Discovery and characterization of a novel class of pyrazolopyrimidinedione tRNA synthesis inhibitors

Article abstract of DOI:10.1038/ja.2014.163

A high-throughput phenotypic screen for novel antibacterial agents led to the discovery of a novel pyrazolopyrimidinedione, PPD-1, with preferential activity against methicillin-resistant Staphylococcus aureus (MRSA). Resistance mapping revealed the likely target of inhibition to be lysyl tRNA synthetase (LysRS). Preliminary structure-activity relationship (SAR) studies led to an analog, PPD-2, which gained Gram-negative antibacterial activity at the expense of MRSA activity and resistance to this compound mapped to prolyl tRNA synthetase (ProRS). These targets of inhibition were confirmed in vitro, with PPD-1 showing IC 50 s of 21.7 and 35 μM in purified LysRS and ProRS enzyme assays, and PPD-2, 151 and 0.04 μM, respectively. The highly attractive chemical properties of these compounds combined with intriguing preliminary SAR suggest that further exploration of this compelling novel series is warranted.

Full text of DOI:10.1038/ja.2014.163

The Journal of Antibiotics (2014), 1–7
2014 Japan Antibiotics Research Association All rights reserved 0021-8820/14
www.nature.com/ja
&
ORIGINAL ARTICLE
Discovery and characterization of a novel class of
pyrazolopyrimidinedione tRNA synthesis inhibitors
Justin I Montgomery1, James F Smith2, Andrew P Tomaras2,3, Richard Zaniewski2, Craig J McPherson2,
Laura A McAllister1, Sandra Hartman-Neumann2, Joel T Arcari1, Marykay Lescoe2, Jemy Gutierrez2,3,
Ying Yuan2,4, Chris Limberakis1 and Alita A Miller25
A high-throughput phenotypic screen for novel antibacterial agents led to the discovery of a novel pyrazolopyrimidinedione,
PPD-1, with preferential activity against methicillin-resistant Staphylococcus aureus (MRSA). Resistance mapping revealed the
likely target of inhibition to be lysyl tRNA synthetase (LysRS). Preliminary structure–activity relationship (SAR) studies led to an
analog, PPD-2, which gained Gram-negative antibacterial activity at the expense of MRSA activity and resistance to this
compound mapped to prolyl tRNA synthetase (ProRS). These targets of inhibition were confirmed in vitro, with PPD-1 showing
IC50s of 21.7 and 35 μM in purified LysRS and ProRS enzyme assays, and PPD-2, 151 and 0.04 μM, respectively. The highly
attractive chemical properties of these compounds combined with intriguing preliminary SAR suggest that further exploration of
this compelling novel series is warranted.
The Journal of Antibiotics advance online publication, 3 December 2014; doi:10.1038/ja.2014.163
INTRODUCTION
anti-methicillin-resistant Staphylococcus aureus activity. PPD-1 was
found to inhibit bacterial translation at the tRNA synthesis step.
Preliminary structure–activity relationship studies led to an analog,
PPD-2, whose antibacterial spectrum was found to be distinct from
the parent compound, which correlated to changes in target specificity.
The highly attractive chemical properties of these compounds,
including potential for extensive medicinal chemistry exploration,
are fairly unusual at this stage of discovery. This fact, combined with
intriguing preliminary structure–activity relationship results, suggests
that further characterization of the PPD series may be warranted as
part of the global search for novel antibiotics.
The importance of identifying new antibacterial agents employing
novel mechanisms of action cannot be overstated. Bacterial resistance
to currently available therapeutic agents continues to be a serious and
growing problem, with calls for action now being made routinely by
both private and public sectors, including the Centers for Disease
Control and the World Health Organization.1–4 In the search for novel
compounds, it is important to remember that nearly every known
antibiotic class was first identified by its ability to kill bacteria; in some
cases, the mode of action was not determined until many years later.5
In contrast, target-based screening strategies to identify inhibitors of
essential proteins have been less successful in generating novel
antibiotics with desirable attributes.6 Although potent inhibitors of a
number of attractive targets have been identified with this
approach, their development into drugs has often been hindered by
the inability of medicinal chemistry to ‘design in’ bacterial cell
penetration. Accordingly, a simple phenotypic screen for inhibition
of bacterial growth was conducted to identify compounds, which were
preselected for their promising physical chemical properties, for
whole-cell antibacterial activity; the mode of action of hit compounds
was subsequently determined using a ‘reverse genomics’ platform
that combined several biochemical and genetic approaches.7
The following describes the discovery and characterization of one
MATERIALS AND METHODS
Whole-cell antibacterial high throughput screen
Single-use frozen stocks of the P. aeruginosa lab strain wild-type PAO1
(American Type Culture Collection, Manassas, VA, USA) were thawed and
diluted to 1x105 CFU ml− 1 in Mueller Hinton Broth. Twenty μl was then added
to individual wells of 384-well Costar plates (catalog #3701, VWR, Radnor, PA,
USA) with each well containing 0.5 μl of test compound at 2 mM in dimethyl
sulfoxide (DMSO), resulting in a final screening concentration of 97 μM in 2.4%
DMSO. Plates were covered and placed in a humidified 37 °C incubator for
18–22 h, after which the OD620nm was determined using a Spectramax
multiscan plate reader (Molecular Devices, Sunnyvale, CA, USA). Any well
with an OD620nm value of o30% that of growth control wells (containing no
such compound, PPD-1, a novel pyrazolopyrimidinedione with compound) was considered a ‘hit’. Approximately 1.3 million compounds
1Department of Medicinal Chemistry, Antibacterials Research Unit, Pfizer Worldwide Research and Development, Groton, CT, USA and 2Department of Discovery Biology,
Antibacterials Research Unit, Pfizer Worldwide Research and Development, Groton, CT, USA
3Current address: Pfizer Global Biotherapeutics, Cambridge, MA, USA.
