Fragment-Based Discovery of Novel Non-Hydroxamate LpxC Inhibitors with Antibacterial Activity

Article abstract of DOI:10.1021/acs.jmedchem.0c01215

UDP-3-O-acyl-N-acetylglucosamine deacetylase (LpxC) is a zinc metalloenzyme that catalyzes the first committed step in the biosynthesis of Lipid A, an essential component of the cell envelope of Gram-negative bacteria. The most advanced, disclosed LpxC inhibitors showing antibacterial activity coordinate zinc through a hydroxamate moiety with concerns about binding to other metalloenzymes. Here, we describe the discovery, optimization, and efficacy of two series of compounds derived from fragments with differing modes of zinc chelation. A series was evolved from a fragment where a glycine moiety complexes zinc, which achieved low nanomolar potency in an enzyme functional assay but poor antibacterial activity on cell cultures. A second series was based on a fragment that chelated zinc through an imidazole moiety. Structure-guided design led to a 2-(1S-hydroxyethyl)-imidazole derivative exhibiting low nanomolar inhibition of LpxC and a minimum inhibitory concentration (MIC) of 4 μg/mL against Pseudomonas aeruginosa, which is little affected by the presence of albumin.

Full text of DOI:10.1021/acs.jmedchem.0c01215

pubs.acs.org/jmc
Article
Fragment-Based Discovery of Novel Non-Hydroxamate LpxC
Inhibitors with Antibacterial Activity
Yousuke Yamada,* Hajime Takashima, David Lee Walmsley, Fumihito Ushiyama, Yohei Matsuda,
Harumi Kanazawa, Toru Yamaguchi-Sasaki, Nozomi Tanaka-Yamamoto, Junya Yamagishi,
Risa Kurimoto-Tsuruta, Yuya Ogata, Norikazu Ohtake, Hayley Angove, Lisa Baker, Richard Harris,
Alba Macias, Alan Robertson, Allan Surgenor, Hayato Watanabe, Koichiro Nakano, Masashi Mima,
Kunihiko Iwamoto, Atsushi Okada, Iichiro Takata, Kosuke Hitaka, Akihiro Tanaka, Kiyoko Fujita,
Hiroyuki Sugiyama, and Roderick E. Hubbard*
Cite This: https://dx.doi.org/10.1021/acs.jmedchem.0c01215
Read Online
ACCESS
Metrics & More
Article Recommendations
*
sı Supporting Information
ABSTRACT: UDP-3-O-acyl-N-acetylglucosamine deacetylase (LpxC) is a
zinc metalloenzyme that catalyzes the first committed step in the
biosynthesis of Lipid A, an essential component of the cell envelope of
Gram-negative bacteria. The most advanced, disclosed LpxC inhibitors
showing antibacterial activity coordinate zinc through a hydroxamate moiety
with concerns about binding to other metalloenzymes. Here, we describe the
discovery, optimization, and efficacy of two series of compounds derived
from fragments with differing modes of zinc chelation. A series was evolved from a fragment where a glycine moiety complexes zinc,
which achieved low nanomolar potency in an enzyme functional assay but poor antibacterial activity on cell cultures. A second series
was based on a fragment that chelated zinc through an imidazole moiety. Structure-guided design led to a 2-(1S-hydroxyethyl)-
imidazole derivative exhibiting low nanomolar inhibition of LpxC and a minimum inhibitory concentration (MIC) of 4 μg/mL
against Pseudomonas aeruginosa, which is little affected by the presence of albumin.
INTRODUCTION
non-hydroxamate metal binding groups. Hydroxypyrone-
containing compounds such as 2 have been discussed and
■
3
There has been growing concern over the emergence of
bacterial resistance to existing therapeutics and the
challenges of discovering new antibiotics. An attractive
target for the development of novel therapeutic agents is the
1
are proposed to form a bidentate complex with the zinc ion
(patent WO 2015/085238 from Forge Therapeutics), and
further examples of this motif have been disclosed in pyrone
compounds such as 3 (Example 49 in patent WO 2017/
083434 from Forge Therapeutics). However, no details of the
pyrone-containing compounds have yet been published.
2
bacterial enzyme UDP-3-O-acyl-N-acetylglucosamine deacety-
3
−5
lase (LpxC).
The enzyme is essential to most Gram-
6
,7
negative pathogens, catalyzing the removal of an N-acetyl
group from UDP-3-O-acyl-N-acetylglucosamine, which forms
the core of Lipid A, the first committed step in biosynthesis
of the lipopolysaccharide (LPS), which is an essential
component of the bacterial cell wall (Figure 1). There are
no mammalian homologues, and several series of inhibitors
3
Compound 2 is referred to as LpxCi-1 in ref; this should
not be confused with a hydroxamate-containing compound
named LpxCi-1 in the Achaogen patent WO 2011/005355.
The published crystal structure of 1 (PF-5081090) in the
12
active site of LpxC from Pseudomonas aeruginosa (PaLpxC)
3
−5
illustrates key features of inhibitor binding as shown in Figure
(
summarized in ref ) have been developed (representatives
3
. The protein coordinates the catalytic zinc ion through side
in Figure 2). The most advanced disclosed compounds
contain a hydroxamate functionality that complexes the zinc
ion at the core of the active site of the metalloenzyme. One
chains of D241, H78, and H237 with a trigonal pyramidal
geometry. In 1, the hydroxamate makes a bidentate
interaction to complete a trigonal bipyramidal geometry at
8
,9
of these compounds (ACHN-975 ) entered clinical trials
but was discontinued due to side effects not thought to be
target-related. Although other properties of ACHN-975 may
9
Received: July 29, 2020
be important, the hydroxamate moiety has been linked to
1
0
mutagenicity in HDAC inhibitors
and also to be
responsible for off-target effects due to the lack of specificity
11
in many matrix metalloprotease (MMP) inhibitors. There is
therefore interest in identifying LpxC inhibitors containing
©
XXXX American Chemical Society
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
A
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Figure 1. LpxC catalyzes the first committed step of LPS biosynthesis (shown is the substrate for Pseudomonas aeruginosa).
Figure 2. Representative LpxC inhibitors.
Figure 3. Detail of the crystal structure (PDB code: 5UPG) of 1 (PF-5081090) binding to the active site of PaLpxC. Unless otherwise stated,
in all structural figures, compounds are shown by sticks with orange carbons, the zinc ion as a van der Waals sphere, and protein amino acids
(side chains only) in ball and stick. Except for (b) and 4(a), all other figures of compounds binding to LpxC are presented with the same view
and convention. (a) Inhibitor binding site, with the inside of the solvent surface in gray and the outside in yellow. (b) Detail of inhibitor and
protein binding to a zinc ion, showing just the hydroxamate moiety of the ligand, the NE2 atoms of the imidazole of histidines, and one oxygen
of the side chain of aspartic acid. Note that 1 also binds in a second site, described as the F181 site later in this manuscript.
the zinc ion, and the pendant sulfone occupies the pocket
that binds a phosphate of UDP (see the structure with the
RESULTS AND DISCUSSION
■
Establishing a Structure-Based Discovery Platform.
13
PDB code: 4MDT ) adjacent to K238. The pyridone ring
makes only hydrophobic interactions with the protein binding
site, acting as a linker that connects the core zinc-chelating
unit to groups that mimic the acyl chain in the natural
substrate that binds in a hydrophobic tunnel that leads to the
solvent. These features of the binding site are exploited by
most published inhibitors: (a) a zinc-chelating motif, (b) a
polar group occupying the UDP binding pocket, (c) a linker
that is mainly hydrophobic but sometimes makes at least one
hydrogen bond interaction, (d) an extended hydrophobic
moiety that binds in the tunnel, and (e) often addition of
groups out into the solvent at the end of the tunnel to
modulate compound properties. Crystal structures of different
compounds bound to LpxC from various organisms show
variation in the extent and nature of the tunnel, which leads
to varying specificity for some compounds for different
Two constructs of the enzyme were used in the work
reported here. A construct for residues 1 to 303 with a C-
terminal his-tag (hereafter PaLpxC (1−303-His)) was used
for the functional assay. A construct for residues 1 to 299
(
based on the published structure, PDB code: 5UPG) was
generated with the surface mutation C40S to reduce protein
aggregation (hereafter PaLpxC) and used for all structural
and biophysical measurements, including the binding assay.
Initial attempts to express and purify N-terminal poly-
histidine-tagged PaLpxC produced insoluble aggregates.
Removing the tag with proteolytic cleavage did not improve
the solubility. Expression of untagged PaLpxC with a reduced
induction time resulted in a protein that was more soluble
1
15
but did not generate crystals. Analysis by H− N HSQC
1
5
NMR of a N-labeled sample of PaLpxC (Figure S1a)
showed fewer amide peaks than expected, suggesting that the
protein is partially unfolded; adding zinc improved the
3
,14
bacterial species.
1
15
Here, we report the fragment-based discovery, optimiza-
tion, and characterization of a potent, non-hydroxamate-
containing inhibitor of PaLpxC, which shows efficacy in
various measures of antibacterial activity.
H− N HSQC NMR (Figure S1b), but the protein still did
not form crystals. Thermal shift analysis (TSA) suggested
there was still heterogeneity; maintaining zinc in the
expression medium resulted in a protein with improved
B
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Figure 4. Fragments are found that bind to different sites on LpxC. (a). Schematic of the protein structure, with F181 in ball and stick, with
ligand in CPK for the crystal structures of a zinc-coordinating glycine fragment (6, yellow), a tunnel binder (7, brown), and binding to the
F181 site (8, green). A second copy (not shown) of 8 binds in the tunnel. (b) Detail of the binding of glycine fragment (6) with just the main
chain of M62 shown (see Figure 3 for the legend). (c). Representative fragment hits.
Table 1. Glycine Series: SAR for Interacting with Zinc and the Substrate Binding Site
behavior in TSA (Figure S2) and that did generate crystals.
This protein was used in all subsequent experiments.
A binding assay and a functional assay were used to
characterize inhibition by compounds. The binding assay
used fluorescence polarization (FP) to measure the displace-
ment of a novel fluorescently labeled hydroxamate compound
different functional assays were used. For the imidazole series,
the functional assay measured activity by the reaction of
fluorescamine with the free amino group generated on the
substrate by LpxC. The glycine series compounds gave
interference in the fluorescamine assay, so a liquid
chromatography mass spectrometry (LCMS) assay was
developed to measure the amount of the product formed.
For all assays, compound inhibition or binding is reported as
5
(see SI) from PaLpxC. This probe resulted from careful
optimization of the labeling position and the fluorophore to
minimize artifacts from slow binding kinetics probably due to
the architecture of the active site of the enzyme. The
fluorophore is attached to the sulfone, which occupies the
UDP binding pocket and extends out into the solvent. Two
IC50 values and ligand efficiency (LE) as −RT ln IC /heavy
50
atom count.
