Design of inhibitors of Helicobacter pylori glutamate racemase as selective antibacterial agents: Incorporation of imidazoles onto a core pyrazolopyrimidinedione scaffold to improve bioavailabilty

Article abstract of DOI:10.1016/j.bmcl.2012.07.004

Structure-activity relationships are presented around a series of pyrazolopyrimidinediones that inhibit the growth of Helicobacter pylori by targeting glutamate racemase, an enzyme that provides d-glutamate for the construction of N-acetylglucosamine-N-acetylmuramic acid peptidoglycan subunits assimilated into the bacterial cell wall. Substituents on the inhibitor scaffold were varied to optimize target potency, antibacterial activity and in vivo pharmacokinetic stability. By incorporating an imidazole ring at the 7-position of scaffold, high target potency was achieved due to a hydrogen bonding network that occurs between the 3-position nitrogen atom, a bridging water molecule and the side chains Ser152 and Trp244 of the enzyme. The lipophilicity of the scaffold series proved important for expression of antibacterial activity. Clearances in vitro and in vivo were monitored to identify compounds with improved plasma stability. The basicity of the imidazole may contribute to increased aqueous solubility at lower pH allowing for improved oral bioavailability.

Full text of DOI:10.1016/j.bmcl.2012.07.004

Bioorganic & Medicinal Chemistry Letters 22 (2012) 5600–5607
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design of inhibitors of Helicobacter pylori glutamate racemase
as selective antibacterial agents: Incorporation of imidazoles onto a core
pyrazolopyrimidinedione scaffold to improve bioavailabilty
Gregory S. Basarab a, , Pamela Hill a, Charles J. Eyermann a, Madhu Gowravaram a, Helena Käck b,
⇑
Ekundayo Osimoni a
a Infection Innovative Medicines, AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
b AstraZeneca Discovery Sciences, Pepparedsleden 1, 431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Structure–activity relationships are presented around a series of pyrazolopyrimidinediones that inhibit
Received 3 May 2012
Revised 28 June 2012
Accepted 2 July 2012
Available online 20 July 2012
the growth of Helicobacter pylori by targeting glutamate racemase, an enzyme that provides D-glutamate
for the construction of N-acetylglucosamine-N-acetylmuramic acid peptidoglycan subunits assimilated
into the bacterial cell wall. Substituents on the inhibitor scaffold were varied to optimize target potency,
antibacterial activity and in vivo pharmacokinetic stability. By incorporating an imidazole ring at the 7-
position of the scaffold, high target potency was achieved due to a hydrogen bonding network that occurs
between the 3-position nitrogen atom, a bridging water molecule and the side chains Ser152 and Trp244
of the enzyme. The lipophilicity of the scaffold series proved important for expression of antibacterial
activity. Clearances in vitro and in vivo were monitored to identify compounds with improved plasma
stability. The basicity of the imidazole may contribute to increased aqueous solubility at lower pH allow-
ing for improved oral bioavailability.
Keywords:
MurI
Glutamate racemase
Helicobacter pylori
Antibacterial
Pyrazolopyrimidinedione
Ó 2012 Elsevier Ltd. All rights reserved.
It is estimated that upwards of 50% of the world’s population is
infected with the Gram-negative bacterium Helicobacter pylori, a
pathogen that colonizes the mucosal lining of the stomach.1
Though many H. pylori infections are asymptomatic, they often
lead to gastric disturbances such as ulcers and esophageal efflux
and perhaps even to gastric cancer.2 Currently a cocktail of three
drugs including two broad-spectrum antibacterials is most often
used over a two week period for the treatment of H. pylori infec-
tions, a treatment plagued by poor patient compliance due to diar-
rhea and other side effects resulting from the suppression of
commensal bacteria. Furthermore, resistance of H. pylori to current
therapies prompts the need for an alternative therapy with a new
mode-of-action.1,3 For more severe and life-threatening infections,
empiric therapy with broad-spectrum antibacterials constitutes
the initial course of action due to the limited capabilities of deter-
mining the culpable pathogen in a timely fashion to tailor the
choice of drug. However for the eradication of chronic, slow grow-
ing pathogens such as H. pylori, precise timing is less important and
selectivity is desirable to eliminate the suppression of bacterial
flora by broad-spectrum agents. The need for an improved therapy
against H. pylori prompted our efforts to identify an orally active
drug that operates by inhibition of glutamate racemase encoded
by the MurI gene essential for the biosynthesis of bacterial cell wall
peptidoglycan. MurI is one of a series of genes responsible for
building N-acetylglucosamine-N-acetylmuramic acid peptidogly-
can subunits that are incorporated into the bacterial cell wall.4
D-glutamate (along with D-alanine and meso-diaminopimelic acid)
is a necessary peptide component of bacterial peptidoglycan that
confers a measure of proteolytic stability to the cell wall due to
its atypical configuration. Deletion of MurI prevents peptidoglycan
construction and bacterial viability by disrupting the supply of
D
-glutamate and thereby represents a worthwhile target for the
design of antibacterial drugs.5 Comparative genomics showed a
divergence for the MurI gene for H. pylori from other bacteria,6
both Gram-negative and Gram-positive, offering the expectation
that a selective anti-H. pylori agent could be identified.
