Antibacterial inhibitors of gram-positive thymidylate kinase: Structure-activity relationships and chiral preference of a new hydrophobic binding region

Article abstract of DOI:10.1021/jm500463c

Thymidylate kinase (TMK), an essential enzyme in bacterial DNA biosynthesis, is an attractive therapeutic target for the development of novel antibacterial agents, and we continue to explore TMK inhibitors with improved potency, protein binding, and pharmacokinetic potential. A structure-guided design approach was employed to exploit a previously unexplored region in Staphylococcus aureus TMK via novel interactions. These efforts produced compound 39, with 3 nM IC₅₀ against S. aureus TMK and 2 μg/mL MIC against methicillin-resistant S. aureus (MRSA). This compound exhibits a striking inverted chiral preference for binding relative to earlier compounds and also has improved physical properties and pharmacokinetics over previously published compounds. An example of this new series was efficacious in a murine S. aureus infection model, suggesting that compounds like 39 are options for further work toward a new Gram-positive antibiotic by maintaining a balance of microbiological potency, low clearance, and low protein binding that can result in lower efficacious doses.

Full text of DOI:10.1021/jm500463c

Article
pubs.acs.org/jmc
Antibacterial Inhibitors of Gram-Positive Thymidylate Kinase:
Structure−Activity Relationships and Chiral Preference of a New
Hydrophobic Binding Region
Sameer P. Kawatkar,*,†,§ Thomas A. Keating,†,∥ Nelson B. Olivier,‡ John N. Breen,†,‡ Oluyinka M. Green,†
Satenig Y. Guler,† Martin F. Hentemann,†,⊥ James T. Loch,† Andrew R. McKenzie,† Joseph V. Newman,†
Linda G. Otterson,† and Gabriel Martínez-Botella†,#
†Infection Innovative Medicines and ‡Discovery Sciences, AstraZeneca, 35 Gatehouse Drive, Waltham, Massachusetts 02451, United
States
S
* Supporting Information
ABSTRACT: Thymidylate kinase (TMK), an essential enzyme
in bacterial DNA biosynthesis, is an attractive therapeutic target
for the development of novel antibacterial agents, and we
continue to explore TMK inhibitors with improved potency,
protein binding, and pharmacokinetic potential. A structure-
guided design approach was employed to exploit a previously
unexplored region in Staphylococcus aureus TMK via novel
interactions. These efforts produced compound 39, with 3 nM
IC50 against S. aureus TMK and 2 μg/mL MIC against
methicillin-resistant S. aureus (MRSA). This compound exhibits
a striking inverted chiral preference for binding relative to earlier
compounds and also has improved physical properties and
pharmacokinetics over previously published compounds. An
example of this new series was efficacious in a murine S. aureus infection model, suggesting that compounds like 39 are
options for further work toward a new Gram-positive antibiotic by maintaining a balance of microbiological potency, low
clearance, and low protein binding that can result in lower efficacious doses.
high single dose of 150 mg/kg to achieve infection stasis in the
mouse. As we continue to explore inhibitors of Gram-positive
TMK, we aim for overall improvements in potency, protein
binding, and/or clearance through specific design and testing of
new compounds.
INTRODUCTION
■
Resistance to antibiotic agents represents one of the most
pressing challenges to the infectious disease community1,2 as it
has rendered marketed antibiotics less effective against many
common bacterial pathogens. The problem of widespread and
increasing antibiotic resistance is only getting worse with the
dearth of novel classes of antibiotics, creating a significant
unmet medical need in the treatment of serious bacterial
infections.
One of the strategies to effectively tackle the problem of
antibacterial resistance is to develop agents that have a novel
mode of action or act through a novel target. In this respect,
bacterial thymidylate kinase (TMK), which catalyzes the
phosphorylation of dTMP to dTDP using ATP as a
cosubstrate, represents a very attractive therapeutic target that
is essential for bacterial survival.3−5 This is also an area of active
antiviral research.6 Recently, we described the discovery of a
potent and specific inhibitor of TMK (1, TK-666; Figure 1)
from a thymine-based lead (2).7 Compound 1 has excellent
solubility, moderate clearance, broad-spectrum Gram-positive
TMK enzyme inhibition, and rapid bactericidal activity. It was
also efficacious in a murine model of Staphylococcus aureus
infection.4 However, the series was also highly protein bound,
and this, coupled with moderate clearance, led to a relatively
Herein, we describe the design, synthesis, and structure−
activity relationships for a separate chemical series derived from
the core scaffold of 2 by targeting a previously unexplored
region adjacent to ring C (Figure 1B). This new series is
distinct from previous compounds in three ways. First, X-ray
crystal structures of bound complexes showed newly formed
interactions, as well as the steric limits of the functional groups
that could be added to the inhibitor scaffold at position C-2.
Second, selected examples also possessed significantly reduced
protein binding compared to previously reported compounds.
Last, and most strikingly, the lead compounds reported here
also have a different diastereomeric preference for optimal
enzyme potency as compared to the earlier TMK inhibitor
series. These features offer new avenues for exploration of
quality inhibitors that can be optimized for clinical candidacy.
In profiling, this new series yielded TMK inhibitors with
Received: December 19, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 1. Activity of C-2-Substituted Analogues of 7 and 8
compd
R
R1
R2
S. aureus TMK IC50 (nM)
7
8
H
F
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
1170
2060
9260
67400
474
H
9
F
Me
Pr
10
11
12
13
14
15
16
F
F
OMe
OEt
F
F
625
H
F
86
Cl
779
F
Br
581
H
CN
1190
a
R2, the C-2 substituent, is highlighted in blue.
Table 2. Activity of Carboxylic Acid Analogues of C-2-
a
Methoxy- and C-2-Fluoro-Substituted Compounds
Figure 1. (A) Structures of thymidylate kinase (TMK) inhibitor 1
(TK-666), an efficacious inhibitor of Gram-positive TMK, and lead
compound 2.4,7 The rings are lettered from A to D for the discussion
in the text. Compounds with C-2 substitution are the focus of this
work. (B) X-ray structure of 2 in complex with S. aureus TMK. The
hydrophobic patch formed by conserved residues Val51, Leu52, and
Phe66 was targeted with C-2 substitutions (right panel). In particular,
additional hydrophobic interactions (left arrow) and novel dipole−
dipole/C−H···O interactions with Phe66 (right arrow) were sought.
b
S. aureus TMK IC50
Sau/SauMRQR MIC
(μg/mL)
compd
R2
(nM)
log D
17
18
19
H
73
10
12
32/32
8/16
−0.55
−0.77
−0.59
OMe
F
32/16
b
a
R2, the C-2 substituent, is highlighted in blue. Bacterial strains from
AstraZeneca collection: S. aureus ARC516/S. aureus ARC517
methicillin-resistant (MR) and quinolone-resistant (QR).
excellent activity against Gram-positive TMK, efficacy in a S.
aureus murine thigh infection model, and somewhat improved
pharmacokinetics versus compounds reported earlier.7
CHEMISTRY
introduced in the initial step of the synthesis by metalation and
addition of an aldehyde to 1,4-dibromo-2,3-difluorobenzene
(27). Mesylate formation from the resulting alcohol 28,
followed by alkylation of amine 6, yielded the bromide
intermediate 30, which was converted to the corresponding
cyano intermediate 31. Activation of the C-3 fluoride by the
adjacent cyano group in 31 allowed the introduction of ring D
in 32 by microwave-mediated etherification. Standard basic
hydrolysis of the nitrile provided the carboxylic acid 33. Finally,
single diastereomers of analogues in Table 2 were obtained by
chiral separation. While the individual chemical steps described
here are not optimized for yield, all yields were sufficient to
obtain the desired compounds in amounts suitable for the
profiling described.
■
The overall approach to prepare the analogues in Tables 1 and
2 required the formation of the carbon−nitrogen bond linking
rings B and C.7,8 The key chiral amine 6 and the appropriate
aldehyde or mesylate were the main synthetic targets (Schemes
1−3). The chiral amine 6 has previously been described.7,8 For
the analogues in Table 1, variation of the C-2 substituent was
the main goal, and those compounds were readily accessed by
starting with commercially available 3-methoxybenzaldehyde 3
and then by installing ring D via copper-mediated coupling of
phenylboronic acids (Scheme 1). The final compound was
assembled by standard reductive amination with amine 6. The
approach to install the carboxylic acid in ring C for both
methoxy and fluoride analogues was analogous (Schemes 2 and
3), involving a palladium-mediated formation of the cyano-
phenyl ring using zinc cyanide and the corresponding
bromophenyl intermediate 24 or 30. Standard alkaline
hydrolysis provided the carboxylic acids in good yields. For
the C-2 methoxy analogues, the installation of ring A and the
alkyl side chain on the carbon linking rings B and C was
analogous to the routes described by us previously7,8 (Scheme
2). In this instance, the route was modified to prepare the C-2
fluoro analogues (Scheme 3). First, the aliphatic side chain was
RESULTS AND DISCUSSION
■
As disclosed previously,7 our efforts to design inhibitors of
Gram-positive TMK originated in thymine-based lead com-
pound 2, which showed moderate enzymatic activity against S.
aureus thymidylate kinase (Figure 1). Further exploration of
ring D substituents led to compounds 7 and 8 with improved
biochemical potency against S. aureus TMK (Table 1). The X-
ray structure7 of 2 bound to S. aureus TMK (Figure 1) showed
B
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Compound 13
a
Reagents and conditions: (a) HBr, 80 °C, 45 min, 46%; (b) 3-chlorophenylboronic acid, Cu(OAc)2, DEA, 3 Å molecular sieves, rt, 72 h, 48%; (c)
polymer-supported cyanoborohydride, DMF/AcOH (18:1), rt, 18 h, 35%.
a
Scheme 2. Synthesis of Compounds 38 and 39
a
Reagents and conditions: (a) 3-chlorophenylboronic acid, Cu(OAc)2, pyridine, 1,2-DCE, 72 h, 55 °C, 35%; (b) EtMgBr, Et2O, 2 h, −78 °C, 36%;
(c) MsCl, Et3N, DCM, 2 h, 0 °C, >95%; (d) 6, ACN, 12−36 h, 80 °C, 16%; (e) Zn(CN)2, Pd(PPh3)4, DMF, 45 min, microwave, 110 °C; (f)
NaOH, EtOH, 12 h, reflux, 31% (two steps); (g) chiral separation, 30% each.
the thymine moiety (ring A) making hydrogen bond
interactions with residues Arg70 and Gln101 and π−π stacking
interactions with Phe66. Ring D also makes an edge-to-face
interaction with Phe66. The phenyl (ring C) and piperidine
(ring B) rings did not appear to make any specific interactions
with the enzyme and instead executed a U-turn to properly
position rings A and D for their direct interactions.
