Discovery of inhibitors of 4'-phosphopantetheine adenylyltransferase (PPAT) to validate PPAT as a target for antibacterial therapy

Article abstract of DOI:10.1128/AAC.01661-13

Inhibitors of 4'-phosphopantetheine adenylyltransferase (PPAT) were identified through high-throughput screening of the AstraZeneca compound library. One series, cycloalkyl pyrimidines, showed inhibition of PPAT isozymes from several species, with the most potent inhibition of enzymes from Gram-positive species. Mode-of-inhibition studies with Streptococcus pneumoniae and Staphylococcus aureus PPAT demonstrated representatives of this series to be reversible inhibitors competitive with phosphopantetheine and uncompetitive with ATP, binding to the enzyme-ATP complex. The potency of this series was optimized using structure-based design, and inhibition of cell growth of Gram-positive species was achieved. Mode-of-action studies, using generation of resistant mutants with targeted sequencing as well as constructs that overexpress PPAT, demonstrated that growth suppression was due to inhibition of PPAT. An effect on bacterial burden was demonstrated in mouse lung and thigh infection models, but further optimization of dosing requirements and compound properties is needed before these compounds can be considered for progress into clinical development. These studies validated PPAT as a novel target for antibacterial therapy. Copyright

Full text of DOI:10.1128/AAC.01661-13

Discovery of Inhibitors of 4=-Phosphopantetheine Adenylyltransferase
(PPAT) To Validate PPAT as a Target for Antibacterial Therapy
Boudewijn L. M. de Jonge,a Grant K. Walkup,a Sushmita D. Lahiri,a Hoan Huynh,b Georg Neckermann,a Luke Utley,c* Tory J. Nash,a
Jesse Brock,a Maryann San Martin,a Amy Kutschke,a Michele Johnstone,a Valerie Laganas,a Laurel Hajec,a Rong-Fang Gu,a
Haihong Ni,b* Brendan Chen,b Kim Hutchings,b* Elise Holt,b David McKinney,b Ning Gao,a Stephania Livchak,a Jason Threshera
Infection Bioscience, AstraZeneca R&D Boston, Waltham, Massachusetts, USAa; Infection Chemistry, AstraZeneca R&D Boston, Waltham, Massachusetts, USAb; Infection
DMPK, AstraZeneca R&D Boston, Waltham, Massachusetts, USAc
Inhibitors of 4=-phosphopantetheine adenylyltransferase (PPAT) were identified through high-throughput screening of the
AstraZeneca compound library. One series, cycloalkyl pyrimidines, showed inhibition of PPAT isozymes from several species, with
the most potent inhibition of enzymes from Gram-positive species. Mode-of-inhibition studies with Streptococcus pneumoniae
and Staphylococcus aureus PPAT demonstrated representatives of this series to be reversible inhibitors competitive with phos-
phopantetheine and uncompetitive with ATP, binding to the enzyme-ATP complex. The potency of this series was optimized
using structure-based design, and inhibition of cell growth of Gram-positive species was achieved. Mode-of-action studies, using
generation of resistant mutants with targeted sequencing as well as constructs that overexpress PPAT, demonstrated that growth
suppression was due to inhibition of PPAT. An effect on bacterial burden was demonstrated in mouse lung and thigh infection
models, but further optimization of dosing requirements and compound properties is needed before these compounds can be
considered for progress into clinical development. These studies validated PPAT as a novel target for antibacterial therapy.
acterial infections remain a significant cause of mortality and cacy models through inhibition of PPAT, validating for the first
Bmorbidity worldwide, in main part due to the emergence of time PPAT as a novel target for antibacterial therapy.
resistance to clinically approved antibacterial drugs (1, 2, 3, 4). As
MATERIALS AND METHODS
a result, novel drugs are needed to treat infections caused by these
resistant isolates (5, 6). Rather than optimizing existing classes of
antibacterial agents, an alternative avenue to pursue new drugs is
(Groningen, Netherlands), was prepared according to a published
Reagents and compounds. All chemicals were from Sigma-Aldrich unless
otherwise specified. 4=-Phosphopantetheine, obtained from Syncom
to seek out targets that have not been used previously in antibac-
terial therapy, as this approach avoids cross-resistance on the
compound level as well as the target level. Advancements in
genomics, high-throughput screening of compound libraries, and
structure-based design have allowed the exploration of many
novel targets for antibacterial therapy (7, 8, 9). One of the targets
explored was 4=-phosphopantetheine adenylyltransferase (PPAT)
(8, 10). Validation of this target for antibacterial therapy remained
uncertain, however, since none of the efforts using this target re-
sulted in inhibitors that suppressed bacterial growth (8, 10, 11).
PPAT, encoded by the gene coaD, is involved in the biosynthe-
sis of coenzyme A (CoA), a cofactor that functions as an acyl group
carrier in a number of central biochemical transformations, such
as the tricarboxylic acid (TCA) cycle and fatty acid metabolism
(12, 13, 14). It forms a hexamer and catalyzes (in the biological
forward direction) the reversible transfer of the adenylyl group of
ATP to 4=-phosphopantetheine, yielding dephospho-coenzyme A
and inorganic pyrophosphate (Fig. 1) (15, 16). PPAT was shown
to be essential for growth through genetic knockout experiments
(17). It is present with high sequence conservation in most bacte-
ria (18). Since bacterial PPAT has low sequence homology with
human PPAT, identification of selective, broad-spectrum bacte-
rial PPAT inhibitors appears to be feasible (18, 19, 20). To that end,
a campaign of high-throughput screening with the AstraZeneca cor-
porate library was initiated using Streptococcus mutans PPAT. Sev-
eral hits were identified, but only one series showed potential for
broad-spectrum PPAT inhibition. This series was optimized and
yielded analogs that suppress growth in vitro and in animal effi-
method (21). Compounds used in this study (Fig. 2) were synthesized
in-house (see the supplemental material).
Bacterial strains and genetic constructs. Strains, plasmids, and oligo-
nucleotides used in this study are listed in Table 1. Overexpression of
Streptococcus pneumoniae coaD in S. pneumoniae was achieved by driving
expression from the galU promoter (22).
Measurement of cellular activity. MICs were measured according to
CLSI guidelines (23). As trailing was observed with some strains, MICs
were defined as the lowest concentration that inhibited Ͼ80% of cell
growth. MIC90s were determined on large panels of clinical strains iso-
lated from various geographic locations with different resistance geno-
and phenotypes and were defined as the lowest concentrations that inhib-
ited growth of Ն90% of the strains.
Inhibition of a human lung carcinoma cell line, A549, was measured
by using the CellTiter 96 Aqueous One Solution cell proliferation assay
Received 30 July 2013 Returned for modification 18 August 2013
Accepted 13 September 2013
Published ahead of print 16 September 2013
Address correspondence to Boudewijn L. M. de Jonge,
boudewijn.dejonge@astrazeneca.com.
* Present address: Luke Utley, Agios Pharmaceuticals, Cambridge, Massachusetts,
USA; Haihong Ni, BioDuro, Beijing, China; Kim Hutchings, Department of
Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor,
Michigan, USA.
Supplemental material for this article may be found at http://dx.doi.org/10.1128
/AAC.01661-13.
Copyright © 2013, American Society for Microbiology. All Rights Reserved.
doi:10.1128/AAC.01661-13
December 2013 Volume 57 Number 12
Antimicrobial Agents and Chemotherapy p. 6005–6015
aac.asm.org 6005

de Jonge et al.
