DXR inhibition by potent mono- and disubstituted fosmidomycin analogues

Article abstract of DOI:10.1021/jm4006498

The antimalarial compound fosmidomycin targets DXR, the enzyme that catalyzes the first committed step in the MEP pathway, producing the essential isoprenoid precursors, isopentenyl diphosphate and dimethylallyl diphosphate. The MEP pathway is used by a number of pathogens, including Mycobacterium tuberculosis and apicomplexan parasites, and differs from the classical mevalonate pathway that is essential in humans. Using a structure-based approach, we designed a number of analogues of fosmidomycin, including a series that are substituted in both the Cα and the hydroxamate positions. The latter proved to be a stable framework for the design of inhibitors that extend from the polar and cramped (and so not easily druggable) substrate-binding site and can, for the first time, bridge the substrate and cofactor binding sites. A number of these compounds are more potent than fosmidomycin in terms of killing Plasmodium falciparum in an in vitro assay; the best has an IC₅₀ of 40 nM.

Full text of DOI:10.1021/jm4006498

Article
pubs.acs.org/jmc
DXR Inhibition by Potent Mono- and Disubstituted Fosmidomycin
Analogues
Anna M. Jansson,†,∥ Anna Więckowska,‡,∥ Christofer Bjorkelid,†,∥ Samir Yahiaoui,‡
̈
Sanjeewani Sooriyaarachchi,† Martin Lindh,‡ Terese Bergfors,† Shyamraj Dharavath,‡ Matthieu Desroses,‡
Surisetti Suresh,‡ Mounir Andaloussi,‡ Rautela Nikhil,§ Sharma Sreevalli,§ Bachally R. Srinivasa,§
Mats Larhed,‡ T. Alwyn Jones,† Anders Karlen,‡,* and Sherry L. Mowbray*,†
́
†Department of Cell and Molecular Biology, Uppsala University, Biomedical Center, Box 596, SE-751 24 Uppsala, Sweden
‡Department of Medicinal Chemistry, Uppsala University, Biomedical Center, Box 574, SE-751 23 Uppsala, Sweden
§AstraZeneca India Private Limited, Bellary Road, Hebbal, Bangalore 560024, India
S
* Supporting Information
ABSTRACT: The antimalarial compound fosmidomycin
targets DXR, the enzyme that catalyzes the first committed
step in the MEP pathway, producing the essential isoprenoid
precursors, isopentenyl diphosphate and dimethylallyl diphos-
phate. The MEP pathway is used by a number of pathogens,
including Mycobacterium tuberculosis and apicomplexan para-
sites, and differs from the classical mevalonate pathway that is
essential in humans. Using a structure-based approach, we
designed a number of analogues of fosmidomycin, including a
series that are substituted in both the Cα and the hydroxamate
positions. The latter proved to be a stable framework for the
design of inhibitors that extend from the polar and cramped (and so not easily druggable) substrate-binding site and can, for the
first time, bridge the substrate and cofactor binding sites. A number of these compounds are more potent than fosmidomycin in
terms of killing Plasmodium falciparum in an in vitro assay; the best has an IC50 of 40 nM.
low bioavailability, rapid clearance from the parasite, and
recrudescent infection,12 although the compound has been used
more successfully in combination with clindamycin.13 Failure to
obtain biological activity against mycobacteria appears to result
from poor uptake.11
INTRODUCTION
■
Fosmidomycin (1) and its acetyl derivative, FR-900098 (2), are
natural product antibiotics with activity against a number of
important pathogens.1,2 They work by blocking the pathway for
biosynthesis of isopentenyl diphosphate and dimethylallyl
diphosphate that proceeds via the intermediate 2-C-methyl-D-
erythritol 4-phosphate (MEP). This MEP pathway is used in
Gram-negative and some Gram-positive bacteria as well as in
plant chloroplasts, algae, and apicomplexan protozoa.3 The
enzymes involved are essential to the organisms possessing
them yet completely absent in humans, and they have therefore
received considerable attention as potential targets for
antimicrobial drug discovery. Fosmidomycin and 2 act by
inhibiting the second (and first committed) step in the
pathway,4,5 in which 1-deoxy-D-xylulose 5-phosphate (DXP)
is converted to MEP by the enzyme 1-deoxy-D-xylulose-5-
phosphate reductoisomerase (DXR, also called IspC). A
number of major pathogens are dependent upon the MEP
pathway, although not all are sensitive to fosmidomycin and its
analogues. For example, while both 1 and 2 inhibit the DXRs of
the protozoan Plasmodium falciparum (PfDXR)6 and Myco-
bacterium tuberculosis (MtDXR),7−10 the Plasmodium species
are sensitive to these antibiotics in vitro and in vivo,6 but the
mycobacteria are not.11 The use of fosmidomycin as a single-
drug treatment for P. falciparum malaria has been hampered by
Considerable efforts have been made to improve the efficacy
of fosmidomycin, for example, modifications to the phospho-
nate group,14,15 extensions to the hydroxamic acid group,16 and
changes in the linker connecting them.15,17 Extensive crystallo-
graphic studies on DXRs from different species have provided
the structural framework needed to understand the binding of
these inhibitors10,18−21 as well as their α-aryl-substituted
analogues.7,22,23 Despite the lack of biological activity of
fosmidomycin itself against M. tuberculosis, the well-conserved
nature of the active site of DXR,21 combined with the
antiparasite activity of some DXR inhibitors, led us to believe
that our medicinal chemistry effort based on MtDXR would
produce interesting compounds for use against other pathogens
where the MEP pathway is essential. It also seemed possible
that some fosmidomycin analogues would in fact be active
against M. tuberculosis. As part of this same-target-other-
pathogen (STOP) strategy, we determined the structure of
Received: May 2, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
MtDXR as an apoenzyme,9 and together with fosmidomycin10
and 221 in ternary complexes with NADPH, as well as
structures with two 3,4-dichlorophenyl-substituted fosmidomy-
cin analogues (3 and 4, Figure 1).7 As expected, these Cα-
binding pocket further with new analogues. Crystallographic
studies on a select group of enzyme−inhibitor complexes were
used to guide our design efforts.
Design, Synthesis, and Biochemical Evaluation of
Hydroxamate-Substituted Analogues of Fosmidomycin.
The synthetic route to the compounds shown in Table 1 was
Table 1. Inhibition Constants for Fosmidomycin Analogues
(Literature Values Are Included for Comparative Purposes)
Figure 1. Structures of fosmidomycin (1), 2, and the 3,4-
dichlorophenyl-substituted fosmidomycin analogues (3 and 4).
substituted dichlorophenyl compounds bound to the substrate
site, coordinating the metal ion in the same way as
fosmidomycin, but they otherwise showed an unexpected
mode of binding that forces disorder in an active-site flap. To
prove that this mode of binding was not a crystallographic
artifact, we determined the structure of MtDXR with one of
these analogues in a new crystal form; it showed an essentially
identical mode of binding.7 Our expression and crystallization
conditions were used elsewhere to determine the structure with
one of these analogues, confirming our conclusions.22 This class
of compounds had previously been reported to have high
activity against P. falciparum growth in a blood assay,24
although the most promising was later found to be inactive
in an in vivo mouse model.25 We conclude that novel
compounds need to be explored and that existing compounds
need to be tested against other organisms.
