Sulfonylpiperidines as novel, antibacterial inhibitors of Gram-positive thymidylate kinase (TMK)

Article abstract of DOI:10.1016/j.bmcl.2012.10.128

Thymidylate kinase (TMK) is an essential enzyme for DNA synthesis in bacteria, phosphorylating deoxythymidine monophosphate (dTMP) to deoxythymidine diphosphate (dTDP), and thus is a potential new antibacterial drug target. Previously, we have described the first potent and selective inhibitors of Gram-positive TMK, leading to in vivo validation of the target. Here, a structure-guided design approach based on the initial series led to the discovery of novel sulfonylpiperidine inhibitors of TMK. Formation of hydrogen bonds with Arg48 in Staphylococcus aureus TMK was key to obtaining excellent enzyme affinity, as verified by protein crystallography. Replacement of a methylene linker in the series by a sulfonamide was accomplished with retention of binding conformation. Further optimization of log D yielded phenol derivative 11, a potent inhibitor of TMK showing excellent MICs against a broad spectrum of Gram-positive bacteria and >10⁵ selectivity versus the human TMK homologue.

Full text of DOI:10.1016/j.bmcl.2012.10.128

Bioorganic & Medicinal Chemistry Letters 23 (2013) 169–173
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Sulfonylpiperidines as novel, antibacterial inhibitors of Gram-positive
thymidylate kinase (TMK)
Gabriel Martínez-Botella a, , James T. Loch a, Oluyinka M. Green a, Sameer P. Kawatkar a,
⇑
,
Nelson B. Olivier b, P. Ann Boriack-Sjodin a,b, Thomas A. Keating a,
⇑
a AstraZeneca Infection Innovative Medicines, 35 Gatehouse Drive, Waltham, MA 02451, USA
b AstraZeneca Discovery Sciences, 35 Gatehouse Drive, Waltham, MA 02451, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Thymidylate kinase (TMK) is an essential enzyme for DNA synthesis in bacteria, phosphorylating deoxy-
thymidine monophosphate (dTMP) to deoxythymidine diphosphate (dTDP), and thus is a potential new
antibacterial drug target. Previously, we have described the first potent and selective inhibitors of Gram-
positive TMK, leading to in vivo validation of the target. Here, a structure-guided design approach based
on the initial series led to the discovery of novel sulfonylpiperidine inhibitors of TMK. Formation of
hydrogen bonds with Arg48 in Staphylococcus aureus TMK was key to obtaining excellent enzyme affinity,
as verified by protein crystallography. Replacement of a methylene linker in the series by a sulfonamide
was accomplished with retention of binding conformation. Further optimization of logD yielded phenol
derivative 11, a potent inhibitor of TMK showing excellent MICs against a broad spectrum of Gram-posi-
tive bacteria and >105 selectivity versus the human TMK homologue.
Received 14 September 2012
Revised 25 October 2012
Accepted 29 October 2012
Available online 5 November 2012
Keywords:
Thymidylate kinase
TMK
Nucleotide kinase
Antibacterials
Ó 2012 Elsevier Ltd. All rights reserved.
Antibiotics
Structure-based drug discovery
Thymidylate kinase (TMK) is an essential enzyme that catalyzes
the ATP-mediated phosphorylation of deoxythymidine monophos-
phate (dTMP) to deoxythymidine diphosphate (dTDP). This is a
necessary step in the biosynthesis of deoxythymidine triphosphate
(dTTP) for DNA synthesis since dTTP and dTDP cannot be imported
into the cell nor biosynthesized through an alternate path. There-
fore TMK is considered an attractive potential target for antibacte-
rial therapy. There have been several prior reports of novel TMK
inhibitors targeting the enzymes from Mycobacterium tuberculosis,1
Pseudomonas aeruginosa,2 and Bacillus anthracis.3 However, pro-
gressing biochemical inhibitors to whole-cell activity and then to
in vivo efficacy has been difficult for TMK, as for all novel antibac-
terial targets. We have recently reported the first highly potent
inhibitors of bacterial Gram-positive TMK4 and in vivo validation
of TMK as an antibacterial target against Staphylococcus aureus in
a mouse model of infection.5
The crystal structure of benzylpiperidine 1 bound to S. aureus
