Design, synthesis, and molecular modeling of novel pyrido[2,3-d]pyrimidine analogues as antifolates; Application of buchwald-hartwig aminations of heterocycles

Article abstract of DOI:10.1021/jm400086g

Opportunistic infections caused by Pneumocystis jirovecii (P. jirovecii, pj), Toxoplasma gondii (T. gondii, tg), and Mycobacterium avium (M. avium, ma) are the principal causes of morbidity and mortality in patients with acquired immunodeficiency syndrome (AIDS). The absence of any animal models for human Pneumocystis jirovecii pneumonia and the lack of crystal structures of pjDHFR and tgDHFR make the design of inhibitors challenging. A novel series of pyrido[2,3-d]pyrimidines as selective and potent DHFR inhibitors against these opportunistic infections are presented. Buchwald-Hartwig coupling reaction of substituted anilines with pivaloyl protected 2,4-diamino-6-bromo-pyrido[2,3-d] pyrimidine was successfully explored to synthesize these analogues. Compound 26 was the most selective inhibitor with excellent potency against pjDHFR. Molecular modeling studies with a pjDHFR homology model explained the potency and selectivity of 26. Structural data are also reported for 26 with pcDHFR and 16 and 22 with variants of pcDHFR.

Full text of DOI:10.1021/jm400086g

Subscriber access provided by MONASH UNIVERSITY
Article
Design, synthesis and molecular modeling of novel
pyrido[2,3-d]pyrimidine analogs as antifolates: Application
of Buchwald-Hartwig aminations of heterocycles
Aleem Gangjee, Ojas A. Namjoshi, Sudhir Raghavan, Sherry F. Queener, Roy L. Kisliuk, and Vivian Cody
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm400086g • Publication Date (Web): 29 Apr 2013
Downloaded from http://pubs.acs.org on May 1, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 67
Journal of Medicinal Chemistry
1
Design, synthesis and molecular modeling of novel
pyrido[2,3-d]pyrimidine analogs as antifolates:
Application of Buchwald-Hartwig aminations of
heterocycles
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
a*
a
a
b
c
Aleem Gangjee, Ojas A. Namjoshi, Sudhir Raghavan, Sherry F. Queener, Roy L. Kisliuk, Vivian
19
2
2
d
Cody
2
23
24
25
a
Division of Medicinal Chemistry, Graduate School Pharmaceutical Sciences, Duquesne University,
26
27
28
6
00 Forbes Avenue, Pittsburgh, PA 15282.
2
30
b
Department of Pharmacology and Toxicology, Indiana University School of Medicine, 635 Barnhill
31
3
3
34
Drive, Indianapolis, IN 46202.
c
Department of Biochemistry, Tufts University Health Science Campus, 136 Harrison Avenue, Boston,
35
36
37
MA 02111.
38
3
d
Hauptman-Woodward Medical Research Institute, 700 Ellicott Street, Buffalo, NY 14203.
40
41
42
4
4
*
Corresponding author: gangjee@duq.edu; 412-396-6070.
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 67
1
2
3
4
5
6
7
8
9
1
1
12
13
Abbreviations: PCP, Pneumocystis pneumonia; pj, Pneumocystis jirovecii; pc, Pneumocystis carinii;
tg, toxoplasma gondii; ma, Mycobacterium avium; rl, rat liver; DHFR, dihydrofolate reductase; TS,
thymidylate synthase; HIV, human immunodeficiency virus; AIDS, acquired immunodeficiency
syndrome; TMP, trimethoprim; PTX, piritrexim; TMQ, trimetrexate; LiHMDS, lithium
hexamethyldisilazide; X-Phos, 2-dicyclohexylphosphino-2',4',6'-triisopropylbiphenyl; S-Phos, 2-
Dicyclohexylphosphino-2',6'-dimethoxybiphenyl; rac-BINAP, racemic-2,2'-Bis(diphenylphosphino)-
1,1'-binaphthyl; XantPhos, 9,9-dimethyl-4,5-bis(diphenylphosphino)xanthene; DCPB,
14
15
16
17
18
19
20
2
2
23
24
25
dicyclohexylphospho biphenyl; t-Butyl-X-Phos, 2-di-t-butylphosphino-2',4',6'-tri-i-propyl-1,1'-biphenyl;
S-Phos, 2-dicyclohexylphosphino-2',6'-dimethoxybiphenyl; Ru-Phos, 2-dicyclohexylphosphino-2',6'-di-
i-propoxy-1,1'-biphenyl; DavePhos, 2-(dicyclohexylphosphino)-2'-(N,N-dimethylamino)biphenyl.
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
2

Page 3 of 67
Journal of Medicinal Chemistry
1
2
Abstract:
3
4
5
6
7
8
9
Opportunistic infections caused by Pneumocystis jirovecii (P. jirovecii, pj), Toxoplasma gondii
T. gondii, tg) and Mycobacterium avium (M. avium, ma) are the principal causes of morbidity and
(
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
mortality in patients with acquired immunodeficiency syndrome (AIDS). The absence of any animal
models for human Pneumocystis jirovecii pneumonia and the lack of crystal structures of pjDHFR and
tgDHFR make the design of inhibitors challenging. A novel series of pyrido[2,3-d]pyrimidines as
selective and potent DHFR inhibitors against these opportunistic infections are presented. Buchwald-
Hartwig coupling reaction of substituted anilines with pivaloyl protected 2,4-diamino-6-bromo-
pyrido[2,3-d]pyrimidine was successfully explored to synthesize these analogs. Compound 26 was the
most selective inhibitor with excellent potency against pjDHFR. Molecular modeling studies with a
pjDHFR homology model explained the potency and selectivity of 26. Structural data are also reported
for 26 with pcDHFR and 16 and 22 with variants of pcDHFR.
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 67
1
2
3
4
5
6
7
8
9
Introduction:
Dihydrofolate reductase (DHFR) converts dihydrofolate into tetrahydrofolate, a cofactor
required for the de novo synthesis of purines, thymidylate, and certain amino acids. DHFR has been
1
successfully exploited in the design of anticancer and antimicrobial agents.
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
Figure 1: Structures of dihydrofolate reductase inhibitors.
3
34
35
3
37
38
39
40
41
With the advent of highly active antiretroviral therapy (HAART) the incidence of opportunistic
infections has reduced dramatically, but non-compliance, toxicity with current treatments, emergence of
drug resistant strains, late diagnosis of HIV, and the rise of HIV cases in developing countries all
42
4
4
45
46
47
48
2
-5
contribute to the persistence of opportunistic infections as a clinical problem. The Centers for Disease
Control report in 2009 has stated that for patients with HIV, under certain circumstances, prophylaxis of
opportunistic infections may need to be continued for life and that, despite HAART, opportunistic
49
5
51
6
, 7
infections remain a leading cause of morbidity and mortality in HIV infected patients.
Pneumocystis
52
53
5
5
56
57
58
pneumonia (PCP) [caused by the fungus Pneumocystis jirovecii (pj)], toxoplasmosis [caused by the
protozoan Toxoplasma gondii (tg)] and Mycobacterium avium-intracellulare infection (MAI) [caused
by the bacteria Mycobacterium avium and Mycobacterium intracellular] are the most common
5
60
6
opportunistic infections.
ACS Paragon Plus Environment
4