15,16
Hit Identification. The Vernalis fragment library
of
1152 compounds was screened against PaLpxC using three
C
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Table 2. Glycine Series: SAR for Extension into the Hydrophobic Tunnel
2
+
ligand-observed NMR experiments (STD, water-LOGSY, and
structures). The zinc-chelating amine group (N−Zn
1
7,18
CPMG)
in pools of six compounds. A binding signal was
distance of ∼2.1Å in all glycine series crystal structures) of
the amino acid also coordinates to H264 and E77. As
expected, modification of that amine group to either
dimethylamino, methylamino, hydroxyl, or morpholino led
to loss of both binding and enzyme inhibitory activity (data
not shown).
We first focused on modification of the amino acid side
chain of 6 to optimize interactions in the substrate binding
site. Inspection of the crystal structure predicted that the
activity would be improved in a D-Ser derivative, which
introduces a hydroxyl group to interact with the side chain of
T190. This was confirmed in 10 (Table 1 and Figure S4a)
with an additional interaction seen in the crystal structure
with D241 that also chelates the zinc ion. As predicted, the
L-Ser derivative 11 showed a significant decrease in activity;
however, the replacement of the hydroxyl of D-Ser with an
amino group 12 was less successful. This could be for two
reasons. First, the calculated pKa of the chelating NH is
reduced from 6.5 in 10 (hydroxyl) to 4.4 in 12 (amino). The
pKa is a measure of the strength of bonding of H to the
NH of the amino acid; this can be taken as a surrogate for
the chelating strength of the NH to the positively charged
zinc ion, so a decrease in pKa will mean a decrease in
chelating strength. The other reason for the reduction in
binding could be due to changes in hydrogen bonding. As
shown in Figure S4a, the hydroxyl in 10 donates one
hydrogen bond (D241) and accepts one hydrogen bond
observed for 252 compounds in one or more NMR
experiments. Putative hits were retested as singletons and
for competition with either 1 or 4, chosen to occupy more of
the UDP binding pocket (see the modeled structure in Figure
S3). Sixteen fragments were competitive with 1 (four of
which were positive in all three NMR experiments).
Representative hits are shown in Figure 4 (all listed in
Table S1). The remaining noncompetitive fragment hits were
assessed for competition with 4, resulting in a further 12
confirmed fragment hits, 4 in all three NMR experiments
(
structures not shown). These 28 fragment hits were assessed
in the FP binding assay. Most bound with lower affinity than
the nominal IC50 threshold of 2.5 mM, but some showed
clear inhibition in the functional assay such as 6 with an IC50
of 41.9 μM with a similar IC50 in the FP assay.
As shown in Figure 4, crystal structures of representative
fragments bound to PaLpxC revealed that molecules that
were competitive with the hydroxamate tool compounds in
the NMR experiments bound to two sites−coordinated to the
zinc (6) or in the tunnel (7). The tunnel is collapsed
somewhat when small zinc-coordinating compounds are
bound but expands when compounds are in the tunnel.
Some fragments (such as 8) that are noncompetitive with the
hydroxamate tool compounds were found to bind to a site
distant from the active site (more than 14 Å through the
protein to the zinc ion away from the substrate binding site).
This site is preformed in all crystal structures and was also
observed in the published crystal structure of 1 (PF-
2
+
2
2
19
(T190). Changing the hydroxyl for a positively charged
nitrogen (NH ) with three polar hydrogen donors would
+
3
5
081090) bound to PaLpxC. The binding site is provided
require T190 to rotate and become an acceptor and still leave
one buried unsatisfied hydrogen bond donor.
by side chains from several hydrophobic residues (F152 and
F176) with F181 at the entrance. Such F181 site compounds
were not active in the functional assay and not pursued
further.
Two zinc-chelating fragments were selected for inves-
tigation, a glycine fragment 6 and an imidazole fragment 9,
both of which were competitive with 1. In the TSA, 6 and 9
stabilize PaLpxC by 4.0 °C and 0.78 °C, respectively, at 2
mM compound as an additional confirmation of binding.
Fragment to Hit for the Glycine Series. As seen in the
crystal structure (Figure 4b), the phenyl ring of the aniline of
Modeling from the co-crystallized structure of 10 suggested
that a sulfonyl group linked by two carbons but not by one
carbon could interact with K238, similar to what is seen for
1. The increase in potency was confirmed in 14 (compared
to 13) but not improved by replacing the sulfone with a
sulfonamide in 15, perhaps because the chelating ability of
the primary amino group was reduced (pKa = 6.2 in 14, but
decreased to pKa = 5.5 in 15) or perhaps again because of a
slight movement of side chains altering the detail of hydrogen
bonding. Further modeling suggested that the sulfone could
be replaced by a carboxylic acid with a 1-carbon linker to
make additional interactions with H264. This carboxylic acid
derivative 16 retained potency, and the predicted binding
modes of 14 and 16 to PaLpxC were confirmed in crystal
structures (Figure S4b,c). However, these modifications
6
forms the linker toward the entrance to the tunnel and is
sandwiched between hydrophobic amino acid side chains
L18 and A214), while the NH of the aniline makes a water-
(
bridged hydrogen bond to the main chain of M62 (the side
chain of this residue is disordered in some of the crystal
D
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Table 3. Glycine Series: Antibacterial Activity
MIC (μg/mL)
P. aeruginosa+ PAβN
b
c
no.
R
functional IC50 (μM)
E. coli+ PAβN
E. coli
P. aeruginosa
2
1
H
0.0221
0.00603
16
1
>32
64
32
8
>32
64
a
2
2
CH SO CH
3
2
2
a
b
c
Racemic mixture. Escherichia coli ATCC25922. Pseudomonas aeruginosa ATCC27853.
Figure 5. (a). Detail of the crystal structure of 22 binding to the active site of PaLpxC(1-299). See the legend of Figure 3 for conventions used.
b). Different view showing zinc coordination to the selected ligand and protein side-chain atoms showing CD1, NE2, and CD2 of the
(
imidazole of histidines and the carboxylic acid groups of the glutamic and aspartic acids.
a
Table 4. Imidazole Series: Improving Zinc Chelation
a
*
denotes racemic mixture.
degraded ligand efficiency, so subsequent SAR fixed the D-
Ser as the zinc-chelating moiety and explored extension into
the hydrophobic tunnel region.
The bound structure of 6 showed limited opportunities for
further interactions from the aniline. Replacing the trifluor-
omethoxy group with extended ethers could not provide a
E
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
good vector into the tunnel region. In addition, the amide
NH interacts with the protein only through a water-bridged
hydrogen bond. We modeled substituted piperidines and
pyrrolidines incorporating additional sp3 carbons to explore
the vectors toward the tunnel and also reduce the number of
polar H-bond donors. Replacing the aniline with 4-benzyl
piperidine (as in 17) provides a linker with the benzyl group
in the entrance of the tunnel with a synthetically accessible
vector into the tunnel. Various hydrophobic groups (Table 2)
were appended to the para position of the benzyl, which
successfully increased potency with retained ligand efficiency.
The length and rigidity of the linker between aromatic groups
had an effect on activity (as seen comparing 18, 19, and 20),
emphasizing the importance of achieving the correct
placement in the hydrophobic tunnel.
Figure 6. Detail of the crystal structure of 29 binding to the active
site of PaLpxC. See the legend of Figure 3 for conventions used.
Introducing an alkyne linker places the second phenyl ring
centrally into the hydrophobic tunnel formed by L18, A214,
A206, and I197, and addition of a para-hydroxymethyl (in
enantiomer of the 2-(1-hydroxyethyl) group is the most
potent.
This exploration of the SAR of the initial imidazole
fragment hit illustrates two important features of fragment-
2
1) on this second ring extends into the solvent at the exit of
20,21
the tunnel. This further improved enzyme activity with some
loss of ligand efficiency. Although 21 inhibited the enzyme
with an IC50 of 20 nM, this did not translate into
antibacterial activity (Table 3). Addition of a sulfonyl group
based discovery.
First, the fragment library typically
consists of a small number of compounds, selected on
physicochemical properties and some measure of chemical
16
diversity. As with any screening approach, the hope is that
the library contains something that binds well enough to be
characterized. It is then important to optimize the features of
the fragment that are making the key interaction with the
target binding site, before embarking on large-scale
optimization of druglike properties. The second feature is
that fragments can provide a diversity of start points for
(
as in 14 to interact with K238) further improved the
enzyme activity to single digit nanomolar (22). The crystal
structure of 22 bound to PaLpxC (Figure 5) confirmed the
designed interactions with the protein and the placement of
the aromatic rings in the tunnel. The geometry of zinc
binding is essentially conserved across the series (Figures S4
and 5b), with a trigonal bipyramidal coordination of the zinc
ion, with two oxygens (OD of D241 and ligand carbonyl O)
defining the z-axis. However, this most potent of the glycine
series showed only weak antibacterial activity in the presence
of phenylalanine-arginine β-naphthylamide (PAβN) (an efflux
pump inhibitor), so the series was not taken further.
22,23
optimization.
In this study, the zinc-chelating groups of
the glycine and imidazole series bind so that the remainder of
the compound provides different vectors and opportunities
for reaching into the linker and tunnel regions, reflected in
the different amino acid groups that are contacted.
Hit to Lead for the Imidazole Series. The co-crystal
structure of 29 bound to PaLpxC (Figure 6) suggested how
to extend the compound into the hydrophobic tunnel of
LpxC with moieties similar to those in the glycine series and
other LpxC inhibitors. The exploration of phenyl rings and
lengths of a linker (Table 5) showed that, as for the glycine
series, the enzyme inhibitory activity tended to improve with
elongation of the aromatic portion of the molecule. The
ligand efficiency for the placement of a biphenyl moiety into
the tunnel was optimal for a 1-carbon (32) or a 3-carbon
(34) rather than a 2-carbon linker (34), similar to what is
observed for a 2-carbon vs 3-carbon linker to chlorophenyl,
comparing 30 to 29 and 31. This feature was retained with
further extension through the hydrophobic tunnel with a Ph-
alkyne-Ph moiety (compare 36 with 35 and 37) with 35 and
37 showing submicromolar functional activity. Although co-
crystal structures of this subset of compounds bound to
PaLpxC could not be obtained (poor compound behavior),
modeling suggested that the variation in the linker length
affected the vector between the hydrophobic tunnel binding
group and the zinc-chelating groups with a consequent effect
on activity. A 1-carbon linker was therefore chosen as optimal
for ligand efficiency.
Fragment to Hit for the Imidazole Series. The initial
imidazole fragment hit 9 bound weakly to PaLpxC.
Exploration of the activity of closely related (near neighbor)
compounds established that addition of chlorine to the
phenyl moiety (23) enhanced activity. The replacement of
the hydroxyl by amino or methoxy groups did not lead to
improvement (data not shown), but removal of the hydroxyl
(24) and the oxygen linker (25) was beneficial (Table 4)
with a much-improved ligand efficiency. Reducing the length
of the 4-carbon aliphatic linker between imidazole and
chlorophenyl to 2-carbons as in 26 had an impact on activity;
3
-carbons retained the same ligand efficiency (27), and this
was the first imidazole compound to give a crystal structure
Figure S5). This confirmed that the imidazole is a
(
monodentate chelator of the zinc ion, placing the aliphatic
chain and chlorophenyl toward the hydrophobic tunnel.
