Enantiomeric atropisomers inhibit HCV polymerase and/or HIV matrix: Characterizing hindered bond rotations and target selectivity

Article abstract of DOI:10.1021/jm401202a

An anthranilic acid series of allosteric thumb pocket 2 HCV NS5B polymerase inhibitors exhibited hindered rotation along a covalent bond axis, and the existence of atropisomer chirality was confirmed by NMR, HPLC analysis on chiral supports, and computational studies. A thorough understanding of the concerted rotational properties and the influence exerted by substituents involved in this steric phenomenon was attained through biophysical studies on a series of truncated analogues. The racemization half-life of a compound within this series was determined to be 69 min, which was consistent with a class 2 atropisomer (intermediate conformational exchange). It was further found by X-ray crystallography that one enantiomer of a compound bound to the intended HCV NS5B polymerase target whereas the mirror image atropisomer was able to bind to an unrelated HIV matrix target. Analogues were then identified that selectively inhibited the former. These studies highlight that atropisomer chirality can lead to distinct entities with specific properties, and the phenomenon of atropisomerism in drug discovery should be evaluated and appropriately managed.

Full text of DOI:10.1021/jm401202a

Subscriber access provided by UNIVERSITY OF CALGARY
Article
Enantiomeric atropisomers inhibit HCV polymerase and/or HIV
matrix: Characterizing hindered bond rotations and target selectivity
Steven R. LaPlante, Pat Forgione, Colette Boucher, René Coulombe, James R.
Gillard, Oliver Thomas Hucke, araz jakalian, Marc-Andre joli, George Kukolj, Chris
Lemke, Robert McCullom, Steve Titolo, Pierre Beaulieu, and Timothy stammers
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm401202a • Publication Date (Web): 11 Sep 2013
Downloaded from http://pubs.acs.org on September 16, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Enantiomeric atropisomers inhibit HCV polymerase and/or HIV
matrix: Characterizing hindered bond rotations and target selectivity
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
Steven R. LaPlante,* Pat Forgione, Colette Boucher, René Coulombe, James Gillard,
Oliver Hucke, Araz Jakalian, MarcꢀAndré Joly, George Kukolj, Christopher Lemke,
Robert McCollum, Steve Titolo, Pierre L. Beaulieu, Timothy Stammers
Departments of Chemistry and Biological Sciences, Boehringer Ingelheim (Canada) Ltd.,
2100 Cunard St., Laval, Quebec, H7S 2G5, CANADA
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 25
1
2
3
4
ABSTRACT
5
6
7
8
An anthranilic acid series of allosteric thumb pocket 2 HCV NS5B polymerase
inhibitors exhibited hindered rotation along covalent bond axis’ and the existence of
atropisomer chirality was confirmed by NMR, HPLC analysis on chiral supports, and
computational studies. A thorough understanding of the concerted rotational properties
about these bonds and the influence exerted by substituents involved in this steric
phenomenon was attained through biophysical studies on a series of truncated analogues.
The racemization halfꢀlife of a compound within this series was determined to be 69
minutes, which is consistent with a class 2 atropisomer (intermediate conformational
exchange). It was further found by Xꢀray crystallography that one enantiomer of a
compound bound to the intended HCV NS5B polymerase target while the mirror image
atropisomer was able to bind to an unrelated HIV matrix target. Analogues were then
identified that selectively inhibited the former. These studies highlight that atropisomer
chirality can lead to distinct entities with specific properties, and the phenomenon of
atropisomerism in drug discovery should be considered similar to atomꢀcentered
chirality.
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
INTRODUCTION
An often overlooked source of chirality is atropisomerism, which results from
slow rotation along a bond axis due to steric hindrance and/or electronic factors.1
Atropisomerism often goes both undetected and unsuspected given that enantiomers
1
cannot be distinguished using routine analytical tools (e.g. H NMR, HPLC, etc.).2
Unlike atomꢀcentric chiral centers, hindered bond rotations are not clearly identified by
2
ACS Paragon Plus Environment

