Cell permeable vanX inhibitors as vancomycin re-sensitizing agents

Article abstract of DOI:10.1016/j.bmcl.2014.03.097

VanX is an induced zinc metallo d-Ala-d-Ala dipeptidase involved in the viable remodeling of bacterial cell wall that is essential for the development of VREF. Here we report two cyclic thiohydroxamic acid-based peptide analogs that were designed, synthes

Full text of DOI:10.1016/j.bmcl.2014.03.097

Bioorganic & Medicinal Chemistry Letters 24 (2014) 2535–2538
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Cell permeable vanX inhibitors as vancomycin re-sensitizing agents
Ramaiah Muthyala a,b, , Namrata Rastogi a,b, Woo Shik Shin c, Marnie L. Peterson b, Yuk Yin Sham c,
⇑
⇑
a Center for Orphan Drug Research, University of Minnesota, 717 Delaware St SE, Minneapolis, MN 55455, United States
b Department of Experimental & Clinical Pharmacology, University of Minnesota, 717 Delaware St SE, Minneapolis, MN 55455, United States
c Center for Drug Design, Academic Health Center, University of Minnesota, 516 Delaware St SE, Minneapolis, MN 55455, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
VanX is an induced zinc metallo
D
-Ala-
D-Ala dipeptidase involved in the viable remodeling of bacterial
Received 13 March 2014
Revised 27 March 2014
Accepted 28 March 2014
Available online 8 April 2014
cell wall that is essential for the development of VREF. Here we report two cyclic thiohydroxamic
acid-based peptide analogs that were designed, synthesized and investigated as vancomycin re-sensitiz-
ing agents. These compounds exhibit low micromolar inhibitory activity against vanX, with low cytotox-
icity and were shown to increase vancomycin sensitivity against VREF. The improved pharmacological
properties of these novel inhibitors over previous transition state mimics should provide an enhanced
platform for designing potent vanX inhibitors for overcoming vancomycin resistance.
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Vancomycin
Drug resistance
VanX
Pyrithione
Thiohydroxamic acid
Antibiotics
Vancomycin is the last line of antibiotic defense against serious
Gram-positive infections when all other therapeutic agents have
failed. Continual emergence of vancomycin resistant bacterial
pathogens such as vancomycin resistant Enterococcus faecium
(VREF) and Staphylococcus aureus (VRSA) poses a major health
threat to children, immune deficient and post-surgical patients in
hospitals.1–3 There is an urgent need to develop alternative therapy
for combating vancomycin resistant pathogens.
incorporation of
decrease in affinity of vancomycin for
This variation circumvents vancomycin inhibitory action against the
bacterial PG biosynthesis pathway and renders vancomycin inactive
as an antibacterial agent.8
D
-Ala-
D
-Lac into the PG layer with a thousand-fold
-Ala- -Lac over -Ala-
-Ala.7
D
D
D
D
Over the past two decades, vanX has been viewed as an attractive
drug target for re-establishing vancomycin sensitivity.9 VanX is a
zinc metallo D-Ala-D-Ala dipeptidase that catalyzes the hydrolytic
Peptidoglycan (PG) is an essential component of bacterial cell
wall and has been widely exploited in the development of effective
antibacterial agents.4 Vancomycin inhibits PG biosynthesis by
cleavage of the dipeptide amide bond.9,10 Transition state analogs,
tetrahedral intermediates, and other mechanism-based irreversible
inhibitors11 have been investigated to understand its catalytic
mechanism.10 Subsequently, several classes of phosphinates,12
phosphothioate,13 phosphonates,12 and phosphoramidates13 tran-
sition state mimics (TSM) with low micromolar inhibition biochem-
ical activities have been reported as vanX inhibitors (Fig. 1).