Page 5 of 67
Journal of Medicinal Chemistry
Lipophilic, nonclassical antifolates (Figure 1) like trimethoprim (TMP), pyrimethamine and
1
2
trimetrexate (TMQ) are approved agents for the treatment and prophylaxis of opportunistic infectionsin
3
4
5
6
7
8
9
8
immunocompromised patients, including MAI. TMP is a selective, but weak inhibitor of
microbial/fungal DHFR. TMP with sulfamethoxazole (SMX) is the most commonly prescribed regimen
6
for opportunistic infections in AIDS patients. SMX, an inhibitor of dihydropteroate synthase, is
1
1
12
13
necessary to compensate for the weak DHFR inhibitory activity of TMP and to decrease the appearance
1
15
9-11
of resistant strains. Side effects associated with sulfonamides and dapsone
often leads to
16
1
18
19
1
2
discontinuation of therapy. TMQ is a potent, but non-specific DHFR inhibitor, used in the treatment of
13, 14
moderate to severe PCP and causes a high rate of myelosuppression
and is co-administered with
20
2
2
8
, 13, 15
leucovorin (5-formyltetrahydrofolate) as a rescue agent to prevent host cell toxicity.
However, this
23
2
25
26
27
dual therapy increases costs and the rescue is not always successful. Piritrexim (PTX) is another potent,
non-selective DHFR inhibitor that has been studied along with leucovorin in clinical trials as an
28
29
8
antiopportunistic agent in AIDS patients.
30
31
3
Given the limitations of both the TMP/sulfonamide combinations, regarding side effects and
3
34
11, 15
resistance due to the sulfa drug component
and TMQ/leucovorin regimens that do not work under
35
36
37
38
39
40
41
42
4
4
45
46
several clinical settings, a single DHFR inhibitor that combines the potency of TMQ or PTX with the
species selectivity of TMP would be highly desirable, in that it would eliminate the need to co-
administer either a sulfonamide or leucovorin with the DHFR inhibitor. However, to our knowledge, the
compounds reported thus far in the literature are not both selective and potent enough to achieve the
7
goal of producing a clinically viable single therapeutic agent.
47
48
.
49
5
51
Design of novel inhibitors:
52
53
PCP was originally thought to be caused by the fungus Pneumocystis carinii (pc), but, it is now
5
5
1
6-18
known that Pneumocystis jirovecii (pj) is responsible for PCP in humans.
The DHFR versions from
56
57
58
1
6
these two organisms differ by 38% in amino acid sequence and differ in sensitivity to TMP and MTX.
59
60
No crystal structure of pjDHFR has been reported thus far. Early analogs in this study were initially
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 67
designed as pcDHFR inhibitors that served as a surrogate for pjDHFR (albeit not a good surrogate) and
1
2
3
4
5
6
7
8
9
1
7
only in the latter stages of this work, following our isolation of pjDHFR in sufficient quantities, was
the inhibition of pjDHFR included in this study.
19-27
28-34
Gangjee et al.,
as well as others,
have reported several structural classes of DHFR
inhibitors in a quest for high potency and high selectivity against microbial/fungal DHFR over human
1
1
12
13
1
8
6
DHFR. One series of bicyclic pyrido[2,3-d]pyrimidines that is absent in the literature is the N -
14
15
(
substituted phenyl)pyrido[2,3-d]pyrimidine-2,4,6-triamines in which the substituted aniline is directly
16
1
18
19
20
2
2
attached to the 6-position of the pyrido[2,3-d]pyrimidine ring. These are analogs of PTX (Fig. 1), a
potent but non-selective DHFR inhibitor. The 5-methyl moiety of the 5-deaza-5-methyl classical as well
as nonclassical antifolates such as PTX and TMQ (Figure 1) make hydrophobic contact with Val115 in
23
2
25
3
5
36
hDHFR. Champness et al. indicated that the 5-methyl group of PTX is in van der Waals contact (≤ 4
26
27
Å) with Ile123 of pcDHFR. PTX lacks selectivity but is a potent inhibitor of both pcDHFR and
28
29
2
0
17
hDHFR. The amino acid sequence of pjDHFR has a corresponding Ile123 as well, and results in the
30
31
3
observed high potency but not selectivity (since hDHFR has a Val115).
3
34
35
36
37
38
39
40
41
42
4
4
45
46
Figure 2. Selective pathogen DHFR inhibitors.
47
48
49
5
51
37
In 2002, Cody et al. showed from X-ray crystal structure that the 5'-methoxy of 1 (Figure 2)
52
53
fits deep within the hydrophobic pocket in pcDHFR lined by Ile65, Pro66 and Phe69 (Figure 3).
5
5
56
57
58
59
60
ACS Paragon Plus Environment
6