Page 3 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
simply viewing a drug’s primary structure. Furthermore, awareness of this timeꢀ
dependent phenomenon has been known to raise concerns for drug discovery and
development projects. Recent reviews have categorized atropisomers into three classes
based on their slow, intermediate and fast rotational properties,4 and it has been
acknowledged that some distinct challenges need to be considered for each of the classes.
Recently, guidelines have been proposed that are consistent with practical and regulatory
perspectives.3,4 Overall, the reviews have espoused the view that atropisomeric
compounds can be developed similar to compounds having classical chiral centers, but
with early detection and proper management.2ꢀ4
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
In this work, we describe methods for detecting and monitoring class 2
atropisomers using NMR, HPLC analysis on chiral supports and the design of related
enantiomeric tool compounds. We also apply a dissection strategy involving truncated
analogues to better understand concerted multiꢀbond rotation properties and the steric
involvement of various substituents. In addition, we also encountered an interesting
observation where one atropisomer binds to the intended target (HCV NS5B polymerase)
and its mirror image enantiomer binds to an unrelated protein (HIV matrix). Analogues
were subsequently identified that inhibited only the former. Overall, this work provides
an example of methods and strategies for addressing the challenges associated with class
2 atropisomers.
RESULTS AND DISCUSSION
The NS5B polymerase of the hepatitis C virus is an attractive target for the
discovery of HCV antivirals.5 During lead optimization efforts on an anthranilic acid
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 25
1
2
3
4
series of thumb pocket 2 allosteric inhibitors, we suspected that compound 1 exhibited
5
6
hindered rotation along the anilide bond (the red colored arrow along the arylꢀamine bond
7
8
9
1
in Figure 1). As expected, simple H NMR spectra and temperature studies could not
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
detect the existence of mirror image enantiomers (Figure 1A). A single set of resonances
was observed for aromatic hydrogens H5 and H6 which shifted in opposite directions
upon heating. A tool compound was therefore synthesized whereby the symmetric
isopropyl moiety (colored blue in Figure 1A for 1) was replaced by a chiral 2ꢀmethylꢀ1ꢀ
methoxy ethyl as in compound 2 (colored blue in Figure 1B). As a result, the
manifestation of atropisomers in 2 combined with the presence of a second chiral element
in the molecule should now lead to the detectable occurrence of diastereomeric species.
1
Indeed, this was confirmed by the H NMR spectrum of 2 which featured two sets of
resonances (e.g. H5A and H5B, H6A and H6B). By inference, compound 1 was thus
1
predicted to display atropisomeric chirality. Given H NMR coalescence temperatures ≥
120 °C, the energy barrier to rotation was estimated to be >18 kcal/mol (Figure 1B).
4
ACS Paragon Plus Environment

Page 5 of 25
Journal of Medicinal Chemistry
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) Shown is the primary structure of the racemic, enantiomeric compound 1,
its H NMR spectrum at 25 °C, and as a function of temperature. The red arrow
1
identifies the bond that experiences atropisomeric hindered rotation. The green arrow
points out the conformeric tertiary amide bond. The blue colored atoms and bonds
highlight substituent differences between compounds 1 and 2. (B) Shown is the primary
structure of compound 2 which is atropisomeric and a diastereomeric mixture of two
compounds. The labels 5A and 6A are resonance assignments for one diasteriomeric
conformation whereas 5B and 6B are for the other. The magenta circles denote the
temperature at which resonances coalescence. (C) Shown is the primary structure of the
related analogue 3 that lacks the carboxylic acid. The magenta circle denotes the
temperatures at which the resonances coalesce.
The existence of atropisomer chirality was also confirmed by HPLC on a chiral
support (as shown in Figure 2A) where two peaks (A and B) were observed for a related
analog 4. Subsequently, the halfꢀlife for atropisomer interconversion was measured by
isolating the single enantiomer from peak A followed by reꢀinjection to measure the rate
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 25
1
2
3
4
5
6
7
8
9
of conversion to the other enantiomer (peak B). The decay of A was measured over time
(Figure 2), to calculate a halfꢀlife of 69 minutes and a rotational energy barrier of 22ꢀ23
kcal/mol (Figure 2B).
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. HPLC on a chiral support was used to determination of atropisomer halfꢀlife
for compound 4. Following isolation of the enantiomers of 4 using chiral HPLC, peak A
was reꢀinjected and HPLC chromatograms acquired as a function of time. The ratio of
observed peak areas for A and B versus time is shown on the right of Figure 2A and
plotted in Figure 2B. An apparent halfꢀlife of 69 minutes was calculated using the
exponential decay equation provided in Figure 2B (t1/2 = ln 2 / decay constant = 0.69 /
0.01).
Possible sources of the observed hindered rotation along the amideꢀaryl bond
were then investigated. The less sterically hindered analogue 3 was synthesized to study
the potential steric influence of the orthoꢀacid group (Figure 1). This served as a control
compound that was highly related and likely not atropisomeric. One purpose was to help
focus on and evaluate the properties of the teriary amide bond. The 1H NMR spectrum of
6
ACS Paragon Plus Environment

