Design and evaluation of analogues of the bacterial cell-wall peptidoglycan motif l-Lys-d-Ala-d-Ala for use in a vancomycin biosensor

Article abstract of DOI:10.1016/j.bmcl.2007.08.055

Four small molecular receptors of vancomycin have been designed to make part of a novel biosensor device based on the FTIR-ATR detection: N-Boc (2a) or N-Ac (2b)-6-aminocaproyl-d-Ala-d-Ala and N-Boc (3a) or N-Ac (3b)-6-aminocaproyl-d-Ala-d-Ser. Using an original microbiological approach to assess the competition of compounds with the natural target of vancomycin in bacteria, EC₅₀ values of 6.3-8.0 × 10^-5 M (2a-b) and 7.1-9.3 × 10^-4 M (3a-b) were determined. Vancomycin:2b complex was characterized by MS.

Full text of DOI:10.1016/j.bmcl.2007.08.055

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5758–5762
Design and evaluation of analogues of the bacterial
cell-wall peptidoglycan motif L-Lys-D-Ala-D-Ala for use
in a vancomycin biosensor
a
b
b
b
Olivier Hernout, Karine Berthoin, Isabelle Delattre, Paul M. Tulkens,
b,c
a,
*
St e´ phane Carryn and Jacqueline Marchand-Brynaert
a
Unit e´ de Chimie organique et m e´ dicinale, Universit e´ catholique de Louvain, B aˆ timent Lavoisier, Place L. Pasteur 1,
B-1348 Louvain-la-Neuve, Belgium
b
Unit e´ de Pharmacologie cellulaire et mol e´ culaire, Universit e´ catholique de Louvain 73.70, Avenue E. Mounier 73,
B-1200 Brussels, Belgium
c
Eumedica s.a. Chemin de Nauwelette 1, B-7170 Manage, Belgium
Received 18 July 2007; revised 24 August 2007; accepted 24 August 2007
Available online 28 August 2007
Abstract—Four small molecular receptors of vancomycin have been designed to make part of a novel biosensor device based on the
FTIR-ATR detection: N-Boc (2a) or N-Ac (2b)-6-aminocaproyl-D-Ala-D-Ala and N-Boc (3a) or N-Ac (3b)-6-aminocaproyl-D-Ala-
D-Ser. Using an original microbiological approach to assess the competition of compounds with the natural target of vancomycin in
À5
À4
bacteria, EC50 values of 6.3–8.0 · 10 M (2a–b) and 7.1–9.3 · 10 M (3a–b) were determined. Vancomycin:2b complex was char-
acterized by MS.
ꢀ
2007 Elsevier Ltd. All rights reserved.
Immobilisation of bioactive molecules on solid supports
has gained a growing interest because the resulting
devices can be used in various detection systems called
In the course of a programme devoted to the develop-
ment of biosensors based on the FTIR-ATR spectros-
copy detection, we are engaged in the detection and
3
1
biosensors. These allow the specific recognition of a
quantification of the glycopeptide antibiotic vancomycin
in human fluids. Vancomycin is of large clinical impor-
tance as it is currently the most often recommended
antibiotic for treating infections caused by methicillin-
resistant Staphylococcus aureus (MRSA) in hospitals.
Yet, it requires repeated blood level monitoring and
rapid delivery of the results to the clinician to ensure
free analyte (the ligand) by a target (the receptor) which
is tightly bound to the device surface. Sensors based on
the molecular recognition of biomolecules have already
attracted intensive interest in many fields such as
environmental analysis, monitoring of biotechnological
2
processes, and medical diagnosis and control. Among
4
the different surface-sensitive techniques applied to
detect ligand–receptor interaction, the FTIR-ATR
method (Fourier transform infrared spectroscopy in
the attenuated total internal reflection mode) is of
particular interest since it allows for high-sensitive
optimal efficacy and avoid undue toxicity.
The antibiotic activity of vancomycin results from
strong non-covalent interactions (five hydrogen bonds)
between the drug and the C-terminal motif of the penta-
peptide L-Ala-D-Glu-L-Lys-D-Ala-D-Ala present in the
2
label-free detection.
5
cell wall peptidoglycan precursor of procaryotes.
Recently, vancomycin has been used in single-molecule
force spectroscopy to detect and image the localization
6
of free D-Ala-D-Ala termini on the surface of bacteria.
In bacteria in which this bonding motif is terminated
by D-Lac or D -Ser, the vancomycin affinity is markedly
Keywords: Vancomycin; D-Ala-D-Ala derivatives; Microbiological
assay; Biosensor.