The first optimization was of the imidazole to improve
chelation by conversion into a bidentate zinc binder. The
crystal structure (Figure S5) suggested that extending from
the 2 position of the imidazole would be effective and 2-
hydroxymethyl (28) retained ligand efficiency, which was
further enhanced in 29, with a 2-(1-hydroxyethyl) group. The
co-crystal structure of 29 with PaLpxC (Figure 6) confirmed
that the compound forms a bidentate complex with trigonal
bipyramidal geometry to the zinc through the hydroxyl
oxygen atom and an imidazole nitrogen (see also Figure 8b
for details of coordination) to give an optimized hit
compound. The structure also suggested that the S
Compound 35 showed weak antibacterial activity against E.
coli only in the presence of PAβN (data not shown). This
was probably associated with the high lipophilicity of the
compound, and so the next step in optimization was to
reduce the lipophilicity by introduction of a heterocycle ring
in place of the first phenyl ring attached to the linker.
F
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Table 5. Imidazole Series: SAR for Extension into the
a
Hydrophobic Tunnel
Figure 7. Detail of the crystal structure of 38 binding to the active
site of PaLpxC. The carbonyl of M62 is alone; otherwise, see the
legend of Figure 3 for conventions used.
as seen in 22 of the glycine series extends through the end of
the tunnel into the solvent to give compounds with low
nanomolar inhibition of the enzyme activity and with
compound properties that provide antibacterial activity.
Analysis of 41 and 42 after chiral separation of the racemic
4
0 confirmed that the S enantiomer is more potent (>100×
enzyme inhibition). The S enantiomer shows antibacterial
activity against a range of Gram-negative bacteria, though the
most interesting activity against P. aeruginosa is abolished in
the presence of human serum albumin (HSA). Further
exploration of the solubilizing group produced compounds 44
to 46, which show similar levels of enzyme inhibition but
with more potent antibacterial activity. Compound 46 has the
most potent profile of antibacterial activity across a range of
bacterial species, but 43 (the structure in Figure 8) shows the
least influence of HSA.
a
*
denotes racemic mixture.
Selectivity of Lead Glycine (22) and Imidazole (43)
Compounds. Table 7 lists the inhibition measured for the
lead compound of each series and a representative
hydroxamate compound (BB-78485 ) against a small panel
of matrix metalloprotease enzymes, MMP-2, MMP-3, and
MMP-9. The results illustrate that the structure-guided design
of each series has generated compounds with selectivity for
LpxC over some representative zinc metalloenzymes, in
contrast to previously described hydroxamate LpxC inhib-
itors.
Modeling also suggested that a 5-membered hetero-aromatic
ring would favorably adjust the vector between the hydro-
phobic tunnel and zinc atom binding groups. A number of
different heterocycles such as variously substituted 1,3-
oxazoles, pyrazoles, triazoles, oxadiazoles, and thiazoles were
explored (data not shown), but replacement with a 1,2-
oxazole was particularly effective, reducing the cLogP (0.9 for
24
38 compared to 2.6 for 32) but also improving the enzyme
inhibitor activity by a factor of 4 with a significant
improvement in ligand efficiency in both binding and
functional assays. This jump in potency can be explained
by inspection of the co-crystal structure of 38, which shows
that the five-membered ring creates a turn in the ligand. This
change in shape matches the binding site of PaLpxC (Figure
Synthesis of 22 and 43. The synthesis of the advanced
compounds 22 and 43 illustrates the general scheme for
synthesis of the glycine and imidazole series, respectively. The
experimental details are provided in the Experimental Section,
and the schemes and synthetic details for all tested
compounds are provided in the Supporting Information
(SI). These compounds have previously been described in
the patents WO 2018/216822, WO 2018/216823, and
WO2020/105660. The purity of all compounds was
measured by LC, and the identity of compounds was
confirmed by MS. The purity of all compounds was >
95%, except for compounds 10 and 15, which were 92%
pure. LC and mass spectrometric traces for the six key
compounds 10, 25, 29, 38, 41, and 43 are provided in SI.
The synthesis of compound 22 is outlined in Schemes 1
and 2. Scheme 1 shows the synthesis of the amino acid
portion of the molecule, intermediate 53. The Aldol reaction
using lithium diisopropylamide and 2-methylsulfanylacetalde-
hyde gave the intermediate thioether as a racemic mixture of
diastereomers 49a and 49b. The diastereomers were
separated via normal-phase silica gel column chromatography
to give the rac-(R*,R*) diastereomer 49a and the rac-(R*,S*)
7
), and the ligand makes an additional interaction with H19
of the enzyme. The angle and distance to the NE2 atom of
H19 are not consistent with a strong hydrogen bond. We
noted that the orientation of this histidine does vary between
structures depending on the opportunities for alternate
hydrogen bonding interactions, notably between H19 and
M62 with small adjustments to the main chain conformation.
Compound 39 is a compound with sub-50 nM functional
activity, which combines the features learned from the various
series to be important for binding, affinity, and antibacterial
activitya 2-(1S-hydroxyethyl)imidazole moiety for zinc
chelation, 1,2-oxazole for interactions in the linker region,
and alkyne to provide a suitable vector to place a phenyl
group centrally in the hydrophobic tunnel.
A series of compounds were synthesized based on 39
Table 6), varying a solubilizing group on the phenyl, which
(
G
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Table 6. Imidazole Series: Antibacterial Activity for Various Gram-Negative Bacteria with a variation in the Solubilizing
,
a b
Group
a
b
rac = racemic mixture. 1: Functional enzyme activity; 2: E. coli ATCC25922; 3: P. aeruginosa ATCC27853; 4: K. pneumoniae ATCC13883;
5: P. aeruginosa 5567.
Figure 8. (a) Detail of the crystal structure of 43 binding to the active site of PaLpxC. Clipped, semitransparent solvent accessible surface,
yellow outside and gray inside. See the legend of Figure 3 for other conventions used. (b) Different view showing zinc coordination to the
selected ligand and protein side-chain atoms as described in the legend to Figure 5(b).
a
Table 7. Inhibition of MMP Enzymes
triflate 54 (synthesis described in WO 2018/216823)
followed by deprotection afforded the intermediate piperidine
compound
MMP-2
MMP-3
MMP-9
5
6 using HCl in 1,4-dioxane. Amide coupling was performed
2
4
2
3
>100
>100
>100
>100
0.416
0.0150
>100
>100
0.0972
0.000177
using HATU, and the final boc-deprotection was facilitated
by formic acid.
The synthesis of compound 43 is outlined in Schemes
BB-78485
Ilomastat (MMP inhibitor)
0.0175
0.000422
3
−5.
a
Values are IC50 (μM).
Scheme 3 shows the preparation of intermediate 60 via
nucleophilic displacement of the tosylate with 4-iodophen-
oxide, followed by reduction of the nitro group to the
primary amine.
diastereomer 49b in a 1:1.3 ratio. The relative stereo-
chemistry was determined by X-ray crystallography of ethyl
rac-(R*,S*)-2-(dibenzylamino)-3-hydroxy-4-(methylsulfonyl)-
butanoate (50b). The rac-(R*,R*) diastereomer 49a was
subsequently oxidized to the sulfone 50a using m-
chloroperbenzoic acid.
Deprotection of the amine by hydrogenation was followed
by tert-butoxycarbonyl (boc)-protection and hydrolysis of the
ethyl ester to afford the intermediate acid 53. Scheme 2
shows the preparation of the acetylenic portion of the
compound and convergent amide coupling with intermediate
Scheme 4 shows the synthesis of chiral imidazole
intermediate 62. Condensation cyclization of (2S)-2-tetrahy-
dropyran-2-yloxypropanal 61 with ammonia and glyoxal
afforded the tetrahydropyran-protected intermediate that
was subsequently used in the synthesis of several example
compounds. Scheme 5 shows the synthesis of compound 43
starting from commercial iodo-1,2-oxazole ester 63. Sonoga-
shira coupling installed the triisopropylsilyl (TIPS)-protected
terminal acetylene group, and a borohydride reduction of the
53 to make compound 22. Sonogashira coupling with the
H
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
Scheme 1. Synthesis of rac-(R*,R*) Acid as the Sodium Salt (53)
a
Reagents and conditions: (a) N,N-diisopropylethylamine (DIPEA), n-BuLi, 2-methylsulfanylacetaldehyde, tetrahydrofuran (THF), −78 °C to RT,
3
0 min, chromatography EtOAc/n-hexane, diastereoisomer rac-(R*,R*) 49a, 15%, diastereoisomer rac-(R*,S*) 49b, 25%; (b) m-chloroperbenzoic
acid, CHCl , 0 °C to RT, 16 h, 47%; (c) hydrogen, palladium hydroxide on carbon, EtOH, RT, 16 h, quant; (d) di-tert-butyl dicarbonate
3
((BOC) O), K CO , THF, EtOH, RT, 3 h, quant; (e) NaOH (10 M aq), EtOH, RT, 3 h, 73%.
2 2 3
a
Scheme 2. Synthesis of 2-Amino-3-hydroxy-4-(methanesulfonyl)butan-1-one (22)
a
Reagents and conditions: (a) 4-ethynylbenzenemethanol, PdCl (PPh ) , CuI, triethylamine (TEA), dimethylformamide (DMF), 50 °C, 30 min,
2
3 2
4
0%; (b) 1,4-dioxane, 4N HCl in 1,4-dioxane, RT, 3 h, quant; (c) compound 53, O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium
hexafluorophosphate (HATU), DIPEA, DMF, RT, 3 h, 96%; (d) formic acid, RT, 2 h, 35%.
a
Scheme 3. Synthesis of Compound 60
a
Reagents and conditions: (a) 4-iodophenol, NaH, DMF, 100 °C, 16 h, 75%; (b) iron powder, ammonium chloride, EtOH, Water, 70 °C, 15 h,
8
8%.
a
Scheme 4. Synthesis of Imidazole Compound 62
Mitsunobu conditions. Following on from deprotection of
the TIPS group to form the terminal acetylene using TBAF, a
second Sonogashira coupling was performed with compound
6
0. The final deprotection of the tetrahydropyranyl (THP)-
protected alcohol and acetonide was performed using p-
toluenesulfonic acid to obtain compound 43.
a
Reagents and conditions: (a) 28% aqueous ammonia, glyoxal,
MeOH, RT, 16 h, quant.
CONCLUSIONS
■
Fragment-based discovery methods have identified two non-
hydroxamate series of compounds that are selective inhibitors
of the zinc metalloenzyme, LpxC. Optimization of both series
was guided by structure-based design to modify zinc
ester afforded the intermediate alcohol 65. The convergent
coupling step with imidazole 62 was achieved under
I
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
Scheme 5. Synthesis of Compound 43
a
Reagents and conditions: (a) ethynyltri(propan-2-yl)silane, TEA, Pd(PPh ) , CuI, DMF, 80 °C, 1 h, 69%; (b) NaBH , EtOH, 0 °C, 4 h, 89%; (c)
3
4
4
6
2, 3-(dimethylcarbamoylimino)-1,1-dimethylurea (TMAD), tributylphosphine, THF, RT, 2 h, 93%; (d) tetra-n-butylammonium fluoride (TBAF),
THF, 0 °C, 1 h, 82%; (e) 60, TEA, PdCl (PPh ) , CuI, DMF, 80 °C, 0.5 h, 55%; (f) p-TsOH, MeOH, RT, 1 h, 87%.
