Structure-based design and development of functionalized mercaptoguanine derivatives as inhibitors of the folate biosynthesis pathway enzyme 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase from staphylococcus aureus

Article abstract of DOI:10.1021/jm501417f

6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase (HPPK), an enzyme from the folate biosynthesis pathway, catalyzes the pyrophosphoryl transfer from ATP to 6-hydroxymethyl-7,8-dihydropterin and is a yet-to-be-drugged antimicrobial target. Building on ou

Full text of DOI:10.1021/jm501417f

Article
pubs.acs.org/jmc
Structure-Based Design and Development of Functionalized
Mercaptoguanine Derivatives as Inhibitors of the Folate Biosynthesis
Pathway Enzyme 6‑Hydroxymethyl-7,8-dihydropterin
Pyrophosphokinase from Staphylococcus aureus
Matthew L. Dennis,†,‡ Sandeep Chhabra,†,‡ Zhong-Chang Wang,†,§ Aaron Debono,† Olan Dolezal,‡
Janet Newman,‡ Noel P. Pitcher,† Raphael Rahmani,† Ben Cleary,† Nicholas Barlow,† Meghan Hattarki,‡
Bim Graham,† Thomas S. Peat,‡ Jonathan B. Baell,† and James D. Swarbrick*,†
†Monash Institute of Pharmaceutical Sciences, Monash University, Parkville, Victoria 3052, Australia
‡CSIRO Biosciences Program, Parkville, Victoria 3052, Australia
§State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210093, People’s Republic of China
S
* Supporting Information
ABSTRACT: 6-Hydroxymethyl-7,8-dihydropterin pyrophos-
phokinase (HPPK), an enzyme from the folate biosynthesis
pathway, catalyzes the pyrophosphoryl transfer from ATP to 6-
hydroxymethyl-7,8-dihydropterin and is a yet-to-be-drugged
antimicrobial target. Building on our previous discovery that 8-
mercaptoguanine (8MG) is an inhibitor of Staphylococcus
aureus HPPK (SaHPPK), we have identified and characterized
the binding of an S8-functionalized derivative (3). X-ray
structures of both the SaHPPK/3/cofactor analogue ternary
and the SaHPPK/cofactor analogue binary complexes have
provided insight into cofactor recognition and key residues
that move over 30 Å upon binding of 3, whereas NMR
measurements reveal a partially plastic ternary complex active
site. Synthesis and binding analysis of a set of analogues of 3 have identified an advanced new lead compound (11) displaying
>20-fold higher affinity for SaHPPK than 8MG. A number of these exhibited low micromolar affinity for dihydropteroate
synthase (DHPS), the adjacent, downstream enzyme to HPPK, and may thus represent promising new leads to bienzyme
inhibitors.
design has therefore sought new generations of lead compounds
effective on DHFR mutant strains11 as well as the possibility of
more tailored sulfa drugs,12 novel pterin-site binding motifs,9,13
and allosteric inhibitors of DHPS.14 Recent insights into the
structural basis for the off-target side effects of the sulfa drugs15
are likely to assist in improving the selectivity and efficacy of this
class of drug.
Notwithstanding these efforts, the magnitude of the problem
of antibiotic resistance is highlighted by the emergence of the
methicillin-resistant strains of S. aureus (MRSA) (“the super
bug”). Originally confined within the hospital setting, more
recently it has spread to the community and begun affecting
those without any risk factors.16,17 The search for effective future
treatments has resulted in increasing interest into alternative
enzyme targets for new antibacterials18 as well as novel
immunization strategies.19 In this regard, our own research20,21
has focused on the structure-based development of inhibitors of
INTRODUCTION
■
Folate is an essential vitamin for the growth of all living
organisms. The reduced form, tetrahydrofolate, is a critical
cofactor for one-carbon transfer reactions required for the
synthesis of purines, amino acids, S-adenosylmethionine,
thymidine monophosphate, and formyl-methionine.1,2 Mam-
mals and higher eukaryotes depend on dietary folate, whereas
plants and most microorganisms synthesize it de novo. The sulfa
drugs, targeting the folate pathway enzyme dihydropteroate
synthase (DHPS), are still used in the clinic after several decades
and vindicate the pathway as an attractive source of antimicrobial
targets.3 Even today, sulfonamides are coadministered with a
bacterial dihydrofolatereductase (DHFR) inhibitor as a syner-
gistic broad-spectrum cocktail to prevent or treat a range of
diseases and infections, including malaria, Toxoplasma gondii
encephalitis, Pneumocystis carinii pneumonia, Tuberculosis, and
Staphylococcus aureus infections.4−8
Point mutations in both DHPS9 and DHFR10 genes have led
to resistance to sulfamethoxazole (SMX)- and trimethoprim
(TMP)-based therapies, respectively. Structure-based rational
Received: August 21, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. (A) HPPK catalysis. (B) Structures of selected HPPK inhibitors. (C) Superposition of EcHPPK structures (gray) with the SaHPPK/8MG
structure (PDB: 3QBC, yellow) showing active site loop conformations in response to a variety of bound ligands along with the bound substrate
(HMDP) and cofactor analogue (AMPCPP). Observed loop 3 conformational changes for EcHPPK are highlighted in green.
6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase
(HPPK), the enzyme directly preceding DHPS in the folate
pathway, responsible for catalyzing pyrophosphoryl transfer
from a magnesium-bound ATP cofactor to 6-hydroxymethyl-7,8-
dihydro-pterin (HMDP) (Figure 1A). HPPK is not a known
target for any existing antibiotic, has no relevant human
homologues, and has druggable sites to develop antimicrobials
effective on current and future SMX/TMP-resistant pathogens.
HPPK is an 18 kDa enzyme composed of a three-layered
α−β−α, thioredoxin-like fold (Figure 1).22 Catalysis follows an
ordered mechanism, whereby the ATP cofactor binds initially to
a comparatively open pocket in which the triphosphate is
complexed by two magnesium ions that are additionally
coordinated by two highly conserved aspartic acid residues.23
Binding of the HMDP substrate then follows, stacking between
two conserved aromatic rings. Three loops (loops 1−3, Figure 1)
interact and seal the active site to allow pyrophosphate transfer
from the cofactor to the substrate. A large body of structural
data,22 predominantly for the Escherichia coli enzyme (EcHPPK)
with the nonhydrolyzable cofactor analogue, AMPCPP, reveals a
relatively rigid core structure and flexible loops, particularly the
cofactor loop 3, which undergoes a mechanistically important 20
Å positional change during the catalytic cycle.23,24 X-ray
structures of HPPK from several species21,25−28 reveal highly
conserved residues in the active site, indicating that suitably
designed inhibitors could display broad-spectrum activity.
Only a few inhibitors of HPPK have been reported to date.
Early work, predating the structural characterization of HPPK,
led to the identification of the gem dimethyl- (1) and phenethyl-
substituted (2) substrate analogues as inhibitors (Figure 1).29
Rationally designed bisubstrate analogues based on these have
also been reported.30−32 Recently, adopting a rapid overlay of
chemical similarity (ROCS) scaffold hopping screening method,
we measured the binding of a series of commercially available
substrate-like compounds and identified 8-mercaptoguanine
(8MG) (Figure 1B) as a novel inhibitor of the S. aureus enzyme
(SaHPPK) (KD = 11 μM, IC50 = 41 μM).21 Unlike other
inhibitors of HPPK, binding of this compound was shown to be
independent of cofactor and Mg2+ ions. 8MG benefits from a
high level of steric and electronic complementarity to the
substrate site and being small, resulting in a high ligand efficiency
rating (ΔG/number of heavy atoms = 0.63 kcal mol−1 heavy
atom−1). On the basis of this discovery, we initiated an ongoing
SAR program to structurally elaborate 8MG into a more potent
inhibitor.20 Because of the demonstrated importance of the
sulfur substituent, we initially focused on chemical extension
from the N9 and N7 positions, revealing that only the latter
strategy was viable. An N7-ethyl alcohol variant exhibited
comparable affinity to that of the parent compound, but it was
competitive with the cofactor.
In view of the above findings, we decided to revisit extension
from the sulfur atom of 8MG. This article reports the results of
this latest work, which has culminated in the identification of
several novel 8MG analogues displaying significantly higher
affinities (KD ∼ 0.45 μM) for SaHPPK compared to that of the
parent compound in the presence of saturating levels of cofactor.
The SaHPPK binding ability and inhibitory activity of an initial
lead compound, as well as a series of synthesized analogues, have
been quantified by surface plasmon resonance (SPR) experi-
ments and an in vitro luminescent kinase coupled-enzyme assay,
respectively, and important structural and dynamic features of
the lead/cofactor/SaHPPK ternary complex have been revealed
by heteronuclear NMR spectroscopy. By solving the X-ray
structure of this complex, the intermolecular interactions
between the lead inhibitor and the enzyme have also been
delineated. Additionally, the X-ray structure of the wild-type
SaHPPK with a cofactor analogue bound has been determined,
revealing key residues required for cofactor recognition and
those that have moved by over 30 Å upon binding of the
inhibitor. Finally, SPR data is presented that show that a number
of the new 8MG analogues are also able to bind to DHPS with
low micromolar affinity, indicating the potential of this class of
compound to be developed into dual-action enzyme inhibitors.33
RESULTS AND DISCUSSION
■
A simple similarity search based on the TimTec (www.timtec.
net) catalogue was performed, and the four commercially
available 8MG analogues (out of seven hits) were purchased: 8-
N-morpholinoguanine, 7-methylguanine, 8-bromoguanine, and
8-((2-(4-methoxyphenyl)-2-oxoethyl)thio)guanine (3). The
SaHPPK binding properties of the first three of these compounds
were reported in our earlier work.20 Compound 3 is reminiscent
B
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
of the known 7-methyl-7-phenethylpterin analogue (2) of Wood
et al.29 (Figure 1B), which was crystallized, along with the 7-
gemdimethyl variant (1), in the first EcHPPK structure by
Stammers et al.34
Binding of 3 to SaHPPK by SPR. Binding of 3 to SaHPPK
was initially quantitatively analyzed by SPR, with SaHPPK
immobilized on a nitrilotriacetic acid (NTA) sensor chip surface.
