Structural modifications of the active site in teicoplanin and related glycopeptides. 1. Reductive hydrolysis of the 1,2- and 2,3-peptide bonds

Article abstract of DOI:10.1021/jo941809v

Reaction of teicoplanin glycopeptides with sodium borohydride in aqueous ethanol solutions produced open pentapeptide derivatives in which the amide bond between amino acids 2 and 3 was hydrolyzed and the carboxyl group of amino acid 2 was reduced to a primary alcohol. Other glycopeptides of the dalbaheptide family, such as vancomycin, ristocetin, and A-40,926, underwent selective reductive hydrolysis (RH) of the heptapeptide backbone at the same position as in teicoplanins, while antibiotic A-42,867 and vancomycin hexapeptide were resistant. Also, teicoplanin and vancomycin were resistant to RH-treatment when the N-terminus was protected as carbamate. In contrast, open hexapeptides in which the 1,2-peptide bond was hydrolyzed and the carboxyl group of amino acid 1 was reduced to hydroxymethyl were obtained from carbamate derivatives of sugar-free compounds deglucoteicoplanin (TD) and vancomycin-aglycon (VA) under RH-conditions. Limited to BOC or CBZ-TD, the 3,4-amide bond was also affected. A possible RH-mechanism is proposed for natural glycopeptides and their derivatives. Teicoplanin-derived RH penta- and hexapeptides maintained residual antibacterial activity. As other analogous RH-glycopeptides, they are key intermediates for the synthesis of new members of this family of antibiotics. A synthetic approach to ring-closed derivatives of TD hexapeptide alcohol (TDHPA) and their activities are also reported.

Full text of DOI:10.1021/jo941809v

J . Org. Chem. 1996, 61, 2137-2150
2137
Str u ctu r a l Mod ifica tion s of th e Active Site in Teicop la n in a n d
Rela ted Glycop ep tid es. 1. Red u ctive Hyd r olysis of th e 1,2- a n d
2,3-P ep tid e Bon d s
Adriano Malabarba,* Romeo Ciabatti, J u¨rgen Kettenring, Pietro Ferrari, Ka´roly Ve´key,†
Elvio Bellasio, and Maurizio Denaro
Marion Merrell Dow Research Institute, Lepetit Center, Via R. Lepetit 34, 21040 Gerenzano (Varese), Italy
Received October 27, 1994 (Revised Manuscript Received J anuary 4, 1996X)
Reaction of teicoplanin glycopeptides with sodium borohydride in aqueous ethanol solutions produced
open pentapeptide derivatives in which the amide bond between amino acids 2 and 3 was hydrolyzed
and the carboxyl group of amino acid 2 was reduced to a primary alcohol. Other glycopeptides of
the dalbaheptide family, such as vancomycin, ristocetin, and A-40,926, underwent selective reductive
hydrolysis (RH) of the heptapeptide backbone at the same position as in teicoplanins, while antibiotic
A-42,867 and vancomycin hexapeptide were resistant. Also, teicoplanin and vancomycin were
resistant to RH-treatment when the N-terminus was protected as carbamate. In contrast, open
hexapeptides in which the 1,2-peptide bond was hydrolyzed and the carboxyl group of amino acid
1 was reduced to hydroxymethyl were obtained from carbamate derivatives of sugar-free compounds
deglucoteicoplanin (TD) and vancomycin-aglycon (VA) under RH-conditions. Limited to BOC or
CBZ-TD, the 3,4-amide bond was also affected. A possible RH-mechanism is proposed for natural
glycopeptides and their derivatives. Teicoplanin-derived RH penta- and hexapeptides maintained
residual antibacterial activity. As other analogous RH-glycopeptides, they are key intermediates
for the synthesis of new members of this family of antibiotics. A synthetic approach to ring-closed
derivatives of TD hexapeptide alcohol (TDHP A) and their activities are also reported.
A current challenge for glycopeptides of the dalbahep-
tide group,1 to which teicoplanin (CTA, Figure 1)2
belongs, is the emerging resistance in VanA enterococci3
due to the replacement of antibiotic’s target peptide
D-Ala-D-Ala by a D-Ala-D-hydroxy acid depsipeptide (Chart
1)4 which is not sufficiently bound by glycopeptides to
prevent the bacterial growth. Considering that the
binding properties of these antibiotics in part depend on
the structure of amino acids 1 and 3,5 activity against
glycopeptide-resistant enterococci could be pursued by
replacement of these amino acids with new amino acids
or other moieties suitably selected to interact with the
modified target.
Chemical6 and biological7 methods for the removal of
amino acid 1 of vancomycin (V, Figure 2) have been
reported, but these procedures could not be applied to
V8 or resulting hexapeptide (VHP ) for the removal of
amino acid 3. In teicoplanin, the N-terminal amino acid
† Central Research Institute for Chemistry, Hungarian Academy of
Science, H-1025 Budapest, II. Pusztaszeri u`t 59-67, Hungary.
X Abstract published in Advance ACS Abstracts, March 1, 1996.
(1) Parenti, F.; Cavalleri, B. Drugs Future 1990, 15, 57.
(2) Coronelli, C.; Gallo, G. G.; Cavalleri, B. Il Farmaco, Ed. Sci. 1987,
10, 767.
(3) J ohnson, A. P.; Uttley, A. H. C.; Woodford, N.; George, R. C. Clin.
Microbiol. Rev. 1990, 3, 280.
(4) Bugg, T. D. H.; Walsh, C. T. Nat. Prod. Rep. 1992, 9, 199.
F igu r e 1. Structure of teicoplanin-like dalbaheptides (W )
-NHCO-) and their 2,3-RH-derivatives (W ) -NH2, HOCH2-).
(5) Naturally occurring glycopeptides are highly modified linear
peptides made up of seven amino acids, five of which are aryl amino
acids and are common to all members of the group. The differentiation
of glycopeptides into four main classes is due to the remaining two
amino acids in positions 1 and 3 which can be both aromatic (free or
linked together through a diphenyl ether bridge) or one aliphatic and
one aromatic.
(6) (a) Booth, P. M.; Stone, D. M.; Williams, D. H. J . Chem. Soc.,
Chem. Commun. 1987, 1694. (b) Nagarajan, R.; Schabel, A. A. J . Chem.
Soc., Chem. Commun. 1988, 1306. (c) Booth, P.; Williams, D. H. J .
Chem. Soc., Perkin Trans. 1 1989, 2335.
(7) Zmijewski, M. J .; Logan, R. M.; Marconi, G.; Debono, M.; Molloy,
R. M.; Chadwell, F.; Briggs, B. J . Nat. Prod. 1989, 52, 203.
(8) The probability to enzymatically remove amino acid 3 is actually
negligible considering its positioning inside the binding pocket hardly
approachable by enzymes.
also cannot be removed by Edman degradation9 since
amino acids 1 and 3 are linked together through a
diphenyl ether bridge. Therefore, a chemoselective pro-
cess for the displacement of these amino acids following
Edman degradation implies preliminary hydrolysis of the
peptide bond between amino acids 2 and 3.10 The
reductive hydrolysis (RH) of the 2,3-amide in CTA, its
(9) Edman, P. Acta Chem. Scand. 1950, 4, 277.
(10) Malabarba, A.; Ciabatti, R. J . Med. Chem. 1994, 37, 2988.
0022-3263/96/1961-2137$12.00/0 © 1996 American Chemical Society

2138 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
Ch a r t 1. Teicop la n in - (DiAc-Lys)-D-Ala -D-Ala
Com p lex
F igu r e 2. Structure of vancomycin-like dalbaheptides (W )
-NHCO-), their hexapeptide derivatives (W ) -NHCO-, Z
) H) and the corresponding 2,3-RH-products (W ) -NH2,
HOCH2-).
respect to V. As reported below, only A-42 and VHP
were resistant.
On the basis of these findings, a possible RH-mecha-
nism was thought (Scheme 1A) which is strictly related
to the particular conformation of the heptapeptide chain
of “natural” glycopeptides. In this case, the opening of
the 2,3-amide would be favored by the formation, under
weakly basic conditions, of a series of hydrogen-bonding
systems whose direction is driven by the presence of the
N-terminus as free base. Accordingly, amide bonds in
enolic form are not susceptible to reductive hydrolysis.
A similar mechanism was expected when the N-terminal
amino group is protected as carbamate, but with different
orientation of the hydrogen bonds (Scheme 1C,D) so as
to allow the reduction of one or both of the 1,2- and 3,4-
amides. Results achieved with carbamate derivatives of
TD and VA upon treatment with NaBH4 in different
EtOH/H2O solutions are discussed. Most experiments
have been done on BOC and CBZ-TD (Figure 3) and
pseudoaglycons (TB, TC), and aglycon (TD) upon treat-
ment of aqueous ethanol solutions of teicoplanin glyco-
peptides with NaBH4 is described in this paper. Under
these conditions, open pentapeptide derivatives (RH-
teicop la n in s, Figure 1) were obtained in which the
carboxyl group of amino acid 2 was reduced to hydroxy-
methyl.
To assess whether this RH-method was generally
applicable to natural dalbaheptides and their derivatives,
six compounds were properly selected and treated with
NaBH4 under conditions similar to those used for the
preparation of RH-teicoplanins from teicoplanins,
namely: glycopeptides ristocetin (R)11 and A-40,926 (A-
40)12 (Figure 1), which are structurally related to teico-
planin but significantly differ from CTA in the sugar
content and distribution, V and its aglycon (VA),6 in
which amino acids 1 and 3 are aliphatic, and one vanco-
mycin-like glycopeptide A-42,867 (A-42)13 and vancomy-
cin-derived hexapeptide (VHP )6,7 (Figure 2), which show
conformational changes14 in their peptide backbone with
(14) A significant difference between A-42 and V is the opposite
orientation of the 3,4 and 4,5-amide bonds. In V, as in the majority of
the other dalbaheptides, the amide-COs of residues 2 and 3 are both
pointing backward on the rear face of the molecule and are close to
each other, while the amide-CO of residue 4 points forward on the front
face. In A-42, the 4,5-amide is stabilized in the enolic form by a
hydrogen bond with the amino group of the sugar vancosamine on
amino acid 6. The OH resulting from this tautomerism is on the back
face of the molecule. It follows a change in the orientation of the 3,4-
amide whose CO is on the front face, while the amide-CO of residue 2
still lies on the back face and so its oxygen results far from that of the
amide-CO of residue 3. In VHP , the absence of the N-terminal
methylleucine causes (at basic pH) a change in conformation of the
N-terminal region in which the 2,3-trans amide unit exhibits a 180°
flipping motion, from having the NH-proton at the front of the
molecule, as in V, to having it at the rear. This results in the opposite
orientation of the amide-COs of residues 2 and 3.
(11) Kalman, J . R.; Williams, D. H. J . Am. Chem. Soc. 1980, 102,
897.