2
3 2
chelation, make interactions with the substrate binding cleft,
and introduce improved compound properties. The 2-(1S-
hydroxyethyl)imidazole 43 is an advanced lead compound for
which further optimization and in vivo efficacy will be
described in a future publication.
(ionization method: electron spray ionization (ESI)), and Agilent
3
85-ELSD was used when the ELSD detector was attached.
r a c - ( R * , R * ) - 2 - A m i n o - 3 - h y d r o x y - 1 - { 4 - [ ( 4 - { [ 4 -
(
hydroxymethyl)phenyl] ethynyl}phenyl)methyl]piperidin-1-
yl}-4-(methanesulfonyl)butan-1-one (22). Step 1.
Ethyl rac-(R*,R*)-2-(Dibenzylamino)-3-hydroxy-4-
(
methylthio)butanoate (49a) and Ethyl rac-(R*,S*)-2-(Diben-
zylamino)-3-hydroxy-4-(methylthio)butanoate (49b). To a
EXPERIMENTAL SECTION
■
stirred solution of diisopropylamine (4.7 mL, 65 mmol) in THF
All chemicals were used directly as received from commercial
(0.21 L) was added 1.6 mol/L n-butyllithium in hexane (20 mL, 32
1
13
suppliers. H and C NMR spectra were recorded on JNM-ECA600
mmol) at −78 °C. The mixture was allowed to warm to 0 °C,
stirred for 30 min, then cooled to −78 °C, and 48 (6 g, 21 mmol)
in THF was added into the mixture. After stirring for 30 min, 2-
methylsulfanylacetaldehyde (2.3 g, 25 mmol) in THF was added
into the mixture. The reaction mixture was allowed to warm to
room temperature, stirred for further 3 h, quenched with a saturated
1
1
(
600 MHz for H, JEOL), JNM-ECA500 (500 MHz for H and 125
13
1
MHz for C, JEOL) or AVANCE III HD 400 (400 MHz for H,
1
13
BRUKER). Chemical shifts for H and C NMR are reported in
parts per million (ppm) referenced to the residual solvent peak
(
dimethyl sulfoxide (DMSO)-d , δ = 2.50 ppm; CDCl , δ = 7.27
6
3
ppm; CD OD, δ = 3.30 ppm). Abbreviations used to describe the
3
1
aqueous sodium bicarbonate, and extracted with CHCl three times.
3
peak signals in H NMR data are s = singlet, br = broad, d =
The organic layer was passed through a phase separator and
concentrated in vacuo onto ISOLUTE HM-N. The residue was
purified, and diastereomers were separated with OH-type silica gel
column chromatography (8−31% EtOAc in n-hexane) to afford rac-
doublet, dd = double doublet, dt = double triplet, t = triplet, q =
quartet, quin = quintet, m = multiplet. All final compounds were
purified to >95% by reverse-phase high-performance liquid
chromatography (HPLC) or normal-phase silica gel column
chromatography. The purity was assessed by reverse-phase UPLC
with a gradient of 5−98% acetonitrile in water (with either an acid
or a base modifier) and monitored by UV absorption at 210−450
nm.
(
(
R*,R*) diastereomer (1.2 g, 15% yield) 49a as yellow oil and rac-
R*,S*) diastereomer (2.0 g, 25% yield) 49b as a yellow oil. The
relative stereochemistry of both diastereomers was determined by X-
ray crystallography of ethyl rac-(R*,S*)-2-(dibenzylamino)-3-hy-
droxy-4-(methylsulfonyl)butanoate 49b. rac-(R*,R*) diastereomer
Unless otherwise described, the following conditions were used
for compound synthesis, purification, and analysis. A microwave
1
4
4
9a: H NMR (400 MHz, CDCl ) δ 7.20−7.35 (m, 10H), 4.24−
3
+
.44 (m, 2H), 4.06−4.16 (m, 1H), 3.92 (d, J = 13.57 Hz, 2H), 3.48
(d, J = 13.57 Hz, 2H), 3.30 (d, J = 9.05 Hz, 1H), 3.02 (dd, J =
2.87, 14.12 Hz, 1H), 2.81 (d, J = 3.67 Hz, 1H), 2.30 (dd, J = 9.11,
14.12 Hz, 1H), 2.03 (s, 3H), 1.42 (t, J = 7.15 Hz, 3H). MS (ESI ):
m/z 374 [M + H] . rac-(R*,S*) diastereomer 49b: H NMR (400
synthesizer Initiator (Biotage) was used for the microwave reaction
conditions. OH-type silica gel column chromatography was
performed using SNAP Ultra (Biotage) or REVELERIS 40 μm
+
(Grace), and amino-type silica gel chromatography was performed
+
1
using SNAP Cartridge ISOLUTE Flash-NH2 (Biotage) or Grace
REVELERIS Amino 40 μm (Grace). The prep-HPLC purifications
were performed using Agilent 1260 Infinity or Agilent 6130
MHz, CDCl ) δ 7.22−7.39 (m, 10H), 4.22−4.40 (m, 2H), 4.10−
3
4.20 (m, 1H), 4.02 (d, J = 13.20 Hz, 2H), 3.66 (s, 1H), 3.53 (d, J =
J
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
1
3.20 Hz, 2H), 3.39 (d, J = 9.17 Hz, 1H), 2.63 (dd, J = 2.81, 13.94
tert-Butyl 4-(4-((4-(Hydroxymethyl)phenyl)ethynyl)benzyl)-
piperidine-1-carboxylate (55). A mixture of 54 (0.81 g, 1.9
mmol), (4-ethynylphenyl)methanol (0.76 g, 5.7 mmol), CuI (36
Hz, 1H), 2.48 (dd, J = 6.85, 14.06 Hz, 1H), 2.02 (s, 3H), 1.40 (t, J
=
+
+
7.15 Hz, 3H). MS (ESI ): m/z 374 [M + H] .
Ethyl rac-(R*,R*)-2-(Dibenzylamino)-3-hydroxy-4-
mg, 0.19 mmol), Pd(PPh ) Cl (0.13 g, 0.19 mmol), and Et N (10
3 2 2 3
(
(
methylsulfonyl)butanoate (50a). To a stirred solution of 49a
1.2 g, 3 mmol) in CHCl (10 mL) was added m-chloroperbenzoic
mL, 72 mmol) in DMF (10 mL) was stirred under a nitrogen
atmosphere at 60 °C for 7 h. After cooling to RT, the mixture was
passed through the Celite and washed with EtOAc. The filtrate was
concentrated under reduced pressure. The residue was purified with
OH-type silica gel chromatography (7−30% EtOAc/n-hexane) to
3
acid (2.0 g, 11.6 mmol) at 0 °C. The reaction mixture was stirred at
room temperature for 16 h, quenched with a saturated aqueous
sodium thiosulfate, and extracted with CHCl twice. The organic
3
1
layer was passed through a phase separator and concentrated in
vacuo onto ISOLUTE HM-N. The residue was purified with OH-
type silica gel column chromatography (8−100% EtOAc in n-
give the title compound (0.16 g, 20% yield) as a beige solid. H
NMR (400 MHz, CDCl ) δ 7.52 (d, J = 8.07 Hz, 2H), 7.45 (d, J =
3
8.07 Hz, 2H), 7.35 (d, J = 8.07 Hz, 2H), 7.12 (d, J = 8.07 Hz, 2H),
4.72 (br s, 2H), 3.95−4.18 (m, 2H), 2.58−2.70 (m, 2H), 2.55 (d, J
= 6.97 Hz, 2H), 1.58−1.70 (m, 3H), 1.45 (s, 9H), 1.07−1.22 (m,
hexane) to give the title compound (0.58 g, 47% yield) as a
1
colorless powder. H NMR (400 MHz, CDCl ) δ 7.26−7.40 (m,
1
Hz, 2H), 3.55 (br d, J = 15.04 Hz, 1H), 3.46 (d, J = 12.96 Hz, 2H),
3
+
+
0H), 4.58−4.66 (m, 1H), 4.28−4.48 (m, 2H), 3.87 (d, J = 13.08
2H). MS (ESI ): m/z 428 [M + Na] . LCMS condition: C. RT =
1.024 min.
3
.21 (d, J = 9.90 Hz, 1H), 3.07−3.15 (m, 1H), 2.93 (s, 3H), 2.62
(4-((4-(Piperidin-4-ylmethyl)phenyl)ethynyl)phenyl)-
methanol Hydrochloride (56). To a solution of 55 (0.14 g, 0.35
mmol) in 1,4-dioxane (4 mL) was added 4 mol/L HCl in 1,4-
dioxane (8 mL) at RT. The reaction mixture was stirred at RT for 4
h and concentrated in vacuo to give the title compound (0.13 g,
(
(
dd, J = 10.39, 15.04 Hz, 1H), 1.46 (t, J = 7.15 Hz, 3H). MS
ESI ): m/z 406 [M + H] . LCMS condition: C. RT = 0.847 min.
Ethyl rac-(R*,S*)-2-(Dibenzylamino)-3-hydroxy-4-
+
+
(methylsulfonyl)butanoate (50b). To a stirred solution of 49b
quant.) as a pale-yellow solid.
(
0.12 g, 0.32 mmol) in acetonitrile (10 mL) and water (5 mL) was
1
H NMR (400 MHz, DMSO-d ) δ 7.49 (d, J = 8.07 Hz, 2H),
added oxone (0.22 g, 0.35 mmol) at 0 °C. The reaction mixture was
stirred at room temperature for 16 h, quenched with a saturated
6
7
8
2
.44 (d, J = 8.07 Hz, 2H), 7.36 (d, J = 8.07 Hz, 2H), 7.20 (d, J =
.07 Hz, 2H), 5.27 (br t, J = 4.58 Hz, 1H), 4.52 (br d, J = 4.28 Hz,
H), 2.87 (br d, J = 12.23 Hz, 2H), 2.32−2.41 (m, 2H), 1.39−1.62
aqueous sodium thiosulfate, and extracted with CHCl twice. The
3
organic layer was passed through a phase separator and concentrated
in vacuo. The residue was purified with OH-type silica gel column
chromatography (0−100% EtOAc in n-hexane) to give the title
(
m, 3H), 1.03 (dq, J = 3.91, 11.98 Hz, 2H), (two proton signals
+
overlap with the DMSO solvent peak). MS (ESI ): m/z 306 [M +
+
1
H] . LCMS condition: B. Retention time (RT) = 0.543 min.