4Current address: Global Therapeutics Research, Zoetis, Kalamazoo, MI, USA.
5Current address: AstraZeneca Infection Innovative Medicines, Waltham, MA, USA.
Correspondence: Dr AA Miller, AstraZeneca Infection Innovative Medicines, 35 Gatehouse Drive, Waltham, MA 02451, USA.
E-mail: alita.miller@astrazeneca.com
Received 28 July 2014; revised 27 October 2014; accepted 2 November 2014

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
2
proprietary to Pfizer were screened, resulting in an average Z-value of 0.81 and containing appropriately diluted test compounds (ranging from 300 to
a hit rate of ~ 0.02%. Hit confirmation and preliminary assessment of the
0.03 μM). All compounds were initially solubilized in 100% DMSO and tested
spectrum were performed by testing each compound for activity over eight in triplicate. The final concentration of DMSO in each well was 1.0%.
twofold-concentration dilutions against the following common laboratory Following a 72 -h exposure to compounds at 37 °C and 5% CO2, cell viability
strains in the same manner as described above: Pseudomonas aeruginosa in each well was determined by measuring the concentration of cellular ATP
(PAO1), Staphylococcus aureus (RN4220), Klebsiella pneumonaie (KP3700)
and Acinetobacter baumannii (AB3176).
using the Cambrex Vialight Plus Cell Proliferation/Cytoxicity Kit as described
by the manufacturer (Lonza, Allendale, NJ, USA). ATP concentration was
determined by reading luminescence in an Envision plate reader (Perkin Elmer,
Waltham, MA, USA). Percent of viable cells relative to untreated controls was
determined for each well. Final IC50 value corresponds to the dose projected to
kill 50% of the cells following a 72-h exposure.
MIC determination
The MIC necessary to inhibit bacterial growth was determined for compounds
of interest by broth microdilution according to standard protocols.8
Cytotoxicity assay
Macromolecular synthesis assay
ThlE-2 cells (American Type Culture Collection) were seeded at a density of Inhibition of specific macromolecular synthesis pathways was assessed by
2500 cells per well in 384-well clear-bottom tissue culture plates containing measuring the incorporation of radiolabelled precursors in a growing culture of
25 μl of standard cell culture medium. After 24 h of incubation at 37 °C and 5% Eschericia coli JL553 (tolC, lysE) as follows. MOPS-defined media (Teknova,
CO2, the medium was removed and replaced with 25 μl of fresh medium
Hollister, CA, USA) was inoculated with an overnight culture of E. coli JL553
Macromolecular Synthesis Assay
H
N
O
N
NH
N
O
100
50
0
Fatty Acid
DNA
RNA
Cell Wall
Protein
0
2
6
18
-1
52
160
µg/mL PPD
TnT Staging Assay
Figure 1 PPD-1 is a protein synthesis inhibitor. (a) E. coli treated with increasing amounts of PPD-1 resulted in a dose-dependent inhibition of incorporation of
amino-acid precursors in a macromolecular synthesis assay (Y axis corresponds to percent of wild-type control). Inset shows the structure of PPD-1. (b) PPD-1
was compared with known RNA and protein synthesis inhibitors in a transcription–translation (TnT) staging assay to assess the point at which inhibition likely
occurs. Activity is shown as relative luminescence units (RLUs). A full color version of this figure is available at The Journal of Antibiotics journal online.
The Journal of Antibiotics

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
3
and incubated at 200 r.p.m. at 37 °C to an OD600 of 0.4–0.5. A 100-μl aliquot of The reaction was started by adding 25 μl of the reaction mix (containing the
this culture was added to a 96-well plate containing 5 μl of antibiotic stocks to
result in a final concentration range of 0.5–20× the MIC and incubated for
10 min with gentle shaking at 35 °C. Ten μl of 14C-radiolabelled precursors for
each macromolecular pathway were added as follows: DNA, 1:4 dilution of
thymidine (50 μCi ml− 1, Amersham, GE Healthcare, Pittsburgh, PA, USA, cat #
CFA-219); RNA, 1:100 dilution of uridine (50 μCi ml− 1, Amersham cat #
CFB-51); protein, 1:55 dilution of l-leucine (50 μCi ml− 1, Amersham cat #
CFB-183); fatty acid, 1:20 dilution of sodium acetate (200 μCi ml− 1, Amersham
cat # CFA-14) and cell wall, 1:20 dilution of diaminopimelic acid (1000
μCi ml− 1, Moravek Biochemicals, Brea, CA, USA, cat # MT-1556). Plates were
incubated for an additional 25 min at 35 °C with gentle shaking. Following
incubation, 100 μl of ice-cold 25% trichloroacetic acid was added to each well,
and the plates were incubated on ice for 1 h. Incorporated counts were
harvested with the Packard FilterMate-96 harvester (Perkin Elmer) using
‘UniFilter GF/B’ filter plates (Perkin Elmer). The filter plates were pre-
washed with 5% trichloroacetic acid, samples were filtered through and then
washed twice with ice-cold 5% trichloroacetic acid. A final wash with 10%
ethanol was performed, and the plates dried under a heat lamp for 5 min and
then allowed to dry for an additional ∼ 10 min at room temperature. The
bottoms of the plates were sealed, and 40 μl of MicroScint scintillation fluid
(PerkinElmer) was added per well. The plates were then top-sealed with
MultiScreen sealing tape (Thermo Fisher, Rockville, MD, USA) and counted on
the TopCount (PerkinElmer), 1 min per well. Experiments were conducted at
least in triplicate. Data analysis was performed using GraphPad Prism (La Jolla,
CA, USA).
S30 fraction and extra cofactors as described above) to all the wells. Five μl of
inhibitor stock was then added to the other wells at various time points as
indicated. After a total of 60 min of incubation at room temperature, the
reaction was stopped with the addition of 35 μl of LucLite solution (Perki-
nElmer) per well. After 7 min of dark adjustment, the amount of luminescence
was read on a Micro-Beta Trilux (Wallach, Sweden). Data were generated in at
least duplicate.
Cloning, expression and purification of aminoacyl tRNA
synthetases
For multicopy expression experiments, the lysRS gene from RN4220 was cloned
into the xylose-inducible expression vector pHIS1525 and transformed into
S. aureus RN4220. To generate purified tRNA synthetases, S. aureus lysRS and
E. coli proRS (the latter of which had been amplified from purified E. coli
W3110 genomic DNA) were cloned into the expression vector pET15b.
N-terminally his-tagged lysyl or prolyl tRNA synthetases (ProRS) were over-
expressed upon isopropyl β-D-1-thiogalactopyranoside induction and purified
using standard conditions as described by the manufacturer (Clontech,
Mountainview, CA, USA).
Aminoacyl tRNA synthetase assay
Aminoacyl tRNA synthetase activity was measured in vitro by determining the
amount of radiolabelled lysyl or prolyl tRNA that was incorporated over time to
a mixture of E. coli tRNAs as follows. First, the following reaction mix was
preincubated for 10 min at room temperature: 50 mM HEPES, pH 7.6, 10 mM
MgCl2, 2 mM ATP, 10 mM βME, 2 mg ml− 1 purified E. coli tRNAs (Sigma-
Aldrich, St Louis, MO, USA), 70 nM purified prolyl or lysysl tRNA synthetase
and 0.5 μM inhibitor in a total volume of 50 μl. The reaction was initiated by
addition of 0.05 μCi (3.12 μM) [C14] labeled Proline or Lysine (Amersham). The
reaction was carried out at 35 °C for 20 min and quenched by addition of 200 μl
ice-cold ethanol, then placed at 4 °C for an additional 30 min to complete
precipitation. RNA was filtered through a 96-well GF/B filter plate, which was
pre-wet with 10 mM unlabeled lysine or proline, and the plate was washed twice
with cold ethanol to remove unbound radiolabeled substrate. Scintillation fluid
(50 μl) was added, and 14C counts were measured on the TopCount reader for
120 s. Activity was reported as percent control (for example, number of counts
obtained for inhibitor-treated samples as compared with the uninhibited
control sample).
Transcription–translation (TnT) assay
Inhibitors of bacterial protein synthesis were identified and characterized by
their ability to inhibit the production of functional luciferase in a cell-free
coupled transcription and translation system. The assay system employs
plasmid DNA containing the firefly luciferase gene, an S30 fraction (E. coli
cytosol enzymes and ribosomes) and a buffer mix containing necessary
transcription and translation reagents and cofactors. Although this assay is
commercially available, the results described here were generated using reagents
produced in-house, as previously described,9 with the following modifications:
E. coli BL-21(DE3) grown in M9 minimal media was used to generate S30
fractions and E. coli JM109/pBestLuc plasmid grown in LB media was used for
plasmid maintenance and purification. Where indicated, increasing amounts of
excess lysine or proline were added up to 150 μM. Each compound was tested in
a 10-point, 1:2 serial dilution format starting at 40 μM and diluting down to
0.078 μM, in at least duplicate.
Resistance mapping of PPD-1 in Neisseria gonorrhoeae
An error-prone PCR library using 10-kb PCR amplicons encompassing the
entire genome was generated in Neisseria gonorrhoeae strain NG2888 (ATCC) as
described.10 Briefly, these amplicons were then used as donor DNAs to
transform NG2888 in 96-well plates, and cells were plated on 2, 4 and 8 times
the MIC. These studies identified two overlapping regions of the genome, both
containing lysRS. To identify the mutation responsible for resistance, the
overlapping region was PCR-amplified using individual resistant mutants as
templates, and the DNA sequence was compared with wild-type NG2888.
Finally, these individual amplicons were used as donor DNAs to create isogenic
strains containing the mutation of interest.
Staging assay
The staging assay is a modification of the TnT assay, where excess inhibitor is
added at various times during the incubation of a normal TnT assay and the
amount of reporter enzyme is determined at the end of incubation. The activity
of different types of inhibitors (transcription, translation initiation, translation
elongation, reporter enzyme inhibition, etc.) can be discerned when reporter
activity is graphed vs inhibitor time of addition, because the earlier an inhibitor
acts in the protein synthesis process, the greater the amounts of enzyme that are
present at the end of a given incubation period (Figure 1b). Exploratory
compounds and known inhibitors were first tested in the TnT assay described
above to generate an IC50 value. The compounds were then diluted with an
appropriate volume of a H2O/DMSO mixture to generate a compound
concentration of 140 times the IC50 value in o14% DMSO (for example,
the IC50 of azithromycin in the TnT assay is 0.19 μM, so a stock solution
(14 mM) was diluted to 22.4 μM in 13% DMSO). This concentration corre-
sponded to a final azithromycin concentration of 3.2 μM during the staging
assay (that is, 20 times the IC50 value), when 5 μl of compound was added to a
final assay volume of 35 μl. Five μl of 0.2 mg ml− 1 pBestLuc plasmid (1 μg of
plasmid) was added to each of the 8 wells per compound being assayed in a
96-well plate. Five μl of the inhibitor stock was added to well 1 (such that it will
result in 20× the IC50 in the final assay volume) and was used as the time 0
control. Data generated for this well was used to normalize (by subtraction) all
Resistance mapping of PPD-2 by multicopy suppression in E. coli
Spontaneous mutants resistant to PPD-2 were generated using E. coli W3110
tolC::Tn10. Briefly, 2x109 cells were plated on cation-adjusted Mueller Hinton
Agar plates (Difco, Becton Dickinson, Houston, TX, USA) containing a range
of PPD-2 concentrations. Plates were incubated at 37 °C for 40 h before
counting colonies and calculating resistance frequencies. Genomic DNA from a
colony recovered from the 8 μg ml− 1 PPD-2 plate was extracted, partially
digested with Sau3AI and ligated to BamHI-digested pUC19 DNA. The
resulting genomic library was electroporated into susceptible E. coli W3110
tolC::Tn10 cells and plated on Mueller Hinton Agar plates containing increasing
concentrations of PPD-2. Plasmid DNA from the resulting PPD-2 colonies was
the other data points corresponding to increasing times of inhibitor additions. extracted and sequenced using M13 primers.
The Journal of Antibiotics

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
4
Table 1 The MIC (in μg ml− 1) of PPD-1 antibacterial activity as
compared with ciprofloxacin and linezolid against the laboratory
S. aureus strain RN4220 and various drug-resistant S. aureus clinical
isolates
S. aureus
strain
Resistance genotype
Ciprofloxacin Linezolid PPD-1
RN4220
None
MRSA, quinR
0.125
16
1
1
8
4
8
8
8
8
Scheme 1 Synthesis of PPD-1 (a) Methylethyl ketone, NaOEt/EtOH, reflux;
(b) (cyclobutylmethyl)hydrazine, AcOH; (c) HNO3, H2SO4, 40 °C ; (d) NH3,
MeOH; (e) H2, Pd/C, EtOH; (f) CDI, NEt3, MeCN, reflux.
USA300
SA1900-10
SA1281-07
SA1851-09
SA1609-09
βla, MRSA, mecA, macrolideR
MRSA, quinR, cfr+ (linR)
MRSA, quinR, cfr+ (linR)
MRSA, quinR, cfr+ (linR)
macrolideR
464
464
464
464
4
16
8
16
Abbreviations: MRSA, methicillin-resistant Staphylococcus aureus ; PPD,
pyrazolopyrimidinedione.
LysRS
250
260
270
280
EC orthologAPELYLKRLVVGGFERVFEINRNFRNEGISVRHNPEFTMMELYMAYADY
SA ortholog
AIELHLKRLIVGGLEKVYEIGRVFRNEGVSTRHNPEFTMIELYEAYADY
NG ortholog
APELYLKRLVVGGLERVFEINRSFRNEGMSVRHNPEFTMIEFYEAFSDY
APELYLKRLVVGGLERVFEINR FRNEGISVRHNPEFTMIELYEAYADY
M266L
Scheme 2 Synthesis of PPD-2 (a) CH3COCl, MeOH,
0 °C-rt, 20 h;
ProRS
130
140
150
160
170
(b) K2CO3, KI, (bromomethyl)cyclobutane, DMF, 60 °C , 24 h; (c) 7 M NH3,
MeOH, 80 °C , 8 h; (d) H2,10% Pd/C, EtOH, rt, 8 h; (e) CDI, TEA, MeCN,
reflux; (f) NBS, AcOH, MeCN, reflux; (g) Pd2(dba)3, PCy3, K3PO4, dioxane/
water,pyridine-4-boronic acid, reflux, 1.5 h. rt, room temperature.
SYKQLPLNFYQIQTKFRDEVRPRFGVMRSREFLMKDAYSFHTSQESLQETYDAM
EC ortholog
NSYKQLPINFYQIQTKFRDEIRPRFGLMRGREFIMKDAYSFHLSQDSLQQTYDGM
SA ortholog
KSYKQLPMTLFQIQSKFRDEKRPRFGLLRGREFIMKDAYSFHADEASLDQTYQDM
NG ortholog
SYKQLPINFYQIQTKFRDEIRPRFGLMRGREFIMKDAYSFH SQDSLQQTYDAM
Generation of PPD analogs
F155L
The chemical syntheses of PPD-1 and PPD-2 are described in Schemes 1 and 2,
respectively. The experimental procedures and subsequent characterization of
these molecules are summarized in detail in the Supplementary Data. Other
PPD analogs were prepared using analogous routes. For PPD-1 (Scheme 1),
substituted pyrazole Int-3 was constructed by reaction of diketoester Int-2 with
cyclobutylmethyl hydrazine. Nitration of the pyrazole ring, followed by
amidation of the ester gave Int-5. Treatment of Int-5 with carbonyldiimidazole
under basic conditions gave PPD-1. Compounds 3–11 can be made in a similar
fashion by first constructing the appropriately substituted pyrazole ring system.
For PPD-2 (Scheme 2), Int-8 was prepared by alkylation of pyrazole Int-7 with
bromomethyl cyclobutane. The desired regioisomer was advanced to ring
closed product Int-11 using identical conditions as those described for the
synthesis of PPD-1. Treatment of pyrazole Int-11 with N-bromosuccinimide
gave bromide Int-12, which can be converted to PPD-2 via Suzuki coupling.
Analogs 12–19 can be synthesized similarly by Suzuki coupling of bromide Int-
12 with the appropriate coupling partner.
Figure 2 Mutations in distinct tRNA synthetase genes confer resistance to
PPD-1 and PPD-2. Resistance to PPD-1 in N. gonorrhoeae (NG) results from
a M266L mutation in the active site of lysyl tRNA synthetase (LysRS) (a),
whereas resistance to PPD-2 in E. coli (EC) results from a F155L mutation
in prolyl tRNA synthetase (ProRS) (b). These targets were subsequently
confirmed in tRNA-incorporation assays using purified S. aureus (SA) LysRS
or E. coli ProRS enzymes (Tables 3A and B).
(~16–32 μg ml− 1). PPD-1 was also tested in a standard cytotoxicity
(ThlE) assay (Thermo Scientific, Rockville, IL, USA) for its ability to
inhibit eukaryotic cell proliferation and was found to have no effect up
to 300 μM (data not shown). Further assessment of its activity showed
that PPD-1 inhibited the growth of a number of multidrug-resistant S.
aureus strains, including those bearing the cfr plasmid,11 which confers
resistance to linezolid and other antibiotics (Table 1).
RESULTS
PPD-1 is a protein synthesis inhibitor
Discovery and antibacterial spectrum of PPD-1
Investigation into the mechanism of action of PPD-1 began with an
evaluation of its ability to inhibit a variety of macromolecular synthesis
pathways using the well-established assay in E.coli. PPD-1 was found
to specifically inhibit protein synthesis with little-to-no effect on DNA,
RNA, cell wall or fatty acid biosynthesis pathways (Figure 1a). PPD-1
was then tested in an E. coli TnT assay and found to be active with an
IC50 of 60.7 μM, which correlates well with its MIC. To further define
the point at which PPD-1 inhibits translation, the activity of the
compound was compared with other known translation inhibitors in a
TnT staging assay, which divides the transcription and translation
reaction into biochemically distinct steps (Figure 1b). This assay can
A high-throughput antibacterial screen was conducted to identify
compounds that inhibited the growth of PAO1, a wild-type strain of
Pseudomonas aeruginosa. PPD-1, a pyrazolopyrimidinedione with MW
of 248.3 and ClogP of 1.5 (Figure 1a, inset), was found to be active in
this screen, showing 63.8% inhibition of growth at 98 μM. Upon hit
confirmation, PPD-1 was found to have an IC50 of 46.4 μM against
wild-type PAO1. Expanded assessment of antibacterial activity revealed
the compound was also active against S. aureus, demonstrating an IC50
of 20 μM (~5 μg ml− 1) against the laboratory strain RN4220. Although
no antibacterial activity was observed up to 100 μM vs K. pneumoniae
and A. baumannii, PPD-1 had modest activity against laboratory
strains of pump-deficient E. coli (tolC− ) with MICs of ~ 64–128 μM distinguish
a general protein synthesis inhibitor as either a
The Journal of Antibiotics

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
5
transcription or a translation inhibitor, and indeed, can further classify Resistance to PPD-1 maps to lysRS
it as inhibiting earlier or later in the translation process. This is An error-prone PCR platform in Neisseria gonorrhoeae called RATE
accomplished by adding the inhibitor at 20–25 times the IC50 (for Rapid Antibacterial Target Elucidation)10 had been established by
concentration (effectively shutting down translation) at various times our group as one component of our standard mode of action
during the TnT assay incubation and measuring how much translation determination strategy, owing to its rapid turnaround time and
product is made (by measuring the amount of reporter activity) minimal compound requirement. This approach was therefore used
relative to control compounds. The activity of different types of as a first step to genetically identify the molecular target of PPD-1
inhibitors (for example, those that inhibit transcription, translation inhibition. Stable resistant mutants in N. gonorrhoeae were generated
initiation, translation elongation, reporter enzyme inhibition, etc.) can that had a fourfold increase in the MIC to PPD-1 as compared with
be discerned when reporter activity is graphed as a function of the the parent strain. Sequence analysis showed that the methionine at
time of addition of the inhibitor, because the amount of enzyme residue 266 in the active site of lysyl tRNA synthetase (LysRS) had
activity after a given incubation time is inversely proportional to been replaced with a leucine (Figure 2, upper panel), suggesting LysRS
the stage at which the inhibitor acts. In other words, the later in the was the target of inhibition of PPD-1. To corroborate these results in
protein synthesis process that an inhibitor acts, the lower is the S. aureus, the lysRS gene from S. aureus was cloned in the xylose-
reporter activity that is observed (Figure 1b). The activity of PPD-1 in inducible expression vector pHIS1525 and transformed into S. aureus
this staging assay was found to correlate closely to that of azithromycin RN4220. Upon induction with 0.5% xylose, overexpression of LysRS
and chloramphenicol, which inhibit peptidyl transferase by binding at resulted in a fourfold increase in the MIC of PPD-1 (but not other
different locations on the 50S ribosomal subunit.12 These data protein synthesis inhibitors) as compared with vector alone (data not
suggested that the mechanism of action of PPD-1 may involve a shown), suggesting that LysRS is also a target of PPD-1 inhibition in
similar stage of protein synthesis.
S. aureus. PPD-1 was then tested for its effect on the activity of
Table 2 Structure–activity relationships of PPD analogs
aEC¼ MIC (mg mlꢀ1) for either E. Coli MC4100 or EC-2880.
bSA¼ MIC (mg mlꢀ1) for either S. aureus RN4220 or USA300.
The Journal of Antibiotics

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
6
Table 3A Chemical properties and biological activities of PPD-1 and Table 3B Chemical properties and biological activities of PPD-1 and
PPD-2 (summary of the chemical properties of PPD-1 and PPD-2)
PPD-2 (relative ability of PPD-1 and PPD-2 to inhibit bacterial growth
and various in vitro activity assays)
Chemical property
PPD-1
PPD-2
Biological activitya
PPD-1
PPD-2
MW
248.13
1.51
0.75
84
297.12
1.32
0.28
96
ClogP
ClogD 7.4
TPSA
MIC, S. aureus
8
464
1
MIC, E. coli (tolC−
)
32
IC50 TnT (E. coli)
53
9
IC50 TnT (E. coli)+excess lysine
55.5
54.2
21.7
143.4
35
10.75
46.8
151
216
0.044
0.215
Abbreviations: PPD, pyrazolopyrimidinedione; TPSA, topological polar surface area.
IC50 TnT (E. coli)+excess proline
IC50, lysyl tRNA synthetase (S. aur)
purified S. aureus LysRS in vitro and was found to inhibit that enzyme
with an IC50 of 21.7 μM, confirming the resistance mapping and
multicopy suppression results.
IC50, lysyl tRNA synthetase (S. aur)+excess ATP
IC50, prolyl tRNA synthetase (E. coli)
IC50, prolyl tRNA synthetase (E. coli)+excess ATP
90.1
Abbreviations: PPD, pyrazolopyrimidinedione; TnT, transcription–translation.
Structure–activity relationship of pyrazolopyrimidinedione analogs
The PPD template is highly amenable to derivatization through a
variety of chemical transformations, and various analogs were synthe-
sized and tested (Table 2). Initially, the R4 side chain was varied to
larger (compounds 3 and 4) and smaller (compound 5) carbocycles,
and while 5- and 6-membered rings were equally active for E. coli, the
cyclobutyl derivative PPD-1 proved superior for S. aureus. Next, methyl
caps were added to the pyrimidinone nitrogens, but compounds 6 and
7 were both inactive. The R3 substituent was removed in compound 8
and varied in size in compounds 9–11, but the ethyl substituent at R3 in
PPD-1 was the most active for S. aureus. A breakthrough in E. coli
activity came when heteroaromatic groups (specifically with a nitrogen
in the 4th position) were introduced. A simple pyridine (PPD-2)
reduced the E. coli MIC to 1 μg ml− 1 and pyrrolopyridine 16 had an
MIC of 0.25 μg ml− 1; however, the aryl derivatives (14–19) had
reduced S. aureus activity. Finally, the pyrazole alkylation regioisomer
(compound 20) and some pyrazolopyrimidinones (21–23) were
synthesized, but were completely devoid of activity.
aValues shown are averages of at least three replicates (shown in μg ml− 1 for MICs and μM for
TnT and tRNA synthetase assays). Where indicated, excess lysine or proline was added to the
TnT assay to a final concentration of 150 μM. Increasing amounts of ATP was added to the
tRNA synthetase assays as described in the text; data shown in the table correspond to the
highest amount of ATP added, or 1.5 mM (12-fold excess).
Preliminary analysis supports differential modes of inhibition
by PPD compounds
As described above and summarized in Table 3B, the relative potency
in antibacterial activity for PPD-1 and PPD-2 correlates with their
in vitro inhibitory activities. However, although resistance mapping
indicated that the primary target of PPD-1 is LysRS, it also demon-
strated some activity against ProRS, whereas PPD-2 preferentially
inhibited ProRS, suggesting a somewhat divergent mechanism of
action. In order to confirm this hypothesis, both compounds were
tested in the TnT assay in the presence and absence of excess amounts
of either proline or lysine. As shown in Table 3B, the inhibitory
activity of PPD-1 in the TnT assay was not affected by addition of
either of these amino acids in excess, whereas the activity of PPD-2
decreased significantly by the addition of excess proline but not lysine,
which suggests that proline specifically competes with PPD-2 at the
active site. The compounds were also evaluated for their ability to
compete with ATP in each of the aminoacyl tRNA synthesis assays
described above. Addition of excess ATP led to a loss of activity for
both compounds against both types of enzymes with the relative effect
corresponding to relative potency for each compound (Table 3B).
Therefore competition with ATP is likely to have a role in the
mechanism of inhibition of both these compounds.
As this investigation led to a number of compounds that gained
Gram-negative activity at the expense of Gram-positive activity, we
selected one of the most potent compounds, PPD-2 (TnT IC50 = 9
μM), for further characterization. Table 3A shows a comparison of the
chemical properties of PPD-1 and PPD-2.
Resistance to PPD-2 maps to proRS
The increased Gram-negative antibacterial activity of PPD-2 was
exploited to map resistance directly in E. coli using a mutant genomic
library. Spontaneously resistant mutants were generated by plating
E. coli W3110 tolC::Tn10 on increasing amounts of compound. These
stably resistant mutants were generated at a frequency of 3.3 × 10− 9 at
8 × MIC, and demonstrated an 8–16-fold increase over the MIC of
the parent strain. Plasmid libraries were then made from genomic
DNA, which was purified from either the sensitive parent strain or a
resistant isolate. The library ligations were transformed into W3110
tolC::Tn10 and plated on selective agar containing increasing concen-
trations of PPD-2. The parental genomic library yielded no colonies,
but the resistant mutant genomic library gave two resistant clones at
8 μg ml− 1 of compound. These plasmids were isolated and retrans-
formed into naive cells to confirm that resistance was plasmid
mediated. The two clones were then sequenced and found to have
overlapping inserts, both of which containing a mutated proRS
(encoding a prolyl tRNA synthetase) with a phenylalanine replacing
DISCUSSION
The results presented here describe the identification of a novel series
of whole-cell active aminoacyl tRNA synthetase inhibitors using a
reverse genomics approach. Although the hit compound, PPD-1, was
more potent against S. aureus and showed a more balanced activity in
the purified LysRS and ProRS assays in vitro, structure–activity
relationship studies led to an interesting divergence of activity,
with PPD-2 gaining Gram-negative antibacterial potency and sub-
micromolar activity against purified E. coli ProRS. These preliminary
results suggest there may be potential within this series to engineer the
desired spectrum of activity, potentially both at the target and the
species level. In addition to their antibacterial properties, compounds
in this series are both ‘rule-of-five’ compliant,13 suggesting the
potential for oral bioavailability, and are highly amenable to additional
the leucine at residue 155 (Figure 2, lower panel). The putative target medicinal chemistry investigation. Finally, preliminary computational
in E. coli was confirmed in vitro using a purified ProRS activity assay, modeling of the inhibitor binding to the active site (data not shown)
with potent inhibition by PPD-2 observed, corresponding to an IC50 suggests additional opportunities for exploration by medicinal chem-
of 0.044 μM.
istry; however, further analysis of this type would be required to
The Journal of Antibiotics

Novel pyrazolopyrimidinediones inhibit tRNA synthesis
JI Montgomery et al
7
confirm this hypothesis. Unfortunately, this and any other additional
research on this series was not possible given the decision by Pfizer to
deprioritize antibacterial discovery in 2011.
1
2
Boucher, H. et al. Bad bugs no drugs, no ESKAPE. Clin. Infect. Dis. 48, 1–12
(2009).
Boucher, H. et al. 10 x '20 Progress–development of new drugs active against gram-
negative bacilli: an update from the Infectious Diseases Society of America. Clin. Infect.
Dis. 256, 1685–1694 (2013).
There are several reasons why aminoacyl tRNA synthetases have the
potential to be good targets for inhibition by novel antibacterial
agents14,15 and a number of discovery efforts have indeed been
conducted around tRNA synthetase inhibitors over the years with
varying success.16 These enzymes have a pivotal role in protein
biosynthesis by catalyzing the transfer of amino acids to their cognate
tRNA and therefore are necessary for bacterial growth and survival.
Despite their conserved functionality, considerable divergent evolution
has resulted in significant structural differences between prokarytic
and eukaryotic orthologs, which can be exploited for selectivity.17 In
addition, a number of natural products, such as albomycin,18
indolmycin19 or microcin C20 have been shown to kill bacteria via
inhibition of tRNA synthetases, suggesting that these enzymes serve as
an important Achilles’ heel for bacterial survival. Finally, the tRNA
synthetases have been clinically validated for Gram-positive skin
infections by the topical use of mupirocin (pseudomonic acid) for
many decades.21,22 However, as for inhibitors of other highly
conserved single-gene targets, the potential liability of rapid resistance
emergence cannot be overlooked. The most notable recent example
was the very promising oxaborole-containing LeuRS inhibitor,
GSK2251052, which demonstrated potent preclinical activity against
MDR Gram-negative bacteria23 and promising exposure and safety in
phase I clinical trials.24 Unfortunately, facile emergence of resistance in
the clinic led to an abrupt termination of phase II studies after several
patients developed resistant bacteria during treatment for complicated
urinary tract infections.25 Because the PPD series described above
likely has similar levels of resistance emergence as was observed for
GSK2251052 preclinically23 (3.3× 10− 9 in an E. coli tolC− strain for
PPD-2 vs 1 in 10− 7 to 10− 8 across all species tested for GSK2251052),
overcoming this phenotype would obviously be a key objective in
future project advancement. One approach to address this concern
would be to design potent inhibition of multiple tRNA synthetases
within a single compound, a strategy whose feasibility is supported by
the preliminary data presented above.
3
Centers for Disease Control and Prevention Executive Report: Antibiotic Resistance
Threats in the United States (2013), http://www.cdc.gov/drugresistance/threat-report-
2013/.
World Health Organization Antimicrobial resistance: global report on surveillance
(2014), http://www.who.int/drugresistance/documents/surveillancereport/en/.
Silver, L. L. Challenges of antibacterial drug discovery. Clin. Microbiol. Rev. 24,
71–109 (2011).
Payne, D., Gwynn, M. N., Holmes, D. J. & Pompliano, D. L. Drugs for bad bugs,
confronting the challenges of antimicrobial discovery. Nat. Rev. Drug Discov. 1,
29–40 (2007).
Miller, A. et al. Discovery and characterization of QPT-1, the progenitor of a new
class of bacterial topoisomerase inhibitors. Antimicrob. Agents Chemother. 52,
2806–2812 (2008).
Clinical and Laboratory Standards Institute. Methods for Dilution Antimicrobial
Susceptibility Tests for Bacteria that Grow Aerobically 9th edn Vol. 32 (CLSI: Wayne,
PA, USA, 2012).
Murray, R. W., Melchior, E. P., Hagadorn, J. C. & Marotti, K. R. S. aureus cell extract
transcription-translation assay: firefly luciferase reporter system for evaluating protein
translation inhibitors. Antimicrob. Agents Chemother. 45, 1900–1904 (2001).
4
5
6
7
8
9
10 Hartman-Neumann, S., Dunham, S., Gosink, M., Kuhn, M.
& Miller, A. Rapid
elucidation of antibacterial targets and resistance determinants using an error-prone
genomic library of N. gonorrhoeae, Poster F2-869, Interscience Conference on
Antimicrobial Agents and Chemotherapy, Boston, MA, USA (2010).
11 Shen, J., Wang, Y. & Schwarz, S. Presence and dissemination of the multiresistance
gene cfr in Gram-positive and Gram-negative bacteria. J. Antimicrob. Chemother. 68,
1697–1706 (2013).
12 Dunkle, J. A., Xiong, L., Mankin, A. S. & Cate, J. H. Structures of the Escherichia coli
ribosome with antibiotics bound near the peptidyl transferase center explain spectra of
drug action. Proc. Natl Acad. Sci. USA 107, 17152–17157 (2010).
13 Pollastri, M. P. Overview of the rule of five. Curr. Protoc. Pharmacol Chapter 9, Unit
9.12 (2010).
14 Lv, I. C. & Zhu, H. L. Aminoacyl-tRNA synthetase inhibitors as potent antibacterials.
Curr. Med. Chem. 19, 3550–3563 (2012).
15 Pohlmann, J. & Brotz-Osterheltz, H. New aminoacyl tRNA synthetase inhibitors as
antibacterial agents. Curr. Drug. Tar. Infect. Dis. 4, 261–272 (2004).
16 Gadakh, B. & VanAerschot, A. Aminoacyl-tRNA synthetase inhibitors as antimicrobial
agents:
a patent review from 2006 till present. Expert Opin. Ther. Pat. 22,
1453–1465 (2012).
17 Raczniak, G., Ibba, M.
& Söll, D. Genomics-based identification of targets in
pathogenic bacteria for potential therapeutic and diagnostic use. Toxicology 160,
181–189 (2001).
18 Zeng, Y. et al. Biosynthesis of albomycin δ2 provides a template for assembling
siderophore and aminoacyl-tRNA synthetase inhibitor conjugates. ACS Chem. Biol. 7,
1565–1575 (2012).
19 Hurdle, J. G., O’Neill, A. J. & Chopra, I. Anti-staphylococcal activity of indolmycin,
a
potential topical agent for control of staphylococcal infections. J. Antimicrob.
Chemother. 54, 549–552 (2004).
20 Severinov, K. & Nair, S. K. Microcin C: biosynthesis and mechanisms of bacterial
resistance. Future Microbiol. 7, 281–289 (2012).
CONFLICT OF INTEREST
All co-authors are current or former employees of Pfizer and thereby may own
21 Ward, A. & Campoli-Richards, D. M. Mupirocin: a review of its antibacterial activity,
pharmacokinetic properties and therapeutic use. Drugs 32, 425–444 (1986).
22 Sutherland, R. et al. Antibacterial activity of mupirocin (pseudomonic acid), a new
antibiotic for topical use. Antimicrob. Agents Chemother. 27, 495–498 (1985).
23 Sutcliffe, J. A. Antibiotics in development targeting protein synthesis. Ann. NY Acad.
Sci. 1, 122–152 (2011).
24 Tenaro, D. et al. Intrapulmonary pharmacokinetics of GSK2251052 in healthy
volunteers. Antimicrob. Agents Chemother. 57, 3334–3337 (2013).
25 Pucci, M. J. & Bush, K. Investigational antimicrobial agents of 2013. Clin. Microbiol.
Rev. 26, 792–821 (2013).
shares in the company.
ACKNOWLEDGEMENTS
We wish to acknowledge the former leaders of the Pfizer Antibacterial Research
Unit for their unfailing dedication and support during very challenging times.
We would also like to acknowledge our former colleague, Zuoyu Xu, for
development of the staging assay during his time at Pfizer.
Supplementary Information accompanies the paper on The Journal of Antibiotics website (http://www.nature.com/ja)
The Journal of Antibiotics