Total Syntheses of Vancomycin and Eremomycin Aglycons

Full text of DOI:10.1002/(SICI)1521-3773(19981016)37:19<2700::AID-ANIE2700>3.0.CO;2-P

COMMUNICATIONS
OH
M(4-6)
Total Syntheses of Vancomycin and
Eremomycin Aglycons**
M(2-4)
Cl
O
O
6
4
2
X
HO
OH
David A. Evans,* Michael R. Wood, B. Wesley Trotter,
Timothy I. Richardson, James C. Barrow, and
Jeffrey L. Katz
O
O
O
H
N
H
N
H
N
NHMe
H
N
H
O
N
H
H
O
O
O
CH2CH(CH3)2
NH
5
M(5-7)
Vancomycin, isolated in 1956 from Streptomyces orienta-
lis,[1] is the prototypical member of a large family of antibiotics
whose structures consist of an arylglycine-rich heptapeptide
aglycon to which are appended an array of sugar residues.
Structural diversity within this family of natural products is
created by variation in the aglycon amino acid constituents
and in the nature, position, and number of incorporated sugar
residues.[2] Clinical applications of vancomycin to the treat-
ment of Gram-positive bacterial infections and its efficacy
against methicillin-resistant Staphylococcus aureus have es-
tablished vancomycin as the antibiotic of last resort against
infections by this deadly pathogen.[3] Recently, emergence of
bacterial strains resistant to treatment by this family of
antibiotics[4] has provided the stimulus for us and others[5] to
develop reaction methodology and strategies relevant to the
synthesis of the individual members of this family of natural
products. In this communication we report the first syntheses of
the vancomycin aglycon (1) and the eremomycin aglycon (2).[6]
The initial phase of our synthesis program in this area focused
on the development of reactions critical to the synthesis of this
family of target structures. Relevant methodology includes
new asymmetric amino acid syntheses[7] and new macrocycli-
zation reactions amenable to the construction of macrocyclic
diaryl ether[8] and biaryl-containing[9] tripeptides. Application
of this methodology to the synthesis of bis-dechlorovanco-
mycin (orienticin C) aglycon has recently been reported.[10]
The most formidable synthesis challenges that the vanco-
mycin skeleton presents are the three stereochemical ele-
ments of atropisomerism present in the structure as a
consequence of hindered rotation in each of the cyclic
tripeptide subunits. Our recent efforts have been directed
toward the development of strategies for controlling these
architectural features during the assemblage of the vancomycin
skeleton. A detailed account of studies that have culminated in
the current vancomycin synthesis is presented in the following
communication.[11] These investigations have dictated both
the general synthesis plan (Figure 1) and the strategy for
controlling atropselectivity in each of the cyclic tripeptide
subunits designated as M(2 ± 4), M(4 ± 6), and M(5 ± 7).[12]
The amino acids required for the (4 ± 7)-tetrapeptide were
synthesized by our chiral auxiliary-based methodology.[7, 13]
We envisioned two potential analogues of amino acid 6
(Figure 1, X ˆ H, X ˆ Cl) that could afford the desired M(4 ±
HO2C
NH2
7
OH
OH
1 X = Cl vancomycin aglycon
2 X = H eremomycin aglycon
HO
M(2-4)
Oallyl
O
OH
F
6
4
2
HO
Cl
HO
O
O
NO2
H
N
H
N
H
N
H
NH
N
H
O
H
O
O
NMeBoc
O
NH
O
O
5
CH2CH(CH3)2
NHDdm
MeHN
7
OBn
OBn
BnO
Oallyl
Oallyl
NO2
HO
F
O
OMs
OH
4
6
4
6
Cl
HO
OH
O
X
O
H
H
N
H
N
N
H
N
H
H
NHBoc
NH2
H
H
O
O
M(4-6)
O
O
NH
NH
5
5
HO2C
HO2C
7
7
OH
OH
OH
OH
HO
HO
Figure 1. Assemblage strategy for vancomycin and eremomycin aglycons.
See ref. [6] for abbreviations.
