Discovery and SAR of a series of 4,6-diamino-1,3,5-triazin-2-ol as novel non-nucleoside reverse transcriptase inhibitors of HIV-1

Article abstract of DOI:10.1016/j.bmcl.2010.09.034

The discovery and SAR study of a series of 4,6-diamino-1,3,5-triazin-2-ol compounds as novel HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs) are reported. The lead compounds in this series showed excellent activity against wild-type and drug-resistant RT enzymes and viral strains. In addition, compounds from this series demonstrated favorable pharmacokinetic profile in rat. A preliminary modeling study was conducted to understand the binding mode of this series of compounds.

Full text of DOI:10.1016/j.bmcl.2010.09.034

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6592–6596
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and SAR of a series of 4,6-diamino-1,3,5-triazin-2-ol as novel
non-nucleoside reverse transcriptase inhibitors of HIV-1
⇑
Bin Liu , Younghee Lee, Jinming Zou, H. Michael Petrassi, Rhoda W. Joseph, Wenchun Chao,
Enrique L. Michelotti, Marina Bukhtiyarova, Eric B. Springman, Bruce D. Dorsey
Ansaris, Four Valley Square, 512 E. Township Line Rd, Blue Bell, PA 19422, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery and SAR study of a series of 4,6-diamino-1,3,5-triazin-2-ol compounds as novel HIV-1 non-
nucleoside reverse transcriptase inhibitors (NNRTIs) are reported. The lead compounds in this series
showed excellent activity against wild-type and drug-resistant RT enzymes and viral strains. In addition,
compounds from this series demonstrated favorable pharmacokinetic profile in rat. A preliminary mod-
eling study was conducted to understand the binding mode of this series of compounds.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 3 August 2010
Revised 1 September 2010
Accepted 7 September 2010
Available online 21 September 2010
Keywords:
NNRTIs
SAR
Human immunodeficiency virus (HIV) reverse transcriptase
(RT) is a DNA polymerase enzyme that converts single stranded
viral genomic RNA into double stranded DNA. HIV RT is one of
three key enzymes in the HIV life cycle and the primary target of
numerous anti-viral drug discovery efforts.1 There are two major
classes of HIV RT inhibitors: nucleoside reverse transcriptase
inhibitors (NRTIs) that bind to the substrate binding site, and
non-nucleoside reverse transcriptase inhibitors (NNRTIs) that bind
to the allosteric site, distinct from the catalytic binding site. Highly
active antiretroviral therapy (HAART) combination regimens,
which typically include both NRTIs and NNRTIs as essential com-
ponents, have dramatically decreased the morbidity and mortality
among patients with HIV infections. In general, non-nucleoside re-
verse transcriptase inhibitors are highly specific to HIV-1 RT and
relatively less toxic compared with nucleoside reverse transcrip-
tase inhibitors (NRTIs).2
significantly limited over time.5 Efforts to develop the next-gener-
ation NNRTIs have focused on the design of compounds with an
improved drug resistance profile. Etravirine (TMC 125), an NNRTI
that was approved recently by the U.S. Food and Drug Administra-
tion, demonstrated activity against HIV-1 viral strains that devel-
oped resistance to the first generation NNRTI drugs. Etravirine
belongs to the diaminopyrimidine (DAPY) class.6 A closely related
structure, diaryltriazine (DATA), also has been identified as highly
potent NNRTIs against wild-type HIV-1 RT (Fig. 1).7 Recently, a ser-
ies of pyridinone derivatives has been reported as potent HIV-1
NNRTIs.8
In the process of our anti-viral drug discovery program, we dis-
covered that 4,6-diamino-1,3,5-triazin-2-ol compound 1 (Fig. 2)
possessed low micromolar potency in HIV cytoprotection assay.
Subsequent mechanism of action studies, in which compound 1
was tested against the various anti-HIV targets including protease,
NNRTIs bind in the hydrophobic pocket of HIV-1 RT approxi-
mately 10 Å from the catalytic site. Upon binding, NNRTIs inhibit
RT catalytic function by causing a large conformational change in
the protein and a distortion of the geometry of the substrate bind-
ing site.3 This binding pocket, which is induced by binding of the
NNRTIs, is quite flexible and can accommodate many structurally
diverse inhibitors.4
The first generation NNRTIs which include Efavirenz (EFZ),
Nevirapine (NVP), and Delavirdine (DLV) have suffered from a
relatively low genetic barrier to resistance mutations, and
consequently the therapeutic effectiveness of these agents was
N
N
R1
R2
R1
R2
O
N
NH
X
N
NH
X
N
NH
N
N
N
Br
N
NH2
R3
NH2
X=O, N
TMC 125 (etravirine)
X=C, O, and N
DAPYs
DATAs
⇑
Corresponding author. Tel.: +1 215 358 2063; fax: +1 215 358 2010.
