Targeting species specific amino acid residues: Design, synthesis and biological evaluation of 6-substituted pyrrolo[2,3-d]pyrimidines as dihydrofolate reductase inhibitors and potential anti-opportunistic infection agents

Article abstract of DOI:10.1016/j.bmc.2018.04.032

To combine the potency of trimetrexate (TMQ) or piritrexim (PTX) with the species selectivity of trimethoprim (TMP), target based design was carried out with the X-ray crystal structure of human dihydrofolate reductase (hDHFR) and the homology model of Pneumocystis jirovecii DHFR (pjDHFR). Using variation of amino acids such as Met33/Phe31 (in pjDHFR/hDHFR) that affect the binding of inhibitors due to their distinct positive or negative steric effect at the active binding site of the inhibitor, we designed a series of substituted-pyrrolo[2,3-d]pyrimidines. The best analogs displayed better potency (IC₅₀) than PTX and high selectivity for pjDHFR versus hDHFR, with 4 exhibiting a selectivity for pjDHFR of 24-fold.

Full text of DOI:10.1016/j.bmc.2018.04.032

Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Targeting species specific amino acid residues: Design, synthesis and
biological evaluation of 6-substituted pyrrolo[2,3-d]pyrimidines as
dihydrofolate reductase inhibitors and potential anti-opportunistic
infection agents
Khushbu Shah a, Xin Lin a, Sherry F. Queener b, Vivian Cody c, Jim Pace c, Aleem Gangjee a,
⇑
a Division of Medicinal Chemistry, Graduate School of Pharmaceutical Sciences, Duquesne University, Pittsburgh, PA 15282, United States
b Department of Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, IN 46202, United States
c Hauptman-Woodward Medical Research Institute, 700 Ellicott Street, Buffalo, NY 14203, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
To combine the potency of trimetrexate (TMQ) or piritrexim (PTX) with the species selectivity of
trimethoprim (TMP), target based design was carried out with the X-ray crystal structure of human dihy-
drofolate reductase (hDHFR) and the homology model of Pneumocystis jirovecii DHFR (pjDHFR). Using
variation of amino acids such as Met33/Phe31 (in pjDHFR/hDHFR) that affect the binding of inhibitors
due to their distinct positive or negative steric effect at the active binding site of the inhibitor, we
designed a series of substituted-pyrrolo[2,3-d]pyrimidines. The best analogs displayed better potency
(IC50) than PTX and high selectivity for pjDHFR versus hDHFR, with 4 exhibiting a selectivity for
pjDHFR of 24-fold.
Received 13 February 2018
Revised 4 April 2018
Accepted 14 April 2018
Available online xxxx
Keywords:
DHFR inhibitors
pjDHFR
Ó 2018 Elsevier Ltd. All rights reserved.
hDHFR
Pneumocystis pneumonia
Opportunistic infections
Pyrrolo[2,3-d]pyrimidines
1. Introduction
strains, late diagnosis of HIV and the rise in the number of cases
in developing countries.6,7 Thus PCP continues to be a significant
Pneumocystis jirovecii (pj) is a fungus that infects the lungs of a
majority of humans around the world. However, the immune sys-
tem in healthy individuals keeps the infection under control. In
immunocompromised patients, pj infection causes Pneumocystis
pneumonia (PCP).1,2 PCP can be fatal for patients with HIV/AIDS
(most common), patients undergoing chemotherapy for cancer,
patients on immunosuppressive medications, patients undergoing
organ or bone-marrow transplantation or those who are malnour-
ished.3,4 PCP presents itself when the patients’ CD4 count is below
200 cells/mm3,5 Although PCP prophylaxis and antiretroviral ther-
apy (ART) have changed the face of the HIV/AIDS epidemic, the
incidence of HIV cases persist due to non-adherence to the medica-
tion, toxicity to the medications, emergence of drug resistant HIV
public health concern. In the US, 9% of the hospitalized HIV/AIDS
and 1% of organ transplant patients develop PCP infection.8 In these
patients, the mortality rate is from 5 to 40% while being treated for
PCP and approaches 100% if left untreated.8
Both the prophylaxis and treatment for PCP involves the combi-
nation of trimethoprim (TMP)-sulfamethoxazole (SMX) (co-tri-
moxazole).9,10 TMP (Fig. 1) is a selective, but weak inhibitor of
dihydrofolate reductase (DHFR), the enzyme necessary for the
reduction of dihydrofolate to tetrahydrofolate,11 while SMX is an
inhibitor of the dihydropteroate synthase (DHPS), the enzyme nec-
essary for the synthesis of folates in fungi.12 The low activity of
TMP against DHFR is augmented by SMX, in the treatment regi-
men. The efficacy, low cost and activity against a variety of infec-
tions has propelled co-trimoxazole to be used indiscriminately.
Due to the rampant use, mutations in the DHPS locus of P. jirovecii
(the fungal species that causes PCP in humans) encoding DHPS
have been documented as the cause of TMP/SMX resistant strains
of PCP.12–14 Various studies have also reported mutations
discovered in pjDHFR after treatment or prophylaxis using DHFR
Abbreviations: Pj, Pneumocystis jirovecii; PCP, Pneumocystis pneumonia; ART,
antiretroviral therapy; TMP, trimethoprim; SMX, sulfamethoxazole; DHFR, dihy-
drofolate reductase; DHPS, dihydropteroate synthase; PTX, piritrexim; TMQ,
trimetrexate; PC, Pneumocystis carinii.
⇑
Corresponding author.
E-mail address: gangjee@duq.edu (A. Gangjee).
https://doi.org/10.1016/j.bmc.2018.04.032
0968-0896/Ó 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

2
K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
Fig. 1. DHFR inhibitors for treatment for PCP.
inhibitors.15–19 Treatment failure and discontinuation of co-tri-
moxazole occurs in several cases due to such resistant strains or
toxicity/allergy caused by SMX.20–24 When treatment fails with
TMP/SMX, the second-line treatment in mild to moderate PCP is
TMP-dapsone or clindamycin-primaquine, which also leads to
low efficacy and often lethal side-effects.9,25–27 Piritrexim (PTX)
and trimetrexate (TMQ) are potent, but non-selective inhibitors
of pjDHFR, which cause dose-limiting toxicities and have been dis-
continued.9,28,29 For patients that do not respond to first line treat-
ment as well the inevitable appearance of resistance, new drugs for
the treatment of PCP are critically needed.
an animal model for in vivo evaluation of the synthesized com-
pounds. Due to this drawback, isolation and use of pjDHFR enzyme
is currently the only direct indicator that a compound could be
effective (or ineffective) in the treatment of PCP infection in
humans.
2. Synthesis
Synthesis of 1–18 utilized a modification of the literature
method.34 To a solution of hydroxyacetone 19 and malononitrile
in ethanol, triethylamine was added and stirred overnight under
argon to afford 20 (Scheme 1). The cyclisation of 20 without purifi-
cation was carried out with guanidine and sodium methoxide at
reflux to obtain 21 (10–35%). To a solution of iodine and the appro-
priate thiophenol (2:1 ethanol:water), 21 was added and main-
tained at reflux to afford 1 and 7–12. The pyrrole nitrogen on 1
and 7–12 was methylated using sodium hydride and methyl iodide
in DMF to afford 2 and 13–18. For the N7-alkylated series, 1 was
alkylated using appropriate alkyl halides to afford 3–6 (Scheme 2).
Synthesis and characterization of compounds 1, 7, 8, and 10 has
been presented previously by Gangjee et al.34
One of the most efficient strategies to treat PCP infection is to
target P. jirovecii DHFR (pjDHFR).30 DHFR catalyzes the reduction
of 7,8-dihydrofolate to the 5,6,7,8-tetrahydrofolate. Inhibition of
DHFR interferes with folate cofactor requiring transformations
including thymidylate and purine biosynthesis and results in inhi-
bition of DNA synthesis.11 This inhibition causes a disruption in
DNA, RNA and protein synthesis of the organism and eventually
leads to death of the fungus. Pneumocystis infection is host spe-
cific. Pneumocystis carinii, however, is a distinct species that infects
rats, different from P. jirovecii, responsible for human infections.
The amino acid sequence of the DHFR of Pneumocystis carinii
(pcDHFR) differs by 38% when compared to the DHFR of Pneumo-
cystis jirovecii (pjDHFR).31 Hence, drug’s activity against the surro-
gate pcDHFR in-vitro may not translate into activity in the
treatment of PCP infection in humans caused by P. jirovecii.
