Design, synthesis and evaluation of 2,4-diaminoquinazolines as inhibitors of trypanosomal and leishmanial dihydrofolate reductase

Article abstract of DOI:10.1016/j.bmc.2005.01.025

This paper describes the design, synthesis and evaluation of a series of 2,4-diaminoquinazolines as inhibitors of leishmanial and trypanosomal dihydrofolate reductase. Compounds were designed by a generating virtual library of compounds and docking them into the enzyme active site. Following their synthesis, they were found to be potent and selective inhibitors of leishmanial dihydrofolate reductase. The compounds were also found to have potent activity against Trypanosoma brucei rhodesiense, a causative organism of African trypanosomiasis and also against Trypanosoma cruzi, the causative organism of Chagas disease. There was significantly lower activity against Leishmania donovani, one of the causative organisms of leishmaniasis.

Full text of DOI:10.1016/j.bmc.2005.01.025

Bioorganic & Medicinal Chemistry 13 (2005) 2637–2649
Design, synthesis and evaluation of 2,4-diaminoquinazolines as
inhibitors of trypanosomal and leishmanial dihydrofolate reductase
Soghra Khabnadideh,a,b, Didier Pez,a, Alexander Musso,c Reto Brun,d
c
Luis M. Ruiz Perez, Dolores Gonzalez-Pacanowska and Ian H. Gilbert
c
a,
*
´
´
aWelsh School of Pharmacy, Cardiff University, Redwood Building, King Edward VII Avenue, Cardiff, CF10 3XF, UK
bDepartment of Pharmacy, University of Shiraz, Shiraz, Iran
c
´
´
Instituto de Parasitologıa y Biomedicina ‘Lopez-Neyra’, CSIC, Avda. del Conocimiento s/n,
´
Parque Tecnologico de Ciencias de la Salud, 18100 Armilla, Granada, Spain
dSwiss Tropical Institute, Socinstrasse 57, CH-4002 Basel, Switzerland
Received 3 August 2004; accepted 14 January 2005
Abstract—This paper describes the design, synthesis and evaluation of a series of 2,4-diaminoquinazolines as inhibitors of leishman-
ial and trypanosomal dihydrofolate reductase. Compounds were designed by a generating virtual library of compounds and docking
them into the enzyme active site. Following their synthesis, they were found to be potent and selective inhibitors of leishmanial dihy-
drofolate reductase. The compounds were also found to have potent activity against Trypanosoma brucei rhodesiense, a causative
organism of African trypanosomiasis and also against Trypanosoma cruzi, the causative organism of Chagas disease. There was sig-
nificantly lower activity against Leishmania donovani, one of the causative organisms of leishmaniasis.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
however relatively little has been done to develop this as
a drug target in leishmania and trypanosomes. Most
clinically used DHFR inhibitors (for example, pyrimeth-
amine, cycloguanil, methotrexate and trimethoprim) are
not selective or only showpoor selectivity for the tryp-
anosomal and leishmanial enzymes. Relatively few
selective inhibitors for leishmanial or trypanosomal
DHFR have been reported in the literature.1 However,
Sirawaraporn et al.9 described some substituted 5-benz-
yl-2,4-diaminopyrimidines as selective inhibitors of
the leishmanial enzyme. In particular the 30-octyloxy
derivative (1, Fig. 1) showed promise.
Parasitic protozoa cause a number of diseases, including
leishmaniasis (caused by species of Leishmania), Chagas
disease (caused by Trypanosoma cruzi) and African tryp-
anosomiasis (caused by Trypanosoma brucei rhodesiense
and T. b. gambiense). These diseases are found predom-
inantly in the developing world, where they cause signif-
icant rates of morbidity and mortality. The current
drugs available to treat these diseases suffer from a
number of disadvantages, including toxic side effects,
poor clinical efficacy, parenteral administration and
increasing problems with resistance. Therefore there is
an urgent need for newdrugs acting at newtherapeutic
targets to tackle these diseases.
We have previously reported a detailed structure–activ-
ity relationship study of 2,4-diaminopyrimidines,2,7,8
varying substituents at the 6-, 30- and 40-positions
(Structure 2, Fig. 1). These studies have led to some po-
tent inhibitors of the enzymes. In particular alkoxy sub-
stituents at the 30- or 40-position gave selective inhibition
of the parasite enzyme, with maximum potency and
selectivity when the chain length was 2–6 carbon atoms.
Compounds also showed activity against T. brucei and
T. cruzi, parasites with maximum activity with longer
chain lengths (6–10 carbons atoms).
We have been interested in dihydrofolate reductase as a
drug target in these organisms.1–8 This enzyme, which
reduces dihydrofolate to tetrahydrofolate (Fig. 1) has
successfully been used as a drug target in other diseases;
*
Corresponding author. Fax: +44 29 2087 4149; e-mail: gilbertih@
cardiff.ac.uk
These authors contributed equally to the work published.
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.01.025

2638
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
O
CO H
2
O
N
N
H
CO H
2
N
HN
H N
N
H
N
H
2
dihydrofolate
NH
N
NH
N
2
2
R
= H, OH, O(CH ) CH
2 n 3
3'-position
4'-position
R
R
O(CH ) CH
2
2
3
2 7
3
N
N
H N
2
R
1
H N
2
R
= H, OH, O(CH ) CH
2 n 3
3
2
1
R
= H, Me, Et, Pr, Bn
6-position
1
IC (L. major DHFR) 0.2µM.
50
IC (human DHFR) 26µM.
50
IC (L. donovani intracellular amastigotes) 4.6µM
50
Figure 1.
We have carried out some modelling studies to try and
explain the activity and selectivity of our compounds.
We compared the active sites of L. major, T. cruzi and
T. brucei DHFR and compared them with the human
enzyme.2,4,8 This data suggests that the active sites of
the protozoal enzymes are relatively similar to one an-
other and have significant differences compared to the
human enzyme. Further the protozoal enzymes have
more lipophilic residues than the human enzyme in the
active site. In particular the parasitic enzymes have a
phenylalanine residue, which interacts with the gluta-
mate part of the substrate, whilst the human enzyme
has an asparagine residue. Our modelling studies indi-
cate that compounds with a lipophilic substituent (com-
pound 2, R2 or R3 = O-Alkyl) interact with this
phenylalanine residue, in the case of the Leishmania
enzyme.
2. Modelling
Literature precedence from the work of BermanÕs group
and others suggested that the 2,4-diaminoquinazoline
may be a good starting point for design of potent anti-
trypanosomal or anti-leishmanial activity.11 Some 2,4-
diaminoquinazolines have been shown to have in vivo
activity in a rodent model of Chagas disease.12–14 We
wanted to combine the 2,4-diaminoquinazoline or re-
lated functionality with a lipophilic side chain, as we
have shown compounds with a lipophilic side chain to
give rise to selectivity for the leishmanial and trypanos-
omal enzymes. We decided to use modelling to optimise
the design of these newcompounds. From our previous
modelling studies (Fig. 2),2,4,15 the key interactions in
the active site are: (i) hydrogen bonds between the dia-
minopyrimidine moiety and an acidic residue in the ac-
tive site (Glu30, Asp52 or Asp48 in the human, L.
major and T. cruzi enzymes, respectively); (ii) the lipo-
philic side chain which interacts with a phenylalanine
residue in the case of L. major, T. cruzi or T. brucei.
In the human enzyme this phenylalanine is replaced by
an asparagine, which does not undergo a favourable
interaction with the lipophilic side chain.
In addition we investigated the active site of the L. major
and human enzymes with the programme GRID,10
using a hydrophobic probe. In the leishmanial enzyme
active site, the regions where the hydrophobic probe
interacts favourably with the enzyme active site has a
larger volume and surface, compared to the human ac-
tive site. This shows that there are more regions of the
leishmanial active site, which can form thermodynami-
cally favourable interactions with hydrophobic substitu-
ents than in the case of the human enzyme. Thus
appropriately sized hydrophobic substituents will under-
go favourable interactions with the leishmanial enzyme,
but unfavourable interactions with the human enzyme.8
This is consistent with the experimental results.
Many of the 2,4-diaminopyridimine derivatives that we
have prepared are potent and selective inhibitors of
the leishmanial and trypanosomal enzymes and some
of the compounds showpotent growth inhibition of
the clinically relevant life-cycle stage of the intact para-
site.2,7,8 However the compounds that we prepared have
shown very limited in vivo activity in rodent models of
infection. We wanted to see if we could prepare a library
of more potent inhibitors and to see if we could eventu-
ally derive inhibitors with in vivo activity.
Figure 2. Important components in the compound structure.