Our structure-based approach to improving DXR inhibitors
focused initially on modifying the 3,4-dichlorophenyl-substi-
tuted analogues in an attempt to utilize a conserved water-filled
pocket we had earlier identified. However, this did not result in
improved IC50 values.7 Here we describe a new strategy in
which we extend from the hydroxamic acid group of
fosmidomycin and related analogues. While we were still
unsuccessful in terms of obtaining compounds active against M.
tuberculosis, some were highly active against P. falciparum
growth in an in vitro blood cell assay. Thus, the STOP
approach has allowed us to design inhibitors that, for the first
time, bridge the substrate and cofactor binding sites and
provide useful new leads in the search for antimalarial drugs.
a
b
c
d
Reference 10. Reference 43. Reference 26. Reference 44.
based on the work of Puyau et al. and Kurz et al.27,28 Synthesis
started from the O-benzylhydroxylamine compound 5 (Scheme
1, which was prepared in three steps according to published
procedures).29 Next, acylation was performed by adding an
excess of acyl chloride to a solution of 5 in dichloromethane
(DCM) or via a HATU-mediated amide coupling (6i−6j).29
Hydrogenolysis of the benzyl group was conducted in the
presence of 10% Pd/C with methanol as solvent under
hydrogen atmosphere at room temperature. Finally, phospho-
nate ester hydrolysis was performed using trimethylsilyl
bromide (TMSBr) in dry DCM to obtain the target
compounds 8a−8j in moderate to good yields. The inhibitory
capacity of the analogues was evaluated in a spectrophotometric
assay, in which the MtDXR-catalyzed NADPH-dependent
rearrangement and reduction of DXP to form MEP was
monitored at 340 nm.
Table 1 shows that none of these compounds inhibited
MtDXR more effectively than fosmidomycin. The acetyl
derivative, 2, showed a small drop in activity,7,10 which we
attribute to minor changes in the interaction with the indole
ring of Trp203 in the active site of MtDXR. In the first crystal
structure of a DXR ternary complex in a conformational state
suitable for catalysis,10 we observed a well-defined active-site
flap that covered the fosmidomycin bound in the substrate site.
The relative orientation of the NADPH-binding domain
suggested that the conformation would allow hydride transfer
from the nicotinamide ring to the substrate. In this state, the
RESULTS AND DISCUSSION
■
General Considerations. The active site of DXR is made
up of two main components, the substrate (DXP/MEP) and
NADPH-binding sites. The substrate-binding site is quite small
and polar/charged and so has limitations when trying to
develop improved drug-like molecules; however, this has been
the target area of essentially all inhibitor design to date.
Structural work has demonstrated the dynamic nature of an
active-site flap involved in the binding of fosmidomycin and
analogues substituted in the Cα position. Attempts have been
made to extend beyond the hydroxamate group; these changes
often have adverse effects on activity, although a few analogues
have been obtained with submicromolar IC50-values for EcDXR
and Pf DXR.26 In our search for DXR inhibitors with in vivo
activity, we decided to evaluate some of these acyl-substituted
analogues on MtDXR and M. tuberculosis and to probe this
B
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Fosmidomycin Analogues 8a−8j
a
Reagents and conditions: (i) 6a−6h, acid chloride, DCM, rt, or 6i−6j, carboxylic acid, HATU/DIPEA, DMF, rt; (ii) H2, 10% Pd/C, MeOH, rt;
(iii) TMSBr, DCM, rt.
indole ring of Trp203 from the flap forms a close set of
interactions with the fosmidomycin backbone. The MtDXR−
Mn−2−NADPH ternary complex showed a similar active-site
structure and enzyme−inhibitor interactions, with the newly
introduced methyl group of the analogue forming good van der
Waals interactions with the indole ring.21 Equivalent studies on
ternary complexes of Pf DXR18 confirm the generality of these
interactions, although there are subtle differences, particularly
in the interaction of the phosphonate group with the enzyme,21
that probably underlie small differences in relative affinity of the
two compounds with EcDXR and Pf DXR (Table 1).
The Disubstituted Fosmidomycin Analogues. The
approximate coplanarity of the hydroxamate and dichlorophen-
yl groups in the experimentally determined enzyme−inhibitor
structures of 3 and 4 with MtDXR,7 with both groups aimed
out toward solvent, suggested that we could combine the Cα-
substituent of 3 and the phenyl substituent of the hydroxamate
in 8d to produce the disubstituted compound, 14a (Table 2).
Compound 14a served as the lead for our second series in
which various substituents were added to the phenyl ring with
the aim of reaching into the NADPH-binding site (see below).
Groups were selected among commercially available acyl
chlorides and carboxylic acids with a variety of size and other
properties, guided by molecular modeling using the exper-
imental MtDXR−Mn−14a structures (described below). This
limited set of related analogues primarily investigated
substitutions at the ortho position (see Table 2). All
compounds were synthesized as racemic mixtures, as outlined
in Scheme 2.
The key intermediate 11 was obtained in a previously
reported, three-step synthetic route,7,24 starting with an
oxidative Heck reaction between acrolein and 3,4-dichlor-
ophenyl boronic acid,30,31 to furnish (E)-3-(3,4-
dichlorophenyl)acrylaldehyde (9). This was followed by 1,4-
addition of triethyl phosphate, and finally reductive amination
with O-benzylhydroxylamine in the presence of sodium
cyanoborohydride as reducing reagent, to produce 11.
Depending on availability of starting materials, the next step
was performed with acyl chlorides in DCM and triethylamine,
furnishing compounds 12a−12d in 46−92% yield, or with
suitable carboxylic acids using propane phosphonic acid
anhydride (T3P) in ethyl acetate as a coupling reagent
(compounds 12f−12k, 43−75% yields). The 2-(2-
thienylmethyl)benzoic acid used to obtain compound 12e
was prepared from 2-(2-thienylcarbonyl)benzoic acid by
refluxing for 48 h in ammonia with zinc powder. The benzyl
protecting group in compounds 12a−12d and 12f−12k was
removed via palladium-catalyzed (10% Pd/C) hydrogenation
under atmospheric hydrogen pressure. Probably due to catalyst
poisoning, the reaction did not give a product with the
thiophene derivative (12e), even in harsher, acidic conditions.
Debenzylation of 12e was easily achieved with trifluoroacetic
acid (TFA), TMSBr, and thioanisole. Final compounds were
obtained by hydrolysis of the diethylphosphonates 13a−13k by
The decreasing potency of the ethyl (8a), 2-furyl (8b), and
cyclopropyl (8c) analogues can be attributed to clashes with the
indole ring of Trp203. However, the introduction of larger rings
with only one extra dihedral degree of freedom (8d, 8e, 8f, and
8g) resulted in IC50s in the low micromolar range. Compounds
8i and 8j are reasonably potent on EcDXR and PfDXR,26 and
they were therefore resynthesized and evaluated on MtDXR.
Unfortunately, they have IC50s greater than 200 μM on that
enzyme. The active compounds in this series are unlikely to be
accommodated in a ternary complex containing an ordered flap.
Instead, we expect the flap to be disordered, as we observed in
complexes of 3 and 4 with MtDXR;7 unfortunately, we could
not obtain an experimental structure to confirm this hypothesis.
It is striking that the benzyl derivative (8h) in this series shows
much decreased activity as judged by IC50, which could indicate
an inability to form suitable interactions in the opened active
site that we describe in more detail below. An alternative
hypothesis (also discussed below) is that these compounds
extend into the cofactor-binding site at the position of the
nicotinamide ring. In such a case, the measured IC50 would be
higher than the binding constant due to competition with
NADPH. Because the concentrations of NADPH used in the
assay are high (200 μM, compared to an estimated Km value of
7.2 μM),9 this effect would only be seen when an inhibitor
binds effectively in the NADPH site. We note also that our IC50
values for MtDXR were systematically higher than those
published for the EcDXR and PfDXR enzymes, see Table 1.
This results from our choice of more stringent assay conditions,
which should allow better discrimination among tight-binding
compounds.
Unfortunately, none of these compounds had significant in
vitro activity against M. tuberculosis; they were not tested
against P. falciparum.
C
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(Table 2). Introduction of progressively larger substituents at
the ortho position of the phenyl ring resulted in a decline of
MtDXR inhibitory activity. Surprisingly, all of the ortho-
substituted compounds were inactive in the P. falciparum
parasite growth assay. When the triazole substituent was moved
from the ortho (14i) to the meta (14j) and para positions
(14k), the MtDXR IC50 values were improved by 2 and 1
orders of magnitude, respectively. The activity in the P.
falciparum parasite assay was also recovered in these two
compounds. In fact, 14j had an MtDXR IC50 value that is
comparable to 2. The best in vitro growth inhibition of P.
falciparum, however, is shown by the lead compound, 14a (0.04
μM); it has five times the activity of 14k (0.19 μM), and 10
times the activity of 14j (0.39 μM). 14a also has eight times the
published activity of 2 (0.32 μM)24 and at least an 8-fold
improvement over fosmidomycin in our in vitro growth
inhibition assays (fosmidomycin was used as a control in our
assays, and had IC50 values in the range 0.2−0.4 μM, compared
to values of 0.3−1.1 μM found in the literature).6,24
Table 2. Inhibition of MtDXR Activity and Plasmodium
Growth by Disubstituted Fosmidomycin Analogues
General Comments on Crystallographic Results. De-
tails of the crystallographic work are included in the Supporting
Information. All of the complexes represent monoclinic and
tetragonal crystal forms that we have described previously,
although there is a large variation in the unit cell constants for
crystals of the monoclinic form. In the monoclinic crystals
described here, as for most of our MtDXR structures, the
asymmetric unit contains a complete dimer where the
NADPH-binding site is more closed in one subunit (chain A;
conformational state being described using a yardstick10 to
describe relative orientations of the N- and C-terminal
domains) and more open in the other (chain B). Somewhat
counterintuitively, the active-site flap in the closed (inhibitor-
bound) site is disordered in each complex, while the flap of the
more open (empty) active site is generally well ordered and
directed away form the active site because of interactions with
the closed-chain active site of a crystallographically related
dimer. The A chain in this crystal form has a higher overall
temperature factor than the B chain, particularly in the edge
helix of the NADPH-binding domain. Our concern that
conclusions about ligand binding could be influenced by crystal
contacts had led us to carry out extensive screening of
crystallization conditions that ultimately identified a second
crystal type. This tetragonal form requires NADPH, and the
asymmetric unit also contains a dimer.21 In this case, the two
subunits adopt the same active-site conformation and are
related by an almost perfect noncrystallographic 2-fold axis. In
the MtDXR−Mn−2−NADPH ternary complex studied earlier
in the tetragonal form (Protein Data Bank (PDB) entry 4A03),
the active-site flap is in a well-defined closed substrate-binding
conformation that makes extensive interactions with the
inhibitor.21 However, the flaps were not well ordered in the
MtDXR−Mn−14a−NADPH complex studied here. Overall,
therefore, the behavior of the active-site flap in structures of the
complexes described in the present study was more similar to
that seen in complexes with the dichlorophenyl Cα-substituted
fosmidomycin analogues 3 and 4 than to the structures in
complex with fosmidomycin or 2.
a
b
Reference 7. Reference 24.
TMSBr in DCM and purified on reversed phase HPLC using
gradient elution with acetonitrile in 0.1% aqueous TFA.
Biochemical/Biological Evaluation of the 14a Series.
Table 2 summarizes results for the compounds in the second
series. The fact that these compounds were tested as racemic
mixtures should be remembered in the interpretation of the
data. While the crystallographic work (see below) indicates that
the S-enantiomer is the predominant form bound to the
enzyme, we have no proof at this time that activity of the R-
enantiomer is entirely lacking. Therefore, the reported IC50s
should be viewed as upper limits, with the true value probably
closer to one-half of this. The lead compound 14a had
submicromolar MtDXR activity (measured IC50 = 0.32 μM)
and remarkably good in vitro growth inhibition in the P.
falciparum parasite assay (IC50 = 0.04 μM). Enzyme inhibition
by this compound was less efficient than the Cα-substituted
formyl derivative 3 but better than the acetyl derivative 4
Mode of Binding to 14a. We were able to crystallize 14a
in complex with MtDXR in both of the known space groups,
with and without the NADPH cofactor, and so can be certain
that our conclusions are unbiased by any crystallographic
artifacts. Figure 2a shows a superposition of the three
independent views we obtained of 14a binding to MtDXR.
D
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Disubstituted Analogues
a
Reagents and conditions: (i) Pd(OAc)2, acrolein, dmphen, p-benzoquinone, 100 °C, MW; (ii) (a) TEP, phenol, 100 °C, (b) 2 M HCl, acetone, 100
°C; (iii) (a) O-benzylhydroxylamine hydrochloride, pyridine, EtOH, rt, (b) NaCNBH3, MeOH, HCl, rt; (iv) acyl chloride, TEA, DCM, rt or
carboxylic acid, T3P, TEA, EtOAc, 0 °C then rt; (v) H2, 10% Pd/C or TMSBr, TFA, thioanisole; (vi) TMSBr, DCM, rt.
Figure 2. (a) Superposition of MtDXR−14a structures from the monoclinic (silver carbons) and tetragonal (carbons in darker shades of gray)
crystal forms. Side chains from residues that coordinate the active-site metal ion are included, as well as the nicotinamide ring of the NADPH
cofactor from each active site in the MtDXR−Mn−14a−NADPH ternary complex structure. Interactions with the metal are shown with small
bubbles. (b) Interactions at the active site of the 14a complex in the monoclinic form. Green bubbles indicate interactions of the phosphonate group
with the enzyme, gold are interactions at the active site metal, and blue are intramolecular ring−ring interactions in the inhibitor. (c) Superposition
of 14a (monoclinic form, gray carbons), 3 (light-red carbons, PDB entry 2Y1D), and 4 (yellow carbons, PDB entry 2Y1G). Superpositions were
performed using the enzyme coordinates as described in the Experimental Section.
The ligand conformations are very similar, as are most of the
interactions with the enzyme (Figure 2b). Metal-ion coordina-
tion is essentially unchanged with distances in the expected
range; the hydroxamic acid group in all cases adopts the
synperiplanar conformation. The largest difference observed is
that in the MtDXR−Mn−14a complex (i.e., that lacking
NADPH), a small change in the side chain of Asp151 (Figure
2a) modifies its coordination to the metal ion, with small
rearrangements of the next shell of side chains, in particular the
Lys128-Asp151-His154-Glu225 cluster. The phosphonate
group is involved in an extensive set of hydrogen-bond and
salt-link interactions that we have described earlier for
complexes with 1, 2, 3, and 4. Despite the differences in
substitution at the hydroxamate, the conformations of 14a, 3,
and 4 are also extremely similar (Figure 2c). The orientation
and interactions of each dichlorophenyl ring within the enzyme
are identical within the limits of the diffraction experiments.
The conformation of the phenyl ring at the hydroxamate in 14a
allows an intramolecular interaction between the two aromatic
groups, which could enhance binding by increasing the
population of the “correct” conformation of the ligand in
solution (Figure 2b). The ring planes are inclined to one
another by ∼47°, and their centers are separated by ∼4.3 Å.
Some ring-edge atoms make van der Waals contact (3.3−3.5
Å), while others are further apart (5−5.5 Å). The other face of
the phenyl ring packs against the side chain of Met267, with
distances to the side-chain sulfur atom in the range 3.7−4.3 Å.
In this conformation, the faces of both phenyl rings are
inaccessible to solvent, while portions of the edge of each ring
are exposed. The proximity of the phenyl ring and the
nicotinamide ring of NADPH (Figure 2a) inspired us to design
analogues that could reach into the cofactor-binding pocket.
Crystal Structures of MtDXR in Complex with 14f and
14i. Despite their relatively high IC50 values (∼15 μM), good
solubility of the compounds allowed us to obtain crystal
structures for complexes of MtDXR with 14f and 14i. Both
structures lack electron density for a metal ion in their A-chain
active sites, although MnCl2 was included in the soak solution
of the former and the cocrystallization experiments of the latter.
The restraints of metal coordination via their hydroxamic acid
E
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (a) Superposition of 14a (monoclinic form, gray carbons), 14f (light-blue carbons), and 14i (magenta carbons), shown in stereo. The gold
sphere represents the active-site metal ion in the 14a complex. The superposition was performed using the enzyme coordinates (as described in the
Experimental Section) and therefore shows differences in ligand position within the active site. (b) Superposition of the same three structures was
performed using the fosmidomycin backbone atoms to highlight differences in the inhibitor conformation.
groups have thus been removed, and the three metal-
coordinating side chains of the enzyme are free to take on
other conformations. Despite their close chemical similarity,
14f and 14i show striking differences in their mode of binding
to the enzyme (Figure 3a). Their propyl chain conformations
are very similar (Figure 3b, root-mean-squares fit for these
atoms 0.13 Å), and the differences arise from a rigid body
rotation (17°) around and translation (0.9 Å) along an axis
close to the phosphonate groups. As a result, the hydroxamate
group of 14f is pulled significantly closer to the metal-binding
site (∼2 Å at the hydroxamate carbonyl oxygen atom). Close to
the rotation axis, the interactions with the enzyme are
essentially identical, with the same residues involved in
interactions with the phosphonate (hydrogen bonds with
main-chain nitrogen of residue 177, and side chains of Ser177,
Ser213, Asn218, and Lys219, as well as two water molecules),
and with the dichlorophenyl ring (Figure 2b). In neither
structure is the phenyl ring completely buried. At their
respective hydroxamate groups, the differences in binding
mean that quasi-equivalent sets of interactions are seen. For
example, the hydroxyl group of Ser152 interacts with the
hydroxamate oxygen in each complex, although these inhibitor
oxygens are separated by 2.8 Å. Likewise, the N-hydroxyl
oxygen of each analogue interacts with the side chain of
Glu153, although the detailed sets of interactions differ. In the
MtDXR−14i structure, a portion of the active-site flap in chain
B is disordered (corresponding to residues A204−208),
although it is ordered in the structure with 14f.
The ortho-substituted phenyl ring in both metal-free
complexes (14f and 14i) makes intramolecular ring−ring
interactions with the dichlorophenyl ring. In each case, the
rings are closer to coplanarity than is observed in the MtDXR−
Mn−14a complex (Figure 3b). The ortho groups of the two
compounds are directed toward, and entering, the NADPH-
binding site. Both have an intramolecular contact to the para-
chlorine atom (in the range 3.5−3.7 Å), but no intramolecular
hydrogen-bonding interactions are observed.
Direct Measurement of KD Using Fluorescence Spec-
troscopy. As the structures of complexes with 14f and 14i
showed that their ortho substituents impinge on the NADPH
site, it seemed likely that the measured IC50 would not be a
good measure of ligand affinity, as discussed above for
compound 8h. This led us to seek other methods for
determining the affinity of MtDXR for the various compounds.
A direct fluorescence assay, based on changes in the
environment of one or more of the six tryptophan residues
of the protein, proved a feasible option; the results with a
number of compounds are presented in Table 3. For five
Table 3. Binding Constants Determined by Fluorescence
Spectroscopy for a Selection of Compounds against MtDXR
a
in the Presence of 1.5 mM MnCl2
Kd with MnCl2 (μM)
MtDXR IC50 (μM)
1
0.04
0.12
0.36
0.32
0.21
0.09
0.09
0.04
0.17
0.20
0.08
0.15
0.70
67.7
0.32
8
3
4
8h
14a
14d
14f
14i
14j
14k
19
13
0.14
1.2
a
IC50 values for MtDXR are also shown for easier comparison.
compounds (1, 3, 4, 14a, and 14j), the measured dissociation
constant (KD) in the presence of MnCl2 agrees well with the
degree of enzyme inhibition as indicated by the IC50s. The
three compounds with ortho substitutions (14d, 14f, and 14i)
showed tighter binding than the unsubstituted lead (14a). The
IC50s of these compounds, however, were much higher than
their KD values (Table 2), and so we conclude that, like 14f and
14i, other compounds with ortho substituents are likely to
interfere with the binding of NADPH. The one para-substituted
compound in the study, 14k, seems to have similar behavior,
but more compounds will need to be synthesized and tested to
determine if the effect is a general one.
Structure−Activity Relationship (SAR) of Analogues
14a−14k. None of the new compounds inhibited the growth
of M. tuberculosis strain H37Rv in a microplate Alamar blue
assay (data not shown). In contrast, the disubstituted analogues
(14a−14k) did show a range of activities against the P.
falciparum 3D7 strain in an in vitro growth assay (Table 2).
Having obtained the crystal structures of three of the analogues
in Table 2, we could address the SAR of this set of compounds.
The MtDXR inhibition of 14a is poorer than that of the formyl
derivative 3 but better than the acetyl derivative 4 described
F
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
earlier (Table 2).7 The structures of complexes with these three
compounds (3, PDB entries 2Y1D and 2Y1F; 4, 2Y1G; 14a,
3ZHX and 3ZHY) all show a disordered active-site flap and
otherwise share a great deal of similarity. In particular, the
fosmidomycin backbone, dichlorophenyl ring, and metal-ion
coordination are highly conserved in these structures (Figure
2c). Despite the poorer enzyme inhibition relative to 1, 2, and
3, the activity of 14a on P. falciparum growth is superior (40
nM, Table 2) and closely comparable to the reference
compound artemisinin (IC50 in the range 5−25 nM in our
assays, similar to the value of 10 nM reported by Baniecki et
al.32
Introducing groups at the ortho position in the phenyl ring at
the hydroxamic acid resulted in compounds that have poorer
MtDXR inhibition as judged by IC50 values; only the analogue
with a methyl substituent (14b) had an IC50 that is
submicromolar. In the MtDXR−Mn−14a−NADPH structure,
the ortho position and the edge of the nicotinamide ring make
van der Waals contact (∼3.5 Å). If substituents are introduced
here, clashes would become severe in this orientation of the
ring. If the substituted phenyl ring is flipped by 180°, clashes
with the dichlorophenyl ring or protein side chains would
develop. Similar conclusions follow if we superimpose the
MtDXR−Mn−14a−NADPH and MtDXR−14f structures to
look for potential clashes between inhibitor and cofactor.
However, the high concentration of NADPH needed for the
activity assay dictates that the IC50 measurements for
compounds entering the cofactor site will underestimate their
actual binding constants. Our modeling of meta and para
substitutions suggested that these would be easier to
accommodate, which is borne out by the results with a limited
number of modifications (involving a triazole ring) at each of
the three positions (Table 2). Although both 14j and 14k show
good inhibition of MtDXR activity and P. falciparum growth,
neither improved on 14a.
compared to fosmidomycin and is 2.5-fold more effective
than 3, the best compound described by Haemers et al.24 The
improved in vitro activity of 14a may reflect differences
between MtDXR and PfDXR or properties of the compound
that increase uptake or disfavor export and metabolism. A series
of decorations at the ortho position on the ring did not improve
inhibitory properties in either enzymatic or parasite growth
assays (Table 2), although substitutions at the meta and para
positions gave superior results. The micromolar or better
enzymatic and in vitro activities of compounds 14j and 14k
suggest that these triazole substitutions may be making specific
interactions with the enzyme.
The structures of the metal-free enzyme in complex with two
closely related inhibitors indicate new binding poses that may
help in the design of additional inhibitors. Restraining a
hydroxamate group to coordinate with the active-site metal ion
while maintaining the extensive set of interactions in the
phosphonate binding site may limit the ability of this series of
compounds to form optimal intra- and intermolecular
interactions within the boundaries of the MtDXR active site.
The propyl chain conformations of 14f and 14i are very similar
in the metal-free active sites despite the different detailed set of
interactions and show only small changes when coordinating
the metal ion (Figure 3b). This suggests that these backbone
conformations are close to a local energy minimum. The metal-
free complexes show more pronounced face-to-face ring
interactions, which are not associated with optimal electrostatic
interactions between aromatic rings in proteins. This may be a
consequence of accommodating the extensions associated with
these compounds as well as the optimization of protein−
inhibitor interactions. Further, the structures indicate that ortho
substitutions to the phenyl ring at the hydroxamate enter the
NADPH binding site and, therefore, make it difficult to
interpret the enzymatic IC50 measurements as indicators of
tight binding. The fluorescence-based binding data, however,
indicate that 14f and 14i have comparable dissociation
constants to fosmidomycin and that these compounds bind
more tightly than 14a. The relatively poor in vitro IC50s of 14f
and 14i relative to 14a, therefore, we attribute to competition
with the NADPH cofactor. The good enzymatic and in vitro
activity of the meta- and para-substituted triazoles, 14j and 14k,
on the other hand, we attribute to a 14a-like mode of binding
where the substitutions are directed away from the
nicotinamide ring binding site of the cofactor into the
additional volume created by the disordered active-site flap.
This space is also the likely binding site for the extensions
showing good enzyme inhibition in the series shown in Table 1.
Our work provides new insights on how DXR interacts with
a promising series of novel compounds, and suggests new
routes for the creation of even better, biologically active
molecules.
CONCLUSIONS
■
The natural product antibiotic fosmidomycin (1) and its acetyl
derivative 2 have attracted a great deal of attention because they
inhibit the growth of Plasmodium species in vitro and in vivo.6
Haemers et al.24 synthesized a number of fosmidomycin
analogues with an α-aryl substituent, several of which showed
improved activity against P. falciparum in an in vitro growth
assay. Complexes with the most active (3) and its slightly less
potent acetyl derivative (4)7,22 showed that these inhibitors
destabilize the active-site flap that packs against the inhibitors in
the metal-free complexes of EcDXR20 and ternary complexes of
PfDXR18 and MtDXR.10,21 The Cα-substituents of 3 and 4
bind in a manner that leaves one face of the phenyl ring
accessible to solvent. The coplanarity of the phenyl ring and the
hydroxamic acid in these structures suggested new designs that
would optimize this mode of binding by replacing the methyl
group of 4 with a phenyl group. We have now tested this
hypothesis by preparing compound 14a, which indeed has
submicromolar MtDXR activity (IC50 = 0.32 μM) and
remarkably good in vitro growth inhibition in a P. falciparum
parasite assay (IC50 = 0.04 μM). The crystal structures of
MtDXR together with 14a confirm the success of the design
aim.
EXPERIMENTAL SECTION
■
Chemistry. The microwave reactions were performed in a Smith
synthesizer producing controlled irradiation at 2450 MHz with a
power of 0−300 W. The reaction temperature was determined using
the built-in online IR sensor. Flash column chromatography was
performed on Merck silica gel 60 (40−63 μm). Analytical thin layer
chromatography was done using aluminum sheets precoated with silica
gel 60 F254. Analytical RPLC-MS was performed on a Gilson HPLC
system with a Finnigan AQA quadrupole mass spectrometer with
detection by UV (DAD) using an Onyx Monolithic C18 column (50
mm × 4.6 mm). Acetonitrile in 0.05% aqueous HCOOH at a flow rate
of 4 mL/min was used as a mobile phase. Preparative RP-HPLC was
Direct KD measurements, as well as the IC50s, show the
binding to MtDXR is ordered 1, 3, 14a, and 4 with
fosmidomycin the tightest binder. Compound 14a, however,
has improved inhibition of the growth of P. falciparum
G
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
performed on a system equipped with a Zorbax SB-C8 column (150
mm × 21.2 mm) in H2O/MeCN gradient with 0.1% CF3COOH or
with 0.05% HCOOH as mobile phase at a flow rate of 5 mL/min,
unless stated otherwise in experimental details. Purity of the final
compounds was determined by RP-HPLC on a Restec Allure biphenyl
column (5 μm, 4.6 mm × 50 mm) and an ACE 5 C18 column (5 μm,
4.6 mm × 50 mm) in a H2O/MeCN gradient with 0.1% CF3COOH
and UV detection at 220 nm. All the compounds showed purity above
therefore, followed our earlier policy of depositing the best fitting
enantiomer (in all cases, S), although both enantiomers were evaluated
during the refinements. Water molecules were added using the profile-
based methods implemented in O, and structural comparisons were
made with the Lsq commands in O with default or close-pair Cα cut-
offs.40 Ring-to-ring orientations were calculated with the Ring
commands available in O version 14. Figures were created in O and
rendered with Molray.41 Secondary structure assignments were made
with the yasspa command in O and drawn with red−blue rainbow
coloring for one chain. Data collection and refinement statistics for the
various new complexes are given in Supporting Information, Table S2.
