The design and synthesis of diaryl ether second generation HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs) with enhanced potency versus key clinical mutations

Article abstract of DOI:10.1016/j.bmcl.2008.03.064

Using a combination of traditional Medicinal Chemistry/SAR analysis, crystallography, and molecular modeling, we have designed and synthesized a series of novel, highly potent NNRTIs that possess broad antiviral activity against a number of key clinical mutations.

Full text of DOI:10.1016/j.bmcl.2008.03.064

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2959–2966
The design and synthesis of diaryl ether second generation
HIV-1 non-nucleoside reverse transcriptase inhibitors
(NNRTIs) with enhanced potency versus key clinical mutations
Thomas J. Tucker,a,* Sandeep Saggar,a John T. Sisko,a Robert M. Tynebor,a
Theresa M. Williams,a Peter J. Felock,b Jessica A. Flynn,c Ming-Tain Lai,b Yuexia Liang,d
Georgia McGaughey,a Meiquing Liu,b Mike Miller,b Gregory Moyer,c Vandna Munshi,b
Rebecca Perlow-Poehnelt,a Sridhar Prasad,a Rosa Sanchez,d Maricel Torrent,a
Joseph P. Vacca,a Bang-Lin Wana and Youwei Yana
aDepartments of Medicinal Chemistry and Structural Biology, Merck Research Laboratories, WP14-3,
770 Sumneytown Pike, PO Box 4, West Point, PA 19486-0004, USA
bDepartment of Antiviral Research, Merck Research Laboratories, WP14-3, 770 Sumneytown Pike,
PO Box 4, West Point, PA 19486-0004, USA
cDepartment of Vaccines and Biologics Research, Merck Research Laboratories, WP14-3,
770 Sumneytown Pike, PO Box 4, West Point, PA 19486-0004, USA
dDepartment of Drug Metabolism, Merck Research Laboratories, WP14-3, 770 Sumneytown Pike,
PO Box 4, West Point, PA 19486-0004, USA
Received 31 January 2008; revised 20 March 2008; accepted 21 March 2008
Available online 25 March 2008
Abstract—Using a combination of traditional Medicinal Chemistry/SAR analysis, crystallography, and molecular modeling, we
have designed and synthesized a series of novel, highly potent NNRTIs that possess broad antiviral activity against a number of
key clinical mutations.
Ó 2008 Elsevier Ltd. All rights reserved.
Since their discovery in the early 1990s, non-nucleoside
reverse transcriptase inhibitors (NNRTIs) have been
shown to be a key component of highly active anti-ret-
roviral therapy (HAART). These compounds target an
allosteric binding pocket that lies close to the catalytic
site, and their binding is non-competitive with respect
to dNTP’s and template/primer.1 There are currently
three commercially available NNRTIs: efavirenz, nevi-
rapine, and delavirdine (Fig. 1). The use of efavirenz
and nevirapine has become a part of standard combina-
tion antiviral therapies producing clinical outcomes with
efficacy comparable to other antiviral regimens.2 Despite
the demonstrated clinical efficacy of combination antivi-
ral regimens using NNRTIs, the emergence of clinical
resistance has become a key issue for this class of com-
pounds and has become a major cause of treatment fail-
ure. A number of important resistant variants have
Keywords: NNRTI; HIV; Mutations.
*
Corresponding author. Tel.: +1 215 699 3437; fax: +1 215 652
3971; e-mail: tom_tucker@merck.com
Figure 1. Currently marketed NNRTIs.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.064

2960
T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
emerged, with the Lys103Asn (K103N) mutation con-
ferring clinical resistance to most known NNRTIs and
the Tyr181Cys (Y181C) mutation conferring clinical
resistance to nevirapine.2 Therefore, an urgent need
has arisen for NNRTIs that possess both a broad spec-
trum of antiviral activity against key mutant strains and
a high genetic barrier to the selection of new mutant
strains. Using a combination of traditional Medicinal
Chemistry/SAR analysis, molecular modeling, and crys-
tallographic studies, we have designed and synthesized a
series of novel, highly potent NNRTIs that possess
broad spectrum antiviral activity. In this manuscript,
we describe the design, synthesis, and biological profile
of this series of novel NNRTIs.