We have recently isolated pjDHFR31 and used it to evaluate clin-
ically used agents such as TMP, PTX and novel DHFR inhibitors.32
These studies demonstrated that the inhibition of human(h)DHFR
compared with pjDHFR allows the calculation of a selectivity ratio
(IC50 hDHFR/IC50 pjDHFR) that provides a measure of the selective
inhibition of the agent for pjDHFR over hDHFR. Compounds, such
as PTX and TMQ, though highly potent, show poor selectivity for
pjDHFR over hDHFR and are much too toxic in vivo; this lack of
selectivity is responsible for their discontinuation for the treat-
ment of infections caused by P. jirovecii. The selectivity of TMP
however, for pjDHFR over hDHFR is 266-fold and contributes to
its clinical success in PCP treatment. Besides the selectivity for
pjDHFR another aspect that is highly desirable in an agent is
potency for pjDHFR. TMP has a low potency as an inhibitor of
pjDHFR and must be used with SMX for clinical efficacy. Our
long-term goal is to provide analogs with excellent potency along
with high selectivity for pjDHFR. Such agents could be used alone
as well as with sulfonamides and other drugs for PCP infections
in humans.
3. Design and docking studies
We published the X-ray crystal structure of hDHFR and pcDHFR
with several pyrido[2,3-d]pyrimidines.33 In addition, using the
published crystal structures for pcDHFR,33 a homology model of
pjDHFR32 was refined to include the cofactor, nicotinamide ade-
nine dinucleotide phosphate (NADPH). This refined pjDHFR homol-
ogy model was utilized to evaluate the docking of the proposed
compounds 1–18. The isolation of pjDHFR along with the develop-
ment of the homology model for pjDHFR provided insight regard-
ing the amino-acid sequence differences between active site of
pjDHFR and pcDHFR, as well as that of hDHFR. The superimposition
of the active site of pcDHFR and pjDHFR (Fig. 2) displays the amino
acid differences present in the active sites of the two enzymes. The
active sites of pcDHFR has Glu32, Ile33, Ile65, and Phe69, which can
affect the binding of the ligands designed. The active site of pjDHFR
at the same positions in the active site has Asp32, Met33, Leu65,
and Ser69. These amino acids differ in their size and electrostatics
and thus would significantly influence the binding of the designed
compounds considerably. These amino acid differences highlight
the futility of designing and evaluating activity against the surro-
gate pcDHFR as inhibitors of pjDHFR.
Gangjee et al.35 reported 6-substituted pyrrolo[2,3-d]pyrimidi-
nes as inhibitors of pcDHFR. Re-evaluation of compound 1 (Table 1)
from this previous study35 in pjDHFR and hDHFR enzymes, indi-
cated a moderate inhibitory potency for pjDHFR and marginal
selectivity for pjDHFR over hDHFR. The reevaluation of 1 in isolated
pjDHFR provided a lead analog for optimization of both potency
and selectivity. We recognized that 1 was overall not as selective
or potent as TMP; however, it was an improvement in its selectiv-
ity over PTX for pjDHFR and a good starting structure for improve-
ment in both potency and selectivity.
Rational design of pjDHFR inhibitors is hampered due to a lack
of crystal structure information for pjDHFR. However, homology
models can be used with refinement to model pjDHFR in the
absence of crystal structures.32 Thus along with known hDHFR X-
ray crystal structures,33 pjDHFR homology models can be used to
design and predict potent and selective pjDHFR inhibitors. Another
significant impediment in the drug discovery of inhibitors of
pjDHFR is the inability to grow the organism outside the human
lung and hence to develop a tissue culture for in vitro studies or
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
3
Scheme 1. a) Malononitrile, TEA, EtOH, rt, 12 h; b) NaOMe, guanidine HCl, EtOH, reflux, 24 h; c) thiophenol, I2, 2:1 EtOH:H2O, reflux, 24 h; d) CH3I, NaH, DMF, rt, 0.5–2 h.
Scheme 2. R-Br, NaH, DMF, rt, 0.5–2 h.
Table 1
Inhibition concentrations (IC50) against pjDHFR and hDHFR and selectivity ratios.
#
pjDHFR (nM)
hDHFR (nM)
Selectivity ratio [hDHFR/pjDHFR]
1
TMP
PTX
213
92
41
970
24,500
2
5
266
0.05
molecular modeling program LeadIt 2.1.636 and the parameters
specified in the Experimental Section. Multiple low energy confor-
mations were obtained on docking 1 in the active site of hDHFR
and pjDHFR. As a representative example, Fig. 4a shows the best
docked conformation of 1 in the pjDHFR homology model. It dis-
plays a bi-dentate ionic bond between protonated N1 and 2-NH2
of 1 with Asp32. This interaction is most commonly observed in
ligands in DHFR crystal structures.37 The 4-NH2 moiety of 1 forms
hydrogen bonds with the backbones of Ile10 and the pyrrolo[2,3-d]
pyrimidine scaffold is stabilized by pi-stacking interaction with
Phe36. The 30-methoxyphenyl moiety of 1 is oriented in the pocket
formed by Leu 25 (not displayed), Met33, Ser64 and Leu65. The 30-
methoxyphenyl oxygen forms a hydrogen bonding interaction
with Ser64 in the pocket. This docked pose generated a docking
score of ꢀ34 kJ/mol. Fig. 4b displays the best docked conformation
of 1 in hDHFR crystal structure (PDB: 4QJC, 1.61 Å).33 It also exhi-
bits a bi-dentate ionic interaction of the protonated N1 and 2-NH2
with Glu30. The 4-NH2 displayed a hydrogen bonding interaction
of with the backbone of Val8 and Val15. The 30-methoxyphenyl
moiety is oriented in the pocket formed by Leu22 (not displayed),
Phe31 and Ser59. The scaffold is similarly stabilized by pi-stacking
interactions with Phe34. This docked pose generated a docking
score of ꢀ29 kJ/mol in the hDHFR crystal structure. The docking
score comparison between pjDHFR and hDHFR shows a difference
Fig. 2. Superimposition of active sites of pcDHFR and pjDHFR. The blue ribbon and
amino acid residues represent the active site of pcDHFR (PDB: 4QJZ, 1.61 Å).33 The
pink ribbon and amino acid residues represent the homology model of pjDHFR
active site. The ligand N6-methyl-N6-(naphthalen-2-yl)pyrido[2,3-d]pyrimidine-
2,4,6-triamine (magenta) was co-crystallized with hDHFR.33
In order to determine the amino acid differences in the active
site of pjDHFR and hDHFR, the pjDHFR homology model sequence
was superimposed on the hDHFR X-ray crystal published with a
pyrido[2,3-d]pyrimidine ligand (Fig. 3).32,33 The active site of
hDHFR is partially composed of Glu30, Phe31, Asn64, and Val115.
Analogous to these, the active site of pjDHFR is partially composed
of the corresponding amino acids Asp32, Met33, Ser69, and Ile123.
The side chains of these amino acids are different in shape, size and
electronic properties, which allows the design of inhibitors with
selectivity and potency for pjDHFR over hDHFR.
Following the evaluation of 1 in pjDHFR and hDHFR, we con-
ducted docking studies of 1 in the pjDHFR homology model and
in the hDHFR crystal structure (PDB: 4QJC, 1.62 Å)33 using the
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

4
K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
with Met33 in the pjDHFR active site and an unfavorable steric
clash with the Phe31 in the hDHFR active site. To further validate
our hypothesis, the N7-methyl analog of 1, 2 was also docked, syn-
thesized, and evaluated.
The introduction of the N7-methyl moiety affords two signifi-
cant changes in the molecule. First, it increases the hydrophobic
interactions in both pjDHFR and hDHFR active sites. The second
change is the decrease in the number of low energy conformations
possible for 2 within 1 kcal/mol, compared to 1. This is a direct
consequence of the further restricted rotation of the 6-aryl moiety
due to the presence of the 5,7-dimethyl groups. The number of
conformations possible for 1 and 2 were calculated using Sybyl38
and were found to be 122 and 72, respectively. Thus, conforma-
tional restriction induced by the N7-methyl group could afford
the bioactive conformation or, at least, easier access to the bioac-
tive conformation of 2 in pjDHFR. These two attributes resulting
from the addition of the N7-methyl group could be responsible
for an increase in potency of 2 over 1. The docking studies of 2 in
the pjDHFR homology model and the hDHFR crystal structure
(PDB: 4QJC, 1.62 Å);33 displays the interactions as expected (Fig.
5a and b). The N7-methyl group is indeed oriented towards the
hydrophobic pocket in both pjDHFR and hDHFR active sites. The
docking scores of 2 in the pjDHFR homology model and the hDHFR
crystal structure were ꢀ36 kJ/mol and ꢀ25 kJ/mol respectively.
The difference in the docked scores between pjDHFR and hDHFR,
of 11 kJ/mol also predicts an increased selectivity for pjDHFR.
