selective inhibitors of pfldh

Azole-based inhibitors of P. falciparum LDH

1

Identification and activity of a series of azole-based compounds with

lactate dehydrogenase-directed anti-malarial activity.

Angus Cameron1, Jon Read1, Rebecca Tranter1, Victoria J. Winter1, Richard B. Sessions1,

R. Leo Brady1#, Livia Vivas2, Anna Easton2, Howard Kendrick2, Simon L. Croft2, David

Barros3, Jose Luis Lavandera3, José Julio Martin3, Felix Risco3, Silvestre García-Ochoa3,

Fracisco Javier Gamo3, Laura Sanz3, Luisa Leon3, Jose R. Ruiz3, Raquel Gabarró3, Araceli

Mallo3 and Federico Gómez de las Heras3.

1Department of Biochemistry and Molecular Recognition Centre, University of Bristol,

Bristol BS8 1TD U.K..

2London School of Hygiene & Tropical Medicine, Keppel Street, London WC1E 7HT,

UK.

3GlaxoSmithKline, Parque Tecnológico de Madrid, Severo Ochoa, 2, 28760 – Tres

Cantos, Madrid, Spain

#Corresponding author:

R. Leo Brady

Tel: 44-117-928 7436

Fax: 44-117-928 8274

Email: [email protected]

Running title: Azole-based inhibitors of P. falciparum LDH

JBC Papers in Press. Published on April 26, 2004 as Manuscript M402433200

Copyright 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

mailto:[email protected]

http://www.jbc.org/


2

*This work was in part supported by a grant from the Medicines for Malaria Venture

(MMV, Geneva).

The atomic coordinates and structure factors for the crystal structures described are

available in the Research Collaboratory for Structural Bioinformatics Protein Databank

under the accession codes PDB # 1T24, 1T25, 1T26, 1T2C, 1T2D, 1T2E and 1T2F.

The abbreviations used are:

LDH: lactate dehydrogenase

pfLDH: Plasmodium falciparum lactate dehydrogenase

hsLDH: human lactate dehydrogenase

HTS: High-Throughput Screen

TCA: Tricarboxylic acid cycle


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


3

Summary

Plasmodium falciparum, the causative agent of malaria, relies extensively on glycolysis

coupled with homolactic fermentation during its blood-borne stages for energy production.

Selective inhibitors of the parasite lactate dehydrogenase (LDH), central to NAD+

regeneration, therefore potentially provide a route to new anti-malarial drugs directed

against a novel molecular target. A series of heterocyclic, azole-based compounds are

described that preferentially inhibit P. falciparum LDH at sub-micromolar concentrations,

typically at concentrations of about 100-fold lower than required for human lactate

dehydrogenase inhibition. Crystal structures show these competitive inhibitors form a

network of interactions with amino acids within the active site of the enzyme, stacking

alongside the nicotinamide ring of the NAD+ co-factor.

These compounds display modest activity against parasitized erythrocytes, including

parasite strains with known resistance to existing anti-malarials, and against P. berghei in

BALB/c mice. Initial toxicity data suggest the azole derivatives have generally low

cytotoxicity, and preliminary pharmocokinetic data show favourable bioavailability and

circulation times. These encouraging results suggest further enhancement of these

structures may yield candidates suitable for consideration as new therapeutics for the

treatment of malaria. In combination these studies also provide strong support for the

validity of targeting the Plasmodium glycolytic pathway and, in particular, LDH in the

search for novel anti-malarials.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


4

Introduction

Plasmodium parasites are believed to lack a functional Krebs (citric acid) cycle for at least

part of their life cycle, and hence rely extensively on ATP generation via the anaerobic

fermentation of glucose (see (1) for review). The energy requirement of the parasitised

erythrocyte is such that utilization of glucose is up to 100 times greater than in non-

parasitised erythrocytes (2,3), and virtually all glucose can be accounted for by production

of lactate (2). Lactate dehydrogenase (LDH), the last enzyme in the glycolytic pathway in

P. falciparum, is a 2-hydroxy acid oxidoreductase that converts pyruvate to lactate, and

simultaneously the conversion of NADH to NAD+. As a constant supply of NADH is a

prerequisite for glycolysis, and LDH acts as the primary source in Plasmodium for the

regeneration of NADH from NAD+, inhibition of LDH is expected to stop production of

ATP, with subsequent P. falciparum cell death. Any compound that blocks the LDH

enzyme is a potentially potent antimalarial with a different mode of action to existing

drugs. As such, P. falciparum lactate dehydrogenase (pfLDH) has been suggested as a

drug target by several authors (4-6). One well-recognised difficulty is that the drug must

potently inhibit pfLDH yet show much less activity against the three human LDH

(hsLDH) isoforms.

A comparison of the crystal structures of both pf and human LDH (7,8) shows two key

differences: namely, positioning of the NADH factor – reflecting sequence changes in the

cofactor binding pocket that displace the nicotinamide ring by about 1.2 Å - and a change

in the sequence (including a 5-residue insertion) and secondary structure of a loop region

that closes down on the active site during catalysis. These changes combine to produce an

increase in the volume of the active site cleft in pfLDH relative to its human counterparts.

In addition to these structural variations, there are significant kinetic differences between


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


5

the two enzymes; indeed, using the NADH derivative APAD, kinetic differences between

human and pfLDH are so great that the observed LDH activity can be used as an

indication of in vivo parasitemia (9). Together, the structural and kinetic discrepancies

between the mammalian and malarial enzymes suggest that specific and potent pfLDH

inhibitors can be designed or identified.

Several groups are known to have targeted pfLDH in drug discovery studies. Derivatives

of the gossypols have been considered as pfLDH inhibitors. Gossypol is a polyphenolic

binaphthyl disequiterpene found in cottonseed oil, has been shown to inhibit LDHs at sub-

micromolar (0.7 µM) levels (10), is competitive for NADH, and exhibits in vitro anti-

malarial activity (4) with an IC50 of 10 µM. However, gossypol is cytotoxic. Attempts to

derivatize gossypol have produced several compounds retaining activity against both the

target enzyme and the parasite, but without significantly improved selectivity or

parasiticidal activity. The synthesis of derivatives of 8-deoxyhemigossylic acid has also

been reported (11). This class of chemicals has been developed in an attempt to reduce the

toxicity believed to be associated with the aldehyde group present in gossypol, and to

increase the specificity of inhibition. The hemi-gossypols exhibit low micromolar

inhibition of LDHs, are competitive for NADH and some of these compounds are

selective inhibitors of pfLDH with respect to human LDH (11). One such compound, 7-p-

trifluoromethylbenzyl-8-deoxyhemigossylic acid, has a reported Ki of 13, 81 and 4 µM

respectively for human heart, muscle and sperm LDH (12) and 0.2 µM for pfLDH (4).

Vander Jagt’s group have also investigated various N-substituted hydroxamic acid

derivatives and reported 35 - 80 µM activity vs. pfLDH (13) and 0.3 - 10 mM activity vs.

human LDH’s (12). Modifications do not appear to significantly affect potency or


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


6

selectivity of these compounds, which appear to be competitive for pyruvate. No

parasiticidal activity has been reported for these compounds.

In this paper we report the discovery of a new class of compounds that inhibit pfLDH and

also display anti-malarial activity. Initially identified in a high-throughput enzymatic

assay, these compounds have been shown to interact directly and preferentially with

pfLDH through X-ray crystallographic and steady-state kinetic analyses. They have been

further characterised in parasiticidal whole cell assays using drug sensitive and resistant

strains of Plasmodium, and demonstrated to have in vivo anti-malarial activity using the P.

berghei rodent model. In combination these results help demonstrate the viability of

targeting pfLDH in the development of novel anti-malarials, and provide examples of

compounds that could be further developed to provide novel therapeutics targeting the

Plasmodium glycolytic pathway.

Experimental Procedures

High throughput enzymatic screen - An LDH enzymatic assay developed for high-

throughput format was used in a High-Throughput Screen (HTS). The dehydrogenase

reaction was run in the reverse (lactate pyruvate) direction and coupled with the ability

of diaphorase to reduce p-iodonitrotetrazolium violet using the NADH generated in the

conversion of lactate to pyruvate (14). The progression of the coupling reaction was

monitored as the increase of absorbance at 492 nm. In the initial screen, potential

inhibition of both pfLDH and human LDH (both expressed as recombinant proteins and

purified as described in (7)) was monitored at single sample points corresponding to 25


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


7

µg/mL of the compound in 5 % DMSO; 0.15 mM NAD+; 1.5 mM lactic acid; 1 mM INT;

18 µg/mL Diaphorase; and 1 µg/mL pfLDH or human LDH. Positive hits were subjected

to additional analysis to determine IC50 values.

Synthesis of Azole Derivatives - The parent compounds of the isooxazole and oxadiazole

families were prepared in multi-gram scale by simple modification of the methods

described in the literature (15). A range of azoles was synthesised by the introduction of

substituents at the hydroxyl and acid moieties of the parent compounds (positions 3 and 4

respectively, Fig. 1). The replacement of heteroatoms within the ring structure was also

considered and the details of the modifications studied are all included in Table 1.

Synthetic routes for these derivatives are described in the Supplementary Material.

Crystallographic analysis of enzyme:azole inhibitory complexes - Crystals of pfLDH and

human LDH were grown using either NADH or NAD+ as cofactor as described in (8) and

(7) respectively. Ligands were introduced to these crystals dissolved in the crystallisation

mother liquor substituted with up to 30 % DMSO. Diffraction data were collected at the

Daresbury SRS synchrotron (station PX14.1), Hamburg DESY synchrotron (station X11)

or using the Nonius® FR591 rotating anode laboratory source (Nonius BV, Netherlands)

and processed using the HKL suite of programs (16). Structures were solved using the

phases from the isomorphous structure of the ternary complex of pfLDH (PDB accession

code: 1LDG) or human LDH (PDB accession code: 1IOZ) and refined using the program

REFMAC5 (17).

Kinetic analysis - Whereas the dehydrogenase reaction was run in the reverse (lactate

pyruvate) direction for the HTS, a more thorough kinetic analysis of selected inhibitors


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


8

was based on the reaction run in the forward direction and monitored by measuring the

change in molar absorbance of NADH at 340 nm as described in (18). For kinetic analysis

in the reverse direction, reactions containing 2 mM phenazine ethosulfate, 1 mg/ml p-

nitrotetrazolium blue, 1 mM NAD+, and varying concentrations of lactate in PET buffer

(50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM EDTA, 3 % (w/v) PEG 6000) were

initiated by the addition of pfLDH to 4.4 nM and monitored by the increase in absorbance

at 655 nm at 25 oC. Data were analysed using non-linear least squares regression with the

software package software package DynaFit(tm) (Biokin Ltd) (19). Calculations of kcat

depended upon the protein concentration as measured at 280 nm using an extinction

coefficient of 1.16 mg/ml.cm for H4-hsLDH proteins, 1.2 mg/ml.cm for M4-hsLDH (20)

and 0.5 mg/ml.cm for pfLDH.

The pKa of the active site histidine was estimated by fitting the variation of KMobs with pH

to KMobs = KM[1 + 10-pK/10-pH] as described in (7).

In vitro anti-plasmodial activity - The drug sensitive 3D7 clone of the NF54 isolate (21)

and the chloroquine, pyrimethamine and cycloguanil resistant K1 strain (Thailand) were

acquired from MR4 (Malaria Research and Reference Reagent Resource Center,

Manassas, Virginia, USA). P. falciparum in vitro culture was carried out following

standard methods (22) with modifications. Briefly, parasites were maintained in tissue

culture flasks in human A Rh+ erythrocytes at 5 % haematocrit in RPMI-1640

supplemented with 25 mM HEPES, 24 mM NaHCO3, 0.2 % (w/v) glucose, 0.03 % L-

glutamine, 150 µM hypoxanthine, and 0.5 % Albumax II ® (Gibco, UK) in a 5 % CO2 /

95 % air mixture at 37 ºC and the medium was changed daily.

The method used to test drug susceptibility was modified from the protocol described by

(23). Briefly, stock drug solutions were dissolved in 100 % DMSO in glass bottles and


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


9

serial dilutions of the drugs were prepared in assay medium (RPMI 1640 supplemented

with 0.5 % Albumax II ® (Life Technologies Inc, USA), 0.2 % w/v glucose, 0.03 % L-

glutamine and 15 µM hypoxanthine) in triplicate, in 96-well plates. This was followed by

the addition of 50 µl asynchronous (65-75 % ring stage) P. falciparum culture (0.5 %

parasitemia) or uninfected erythrocytes at 5 % haematocrit to each well in assay medium.

Plates were incubated at 37 °C, in 5 % CO2 / 95 % air mixture for 24 h, followed by the

addition of 20 µl (0.1 µCi/well) of [3H] hypoxanthine to each well. Plates were mixed for

1 minute using a plate shaker and returned to the incubator. After an additional 24 h

incubation period, the experiment was terminated by placing the plates in a –80 oC freezer.

Plates were thawed and harvested onto glass fibre filter mats using a cell harvester

(Tomtec, USA) and dried. After the addition of Meltilex® solid scintillant (Wallac), the

incorporated radioactivity was counted using a Wallac® 1450 Betaluxscintillation counter

(Wallac Oy). All assays included chloroquine diphosphate as a standard and control wells

with untreated infected and uninfected erythrocytes. Data acquired by the Wallac®

BetaLux scintillation counter were exported into a MICROSOFT® EXCEL

spreadsheet (Microsoft Corporation) and the IC50s / IC90s of each drug calculated using

XLFit® (ID Business Solutions Ltd, UK) line fitting software.

In vivo antimalarial activity –A preliminary experiment was undertaken at a single dose

to evaluate the activity of two leading compounds from the heterocyclic series.

Compounds OXD1 and IOA1 (see Table 1 for structures) were tested in the P.berghei

model using the Peters 4-day suppressive test (24) using chloroquine as a positive control.

Briefly, naïve 18-20 g BALB/C mice were infected intravenously with 2x106 parasitized

red cells on D+0. For administration, compounds were freshly prepared in 10 % DMSO in

sterile PBS the day of use. Two hours post-infection mice received the first treatment by


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


10

the intraperitoneal route. Mice were further treated on D+1 – 3. Blood films from tail

blood were prepared on D+4 and parasitemia determined by microscopic examination of

Giemsa stained blood films. A further experiment was carried out to determine oral

bioavailability and the ED50 of the compounds in an in vivo dose response experiment.

Compounds OXD1, IOA1 and TDA1 were tested at 100, 50, 25 and 12.5 mg/kg/day by

the intraperitoneal route and at 100 mg/kg/day by the oral route. Chloroquine was used as

a positive control at 10 mg/kg/day by the intraperitoneal route. Mice were treated and

levels of parasitemia determined as described for the single dose experiment.

Pharmacokinetic studies -A preliminary study to assess the plasma levels of OXD1 and

IOA1 was performed by intraperitoneal or oral administration of each compound to mice

inoculated two hours earlier with 1 x 107 P. berghei parasitised erythrocytes. Plasma

samples were collected from 3 mice at each time point: 30, 90 and 180 min post

compound administration. Plasma concentrations of OXD1 and IOA1 were determined

by peak integration after separation using standard LC-MS analysis techniques.

In vitro cytotoxicity assays – Cytotoxicity of OXD1 against mammalian cells was

assessed by standard methods (25). Sterile 96-well microtiter plates were seeded with

100 µl of KB cells at 4 x 104 / ml in RPMI 1640 supplemented with 10 % heat inactivated

foetal calf serum (complete medium). Plates were incubated at 37 °C, 5 % CO2 / 95 % air

mixture for 24 h. Compounds were dissolved in 100 % DMSO at 20 mg/ml. For the

assays, serial dilutions of the compounds were prepared in triplicate using complete

medium. Culture supernatants were removed and replaced by the serial dilutions at 300,

30, 3 and 0.3 µg/ml. The positive control drug was podophyllotoxin (Sigma, UK). Plates

were incubated for a further 72 h followed by the addition of 10 µl of AlamarBlue®


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


11

(AccuMed International Inc, USA) to each well and incubation for 2-4 h at 37 °C, 5 %

CO2 / 95 % air mixture before reading at EX/EM 530/585 nm (cut-off 550nm) in a

SPECTRAMAX® GEMINI plate reader (Molecular Devices Inc, USA). ED50 values were

calculated using XLFit® (ID Business Solutions Ltd, UK) line fitting software.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


12

RESULTS

High throughput enzymatic screen - Screening of over 500,000 compounds primarily

from the Glaxo Wellcome proprietary chemical collection identified a number of chemical

families of interest that had a pfLDH IC50 between 0.05 – 0.01 µg/ml and a selectivity

ratio pfLDH/hsLDH IC50 > 10. One of these families consisted of compounds that will be

collectively referred to as the azoles (Fig. 1). All three parent compounds had a consistent

substitution pattern with a hydroxy function and a carboxy moiety in adjacent positions

(3,4-disubsitituion).

Synthesis and activity of azole derivatives - A series of compounds was synthesised with

the aim of exploring the structure activity relationships of the parent azoles. As shown in

Tables 1A-D, we tested examples of derivatives at all five positions of the pentacyclic

ring, as well as heterocyclic variations within the ring structure itself. In summary,

modifications to, or substitution of, the C3 hydroxyl and C4 carboxyl groups resulted in

the loss of at least two orders of magnitude of activity against pfLDH, none of the ethers,

esters or bioisosters of the hydroxyl or acid moieties were active. Modifying the

heterocyclic nature of the ring by substituting the oxygen at position 1 with sulfur

(resulting in the thiadiazole family, Table 1C) increased the potency slightly, but also led

to a slight reduction in specificity. Substitution at the same position with nitrogen (leading

to the triazoles, Table 1D) was non-productive. The rationale for the wide variations in

potency when modifying position 1 could be rationalised as arising from either steric

hindrance (as the presence of nitrogen in position 1 introduces an extra substituent in the

ring) or electronic effects (as the presence of the nitrogen modifies the pKa of the hydroxyl

group). As will be shown later, the acidity of the 2-hydroxyl group is a key factor that


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


13

determines the types of interactions established by the substrate into the active site.

Activity is retained if the ring nitrogens at positions 2 and 5 are substituted with carbon to

produce the isooxazoles (Table 1B), but substitution at position 2 is less well tolerated.

Despite this, the introduction of methyl groups at either of these positions reduces activity

to greater than 100 µM.

Crystallographic analysis of enzyme:azole inhibitory complexes - A summary of the

diffraction data and refinement statistics for complexes of pfLDH with OXD1, TDA1 and

IOA1, and human LDH with OXD1, is shown in Table 2.

The hydroxyacid heterocyclic azole OXD1 binds in the active site of both human and

plasmodial LDH, alongside the NAD+ co-factor (Figs. 2a and 2b). The carboxyl acid

group of the inhibitor forms salt bridges with Arg171 and Arg109 of the active site loop,

mimicking the interactions of the pyruvate substrate (8). The hydroxyl group of the

oxadiazole contributes to a hydrogen bond network with the side chains of residues

Asn140, His195 and Arg109. In both the human and plasmodial crystal structures the

heterocyclic ring of the inhibitor is stacked parallel to the nicotinamide ring of the NAD+

co-factor, while the nitrogen in position 2 of the heterocyclic ring is within Van der Waals

radii of the NAD+ and the side chain amine of Asn140. In the plasmodial structure the

side of the heterocyclic ring harbouring the carboxyl group is in close proximity to Pro246

(threonine in hsLDH), and the Oγ of Ser245 forms a hydrogen bonds with both the ring

oxygen of the inhibitor and the cofactor (via a bound water molecule). In the human LDH

structure, the ring oxygen of the inhibitor is hydrogen bonded to a water molecule, but this

water appears to make no contacts with the protein itself.

The closely related inhibitors TDA1 (Table 1C) and IOA1 (Table 1B) bind in a manner

virtually identical to that of OXD1, although there is small offset of <1Å in the ring


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


14

position when measured at position 1 (Fig. 3). Apart from this offset, all other features of

the complex - including the surrounding protein residues- are essentially identical to those

seen in the pfLDH:OXD1 structure. In all three complexes, the oxygen or sulfur at

position 1 forms a hydrogen bond with the side chain of Ser 245. This interaction is absent

in the crystal structure of OXD1 bound to human LDH, where the equivalent residue is a

tyrosine but projects away from the bound inhibitor. Ser 245 and Pro 246 adopt different

conformations when neither substrate nor inhibitor is bound, as seen in the crystal

structure of the binary complex (Fig. 4). The adjacent ‘active site loop’ is also disordered

in this binary complex as is commonly observed for this class of oxidoreductase enzymes.

Kinetic analysis- We used steady state kinetics to examine the mode of binding of the

azole inhibitor family. All active azoles exhibited mixed inhibition against both NADH

and pyruvate, and competitive inhibition against lactate (Fig. 6). Kinetically, LDH

behaves as a sequential ordered bi-bi enzyme with the cofactor binding prior to substrate

and being released after the product. The azoles hence bind preferentially to the

enzyme:NAD+ complex, mimicking the non-specific inhibitor oxalate (26). The measured

inhibitory constants (Ki) were 210 nM for OXD1, 470 nM for IOA1 and 290 nM for

TDA1.

