Design and evaluation of mechanism-based inhibitors of D-alanyl-D-alanine dipeptidase VanX

Article abstract of DOI:10.1002/1099-0690(200211)2002:21<3573::AID-EJOC3573>3.0.CO;2-V

VanX protein is a D-alanyl-D-alanine dipeptidase essential for vancomycin resistance in Enterococcus. It is also a key drug target in circumventing clinical glycopeptide antibiotic resistance. The dipeptide-like compound D-Ala-D- Gly(SC₆H₄-p-CHF₂) (1) was recently reported as the first mechanism-based inhibitor with high affinity for the enzyme but poor inhibitory efficiency (k_inact/K_irr = 9320 m^-1 s^-1) and a high partition ratio (7600). In order to assess the effects of variations in aromatic substituents on the in vitro inhibition properties of 1, we have prepared a few derivatives and analyzed them using the alternative substrate D-Ala-D-Gly(SPh)-OH (15) designed according to a similar strategy. Whereas moving the electrophilic difluoromethyl group to the ortho position slightly improved the inhibition parameters (k_inact/ K_irr = 1140 M^-1 s^-1) with a small decrease in the partition ratio (7200), introduction of a methoxy group in D-Ala-D-ortho/para-(difluoromethyl)phenylthioglycines resulted in a marked decrease in inhibitory efficiency. These results demonstrate the difficulties in improving the leading compound 1 given the restricted space available at the VanX active site. Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.

Full text of DOI:10.1002/1099-0690(200211)2002:21<3573::AID-EJOC3573>3.0.CO;2-V

FULL PAPER
Design and Evaluation of Mechanism-Based Inhibitors of
Dipeptidase VanX
D
-Alanyl-
D
-alanine
[a]
[a]
[a]
[a]
¨
Loıc Yaouancq, Maria Anissimova, Marie-Ange Badet-Denisot, and Bernard Badet*
Dedicated to Professor Marc Julia on the occasion of his 80th birthday
Keywords: Antibiotics / Resistance / Enzymes / Inhibitors / α-[(Difluoromethyl)arylthio]glycine
VanX protein is a
for vancomycin resistance in Enterococcus. It is also a key
D
-alanyl-
D
-alanine dipeptidase essential
OH (15) designed according to a similar strategy. Whereas
moving the electrophilic difluoromethyl group to the ortho
drug target in circumventing clinical glycopeptide antibiotic
position slightly improved the inhibition parameters (kinact
Kirr = 1140
−1 s−1) with a small decrease in the partition ratio
(7200), introduction of a methoxy group in -Ala- -ortho/
para-(difluoromethyl)phenylthioglycines resulted in
marked decrease in inhibitory efficiency. These results dem-
onstrate the difficulties in improving the leading compound
1 given the restricted space available at the VanX active site.
( Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
2002)
/
resistance.
The
dipeptide-like
compound
D-Ala-D-
M
Gly(SC6H4-p-CHF2) (1) was recently reported as the first
D
D
mechanism-based inhibitor with high affinity for the enzyme
a
but poor inhibitory efficiency (kinact/Kirr = 9320
m
−1 s−1) and a
high partition ratio (7600). In order to assess the effects of
variations in aromatic substituents on the in vitro inhibition
properties of 1, we have prepared a few derivatives and ana-
lyzed them using the alternative substrate D-Ala-D-Gly(SPh)-
Introduction
cently reported the first mechanism-based inhibitor -Ala-
-Gly(SC6H4-p-CHF2) (1) for VanX.[7] As illustrated in
Scheme 1, enzyme-mediated cleavage of 1 generates 4-thio-
quinone fluoromethide, which is able to react covalently
with enzyme nucleophilic residues, resulting in irreversible
inhibition.
While the affinity of VanX towards 1 is strong, the inac-
tivation efficiency is poor and the partition ratio is high
(kinact/Kirr ϭ 9320 Ϫ1 sϪ1, partition ratio ϭ 7600). As part
of our continuing interest in the search for VanX inhibitors,
we have tried to improve the efficiency of this dipeptide
surrogate by chemical modification of its aromatic ring. We
describe in this paper the syntheses of three novel com-
pounds in which the reactivity of each compound has been
tuned by its aromatic substitution. Also included are the
evaluation of the effects of these compounds on purified
VanX using -Ala--Gly(SPh)-OH as the substrate, as its
turnover product can be detected by 5,5Ј-dithiobis(2Ј-nitro-
benzoic acid) (Ellman’s reagent).
During the three decades since the introduction of glyco-
peptide antibiotics into therapy, enterococci have developed
an incredible resistance to vancomycin, and this resistance
is still growing. Vancomycin acts by binding to the peptido-
glycan precursor at the -Ala--Ala terminus, thus
blocking cell-wall formation.[1] Acquired resistance in bac-
teria requires VanX, a dipeptidase that cleaves -Ala--Ala,
allowing for the synthesis of -Ala--Lac mediated by the
coupled actions of dehydrogenase VanH and ligase VanA.[2]
The new -Ala--Lac precursor can then be integrated into
the peptidoglycan layer, resulting in a 1000-fold weaker
binding of vancomycin compared to -Ala--Ala, the con-
stituent of the wild type. This essential role of VanX was
outlined by Reynolds et al.,[3] who demonstrated that inser-
tional inactivation of VanX reestablished vancomycin sus-
ceptibility in an Enterococcus faecalis resistant strain. VanX
has since become a key drug target in circumventing clinical
antibiotic resistance.