Page 7 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
37
Figure 3. Stereoview of quinazoline inhibitor 1 in pcDHFR binding site (pdb id: 1LY3); generated using MOE
2
2
38
2
010.10. (Hydrogen atoms eliminated for clarity).
23
24
25
26
27
The corresponding amino acids in hDHFR are Ile60, Pro61 and Asn64, respectively. The
28
29
replacement of hydrophobic Phe69 in pcDHFR with the more polar Asn64 allows the design of
30
3
3
2
0, 21, 27
selective agents for pcDHFR with hydrophobic substituents.
Additionally, the N9-methyl
3
34
moiety of 1 is in van der Waals contact (≤ 4 Å) with the side chain of Ile123 of pcDHFR probably
35
36
contributing to its increased potency.
37
3
39
2
0, 25
Gangjee et al.
have reported pyrido[2,3-d]pyrimidine analogs such as 2a, 2b, and 3 (Figure
40
41
2
). Compounds 2a and 2b were selective for pcDHFR (selectivity ratios = 15.7 and 13, respectively)
42
4
2
5
and tgDHFR (selectivity ratio = 23 and 4.7, respectively) over rlDHFR, while compound 3 exhibited
4
45
46
excellent selectivity for pcDHFR (selectivity ratio = 101) and tgDHFR (selectivity ratio = 303) over
4
48
20, 27
hDHFR.
Crystallographic data of the ternary complex hDHFR-NADPH and 3 was used to explain
49
5
2
7
its selectivity for pathogen DHFR compared to hDHFR.
51
52
53
5
5
56
57
58
PTX analogs without the 5-methyl groups do not interact with hDHFR as effectively and exibit
lower potency as compared to PTX. If the relative loss of potency of these analogs against hDHFR is
greater than the loss of potency against pjDHFR (or other microbial DHFR), these compounds would be
59
60
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 67
more selective inhibitors of pjDHFR than hDHFR. This hypothesis has been validated in a previous
1
2
3
4
5
6
7
8
9
2
0, 27
report for pcDHFR and hDHFR.
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
2
30
Figure 4. Novel DHFR inhibitors 4-15
31
3
3
34
A similar approach was utilized in the design of analogs 4-15 (Figure 4) of this study in which
35
3
37
the 5-methyl moiety of PTX was removed and the 6-methylene moiety of PTX was isosterically
38
39
40
41
replaced by an NH group. Using this hypothesis, variations of electron-donating, electron-withdrawing
and bulky substituents on the phenyl side chain were implemented to explore the SAR of this novel
42
4
20
series of compounds. In the previous reports involving pyrido[2,3-d]pyrimidines and pyrido[3,2-
d]pyrimidines as DHFR inhibitors against opportunistic infections, it has been shown that compounds
with electron rich 3',4',5'-trimethoxy and 2',5'-dimethoxy substituted phenyl rings were usually more
selective (20- to 304-fold for pathogen (tg, E. coli) DHFR. Hence compounds 5-8 were designed as
possible selective inhibitors. These compounds could interact with Phe69 of pcDHFR over Asn64 in
4
45
46
47
48
49
50
51
52
53
5
5
56
3
7
hDHFR and afford selectivity.
2
0, 23
57
58
In addition, compounds with bulky substituents such as 1-naphthyl and 2-naphthyl rings
59
60
were found to be more potent DHFR inhibitors as compared to those with an unsubstituted phenyl ring.
ACS Paragon Plus Environment
8

Page 9 of 67
Journal of Medicinal Chemistry
Hence compounds 9-11 and 13 were included in this study to explore the possibility of improving the
overall potency while maintaining selectivity. The naphthyl group can also interact with Phe69 of
pcDHFR and not with Asn64 in hDHFR, and could afford selective and potent inhibitors of pcDHFR.
1
2
3
4
5
6
7
8
9
25
In the pyrido[2,3-d]pyrimidine series of compounds Gangjee et al. showed that the 2',5'-
dichlorophenyl substituted compound 2a (Figure 2) was the most selective inhibitor of both pcDHFR
1
1
12
13
(
selectivity ratio = 15.7) and tgDHFR (selectivity ratio = 23) versus rlDHFR. Additionally, in a recent
1
15
16
17
18
19
20
2
2
23
24
25
26
27
17
report by Cody et al. this pyrido[2,3-d]pyrimidine analog with a 2',5'-dichlorophenyl ring was
significantly potent and selective for pjDHFR (IC50 = 0.172 µM; selectivity ratio = 33) over rlDHFR. In
the same series of compounds a 3',4',5'-trichlorophenyl substituted pyrido[2,3-d]pyrimidine 2b (Fig. 2)
also exhibited excellent potency and selectivity for pjDHFR (IC50 = 0.0534 µM; selectivity ratio = 23)
over rlDHFR. Hence compounds 12, 14 and 15 were included to determine the effect of similar electron
withdrawing substituents on potency and selectivity against pjDHFR.
28
29
The truncated bridge (one N atom vs. the usual two atom C-N bridge) between the heterocycle
30
3
3
2
0-25
and the aryl ring of these novel series of compounds
has less flexibility as compared to TMQ. This
3
34
35
36
is expected to decrease the number of possible conformations and hence may increase selectivity and/or
potency of these compounds.
37
38
39
40
41
Design of compounds 16-26:
42
4
4
45
Transposition of the 5-methyl group of PTX and TMQ (Figure 1) to the N9-position in
pyrido[2,3-d]pyrimidine analogs (Figure 2), in general improves potency as well as selectivity against
46
47
48
2
0
pathogen (tg, E. coli) DHFR.
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 67
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
Figure 5. N9-methyl analogs 16-26
2
2
23
24
A crystal structure of 3 (Figure 2) complexed with hDHFR showed that the N9-methyl moiety is
25
26
27
28
29
30
31
3
3
34
35
36
2
7
not in hydrophobic contact with any hydrophobic side chain of hDHFR. In addition, in the molecular
model of 3 in pcDHFR the N9-methyl moiety of 3 forms hydrophobic contacts with both Ile123 and
Ile65 of pcDHFR. Additional hydrophobic interactions of the N9-methyl moiety of 3 with pcDHFR
compared to the lack of hydrophobic interactions with hDHFR was one of the possible reasons for the
increased potency and selectivity of 3 for pcDHFR compared with hDHFR.
37
3
39
40
41
42
4
Thus compounds 16-26 (Figure 5) were designed by incorporating a N9-methyl group. This
methyl group was anticipated to interact (Figure 6) with Ile123 of pcDHFR (and also with the
corresponding Ile123 of pjDHFR) and not with the shorter Val115 in hDHFR in a similar way as that of
4
4
2
7
3, thus affording both potency and selectivity for pcDHFR and pjDHFR.
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
1
0