Our early efforts in this series focused on the synthesis of
analogs of 1 (Fig. 2), which were discovered by high
throughput screening.3 However, disappointing phar-
macokinetic results with compounds of this class forced
us to retarget our efforts.3 Using 1 as a lead structure,
and comparing the compound with the known NNRTI
2 (Fig. 2)4 via a molecular modeling analysis,5 we pro-
posed to hold the common features of the two com-
pounds constant and replace the western (A ring)
portion of 2 with a novel biaryl sub-structure. Using a
rapid analog approach, we examined a number of biaryl
replacements, arriving at the novel diphenyl ether sub-
structure of 3 (Fig. 2) that became a next generation lead
for further efforts in our project. Direct molecular mod-
eling comparison of 2 and 3 (Fig. 3)5 suggested that the
biaryl regions of the two molecules likely fit the NNRTI
site in a highly similar manner, and also suggested that
the amide carbonyl of 3 likely maintained a critical
hydrogen bonding interaction with the backbone NH
of Lysine 103 similar to that seen with 2.5 It is thought
that direct interaction with the backbone of Lysine 103
is critical for maintaining good potency against muta-
tions at this position such as K103N.6 This modeling
comparison strongly suggested (by analogy to 25) that
the addition of di-meta substitution on the western aryl
ring (A ring) as well as a 2 or 3 substituent on the central
aryl ring (B ring) of 3 might help to optimize interac-
tions with the binding site and help to further improve
potency in this series. The model further suggested that
the aryl sulfonamide eastern ring (C ring) lied along the
enzyme/water interface with the sulfonamide group fully
Figure 3. Molecular modeling comparison of compound 2 (blue) and
compound 3 (yellow) in the WT-NNRTI binding site. Key active site
residues are labeled in black. Details of the molecular modeling
methodologies are provided in Ref. 5.
solvent exposed, indicating the potential for various po-
lar substitutions in this region. Using this modeling
study as a guide, we moved forward to prepare a series
of analogs of 3 in an effort to enhance the potency of this
series.
Analogs in this series were synthesized via the route de-
tailed in Scheme 1.7 The lead structure 3 was prepared
via a two-step synthesis. Acylation of aniline 4 with acid
bromide 5 provided the alpha-bromo amide intermedi-
ate 6. Alkylation of 3-phenoxyphenol with 6 provided
3 in low yield. SNAR reaction of the commercially avail-
able phenols 7 and fluoride 8 provided the diaryl ether
intermediates 9 in good yield. The methyl protecting
group of intermediates 9 was removed in nearly quanti-
tative yield by treatment with boron tribromide to give
the key diaryl ether–phenol intermediate 10. Alkylation
of 10 with tert-butyl bromoacetate provided the tert-bu-
tyl ester, which was cleaved with TFA and the crude
acid refluxed with thionyl chloride to provide the acid
chloride 11 in 40% yield for the three step transforma-
tion. Treatment of the acid chloride with the requisite
aniline derivative in refluxing benzene gave the desired
amide derivatives 12–17 in moderate to good yield.
Alternatively, the key phenol intermediate 10 was alkyl-
ated with the desired chloromethyl or bromomethyl
Figure 2. Design paradigm for novel diaryl ether lead structures.

T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
2961
Scheme 1. Synthesis of amide analogs 3, 12–17 and heterocyclic analogs 18–22. Reagents and conditions: (a) NaHCO3, CH2Cl2, rt (88%); (b) NaH,
2-phenoxyphenol DMF, rt, 2 h (20%); (c) K2CO3, NMP, 120° (75%); (d) BBr3, CH2Cl2 (95%); (e) i—NaH, tert-butylbromoacetate, DMF; ii—neat
TFA; iii—SOCl2/reflux (40% for the three steps); (f) HNR1R2, benzene, reflux overnight (30–90%); (g) i—NaH, XCH2Het, DMF, 2–4 h; ii—
deprotection (if needed) (70–90% for two steps). Full experimental details and analytical data are provided in Ref. 7.
heterocycle derivative (purchased commercially or pre-
pared according to standard literature methods) to pro-
vide the desired heterocyclic products 18–22. In the case
of the indazole derivative 22, it was necessary to use the
N1-Boc protected 3-bromomethyl indazole in the
alkylation, and the Boc protected product was treated
with neat TFA to remove the Boc group and provide 22.