Phosphorus-containing dipeptide analogs that mimic the
tetrahedral intermediate during hydrolysis of -Ala--Ala
by VanX have been found to behave as potent, reversible
inhibitors of VanX in the micromolar range.[4Ϫ6] We re-
Results
Chemical Synthesis of the Pseudodipeptides
The synthesis of α-thio-substituted glycine from the cor-
responding α-acetoxyglycine was performed using the strat-
egy described by Kingsbury et al.[8] The key step (Scheme 2)
in the syntheses of the α-[(difluoromethyl)arylthio]glycines
12, 13 and 14 resides in the substitution of the acetoxy
[a]
Institut de Chimie des Substances Naturelles,
CNRS-UPR 2301,
1, Avenue de la Terrasse, 91198 Gif-sur-Yvette, France
Fax: (internat.) ϩ 33-1/69077247
E-mail: Bernard.Badet@icsn.cnrs-gif.fr
Eur. J. Org. Chem. 2002, 3573Ϫ3579  2002 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ193X/02/1121Ϫ3573 $ 20.00ϩ.50/0
3573

L. Yaouancq, M. Anissimova, M.-A. Badet-Denisot, B. Badet
FULL PAPER
Scheme 1. Principle of mechanism-based inactivation of VanX with compound 1
Scheme 2. Synthesis of pseudodipeptides
group of the racemic pseudodipeptide 2[7] by the corres- respectively, in 40Ϫ60% yield. Subsequent reverse-phase
ponding thio-substituted benzaldehydes 3, 4 and 5. Com- preparative HPLC gave the analytically pure diastereoiso-
pounds 3 and 4 were obtained by the NewmanϪKwart re- mers, each with an overall yield of 10Ϫ15% from 2 after lyo-
arrangement[11,12] of O-dimethylthiocarbamoylated vanil- philization.
lin[9] and ortho-vanillin,[10] respectively. 2-Thiosalicylal-
Attempts to determine the absolute configuration of the
dehyde (5) was conveniently synthesized in 32% yield by carbon atom carrying the sulfur-containing side chain of
formylation of the thiophenol dilithium salt.[13,14] Nucleo- each diastereoisomer of 12, 13 and 14 by X-ray analysis was
philic displacement of the acetate in 2 was accomplished hampered by the difficulty in obtaining suitable crystals of
under standard conditions (Et3N/DMF)[15] to afford 6 and these compounds. However, the difference in retention time
7 in 45Ϫ55% yield.
in HPLC,[8] variations in 1H NMR spectroscopic data[16]
Unexpectedly, compound 5 reacted differently under and the difference of yields for each stereoisomer,[17] as re-
these conditions, giving only an unidentified compound in ported in the literature, have allowed the tentative assign-
which the aldehyde function has disappeared. However, re- ment of the stereochemistry as shown in Scheme 2. Con-
placing DMF by CHCl3 resulted in the suppression of this sidering the peptide bond as a single bond between the α-
unwanted reaction, leading to the formation of 8 with a carbon atoms, as proposed by Kingsbury et al.,[8] the side
78% yield. Compounds 6,7 and 8 were then fluorinated by chains would have the trans and gauche relationships in the
DAST or deoxo-fluorTM, the resulting products of which ,- and ,-peptides, respectively. Therefore, the ,-pep-
were deprotected with TFA/AcOH to afford 12, 13 and 14, tide would exist with both lateral chains located on the
3574
Eur. J. Org. Chem. 2002, 3573Ϫ3579

Mechanism-Based Inhibitors of D-Alanyl-D-alanine Dipeptidase VanX
FULL PAPER
same side of the molecule, causing it to be better absorbed tions give straight lines that intercept the positive y axis,
to the RP-18 column than the ,-peptide, in which the indicating saturation kinetics (Figure 1). The calculated
lateral chains are situated trans to one another. This would values of these kinetic parameters are given in Table 1.
suggest the first eluting isomers to be the ,-stereoisomers
1
12a, 13a and 14a (Scheme 2). Various H NMR spectro-
scopic studies of stereoisomeric dipeptides containing one
aliphatic and one aromatic residue[16] have shown that the
β-protons of the aliphatic residue are more shielded in the
,- and ,-dipeptides than in their ,- or ,-analogs. In
all three dipeptides 12, 13 and 14 the signals of the methyl
protons of the alanine residue of isomer a are shifted down-
field by 0.32, 0.44 and 0.13 ppm, respectively, with respect
to their counterparts in isomer b, confirming the assigned
configuration. Finally, the difference in isolated yields be-
tween isomer a (45%) and isomer b (ca. 35%) could sim-
ilarly reflect a preferential cyclization, which is possible only
in the b isomers due to the position of its side chain.
kobs ϭ (kinact [I])/{[I] ϩ Kirr (1 ϩ [S]/Km)}
(1)
Biochemical Evaluation of the Effects of Pseudodipeptides
12, 13 and 14 on Purified VanX
Incubation of VanX (ca. 0.5 µ) in the presence of milli-
molar concentrations of 12a or 14a resulted in total disap-
pearance of the compound within minutes. On the other
hand, isomers 12b and 14b remain unaffected by prolonged
Figure 1. Semi-log plot showing time-dependent loss of activity of
incubation with large amounts of the enzyme. On the basis VanX by -Ala-Gly(SC6H4-o-CHF2)-OH (14a) at 0.1 m (•), 0.25
m (ᮀ), 1 m (᭜), 2 m (᭝), 3m (᭿); inset: double-reciprocal plot
of the reported specificity of VanX for ,-peptides, it can
of VanX inactivation by 14a; values of kobs were obtained from the
be concluded that the first eluting isomers 12a and 14a
most likely have the same ,-configuration.