TMK (Fig. 1)4 was used as the starting point to embark on the dis-
covery of a second series of related piperidylthymine-containing
inhibitors. We wished to explore new linkers connecting the mid-
dle rings that could enable the turn required for binding, and in
this way provide broader physical property options and perhaps
a shorter synthetic route. Molecular modeling examined replacing
the simple methylene carbon connecting the piperidine and phen-
oxy rings with a quaternary carbon (both di-alkyl and spiro), an
ethylene linker, and a sulfonamide. Modeling results did not favor
ethylene, and quaternary centers did not show exceptional prom-
ise over the tertiary carbons explored previously.4 However, sul-
fonamides did appear to be a suitable and accessible replacement
(Fig. 1). The synthesis of sulfonylpiperidines 2 and 3 were carried
out easily by coupling commercially available 3-phenoxyben-
zene-1-sulfonyl chlorides with (S)-piperidin-3-ylthymine (V).4
The resulting sulfonylpiperidines 2 and 3 validate the concept by
showing modest binding affinity, 21 and 0.29 lM respectively, to
Streptococcus pneumoniae TMK (Table 1). The much higher affinity
of compound 3 can be explained by the very lipophilic nature of
the dichlorophenoxy ring.4
From this initial success in linker modification, we sought to
build greater potency. We have previously shown that the
highly-conserved residue Arg48 provides a key interaction in the
Gram-positive binding pocket and can be targeted by placing a
Abbreviations: TMK, thymidylate kinase; dTMP, deoxythymidine monophos-
phate; dTDP, deoxythymidine diphosphate; dTTP, deoxythymidine triphosphate;
MIC, minimum inhibitory concentration.
⇑
Corresponding authors. Tel.: +1 617 299 8375; fax: +1 617 299 8379 (G.M.-B.);
tel.: +1 781 839 4420; fax: +1 781 839 4600 (T.A.K.).
E-mail addresses: gabriel@sagerx.com (G. Martínez-Botella), thomas.keating@
astrazeneca.com (T.A. Keating).
Present address: Sage Therapeutics, 215 First Street, Cambridge, MA 02141, USA.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.128

170
G. Martínez-Botella et al. / Bioorg. Med. Chem. Lett. 23 (2013) 169–173
Arg70
Ser97
Phe66
Arg48
Gln101
H
H
O
N
O
O
N
O
H
H
O O
O
N
O
S
N
N
N
(1)
CF3
CF3
Figure 1. X-ray structure of benzylpiperidine 14,5 (orange, left, PDB ID 4HEJ) bound to S. aureus TMK and a model of a corresponding sulfonylpiperidine scaffold (green, right).
Replacement of the methylene linker with a sulfonamide is predicted to enable a very similar turn geometry to preserve the compound binding orientation and protein
contacts.
carboxylic acid at C-4 of the middle phenyl ring.4 This arginine is
present in Staphylococcus, Streptococcus, and Enterococcus enzymes,
as well as Escherichia coli, but not in the human ortholog.5 Com-
pounds 4–6 were synthesized by the route described below
(Scheme 1). The key step in the synthesis was the preparation of
sulfonyl chloride VIII via diazotization,6 followed by straightfor-
ward sulfonylation of piperidinylthymine V. Compounds 4–6
showed more than a 100-fold improvement in binding affinity
against both S. aureus and S. pneumoniae TMK. Compound 57 was
crystallized8 with S. aureus TMK and revealed the formation of
two clear hydrogen bonds between the carboxylate and Arg48
and also confirmed the binding mode predicted for the scaffold
(Figs. 1 and 2A). Unfortunately the minimum inhibitory concentra-
tion (MIC) for bacterial growth suppression by compounds 4–6
was poor against both S. aureus and S. pneumoniae, which may be
due to poor cell membrane penetration resulting from very low
logDs (Table 1). In order to validate this hypothesis, amide 7 was
tested and showed an improvement in the MIC against S. pneumo-
niae with a log D = 2. However, 7 showed relatively poor inhibition
of S. aureus TMK (IC50 = 784 nM).
ment in MICs for both S. aureus and S. pneumoniae with logDs
in a more optimal range of 1.6–2.4. Since it is recognized that
phenols can introduce toxicological risk, we synthesized
hydroxymethyl analog 18. This compound, though, lost signifi-
cant enzymatic and whole-cell potency.
