The identification, analysis and structure-based development of novel inhibitors of 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase

Article abstract of DOI:10.1016/j.bmc.2014.02.022

6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase (HPPK) is an essential enzyme in the microbial folate biosynthetic pathway. This pathway has proven to be an excellent target for antimicrobial development, but widespread resistance to common therapeutics including the sulfa drugs has stimulated interest in HPPK as an alternative target in the pathway. A screen of a pterin-biased compound set identified several HPPK inhibitors that contain an aryl substituted 8-thioguanine scaffold, and structural analyses showed that these compounds engage the HPPK pterin-binding pocket and an induced cryptic pocket. A preliminary structure activity relationship profile was developed from biophysical and biochemical characterizations of derivative molecules. Also, a similarity search identified additional scaffolds that bind more tightly within the HPPK pterin pocket. These inhibitory scaffolds have the potential for rapid elaboration into novel lead antimicrobial agents.

Full text of DOI:10.1016/j.bmc.2014.02.022

Bioorganic & Medicinal Chemistry 22 (2014) 2157–2165
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
The identification, analysis and structure-based development
of novel inhibitors of 6-hydroxymethyl-7,8-dihydropterin
pyrophosphokinase
Mi-Kyung Yun a, Daniel Hoagland b, Gyanendra Kumar a, M. Brett Waddell a, Charles O. Rock c,d
,
Richard E. Lee b,d, , Stephen W. White a,d,
⇑
⇑
a Department of Structural Biology, St. Jude Children’s Research Hospital, Memphis, TN 38105, USA
b Department of Chemical Biology and Therapeutics, St. Jude Children’s Research Hospital, Memphis, TN 38105, USA
c Department of Infectious Diseases, St. Jude Children’s Research Hospital, Memphis, TN 38105, USA
d Department of Microbiology, Immunology and Biochemistry, University of Tennessee Health Science Center, Memphis, TN 38163, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase (HPPK) is an essential enzyme in the microbial
folate biosynthetic pathway. This pathway has proven to be an excellent target for antimicrobial devel-
opment, but widespread resistance to common therapeutics including the sulfa drugs has stimulated
interest in HPPK as an alternative target in the pathway. A screen of a pterin-biased compound set iden-
tified several HPPK inhibitors that contain an aryl substituted 8-thioguanine scaffold, and structural anal-
yses showed that these compounds engage the HPPK pterin-binding pocket and an induced cryptic
pocket. A preliminary structure activity relationship profile was developed from biophysical and bio-
chemical characterizations of derivative molecules. Also, a similarity search identified additional scaf-
folds that bind more tightly within the HPPK pterin pocket. These inhibitory scaffolds have the
potential for rapid elaboration into novel lead antimicrobial agents.
Received 31 December 2013
Revised 4 February 2014
Accepted 14 February 2014
Available online 25 February 2014
Keywords:
Antimicrobial
HPPK
Pterin
Folate
Ó 2014 Elsevier Ltd. All rights reserved.
X-ray crystallography
1. Introduction
since their discovery 80 years ago,3 but their use has been seriously
compromised by widespread microbial resistance and severe aller-
A number of key metabolic reactions, such as the synthesis of
thymine, involve a one-carbon transfer step, and the carbon atom
is supplied by the essential cofactor tetrahydrofolate.1 Higher
organisms derive folate from their diet as vitamin B9, but the vast
majority of eubacteria and many lower eukaryotes synthesize tet-
rahydrofolate de novo. The folate biosynthetic pathway is therefore
an ideal target for the development of antibiotics.2 This is exempli-
fied by the sulfonamide class of drugs (sulfa drugs) that target the
enzyme dihydropteroate synthase (DHPS) in the pathway. Sulfa
drugs have remained important broad spectrum clinical agents
gic reactions.
We recently showed that DHPS from the pathogen Francisella
tularensis is part of a bifunctional enzyme and fused to 6-hydroxy-
methyl-7,8-dihydropterin pyrophosphokinase (HPPK) that cata-
lyzes the previous step in the pathway.4 We also showed that
one of our DHPS pterin-pocket inhibitors engages the HPPK pterin
pocket, even though there is no structural similarity between the
pockets. Despite its high conservation and pivotal role in folate
synthesis, there have been relatively few attempts to develop lead
inhibitory compounds against HPPK as potential novel antibiot-
ics.5–7 This is somewhat surprising because many HPPK crystal
structures are now available4,5,8–11 and the catalytic mechanism
is understood.6,7,12–18 This untapped potential has been noted,19
and there has recently been renewed interest in HPPK as an anti-
microbial drug target.20–23
Abbreviations: HPPK, 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase;
DHP, 6-hydroxymethyl-7,8-dihydropterin; DHPS, dihydropteroate synthase; SPR,
surface plasmon resonance; KD, equilibrium dissociation constant; IC50
, half
maximal inhibitory concentration; AMPCPP,
a
,b-methyleneadenosine 50-triphos-
HPPK is a small (ꢀ18 kDa) highly conserved enzyme with an
ab
phate; EcHPPK, Escherichia coli 6-hydroxymethyl-7,8-dihydropterin pyrophospho-
kinase; BME, 2-mercaptoethanol.
fold that catalyzes the transfer of pyrophosphate from ATP to
6-hydroxymethyl-7,8-dihydropterin (DHP) to form 6-hydroxy-
methyl-7,8–dihydropterin-pyrophosphate (DHPPP), one of the
two substrates of DHPS. The adenosine ring of ATP packs into a
⇑
Corresponding authors. Tel.: +1 901 595 6617; fax: +1 901 595 5715 (R.E.L.);
tel.: +1 901 595 3040; fax: +1 901 595 3032 (S.W.W.).
E-mail addresses: richard.lee@stjude.org (R.E. Lee), stephen.white@stjude.org
(S.W. White).
http://dx.doi.org/10.1016/j.bmc.2014.02.022
0968-0896/Ó 2014 Elsevier Ltd. All rights reserved.

2158
M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
Table 1
Compound 1 variants
conserved cleft, the triphosphate is coordinated by two essential
Mg2+ ions, and DHP binds within an adjacent pocket with the pter-
O
N
in ring p-stacked between two conserved aromatic residues. HPPK
R
N
uses an ordered enzyme mechanism in which the ATP cleft is first
occupied, followed by the binding of DHP. Similar to DHPS,24 HPPK
uses stabilizing loop conformational changes to assemble the com-
plete active site and the DHP binding pocket.13,17,25,26
HN
H2N
S
N
H
b
Compound
R
% Inhibitiona
KD
(
l
M)
IC50
139
254
64
(lM)
Our discovery that a DHPS pterin-pocket inhibitor can also en-
gage the pterin pocket of F. tularensis HPPK (FtHPPK) is not alto-
gether surprising because the product of HPPK, DHP, is the
substrate for DHPS, and the architectures of both pockets have
therefore evolved to engage the same small molecule. This obser-
vation prompted us to screen our library of DHPS pterin-pocket
binding molecules for additional HPPK inhibitors. These studies
yielded two related compounds, and using a structure-based ap-
proach, we have synthesized derivative compounds and derived
an initial SAR pattern. Based on these data, we then performed a
similarity search of the NCI compound repository, and identified
and structurally characterized several inhibitory fragment scaf-
folds for future optimization.