compound (80 mg, 61% yield) as a colorless powder. H NMR
rac-(R*,R*)-tert-Butyl (3-Hydroxy-1-(4-(4-((4-
(
400 MHz, CDCl ) δ 7.24−7.39 (m, 10H), 4.26−4.47 (m, 3H),
3
(
(
hydroxymethyl)phenyl) ethynyl)benzyl)piperidin-1-yl)-4-
4
3
1
.02 (br d, J = 13.08 Hz, 2H), 3.57 (br d, J = 13.20 Hz, 2H), 3.27−
methylsulfonyl)-1-oxobutan-2-yl)carbamate (57). To a stirred
.41 (m, 1H), 3.09 (dd, J = 8.50, 14.86 Hz, 1H), 2.94−3.02 (m,
+
solution of 56 (34 mg, 0.1 mmol) in DMF (2.0 mL), 53 (35 mg,
0.11 mmol), HATU (57 mg, 0.15 mmol), and DIPEA (52 μL, 0.54
mmol) were added. The mixture was stirred for 3 h at RT. The
reaction mixture was purified via prep-HPLC to give the title
H), 2.91 (s, 3H), 1.42 (t, J = 7.09 Hz, 3H). MS (ESI ) m/z 406
+
[
M + H] . X-ray: A single crystal was grown from ethanol. The
crystal details were as follows: triclinic space group P1
̅
, a =
1
1.9376(2) Å, b = 13.4893(3) Å, c = 15.7344(3) Å, α = 69.216(5)
+
compound (56 mg, 96% yield) as an amorphous solid. MS (ESI ):
°
, β = 83.333(6) °; γ = 88.323(6) °; R and wR values were 0.0931
+
m/z 607 [M + Na] . LCMS condition: C. RT = 0.786 min.
and 0.3597, respectively.
r a c - ( R * , R * ) - 2 - A m i n o - 3 - h y d r o x y - 1 - { 4 - [ ( 4 - { [ 4 -
Ethyl rac-(R*,R*)-2-Amino-3-hydroxy-4-(methylsulfonyl)-
butanoate (51). To a solution of 50a (0.40 g, 1 mmol) in
EtOH (9.9 mL) was added 10% palladium hydroxide on carbon (50
mg, catalytic). The mixture was stirred at room temperature for 16 h
under a hydrogen atmosphere. The reaction mixture was filtered,
and the filtrate was concentrated in vacuo to give crude 51 (0.24 g,
(
hydroxymethyl)phenyl]ethynyl} phenyl)methyl]piperidin-1-
yl}-4-(methanesulfonyl)butan-1-one (22). A solution of 57
53 mg, 0.9 mmol) in formic acid (1.5 mL, 40 mmol excess) was
(
stirred for 2 h at RT. The reaction mixture was diluted with CHCl3
and concentrated in vacuo. To the residue was added 7 mol/L
ammonia in MeOH (2 mL, excess), and the mixture was stirred for
quant.) as oil, which was used in the next step without further
purification. H NMR (400 MHz, CDCl ) δ 4.41 (ddd, J = 2.14,
1
0 min at RT. The reaction mixture was concentrated in vacuo and
1
3
purified with prep-HPLC to give the title compound (11 mg, 35%
yield) as a colorless powder. H NMR (500 MHz, CD OD) δ 7.47
5
.50, 9.48 Hz, 1H), 4.22−4.29 (m, 1H), 3.72 (q, J = 6.97 Hz, 1H),
1
3
3.64 (d, J = 5.50 Hz, 1H), 3.27 (dd, J = 9.90, 15.04 Hz, 1H), 3.19
(d, J = 8.23 Hz, 2H), 7.43 (d, J = 7.89 Hz, 2H), 7.36 (d, J = 8.23
(
br d, J = 14.70 Hz, 1H), 3.04 (s, 3H), 1.32 (t, J = 7.15 Hz, 3H).
Hz, 2H), 7.20 (br d, J = 7.89 Hz, 2H), 4.62 (s, 2H), 4.42−4.57 (m,
+
+
MS (ESI ): m/z 226 [M + H] . LCMS condition: B. RT = 0.258
min.
3
2
H), 3.92−4.00 (m, 1H), 3.36−3.47 (m, 1H), 3.11−3.23 (m, 1H),
.94−3.10 (m, 4H), 2.53−2.77 (m, 3H), 1.84−1.98 (m, 1H), 1.75
Ethyl rac-(R*,R*)-2-((tert-Butoxycarbonyl)amino)-3-hy-
droxy-4-(methylsulfonyl) butanoate (52). To a stirred solution
of 51 (0.17 g, 0.75 mmol as the crude material) in THF (3.8 mL)
13
(
br t, J = 12.60 Hz, 2H), 1.09−1.33 (m, 2H). C NMR (125 MHz,
CD OD) δ 163.2, 162.9, 143.4, 142.1, 141.9, 132.6, 132.6, 130.6,
3
1
5
3
30.5, 128.1, 123.7, 123.6, 122.6, 122.5, 90.1, 90.0, 66.4, 66.4, 64.9,
7.1, 57.0, 56.7, 47.4, 47.2, 44.2, 43.6, 43.4, 43.3, 39.2, 39.0, 33.7,
3.7, 33.0, 32.8, 31.1 (35 carbon signals are observed because of the
and EtOH (1.5 mL) were added (BOC) O (0.20 g, 0.92 mmol) and
2
potassium carbonate (0.27 g, 1.95 mmol). After stirring for 3 h at
room temperature, the reaction mixture was quenched with water
and extracted with CHCl3 three times. The organic layer was
concentrated in vacuo onto ISOLUTE HM-N. The residue was
purified with OH-type silica gel column chromatography (30−100%
EtOAc in n-hexane) to give 52 (0.27 g, quant.) as colorless oil. MS
+
+
rotational isomer). MS (ESI ): m/z 485 [M + H] . high-resolution
mass spectrometry (HRMS) (ESI /APCI) calcd. for [C H N O S
+
+
2
6
32
2
5
+
H] : 485.2105; found: 485.2089. LCMS condition: B. RT = 0.590
+
+
(ESI ): m/z 348 [M + Na] . LCMS condition: B. RT = 0.635 min.
Sodium rac-(R*,R*)-2-((tert-Butoxycarbonyl)amino)-3-hy-
droxy-4-(methylsulfonyl) butanoate (53). To a solution of 52
5
-[(4-Iodophenoxy)methyl]-2,2-dimethyl-5-nitro-1,3-diox-
(
0.28 g) in EtOH (2.9 mL) was added 10 mol/L aqueous sodium
ane (59). To a solution of 4-iodophenol (2.1 g, 9.6 mmol) in DMF
(96 mL) was added NaH (0.50 g, 12 mmol) and stirred for 1.5 h at
room temperature. To the reaction mixture was added 58 (3.6 g, 11
mmol) and stirred for an additional 16 h at 100 °C. The resulting
mixture was diluted with EtOAc and washed with water. The
hydroxide (0.26 mL, 2.6 mmol). After stirring for 3 h at room
temperature, the resulting precipitate was collected by filtration and
air-dried to afford the title compound (0.20 g, 73% yield) as a
−
−
colorless solid. MS (ESI ): m/z 296 [M − H] . LCMS condition:
A. RT = 0.879 min.
organic layer was dried over MgSO and concentrated in vacuo. The
4
K
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
residue was triturated with diethyl ether and collected by filtration
maintained below 5 °C. After stirring for 1 h at 0 °C, the mixture
was concentrated, and the residue was diluted with EtOAc followed
by washing with water and brine. The organic phase was separated,
to afford the title compound (2.8 g, 75% yield) as a colorless solid.
1
H NMR (400 MHz, CDCl ) δ 7.57 (d, J = 8.93 Hz, 2H), 6.67 (d,
3
J = 8.93 Hz, 2H), 4.45 (d, J = 12.72 Hz, 2H), 4.42 (s, 2H), 4.15 (d,
J = 12.59 Hz, 2H), 1.46 (s, 3H), 1.43 (s, 3H).
dried over MgSO , and concentrated under reduced pressure. The
4
residue was purified with OH-type silica gel column chromatography
(50−100% EtOAc in n-hexane and then 0−10% MeOH in EtOAc)
5
-[(4-Iodophenoxy)methyl]-2,2-dimethyl-1,3-dioxan-5-
1
amine (60). To a solution of 59 (0.20 g, 0.51 mmol) in EtOH (2.0
to give the title compound (6.5 g, 82% yield) as a brown oil. H
mL) and water (1.0 mL) were added ammonium chloride (41 mg,
NMR (400 MHz, CDCl ) δ 6.99−7.05 (m, 1H), 6.87 (br s, 1H),
3
0.77 mmol) and iron powder (0.17 g, 3.1 mmol). After being stirred
6.25−6.40 (m, 1H), 5.31−5.51 (m, 2H), 3.81−3.98 (m, 1H), 5.06−
5.15 (m, 1H), 4.46−4.75 (m, 1H), 3.61−3.65 (m, 1H), 3.42−3.55
for 15 h at 70 °C, the reaction mixture was filtered through Celite.
The filtrate was concentrated in vacuo. The residue was diluted with
EtOAc and washed with saturated sodium bicarbonate. The organic
+
(m, 1H), 1.70−1.84 (m, 1H), 1.37−1.66 (m, 8H). MS (ESI ): m/z
+
302[M + H] . LCMS condition: B. RT = 0.415 min.
layer was dried over MgSO and concentrated in vacuo. The residue
was purified with amino-type silica gel column chromatography
2,2-Dimethyl-5-[[4-[2-[3-[[2-[(1S)-1-tetrahydropyran-2-
yloxyethyl]imidazol-1-yl]methyl]isoxazol-5-yl]ethynyl]-
phenoxy]methyl]-1,3-dioxan-5-amine (68). A round-bottomed
flask was charged with 67 (110 mg, 0.37 mmol), 60 (110 mg, 0.3
4
(
8
10−40% EtOAc in n-hexane) to afford the title compound (0.16 g,
1
8% yield) as a colorless solid. H NMR (400 MHz, CDCl ) δ 7.56
3
mmol), PdCl (PPh ) (11 mg, 0.015 mmol), and CuI (5.8 mg, 0.03
(d, J = 8.80 Hz, 2H), 6.71 (d, J = 8.80 Hz, 2H), 3.96 (s, 2H), 3.89
2
3 2
mmol). The flask was evacuated and backfilled with nitrogen
followed by addition of DMF (3.0 mL) and TEA (0.52 mL, 0.7
mmol). The mixture was placed into a preheated oil bath (50 °C)
and stirred for 30 min. The reaction mixture was diluted with
(
1
d, J = 11.62 Hz, 2H), 3.63 (d, J = 11.62 Hz, 2H), 1.47 (s, 3H),
.42 (s, 3H).
