Synthesis and characterization of potent inhibitors of Trypanosoma cruzi dihydrofolate reductase

Article abstract of DOI:10.1016/j.bmc.2010.04.020

Dihydrofolate reductase (DHFR) of the parasite Trypanosoma cruzi (T. cruzi) is a potential target for developing drugs to treat Chagas' disease. We have undertaken a detailed structure-activity study of this enzyme. We report here synthesis and characterization of six potent inhibitors of the parasitic enzyme. Inhibitory activity of each compound was determined against T. cruzi and human DHFR. One of these compounds, ethyl 4-(5-[(2,4-diamino-6-quinazolinyl)methyl]amino-2-methoxyphenoxy)butanoate (6b) was co-crystallized with the bifunctional dihydrofolate reductase-thymidylate synthase enzyme of T. cruzi and the crystal structure of the ternary enzyme:cofactor:inhibitor complex was determined. Molecular docking was used to analyze the potential interactions of all inhibitors with T. cruzi DHFR and human DHFR. Inhibitory activities of these compounds are discussed in the light of enzyme-ligand interactions. Binding affinities of each inhibitor for the respective enzymes were calculated based on the experimental or docked binding mode. An estimated 60-70% of the total binding energy is contributed by the 2,4-diaminoquinazoline scaffold.

Full text of DOI:10.1016/j.bmc.2010.04.020

Bioorganic & Medicinal Chemistry 18 (2010) 4056–4066
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and characterization of potent inhibitors of Trypanosoma cruzi
dihydrofolate reductase
Norbert Schormann a, Sadanandan E. Velu b, Srinivasan Murugesan b, Olga Senkovich a, Kiera Walker a,
Bala C. Chenna b, Bidhan Shinkre b, Amar Desai a, Debasish Chattopadhyay a,
*
a Department of Medicine and Center for Biophysical Sciences and Engineering, University of Alabama at Birmingham, Birmingham, AL 35294, USA
b Department of Chemistry, University of Alabama at Birmingham, Birmingham, AL 35294, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Dihydrofolate reductase (DHFR) of the parasite Trypanosoma cruzi (T. cruzi) is a potential target for devel-
oping drugs to treat Chagas’ disease. We have undertaken a detailed structure–activity study of this
enzyme. We report here synthesis and characterization of six potent inhibitors of the parasitic enzyme.
Inhibitory activity of each compound was determined against T. cruzi and human DHFR. One of these
compounds, ethyl 4-(5-[(2,4-diamino-6-quinazolinyl)methyl]amino-2-methoxyphenoxy)butanoate (6b)
was co-crystallized with the bifunctional dihydrofolate reductase-thymidylate synthase enzyme of T. cru-
zi and the crystal structure of the ternary enzyme:cofactor:inhibitor complex was determined. Molecular
docking was used to analyze the potential interactions of all inhibitors with T. cruzi DHFR and human
DHFR. Inhibitory activities of these compounds are discussed in the light of enzyme–ligand interactions.
Binding affinities of each inhibitor for the respective enzymes were calculated based on the experimental
or docked binding mode. An estimated 60–70% of the total binding energy is contributed by the 2,4-dia-
minoquinazoline scaffold.
Received 20 February 2010
Revised 5 April 2010
Accepted 6 April 2010
Available online 9 April 2010
Keywords:
Parasitic disease
Chagas disease
DHFR
Antifolate
Drug development
Crystal structure
Trypanosoma cruzi
Protozoan
Ó 2010 Elsevier Ltd. All rights reserved.
Kinetoplast
1. Introduction
tional drug design. DHFR has a proven track record as a drug target
in cancer chemotherapy. More importantly, DHFR inhibitors are
Chagas’ disease is caused by the protozoan parasite Trypano-
soma cruzi, which is primarily transmitted by an insect vector
but also via blood transfusion, organ transplant, and from mother
to child. The disease affects millions of people in Latin America
where the vector and the parasite are endemic.1 Nevertheless,
since the infected hosts, if untreated can carry the parasite through
their life increased travel and immigration give rise to an emerging
threat in many countries outside the endemic regions. But the
drugs used for the treatment of T. cruzi infection cause serious side
effects and are not effective in the chronic stage of the disease.
Therefore, validating potential drug targets and identifying novel
drug candidates is of considerable global importance. Since the
dihydrofolate reductase (DHFR) activity of T. cruzi (TcDHFR) is
essential for the parasite, it represents a potential target for ra-
successfully used in the treatment of bacterial and parasitic
infections.2–5 While DHFR is a monofunctional protein in mam-
mals, T. cruzi and other protozoan parasites carry a bifunctional
form of the enzyme in which the DHFR domain is linked to the thy-
midylate synthase (TS) domain with a linker sequence whose
length varies from one parasite to the other. Earlier data from
our laboratory showed that an inhibitor of TcDHFR, namely the
antifolate drug trimetrexate (TMQ), inhibits the growth of amasti-
gote and trypomastigote forms of T. cruzi in vitro.6 However, very
few, if any, high affinity inhibitors of TcDHFR have been reported.
In order to facilitate rational design of a selective inhibitor of the
T. cruzi enzyme we have initiated a comprehensive structure–
activity study using various antifolate molecules through selective
library search and limited chemical synthesis. As part of this study
we previously characterized the structure of the biologically rele-
vant bifunctional form of the parasitic enzyme, (TcDHFR–TS), and
determined the structure of the enzyme with TMQ bound to the
DHFR active site.7 We also identified a number of highly active
inhibitors of TcDHFR from libraries of antifolate compounds and
developed a three-dimensional model for quantitative structure–
activity (3D-QSAR) analysis of inhibitors of TcDHFR activity.8
Abbreviations: TMQ, trimetrexate (5-methyl-6-{3,4,5-trimethoxyphenyl)
amino]- methyl} quinazoline-2,4-diamine); CO4, 5-methyl-6-{[methyl(3,4,5-trime-
thoxyphenyl)amino]methyl} pyrido[2,3-d] pyrimidine-2,4-diamine; DHF, dihydro-
folate; NADP/NADPH, oxidized and reduced nicotinamide adenine dinucleotide
phosphate; dUMP, deoxyuridine monophosphate; EDO, ethylene glycol.
* Corresponding author. Tel.: +1 205 934 0124; fax: +1 205 934 0480.
E-mail address: debasish@uab.edu (D. Chattopadhyay).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.04.020

N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
4057
Based on earlier results of Davoll et al.9 and our data we have
recently synthesized six compounds that contain the 2,4-diamino-
quinazoline scaffold similar to TMQ but the methyl group at the
C5-position in the quinazoline ring is absent (Fig. 1). Moreover,
structural modifications were made to the phenyl ring and the lin-
ker nitrogen atom of TMQ. We report here the synthesis and char-
acterization of these compounds and their inhibitory activity
against TcDHFR and human DHFR. In order to determine the bind-
ing mode of the synthesized compounds in the enzyme active site,
we co-crystallized one inhibitor, ethyl 4-(5-[(2,4-diamino-6-qui-
nazolinyl)methyl]amino-2-methoxyphenoxy)butanoate (6b) with
the T. cruzi bifunctional enzyme and NADPH and determined the
structure of the ternary complex. We used molecular docking to
predict the binding of all inhibitors in T. cruzi and human DHFR
(hDHFR). The binding poses were used in the calculation of binding
affinities of these inhibitors for both enzymes. Experimentally
determined enzyme inhibition data are explained in the light of
these calculations. Throughout the work, we used a recombinant
preparation of bifunctional TcDHFR–TS enzyme for enzyme inhibi-
tion assays and structural studies.
of this enzyme for developing drugs against Chagas’ disease and
underscore the need for further investigation.11–13
TMQ is an excellent inhibitor of TcDHFR and is highly effective
in inhibiting growth of T. cruzi in vitro in tissue culture. However,
TMQ, despite being an FDA approved drug (for treatment of Pneu-
mocystis carinii infection in AIDS patients), suffers from serious lim-
itations. TMQ is also a potent inhibitor of hDHFR and therefore
requires co-administration with leukovorin in order to rescue the
host cells from the toxic effect of the drug. However, the potent
activity of TMQ against the parasite combined with previous stud-
ies indicating antiparasitic activity of 2,4-diaminoquinazoline
antifolates emphasizes the importance of further structure–activ-
ity studies on this class of compounds. We have now synthesized
six compounds that retain the 2,4-diaminoquinazoline scaffold
but contain modifications in the phenyl ring and the N-linker
(Fig. 1) to examine if new contacts with the parasitic enzyme, espe-
cially near the active site residue Met49, can be achieved. Since the
corresponding residue in hDHFR is Phe31, we hoped that the mod-
ifications may have adverse effect on the binding of these com-
pounds in hDHFR.