7
*
Corresponding author. Tel.: +32 0 10 472740; fax: +32 0 10
74168; e-mail: marchand@chim.ucl.ac.be
reduced, resulting in resistance to this antibiotic.
The usual simplified model to study the interaction of
4
0
960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.055

O. Hernout et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5758–5762
5759
vancomycin and its bacterial target is the N-a-Ac-L-Lys-
D-Ala-D-Ala tripeptide (1) for which the molecular basis
here by masking the e-NH with t-butyloxycarbonyl or
acetyl groups.
2
1
c
of vancomycin affinity is well established (Scheme 1)
and which has been used, after immobilisation on aga-
The target molecules 2–3 were prepared as usual in
peptide chemistry (Scheme 2). Briefly, the amine
function of 6-aminocaproic acid was protected with
8
rose, to purify vancomycin from fermentation broth.
In view of immobilising vancomycin binding motifs
(
tert-butyloxycarbonyl group (Boc O, NaOH (1 M),
2
receptors) on an ATR optical element, two key ques-
dioxane/H O (2:1), 0–20 ꢁC, 17 h.) and the acid func-
2
tions need to be addressed, namely: (i) the possibility
of replacing L-Lys with 6-aminocaproic acid (this struc-
tural modification has the advantage of suppressing a
chiral centre and simplifying the synthesis of the biosen-
sor); (ii) the possibility of recycling the biosensor thanks
to the formation of weaker complexes than the one
formed with L-Lys-D-Ala-D-Ala.
tion was activated as N-hydroxysuccinimidyl ester
(NHS, DMAP, DCC, CH Cl , 0–20 ꢁC, 17 h.). This es-
ter 4a was reacted with commercial D-Ala-D-Ala using
2
2
PyBOP (benzotriazol-1-yl-oxytripyrrolidinophosphoni-
9
um hexafluorophosphate) as coupling reagent. The
resulting N-Boc-6-aminocaproyl-D-Ala-D-Ala peptide
2a was purified by chromatography. D-Ala-D-Ser (5b)
was obtained by coupling commercial H-D-Ser-(Ot-
Bu)-Ot-Bu with Boc-D-Ala-OH (PyBOP, TEA, MeCN,
20 ꢁC, 2 h.) followed by hydrolysis of the tert-butyl es-
ters (TFA/CH Cl (1:1), 20 ꢁC, 1 h.). N-Boc-6-amino-
In this communication, we have selected four potential
synthetic targets (2a–b and 3a–b) ending, respectively,
with D-Ala or D-Ser (Scheme 2). Before their immobili-
sation on solid supports via the NH -aminocaproyl end-
2
ing, those molecules have been subjected to HPLC
2
2
caproyl-D-Ala-D-Ser (3a) was prepared in a same
manner as for 2a.
(
(
High Performance Liquid Chromatography), MS
Mass Spectrometry) and microbiological studies to
From commercially available 6-acetamidocaproic acid
(4b), the two analogous mimics 2b and 3b were obtained
in the same conditions.
determine their ability to bind to vancomycin in com-
parison with compound 1. Since neither the a-N-Ac
1
0
nor the e-NH groups of 1 are considered critical in its
2
1
1
binding to vancomycin, we speculated that the first
group could be removed and that the second one could
be used for surface anchoring. This has been modelled
HPLC (RP C18 column and UV detection) was used
to follow the formation of complexes between vancomy-
cin (obtained as Vancocin 500 from GlaxoSmithKline,
ꢂ
Genval, Belgium) and selected target compounds. Van-
comycin (67 lM final concentration) was mixed with
increasing concentrations of compounds 1 or 2a in aque-
ous solution (pH 7.4). A 52.2% and a 73.9% reduction of
the free vancomycin concentration were observed at a
vancomycin:target compound molar ratio of 1:25 for 1
and 2a, respectively. We, however, were not able to iden-
tify or to isolate the corresponding vancomycin com-
plexes by (semi-preparative) HPLC. Yet, direct
injection of the mixtures in a mass spectrometer led to
unambiguous evidence of complex formation (analysis
performed with electrospray ionization [ESI]) using the
negative ion mode (data processed by Excalibur
TM
ver-
Scheme 1. Interaction between vancomycin and 1.
sion 1.2 software). Figure 1 shows a typical collision-in-
duced dissociation spectrum of vancomycin:2b
mixture (1:25 molar ratio). A similar finding was made
a
1
2
with a vancomycin:1 mixture.