Briefly, two distinct 14a complexes were solved in different space
groups, with and without NADPH, and refined at resolutions of 2.3
and 2.0 Å, with crystallographic R-factors of 20.1% and 18.4%,
respectively. MtDXR 14f and 14i complexes were refined at
resolutions of 2.25 and 1.90 Å, with crystallographic R-factors of
20.0% and 20.5%, respectively. More structural details, and figures
showing the electron density around each of the ligands, are included
in the Supporting Information. The electron density for each PDB
entry is available at the Uppsala Electron Density Server.42
1
95%, except compound 14h (93% pure). H and 13C NMR spectra
1
were recorded on Varian Mercury Plus instrument; H at 399.8 MHz
1
and 13C at 100.5 MHz. The chemical shifts for H NMR and 13C
NMR were referenced to TMS via residual solvent signals (1H, CDCl3
at 7.26 ppm, CD3OD at 3.31 ppm; 13C, CDCl3 at 77.16 ppm, CD3OD
at 49.00 ppm). 31P spectra were recorded on a Varian Mercury 300
Plus instrument at 121.4 MHz; the chemical shifts were referenced to
85% H3PO4 which was used as an external standard. Molecular masses
(HR-ESI-MS) were determined on a Micromass Q-Tof2 mass
spectrometer equipped with an electrospray ion source. Compounds
9, 10, and 11 were synthesized as previously reported.7,24
Materials. Most reagents were purchased from commercial
suppliers and used without further purification. DCM was distilled
under nitrogen immediately before use; calcium hydride was used as a
drying agent.
ASSOCIATED CONTENT
■
S
* Supporting Information
Diethyl-3-(N-(Benzyloxy)phenylamido)-1-(3, 4-
dichlorophenyl)propylphosphonate (12a). [3-Benzyloxyamino-1-
(3,4-dichlorophenyl)-propyl]-phosphonic acid diethyl ester (11) (0.10
g, 0.22 mmol), benzoyl chloride (51 μL, 0.44 mmol), and
triethylamine (94 μL, 0.67 mmol) in DCM (3.0 mL) were stirred at
room temperature for 26 h. When the reaction was completed, the
mixture was diluted with water and the aqueous layer was extracted
with DCM. The combined organic fractions were dried over
anhydrous MgSO4 and filtered; the solvent was removed under
vacuum, and the product was purified by column chromatography
using ethyl acetate/isohexane (1/4), then DCM to yield 0.056 g
(46%) of the desired product.
Additional experimental details concerning synthesis of
compounds, spectroscopic data, protein expression and
purification, enzyme activity/inhibition/binding studies, and
crystallization. This material is available free of charge via the
Internet at http://pubs.acs.org.
Accession Codes
Coordinates and structure factor data have been deposited at
the PDB with entry codes: MtDXR−Mn−14a (3ZHX),
MtDXR−Mn−14a−NADPH (3ZHY), MtDXR−14f (3ZHZ),
and MtDXR−14i (3ZI0).
1-(3,4-Dichlorophenyl)-3-(N-hydroxyphenylamido)-
propylphosphonic Acid (14a). A mixture of 12a (0.06 g, 0.10
mmol) and 10% Pd/C (0.01 g, 0.10 mmol) in methanol (1.5 mL) was
hydrogenated at atmospheric pressure for 5 h. After reaction was
finished, the mixture was filtered through a Celite pad and
concentrated under vacuum to give compound 13a. To a solution of
crude phosphonate diethyl ester (13a) (0.05 g, 0.10 mmol) in dry
DCM (1 mL), TMSBr (68.5 μL, 0.50 mmol) was added dropwise
under N2 at room temperature. After 10 h, the volatiles were removed
under vacuum to give the corresponding phosphonic acid derivative.
The product was purified on preparative RP-HPLC using gradient
elution (15−60% acetonitrile in 0.1% aqueous TFA) to yield 0.02 g
AUTHOR INFORMATION
■
Corresponding Author
*Structural biology and enzymology (S.L.M.): phone, +46-18-
4714990; fax, +46-18-530396; E-mail, mowbray@xray.bmc.uu.
se. Medicinal chemistry (A.K.): phone, +46-18-4714293; fax,
+46-18-4714474; E-mail, anders.karlen@orgfarm.uu.se.
Author Contributions
∥Christofer Bjorkelid and Anna Jansson contributed equally to
̈
the structural biology and enzymology. Bjorkelid, Jansson, and
̈
1
(53%) of 14a. H NMR (400 MHz, CD3OD) δ ppm: 2.25 (m, 1H),
Wieckowska contributed equally to the work as a whole.
̨
2.61 (m, 1H), 3.11 (m, 1H), 3.62 (m, 2H), 7.23 (m, 1H), 7.42 (m,
7H).
Notes
Crystallography. Diffraction data were collected from single
crystals at the European Synchrotron Radiation Facility (ESRF,
Grenoble, France) or at MAXlab (Lund, Sweden). The diffraction
images were indexed and integrated using MOSFLM33 and scaled with
SCALA,34 elements of the CCP4 package (CCP4, 1994).35 The
crystallized complexes belonged to space groups we have previously
described (P217 or P4321221). The MtDXR−14i complex, however, had
a rather large change in the unit cell parameters compared to our
earlier published values, and so the structure was solved by molecular
replacement using Phaser.36 All complexes were refined with
alternating reciprocal-space refinement with REFMAC537 and manual
rebuilding with O.38 Stereochemical restraints for each ligand were
generated in O with the qds tools, and the fitted model was then used
to generate restraints for the refinement with a suitably edited
dictionary generated by REFMAC5. The hydroxamate group was
restrained to planarity with target standard deviations of 0.02 Å. Metal
coordination target distances were taken from Harding39 and used
with standard deviations of 0.02 Å. All inhibitors were synthesized as
racemic mixtures and, at the resolution of these studies, we cannot
unequivocally state that one enantiomer binds exclusively. We have,
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Jerry Stahlberg, Swedish University of Agricultural
̊
Sciences, for help with the fluorescence data analysis. This work
was supported by the Swedish Foundation for Strategic
Research (SSF), the European Union Sixth Framework
Program NM4TB (CT:018923), the Swedish Research Council
and Uppsala University.