Binding sensorgrams (Figure 2) were of good quality and
AMPCPP bound SaHPPK were distributed mainly over two
regions that, when mapped to the structure (Figure 3D), are
consistent with pterin-site binding. Amides 8−12 on the sheet
lining the pterin pocket showed small absolute CSPs around the
one-standard deviation (σ) value (Figure 3B). Amides from
residues 44−53 in loop 2 that are more solvent-exposed (see
green or surface representation) and the more buried ring
stacking amide from Phe123 showed even larger CSPs, over 3σ in
magnitude. Other large CSPs were also observed for Gly90,
Cys80, His82, and the side-chain Hε2 of His115 in the loop 3
hinge region, which probably reflect changes in loop structure
between the AMPCPP/SaHPPK binary and 3/AMPCPP/
SaHPPK ternary complexes. Although the structure of the
8MG/AMPCPP/SaHPPK complex has yet to be determined, a
similarly large CSP is observed for Gly90 upon the addition of
8MG (Figure 3A), suggesting that the position and environment
of the tip of loop 3 in this structure is similar to that in the 3/
AMPCPP/SaHPPK complex, at least around the Gly90 position.
In order to specifically probe the binding mode of the S8-
substituted pendant in 3, together with any associated conforma-
tional changes in the SaHPPK structure, the CSPs for the 3/
AMPCPP/SaHPPK complex were compared to those derived
from the 8MG/AMPCPP/SaHPPK spectra (Figure 3C) and
depicted in a model constructed from the AMPCPP/SaHPPK X-
ray structure (vide infra). As part of this initial analysis, it was
assumed that the guanine moieties of 8MG and 3 superpose in
the substrate pocket, and the loop 2 conformation was derived
from that observed in the 8MG/SaHPPK binary complex.21
Differential CSP data is quite powerful for determining the pose
of bound ligands by NMR,36 and this analysis showed that the
pendant of 3 interacts closely with residues in loop 2, specifically
around Tyr48, extending up and out from the base of the
substrate-binding site (Figure 3D). The differential CSP
observed for Arg121 may be suggestive of a direct interaction
with the pendant. The differential CSPs that mapped to those
residues under the pterin pocket (amides 8−10), remote from
the pendant, may support a small change in the binding
orientation of the guanine moiety compared to that in the 8MG/
AMPCPP/SaHPPK case. Although Asp95 and His115 are not
solvent-exposed, the changes in the CSPs reflect a change in the
environment near the gamma phosphate, potentially due to loop
3 structural changes (vide infra).
Figure 2. SPR sensorgrams (top panels) and steady-state affinity fits
(bottom panels) for the binding of compound 3 to SaHPPK in the (A)
presence or (B) absence of 1 mM ATP.
consistent with 1:1 stoichiometric binding. The binding affinities
(KD values) were derived by globally fitting the steady-state data
sets to a single-site binding model. Thus, in the presence of ATP,
3 was estimated to bind SaHPPK with an affinity of 1.09 0.12
μM. On the other hand, in the absence of ATP, steady-state
fitting using 8MG as a reference (KD = 10.8 0.4 μM)21 revealed
far weaker binding (KD = 77 16 μM).
Binding of 3 to SaHPPK by NMR Spectroscopy. The
chemical shift of a nucleus is highly sensitive to changes in its
local environment and is thus a convenient site-specific probe for
analyzing ligand-binding events, including conformational and
dynamic changes during complex formation.35 Incremental
addition of 3 to a sample of apo SaHPPK revealed only moderate
strength binding, as evidenced by signal broadening (character-
istic of intermediate exchange) and very minor chemical shift
perturbations (CSPs) (data not shown). Broadened resonances
mapped approximately to the substrate-binding site (data not
shown). In stark contrast, titration of 3 into a solution of
SaHPPK with either ATP or the nonhydrolyzable ATP analogue,
AMPCPP, fully bound resulted in widespread chemical shift
changes (Figure 3A), and all CSPs exhibited slow exchange on
the NMR time scale. Moreover, in contrast to the case with
8MG,21 saturation was achieved at close to a 1:1 ligand-to-
enzyme ratio, consistent with the higher affinity of 3 for SaHPPK
measured by SPR (vide supra) in the presence of the nucleotide.
To investigate the binding of 3 in more detail, we assigned the
1HN, 15N, and 13CA backbone resonances for the 3/AMPCPP/
SaHPPK ternary complex using a triple-resonance 3D NMR
experiment. The absolute CSPs induced by binding of 3 to the
X-ray Structures of 3/AMPCPP/SaHPPK and AMPCPP/
SaHPPK. The X-ray structure of SaHPPK in complex with 3 and
AMPCPP (Figure 4) was solved at 2.0 Å resolution using
molecular replacement (Table 1). The ternary complex crystal-
lized in the P61 space group, with a single protein molecule in the
asymmetric unit. Backbone density was observed for all 158
amino acid residues of the protein. The guanine moiety of 3 is
positioned similarly to that of 8MG in the 8MG/SaHPPK binary
complex,21 making a total of six hydrogen bonds with the protein
and π-stacking between the aromatic rings of Phe54 and Phe123.
The aromatic ring of the pendant projects out and away from the
substrate pocket into the loop 2/loop 3 region, making favorable
hydrophobic interactions with Val46 and Gly47. The adjacent
ketone group stacks against the guanidinium group of Arg121
and also interacts with the phenyl ring of Phe123. These features
are also consistent with the large CSPs observed for these
residues upon formation of the ternary complex (vide supra).
The Mg2+ ions in the 3/AMPCPP/SaHPPK structure
superpose closely with those in the HMDP/AMPCPP/EcHPPK
structure (PDB: 1QON),22,23 differing in position by only 0.22
(Mg1) and 0.35 Å (Mg2). Both ions are coordinated to residues
C
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. NMR data of SaHPPK binding to 3. (A) Superposition of the 2D 15N HSQC spectra recorded in a sample of ∼120 μM SaHPPK in 10 mM
Mg2+ 50 mM HEPES, pH 8, in the presence of saturating amounts of AMPCPP (blue), 8MG (green), and compound 3 (red). (B) Raw CSP (black) for
the change in weighted averaged chemical shifts for 3/AMPCPP compared to the AMPCPP 2D 15N HSQC spectra. CSP plot weighted by solvent
accessibility (green). CSPs for the side chain of His115 and Trp89 are shown in gray. (C) Raw weighted average chemical shifts (black) derived from the
change in the 3/AMPCPP and the 8MG/AMPCPP 2D NMR spectra and weighted for solvent accessibility (green). In both panels B and C, pink, blue,
and red horizontal lines signify 1, 2, and 3 standard deviations of the CSPs. (D, E) CSPs greater than the values in panels B and C are mapped to the
structure of SaHPPK in panels D and E, respectively.
To help rationalize the conformational changes accompanying
binding of 3, as well as the ∼70-fold enhanced affinity of 3 for the
cofactor/SaHPPK complex relative to that for SaHPPK alone,
the X-ray structure of SaHPPK in complex with AMPCPP was
also determined (Figure 5). The complex crystallized in the space
group P21 (Table 1) and was solved via molecular replacement to
a resolution of 2.7 Å. In contrast to the 8MG/SaHPPK complex,
four rather than two protein molecules were found in the
asymmetric unit (Figure 5A).21 There are also differences
between the two crystal forms in terms of the nature of the
protein−protein interface. The existence of an intermolecular
disulfide bond between the solvent-exposed Cys80 residues of
neighboring proteins in the AMPCPP/SaHPPK crystal lattice is
particularly notable (Figure 5B). Despite the presence of
AMPCPP in the crystallizing solution used to grow crystals of
the 8MG/SaHPPK binary complex, AMPCPP did not bind in
this case, which was rationalized in terms of the binding site being
partly occluded by the interface of the protomers in the
Figure 4. X-ray structure of SaHPPK in complex with AMPCPP and 3.
(A) Detail of the active site. Loops 2 and 3 are shown in magenta, and
the two magnesium ions, in green. (B) mFo − DFc difference density
map of AMPCPP and 8MG contoured at 3.0σ.
asymmetric unit. The change in the nature of the protein−
protein interface associated with intermolecular disulfide bond
formation appears to be more compatible with AMPCPP
binding.
As anticipated, the AMPCPP occupies the cofactor site, with
phosphate oxygens hydrogen-bonded to Arg121, Arg117,
His115, and Arg92 (one of three arginines in loop 3). The
adenine base is hydrogen-bonded to the amide backbone of Ile98
D95 and D97 and share coordination to a β-phosphate oxygen.
One magnesium center is coordinated by the α-phosphate,
whereas the other is bound by the γ-phosphate. The fifth and
sixth coordination sites are occupied by water molecules in each
case.
D
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. X-ray Statistics
AMPCPP (4CYU)
AMPCPP/3 (4CRJ)
AMPCPP/7 (4CWB)
spacegroup
P21
P61
P61
wavelength (Å)
1.0080
0.9537
0.9537
unit-cell parameters (Å, deg)
a = 62.00, b = 94.26, c = 62.99, α = γ = 90.00, a = b = 82.48, c = 52.17, α = β = 90.00, a = b = 83.82, c = 52.04, α = β = 90.00,
β = 112.44
γ = 120.00
γ = 120.00
Diffraction Data
resolution range (Å)
no. of unique reflections
no. of observed reflections
Matthews coefficient,
47.13−2.70 (2.77−2.70)
18 506 (2455)
139 533
42.17−2.00 (2.05−2.00)
13 792 (1962)
296 051
42.33−1.56 (1.60−1.56)
29 793 (1419)
332 946
2.39
2.83
2.93
V
M (Å3 Da1−
)
solvent content (%)
completeness (%)
data redundancy
mean I/σ(I)
Rmerge
48.6
58.6
58.1
100.0 (100.0)
7.5 (7.6)
99.7 (98.0)
21.5 (19.7)
12.2 (3.6)
0.215 (1.02)
0.047 (0.232)
99.7 (96.9)
11.2 (10.3)
14.0 (2.7)
0.091 (0.673)
0.041 (0.315)
9.4 (2.1)
0.139 (0.899)
0.082 (0.528)
Rp.i.m.