binding to the terminal amino acid residues,
D-alanyl-D-alanine
(D
-Ala- -Ala) of the NAM/NAG-peptide subunits. The formation of
D
a tightly bound complex prevents trans-peptidation and cross-link-
ages among glycan strands, leading to the collapse of the compro-
mised bacterial cell wall and ultimately cell death.5
High-level vancomycin resistance is conferred by the acquisition
and over expression of a five cassette gene (vanR, vanS, vanH, vanA
and vanX)6 that results in the accumulation of
(D-Ala-D
D
-alanyl-
-Lac) as an alternate peptidoglycan precursor for bacterial
cell wall remodeling. Removal of the endogenous -Ala- -Ala by
D-lactate
D
D
the vanX gene product, vanX enzyme allows predominant
⇑
Corresponding authors. Tel.: +1 612 624 7120 (R.M.), tel.: +1 612 625 8255
(Y.Y.S.).
E-mail addresses: muthy003@umn.edu (R. Muthyala), shamx002@umn.edu
(Y.Y. Sham).
Figure 1. Transition state vanX inhibitors.
http://dx.doi.org/10.1016/j.bmcl.2014.03.097
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

2536
R. Muthyala et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2535–2538
Unfortunately, these inhibitors exhibit no cellular penetration prop-
erties required for cell based inhibition activity, rendering them
unattractive for further therapeutic development. In this study,
we explore the design of novel cyclic thiohydroxamic acid peptide
analogues that, for the first time, exhibit both low micromolar bio-
chemical inhibitory activity against vanX and vancomycin re-sensi-
tizing activity against VREF in cell-based assay.
VanX is a zinc dependent dipeptidase that can only hydrolyze
D,D-dipeptide substrates, and does not exhibit any activity against
tripeptides, terminally capped dipeptides and specific stereoiso-
meric dipeptides.10 Many designs of metallo enzyme inhibitors
incorporate a hydroxamic acid as the bidentating group of choice,
but we purposely avoided this moiety because of its potential lia-
bilities in their selectivity toward other divalent metals that can
lead to off-target cytotoxicity. While simple acyclic hydroxamic
acids are not suitable as drug candidates, cyclic hydroxamic acids
or their isosteres such as cyclic
a-hydroxy ketones and non-toxic
thioketones constitute viable alternatives.14 Heterocyclic hydroxa-
mic acid-containing compounds have been studied extensively as
antiviral and antibacterial agents with favorable pharmacological
properties.15 Hydroxy pyridones such as deferiprone have been
used clinically to treat b-thalassaemia.
Figure 2. Design of ZBG-AA. Expected mode of chelation by ionized 4 and 5 with
vanX Zn+2 cofactor.
Based on established low molecular weight of previous TSM
vanX inhibitors, our earlier designs of metallo binding inhibi-
tors16,17 and docking studies, we hypothesized that an effective
vanX inhibitor should consist only of a zinc binding group (ZBG)
and a single D-amino acid group (AA) in order to bind within the
catalytic site of vanX (Fig. 2). Our design incorporates pyrithione,
a heterocyclic thiohydroxamic acid moiety as the zinc specific
binding group with a single D-amino acid. Pyrithione is a monoan-
ionic zinc specific bidentating chelator that forms a five-membered
complex with Zn2+ via its oxygen and sulfur atomic centers. Zinc
pyrithione (ZPT) has been isolated from the Polyalthia nemoralis
roots used in Chinese herbal remedy18 and shown to possess anti-
malarial properties.19 Pyrithiones was first synthesized in 1950’s as
an analogue to aspergillic acid, a naturally occurring antibiotic.20
Today, pyrithiones (sodium, zinc and copper) have been well
established as active antidandruff21 and antifouling22 agents. More
recently, pyrithiones have been shown to exhibit both anticancer23
and antiviral activities.24,25 Furthermore, they possess desired non-
toxic14, ionophoric properties to overcome the cellular penetration
problems experienced by TSM compounds 1–3 (Fig. 1), making the
design ideal for the last line of antibiotic combination therapy.
