Synthesis of modified peptidoglycan precursor analogues for the inhibition of glycosyltransferase

Article abstract of DOI:10.1021/ja302099u

The peptidoglycan glycosyltransferases (GTs) are essential enzymes that catalyze the polymerization of glycan chains of the bacterial cell wall from lipid II and thus constitute a validated antibacterial target. Their enzymatic cavity is composed of a donor site for the growing glycan chain (where the inhibitor moenomycin binds) and an acceptor site for lipid II substrate. In order to find lead inhibitors able to fill this large active site, we have synthesized a series of substrate analogues of lipid I and lipid II with variations in the lipid, the pyrophosphate, and the peptide moieties and evaluated their biological effect on the GT activity of E. coli PBP1b and their antibacterial potential. We found several compounds able to inhibit the GT activity in vitro and cause growth defect in Bacillus subtilis. The more active was C16-phosphoglycerate-MurNAc-(l-Ala-d-Glu)-GlcNAc, which also showed antibacterial activity. These molecules are promising leads for the design of new antibacterial GT inhibitors.

Full text of DOI:10.1021/ja302099u

Article
pubs.acs.org/JACS
Synthesis of Modified Peptidoglycan Precursor Analogues for the
Inhibition of Glycosyltransferase
Shrinivas Dumbre,† Adeline Derouaux,‡ Eveline Lescrinier,† Andre Piette,‡ Bernard Joris,‡
́
Mohammed Terrak,‡ and Piet Herdewijn*,†
†Laboratory of Medicinal Chemistry, Rega Institute for Medical Research, University of Leuven (KU Leuven), Minderbroedersstraat
10, 3000 Leuven, Belgium
‡Centre d’Ingen
́
́ ́ ̀ ́ ̀
ierie des Proteines, Universite de Liege, Allee de la chimie, B6a, B-4000, Sart Tilman, Liege, Belgium
S
* Supporting Information
ABSTRACT: The peptidoglycan glycosyltransferases (GTs)
are essential enzymes that catalyze the polymerization of
glycan chains of the bacterial cell wall from lipid II and thus
constitute a validated antibacterial target. Their enzymatic
cavity is composed of a donor site for the growing glycan chain
(where the inhibitor moenomycin binds) and an acceptor site
for lipid II substrate. In order to find lead inhibitors able to fill
this large active site, we have synthesized a series of substrate
analogues of lipid I and lipid II with variations in the lipid, the pyrophosphate, and the peptide moieties and evaluated their
biological effect on the GT activity of E. coli PBP1b and their antibacterial potential. We found several compounds able to inhibit
the GT activity in vitro and cause growth defect in Bacillus subtilis. The more active was C16-phosphoglycerate-MurNAc-(L-Ala-
D-Glu)-GlcNAc, which also showed antibacterial activity. These molecules are promising leads for the design of new antibacterial
GT inhibitors.
structure−activity relationship studies have been carried out
on moenomycin8−13 which identified the phosphoglycerate
moenocinol trisaccharide as the minimal structural feature
required for antibacterial and GT inhibitory activities.
INTRODUCTION
■
The problem of multidrug resistant (MDR) bacteria is on the
rise.1−3 The difficulty of treating infections due to MDR strains
causes a longer incubation time with antibiotics, during which
microbial resistance is spread to others. This inflicts huge costs
on health care, and potentially increases mortality and
morbidity. To face this threat, new antiobiotics are urgently
needed.
Recently the first structures of GTs and their complexes with
moenomycin became available.14−18 The GT domain is
composed of two lobes (head and jaw domains). The fold of
the head domain shows structural similarities with that of the
phage λ lysozyme (λL). The jaw domain is specific to the GT51
family and has a hydrophobic region partly embedded in the
cytoplasmic membrane to presumably facilitate the access to
the membrane bound lipid II substrate. The active site of GT is
located at the interface of the two lobes and can be divided into
two substrate binding sites: a donor site for the growing glycan
chain and where moenomycin binds, and an acceptor site for
lipid II (Figure 1A).
The length of the lipid moiety is critical for the inhibitory
activity of moenomycin and the processive turnover of the
substrate by GT.19 For the substrate, this property is
determined essentially by the donor site which requires a
lipid longer than 20 carbons (maximum activity observed with
betulaheptaprenyl, C3520), whereas the acceptor site can accept
a shorter lipid.19 Therefore the use of substrate analogues with
reduced lipid chain length and targeting the acceptor site with
high affinity could be a new approach in the design of active GT
inhibitors. As prerequisite to this we need to better understand
Bacteria have a unique cellular feature called the
peptidoglycan (PG) sacculus, a polymeric layer that is the
main constituent of bacterial cell walls and is absent in animal
cells.4 The monomeric unit precursor of the PG consists of N-
acetylmuramic acid (MurNAc) and N-acetylglucosamine
(GlcNAc) linked to an undecaprenyl pyrophosphate lipid
carrier, and a pentapeptide (lipid II, 1, Figure 1A). This
monomeric unit is assembled in the cytoplasm and then flipped
out into the periplasm,5,6 where it is used as a substrate by the
membrane bound glycosyltransferases (GTs) to synthesize the
glycan chains (nascent PG). The growing glycan chains are
cross-linked and attached to a preexisting PG sacculus via an
amide bond formation between the peptide chains by
transpeptidases; this reaction is the target of β-lactam
antibiotics.7 Inhibition of the GT step by the specific inhibitor
moenomycin leads to bacterial death. This makes the GTs a
validated and attractive target for the development of new
antibacterials against antibiotic multiresistant strains. Moeno-
mycin (2, Figure 1B) is not used clinically because of its poor
pharmacokinetic properties, which is related to its phosphogly-
cerate lipid tail. To overcome this limitation, extensive
Received: March 2, 2012
Published: May 2, 2012
© 2012 American Chemical Society
9343
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
Figure 1. Membrane steps of bacterial cell wall biosynthesis and
structure of the GT ligands. (A) Schematic representation of the GT
domain catalyzing the glycan chain elongation. The elongating glycan
chain binds to the donor site, and lipid II (1) binds to the acceptor
site. Catalysis occurs by deprotonation of 4-OH of lipid II by the
catalytic glutamate (E233 in PBP1b) followed by nucleophilic attack
on the C1 of the growing chain and departure of the
undecaprenylpyrophosphate. (B) Structure of moenomycin A (2),
an analogue of the growing glycan chain which binds to the donor site.
