Bio-inspired synthetic receptor molecules towards mimicry of vancomycin

Article abstract of DOI:10.1016/S0960-894X(01)00237-2

A 512-member library of bio-inspired synthetic receptor molecules was prepared featuring a triazacyclophane scaffold. The purpose of this scaffold was to orient three (identical) peptide 'binding arms' in order to mimic an antibiotic binding cavity as is present in the vancomycin antibiotics. The library was screened with D-Ala-D-Ala and D-Ala-D-Lac containing ligands, which are present in the cell wall precursors of pathogenic bacteria. Screening and validation led to identification of a synthetic receptor capable of binding these ligands.

Full text of DOI:10.1016/S0960-894X(01)00237-2

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1521–1525
Bio-inspired Synthetic Receptor Molecules
Towards Mimicry of Vancomycin
Menno C. F. Monnee, Arwin J. Brouwer, Linda M. Verbeek,
Andre M. A. van Wageningeny and Rob M. J. Liskamp*
Department of Medicinal Chemistry, Utrecht Institute for Pharmaceutical Sciences,
Utrecht University, PO Box 80082, 3508 TB Utrecht, The Netherlands
Received 26 January 2001; accepted 21 March 2001
Abstract—A 512-member library of bio-inspired synthetic receptor molecules was prepared featuring a triazacyclophane scaffold.
The purpose of this scaffold was to orient three (identical) peptide ‘binding arms’ in order to mimic an antibiotic binding cavity as is
present in the vancomycin antibiotics. The library was screened with d-Ala-d-Ala and d-Ala-d-Lac containing ligands, which are
present in the cell wall precursors of pathogenic bacteria. Screening and validation led to identification of a synthetic receptor
capable of binding these ligands. # 2001 Elsevier Science Ltd. All rights reserved.
The antibacterial activity of glycopeptide antibiotics
such as vancomycin probably largely results from bind-
ing to the terminus of the bacterial cell wall precursor,
namely d-alanine-d-alanine, thereby inhibiting forma-
tion of the cell wall in growing bacteria.1,2,3 In view of
the increasing number of bacteria resistant against van-
comycin antibiotics, it is imperative to uncover new
molecules capable of binding d-alanine-d-alanine and/
or d-alanine-d-lactic acid, which is the terminus of the
cell wall precursor present in resistant bacteria.4
To construct these synthetic receptor molecules, we
have chosen for covalent control by using a scaffold, to
which three peptide arms can be attached and oriented
in space. The tripodal hinge for these receptor molecules
was based on the best hinge we have recently found for
tweezer-like receptors.8 A bridge between the two
nitrogen atoms containing an additional nitrogen atom
then leads to a triazacyclophane scaffold to which three
binding arms can be attached (Fig. 1).
Thus, starting from bisbromide 1,9 triazacyclophane
scaffold 3 was prepared by alkylation with sulfonamide
We are involved in two programs towards realizing this
goal. In the first program of biomimetic synthetic recep-
tor molecules, we aim at vancomycin mimics in which
analogues of crucial parts of the vancomycin structure
are prepared to dissect and further improve the biologi-
cal activity due to these parts.5,6 In the second program
of bio-inspired synthetic receptor molecules, completely
different molecules are constructed.7 However, in devel-
oping these molecules, we are guided—if not inspired—
by the basic characteristics of vancomycin: (1) It is a
natural receptor molecule for the terminus of the bac-
terial cell wall precursor. (2) It derives its rigidity by
covalent control of its shape. (3) Its building blocks are
amino acids.
2
10 in 64% yield. Removal of the o-nitrobenzenesulfonyl
and Boc protecting groups was followed by introduc-
tion of a spacer by alkylation with 3-bromo N-Boc
propylamine leading to ester 4.11 This spacer facilitates
the introduction of buildings blocks of the binding arms
(Scheme 1).
Following saponification of the methylester, coupling to
a solid-phase resin containing glycine12 and removal of
the Boc protecting groups, a library was prepared by the
split-mix method.13 To facilitate screening, a repre-
sentative set consisting of eight, both non-functionalized
and functionalized amino acids, namely Gly, Ala, Val,
Leu, Phe, Lys(Boc), Glu(Ot-Bu) and Ser(t-Bu) was
used. Although this set is limited, it reflects the amino
acid diversity to a considerable extent. The peptidic
binding arms consisted each of three amino acid resi-
dues that were introduced in three subsequent BOP-
*Corresponding author. Tel.: +31-30-2537396; fax: +31-30-2536655;
e-mail: r.m.j.liskamp@pharm.uu.nl
yPresent address: DSM-Research, PO Box 18, 6160 MD Geleen, The
Netherlands.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00237-2