(12) (a) Goldstein, B. P.; Selva, E.; Gastaldo, L.; Berti, M.; Pallanza,
R.; Ripamonti, F.; Ferrari, P.; Denaro, M.; Arioli, V.; Cassani, G.
Antimicrob. Agents Chemother. 1987, 31, 1961. (b) Selva, E.; Gastaldo,
L.; Beretta, G.; Borghi, A.; Goldstein, B. P.; Arioli, V.; Cassani, G.;
Parenti, F. U.S. Patent No. 4,935,238 (J une 19, 1990).
(13) Riva, E.; Gastaldo, L.; Beretta, M. G.; Ferrari, P.; Zerilli, L. F.;
Cassani, G.; Selva, E.; Goldstein, B. P.; Berti, M.; Parenti, F.; Denaro,
M. J . Antibiot. 1989, 42, 497.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2139
Sch em e 1. P r op osed RH-Mech a n ism
Sch em e 2. Syn th esis of RH-Glycop ep tid e
Der iva tives
hexapeptide (VAHP , Figure 2) with higher yields with
respect to Edman degradation.6 The N-acetyl derivative
(Ac-VAHP ) of VAHP was also treated with NaBH4
under the above conditions to achieve additional infor-
mation on the RH-mechanism.
Hexapeptides VAHP and TDHP A are useful interme-
diates for the preparation of new structurally different
dalbaheptides. In particular, from VAHP there is the
possibility to obtain a number of VA-derived glycopep-
tides by direct condensation of the terminal amino group
with a series of different amino acids. Also, the avail-
ability of BOC or CBZ-TDHP A, more than the unpro-
tected TDHP A, provides the opportunity for the direct
synthesis of new families of related dalbaheptides with
a common diphenyl ether-amino acid fragment 316 and
different amino acid 1-residues. In fact, in TDHP A
carbamates the only remaining free amino group is
suitable to this goal. They are also useful intermediates
to obtain TD derivatives with an enlarged 1,2,3-macro-
cyclic ring. To possibly establish further structure-
activity relationships, two ring-closed compounds (RC-1
BOC-VA (Figure 2), since carbamate derivatives of
parent sugar-containing glycopeptides CTA and V proved
to be stable under all reaction conditions.15 The aim of
this work was to assess the actual possibility to selec-
tively reduce the 1,2-amide in sugar-free glycopeptide
carbamates so as to provide a suitable synthetic proce-
dure to obtain TD-derived hexapeptide alcohol (TDHP A,
Figure 3) and to propose a new method to produce VA
(15) The resistance of CTA and V carbamates to reductive hydrolysis
might be ascribed to a possible steric effect of their sugar moieties on
aromatic ring 4 on the formation and structure of hydrogen bonding/
boron complex systems. This might favor the reduction of the 2,3-amide
in CTA and V, while preventing the reduction of the 1,2-peptide bond
in these compounds and their carbamate derivatives.
(16) In new semisynthetic dalbaheptides derived from TDHP A,
amino acid 3 is a double amino acid residue formed by the diphenyl
ether moiety originally belonging to amino acids 1 and 3 of TD.

2140 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
at C-3,18 which generally occurs in teicoplanins under
prolonged basic treatment, was only observed with TD.
The different conversion rate (CTA > TB > TC . TD)
seems to play a critical role in the epimerization of
teicoplanins. Epimers, once formed, were not susceptible
to reductive hydrolysis upon treatment with NaBH4. This
is likely the reason why RH-TD was obtained with lower
yields than the other RH-teicop la n in s. Also, the pres-
ence of the sugar moieties seems to favor this transfor-
mation while preventing epimerization. In order to
increase yields in RH-TD while decreasing epimeriza-
tion, new reaction conditions were studied.19 It was
found that the formation of the C3-epimer (epi-TD)
increased with increasing reaction time which was
inversely related to the concentration of NaBH4, while
the conversion rate of TD into RH-TD was somewhat
proportional to the concentration of NaBH4. It was also
assessed that the concentration of water in the aqueous
alcohol mixtures was critical, since the transformation
yields of TD into RH-TD were lower when the EtOH/
H2O ratio was close to or higher than 1. In particular,
when a large excess of NaBH4 was used in a EtOH/H2O
6/4 solution (method a′), deglucoteicoplanin hexapeptide
alcohol (TDHP A) formed as byproduct. Yields in TDH-
P A were increased using a EtOH/H2O 9/1 mixture
(method a′′). Under these conditions, a small (∼5%)
amount of C3-epimer (epi-TDHP A) was also obtained.
Among teicoplanin dalbaheptides, only TD was suscep-
tible to reduction of the 1,2-peptide bond. Pseudoagly-
cons RH-TB, RH-TC, and aglycon RH-TD were also
prepared from RH-CTA by hydrolysis of the sugars
under controlled acidic conditions (methods b, c, and d).
F igu r e 3. Structure of (R ) H) TD (Y ) W ) Z ) -NHCO-),
RH-TD (W ) -NH2, HOCH2-, Y ) Z ) -NHCO-), TDHP A
(Y ) W ) -NHCO-, Z ) -NH2, HOCH2-), and TDTP DA (Y
) Z ) -NH2, HOCH2-, W ) -NHCO-) and their BOC (R )
CO2CMe3) and CBZ (R ) CO2CH2Ph) carbamate derivatives
(with nomenclature of aromatic, benzylic, and HC-R protons.
Following method a, analogous RH-derivatives (Table
2) were obtained from vancomycin (V) and its aglycon
(VA), ristocetin (R), and A-40,926 (A-40). Under these
conditions, the C-terminal methyl ester of R was previ-
ously reduced to hydroxymethyl, so that the final com-
pound corresponded to the RH-derivative (RH-HR) of
decarboxyhydroxymethylristocetin (HR ).20 Unexpect-
edly, A-42,867 (A-42) and V-derived hexapeptide (VHP )
were resistant to reduction with NaBH4. These results
indicated that only glycopeptides which possess the
“natural” dalbaheptide conformation14 are susceptible to
reductive hydrolysis.
The reduction of an amide with NaBH4 to give an
amino alcohol is a rather uncommon reaction which
might occur in dalbaheptides according to a specific
mechanism of activation of the 2,3-peptide bond due to
the particular conformation of their heptapeptide chain.
The following RH-mechanism is supported by a possible
analogy with that proposed by Micovic and Mihailovic21
for the reduction of amides to alcohols with lithium
aluminum hydride.
F igu r e 4. Structure and formation of TDHP O.
and R C-2, Figure 5) were synthesized in which the
original 1,2,3-macrocyclic ring of TD was enlarged by a
spacer of 1 and 2 methylene groups, respectively.
Resu lts a n d Discu ssion
P r op osed RH-Mech a n ism for Glycop ep tid es a n d
Th eir Der iva tives. The conformation of the heptapep-
tide chain in most of glycopeptides is such as to allow
the formation of a series of hydrogen bonding systems
Red u ctive Hyd r olysis of th e 2,3-P ep tid e Bon d .
Treatment of teicoplanin antibiotics CTA,17 TB, TC, and
TD (Figure 1) at room temperature with a large excess
of NaBH4 in a EtOH/H2O 35/65 solution (Scheme 2)
afforded corresponding pentapeptide amino alcohol de-
rivatives RH-CTA, RH-TB, RH-TC, and RH-TD with
relatively good yields (method a, Table 1). Epimerization
(18) Barna, J . C. J .; Williams, D. H.; Strazzolini, P.; Malabarba, A.;
Leung, T.-W. C. J . Antibiot. 1984, 37, 1204.
(19) Different concentrations of NaBH4 and H2O in various aqueous
alcohol mixtures were investigated. The detailed experimental results
are not reported.
(17) Teicoplanin (CTA, Figure 1) is a complex of five strictly related
components (namely, factors A2-1, 2, 3, 4, 5) which only differ in the
structure of the acyl chain of the N-acylglucosamine unit on aromatic
residue 4. The acyl COR moieties are: R ) (CH2)2CHdCH(CH2)4CH3
(A2-1); R ) (CH2)6CH(CH3)2 (A2-2); R ) n-C9H19 (A2-3); R ) (CH2)6-
CH(CH3)CH2CH3 (A2-4); R ) (CH2)7CH(CH3)2 (A2-5).
(20) The selective reduction of the C-terminal methyl ester of R to
give HR occurred before starting the reductive hydrolysis of the 2,3-
amide bond. Upon treatment of R with NaBH4 under the same above
RH conditions (method a) the conversion of R into HR was completed
within 90 min.
(21) Micovic, V. M.; Mihailovic, M. L. J . Org. Chem. 1953, 18, 1190.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2141
Ta ble 1. RH-Teicop la n in Der iva tives
titration
yield,
% (method)
HPLCa
tR, min
FAB MS
compd
pKa-1
pKa-2
pKa-3
EWb
(M + H)+
formula
MW
RH-CTAc
RH-TB
RH-TC
RH-TD
82 (a)d
15.3
10.2
10.9
11.6
4.7
4.7
4.6
4.5
5.9
5.9
5.8
5.5
7.3
7.3
7.5
7.4
637
549
477
425
1882
1567
1405
1202
C88H101N9Cl2O33
C72H72N8Cl2O28
C66H62N8Cl2O23
C58H49N7Cl2O18
1883.7
1568.3
1406.2
1203.0
75 (a),d 75e (b)
62 (a),d 55e (c)
47 (a),f 20e (d)
a
HPLC (method A). Relative (vs. parent unmodified teicoplanin) tR’s: 0.96 (RH-CTA), 0.98 (RH-TB), 0.93 (RH-TC), 0.92 (RH-TD).
Equivalent weight. The values given are corrected for solvent content (TG). c Data are referred to factor A2-2. Molar ratio NaBH4/
b
d
substrate: 75; reaction time: 2 h (3 h for RH-TC). e From RH-CTA. f Molar ratio NaBH4/substrate: 250; reaction time: 16 h.