Figure 2. Concept and design. Structures of the peptidoglycan
precursor lipid II (1), disaccharide fragment (EFG, 3), and
monosaccharide fragment (FG, 4) of moenomycin. Structures of the
lipid I and lipid II monophosphate analogues that we have synthesized
(18 analogues); these compounds have three main structural features:
(a) a monophosphate lipid chain, as monophosphates are less prone to
hydrolysis than pyrophosphate, (b) with or without a shorter peptide
unit, and (c) a phosphoglycerate as a lipid carrier, as in moenomycin
which mimics the pyrophosphate.
the contribution of the different moieties (lipid, pyrophosphate,
sugars, and peptide) of the lipid II substrate for binding to the
GT enzymatic cavity.
In early 2000, the semisynthesis and total synthesis of lipid I
and lipid II substrates was reported.20−22 However, there have
been limited efforts in designing and synthesizing lipid I/II
based inhibitors for GT.23−26 This is probably due to the
structural complexity and chemical instability of lipid I/II,
which complicates the synthetic efforts of modified lipid I and
lipid II.
In this work we aim to understand the substrate specificity of
E. coli PBP1b and identify substrate analogues able to inhibit
the GT activity of the enzyme. These analogues could be used
as templates for structure-based drug design to optimize
substrates for the donor site, acceptor site, or both. We have
synthesized a series of substrate analogues of lipid I and lipid II
with variations in the lipid, the pyrophosphate, and the peptide
moieties and evaluated their biological effect on the GT activity
of E. coli PBP1b and their antibacterial potential.
carboxamate on position 5 (instead of hydroxymethyl group),
and an hydroxyl group on position 2 (instead of the N-acetyl
group).
We have prepared 18 analogues of lipid I/II as outlined in
Figure 2. Analogues of lipid I (17, 18) and lipid II (19, 20, and
21) were synthesized following the synthetic strategy shown in
Scheme 3. Compound 627 was obtained starting from N-acetyl
glucosamine (5). The benzyl glycoside of N-acetyl glucosamine
was acetylated, and then the benzyl group was removed by
hydrogenolysis to afford 6 (Scheme 1). Compound 6 is the
starting material for the synthesis of lipid I analogues 17 and
18. Phosphorylation of 6 by reaction with dibenzyl N,N-
diisopropylphosphoramidite and 1H-tetrazole, followed by
Scheme 1
RESULTS AND DISCUSSION
■
Design and Synthesis of Lipid I and Lipid II
Monophosphate Analogues. We have synthesized several
lipid I and lipid II analogues, related to the structure of
moenomycin, with minimal molecular fragments. Figure 2
shows the resemblance between the moenomycin EFG
fragment (3, Figure 2) and lipid II (1): both have a GlcNAc
residue as second sugar. The undecaprenyl pyrophosphate
moiety in 1 is replaced by an isoelectronic phosphoglycerate
moenocinol unit in 3. Besides the presence of a phosphogly-
cerate function, the major difference between compound FG
fragment (4) and lipid I is the presence of carbamoyl group on
the 3-hydroxyl group (instead of lactoyl pentapeptide), a
9344
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
Scheme 2
Scheme 3
oxidation with hydrogen peroxide and debenzylation, afforded
728 as a triethylamine salt. Compound 7, after activation with
N,N′-dicyclohexylcarbodiimide or tosyl chloride or 2,4,6-
triisopropylbenzenesulfonyl chloride29 and treatment with
prenol, failed to give the desired alkylated phosphate analogue:
invariably a pyrophosphate dimer was formed. Therefore, an
alternative approach was considered in which the nucleophil-
icity of the phosphate group is enhanced by using bulky tetra-n-
butylammonium as counterion.30 Thus compound 7 was mixed
with 2.1 equiv of tetra-n-butylammonium hydroxide and
lypholized to obtain compound 8 as a bis(tetra-n-butyl)-
ammonium salt. This salt underwent smooth alkylation of the
phosphate function with prenyl bromide in the presence of 4 Å
molecular sieves to afford the prenylated phosphate compound,
which upon deacetylation via saponification afforded 17 in 87%
yield.
