Design, synthesis, and biological evaluations of novel quinolones as HIV-1 non-nucleoside reverse transcriptase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.05.073

A novel series of quinolones was discovered as HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs) using a structure-based approach. The lead quinolones exhibited single digit nanomolar potency in the HIV-1 replication assays. The preliminary S

Full text of DOI:10.1016/j.bmcl.2006.05.073

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4246–4251
Design, synthesis, and biological evaluations of novel
quinolones as HIV-1 non-nucleoside reverse transcriptase inhibitors
David Ellis, Kelli L. Kuhen, Beth Anaclerio, Baogen Wu, Karen Wolff, Hong Yin,
Badry Bursulaya, Jeremy Caldwell, Donald Karanewsky and Yun He*
Genomics Institute of the Novartis Research Foundation (GNF), 10675 John Jay Hopkins Drive, San Diego, CA 92121, USA
Received 7 April 2006; revised 20 May 2006; accepted 22 May 2006
Available online 16 June 2006
Abstract—A novel series of quinolones was discovered as HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs) using a
structure-based approach. The lead quinolones exhibited single digit nanomolar potency in the HIV-1 replication assays. The
preliminary SAR of these quinolones was also established via systematic structural modifications. These novel and potent
quinolones could serve as advanced leads for further optimization.
Ó 2006 Elsevier Ltd. All rights reserved.
HIV-1 reverse transcriptase (RT) is a key enzyme in
Et
HIV replication and is a well-established target for
anti-HIV drug discovery.1–3 Three non-nucleoside re-
verse transcriptase inhibitors (NNRTIs) have been ap-
proved by FDA. They are nevirapine (ViramuneÒ),
delavirdine (RescriptorÒ) and efavirenz (SustivaÒ), and
are key components of the combination therapy.4–6
Since significant resistance has been developed against
these drugs, there is an urgent need to develop new
NNRTIs that would overcome the current drug resis-
tance.7 Previously, we have reported the discovery of
novel NNRTIs via HTS and SAR studies.8–11 In this let-
ter, we wish to report our discovery of a series of quin-
olones as NNRTIs based on rational design.
H
O
O
CF3
O
O
Cl
O
N
Cl
S
O
N
O
N
H
O
N
H
H
1: Efavirenz
2: HBY-097
3
O
R
Br
OEt
N
O
H
4
Figure 1.
Among the reported NNRTIs, a number of them share
common structural features. For example, efavirenz (1),
HBY-097 (2),12 and oxindole 38,9 all share a fused aro-
matic core, which contains a hydrophobic phenyl moie-
ty, and an amide or carbamide moiety. By carefully
examining these structures, we designed a quinolone
scaffold (4). Based on molecular modeling studies,
quinolone 4 fits into the NNRTI binding site and aligns
nicely with efavirenz in space (Fig. 1). The quinolone
core of 4 and the dihydrobenzoxazinone core of efavi-
renz almost overlap in the center of the binding pocket.
The halophenyl moieties in both molecules occupy the
same hydrophobic pocket. An alkyl R group in 4 and
the cyclopropylethynyl group in efavirenz could occupy
the same hydrophobic pocket. The cyclopropyl moiety
in 4 and the trifluoromethyl group in efavirenz fill into
another small hydrophobic space. Both molecules could
form critical hydrogen bonds between the lactam moie-
ties and K101 of the enzyme. The ethyl ester group in 4
extends out into the hydrophobic pocket and points to
the solvent exposed regions.
Keywords: Quinolones; HIV-1 non-nucleoside reverse transcriptase
inhibitor (NNRTI); SAR; Molecule modeling; HIV replication.