2.1. Enzyme inhibition
2. Results and discussion
All six compounds reported here are potent inhibitors of
TcDHFR with IC50 values in the nanomolar range. Inhibitory activ-
ities against TcDHFR and hDHFR are summarized in Table 1. We
determined experimental Ki values for 6a and 6b using Linewe-
aver–Burk plots (Fig. 2). These values agree well with the Ki values
calculated from the average IC50 values. However, all compounds
show excellent inhibition of hDHFR, although the IC50 values are
three to sixfold higher. As a consequence, the selectivity indices
of these inhibitors are comparable to that of TMQ (SI = 4).
All organisms require folate. Reduced forms of folate and its
derivatives are necessary for synthesis of essential biomolecules
such as protein, DNA, and RNA. As a result, metabolic steps leading
to the biosynthesis and modification of folates have been exten-
sively studied and enzymes involved in these pathways have been
targets of many clinically used drugs. Among these enzymes,
DHFR, in addition to being one of the most common targets for
cancer chemotherapy, has been exceptionally successful as a target
for antibacterial and antiparasitic agents. This is remarkable con-
sidering that DHFR is an evolutionarily conserved enzyme and
many residues involved in the reaction mechanism and substrate
binding are also highly conserved among various organisms. Since
DHFR is an essential enzyme in humans, any cross-reactivity with
hDHFR renders the drugs toxic to the host. Nevertheless, certain
drugs show high degree of selectivity towards DHFR of different
pathogens as compared to the human enzyme. Therefore, it is con-
ceivable that a therapeutically useful inhibitor of TcDHFR may be
developed. However, as Gilbert10 noted in his review on inhibitors
of leishmanial and trypanosomal DHFRs, reports of such inhibitors
have been limited. Nevertheless, results of limited medicinal
chemistry as well as in silico docking efforts reiterate the promise
2.2. Crystal structure
In an attempt to analyze the effect of modifications on the bind-
ing of the inhibitors within the enzyme active site, we determined
the structure of the ternary enzyme:cofactor:inhibitor complex
using one of the inhibitors, namely 6b. The bifunctional TcDHFR–
TS enzyme was used in crystallization. The structure of this com-
plex was refined to a resolution of 2.5 Å (R of 20.0%, Rfree of
23.4%; see Table 2). The asymmetric unit contains four TcDHFR–
TS subunits (A, B, C and D) with one cofactor (NADP) and the inhib-
itor 6b bound to each DHFR domain and one sulfate ion at the
phosphate binding site in each TS domain. In addition to the
OCH3
OCH3
NH2
OCH3
NH2 CH3
N
N
OR1
H
N
N
OCH3
H
H2N
N
6a-d
H2N
N
TMQ
OCH3
NH2
6a, R1 = CH3
6b, R1 =
N
N
O(CH2)4COOMe
O
H2N
N
O
O
8
CH3
6c, R1 =
OCH3
O
NH2
O
O
N
N
CH2
O(CH2)3COOH
6d, R1 =
H
H2N
N
9
Figure 1. Target structures of the synthesized analogs.

4058
N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
Table 1a
Inhibitory activity against TcDHFR and hDHFR
TcDHFR
hDHFR
SI (IC50
)
SI (Ki)
IC50 (nM)
Ki (nM)
IC50 (nM)
Ki (nM)
6a
6b
6c
6d
8
27.1 4.7
23.8 7.0
27.0 4.0
57.5 4.6
422 64
1.6 0.4
1.3 0.2
1.6 0.2
3.3 0.6
24.9 6.4
3.6 0.8
6.6 0.6
85.5 11.2
68.3 5.6
11.8 0.8
9.6 1.5
13.0 1.6
24.5 4.3
188 32
49.8 7.0
3.2
2.9
3.5
3.0
3.2
5.8
4.0
7.4
7.4
8.1
7.4
7.6
93.5 10.8
174.6 15.2
1338 167
356.6 28.1
80.9 12.5
9
61.0 6.6
20.2 6.6
13.8
TMQa
IC50 = ci/(vo/vi À 1) and Ki = IC50/(S/Km + 1)b
Ki calculation uses the Michaelis–Menten equation for competitive inhibition: for S » Km, Ki « IC50
SI (selectivity index) defined as IC50 (hDHFR)/IC50 (TcDHFR) or Ki (hDHFR)/Ki (TcDHFR)
a
Experimental activity data for TMQ are taken from Ref. 6.
Km (DHF) values used for calculation of Ki are 1.2 and 2.8 lM for TcDHFR and hDHFR, and were taken from Refs. 28,29, respectively.
b
Table 1b
Experimental Ki for compounds 6a and 6b.
6a: Ki = 4.75 0.98 nM (Km = 3.25 0.82 nM; R2 = 0.990)
6b: Ki = 5.68 1.13 nM (Km = 1.45 0.32 nM; R2 = 0.998)
C-terminal 6 or 7 residues several residues in a loop region (residue
number 110–121) are missing in each subunit. The final model con-
tains 99.8% of all protein residues in the allowed regions of the
Ramachandran plot. Quality of electron density for placing the inhib-
itor was excellent except for the ethylbutanoate chain (Fig. 3). Sev-
eral atoms of this chain were disordered in all four subunits.
Electron density for the NADP molecule was also of excellent quality.
Overall structure of the protein in this complex is very similar to
structures described earlier for inhibitor-free TcDHFR–TS and the
TMQ complex. No major change in the protein structure is noticed
in the DHFR domain upon binding of the inhibitor except some
movement of the Met49 side chain. The inhibitor is bound in the
hydrophobic pocket with the pyrimidine ring of the 2,4-diamino-
quinazoline scaffold stacked against residue Phe52. The cofactor
molecule is involved in extensive interactions with the enzyme
but provides only weak stacking interaction with the quinazoline
ring of the inhibitor. Ternary (enzyme:cofactor:inhibitor) com-
plexes of DHFRs with both the oxidized and reduced forms of the
cofactor have been reported suggesting that the oxidation state
of the cofactor may not determine inhibitor binding.7,14 Carboxyl
oxygen atoms of the side chain of Asp48 and the main chain O
atom of Val26 provide major hydrophilic interactions for the inhib-
itor. As predicted, the ethylbutanoate chain packs against residue
Met49. Also, the methoxy group of the inhibitor makes a hydro-
phobic interaction with the methyl group of Met49 side chain
(Fig. 4).
There was one unexplained density in a pocket near the inter-
face of the DHFR and TS domains in each subunit. In the crystal
structures of TcDHFR–TS reported earlier, sulfate ions or ethylene
glycol molecules were placed in this region. However, attempts
to refine sulfate ions or ethylene glycol in the present structure
were unsuccessful and resulted in ambiguous density and unrea-
sonable temperature factors. Moreover, Cys403 in the TS domain
seems to have been modified at least in the A subunit, although
the nature of the modification could not be resolved at this time.
2.3. Docking and binding affinity
We used molecular docking for further analysis of the results.
The experimental binding modes of TMQ (PDBID: 3HBB) and 6b
(PDBID: 3KJS) provide a framework for predicting binding modes
Figure 2. Lineweaver–Burk plots. (A) Compound 6a. (B) Compound 6b.