Compound 11 was more broadly profiled against pathogens S.
aureus, S. pneumoniae, Streptococcus pyogenes, Staphylococcus epide-
rmidis and Enterococcus spp. and showed potent Gram-positive
activity (Table 2). MIC90 values were obtained for small (ꢀ20
strain) panels of the first three organisms. These panels encom-
passed a range of sensitive and methicillin- and quinolone-resis-
tant strains. The MIC90 values did not shift appreciably from
single-strain MICs. This indicates a lack of pre-existing resistance
and a high degree of target conservation, and demonstrates the va-
lue of a novel antibacterial target and mechanism. Compound 11
was also tested in a human TMK IC50 assay and showed excellent
selectivity (>105) against the human ortholog. Finally, 11 showed
no activity against eukaryotic cells, including human A549 (lung
epithelial adenocarcinoma) and yeast, nor against Gram-negative
bacteria (Table 2). This is not surprising; the selectivity of analogs
of parent scaffold 1 has been examined previously,5 and we simi-
larly conclude here that the structural differences among the
Gram-positive, Gram-negative, and human TMK orthologs make
it difficult to achieve the broadest possible spectrum from this
new series of inhibitors, although this also results in little unde-
sired activity against the human enzyme. The physical properties
of 11 are also in Table 2; aqueous solubility is good, but as with
the original series,5 plasma protein binding is high. In vitro clear-
ance through microsomes or heptatocytes is moderate to high
(40% to 75% of hepatic blood flow). Finally, there is no evidence
of inhibition of the most prevalent cytochrome P450 enzymes
(1A2, 2C9, 2C19, 2D6, 3A4) which could lead to drug–drug interac-
tions, and only weak inhibition of the hERG ion channel. However,
because 11 possesses a phenol/masked catechol, an investigation
into the reactive metabolite liability of the series is required before
further progression.
Replacement of the carboxylic acid with another group com-
petent for interaction with Arg48 was explored by modeling.
While isosteric tetrazoles did not appear optimal, phenols were
chosen to provide a good balance between logD and ability to
interact specifically with Arg48.9 Compound 10 was prepared
as shown in Scheme 1 and showed promising IC50s against both
S. aureus and S. pneumoniae enzymes, 255 and 14 nM, respec-
tively.
A a >20-fold
rapid survey of phenoxy rings4 led to
improvement in binding affinity. The potent inhibition exhibited
by 11 (Table 1) yields a very attractive calculated ligand effi-
ciency (LE; [À1.4LogKi]/[#heavy atoms]) of 0.41, as compared
to an LE of 0.24–0.28 for the original lead 3. Compound 1110
was crystallized11 with S. aureus TMK and showed the formation
of one clear hydrogen bond between the phenolic group and the
Arg48 sidechain with the overall binding mode retained
(Fig. 2B). Compounds 11–17 also showed a >10-fold improve-

G. Martínez-Botella et al. / Bioorg. Med. Chem. Lett. 23 (2013) 169–173
171
Table 1
Enzyme inhibition (IC50) and antibacterial activity (MIC) of sulfonylpiperidine compounds versus Gram-positive pathogens S. pneumoniae and S. aureus12,13
H
O
N
O
O
O
N
O
S
N
R2
R1
Compound
R1
H
R2
TMK IC50 (nM)
Sau
LogD
MIC (l
g/mL)13
Spn
Spn
>64
SauMSQS/SauMRQR
>64/ND
2
Ph
20900
81500
2.3
Cl
Cl
Cl
3
H
292
2780
>3.5
4
4/4
Cl
4
5
6
CO2H
CO2H
CO2H
<6
6
23
174
2
À0.9
À1.6
À1.3
4
>64/>64
>64/>64
>64/>64
16
2
Cl
1
Cl
Cl
Cl
7
8
CONH2
CN
10
23
784
530
2.0
1
2
2
>64/>64
>64/>64
9
OMe
OH
6
358
255
2.7
1.5
2
>64/16
16/16
10
Ph
14
0.5
Cl
11
12
OH
OH
0.5
4
0.5
57
2.1
1.6
0.03
0.25
0.25/0.5
16/16
F
Cl
Cl
13
14
OH
OH
8
18
3
2.1
2.4
0.5
1/2
18
0.25
1/0.5
Br
15
16
OH
OH
0.7
0.7
2
2.3
2.3
0.1
1/0.5
1/1
Cl
12
0.05
F
CF3
Cl
17
18
OH
3
64
2.3
2.7
0.1
2
2/2
CH2OH
28
610
32/64

172
G. Martínez-Botella et al. / Bioorg. Med. Chem. Lett. 23 (2013) 169–173
Figure 2. X-ray crystal structures of sulfonylpiperidine compounds bound to S. aureus TMK, illustrating retention of binding conformation from the parent benzylpiperidines
and clear interactions with the highly conserved residue Arg48. (A) Sulfonylpiperidine 5 at 1.9 Å (PDB ID 4HLC). The carboxylic acid shows a bidentate interaction. (B)
Sulfonylpiperidine 11 at 2.0 Å (PDB ID 4HLD). The phenolic group shows a monodentate interaction with Arg48.
Cl
O
(a)
O
(b)
O
+
HO
NO2
B(OH)2
Cl
O
NH2
Cl
O
SO2Cl
(I)
(II)
(III)
(IV)
O
N
O
O
HN
O
HN
HN
O
N
O
N
(c)
O
O
HO
O
(d)
N
S
O
N
HN
Cl
Cl
S
O
O
O
(8)
(11)
(V)
(a) Cu(OAc)2, pyridine, DCE, 50°C; Fe, NH4Cl, MeOH, H2O, 70°C; (b) NaNO2, AcOH, CuCl, SO2, -10°C;
(c) V, DIEA, DCE, 0°C to RT; (d) Py.HCl, 165°C
F
Cl
Cl
MeO2C
O
Cl
MeO2C
O
HO2C
(e)
(f)
+
NO2
OH
NH2
SO2Cl
(VI)
(II)
(VII)
(VIII)
O
N
O
N
HN
HN
O
O
Cl
MeO2C
O
Cl
O
ROC
(g)
(h)
N
N
S
S
O
O
O O
(IX)
(6) R = OH
(7)
R = NH2
(e) (i) SOCl2, MeOH; (ii) CuI, K2CO3; (iii) Fe, NH4Cl; (f) NaNO2, AcOH, CuCl, SO2, -10°C;
(g) V, TEA, DCE, 0°C to RT; (h) LiOH R = OH; NH3 R = NH2
Scheme 1. Synthesis of sulfonylpiperidine scaffold analogs.
In this report a novel series of selective, rationally-designed
antibacterial compounds targeting TMK have been described. The
change of a benzylpiperidine present in the original inhibitor ser-
ies4 to a sulfonylpiperidine scaffold resulted in potent inhibitors
of TMK retaining a similar binding mode. Conserved Arg48 is a
key residue that drives potency improvements when hydrogen
bonding was achieved using either carboxylic acid or hydroxyl
groups. Optimal cell permeability for these novel agents was

G. Martínez-Botella et al. / Bioorg. Med. Chem. Lett. 23 (2013) 169–173
173
Table 2
Medicines, AstraZeneca Boston, for determination of IC50 and MIC
values, and the Drug Metabolism & Pharmacokinetics Department,
AstraZeneca Boston, for physical property determinations.
Enzyme inhibition, microbiological, and physio-chemical profile of 11
Enzyme
IC50 (nM)
S. pneumoniae TMK
S. aureus TMK
E. coli TMK
0.5
0.5
4500
>200000
References and notes
1. Vanheusden, V.; Van Rompaey, P.; Munier-Lehmann, H.; Pochet, S.; Herdewijn,
P.; Van Calenbergh, S. Bioorg. Med. Chem. Lett. 2003, 13, 3045.
2. 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. J. Med. Chem. 2012, 55, 852.
3. Byun, Y.; Vogel, S. R.; Phipps, A. J.; Carnrot, C.; Eriksson, S.; Tiwari, R.; Tjarks, W.
Nucleosides Nucleotides Nucl. 2008, 27(3), 244.
4. Martínez-Botella, G.; Breen, J. N.; Duffy, J. E. S.; 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. J. Med. Chem. doi: http://dx.doi.org/10.1021/
jm3011806, published online: Oct 8, 2012.
5. 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.; Martínez-Botella, G. ACS
Chem. Biol. 2012, 7, 1866.
Human TMK
Organism13
MIC (lg/mL)
S. aureus (MIC90, n = 23)a
S. pneumoniae (MIC90, n = 21)a
Streptococcus pyogenes (MIC90, n = 21)a
Streptococcus agalactiae
Staphylococcus epidermidis
Staphylococcus haemolyticus
Staphylococcus lugdunensis
Enterococcus faecium (LRE)
Enterococcus faecalis (VRE)