Encouraged by these modeling studies, we carried out
the synthesis of the designed quinolones (Scheme 1).13
Starting from the commercially available aniline 5, the
*
Corresponding author. Tel.: +1 858 332 4706; fax: +1 858 332
4513; e-mail: yhe@gnf.org
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.073

D. Ellis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4246–4251
4247
R
Cl
O
Cl
CN
Cl
CN
a
b
NH
NH
NH2
PMB
PMB
7
6
5
c
R
R
R
Cl
CO2Et
O
Cl
CO2Et
CO2Et
Cl
e
d
N
N
O
N
O
H
PMB
PMB
4
9
8
4a:
4b:
4c:
4d:
4e:
4f:
4i:
R = Me
R = Et
R = nPr
R = nBu
R = (CH2)2CHCH2
R = nPent
R = (CH2)5CH3
R = iPr
R = CH2CH(CH3)2
R = CH2Ph
R = (CH2)2Ph
R = cPr
R = cPentyl
R = cHexyl
4j:
4k:
4l:
4m:
4o:
4p:
4g:
4h:
R = 4-PhPh
Scheme 1. Synthesis of quinolones 4a–4p. Reagents and conditions: (a) PMBOH (1.5 equiv), p-TsOH (cat.), CH3CN, 70 °C, 12 h, 99%; (b) RMgCl
(2.0 equiv), THF, 0–25 °C, 1 h, 75–92%; (c) ClCOCH2CO2Et (2.0 equiv), benzene, 80 °C, 4 h, 73–88%; (d) (CH3)2SOCH3I (2.0 equiv), NaH (3.0
equiv), DMSO, 0–90 °C, 51–78%; (e) CAN (2.0 equiv), CH3CN/H2O = 9:1, 0–25 °C, 12 h, 74–90%.
amino group was first protected with a PMB group, and
the nitrile was then converted into a ketone by reacting
with various Grignard regents. Treatment of the result-
ing ketone with ethyl chlorocarbonylacetate led to the
formation of the quinolone core 8. The cyclopropane
moiety was installed by reacting with dimethyloxosulfo-
nium methylide.14 Removal of the PMB group gave rise
to quinolones 4a–4p. These quinolones were tested
against the WT HIV-1 virus, and their antiviral activities
are shown in Table 1. As suggested by molecular model-
ing studies, a number of these quinolones exhibited po-
tent antiviral activity, and a clear SAR emerged. The R
group prefers straight aliphatic chains, and the best R
groups are n-butyl (4d), n-4-pentenyl (4e), and n-pentyl
(4f), and all three compounds exhibited low single digit
nanomolar potency. Either longer or shorter R groups
resulted in reduced activity (4a–4c, 4g). With the excep-
tion for the i-butyl analog 4i and phenethyl analog 4k,
all other quinolones with branched alkyl, phenyl alkyl
or phenyl groups (4h, 4j, and 4l–4p) showed poor antivi-
ral activity. These data are consistent with the modeling
studies, which suggest a narrow hydrophobic pocket in
the enzyme which can interact with this region of the
molecule (Fig. 2).
the final products 13 and 14, respectively. Antiviral
testing indicated that both 13 and 14 lost activity sig-
nificantly, representing more than a 60- and 30-fold
reduction in potency, respectively (Table 1). These
data demonstrate the important role of the ester moi-
ety in the interaction with the enzyme, as well as the
cyclopropane moiety in anchoring the molecule for
optimal interactions.
To further explore the SAR in the ester region, a li-
brary of esters and amides (16a–16y) were prepared
via the corresponding acid intermediate 15 (Scheme
3). Biological testing suggested that small esters are
preferred with the methyl (16a) and allyl (16b) esters
being among the most potent. The slightly larger es-
ters, cyanoethyl (16c), isobutenyl (16d), and phenyl
analogs (16e–16f), suffered a dramatic loss in antiviral
activity. In the case of 16c, the polar nitrile moiety
could have resulted in unfavorable interactions with
the enzyme. Replacement of the ethyl ester with a pri-
mary amide still gave a potent inhibitor (16g),
although it suffered some loss of activity.
Consistent with the observations in the ester series, all
tertiary amides, phenyl and pyridyl amides (16q–16y),
showed significantly less activity. Only the 3-cyanophe-
nyl analog 16t displayed good antiviral potency
(160 nM). The activity difference between 16s and 16t
could be the result of the steric interactions of the ortho
nitrile group in 16s with the quinolone core that led to
less favorable interactions with the enzyme. As the com-
pounds were tested in a cell-based assay, different cell
permeability of these analogs might have played a role
in their reduced antiviral activity.
To explore the role of the ethyl ester moiety, a nitrile
analog 13 was prepared (Scheme 2). Ketone 7d was
first converted into the corresponding cyanoquinolone
10 in excellent yield by reacting with cyanoacetyl chlo-
ride. Treatment of 10 with dimethyloxosulfonium
methylide led to the isolation of the desired cyclopro-
pane analog 11. In addition to 11, another product
was isolated, which had a molecular weight of 14
mass units more than that of 11. Spectroscopic studies
established its structure as a cyclobutane analog 12,
which was likely due to the reaction of 11 with an
additional equivalent of dimethyloxosulfonium methy-
lide. Removal of the PMB groups in 11 and 12 led to
Alcohol 19a and ethers 19b–19e were also prepared to
further understand the SAR in this region (Scheme 4).
Ester 9d was first selectively reduced to alcohol 17,

4248
D. Ellis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4246–4251
Table 1. Anti-HIV activities of quinolones.15
Compound
EC50 (lM)
CC50 (lM)
Efavirenz
Nevirapine
2
0.001
0.050
0.002
>10
ꢀ10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
8.822
>10
>10
>10
>10
>10
6.088
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
4a
4b
5.039
0.045
0.003
0.001
0.001
0.025
>10
4c
4d
4e
4f
4g
4h
4i
0.057
>10
4j
4k
0.221
>10
4l
4m
4n
3.248
1.462
>10
Figure 2. Docking of 4d and efavirenz in the NNRTI binding site.15
4o
4p
>10
13
14
0.197
0.102
0.001
0.001
0.113
2.361
>10
These data also illustrate the importance of the ester car-
bonyl oxygen of 4d in the interaction with the enzyme.
16a
16b
16c
16d
16e
16f
16g
16h
16i
16j
16k
16l
16m
16n
16o
16p
16q
16r
16s
16t
16u
16v
16w
16x
16y
19a
19b
19c
19d
19e
23
A ketone analog (23) of 4d was synthesized according to
Scheme 5. Hydrolysis of the ester in 9d with lithium
hydroxide gave the corresponding acid 20, which was
converted into Weinreb amide 21. Reaction of amide
21 with propyl magnesium chloride furnished ketone
22. Ketone 23 was obtained after the removal of the
PMB group using CAN. Although 23 still exhibited
good antiviral activity, its EC50 suffered more than 20-
fold loss compared to that of 4d, demonstrating the role
of the ester oxygen in the interaction with the enzyme.
Molecular modeling was carried out to understand the
interaction of the ester region with the enzyme. The
activity of ester (4d), ether (19c), amide (16g), keto
(23), and ethylamide (16h) quinolones correlates quite
well with their corresponding free energies of binding,
which are À33.7, À28.0, À32.2, À29.8, and À24.6 kcal/
mol, respectively. The q2 value for this correlation is
0.54. The lower potency of keto-quinolone (23) could
be due to the lack of a polar atom (oxygen in ester or
nitrogen in amide) that occupies a polar sub-pocket in
mostly hydrophobic binding site as identified with Site-
map, Schrodinger, LLC.16
1.095
0.012
0.747
0.489
6.300
2.775
5.691
2.821
5.205
>10
>10
>10
>10
1.713
0.160
1.327
4.551
2.695
8.259
>10
0.743
0.045
0.065
0.240
0.037
0.265
0.166
1.398
Quinolones 24 and 25 were prepared to further under-
stand the SAR in the cyclopropane region (Scheme 6).