E-mail address: bliu@ansarisbio.com (B. Liu).
Figure 1. Chemical structures of DATA and DAPY as NNRTIs.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.09.034

B. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6592–6596
6593
OH
transcriptase. This series of compounds could be considered as lead
molecules for the discovery of novel NNRTIs with diversified
structures.
N
N
H
N
N
H
N
N
H
The synthesis of 4,6-diamino-1,3,5-triazin-2-ol compounds has
been reported previously.10 In order to expedite the SAR study, both
solution phase synthesis and solid phase synthesis were utilized. A
general solution phase synthetic route is illustrated with the prep-
aration of compound 1 in Scheme 1. To this end, cyanuric chloride
was first treated with the biphenyl-2-amine to afford the interme-
diate N-(biphenyl-2-yl)-4,6-dichloro-1,3,5-triazin-2-amine, which
was treated subsequently with the more nucleophilic primary
amine, (1H-indol-6-yl)methanamine. Finally, hydrolysis of the
remaining chlorine was accomplished by heating with aqueous
NaOH and H2O2 to yield the desired 1,3,5-triazin-2-ol analog in rel-
atively good yield. Previously, we reported an efficient synthesis of
the biaryl building blocks which are analogs of biphenyl-2-amine.11
These building blocks were used in the synthesis of compounds
described herein.
For solid phase synthesis, a modified Wang resin was utilized
following literature precedent.12 Scheme 2 shows a typical solid
phase synthesis we applied for the preparation of analog 2. Briefly,
the modified Wang resin was treated with an excess of cyanuric
chloride in anhydrous THF in the presence of DIEA to obtain the re-
sin-bound dichlorotriazine. The resin-bound dichlorotriazine inter-
mediate was treated with 3-bromoaniline. The resin was washed
Compound 1
Figure 2. Structure of compound 1.
fusion process, integrase, and RT, indicated that RT was the pri-
mary target for this compound with an IC50 of 1.3 l
M.9
A medicinal chemistry optimization was carried out in order to
improve the potency of this initial hit compound. The SAR study
ultimately resulted in a series of 4,6-diamino-1,3,5-triazin-2-ol
compounds with low nanomolar activity against wild-type reverse
transcriptase and highly potent anti-viral activity. This series of
compounds also demonstrated good preliminary pharmacokinetics
including favorable bioavailability. In addition, these 4,6-diamino-
1,3,5-triazin-2-ol inhibitors demonstrated a SAR pattern different
from the DATA NNRTIs,7 suggesting a possible different binding
mode. In this Letter, we describe structure–activity relationship
studies, initial in vivo PK properties and the putative binding mode
of the 4,6-diamino-1,3,5-triazin-2-ol compounds with reverse
Cl
C
l
H
N
N
N
a
N
N
H2N
+
H2N
+
Cl
N
N
H
Cl
N
Cl
OH
Cl
N
N
N
N
N
H
H
c
N
N
b
N
N
N
H
N
H
N
H
H
1
Scheme 1. Reagents and conditions: (a) DIEA, dioxane, 0 °C; (b) DIEA, dioxane, room temperature; (c) NaOH, H2O2, MeOH, dioxane, 80 °C.
Wang
Resin
∗
O
N
O
N
OMe
b
a
O
N
N
N
N
OH
Cl
N
H
Br
Cl
Cl
MeO
cleavage site
O
N
OH
N
N
d
N
N
c
H
N
H
N
H
N
Br
N
N
H
N
N
H
Br
H
2
Scheme 2. Reagents and conditions: (a) Cyanuric chloride, DIEA, THF, 0 °C; (b) 3-bromoaniline, DIEA, THF, room temperature; (c) (1H-indol-6-yl)methanamine, DIEA, THF,
50 °C; (d) TFA/DMS, DCM.

6594
B. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6592–6596
with THF and DCM, which was subsequently reacted with (1H-in-
dol-6-yl)methanamine before it was treated with 5% TFA and 5%
DMS in DCM. The current 1,3,5-triazin-2-ol structures appear to
exist as a mixture of enol and amide forms based on NMR spectrum
analysis.