N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
4059
Table 2
Crystal data, data collection, and refinement statistics for TcDHFR–TS:6b complex
Crystal data
Space group
P43212
Unit cell parameters (Å)
Resolution range (Å)
Redundancy
a = 175.93, b = 175.93, c = 250.35
19.99–2.50 (2.59–2.50)*
4.0 (3.6)
Completeness (%)
Rmerge (%)
99.6 (98.9)
6.0 (25.6)
I/
r
13.0 (4.0)
Refinement statistics
Resolution (Å)
No. of reflections
Completeness (%)
Rall (%)
19.99–2.50 (2.56–2.50)
126968 (9574)
99.6 (99.7)
20.2
Rwork (%)
20.0 (27.0)
Rfree (%)
23.4 (30.9)
No. of atoms
Total
Protein
17053
15908
316
Ligand
Ethylene glycol
Ion
52
40
737
35.3
Figure 4. Location of inhibitor 6b in the active site of TcDHFR. Inhibitor molecule 6b
is shown as stick model (color code: yellow-C, blue-N, red-C); protein residues are
in surface representation with selected labeled residues (Val26, Asp48, and Met49)
as stick models (color code: white-C, blue-N, and red-O). Hydrophobic interactions
between the inhibitor molecule and side chain atoms of Met49 are also shown.
Water
Wilson B-factor (Å2)
Average B-factors (Å2)
Overall
28.7
28.2
44.8
56.1
81.4
28.3
Protein
Ligand (inhibitor + NADPH)
Ethylene glycol
Ion
Water
R.m.s. deviations
Bonds (Å)
and amino groups in the 2,4-diaminoquinazoline ring and the
backbone O atoms of residues Val26, Val27, and Ile154 as well as
the side chain O atoms of Asp48 in the TcDHFR active site. A similar
hydrogen bonding pattern is observed for the remaining five com-
pounds with the differences being attributed to the degree of
superimposition of the 2,4-diaminoquinazoline scaffolds.
0.007
1.18
0.15
Angles (°)
Coordinate error ESU
Correlation coefficient
FoFc
Docked poses of the inhibitors in the TcDHFR binding pocket
(Fig. 5A) reveal that the introduced modifications do not involve
additional interactions with the protein except enhanced hydro-
phobic interaction with Met49 for some inhibitors. Rather most
inhibitors extend into an open area of the protein structure. Inher-
ent flexibility in the structure of DHFRs has been described partic-
ularly in studies on Escherichia coli DHFR.15,16 These studies
showed that during transition between different states of the en-
zyme the largest conformational change occurs in the Met20 loop
of DHFR (corresponds to Ile41 loop in T. cruzi DHFR) and is accom-
panied by the motion of the cofactor into and out of the binding
pocket. Conformational changes involving ligand (substrate or
inhibitor) and cofactor binding play important roles in determining
the functional outcome for enzyme catalysis and inhibitor activ-
ity.17 In our studies with various inhibitors of TcDHFR, we noticed
movement in the protein structure.8 Flexibility in the target active
site makes it possible for even large substituent groups to pack in
the open space as shown in the close up views of 6b and 6d in the
pocket (Fig. 5B and C). In the crystal structure of the enzyme-inhib-
itor complex, part of the ethylbutanoate ester chain of 6b shows no
interaction with the protein or solvent and remains disordered
(Fig. 3). On the other hand, this structure suggests that the Phe88
residue in TcDHFR may be targeted for selective hydrophobic inter-
action since the amino acid residue in the equivalent position in
hDHFR is Asn64 (Fig. 6).
0.95
0.93
FoFc free
Ramachandran plot
Allowed (%)
Disallowed (%)
99.8
0.2 [1 residue per subunit]
*
Numbers in parentheses represent highest resolution shell.
The binding (docking) modes of all compounds in the hDHFR
active site are shown in Figure 7A. The docked pose of CO4 for
which a co-crystal structure with hDHFR is available was used as
a guide to compare the reliability of the docked conformations.18
The 2,4-diaminoquinazoline moiety of the inhibitors packs tightly
in the hDHFR active site providing major binding interactions. The
quinazoline ring stacks against the phenyl ring of residue Phe32. As
in TcDHFR active site, the inhibitors extend to the open solvent ex-
posed region of the molecule. If the inhibitors would have adopted
the same conformation in the active sites of both TcDHFR and
hDHFR, several inhibitors would be expected to clash in hDHFR
Figure 3. Electron density map (Fo À Fc, contoured at 3.0
r) for compound 6b.
of the remaining compounds using molecular docking. Previously,
we showed that the docking poses quite accurately capture the
binding mode in the crystal structure with r.m.s. deviations of
1.17 and 1.53 Å between the X-ray pose and docking pose for
TMQ and 6b, respectively.8 The calculated total estimated li-
gand–receptor interaction energy and the hydrogen bonding pat-
tern are essentially the same for both poses of each molecule
(see Table 3). Hydrogen bonds are observed between ring N-atoms

4060
N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
Table 3
Estimated binding affinities based on docked poses of inhibitors within TcDHFR and hDHFR active site
G (kcal molÀ1
)
MW
Nheavy
LEb
Kd (nM)
Àlog Kd
a
D
TcDHFR
6a
6b
6c
6d
8
9
TMQ
À12.07
325
425
397
459
468
425
369
24
31
29
34
34
31
27
0.50
0.38 (0.37)
0.43
0.36
0.25
1.8
2.4 (4.2)
1.0
1.5
889
8.74
8.62 (8.38)
9.00
8.84
6.05
À11.92 (À11.58)c
À12.43
À12.21
À8.36
À9.13
0.29
0.37 (0.36)
246
50 (92)
6.61
7.30 (7.04)
À10.09 (À9.72)
hDHFR
6a
6b
6c
6d
8
9
CO4
À11.32
325
425
397
459
468
425
384
24
31
29
34
34
31
28
0.47
0.43
0.43
0.34
0.38
0.43
0.48 (0.42)
6.4
0.2
0.8
4.1
0.4
0.2
0.2 (3.3)
8.19
9.65
9.08
8.39
9.35
9.76
9.79 (8.48)
À13.33
À12.55
À11.59
À12.92
À13.48
À13.52 (À11.71)
TcDHFR
hDHFR
D
G [kcal molÀ1
]
LEd
Ki [nM]
D
G [kcal molÀ1
]
LEd
Ki [nM]
6a
6b
6c
6d
8
À12.15
À12.28
À12.15
À11.72
À10.51
À11.67
À11.30
0.51
0.40
0.42
0.34
0.31
0.38
0.42
1.6
1.3
1.6
3.3
24.9
3.6
À10.95
À11.08
À10.89
À10.51
À9.29
0.46
0.36
0.38
0.31
0.27
0.33
11.8
9.6
13.0
24.5
188
49.8
9
TMQ
À10.09
6.6
Based on results from PEARLS web server (http://ang.cz3.nus.edu.sg/cgi-bin/prog/rune.pl).26
D
G = ÀRT ln Kd (RT ꢀ 0.6 kcal molÀ1 for T = 298 K).
a
Nheavy = Non-hydrogen atoms.
b
LE =
D
G/Nheavy (ligand efficiency) [based on affinity in docked poses; Kd].
c
Values based on X-ray crystal structures are shown within parenthesis.
d
LE =
DG/Nheavy (ligand efficiency) [based on activity; Ki].
Figure 5. Docking of the inhibitors into the active site of TcDHFR. (A) Docked poses of inhibitors in TcDHFR active site. Protein structure is shown as a cartoon diagram,
ligands, and NADP are shown as stick models with selected residues (Met49, Phe52, Phe88) shown in line representation. (B) Docked pose of compound 6b is shown in yellow
(C), red (O), and blue (N). For comparison TMQ is also shown as stick model (color code: cyan-C, red-O, blue-N). Residues Met49, Phe52, and Phe88 are shown in line
representation and are labeled. (C) Docked pose of compound 6d shown in yellow (C), red (O), and blue (N). For comparison TMQ is also shown as stick model (color code:
cyan-C, red-O, blue-N). Protein residues Met49 and Phe52 shown in line representation are labeled.
in particular with Leu22 (Fig. 7B). Instead, both the inhibitor and
protein residues undergo conformational changes. In addition,
the side chain amino group of Gln35 in hDHFR is positioned within
hydrogen bonding distance from acceptor oxygen atoms of several
inhibitors, thereby enhancing the affinity of these inhibitors for
hDHFR and resulting in a decrease in selectivity (Fig. 7C).
zyme’s active site (Table 3). For comparison, we also estimated
binding affinities of TMQ and 6b for TcDHFR based on the crystal
structures of these inhibitors with TcDHFR–TS and the binding
affinity of CO4 based on its crystal structure with hDHFR. As shown
in Table 4, energy analysis of binding for these lipophilic inhibitors
demonstrates that the van der Waals energy portion provides the
main contribution to the total energy of binding. The contribution
of hydrogen bonding energy to the total energy varies between 5%
We calculated the binding affinity of each inhibitor for TcDHFR
and hDHFR based on their docking poses in the respective en-

N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
4061
without significant clashes. Therefore, modifications outside the
2,4-diaminoquinazoline scaffold were not effective in enhancing
selectivity for the parasitic enzyme over the human host enzyme.
The results presented here underscore the need for identifying
high affinity scaffolds that are able to discriminate between the
target in host and pathogen. At the same time, this study offers in-
sight about the enzyme active site for further exploration of the
chemical space. Although the docking technique we used quite
reliably predicts the binding conformation of ligands, our results
point out the need to account for ligand induced conformational
changes in the DHFR structure for more accurate prediction of pro-
tein–ligand interactions.
Despite the high degree of overall homology in primary se-
quences and in three-dimensional structures of DHFR from differ-
ent organisms, DHFR remains a potentially attractive target for
drugs to treat microbial infection. For example, trimethoprim
exhibits a 105-fold selectivity for bacterial DHFR as compared to
vertebrate enzymes. Probing the active sites of both the host and
pathogen enzyme with diverse scaffolds will lead to the identifica-
tion of hot spots for selectivity.
4. Experimental
4.1. General methods for synthesis
Solvent evaporations were carried out in vacuo with a rotary
evaporator. Thin layer chromatography (TLC) was performed on
silica gel plates with fluorescent indicator (Whatmann, silica gel,
Figure 6. Superimposition of active sites in TcDHFR and hDHFR. Crystal structure of
TcDHFR domain (residues 20–232) in the ternary complex of TcDHFRTS:NADP:6b
was superimposed on the crystal structure of hDHFR in the CO4 complex. Protein
residues are shown in line representation. Color code for TcDHFR: green (C), blue
(N), red (O); color code for hDHFR: rose (C), blue (N), red (O). Inhibitor 6b is shown
as stick model with the same color code as TcDHFR. Location of one of the dissimilar
residues near the active site (Phe88 in TcDHFR and Asn64 in hDHFR) is labeled.
UV254, 25
and 365 nm). Purification by column and flash chromatography
was carried out using ‘BAKER’ silica gel (40 m) in the solvent sys-
lm plates). Spots were visualized by UV light (254
l
tems indicated. The amount (weight) of silica gel for column chro-
matography was in the range of 50–100 times the amount (weight)
of the crude compounds being separated. Proton nuclear magnetic
and 20% among these compounds. The 2,4-diaminoquinazoline
scaffold provides 60–70% of the total ligand-receptor binding en-
ergy (ꢀ70% of the total van der Waals energy and 80–100% of
the total hydrogen bond energy for each inhibitor). We also calcu-
lated the ligand efficiency (LE) of the six inhibitors for TcDHFR and
hDHFR based on experimental activities and estimated binding
affinities. Ligand efficiency is defined as the free energy contribu-
resonance (1H NMR) and carbon nuclear magnetic resonance (13
C
NMR) spectra were recorded on a Bruker DPX-300 spectrometer
using TMS as internal standard. The values of chemical shifts (d)
are given in ppm and coupling constants (J) in hertz. The chemical
shift values are reported as parts per million (ppm) relative to tet-
ramethylsilane as internal standard. Mass spectra were recorded
on Micromass Platform LCC instrument. Anhydrous solvents used
for reactions were purchased in Sure-Seal™ bottles from Aldrich
Chemical Company. Other reagents were purchased from Aldrich,
Lancaster or Fisher chemical companies and used as received.
tion per ‘heavy atom’. The formula ‘
D
G = ÀRTlnK’ correlates the
free energy of binding with either the inhibitory activity (Ki) or
the binding affinity (Kd). Although activities and binding affinities
for enzyme inhibitors are defined differently, we observe a very
good correlation between LE values based on experimental inhibi-
tory activities and estimated affinities of inhibitors 6a–d and TMQ
for TcDHFR, while the correlation for inhibitors 8 and 9 is not good
(values of Ki and Kd differ 36–68-fold, see Table 3). In the case of
hDHFR we only observe a good correlation for inhibitors 6a and
6c (Table 3).
4.2. Syntheses
The chemical structures of all target compounds including TMQ
for comparison are shown in Figure 1.
Synthesis of target compounds 6a–d is outlined in Scheme 1.
The synthesis started with commercially available 2-fluoro-5-
formylbenzonitrile (1). The 2-fluoro-5-formylbenzonitrile 1 was
aminated with 5-amino-2-methoxyphenol 2 in presence of NaC-
NBH3 and ZnCl2 in methanolic medium to afford the aminated
product 3 in 77% yield. The alkylation of hydroxy compound 3 with
alkyl halides in the presence of NaH in anhydrous DMF gave the
compounds 4a–d in 81–90% yield. The target compounds, 6a–d,
were prepared from the compounds 4a–d in 44–83% yield by
refluxing with guanidine carbonate (5) in N,N-dimethylacetamide.
Synthesis of target compound 8 is outlined in Scheme 2. Reduc-
tive N-alkylation of compound 4c with propionaldehyde using
NaCNBH3 and few drops of conc. HCl afforded the N-propyl com-
pound 7 in 93% of yield. Compound 7 was then converted to the
target quinazoline 8 in 38% yield by refluxing with guanidine car-
bonate 5 in N,N-dimethylacetamide.
3. Conclusion
The synthesized molecules are potent inhibitors of T. cruzi DHFR
activity. However, these compounds also show high degree of
inhibitory activity against the human enzyme. Introduced chemi-
cal modifications were not effective in either providing sufficient
additional favorable contacts with the parasitic enzyme or unfa-
vorable contacts with the host protein. Some of the substituent
groups showed hydrophobic interactions with Met49 of TcDHFR.
However, these interactions did not provide a net gain in selectiv-
ity. The 2,4-diaminoquinazoline scaffold makes a large contribu-
tion towards the energy of binding of these molecules in the
active sites of both enzymes. Moreover, flexibility in the structure
of the DHFR active site allows binding of even very large inhibitors

4062
N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
Figure 7. Docking of the inhibitors into the active site of hDHFR. (A) Docked poses of inhibitors in hDHFR active site. Protein model is shown in cartoon representation.
Ligands and NADP are shown as stick models. Selected residues (Phe31, Phe 34, and Gln 35) are shown in line representation. (B) Stereoscopic view showing docked poses of
several inhibitors as stick models as placed in the TcDHFR active site superimposed onto the hDHFR structure (shown as cartoon). Residues Pro26, Phe31, and Pro61 in hDHFR
close to the inhibitors are shown as stick models (color code: pink-C, blue-N, red-O). (C) Binding of inhibitor 6b in the hDHFR active site. The protein is shown as a cartoon
diagram. Docked conformation of 6b is shown as stick model (color code: yellow-C, blue-N, red-O). For comparison, the compound CO4 is shown as stick model (color code:
green-C, blue-N, red-O). NADP is also shown as stick model (color code: white-C, blue-N, red-O). Amino acid residues Phe31, Phe34, and Gln35 shown in line representation
are labeled. Gln35 side chain N atom is within hydrogen bonding distance from the methoxy oxygen atom on the ligand.