For the preparation of the lipid II disaccharide mono-
phosphate prenyl analogue 19, the synthesis of the disaccharide
monophosphate 16 is required. This synthesis was carried out
by following a modified version of a published protocol,31,32 as
illustrated in Scheme 2. This involves the synthesis of the
trichloroacetimidate-activated donor (GlcNAc) 11 and accept-
or (MurNAc) 14. Schmidt’s glycosylation33 reaction between
11 and 14, followed by removal of the 6-O-benzyl and N-Troc
9345
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
groups, their reprotection as acetates, and the removal of the
anomeric O-benzyl group by hydrogenolysis, afforded the lipid
II disaccharide precursor 15. Compound 15 was phosphory-
lated, debenzylated, and converted to its bis(tetra-n-butyl)-
ammonium salt (16) in analogy to the procedure described for
8. We further utilized the bis(tetra-n-butyl)ammonium salts 8
and 16 to obtain the farnesyl lipid I (18), the prenyl lipid II
(19), the farnesyl lipid II (20), and the cetyl lipid II (21)
monophosphate analogues (see Scheme 3).
Scheme 5
The GT substrate analogues 17 and 19 bearing a prenyl
chain and the compounds 18 and 20 did not show inhibition of
E. coli PBP1b GT activity (data not shown). Compounds 18
and 20 have a tendency to aggregate, which may contribute to
their inactivity. Compound 21 with a saturated C16 lipid chain
shows a Ki value of 26 6 μM (see the section GT Inhibition
and Antibacterial Activities of Lipid I and Lipid II Analogues
and Table 2). This promising initial result encouraged us to
increase the molecular complexity of lipid I/II analogues by
incorporating suitable peptides and phosphoglycerate saturated
C16 moieties.
The Boc protected L-Ala-D-Glu (22) and L-Ala-γ-D-Glu-L-Lys
(25) peptides, having base labile protecting groups in the side
chains, were synthesized using standard solution phase
protocols as shown in Scheme 4. The dimethyl ester
sodium hydride and cetyl bromide to afford 28 in 87% yield. It
is worth noting that the initial efforts to remove the PMB group
at room temperature and dichloromethane as a solvent was
found to be poor yielding. However, the reaction when carried
out in the presence of pH 7 phosphate buffer at 5−10 °C,
followed by careful removal of DDQ-2H and work up by
precipitation with diethyl ether, afforded the required product
in 84% yield. This PMB deprotected compound was
subsequently oxidized and methylated and TBDPS depro-
tection afforded the desired glycerate cetyl chain (30) in 47%
yield over four steps. The glycerate geranyl compound 31 was
prepared in the same way.
Scheme 4
The alkylation strategy used to prepare 21 is poor, yielding
10% (Scheme 3), and the strategy does not allow peptide
coupling regioselectively, as the free phosphate form will
compete for the peptide coupling reaction. A new synthetic
strategy was devised, which utilizes phosphoramidite chemistry
for the preparation of monophosphate triester.36
Bisdiisopropyl amine chlorophosphine was treated with
benzyl alcohol to obtain benzyl diisopropylamine phosphor-
amidite.37 This phosphoramidite reagent was used to prepare
the phosphoramidite 33 and the phosphoglycerate 34 (Scheme
6). For the preparation of lipid I monophosphate analogues
bearing a peptide moeity, MurNAc sugar 37 was obtained from
compound 13b (Scheme 7). Acid mediated benzylidene acetal
removal and reprotection of the two hydroxyl functionalities
with acetates followed by hydrogenolysis of the anomeric
hydrochloride of D-glutamic acid (22) was obtained from D-
glutamic acid with trimethylsilyl chloride in methanol34 and was
coupled to Boc-L-Ala-OH under standard 1-ethyl-3-(3′-
dimethylaminopropyl)carbodiimide (EDCI) conditions to
afford the dipeptide 23 in 78% yield over two steps. Boc-D-
Glu-OBzl was coupled with N6-Fmoc protected methyl ester of
L-Lys to afford the dipeptide 24 in 89% yield. Deprotection,
followed by EDCI mediated coupling with Boc-L-Ala-OH, gave
the protected tripeptide 25 in 83% yield.
Scheme 6
For the synthesis of phosphoglycerates,12,23 we have
developed a protocol starting from D-mannitol (Scheme 5)
that serves as a chiral pool for the glycerate moiety and was
used to obtain acetonide protected alcohol 26 on a multigram
scale.35 The (S)-acetonide protected glycerol 26 was subjected
to p-methoxybenzyl (PMB) protection as it can be selectively
removed under neutral conditions with 2,3-dichloro-5,6-
dicyanobenzoquinone (DDQ). Acetonide deprotection and
regioselective protection of the primary hydroxyl group with t-
butyldiphenylsilyl (TBDPS) gave the key intermediate 27a.
This key intermediate (27a) was cetylated by treatment with
9346
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
Scheme 7
Scheme 8
Scheme 9
Table 1. R1, R2, and Peptide Moieties in Compounds 49−55 and 56−63 (See Scheme 9)
#
*
Yield over three steps. Conversion of 49a to 58 was carried out by saturated NH3 in MeOH, with flash chromatography purification.
benzyl protecting group afforded 37 in 67% yield over three
steps.
The synthesis of lipid I and lipid II monophosphate
analogues without a peptide unit is depicted in Scheme 8.
