Structure-guided design, synthesis and biological evaluation of novel DNA ligase inhibitors with in vitro and in vivo anti-staphylococcal activity

Article abstract of DOI:10.1016/j.bmcl.2012.08.094

A series of 2-amino-[1,8]-naphthyridine-3-carboxamides (ANCs) with potent inhibition of bacterial NAD⁺-dependent DNA ligases (LigAs) evolved from a 2,4-diaminopteridine derivative discovered by HTS. The design was guided by several highly resolved X-ray structures of our inhibitors in complex with either Streptococcus pneumoniae or Escherichia coli LigA. The structure-activity-relationship based on the ANC scaffold is discussed. The in-depth characterization of 2-amino-6-bromo-7-(trifluoromethyl)-[1,8]- naphthyridine-3-carboxamide, which displayed promising in vitro (MIC Staphylococcus aureus 1 mg/L) and in vivo anti-staphylococcal activity, is presented.

Full text of DOI:10.1016/j.bmcl.2012.08.094

Bioorganic & Medicinal Chemistry Letters 22 (2012) 6705–6711
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure-guided design, synthesis and biological evaluation of novel DNA
ligase inhibitors with in vitro and in vivo anti-staphylococcal activity
⇑
Jean-Philippe Surivet , Roland Lange, Christian Hubschwerlen, Wolfgang Keck, Jean-Luc Specklin,
Daniel Ritz, Daniel Bur, Hans Locher, Peter Seiler, Daniel Stefan Strasser, Lars Prade, Christopher Kohl,
Christine Schmitt, Gaëlle Chapoux, Eser Ilhan, Nadia Ekambaram, Alcibiade Athanasiou,
Andreja Knezevic, Daniela Sabato, Alain Chambovey, Mika Gaertner, Michel Enderlin, Maria Boehme,
Virginie Sippel, Pierre Wyss
Actelion Pharmaceuticals Ltd, Gewerbestrasse 16, CH-4123 Allschwil, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 2-amino-[1,8]-naphthyridine-3-carboxamides (ANCs) with potent inhibition of bacterial
NAD+-dependent DNA ligases (LigAs) evolved from a 2,4-diaminopteridine derivative discovered by
HTS. The design was guided by several highly resolved X-ray structures of our inhibitors in complex with
either Streptococcus pneumoniae or Escherichia coli LigA. The structure–activity-relationship based on the
ANC scaffold is discussed. The in-depth characterization of 2-amino-6-bromo-7-(trifluoromethyl)-[1,8]-
naphthyridine-3-carboxamide, which displayed promising in vitro (MIC Staphylococcus aureus 1 mg/L)
and in vivo anti-staphylococcal activity, is presented.
Received 18 July 2012
Revised 23 August 2012
Accepted 25 August 2012
Available online 4 September 2012
Keywords:
Antibacterials
DNA ligase
Ó 2012 Elsevier Ltd. All rights reserved.
Selective inhibitors
2-amino-[1,8]-naphthyridine-3-
carboxamides
In vivo active
Over the last seventy years, antibacterial therapies have revolu-
tionized medical practice and have contributed to the increase of
the average lifespan. Unfortunately, the emergence of multidrug
resistant bacteria has come along with the use of antibacterial
drugs. The prevalence of such resistant pathogens has reached such
an alarming level that immediate and drastic actions are urgently
needed to ensure that effective therapeutic options remain avail-
able.1 Beyond any economical and strategic considerations,2 the
scarcity of novel validated biological targets amenable to antibac-
terial therapy and the lack of new chemical entities not subject
to cross-resistance with established anti-infective drugs, make
the discovery and the clinical development of a novel antibiotic
acting through a novel mode of action, one of the most fascinating
scientific challenges.3
little attention despite the potential to reach broad-spectrum
activity.5 LigA isoenzymes are encoded by essential genes that
show a high degree of conservation across a relevant spectrum of
Gram-positive and Gram-negative pathogens.6 The ligation mech-
anism consists first in the adenylation of a lysine in the active site
of LigA by NAD+ with release of nicotinamide mononucleotide.
Then, the adenosine monophosphate (AMP) is transferred from
the LigA-Lys to the 50-end of the 50-PO42ꢀ-terminated DNA strand.
Finally, the phosphodiester bond is formed via the nucleophilic at-
tack of the 30-OH end onto the activated 50-PO42ꢀ, with release of
AMP.7
Bacterial LigA isoenzymes utilize NAD+ as a cofactor and belong
to a distinct phylogenetic cluster strictly separated from the cluster
of mammalian ligases consuming ATP.8 The dissimilarity between
the NAD+-dependent bacterial and the ATP-dependent human DNA
ligases increases the chance to design selective inhibitors of the
bacterial enzyme thereby minimizing the risk of mechanism based
toxicity owing to human ligase inhibition.9
When we initiated our investigations, only two classes of selec-
tive LigA inhibitors displaying target related antibacterial activity
had been reported: one derived from adenosine such as compound
110 and the other belonging to the pyridochromanone family,
exemplified by compound 211 (Fig. 1). Recently, 4-amino-pyr-
DNA ligases are essential for the maintenance of genome integ-
rity as they are indispensable for DNA replication, recombination
2ꢀ
and repair. During these vital processes, 5’-PO4 30-OH polynucle-
otide nicks are formed in double stranded DNA and must be re-
sealed by ligase. NAD+-dependent bacterial DNA ligase (LigA; EC
6.5.1.2) is a validated antibacterial target4 but has so far received
⇑
Corresponding author.