1522
M. C. F. Monnee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1521–1525
coupling reactions, leading to a library of 83 (=512)
synthetic receptors (Scheme 2). Finally, Fmoc, Boc and
tert-butyl protecting groups were removed by piperidine
and TFA, to give the deprotected library.
10 mg of the library (ca. 104 beads) with a chloroform
solution of Ds-Gly-d-Ala-d-Lac-OH 8 or Ds-Gly-d-
Ala-d-Ala-OH 717 and visual inspection of the beads
using a Leica fluorescence stereomicroscope. At con-
centrations >50 mM, no selectivity of binding was
observed and >50% of the beads were fluorescent.
Screening of this library was carried out by incubating
Figure 1. Vancomycin and bioinspired synthetic receptor molecules having a triazacyclophane scaffold.
Scheme 1. Synthesis and solid-phase attachment of the triazacyclophane scaffold.
Scheme 2. Preparation of the 512-member synthetic receptor library.

M. C. F. Monnee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1521–1525
1523
However, using concentrations of 10 mM resulted in
selective binding and <10% of the beads were fluor-
escent. In general, the observed fluorescence intensity
was higher for the d-lactate-containing ligand 8 as
compared to the d-alanine-containing ligand 7 which
might be a qualitative indication of stronger binding.
After incubation with the ligands three fluorescent
beads were selected from each incubation and subjected
to Edman degradation. The identity of the peptide arms
of the corresponding synthetic receptors, attached to the
solid-phase resin bead, is shown in Table 1.
basic amino acid (Lys, three out of three selected library
members). The differences of only one amino acid
between the former library members in determining the
selectivity prompted us to resynthesize these library mem-
bers and determine in detail the binding characteristics.
Resynthesis was carried out analogously to the synthesis
depicted in Scheme 1 except for the use of trisulfona-
mide 914 instead of disulfonamide 2. The former is more
easily accessible, in higher yields, and more suitable for
synthetic receptor molecules containing three identical
arms15 (Scheme 3). Furthermore, Boc-Val-OH was used
instead of Fmoc-Val-OH, thereby avoiding an addi-
tional Fmoc-deprotection step. After resynthesis the
protected compounds were cleaved from the resin for
characterization. Electron spray mass spectrometry was
in agreement with the identity of the compounds and
TLC showed that the purity was quite good considering
the number of steps (10) which were carried out for the
solid-phase synthesis of these compounds.
Table 1.
Screening with:
Sequence found
by Edman
degradation:
AA3-AA2-AA1
Resynthesized
synthetic
receptor
Ds-Gly-d-Ala-d-Ala-OH 7
Ds-Gly-d-Ala-d-Lac-OH 8
Val-Glu-Phe
Phe-Lys-Ser
Glu-Gly-Gly/Phea
Val-Lys-Phe
Lys-Lys-Gly
12
11
For establishing the binding characteristics, it is of
utmost importance to determine the binding constants
as accurately and rigorously as possible. To this end,
aliquots of resin containing the deprotected receptors 11
and 12, respectively, were accurately weighed and incu-
bated with varying concentrations of dansylated ligand
(Ds-Gly-d-Ala-d-Ala-OH 7 or Ds-Gly-d-Ala-d-Lac-
OH 8) in chloroform. Around 10 concentrations were
chosen in order to obtain a number of evenly dis-
tributed data points for the generation of a good
regression line. After incubation overnight in the dark
and filtration of the beads, fluorescence intensity was
measured in the filtrate, thereby allowing determination
of the concentration of the unbound ligand. A calibra-
tion curve was used for this purpose since at higher
concentrations fluorescence is quenched, giving rise to a
non-linear relationship. The obtained values were also
corrected for volume loss due to the evaporation of the
solvent. For establishing the concentration of synthetic
receptors on the resin ([R0]), it was important to correct
the mass of the weighed resin for the mass of the syn-
thetic receptor, which at the particular loading of
the resin (ca. 0.3 mmol gÀ1), amounts to ca. 30%!
Lys-Leu-Val
aThe sequencing cycle was not completely unambiguous with respect
to this amino acid residue.
Validation of ‘hits’ by resynthesis of the selected library
members and reproducing—in a more accurate way—
the properties for which the library was screened is
essential in any combinatorial approach towards finding
compounds with desired properties. In these screenings,
it was found that the peptidic arms of a synthetic
receptor library member capable of binding the fluor-
escent d-Ala-d-Ala ligand 7 consisted of Val-Glu-Phe,
whereas the peptidic arms in a library member capable
of binding d-Ala-d-Lac 8 had the sequence Val-Lys-
Phe. The differences between the other library members
capable of binding d-Ala-d-Ala or d-Ala-d-Lac were
greater, but even judging from these small sets, it
seemed that binding of d-Ala-d-Ala involved an acidic
amino acid (Glu, two out of three selected library
members), whereas binding of d-Ala-d-Lac involved a
Scheme 3. Resynthesis of two members, i.e. 11 and 12 of the synthetic receptor library.