Analysis of the X-ray structure also revealed an additional
small hydrophobic region formed by Val51 and Leu52, close to
the C-2 carbon of ring C (Figure 1B). We hypothesized that
small aliphatic residues could form favorable, buried hydro-
phobic interactions with these side chains. In addition to
targeting this hydrophobic patch with aliphatic substituents, the
proximity of Phe66 to C-2 also prompted us to evaluate alkoxy
and halogen substituents for potential interactions with Phe66.
We thus embarked on a design and modeling effort to explore
this opportunity. In the design and testing phases, we paid close
attention to log D as this parameter was found to be critical for
S. aureus microbiological activity.7
Molecular modeling of C-2 alkyl substituents suggested that
small, aliphatic moieties could positively interact with the side
chains of Val51 and Leu52 without altering the position of the
remainder of the inhibitor. Many studies have shown that
hydrophobic interactions play a crucial role in stabilizing
protein structure,9,10 protein−protein interactions,11 and
protein−ligand complexes.12−14 While increasing hydrophobic-
C
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 3. Synthesis of Compounds 46 and 47
a
iPrMgCl−LiCl, 3-methylbutanal, THF, −20 °C, 100%; (b) MsCl, TEA, DCM; (c) DIEA, MeCN, 80 °C, 16% (for two steps); (d) Zn(CN)2,
Pd(PPh3)4, 140 °C, microwave, 67%; (e) 3-chlorophenol, K2CO3, 160 °C, microwave, 71%; (f) NaOH, MeOH, H2O, 80 °C, 62%; (g) chiral
separation, 38% each.
ity too much can be deleterious to solubility and protein
binding and increase off-target effects, our TMK inhibitors are
generally extremely soluble and very specific for the target.4,7
Past efforts required the addition of hydrophobic groups to
moderate a very low log D and gain microbiological activity
aganst S. aureus.7 These compounds had high serum protein
binding, and here it was hoped that by altering the distribution
of hydrophobic moieties in the inhibitor, a different protein-
binding profile could be realized. For the specific designs,
modeling also indicated that aliphatic substitutions much larger
than a methyl group, including branched aliphatics, would not
be tolerated. Thus, only methyl and propyl substitutions were
prioritized for synthesis and evaluation.
In the case of proposed C-2 alkyl ethers, the oxygen atom
was predicted to form a new interaction with aromatic protons
of Phe66, with the alkyl moiety contributing an interaction with
Val51 and Leu52. The interaction between aromatic C−H and
oxygen atoms, traditionally termed a C−H···O hydrogen bond,
has been studied theoretically15−17 as well as experimen-
tally18−21 and is estimated to contribute between 0.3 and 1.5
kcal/mol in favorable interaction energy depending on the
electronic environment of the atoms involved.15−17 Methoxy
and ethoxy substitutions were selected as, again, modeling
predicted potential steric clashes between longer aliphatic
moieties in alkyl ethers and protein residues in the targeted
region (Supporting Information). Specificially, methoxy was
accommodated in the model, but ethoxy appeared too large and
forced a flip in modeled ring C.
Similar to the alkyl ethers, modeling predicted favorable
interactions between halogen atoms and Phe66 protons. While
such interactions are popularly known as “halogen bonds”,
theoretical analysis has revealed that the interaction is more
similar to a dipole−dipole interaction than a hydrogen bond22
and that C−F···H−C interactions are favorable with an
interaction energy of up to 1.6 kcal/mol.22 Moreover, a
systematic study of several protein−ligand complexes involving
C−F bonds interacting with the phenyl side chain of
phenylalanine and tyrosine residues showed that the favored
interaction of fluorine is with aromatic protons (edge
interaction), with very few systems showing fluorine approach-
ing the π-cloud of the aromatic ring (face interaction).22
To test these potential interactions with C-2 substituents (R2
in Table 1), a matched set of analogues of 7 and 8 were tested
for biochemical inhibition of S. aureus TMK. As shown in Table
1, replacement of hydrogen with alkyl groups at C-2 was
disfavored and resulted in a loss of activity (compounds 9 and
10). As predicted by modeling, the size of the substituent was
important for the enzymatic potency. While the loss in activity
D
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. X-ray structures of compounds 18 (A) and 19 (B) in complex with S. aureus TMK at 1.7 and 1.9 Å resolution, respectively. In addition to
interactions observed previously (hydrogen bonds with Arg70 and Gln101 and a salt-bridge interaction with Arg48, left panels), the C-2 substituents
form predicted interactions with Val51, Leu52, and Phe66 (right panels). In the case of 18, the ether oxygen forms a C−H···O hydrogen bond with
the aromatic proton of C-4′Phe66, while the methyl group forms hydrophobic interactions with Val51 and Leu52 (A). In the case of 19, the fluorine
atom forms a dipolar interaction with the aromatic proton of C-4′Phe66 (B). The distances shown are between heavy atoms.
for 9 was only about 5-fold, compound 10 with a much larger
propyl substituent was more than 30-fold less potent compared
to 8. We believe that the loss of potency for the methyl
substituent was a result of unfavorable steric and electronic
interactions with Phe66 (methyl hydrogens too closely
approach aromatic protons on Phe66). Thus, the proximity
of the methyl substituent to Phe66 resulted in an unfavorable
interaction, and the propyl substituent in addition was too large
for the available space. Alkoxy groups and halogens at C-2,
however, led to a 3−10-fold improvement in activity, suggesting
favorable interactions with Phe66. It was somewhat surprising
to see that the ethoxy analogue 12 was about 3-fold better than
8, since the docking model indicated that an ethoxy group, like
the propyl substitution, was too large for the space. Among
halogens, chlorine and bromine substitution resulted in a 3−5-
fold improvement over the parent compound (8). Fluorine
substitution resulted in the best example with a more than 10-
fold improvement in inhibition. Cyano substitution (16) was
neutral with regard to potency. These results allowed us to
focus on the methoxy and fluorine analogues for further
development.
physical properties and avoided human ether-a-go-go-related
gene (hERG) ion channel inhibition. Thus, maintenance of this
interaction with Arg48 in conjunction with any C-2 substitution
was paramount. We also at this time decided on the 3-chloro
substituent for ring D over the corresponding 3-bromo of
advanced compound 1. While 1 was extremely useful for in vivo
target validation,4 bromine is not favored nor well-represented
in drug candidates and in the case of 1 offered only small
advantages in potency and protein binding over the
corresponding 3-chloro analogues.
As anticipated, the carboxylates 18 and 19 resulted in
moderate improvements (up to 7-fold, Table 2) in IC50 and
represent an advance over the prior compound 17. To verify
the interactions afforded by the C-2 substituents, X-ray
structures of 18 and 19 bound to S. aureus TMK were solved.
The structures of both compounds confirmed previously
observed binding modes with all the interactions formed by
the A and D rings retained (Figure 2, left panels). Importantly,
the X-ray structures showed the C-2 substituents in 18 and 19
forming predicted interactions with Val51, Leu52, and Phe66
(Figure 2, right panels). In the case of 18, the ether oxygen was
3.3 Å from the nearest aromatic carbon of Phe66 (C-4′Phe66),
forming a C−H···O hydrogen bond, whereas the methyl carbon
was equidistant at 3.7 Å from the Val51 and Leu52 side chains.
A recent analysis17 of observed C−H···O interactions in
systems deposited in the Cambridge Structural Database
suggests that the optimal distance of oxygen from the centroid
To test this advance in more elaborate compounds, we
incorporated a previously identified C-4 carboxylate substitu-
ent7 (Table 2). As described previously,7 the carboxylate
moiety was particularly critical for improving enzyme potency
by forming a salt-bridge interaction with Arg48. Moreover,
introduction of a carboxylate also improved the MICs and
E
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
of the aromatic ring is 4.5 Å and the optimal C−H−O angle is
140°. In the X-ray structure of 18, these parameters were 4.6 Å
and 143°, respectively, consistent with the existence of a C−
H···O hydrogen bond. Ab initio calculations on similar systems
predict such interactions to be favorable, contributing up to 1.5
kcal/mol in interaction energy.17 In the case of 19, the fluorine
atom was 3.2 Å away from C-4′Phe66 and the angle C−F···C-
4′Phe66 was 149°. This geometry is in agreement with previously
reported analysis of fluorinated ligands and protein complexes
from the Protein Data Bank (PDB)22 in which such C−F···
Caromatic contacts were shown to have a mean distance of 3.2 Å
and mean angle of 130° (standard deviation 38°) and is
estimated to have 1.6 kcal/mol of positive interaction energy.
We next incorporated lipophilic substituents off the
methylene linker between the B and C rings (Table 3).
aureus. For example, compound 45 with a log D value of 0.51
had an MIC = 0.5 μg/mL against methicillin-resistant/
quinolone-resistant (MRQR) S. aureus. On the other hand,
compound 41 (log D = 0.13) had an MIC 8-fold less potent
than that of 45 in spite of a similar IC50. The best compounds,
46 and 47, showed excellent MICs of 0.25−0.5 μg/mL against
S. aureus.
It was striking that the compounds in Table 3 also revealed
an unexpected, altered preference for chirality of the R3
substituent. In the case of previously reported compounds7
(and including 34−37) with no substitution at C-2, R,S
diastereomers were consistently 2−10-fold more potent than
the corresponding S,S diastereomers. In the specific instance of
1 and its S,S diastereomer, 1 possessed a 3.5-fold better IC50
against S. aureus TMK and a 4-fold better MIC against MRQR
S. aureus.7 However, for the C-2-methoxy-substituted com-
pounds, the S,S diastereomer seemed to be significantly more
potent than the respective R,S diastereomer (Table 3). For
example, compound 39 has a ∼20-fold better IC50 than 38 and
has a >8-fold improvement in MIC against MRQR S. aureus.
The high-resolution 1.83 Å X-ray structure of 38 in complex
with S. aureus TMK displayed a very conserved binding mode
relative to the 1.67 Å structure of complexed 18
(RMSDligand heavy atoms = 0.16 Å, Figure 3). The methoxy moiety
formed expected interactions with Val51, Leu52, and Phe66;
the oxygen of the methoxy group was 3.5 Å from the nearest
aromatic carbon of Phe66, whereas the methyl carbon was 3.8
and 3.6 Å from Leu51 and Val52, respectively. The X-ray
structure also revealed a very small movement of the methoxy
carbon compared to that in 18, resulting in a slightly shorter
distance (3.6 Å compared to 3.8 Å) between the methoxy
carbon and Val51 side chain. The ethyl moiety at R3, however,
did not form the anticipated hydrophobic interactions with
Val51 and Leu52; the terminal carbon of the ethyl group was
4.6 Å from the Leu52 side chain.
a
Table 3. Activity of Analogues of 18 and 19
Since 38 bound in an unperturbed binding mode (relative to
18), we hypothesize that the observed lower biochemical
potency of 38 relative to its diastereomer 39 was due to the
strain arising from a steric clash between the methoxy group
and the R3 substituents in the R configuration (we did consider
that these di-ortho-substituted compounds could possibly be
subject to restricted rotation and atropisomerism, but the
relatively small differences in potency observed between
diastereomers 38 and 39, along with the very conserved
binding mode of 38, argued against this phenomenon;
atropisomerism is typically associated with much higher barriers
to rotation23). Such a clash can be avoided in the S
configuration while retaining the entropic gain achieved via
stabilization of the ligand conformation. The R-substituted C-2
fluoro analogues showed improved activity compared to their
OMe counterparts, which we propose was due to the much
smaller size of fluorine. In fact, the diastereomeric pairs of
fluoro analogues have similar IC50 values against S. aureus TMK
(Table 2), suggesting that the chirality of the R3 substituents
does not affect the binding and potency of the fluoro analogues.