Page 7 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3 in Figure 1C lacks an extra set of resonances expected from atropisomerism, and had
two broadened resonances for hydrogen 5 due to cis and trans amideꢀbond orientations.
Refer to Figure 5 for visual models of the cis (e.g. conformer A) and trans (e.g.
conformer B) amide orientations. Resonance coalescence was observed at a much lower
temperature for compound 3 (Figure 1C) which is consistent with a lower energy barrier
in the range of 12ꢀ16 kcal/mol for the tertiary amideꢀbond rotation. This dissection and
characterization of the rotational behavior of the tertiary amide bond was also
corroborated by ROESY NMR data of compound 3 in Figure 3B where crossꢀpeaks were
noted between the 5cis and 5trans resonances. These crossꢀpeaks report rotational exchange
on the order of the msec to sec timescale.
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
The dynamic properties and interplay of both the anilide and the tertiary amide
bonds could also be revealed. Figure 3A shows two clear sets of distinct resonances in
1
the 1D H NMR spectrum that represent atropisomers for compound 2 Furthermore, no
ROESY crossꢀpeaks were detected between the sets of atropisomer resonances, which
was consistent with slow bond rotation (minute timescale or longer). As expected,
compound 2 also displayed a low population of the transꢀconformation of the tertiary
amide in its 1D 1H NMR spectrum. The presence of ROESY crossꢀpeaks is indicative of
relatively fast bond rotation (msec to sec timescale). Thus, the NMR data are consistent
with compound 2 having a predominantly cisꢀamide orientation and atropisomerically
hindered rotation along the anilide bond. Compound 3, on the other hand, displays both
cisꢀ and transꢀamide orientations in equivalent populations (Figure 1C). This hindered
bond has a lower rotation barrier, and the compound has no atropisomeric properties
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 25
1
2
3
4
5
6
along its anilide bond. The above analyses shows that both bonds in this series of
compounds experience hindered rotation but on different timescales.
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 3. 2D NMR ROESY spectra of compounds 2 and 3.
The impact of these conformational properties on biological activities, bioactive
conformation, selectivity and chirality was then addressed. This involved an analysis of
the influence of the anilide Nꢀalkyl (R) group using a series of related control analogs that
differed only at the amide R group (blue in Figure 4). As expected, the least hindered,
unsubstituted amide compound 5 (R=H) had no evidence of atropisomerism. The weak
activity (IC50=15 ꢀM against HCV polymerase) was largely explained by ROESY NMR
data that showed the predominant freeꢀstate conformation did not correspond the
bioactive Lꢀshape conformation that is required for HCV polymerase activity (vide infra,
Figure 7A). When free in solution, the ROESY data lacked the reporter crosspeaks (i.e.
between resonances of the cyclohexyl and the H3 aromatic proton of the benzoic acid
group) that were signatory of the bioactive conformation (see the top right of Figure 4).
8
ACS Paragon Plus Environment