Evaluation of 2 in enzyme assays displayed increased potency
towards pjDHFR and a 8-fold selectivity in inhibition of pjDHFR
over hDHFR (Table 2). Compared to the activity of 1, the increase
in potency towards pjDHFR can be attributed, in part, to the
hydrophobic binding of the N7-methyl group in the pocket formed
by Met33 and Leu25 (not displayed) and the easier access to the
bound conformation, whereas the selectivity increase could be
due to a probable steric clash between the N7-methyl group of 2
with Phe31 in hDHFR thus making the binding of 2 less favorable
in hDHFR than 1. These evaluations (in vitro IC50) validate our
homology model and docking methods. Owing to a large size of
the pocket, the N7-methyl did not create a substantial increase
in potency and/or selectivity, as expected. Hence it was of interest
to synthesize and evaluate longer chain N7-substituetnts of 2, Ser-
ies 1 (Table 2, compounds 3–6). These longer alkyl chains at the N7
exhibited an increased potency for pjDHFR, but 4 afforded a selec-
tivity of 24-fold for pjDHFR over hDHFR, which was the highest
Fig. 3. Superimposition of active sites of hDHFR and pjDHFR. The amino acid
residues shown are the residues that are different in the active site of the
two species. The grey ribbon and amino acid residues co-crystallized with the
ligand, N6-methyl-N6-(3,4,5-trifluorophenyl)pyrido[2,3-d]pyrimidine-2,4,6-triamine
(magenta), represent the active site of hDHFR (PDB: 4QJC, 1.62 Å).33 The pink ribbon
and amino acid residues represent the homology model of pjDHFR active site.
of approximately 4 kJ/mol, suggesting selectivity of compound 1
for pjDHFR over hDHFR. This gain in selectivity could be a conse-
quence of the steric clash of the side chain phenyl ring of 1 with
Phe31 (in hDHFR) which is absent with Met33 (in pjDHFR). The
N7-H of 1 presents itself towards a hydrophobic pocket formed
by Phe31 in hDHFR and Met33 in pjDHFR (Fig. 4a). We reason that
this amino acid variation of Phe31 (in hDHFR) and Met33 (in
pjDHFR) in the active sites can be further exploited to obtain selec-
tivity for pjDHFR. Met33 is comparatively more flexible than Phe31
and hence can better accommodate larger inhibitors compared to
Phe31. Thus appropriate substitutions on the N7 of the pyrrolo
[2,3-d]pyrimidine scaffold of 1 could target this amino acid differ-
ence. The predicted distances of the N7 in 1 is approximately 4.89
Å from Met33 in the pjDHFR docked pose and approximately 3.57
Å from Phe31 in the hDHFR docked pose. Thus, a methyl substitu-
tion on N7 of 1 could create favorable hydrophobic interactions
Fig. 4. Docked pose of 1 (cyan) in (a) homology model of pjDHFR and (b) crystal structure of hDHFR (PDB: 4QJC, 1.62 Å).33
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
5
Fig. 5. Docked pose of 2 (cyan) in (a) homology model of pjDHFR and (b) crystal structure of hDHFR (PDB: 4QJC, 1.62 Å)33
Table 2
Inhibition concentrations (IC50) against pjDHFR and hDHFR and selectivity ratios.
#
R
pjDHFR (nM)
hDHFR (nM)
Selectivity ratio [hDHFR/pjDHFR]
1
2
3
4
5
6
H
CH3
213
160
35
84
74
970
1200
511
2046
579
1130
5
8
15
24
8
CH2CH3
CH2CH2CH3
CH(CH3)2
CH2CH2CH2CH3
73
15
observed for the pyrrolo[2,3-d]pyrimidine series. To structurally
explain these results, we performed docking of 4 in the pjDHFR
homology model (Fig. 6a) and the hDHFR crystal structure (Fig.
6b). In the homology model of pjDHFR, the terminus of the propyl
chain was at a distance of 3.54 Å from Met33, in pjDHFR. This pose
showed an excellent docking score of ꢀ36 kJ/mol, and the com-
pound displayed an inhibitory potency IC50 of 74 nM for pjDHFR.
The modelling of 4 in hDHFR showed the docked pose as depicted
in Fig. 6b and the docked score obtained was ꢀ24 kJ/mol. The low
docking score suggested a less than appropriate fit of 4 in active
site of hDHFR. The low score observed, also, reinforces the possibil-
ity of a steric clash of the propyl moiety with Phe31 (as observed in
Fig. 6c), which explains a decreased potency of 4 in the hDHFR. On
homologation to a N7-n-butyl and branching to i-propyl, the activ-
ity and selectivity against pjDHFR does not increase significantly,
indicating that the propyl chain is optimal at N7-position for this
series.
Our efforts at targeting the hydrophobic pocket containing
Met33 (in pjDHFR) and Phe31 (in hDHFR) led to 4. It was of interest
to study the effect of other amino acid differences within the active
site where the side chain aryl group binds. The amino acids at a
distance of 4.5 Å around the ligand were studied (Fig. 6). The
pocket in pjDHFR is composed of Met33, Ser64 and Leu65 and
Ile123 and Asp21, Phe31, Ser59 and Val115 in hDHFR. Thus, the
active sites have different electronics, shape and size which could
affect the binding properties of the pocket. To achieve potency
and selectivity by targeting these differing residues, the side-chain
aryl substituents with electron withdrawing, electron donating
and sterically bulky groups, as replacements for the 30-methoxy-
phenyl group were attempted (Table 3, Compounds 7–12). Evalua-
tion of 7–12 led to potent and selective compounds 9 and 12
(Table 3). The 2-napthyl and 4-trifluromethoxyphenyl substitu-
tions showed a 2-fold increase in potency and a 2-fold increase
in selectivity, compared to 1. The gain in potencies of 9 and 12 in
pjDHFR, compared to 1, could be due to productive shape comple-
mentarity of the side chain aryl group and the pocket formed by
Met33, Ser64 and Leu65. The gain in selectivity of 9 and 12 for
pjDHFR over hDHFR, compared to 2, could be due to the steric clash
of the bulkier side chain aryl moiety with Phe31 (in hDHFR), which
is absent with Met33 (in pjDHFR). The high probability of a steric
hindrance between the bulky side chain aryl group in these com-
pounds with Phe31 in hDHFR is evident in the docking studies of
9 in the crystal structure of hDHFR (PDB: 4QJC, 1.62 Å)33 in
Fig. 7. The Phe31 in the hDHFR active site limits the movement
of the side chain and forces a steric hindrance which decreases
the potency of the larger side chain aryl groups, for hDHFR.
Since N7-methylation of 1 afforded an increase in potency and
selectivity, we methylated the N7-position of 7–12 to afford 13–18
(Table 4). The N7-methlyation with varied side chain aryl group
did not afford an increase in potency or selectivity and 4 remained
the most selective compound in this series.
4. X-ray crystal structures (PDB accession numbers hDHFR-3
(5HT4); hDHFR-14 (5HT5) for compounds 3 and 14)
Structural data were measured for the ternary complexes of
NADPH and native human DHFR with inhibitors 3 (Table 2) and
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

6
K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
Fig. 6. (a) Docked pose of 4 (cyan) in the homology model of pjDHFR; (b) docked pose of 4 (cyan) in the crystal structure of hDHFR (PDB: 4QJC, 1.62 Å)33 and (c) space-filled
representation of Phe31 residue and N7-propyl group in the docked pose of 4 (cyan) in the crystal structure of hDHFR (PDB: 4QJC, 1.62 Å)33 to illustrate the high probability of
steric clash.
14 (Table 4), respectively, to validate the binding interactions of
these inhibitors in the active site of hDHFR (Fig. 8). Compound 3
was selected for its high selectivity and potency in pjDHFR. These
data reveal that the presence of the N7-ethyl group of 3 causes
the conformation of Phe31 to differ from that observed in the
hDHFR complex with 14; Phe31 adopts alternate positions with
partial occupancy. It is also interesting to note that the small shift
in the binding orientation of inhibitors 3 and 14 allows the 30-
methoxy and the 40-methoxy of 3 and 14 respectively to occupy
similar positions in the binding site.
and the 40-methoxy substituents occupy the same binding pocket.
In 3, the amino group of the side chain of Asn64 is within hydrogen
bonding distance to the 30-methoxy oxygen (2.9 Å) and the 40-
methoxy oxygen is within 3.4 Å of the Asn64 amino moiety in
14. The interactions of the N7-methyl substituent of 3 and N7-
ethyl substituent of 14 results in Phe31 having two alternate con-
formations. Analysis of the intermolecular interactions involving
the C5-methyl substituent of 3 and 14 shows hydrophobic contacts
(4.3 and 4.6 Å, respectively) with the C5 of Val115. The 4-NH2
group of the inhibitors 3 and 14 form a hydrogen bond with the
backbone carbonyl of Val115 (3.0 and 3.3 Å, respectively). Docking
studies of 3 and 14 in the hDHFR crystal structure (PDB ID: 4QJC)33
afforded poses which mimic the conformation obtained from the
The overall structures of hDHFR in complex with 3 and 14 are
similar to those reported for other hDHFR inhibitor com-
plexes.19,32,33 As observed in Fig. 8, for 3 and 14, the 30-methoxy
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
7
Table 3
Inhibition concentrations (IC50) against pjDHFR and hDHFR and selectivity ratios.
#
R
pjDHFR (nM)
hDHFR (nM)
Selectivity ratio [hDHFR/pjDHFR]
1
7
8
9
10
11
12
30-OCH3Ph
20-OCH3Ph
40-OCH3Ph
2-Naph
1-Naph
30,40-diFPh
40-OCF3Ph
213
177
252
101
167
240
81
970
624
5
4
6
12
7
1410
2100
1216
2318
811
10
10
5. Summary
The X-ray crystal structures of 3 and 14 in hDHFR validate our
hypothesis that bulk at the N7-position of the pyrrolo[2,3-d]pyrim-
idine scaffold results in a steric clash with Phe31. We have success-
fully designed, synthesized and evaluated novel series of analogs to
explore active site amino acid residue differences in hDHFR and
pjDHFR enzymes in our attempt to afford selective inhibitors of
pjDHFR over hDHFR. This effort led to several compounds (3–6
and 12) exhibiting potency greater than TMP (92 nM) and selectiv-
ity greater than PTX (0.05-fold). The docking studies and crystal
structures reveal the importance of targeting the differences in
amino acid residues in the active site of pjDHFR and hDHFR. These
predictions from the docking studies and the X-ray crystal studies
were corroborated by the biological evaluation results. Compound
4 afforded the best selectivity for pjDHFR over hDHFR (24 fold)
with a potency of 84 nM for pjDHFR. This suggested that the opti-
mum bulk at the N7-position of the pyrrolo[3,2-d]pyrimidine scaf-
fold, that can be tolerated for increased binding to pjDHFR active
site and causing steric clash with hDHFR active site, is equivalent
to a propyl group. Compound 4 maintained the potency equivalent
to PTX, but exhibited a 48-fold increase in selectivity for pjDHFR
over hDHFR. Utilizing the information provided in this study has
allowed the design and synthesis of potentially more potent and
selective compounds that are currently underway and will be the
subject of future publications.
Fig. 7. Docked pose of 9 (cyan) in the crystal structure of hDHFR (PDB ID: 4QJC).33
The Phe 31 residue is shown in a space fill view to illustrate the high probability of
steric clash with the side chain aryl group of 9.
6. Experimental section
6.1. Synthesis
crystal structures of hDHFR as a ternary complex with 3 and 14
(Fig. 8). This further validates our docking protocols.
All evaporations were carried out in vacuo with a rotary evapo-
rator. Analytical samples were dried in vacuo (0.2 mm Hg) in a
Table 4
Inhibition concentrations (IC50) against pjDHFR and hDHFR and selectivity ratios.
#
R
pjDHFR (nM)
hDHFR (nM)
Selectivity ratio [hDHFR/pjDHFR]
2
30-OCH3Ph
20-OCH3Ph
40-OCH3Ph
2-Naph
1-Naph
30,40-diFPh
40-OCF3Ph
160
210
219
130
177
247
110
1200
1400
1372
970
1104
1917
1101
8
7
6
7
6
8
10
13
14
15
16
17
18
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

8
K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
Fig. 8. (a) Comparison of the crystal structures of human DHFR as a ternary complex with 3 (yellow) and 14 (green) showing the electron density for the complex with
hDHFR-3 (2Fo-Fc, 1 , blue, 3 , green) and (b) Comparison of the binding pocket for hDHFR-3 (yellow) and 14 (green). Note that Phe31 occupies two alternative
conformations in these two structures. This change is in response to the larger N7-ethyl substituent of the inhibitor 3 as compared to N7-methyl substituent of 14.
r
r
CHEM-DRY drying apparatus over P2O5 at 70 °C. Melting points
were determined on a MEL-TEMP II melting point apparatus with
FLUKE 51 K/J electronic thermometer and are uncorrected. Nuclear
magnetic resonance spectra for proton (1H NMR) was recorded on
a Bruker 400/500 MHz NMR spectrometer. The chemical shift val-
ues are expressed in ppm (parts per million) relative to tetram-
ethylsilane as an internal standard: s, singlet; d, doublet; t,
triplet; q, quartet; m, multiplet; br, broad singlet. The relative inte-
grals of peak areas agreed with those expected for the assigned
structures. High-resolution mass spectra (HRMS) were recorded
on a MICROMASS AUTOSPEC (EBE Geometry) double focusing mass
The filtrate was evaporated in vacuo, and the residue was chro-
matographed on silica gel with 10% MeOH/CHCl3 as the eluent.
Fractions containing the product were combined and evaporated
to give 24 (7.3 g, 55%) as a light brown solid; TLC Rf 0.63 (MeOH/
CHCl3/NH4OH, 1:5:0.5); mp, 166–168 °C. 1H NMR (400 Hz)
(Me2SO-d6) d 2.23 (s, 3H, 5-CH3), 5.25–5.78 (br, 2H, 2-NH2, exch.),
6.19 (s, 2H, 4-NH2, exch.), 6.42 (s, 1H, 6-H), 10.43 (s, 1H, 7-H,
exch.).
6.1.1.3.
5-Methyl-6-(naphthalen-2-ylthio)thieno[2,3-d]pyrimidine-
spectrometer (Electron Impact
–
EI) or Waters Q-TOF
2,4-diamine (9). Compound 9 (0.32 g, 21%) was obtained from 21
(0.8 g, 4.4 mmol), 2-thionapthalene (1.42 g, 8.8 mmol), and iodine
(2.25 g, 8.8 mmol); TLC Rf 0.50 (MeOH/CHCl3/NH4OH, 1:5:0.5);
mp, 191.8–193.5 °C. 1H NMR (400 Hz) (Me2SO-d6) d 2.09 (s, 3H,
5-CH3), 5.66 (s, 2H, 2-NH2, exch.), 6.32 (s, 2H, 4-NH2, exch.), 7.17
(dd, 1H, C6H4, J = 1.8Hz, J = 8.7Hz), 7.45 (m, 3H, C6H4), 7.76 (d,
1H, J = 8.3Hz), 7.82 (s, 1H, C6H4), 7.84 (m, 1H, C6H4), 11.05 (s, 1H,
7-H, exch.). Anal. Calcd. for C17H15N5S: C, 63.53; H, 4.70; N,
21.79; S, 9.98. Found: C, 62.22; H, 4.86; N, 20.99; S, 9.60.
(quadrupole/time-of-flight tandem instrument) mass spectrome-
ter (Electro-Spray Ionization – ESI). Elemental analyses were per-
formed by Atlantic Microlab, Inc., Norcross, GA. Element
compositions are within 0.4% of the calculated values. Fractional
moles of water or organic solvents frequently found in some ana-
lytical samples of antifolates could not be prevented in spite of
24–48 h of drying in vacuo and were confirmed where possible
by their presence in the 1H NMR spectra. Thin-layer chromatogra-
phy (TLC) was performed on WHATMAN UV254 silica gel plates
with a fluorescent indicator, and the spots were visualized under
254 and/or 365 nm illumination. Proportions of solvents used for
TLC are by volume. Column chromatography was performed on a
230–400 mesh silica gel purchased from Fisher Scientific. All sol-
vents and chemicals were purchased from Sigma-Aldrich Chemical
Co. or Fisher Scientific.
6.1.1.4.
6-((3,4-Difluorophenyl)thio)-5-methylthieno[2,3-d]pyrim-
idine-2,4-diamine (11). Compound 11 (0.35 g, 24.3%) was obtained
from 21 (0.8 g, 4.4 mmol), 3,4-difluorothiophenol (1.30 g, 8.8
mmol), and iodine (2.25 g, 8.8 mmol); TLC Rf 0.50 (MeOH/CHCl3/
NH4OH, 1:5:0.5); mp, 291.7–295.8 °C. 1H NMR (500 Hz) (Me2SO-
d6) d 2.33 (s, 3H, 5-CH3), 5.65 (s, 2H, 2-NH2, exch.), 6.28 (s, 2H,
4-NH2, exch.), 6.81 (d, 1H, C6H4, J = 8.7 Hz), 7.06 (d, 1H, C6H4, J =
19.2 Hz), 7.37 (dd, 1H, C6H4, J = 8.6 Hz, J = 19.2 Hz), 10.99 (s, 1H,
7-H, exch.). Anal. Calcd. for C13H11F2N4S 0.34 CH3OH: C, 50.87; H,
4.05; F, 11.76; N, 21.68; S, 9.93. Found: C, 50.97; H, 3.88; F,
11.55; N, 21.65; S, 10.06.
6.1.1. Procedure for the synthesis of compounds 1, 7–12
6.1.1.1. 2-Amino-4-methyl-furan-3-carbonitrile (20). To a solution of
acetol (10 g, 135 mmol) in methanol (200 mL) at room tempera-
ture was added malononitrile (8.9 g, 135 mmol) and triethylamine
(13.7 g, 135 mmol). The resulting mixture was stirred at room tem-
perature overnight. The reaction mixture was then stripped of sol-
vent in vacuo. The residue was washed with hexane-ethyl acetate
(5:1) (250 mL ꢁ 5). The resulting hexane-ethyl acetate solution of
the product was collected. After the evaporation of solvent under
reduced pressure, 13 g (79%) of the crude product was obtained
as an orange powder and was used directly in the next reaction
without analysis.
6.1.1.5.
6-((4-Trifluromethoxyphenyl)thio)-5-methylthieno[2,3-d]
pyrimidine-2,4-diamine (12). Compound 12 (0.38 g, 28%) was
obtained from 21 (0.8 g, 4.4 mmol), 4-trifluromethoxythiophenol
(1.26 g, 8.8 mmol), and iodine (2.25 g, 8.8 mmol); TLC Rf 0.50
(MeOH/CHCl3/NH4OH, 1:5:0.5); mp, 291.7–295.8 °C. 1H NMR
(500 Hz) (Me2SO-d6) d 2.33 (s, 3H, 5-CH3), 5.66 (s, 2H, 2-NH2,
exch.), 6.28 (s, 2H, 4-NH2, exch.), 7.08 (d, 2H, C6H4, J = 8.9 Hz),
7.30 (d, 2H, C6H4, J = 8.1 Hz), 10.98 (s, 1H, 7-H, exch.). Anal. Calcd.
for C14H12F3N5OS: C, 47.32; H, 3.40; F, 16.04; N, 19.71; O, 4.50; S,
9.02. Found: C, 47.13; H, 3.50; F, 15.99; N, 19.57; S, 8.91
6.1.1.2. 2,4-Diamino-5-methyl-pyrrolo[2,3-d]pyrimidine (21). To a
solution of guanidine free base (from 82 mmol of NaOMe) in anhy-
drous ethanol (150 mL) was added aminonitrile 20 (10.0 g, 82
mmol). The mixture was refluxed for 24 h, cooled, and filtered.
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
9
6.1.2. General procedure for the synthesis of compounds 2–6, 13–18
Mixture of 6-substituted pyrrolo[2,3-d]pyrimidine and sodium
hydride was added to a three neck RBF. The RBF was made anhy-
drous using argon gas balloon. To this mixture, anhydrous DMF
(10 mL) was added and stirred for 30 min with vigorous stirring.
Subsequently, appropriate alkyl halide was injected in the reaction
mixture, and the resulting reaction mixture was stirred and mon-
itored by TLC until reaction was completed. The DMF was evapo-
rated under vacuum and silica plug was prepared. The final
compound was purified by flash chromatography using metha-
nol-chloroform gradient elution.
Calcd. for C17H21N5OS: C, 59.45; H, 6.16; N, 20.39; O, 4.66; S,
9.34. Found: C, 59.12; H, 6.08; N, 19.65; S, 8.87.
6.1.2.5. 7-Butyl-6-((3-methoxyphenyl)thio)-5-methyl-7H-pyrrolo[2,3-
d]pyrimidine-2,4-diamine (6). Reaction of 1 (0.100 g, 0.33 mmol),
sodium hydride (0.010 g, 0.4 mmol) and 1-bromobutane (55
mmol, 0.5 mmol) using the general procedure described above
gave 6 (0.065 g, 55%) as white solid; TLC Rf 0.62 (MeOH/CHCl3/
NH4OH, 1:5:0.5); mp, 282.4–284.2 °C; 1H NMR (400 Hz) (Me2SO-
d6) d 0.76 (t, 3H, J = 7.3 Hz, ACH3), 1.14 (qd, 2H, J = 7.2 Hz, J =
14.6 Hz, ACH2A), 1.45 (td, 2H, J = 7.5 Hz, J = 14.7 Hz, ACH2A),
2.38 (s, 3H, 5-CH3), 3.68 (s, 3H, 3-OCH3), 3.87 (t, 2H, J = 14.5 Hz,
ACH2A),3.92 (q, 2H, J = 7.0 Hz, ACH2A), 5.80 (s, 2H, 2-NH2, exch.),
6.49 (s, 2H, 4-NH2, exch.), 6.49 (d, 2H, J = 7.2 Hz, C6H4), 6.72 (dd,
1H, C6H4, J = 2.1 Hz, J = 7.2 Hz), 7.19 (t, 1H, C6H4, J = 8.2 Hz). Anal.
Calcd. for C18H23N5OS 0.03 CHCl3: C, 59.98; H, 6.43; N, 19.40; O,
4.48; S, 8.88. Found: C, 59.94; H, 6.25; N, 19.38; S, 8.78.
6.1.2.1. 6-((3-Methoxyphenyl)thio)-5,7-dimethyl-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (2). Reaction of 1 (0.150 g, 0.32 mmol),
sodium hydride (0.012 g, 0.5 mmol) and iodomethane (31 mmol,
0.5 mmol) using the general procedure described above gave 2
(0.120 g, 76.44%) as white solid; TLC Rf 0.58 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 277.4–279.4 °C; 1H NMR (400 Hz) (Me2SO-d6) d 2.38
(s, 3H, 5-CH3), 3.37 (s, 3H, 7-CH3), 3.69 (s, 3H, 3-OCH3), 5.81 (s,
1.68H, 2-NH2, exch.), 6.37 (s, 1.58H, 4-NH2, exch.), 6.49 (dd, 2H,
C6H4, J = 1.76, 10.55 Hz), 6.73 (dd, 1H, C6H4, J = 2.35, 8.20 Hz),
6.1.2.6. 6-((2-Methoxyphenyl)thio)-5,7-dimethyl-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (13). Reaction of 7 (0.150 g, 0.5 mmol),
sodium hydride (0.012 g, 0.5 mmol) and iodomethane (31 mmol,
0.5 mmol using the general procedure described above gave 13
(0.1 g, 64%) as a white solid; TLC Rf 0.58 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 208.6–209.6 °C; 1H NMR (400 Hz) (Me2SO-d6) d
2.34 (s, 3H, 5-CH3), 3.37 (s, 3H, 7-CH3), 3.89 (s, 3H, 2-OCH3), 5.76
(s, 2H, 2-NH2, exch.), 6.35 (s, 2H, 4-NH2, exch.), 6.81 (m, 2H,
C6H4), 7.01 (d, 1H, J = 6.7 Hz, C6H4), 7.11 (d, 1H, J = 7.8 Hz, C6H4).
HRMS (ESI) calculated for C15H17N5OS [M+H]+, 316.12266. Found:
316.12402. HPLC analysis: retention time, 21.99 min; peak area,
97.37%; eluent A, H2O: eluent B, ACN; gradient elution (100%
H2O to 10% H2O) over 60 min with flow rate of 0.5 mL/min and
detection at 245 nm; column temperature, rt.
7.19 (t, 1H, C6H4, J = 7.96, 7.96 Hz). HRMS (ESI) calculated for C15
-
H17N5OS [M+H]+, 316.12266. Found: 316.12198. Found:
316.12198. HPLC analysis: retention time, 22.79 min; peak area,
96.08%; eluent A, H2O: eluent B, ACN; gradient elution (100%
H2O–10% H2O) over 60 min with flow rate of 0.5 mL/min and
detection at 245 nm; column temperature, rt.
6.1.2.2. 7-Ethyl-6-((3-methoxyphenyl)thio)-5-methyl-7H-pyrrolo[2,3-
d]pyrimidine-2,4-diamine (3). Reaction of 1 (0.150 g, 0.50 mmol),
sodium hydride (0.012 g, 0.5 mmol) and bromoethane (53 mmol,
0.5 mmol) using the general procedure described above gave 3
(0.095 g, 60.5%) as white solid; TLC Rf 0.58 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 136.9–139.4 °C; 1H NMR (400 Hz) (Me2SO-d6) d
1.04 (t, 3H, J = 7.0 Hz, ACH3), 2.37 (s, 3H, 5-CH3), 3.68 (s, 3H, 3-
OCH3), 3.92 (q, 2H, J = 7.0 Hz, ACH2A), 5.80 (s, 2H, 2-NH2, exch.),
6.35 (s, 2H, 4-NH2, exch.), 6.49 (d, 2H, J = 7.2 Hz, C6H4), 6.72 (dd,
1H, C6H4, J = 2.1 Hz, J = 7.2 Hz), 7.19 (t, 1H, C6H4, J = 8.2 Hz). Anal.
Calcd. for C16H19N5OS: C, 58.34; H, 5.81; N, 21.26; O, 4.86; S,
9.73. Found: C, 58.0.; H, 5.97; N, 21.05; S, 9.52.
6.1.2.7. 6-((4-Methoxyphenyl)thio)-5,7-dimethyl-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (14). Reaction of 8 (0.150 g, 0.32 mmol),
sodium hydride (0.012 g, 0.5 mmol) and iodomethane (31 mmol,
0.5 mmol) using the general procedure described above gave 14
(0.135 g, 86%) as white solid; TLC Rf 0.58 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 266.0–267.8 °C; 1H NMR (400 Hz) (Me2SO-d6) d
2.38 (s, 3H, 5-CH3), 3.37 (s, 3H, 7-CH3), 3.69 (s, 3H, 3-OCH3), 5.81
(s, 2H, 2-NH2), 6.37 (s, 2H, 4-NH2), 6.49 (dd, J = 1.76, 10.55 Hz,
2H, C6H4), 6.73 (dd, J = 2.35, 8.20 Hz, 1H, C6H4), 7.19 (t, J = 7.96,
7.96 Hz, 1H, C6H4). Anal. Calcd. for C15H17N5OS: C, 57.12; H, 5.43;
N, 22.21; O, 5.07; S, 10.17. Found: C, 56.90; H, 5.48; N, 21.94; S,
10.01.
6.1.2.3.
6-((3-Methoxyphenyl)thio)-5-methyl-7-propyl-7H-pyrrolo
(0.120 g, 0.50
[2,3-d]pyrimidine-2,4-diamine (4). Reaction of
1
mmol), sodium hydride (0.012 g, 0.5 mmol) and 1-bromopropane
(62 mmol, 0.5 mmol) using the general procedure described above
gave 4 (0.050 g, 37%) as white solid; TLC Rf 0.60 (MeOH/CHCl3/
NH4OH, 1:5:0.5); mp, 266.4–268.2 °C; 1H NMR (400 Hz) (Me2SO-
d6) d 0.73 (t, 3H, J = 7.4 Hz, ACH3), 1.51 (qd, 2H, J = 7.2 Hz, J =
14.5 Hz, ACH2A), 2.37 (s, 3H, 5-CH3), 3.35 (s, 3H, 3-OCH3), 3.83
(t, 2H, J = 14.5 Hz, ACH2A), 3.92 (q, 2H, J = 7.0 Hz, ACH2A), 5.79
(s, 2H, 2-NH2, exch.), 6.35 (s, 2H, 4-NH2, exch.), 6.49 (d, 2H, J =
7.2 Hz, C6H4), 6.72 (dd, 1H, C6H4, J = 2.1 Hz, J = 7.2 Hz), 7.19 (t,
1H, C6H4, J = 8.2 Hz).). Anal. Calcd. for C17H21N5OS: C, 59.45; H,
6.16; N, 20.39; O, 4.66; S, 9.34. Found: C, 58.72; H, 6.23; N,
19.97; S, 8.98.
6.1.2.8.
5,7-Dimethyl-6-(naphthalen-2-ylthio)-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (15). Reaction of 9 (0.18 g, 0.56 mmol),
sodium hydride (0.016 g, 0.67 mmol) and iodomethane (40 mmol,
0.64 mmol) using the general procedure described above gave 15
(0.11 g, 59%) as white solid; TLC Rf 0.57 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 266.0–267.8 °C 1H NMR (400 Hz) (Me2SO-d6) d 2.43
(s, 3H, 5-CH3), 3.40 (s, 3H, 7-CH3), 5.82 (s, 2H, 2-NH2, exch.), 6.39
(s, 2H, 4-NH2, exch.), 7.13 (d, 1H, J = 8.7 Hz), 7.45 (dd, 3H, J = 6.7
Hz, J = 12.8 Hz)), 7.78 (d, 1H, J = 7.8 Hz), 7.85 (d, 2H, J = 8.6 Hz).
Anal. Calcd. for C18H17N5S: C, 64.45; H, 5.11; N, 20.88; S, 9.56.
Found: C, 64.21; H, 5.25; N, 20.68; S, 9.29.
6.1.2.4. 7-Isopropyl-6-((3-methoxyphenyl)thio)-5-methyl-7H-pyrrolo
[2,3-d]pyrimidine-2,4-diamine (5). Reaction of
1 (0.090 g, 0.30
mmol), sodium hydride (0.009 g, 0.36 mmol) and 2-bromopropane
(38 mmol, 0.36 mmol) using the general procedure described
above gave 5 (0.060 g, 59%) as white solid; TLC Rf 0.60 (MeOH/
CHCl3/NH4OH, 1:5:0.5); mp, 157.4–160.1 °C; 1H NMR (400 Hz)
(Me2SO-d6) d 1.40 (d, 6H, J = 6.7 Hz, ACH3), 2.38 (s, 3H, 5-CH3),
3.68 (s, 3H, 3-OCH3), 4.13(m, 1H, ACHA), 5.80 (s, 2H, 2-NH2, exch.),
6.49 (s, 2H, 4-NH2, exch.), 6.49 (d, 2H, J = 7.2 Hz, C6H4), 6.72 (dd,
1H, C6H4, J = 2.1 Hz, J = 7.2 Hz), 7.19 (t, 1H, C6H4, J = 8.2 Hz). Anal.
6.1.2.9.
5,7-Dimethyl-6-(naphthalen-1-ylthio)-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (16). Reaction of 10 (0.20 g, 0.62 mmol),
sodium hydride (0.017 g, 0.75 mmol) and iodomethane (46 mmol,
0.72 mmol) using the general procedure described above gave 16
(0.12 g, 57%) as white solid; TLC Rf 0.57 (MeOH/CHCl3/NH4OH
1:5:0.5); mp, 232.6–235.6 °C 1H NMR (400 Hz) (Me2SO-d6) d 2.51
(s, 3H, 5-CH3), 3.39 (s, 3H, 7-CH3), 5.89 (s, 2H, 2-NH2, exch.), 6.49
(s, 2H, 4-NH2, exch.), 6.64 (d, 1H, C6H4, J = 7.3 Hz), 7.34 (t, 1H, J =
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

10
K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
7.8 Hz), 7.63 (td, 2H, J = 6.9 Hz, J = 14.9 Hz,), 7.72 (d, 1H, J = 8.1 Hz),
7.96 (d, 1H, J = 8.0 Hz), 8.28 (d, 1H, J = 8.3 Hz). Anal. Calcd. for
6.4. Crystallization and X-ray data collection and refinement
C
18H17N5S: C, 64.45; H, 5.11; N, 20.88; S, 9.56. Found: C, 64.68;
Expression and purification of wild type human dihydrofolate
reductase (hDHFR) were carried out as previously described.41
Recombinant hDHFR was washed in a Centricon-10 with 100 mM
K2HPO4 buffer pH 6.9 with 30% saturated ammonium sulfate and
concentrated to 7.9 mg mlꢀ1. The hDHFR samples were incubated
for 1 h on ice with a tenfold excess of NADPH and compounds 3
and 14, respectively, prior to crystallization using the hanging-
drop vapor diffusion method using siliconized glass cover slips
and storage at 14 °C. Protein droplets of the hDHFR complexes con-
tained K2HPO4 pH 6.9 with 30% saturated ammonium sulfate equi-
librated against a reservoir solution consisting of 100 mM K2HPO4
pH 6.9 with 60% saturated ammonium sulfate, 3% (v/v) ethanol.
Crystals of hDHFR-3-NADPH and hDHFR-14-NADPH ternary com-
plex were hexagonal and belonged to the space group H3. Data
were collected at 100 K to 1.46 Å resolution for both crystals using
the remote access robot on beamline 14.7 at the Stanford Syn-
chrotron Radiation Laboratory.19,42–44 The data were processed
using HKL2000 program package.45 The diffraction statistics are
shown in Table 5. Both crystal structures were solved by molecular
replacement methods using the coordinates for hDHFR (1u72)46 in
the program Molref.47 Inspection of the resulting difference elec-
tron density maps made using COOT48 running on an iMac work-
station revealed density for the ternary complex of both crystals.
The final cycles of refinement were carried out using the program
Refmac5 in the CCP4 suite of programs.47 The Ramachandran con-
formational parameters from the last cycle of refinement gener-
H, 4.91; N, 20.82; S, 9.59.
6.1.2.10. 6-((3,4-Difluorophenyl)thio)-5,7-dimethyl-7H-pyrrolo[2,3-d]
pyrimidine-2,4-diamine (17). Reaction of 11 (0.120 g, 0.39 mmol),
sodium hydride (0.012 g, 0.5 mmol) and iodomethane (31 mmol,
0.5 mmol) using the general procedure described above gave 17
(0.08 g, 64%) as white solid; TLC Rf 0.57 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 266.0–267.8 °C 1H NMR (500 Hz) (Me2SO-d6) d 2.38
(s, 3H, 5-CH3), 3.34 (s, 3H, 7-CH3), 5.83 (s, 2H, 2-NH2, exch.), 6.38
(s, 2H, 4-NH2, exch.), 6.75 (d, 1H, C6H4, J = 8.7 Hz), 7.06 (ddd, 1H,
C6H4, J = 2.3 Hz, J = 7.4 Hz, J = 10.8 Hz), 7.37 (ddd, 1H, C6H4, J =
2.3 Hz, J = 7.4 Hz, J = 10.8 Hz). Anal. Calcd. for C14H13F2N5S 0.04
CHCl3: C, 51.71; H, 4.03; F, 11.65; N, 21.47; S, 9.64. Found: C,
51.75; H, 4.01; F, 11.47; N, 21.30; S, 9.74
6.1.2.11. 5,7-dimethyl-6-((4-(trifluoromethoxy)phenyl)thio)-7H-pyr-
rolo[2,3-d]pyrimidine-2,4-diamine (18). Reaction of 12 (0.150 g,
0.42 mmol), sodium hydride (0.012 g, 0.5 mmol) and iodomethane
(32 mmol, 0.5 mmol) using the general procedure described above
gave 18 (0.1 g, 48%) as white solid; TLC Rf 0.57 (MeOH/CHCl3/NH4OH,
1:5:0.5); mp, 282.0–283.8 °C 1H NMR (400 Hz) (Me2SO-d6) d 2.39
(s, 3H, 5-CH3), 3.38 (s, 3H, 7-CH3), 5.88 (s, 2H, 2-NH2, exch.), 6.47
(s, 2H, 4-NH2, exch.), 7.06 (d, 2H, C6H4, J = 8.2Hz), 7.29 (d, 2H,
C6H4, J = 8.3Hz). Anal. Calcd. for C15H14F3N5OS: C, 48.78; H, 3.82;
F, 15.43; N, 18.96; O, 4.33; S, 8.68. Found: C, 49.03; H, 4.01; F,
18.93; N, 18.93; S, 8.61.
Table 5
Crystal properties and refinement statistics of 3 and 14 bound to hDHFR.
6.2. Molecular modeling
Compound
(B2-282) 3
(A6-283) 14
PDB accession
Space group
Lattice constants (Å)/°
a
b
c
a
b
5HT4
H3
5HT5
H3
Docking of compounds 1–18 was carried out using the pub-
lished X-ray crystal structure of N6-methyl-N6-(3,4,5-trifluo-
rophenyl)pyrido[3,2-d]pyrimidine-2,4,6-triamine in hDHFR (PDB:
4QJC, 1.62 Å)33 and in the homology model of pjDHFR32 using Lea-
dIT 2.1.6.36 The docking in LeadIT was constrained to the active site
of the protein. Polar hydrogen atoms of amino acids were not con-
strained, thereby permitting them free rotation. Base placement of
fragments for docking was carried out using triangle docking.
Default parameters were used for scoring and clash handling. The
maximum number of solutions per iteration and maximum num-
ber of fragmentation were set to 200. Ten poses were obtained
per molecule. The docked poses were exported to MOE 2016.08
for visualization.39 The validation of LeadIT as a suitable docking
system for pjDHFR and hDHFR was carried out by re-docking the
native ligands in the X-ray crystal structures of pcDHFR (PDB:
2FZI)40 and hDHFR (PDB: 4QJC). The ligands were sketched in
MOE 2016.0839 and docking was carried out with LeadIT 2.1.6 as
described above. The best docked pose of the ligands had RMSD
of 0.7060 Å in pjDHFR and 0.8860 Å in hDHFR. Thus, LeadIT 2.1.6
was validated and chosen for the docking studies.
85.68
85.68
77.03
90.0
90.0
120.0
SSRL 14-1
1.46 (1.49)
0.979
0.05 (0.067)
92.6 (46.2)
121,242
37,327
35.0 (0.90)
2.0 (1.4)
25,428
34.2–1.60
0.24
0.30
1677
78
28.2
85.45
85.45
77.69
90.0
90.0
120.0
SSRL 14-1
1.46 (1.49)
0.979
0.174 (0.136)
78.9 (70.6)
49,944
36,629
34.5 (2.4)
1.3 (1.2)
12,777
26.3–1.90
0.19
0.28
1653
75
33.2
c
Beamline
Resolution Å
Wavelength Å
Rmerge %a,b
Completeness %
Observed Reflect
Unique Reflections
I/r(I)
Multiplicitya
Reflections used
Resolution Å
R-factor
Rfree
Total protein atoms
Total water atoms
Average B-factor Å2
Error in Luzzati plot
Rms deviation from ideal
Bond length Å
0.27
0.24
6.3. Pharmacological assay
0.021
2.38
0.021
2.23
Bond angle
The expression and purification of recombinant pj- and hDHFR
was carried out as previously described.18 Standard DHFR assays
were conducted at 37 °C with continuous recording of change of
absorbance at 340 nM. The assay contained 41 mM sodium phos-
phate buffer at pH 7.4, 8.9 mM 2-mercaptoethanol, 150 mM KCl,
Ramachandran plot
Most favored %
Additional allowed %
Disallowed %
96.7
2.7
0.5
95.7
3.3
1.1
aThe values in parentheses refer to data in the highest resolution shell.
bRsym
RhRi|Ih,i ꢀ hIhi|/RhRi|Ih,i|, where hIhi is the mean intensity of
equivalent reflections.
cR-factor =
|Fobs ꢀ Fcalc|/
structure factor amplitudes.
dRfree-factor was calculated for R-factor for a random 5% subset of all reflections.
=
a set of
and saturating concentrations of NADPH (117
l
M). Dihydro folic
M. The
acid (DHFA) was used at an optimum concentration of 9
l
R
RFobs, where Fobs and Fcalc are observed and calculated
results reported previously by Namjoshi et al.31 were carried out
at 18 M of DHFA.
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032

K. Shah et al. / Bioorganic & Medicinal Chemistry xxx (2018) xxx–xxx
11
ated by RAMPAGE49 showed that more than 96% of the residues
refined have the most favored conformation and none are in the
disallowed regions. Coordinates for these structures have been
deposited with the Protein Data Bank.
associated factors among HIV-infected patients treated for Pneumocystis
jirovecii pneumonia. PLoS One. 2014;9 e106141.
21. Gordin FM, Simon GL, Wofsy CB, Mills J. Adverse reactions to trimethoprim-
sulfamethoxazole in patients with the acquired immunodeficiency syndrome.
Ann Intern Med. 1984;100:495–499.
22. Masters PA, O’Bryan TA, Zurlo J, Miller DQ, Joshi N. Trimethoprim-
sulfamethoxazole revisited. Arch Intern Med. 2003;163:402–410.
23. Floris-Moore MA, Amodio-Groton MI, Catalano MT. Adverse reactions to
Acknowledgements
trimethoprim/sulfamethoxazole
2003;37:1810–1813.
24. Ho JM, Juurlink DN. Considerations when prescribing trimethoprim-
sulfamethoxazole. CMAJ. 2011;183:1851–1858.
in
AIDS.
Ann
Pharmacother.
This work was supported, in part, by grants from the National
Institute of Health (NIH), National Institute of Allergy and Infec-
tious Diseases (NIAID), RO1AI098458, Adrian van Kaam Chair in
Scholarly Excellence (AG) Duquesne University, and NSF for NMRs.
We would like to acknowledge Dr. Sudhir Raghavan for his input in
conducting the molecular modeling.
25. Sangiolo D, Storer B, Nash R, et al. Toxicity and efficacy of daily dapsone as
Pneumocystis jiroveci
prophylaxis after hematopoietic stem
cell
transplantation:
a
case-control study. Biol Blood Marrow Transplant.
2005;11:521–529.
26. White NJ. Cardiotoxicity of antimalarial drugs. Lancet Infect Dis.
2007;7:549–558.
27. Nickel P, Schurmann M, Albrecht H, et al. Clindamycin-primaquine for
Pneumocystis jiroveci pneumonia in renal transplant patients. Infection.
2014;42:981–989.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmc.2018.04.032.
28. Kovacs JA, Allegra CJ, Swan JC, et al. Potent antipneumocystis and
antitoxoplasma activities of piritrexim, a lipid-soluble antifolate. Antimicrob
Agents Chemother. 1988;32:430–433.
29. Short C-ES, Gilleece YC, Fisher MJ, Churchill DR. Trimetrexate and folinic acid: a
valuable salvage option for Pneumocystis jirovecii pneumonia. AIDS.
2009;23:1287–1290.
References
30. Gangjee A, Kurup S, Namjoshi O. Dihydrofolate reductase as
a target for
1. Kovacs JA. 349 – Pneumocystis pneumonia A2 – Goldman, Lee. In: Schafer AI,
ed. Goldman’s Cecil Medicine (Twenty-Fourth Edition). Philadelphia: W.B.
Saunders; 2012:1997–2005.
chemotherapy in parasites. Curr Pharm Des. 2007;13:609–639.
31. Cody V, Chisum K, Pope C, Queener SF. Purification and characterization of
human-derived Pneumocystis jirovecii dihydrofolate reductase expressed in
Sf21 insect cells and in Escherichia coli. Protein Expr Purif. 2005;40:417–423.
32. Gangjee A, Namjoshi OA, Raghavan S, Queener SF, Kisliuk RL, Cody V. Design,
synthesis, and molecular modeling of novel pyrido[2,3-d]pyrimidine analogues
as antifolates; application of Buchwald-Hartwig aminations of heterocycles. J
Med Chem. 2013;56:4422–4441.
2. Huang YS, Yang JJ, Lee NY, et al. Treatment of Pneumocystis jirovecii pneumonia
in HIV-infected patients:
2017;15:873–892.
a
review. Expert Rev Anti Infect Ther.
3. Sokulska M, Kicia M, Wesołowska M, Hendrich AB. Pneumocystis jirovecii—from
commensal to pathogen: clinical and diagnostic review. Parasitol Res.
a
2015;114:3577–3585.
33. Cody V, Pace J, Namjoshi OA, Gangjee A. Structure-activity correlations for
three pyrido[2,3-d]pyrimidine antifolates binding to human and Pneumocystis
4. Yiannakis EP, Boswell TC. Systematic review of outbreaks of Pneumocystis
jirovecii pneumonia: evidence that P. jirovecii is a transmissible organism and
the implications for healthcare infection control. J Hosp Infect. 2016;93:1–8.
5. Goldman L, Schafer AI. Goldman-Cecil Medicine. Elsevier Health Sciences; 2015.
6. Bernheimer JM, Patten G, Makeleni T, et al. Paediatric HIV treatment failure: a
silent epidemic. J Intl AIDS Soc. 2015;18:20090.
7. Armstrong-James D, Meintjes G, Brown GD. A neglected epidemic: fungal
infections in HIV/AIDS. Trends Microbiol. 2014;22:120–127.
8. Harris JR, Balajee SA, Park BJ. Pneumocystis jirovecii pneumonia: current
knowledge and outstanding public health issues. Curr Fungal Infect Rep.
2010;4:229–237.
carinii dihydrofolate reductase. Acta Crystallogr
F Struct Biol Commun.
2015;71:799–803.
34. Taylor EC, Patel HH, Jun J-G. A one-step ring transformation/ring annulation
approach to pyrrolo[2,3-d]pyrimidines. new synthesis of the potent
a
dihydrofolate reductase inhibitor TNP-351. J Org Chem. 1995;60:6684–6687.
35. Gangjee A, Lin X, Queener SF. Design, synthesis, and biological evaluation of
2,4-diamino-5-methyl-6-substituted-pyrrolo[2,3-d]pyrimidines
dihydrofolate reductase inhibitors. J Med Chem. 2004;47:3689–3692.
36. LeadIT 2.1.6; Biosolve IT: St. Augustin, Germany; www.biosolveit.de.
as
37. Cody V, Schwalbe CH. Structural characteristics of antifolate dihydrofolate
reductase enzyme interactions. Crystallogr Rev. 2006;12:301–333.
38. SYBYL-X 2.1.1, Tripos International, 1699 South Hanley Rd., St. Louis, Missouri,
63144, USA.
39. Molecular Operating Environment (MOE), 2016.08; Chemical Computing
Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A
2R7, 2017.
9. Masur H, Brooks JT, Benson CA, Holmes KK, Pau AK, Kaplan JE. Prevention and
treatment of opportunistic infections in HIV-infected adults and adolescents:
updated Guidelines from the Centers for Disease Control and Prevention,
National Institutes of Health, and HIV Medicine Association of the Infectious
Diseases Society of America. Clin Infect Dis. 2014;58:1308–1311.
10. Miller RF, Huang L, Walzer PD. Pneumocystis pneumonia associated with human
immunodeficiency virus. Clin Chest Med. 2013;34:229–241.
40. Cody V, Pace J, Chisum K, Rosowsky A. New insights into DHFR interactions:
analysis of Pneumocystis carinii and mouse DHFR complexes with NADPH and
two highly potent 5-(omega-carboxy(alkyloxy) trimethoprim derivatives
reveals conformational correlations with activity and novel parallel ring
stacking interactions. Proteins. 2006;65:959–969.
41. Cody V, Pace J, Makin J, Piraino J, Queener SF, Rosowsky A. Correlations of
inhibitor kinetics for Pneumocystis jirovecii and human dihydrofolate reductase
with structural data for human active site mutant enzyme complexes.
Biochemistry. 2009;48:1702–1711.
42. Cohen AE, Ellis PJ, Miller MD, Deacon AM, Phizackerley RP. An automated
system to mount cryo-cooled protein crystals on a synchrotron beam line,
using compact sample cassettes and a small-scale robot. J Appl Crystallogr.
2002;35:720–726.
11. Hawser S, Lociuro S, Islam K. Dihydrofolate reductase inhibitors as antibacterial
agents. Biochem Pharmacol. 2006;71:941–948.
12. Yoon C, Subramanian A, Chi A, et al. Dihydropteroate synthase mutations in
pneumocystis pneumonia: impact of applying different definitions of
prophylaxis, mortality endpoints and mutant in a single Cohort. Med Mycol.
2013;51:568–575.
13. Huang L, Crothers K, Atzori C, et al. Dihydropteroate synthase gene mutations
in pneumocystis and sulfa resistance. Emerg Infect Dis. 2004;10:1721–1728.
14. Ponce CA, Chabe M, George C, et al. High prevalence of pneumocystis jirovecii
dihydropteroate synthase gene mutations in patients with a first episode of
Pneumocystis pneumonia in Santiago, Chile, and clinical response to
trimethoprim-sulfamethoxazole therapy. Antimicrob Agents Chemother.
2017;61:e01290–16.
15. Queener SF, Cody V, Pace J, Torkelson P, Gangjee A. Trimethoprim resistance of
dihydrofolate reductase variants from clinical isolates of Pneumocystis jirovecii.
Antimicrob Agents Chemother. 2013;57:4990–4998.
43. González A, Moorhead P, McPhillips SE. Web-Ice: integrated data collection and
analysis
for
macromolecular
crystallography.
J
Appl
Crystallogr.
2008;41:176–184.
44. McPhillips TM, McPhillips SE, Chiu HJ, et al. Blu-Ice and the Distributed Control
System: software for data acquisition and instrument control at
16. Nahimana A, Rabodonirina M, Bille J, Francioli P, Hauser PM. Mutations of
Pneumocystis jirovecii dihydrofolate reductase associated with failure of
prophylaxis. Antimicrob Agents Chemother. 2004;48:4301–4305.
17. Siripattanapipong S, Leelayoova S, Mungthin M, Worapong J, Tan-Ariya P. Study
of DHPS and DHFR genes of Pneumocystis jirovecii in Thai HIV-infected patients.
Med Mycol. 2008;46:389–392.
macromolecular
2002;9:401–406.
crystallography
beamlines.
J
Synchrotron
Radiat.
45. Otwinowski Z, Minor W. Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 1997;276:307–326.
46. Cody V, Luft JR, Pangborn W. Understanding the role of Leu22 variants in
methotrexate resistance: comparison of wild-type and Leu22Arg variant mouse
and human dihydrofolate reductase ternary crystal complexes with
methotrexate and NADPH. Acta Crystallogr D Biol Crystallogr. 2005;61:147–155.
47. Winn MD, Ballard CC, Cowtan KD, et al. Overview of the CCP4 suite and current
developments. Acta Crystallogr D Biol Crystallogr. 2011;67:235–242.
48. Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot.
Acta Crystallogr D Biol Crystallogr. 2010;66:486–501.
49. Lovell SC, Davis IW, Arendall WB, et al. Structure validation by Calpha
geometry: phi, psi and Cbeta deviation. Proteins. 2003;50:437–450.
18. Cody V, Pace J, Queener SF, Adair OO, Gangjee A. Kinetic and structural analysis
for potent antifolate inhibition of Pneumocystis jirovecii, Pneumocystis carinii,
and human dihydrofolate reductases and their active-site variants. Antimicrob
Agents Chemother. 2013;57:2669–2677.
19. Cody V, Pace J, Makin J, Piraino J, Queener SF, Rosowsky A. Correlations of
inhibitor kinetics for Pneumocystis jirovecii and human dihydrofolate reductase
with structural data for human active site mutant enzyme complexes.
Biochemistry. 2009;48:1702–1711.
20. Yang JJ, Huang CH, Liu CE, et al. Multicenter study of
trimethoprim/sulfamethoxazole-related
hepatotoxicity:
incidence
and
Please cite this article in press as: Shah K., et al. Bioorg. Med. Chem. (2018), https://doi.org/10.1016/j.bmc.2018.04.032