Refinement
R
R
free (%)
cryst (%)
24.2
19.4
4.5
4
20.0
16.4
5.0
1
17.8
16.0
5.1
1
size of Rfree set (%)
protein molecules in the
asymmetric unit
inhibitor molecules
cofactor-analogue molecules
water molecules
1
1
4
1
1
13
83
141
RMSD from Ideal Values
bond lengths (Å)
bond angles (deg)
Mean B factors (Å2)
Ramachandran Plot
favored (%)
0.010
1.49
54.2
0.006
1.36
27.4
0.009
1.60
15.5
99.0
0.2
98.7
0.0
98.2
0.0
outliers (%)
Figure 5. X-ray structure of SaHPPK in complex with AMPCPP. (A) X-ray structure showing the spatial arrangement of the four protomers in the
asymmetric unit. (B) 2mFo − DFc standard density map (green mesh) contoured at 1.5σ showing a disulfide bond observed between protomers in
adjacent unit cells. (C) mFo − DFc difference density map (green mesh) of AMPCPP from protomers A (left) and B (right) contoured at 3.0σ. (D)
Superposition of the four protomers and detail of the interactions of the bound AMPCPP. In panels A and C, alternate conformations of AMPCPP are
omitted for clarity, and magnesium ions are colored green.
and Ser112 and forms hydrophobic contacts with the side chains
of Leu111, Ile98, and Leu71. The ribose O2′ is hydrogen-bonded
to Lys110. Two magnesium ions are again present, coordinated
by Asp95, Asp97, and either the α- and β-phosphates or the β-
and γ-phosphates of AMPCPP. Alternate conformations of
AMPCPP are present in protomers B and D. In these
conformations, the γ-phosphate is oriented orthogonal to the
γ-phosphate site (Figure 5D), forming additional interactions to
E
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 6. Conformational changes in SaHPPK. (A) Superposition of the AMPCPP (green) and the AMPCPP/3 (magenta) X-ray structures illustrating
the change in conformation of the loop 3 and selected arginine side chains. (B) Superposition of the three SaHPPK X-ray structures shown in panel C
highlighting the changes in the loop regions. (C) Comparison of the X-ray structure of SaHPPK in complex with AMPCPP/3, AMPCPP, and 8MG.
Selected side chains are shown to illustrate the positional and conformational changes.
Arg92 and Gln3 of an adjacent protomer, at the expense of loss of
hydrogen bonds to His115, Arg118, and Arg121 as well as
coordination to a magnesium ion.
phosphate. In the 3/AMPCPP/SaHPPK complex, the Arg92
headgroup is also rotated by 90° and is oriented approximately
parallel to that of Arg83; it hydrogen bonds to the β-phosphate as
before. Gly90 at the tip of loop 3 moves ca. 17 Å closer to the
active site upon the binding of 3 to the AMPCPP/SaHPPK
complex. The conformations of the side chains of the other
residues involved in nucleotide binding (Leu71, Leu75, Glu78,
Asp95, Asp97, Ile98, Lys110, Leu111, Ser112, Val113, His115,
and Arg117) do not change between the two structures.
The AMPCPP lies over 4 Å away from 3 in the ternary
complex; therefore, the significantly enhanced binding affinity of
3 toward SaHPPK in the presence of the cofactor must be due to
cofactor-induced intermolecular interactions rather than specific
interactions with the cofactor itself. In the 8MG/SaHPPK
structure,21 residue Arg121 exists in two conformations (shown
in orange in Figure 6C), one oriented toward and one
perpendicular to the cofactor, suggesting that this residue is
likely mobile in the apo enzyme. In both the AMPCPP/SaHPPK
and 3/AMPCPP/SaHPPK structures, the conformation appears
to be locked by the hydrogen bond to the γ-phosphate of the
AMPCPP, possibly assisted by a π interaction between the
guanidinium group of Arg121 and the ketone of 3, resulting in
the observed cofactor-mediated improvement in binding affinity.
From the 3/AMPCPP/SaHPPK structure, the large, yet
similar, chemical shift change observed in the NMR spectra for
Gly90 upon binding of 3 (Figure 3A) or 8MG to AMPCPP/
SaHPPK can be attributed to hydrogen bonding to the carbonyl
of Glu87 combined with a favorable interaction between the
sulfur atom of each ligand and the carbonyl group of Trp89,39
rather than a previously hypothesized hydrogen bond between
the SH of 8MG and the carbonyl of Trp89.21
Density was either weak or not observed for several of the loop
2 residues, indicative of mobility therein, concordant with prior
NMR relaxation studies on EcHPPK37 and SaHPPK21 in which
loop 2 was shown to be dynamic in the absence of bound
substrate or substrate inhibitor. The observed loop 3
conformations (Figure 5C) overlap well with those previously
found in the ensemble of NMR structures of AMPCPP/EcHPPK
(data not shown),24 and the existence of multiple conformations
in the structure is again in line with the broadened NMR
resonances observed for residues 84−89 and 92 (vide supra). In
the AMPCPP complex, the side chain of Arg92 is extended and
makes an end-on salt bridge to the α- and β-phosphates, which
caused loop 3 to adopt a more open conformation relative to that
observed in the crystal structure of apo EcHPPK.
It is noted that the side-chain orientation of Arg92 is not at all
defined in the NMR structure of EcHPPK, which has been used
in the past to represent the EcHPPK/AMPCPP structure in the
absence of a suitable EcHPPK X-ray structure,23 although this
may reflect, to some degree, the difficulty and limitation of
employing short-range NOEs to characterize hydrophilic
interactions. The important role of Arg92 in cofactor recognition
is underscored by the current AMPCPP/SaHPPK complex,
which resembles the predicted AMPCPP/EcHPPK model
derived from locally enhanced sampling and molecular dynamics
simulations.38
A comparison of the two structures presented here with the
previously reported 8MG/SaHPPK structure (PDB: 3QBC)
(Figures 6A−C)21 reveals marked differences in the conforma-
tion of the cofactor loop 3 and the substrate loop 2 and more
subtle change in loop 5 (Figure 6B). There is a notable change in
position of Arg92 upon binding of 3. In the 8MG/SaHPPK
structure, the guanidinium group of Arg92 is displaced by that of
Arg83 and is translated 3.5 Å toward the substrate pocket. Arg83
moves 13 Å to make a hydrogen-bond contact to the α-
Insight into the Dynamics of the 3/AMPCPP/SaHPPK
Complex by NMR. To investigate the dynamic properties of the
3/AMPCPP/SaHPPK complex on the fast (pico- to nano-
second) time scale, 15N heteronuclear NOEs were measured, and
those amides with NOE values chosen to be less than 0.75 were
mapped onto the surface of the structure (Figure 7) to highlight
F
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 7. 600 MHz 15N heteronuclear NOE values for SaHPPK in complex with AMPCPP/3 (red) and AMPCPP/8MG (black). The 15N NOE for the
side-chain Hε1 for Trp89 is shown as a large shaded circle. Residues with 15N NOE values less than 0.75 are mapped onto the surface and ribbon
representation of the SaHPPK/AMPCPP/3 X-ray structure. Amides not observed in the 15N HSQC spectra due to severe broadening in loop 3 are
colored magenta, and proline residues, cyan.
Figure 8. Comparison of the SaHPPK AMPCPP/3 structure with the EcHPPK/AMPCPP/2 structure. (A) The substrate pocket in the SaHPPK/
AMPCPP/3 complex (magenta) is more open than that in EcHPPK/AMPCPP/2 (blue) or in EcHPPK/AMPCPP/HMDP (green), which is
completely sealed. (B) Comparison of the active site of SaHPPK/AMPCPP/3 (magenta) with that of EcHPPK/AMPCPP/2 (blue). (C) Surface
representation of the EcHPPK/AMPCPP/HMDP (left), EcHPPK/AMPCPP/2 (middle), and SaHPPK/AMPCPP/3 complexes (right).
statistically relevant regions with increased mobility compared to
the global average (0.81). Residues 1 and 158 at the termini are
previously achieved for 8MG/AMPCPP/SaHPPK21 and yielded
very similar results for both ternary complexes, with the standard
deviation of the average value noticeably smaller than that
reported earlier (0.08 vs 0.12). The 15N NOE for the side chain of
Trp89 appears to be essentially the same for both the 3/
AMPCPP/SaHPPK and 8MG/AMPCPP/SaHPPK complexes,
with a degree of residual fast motion similar to that observed for
loop 2, indicating that the interaction of the pendant of 3 with
this residue is not strong enough to dampen Trp89’s fast time
scale motion entirely. From a structural perspective, this
observation may be more consistent with the position of the
Trp89 side chain observed in the X-ray structure compared to
that observed in the HMDP/AMPCPP/EcHPPK ternary
complex.40
highly mobile (15N NOE < 0.5). A characteristic dip in the 15
N
NOE values from the global average for amides 47−51 is
indicative of some residual fast-time scale motion centered
around the interface of loops 2 and 3. Other amides experiencing
limited motion include Arg92 in the hinge of loop 3 and Glu103,
located in a loop leading into the β hairpin, within which Asp107
and Leu111 are also partly mobile.
The data were compared with that for the 8MG/AMPCPP/
SaHPPK complex (shown in black) by recording the sample
under identical conditions and with the same pulse sequence. A
combination of a newer cryoprobe and a more optimal pulse
sequence afforded much better water suppression than
G
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 9. (A) Overview of the active sites of the SaHPPK/AMPCPP/3 (pink) with SaHPPK/AMPCPP/7 complexes and (B) the mFo − DFc difference
density map of the ligands in the SaHPPK/AMPCPP/7 complex contoured at 3.0σ. (C) Superposition of the bound poses for 3 (pink) and 7 (magenta)
with (D) the modeled bound pose of 11 (white). (E) 3 and 11 shown within a surface representation of the substrate pocket. (F) Raw CSPs
corresponding to the change in the weighted average chemical shifts observed for 11/AMPCPP/SaHPPK relative to 8MG/AMPCPP/SaHPPK in the
2D 15N HSQC NMR spectra, mapped onto the docked 11/AMPCPP/SaHPPK structure. The CSPs are colored pink, blue, and red for those greater
than the 1, 2, or 3 times the standard deviation of the CSP values shown in Figure 3C. The broadened Gly47 is colored yellow. (G) The bound pose of 3
in the EcHPPK/AMPCPP/3 (PDB: 4M5J, orange) is different from that in the SaHPPK/AMPCPP/3 structure (pink).
Even though 3 is larger than 8MG, binds more tightly to the
AMPCPP/SaHPPK complex (by over an order of magnitude),
and protrudes into the loop 2/3 region, amide signals for residues
84−89 in loop 3 were not observed in the NMR spectra of 3/
AMPCPP/SaHPPK (magenta in Figure 7), revealing large
amplitude motion on the slower (micro- to millisecond) time
scale. Thus, the addition of the pendant to the 8MG parent
scaffold appears to have limited large-scale impact on the overall
micro- to millisecond backbone dynamics of loop 3. However,
given that the pendant forms few intermolecular interactions
with side-chain atoms, this is not entirely unexpected.