Figure 1 shows the structure for pyrazolopyrimidinedione
inhibitor 1 of the H. pylori MurI isozyme that was discovered by
high throughput screening against H. pylori MurI and of analogs
2 and 3 that were designed in subsequent analog optimization pro-
grams.6b,7 Compounds 1, 2 and 3 show exquisite selectivity for the
H. pylori MurI isozyme (IC50s of 1400, 26 and 34 nM, respectively)
relative to three other MurI isozymes (derived from Staphylococcus
aureus, Enterococcus faecalis and Escherichia coli) where IC50s
exceeded the assay limit of 400
translated into the suppression of H. pylori growth in vitro as the
three compounds demonstrated MICs ranging from 0.5–8 g/ml;
lM. The enzyme inhibitory potency
⇑
Corresponding author. Tel.: +1 781 839 4689.
E-mail address: greg.basarab@astrazeneca.com (G.S. Basarab).
l
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.07.004

G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
5601
O
O
N
N
S
N
O
N
O
O
N
3
4
Cl
Cl
N 5
6
N
N
N
N
N
7
2
N
O
N
N
N
1
O
N
O
N
N
1
2
3
MurI IC50 = 1400 nM
MurI IC50 = 26 nM
MurI IC50 = 34 nM
MIC H. pylori (wild-type) = 8 µg/ml
MIC H. pylori (wild-type) = 0.5 µg/ml
in vivo Cl (mouse) = 18 ml/min/kg
bioavailabilty (mouse) < 10%
MIC H. pylori (wild-type) = 2.5 µg/ml
in vivo Cl (mouse) = 15 ml/min/kg
bioavailabilty (mouse) = 13%
Figure 1. Biological properties of pyrazolopyrimidinedione inhibitors of glutamate racemase (MurI).
no MICs below the assay limit of 64
l
g/ml were seen against a
hydrazones 5 are condensed with a second aldehyde catalyzed by
piperidine and heated to effect cyclization and rearrangement
to compound 6 via a sequence first described by Yamada and co-
workers and exemplified in the patent literature.10
broad-spectrum of Gram-negative and Gram-positive pathogens.
As a practical matter, an anti-H. pylori therapy would be admin-
istered orally over a likely time course of treatment for 1–2 weeks.8
Therefore, the drug needs to achieve high blood levels by oral
administration with sufficient exposure (after accounting for bind-
ing to endogenous protein) to exceed the MIC several fold over a
period of time depending on the pharmacodynamic driver and
magnitude. The free levels of 2 and 3 on oral dosing at 40 mg/kg
in mouse models fell well below the MICs due, in large part, to
low bioavailabilities. An exposure of 3 above its MIC = 2 could be
maintained for 4 h by oral dosing 40 mg/kg of the pro-drug sulfox-
ide analog of 3. However if the exposure is adjusted for the plasma
protein binding (ppb of 3 is 97% in the mouse), the expectation is
that the drug would not suppress H. pylori proliferation in vivo.7
We therefore set out to improve upon 2 and 3 by designing ana-
logues with improved systemic exposure on oral dosing to the
point that sustained free drugs levels well above MICs could be
attained.
The various hydrazides 4 are synthesized by methods set out in
the literature wherein a monosubstituted urea is treated with
diethylmalonate and base to form an intermediate pyrimdinetri-
one that is regioselectively converted to a pyrimidinedione chlo-
ride 7a with POCl3 (Scheme 2).6b,10,11 The NH of 7a is alkyated on
nitrogen leading to 8. Alternatively, the 7-position substituent
can be introduced first by alkylation of 6-chloro-1H-pyrimidine-
2,4-dione to form 7b; subsequent alkylation affixes the 5-position
substituent. The chloride is then displaced with hydrazine to afford
4.11 Analogously, the compounds with the isobutyl substituent at
the 7-position were made with isobutyl bromide in place of the
cyclopropylmethyl bromide in the reaction sequences. Here we
investigate the influence of 3- and 5-positions as opportunities to
incorporate more polar functionalities deemed helpful for increas-
ing solubility and improving physical properties that were, in turn,
deemed important for bioavailability, clearance and plasma pro-
tein binding (ppb).