Table 1 summarizes the SAR of the left side RL and the right side
RR modifications while maintaining the central 1,3,5-triazin-2-ol
core. In a wild-type (WT) RT enzymatic assay, a few notable SAR
trends were observed. First, introduction of a methyl group in
the 2-position of phenyl ring A (see entry 1 for ring and atom
Table 1
SAR of left side and right side of the inhibitor
OH
N
N
N
RL
RR
Compd Left side RL
Compd Left side RL
Right side RR
WT RT
IC50
Anti HIV
Right side RR
WT RT
IC50
Anti HIV
EC50
a
a,b
a
a,b
EC50
H
N
3
NH
4
5
2
NH
NH
A
O
N
HN
HN
1
1
3.5
40
7.2
10
11
5.1
4.1
1.5
nd
6
B
H
N
N
N
O
O
HN
N
Br
O
2
>10
O
F
H
N
NH
NH
NH
HN
HN
HN
HN
3
>100
0.5
>10
0.9
nd
12
13
14
>100
0.083
0.008
nd
H
N
N
N
O
O
4
0.12
0.005
F
F
O
HN
HN
5
>100
O
F
F
N
NH
O
N
S
HN
HN
O
6
>100
>10
15
0.07
0.01
O
O
N
N
N
NH
O
O
HN
HN
HN
HN
HN
7
8
4.4
2.8
0.2
16
17
0.08
0.007
0.003
O
O
NH
O
0.07
0.055
F
O
O
NH
O
HN
9
0.09
0.4
18
0.09
0.13
a
IC50 and EC50 are in
lM.
b
Test derivatives were evaluated for inhibitory effect versus the wild-type RT enzyme and anti-retroviral activity against HIV-1 WT strain. The efficacy of test compounds
against HIV-1 WT strain was determined at EC50 which is the 50% inhibitory concentration for inhibition of the cytopathic effect of HIV-1RF in CEM-SS cells or HIV-1IIIB in
MT-4 cells.

B. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6592–6596
6595
Table 2
linkers (X and Y). Analogs 21–25 were synthesized using similar
chemistry (Scheme 3). The SAR suggested that replacement of
the NH linker with an oxygen linker on either side of the 1,3,5-tria-
zin-2-ol core resulted in reduced potency (21, 22 vs 8). The SAR
clearly indicated that replacement of the hydroxyl group in com-
pound 8 with an amino group (23) or chloride (24) was detrimen-
tal and resulted in significant loss of activity, implying a possible
loss of a favorable interaction. This SAR observation is opposite
to the reported SAR of the DATA series, where replacement of the
amino group with hydroxyl group resulted in much less active
derivatives.7 Replacement of the hydroxyl group with methoxy
diminished the activity by about 30-fold (25 vs 8). Taken together,
the SAR demonstrated the importance of the 1,3,5-triazin-2-ol core
and amino linkers.
Analysis of data generated in HIV viral cytoprotection assay
indicated that the potency rank of tested compounds tends to cor-
relate well with enzymatic RT activity. Notably, although com-
pounds 16 and 17 displayed inhibitory activity against WT RT in
the medium to high double digit nanomolar range, both com-
pounds showed excellent cellular potency. Overall, compounds
14 and 17 demonstrated the best potency in anti-HIV cellular
assay.13
SAR study of hydroxyl and amino linker replacements
Z
R
N
N
X
N
Y
O
Compd
Structures
Z
WT RT Anti-HIV-1
a
a
X
Y
R
IC50
EC50
8
21
22
23
24
25
NH
NH
O
NH
NH
NH
NH
O
NH
O
NH
NH
OH
OH
OH
NH2
Cl
Me
Me
Me
Me
F
0.07
9.7
23.6
>100
>100
2.57
0.2
12.6
4
>50
ND
3.0
OMe
Me
a
IC50 and EC50 are in lM.
2numbering) in compound 1 resulted in sevenfold increase of po-
tency (4 vs 1). Secondly, replacement of (3H-indol-6-yl)methan-
amine with benzofuran-5-ylmethanamine led to another
sevenfold increase in activity (8 vs 4). However, both substituted
indole (6) and (3,4-dihydro-2H-benzo[b][1,4]dioxepin-7-yl)meth-
anamine (7) resulted in loss of the activity in enzymatic assay.
Fluoride substitution at the 2-position of phenyl ring A offered im-
proved potency similar to methyl substitution (9 vs 8). Interest-
ingly, methylation of the left side NH linker displayed a 10-fold
increase of potency (14 vs 9) in one case, while equal potency
was observed in another case (13 vs 8). On the other hand, ethyla-
tion of the same NH linker resulted in almost complete loss of the
potency (12). Taken together, the SAR suggested that either a tight
binding pocket or a favorable conformational change induced by
the NH methyl group caused this ‘methyl effect.’ Conversely, N-
methylation on the right side NH resulted in significant loss of po-
tency (10 vs 8) suggesting the potential involvement of the NH
group in binding with RT. Introduction of a second fluorine on ring
A resulted in a 10-fold loss of activity (15 vs 14). In addition,
replacement of ring B by other aromatic rings such as furan pro-
vided similar potency (16 vs 13). Two analogs with substitutions
at the 5-position on ring A (2 and 3) displayed diminished activity.