FIG 1 Pathway of CoA biosynthesis and reaction of PPAT.
with MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)- zae. Detection of ATP formation was measured either through a hexoki-
2-(4-sulfophenyl)-2H-tetrazolium, inner salt] from Promega Corp. nase/glucose-6-phosphate dehydrogenase coupling system analogous to
(Madison, WI). A549 cells were grown at 37°C under 5% CO2 conditions that reported previously for E. coli PPAT (16) (S. mutans PPAT) or
in RPMI 1640 medium supplemented with 10% fetal bovine serum and 1 through the ATPlite detection solution system (PerkinElmer) (E. coli and
mM glutamine (Invitrogen, Carlsbad, CA) for 72 h. Cytotoxicity was H. influenzae PPAT). Levels of inhibitory activity of compounds against S.
quantified by determining the lowest compound concentration at which pneumoniae, S. aureus, and Homo sapiens PPAT were measured in the
transmission was increased by 50%.
forward direction. For S. aureus and H. sapiens, formation of pyrophos-
Cidality of compounds was measured with killing kinetic assays (24) phate was measured (after conversion to orthophosphate) using a mala-
using Staphylococcus aureus ARC516 and a LytAϪ strain for S. pneumoniae chite green molybdate solution, prepared as previously described (27).
(25).
For S. pneumoniae PPAT, formation of dephospho-coenzyme A was mea-
Spontaneous resistance frequencies for S. aureus ARC516 and S. pneu- sured through mass spectrometry analysis. (For experimental details on
moniae D39 were measured by plating in triplicate high inocula of cells on each of the assays, see the supplemental material.)
compound-containing plates at up to 8ϫ MIC. Plates were incubated at
S. aureus and S. pneumoniae PPAT mechanism of inhibition. Reac-
37°C, and colonies were counted after 48 h. Spontaneous resistance fre- tions were conducted in quadruplicate at room temperature in buffer
quencies were calculated as the ratios of average CFU in the presence and containing 50 mM MOPS (morpholinepropanesulfonic acid)-NaOH
absence of compound.
(pH 7.0), 8 mM MgCl2, 1 mM dithiothreitol (DTT), 75 mM NaCl, 0.5 mM
Cloning and production of bacterial and human PPAT. PPAT from EDTA, 0.01% (wt/vol) Brij 35, 0.4 U/ml yeast inorganic pyrophosphatase
the different organisms was cloned and overexpressed in Escherichia coli (Roche), and either 4 nM S. aureus or 3 nM S. pneumoniae PPAT (calcu-
and purified through column chromatography (see the supplemental ma- lated as monomers). A matrix of final ATP and 4=-phosphopantetheine
terial). Final proteins were determined to be Ն95% pure by SDS gel elec- concentrations was studied with 2-fold dilutions of each substrate; ATP
trophoresis, except for the S. aureus isozyme, which was obtained at was adjusted from 1,000 ␮M to 15.6 ␮M in both studies. For S. aureus, the
Ն90% purity.
4=-phosphopantetheine concentration was adjusted between 150 ␮M and
PPAT biochemical inhibition. Values for 50% inhibitory concentra- 2.3 ␮M and the inhibitor was used at concentrations of 8.1, 2.7, 0.9, 0.3,
tions (IC50s) in activity against Streptococcus mutans, E. coli, and Haemo- 0.1, and 0 ␮M. For S. pneumoniae, the 4=-phosphopantetheine concentra-
philus influenzae PPAT were measured in the reverse direction. The S. tion was adjusted with doubling dilutions between 800 ␮M and 50 ␮M,
mutans isozyme was assayed in the reverse direction because a sensitive and the inhibitor was tested at 150, 83, 47, 26, 15, and 0 nM. Reactions
and balanced (26) high-throughput screening (HTS) assay could be for- were performed in untreated clear polystyrene 384-well plates (Costar) in
matted that way at acceptable cost. Assays for E. coli and H. influenzae a 30-␮l volume with a 45-␮l malachite green molybdate solution quench.
enzymes were assayed in the reverse direction to maintain consistency of Quenched reactions were allowed to incubate for 5 min, and then absor-
analysis for the Gram-negative isozymes and to avoid issues arising from bance at 650 nm was read on a Spectramax plate reader (Molecular De-
the 4=-phosphopantetheine substrate inhibition observed for H. influen- vices). Detection standard curves were made between 0 and 30 ␮M phos-
FIG 2 Structures of compounds used in this study.
6006 aac.asm.org
Antimicrobial Agents and Chemotherapy

Inhibition of Phosphopantetheine Adenylyltransferase
TABLE 1 Strains, plasmids, and oligonucleotides used in this study
Strain, plasmid, or
oligonucleotide
Description
Source or reference
Strains
E. coli BL21(DE3)
E. coli BL21 pLysS(DE3)
S. mutans UAB159
S. aureus RN4220
E. coli strain used for protein overexpression
E. coli strain used for protein overexpression with pET23a-Spn CoaD
Source for coaD of species
EMD Chemicals
EMD Chemicals
ATCC
22
Source for coaD of species
S. pneumoniae R6
H. influenzae Rd (KW20)
E. coli MG1655 (K-12)
S. pneumoniae D39
S. pneumoniae ATCC 10813
S. pyogenes ARC838
S. aureus ARC516
E. faecium ARC521
H. influenzae ATCC 51907
E. coli W3110
Source for coaD of species
Source for coaD of species
Source for coaD of species
Used in susceptibility testing
Clinical isolate used in in vivo studies
Clinical isolate used in susceptibility testing
Clinical isolate used in susceptibility testing and in vivo studies
Used in susceptibility testing
22
22
22
22
ATCC
AZ culture collectiona
AZ culture collection
AZ culture collection
ATCC
Used in susceptibility testing
E. coli W3110 tolC gene knockout, used in susceptibility testing
Clinical isolate used in susceptibility testing
Clinical isolate used in susceptibility testing
AZ culture collection
AZ culture collection
AZ culture collection
E. coli ARC534
C. albicans ARC527
Plasmids
pET23a
pET30a
pCR4-TOPO
General overexpression plasmid
General overexpression plasmid
Subcloning plasmid
EMD Chemicals
EMD Chemicals
Invitrogen
pET30a-Eco coaD
pET30a-Hinf coaD
pET30a-Hsap coaDE
pET30a-Sau coaD
pET30a-Smu coaD
pET23a-Spn coaD
pBA467-Spn coaD
Expression plasmid for E. coli PPAT
Expression plasmid for H. influenzae PPAT
Expression plasmad for H. sapiens PPAT
Expression plasmid for S. aureus PPAT
Expression plasmid for S. mutans PPAT
Expression plasmid for S. pneumoniae PPAT
Plasmid to construct overexpression of PPAT in S. pneumoniae
This study
This study
This study
This study
This study
This study
This study
Oligonucleotides
Eco coaD F
Eco coaD R
Eco coaD mutF
Eco coaD mutR
Hinf coaD F
Hinf coaD R
Hsap coaDE F
Hsap coaDE R
Sau coaD F
5=-GACTCATATGCAAAAACGGGCGATTTA-3= (NdeI site underlined)
5=-GTCAGTCGACCTACGCTAACTTCGCCATCA-3= (SalI site underlined)
5=-GAAATGCAGCTGGCGCACATGAATCGCCACTTAATGCCGG-3=
5=-CCGGCATTAAGTGGCGATTCATGTGCGCCAGCTGCATTTC-3=
5=-GCTACATATGACGAGCGTGATTTATCC-3= (NdeI site underlined)
5=-GCTAGTCGACTCATCGTGCTTTTAACGCAT-3= (SalI site underlined)
5=-CACCATGGCCGTATTCCGGTCGGGTCTCC-3= (NcoI site underlined)
5=-ACGTGTCGACTCAGTCGAGGGCCTGATGAGTCTTGG-3= (SalI site underlined)
5=-ACGTCATATGGAACATACAATAGCGGTC-3= (NdeI site underlined)
5=-ACGTGTCGACTTACTTAAATTTCTTCTTCAATGCC-3= (SalI site underlined)
5=-GACTCATATGTCAGATAGAATTGGACTC-3= (NdeI site underlined)
5=-GATCGTCGACTTAAATTTTTTGTTTGTTTT-3= (SalI site underlined)
5=-ACGTCATATGTCAGATAAGATTGGCTTATTC-3= (NdeI site underlined)
5=-ACGTGTCGACCTAATCTTTTTTTTCATTTCTTATTTCC-3= (SalI site underlined)
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
Sau coaD R
Smu coaD F
Smu coaD R
Spn coaD F
Spn coaD R
a AZ, AstraZeneca R&D Boston.