Ta ble 2. Oth er RH-Glycop ep tid e Der iva tives
reactiona
time, h yield, % tR, min
FAB MSb (M + H)+
molar ratio (NaBH4
vs. substrate)
HPLCd titration
1H NMRc (CH2OH)
compd
EWe
found
calcd
δ, ppm
formula
MW
RH-A-40
RH-V
RH-VA
RH-HR
80
200
250
120
24
48
36
16
68
61
44
59
20.2
7.6
12.8
8.5
470
430
390
750
1735.5
1452.5
1146.2
2043.6
1735.5
1452.5
1146.3
2043.7
3.27
3.28
3.35
3.38f
C83H92N8Cl2O29 1736.6
C
C
66H79N9Cl2O24 1453.3
53H56N8Cl2O17 1148.0
C94H114N8O43
2044.0
(HR
50
3
86
8.8
1090
2039.6
2039.7
f
C
94H110N8O43
2040.0)
a
b
Method a. Mass numbers refer to the lowest mass isotope of a cluster. c Data referred to the new methylene protons of residue 2.
d
HPLC (method A). Relative (vs. parent unmodified glycopeptide) tR’s: 0.91 (RH-A-40), 0.79 (RH-V), 0.73 (RH-VA), 0.87 (RH-HR), 0.86
(HR). e Equivalent weight. The values given are corrected for solvent content (TG). f The new methylene protons of the C-terminal CH2OH
of RH-HR and HR resonate at δ 3.51 and 3.50 ppm, respectively.
in the right-hand side of these molecules that involve the
OHs of the amides between residues 1,2, and 3,4, both
postulated in the enolic form to be unsusceptible to
reductive hydrolysis, and the N-terminus of residue 1 and
the CO of the 2,3-amide, respectively (Scheme 1A). In
these systems, which are favored by the weak basic pH
of the reaction medium, the direction of hydrogen bonds
is driven as depicted in structure I by the presence of
the N-terminal amino group as free base. As a conse-
quence, the formation of a hydrogen bond between the
NH of the 2,3-amide and the adjacent -Nd atom of the
vinylog form of the 1,2-amide is hypothesized to stabilize
the -NHCO- structure of the 2,3-amide. The formation
of a seven-membered hydrogen-bonding system involving
the OH of the 3,4-amide, in the enolic form, and the
amide-CO of residue 2, is allowed by the proximity of both
of these carbonyl groups on the rear face of the dalba-
heptide molecule. An initial interaction between the
borohydride and the enolic-OH of residue 3 in structure
I would give adduct II, followed by the formation of
partially reduced cyclic boron complex III, as the result
of an unusual intramolecular nucleophilic addition of an
hydride ion to the CO-carbon atom of residue 2 and the
formation of a boron-oxygen bond involving the carbo-
nyl-oxygen of residue 2 and the boron atom of adduct
II. A further nucleophilic attack by water at the C-atom
of the carbinol-boron complex (III), with the simulta-
neous displacement of the borate anion, gives carbino-
lamine IV. This hemiaminal is in equilibrium with the
amino aldehyde form V which is readily reduced to the
final amino alcohol (RH) by the excess borohydride.
The resistance of A-42 and VHP to reductive hydroly-
sis demonstrates the need for the “natural” conformation
of the heptapeptide chain to promote the RH-mechanism
in glycopeptides. In this sense, a critical role is played
by the N-terminal amino acid 1. In fact, in the absence
of this amino acid, as in VHP , a different hydrogen
bonding system (Scheme 1B) is hypothesized due to a
change in the orientation of the 2,3-amide-CO at basic
pH. In glycopeptide-derived hexapeptides, the formation
of a hydrogen bond between the N-atom of the terminal
amino group and the enolic-OH of the adjacent amide
would be favored by the presence of the unshared electron
pair on the nitrogen of free NH2. A further support is
given by the resistance of teicoplanin epimers to reductive
hydrolysis. With respect to the corresponding unmodified
antibiotics, these compounds show opposite configuration
at C-3.18 As a consequence, the amide-CO of residue 3
points forward on the front face while the amide-CO of
residue 2 is still pointing backward.
In glycopeptide antibiotics, the possible existance of the
1,2- and 3,4-peptide linkages in the enolic form (Scheme
1A) at basic pH could explain the different binding
strength to synthetic models of the antibiotic’s target
peptide at pH 5 and 9.22 In fact, the values of the
association constants are generally about 10 times higher
when determined at pH 5 with respect to those found at
pH 9. Given that the NH-protons of residues 2 and 4,
besides that of residue 3, are directly involved in the
hydrogen bonding interaction with the carboxylate anion
of the synthetic dipeptide model, the contribution of the
NH-protons of amides 1,2 and 3,4 to the complex forma-
tion (Chart 1) would be lower at pH 9 for the presence of
these amides mainly in the vinylog form.
As mentioned above, the driving force which possibly
determines the direction of hydrogen bonds in “natural”
dalbaheptides is the basic character of the N-terminal
amino or methylamino group. It follows that different
situations are expected when amino acid 1 is present but
the N-terminus is acylated or protected as carbamate.
Accordingly, in carbamate derivatives of glycopeptides
possessing a terminal primary amino group, as TD-
carbamates (Scheme 1C), the carbonyl oxygen of the 1,2-
amide should act as the base and the carbamate-NH as
the acid in the hydrogen bonding formation. In carbam-
ate derivatives of vancomycin-like dalbaheptides, as VA-
carbamates, because of the substitution of the carbamate
NH-proton by a methyl group, a further different hydro-
gen bonding system (Scheme 1D) was hypothesized. This
allowed us to predict the reductive hydrolysis of the 1,2-
and 3,4-amides in BOC or CBZ-TD and of the 1,2-amide
in BOC-VA.
(22) (a) Barna, J . C. J .; Williams, D. H.; Williamson, M. P. J . Chem.
Soc., Chem. Commun. 1985, 254. (b) Malabarba, A.; Trani, A.; Tarzia,
G.; Ferrari, P.; Pallanza, R.; Berti, M. J . Med. Chem. 1989, 32, 783.

2142 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
Sch em e 3. Red u ctive Hyd r olysis of
TD-Ca r ba m a tesa a n d Syn th esis of TDHP A a n d
TDHP O
In contrast to BOC-TD, BOC-VA was completely
transformed into hexapeptide VAHP even under condi-
tions (EtOH/H2O 2/8) favorable to the opening of the 2,3-
peptide bond (Scheme 4, Table 4). As expected, the
selectivity of reduction of the 1,2-amide was markedly
higher in BOC-VA than in BOC-TD.
Acetylation of VAHP at the N-terminus with MeCOCl/
TEA (DMF) yielded Ac-VAHP which was treated with
NaBH4 in EtOH/H2O 3/7 solution. A 6/4 mixture of
VAHP and the N-acetyl derivative Ac-RH-VAHP was
obtained (Scheme 4, Table 4). This result further sup-
ported the overall RH-mechanism proposed for glycopep-
tides and their derivatives. Accordingly, in the case of
N-acylated glycopeptide-derived hexapeptides, both the
1,2-mimicking amide and the 2,3-peptide bonds are
susceptible to reduction (Scheme 5).
a
Reaction conditions and yields in Table 3.
Red u ctive Hyd r olysis of th e 1,2- a n d 3,4-P ep tid e
Bon d s. For TD-carbamates, the experiments were run
using N-tert-butyl (BOC-TD) and benzyl (CBZ-TD)
oxycarbonyl derivatives of TD, which were treated with
NaBH4 (Scheme 3) in various aqueous alcohol media
containing decreasing percentages (from 80 to 10%) of
water. The amount of the reductive agent was also
increased to reduce the reaction time while minimizing
epimerization at C-3. The results (Table 3) indicated that
all of the 1,2-, the 2,3-, and the 3,4-amide bonds were
susceptible to reduction, though to a different extent
depending on the relative percentage of water in the
solvent mixture. The opening of the 1,2-peptide linkage
to give BOC- or CBZ-TDHP A (Figure 3) was favored
with decreasing percentages of water, while BOC- or
CBZ-RH-TD were the main products when the relative
amount of water was higher than 40%. The highest RH-
selectivity (60%) of the 1,2-amide was achieved (from
BOC-TD) in a EtOH/H2O 9/1 mixture. The reduction of
both the 1,2- and 3,4-amides was also observed (5-10%).
Resulting carbamate (BOC or CBZ) derivatives of de-
glucoteicoplanin tetrapeptide dialcohol TDTP DA (Figure
3) were obtained as byproducts under all reaction condi-
tions. In this case, the opening of the 1,2-peptide bond
likely occurred simultaneously with the opening of the
3,4-amide, or at faster rate after the 3,4-cleavage, since
there was no chromatographic (HPLC) evidence of the
formation of an intermediate RH-product at the 3,4-
position. On the other hand, the double RH-products
cannot be derived from the TDHP A carbamates since
these latter compounds were stable to RH-treatment
under all described conditions. When i-PrOH instead of
EtOH was used,23 similar results were observed as above,
but relatively longer (∼30%) times were necessary to
complete the reactions. Water was always necessary
since no transformation occurred in absolute MeOH,
EtOH, or i-PrOH. Between the two carbamates selected
for this investigation,24 BOC-TD was more suitable for
the preparation of TDHP A (via intermediate BOC-
TDHP A).25 In fact, BOC-TDHP A was stable once
formed while CBZ-TDHP A underwent a further, though
slow, transformation into oxazolidinone derivative TDH-
P O (Figure 4), as the result of an intramolecular nucleo-
philic cyclization involving the newly formed primary
hydroxyl group of residue 1 and adjacent CBZ-carbonyl
group.
Some physicochemical and analytical data of RH-
products obtained from carbamate and acetyl derivatives
are given in Table 5.
TDHP A-Der ived Da lba h ep tid es. The enlargement
of the 1,2,3-macrocyclic ring of TD by one and two
methylene groups was achieved by acylation of the free-
terminal amino group of the hexapeptide chain of BOC-
TDHP A with bromoacetyl and 3-bromopropionyl chlo-
ride, respectively, in Me2CO/H2O 1/1 solution in the
presence of NaHCO3 (Scheme 6). The BOC protective
group was then removed from the resulting bromoacetyl
(BOC-Ac-1) and bromopropionyl (BOC-Ac-2) derivatives
with TFA at room temperature. Final cyclization to give
RC-1 and RC-2 (Figure 5) was carried out upon treat-
ment of intermediates Ac-1 and Ac-2, respectively, with
K2CO3 at room temperature in DMF solution. Reaction
yields are given in Table 6.
Str u ctu r e Elu cid a tion . Most of the NMR work was
carried out on RH-TD, TDHP A, and epi-TDHP A. The
structures of the other RH-derivatives were determined
1
by comparison of their H NMR spectra with those of the
above compounds and corresponding unmodified glyco-
peptides. The MWs of all derivatives were confirmed by
mass spectrometry. The FAB MS spectra of RH-teico-
p la n in s were particulary studied.
NMR Str u ctu r a l Stu d ies of RH-TD. A first inspec-
1
tion of its H NMR spectrum compared to that of TD26
shows two main differences: (i) there is a strong upfield
shift of proton x2,27 and (ii) one of the two aromatic high-
field protons does not belong to aromatic ring 4. In a
resolution-enhanced spectrum, the signal at 5.83 ppm
shows triplet multiplicity, indicating meta-coupling to
two aromatic protons, whereas the aromatic protons of
ring 4 appear as small doublets, due to the substitution
pattern of this system. The through-bond connectivities
for the heptapeptide core were easily determined in a
DQF-COSY experiment,28 giving the NH,HCR connec-
tivities. The proton assigned as x2 shows cross-peaks to
two different aliphatic CH2 groups, namely the protons
z2, z2′, and the newly introduced CH2 moiety deriving
from the reductive cleavage of the 2,3-amide bond. The
13C chemical shift of the new CH2 at 61 ppm, determined
in an inverse H,C correlation, indicates the substitution
by a hydroxyl group. The assignment for the aromatic
(23) The use of MeOH instead of EtOH was unsuitable for these
studies due to the faster decomposition of the reducing agent, the lower
selectivity, and relatively poor reaction yields.