atom on the central ring fit tightly in a small lipophilic
pocket below the Y181 residue and behind V179. The
Y181 side chain is rotated by approximately 90° relative
to its position in most first generation NNRTI crystal
structures,6 and is sandwiched between the meta chlo-
rine atom of the A western ring and the 3-chlorine atom
of the central B ring. The observed rotation of the Y181
side chain may be important for attaining the high levels
of potency versus the Y181C mutant seen with this ser-
ies, as the direct stacking interaction between the A ring
of the inhibitor and the aryl ring of the Y181 side chain
is minimized. In some first generation NNRTIs such as
nevirapine, this pi-stacking interaction is critical for
attaining potency but becomes a key liability versus
the Y181C mutant.6 The amide carbonyl group of the
inhibitor makes a direct H-bonding interaction with
the backbone NH of K103, another key interaction that
contributes to the high levels of potency observed
against the K103N mutation. The C ring and its ortho-
chlorine substituent make lipophilic interactions with
Y318, F227, P236, and V106. The entire eastern edge
of the C ring lies along the solvent interface, with the
4-sulfonamide group completely solvent exposed. The
crystal structure suggested that moving the chlorine
atom on the B ring to the neighboring 2-position might
provide an even better fit in the active site, since the
chlorine atom of 14 appeared to be making very close
contact with the side chain of V179. Moving the 3-chlo-
rine substituent on the central aryl B ring one position to
the 2-position (15) resulted in an increase in potency ver-
sus WT and the key mutant enzymes, providing a com-
pound with sub-nanomolar enzyme inhibitory potency
and high levels of antiviral potency in our cell based as-
say. Replacement of the aryl sulfonamide of 15 with a
Table 1 details the data for the amide series of analogs.
Compounds were evaluated for intrinsic enzyme inhibi-
tory potency versus WT RT (wild type HIV-1 reverse
transcriptase), as well as the K103N and Y181C mutant
RTs.8 Compounds were also evaluated for antiviral po-
tency (CIC95) against WT and the key mutant viruses in
the presence of 10% FBS (Fetal Bovine Serum; WT and
mutants) as well as 50% NHS (Normal Human Serum;
WT) to evaluate the effects of protein binding.9 Adding
the 3-chloro substituent to the central aryl ring (12) re-
sulted in an increase in activity versus WT and the
Y181C mutant RT enzyme. Addition of a meta cyano
substituent to the A ring resulted in a large enhancement
of potency (13) versus WT and both the K103N and
Y181C enzymes, and the further addition of a second
meta chloro substituent provided a more moderate
enhancement in potency versus the K103N mutant
(14). This di-meta substitution on the A ring appears
to provide excellent potency versus WT and the key mu-
tant enzymes, similarly to the SAR seen with 2.4 Crystal-
lographic studies with 14 (Fig. 4)10 confirmed the
predictions of our earlier molecular modeling studies.
The crystal structure of 1410 demonstrated (as suggested
by the modeling study depicted in Fig. 3) that the A ring
was indeed filling the large lipophilic Y181–Y188–W229
binding pocket of the NNRTI site, while the chlorine

2962
T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
Table 1. Data for reference compounds 1–2 and amide derivatives 3, 12–17
X
Y
3
O
O
O
R
2
Z
3, 12-17
Compound
X
Y
Z
R
Inhibition of RT
polymerase, IC50 (nM)a
Antiviral activity in cell culture, CIC95 (nM)b
WT K103N Y181C WT
K103N
Y181C
WT
(10% FBS) (10% FBS) (10% FBS) (50% NHS)
1
2
3
—
—
H
—
—
H
—
—
H
—
—
4.05
2.16
10.37
3.16
34.52
2.48
722
16
ND
6.98
ND
ND
3.28
ND
62
2.65
ND
30.48
ND
SO2NH2
SO2NH2
SO2NH2
SO2NH2
SO2NH2
SO2CH3
51.70 192.4
10.50 218
N
H
Cl
Cl
Cl
Cl
Cl
Cl
12
13
14
15
16
17
H
H
H
3-Cl
25
ND
8
ND
ND
ND
9.52
ND
5.26
ND
ND
160
N
H
CN 3-Cl
0.96
2.11
0.14
0.99
0.86
14.4
3.48
0.21
1.33
0.33
1.2
ND
N
H
Cl CN 3-Cl
Cl CN 2-Cl
Cl CN 2-Cl
Cl CN 2-Cl
2.66
0.28
0.91
1.7
16
ND
125
N
H
7.56
31
14.85
ND
23.91
78
N
H
N
H
SO2CH3
3.68
14.26
29.74
N
a Compounds were evaluated in a standard SPA assay. Values are the geometric mean of at least two determinations. ND, not determined. Assay
protocols are detailed in Ref. 8.