The x-ray derived crystal structure of 1 with the MurI enzyme6b
helped guide much of the inhibitor design strategies. Introducing
polarity on the 2- and 7-position substituents (see Fig. 1 for num-
bering) is detrimental to inhibitory potency as both positions
reside in hydrophobic environments of the enzyme binding pocket.
By contrast, the 5-position extends along a hydrophobic surface
that is exposed otherwise to solvent. The 3-position lies in a par-
tially hydrophobic environment and a partially hydrophilic envi-
ronment that extends out to solvent. The pyridine nitrogen of 1,
the nitrile of 2 and the sulfone of 3 lie within the hydrophilic re-
gion. Indeed, even more polar functionality such as sulfonamide
and sulfoxide in place of the nitrile of 2 or the sulfone of 3 are
well-tolerated for MurI inhibitory potency. However, polar substit-
uents diminish bacterial membrane permeability resulting in high-
er MICs versus H. pylori.7 Indeed, a problem confounding design
work is that hydrophobicity is best suited for membrane perme-
ability and lower MICs while hydrophilicity oftentimes correlates
with lower ppb and higher solubility that will translate into higher
bioavailability properties that make for better drugs.
Many of the R3 aldehydes of Scheme 1 are commercially avail-
able or known in the literature while others required de novo
synthesis.12 The cyanoimidazole aldehyde 12 for incorporation into
compounds 24 and 58–62 was prepared via sequence in Scheme 3.
Initial protection of the starting imidazole carboxaldehyde was fol-
lowed by dibromination to afford 10. The bromine atom adjacent
to the N-methylimidazole nitrogen was selectively removed by
metal halogen exchange and quench with acid. Palladium cata-
lyzed Zn(CN)2 cross-coupling and hydrolysis affords 12.
The cyanofuran aldehyde incorporated into 28 was prepared in
5 steps by the sequence outlined in Scheme 4. Reduction of ester
13 and protection of the alcohol to form 14 allows electrophilic
cyanation with chlorosulfonyl isocyante13 affording 15. Deacyla-
tion and oxidation to the aldehyde finishes the sequence.
The isomeric cyanofuran aldehyde 20 for incorporation into 25
was prepared by an analgous sequence of steps (Scheme 5).
Results and discussion
Chemistry
The cyclopropylmethyl substituent is optimal at the 7-position
of the pyrazolopyrimidine scaffold as even a slight perturbation,
the replacement of the cyclopropylmethyl group with an isobutyl
group lowers potency 3-fold and leads to metabolically less stable
compounds as determined by microsomal stability determinations.
Compare compounds 2 and 23 with 26 and 27, respectively
(Table 1). Hence, the majority of the compounds described here
were made with the 7-position cyclopropylmethyl substituent.
The 2-, 3-, 5- and 7-positions of the pyrazolopyrimidinedione
scaffold can be systematically varied (see 1 for numbering) for
developing structure–activity relationships via a single pot sequence
that invokes two condensation reactions onto a pyrimidinedione
hydrazide followed by a cyclization-rearrangement reaction as out-
lined in Scheme 1. Hydrazides 4 are condensed with 6-chloroquin-
oline-4-carboxaldehyde9 to form hydrazones 5. Without isolation,

5602
G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
CHO
N
Cl
O
O
N
O
N
R3
N
Cl
R3
R5
O
R5
R5
O
Cl
O
N
N
N
H
N
NH2
N
N
N
H
O
N
N
H
N
N
H
4
5
6
Scheme 1. General synthesis of the pyrazolopyrimidine scaffold.
O
O
R5
O
1) NaOH,
1) NaOH
2) POCl3
Br
N
R5
EtO
+
DMF, 80 °C, 16 h
N
H
NH2
O
N
H
Cl
EtO
O
O
O
N
R5
7a
R5
O
NH2NH2, MeOH
RT, 16 hr
N
N
NH
O
N
N
H
Cl
O
R5
DMF, 80 °C, 16 h
, K CO
3
X
O
2
HN
Br
60-70% - 2 steps
HN
4
8
O
N
Cl
K2CO3, NaI,DMF
80 °C, 16 hr
O
N
H
Cl
43%
7b
Scheme 2. General syntheses of key hydrazide intermediates 4 enabling variation of the R5 substituent.