Comparison of the Structures of 3/AMPCPP/SaHPPK,
2/AMPCPP/EcHPPK, and 3/AMPCPP/EcHPPK. Given the
broad similarity in the structures of inhibitors 2 and 3 and the
sequence similarity between EcHPPK and SaHPPK, it is not
surprising that the 3/AMPCPP/SaHPPK structure is generally
quite similar to that of the previously reported 2/AMPCPP/
EcHPPK ternary complex (PDB: 1DY3) (Figure 8A,B),34 with
the cofactor loop 3 closed in over the active site in both cases,
compared with the extended conformation observed in the
8MG/SaHPPK binary complex.21
In the case of the E. coli enzyme, a comparison of the 2/
AMPCPP/EcHPPK complex with the substrate/AMPCPP/
EcHPPK complex40 shows that the binding of 2 leads to an
opening of the substrate pocket lid, revealing a solvent-exposed
active site pocket, filled by the phenethyl pendant of 2. This
widening to accommodate 2 causes a rotation and 3.6 Å
positional change in the side-chain methyl groups of Leu45 and a
3.1 Å sized hinge movement of Trp89 (Figure 8A). Although the
equivalent SaHPPK structure in complex with HMDP and
AMPCPP has not been crystallized, a comparison with the 8MG
structure21 reveals that the rotation of the equivalent Val46
methyl groups is significantly smaller (<1 Å) in comparison
(Figure 8A).
Although the phenyl rings of 2 and 3 in the enzyme complexes
occupy similar positions, the position of the side chain of the
Trp89 residue is markedly different in the two structures (Figure
8B). In the 2/AMPCPP/EcHPPK structure, the Trp89 indole
ring orients toward the binding site, forming an edge-on π-
stacking interaction with the phenyl ring of 2 on one side and the
Arg88 residue on the other, and the indole ring Hε2 atom makes
a hydrogen-bond contact to the γ-phosphate of ATP. In the
SaHPPK structure, however, Trp89 orients away from the
binding site to form a hydrophobic contact with the loop 2
residue, Tyr48 (Pro in EcHPPK). The Arg88 guanidinium group
is translated 5 Å deeper into the cofactor site, where it appears to
hydrogen bond to the heterocyclic ribose oxygen of AMPCPP.
For EcHPPK, formation of the HMDP/AMPCPP ternary
complex involves a hydrogen-bond network that depends on the
interaction of residues Asn11 and Gln51 (equivalent to Asn10
and Gln50 in SaHPPK) to draw all three loop regions into the
fleeting transition state conformation.23 While we have not yet
solved the structure of the HMDP/AMPCPP/SaHPPK
complex, it is possible that a Trp89-Tyr48 interaction helps to
stabilize the interaction between loops 2 and 3 within this
complex. A surface representation of all three ternary complexes
(Figure 8C) illustrates that the positioning of the Trp89 and
Arg88 residues in the 3/AMPCPP/SaHPPK complex leads to a
larger, more solvent-accessible binding pocket than that in the 2/
AMPCPP/EcHPPK complex. In contrast, the substrate-binding
site is completely shielded from solvent in the HMDP/
AMPCPP/EcHPPK complex.40
Comparing the cofactor sites, in EcHPPK, three arginines
(Arg82, Arg84, and Arg92) bind the phosphates of AMPCPP in
2/AMPCPP/EcHPPK. For SaHPPK, Arg83 and Arg92 replace
the roles of Arg84 and Arg92 in EcHPPK; however, Arg85 points
in the opposite direction to the equivalent Arg84 in EcHPPK
(Figure 8B). This difference may be due to the single-residue
H
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 2. SAR for Compounds 3−13
a
b
c
With 1 mM ATP present. No ATP present. Chhabra et al.21
insertion in loop L3 in the E. coli enzyme (Ala86). Arg84 in
EcHPPK hydrogen bonds to the α-phosphate of AMPCPP,
resulting in a difference of 1.2 Å in the positioning of the O5′
atoms in the two ternary structures. The position of the O3′
atoms also differs by 2.3 Å, likely due to interaction with Gln74 in
EcHPPK and Leu75 in SaHPPK, which alters the ring pucker
(C2′ endo in SaHPPK and C3′ endo in EcHPPK). The change in
the O2′ positions in the two puckers and the hydrogen bond
from O2′ to the backbone of Lys110 in SaHPPK (Arg110 in
EcHPPK) may explain the observed 2.5 Å difference in the
positioning of the tip of the cofactor loop 5 (Figure 8B).
EcHPPK structures reveals some interesting differences
pertaining to the active site region. Most notably, the active
site is more solvent-exposed in the SaHPPK structure, and the
pendant of 3 adopts different poses, with the ketone oxygen and
para carbon of the pendant ring differing in position by 1.6 and
1.2 Å, respectively, between the two structures (Figure 9F).
Given that the NMR data for the 3/AMPCPP/SaHPPK complex
indicates that both residues are in close contact with the ligand
and that the loops surrounding the pendant of 3 are not
completely rigidified on either the fast or slow time scale, it would
be interesting to conduct a similar study on 3/AMPCPP/
EcHPPK.
During the later stages of this work, Yun et al.41 also reported
the X-ray structure of 3 in complex with AMPCPP/EcHPPK. A
comparison of the 3/AMPCPP/SaHPPK and 3/AMPCPP/
SAR of Compound 3 and Other S8-Substituted
Guanine Analogues. To probe the contribution of specific
I
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 10. SPR sensorgrams (top panels) and steady-state affinity fits (middle panels) for the binding of all compounds to SaHPPK in the (A) presence
or (B) absence of 1 mM ATP and to (C) EcDHPS. (D) Kinase-Glo assay results for compound 11. Error bars indicate SEM.
groups to the binding affinity of 3 toward SaHPPK, we embarked
on a SAR investigation and synthesized a series of S8-substituted
8MG analogues. The affinity and activities for the series of
compounds are shown in Table 2 and Figure 10.
the 2-position (6) led to complete abrogation of activity, which is
again suggestive of a more open active site around the aryl group
in SaHPPK. Furthermore, they found that large groups at the 4-
position were not tolerated, whereas we have shown that they can
be for SaHPPK. Indeed, we were able to crystallize and solve the
structure of the biphenyl analogue 7 in complex with AMPCPP
(Figure 9A,B). The structure is remarkably similar to that of the
3/AMPCPP/SaHPPK complex. The largest differences are
found in the orientation of the Arg88 side chain, which is split
between two conformations. The difference of most relevance to
the SAR analysis, however, is the change in the position of the
ketone group (2.2 Å), which maintains the stacking between the
Arg121 guanidinium group and the sulfur atom (Figure 8A,C).
The position of the sulfur in the biphenyl analogue is essentially
the same as that in the 8MG/SaHPPK binary structure.
In view of the observed changes in the sulfur and carbonyl
positions (Figure 9C) in the cofactor-bound SaHPPK complexes
of 3 and 7, it was reasoned that removal of a carbon from the
linker in 3 might serve to draw the aryl ring down into the
position observed in the 7/AMPCPP/SaHPPK structure. We
reasoned that π stacking would be possible via the phenyl ring
tethered to a shortened linker and that this could replace the
ketone π system. Synthetic efforts were thus directed toward
shortening the linker in 3 to a substituted benzyl group. Listed in
Table 2 are the results for the unsubstituted compound 10, which
effectively represents a direct replacement of the acetyl group of 9
with a phenyl ring. Gratifyingly, activity was maintained (KD =
1.8 μM). A limited set of analogues was then assembled, the most
We first investigated the likely contributions of the methylene,
aryl, carbonyl and methoxy groups of 3 (KD = 1.1 μM) toward its
affinity for the cofactor-bound enzyme. Removal of the 4-
methoxy (compound 4, KD = 1.4 μM) had little effect on binding,
an observation predicted from the solvent-exposed nature of this
group in the X-ray structure (Figure 3A). A branched methylene
was also tolerated (5), furnishing a KD = 1.8 μM, whereas moving
the 4-methoxy group to the 2-position (6) gave a slight loss of
affinity (KD = 2.8 μM). Furthermore, replacement of the 4-
methoxy group with a phenyl group to give biphenyl 7 did not
significantly alter affinity (KD = 0.81 μM). Saturation of the
ketone group to give the methylene analogue 8, on the other
hand, led to a significant decrease in affinity (KD = 7.8 μM),
which is again supported by the structural data, which revealed a
specific interaction between the ketone group and the side chain
of Arg121 in 3. Replacement of the aryl ring with a methyl group
(9) led to only a slight decrease in affinity (KD = 1.9 μM),
suggesting that the interactions of the aryl ring with the enzyme
are not strong, consistent with the relatively open or dynamic
nature of the active site around the aryl group.
In their recent study with EcHPPK, Yun et al.41 also found that
the 4-OMe group in 3 was not important for binding to EcHPPK.
However, in contrast to our findings for SaHPPK, they found
that a branched methylene linker and moving the 4-methoxy to
J
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
interesting of which are shown in Table 2. Here, it can be seen
that a 2-CF3 group results in a relative loss of affinity (12, KD =
2.8 μM). However, the 2-fluorobenzyl analogue 11 displayed an
increase in affinity toward SaHPPK, with a KD of 453 nM.
Another compound of interest was the 4-cyano analogue 13,
which also displayed strong affinity toward SaHPPK with a KD of
660 nM. Compared with 3, compounds 11 and 13 benefit from a
reduced molecular weight and number of rotatable bonds. For
these reasons, in addition to the comparative ease of synthesis of
these and other S-benzyl-substituted compounds, compounds
such as 11 and 13 represent promising leads for further
development. Interestingly, 11 also binds appreciably to the apo
enzyme and does so to a far greater extent than that of 3, with
respective KD values of 4.3 and 77 μM. It may therefore be a good
synthetic starting point for investigating cofactor-competitive
binders or moieties that bind at the metal site, perhaps by
combining SAR from our previously reported cofactor-
competitive N7 ethyl alcohol 8MG analogue (KD = 10 μM for
the apo enzyme).20 Compound 13 also corresponds to the only
S-benzyl-substituted analogue of 8MG tested by Yun et al.41 In
contrast to our findings for SaHPPK, this compound exhibits
very poor affinity for EcHPPK.