1
2
3
75
53
80
9.7 0.4
O
N
29
2
6.7 0.2
5.8 0.2
O
O
O
OH
F
4
5
77
73
82
79
11.2 0.5
2. Results
N
6
7
68
39
3
88
44
2.1. Initial screen of DHPS pterin pocket inhibitors
O
O
During the course of our drug discovery studies on DHPS, we
have generated a library of ꢀ230 potential pterin pocket binding
molecules. Using an endpoint HPPK assay that monitors unpro-
cessed ATP substrate as a direct readout of inhibition, we screened
these compounds against Escherichia coli HPPK (EcHPPK). The
screen revealed 2 compounds that significantly inhibit HPPK at
74
77
14.5 0.6
O
O
O
8
9
5.9 0.3
51
250 lM (compounds 1 and 2, Table 1). To characterize the binding
48
18
18.9 0.9
243
of 1 and 2 to EcHPPK in more detail, we used surface plasmon res-
onance (SPR) to measure their binding characteristics. EcHPPK was
immobilized on the sensor chip, and binding was measured in the
H
10
DHPc
>100
1.31 0.03
absence and presence of 2 lM of the non-hydrolysable ATP analog
a
% Inhibition of enzyme activity was tested at 250
l
M of the compound.
KD value was determined by SPR with equilibrium affinity analysis in the
M AMPCPP.
AMPCPP. The sensorgrams and binding isotherms are shown in Fig-
ures S1a and S1b, and it is clear that the compounds showed no
appreciable binding in the absence of AMPCPP but robust binding
in the presence of AMPCPP. In the HPPK enzyme mechanism, the
assembly of the pterin-binding pocket depends on ATP-dependent
conformational changes in the three active site loops,13 and the
SPR data are therefore consistent with 1 and 2 both engaging the
pterin pocket. The rapid dissociation rates (kd) of 1 and 2 precluded
the measurement of KD values using the SPR kinetic approach, but
it was possible to obtain the KD values using the alternative equi-
librium affinity method, and these are shown in Table 1. We also
used the identical SPR protocol to measure the KD of DHP (Table 1,
Fig. S2), and this confirmed that 1 and 2 represent weak initial hit
molecules that require significant optimization to efficiently com-
pete with the HPPK pterin substrate. Note that it was possible to
measure the KD of DHP by SPR using the kinetic approach, and
the value agrees well with that obtained by the equilibrium affinity
b
presence of 2
l
c
The pterin substrate of HPPK, 6-hydroxymethyl-7,8-dihydropterin.
(SaHPPK).20 Those studies included structural characterizations
using X-ray crystallography which revealed that free 8-thiogua-
nine is sandwiched between the two conserved aromatic side
chains with the ring oxygen and nitrogen atoms engaging in
hydrogen bonding interactions similar to those of the natural
pterin substrate. However, compounds 1 and 2 appear to be more
potent than either 8-thioguanine itself or derivatives that were re-
cently reported.23 We therefore analyzed their ternary complexes
with AMPCPP by X-ray crystallography to understand the struc-
tural basis of this potency. EcHPPK forms high quality crystals in
well characterized conditions,8 and it was possible to obtain high
resolution structures for both ternary complexes (Table 2,
Table S1, Fig. 1, Fig. S3). It should be noted that the crystallization
conditions for 1 and 2 contain 0.2 M CaCl2, and based on the high
quality refined electron density (Fig. S4), it was apparent that Ca2+
ions had replaced Mg2+ ions at the active site. However, the two-
metal site is not part of the pterin pocket, and this crystallization
artifact does not appear to impact the binding of these compounds
or those described later in this study. The identity of the metals in
each of the structures is noted in Tables S1 and S2.
method (1.31 0.03 lM KD(eq) versus 1.05 0.01 lM KD(kin);
Fig. S2). Note also that our measured KD for DHP is significantly
higher than that previously reported by Li et al.15 but we attribute
this discrepancy to the much lower concentration of AMPCPP used
in our experiment (2 lM vs 100 lM). Isothermal titration calorim-
etry experiments (not shown) support this explanation.
2.2. The 8-thioguanine scaffold
Overall, both structures are very similar to that of the previ-
ously reported HPPKꢁAMPCPPꢁDHP substrate ternary complex,13
and Figure 1a shows a detailed stereo view of the compound 1 ter-
nary complex. The 8-thioguanine scaffold occupies the pterin
Compounds 1 and 2 both contain an 8-alkylated-8-thioguanine
scaffold, and its ability to mediate the inhibition of EcHPPK is
consistent with similar studies on Staphylococcus aureus HPPK

M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
2159
Table 2
Summary of crystallographic refinement data
a
b
Compound
Space group
Cell dimensions
, b, c (°)
Resolution range (Å)
Rwork/Rfree
PDB accession code
a, b, c (Å)
a
1
2
3
7
8
9
18
21
P21
P21
P21
P21
P21
P21
P21
P31
35.9, 57.4, 38.3
35.8, 57.6, 38.5
35.9, 58.1, 38.4
36.5, 58.4, 38.4
36.0, 57.9, 38.5
35.9, 58.0, 38.6
35.9, 57.9, 38.7
107.3, 107.3, 41.8
90.0, 114.0, 90.0
90.0, 115.4, 90.0
90.0, 115.6, 90.0
90.0, 113.8, 90.0
90.0, 115.4, 90.0
90.0, 115.5, 90.0
90.0, 115.6, 90.0
90.0, 90.0, 120.0
22.2–1.3
32.3–1.1
34.7–1.1
33.4–1.7
34.8–1.3
19.9–1.1
34.9–1.1
35.1–2.0
0.144/0.177
0.134/0.149
0.152/0.167
0.168/0.213
0.149/0.174
0.173/0.186
0.170/0.191
0.159/0.192
4M5G
4M5H
4M5I
4M5J
4M5K
4M5L
4M5M
4M5N
a
Rfree is the R value obtained for a test set of reflections consisting of randomly selected 5% subset of the data set excluded from refinement.
Coordinates and structure factors for all the structures have been deposited in the Protein Data Bank.
b
Figure 1. Crystal structures of the HPPK ternary complexes with compounds 1 and 2. (a) Detailed stereo view of the 1 ternary complex. See text for details of the interaction.