Page 9 of 25
Journal of Medicinal Chemistry
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
(ppm)
OH
CF3
2.0
2.0
O
O
No bioactive
4.0
conformation
R
N
6.0
6.0
bioactive
conformation
O
8.0
8.0
(ppm)
8.0
6
8.0 6.0
HCV
HCV
(ppm)
polymerase replicon
HIV matrix
Compound
R
2.0
2.0
IC50 (ꢀM) EC50 (ꢀM)
IC50 (ꢀM)
Bioactive
4.0
conformation
6.0
6.0
H
15.0
-
5
6
>95
8.0
75-fold
8.0
(ppm)
6.0
8.0 6.0
CH3
0.20
0.19
0.13
1.1
>100
(ppm)
2.0
2.0
Bioactive
0.64
7
8
>175
-
4.0
conformation
9-fold
0.070
6.0
6.0
N
8.0
(ppm)
6.0
.0
8.0 6.0
Figure 4. Shown are analogs 5ꢀ8 which differ by the R groups shown in blue.
Enzymatic activities against HCV polymerase and HIV matrix are provided, along with
activities in the cellꢀbased HCV g1b replicon assay. ROESY NMR data are shown on
the right for three compounds. The predominance of the bioactive conformation in the
free state results in ROESY crosspeaks between the aryl and cyclohexyl group as shown
above for compounds 6 and 8.
However, with R=CH3 as in compound 6, the activity improved 75ꢀfold (IC50 =
0.2 ꢀM) and the predominant freeꢀstate conformation corresponded to the bioactive Lꢀ
shape. Crosspeaks were noted between the cyclohexyl and the benzoic acid group (see
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 25
1
2
3
4
5
6
7
8
9
circle at the center right in Figure 4). A general model of this bioactive cis tertiary amide
conformation, which is required for HCV polymerase activity, is displayed as conformer
A in Figure 5, vide infra). Thus, the methyl substitution introduced a rotational barrier
along the tertiaryꢀamide bond that promoted both the bioactive conformation and the
biological activity.
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
Notably, further bulk at this position resulted in a similar promotion of the
bioactive conformation and biological activity. The enzymatic activities (IC50) of
compounds 6, 7 and 8 were similar suggesting that all bulky Rꢀgroups force the tertiary
amide into the bioactive cis conformation. For example, similar ROESY peaks are
shown for compounds 6 and 8 in Figure 4. The data are consistent with a rigidification
role for the Rꢀgroup that is solvent exposed in the bound state. For example, the large R
group of compound 8 had comparable potency to analogs with smaller substituents and
did not cause steric clashes within the receptor pocket (refer to Figure 7A). The R
position was permissive and provided opportunities for improvement of the biological
properties of the molecules, such as cellꢀbased replicon potency as shown for the
piperidine analog 8. Overall, this study determined that the steric properties of the Rꢀ
group, acid, and cyclohexyl together contribute to hinder rotation along the anilide and
tertiary amide bonds and favor the bioactive conformation in the free state.
The aboveꢀdescribed exchange phenomena were further explored by calculations
of the energy profiles along the anilide bond (torsion angle 1) and the tertiary amide bond
(torsion angle 2) (see Figure 5). Using the same method as described elsewhere,2 a
torsion scan was computed and model views of the lower energy states are displayed as
conformers AꢀD. These calculations served to provide valuable predictions of the
10
ACS Paragon Plus Environment

Page 11 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
rotation energy barriers. The lowest barrier to rotation determined for torsion 2 was 16.2
kcal/mol (green in Figure 5). The calculations also predicted the cis conformation to be
0.3 kcal/mol lower in energy than trans. Although the energy difference of 0.3 kcal/mol
is likely not significant given the precision of the method, experimental evidence by
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
1
NMR confirms the significant predominance of the cis orientation in solution (1D H
NMR spectrum in Figure 3A). This cisꢀamide orientation for torsion 2 must be forced
and stabilized by steric bumping of the COOH with the cyclohexyl resulting in a barrier
of 27.3 kcal/mol (Figure 5, direct conformer A to D racemization). An energy minimum
for torsion 2 occurs when the amide oxygen is anti to the phenyl. However, given a trans
orientation along torsion 2, rotation along torsion 1 would involve a lower barrier to
rotation of 21.8 kcal/mol.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 25
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. Displayed are four model conformers AꢀD that result from hindered rotation
along torsion angles 1 (red) and 2 (green) as defined above. The calculated lowest energy
barriers to rotation are provided. The calculation were applied on the simplified model
compounds shown above, and the same procedure/method was described elsewhere.2
The combination of the above theoretical and experimental data was useful for
formulating a plausible interpretation of the dynamic interplay of torsion angles 1 and 2
(Figure 5). For this series of compounds, the process of atropisomer racemization along
torsion angle 1 implicates two rotations along torsion angle 2. The preferred path from
conformer A to D (Figure 5) likely proceeds via the lower rotational barriers involving
conformers B and C. See the concerted rotation path as illustrated. First, the amide bond
must rotate from the predominant cis orientation to the trans orientation (conformer A to
B). The anilide bond can then rotate more fluidly and frequently (msec to sec) given less
12
ACS Paragon Plus Environment

Page 13 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
steric hindrance and lower energy barrier (conformer B to C). This is then followed by
rotation of the amide bond back to the more stable cis orientation (conformer C to D).
This concerted bond mechanism involves a barrier of 21.8 kcal/mol and is
consistent with the experimentally derived rotational atropisomeric energy barrier of 22ꢀ
23 kcal/mol. Otherwise, a direct rotation of conformer A to D would be calculated to be
much higher at 27.3 kcal/mol. Thus, atropisomer interconversion likely occurs via the
lessꢀpopulated transꢀamide conformation, which gives access to the lower barrier of
rotation of torsion 1. Thus, the lowerꢀenergy path to rotation is preferred and implicates
three bond rotations.
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
In addition to the mechanistic studies described above, we also noted an
intriguing feature of anthranilic acid thumb pocket 2 polymerase inhibitors that resulted
from the atropisomeric properties inherent to some of the analogs. A highꢀthroughput
screen of our corporate compound collection that aimed to identify inhibitors of HIV
matrix in virion morphogenesis, an essential process for viral replication (see Supporting
Information),6,7 uncovered hits from a subset of the anthranilic acid series. Figure 6
shows anthranilic acid analog 9 (HIV matrix IC50 = 3.4 ꢁM) bound to HIV matrix as
detected by 15N TROSY NMR, which was an effective filtering step in our hit triaging
15
protocol. An overlay of the N TROSY data of apo matrix (blue) with a spectrum after
addition of compound 9 (red) demonstrated that specific peaks shifted positions (blue
peaks not overlapped with red peaks). Many of these peaks were from residues located
near the phosphatidylinositol 4,5ꢀbisphosphate (PIP2) substrate binding site of matrix
(see Supporting Information). This technique was used to illustrate that a subset of
anthranilic acid derivatives were also binders and inhibitors of HIV matrix. The binding
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 25
1
2
3
4
site and Lꢀshaped ligand bioactive conformation of compound 9 bound to HIV matrix
5
6
were corroborated by an Xꢀray structure of the complex (bottom right of Figure 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
15
Figure 6. Binding to HIV matrix. On the left is a superposition of N TROSY spectra
of apo matrix (blue) with compounds 9 added to matrix (red). On the right is an Xꢀray
structure of compound 9 bound to matrix at the PIP binding site (see Supporting
Information).
An intriguing atropisomeric feature of these compounds resulting from
atropisomerism was noted when comparing the Xꢀray structure of compound 9 bound to
HCV polymerase versus HIV matrix (Figure 7). At the bottom of Figure 7B, the chiral
conformation labeled “acid on the left” was found to bind HIV matrix, whereas the
opposite atropiomeric orientation labeled “acid on the right” was noted to bind to HCV
polymerase (bottom of Figure 7A). Thus, one atropisomer of compound 9 bound to an
HCV target, and the mirror image bound to an unrelated HIV target. Note that both
chiral compounds had comparable enzymatic and cell culture activity in the two unrelated
14
ACS Paragon Plus Environment

Page 15 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
biological systems and bound to the receptors in the bioactive Lꢀshape (cis amide
conformation).
7
8
9
Our awareness and study of atropisomerism properties of tertiary anthranilic acid
amide derivatives proved beneficial in the context of these observations. A subset of
analogs was identified that were specific inhibitors of HCV NS5B and did not inhibit
HIV matrix. This divergent SAR was exemplified in Figure 4 where compounds 5ꢀ7
were inactive in the matrix assay (IC50 >100 ꢀM).
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 7. Compound 9 is shown along with inhibition activities against HCV
polymerase (A) and HCV matrix (B). Xꢀray crystal structures are shown with compound
9 and the receptors (top), and the bound conformation of the compound for visual
purposes (bottom). Coordinates have been deposited in the Protein Data Bank with
accession code 4JVQ for HCV polymerase and 4JMU for HIV matrix.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 25
1
2
3
4
5
6
7
8
9
The work described above presents a good example of the evaluation of a class 2
atropisomeric compound series and its properties. This class represents atropisomeric
compound types that can cover a wide range of rotational rates. Compounds that have
intermediate rotational rates can be challenging, and the relevant drug discovery
implications must be considered on a compound to compound basis. In this case, being
able to inhibit both HCV and HIV might be a desirable and beneficial feature, but one
should consider that specific binding of an atropisomer to an unwanted target is also
possible and may lead to unwanted adverse effects depending on the target. It is also
worth mentioning that unseparated nonꢀinterconverting or rapidly interconverting
atropisomers should be considered as racemic mixtures (i.e. 50% pure compounds) with
all the implications (e.g. potential toxicity by binding of one of the isomers to unwanted
targets, difficulties in generating crystal forms in development, or stability and challenges
in interpreting ADMEꢀPK data where one enantiomer could affect the properties of the
other: e.g. one enantiomer may inhibit CYP and the other inadvertently appear more
stable). However, the development of atomicꢀcentric racemic drugs has been a successful
strategy and can serve as a reference for atropisomeric racemates.4
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
It is also valuable to remark that when atropisomerism is detected early during
lead optimization, a production medicinal chemistry evaluation can be beneficial. In this
case, the isomers are in intermediate conformational exchange and the receptor is able to
select the preferred one out of the mixture. However, even though many of the
compounds in this series had the potential to bind to both targets, it was still possible to
generate divergent SAR and to identify analogs that were specific for the intended target
by capitalizing on other structural elements.
16
ACS Paragon Plus Environment

Page 17 of 25
Journal of Medicinal Chemistry
1
2
3
4
CONCLUSIONS
5
6
7
8
This work summarizes the detection and mechanistic elucidation of atropisomeric
chirality in a series of anthranilic acid thumb pocket 2 HCV NS5B allosteric inhibitors.
Based on the ensemble of information, a forward perspective could be developed. The
compounds discussed in this work had rotational t1/2 of approximately 69 minutes and
therefore were considered Class 2 atropisomers that experience “intermediate
interconversion”.4 Based on practical limitations, the isomers could not be isolated and
tested separately due to this interconversion under physiologically relevant conditions.
Thus, further studies should employ racemic mixtures. This work can serve as a model
example for how drug discovery groups may approach compounds or series that are class
2 atropisomers.
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
MATERIALS AND METHODS
General compound characterization. Procedures for compound syntheses and
characterization are described in the Supporting Information and elsewhere.11 NMR
1
spectra were recorded on a Bruker AVANCE II (400 MHz for H NMR) spectrometer
and were referenced to either DMSOꢀd6 (2.50 ppm) or CDCl3 (7.27 ppm). Mass spectra
were obtained from a Micomass AutoSpec instrument using ES as ionization mode.
Purification of crude material was performed either by flash column chromatography or
by using a CombiFlash Companion using RediSep silica or SilicaSep columns according
to preprogrammed gradient and flow rate separation conditions in hexane/EtOAc or
DCM/MeOH. The final compounds were purified by preparative HPLC on a Waters 2767
sample manager with pumps 2525, column fluidics organizer (CFO), PDA detector 2996,
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 25
1
2
3
4
5
6
7
8
9
and MassLynx 4.1 using either a Whatman Partisil 10ꢀODSꢀ3 column, 2.2 cm × 50 cm or
a YMC CombiꢀPrep ODSꢀAQ column, 50 mm × 20 mm i.d., S, 5 ꢂm, 120 Å, and a linear
gradient program from 2% to 100% AcCN/water (0.06% TFA). Fractions were analyzed
by analytical HPLC, and the pure fractions were combined, concentrated, frozen, and
lyophilized to yield the desired compound as a neutral entity or the trifluoroacetate salt
for basic analogues. Inhibitor HPLC purity was measured by using a Waters Alliance
2695 separation module with a Waters TUV 2487 UV detector. The column was a
Combiscreen ODSꢀAQ, 5 ꢂm, 4.6 mm × 50 mm, linear gradient from 5% to 100%
ACN/H2O + 0.06% TFA in 10.5 min, detection at 220 nm. All final inhibitors had HPLC
homogeneity of ≥95% unless noted otherwise (see Supporting Information). High
resolution exact mass determinations (HRMS) were performed using either an 6220
AccurateꢀMass TimeꢀOfꢀFlight LC/MS with Atmospheric Pressure Photoionization
(APPI+) via flowꢀinjection analysis or a Bruker microOTOFꢀQ II with electrospray
ionization (ESI+) via LCꢀMS
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
NMR spectroscopy. Samples used for Figures 1, 3 and 4 were prepared by
dissolving 0.5 mg in DMSOꢀd6. All NMR spectra were acquired on Bruker AVANCE III
600 or 400 MHz spectrometers using standard pulse programs and parameters.
1
Assignment of the inhibitor H resonances were secured by careful analysis of standard
1
1D H COSY, ROESY, HSQC and HMBC experiments. ROESY data were acquired
with a continuous spinꢀlock during mixing (300 msec), a 2.5 kHz spinꢀlock field, the
carrier set at 4.7 ppm, 400 t1 data points from the addition of 32 transients, a 5200 Hz
sweep width and a relaxation delay of 1.2 sec
18
ACS Paragon Plus Environment