ABBREVIATIONS USED
■
DXP, 1-deoxy-D-xylulose 5-phosphate; DXR, 1-deoxy-D-xylu-
lose-5-phosphate reductoisomerase; EcDXR, Escherichia coli
DXR; MEP, 2-C-methyl-D-erythritol 4-phosphate; MtDXR,
Mycobacterium tuberculosis DXR; PDB, Protein Data Bank;
Pf DXR, Plasmodium falciparum DXR; SAR, structure−activity
relationship; STOP, same-target-other-pathogen
H
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and spacer of fosmidomycin, a potent inhibitor of deoxyxylulose
phosphate reductoisomerase. J. Org. Chem. 2010, 75, 3203−3207.
(16) Ortmann, R.; Wiesner, J.; Silber, K.; Klebe, G.; Jomaa, H.;
Schlitzer, M. Novel deoxyxylulosephosphate-reductoisomerase inhib-
itors: fosmidomycin derivatives with spacious acyl residues. Arch.
Pharm. 2007, 340, 483−490.
(17) Bodill, T.; Conibear, A. C.; Blatch, G. L.; Lobb, K. A.; Kaye, P.
T. Synthesis and evaluation of phosphonated N-heteroarylcarbox-
amides as DOXP-reductoisomerase (DXR) inhibitors. Biorg. Med.
Chem. 2011, 19, 1321−1327.
(18) Umeda, T.; Tanaka, N.; Kusakabe, Y.; Nakanishi, M.; Kitade, Y.;
Nakamura, K. T. Molecular basis of fosmidomycin’s action on the
human malaria parasite Plasmodium falciparum. Sci. Rep. 2011, 1, 9.
(19) Steinbacher, S.; Kaiser, J.; Eisenreich, W.; Huber, R.; Bacher, A.;
Rohdich, F. Structural basis of fosmidomycin action revealed by the
complex with 2-C-methyl-D-erythritol 4-phosphate synthase (IspC). J.
Biol. Chem. 2003, 278, 18401−18407.
(20) Mac Sweeney, A.; Lange, R.; Fernandes Roberta, P. M.; Schulz,
H.; Dale Glenn, E.; Douangamath, A.; Proteau Philip, J.; Oefner, C.
The crystal structure of E. coli 1-deoxy-D-xylulose-5-phosphate
reductoisomerase in a ternary complex with the antimalarial
compound fosmidomycin and NADPH reveals a tight-binding closed
enzyme conformation. J. Mol. Biol. 2005, 345, 115−127.
REFERENCES
■
(1) Okuhara, M.; Kuroda, Y.; Goto, T.; Okamoto, M.; Terano, H.;
Kohsaka, M.; Aoki, H.; Imanaka, H.; Okuhara, M.; Kuroda, Y.; Goto,
T.; Okamoto, M.; Terano, H.; Kohsaka, M.; Aoki, H.; Imanaka, H.
Studies on new phosphonic acid antibiotics. III. Isolation and
characterization of FR-31564, FR-32863 and FR-33289. J. Antibiot.
(Tokyo) 1980, 33, 24−28.
(2) Davey, M. S.; Tyrrell, J. M.; Howe, R. A.; Walsh, T. R.; Moser, B.;
Toleman, M. A.; Eberl, M. A promising target for treatment of
multidrug-resistant bacterial infections. Antimicrob. Agents Chemother.
2011, 55, 3635−3636.
(3) Rohmer, M. The discovery of a mevalonate-independent pathway
for isoprenoid biosynthesis in bacteria, algae and higher plants. Nat.
Prod. Rep. 1999, 16, 565−574.
(4) Kuzuyama, T.; Shimizu, T.; Takahashi, S.; Seto, H.
Fosmidomycin, a specific inhibitor of 1-deoxy-D-xylulose 5-phosphate
reductoisomerase in the nonmevalonate pathway for terpenoid
biosynthesis. Tetrahedron Lett. 1998, 39, 7913−7916.
(5) Zeidler, J.; Schwender, J.; Mueller, C.; Wiesner, J.; Weidemeyer,
C.; Beck, E.; Jomaa, H.; Lichtenthaler, H. K. Inhibition of the non-
mevalonate 1-deoxy-D-xylulose-5-phosphate pathway of plant iso-
prenoid biosynthesis by fosmidomycin. Z. Naturforsch., C 1998, 53,
980−986.
(6) Jomaa, H.; Wiesner, J.; Sanderbrand, S.; Altincicek, B.;
Weidemeyer, C.; Hintz, M.; Turbachova, I.; Eberl, M.; Zeidler, J.;
Lichtenthaler, H. K.; Soldati, D.; Beck, E. Inhibitors of the
nonmevalonate pathway of isoprenoid biosynthesis as antimalarial
drugs. Science 1999, 285, 1573−1576.
(21) Bjorkelid, C.; Bergfors, T.; Unge, T.; Mowbray, S. L.; Jones, T.
A. Structural studies on Mycobacterium tuberculosis DXR in complex
with the antibiotic FR-900098. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2012, 68, 134−143.
(22) Deng, L.; Diao, J.; Chen, P.; Pujari, V.; Yao, Y.; Cheng, G.;
Crick, D. C.; Prasad, B. V. V; Song, Y. Inhibition of 1-deoxy-D-
xylulose-5-phosphate reductoisomerase by lipophilic phosphonates:
SAR, QSAR, and crystallographic studies. J. Med. Chem. 2011, 54,
4721−4734.
(7) Andaloussi, M.; Henriksson, L. M.; Więckowska, A.; Lindh, M.;
Bjorkelid, C.; Larsson, A. M.; Suresh, S.; Iyer, H.; Srinivasa, B. R.;
Bergfors, T.; Unge, T.; Mowbray, S. L.; Larhed, M.; Jones, T. A.;
́
Karlen, A. Design, synthesis, and X-ray crystallographic studies of α-
aryl substituted fosmidomycin analogues as inhibitors of Mycobacte-
rium tuberculosis 1-deoxy-D-xylulose 5-phosphate reductoisomerase. J.
Med. Chem. 2011, 54, 4964−4976.
̈
(23) Behrendt, C. T.; Kunfermann, A.; Illarionova, V.; Matheeussen,
A.; Pein, M. K.; Grawert, T.; Kaiser, J.; Bacher, A.; Eisenreich, W.;
̈
Illarionov, B.; Fischer, M.; Maes, L.; Groll, M.; Kurz, T. Reverse
fosmidomycin derivatives against the antimalarial drug target IspC
(Dxr). J. Med. Chem. 2011, 54, 6796−6802.
(8) Dhiman, R. K.; Schaeffer, M. L.; Bailey, A. M.; Testa, C. A.;
Scherman, H.; Crick, D. C. 1-Deoxy-D-xylulose 5-phosphate
reductoisomerase (IspC) from Mycobacterium tuberculosis: towards
understanding mycobacterial resistance to fosmidomycin. J. Bacteriol.
2005, 187, 8395−8402.