Previous structural study has revealed details of the vanX active
site.26 Its substrate specificity can be attributed to the divalent zinc
cofactor, the overall compactness of the binding pocket, and the
pre-organized amino acid sidechains inside the active site for the
stabilizing the zwitterionic termini of the dipeptide substrates
(Fig. 3A). Docking of the stereoisomeric dipeptides into the active
site reveals zwitterionic D,D-dipeptides are structurally preferred
over stereoisomeric dipeptides (not shown). The negatively
Figure 3. Model of the vanX active site with (A) D-Ala-D-Ala substrate, (B) 4 and (C)
5. The translucent molecular surface colored by electrostatic potential surface
highlights the stereoselectivity and the diminutive size of the binding pocket. The
measured distances in angstroms are shown in yellow. Both S114 and S115 are
essential is stabilizing the amide oxygen and carboxylic C-terminus of compounds 4
and 5.
charged side chains of D123 and D142 and the phenol group of
Y21 provide the key stabilization interactions for the positively
charged N-terminus while the S114 hydroxyl group is essential
for stabilizing the negatively charged C-terminal carboxylate
group. The small binding pocket can accommodate only low

R. Muthyala et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2535–2538
2537
ClH.H2N
COOMe
30%H2O2
TFA, 800C
NaHS
COOH
Na2S.9H2O
H
N
COOMe
Br
N
Br
N
COOH
S
N
COOH
S
N
HATU
DIPEA
OH
7
OH
OH
O
6
8
9
NH2.HCl
COOMe
H O:MeOH LiOH.H2O
THF:
2
HATU
2:1:1
Ph
DIPEA
H
O:MeOH
THF:H2
:1:1
COOH
N
H
H
N
S
N
2
N
COOMe
COOH
S
N
N
OH
S
O
Ph
LiOH.H2O
OH
O
Ph
O
OH
5
4
10
Scheme 1. Synthesis of 4 and 5.
molecular weight inhibitors, similar to that of the
dipeptide substrate.
D
-Ala-
D
-Ala
ester or phenyl alanine methyl ester followed by hydrolysis with
lithium hydroxide in methanol–THF–water-mixture, gave 4 or 5
in 40%, and 62% yield respectively. Extended periods of hydrolysis
at room temperature gave similar yields but racemization
occurred.
To examine the design of our vanX inhibitors, docking studies of
4 and 5 were carried out using the Schrodinger’s modeling suite
package.27 Our study began with the model of the
D-Ala-D-Ala vanX
complex (Fig. 3A) based on the reported apo X-ray structure of
vanX (PDB: 1R44) and structural details of the transition state ana-
logue described by Park and co-workers.26 The Docking study has
been extensively described elsewhere.28,29 Briefly, energy minimi-
zation with an implicit generalized solvent model was used to
account for consistent structural relaxation of the final complex
model. Based on this model, docking of 4 and 5 were carried out
using Glide27 at Standard Precision (SP) with Zn2+ defined as the
H
N
COO-
+H3
N
KM = 830 µM
O
S
kcat = 76 s-1
11
required constraint. For comparison, docking of the
D-Ala-D-Ala
Figure 4. Reported kinetic parameters for D-alanyl-a-R-phenylthioglycine as the
reporting substrate for the biochemical assay (see Ref. 31).
was carried out to reproduce the expected mode of substrate bind-
ing (Fig. 3A). Subsequent docking of the design inhibitors further
revealed the preferred mode of binding for 4 and 5 involving the
expected ionized pyrithione within the expected distances neces-
sary for forming bidentating chelate with the zinc cofactor
(Fig. 3B and C). The potential mode of binding for both 4 and 5 is
observed to be slightly shifted toward the opening of the active site
A previously described coupled colorimetric assay using D-ala-
nyl-
mine the inhibition constant, Ki, of 4 and 5. The colorimetric
substrate needed for enzymatic biochemical assay, -phenylthio
a
-R-phenyl thioglycine substrate 1131 was employed to deter-
a
as compared to the D-Ala-D-Ala natural substrate. This is likely due
containing peptide (H-d-Ala-Psg-OH, 11) and the corresponding
diastereomer, H-d-Ala-dl-Psg-OH were prepared according to the
literature procedure.31 The expression and purification of vanX
was carried out according to earlier studies of Walsh and co-work-
ers.12 The initial rates were calculated for the first 300 s, and absor-
bance changes were converted to concentration change using the
molar extinction coefficient for the liberated thionitrobenzoic acid.