9347
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
Scheme 10
The tricoordinated phosphorus cetyl compound 33 upon
activation with 1H-tetrazole and treatment with MurNAc sugar
37 followed by oxidation with hydrogen peroxide afforded the
α-phosphorylated compound 38 in 78% yield. Compound 38
on hydrogenolytic removal of the benzyl group and careful
saponification with lithium hydroxide afforded the lipid I
analogue 42 in 65% yield after flash column chromatographic
purification. Peptidoglycan analogues 43, 21, and 44 were
synthesized in the same way. β-Elimination of the ethyl
phenylsulfonyl moiety in compound 38 by 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU) gave compound 45 in
78% yield, affording the free lactoyl moiety for the peptide
coupling reaction.
The synthesis of lipid I and lipid II peptidoglycan
monophosphate analogues containing L-Ala-D-Glu and L-Ala-
γ-D-Glu-L-Lys is shown in Scheme 9. Removal of the Boc group
in dipeptide 23 and coupling of the resultant free amino group
with the lactoyl carboxylic acid group of 45 were performed
using EDCI and 1-hydroxybenzotriazole (HOBt) to afford 49a
in 67% yield. Removal of benzyl and careful saponification of
49a afforded the lipid I monophosphate analogue 56 in 46%
yield over two steps. In a similar manner, compounds 57−63
were obtained from their respective precursors as shown in
Scheme 9 and Table 1. During hydrogenolysis of 53a, 54a, and
55a, apart from the debenzylation reaction (phosphate and
glutamate), also partial removal of the Fmoc group on N6-Lys
was observed. The remaining Fmoc groups were removed
during the saponification reaction. Once the peptide is coupled
to the sugar, the following reactions need to be carefully
monitored by low-resolution mass spectroscopy, 31P NMR, and
thin layer chromatography, in order to obtain good yields. The
purification of the final monophosphate diester compounds was
found to be cumbersome. Flash column chromatography and
BioGel-P2 purification yielded the final compounds in
moderate yield. HPLC purification of these compounds on a
reverse phase column under various conditions resulted in a
complete loss of material (57, 58, 59, and 60) or the
recuperation of 51 and 52 in very low yields.
followed by mild oxidation in the presence of N-methylmor-
pholine (NMM)−CCl4 and careful saponification, gave the
desired analogue 64. In a similar way, the disaccharide
monophosphate geranyl glycerate analogue 65 was obtained
from 15.
GT Inhibition and Antibacterial Activities of Lipid I
and Lipid II Analogues. Monophosphate Analogues.
Monophosphate analogues of lipid I and lipid II carrying a
saturated 16 carbon lipid chain without peptide stems or having
a dipeptide or tripeptide (Schemes 8 and 9; 42, 21, 56, 57, 58,
59, and 60) have been evaluated for the inhibition of the GT
activity and antibacterial effect. They were all able to inhibit the
GT activity of PBP1b in vitro to some extent (residual activities
between 7 and 84% at 500 μM compound). The disaccharide
analogues (21, 57) are about 2-fold better inhibitors than their
cognate monosaccharide forms (42, 56), and the inhibition of
GT activity by lipid I and lipid II analogues was found to
decrease in a similar manner as the length of the peptide
increased (zero to three residues, 42, 56, and 59 or 21, 57, and
60; Table 2). The inhibitory activity of the mono- or
disaccharide C16-phosphate without peptide was about 6- or
2.5-fold higher than that with tripeptide and dipeptide,
Table 2. Peptidoglycan Glycosyltransferase Inhibition and
Antibacterial Activities of the Peptidoglycan Precursor
a
Analogues
compd
RA (%)
Ki (μM)
MIC (μg/mL)
cell shape
42
56
58
59
43
61
21
57
60
44
62
63
64
65
14
43
41
84
3.5
15
7
51.5 10.5
>128
>128
>128
>128
>128
>128
>128
>128
>128
>128
128
C
N
N
N
C
N
C
C
N
C
L
nd
nd
nd
21
46 7.5
26
9
6
16
40
2.3
1.7
6.5
102
89
48 11
nd
32.7
17.6
36
5
All efforts to couple 35 with lipid I precursor 37 failed to give
the corresponding monophosphate triester. The PMB protect-
ing group for 35 was selected as it could have been easily
removed using DDQ without affecting the double bond.
Likewise the reaction between 37/15 and 36 failed. In both
cases only starting material could be recovered. Also the
reaction between the phosphoramidite of 37 and the glycerate
geranyl alcohol 31 failed in our hands. Therefore, H-
phosphonate chemistry was employed38 to synthesize the
desired lipid I and lipid II phosphoglycerate geranyl analogues
as shown in Scheme 10. Compound 37 upon treatment with 2-
chloro-4H-1,3,2-benzodioxaphosphorin-4-one gave the H-phos-
phonate which upon capping and activation with adamantane-
carbonyl chloride39 and nucleophilic substitution with 31,
2
8
>128
>128
>128
N
nd
nd
nd
nd
a
GT inhibition assays were performed with E. coli PBP1b. RA = GT
residual activity of PBP1b at 500 μM compounds. The data represent
mean values of three independent experiments. Standard deviations
were within 10% of these mean values (RA). Minimum inhibitory
concentration (MIC) values and cell shapes are for B. subtilis 168. Cell
shapes: C, chaining cells; N, normal cells; L, cell Lysis. nd, not
determined. Similar MIC values were obtained with S. aureus ATCC
25923. The MIC value of moenomycin disaccharide EF was >100 μg/
mL against several Gram-positive bacteria and IC50 0.24 μM using E.
coli membrane.40
9348
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
Figure 3. Cellular effect of the compounds. The cells (OD600 nm = 0.1−0.2) were incubated with 256 μg/mL compounds and the samples were taken
at 1 h interval over 3 h and stained with the membrane dye FM-1-43-FX. The images show the effect of the compounds indicated by their code
number after 3 h on B. subtilis cells observed by using both phase contrast and fluorescence microscopy.
respectively. These results show that GlcNAc plays a significant
role in binding to the enzyme and that in contrast the di- or
tripeptide has a negative effect. Note that the monosaccharide
dipeptide with amidated glutamate (58) behaves as the free
carboxyl form (56).
against E. coli and have no effect, probably because of the outer
membrane barrier (data not shown).