b CIC95 (Cell culture inhibitory concentration) is defined as the concentration at which the spread of virus is inhibited by >95% in MT-4 human
T-lymphoid cells maintained in RPMI 1640 medium containing either 10% FBS or 50% NHS. Details of the assay protocols are provided in Ref. 9.
Values are the geometric mean of at least two determinations. ND, not determined. No cytotoxicity was observed for any of the compounds up to
the upper limit of the assay (8.3 lM).
methyl sulfone provided 16, which was essentially equi-
potent with 15. Unfortunately, 15 and 16 are rapidly
cleared in rats and dogs after iv administration, and
metabolic profiling confirmed that the anilide underwent
rapid metabolic hydrolysis.11
back to the nitrogen of the amide appeared to be a rea-
sonable approach, allowing the aliphatic carbons to fill
the same space as the chloro substituent, while also
allowing the amide carbonyl to maintain the key hydro-
gen bonding interaction with the NH of K103. The cyc-
lic amide analog 17 (Table 1) retained excellent intrinsic
and cell based potency, and showed enhanced stability,
longer half-live and lowered clearance after iv dosing
to rats and dogs.11 Unfortunately, the compound also
In an effort to prevent the metabolic hydrolysis of this
critical amide, we examined cyclic amide derivatives.
Cyclization from the ortho position of the aromatic ring

T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
2963
(10.8 h) after iv dosing in rats.11 These results taken in
combination suggested that there was indeed some via-
bility to the concept of replacing the amide functionality
with a heterocycle, however, it was clear that more po-
tency and solubility combined with less lipophilicity
would be needed to make molecules of this type attrac-
tive for further investigation. In an effort to develop no-
vel amide replacements with better physical properties
and the potential for increased potency, we began to
look toward other heterocyclic motifs that might be suit-
able as replacements for the aryl amide. After careful
consideration of the modeling and crystallographic data,
we rationalized that an indazole moiety (Fig. 6) might
provide a novel replacement for this region of the mol-
ecule. The indazole nucleus would have the added po-
tential of making two hydrogen bonding interactions
with both the backbone NH and the backbone carbonyl
of K103 while maintaining many of the lipophilic inter-
actions provided by the aryl amide moiety. Incorpora-
tion of the indazole moiety into the molecule provided
22 (Table 2), which showed high levels of enzyme and
cell-based potency. Compound 22 (Table 3) demon-
strated moderate pharmacokinetics after iv dosing to
rats, dogs, and rhesus macaques and showed low levels
of oral bioavailability after dosing in rats. Various
manipulations of the dosing vehicle formulation have
enhanced this number to 10%, however, this is still less
than desired for further development. The compound
possesses low solubility which appears to be responsible
for the low oral bioavailability. Compound 22 was also
evaluated for its antiviral potency versus a panel of clin-
ical mutations13 (Table 4) in the presence of 40% NHS,
and was shown to possess broad potency versus most of
the mutations tested. Only the rare mutation Y188L (oc-
curs in 0.6% of clinical isolates)14 provided notable de-
creases in potency; this is presumably due to the
intimate interaction between the Western A ring of 22
and the aryl ring of the Y188 residue. When the residue
is mutated to a leucine, much of the interaction between
the inhibitor A ring and the sidechain is lost due to the
decreased size of the sidechain as well as the loss of the
pi-stacking interaction. Crystallographic studies with 22
(Fig. 7)15 confirmed the binding mode proposed earlier
(Fig. 6) for the compound. All of the key lipophilic
interactions of the western A ring and central B ring
are maintained as in previous compounds. The Y181
side chain is rotated by 90° as seen in the earlier crystal
structure of 14. As theorized, the two pyrrolo nitrogens
of the indazole moiety appear to make two direct hydro-
gen bonding interactions with the NH and carbonyl of
Figure 4. The X-ray crystal structure of compound 14 (salmon) bound
˚
to the NNRTI binding site (2.7 A resolution). Details of the crystal-
lographic studies are provided in Ref. 10.
showed low oral bioavailability in both rats and dogs
(<3%), and had low solubility.