The kinetic parameters for the colorimetric substrate 11 (see Fig. 4)
were in accord with the earlier reported KM value of
0.83 0.08 mM,31 which was also comparable to the kinetic
to the placement of the ZBG in the terminal end of the molecule.
This mode of zinc chelation lead to the removal of the hydrogen
bond between Y109(O) and the amide hydrogen observed in the
dipeptide substrate and a shift in the hydrogen bonding network
involving S114 and S115 for stabilizing the amide oxygen and
the carboxylic C-terminus of compounds 4 and 5. This compensat-
ing effect could explain the observed G-scores for 4 and 5 (À7.6
and, À8.3, respectively) as compared to
D
-Ala-D-Ala (À8.8).
To test the validity of our design framework, we have synthe-
sized 4 and 5 as model compounds (see Scheme 1). 6-Thio
N-hydroxy pyridine 2-carboxylic acid, 8, was synthesized from
2-bromo pyridine 6-carboxylic acid, 6, according to literature pro-
cedures.30 Initially, a variety of usual peptide coupling agents, for
example EDCI or CDI, or benzotriazole in combination with various
of bases and conditions were attempted for coupling of N-benzyl-
parameters of the natural D-Ala-D-Ala substrate (KM = 0.11 0.01 -
mM). In contrast, the diastereomer of colorimetric substrate 11
was found to be enzymatically inactive—a fact that supports the
oxy 8 with amino acids such as D-alanine methyl esters or D-phenyl
H
N
H
N
COOH
COOH
alanine methyl esters. In all cases the coupling reaction gave unac-
ceptable low yields and debenzylation of coupled product also
became troublesome. Peptide coupling of 8 was best accomplished
(without protection of N-hydroxy group) with HATU
(1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo [4,5-b]pyrid-
inium-3-oxido hexafluoro phosphate) to give good yields. The cou-
N
S
N
S
O
OH
OH
O
Ph
Ki = 6.84 0.10 µM
Ki = 2.74 0.02 µM
4
5
pling of 8 (no need to protect N-OH group) with D-alanine methyl
Figure 5. Low lM inhibition constants for 4 and 5 against vanX.

2538
R. Muthyala et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2535–2538
stereo selectivity of vanX enzyme for certain dipeptides. After
reproducing these kinetic parameters, the Ki for 4 and 5 were
determined by least-squares fit to the Michaelis–Menten equation.
these data strongly support and demonstrate the feasibility toward
further explore the biochemical and biological properties of vanX
inhibitors 4 and 5 as lead candidates for rational modification to
further improve their efficacy.
The determined Ki’s were 6.84 and 2.74 lM for compounds 4 and 5,
respectively (Fig. 5). These are comparable to the low micromolar
inhibitory activities of the earlier reported transition state mimics
1–3 (Fig. 1).