Page 19 of 25
Journal of Medicinal Chemistry
1
2
15
15
3
Samples for the N TROSY experiments with N matrix (20 mM), 50 mM sodium
phosphate pH 5.5. Inhibitor 9 or PIP2 were added via DMSO stocks such that a ligand to
matrix of 10:1 was used with a total of 1% DMSO. Resonance assignments were
secured (Figure S1, Supporting Information) from the literature.7,10
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
Chiral HPLC determination of the atropisomer half-life of compound 4. The
racemate of compound 4 was analyzed by chiral HPLC using the column ChiralꢀAGP,
5ꢁm, 4.0 x 150 mm. The eluent was 19% (CH3CN) : 81% (0.1%CH3COOH, pH 4.3).
The injection volume was 50 ꢁL, flow was 1mL/min, run time was 15 min, detection was
PDA (192ꢀ320nm) Res. 1.2nm, and integration at 225nm. Two wellꢀseparated peaks (A
and B) corresponding to compound 4 were observed. Peak A was isolated and quickly
reꢀinjected into the chiral HPLC and peaks A and B were integrated. Reinjections were
performed from the initial isolated peak A and peaks integrated as indicated in Figure 2.
The ratios of peaks A/B were determined at time points of 20, 50, 100, 150, 165 and 200
minutes. The percent of A as a function of time is plotted in Figure 2. The equation
employed to determine the atropisomer halfꢀlife was the following: y = 24.82*exp)ꢀ
0.01*x)+51.24. An R2=0.9971 was calculated for the data in Figure 2. The halfꢀlife of
peak A (rotation from A to an equilibrium of A and B) was 69 minutes.
HIV matrix construct, expression and assay. Matrix protein (amino acid sequence
1ꢀ132) from HIVꢀ1 strain NL4.3 was inserted in pET11a vector and expressed in
BL21(DE3) cells or in 15N labeled bacterial growth medium for use in TROSY
experiments. Purification was as described previously7 with modifications. Cell pellet
was lysed in 20 mM TrisꢀCl pH 8.0, 500 mM NaCl, 5 mM imidazole, protease inhibitor
tablet (without EDTA) and sonicated before centrifugation at 39,000 x g for 30 min.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 25
1
2
3
4
5
6
7
8
9
Supernatant was passed on a HisTrap HP column and eluted with 1 M imidazole. Eluate
was diluted to 100 mM NaCl and passed on sequential HiTrap DEAE FF and HiTrap SP
columns. Columns were disconnected and protein eluted from SP column with 500 mM
NaCl. The eluted protein was dialyzed against 20 mM TrisꢀCl pH 8.0, 200 mM NaCl,
centrifuged 20,000 x g for 30 min. and stored at ꢀ80°C. Protein concentration was
determined in 6 M GuHCl (1:20 dilution) using ε=16740 and MW=15665 yielding a
stock of 1.04 mg/ml.
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
The fluorescence polarization assay was based on the reported interaction of matrix
with phosphatidylinositol (PI) 4,5 bisphosphate [PI(4,5)P2]2 and established in order to
measure this binding and test compounds which interfere with the interaction. Corning
384ꢀwell black plates (#3654) contained 500 nM MAꢀHis and 15 nM BodipyꢀTMR
PI(4,5)P2ꢀC6 (Echelon Biosciences) with or without compound in 50 mM TrisꢀCl pH
8.0, 50 mM NaCl, 0.01% Igepal CAꢀ630, 1 mM DTT, 1% DMSO. Plate was incubated
for 30 minutes at RT and then mP signal read on Perkin Elmer Wallac Plate Reader (0.2
sec). Apparent affinity of MAꢀHis WT with BodipyꢀPI(4,5)P2ꢀC6 was ~1 ꢀM. See also
refences 6 and 7 for more details.
HCV NS5B polymerase constructs for IC50 activity assays and X-ray
crystallography. HCV genotype 1b NS5B strain 1bꢀ40 and J4 were used to clone into
baculovirus and E. coli expression vectors.8 The full length HCV NS5B polymerase used
for compound IC50 determination was expressed as a hexaꢀhistidine fusion protein in
baculovirus infected cells and purified by metal chelating chromatography on Ni2+
column followed by DEAEꢀsepharose and heparinꢀsepharose chromatography. All other
details are provided elsewhere.9 The genotype 1b J4 strain NS5B lacking the Cꢀterminal
20
ACS Paragon Plus Environment