A microbiological approach was then used to directly
assess the competition of compounds 2a, 2b, 3a and
3b with the natural target of vancomycin in susceptible
bacteria. For this purpose, 10 viable bacteria/mL (col-
6
ony forming units [CFU]) of a fully sensitive S. aureus
(ATCC 25923) were exposed to a constant concentra-
tion of vancomycin (1 mg/L [0.69 lM] corresponding
to its minimal inhibitory concentration as determined
1
3
by broth microdilution technique ) and increasing
concentrations of the target compounds (molar ratios
1:1 to 1:100,000). The mixtures were then incubated
for 5 h at 37 ꢁC in Mueller–Hinton cation-adjusted
broth. Bacterial killing or growth was then evaluated
1
3
by colony counting (in this system, vancomycin alone
caused a 1 log CFU decrease, whereas cultures made
in the absence of vancomycin showed a 2 log CFU
Scheme 2. Different 6-aminocaproyl mimics of the D-Ala-D-Ala
receptor.

5760
O. Hernout et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5758–5762
Figure 1. Fragmentation spectrum of the monoanionic vancomycin + 2b at 1:25 molar ratio. The arrowhead points to an m/z value consistent with a
complex vancomycin + 2b-2H (1762.87).
increase; the diluent used to solubilize compounds 2a
and 2b [DMSO 1% final concentration] was without
influence).
14
Results are shown in Figure 2 and Table 1. Com-
pound 1 proved competitive with an apparent EC50
(
half maximal effective concentration) value of
.6 · 10 M, in agreement with previous, independent
À6
6
measures of its affinity constant with vancomycin
À6
15
(
2.1 · 10 M). In contrast, the dipeptide D-Ala-D-
Ala was poorly competitive, demonstrating the critical
role of the far-left hydrogen bound (see Scheme 1) in
the binding of compound 1 with vancomycin. Replace-
ment of N-a-Ac-L-Lys by N-protected-6-aminocaproyl
(
2) increased the EC50 of about 1 log, suggesting that
the a-N-Ac plays a so far undescribed, albeit minor,
role in vancomycin binding to its target. Conversely,
an acetyl or a more bulky group such as t-Boc on
Figure 2. Competition between the vancomycin natural target in
S. aureus and the target analogues 1, 2a/2b, 3a/3b, or the dipeptide
D-Ala-D-Ala.
the e-NH group was not critical. Replacing the termi-
2
nal D-Ala by D-Ser (3) further increased the EC50
(
about one log for 3a/3b compared to 2a/2b) in
accordance with published values for the difference in
Table 1. EC50 values of the compounds tested in competition with the
vancomycin natural target in S. aureus
affinities between N-a-Ac-L-Lys-D-Ala-D-Ala and
1
6
N-a-Ac-L-Lys-D-Ala-D-Ser.
Compound
EC50
This simple and inexpensive competition assay thus pro-
vided useful information for the selection of the candi-
date targets. Our microbiological method could be
routinely used to determine the binding affinity of syn-
thetic small molecules towards vancomycin.
Observed log
(vancomycin:target)
molar ratio
Calculated target
concentration (M)
a
À6
1
2a
0.984 ± 0.014
1.964 ± 0.011
2.066 ± 0.007
3.011 ± 0.010
3.129 ± 0.004
4.101 ± 0.009
6.6 · 10
À5
6.4 · 10
À5
2
3
3
b
a
b
8.0 · 10
À4
7.1 · 10
In conclusion, we found that N-protected-6-aminocap-
royl-D-Ala-D-Ala (2) still binds vancomycin with an
apparent dissociation constants of the order of
À4
9.3 · 10
À3
D-Ala-D-Ala
8.7 · 10
À5
a
log(vancomycin:target molar ratio)
À6
1
0
M. As the plasma concentrations of vancomycin
10
· 0.69 · 10
.

O. Hernout et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5758–5762
5761
1
7
that need to be monitored in patients (15–50 mg/L) are
within the same molar range, these analogues could be
used as potential receptors for FTIR-ATR-based detec-
tion of vancomycin in the clinics. 6-Aminocaproyl-D-
Ala-D-Ser derivatives (3) could provide an alternative
if a less tight binding was necessary to obtain faster
and more complete dissociation in washout-phases.