PfLDH Ser245Ala mutant – Crystal structures of the azoles bound to pfLDH suggested

the hydrogen bond formed between the oxygen or sulphur in position 1 of the azole and

the Oγ of Ser 245 was likely to be a major determinant of selectivity for the malarial form

of the enzyme. In order to test this hypothesis, we used site directed mutagenesis to

change the side chain at position 245 to alanine. Co-crystallisation of the pfLDH Ser 245


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


15

Ala mutant with oxamate (a substrate analogue) and NADH shows that the “active site

loop” occupies the “closed” position previously seen in the wild type ternary complex, but

the Ala 245-Pro 246 region is in the “open” conformation normally seen in the binary

(enzyme:cofactor) complex (Fig 6). In the structure of the wild type enzyme co-

crystallised with NADH and oxamate, the Oγ of Ser 245 forms a hydrogen bond with a

water molecule that is in turn bonded to the cofactor ribose group. This water molecule is

not present in the crystal structure of the mutant. The inability of the alanine side chain in

the mutant to interact with the ribose group appears to correlate with the Ala 245-Pro 246

region remaining in the “open” conformation, in turn probably explaining the reduced

activity of this mutant enzyme (Table 3). The critical interaction formed between the

serine in wild-type pfLDH and the azole inhibitors raises the possibility that resistance to

this class of inhibitor could develop by mutating this single residue. Nonetheless, it is

reassuring that the mutant enzyme produced is kinetically crippled, with a kcat over two

orders of magnitude lower than the wild type. This implies that mutation of the serine is

likely to be fatal to the plasmodium. Table 3 suggests the major reason for low activity in

the absence of the serine is the reduced ability of the mutant enzyme to bind pyruvate

(KM.WT/KM.S245A = 114) rather than NADH (KM.WT/KM.S245A = 4.1). One explanation is

that the optimal binding site for the substrate may result from subtle rearrangements that

follow binding of the cofactor, and may involve the Ser 245 – Pro 246 region.

In vitro anti-plasmodial activity - The anti-plasmodial activity of a subset of the azole

series was determined against the P. falciparum 3D7 drug sensitive clone and the K1 drug

resistant strain, and the measured IC50 / IC90 values are included in Tables 1A-D. Only a

small number of the compounds had demonstrated activity within the range of

concentrations measured against whole cells, although this is likely in part to reflect the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


16

limited potency observed with the enzyme. The whole cell activity results broadly

correlated with the IC50 values determined against the pfLDH enzyme.

In vivo antimalarial activity and toxicity - In the Peters 4-day suppressive test, both

compounds OXD1 and IOA1 showed significant inhibition compared to untreated control

mice. The percentage of inhibition observed was 41 % for OXD1 and 30 % for IOA1

compared with untreated controls, whereas chloroquine showed 100% inhibition. Figure 9

illustrates the inhibitory activity observed in a dose response experiment using the

oxazoles OXD1 and IOA1 and the thiazole TDA1. P. berghei has previously been shown

to contain a single gene for LDH, the product of which is kinetically and structurally

virtually identical to pfLDH (18).

Pharmacokinetic studies of OXD1 and IOA1 - Plasma concentration time curves (AUC)

are shown for OXD1 and IOA1 in Fig 10. After a single intravenous dose of 5 mg/kg, the

pharmacokinetic parameters evaluated for each compound were similar. For IOA1 and

OXD1, both the AUC0-8 and elimination half-life (t1/2) were 4.8 and 4.6 µg.h/ml and 0.38

and 0.39 h, respectively.

Absorption of IOA1 and OXD1 after intraperitoneal administration of 20 mg/kg was very

rapid, with peak plasma levels of 15.3 and 14.6 µg/ml respectively, reached within 25

min. However, the t1/2 of intraperitoneal administration of IOA1 (0.56 h.) was

approximately three times lower than that of OXD1 (1.46 h).

The time courses of drug levels in plasma for OXD1, IOA1, and TDA1 in mice after

intraperitoneal or oral administration (dose, 100 mg/kg) are shown in Table 4. All


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


17

compounds tested - OXD1, IOA1, and TDA1 - showed consistently high plasma levels

after oral administration with peak levels of 17.9, 19.8 and 35.2 µg/ml, respectively,

measured 30 minutes post-administration. Plasma peak levels obtained after

intraperitoneal administration of OXD1, IOA1, and TDA1 (25.5, 20.4, and 41.0 µg/ml,

respectively) were similar to results obtained after oral administration.

In vitro cytotoxicity assays - No overt toxicity was associated with OXD1 when tested

against the mammalian KB (ED50 = 0.64 mM) and HeLa (ED50 = 1.3 mM) cell lines.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


18

Discussion

Following early reports of enhanced rates of glycolytic turnover in Plasmodium-infected

erythrocytes, and biochemical data suggesting the absence of a complete citric acid cycle

in at least the blood-borne stages of the parasite lifecycle, it was recognised that

homolactic fermentation – the conversion of pyruvate to lactate with concomitant

reduction of NADH – plays a special role in survival of the malarial parasite. Although

sequence (27) and structural (7,8,18) differences have been noted between the plasmodial

and human forms of the enzymes that support the theoretical viability of targeting pfLDH

for anti-malarial activity, there have been only a very limited number of compounds

reported as specific inhibitors of this enzyme (4,10,11,13,28), and none readily amenable

to simple synthesis or free from toxicological problems. The azoles described in this

paper form the first example of compounds that are readily synthetically accessible, form

specific inhibitors of pfLDH and have good drug-like properties.

Each of the three parent compounds – the oxadiazole OXD1, the isoxazole IAO1, and the

thiadiazole TDA1 – were initially discovered through an enzyme assay-based high-

throughput screen, and inhibit pfLDH with low or sub-micromolar IC50 values, in contrast

to the ~ 50 µM IC50 values observed with human LDH. The data presented in this paper

provide a detailed and consistent description of the interaction of these compounds with

their enzyme target.

The molecular basis of this inhibition is illustrated by the crystal structures of the

enzyme:inhibitor complexes. In each case, the azole inhibitor binds directly within the

active site of the enzyme, effecting closure of the active site loop which normally makes

direct interactions with the keto-acid (pyruvate) substrate. The concerted interactions of


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


19

the substrate with this loop and its subsequent closure have previously been shown to be a

key factor in the exquisite selectivity of the lactate and, particularly, malate

dehydrogenase enzymes (29). Replication of this complementarity with both the open and

closed form of the enzyme provides a powerful means of generating inhibitor selectivity.

Nonetheless, although the moderate level of activity of these compounds is encouraging,

higher potency needs to be achieved to use these compounds as therapeutics. The

relatively small size (MW ~ 140) of the azoles suggests there is plenty of scope to enlarge

these structures to enhance their contacts with the enzyme and hence increase the affinity

of binding. However, their almost complete burial within the closed active site makes this

problematic.

From the crystal structures a number of common interactions can be identified. The

carboxylic acid group of each inhibitor forms a bifurcated salt bridge with Arg171 and

Arg109, mimicking the interaction with the pyruvate substrate as has previously been

documented (8) from ternary complexes of pfLDH with oxamate – a substrate analogue.

The hydroxyl group of each azole forms a common set of hydrogen bonds with Leu140,

His195 and Arg109. Taken together these interactions explain why the hydroxyl-carboxyl

motif is vital for the activity of these inhibitors, and why extending the inhibitor from the

carboxyl or hydroxyl groups consistently results in a reduction or complete loss of activity

(Table 1). These interactions also support the proposition that for activity both the acid

and hydroxyl groups must have suitable pKa values to leave these groups ionised in vivo.

The heterocyclic ring of each inhibitor stacks parallel to the nicotinamide ring of the

NAD+ co-factor, separated by about 2 Å where π- π cloud stacking interactions would

contribute to ligand binding. Similar interactions have also been noted in crystal


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


20

structures of other anti-microbials active against plasmodial NADP(H)-dependent

enzymes, such as the NADPH-dependent dihydrofolate reductase (30), and Triclosan

binding to the parasite NADH-dependent enoyl acyl-carrier protein reductase (31).

Pro246 lies in adjacent to the carboxylate function of the heterocyclic ring, leaving little

space to extend these inhibitors from the nitrogen in position 5 of the ring. This is

consistent with the decrease in activity associated with the introduction of substituents at

this position (Table 1). Even the introduction of a simple methyl group at position 5,

which the structures suggests might profitably make van der Waals contacts with the

proline, is not tolerated (IOA3, Table 1B). The nitrogen in position 2 of the heterocyclic

ring is already within Van der Waals radii of the NAD+ and Asn126; and extending the

inhibitor in this direction is unlikely to result in tighter binding. Hence, from the crystal

structures, extending the ring from position 1 is seen as the modification most likely to

result in new interactions.

Importance of Ser 245

In the oxadiazole structures, the oxygen at position 1 forms a hydrogen bond with the side

chain of Ser 245 (Fig 2a). This interaction is likely to account for much of the selectivity

of these compounds for the malarial rather than human forms of LDH, as in the latter this

serine is absent – replaced by a tyrosine that is oriented away from the active site (Fig 2b).

Extending the azole moiety from this position would require movement of the serine side

chain from its current ‘in’ position. We noted that this was indeed possible, as in the

crystal structure of the binary complex (enzyme + co-factor, Fig 4), where the active site

loop is disordered, the Ser 245 / Pro 246 main chain segment adopts a more open

conformation that we termed the ‘out’ position. With the serine in this conformation, it

appears there is plenty of scope to extend the azole inhibitors from the 1-position of the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


21

ring to increase contacts with residues distal from serine 245. However, as mentioned

above, binding of the azole inhibitors is characterised by the formation of hydrogen bonds

between arginines 109 and 171 from the enzyme with the carboxy and hydroxy groups

seen as essential to this class of inhibitor. As arginine 109 is part of the active site loop,

the loop must therefore be closed for these bonds to be formed. We observed that in most

structures of ternary or inhibitor complexes of pfLDH closure of the active site loop was

accompanied by movement of the Ser 245-Pro 246 loop to the “in” position. In order to

probe whether this apparently concerted movement was essential for, or subsidiary to,

active site loop closure we prepared a mutant form of pfLDH with Ser 245 changed to

alanine.

We assessed the kinetics and crystal structure (in the presence of NADH and oxamate) of

this mutant (S245A) form of pfLDH. As expected, the only significant differences seen

between the mutant and wild type structures are in the immediate region of the Ser 245

Ala mutation. The 245-246 loop region is in the ‘out’ position, in an identical

conformation to the wild-type binary complex, although unlike the binary structure the

active site loop of the mutant is closed. This indicates that azole binding need not

necessarily require closure of Ser 245 to the ‘in’ position. The loss of the hydrogen

bonding capability of Ser 245 leaves the Cα atom of the mutant Ala 245 approximately

3Å distant from the equivalent Cα in the wild type ternary structure, leaving plenty of

scope for inclusion of 1-position substituents in the azole ring.

Intriguingly, the kinetic data show that while the binding affinity of the S245A mutant for

NADH is decreased by only a factor of three, the ability to bind pyruvate is much more

severely reduced (a 114 fold increase in KM for pyruvate, Table 3). The Oγ of serine 245

in the wild type ternary structure hydrogen bonds indirectly via a water molecule to the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


22

ribose of NADH so it might be anticipated that NADH binding would be affected more by

the serine to alanine mutation. However, this type of behaviour has been observed in

mammalian S163L LDH mutants (32,33) where a hydrogen bond between S163 and the

nicotinamide amide group is removed. The conclusion of that study suggests that the

hydrogen bond was more important for orienting the nicotinamide head group than for

contributing to binding. The NADH binds before pyruvate in the bi-bi reaction

mechanism and, in turn, forms part of the pyruvate binding site such that misalignment of

the nicotinamide-ribose group of NADH in the active site will compromise pyruvate

binding. Unlike the mammalian S163L LDH mutants, the pf-LDH S245A mutant also

shows a greatly reduced kcat (224 fold). This indicates that the hydride transfer step in the

ES complex is compromised, perhaps by misalignment of the pyruvate while the 245-264

is in its open conformation.

These studies show that serine 245, absent in human LDH, is critical to the correct

functioning of the pfLDH enzyme due to its role in the creation of the pyruvate binding

site, and to a lesser extent in the binding of NADH in the active site. Designing an

antimalarial compound that interacts with this residue may therefore reduce the likelihood

of viable resistant strains of parasite being formed, since mutation of S245 severely

reduces the catalytic ability of the enzyme. The structural data supported the notion that,

in principle, the azole ring could be extended from the 1-position by moving Ser 245 to

the ‘out’ position, while still maintaining interactions of the hydroxyl and carboxyl

functions of the ring with residues from an enzyme closed active site loop.

Azoles bind as the dianion

A limitation in the expansion of the azoles from the 1-position arises, however, from the

delocalised electronic character of these pentacyclic rings along with potential keto-


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


23

enolisation of the 3-hydroxyl substituent. Kinetic analysis (Fig. 6) shows that OXD1 is

essentially uncompetitive with pyruvate, i.e. it does not compete with pyruvate for the

pfLDH-NADH pyruvate-binding site. Binding to pfLDH-NAD+ is instead preferred by

about three orders of magnitude (i.e. OXD1 is competitive with lactate), a result consistent

with the crystallographic studies in which we observed significantly increased resolution

of diffraction and order within the binding site when using complexes with NAD+ rather

than NADH (data not shown).

This strongly indicates that OXD1 binds as the dianion, which is consistent with the pKa

of its hydroxyl group (estimated, from chromatography studies, to be approximately 5)

and is analogous to the behaviour of oxalate. It has been previously noted that LDH only

forms stable ternary complexes when the charges on the active site histidine, the substrate

and the coenzyme nicotinamide ring sum to zero (29,34). This requirement is satisfied if

OXD1 is dianionic and His 195 is protonated, as NAD+ is the preferred form of the co-

factor. The pKa of the active site histidine was estimated to be 8.1 ± 0.2, close to that of

the human LDHs (7). Observed bond lengths in the high resolution (1.1 Å) OXD1-NAD+-

pfLDH complex support this arrangement, with the length of the bond between His195

and the oxadiazole 3-hydroxy group being 2.6 Å, consistent with a charged hydrogen bond

between a protonated histidine and anionic hydroxyl. The requirement for the pKa of the

azole hydroxyl group to be close to or below physiological pH considerably restrains the

design of other 5-membered ring heterocyclic compounds which might be derivatised at

position 1. One explanation for the lack of inhibitory activity associated with the triazole

(TRZ) series (Table 1D) is that the pKa of the hydroxyl group in these compounds is

expected to be very high (> 9.5). On this basis, synthesis of thiazoles might offer a better

route forward to derivatisation of this family of compounds.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


24

Anti-malarial activity of the azoles

The activity of the azole series against pfLDH is largely paralleled in their activity against

Plasmodium falciparum-infected erythrocytes, in both chloroquine-resistant and

chloroquine-sensitive strains (Table 1). All compounds that were inactive against the

target enzyme also failed to kill the parasite in vitro; similarly, the relative activity (as

reflected by the IC50 values) of the small number of pfLDH inhibitors is preserved in the

whole-cell assays. Nonetheless, these are low levels of activity. This is particularly evident

from the concentrations required to kill 90% of infected cells (IC90 values), which are

considerably greater than the equivalent IC50 values in all cases for which the data are

available. For OXD1, for example, a concentration of about 75 µM was needed to kill

90% of infected cells. This is a lower level of activity than would normally be considered

appropriate for a drug candidate.

This low level of cellular activity undoubtedly arises in part from the modest levels of

activity against the target enzyme. OXD1, for example, has an IC50 value of just under 1

µM in the PfLDH enzymatic assay, whereas typical drug candidates show low nanomolar

activity in equivalent assays. It is therefore unsurprising that, for most of the compounds

studied, complete killing of infected parasites could not be achieved within the normal

concentrations (25 µg/ml) used for the in vitro Plasmodium assays. Drug uptake is an

issue of particular concern with highly charged, anionic compounds such as the azoles

described in this study, and is especially critical for anti-malarials where there is a need to

cross at least three membranes (erythrocyte, parasitophorous vacuole and parasite) before

parasite targets can be reached. In preliminary drug uptake studies we have found no

evidence for preferential uptake of OXD1 within either infected or uninfected

erythrocytes, in contrast to chloroquine controls (data not shown). This is consistent with


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


25

only low levels of the OXD1 compound penetrating the cells – hence contributing to the

low levels of cellular activity – or the compound is actively being extruded by an

unknown transporter. There was no evidence that compounds in which the carboxylate

function was esterified, hence reducing its anionic character, were more active in the

whole cell (e.g. OXD6, OXD7, OXD9). However, as the esters are inactive against the

PfLDH enzyme (Table 1A) they would need to be hydrolysed within the cells, a reaction

that has proven difficult in our experience. Further uptake studies are required if this

family of PfLDH inhibitors is to be developed.

Plasmodium berghei, frequently used as an animal model for human malaria, shares a

highly homologous form of LDH (97% identical to pfLDH) for which the crystal structure

has recently been determined (18). As the special features of the active site that

distinguish pfLDH from its human counterparts are completely preserved in P. berghei

LDH, it is anticipated that compounds active against pfLDH would also be effective

against the P. berghei parasite. This indeed proved to be the case. In the Peters 4 day test,

both compounds showed significant anti-malarial activity in P. berghei-infected mice

when compared to untreated control mice. The percentage of inhibition observed was 41

% for OXD1 and 30 % for IOA1 compared with untreated controls, whereas chloroquine

showed 100 % inhibition. The lower in vivo inhibitory activity of compound IOA1

compared to OXD1 correlated with their relative anti-plasmodial activity observed in vitro

– and their ability to inhibit pfLDH in the enzymatic assays. This trend is also seen when

the inhibitory activity was measured in a dose response experiment (Figure 7) where both

OXD1 and TDA1 both proved more effective at killing the parasite than the isoxazole

IOA1. However, all three azoles were less effective than the chloroquine control. This is

consistent with the moderate inhibitory levels of activity measured for the azoles against

their enzyme target.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


26

From the preliminary pharmacokinetic studies, it is clear that these three compounds

showed good plasma levels but short half-lives in the circulation (see pharmacokinetic

studies). This may also contribute to the lack of complete suppression of parasitemia at

doses of 50 or 100 mg/kg/d either by the oral or intraperitoneal routes. However, the good

dose response observed with TDA1 and partly with OXD1 supports the notion that these

compounds have specific anti-plasmodial activity. Although the solubility of IOA1 in

aqueous solutions for oral dosing was poor compared to the other two compounds, despite

reasonable plasma levels its anti-malarial activity showed an erratic dose response. These

results show that despite these compounds exhibiting only moderate pfLDH activity in

vitro and in the absence of preferential uptake to infected cells, suppression of parasitemia

in vivo is nevertheless possible.

The unusual route for energy (ATP) generation in Plasmodium relative to its human host

suggests targeting of glycolytic enzymes should prove a valuable source of compounds

with anti-parasitic activity. This demonstration that inhibition of pfLDH is fatal to the

parasite supports the view that, despite the recent identification of all the genes necessary

for a complete TCA cycle in P.falciparum (35), an effective respiratory chain is unlikely

to be functional in at least the blood-borne stages of the parasite lifecycle. Although

previous workers have described gossypol-like compounds that inhibit pfLDH, these

molecules have been difficult (expensive) to synthesise and many are likely to have poor

cytotoxicity profiles. In this paper we have described a new class of compounds that

specifically inhibit pfLDH and also display anti-malarial activity. X-ray crystallographic

analyses illustrate these compounds interact directly and preferentially with pfLDH. They

have been further characterised in parasiticidal whole cell assays using drug sensitive and

resistant strains of Plasmodium, and demonstrated to have modest in vivo anti-malarial


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


27

activity using the P. berghei rodent model. In combination, we believe these results make

a substantial case for the validation of pfLDH as a viable target for anti-malarials.

Although the azoles are limited in opportunities for further derivatisation because of their

intimate contacts made within the active site of the enzyme, and appear to have limited

cellular uptake in their current form, further development of extended azole-like

compounds might prove a profitable route for the development of novel anti-malarials.

Finally, this study also demonstrates the feasibility of developing specific inhibitors

against microbial targets that have direct human homologues. Small structural differences

between pfLDH and human LDH allow sufficient discrimination for preferential inhibition

of the parasite enzyme. This principle vastly increases the pool of proteins that could be

used as viable drug targets.

Acknowledgements

Acknowledgements: We gratefully acknowledge financial support from the Medicines

for Malaria Venture (www.mmv.org) for these studies. VJW was supported by a

studentship from the UK BBSRC. We are grateful to the staff at the Daresbury SRS and

Hamburg DESI synchrotrons for access to X-ray facilities used for this study.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Azole-based inhibitors of P. falciparum LDH References

1. Lang-Unnasch, N, and Murphy, AD. (1998) Annu Rev Microbiol 52, 561-590

2. Scheibel, LW, Adler, A, and Trager, W. (1979) Proc Natl Acad Sci U S A 76,

5303-5307

3. Roth, EF, Jr., Raventos-Suarez, C, Perkins, M, and Nagel, RL. (1982) Biochem

Biophys Res Commun 109, 355-362

4. Gomez, MS, Piper, RC, Hunsaker, LA, Royer, RE, Deck, LM, Makler, MT, and

Vander Jagt, DL. (1997) Mol Biochem Parasitol 90, 235-246

5. Menting, JG, Tilley, L, Deady, LW, Ng, K, Simpson, RJ, Cowman, AF, and Foley,

M. (1997) Mol Biochem Parasitol 88, 215-224

6. Ridley, RG. (1997) Exp Parasitol 87, 293-304

7. Read, JA, Winter, VJ, Eszes, CM, Sessions, RB, and Brady, RL. (2001) Proteins

43, 175-185

8. Dunn, CR, Banfield, MJ, Barker, JJ, Higham, CW, Moreton, KM, Turgut-Balik, D,

Brady, RL, and Holbrook, JJ. (1996) Nat Struct Biol 3, 912-915

9. Makler, MT, Ries, JM, Williams, JA, Bancroft, JE, Piper, RC, Gibbins, BL, and

Hinrichs, DJ. (1993) Am J Trop Med Hyg 48, 739-741

10. Royer, RE, Deck, LM, Campos, NM, Hunsaker, LA, and Vander Jagt, DL. (1986)

J Med Chem 29, 1799-1801

11. Deck, LM, Royer, RE, Chamblee, BB, Hernandez, VM, Malone, RR, Torres, JE,

Hunsaker, LA, Piper, RC, Makler, MT, and Vander Jagt, DL. (1998) J Med Chem

41, 3879-3887

12. Yu, Y, Deck, JA, Hunsaker, LA, Deck, LM, Royer, RE, Goldberg, E, and Vander

Jagt, DL. (2001) Biochem Pharmacol 62, 81-89


http://ww

w.jbc.org/

Dow

nloaded from

28

http://www.jbc.org/


29

13. Dando, C, Schroeder, ER, Hunsaker, LA, Deck, LM, Royer, RE, Zhou, X,

Parmley, SF, and Vander Jagt, DL. (2001) Mol Biochem Parasitol 118, 23-32

14. Allain, CC, Henson, CP, Nadel, MK, and Knoblesdorff, AJ. (1973) Clin Chem 19,

223-227

15. Wieland, H, Kitasato, Z, and Utzino, S. (1930) Liebigs Ann. Chem. 478, 43

16. Otwinowski, Z, and Minor, W. (1997) in Macromolecular Crystallography, part

A, (Carter, J., C.W., and Sweet, R. M., eds) Vol. 276, pp. 307-326, Academic

Press, New York

17. Collaborative Computational Project, N. (1994) Acta Crystallogr D Biol

Crystallogr D50, 760-763

18. Winter, VJ, Cameron, A, Tranter, R, Sessions, RB, and Brady, RL. (2003) Mol

Biochem Parasitol In Press

19. Kuzmic, P. (1996) Anal Biochem 237, 260-273

20. Barstow, DA, Black, GW, Sharman, AF, Scawen, MD, Atkinson, T, Li, SS, Chia,

WN, Clarke, AR, and Holbrook, JJ. (1990) Biochim Biophys Acta 1087, 73-79

21. Ponnudurai, T, Leeuwenberg, AD, and Meuwissen, JH. (1981) Trop Geogr Med

33, 50-54

22. Trager, W, and Jensen, JB. (1976) Science 193, 673-675

23. Desjardins, RE, Canfield, CJ, Haynes, JD, and Chulay, JD. (1979) Antimicrob

Agents Chemother 16, 710-718

24. Peters, W. (1975) Ann Trop Med Parasitol 69, 155-171

25. Palomino, JC, and Portaels, F. (1999) Eur J Clin Microbiol Infect Dis 18, 380-383

26. Novoa, WB, Winer, AD, Glaid, AJ, and Schwert, GW. (1959) J Biol Chem 234,

1143-1148


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


30

27. Sessions, RB, Dewar, V, Clarke, AR, and Holbrook, JJ. (1997) Protein Eng 10,

301-306

28. Vander Jagt, DL, Caughey, WS, Campos, NM, Hunsaker, LA, and Zanner, MA.

(1989) Prog Clin Biol Res 313, 105-118

29. Chapman, AD, Cortes, A, Dafforn, TR, Clarke, AR, and Brady, RL. (1999) J Mol

Biol 285, 703-712

30. Yuvaniyama, J, Chitnumsub, P, Kamchonwongpaisan, S, Vanichtanankul, J,

Sirawaraporn, W, Taylor, P, Walkinshaw, MD, and Yuthavong, Y. (2003) Nat

Struct Biol 10, 357-365

31. Perozzo, R, Kuo, M, Sidhu, ABS, Valiyaveettil, JT, Bittman, R, Jacobs, WR,

Fidock, DA, and Sacchettini, JC. (2002) Journal of Biological Chemistry 277,

13106-13114

32. Hewitt, CO, Eszes, CM, Sessions, RB, Moreton, KM, Dafforn, TR, Takei, J,

Dempsey, CE, Clarke, AR, and Holbrook, JJ. (1999) Protein Eng 12, 491-496

33. Eszes, CM, Sessions, RB, Clarke, AR, Moreton, KM, and Holbrook, JJ. (1996)

FEBS Lett 399, 193-197

34. Wilks, HM, Hart, KW, Feeney, R, Dunn, CR, Muirhead, H, Chia, WN, Barstow,

DA, Atkinson, T, Clarke, AR, and Holbrook, JJ. (1988) Science 242, 1541-1544

35. Gardner, MJ, Hall, N, Fung, E, White, O, Berriman, M, Hyman, RW, Carlton, JM,

Pain, A, Nelson, KE, Bowman, S, Paulsen, IT, James, K, Eisen, JA, Rutherford, K,

Salzberg, SL, Craig, A, Kyes, S, Chan, MS, Nene, V, Shallom, SJ, Suh, B,

Peterson, J, Angiuoli, S, Pertea, M, Allen, J, Selengut, J, Haft, D, Mather, MW,

Vaidya, AB, Martin, DM, Fairlamb, AH, Fraunholz, MJ, Roos, DS, Ralph, SA,

McFadden, GI, Cummings, LM, Subramanian, GM, Mungall, C, Venter, JC,


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


31

Carucci, DJ, Hoffman, SL, Newbold, C, Davis, RW, Fraser, CM, and Barrell, B.

(2002) Nature 419, 498-511


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


32

Figure legends

Figure 1. Azole-based inhibitors of pfLDH. Schematic showing chemical structures

of OXD1, IOA1 & TAD1 parent compounds.

Figure 2. OXD1 binds in the active site of pfLDH. Stereo figure showing the

bound conformation of OXD1 in the active site of (a, top) pfLDH, and (b, bottom) human

LDH. In each case the co-factor NAD+ is also shown, and key amino acids are labelled.

Distances of contacts between OXD1 and polar atoms from the enzyme are given in

Angstroms.

Figure 3. Azole-based inhibitors in the active site of pfLDH. Figure shows the

active site region in the crystal structures of pfLDH co-crystallised with NAD+ and OXD1

(green), NADH and IOA1 (blue), and NADH with TDA1 (magenta).

Figure 4. Movement of Ser 245 / Pro 246 between binary and ternary complexes

of pfLDH. Figure shows a stereoview of the crystal structure of the pfLDH:NADH binary

complex (blue carbons) overlaid on the pfLDH:NAD+:oxalate ternary complex (green

carbons except NADH carbons which are shown in magenta). Note that Arg109 and the

water molecule are both disordered in the binary structure, a consequence of the active site

loop (residues 102-109) not being ordered in the absence of substrate. The arrow indicates

the direction of movement of the Ser 245 / Pro 246 during the transition from the binary to

the ternary structure. Distances of key contacts are given in Ångstroms for the ternary

structure.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


33

Figure 5. OXD1 competes with lactate for binding to pfLDH. Figure shows

Lineweaver-Burke plots (1/V vs 1/[S]) for inhibition of pfLDH activity when titrated

against varied amounts of (A) lactate (OXD1 concentrations ▲, 1.0 µM; ▼, 0.5 µM; ,

0.25 µM; ■, 0 µM) (B) NADH (OXD1 concentrations ▲, 12.5 µM; ▼, 6.25 µM; , 3.125

µM; ■, 0 µM) and (C) pyruvate (OXD1 concentrations ▲, 10.0 µM; ▼, 5.0 µM; , 2.5

µM; ■, 0 µM). The inhibition is classic competitive against lactate, and mixed against

NADH and pyruvate.

Figure 6. Crystal structure of the Ser245Ala mutant. Figure shows a stereoview

of the active site region from the crystal structure of the Ser 245Ala mutant pfLDH (cyan

carbons) co-crystallised with NADH and oxamate, overlaid on the structure of the wild-

type enzyme (green carbons except NADH carbons which are shown in magenta) co-

crystallised with the same ligands. The structures are overlaid on oxamate. Note that the

substrate analogue oxamate has an amide nitrogen in place of a methyl group in the

natural substrate, pyruvate. Distances between polar neighbours are given in Angstroms

for the wild-type enzyme.

Figure 7: In vivo dose response of P. berghei infected mice treated with azole-

based inhibitors. Chart shows the percentage inhibition of P.berghei in mice when

treated with azole-based inhibitors administered orally (solid bars) and via intra peritoneal

(IP) injection (all other bars). Chloroquine treatment at 10 mg/kd/day was included as a

control and resulted in complete inhibition (data not shown).


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


34

Figure 8: Pharmacokinetic data for azole-based inhibitors. Graphs show plasma

concentration time curves (AUC) for OXD1 and IOA1. Circulating levels of each

compound were determined at time intervals after (A) intra-venous injection of OXD1, 5

mg/kg, (B) intra-peritoneal injection of OXD1, 20mg/kg, (C) intra-venous injection of

IOA1, 5 mg/kg, (D) intra-peritoneal injection of IOA1. AUC0-8 and elimination half-lives

(t1/2) were calculated from each curve. Data shown are fitted to a one phase exponential

decay for display purposes.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


35

Table 1A. Activity of 1,2,5-oxadiazole (OXD) series

NO

O

N

HO

O

Cl

Cl

NN

NH2 OH

O

O

NN

NH

O

OHO

OH

NN

OH

O

OH

O

Compound Structure IC50 pfLDH(µM)

IC50 hsLDH(µM)

IC50 [IC90] 3D7 (µM)

IC50 [IC90] K1

(µM)

OXD1

0.65 72.05 22.5 [75.9]

18.6 [143]

OXD2

>200 >200 N/D N/D

OXD3

>200 >200 N/D N/D

OXD4

>200 >200 N/D N/D

OXD5

>200 >200 N/D N/D

NN

OHOH

O

O

NN

OH

O

O

N

O

NO

O

O

O

OXD6

>200 >200 54.4 [> 87] N/D

OXD7

>100 >100 89.8 [>159] N/D

OXD8

N

O

NO

OH

O

>200 >200 69.4 [>102] N/D

OXD9

>200 >200 N/D N/D

N

O

N

HO

O

OEt

N/D: not determined


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


36

Table 1B. 1,2/1,5-isoxazole (IOA) series

Compound Structure IC50

pfLDH (µM)

IC50 hsLDH (µM)

IC50 [IC90] 3D7 (µM)

IC50 [IC90]

K1 (µM)

IOA1 N

OHOH

O

O

1.1 54

74.4 [>194]

N/D

IOA2 N

OHOH

O

O

16 >100 >194 [>194] N/D

IOA3

N

OH

O

O

OH

>100 >100 N/D N/D

IOA4

N

O

O

OH

O

>100 >100 N/D N/D

IOA5

N

O

O

OH

O

>100 >100 N/D N/D

IOA6

N

OH

O

O

OEt

>100 >100 >59 [>159] N/D

IOA7

NO

OHO

OMe

>100 >100 N/D N/D

IOA8

O

OH

N

OOEt

>100 >100 >59 [>159] N/D

IOA9

N

OH

OH

O

O

OH

O

81.01 >100 N/D N/D

N/D: not determined


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


37

Table 1C. 1,2,5-thiadiazole (TDA) series


IC50 hsLDH(µM)

IC50 [IC90] 3D7 (µM)

IC50 [IC90] K1

(µM)

TDA1

NN

OHOH

O

S

0.14 10.27 75.3 [>171]

144 [>171]

TDA2

NN

NH2 OH

O

S

>100 >100 N/D N/D

TDA3

N

O

NS

OHO

>100 >100 N/D N/D

N/D: not determined


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


38

Table 1D. Triazole (TRZ) series


IC50 hsLDH(µM)

IC50 [IC90] 3D7 (µM)

IC50 [IC90] K1

(µM)

TRZ1 NN

OH

N

OH

O

>100 >100 N/D N/D

TRZ2 N

NN

OHOH

O

>100 >100 N/D N/D

TRZ3

NN

N

OHO

OH

Ph

>100 >100 N/D N/D

TRZ4 N

NN

OHOH

O

Ph

>100 >100 N/D N/D

TRZ10 N

H

O

OH

NN

98 >200 N/D >884 [> 221]

N/D: not determined by guest on April 9, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Table 2. Data collection and refinement statistics

Data collection pfLDH: OXD1:NAD+

hsLDH: OXD1:NAD+

pfLDH: IOA1:NADH

pfLDH: TDA1:NADH

pfLDH: NADH

pfLDH: OXM:NADH

PfLDH S245A:OXM:NADH

Xray source Nonius FR591 Nonius FR591 SRS PX14.1 SRS PX14.1 SRS PX14.1 DESY X11 SRS PX14.1 Wavelength 1.5418 1.5418 1.488 1.488 1.488 0.9091 1.488 Resolution range (Ă) 30 – 1.7 30 – 3.0 30 – 1.9 30 – 1.8 30 – 2.0 25 – 1.1 20 – 1.85 No. unique reflections 33448 27214 26093 29685 21521 127177 25934Redundancy 2.7 2.7 4.1 5.5 4.0 5.6 4.9Completeness (%) 96 (90) 91 (80) 99.8 (99.2) 98.5 (97.1) 98.3 (99.8) 99.7 (95.6) 98.8 (96.2) Rsym 0.058 (0.188) 0.166 (0.302) 0.066 (0.227) 0.064 (0.141) 0.044 (0.105) 0.072 (0.473) 0.059 (0.289)

Average I/σ(I) 18 (5.1) 6.4 (2.6) 19.5 (4.7) 28.3 (9.2) 28.5 (10.1) 22.5 (2.1) 21.3 (3.7)

Refinement

Rfree (%) 16.7 27.8 18.6 20.9 18.4 15.2 19.1Rcryst (%) 14.3 24.0 15.2 17.4 14.3 14.3 15.6R.M.S. bond length (Ă) 0.011 0.013 0.016 0.008 0.024 0.007 0.023R.M.S. bond angle (º) 1.544 2.103 1.667 1.300 1.815 1.287 2.375PDB Accession Code 1T24 1T2F 1T25 1T26 1T2C 1T2D 1T2E

Table 3. Kinetic properties of pfLDH Ser245Ala.

S245A pfLDH Wild Type pfLDH* KM NADH (µM) 66 (±2.7) 16 (±1) KM Pyruvate (µM) 5566 (±185) 49 (±6) kcat (s-1) 0.42 (±0.01) 94 (±3) * data from (18)

39

by guest on April 9, 2018 http://www.jbc.org/ Downloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Table 4. Time course of drug levels in serum for OXD1, IOA1, and TDA1.

0.5 17.89 12.2

1.5 4.77 6.2

3 4.4 5.1

0.5 25.48 2.5

1.5 11.99 13.4

3 2.64 19.8

0.5 19.84 4.1

1.5 9.12 14.7

3 4.9 15.9

0.5 20.41 12.1

1.5 7.35 27.6

3 1.15 26.8

0.5 35.19 6.7

1.5 19.82 17.9

3 16.09 4.6

0.5 41.03 4.6

1.5 26.6 17.2

3 7.56 9.1

OXD1

100 mg/kg i.p.

100 mg/kg p.o.

100 mg/kg i.p.

100 mg/kg p.o.

IOA1

TDA1

100 mg/kg p.o.

100 mg/kg i.p.

41


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Fig 1:

ON

OHO

OH

ON N

OHO

OH

SN N

OHO

OH

AAA A

A

OHO

OH

1

3425

IOA1 OXD1 TAD1 Fig 2a

Fig 2b

42


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Fig 3

Fig 4

43


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Fig 5a

Fig 5b

Fig 5c

44


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Fig 6

45


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Fig 7

Fig 8

46


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Supplementary Material

Synthesis of Derivatives of Parent Azole Compounds

In order to establish a structure-activity relationship (SAR) for the azole family, a wide range of

derivatives were prepared. The parent compounds (hydroxyacid azoles) and their hydroxyesters

derivatives from the oxadiazole (OXD1) and isoxazole groups (IOA1-IOA3, IOA6 and IOA8), were

prepared using the methods previously described in the literature, with slight modifications to facilitate

gram scale process (see references in Table S1). The compoundsfrom the thiadiazole (TDA1 - TDA2)

and triazole groups (TRZ1-TRZ5), were acquired from external sources for screening purposes. The

experimental procedures for the synthesis of novel compounds described in the text, and their

corresponding NMR spectral data, are described below.

S1.1: 4-hydroxy-1,2,5-oxadiazole-3-carboxylic acid, ethyl ester (OXD7)

A solution of 1.70 g 4-hydroxy-1,2,5-oxadiazole-3-carboxylic acid (OXD1) (13 mmol) in 30 ml of

ethanol in the presence of catalytic amounts of H2SO4 (300µL), was heated to reflux. The progress of

the reaction was monitored by HPLC. After 8 hours of heating, the starting material was consumed.

The reaction mixture was cooled to room temperature and the solvent was removed under reduce

pressure. The mixture was re-dissolved in water (pH of the solution = 11) and extracted with Et2O to

remove any non acidic residue formed in the reaction. The aqueous phase was acidified to pH=1 using

concentrated HCl and then extracted with Et2O. The combined ether extracts were washed with brine

and the solvent removed under reduce pressure to yield 1.04 g of the expected compound as a white

47


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


solid (Yield 50%). 1H-RMN (DMSO, δ): 1.30 (t, J=7.15, 3H), 4.36 (q, J=7.15, 2H), 13.4 (bs, 1H, OH).

13C-RMN (DMSO, δ): 14.0, 62.3, 140.5, 157.34, 162.8.

S1.2: 4-hydroxy-1,2,5-oxadiazole-3-carboxilic acid, (3',4'-dichloro)benzyl ester (OXD6 )

A solution was prepared from 272 mg of 4-hydroxy-1,2,5-oxadiazole-3-carboxylic acid (OXD1) in 3

ml of DMF, to which were successively added 31 mg of dimethyl aminopyridine (0.25 mmol), 370 mg

of 3,4-dichlorobenzyl alcohol (2.1 mmol), 1.07 g (2.25 mmol) of methyl-p-toluenesulphonate and 354

mg of N-cyclohexyl-N’-(2-morpholinoethyl)carbodiimide. The progression of the reaction was

monitored by TLC. After 3H of reaction no starting material was detected. The precipitant formed

during the reaction was filtered, and the mother liquor was concentrated under reduced pressure,

obtaining a thick melange that was re-dissolved in EtOAc and extracted with NH4Cl (1N). The

combined organic extracts were washed with brine, dried over Na2SO4, and concentrated under reduced

pressure. The yellowish residue obtained was purified by flash chromatography using AcOEt:Hexane

(1:1), which was further purified by recrystallisation with ethyl ether: hexane. Finally 150 mg of the

pure expected compound was obtained in a 24% yield of a white crystalline solid. 1H-RMN (DMSO, δ):

5.29 (s, 2H), 7.44 (d, 1H, J=8.25), 7.65 (d, 1H, J=8.25), 7.73 (s, 1H). 13C-RMN (DMSO δ): 64.3, 128.1,

129.8, 130.7, 130.6, 131.0, 137.0, 142.5, 160.7, 168.2

S1.3: 3-hydroxy-5-methyl-4-isoxazolecarboxilic acid, ethyl ester (IOA7)

A solution of 3.89 g hydroxyl amine hydrochloride (56 mmol) in 25 ml of methanol was cooled to 0ºC

and, under stirring, 20 ml of MeONa (30% in MeOH) were slowly added. The formation of an

abundant white precipitate (NaCl) was observed during the neutralisation. Then, 11 g of ethyl dimethyl

48


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


(1-ethoxyethylidene) malonate 1 (51 mmol) dissolved in 100 ml of methanol were added to the reaction

media. After 25 min of reaction no starting material could be detected by TLC. The product was

purified by initially filtering the NaCl formed, and removing the methanol used as solvent under

reduced pressure. The brown mixture obtained was redissolved in a 1:1 mixture of ethyl acetate and

NH4Cl, and the phases were separated. The organic phase was conserved and the aqueous layer was

treated with concentrated HCl to adjust the pH to pH=4. The mixture was further extracted with ethyl

acetate (3x200 ml) to remove any traces of the isomer from the media. The combined organic extracts

were collected, washed with brine, dried over Na2SO4 and evaporated under reduced pressure to yield a

mixture that contained the major isomer, which was purified by crystallisation from methanol to yield

45% of IOA7. 1H-RMN (DMSO, δ): 2.52 (t, s, 3H), 3.73 (s, 3H). 13C-RMN (CD3OD, δ): 14.6, 61.2,

88.3, 152.5, 163.1, 172.1 The corresponding data are given below. The aqueous phase that was

originally adjusted to pH=4 contained the minor isomer, which was isolated in 37% yield after

acidification to pH=1: IOA3e: 1H-RMN (DMSO, δ): 2.28 (t, s, 3H), 3.62 (s, 3H).