(24) The BOC and CBZ protective groups were selected for these
studies considering that the other functional groups of the aglycon
structure of glycopeptides are generally stable to the reaction conditions
(TFA, room temperature, for BOC; H2, Pd/C, for CBZ) required for their
removal from corresponding carbamate derivatives.
(25) The removal (TFA, room temperature) of the BOC protective
group from BOC-TDHP A gave TDHP A with high (95%) yields.
(26) Malabarba, A.; Ferrari, P.; Gallo, G. G.; Kettenring, J .; Caval-
leri, B. J . Antibiot. 1986, 39, 1430.
(27) By analogy with the proton nomenclature adopted for glyco-
peptides, HCR protons of amino acid fragments (n ) 1-7) are indicated
with the symbol xn, CH(OH) or CH2 benzyl protons are defined as zn
or zn, zn′, while symbol wn is used for NH or NH2 protons. Aromatic,
benzylic, and HCR protons are defined as shown in Figure 3.
(28) Marion, D.; Wuethrich, K. Biochem. Biophys. Res. Commun.
1983, 113, 967.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2143
Ta ble 3. Red u ctive Hyd r olysis of TD-Ca r ba m a tes
reductive hydrolysis productsa
CBZ-RH-TD (%) CBZ-TDTP DA (%)
solvent mixture
EtOH (%) H2O (%)
starting compd
CBZ-TDHP A (%)
epimers (%)
TDHP O (%)
RH-POSITION
1,2
∼50c
∼35c
∼10c
2,3
30
35
45
1,2 and 3,4
90b
60
10
40
80
CBZ-TD
CBZ-TD
CBZ-TD
10
10
5
10
20
40
45d
27d
9d
20b
solvent mixture
EtOH (%) H2O (%) starting compd
reductive hydrolysis productsa
BOC-TDHP A (%) BOC-RH-TD (%) BOC-TDTP DA (%) epimers (%) TDHP O (%)
90b
60
10
40
80
BOC-TD
BOC-TD
BOC-TD
60
40
15
25
30
40
7
10
5
∼8
20
40
absent
absent
absent
20b
a
b
The percent product composition of final reaction mixtures was determined by HPLC. When i-PrOH was used instead of EtOH, a
similar behavior was observed, but longer reaction times. c Highest percentage of CBZ-TDHP A observed (first step) before rearrangement
into TDHP O. Percentage of TDHP O present in the final reaction mixture (second step); transformation yield of CBZ-TDHP A into
d
TDHP O: ∼90%.
Sch em e 4. Red u ctive Hyd r olysis of
VA-Ca r ba m a tes a n d Ac-VAHP
conditions. This is the best way to explain the strong
NOE between proton x3 and aromatic protons 3f and 3b
and amidic w4. Furthermore, the NOE’s show that the
right-hand part of the molecule is folded backward.
Instead of having the typical pocket-like conformation of
teicoplanin,31 the molecule forms in R H-TD two half-
pockets pointing in opposite directions (Figure 6).
Ta ble 4. Red u ctive Hyd r olysis of BOC-VA a n d
N-Ac-VAHP
NMR Sp ectr a of th e Oth er 2,3-RH-Der iva tives.
All RH-teicop la n in s show the same NOE pattern in
their NOESY spectra like RH-TD, thus confirming their
solvent mixture
reductive hydrolysis products
starting
compd
EtOH (%) H2O (%)
VAHP
Ac-RH-VAHP
1
structural and conformational similarity. The H NMR
80
20
30
20
80
70
BOC-VA
BOC-VA
Ac-VAHP
100
100
60
spectra of the other RH-glycop ep tid es are similar to
those of RH-teicop la n in s. In their DQF-COSY spectra,
the proton assigned as x2 exhibits cross-peaks to two
different aliphatic CH-groups, to the proton (s) assigned
as z2 and the newly introduced CH2 moiety deriving from
the reductive cleavage of the amide bond between amino
acids 2 and 3. Furthermore, the NOESY spectra indicate
the same conformational changes as in R H -t eico-
p la n in s.
40
protons of RH-TD was done on the basis of their typical
spin systems. The following spin systems are as ex-
pected: four ABX (rings 1, 2, 5, 6), two AB (rings 4, 7),
and one AMN (ring 3). They are identified through their
typical coupling pattern in the DQF-COSY spectra. So,
for example, the four ABX-spin systems each show a
strong cross-peak due to the ortho-coupled protons and
a weak cross-peak for the meta-coupled protons. The
connection between the individual aromatic rings and
their benzylic protons was achieved by a carefully tuned
COSY experiment with enhancement of long-range cou-
plings,29 showing weak cross-peaks between the HCR
protons and the aromatic ortho protons of the attached
aromatic moieties. These results were confirmed by NOE
measurements that gave the through space connectivities
between these protons.
NMR Sp ectr a of TDHP A a n d epi-TDHP A. A first
inspection the 1H NMR spectra of the two compounds
compared to that of TD shows that for both the left-hand
part of the molecule is unchanged. The DQF-COSY
spectra allowed the assignment of most of the NH,HCR
connectivities. In both spectra only crosspeaks from the
HCR proton to the newly introduced alcoholic function
are present, so that the reductive hydrolysis has occurred
between amino acids 1 and 2. Only for TDHP A was the
proton spectrum completely assigned by homonuclear
experiments since the chemical shifts, particularly those
of protons of the aromatic region, are quite similar to
those of TD. For epi-TDHP A further studies were
necessary due to a drastic change in the chemical shifts
of the aromatic protons which leads to severe overlapping
in the 1H NMR spectra. However, the COSY spectra
show that the two compounds have the same primary
structure and that TDHP A has a conformation similar
to that of TD. For a detailed conformational study of
epi-TDHP A with the help of NOESY32 experiments the
whole proton spectrum had to be assigned, which was
possible only by heteronuclear experiments. In the
1H NMR Con for m a tion a l Stu d ies of RH-TD.
A
first indication of the conformational changes in RH-TD
with respect to TD results from the two aromatic high-
field protons. In TD, aromatic protons 4b and 4f are
influenced by the anisotropic effect of aromatic moieties
6 and 2, documented by their highfield shift. In R H-
TD, only 4f is shifted to a considerably high field (5.49
ppm), while 4b resonates at 6.45 ppm. On the other
hand, aromatic proton 3b experiences the anisotropy of
aromatic ring 1, thus shifted by 0.5 ppm to higher field
than its counterpart in TD. The NOESY30 spectrum
reveals further details of the conformation of RH-TD.
The protons on the left-hand part of the molecule up to
the x4 position show the same NOE pattern and therefore
the same conformation as TD. At the center x3 epimer-
ization seems to have occurred, likely due to the basic
(31) (a) Barna, J . C. J .; Williams, D. H.; Stone, D. J . M.; Leung,
T.-W. C.; Doddrell, D. M. J . Am. Chem. Soc. 1984, 106, 4895. (b) Hunt,
A. H.; Molloy, R. M.; Occlowitz, J . L.; Marconi, G. C.; Debono, M. J .
Am. Chem. Soc. 1984, 106, 4891.
(29) Bax, A.; Freeman, R. J . Magn. Reson. 1981, 44, 542.
(30) Macura, S.; Wuethrich, K.; Ernst, R. R. J . Magn. Reson. 1982,
47, 351.
(32) (a) Bodenhausen, G.; Kogler, H.; Ernst, R. R. J . Magn. Reson.
1984, 58, 370. (b) Macura, S.; Wuethrich, K.; Ernst, R. R. J . Magn.
Reson. 1982, 47, 351.

2144 J . Org. Chem., Vol. 61, No. 6, 1996
Sch em e 5. P r op osed RH-Mech a n ism in Ac-VAHP
Malabarba et al.
Ta ble 5. Hexa p ep tid e a n d Ca r ba m a te Der iva tives
HPLCa tR,
FAB MS
identified by their common crosspeaks with the already
assigned NH proton resonances. The two NH2-groups are
shifted in comparison with the amide resonances about
80 ppm to higher field, again confirming the primary
structure. The NOE’s of TDHP A correspond to the
NOE’s found in TD, except for the hydrolyzed part. In
epi-TDHP A, additional NOE interactions are detected
between proton x3 and the amide proton w4 and between
w3 and proton x2. The same NOE’s were detected in epi-
TD18 where epimerization occurred at the chiral center
x3. Accordingly, the molecule of epi-TDHP A, instead of
having the typical pocket-like conformation of TD and
TDHP A, is formed by two half-pockets pointing in
opposite directions as RH-TD.
Ma in 1H NMR Assign m en ts for BOC- a n d CBZ-
RH-TD, BOC-TDHP A, TDHP O, a n d BOC-TDTP DA.
The assignments have been derived from the two-
dimensional experiments COSY phase-sensitive double
quantum filter (COSYPHDQ) and COSY, relayed coher-
ence transfer (COSYRCT),35 and NOE phase-sensitive
(NOESYPH) experiments, and were based on the well-
established spectra-structure correlations in the teico-
planin field.26,31,36
compd
min (method) (M + H)+
formula
MW
BOC-TDHP A 16.1 (A), 6.8 (B)
1302
N.D.b
1202
N.D.
1336
1228
1306
1016
1062
C63H57N7Cl2O20 1303.0
CBZ-TDHP A
TDHP A
18.4 (A)
C66H55N7Cl2O20 1337.1
C58H49N7Cl2O18 1203.0
C63H57N7Cl2O20 1303.0
C66H55N7Cl2O20 1337.1
C59H47N7Cl2O19 1229.0
C63H61N7Cl2O20 1307.1
C46H39N7Cl2O16 1016.8
C48H45N7Cl2O17 1062.8
12.1 (A)c
BOC-RH-TD
CBZ-RH-TD
TDHP O
17.7 (A), 7.3 (B)
21.2 (A)
12.6 (A)
BOC-TDTP DA 13.9 (A)
VAHP 11.3 (A)
Ac-RH-VAHP 10.7 (A)
a
HPLC (method A), tR: BOC-TD, 24.0 min; CBZ-TD, 29.0 min.
N.D. ) not determined c For epi-TDHP A, tR 12.3 min.
b
HMQC33 spectrum of epi-TDHP A, the crosspeaks of the
protonated carbon atoms appear as doublets since no
decoupling was done during acquisition. As expected, in
the aromatic region 19 correlations are detected and only
few can be immediately assigned, viz., 4f, 4b and 3b, 3d,
and 3f. The HMBC34 experiment show correlations via
4
2J , 3J , and J couplings which allow, with the help of the
aforementioned homonuclear experiments, the assign-
1
ment of all H and 13C resonances of TDHP A and epi-
TDHP A. The assignment of the aromatic protons of ring
2 of epi-TDHP A is given here as an example. The z2
proton shows long-range couplings to one quaternary
carbon at 134.94 ppm and to one protonated carbon atom
at 131.43 ppm, which is correlated in the HMQC spec-
trum to a small doublet (3J ) 2 Hz) at 7.84 ppm in the
proton spectrum and can be assigned to the aromatic
proton 2b. In the DQF-COSY spectrum, this signal is
correlated via a small crosspeak (meta coupling) to 2f,
6.94 ppm, and this proton shows a strong ortho coupling
to a doublet at 7.12 ppm, which has to be assigned to 2e.
In the HMBC spectrum, 2b shows long-range correlations
to quaternary carbon atoms at 134.94, 128.47, and 150.10
ppm, and 2e to the signals at 134.94 and 128.47 ppm.
The resonance at 134.94 ppm has to be assigned to 2a,
as this signal is also correlated to z2. The carbon atom
at 150.10 ppm due to the chemical shift value has to be
2d, and one remaining resonance at 128.47 ppm has to
be assigned to 2c. It is worth mentioning that in both
cases long range crosspeaks are detected between 5b and
7b, thus confirming the presence of the diphenyl linkage
between these two aromatic rings. The HMQC experi-
ment allowed the assignment of all nitrogen resonances
of epi-TDHP A and most of those of TDHP A. For epi-
TDHP A, five amidic nitrogens are detected and readily
As expected, the spectra of BOC and CBZ-RH-TD
show a downfield shift (∆δ ∼+0.8 ppm) of proton x1 due
to the involvement of w1 into the carbamate formation.
A similar trend (∆δ +0.3 ppm) is found for BOC-
TDHP A. The 1H NMR spectrum of TDHP O is very
similar to that of TDHP A except for the presence of
signals due to a CH2CH system in an oxazolidinone
moiety37 instead of a HOCH2CH system as in TDHP A.
The structure of BOC-TDTP DA was clearly deduced
from the above 1H NMR experiments carried out in
DMSO-d6 solution added with traces of CF3COOH. The
reduction of the 1,2- and 3,4-amide bonds could be defined
+
by the coupling of the two newly formed NH3 groups
with their adjacent CH’s (x2 and x4) which, in turn, were
identified by NOE or long-range coupling interactions
with the protons of the surrounding part of the molecule.
Actually, the 1H NMR data allow the buildup of the whole
molecule of BOC-TDTP DA. Starting from the left-hand
side of the molecule, the chemical shift values and NOE
interactions concerning amino acid fragments 7-5 do not
show any difference as compared to those of TD‚HCl.
Considering amino acid residue 4, the x4 signal is
(35) Wagner, G. J . Magn. Reson. 1983, 55, 151.
(36) Malabarba, A.; Trani, A.; Ferrari, P.; Pallanza, R.; Cavalleri,
B. J . Antibiot. 1987, 40, 1572.
(33) Bax, A.; Subramanian, S. J . Magn. Reson. 1986, 67, 565.
(34) (a) Bax, A.; Summers, M. F. Chem. Soc. 1986, 108, 2093. (b)
Bax, A.; Marion, D. J . Magn. reson. 1988, 78, 186.
(37) The presence of an oxazolidinone moiety was confirmed by the
νCdO band at 1755 cm-1 in the IR spectrum of TDHP O.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2145
Sch em e 6. Syn th esis of RC-1 a n d RC-2 (F igu r e 5)
Ta ble 6. TDHP A-Der ived Da lba h ep tid es a n d Th eir In ter m ed ia tes
yielda
(%)
HPLC tR,
min (method)
FAB MS
compd
(M + H)+
formula
MW
BOC-Ac-1
BOC-Ac-2
Ac-1
Ac-2
RC-1
70
95
70
95
60
85
10.8 (B)
11.9 (B)
14.75 (A), 5.9 (B)
15.8 (A), 6.3 (B)
12.7 (A)
1423
1437
1323
1337
1242
1256
C65H58N7BrCl2O21
1424.0
1438.0
1323.9
1337.9
1243.0
1257.0
C
C
C
C
C
66H60N7BrCl2O21
60H50N7BrCl2O19
61H52N7BrCl2O19
60H49N7Cl2O19
RC-2
13.1 (A)
61H51N7Cl2O19
a
From BOC-TDHP A.
identified on the basis of its spatial proximity, revealed
by NOE interactions, to w5 and 4f, which in turn is
identified by its signal generally at very high field in the
teicoplanin family with respect to the other aromatic
protons. Compared with that of TD‚HCl, the chemical
shift of the x4 signal indicates some changes in the x4
the 3,4-amide underwent a reductive hydrolysis. Residue
3 is characterized by the mutual meta coupling of three
aromatic protons, one of which shows NOE interactions
with the aliphatic CH, CH2, and NH groups. The signals
of the CH2 protons of residue 2 show a NOE effect with
+
the 2b proton and coupling with the CHNH3 system.
+
region. Its coupling with an NH3 group, whose signal
resonates at 8.50 ppm, shows that the 3,4-peptide linkage
has been hydrolyzed. In addition, the presence of the
CH2OH group in the structure of residue 3 confirms that
F igu r e 5. Structure of RC-1 (n ) 1) and RC-2 (n ) 2).
F igu r e 6. Stereostructure model of RH-TD.

2146 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
Ta ble 7. Assign m en ts of Ch a r a cter istic Sign a ls of th e 1H
NMR Sp ectr a of RC-1 a n d RC-2 in Com p a r ison w ith TD
(in DMSO-d 6, TMS In ter n a l Refer en ce, δ in p p m )
in addition to the (MH - 16)+ peak, there is a much
smaller one formed by H2O loss which is a “real” gas
phase process, as indicated by a corresponding meta-
stable peak. A very interesting correlation was observed
between the facility to lose a sugar unit in FAB and the
reactivity of the same sugar unit in solution chemistry.
Hydrolysis of these compounds proceeds as follows:39 RH-
CTA easily loses the acylglucosamine, and then the
mannose is lost, but this reaction requires much stronger
conditions. The acetylglucosamine is lost only in the last
step. The fragment ion abundances, as observed in FAB
(reactivity of gas phase ions), are parallel with this: the
loss of the acylglucosamine unit gives a very abundant
fragment, while the loss of either mannose or acetylglu-
cosamine gives small peaks, but mannose loss seems to
be slightly favored. This correlation is interesting for
fundamental studies on the reactivity of ions and neu-
trals but also has practical importance in estimating the
reactivity of certain groups based on FAB spectra.
The FAB spectra of the other RH-d er iva tives are
analogous to those of RH-teicop la n in s. All compounds
give abundant molecular ion cluster four mass units
higher than the starting material. These derivatives,
similar to those of teicoplanin, give nucleophilic substitu-
tion products at 16 (OH f H) and, except for RH-HR,
34 (Cl f H) Da lower masses under FAB conditions. It
is noteworthy that sensitivity for RH-HR is much higher
than that for the other compounds. This might indicate
that the chlorine substitution in glycopeptides reduces
FAB sensitivity. The spectrum of BOC-TDTP DA gives
a molecular ion 8 mass units higher than that of BOC-
TD, in accordance with the reduction of two amide bonds,
while the MW of TDHP O was found to be 26 mass units
higher than that of TDHP A, which is in agreement with
the introduction of one CO and loss of two H and
consequently with the oxazolidinone structure of this
compound. The MWs found for RC-1 and RC-2 are in
accordance with their structure of ring closed compounds.
P ep tid e Bin d in g Stu d ies a n d An tiba cter ia l Activ-
it y. The ability of R H-t eicop la n in s and related 2,3-
RH-glycop ep tid e d er iva tives, as well as that of TDH-
P A, its C3-epimer epi-TDHP A, and ring-closed compounds
RC-1 and RC-2 to complex with the antibiotic’s target
peptide D-Ala-D-Ala, was investigated by measuring their
binding to the synthetic analog Ac2-L-Lys-D-Ala-D-Ala by
both UV and NMR methods. Using the diffential UV
assay,40 no detectable variation in absorbance was ob-
served on adding the test peptide. Nevertheless, with
the only exception of epi-TDHP A, the compounds were
retained by a resin of D-Ala-D-Ala covalently linked with
an inert matrix,41 thus showing affinity for the dipeptide.
This apparent discrepancy could be interpreted as being
due to a different complexation which does not cause any
detectable difference in the UV absorbance of the open
molecule. Limited to 2,3-RH-d er iva tives, preliminary
investigation using the 2D NMR spectroscopy seems to
indicate that no binding exists between the two amide-
NH groups of residues 2 and 4, and the free NH2 of
residue 3 and the carboxylate anion of target peptide
model, but that a weak binding still exists which involves
the amide-NH proton of residue 7 and the lysyl-CO of
the synthetic tripeptide. Further NMR studies are in
progress to assess this unusual mechanism. In fact, for
assignment
RC-1a
RC-2a
TDa
CH2OH
3.72
3.64
CHCH2OH
CH2CH2-CdO
H2N+CH
5.55
5.46
3.30
8.19
4.13, 3.60
4.71
2.76
5.66
5.72
4.34
4.12
4.42
5.47
5.33
7.65
4.40, 4.26
4.85
2.75
5.34
5.57
4.33
4.12
4.45
+
CH2NH2
x2
z2
x3
x4
x5
x6
x7
4b
4f
4.92
2.87
5.35
5.60
4.33
4.10
4.42
5.50
5.08
5.63
5.10
a
‚HCl.
This finding identifies the amide bond linking amino acid
fragments 1 and 2 as the second hydrolysis site. The
reduction of the 1,2-peptide linkage is confirmed by the
structure of residue 1, which was identified by decou-
plings, obtained from the COSY spectrum in the system
CH2CHNH, where the NH group is part of a carbamate
1
moiety, as shown by its H NMR signal which does not
change on adding CF3COOH.
NMR Sp ectr a of RC-1 a n d RC-2. The structures of
RC-1 and RC-2 were defined on the basis of their COSY
spectra and by comparing their 1H NMR spectra with
that of TD‚HCl (Table 7). The two ring-closed com-
pounds are in the form of hydrochlorides as shown by
1
the NH2+ signal in their H NMR spectrum. As expected,
the main changes are limited to the right-hand side of
the molecule while the overall conformation of the core
structure appears to be almost unmodified in RC-1.
Some minor changes are observed in RC-2.
F AB MS Sp ectr a . The FAB MS spectra of RH-
teicop la n in s show abundant molecular ion clusters. The
lowest mass isotopes of the protonated molecular ions
were found to be 1882.7 (RH-CTA, factor A2-2),17 1567.4
(RH-TB), 1405.4 (RH-TC), and 1202.3 (RH-TD), in
agreement within 0.1 Da with the values calculated for
the structures assigned. The molecular masses calcu-
lated for average isotope content are about 1.0 Da higher.