Since the Boc-compounds are easily N-deprotected for
2H); 3.04 (t, 2H, J = 7.5 Hz); 4.18 (q, J = 7.5 Hz, 1H); 4.29
(
1
q, J = 7.5 Hz, 1H). C NMR (75 MHz, CDCl
3
3
) d 181.4;
1
2
79.3; 176.7; 160.9; 83.3; 52.8; 51.9; 42.3; 37.7; 30.2; 27.9;
+
31 3 6-
7.4; 22.4; 19.7; 19.1. HRMS (ES ) calcd for C17H N O
+
1
Na [M+Na] 396.2111: found 396.2107. (3a) H NMR
300 MHz, D O) d 1.20–1.66 (m, 18H); 2.28 (t, J = 7.5 Hz,
H); 3.03 (t, J = 7.5 Hz, 2H); 3.86 (t, J = 4.7 Hz, 2H);
(
2
2
4
1
13
.29–4.41 (m, 2H). C NMR (75 MHz, D
77.4; 177.3; 160.9; 83.3; 64.2; 58.6; 52.0; 42.3; 37.8; 31.1;
2
O) d 179.4;
3
+
further linking to germanium devices, they are cur-
rently used for the effective development of a vancomy-
cin biosensor for medical applications.
30.2; 27.9; 27.4; 19.2. HRMS (ES ) calcd for C H N O
17 31 3 7-
+
1
Na [M+Na] 412.2060: found 412.2049. (2b) H NMR
(300 MHz, D O) d 1.17–1.61 (m, 12H); 1.95 (s, 3H); 2.27
t, J = 6.9 Hz, 2H); 3.11 (t, J = 6.9 Hz, 2H); 4.14 (q,
2
(
1
3
J = 5.8 Hz, 1H); 4.32 (q, J = 5.4 Hz, 1H). C NMR
75 MHz, D O) d 180.0; 177.6; 175.0; 174.7; 51.3; 50.3;
0.1; 36.0; 28.7; 26.0; 25.7; 22.7; 18.1; 17.4. HRMS (ES )
(
4
2
Acknowledgments
+
+
calcd for C14
3
H
25
3
N O
5
Na [M+Na] 338.1692: found
We thank J.L. Habib-Jiwan for help in the MS studies.
J.M. is Senior Research Associate of the Belgian Fonds
National de la Recherche Scientifique, and S.C. benefi-
ciary of a First Entreprise programme of the R e´ gion
Wallonne. This work was supported by the Region
Wallonne (MED-ATR project, Contract Nos.
1
38.1690. (3b) H NMR (300 MHz, D O) d 1.23–1.69
2
(
m, 9H); 1.96 (s, 3H); 2.28 (t, J = 7.5 Hz, 2H); 3.14 (t,
J = 7.3 Hz, 2H); 3.79–3.96 (m, 2H); 4.29–4.45 (m, 2H).
+
HRMS (ES ) calcd for C H N O Na [M+Na]
+
1
4
25
3
6
354.1641: found 354.1642.
All the experimental protocols, for intermediates and final
products, with structural characterisations and figures of
0
516040-0516263-0516264-0516265).
1
13
selected H and C NMR spectra are available as
Supporting Information.
1. We used Waters equipment (HPLC system 2690, diode
1
Supplementary data
ꢂ
array detector 996, Millenium 32 program) and Symme-
try Shield
TM
RP C18 column from Waters (4.6 · 100 mm;
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.08.055.
5 lm). The analysis conditions are as follows: mobile
phase = acetonitrile and ammonium acetate buffer, 70
mM, pH 5.0; gradient from 5% CH CN + 95% aqueous
3
buffer to 95% CH CN + 5% aqueous buffer; flow
3
rate = 1 mL/min; T = 20 ꢁC; injected volume = 10 lL;
analysis time = 20 min; UV detection = 280 nm.
2. Vancomycin:2b complex: m/z = 1762.87 (vancomy-
References and notes
1
1
. (a) Kumar, P.; Agrawal, S. K.; Misra, A.; Gupta, K. C.
Bioorg. Med. Chem. Lett. 2004, 14, 1097; (b) Lahiri, J.;
Isaacs, L.; Tien, J.; Withesides, G. M. Anal. Chem. 1999,
cine + 2b-2H = C80
cine–CO –H); 943.87 (trimer of 2b); 629.02 (dimer of
2b); 314.22 (2b) (see Fig. 1).
Vancomycin:1 complex: m/z = 1776.96 (Vanco + 1-2H =
C H100Cl N O29); 1318.83 (tetramer of 1); 988.93 (tri-
80 2 13
mer of 1); 659.17 (dimer of 1); 329.26 (1) (spectrum not
shown).