Finally, the SAR suggested the replacement of benzofuran in com-
pound 13 with (2,3-dihydrobenzofuran-5-yl)methanamine (17) or
benzo[d][1,3]dioxol-5-ylmethanamine (18) resulted in comparable
activity in RT inhibition.
Compounds 14 and 17 were selected for testing against three
key NNRTI resistance RT mutants (L100I, V106A, and Y188L) in
enzymatic assay, and the results are shown in Table 3. Both com-
pounds showed good activity against L100I RT mutant, but were
not effective against V106A and Y188L RT mutants. Next, the activ-
ity of compounds 14 and 17 against mutant HIV viruses was also
Table 3
Compound mutation resistance in RT enzymatic assaya
Compd
WT
L100I
V106A
Y188L
14
17
0.008
0.055
0.047
0.070
9.1
36.7
6.1
9.5
a
IC50 in lM.
Table 4
Compound mutation resistance in Anti-HIV cytoprotection evaluation
Compd CeMSS/NL4-3 CEMSS CeMSS/NL4-3 CeMSS/NL4-3
EC50
(
lm)
TC50
(lm) K103N EC50
(
lm)
Y181C EC50 (lm)
NVP
EFZ
14
0.2
>1
>1
>1
>1
>1.0
>1.0
>1.0
>1.0
>1.0
0.005
0.1
0.004
0.006
0.050
Table 2 summarizes the SAR study which explored the role of
the central 1,3,5-triazin-2-ol core substituent (Z) and different
17
0.2
Cl
R
N
N
Cl
N
O
a
b
Cl
Z
R
R
Y
Cl
N
N
N
N
N
c
N
N
X
N
Y
X
N
Cl
Cl
O
O
Cl
N
a
N
N
b
O
Cl
O
Scheme 3. Reagents and conditions: (a) benzofuran-5-ylmethanol or biphenyl-2-ol, DIEA, THF, room temperature; (b) benzofuran-5-ylmethanamine or biphenyl-2-amine
derivatives, DIEA, THF, 50 °C; (c) NaOH, H2O2, MeOH, dioxane, or MeONa, MeOH, or NH3, MeOH, 80 °C.

6596
B. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6592–6596
Table 5
compound 14. It was suggested that the binding of this series
Rat pharmacokineticsa
of compounds can be better explained by the amide tautomer of
the 1,3,5-triazin-2-ol (the structure shown in Fig. 3). The carbonyl
oxygen and NH of the triazine ring form two hydrogen bonds with
the Lys101 backbone NH and carbonyl, respectively, which is con-
sistent with the SAR. The B ring of the biaryl moiety sits in the
pocket formed by Leu100, Leu234, and Trp229. The benzofuran
moiety of compound 14 sits between the lipophilic pocket formed
by Trp229, Tyr188, Tyr181, Phe227, and Val106. The ligand also
Parameter
14
17
Cmax
AUC (
t1/2 (h)
Vdss (L/kg)
Cl (L/h/kg)
F (%)
(
lM)
3.8
6.1
lM h)
12.5
1.45
0.35
0.31
38
15.6
1.40
0.32
0.27
44
a
PO dose: 4 mg/kg. IV dose: 2 mg/kg.
forms favorable
p p interaction with Tyr181 and Trp229. This pro-
À
posed binding mode is different from the binding mode of the
DATA class of NNRTIs,15 presumably due to the substituents on
the current 1,3,5-triazin-2-ol core which are bulkier than those
in the DATA series.
Trp229
Tyr188
In summary, a series of 4,6-diamino-1,3,5-triazin-2-ol com-
pounds was discovered as novel NNRTIs, and the preliminary SAR
of this series of compounds was established. This study led to
the identification of inhibitors with single digit nanomolar potency
both in enzymatic and cellular assays. Preliminary in vivo rat PK
study showed that this class of compounds possesses favorable
PK properties suitable for oral administration. Molecular modeling
studies were employed to understand the binding mode between
these inhibitors and the wild-type HIV reverse transcriptase. The
current results indicate that this series of compounds could serve
as potential leads to further develop novel and potent NNRTIs.