Ethyl 5-{[Tri(propan-2-yl)silyl]ethynyl}-1,2-oxazole-3-car-
boxylate (64). A round−bottomed flask was charged with 63
(18 g, 67 mmol), ethynyltri(propan-2-yl)silane (12 g, 67 mmol),
CHCl
hydroxide and brine. The organic phase was separated, dried over
MgSO , and concentrated under reduced pressure. The residue was
3
, followed by washing with 10% aqueous ammonium
DMF (52 mL), and TEA (73 mL, 1 mol, in excess). The flask was
evacuated and backfilled with nitrogen, followed by addition of
Pd(PPh ) (2.4 g, 2 mmol) and CuI (0.80 g, 4.2 mmol). After being
stirred for 1 h at 80 °C, the mixture was poured into a half-saturated
aqueous ammonium chloride and extracted three times with EtOAc-
hexane (1:1). The combined organic layer was washed with water,
dried over MgSO , and concentrated in vacuo. The residue was
purified with OH-type silica gel column chromatography (0−15%
EtOAc in n-hexane) to give the title compound (12 g, 69% yield) as
4
purified with amino-type silica gel chromatography (20−100%
3
4
EtOAc in n-hexane) to afford the title compound (90 mg, 55%
1
yield) as a yellow oil. H NMR (400 MHz, CDCl
) δ 7.46−7.54
3
(m, 2H), 6.99−7.08 (m, 1H), 6.88−6.98 (m, 3H), 6.19−6.34 (m,
1H), 5.32−5.49 (m, 2H), 5.13 (quin, J = 6.85 Hz, 1H), 4.47−4.77
(m, 1H), 4.04 (s, 2H), 3.43−3.95 (m, 6H), 1.37−1.89 (m, 15H).
MS (ESI ): m/z 537 [M + H] . LCMS condition: A. RT = 0.835
min.
4
+
+
1
a pale-yellow oil. H NMR (400 MHz, CDCl ) δ 6.81 (s, 1H), 4.45
3
(
q, J = 7.1 Hz, 2H), 1.42 (t, J = 7.1 Hz, 3H), 1.09−1.23 (m, 21H).
2-Amino-2-[(4-{[3-({2-[(1S)-1-hydroxyethyl]-1H-imidazol-1-
+
+
MS (ESI /APCI): m/z 322 [M + H] .
5-{[Tri(propan-2-yl)silyl]ethynyl}-1,2-oxazol-3-yl)methanol
65). To a cooled (0 °C) solution of 64 (10 g, 32 mmol) in EtOH
53 mL) and THF (103 mL) was added NaBH (2.4 g, 64 mmol)
yl}methyl)-1,2-oxazol-5-yl]ethynyl}phenoxy)methyl]propane-
1
,3-diol (43). To a solution of 68 (0.11 g, 0.20 mmol) in MeOH
(
(
(2.0 mL) and water (0.20 mL) was added p-TsOH monohydrate
(91 mg, 0.48 mmol) at RT. After being stirred for 20 h, the reaction
mixture was basified with saturated aqueous sodium bicarbonate at 0
°C and extracted with chloroform. The organics were combined and
passed through a phase separator and concentrated in vacuo. The
residue was purified with amino-type silica gel chromatography
(10% MeOH in chloroform) to give the title compound (61 mg,
(
4
in portions while keeping the internal temperature below 5 °C. After
stirring for 3 h at 0 °C, half-saturated aqueous ammonium chloride
was added to the reaction mixture while keeping the internal
temperature below 15 °C. The mixture was stirred for 30 min at
room temperature and concentrated under reduced pressure, and
the resulting aqueous solution was extracted three times with
75% yield) as a colorless amorphous solid.
1
CHCl . The combined organic layer was washed with brine, dried
H NMR (500 MHz, DMSO-d ) δ 7.57 (d, J = 9.17 Hz, 2H),
6
3
over MgSO , and concentrated in vacuo. The residue was purified
7.16 (d, J = 1.15 Hz, 1H), 7.03 (d, J = 8.79 Hz, 2H), 6.81 (d, J =
1.15 Hz, 1H), 6.79 (s, 1H), 5.38−5.47 (m, 3H), 4.85 (quin, J =
6.40 Hz, 1H), 4.61 (br s, 2H), 3.82 (s, 2H), 3.33−3.40 (m, 2H),
4
with OH-type silica gel column chromatography (10−50% EtOAc in
n-hexane) to give the title compound (7.9 g, 89% yield) as a pale-
1
yellow oil. H NMR (400 MHz, CDCl ) δ 6.49 (s, 1H), 4.77 (s,
2
1.46 (d, J = 6.88 Hz, 3H).
3
+
+
13
H), 1.02−1.25 (m, 21H). MS (ESI ): m/z 280 [M + H] . LCMS
condition: C. RT = 1.078 min.
-[(2-{(1S)-1-[(Oxan-2-yl)oxy]ethyl}-1H-imidazol-1-yl)-
C NMR (125 MHz, DMSO-d ) δ 161.6, 160.6, 152.9, 149.7,
6
133.6, 126.5, 120.9, 115.2, 111.1, 107.5, 99.0, 74.1, 69.6, 63.0, 61.4,
+
+
3
56.4, 40.5, 21.8. HRMS (ESI /APCI) calcd for [C21
H
24
N
4
O
5
+ H] :
methyl]-5-{[tri(propan-2-yl)silyl]ethynyl}-1,2-oxazole (66). To
a stirred solution of 65 (1.4 g, 4.8 mmol) and 62 (1.1 g, 5.8 mmol)
in THF (19 mL) were added TMAD (1.1 g, 6.3 mmol) and
tributylphosphine (1.3 mL, 6.3 mmol) at 0 °C under a nitrogen
atmosphere, and the mixture was stirred for 3.5 h at RT. The
reaction mixture was quenched by water and extracted three times
with EtOAc. The organic layer was dried over MgSO4 and
concentrated in vacuo onto ISOLUTE HM-N. The residue was
purified with OH-type silica gel column chromatography (40−80%
413.1819; found: 413.1800.
General Experimental Details for Assays and Structure
Determination. UDP-3-O-(R-3-hydroxydecanoyl)-N-acetylglucos-
amine was obtained from Alberta Research Council (Alberta,
Canada) and stored as 100mM stock dissolved in DMSO at −80
°C.
Fluorescamine was obtained from FUJIFILM Wako Pure
Chemical Corporation or Sigma-Aldrich.
Uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) was
stored as 200 μM stock dissolved in methanol−acetonitrile (1:1) at
−20 °C.
Compounds were stored as 10 mM, 20 mM, 50 mM, or 200 mM
stock in DMSO.
Protein Production. PaLpxC(1-299)C40S. Plasmids encoding P.
aeruginosa LpxC(1-299) with a mutation, C40S, were obtained from
ATUM, in the T7 expression vector pJ411 (ATUM).
MIKQRTLKNIIRATGVGLHSGEKVYLTLKPAPVDT-
GIVFSRTDLDPVVEIPARAENVGETTMSTTLVKGDVKVDT-
VEHLLSAMAGLGIDNAYVELSASEVPIMDGSAGPFVFLIQSAGL-
EtOAc in n-hexane) to afford the title compound 66 (1.5 g, 68%
1
yield) as a yellow oil. H NMR (600 MHz, CDCl ) δ 6.85−7.08
3
(
m, 2H), 6.16−6.35 (m, 1H), 5.29−5.50 (m, 2H), 5.06−5.19 (m,
1
1
H), 4.46−4.75 (m, 1H), 3.81−3.95 (m, 1H), 3.42−3.56 (m, 1H),
+
+
.05−1.85 (m, 30H). MS (ESI ): m/z 458 [M + H] . LCMS
condition: B. RT = 1.096 min.
-Ethynyl-3-[(2-{(1S)-1-[(oxan-2-yl)oxy]ethyl}-1H-imidazol-
-yl)methyl]-1,2-oxazole (67). To a cooled (0 °C) solution of 66
12 g, 26 mmol) in THF (0.18 L) was added TBAF (28 mL, 1
mol/L solution in THF), while the internal temperature was
5
1
(
L
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
QEQEAAKKFIRIKREVSVEEGDKRAVFVPFDGFKVS-
FEIDFDHPVFRGRTQQASVDFSSTSFVKEVSRARTFGFMR-
DIEYLRSQNLALGGSVENAIVVDENRVLNEDGLRYEDEFVKH-
KILDAIGDLYLLGNSLIGEFRGFKSGHALNNQLLRTLIADKDA-
WEVVTFEDARTAPISYMRP
Data Collection. Diffraction data of both the soaked and co-
crystals were collected at 100 K at either Diamond Light Source
(UK), Soleil Synchrotron (France), in-house on a Bruker D8
Venture TXS Generator with a Bruker Photon 100 detector, or in-
house on a Micro Max 007 HF Generator (Rigaku) with an R-AXIS
VII detector (Rigaku) or PILATUS 200K (Rigaku). Data processing
Plasmids were transformed into E. coli BL21(DE3) pLysS cells.
Cultures were grown at 37 °C in Terrific Broth with Kanamycin
25
was carried out with XDS , SAINT and SADABS (Bruker AXS
Inc., Madison, Wisconsin), CrystalClear (Rigaku) or CrysAlisPro
(Rigaku). Data collection statistics are summarized in Table S2.
Structure Determination. Crystallographic data was processed
(
100 μg/mL) until the OD600 was 0.7 at which time the
temperature was reduced to 30 °C and expression induced by
addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) to a final
concentration of 1 mM in the presence of 0.1 mM ZnCl or ZnSO .
2
6
and structures determined and refined within the CCP4 package
2
4
27
Cells were harvested by centrifugation after 2 h.
using Sketcher (CCP4) or ProDrg to build models of each
compound. The structures were determined by molecular replace-
ment using the published structure (PDBcode 3UHM) to calculate
For 15N-labeled PaLpxC, the transformed cells were grown in
EnpressoB, Nitrogen Free (Enpresso), supplemented with 2.5 g/L
15
28
N-(NH ) SO (Goss Scientific) induced as above and harvested
model phases and subsequently refined using REFMAC5.
4
2
4
after 3 h.
Interactive graphical model building was carried out with
29
Cells were resuspended in 3 volumes of 20 mM 4-(2-
hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.0)
and 2 mM DTT and lysed by passing three times through a
Stanstead Homogenizer (Stanstead). The lysate was clarified by
centrifugation and applied to CaptoQ, eluting in a gradient to 1 M
NaCl. (NH ) SO was added to a final concentration of 1 M, and
the LpxC was further purified by hydrophobic interaction
chromatography using HiTrap Butyl HP (GE Healthcare) with a
gradient of 1−0 M (NH ) SO and finally Superdex200pg (GE
Healthcare), equilibrated in 20 mM HEPES (pH 7.4), 150 mM
NaCl, and 2 mM DTT.
PaLpxC(1-303-His). The gene fragment encoding residues 2-303
of PaLpxC was amplified using primers with NdeI and NotI sites
and subcloned into pET21a vectors. The protein was expressed in E.
coli BL21(DE3)pLysS cells by inducing with 1.0 mM IPTG and 0.2
mM ZnSO at OD = 0.5 for 7 h at 18 °C. The protein was
purified using TALON metal affinity resin (TaKaRa) and buffer-
exchanged against 40 mM HEPES (pH 8.0) using PD-10 (GE
Healthcare).
COOT. In all structures, the respective ligands were clearly
defined by the initial electron density maps; Table S2 includes an
image of the 2Fo-Fc map around each ligand.
NMR Methods. All protein NMR spectra were recorded on a 600
MHz Bruker Avance III spectrometer fitted with a cryoprobe.
1
15
15
H− N HSQC NMR spectra were recorded with N-labeled
PaLpxC at 250 μM in 20 mM HEPES (pH 7.4), 150 mM NaCl,
and 5% D O.