6) chlorine atropisomer from either stereochemical outcome
of an SNAr-based macrocyclization given the versatility of the
Sandmeyer reaction (NO2 !H, or NO2 !Cl). Preparation of
the amino acids 6 commenced with (isothiocyanoacetyl)-
oxazolidinone 3 (Scheme 1).[7c] Stannous triflate mediated
aldol reaction of 3 with aldehyde 4a or 4b[14] proceeded with
good diastereoselection (drˆ 95:5) to afford the derived syn
aldol adducts 5a and 5b. In subsequent manipulations of these
adducts, the presence of the electrophilic halogen-containing
nitroaromatic nucleus, particularly in the chlorinated series
4b ± 14b, necessitated careful optimization of reaction param-
eters in order to minimize undesirable SNAr-based fluoride
displacement reactions. After Boc protection and transfor-
mation to the oxazolidinone in a one-pot sequence, selective
removal of the chiral auxiliary was effected under carefully
controlled conditions[15] to afford carboxylic acids 6a and 6b.
Peptide coupling of 6a and 6b with amino acid 7 (EDCI ´
HCl, HOBt, 4/1 CH2Cl2/DMF) provided dipeptides 7a and
7b. While endocyclic oxazolidinone cleavage of 7a could be
accomplished with LiOH (MeOH/H2O), the corresponding
transformation of chloro-containing 7b was hampered by
competing SNAr-based displacement of the highly reactive
aromatic fluoride on ring 6. Use of Li2CO3 in MeOH
minimized this side reaction and provided 8b, which was
used in unpurified form after aqueous workup. Following Boc
cleavage, coupling of 8a and 8b to amino acid 5 and
subsequent protecting group exchange afforded tripeptides
[*] Prof. D. A. Evans, M. R. Wood, B. W. Trotter, T. I. Richardson,
J. C. Barrow, J. L. Katz
Department of Chemistry & Chemical Biology, Harvard University
Cambridge, MA 02138 (USA)
Fax: (‡1)617-495-1460
E-mail: evans@chemistry.harvard.edu
[**] Financial support has been provided by the National Institutes of
Health (NIH).
Supporting information for this article is available on the WWW
under http://www.wiley-vch.de/home/angewandte/ or from the author.
2700
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3719-2700 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 19

COMMUNICATIONS
X
6
X
O
O
F
F
NCS
a
X
X
O
N
6
F
F
OH
O
b, c
O2N
e (X=Cl)
d
O2N
6
O
6
Bn
O
O
O
O
f (X=H)
N
3
O2N
O2N
O
NHBoc
O
O
S
X
Boc
XP
NH
N
NH
N
H
HO2C
MeHN
MeHN
F
Boc
O
6
O
7
77%
46%
5a X = H
5b X = Cl
6a X = H
6b X = Cl
8a X = H
8b X = Cl
7
O2N
MeO
OMe
81%
MeO
X
OMe
H
4a X = H
4b X = Cl
7a X = H
7b X = Cl
91% from 7a
82% from 7b
72%
g
X
F
Oallyl
F
X
6
OH
RO
OMs
6
F
O2N
O
OH
H
4
O2N
6
O
NHR
OBn
OH
k (X=Cl)
O
O
N
H
NHR
OH
O2N
N
H
N
H
65% from 11b
O
i
NH
O
N
NH
MeHN
NHBoc
O
l (X=H)
O
5
65%
MeHN
O
NH
5
OH
O
77% from 11a
MeHN
5
7
dr >95:5
7
OMe
OMe
OMe
MeO
OMe
MeO
j
7
OMe
OMe
MeO
m
85%
13a,b R = TBS
14a,b R = H
h
9a,b R = Boc
10,ab R = Tfa
11a,b R = Tfa
12a,b R = H
88-98%
Scheme 1. Synthesis of the M(5 ± 7)-tetrapeptide. a) Sn(OTf)2, N-ethylpiperidine, THF, 788C, then 4; b) Boc2O, DMAP, CH2Cl2, then 1:1 HCO2H/30%
H2O2; c) 3.1 equiv LiOOH, 58C (X ˆ Cl), 1.2 equiv LiOOH, 108C (X ˆ H); d) amino acid 7, TFA, DMS, CH2Cl2, 08C, then 6, EDCI ´ HCl, HOBt,
CH2Cl2/DMF (X ˆ Cl), THF (X ˆ H), 08C; e) Li2CO3, MeOH, room temperature; f) LiOH, MeOH/H2O; g) TFA, DMS, CH2Cl2, 08C, then amino acid 5,
EDCI ´ HCl, HOBt, THF, 08C; h) TFA, DMS, CH2Cl2, 08C, then TFAA, 2,6-lutidine, CH2Cl2, room temperature; i) VOF3, BF3 ´ Et2O, AgBF4, TFA/CH2Cl2,
08C, then NaHB(OAc)3; j) NaHCO3, MeOH, H2O, room temperature, 6 d (X ˆ Cl), 9 ± 14 d (X ˆ H); k) amino acid 4, HATU, HOAt, collidine, CH2Cl2/
DMF, 208C, 16 h; l) isobutyl chloroformate, N-methylmorpholine, EtOAc, 10 ! 58C, 30 min, then amino acid 4 in EtOAc/DMF, 208C ! rt, 2 h;
m) HF ´ pyridine, THF, room temperature, 1 h (X ˆ Cl), TBAF, CH2Cl2, 08C, 15 min (X ˆ H). See ref. [6] for abbreviations.
