Thiazolopyrimidine inhibitors of 2-methylerythritol 2,4-cyclodiphosphate synthase (IspF) from Mycobacterium tuberculosis and Plasmodium falciparum

Article abstract of DOI:10.1002/cmdc.201000083

A library of 40 000 compounds was screened for inhibitors of 2-methylerythritol 2,4-cyclodiphosphate synthase (IspF) protein from Arabidopsis thaliana using a photometric assay. A thiazolopyrimidine derivative resulting from the high-throughput screen was

Full text of DOI:10.1002/cmdc.201000083

MED
DOI: 10.1002/cmdc.201000083
Thiazolopyrimidine Inhibitors of 2-Methylerythritol 2,4-
Cyclodiphosphate Synthase (IspF) from Mycobacterium
tuberculosis and Plasmodium falciparum
Julie G. Geist,[a] Susan Lauw,[b] Victoria Illarionova,[c] Boris Illarionov,[d] Markus Fischer,*[c]
Tobias Grꢀwert,[b] Felix Rohdich,[b] Wolfgang Eisenreich,[b] Johannes Kaiser,[b] Michael Groll,*[b]
Christian Scheurer,[e] Sergio Wittlin,[e] Josꢁ L. Alonso-Gꢂmez,[a] W. Bernd Schweizer,[a]
Adelbert Bacher,[b] and FranÅois Diederich*[a]
A library of 40000 compounds was screened for inhibitors of
2-methylerythritol 2,4-cyclodiphosphate synthase (IspF) protein
from Arabidopsis thaliana using a photometric assay. A thiazo-
lopyrimidine derivative resulting from the high-throughput
screen was found to inhibit the IspF proteins of Mycobacterium
tuberculosis, Plasmodium falciparum, and A. thaliana with IC50
values in the micromolar range. Synthetic efforts afforded de-
rivatives that inhibit IspF protein from M. tuberculosis and P. fal-
ciparum with IC50 values in the low micromolar range. Several
compounds act as weak inhibitors in the P. falciparum red
blood cell assay.
Introduction
The non-mevalonate pathway for the biosynthesis of isopre-
noids is a clinically validated target for the therapy of malaria.[1]
The pathway is also essential for many pathogenic bacteria in-
cluding Mycobacterium tuberculosis. Because animals obtain
isoprenoids via the mevalonate pathway and do not express
the enzymes of the alternative non-mevalonate pathway, any
potential anti-infective drugs directed against the non-mevalo-
nate pathway enzymes should be devoid of target-related tox-
icity.[2]
by aromatic modules, and the enzyme–inhibitor complexes
were studied by X-ray crystallography.[16] More recently, a series
of non-phosphate, cytosine-type inhibitors with IC50 values in
the micromolar range has been reported.[17]
Earlier, we showed that fluorescent ligands for IspF protein
can serve as a basis for a fluorescence-monitored ligand bind-
ing competition assay for inhibitor screening.[16] More recently,
we developed a dynamically monitored kinetic assay that is
suitable for the screening of large compound libraries
(Scheme 1).[18]
Although the non-mevalonate pathway was discovered rela-
tively recently, its enzymes have now been studied in some
detail. Specifically, IspF protein (2-methylerythritol 2,4-cyclodi-
phosphate synthase, EC 4.6.1.12) converts diphosphocytidyl-2-
methylerythritol 2-phosphate (1) into 2-methylerythritol 2,4-cy-
clodiphosphate (2) (Scheme 1). The structures of IspF proteins
from Escherichia coli,[3–5] M. smegmatis,[6] Haemophilus influen-
zae,[7] Salmonella typhimurium,[8] Thermus thermophilus,[9] She-
wanella oneidensis,[10] Burkholderia pseudomallei,[11] Yersinia
pestis,[12] Plasmodium vivax,[13] and Arabidopsis thaliana[14] have
been determined by X-ray crystallography. They are all C3-sym-
metric homotrimers whose active sites are located at the inter-
faces of adjacent subunits. The structure of a bifunctional
IspDF protein of Campylobacter jejuni has also been deter-
mined by X-ray crystallography.[15] The structural studies were
predominantly focused on IspF orthologues from major patho-
gens. Presently, over 40 structures have been deposited in the
RCSB Protein Data Base. This massive body of structural data
reflects the consensus in the scientific community that IspF
protein could serve as a target for the development of novel
anti-infective drugs.
[a] J. G. Geist, Dr. J. L. Alonso-Gꢀmez, Dr. W. B. Schweizer, Prof. Dr. F. Diederich
Laboratorium fꢁr Organische Chemie, ETH Zꢁrich
Hçnggerberg, HCI, 8093 Zꢁrich (Switzerland)
Fax: (+41)44-632-1109
E-mail: diederich@org.chem.ethz.ch
[b] Dr. S. Lauw, Dr. T. Grꢂwert, Dr. F. Rohdich, Dr. W. Eisenreich, Dr. J. Kaiser,
Prof. Dr. M. Groll, Prof. Dr. A. Bacher
Lehrstuhl fꢁr Biochemie, Technische Universitꢂt Mꢁnchen
Lichtenbergstraße 4, 85748 Garching (Germany)
Fax: (+49)89-289-13363
E-mail: michael.groll@ch.tum.de
[c] Dr. V. Illarionova, Prof. Dr. M. Fischer
Institut fꢁr Lebensmittelchemie, Universitꢂt Hamburg
Grindelallee 117, 20146 Hamburg (Germany)
Fax: (+49)40-42838-4342
E-mail: markus.fischer@chemie.uni-hamburg.de
[d] Dr. B. Illarionov
Ikosatec GmbH, Grindelallee 117, 20146 Hamburg (Germany)
[e] C. Scheurer, Dr. S. Wittlin
Swiss Tropical and Public Health Institute
Socinstrasse 57, 4051 Basel (Switzerland)
A first series of synthetic inhibitors for IspF protein of E. coli
has been obtained earlier by replacing the polyol moiety of 1
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000083.
1092
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1092 – 1101

Thiazolopyrimidine Inhibitors of IspF
Scheme 1. Reaction sequence for high-throughput screening of IspF protein.
The selection of a target protein with favorable properties is
essential for efficient high-throughput screening. The IspF pro-
tein of P. falciparum[19] was not particularly suitable with regard
to solubility and stability. In the present study, we cloned and
expressed IspF protein from M. tuberculosis (which has also
been cloned and expressed by other authors[6]), but found it
poorly suited for the screening of large compound libraries
due to solubility and stability limitations. Therefore, we consid-
ered the use of a surrogate target for the purpose of high-
throughput screening (see figure 1 of the Supporting Informa-
tion for the amino acid sequence alignment of IspF proteins).
In plants and apicomplexan protozoa, the enzymes of the
non-mevalonate pathway are located in the chloroplast and
apicoplast compartments, respectively; notably, apicoplasts are
believed to be homologues of plant chloroplasts. This suggest-
ed the use of a plant orthologue as a surrogate for a protozo-
an pathogen enzyme.
Table 1. Properties of IspF enzymes from various organisms.
Organism
k
cat [secꢀ1
]
KM [mm] pH optimum
Metal ion
preference
A. thaliana[a]
P. falciparum[19]
E. coli[a]
2.5
1.6
2.7
1.8
483
252
410
6.0
7.0
–
Co2+, Mn2+, Mg2+
Mg2+, Mn2+
Mg2+
M. tuberculosis[a]
2000
–
Mg2+
[a] As measured in this study.
Cloning, expression, and properties of IspF protein from
A. thaliana
Because neither IspF proteins from M. tuberculosis and P. falci-
parum were satisfactory for high-throughput screening, IspF
protein of A. thaliana was cloned and expressed as a potential
surrogate target. Specifically, a DNA segment specifying the
predicted amino acid sequence of the putative mature form of
A. thaliana IspF protein (amino acids number 76–223 of the
open reading frame NC_003070.9, chromosome 1, accession
no. NP-564819), without the putative chloroplast targeting se-
quence, was synthesized from ten oligonucleotides by step-
wise PCR amplification. The sequence was adapted to the
codon preferences of E. coli. The synthetic open reading frame
was cloned into the plasmid pQE30 where it was preceded by
a DNA segment specifying an N-terminal polyhistidine tag. In
an E. coli host strain, the recombinant plasmid directed the
synthesis of abundant amounts of recombinant IspF protein
that was isolated from crude cell extract by metal-chelation
chromatography with greater than 95% purity. The molecular
mass of the recombinant protein, as determined by MALDI–
TOF mass spectrometry, was 18263 Da in good agreement
with the predicted value of 18261 Da. Partial Edman degrada-
tion afforded the N-terminal sequence MRGSHHHHHH
GSRIGHGFDLHRLEP. Some properties of the enzyme are sum-
marized in Table 1.
Using a recombinant, pseudomature IspF protein from the
plant, A. thaliana, we screened a library of 40000 compounds
and identified a thiazolopyrimidine derivative that inhibits IspF
proteins of M. tuberculosis and P. falciparum. Synthesis of deriv-
atives afforded a small family of compounds inhibiting the
pathogen enzymes with IC50 values in the low micromolar
range and had weak activity against P. falciparum in the red
blood cell assay.
Results
Cloning, expression, and properties of IspF protein from
M. tuberculosis
The putative open reading frame NC_000962.2 of M. tubercu-
losis H37Rv (Genebank accession no. NP_218098.1) was ampli-
fied by polymerase chain reaction (PCR) and cloned into a plas-
mid which directed the synthesis of a protein with an apparent
molecular weight of 17 kDa as judged by polyacrylamide gel
electrophoresis in a recombinant E. coli strain. The recombinant
protein was isolated by metal chelation chromatography.
Some properties are summarized in Table 1. The protein had a
marked tendency to precipitate even after relatively short stor-
age.
ChemMedChem 2010, 5, 1092 – 1101
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1093

F. Diederich, M. Fischer, M. Groll, et al.
MED
Library screening
The recombinant IspF protein of
A. thaliana appeared favorable
with regard to expression level,
solubility, and stability, and was
used to screen a library of about
40000 synthetic compounds
with a molecular weight range
from 150–650 Da. The photo-
metric assay method used was
reported previously (Scheme
1).[18] Briefly, cytidine 5’-mono-
phosphate (CMP) obtained as
the second product of the IspF-
catalyzed reaction is converted
into cytidine 5’-diphosphate
(CDP) by cytidylate kinase using
ATP as substrate. The resulting
ADP is recycled under regenera-
tion of ATP using phosphoenol-
Scheme 2. Synthesis of compounds (ꢁ)-3–(ꢁ)-12. Reagents and conditions: a) HCl, EtOH, reflux, 6 h; or PPE, THF,
reflux, 6 h; b) Et2NH, [Pd(PPh3)4], THF, room temperature, 3 h; c) ClCH2CO2H, benzaldehyde, NaOAc, Ac2O/AcOH,
reflux, 8 h; d) K2CO3, DMF, 808C, 3 h. PPE=polyphosphate ester. See Table 3 for the definition of R1, R2, R3, R4, and
R5 in (ꢁ)-3–(ꢁ)-12.
pyruvate as the substrate and adenylate kinase as the catalyst.
The pyruvate resulting from cleavage of phosphoenolpyruvate
is reduced by lactate dehydrogenase using NADH as co-sub-
strate whose consumption is monitored photometrically at
340 nm.
tecting group was performed in N,N-dimethylformamide with
K2CO3, leading to compounds (ꢁ)-3–(ꢁ)-12. Thiazolopyrimi-
dines
(ꢁ)-3–(ꢁ)-12 were isolated as yellow solids with high melting
points (>2008C), and feature poor water solubility. Their stor-
age at room temperature did not lead to decomposition.
Titration of (ꢁ)-9 with NaOH and photometric titration of
(ꢁ)-9 in buffer solutions with different pH were used to evalu-
ate the pKa values of the dibromophenol moiety in compound
(ꢁ)-9. Both methods estimated a pKa value of around 7–8 (see
Supporting Information).
Hits were confirmed using the photometric assay that had
been used in the primary assay. Confirmed hits were checked
again using an assay that monitors the conversion of [1,3,4-
13C3]-1 into [1,3,4-13C3]-2 by 13C NMR spectroscopy.
The most active inhibitor identified by the library screen was
the thiazolopyrimidine derivative (ꢁ)-3 (for the substitution
pattern, see Table 3 below). Preliminary kinetic experiments in-
dicated that the compound also inhibits IspF orthologues from
P. falciparum and M. tuberculosis. On the basis of these prelimi-
nary results, compound (ꢁ)-3 and the derivatives (ꢁ)-4–(ꢁ)-
12 were synthesized and their structures unequivocally con-
firmed (Scheme 2).
An X-ray crystal structure of compound (ꢁ)-27 was obtained
(Figure 1), confirming the Z-conformation of the exocyclic
double bond. In the crystal packing, the dibromophenol
moiety undergoes a p–p stacking interaction with the pyrimi-
dine ring of a neighboring molecule (dC14···N4’ =3.57 ꢄ; see
figure 2 in the Supporting Information).
The Z-conformation of the exocyclic double bond in com-
pound (ꢁ)-21 was also investigated by HMBC and 13C NMR
spectroscopy (see figures 3 and 4 of the Supporting Informa-
tion). The 13C NMR (without decoupling) spectrum of (ꢁ)-21
Synthesis
The preparation of the thiazolopyrimidine ligands involved
three or four reaction steps (Scheme 2).[20] The first step was
the synthesis of dihydropyrimidine (DHPM) derivatives (ꢁ)-13–
(ꢁ)-18 using Biginelli conditions. Compounds (ꢁ)-13–(ꢁ)-18
were obtained in a one-pot multicomponent reaction by react-
ing a b-keto ester, thiourea, and the corresponding aldehyde
in ethanol or tetrahydrofuran at reflux; hydrochloric acid or
polyphosphate ester (PPE) was used as a catalyst.[21,22] The di-
hydropyrimidine carboxylic acids (ꢁ)-19 and (ꢁ)-20 were ob-
tained after deallylation of the allyl esters (ꢁ)-17 and (ꢁ)-18,
using [Pd(PPh3)4] and diethylamine in tetrahydrofuran.[23] Thia-
zolopyrimidines (ꢁ)-21–(ꢁ)-29 (for the definition of R1–R5, see
Table 3) were obtained by condensation of DHPMs with chloro-
acetic acid and the corresponding benzaldehyde in the pres-
ence of sodium acetate in acetic anhydride (Ac2O)/acetic acid
(AcOH) medium at reflux. Finally, removal of the phenolic pro-
3
showed a coupling constant of JCH =6.4 Hz between the vinyl-
ic proton and the carbonyl carbon, which is a typical value for
3
a Z-conformation. In the E-conformation, the JCH coupling con-
stant is expected to be higher.
All target compounds have one single (Z) configuration of
the double bond and possess one stereogenic center; they are
thus obtained as racemates. Therefore, we decided to separate
the racemic mixture of compound (ꢁ)-9 in order to investigate
the biological properties of each enantiomer.
Optical resolution
The enantiomers of 9 were obtained in enantiopure form by
HPLC on a chiral stationary phase. In order to assess the ten-
dency of the compound to racemize, solutions of (+)-9 and
1094
www.chemmedchem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1092 – 1101

Thiazolopyrimidine Inhibitors of IspF
the thermal energy window
(0.5 kcal), each of them is consid-
ered to be significantly present
in solution. CD spectra were
computed for these five con-
formers by considering the first
20 vertical electronic transitions
at the time dependent (TD)-
B3LYP/6-31G(d) level of theory
(see figure 6 of the Supporting
Information). The calculated CD
spectra for the five conformers
were similar in terms of the
energy and sign of the Cotton
effects. Then, all calculated CD
curves were Boltzmann-weight-
ed and summed to afford the
average curve in Figure 2. The
predicted CD spectrum for (R)-9
has a negative Cotton effect at
Figure 1. ORTEP at the 50% probability level of the crystal structure of inhibitor (ꢁ)-27 at 223 K with arbitrary
numbering scheme. The dihydropyrimidine ring adopts a half-chair conformation with C5 sticking out of the
plane and the thiophene ring in axial position.
(ꢀ)-9 in Tris hydrochloride (pH 8) were incubated at room tem-
perature (i.e., under the conditions used for enzyme assays),
and circular dichroism (CD) was monitored at intervals. The
enantiomers were found to be stable for extended periods as
no change in the CD spectra was observed after one week at
room temperature.
380 nm and a positive one at 280 nm, reproducing the experi-
mental result very well in terms of wavelength and relative in-
tensity of the bands. The comparison of the calculated and the
experimental CD curves suggests that the first-eluted enantio-
mer (ꢀ)-9 is the R isomer.
In order to assign the absolute configuration of the enantio-
mers of 9, we applied a quantum-mechanical method to simu-
late the CD spectrum for comparison with the experimental
data.[24] The calculations were performed on conformers of (R)-
9a. At first, we scanned the potential energy surface by sys-
tematic variation of the torsion angles C15ꢀC20ꢀC44ꢀO45 and
N16ꢀC15ꢀC13ꢀO14 (see figure 5 of the Supporting Informa-
tion). The initial three-dimensional models were optimized at
the B3LYP/6-31G(d) level of theory without geometric con-
straints, resulting in five conformers (R)-9a–e within an energy
window of 0.2 kcalmolꢀ1. The absence of imaginary frequen-
cies from the harmonic vibrational analysis confirmed their
identity as true minima. Because all the conformers are within
Inhibition of IspF proteins
Derivatives of compound (ꢁ)-3 were obtained by modifica-
tions at five positions: at the pyrimidine ring (R1; see Scheme 2
and Table 3 for details); at the ester (R2), and at the phenolic
ring (substituents R3, R4, and R5). IC50 values for the compounds
under study were obtained by the photometric assay that had
been used for the primary screen. Dilution series were assayed
in 96-well plates that were monitored continuously. Initial rates
were extracted computationally and were used as input for
computational assessment of the IC50 values.