Phosphoglycerate Analogues. The structures of E. coli
PBP1b and other GTs in complexes with moenomycin show
that the EF ring phosphoglycerate portion makes the major
interactions with the active site residues of the GT domain.17 It
is believed that the phosphoglycerate moiety of moenomycin
could mimic the pyrophosphate group of the substrate. In order
to improve the inhibitory activity of our compounds, we have
prepared a new series of compounds with C16-phosphogly-
cerate (Schemes 9 and 10) as a C-1 substituent.
The biological evaluation shows that all phosphoglycerate
compounds (43, 44, 61, 62, and 63) had increased inhibitory
activity of the GT (3−10-fold approximately) compared to
their cognate monophosphate analogues (Table 2). Interest-
ingly, the disaccharide dipeptide C16-phosphoglycerate (62)
was about 2-fold more active than the peptide-free disaccharide
C16-phosphoglycertae (44, Ki = 17.6 vs 32.7 μM) in contrast to
the monophosphate forms where the peptide-free compound
was more active. On the other hand, the antibacterial activity of
all the compounds has virtually not changed (MIC > 128 μg/
mL) except that of the disaccharide dipeptide (62), which has a
MIC value of 128 μg/mL in accordance with the in vitro
activity of this compound. This result was further confirmed by
looking at the cells under the microscope; the disaccharide
dipeptide did not induce cell chaining but instead provoked cell
lysis at 1−2 times the MIC (data not shown). The disaccharide
tripeptide (63) had a minor or no effect on cell morphology,
while the disaccharide phosphoglycerate without peptide (44)
induced cell chaining (Figure 3). For the monosaccharide C16-
phosphoglycerate compounds only the one without peptide
In terms of antibacterial activity, all the compounds (mono-
and disaccharide C16-phosphate with or without peptide)
showed MIC values higher than 128 μg/mL against Staph-
ylococcus aureus ATCC 25923 and Bacillus subtilis 168 (Table
2). This activity is comparable to the MIC value (>100 μg/mL)
of moenomycin disaccharide EFG derivative (close analogue of
compound 21, Figure 2 and Scheme 8) for Gram-positive
bacteria tested40 despite the presence of a C16-phosphate lipid
tail in compound 21 instead of C25-phosphoglycerate in
moenomycin which is known to a play important role in its
activity. Surprisingly, observation of the B. subtilis cells in the
microscope after 16−18 h of incubation with compounds
shows that the most active compounds (42, 21, and 57)
induced cell chaining (Figure 3). Dead cells and curved
filament have been also observed in some cases. The
monosaccharide dipeptide (56) and monosaccharide tripeptide
(59) (the less active) do not have any effect on cell
morphology. In contrast, the cells treated with the disaccharide
dipeptide (57) grow as long filaments of unseparated cells,
while the disaccharide tripeptide (60) had no effect.
Membrane labeling of the cells with the fluorescent dye FM-
1-43-FX shows that most of the cells are in an advanced stage
of cytokinesis but remained attached to each other by their
poles. Bright fluorescent spots and probably uncompleted septa
have been also observed (Figure 3). These observations show
that cell division is perturbed and that cell separation is blocked
by these compounds. The compounds have been also tested
9349
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
(43) produced long cell chains while the dipeptide (61) form
had a minor or no effect.
Overall, the data show a nice correlation between GT
inhibition activity of a compound and its effect on bacterial
cells, indicating that these morphological defects are due to GT
inhibition.
investigation to understand the exact mode of action of these
compounds including the determination of the crystal
structures of their complexes with a GT.
This will provide a good template for structure-based design
of new GT inhibitors that encompass the extended polar sugar
binding sites of the enzyme cavity. The affinity of such
compounds for the enzyme could be further improved by
bridging the donor and acceptor sites with sugar ligands with
increased complexity, which could compensate for the loss or
reduction of the lipid chain length, thus circumventing the
negative properties of the long lipid chain observed for
moenomycin.
In addition to C16 analogues, two phosphoglycerate geranyl
substituted compounds, a monosaccharide and a disaccharide
without peptide (64, 65), have been prepared and compared
with their cognate monophosphate C16 analogues (42, 21;
Table 2). Interestingly, the disaccharide phosphoglycerate
geranyl exhibits weak activity (11% of inhibition at 500 μM)
while the geranyl monophosphate was completely inactive. This
suggests that the phosphoglycerate improves binding by
mimicking the pyrophosphate but that the length/substitution
of the lipid is a more determinant factor.