The complex contains many hydrogen bonds (dotted red lines) that involve a number of ordered water molecules (red spheres). Loop2 and loop3 that become ordered in the
complex are colored cyan and magenta, respectively. Only the phosphate groups (orange/red) of AMPCPP are shown for clarity (the a-b carbon bond is shown in green). (b)
Comparison of the 1 (yellow carbons) and DHP (grey carbons) ternary complexes. Note the movements (indicated by arrows) of Leu45, Trp89 and Arg92. (c) Comparison of
the 1 and 2 ternary complexes showing the different orientations of the two substituents within the pocket bounded by Leu45 and Trp89.
pocket in much the same way as the pterin ring of DHP. It is sand-
wiched between Tyr53 and Phe123, and ring oxygen and nitrogen
atoms engage in specific hydrogen bonding interactions with the
side chains of Thr42 and Asn55, the main chains of Pro43 and
Leu45, and a number of structured water molecules. Loop2 and
loop3 both adopt ordered structures to facilitate these interactions
that closely resemble their conformations in the substrate ternary
complex. The substrate ternary complex suggests that substituents
at the 8-thio position would not be tolerated because of steric
clashes with the side chains of Leu45 and Trp89. However, 8-thio
substitution creates a new cryptic binding pocket by an induced
fit mechanism in which these side chains rotate to generate the re-
quired space (Fig. 1b). Also, the inherent flexibilities of loop2 and
loop3 that contain these two residues are exploited by small addi-
tional conformational adjustments. Although Trp89 adopts a new
conformation, the ring nitrogen atom is still able to form the
complex, Arg84 and Arg92 engage the a- and a/b-phosphates of
the nucleotide, respectively, and Arg82 forms a salt bridge with
Glu77. However, in the inhibitor structures (Fig. 1a), Arg82 and
Arg92 move to form salt bridges with Asp95 and to engage the
a- and b-phosphates, respectively, while Arg84 maintains its inter-
action with the -phosphate. These movements appear to be re-
a
lated to the absence of the hydroxyl group that is present in DHP
and to the ‘pulling’ effect of the substituents at the 8-thio position.
Together, these create additional space adjacent to the pterin pock-
et that becomes occupied by the side chain of Arg92 (Fig. 1b),
which then allows Arg82 and Arg84 to relocate into their new posi-
tions (Fig. 1a).
2.3. Analysis of 8-substituted-8-thioguanine derivatives
The enzyme end-point assay and the SPR data both suggest that
1 is a marginally better inhibitor of EcHPPK than 2 (Table 1).
Although the crystal structures are very similar, the source of this
potency appears to be a key interaction in 1 that involves the car-
bonyl group on the linker between the phenyl and 8-thioguanine
moieties (Fig. 1a). Specifically, the carbonyl group stacks onto the
guanidinium group of Arg121 that, in turn, forms a stabilizing salt
bridge interaction with the c-phosphate of AMPCPP. The carbonyl
group is not present in 2, and the terminal phenyl group moves fur-
ther into the pocket to partially stack with Arg121 (Fig. 2a). How-
hydrogen bonding interaction with the c-phosphate oxygen atoms
of AMPCPP that is observed in the substrate ternary complex. The
structure of the compound 2 ternary complex is virtually identical
to that of 1 apart from the orientation of the substituent within the
induced pocket (Fig. 1c).
Movements of Arg82 and Arg92 adjacent to the phosphate
groups of the bound ATP play an essential role in HPPK cataly-
sis,16,27 and compounds 1 and 2 induce similar conformational
changes in these conserved basic residues. In the substrate ternary

2160
M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
ever, this movement requires an additional rotation of the Leu45
side chain (Fig. 2a), and appears to result in a less optimal docking
with the newly created pocket.
other structures. This explains the lower activity of 9, and the
much lower activity of 13 in which the ortho-substituent is some-
what larger (Table S3). It should be noted that there is a crystallo-
graphic caveat to these structural observations, namely, a close
contact with an adjacent molecule in the crystal lattice that im-
pacts the para substitution more than the ortho and meta
substitutions.
Having established the importance of the carbonyl group within
1, we synthesized and tested the activities of a number of variants
with different substitutions on the phenyl ring. Table 1 shows the 7
derivatives (compounds 3–9) with comparable activities to 1 and
2, and Table S3 shows the 7 compounds with significantly reduced
activities (compounds 11–17). Analysis of the latter compounds re-
veals that extension of the phenyl substituent (11), loss of the aro-
matic ring (12), large substitution at the ortho position of the
phenyl ring (13), and changes to the carbonyl-containing linker
group (14–17) all result in loss of activity. The results of the end-
point inhibition assay shown in Table 1 cannot readily distinguish
between the potencies of compounds 1–9, and we therefore deter-
mined their KD values using SPR and also measured their IC50 val-
ues by conducting the enzyme assay at increasing concentrations
2.4. Further studies on the HPPK pterin-binding pocket
In contrast to previous studies,20 we found that 8-thioguanine
(10) is a very poor inhibitor of EcHPPK and shows minimal binding
by SPR in the presence of AMPCPP (Table 1 and Fig. S1j). This
explains why, despite a number of attempts, we have failed to ob-
tain a crystal structure of EcHPPK bound to 10. It was previously
predicted that the sulfur atom in 10 would form a van der Waals
interaction with Trp89,20 and our structures of 1, 2, 3, 7, 8 and 9
have not only confirmed this but revealed the potential for a steric
clash. It appears that our inhibitory compounds that contain this
scaffold are only able to bind because the substituents ‘pull’ the
scaffold away from Trp89 and also rotate the side chain, thereby
relieving the steric clash. This in turn suggests that 10 is actually
a suboptimal scaffold that compromises the binding of all the com-
pounds that contain it. Given this limitation, we decided to search
for alternative scaffolds for optimization.
of compound between 0 and 500 lM (see Section 5 for details).
The SPR data and the IC50 measurements are shown in Figures S1
and S5, respectively. Overall, small substitutions at the para and
meta positions (3, 4, 5, 7 and 8) are preferred over larger para sub-
stitutions (6) and ortho substitutions (9). To investigate this pat-
tern more systematically, the crystal structures of the ternary
complexes with 3, 7, 8 and 9 were determined (Tables 2, S1 and
S2).
All four ternary complex structures are virtually identical to
that of 1, and the only significant differences occur in the immedi-
ate vicinity of the phenyl ring (Fig. 2b–e). In the 3 and 7 ternary
complexes (Fig. 2b and c), the para-substituents extend directly
away from the pocket, and this appears to ‘push’ the phenyl group
into the interior of the pocket by ꢀ1 Å compared to 1. This results
in a more favorable stacking interaction between the planar car-
bonyl and the guanidinium group of Arg121, and may explain
why the para substitutions generally display superior activities.
The 8 structure (Fig. 2d) shows that the meta-substituent has little
effect on the docking conformation, and is able to form a favorable
van der Waals interaction with Leu45. Finally, the 9 structure
(Fig. 2e) reveals that the ortho-methyl group is directly adjacent
to Leu45 which is forced to rotate further away compared to the
We first examined 6-thioguanine (18) in which the sulfur atom
is moved to a position on the scaffold that would not interact with
Trp89. Although solubility issues prevented the measurement of a
meaningful IC50 value, 18 is apparently a far superior inhibitory
scaffold than 8-thioguanine based on percent inhibition (Table 3).
This was confirmed by SPR that revealed a low lM KD value (Table 3)
and showed that it binds only in the context of the ternary complex
with AMPCPP (Fig. S6a). The crystal structure of the ternary com-
plex (Tables 2 and S2, Figures 3A and S3c) shows that 18 engages
the pterin pocket in the same fashion as the 8-thioguanine scaffold,
but there is no steric clash with Trp89 and the sulfur atom occupies
available space adjacent to Asn55. 18 is a low molecular weight effi-
cient HPPK binder that can be considered a fragment ‘hit’, and its li-
gand efficiency can be monitored using the ‘Percent Efficiency
Figure 2. Crystal structures of HPPK ternary complexes with compounds related to 1 (2, 3, 7, 8 and 9). In panels a–e, the 1 structure (grey carbons) is superimposed on the
compound structure (yellow carbons), and only three key residues are shown for clarity. Panel f shows the 3 complex in a surface representation to emphasize the
engagement of the 8-thio substituent within the cryptic pocket created by loop2 (cyan) and loop3 (magenta).