displacement for either 12a or 12b by HPLC. Coupling[16] of
Oallyl
HO2C
NHBoc
OBn
O
12 and amino acid 4 provided tetrapeptides 13. As a prelude
to M(4 ± 6) macrocyclization, silyl deprotection of the phe-
nolic hydroxyl in ring 4 was accomplished for 13a under basic
conditions (TBAF, THF), while acidic deprotection condi-
tions (HF ´ pyridine) proved necessary for substrate 13b.[17]
Atropisomer control during the construction of the M(4 ± 6)
macrocycle has posed a major challenge to the development
of a successful vancomycin synthesis. In conjunction with
evaluating potential strategies for achieving the proper
disposition of the chloro substituent in ring 6, substrates 14a
and 14b were examined with respect to the diastereoselec-
tivity of their individual cyclization processes (Scheme 2).
Optimal conditions for intramolecular SNAr cyclization
(Na2CO3, DMSO, room temperature),[18] when applied in
the dechloro series starting from 14a, provided the undesired
nitro atropisomer (drˆ 10:1); in situ derivatization of the
phenolic hyxdoxy group in ring 5 afforded 15. Reduction of
the nitro group in 15 provided the major atropisomer 16 as a
crystalline aniline whose structure was established by X-ray
analysis.[19, 20] These results led to the speculation that the
analogue 14b with a chloro substituent in ring 6 might also
cyclize with the same stereochemical outcome (14b !17) if
the stereochemical bias for the nitro group were maintained.
In the event, cyclization of 14b afforded the desired atro-
pisomer 17 as a 5:1 mixture of atropdiastereomers in 79%
yield (Scheme 2). The enhanced electrophilicity of 14b was
evident in the striking rate difference for the two cyclizations
(14a: 66 h; 14b: 1.5 h). More remarkable was the fact that 14b
TBSO
OMs
NHBoc
MeHN
4
5
7
HO2C
NHBoc
OMe
MeO
OMe
amino acid 4
amino acid 5
amino acid 7
10a and 10b in excellent overall yield for the hydrolysis and
coupling events (see Scheme 1).
Oxidative cyclization of 10a and 10b according to our
previously reported conditions (VOF3)[9, 10] provided the
M(5 ± 7) cyclic tripeptides 11a and 11b as their unnatural R
atropisomers (dr> 95:5). The reductive quench for this
oxidation was modified to employ NaBH(OAc)3 in place of
zinc dust to avoid the unwanted reduction of the nitro group
in ring 6. As noted previously, the desired cleavage of the aryl
benzyl ether in ring 5 was also observed under the reaction
conditions.[9, 10] We also investigated the analogous oxidative
cyclizations of more conventional amino acid 7 ester ana-
logues of tripeptide 10. However, these substrates exhibited
significant levels of epimerization of the C-terminal amino
acid residue during the cyclization process, a side reaction that
was suppressed through use of the illustrated N-methyl amides.