4
2
4
2
Ligand-observed fragment screening was performed with 500 μM
of each compound, 10 μM PaLpxC in 20 mM HEPES (pH 7.4)
and, 15 μM NaCl, in mixtures of six compounds per sample as
4
2
4
17,18
described previously.
Computational Methods. Compounds and proteins were
modeled using the Schrodinger Suite of programs (mmshare-
v32018:mmshare-v3.5, Schrodinger, LLC, New York, NY). Protein
̈
̈
structure preparation (PrepWizard), ligand preparation (Ligprep),
and generation of docking grids were run with defaults. Docking was
carried out using Glide, with and without constraints in the
positions of the atoms chelating the zinc.
4
600
TSA Methods. Fluorescence was measured on an Agilent
MX300P Real-Time PCR with a heating rate of 1 °C per minute
and Ex/Em 492/610 nm. Final concentrations used were 5 μM
PaLpxC, 5× Sypro Orange, 40 mM HEPES (pH 8.0), and 2%
DMSO in a final volume of 40 μL. Protein denaturation curves were
measured in the absence and presence of low nM affinity
compounds 1 (PF-5081090), which were added at a 10-fold excess
of the protein concentration.
Calculations of pKa were performed with the ACD/Percepta
software (ACD/Labs 2017.13, Advanced Chemistry Development,
Inc) packaged in Classic mode, not using GALAS.
Fluorescence Polarization (FP) Assay. Compounds were
screened in a 50 μL assay containing 20 nM PaLpxC or
PaLpxC(1-303-His), 5 nM fluorescein probe 5, 40 mM HEPES
(pH 7.5), 150 mM NaCl, 0.02% Tween 20, with or without 25 nM
ZnCl , and 5% DMSO. PaLpxC and compounds were preincubated
2
When evaluating compounds, fluorescence was measured on an
Applied Biosystems 7500 Fast Real-Time PCR system with a
heating rate of 0.96 °C per minute (ramp rate: 1%) and the ROX
for 30 minutes prior to addition of the fluorescein probe, and the
assay was then incubated for a further 2 h at room temperature.
Plates were read on either a Biotek Synergy Neo Plate reader Ex/
Em 485(20)/528(20) nm, Envision 2101 (Ex/Em 480/535 nm), or
PHERAStar FSX (Ex/Em 485/535 nm) in conjunction with
polarizing filters.
(
carboxy-X-rhodamine) detector. Final concentrations used were 5
μM PaLpxC(1-303-His), 1× Protein Thermal Shift Dye (Life
Technologies), 40 mM HEPES (pH 8.0), 10 or 40 μM ZnCl , and
2
5
% DMSO in a final volume of 20 μL. Protein denaturation curves
Inhibition values were calculated from control DMSO samples
containing no compound (0% inhibition) and samples containing no
enzyme (100% inhibition). Data reported in the tables is the
geometric mean of at least two determinations; compound 69 or
BB-78485 was included in assays as a control with IC50 values in
the range 0.10−0.40 and 0.015−0.060 μM, respectively. The
synthesis of 69 is described in the Supporting Information.
Fluorescamine-Based Functional Activity Assay. Imidazole series
compounds were assayed with 3.6 nM PaLpxC(1-303-His), 20 μM
UDP-3-O-(R-3-hydroxydecanoyl)-N-acetylglucosamine (substrate),
were measured in the absence and presence of compounds which
were added at a more than 10-fold excess of the protein
concentration.
Crystallization and Structure Determination by X-ray Crystal-
lography. Co-Crystallization to Establish Initial Crystals. PaLpxC
at 10.6 mg/mL was mixed with compound 47 (see SI) at 4-fold
molar excess. The complex crystallized in 0.1 M Tris (pH 8.5), 30%
24
poly(ethylene glycol) 4 K, and 0.2 M MgCl using the sitting-drop
2
vapor-diffusion method at 19 °C. The co-crystals took 2 weeks to
grow. The compound binds in the tunnel region, making no contact
with zinc.
40 mM HEPES (pH 8.0), 25 nM ZnCl , 0.02% Brij-35, and 5%
2
DMSO. Assay plates were incubated for 60 min at room
temperature and terminated by the addition of 1.0 mg/mL
fluorescamine in 1:1 dimethylformamide/acetonitrile followed by
0.2 M sodium phosphate buffer (pH 8.0) and incubated. Plates were
read on an Enspire: Ex/Em 390/495 nm.
Protein−Ligand Complexes Containing 4, 5, and 24. Co-
crystals containing 47 were soaked out and compounds 4, 5, or 24
subsequently soaked in, also overnight. The protein crystals were
flash-frozen at 100 K using a cryo-protection buffer consisting of
mother liquor with 20−25% glycerol added.
Protein−Ligand Complexes Containing Other Compounds.
Details of the conditions under which crystals were obtained by
co-crystallization of individual compounds are in SI.
Inhibition values were calculated from control DMSO samples
containing no compound (0% inhibition) and samples containing no
PaLpxC(1-303-His) (100% inhibition); compound 69 or BB-
24
78485 was included in assays with IC50 values in the range
M
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
0
.30−1.2 and 0.010−0.040 μM, respectively. Data reported in tables
Roderick E. Hubbard − Vernalis (R&D) Ltd., Cambridge
CB21 6GB, U.K.; Email: r.hubbard@vernalis.com
are the geometric mean of at least two determinations.
This assay measures the extent to which LpxC catalyzes the
removal of the acetyl group from the acetylglucosamine by the
reaction of the resulting free amino group with fluorescamine. It is
therefore not suitable for the glycine series.
Authors
Matrix Metalloprotease (MMP) Assays. The MMP inhibitor
profiling kit, Fluorometric RED (Enzo Life Sciences, Inc., Farm-
ingdale, NY), was used for the human MMP inhibition assays. The
inhibitory activities of compounds against MMP-2, 3, and 9 were
evaluated as described in the kit protocol with some modifications.
In brief the assay volume was modified to 50 μL and final enzyme
concentrations of MMP-2, 3, and 9 were modified to 18.3, 21.0, and
Hajime Takashima − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
David Lee Walmsley − Vernalis (R&D) Ltd., Cambridge
CB21 6GB, U.K.
Fumihito Ushiyama − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
5
.67 mU/μL, respectively. Fluorescence (Ex/Em 545/576 nm) was
detected using the CLARIOstar (BMG Labtech). Ilomastat (GM-
30
Yohei Matsuda − Taisho Pharmaceutical Co., Ltd., Saitama
6001) is an MMP inhibitor used as a positive control. Values
331-9530, Japan
reported are the geometric mean of two determinations.
Mass-Spectrometry-Based Functional Activity Assay. Glycine
series compounds were screened with 3.6 nM PaLpxC(1-303-His),
Harumi Kanazawa − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Toru Yamaguchi-Sasaki − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Nozomi Tanaka-Yamamoto − Taisho Pharmaceutical Co.,
Ltd., Saitama 331-9530, Japan
Junya Yamagishi − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Risa Kurimoto-Tsuruta − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Yuya Ogata − Taisho Pharmaceutical Co., Ltd., Saitama
331-9530, Japan
Norikazu Ohtake − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
20 μM UDP-3-O-(R-3-hydroxydecanoyl)-N-acetylglucosamine (sub-
strate), 40 mM HEPES (pH 8.0), 25 nM ZnCl , 0.02% Brij-35, and
2
5
% DMSO. Samples were incubated for 60 min at room
temperature, and the reaction was terminated by the addition of
four times the volume of acetonitrile containing 25 μM UDP-
GlcNAc as the internal standard. The mixture was centrifuged, and
the supernatant was injected into liquid chromatography-tandem
mass spectrometry (LC/MS/MS). The samples were separated with
Inertsil Amide (3.0 μm, 50 mm × 2.1 mm I.D., GLScience, Japan)
with a Shimadzu HPLC system. The mobile phase was 8 mM
ammonium acetate containing 72% acetonitrile, and the flow rate
was 0.2 mL/min. MS/MS detection of each component (m/z 734
and m/z 385) was performed using a TSQ Quantum system
(Thermo Fisher Scientific) or a QTRAP-5500 system (Sciex) with
Hayley Angove − Vernalis (R&D) Ltd., Cambridge CB21
the electrospray interface in negative ion detection mode.
Inhibition values were calculated from product signals of control
DMSO samples containing no compound (0% inhibition) and
samples containing no substrate (100% inhibition). The control
compound 69 was included in assays with IC50 values in the range
of 0.25−1.0 μM. Data reported in tables are the geometric mean of
at least two determinations.
6GB, U.K.
Lisa Baker − Vernalis (R&D) Ltd., Cambridge CB21 6GB,
U.K.
Richard Harris − Vernalis (R&D) Ltd., Cambridge CB21
6GB, U.K.
Alba Macias − Vernalis (R&D) Ltd., Cambridge CB21 6GB,
U.K.
Evaluation of Antibacterial Activity. The minimum inhibitory
concentration (MIC) was determined using the Clinical and
Laboratory Standards Institute (CLSI) methodology, described in
Alan Robertson − Vernalis (R&D) Ltd., Cambridge CB21
6GB, U.K.
31
a CLSI document. P. aeruginosa 5567 was a clinical isolate
obtained in Japan. Where used, the PAβN concentration was 50 μg/
mL for P. aeruginosa and 200 μg/mL for E. coli; the HSA
concentration was 5% w/v.
Allan Surgenor − Vernalis (R&D) Ltd., Cambridge CB21
6GB, U.K.
Hayato Watanabe − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Koichiro Nakano − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
ASSOCIATED CONTENT
sı Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jmedchem.0c01215.
■
*
Masashi Mima − Taisho Pharmaceutical Co., Ltd., Saitama
331-9530, Japan
Kunihiko Iwamoto − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan
Synthesis of compounds 9−24 and 26−32; structural
biology data (PDF)
Molecular formula strings (CSV)
Atsushi Okada − Taisho Pharmaceutical Co., Ltd., Saitama
331-9530, Japan
Iichiro Takata − Taisho Pharmaceutical Co., Ltd., Saitama
31-9530, Japan
Kosuke Hitaka − Taisho Pharmaceutical Co., Ltd., Saitama
31-9530, Japan
Akihiro Tanaka − Taisho Pharmaceutical Co., Ltd., Saitama
31-9530, Japan
Kiyoko Fujita − Taisho Pharmaceutical Co., Ltd., Saitama
31-9530, Japan
Accession Codes
3
The refined coordinates of the structures presented in this
manuscript are deposited at the RCSB with the following
codes: 6: 7CIC, 7: 7CIB, 8: 7CIA, 10: 7CIE, 14: 7CI4, 16:
3
7
7
CI5, 22: 7CI7, 27: 7CID, 29: 7CI6, 38: 7CI8, and 43:
CI9. The authors will release the atomic coordinates and
3
experimental data upon article publication
3
Hiroyuki Sugiyama − Taisho Pharmaceutical Co., Ltd.,
AUTHOR INFORMATION
Corresponding Authors
Yousuke Yamada − Taisho Pharmaceutical Co., Ltd.,
Saitama 331-9530, Japan; Email: yo-yamada@taisho.co.jp
■
Saitama 331-9530, Japan
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jmedchem.0c01215
N
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Author Contributions
(10) Shen, S.; Kozikowski, A. P. Why hydroxamates may not be
the best histone deacetylase inhibitorswhat some may have
forgotten or would rather forget? ChemMedChem 2016, 11, 15−21.