The photometric assay uses four auxiliary enzymes in order
to generate a chromogenic signal (Scheme 1). It was therefore
important to check the inhibition data from this assay by an in-
dependent method in order to rule out that the results were
influenced by interaction of the study compounds with auxili-
ary enzymes. As a counter-screen, we used an NMR-based
assay directly monitoring the consumption of 1 and the pro-
duction of 2. The assay mixture contained only buffer, Mg2+
ions, IspF protein, and multiply 13C-labeled substrate ([1,3,4-
13C3]-1). Due to the multiple labeling, the 13C NMR signals of
substrate and product appear as highly characteristic multip-
lets (Table 2). Reaction mixtures containing inhibitors in se-
quential dilutions were incubated for a predetermined time
that was sufficient to reach about 50% turnover in samples
without inhibitor. It should be noted that the data represent
end point measurements, as opposed to the photometric
assays, which were based more rigorously on initial rates (ini-
tial rate measurements are not possible with NMR detection
due to insufficient sensitivity). Due to this technical difference,
Figure 2. CD spectra. Experimental spectra in CHCl3: (ꢀ)-9, blue line; (+)-9,
red line (enantiomeric ratio 100:0). Boltzmann-weighted CD simulation for
(R)-9 at the TD-B3LYP/6-31G(d) level of theory, green line.
ChemMedChem 2010, 5, 1092 – 1101
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1095

F. Diederich, M. Fischer, M. Groll, et al.
MED
parum protein. By comparison, the IC50 values for the E. coli
protein were generally higher (Table 3).
Table 2. 13C NMR data of the reactant used and the product of the IspF-
catalyzed transformation.
Halogen substituents R3 and R5 in the phenolic ring are cru-
cial for activity; maximum inhibitory activity is observed when
both R3 and R5 are bromine atoms (entries 1, 5–9, and 11–12,
Table 3). Of similar importance is the phenolic group. Esterifica-
tion of the phenolic hydroxy group is accompanied by a major
decrease in inhibitory activity (entries 15, 16, Table 3). Within
the limits of experimental accuracy, enantiomers of 9 showed
the same respective IC50 values (entries 8, 9, Table 3). This sug-
gests that the R1 substituent on the pyrimidine ring has no sig-
nificant interaction with the protein. In line with that, variation
of the R1 substituent had little, if any, influence on inhibitory
activity. At present, the structure–activity relationship data do
not allow the proposal of a preferred docking mode of the in-
hibitors in the IspF enzymes.
Atom
position
d [ppm]
J
CC [Hz]
JPC [Hz]
[1,3,4-13C3]-4-diphosphocytidyl-2C-methyl-d-erythritol 2-phosphate (1):[18]
1
3
4
65.9
74.0
67.2
–
42.9
42.9
2.0, 1.8
7.1, 7.1
5.8
[1,3,4-13C3]-2C-methyl-d-erythritol 2,4-cyclodiphosphate (2):[25]
1
3
4
66.9
68.4
65.7
–
42.7
42.7
–
–
6.6
the IC50 values from the two methods should not be expected
to be numerically identical.
As an example, Figure 3 shows enzyme inhibition data for
(ꢁ)-9 obtained by the photometrically monitored multicompo-
nent assay and the NMR assay. The data leave no doubt that
the study compounds target IspF protein and not the auxiliary
enzymes.
Red blood cell assay with P. falciparum
The antimalarial activity of (ꢁ)-3–(ꢁ)-12 was evaluated using
erythrocytes infected by P. falciparum (Table 4). Chloroquine
and artesunate were used as controls, and gave IC50 values in
the expected range. Dihydropyrimidine intermediates (ꢁ)-15
and (ꢁ)-16 were also tested as such scaffolds are known to ex-
hibit antiviral, antibacterial, and anti-inflammatory activities,[20]
but they had no effect on parasite growth. In this assay, com-
pounds (ꢁ)-3, (ꢁ)-5, and (ꢁ)-11 were the most active com-
pounds, exhibiting IC50 values of 1.4–1.6 mm. More polar li-
gands, such as the carboxylic acid derivatives (ꢁ)-8 and (ꢁ)-
12, had much higher IC50 values (>17 mm and >20 mm, respec-
tively).
Some compounds studied ((ꢁ)-3, (ꢁ)-7, (ꢁ)-10, (ꢁ)-11) had
IC50 values in the single-digit micromolar range for IspF protein
of M. tuberculosis and low double-digit values for the P. falci-
Discussion
As stated in numerous recent publications, there is universal
consensus about the dire need for novel anti-infective drugs.
Whereas the currently used therapeutic agents are subject to
rapid attrition by resistance development, most major drug
companies have curtailed or terminated their antibiotics dis-
covery programs for a variety of reasons that have been dis-
cussed elsewhere.[26]
Two diseases of acute concern are malaria and tuberculosis.
The global incidence of malaria is in the range of several hun-
dred million infections per year, and about one third of the
global population is believed to be infected by M. tuberculosis,
the causative agent of tuberculosis. Each of these diseases
causes more than a million deaths per year.[27,28] Multidrug-re-
sistant forms of M. tuberculosis are spreading rapidly, and some
of these infections are resistant to all currently available drugs.
Resistance against all currently available antimalarial agents is
also progressing rapidly, and it has therefore been argued that
novel antimalarial agents would have to be introduced with a
cadence of about five years in order to stem the tide.
Figure 3. Enzyme inhibition by (ꢁ)-9. A) IspF protein of M. tuberculosis;
B) IspF of P. falciparum. Data from the photometric assay (initial reaction
rate) are in red, data from the NMR assay (endpoint measurements) are in
blue. Experimental data are indicated by symbols, fit curves are indicated by
lines. The kinetic data were evaluated with a nonlinear regression method
using the program Dynafit.[34]
It has been argued that the genomic revolution following
the unraveling of the Haemophilus influenzae genome has
failed to deliver the expected dividends in terms of novel anti-
biotics.[29,30] With regard to the topic of this study, it may be
1096
www.chemmedchem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1092 – 1101

Thiazolopyrimidine Inhibitors of IspF
Table 3. Inhibition of IspF enzymes by thiazolopyrimidines (photometric assay).
Entry
Ligand
Substituents
IC50 [mm]
P. falciparum
R1
R2
R3
R4
R5
A. thaliana
M. tuberculosis
E. coli
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
(ꢁ)-3
(ꢁ)-4
(ꢁ)-5
(ꢁ)-6
(ꢁ)-7
(ꢁ)-8
(ꢁ)-9
(S)-9
2-thiophenyl
2-thiophenyl
2-thiophenyl
2-thiophenyl
2-thiophenyl
2-thiophenyl
2-benzofuranyl
2-benzofuranyl
2-benzofuranyl
2-benzofuranyl
4-methoxyphenyl
Me
Et
Et
Et
Et
Bn
H
Et
Et
Et
Et
Et
H
Br
F
Br
H
Br
Br
Br
Br
Br
Cl
Br
Br
–
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
–
Br
F
H
6.8ꢁ0.6
102ꢁ6
130ꢁ12
>300
9.6ꢁ1.5
89ꢁ12
6.1ꢁ0.8
322ꢁ42
125ꢁ18
>300
32ꢁ7
>300
208ꢁ22
212ꢁ19
>300
H
>300
Br
Br
Br
Br
Br
Cl
Br
Br
–
12ꢁ1
35ꢁ5
87ꢁ7
32ꢁ3
32ꢁ2
22ꢁ2
18ꢁ2
11ꢁ2
56ꢁ5
2.1ꢁ0.5
141ꢁ43
13ꢁ2
69ꢁ4
223ꢁ23
58ꢁ9
60ꢁ13
39ꢁ7
36ꢁ6
18ꢁ3
192ꢁ16
–
77ꢁ9
7.9ꢁ0.6
8.9ꢁ1.8
15ꢁ1
14ꢁ2
(R)-9
18ꢁ3
(ꢁ)-10
(ꢁ)-11
(ꢁ)-12
(ꢁ)-15
(ꢁ)-16
(ꢁ)-21
(ꢁ)-27
7.9ꢁ0.9
3.2ꢁ0.4
69ꢁ5
7.4ꢁ0.9
5.1ꢁ0.4
65ꢁ8
4-methoxyphenyl
2-thiophenyl
2-benzofuranyl
2-thiophenyl
Et
Bn
Et
Et
>300
>300
>300
>300
>300
>300
>300
>300
>300
–
Br
H
–
OAc
OMe
–
Br
H
>300
>300
>300
–
>300
–
we decided to use a pseudomature form of a plant enzyme as
a surrogate target. The surrogate target strategy resulted in
the identification of the screening hit (ꢁ)-3 that was found to
inhibit the IspF enzymes from M. tuberculosis and P. falciparum,
besides the A. thaliana enzyme. IC50 values for inhibition of
IspF protein and for proliferation of P. falciparum in human er-
ythrocytes are of similar values in the single-digit micromolar
range (Table 4), although there is no strict parallel between the
two experimental parameters. The lack of strict proportionality
between the enzyme assay and the parasite assay may be
caused by differences in the pharmacological properties in-
cluding take-up and metabolism of different derivatives inside
erythrocytes and parasites. However, we note that there is as
yet no formal proof that the toxicity for the parasites is caused
exclusively or predominantly by inhibition of IspF protein.
To the best of our knowledge, this is the first report on the
identification of heterocyclic inhibitors of IspF protein by
screening of a random compound library. Despite the relatively
small size of the library, IC50 values in the low single-digit
range were reached. The technology has documented its po-
tential and is ready for the screening of larger libraries.
Table 4. Inhibition of P. falciparum strains NF54 as measured by red
blood cell assay.[a]
Compound
IC50 [mm]
IC50 [ngmLꢀ1
]
chloroquine
artesunate
(ꢁ)-3
0.013
0.007
1.5
1.6
3.2
4.3
2.7
858
829
1362
(ꢁ)-5
(ꢁ)-6
(ꢁ)-7
2.0
1304
(ꢁ)-8
>18
3.9
>10000
2417
1012
904
>10000
1941
(ꢁ)-9
(ꢁ)-10
(ꢁ)-11
(ꢁ)-12
(ꢁ)-15
(ꢁ)-16
(ꢁ)-27
1.9
1.5
>20
6.3
17.6
2.8
6061
1246
[a] Values are the mean of two or more independent experiments.
worth noting that the rapid elucidation of the non-mevalonate
isoprenoid pathway would have been hardly possible without
the tools of comparative genomics.[31] However, it is true that
the use of this knowledge for the actual development of novel
chemical entities capable of defeating pathogens by inhibiting
their non-mevalonate pathway enzymes is still in its infancy.
The only potential antimalarial agent directed at an enzyme of
the isoprenoid pathway (IspC protein) that has made it to the
clinical level is fosmidomycin, an antibiotic whose develop-
ment had been abandoned in the 1980, but which has made a
tentative comeback and has been shown to have activity
against human malaria in small clinical studies.[1,32]
Experimental Section
For general details and procedures see the Supporting Informa-
tion.
Library: A library of about 40000 synthetic compounds was pur-
chased from ChemDiv Inc. (San Diego, USA), with molecular
weights ranging from 150–650 Da.
In vitro antimalarial activity: in vitro antimalarial activity was mea-
sured using a [3H]hypoxanthine incorporation assay with the P. fal-
ciparum strain NF54. Results were expressed as the concentration
resulting in 50% inhibition (IC50). See the Supporting Information
for more details.
Conclusions
Due to technical problems with the use of IspF proteins from
P. falciparum and M. tuberculosis for high-throughput screening,
ChemMedChem 2010, 5, 1092 – 1101
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1097

F. Diederich, M. Fischer, M. Groll, et al.
MED
and the mixtures were analyzed by 13C NMR spectroscopy. Sub-
strate/product ratios were obtained from the signals of C-3 of
[1,3,4-13C3]-1 (74.0 ppm) and C-3 of [1,3,4-13C3]-2 (68.4 ppm).
Biology
Materials: [1,3,4-13C3]-1 was prepared as reported earlier.[18] Re-
combinant IspF protein from P. falciparum and from E. coli, respec-
tively, and cytidylate kinase from E. coli were prepared as described
elsewhere.[18] Lactate dehydrogenase and adenylate kinase were
purchased from Sigma (Switzerland). NADH was purchased from
Acros Organics (Switzerland).
Chemistry
The experimental details for the syntheses of compounds (ꢁ)-3–
(ꢁ)-12, (ꢁ)-15, (ꢁ)-16, (ꢁ)-21, and (ꢁ)-27 are described below. All
other experimental data are given in the Supporting Information.
Expression of the A. thaliana IspF gene: A synthetic open reading
frame was generated by a sequence of PCR reactions using pair-
wise combinations of oligonucleotides, shown in table 1 in the
Supporting Information, as primers (starting with the oligonucleo-
tides “Initial Fw” and “Initial Rev”). After each PCR extension, the re-
spective amplification product was purified by agarose gel electro-
phoresis and was then used as template in the next amplification
round. The final amplification product was digested with the re-
striction endonucleases BamHI and SalI and was then cloned into
the plasmid pQE30. The resulting plasmid pQE-ispF-At (GeneBank
accession number GU828015) was transformed into E. coli strain
XL1-Blue.
General procedure 1 for the synthesis of dihydropyrimidines:
1a) A mixture of ester (1.0 equiv), thiourea (1.0 equiv), and alde-
hyde (1.0 equiv) in EtOH (1 mLmmolꢀ1) was treated with concd
HCl (37% in H2O, 30 mLmmolꢀ1) as catalyst and then kept at reflux
for 6 h. 1b) A mixture of ester (1.0 equiv), thiourea (1.0 equiv), and
aldehyde (1.0 equiv) in THF (1 mLmmolꢀ1) was treated with PPE
(150 mgmmolꢀ1) as catalyst and then kept at reflux for 6 h.[23] The
crude products were either purified by column chromatography
on silica gel or washed with 50% aq EtOH.
General procedure 2 for the synthesis of thiazolopyrimidines: A
mixture of dihydropyrimidine (1.0 equiv), chloroacetic acid
(1.0 equiv), benzaldehyde (1.0 equiv), and sodium acetate
(2.5 equiv) in Ac2O (1.5 mLmmolꢀ1) and AcOH (2 mLmmolꢀ1) was
kept at reflux for 8 h. The crude products were purified by column
chromatography on silica gel.
Expression of the M. tuberculosis IspF gene: The putative open
reading frame NC_000962 was amplified using the oligonucleo-
tides IspFMYCvo and IspFMYChi (see table 1 in the Supporting In-
formation) as primers and DNA of M. tuberculosis H37Rv as a tem-
plate. The amplification product was digested with BamHI and PstI
and was ligated into the plasmid pQE30, which had been treated
with the same restriction enzymes. The recombinant plasmid was
electrotransformed into E. coli strain XL1-Blue.[33]
General procedure 3 for the deprotection of thiazolopyrimi-
dines: The protected phenol (1.0 equiv) was stirred with K2CO3
(1.6 equiv) at 808C in DMF (4 mLmmolꢀ1). The mixture was stirred
for 3 h, diluted with H2O (100 mL), and then adjusted to pH 4 with
1n HCl. The aqueous phase was extracted with EtOAc. The com-
bined organic phases were washed with H2O and dried over
MgSO4. The solvent was removed in vacuo. The crude product was
purified by column chromatography on silica gel.
Protein purification: Recombinant IspF proteins were purified by
metal-affinity chromatography using a previously published gener-
al procedure.[19]
High-throughput screening:[18] Assays were conducted in 384-well
polystyrene plates (Nunc, cat. no. 262160). Aliquots (20 mL) of a so-
lution containing 1.2 mm NADH, 6 mm MgCl2, 18 mm phosphoe-
nolpyruvate, 6 mm ATP, pyruvate kinase (3 U), CMP kinase (1.5 U),
and A. thaliana IspF protein (3 mm) in 300 mm Tris HCl, pH 8.0,
were added to wells preloaded with 20 mL of 9% v/v (CH3)2SO con-
taining test compound (60–100 mm). The reaction was started by
adding 20 mL aliquots of 3 mm 1. Each plate was monitored photo-
metrically (340 nm) during a period of 30 min using a photometric
plate reader operating at room temperature. Initial rates were esti-
mated using in-house software.
(ꢁ)-Ethyl (2Z)-2-(3,5-Dibromo-4-hydroxybenzylidene)-7-methyl-
3-oxo-5-(2-thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimi-
dine-6-carboxylate ((ꢁ)-3): General procedure 3 starting from (ꢁ)-
22 (219 mg, 0.35 mmol). Purification by chromatography (SiO2;
hexane/Et2O, 5:3) gave compound (ꢁ)-3 as a yellow solid (108 mg,
53%): Rf =0.30 (hexane/EtOAc, 1:1); mp: 2308C (dec.); 1H NMR
(300 MHz, CDCl3): d=1.21 (t, J=7.1 Hz, 3H, CO2CH2CH3), 2.53 (s,
3H, CH3), 4.11–4.20 (m, 2H, CO2CH2CH3), 6.23 (brs, 1H, OH), 6.52 (s,
1H, CH), 6.92 (t, J=3.6 Hz, 1H, HAr), 7.08 (d, J=3.5 Hz, 1H, HAr),
7.23 (d, J=5.0 Hz, 1H, HAr), 7.59 (s, 2H, C6H2Br2), 7.60 ppm (s, 1H,
CH); 13C NMR (75 MHz, CDCl3): d=14.1, 22.7, 49.8, 60.7, 108.8,
110.2, 120.7, 126.2, 126.9 (2C), 128.3, 130.1, 133.4 (2C), 142.4, 145.8,
IC50 determination using the photometric assay: Assays were
conducted in 96-well plates (Nunc, cat. no. 781602) with transpar-
ent flat bottoms. Assay mixtures (total volume=200 mL) contained
100 mm Tris HCl, pH 8.0, 0.4 mm NADH, 1 mm ATP, 1 mm phos-
phoenolpyruvate, 10 mm MgCl2, 0.02 U of IspF protein, 0.5 U of
CMP kinase, 0.5 U of pyruvate kinase, and 0.5 U of lactate dehydro-
genase. Dilution series (1:2) of inhibitors covered the concentration
range of approximately 200–0.8 mm. The reaction was started by
addition of 1 in 100 mm Tris HCl, pH 8.0, to a final concentration of
1 mm. The reaction was monitored photometrically (room temper-
ature) at 340 nm using a plate reader (Multican Spectrum, Thermo
Electron Corporation). Initial rate values were evaluated with a
nonlinear regression method using the program Dynafit.[34]
~
147.3, 151.0, 153.0, 164.7, 165.2 ppm; IR (ATR): n=2981, 1702,
1542, 1474, 1372, 1296, 1234, 1152, 1080, 1056, 706 cmꢀ1; HRMS-
MALDI: m/z [M+H]+ calcd for C21H17Br2N2O4S2+: 582.8991, found:
582.8994.
(ꢁ)-Ethyl (2Z)-2-(3,5-Difluoro-4-hydroxybenzylidene)-7-methyl-3-
oxo-5-(2-thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-
6-carboxylate ((ꢁ)-4): General procedure 3 starting from (ꢁ)-23
(110 mg, 0.22 mmol). Purification by chromatography (SiO2;
hexane/EtOAc, 1:1!100% EtOAc) gave compound (ꢁ)-4 as a
yellow solid (35 mg, 35%): Rf =0.40 (hexane/EtOAc, 1:1); mp:
2708C (dec.); 1H NMR (300 MHz, (CD3)2SO): d=1.15 (t, J=7.1 Hz,
3H, CO2CH2CH3), 2.39 (s, 3H, CH3), 4.05–4.15 (m, 2H, CO2CH2CH3),
6.34 (s, 1H, CH), 6.95–7.00 (m, 2H, HAr), 7.33 (dd, J=7.6, 1.6 Hz, 2H,
C6H2F2), 7.47 (dd, J=5.0, 1.2 Hz, 1H, HAr), 7.78 ppm (s, 1H, CH);
13C NMR (75 MHz, (CD3)2SO): d=13.9, 22.3, 49.3, 60.3, 108.2, 113.6
(dd, 2JC,F =22.4, 4JC,F =6.0 Hz, 2C), 118.4, 122.6 (t, 3JC,F =8.8 Hz),
IC50 determination by NMR: Reaction mixtures contained 100 mm
Tris HCl, pH 8.0, 2 mm MgCl2, 1 mm [1-13C1]glucose (internal stan-
dard), 3.4 mm [1,3,4-13C3]-1, 0–200 mm inhibitor, 1–10 mg IspF pro-
tein, 3% v/v (CH3)2SO, and 10% v/v D2O in a volume of 500 mL.
They were incubated at 378C for 45 min. The reactions were termi-
nated by the addition of EDTA to a final concentration of 20 mm,
2
126.4, 126.8, 127.1, 132.0 (brs), 137.0 (t, JC,F =16.3 Hz), 142.6, 151.5
1098
www.chemmedchem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1092 – 1101

Thiazolopyrimidine Inhibitors of IspF
(dd, 1JC,F =243.5, 3JC,F =7.9 Hz, 2C), 152.1, 155.0, 164.1, 164.7 ppm;
6.95–7.00 (m, 2H, HAr), 7.47 (d, J=5.0 Hz, 1H, HAr), 7.78–7.79 (m,
3H, CH, C6H2Br2), 12.81 ppm (brs, 1H, OH); 13C NMR (75 MHz,
(CD3)2SO): d=22.2, 49.2, 108.1, 112.2 (2C), 118.8, 126.3, 126.7, 126.9
(2C), 130.6, 133.7 (2C), 142.6, 151.3, 152.9, 154.1, 163.9, 166.3 ppm;
19
~
F NMR (282 MHz, (CD3)2SO): d=ꢀ129.4 ppm (s, 2F); IR (ATR): n=
3027, 2922, 2851, 1719, 1677, 1593, 1522, 1440, 1319, 1231, 1168,
1098, 1015, 790, 706 cmꢀ1; HRMS-MALDI: m/z [M+H]+ calcd for
C21H17F2N2O4S2+: 463.0592, found: 463.0590.
~
IR (ATR): n=2922, 2850, 1682, 1579, 1537, 1474, 1372, 1293, 1248,
1153, 1059, 702 cmꢀ1
;
HRMS-MALDI: m/z [M+H]+ calcd for
C19H13Br2N2O4S2+: 554.8678, found: 554.8653.
(ꢁ)-Ethyl
(2Z)-2-(3-Bromo-4-hydroxybenzylidene)-7-methyl-3-
oxo-5-(2-thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-
6-carboxylate ((ꢁ)-5): General procedure 3 starting from (ꢁ)-24
(136 mg, 0.25 mmol). After filtration, compound (ꢁ)-5 was ob-
tained as a yellow solid (125 mg, 99%): Rf =0.50 (EtOAc/hexane,
2:1); mp: 2408C (dec.); 1H NMR (300 MHz, CDCl3): d=1.21 (t, J=
7.1 Hz, 3H, CO2CH2CH3), 2.54 (s, 3H, CH3), 4.11–4.22 (m, 2H,
CO2CH2CH3), 6.52 (s, 1H, CH), 6.92 (dd, J=5.1, 3.5 Hz, 1H, HAr),
7.08–7.12 (m, 2H, HAr), 7.22 (d, J=5.7 Hz, 1H, HAr), 7.37 (dd, J=8.5,
2.6 Hz, 1H, HAr), 7.62 (d, J=2.3 Hz, 1H, HAr), 7.69 ppm (s, 1H, CH);
13C NMR (75 MHz, CDCl3): d=14.1, 22.7, 49.7, 60.6, 108.7, 111.1,
116.9, 119.1, 126.1, 126.9 (2C), 127.4, 131.1, 131.8, 134.0, 142.6,
(ꢁ)-Ethyl (2Z)-5-(1-Benzofuran-2-yl)-2-(3,5-dibromo-4-hydroxy-
benzylidene)-7-methyl-3-oxo-2,3-dihydro-5H-[1,3]thiazolo[3,2-
a]pyrimidine-6-carboxylate ((ꢁ)-9): General procedure 3 starting
from (ꢁ)-21 (301 mg, 0.45 mmol). Purification by chromatography
(SiO2; hexane/Et2O, 5:3) gave compound (ꢁ)-9 as a yellow solid
(117 mg, 41%): Rf =0.60 (EtOAc/pentane, 2:1); mp: 2308C (dec.);
1H NMR (300 MHz, CDCl3): d=1.26 (t, J=7.1 Hz, 3H, CO2CH2CH3),
2.54 (s, 3H, CH3), 4.18 and 4.20 (2q, J=7.1 Hz, 2H, CO2CH2CH3),
6.49 (s, 1H, CH), 6.74 (s, 1H, HAr), 7.16–7.29 (m, 2H, HAr), 7.42 (d, J=
7.9 Hz, 1H, HAr), 7.52 (d, J=7.8 Hz, 1H, HAr), 7.59 ppm (brs, 3H, CH,
C6H2Br2); 13C NMR (75 MHz, CDCl3): d=14.2, 22.9, 49.1, 60.7, 105.7,
110.7, 111.7, 120.6, 121.4, 122.3, 124.8, 127.8, 128.3, 130.2, 133.5
(2C), 151.0, 153.2, 154.3, 154.8, 155.9, 163.8, 164.7, 165.1 ppm (1
~
153.2, 154.1, 155.4, 165.1, 165.3 ppm; IR (ATR): n=2923, 2851,
1719, 1676, 1585, 1523, 1362, 1323, 1220, 1153, 1122, 813,
708 cmꢀ1; HRMS-MALDI: m/z [M+H]+ calcd for C21H18BrN2O4S2
:
+
504.9886, found: 504.9881.