The olefinic analog 24 was conveniently synthesized by
simply removing the PMB group in 8d. Conjugate addi-
tion of lithium dimethylcuprate17 to the double bond in
8d followed by the removal of the PMB group furnished
25 as a mixture of two diastereoisomers in a ratio of 1:1.
While 24 retained a 166 nM EC50 against the WT HIV
virus, 25 had only micromolar potency. These data
again support the important role of the cyclopropane
moiety in rigidifying the molecules for optimal
interactions.
24
25
which was then treated with methyl, ethyl, and allyl io-
dide in the presence of sodium hydride, giving rise to the
corresponding PMB-protected ethers 18b–18d. Standard
acylation of the hydroxy group in 17 led to acetate 18e.
Deprotection of the PMB groups in 17, 18b–18e with
CAN furnished the alcohol analogs 19a–19e. Interest-
ingly, most of these analogs exhibited good antiviral
activity. In particular, 19b, 19c, and 19e showed EC50s
of 45, 65, and 37 nM, respectively, which are consistent
with the SAR observed for the corresponding esters.
A number of quinolones were assayed against a panel of
NNRTI resistant mutants. These quinolones exhibited
good antiviral activity against a number of these mu-
tants, including E138K, F227L, I135V, L100I, and

D. Ellis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4246–4251
4249
( )n
Cl
Cl
CN
O
Cl
CN
O
O
a
b
NH
PMB
N
N
PMB
PMB
7d
12
10
11: n = 1
12: n = 2
c
Cl
Cl
CN
O
c or d
CO2Et
O
N
H
N
H
14
13
Scheme 2. Synthesis of quinolones 13 and 14. Reagents and conditions: (a) NCCH2COCl (2.0 equiv), benzene, 80 °C, 2 h, 89%; (b) (CH3)2SOCH3I
(3.0 equiv), NaH (4.2 equiv), DMSO, 60 °C, 12 h, 77%, 11:12 = 1:1; (c) 11, CAN (2.0 equiv), CH3CN/H2O = 9:1, 0–25 °C, 12 h, 76% for 13, 82% for
14; (d) Concd H2SO4 (cat.), EtOH, reflux, 24 h, 47%.
Cl
CO2H
O
Cl
CO2Et
O
Cl
COR
O
a
b
N
N
H
N
H
H
4d
15
16
16a: R = OMe
16i: R = NHAllyl
16j: R = nBu
16k: R = NHiPr
16l: R = CH2cPr
16m:R = NHBn
16o: R = NH(CH2)2CO2H
16p: R = N(CH2)3
16q: R = N(CH2)4
16r: R = N(Me)Et
16s: R = NHPh-2-CN
16t: R = NHPh-3-CN
16u: R = NHPh-2-CONH2
16v: R = NHPh-4-CONH2
16w:R = NH-2-Py
16b: R = OAllyl
16c: R = O(CH2)2CN
16d: R = OCH2CHCMe2
16e: R = OPh
16f: R = OPh-4-CN
16g: R = NH2
16h: R = NHEt
16x: R = NH-4-Py
16y: R = NHPh-4-CO2H
Scheme 3. Synthesis of quinolones 16a–16y. Reagents and conditions: (a) LiOH (3.0 N), THF/MeOH/H2O = 3:2:1, 70 °C, 5 h, 90%; (b) EDCI
(2.0 equiv), iPr2NEt (3.0 equiv), CH2Cl2, 25 °C, 12 h, 77–94%.