phate under both low- and high-ATP and -4=-phosphopantetheine
conditions to confirm the linearity of the detection system under all assay
conditions. Initial rates were calculated from reactions of 0, 10, 20, 30, and
40 min. Rates are expressed as nM pyrophosphate product minϪ1 per nM
enzyme (calculated as a monomer) for units of minϪ1. Initial rate data
were fitted by nonlinear least-squares regression to equations in Grafit
5.0.13 (Erithacus Software). Models for a sequential mechanism of sub-
strate binding were fitted, allowing inhibitor to bind to free enzyme
and/or ATP-enzyme complex (see equation 1 below). Here, [A], [B], and
[I] are the concentrations of the substrates and inhibitor, Kma and Kmb are
the Michaelis constants for substrates A and B, and Kia, Ki1, and Ki2 are the
dissociation constants (Kds) for substrate A from complex EA, inhibitor I
Vmax[A][B]
[I]
v ϭ
[I]
K Kmb ϩ Kma[B] 1 ϩ
ϩ K [A] 1 ϩ
ϩ [A][B]
͑
͒
͑
͒
ia
mb
ͩ ͪ
ͩ ͪ
Ki1
Ki2
Isothermal titration calorimetry. Measurements were made with a
VP-ITC instrument (Microcal/GE Healthcare, Piscataway, NJ). Solutions
of PPAT (generally less than 2 ml) were dialyzed against 2 liters of 50 mM
MOPS-NaOH (pH 7.0), 75 mM NaCl, 8 mM MgCl2, and 1 mM Tris(2-
carboxyethyl)phosphine hydrochloride overnight at 4°C. Titrations were
conducted at 25°C. PPAT protein was placed in the instrument cell at a
concentration of 60 ␮M (calculated as a monomer), and compound B was
injected at a concentration of 400 ␮M. Data were analyzed using the
from complex EI, and inhibitor I from complex EAI, respectively, as in- software provided by the manufacturer (Origin) and were fitted to both
dicated in equation 1:
single-site and two-site binding models.
December 2013 Volume 57 Number 12
aac.asm.org 6007

de Jonge et al.
Measurement of physical chemical and pharmacokinetic (PK) TABLE 2 Activity for compoundsa
properties. LogD, plasma protein binding, equilibrium solubility, and
clearance in rats for compounds were measured as described before (28).
X-ray crystallography studies. Initial crystallization conditions were
obtained using sparse matrix crystallization screens (29). The best crystals
were grown from 14% to 19% polyethylene glycol 3350 (PEG 3350)–200
mM ammonium sulfate–0.1 M propionic acid cacodylate Bis-tris pro-
pane buffer (pH 7.5) (Molecular Dimensions) at 20°C by using the vapor
diffusion sitting drop method with a drop size of 2.5 ␮l protein and a
2.5-␮l reservoir. To obtain an inhibitor-complexed PPAT structure, a 2
mM final concentration of the inhibitor was included in the protein buf-
fer. The crystals grew in 5 to 7 days and were cryoprotected with Para-
tone-N and flash-frozen in liquid nitrogen prior to data collection. Data
for the inhibitor protein cocomplex were collected at Structural GenomiX
in Advanced Photon Source. Data processing was done with MOSFLM
and scaled with Scala (CCP4 suite). The autoBUSTER software suite (30)
was used to index, integrate, and scale the data. Crystallographic refine-
ment was performed using REFMAC, and model building and placement
of water molecules was done using COOT (31). The final model quality
was assessed using PROCHECK (32). Structural analysis and figures were
prepared using Pymol.
Value for compound:
HTS
Parameter
hit
A
B
C
D
Biochemical activity
S. mutans IC50 (nM)
S. pneumoniae IC50 (nM)
S. aureus IC50 (nM)
H. influenzae IC50 (nM)
E. coli IC50 (nM)
98
205
ND
ND
ND
40
ND
0.13
0.41
2.6
0.065
0.87
120
7,300 330
8,800 860
490
9,500 740
4,800 330
ND
ND
ND
Ͼ5,000 ND
110
H. sapiens IC50 (nM)
Ͼ200,000 24,000
Cellular activity
S. pneumoniae MIC (␮g/ml) Ͼ64 Ͼ64
Ͼ64
ND
Ͼ64
ND
2
1
1
64
0.03
2
0.12
2
Ͼ64
Ͼ64
Ͼ64
S. pyogenes MIC (␮g/ml)
S. aureus MIC (␮g/ml)
E. faecium (␮g/ml)
ND
Ͼ64 Ͼ64
ND ND
ND
H. influenzae MIC (␮g/ml) Ͼ64 Ͼ64
Ͼ64 Ͼ64
ND Ͼ64
E. coli MIC (␮g/ml)
E. coli MIC (␮g/ml)
efflux-compromised
C. albicans MIC (␮g/ml)
ND
ND
Ͼ64 Ͼ64
Ͼ64 Ͼ64
In vivo efficacy studies. Activity of compounds was assessed in im-
munocompetent mouse lung and thigh infection models (22, 28). All
animal studies were approved by the Institutional Animal Care and Use
Ͼ64 Ͼ64
Ͼ64 Ͼ64
Ͼ64
Committee of AstraZeneca and were conducted in accordance with the A549 human cell line MIC
ND
ND
ND
ND
Ͼ64
32
guidelines of the American Association for Accreditation of Laboratory
Animal Care. Pathogen-free, female CD1 mice, each weighing ca. 20 g
(Charles River Laboratories), received oral administrations of aminoben-
zotriazole (ABT) at 50 mg/kg of body weight 2 h prior and 10 h postin-
fection to inhibit cytochrome P450 (CYP450) activity (33).
Infection in the lung was established by intratracheal inoculation of S.
pneumoniae ATCC 10813 (1 ϫ 106 CFU/mouse) (34). Therapy with com-
pounds C and D (Fig. 1) was started 2 h postinfection. Groups of 10 mice
each were exposed to 100 mg/kg twice daily (b.i.d.) or four times daily
(q.i.d.) every 4 h (q4h). Pyrrolamide compound 26 (35) was dosed orally
at 80 mg/kg b.i.d./q12h as a positive control. Control mice received vehi-
cles 15% SBEbCD (Captisol) for compound C and 10% 2-pyrrolidone–
0.18% polyvinylpyrrolidone-K30–0.225 M SDS for compound D q.i.d./
q4h. All treatments were administered intraperitoneally in a volume of 0.2
ml, and 24 h after the start of compound exposure, mice were euthanized
and CFU/lung were determined.
Thigh infections were established with administration of S. aureus
ARC516 postlaterally into the right thigh muscle to achieve a target start-
ing inoculum of 1 ϫ 106 CFU/thigh. Therapy with compound C was
administered intraperitoneally starting 2 h after infection, and groups of
10 mice each were exposed to 100 mg/kg twice daily (b.i.d.) or four times
daily (q.i.d.) every 4 h (q4h). Linezolid was administered at 160 mg/kg
orally once as a positive control, whereas control mice received 15% Cap-
tisol (vehicle for compound C) q.i.d./q4h. Mice were euthanized 24 h after
the start of treatment, and thighs were homogenized for CFU determina-
tions.
(␮g/ml)
Mode of action
MIC for S. pneumoniae
(␮g/ml)
MIC for S. pneumoniae (␮g/ ND
ml) PPAT-overexpressing
construct
ND
ND
ND
ND
1
8
0.03
0.25
Other
LogD
Solubility (uM)
Human protein binding—fu 4.6
(%)
Mouse protein binding—fu ND
(%)
Rat CL (ml/min/kg)
Ͼ3
8
Ͼ3
Ͻ1
0.3
0.69 0.68
850
1.5
2.4
73
4.2
Ͼ1000
2.0
13
96
ND
ND
ND
ND
11
23
ND
a For structures, see Fig. 1. ND, not determined; fu, fraction unbound.