E-mail address: jean-philippe.surivet@actelion.com (J.-P. Surivet).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.08.094

6706
J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
NH2
hydrogen bonds either with the backbone of the enzyme (Leu114)
or amino-acid side chains (Glu113 and Lys290).16
O
O
NH2
O
N
N
F
N
A superposition of compounds 1 and 3 bound to LigA (Fig. 3) re-
vealed that both compounds are covering the same space as the
AMP unit of NAD+, as found in the structure of NAD+ bound to
Enterococcus faecalis LigA (PDB code 1TAE).19 The cyclopentyl unit
of 1 and the 7-methoxy group of 3 are both occupying a part of a
hydrophobic cave which we dubbed ‘tunnel’. This cavity is not
present in huLig 1 since it is blocked by the side chain of Trp742
(PDB code: 1X9N).9,20 Replacement of the N-5 nitrogen of hit 3
by a carbon atom offered a new and promising exit vector in the
direction of the ribose pocket. The 4-amino group of compound 3
interacted with the backbone of LigA (Lys115) through a water
molecule (Fig. 2). Although compound 3 did not inhibit huDHFR,
the removal of this 4-amino group, combined with the replace-
ment of the N-3 nitrogen of 3 by a carbon atom was a preemptive
measure to avoid undesired interactions with enzymes of the fo-
late pathway, as experienced with related molecules.21 We in-
tended to combine features of compound 1 and a modified
version of 3 (compound 5a, Fig. 3) in order to reach simultaneously
the ribose pocket as well as the hydrophobic ‘tunnel’.
O
N
1
N
NH2
OH
OH
2
O
HO
Figure 1. Reported LigA inhibitors 1 and 2, targeting the ATP-binding site.
ido[2,3-d]pyrimidin-5(8H)-ones, displaying moderate Gram-posi-
tive antibacterial activities through selective LigA inhibition, have
also been reported.12
Herein, we report the discovery of a novel class of DNA LigA
inhibitors based on a 2-amino-[1,8]-naphthyridine-3-carboxamide
(ANC) scaffold and our efforts to improve the biological activity
leading to a compound displaying potent in vivo activity against
Staphylococcus aureus.
A HTS campaign against the Escherichia coli LigA resulted in the
identification of compounds 3 and 4 (Table 2). The two hits dis-
played IC50 in the low micromolar range against S. aureus LigA
and in a higher range against the E. coli enzyme.13 Neither the hu-
man DNA ligase 114 (huLig 1) nor the growth of wild-type bacte-
Compound 6 (Scheme 1) inhibited E. coli LigA (IC50 = 25 lM)
and the growth of sensitized E. coli strain in the same range as
compound 3, but inhibited neither huLig 1 nor huDHFR at this con-
centration. We focused our efforts on exploring a suitable substitu-
tion pattern to fill the ribose site, initially defining R3 as CF3.
Unfortunately, enzymatic inhibition measured for more than 100
compounds of formulae 8 and 9 (Scheme 1) on E. coli LigA re-
ria15 was inhibited by compounds
3
and 4. However, LigA
specific inhibitory activity could be shown with a panel of isogenic
‘permeable’ tolC (efflux knock-out) rfaC (truncated LPS) E. coli
D
D
strains with low or high ligation activity. Thus, a particular ligase
mutation (L15F amino acid exchange in ligA251)16 rendered strains
carrying only the ligA251 allele more sensitive toward specific LigA
inhibitors resulting in a >20-fold reduced ligation activity.17 The
specificity of compound 3 for LigA was shown through the signifi-
cant increase of MIC (minimal inhibitory concentration) when
wild-type (wt) LigA (pECLigA) was over-produced in the isogenic
E. coli strain. Compound 4 showed a similar activity pattern as
compound 3, however to a lower extent as a result of its weaker
inhibitory potency against E. coli LigA (Table 1). The better bio-
chemical potency and whole-cell activity of hit 3 over hit 4, com-
mained in a range between 5 and 25 lM (IC50). The introduction
of hydrophobic groups, such as aromatic or hetero-aromatic
groups were usually not tolerated (compounds not shown), with
the exception of phenyl groups substituted with polar functional
groups such as in compound 8a (Scheme 1). The introduction of
aliphatic polar substitutions as in compound 9a (Scheme 1), did
not lead to improved enzymatic inhibitory activities either. The hy-
droxyl groups of the compound 9a did not interact favorably with
the existing H-bond network defined by water molecules and sur-
rounding amino acid side chains in contrast to the hydroxyl groups
of the ribose unit present in AMP.
The X-ray structure of the pyridochromanone 2 bound to trun-
cated E. faecalis LigA (PDB code 3BA8)22, revealed that an amide
moiety strengthened the interaction with the enzyme backbone.
This observation led us to establish the 2-amino-[1,8]- naphthyri-
dine-3-carboxamide (ANC) series 5b (Fig. 3), and further introduce
a substitution on position 6 (R2 substitution) to increase the inter-
actions with the protein. During the course of our program, X-ray
structures of two pyridopyrimidine derivatives bound to LigA have
bined with
0.4 Kcal/mol/non H-atom, calculated using S. aureus LigA inhibition
data) and higher amenability for structural modifications,
a slightly better Ligand Efficiency (LE, 0.54 vs
a
prompted us to select the pteridine derivative 3 as a starting point
for a medicinal chemistry program.18
In parallel, the X-ray crystal structure of hit 3 (Fig. 2) bound
within the AMP binding pocket of a truncated form of Streptococcus
pneumoniae LigA was solved. The major binding interactions are a
p
-stacking interaction with the side chain of Tyr225 and a series of
Table 1
Biological activities of HTS validated hits 3 and 4 on LigAs and engineered E. coli strains
NH2
N
N
NH
N
N
O
N
H
O
N
N
NH2
1
8
NH2
4
3
MIC wtc,f
MIC Ligꢀ
MIC Lig+
a
a
b
d,f
e,f
Compounds
IC50 S. aureus
IC50 E. coli
KD
3
4
2.5
2.5
25
>25
0.8
20
>16
>16
2
8
>16
16
a
b
c
d
e
f
IC50 on LigA (lM).
Equilibrium dissociation constant measured on Biacore using Streptococcus pneumoniae LigA (lM).
E. coli
E. coli
E. coli
D
D
D
tolC,
tolC,
tolC,
D
D
D
rfaC
rfaC, ligA251.
rfaC, ligA251, pECLigA.
MICs expressed in mg/L.

J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
6707
Table 2
Enzymatic inhibitory and antibacterial activities15 of ANC derivatives 5b (see Fig. 3)
a
c,d
e
Compounds
R1
R2
R3
IC50
(lM)
MIC wtb
MIC Ligꢀ
KD (nM)
MICi
S.af
S.pg
H.ih
13a
13b
13c
13d
13e
13f
H
Me
H
H
H
H
H
H
–Cl
Bn
–CF3
–CF3
–CF3
–CF3
–CF3
–CF3
0.6
0.5–0.6
0.5
10
16
4
0.125
0.25
2
1
1
0.5
152
0.125
101
48
8
8
>16
>16
>16
>16
>16
>16
8
0.5
0.5
4
1
4
–(CH2)2OMe
–(CH2)2OH
–CH2NH2
Ph
–(CH2)2–
Me
>16
>16
>16
32
0.25
4
458
431
102
237
16
>16
>16
>16
>16
>16
4
10
2.5
2.5
8
8
17a
17b
–Cl
0.5
107
4
2
4
17c
Me
–Cl
0.8–1
8
0.125
116
8
>16
17d
Me
–Cl
1–2
>16
0.25
118
>16
8
2
N
20a
20b
H
H
Br
Br
CF3
tBu
0.3
0.3–0.5
2
8
0.063
0.25
16
77
1
4
2
8
0.125
2
20c
20d
H
H
Br
Br
0.5
>5
>16
>16
4
8
184
>16
>16
>16
>16
>16
>16
OBn
4312
20e
22a
H
H
Br
>5
>16
0.5
>16
2608
48
>16
4
>16
8
>16
NBoc
Me
CF3
0.5
0.063
0.25
22b
22c
H
H
CF3
CF3
0.4
0.6
8
8
0.125
0.25
89
75
2
4
>16
>16
1
2
Et
22d
H
CF3
0.8-1
>16
1
152
16
>16
8
22e
22f
H
H
CF3
CF3
0.4
2.5
2
0.125
4
83
1
16
16
0.5
16
>16
150
>16
22g
22h
22i
H
H
H
Bn
–CH2Bn
–CH2OH
CF3
CF3
CF3
0.8–1
>10
0.5
16
>16
>16
0.125
32
2
141
1216
389
>16
>16
>16
>16
>16
>16
4
>16
2
OH
N
N
30a
Br
CF3
0.5
4
0.125
43
16
>16
4
30b
Br
CF3
0.7
4
0.125
31
8
8
8
OH
OH
N
N
30c
30d
Br
Br
CF3
CF3
0.7
0.7
8
0.25
0.5
39
57
>16
8
16
8
>16
4
OH
>16
N
30e
Br
CF3
1–2
>16
1
86
>16
>16
8
NH2
a
b
c
d
e
f
E. coli LigA.
E. coli
E .coli
D
D
tolC,
tolC,
D
D
rfaC.
rfaC, ligA251.
All compounds showed MIC >16 on E. coli
Equilibrium dissociation constant measured on Biacore using S. pneumoniae LigA (nM).
S. aureus ATCC 29213.
S. pneumoniae A-216.
Haemophilus influenzae A921.
MICs expressed in mg/L. None of the compounds reported in the table inhibited huLig1 or huDHFR.
DtolC; DrfaC, ligA251, pECLigA except compound 20a (MIC = 8).
g
h
i
been released.23 These inhibitors share structural features with 5b,
but no biological data were published so far. A series of 5-substi-
tuted-7-trifluoro-ANC 13a–f and 5,6-disubstituted-7-chloro-ANC
17a–d were synthesized (Schemes 2 and 3) and evaluated (Table
2).
Sub-micromolar inhibitory activities against E. coli LigA were
measured for compounds 13a and 13b, confirming the positive im-
pact of the carboxamide that forms two H-bonds with the back-
bone of the protein (Fig. 4). Compounds 13a and 13b bearing
either no or just a small substituent in position 5 displayed strong
growth inhibitory activity on sensitized E. coli strains but lacked
inhibitory activities against huDHFR and huLig1 respectively, con-
firming the excellent specificity and selectivity profile already ob-
served for compound 6. The same findings were observed with
several 7-chloro-ANC (17a–d) featuring small R1, which inhibited
LigA as well as a sensitized E. coli strain. Derivative 17b exhibited
measurable antibacterial effects on wild type S. aureus and Hae-
mophilus influenzae (Table 2). By contrast, increasing the size
and/or the polarity of R1 in position 5 caused a decrease of inhibi-
tion on E. coli LigA and sensitized E. coli strain (compounds 13c–f).

6708
J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
again the detrimental impact of any large R1 substitution filling
the ribose pocket, on the antibacterial activities.
The SAR around the 2-amino[1,8]-naphthyridine scaffold was
explored preparing two series of 7-trifluoromethyl-2-amino[1,8]-
naphthyridines. Naphthyridines of formula 8 were obtained from
the intermediate 6 (obtained by chlorination of the known deriva-
tive 724) via a Suzuki–Miyaura cross coupling reaction25 using a
variety of (hetero)aromatic boronic acids. Naphthyridines of for-
mula 9 were prepared via a substitution reaction of compound 6
with secondary amines (Scheme 1). The synthesis of 5-substi-
tuted-7-trifluoromethyl-ANC 13 (Scheme 2) started with the con-
densation of ethyl diamino-acrylate 10 with b-substituted-b-
methoxyvinyl trifluoromethyl ketones 11 affording the pyridines
12. The reduction of the ester moiety with LAH and subsequent oxi-
dation with MnO2 led to the corresponding aromatic aldehydes that
were further condensed with 2-cyanoacetamide in presence of
TMG leading to derivatives 13a–f. In case R1 = CH2NHBoc or
(CH2)2-OBn, a deprotection step was necessary using either HBr in
AcOH or TFA. 5,6-Disubstituted-7-chloro-ANC 20a–e were pre-
Figure 2. Co-crystal structure of HTS hit 3 with truncated S. pneumoniae LigA16
(PDB code: 4GLW).
pared via a condensation reaction of the a-substituted b-ketoesters
14 with ethyl diamino-acrylate 10.26 The resulting 6-amino-pyri-
din-2-ones 15 were refluxed in POCl3 to afford the chloropyridines
16, which were further converted to the ANC 17 using the usual
three ultimate steps (as outlined in Scheme 2). A series of 2-ami-
no-pyridines 19 and 21, bearing various substitutions in position
6, were obtained by condensation of the sodio enolates 18 with
ethyl diamino-acrylate 10 in presence of acetic acid. The regioselec-
tive bromination was performed with NBS, yielding the derivatives
19. Palladium-catalyzed cross coupling reactions, carried out with
various boronic acids, potassium trifluoroborate salts27 or zinc re-
agents28 provided derivatives 21, bearing a variety of substitutions
in position 6. If necessary, any remaining unsaturation was reduced
by catalytic hydrogenation. The derivatives 20 were directly ob-
tained from the intermediate 19 (R = CF3) using a three steps se-
quence (reduction to the alcohol, oxidation to the aldehyde and
condensation with 2-cyanoacetamide). The same sequence was ap-
plied to generate ANC derivatives 22a–i from the esters 21 (Scheme
3). The synthesis of a library of 5-N-substituted ANC derivatives
30a–e (Scheme 4) was achieved through nucleophilic substitution
of the 5-chloro ANC 29 using various primary and secondary
amines. The synthesis of the intermediate 29 was achieved from
the known precursor 23.29 After chlorination in hot POCl3, sequen-
tial displacement with the sodio anion of 4-methoxybenzylalcohol
The two derivatives 20a and 20b, wherein R3 is trifluoro or tert-bu-
tyl respectively, showed potent inhibition of E. coli LigA, low equi-
librium dissociation constant on S. pneumoniae LigA enzyme, and
strong growth inhibition of sensitized E. coli strains. Furthermore,
compounds 20a and 20b showed significant antibacterial activity
on wild-type S. aureus, S. pneumoniae and H. influenzae. Increasing
either the size and/or the polarity of the R3 (compounds 20c–e) led
to decreased activity.