In preparation for incorporating amino acid 4 into the
M(5 ± 7) tripeptide 12, removal of the trifluoroacetamide
protecting group in 11 was accomplished by treatment with
NaHCO3 in MeOH/H2O. These conditions effected conver-
sion into amines 12 with no detectable aromatic fluoride
Angew. Chem. Int. Ed. 1998, 37, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3719-2701 $ 17.50+.50/0
2701

COMMUNICATIONS
Oallyl
Y
X
Oallyl
Oallyl
NO2
O
OMs
F
HO
OMs
O
OMs
4
6
6
4
4
6
HO
H
OH
Cl
HO
O2N
H
H
H
N
H
H
H
O
N
O
O
N
N
O
N
O
N
O
NHBoc
NHBoc
NHBoc
O
O
O
O
NH
O
O
OTf
NH
NH
OTf
OH
a (X = Cl)
a (X = H)
MeHN
MeO
MeHN
MeHN
MeO
5
5
5
66 h, 78%
dr = 10:1
1.5 h, 79%
dr = 5:1
7
7
7
OMe
OMe
OMe
OMe
OMe
OMe
b, c, d 77%
MeO
b
74%
15 Y = NO2
16 Y = NH2
14a X = H
14b X = Cl
17
OR1
OMs
4
OPiv
Oallyl
O
O
O
OMs
OH
6
6
5
6
4
4
HO
HO
Cl
Cl
Cl
HO
O
O
H
H
g, h
H
i – m
65%
H
H
N
N
H
N
N
O
N
N
O
NHR2
H
H
NHTfa
NH2
54%
O
H
H
O
O
O
NH
O
O
NH
NH
18 R1 = H, R2 = Boc
O
O
MeHN
dr > 95:5
5
5
e, f, 88%
7
MeHN
MeHN
20
7
21
7
OMe
OMe
OH
OH
OBn
OBn
19 R1 = Piv, R2 = Tfa
MeO
HO
BnO
Scheme 2. Synthesis of the M(4 ± 6)(5 ± 7) bicyclic tetrapeptide. a) Na2CO3, DMSO, room temperature, 66 h (X ˆ H), 1.5 h (X ˆ Cl), then Tf2NPh, 1 h;
b) Zn0, HOAc, EtOH, 408C; c) NaNO2, H3PO2, cat. Cu2O, THF/H2O, 08C, 1 h; d) [Pd(dppf)Cl2] ´ CH2Cl2, Et3N, HCO2H, DMF, 758C; e) PivCl, Et3N, THF,
room temperature; f) TFA, DMS, CH2Cl2, 08C, then TFAA, 2,6-lutidine, CH2Cl2, 08C to room temperature; g) AlBr3, then EtSH, 08C; h) MeOH, 558C;
i) BnBr, Cs2CO3, Bu4NI, DMF, 08C; j) LiSEt, THF, 08C; k) allyl-Br, Cs2CO3, DMF, 08C; l) LDA, THF, 788C; m) LiOH, THF/H2O/MeOH, 08C. See
ref. [6] for abbreviations.
cyclized at an appreciable rate upon dissolution in DMSO and
other polar, aprotic solvents without added base. Bicyclic
peptide 17 was submitted to reduction (Zn, HOAc, EtOH,
408C) and diazotization/reduction (NaNO2, H3PO2, Cu2O,
THF/H2O, 08C),[21] to give the desired monochloride in 88%
yield. NOE experiments confirmed that the major atropisom-
er possessed the desired R configuration.[22] Palladium-
catalyzed reduction of 17 removed both the extraneous aryl
triflate on ring 5 and the phenolic allyl protecting group on
ring 4 (Scheme 2),[23] affording phenol 18 in 87% yield.[24]
Protecting group adjustment (18 !19) and cleavage of the
phenolic methyl ethers (AlBr3/CH2Cl2, EtSH, 08C) was
followed by M(5 ± 7) atropisomerization (MeOH, 558C,
24 h) to give the bicyclic peptide 20, which possesses the
desired S biaryl configuration (dr> 95:5). It is noteworthy
that 20 and its analogue without a chloro substitutent in ring 6,
which comprise the M(5 ± 7) aglycon subunits common to
every member of the vancomycin family, contain the 5 ± 6 cis-
amide found in vancomycin.[25] In preparation for the final
peptide coupling, the triphenol was benzylated, the phenolic
hydroxy group in ring 4 was transformed to its allyl ether, and
the mesylate and trifluoroacetamide protecting groups were
removed to provide free amine 21.
The completion of the aglycon synthesis is summarized in
Scheme 3. Coupling of 21 and tripeptide 22[10b] (EDCI ´ HCl,
HOAt, THF, 08C) afforded heptapeptide 23 in 86% yield
with no detectable epimerization (Scheme 3). Under optimal
conditions (CsF, DMSO, room temperature, 5 h), intramolec-
ular SNAr cyclization provided the fully assembled tricyclic
core of vancomycin as a 5:1 mixture of M(2 ± 4) atropisomers
favoring the desired R configuration 24 (95% yield). The
mixture of nitro atropisomers was reduced (Zn, AcOH, EtOH,
408C) to provide two diastereomeric anilines which were
separated by silica gel chromatography, affording a single
atropisomer in 80% yield. Sandmeyer transformation[26] to
the protected vancomycin aglycon 25 proceeded in 81% yield.