Interestingly, none of the isomers of compound 13 were
cleaved by VanX.
slope of semilogarithmic plots of residual activity (A/A0) of VanX
incubated in the presence of 2 m -Ala--Ala and variable
amounts of 14a vs. incubation time
VanX inhibition by 1 has previously been analyzed by
quantification of the residual activity using the coupled ac-
tions of -amino acid oxidase and lactate dehydrogenase in
the presence of NADH. In an effort to improve the assay,
we have developed a convenient and sensitive method based
on the ability of the enzyme to recognize dipeptides con-
taining α-substituted (phenylthio)glycine as the C-terminal
amino acid. This assay, which quantifies the released thio-
phenol using Ellman’s reagent, allows continuous mon-
itoring of enzyme activity and is perfectly suited to high-
throughput screening of inhibitors (Scheme 3).[18]
Table 1. Kinetic parameters of VanX inhibition
Upon incubation of 14a (2 m) with VanX, time-depend-
ent inactivation occurred, resulting in greater than 95% in-
hibition after 1 min. Therefore -Ala--Ala was added to
the incubation mixture to reduce the rate of the enzyme-
inhibitor reaction. Incubation of VanX with compounds
12a and 14a under these conditions exhibited a time-de-
pendent loss of enzyme activity (Figure 1) following first-
order kinetics. The data were fitted to Equation (1) accord-
ing to the method of Knight and Waley,[19] where kobs is a
pseudo-first-order rate constant for inactivation, kinact the
maximal rate of inactivation under inhibitor saturation con-
ditions, [I] the concentration of inhibitor at time zero, Kirr
the irreversible inhibition constant for the compound under
study or the concentration of inhibitor giving half the max-
imum rate of inactivation, [S] the concentration of -Ala-
-Ala, and Km the Michaelis constant for -Ala--Ala.
Double-reciprocal plots of kobs versus inhibitor concentra-
Eur. J. Org. Chem. 2002, 3573Ϫ3579
3575

L. Yaouancq, M. Anissimova, M.-A. Badet-Denisot, B. Badet
FULL PAPER
The number of inhibitor molecules necessary to inactiv- termediate species by changing the fluorine atom’s position
ate VanX was determined by incubating a fixed amount of and/or the substitution of the aromatic ring was conceived
enzyme (43 pmol) and variable amounts of inhibitors for and attempted.
7 min in 50 m HEPES at pH ϭ7 and 37 °C. A plot of the
residual activity (A/A0) against the (I/I0) ratio (Figure 2)
gave the partition ratios, which are listed in Table 1.
Preliminary investigations with the p-nitrophenylthio
equivalent of 1 showed the marked instability of -Ala--
Gly(SC6H4-p-NO2) in neutral solutions. Therefore, we ex-
cluded the possibility of introducing additional electron-
withdrawing groups on the aromatic ring of 1 as a means
of facilitating the decomposition of the substituted 2-
phenylthioglycine intermediate (Scheme 1). The substitu-
tion of the aromatic ring by electron-donating groups could,
on the other hand, promote fluoride elimination from the
substituted (difluoromethyl)thiophenol. Such an approach
has been demonstrated by β-glucosidase inhibition, in
which the addition of a methoxy group to the aromatic ring
of (difluoromethyl)aryl-β--glucosides resulted in an in-
crease of inhibitor efficiency.[20]
The aforementioned instability of -Ala--Gly(SC6H4-p-
NO2) precluded its utilization as a substrate for VanX.
Thus, in our evaluation of enzyme activity, we used -Ala-
-Gly(SPh)-OH as an alternate VanX substrate. Detection
of the thiophenol released by enzyme action was monitored
continuously by Ellman’s reagent derivatization (Scheme 3).
Figure 2. Partition-ratio determination: inactivation of VanX was
performed as described in the Expt. Sect. with variable amounts of
14a; the plot represents the residual activity vs. the ratio [I]/[E0]
The incubation of VanX with compounds 13a or 13b did
not affect the enzyme activity. The catalytic efficiency of
enzyme toward either of these compounds as a substrate
was examined by TLC analysis. Even incubation at 37 °C
of 2 m 13a or 13b for 1 h with an excess of VanX (concen-
tration 10 times higher than usual conditions) did not allow
the detection of -Ala. These two compounds are, there-
fore, not recognized by the enzyme as substrates. The in-
hibition parameters are reported in Table 1.
Analysis of the inhibition data revealed that only com-
pounds 12a and 14a affected VanX activity. The introduc-
tion of two additional groups in the ortho positions of the
aromatic ring generates steric constraints in the small-size
enzyme active site,[21] resulting in loss of recognition of 13a
by VanX. The change of the difluoromethyl position from
para to ortho (Table 1) resulted in a twofold increase in bind-
ing, but a diminished inactivation rate constant without no-
ticeable changes in the partition ratio. The overall efficiency
increased 1.4-fold. On the other hand, addition of a me-
thoxy group in the ortho position (12a) led to a substantial
(10-fold) decrease of inactivation efficiency. The effects ob-
Discussion
In the search for VanX inhibitors, we tried to develop an served in this case are opposite to those reported for glucos-
approach using the enzyme recognition properties for the idase inhibition. This difference may be ascribed to the dis-
alternative substrate -Ala--Phe,[4] while also considering tinctions between our inhibitor activation mode and the
the possibility of generating highly reactive 4-thioquinone one proposed by Danzin’s group for β-glucosidase. In the
fluoromethide through the enzyme-catalyzed peptide cleav- case of β-glucosidase, enzyme-catalyzed hydrolysis of the
age (Scheme 1). Because of the less-than-satisfactory per- glucosidic linkage directly generates (difluoromethyl)-
formance of the initial mechanism-based inhibitor 1, im- phenol, which is assumed to rapidly afford reactive fluorin-
proving its efficiency by increasing the reactivity of the in- ated quinone methide. The substitutions are therefore be-
Scheme 3. Rationale for VanX assay
3576
Eur. J. Org. Chem. 2002, 3573Ϫ3579