This notion is further supported by the 1.62 and 1.78 Å X-ray
structures of 46 and 47, respectively, which revealed very
similar binding modes (RMSDligand heavy atoms = 0.18 Å) for the
core scaffolds (Figure 4). The isobutyl moiety in both X-ray
structures, however, did not form any hydrophobic interactions
with Val51 and Leu52 residues.
a
Diastereomeric pairs are grouped together in the shaded and white
bars. R2, the C-2 substituent, is highlighted in blue. The C-4
substituent is a carboxylic acid. Bacterial strains from AstraZeneca
collection: S. aureus ARC516/S. aureus ARC517 methicillin-resistant
(MR) and quinolone-resistant (QR). log D value calculated from
measured log D values of the diastereomeric mixture and the opposite
diastereomer 38.
b
c
Increasing the size and lipophilicity of substituents at the
methylene carbon (R3 substituents) resulted in a progressive
increase in log D and, as expected, improved MICs against S.
aureus (Table 3). In general, methoxy compounds resulted in
lower log D values as well as poorer MICs compared to the
corresponding fluorinated compounds. We found that changing
R3 from ethyl to propyl increased log D by approximately 0.4
for R2 = H and F, but the log D increase was smaller (∼0.15)
for the corresponding R2 = OMe compounds. This difference
was reflected in the MIC values, which do not improve on
going from R3 = ethyl to R3 = propyl when R2 = OMe. The
data in Table 3 continue to support the notion that the
lipophilicity of the compounds affects cell permeation in S.
Taken together, these data have deepened our understanding
of the role of the R3 substituent in the binding and potency of
F
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (A) X-ray structure of 38 in complex with S. aureus TMK. The X-ray structure confirmed previously observed interactions with Arg48,
Arg70, and Gln101. (B) Methoxy moiety in 38 with Val51 and Leu52, with the methoxy carbon placed 3.6 and 3.8 Å, respectively, from the Val51
and Leu52 side chains. The methoxy oxygen forms a C−H···O hydrogen bond interaction with the proton of C-4′Phe66. (C) Superimposition of X-
ray structures of 38 (green) and 18 (yellow) displaying identical binding modes for the core scaffold (RMSDligand heavy atoms = 0.16 Å). No significant
changes are observed in either protein structure upon binding of ligands (Cα RMSD = 0.18 Å). Distances between the methoxy group in 18 and
Val51, Leu52, and C-4′Phe66 are shown for reference. All distances shown are between heavy atoms.
potency gains from R3 substitution have little to do with direct
hydrophobic interaction with Val51 and Leu52, but instead
arise from (a) stabilization of the ring B/ring C turn that
positions rings A and D for specific interactions and (b) a
general hydrophobic effect that both improves the entropy of
desolvation and allows the more hydrophobic/higher log D
compound to penetrate the cellular membrane of S. aureus
more readily. Thus, with no potential steric clash between C-2
and R3 substituents (34−37), there is a minor gain from the R
configuration because of slightly positive interactions with the
hydrophobic patch. With a small C-2 fluorine in place,
preference for R or S disappears, and with a larger methoxy,
the preference inverts because the steric repulsion more than
cancels any small direct hydrophobic interaction. However, for
all molecules, the proposed dominant turn stabilization and
desolvation effects are preserved.
These results expand the series’ structure−activity relation-
ship (SAR) potential, as new combinations at C-2 and R3 can
be explored to improve the overall profile. The free fraction is
one key parameter: the chirality of substituents at R3 had an
impact on the human plasma protein binding of some C-2-
methoxy-substituted compounds (Table 4). For example,
compound 38 had a free fraction of 1.5% in human plasma,
whereas the other, more potent diastereomer 39 showed
significantly lower protein binding (11% free). Fluoro-
substituted compounds, in general, have higher protein binding
in human plasma than the methoxy-substituted compounds,
and the corresponding fluoro-substituted diastereomeric pair
(42 and 43) had a more uniformly low free fraction. Several
studies have reported the effect of chirality on plasma protein
binding,24 resulting in significant differences in the amount of
free drug present in the plasma.25,26 Moreover, such stereo-
selective plasma protein binding can be species-dependent.26
Indeed, compounds 38 and 39, which have a significant
difference in the human plasma, did not exhibit these
differences in mouse plasma, which is important for perform-
ance in the infection model (Table 4). We also measured
improvements in metabolic clearance; reference compound 1
had a clearance in rat of 42 mL/min/kg (61% of hepatic blood
flow). In the series presented here, clearance was dependent on
the R3 chirality: for example, rat clearance for 44 and 45 was 64
and 18 mL/min/kg, respectively, and for 46 and 47, it was 84
Figure 4. (A) Superimposition of X-ray structures of diastereomers 46
(green) and 47 (yellow) in complex with S. aureus TMK displaying
identical binding modes for the core scaffold of ligands
(RMSDligand heavy atoms = 0.18 Å). No significant changes were observed
in either protein structure upon binding of ligands (Cα RMSD = 0.24
Å). In neither ligand does the butyl group form hydrophobic
interactions with Val51 and Leu52, as the distances between the
isobutyl groups and side chains are greater than 4 Å. (B) The
experimental electron densities of the ligands (2Fo − Fc map at 1.2σ)
clearly show the absolute configuration at the chiral center. The
ligands are rotated 90° relative to those in panel A.
this inhibitor series. While the prior series7 exhibited a small,
but consistent, preference (2−10 fold) for the R configuration
of alkyl substituents at this position, the C-2 methoxy
compounds here yielded optimized compounds with the
opposite stereochemistry, and the C-2 fluoro eliminated the
preference altogether. Taken as a whole, we propose that the
G
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
aureus using the same procedure as described previously.4
Under selection at 4 times the MIC, resistant colonies were
found at a rate of 6 × 10−9. Those mutants found to be stable
upon passage showed an 8- to >64-fold increase in MIC.
Sequencing of the mutants revealed changes in the TMK
protein coding region that produced the single amino acid
changes Gly44Ser and Glu37Asp in separate mutants. These
mutations map to the inhibitor-binding region (Supporting
Information).
Table 4. Effect of the Chirality of the R3 Substituent (Table
3) on Plasma Protein Binding in Diastereomers
a
As a proof-of-principle that efficacy was achievable in this
series of compounds, 47 was evaluated in vivo against
methicillin-sensitive S. aureus ARC516 in a mouse thigh
infection model. The selection of 47 was based on better
MICs against S. aureus and other Gram-negative pathogens and
improved pharmacokinetics (PK) (Table 4) over 39. 47 was
administered once daily via the intraperitoneal route for doses
from 5 to 200 mg/kg (Supporting Information); to achieve a
total daily dose of 400 mg/kg, 200 mg/kg was administered bid,
q12h (twice daily, every 12 h). 47 was found to be efficacious,
inhibiting progression of infection at a dose between 100 and
200 mg/kg/day and achieving a 10-fold reduction in the
number of colony-forming units (CFU) at a dose of 400 mg/
kg/day (Figure 5A). While it is not indicative of a defined
pharmacokinetic/pharmacodynamic relationship, for the 200
and 400 mg/kg/day doses, which resulted in a decrease in
bacterial burden below the starting input, the amount of time
above the free plasma MIC over 24 h ranged from 27% to 60%.
These results were similar to those of 14 when dosed in the
same model (the exposures between 5 and 200 mg/kg for 47
are shown in Figure 5B), yielding an efficacious dose for
infection stasis of about 150 mg/kg/day. This was not
surprising given that 1 and 47 are broadly comparable in
MIC, mouse free fraction, and mouse clearance. As the in vivo
behavior of 47 was consistent with that of 1, these results
validate TMK as an antibacterial target with a second inhibitor
and reinforce our focus on improving the key parameters of
potency, protein binding, and clearance. These results also
underline the potential of C-2-substituted compounds to be
developed into an efficacious antibacterial agent against serious
bacterial infections caused by Gram-positive pathogens, in
particular S. aureus, with low cytotoxicity in humans. The
challenge is improving the cellular potency and pharmacoki-
a
Diastereomeric pairs are grouped together in the shaded and white
bars. ND = not determined.
and 38 mL/min/kg, respectively (Table 4). In each case, the S
diastereomer at R3 yielded an improved clearance rate.
The observed improvement in human plasma protein
binding along with reasonable MIC and MIC90 values (Table
5) against MRQR S. aureus established compounds in this
series as strong alternate leads. These compounds also had
good MICs against Staphylococcus epidermidis and Enterococcus
faecium and excellent selectivity against human TMK (>25000-
fold in IC50) and Gram-negative pathogens Escherichia coli and
Pseudomonas aeruginosa (MIC > 64 μg/mL, Table 5). Other
microbiological advantages to these inhibitors such as rapid
bactericidal activity and a low resistance rate were established
earlier4 and were only verified here using simple analogues.
Specifically, the D ring bromo analogue of 34 was tested in a
time-kill experiment as described previously.4 Against S. aureus,
rapid and concentration-dependent killing (>3 log units in 4 h)
was observed at 0.5 times the MIC and above. Some regrowth
was observed at 24 h at 0.5 and 1 times the MIC. Separately,
compound 36 was used to generate resistance mutants in S.
Table 5. Gram-Positive Spectrum Data, Population MICs, and Selectivity of Selected Compounds
MIC (μg/mL)
a
bacterial strain
1
39
46
47
b
b
b
b
c
S. aureus/S. aureus MRQR
Streptococcus pneumoniae
Streptococcus pyogenes
Staphylococcus epidermidis
Enterococcus faecium
Enterococcus faecalis
MIC90 = 1
MIC90 = 4
MIC90 = 2
0.5/0.5
b
c
MIC90 = 0.03
MIC90 = 2
MIC90 = 0.13
0.13
2
b
c
MIC90 = 0.5
MIC90 = 4
MIC90 = 2
2
4
2
4
0.25
1
8
2
2
>8
>64
>64
8
4
Escherichia coli
>64
>64
>64
>64
>64
>64
Pseudomonas aeruginosa
human TMK IC50 (nM)
62000
>64
>200000
>64
51000
ND
29000
ND
human A549 MIC (μg/mL)
a
Bacterial strains from AstraZeneca collection: S. aureus ARC516; SauMRQR, S. aureus ARC517 methicillin-resistant (MR) and quinolone-resistant
(QR); S. pneumoniae ARC2481; S. pyogenes ARC2834; S. epidermidis ARC323; E. faecium ARC1239 linezolid-resistant (LRE); E. faecalis ARC1618
b
vancomycin-resistant (VRE); E. coli ARC523; P. aeruginosa ARC545. S. aureus MIC90 of 22 strains, S. pneumoniae MIC90 of 19 strains, S. pyogenes
c
MIC90 of 21 strains. S. aureus MIC90 of 23 strains, S. pneumoniae MIC90 of 20 strains, S. pyogenes MIC90 of 21 strains.