M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
2161
Table 3
Fragments
solubility did not allow the measurement of their IC50 values. 19
and 20 both resemble 8-thioguanine, but with the sulfur atom at
the 8-position replaced with a hydrogen atom and a methyl group,
respectively, and their improved potencies and abilities to bind
EcHPPK in a ternary complex support our notion that the bulky sul-
fur atom in 8-thioguanine leads to a steric clash with Trp89 when
loop3 becomes ordered in the presence of nucleotide.
Compound
Structure
% Inhibitiona
51
KD
(
l
M)b
PEIc
S
N
HN
18
19
20
21
18
1
3
1
3.05
N
H
H2N
H2N
H2N
N
O
Compound 21 is distinct from any of the other scaffolds, and we
determined the structure of its ternary complex (Tables 2 and S2,
Figs. 3c and S3d) to establish its mode of binding. It binds within
the pterin pocket, as anticipated, but exploits the available hydro-
gen bonding architecture in a different way compared to the other
scaffolds. The sulfur atom engages the same space as that of 18
adjacent to Asn55, but the side chains of Leu45 and Trp89 adopt
different conformations to accommodate the scaffold. Most nota-
bly, Trp89 rotates by almost ꢀ120° and the ring nitrogen atom
N
HN
HN
48
49
48
38
18
3.18
2.97
2.87
N
H
N
O
N
N
H
N
S
no longer forms a hydrogen bond with the
c-phosphate oxygen
HN
N
7.5 0.2
atoms of AMPCPP. This rotation of Trp89 appears to be caused by
two water molecules (Wat1 and Wat2 in Fig. 3c) that occupy extra
available space in the 21 complex and create a new hydrogen
bonding constellation involving the c-phosphate oxygen atoms of
AMPCPP and the carbonyl oxygen atoms of Arg88 and Gly90.
H2
N
NH
N
a
% Inhibition of enzyme activity was tested at 250
l
M of the compound.
KD value was determined by SPR with equilibrium affinity analysis in the
M AMPCPP.
Percent Efficiency Index (PEI) was calculated as the percent inhibition
(expressed in fraction) at 250 M of the compound divided by the molecular weight
b
presence of 2
l
c
3. Discussion
l
(expressed in kDa) of the inhibitor compound. The PEI value of compound 3 is 2.54.
The bacterial enzymes HPPK and DHPS catalyze adjacent steps
in the folate biosynthetic pathway, and they each have very spe-
cific pockets that engage their pterin substrates. Although the
two pockets have different architectures and have apparently
evolved independently, our recent demonstration that a small mol-
ecule can bind both pockets4 suggested that pterin-like scaffolds
can be used in parallel to seek potent inhibitors of both enzymes.
The successful identification of two HPPK inhibitors from our li-
brary of molecules designed to engage the DHPS pterin-binding
pocket supports this proposal. The two compounds are structurally
related and contain an 8-thioguanine scaffold with two different
groups at the 8-thio position. A preliminary SAR analysis of the
8-thio substituent indicated that this series is not suitable for fur-
ther development, and we showed that the problem is the sulfur
atom that hinders the binding of 8-thioguanine within the HPPK
pterin pocket. We subsequently identified tighter binding frag-
ment molecules and confirmed by X-ray crystallography that they
can specifically engage the pterin pocket without the steric issues
experienced by 8-thioguanine.
Index’ (PEI)28 metric. This value for 18 (Table 3) confirms the suit-
ability of the molecule for future elaboration.
Next, applying best practices from other fragment-based drug
discovery programs,29,30 we decided to optimize the 18 scaffold
to identify the best starting templates for further elaboration. We
did this by conducting a 2D similarity search for new fragments
with high Tanimoto similarity to 18 against the NCI database of
compounds and applying a molecular weight cutoff of 300 Daltons.
25 fragments were shortlisted, and 17 were able to be obtained
from the NCI and verified to be >95% pure using HPLC analysis.
Docking simulations using AutoDock showed that they all have
the potential to bind within the pterin-binding pocket in a similar
manner to 18 (Fig. 3b). When tested, three of these fragments, 19,
20 and 21 (Table 3), showed inhibitory activities around 50% with
PEI values similar to 18. All three are dihydropurine fragments that
also showed AMPCPP-dependent binding by SPR with KD values
comparable to 18 (Table 3 and Figs. S6b-d). Similar to 18, limited
Figure 3. Crystal structures of HPPK ternary complexes with two sulfur-containing inhibitory scaffolds. (a) Compound 18 (6-thioguanine). (b) The chemical fingerprint of 18
in the HPPK pterin-binding pocket in the same orientation as (a). (c) Stereo view of compound 21 in the HPPK ternary complex. Note that the side chains of Leu45 and Trp89
are in different orientations compared to the other reported structures.

2162
M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
Independent studies have previously identified 8-thioguanine
We believe that the structural and biochemical data presented in
this study have provided an excellent basis to enable such efforts
to move forward rapidly.
as an HPPK inhibitor suitable for development,20 and preliminary
attempts to do this were recently reported.23 These studies, using
HPPK from S. aureus, showed that 8-thioguanine can bind in the
absence of nucleotide, but our SPR data show that binding is min-
imal both in the absence and presence of nucleotide. In the pub-
lished structure, 8-thioguanine binds the pterin pocket in the
context of a non-physiological crystal dimer with the small mole-
cule at the dimer interface. In addition, the flexible loop3 contain-
ing Trp89 was displaced and it was not possible to crystallize the
ternary complex with AMPCPP. Our structures have confirmed
the prediction20 that the sulfur atom is directly adjacent to Trp89
in the ternary complex. However, the binding of our 8-thioguanine
derivatives only appears possible because the 8-position substitu-
ents create the necessary space.
4. Conclusions
In summation, the flexibility and druggability of the pterin bind-
ing pocket of HPPK has been structurally and biochemically investi-
gated through detailed enzymatic inhibition assays, X-ray
crystallography, surface plasmon resonance and computational
modeling. These techniques were used to screen a set of compounds
originally procured to target the pterin binding site of DHPS, the
subsequent enzyme in the folate biosynthetic cascade, which iden-
tified 8-substituted-8-thioguanines as potential inhibitors of HPPK.
Key to this discovery was that 8-thiosubstitution induced a new
cryptic binding pocket adjacent to the pterin binding site, and fur-
ther analogs were developed to more optimally engage this pocket.
Free 8-thioguanine was shown to be a suboptimal pterin pocket
binding core, and a series of dihydropurine fragments was identified
that should yield more efficient enzyme inhibitors. The structural
activity relationships derived from these various molecules, when
merged, will facilitate the future design of potent HPPK inhibitors
with the potential of being developed into novel antimicrobial ther-
apeutics suitable to combat the rising threats of drug resistant
bacteria.