Crystallization attempts with 11 were unsuccessful; therefore,
docking of 11 into the SaHPPK crystal structure was undertaken
(Figure 9D, E). The resulting model suggests that the fluorine
atom could position itself relative to the guanidinium group in a
manner similar to that of the ketone group in 3. This could
plausibly confer increased affinity, as the guanidinium group is
known to be highly fluorophilic and can interact strongly with a
negatively polarized fluorine, favoring parallel orientations to the
guanidinium plane,42 as observed in the docked model. The
model indicates that the phenyl ring can overlay with the first ring
in the biphenyl analogue 7, maintaining interactions to Val46. A
reason for the lower binding affinity of 12 could be due to the
likely propensity for the 2-CF3 group to twist the phenyl ring out
of plane. The 15N HSQC spectrum of 11/AMPCPP/SaHPPK
was assigned using the HNCA experiment, as was done for the
spectrum of 3/AMPCPP/SaHPPK. Comparing the differential
CSPs of 8MG/AMPCPP-saturated SaHPPK with either 3/
AMPCPP/SaHPPK (Figure 3C) or 11/AMPCPP/SaHPPK
(Figure 9F and Supporting Information Figure S1) reveals very
similar pattern of CSPs but with notably larger differential CSPs
for Tyr48 and Glu50 amides in loop 2. Of note, the signal for
Gly47, the nearest amide to the pendant of 11 in the model, was
not observed in the spectra (shown in yellow in Figure 9F) but
was in the complex with 3, most likely due to subtle motional
effects of the proximal phenyl ring of the pendant on a micro- to
millisecond time scale. The larger differential CSPs (>3σ) of
Tyr48 and Glu50 (shown in red) may represent a conformational
change in loop 2, which also effects the stacking interaction with
Trp89, as evidenced by differential CSPs of >2σ (shown in blue).
Substantial differences were also noted in the arginine side-chain
region of the spectra (data not shown), which may be consistent
with the Arg121 guanidinium−fluorine interaction predicted by
the model. Although the observed NMR chemical shift data is
supportive of the docked pose of 11, assignment of the NMR
signals of the side-chain atoms will be a prerequisite to determine
the precise bound structure of 11.
polarizes the aryl ring, which would favor the π interaction with
the Arg121, presumably for both enzymes. It is unclear why this
would decrease the affinity compared to that of 3 for the
EcHPPK enzyme. The enhanced affinity for SaHPPK may have
its origins in the different pendant poses for 3 bound to the two
enzymes (Figure 9G) as well as differences in the orientation of
the methyl groups of Leu45 in EcHPPK and Val46 in SaHPPK
(Figure 8A). In order to assess whether enhanced affinity
translated into enhanced functional inhibition of HPPK, we
selected 11 and tested it alongside 8MG (IC50 = 41 μM)20,21 to
obtain the IC50 value. This gave rise to a value of 25 μM (Figure
10), suggesting that, while the higher affinity of 11 for HPPK
gives rise to increased functional inhibition, the increase is not as
great as might have been expected in view of 11’s much higher
affinity for HPPK (in the presence of ATP). However, it is noted
that in terms of drug-likeness, our lead compound 11 has a
number of favorable properties. For a compound with
submicromolar affinity, it has a relatively low molecular weight
of 291 Da, a topological polar surface area (tPSA) of 95 Å2,
suitable for membrane permeability and oral availability,43 and a
cLogP of 3.5. On the downside, it contains four hydrogen-bond
donors, and, accordingly, its membrane penetration ability was
found to be poor (A−B Papp = 0.6 0.2 × 10−6 cm s−1 for Caco-2
monolayers).
Binding of Compounds to DHPS. Given the chemical
similarity of the guanine and pterin scaffolds, the fact that the
pterin core is common to both HPPK and DHPS substrates, and
that 8MG is a known DHPS binder,9 it was decided to measure
the binding of our 8MG analogues to DHPS from E. coli (Figure
10). SPR data showed that compounds 3 and 4 bind DHPS with
a KD of ∼4.0 μM. Furthermore, the best HPPK binder,
compound 11, also binds DHPS with appreciable affinity (KD
∼ 8.0 μM). These KD values are notably better than that of 8MG
(76.1 μM). As for HPPK binding, the para OMe group in
compound 3 was found to have little impact on DHPS binding
affinity, and fluorine substitution at the ortho position of the
benzyl group appears to be beneficial.
CONCLUSIONS
■
Building on our previous structure-based approaches toward
inhibiting HPPK, we have used a combination of biophysical
methods, including SPR, NMR spectroscopy, and X-ray
crystallography, to reveal mechanistically important structural
changes accompanying binding of a series of 8MG-derived
substrate-site inhibitors to cofactor analogue-bound HPPK. In
combination with chemical synthesis, this has resulted in the
development of an advanced new lead compound (11)
displaying an affinity for SaHPPK over 20 times greater than
the previously reported parent compound, 8MG. Active site
structural details for the complexes presented here will assist in
the design and development of species-selective or broad-
spectrum inhibitors of HPPK. In this regard, it is notable that the
binding of 8MG and other analogues to EcHPPK is apparently
much weaker than that to SaHPPK.41 Sequence-related
structural differences, discussed here, may present avenues to
increased selectivity and potency. A number of the 8MG
analogues exhibit appreciable affinity for DHPS, highlighting the
potential for this class of compound to be developed into dual-
target inhibitors. A major focus of our future work will be on
developing analogues of 11 that are able to permeate bacterial
membranes, with the goal of achieving antibacterial activity.
The binding data for 13, 6, and 7 highlights a clearly divergent
SAR trend for SaHPPK and EcHPPK. The higher affinity of the
latter two compounds for SaHPPK can be reconciled on the basis
of the observed increased size/plasticity of the binding pocket for
SaHPPK. In the case of 13, the para-attached cyano group
K
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Crystallization and X-ray Structure Determination. Crystal-
lization experiments were performed as described previously.44 In brief,
co-crystallization was set up in the C3 screens (CSIRO) at 281 K using
sitting-drop vapor-diffusion method with droplets consisting of 150 nL
of protein solution and 150 nL of reservoir solution and using a reservoir
volume of 50 μL. Crystals of the SaHPPK in complex with AMPCPP
were obtained from a solution containing 120 mM magnesium acetate,
12.6% (w/v) PEG 8000, 120 mM Tris, pH 8.5, 1 mM AMPCPP, and the
protein at a concentration of 6.9 mg mL−1. Crystals of the 3/AMPCPP/
SaHPPK complex grew from a solution containing 275 mM ammonium
nitrate and 22.1% PEG 4000 with a protein concentration of 7.5 mg
mL−1. Crystals of the 7/AMPCPP/SaHPPK complex were grown under
similar conditions: 210 mM ammonium nitrate, 22.2% PEG 3350, and a
protein concentration of 6.9 mg mL−1. Data were collected at the MX-2
beamline of the Australian Synchrotron (see Table 1 for statistics) using
an ADSC Quantum 315 detector, with 270 frames obtained with a one-
degree oscillation angle for a complete data set. These data were indexed
using XDS48 and scaled using SCALA.49 The SaHPPK structure
(4AD6) was used to solve the initial phases of the binary and ternary
complexes by molecular replacement using Phaser.50 Refinement was
performed using REFMAC5,51 and the electron density maps were
visualized in Coot.52 After several rounds of manual rebuilding, ligands
and water molecules were added, and the models further refined to a
resolution of 2.7 Å (Rfree (%) = 24.2; Rwork (%) = 19.4) for the AMPCPP
complex; to a resolution of 2.0 Å (Rfree (%) = 20.1; Rwork (%) = 16.1) for
the 3/AMPCPP/SaHPPK ternary complex, and to a resolution of 1.6 Å
(Rfree (%) = 17.8; Rwork (%) = 16.0) for the 7/AMPCPP/SaHHPK
ternary complex.
EXPERIMENTAL SECTION
■
Samples of SaHPPK for NMR Spectroscopy. Isotopically labeled
protein samples for NMR spectroscopy were prepared as described
previously.21 E. coli BL21 (DE3) cells were transformed and grown
overnight in 3 mL of 2× YT medium supplemented with 100 μg mL−1
kanamycin for selection. The overnight culture was subcultured into 50
mL of minimal media that was grown to an OD600 of 0.5−0.7. This was
then added to 1 L of minimal media supplemented with 1.5 g of 15N
ammonium chloride and/or 3 g of 13C glucose and grown at 310 K until
the OD600 was 0.5−0.8. Isopropyl β-D-1-thiogalactopyranoside (IPTG)
was added to a final optimized concentration of 0.5 mM, and expression
was carried out at 293 K for 12 h. Purification was carried out as reported
previously.44
Preparation of DHPS from E. coli. A pET28a plasmid containing
the synthesized EcDHPS sequence (Geneart) was cloned with an N-
terminal hexahistidine tag and a thrombin cleavage site. E. coli BL21
(DE3) cells transformed with the plasmid were grown overnight in 20
mL 2× YT media supplemented with 50 μg mL−1 kanamycin for
selection. The overnight culture was then subcultured into fresh 2× YT
(0.5L) with growth at 310 K for ∼2 h until an OD600 of 0.5−0.7 was
reached. IPTG was added to a final concentration of 0.3 mM, with
expression occurring at 301 K for 22 h. The cultures were centrifuged at
5000 rpm for 10 min, and the cells were resuspended in 50 mL 50 mM
Tris, pH 8.5, 5% glycerol, 5 mM MgCl2. An EDTA-free complete
protease-inhibitor cocktail tablet (Roche) was added together with
lysozyme and DNase to a final concentration of 0.4 and 0.6 mg mL−1,
respectively. After 10 min, the cells were sonicated, and the cell debris
was removed by centrifugation at 18 000 rpm at 277 K for 30 min. The
supernatant was filtered (0.45 μm filter) and loaded onto a Ni-NTA
IMAC column (Qiagen). Unbound protein was washed off with 10 mM
imidazole in 50 mM Tris buffer, pH 8.5, 0.1 M NaCl, 5% glycerol, 2 mM
MgCl2, 1 mM DTT. Protein was eluted from the column with a 500 mM
imidazole, DTT-free variant of the above buffer. The protein was further
purified using a Superdex 75 size-exclusion 16/60 column (GE
Healthcare) and eluted with 50 mM Tris, pH 8.5, 2 mM MgCl2, 2
mM DTT, followed by the use of a MonoQ ion-exchange 16/10 column
(GE Healthcare). The column was equilibrated with 50 mM Tris, pH
8.5, 2 mM MgCl2, 2 mM DTT, with elution of protein using an
equivalent buffer with the addition of 0.25 M NaCl. Fractions were
analyzed using a 15% SDS-PAGE gel with Coomassie staining. Protein
was pooled and concentrated to 2 mg mL−1 using a 3 kDa molecular
weight cutoff ultrafiltration centrifugal device (Amicon). All samples
were snap-frozen and stored at 193 K.