This program was guided by structure-based design based on co-
crystal structures that were obtained with PPAT and dimethoxy-
pyrimidine analogues.
Optimization of PPAT inhibitors. Cocrystal structures ob-
tained with PPAT from S. mutans, H. influenzae, and S. aureus
showed analogues to be bound in the phosphopantetheine site, in
a pocket formed between two monomers. This is exemplified for
S. aureus PPAT through the comparative overlay of a cocrystal
structure of compound A with known substrate-bound structures
(36) (Fig. 3A). Polar and hydrophobic interactions were observed
Protein structure accession numbers. Coordinates and structure
factors have been deposited in the Protein Data Bank under accession
numbers 4NAT, 4NAH, and 4NAU.
RESULTS
Identification of PPAT inhibitors. A high-throughput screen of between compound and enzyme (Fig. 3B). Hydrogen bonds were
the AstraZeneca compound library with S. mutans PPAT identi- observed with the side chains of residues Glu135, Asn107, and
fied a dimethoxypyrimidine as an inhibitor. In addition to potent Tyr99, and the dichlorophenyl and cyclohexyl moieties of the
activity against PPAT from S. mutans (IC50 Ͻ100 nM), this HTS compound were stabilized by van der Waals interactions with
hit (Fig. 2) also showed activity against PPAT isozymes of other Leu75, Leu38, Met106, Leu103, and Ile87. The cocrystal structure
species (Table 2), indicating a potential to achieve inhibition of with compound A, obtained early on in the program, indicated
PPAT isozymes of both Gram-positive and Gram-negative species room for expansion from the pyrimidine ring toward residues in
with this series. A medicinal chemistry program was initiated to the ATP-binding pocket and opportunities to establish interac-
improve the potency of the compounds against PPAT isozymes. tions with residues from the neighboring substrate recognition
6008 aac.asm.org
Antimicrobial Agents and Chemotherapy

Inhibition of Phosphopantetheine Adenylyltransferase
FIG 3 Cocrystal structures of the inhibitors in S. aureus PPAT. (A) Structure with compounds A and B overlaid on the ATP- and phosphopantetheine-bound
structures of E. faecalis PPAT (3ND6 and 3ND7). The image represents the crystal structure of PPAT from E. faecalis in the ligand-unbound state and in complex
with ATP and pantetheine (36). The compounds bound to the S. aureus structure are depicted in the form of a stick model, while substrates ATP (blue) and
4=-phosphopantetheine (black) are shown in a line diagram. The S. aureus protein backbones of the two monomers are depicted in green and gray cartoons,
respectively. (B) View of the interaction in the binding pocket to compounds A and D showing additional interactions resulting in improved potency. Residues
with close interactions are depicted in the form of a stick diagram with the primary monomer colored in green and the adjacent monomer colored in gray.
Compound A is shown in pink and compound D is drawn as yellow sticks, while the ATP is shown in cyan. The protein backbone for compound A cocomplex
is depicted in pink, while that of compound D is depicted in green and gray.
loop (Ser 41 and Lys43) to improve potency in the series. On the (40) produced only results with significant deviations between the
other hand, few improvements in potency could be envisaged with data and fits, so Ki for the inhibitor with this enzyme is not re-
the other rings, as the tight hydrophobic pockets surrounding ported. However, reciprocal plot analysis clearly demonstrated
these groups do not allow for enhanced interactions. Analogues competitive behavior with respect to 4=-phosphopantetheine and
with substitutions on the pyrimidine ring, designed to reach to- a mixed inhibition profile relative to ATP (Fig. 4).
ward the ATP-binding pocket, were synthesized and showed, as
Analysis of S. pneumoniae inhibition data were consistent with
expected, potent inhibition of Gram-positive isozymes, with IC50s inhibitor being competitive with 4=-phosphopantetheine and un-
in the picomolar range (Table 2, compounds C and D). A cocrystal competitive with ATP (Fig. 5). The inhibition data were fitted to
structure of compound D showed that potency was gained equation 1 via nonlinear least-squares analysis, and parameters
through additional interactions with Ser41, Tyr139, and Lys43 from this model are summarized in Table 3. The more potent
(Fig. 3B). As expected, no significant potency enhancements were inhibitor binding mode (Ki2 ϭ 12.2 Ϯ 0.4 nM) was for the inhib-
achieved through modifications on the benzyl and cyclohexyl itor binding to the enzyme-ATP complex. In addition, a weaker
rings, except for an additional chloro substitution on the benzyl inhibitor binding mode (Ki1 ϭ 0.29 Ϯ 0.05 ␮M) was identified for
ring (Table 2; compare HTS hit with compound A). Whereas po- inhibitor binding to the free enzyme. The relatively weak binding
tency improvements were made against all bacterial isozymes, the of compound B to free S. pneumoniae PPAT was confirmed via
improvements were more substantial against Gram-positive isothermal titration calorimetry. In agreement with the kinetic
isozymes (Table 2).
studies, compound B was bound to S. pneumoniae PPAT in the
Mode-of-inhibition studies. Steady-state inhibition by com- absence of ATP. Analysis of binding thermograms produced sig-
pound B of the PPAT enzymes from S. aureus and S. pneumoniae nificantly better fits to a model with two different classes of bind-
was studied in the forward reaction direction using a matrix of ing sites than to a model with only one class (see Fig. S2 in the
ATP, 4=-phosphopantetheine, and inhibitor concentrations. Ini- supplemental material). Parameters determined for the first class
tial inspection of reciprocal plots clearly indicated that the enzyme of binding sites are Kd (dissociation constant) ϭ 0.27 Ϯ 0.05 ␮M,
mechanism proceeds through a ternary complex (37), which is stoichiometry (per hexamer) ϭ 2.2 Ϯ 0.08, and ⌬H ϭ Ϫ19.7 Ϯ
consistent with earlier kinetic and structural studies on the E. coli 0.3 kcal/mole; for the less potent class of binding sites, Kd ϭ 6.3 Ϯ
enzyme (16, 38, 39). For the S. aureus enzyme, the kinetics are 0.9 ␮M, stoichiometry (per hexamer) ϭ 1.9 Ϯ 0.2, and ⌬H ϭ
complex, with significant substrate inhibition observed for both Ϫ9.6 Ϯ 0.2 kcal/mole.
ATP and 4=-phosphopantetheine substrates (see Fig. S1 in the
supplemental material). In contrast, no signs of substrate inhibi- sulted in inhibition of growth for Gram-positive bacteria, with
tion were observed for S. pneumoniae. IC50s of less than 20 nM needed to achieve measurable MICs (Յ64
Cellular activity. Increased potency of enzyme inhibition re-
Attempts to fit the S. aureus data via nonlinear least-squares ␮g/ml), and continued optimization for enzyme inhibition (IC50s Ͻ
methods to common models incorporating substrate inhibition 1 nM) resulted in analogues that showed potent (Յ8 ␮g/ml) MIC
December 2013 Volume 57 Number 12
aac.asm.org 6009

de Jonge et al.
FIG 4 Reciprocal plot analysis of compound B inhibition of S. aureus PPAT. For clarity, plots are shown at only one concentration each for the paired substrates.