The isotope pattern of the molecular ion clusters is also
in agreement with the proposed structures. Due to traces
of Na in the samples, all spectra show various amounts
of molecular ions (M + Na)+. Losses of sugar units can
also be observed, giving further structural information.
All these ions are accompained by peaks formally derived
by losses of 16 and 34 mass units, and these are common
to all teicoplanin derivatives. The latter process, which
occurs in the liquid matrix under FAB conditions, was
also observed in the FAB study of chlorinated nucleosides
and is due to a displacement of a Cl by an H.38 The
isotope distribution of the (MH - 34)+ peak suggests the
loss of one Cl atom. By analogy, a 16 Da loss indicates
displacement of an OH by an H, which may proceed with
a similar mechanism. To our knowledge, this is the first
observation of an OH/H exchange occurring in FAB
conditions. This adds to the growing evidence that
bombardment of a liquid matrix with fast neutrals
produces reactions not unlike to those observed in pho-
todecompositions. It might be interesting to add that,
(39) Malabarba, A.; Strazzolini, P.; Depaoli, A.; Landi, M.; Berti,
M.; Cavalleri, B. J . Antibiot. 1984, 37, 988.
(38) McCloskey, J . A.; Nelson, C. C.; Sethi, S. K. Anal. Chem. 1984,
56, 1975.
(40) Nieto, M.; Perkins, H. R. Biochem. J . 1971, 124, 845.
(41) Corti, A.; Cassani, G. Biochem. Biotechnol. 1985, 11, 101.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2147
Ta ble 8. In Vitr o Activity of RH-Glycop ep tid es in Com p a r ison w ith P a r en t Un m od ified Glycop ep tid es
MIC (µg/mL)
Staphylococcus
Staphylococcus
Streptococcus
Streptococcus
Enterococcus
compd
aureus Tour
epidermidis ATCC 12228
pyogenes C 203
pneumoniae UC 41
faecalis ATCC 7080
RH-CTA
RH-TD
RH-A-40
RH-V
8
4
16
16
2
2
64
32
1
8
2
8
4
16
4
4
16
16
16
8
RH-HR
>128
CTA
TD
A-40
V
0.125
0.063
0.063
0.25
4
0.125
0.016
0.063
0.5
0.063
0.125
0.063
0.125
0.5
0.063
0.125
0.063
0.125
1
0.125
0.125
0.125
0.5
R
4
1
HR
4
8
1
1
2
Ta ble 9. An tiba cter ia l Activity of TDHP A, RC-1, a n d
RC-2 in Com p a r ison w ith TD a n d RH-TD
MICa (µg/mL)
test organisms
TDHP A RC-1 RC-2 RH-TDb
TDb
S. aureus Tour L165
S. epidermidis L147
S. epidermidis L533
S. haemolyticus L602
S. pyogenes L49
2
0.5
1
1
1
0.5
8
4
16
32
8
>128
32
32 (4) 0.063
8 (2) 0.063 (0.016)
16 (8) 0.063
16
16
0.25
8
4
8
32 (8) 0.25 (0.13)
S. pneumoniae L44
E. faecalis L149
4
4
16
32
16
16
0.13
0.13
Escherichia coli L47 >128
>128 >128 >128
64
a
Values given were obtained with final inoculum of 5 × 105
cfu/mL. Values in parentheses refer to inocula of 104 cfu/mL.
b
F igu r e 7. Structure of the 1,2,3-macrocyclic ring in RC-1
compared with that generated by a hydrogen bonding system
in TDHP A.
the antibiotics of the vancomycin family it has been
reported42 that this secondary binding only exists as a
consequence of the primary interaction. However, a
mechanism of action different from that of “natural”
glycopeptides cannot be excluded in principle.
at the rear. This results in the loss of the ability by
VAHP to bind to dalbaheptide’s target peptide D-Ala-D-
Ala6 and hence in the loss of the antibiotic activity. In
epi-TDHP A, the change in the configuration of the 1,3-
diphenyl ether amino acid is the same as that occurring
at the chiral center of amino acid 3 of TD in the
epimerization to epi-TD. Like epi-TD, the molecule of
epi-TDHP A, instead of having the typical pocket-like
conformation in the binding region, has two half-pockets
pointing in opposite directions. It follows a perturbation
in the active site which results in the loss of antibacterial
activity. Also in RH-TD, due to the opening of the 2,3-
amide bond, the right-hand part of the molecule is folded
backward, but in this case the “natural” S configuration
of amino acid 3 is in part maintained.45 This could
explain the residual antibacterial activity of RH-TD. The
comparable activity between TDHP A and its cyclic
derivative RC-1 was expected considering that the
overall conformation of the two compounds is very similar
and not very different from that of TD, as drawn from
NMR conformational studies. One possibility is the
presence in TDHP A of a hydrogen bonding, favored at
the neutral/physiological pH of the in vitro tests, between
the peptide N-terminal amino group of residue 2 and the
newly formed hydroxyl group of residue 1 (Figure 7). The
resulting 16-membered hydrogen bonding system should
have similar conformation and geometry as that of the
corresponding 16-membered 1,2,3-macrocyclic ring present
in the structure of RC-1. However, the lower activity of
these two compounds compared with that of TD indicates
that the enlargement of the 1,2,3-macrocyclic ring has a
negative effect on their binding properties to the target
The drastic structural modification of the binding
pocket accounts for the decreased, though interesting,43
in vitro activity of 2,3-RH-d a lba h ep tid es against a
panel of selected Gram-positive bacteria (Table 8). The
lack of activity of RH-HR (MIC >128 µg/mL) is likely
related to the poor activity of R. Although inactive
against Gram-negative bacteria and less active than TD
against Gram-positive organisms, TDHP A was signifi-
cantly more active than RH-TD against the majority of
tested strains (Table 9). In contrast, epi-TDHP A was
inactive at a concentration of up to 128 µg/mL.44 The
activity of ring-closed compound RC-1 was comparable
to that of TDHP A, while homologous cyclic derivative
RC-2 was markedly less active. The activity of TDHP A
was unexpected considering that the vancomycin-derived
hexapeptide aglycon VAHP was devoid of antibacterial
activity.6 This is very likely due to the different confor-
mation of their hexapeptide backbone in the active site
region. While TDHP A almost maintains the original
conformation of the peptide chain and in particular the
orientation of the 2,3-amide as TD, in VAHP the 2,3-
trans peptide unit exhibits a 180˚ flipping motion, from
having the NH-proton at the front of the molecule, as in
VA, TD, and the other active glycopeptides, to having it
(42) Williams, D. H.; Waltho, J . P. Biochem. Pharmacol. 1988, 37,
133.
(43) The unexpected activity of RH-glycopeptides is of particular
interest considering that their mode of action is very likely based on
the interaction (Chart 1) with a target moiety, the lysyl-CO which is
present in the peptidoglycan portion of the bacterial cell wall of both
glycopeptide-susceptible (L-Lys-D-Ala-D-Ala) and resistant (L-Lys-D-Ala-
D-hydroxy acid) organisms.
(45) RH-TD, as obtained by reductive hydrolysis of TD, is an
unresolved mixture of two diasteroisomers S (RH-TD, as the main
product, ∼60%) and R (epi-RH-TD) at the C3-chiral center.
(44) Data not reported.

2148 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
evaporated at 35 °C under reduced pressure. The jelly residue,
dissolved in 1 L of H2O, was purified by reversed-phase column
chromatography as described above. Fractions containing
pure (HPLC) RH-derivatives were pooled and solvents were
evaporated, at 40 °C under reduced pressure, in the presence
of 1-BuOH to avoid foaming. The solid residue was collected,
washed with 200 mL of Et2O, and dried at room temperature
in vacuo (over KOH) for 3 days, yielding the final product.
dipeptide. This is even more evident with a further ring
enlargement by one additional methylene group, as
shown by the poor activity of RC-2.
Exp er im en ta l Section
The 1H NMR experiments were recorded at 500 MHz in
DMSO-d6 solution, added or not with CF3COOH. The phase-
sensitive double quantum filter (PHDQ) 1H-1H COSY spectra
were run using time-proportional-phase-incrementation in f1.
In the relayed-coherence-transfer (RCT) 1H-1H COSY, the
fixed delay was 25 ms. The phase-sensitive NOE spectra were
obtained with a mixing time of 200 ms. In the HMQC
P r ep a r a tion of RH-TB fr om RH-CTA (Meth od b).
A
solution of 2 g (∼1 mmol) of RH-CTA in 20 mL of dry
trifluoroacetic acid (TFA) was stirred at room temperature for
30 min, and then 130 mL of Et2O was added. The precipitated
solid was collected and redissolved in 100 mL of H2O. The
resulting solution was adjusted at pH 7.5 with 1 N NaOH.
Purification by reversed-phase column chromatography under
usual conditions yielded 1.2 g of the title compound.
P r ep a r a t ion of R H -TC fr om R H -CTA or R H -TB
(Meth od c). Dry HCl was bubbled at room temperature into
a stirred suspension of 1 mmol of RH-CTA, or RH-TB, in 100
mL of dimethoxyethane (DME) for 6 h. Then, the solvent was
evaporated at 40 °C under reduced pressure, and the solid
residue, dissolved at pH 7.5 in 100 mL of H2O, was chromato-
graphed to give (55-60%) the title compound.
P r ep a r a tion of RH-TD fr om RH-CTA, RH-TB, or RH-
TC (Meth od d ). Dry HCl was bubbled at 70 °C into a stirred
suspension of 1 mmol of one of the above RH-teicop la n in s
in 100 mL of trifluoroethanol (TFE) for 16 h. The insoluble
matter was collected and dissolved in 200 mL of H2O at pH
7.0. After purification by column chromatography, the title
compound was obtained (20-25%).
Red u ctive Hyd r olysis of TD-Ca r ba m a tes. The com-
parative experiments (Table 3) were carried out as follows. To
a stirred solution of 4 mmol of BOC or CBZ-TD in 150 mL of
a proper aqueous alcohol mixture, was added 30 g of NaBH4
(pellets) portionwise in 90 min, while the temperature was
maintained at 20-25 °C. Stirring was continued at room
temperature until the starting compound was completely
transformed (the course of the reaction was monitored by
HPLC every 2 h), and then the reaction mixture was slowly
poured into 300 mL of a MeOH/EtOH/AcOH 3/2/1 solution.
After 1-BuOH was added to prevent foaming, the solvents were
evaporated at 45 °C under reduced pressure. The solid residue
was collected and chromatographed to yield pure final prod-
ucts.
Syn th esis of TDHP A (a n d epi-TDHP A). (a ) F r om TD.