Mechanism-Based Inhibitors of D-Alanyl-D-alanine Dipeptidase VanX
FULL PAPER
1.5 H), 3.97 (s, 1.5 H), 4.19 (m, 1 H), 5.03 (d, J ϭ 7.0 Hz, 0.5 H),
5.06 (d, J ϭ 7.0 Hz, 1 H), 6.19 (dd, J ϭ 9.0, J ϭ 2.0 Hz, 1 H), 7.18
(d, J ϭ 9.0 Hz, 1 H), 7.33 (s, 1 H), 7.35 (dd, J ϭ 7.8, J ϭ 1.2 Hz,
1 H), 7.54 (d, J ϭ 7.8 Hz, 1 H), 9.86 (s, 0.5 H), 9.90 (s, 0.5 H) ppm.
13C NMR (CDCl3,75 MHz): δ ϭ 18.00, 28.30, 28.70, 55.60, 57.84,
59.41, 83.26, 123.80, 126.96, 134.59, 136.28, 152.37, 154.88, 165.94,
169.21, 191.59 ppm. MS (ESI): m/z ϭ 469.1 [M ϩ H]ϩ, 491.1 [M
ϩ Na]ϩ.
lieved to facilitate the formation of thioquinone methide or
increase its reactivity towards enzyme nucleophiles. In the
VanX case, the formation of p-(difluoromethyl)thiophenol
requires the initial decomposition of α-(arylthio)glycine de-
rived from the peptide bond cleavage. Although the latter
was reported to be unstable, its lifetime obviously affects
the rate of formation of p-(difluoromethyl)thiophenol,
which is the actual inactivating species. However, in the ab-
sence of data concerning the rate of spontaneous break-
down of the intermediates and without comparing enzyme
active sites, the attribution of the poor catalytic efficiency
of 12a to a slow decomposition of α-(arylthio)glycine can
only be considered as speculation.
Boc-D-Ala-Gly(SC6H4-o-OCH3-o-CHO)-OtBu (7): A solution of 4
(2.9 g, 17.5 mmol) in DMF was added to a solution of 2 (3.5 g,
9.71 mmol) in DMF (15 mL) and Et3N (2.05 mL, 14.6 mmol). The
reaction was carried out as described for 6 giving 7 as a viscous
liquid (2.5 g, 55%). 1H NMR (CDCl3, 300 MHz): δ ϭ 1.27 (d, J ϭ
7.1 Hz, 3 H), 1.43 (s, 9 H), 1.42 (s, 9 H), 3.99 (s, 3 H), 4.17 (m, 1
H), 4.89 (m, 1 H), 5.91 (d, J ϭ 9.0 Hz, 0.5 H), 5.94 (d, J ϭ 9.0 Hz,
0.5 H), 6.81 (d, J ϭ 9.0 Hz, 0.5 H), 6.85 (d, J ϭ 9.0 Hz, 0.5 H),
7.15 (d, J ϭ 8.0 Hz, 1 H), 7.45 (t, J ϭ 7.3 Hz, 1 H), 7.54 (d, J ϭ
8.0 Hz, 1 H), 10.69 (s, 1 H) ppm. 13C NMR (CDCl3, 75 MHz): δ ϭ
17.88, 27.44, 28.10, 55.14, 56.26, 60.22, 83.10, 115.68, 120.10,
123.10, 130.67, 139.33, 155.19, 160.71, 166.34, 171.74, 192.36 ppm.
MS (ESI): m/z ϭ 491 [M ϩ Na]ϩ, 507 [M ϩ K]ϩ.
Conclusion
We have studied the effects of aromatic substitution on
the in vitro inhibition properties of the first mechanism-
based inhibitor of dipeptidase VanX -Ala--Gly(SC6H4-
p-CHF2). From this brief study it appears that both the
narrowness and the selectivity of the VanX active site is a
major obstacle to the design of efficient mechanism-based
dipeptide-like inhibitors. Nevertheless, this study is a good
example of the complexity of inhibitor design, which will
continue to challenge and inspire the ingenuity of bioor-
ganic chemists.
Boc-D-Ala-Gly(SC6H4-o-CHO)-OtBu (8): Compound 5 (844 mg,
6.10 mmol) and Et3N (780 µL, 5.55 mmol) were added at room
temperature to a solution of 2 (2 g, 5.55 mmol) in chloroform
(26 mL). After 18 h, the reaction mixture was dried and separated
by flash chromatography on silica gel (EtOAc/heptane, 3:7) to give
1
8 as a white solid (1.61 g, 78%). H NMR (CDCl3, 300 MHz): δ ϭ
1.29 (d, J ϭ 7.1 Hz, 1.5 H), 1.36 (d, J ϭ 7.1 Hz, 1.5 H), 1.38 (s, 9
H), 1.42 (s, 9 H), 4.18 (m, 1 H), 4.98 (d, J ϭ 7.2 Hz, 0.5 H), 5.06
(d, J ϭ 7.2 Hz, 0.5 H), 5.65 (d, J ϭ 4.4 Hz, 0.5 H), 5.62 (d, J ϭ
4.4 Hz, 0.5 H), 7.22 (br. d, J ϭ 7.2 Hz, 1 H), 7.48 (t, J ϭ 7.4 Hz,
1 H), 7.58 (dt, J ϭ 7.5, J ϭ 1.2 Hz, 1 H), 7.74 (d, J ϭ 7.2 Hz, 1
H), 7.93 (dd, J ϭ 6.4, J ϭ 1.3 Hz, 1 H), 10.54 (s, 0.5 H), 10.57 (s,
0.5 H) ppm. 13C NMR (CDCl3, 75 MHz): δ ϭ 17.60, 17.77, 27.74,
28.23, 57.66, 57.75, 83.89, 129.11, 129.23, 129.43, 134.11, 136.77,
166.63, 166.75, 171.80, 191.99, 192.34 ppm. MS (ESI): m/z ϭ 461.1
[M ϩ Na]ϩ, 477.1 [M ϩ K]ϩ, 899.5 [2 M ϩ Na]ϩ, 915.3 [2 M
ϩ K]ϩ.