Synthesis of ether-derivatives of parent compounds:

All ethers were synthesised from the ethyl esters of the corresponding parent compounds. The synthesis

of each compound required two steps: first, the introduction of the substituent in the hydroxyl

functional group, and secondly the saponification of the ester group to yield the corresponding acid.

49


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


S1.4: 3-methoxy-4-isoxazolecarboxilic acid, ethyl ester (IOA5e)

Using an oven-dried two-neck flask equipped with a reflux condenser and flushed with N2 after the

addition of 2.5g of Cs2CO3 (3.81 mmol), a solution of IOA6 (600 mg, 3.81 mmol) in 10 ml of acetone

was added to the base and the heterogeneous mixture was stirred and heated at 60ºC over one hour

(release of CO2 was observed during the generation of the anion). After this period, the mixture was

left to reach room temperature and 356 ul of MeI (5.72 mmol) was added, and the mixture was stirred

for an additional hour. The progress of the reaction was monitored by TLC. After 1 hour at the same

temperature all the starting material had been consumed. The expected compound was isolated by

addition of 20 ml of saturated NH4Cl and extracted with ethyl ether (3x50 mL). The combined organic

extracts were washed in brine, dried over MgSO4 and the solvent was evaporated under reduced

pressure. The crude obtained was purified by flash chromatography (Hexane: AcOEt, 9:1), to yield the

expected compound in 64% yield. 1H-NMR (CDCl3, δ): 1.33 (t, J=7.14, 3H), 4.31(s, 3H), 4.31, (q,

J=7.14, 2H), 8.67 (s, 1H). 13C-NMR (CDCl3, δ): 14.2, 57.9, 61.0, 105.2, 160.2, 164.6, 169.4

S1.5: N-methyl-3-oxo-2(2H)-4-isoxazolecarboxilic acid, ethyl ester IOA4e

In a dried oven flask, flushed with N2 and under inert atmosphere (N2), was placed a solution of IOA6

(300 mg , 1.9 mmol) in 5ml of dry dichloromethane. The solution was cooled to 0ºC and 1.9 mL of

TMSCHN2 (3.81 mmol of 2M solution in hexane) was added drop-wise to the mixture. The release of

N2 was observed during the reaction. The progress of the reaction was monitored by TLC and after 10

min of reaction no starting material was detected. The reaction was stopped by evaporation of the

solvent under reduced pressure and the crude directly purified by flash chromatography (Hex: AcOEt,

10:1). The expected compound was isolated as the less polar compound in a 45% yield. 1H-NMR

50


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


(CDCl3, δ): 1.33 (t, J=7.14, 3H), 3.55 (s, 3H), 4.31, (q, J=7.14, 2H), 8.42 (s, 1H). 13C-NMR (CDCl3, δ):

14.2, 32.9, 61.1, 108.6, 160.2, 163.0, 175.4

S1.6: General method A: Synthesis of ether derivatives under Mitsunobu conditions

Over a solution of 4-hydroxy-1,2,5-oxadiazole-3-carboxylate ethyl ester (OXD7) in anhydrous THF

and under N2 atmosphere, 1.04 molar equivalents of PPh3 were added. The reaction mixture was cooled

to 0ºC, before the consecutive addition of 1 molar equivalent of diethyl azadicarboxylate and 1 molar

equivalent of the corresponding alcohol. After addition of the alcohol the cooling bath was removed

and the reaction mixture was left stirring overnight. The solvent was removed under reduced pressure.

The crude obtained was washed with hexane and the remaining solid was directly purified by flash

chromatography using the eluent indicated in each case.

S1.7: 4-[(3-phenylprop-2-enyl)-oxi]-1,2,5-oxadiazole-3-carboxylic acid, ethyl ester (OXD8e)

This compound was prepared following the general method A starting from 119 mg of OXD7 (0.75

mmol) and 100 mg of the 3-phenyl-2-propen-1-ol (cinnamyl alcohol, 0.75 mmol) and the

corresponding amounts of PPh3 (273mg, 0.78 mmol) and DEAD (130 mg, 0.75 mmol). The crude from

the reaction mixture was purified by flash chromatography, using a gradient from AcOEt: Hex (1:10) to

pure AcOEt. Yield 42%. 1H-NMR (CDCl3δ): 1.45 (t, J=7.13, 3H), 4.48 (q, J=7.14, 2H), 5.07 (dd,

J1=6.54, J2=1.28, 2H), 6.45 (dt, J1=15.93, J2=6.40, J3=6.40, 1H), 6.83 (d, J=15.74, 1H), 7.29-7.44 (m,

5H).

S1.8: 4-[(p-methoxyphenyl)-methoxy]-1,2,5-oxadiazole-3-carboxilic acid, ethyl ester (OXD9e).

51


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


OXD9e was prepared following the general method A starting from 150 mg of OXD7 (0.95 mmol) and

133 mg of 4-methoxy benzyl alcohol (0.95 mmol) and the corresponding amounts of PPh3 (343mg,

0.99 mmol) and DEAD (165 mg, 0.95 mmol). The crude from the reaction mixture was purified by

flash chromatography, using a gradient from AcOEt: Hex (1:3) to pure AcOEt. Yield 49%. 1H-NMR

(CDCl3, δ): 1.40 (t, J=7.14, 3H), 3.82(s, 3H), 4.44, (q, J=7.14, 2H), 5.36 (s, 2H), 6.92, (d, J= 8.8, 2H),

7.43, (d, J= 8.8, 2H). 13C-NMR (CDCl3, δ): 14.0, 55.3, 62.3, 74.4, 114.0, 126.3, 130.4, 139.6, 157.2,

160.2, 163.8 .

S1.9: General Method B. Hydrolysis of hydroxyesters.

Over a cooled (0ºC) solution of the starting hydroxyester in methanol (2M solution), 4 equivalents of

LiOH (1.25 N) were slowly added while stirring and the mixture was left to reach room temperature on

its own. The reaction was monitored by TLC, and usually a period of 40 to 60 min was enough to

complete the hydrolysis. The isolation procedure was performed by removal of the methanol under

reduced pressure, extraction with dichloromethane to remove any organic component and acidification

to pH=1. The precipitate formed was filtered as the pure acid. When a significant precipitate was not

formed after acidification, the aqueous phase was extracted several times with AcOEt (3x50 ml). The

organic extracts were combined, washed with brine and dried over Na2SO4. The expected compounds

were obtained as pure compounds and usually did not require any further purification.

52


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


S1.10: N-methyl-3-oxo-2(2H)-4-isoxazolecarboxilic acid, IOA4

Following the general procedure C, 87 mg of the ester N-methyl-3-oxo-2(2H)-4-isoxazolecarboxilic

acid (IOA4e) (0.31 mmol) were hydrolysed over 90 minutes. A further chromatographic purification

was necessary (Hex: AcoEt (3:1) to AcOEt) to obtain 38 mg of the pure expected compound in 60%

yield. 1H-NMR (DMSO, δ): 2.93 (s, 3H), 9.09 (s, 1H), 13.0 (s, 1H). 13C-NMR (DMSO, δ): 33.6, 108.1,

160.3, 167.0, 171.3 .

S1.11: 3-methoxy-4-isoxazolecarboxilic acid, ethyl ester IOA5

Following the general procedure B, 87 mg of the ester 3-methoxy-4-isoxazolecarboxilic acid, ethyl

ester (IOA5e) (0.31 mmol), were hydrolysed over 90 minutes. A further chromatographic purification

was necessary (Hex: AcoEt (3:1) to AcOEt) to obtain 38 mg of the pure expected compound in 60%

yield. 1H-NMR (DMSO, δ): 3.93 (s, 3H), 9.29 (s, 1H), 13.0 (s, 1H). 13C-NMR (DMSO, δ): 57.5, 105.0,

160.9, 166.7, 169.1

S1.12: 4-[(3-phenylprop-2-enyl)-oxy]-1,2,5-oxadiazole-3-carboxylic acid (OXD8)

Following the general procedure B, 44 mg of 4-[(3-phenylprop-2-enyl)-oxy]-1,2,5-oxadiazole-3-

carboxylic acid, ethyl ester (OXD8e) (0.16 mmol), were hydrolysed over 90 min in a 64% yield. 1H-

RMN (CD3OD δ): 5.04 (dd, J1=6.42, J2=1.29, 2H), 6.50 (dt, J1=15.94, J2=6.42, J3=6.42, 1H), 6.85 (d,

J=15.94, 1H), 7.25-7.47 (m, 5H). 13C-RMN (CD3OD, δ): 74.4, 122.9, 127.8, 129.4(4xC), 129.7 (4xC),

136.9, 137.5, 141.7, 159.8, 165.3.

53


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


S1.13: 4-[(p-methoxybenzyl)-oxy]-1,2,5-oxadiazole-3-carboxylic acid (OXD9)

Following the general procedure B, 87 mg of the ester 4-[(p-methoxybenzyl)-oxy]-1,2,5-oxadiazole-3-

carboxylic acid, ethyl ester (IOA9) (0.31 mmol), were hydrolysed over 90 minutes. A further

chromatographic purification was necessary (Hex: AcoEt (3:1) to AcOEt) to obtain 38 mg of the pure

expected compound in 60% yield. 1H-RMN (DMSO, δ): 5.37 (s, 2H), 3.81 (s, 3H), 6.92, (d, J= 8.8,

2H), 7.44, (d, J= 8.8, 2H). 13C-RMN (DMSO, δ):55.3, 74.4, 113.0 (4xC), 126.5, 129.4 (4xC), 139.9,

157.5, 160.3, 165.8.

S1.14: 4-methoxy-1,2,5-tiadiazol-3-carboxylic acid (TDA3)2.

Following the general procedure C, 44 mg of ethyl 4-methoxy-1,2,5-tiadiazol-3-carboxylic acid ethyl

ester (TDA3e) (0.16 mmol), were hydrolysed over 90 min in a 60% yield with a purity >95% (HPLC).

1H-RMN (CD3OD, δ): 4.13 (s, 3H). 13C-RMN (CD3OD, δ): 58.6, 140.9, 161.8, 167.2.

Synthesis of Bioisoster derivatives: Replacement of the carboxylic acid functional group

S1.15: 4-(hydroxy-1,2,5-oxadiazole-3-oxo)- N-methylsulphonamide (OXD10)

To an oven-dried flask, flushed with N2 and containing 3Å molecular sieves, were added 182 mg of 4-

hydroxy-1,2,5-oxadiazole-3-carboxylic acid (OXD1) (1.4 mmol) dissolved in 20 ml of dry THF, and

563 mg of 1,1' carbonyl diimidazole. The mixture was heated to 65 ºC and left stirring for 4 hours.

After this period the mixture was cooled to room temperature and 362 uL of methyl sulphonamide (3.5

54


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


mmol) and 524 µL of 1,8-diazabycyclo-[5,4,0]-7-undecene (3.5 mmol) were added. The mixture was

heated again to 65ºC overnight. After this period, the crude was diluted with ethyl acetate and filtered

over Celite®. The Celite® was carefully washed with AcOEt and methanol until all of the substrate had

been eluted from the silica (TLC). The combined organic extracts were dried over Na2SO4, and the

solvent removed under reduced pressure. The oily residue obtained was further purified by flash

chromatography to isolate a main fraction that contained the expected compound together with the

amine used as base. This mixture was redissolved in HCl (1N) and extracted with ACOEt (2x100 mL).

The aqueous phase was discarded. The organic solvent was removed under reduced pressure and the

expected compound was finally isolated as the lithium salt by treatment with a basic solution of NaOH

(pH=6-7) and subsequent lyophilisation. The sulphonamide OXD10 was obtained with >90% purity

(HPLC) in a 67% yield.

1H-RMN (D2O δ): 3.02 (s, 3H). 13C-RMN (CD3OD δ): 40.1, 145.3, 164.7, 165.5

S1.16: N-(3-hidroxy isoxazole-4-oxo)-methylsulphonamide (IOA9)

To an oven-dried flask, flushed with N2 and containing 3Å molecular sieves, were added 138 mg of 3-

hydroxy-isooxazole-4-carboxylic acid (1.07 mmol) dissolved in 6 mL of dry THF, and 563 mg of 1,1'

carbonyl diimidazole. The mixture was heated at 70 ºC and left stirring for 2 hours. After this period

the mixture was cooled to room temperature and 204 mg of methyl sulphonamide (2.14 mmol) and

326mg of 1,8-diazabycyclo-[5,4,0]-7-undecene (2.14 mmol) were added. The mixture was then heated

to 65ºC overnight. After this period the solvent was removed under reduced pressure and the oily

residue was partitioned in dichloromethane and NH4Cl (1N). The aqueous phase was acidified to pH=

3 and then extracted with AcOEt (4x75ml). The different organic extracts were collected, dried over

Na2SO4 and the solvent was removed under reduced pressure to yield a yellowish residue that was

macerated with acetonitrile to yield 20 mg of a white solid with purity > 95 % (HPLC) in a 10% yield.

55


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


The yield of the reaction was not improved since the compound was used just for screening purposes.

1H-RMN (CD3OD, δ): 3.34 (s, 3H, 9.03 (s, 1H). 13C-RMN (CD3OD δ): 41.7, 107.2, 160.5, 166.2, 168.4

S1.17: 3-hydroxyisoxazol-4-hydroxymethyl amide (IOA10)

To an oven-dried flask, flushed with N2, were added 140mg of 3-hydroxy-isooxazole-4-carboxylic acid

(1.08 mmol) and 2 mL of thionyl chloride. The mixture was heated at reflux under stirring for 3 hours.

After this period the mixture was concentrated at reduced pressure and the residue containing the

corresponding acid chloride was redissolved in anhydrous THF (5mL). This solution was added while

stirring to a cooled (0ºC) solution of 360 mg of methyl hydroxylamine (4.4 mmol) and 900 uL of

triethylamine (6.5 mmol) in 5ml of dry THF. The reaction mixture was then left to reach room

temperature and stirred overnight. The expected compound was isolated by acidification of the reaction

mixture with HCl (1N) and extraction with dichloromethane (3x50mL). The organic extracts were

combined, washed with brine and dried over Na2SO4. The solvent was removed under reduced pressure

to obtain a solid that was purified by flash chromatography (dichloromethane:MeOH, 95:5) to obtain

the expected bioisostere in a 30% yield with a purity >95% (HPLC). 1H-RMN (CD3OD, δ): 3.34 (s, 3H,

CH3), 8.95 (s, 1H, CH). 13C-RMN (CD3OD δ): 36.2, 102.7, 163.8, 164.2, 171.7.

56


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


Bibliographic information on compounds prepared and cited in the article

Table S1A: Synthesis of 1,2,5-oxadiazole (OXD) series

NN

NH2 OH

O

O

NN

NH

O

OHO

OH

NN

OH

O

OH

O

Compound Structure Reference for synthesis

OXD1

H. Wieland, Z. Kitasato and S. Utzino. Liebigs Ann. Chem. 478 (1930), p. 43

OXD2

Strelenko, Yu. A.; Sheremetev, A. B.; Khmel'nickii, L. I. Monosubstituted

furazans. I. NMR investigations. Khimiya Geterotsiklicheskikh Soedinenii

(1992), (8), 1101-5.

OXD3

Purchased from external sources: Synthesis details not available

OXD4

Bought from external sources: Catalog Name: Chemical Block Building Blocks

Registry Number: 88598-08-7 , CHEMCATS

OXD5

H. Wieland, Z. Kitasato and S. Utzino. Liebigs Ann. Chem. 478 (1930), p. 43

NN

OHOH

O

O

NN

OH

O

O

57


http://www.jbc.org/


NO

O

N

HO

O

Cl

Cl

N

O

NO

O

O

O

OXD6

Described in Section S1.2

OXD7

Willson, Timothy M.; Charifson, Paul S.; Baxter, Anthony D.; Geddie, Nora G.

Bone targeted drugs. 1. Identification of heterocycles with hydroxyapatite affinity.

Bioorganic & Medicinal Chemistry Letters (1996), 6(9), 1043-1046. Also

described in Section 1.8.

OXD8

Described in Section S1.6, S1.12

OXD9

Described in Section S1.8, S1.9, S1.13

OXD10


N

O

N

HO

O

OEt

N

O

NO

OH

O

N ON

NH

O

OH

S

O

O

Me

58


http://www.jbc.org/


Table S1B: Synthesis of 1,2/1,5-isoxazole (IOA) series:

Compound Structure Reference

IOA1

N

OHOH

O

O

Willson, Timothy M.; Charifson, Paul S.; Baxter, Anthony D.; Geddie, Nora G. Bone targeted

drugs. 1. Identification of heterocycles with hydroxyapatite affinity. Bioorganic & Medicinal

Chemistry Letters (1996), 6(9), 1043-1046

IOA2

N

OHOH

O

O




IOA3

N

OH

O

O

OH

Von Itter, Franz Albert; Steffen, Klaus Dieter. Preparation of dialkyl

alkoxyalkylidenemalonates by aluminosilicate-catalyzed condensation of dialkyl malonates

with ortho esters. Eur. Pat. Appl. (1991), 6 pp. EP 413918

IOA4

N

O

O

OH

O

CH3


IOA5

N

O

O

CH3 OH

O


IOA6

N

OH

O

O

OEt




59


http://www.jbc.org/


IOA7

N

CH3O

OHO

OMe

Von Itter, Franz Albert; Steffen, Klaus Dieter. Preparation of dialkyl

alkoxyalkylidenemalonates by aluminosilicate-catalyzed condensation of dialkyl malonates

with ortho esters. Eur. Pat. Appl. (1991), 6 pp. EP 413918. Modified as described in

Section S1.3

IOA8

O

OH

N

OOEt




IOA9

N

OH

O

O

N

SO

OMe


IOA10

N O

OH

NOOH

Me


60


http://www.jbc.org/


Table S1C: Synthesis of 1,2,5-thiadiazole (TDA) series:


TDA1

NN

OHOH

O

S

J.M. Ross and W.C. Smith. J. Am. Chem. Soc. 86 (1964), p. 2861

TDA2

NN

NH2 OH

O

S

Bought form external sources: Interchim Intermediates. Registry Number: 2829-58-5

Meyer, Rich B., Jr.; Skibo, Edward B. Synthesis of fused [1,2,6] thiadiazine 1,1-dioxides as

potential transition-state analog inhibitors of xanthine oxidase and guanase. Journal of

Medicinal Chemistry (1979), 22(8), 944-8.

TDA3

N

O

NS

CH3 OHO

J.M. Ross and W.C. Smith. J. Am. Chem. Soc. 86 (1964), p. 2861

Modified as described in Section 1.14

61


http://www.jbc.org/


Table S1D: Synthesis of Triazole (TRZ) series


TRZ1

NN

OH

N

OH

O

Screening Compound but the synthesis is described in: Y. A. Rozin, E. A. Saveleva, Y. Y.

Morzherin, W, Dehaen, S. Toppet, L. Van Meervelt, V. A. Bakulev, J. Chem. Soc, Perkin

Transactions 1, 2002, 2, 211-216.

TRZ2

N

NN

OHOH

O

GSK Screening Compound

TRZ3

N

NN

OHOH

O

Ph

D. R. Buckle, C. J. M. Rockell, J. Chem Soc, Perkin I, 1982, 627-63.

TRZ4

N

NN

OHOH

O

Ph

D. R. Buckle, C. J. M. Rockell, J. Chem Soc, Perkin I, 1982, 627-63.

TRZ5

NHOH

ON

N

Scifinder Search: Katritzky, Alan R.; Zhang, Yuming; Singh, Sandeep K. 1,2,3-triazole

formation under mild conditions via 1,3-dipolar cycloaddition of acetylenes with azides.

Heterocycles (2003), 60(5), 1225-1239.

62


http://www.jbc.org/


63


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Federico Gomez de las HerasJavier Gamo, Laura Sanz, Luisa Leon, Jose R. Ruiz, Raquel Gabarro, Aracelli Mallo and

Jose Luis Lavandera, José Julio Martin, Felix Risco, Silvestre García-Ochoa, FraciscoLeo Brady, Livia Vivas, Anna Easton, Howard Kendrick, Simon L. Croft, David Barros,

Angus Cameron, Jon A. Read, Rebecca Tranter, Victoria J. Winter, Richard B. Sessions, R.dehydrogenase-directed anti-malarial activity

Identification and activity of a series of azole-based compounds with lactate

published online April 26, 2004J. Biol. Chem.

10.1074/jbc.M402433200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

http://www.jbc.org/content/suppl/2004/05/06/M402433200.DC1


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M402433200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;M402433200v1&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2004/04/26/jbc.M402433200.citation

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=early/2004/04/26/jbc&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2004/04/26/jbc.M402433200.citation

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/suppl/2004/05/06/M402433200.DC1

http://www.jbc.org/

selective inhibitors of pfldh

Documents