(24) Haemers, T.; Wiesner, J.; Van Poecke, S.; Goeman, J.;
Henschker, D.; Beck, E.; Jomaa, H.; Van Calenbergh, S. Synthesis of
α-substituted fosmidomycin analogues as highly potent Plasmodium
falciparum growth inhibitors. Bioorg. Med. Chem. Lett. 2006, 16, 1888−
1891.
(25) Verbrugghen, T.; Cos, P.; Maes, L.; Van Calenbergh, S.
Synthesis and evaluation of a-halogenated analogues of 3-
(acetylhydroxyamino)propylphosphonic acid (FR900098) as antima-
larials. J. Med. Chem. 2010, 53, 5342−5346.
(26) Giessmann, D.; Heidler, P.; Haemers, T.; Van Calenbergh, S.;
Reichenberg, A.; Jomaa, H.; Weidemeyer, C.; Sanderbrand, S.;
Wiesner, J.; Link, A. Towards new antimalarial drugs: synthesis of
non-hydrolyzable phosphate mimics as feed for a predictive QSAR
study on 1-deoxy-D-xylulose-5-phosphate reductoisomerase inhibitors.
Chem. Biodiversity 2008, 5, 643−656.
(27) Puyau, P. D.; Perie, J. J. Synthesis of substrate analogues and
inhibitors for the phosphoglycerate mutase enzyme. Phosphorus, Sulfur,
Silicon Relat. Elem. 1997, 129, 13−45.
(28) Kurz, T.; Geffken, D.; Wackendorff, C. Hydroxyurea analogues
of fosmidomycin. Z. Naturforsch., B 2003, 58, 106−110.
(29) Suresh, S.; Shyamraj, D.; Larhed, M. Synthesis of antimalarial
compounds fosmidomycin and FR900098 through N- or P-alkylation
reactions. Tetrahedron 2013, 69, 1183−1188.
(30) Lindh, J.; Enquist, P. A.; Pilotti, Å.; Nilsson, P.; Larhed, M.
Efficient palladium(II) catalysis under air. Base-free oxidative Heck
reactions at room temperature or with microwave heating. J. Org.
Chem. 2007, 72, 7957−7962.
(31) Nordqvist, A.; Bjorkelid, C.; Andaloussi, M.; Jansson, A. M.;
́
Mowbray, S. L.; Karlen, A.; Larhed, M. Synthesis of functionalized
cinnamaldehyde derivatives by an oxidative Heck reaction and their
(9) Henriksson, L. M.; Bjorkelid, C.; Mowbray, S. L.; Unge, T. The
̈
1.9 Å resolution structure of Mycobacterium tuberculosis 1-deoxy-D-
xylulose 5-phosphate reductoisomerase, a potential drug target. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2006, 62, 807−813.
(10) Henriksson, L. M.; Unge, T.; Carlsson, J.; Åqvist, J.; Mowbray,
S. L.; Jones, T. A. Structures of Mycobacterium tuberculosis 1-deoxy-D-
xylulose-5-phosphate reductoisomerase provide new insights into
catalysis. J. Biol. Chem. 2007, 282, 19905−19916.
(11) Brown, A. C.; Parish, T. Dxr is essential in Mycobacterium
tuberculosis and fosmidomycin resistance is due to a lack of uptake.
BMC Microbiol. 2008, 8, 78.
(12) Lell, B.; Ruangweerayut, R.; Wiesner, J.; Missinou, M. A.;
Schindler, A.; Baranek, T.; Hintz, M.; Hutchinson, D.; Jomaa, H.;
Kremsner, P. G. Fosmidomycin, a novel chemotherapeutic agent for
malaria. Antimicrob. Agents Chemother. 2003, 47, 735−738.
(13) Borrmann, S.; Adegnika, A. A.; Matsiegui, P. B.; Issifou, S.;
Schindler, A.; Mawili-Mboumba, D. P.; Baranek, T.; Wiesner, J.;
Jomaa, H.; Kremsner, P. G. Fosmidomycin−clindamycin for
Plasmodium falciparum infections in African children. J. Infect. Dis.
2004, 189, 901−908.
(14) Perruchon, J.; Ortmann, R.; Altenkamper, M.; Silber, K.;
Wiesner, J.; Jomaa, H.; Klebe, G.; Schlitzer, M. Studies addressing the
importance of charge in the binding of fosmidomycin-like molecules to
deoxyxylulosephosphate reductoisomerase. ChemMedChem 2008, 3,
1232−1241.
́
(15) Zingle, C.; Kuntz, L.; Tritsch, D.; Grosdemange-Billiard, C.;
Rohmer, M. Isoprenoid biosynthesis via the methylerythritol
phosphate pathway: structural variations around phosphonate anchor
use as starting materials for preparation of Mycobacterium tuberculosis
I
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1-deoxy-D-xylulose-5-phosphate reductoisomerase inhibitors. J. Org.
Chem. 2011, 76, 8986−8998.
(32) Baniecki, M. L.; Wirth, D. F.; Clardy, J. High-throughput
Plasmodium falciparum growth assay for malaria drug discovery.
Antimicrob. Agents Chemother. 2007, 51, 716−723.
(33) Leslie, A. The integration of macromolecular diffraction data.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2006, 62, 48−57.
(34) Evans, P. Scaling and assessment of data quality. Acta Crystallogr.
Sect. D: Biol. Crystallogr. 2006, 62, 72−82.
(35) CCP4. The CCP4 suite: programs for protein crystallography.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 1994, 50, 760−763.
(36) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M.
D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl.
Crystallogr. 2007, 40, 658−674.
(37) 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.
(38) Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard, M. Improved
methods for building protein models in electron density maps and the
location of errors in these models. Acta Crystallogr., Sect. A: Found.
Crystallogr. 1991, 47, 110−119.
(39) Harding, M. Small revisions to predicted distances around metal
sites in proteins. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2006, 62,
678−682.
(40) Kleywegt, G. J.; Jones, T. A. Detecting folding motifs and
similarities in protein structures. Methods in Enzymology; Academic
Press Inc: San Diego, 1997; Vol. 277, p 525−545.
(41) Harris, M.; Jones, T. A. Molraya web interface between O and
the POV-Ray ray tracer. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2001, 57, 1201−1203.
(42) Kleywegt, G. J.; Harris, M. R.; Zou, J. Y.; Taylor, T. C.; Wahlby,
A.; Jones, T. A. The Uppsala electron-density server. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 2004, 60, 2240−2249.
(43) Andaloussi, M.; Lindh, M.; Bjorkelid, C.; Suresh, S.;
́
Wieckowska, A.; Iyer, H.; Karlen, A.; Larhed, M. Substitution of the
phosphonic acid and hydroxamic acid functionalities of the DXR
inhibitor FR900098: An attempt to improve the activity against
Mycobacterium tuberculosis. Bioorg. Med. Chem. Lett. 2011, 21, 5403−
5407.
(44) Silber, K.; Heidler, P.; Kurz, T.; Klebe, G. AFMoC enhances
predictivity of 3D QSAR: A case study with DOXP-reductoisomerase.
J. Med. Chem. 2005, 48, 3547−3563.
J
dx.doi.org/10.1021/jm4006498 | J. Med. Chem. XXXX, XXX, XXX−XXX