Our future goals will be to use these preliminary data to devel-
op more potent inhibitors of HPPK with whole cell activity using
structure-based approaches. Our current results have suggested a
number of avenues to achieve this that will be pursued. First, we
have identified several novel and quite potent pterin pocket
binding scaffolds, and each of these will be elaborated into mini li-
braries for further screening. Second, our analyses of the 8-thio-
guanine derivatives have provided key information on how these
scaffolds can be elaborated to take advantage of a cryptic pocket
adjacent to the active site that is induced upon binding. This pocket
(Fig. 2f) has in fact been previously identified from earlier struc-
tural studies of an HPPK inhibitory molecule that contains the pter-
in scaffold.6 Finally, the nucleoside and triphosphate binding sites
are potentially rich unexploited locales for further optimization of
our evolving molecules. The build-out of our inhibitors into these
adjacent sites offers the possibility of generating tight-binding
nucleotide-independent inhibitors. Similar strategies involving
the synthesis of bisubstrate analogs (pterin-nucleotide adducts)
have already demonstrated the feasibility of this approach for this
target.7,21,22 We recognize that these nucleotide binding sites are
inherently polar and that the build-out process must eventually
incorporate strategies for optimizing membrane permeability and
cell penetration.
A particular challenge in developing potent inhibitors of HPPK is
the functional flexibility of the active site loops and key residues.
Our crystal structures have confirmed the inherent plasticity of
the HPPK active site locale that even extends to the pterin pocket
(Figs. S7 and S8), and inhibitor binding is clearly associated with
a considerable entropic penalty as the flexible loops become or-
dered. This presumably explains why all of our inhibitors only dis-
play modest KD values. However, this provides two advantages that
we intend to exploit. First, our low molecular weight inhibitors are
essentially fragment molecules, and it has been well established
that plasticity within the fragment binding locale can be used to
great effect to ‘evolve’ tight binding derivative molecules.29,30 Sec-
ond, the unavoidable entropic penalty will dictate that enthalpic
binding must be optimized in derivative molecules, and this should
lead to tight binding molecules with slow off rates.
5. Experimental
5.1. Enzyme preparation
The EcHPPK-GST fusion gene was kindly provided by Dr. Hon-
ggao Yan (Michigan State University). The EcHPPK-GST fusion en-
zyme was expressed in E.coli and purified as previously described
in three steps: (1) GSTrap FF column (Amersham Biosciences), (2)
thrombin digestion to remove the GST tag, and (3) size exclusion
chromatography.27,33
5.2. Enzyme assay
HPPK activity was determined by measuring unprocessed ATP
substrate using the Kinase-Glo Luminescent assay kit (Promega).20
The enzyme activity of HPPK was measured in a reaction mixture
of 50
l
l containing 0.75
l
M 6-hydroxymethyl-7,8-dihydropterin
M ATP, 100 mM Tris–
hydrochloride (Schircks Laboratories), 0.75
l
HCl, pH 8.5, 10 mM MgCl2, 10 mM BME, 0.01%(V/V) Tween 20,
0.01%(W/V) BSA, 5% DMSO, and 5 ng EcHPPK. Inhibitor compounds
were dissolved in DMSO, and inhibition was tested at 250 lM. After
20 min of incubation at 23 °C, the unprocessed ATP was detected
using the Kinase-Glo assay kit with a Multi-Detection Microplate
Reader Synergy™ HT(BioTek) according to manufacturer’s sug-
gested protocols. The above assay conditions were established after
a preliminary optimization procedure to find an ideal compromise
between the luminescence signal, the linearity of the reaction, and
the reduction in signal caused by the inhibitors. To determine the
half maximal inhibitory concentration (IC50) values, HPPK activities
were measured in the presence of increasing concentrations of the
Antimicrobial antifolate therapeutics that contain a combina-
tion of sulfa drugs (DHPS) and dihydrofolate inhibitors have proven
to be very effective,31 and HPPK/DHPS pterin-based inhibitors offer
the possibility of alternate combination therapies. The high conser-
vation of the HPPK and DHPS pterin pockets in all micro-organisms
should mean that these therapies have broad spectrum potential
that are less prone to resistance mutations. These agents should
also be suitable for use in patients with sulfa-allergies and in
patient populations with high incidence of sulfa allergies such as
those with HIV undergoing prophylaxis to prevent pneumocystis
pneumonia.32 To obtain such therapeutics, it is clear that these
compounds must be further optimized to enhance target affinity
and microbial penetration, and to maximize exposure in the host.
compounds ranging from 0 to 500 lM in duplicate. IC50 values were
calculated from the dose-response curves using GraphPad Prism
(version 5.00).
5.3. Biotinylation of HPPK
HPPK was minimally biotinylated by reaction with EZ-Link Sul-
fo-NHS-LC-LC-Biotin (Thermo Scientific). The biotin reagent was
added to the protein at a 0.5:1 molar ratio, and the reaction was

M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
2163
incubated on ice for 5 h. Unconjugated biotin was removed by pro-
cessing the samples through two Zeba Spin Desalting Columns
(Thermo Scientific) that had been equilibrated with storage buffer
(20 mM Tris (pH 8.0), 100 mM NaCl, 10% glycerol). Bovine serum
albumin (BSA) was added to the reaction at a final concentration
of 0.1 mg/mL immediately prior to processing through the spin col-
umns to improve recovery.34 The biotinylated protein was aliquot-
ted, flash-frozen and stored at ꢂ80 °C for use in subsequent
binding experiments.
optimized using PHENIX38 and COOT,39 respectively. Data collec-
tion and refinement statistics are summarized in Table 2, and are
fully presented in Supplementary Tables S1 and S2 together with
the PDB accession codes.
5.6. Fragment optimization
Based on the hit fragment 6-thioguanine (18), a 2D chemical
similarity search was conducted against the NCI database
(265,242 compounds) for fragment-like molecules (Molecular
weight cutoff: 300 Da) yielding 25 compounds. The resulting hits
were docked into the pterin site receptor derived from EcHPPK/6-
thioguanine structure using AutoDock 3.05 to find their probable
binding modes and binding energies. Out of the 25 compounds, only
17 were actually available at the NCI; these were acquired from the
NCI (see below) and their activities were tested against EcHPPK.
5.4. Affinity analysis of compound binding by surface plasmon
resonance
SPR experiments were conducted at 20 °C using a SensiQ Pioneer
optical biosensor (SensiQ Technologies). Neutravidin (Thermo Sci-
entific) was covalently immobilized on a polysaccharide hydrogel-
coated gold surface (COOH5 chip; SensiQ Technologies) using rou-
tine amine coupling chemistry in immobilization buffer (10 mM
HEPES pH 7.4, 150 mM NaCl, 0.005% Tween20). Carboxyl groups
on the hydrogel were activated with N-ethyl-N0-(3-dimethylamino-
propyl) carbodiimide (EDC) and N-hydroxysuccinimide (NHS), and
neutravidin was injected until immobilization levels of ꢀ5000–
7000 RU were achieved. Remaining active sites were blocked by
reaction with ethanolamine. The instrument was primed with bind-
ing buffer (20 mM Tris pH 8.5, 100 mM NaCl, 10 mM MgCl2, 1 mM
TCEP, 0.005% Tween20, 10% glycerol, 5% DMSO), and biotinylated
HPPK was injected until ꢀ3000–3700 RU of protein was captured.