The beneficial effect of the R2 substitution on the antibacterial
activities is obvious by comparing the activities of compounds
13a and 20a. Due to an easier chemical access to the ANC scaffold
(Scheme 3), an extensive investigation around R2 was performed
keeping R3 = CF3. Various small alkyl groups were tolerated, pro-
viding compounds with significant antibacterial activity on wild-
type S. aureus and H. influenzae (compounds 22a–e, Table 2) while
larger substituents such as a benzyl (compound 22g) or a phen-
ethyl (compound 22h) displayed weaker or no measurable antibac-
terial activities. To further extend our exploration of the 6-bromo-
7-trifluoromethyl-ANC scaffold, various N-5 substituted deriva-
tives (30a–e) were synthesized (Scheme 4). While sub-micromolar
inhibitory activities against E. coli LigA were recorded for hydrox-
yalkylamine compounds 30a–d, only modest minimal inhibitory
concentrations were obtained on wild-type S. aureus confirming
design of selective
bacterial LigA
inhibitors
R1
N
R1 NH2
R2
R3
R4
N
R3
N
N
NH2
N
NH2
5a R2 = R4 = H
possible
DHFR inhibitor
improving enzymatic
inhibitory activity
ANC 5b R4 = CONH2
Figure 3. Structure guided design of novel ANCs starting from the superimposition of compounds 1 (yellow bonds) and 3 (magenta bonds).

J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
6709
R1
OH
N
Cl
CO2H
a
8a R1=
b
F3C
R5
N
N
NH2
F3C
N
NH2
F3C
N
N
6
NH2
8
7
R6
N
N
HO
9a R5R6N=
c
OH
N
F3C
N
9
NH2
Scheme 1. Reagents and conditions: Preparation of small librairies of 7-trifluoromethyl-2-amino[1,8]naphthyridines 8 and 9: (a) POCl3, reflux; (b) R1-B(OH)2, (PPh3)4Pd
(5 mol %), K2CO3, aq dioxane, reflux; (c) R5R6’NH, NMP, 120 °C.
R1
O
R1
N
NH2
O
R1 OEt
O b, c, d, e
NH2
CO2Et
NH2
O
CF3
11
a
H2N
O
F3C
N
NH2
F3C
N
10
12
13
R1=
H, CH2N3, -(CH2)2OR (R = Me, Ph), Ph, Me
f
EtO2C
R1
R2
14
R1
NH2
O
R1 OEt
R1 OEt
g
R2
Cl
R2
b, c, d
R2
Cl
O
O
N
N
NH2
O
N
H
NH2
N
NH2
17
15
16
Scheme 2. Reagents and conditions: Synthesis of 5-substituted-7-trifluoromethyl-ANC 13 and 5,6-disubstituted-7-chloro- ANC 17: (a) THF, 80 °C; (b) LAH, Et2O, 0 °C, (when
R1 = CH2N3 then Boc2O; (c) MnO2, DCM, 40 °C; (d) 2-cyanoacetamide, TMG, EtOH, 60 °C; (e) when R1 = (CH2)2OBn: HBr, AcOH, rt (to provide 13d); when R1 = NHBoc, TFA, rt
(leading to 13e); (f) pyridine, EtOH,
D
, (g) POCl3, 90°C.
OEt
NH2
O
Br
R3
O
a, b
Br
R3
O
c, d, e
NaO
R3
N
NH2
19
N
N
N
NH2
18
20
R3= CF3, tBu,
,
NBoc
OBn
,
f
(R3 = CF3)
CO2Et
NH2
NH2
O
R2
R2
F3C
N
22
NH2
F3C
N
21
Scheme 3. Reagents and conditions: Synthesis of 6-substituted-7-trifluoromethyl-ANC 20 and 22: (a) 10, THF, 80 °C; (b) NBS, MeCN, rt; (c) LAH, Et2O, 0 °C; (d) MnO2, DCM,
40°C; (e) 2-cyanoacetamide, TMG, EtOH, 60 °C, rt; (f) R2 = Me: MeZnBr, cat. Pd(PPh3)4, dioxane, 80 °C; R2 = cPr, Bn: BnBF3K, cat. Cl2Pd(dppf)2, Cs2CO3, aq THF, ; R2 = Et, cPent,
–(CH2)2Ph: R’B(OH)2, cat. Pd(PPh3)4, K2CO3, aq dioxane, then H2, cat. PtO2, EA, rt.
D
D
and 3,4-dimethoxybenzylamine and cleavage of both protecting
groups in TFA afforded the intermediate 27. Subsequently, a treat-
ment with POCl3 followed by a bromination using NBS gave rise
to the ester 28. The transformation into naphthyridine 29 was then
performed in three steps as described previously.