Deprotection of 25 to vancomycin aglycon (1) followed the
sequence developed for the synthesis of orienticin C.[10b] The
N-methyl amide moiety at the carboxyl terminus was
removed by selective nitrosation in the presence of seven
other amidic functional groups (N2O4, NaOAc, CH2Cl2,
MeCN, 08C) followed by treatment with lithium hydrogen
peroxide to give 26 in 68% yield.[27] Superficially, this
transformation appears improbable in view of the other
competing nitrosation sites; however, the steric effects
associated with amide nitrosation have been documented.[28]
Subsequent deprotection of allyl ether 26 ([Pd(PPh3)4],
morpholine, THF, 08C) proceeded in 62% yield. Benzyl
deprotection in the presence of the aryl chlorides was
achieved using transfer hydrogenation conditions (Pd/C, 1,4-
cyclohexadiene, EtOH, room temperature, 70% yield).[29]
Final deprotection of the acid-labile nitrogen-protecting
groups (Boc, Ddm) was effected through prolonged exposure
to trifluroacetic acid (TFA/CH2Cl2, 3/1, DMS, 83% yield)
to afford vancomycin aglycon (1), which proved identical
(1H NMR, COSY, HPLC, MS, [a]) to a natural sample.[30, 31]
The route detailed here provides vancomycin aglycon (1) in
40 steps (longest linear sequence) from 3,5-dimethoxybenzyl
alcohol (amino acid 7). Eremomycin aglycon (2)[32] was also
successfully synthesized through an identical strategy. These
syntheses provide diastereoselective solutions to each of the
biaryl ether and biaryl macrocycles and define a convergent
assemblage process which can be extended to a variety of
natural and unnatural analogues in the vancomycin series.
Received: August 17, 1998 [Z12295IE]
German version: Angew. Chem. 1998, 110, 2864 ± 2868
2702
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3719-2702 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 19

COMMUNICATIONS
Oallyl
OR
F
O
OH
O
OH
O
F
2
6
4
6
4
2
HO
HO
O
Cl
HO
Cl
HO
O
NO2
O
O
NO2
H
H
H
H
N
N
H
N
H
HO2C
O
N
N
a
N
NH
H
H
NH2
NH
N
H
O
H
H
O
86%
O
N(Me)Boc
O
N(Me)Boc
O
O
NH
NH
O
O
O
5
5
NHDdm
CH2CHMe2
CH2CHMe2
NHDdm
21
22
MeHN
7
MeHN
7
OBn
OBn
OBn
OBn
23
BnO
BnO
Oallyl
Oallyl
X
b
95%
Cl
O
O
O
O
6
4
6
2
4
2
Cl
HO
Cl
HO
OH
OH
O
O
O
O
H
e, f
1
H
H
N
g, h, i
dr = 5:1
H
N
H
N
H
N
vancomycin
aglycon
N
N
H
68%
H
N
H
O
NH
NH
O
N
H
O
H
O
H
O
N(Me)Boc
O
O
N(Me)Boc
O
O
NH
NH
O
O
O
5
5
CH2CHMe2
CH2CHMe2
NHDdm
NHDdm
MeHN
7
HO
OBn
OBn
7
OBn
OBn
24 X = NO2
25 X = Cl
c, d
65%
26
BnO
BnO
Scheme 3. Assemblage of the vancomycin aglycon. a) EDCI, HOAt, THF, 08C; b) CsF, DMSO, room temperature; c) Zn0, HOAc, EtOH, 408C; d) HBF4,
tBuONO, MeCN, then CuCl, CuCl2, H2O; e) N2O4, NaOAc, CH2Cl2/CH3CN, 08C; f) H2O2, LiOH, THF/H2O; g) [Pd(PPh3)4], morpholine, THF; h) 10% Pd/
C, 1,4-cyclohexadiene, EtOH, room temperature; i) TFA, DMS, CH2Cl2, 08C to room temperature. See ref. [6] for abbreviations.
[9] a) D. A. Evans, C. J. Dinsmore, D. A. Evrard, K. M. DeVries, J. Am.