The carboxylic acid derivative 9 is prepared by hydrolysis of
corresponding ethyl ester 6b using 1 N. NaOH in THF with 72%
yield (Scheme 3).
to obtain the crude product. The crude product was purified by
flash chromatography on silica gel using EtOAc–hexanes (1:3) as
the eluent to afford the pure 5-[(3-hydroxy-4-methoxyphenyl-
amino)methyl]-2-fluorobenzonitrile 3 (5.67 g, 77%); mp 113 °C;
1H NMR (CDCl3) d 3.81 (s, 3H), 3.95 (br s, 1H), 4.30 (s, 2H), 5.61
(br s, 1H), 6.04 (dd, 1H, J1 = 2.7 Hz, J2 = 8.6 Hz), 6.23 (d, 1H,
J = 2.7 Hz), 6.70 (d, 1H, J = 8.6 Hz), 7.18 (t, 1H, J = 8.5 Hz) and
7.52–7.65 (m, 2H); 13C NMR (CDCl3) d 47.5, 56.9, 101.0, 101.4
and 101.5 (C–F coupling), 104.1, 112.6, 114.2, 116.6 and 116.8
(C–F coupling), 131.9, 133.9 and 134.0 (C–F coupling), 137.2 and
137.3 (C–F coupling), 139.7, 142.5, 146.8, 161.0 and 163.6 (C–F
coupling); MS (ES+): m/z = 273 [M+H].
4.2.1. 5-[(3-Hydroxy-4-methoxyphenylamino)methyl]-2-
fluorobenzonitrile (3)
To
a solution of 2-fluoro-5-formylbenzonitrile 1 (4.05 g,
27.1 mmol) in anhydrous MeOH (85 mL) 5-amino-2-methoxyphe-
nol 2 (4.16 g, 29.9 mmol) was added and stirred at room tempera-
ture for 15 min. A solution of NaCNBH3 (1.88 g, 29.9 mmol) and
ZnCl2 (1.85 g, 13.6 mmol) in MeOH (60 mL) was added dropwise
to the reaction mixture and stirred at room temperature for 12 h.
After completing the reaction (as indicated by tlc), water (2 mL)
was added to quench the reaction mixture and solvent was com-
pletely removed under reduced pressure. The residue obtained
was dissolved in dichloromethane (60 mL) and washed with water
(3 Â 30 mL), brine (1 Â 30 mL) and dried over Na2SO4. The drying
agent was filtered off and the solvent was concentrated in vacuum
4.2.2. General procedures for the preparation of 4
To a solution of compound 3 (2.47 g, 9.1 mmol) in anhydrous
DMF (4 mL), NaH (0.261 g, 10.9 mmol) was added and stirred at
room temperature for 30 min. Ethyl 4-bromobutyrate (1.44 mL,
10 mmol) was added to the reaction mixture and stirred at room

N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
4063
Table 4
extracted with EtOAc (3 Â 75 mL). The combined organic layer
was washed with water (3 Â 75 mL) and brine (1 Â 75 mL). The ex-
tract was then dried over Na2SO4. After the removal of the drying
agent, solvent was completely removed to obtain the crude prod-
uct which was purified by flash chromatography on silica gel using
EtOAc–hexanes (1:3) as the eluent to afford pure products 4a–d.
Distribution of total energy of ligand–receptor interactions based on docked poses of
inhibitors in the TcDHFR or hDHFR active site
Total (kcal/mol)
van der Waals
(kcal/mol)
Hydrogen bond
(kcal/mol)
TcDHFR
6a
6b
6c
6d
8
9
TMQ
À12.07
À9.45
À2.46
À11.92 (À11.58)a
À12.43
À10.41 (À11.29)
À10.45
À2.29 (À1.76)
À2.42
4.2.2.1. 5-{[(3,4-Dimethoxyphenyl)amino]methyl}-2-fluoroben-
zonitrile (4a). Isolated as a yellow oil; (83%); 1H NMR (CDCl3) d
3.80 (s, 3H), 3.82 (s, 3H), 4.32 (s, 2H), 6.06 (dd, 1H, J1 = 2.7 Hz,
J2 = 8.5 Hz), 6.24 (d, 1H, J = 2.7 Hz), 6.72 (d, 1H, J = 8.5 Hz), 7.19 (t,
1H, J = 8.5 Hz), 7.60–7.66 (m, 2H); 13C NMR (CDCl3) d 48.0, 56.2,
57.0, 99.6, 101.7 and 101.8 (C–F coupling), 103.9, 113.4, 114.4,
116.9, 117.1, 132.3, 134.2, 134.3 (C–F coupling), 137.4, 137.5 (C–F
coupling), 142.4, 142.5 (C–F coupling), 150.4, 161.4, 163.9 (C–F
coupling); MS (ES+): m/z = 287 [M+H].
À12.21
À10.00
À2.22
À8.36
À8.91
À0.46
À9.13
À9.14
À0.79
À10.09 (À9.72)
À10.16 (À9.64)
À2.34 (À0.89)
hDHFR
6a
6b
6c
6d
8
9
CO4
À11.32
À10.09
À2.39
À13.33
À11.43
À2.61
À12.55
À11.03
À1.33
À11.59
À9.47
À1.05
À12.92
À11.07
À1.19
À13.48
À10.96
À2.26
4.2.2.2. Ethyl 4-[5-(3-cyano-4-fluorobenzylamino)-2-methoxy-
phenoxy]butanoate (4b). Isolated as a yellow oil; (81%); 1H
NMR (CDCl3) d 1.25 (t, 3H, J = 7.6 Hz), 2.06–2.15 (m, 2H), 2.51
(t, 2H, J = 7.2 Hz); 3.76 (s, 3H), 3.98 (t, 2H, J = 6.4 Hz), 4.13 (q,
2H, J = 7.6 Hz), 4.30 (s, 2H), 6.06 (dd, 1H, J1 = 2.8 Hz, J2 = 8.4 Hz),
6.26 (d, 1H, J = 2.8 Hz), 6.71 (d, 1H, J = 8.4 Hz), 7.17 (t, 1H,
J = 8.4 Hz), 7.56–7.67 (m, 2H); 13C NMR (CDCl3) d 14.3, 24.5,
30.8, 47.6, 57.0, 60.5, 67.9, 101.0, 101.4 and 101.5 (C–F coupling),
104.2, 114.1, 116.5 and 116.7 (C–F coupling), 132.0, 133.9 and
À13.52 (À11.71)
À10.05 (À10.94)
À2.88 (À1.12)
Based on results from PEARLS web server (http://ang.cz3.nus.edu.sg/cgi-bin/prog/
rune.pl).26
a
Values in parenthesis are based on X-ray crystal structures of enzyme:cofac-
tor:inhibitor complex.
temperature for 12 h. After completion of the reaction (as indicated
by tlc), the reaction mixture was diluted with water (60 mL) and
OCH3
NaH
O
OCH3
DMF
NaCNBH3
NC
F
NC
F
H
N
H
+
R1Br
ZnCl2
MeOH
OH
H2N
OH
2
1
3
OCH3
OR1
OCH3
OR1
5
NH
H2N C
NH2
N
H2CO3
NH2
NC
F
N
N
H
N
H
N,N-Dimethyl-
acetamide
H2N
6
4
6a, R1 = CH3
4a, R1 = CH3
O
O
6b, R1 =
4b, R1 =
O
O
O
O
6c, R1 =
4c, R1 =
O
O
O
O
O
CH2
6d, R1 =
CH2
4d, R1 =
O
Scheme 1.
OCH3
O(CH2)4COOMe
OCH3
O(CH2)4COOMe
CH3CH2CHO
NaCNBH3
NC
F
NC
F
N
N
H
HCl, CH3CN
7
CH3
OCH3
O(CH2)4COOMe
4c
5
NH
H2N C
NH2
H2CO3
NH2
N
N
N,N-Dimethyl-
acetamide
H2N
N
8
CH3
Scheme 2.

4064
N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
OCH3
O(CH2)3COOEt
OCH3
NH2
N
NH2
1N. NaOH
N
N
N
N
O(CH2)3COOH
H
H
THF
H2N
H2N
N
6b
9
Scheme 3.
