Design, synthesis and biological evaluation of potent NAD +-dependent DNA ligase inhibitors as potential antibacterial agents. Part I: Aminoalkoxypyrimidine carboxamides

Article abstract of DOI:10.1016/j.bmcl.2012.04.037

A series of 2,6-disubstituted aminoalkoxypyrimidine carboxamides (AAPCs) with potent inhibition of bacterial NAD⁺-dependent DNA ligase was discovered through the use of structure-guided design. Two subsites in the NAD⁺-binding pocket were explored to modulate enzyme inhibitory potency: a hydrophobic selectivity region was explored through a series of 2-alkoxy substituents while the sugar (ribose) binding region of NAD⁺ was explored via 6-alkoxy substituents.

Full text of DOI:10.1016/j.bmcl.2012.04.037

Bioorganic & Medicinal Chemistry Letters 22 (2012) 3693–3698
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis and biological evaluation of potent NAD+-dependent DNA
ligase inhibitors as potential antibacterial agents. Part I:
Aminoalkoxypyrimidine carboxamides
⇑
Wenxin Gu , Tiansheng Wang, Francois Maltais, Brian Ledford, Joseph Kennedy, Yunyi Wei,
Christian H. Gross, Jonathan Parsons, Leonard Duncan, S.J. Ryan Arends, Cameron Moody,
Emanuele Perola, Jeremy Green, Paul S. Charifson
Vertex Pharmaceuticals, Inc., 130 Waverly St., Cambridge, MA 02139, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 17 April 2012
A series of 2,6-disubstituted aminoalkoxypyrimidine carboxamides (AAPCs) with potent inhibition of
bacterial NAD+-dependent DNA ligase was discovered through the use of structure-guided design. Two
subsites in the NAD+-binding pocket were explored to modulate enzyme inhibitory potency: a hydropho-
bic selectivity region was explored through a series of 2-alkoxy substituents while the sugar (ribose)
binding region of NAD+ was explored via 6-alkoxy substituents.
Keywords:
NAD+-DNA ligase
Antibacterial
Ó 2012 Elsevier Ltd. All rights reserved.
Aminoalkoxypyrimidine carboxamide
Crystal structures
MIC
DNA ligase was discovered in 1967 by the Gellert, Lehman,
Richardson, and Hurwitz laboratories.1 DNA ligase functions in
both DNA repair and DNA replication, and is essential for repairing
nicks during DNA replication and recombination. The mechanism
of this enzyme involves the formation of covalent phosphodiester
bonds between the 30 hydroxyl end of one nucleotide and the 50
phosphate end of another. There are three distinct steps involved
in this overall process: (1) adenylation of a lysine residue in the ac-
tive site of the enzyme with the release of pyrophosphate, (2)
transfer of AMP to the 50 phosphate of the nicked DNA strand,
and (3) attack on the AMP–DNA bond by the 30-hydroxyl of the
nicked DNA sealing the phosphate backbone and releasing AMP.1a
DNA ligases are classified as either ATP-dependent or NAD+-
dependent, depending upon whether the required cofactor is ATP
or NAD. All known eukaryotic DNA ligases are ATP-dependent
while the bacterial DNA ligases are NAD+-dependent.2 These bacte-
rial NAD+-dependent DNA ligases (LigA) are broadly conserved
across bacteria with Escherichia coli LigA as the prototype of this
family.3 X-ray crystal structures of ATP- and NAD+-dependent
DNA ligases demonstrate significant differences in their substrate
binding sites.4–6 The narrow phylogenetic distribution, unique sub-
strate specificity, and distinctive domain structure of LigA com-
pared with ATP-dependent human DNA ligases suggest that
bacterial NAD ligases might be desirable targets for the develop-
ment of new and selective antibacterial drugs. In particular, a
hydrophobic ‘tunnel’ is present in all bacterial LigA domains that
have been crystallized to date. By contrast, there is no such corre-
sponding region present in the adenosine-binding pockets of Hu-
Lig1,6 ChVLig,7 and other eukaroytic ATP-dependent DNA
ligases.3 This situation invites the design of C2-substituted analogs
of adenosine nucleotides and NAD+ as selective inhibitors of LigA.3
Recently, there have been a series of reports describing small mol-
ecule competitive inhibitors of bacterial LigA, discovered either by
experimental high-throughput screening8–10 or by virtual screen-
ing through computational docking of ligands into LigA crystal
structures.11,12 Very recently, a first example of in vivo validation
of LigA as an antibacterial target was reported.13 We report herein
our discovery of novel 2,6-disubstituted aminoalkoxypyrimidine
carboxamides as potent and selective inhibitors of NAD+-depen-
dent DNA ligases from various bacteria.
The initial design began with a structural analysis of published
inhibitors in complex with bacterial LigA.14
Figure 1 illustrates the superposition of an adenosine analog
from AstraZeneca15 and a pyridochromanone analog reported from
Bayer16 in the active site of LigA. We determined the X-ray
structures of compounds 1 and 2 ( Fig. 2) in the course of this
study. Analysis of this overlay suggests that the aminopyrimidine
moiety of 1 binds to the protein through the same hydrogen bond-
ing pattern as the aminopyridine of 2. Additionally, the 2-alkoxy
moiety of 1 and the phenyl portion of the chromanone ring of 2
both occupy the hydrophobic site of LigA. The 2,6-disubstituted
⇑
Corresponding author.
E-mail address: wenxin_gu@vrtx.com (W. Gu).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.04.037

