.
Scheme 2. Synthesis of model (ꢀ)-1. Reagents and conditions:
a) 1. mW, toluene, 1108C, 15 min; 2. THF, NH3, 30 min; 3. NH3 (2m),
MeOH, 5 h; 4. NaOH (1n), MeOH, 4.5 h, 62%; b) TFA, 08C, 1 h,
99%; c) 1. EDCI, HOBt, DMF, 08C, 10 min; 2. (+)-8, Et3N, 08C!RT,
1 h, 88%; d) 1. HCl, MeOH, H2 (5 atm), Pd/C, 1 h; 2. HCl, H2O, H2
(5 atm), Pd/C, 1 h, 75%.
rearrangement and subsequent ammoniolysis). Removal of
the protecting groups of 12 under the reaction conditions that
were previously used in Scheme 1[14] gave (ꢀ)-1 in 75% yield.
The NMR spectra ofsynthetic ( ꢀ)-1 and natural (+)-1
were identical in all respects; co-addition ofnatural ( +)-1 to
topicities and 13C NMR chemical shift differences of a C2-
symmetric diamino tetraol derived from 2.[5] Considering that
the configuration of the amino acid in (+)-1 was unequiv-
1
1
ocally l,[5] the H and 13C NMR signals at C10–C15 showed
(ꢀ)-1 gave a single discrete set of H (Figure 2 d) and 13C
1
negligible differences, and the largest H NMR difference in
signals corresponding to those ofnatural ( +)-zwittermi-
cin A.[1a]
chemical shift occurred at H8, we hypothesized that the
mismatch resulted from the inversion of all configurations in
the diaminopolyol–carboxylate moiety of( ꢀ)-10: C8–C11,
C13, and C14. The latter (inversion ofconfiguration at C14)
would negate the original assumption ofa formal biosynthesis
of 1 derived from an l-Ser starter unit[4a,5] in the NRPS
loading domain and would therefore require the involvement
of d-Ser. To test this hypothesis compound 12, a diastereomer
of 9, was prepared by coupling carboxylic acid 6 with d-a-
aminoamide (+)-8 (88%; Scheme 2, where (+)-8 was derived
from the known acyl azide 11[16] in 2 steps by Curtius
Finally, the specific rotation of synthetic (ꢀ)-1 ([a]D =
ꢀ7.98, H2O) was opposite in sign and equal in magnitude to
natural zwittermicin A. Therefore the configuration of zwit-
termicin A ((+)-1) is 4S,8R,9S,10S,11S,13S,14R as depicted
(Figure 1). The assignment ofconifguration described here
has implications for the biosynthesis of (+)-1. Three scenarios
can be considered to explain the unexpected 14R configura-
tion ofzwittermicin A; 1) direct incorporation of d-Ser at
C13–C15 (path a); Scheme 3), which is similar to that
[17a]
observed for the d-Ala starter residue ofcyclosporine,
Angew. Chem. Int. Ed. 2008, 47, 8086 –8089
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim