Rational design of potent non-nucleoside inhibitors of HIV-1 reverse transcriptase

Article abstract of DOI:10.1021/jm301294g

A new series of non-nucleoside reverse transcriptase inhibitors based on an imidazole-amide biarylether scaffold has been identified and shown to possess potent antiviral activity against HIV-1, including the NNRTI-resistant Y188L mutated virus. X-ray crystallography of inhibitors bound to reverse transcriptase, including a structure of the Y188L RT protein, was used extensively to help identify and optimize the key hydrogen-bonding motif. This led directly to the design of compound 43 that exhibits remarkable antiviral activity (EC₅₀ < 1 nM) against a wide range of NNRTI-resistant viruses and a favorable pharmacokinetic profile across multiple species.

Full text of DOI:10.1021/jm301294g

Subscriber access provided by Brown University Library
Article
Rational Design of Potent Non-Nucleoside
Inhibitors of HIV Reverse Transcriptase
Andrew J. Peat, Pek Yoke Chong, Dulce M Garrido, Huichang Zhang,
Eugene Stewart, Robert Nolte, Liping Wang, Robert G Ferris, Mark Edelstein,
Kurt L Weaver, Amanda Mathis, Michael Youngman, and Paul Sebahar
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 08 Nov 2012
Downloaded from http://pubs.acs.org on November 10, 2012
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 17
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Rational Design of Potent Non-Nucleoside Inhibitors of HIV-1
Reverse Transcriptase
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Pek Chong, Paul Sebahar†, Michael Youngman, Dulce Garrido, Huichang Zhang, Eugene Stewart, Robꢀ
ert T. Nolte, Liping Wang, Robert G. Ferris, Mark Edelstein, Kurt Weaver, Amanda Mathis and Andrew
Peat*
GlaxoSmithKline Research & Development, 5 Moore Drive, Research Triangle Park, NC 27709, USA
KEYWORDS: non-nucleoside reverse transcriptase inhibitor, NNRTI, antiviral, HIV-1, biarylether, substituted imidazole.
Supporting Information Placeholder
ABSTRACT:. A new series of nonꢀnucleoside reverseꢀtranscriptase inhibitors based on an imidazoleꢀamide biarylether scaffold
has been identified and shown to possess potent antiviral activity against HIVꢀ1, including the NNRTI resistant Y188Lꢀmutated
virus. Xꢀray crystallography of inhibitors bound to reverse transcriptase, including a structure of the Y188L RT protein, was used
extensively to help identify and optimize the key hydrogenꢀbonding motif. This led directly to the design of compound 43 that
exhibits remarkable antiviral activity (EC50 < 1nM) against a wide range of NNRTI resistant viruses and a favorable pharmacokinetꢀ
ic profile across multiple species.
■ INTRODUCTION
The discovery and implementation of highly active antiretroviral therapy (HAART) in 1995 revolutionized the treatment of
HIV/AIDS, significantly improving mortality rates to where it is now considered a manageable, chronic disease. Efavirenz (EFV,
Figure 1a) is a key component of HAART and acts as a nonꢀnucleoside inhibitor of HIVꢀ1 reverse transcriptase (RT), the enzyme
responsible for conversion of viral RNA into doubleꢀstranded DNA. A concern with respect to efavirenz is its low genetic barrier
for viral resistance, whereby a singleꢀpoint mutation (e.g. K103N) renders the drug impotent (Table 1).1 The search for secondꢀ
generation NNRTIs active against efavirenzꢀresistant HIVꢀ1 strains and with higher genetic barriers to resistance has yielded two
additional drugs, etravirine and rilpivirine (Figure 1a, Table 1).2 However, viral resistance to this mechanistic class of inhibitors
continues to emerge and there is a clear need to discover new agents with unique resistance profiles.
In the last decade, efforts at GlaxoSmithKline led to the clinical development of 2 (GW634), a prodrug of a potent nonꢀnucleoside
RT inhibitor 1 (GW248, Figure 1).3 Compound 2 lowered HIVꢀ1 viral RNA titres in NNRTIꢀexperienced patients when given as a
monotherapy over 7ꢀdays.4 However, there were several potential issues associated with 2 that the backup program aimed to adꢀ
dress. First, preclinical studies revealed that the primary pathway for metabolic clearance was amide hydrolysis that liberated the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 17
corresponding carboxylic acid and aniline.5 Recent studies have raised concerns around the potential toxicity associated with aroꢀ
matic amines;6 as such, removing the possibility to form an aniline metabolite became a priority. Second, rash remains a major
tolerability issue associated with efavirenz treatment and a liability that future drugs in this class need to address.7 Although the
evidence linking rash to sulfonamide functional groups remains controversial,8 it seemed prudent to avoid this moiety altogether in
a backup compound. Similarly, reports that benzophenoneꢀcontaining drugs can cause photosensitivity responses in patients
prompted us to target analogs lacking this structural motif.9 Finally, the in vitro resistance profile of 2 revealed the emergence of
viruses with mutations at position 188 of HIVꢀ1 RT.3 One mutation in particular, Y188L, renders the virus resistant to many
known NNRTI’s including efavirenz (Table 1).10 Therefore, we focused on identifying inhibitors that would address these clinical
needs.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. a) Structures of three marketed NNRTIs efavirenz, etravirine, and ripivirine; b) modification of undesirable structural
features (yellow) in a prior clinical asset (2) that led to the envisioned target scaffold I.
■ RESULTS AND DISCUSSION
Of the desired structural changes outlined above, removal of the benzophenone group proved to be the most straightforward. In
addition to our efforts,11 other research groups had previously described inhibitors containing the biarylether linker that was incorꢀ
porated into our design.12, 13, 14 With respect to eliminating potential aniline metabolites, a simple solution was to reverse the amide
connectivity, thereby leading to the envisioned target scaffold I (Figure 1). While it was unclear if this spatial arrangement of hyꢀ
drogenꢀbond donor and acceptor would be tolerated, the potential for rapid analog synthesis and the plethora of readily available
carboxylic acids supported this approach. The requisite benzylamine 6 was prepared in eight steps with reasonable overall yield
(26%; Scheme 1) from the commercially available phenol 3. The reactions performed well on multiꢀgram scale, providing easily
isolated intermediates and access to hundreds of grams of amine 6.
ACS Paragon Plus Environment

Page 3 of 17
Journal of Medicinal Chemistry
Benzylamine 6 was coupled with a diverse set of carboxylic acids to give the desired analogs in good yields.15 Initial results with
an aryl sulfonamide similar to the one that had improved potency within the benzophenone scaffold were disappointing, especially
the absence of activity against the important Y188L virus (compound 7, Table 1). The benzoic acid derivatives in general failed to
deliver an attractive antiviral profile but, in contrast, heteroaryl carboxamide analogs proved much more promising. While the 2ꢀ
benzofuranyl analog 8 also displayed poor antiviral activity, a modest improvement was noted with the 2ꢀindole and 2ꢀ
benzimidazole derivatives (9 and 10, respectively), suggesting that a hydrogenꢀbond donor adjacent to the amideꢀlinker was advanꢀ
tageous. This observation was supported by analogs 11 and 12, in which removal of the hydrogen bond donor via N(1)ꢀmethylation
of the imidazole (11) resulted in a >10ꢀfold loss in potency versus the desꢀMe analog (12).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Synthesis of the requisite benzylamine intermediate 6.
Reagents and conditions: (a) 1,2,3ꢀtrifluoroꢀ4ꢀnitrobenzene, NaH, THF (74%); (b) diꢀtꢀbutylmalonate, NaH, THF; (c) TFA, DCM;
(d) CuO, ∆, MeCN (87% over 3 steps); (e) Na2S2O4, water/THF (92%); (f) CuCl2, tBuONO, MeCN (67%); (g) NBS, AIBN, CCl4
(79%); (h) NH3, MeOH (84%).
The similar activity between the benzimidazole 10 and the imidazole 12 was encouraging since the latter possessed improved physꢀ
icochemical properties with respect to molecular weight and lipophilicity (cLogP, Table 1). A strong emphasis has been placed on
improving the physicochemical properties of drug candidates as a means of reducing attrition in late stage clinical development.16
Substitution at the 4ꢀposition of the imidazole (R1) with a –Me group (13) led to a slight improvement in potency compared to the
protioꢀanalog 12; however, introduction of an electronꢀwithdrawing –CF3 group (14) afforded a compound with subꢀnanomolar
potency versus WT and K103N viruses. More importantly, activity against the Y188L virus was also achieved. Replacing the –
CF3 substituent with other small electronꢀwithdrawing groups such as –Cl and –Br (15 and 16, respectively) led to >45ꢀfold inꢀ
creases in Y188L activity and antiviral profiles that surpassed that of the previous clinical candidate 1. Both compounds display
subꢀnanomolar activity against another common mutation in HIVꢀ1, Y181C, which confers NNRTIꢀresistance.
An Xꢀray crystal structure of analog 15 bound to HIVꢀ1 reverse transcriptase (Figure 2) shows the existence of two optimal Hꢀ
bonds between a) the carbonyl of the imidazole carboxamide and the amide N–H of Lys103 (2.9 Å); and b) the imidazole N–H and
the carbonyl of Lys103 (2.8 Å). The twoꢀpoint binding motif explains the enhanced potency observed for heterocycles bearing an
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 17
N–H group adjacent to the amide carbonyl versus those without (compare compounds 9, 10, 12–16 versus compounds 8 and 11). It
also explains the lack of resistance to the K103N mutation since these inhibitors hydrogen bond to N–H and C=O groups located
along the protein backbone and not with the amino acid sideꢀchain. As shown in Figure 2, the Lys103 side chain is oriented away
from the inhibitor and shows little change in orientation upon mutating to Asn103.
1
2
3
4
5
6
7
8
9
Table 1. Antiviral potency and cLogP of aryl and heteroaryl analogs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EC50 (nM)a
Cmpd
EFV
Ar, Het
R1
WT
K103N
Y181C
Y188L
cLogPb
0.5
20
1.0
157
3.7
ETR
0.6
0.6
2.4
1.0
5.0
TMC278
1
0.5
1.8
0.5
1.9
1.3
0.7
3.7
18
5.5
3.4
7
13
60
180
>1000
4.2
8
9
52
380
26
340
58
>1000
>1000
6.1
6.1
8.9
10
11
5.1
18
27
>1000
>1000
6.4
4.4
120
460
1200
12
H
9.7
48
260
>1000
4.4
13
14
Me
2.4
0.1
7.2
0.7
15
>1000
320
4.4
4.2
CF3
1.4
ACS Paragon Plus Environment

Page 5 of 17
Journal of Medicinal Chemistry
15
16
Cl
Br
0.1
0.1
0.1
0.1
0.2
0.2
5.6
7.1
4.4
4.4
1
2
3
4
5
6
a For details of the assay see Reference 3. Standard deviations are listed in the Supporting Information.
b cLogP was calculated using the model provided in the Daylight/Biobyte software.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Two key binding interactions exist between compound 15 and the backbone N–H and C=O of Lys103 of HIVꢀ1 RT.
The beneficial effect on potency resulting from substitution at the imidazole 4ꢀposition appears multifaceted. Small, lipophilic
groups such as –Me, –CF3, –Cl, –Br (examples 13–16) occupy a small, hydrophobic pocket created by P225, F227, and P236 (Figꢀ
ure 2), thereby providing enhanced binding affinity over the 4–H analog 12. The superior activity of the analogs bearing electronꢀ
withdrawing groups (14–16) versus the analog with an electronꢀdonating group 13 may be in part attributed to the lower pKa of the
imidazole (N–H) proton. The measured pKa’s of the imidazoles of analogs 15 and 13 are 8.9 and >12, respectively.17 Increased
acidity facilitates hydrogenꢀbonding and results in stronger binding of the inhibitor to the protein backbone.
Another factor contributing to the increased activity observed with electronꢀwithdrawing versus electronꢀdonating groups at the
imidazole 4ꢀposition is an energetic preference for a conformation conducive for binding to reverse transcriptase. A number of
rotameric and tautomeric conformations are available to the imidazoleꢀ5ꢀcarboxamide derivatives; however, binding to RT requires
a linear arrangement of hydrogenꢀbond acceptor (C=O) and donor (N–H). Experimental and computational methods have shown
that 4ꢀsubstitution of imidazoles with electronꢀwithdrawing groups favor the comformation desired for RT binding.18, 19 To explore
this hypothesis in the context of our lead series, ab initio calculations were conducted on a related molecule, 4ꢀchloroꢀNꢀethylꢀ
imidazoleꢀ5ꢀcarboxamide, to assess the relative energies of probable conformational and tautomeric forms (Figure 3a). The results
show that the conformation required for RT binding, A, is the lowest energy conformation being 3.4, 8.6, and 10.6 kcal/mol lower
in energy than conformations C, B, and D, respectively.20 This is expected as conformer A appears to minimize destabilizing steric
and electronic interactions apparent in the other conformations. Conversely, ab initio calculations on the analogous Meꢀsubstituted
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 17
analog reveal that the RT active conformation E is not the most stable but is 3.4 kcal/mol higher in energy than conformation G
(Figure 3b). Taken together, the results may account for the 24ꢀ72 fold decrease in activity of the 4ꢀMeꢀcompound 13 relative to
the 4ꢀchloroꢀanalog 15 against wildtype and K103N viruses.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figures 3. Relative energies for probable conformations of a) 4ꢀchloro and b) 4ꢀmethyl Nꢀethylꢀimidazoleꢀ5ꢀcarboxamide as calcuꢀ
lated using ab initio methods.19
Further investigation of the SAR required the syntheses of 4ꢀchloroꢀimidazole 5ꢀcarboxylic acids with substitution at positionꢀ2.
Simple alkyl groups were introduced via de novo construction of the chloroꢀimidazole ring system according to published reports
(Scheme 2).21 Benzyloxyacetaldehyde 17 was subjected to standard Strecker reaction conditions and the resulting aminonitrile was
Nꢀacylated with several acid chlorides in good yields. Heating the Nꢀacylatedꢀαꢀaminonitriles 18–20 in the presence of PPh3 and
CCl4 cleanly afforded the chloroꢀimidazoles 22–24, presumably via intermediate 21.21 Acidꢀcleavage of the Oꢀbenzyl protecting
groups liberated the primary alcohols that were converted to the desired carboxylic acids 25–27 via a twoꢀstep protocol that avoided
issues of imidazole oxidation. The acids were coupled with benzylamine 6 to give the desired final products 28–30 (Table 2).
ACS Paragon Plus Environment

Page 7 of 17
Journal of Medicinal Chemistry
Scheme 2. Syntheses of 2ꢀalkylꢀ4ꢀchloroꢀimidazole carboxylic acids.
1
2
3
4
5
6
7
8
9
Ph
Ph
N
Ph
O
O
a - b
P
CN
O
c
BnO
H
N
R2
R2
Cl
H
N
BnO
18 R2 = Me (65%)
19 R2 = Et (78%)
20 R2 = nPr (81%)
BnO
21
17
Cl
N
Cl
N
R2
R2
d - f
HO
N
H
BnO
N
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
22 R2 = Me
23 R2 = Et
24 R2 = nPr
25 R2 = Me (14% from 18)
26 R2 = Et (26% from 19)
27 R2 = nPr (48% from 20)
Reagents and conditions: (a) NaCN, NH4Cl, NH4OH; (b) R2C(O)Cl; (c) PPh3, CCl4, MeCN, ∆; (d) MsOH, CHCl3; (e) MnO2,
DCM, 1,4ꢀdioxane, ∆, (f) NaClO2, NaH2PO4.H2O, 2ꢀMeꢀ2ꢀbutene, tBuOH, THF.
Unfortunately, the cyclization of αꢀaminonitriles was found to have limited scope and could not be adapted to incorporate a number
of other functional groups at the 2ꢀposition. To further expand the SAR within this series, a more straightforward approach involvꢀ
ing the direct functionalization of the imidazole ring system was explored. Commercially available 2ꢀbromoꢀ4,5ꢀdichloroꢀ
imidazole was SEMꢀprotected (compound 31) then subjected to a oneꢀpot, sequential lithiation/trapping protocol (Scheme 3). Adꢀ
dition of nꢀBuLi at –78 °C affected halogenꢀmetal exchange at the 2ꢀposition and the resulting anion was trapped with one equivaꢀ
lent of an electrophile (TMSꢀCl22 or MeSSMe). Addition of another equivalent of nꢀBuLi afforded a second halogenꢀmetal exꢀ
change that was directed exclusively to the chlorine adjacent to the SEMꢀprotecting group.23 Subsequent treatment with DMF gave
good isolated yields of aldehydes 32 and 33, the latter of which was oxidized to the corresponding carboxylic acid 34. Although
two lowꢀtemperature metallation steps were required, this procedure was used to prepare multigram quantities of 34 (and ultimately
15) and proved to be superior to other methods aimed at directly chlorinating the imidazole. For example, attempts to chlorinate
Scheme 3. Oneꢀpot sequential lithiation/trapping of SEMꢀprotected 2ꢀbromoꢀ4,5ꢀdichloroꢀimidazole.
Reagents and conditions: (a) nꢀBuLi, THF, ꢀ78 °C; (b) TMSꢀCl or MeSSMe; (c) nꢀBuLi, THF, ꢀ78 °C; (d) DMF; (e) NaClO2,
NaH2PO4.H2O, 2ꢀMeꢀ2ꢀbutene, tBuOH, THF; (f) CHO)n; (g) TBSCl, imidazole, DCM.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 17
imidazoleꢀ5ꢀcarboxylic acid methyl ester using NCS gave low yields (<20%) of desired product. The sequential lithiation protocol
also provided access to 4ꢀbromoꢀimidazoleꢀ5ꢀcarboxylic acids (e.g. compound 16) by employing SEMꢀprotected 2,4,5ꢀtribromoꢀ
imidazole as the starting material.
1
2
3
4
5
6
7
8
The treatment of the 2ꢀlithiated imidazole with other electrophiles allowed for the incorporation of polar functionality as in the case
of paraformaldehyde, whereby the corresponding methyl alcohol 35 was formed in 65% yield (Scheme 3). The alcohol was proꢀ
tected as the TBDMSꢀether then treated with nꢀBuLi/DMF to give aldehyde 36. Oxidation to the corresponding acid 37, coupling
to benyzlamine 6, and removal of protecting groups afforded the final compound 38 (Table 2).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Antiviral activity and cLogP of imidazoles substituted at R2.
EC50 (nM)a
Cmpd
15
X
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Br
R2
H
WT
0.1
0.2
0.3
0.3
0.3
5.3
0.1
0.1
K103N
0.1
Y181C
0.2
Y188L
5.6
cLogPb
4.4
28
Me
0.2
0.3
11
4.6
29
Et
0.3
0.8
76
5.2
30
nPr
0.4
1.3
240
24
5.7
38
CH2OH
OH
0.3
1.2
3.3
40
14
57
>1000
0.9
4.7
42
NH2
NH2
0.1
0.1
4.1
43
0.1
0.1
0.5
4.2
a For details of the assay see Reference 3. Standard deviations are listed in the Supporting Information.
b cLogP was calculated using the model provided in the Daylight/Biobyte software.
As shown in Table 2, increasing the size of the alkyl substituents at the 2ꢀposition of the imidazole (–Me, –Et, –nPr) resulted in a
progressive decrease in activity against the Y188L virus (examples 28, 29, and 30, respectively). The deleterious effects of hydroꢀ
phobic groups in this region of the molecule were not surprising given that these substituents project directly toward solvent. Unꢀ
fortunately, appending a polar 2ꢀhydroxymethyl group (38), did not improve activity over the 2ꢀmethyl variant (28). The Xꢀray
crystal structure of 38 bound to HIVꢀ1 RT showed the methylalcohol projecting into solvent with no direct interactions to the proꢀ
tein (Figure 4), suggesting that while such substitution might provide a means to improve aqueous solubility, it was unlikely to sigꢀ
nificantly contribute to the antiviral potency. However, it appeared that an Hꢀbond donor directly attached to the 2ꢀposition of the
imidazole might be able to interact with the backbone carbonyl group of residue 104. Based on this hypothesis, the 2ꢀOH and 2ꢀ
NH2 analogs were synthesized (Scheme 4).
ACS Paragon Plus Environment

Page 9 of 17
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Figure 4. The polar hydroxymethyl group of 38 extends past the carbonyl of K104 and projects into solvent.
Scheme 4. Synthesis of 2ꢀOH and 2ꢀNH2 substituted imidazoles 40 and 42.
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cl
Cl
N
N
N
a - c
d - g
OMe
H
N
OH
31
33
H
N
H
R3
SEM
O
O
39
40
Cl
Cl
N
N
h - i
j - m
N3
H
N
NH2
H
N
N
H
R3
Cl
SEM
O
O
42
41
Cl
R3
=
NC
O
F
Reagents and conditions: Synthesis of 40: (a) NaOMe, CuBr (67%); (b) nꢀBuLi, THF, ꢀ78 °C; (c) DMF (80%); (d) NaClO2,
NaH2PO4.H2O, 2ꢀMeꢀ2ꢀbutene, tBuOH, THF (quant.); (e) EDC, HOBT, 6, DMF; (f) TFA, DCM (56% over 2 steps); (g) BBr3,
DCM (15%). Synthesis of 42: (h) mCPBA; (i) NaN3, DMF (69% over 2 steps); (j) NaClO2, NaH2PO4.H2O, 2ꢀMeꢀ2ꢀbutene,
tBuOH, THF (quant.); (k) TFA, DCM; (l) EDC, HOBT, 6, DMF; (m) Lindlar catalyst, H2 (50 psi), EtOAc (53% from 41).
Initial results were perplexing in that the 2ꢀOH compound 40 displayed no activity against the Y188L virus while the 2ꢀNH2 derivaꢀ
tive 42 remains one of the most potent compounds we have tested. However, upon further examination, solutionꢀphase FTꢀIR exꢀ
periments revealed that the 2ꢀOH analog 40 exists almost entirely as the ketoꢀtautomer.24 Consequently, the imidazolone oxygen is
in close proximity to the carbonyl oxygen of residue 104 and the ensuing electronic repulsion likely accounts for the large drop in
potency. In contrast, the 2ꢀNH2 group exists as the aminoꢀtautomer that is capable of hydrogen bonding with residue 104, leading
to an improvement in Y188L antiviral potency versus the 2ꢀprotio analog (compare Entry 15 to 42, Table 2).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 17
The optimized 2ꢀNH2ꢀimidazole moiety was also combined with other modified benzylamine intermediates (see Supporting Inforꢀ
mation for synthetic details). One minor modification in which the chloro substituent was replaced with bromo on the central aroꢀ
matic ring led to a further, albeit modest, improvement in potency versus the Y188L virus (compound 43, Table 2). An Xꢀray crysꢀ
tal structure of compound 43 bound to RT protein shows the existence of three optimal Hꢀbonds (Figure 5a). In addition to the two
hydrogen bonds to Lys103 described above (3.0 and 2.9 Å), a third interaction exists between the imidazole 2ꢀamino group and the
carbonyl of Lys104 (2.7 Å).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Xꢀray structure of compound 43 with HIVꢀ1 RT highlights: a) threeꢀpoint binding of the aminoꢀimidazoleꢀamide with
residues K103 and K104 (wildtype RT); b) π–stacking and Tꢀstacking of the terminal aryl ring of compound 43 with Tyr188 and
Trp229 (wildtype RT); c) loss of the favorable π–stacking interaction upon mutation of tyrosine to leucine at position 188 (Y188L
RT).
The inhibitorꢀprotein interactions appear highly optimized, to the point that potency is maintained against the challenging Y188L
virus. For the biarylether class of NNRTI’s, significant binding energy is gained through πꢀstacking (3.5 Å) of the terminal aryl
ring of the inhibitor with the phenol of Tyr188 (Figure 5b). Mutation of tyrosine to leucine eliminates this favorable interaction
resulting in decreased activity for the inhibitor, often by as much as 100ꢀ1000ꢀfold. We successfully obtained a Xꢀray crystal strucꢀ
tures of Y188L RT protein,25 including one with compound 43 bound to the NNRTI allosteric site which clearly shows Leu188
packing against the inhibitor aryl ring (Figure 5c). The fact that compound 43 exhibits a small shift in potency (5ꢀfold) against the
Y188L virus (relative to WT virus) suggests that the strong Hꢀbonds made through the imidazole carboxamide moiety are capable
of compensating for the loss of the favorable πꢀstacking interactions.
The Xꢀray crystal structure of 43 bound to the wildtype RT allosteric pocket also reveals a Tꢀstacking interaction (3.7 Å) between
the terminal aryl ring of the inhibitor and the tryptophan ring of W229 (Figure 5b), a residue that is conserved across the NNRTIꢀ
resistant viruses. The favorable Tꢀstacking interaction (3.7 Å) is maintained upon mutation of tyrosine 188 to leucine (Figure 5c);
furthermore, the large volume of the triꢀsubstituted terminal aryl ring helps occupy the void created by the smaller leucine
sidechain. These interactions between the aryl ring and the protein may explain the improved Y188L activity of 43 and the diaꢀ
rylpyridine (DAPY) analogs (e.g. etravirine, rilpivirine) relative to efavirenz which places a nonꢀaromatic cyclopropyl ring in this
ACS Paragon Plus Environment

Page 11 of 17
Journal of Medicinal Chemistry
portion of the binding pocket. The greater torsional flexibility of the biarylether 43 compared to efavirenz may also allow the inꢀ
hibitor to adapt to conformational changes within the NNRTI binding pocket that arise from mutations in the RT protein and thereꢀ
by maintain key binding interactions and antiviral activity.
1
2
3
4
5
6
7
8
Compound 43 is active against a diverse set of NNRTI resistant viruses containing singleꢀ, doubleꢀ, and tripleꢀpoint mutations that
have been reported from clinical or in vitro studies. With the exception of the double Y181CꢀF227C mutated virus (EC50 of 3.2
nM), compound 43 exhibits subꢀnanomolar potency (EC50) against all of the NNRTI resistant viruses tested (33 of 34 viruses; see
Supporting Information).26 However, the antiviral efficacy of NNRTIs in vivo is also dependent upon the free fraction of drug
available in plasma and, therefore, the effect of plasma protein binding on in vitro potency was evaluated. The EC50’s of 43 and
other marketed NNRTIs versus wildtype HIVꢀ1 were shifted in the presence of physiological concentrations of human serum albuꢀ
min (40 mg/mL) and human alphaꢀ1 acid glycoprotein (1 mg/mL) with a rank order of etravirine (13ꢀfold) < efavirenz (23ꢀfold) <
43 (71ꢀfold) < rilpivirine (116ꢀfold).26 A comparison of the proteinꢀadjusted EC50’s (PAꢀEC50) for the four chemotypes across a
diverse set of NNRTI resistant viruses is shown in Figure 6.26 Overall, 43 displays a remarkable resistance profile that is far superiꢀ
or to that of the goldꢀstandard, efavirenz, especially against the most prevalent NNRTIꢀresistant mutation K103N. Furthermore, the
activity of 43 is maintained against viruses bearing mutations at position 181 which have been shown to confer resistance to the
DAPY analogs etravirine and rilpivirine.27 For example, 43 showed less than a 3ꢀfold drop in potency against Y181I and Y181V
(relative to wildtype) while etravirine and rilpivirine exhibited 83ꢀfold and 76ꢀfold changes, respectively. Compound 43 is also
active against viruses bearing Y181 mutations and an additional mutation, including Y181CꢀK103N, Y181CꢀV179F, and Y181Iꢀ
V179F. In contrast, etravirine and rilpivirine are less effective at inhibiting viruses with the Y181ꢀdouble mutations. To the best of
our knowledge, 43 represents one of the most potent NNRTI’s ever described and highlights the level of antiviral activity that can
be achieved with this mechanistic class of inhibitors.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. The proteinꢀadjusted EC50’s (PAꢀEC50) of compound 43 relative to the marketed NNRTI drugs efavirenz, etravirine, and
rilpivirine against a diverse set of NNRTI resistant HIVꢀ1 viruses. The EC50 values and protein shift values are reported in the
Supporting Information.
The in vivo pharmacokinetic profile of compound 43 was examined across multiple species (Table 3). In general, the compound
exhibited a favorable PK profile with low clearance in mouse, dog, and cynomolgus monkey and oral bioavailability of >80% in
mouse and rat. The oral exposure from a 5 mg/kg solution dose was high across all four species resulting in drug plasma concentraꢀ
tions at 24 hours post dose (C24) above the EC50 for all tested NNRTI resistant viruses.
In summary, a new series of nonꢀnucleoside reverseꢀtranscriptase inhibitors based on an imidazoleꢀamide scaffold has been identiꢀ
fied and shown to possess potent antiviral activity against HIVꢀ1, including the resilient Y188Lꢀmutated virus. Xꢀray crystallogꢀ
raphy of inhibitors bound to reverse transcriptase, including the first reported structure of the Y188L protein, was used extensively
to help identify and optimize the key Hꢀbonding motif. This led directly to the design of compound 43 that exhibits remarkable
antiviral activity against a wide range of NNRTI resistant viruses and a favorable pharmacokinetic profile across multiple species.
Subsequent studies with this compound as well as further optimization within this scaffold will be reported shortly.
ACS Paragon Plus Environment

Page 13 of 17
Journal of Medicinal Chemistry
Table 3. In Vivo Pharmacokinetic Profile of Compound 43.
1
2
3
4
5
i.v.a
Vdss
p.o.b
AUC0ꢀα
(ng*h/mL) %F
Cl
t1/2
(h)
3.1
Species (mL/min/kg)
(L/kg)
Mousec
8.2
6
2.1
9605
94
7
8
Ratd
28.6±5.4
3.6±0.4 2.2±0.3
2509
84
9
Doge
2.8±0.1
0.9±0.03 4.3±1.0 14277±939 47±3
1.8±0.1 6.1±1.1 6458±883 34±5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cynoe
4.7±1.3
a Dosed i.v. at 1 mg/kg in DMSO:Solutol:5% mannitol (10:10:80).
b Dosed orally at 5 mg/kg in DMSO:Solutol:5% mannitol (10:10:80).
c Composite sampling; each time point represents n=3 mice.
d i.v.( n=3); p.o.( n=2).
e i.v.( n=3); p.o.( n=3).
ASSOCIATED CONTENT
Experimental details for the syntheses and spectroscopic characterization of the compounds in this paper, as well as details related
to cellular and in vivo studies. This material is available free of charge via the Internet at http://pubs.acs.org. The Xꢀray crystal
structure coordinates for compounds 15 and 38 (wildtype RT) and compound 43 (wildtype and Y188L RT) have been deposited
into the RCSB Protein Data Bank (PDB) at http://www.pdb.org with the following deposition codes: 2YNF, 2YNG, 2YNH, 2YNI.
AUTHOR INFORMATION
Corresponding Author
*Eꢀmail: andy.j.peat@gsk.com. Phone: 919ꢀ483ꢀ6137
Present Addresses
† Myrexis, Inc., 305 Chipeta Way, Salt Lake City, UT, 84108, USA.
ACKNOWLEDGMENT
The author would like to thank Doug Minnick in the Analytical Chemistry Department at GlaxoSmithKline for conducting FTꢀIR
experiments to support this manuscript.
ABBREVIATIONS
HAART, highly active antiretroviral therapy; NNRTI, nonꢀnucleoside reverse transcriptase inhibitor; RT, reverse transcriptase;
EFV, efavirenz.
REFERENCES
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 17
(1) Vrouenraets, S. M. E.; Wit, F.; Van Tongeren, J.; Lange, J. M. A. Efavirenz: a review Expert Opin. Pharmacother. 2007, 8, 851ꢀ871.
1
2
3
4
(2) de Bethune, M.ꢀP. Nonꢀnucleoside reverse transcriptase inhibitors (NNRTIs), their discovery, development, and use in the treatment of
5
HIVꢀ1 infection: A review of the last 20 years (1989ꢀ2009) Antiviral Research 2010, 85, 75ꢀ90.
6
7
8
9
(3) (a) Ferris, R. G.; Hazen, R. J.; Roberts, G. B.; St. Clair, M. H.; Chan, J. H.; Romines, K. R.; Freeman, G. A.; Tidwell, J. H.; Schaller, L. T.;
Cowan, J. R.; Short, S. A.; Weaver, K. L.; Selleseth, D. W.; Moniri, K. R.; Boone, L. R Antiviral activity of GW678248, a novel benzophenone
nonnucleoside reverse transcriptase inhibitor. Antimicrob. Agents Chemother. 2005, 49, 4046ꢀ4051. (b) Andrews, C. W.; Chan, J. H.; Freeman, G.
A.; Romines, K. R.; Tidwell, J. H. Benzophenones as inhibitors of reverse transcriptase PCT Int. Appl. 2001, WO 2001017982A1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(4) Becker, S.; Lelezari, J.; Walworth, C.; Kumar, P.; Cade, J.; NgꢀCashin, J.; Kim, Y.; Scott, J.; St. Clair, M.; Jones, L.; Symonds, W. Antiviral
Activity and Safety of GW695634, a Novel Next Generation NNRTI in NNRTI-resistant HIV-1 Infected Patients, 3rd IAS Conference on HIV Pathꢀ
ogenesis and Treatment, Rio de Janeiro, Brazil, 2005, WePe6.2C03.
(5) de Serres, M.; Moss, L.; Sigafoos, J.; Sefler, A.; Castellino, S.; Bowers, G.; SerabjitꢀSingh, C. The disposition and metabolism of
GW695634: A nonꢀnucleoside reverse transcriptase inhibitor (NNRTI) for treatment of HIV/AIDS Xenobiotica, 2010, 40, 437ꢀ445.
(6) Turesky, R. J. Metabolism and Biomarkers of Heterocyclic Aromatic Amines in Molecular Epidemiology Studies: Lessons Learned from
Aromatic Amines Chem. Research Toxicol. 2011, 24, 1169ꢀ1214.
(7) BlasꢀGarcia, A.; Esplgues, J. V.; Apostolova, N. Twenty years of HIVꢀ1 nonꢀnucleoside reverse transcriptase inhibitors: time to reevaluate
their toxicity Curr. Med. Chem. 2011, 18, 2186ꢀ2195.
(8) Johnson, K. K.; Green, D. L.; Rife, J. P.; Limon, L. Sulfonamide crossꢀreactivity: fact or fiction? Ann. Pharmacother. 2005, 39, 290ꢀ301.
(9) Bosca, F.; Miranda, M. Photosensitizing drugs containing the benzophenone chromophore J. Photochem. Photobiol., B. 1998, 43, 1ꢀ26.
(10) Zhang, Z.; Xu, W.; Koh, Y.; Shim, J.H.; Girardet, J.; Yeh, L.; Hamatake, R.K.; Hong, Z A novel nonnucleoside analogue that inhibits huꢀ
man immunodeficiency virus type 1 isolates resistant to current nonnucleoside reverse transcriptase inhibitors. Antimicrob. Agents Chemother.
2007, 51, 429ꢀ437.
(11) Peat, A. J.; Sebahar, P. R.; Youngman, M.; Chong, P. Y.; Zhang, H. Preparation of Nꢀ(cyanophenoxybenzyl) (hetero)aryl carboxamides as
a nonꢀnucleoside reverse transcriptase inhibitors for treating HIV infection PCT Int. Appl. 2008, WO 2008154271A1.
(12) (a) Sweeney, Z. K; Harris, S. F.; Arora, N.; Javanbakht, H.; Li, Y.; Fretland, J.; Davidson, J. P.; Billedeau, J. R.; Gleason, S. K.; Hirschꢀ
feld, D.; KennedyꢀSmith, J. J.; Mirzadegan, T.; Roetz, R.; Smith, M.; Sperry, S.; Suh, J. M.; Wu, J.; Tsing, S.; Villasenor, A. G.; Paul, A.; Su, G.;
ACS Paragon Plus Environment

Page 15 of 17
Journal of Medicinal Chemistry
Heilek, G.; Hang, J. Q.; Zhou, A. S.; Jernelius, J. A.; Zhang, F.; Klumpp, K. Design of Annulated Pyrazoles as Inhibitors of HIVꢀ1 Reverse Tranꢀ
scriptase J. Med. Chem. 2008, 51, 7449ꢀ7458. (b) Dunn, J. P.; Swallow, S.; Sweeney, Z. K. Nonꢀnucleoside Reverse Transcriptase Inhibitors PCT
US 2004, US20040192704A1.
1
2
3
4
5
6
7
8
9
(13) (a) Tucker, T. J.; Sisko, J. T.; Tynebor, R. M.; Williams, T. M.; Felock, P. J.; Flynn, J. A.; Lai, M.; Liang, Y.; McGaughey, G.; Liu, M.;
Miller, M.; Moyer, G.; Munshi, V.; PerlowꢀPoehnelt, R.; Prasad, S.; Reid, J. C.; Sanchez, R.; Torrent, M.; Vacca, J. P.; Wan, B.; Yan, Y. Discovꢀ
ery of 3ꢀ{5ꢀ[(6ꢀaminoꢀ1Hꢀpyrazolo[3,4ꢀb]pyridineꢀ3ꢀyl)methoxy]ꢀ2ꢀchlorophenoxy}ꢀ5ꢀchlorobenzonitrile (MKꢀ4965): a potent, orally bioavailable
HIVꢀ1 nonꢀnucleoside reverse transcriptase inhibitor with improved potency against key mutant viruses J. Med. Chem. 2008, 51, 6503ꢀ6511. (b)
Saggar, S. A.; Sisko, J. T.; Tucker, T. J.; Tynebor, R. M.; Su, DꢀS.; Anthony, N. J. PCT US 2007, US20070021442.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(14) Jones, L. H.; Middleton, D. S.; Mowbray, C. E.; Newman, S. D.; Williams, D. H. Chemical Compounds. PCT Int. Appl. 2006, WO
2006067587A1.
(15) The structures of the final compounds were supported by 1H NMR and LCMS. Yields refer to isolated compounds in >95% purity as deꢀ
termined by 1H NMR and LCMS.
(16) Meanwell, N. A. Improving Drug Candidates by Design: A Focus on Physicochemical Properties As a Means of Improving Compound
Disposition and Safety Chem. Res.Toxicol. 2011, 24, 1420ꢀ1456.
(17) The pKa’s were measured using a Sirius T3 spectrophotometer according to standard procedure. The experimentally determined pKa valꢀ
ues for compounds 13 and 15 were in good agreement with the calculated values (12.2 and 9.5, respectively) obtained using the ACD Labs pKa
Database version 12 software.
(18) Charton, M. Electrical effects of ortho substituents in imidazoles and benzimidazoles J. Org. Chem. 1965, 30, 3346ꢀ3350.
(19) Ogretir, C.; Yarligan, S. AM1 and PM3 study of the protonation tautomerization and valence tautomerization of some 4ꢀsubstituted imidꢀ
azoles J. Mol. Struct. 1998, 425, 249ꢀ254.
(20) Initial conformers and tautomers were generated by the 3D conversion of 2D SMILES strings using the CONCORD v6.1.1 program folꢀ
lowed by manual adjustment of the relevant dihedrals to provide the starting geometries. These four initial structures were fully optimized using
the GAUSSIAN03 program and utilizing the standard 6ꢀ31G** basis set at the restricted HartreeꢀFock (RHF) level of theory. Stationary points
were confirmed via a singleꢀpoint frequency calculation on the fully optimized geometries using the same basis set and level of theory. Examinaꢀ
tion of the frequencies revealed that all stationary points represented minimum energy geometries.
(21) Zhong, Y.ꢀL.; Lee, J.; Reamer, R. A.; Askin, D. New Method for the Synthesis of Diversely Functionalized Imidazoles from NꢀAcylated
αꢀAminonitriles Org. Lett. 2004, 6, 929ꢀ931.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 17
1
2
(22) In the case of TMSCl, the TMS group acts as a transient protecting group that is cleaved during workup to afford the protioꢀanalog.
3
4
5
6
7
(23) Groziak, M. P.; Wei, L. Multifunctionalization of imidazole via sequential halogenꢀmetal exchange: a new route to purineꢀring analogs J.
Org. Chem. 1992, 57, 3776ꢀ3780.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(24) Spectroscopic FTꢀIR data supporting the predominance of the keto tautomer is available in the Supporting Information.
(25) The first Xꢀray crystal structure of Y188L RT protein was reported by: Hsiou, Y.; Das, K.; Ding, J.; Clark, A. D.; Kleim, J.; Rosner, M.;
Winkler, I.; Riess, G.; Hughes, S. H.; Arnold, E. Structures of Tyr188Leu Mutant and Wildꢀtype HIVꢀ1 Reverse Transcriptase Complexed with
the Nonꢀnucleoside Inhibitor HBY 097: Inhibitor Flexibility is a Useful Design Feature for Reducing Drug Resistance J. Mol. Biol. 1998, 284, 313ꢀ
323.
(26) The EC50 values, protein binding values and standard deviations for compound 43, efavirenz, etravirine, and rilpivirine are listed in the
Supporting Information.
(27) Azijn, H.; Tirry, I.; Vingerhoets, J.; de Béthune, M.ꢀP.; Kraus, G.; Boven, K.; Jochmans, D.; Van Craenenbroeck, E.; Picchio, G.; Rimsky,
L. T. TMC278, a NextꢀGeneration Nonnucleoside Reverse Transcriptase Inhibitor (NRRTI), Active against WildꢀType and NNRTIꢀResistant HIVꢀ
1 Antimicrob. Chemother. 2010, 54, 718ꢀ727.
ACS Paragon Plus Environment

Page 17 of 17
Journal of Medicinal Chemistry
Table of Contents graphic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment