α-Galactose based neoglycopeptides. Inhibition of verotoxin binding to globotriosylceramide

Article abstract of DOI:10.1016/S0968-0896(99)00226-6

Solution and solid phase strategies for the synthesis of α-galactose based neoglycopeptide derivatives 2-13 were developed. Neoglycopeptides generated were tested for the inhibition of verotoxin binding to globotriosylceramide (Gb3) using ELISA. Among all

Full text of DOI:10.1016/S0968-0896(99)00226-6

Bioorganic & Medicinal Chemistry 7 (1999) 2823±2833
ꢀ-Galactose Based Neoglycopeptides. Inhibition of Verotoxin
Binding to Globotriosylceramidey
Prabhat Arya, a,* Kristina M. K. Kutterer, a Huiping Qin, a Johanne Roby, a
Michael L. Barnes, a Shuqiong Lin, a Cli€ord A. Lingwood b and Markus G. Peter b
aChemical Biology Program, Steacie Institute for Molecular Sciences, National Research Council of Canada, Ottawa, ON,
Canada, K1A 0R6
bDepartment of Microbiology, The Hospital for Sick Children, Toronto, ON, Canada, M5G 1X8
Received 20 April 1999; accepted 15 July 1999
AbstractÐSolution and solid phase strategies for the synthesis of a-galactose based neoglycopeptide derivatives 2±13 were devel-
oped. Neoglycopeptides generated were tested for the inhibition of verotoxin binding to globotriosylceramide (Gb3) using ELISA.
Among all of the compounds tested, only the lipid derivatives of neoglycopeptides, 11, 12 and 13 were found to be inhibitors,
IC50=2.0 mM (11b and 12c) and 0.2 mM (11c and 13c). All of the inhibitors (11b, 11c, 12c and 13c) have a similar branching of the
two a-galactosyl units at the N-terminal glycine residue of a short peptide and a lipid moiety attached at the C-terminal site. Both of
these factors seem to be crucial for the inhibition. It is interesting to note that the inhibitors have only a portion of the natural
trisaccharide ligand. The secondary groups either may contribute in sub-site oriented interactions with the protein receptors or may
mimic the internal sugar units of the cell-surface ligand, Gb3. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
mimics which can be synthesized with relative ease and
possess acceptable pharmacokinetic properties. Mimics
of cell-surface carbohydrates capable of competing with
natural carbohydrate±protein interactions would be of
interest in potentially developing anti-microbial agents.9
The importance of cell-surface carbohydrates in initiat-
ing a wide variety of biological and pathological pro-
cesses is now well recognized.1±3 Since carbohydrate±
protein interactions are responsible for initiating the
early stages of several microbial infections, the use of a
soluble form of the natural ligands as anti-adhesives
may serve as useful agents in developing new therapies.
For the success of this approach, there are two major
challenges to consider and to overcome. First, it is well
accepted that the individual interactions between car-
bohydrate ligands and protein receptors are weak (i.e.
millimolar region), and in natural systems such interac-
tions are applied in a cooperative manner.4±6 Second, as
we have learned from the poor in vivo stability and
bioavailability of peptides and nucleotides as ther-
apeutics, carbohydrates are also likely to su€er from the
same drawbacks and are not ideal candidates.7,8 How-
ever, unlike with carbohydrates, tremendous progress
has been made in developing peptide and nucleotide
In the past several years, as an alternative to the use of
antibiotics for the treatment of bacterial infections, sig-
ni®cant e€orts have been made in designing soluble
multivalent carbohydrate ligands as anti-adhesive
agents. Some of these approaches include carbohydrate-
based polymers, dendrimers, as well as anchoring of
carbohydrate ligands on various solid supports.10±13
Due to the high molecular weight of glyco-based deri-
vatives, it is likely that such compounds are not rapidly
transported into various tissues and possess poor phar-
macokinetic properties.
Results and Discussion
In recent years, there has been much emphasis on the
synthesis of small molecules as analogues of cell-surface
carbohydrate ligands.14±17 With a similar goal of devel-
oping compounds that are capable of interfering with
the natural carbohydrate±protein interactions, we pro-
posed a ¯exible and control-oriented model for the
Key words: Neoglycopeptides; glycopeptides; glycomimetics; carbo-
hydrate-protein interactions; verotoxin; globotriosylceramide mimics.
* Corresponding author. Tel.: +1-613-993-7014; fax: +1-613-952-
0068; e-mail: prabhat.arya@nrc.ca
y NRCC publication No. 42181.
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00226-6

2824
P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
synthesis of neoglycopeptides.18,19 This model is based
upon presenting terminal or exposed glycosides on a
peptide/pseudopeptide template and can be achieved in
a controlled manner, to allow for a high degree of ¯ex-
ibility in the design. For the initial studies, a-galactose
derivatives were incorporated as the terminal glycoside
moiety of the neoglycopeptides. a-Galactose is the
terminal glycoside of the trisaccharide portion of glo-
botriosylceramide (Gb3, Fig. 1, 1). The glycosphingo-
lipid ligand, Gb3, is present on cell-surfaces of human
epithelial cells and is involved in the binding of Shiga
toxin and Shiga-like toxin receptors produced
by Shigella dysenteriae or by certain enterohemorrhagic
E. coli.20±25 Shiga and Shiga-like toxins have similar
structural and functional proteins, and are commonly
known as verotoxins. It is known that infection with
verotoxin producing E. coli leads initially to hemor-
rhagic colitis and further to hemolytic uremic syndrome
in humans.26,27 The binding subunits (e.g. B subunits,
see Fig. 1) of verotoxin are lectins (i.e. homopentameric
protein) that recognize the galabiose moiety (Gala1-
4Gal) of glycosphingolipid derivatives. It has been
shown that cells which do not express the galabiose
moiety on their surfaces are not a€ected by the toxins.28
Following the carbohydrate ligand±lectin receptor
recognition, in which the B subunits are involved, the
subsequent step is the endocytosis of the toxic subunit,
subunit A, of verotoxin.29,30 It is, therefore, reasonable
to believe that novel compounds that are capable of
interfering with cell-surface glycoconjugate ligand±toxin
receptor interactions may o€er new therapies for the
treatment of the disease initiated by ingestion of enter-
ohemorrhagic E. coli. It is interesting to note that the
binding constant for the soluble trisaccharide of Gb3 to
the soluble form of the pentameric subunit B of a simi-
to be inhibitors, even in the millimolar concentration
range.23
Design and synthesis of ꢀ-galactose based
neoglycopeptides
With the goal of developing inhibitors of verotoxin
binding to Gb3, several a-galactose based neoglycopep-
tides were synthesized (see Figs. 2 and 3) and tested for
the inhibition of verotoxin binding to Gb3 by ELISA.
Such assays have been described in detail in the litera-
ture.29 The ®rst strategy was developed to present
bivalent carbon linked a-galactoside derivatives on dif-
ferent peptide/pseudopeptide based templates (Fig. 2).
In the long term, this approach could be extended to
obtain bivalent forms of the terminal or exposed sac-
charide moiety on a library of peptide/pseudopeptide
templates. A combinatorial approach to optimize the
presentation of bivalent a-galactoside derivatives may
be used to reach two di€erent binding domains of a
multimeric protein receptor. We are working with C-
linked neoglycopeptides to bene®t from the advantages
of structural stability in vivo, in acidic and basic media,
and ease of synthesis.31 Bivalent, a-galactose based
neoglycopeptides (2±3, Fig. 2) were synthesized by
solution phase chemistry, whereas
a solid phase
approach was developed for the synthesis of compounds
4±7 (Fig. 2). As a second strategy, we plan to couple one
or two a-galactoside derivatives onto the N-terminal
side of short peptides (see compounds 9±13, Fig. 3). It is
anticipated that the additional peptide/pseudopeptide/
neoglycopeptide moiety may provide sub-site oriented
secondary interactions with the protein receptors.
Alternatively, the additional groups may mimic the
internal sugars of the cell-surface ligand, Gb3. In the
long term, this approach could also be developed in a
combinatorial manner. Part of the second strategy con-
sisted of coupling long-chain hydrophobic amines [i.e.
NH2(CH2)10Me and NH2(CH2)13Me] to the carboxyl
lar toxin is weak (e.g. millimolar region, Ka 0.5±1Â103
1
M
for subunit monomer).29 Also, derivatives of the
natural ligand which contain the galabiosyl moiety
attached to a short chain hydrocarbon were shown not
Figure 1. Model for the interaction of verotoxin to globotriosylcer-
amide (Gb3) on cell-surfaces.
Figure 2. Bivalent species of a-galactose based neoglycopeptides.

P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
2825
Figure 3. a-Galactose based neoglycopeptides.
terminal of the peptide moiety. The attachment of the
lipid moiety was expected to provide a ``cluster e€ect''
to enhance the inhibitory e€ect.
dipeptide 21 using a HATU coupling method to give
compound 22a. Following treatment with NaOMe/
MeOH, neoglycopeptide 2 was obtained from 22a in
60% yield, after puri®cation. Using the same approach,
compound 3 was obtained from 19b.
The synthesis of a-galactose based neoglycopeptides
9a±13a was developed. Compound 9a was synthesized
by attaching a C-linked a-galactose derivative having a
carboxyl group, 18, onto the side chain of the dipeptide
lysine-glycine as a key step. Compounds 10a and 11a
are branched a-galactose based neoglycopeptide deriva-
tives. Using a bis-reductive amination approach, two
units of a-C-galactosyl derivatives 28 were attached to
the side chain of lysine in the peptide backbone, leading
to the synthesis of compound 10a. For the synthesis of
compound 11a, sequential branching of the amino
group of glycine benzyl ester 35 was accomplished by
reductive amination using an a-C-galactosyl derivative
having an aldehyde group, 28, followed by amide
coupling of the monosubstituted glycine derivative to
an a-C-galactosyl derivative with a carboxyl group, 18.
Coupling of di-substituted a-C-galactosyl glycine 38 to
dipeptide 27 was achieved using standard peptide cou-
pling methods, generating neoglycopeptide 11a.
Solid phase, parallel synthesis of neoglycopeptides 4±7
(Scheme 2)
Fmoc±Rink Amide MBHA resin (Novabiochem, load-
ing 0.56 mmol/g, 23) was treated with 20% piperidine to
e€ect Fmoc removal. The free amino group was coupled
with neoglycopeptide building block 30 using a HATU
coupling method (4.0 eq 30, 4.0 eq HATU, 8.0 equiv
DIEA). The product, 24a, after treatment with 20%
piperidine, gave compound 24b with a free amino group
on the resin. At this stage, the resin was divided into two
portions. Each portion was subjected to parallel cou-
pling either with FmocHN-ala-gly-OH (25) or
FmocHN-pro-gly-OH (26) using a HATU method, fol-
lowed by the removal of the Fmoc group to give resin
A1 and A2. Resin A1 and A2 was divided further into
two portions, and was subjected to parallel couplings.
Neoglycopeptide 4 was obtained from the ®rst half of
the A1 batch in a sequential order: (i) coupling of the
building block, 30, using a HATU method, (ii) 20%
piperidine, (iii) N-acetylation by Ac2O, (iv) 95% TFA,
for the resin cleavage, and (iv) NaOMe in MeOH for
the deacetylation. Neoglycopeptide 4 was obtained in
42% overall yield after puri®cation by reverse phase
HPLC. From the second half of the A1 batch, neogly-
copeptide 2 was obtained as follows: (i) coupling of
Fmoc-ala-gly-OH (25) using a HATU method, (ii) 20%
piperidine, (iii) coupling of compound 30 using a
HATU method, (iv) 20% piperidine, (v) N-acetylation
by using Ac2O, (vi) cleavage from the resin (95% TFA),
and (vii) treatment with NaOMe in MeOH for the de-
acetylation. Neoglycopeptide 5 was obtained in 40%
overall yield after puri®cation by reverse phase HPLC.
Synthesis of neoglycopeptides 2 and 3 (Scheme 1)
The free amino groups of the benzyl ester of 3,5-dia-
minobenzoic acid (14) were coupled with BocHN-ala-
gly-OH (15) and BocHN-pro-gly-OH (16) respectively,
using a DCC±HOBt method, to give compounds 17a
and 17b in 90 and 93% yields after puri®cation.
Compounds 17a and 17b were treated with TFA to
remove the Boc-groups, and independently coupled
with a-galactosyl carboxylic acid 18 to obtain com-
pounds 19a (70%) and 19b (98%). Bivalent galactosyl
derivative 19a was hydrogenated, using 10% Pd/C as a
catalyst, to give the free carboxylic acid 20a. Acid 20a
was then coupled to the side-chain amino group of

2826
P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
Scheme 1. (a) 14 (1.0 mmol), Boc-NH-ala-gly-OH (15, 2.0 mmol) or Boc-NH-pro-gly-OH (16, 2.0 mmol), DCC (2.5 mmol), HOBt (2.5 mmol), DIEA
(2.5 mmol), CH2Cl2, room temp, 17±20 h, 90±93%. (b) (i) TFA/CH2Cl2, room temp, 1 h; (ii) 18 (2.0 mmol), DCC (2.5 mmol), HOBt (2.5 mmol),
DIEA (2.5 mmol), CH2Cl2, room temp, 17±20 h, 70% (19a), 98% (19b). (c) 10% Pd/C, EtOH, room temp, 1.5 h, 97%. (d) 21 (1.0 mmol), HATU
(1.4 mmol), DIEA (1.2mmol), DMF, room temp, 18h for 22a (20%) and 5 h for 22b (30%). (e) NaOMe, MeOH, room temp, 5 h, 60% (2) and 33% (3).
A similar strategy was utilized for the synthesis of neo-
glycopeptides 6 (51%) and 7 (46%) using resin batch
A2.
with aldehyde 28 gave mono-glycosylated compound 36
in 85% yields after puri®cation. Compound 36 was then
coupled with a-galactosyl carboxylic acid 18 using the
DCC/HOBt method. The coupled product, 37, was
obtained in 80% yield after puri®cation.
Synthesis of neoglycopeptides 8±10 (Scheme 3)
a-Galactoside amide derivatives 8a±c were synthesized
from compound 18 as follows. a-Galactosyl carboxylic
acid 18 was coupled with three di€erent amines (e.g.
NHEt2, NH(CH2)10Me and NH(CH2)13Me) using the
DCC/HOBt coupling method, followed by de-O-acet-
ylation (NaOMe/MeOH). Neoglycopeptides 9a±c were
obtained from dipeptide 27. Coupling of dipeptide 27
with a-galactosyl carboxylic acid 18 using the DCC/
HOBt method gave compound 29 in 87% yield, after
puri®cation. Neoglycopeptides 9a±c (60±73%) were
obtained from 29 in a number of steps: (i) hydrogena-
tion over 10% Pd/C to give the free carboxylic acid, (ii)
coupling to di€erent amines (DCC/HOBt), (iii) 20%
piperidine for Fmoc removal, (iv) Ac2O for N-acetyl-
ation, and (v) deacetylation of sugar hydroxyls using
NaOMe/MeOH. Dipeptide 27 was subjected to bis-
reductive amination with a-galactosyl aldehyde 28 to
obtain compound 32 in 65% yields. Neoglycopeptides
10a±c were obtained from 32 by the same series of steps
discussed above.
Hydrogenation of 37 gave the carboxylic acid 38. Cou-
pling of acid 38 with either dipeptide 27 or the benzyl
ester of p-aminobenzoic acid using DCC/HOBt gave
compounds 40 and 43 respectively. Neoglycopeptides
12a,c were obtained from compound 38 as follows.
Compound 38 was coupled with two di€erent amines
using DCC/HOBt to give 39a,c and then subjected to
de-O-acetylation conditions. To generate neoglycopep-
tides 11a±c and 13a,c, compounds 41 and 44 underwent
lipid coupling and de-O-acetylation in a manner analo-
gous to that described for compounds 12a,c.
Inhibition of verotoxin binding to Gb3 by ELISA
To screen for the inhibition of verotoxin binding to Gb3
using a-galactose based neoglycopeptides, a microtiter
plate assay was established. Gb3 isolated from human
kidney was immobilized onto 96-well plates and the
amount of verotoxin bound was measured against
increasing concentrations of neoglycopeptide deriva-
tives using a receptor±ELISA based system. None of the
a-galactose based bivalent neoglycopeptides (2±7, for
inhibition results see Table 1) showed any inhibition.
From all of the derivatives (8±11) tested, only the lipid
derivatives of neoglycopeptide 11 exhibited inhibitory
Synthesis of neoglycopeptides 11±13 (Scheme 4)
Neoglycopeptides 11±13 were synthesized as follows.
Monoreductive amination of glycine benzyl ester 35

P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
2827
Scheme 2. (a) (i) 20% piperidine in DMF; (ii) 4.0 equiv 30, 4.0 equiv HATU, 8.0 equiv DIEA in DMF, 16 h; (iii) same as (a) (i). (b) (i) 4.0 equiv 25,
4.0 equiv HATU, 8.0 equiv DIEA in DMF, 6 h; (ii) same as (a) (i). (c) (i) 4.0 equiv 26, 4.0 equiv HATU, 8.0 equiv DIEA in DMF, 6 h; (ii) same as (a)
(i). (d) (i) 4.0 equiv 30, 4.0 equiv HATU, 8.0 equiv DIEA in DMF, 16 h; (ii) same as (a) (i); (iii) Ac2O, DMF, 3 h; (iv) 95% TFA; (v) NaOMe, MeOH.
(e) (i) 4.0 equiv 25, 4.0 equiv HATU, 8.0 equiv DIEA in DMF, 6 h; (ii) same as (a) (i). (f) (i) 4.0 equiv 30, 4.0 equiv HATU, 8.0 equiv DIEA in DMF,
16 h; (ii) same as (a) (i); (d) iii±v. (g) (i) 4.0 equiv 26, 4.0 equiv HATU, 8.0 equiv DIEA in DMF, 16 h; (ii) same as (a) (i).
Scheme 3. (a) 18 (1.0 equiv), DCC (1.0 equiv), HOBt (1.0 equiv), DIEA (1.0 equiv), 87%. (b) H2, 10% Pd/C, 95% EtOH. (c) NHEt2 or
NH2(CH2)10Me or NH2(CH2)13Me (1.5 equiv), DCC (1.0 equiv), HOBt (1.0 equiv), DIEA (1.0 equiv), 60±73% for 9a±c from 29. (d) (i) 20% piper-
idine in DMF; (ii) Ac2O, DMAP. (e) NaOMe, MeOH. (f) 28 (2.2 equiv), THF, NaB(OAc)3H, AcOH, pH 5.0, 65%.

2828
P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
Scheme 4. (a) 28 (0.9 equiv), THF, NaB(OAc)3H, AcOH, pH 5.0, 0ꢀC, 85%. (b) 18 (1.0 equiv), DCC (1.0 equiv), HOBt (1.0 equiv), DIEA
(1.0 equiv), 80%. (c) H2 10% Pd/C, 95% EtOH. (d) NHEt2 or NH2(CH2)10Me or NH2(CH2)13Me (1.5 equiv), DCC (1.0 equiv), HOBt (1.0 equiv),
DIEA (1.0 equiv). (e) 27 (1.0 equiv), DCC (1.0 equiv), HOBt (1.0 equiv), DIEA (2.0 equiv), 83±85%. (f) (i) 20% piperidine in CH2Cl2; (ii) Ac2O,
DMAP. (g) NaOMe, MeOH. (h) Benzyl ester of p-aminobenzoic acid (1.0 equiv), DCC (1.0 equiv), HOBt (1.0 equiv), DIEA (1.0 equiv).
Table 1. ELISA-inhibition of verotoxin binding to immobilized Gb3
glycine structure, but to which an unnatural amino acid
residue, p-NH-Ph-CO-, has been added between glycine
and the lipid moiety. These two analogues were
designed to explore the importance of the tripeptide
moiety of neoglycopeptide 11c for the inhibition of ver-
otoxin/Gb3 bindings. As with compound 11a, com-
pound 12a did not show any inhibition. The lipid
derivative of the branched a-galactosyl glycine amino
acid, 12c, inhibited the Gb3±verotoxin interaction
(IC50=2 mM). Interestingly, removal of the lys-gly
dipeptide moiety from neoglycopeptide 11c showed a
weak inhibition. The addition of an unnatural amino
acid at the C-terminal of compound 12 enhanced the
inhibition (IC50=0.2 mM for compound 13c). Com-
pound 13c has a similar inhibitory e€ect to that shown
by neoglycopeptide 11c but it is a simpler compound,
with a more facile synthesis.
Compound
IC50
2±7
No inhibition
No inhibition
No inhibition (11a), 2.0 mM (11b), 0.2 mM (11c)
No inhibition (12a), 2.0 mM (12c)
No inhibition (13a), 0.2 mM (13c)
8a±c to 10a±c
11a±c
12a,c
13a,c
e€ects (IC50=2 mM for 11b and 0.2 mM for 11c) on
verotoxin binding to Gb3. It is interesting to note that a
subtle change in the lipid moiety can in¯uence the inhi-
bitory action. Lipid derivatives 8a±c, 9a±c and 10a±c
were not found to be inhibitors. It seems likely that the
appropriate branching of the two a-galactosyl units and
the attachment of the lipid moiety in compounds 11b
and 11c are critical for the inhibition. Alternatively, the
secondary groups may mimic the internal sugars of the
natural ligand, Gb3.
Conclusion
Several solution and solid phase approaches for the
synthesis of a-galactose based neoglycopeptides were
developed and neoglycopeptides generated were tested
for the inhibition of verotoxin binding to Gb3. Some of
these approaches can be modi®ed for the assembly of
combinatorial libraries of neoglycopeptides. Of all the
compounds tested, 11c and 13c were found to be the
Lead compound derivatives and inhibition studies
To determine the elements of compound 11c that are
necessary for the inhibitory e€ect, neoglycopeptides
12a,c and 13a,c were synthesized. Compound 12c is the
lipid derivative of the branched a-galactosyl based gly-
cine. Analogue 13c has the same branched a-galactosyl

P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
2829
best inhibitors (IC50=0.2 mM). To our knowledge, this
is the ®rst example of small-molecule inhibitors of
verotoxin binding to Gb3 that have only a portion of
the recognition element of the trisaccharide ligand, and
are capable of exhibiting inhibitory e€ects at a sub-mil-
limolar concentration. Although it is too early to pro-
vide any explanation for this behavior, it is reasonable
to believe that the appropriate positioning of the two a-
C-galactose derivatives may play a crucial role in gen-
erating this e€ect. The secondary groups may contribute
to the overall binding by participating in sub-site orien-
ted interactions with the protein receptors or may mimic
the internal sugars of the cell-surface ligand, Gb3. Fur-
ther work is in progress to develop a better under-
standing of the importance of the peptide backbone and
of the lipid moiety in their assistance in amplifying
inhibitory e€ects.