H
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
ences at this position and yielded equipotent diastereomeric
compound pairs. These results have shed light on the specific
role of the R3 alkyl substituent, showing it to be sensitive to an
intramolecular steric clash with C-2 substituents when bound
and indicating that a direct hydrophobic interaction with
conserved Val51 and Leu52 is likely to be of minor value.
These results also open up new options for SAR development,
while preserving broad Gram-positive coverage with good
MICs against S. aureus, S. pyogenes, S. epidermidis, and
Enterococcus species and excellent selectivity against human
TMK. An example of this new series was successful in the S.
aureus mouse infection model, proving that new C-2- and R3-
substituted compounds can be efficacious in vivo. We continue
to develop SARs for these inhibitors to find the proper balance
of MIC, low clearance, and low protein binding that will result
in lower efficacious doses. The work presented here not only
illustrates an example of a structure-guided design approach,
but also highlights how taking advantage of a small region
observed in the X-ray structures led to new SAR directions and
new insights into the sources of binding potency.
EXPERIMENTAL SECTION
■
General Experimental Details. All commercial reagents and
anhydrous solvents were obtained from commercial sources and were
used without further purification, unless otherwise specified.
LC−MS Conditions. Method 1. Samples were analyzed by reversed-
phase LC−MS using Varian Polaris C18A, 2 mm × 50 mm, 3 μm
particle size columns. An Agilent HP1100 (Wilmington, DE) LC
system was used with a gradient elution profile of 5−95% B over 4.5
min at 1 mL/min and then re-equilibration at the initial conditions to
6 min. The injection volume was 2 μL and column temperature 30 °C.
Mobile phase A was 0.1% formic acid in water and mobile phase B
0.1% formic acid in acetonitrile. Detection was based on electrospray
ionization (ESI) in positive and negative polarity using a Waters ZQ
mass spectrometer (Milford, MA), diode-array UV detector from 210
to 400 nm, and evaporative light-scattering detector (Sedex 75, Sedere,
Alfortville Cedex, France).
Method 2. This method was the same as method 1, except mobile
phase A was 10 mM ammonium acetate in 5:95 acetonitrile/water and
mobile phase B was acetonitrile.
Method 3. This method was the same as method 1, except the
gradient elution profile was 0−95% B.
Method 4. This method was the same as method 2, except the
gradient elution profile was 0−95% B.
Accurate mass was measured using a hybrid quadrupole time-of-
flight mass spectrometer (microTOFQ, Bruker Daltonics) in ESI+
mode. Method: 5−95% mobile phase B from 0.0 to 5.0 min, hold at
95% mobile phase B to 5.1 min (mobile phase A, 0.1% formic acid in
water; mobile phase B, 0.1% formic acid in acetonitrile), 1.0 mL/min
flow, column Varian Polaris C18-A 2.0 mm × 50 mm, 3.0 μm particle
size. A 1 μL volume of sample was injected into an HPLC instrument
using an externally calibrated instrument for accurate mass measure-
ment. A 50 μL/min flow of eluent from the LC instrument and a 50
μL/min flow of make-up solution (CH3CN/H2O, 50:50) went into
the MS instrument.
Figure 5. (A) Dose−response effect on the treatment of mice with 47
in a S. aureus thigh infection model. Error bars represent standard
errors in CFU measurements. (B) PK profile of 47 representing a
nonlinear exposure in mice with predosed aminobenzotriazole (ABT).
netic profile of compounds like 39 while maintaining the gains
achieved in free fraction. We believe that taking advantage of
the vector offered by the ethyl group in 39 or substitutions in
other parts of the scaffold will allow desired improvements in
cellular potency while maintaining or bettering the protein
binding and metabolic profile of 39, and new results along these
lines will be disclosed in due course.
CONCLUSIONS
■
In summary, we employed a structure-guided design approach
to exploit a previously unexplored binding region in the S.
aureus TMK enzyme. Our goals were to provide new options
for SAR development that could improve these Gram-positive
TMK inhibitors in the areas of potency, serum free fraction,
and/or metabolic clearance, which we have identified as the key
parameters driving in vivo efficacy.4,7 Compounds with
halogens as well as small alkyl and alkyl ether substituents at
C-2 were synthesized and tested to probe this area of the
enzyme. We found that fluoro and methoxy substituents at C-2
improved potency through new interactions with aromatic
hydrogens and hydrophobic side chains in the vicinity. Alkyl
groups or larger alkoxy groups did not provide the same
benefit. An unexpected discovery was that the methoxy
substituent inverted the chiral preference at the R3 (linker)
position for binding, a result we attribute to an unfavorable
intramolecular steric interaction in the bound form of the
inhibitor that overrides the small preference for a direct
hydrophobic interaction in the opposite diastereomer in C-2-
unsubstituted analogues. On the other hand, the best halogen
substituent, fluorine, largely eliminated diastereomeric differ-
1H NMR spectra (δ, ppm) were recorded using a Bruker Advance
Ultrashield 300 MHz or Bruker DPX 400 MHz instrument.
Column chromatography was performed using Silcycle FLH-
R10030B Silisep cartridges (12−330 g). Preparative reversed-phase
HPLC chromatography was carried out using a Waters Atlantis T3-
C18 column, 19 × 100 mm, 5 μm, a linear gradient from 10% to 90%
CH3CN in H2O over 12 min (0.1% trifluoroacetic acid), and a flow
rate of 20 mL/min. Preparative chiral chromatography was carried out
as follows: HPLC, Chiralpak IC column, 30 × 250 mm, 5 μm, hexane
(60%), methanol/ethanol (1:1) (40%), 0.5% diethylamine, flow rate
40 mL/min; SFC, Chiralpak AD column, 30 × 250 mm, 5 μm, carbon
dioxide (60%), 2-propanol (40%), flow rate 120 mL/min; chiral
I
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
analytical HPLC, Chiralpak IC column, 4.6 × 250 mm, 5 μm, hexane
(60%), methanol/ethanol (1:1) (40%), 0.5% diethylamine, flow rate 1
mL/min; chiral analytical SFC, Chiralpak AD column, 4.6 × 250 mm,
5 μm, carbon dioxide (60%), 2-propanol (40%), flow rate 2.8 mL/min.
The purity of the final compounds was assessed on the basis of
analytical LC−MS, and the results were greater than 95% unless
specified otherwise. The diastereomeric ratio of specific compounds
was measured by chiral analytical HPLC and was >95:5 unless
otherwise indicated.
Molecular Modeling. Molecular docking was performed using
Glide version 5.027,28 in standard flexible docking mode. X-ray
structures of S. aureus TMK protein in complex with related ligands (2,
18, and 19; PDB IDs 4HEJ, 4QG7, and 4QGA, respectively) were
used as templates for docking studies. The protein structures were
prepared using Protein Preparation Wizard, whereas the ligands were
prepared using the Ligprep utility in Maestro 8.5 (Schrodinger, LLC,
New York, 2008). Docked poses were minimized in the OPLS 200129
force field as implemented in Maestro 8.5.
General experimental procedures for biochemical assays, protein
crystallography, and the in vivo infection model are described in the
Supporting Information.
Screening Pharmacokinetics. In vivo screening pharmacokinetic
experiments were performed in male Wistar rats dosed at 5−10 mg/kg
iv bolus in a vehicle of DMA/TEG/saline (40:40:20). Bioanalysis of
plasma levels was performed as previously described.4
Experimental Synthetic Procedures and Characterization
Data of the Compounds. Experimental procedures and character-
ization data for compounds 1, 2, 17, and 34−37 and intermediate 6
can be found elsewhere.4,7
and then filtered to remove the solids. The polymer was rinsed with
additional DMF (5 mL), and the combined organic layers were
concentrated. The crude residue was redissolved in DMSO and
purified by HPLC employing a 10−75% acetonitrile gradient in water
(0.1% TFA). The desired fractions were pooled and lyophilized to
furnish 80 mg (35%) of 13 as an off-white solid. LC−MS (method 3):
tR = 2.00 min, m/z = 444 (M + H). 1H NMR (300 MHz, MeOH-d4):
δ 7.44 (m, 2H), 7.33 (m 3H), 7.14 (m, 1H), 7.01 (t, 1H), 6.92 (dd,
1H) 4.73 (m, overlapping w/CD3OD, ∼1H), 4.53 (m, 2H), 3.60 (d,
1H), 3.40 (t, 1H), 3.10 (t, 1H), 1.90−2.25 (m, 4H), 1.87 (s, 3H) ppm.
HRMS: m/z 444.1467, calcd 444.1491 (M + H).
(S)-1-(1-(3-(3-Chloro-5-fluorophenoxy)-2-methylbenzyl)piperidin-
3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (9). Yield: 75 mg (13%)
as an off-white solid. LC−MS (method 3): tR = 2.30 min, m/z = 460
1
(M + H). H NMR (300 MHz, MeOH-d4): δ 7.42 (m, 3H) 7.17 (m,
1H) 6.93 (dt, 1H), 6.72 (m, 1H) 6.62 (dt, 1H), 4.55 (m, 1H), 4.50
(m, 2H) 3.53 (m, 2H), 3.45 (m, 1H), 3.15 (m, 1H) 2.33 (s, 3H),
1.80−2.25 (overlapping s and m, 7H) ppm. HRMS: m/z 458.1641,
calcd 458.1647 (M + H).
(S)-1-(1-(3-(3-Chloro-5-fluorophenoxy)-2-propylbenzyl)piperidin-
3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (10). Yield: 19.1 mg
(24%) as an off-white solid. LC−MS (method 3): tR = 2.50 min,
1
m/z = 488 (M + H). H NMR (300 MHz, MeOH-d4): δ 7.73 (br s,
1H), 7.22 (overlapping m’s, 2H), 6.92 (m, 1H), 6.88 (dt, 1H), 6.67
(m, 1H), 6.57 (dt, 1H), 4.57 (m, 1H), 3.60 (m, 2H), 2.85 (m, 2H),
2.65 (m, 2H), 2.33 (dt, 2H), 1.45−1.90 (s overlapping m’s, 9H), 0.97
(t, 3H) ppm.
(S)-1-(1-(3-(3-Chloro-5-fluorophenoxy)-2-methoxybenzyl)-
piperidin-3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (11). Yield: 44
mg (93%) as an off-white solid. LC−MS (method 4): tR = 3.00 min,
Compounds presented in Table 1 (9−16) were synthesized using a
general procedure exemplified for compound 13 (Scheme 1).