Escherichia coli
Haemophilus influenzae
Pseudomonas aeruginosa
Klebsiella pneumoniae
Candida albicans
Human A549
0.5
0.06
0.5
0.5
2
0.03
0.016
0.13
0.5
>64
>64
>64
>64
>64
>64
>64
6. Hoffman, R. V. Org. Synth. 1981, 60, 121.
7. Compound 5: LC–MS (water (A)/acetonitrile (B)/formic acid 0.1% from 5% to 95%
B
1.8
in 2.5 min, flow 1 mL/min, Acquity HSS T3 C18 column (2.1 Â 50 mm,
Sheep erythrocyte lysis
l
m)) tR = 1.15 min, m/z = 500.5 (M+1), 498.5 (MÀ1). 1H NMR (300 MHz,
Physical properties and selectivity
DMSO-d6): d 1.73 (s, 3H), 1.79–1.84 (m, 2H), 2.25 (s, 3H), 2.29–2.34 (m, 2H),
3.15–3.20 (m, 2H), 3.51–3.56 (m, 2H), 3.79 (s, 3H), 4.34–4.37 (m, 1H), 6.84 (d,
J = 7.77 Hz, 1H), 6.89 (s, 1H), 7.00 (d, J = 8.46 Hz, 1H), 7.17 (s, 1H), 7.24 (t,
J = 7.89 Hz, 1H), 7.53 (s, 1H), 7.60 (d, J = 8.10 Hz, 1H), 8.03 (d, J = 8.16 Hz, 1H),
11.29 (m, 1H). Purity >95%.
Molecular weight (g/mol)
Equilibrium solubility pH 7.4 (
LogD
Plasma percent free (human)
Rat microsome CLint (
Rat hepatocyte CLint (
492
193
2.1
1
27
45
lM)
8. X-ray coordinates for 5 bound to S. aureus TMK are deposited in the RCSB
Protein Data Bank; PDB ID code: 4HLC.
l
l
L/min/mg)
L/min/106 cells)
9. Henriksen, A.; Smith, A. T.; Gajhede, M. J. Biol. Chem. 1999, 274, 35005.
10. Compound 11: LC–MS (water (A)/acetonitrile (B)/formic acid 0.1% from 5% to
95% B in 2.5 min, flow 1 mL/min, Acquity HSS T3 C18 column (2.1 Â 50 mm,
CYP 1A2, 2C9, 2C19, 2D6, 3A4 (IC50
hERG (IC50 M)
,
lM)
>20
27
,
l
1.8
l
m)) tR = 1.26 min, m/z = 492.4 (M+1), 490.3 (MÀ1). 1H NMR (400 MHz,
a
MIC90 = MIC of 90% of n strains tested (MIC90’s for levofloxacin in the panels
CD3OD): d 1.67–1.71 (m, 1H), 1.80 (d, J = 3.20 Hz, 1H), 1.86 (s, 3H), 1.88–1.93
(m, 2H), 2.38–2.45 (m, 1H), 2.57 (t, J = 10.84 Hz, 1H), 3.64–3.66 (m, 1H), 3.73–
3.76 (m, 1H), 4.45–4.50 (m, 1H), 6.88–6.91 (m, 1H), 6.97 (t, J = 2.12 Hz, 1H),
7.09–7.12 (m, 1H), 7.15 (d, J = 8.56 Hz, 1H), 7.31 (d, J = 8.20 Hz, 1H), 7.33–7.35
(m, 1H), 7.48 (s, 1H), 7.54 (dd, J = 2.24, 8.52 Hz, 1H). Purity >95%.
11. X-ray coordinates for 11 bound to S. aureus TMK are deposited in the RCSB
Protein Data Bank; PDB ID code: 4HLD.
were 16, 2, and 0.25 lg/mL for S. aureus, S. pneumoniae, and S. pyogenes, respec-
tively). All other MICs are of single sentinel isolates.12,13
achieved with the phenol-containing sulfonylpiperidines and
excellent MICs were observed across Gram-positive bacteria spe-
cies as well as excellent selectivity against the human TMK
homologue.
12. Methods for biochemical activity against S. aureus and S. pneumoniae TMK,
minimum inhibitory concentration (MIC) determination, and physical property
measurements can be found in Ref. 5.
13. Bacterial strains from AstraZeneca collection: Streptoccocus pneumoniae
ARC548; Streptococcus pyogenes ARC838; Staphylococcus aureus ARC516
methicillin (MS) and quinolone (QS) sensitive; Staphylococcus aureus ARC517
methicillin (MR) and quinolone (QR) resistant; Staphylococcus epidermidis
ARC323; Enterococcus faecium ARC1239 linezolid-resistant (LRE); Enterococcus
faecalis ARC1618 vancomycin-resistant (VRE); Escherichia coli ARC523;
Haemophilus influenzae ARC446; Pseudomonas aeruginosa ARC545; Klebsiella
pneumoniae ARC1865.
Acknowledgments
We thank Dr. Mahalingam Kannan and team at Syngene
(Bangalore, India) for chemistry support. We also thank the follow-
ing groups: the Bioscience Department of Infection Innovative