Experimental Section
General Remarks: Where appropriate, reactions were carried out in
dried glassware under argon. All solvents were dried by standard
methods before use. Dichloromethane was distilled from calcium
hydride under argon. Triethylamine was kept dry over sodium un-
der argon; DMF and TMEDA over molecular sieves. 60 SDS silica
gel was employed for flash chromatography. 1H and 13C NMR
spectra were recorded with Bruker AC 250 or 300 spectrometers.
Low-resolution MS (ESI) data were obtained with a Navigator
(Thermoquest) instrument and high-resolution mass spectra were
recorded by the mass spectroscopic service at the Institut de Chimie
Boc-D-Ala-Gly(SC6H4-o-OCH3-p-CHF2)-OtBu (9):
DEOXO-
FLUOR (286 µL, 1.55 mmol) was slowly added under argon to a
solution of 6 (427 mg, 0.91 mmol) in anhydrous CH2Cl2 (5 mL) at
0 °C. The reaction mixture was stirred at room temperature for
12 h, then absolute ethanol was added (1 mL). After 30 min, the
solvent was evaporated and the crude residue separated by silica
gel chromatography (EtOAc/heptane, 4:6) to give 9 as a white solid
(210 mg, 47%). 1H NMR (CDCl3, 300 MHz): δ ϭ 1.35 (d, J ϭ
3.3 Hz, 1.5 H), 1.37 (d, J ϭ 3.3 Hz, 1.5 H), 1.44 (s, 9 H), 1.48 (d,
J ϭ 0.6 Hz, 9 H), 3.98 (s, 3 H), 4.20 (m, 1 H), 5.04 (m, 1 H), 5.89
(dd, J ϭ 9.0, J ϭ 4.0 Hz, 1 H), 6.23 (dd, J ϭ 9.0, J ϭ 4.0 Hz, 1
H), 6.61 (t, J ϭ 56.3 Hz, 1 H), 7.01 (s, 1 H), 7.02 (d, J ϭ 8.0 Hz,
1 H), 7.53 (d, J ϭ 8.0 Hz, 1 H) ppm. 13C NMR (CDCl3, 75 MHz):
δ ϭ 18.18, 27.65, 28.23, 54.92, 56.07, 66.35, 83.11, 107.72, 112.88
(t, J ϭ 236 Hz, CHF2), 114.04, 118.00, 135.85, 159.71, 166.87,
171.64 ppm. MS (ESI): m/z ϭ 491.1 [M ϩ H]ϩ, 513.0 [M ϩ Na]ϩ,
1003.4 [2 M ϩ Na]ϩ.
´
Moleculaire (Orsay). HPLC analyses were performed with a Waters
600 E instrument, equipped with a Waters 490 E UV detector, on
˚
analytical Delta Pack C18 Milligen (6 µm, pore size 60 A, 0.8 ϫ
10 cm) and preparative Waters RCM (2.5 ϫ 10 cm) columns. Ana-
lytical runs were performed under 30 min gradient conditions from
water to acetonitrile (both solvents contained 0.01% TFA). α-Ace-
toxyglycine (2) was synthesized as described earlier.[7] 4-Mercapto-
3-methoxybenzaldehyde (3),[9,11,12] 2-mercapto-3-methoxybenzal-
dehyde (4),[10Ϫ12] and 2-thiosalicylaldehyde (5)[13,14] were prepared
according to literature methods.
Boc-D-Ala-Gly(SC6H4-o-OCH3-p-CHO)-OtBu (6): Et3N (280 µL,
1.99 mmol) and a solution of 3 (481 mg, 2.88 mmol) in DMF were
added to 2 (580 mg, 1.6 mmol) in DMF (13 mL). The initial red
solution became orange after 1 h. Stirring was continued for 16 h
at room temperature. DMF was then evaporated under vacuum
and the residue was separated by silica gel chromatography
Boc-D-Ala-Gly(SC6H4-o-OCH3-o-CHF2)-OtBu (10): DEOXO-
FLUOR (944 µL, 5.12 mmol) was slowly added under argon to a
solution of 7 (1.6 g, 3.42 mmol) in anhydrous CH2Cl2 (50 mL) at 0
°C. The solution was stirred for 18 h, then absolute ethanol was
(EtOAc/heptane, 4:6) to give 6 as a colorless viscous liquid (337 mg,
1