2-Fluoro-3-hydroxybenzaldehyde (4). The title compound was
prepared by microwave irradiation of a solution of 2-fluoro-3-
methoxybenzaldehyde (3; 600 mg, 3.89 mmol) in a solution of
hydrogen bromide and acetic acid (33% (w/w), 7.2 mL, 39.64 mmol)
at 80 °C for 45 min. The reaction mixture was poured into water (20
mL) and quenched by the addition of excess saturated sodium
bicarbonate solution (50 mL). The resulting mixture was extracted
with EtOAc (3 × 50 mL), and the combined organic layers were dried
with magnesium sulfate, filtered, and concentrated. The crude residue
was chromatographed on silica gel employing a 0−20% EtOAc
gradient in chloroform. Desired fractions were pooled and
concentrated to furnish 250 mg (46%) of 4 as a colorless solid.
1
m/z = 474 (M + H). H NMR (300 MHz, CD3OD-d4): δ 7.46 (br s,
1H), 7.43 (m, 1H), 7.29 (m, 2H), 6.95 (dt, 1H), 6.80 (br s, 1H), 6.70
(dt, 1H), 4.75 (m, 1H) 4.45 (m, 2H), 3.93 (s, 3H), 3.57 (m, 2H), 3.07
(m, 1H), 1.93−2.23 (m, 4H), 1.88 (s, 3H) ppm. HRMS: m/z
474.1521, calcd 474.1597 (M + H).
(S)-1-(1-(3-(3-Chloro-5-fluorophenoxy)-2-ethoxybenzyl)piperidin-
3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (12). Yield: 72 mg (44%)
as an off-white solid. LC−MS (method 3): tR = 2.32 min, m/z = 490
1
(M + 3). H NMR (300 MHz, MeOH-d4): δ 7.43 (m, 2H), 7.29 (m,
2H), 6.95 (dt, 1H), 6.79 (m, 1H), 6.69 (dt, 1H), 4.73 (m, 1H), 4.45
(m, 1H), 4.17 (q, 2H), 3.57 (m, 2H), 3.37 (m, 1H), 3.10 (m, 1H),
1.83−2.23 (m and s 7H) ppm. HRMS: m/z 488.1761, calcd 488.1753
(M + H).
1
LC−MS (method 1): tR = 1.23 min, m/z = 139 (M − 1). H NMR
(300 MHz, DMSO-d6): δ 10.35 (s, 1H), 10.23 (s, 1H), 7.20 (m, 3H)
ppm.
(S)-1-(1-(2-Chloro-3-(3-chloro-5-fluorophenoxy)benzyl)piperidin-
3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (14). Yield: 32 mg (17%)
as an off white solid. LC−MS (method 3): tR = 2.31 min, m/z = 478
3-(3-Chlorophenoxy)-2-fluorobenzaldehyde (5). The title com-
pound was prepared by stirring a mixture of 4 (150 mg, 1.07 mmol), 3-
chlorophenylboronic acid (250 mg, 1.6 mmol), copper(II) acetate
(427 mg, 2.14 mmol), and 3 Å molecular sieves in a 40 mL vial in
anhydrous dichloromethane (30 mL). Hunig’s base (0.932 mL, 5.35
mmol) was added in one portion, and the reaction was stirred at room
temperature for 72 h. The reaction was diluted with additional
dichloromethane (20 mL) and water (20 mL) and then acidified by
addition of 2 M HCl. The biphasic mixture was stirred vigorously and
filtered, and the organic layer was collected. The acidic layer was
extracted twice more with dichloromethane (2 × 30 mL), and then the
combined organic layers were dried over magnesium sulfate, filtered,
and concentrated. The crude residue was chromatographed employing
a 0−100% chloroform gradient in hexanes. The desired fractions were
concentrated to provide 129 mg (48%) of 5 as a colorless oil. LC−MS
(method 1): tR = 2.88 min, m/z = 249 (M − 1). 1H NMR (300 MHz,
MeOH-d4): δ 10.27 (s, 1H), 6.65−7.55 (m, 7H) ppm.
(S)-1-(1-(3-(3-Chlorophenoxy)-2-fluorobenzyl)piperidin-3-yl)-5-
methylpyrimidine-2,4(1H,3H)-dione (13). The title compound was
prepared by stirring polymer-supported cyanoborohydride (515 mg,
2.0 mmol/g, 1.03 mmol), 5 (129 mg, 0.51 mmol), and (S)-5-methyl-1-
(piperidin-3-yl)pyrimidine-2,4(1H,3H)-dione (6; 194 mg, 0.93 mmol)
in a 50 mL flask in an 18:1 mixture of DMF (9 mL) and acetic acid
(0.50 mL). The suspension was stirred for 18 h at room temperature
1
(M + H). H NMR (300 MHz, MeOH-d4): δ 7.44−7.62 (m, 3 H)
7.36 (dd, 1H) 6.99 (dt, 1H) 6.74−6.82 (m, 1H) 6.70 (dt, 1H) 4.66−
4.77 (m, 1H) 4.51−4.66 (m, 2H) 3.51−3.67 (m, 2H) 3.35−3.48 (m,
1H) 3.04−3.25 (m, 1H) 1.92−2.23 (m, 4H) 1.89 (d, 3H) ppm.
HRMS: m/z 478.1114 calcd 478.1101 (M + H).
(S)-1-(1-(2-Bromo-3-(3-chloro-5-fluorophenoxy)benzyl)piperidin-
3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (15). Yield: 53 mg (24%)
as an off-white solid. LC−MS (method 3): tR = 2.34 min, m/z = 522
(M + H). 1H NMR (300 MHz, MeOH-d4): δ 7.57−7.65 (m, 2H) 7.52
(d, 1H) 7.31−7.42 (m, 1H) 7.03 (dt, 1H) 6.78−6.85 (m, 1H) 6.73
(dt, 1H) 4.71−4.82 (m, 1H) 4.62−4.70 (m, 2H) 3.59−3.73 (m, 2H)
3.44−3.56 (m, 1H) 3.16−3.30 (m, 1H) 1.95−2.29 (m, 4H) 1.94 (d,
3H) ppm. HRMS: m/z 522.0605, calcd 522.0517 (M + H).
(S)-2-(3-Chlorophenoxy)-6-((3-(5-methyl-2,4-dioxo-3,4-dihydro-
pyrimidin-1(2H)-yl)piperidin-1-yl)methyl)benzonitrile (16). Yield: 40
mg (24%) as an off-white solid. LC−MS (method 3): tR = 2.16 min,
1
m/z = 451 (M + H). H NMR (300 MHz, MeOH-d4): δ 8.46 (br s,
1H), 7.83 (br s, 1H), 7.59 (t, 1H), 7.42 (t, 1H), 7.36 (d, 1H), 7.25 (m,
1H), 7.12 (t, 1H), 7.02 (dd, 1H) 6.94 (m, 1H) 4.58 (m, 1H) 3.77 (s,
1H), 2.88 (dd, 1H), 2.77 (d, 1H), 2.48 (m, 2H), 1.63−1.93 (s
overlapping m, 7H) ppm. HRMS: m/z 451.1531, calcd 451.1538 (M +
H).
J
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Compounds presented in Table 2 were synthesized using a general
procedure exemplified for compounds 38, 39, 46, and 47 (Schemes 2
and 3).
and the resulting mixture was stirred at reflux for 12 h. The mixture
was concentrated under reduced pressure, and the residue was
dissolved in water (25 mL) and extracted once with diethyl ether to
remove neutral impurities. The aqueous portion was acidified with 1.5
N HCl and extracted with ethyl acetate (3 × 20 mL). The combined
organic layers were washed with brine (10 mL), dried over sodium
sulfate, and evaporated. The crude residue was purified by flash
chromatography using 5% methanol in chloroform to furnish 26 (10
mg, 19%). LC−MS (method 3): tR = 2.29 min, m/z = 528 (M + H).
1H NMR (400 MHz, DMSO-d6): δ 0.65−0.72 (m, 3H), 1.62−1.98
(m, 8H), 2.08−2.20 (m, 1H), 2.24−2.36 (m, 1H), 2.62−2.74 (m, 1H),
2.90−3.01 (m, 1H), 3.60−3.74 (m, 4H), 4.52−4.60 (m, 1H), 4.70−
4.94 (m, 1H), 6.79−6.98 (m, 2H), 7.09 (d, J = 7.96 Hz, 1H), 7.24−
7.34 (m, 1H), 7.51−7.57 (m, 1H), 7.64−7.78 (m, 2H), 11.36−11.40
(m, 1H), 13.20 (br s, 1H) ppm.
4-Bromo-3-(3-chlorophenoxy)-2-methoxybenzaldehyde (21). To
a stirred suspension of bromo-3-hydroxy-2-methoxybenzaldehyde (20;
5 g, 21.645 mmol), 3-chlorophenylboronic acid (6.77 g, 43.290
mmol), and pyridine (5.14 g, 64.935 mmol) in 1,2-dichloroethane (50
mL) was added copper(II) acetate (9.8 g, 54.112 mmol). The mixture
was then heated at 55 °C for 3 days with continuous oxygen bubbling.
The reaction mixture was filtered through a Celite bed, and the filtrate
was washed successively with water (2 × 25 mL) and brine (25 mL).
The organic layer was dried over Na2SO4 and evaporated under
reduced pressure. The crude residue was purified by flash
chromatography using 20% ethyl acetate in hexane to yield 2.6 g
(35%) of the title product 21. LC−MS (method 1): tR = 3.28 min, m/
1
2-(3-Chlorophenoxy)-3-methoxy-4-{(1R)-1-[(3S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl]propyl}benzoic
Acid (38). A 36 mg sample of 26 was subjected to HPLC preparative
chiral separation eluting with a mixture of hexanes (50%), methanol/
ethanol (1:1) (50%), and diethylamine (0.1%). The title compound
eluted first. After removal of the solvent, the sample was lyophilized to
provide 10 mg of 38 (30%) as a white solid. LC−MS (method 4): tR =
z = 341 (M + H). H NMR (400 MHz, CD3OD): δ 3.80 (s, 3H),
6.74−6.80 (m, 2H), 7.02−7.06 (m, 1H), 7.26−7.34 (m, 2H), 7.89−
7.94 (m, 1H), 10.32 (s, 1H) ppm.
1-[4-Bromo-3-(3-chlorophenoxy)-2-methoxyphenyl]propan-1-ol
(22). To a stirred solution of aldehyde 21 (700 mg, 2.049 mmol) in
diethyl ether (10 mL) was added ethylmagnesium bromide (6.15 mL,
6.147 mmol, 1.0 M solution in diethyl ether) at −78 °C, and the
resulting mixture was allowed to stir for 2 h. The reaction was
quenched with saturated aqueous NH4Cl (10 mL) and then diluted
with ethyl acetate (60 mL). The organic layer was washed with water
(2 × 20 mL) and brine (20 mL) and dried over Na2SO4. The solvent
was evaporated under reduced pressure to obtain the crude product,
which was purified by flash chromatography using 20% ethyl acetate in
hexane to yield 275 mg (36%) of the title compound 22. LC−MS
(method 1): tR = 3.27 min, m/z = 371 (M + H). 1H NMR (400 MHz,
CD3OD): δ 0.93 (t, 3H), 1.71−1.76 (m, 2H), 3.80 (s, 3H), 4.88−4.82
(m, 1H), 6.73−6.77 (m, 2H), 7.04−7.06 (m, 1H), 7.25−7.33 (m, 2H),
7.44−7.47 (m, 1H) ppm.