Br
Br
Br
CN
1) Zn(CN)2
Pd2(dba)3, DPPF
1) HC(OEt)3
TsOH
N
n-BuLi, H+
70%
N
N
N
N
N
N
N
2) H3O+
36%
2) NBS
83%
O
O
O
O
O
H
O
H
Et
Et
Et
Et
10
9
11
12
Scheme 3. Synthesis of cyanoimidazole 12 for construction of compounds 24 and 58–62.
NC
NC
O
O
ClSO2NCO, DMF
CH3CN, -40 °C
1) NH3, MeOH
1) LAH, THF, -70 °C
2) AcCl, TEA, THF
O
O
O
O
O
O
2) PCC, CH2Cl2
27%
Ac
Ac
O
13
14
15
16
67%
54%
Scheme 4. Synthesis of cyanofuran 16 for construction of compound 28.
NC
NC
O
ClSO2NCO, DMF
1) NH3, MeOH
2) PCC, CH2Cl2
O
1) LAH, THF, -70 °C
O
O
O
O
2) AcCl, TEA, THF
CH3CN, -40 °C
68%
O
O
Ac
Ac
O
17
18
19
20
72%
18%
Scheme 5. Synthesis of cyanofuran 20 for construction of compound 25.
Additionally, the chloroquinolinemethyl substituent at the 2-
position is optimal6b and was kept constant through the analogs
herein. Enzyme bound compounds are likely pre-organized into
the binding conformation due to p–p stacking of the chloroquino-
line with 3-position aromatic groups6b leading to a hydrophobic
collapse in aqueous solution.

G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
5603
Table 1
Influence of 3-position substituents on IC50, MIC and Clint
O
R3
Cl
N
N
N
O
N
R7
N
20
c
Compd
R7
R3
IC50 (nM)
MIC (
Wild typea
lg/ml)
Clint
Human
(l
l/min/mg)
hefCÀb
Mouse
N
N
2
25
0.5
0.8
14
50
O
22
23
24
39
41
87
3
2
4
—
19
<17
<14
53
21
87
N
N
N
0.25
2
N
N
N
N
25
26
60
86
1
4
0.25
—
59
94
>100
>100
O
N
N
N
N
27
28
170
220
4
2
—
2
17
38
>100
63
N
O
NH
29
30
220
230
8
8
—
—
20
>100
64
N
N
N
N
<14
N
31
32
33
260
280
740
18
—
—
—
>100
34
>100
>100
61
S
N
NH2
8
N
S
NH2
>32
<14
N
N
34
35
810
900
8
4
32
92
<14
N
N
32
—
>100
N
H
a
b
c
H. pylori strain 055.
Ref. 14.
Liver microsome clearance.
Placing a 5-atom versus 6-atom aromatic heterocycle at the
Ser152, the indole NH of Trp244 and other hydrogen bonding res-
idues that lie in the surrounding region of the enzyme in a channel
that extends to solvent. However, the more polar functionality de-
creases antibacterial activity presumably due to lower membrane
permeability. Previous work had also shown that a 1,2-orientation
between the scaffold attachment and a small lipophilic group
(methyl group or chlorine atom) improves inhibitory potency.7
3-position of the pyrazolopyrimidinedione scaffold affords the
optimal inhibitory potency versus H. pylori MurI. Previous work
had shown that potency is enhanced if there is a 1,3-orientation
between the scaffold attachment and a polar group (e.g., sulfon-
amide) on the heterocycle.7 The polar group interacts directly or
indirectly (through bridging water molecules) with hydroxyl of

5604
G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
Table 2
nitrogen atom (the latter exemplified by 23 and 27, Table 1 and
the compounds of Table 2). A model derived by x-ray diffraction
of crystals of a complex of 21 with the enzyme (Fig. 2) shows the
imidazole lone pair of electrons associated with a water molecule
that is positioned to bridge Ser152 and Trp244 in a hydrogen bond-
ing array.15 Compound 21 was one of the first compounds made
in this series and has the non-optimized isobutyl group at the 7-
position and a non-optimized nathphylenemethyl substituent at
the 2-position. Positioning a carbon atom with a hydrogen substi-
tuent instead of the imidazole lone pair of electrons towards this
region as in pyrrole 31, triazole 34, and imidazole 35 confers con-
siderably lower potency. The small amino substituent of 32 and 33
is also detrimental to activity. Slightly larger substituents such as
the nitrile, by contrast, dislodge the bringing water molecule and
can form direct interactions with the enzyme, hence accounting
for the better potencies that must be attributed in part to the over-
all entropy gain. Examples of a nitrile displacing a crystallographic
water molecule to increase inhibitory potency have been docu-
mented.16 Decreased inhibitor potency was seen if a heteroatom
is placed in any of the other positions of the heterocycle as exem-
plified by comparing 23–30 where the transformation from imid-
R3 Methylimidazoles: influence of 5-position substituents on IC50’s, MICs and Clint
N
N
O
R5
O
Cl
N
N
N
N
N
Compd R5
IC50
MIC (
l
g/ml)
Clint (ll/min/kg)
c
(nM)20
Wild
typea
hefCÀb Human Mouse
36
37
38
HC„CCH2–
HOCH2C„CCH2–
CH2@CHCH2–
24
54
56
2
64
1
0.5
16
0.8
<14
>100
25
18
>100
49
39
63
8
2
47
44
40
41
69
72
2
4
0.9
1
<14
<14
44
15
CH3CH2–
azole to triazole decreases inhibitory potency
7 to 8-fold.