45%). H NMR (CDCl3, 300 MHz): δ ϭ 1.32 (d, J ϭ 7.2 Hz, 1.5 added (1 mL) as described for 9. The reaction solvents were evapor-
H), 1.34 (d, J ϭ 7.2 Hz, 1.5 H), 1.40 (s, 9 H), 1.42 (s, 9 H), 3.91 (s,
Eur. J. Org. Chem. 2002, 3573Ϫ3579
ated and the residue separated by silica gel chromatography
3577

L. Yaouancq, M. Anissimova, M.-A. Badet-Denisot, B. Badet
FULL PAPER
(EtOAc/heptane, 3:7). Compound 10 was obtained as a white solid 160.47, 168.70, 168.82 ppm. MS (ESI): m/z ϭ 334.9 [M ϩ H]ϩ,
(1.61 g, 70%). 1H NMR (CDCl3, 300 MHz): δ ϭ 1.41 (d, J ϭ 356.9 [M ϩ Na]ϩ. HRMS (ESI): calcd. for C13H16F2N2NaO4S [M
7.1 Hz, 1.5 H), 1.44 (d, J ϭ 7.1 Hz, 1.5 H), 1.50 (s, 9 H), 1.51 (d,
J ϭ 8.0 Hz, 4.5 H), 1.54 (d, J ϭ 8.0 Hz, 4.5 H), 4.04 (s, 3 H), 4.20
(m, 1 H), 4.94 (br. d, J ϭ 7.4 Hz, 0.5 H), 5.04 (br. d, J ϭ 7.4 Hz,
ϩ Na]ϩ m/z ϭ 357.06965, found 357.07022. 2nd Eluted Diastereo-
isomer 13b: 1H NMR ([D6]DMSO, 300 MHz): δ ϭ 0.68 (d, J ϭ
6.5 Hz, 3 H), 3.59 (q, J ϭ 6.5 Hz, 1 H), 3.87 (s, 3 H), 5.89 (d, J ϭ
0.5 H), 6.02 (dd, J ϭ 9.0, J ϭ 4.0 Hz, 1 H), 7.08 (d, J ϭ 8.2 Hz, 1 9.7 Hz, 1 H), 7.08 (d, J ϭ 8.5 Hz, 1 H), 7.12 (d, J ϭ 8.0 Hz, 1 H),
H), 7.20 (d, J ϭ 9.0 Hz, 1 H), 7.23 (t, J ϭ 55.2 Hz, 1 H), 7.36 (d, 7.34 (t, J ϭ 55.6 Hz, 1 H), 7.41 (t, J ϭ 8.0 Hz, 1 H), 8.10 (br. s, 1
J ϭ 8.1 Hz, 1 H), 7.51 (t, J ϭ 8.1 Hz, 1 H) ppm. 13C NMR (CDCl3, H), 8.61 (d, J ϭ 9.7 Hz, 1 H) ppm. 13C NMR ([D6]DMSO,
75 MHz): δ ϭ 17.97, 27.48, 28.20, 54.83, 56.19, 66.50, 66.64, 83.16,
109.45, 112.60 (t, J ϭ 236 Hz, CHF2), 112.83, 117.88, 130.94,
160.43, 166.52 ppm. MS (ESI): m/z ϭ 513 [M ϩ Na]ϩ.
75 MHz): δ ϭ 17.11, 47.73, 55.87, 57.81,113.00, 113.29 (t, J ϭ
240 Hz, CHF2), 129.93, 138.53, 160.54, 168.66, 168.37 ppm. MS
(ESI): m/z ϭ 334.9 [M ϩ H]ϩ, 356.9 [M ϩ Na]ϩ, 372.8 [M Ϫ H
ϩ 2 Na]ϩ. HRMS (ESI): calcd. for C13H16F2N2NaO4S [M ϩ Na]ϩ
m/z ϭ 357.06965, found 357.07015.
Boc-D-Ala-Gly(SPh-o-CHF2)-OtBu
(11):
DAST
(1.54 mL,
11.62 mmol) was added under argon to a solution of 8 (3 g,
6.84 mmol) in anhydrous CH2Cl2 (100 mL) at 0 °C. After 16 h, the
reaction mixture was washed with water until neutral (7 ϫ 50 mL),
then dried with MgSO4 and the solvents were evaporated. Separa-
tion by silica gel chromatography (EtOAc/heptane, 3:7) gave 11 as
a white viscous solid (1.6 g, 51%). 1H NMR (CDCl3, 300 MHz):
δ ϭ 1.25 (d, J ϭ 5.5 Hz, 1.5 H), 1.28 (d, J ϭ 5.5 Hz, 1.5 H), 1.37
(s, 9 H), 1.43 (d, J ϭ 2.4 Hz, 9 H), 4.16 (m, 1 H), 4.89 (d, J ϭ
7.4 Hz, 0.5 H), 4.99 (d, J ϭ 7.4 Hz, 0.5 H), 5.54 (d, J ϭ 8.4, J ϭ
3.1 Hz, 1 H), 5.56 (d, J ϭ 8.4, J ϭ 3.1 Hz, 1 H), 7.17 (dt, J ϭ 55.1,
J ϭ 1.9 Hz, 1 H), 7.17 (br. s, 1 H), 7.45 (m, 2 H), 7.70 (dt, J ϭ 7.0,
J ϭ 1.4 Hz, 2 H) ppm. 13C NMR (CDCl3,75 MHz): δ ϭ 17.75,
27.68, 28.23, 57.83, 57.92, 83.80, 112.77 (t, J ϭ 236 Hz, CHF2),
126.11, 129.75, 131.02, 136.74, 136.94, 166.80, 166.92, 171.72 ppm.
MS (ESI): m/z ϭ 483.1 [M ϩ Na]ϩ, 499.1 [M ϩ K]ϩ, 943.3 [2 M
ϩ Na]ϩ.
D
-Ala-Gly(SC6H4-o-CHF2)-OH (14a, 14b): TFA (28 mL) was ad-
ded to a solution of 11 (1.6 g, 3.47 mmol) in acetic acid (7 mL).