The concentrations of inhibitor were 0 ␮M (open circles), 0.1 ␮M (filled circles), 0.3 ␮M (open squares), 0.9 ␮M (filled squares), 2.7 ␮M (open triangles), and
8.1 ␮M (filled triangles). Panel A shows the results obtained with 37.5 ␮M 4=-phosphopantetheine; panel B shows those obtained with 0.25 mM ATP.
values (Table 2). Such analogues were active (MIC90s Յ 2 ␮g/ml) shown). Sequencing of the coaD gene from these mutants showed
against a variety of clinical Gram-positive isolates, including me- mutations resulting in amino acid modifications in PPAT (Table
thicillin-resistant and quinolone-resistant S. aureus, macrolide- 6), indicating that growth suppression in S. aureus and S. pneu-
resistant S. pneumoniae, and vancomycin-resistant enterococci, as moniae was due to inhibition of PPAT. This was confirmed for S.
illustrated for compound D (Table 4). Despite the potent activity pneumoniae through genetic transformation of a PCR product
against Gram-positive species, no measurable inhibition of containing altered coaD or chromosomal DNA from a S. pneu-
growth (MIC Ͼ 64 ␮g/ml) of Gram-negative species was seen, moniae coaD-resistant mutant into a susceptible strain, which re-
even with efflux-compromised strains (Table 2). The compounds sulted in transformants with mutated PPAT and elevated MICs
did not inhibit growth of yeast or of a mammalian cell line (MIC Ͼ against the dimethoxypyrimidines (data not shown). In addition,
64 ␮g/ml), with the exception of compound D, which showed relevant project compounds showed elevated MIC values with a
some effect against a human cell line (Table 2).
PPAT-overexpressing S. pneumoniae construct, supporting the
To confirm that bacterial growth suppression was due to PPAT notion that growth suppression occurred through PPAT inhibi-
inhibition, mutants resistant to compound D were isolated, and tion in this species (Table 2).
the coaD gene of these mutants was sequenced. Resistant mutants
Compounds in this series were bacteriostatic, with almost no
could be generated for S. aureus and S. pneumoniae at frequencies reduction in culture viability for S. aureus, even after 24 h of com-
of 10Ϫ7 to 10Ϫ9, depending on the species and the selection con- pound exposure (Fig. 6). Although sterilization was achieved for
centration (Table 5), and the MICs of these mutants were elevated S. pneumoniae in the killing kinetics experiment, the rate of killing
at least 8-fold (Table 6). Cross-resistance only to relevant project was not accelerated compared to the rate seen with an untreated
compounds and not to other chemical classes was seen (data not culture (Fig. 6). For both species, the compounds showed a delay
FIG 5 Mode-of-inhibition studies with compound B and S. pneumoniae PPAT. The dominant inhibition modality of compound B with S. pneumoniae PPAT
is uncompetitive with ATP and competitive with 4=-phosphopantetheine. Lines indicate the best fit of data to equation 1. (A) Double reciprocal plot of 1/velocity
versus 1/[ATP]. In the plot, only data for 800 ␮M 4=-phosphopantetheine are shown. The ATP concentration ranged from 15 to 1,000 ␮M, with concentrations
of inhibitor fixed at 0 nM (open circles), 15 nM (filled circles), 26 nM (open squares), 47 nM (filled squares), 83 nM (open triangles), and 150 nM (filled triangles).
(B) Double reciprocal plot of 1/velocity versus 1/[PPant]. Data for 1,000 ␮M ATP are shown. The 4=-phosphopantetheine concentration was adjusted from 50
to 800 ␮M, and inhibitor concentrations were fixed as described for panel A. The data were fitted by the nonlinear least-squares method to a sequential kinetic
mechanism, with inhibitor binding to both enzyme and enzyme-ATP complex. Initial velocity data were calculated as nM pyrophosphate produced minϪ1 per
nM of enzyme monomer minϪ1
.
6010 aac.asm.org
Antimicrobial Agents and Chemotherapy

Inhibition of Phosphopantetheine Adenylyltransferase
TABLE 3 Summary of mechanism-of-inhibition parameters for
compound B with S. pneumoniae PPAT
TABLE 5 Spontaneous resistance frequencies for compound D for S.
pneumoniae and S. aureus
Parametera
Fitted valueb
Organism
Selection condition
Frequency of resistance
kcatc (minϪ1
)
125 Ϯ 2
121 Ϯ 9
54 Ϯ 4
223 Ϯ 9
0.29 Ϯ 0.05
0.0122 Ϯ 0.0004
S. pneumoniae
1ϫ MIC
2ϫ MIC
4ϫ MIC
2.6E-7
3.7E-9
Ͻ3.1E-10
Kia (␮M)
Kma (␮M)
Kmb (␮M)
Ki1 (␮M)
Ki2 (␮M)
S. aureus
1ϫ MIC
2ϫ MIC
4ϫ MIC
1.8E-9
2.7E-9
7.7E-10
a Substrate A (ma), ATP; substrate B (mb), 4=-phosphopantetheine.
b Error data represent standard errors based on the fit of data to equation 1.
c Determined as Vmax/Etot, with Etot being the enzyme monomer concentration (3 nM).
metabolism of the compounds in order to maximize exposure
in the onset of growth inhibition (growth continued upon initial (33). While this pretreatment had limited effect on the clearance
exposure to compound; Fig. 6), which probably reflects the time of compound C, clearance of compound D was reduced 6-fold
required for depletion of an existing coenzyme A pool.
from 64 to 11 ml/min/kg of body weight, indicative of P450 me-
Efficacy in mouse infection models. Compounds C and D tabolism for this compound. For both compounds, dosing at 100
were tested in an immunocompetent mouse lung S. pneumoniae mg/kg four times a day resulted in stasis compared to the start of
infection model. Mice were pretreated with ABT to reduce hepatic treatment (Fig. 7). PK analysis of these doses showed that the free
TABLE 4 Activity of compound D against Gram-positive clinical isolatesa
Value(s)
Organism and parameter
Compound D
Linezolid
Erythromycin
Levofloxacin
Ceftriaxone
Vancomycin
S. pneumoniae (n ϭ 70)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
0.004
0.008
Յ0.001 to 0.016
1
1
0.125
Ͼ16
0.031 to Ͼ16
1
1
ND
ND
ND
ND
ND
ND
MIC range (␮g/ml)
0.5 to 2
0.5 to 16
S. pyogenes (n ϭ 69)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
0.125
0.5
0.031 to 1
2
2
ND
ND
ND
0.5
1
0.25 to 2
Յ0.063
Յ0.063
Յ0.063 to Յ0.063
ND
ND
ND
MIC range (␮g/ml)
1 to 32
S. aureus (n ϭ 105)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
0.031
0.062
0.016 to 0.125
2
4
ND
ND
ND
0.25
Ͼ8
0.0625 to Ͼ8
8
ND
ND
ND
Ͼ64
1 to Ͼ64
MIC range (␮g/ml)
1 to 4
MRSA (n ϭ 25)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
MIC range (␮g/ml)
0.062
0.125
0.016 to 0.025
2
4
ND
ND
ND
Ͼ8
Ͼ8
4 to Ͼ8
Ͼ64
Ͼ64
Ͼ64 to Ͼ64
ND
ND
ND
1 to 4
Enterococci (n ϭ 32)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
1
2
2
8
ND
ND
ND
16
Ͼ16
1 to Ͼ16
ND
ND
ND
2
Ͼ32
0.5 to Ͼ32
MIC range (␮g/ml)
0.125 to 4
1 to Ͼ16
E. faecalis (n ϭ 16)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
0.5
1
0.125 to 1
2
2
ND
ND
ND
1
ND
ND
ND
2
Ͼ32
0.5 to Ͼ32
Ͼ16
1 to Ͼ16
MIC range (␮g/ml)
2 to 2
E. faecium (n ϭ 16)
MIC50 (␮g/ml)
MIC90 (␮g/ml)
1
2
2
8
ND
ND
ND
Ͼ16
Ͼ16
1 to Ͼ16
ND
ND
ND
Ͼ32
Ͼ32
0.5 to Ͼ32
MIC range (␮g/ml)
0.5 to 4
1 to Ͼ16
Other species
S. agalactiae MIC (␮g/ml)
S. milleri MIC (␮g/ml)
S. mitis MIC (␮g/ml)
S. mutans MIC (␮g/ml)
S. sallvarius MIC (␮g/ml)
S. sanguis MIC (␮g/ml)
S. epidermis MIC (␮g/ml)
S. saprophyticus MIC (␮g/ml)
4
2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2
1
2
1
2
2
0.5
0.5
ND
ND
ND
ND
ND
ND
ND
ND
16
16
8
8
8
8
8
32
Յ0.016
0.25
0.25
Յ0.016
Յ0.016
0.25
a MRSA, methicillin-resistant S. aureus; ND, not determined.
December 2013 Volume 57 Number 12
aac.asm.org 6011

de Jonge et al.
TABLE 6 Characterization of compound D-resistant mutants of S.
aureus and S. pneumoniae
Selection
condition
Compound D
MIC (␮g/ml)
PPAT
mutation
Organism
S. pneumoniae wild type
S. pneumoniae mutant 1
S. pneumoniae mutant 2
S. pneumoniae mutant 3
NAa
2ϫ MIC
2ϫ MIC
2ϫ MIC
0.031
1
0.125
0.5
NA
N107H
F8L
T117N
S. aureus wild type
S. aureus mutant 1
S. aureus mutant 2
S. aureus mutant 3
NA
0.06
NA
2ϫ MIC
2ϫ MIC
2ϫ MIC
4
4
1
N106H
N106Y
V136F
FIG 7 Efficacy in mouse S. pneumoniae lung infection model (left) for com-
pounds C (dark gray bars) and D (light gray bars) and in mouse S. aureus thigh
model (right) for compound C. Mice were treated at 100 mg/kg of body weight
b.i.d. q4h (dashed outlines) or q.i.d. q4h (solid outlines). Data from pretreat-
ment (white bars), vehicle growth controls for compounds C and D (no out-
line), and positive controls (black bars; see Materials and Methods) are in-
cluded for comparison. The horizontal black dashed line indicates the level of
stasis (left) or level of vehicle growth control (right). *, P Ͻ 0.05 compared to
vehicle growth control (Mann-Whitney test).
a NA, not applicable.
plasma concentrations of the compounds were above the MICs
for at least half the duration of the experiment (see Fig. S3 in the
supplemental material). A weaker response was observed in the
mouse thigh S. aureus infection model. In that model, the same
dose of compound C did not provide stasis compared to the start
of treatment but a reduction of the bacterial burden of only about
1 log compared to the untreated control results (Fig. 7). Dosing at
higher concentrations could not be attempted due to lack of sol-
ubility (compound D) or adverse effects (compound C).
Drug-like properties of compounds. Whereas activity of the
molecules could be optimized, despite numerous efforts, other
important compound properties, such as low protein binding,
high solubility, sufficient in vivo stability, and low toxicity, could
not be reconciled favorably with good biological activity. For ex-
ample, compounds C and D both showed potent MIC values, but
while compound C was highly soluble, it was rapidly cleared, and
while compound D was improved in clearance, its solubility was
low and signs of toxicity became apparent through weak activity
against a mammalian cell line (Table 2). The potency against the
mammalian cell line may be masked by the relatively high protein
binding of the analogues (90% to 95%, depending on the species;
Table 2), since mammalian cytotoxicity is measured in the pres-
ence of protein. Protein binding appears to affect the potency of
the compounds, as the drug MIC for S. aureus measured in the
presence of 50% human serum increased 16-fold.
thesis. Whereas inhibitors of this enzyme were identified previ-
ously (10, 11), none of these showed activity at the cellular level.
The series of PPAT inhibitors described in this report represent
the first reported set of compounds that not only inhibit the en-
zyme but, as a result, also suppress bacterial growth in vitro and
affect bacterial burden in animal models, validating PPAT as a
legitimate target for therapy against Gram-positive bacteria. These
inhibitors were discovered through high-throughput screening
with the isolated enzyme followed by optimization through me-
dicinal chemistry efforts guided by X-ray crystallography. Al-
though this approach may not always yield success (7, 8), this
study provides another example of the ability of target-based
screening to identify novel antibacterial agents. Unfortunately,
while the activities led to a series with desirable activity, further
progress into clinical development was hampered by the inability
to reconcile this with favorable pharmacokinetics, high solubility,
and low protein binding and/or low toxicity, as well as the poten-
tial for resistance development that is often encountered with
agents directed against single gene targets (7).
Biochemical characterization of PPAT inhibitors. Kinetic
and isothermal titration calorimetry studies with S. pneumoniae
PPAT were in agreement, showing compound B of the cycloalkyl
DISCUSSION
This report describes efforts to identify novel antibacterials acting pyrimidine series to be competitive with 4=-phosphopantetheine and
through inhibition of PPAT, an essential enzyme in CoA biosyn- cobinding with ATP in an uncompetitive manner, in addition to a
FIG 6 Killing kinetics for compound D with S. pneumoniae (left) and S. aureus (right). Cells were exposed to 0ϫ (open circles), 1ϫ (filled triangles), 2ϫ (filled
squares), 4ϫ (filled circles), and 8ϫ (filled diamonds) the MIC, and killing of cells was measured by plotting CFU/ml over time. Compound D was added at 0 h.
6012 aac.asm.org
Antimicrobial Agents and Chemotherapy

Inhibition of Phosphopantetheine Adenylyltransferase
second weaker mode of inhibition with the compound binding to pound penetration. Compound penetration into Gram-negatives
is a general issue for many compounds (42).
free enzyme. Analysis of kinetic data for the S. aureus enzyme showed
similarities to S. pneumoniae PPAT, with an inhibition modality for
compound B competitive with 4=-phosphopantetheine but with a
mixed-type inhibition profile with respect to ATP.
The compounds did not accelerate killing kinetics for S. aureus
or S. pneumoniae compared to the results for untreated cultures
that reached stationary phase. In addition, onset of growth sup-
pression was delayed, most likely due to the time required for
depletion of an existing CoA pool (14). These findings are similar
to those from another class of cofactor biosynthesis inhibitors, the
sulfonamides, which inhibit dihydropteroate synthetase, an en-
zyme involved in biosynthesis of the cofactor folate (43). It can be
anticipated that inhibitors of biosynthesis of other cofactors, such
as riboflavin and biotin, may show similar microbiological prop-
erties. Although a delay in the onset of growth inhibition and slow
killing kinetics might be expected to negatively impact the efficacy
of treatment, compounds with such properties have been success-
fully used in the clinic, as evidenced by the sulfonamides. Further-
more, the consensus is that there is no evidence supporting use of
cidal over static drugs for bacterial diseases aside from endocardi-
tis and meningitis (44).
Resistant mutants could be isolated with high frequency but
only at concentrations of compound of less than 4-fold the MIC
(Table 5). The drug MIC values for the mutants were significantly
elevated for PPAT inhibitors due to target modifications (Table
6), and coverage of such mutants would require significant higher
compound exposures. Moreover, at least for S. aureus, inhibition
of PPAT is not a rapidly cidal event and allows selection of resis-
tant mutants when compound exposure drops below therapeutic
levels. However, the reduced isolation of spontaneous resistant
mutants at elevated compound concentrations, and the fact that
no resistant mutants emerged during killing kinetics experiments,
even upon prolonged incubation, may indicate a low mutant pre-
vention concentration (MPC) (45) for these compounds. Thus,
the disadvantage of PPAT as a target with poor cidality that allows
for selection of mutants with significantly elevated drug MIC val-
ues may be offset by a low MPC. Further experimentation, using
multiple strains, would be required to test this hypothesis.
Efficacy in animal models. The lead compounds in the series
showed an effect on bacterial burden in two animal models, with
stasis achieved in the S. pneumoniae lung model (Fig. 7). Higher
doses and/or prolonged duration of dosing, to take into consider-
ation the observed delayed onset of growth inhibition and poor
killing kinetics seen in vitro, would likely lead to an improved
response. However, these could not be attempted with the current
compounds due to poor solubility and lack of tolerability. In the S.
aureus thigh model, even stasis was not achieved at the same doses,
but given the even slower killing kinetics observed in vitro with this
species compared to S. pneumoniae, this outcome could have been
expected. Alternatively, poor compound penetration into thigh,
and/or higher CoA pool levels in S. aureus, may have contributed
to the poorer response in this model.
The results from the kinetic and isothermal titration calorimet-
ric experiments were supported by analyses of cocrystal struc-
tures, which showed that the compounds bind to the 4=-phospho-
pantetheine pocket, with an accompanying ATP molecule
present. While the most thorough kinetic studies were done with
the S. pneumoniae enzyme and crystallography studies succeeded
with the S. aureus enzyme, the high degree of PPAT sequence
homology between these isozymes (see Fig. S4 in the supplemental
material) reduces concerns that general findings for one species
cannot be translated to others.
The cycloalkyl pyrimidines represent a new class of PPAT in-
hibitors, with activity against multiple bacterial PPAT isozymes.
The compounds did not inhibit human PPAT enzyme, thereby
showing desired antibacterial drug selectivity. Although members
of this scaffold were more potent inhibitors of Gram-positive
PPAT enzymes, they also had submicromolar IC50s against PPAT
of Gram-negative species (Table 2). Earlier studies reported on
other 4=-phosphopantetheine-competitive PPAT inhibitors. Di-
peptide-based inhibitors were designed to act as 4=-phospho-
pantetheine mimics, and crystallographic studies confirmed them
to be bound in the 4=-phosphopantetheine site (11). This series
was optimized for inhibition of E. coli PPAT, and the most potent
compound in that series (compound 8) had a 6 nM IC50 but all
lacked activity against bacteria. Pyrazoloquinolones (10) were
identified by screening with E. coli PPAT and were characterized as
ATP-competitive inhibitors. Although balanced activity against E.
coli and E. faecalis PPAT was achieved, these compounds showed
relatively low potency only, with Ki values at the single-digit mi-
cromolar level.
Activity of PPAT inhibitors against bacteria. Cycloalkyl py-
rimidines represent the first reported series of compounds to
show suppression of bacterial growth through inhibition of PPAT.
The mode of action of the compound series was unequivocally
established through genetic means, by generation of resistant mu-
tants that were shown to have mutations in coaD, and by showing
reduced susceptibility for a PPAT-overexpressing construct (Ta-
ble 2). To achieve growth inhibition, very potent enzyme inhibi-
tion (IC50s Ͻ 20 nM) was required. The reason for the high bio-
chemical potency requirement is unclear, but it may have been
due to poor compound penetration into the cell and/or may have
been the result of a requirement to inhibit this enzyme to a high
extent in order to achieve growth suppression, similar to the
mechanism proposed for inhibitors of DNA ligase (41). It would
be highly beneficial for a bacterium to express excess biosynthetic
capacity for CoA, a cofactor that plays a central role in many
cellular processes, to absorb minor disturbances in biosynthesis
without disrupting growth.
In conclusion. A series of cycloalkyl pyrimidines was identified
through high-throughput screening and optimized through a me-
dicinal chemistry program guided by crystallography. The lead
compounds of the series suppressed growth of Gram-positive bac-
teria through inhibition of PPAT and demonstrated effects on
bacterial burden in two mouse efficacy models, thereby validating
PPAT as a novel target for antibacterial therapy. Reconciling these
favorable biological properties with other properties needed for
drugs appeared not to be possible, preventing progression of this
The compounds showed a spectrum of activity limited to
Gram-positive species. As expected for compounds with a novel
mode of action, cycloalkyl pyrimidines were equally active against
strains susceptible to and strains resistant to clinically approved
drugs (Table 4). The lack of growth inhibition for Gram-negative
species was probably due to lower potency for inhibition of Gram-
negative PPAT enzymes (Table 2), combined with poor com- series into clinical development. Additional efforts will be re-
December 2013 Volume 57 Number 12
aac.asm.org 6013

de Jonge et al.
coenzyme A biosynthetic pathway via comparative genomics. J. Biol.
Chem. 277:21431–21439.
20. Genschel U. 2004. Coenzyme A biosynthesis: reconstruction of the path-
way in archaea and an evolutionary scenario based on comparative
genomics. Mol. Biol. Evol. 21:1242–1251.
21. Mandel AL, La Clair JJ, Burkart MD. 2004. Modular synthesis of
pantetheine and phosphopantetheine. Org. Lett. 6:4801–4803.
22. Mills SD, Eakin AE, Buurman ET, Newman JV, Gao N, Huynh H,
Johnson KD, Lahiri S, Shapiro AB, Walkup GK, Yang W, Stokes SS.
2011. Novel bacterial NADϩ-dependent DNA ligase inhibitors with
broad-spectrum activity and antibacterial efficacy in vivo. Antimicrob.
Agents Chemother. 55:1088–1096.
23. Clinical and Laboratory Standards Institute. 2009. Methods for dilution
antimicrobial susceptibility tests for bacteria that grow aerobically; ap-
proved standard, 9th ed. M07-A8, vol 29, no. 2. Clinical and Laboratory
Standards Institute, Wayne, PA.
quired to identify PPAT inhibitors that can progress into the
clinic, hopefully leading to other potent compounds with low re-
sistance potential that can be used in treatment of infections for
which current therapies have become limited as a result of resis-
tance development.
ACKNOWLEDGMENTS
We thank members of the microbiology susceptibility testing, pharmacol-
ogy, and DMPK groups for performing MIC determinations, executing
efficay studies, and physical chemical measurements on the compounds,
respectively, and Kathy MacCormack for DNA sequencing.
REFERENCES
24. National Committee for Clinical Laboratory Standards. 1999. Methods
for determining bactericidal activity of antimicrobial agents; approved
guideline. M26-A, vol 19, no. 18. National Committee for Clinical Labo-
ratory Standards, Wayne, PA.
25. Tomasz A, Moreillon P, Pozzi G. 1988. Insertional inactivation of the
major autolysin gene of Streptococcus pneumoniae. J. Bacteriol. 170:5931–
5934.
26. Yang J, Copeland RA, Lai Z. 2009. Defining balanced conditions for
inhibitor screening assays that target bisubstrate enzymes. J. Bimol.
Screen. 14:111–120.
27. Shapiro AB, Gao N, Thresher J, Walkup GK, Whiteaker J. 2011. A
high-throughput absorbance-based assay for methionine produced by
methionine aminopeptidase using S-adenosyl-L-methionine synthetase.
J. Biomol. Screen. 16:494–505.
28. Keating TA, Newman JV, Olivier NB, Otterson LG, Andrews B, Boriack-
Sjodin PA, Breen JN, Doig P, Dumas J, Gangl E, Green OM, Guler SY,
Hentemann MF, Joseph-McCarthy D, Kawatkar S, Kutschke A, Loch JT,
McKenzie AR, Pradeepan S, Prasad S, Martínez-Botella G. 2012. In vivo
validation of thymidylate kinase (TMK) with a rationally-designed, selective
antibacterial compound. ACS Chem. Biol. 7:1866–1872.
29. Jancarik J, Kim SH. 1991. Sparse matrix sampling: a screening method for
crystallization of proteins. J. Appl. Crystallogr. 24:409–411.
30. Smart OS, Womack TO, Flensburg C, Keller P, Paciorek W, Sharff A,
Vonrhein C, Bricogne G. 2012. Exploiting structure similarity in refine-
ment: automated NCS and target-structure restraints in BUSTER. Acta
Crystallogr. D Biol. Crystallogr. 68(Pt 4):368–380.
1. Fauci AS, Morens DM. 2012. The perpetual challenge of infectious dis-
eases. N. Engl. J. Med. 366:454–461.
2. Roberts RR, Hota B, Ahmad I, Scott RD, Foster SD, Abbasi F,
Schabowski S, Kampe LM, Ciavarella GG, Supino M, Naples J, Cordell
R, Levy SB, Weinstein RA. 2009. Hospital and societal costs of antimi-
crobial-resistant infections in a Chicago teaching hospital: implications
for antibiotic stewardship. Clin. Infect. Dis. 49:1175–1184.
3. Arias CA, Murray BE. 2009. Antibiotic-resistant bugs in the 21st century—a
clinical super-challenge. N. Engl. J. Med. 360:439–443.
4. Nordmann P, Poirel L, Toleman MA, Walsh TR. 2011. Does broad-
spectrum ␤-lactam resistance due to NDM-1 herald the end of the anti-
biotic era for treatment of infections caused by Gram-negative bacteria? J.
Antimicrob. Chemother. 66:689–692.
5. Bush K, Courvalin P, Dantas G, Davies J, Eisenstein B, Huovinen P,
Jacoby GA, Kishony R, Kreiswirth BN, Kutter E, Lerner SA, Levy S,
Lewis K, Lomovskaya O, Miller JH, Mobashery S, Piddock LJV, Projan
S, Thomas CM, Tomasz A, Tulkens PM, Walsh TR, Watson JD,
Witkowski J, Witte W, Wright G, Yeh P, Zgurskaya HI. 2011. Tackling
antibiotic resistance. Nat. Rev. Microbiol. 9:894–896.
6. Levy SB, Marshall B. 2004. Antibacterial resistance worldwide: causes,
challenges and responses. Nat. Med. 10:S122–S129.
7. Silver LL. 2011. Challenges of antibacterial discovery. Clin. Microbiol.
Rev. 24:71–109.
8. Payne DJ, Gwynn MN, Holmes DJ, Pompliano D. 2007. Drugs for bad
bugs: confronting the challenges of antibacterial discovery. Nat. Rev. Drug
Discov. 6:29–40.
31. Emsley P, Cowtan K. 2004. Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60(Pt 12 Pt 1):2126–2132.
32. Laskowski RA, MacArthur MW, Moss DS, Thornton JM. 1993.
PROCHECK—a program to check the stereochemical quality of pro-
tein structures. J. App. Cryst. 26:283–291.
33. Balani SK, Li P, Nguyen J, Cardoza K, Zeng H, Mu D-X, Wu J-T, Gan
L-S, Lee FW. 2004. Effective dosing regimen of 1-aminobenzotriazole for
inhibition of antipyrine clearance in guinea pigs and mice using serial
sampling. Drug Metab. Dispos. 32:1092–1095.
34. Azoulay-Dupuis E, Bedos JP, Vallée E, Hardy DJ, Swanson RN, Po-
cidalo JJ. 1991. Antipneumococcal activity of ciprofloxacin, ofloxacin,
and temafloxacin in an experimental mouse pneumonia model at various
stages of the disease. J. Infect. Dis. 163:319–324.
35. Sherer BA, Hull K, Green O, Basarab G, Hauck S, Hill P, Loch JT,
Mullen G, Bist S, Bryant J, Boriack-Sjodin A, Read J, DeGrace N,
Uria-Nickelsen M, Illingworth RN, Eakin AE. 2011. Pyrrolamide DNA
gyrase inhibitors: optimization of antibacterial activity and efficacy.
Bioorg. Med. Chem. Lett. 21:7416–7420.
9. Barrett JF. 2006. Impact of genomics—emerging targets for antibacterial
therapy. Comp. Med. Chem. II 7:731–748.
10. Miller JR, Thanabal V, Melnick MM, Lall M, Donovan C, Sarver RW,
Lee D-Y, Ohren J, Emerson D. 2010. The use of biochemical and bio-
physical tools for triage of high-throughput screening hits—a case study
with Escherichia coli phosphopantetheine adenylyltransferase. Chem. Biol.
Drug Des. 75:444–454.
11. Zhao L, Allanson NM, Thomson SP, Maclean JKF, Barker JJ, Primrose
WU, Tyler PD, Lewendon A. 2003. Inhibitors of phosphopantetheine
adenylyltransferase. Eur. J. Med. Chem. 38:345–349.
12. Zhang YM, White SW, Rock CO. 2006. Inhibiting bacterial fatty acid
synthesis. J. Biol. Chem. 281:17541–17544.
13. Spry C, Kirk K, Saliba KJ. 2008. Coenzyme A biosynthesis: an antimi-
crobial drug target. FEMS Microbiol. Rev. 32:56–106.
14. Leonardi R, Zhang Y-M, Rock CO, Jackowski S. 2005. Coenzyme A:
back in action. Progr. Lipid Res. 44:125–153.
15. Izard T, Geerlof A. 1999. The crystal structure of a novel bacterial adeny-
lyltransferase reveals half of sites reactivity. EMBO J. 18:2021–2030.
16. Geerlof A, Lewendon A, Shaw WV. 1999. Purification and characteriza-
tion of phosphopantetheine adenylyltransferase from Escherichia coli. J.
Biol. Chem. 274:27105–27111.
36. Yoon HJ, Kang JY, Mikami B, Lee HH, Suh SW. 2011. Crystal structure
of phosphopantetheine adenylyltransferase from Enterococcus faecalis in
the ligand-unbound state and in complex with ATP and pantetheine. Mol.
Cells 32:431–435.
17. Freiberg C, Wieland B, Spaltmann F, Ehlert K, Brötz H, Labischinski H.
2001. Identification of novel essential Escherichia coli genes conserved 37. Leskovac V. 2003. Comprehensive enzyme kinetics. Kluwer Academic
among pathogenic bacteria. J. Mol. Microbiol. Biotechnol. 3:483–489.
Press, Boston, MA.
18. Gerdes SY, Scholle MD, D’Souza M, Bernal A, Baev MV, Farrell M, 38. Izard T. 2002. The crystal structures of phosphopantetheine adenylyl-
Kurnasov OV, Daugherty MD, Mseeh F, Polanuyer BM, Campbell JW,
Anantha S, Shatalin KY, Chowdhury SAK, Fonstein MY, Osterman AL.
transferase with bound substrates reveal the enzyme’s catalytic mecha-
nism. J. Mol. Biol. 315:487–495.
2002. From genetic footprinting to antimicrobial drug targets: examples in 39. Miller JR, Ohren J, Sarver RW, Mueller WT, de Dreu P, Case H,
cofactor biosynthetic pathways. J. Bacteriol. 184:4555–4572.
19. Daugherty M, Polanuyer B, Farrell M, Scholle M, Lykidis A, de Crecy-
Lagard V, Osterman A. 2002. Complete reconstitution of the human
Thanabal V. 2007. Phosphopantetheine adenylyltransferase from Esche-
richia coli: investigation of the kinetic mechanism and role in regulation of
coenzyme A biosynthesis. J. Bacteriol. 189:8196–8205.
6014 aac.asm.org
Antimicrobial Agents and Chemotherapy

Inhibition of Phosphopantetheine Adenylyltransferase
40. Segel IH. 1975. Enzyme kinetics. Wiley-Interscience, New York, NY.
41. Korycka-Machala M, Rychta E, Brzostek A, Sayer HR, Rumijowska-
Galewicz A, Bowater RP, Dziadek J. 2007. Evaluation of NADϩ-
dependent DNA ligase of mycobacteria as a potential target for antibiotics.
Antimicrob. Agents Chemother. 51:2888–2897.
42. Kumar A, Schweizer HP. 2005. Bacterial resistance to antibiotics: active
efflux and reduced uptake. Adv. Drug Deliv. Rev. 57:1486–1513.
43. Seydel JK, Wempe E, Miller GH, Miller L. 1973. Quantification of the
antibacterial action of trimethoprim alone and in combination with
sulfonamides by bacterial growth kinetics. J. Infect. Dis. 128(Suppl):
463–469.
44. Finberg RW, Moellering RC, Tally FP, Craig WA, Pankey GA, Dellinger
EP, West MA, Joshi M, Linden PK, Rolston KV, Rotschafer JC, Rybak
MJ. 2004. The importance of bactericidal drugs: future directions in in-
fectious disease. Clin. Infect. Dis. 39:1314–1320.
45. Blondeau JM. 2009. New concepts in antimicrobial susceptibility testing:
the mutant prevention concentration and mutant selection window ap-
proach. Vet. Dermatol. 20:383–396.
December 2013 Volume 57 Number 12
aac.asm.org 6015