Acknowledgments
Based on these findings, we determined the ability of com-
pounds 4 and 5 to re-sensitize VREF (E12-2) to vancomycin. A
checkerboard MIC analysis was performed to determine their addi-
tive or synergistic activities with vancomycin. Individually, 4 and 5
did not inhibit bacterial growth of VREF (not shown). The MIC of
We thank Prof. Christopher Walsh for providing the plasmid
used in the expression of vanX. This research was supported by
the faculty development grant from the University of Minnesota
Academic Health Center. We thank the University of Minnesota
Supercomputing Institute and the Center for Drug Design for
providing the necessary computational and instrumentation
resources. We thank Dr. Fred Schendel, Biotechnology Institute,
University of Minnesota for vanX protein preparation. We also
thank Jesse Meyer for performing the biochemical assays.
vancomycin against VREF was reduced from 256
lg/ml to
125 g/ml in the presence of 4 (212 g/ml) and 5 (72 l
l
l
g/ml)
(Table 1). This corroborate our design strategy that both com-
pounds were capable of penetrating the bacterial cell wall, inhibit
vanX, and, thereby, reduce vancomycin MICs. VanX inhibitor 4 was
3 times more potent than 5 in parallel to their observed biochem-
ical Ki values.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2014.03.
097.
Table 1
Checkerboard MIC analysis of E12-2 against vancomycin
Compound
MIC (
lg/ml)
MIC (lM)
References and notes
4
5
212
72
875
236
1. Hanna, H.; Umphrey, J.; Tarrand, J.; Mendoza, M.; Raad, I. Infect. Control Hosp.
Epidemiol. 2001, 22, 217.
2. Sader, H. S.; Jones, R. N. Diagn. Microbiol. Infect. Dis. 2009, 65, 158.
3. Courvalin, P. Clin. Infect. Dis. 2006, 42, S25.
4. Antibiotics: Actions, Origins, Resistance; Walsh, C., Ed.; ASM Press, 2003.
5. Barna, J. C. J.; Williams, D. H. Annu. Rev. Microbiol. 1984, 38, 339.
6. Arthur, M.; Molinas, C.; Depardieu, F.; Courvalin, P. J. Bacteriol. 1993, 175, 117.
7. Bugg, T. D. H.; Dutkamalen, S.; Arthur, M.; Courvalin, P.; Walsh, C. T.
Biochemistry-US 1991, 30, 2017.
8. Williams, D. H.; Bardsley, B. Angew. Chem., Int. Ed. 1999, 38, 1173.
9. Reynolds, P. E.; Depardieu, F.; Dutkamalen, S.; Arthur, M.; Courvalin, P. Mol.
Microbiol. 1994, 34, 1065.
10. Wu, Z.; Wright, G. D.; Walsh, C. T. Biochemistry-US 1995, 34, 2455.
11. Yaouancq, L.; Anissimova, M.; Badet-Denisot, M. A.; Badet, B. Eur. J. Org. Chem.
2002, 3573.
12. Wu, Z.; Walsh, C. T. Proc. Nat. Acad. Sci. U.S.A. 1995, 92, 11603.
13. Yang, K. W.; Brandt, J. J.; Chatwood, L. L.; Crowder, M. W. Bioorg. Med. Chem.
Lett. 2000, 10, 1085.
We next assessed the cytotoxicity of compounds 4 and 5 against
human embryonic kidney HEK 293 cell lines by MTT assay. Cultur-
ing and maintenance of HEK293 cells were performed as described
by ATCC. The CC50’s for both compounds 4 and 5 were 76
lM and
64 M after 72 h treatment, comparable to existing drugs such as
l
amoxicillin (Table 2). Interestingly, vancomycin was shown to be
non-toxic against the HEK293 cells. The observed cytotoxicity of
these compounds further support our design strategy of using a
more specific zinc binding group to limit inhibition of off-targets
proteins that can lead to unwanted cytotoxicity.
14. Puerta, D. T.; Griffin, M. O.; Lewis, J. A.; Romero-Perez, D.; Garcia, R.; Villarreal,
F. J.; Cohen, S. M. J. Biol. Inorg. Chem. 2006, 11, 131.
15. Agrawal, A.; de Oliveira, C. A. F.; Cheng, Y. H.; Jacobsen, J. A.; McCammon, J. A.;
Cohen, S. M. J. Med. Chem. 2009, 52, 1063.