Inactivation of all GTs (four class A PBPs) in B. subtilis was
not lethal, and since the mutant was still able to produce
peptidoglycan and does not have a monofunctional GT, it has
been proposed that B. subtilis contains an unknown GT
different in sequence from the GT51 family.41 PG synthesis by
the quadruple mutant was demonstrated in vitro and was
shown to be sensitive to moenomycin, confirming the GT
reaction as an antibacterial target.
Interestingly, the quadruple mutant cells have been shown to
be filamentous and bent with irregularly spaced and
uncompleted septa visualized by membrane staining and
fluorescence microscopy. Similarly, the exposure of the wild-
type strain to moenomycin resulted in a phenotype similar to
that of the quadruple mutant. The similarities between our
results and those of McPherson and Popham41 support the
conclusion that the cellular defects we observed are the result of
the GT inhibition. In the meantime we cannot exclude that, in
addition to the GTs, these compounds also inhibit other
enzymes of the peptidoglycan metabolism.
MATERIALS AND METHODS
■
Reagents and Protein. Radiolabeled [14C-D-Ala]-lipid II (0.078
μCi nmol−1), [N-acetylglucosaminyl-N-acetylmuramoyl (L-Ala-γ-D-
Glu-(L)-meso-A2pm-(L)-D-Ala-D-Ala)-pyrophosphate-undecaprenol]
meso-diaminopimelic acid, meso-A2pm, was prepared essentially as
previously described in ref 42. All compounds were solubilized in
dimethyl sulfoxide (DMSO; 5−10 mM). Escherichia coli PBP1b was
produced and purified as described in ref 43.
In Vitro Glycosyltransferase Activity and Inhibition Assays.
In vitro peptidoglycan polymerization assay was performed by
incubation of 2.5 μM D-[14C]Ala-D-[14C]Ala-labeled lipid II (0.078
μCi nmol−1) and purified PBP1b enzyme (100 nM) at 30 °C in 50
mM Tris·HCl, pH 7.5, 200 mM NaCl, 0.2% decylPEG, 10 mM CaCl2,
and 20% DMSO. The reactions were stopped after 7 min with
moenomycin (10 μM), and the products (unused lipid II substrate and
polymerized peptidoglycan) were separated by TLC on silica gel plates
(Fluka) using 2-propanol/ammonium hydroxide/water (6:3:1; v/v/v)
as a mobile phase.44 Lipid II moves with the solvent (Rf = 0.65), and
all the polymerized material remains at the origin in one spot (Rf = 0).
The TLC plates were exposed to a storage phosphor screen (GE
Healthcare) for 16 h. The images were revealed using a Typhoon Trio
+ imager and analyzed using the ImageQuant TL software (GE
Healthcare).
The initial GT inhibition assay was carried out by measuring the
residual activity (RA) of E. coli PBP1b in the presence of 500 μM
compounds. All assays were repeated at least three times. The IC50
values were determined when the RAs were <20% at 500 μM
compounds. Using variable concentrations of inhibitor (0−500 μM),
the concentration that caused 50% of reduction in the RA of an
enzyme was considered as IC50. Ki values were determined using the
equation IC50 = Ki(1+ [s]/Km) and a Km value of 1.8 μM.42
MIC Determination and Microscopy. Minimum inhibitory
concentration (MIC) determinations were performed by the broth
microdilution method against S. aureus ATCC 25923 and B. subtilis
168 according to the EUCAST (European Committee on Anti-
microbial Susceptibility Testing)/CLSI (Clinical and Laboratory
Standard Institute) recommended procedures. The compounds were
solubilized in 100% DMSO at a concentration of 5−10 mg/mL, and
diluted in Mueller−Hinton broth (MHB), just before utilization.
Cells from the samples used for MIC determination were fixed on
glass slabs coated with 1% agarose and examined in the micro-
scope.45,46 This experiment was repeated by incubation of new
cultured cells (DO600 nm = 0.1−0.2) with 256 μg/mL compounds, and
the samples were taken at 1 h interval over 3 h and incubated with the
fluorescent dye FM-1-43-FX (Invitrogen) to label the membranes. The
cells were then fixed with formaldehyde and glutaraldehyde, washed
with PBS, and observed under a fluorescent microscope (Zeiss Axio
Imager.Z1). Photographs were taken with a cooled AxioCam MRm
(Zeiss) mounted on the fluorescent microscope through an EC Plan-
Neofluar 100 × 1.3 oil immersion objective in bright field and
fluorescence using a filter set 37 (425−475-nm band-pass excitation
and 485−535-nm band-pass emission; Zeiss). Images were analyzed
using the AxioVision Rel. 4.5 (Zeiss) software as previously
described.46
CONCLUDING REMARKS
■
A systematic investigation of a GT enzyme inhibition by
peptidoglycan precursor analogues with variations in the
carbohydrate, lipid, pyrophosphate, and peptide moieties
showed that saturated C16 phosphoglycerate linked to
MurNAc-GlcNAc (44) and to the disaccharide dipeptide (L-
Ala-D-Glu) (62) are good inhibitors of the GT activity in vitro
and are able to induce growth defects or lysis of bacterial cells.
The antibacterial activities observed with the precursor
analogues in this work correlate with those of the inactivation
of the GT or their inhibition by moenomycin in B. subtilis41 and
suggest that these compounds inhibit the GT activity in vivo.