To a stirred solution of 100 g (∼83 mmol) of TD in 3.5 L of a
H2O/EtOH 6/4 mixture was added 600 g (∼16 mol) of NaBH4
(pellets) portionwise in 2 h, with cooling to 15-25 °C. The
reaction mixture was stirred at room temperature for 22 h,
and then it was poured into a solution of 960 mL of glacial
AcOH in 5.5 L of a MeOH/EtOH 65/35 mixture, with cooling
to 10 °C. The resulting suspension was centrifuged, and the
insoluble matter (5 g, epi-TD)46 was separated. The clear
solution was concentrated at 45 °C under reduced pressure to
a final volume of ∼500 mL in the presence of 1-BuOH to
prevent foaming. The precipitated solid (mainly boron salts)
was filtered off, and the filtrated solution was loaded on a
column of 2.5 kg of silanized silica gel in H2O. The column
was eluted with 10 L of H2O and then with 10 L of a MeCN/
H2O 1/1 mixture while 500 mL fractions were collected.
Fractions 9-16 contained crude RH-TD (∼55 g, HPLC titer
63%). Fractions 21-30, containing a mixture of the title
compound and its C3-epimer, were pooled, and 9 L of 1-BuOH
was added. The resulting mixture was concentrated at 40 °C
under reduced pressure to a small (∼200 mL) volume. Upon
addition of Et2O (500 mL), the precipitated solid was collected
and washed with 100 mL of Et2O, yielding 30 g of a powder
containing TDHP A (83%) and epi-TDHP A (17%). Resolution
of the above mixture by reversed-phase chromatography
yielded 18 g of pure TDHP A and 0.7 g of pure epi-TDHP A.
(b) F r om BOC-TD. To a stirred solution of 10.5 g (∼8
mmol) of BOC-TD in 300 mL of a EtOH/H2O 9/1 mixture was
added 60 g of NaBH4 (powder) portionwise at 20 °C in 2 h.
1
experiments for H-13C correlations, the sweep width in t2 was
5500 Hz and t1 11300 Hz; 512 increments with 64 transitions
each were collected. For the HMBC experiments with epi-
TDHP A, 196 transitions were collected. For the 1H-15N
correlations, the sweep width in t1 was 2000 Hz; 128 incre-
ments with 200 experiments each were collected.
FAB MS positive ion spectra were obtained using 8 kV
accelerating voltage. The samples were dissolved in DMF or
H2O, and then 1 µL of the solution was mixed with 1 µL of
thioglycerol matrix containing 0.1 N AcOH on the target.
Reaction products were purified by reversed-phase column
chromatography on silanized silica gel (0.063-0.2 mm), ac-
cording to the following procedure: 1 g of crude compound was
dissolved in 20-30 mL of a MeCN/H2O (1/1) mixture, the
solution was adjusted at pH 5.5 with solid HCO2NH4, and then
H2O was added dropwise under stirring until precipitation
started; after a few drops of MeCN were added, the resulting
cloudy solution was loaded on a column of 50 g of silanized
silica gel in a same solvent mixture; elution was carried out
according to a linear step-gradient from 5-10% to 40-70% of
MeCN in H2O, in 10-15 h, at a flow rate of 200-300 mL/h,
while 20 mL fractions were collected; those containing pure
compound were pooled, and enough 1-BuOH was added to
obtain, after evaporation of most solvent at 40 °C under
reduced pressure, a concentrated dry butanolic solution (or
suspension); on adding Et2O the precipitated solid was col-
lected, washed with Et2O, and dried in vacuo at room tem-
perature overnight.
Reactions, column eluates, and final products were checked
by HPLC performed on a column (125 × 4 mm) prepacked with
LiChrospher RP-8 (5 µm). Chromatograms were recorded at
254 nm. Elutions were carried out by mixing eluent a, MeCN,
with eluent b, 0.2% aqueous HCO2NH4, according to linear step
gradients programmed as follows:
time (min)
0
10
20
30
35
40
Meth od A: % b in a
Meth od B: % b in a
5
20
23
33
26
46
35
60
75
75
5
20
All RH-derivatives were analyzed for C, H, N, and Cl on
samples previously dried at 140 °C under N2 atmosphere. The
analytical results obtained for the above elements were within
(0.4% of the theoretical values. The solvent content (<12%,
mainly water) and inorganic residue (<0.2%) were determined
by thermogravimetry (TG), at 140 °C, after the samples were
heated at 900 °C in O2 atmosphere, respectively.
The V-aglycon (VA) was prepared by reaction of V with TFA,
according to the procedure described by Nagarajan and Scha-
bel.6 BOC and CBZ-TD were prepared as previously de-
scribed.36
Red u ctive Hyd r olysis of Glycop ep tid es Teicop la n in
(CTA a n d Its Sin gle F a ctor A2-2), Its P seu d oa glycon s (TB
a n d TC) a n d It s Aglycon (TD), Va n com ycin (V), It s
Aglycon (VA), Ristocetin (R), a n d A-40,926 (A-40) (Gen -
er a l P r oced u r e, Meth od a ). A suspension of 10 mmol of a
dalbaheptide in 600 mL of a H2O/EtOH 65/35 mixture was
stirred at 10-15 °C for 90 min, while a suitable amount
(Tables 1 and 2) of NaBH4 pellets was added portionwise. A
clear solution formed which was stirred at room temperature
over a period of time depending on the starting compound.
After 1 L of MeOHand 0.5 L of EtOH were added, the resulting
solution was slowly poured into a solution of an excess (10%)
of AcOH in 0.5 L of MeOH, and then the solvents were
(46) Precipitated epi-TD (20% over the total amount formed) was
enough pure for analysis.

Hydrolysis of the 1,2- and 2,3-Peptide Bonds in Teicoplanin
J . Org. Chem., Vol. 61, No. 6, 1996 2149
After being stirred at room temperature for 72 h, the reaction
mixture was worked up as described above, yielding 5.6 g of
BOC-TDHP A. This compound was dissolved at room tem-
perature in 50 mL of dry TFA. After 5 min of stirring, the
solvent was evaporated and the oily residue was slurried with
50 mL of EtOAc and then with 150 mL of Et2O. The solid
which separated was chromatographed under the usual condi-
tions, yielding 4.9 g of pure TDHP A.
P r ep a r a tion of TDHP O via In ter m ed ia te CBZ-TDH-
P A. A stirred solution of 10.7 g (∼8 mmol) of CBZ-TD in 300
mL of a EtOH/H2O 9/1 mixture was treated with 60 g of NaBH4
as described above. After 48 h, the reaction mixture was
divided into two portions of 150 mL each. One portion was
worked up as usual to yield 1.7 g of pure CBZ-TDHP A. The
other aliquot was allowed to react for additional 72 h until
CBZ-TDHP A completely transformed into TDHP O, which
was recovered and chromatographed,47 obtaining 1.4 g of pure
title compound.
Syn th esis of VAHP fr om BOC-VA. (a ) P r ep a r a tion of
BOC-VA. A stirred solution of 1.15 g (∼1 mmol) of VA in 25
mL of a dioxane/water 1/1 mixture was adjusted at pH 6.5 with
solid NaHCO3, and then a solution of di-tert-butyl dicarbonate
in 3 mL of dioxane was added dropwise in 15 min. After being
stirred at room temperature for 24 h, the reaction mixture was
adjusted at pH 3 with 1 N HCl and 50 mL of H2O was added.
Extraction with 50 mL of a 1-BuOH/EtOH 1/1 mixture and
evaporation of the organic solvents yielded 1.2 g of BOC-VA
pure enough for the next step.
(b) Red u ctive Hyd r olysis of BOC-VA. The compound
(BOC-VA, 1.2 g) was dissolved in 50 mL of a EtOH/H2O 8/2
(or 2/8) mixture and then treated with 6 g of NaBH4 as
described above for the reductive hydrolysis of BOC-TD,
yielding 0.85 g of pure title compound.
Syn th esis a n d Red u ctive Hyd r olysis of Ac-VAHP . The
compound was prepared (98%) by reaction of 1 mmol of VAHP
with 1.1 mmol of MeCOCl in 20 mL of DMF at room
temperature (4 h), in the presence of 1 mmol of TEA. Then it
was submitted to reductive hydrolysis (reaction time: 35 h)
under the same conditions described above for BOC-VA,
yielding48 VAHP (36%) and Ac-RH-VAHP (24%).
Syn th esis of RC-1 a n d RC-2. Step a : BOC-Ac-1 a n d
BOC-Ac-2. To a stirred solution of 1.3 g (∼1 mmol) of BOC-
TDHP A and of 84 mg of NaHCO3 in 30 mL of a Me2CO/H2O
1/1 mixture was added dropwise a solution of 1 mmol of
bromoacetyl (for BOC-Ac-1) or 3-bromopropionyl (for BOC-
Ac-2) chloride in 2 mL of Me2CO with cooling to 0-3 °C. After
15 min, the reaction mixture was poured into 40 mL of H2O,
and after the pH was adjusted to 4.5 with glacial AcOH, the
resulting mixture was extracted with 100 mL (2 × 50 mL) of
EtOAc (or 1-BuOH). The organic layer was washed with 25
mL of H2O, and then it was concentrated at 35 °C (or 45 °C)
under reduced pressure to a small volume (∼5 mL). On
addition of Et2O (∼50 mL), the precipitated solid was collected
and dried in vacuo at room temperature overnight, yielding
BOC-Ac-1 (0.79 g) or BOC-Ac-2 (1.3 g).
Step b: Ac-1 a n d Ac-2. A solution of 0.7 mmol of one of
the above BOC-acyl derivatives in 10 mL of dry TFA was
stirred at room temperature for 10 min, and then the solvent
was evaporated (30 °C, reduced pressure). The oily residue
was redissolved in 50 mL of H2O, and the resulting solution
was adjusted at pH 7 with 1 N NaOH then extracted with 50
mL of 1-BuOH. The organic layer was washed with 15 mL of
H2O and concentrated at 40 °C under reduced pressure to a
small (∼5 mL) volume. Upon addition of 50 mL of Et2O, the
precipitated solid was collected, washed with 20 mL of Et2O,
and then dried in vacuo at room temperature overnight,
yielding (100%) Ac-1 or Ac-2.
Step c: RC-1 a n d RC-2. To a stirred solution of 0.5 mmol
of one of the above compounds in 40 mL of DMF was added
100 mg (∼0.8 mmol) of K2CO3, and the resulting suspension
was stirred at room temperature for 2 h. Then the reaction
mixture was poured into 100 mL of H2O. The resulting
solution was adjusted at pH 3 with 1 N HCl, and then it was
extracted with 70 mL of 1-BuOH. The organic layer was
washed with 50 mL of H2O, and then it was concentrated at
35 °C under reduced pressure to a small (∼10 mL) volume.
On addition of 50 mL of Et2O, the precipitated solid was
collected and purified by reversed-phase chromatography
under the usual conditions, yielding pure RC-1 (85%) or RC-2
(90%).