Page 11 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
1
18
3
7
Figure 6: Stereoview. Docked pose of 16 in pcDHFR binding pocket (PDB: 1LY3 ).
19
20
2
2
To validate our hypothesis prior to biological evaluation, we docked 16 into the pcDHFR active
23
2
3
9
site (PDB: 1LY3) using LeadIT 1.3.0 (details in Experimental). Figure 6 shows the docked pose of 16
in the pcDHFR active site. In Figure 6 the N9-methyl group is 3.94 Å from the terminal methyl group of
Ile123 of pcDHFR. In hDHFR, the corresponding Val115, being shorter by one carbon, may not interact
with the N9-methyl group. This should improve selectivity as well as potency of these compounds
against pcDHFR (and pjDHFR) over hDHFR. In addition the N9-methyl group of 16 is only 3.8 Å away
from side chain Ile65 in pcDHFR (Fig. 6), and may improve potency by hydrophobic interactions.
Further, the N9-methyl group restricts the number of possible conformations of the side chain phenyl
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
2
0, 40
group, thus perhaps contributing to increasing potency and selectivity.
Compounds with an N9-
42
4
4
45
46
47
48
49
50
methyl group are more lipophilic than the corresponding N9-H analogs; thus improving hydrophobicity
4
0
and perhaps cell penetration. Thus compounds 16-26 with the N9-methyl group and various electron
donating, electron withdrawing and bulky hydrophobic substituents were included to explore the SAR
and to afford potent and selective inhibitors for pcDHFR and pjDHFR.
51
52
53
5
5
Chemistry:
56
57
58
The synthesis of target compounds 4-26 were envisioned from the dipivaloyl-protected 6-
59
60
bromopyrido[2,3-d]pyrimidine intermediate 30 (Scheme 1).
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 67
Scheme 1: Synthesis of 30
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
Compound 30 was synthesized from 2,4,6-triaminopyrimidine 27 via the reported method of
17
1
19
4
1, 42
Gangjee et al.
(Scheme 1). Compound 27 was dissolved in absolute ethanol at reflux with a drop
20
2
wise addition of concentrated HCl (just sufficient to dissolve 27) followed by the addition of bromo-
2
23
24
25
malonaldehyde 28 at reflux to obtain 29. Both amino groups of 29 were converted to pivaloyl amides by
reacting 29 with trimethylacetic anhydride in pyridine at reflux to afford 30. Since both amino groups of
26
27
28
3
0 are protected as pivaloyl amides and since these pivaloyl groups afforded excellent solubility in non-
29
3
polar solvents, 30 was an excellent substrate for metal-catalyzed coupling reactions for the synthesis of
31
3
3
the target compounds.
34
35
4
3, 44
Initial attempts at amination of 30 using a standard Buchwald-Hartwig coupling protocol
36
3
38
39
4
5
[Pd(OAc)
DMF, Na
2
, (±)-BINAP, NaOt-Bu, Toluene, reflux] or copper (I) catalyzed aminations (CuI, L-Proline,
2
CO , 120 °C) were unsuccessful. Hence a systematic screening approach was undertaken to
3
40
41
42
4
4
determine if the Buchwald-Hartwig amination of dipivaloyl-protected 6-bromopyrido[2,3-d]pyrimidine
30 was possible and if so, to determine the optimum conditions.
45
46
47
48
49
50
51
52
53
5
5
56
57
58
59
60
ACS Paragon Plus Environment
1
2

Page 13 of 67
Journal of Medicinal Chemistry
Development of a coupling protocol for Buchwald-Hartwig coupling of 30 with substituted
anilines:
1
2
3
4
5
6
7
8
9
PPh2
P(C H
)
P(C H
)
6
11 2
6
11 2
PPh2
i-Pr
i-Pr
O
PPh2
PPh2
1
1
12
13
i-Pr
X-Phos
rac-BINAP
Xantphos
DCPB
14
15
16
17
18
P[CH ) ]
P(C H
)
3
3 2
P(C H
)
6
11 2
P(C H )
6
6
11 2
11 2
N(Me)2
MeO
OMe
i-PrO
Oi-Pr
i-Pr
i-Pr
19
20
2
2
i-Pr
t-ButylX-Phos
S-Phos
RuPhos
DavePhos
23
24
25
Figure 7: Ligands for amination reaction
26
27
28
29
A series (Figure 7) of bidentate [racemic-2,2'-bis(diphenylphosphino)-1,1'-binaphthyl (rac-
30
31
3
BINAP), 9,9-dimethyl-4,5-bis(diphenylphosphino)xanthene (XantPhos)] and monodentate-biphenyl
3
34
[dicyclohexylphospho biphenyl (DCPB), 2-dicyclohexyl-phosphino-2',4',6'-triisopropylbiphenyl (X-
35
3
Phos), 2-di-t-butylphosphino-2',4',6'-tri-i-propyl-1,1'-biphenyl (t-Butyl-X-Phos), 2-
37
38
39
dicyclohexylphosphino-2',6'-dimethoxybiphenyl (S-Phos), 2-dicyclohexylphosphino-2',6'-di-i-propoxy-
40
41
1
,1'-biphenyl (Ru-Phos), 2-(dicyclohexylphosphino)-2'-(N,N-dimethylamino)biphenyl (DavePhos)]
42
4
4
4
46
were selected on the basis of previous successful literature reports and to vary the bulk, electronics and
46-48
chelating abilities of the ligands.
47
48
49
50
51
52
53
5
Reaction of 30 with p-anisidine 31b to form the coupled product 32b was used as a model
reaction for ligand-screening (Scheme 2). It was gratifying to note that the biphenyl ligands were
effective and X-Phos was the most suitable ligand for carrying out this conversion. The results of the
ligand-screening experiments are shown in Chart 1.
5
56
57
58
59
60
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 67
1
2
3
4
5
6
7
8
9
Scheme 2: Amination reactions of 30 with substituted anilines – Method development
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
Chart 1: Ligand-screening experiments
36
3
38
3
40
41
42
a
Isolated yields. [in all cases except with DCPB (resulted in debromination), unreacted starting material was also
dba ; 8 mol % monodentate
ligand / 4 mol % bidentate ligand; 1.4 equivalents NaOt-Bu; Toluene, 100°C, 20 h).
isolated]. (conditions: 1 equivalent 30; 1.2 equivalent p-anisidine 31b; 2 mol % Pd
2
3
4
4
45
46
47
48
49
50
51
52
53
5
5
56
For rac-BINAP, XantPhos and DavePhos, unreacted starting material was isolated (90-95 %);
while when DCPB was used, a majority of the staring material underwent debromination. For X-Phos,
S-Phos, t-ButylX-Phos and Ru-Phos, unreacted starting material was isolated (40-85 %) even after 48
hours of stirring; which suggests that the catalyst degraded before complete consumption of the starting
material.
57
58
59
60
ACS Paragon Plus Environment
1
4

Page 15 of 67
Journal of Medicinal Chemistry
Hence it was decided to optimize the yields by increasing the catalyst loading. Increasing the
1
2
catalyst loading up to 15 mol % (with respect to Pd
Variation of the base used was attempted next in an effort to improve the reaction yields (Table 1).
Cesium carbonate (Cs CO ) did not aid in the reaction and starting material was isolated almost
2 3
dba ) improved the yield from 45 % to 60 %.
3
4
5
6
7
8
9
2
3
quantitatively (entries 4 and 5). Addition of 2.2 equivalents of lithium hexamethyldisilazide (LiHMDS)
led to slight improvement (entry 6) in yield (53%). An additional one equivalent of LiHMDS improved
the yield to 91% (entry 7).
1
1
12
13
14
15
16
17
18
19
20
Table 1: Selection of a suitable base
2
2
Entry Base
Equivalents Isolated yield (%)
23
2
25
26
27
28
29
30
31
3
3
34
35
3
37
38
39
1
2
3
4
5
6
7
8
NaOt-Bu 1.5
NaOt-Bu 2.5
45
48
43
0
NaOt-Bu
3
Cs
Cs
2
2
CO
3
3
1.6
2.6
CO
0
LiHMDS 2.2
LiHMDS 3.2
LiHMDS 3.7
53
91
90
40
41
42
4
4
45
46
(
Conditions: 1 equivalent 30; 1.2 equivalent p-anisidine 31b; 2 mol % Pd
2 3
dba ; 8 mol % X-Phos; # of equivalents
47
48
of Base; Toluene, 100°C, 20 h).
49
50
51
52
53
5
5
56
Additional NaOt-Bu, however, did not improve the yield significantly (entries 2 and 3; Table 1).
This also indicated that deprotonation is not the rate determining step of this reaction. These results
suggested that premature deactivation of the catalyst probably occurs during the reaction and that the
addition of an extra equivalent of LiHMDS prevents this deactivation. Similar observations have been
57
58
5
60
4
9
reported by Buchwald et al. and suggested that the lithiate or a lithium-amide aggregate of the
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 67
deprotonated amide from the substrate functions as a protecting group, thus inhibiting oxygen from
binding to the palladium center, and preventing its deactivation.
1
2
3
4
5
6
7
8
9
This strategy was applied for the synthesis of 32a-l (Table 2) in significantly better yields.
Table 2: Amination with substituted anilines
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
Entry Comp. No. R
1
Isolated
26
27
Yield (%)
28
29
30
31
3
3
3
35
36
37
38
39
40
4
42
4
4
45
4
47
48
49
50
51
52
5
5
5
56
1
2
3
4
5
6
7
8
9
1
32a
32b
32c
32d
32e
32f
32g
32h
32i
H
85
88
56
55
4'-OMe
2'-OMe
2',5'-diOMe
3',4',5'-triOMe 78
4'-Me
93
91
75
60
65
78
75
4'-i-Pr
3',4'-(C
4
H
4
4
)
)
4'-NO
2
0
32j
32k
32l
2',3'-(C
4
H
11
2',5'-diF
3',4',5'-triF
1
2
57
58
59
60
ACS Paragon Plus Environment
1
6

Page 17 of 67
Journal of Medicinal Chemistry
Anilines with electron-donating substituents (entries 2, 6-7; Table 2) afforded better yields;
1
2
while those with electron-withdrawing substituents (entries 9, 11-12) led to slightly lower yields. Bulky
ortho substituents on the anilines (entries 3-4, 10) probably retard the rate of the reaction, leading to
inferior yields.
3
4
5
6
7
8
9
1
In all cases, the H NMR spectra of the products showed the appearance of a new broad singlet
1
1
12
13
2
(D O exchangeable) between δ 5.9- δ 6.4 integrating for 1 proton and corresponding to the 9-NH
14
15
16
17
18
19
20
2
2
23
24
25
proton. Additionally aromatic protons of the coupled arylamine, in the aromatic region, confirmed the
formation of the products. All the products were confirmed by HRMS analysis.
Amination of 30 with N-methylanilines using above protocol, however, did not lead to complete
consumption of the starting material even after prolonged reaction time (entries 1 and 2; Table 3).
Table 3: Amination using N-methyl-4-methoxy-anilines
26
27
28
29
30
31
3
3
34
35
3
37
38
39
Entry Ligand
Time (h) Results; Isolated
yield
40
41
1
X-Phos
20
Incomplete;
42
4
4
Inseparable
4
46
47
48
a
2
3
X-Phos
S-Phos
48
20
Incomplete; 57 %
Complete; 88 %
49
50
51
5
53
a
An additional aliquot of Pd
2
dba
3
(2 mol %) and X-Phos (8 mol %) premixed with
5
5
toluene was added after 24 h.
56
57
58
59
60
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 67
It was hypothesized that the methyl group on the aniline nitrogen may lead to steric hindrance to
the approach of the palladium complex involving the bulky X-Phos ortho-isopropyl groups. Hence S-
Phos with smaller ortho-methoxy groups was employed. Indeed, this led to the complete conversion of
the starting material with excellent isolated yield of the desired 32n. (Table 3, entry 3).
Table 4: Amination with substituted N-methylanilines
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
2
24
25
Entry Comp. No.
R
1
Isolated yield
%)
(
26
27
28
29
3
31
3
3
34
35
36
37
38
39
40
41
4
4
4
45
46
47
48
49
50
1
2
3
4
5
6
7
8
9
10
32m
32n
32o
32p
32q
32r
32s
H
83
90
59
55
81
96
93
63
56
93
90
4'-OMe
2'-OMe
2',5'-diOMe
3',4',5'-triOMe
4'-Me
4'-i-Pr
32t
4'-NO
2',3'-(C
2',5'-diF
3',4',5 '-triF
2
32u
32v
32w
4 4
H )
51
52
1
1
53
5
5
5
57
This strategy was applied for the amination reactions of 30 with substituted N-methylanilines to
5
59
1
synthesize 32m-w in good to excellent yields (Table 4). The H NMR spectra of the products showed
60
ACS Paragon Plus Environment
1
8

Page 19 of 67
Journal of Medicinal Chemistry
3
the presence of a sharp singlet between δ 3.4 – δ 3.6 integrating for 3 protons of the NCH group. All
1
2
3
4
5
6
7
8
9
the products were confirmed using HRMS analysis.
Scheme 3: Synthesis of target compounds 4-26
1
1
12
13
14
15
16
17
18
19
20
2
2
23
3
Pivaloyl deprotection of compounds 32a-w was carried out with liquid NH in a Parr pressure
24
25
1
vessel at room temperature to afford target compounds 4-26 (Scheme 3). The H NMR spectra of the
26
27
28
products showed the disappearance of a doublet (~ δ 1.3) corresponding to the pivaloyl groups and the
29
30
2
presence of two broad D O exchangeable peaks corresponding to the 2,4-diamino groups.
31
3
3
3
35
Scheme 4: Synthesis of substituted N-methylanilines
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
Substituted N-methylanilines 31p-q, 31s and 31v-w (Scheme 4) were synthesized via a one pot
reaction in which the corresponding substituted anilines 31d-e, 31g and 31k-l respectively were reacted
with paraformaldehyde in the presence of sodium methoxide followed by treatment with sodium
5
5
50
1
borohydride in methanol at reflux. The H NMR spectra of the products showed the presence of a
5
5
singlet integrating for 3 protons at around δ 2.8 corresponding to the N-methyl moiety.
57
58
59
60
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 67
1
2
3
4
5
6
7
8
9
Biological evaluation and Discussion:
Compounds 4-26 were evaluated as inhibitors of recombinant pjDHFR, recombinant hDHFR,
pcDHFR, rlDHFR and maDHFR. Selectivity ratios were determined using recombinant hDHFR or
1
1
12
13
14
15
using rlDHFR as a mammalian DHFR source. MTX, TMP, PTX, pyrimethamine, and TMQ were used
as positive controls in the assays.
16
17
18
1
20
2
2
Table 5: Inhibitory concentrations (IC50, in µM) against recombinant DHFR from P. jirovecii (pj); human (h) and
a
selectivity ratios
23
24
Compound
pjDHFR hDHFR h/pj
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
µM
µM
IC50
0.19
3.65
0.25
1.0
IC50
0.30
0.40
0.44
0.18
0.31
0.62
1.1
4
0.6
9.1
0.6
5.6
13
5
6
7
8
3.9
41
42
4
4
4
46
9
2.1
3.3
5.6
5.4
2.0
8.2
0.5
3.5
10
5.9
47
48
49
5
51
52
53
5
5
56
5
58
1
1
0.40
3.6
2.2
12
7.2
1
1
3
4
0.25
0.95
0.87
2.1
0.45
3.1
15
59
60
ACS Paragon Plus Environment
2
0