The disaccharide compounds (44 and 62) are lipid II substrate
mimics and are expected to bind to both donor and acceptor
sites of GT. They would have binding similar to that of the
enzyme as natural substrates with the peptide (in compound
62) pointing away from the active site.15 During the initiation
of a glycan chain elongation by the GT, lipid II binds to both
the donor and acceptor sites; subsequently, during the
elongation phase and due to the processivity of the enzyme,
chain elongation occurs in the donor site by successive
additions of a disaccharide unit from the acceptor site bound
lipid II. Therefore, the analogue compounds have to compete
with lipid II for binding to the donor site and acceptor site at
the initiation phase, and then during the elongation phase the
compounds have to compete also with the elongating chain.
This shows the complexity of the system and will require more
9350
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351

Journal of the American Chemical Society
Article
(19) Perlstein, D. L.; Andrew Wang, T.-S.; Doud, E. H.; Kahne, D.;
Walker, S. J. Am. Chem. Soc. 2010, 132, 48.
(20) Ye, X. Y.; Lo, M. C.; Brunner, L.; Walker, D.; Kahne, D.; Walker,
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and characterization data for all new
compounds used in this study. This material is available free of
charge via the Internet at http://pubs.acs.org.
S. J. Am. Chem. Soc. 2001, 123, 3155.
(21) VanNieuwenhze, M. S.; Mauldin, S. C.; Zia-Ebrahimi, M.;
Aikins, J. A.; Blaszczak, L. C. J. Am. Chem. Soc. 2001, 123, 6983.
(22) VanNieuwenhze, M. S.; Mauldin, S. C.; Zia-Ebrahimi, M.;
Winger, B. E.; Hornback, W. J.; Saha, S. L.; Aikins, J. A.; Blaszczak, L.
C. J. Am. Chem. Soc. 2002, 124, 3656.
AUTHOR INFORMATION
■
Corresponding Author
Piet.Herdewijn@rega.kuleuven.be
(23) Hecker, S. J.; Minich, M. L.; Lackey, K. J. Org. Chem. 1990, 55,
4904.
(24) Qiao, L.; Vederas, J. C. J. Org. Chem. 1993, 58, 3480.
(25) Cudic, P.; Behenna, D. C.; Yu, M. K.; Kruger, R. G.; Szewczuk,
L. M.; McCafferty, D. G. Bioorg. Med. Chem. Lett. 2001, 11, 3107.
(26) Garneau, S.; Qiao, L.; Chen, L.; Walker, S.; Vederas, J. C. Bioorg.
Med. Chem. 2004, 12, 6473.
(27) Mohan, H.; Vasella, A. Helv. Chim. Acta 2000, 83, 114.
(28) Heidlas, J. E.; Lees, W. J.; Pale, P.; Whitesides, G. M. J. Org.
Chem. 1992, 57, 146.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the IAP 6/19 project, and the
Fonds de la Recherche Fondamentale Collective (FRFC Nos.
2.4506.08 and 2.4543.12). M.T. is a Research Associate of the
National Fund for Scientific Research (F.R.S_FNRS, Belgium).
(29) Itakura, K.; Katagiri, N.; Bahl, C. P.; Wightman, R. H.; Narang,
S. A. J. Am. Chem. Soc. 1975, 97, 7327.
̀
We thank Martine Nguyen-Disteche for helpful discussion and
(30) Davisson, V. J.; Woodside, A. B.; Neal, T. R.; Stremler, K. E.;
Muehlbacher, M.; Poulter, C. D. J. Org. Chem. 1986, 51, 4768.
(31) Lioux, T.; Busson, R.; Rozenski, J.; Nguyen-Disteche, M.; Frere,
J. M.; Herdewijn, P. Collect. Czech. Chem. Commun. 2005, 70, 1615.
(32) Fraipont, C.; Sapunaric, F.; Zervosen, A.; Auger, G.; Devreese,
B.; Lioux, T.; Blanot, D.; Mengin-Lecreulx, D.; Herdewijn, P.; Van
Beeumen, J.; Frere, J. M.; Nguyen-Disteche, M. Biochemistry 2006, 45,
4007.
Marie Schloesser for technical assistance. We thank Prof. Jef
Rozenski for mass spectrometric analysis and Luc Kerremans
for technical assistance. Mass spectrometry was made possible
by the support of the Hercules Foundation of the Flemish
Government (Grant 20100225-7).
REFERENCES
■
(33) Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. Engl. 1980, 19,
731.
(1) Levy, S. B.; Marshall, B. Nat. Med. 2004, 10, S122.
(2) WHO, World Health Organization. Antimicrobial Resistance
[Online]. 2011.http://www.who.int/mediacentre/factsheets/fs194/
en/index.html, No. 194.
(34) Li, J. B.; Sha, Y. W. Molecules 2008, 13, 1111.
(35) Thomas, B. N.; Lindemann, C. M.; Corcoran, R. C.; Cotant, C.
L.; Kirsch, J. E.; Persichini, P. J. J. Am. Chem. Soc. 2002, 124, 1227.
(36) Michalski, J.; Dabkowski, W. Top. Curr. Chem. 2004, 232, 93.
(37) Dreef-Tromp, C. M.; Lefeber, A. W. M.; van der Marel, G. A.;
van Boom, J. H. Synthesis 1992, 1992, 1269.
(3) Zhang, Q.; Lambert, G.; Liao, D.; Kim, H.; Robin, K.; Tung, C.
K.; Pourmand, N.; Austin, R. H. Science 2011, 333, 1764.