Cl
Cl
Cl
CO2Et
O
OH
a
OR
b
N
O
N
O
N
PMB
PMB
PMB
18b-18e
17
9d
c
a:
b:
c:
d:
e:
R = H
R = Me
R = Et
R = Allyl
R = Ac
Cl
OR
N
H
O
19a-19e
Scheme 4. Synthesis of quinolones 19a–19e. Reagents and conditions: (a) LAH (1.1 equiv), THF, 0 °C, 1 h, 96%; (b) for 18b–18d: RI (1.3 equiv),
NaH (2.0 equiv), DMSO, 77–86%; for 18e: Ac2O (2.0 equiv), Et3N (3.0 equiv), CH2Cl2, 0–25 °C, 1 h, 82%; (c) CAN (2.0 equiv), CH3CN/H2O = 9:1,
0–25 °C, 12 h, 82–91%.
Y181H. However, they showed only moderate or weak
activity for the rest of the mutant viruses (Table 2).
propane moiety in positioning the substituents on the
quinolone nucleus for optimal interactions with the
enzyme. The ester moiety also plays an important role
for the antiviral activity. Several of these quinolones
exhibited potent inhibitory activity against the WT virus
and showed promising activity against several NNRTI
In summary, using a structure-based approach, a series
of novel quinazoline NNRTIs was designed and synthe-
sized. SAR studies revealed the critical role of the cyclo-

4250
D. Ellis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4246–4251
O
O
N
O
Cl
CO2H
O
Cl
CO2Et
O
Cl
a
b
N
N
N
PMB
PMB
PMB
9d
20
21
c
Cl
d
Cl
O
O
O
N
O
N
H
PMB
23
22
Scheme 5. Synthesis of quinolone 23. Reagents and conditions: (a) LiOH (3.0 N), THF/MeOH/H2O = 3:2:1, 70 °C, 5 h, 88%; (b) MeSO2Cl (1.1
equiv), Et3N (3.0 equiv), HN(Me)OMe, CH2Cl2, 0–25 °C, 1 h, 88%; (c) nPrMgCl (1.0 equiv), THF, 0–25 °C, 12 h, 24%; (d) CAN (2.0 equiv),
CH3CN/H2O = 9:1, 0 °CÀ25 °C, 12 h, 40%.
Cl
CO2Et
O
Cl
CO2Et
O
Cl
CO2Et
O
a
b,a
N
N
N
H
H
PMB
24
8d
25
Scheme 6. Synthesis of quinolones 24 and 25. Reagents and conditions: (a) CAN (2.0 equiv), CH3CN/H2O = 9:1, 0–25 °C, 12 h, 62–71%; (b) MeLi
(5.5 equiv), CuI (5.0 equiv), Et2O, 0 °C, 5 min, then 8d, 0–25 °C, 1 h, 53%.
Table 2. Antiviral activities (EC50, lM) of selected quinolones against HIV-1 resistant mutants15