Keywords: antibiotics ´ eremomycin ´ natural products ´
total synthesis ´ vancomycin
Chem. Soc. 1993, 115, 6426 ± 6427; b) D. A. Evans, C. J. Dinsmore,
Tetrahedron Lett. 1993, 34, 6029 ± 6032.
[10] For the synthesis of orienticin C aglycon see: a) D. A. Evans, C. J.
Dinsmore, A. M. Ratz, D. A. Evrard, J. C. Barrow, J. Am. Chem. Soc.
1997, 119, 3417 ± 3418; b) D. A. Evans, J. C. Barrow, P. S. Watson,
A. M. Ratz, C. J. Dinsmore, D. A. Evrard, K. M. DeVries, J. A.
Ellman, S. D. Rychnovsky, J. Lacour, J. Am. Chem. Soc. 1997, 119,
3419 ± 3420.
[1] M. H. McCormick, W. M. Stark, G. E. Pittenger, R. C. Pittenger,
G. M. McGuire, Antibiot. Annu. 1955 ± 56, 606 ± 611.
[2] For a comprehensive discussion of the vancomycin antibiotics see:
Glycopeptide Antibiotics (Ed.: R. Nagarajan), Marcel Dekker, New
York, 1994.
[3] a) D. H. Williams, Natl. Prod. Reports 1996, 469 ± 477; b) M. Foldes, R.
Munro, T. C. Sorrell, S. Shankar, M. Toohey, J. Antimicrob. Chemo-
ther. 1983, 11, 21 ± 26.
[4] a) S. Tabaqchali, Lancet, 1997, 350, 1644 ± 1645; b) R. C. Moellering,
Clinical Infectious Diseases 1998, 26, 1196 ± 1199; c) C. T. Walsh, S. L.
Fisher, I.-S. Park, M. Prahalad, Z. Wu, Chem. Biol. 1996, 3, 21 ± 28, and
references therein.
[11] D. A. Evans, C. J. Dinsmore, P. S. Watson, M. R. Wood, T. I. Richard-
son, B. W. Trotter, J. L. Katz, Angew. Chem. 1998, 111, 2868 ± 2872;
Angew. Chem. Int. Ed. 1998, 37, 2704 ± 2708.
[12] The seven amino acid residues are numbered consecutively starting
from the amino terminus. The M(X ± Y) nomenclature refers to the
macrocycle containing an oxidative cross-link between aryl groups of
residues X and Y. Bicyclic moieties will be identified as M(W± Y)(X ±
Z).
[13] Amino acid 4 was derived from (R)-para-hydroxy-phenylglycine.
[14] Compound 4b was prepared by the chlorination of 4a: F. D. March,
W. B. Farnham, D. J. Sam, B. E. Smart, J. Am. Chem. Soc. 1982, 104,
4680 ± 4682. 4-Fluoro-3-nitrobenzaldehyde (4a) was prepared from 4-
fluorobenzaldehyde: F. Michael, D. Noffz, Chem. Ber. 1957, 90, 1586 ±
1589. Dr. S. Rubenstein is acknowledged for the preparation of this
material.
[15] Alternate conditions and reagents were plagued by competitive
endocyclic oxazolidinone hydrolysis and/or undesired nucleophilic
displacement of the fluoride. D. A. Evans, T. C. Britton, J. A. Ellman,
Tetrahedron Lett. 1987, 28, 6141 ± 6144.
[16] HATU-mediated peptide coupling: L. A. Carpino, A. El-Fahan, C. A.
Minor, F Albericio, J. Chem. Soc. Chem. Commun. 1994, 201 ± 203.
[17] The removal of the TBS protecting group highlighted the significant
reactivity difference between the chloro and des-chloro substrates.
Treatment of 13b with TBAF resulted in immediate M(4 ± 6)
cyclization with reduced diastereoselection. See also ref. [24].
[18] Studies of the cyclization of 14a revealed that the observed
atropselectivity depended significantly on the base used, varying
from 4:1 (KF) to 10:1 (Na2CO3) in favor of the unnatural atropisomer.
It was further observed that cyclization of substrates in which the
meta-phenols of amino acid 4 were both unmasked proceeded with
negligible selectivity under a variety of conditions, necessitating the
synthesis of a fully differentiated triphenol.