1524
M. C. F. Monnee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1521–1525
Furthermore, the loading, which is ultimately the num-
ber of synthetic receptors, has to be determined pre-
cisely, since the binding constant Ka is very sensitive to
variations in this number.16 As a reference we used resin
containing an acetylated glycine residue, that is Ac-Gly-
O-ArgogelTM which is used as a linker for attachment of
the synthetic receptors (Scheme 2). The values for the
amount of binding obtained with this resin—referred to
as non-specific binding—were always subtracted from
the values obtained with the resin containing the
synthetic receptors.
Table 2. Determination of the binding constant Ka of the resynthe-
sized receptors 11 and 12
Val-Lys-Phe (11)
Val-Glu-Phe (12)
Ds-Gly-d-Ala-d-Ala-OH 7
Ds-Gly-d-Ala-d-Lac-OH 8
26,600 MÀ1
10,100 MÀ1
—
—
requirement for validation of hits found by screening
(Table 2).
In conclusion, we have prepared a library of bio-
inspired receptors aimed at mimicry of vancomycin.
Although screening indicated that we had selectivity for
binding to the cell wall precursor present in resistant
bacteria, validation of the hits showed that our synthetic
receptor was even better in binding d-Ala-d-Ala present
in non-resistant bacteria. Nevertheless, affinity con-
stants for binding by synthetic receptor 11 of either one
of both cell wall precursors are very promising. There-
fore, we think that this approach is an attractive one to
pursue further in order to find synthetic receptors with
binding properties in more polar solvents including
water as well as for uncovering synthetic receptors cap-
able of binding other ligands.
Resynthesis of resin containing synthetic receptor 11
having the amino acid sequence Val-Lys-Phe in its pep-
tidic binding arms and which in the screening experi-
ment was selected as a binder of Ds-Gly-d-Ala-d-Lac-
OH (8), followed by determination of the binding con-
stant resulted in quite a satisfactory Ka of 10,100 MÀ1
However, incubating this receptor with Ds-Gly-d-Ala-
d-Ala-OH (7) gave the even better Ka of 26,600 MÀ1
.
,
although this synthetic receptor was not identified upon
screening the library of synthetic receptors with Ds-Gly-
d-Ala-d-Ala-OH! Even more surprisingly were the
results found after resynthesis of resin containing syn-
thetic receptor 12 having the amino acid sequence Val-
Glu-Phe in its peptidic binding arms followed by incu-
bation with ligand Ds-Gly-d-Ala-d-Ala-OH used in the
screening to uncover this synthetic receptor. No binding
was detected whatsoever; binding to the reference resin
Ac-Gly-O-ArgogelTM was in fact better than binding to
the synthetic receptor! It seemed that the presence of the
synthetic receptor led to an increasing incliness to bind
less ligand. Clearly no Ka could be determined. Also, no
binding was detected with the other ligand Ds-Gly-d-
Ala-d-Lac-OH 8. Although at first glance these data
appeared difficult to explain, they may point at subtle
differences between screening conditions and determina-
tion of the binding constants, thereby underlining the
importance of validating any hits which are obtained after
screening. Determination of the binding characteristics
was carried out on the deprotected synthetic receptor,
which was obtained after deprotection by acidolysis
with TFA to remove Boc and t-Bu protective groups.
To remove residual acid, the resin bound deprotected
receptor was rigorously washed with triethylamine, fol-
lowed by washings to remove excess of triethylamine.
However, washing with triethylamine will convert the
side-chain carboxylic acid moieties of glutamic acid
residues into the carboxylates. The resulting negatively
charged residues may discourage binding of carboxyl-
ate-containing ligands such as 7 and 8. This may explain
the observed absence of binding by these synthetic
receptor containing resins and in fact a higher level of
binding by the uncharged reference resin. It was found
that binding to the synthetic receptor was observed
when the pH was lowered, for example by adding acetic
acid (data not shown). It was therefore assumed that
screening of the library of receptors, which led to selec-
tion of the resynthesized receptor 12, had been carried
out under slightly acidic conditions either because not
entirely acid-free chloroform was used or the depro-
tected synthetic library was not completely neutralized.
Be this as it may, these results emphasize the absolute
Acknowledgements
The Sequence Center Utrecht is gratefully acknowledged
for carrying out the sequence analyses. These investiga-
tions were supported (M.C.F.M. and A.M.A.W.) by the
council for Chemical Sciences of The Netherlands
Organization for Scientific Research (CW-NWO).
References and Notes
1. Foldes, M.; Munro, R.; Sorrell, T. C.; Shankar, S.; Toohey,
M. J. Antimicrob. Chemother. 1983, 11, 2 1.
2. Williams, D. H.; Bardsley, B. Angew. Chem., Int. Ed. 1999,
38, 1172. Williams, D. H.; Bardsley, B. Angew. Chem. 1999,
111, 1264.
3. Nicolaou, K. C.; Boddy, C. N. C.; Brase, S.; Winssinger, N.
Angew. Chem., Int. Ed. 1999, 38, 2096. Nicolaou, K. C.;
Boddy, C. N. C.; Brase, S.; Winssinger, N. Angew. Chem.
1999, 111, 2230.
4. Neu, H. C. Science 1992, 257, 1064. Walsh, C. T.; Fischer,
S. L.; Park, I.-S.; Prahalad, M.; Wu, Z. Chem. Biol. 1996, 3,
21. Walsh, C. Science 1999, 284, 442.
5. Xu, R.; Greiveldinger, G.; Marenus, L. E.; Cooper, A.;
Ellman, J. A. J. Am. Chem. Soc. 1999, 121, 4898.
6. Pieters, R. J. Tetrahedron Lett. 2000, 41, 7541.
7. For earlier examples of molecules structurally unrelated to
vancomycin capable of binding Ala-Ala containing sequences
see: Henley, P. D.; Waymark, C. P.; Gillies, I.; Kilburn, J. D.
J. Chem. Soc., Perkin Trans. 1 2000, 1021. Henley, P. D.; Kil-
burn, J. D. Chem. Commun. 1999, 1335. Dowden, J.; Edwards,
P. D.; Flack, S. S.; Kilburn, J. D. Chem. Eur. J. 1999, 79.
Dowden, J.; Edwards, P. D.; Killburn, J. D. Tetrahedron Lett.
1997, 38, 1095. Liu, R.; Sanderson, P. E. J.; Still, W. C. J. Org.
Chem. 1990, 55, 5184.