Compound
WT
E138K
F227L
I135V
K103N
L100I
V106A
Y181C
Y188L
V179E
Y181H
Y181S
Efavirenz
Nevirapine
0.001
0.050
0.045
0.003
0.001
0.001
0.025
0.057
0.001
0.001
0.012
0.045
0.065
0.037
0.002
0.031
0.132
0.047
0.057
0.018
0.098
0.091
0.008
0.032
0.046
0.042
0.094
0.054
0.001
0.148
0.026
0.007
0.009
0.005
0.041
0.022
>10
0.001
0.079
0.097
0.028
0.033
0.017
0.066
0.085
0.012
0.023
0.015
0.046
0.074
0.064
0.032
5.053
>10
0.010
0.164
0.041
0.014
0.035
0.012
0.053
0.063
0.004
0.001
0.013
0.033
0.039
0.045
0.004
8.458
1.013
0.279
0.260
0.238
0.878
2.941
0.057
1.220
2.178
0.726
1.124
4.725
0.001
>10
>10
0.280
>10
>10
>10
>10
>10
>10
>10
7.339
>10
>10
>10
>10
>10
0.006
0.116
1.889
0.914
0.304
0.234
1.597
3.953
0.244
0.233
0.439
0.721
1.034
0.735
0.002
0.112
0.210
0.009
0.089
0.016
0.061
0.210
0.008
0.020
0.063
0.121
0.216
0.059
0.001
3.001
1.213
0.684
1.681
0.369
0.929
0.901
0.581
0.932
2.525
2.704
4.180
2.354
4c
4d
0.884
1.247
0.669
3.075
>10
1.653
>10
4e
4f
4g
0.926
2.249
>10
4i
16a
16b
16g
19b
19c
19e
0.682
1.081
3.827
2.916
3.955
2.618
0.727
2.945
7.957
5.029
6.237
7.463
0.001
0.001
0.020
0.251
0.001
5. Maggiolo, Franco; Ripamonti, Diego; Suter, Fredy
J. Antimicrob. Chemother. 2005, 5, 821.
6. Marks, K.; Gulick, R. M. HIV Chemother. 2005, 1.
7. Wainberg, M. A.; Sawyer, J. P. C.; Montaner, J. S. G.;
Murphy, R. L.; Kuritzkes, D. R.; Raffi, F. Antiviral Ther.
2005, 10, 13.
resistant mutants. These novel quinolones could serve as
advanced leads for further optimizations, the goal of
which will be focused on overcoming the NNRTI
resistance.
8. Jiang, T.; Kuhen, K. L.; Wolff, K.; Yin, H.; Bieza, K.;
Caldwell, J.; Bursulaya’, B.; Wu, T. Y.-H.; He, Y. Bioorg.
Med. Chem. Lett. 2006, 16, 2105.
References and notes
1. Locatelli, G. A.; Cancio, R.; Spadari, S.; Maga, G. Curr.
Drug Metab. 2004, 5, 283.
2. De Clercq, E. Chem. Biodiversity 2004, 1, 44.
3. Gotte, M. Expert Rev. Anti-Infect. Ther. 2004, 2, 707.
4. Barbaro, G.; Scozzafava, A.; Mastrolorenzo, A.; Supuran,
C. T. Curr. Pharm. Des. 2005, 11, 1805.
9. Jiang, T.; Kuhen, K. L.; Wolff, K.; Yin, H.; Bieza, K.;
Caldwell, J.; Bursulaya’, B.; Tuntland, T.; Zhang, K.;
Karanewsky, D.; He, Y. Bioorg. Med. Chem. Lett. 2006,
16, 2109.
10. Wu, B.; Kuhen, K.; Nguyen, T. N.; Ellis, D.; Anaclerio,
B.; He, X.; Yang, K.; Karanewsky, D.; Yin, H.; Wolff, K.;

D. Ellis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4246–4251
4251
Bieza, K.; Caldwell, J.; He, Y. Bioorg. Med. Chem. Lett.
2006, 16, 3430.
14. Meyers, A. I.; Romine, J. L.; Fleming, S. A. J. Am. Chem.
Soc. 1988, 110, 7245.
11. Wang, Z.; Wu, Baogen; Kuhen, K. L.; Nguyen, T. N.;
Nguyen, D. G.; Bursulaya, B.; He, Y. Bioorg. Med. Chem.
Lett. 2006, 16, 0000.
12. Kleim, J.-P.; Bender, R.; Kirsch, R.; Meichsner, C.;
Paessens, A.; Roesner, M.; Ruebsamen-Waigmann, H.;
Kaiser, R.; Wichers, M. Antimicrob. Agents Chemother.
1995, 39, 2253.
13. All compounds reported are racemic, and only one
enantiomer is shown. Data are based on the measurement
of the racemic mixture.
15. Molecular modeling and assays were carried out as
described in Refs. 8 and 9.
16. The free energy of binding was obtained by subtracting
the ligand and receptor energies from that of the
complex. For the free energy calculations the OPLS-AA
force field (Macromodel, Schrodinger Inc) was employed.
The solvation was accounted for via Generalized Born
method.
17. Irie, H.; Matsumoto, R.; Nishimura, M.; Zhang, Y. Chem.
Pharm. Bull. 1990, 38, 1852.