Page 21 of 67
Journal of Medicinal Chemistry
1
1
1
1
6
7
8
9
0.0022
0.0019
0.0016
0.0021
0.0016
0.0024
0.0024
0.028
0.057
0.012
0.014
26
1
2
3
4
5
6
7
8
9
1
1
12
13
14
1
16
17
18
19
20
2
2
23
24
25
26
2
28
29
30
31
3
3
34
35
36
37
6.6
9.3
0.0047
0.0046
0.032
0.052
0.34
2.2
2.8
13
20
2
1
22
21
2
2
3
4
12
0.0025
0.0034
0.0042
0.12
0.0056
0.046
0.15
2.2
14
25
2
6
35
TMP
MTX
PTX
32.2
268
0.34
1.9
1.2
3.1
0.0051
0.0016
0.0021
0.0017
0.003
0.0026
0.40
TMQ
pyrimethamine 0.13
38
39
40
41
42
4
4
45
4
a
These assays were carried out at 37 °C under 18 µM dihydrofolic acid concentration. The assay also
47
48
49
contains 117 µM NADPH, 8.9 mM 2-mercaptoethanol, 150 mM KCl, 41 mM Na phosphate buffer pH
50
51
7
.4 and sufficient enzyme to cause a change in OD340 of 0.005/minute. The standard error of the means
52
53
5
for these values are 12% or less than the mean value.
5
56
57
58
59
60
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 67
The IC50 values and selectivity ratios of compounds 4-26 against recombinant pjDHFR and
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
recombinant hDHFR are shown in Table 5. Standard compounds show the properties expected, based on
prior studies: TMP (IC50 0.12 M) and pyrimethamine are not potent inhibitors of pjDHFR but TMP is
the most selective compound, whereas MTX, PTX and TMQ are all potent but relatively non-selective.
Experimental compounds with N9-H exhibited submicromolar to micromolar inhibition of pjDHFR and
hDHFR. Compounds with electron-donating substituents at the meta- and/or para- positions on the
phenyl side chain (5, 7-10) were less potent against hDHFR compared to the unsubstituted phenyl
compound 4. While against pjDHFR these compounds retained potency (5, 7, 8) or exhibited relatively
less loss of potency than that against hDHFR (9-10) as compared to 4 thus affording improved
selectivity ratio (8; selectivity ratio = 12.7).
2
23
24
25
26
27
28
29
Compounds with bulky side chains (11 and 13) retained potency against pjDHFR, and exhibited
a 10-fold loss in potency against hDHFR leading to improved selectivity for the pathogen DHFR
compared to the unsubstituted phenyl compound 4.
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
Compounds with electron deficient side chains exhibited diminished potency against pjDHFR
indicating the importance of electronics for potency. Compound 15 (3',4',5'-trifluorophenyl) showed
slightly better selectivity for pjDHFR (selectivity ratio = 3.5) compared to the unsubstituted phenyl
compound 4. Compound 14 (2',5'-difluorophenyl) showed poor selectivity and 12 (4-nitrophenyl)
exhibited diminished inhibitory activities against both pjDHFR and hDHFR. Substitution of electron-
withdrawing groups on the side chain phenyl ring were not conducive for potency or selectivity as
compared to substitutions with electron-donating groups.
47
48
49
50
51
52
53
5
5
56
57
58
Inclusion of a methyl group at the N9- position led to a remarkable 100-fold improvement in
potency with superior selectivity. Compound 26 (3',4',5'-trifluorophenyl side chain) was the most
selective compound of the series. The selectivity may be attributed, in part, to the possible interaction of
the N9-methyl with the longer side chain of Ile123 of pjDHFR; while interaction with the shorter
Val115 in hDHFR may not be as productive.
59
60
ACS Paragon Plus Environment
2
2

Page 23 of 67
Journal of Medicinal Chemistry
A similar trend in selectivity was observed for compounds 16, 18, 21-23, and 25. However
1
2
compounds 17, 19, 20 and 24 exhibited slightly diminished selectivity compared to the corresponding
N9-H compounds 5, 7, 8 and 11. Thus, the presence of the methyl group at the N9- position perhaps
dictates the conformation of the bound inhibitor and its binding energy and is different for 18, 21 – 23
and 25 – 26 compared with 17, 19, 20, and 24.
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
Compounds with methoxy groups at ortho-, meta- and para- positions of the side chain phenyl
ring exhibited improved potency against pjDHFR. However, these compounds also showed marked
improvement in potency against hDHFR leading to diminished overall selectivity for the pathogen
DHFR as compared to the unsubstituted phenyl compound 16. Compound 18 (2'-methoxyphenyl side
chain) was the most potent compound against pjDHFR with good selectivity (IC50 = 1.55 nM;
selectivity ratio = 9.3). Compounds with alkyl substituents as well as bulky substituents retained
potency, while those with electron-withdrawing substituents exhibited a slight loss of potency against
pjDHFR. Compound 26 (3',4',5'-triflurophenyl side chain), exhibited a marked improvement in
selectivity for pjDHFR.
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
Analogs 16-26 have high potencies against pjDHFR, similar to the most potent standard
compounds MTX, PTX, but have the advantage of significantly enhanced selectivity (16, 22, 26) against
pjDHFR. The most selective compounds are about 10-fold less selective than TMP but are about 50-fold
more potent (16). Analogs 16, 22 and 26 are lead compounds for the development of potent and
selective inhibitors of pjDHFR.
45
46
47
48
Docking Studies with pjDHFR
49
50
51
52
There is currently no known crystal structure of pjDHFR. A homology model was built using the
53
5
5
5
1
1
.60 Å pcDHFR crystal structure (PDB: 2FZI , chain A) for evaluating the binding of 26 in pjDHFR.
5
57
39
Docking studies were performed using LeadIT 1.3.0. The docked poses were exported to MOE
58
5
60
3
8
2010.10, rescored using the affinity dG scoring system, refined using the forcefield system and
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
rescored using London dG scoring system. The binding poses were also visualized using the ligplot
utility in MOE and the Poseview utility in LeadIT 1.3.0.
Page 24 of 67
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
Figure 8: Stereoview. Docked pose of 26 in pjDHFR homology model.
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
Figure 9: Stereoview. N-Me of 26 interacts with Ile123 (pjDHFR, red) and not as well with Val115
4
47
35
(
superimposed hDHFR, green).
48
49
50
51
Figure 8 shows the best docked pose of 26 (white) in the folate binding site of the pjDHFR
52
53
5
2
model. In this pose the protonated N1 and 2-NH of 26 interact in an ionic bond with Asp32. This
5
56
bidentate ionic bond with a conserved acid residue has been observed in most DHFR crystal
57
5
5
2
structures. The 4-NH
2
moiety forms hydrogen bonds with the backbone of Ile10. The pyrido[2,3-
59
60
ACS Paragon Plus Environment
2
4