134.0 (C–F coupling), 137.2 (2C, C–F coupling), 142.2, 142.5, 149.5,
4.2.3.3. Methyl 5-(5-[(2,4-diamino-6-quinazolinyl)methyl]amino-
2-methoxyphenoxy) pentanoate (6c). Isolated as a yellow oil;
(48%); 1H NMR (DMSO-d6) d 1.61–1.66 (m, 4H), 2.35 (t, 2H,
J = 6.1 Hz), 3.57 (s, 3H), 3.59 (s, 3H), 3.82 (t, 2H, J = 6.1 Hz), 4.20
(s, 2H), 5.75 (br s, 1H), 6.05 (dd, 1H, J1 = 2.7 Hz, J2 = 8.7 Hz), 6.31
(d, 1H, J = 2.7 Hz), 6.66 (d, 1H, J = 8.7 Hz), 6.91 (br s, 2H), 7.28 (d,
1H, J = 8.4 Hz), 7.63 (d, 1H, J = 8.4 Hz), 8.09 (br s, 2H), 8.15 (s,
1H); 13C NMR (DMSO-d6) d 22.3, 29.2, 34.6, 48.5, 57.5, 68.4,
101.0, 103.9, 110.6, 115.5, 123.4, 124.0, 133.1, 133.6, 141.3,
144.8, 150.0, 150.7, 160.8, 160.78, 163.3, 175.9; MS (ES+):
m/z = 426 [M+H].
161.0 and 163.5 (C–F coupling), 173.3; MS (ES+): m/z = 387 [M+H].
4.2.2.3. Methyl 4-[5-(3-cyano-4-fluorobenzylamino)-2-methoxy-
phenoxy]pentanoate (4c). Isolated as a yellow oil; (90%); 1H NMR
(CDCl3) d 1.81À1.86 (m, 4H), 2.41 (t, 2H, J = 6.8 Hz), 3.68 (s, 3H),
3.79 (s, 3H), 3.96 (t, 2H, J = 6.8 Hz), 4.32 (s, 2H), 6.08 (dd, 1H,
J1 = 2.7 Hz, J2 = 8.5 Hz), 6.24 (d, 1H, J = 2.7 Hz), 6.73 (d, 1H,
J = 8.5 Hz), 7.20 (t, 1H, J = 8.5 Hz), 7.60–7.66 (m, 2H); 13C NMR
(CDCl3) d 21.9, 29.0, 34.1, 48.0, 52.0, 57.4, 68.8, 101.2, 101.8,
102.0 (C–F coupling), 104.4, 114.4, 116.9, 117.1 (C–F coupling),
132.3, 134.2, 134.3 (C–F coupling), 137.5 (2C, C–F coupling),
142.5, 142.9, 150.0, 161.4, 163.9 (C–F coupling), 174.3; MS (ES+):
m/z = 387 [M+H].
4.2.3.4. Methyl 4-[(5-[(2,4-diamino-6-quinazolinyl)methyl]amino-
2-methoxyphenoxy) methyl]benzene carboxylate (6d). Isolated as
an oil; (54%); 1H NMR (DMSO-d6) d 3.63 (s, 3H), 3.85 (s, 3H), 4.17
(br s, 2H), 5.07 (s, 2H), 5.81 (br s, 1H), 6.12 (dd, 1H, J1 = 8.4 Hz,
J2 = 2.3 Hz), 6.34–6.44 (m, 1H), 6.62–6.86 (m, 3H), 7.24 (d, 1H,
J = 8.4 Hz), 7.42–7.68 (m, 3H), 7.81–8.04 (m, 4H), 8.09–8.24 (m,
1H); 13C NMR (DMSO-d6) d 47.9, 52.6, 57.3, 69.7, 101.2, 104.7,
110.0, 115.5, 121.6, 123.3, 127.8, 129.3, 129.7, 133.8, 134.2,
141.2, 143.5, 144.2, 146.5, 149.1, 158.6, 163.1, 166.5; MS (ES+)
m/z 460 (M+H).
4.2.2.4. Methyl 4-({5-[(3-cyano-4-fluorobenzyl)amino]-2-methoxy-
phenoxy}methyl)benzoate (4d). Isolated as a yellow oil; (86%); 1H
NMR (CDCl3) d 3.81 (s, 3H), 3.91 (s, 3H), 4.02 (br s, 1H), 4.21 (s, 2H),
5.12 (s, 2H), 6.06–6.17 (m, 2H), 6.76 (d, 1H, J = 8.4 Hz), 7.11 (t, 1H,
J = 8.7 Hz), 7.44 (d, 2H, J = 8.4 Hz), 7.46–7.57 (m, 2H), 7.96–8.04 (m,
2H); 13C NMR (CDCl3) d 47.4, 52.1, 57.0, 70.5, 101.3 and 101.5 (C–F
coupling), 101.7, 105.1, 114.0, 114.3, 116.4 and 116.6 (C–F cou-
pling), 126.6, 129.5, 129.8, 131.7, 133.6 and 133.7 (C–F coupling),
137.0 (2C, C–F coupling), 142.1, 142.5, 142.6 (2C, C–F coupling),
149.0, 160.4 and 163.9 (C–F coupling), 166.8; MS (ES+) m/z 421
(M+H).
4.2.4. Preparation of methyl 5-{5-[(3-cyano-4-fluorobenzyl)
(propyl)amino]-2-methoxyphenoxy}pentanoate (7)
To a stirred solution of methyl 4-[5-(3-cyano-4-fluorobenzyl-
amino)-2-methoxyphenoxy] pentanoate 4c (0.055 g, 0.142 mmol)
in CH3CN (10 mL) was added propionaldehyde (0.017 g,
0.28 mmol) followed by NaCNBH3 (0.027 g, 0.43 mmol). It re-
mained as a suspension. The pH of the suspension was adjusted
to 2–3 by drop wise addition of concentrated HCl. As the mixture
was acidified, the suspended starting material began to dissolve
into solution, followed by precipitation of the crude product. The
suspension containing the crude product was stirred at room tem-
perature for an additional 0.5 h. The crude product was filtered,
stirred with 2 N Na2CO3 and washed with water, brine and dried
with Na2SO4 and concentrated in vacuo. The residue was purified
by column chromatography using silica gel with hexane–EtOAc
(85:15) as eluent to obtain the pure product 7 as a yellow oil
(0.057 g, 93%); 1H NMR (CDCl3) d 0.94 (t, 3H, J = 7.3 Hz), 1.61–
1.67 (m, 2H), 1.79–1.83 (m, 4H), 2.40 (t, 2H, J = 6.7 Hz), 3.26 (t,
2H, J = 7.3 Hz), 3.68 (s, 3H), 3.79 (s, 3H), 3.95 (t, 2H, J = 6.7 Hz),
4.43 (s, 2H), 6.19 (dd, 1H, J1 = 2.5 Hz, J2 = 8.6 Hz), 6.30 (d, 1H,
J = 2.5 Hz), 6.76 (d, 1H, J = 8.8 Hz), 7.16 (t, 1H, J = 8.6 Hz), 7.47–
7.52 (m, 2H); 13C NMR (CDCl3) d 12.0, 20.9, 22.0, 29.1, 34.1, 52.0,
54.7, 55.0, 57.3, 69.0, 101.7, 101.9 (C–F coupling), 102.1, 106.3,
114.4, 114.5, 116.8, 117.0 (C–F coupling), 132.0, 133.8, 133.9 (C–
F coupling), 137.2, (2C, C–F coupling), 142.6, 143.7, 149.8, 161.2,
163.8 (C–F coupling), 174.3; MS (ES+): m/z = 429 [M+H].
4.2.3. General procedures for the preparation of 6a–d
To the solution of compound 4 (7 mmol) in anhydrous N,N-
dimethylacetamide (25 mL), guanidine carbonate (1.26 g, 7 mmol)
was added and the resulting mixture was heated at 140 °C for 5 h.
The solvent was completely removed under vacuum and the resi-
due obtained was purified by flash column chromatography over
silica gel using MeOH–NH3 saturated CHCl3 (1:9) to furnish the
pure products 6a–d.
4.2.3.1. 6-[(3,4-Dimethoxyanilino)methyl]-2,4-quinazolinedi-
amine (6a). Isolated as a yellow oil; (83%); 1H NMR (DMSO-d6)
d 3.59 (s, 3H, 3.65 (s, 3H), 4.16 (d, 1H, J = 5.5 Hz), 5.72 (br t, 1H,
J = 5.5 Hz), 6.01 (br s, 2H), 6.07 (dd, 1H, J1 = 2.5 Hz, J2 = 8.6 Hz),
6.34 (d, 1H, J = 2.5 Hz), 6.65 (d, 1H, J = 8.6 Hz), 7.16 (d, 1H,
J = 8.6 Hz), 7.30 (br s, 2H), 7.51 (dd, 1H, J1 = 1.8 Hz, J2 = 8.6 Hz);
13C NMR (DMSO-d6) d 47.7, 55.2, 56.6, 99.0, 102.8, 109.8, 114.3,
122.5, 123.9, 131.9, 132.5, 140.2, 143.9, 149.8, 160.3, 162.4; MS
(ES+): m/z = 326 [M+H].
4.2.3.2. Ethyl 4-(5-[(2,4-diamino-6-quinazolinyl)methyl]amino-
2-methoxyphenoxy) butanoate (6b). Isolated as a yellow solid;
(44%); mp 194 °C; 1H NMR (DMSO-d6) d 1.17 (t, 3H, J = 7.1 Hz),
1.86–1.97 (m, 2H), 2.43 (t, 2H, J = 7.3), 3.60 (s, 3H), 3.86 (t, 2H,
J = 6.3 Hz), 4.04 (q, 2H, J = 7.1 Hz), 4.15 (d, 2H, J = 5.1 Hz), 5.69 (t,
1H, J = 5.1 Hz), 6.04 (br s, 2H), 6.09 (dd, 1H, J1 = 2.4 Hz,
J2 = 8.6 Hz), 6.33 (d, 1H, J = 2.4 Hz), 6.68 (d, 1H, J = 8.6 Hz), 7.17
(d, 1H, J = 8.6 Hz), 7.30 (br s, 2H), 7.50 (dd, 1H, J1 = 1.6 Hz,
J2 = 8.6 Hz), 7.99 (s, 1H); 13C NMR (DMSO-d6) d 14.1, 24.4, 30.1,
47.7, 56.7, 59.9, 67.0, 100.2, 103.3, 109.9, 114.7, 122.4, 124.1,
131.8, 132.5, 140.5, 144.0, 149.0, 151.3, 160.4, 162.4, 172.6; MS
(ES+): m/z = 426 [M+H].
4.2.5. Preparation of methyl 5-{5-[[(2,4-diamino-6-quinazolinyl)
methyl](propyl)amino]-2-methoxyphenoxy} pentanoate (8)
Compond 8 was synthesized from 7 (0.1 g, 0.23 mmol) and gua-
nidine carbonate (0.063 g, 0.35 mmol) in N,N-dimehtylacetamide
(5 mL) according to the procedure described for 6. Isolated as a yel-
low oil (0.041 g, 38%); 1H NMR (DMSO-d6) d 0.87 (t, 3H, J = 7.3 Hz),
1.53–1.63 (m, 6H), 2.33 (t, 2H, J = 6.7 Hz), 3.17 (s, 2H), 3.58 (s, 3H),
3.61 (s, 3H), 3.84 (br t, 2H), 4.14 (br s, 1H), 4.42 (s, 2H), 6.20 (dd,

N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
4065
1H, J1 = 2.6 Hz, J2 = 8.8 Hz), 6.35–6.38 (m, 3H), 6.73 (d, 1H,
J = 8.8 Hz), 7.19 (d, 1H, J = 8.5 Hz), 7.42 (d, 1H, J = 8.5 Hz), 7.64 (br
s, 2H), 7.95 (s, 1H); 13C NMR (DMSO-d6) d 12.3, 20.6, 22.1, 29.0,
33.7, 49.5, 52.1, 57.4, 68.5, 80.0, 101.8, 105.8, 110.6, 115.4, 122.1,
122.9, 128.3, 127.2, 133.2, 141.5, 144.5, 149.8, 163.3, 174.1; MS
(ES+): m/z = 468 [M+H].
zyme. However, diffraction quality crystals were obtained only
with 6b. Complex of 6b with T. cruzi enzyme was obtained by incu-
bating bifunctional TcDHFR–TS enzyme (10 mg/ml) with 1 mM
NADP, 1 mM dUMP and 1 mM inhibitor for 3 h at 4 °C. Crystals
were grown from 50 mM HEPES, pH 7.0, 10 mM magnesium chlo-
ride and 5–10% ethylene glycol (EDO) using 1.5–1.7 M ammonium
sulfate as precipitant. Crystals appeared in 2–3 days and reached
their maximum size (0.2 Â 0.2 Â 0.15 mm) in two weeks. These
crystals belong to tetragonal space group P43212 with four mole-
cules in the asymmetric unit. For cryo-protection we used 30%
EDO in the mother liquor. Intensity data for the complex were col-
lected at the Advanced Photon Source (APS) synchrotron IMCA
4.2.6. Preparation of 4-(5-[(2,4-diamino-6-quinazolinyl)
methyl]amino-2-methoxyphenoxy) butanoic acid (9)
To a stirred solution of ethyl 4-(5-[(2,4-diamino-6-quinazoli-
nyl)methyl]amino-2-methoxy phenoxy)butanoate (6b) (0.1 g,
0.23 mmol) in THF (5 mL) and MeOH (2 mL) was added 1 N. NaOH
solution (2.5 mL) at room tempareture for 4 h. Then it was acidified
with 1 N HCl and solvent was completely removed under reduced
pressure. The residue obtained was dissolved in dichloromethane
(25 mL) and washed with water (3 Â 15 mL), brine (1 Â 15 mL)
and dried over Na2SO4. The drying agent was removed and the sol-
vent was concentrated in vacuum to obtain the crude product. The
crude product was purified by flash chromatography on silica gel
using CHCl3–methanol (9:1) as the eluent to give 4-(5-[(2,4-diami-
*
17ID beam line and were processed using HKL2000 and D TREK
program packages.19,20
The crystal structure was solved by molecular replacement with
the structure of one molecule of ligand-free TcDHFR–TS (PDB Id:
2H2Q) as search model using the program PHASER.21 Difference
electron density maps allowed the placement of cofactor and
inhibitor in all four subunits (A, B, C, and D) of the inhibitor com-
plex. Several atoms in the ethylbutanoate chain were disordered
and the occupancy of these atoms was kept at zero. No density
for the dUMP molecule was observed and instead one sulfate ion
(from the crystallization medium) was located at the position nor-
mally occupied by the phosphate group of dUMP. Several rounds of
map fitting and structure refinement were carried out with pro-
grams COOT22 and REFMAC5.23
no-6-quinazolinyl)methyl]amino-2-methoxy
phenoxy)butanoic
acid (9) (0.067 g, 72%) as a yellow oil; 1H NMR (DMSO-d6) d 1.88
(q, 2H, J = 6.5 Hz), 2.32 (t, 2H, J = 6.5 Hz), 3.59 (s, 3H), 3.85 (t, 2H,
J = 6.5 Hz), 4.16 (s, 2H), 6.04 (br s, 1H), 6.05 (d, 1H, J = 7.7 Hz),
6.33 (s, 1H), 6.66 (d, 1H, J = 8.5 Hz), 6.92 (br s, 2H), 7.21–7.24 (m,
1H), 7.58 (d, 1H, J = 7.7 Hz), 7.84 (br s, 2H), 8.07 (s, 1H); 13C NMR
(DMSO-d6) d 24.5, 30.4, 47.5, 56.7, 67.1, 100.2, 103.2, 109.2,
109.6, 114.7, 122.7, 133.2, 140.6, 143.8, 149.1, 162.6, 174.5; MS
(ES+): m/z = 398[M+H].
4.5. Molecular docking and energy analysis
Ligand structures generated with the program package Marvin-
Beans/JChem (ChemAxon; see: http://www.chemaxon.com/) pro-
vided the input files for docking into the active sites of TcDHFR
and hDHFR using the program iGemDock.24,25 The program in-
cludes routines for target and database preparation, molecular
docking, and post-docking analysis. The active site of PDB entry
3CLB (TcDHFR–TS in complex with NADP and TMQ) was used for
docking of the inhibitors in case of TcDHFR. The ligand binding area
was defined by including the cofactor and all protein residues
within 8 Å distance from the ligand (TMQ). Atom formal charges
and atom types are then added. During docking the protein was
kept rigid since the docking program does not allow the inclusion
of flexibility of protein residues or defined side chain rotamers. On
the other hand the docking procedure executes flexible docking for
each ligand in the ligand database. For scoring we used the Gem-
dock scoring function that consists of a simple empirical scoring
function and a pharmacophore-based scoring function (includes
electrostatic, steric, and hydrogen bonding potentials), which is de-
scribed in the literature.24,25 In a final step, iGemDock re-ranks and
sorts all docked ligand conformations that can be displayed in the
post-docking analysis.