Unoccupied biotin-binding sites of neutravidin on both the refer-
ence and HPPK surfaces were blocked with amine-PEG2-biotin
(Thermo Scientific) to minimize potential non-specific binding by
the compounds.35 For the binding of compounds 5, 6, 7, 8, 10 and
20 to HPPK in the absence of AMPCPP, the data were collected in
5.7. Compound procurement
Compounds 1, 2, 10 and 18–21 were either obtained from com-
mercial vendors or compound repositories. The details are as fol-
lows: 1, 2, Scientific Exchange, Inc.; 10, Ryan Scientific, Inc.; 18,
Sigma–Aldrich Co.; 19–21, Drug Synthesis and Chemistry Branch,
Developmental Therapeutics Program, Division of Cancer Treat-
ment and Diagnosis, NCI.40 Compounds 3–9 and 11–17 were syn-
thesized as described below.
5.8. Chemistry
All chemicals and solvents were purchased from Sigma–Aldrich
chemical company (St Louis, MO) and used as received. All chem-
ical reactions were tracked by TLC using Silicycle Silica Gel 60F254
plates and spots were visualized by UV lamp or I2 condensation. 1H
NMR spectra were recorded on a 400 MHz Bruker NMR and chem-
ical shifts were reported relative to solvent peak. Analytical RP-
screening mode (i.e. a single injection at 100
experiments, the compounds were prepared in running buffer as a
3-fold dilution series starting at 100 M and were injected at a flow
rate of 80 L/min. A series of buffer-only (blank) injections was in-
lM). For all other
l
l
HPLC was determined on both
equipped with an Acquity BEH C18 column (1.7
0.5 mL/min and a gradient of solvent A (water with 0.1% formic
acid) and solvent (acetonitrile with 0.1% formic acid):
a
Waters Acquity UPLC–MS
cluded throughout the experiment to account for instrumental
noise. The data were processed, double-referenced, solvent cor-
rected and analyzed35,36 using the software package Qdat (version
2.1.0.21, BioLogic Software). The equilibrium dissociation constants
were determined by fitting the data to a 1:1 interaction model.
l
m), flow rate of
B
0–0.25 min 97% A; 0.25–3.0 min 3–100% B (linear gradient);
3.0–4.5 min 100% B; 4.5–4.75 min 0–97% A (linear gradient);
4.75–5.0 min 97% A. Purity was monitored by both UV absorbance
(monitored at 225–475 nm) and ELSD detection methods. All com-
pounds were found to have >95% purity with the described analyt-
ical methods. High resolution mass spectra were determined using
a Waters Xevo G2 QTof.
5.5. Crystallographic analyses
Crystals were grown by the vapor diffusion method at 18 °C.
The protein solution (6.6 mg/ml in 20 mM Tris–HCl, pH 8.0, and
0.1 M NaCl) was incubated with 4 mM AMPCPP (a,b-methylenead-
enosine 50-triphosphate lithium salt, Sigma), 20 mM MgCl2, and
1ꢀ2 mM compound (final 4% DMSO) to form a complex. An addi-
tional 5 mM TCEP (Tris(2-carboxyethyl)phosphine hydrochloride)
was added to the complex solutions of compounds 2 and 18. Initial
crystals were identified robotically using the PEGs suite, PEGs II
suite, and JCSG+ suite crystallization screening kits (Qiagen), which
were manually optimized. The co-crystallization conditions for
compounds 1, 3, 7, 8 and 9 are 0.1 M Tris–HCl, pH 8.5, 0.2 M CaCl2,
and 25% PEG 4 K, in a 1:1 ratio with the protein complex solution.
The crystals of compounds 2 and 18 were obtained with the reser-
voir solution containing 0.1 M Hepes-NaOH, pH 7.5, 0.2 M CaCl2,
and 25–30 % PEG 4 K. The crystal of compound 21 was grown with
the reservoir solution that contained 0.1 M sodium cacodylate, pH
6.5, and 1 M tri-sodium citrate. Crystals were cryoprotected with
30% glycerol, and then flash frozen in liquid nitrogen. All diffraction
data were collected at the SERCAT beam lines 22-ID and 22-BM at
the Advanced Photon Source, and processed using HKL2000.37
Structures were solved by molecular replacement using the EcH-
PPK structure as the search model. Structures were refined and
5.8.1. General procedure for the preparation of compounds 3–9
and 12–17
In a round bottom flask equipped with a stir bar was dissolved
2-amino-6-hydroxy-8-mercaptopurine (1–2 mmol, 1.0 equiv) in
aqueous sodium hydroxide (0.4 N, 10 mL) and corresponding
phenyl substituted 2-bromoacetophenones (1.25–2.5 mmol,
1.25 equiv) dissolved in EtOH (2 mL) was added. The mixtures
were stirred for 1 h at room temperature before the solution was
neutralized by drop wise addition of HCl (1 N) to yield
a
precipitate. The precipitate was collected via vacuum filtration,
washed with diethyl ether and verified to be pure via LC–MS and
1H NMR.
5.8.1.1. 2-Amino-8-((2-oxo-2-(p-tolyl)ethyl)thio)-1H-purin-6(7H)-
one: (3).
Yield: 769.5 mg (98%) of off white powder 1H NMR
(400 MHz, DMSO-d6) d 2.38 (s, 3H), 2.55 (s, 4H), 4.67 (s, 2H), 6.05
(s, 2H), 7.32 (d, J = 7.94 Hz, 2H), 7.87–8.00 (m, 2H), 10.57 (s, 1H).
ESI-HRMS: m/z: calcd: 316.0824 found: 316.0890.

2164
M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
5.8.1.2. 2-Amino-8-((2-(4-hydroxyphenyl)-2-oxoethyl)thio)-1H-
12.49 (d, J = 31.18 Hz, 1H). ESI-HRMS: m/z: calcd: 330.0980 found:
purin-6(7H)-one: (4).
Yield: 188.2 mg (59.3%) of off white
330.1039.
powder. 1H NMR (400 MHz, DMSO-d6) d 4.71 (s, 2H), 6.36 (s, 2H),
6.70–6.79 (m, 2H), 7.79–7.88 (m, 2H). ESI-HRMS: m/z: calcd:
318.0616 found: 318.0673.
5.8.1.12. S-(2-Amino-6-oxo-6,7-dihydro-1H-purin-8-yl) 4-fluor-
obenzothioate: (16).
Yield: 955.0 mg (88%) of white powder.
1H NMR (400 MHz, DMSO-d6) d 6.45–6.61 (m, 2H), 7.20–7.46 (m,
2H), 7.89–8.07 (m, 2H), 10.73 (s, 1H), 12.97 (s, 1H). ESI-MS:
[M+H]+ Found: 306.2.
5.8.1.3. 2-Amino-8-((2-(4-fluorophenyl)-2-oxoethyl)thio)-1H-
purin-6(7H)-one: (5).
Yield: 416.8 (52.2%) of red solid 1H
NMR (400 MHz, DMSO-d6) d 4.75 (s, 2H), 6.41 (s, 2H), 7.27–7.34
(m, 2H), 8.05 (dd, J = 5.63, 8.71 Hz, 2H). ESI-HRMS: m/z: calcd:
320.0573 found: 320.0627.
5.8.1.13. 2-Amino-8-((2-morpholinoethyl)thio)-1H-purin-6(7H)
-one: (17).