NMR Spectroscopy. All NMR experiments were recorded at 295 K
on a Bruker Avance 600 MHz NMR spectrometer equipped with a
cryoprobe and Z-axis gradient. Triple-resonance NMR experiments
were performed on a sample of ∼0.25 mM 15N/13C-labeled SaHPPK
dissolved in a 90%/10% H2O/D2O HEPES buffer, 1% sorbitol, and 2%
DMSO-d6 at pH 8.0 in the presence of 10 mM MgCl2 and saturating
amounts of AMPCPP (0.5 mM). Titrations of compound 3 or 11 to
saturation was performed from a 50 mM stock dissolved in DMSO-d6.
Backbone assignments were obtained using the HNCA and HN(CO)-
CA experiments, and assignments were further confirmed using a 3D
15N-edited NOESY experiment recorded with a mixing time of 120 ms.
3D experiments used a WATERGATE sequence for solvent
suppression. 15N heteronuclear NOE spectra were recorded on a
∼0.36 mM 15N-labeled sample of SaHPPK in the presence of 1 mM
AMPCPP and either 600 μM 8MG or ∼400 μM 3 using gradients for
coherence selection and sensitivity enhancement. Three seconds of
saturation was applied using a binomial train of pulses separated by a
delay of 5 ms to generate the desired heteronuclear NOE and was
applied off- and on-resonance in an interleaved manner, in addition to 1
s of relaxation delay. Errors were calculated from the baseplane noise
level. Spectra were processed using NMRPipe45 and analyzed with
XEASY46 or SPARKY.47 2D 15N HSQC and 15N NOE experiments
were typically acquired with t1max (15N) = 51−62 ms and t2max (1H) =
142 ms, whereas triple-resonance experiments were acquired with t1max
(15N) = 23.3 ms, t2max (13C) = 10.4 ms, t2max (1H) = 15.1 ms, and t3max
(1H) = 142 ms.
The coordinates of SaHPPK in complex with AMPCPP, in complex
with 3/AMPCPP, and in complex with 7/AMPCPP have been
deposited at the Protein Data Bank with accession numbers 4CYU,
4CRJ, and 4CWB, respectively.
Surface Plasmon Resonance (SPR). All SPR experiments were
performed using Biacore T200 biosensor (GE Healthcare). Immobiliza-
tions were performed in HBS-EP+ running buffer (10 mM HEPES, pH
7.4, 150 mM NaCl, 50 μM EDTA, 0.05% [v/v] Tween-20) at 298 K with
a constant flow-rate of 10 μL min−1. SaHPPK and EcDHPS proteins
were covalently coupled to the NTA chip (GE Healthcare) surface using
a previously described method.53 Briefly, a single flow cell on the chip
surface was sequentially activated by injecting (1) 40 μL of nickel sulfate
and (2) 70 uL of a 1:1 mixture of NHS/EDC (N-hydroxysuccinimide/
N-ethyl-N′-(3-diethylaminopropyl)carbodiimide). Recombinant pro-
tein was diluted in the running buffer (SaHPPK to 225 μg mL−1;
EcDHPS to 80 μg mL−1) and injected over an activated flow cell for 20
min (200 μL). Amine-coupled surface was subsequently blocked with 70
μL of 1 M ethanolamine, pH 8.0, and then further regenerated with two
10 μL injections of 350 mM EDTA prepared in running buffer. Using
this coupling approach, average immobilization levels achieved were
5400 RU for SaHPPK and 7200 RU for EcDHPS. Additionally,
ubiquitin-specific-processing protease 7 (USP7) was coupled in a similar
fashion to provide for an unrelated negative control surface (6600 RU).
All SPR binding experiments were performed at 293 K in SPR binding
buffer (50 mM HEPES, pH 8.0, 150 mM NaCl, 5 mM DTT, 10 mM
MgCl2, 0.05% [v/v] Tween-20, 5% [v/v] DMSO). Analytes were
serially diluted (3-fold) in SPR binding buffer, injected for 30 s contact
time at 60 μL min−1, and then allowed to dissociate for 60 s. Each analyte
titration was performed in duplicate or greater. Binding sensorgrams
were processed, solvent-corrected, and double-referenced using
Scrubber software (BioLogic Software, Australia). SPR binding analysis
of several of the compounds investigated in this study revealed
dissociation rates that were not sufficiently slow to allow global fitting to
a kinetic binding model, for which the kd (dissociation rate constant)
must typically be <0.5 s−1 for SPR instruments to be able to capture
sufficient data points during the dissociation phase. Therefore, to
determine binding affinities (KD values), responses at equilibrium for
each analyte were fitted to a 1:1 steady-state affinity model available
within Scrubber using 8MG as a reference, as previously described.20
KinaseGlo Biochemical Assay. HPPK activity was quantified using
a KinaseGlo assay kit (Promega) as previously reported.21 In this, firefly
luciferase utilizes the remaining ATP after HPPK catalysis, producing a
L
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
luminescence signal that is directly proportional to ATP concentration.
The enzyme activity and optimum concentration to define kinetic
parameters were optimized as described previously.21 For kinetic
measurements, an optimized HPPK concentration of 0.4 ng μL−1 assay
volume was used, which allowed for monitoring of the first 10% of
reaction turnover within a reasonable time period (20 min).
Measurements were performed in 96-well plates using assay buffer
(100 mM Tris-HCl, pH 8.5, 10 mM MgCl2, 0.01% (w/v) BSA, 0.01%
(v/v) Tween 20 and 10 mM β-mercaptoethanol). Typically, 5 μL of test
compound (dissolved in 50% DMSO) and 20 μL of enzyme were added
to each well followed by 25 μL of assay buffer, giving 0.3 μM pterin and
0.2 μM ATP in a total reaction volume of 50 μL. After 20 min incubation
at room temperature, the enzymatic reaction was stopped with 50 μL of
KinaseGlo reagent. Luminescence was recorded after a further 10 min
using a FLUOstar Optima plate reader (BMG, Labtech Ltd.). Reactions
were performed in triplicate. Kinetic data and inhibition data were fit to
Michaelis−Menten and sigmoidal dose−response equations, respec-
tively, using GraphPad Prism.
7.64 (dd, J = 7.7, 1.8 Hz, 1H), 7.59 (m, 1H), 7.21 (d, J = 8.3 Hz, 1H),
7.08−7.01 (m, 1H), 6.28 (bs, 2H), 4.67 (s, 2H), 3.92 (s, 3H). HRMS:
m/z calcd for [M + H]+ C14H13N5O3S, 332.0812; found, 332.0821.
8-((2-([1,1′-Biphenyl]-4-yl)-2-oxoethyl)thio)-2-amino-1,9-di-
hydro-6H-purin-6-one (7). Compound 7 was synthesized using
1
general procedure A. Yield 9%. H NMR (DMSO-d6, 400 MHz): δ
10.97 (s, 1H), 8.11 (d, J = 8.5 Hz, 2H), 7.82 (d, J = 8.5 Hz, 2H), 7.77−
7.74 (m, 2H), 7.50 (m, 2H), 7.45−7.43 (m, 1H), 6.39 (s, 2H), 4.80 (s,
2H). HRMS: m/z calcd for [M + H]+ C19H15N5O2S, 378.1019; found,
378.1028.
2-Amino-8-((2-oxopropyl)thio)-1,9-dihydro-6H-purin-6-one
(9). Compound 9 was synthesized using general procedure A using 4
equiv of chloroacetone and 8 mL of 0.4 M NaOH. Yield 55%. 1H NMR
(DMSO-d6, 400 MHz): δ 12.60 (bs, 1H), 10.65 (bs, 1H), 6.33 (s, 2H),
4.16 (s, 2H), 2.23 (s, 3H). HRMS: m/z calcd for [M + H]+ C8H9N5O2S,
240.0550; found, 240.0553.
General Procedure B for the Synthesis of Compounds 8 and
10−13. 8-Mercaptoguanine (0.10 g, 0.55 mmol) was dissolved in 0.5 M
NaOH (3 mL), and to the solution was added substituted benzyl or
phenethyl bromide (0.61 mmol). The mixture was stirred at room
temperature for 5−24 h and then diluted with 0.5 M NaOH (8 mL). The
aqueous layer was extracted with EtOAc (3 × 20 mL), acidified with
acetic acid (pH 5.0), and stirred for 15−20 min. The precipitated solid
was collected by filtration, washed thoroughly with water and ethanol,
and dried to give the solid product.
2-Amino-8-((4-methoxyphenethyl)thio)-1,9-dihydro-6H-
purin-6-one (8). Compound 8 was synthesized using general
procedure B. Yield 73%. 1H NMR (DMSO-d6, 400 MHz) δ: 12.5 (bs,
1H), 10.5 (bs, 1H), 7.18−6.84 (m, 4H), 6.55 (s, 2H), 3.72 (s, 3H), 3.33
(t, J = 7.7 Hz, 2H,), 2.88 (t, J = 7.1 Hz).HRMS: m/z calcd for [M + H]+
C14H15F3N5O2S, 318.1019; found, 318.1021.
Molecular Modeling. Molecular modeling was performed using the
Schrodinger Suite 2014 (www.schrodinger.com) through the Maestro
̈
interface (Maestro, version 9.7).54 Protein preparation of 4CWB was
performed with the Protein Preparation Wizard workflow implemented
by Schrodinger (Epik, version 2.7),55 with deletion of all waters. In order
̈
to eliminate any bond length or bond angle biases in the structures,
compound 11 was subjected to a full minimization prior to docking
using LigPrep (LigPrep, version 2.9).56 Docking was carried out with
Glide,57 version 6.2, using Extra Precision (XP) mode.
Compound Procurement and Analysis. 8MG and compound 3
were purchased from Sigma-Aldrich and TimTec, respectively. All other
compounds were synthesized as described below. In all cases, 1H NMR
spectra were recorded on a 400 MHz Bruker NMR spectrometer, and
chemical shifts were referenced to the solvent peak. Analytical reversed-
phase high-performance liquid chromatography (RP-HPLC) was
performed on an Agilent 1260 Infinity HPLC system using an Agilent
Eclipse Plus C18 column (100 × 4.6 mm i.d., 3.5 μm) with a flow rate of
1 mL min−1 and UV detection at 214 and 254 nm. Elution was achieved
with standard HPLC buffers (buffer A: 99.9% H2O/0.1% TFA; buffer B:
99.9% CH3CN/0.1% TFA) using a gradient from 5% B/95% A to 100%
B over 10 min. All compounds were determined to be >95% purity by
this method.