16. Chen, L. Q.; Petrelli, R.; Gao, G. Y.; Wilson, D. J.; McLean, G. T.; Jayaram, H. N.;
Sham, Y. Y.; Pankiewicz, K. W. Bioorg. Med. Chem. 2010, 18, 5950.
17. Tang, J.; Maddali, K.; Metifiot, M.; Sham, Y. Y.; Vince, R.; Pommier, Y.; Wang, Z.
Q. J. Med. Chem. 2011, 54, 2282.
18. Yao, J. Z.; Liang, H. Q.; Liao, S. X. Acta Pharm. Sin. 1994, 29, 845.
19. Han, G. Y.; Xu, B. X.; Wang, X. P.; Liu, M. Z.; Xu, X. Y.; Meng, L. N.; Chen, Z. L.;
Zhu, D. Y. Acta Chim. Sinica 1981, 39, 433.
Table 2
MTT cytotoxicity study against human embryonic kidney HEK 293 cells
Compound
CC50 (lM)
Vancomycin
Amoxicillin
4
5
>100
24
76
64
20. Shaw, E.; Bernstein, J.; Losee, K.; Lott, W. A. J. Am. Chem. Soc. 1950, 72, 4362.
21. Schwartz, J.; Mizoguchi, H.; Bacon, R. J. Am. Acad. Dermatol. 2013, 68, Ab46.
22. Marcheselli, M.; Azzoni, P.; Mauri, M. Aquat. Toxicol. 2011, 102, 39.
23. Tailler, M.; Senovilla, L.; Lainey, E.; Thepot, S.; Metivier, D.; Sebert, M.; Baud, V.;
Billot, K.; Fenaux, P.; Galluzzi, L.; Boehrer, S.; Kroemer, G.; Kepp, O. Oncogene
2012, 31, 3536.
24. Qiu, M.; Chen, Y.; Chu, Y.; Song, S. W.; Yang, N.; Gao, J.; Wu, Z. W. Antiviral Res.
2013, 100, 44.
25. Krenn, B. M.; Gaudernak, E.; Holzer, B.; Lanke, K.; Van Kuppeveld, F. J. M.;
Seipelt, J. J. Virol. 2009, 83, 58.
26. Bussiere, D. E.; Pratt, S. D.; Katz, L.; Severin, J. M.; Holzman, T.; Park, C. H. Mol.
Cell 1998, 2, 75.
27. Maestro v9.3, Glide v5.8, Impact v5.8, Liason v5.8, Macromodel v9.9., Prime
v3.1, Schrödinger, LLC.: New York, NY 2012.
28. Lu, D.; Sham, Y. Y.; Vince, R. Bioorg. Med. Chem. 2010, 18, 2037.
29. Tang, J.; Maddali, K.; Pommier, Y.; Sham, Y. Y.; Wang, Z. Q. Bioorg. Med. Chem.
Lett. 2010, 20, 3275.
30. Abudari, K.; Raymond, K. N. Inorg. Chem. 1991, 30, 519.
31. Anissimova, M.; Yaouancq, L.; Badet-Denisot, M. A.; RamaBadet, B. J. Pept. Res.
2003, 62, 88.
We have synthesized two pyrithione amino acid compounds
using a novel vanX inhibitor design frame-work. These compounds
exhibit low micromolar biochemical inhibitory activity that were
comparable to previously reported transition state mimic inhibi-
tors, low cytotoxicity against human cells and the ability to pene-
trate and re-establish vancomycin sensitivity against VREF cell
culture. Since neither compounds 4 nor 5 exhibits any antibacterial
activity by themselves, they act as true re-sensitizing agents of
vancomycin. We recognize a one to two fold reduction in vanco-
mycin concentration may not appear to be significant; however,
this is the first proof of concept that a cell permeable vanX inhib-
itor can re-sensitize vancomycin against a vancomycin resistant
bacterial pathogens. To the best of our knowledge, we have over-
come a major drug design problem over the last two decades by
utilizing heterocyclic thiohydroxamates in our design. Finally,