Phe227
Tyr181
Leu234
Val106
Leu100
Val179
Lys103
Acknowledgments
O
H
N
N
N
O
We thank Mr. Brandon Campbell and Dr. Robert Schiksnis for
their analytical assistance.
N
NH
F
Lys101
References and notes
Figure 3. Proposed binding mode of compound 14 with wild-type RT.
1. De Clercq, E. Med. Chem. Res. 2004, 13, 439.
2. Balzarini, J. Curr. Top. Med. Chem. 2004, 4, 921.
evaluated (Table 4). Both compounds showed similar activity and
trends as Nevirapine (NVP) and Efavirenz (EFZ).
In a pharmacokinetic study (Table 5), both compounds 14 and
17 are well absorbed when administered orally in rats (4 mg/kg),
3. Sahlberg, C.; Zhou, X. Anti-Infect. Agents Med. Chem. 2008, 7, 101.
4. Zhou, Z.; Madrid, M.; Evanseck, J.; Madura, J. J. Am. Chem. Soc. 2005, 127, 17253.
5. Ren, J.; Stammers, D. K. Virus Res. 2008, 134, 157.
6. Andries, K.; Azijn, H.; Thielemans, T.; Ludovici, D.; Kukla, M.; Heeres, J.; Janssen,
P.; De Corte, B.; Vingerhoets, J.; Pauwels, R.; de Bethune, M. P. Antimicrob.
Agents Chemother. 2004, 48, 4680.
7. Ludovici, D. W.; Kavash, R. W.; Kukla, M. J.; Ho, C. Y.; Ye, H. Bioorg. Med. Chem.
Lett. 2001, 11, 2229.
8. Le Van, K.; Cauvin, C.; de Walque, S.; Georges, B.; Boland, S.; Martinelli, V.;
Demonté, D.; Durant, F.; Hevesi, L.; Van Lint, C. J. Med. Chem. 2009, 52, 3636.
9. In this mechanism of action study, compound 1 showed the following EC50’s:
with a Cmax of 3.8 lM and 6.1 lM, respectively. Following intrave-
nous dosing (2 mg/kg), both compounds showed reasonable half-
life (ꢀ1.4 h), and the apparent bioavailability was determined to
be around 40% for both compounds.
A computational modeling study was carried out to understand
the possible binding mode of this series of compounds and interac-
tions with the reverse transcriptase. Briefly, the prediction of the
binding mode was performed with a pharmacophore search fol-
lowed by an energy minimization within the protein. The protein
structure was obtained from the Protein Brookhaven Database
(PDB code 1SUQ). A four point pharmacophore was defined based
on 1SUQ ligand, which contains three aromatic features and one
hydrogen bond donor interaction with Lys101 backbone carbonyl
group. The shape of the binding site is defined with 190 exclusive
volume points based on the 1SUQ protein structure. Ligand confor-
mations were generated with a systematic search. The binding
modes of the compounds were selected by visual inspection of
the poses.14 Figure 3 shows the predicted binding mode of
Fusion inhibition >25 lM; Integrase >1000 lM; Protease >100 lM.
10. Baindur, N.; Chadha, N.; Brandt, B.; Asgari, D.; Patch, R.; Schalk-HiHi, C.; Carver,
T.; Petrounia, I.; Baumann, C.; Ott, H.; Manthey, C.; Springer, B.; Player, M. J.
Med. Chem. 2005, 48, 1717.
11. Liu, B.; Moffet, K.; Joseph, R.; Dorsey, B. Tetrahedron Lett. 2005, 46, 1779.
12. Kaval, N.; Van der Eycken, J.; Dehaen, W.; Strohmeier, G.; Kappe, C.; Van der
Eycken, E. J. Comb. Chem. 2003, 5, 560.
13. Compound 14 was also screened against other anti-HIV targets and showed the
following EC50’s: Fusion inhibition = 20
M.
lM; Integrase >100 lM; Protease >100
l
14. Pharmacophore generation and searching, generation of ligand conformation
and energy minimization were performed with MOEÒ (Chemical Computing
Group).
15. Das, K.; Clark, D., Jr.; Lewi, P.; Heeres, J.; De Jonge, M.; Koymans, L.; Vinkers, H.;
Daeyaert, F.; Ludovici, D.; Kukla, M.; De Corte, B.; Kavash, R.; Ho, C.; Ye, H.;
Lichtenstein, M.; Andries, K.; Pauwels, R.; De Bethune, M.; Boyer, P.; Clark, P.;
Hughes, S.; Janssen, P.; Arnold, E. J. Med. Chem. 2004, 47, 2550.