42
79
8
2
<14
6
Cyanoimidazole 24 shows decreased potency (2 to 3-fold) relative
to cyanopyrrole 2 as do cyanofurans 25 and 28 (2- and 8-fold,
respectively). However, due the lipophilic nature of the cyanohet-
erocycles, membrane permeability is increased and antimicrobial
activity does not deteriorate correspondingly.
O
N
43
170
64
16
>100
>100
44
45
(CH3)2CH–
HC„C(CH3)CH–
220
460
16
8
16
8
53
21
51
65
Maintaining the better heterocycles at the 3-position (imidaz-
ole, cyanopyrrole and cyanoimidazole, Tables 2 and 3) allows
elaboration of the scaffold 5-position to further explore struc-
ture–activity insights. As mentioned, the 5-position abuts a hydro-
phobic patch and otherwise opens to bulk solvent (Fig. 2). Hence,
both lipophilic and polar groups can be tolerated in this position,
though in line with previous observations, higher polarity compro-
mises MICs even when potencies are only reduced a few fold (com-
pare 37 and 43 with 23, for example). Compound 46 registers one
of the higher inhibitory potencies against the MurI enzyme, yet
MICs are higher due to the polar triazole group. Similarly, the MICs
of 48–50 trend upward due to the polar substituents. Oftentimes,
much better MICs are recorded for the more polar compounds
against the pump debilitated hefCÀ strain of H. pylori. This is partic-
ularly evident from previous work in which particularly polar sul-
fonamide and sulfoxide substituents afforded efflux ratios (the
ratio between the MIC of wild-type H. pylori and the hefCÀ strain)
of 32–64 or more and resulted with exceedingly high MICs versus
the wild-type strain.7 Consequently, much of the work herein fo-
cussed on less polar substituents in compound design. A working
hypothesis that drives the design is that the efflux pump recog-
nizes the various more polar and less polar compounds somewhat
equally, and therefore, elimination of the pump would lead to a
more pronounced lowering of MICs for more polar compounds
with the reduced capability to cross bacterial membranes. More
a
b
c
H. pylori strain 055.
Ref. 14.
Liver microsome clearance.
The lipophilic group lies correspondingly in a complementary lipo-
philic pocket of the enzyme and enhances the staggered rotational
conformation of the heterocycle relative to the scaffold. As might
be expected, the increased potency along with increased lipophil-
icity enhancing membrane permeability improves antibacterial
activity. In addition to MICs against a wild-type H. pylori strain,
MICs are shown versus an H. pylori construct (designated hefCÀ)
that is compromised by knock-out of the AcrB component of the
AcrAB-TolC efflux pump machinery.14 This allows for interpreta-
tion of the contribution of permeation though the outer and inner
membranes of the pathogen in limiting the expression of antimi-
crobial activities of the inhibitors.
So herein, we have incorporated a series of 5-membered ring
heterocycles with a 1,3-orientation relative to the scaffold attach-
ment for displaying a substituent towards the Ser152/Trp244
channel. A methyl group was additionally incorporated in a 1,2-
orientation to the scaffold for an overall 1,2,4-orientation of methyl
group, scaffold attachment, and substituent, respectively. The
substituents of choice include those of intermediate polarity such
as ketones and nitriles as well as the lone pair of a heterocyclic
Figure 2. Surface surrounding 5-position substituent (left) and H-bonding array with a crystallographic water molecule, Ser152 and Trp244 (right).