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
2639
To facilitate the modelling, we divided the potential
inhibitors into three components:
Component C is a spacer. It plays an important role be-
cause it places the lipophilic side chain in optimal posi-
tion relative to Met53 and Phe91 of the L. major active
site or relative to Met49 and Phe88 of the T. cruzi active
site. Different potential spacers were used, including
compounds with either one or two carbon atoms. In
some cases oxygen or nitrogen were included to provide
potential hydrogen bond donors and acceptors. In total,
eight different linkers were designed including four link-
ers with one atom and fours linker with two atoms.
• Component A is the 2,4-diaminoquinazoline or 2,4-
diaminoquinazoline derivative, which makes hydro-
gen bonds with an acidic residue in the active site
(Glu30, Asp52 or Asp48 in the human, L. major
and T. cruzi enzymes, respectively).
• Component B is a lipophilic substituent, which gives
selectivity for the leishmanial and trypanosomal
enzyme over the human enzyme.
• Component C is the linker between components A
and B (Fig. 2).
Component B is the lipophilic side chain. Different lipo-
philic functions were tested such as a linear or substi-
tuted alkyl chain or aromatic rings with or without
hydroxyl group to see a possible difference in the confor-
mation after the docking. In total, thirteen different side
chains were designed.
We then chose a number of different possibilities for
components A, B and C and combined these. The frag-
ments chosen are shown in Table 1.
Component A was a heterocycle based on the 2,4-diami-
nopyrimidine moiety fused to a five- or six-member ring
with or without nitrogen atoms. These structures would
be analogous to other known DHFR inhibitors. In to-
tal, nine different structures were considered including
2,4-diaminoquinazoline, pteridine, pyridopyrimidine
and pyrrolopyrimidine structures.
It is also important to select compounds which have
Ôdrug-likeÕ properties.16 One criteria that can be used
are LipinskiÕs rules for oral bioavailability.17
By combination of the different components, A, B and
C, a library of 936 molecules was generated. All these
compounds have a molecular weight ranging from 285
Table 1. Fragments used for generating compounds for docking
Entry
Component A
Component C
Component B
NH
2
P
X
N
1
OR
H N
2
N
Q
P, Q = CH or N
X = CH2, O, NH, NMe
R = H, Et, Bu, Hex, Octyl, Benzyl
NH NH
2
2
2
3
4
5
N
N
X
O
O
O
H N
2
N
X = CH2, O, NH, NMe
NH
2
H N
2
N
NH
2
NH
2
OH
N
N
O
O
H N
2
N
N
H
NH
2
N
H
H N
2
N
6
7
O
O

2640
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
to 409, between 2 and 5 hydrogen bond donors, and be-
tween 5 and 9 hydrogen bond acceptors, as indicated by
LipinskiÕs rules. The ClogP range of the molecules was
between ꢀ0.72 and 7.12. However, this parameter was
not considered for the filtering protocol.
The compounds were drawn under the Sybyl pro-
gramme. Their formal charges computed using the Gast-
inger method and their conformation energy minimised
with the Powell method. The compounds were then
docked into the human, L. major and T. cruzi DHFR
active sites using FlexX programme. The compounds
were docked into the human enzyme, in order to esti-
mate the possibility of selectivity.
Figure 4. The predicted binding of one of the designed inhibitors
(compound 3) in the active site of the human enzyme.
O
NH
N
R
There was a lack of correlation between the FlexX scor-
ing function and what was observed visually; therefore
visual inspection was used to assess binding of ligands.
Firstly, the heterocyclic moiety was evaluated for its
ability to bind Glu30, Asp52 or Asp48 of, respectively,
the human, L. major and T. cruzi enzymes. Secondly,
the spacer units were assessed for their ability to posi-
tion the aromatic ring of the side chain in a conforma-
tion to give a lipophilic contact with Met53 or Met49
of, respectively, the L. major and T. cruzi enzymes. Fi-
nally, the side chain was evaluated for its capacity to of-
fer the maximal lipophilic contact with Phe91 or Phe88
of, respectively, the L. major and T. cruzi enzymes.
2
N
H
2
N
4
Figure 5. General structure of designed compounds.
favourable interactions of a hydrophobic probe with
the active site.
The predicted binding of one of the compounds, com-
pounds 3, is shown in the active site of the L. major en-
zyme in Figure 3. The 2,4-diaminoquinazoline formed
hydrogen bonds with Asp52, one of the aromatic rings
had a lipophilic interaction with Met53 and the terminal
aromatic ring had a good lipophilic interaction with
Phe91. Moreover, the terminal (phenolic) hydroxyl
group showed a strong interaction with the side chain
In general, all the heterocyclic rings (subunits A) gave a
satisfactory binding with Glu30, Asp52 or Asp48. How-
ever, the heterocyclic rings (entries 2–5) were eliminated
because they did not offer any binding improvement.
These fragments contained extra hydrogen bond do-
nors, but no additional hydrogen bonds could be de-
tected with the protein active site.
˚
amino of Lys57 (2.93 A), which could either be a hydro-
gen bond, or an electrostatic interaction.
The two-unit linkers (component C, entry 2) showed
that they could place the aromatic ring of the side chain
in a better position than the one-unit linkers (compo-
nent C, entry 1). The heteroatoms did not contribute
to increase the number of hydrogen bonds whether the
compound had a one- or two-unit linker. One of the best
and simplest spacer would therefore be the ethylene
linker.
The result of the docking of compound 3 into the human
DHFR active site was also analysed. In this case, it was
observed that the 2,4-diaminoquinazoline ring and the
middle aromatic ring exhibited the same kind of interac-
tions with Glu30 and Phe31 as those in the L. major
DHFR active site with Asp52 and Met53. However,
the aromatic ring at the end of chain C was more bent
due to a hydrogen bond between the terminal hydroxyl
˚
group of the molecule and Arg70 (2.89 A, Fig. 4).
All the lipophilic side chains gave a satisfactory confor-
mation. The lipophilic side chains were found in areas of
the active site, where GRID predicted energetically
The docking of compound 3 into the T. cruzi DHFR ac-
tive site did not give a satisfactory conformation. This
failure could not be explained.
The analysis of the docked conformations of all the
compounds of the newlibrary showed that compounds
of general structure 4 (Fig. 5), where R is a lipophilic
group, could be potentially newDHFR inhibitors. Their
potential selectivity for the parasite enzymes over the
human enzyme has also been illustrated.
3. Chemistry
The compounds prepared are shown in Figure 6. The
synthetic route to these compounds are shown in
Figure 3. The predicted binding of one of the designed inhibitors
(compound 3) in the active site of the L. major enzyme.

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
2641
O
R
O
O
NH
N
2
NH
N
R
NH
N
R
2
2
N
N
N
H N
2
H N
2
H N
2
5, R = H
12, R = CH2Ph
13, R = CH2Ph
6, R = Et
7, R = Bu
8, R = Hex
9, R = Octyl
10, R = Decyl
11, R = CH2Ph
Figure 6. Compounds prepared for evaluation.
NH
N
Schemes 1–3. Our strategy is a modification of method-
ology developed by Harris et al.18,19 Essentially, the
molecules were prepared by preparation of a 2,4-diami-
noquinazoline (Scheme 1) and then coupling it to a
phenoxy derivative by either a Sonogoshira (Scheme 2)
or Heck coupling (Scheme 3).
2
NC
H N
(a)
NC
H N
I
(b)
I
N
2
2
88%
90%
H N
2
14
15
16
Scheme 1. Reagents and conditions: (a) ICl, AcOH, rt; (b) guanidine,
150 ꢁC, 2 h.
The 2,4-diaminoquinazoline moiety was prepared in two
steps.18 Firstly 2-aminobenzonitrile (14) was iodinated
with iodine monochloride to give the iodide 15. The next
step was quinazoline formation by cyclisation with gua-
nidine. Initially this step proved to be problematic in
terms of reproducibility and yield. However by carrying
out the reaction in the absence of solvent, it was possible
to form the product 16 in reliably in high yield.
O
Ph
O
Ph
OH
O
I
Ph
(a)
(b)
(c)
92%
94%
62%
I
SiMe
H
3
17
18
19
20
(d)
40%
O
O
NH
N
NH
2
2
(e)
N
N
43%, after 6h
H N
2
H N
2
N
13
11
OH
NH
N
2
N
H N
2
50%, after 6h
5
Scheme 2. Reagents and conditions: (a) BnBr, K2CO3, 2-butanone; (b) Me3SiC„CH, (Ph3P)2PdCl2, CuI, Et3N; (c) K2CO3, MeOH; (d) 16,
(Ph3P)2PdCl2, CuI, Et3N, DMF, 50 ꢁC; (e) H2, Pd(OH)2, DMF.