3694
W. Gu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3693–3698
Figure 1. Superimposition of X-ray structures of 1 and 2 in the active site of Enterococcus faecalis LigA.
NH2
as sulfonamide and amidine were explored as well (Fig. 3, synthe-
N
N
ses not shown).
N
O
O
NH2
O
Further efforts directed towards filling both the hydrophobic (2-
substituents) and sugar sites built on SAR from the 2-position
exploration. As shown in Scheme 2, condensation of 2-(bis(methyl-
thio)methylene)malononitrile, 7 and cyanamide under basic condi-
tions generated the desired pyrimidine, which was followed by
bromination to afford compound 8. The hydrophobic 2-alkoxy
groups were then introduced by nucleophilic displacement of the
bromide with various primary alcohols under basic conditions to
generate ethers 9. The nitrile was then oxidized to carboxamide
10 and the methyl sulfide of 10 was converted to the convenient
sulfoxide leaving group, 11, which is well-positioned for the final
6-substitution. Reaction of intermediate 11 with different alcohols
O
N
O
HO
N
NH2
OH
HO
1
2
NH2
O
Z
N
NH2
X: CH, N
Y: O, S, NH
Z: O, S, NH
Hydrophobic
Moiety
Y
X
yielded
the
2,6-disubstituted
aminoalkoxypyrimidine
carboxamides.
Sugar Site
Substituents at the 2-position were extensively investigated to
explore the hydrophobic site for potential to modulate enzyme
inhibitory activity and antibacterial potency as well as selectivity.
As shown in Table 1, compounds substituted with small hydro-
phobic cyclic alkoxy groups, such as cyclopentyl, 13 and cyclo-
hexyl, 14, exhibited sub-micromolar enzyme inhibitory activity.
However, these analogs demonstrated no significant antibacterial
potency. Smaller alkoxy groups such as isopropyl, 16 and fluoro-
ethyl ether, 17 maintain enzyme potency at the same level, and
begin to show some antibacterial potency against a tempera-
ture-sensitive ligase knockdown E. coli strain. Interestingly, 16
and 17 maintain this albeit weak antibacterial potency even when
the temperature sensitive mutant strain possesses a wild-type ef-
flux phenotype, suggesting that these compounds bearing smaller
substituents somehow help evade efflux. As expected, when the
hydrophobic groups were replaced with the hydrophilic carbox-
ylic acid 18, the enzyme affinity dropped more than 100-fold
(IC50 >10 M). In addition to the 2-ether linkage, sulfide, amine
as well as carbon linkers were investigated but none of them
showed promising results. Substitution of the amine and carbox-
amide, and replacement of the carboxamide with sulfonamide or
amidine (compounds 22–27) also caused a significant loss of
activity. An X-ray crystal structure of AAPC, 13 bound to the ac-
tive site of LigA indicated that the aminopyrimidyl carboxamide
occupied identical position as the adenine portion of compound
1 (Fig. 4).
Aminoalkoxy Pyrimidine Carboxamide
(AAPCs)
Figure 2. Aminoalkoxypyrimidine carboxamides (AAPCs).
aminoalkoxypyrimidine carboxamides (AAPCs) emerged as a hy-
brid design based on a central ring-opened pyridochromanone, 2
(minus the ketone), but with a 2-alkoxy aminopyrimidine scaffold
similar to 1 (Fig. 2).
We explored the impact of AAPC substitution on binding to the
hydrophobic and sugar subsites of LigA, respectively. The explora-
tion initially proceeded with mono substitution at the 2-position
(i.e., 2-substituted aminoalkoxypyrimidine carboxamides). As
shown in Schemes 1 and 2-chloro-4-amino-5-cyanopyrimidine
was converted to the corresponding 2-alkoxy ether 4, which was
then further oxidized to the carboxamide 5 by treating the nitrile
with hydroperoxide under basic conditions.17An alternative meth-
od was developed later and involved direct conversion of 4-amino-
2-chloropyrimidine-5-carboxamide, 6 to 5. Other linkages, such as
sulfide 19 and amine 20, were obtained with the same method. In
addition, starting from a simple terminal alkene and reducing it to
its corresponding boron Suzuki precursor followed by standard Su-
zuki coupling, we also synthesized 21 with a simple alkyl linker.
Substitution of the 4-amino and 5-carboxamide groups and the
replacement of carboxamide with different functional groups such
Considering the relatively flat SAR at the mono-C-2 substituted
AAPCs, we explored simultaneous C-6 position substitution with
predominantly cyclic groups of varying hydrophilicity intended