[5] For reviews on methodology relevant to the synthesis of the
vancomycin aglycon structure see: a) D. A. Evans, K. D. DeVries in
ref. [2], pp. 63 ± 103; b) A. V. Rama Rao, M. K. Gurjar, K. L. Reddy,
A. S. Rao, Chem. Rev. 1995, 95, 2135 ± 2168; c) J. Zhu, Synlett. 1997,
133 ± 144; d) R. M. Williams, J. A. Hendrix, Chem. Rev. 1992, 92, 889 ±
917.
[6] Abbreviations: dr: diastereomer ratio; Xp: (4S)-4-(phenylmethyl)-2-
oxazolidinone; Tfa: trifluoroacetyl TFA: trifluoroacetic acid; Boc:
tert-butoxycarbonyl; NOE: nuclear Overhauser effect; Tf: trifluoro-
methanesulfonyl; Ms: methanesulfonyl; Bn: benzyl; Piv: pivaloyl;
Ddm: 4,4'dimethoxydiphenylmethyl; EDCI ´ HCl: 1-(3-(dimethylami-
no)propyl)-3-ethylcarbodiimide hydrochloride; HOBt: 1-hydroxy-
benzotriazole; DMAP: 4-dimethylaminopyridine; DMS: dimethyl-
sulfide; TFAA: trifluoroacetic anhydride; HATU: O-(7-azaben-
zotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate; HOAt:
7-aza-1-hydroxybenzotriazole; LDA: lithium diisopropylamide; dppf:
1,1'-bis(diphenylphosphanyl)ferrocene.
[7] a) D. A. Evans, D. A. Evrard, S. D. Rychnovsky, T. Früh, W. G.
Wittingham, K. M. DeVries, Tetrahedron Lett. 1992, 33, 1189 ± 1192;
b) D. A. Evans, T. C. Britton, J. A. Ellman, R. L. Dorow, J. Am. Chem.
Soc. 1990, 112, 4011 ± 4030; c) D. A. Evans, A. E. Weber, J. Am. Chem.
Soc. 1987, 109, 7151 ± 7157.
[8] a) D. A. Evans, J. A. Ellman, K. M. DeVries, J. Am. Chem. Soc. 1989,
111, 8912 ± 8914; b) D. A. Evans, C. J. Dinsmore, A. M. Ratz,Tetrahe-
dron Lett. 1997, 38, 3189 ± 3192; c) D. A. Evans, P. S. Watson,
Tetrahedron Lett. 1996, 37, 3251 ± 3254.
[19] Following diazotization/reduction of aniline 16, eremomycin aglycon
(2) was synthesized by a route analogous to that presented for
Angew. Chem. Int. Ed. 1998, 37, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3719-2703 $ 17.50+.50/0
2703

COMMUNICATIONS
vancomycin. Synthetic 2 gave 1H NMR, HPLC, and mass spectral data
identical to a natural comparison sample (refs. [23] and [27]).
[20] We thank Mr. Kevin Campos for determining the X-ray structure of
this intermediate.
[21] H. J. Shine, S. E. Rhee, J. Am. Chem. Soc. 1986, 108, 1000 ± 1006.
[22] This assignment was confirmed by chemical correlation to a minor
atropisomer corresponding to 16.
M(4-6)
M(2-4)
OH
Y
O
O
6
4
2
X
HO
OH
O
O
O
H
N
H
N
H
N
NHMe
H
N
H
N
H
H
O
[23] S. Cacci, P. G. Ciattini, E. Morera, G. Ortar, Tetrahedron Lett. 1986, 27,
5541 ± 5544.
[24] Separation of the undesired M(4 ± 6) atropisomer was most conven-
iently achieved at this stage; the desired isomer was isolated in 69%
yield.
O
O
O
CH2CH(CH3)2
NH
5
M(5-7)
HO2C
NH2
7
OH
OH
HO
vancomycin aglycon
eremomycin aglycon
orienticin C aglycon
1a X = Cl, Y = Cl
1b X = H, Y = Cl
1c X = H, Y = H
[25] The stereochemistry of 20 was confirmed by ROESY analysis
(500 MHz).
[26] D. L. Boger, R. M. Borzilleri, S. Nukui, R. T. Beresis, J. Org. Chem.
1997, 62, 4721 ± 4736.