~
signal not visible); IR (ATR): n=3294, 2982, 1707, 1541, 1473, 1452,
1372, 1247, 1153, 1058, 748 cmꢀ1; HRMS-MALDI: m/z [M+H]+ calcd
for C25H19Br2N2O5S++: 616.9376; found: 616.9363. Enantiomeric sep-
aration was performed by recycling HPLC using a WHELK-O
column (CH2Cl2/hexane/EtOAc, 70:20:10). (ꢀ)-(R)-9 [a]2D0 =ꢀ410
(c=0.5, CHCl3), (+)-(S)-9 [a]2D0 =+405 (c=0.5, CHCl3).
(ꢁ)-Ethyl
(2Z)-2-(4-Hydroxybenzylidene)-7-methyl-3-oxo-5-(2-
thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-6-carbox-
ylate ((ꢁ)-6): General procedure 3 starting from (ꢁ)-25 (100 mg,
0.21 mmol). Purification by chromatography (SiO2; hexane/EtOAc,
2:1) gave compound (ꢁ)-6 as a yellow solid (30 mg, 33%): Rf =
0.30 (hexane/EtOAc, 2:1); mp: 242–2448C; 1H NMR (300 MHz,
CDCl3): d=1.21 (t, J=7.1 Hz, 3H, CO2CH2CH3), 2.54 (s, 3H, CH3),
4.11–4.22 (m, 2H, CO2CH2CH3), 5.59 (s, 1H, OH), 6.53 (s, 1H, CH),
6.90–6.94 (m, 3H, HAr, C6H4), 7.09 (d, J=3.5 Hz, 1H, HAr), 7.22 (d, J=
5.1 Hz, 1H, HAr), 7.40 (d, J=8.7 Hz, 2H, C6H4), 7.77 ppm (s, 1H, CH);
13C NMR (75 MHz, CDCl3): d=14.1, 22.6, 49.6, 60.6, 108.4, 116.4 (2C),
117.2, 125.9, 126.1, 126.8, 126.9 (2C), 132.4 (2C), 133.9, 142.8, 153.3,
(ꢁ)-Ethyl
(2Z)-5-(1-Benzofuran-2-yl)-2-(3,5-dichloro-4-hydroxy-
benzylidene)-7-methyl-3-oxo-2,3-dihydro-5H-[1,3]thiazolo[3,2-
a]pyrimidine-6-carboxylate ((ꢁ)-10): General procedure 3 starting
from (ꢁ)-28 (1.0 g, 1.7 mmol). Purification by chromatography
(SiO2; hexane/CH2Cl2/EtOAc, 6:3:1) gave compound (ꢁ)-10 as a
yellow solid (424 mg, 46%): Rf =0.60 (EtOAc); mp: 2108C (dec.);
1H NMR (400 MHz, CDCl3): d=1.26 (t, J=7.1 Hz, 3H, CO2CH2CH3),
2.54 (s, 3H, CH3), 4.18 and 4.20 (2q, J=7.2 Hz, 2H, CO2CH2CH3),
6.49 (s, 1H, CH), 6.74 (s, 1H, HAr), 7.19 (t, J=7.9 Hz, 1H, HAr), 7.24–
7.28 (m, 1H, HAr), 7.41 (s, 2H, C6H2Cl2), 7.41–7.23 (m, 1H, HAr), 7.52
(d, J=7.6 Hz, 1H, HAr), 7.60 ppm (s, 1H, CH); 13C NMR (100 MHz,
CDCl3): d=14.2, 22.9, 49.1, 60.8, 105.7, 105.8, 111.7, 120.6, 121.4
(2C), 122.2, 122.9, 124.8, 127.0, 127.8, 129.7 (2C), 130.5, 149.5,
~
156.3, 157.9, 165.4 ppm; IR (ATR): n=2921, 1709, 1692, 1573, 1534,
1511, 1290, 1242, 1148, 1056, 829, 700 cmꢀ1; HRMS-MALDI: m/z
[M+H]+ calcd for C21H19N2O4S2+: 427.0781, found: 427.0785; Anal.
calcd for C21H18N2O4S2: C 59.14, H 4.25, N 6.57, found: C 59.22, H
4.43, N 6.30.
(ꢁ)-Benzyl (2Z)-2-(3,5-Dibromo-4-hydroxybenzylidene)-7-methyl-
3-oxo-5-(2-thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimi-
dine-6-carboxylate ((ꢁ)-7): General procedure 3 starting from (ꢁ)-
26 (78 mg, 0.11 mmol). Purification by chromatography (SiO2;
hexane/Et2O, 7:5) gave compound (ꢁ)-7 as a brown solid (25 mg,
34%): mp: 2108C (dec.); Rf =0.60 (EtOAc); 1H NMR (300 MHz,
CDCl3): d=2.54 (s, 3H, CH3), 5.08 (d, J=12.4 Hz, 1H, CH2Ph), 5.20
(d, J=12.4 Hz, 1H, CH2Ph), 6.54 (s, 1H, CH), 6.91 (dd, J=5.1, 3.6 Hz,
1H, HAr), 7.06 (d, J=3.6 Hz, 1H, HAr), 7.14–7.17 (m, 2H, HAr), 7.23
(dd, J=5.0, 1.1 Hz, 1H, HAr), 7.29–7.31 (m, 3H, HAr), 7.58 (s, 2H,
C6H2Br2), 7.60 ppm (s, 1H, CH); 13C NMR (75 MHz, CDCl3): d=22.9,
49.8, 66.4, 108.2, 110.6 (2C), 120.6, 126.3, 126.9, 127.2, 127.8 (2C),
128.0, 128.1, 128.4 (2C), 130.1, 133.3, 135.5 (2C), 142.2, 150.9, 153.7,
~
153.2, 154.3, 154.8, 155.3, 164.7, 165.1 ppm; IR (ATR): n=1718,
1681, 1532, 1367, 1301, 1235, 1156, 1060, 735 cmꢀ1; HRMS-MALDI:
m/z [M+H]+ calcd for C25H19Cl2N2O5S+: 529.0386, found: 529.0385.
(ꢁ)-Ethyl
(2Z)-2-(3,5-Dibromo-4-hydroxybenzylidene)-5-(4-me-
thoxyphenyl)-7-methyl-3-oxo-2,3-dihydro-5H-[1,3]thiazolo[3,2-
a]pyrimidine-6-carboxylate ((ꢁ)-11): General procedure 3 starting
from (ꢁ)-29 (920 mg, 1.4 mmol). Purification by chromatography
(SiO2; hexane/EtOAc/CH2Cl2, 6:2:1) gave compound (ꢁ)-11 as a
yellow solid (560 mg, 65%): Rf =0.50 (EtOAc/pentane, 2:1); mp:
172–1748C; 1H NMR (300 MHz, CDCl3): d=1.19 (t, J=7.1 Hz, 3H,
CO2CH2CH3), 2.51 (s, 3H, CH3), 3.76 (s, 3H, OCH3), 4.09 and 4.11 (2q,
J=7.1 Hz, 2H, CO2CH2CH3), 6.14 (s, 1H, CH), 6.82 (d, J=8.8 Hz, 2H,
C6H4), 7.32 (d, J=8.8 Hz, 2H, C6H4), 7.51 (s, 1H, CH), 7.56 ppm (s,
2H, C6H2Br2); 13C NMR (75 MHz, CDCl3): d=14.3, 22.9, 55.2, 55.3,
60.6, 109.4, 110.5 (2C), 113.7 (2C), 120.7, 128.2, 129.1 (2C), 129.4
(2C), 131.8; 133.1 (2C), 150.6, 151.6, 159.4, 164.6, 165.1 ppm; IR
~
154.9, 164.5, 164.8 ppm; IR (ATR): n=3064, 2921, 1709, 1605, 1538,
1475, 1374, 1227, 1153, 1054, 973, 785, 695 cmꢀ1; HRMS-MALDI:
m/z [M+H]+ calcd for C26H19Br2N2O4S2+: 644.9147, found: 644.9151.
(ꢁ)-(2Z)-2-(3,5-Dibromo-4-hydroxybenzylidene)-7-methyl-3-oxo-
5-(2-thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-6-car-
boxylic acid ((ꢁ)-8): General procedures 2 and 3 starting from (ꢁ)-
19 (141 mg, 0.50 mmol). Purification by chromatography (SiO2;
hexane/EtOAc/CH2Cl2, 5:1:1) gave compound (ꢁ)-8 as a brown
solid (35 mg, 12%, two steps): Rf =0.10 (EtOAc); mp: 2208C (dec.);
1H NMR (300 MHz, (CD3)2SO): d=2.40 (s, 3H, CH3), 6.33 (s, 1H, CH),
~
(ATR): n=2361, 1705, 1602, 1576, 1538, 1473, 1386, 1302, 1239,
1159, 1084, 1059, 1034 cmꢀ1; HRMS-MALDI: m/z [M+H]+ calcd for
C24H21Br2N2O5S+: 606.9532, found: 606.9538; Anal. calcd for
C24H20Br2N2O5S: C 47.39, H 3.31, N 4.61, found: C 47.89, H 3.55, N
4.53.
ChemMedChem 2010, 5, 1092 – 1101
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1099

F. Diederich, M. Fischer, M. Groll, et al.
MED
(ꢁ)-(2Z)-2-(3,5-Dibromo-4-hydroxybenzylidene)-5,7-dimethyl-3-
oxo-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-6-carboxylic
acid ((ꢁ)-12): General procedures 2 and 3 starting from (ꢁ)-20
(300 mg, 1.6 mmol). Purification by chromatography (SiO2; hexane/
EtOAc/CH2Cl2, 6:2:1) gave compound (ꢁ)-12 as a yellow solid
(278 mg, 35%, two steps): Rf =0.37 (EtOAc/MeOH 9:1); mp: 233–
2358C; 1H NMR (400 MHz, (CD3)2SO): d=1.25 (d, J=6.2 Hz, 3H,
CH3), 2.29 (s, 3H, CH3), 5.08 (q, J=6.2 Hz, 1H, CH), 7.77 (s, 1H, CH),
7.81 ppm (s, 2H, C6H2Br2); 13C NMR (100 MHz, (CD3)2SO): d=20.0,
22.2, 47.7, 110.1, 112.3 (2C), 127.2, 129.4, 133.3, 113.5 (2C), 150.7,
(ꢁ)-Ethyl
(2Z)-2-(4-Methoxybenzylidene)-7-methyl-3-oxo-5-(2-
thienyl)-2,3-dihydro-5H-[1,3]thiazolo[3,2-a]pyrimidine-6-carbox-
ylate ((ꢁ)-27): General procedure 2 starting from (ꢁ)-13 (565 mg,
2.0 mmol). Purification by chromatography (SiO2; pentane/EtOAc/
CH2Cl2, 11:1:1) gave compound (ꢁ)-27 as a yellow solid (129 mg,
15%): Rf =0.20 (pentane/EtOAc/CH2Cl2, 11:1:1); mp: 152–1558C;
1H NMR (300 MHz, CDCl3): d=1.21 (t, J=7.1 Hz, 3H, CO2CH2CH3),
2.53 (s, 3H, CH3), 3.86 (s, 3H, OCH3), 4.11–4.21 (m, 2H, CO2CH2CH3),
6.53 (s, 1H, CH), 6.91 (dd, J=5.1, 3.6 Hz, 1H, HAr), 6.98 (d, J=8.8 Hz,
2H, C6H4), 7.09 (d, J=3.6 Hz, 1H, HAr), 7.21 (dd, J=5.1, 1.2 Hz, 1H,
HAr), 7.44 (d, J=9.0 Hz, 2H, C6H4), 7.78 ppm (s, 1H, CH); 13C NMR
(75 MHz, CDCl3): d=14.1, 22.7, 49.6, 55.5, 60.6, 108.3, 114.8 (2C),
117.1, 126.1, 126.7, 126.8 (2C), 132.1 (2C), 133.8, 142.8, 153.4, 156.1,
~
152.7, 154.7, 164.3, 166.7 ppm; IR (ATR): n=3347, 2818, 1681, 1534,
1476, 1374, 1294, 1252, 1143, 1142 cmꢀ1; HRMS-ESI: m/z [MꢀH]+
calcd for C16H11Br2N2O4S+: 484.8812, found: 484.8800.
~
161.5, 165.3, 165.4 ppm; IR (ATR): n=2979, 1704, 1589, 1537, 1508,
1384, 1361, 1297, 1257, 1225, 1153, 1057, 1019, 822 cmꢀ1; HRMS-
MALDI: m/z [M+H]+ calcd for C22H21N2O4S2+: 441.0937, found:
441.0944.
(ꢁ)-Ethyl 4-(4-Methoxyphenyl)-6-methyl-2-thioxo-1,2,3,4-tetrahy-
dro-pyrimidine-5-carboxylate ((ꢁ)-15): General procedure 1a
starting from thiourea (815 mg, 10.6 mmol). Compound (ꢁ)-15
was obtained as a colorless solid (2.5 g, 78%): Rf =0.60 (EtOAc/pen-
tane, 2:1); mp: 137–1408C; 1H NMR (300 MHz, CDCl3): d=1.17 (t,
J=7.1 Hz, 3H, CO2CH2CH3), 2.35 (s, 3H, CH3), 3.78 (s, 3H, OCH3),
4.09 (q, J=7.1 Hz, 2H, CO2CH2CH3), 5.33 (d, J=2.9 Hz, 1H, CH), 6.84
(d, J=8.6 Hz, 2H, C6H4), 6.84 (d, J=8.6 Hz, 2H, C6H4), 7.41 (brs, 1H,
NH), 8.03 ppm (brs, 1H, NH); 13C NMR (75 MHz, CDCl3): d=14.1,
18.3, 55.6, 55.7, 60.3, 103.2, 114.1 (2C), 128.0 (2C), 134.7, 142.7,
X-Ray analysis of (ꢁ)-27: Crystal data: C22H20N2O4S2, Mr =440.540,
ꢀ
triclinic space group P1, a=8.2219 (2), b=9.4602 (2), c=13.6744
(3) ꢄ, a=80.193 (1), b=84.872 (1), g=84.647 (1)8, z=2, Dx =
1.406 Mgmꢀ3. Data collection: From a yellow crystal of size 0.36ꢅ
0.28ꢅ0.008 mm, 8375 reflections were measured at 223 K on a
Bruker Kappa CCD diffractometer with Mo Ka radiation (l=
0.71073). q=2.425–27.4858. A total of 4066 of the 4752 independ-
ent reflections (Rint =0.03) were marked as observed (>2s(I)). The
structure was solved by direct methods with SIR97[35] and 332 pa-
rameters were refined with SHELXL97[36] by full matrix least-squares
(nonhydrogen atoms with anisotropic ADPs, hydrogen atoms with
isotropic U). Hydrogen atoms from the slightly disordered ethyl
group were calculated and included in the structure factor calcula-
tion. Robs =0.0514, Rall =0.0597, no absorption correction (m=
0.288 mmꢀ1). CCDC-765449 contains the supplementary crystallo-
graphic data for this paper and is available free of charge from the
Cambridge Crystallographic Data Centre (CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK; Tel.: (+44) 1223-336-408; Fax: (+44)
1223-336-033; E-mail: deposit@ccdc.cam.ac.uk).
~
159.5, 165.2, 174.3 ppm; IR (ATR): n=3309, 3167, 2982, 1665, 1573,
1508, 1455, 1267, 1194, 1169, 1120, 1026, 764, 653 cmꢀ1; HRMS-
MALDI: m/z [M+H]+ calcd for C15H19N2O3S+: 307.1111, found:
307.1114; Anal. calcd for C15H18N2O3S: C 58.80, H 5.92, N 9.14;
found: C 59.00, H 5.83, N 8.95.
(ꢁ)-Benzyl
6-Methyl-4-(2-thienyl)-2-thioxo-1,2,3,4-tetra-hydro-
pyrimidine-5-carboxylate ((ꢁ)-16): General procedure 1a starting
from thiourea (230 mg, 3.0 mmol). Purification by chromatography
(SiO2; cyclohexane/EtOAc/CH2Cl2, 4:1:1) gave compound (ꢁ)-16 as
a white powder (380 mg, 37%): Rf =0.20 (cyclohexane/EtOAc/
1
CH2Cl2, 4:1:1); mp: 153–1548C; H NMR (300 MHz, CDCl3): d=2.36
(s, 3H, CH3), 5.14 (d, J=12.4 Hz, 1H, CH2Ph), 5.15 (d, J=12.4 Hz,
1H, CH2Ph), 5.69 (d, J=3.4 Hz, 1H, CH), 6.88–6.93 (m, 2H, HAr),
7.18–7.23 (m, 3H, Ph), 7.28–7.33 (m, 3H, Ph, HAr), 7.72 (brs, 1H,
NH), 8.29 ppm (brs, 1H, NH); 13C NMR (75 MHz, CDCl3): d=18.3,
50.9, 66.4, 102.7, 124.9, 125.6, 126.9, 128.1, 128.2 (2C), 128.5 (2C),
Acknowledgements
~
135.6, 143.9, 145.5, 164.7, 174.8 ppm; IR (ATR): n=3153, 2988,
This research was supported by the Swiss National Science Foun-
dation, a graduate student fellowship from Novartis (to J.G.G.),
and by the Hans-Fischer-Gesellschaft. We acknowledge CESGA
(Santiago de Compostela, Spain) for the allocation of computa-
tional resources.
1708, 1644, 1592, 1454, 1424, 1316, 1186, 1106, 1082, 753,
695 cmꢀ1; HRMS-MALDI: m/z [M+H]+ calcd for C17H17N2O2S2
345.0726, found: 345.0724; Anal. calcd for C17H16N2O2S2: C 59.28, H
4.68, N 8.13, found: C 59.29, H 4.80, N 8.11.
+
:
(ꢁ)-Ethyl
(2Z)-2-[4-(Acetyloxy)-3,5-dibromobenzylidene]-5-(1-
Keywords: anti-infective agents · inhibitors · non-mevalonate
pathway · thiazolopyrimidines
benzofuran-2-yl)-7-methyl-3-oxo-2,3-dihydro-5H-[1,3]thiazolo-
[3,2-a]pyrimidine-6-carboxylate ((ꢁ)-21): General procedure 2
starting from (ꢁ)-14 (285 mg, 0.90 mmol). After filtration, com-
pound (ꢁ)-21 was obtained as a yellow solid (334 mg, 56%): Rf =
0.40 (hexane/EtOAc, 2:1); mp: 220–2228C; 1H NMR (300 MHz,
CDCl3): d=1.26 (t, J=7.1 Hz, 3H, CO2CH2CH3), 2.41 (s, 3H, COCH3),
2.54 (s, 3H, CH3), 4.19 (q, J=6.9 Hz, 2H, CO2CH2CH3), 6.49 (s, 1H,
CH), 6.74 (s, 1H, HAr), 7.18–7.30 (m, 2H, HAr), 7.43 (d, J=8.3 Hz, 1H,
HAr), 7.52 (d, J=7.7 Hz, 1H, HAr), 7.61 (s, 1H, CH), 7.66 ppm (s, 2H,
C6H2Br2); 13C NMR (75 MHz, CDCl3): d=14.3, 20.6, 22.9, 49.2, 60.8,
105.7, 105.9, 111.6, 118.6 (2C), 121.3, 122.9, 123.2, 124.7, 129.2,
133.1 (2C), 133.2, 133.5, 147.2, 152.8, 153.9, 154.7, 155.4, 164.2,
[1] M. A. Missinou, S. Borrmann, A. Schindler, S. Issifou, A. A. Adegnika, P. B.
Matsiegui, R. Binder, B. Lell, J. Wiesner, T. Baranek, H. Jomaa, P. G. Krems-
ner, Lancet 2002, 360, 1941–1942.
[2] a) W. Eisenreich, F. Rohdich, A. Bacher, Trends Plant Sci. 2001, 6, 78–84;
b) W. Eisenreich, A. Bacher, D. Arigoni, F. Rohdich, Cell. Mol. Life Sci.
2004, 61, 1401–1426; c) W. N. Hunter, J. Biol. Chem. 2007, 282, 21573–
21577.
[3] L. E. Kemp, C. S. Bond, W. N. Hunter, Proc. Natl. Acad. Sci. USA 2002, 99,
6591–6596. PDB code: 1GX1.
[4] S. B. Richard, J. L. Ferrer, M. E. Bowman, A. M. Lillo, C. N. Tetzlaff, D. E.
Cane, J. P. Noel, J. Biol. Chem. 2002, 277, 8667–8672. PDB codes: 1KNK,
1KNJ.
~
166.7, 197.6 ppm; IR (ATR): n=3298, 2982, 1776, 1685, 1556, 1453,
1363, 1320, 1177, 1158, 1085, 759 cmꢀ1; HRMS-MALDI: m/z [M+H]+
calcd for C27H21Br2N2O6S+: 658.9482, found: 658.9481.
1100
www.chemmedchem.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1092 – 1101

Thiazolopyrimidine Inhibitors of IspF
[5] S. Steinbacher, J. Kaiser, J. Wungsintaweekul, S. Hecht, W. Eisenreich, S.
Gerhardt, A. Bacher, F. Rohdich, J. Mol. Biol. 2002, 316, 79–88. PDB
codes: 1JY8, 1U40, 1U43, 1U3P, 1U3L.
[6] L. Buetow, A. C. Brown, T. Parish, W. N. Hunter, BMC Struct. Biol. 2007, 7,
68. PDB code: 2UZH.
[19] F. Rohdich, W. Eisenreich, J. Wungsintaweekul, S. Hecht, C. A. Schuhr, A.
Bacher, Eur. J. Biochem. 2001, 268, 3190–3197.
[20] B. Tozkoparan, M. Yarim, S. SaraÅ, M. Ertan, P. Kelicen, G. Altinok, R.
Demirdamar, Arch. Pharm. 2000, 333, 415–420.
[21] C. Kappe, Tetrahedron 1993, 49, 6937–6963.
[7] C. Lehmann, K. Lim, J. Toedt, W. Krajewski, A. Howard, E. Eisenstein, O.
Herzberg, Proteins 2002, 49, 135–138. PDB code: 1JN1.
[8] J. Osipiuk, M. Gu, S. Peterson, W. F. Anderson, A. Joachimiak, unpublish-
ed results. PDB code: 3GHZ.
[9] H. Kishida, T. Wada, S. Unzai, T. Kuzuyama, M. Takagi, T. Terada, M. Shir-
ouzu, S. Yokoyama, J. R. H. Tame, S.-Y. Park, Acta Crystallogr. Sect. D
2002, 59, 23–31. PDB codes: 1IV1, 1IV2, 1IV3, 1IV4.
[22] F. S. Falsone, C. Kappe, ARKIVOC 2001, (ii), 122–134.
[23] B. Desai, D. Dallinger, C. Kappe, Tetrahedron 2006, 62, 4651–4664.
[24] N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev. 2007, 36, 914–931.
[25] S. Herz, J. Wungsintaweekul, C. A. Schuhr, S. Hecht, H. Lꢆttgen, S.
Sagner, M. Fellermeier, W. Eisenreich, M. H. Zenk, A. Bacher, F. Rohdich,
Proc. Natl. Acad. Sci. USA 2000, 97, 2486–2490.
[26] G. Taubes, Science 2008, 321, 356–361.
[10] S. Ni, H. Robinson, G. C. Marsing, D. E. Bussiere, M. A. Kennedy, Acta
Crystallogr. Sect. D 2004, 60, 1949–1957. PDB code: 1TOA.
[11] T. E. Edwards, D. R. Davies, R. Hartley, W. Zeller, unpublished results. PDB
code: 3KE1.
[12] Y. Kim, N. Maltseva, J. Stam, W. F. Anderson, A. Joachimiak, unpublished
results. PDB codes: 3F6M, 3FPI.
[13] A. K. Wernimont, J. Lew, I. Kozieradzki, Y. H. Lin, X. Sun, C. Khuu, Y. Zhao,
M. Schapira, C. H. Arrowsmith, A. M. Edwards, J. Weigelt, A. Bochkarev,
R. Hui, J. D. Artz, T. Xiao, unpublished results. PDB code: 3B6N.
[14] B. M. Calisto, J. Perez-Gil, M. Bergua, J. Querol-Audi, I. Fita, S. Imperial,
Protein Sci. 2007, 16, 2082–2088. PDB code: 2PMP.
[27] WHO: Malaria, fact sheet No. 94, World Health Organization, 2009.
[28] WHO: Tuberculosis, fact sheet No. 104, World Health Organization,
2007.
[29] D. J. Payne, M. N. Gwynn, D. J. Holmes, D. L. Pompliano, Nat. Rev. Drug
Discovery 2007, 6, 29–40.
[30] C. E. Barry, N. Engl. J. Med. 2009, 360, 2466–2467.
[31] D. E. Cane, Science 2000, 287, 818–819.
[32] A. R. Odom, W. C. Van Voorhis, Mol. Biochem. Parasitol. 2010, 170, 108–
111.
[33] W. O. Bullock, J. M. Fernandez, J. M. Short, BioTechniques 1987, 5, 376–
379.
[15] M. Gabrielsen, F. Rohdich, W. Eisenreich, T. Grꢀwert, S. Hecht, A. Bacher,
W. N. Hunter, Eur. J. Biochem. 2004, 271, 3028–3035. PDB codes: 1W55,
1W57.
[16] C. M. Crane, J. Kaiser, N. L. Ramsden, S. Lauw, F. Rohdich, W. Eisenreich,
W. N. Hunter, A. Bacher, F. Diederich, Angew. Chem. 2006, 118, 1082–
1087; Angew. Chem. Int. Ed. 2006, 45, 1069–1074. PDB codes: 2AMT,
2GZL.
[34] P. Kuzmic, Anal. Biochem. 1996, 237, 260–273.
[35] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giacovazzo, A.
Guagliardi, A. G. G. Moliterni, R. Spagna, J. Appl. Crystallogr. 1999, 32,
115 – 119.
[36] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[17] C. Baumgartner, C. Eberle, F. Diederich, S. Lauw, F. Rohdich, W. Eisen-
reich, A. Bacher, Helv. Chim. Acta 2007, 90, 1043–1068.
[18] V. Illarionova, J. Kaiser, E. Ostrozhenkova, A. Bacher, M. Fischer, W. Eisen-
reich, F. Rohdich, J. Org. Chem. 2006, 71, 8824–8834.
Received: February 25, 2010
Revised: April 7, 2010
Published online on May 17, 2010
ChemMedChem 2010, 5, 1092 – 1101
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1101