1-[(3S)-1-{1-[4-Bromo-3-(3-chlorophenoxy)-2-methoxyphenyl]-
propyl}piperidin-3-yl]-5-methylpyrimidine-2,4(1H,3H)-dione (24).
To a stirred solution of 22 (450 mg, 1.32 mmol) in dry
dichloromethane (5 mL) was added triethylamine (400 mg, 3.96
mmol) followed by mesyl chloride (227 mg, 1.050 mmol) at 0 °C, and
the mixture was stirred at room temperature for 2 h. The mixture was
diluted with additional dichloromethane and then washed successively
with 10% NaHCO3 (10 mL), water (10 mL), and brine (20 mL), dried
over sodium sulfate, and filtered. The resulting solution was
concentrated under reduced pressure to obtain the crude mesyl
derivative 23 (0.700 mmol), which was dissolved in CH3CN (5 mL).
Next 6 (673 mg, 3.30 mmol) was added in one portion, and the
reaction mixture was stirred at 80 °C for 12−36 h. The reaction
solvent was evaporated, and the mixture was diluted with ethyl acetate
(25 mL) and then washed with water (25 mL) and brine (25 mL),
dried over Na2SO4, filtered, and concentrated. The obtained crude
residue was purified by flash chromatography (50% ethyl acetate in
hexane) to yield the title product 24 (100 mg, 16%). LC−MS
(method 3): tR = 2.41 min, m/z = 562 (M + H). 1H NMR (400 MHz,
DMSO-d6): δ 0.77 (t, 3H), 1.60−1.76 (m, 7H), 1.80−2.02 (m, 4H),
2.74−2.82 (m, 2H), 3.65−3.69 (m, 3H), 3.82−3.88 (m, 1H), 4.30−
4.40 (m, 1H), 6.71−6.92 (m, 2H), 7.10−7.14 (m, 1H), 7.20−7.23 (m,
1H), 7.27−7.38 (m, 1H), 7.52−7.55 (m, 1H), 7.63−7.67 (m, 1H),
11.20−11.22 (m, 1H) ppm.
1
1.55 min, m/z = 528 (M + H). H NMR (400 MHz, MeOH-d4): δ
7.83 (d, 1H), 7.50 (d, 1H), 7.40 (br s, 1H), 7.25 (t, 1H), 7.04 (dd,
1H), 6.85 (br s, 1H), 6.74 (dd, 1H), 4.73 (probably overlapping m’s,
also overlapping CD3OD, ∼2H), 3.89 (s, 3H), 3.82 (m, 1H), 3.42 (m,
1H), 3.13 (m, 1H), 2.87 (m, 1H), 2.40−1.70 (s overlapping m’s, 9H),
0.85 (t, 1H). HRMS: m/z 528.1896, calcd 528.1902 (M + H).
2-(3-Chlorophenoxy)-3-methoxy-4-{(1S)-1-[(3S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl]propyl}benzoic
Acid (39). A 36 mg sample of 26 was subjected to HPLC preparative
chiral separation eluting with a mixture of hexanes (50%), methanol/
ethanol (1:1) (50%), and diethylamine (0.1%). The title compound
eluted second. After removal of the solvent, the sample was lyophilized
to provide 11 mg of 39 as a white solid. LC−MS (method 4): tR = 1.65
min, m/z = 528 (M + H). 1H NMR (400 MHz, MeOH-d4): δ 7.83 (d,
1H), 7.52 (d, 1H), 7.43 (br s, 1H), 7.25 (t, 1H) 7.04 (dd, 1H) 6.85 (br
s, 1H), 6.74 (dd, 1H) 4.70 (probably overlapping m’s, also overlapping
CD3OD, 2H), 3.92 (s, 3H), 3.77 (m, 1H), 3.44 (m, 1H), 3.10 (m,
1H), 2.87 (m, 1H), 2.40−1.70 (s overlapping m’s, 9H, 0.87 (t, 1H).
HRMS: m/z 528.1905, calcd 528.1902 (M + H).
2-(3-Chlorophenoxy)-3-methoxy-4-((R)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)benzoic
Acid (40). A 30 mg sample of racemic material was subjected to HPLC
preparative chiral separation eluting with a mixture of hexanes (50%),
methanol/ethanol (1:1) (50%), and diethylamine (0.1%). The title
compound eluted second. After removal of the solvent, the sample was
lyophilized to provide 40 as a solid (98:2 diastereomeric ratio) (5.3
mg, 18%). LC−MS (method 4): tR = 2.18 min, m/z = 542 (M + H).
2-(3-Chlorophenoxy)-3-methoxy-4-((S)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)benzoic
Acid (41). A 30 mg sample of racemic material was subjected to HPLC
preparative chiral separation eluting with a mixture of hexanes (50%),
methanol/ethanol (1:1) (50%), and diethylamine (0.1%). The title
compound eluted first. After removal of the solvent, the sample was
lyophilized to provide 41 as a gum (>99:1 diastereomeric ratio) (9.7
mg, 32%). LC−MS (method 4): tR = 1.98 min, m/z = 542 (M + H).
(S)-2-(3-Chlorophenoxy)-3-methoxy-4-((3-(5-methyl-2,4-dioxo-
3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)methyl)benzoic Acid
(18). Yield: 18 mg (28.9%) as an off-white solid. LC−MS (method
3): tR = 2.17 min, m/z = 500 (M + H). 1H NMR (300 MHz,
CD3OD): δ 7.76 d 1H; 7.53 d 1H; 7.42 d 1H; 7.25 t 1H; 7.02 dm 1H;
6.84 (t,1H); 6.74 (dm, 1H); 4.70 (m, 1H); 4.47 (q, 1H); 3.90 (s, 3H);
3.53 (m, 2H); 3.33 (m, overlapping CD3OD, 1H); 3.13 (t, 1H); 2.20−
1.90 (m, 4H); 1.87 (s, 3H) ppm.
2-(3-Chlorophenoxy)-3-methoxy-4-{1-[(3S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl]propyl}benzoic
Acid (26). In a 2−5 mL microwave vial a solution of 24 (90 mg, 0.159
mmol) in DMF (3 mL) was purged with nitrogen for 10 min.
Zn(CN)2 (56 mg, 0.479 mmol) and Pd(PPh3)4 (8.3 mg, 0.007 mmol)
were added at room temperature, then the reaction vial was capped,
and the reaction mixture was heated at 110 °C for 45 min under
microwave irradiation. After being cooled to room temperature, the
reaction was quenched with ice−water (20 mL) and extracted with
diethyl ether (3 × 10 mL). The combined organic layers were washed
with brine (10 mL), dried over anhydrous sodium sulfate, and
evaporated under reduced pressure. The crude 25 was dissolved in
ethanol (8 mL), then solid NaOH (79 mg, 1.972 mmol) was added,
1-(4-Bromo-2,3-difluorophenyl)-3-methylbutan-1-ol (28). 1,4-Di-
bromo-2,3-difluorobenzene (27; 4.35 g, 16.00 mmol) was dissolved in
THF (70.8 mL) and the solution cooled to −40 °C. Isopropylmagne-
sium chloride−lithium chloride (17.23 mL, 22.40 mmol) was added
slowly, keeping the temperature below −30 °C. The mixture was
stirred at about −35 °C for 1 h and warmed to −7 °C for an additional
K
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1 h. The reaction mixture was cooled to −40 °C, 3-methylbutanal
(3.43 mL, 32.00 mmol) was added in one portion (the temperature
rose to −30 °C), and stirring was continued for 4 h at −25 °C. The
reaction mixture was poured into 1 M H2SO4 and ice, and the resulting
mixture was saturated with solid NaCl and then extracted twice with
tert-butyl methyl ether (2 × 150 mL). The combined extracts were
dried over sodium sulfate and concentrated in vacuo. After silica gel
chromatography (0−100 ethyl acetate in hexane) 28 was isolated as a
colorless oil (5.50 g, 123%). The material was taken on to the next
step with no additional purification. LC−MS (method 3): tR = 3.24
min, m/z = 279.0 (M − H). 1H NMR (300 MHz, CDCl3): δ 0.83 (dd,
6H), 1.39 (m, 1H), 1.58 (m, 2H), 2.59 (br s, 1H), 4.92 (m, 1H), 7.12
(m, 1H), 7.43 (m, 1H) ppm.
1-(4-Bromo-2,3-difluorophenyl)-3-methylbutyl Methanesulfo-
nate (29). 28 (5.5 g, 19.70 mmol) was dissolved in dichloromethane
and the resulting solution then cooled to 0 °C. TEA (8.24 mL, 59.11
mmol) was added followed by methanesulfonyl chloride (2.303 mL,
29.56 mmol), and the reaction mixture was stirred for 2 h at 0 °C.
After this time, the reaction mixture was warmed to ambient
temperature (starting material did not ionize in LC−MS) and stirred
for an additional 12 h at this temperature. At that time, it was washed
successively with 1 N HCl, satd bicarbonate solution, and then brine.
The organic portion was dried over magnesium sulfate, filtered, and
concentrated to give 29 as a yellow oil which was used without further
purification.
2-(3-Chlorophenoxy)-3-fluoro-4-(3-methyl-1-((S)-3-(5-methyl-
2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)-
benzoic Acid (33). 32 (1170 mg, 2.23 mmol) was taken and diluted
with water (15 mL) and MeOH (5 mL). NaOH (891 mg, 22.29
mmol) was added, and the mixture was heated at 80 °C for 12 h. At
this time, the reaction was cooled to ambient temperautre, then taken
up in EtOAc (100 mL), and carefully acidified to pH 5 with
concentrated HCl. The aqeuous portion was washed six times with
EtOAc, and the combined organic layers were dried over magnesium
sulfate, filtered, and concentrated. The residue was taken up in 1.5 mL
of MeOH and purified via reversed-phase chromatography (0−100%
ACN). Clean fractions were collected and after lyophilization afforded
33 (750 mg, 61.9%) as a white solid. LC−MS (method 3): tR = 1.99
1
min, m/z = 544.0 (M + H). H NMR (300 MHz, CD3OD): δ 0.82
(dd, 6H), 1.19 (m, 1H), 1.66−1.95 (br m, 7H), 2.04 (m, 1H), 2.19
(dt, 1H), 2.78 (m, 1H), 3.06 (m, 1H), 3.38 (m, 1H), 3.68 (m, 1H),
4.65 (m, 2H), 6.70 (m, 1H), 6.76 (m, 1H), 6.96 (m, 1H), 7.18 (app t,
1H), 7.32 (br s, 1H), 7.51 (dd, 1H), 7.83 (m, 1H) ppm.