oral bioavailability (F = 80%) while subcutaneous administration
(10 mg/kg) to mice resulted in high exposure (see Table 3). As sig-
nificant free concentration was measured in mouse plasma for
the compound 20a, it was evaluated in a neutropenic murine thigh
infection model31 using a quinolone sensitive laboratory strain
descendent of S. aureus DSM 11823 as infecting pathogen (20a;
MIC = 0.5mg/L). When administered at a dose of 100 mg/kg (s.c.;
2h post-infection and once per experiment), 20a provided a net sta-
tic effect 24 h after infection (inoculum size 2 ꢁ 103 cfu i.m.; com-
pound formulated in 5% DMSO/15% solutol HS15 in water). In
vivo efficacy of a LigA inhibitor was already demonstrated on a sim-
ilar aminal model with a metabolically labile adenosine analog.10c
Compound 20a inhibited growth of various strains of S. aureus
even in presence of human serum (Table 4), and displayed some
The physicochemical properties and metabolic stability profile
of the most promising LigA inhibitor 20a were determined in more
detail. The aqueous solubility remained in the low single digit lg/
mL range at all pHs.30 Major human cytochrome P450 isoenzymes
2C9, 2D6 and 3A4 were not inhibited by compound 20a
(IC50 > 50 lM). It was stable in rat microsomes and hepatocytes
with respective intrinsic clearances of 20
l
L/(min⁄mg) and
1.8
l
L/(min⁄106cells). Rat PK performed with compound 20a re-
vealed low clearance, moderate volume of distribution and high

6710
J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
Table 4
Antibacterial activity15 of compound 20a against selected pathogens32
Gram-positive strains
MIC (mg/ Gram-negative strains
L)
MIC (mg/L)
S. aureus A798
1
1
2
>32
8
H. influenzae A921
M. catarrhalis A894
P. aeruginosa ATCC27853 >32
K. pneumoniae T6474
A. baumannii T6474
E. coli ATCC25922
0.125
8
S. aureus ATCC25923
S. aureus ATCC25923a
E. faecium A949
E. faecalis ATCC29212
S. pneumo.ATCC49619
C. difficile ATCCA1148
>32
>32
>32
16
8
S. cerevisiae (yeast) A136 >32
a
Measured in presence of 50% human serum.
potent inhibitors displayed in vitro antibacterial activities on
Gram-positive S. aureus as well as fastidious Gram-negative H.
influenzae. Target specificity of these inhibitors was established
by comparing whole-cell assays using E. coli strains with low or
high DNA LigA activity respectively. The establishment of the
SAR required the development of synthetic pathways leading to
highly functionalized 2-aminopyridines, a process not suited for
high-throughput chemistry. Our investigations demonstrated 1)
that substitution of position 5 did not lead to an increase of binding
affinity and 2) that substituents in positions 6 and 7 have a strong
influence on SAR.
In depth characterization of compound 20a showed that its
good metabolic stability (combined with an acceptable mouse
plasma protein binding) led to sufficient free exposure to ensure
in vivo efficacy in a murine infection model, using a S. aureus strain
as an infecting pathogen. Complete characterization of biochemical
and antibacterial properties of inhibitor 20a will be discussed else-
where. Further investigations of the ANC scaffold will focus on 1)
improving the biochemical potency on various LigA enzymes to ex-
pand the antibacterial spectrum, 2) decreasing resistance frequen-
Figure 4. X-ray structure of ANC 20a bound to E. coli LigA (PDB code: 4GLX).
antibacterial potency against other Gram-positive (E. faecalis) and
RTI pathogens such as S. pneumoniae and Moraxella catarrhalis
but it did not inhibit growth of Enterococcus faecium, and in general
Gram-negative organisms. The IC50 against S. pneumoniae LigA was
two to threefold higher than for S. aureus LigA. Not surprisingly for
Gram-negative strains, the lack of activity can be attributed to poor
permeability and high efflux (see permeable E. coli activity in Table
3). The complete absence of activity against Saccharomyces cerevi-
siae which has an ATP-dependent ligase, confirmed the excellent
selectivity profile of our ANC inhibitors for the bacterial class of
these enzymes.novel class of NAD+-dependent LigA inhibitors, i.e.
functionalized 2-amino-[1,8]-naphthyridine-3-carboxamides, has
been obtained in a structure guided optimization approach starting
from a HTS hit featuring a 2,4-diaminopteridine scaffold. The most
R5
R6
R2
Cl
Cl
N
NH2
O
N
N
NH2
O
CO2Et
R1
Br
CO2Et
NH2
Br
Br
f, g, h
a, e
i
F3C
N
F3C
N
F3C
N
NH2
F3C
N
NH2
29
23 R1 = R2 = OH
24 R1 = R2 = Cl
28
30
a
b
25 R1 = Cl, R2 = PMBO
26 R1 = DMBnNH, R2 = PMBO
27 R1 = NH2, R2 = OH
c
d
Scheme 4. Reagents and conditions: Synthesis of various 5-substituted 6-bromo- ANC 30: (a) POCl3, DMF, 90 °C; (b) 4-MeO-BnOH (PMB-OH), NaH, DMF, 0 °C; (c) 3,4-di-MeO-
BnNH2 (DMBNH2), 100 °C; (d) TFA, 70 °C, 2 h; (e) NBS, MeCN, rt; (f) LAH, Et2O, 0 °C; (g) MnO2, DCM, 40 °C; (h) 2-cyanoacetamide, TMG, EtOH, 60 °C; (i) HNR5R6, Et3N, NMP,
100 °C.
Table 3
PK parameters determined for compound 20a
d
e
f
CLc
Vss
t1/2
Cmax
AUCg
fAUCg
Rat (0.1 mg/kg, iv.a or 1 mg/kg po.b)
Mouse (10 mg/kg, s.c.b)
11.0
0.71
n/a
2.3
0.41h
7.05
0.15
1.22h
84.5
0.021i
0.171h,i
3.89j
n/a
2.9
a
b
c
d
e
f
5% DMSO/28.5% HPbCD in water.