General Procedure A for the Synthesis of Compounds 4−7
and 9. 8-Mercaptoguanine (0.200 g, 1.09 mmol) was dissolved in 0.4 M
NaOH (5.5 mL), and to the solution was added phenacyl bromide (or
analogue) (0.24 g, 1.2 mmol) in ethanol (0.9 mL). The reaction was
allowed to stir for 2 h, following which a white precipitate formed in
solution, which was collected by vacuum filtration to give the title
compound as a white amorphous solid.
2-Amino-8-(benzylthio)-1,9-dihydro-6H-purin-6-one (10).
Compound 10 was synthesized using general procedure B. Yield 36%.
1H NMR (DMSO-d6, 400 MHz) δ: 12.5 (bs, 1H, NH), 10.5 (bs, 1H),
7.29−7.21 (m, 5H), 6.57 (s, 2H), 4.34 (s, 2H). HRMS: m/z calcd for [M
+ H]+ C12H11N5OS, 274.0757; found, 274.0760.
2-Amino-8-((2-fluorobenzyl)thio)-1,9-dihydro-6H-purin-6-
one (11). Compound 11 was synthesized using general procedure B.
Yield 66%. 1H NMR (DMSO-d6, 400 MHz) δ: 12.5 (bs, 1H), 10.5 (bs,
1H), 7.37−7.03 (m, 4H), 6.30 (s, 2H), 4.30 (s, 2H). HRMS: m/z calcd
for [M + H]+ C12H10FN5OS, 292.0663; found, 292.0666.
2-Amino-8-((2-(trifluoromethyl)benzyl)thio)-1,9-dihydro-6H-
purin-6-one (12). Compound 12 was synthesized using general
procedure B. Yield 52%. 1H NMR (DMSO-d6, 400 MHz) δ: 12.6 (bs,
1H), 10.6 (bs, 1H), 7.75−7.48 (m, 4H), 6.36 (s, 2H), 4.53 (s, 2H).
HRMS: m/z calcd for [M + H]+ C13H10F3N5OS, 342.0631; found,
342.0636.
4-(((2-Amino-6-oxo-6,9-dihydro-1H-purin-8-yl)thio)methyl)-
benzonitrile (13). Compound 13 was synthesized using general
procedure B. Yield 73%. 1H NMR (DMSO-d6, 400 MHz) δ 10.89 (bs,
1H), 7.77 (d, J = 8.3 Hz, 2H), 7.55 (d, J = 8.3 Hz, 2H), 6.43 (bs, 2H),
4.45 (s, 2H). HRMS: m/z calcd for [M + H]+ C13H10N6OS, 299.0710;
found, 299.0699.
2-Amino-8-((2-oxo-2-phenylethyl)thio)-1,9-dihydro-6H-
purin-6-one (4). Compound 4 was synthesized using general
1
procedure A. Yield 50%. H NMR (DMSO-d6, 400 MHz): δ 12.51
(bs, 1H), 10.90 (bs, 1H), 8.03−8.01 (m, 2H), 7.67 (t, J = 7.4 Hz, 1H),
7.55 (t, J = 7.7 Hz, 2H), 6.45 (s, 2H), 4.86 (s, 2H). HRMS: m/z calcd for
[M + H]+ C13H11N5O2S, 302.0706; found, 302.0710.
2-Amino-8-((1-oxo-1-phenylpropan-2-yl)thio)-1,9-dihydro-
6H-purin-6-one (5). Compound 5 was synthesized using general
procedure A including the following work up. The reaction mixture was
diluted with 0.5 M NaOH (7 mL). The aqueous layer was extracted with
EtOAc (3 × 3 mL) and acidified with acetic acid (2 mL). The precipitate
solid was collected by filtration and dried to give solid product. Yield
28%. 1H NMR (DMSO-d6, 400 MHz) δ 10.91 (bs, 1H), 8.01 (d, J = 7.4
Hz, 2H), 7.65 (t, J = 7.3 Hz, 1H), 7.52 (t, J = 7.7 Hz, 2H), 6.46 (bs, 2H),
5.43 (q, J = 6.7 Hz, 1H), 1.53 (d, J = 6.9 Hz, 3H). HRMS: m/z calcd for
[M + Na]+ C14H13N5O2S, 338.0682; found, 338.0688.
2-Amino-8-((2-(2-methoxyphenyl)-2-oxoethyl)thio)-1,9-di-
hydro-6H-purin-6-one (6). Compound 6 was synthesized using
general procedure A including the following work up. The reaction
mixture was diluted with 0.5 M NaOH (7 mL). The aqueous layer was
extracted with EtOAc (3 × 3 mL) and acidified with acetic acid (2 mL).
The precipitate solid was collected by filtration and dried to give solid
product. Yield 63%. 1H NMR (DMSO-d6, 400 MHz) δ 10.58 (bs, 1H),
ASSOCIATED CONTENT
* Supporting Information
CSP data for the binding of 11 to AMPCPP saturated SaHPPK.
This material is available free of charge via the Internet at http://
pubs.acs.org.
■
S
Accession Codes
The coordinates of SaHPPK in complex with AMPCPP, in
complex with 3/AMPCPP, and in complex with 7/AMPCPP
have been deposited at the Protein Data Bank with accession
numbers 4CYU, 4CRJ, and 4CWB, respectively.
AUTHOR INFORMATION
■
Corresponding Author
*Tel.: ++61 3 9903 9543. E-mail: james.swarbrick@monash.edu.
M
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(13) Zhao, Y.; Hammoudeh, D.; Yun, M. K.; Qi, J.; White, S. W.; Lee,
R. E. Structure-based design of novel pyrimido[4,5-c]pyridazine
derivatives as dihydropteroate synthase inhibitors with increased
affinity. ChemMedChem. 2012, 7, 861−870.
(14) Hammoudeh, D. I.; Date, M.; Yun, M. K.; Zhang, W.; Boyd, V. A.;
Viacava Follis, A.; Griffith, E.; Lee, R. E.; Bashford, D.; White, S. W.
Identification and characterization of an allosteric inhibitory site on
dihydropteroate synthase. ACS Chem. Biol. 2014, 1294−1302.
(15) Haruki, H.; Pedersen, M. G.; Gorska, K. I.; Pojer, F.; Johnsson, K.
Tetrahydrobiopterin biosynthesis as an off-target of sulfa drugs. Science
2013, 340, 987−91.
(16) Kluytmans-Vandenbergh, M. F.; Kluytmans, J. A. Community-
acquired methicillin-resistant Staphylococcus aureus: current perspec-
tives. Clin. Microbiol. Infect. 2006, 12, 9−15.
(17) Kollef, M. H.; Micek, S. T. Methicillin-resistant Staphylococcus
aureus: a new community-acquired pathogen? Curr. Opin. Infect. Dis.
2006, 19, 161−168.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
All SaHPPK crystals were grown at the C3 Crystallization Centre
at CSIRO, Parkville, Australia, and X-ray data were obtained at
the MX2 beamline at the Australian Synchrotron, Victoria,
Australia. We thank OpenEye for a license to use their software.
J.B.B. was supported by an NHMRC Research Fellowship
(1020411).
ABBREVIATIONS USED
■
HMDP, 6-hydroxymethyl-7,8-dihydro-pterin; SMX, sulfame-
thoxazole; TMP, trimethoprim; SaHPPK, 6-hydroxymethyl-
7,8-dihydropterin pyrophosphokinase from S. aureus; EcHPPK,
6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase from E.
coli; EcDHPS, dihydropteroate synthase from E. coli; MRSA,
methicillin-resistant S. aureus; 8MG, 8-mercaptoguanine; SPR,
surface plasmon resonance; HSQC, heteronuclear single
quantum coherence; SAR, structure−activity relationships
(18) Ohlsen, K.; Lorenz, U. Novel targets for antibiotics in
Staphylococcus aureus. Future Microbiol. 2007, 2, 655−666.
(19) Spaulding, A. R.; Salgado-Pabon, W.; Merriman, J. A.; Stach, C. S.;
Ji, Y.; Gillman, A. N.; Peterson, M. L.; Schlievert, P. M. Vaccination
against Staphylococcus aureus pneumonia. J. Infect. Dis. 2014, 209, 1955−
1962.
(20) Chhabra, S.; Barlow, N.; Dolezal, O.; Hattarki, M. K.; Newman, J.;
Peat, T. S.; Graham, B.; Swarbrick, J. D. Exploring the chemical space
around 8-mercaptoguanine as a route to new inhibitors of the folate
biosynthesis enzyme HPPK. PLoS One 2013, 8, e59535.
(21) Chhabra, S.; Dolezal, O.; Collins, B. M.; Newman, J.; Simpson, J.
S.; Macreadie, I. G.; Fernley, R.; Peat, T. S.; Swarbrick, J. D. Structure of
S. aureus HPPK and the discovery of a new substrate site inhibitor. PLoS
One 2012, 7, e29444.
(22) Derrick, J. P. The structure and mechanism of 6-hydroxymethyl-
7,8-dihydropterin pyrophosphokinase. Vitam. Horm. 2008, 79, 411−
433.
(23) Blaszczyk, J.; Shi, G.; Li, Y.; Yan, H.; Ji, X. Reaction trajectory of
pyrophosphoryl transfer catalyzed by 6-hydroxymethyl-7,8-dihydrop-
terin pyrophosphokinase. Structure 2004, 12, 467−475.
(24) Xiao, B.; Shi, G.; Gao, J.; Blaszczyk, J.; Liu, Q.; Ji, X.; Yan, H.
Unusual conformational changes in 6-hydroxymethyl-7,8-dihydropterin
pyrophosphokinase as revealed by X-ray crystallography and NMR. J.
Biol. Chem. 2001, 276, 40274−40281.
(25) Garcon, A.; Levy, C.; Derrick, J. P. Crystal structure of the
bifunctional dihydroneopterin aldolase/6-hydroxymethyl-7,8-dihydrop-
terin pyrophosphokinase from Streptococcus pneumoniae. J. Mol. Biol.
2006, 360, 644−653.
(26) Hennig, M.; Dale, G. E.; D’Arcy, A.; Danel, F.; Fischer, S.; Gray, C.
P.; Jolidon, S.; Muller, F.; Page, M. G.; Pattison, P.; Oefner, C. The
structure and function of the 6-hydroxymethyl-7,8-dihydropterin
pyrophosphokinase from Haemophilus influenzae. J. Mol. Biol. 1999,
287, 211−219.
(27) Lawrence, M. C.; Iliades, P.; Fernley, R. T.; Berglez, J.; Pilling, P.