2642
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
O
H
H
O
H
(c)
(b)
(a)
45%
75%
75%
OH
O
Ph
O
Ph
21
22
23
OH
NH
2
(d)
O
NH
2
N
N
H N
2
N
70%
H N
N
2
12
5
Scheme 3. Reagents and conditions: (a) BnBr, K2CO3, 2-butanone, reflux; (b) Ph3PMe+Brꢀ, K2CO3, dioxane, water, reflux; (c) 16, Pd(OAc)2,
(MePh)3P, piperidine, DMF, 120 ꢁC; (d) H2, Pd(OH)2, DMF.
The next stage in the synthesis is to add on a phenoxy
derivative. This was done using two methods. In the first
method, the 2,4-diaminoquinazoline 16 was coupled to a
phenoxy derivative using a Sonogoshira coupling.20 In
order to perform a Sonogoshira coupling, an acetylene
derivative was required. The starting point was iodophe-
nol (17) (Scheme 2). The phenol group was protected as
the benzyl ether (18), followed by replacement of the
ethylene derivative 23. A Heck coupling was then car-
ried out with 6-iodo-2,4-diaminoquinazoline (16) to give
12, similarly to the conditions described by Graffner-
Nordberg et al.21 This double bond could then be re-
duced and the benzyl group removed using hydrogen
in the presence of PearlmanÕs catalyst.
The final step of the synthetic scheme was alkylating the
free hydroxyl analogue 5. This was carried out using the
required alkyl iodide with potassium carbonate in etha-
nol (Scheme 4).
iodo-group with
a trimethylacetylene functionality
(19). The trimethylsilyl group was removed using potas-
sium carbonate in methanol (20), followed by coupling
to the 6-iodo-2,4-diaminoquinazoline using Sonogoshira
conditions (13). The final step was a simultaneous
hydrogenation of the acetylene functionality and
hydrogenolysis of the benzyl group. If the reaction was
not left to go to completion, it was possible to isolate
an intermediate compound, in which the acetylene
group had been reduced, but the benzyl group had not
been removed. Leaving the reaction for 24 h, led to
increasing levels of the required product (5) compared
to partially reduced product (11). Careful choice of con-
ditions was necessary; the solvent used was DMF and
the catalyst PearlmanÕs catalyst.18 Use of other catalysts
or solvents failed to yield the desired product.
4. Enzyme assays
Compounds were assayed against recombinant L. major
and human DHFR. The results of these assays are pre-
sented in Table 2. There was no clear trend for the alkyl
derivatives, all showing potent inhibition against the L.
major enzyme. The decyl derivative (10) showed the
weakest inhibition. Comparing compounds 11 and 12,
which both have a benzyl group at the terminal position,
11 is the most potent. This may be due to the alkene
group in 12 preventing the compound from adopting
an optimal conformation, in particular preventing the
terminal benzyl group from forming a p-stacking inter-
action with Phe91.
In order to carry out a Heck coupling, the ethylene
derivative of the phenoxy compound was required.
The starting point was 4-hydroxybenzaldehyde, which
was protected as the benzyl ether (22) (Scheme 3). A
Wittig reaction was then performed to give the required
Compounds gave a weaker inhibition of the human en-
zyme than of the L. major enzyme, but followed similar
OR
OH
NH
N
2
NH
N
2
R-I, K CO , EtOH
2
3
N
N
H N
2
H N
2
6, R = Et, 56%
5
7, R = Bu, 64%
8, R = Hex, 75%
9, R = Oct, 70%
10, R = Decyl, 45%
Scheme 4.

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
Table 2. Inhibition of recombinant L. major and human DHFR
2643
O
NH
N
R
2
X
N
Y
H N
2
Compound
X–Y
R
L. major
IC50 (lM)
Human
Selectivity
Ki (lM)
Ki (lM)
IC50 (lM)
5
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH@CH
H
0.0014
0.0078
0.0019
0.0076
0.0053
0.17
0.034
0.18
0.034
0.048
0.042
0.072
0.269
1.288
0.038
0.908
1.65
2.34
2.04
3.47
12.89
61.7
1.85
43.5
24.2
6.15
22.1
9.47
50.7
7.57
6
Ethyl
Butyl
Hexyl
Octyl
Decyl
CH2Ph
CH2Ph
7
0.044
0.177
0.123
3.97
8
9
10
11
12
0.0052
0.212
0.12
4.89
7.3
4.28
trends. Thus the alkyl derivatives (5–10) gave similar lev-
els of inhibition, although the octyl (9) and decyl (10)
derivatives were slightly less active. For those compounds
that have a benzyl group at the terminal position, 12 with
an alkene was less effective than the reduced analogue 11.
compounds showed similar growth inhibition, except
the hydroxyl analogue (5), which showed low activity.
The benzyl analogue 11 was poorly active, although
interestingly the conformationally constrained ana-
logues of this compound showed more potent inhibition
of growth, despite the fact that 12 was much less active
against the enzyme compared to 11. With the exception
of these conformationally constrained analogues (12,
13) none of the compounds were selective for the leish-
mania parasite.
The compounds were all moderately selective for the
parasite enzyme; the highest selectivity index obtained
was 51.
5. Anti-parasite assays
All compounds were potent inhibitors in the nanomolar
range of the growth of T. b. rhodesiense. No clear trend
is observable. Of the alkyl derivatives, the unsubstituted
analogue (5) shows the most potent inhibition. The ben-
zyl derivative (11) also shows potent growth inhibition,
along with the alkene derivative, 12. The alkyne ana-
logue (13) shows weaker inhibition. The compounds
showa 4–16-fold selectivity for the parasite over the
mammalian cells.
Compounds were evaluated against axenic L. donovani
amastigotes. T. cruzi and T. brucei are parasites related
to Leishmania. An analysis of the homology models of
the T. cruzi and T. brucei DHFRs, revealed a great sim-
ilarity between the enzyme active sites. Therefore com-
pounds were also screened against intracellular T.
cruzi amastigotes and T. b. rhodesiense trypomastigotes
(Table 3). The stages of the life-cycle studied here are
the clinically relevant stages of the life-cycle.
The compounds also showed potent inhibition of the
growth of T. cruzi, w ith IC value in the nanomolar
50
Against L. donovani, the alkyl derivatives (5–10) showed
inhibition of growth at the micromolar level. All the
range. The activity seems to increase as the alkyl chain
is lengthened, with the ethyl (6) and butyl (7) derivatives
Table 3. The effect of the compounds on the growth of axenic L. donovani amastigotes, intracellular T. cruzi amastigotes and T. b. rhodesiense
trypomastigotes
O
NH
R
2
X
N
Y
H N
N
2
Compound
X–Y
R
T. b. rhodesiense IC50 (lM)
T. cruzi IC50 (lM)
L. donovani IC50 (lM)
L6-cells IC50 (lM)
5
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH@CH
C„C
H
0.054 (12)a
0.081 (10)
0.23 (5)
0.078 (8)
0.041 (20)
0.042 (26)
0.11 (14)
0.22 (6)
79.9 (0.01)
2.0 (0.4)
1.7 (0.6)
2.1 (0.7)
2.0 (0.6)
5.5 (0.7)
13.8 (0.1)
0.53 (3)
1.6 (3)
0.82
0.84
1.1
1.5
1.2
3.8
1.2
1.6
4.9
6
Ethyl
Butyl
Hexyl
Octyl
Decyl
CH2Ph
CH2Ph
CH2Ph
7
8
0.095 (16)
0.35 (4)
0.67 (6)
9
10
11
12
13
0.15 (25)
0.059 (20)
0.18 (9)
0.10 (12)
0.17 (9)
0.42 (12)
0.50 (10)
Standard drugs: T. b. rhodesiense, melarsoprol, IC50 0.004 lM; T. cruzi, benznidazole, IC50 1.5 lM; L. donovani, miltefosine, IC50 0.47 lM; L6-cells,
podophyllotoxin, IC50 0.005 lM.
a Selectivity data is shown in parentheses, where selectivity is defined: IC50 (L-6 cells)/IC50 (parasite).

2644
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
showing the most potent inhibition. As the chain length
is increased, the growth inhibition decreased (8–10). The
compounds showed a selectivity of 6–26-fold compared
to mammalian cells.
inhibition of the enzyme and also showrelatively weak
growth inhibition of T. b. rhodesiense and T. cruzi. This
implies that the compounds may be killing the parasites
by inhibition of DHFR. In other compounds there was
a lower correlation between inhibition of enzyme and
growth inhibition. However, it is unlikely that there
would be a direct correlation between enzyme inhibition
and growth inhibition in all cases, as in cellular assays
there are factors other than enzyme inhibition, which
may be important, such as penetration of compounds
into the cell, metabolism of compounds, etc.
The compounds appear to have similar levels of toxicity
against mammalian cells, with IC50 values at about
1 lM. For L. donovani, the compounds are less active
against the parasites than the mammalian cells.
6. Discussion
These compounds represent newleads in the search for
anti-parasitic agents, and this class of compound merits
further work for development as potential anti-trypano-
somal compounds.
DHFR is a potential drug target against Leishmania
and trypanosomes. However, it is necessary to find
inhibitors, which are sufficiently potent and selective
for the enzyme from these organisms, and which show
activity against the parasites in vitro and in vivo. In pre-
vious papers we have reported a number of compounds
based on 2,4-diaminopyrimidines, which are potent
inhibitors of the enzyme and showactivity in vitro,
but only showlimited activity in vivo. The use of a lipo-
philic substituent appears to give selectivity for the par-
asite enzyme over the human enzyme.
7. Experimental
All solvents and chemicals were purchased from Aldrich
Chemical Co. Reactions were monitored by TLC using
pre-coated silica gel 60 F254 plates. Chromatography
was performed with silica gel (220–240 mesh) prepacked
columns or by flash master chromatography. All reac-
tions requiring anhydrous conditions or an inert atmo-
sphere were performed under a positive pressure of
2,4-Diaminoquinazoline-based compounds have been
shown to have in vivo activity in a rodent model of Cha-
gas disease, extending the life-span of infected animals,
although not giving radical cure (Davoll et al., Elslager
et al. and Thompson et al.).12–14 We hypothesised that
coupling our lipophilic substituents to the 2,4-diamino-
quinazoline or a related moiety may give rise to com-
pounds with good enzyme selectivity and in vivo
activity.
1
N2. H NMR and 13C NMR spectra were recorded at
300 and 75 MHz, respectively.
The numbering system used in Figure 7 is used in inter-
preting the NMR spectra.
7.1. 2-Amino-5-iodobenzonitrile (15)19
In this paper we report the use of a combinatorial
docking routine to design a series of potential inhibi-
tors of DHFR based on 2,4-diaminoquinazolines. We
report the development of methodology to prepare
these compounds. The compounds prepared showed
potent activity against L. major DHFR. Surprisingly
there was relatively weak inhibition of the intact L.
donovani axenic amastigotes, and little correlation be-
tween inhibition of the enzyme and growth of L. dono-
vani. We cannot explain the lack of activity against the
L. donovani. Access of compounds to the parasite
should not be a problem, as in this model the parasite
is cultured axenically. However the low pH of the med-
ium (pH 5.4) may reduce diffusion of these basic com-
pounds into the parasites.
A solution of iodine monochloride (1.374 g, 8.46 mmol,
1 equiv) in acetic acid (10 mL) was added to a solution
of 14 (1 g, 8.46 mmol, 1 equiv) in acetic acid (10 mL).
The mixture was stirred at room temperature for 5 h.
Then the reaction mixture was poured into 50 mL cold
water. The solid was filtered off, washed with cold water
and dried. A recrystallisation with toluene and hexane
gave a light brown powder (1.82 g, 88%, mp = 85 ꢁC).
Rf = 0.45 (40% EtOAc/hexane); 1H NMR (CDCl3) d
7.71 (s, 1H, H6), 7.62 (d, 1H, J 8.7 Hz, H4), 6.6 (d,
1H, J 8.8 Hz, H3), 4.6 (s, 2H, NH2); 13C NMR (CDCl3)
d 149.51 (C2), 142.95 (C6), 140.41 (C4), 117.56 (C3),
116.51 (CN), 98.69 (C1), 98.68 (C5); MS (EI+) m/z
244.0 [M, 90%].
7.2. 6-Iodo-2,4-quinazolinediamine (16)19
Structurally the enzymes from T. brucei and T. cruzi are
very similar to those from L. major. Therefore if a com-
pound inhibits the L. major enzyme it will almost cer-
tainly inhibit the T. brucei or T. cruzi enzymes. Thus
compounds were evaluated against the parasites T. b.
rhodesiense and T. cruzi and showed potent growth inhi-
bition. There is some correlation between inhibition of
the enzyme and growth inhibition of T. b. rhodesiense
and growth inhibition of T. cruzi. Thus, for example,
compound 5 shows potent inhibition of the enzyme
and potent growth inhibition of T. b. rhodesiense and
T. cruzi; and compounds 10 and 12 showrelatively weak
Sodium ethoxide (4.858 g, 71.4 mmol, 3 equiv) was
added to
a
solution of guanidine hydrochloride
2"
3"
3'
O
4'
2'
2"
1"
4"
NH
2
a
5
1"
4
10
O
1'
N
1'
6
b
2
2
7
2'
1
H N
N
1
4'
2
9
3'
8
Figure 7. The numbering system used in interpreting NMR spectra.

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
2645
(6.82 g, 71.4 mmol, 3 equiv) in dry ethanol (30 mL) and
stirred for 20 min at room temperature. The suspension
was filtered through Celite and the solid washed with
ethanol. The filtrate evaporate in reduced pressure and
the viscose oil was heated for 2 h at 150 ꢁC with com-
pound 15 (5.8 g, 23.8 mmol, 1 equiv). The resulting solid
was cooled to 80 ꢁC, ethanol (30 mL) was added and re-
fluxed for 24 h. The flask was kept in the refrigerator
overnight. The precipitate was filtered off, washed with
cold ethanol and dried by high vacuum (6.12 g, 90%,
3.5 equiv) in dry methanol (50 mL) was stirred at room
temperature for 6 h. The reaction mixture filtered off
and the filtrate evaporate under reduced pressure. The
crude product extracted by ethyl acetate and washed
with water. The organic layer dried by Na2SO4 and
evaporate in vacuum to get a brown solid (2.72 g,
62%, mp = 65 ꢁC); Rf = 0.4 (10% EtOAC/hexane); 1H
NMR (CDCl3) d 7.6–7.4 (m, 7H, H30 + H50, H200–
H600), 6.95 (d, 2H, J 8.9 Hz, H20 + H60), 5.10 (s, 2H,
CH2), 3.05 (s, 1H, „CH); 13C NMR (CDCl3) d 159.06
(C10), 138.69 (C30 + C50), 134.07 (C100), 129.09
(C200 + C600), 128.56 (C400), 127.88 (C300 + C500), 117.74
(C20 + C60), 115.29 (C40), 83.50 (C2), 76.34 (C1),
70.46, (CH2); MS (ESꢀ) m/z 209 [M+1, 9%].
1
mp = 265 ꢁC); Rf = 0.2 (50% MeOH/CHCl3); H NMR
(DMSO-d6) d 8.37 (s, 1H, H5), 7.72 (d, 1H, J 8.8 Hz,
H7), 7.36 (s, 2H, NH2), 6.98 (d, 1H, J 8.8 Hz, H8),
6.17 (s, 2H, NH2); 13C NMR (CDCl3) d 167.79 (4),
161.35 (2), 152.19 (9), 140.82 (7), 132.33 (5), 126.98
(8), 112.95 (10), 82.57 (6); MS (ES+ and CI+) m/z 287
[M+1, 100%].
7.6. 6-{2-[4-(Benzyloxy)phenyl]-1-ethynyl}-2,4-quinazol-
inediamine (13)
7.3. 1-Benzyloxy-4-iodobenzene (18)
7.6.1. Method 1—Sonogoshira reaction. A solution of 16
(1.75 g, 6.12 mmol, 1 equiv), 20 (1.78 g, 8.56 mmol,
1.4 equiv), copper iodide (234 mg, 1.22 mmol,
0.2 equiv), bis(triphenylphosphine)palladium chloride
(430 mg, 0.61 mmol, 0.1 equiv) and triethylamine
(4.99 g, 7.9 mL, 49 mmol, 8 equiv) in DMF (20 mL) stir-
red at 50 ꢁC for 48 h. Solvent removed in vacuum and
the residue stirred in dilute aqueous ammonia (30 mL)
and ether (30 mL) for 30 min. The mixture filtrated
and purified by flash chromatography. A yellowsolid
obtained (0.9 g, 40%, mp = 260 ꢁC).
A solution of iodobenzene 17 (6 g, 27.3 mmol, 1 equiv),
anhydrous potassium carbonate (11.320 g, 81.9 mmol,
3 equiv), benzyl bromide (14 g, 9.8 mL, 81.9 mmol,
3 equiv) in 2-butanone (50 mL) was stirred under reflux
for 48 h. Then the reaction mixture was extracted with
ethyl acetate and washed with water. The organic layer
was dried by Na2SO4, evaporated in lowpressure and
dried by high vacuum. A white solid was obtained
(7.78 g, 92%); mp = 58 ꢁC (lit. 6122); Rf = 0.8 (10%
1
EtOAC/hexane); H NMR (CDCl3) d 7.52 (d, 2H, J
8.9 Hz, H3 + H5), 7.4–7.2 (m, 5H, H20–H60), 6.7 (d,
2H, J 8.9 Hz, H2 + H6), 5.0 (s, 2H, CH2); 13C NMR
(CDCl3) d 159.11 (C1), 138.74 (C3 + C5), 137 (C10),
129.14 (C20 + C60), 128.61 (C40), 127.94 (C30 + C50),
117.79 (C2 + C6), 83.59 (C4), 70.54 (CH2); MS (ES+)
m/z 311.4 [M+1, 6%].
7.6.2. Method 2—Heck coupling. A solution of 16
(0.418 g, 1.46 mmol, 1 equiv), 20 (0.458 g, 2.2 mmol,
1.5 equiv), copper iodide (57 mg, 0.3 mmol, 0.2 equiv),
palladium acetate (33 mg, 0.15 mmol, 0.1 equiv), tri-o-
tolylphosphine (0.09 g, 0.3 mmol, 0.2 equiv) and piper-
idine (1.02 g, 1.2 mL, 12 mmol, 8 equiv) in DMF
(10 mL) stirred at 90 ꢁC for 24 h. Solvent removed in
vacuum and purified by flash chromatography. A yellow
solid obtained (0.24 g, 45%, mp = 260 ꢁC); Rf = 0.5
7.4. 2-[4-(Benzyloxy)phenyl]-1-ethynyl(trimethyl)silane (19)
1
To a solution of 18 (7 g, 22.6 mmol, 1 equiv) in triethyl-
amine (50 mL), copper iodide (43 mg, 0.226 mmol,
0.01 equiv), bis(triphenylphosphine)palladium chloride
(159 mg, 0.226 mmol, 0.01 equiv) and trimethylsilylacet-
ylene (4.44 g, 6.4 mL, 45.2 mmol, 2 equiv) were added
and the reaction mixture was stirred at room tempera-
ture for 6 h. Then the solution was filtered off and con-
centrated under reduced pressure. The crude product
was purified by flash master chromatography with hex-
ane and ethylacetate. A yellowsolid product was ob-
tained (5.95 g, 94%, mp = 50 ꢁC); Rf = 0.5 (10%
(25% MeOH/CHCl3); H NMR (DMSO-d6) d 8.23 (s,
1H, H5), 7.50 (d, 1H, J 8.23 Hz, H7), 7.46–7.35 (m,
8H, H8, H2 + H60, H200–H600), 7.09 (d, 2H, J 8.8 Hz,
H30 + H50), 5.15 (s, 2H, CH2); 13C NMR (CDCl3) d
162.42 (C10), 161.70 (C40), 158.76 (C4 + C2), 152.83
(C9), 137.07 (C5), 135.15 (C7), 133.03 (C20 + C60),
128.84 (C200 + C600), 128.32 (C400), 128.15 (C300 + C500),
127.39 (C100), 124.99 (C8), 115.62 (C30 + C50), 115.21
(C6), 113.65 (C10), 89.06 (b), 88.42 (a), 69.71 (CH2);
MS (ES+) m/z 367.1 [M+1, 100%], HRMS (ES+)
(M+H)+ C23H19N4O+ requires 367.1553, found
367.1552.
1
EtOAC/hexane); H NMR (CDCl3) d 7.5–7.3 (m, 7H,
H30 + H50, H200–H600), 6.85 (d, 2H,
J
8.9 Hz,
H20 + H60), 5.10 (s, 2H, CH2), 0.3 (s, 9H, (CH3)3); 13C
NMR (CDCl3) d 158.53 (C10), 138.16 (C30 + C50),
133.4 (C100), 128.56 (C200 + C600), 128.03 (C400), 127.39
(C300 + C500), 117.21 (C20 + C60), 114.65 (C40), 105.06
(C2), 82.98 (C1), 69.97 (CH2), 0.0 ((CH3)3); MS (ES+)
m/z 281.1 [M+1, 9%].
7.7. 4[2(2,4-Diamino-6-quinazolinyl)ethyl]phenol (5), and
6-[4-(benzyloxy)phenethyl]2,4-quinazolindiamine (11)
Compound 13 (300 mg, 0.82 mmol) and palladium
hydroxide (300 mg) stirred in dry DMF (20 mL) under
H2 gas for 6 h at room temperature. The reaction mix-
ture filtrate on Celite, concentrated in vacuum and puri-
fied by column chromatography. Two distinctive
compounds were proved by spectroscopic methods: (5,
115 mg, 50%, mp = 275 ꢁC and 11, 130 mg, 43%,
mp = 215 ꢁC). We can increase the ratio of compound
7.5. 1-(Benzyloxy)-4-(1-ethynyl)benzene (20)
A solution of 19 (5.9 g, 21 mmol, 1 equiv) and anhy-
drous potassium carbonate (10.16 g, 73.5 mmol,