Yield: 335.6 mg (45.3%) of white powder. 1H NMR
(400 MHz, DMSO-d6) d 2.35–2.47 (m, 4H), 2.61 (t, J = 6.88 Hz, 2H),
3.56 (t, J = 4.62 Hz, 4H), 6.32 (s, 2H), 10.60 (s, 1H), 12.53 (s, 1H).
ESI-HRMS: m/z: calcd: 297.1089 found: 297.1153.
5.8.1.4. 4-(2-((2-Amino-6-oxo-6,7-dihydro-1H-purin-8-yl)thio)
acetyl)benzonitrile: (6).
Yield: 260.2 mg (80%) of brown
powder. 1H NMR (400 MHz, DMSO-d6) d 4.89 (s, 2H), 6.29 (s, 2H),
8.03–8.06 (m, 2H), 8.15–8.19 (m, 2H), 10.60 (d, J = 27.50 Hz, 2H),
12.35–12.64 (m, 1H). ESI-HRMS: m/z: calcd: 327.0619 found:
327.0682.
5.8.2. General procedure for the preparation of compound 11
To a stirred solution of 16 (1 mmol, 1.0 equiv) in aqueous so-
dium hydroxide (0.4 N, 10 mL) was added 4-(2-chloroethyl)mor-
pholine (1.25-2.5 mmol, 1.25 equiv) and the mixture stirred for
overnight at room temperature. The solution was then neutralized
by drop wise addition of HCl (1 N) to yield a precipitate, which was
collected via vacuum filtration, washed with diethyl ether and ver-
ified to be pure via LC–MS and 1H NMR.
5.8.1.5.
2-Amino-8-((2-(4-methoxyphenyl)-2-oxoethyl)thio)-
1H-purin-6(7H)-one: (7).
Yield: 278.5 mg (42%) of white
powder. 1H NMR (400 MHz, DMSO-d6) d 3.87 (s, 3H), 4.82 (d,
J = 19.12 Hz, 2H), 6.32 (s, 1H), 6.96–7.27 (m, 2H), 8.00 (dd,
J = 6.36, 8.71 Hz, 2H), 10.48 (s, 1H), 12.53 (s, 1H). ESI-HRMS: m/z:
calcd: 332.0773 found: 332.0806.
5.8.2.1. 2-Amino-8-((2-(4-(2-morpholinoethoxy)phenyl)-2-oxo-
ethyl)thio)-1H-purin-6(7H)-one: (11).
Yield: 32.6 mg (7.6%)
2.42 (d,
5.8.1.6.
6(7H)-one: (8).
2-Amino-8-((2-oxo-2-(m-tolyl)ethyl)thio)-1H-purin-
Yield: 367.3 mg (46.6%) of red solid. 1H NMR
of brown powder. 1H NMR (400 MHz, DMSO-d6)
d
J = 5.07 Hz, 4H), 2.60 (dt, J = 6.42, 30.99 Hz, 2H), 3.53 (t,
J = 4.51 Hz, 4H), 4.01 (t, J = 6.27 Hz, 1H), 4.25 (t, J = 6.30 Hz, 1H),
4.70–4.90 (m, 2H), 6.13 (s, 1H), 6.86 (dd, J = 2.73, 8.78 Hz, 2H),
7.89 (dd, J = 4.31, 7.68 Hz, 2H), 10.76 (s, 1H). ESI-HRMS: m/z: calcd:
431.1457 found: 431.1518.
(400 MHz, DMSO-d6) d 2.39 (s, 3H), 4.87 (s, 1H), 6.28–6.39 (m,
1H), 7.41–7.52 (m, 2H), 7.81–7.94 (m, 2H), 10.57 (s, 1H). ESI-
HRMS: m/z: calcd: 316.0824 found: 316.0890.
5.8.1.7.
6(7H)-one: (9).
2-Amino-8-((2-oxo-2-(o-tolyl)ethyl)thio)-1H-purin-
Yield: 233.2 mg (35.2%) of tan powder. 1H
Acknowledgments
NMR (400 MHz, DMSO-d6) d 2.37 (s, 3H), 4.71 (s, 2H), 6.33 (s,
1H), 7.28–7.38 (m, 2H), 7.46 (td, J = 1.38, 7.57 Hz, 1H), 7.91 (d,
J = 7.60 Hz, 1H), 10.55 (d, J = 42.42 Hz, 1H), 12.69 (d,
J = 132.88 Hz, 1H). ESI-HRMS: m/z: calcd: 316.0824 found:
316.0416.
This work was supported by National Institutes of Health Grant
AI070721 (S.W.W. and R.E.L.), Cancer Center (CORE) Support Grant
CA21765, and the American Lebanese Syrian Associated Charities
(ALSAC). We thank Zhenmei Li, John Bollinger and the St. Jude
Molecular Interactions Core for technical assistance, and Dr.Hon-
ggao Yan (Michigan State University) for reagents. Data were
collected at Southeast Regional Collaborative Access Team (SER-
CAT) 22-ID and 22-BM beamlines at the Advanced Photon Source,
Argonne National Laboratory, and we thank SER-CAT staff for their
assistance. Supporting SER-CAT institutions may be found at
www.ser-cat.org/members.html. Use of the Advanced Photon
Source was supported by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences, under Contract No. W-
31-109-Eng-38.
5.8.1.8. 2-Amino-8-((2-morpholino-2-oxoethyl)thio)-1H-purin-
6(7H)-one: (12).
Yield: 166.8 mg (43%) of white powder. 1H
NMR (400 MHz, DMSO-d6) d 3.45 (d, J = 4.97 Hz, 3H), 3.49–3.59
(m, 6H), 3.61 (d, J = 4.71 Hz, 3H), 4.22 (s, 3H), 6.32 (s, 2H), 10.52
(s, 1H), 12.52 (s, 1H). ESI-HRMS: m/z: calcd: 311.0882 found:
311.0907.
5.8.1.9.
2-Amino-8-((2-(2-methoxyphenyl)-2-oxoethyl)thio)-
1H-purin-6(7H)-one: (13).
Yield: 431.7 mg (65.1%) of off-
white powder 1H NMR (400 MHz, DMSO-d6) d 3.93 (s, 3H), 4.68
(s, 2H), 6.28 (s, 2H), 6.89–7.39 (m, 2H), 7.47–7.86 (m, 2H), 10.57
(s, 1H), 12.58 (s, 1H). ESI-HRMS: m/z: calcd: 332.0773 found:
332.0806.
A. Supplementary data
Supplementary data (X-ray crystallography data collection and
refinement statistics, weakly binding compounds, surface plasmon
resonance data, selected electron densities, confirmation of Ca2+
ions in the HPPK two-metal ATP-binding site, IC50 measurements
and plasticity within in the HPPK pterin-binding pocket) associated
with this article can be found, in the online version, at http://
dx.doi.org/10.1016/j.bmc.2014.02.022.
5.8.1.10.
purin-6(7H)-one: (14).
2-Amino-8-((1-oxo-1-phenylpropan-2-yl)thio)-1H-
Yield: 194.4 mg (30.8%) of yellow
powder. 1H NMR (400 MHz, DMSO-d6) d 1.54 (d, J = 6.84 Hz, 3H),
5.44 (q, J = 6.89 Hz, 1H), 6.35 (s, 2H), 7.52 (t, J = 7.74 Hz, 2H),
7.60–7.70 (m, 1H), 7.96–8.06 (m, 2H), 10.68 (s, 1H), 12.66 (s, 1H).
ESI-HRMS: m/z: calcd: 316.0824 found: 316.0890.
5.8.1.11.
purin-6(7H)-one: (15).
powder. 1H NMR (400 MHz, DMSO-d6)
7.01 Hz, 3H), 2.38 (d, J = 2.05 Hz, 3H), 5.43 (s, 1H), 6.34 (s, 2H),
7.27–7.43 (m, 2H), 7.91 (dd, J = 2.04, 8.33 Hz, 2H), 10.60 (s, 1H),
2-Amino-8-((1-oxo-1-(p-tolyl)propan-2-yl)thio)-1H-
Yield: 217.4 mg (26.4%) of yellow
1.53 (dd, J = 2.07,
References and notes
d
1. Blakley, R. L. In Folates and Pterins; John Wiley and Sons: New York, 1984; Vol.
1, pp 628.
2. Bermingham, A.; Derrick, J. P. BioEssays 2002, 24, 637.

M.-K. Yun et al. / Bioorg. Med. Chem. 22 (2014) 2157–2165
2165
3. Anand, N. In Burger’s Medicinal Chemistry and Drug Discovery; Wiley-
Interscience: New York, 1996; Vol. 2, pp 527–573.
4. Pemble, C. W. t.; Mehta, P. K.; Mehra, S.; Li, Z.; Nourse, A.; Lee, R. E.; White, S. W.
PLoS One 2010, 5, e14165.
5. 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. J. Mol. Biol. 1999, 287, 211.
6. Stammers, D. K.; Achari, A.; Somers, D. O.; Bryant, P. K.; Rosemond, J.; Scott, D.
L.; Champness, J. N. FEBS Lett. 1999, 456, 49.
7. Shi, G.; Blaszczyk, J.; Ji, X.; Yan, H. J. Med. Chem. 2001, 44, 1364.
8. Xiao, B.; Shi, G.; Chen, X.; Yan, H.; Ji, X. Structure 1999, 7, 489.
9. Lawrence, M. C.; Iliades, P.; Fernley, R. T.; Berglez, J.; Pilling, P. A.; Macreadie, I.
G. J. Mol. Biol. 2005, 348, 655.
10. Garcon, A.; Levy, C.; Derrick, J. P. J. Mol. Biol. 2006, 360, 644.
11. Blaszczyk, J.; Li, Y.; Cherry, S.; Alexandratos, J.; Wu, Y.; Shaw, G.; Tropea, J. E.;
Waugh, D. S.; Yan, H.; Ji, X. Acta Crystallogr. D Biol. Crystallogr. 2007, 63, 1169.
12. Bermingham, A.; Bottomley, J. R.; Primrose, W. U.; Derrick, J. P. J. Biol. Chem.
2000, 275, 17962.
22. Shi, G.; Shaw, G.; Li, Y.; Wu, Y.; Yan, H.; Ji, X. Bioorg. Med. Chem. 2012, 20, 4303.
23. Chhabra, S.; Barlow, N.; Dolezal, O.; Hattarki, M. K.; Newman, J.; Peat, T. S.;
Graham, B.; Swarbrick, J. D. PLoS One 2013, 8, e59535.
24. Yun, M. K.; Wu, Y.; Li, Z.; Zhao, Y.; Waddell, M. B.; Ferreira, A. M.; Lee, R. E.;
Bashford, D.; White, S. W. Science 2012, 335, 1110.
25. Xiao, B.; Shi, G.; Gao, J.; Blaszczyk, J.; Liu, Q.; Ji, X.; Yan, H. J. Biol. Chem. 2001,
276, 40274.
26. Li, G.; Felczak, K.; Shi, G.; Yan, H. Biochemistry 2006, 45, 12573.
27. Li, Y.; Wu, Y.; Blaszczyk, J.; Ji, X.; Yan, H. Biochemistry 2003, 42, 1581.
28. Abad-Zapatero, C.; Metz, J. T. Drug Discovery Today 2005, 10, 464.
29. Hajduk, P. J.; Greer, J. Nat. Rev. Drug Disc. 2007, 6, 211.
30. Congreve, M.; Chessari, G.; Tisi, D.; Woodhead, A. J. J. Med. Chem. 2008, 51,
3661.
31. Masters, P. A.; O’Bryan, T. A.; Zurlo, J.; Miller, D. Q.; Joshi, N. Arch. Intern. Med.
2003, 163, 402.
32. Merali, S.; Zhang, Y.; Sloan, D.; Meshnick, S. Antimicrob. Agents Chemother.
1990, 34, 1075.
13. Blaszczyk, J.; Shi, G.; Yan, H.; Ji, X. Structure 2000, 8, 1049.
14. Shi, G.; Gong, Y.; Savchenko, A.; Zeikus, J. G.; Xiao, B.; Ji, X.; Yan, H. Biochim.
Biophys. Acta 2000, 1478, 289.
15. Li, Y.; Gong, Y.; Shi, G.; Blaszczyk, J.; Ji, X.; Yan, H. Biochemistry 2002, 41, 8777.
16. Blaszczyk, J.; Li, Y.; Shi, G.; Yan, H.; Ji, X. Biochemistry 2003, 42, 1573.
17. Blaszczyk, J.; Li, Y.; Wu, Y.; Shi, G.; Ji, X.; Yan, H. Biochemistry 2004, 43, 1469.
18. Blaszczyk, J.; Shi, G.; Li, Y.; Yan, H.; Ji, X. Structure 2004, 12, 467.
19. Swarbrick, J.; Iliades, P.; Simpson, J. S.; Macreadie, I. Open Enzyme Inhib. J. 2008,
1, 12.
33. Hevener, K. E.; Yun, M. K.; Qi, J.; Kerr, I. D.; Babaoglu, K.; Hurdle, J. G.;
Balakrishna, K.; White, S. W.; Lee, R. E. J. Med. Chem. 2010, 53, 166.
34. Papalia, G.; Myszka, D. Anal. Biochem. 2010, 403, 30.
35. Giannetti, A. M. Methods Enzymol. 2011, 493, 169.
36. Myszka, D. G. J. Mol. Recognit. 1999, 12, 279.
37. Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
38. Adams, P. D.; Grosse-Kunstleve, R. W.; Hung, L. W.; Ioerger, T. R.; McCoy, A. J.;
Moriarty, N. W.; Read, R. J.; Sacchettini, J. C.; Sauter, N. K.; Terwilliger, T. C. Acta
Crystallogr. D Biol. Crystallogr. 2002, 58, 1948.
20. Chhabra, S.; Dolezal, O.; Collins, B. M.; Newman, J.; Simpson, J. S.; Macreadie, I.
G.; Fernley, R.; Peat, T. S.; Swarbrick, J. D. PLoS One 2012, 7, e29444.
21. Shi, G.; Shaw, G.; Liang, Y. H.; Subburaman, P.; Li, Y.; Wu, Y.; Yan, H.; Ji, X.
Bioorg. Med. Chem. 2012, 20, 47.
39. Emsley, P.; Cowtan, K. Acta Crystallogr. D Biol. Crystallogr. 2004, 60, 2126.
40. Hergenrother, P. J. Curr. Opin. Chem. Biol. 2006, 10, 213.