Docked poses of TMQ and 6b in TcDHFR were compared with
the experimental binding modes in the crystal structure (3HBB
and 3KJS, respectively). In the absence of a crystal structure of
hDHFR with TMQ we used the structure of hDHFR with an analog
of TMQ (PDBID: 1PD8). The docked poses that provided the closest
structural alignment of the 2,4-diaminoquinazoline ring of each
inhibitor with the same moiety in TMQ, 6b or CO4 were considered
as most likely binding modes.
We also calculated estimated binding affinities based on docked
poses of all six compounds for TcDHFR and hDHFR using the
PEARLS web server (http://ang.cz3.nus.edu.sg/cgi-bin/prog/ru-
ne.pl)26 and compared these with data obtained for the X-ray poses
of TMQ, 6b, and CO4 (Table 3). An energy analysis for this set of
compounds shows the contribution of van der Waals and hydrogen
bond energy to the total free energy (Table 4). Figures have been
prepared using PyMOL.27
4.3. Enzyme inhibitory assay
Recombinant bifuntional enzyme (TcDHFR–TS) was used
throughout this study. Inhibitory activity of the reported com-
pounds was measured in a spectrophotometric assay using the
bifunctional enzyme as described previously.6 Briefly, the reaction
mixture for TcDHFR contained 1 ml of 50 mM Tris–HCl buffer, pH
7.0, dihydrofolic acid (DHF), 100 lM NADPH and 0.5–1.0 lg of en-
zyme, and the reaction velocity was measured by monitoring
absorbance at 340 nm for 30 s at 22.6 °C. Human DHFR (hDHFR)
activity was measured in 50 mM KH2PO4 buffer, pH 7.3, containing
250 mM KCl and 5 mM 2-mercaptoethanol. The IC50 value (concen-
tration of inhibitor required for 50% inhibition) of each inhibitor
was determined by measuring reaction velocity at several inhibitor
concentrations for two (hDHFR) or three (TcDHFR) different DHF
concentrations, and then averaged. All measurements were con-
ducted in triplicate. The reproducibility and stability of the com-
pound assays were verified by repeating the assays on different
days.
In general, Ki values for each inhibitor were calculated from the
corresponding IC50 values. The formulas used for calculation of IC50
and Ki values from the observed experimental reaction rates are
shown in Table 1. However, in order to judge the accuracy of this
calculation we experimentally determined Ki values for com-
pounds 6a and 6b (for TcDHFR) for comparison with calculated val-
ues. For these measurements reaction velocities were obtained
using five different concentrations of DHF at three different inhib-
itor concentrations. The experimental Ki values were determined
from Lineweaver–Burk plots using the ‘Enzyme Kinetics’ module
in SigmaPlot (Systat Software Inc.).
4.4. Crystallization of ternary complex with 6b, intensity data
collection, structure solution
Attempts were made to crystallize ternary (inhibitor:cofac-
tor:enzyme) complexes of 6a and 6b using the bifunctional en-

4066
N. Schormann et al. / Bioorg. Med. Chem. 18 (2010) 4056–4066
12. Zuccotto, F.; Brun, R.; Gonzalez Pacanowska, D.; Ruiz Perez, L. M.; Gilbert, I. H.
Acknowledgments
Bioorg. Med. Chem. Lett. 1999, 9, 1463.
13. Chowdhury, S. F.; Villamor, V. B.; Guerrero, R. H.; Leal, I.; Brun, R.; Croft, S. L.;
Goodman, J. M.; Maes, L.; Ruiz-Perez, L. M.; Pacanowska, D. G.; Gilbert, I. H. J.
Med. Chem. 1999, 42, 4300.
14. Anderson, A. C. Acta Crystallogr. F Struct. Biol. Cryst. Commun. 2005, 61, 258.
15. Pan, H.; Lee, J. C.; Hilser, V. J. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 12020.
16. Schnell, J. R.; Dyson, H. J.; Wright, P. E. Ann. Rev. Biophys. Biomol. Struct. 2004,
33, 119.
17. Matthews, D. A.; Bolin, J. A.; Burridge, J. M.; Filman, D. J.; Volz, K. W.; Kaufman,
B. T.; Beddell, C. R.; Champness, J. N.; Stammers, D. K.; Kraut, J. J. Biol. Chem.
1985, 260, 381.
18. Cody, V.; Luft, J. R.; Pangborn, W.; Gangjee, A. Acta Crystallogr. D Biol. Crystallogr.
2003, 59, 1603.
19. Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
20. Pflugrath, J. W. Acta Crystallogr. D Biol. Crystallogr. 1999, 55, 1718.
21. McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.;
Read, R. J. J. Appl. Crystallogr. 2007, 40, 658.
22. Emsley, P.; Cowtan, K. Acta Crystallogr. D Biol. Crystallogr. 2004, 60, 2126.
23. Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr. D Biol. Crystallogr.
1997, 53, 240.
This work was in part supported by funds from Mousetrap
Foundation. K.W. and D.C. are also supported by a grant-in-aid
(award ID 0855076E) from American Heart Association.
References and notes
1. Moncayo, A.; Silveira, A. C. Mem. Inst. Oswaldo Cruz 2009, 104, 17.
2. Ferone, R.. In Antimalarial Drugs; Peter, W., Richards, W. H. G., Eds.; Springer:
New York, 1984; Vol. 68, pp 207–221.
3. Olliari, P. L.; Yuthavong, Y. Pharmacol. Ther. 1999, 81, 91.
4. Chan, D. C. M.; Anderson, A. C. Curr. Med. Chem. 2006, 13, 377.
5. Yuthavong, Y.; Yuvaniyama, J.; Chitnumsub, P.; Vanichtanankul, J.;
Chusacultanachai, S.; Tarnchompoo, B.; Vilaivan, T.; Kamchonwongpaisan, S.
Parasitology 2005, 130, 249.
6. Senkovich, O.; Bhatia, V.; Garg, N.; Chattopadhyay, D. Antimicrob. Agents
Chemother. 2005, 49, 3234.
7. Senkovich, O.; Schormann, N.; Chattopadhyay, D. Acta Crystallogr.
Crystallogr. 2009, 65, 704.
D Biol.
24. Yang, J.-M.; Chen, C. C. Proteins 2004, 55, 288.
25. Yang, J.-M. J. Comput. Chem. 2004, 25, 843.
8. Schormann, N.; Senkovich, O.; Walker, K.; Wright, D. L.; Anderson, A. C.;
Rosowsky, A.; Ananthan, S.; Shinkre, B.; Velu, S.; Chattopadhyay, D. Proteins
2008, 73, 889.
9. Davoll, J.; Johnson, A. M.; Davies, H. J.; Bird, O. D.; Clarke, J.; Elslager, E. F. J. Med.
Chem. 1972, 15, 812.
10. Gilbert, I. H. Biochim. Biophys. Acta 2002, 1587, 249.
11. Zuccotto, F.; Zvelebil, M.; Brun, R.; Chowdhury, S. F.; Di Lucrezia, R.; Leal, I.;
Maes, L.; Ruiz-Perez, L. M.; Gonzalez Pacanowska, D.; Gilbert, I. H. Eur. J. Med.
Chem. 2001, 36, 395.
26. Han, L. Y.; Lin, H. H.; Li, Z. R.; Zheng, C. J.; Cao, Z. W.; Xie, B.; Chen, Y. Z. J. Chem.
Inf. Model. 2006, 46, 445.
27. DeLano, W. L. DeLano Scientific; San Carlos: CA, 2002.
28. Reche, P.; Arrebola, R.; Olmo, A.; Santi, D. V.; Gonzalez-Pacanowska, D.; Ruiz-
Perez, L. M. Mol. Biochem. Parasitol. 1994, 65, 247.
29. White, E. L.; Ross, L. J.; Cunningham, A.; Escuyer, V. F. E. M. S. Microbiol. Lett.
2004, 232, 101.