After 4 h, the solvents were evaporated and the diastereoisomers
purified by preparative HPLC, then lyophilized to give 14a
(486 mg, 46%) and 14b (275 mg, 26%) as white powders. 1st Eluted
1
Diastereoisomer 14a: H NMR ([D6]DMSO, 300 MHz): δ ϭ 1.28
(d, J ϭ 6.8 Hz, 3 H), 3.94 (q, J ϭ 6.8 Hz, 1 H), 5.17 (d, J ϭ 4.0 Hz,
1 H), 7.32 (t, J ϭ 55.3 Hz, 1 H), 7.49 (t, J ϭ 6.5 Hz, 1 H), 7.50
(dt, J ϭ 6.3, J ϭ 2.3 Hz, 1 H), 7.62 (dd, J ϭ 6.3, J ϭ 2.3 Hz, 1
H), 7.80 (d, J ϭ 6.5 Hz, 1 H), 8.25 (br. s, 2 H), 8.90 (d, J ϭ 4.0 Hz,
1 H) ppm. 13C NMR ([D6]DMSO, 75 MHz): δ ϭ 18.05, 50.63,
59.76, 115.04 (t, J ϭ 234 Hz, CHF2), 117.53, 119.37, 128.15,
132.40, 133.35, 138.84, 163.82, 171.21, 171.42 ppm. MS (ESI): m/
z ϭ 304.9 [M ϩ H]ϩ, 326.8 [M ϩ Na]ϩ, 342.8 [M ϩ K]ϩ. HRMS
(ESI): calcd. for C12H15F2N2O3S [M ϩ H]ϩ m/z ϭ 305.0771, found
305.07686. 2nd Eluted Diastereoisomer 14b: 1H NMR (D2O,
300 MHz): δ ϭ 1.15 (d, J ϭ 7.1 Hz, 3 H), 3.88 (q, J ϭ 7.1 Hz, 1
H), 5.75 (s, 1 H), 7.10 (t, J ϭ 55.0 Hz, 1 H), 7.47 (m, 2 H), 7.53
(d, J ϭ 7.3 Hz, 1 H), 7.62 (d, J ϭ 7.2 Hz, 1 H) ppm. 13C NMR
(D2O, 75 MHz): δ ϭ 18.00, 50.70, 59.19, 115.10 (t, J ϭ 235 Hz,
CHF2), 128.30, 132.49, 133.44, 138.99, 157.33, 171.57 ppm. MS
(ESI): m/z ϭ 304.9 [M ϩ H]ϩ, 326.9 [M ϩ Na]ϩ, 342.9 [M ϩ K]ϩ.
D-Ala-Gly(SC6H4-o-OCH3-p-CHF2)-OH
(12a,
12b):
TFA
(10.7 mL) was added to a stirred solution of 9 (800 mg, 1.63 mmol)
in acetic acid (2.7 mL). After 6 h, the solvents were evaporated.
The crude residue was redissolved in acetic acid (8 mL) and filtered.
The two diastereoisomers were separated by preparative HPLC
(water/acetonitrile, 95:5), then lyophilized to give 12a (239 mg,
44%) and 12b (207 mg, 38%) as white powders. 1st Eluted Diaster-
1
eoisomer 12a: H NMR ([D6]DMSO, 250 MHz): δ ϭ 1.29 (d, J ϭ
VanX Preparation: The expression vector pIADL14 (maltose-bind-
ing-proteinϪVanX [MBP-VanX]) was a gift from Prof. C. T. Walsh
(Harvard Medical School, Boston, USA). VanX was expressed in
BL21(DE3) Escherichia coli cells and purified as described earlier.[7]
6.5 Hz, 3 H), 3.78 (q, J ϭ 6.5 Hz, 1 H), 3.83 (s, 3 H), 5.42 (s, 1 H),
6.96 (t, J ϭ 55.8 Hz, 1 H), 7.06 (d, J ϭ 7.4 Hz, 1 H), 7.09 (s, 1 H),
7.69 (d, J ϭ 7.4 Hz, 1 H), 8.83 (br. s, 1 H) ppm. MS (ESI): m/z ϭ
334.9 [M ϩ H]ϩ, 356.9 [M ϩ Na]ϩ, 372.7 [M ϩ K]ϩ. 2nd Eluted
1
Inhibition Assays: To ensure the inhibitor’s stability, all assays re-
garding VanX inhibition were carried out at pH ϭ 7. VanX
(43 pmol, 1 µg) was incubated in 100 µL of 50 m HEPES (pH ϭ
7) in the presence of various concentrations of inhibitor and 2 m
-Ala--Ala in the case of 14a at 37 °C. At different time intervals,
3-µL aliquots were withdrawn and diluted 333-fold in 1-mL cu-
vettes containing 50 m HEPES, 2.5 m Ellman’s reagent and
5 m of -Ala--Gly(SPh)-OH (15) as a substrate at 37 °C.[18]
After the initial burst due to the presence of 4-thioquinone fluoro-
methide in the incubation mixture, the slope of thionitrobenzoate
production vs. time, reflecting the VanX residual activity, was mon-
itored continuously at 412 nm with a CARY 100 spectrophoto-
meter. The initial rates were calculated for the first 300 s, and ab-
sorbance changes were converted into concentration changes using
the molar extinction coefficient of liberated nitrothiobenzoic acid
[ε(412 nm) ϭ 13600 Ϫ1 cmϪ1].
Diastereoisomer 12b: H NMR ([D6]DMSO, 250 MHz): δ ϭ 0.97
(d, J ϭ 7.0 Hz, 3 H), 3.80 (q, J ϭ 7.0 Hz, 1 H), 3.84 (s, 3 H), 5.61
(d, J ϭ 8.0 Hz, 1 H), 6.95 (t, J ϭ 55.8 Hz, 1 H), 7.02 (d, J ϭ
7.8 Hz, 1 H), 7.07 (s, 1 H), 7.52 (d, J ϭ 7.8 Hz, 1 H), 8.20 (br. s, 2
H), 8.69 (d, J ϭ 8.7 Hz, 1 H) ppm. 13C NMR ([D6]DMSO,
75 MHz): δ ϭ 17.40, 47.96, 55.76, 56.94, 107.43, 114.74 (t, J ϭ
238 Hz, CHF2), 117.86, 130.69, 157.03, 168.06, 168.93 ppm. MS
(ESI): m/z ϭ 334.9 [M ϩ H]ϩ, 356.9 [M ϩ Na]ϩ. HRMS (ESI):
calcd. for C13H16F2N2NaO4S [M ϩ Na]ϩ m/z ϭ 357.06965,
found 357.07021.