Given the issues we had encountered with amides 12–17
(Table 1), we proceeded to alter the structure to obtain
compounds with more favorable profiles. We hypothe-
sized that it might be possible to replace the C-ring
amide portion of the molecules with a series of heterocy-
cles that would use a heteroatom to maintain the key
hydrogen bond acceptor role of the amide carbonyl
(Fig. 5).12 Replacement of the amide functionality with
a series of heterocycles provided 18–22 (Table 2). The
simple benzthiazole analog 18 showed somewhat re-
duced but still encouraging enzyme potency in compar-
ison to the previously prepared amide analogs. Addition
of the chlorine atom (19) to mimic the chlorine of the
amide series resulted in an increase in enzyme potency;
however, antiviral potency remained low. The com-
pound was shifted in the presence of 50% NHS, suggest-
ing high levels of protein binding. In an effort to lower
the lipophilicity and protein binding of these com-
pounds, the benzoxazole analog 20 was prepared, and
showed excellent enzyme potency; however, the modest
cell based potency remained an issue. Moving to the
more polar benzimidazole core (21) provided a com-
pound with enhanced physical properties. However,
the more polar heterocyclic nucleus was not well toler-
ated by the enzyme active site, leading to losses in en-
zyme and antiviral potency. Interestingly, 21 showed
moderate clearance (13.6 ml/min/kg) and a long half life
Figure 5. Proposed binding mode for heterocyclic derivatives 18–21.

2964
T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
Table 2. SAR for heterocyclic derivatives 18–22
Cl
CN
O
O
Het
Cl
18-22
Compound
Het
Inhibition of RT
polymerase, IC50 (nM)a
Antiviral activity in cell culture, CIC95 (nM)b
WT
K103N
79.24
Y181C
WT(10% FBS)
312
K103N(10% FBS)
ND
Y181C(10% FBS)
ND
WT(50% NHS)
1250
S
N
18
19
19.29
3.06
54.92
5.15
Cl
Cl
5.98
1.41
156
78
ND
ND
ND
ND
1250
312
S
N
20
21
22
0.59
3.2
6.38
50
O
N
Cl
H
N
7.30
103
ND
ND
2778
N
1.35
1.12
2.62
22.57
33.44
101.5
114.4
NH
N
a Compounds were evaluated in a standard SPA assay; values are the geometric mean of multiple determinations. ND, not determined. Assay
protocols are detailed in Ref. 8.
b CIC95 (Cell culture inhibitory concentration) is defined as the concentration at which the spread of virus is inhibited by >95% in MT-4 human T-
lymphoid cells maintained in RPMI 1640 medium containing either 10% FBS or 50% NHS. ND, not determined. Details of the assay protocols are
provided in Ref. 9. Values are the geometric mean of multiple determinations. No cytotoxicity was observed for any of the compounds up to the
upper limit of the assay (8.3 lM).
Figure 6. Proposed binding mode for indazole derivative 22.
Table 3. Pharmacokinetic data for compound 22
the backbone of K103. The phenyl ring of the indazole
makes a stacking interaction with Y318, and also ap-
Rata
Dogb
Rhesusc
pears to make lipophilic interactions with the side chains
of F227, P236, and V106A. The bottom edge of this phe-
nyl ring lies directly along the enzyme–solvent interface
and appears to be amenable to further polar
substitution.
T1/2 (h)
4.7
13.6
6.2
2.9
3.3
1.5
8.10
24
Cl (ml/min/kg)
AUC (lM h)
Vd (l/kg)
28.9
0.4
1.7
7.1
ND
1.2
ND
F (%)
a Average of three Sprague–Dawley rats dosed at 10 mpk po (methocel
susp.) and 2 mpk iv (DMSO). All values are within 25% of the mean.
b Average of two beagle dogs dosed at 0.25 mpk iv (DMSO).
c Average of two rhesus macaques dosed at 1 mpk iv (DMSO).