2 12
H100Cl N O29); 1404.09 (vancomy-
2
7
1, 777; (c) Cooper, M. A.; Fiorini, M. T.; Abell, C.;
Williams, D. H. Bioorg. Med. Chem. Lett. 2000, 8, 2609;
d) Leech, D. Chem. Soc. Rev. 1994, 23, 205; (e) Andree-
(
scu, S.; Sadik, O. A. Pure Appl. Chem. 2004, 76, 861.
. Liao, W.; Wei, F.; Liu, D.; Qian, M. X.; Yuan, G.; Zhao,
X. S. Sensors and Actuators B 2006, 114, 445.
. (a) Marchand-Brynaert, J.; Goormaghtigh, E.; Hombl e´ ,
F.; Vou e´ , M.; De Coninck, J., 2002, PCT WO
2
3
13. Samples were plated on tryptic soy agar (TSA) after
appropriate dilutions, and incubated for 24 h at 37 ꢁC.
Colonies were enumerated using a Gel Doc 2000 appara-
tus (Bio-Rad laboratories, Hercules, CA) operated with
Quantity One software (Bio-Rad Laboratories). The
actual number of colonies counted was typically between
200 and 800, and was never less than 80. See Carryn, S.;
Van de Velde, S.; Van Bambeke, F, Mingeot-Leclercq,
M.-P.; Tulkens, P.M. J. Infect. Dis. 2004, 189, 2101 and
Barcia-Macay, M.; Lemaire, S.; Mingeot-Leclercq, M. P.;
Tulkens, P. M.; Van Bambeke, F. J Antimicrob Chemother
2006, 58, 1177 for details and validation.
0
2/056018A1; (b) Devouge, S.; Salvagnini, C.;
Marchand-Brynaert, J. Bioorg. Med. Chem. Lett. 2005,
5, 3252; (c) Vou e´ , M.; Goormaghtigh, E.; Hombl e´ , F.;
1
Marchand-Brynaert, J.; Conti, J.; Devouge, S.; De
Coninck, J. Langmuir 2007, 23, 949.
4
. (a) Begg, E. J.; Barclay, M. L.; Kirkpatrick, C. M. Br. J.
Clin. Pharmacol. 2001, 52, 35S; (b) Davani, S.; Muret, P.;
Royer, B.; Hoen, B.; Kantelip, J. P. Ann. Biol. Clin.
(
Paris) 2002, 60, 655.
. Williams, D. H. Nat. Prod. Rep. 1996, 13, 469.
14. The ordinate shows the variation in the number of viable
bacteria after 5 h of incubation compared to the original
inoculum (horizontal dotted line). Negative values indi-
cate bacterial killing due to the action of vancomycin (the
values of the negative plateaus are similar to what was
observed with vancomycin alone [not shown]). Positive
values correspond to bacterial growth due to the decreased
efficacy of vancomycin in the presence of the tested
compounds (the values of the positive plateaus are similar
to what was observed in the absence of vancomycin [not
shown]. Curves were generated by fitting a variable slope
sigmoidal dose-response equation using GrapPad Prism
(version 4.03, GraphPad Software Inc., San Diego, CA) to
the experimental points. EC50 values (shown on Table 1)
5
6
. Gilbert, Y.; Deghorain, M.; Wang, L.; Xu, B.; Pollheimer,
P. D.; Gruber, H. J.; Errington, J.; Hallet, B.; Haulot, X.;
Verbelen, C.; Hols, P.; Dufr eˆ ne, Y. F. Nano Lett. 2007, 7,
796.
. Evers, S.; Quintiliani, R., Jr.; Courvalin, P. Microb. Drug
Resist. 1996, 2, 219.
. Folena-Wasserman, G.; Sitrin, R. D.; Chapin, F.; Snader,
K. M. J. Chromatogr. 1987, 392, 225.
. Montalbetti, C. A. G.; Falque, V. Tetrahedron 2005, 61,
7
8
9
1
0827.
0. Spectral characteristics of final products. (2a) H NMR
300 MHz, D O) d 1.09–1.67 (m, 21H); 2.27 (t, J = 7.5 Hz,
1
1
(
2

5762
O. Hernout et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5758–5762
and slope factors (spanning from 2.83 to 3.92) were
calculated with the same software for each compound.
Each point represents the arithmetic mean of triplicates
15. Nieto, M.; Perkins, H. R. Biochem. J. 1971, 23, 789.
16. Billot-Klein, D.; Blanot, D.; Gutmann, L.; van Hei-
jenoort, J. Biochem. J. 1994, 304, 1021.
±SD and the experiment was repeated twice with similar
results.
17. Wenk, M.; Vozeh, S.; Follath, F. A review. Clin. Phar-
macokinet. 1984, 9, 475.