Page 25 of 67
Journal of Medicinal Chemistry
d]pyrimidine scaffold is stabilized by a pi-stacking interaction with Phe36 and with side chain carbon
atoms of Met33 (not shown) and Leu25. The 3′,4′,5′-trifluorophenyl moiety of 26 can form hydrophobic
interactions with the side chain atoms of Leu25, Thr61, Ser64, and Leu65. The 3′-F of 26 interacts with
the hydroxyl moiety of Ser64. The N9-Me moiety of 26 is oriented towards the hydrophobic pocket
formed by Ile123, Leu72 and Leu65 and interacts specifically with Ile123. The docking score of 26 was
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
-
(
35.98 kJ/mol for pjDHFR compared with -35.41 kJ/mol for hDHFR. The docking scores for TMP
-23.23 kJ/mol), MTX (-58.24 kJ/mol) and pyrimethamine (-29.36 kJ/mol) are consistent with the
16
1
18
19
20
2
2
23
24
25
26
27
28
29
corresponding observed inhibitory activities against pjDHFR. Compound 16 makes similar binding
with pjDHFR except for the trifluoro groups in 26. It was of interest to attempt to explain the potency
and selectivity of 16 and 26 (compared to 4 and 15 respectively) for pjDHFR. From the IC50 values in
Table 5 for the two pairs, 16 / 4 and 26 / 15, it is evident that the N9-Me moiety makes about a 300- and
a 200-fold difference in potency for pjDHFR respectively. There are two important consequences of
adding the Me group on the N9. The first one is the interaction of the N9-Me with Ile123 in pjDHFR
30
31
3
(Fig. 9). This interaction is absent in the N9-H analogs, 4 and 15 from modeling studies. The second
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
significance from modeling studies (SYBYL x1.3,; around the C6-C9 bond, using 5 increments) is that
the N9-Me restricts the available low energy conformations (50 conformations, within a 3 k cal/mol cut
off) that the molecule can adopt compared with the unhindered N9-H (94 conformations). Thus it is
perhaps easier for 16 and 26 to adopt the bound conformation than it is for 4 and 15 and results, in part,
in the improved IC50 of 16 and 26 over 4 and 15. In addition to potency, the selectivity of 16 and 26 for
pjDHFR over hDHFR are superior to that of 4 and 15. Once again, the N-Me moiety must play a
significant role in the high selectivity of 16 and 26 for pjDHFR over hDHFR compared to 4 and 15. In
pjDHFR the N9-Me moiety is 3.67 Å away from the longer Ile123 compared to hDHFR where it is 4.64
Å away from the shorter Val115. These superimpositions of the docked structure of 26 in the pjDHFR
homology model and hDHFR crystal structure (Fig. 9) and the highly productive interaction of the N9-
Me with Ile123 at 3.67 Å and a somewhat less of a productive interaction of the N9-Me with the shorter
57
58
59
60
Val115 of hDHFR affords, in part, a molecular explanation of the pjDHFR selectivity of 26 over 15.
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 67
Confirmation of these homology model predictions await the X-ray crystal structures of these analogs
1
2
3
4
5
6
7
8
9
with pjDHFR and hDHFR, currently in progress.
Table 6: Inhibitory concentrations (IC50, in µM) against DHFR from P. carinii (pc); rat liver (rl); M.
a
avium (ma) and selectivity ratios
1
1
12
13
14
1
Compd.
pcDHFR
rlDHFR
µM IC50
maDHFR
µM IC50
µM IC50
#
rl/pc
0.2
0.7
0.1
0.5
0.9
0.5
2.4
1.3
0.8
1.9
0.1
0.6
1.6
2.5
1.0
1.0
3.0
1.1
rl/ma
0.3
1.6
0.3
1.0
2.7
0.9
1.4
1.8
1.5
2.9
0.4
1.4
4.6
2.8
3.5
1.7
12
16
17
1
19
20
2
2
23
24
2
26
27
28
29
30
31
3
3
34
35
36
37
38
39
40
41
42
4
4
45
46
47
48
4
2.3
0.42
3.9
1.2
5
5.4
2.5
6
7.3
0.76
1.7
2.4
7
3.3
1.7
8
4.6
4.0
1.5
9
7.7
4.0
4.4
10
4.9
12
8.7
1
1
1
1
1
2
3
4
2.5
3.3
1.8
16
12
7.9
1.5
2.9
1.0
7.7
1.1
3.3
15
9.0
5.2
3.6
1
1
6
7
0.076
0.0024
0.023
0.0036
0.0023
0.025
0.12
0.0059
0.024
0.0037
0.0068
0.028
0.027
0.0021
0.0067
0.0022
0.00057
0.027
49
50
5
52
53
5
5
56
18
1
2
9
0
57
5
59
21
1.1
60
ACS Paragon Plus Environment
2
6