In summary, we have designed and synthesized a series
of novel diaryl ether based NNRTI lead structures
which have excellent potency versus WT and key mutant

T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
2965
Table 4. Antiviral potency of compound 22 versus various clinically
isolated mutant viruses (NL4-3 isolate)a
R.; Miller, M. D.; Moyer, G.; Munshi, V.; Orvieto, F.;
Palumbi, M. C.; Pescatore, C.; Rowley, M.; Williams, P.
D.; Summa, V. Bioorg. Med. Chem. Lett. 2006, 16, 2748;
(b) De La Rosa, M.; Kim, H. W.; Gunic, E.; Jenket, C.;
Boyle, U.; Koh, Y.; Korboukh, I.; Allan, M.; Zhang, W.;
Chen, H.; Xu, W.; Nilar, S.; Yao, N.; Hamatake, R.;
Lang, S. A.; Hong, Z.; Zhang, Z.; Girardet, J. Bioorg.
Med. Chem. Lett. 2006, 16, 4444; (c) Gagnon, A.; Amad,
M. H.; Bonneau, P. R.; Coloumbe, R.; DeRoy, P. L.;
Doyon, L.; Duan, J.; Garneau, M.; Guse, I.; Jakalian, A.;
Jolicoeur, E.; Landry, S.; Malenfant, E.; Simoneau, B.;
Yoakim, C. Bioorg. Med. Chem. Lett. 2007, 17, 4437.
4. Romines, K. R.; Freeman, G. A.; Schaller, L. T.; Cowan,
J. R.; Gonzales, S. S.; Tidwell, J. H.; Andrews, C. W.;
Stammers, D. K.; Hazen, R. J.; Ferris, R. G.; Short, S. A.;
Chan, J. H.; Boone, L. R. J. Med. Chem. 2006, 49, 727.
5. Compounds 2 and 3 were placed in the NNRTI binding
pocket based on prior knowledge. Approximately 30
conformations of each compound were generated using
two algorithms: a distance geometry (1) and a knowledge
based (2) method. All conformations were then subjected to
energy minimization using a distance dependent dielectric
constant of 2 and applying the MMFFs forcefield (3). The
depicted pose of compound 2 was chosen because it
exhibited a bioactive energy conformation and the lowest
overall interaction energy scoreas computed byMMFFs. In
addition, the pose reflected the geometry shown in the
poster presentation in Ref. 4 below (4). The comparison
pose of compound 3 is also a low energy pose with a
concomitant good overall score. In addition, it most closely
resembles the pose of compound 2. These figures were
created in PyMol (5). Modeling discussion references: (1);
(a) Crippen, C. M.; Havel, T. F. In Distance Geometry and
Molecular Conformation; Bawden, E., Ed.; Research Stud-
ies Press: Wiley, New York, 1988; (b) Kuszewski, J.; Nilges,
M.; Brunger, A. T. J. Biomol. NMR 1992, 2, 33; (2) Feuston,
B. P.; Miller, M. D.; Culberson, J. C.; Nachbar, R. B.;
Kearsley, S. L. J. Chem. Inf. Comput. Sci. 2001, 41, 754; (3)
Halgren, T. A. J. Comput. Chem. 1996, 17, 490, 520, 553,
616; Halgren, T. A.; Nachbar, R. B. J. Comput. Chem. 1996,
17, 587; (4) Freeman, G.; Romines, K.; Schaller, L.; Ferris,
R.; Roberts, G.; Short, S.; Weaver, K.; Hazen, R.; Creech,
K.; St. Clair, M.; Tidwell, M.; Cowan, J.; Chamberlain, P.;
Rena, J.; Stuart, D.; Stammers, D.; Andrews, C.; Koszalka,
G.; Burnette, T.; Chan, J.; Boone, L. Abstracts of Papers,
2nd IAS Conference on HIV Pathogenesis and Treatment,
Paris, France, July 13–16, 2003; Poster # 538; (5) PyMol,
v.99 distributed by DeLano Scientific LLC was utilized to
generate this image.