A.; Macreadie, I. G. The three-dimensional structure of the bifunctional
6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase/dihydropter-
oate synthase of Saccharomyces cerevisiae. J. Mol. Biol. 2005, 348, 655−
670.
(28) Xiao, B.; Shi, G.; Chen, X.; Yan, H.; Ji, X. Crystal structure of 6-
hydroxymethyl-7,8-dihydropterin pyrophosphokinase, a potential target
for the development of novel antimicrobial agents. Structure 1999, 7,
489−496.
REFERENCES
■
(1) Bermingham, A.; Derrick, J. P. The folic acid biosynthesis pathway
in bacteria: evaluation of potential for antibacterial drug discovery.
BioEssays 2002, 24, 637−648.
(2) Swarbrick, J. D.; Iliades, P.; Simpson, J. S.; Macreadie, I. Folate
biosynthesis - reappraisal of old and novel targets in the search for new
antimicrobials. Open Enzyme Inhib. J. 2008, 1, 12−33.
(3) Masters, P. A.; O’Bryan, T. A.; Zurlo, J.; Miller, D. Q.; Joshi, N.
Trimethoprim-sulfamethoxazole revisited. Arch. Int. Med. 2003, 163,
402−410.
(4) Adra, M.; Lawrence, K. R. Trimethoprim/sulfamethoxazole for
treatment of severe Staphylococcus aureus infections. Ann. Pharmacother.
2004, 38, 338−341.
(5) Forgacs, P.; Wengenack, N. L.; Hall, L.; Zimmerman, S. K.;
Silverman, M. L.; Roberts, G. D. Tuberculosis and trimethoprim-
sulfamethoxazole. Antimicrob. Agents Chemother. 2009, 53, 4789−4793.
(6) Martin, S. I.; Fishman, J. A. Pneumocystis pneumonia in solid organ
transplant recipients. Am. J. Transplant. 2009, 9, S227−S233.
(7) Pashley, T. V.; Volpe, F.; Pudney, M.; Hyde, J. E.; Sims, P. F.;
Delves, C. J. Isolation and molecular characterization of the bifunctional
hydroxymethyldihydropterin pyrophosphokinase-dihydropteroate syn-
thase gene from Toxoplasma gondii. Mol. Biochem. Parasitol. 1997, 86,
37−47.
(8) Thera, M. A.; Sehdev, P. S.; Coulibaly, D.; Traore, K.; Garba, M. N.;
Cissoko, Y.; Kone, A.; Guindo, A.; Dicko, A.; Beavogui, A. H.; Djimde, A.
A.; Lyke, K. E.; Diallo, D. A.; Doumbo, O. K.; Plowe, C. V. Impact of
trimethoprim-sulfamethoxazole prophylaxis on falciparum malaria
infection and disease. J. Infect. Dis. 2005, 192, 1823−1829.
(9) Hevener, K. E.; Yun, M. K.; Qi, J.; Kerr, I. D.; Babaoglu, K.; Hurdle,
J. G.; Balakrishna, K.; White, S. W.; Lee, R. E. Structural studies of pterin-
based inhibitors of dihydropteroate synthase. J. Med. Chem. 2010, 53,
166−177.
(10) Frey, K. M.; Liu, J.; Lombardo, M. N.; Bolstad, D. B.; Wright, D.
L.; Anderson, A. C. Crystal structures of wild-type and mutant
methicillin-resistant Staphylococcus aureus dihydrofolate reductase reveal
an alternate conformation of NADPH that may be linked to
trimethoprim resistance. J. Mol. Biol. 2009, 387, 1298−1308.
(11) Bourne, C. R.; Bunce, R. A.; Bourne, P. C.; Berlin, K. D.; Barrow,
E. W.; Barrow, W. W. Crystal structure of Bacillus anthracis dihydrofolate
reductase with the dihydrophthalazine-based trimethoprim derivative
RAB1 provides a structural explanation of potency and selectivity.
Antimicrob. Agents Chemother. 2009, 53, 3065−3073.
(29) Wood, H. C. S. Specific inhibition of dihydrofolate biosynthesis
a new approach to chemotherapy. In Chemistry and Biology of Pteridines;
Pfleiderer, W., Ed.; Walter de Gruyter: New York, 1975.
(30) Shi, G.; Blaszczyk, J.; Ji, X.; Yan, H. Bisubstrate analogue inhibitors
of 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase: synthesis
and biochemical and crystallographic studies. J. Med. Chem. 2001, 44,
1364−1371.
(12) Yun, M. K.; Wu, Y.; Li, Z.; Zhao, Y.; Waddell, M. B.; Ferreira, A.
M.; Lee, R. E.; Bashford, D.; White, S. W. Catalysis and sulfa drug
resistance in dihydropteroate synthase. Science 2012, 335, 1110−1114.
(31) Shi, G.; Shaw, G.; Li, Y.; Wu, Y.; Yan, H.; Ji, X. Bisubstrate analog
inhibitors of 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase:
N
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
new lead exhibits a distinct binding mode. Bioorg. Med. Chem. 2012, 20,
4303−4309.
(52) Emsley, P.; Cowtan, K. Coot: model-building tools for molecular
graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 2126−
2132.
(32) Shi, G.; Shaw, G.; Liang, Y. H.; Subburaman, P.; Li, Y.; Wu, Y.;
Yan, H.; Ji, X. Bisubstrate analogue inhibitors of 6-hydroxymethyl-7,8-
dihydropterin pyrophosphokinase: new design with improved proper-
ties. Bioorg. Med. Chem. 2012, 20, 47−57.
(53) Kimple, A. J.; Muller, R. E.; Siderovski, D. P.; Willard, F. S. A
capture coupling method for the covalent immobilization of
hexahistidine tagged proteins for surface plasmon resonance. Methods
Mol. Biol. 2010, 627, 91−100.
(33) Bourne, C. R. Utility of the biosynthetic folate pathway for targets
in antimicrobial discovery. Antibiotics 2014, 3, 28.
(54) Maestro, version 9.7; Schrodinger, LLC: New York, 2014.
̈
(55) Schrodinger Suite 2014-1: Protein Preparation Wizard, Epik version
̈
(34) Stammers, D. K.; Achari, A.; Somers, D. O.; Bryant, P. K.;
Rosemond, J.; Scott, D. L.; Champness, J. N. 2.0 Å X-ray structure of the
ternary complex of 7,8-dihydro-6-hydroxymethylpterinpyrophosphoki-
nase from Escherichia coli with ATP and a substrate analogue. FEBS Lett.
1999, 456, 49−53.
(35) Williamson, M. P. Using chemical shift perturbation to
characterise ligand binding. Prog. Nucl. Magn. Reson. Spectrosc. 2013,
73, 1−16.
(36) Medek, A.; Hajduk, P. J.; Mack, J.; Fesik, S. W. The use of
differential chemical shifts for determining the binding site location and
orientation of protein-bound ligands. J. Am. Chem. Soc. 2000, 122,
1241−1242.
2.7, Impact version 6.2, Prime version 3.4; Schrodinger, LLC: New York,
̈
2014.
(56) LigPrep; Schrodinger, LLC: New York, 2014.
̈
(57) Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.;
Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. Extra
precision glide: docking and scoring incorporating a model of
hydrophobic enclosure for protein-ligand complexes. J. Med. Chem.
2006, 49, 6177−6196.
(37) Lescop, E.; Lu, Z.; Liu, Q.; Xu, H.; Li, G.; Xia, B.; Yan, H.; Jin, C.
Dynamics of the conformational transitions in the assembling of the
Michaelis complex of a bisubstrate enzyme: a 15N relaxation study of
Escherichia coli 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase.
Biochemistry 2009, 48, 302−312.
(38) Yang, R.; Lee, M. C.; Yan, H.; Duan, Y. Loop conformation and
dynamics of the Escherichia coli HPPK apo-enzyme and its binary
complex with MgATP. Biophys. J. 2005, 89, 95−106.
(39) Iwaoka, M.; Isozumi, N. Hypervalent nonbonded interactions of a
divalent sulfur atom. Implications in protein architecture and the
functions. Molecules 2012, 17, 7266−7283.
(40) Blaszczyk, J.; Shi, G.; Yan, H.; Ji, X. Catalytic center assembly of
HPPK as revealed by the crystal structure of a ternary complex at 1.25 Å
resolution. Structure 2000, 8, 1049−1058.
(41) Yun, M. K.; Hoagland, D.; Kumar, G.; Waddell, M. B.; Rock, C.
O.; Lee, R. E.; White, S. W. The identification, analysis and structure-
based development of novel inhibitors of 6-hydroxymethyl-7,8-
dihydropterin pyrophosphokinase. Bioorg. Med. Chem. 2014, 22,
2157−2165.
(42) Zhou, P.; Zou, J. W.; Tian, F. F.; Shang, Z. C. Fluorine bonding -
how does it work in protein-ligand interactions? J. Chem. Inf. Model.
2009, 49, 2344−2355.
(43) Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K.
W.; Kopple, K. D. Molecular properties that influence the oral
bioavailability of drug candidates. J. Med. Chem. 2002, 45, 2615−2623.
(44) Chhabra, S.; Newman, J.; Peat, T. S.; Fernley, R. T.; Caine, J.;
Simpson, J. S.; Swarbrick, J. D. Crystallization and preliminary X-ray
analysis of 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase
from Staphylococcus aureus. Acta Crystallogr., Sect. F: Struct. Biol. Cryst.
Commun. 2010, 66, 575−578.
(45) Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeifer, J.; Bax,
A. NMRPipe: a multidimensional spectral processing system based on
UNIX pipes. J. Biomol. NMR 1995, 6, 277−293.
(46) Bartels, C.; Xia, T. H.; Billeter, M.; Guntert, P.; Wuthrich, K. The
program XEASY for computer-supported NMR spectral analysis of
biological macromolecules. J. Biomol. NMR 1995, 6, 1−10.
(47) Goddard, T. D.; Kneller, D. G. SPARKY 3; University of
California: San Francisco, CA.
(48) Kabsch, W. XDS. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010,
66, 125−32.
(49) Collaborative Computational Project, Number 4. The CCP4
suite: programs for protein crystallography. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 1994, 50, 760−763.
(50) Storoni, L. C.; McCoy, A. J.; Read, R. J. Likelihood-enhanced fast
rotation functions. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60,
432−438.
(51) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 1997, 53, 240−255.
O
dx.doi.org/10.1021/jm501417f | J. Med. Chem. XXXX, XXX, XXX−XXX