D-Ala-Gly(SC6H4-o-OCH3-o-CHF2)-OH
(13a,
13b):
TFA
(28.4 mL) was added to a stirred solution of 10 (2.12 g, 4.32 mmol)
in acetic acid (7.1 mL). After 5 h, the solvents were evaporated and
the residue redissolved in acetic acid, then the diastereoisomers
were separated by preparative HPLC to give 13a (620 mg, 43%)
and 13b (505 mg, 35%) as white powders after lyophilization. 1st
Eluted Diastereoisomer 13a: 1H NMR ([D6]DMSO, 300 MHz): δ ϭ
Partition Ratio Determination: VanX (43 pmol) was incubated in
1.12 (d, J ϭ 6.9 Hz, 3 H), 3.57 (q, J ϭ 6.9 Hz, 1 H), 3.86 (s, 3 H), 100 µL of 50 m HEPES (pH ϭ 7) at 37 °C in the presence of
5.50 (br. s, 1 H), 7.13 (d J ϭ 7.5 Hz, 1 H), 7.17 (d, J ϭ 5.7 Hz, 1 various concentrations of inhibitors (0Ϫ3 m) and 2 m -Ala--
H), 7.37 (t, J ϭ 55.6 Hz, 1 H), 7.44 (t, J ϭ 8.0 Hz, 1 H), 8.54 (br. Ala in the case of 14a. After 10 min, a 3-µL aliquot was removed
s, 1 H) ppm. 13C NMR ([D6]DMSO, 62.5 MHz): δ ϭ 17.32, 48.08, and tested for activity as described above. The partition ratio was
56.02, 68.13, 109.56, 113.33 (t, J ϭ 236 Hz, CHF2), 117.56, 130.44, obtained as described earlier.[7]
3578
Eur. J. Org. Chem. 2002, 3573Ϫ3579

Mechanism-Based Inhibitors of D-Alanyl-D-alanine Dipeptidase VanX
FULL PAPER
[8]
W. D. Kingsbury, J. Bohem, Int. J. Peptide Protein Res. 1986,
27, 659Ϫ665.
S. Wong, S. Sasso, H. Jones, J. J. Kaminski, J. Med. Chem.
1984, 27, 20Ϫ27.
L. K. A. Rahman, R. M. Scrowston, J. Chem. Soc., Perkin
Trans. 1 1983, 12, 2973Ϫ2978.
H. Kwart, E. R. Evans, J. Org. Chem. 1966, 31, 410Ϫ413.
M. S. Newman, H. A. Karnes, J. Org. Chem. 1966, 31,
3980Ϫ3984.
W. J. Still, R. Natividad-Preyra, F. Dean Toste, Can. J. Chem.
1999, 77, 113Ϫ121.
D. A. Nation, M. R. Taylor, K. P. Wainwright, J. Chem. Soc.,
Dalton Trans. 1966, 3001Ϫ3009.
G. Apitz, M. Jäger, S. Jaroch, M. Kratzel, L. Schäffeler, W.
Steglich, Tetrahedron. 1993, 49, 8223Ϫ8232.
R. Gonzalez-Muniz, F. Cornille, F. Bergeron, D. Ficheux, J.
Pothier, C. Durieux, B. P. Roques, Int. J. Peptide Res. 1991,
37, 331Ϫ340.
J. L. Kraus, G. Attardo, Eur. J. Med. Chem. 1992, 27, 19Ϫ26.
M. Anissimova, L. Yaouancq, M.-A. Badet-Denisot, B. Badet,
Anal. Biochem., submitted.
Acknowledgments
[9]
We thank S. Gueddoudj and F. Noor for their assistance in the
synthesis of compounds 12 and 13 and F. Perez for performing the
[10]
´
HRMS analysis. Helpful comments from Prof. H.-W. Liu, Dr.
P. Durand and one of the referees are gratefully acknowledged.
This work was supported by a grant from Centre National de la
Recherche Scientifique, France (Programme Physique et Chimie du
Vivant) and a CNRS fellowship to M. A.
[11]
[12]
[13]
[14]
[15]
[16]
[1]
J. C. J. Barna, D. H. Williams, Annu. Rev. Microbiol. 1984,
38, 339Ϫ357.
[2]
M. Arthur, P. Couvalin, Antimicrob. Agents Chemother. 1993,
37, 1563Ϫ1571.
P. E. Reynolds, F. Depardieu, S. Dutka-Malen, M. Arthur, P.
[3]
[17]
[18]
Courvalin, Mol. Microbiol. 1994, 13, 1065Ϫ1070.
[4]
Z. Wu, G. D. Wright, C. T. Walsh, Biochemistry 1995, 34,
[19]
[20]
2455Ϫ2463.
G. G. Knight, S. G. Waley, Biochem. J. 1985, 225, 435Ϫ439.
S. Halazy, A. E. Berger, A. Ehrhard, C. Danzin, Bioorg. Chem.
1990, 18, 330Ϫ344.
D. E. Bussiere, S. D. Pratt, L. Katz, J. M. Severin, T. Holzman,
C. H. Park, Mol. Cell 1998, 2, 75Ϫ84.
[5]
Z. Wu, C. T. Walsh, Proc. Natl. Acad. Sci. USA 1995, 92,
11603Ϫ11607.
K.-W. Yang, J. J. Brandt, L. L. Chatwood, M. W. Crowder,
[6]
[21]
Bioorg. Med. Chem. Lett. 2000, 10, 1085Ϫ1087.
[7]
´
R. Araoz, E. Anhalt, L. Rene, M.-A. Badet-Denisot, P. Courv-
Received July 4, 2002
alin, B. Badet, Biochemistry 2000, 39, 15971Ϫ15979.
[O02364]
Eur. J. Org. Chem. 2002, 3573Ϫ3579
3579