Mutation
IC50 (nM, 40% NHS)
WT
G190A
4.7
9.4
G190S
L100I
23.4
2.6
K103N
Y181C
Y188L
7.9
13.8
>1000
21
K101E/G190A
K103N/G190A
K103N/P225H
K103N/V179E
K103N/Y181C
Y181C/G190A
K103N/Y181C/G190A
21.7
29.3
20.3
141
35.8
171
a Compounds were analyzed in a Monogram Bioscience Phenoscreen
assay in the presence of 40% NHS. The IC50 is defined as the con-
centration of compound in cell culture required to block 50% of viral
replication. Details provided in Ref. 13.
Figure 7. The X-ray crystal structure of compound 22 (blue) in the
˚
WT-NNRTI binding site (2.7 A resolution). Details of the crystallo-
graphic studies are provided in Ref. 15.
viruses. Further systematic manipulation of the lead
structures guided by SAR, modeling, and crystallogra-
phy have resulted in the discovery of compound 22,
which has become the prototype of a potent and novel
NNRTI platform. Synthesis in this series is continuing
in an effort to modulate physical properties and prepare
a more soluble and bioavailable analog of compound
22. Taking advantage of the solvent exposed region of
this molecule as a means to incorporate solubilizing
groups is a key ongoing effort in our laboratories, and
our further efforts in this area will be detailed in subse-
quent publications.
6. Ren, J.; Nichols, C.; Bird, L. E.; Fujiwara, T.; Sugimoto,
H.; Stuart, D. I.; Stammers, D. K. J. Biol. Chem. 2000,
275, 14316.
7. Full experimental procedures and characterization data
are available in the following patent application: Saggar,
S. A.; Sisko, J. T.; Tucker, T. J.; Tynebor, R. M.; Su, D.;
Anthony, N. J. U.S. Patent 2007/0021442, 2007; Chem.
Abstr. 2007, 146, 163149.
8. Full experimental protocols for the HIV-1 RT polymerase
SPA assay is described in the patent application cited in
Ref. 7.
9. Vacca, J. P.; Dorsey, B. D.; Schleif, W. A.; Levin, R. B.;
McDaniel, S. L.; Darke, P. D.; Zugay, J.; Quintero, J. C.;
Blahy, O. M.; Roth, E.; Sardana, V. V.; Schlabac, A. J.;
Graham, P. I.; Condra, J. H.; Gotlib, L.; Holloway, M.
K.; Lin, J.; Chen, I.-W.; Vastag, K.; Ostovic, D.; Ander-
son, P. S.; Emini, E. E.; Huff, J. R. Proc. Natl. Acad. Sci.
U.S.A. 1994, 91, 4096.
References and notes
1. Demeter, L. M.; Doamaoal, R. A. Int. J. Biochem. Cell
Biol. 2004, 36, 1735.
2. Tarby, C. M. Curr. Top. Med. Chem. 2004, 4, 1045.
3. A number of groups have reported the NNRTI activity of
similar structures: (a) Muraglia, E.; Kinzel, O. D.; Laufer,
10. The X-ray diffraction data of the wild type RT and
˚
inhibitor 14 complex crystal were collected at 2.7 A

2966
T. J. Tucker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2959–2966
resolution with Rsym of 0.086% and 86.4% completeness.
Assays performed by Monogram Bioscience, South San
Francisco, CA. Details of the assay protocols are
available: http://www.monogramhiv.com/assays/hcp/
phenoHIVTechnology.aspx.
The structure was refined to an R-factor of 0.168. The
structure has been deposited in the PDB; PDB code is
3C6T.
11. Unpublished results, Drug Metabolism Department,
Merck Research Laboratories.
12. In the paper cited in Ref. 4, there is an example of the
attempted use of a benzimidazole to perform a similar
function.
14. Dykes, C.; Demeter, L. M. Abstracts of Papers, 14th
Conference on Retroviruses and Opportunistic Infections,
Los Angeles, CA, February 25–28, 2007; Poster #633.
15. The X-ray diffraction data of the wild type RT and
˚
inhibitor 22 complex crystal were collected at 2.7 A
13. The compound was evaluated in a Monogram Bioscience
PhenoSense Assay versus a panel of clinically derived
RT mutants in the presence of 40% normal human
serum. Values are the average of two determinations.
resolution with Rsym of 0.103% and 93.4% completeness.
The structure was refined to an R-factor of 0.203. The
structure has been deposited in the PDB; PDB code is
3C6U.