The manuscript was written by R.E.H., D.L.W., and Y.Y. with
contributions from other authors. All authors have given
approval to the final version of the manuscript.
(11) Vandenbroucke, R. E.; Libert, C. Is there new hope for
therapeutic matrix metalloproteinase inhibition? Nat. Rev. Drug
Discovery 2014, 13, 904−927.
Notes
The authors declare no competing financial interest.
(12) Montgomery, J. I.; Brown, M. F.; Reilly, U.; Price, L. M.;
Abramite, J. A.; Arcari, J.; Barham, R.; Che, Y.; Chen, J. M.; Chung,
S. W.; Collantes, E. M.; Desbonnet, C.; Doroski, M.; Doty, J.;
Engtrakul, J. J.; Harris, T. M.; Huband, M.; Knafels, J. D.; Leach, K.
L.; Liu, S.; Marfat, A.; McAllister, L.; McElroy, E.; Menard, C. A.;
Mitton-Fry, M.; Mullins, L.; Noe, M. C.; O’Donnell, J.; Oliver, R.;
Penzien, J.; Plummer, M.; Shanmugasundaram, V.; Thoma, C.;
Tomaras, A. P.; Uccello, D. P.; Vaz, A.; Wishka, D. G. Pyridone
methylsulfone hydroxamate Lpxc inhibitors for the treatment of
serious gram-negative infections. J. Med. Chem. 2012, 55, 1662−
ACKNOWLEDGMENTS
■
The authors thank Kanako Hatanaka, Ryo Takahashi, Maho
Honda-Yamagishi, Takanori Kawaguchi, Mitsukane Yoshina-
ga, Takashi Yoshizumi, Shingo Sekine, Shinichi Nishimoto,
Takehiko Kato, and Tomoyuki Hata of Taisho Pharmaceut-
icals for technical assistance; Sean McKenna and Yikang
Wang of Vernalis for support in synthesis; and the staff at
Diamond Light Source and Soleil synchrotrons for help with
crystallographic data collection.
1
(
670.
13) Clayton, G. M.; Klein, D. J.; Rickert, K. W.; Patel, S. B.;
Kornienko, M.; Zugay-Murphy, J.; Reid, J. C.; Tummala, S.; Sharma,
S.; Singh, S. B.; Miesel, L.; Lumb, K. J.; Soisson, S. M. Structure of
the bacterial deacetylase Lpxc bound to the nucleotide reaction
product reveals mechanisms of oxyanion stabilization and proton
transfer. J. Biol. Chem. 2013, 288, 34073−34080.
ABBREVIATIONS
■
LpxC, UDP-3-O-acyl-N-acetylglucosamine deacetylase; MIC,
minimum inhibitory concentration; PAβN, phenylalanine-
arginine β-naphthylamide; UDP-GlcNAc, uridine diphosphate
N-acetylglucosamine; IPTG, isopropyl β-D-1-thiogalactopyr-
anoside; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesul-
fonic acid; RT, retention time; MMP, matrix metalloprotease;
(14) Zhang, J.; Chan, A.; Lippa, B.; Cross, J. B.; Liu, C.; Yin, N.;
Romero, J. A.; Lawrence, J.; Heney, R.; Herradura, P.; Goss, J.;
Clark, C.; Abel, C.; Zhang, Y.; Poutsiaka, K. M.; Epie, F.; Conrad,
M.; Mahamoon, A.; Nguyen, K.; Chavan, A.; Clark, E.; Li, T. C.;
Cheng, R. K.; Wood, M.; Andersen, O. A.; Brooks, M.; Kwong, J.;
Barker, J.; Parr, I. B.; Gu, Y.; Ryan, M. D.; Coleman, S.; Metcalf, C.
A., 3rd Structure-based discovery of Lpxc inhibitors. Bioorg. Med.
Chem. Lett. 2017, 27, 1670−1680.
boc, tert-butoxycarbonyl; (BOC) O, di-tert-butyl dicarbonate;
2
HATU, O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluro-
nium hexafluorophosphate; TBAF, tetra-n-butylammonium
fluoride; THP, tetrahydropyranyl; TMAD, 3-(dimethylcarba-
moylimino)-1,1-dimethylurea; TIPS, triisopropylsilyl
(15) Baurin, N.; Aboul-Ela, F.; Barril, X.; Davis, B.; Drysdale, M.;
Dymock, B.; Finch, H.; Fromont, C.; Richardson, C.; Simmonite,
H.; Hubbard, R. E. Design and characterization of libraries of
molecular fragments for use in nmr screening against protein targets.
J. Chem. Inf. Comput. Sci. 2004, 44, 2157−2166.
REFERENCES
■
(16) Chen, I. J.; Hubbard, R. E. Lessons for fragment library
(
1) Hofer, U. The cost of antimicrobial resistance. Nat. Rev.
Microbiol. 2019, 17, 3.
design: Analysis of output from multiple screening campaigns. J.
Comput. Aided Mol. Des. 2009, 23, 603−620.
(
2) Brown, E. D.; Wright, G. D. Antibacterial drug discovery in the
resistance era. Nature 2016, 529, 336−343.
3) Chen, A. Y.; Adamek, R. N.; Dick, B. L.; Credille, C. V.;
(17) Hubbard, R. E.; Davis, B.; Chen, I.; Drysdale, M. J. The seeds
approach: Integrating fragments into drug discovery. Curr. Top Med.
Chem. 2007, 7, 1568−1581.
(
Morrison, C. N.; Cohen, S. M. Targeting metalloenzymes for
(18) Hubbard, R. E.; Murray, J. B. Experiences in fragment-based
therapeutic intervention. Chem. Rev. 2019, 119, 1323−1455.
lead discovery. Methods Enzymol. 2011, 493, 509−531.
(4) Erwin, A. L. Antibacterial drug discovery targeting the
(19) Credille, C. V.; Dick, B. L.; Morrison, C. N.; Stokes, R. W.;
lipopolysaccharide biosynthetic enzyme Lpxc. Cold Spring Harb.
Perspect. Med. 2016, 6, No. a025304.
Adamek, R. N.; Wu, N. C.; Wilson, I. A.; Cohen, S. M. Structure−
activity relationships in metal-binding pharmacophores for influenza
endonuclease. J. Med. Chem. 2018, 61, 10206−10217.
(
2
(
5) Kalinin, D. V.; Holl, R. Lpxc inhibitors: A patent review (2010-
016). Expert Opin. Ther. Patents 2017, 27, 1227−1250.
6) Raetz, C. R.; Reynolds, C. M.; Trent, M. S.; Bishop, R. E.
(20) Lamoree, B.; Hubbard, R. E. Current perspectives in
Lipid a modification systems in gram-negative bacteria. Annu. Rev.
Biochem. 2007, 76, 295−329.
fragment-based lead discovery (fbld). Essays Biochem. 2017, 61,
453−464.
(7) Raetz, C. R. H.; Whitfield, C. Lipopolysaccharide endotoxins.
(21) Erlanson, D. A.; Fesik, S. W.; Hubbard, R. E.; Jahnke, W.;
Jhoti, H. Twenty years on: The impact of fragments on drug
discovery. Nat. Rev. Drug Discovery 2016, 15, 605−619.
(22) Hubbard, R. E. Fragment Based Lead Discovery, In The
Handbook of Medicinal Chemistry: Principles and Practice; The Royal
Society of Chemistry, 2015; pp 122−153.
Annu. Rev. Biochem. 2002, 71, 635−700.
(8) Krause, K. M.; Haglund, C. M.; Hebner, C.; Serio, A. W.; Lee,
G.; Nieto, V.; Cohen, F.; Kane, T. R.; Machajewski, T. D.;
Hildebrandt, D.; Pillar, C.; Thwaites, M.; Hall, D.; Miesel, L.;
Hackel, M.; Burek, A.; Andrews, L. D.; Armstrong, E.; Swem, L.;
Jubb, A.; Cirz, R. T. Potent Lpxc inhibitors with in vitro activity
against multidrug-resistant pseudomonas aeruginosa. Antimicrob.
Agents Chemother. 2019, 63, e00977.
(23) Hubbard, R. E. The Role of Fragment-Based Discovery in
Lead Finding, In Fragment-Based Drug Discovery Lessons and Outlook,
Erlanson, D. A.; Jahnke, W., Eds.; Wiley, 2016; pp 1−36.
(24) Clements, J. M.; Coignard, F.; Johnson, I.; Chandler, S.;
Palan, S.; Waller, A.; Wijkmans, J.; Hunter, M. G. Antibacterial
activities and characterization of novel inhibitors of Lpxc. Antimicrob.
Agents Chemother. 2002, 46, 1793−1799.
(25) Kabsch, W. Xds. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2010, 66, 125−132.
(26) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.;
Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A.
G. W.; McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N.
(9) Cohen, F.; Aggen, J. B.; Andrews, L. D.; Assar, Z.; Boggs, J.;
Choi, T.; Dozzo, P.; Easterday, A. N.; Haglund, C. M.; Hildebrandt,
D. J.; Holt, M. C.; Joly, K.; Jubb, A.; Kamal, Z.; Kane, T. R.;
Konradi, A. W.; Krause, K. M.; Linsell, M. S.; Machajewski, T. D.;
Miroshnikova, O.; Moser, H. E.; Nieto, V.; Phan, T.; Plato, C.;
Serio, A. W.; Seroogy, J.; Shakhmin, A.; Stein, A. J.; Sun, A. D.;
Sviridov, S.; Wang, Z.; Wlasichuk, K.; Yang, W.; Zhou, X.; Zhu, H.;
Cirz, R. T. Optimization of Lpxc inhibitors for antibacterial activity
and cardiovascular safety. ChemMedChem 2019, 14, 1560−1572.
O
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
S.; Potterton, E. A.; Powell, H. R.; Read, R. J.; Vagin, A.; Wilson, K.
S. Overview of the ccp4 suite and current developments. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 235−242.
(
27) Schuttelkopf, A. W.; van Aalten, D. M. F. Prodrg: A tool for
̈
high-throughput crystallography of protein-ligand complexes. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 1355−1363.
(
28) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 1997, 53, 240−255.
(29) Emsley, P.; Cowtan, K. Coot: Model-building tools for
molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,
0, 2126−2132.
30) Levy, D. E.; Lapierre, F.; Liang, W.; Ye, W.; Lange, C. W.; Li,
6
(
X.; Grobelny, D.; Casabonne, M.; Tyrrell, D.; Holme, K.; Nadzan,
A.; Galardy, R. E. Matrix metalloproteinase inhibitors: A structure-
activity study. J. Med. Chem. 1998, 41, 199−223.
(31) CLSI. Performance Standards for Antimicrobial Susceptibility
Testing. In CLSI Supplement M100-S26, 27th ed.; Wayne, P. A., Ed.;
Clinical and Laboratory Standards Institute, 2017.
P
https://dx.doi.org/10.1021/acs.jmedchem.0c01215
J. Med. Chem. XXXX, XXX, XXX−XXX