2-(3-Chlorophenoxy)-3-fluoro-4-((R)-3-methyl-1-((S)-3-(5-methyl-
2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)-
benzoic Acid (46). A diastereomeric mixture of 33 (328 mg, 0.60
mmol) was subjected to HPLC preparative chiral separation eluting
with a mixture of hexanes (50%), methanol/ethanol (1:1) (50%), and
diethylamine (0.1%). The title compound 46 eluted first. After
removal of the solvent, the title compound 46 was obtained as an off-
1
white solid (125 mg, 38.1%). H NMR (300 MHz, CD3OD): δ 0.82
1-((3S)-1-(1-(4-Bromo-2,3-difluorophenyl)-3-methylbutyl)-
piperidin-3-yl)-5-methylpyrimidine-2,4(1H,3H)-dione (30). 29 (7 g,
19.60 mmol) and 6 (6.15 g, 29.39 mmol) were combined in
acetonitrile (88 mL), and then Hunig’s base (10.27 mL, 58.79 mmol)
was added. The reaction mixture was heated at 80 °C for 24 h, at
which time LC−MS indicated that the reaction was clean. At this time,
the reaction was cooled to ambient temperature, then diluted with
EtOAC, and washed twice with water and once with brine. The
organic portion was dried over sodium sulfate, filtered, and
concentrated. The resulting oil was purified via silica chromatogaphy
(gradient 0−100% ethyl acetate, 80 g column) to give 30 (1.500 g,
16.27%) as a white solid. LC−MS (method 3): tR = 3.10 min, m/z =
(dd, 6H), 1.22 (br s, 1H), 1.87 (m, 4H), 2.03 (m, 1H), 2.18 (app t,
1H), 2.75 (br t, 1H), 2.95 (br t, 1H), 3.44 (br d, 1H), 3.56 (br d, 1H),
4.50−4.68 (m, 3H), 6.70 (m, 1H), 6.75 (m, 1H), 6.97 (m, 1H), 7.17
(app t, 1H), 7.31 (br s, 1H), 7.50 (dd, 1H), 7.82 (dd, 1H) ppm.
HRMS: m/z 544.2015, calcd 544.2015 (M + H).
2-(3-Chlorophenoxy)-3-fluoro-4-((S)-3-methyl-1-((S)-3-(5-methyl-
2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)-
benzoic Acid (47). A diastereomeric mixture of 33 (328 mg, 0.60
mmol) was subjected to HPLC preparative chiral separation eluting
with a mixture of hexanes (50%), methanol/ethanol (1:1) (50%), and
diethylamine (0.1%). The title compound 47 eluted second. After
removal of the solvent, the title compound 47 was obtained as an off
white solid (125 mg, 38.1%). LC−MS (method 3): tR = 2.0 min, m/z
1
470/472 (M + 1). H NMR (300 MHz, CD3OD): δ 0.93 (dd, 6H),
1.38−1.91(br m, 10H), 1.95−2.33 (br m, 3H), 2.83 (m, 1H), 2.99 (m,
1H), 4.53 (m, 1H), 7.12 (m, 1H), 7.43 (m, 1H), 7.62 (br s, 1H) ppm.
2,3-Difluoro-4-(3-methyl-1-((S)-3-(5-methyl-2,4-dioxo-3,4-dihy-
dropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)benzonitrile (31). 30 (1.5
g, 3.19 mmol), dicyanozinc (1.5 g, 12.76 mmol), and tetrakis-
(triphenylphosphine)palladium(0) (37 mg, 0.03 mmol) were
suspended in acetonitrile (10 mL), and the reaction mixture was
heated under microwave irradiation at 140 °C for 105 min. The
mixture was filtered, and the resulting filtrate was concentrated via
rotary evaporation and then purified via normal-phase chromatography
(0−100% ethyl acetate, 40 g column) to isolate the desired product as
well as unreacted starting material. The presence of 31 (0.895 g,
67.4%) was confirmed by LC−MS with some unreacted starting
material. The material was taken forward without further purification.
LC−MS (method 3): tR = 1.93 min, m/z = 416.0 (M − 1). 1H NMR
(300 MHz, DMSO-d6): δ 0.86 (dd, 6H), 1.32−2.04 (br m, 12H),
2.69−2.96 (br m, 2H), 4.07 (m, 1H), 4.35 (m, 1H), 7.45 (app t, 1H),
7.59 (m, 1H), 7.77 (app t, 1H) 11.20 (s, 1H) ppm.
2-(3-Chlorophenoxy)-3-fluoro-4-(3-methyl-1-((S)-3-(5-methyl-
2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)-
benzonitrile (32). 31 (0.895 g, 2.15 mmol) was dissolved in NMP (2
mL), and 3-chlorophenol (0.359 g, 2.79 mmol) was added in a
microwave vial. Potassium carbonate (0.594 g, 4.30 mmol) was added,
and the reaction mixture heated at 160 °C for 15 min under
microwave irradiation. At this time, the reaction was diluted with
EtOAc (25 mL) and then washed twice with water and once with
brine. The organic portion was dried over magnesium sulfate, filtered,
concentrated, and then purified via normal-phase chromatography (40
g, 0−100% ethyl acetate). After concentration of the appropriate
fractions via rotary evaporation, LC−MS confirmed the presence of 32
(0.800 g, 70.9%). LC−MS (method 3): tR = 2.62 min, m/z = 525.0 (M
+ H).
1
= 545.0 (M + H). H NMR (300 MHz, CD3OD): δ 0.87 (dd, 6H),
1.21 (br s, 1H), 1.87 (m, 4H), 2.02 (m, 1H), 2.19 (app t, 1H), 2.75 (br
t, 1H), 2.95 (br t, 1H), 3.49 (m, 2H), 4.50−4.68 (m, 3H), 6.73 (m,
2H), 6.96 (m, 1H), 7.18 (app t, 1H), 7.32 (br s, 1H), 7.50 (dd, 1H),
7.82 (dd, 1H) ppm. HRMS: m/z 544.2027, calcd 544.2015 (M + H).
2-(3-Chlorophenoxy)-3-fluoro-4-((R)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)propyl)benzoic
Acid (42). Yield: 28 mg (23.3%) as an off-white solid. LC−MS
1
(method 3): tR = 1.75 min, m/z = 516.2 (M + 1). H NMR (300
MHz, CD3OD): δ 0.78 (t, 3H), 1.06 (d, 2H), 1.54−1.86 (br m, 8H),
1.97 (m, 1H), 2.15 (br t, 2H), 2.80 (m, 1H), 2.91 (m, 1H), 3.89 (m,
1H), 4.42 (m, 1H), 6.64−6.75 (m, 2H), 6.92 (m, 1H), 7.15 (t, 1H),
7.27 (dd, 1H) 7.54 (br s, 1H), 7.64 (app d, 1H) ppm. HRMS: m/z
516.1677, calcd 516.1702 (M + H).
2-(3-Chlorophenoxy)-3-fluoro-4-((S)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)propyl)benzoic
Acid (43). Yield: 12 mg (10%) as an off-white solid. LC−MS (method
3): tR = 1.72 min, m/z = 516.2 (M + 1). 1H NMR (300 MHz,
CD3OD): δ 0.75 (t, 3H), 1.47−2.11 (br m, 11H), 2.24 (br t, 1H), 2.76
(m, 1H), 2.96 (m, 1H), 3.82 (m, 1H), 4.47 (m, 1H), 6.70 (m, 2H),
6.92 (m, 1H), 7.13 (m, 1H), 7.28 (m, 1H), 7.51 (br s, 1H), 7.65 (app
d, 1H) ppm. HRMS: m/z 516.1675, calcd 516.1702 (M + H).
2-(3-Chlorophenoxy)-3-fluoro-4-((R)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)benzoic
Acid (44). Yield: 25 mg (23.8%) as an off-white solid. LC−MS
1
(method 3): tR = 1.84 min, m/z = 532.2 (M + H). H NMR (300
MHz, CD3OD): δ 0.77 (br t, 3H), 1.06 (d, 2H), 1.56−1.87 (br m,
8H), 1.97 (m, 1H), 2.17 (br t, 2H), 2.82 (m, 1H), 2.92 (m, 1H), 3.91
(m, 1H), 4.43 (m, 1H), 4.81 (br s, 1H), 6.74 (m, 1H), 6.87 (m, 1H),
7.08 (m, 2H), 7.28 (m, 1H), 7.53 (br s, 1H), 7.66 (br d, 1H) ppm.
HRMS: m/z 530.1853, calcd 530.1859 (M + H).
L
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
2-(3-Chlorophenoxy)-3-fluoro-4-((S)-1-((S)-3-(5-methyl-2,4-
dioxo-3,4-dihydropyrimidin-1(2H)-yl)piperidin-1-yl)butyl)benzoic
Acid (45). Yield: 26 mg (24.7%) as an off-white solid. LC−MS
Prasad for human TMK IC50 values, and Selvi Pradeepan for
coordinating the log D measurements.
1
(method 3): tR = 1.86 min, m/z = 532.2 (M + H). H NMR (300
ABBREVIATIONS
MHz, CD3OD): δ 0.75 (br t, 3H), 1.05 (d, 2H), 1.50−2.09 (br m,
10H), 2.25 (t, 1H), 2.77 (m, 1H), 2.98 (m, 1H), 3.83 (m, 1H), 4.48
(m, 1H), 4.78 (m, 1H), 6.72 (m, 1H), 6.88 (m, 1H), 7.08 (m, 2H),
7.29 (dt, 1H), 7.52 (br s, 1H), 7.66 (dd, 1H) ppm.
■
TMK, thymidylate kinase; MIC, minimum inhibitory concen-
tration
(S)-2-(3-Chlorophenoxy)-3-fluoro-4-((3-(5-methyl-2,4-dioxo-3,4-
dihydropyrimidin-1(2H)-yl)piperidin-1-yl)methyl)benzoic Acid (19).
(S)-2-(3-Chlorophenoxy)-3-fluoro-4-((3-(5-methyl-2,4-dioxo-3,4-di-
hydropyrimidin-1(2H)-yl)piperidin-1-yl)methyl)benzonitrile was
added to ethanol (3 mL) and water (3 mL) followed by NaOH
(32.4 mg, 0.81 mmol) pellets. The suspension was heated under
nitrogen at reflux for 2 days. The solvents were stripped, and any
residual EtOH was chased with water. The residue was then taken up
in water, and the pH was adjusted to ∼5 with 2 N HCl. The aqueous
portion was then extracted three times with EtOAc. The combined
organic layers were dried over MgSO4 and stripped. The crude white
solid was taken up in DMSO and a little acetonitrile and purified via
preparative HPLC eluting with 15−50% acetonitrile with 0.1% TFA.