(4) Vollmer, W.; Blanot, D.; de Pedro, M. A. FEMS Microbiol. Rev.
2008, 32, 149.
(5) Mohammadi, T.; van Dam, V.; Sijbrandi, R.; Vernet, T.; Zapun,
A.; Bouhss, A.; Diepeveen-de Bruin, M.; Nguyen-Disteche, M.; de
Kruijff, B.; Breukink, E. EMBO J. 2011, 30, 1425.
(6) Bouhss, A.; Trunkfield, A. E.; Bugg, T. D.; Mengin-Lecreulx, D.
FEMS Microbiol. Rev. 2008, 32, 208.
(7) Sauvage, E.; Kerff, F.; Terrak, M.; Ayala, J. A.; Charlier, P. FEMS
Microbiol. Rev. 2008, 32, 234.
(8) Fuse, S.; Tsukamoto, H.; Yuan, Y.; Wang, T. S.; Zhang, Y.; Bolla,
M.; Walker, S.; Sliz, P.; Kahne, D. ACS Chem. Biol. 2010, 5, 701.
(9) Ostash, B.; Walker, S. Nat. Prod. Rep. 2010, 27, 1594.
(10) Welzel, P. Chem. Rev. 2005, 105, 4610.
(11) Ritzeler, O.; Hennig, L.; Findeisen, M.; Welzel, P.; Muller, D.
Tetrahedron 1997, 53, 1665.
(12) Kahne, D. E.; Adachi, M.; Zhang, Y.; Leimkuhler, C.; Sun, B. Y.;
LaTour, J. V. J. Am. Chem. Soc. 2006, 128, 14012.
(13) Sofia, M. J.; Allanson, N.; Hatzenbuhler, N. T.; Jain, R.; Kakarla,
R.; Kogan, N.; Liang, R.; Liu, D. S.; Silva, D. J.; Wang, H. M.; Gange,
D.; Anderson, J.; Chen, A.; Chi, F.; Dulina, R.; Huang, B. W.; Kamau,
M.; Wang, C. W.; Baizman, E.; Branstrom, A.; Bristol, N.; Goldman,
R.; Han, K. H.; Longley, C.; Midha, S.; Axelrod, H. R. J. Med. Chem.
1999, 42, 3193.
(38) Stawinski, J.; Stroberg, R. Oligonucleotide Synthesis. Methods and
̈
Applications; Herdewijn, P., Ed.; Methods in Molecular Biology 288;
Humana Press: Totowa, NJ, USA, 2005; p 81.
(39) Andrus, A.; Efcavitch, J. W.; McBride, L. J.; Giusti, B.
Tetrahedron Lett. 1988, 29, 861.
(40) Goldman, R. C.; Baizman, E. R.; Branstrom, A. A.; Longley, C.
B. Bioorg. Med. Chem. Lett. 2000, 10, 2251.
(41) McPherson, D. C.; Popham, D. L. J. Bacteriol. 2003, 185, 1423.
(42) Terrak, M.; Ghosh, T. K.; Van Heijenoort, J.; Van Beeumen, J.;
Lampilas, M.; Aszodi, J.; Ayala, J. A.; Ghuysen, J.-M.; Nguyen-
Disteche, M. Mol. Microbiol. 1999, 34, 350.
̀
(43) van Heijenoort, Y.; Gomez, M.; Derrien, M.; Ayala, J.; van
Heijenoort, J. J. Bacteriol. 1992, 174, 3549.
(44) Derouaux, A.; Turk, S.; Olrichs, N. K.; Gobec, S.; Breukink, E.;
Amoroso, A.; Offant, J.; Bostock, J.; Mariner, K.; Chopra, I.; Vernet,
T.; Zervosen, A.; Joris, B.; Frere, J. M.; Nguyen-Disteche, M.; Terrak,
M. Biochem. Pharmacol. 2011, 81, 1098.
(45) Koppelman, C.-M.; Aarsman, M. E. G.; Postmus, J.; Pas, E.;
Muijsers, A. O.; Scheffers, D.-J.; Nanninga, N.; Den Blaauwen, T. Mol.
Microbiol. 2004, 51, 645.
̈
(46) Pastoret, S.; Fraipont, C.; den Blaauwen, T.; Wolf, B.; Aarsman,
M. E.; Piette, A.; Thomas, A.; Brasseur, R.; Nguyen-Disteche, M. J.
Bacteriol. 2004, 186, 8370.
(14) Lovering, A. L.; de Castro, L. H.; Lim, D.; Strynadka, N. C.
Science 2007, 315, 1402.
(15) Lovering, A. L.; Gretes, M.; Strynadka, N. C. Curr. Opin. Struct.
Biol. 2008, 18, 534.
(16) Heaslet, H.; Shaw, B.; Mistry, A.; Miller, A. A. J. Struct. Biol.
2009, 167, 129.
(17) Wong, C. H.; Sung, M. T.; Lai, Y. T.; Huang, C. Y.; Chou, L. Y.;
Shih, H. W.; Cheng, W. C.; Ma, C. Proc. Natl. Acad. Sci. U.S.A. 2009,
106, 8824.
(18) Yuan, Y.; Barrett, D.; Zhang, Y.; Kahne, D.; Sliz, P.; Walker, S.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 5348.
9351
dx.doi.org/10.1021/ja302099u | J. Am. Chem. Soc. 2012, 134, 9343−9351