[27] D. A. Evans, P. H. Carter, C. J. Dinsmore, J. C. Barrow, J. L. Katz,
D. W. Kung, Tetrahedron Lett. 1997, 38, 4535 ± 4538.
[28] J. Garcia, J. Gonzalez, R. Segura, J. Vilarrasa, Tetrahedron 1984, 40,
3121 ± 3127.
[29] A. M. Felix, E. P. Heimer, T. J. Lambros, C. Tzougraki, J. Meienhofer,
J. Org. Chem. 1978, 43, 4194 ± 4196.
[30] We thank Dr. Bret Huff of Eli Lilly & Company for generous samples
of vancomycin ´ HCl and eremomcyin ´ HOAc.
[31] Vancomycin aglycon was prepared from vancomycin ´ HCl according
to: R. Nagarajan, A. A. Schabel, J. Chem. Soc. Chem. Comm. 1988,
1306 ± 1307; see also F. J. Marshall, J. Med. Chem. 1965, 8, 18 ± 24.
[32] Eremomycin aglycon was prepared from eremomycin ´ HOAc accord-
ing to: G. F. Gause, M. G. Brazhnikova, N. N. Lomakina, T. F.
Berdnikova, G. B. Fedorova, N. L. Tokareva, V. N. Borisova, G. Y.
Batta, J. Antibiotics 1989, 42, 1790 ± 1799.
Figure 1. Space-filling representation of the vancomycin aglycon.[3b]
The high barrier to rotation about the biaryl bond[4] in
vancomycin introduces an element of axial chirality into the
structure, while hindered rotation about the axes defined by
the para-oriented CH(OH) and O-aryl substituents in ring 2
and ring 6 incorporates two examples of planar chirality.
Collectively, these three features of the aglycon architecture
present the significant challenge of controlling atropisomer-
ism in the construction of each of the three macrocyclic
tripeptide subunits designated as M(2 ± 4), M(4 ± 6), and
M(5 ± 7).[5] Hence, even with asymmetric syntheses of the
amino acid constituents[6] and an assemblage strategy in
hand,[7] one is still faced with the problem of producing the
vancomycin aglycon skeleton as only one of eight possible
atropdiastereomers.
The vancomycin aglycon skeleton (1a) consists of three
interlocking cyclic tripeptides that collectively afford a con-
formationally rigid cup-shaped structure (Figure 1).[3] It is
evident that the biaryl bond connecting amino acids 5 and 7 is
the pivotal rigidifying amino acid crosslink. We therefore
adopted the premise that macrocyclization model studies for
the individual rings lacking the M(5 ± 7) tripeptide subunit,
while informative in identifying the local contributions to
atropdiastereoselection, could prove unreliable stereochem-
ical predictors for more complex cyclization substrates con-
taining the M(5 ± 7) fragment.
Nonconventional Stereochemical Issues in the
Design of the Synthesis of the Vancomycin
Antibiotics: Challenges Imposed by Axial and
Nonplanar Chiral Elements in the
Heptapeptide Aglycons**
David A. Evans,* Christopher J. Dinsmore,
Paul S. Watson, Michael R. Wood,
Timothy I. Richardson, B. Wesley Trotter, and
Jeffrey L. Katz
In the preceding communication, we described the first
syntheses of the heptapeptide aglycons of vancomycin (1a)
and eremomycin (1b).[1] This contribution focuses on the
development of stereoselective methods for the synthesis of
the three stereochemical elements of atropisomerism present
in vancomycin.[2] The development of a strategy for control-
ling these architectural features is one of the principal
challenges presented by this family of natural products. As
an aid in the ensuing discussion, a space-filling representation
of the vancomycin aglycon, taken from the X-ray structure by
Sheldrick et al.,[3] is provided (Figure 1).
[*] Prof. D. A. Evans, C. J. Dinsmore, P. S. Watson, M. R. Wood,
T. I. Richardson, B. W. Trotter, J. L. Katz
Department of Chemistry & Chemical Biology, Harvard University
Cambridge, MA 02138 (USA)
The construction of the M(5 ± 7) biaryl subunit[8] forms the
basis of the synthesis plan. After the incorporation of an
additional ortho benzyloxy substituent on ring 5, high levels of
kinetic atropdiastereoselection for the unnatural atropiso-
meric product 3(R) were observed in the oxidative cyclization
Fax: (‡1)617-495-1460
E-mail: evans@chemistry.harvard.edu
[**] Financial support has been provided by the National Institutes of
Health (NIH).
2704
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3719-2704 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 19