G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
5605
Table 3
R3 Cyanopyrroles and cyanoimidazoles: influence of 5-position substitution
N
N
Y
O
N
R5
N
Cl
N
N
O
N
20
c
Compd
R5
Y
IC50 (nM)
MIC (
l
g/ml)
Clint
Human
(l
l/min/mg)
Wild Typea
hefCÀb
Mouse
N
N
N
46
47
CH
CH
13
28
8
16
>100
<14
>100
16
HC„CCH2–
0.75
0.3
O
48
CH
51
16
4
>100
>100
N
49
50
NCCH2–
HOCH2C„CCH2–
CH
CH
55
57
8
32
1
16
<14
>100
96
>100
51
CH
62
2
2
23
<14
52
53
54
CH3CH2–
CH2@CHCH2–
CH3C„CCH2–
CH
CH
CH
70
70
74
0.75
1.5
2
0.6
1
2
25
33
99
17
76
>100
55
56
CH
CH
80
88
2
0.13
16
30
33
HOCH2CH2
64
>100
>100
N
N
57
CH
120
8
4
>100
>100
58
59
HC„C(CH3)CH–
(CH3)2CH–
CH
CH
160
240
2
4
8
8
91
26
>100
>100
N
60
CH
260
64
4
41
>100
N
N
N
61
62
CH
CH
470
670
32
16
32
16
>100
>100
>100
>100
N
N
N
63
64
HC„CCH2
N
N
89
1
2
2
2
<14
ND
28
150
ND
65
66
CH3CH2
CH2@CHCH2
N
N
180
250
2
2
2
4
51
>100
<14
>100
67
N
250
8
4
<14
<14
a
b
c
H. pylori strain 055.
Ref. 14.
Liver microsome clearance.21
lipophilic methyl, ethyl, cyclopropyl, cyclopropylmethyl, allyl and
propargyl R3 substituents are favored for wild-type MICs. Indeed,
the efflux ratios of compounds are about two or less (compounds
2, 28, 38, 40, 44–45, 47, 51–54, 58–59, and 63–66) with more lipo-
philic functionality. Recognizing that the individual MICs can have
a 2-fold variations, the trend holds that efflux ratios increase with
greater polarity and decrease with greater lipopholicity, though
subtle exceptions to the observations may be seen. Branching at
the attachment point of the R5 substituent reduces inhibitory po-
tency as seen with the incorporation of isopropyl groups of 44
and 59 and a-methyl propargyl groups of 45 and 58. The reduction
in binding potency is less pronounced (about 3-fold) for the cyclo-
propyl substituent seen in 37, 51 and 64 relative to the methyl sub-
stituent of 23, 2 and 24, respectively.
Mouse and human microsomal clearances were determined in
parallel with MICs; compounds with lower MICs and Clint values

5606
G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
Table 4
incorporating an imidazole ring at the 7-position of scaffold. The
PK parameters in the mouse22
imidazole was shown to impart higher target potency through a
hydrogen bonding network with a bridging water molecule to
the protein. The basicity of the imidazole may also contribute to in-
creased aqueous solubility that allows for improved adsorption.
Adjustment of substituents otherwise help to optimize target po-
tency, antibacterial activity and in vivo stability. However, more
work needs to be done to understand the limiting factors for
translation to good efficacy in an animal model and for utility as
a therapy for treatment of H. pylori infections in humans.
Compound 23
Compound 36
Cl (ml/min/kg)
t1/2 (h)
14
20
2.1
3.6
100
4
0.7
1.2
76
Vss (L/kg)
F (%)
ppb23
<3.1
less than 50 ml/min/mg were evaluated in vivo for PK characteris-
tics in the mouse. The highest drug exposure (AUC on oral dosing
of 40 mg/kg of drug) was recorded by the compounds where
R7 = imidazole; in particular, 23 and 36 (Table 4) exhibit lower
clearances and higher bioavailabilities combined to give sustained
blood levels that rise well above the MICs as exemplified with the
time course plot for 36 in Chart 1. The improvement in bioavail-
ability with the imidazole is believed to relate to its weak basicity
enhancing solubilization in the acidic conditions of the stomach
before passing to the intestine for absorption. The pKa of the 36
was measured to be 7.2 in line with the weakly basic imidazole
substituent.
Acknowledgments
Credit should be given to Tomas Lundqvist of AstraZeneca Dis-
covery Sciences for assistance in generating the crystal structure of
the MurI enzyme–inhibitor complex. The AstraZeneca Infection
Innovative Medicines Biosciences Department determined the
MICs and carried out the in vivo efficacy studies. The AstraZeneca
Infection Innovative Medicines DMPK group performed the in vitro
and in vivo PK assessments.
Compound 36 was tested for efficacy in a murine H. pylori infec-
tion model dosing for 2 days post inoculation at 40 mg/kg qid, in
part to determine the pharmacodynamic (PD) driver. A modest,
statistically insignificant 1 log reduction in the CFU count versus
the untreated control was seen despite achieving blood levels 4X
the MIC. As a positive control, Y-34867, a fluoroquinolone with
anti-H. pylori activity,17 was dosed for 2 days at 10 mg/kg bid
showing complete eradication of the pathogen (nearly 5log reduc-
tion in CFU counts). The reasons for the compound failing in the
efficacy experiment are many-fold. A more potent drug in this class
with a higher free fraction would be better suited for showing
eradication of H. pylori in the model as a test in principle whether
MurI inhibition would be a viable mode-of-action in vivo. This
References and notes
1. Jones, K. R.; Cha, J.-H.; Merrell, D. S. Curr. Drug Ther. 2008, 3, 190.
2. (a) Malfertheiner, P.; Bornschein, J.; Selgrad, M. J. Digestive Dis. 2010, 11, 2; (b)
Kim, S. S.; Ruiz, V. E.; Carroll, J. D.; Moss, S. F. Cancer Lett. 2011, 305, 228.
3. (a) O’Connor, A.; Gisbert, J. P.; McNamara, D.; O’Morain, C. Helicobacter 2010,
15, 46; (b) Graham, D. Y.; Fischbach, L. Gut 2010, 59, 1143; (c) Suzuki, H.;
Nishizawa, T.; Hibi, T. Future Microb. 2010, 5, 639.
4. (a) Van Heijenoort, J. J. Nat. Prod. Rep. 2001, 18, 503; (b) Doublet, P.; van
Heijenoort, J.; Mengin-Lecreulx, D. J. Bacteriol. 1992, 174, 5772; (c) Glavas, S.;
Tanner, M. E. Biochemistry 2001, 40, 6199; (d) Doublet, P.; van Heijenoort, J.;
Mengin-Lecreulx, D. Biochemistry 1994, 33, 5285.
5. Fisher, S. L. Microbiol. Biotechnol. 2008, 1, 345.
6. (a) Supuran, C. T. Future Microbiol. 2007, 2, 111; (b) Lundqvist, T.; Fisher, S. L.;
Kern, G.; Folmer, R. H. A.; Xue, Y.; Newton, D. T.; Keating, T. A.; Alm, R. A.; de
Jonge, B. L. M. Nature 2007, 447, 817; (c) Doig, P.; de Jonge, B. L.; Alm, R. A.;
Brown, E. D.; Uria-Nickelsen, M.; Noonan, B.; Mills, S. D.; Tummino, P.; Carmel,
G.; Guild, B. C.; Moir, D. T.; Vovis, G. F.; Trust, T. J. Microbiol. Mol. Biol. Rev. 1999,
63, 675.
7. Basarab, G. S.; Hill, P.; Rastagar, A.; Webborn, P. H. J. Bioorg. Med. Chem. Lett.
2008, 18, 4716.
8. Graham, D. Y.; Lu, H.; Yamaoka, Y. Curr. Opin. Drugs 2008, 68, 725.
9. Campbell, K. N.; Sommers, A. H.; Kerwin, J. F.; Campbell, B. K. J. Am. Chem. Soc.
1946, 68, 1851.
10. (a) Yamada, F. J. Am. Chem. Soc. 1974, 96, 5607; (b) Basarab, G.; Eyermann, J.;
Gowravaram, M.; Green, O.; MacPherson, L.; Morningstar, M; Nguyen, T. U.S.
Patent 7285558, 2004; Chem. Abstr. 2004, 139, 89821; (c) Cheshire, D.; Cooper,
M.; Donald, D.; Thorne, P. U.S. Patent, 2000, 6028074; Chem. Abstr. 2000, 129,
302556.
would hold in particular if the PD driver were AUC/MIC or Cmax
/
MIC versus Time/MIC, the latter being less sensitive to protein
binding effects. Unfortunately, the efficacy experiment with 36
falls short of being able to delineate the PD driver that would aide
future compound design. Additionally, it would be of interest to
determine whether the blood levels of drug correlate with levels
that might meet the stomach region of colonization. H. pylori is a
relatively slow growing pathogen that evades the body’s immune
system colonizing the stomach mucosa for years or decades. A
compound with cidal activity is best suited in such a situation,18
and the class of compounds described herein is static.19 The
question arises as to whether the target site is limiting towards
achieving cidality or the whether the pyrazolopyrimidine class of
inhibitors is not cidal in itself; additional work is needed to
improve the understanding.
11. (a) Naka, T.; Furukawa, Y. Chem. Pharm. Bull. 1979, 6, 1328; (b) Isobe, Y.; Hirota,
K. Chem. Pharm. Bull. 2003, 12, 1451.
12. The corresponding aldehydes for formation of the 3-position substituents in
compounds 17, 18, 23 and 26–40 are commercially available. The synthesis of
the aldehyde for the formation of the 3-position substituents of compounds 2,
21, 41–57 is described in: Loader, C. E.; Anderson, H. J. Can. J. Chem. 1981, 59,
2673 and Ref. 10b. The synthesis of the corresponding aldehyde for the
formation of the 3-position substituent of compounds 25 is described in:
Moderhack, D.; Hoppe-Tichy, T. J. Prakt. Chem., Chem. Zeit. 1996, 338, 169 and
Ref. 10b.
In summary, progress was made towards achieving good PK
properties (clearance and bioavailability) for the pyrazolopyrimi-
dine class of compounds that inhibit the growth of H. pylori by
13. Floyd, A. J.; Kinsman, R. G.; Roshan-Ali, Y.; Brown, D. W. Tetrahedron 1983, 39,
3881.
14. (a) de Jonge, B. L. M.; Kutschke, A.; Uria-Nickelsen, M.; Kamp, H. D.; Mills, S. D.
Antimicrob. Agents Chemother. 2009, 53, 3331; (b) Kutschke, A.; de Jonge, L. M.
B. Antimicrob. Agents Chemother. 2005, 7, 3009.
15. Crystallization of compound 21 with MurI was carried out as described in Ref.
6b. The data and coordinates are deposited in the Protein Data Bank; the
accession code is 4b1f.
16. (a) Chen, J. M.; Xu, S. L.; Wawrzak, Z.; Basarab, G. S.; Jordan, D. B. Biochemistry
1998, 37, 17735; (b) Liu, C.; Wrobleski, S. T.; Lin, J.; Ahmed, G.; Metzger, A.;
Wityak, J.; Gillooly, K. M.; Shuster, D. J.; McIntyre, K. W.; Pitt, S.; Shen, D. R.;
Zhang, R. F.; Zhang, H.; Doweyko, A. M.; Diller, D.; Henderson, I.; Barrish, J. C.;
Dodd, J. H.; Schieven, G. L.; Leftheris, K. J. Med. Chem. 2005, 48, 6261.
17. Sakurai, N.; Sano, M.; Hirayama, F.; Kuroda, T.; Uemori, S.; Moriguchi, A.;
Yamamoto, K.; Ikeda, Y.; Kawakita, T. Bioorg. Med. Chem. Lett. 1998, 8, 2185.
18. (a) Suzuki, H.; Nishizawa, T.; Hibi, T. Future Microbiol. 2010, 5, 639; (b) Gisbert,
J. P.; Calvet, X.; O’Connor, J. P. A.; Megraud, F.; O’Morain, C. A. Exp. Opin.
Pharmacother. 2010, 11, 905.
Chart 1. Mouse PK of compound 36.

G. S. Basarab et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5600–5607
5607
19. Unpublised work, B.M.L. de Jonge, AstraZeneca Infection Innovative Medicines.
20. IC50 is the mean of duplicate determinations by the method in Ref. 6b. MurI
protein was expressed from murI gene of chromosomal DNA of wild-type
22. Compounds for oral and iv dosing in mice were formulated in 25% PEG/water
and 4:4:2 dimethylacetamide-PEG400–water, respectively. Compound
concentrations in plasma at different time points were quantified by LC/MS.
Pharmacokinetic (PK) parameters were calculated by using WinNonlin. Mean
results were determined from experiments run in triplicate.
Helicobacter pylori strain
J 99 (human isolate, Tenn. Genome Therapeutic
Corporation, now Oscient Pharmaceuticals, Inc.)
21. Metabolic stability was determined by incubating a 2
lM compound solution
23. Plasma protein binding (ppb) was determined from a 10 lM compound
in 0.5 mg/ml human or mouse microsomes and adding 1 mM NADPH.
Compound concentrations at different time points were quantified by LC/MS.
Intrinsic clearance (Clint) was calculated from disappearance of the compound.
solution in a Dianorm plasma well incubating at 37 °C for 16 h. Free fractions
were calculated from ratio between drug concentration in buffer and drug
concentration in buffer plus plasma wells as determined by LC/MS.