2646
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
5 up to 65% by increasing the time of reaction until
24 h.
5.66 (d, 1H, J 17.56 Hz, @CHH), 5.16 (d, 1H, J
10.98 Hz, @CHH), 5.08 (s, 2H, –CH2–); 13C NMR
(CDCl3) d 159 (C1), 137.37 (C10), 136.63 (@CH),
131.12 (C4), 129.04 (C20 + C60), 128.42 (C40), 127.90
(C30 + C50), 127.84 (C3 + C5), 115.31 (C2 + C6),
112.16 (@CH2), 70.46 (–CH2–); MS (CI+) m/z 211.1
[M+1, 100%], (EI+) m/z 210.1 [M, 9%].
1
Compound 5: Rf = 0.1 (40% MeOH/CHCl3); H NMR
(DMSO-d6) d 7.9 (s, 1H, H5), 7.45 (d, 1H, J 8.42 Hz,
H8), 7.18 (d, 1H, J 8.42 Hz, H7), 7.05 (d, 2H, J
8.6 Hz, H20 + H60), 6.7 (d, 2H, J 8.42 Hz, H30 + H50),
2.8 (s, 2H, a), 2.5 (s, 2H, b); 13C NMR (CDCl3) d
162.82 (C9), 159.14 (C2), 155.72 (C4 + C40), 134.92
(C6), 134.50 (C10), 131.86 (C5), 129.52 (C20 + C60),
123.05 (C7), 122.32 (C8), 115.40 (C30 + C50 + C10),
37.53 (a), 36.53 (b); MS (ES+) m/z 281.2 [M+1, 100%],
HRMS (ES+) (M+H)+ C16H17N4O+ requires 281.1397,
found 281.1404.
7.10. 6-{2-[4-(Benzyloxy)phenyl]-1-ethenyl}-2,4-quinazol-
inediamine (12)
Compound 16 (1.96 g, 6.85 mmol, 1 equiv), compound
23 (2.16 g, 10.26 mmol, 1.5 equiv) palladium acetate
(34 mg, 0.14 mmol, 0.02 equiv), tri-o-tolyl phosphine
(84 mg, 0.27 mmol, 0.04 equiv), piperidine (5.5 mL) in
dry DMF (9 mL) were stirred at 120 ꢁC for 60 h. Solvent
was removed in vacuo and the dark residue stirred with
10% KOH for 1 h. The precipitate filtrated, washed with
water and purified by column chromatography. A yellow
solid was obtained (1.13 g, 45%, mp = 275 ꢁC); Rf = 0.1
1
Compound 11: TLC Rf = 0.2 (40% MeOH/CHCl3); H
NMR (DMSO-d6) d 7.85 (s, 1H, H5), 7.5–7.3 (m, 6H,
H7, H200–H600), 7.2–7.1 (m, 4H, H20–H60), 6.9 (d, 1H,
J 8.8 Hz, H8), 5.1 (s, 2H, c), 2.9 (s, 2H, a), 2.5 (s, 2H,
b); 13C NMR (CDCl3) d 162.65 (C9), 160.55 (C4),
156.87 (C2 + C40), 137.59 (C5), 134.08 (C100), 133.87
(C6), 133.66 (C10), 129.65 (C20 + C60), 128.76
(C200 + C600), 128.11 (C400), 127.99 (C300 + C500), 124.22
(C7), 122.67 (C8), 114.94 (C30 + C50 + C10), 69.48 (c)
37.47(a), 36.46 (b); MS (ES+) m/z 371.2 [M+1, 100%],
HRMS (ES+) (M+H)+ C23H23N4O+ requires 371.1866,
found 371.1863.
1
(20% MeOH/CHCl3); H NMR (DMSO-d6) d 8.15 (s,
1H, H5), 7.75 (d, 1H, J 8.8 Hz, H8), 7.50–7.30 (m, 8H,
H7 + Ha + Hb + H20 + H60, H300–H500), 7.2 (d, 2H, J
8.23 Hz, 200 + 600), 7 (d, 2H, J 8.9, H30 + H50), 5.15 (s,
2H, CH2); 13C NMR (CDCl3) d 167.82 (C4), 162.83
(C40), 161.19 (C9), 158.16 (C2), 137.40 (C100), 131.11
(C5), 130.58 (C7), 130.22 (a), 129.26 (C20 + C60), 128.8
(b), 128.56 (C200 + C600), 128.13 (C300 + C500), 127.71
(C8), 126.44 (C10), 121.44 (C400), 115.52 (C30 + C50),
114.93 (C6), 110.68 (C10), 69.60 (CH2); MS (ES+) m/z
369.0 [M+1, 98%], HRMS (ES+) (M+H)+ C23H21N4O+
requires 369.1710, found 369.1712.
7.8. 4-(Benzyloxy)benzaldehyde (22)
A solution of 4-hydroxybenzaldehyde 21 (5 g, 41 mmol,
1 equiv), anhydrous potassium carbonate (17 g, 123
mmol, 3 equiv), benzyl bromide (21.04 g, 14.6 mL,
123 mmol, 3 equiv) in 2-butanone (50 mL) were stirred
under reflux for 24 h. Then the reaction mixture was ex-
tracted with ethyl acetate and washed with water. The
organic layer dried by Na2SO4, evaporated in lowpres-
sure and purified by flash chromatography. A yellowso-
lid was obtained (6.52 g, 75%, mp = 60 ꢁC); Rf = 0.4
(10% EtOAC/hexane); 1H NMR (CDCl3) d 9.95 (s,
1H, CHO), 7.9 (d, 2H, J 8.6 Hz, H2 + H6), 7.49–7.47
(m, 5H, H20–H60), 7.15 (d, 2H, J 8.8 Hz, H3 + H5),
5.21 (s, 2H, CH2); 13C NMR (CDCl3) d 191.21
(CHO), 164.17 (C4), 136.38 (C10), 132.44 (C2 + C6),
130.58 (C1), 129.18 (C20 + C60), 128.78 (C40), 127.93
(C30 + C50), 115.59 (C3 + C5), 70.71 (CH2); MS (CI+)
m/z 213.1 [M+1, 100%].
7.11. 4[2(2,4-Diamino-6-quinazolinyl)ethyl]phenol (5).
{Alternate method}
Compound 12 (300 mg, 0.82 mmol) and palladium
hydroxide (300 mg) stirred in dry DMF (20 mL) under
H2 gas for 24 h at room temperature. The reaction mix-
ture filtrate on Celite, concentrated in vacuum and puri-
fied by column chromatography to get the product
(160 mg, 70%, mp = 278 ꢁC).
7.12. 6[4-(Ethoxy)phenethyl]-2,4-quinazolinediamine (6)
Compound 5 (130 mg, 0.46 mmol, 1 equiv), 1-iodo-
ethane (217 mg, 0.12 mL, 1.39 mmol, 3 equiv) and
potassium carbonate (160 mg, 1.16 mmol, 2.5 equiv) in
ethanol (10 mL) refluxed for 24 h. Solvent removed
under reduced pressure and purified by column chroma-
tography to get the product (80 mg, 56%, mp = 230 ꢁC);
7.9. 1-Benzyloxy-4-vinylbenzene (23)
A solution of 22 (3.5 g, 16.5 mmol, 1 equiv) in dioxane
(20 mL) was added to a solution of methyltriphenyl-
phosphonium bromide (8.85 g, 24.8 mmol, 1.5 equiv),
anhydrous potassium carbonate (3.43 g, 24.8 mmol,
1.5 equiv) in dioxane (30 mL) and water (2 mL). The
reaction mixture was stirred under reflux for 48 h. The
crude product was extracted with DCM and washed
with water. The organic layer dried by Na2SO4, concen-
trated in vacuum and purified by flash chromatography
to get a white solid (2.6 g, 75%, mp = 65 ꢁC); Rf = 0.8
1
Rf = 0.2 (25% MeOH/CHCl3); H NMR (DMSO-d6) d
7.9 (s, 1H, H5), 7.4 (d, 1H, J 8.23 Hz, H7), 7.2 (d, 1H,
J 8.23 Hz, H8), 7.1 (d, 2H, J 8.6 Hz, H20 + H60), 6.8
(d, 2H, J 8.6 Hz, H30 + H50), 3.98 (q, 2H, J 6.9 Hz,
H100), 2.85 (s, 2H, a), 2.5 (s, 2H, b), 1.3 (t, 3H, J
6.9 Hz, CH3); 13C NMR (DMSO-d6)
d
162.74
(C40 + C9), 159.74 (C2), 157.06 (C4), 134.24
(C5 + C6), 133.59 (C10), 129.60 (C20 + C60), 123.18
(C7), 122.90 (C8), 114.52 (C30 + C50), 110.30 (C10),
63.17 (C100), 37.45 (a), 36.44 (b), 15.07 (CH3); MS
(ES+) m/z 309.1 [M+1 = 100%], HRMS (ES+) (M+H)+
C18H21N4O+ requires 309.1710, found 309.1722.
1
(10% EtOAC/hexane); H NMR (CDCl3) d 7.46–7.35
(m, 7H, H3 + H5, H20–H60), 6.96 (d, 2H, J 8.8 Hz,
H2 + H6), 6.69 (q, 1H, J 10.98 Hz, J 6.59 Hz, @CH),

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
2647
7.13. 6[4-(Butyloxy)phenethyl]-2,4-quinazolinediamine (7)
C20 + C60), 123.32 (C9), 120.92 (C5), 114.58
(C30 + C50), 67.65 (C100), 38.31 (a), 37.31 (b), 31.59
(C600), 29.10 (C400 + C500), 29.02 (C200), 25.90 (C300),
22.44 (C700), 14.31 (CH3); MS (CI+) m/z 393.4
[M+1 = 100%], HRMS (ES+) (M+H)+ C24H33N4O+ re-
quires 393.2649, found 393.2653.
Compound 5 (130 mg, 0.46 mmol, 1 equiv), 1-iodobu-
tane (256 mg, 0.16 mL, 1.39 mmol, 3 equiv) and potas-
sium carbonate (160 mg, 1.16 mmol, 2.5 equiv) in
ethanol (10 mL) refluxed for 24 h. Solvent removed
under reduced pressure and purified by column chro-
matography to get the product (100 mg, 64%,
7.16. 6[4-(Decyloxy)phenethyl]-2,4-quinazolinediamine
(10)
1
mp = 180 ꢁC); Rf = 0.4 (30% MeOH/CHCl3); H NMR
(DMSO-d6) d 7.85 (s, 1H, H5), 7.4 (d, 1H, J 8.4 Hz,
H7), 7.2–7.1 (m, 3H, H8 + H20 + H60), 6.85 (d, 2H, J
8.4 Hz, H30 + H50), 3.9 (t, 2H, J 6.5 Hz, H100), 2.85 (s,
2H, a), 2.5 (s, 2H, b), 1.7 (m, 2H, H200), 1.4 (m, 2H,
H300), 0.9 (t, 3H, J 7.3 Hz, CH3); 13C NMR (DMSO-
d6) d 162.65 (C9), 160.49 (C2), 157.22 (C4 + C40),
133.85 (C5), 133.65 (C6 + C10), 129.59 (C20 + C60),
124.09 (C7), 122.70 (C8), 114.56 (C30 + C50), 110.43
(C10), 67.34 (C100), 37.52 (a), 36.51 (b), 31.17 (C200),
19.12 (C300), 14.07 (CH3); MS (ES+) m/z 337.2
[M+1 = 100%], HRMS (ES+) (M+H)+ C20H25N4O+ re-
quires 337.2023, found 337.2027.
Compound 5 (100 mg, 0.36 mmol, 1 equiv), 1-iodode-
cane (290 mg, 0.23 mL, 1.07 mmol, 3 equiv) and potas-
sium carbonate (125 mg, 0.89 mmol, 2.5 equiv) in
ethanol (10 mL) refluxed for 24 h. Solvent removed un-
der reduced pressure and purified by column chroma-
tography to get the product (67 mg, 45%, mp =
300 ꢁC); Rf = 0.45 (25% MeOH/CHCl3); 1H NMR
(DMSO-d6) d 8 (s, 1H, H5), 7.5 (d, 1H, J 8.23 Hz,
H7), 7.2 (d, 1H, J 8.05 Hz, H8), 7.05 (d, 2H, J 8.6 Hz,
H20 + H60), 6.8 (d, 2H, J 8.6 Hz, H30 + H50), 3.85 (t,
2H, J 6.4 Hz, H100), 2.8 (s, 2H, a), 2.45 (s, 2H, b), 1.55
(m, 2H, H200), 1.4–1.1 (m, 14H, H300–H900), 0.8 (t, 3H,
J 6.4 Hz, CH3); 13C NMR (DMSO-d6) d 163.19 (C9),
157.29 (C2), 156 (C4), 140.74 (C40), 137.2 (C5), 135.81
(C6), 133.21 (C10), 129.6 (C20 + C60), 123.75 (C7),
118.68 (C8), 114.6 (C30 + C50), 109.71 (C10), 67.65
(C100), 37.19 (a), 36.13 (b), 31.65 (C800), 29.36 (C600),
29.30 (C700), 29.13 (C300), 29.08 (C500), 29.05 (C900),
25.89 (C400), 22.45 (C200), 14.32 (CH3); MS (ES+) m/z
421.1 (M+1, 22%), HRMS (ES+) (M+H)+
C26H37N4O+ requires 421.2962, found 421.2974.
7.14. 6[4-(Hexyloxy)phenethyl]-2,4-quinazolinediamine
(8)
Compound 5 (100 mg, 0.36 mmol, 1 equiv), 1-iodohex-
ane (227 mg, 0.16 mL, 1.07 mmol, 3 equiv) and potas-
sium carbonate (125 mg, 0.89 mmol, 2.5 equiv) in
ethanol (10 mL) refluxed for 24 h. Solvent removed
under reduced pressure and purified by column chroma-
tography to get the product (98 mg, 75%, mp = 175 ꢁC);
1
Rf = 0.4 (30% MeOH/CHCl3); H NMR (DMSO-d6) d
7.9 (s, 1H, H5), 7.35 (d, 1H, J 8.23 Hz, H7), 7.2–7.1
(m, 3H, H8 + H20 + H60), 6.8 (d, 2H, J 8.4 Hz,
H30 + H50), 3.95 (t, 2H, J 6.4 Hz, H100), 2.85 (s, 2H,
a), 2.5 (s, 2H, b), 1.7 (m, 2H, H200), 1.5–1.3 (m, 6H,
H300–H500), 0.9 (t, 3H, J 6.6 Hz, CH3); 13C NMR
(DMSO-d6) d 162.66 (C4), 160.35 (C9), 157.22 (C2 +
C40), 133.90 (C5), 133.75 (C7), 133.64 (C10 + C6),
129.59 (C20 + C60), 123.94 (C8), 122.74 (C10), 114.56
(C30 + C50), 67.64 (C100), 37.51 (a), 36.50 (b), 31.38
(C400), 29.07 (C200), 25.58 (C300), 22.44 (C500), 14.27
(CH3); MS (ES+) m/z 365.2 [M+1 = 100%], HRMS
(ES+) (M+H)+ C22H29N4O+ requires 365.2336, found
365.2348.
7.17. Modelling studies
Modelling was performed on Silicon Graphics O2
workstations using Sybyl 6.7 and GRID19. For dock-
ing, water molecules were removed from the protein.
For FlexX the active site was defined as all residues
within a radius of 6.5 A from the folate ligand (in
˚
the case of the human enzyme) or 6.5 A from the metho-
trexate ligand (in case of the L. major enzyme). The
compounds for docking were drawn in Sybyl. Before
docking, the compounds were subjected to energy min-
imisation using the Powell method using Sybyl with a
gradient cut-off of 0.01 kcal/mol or a limit of 1000
iterations.
˚
7.15. 6[4-(Octyloxy)phenethyl]-2,4-quinazolinediamine
(9)
7.18. Enzyme assays
Compound 5 (100 mg, 0.36 mmol, 1 equiv), 1-iodo-
octane (260 mg, 0.2 mL, 1.07 mmol, 3 equiv) and potas-
sium carbonate (125 mg, 0.89 mmol, 2.5 equiv) in
ethanol (10 mL) refluxed for 24 h. Solvent removed
under reduced pressure and purified by column chroma-
tography to get the product (98 mg, 70%, mp = 230 ꢁC);
DHFR catalyses the NADPH dependant reduction of
dihydrofolate to tetrahydrofolate. Assays in the absence
and presence of inhibitor were performed spectrophoto-
metrically by measuring the rate of decrease in NADPH
absorbance at 340 nm at 25 ꢁC. The inhibition constant
(Ki) was calculated from Km and IC50 values assuming
competitive inhibition. Standard reactions contained
dihydrofolate (DHF) 30 lM, NADPH 100 lM, bovine
serum albumin (BSA) (1 mg/mL) and enzyme in
50 mM TES pH 7.0, 75 mM b-mercaptoethanol, 1 mM
EDTA. Enzyme concentrations used were 2.72 nM for
L. major DHFR and 8.5 nM for the human enzyme.
Routinely, a range of at least five different inhibitor con-
centrations was tested for IC50 calculations. Absorbance
1
Rf = 0.4 (30% MeOH/CHCl3); H NMR (DMSO-d6) d
7.95 (s, 1H, H5), 7.50 (d, 1H, J 8.23 Hz, H7), 7.25 (d,
1H, J 8.23 Hz, H8), 7.15 (d, 2H, J 8.4 Hz, H20 + H60),
6.8 (d, 2H, J 8.4 Hz, H30 + H50), 3.90 (t, 2H, H100), 2.9
(s, 2H, a), 2.5 (s, 2H, b), 1.7 (m, 2H, H200), 1.45–1.25
(m, 10H, H300–H700), 0.9 (t, 3H, J 6.68 Hz, CH3); 13C
NMR (DMSO-d6) d 162.97 (C2 + C10), 157.25 (C4 +
C40), 131.50 (C6), 133.39 (C7 + C10), 129.60 (C8 +

2648
S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
was recorded in a Hewlett–Packard model 8543 spectro-
photometer. Dihydrofolate was prepared from folic acid
by reduction with sodium dithionite according to the
method of Futterman.23 The following Km values were
used for calculation of Ki values: L. major DHFR,
1.3 lM; human DHFR, 0.64 lM. NADPH and BSA
were purchased from SIGMA. Once prepared, dihydro-
folate was stored at ꢀ80 ꢁC until use. The concentration
of dihydrofolate was measured spectrophotometrically
at 282 nm.
used covering a range from 30 to 0.041 lg/mL. Each
drug was tested in duplicate and each assay was repeated
at least once. After 72 h of incubation the plates were in-
spected under an inverted microscope to assure growth
of the controls and sterile conditions. 10 lL of Alamar
Blue (12.5 mg resazurin dissolved in 100 mL distilled
water) were then added to each well and the plates incu-
bated for another 2 h. Then the plates were read with a
Spectramax Gemini XS microplate fluorometer (Mole-
cular Devices Cooperation, Sunnyvale, CA, USA) using
an excitation wave length of 536 nm and an emission
wave length of 588 nm. Data were analysed using the
software Softmax Pro (Molecular Devices Cooperation,
Sunnyvale, CA, USA). Decrease of fluorescence (= inhi-
bition) was expressed as percentage of the fluorescence
of control cultures and plotted against the drug concen-
trations. From the sigmoidal inhibition curves the IC50
values were calculated.
7.19. Anti-parasite assays
7.19.1. Trypanosoma b. rhodesiense. Minimum essential
medium (50 lL) supplemented according to Baltz
et al.24 with 2-mercaptoethanol and 15% heat-inacti-
vated horse serum was added to each well of a 96-well
microtitre plate. Serial drug dilutions were prepared
covering a range from 90 to 0.123 lg/mL. Then 104
bloodstream forms of T. b. rhodesiense STIB 900 in
50 lL were added to each well and the plate incubated
at 37 ꢁC under a 5% CO2 atmosphere for 72 h. Alamar
Blue (10 lL, 12.5 mg resazurin dissolved in 100 mL dis-
tilled water) were then added to each well and incuba-
tion continued for a further 2–4 h. The plate was then
read in a Spectramax Gemini XS microplate fluorometer
(Molecular Devices Cooperation, Sunnyvale, CA, USA)
using an excitation wavelength of 536 nm and emission
wavelength of 588 nm.25 Fluorescence development
was measured and expressed as percentage of the con-
trol. Data were transferred into the graphic programme
Softmax Pro (Molecular Devices), which calculated IC50
values. Cytotoxicity was assessed using the same assay
and rat skeletal myoblasts (L-6 cells).
Acknowledgements
We would like to acknowledge the Welsh School of
Pharmacy, the University of Shiraz, and the FSI Net-
work RICET/C03 for funding. This investigation re-
ceived financial support from the UNDP/World Bank/
WHO Special Programme for Research and Training
in Tropical Diseases (TDR) for the drug screening part.
The EPSRC National Mass Spectrometry service centre
Swansea is acknowledged for accurate mass
spectrometry.
References and notes
1. Gilbert, I. H. Biochim. Biophys. Acta 2002, 1587, 249.
2. 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.
3. Zuccotto, F.; Brun, R.; Pacanowska, D. G.; Perez, L. M.
R.; Gilbert, I. H. Bioorg. Med. Chem. Lett. 1999, 9, 1463.
4. Zuccotto, F.; Martin, A. C. R.; Laskowski, R. A.;
Thornton, J. M.; Gilbert, I. H. J. Comput.-Aided Mol.
Des. 1998, 12, 241.
5. 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.
6. Chowdhury, S. F.; Di Lucrezia, R.; Guerrero, R. H.;
Brun, R.; Goodman, J.; Ruiz-Perez, L. M.; Pacanowska,
D. G.; Gilbert, I. H. Bioorg. Med. Chem. Lett. 2001, 11,
977.
7.19.2. Trypanosoma cruzi. Rat skeletal myoblasts (L-6
cells) were seeded in 96-well microtitre plates at
2000 cells/well in 100 lL RPMI 1640 medium with
10% FBS and 2 mM L-glutamine. After 24 h the medium
was removed and replaced by 100 lL per well containing
5000 trypomastigote forms of T. cruzi Tulahuen strain
C2C4 containing the b-galactosidase (Lac Z) gene. After
48 h the medium was removed from the wells and re-
placed by 100 lL fresh medium with or without a serial
drug dilution. Seven 3-fold dilutions were used covering
a range from 90 to 0.123 lg/mL. Each drug was tested in
duplicate. After 96 h of incubation the plates were in-
spected under an inverted microscope to assure growth
of the controls and sterility. Then the substrate
CPRG/Nonidet (50 lL) was added to all wells. A colour
reaction developed within 2–6 h and could be read
photometrically at 540 nm. Data were transferred into
the graphic programme Softmax Pro (Molecular De-
vices), which calculated IC50 values.
7. Chowdhury, S. F.; Guerrero, R. H.; Brun, R.; Ruiz-Perez,
L. M.; Pacanowska, D. G.; Gilbert, I. H. J. Enzym. Inhib.
Med. Chem. 2002, 17, 293.
8. Pez, D.; Leal, I.; Zuccotto, F.; Boussard, C.; Brun, R.;
Croft, S. L.; Yardley, V.; Ruiz-Perez, L. M.; Gonzalez
Pacanowska, D.; Gilbert, I. H. Bioorg. Med. Chem. 2003,
11, 4693.
9. Sirawaraporn, W.; Sertsrivanich, R.; Booth, R. G.;
Hansch, C.; Neal, R. A.; Santi, D. V. Mol. Biochem.
Parasitol. 1988, 31, 79.
10. Goodford, P. J. J. Med. Chem. 1985, 28, 849.
11. Berman, J. D.; King, M.; Edwards, N. Antimicrob. Agents
Chemother. 1989, 33, 1860.
7.19.3. Leishmania donovani. Amastigotes of L. donovani
strain MHOM/ET/67/L82 were grown in axenic culture
at 37 ꢁC in SM medium26 at pH 5.4 supplemented with
10% heat-inactivated foetal bovine serum under an
atmosphere of 5% CO2 in air. Hundred microlitres of
culture medium with 105 amastigotes from axenic cul-
ture with or without a serial drug dilution were seeded
in 96-well microtitre plates. Seven 3-fold dilutions were

S. Khabnadideh et al. / Bioorg. Med. Chem. 13 (2005) 2637–2649
2649
12. Davoll, J.; Johnson, A. M.; Davis, H. J.; Bird, O. D.;
Clarke, J.; Elslager, E. F. J. Med. Chem. 1972, 15, 812.
13. Elslager, E. F.; Bird, O. D.; Clarke, J.; Perricone, S. C.;
Worth, D. F. J. Med. Chem. 1972, 15, 1138.
19. Harris, N. V.; Smith, C.; Bowden, K. Eur. J. Med. Chem.
1992, 27, 7.
20. Sonogashira, K.; Tohda, Y.; Hagihara, N. Tetrahedron
Lett. 1975, 16, 4467.
14. Thompson, P. E.; Bayles, A.; Olszewski, B. Exp. Parasitol.
1969, 25, 32.
21. Graffner-Nordberg, M.; Kolmodin, K.; Aqvist, J.; Que-
ener, S. F.; Hallberg, A. J. Med. Chem. 2001, 44, 2391.
22. Cativiela, C.; Serrano, J. L.; Zurbano, M. M. J. Org.
Chem. 1995, 60, 3074.
23. Futterman, S. J. Biol. Chem. 1957, 28, 1031.
24. Baltz, T.; Baltz, D.; Giroud, C.; Crockett, J. EMBO J.
1985, 4, 1273.
15. Pez, D.; Leal, I.; Zuccotto, F.; Boussard, C.; Brun, R.;
Croft, S. L.; Yardley, V.; Perez, L. M. R.; Pacanowska, D.
G.; Gilbert, I. H. Bioorg. Med. Chem. 2003, 11, 4693.
16. Walters, W. P.; Murcko, M. A. Adv. Drug Deliv. Rev.
2002, 54, 255.
17. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney,
P. J. Adv. Drug Deliv. Rev. 1997, 23, 3.
18. Harris, N. V.; Smith, C.; Bowden, K. Synlett 1990, 577.
25. Raz, B.; Iten, M.; Grether-Buhler, Y.; Kaminski, R.;
Brun, R. Acta Trop. 1997, 68, 139.
26. Cunningham, I. J. Protozool. 1977, 24, 325.