5% DMSO/15% Solutol HS15 in water.
Total plasma clearance (mL/min/kg).
Volume of distribution at steady state (L/kg).
Terminal half-life (h).
Highest observed plasma concentration (lg/mL).
g
h
i
Area under the concentrationꢀtime curve from time 0 to 1 (iv.) or 24 h (po./sc., mg*h/L).
Determined from the po. experiment.
Unbound fraction fu = 0.14.
Unbound fraction fu = 0.046. n/a: not applicable.
j

J.-P. Surivet et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6705–6711
6711
O’Dowd, H.; Wei, Y.; Perola, E.; Parsons, J.; Gross, C. H.; Moody, C. S.; Arends, S.
J. R.; Charifson, P. S. Bioorg. Med. Chem. Lett. 2012, 22, 3699.
13. IC50s were measured and calculated as recommended in: Copeland, R. A.
Evaluation of Enzyme Inhibitors in Drug Discovery; Wiley and Sons: New Jersey,
2005. Chapter 5.
cies 3) optimizing the key physicochemical properties (solubility,
plasma protein binding).
Acknowledgments
14. Chafin, D. R.; Vitolo, J. M.; Henricksen, L. A.; Bambara, R. A.; Hayes, J. J. EMBO J.
2000, 19, 5492.
We would like to acknowledge Dr Naomi Tidten-Luksch for
finalizing and depositing the protein X-ray structures, Markus
Gude (High Throughput Medicinal Chemistry group), and the
members of the the Preformulation and DMPK groups at Actelion
Pharmaceuticals Ltd for the excellent support they provided.
15. Clinical Laboratory and Standards Institute. Methods for Dilution Susceptibility
Test for Bacteria That Grow Aerobically; Approved Standard Seventh Edition,
document. M7-A7, Clinical Laboratory and Standards Institute, Wayne, PA,
USA, 2006.
16. For consistency, E. coli amino acid numbering is used throughout the
publication.
17. (a) Lavesa-Curto, M.; Sayer, H.; Bullard, D.; MacDonald, A.; Wilkinson, A.;
Smith, A.; Bowater, L.; Hemmings, A.; Bowater, R. P. Microbiology 2004, 150,
4171; (b) Dermody, J. J.; Robinson, G. T.; Sternglanz, R. J. Bacteriol. 1979, 139,
701.
References and notes
1. Boucher, H. W.; Talbot, G. H.; Bradley, J. S.; Edwards, J. E.; Gilbert, D.; Rice, L. B.;
Scheld, M.; Spellberg, B.; Bartlett, J. Clin. Infect. Dis. 2009, 48, 1.
2. (a) Katz, M. L.; Mueller, L. V.; Polyakov, M.; Weinstock, S. F. Nat. Biotechnol.
2006, 24, 1529; (b) Kraus, C. N. Curr. Opin. Microbiol. 2008, 11, 434; (c) Projan, S.
J. Curr. Opin. Microbiol. 2003, 6, 427.
18. Gajiwala, K. S.; Pinko, C. Structure 2004, 12, 1449.
19. Structurally related 2,4-diamino-7-dimethylamino-pyrimido[4,5-d]pyrimidine
was also identified as a selective DNA LigA inhibitor at Eli Lilly & Co; see Meier,
T. I.; Yan, D.; Peery, R. B.; McAllister, K. A.; Zook, C.; Peng, S.-B.; Zhao, G. FEBS J.
2008, 275, 5258.
20. Pascal, J. M.; O’Brien, P. J.; Tomkinson, A. E.; Ellenberger, T. Nature 2004, 432,
473.
3. (a) Silver, L. L. Clin. Microbiol. Rev. 2011, 24, 71; (b) Payne, D. J.; Gwynn, M. N.;
Holmes, D. J.; Pompliano, D. L. Nat. Rev. Drug Disc. 2007, 6, 29.
4. (a) Olivera, B. M.; Lehman, I. R. Proc. Natl. Acad. Sci. U.S.A. 1967, 57, 1426; (b)
Lehman, I. R. Science 1974, 186, 790; (c) Gellert, M. Proc. Natl. Acad. Sci. U.S.A.
1967, 57, 148; (d) Weiss, B.; Richardson, C. C. Proc. Natl. Acad. Sci. U.S.A. 1967,
1021, 57; (e) Gefter, M. L.; Becker, A.; Hurwitz, J. Proc. Natl. Acad. Sci. U.S.A. 1967,
58, 240.
5. (a) Swift, R. V.; Amaro, R. E. Expert Opin. Drug Discov. 2009, 4, 1281; (b) Dwivedi,
N.; Dube, D.; Pandey, J.; Singh, B.; Kukshal, V.; Ramachandran, R.; Tripathi, R. P.
Med. Res. Rev. 2008, 28, 545.