W. Gu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3693–3698
3695
NH2
NH2
NH2
O
NH2
O
CN
CN
R1XH, NaH,
DMF
H2O2, K2CO3
DMSO
R1OH, NaH,
DMF
N
N
N
NH2
N
NH2
R1
R1
Cl
N
X
N
X
N
Cl
N
3
4
5
6
O
HO
X= O; R1=
F
13
14
16
15
18
17
X= S; R1=
X= NH; R1=
19
20
NH2
O
NH2
O
1. 9-BBN, THF
2. Pd(PPh3)4, Na2CO3, DMF
N
NH2
N
NH2
+
X= CH; R1=
N
Cl
N
21
6
Scheme 1. Synthesis of AAPCs.
NH2
N
NH2
N
N
1. NaH, NH2CN,
EtOH
2. HBr
N
H2O2, DMSO
K2CO3, 50 o
R1OH
N
N
C
NaH, THF
R1
S
S
Br
N
S
O
N
S
8
7
9
NH2
O
NH2
O
NH2
O
S
N
N
NH2
m-CPBA/
NH2
N
NH2
R2OH
DIEA, THF
N
R1
Oxone
R1
O
O
R1
O
N
S
O
N
R2
O
10
11
12
O
OH
R1=
R2 =
HO
HO
F
F
32
31
28
29
30
H
N
OMe
HO
33
35
34
36
OMe
R1=
R2 =
R1=
R2
=
HO
HO
37
Scheme 2. Synthesis of 2,6-disubstituted aminoalkoxypyrimidine carboxamides.
to mimic the ribose of NAD+. The introduction of either a 2-
hydroxycyclopentyl (28) or a 2-hydroxycyclohexyl (29) group at
the C-6 position of AAPCs decreases the enzyme inhibitory IC50
approximately 10-fold as compared with the mono 2-substituted
scaffold (compare 13 with 28 or 29, Table 2).
As illustrated in Figure 5, the X-ray crystal structure of 28
bound to Enterococcus faecalis ligase showed that the pyrimidine
core makes favorable hydrophobic
p-stacking interactions with
the phenyl ring of Tyr-227; the cyclopentyl group occupies the
hydrophobic pocket; the ether 2-oxygen atom and the 3-nitrogen