NMR Da ta (DMSO-d6, δ in p p m ) for Som e Rep r esen ta -
tive Com p ou n d s. RH-TD: 1H NMR δ 2.67 (m, 1H), 2.87
(m, 1H), 3.51 (br s, 2H), 3.96 (m, 1H), 4.17 (br s, 1H), 4.20 (dd,
1H), 4.33 (br s, 1H), 4.38 (d, 1H), 4.59 (br s, 1H), 5.11 (s, 1H),
5.49 (s, 1H), 5.65 (d, 1H), 5.83 (s, 1H), 6.36 (s, 1H), 6.39 (s,
1H), 6.45 (s, 3H), 6.58 (d, 1H), 6.62 (s, 1H), 6.70 (d, 2H), 6.77
(d, 1H), 6.83 (d, 1H), 6.93 (d, 1H), 6.96 (d, 1H), 7.16 (d, 2H),
7.23 (s, 1H), 7.42 (d, 1H), 7.74 (d, 1H), 7.89 (s, 1H), 8.39 (d,
1H), 8.62 (br s, 1H), 8.64 (d, 1H).
BOC-RH-TD (‚TFA): 1H NMR δ 1.38 (s, 9H), 2.70 (m, 1H),
2.90 (m, 1H), 3.40-3.50 (m, 2H), 4.00 (m, 1H), 4.23 (dd, 1H),
4.42 (d, 1H), 4.62 (br s, 1H), 4.85 (br s, 1H), 4.92 (br s, 1H),
5.11 (s, 1H), 5.50 (s, 1H), 6.32 (s, 1H), 6.75 (br s, 1H), 7.18 (s,
1H), 7.31 (br s, 1H), 7.42 (d, 1H), 7.89 (s, 1H), 8.60 (d, 1H, and
NH3+), 8.90 (br s, 1H), 9.20 (d, 1H).
CBZ-RH-TD (‚TFA): 1H NMR δ 2.75 (m, 1H), 2.90 (m, 1H),
3.38-3.48 (dd, 2H), 4.00 (m, 1H), 4.28 (dd, 1H), 4.46 (d, 1H),
4.65 (br s, 1H), 4.90 (br s, 1H), 5.00 (br s, 1H), 5.05 (br s, 2H),
5.14 (s, 1H), 5.50 (s, 1H), 5.64 (d, 1H), 6.00 (s, 1H), 6.32 (s,
1H), 6.48 (s, 1H), 6.70 (s, 1H), 6.80 (br s, 1H), 7.35 (m, 5H),
7.45 (d, 1H), 7.85 (br s, 1H), 7.90 (s, 1H), 8.60 (d, 1H, and
NH3+), 8.88 (br s, 1H), 9.22 (d, 1H).
TDHP A: 1H NMR δ 2.74 (dd, 1H), 3.21 (dd, 1H), 3.58 (d,
1H), 3.62 (d, 1H), 3.71 (m, 1H), 4.10 (m, 1H), 4.12 (dd, 1H),
4.37 (d, 1H), 4.42 (d, 1H), 5.08 (d, 1H), 5.25 (s, 1H), 5.48 (d,
2H), 5.96 (d, 2H), 6.26 (s, 2H), 6.35 (s, 1H), 6.40 (s, 1H), 6.63
(d, 1H), 6.68 (d, 1H), 6.72 (d, 1H), 6.94 (d, 1H), 7.07 (s, 1H),
7.10 (d, 1H), 7.12 (d, 2H), 7.16 (s, 1H), 7.28 (d, 1H), 7.43 (d,
1H), 8.42 (d, 1H), 8.42 (br s, 1H), 8.52 (br s, 1H); 13C NMR δ
39.72 (t), 54.14 (d), 54.70 (d), 55.99 (d), 56.36 (d), 57.29 (d),
58.03 (d), 62.09 (d), 63.76 (t), 71.89 (d), 102.40 (d), 102.95 (d),
104.80 (d), 106.47 (d), 107.36 (d), 107.69 (d), 116.64 (d), 117.56
(d), 118.49 (s), 121.63 (d), 123.36 (d), 124.96 (d), 125.69 (s),
125.88 (d), 126.25 (s), 127.36 (d), 127.54 (s), 127.91 (d), 128.64
(s), 131.24 (d), 134.39 (s), 135.28 (s), 136.45 (s), 136.73 (d),
137.90 (s), 142.15 (s), 148.31 (s), 149.73 (s), 150.47 (s), 155.46
(s), 156.76 (s), 157.68 (s), 158.56 (s), 159.16 (s), 161.56 (s),
167.85 (s), 168.96 (s), 169.33 (s), 169.63 (s), 172.84 (s), 173.03
(s); 15N NMR δ 87.7 (d), 95.8 (d), 105.2 (d), 107.5 (d).
epi-TDHP A: 1H NMR δ 2.69 (dd, 1H), 3.42 (dd, 1H), 3.60
(m, 1H), 3.67 (dd, 2H), 4.11 (dd, 1H), 4.14 (br s, 1H), 4.31 (d,
1H), 4.42 (d, 1H), 4.85 (s, 1H), 5.12 (br s, 2H), 5.51 (s, 1H),
5.87 (d, 1H), 5.92 (s, 1H), 6.00 (d, 1H), 6.19 (s, 2H), 6.25 (s,
1H), 6.47 (s, 1H), 6.65 (d, 1H), 6.71 (d, 2H), 6.80 (br s, 1H),
6.94 (d, 1H), 6.97 (d, 1H), 7.06 (d, 1H), 7.12 (d, 3H), 7.24 (d,
1H), 7.43 (d, 1H), 7.82 (s, 1H), 7.84 (s, 1H), 8.31 (br s, 1H),
8.48 (br s, 1H), 8.59 (d, 1H), 8.95 (d, 1H); 13C NMR δ 36.58 (t),
53.22 (d), 54.11 (d), 54.33 (d), 56.92 (d), 60.25 (d), 62.09 (d),
63.02 (t), 71.71 (d), 102.58 (d), 103.69 (d), 104.25 (d), 105.91
(d), 106.10 (d), 107.58 (d), 109.24 (d), 116.45 (d), 117.38 (d),
118.12 (s), 121.44 (d), 123.48 (d), 124.96 (d), 125.14 (s), 125.70
(s), 125.76 (d), 126.06 (s), 126.44 (s), 127.54 (d), 127.91 (d),
128.47 (s), 131.43 (d), 131.98 (d), 134.20 (s), 134.94 (s), 136.05
(d), 136.23 (s), 137.90 (s), 142.52 (s), 147.51 (s), 148.25 (s),
149.18 (s), 149.55 (s), 150.10 (s), 155.28 (s), 156.76 (s), 157.50
(s), 158.97 (s), 158.98 (s), 165.03 (s), 167.85 (s), 168.40 (s),
169.22 (s), 172.43 (s); 15N NMR δ 21.0 (t), 22.5 (t), 87.0 (d),
100.2 (d), 103.0 (d), 106.5 (d), 107.3 (d).
BOC-TDHP A: 1H NMR δ 1.34 (s, 9H), 2.70 (dd, 1H), 3.22
(dd, 1H), 3.42 (m, 2H), 3.69 (m, 1H), 4.12 (dd, 1H), 4.38 (d,
1H), 4.40 (m, 2H), 5.10 (d, 1H), 5.26 (s, 1H), 5.47 (s, 1H), 5.51
(d, 1H), 5.90 (s, 1H), 5.96 (d, 1H), 6.22 (s, 1H), 6.30 (s, 1H),
6.55 (d, 1H), 7.81 (s, 1H), 8.35 (br s, 1H), 8.43 (d, 1H), 8.50 (br
s, 1H).
TDHP O: 1H NMR δ 2.76 (dd, 1H), 3.21 (dd, 1H), 3.69 (m,
1H), 3.94 (m, 1H), 4.12 (dd, 1H), 4.39 (d, 2H), 4.58-4.80 (dd,
2H), 5.09 (s, 1H), 5.26 (s, 1H), 5.47 (s, 1H), 5.50 (d, 1H), 5.70
(d, 1H), 5.92 (s, 1H), 6.24 (s, 1H), 6.33 (s, 1H), 6.63 (d, 1H),
(47) Purification yield, ∼60%.
(48) Overall yields from VAHP .

2150 J . Org. Chem., Vol. 61, No. 6, 1996
Malabarba et al.
7.29 (d, 1H), 7.43 (d, 1H), 7.81 (s, 1H), 8.05 (s, 1H), 8.37 (d,
1H), 8.43 (d, 1H), 8.52 (br s, 1H).
for the determination of antimicrobial activity. Dr. L.
Colombo is acknowledged for his contribution (FAB MS
spectra) in the structure elucidation of some RH-
products.
BOC-TDTP DA (‚TFA): 1H NMR δ 2.78 (dd, 1H), 3.18 (dd,
1H), 3.42-3.60 (dd, 4H), 4.13 8m, 1H), 4.23 (dd, 1H), 4.41 (m,
1H), 4.46 (d, 1H), 4.63 (d, 1H), 4.78 (m, 1H), 5.14 (br s, 1H),
5.57 (s, 1H), 6.04 (s, 1H), 6.28 (s, 1H), 6.39 (s, 1H), 6.42 (s,
1H), 6.49 (s, 1H), 6.85 (d, 1H), 6.91 (s, 1H), 7.12 (br s, 1H),
7.27 (s, 1H), 7.56 (s, 1H), 7.89 (s, 1H), 8.10 (br s, NH3+), 8.50
(br s, NH3+), 8.60 (d, 1H), 9.01 (d, 1H), 9.18 (d, 1H).
Su p p or tin g In for m a tion Ava ila ble: Combustion analy-
sis data; abundance of fragment ions formed by sugar losses,
as observed in FAB MS of RH-CTA, RH-TB, and RH-TC;
assignments of the 1H NMR spectra of RH-TD, BOC and
CBZ-RH-TD, BOC-TDHP A, TDHP O, and BOC-TDHP A in
comparison with TD; NOE’s observed in RH-TD; assignments
of the 1H, 13C, and 15N NMR spectra of TDHP A and epi-
TDHP A; long-range couplings observed by HMBC, and NOE’s
for TDHP A and epi-TDHP A (13 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
Deter m in a tion of An tiba cter ia l Activity. Antibacterial
activity expressed as MIC (minimal inhibitory concentration
in µg/mL) was determined by the microdilution method in
Difco Todd-Hewitt broth (streptococci) or Oxoid Iso-Sensitest
broth (other organisms). The final inoculum was about 5 ×
105 cfu49 /mL. MIC was read as the lowest concentration which
showed no visible growth after 18-24 h incubation at 37 °C.
Ack n ow led gm en t. We are indebted to G. Tamborini
for acid-base titrations and IR spectra and to R. Scotti
(49) Colony forming unit.
J O941809V