The fractions containing the major peak were combined and
concentrated on the Rotovap to remove acetonitrile. The remaining
solution was then frozen on dry ice and placed on the lyophilizer to
give the title compound (0.02 g, 50.6%) as a solid. LC−MS (method
3): tR = 1.89 min, m/z = 488 (M + H). 1H NMR (300 MHz,
CD3OD): δ 7.87 (dd, 1H), 7.55 (m, 1H), 7.48 (br s, 1H), 7.27 (t,
1H), 7.06 (dm, 1H); 6.88 (t, 1H), 6.80 (dd, 1H), 4.67 (m, 1H), 4.33
(m, 1H), 3.43 (m, 1H), 3.17 (t, 1H), 2.95 (t, 1H), 2.20 1.80
(overlapping m and s, 6H) ppm. HRMS: m/z 488.1383, calcd
488.1389 (M + H).
REFERENCES
■
(1) Fischbach, M. A.; Walsh, C. T. Antibiotics for Emerging
Pathogens. Science 2009, 325, 1089−1093.
(2) Payne, D. J.; Gwynn, M. N.; Holmes, D. J.; Pompliano, D. L.
Drugs for Bad Bugs: Confronting the Challenges of Antibacterial
Discovery. Nat. Rev. Drug Discovery 2007, 6, 29−40.
(3) Choi, J. Y.; Plummer, M. S.; Starr, J.; Desbonnet, C. R.; Soutter,
H.; Chang, J.; Miller, J. R.; Dillman, K.; Miller, A. A.; Roush, W. R.
Structure Guided Development of Novel Thymidine Mimetics
Targeting Pseudomonas aeruginosa Thymidylate Kinase: From Hit to
Lead Generation. J. Med. Chem. 2012, 55, 852−870.
(4) Keating, T. A.; Newman, J. V.; Olivier, N. B.; Otterson, L. G.;
Andrews, B.; Boriack-Sjodin, P. A.; Breen, J. N.; Doig, P.; Dumas, J.;
Gangl, E.; Green, O. M.; Guler, S. Y.; Hentemann, M. F.; Joseph-
McCarthy, D.; Kawatkar, S.; Kutschke, A.; Loch, J. T.; McKenzie, A.
R.; Pradeepan, S.; Prasad, S.; Martinez-Botella, G. In Vivo Validation
of Thymidylate Kinase (TMK) with a Rationally Designed, Selective
Antibacterial Compound. ACS Chem. Biol. 2012, 7, 1866−1872.
(5) Vanheusden, V.; Van Rompaey, P.; Munier-Lehmann, H.; Pochet,
S.; Herdewijn, P.; Van Calenbergh, S. Thymidine and Thymidine-5′-
O-Monophosphate Analogues as Inhibitors of Mycobacterium tuber-
culosis Thymidylate Kinase. Bioorg. Med. Chem. Lett. 2003, 13, 3045−
3048.
(6) Deville-Bonnea, D.; El Amria, C.; Meyer, P.; Chen, Y.;
Agrofoglio, L. A.; Janin, J. Human and Viral Nucleoside/Nucleotide
Kinases Involved in Antiviral Drug Activation: Structural and Catalytic
Properties. Antiviral Res. 2010, 86, 101−120.
(7) Martinez-Botella, G.; Breen, J. N.; Duffy, J. E.; Dumas, J.; Geng,
B.; Gowers, I. K.; Green, O. M.; Guler, S.; Hentemann, M. F.;
Hernandez-Juan, F. A.; Joseph-McCarthy, D.; Kawatkar, S.; Larsen, N.
A.; Lazari, O.; Loch, J. T.; Macritchie, J. A.; McKenzie, A. R.; Newman,
J. V.; Olivier, N. B.; Otterson, L. G.; Owens, A. P.; Read, J.; Sheppard,
D. W.; Keating, T. A. Discovery of Selective and Potent Inhibitors of
Gram-Positive Bacterial Thymidylate Kinase (TMK). J. Med. Chem.
2012, 55, 10010−10021.
ASSOCIATED CONTENT
* Supporting Information
■
S
Protocols for enzymatic, MIC, protein binding, and log D
assays, protein crystallography methods and statistics tables,
details of animal care and pharmacokinetic experiments, and S.
aureus mouse thigh infection model. This material is available
free of charge via the Internet at http://pubs.acs.org.
Accession Codes
4QG7 (compound 18), 4QGA (19), 4QGF (38), 4QGG (46),
and 4QGH (47).
AUTHOR INFORMATION
Corresponding Author
*E-mail: sameer.kawatkar@astrazeneca.com. Phone: (781) 839
4694.
(8) Martinez-Botella, G.; Loch, J. T.; Green, O. M.; Kawatkar, S. P.;
Olivier, N. B.; Boriack-Sjodin, P. A.; Keating, T. A. Sulfonylpiperidines
as Novel, Antibacterial Inhibitors of Gram-Positive Thymidylate
Kinase (TMK). Bioorg. Med. Chem. Lett. 2013, 23, 169−173.
(9) Pace, C. N. Contribution of the Hydrophobic Effect to Globular
Protein Stability. J. Mol. Biol. 1992, 226, 29−35.
(10) Pace, C. N.; Fu, H.; Fryar, K. L.; Landua, J.; Trevino, S. R.;
Shirley, B. A.; Hendricks, M. M.; Iimura, S.; Gajiwala, K.; Scholtz, J.
M.; Grimsley, G. R. Contribution of Hydrophobic Interactions to
Protein Stability. J. Mol. Biol. 2011, 408, 514−528.
■
Present Addresses
§S.P.K.: AstraZeneca Oncology Innovative Medicines, 35
Gatehouse Dr., Waltham, MA 02451.
∥T.A.K.: ImmunoGen, Inc., 830 Winter St., Waltham, MA
02451.
⊥M.F.H.: Novartis Institutes for BioMedical Research, 250
Massachusetts Ave., Cambridge, MA 02139.
#G.M.-B.: Sage Therapeutics, 215 First St., Cambridge, MA
02141.
(11) Vallone, B.; Miele, A. E.; Vecchini, P.; Chiancone, E.; Brunori,
M. Free Energy of Burying Hydrophobic Residues in the Interface
between Protein Subunits. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 6103−
6107.
(12) Bissantz, C.; Kuhn, B.; Stahl, M. A Medicinal Chemist’s Guide
Notes
to Molecular Interactions. J. Med. Chem. 2010, 53, 5061−5084.
(13) Bohm, H.-J.; Klebe, G. What Can We Learn from Molecular
The authors declare no competing financial interest.
̈
Recognition in Protein-Ligand Complexes for the Design of New
Drugs? Angew. Chem., Int. Ed. Engl. 1996, 35, 2588−2614.
(14) Leung, C. S.; Leung, S. S.; Tirado-Rives, J.; Jorgensen, W. L.
Methyl Effects on Protein-Ligand Binding. J. Med. Chem. 2012, 55,
4489−4500.
(15) Gu, Y.; Kar, T.; Scheiner, S. Fundamental Properties of the
CH···O Interaction: Is It a True Hydrogen Bond? J. Am. Chem. Soc.
1999, 121, 9411−9422.
ACKNOWLEDGMENTS
■
We thank Mahalingam Kannan and the team at Syngene
(Bangalore, India) for chemistry support. We thank AstraZe-
neca Infection colleagues Christina Blinn and Krista Farrington
for in vivo pharmacology support, Nancy DeGrace and
Natascha Bezdenejnih-Snyder for analytical support, Swati
M
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(16) Kollman, P.; McKelvey, J.; Johansson, A.; Rothenberg, S.
Theoretical Studies of Hydrogen-Bonded Dimers. Complexes
Involving HF, H2O, NH3, HCl, H2S, PH3, HCN, HNC, HCP,
CH2NH, H2CS, H2CO, CH4, CF3,H, C2H2, C2H4, C6H6, F−, and
H3O+. J. Am. Chem. Soc. 1975, 97, 955−965.
̌
́ ́ ́
(17) Veljkovic, D. Z.; Janjic, G. V.; Zaric, S. D. Are C-H···O
Interactions Linerar? The Case of Aromatic CH Donors. CrystEng-
Comm 2011, 13, 5005−5010.
(18) Alshahateet, S. F.; Bishop, R.; Craig, D. C.; Scudder, M. L. The
Ether-1,3-Peri Aromatic Hydrogen Interaction. CrystEngComm 2001,
3, 107−110.
(19) Desiraju, G. R. The C−H···O Hydrogen Bond In Crystals: What
Is It? Acc. Chem. Res. 1991, 24, 290−296.
(20) Steiner, T.; Saenger, W. Role of C−H···O Hydrogen Bonds in
the Coordination of Water Molecules. Analysis of Neutron Diffraction
Data. J. Am. Chem. Soc. 1993, 115, 4540−4547.
(21) Taylor, R.; Kennard, O. Crystallographic Evidence for the
Existence of CH···O, CH···N and CH···Cl Hydrogen Bonds. J. Am.
Chem. Soc. 1982, 104, 5063−5070.
(22) Zhou, P.; Zou, J.; Tian, F.; Shang, Z. Fluorine BondingHow
Does It Work in Protein−Ligand Interactions? J. Chem. Inf. Model.
2009, 49, 2344−2355.
(23) Alkorta, I.; Elguero, J.; Roussel, C.; Vanthuyne, N.; Piras, P.
Atropisomerism and Axial Chirality in Heteroaromatic Compounds.
Adv. Heterocycl. Chem. 2012, 105, 1−188.
(24) Shen, Q.; Wang, L.; Zhou, H.; Jiang, H.; Yu, L.; Zeng, S.
Stereoselective Binding of Chiral Drugs to Plasma Proteins. Acta
Pharmacol. Sin. 2013, 34, 998−1006.
(25) Brocks, D. R. Drug Disposition in Three Dimentions: An
Update on Stereoselectivity in Pharmacokinetics. Biopharm. Drug
Dispos. 2006, 27, 387−406.
(26) Fitos, I.; Visy, J.; Simonyi, M. Species-Dependency in Chiral-
Drug Recognition of Serum Albumin Studied by Chromatographic
Methods. J. Biochem. Biophys. Methods 2002, 54, 71−84.
(27) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.;
Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.; Shelley, M.;
Perry, J. K.; Shaw, D. E.; Francis, P.; Shenkin, P. S. Glide: A New
Approach for Rapid, Accurate Docking and Scoring. 1. Method and
Assessment of Docking Accuracy. J. Med. Chem. 2004, 47, 1739−1749.
(28) Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H. S.;
Frye, L. L.; Pollard, W. T.; Banks, J. L. Glide: A New Approach for
Rapid, Accurate Docking and Scoring. 2. Enrichment Factors in
Database Screening. J. Med. Chem. 2004, 47, 1750−1759.
(29) Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen, W.
L. Evaluation and Reparametrization of the OPLS-AA Force Field for
Proteins via Comparison with Accurate Quantum Chemical
Calculations on Peptides. J. Phys. Chem. B 2001, 105, 6474−6487.
N
dx.doi.org/10.1021/jm500463c | J. Med. Chem. XXXX, XXX, XXX−XXX