3696
W. Gu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3693–3698
O
NH2
N
O
NH2
O
NH
N
O
NH
N
O
N
NH
N
NH
N
NH2
N
NH2
O
O
O
N
O
O
N
22
23
24
25
O
NH2
O
O
NH2 NH
NH2
S
N
NH2
N
O
N
O
N
26
27
Figure 3. Substitution of amine and carboxamide, replacement of carboxamide.
Table 1
NAD+-dependent DNA ligase activity and MICs of 2-substituted AAPCs
Compound
Enzyme inhibition IC50
(l
M)
MIC (lM)
E. coli NAD+-dependent DNA
ligase
Human
Lig1a
E. coli 1260b (LigAts
D
efflux) E. coli 1201c (LigAts
)
E. coli 1410d (wild type) S. aureuse (wild type)
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
0.61
0.56
0.24
0.41
0.92
>10
6.9
>55
>55
>55
>55
>55
>55
>40
8.1
>55
>55
>55
>55
>55
>55
NDf
ND
ND
ND
ND
ND
ND
ND
ND
9.4
9.8
3.1
15
28
>100
33
>100
>100
>100
>100
>100
>100
>100
100
100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
100
>100
>100
27
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
23
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
a
Human ATP-dependent DNA ligase 1.
b
E. coli strain #1260 (relevant genotype: lig-251 tolC210::Tn10) was derived by P1-mediated transduction of strain #1201 using transducing lysate from CAG12184 (Singer
et al., 1989). lig-251 encodes a temperature sensitive LigA.
c
E. coli strain #1201 (relevant genotype: lig-251 tolC+) is strain GR501 (Dermody et al., 1979).
d
E. coli strain #1410 (relevant genotype: ligA+ tolC+) is strain GR523 (Dermody et al., 1979).
e
S. aureus wild-type strain (ATCC 29213).
ND, not done.
f
ring atom form hydrogen bonds with the side chain amino group of
Lys-291; the carbonyl group of the carboxamide forms a hydrogen
bond with the backbone amide nitrogen of Ile-121 while the nitro-
gen atom of the carboxamide group participates in an extensive
hydrogen bonding network involving the side chain carboxylate
of Glu-175, the backbone carbonyl oxygen of Ile-121, and a water
molecule, which together with another water molecule in the su-
gar site form two hydrogen bonds with the hydroxyl group from
the 6-cyclopentyl ring.
Fluoro substitution on a C-6 cyclohexyl (compound 30) im-
proved the LigA IC50 to 6 nM. Analysis of the crystal structure of
LigA shows a well-defined region for the ribose sugar moiety of
NAD+. It is believed that compounds such as 30 fill this subsite rea-
sonably well. This provided impetus to fill this space even more
effectively. Compound 34 was designed to occupy the entire
sugar-binding region; however, enzyme potency was reduced
ꢀ10-fold relative to 30. A number of other cyclic and acyclic sub-
stituents at the C-6 position were explored but no further potency
enhancement was achieved. When the C-2 cyclopentyloxy group
was replaced with the trans-4-methyl cyclohexyloxy group and
Figure 4. Overlay of AAPC analog, 13 with compound 1. Crystallographic data for
the structures has been deposited with the RCSB Protein Data Bank. Compound 13,
the 6-substituent was 2-hydroxycyclopentyl, compound 37
showed an MIC of 100 M versus a wild type Gram positive
PDB accession code: 4EEQ.