6. (a) Gottesman, M. M.; Hicks, M. L.; Gellert, M. J. Mol. Biol. 1973, 77, 531; (b)
Kaczmarek, F. S.; Zaniewski, R. P.; Gootz, T. D.; Danley, D. E.; Mansour, M. N.;
Griffor, M.; Kamath, A. V.; Cronan, M.; Mueller, J.; Sun, D.; Martin, P. K.; Benton,
B.; McDowell, M.; Biek, D.; Schmid, M. B. J. Bacteriol. 2001, 183, 3016; (c) Gong,
C.; Martins, A.; Bongiorno, P.; Glickman, M.; Shuman, S. J. Biol. Chem. 2004, 279,
20594; (d) Petit, M. A.; Ehrlich, S. D. Nucleic Acids Res. 2000, 28, 4642.
7. Tomkinson, A. E.; Vijayakumar, S.; Pascal, J. M.; Ellenberger, T. Chem. Rev. 2006,
106, 687.
21. Song, Y.; Hou, Y.; Gupta, S.; Reed, A.D.; Ghosh, D. US 2007-0082920. Some 2,4-
diamino-pyrido[2,3-d]pyrimidines, claimed to act as NAD+-dependent DNA
Ligase inhibitors in this patent application, displayed in our hands, strong
inhibition of huDHFR and potent antibacterial activity. The latter effect, readily
antagonized by addition of thymidine in the growth medium, was likely due to
bacterial DHFR inhibition.
22. Pinko, C.; Borchardt, A.; Nikulin, V.; Su, Y. available at http://
www.ncbi.nlm.nih.gov/Structure/mmdb/mmdbsrv.cgi?uid=67865.
23. PDB IDs: 3BAA and 3BAB.
24. Ferrarini, P. L.; Mori, C.; Badawneh, M.; Calderone, V.; Calzolari, L.; Loffredo, T.;
Martinotti, E.; Saccomanni, G. Eur. J. Med. Chem. 1998, 33, 383.
25. Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.
26. The condensation reaction between compounds 10 and 14 is sluggish, leading
to ethyl 2-(4-R1-5-R2-6-oxo-1,6-dihydropyrimidin-2-yl)acetate as the major or
sole product.
27. Molander, G. A.; Bernardi, C. R. J. Org. Chem. 2002, 67, 8424.
28. Negishi, E-i.; Hu, Q.; Huang, Z.; Qian, M.; Wang, G. Aldrichim. Acta 2005, 38, 71.
29. Adam, F. M.; Bish, G.; Calo, F.; Carr, C. L.; Castro, N.; Hay, D.; Hodgson, P. B.;
Jones, P.; Knight, C. J.; Paradowski, M.; Parsons, G. C.; Proctor, K. J. W.; Pryde, D.
C.; Rota, F.; Smith, M. C.; Smith, N.; Tran, T. D.; Hitchin, J.; Dixon, R. Org. Process
Res. Dev. 2011, 15, 788.
8. (a) Pascal, J. M. Curr. Opin. Struct. Biol. 2008, 18, 96; (b) Wilkinson, A.; Day, J.;
Bowater, R. Mol. Microbiol. 2001, 40, 1241.
9. Shuman, S. J. Biol. Chem. 2009, 284, 17365.
10. (a) Cavero-Tomas, M.; Gowravaram, M.; Huynh, H.; Ni, H.; Stokes, S. S. WO
2006-0405558.; (b) Stokes, S. S.; Huynh, H.; Gowravaram, M.; Albert, R.;
Cavero-Tomas, M.; Chen, B.; Harang, J.; Loch, J. T., III; Lu, M.; Mullen, G. B.;
Zhao, S.; Liu, C. F.; Mills, S. D. Bioorg. Med. Chem. Lett. 2011, 21, 4556; (c) Mills, S.
D.; Eakin, A. E.; Buurman, E. T.; Newman, J. V.; Gao, N.; Huynh, H.; Johnson, K.
D.; Lahiri, S.; Shapiro, A. B.; Walkup, G. K.; Yang, W.; Stokes, S. S. Antimicrob.
Agents Chemother. 2011, 1088, 55.
11. Broetz-Oesterhelt, H.; Knezevic, I.; Bartel, S.; Lampe, T.; Warnecke-Eberz, U.;
Ziegelbauer, K.; Haebich, D.; Labischinski, H. J. Biol. Chem. 2003, 278, 39435.
12. (a) Gu, W.; Wang, T.; Maltais, F.; Ledford, B.; Kennedy, J.; Wei, Y.; Gross, C. H.;
Parsons, J.; Duncan, L.; Arends, R.; Cameron, M.; Perola, E.; Green, J.; Charifson,
P. Bioorg. Med. Chem. Lett. 2012, 22, 3693; (b) Wang, T.; Duncan, L.; Gu, W.;
30. Thermodynamic solubilities were measured after 24 h equilibration in the
studied medium.
31. Andes, D.; Craig, W. A. Int. J. Antimicrob. Agents 2002, 19, 261.
32. S. aureus: Staphylococcus aureus; E. faecium: Enterococcus faecium; E. faecalis:
Enterococcus faecalis; S. pneumoniae or S. pneumo: Streptococcus pneumoniae; C.
difficile: Clostridium difficile; K. pneumoniae: Klebsiella pneumoniae; M.
catarrhalis: Moraxella catarrhalis; P. aeruginosa: Pseudomonas aeruginosa;
E. coli: Escherichia coli; A. baumannii: Acinetobacter Baumannii; S. cerevisiae:
Saccharomyces cerevisiae.