Page 27 of 67
Journal of Medicinal Chemistry
2
2
2
2
2
3
4
5
0.021
0.58
0.072
0.61
3.4
1.1
9.9
0.7
1.9
14
0.026
2.7
0.9
10
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
1
18
19
20
2
0.15
0.0021
0.098
0.23
0.021
0.070
0.43
0.0020
0.017
3.9
15
26
0.029
TMP
MTX
PTX
TMQ
12.5
180
0.0015
0.00022
0.00061
120000
11
0.0013
0.013
0.042
2.4
0.0025
0.0033
0.003
1.5
1.9
0.3
5
0.07 0.0015
0.6 0.15
2
PYR
10
2
23
2
a
These assays were carried out as described for Table 5, except 90 µM dihydrofolic acid was used.
25
26
27
The standard error of the means for these values are 12% or less than the mean value.
28
29
30
31
3
3
34
35
36
37
Compounds 4-26 were also evaluated in an earlier screen against rlDHFR, pcDHFR, and
maDHFR (Table 6). These data are presented here in order to evaluate the potential of the experimental
compounds as single selective agents with activity against another important opportunistic pathogen
38
39
(MAI) and to demonstrate the differences in the original screen using pcDHFR with rlDHFR, versus the
40
41
modern screen using pjDHFR with hDHFR.
42
4
4
45
46
47
48
49
50
51
52
53
5
5
56
57
58
In the screen using pcDHFR and rlDHFR, compound 24 was the most selective compound
against pcDHFR (selectivity ratio = 9.9). Compounds with bulkier side chains (10, 11, 13, 22 and 24)
exhibited marginal selectivity for pcDHFR. Inclusion of a methyl group at N9- position improved
potency against both pcDHFR and rlDHFR. However, relative improvement in potency against
pcDHFR was greater compared to that against rlDHFR leading to a somewhat improved selectivity of
the N9-methyl compounds 16-26 compared to the corresponding N9-H compounds (4-10, and 12-15).
In the N9-H series of compounds (4-15) substitution of electron-donating as well as electron-
59
60
withdrawing groups on the phenyl side chain led to diminished potency against pcDHFR. Compounds
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 67
with bulky substituents on the side chain phenyl ring retained potency against pcDHFR compared to 4.
Additionally these compounds with bulky substituents exhibited improved selectivity for pcDHFR over
rlDHFR (10, 11, and 13). In the N9-methyl series of compounds (16-26) substitution of electron-
donating or bulky groups on the side chain phenyl ring afforded compounds with marked improvement
in potency (3- to 35-fold) against pcDHFR compared with analogs lacking these substitutions.
Compounds with electron withdrawing groups, however, exhibited decreased inhibitory activity against
pcDHFR. Generally, compounds with bulkier substituents on the phenyl side chain exhibited a
combination of superior potency as well as selectivity (13, 22 and 24). This potency and selectivity can
be attributed to the hydrophobic interaction with Phe69 in pcDHFR; while a similar interaction with
mammalian DHFR is not condusive because of the presence of a corresponding Asn64 in mammalian
DHFR. Compound 24 (2-naphthyl side chain) was the most potent as well as the most selective
compound against pcDHFR (IC50 = 2.1 nM; selectivity ratio = 9.9). This compound is as selective as
TMP and is additionally 5952-fold more potent than TMP against pcDHFR (IC50 12.5 M). It is also 6-
fold more potent than PTX against pcDHFR. These differences of potency and selectivity of analogs
against pjDHFR and pcDHFR are included in this study to further underscore the differences between
rat derived pathogen DHFR (pc) and the human derived pathogen DHFR (pj).
1
2
3
4
5
6
7
8
9
1
1
12
13
14
15
16
17
18
19
20
2
2
23
24
25
26
27
28
29
30
31
3
3
34
35
36
37
38
39
Against maDHFR, all the compounds from the N9-H series (4-15) exhibited micromolar
40
41
42
4
4
45
46
47
48
49
50
51
52
53
5
5
inhibitory potency. Any variation from the unsubstituted phenyl ring led to slight decrease in potency.
Exception to this trend was 13 substituted with 2',3'-naphthyl side chain, which exhibited slightly
improved inhibitory potency against pcDHFR compared to compound 4 with an unsubstituted phenyl
ring. Electron-donating as well as electron-withdrawing groups at the ortho position of the side chain
phenyl group led to diminished selectivity for maDHFR (6, 7 and 14). Compounds with substitution on
the side chain phenyl group at the para and meta positions showed improved selectivity for maDHFR
(5, 8, 10-12 and 15). Compounds with bulky substituents exhibited good potency as well as selectivity
56
57
58
(
13). Compounds from the N9-methyl series (16-26) were potent inhibitors of maDHFR. Compounds
59
60
with electron-donating substituents on the side chain phenyl ring (17-20) exhibited marked
ACS Paragon Plus Environment
2
8

Page 29 of 67
Journal of Medicinal Chemistry
improvement (2- to 45-fold) in potency against maDHFR. Compound 23 with an electron-withdrawing
-NO2 group on the side chain phenyl ring showed diminished potency against maDHFR as compared
1
2
3
4
5
6
7
8
9
1
1
12
13
4
to the unsubstituted compound 16. Compounds with bulkier side chain retained potency. It appears that
substitution of electron-donating groups at the para- and meta- positions, (17 and 20), on the side chain
phenyl ring leads to improvement in potency as well as selectivity. In this series, compound 20 (3',4',5'-
trimethoxyphenyl side chain) was the most potent compound against maDHFR with subnanomolar
potency (IC50 = 0.573 nM) as well as good selectivity for maDHFR.
14
15
16
1
18
19
20
2
2
23
24
25
26
27
28
29
30
31
While some similarities were noted in the order of potency observed between the screen using
pjDHFR/hDHFR and the screen using pcDHFR/rlDHFR, there were also clear differences. For
example while compound 22 rates highly in both screens for selectivity, compound 24 would have been
selected as a candidate based upon the pcDHFR/rlDHFR screen, but does not rank high in the
pjDHFR/hDHFR screen. These results confirm that the differences in structure between pjDHFR and
pcDHFR significantly impact drug design and argue for rescreening of earlier libraries of compounds,
an activity that is now underway.
3
3
34
35
3
37
38
X-ray crystal structures (PDB Accession Numbers: pcDHFR-K37S/F69N-#16 (4IXG); pcDHFR-
F69N-#22 (4IXF); pcDHFR-#26 (4IXE))
39
40
41
42
4
Structural data were measured for ternary complexes of NADPH and native Pneumocsytis
carinii (pc) DHFR, as well as their single (F69N) and double (K37S/F69N) variants, with inhibitors 26,
4
45
46
1
6 and 22, respectively, to validate the binding interactions of these inhibitors in the active site of
47
48
pcDHFR (Figs. 10 and 11 electron density). The overall characteristics of the two variant pcDHFR
49
5
51
ternary complexes with 16 and 22 are similar to those reported for 1 and preserve the overall pattern of
5
53
37, 52
contacts with invariant active site residues (Figures 1S and 2S)
The overall structure of pcDHFR
5
5
56
57
with 26 differs in the conformation of the flexible loop regions encompassing residues 43-49 and 82-91
Figure 12). Structural data for the majority of pcDHFR inhibitor complexes reveals that loop 83-89 is
(
58
59
60
disordered with poor electron density indicating the high flexibility of this region. In the case of the
ACS Paragon Plus Environment
2
9