W. Gu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3693–3698
3697
Table 2
NAD+-dependent DNA ligase activity and MICs of 2,6-disubstituted AAPCs
Compound
Enzyme inhibition IC50
(
lM)
MIC (lM)
E. coli NAD+-dependent DNA
ligase
Human
Lig1a
E. coli 1260b (LigAts
D
efflux) E. coli 1201c (LigAts
)
E. coli 1410d (wild type) S. aureuse (wild type)
28
29
30
31
32
33
34
35
36
37
0.12
>55
>55
>55
>55
>55
>55
>55
>55
>55
>55
9.4
1.6
1.2
9.4
1.2
1.2
1.6
4.7
6.3
1.2
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
45
>100
>100
83
0.046
0.019
0.064
0.078
0.014
0.080
0.63
>100
>100
>100
>100
>100
>100
>100
>100
>100
100
>100
>100
>100
100
0.061
0.18
a,b,c,d,e
As in Table 1.
Figure 5. X-ray structure of 28 binding to E. faecalis LigA. Crystallographic data for the structures has been deposited with the RCSB Protein Data Bank. Compound 28, PDB
accession code: 4EFB.
Staphylococcus aureus strain (Table 2), and an MIC of 75
l
M and
of the AAPCs presented here showed promising Gram positive
MICs, thus warranting further investigation for this series.
12.5 M against wild type Haemophilus influenzae (ATCC 51907)
l
and wild type Staphylococcus pneumoniae (ATCC 10015), respec-
tively (data not shown). Tables 1 and 2 also show that none of
the tested AAPC analogs presented in this work possess any
measurable activity versus human ATP-dependent ligase. This is
presumably due to the presence of the 2-alkoxy substituents that
effectively fill the larger hydrophobic sub-pocket present in the
bacterial NAD+ dependent DNA ligases.
Acknowledgements
We would like to thank D. Deiniger, S. Nanthakumar, D. Lowe
for their contributions. We thank E. Nelson and R. Iyer for early
MIC data.
As shown in Tables 1 and 2, the 2,6-disubstituted AAPCs reach
single digit micromolar MICs against the E. coli temperature sensi-
tive ligase knockdown/tolC- double mutant. However, none of
these inhibitors showed MIC <100 M against wild type E. coli. Inter-
estingly, some of these analogs showed promising activity against
wild type Gram positive S. aureus.
In conclusion, the aminoalkoxypyrimidine carboxamides
(AAPC) were evolved into potent NAD+-dependent DNA ligase
inhibitors via structure-guided optimization. Interactions with
two critical regions of the ligase active site were exploited via sub-
stituents at the 2-position that fill the hydrophobic selectivity re-
gion of the active site and substituents at the 6-position that
occupy the ribose-binding region of NAD+. Occupation of both sub-
sites is required to achieve potent enzyme inhibitory activity. Some
References and notes
1. (a) Olivera, B. M.; Lehman, I. R. Proc. Natl Acad. Sci. U.S.A. 1967, 57, 1426;
Lehman, I. R. Science 1974, 186, 790; (b) Gellert, M. Proc. Natl. Acad. Sci. U.S.A.
1967, 57, 148; (c) Weiss, B.; Richardson, C. C. Proc. Natl. Acad. Sci. U.S.A. 1967, 57,
1021; (d) Gefter, M. L.; Becker, A.; Hurwitz, J. Proc. Natl. Acad. Sci. U.S.A. 1967,
58, 240.
2. Ellenberger, T.; Tomkinson, A. E. Annu. Rev. Biochem. 2008, 77, 313.
3. Shuman, S. J. Biol. Chem. 2009, 284, 17365.
4. Gajiwala, K. S.; Pinko, C. Structure 2004, 12, 1449.
5. Han, S.; Chang, J. S.; Griffor, M. Acta Crystallogr. Sect. F Struct. Biol. Cryst.
Commun. 2009, 1078, 65.
6. Nandakumar, J.; Nair, P. A.; Shuman, S. Mol. Cell 2007, 26, 257.
7. Odell, M.; Sriskanda, V.; Shuman, S.; Nikolov, D. Mol. Cell 2000, 6, 1183.
8. BrÖtz-Oesterhelt, H.; Knezevic, I.; Bartel, S.; Lampe, T.; Warnecke-Eberz, U.;
Ziegelbauer, K.; Habich, D.; Labischinski, H. J. Biol. Chem. 2003, 278, 39435.
9. Miesel, L.; Kravec, C.; Xin, A. T.; McMonagle, P.; Ma, S.; Pichardo, J.; Feld, B.;
Barrabee, E.; Palermo, R. Anal. Biochem. 2007, 366, 9.

3698
W. Gu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3693–3698
10. Meier, T. I.; Yan, D.; Peery, R. B.; McAllister, K. A.; Zook, C.; Peng, S. B.; Zhoa, G.
FEBS J. 2008, 275, 5258.
11. Srivastava, S. K.; Dube, D.; Tewari, N.; Dwivedi, N.; Tripathi, R. P.;
Ramachandran, R. Nucleic Acids Res. 2005, 33, 7090.
12. Srivastava, S. K.; Dube, D.; Kukshal, V.; Jha, A. K.; Hajela, K.; Ramachandran, R.
Proteins 2007, 69, 97.
Mills, S. D.; Eakin, A. E.; Buurman, E. T.; Newman, J. V.; Gao, N.; Huyhn, H.;
Johnson, K. D.; Karantzeni, I.; Lahiri, S.; Shapiro, A.; Walkup, G. K.; Yang, W.;
Stokes, S. S. Antimicrob. Agents Chemother. 2011, 55, 1088.
15. 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.
13. Mills, S. D. Antimicrob. Agents Chemother. 2011, 1088, 55.
14. (a) Brotz-Oesterhelt, H.; Knezevic, I.; Bartel, S.; Lampe, T.; Warnecke-Eberz, U.;
Ziegelbauer, K.; Habich, D.; Labischinski, H. J. Biol. Chem. 2003, 278, 39435; (b)
16. Brotz-Oesterhelt, H.; Knezevic, I.; Bartel, S.; Lampe, T.; Warnecke-Eberz, U.;
Ziegelbauer, K.; Habich, D.; Labischinski, H. J. Biol. Chem. 2003, 278, 39435.
17. Katritzky, A. R.; Pilarski, B.; Urogdi, L. Synthesis 1989, 949.