Compounds 22a and 22b. To a solution of the carboxylic
acid derivative 20a (1.0 mmol) in DMF (20 mL) was
added HATU (1.2 mmol), DIEA (1.2 mmol) and dipep-
tide 21 (1.0 mmol). The mixture was stirred at room
temperature for 18 h. Following the work-up, puri®ed
product 22a was obtained in 20% yield after ¯ash
chromatography. Compound 22b was synthesized using
1
similar reaction conditions. 22a: H NMR (400 MHz,
MeOH-d4): d 1.30±2.20 (m, 39H), 2.45±3.10 (m, 4H),
3.30±3.45 (m, 2H), 3.80±4.45 (m, 15H), 4.65±4.75 (m,
2H), 5.00±5.50 (m, 8H), 7.15±7.35 (m, 5H), 7.80±8.32
(m, 3H). 13C NMR (100 MHz, MeOH-d4): d 16.2, 16.4,
19.6, 19.7, 21.4, 23.2, 29.1, 31.0, 33.6, 42.7, 42.9, 43.3,
50.4, 54.8, 60.6, 61.5, 68.1, 68.3, 68.8, 70.3, 74.1, 74.7,
114.6, 114.9, 127.0, 129.1, 136.3, 138.8, 139.1, 168.9,
174.8. LRMS (Electrospray, positive ion mode, m/z) for
C66H88 N10O28: 1469.3 (MH+). 22b: 1H NMR
(400 MHz, MeOH±d4): d 1.35±2.27 (m, 41H), 2.45±3.00
(m, 4H), 3.20±3.55 (m, 2H), 3.65±4.48 (m, 21H), 5.10±
5.55 (m, 8H), 7.15±7.35 (m, 5H), 7.85±8.25 (m, 3H). 13C
NMR (100 MHz, MeOH-d4): d 19.6, 19.7, 19.8, 21.4,
23.2, 25.0, 29.2, 29.5, 31.0, 32.5, 39.6, 42.7, 42.9, 54.8,
61.2, 61.5, 68.3, 69.2, 72.4, 74.0, 113.6, 114.9, 127.1,
128.4, 136.2, 139.1, 168.4, 174.4. LRMS (Electrospray,
positive ion mode, m/z) for C70H92N10O28: 1521.7
(MH+). 2: 1H NMR (400 MHz, D2O): d 1.45 (br d,
J=7.2 Hz, 8H), 1.59±1.67 (m, 2H), 1.72±1.85 (m, 2H),
1.95 (s, 3H), 2.62±2.68 (dd, J=14.8 Hz, 2H), 2.74±2.80
(dd, J=14.8 Hz, 2H), 3.38 (t, J=6.8 Hz, 2H), 3.62±3.79
(m, 6H), 3.88±3.92 (m, 4H), 3.98 (br s, 2H), 4.01±4.04
(m, 2H), 4.05±4.11 (m, 4H), 4.25 (t, J=7.1 Hz, 1H),
4.26±4.39 (m, 4H), 4.47±4.53 (m, 2H), 7.23 (d,
J=7.4 Hz, 2H), 7.31 (t, J=7.3 Hz, 1H), 7.37 (t,
J=7.5 Hz, 2H), 7.65 (d, J=1.7 Hz, 2H), 7.83±7.89 (m,
1H). 13C NMR (100 MHz, D2O): d 17.2, 22.0, 23.1,
28.7, 32.0, 33.0, 40.5, 43.6, 43.9, 51.3, 55.1, 62.1, 73.7,
118.0, 120, 127.9, 128.3, 129.6, 136.0, 141.0, 170.7,
176.8. LRMS (Electrospray, H2O, positive ion mode, m/
z) for C50H72 N10O20: 1132.3 (M+), 1155.5 (MNa+). 3:
1H NMR (400 MHz, D2O): d 1.25±1.45 (m, 2H), 1.50±
1.60 (m, 2H), 1.62±1.80 (m, 2H), 1.84 (s, 3H), 1.85±2.18
(m, 6H), 2.20±2.35 (m, 2H), 2.40±3.05 (m, 4H), 3.25±
3.55 (m, 31H), 7.10±7.90 (m, 8H). 13C NMR (100 MHz,
D2O): d 21.9, 22.6, 24.8, 28.2, 30.0, 30.7, 31.4, 40.0, 43.1,
43.4, 48.7, 61.4, 68.2, 69.1, 70.2, 72.9, 73.6, 117.0, 117.2,
127.3, 127.8, 129.1, 135.9, 138.0, 170.2, 175.7. LRMS
(Electrospray, positive ion mode, m/z) for C54H76
N10O20: 1185.6 (M+), 1207.5 (MNa+).
Experimental
Synthesis of 2 and 3 (Scheme 1). Compounds 17a and
17b. To a solution of Boc-NH-ala-gly-OH (15, 2.0
mmol) in CH2Cl2 (50 mL) was added DCC (2.5 mmol)
and HOBt (2.5 mmol). The reaction mixture was stirred
for 15 min, followed by the addition of the benzyl ester
of 3,5-diaminobenzoic acid 14 (1.0 mmol) and DIEA
(2.5 mmol). It was further stirred at room temperature
for 17 h, solvent evaporated in vacuo and the product
puri®ed by ¯ash chromatography over silica gel. Com-
pound 17a was obtained in 90% yield after the removal
of the Boc group. Similar reaction conditions were uti-
lized to obtain compound 17b from Boc-NH-pro-gly-
OH (16)
Compounds 20a and 20b. DCC (2.5 mmol) and HOBt
(2.5 mmol) were added to a solution of a-galactosyl
carboxylic acid, 18 (2.0 mmol) in CH2Cl2 (50 mL) and
stirred for 15 min. To this mixture, compound 17a
(1.0 mmol) and DIEA (2.5 mmol) was added and then
stirred for 17 h. After the removal of the solvent in
vacuo, the product was puri®ed (19a, 70% yield), and
subjected to hydrogenation to remove the benzyl group
leading to compound 20a in 97% yield. Synthesis of
compound 20b was achieved in a similar manner. 19a: 1H
NMR (400 MHz, MeOH-d4): d 1.35±1.43 (d, J=7.0 Hz,
6H), 1.91, 1.98, 2.05, 2.07 (4s, 24H), 2.45±2.90 (m, 4H),
3.85±4.38 (m, 12H), 4.63±4.77 (m, 2H), 5.25±5.50 (m,
8H), 7.30±7.50 (m, 5H), 8.18±8.65 (m, 3H). 13C NMR
(100 MHz, MeOH-d4): d 14.9, 18.3, 18.4, 32.4, 42.0, 49.3,
60.1, 65.7, 66.8, 67.8, 69.1, 71.1, 72.8, 73.4, 114.9, 115.8,
126.8, 127.0, 127.4, 130.0, 135.5, 138.0, 165.1, 167.7,
169.1, 169.4, 170.6, 173.6. LRMS (Electrospray, positive
ion mode, m/z) for C56H70N6O26: 1243.4 (MH+). 20a:
1H NMR (400 MHz, MeOH-d4): d 1.25±1.40 (m, 8H),
1.85±2.25 (m, 24H), 2.60±2.90 (m, 4H), 3.15±4.35 (m,
18H), 5.20±5.52 (m, 8H), 7.30±7.60 (m, 5H), 7.90±8.60
(m, 3H). 13C NMR (100 MHz, MeOH-d4): d 12.0, 17.2,
17.5, 19.8, 20.5, 32.5, 43.5, 48.5, 49.2, 55.2, 61.3, 62.1,
68.2, 70.8, 114.5, 115.6, 127.0, 138.2, 165.3, 173.5.
LRMS (Electrospray, positive ion mode, m/z) for
C60H74N6O26: 1295.4 (MH+).
Synthesis of neoglycopeptides 4±7 (Scheme 2). Neoglyco-
peptide benzyl ester 29. Dipeptide 27 (10.0 mmol), DCC
(10.0 mmol), HOBt (10.0 mmol) and DIEA (10.0 mmol)
were added to a solution of a-galactosyl carboxyl acid
(18, 10.0 mmol), in CH2Cl2 (100 mL). The solution was
stirred at the room temperature for 20 h. The solvent
was evaporated in vacuo, and the residue puri®ed over
silica gel by ¯ash column chromatography to obtain
compound 29 in 87% yield. 1H NMR (600 MHz,
CDCl3): d 1.40±1.90 (m, 6H), 2.04±2.10 (2s, 12H), 2.37±
2.60 (ddd, J=4.1, 9.8, 15.3 Hz, 2H), 3.25 (m, 2H), 4.00±
4.29 (m, 6H), 4.40 (d, J=6.2 Hz, 2H), 4.70 (m, 1H),
5.00±5.60 (m, 6H), 5.64 (d, J=7.4 Hz, 1H), 6.63 (bs,
1H), 6.68 (bs, 1H), 7.14±7.32 (m, Ar-H, 13H); 13C NMR

2830
P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
(CDCl3): d 172.3, 170.7, 169.9, 169.8, 169.6, 156.3,
143.7, 141.2, 135.0, 129.0, 128.5, 127.7, 127.0, 126.7,
125.2, 125.0, 119.9, 69.2, 69.0, 68.7, 67.8, 67.0, 61.2,
54.3, 47.0, 41.2, 38.8, 31.2, 31.1, 28.6, 22.2, 20.6. LRMS
(FAB, TG/Gly 1:1+1% HCl, positive ion mode, m/z)
for C46H53N3O15: 910.2 (MNa+).
2H), 4.40±4.50 (m, 4H). 13C NMR (100 MHz, D2O): d
22.1, 22.7, 24.7, 28.2, 29.9, 30.8, 32.7, 39.5, 42.1, 42.2,
42.5, 42.8, 47.4, 54.2, 61.0, 61.2, 67.9, 69.1, 70.0, 72.8,
73.3, 169.6, 169.7, 172.0, 173.8, 174.5, 174.8, 175.0,
175.1, 175.5. LRMS (Electrospray, H2O, positive ion
mode, m/z) for C48H79N11O21: 1146.5 (MH+), 1168.4
(MNa+), 574.0 (M+/2).
Neoglycopeptide building block 30. A solution of benzyl
ester 29 in 95% EtOH was hydrogenated at atmospheric
pressure in the presence of 10% Pd/C (10 mol%) for
45 min. The catalyst was removed by ®ltration through
a pad of Celite, providing compound 30, after evapora-
Synthesis of compound 33 from dipeptide 27 (Scheme 3).
Neoglycopeptide benzyl ester 32. a-Galactosyl aldehyde
(28, 2.2 mmol) was added to a solution of dipeptide 27
(1.0 mmol) in THF (20 mL). The solution was acidi®ed
with AcOH to pH 5.0. After stirring at 0ꢀC for 20 min,
NaB(OAc)3H (2.5 mmol) was added to the solution and
stirred at room temp for 12 h. Puri®ed compound 32
was obtained in 65% yield. LRMS (Electrospray) for
C62H78N3O23: 1232.4 (MH+). 1H NMR (600 MHz,
CDCl3): d 1.40±2.00 (m, 12H), 2.00±2.10 (4s, 24H),
2.40±2.53 (m, 6H), 4.00±4.43 (m, 12H), 5.06±5.10 (m,
1H), 5.17 (s, 2H), 5.20±5.28 (m, 4H), 5.40±5.49 (m, 4H),
7.26±7.78 (m, 13H); 13C NMR (CDCl3): d 172.1, 170.9,
170.3, 169.7, 169.6, 169.1, 156.2, 143.8, 143.6, 141.1,
135.0, 128.4, 128.3, 128.1, 127.5, 126.9, 125.0, 121.9,
119.0, 68.0, 67.9, 67.1, 66.9, 61.1, 60.7, 54.1, 49.7, 49.1,
46.9, 33.5, 32.2, 28.4, 25.3, 22.4, 21.8, 21.2, 21.1, 20.4,
20.1.
1
tion of the solvent, in 97% yield. H NMR (400 MHz,
D2O): d 1.30±1.47 (m, 2H), 1.47±1.57 (m, 2H), 1.65±1.88
(m, 2H), 2.04 (s, 3H), 2.53±2.71 (m, 2H), 3.19 (t,
J=6.5 Hz, 2H), 3.64±3.86 (m, 4H), 3.89 (d, J=9.2 Hz,
2H), 3.94±4.06 (m, 2H), 4.19±4.25 (m, 1H), 4.43±4.51
(m, 1H). 13C NMR (100 MHz, D2O): d 21.9, 22.7, 28.3,
30.7, 32.7, 39.5, 42.4, 54.6, 61.2, 68.0, 69.1, 70.0, 72.8,
73.3, 173.9, 174.5, 175.1, 175.5. LRMS (Electrospray,
H2O, positive ion mode, m/z) for C18H32N4O9: 449.3
(MH+), 471.3 (MNa+).
1
Neoglycopeptide 4. H NMR (400 MHz, D2O): d 1.00±
1.90 (m, 12H), 1.41 (d, J=7 Hz, 3H), 2.26 (s, 3H), 2.54±
2.82 (m, 4H), 3.20 (t, J=7 Hz, 4H), 3.58±3.72 (m, 4H),
3.72±3.78 (m, 2H), 3.78±3.88 (m, 2H), 3.88±3.92 (m,
2H), 3.92±4.06 (m, 8H), 4.22±4.38 (m, 3H), 4.46±4.54
(m, 2H). 13C NMR (100 MHz, D2O): d 16.8, 22.0, 22.7,
28.2, 30.7, 32.7, 39.5, 42.5, 42.8, 50.2, 54.4, 61.2, 67.9,
69.1, 70.0, 72.8, 73.3, 171.7, 172.1, 173.8, 174.5, 174.9,
175.1, 175.6, 176.0. LRMS (Electrospray, H2O, positive
ion mode, m/z) for C39H67N9O19: 966.4 (MH+), 988.5
(MNa+), 483.9 (M+/2).
Synthesis of neoglycopeptides 8a±c to 10a±c from 18, 30
and 33 (Scheme 3). General procedure. To a solution of
the carboxylic acid (1.0 mmol) in CH2Cl2 was added
DCC (1.0 mmol) and HOBt (1.0 mmol). The reaction
mixture was then left to stir at room temperature for
15 min, at which time the amine (2.0 mmol) component
was then added and the reaction mixture stirred a fur-
ther 4±6 h. In the case of NHEt2, 1.0 equiv of DIEA was
added to the reaction mixture. The solvent was evapo-
rated in vacuo and the product puri®ed by ¯ash chro-
matography over silica gel. The amine derivative was
treated with 20% piperidine in DMF (30 mL) for 1 h, at
which point TLC indicated complete removal of the
Fmoc protecting group. Acetic anhydride and a cataly-
tic amount of DMAP were then added to the reaction
mixture and stirred for a further 4±5 h. After removal of
DMF, the reaction mixture was puri®ed by ¯ash chro-
matography over silica gel. The fully acetylated product
was dissolved in MeOH, 0.5 M NaOMe solution was
added until the pH of the solution was 10±11, and stir-
red at room temperature for 3±4 h. The solution was
neutralized with solid CO2, generating neoglycopeptides
8a±c to 10a±c in 60±75% yields after puri®cation by
1
Neoglycopeptide 5. H NMR (400 MHz, D2O): d 0.80±
1.90 (m, 12H), 1.39 (d, J=7.3 Hz, 6H), 2.03 (s, 3H),
2.52±2.72 (m, 4H), 3.19 (t, J=6.8 Hz, 4H), 3.62±3.72
(m, 4H), 3.72±3.78 (m, 2H), 3.78±3.88 (m, 2H), 3.88±
3.92 (m, 2H), 3.92±4.04 (m, 10H), 4.22±4.38 (m, 4H),
4.44±4.52 (m, 2H). 13C NMR (100 MHz, D2O): d 16.7,
22.0, 22.7, 28.2, 30.9, 32.7, 39.5, 42.6, 42.9, 50.3, 54.4,
61.2, 67.9, 69.1, 70.0, 72.8, 73.3, 171.7, 171.8, 172.1,
173.8, 174.5, 174.9, 175.1, 175.6, 176.0. LRMS (Elec-
trospray, H2O, positive ion mode, m/z) for
C44H75N11O21: 1094.4 (MH+), 1116.5 (MNa+).
1
Neoglycopeptide 6. H NMR (400 MHz, D2O): d 0.90±
2.44 (m, 16H), 2.03 (s, 3H), 2.52±2.82 (m, 4H), 3.12±
3.22 (m, 4H), 3.54±3.70 (m, 8H), 3.70±3.78 (m, 2H),
3.78±3.86 (m, 4H), 3.86±4.06 (m, 6H), 4.06±4.16 (m,
1H), 4.22±4.32 (m, 2H), 4.41±4.52 (m, 2H); 13C NMR
(100 MHz, D2O): d 22.1, 22.7, 24.8, 28.2, 29.7, 30.6,
31.2, 32.7, 39.5, 42.1, 42.5, 42.8, 47.5, 54.0, 54.2, 61.2,
67.9, 69.1, 70.0, 72.8, 73.4, 169.7, 172.0, 173.8, 174.5,
174.7, 175.0, 175.5. LRMS (Electrospray, H2O, positive
ion mode, m/z) for C41H69N9O19: 992.3 (MH+), 1014.4
(MNa+).
1
reverse phase HPLC. 8a: H NMR (400 MHz, D2O): d
0.98 (t, J=7.2 Hz, 3H), 1.09 (t, J=7.2 Hz, 3H), 2.62 (a
of ABX, J=9.3 Hz, J= 15.9 Hz, 1H), 2.71 (b of ABX,
J=4.4 Hz, J= 15.9 Hz, 1H), 3.26 (q, J=7.2 Hz, 2H),
3.34 (q, J=7.2 Hz, 2H), 3.56 (A of ABX, J=7.6 Hz,
J= 11.6 Hz, 1H), 3.65 (B of ABX, J=4.4 Hz,
J= 11.6 Hz, 1H), 3.69 (dd, J=3.2 Hz, 1H), 3.72±3.74
(m, 1H), 3.85±3.90 (m, 1H), 3.90 (dd, J=9.6 Hz, 1H),
4.46 (dt, J=5.6 Hz, 1H); 13C NMR (100 MHz, D2O): d
12.4, 13.7, 29.6, 41.4, 43.4, 61.0, 68.3, 68.9, 70.1, 72.7,
73.4, 172.8; LRMS (Electrospray, H2O, positive ion
mode, m/z) for C12H23NO6: 278.1 (MH+), 300.1
1
Neoglycopeptide 7. H NMR (400 MHz, D2O): d 1.24±
1.60 (m, 8H), 1.60±2.44 (m, 12H), 2.03 (s, 3H), 2.50±
2.82 (m, 4H), 3.12±3.22 (m, 4H), 3.48±3.76 (m, 12H),
3.76±3.90 (m, 4H), 3.90±4.16 (m, 8H), 4.22±4.32 (m,
1
(MNa+). 8c: H NMR (400 MHz, CD3OD): d 0.92 (t,

P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
2831
J=6.8 Hz, 3H), 1.31 (br s, 22H), 1.49±1.55 (m, 2H), 2.49
(A of ABX, J=3.7 Hz, J= 15.2 Hz, 1H), 2.63 (B of
ABX, J=10.5 Hz, J= 15.2 Hz, 1H), 3.19 (t, J=7.1 Hz,
2H), 3.64±3.70 (m, 2H), 3.79±3.83 (m, 1H), 3.86±3.92
(m, 2H), 3.97 (t, J=3.0 Hz, 1H), 4.43 (dt, J=3.4 Hz,
J=3.7 Hz, J=10.4 Hz, 1H); 13C NMR (100 MHz,
CD3OD): d 13.4, 22.7, 27.0, 29.4, 29.5, 29.7, 29.8, 32.1,
33.5, 39.5, 60.6, 68.6, 69.2, 71.1, 71.4, 73.9, 173.1;
LRMS (FAB, TG/Gly 1:1+1% HCl, positive ion
mode, m/z) for C22H43NO6: 418.2 (MH+). 9c: 1H NMR
(400 MHz, CD3OD): d 0.92 (t, J=6.7 Hz, 3H), 1.25±
1.64 (m, 29H), 1.64±1.76 (m, 1H), 1.76±1.88 (m, 1H),
2.03 (s, 3H), 2.50 (A of ABX, J=3.5, 15.1 Hz, 1H), 2.64
(B of ABX, J=10.7, 15.1 Hz, 1H), 3.16±3.29 (m, 4H),
3.63±3.98 (m, 7H), 4.12±4.19 (m, 1H), 4.40 (dt, J=3.5,
10.7 Hz, 1H); 13C NMR (100 MHz, CD3OD): d 13.4,
21.5, 22.7, 23.1, 26.9, 29.0, 29.3, 29.5, 29.7, 29.8, 30.9,
32.1, 33.4, 39.0, 39.5, 42.5, 54.8, 60.8, 68.7, 69.1, 71.1,
71.4, 73.9, 170.4, 173.0, 173.3, 174.2; LRMS (Electro-
spray, H2O, positive ion mode, m/z) for C32H60N4O9:
645.5 (MH+), 323.4 (MH+/2). 10c: 1H NMR
(400 MHz, D2O): d 0.76 (t, J=6.6 Hz, 3H), 1.17 (bs,
17H), 1.38±1.45 (m, 4H), 1.61±1.81 (m, 4H), 1.92±2.22
(m, 4H), 1.95 (s, 3H), 3.03±3.18 (m, 4H), 3.50±3.72 (m,
11H), 3.75 (s, 2H), 3.84±3.94 (m, 4H), 4.00±4.08 (m,
2H), 4.10±4.15 (m, 1H); 13C NMR (100 MHz, D2O): d
13.8, 19.7, 22.0, 22.4, 22.5, 23.3, 26.3, 28.6, 28.7, 28.9,
29.0, 29.1, 29.2, 30.5, 31.6, 39.7, 42.9, 51.1, 53.4, 54.4,
61.4, 68.1, 69.1, 70.0, 73.1, 73.2, 171.2, 175.0; LRMS
(Electrospray, H2O, positive ion mode, m/z) for
C37H70N4O13: 779.6 (MH+), 390.5 (MH+/2).
centrated and puri®ed by ¯ash chromatography (CH2Cl2
to 5% MeOH in CH2Cl2) to give compound 39 in 85%
yield. This was subjected to deacetylation reaction con-
ditions followed by puri®cation using reverse phase
HPLC to obtain pure compound 12c in 86% yield. 12c:
1H NMR (400 MHz, D2O): d 0.91 (t, J=6.8 Hz, 3H),
1.29 (bs, 22H), 1.42±1.67 (m, 2H), 1.69±2.06 (m, 2H),
2.53±2.98 (m, 2H), 3.16±3.35 (m, 2H), 3.36±3.55 (m,
2H), 3.56±4.27 (m, 14H); LRMS (Electrospray, H2O,
positive ion mode, m/z) for C32H60N3O12: 665.3
(MH+).
Compound 40. To
a solution of compound 38
(1.0 mmol) in CH2Cl2 (10 mL) was added DCC
(1.0 mmol) and HOBt (1.0 mmol) and stirred for 15 min.
A solution of the dipeptide (1.0 mmol) and DIEA
(2.0 mmol) in CH2Cl2 (2 mL) was then added. Neogly-
copeptide derivative 40 was obtained in 83% yield after
puri®cation. LRMS (Electrospray) for C64H79N4O25:
1
1303.4 (MH+); H NMR (600 MHz, CDCl3): d 1.04±
1.76 (m, 10H), 1.80±1.99 (m, 4H), 2.0, 2.02, 2.03, 2.06,
2.08, 2.10, 2.12 (7s, 24H), 2.92±3.61 (m, 7H), 3.14±3.72
(m, 14H), 5.13±5.17 (m, 3H), 5.22±5.28 (m, 1H), 5.30±
5.35 (m, 1H), 5.39±5.41 (m, 2H), 7.24±7.76 (m, 13H);
13C NMR (CDCl3): d 172.4, 170.4, 169.9, 169.6, 169.5,
168.9, 156.1, 143.6, 141.1, 135.1, 128.4, 128.1, 127.5,
126.9, 126.2, 125.0, 119.8, 70.7, 69.1, 68.6, 68.3, 68.2,
67.7, 66.3, 66.2, 62.3, 61.3, 60.4, 54.2, 51.2, 48.9, 46.9,
45.2, 41.1, 33.6, 32.0, 31.7, 28.3, 24.7, 23.5, 22.1, 20.5.
Synthesis of 11a±c was achieved from compound 40 in a
number of steps. 11a: 1H NMR (400 MHz, D2O): d
0.97 (t, J=7.1 Hz, 3H), 1.08 (t, J=7.1 Hz, 3H), 1.24±
1.30 (m, 2H), 1.37±1.43 (m, 2H), 1.58±1.76 (m, 2H),
1.83±1.93 (m, 2H), 1.93 (s, 3H), 2.58±2.80 (m, 2H),
3.07±3.14 (m, 2H), 3.24 (q, J=7.1 Hz, 2H), 3.26 (q,
J=7.1 Hz, 2H), 3.42±3.52 (m, 2H), 3.54±3.74 (m, 9H),
3.83±3.98 (m, 4H), 3.94 (s, 2H), 3.98 (s, 2H), 4.18 (dd,
J=5.2 and 8.8 Hz, 1H), 4.47 (t, J=6.1 Hz, 1H); 13C
NMR (100 MHz, D2O): d 12.4, 13.2, 22.0, 22.7, 23.2,
28.3, 30.5, 31.0, 39.4, 41.1, 41.4, 42.2, 47.1, 50.2, 54.1,
61.0, 61.7, 68.1, 68.8, 69.5, 70.0, 70.2, 72.4, 72.6, 73.2,
73.4, 73.8, 169.3, 171.1, 174.3, 174.8, 174.9; LRMS
(Electrospray, H2O, positive ion mode, m/z) for
C32H57N5O15: 752.3 (MH+), 774.2 (MNa+), 376.8
Synthesis of 11a±c and 12a,c (Scheme 4). Compound 36.
a-Galactosyl aldehyde (28, 0.9 mmol) was added to a
solution of glycine benzyl ester 35 (1.0 mmol) in THF
(20 mL). The solution was acidi®ed with AcOH to pH
5.0. After stirring at 0ꢀC for 20 min, NaB(OAc)3H
(1.2 mmol) was added to the solution and stirred at
room temp for 12 h. Puri®ed compound 36 was
obtained in 50% isolated yield.
Compound 37. To a solution of a-galactosyl acid 18
(5.0 mmol) in CH2Cl2 (40 mL) was added DCC
(5.0 mmol) and HOBt (5.0 mmol). A solution of the
glycine benzyl ester derivative 36 (5.0 mmol) and DIEA
(5.0 mmol) in CH2Cl2 (10 mL) was then added and the
reaction mixture stirred at room temp for 16 h. DCU
produced was then ®ltered o€ and the solution evapo-
rated to dryness. Further puri®cation was e€ected by
¯ash chromatography, giving compound 37 in 80%
yield. As in previous cases, compound 37 was subjected
to hydrogenation to obtain the carboxylic derivative
38.
1
(MH+/2). 11b: H NMR (400 MHz, D2O): d 0.74 (t,
J=7.1 Hz, 3H), 1.15 (br s, 16H), 1.29±1.39 (m, 6H),
1.62±1.80 (m, 2H), 1.87±1.92 (m, 5H), 2.61±2.86 (m,
2H), 3.07±3.16 (m, 4H), 3.42±3.79 (m, 12H), 3.83±4.02
(m, 7H), 4.07 (br s, 1H), 4.45±4.52 (m, 1H); 13C NMR
(100 MHz, D2O): d 14.2, 22.3, 23.0, 27.3, 28.6, 29.2,
29.8, 30.1, 30.8, 32.3, 39.9, 42.9, 45.4, 47.0, 50.1, 54.4,
61.0, 61.7, 68.2, 68.5, 68.7, 69.5, 70.1, 70.3, 72.6, 73.1,
73.9, 78.9, 170.8, 170.9, 173.9, 174.2, 174.9; LRMS
(Electrospray, H2O, positive ion mode, m/z) for
1
Compounds 12a and 12c from 38. The synthesis of 12a
and 12c was accomplished from compound 38 in a ser-
ies of steps as discussed earlier.
C39H71N5O15: 850.4 (MH+), 426.0 (MH+/2). 11c: H
NMR (400 MHz, D2O): d 0.74 (t, J=6.5 Hz, 3H), 1.14±
1.19 (m, 24H), 1.27±1.44 (m, 4H), 1.51±1.79 (m, 4H),
1.90 (s, 3H), 2.61±2.88 (m, 2H), 3.03±3.09 (m, 4H),
3.33±3.73 (m, 12H), 3.86±3.96 (m, 7H), 4.06 (br s, 1H),
4.45±4.52 (m, 1H); 13C NMR (100 MHz, D2O): d 14.2,
21.8, 22.4, 22.6, 23.0, 27.4, 29.3, 30.0, 30.3, 30.5, 32.4,
39.9, 42.9, 44.9, 47.0, 50.1, 55.0, 61.0, 61.8, 68.2, 68.6,
68.7, 69.5, 70.2, 70.3, 72.6, 73.1, 73.6, 78.9, 170.7, 170.9,
Compound 12c. To a solution of acid 38 (0.062 mmol) in
CH2Cl2 (2 mL) was added DCC (0.093 mmol) and
HOBT (0.093 mmol). After stirring for 0.5 h, 1-tetra-
decylamine (0.093 mmol) was added and the reaction
stirred overnight. The reaction mixture was ®ltered, con-

2832
P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
173.8, 174.2, 174.8; LRMS (Electrospray, H2O, positive
ion mode, m/z) for C42H77N5O15: 892.4 (MH+), 446.9
(MH+/2).
50 mL/well of the pre-incubated inhibitor±VT dilutions
were added in triplicates, plates covered with para®lm
and left on the bench for 120 min.
Synthesis of 13a,c from 38 (Scheme 4). Compound 43.
Diisopropylethylamine (0.38 mmol) was added to a
solution of the benzyl ester (0.19 mmol), acid 38
(0.19 mmol), and HATU (72.3 mg, 0.19 mmol) in dry
CH2Cl2 (4 mL). After stirring overnight, the reaction
mixture was puri®ed by ¯ash chromatography (CH2Cl2
to 5% MeOH in CH2Cl2) to give compound 43 (97%
yield). 43: 1H NMR (400 MHz, CDCL3): d 1.23 (s, 1H),
1.61 (s, 3H), 1.85±2.12 (m, 24H), 2.52±2.85 (m, 2H),
3.38-3.61 (m, 2H), 3.86±4.34 (m, 9H), 7.28±8.06 (m,
9H); LRMS (Electrospray, H2O, positive ion mode, m/z)
for C48H58N3O22: 1015.4 (MH+), 527.3 (MH+/2). 13a:
1H NMR (400 MHz, D2O): d 1.00 (t, J=7.1 Hz, 3H),
1.13 (t, J=7.1 Hz, 3H), 1.79±2.03 (m, 2H), r1 2.68 and r2
2.85 (d's, J=6.0 Hz, J=10.0 Hz for two rotamers, r1
and r2, 2H), 3.21 (q, J=7.1 Hz, 2H), 3.42 (q, J=7.1 Hz,
2H), 3.47±3.82 (m, 10H), 3.83±4.08 (m, 5H), r1 4.17 and
r2 4.37 (bs's, 2H), 4.46±4.57 (m, 1H), 7.28±7.36 (m, 2H),
7.42±7.51 (m, 2H); 13C NMR (100 MHz, D2O): d 12.3,
13.5, 22.2, 23.3, 29.9, 40.5, 44.8, 45.7, 47.2, 50.8, 52.4,
61.0, 61.5, 61.7, 68.1, 68.2, 68.4, 68.7, 68.8, 69.4, 69.5,
70.1, 70.2, 72.2, 72.4, 72.6, 73.2, 73.3, 73.8, 77.2, 78.9,
121.9, 122.0, 127.4, 127.5, 132.8, 133.0, 138.1, 138.3,
169.7, 170.0, 173.2, 173.3, 174.4, 174.5; LRMS (Electro-
spray, H2O, positive ion mode, m/z) for C29H45N3O13:
Detection and analysis of VT bound. After incubation
with the inhibitor±VT mixture the solution was shaken
out of the wells into the sink and the wells washed three
times with 150 mL of 0.2% BSA±PBS. Following the last
wash step, 50 mL/well of a 1 mg/ml solution of the
monoclonal antibody PH1 in 0.2% BSA±PBS was
added. Plates were covered with para®lm and left on the
bench for 60 min. Again, after the incubation, wells
were washed three times with 150 mL of 0.2% BSA±
PBS. After the last wash, 50 mL of a 1:2000 dilution of
GAM±HRP in 0.2% BSA±PBS was added to each well.
Plates were covered with para®lm and left on the bench
for 60 min. After this incubation, wells were washed
three times with 150 mL of 0.2% BSA±PBS leaving the
third wash in the wells for 3 min before a ®nal quick
wash with PBS only. 100 mL/Well of ABTS solution
(2,20-azino-bis[3-ethylbenzthiazoline-6-sulfonic
acid];
0.5 mg/mL containing 3 mL of 30% H2O2 per 10 mL)
was added and the plates covered with tin-foil. For
color development, plates were agitated gently on a
shaking table. After sucient color had developed,
plates were read using an ELISA plate reader at 410 nm
wavelength at intervals until the desired absorption
value was reached (O.D. <2.0). The percentage of spe-
ci®c bound VT was calculated using the value for total
bound VT (i.e.: without any inhibitors present) after
subtracting background activity.
1
644.3 (MH+). 13c: H NMR (400 MHz, DMSO-d6): d
0.84 (t, J=6.8 Hz, 3H), 1.24 (bs, 22 H), 1.45±1.55 (m,
2H), 1.66±1.94 (m, 2H), 3.17±3.30 (m, 2H), 3.40±3.86
(m, 13H), 4.02±4.98 (m, 8H), 7.64 (t, J=9.0, 2H), 7.80
(t, J=8.6, 2H); LRMS (Electrospray, H2O, positive ion
mode, m/z) for C39H65N3O13: 784.4 (MH+), 392.9
(MH+/2).
Acknowledgements
Mr. Donald Leek and Ms. Malgosia Daroszewska
(SIMS, NRC) are thanked for providing NMR and MS
technical support.
Puri®cation of verotoxin (VT) and isolation and puri®ca-
tion of Gb3.23 Verotoxin was isolated and puri®ed from
the high expression recombinant E. coli pJB28 strain.
The glycosphingolipid Gb3 was isolated from the neutral
glycolipid fraction from human renal tissue as described.
References and Notes
1. Varki, A. Glycobiology 1993, 3, 97.
2. Dwek, R. A. Chem. Rev. 1996, 96, 683.
3. Sears, P.; Wong, C.-H. Proc. Natl. Acad. Sci. USA 1996, 93,
12086.
4. Kiessling, L. L.; Pohl, N. L. Chem. Biol. 1996, 3, 71 and
references therein.
5. Lee, Y. C.; Lee, R. T. Acc. Chem. Res. 1995, 28, 321.
6. Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew.
Chem. Int. Ed. Engl. 1998, 37, 2754.
Immobilization of Gb3 onto 96-well plates. A stock solu-
tion of 1.2 mg Gb3/mL in ethanol was prepared and
50 mL/well was aliquoted into 96-well assay plates. The
plates were then left on the bench at room temperature
all day or overnight to allow the MeOH to evaporate.
Once dry, plates were stored desiccated for up to a week.
7. McAuli€e, J. C.; Hindsgaul, O. Chem. and Ind. 1998, 170
8. Roy, R. In Carbohydrates in Drug Design; Witczak, Z. J.,
Niefort, K. A., Eds; Marcel Dekker, Inc.: New York, 1997; pp
83±135.
9. Lingwood, C. A. Trends in Microbiology 1996, 4, 147 and
references therein.
10. Mortell, K. H.; Gingras, M.; Kiessling, L. L. J. Am. Chem.
Soc. 1994, 116, 12053.
11. Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G.
M. J. Am. Chem. Soc. 1996, 118, 3789.
Preparation of inhibitors. Soluble inhibitors of VT bind-
ing were prepared in phosphate bu€ered saline (PBS)
and then mixed 1:1 with a 2Â solution of VT (8 ng per
50 mL in 0.4% BSA±PBS) and incubated for 120 min at
room temperature. The VT±inhibitor mixture was
added to the wells in triplicate at 50 mL per well. 11
Serial dilutions of the inhibitors were made to give a
®nal concentration range of 30 nM to 3 mM.
12. Spevak, W.; Nagy, J. O.; Charych, D. H.; Schaefer, M. E.;
Gilbert, J. H.; Bednarski, M. D. J. Am. Chem. Soc. 1993, 115,
1146.
13. Kingery-Wood, J. E.; Williams, K. W.; Sigal, G. B.;
Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 7303.
Incubation. 96-Well plates were pre-incubated with
150 mL 0.2% BSA in PBS per well for 60 min. Before
adding the inhibitor and verotoxin mixture, wells were
washed twice with 150 mL/well of 0.2% BSA±PBS.

P. Arya et al. / Bioorg. Med. Chem. 7 (1999) 2823±2833
2833
14. Wang Z.-H.; Hindsgaul, O. In Glycoimmunology 2; Ple-
num: New York, 1998 and references therein.
15. Wong, C.-H.; Moris-Varas, F.; Hung, S.-C.; Marron, T.
G.; Lin, C.-C.; Gong, K. W.; Weitz-Schmidt, G. J. Am. Chem.
Soc. 1997, 119, 8152.
16. Hansen, H. C.; Haataja, S.; Finne, J.; Magnusson, G. J.
Am. Chem. Soc. 1997, 119, 6974.
17. Siamanek, E. E.; McGarvey, G. J.; Jablonoski, J. A.;
Wong, C.-H. Chem. Rev. 1998, 98, 833.
18. Arya, P.; Dion, S.; Shimizu, G. K. H. Bioorg. Med. Chem.
Lett. 1997, 7, 1537.
19. Arya, P.; Kutterer, K. M. K.; Qin, H.; Roby, J.; Barnes,
M. L.; Kim, J. M.; Roy, R. Bioorg. Med. Chem. Lett. 1998, 8,
1127.
24. For a recent example of Gb3 mimics, see: Mylvaganam,
M.; Lingwood, C. A. Biochem. Biophys. Res. Commun. 1999,
257, 391.
25. A multivalent form of Gb3 has been found as sub-nano-
molar inhibitor for shiga toxins in Bundle's group, unpub-
lished results.
26. Riley, L. W.; Remis, R. S.; Helgerson, S. D.; McGee, H.
B.; Wells, J. G.; Davis, B. R.; Hebert, R. J.; Olcott, E. S.;
Johnson, L. M.; Hargrett, N. T.; Blake, P. A.; Cohen, M. C.
N. Engl. J. Med. 1983, 308, 681.
27. Karmali, M. A.; Petric, M.; Lim, C.; Fleming, P. C.;
Arbus, G. S.; Lior, H. J. Infect. Dis. 1985, 151, 775.
28. Weinstein, D. L.; Jackson, M. P.; Perera, L. P.; Holmes, R.
K.; O'Brien, A. D. Infect. Immun. 1989, 57, 3743.
29. St. Hilaire, P. M.; Boyd, M. K.; Toone, E. J. Biochemistry
1994, 33, 14452.
20. Jones, D. H.; Lingwood, C. A.; Barber, K. R.; Grant, C.
W. M. Biochemistry 1997, 36, 8539.
21. Lingwood, C. A. Glycoconjugate Journal 1996, 13, 495.
22. Nyholm, P.-G.; Magnusson, G.; Zheng, Z.; Norel, R.;
Boyd, B. B.; Lingwood, C. A. Chem. Biol. 1996, 3, 263.
23. Boyd, B.; Magnusson, G.; Zhiuyan, Z.; Lingwood, C. A.
Eur. J. Biochem. 1994, 223, 873 and references therein.
30. Ling, H.; Boodhoo, A.; Hazes, B.; Cummings, M. D.;
Armstrong, G. D.; Brunton, J. L.; Reed, R. J. Biochemistry
1998, 37, 1777.
31. Kihlberg, J.; Elofsson, M. Curr. Med. Chem. 1997, 4, 85±
116.