M. C. F. Monnee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1521–1525
1525
8. Lowik, D. W. P. M.; Weingarten, M. D.; Broekema, M.;
Brouwer, A. J.; Still, W. C.; Liskamp, R. M. J. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 1846. Lowik, D. W. P. M.; Weingar-
ten, M. D.; Broekema, M.; Brouwer, A. J.; Still, W. C.; Lis-
kamp, R. M. Angew. Chem. 1998, 110, 1947.
Abstracts of Papers, Vol. 5, p. 47; 14th Int. Congr. Biochem.
Prague, Czechoslavakia. Furka, A.; Sebestyen, F.; Asgedom,
M.; Dibo, G. Int. J. Pept. Protein Res. 1991, 37, 487.
14. Wang, T.; An, H.; Vickers, T. A.; Bharadwaj, R.; Cook,
P. D. Tetrahedron 1998, 54, 7955.
9. Kurz, K.; Gobel, M. W. Helv. Chim. Acta 1996, 79, 1967.
10. An, H.; Cummins, L. C.; Griffey, R. H.; Bharadwaj, R.;
Haly, B. D.; Fraser, A. S.; Wilson-Lingardo, L.; Rinsen,
L. M.; Wyatt, J. R.; Cook, P. D. J. Am. Chem. Soc. 1997,
119, 3696.
15. Scaffold 3 obtained from sulfonamide 2, containing two
orthogonal protecting groups can be used for the preparation
of receptor molecules containing two identical and a third
different arm.
16. This is apparent from the used equation:
11. Mulders, S. J. E.; Brouwer, A. J.; van den Meer, P. G. J.;
Liskamp, R. M. J. Tetrahedron Lett. 1997, 38, 631. Brouwer,
A. J.; Mulders, S. J. E.; Liskamp, R. M. J. Eur. J. Org. Chem.
2001, in press.
12. Introduction of an Fmoc-glycine residue enables a reliable
determination of the loading as well as cleavage of the syn-
thetic receptor under mildly basic conditions.
À
Â
ÃÁ
ÀÂ
Ã
Á
1= ½L0Š À Leq ¼ 1= Leq ½R0ŠKa þ 1=½R0Š
Â
Ã
in which ½L0Š is the initial concentration of ligand 7 or 8; Leq
is the equilibrium concentration of the ligand in solution and
[R0] is the ‘concentration’ of synthetic receptors on the resin.
17. Rothman, J. H.; Still, W. C. Bioorg. Med. Chem. Lett.
1997, 7, 3159.
13. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G.