Page 21 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
21 amino acids and encoding a Cꢀterminal hexaꢀhistidine tag (HCVNS5B1bJ4ꢀꢃC21ꢀH6)
was used for Xꢀray crystallography. This construct was expressed, purified and
concentrated as previously described.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
Determination of X-ray structures. Crystallization of apo HCV NS5B
polymerase and inhibitor soaking studies proceeded as described elsewhere.9 Also, the
crystallization and structure determination of HIV matrix were based on previous
reports.10 Coordinates have been deposited in the Protein Data Bank with accession code
4JVQ for HCV polymerase and 4JMU for HIV matrix.
ASSOCIATED CONTENT
Supporting Information
Supporting Information available that includes information for the HIV matrix assay,
NMR data that employed HIV matrix, and compound characterization data. This
material is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
Eꢀmails: stevenrlaplante@gmail.com and RESGeneral.LAV@boehringerꢀingelheim.com
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 25
1
2
3
4
5
6
7
8
9
We are grateful to S. Mason, N. Aubry, M. Cordingley, P. Edwards, W. Watkins and R.
Bethel for encouragement and support. The contributions and teamwork from our
colleagues at Boehringer Ingelheim were also appreciated.
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
ABBREVIATIONS USED
ADMEꢀPK, adsorption, distribution, metabolism, excretion, pharmacokinetics; COSY,
correlated spectroscopy; DMSO, dimethylsulfoxide; EDTA, ethylenediaminetetraacetic
acid; HCV, hepatitis C virus; HIV, human immunodeficiency virus; HPLC, highꢀpressure
liquid chromatography; HMBC, heteronuclear multipleꢀbond correlation; HMQC
heteronuclear multiple quantum coherence; NMR, nuclear magnetic resonance; NS,
nonstructural; PIP2, phosphatidylinositol 4,5ꢀbisphosphate; ROESY, rotatingꢀframe
nuclear Overhauser effect; SAR, structureꢀactivity relationship; relaxation time; TROSY,
transverse relaxation optimized spectroscopy.
REFERENCES
(1) Oki, M., Topics in Stereochemistry, vol. 14; Ed.; Allinger, N. L.; Eliel, N. L.;
Wilen, S. H., 1983, 1ꢀ81; b) Christie, G. H.; Kenner, J., The Molecular
Configurations of Polynuclear Aromatic Compounds. Part I. The Resolution of
6:6'ꢀDinitroꢀ and 4:6:4':6'ꢀTetranitroꢀ diphenic Acids in Optically Active
Compounds, J. Chem. Soc. 1922, 121, 614ꢀ620; c) Kuhn, R., Molekular
Asymmetrie, In Stereochemie, Freudenberg, H. (ed)., Franz Deutike, Leipzigꢀ
Wein, 1933, 803ꢀ24.
22
ACS Paragon Plus Environment

Page 23 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(2) LaPlante, S.R.; Edwards, P.J.; Fader, L.D.; Jakalian, A.; Hucke, O., Revealing
atropisomer axial chirality in drug discovery, ChemMedChem, 2011, 6, 3, 505ꢀ
513.
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
(3) Clayden, J.; Moran, W.J.; Edwards, P.J.; LaPlante, S.R., The challenge of
atropisomerism in drug discovery, Angewandte Chemie, International Edition,
2009, 48, 35, 6398ꢀ6401.
(4) LaPlante, S.R., Fader, L.D.; Fandrick, K.R.; Fandrick, D.R.; Hucke, O.r; Kemper,
R.; Miller, S.P.F.; Edwards, P.J., Assessing Atropisomer Axial Chirality in Drug
Discovery and Development, Journal of Medicinal Chemistry 2011, 54, 20, 7005ꢀ
7022.
(5) a) Sofia, M. J.; Chang, W.; Furman, P. A.; Mosley, R. T.; Ross, B. S., Nucleoside,
nucleotide, and nonꢀnucleoside inhibitors of hepatitis C virus NS5B RNAꢀ
dependent RNAꢀpolymerase, J. Med. Chem. 2012, 55, 2481. b) Beaulieu, P.L.,
“Design and Development of NS5B Polymerase NonꢀNucleoside Inhibitors for
the Treatment of Hepatitis C Virus Infection” in Successful Strategies for the
Discovery of Antiviral Drugs; Desai, M. C.; Meanwell, N. A. Eds.; RSC:
Cambridge, 2013; RSC Drug Discovery Series Vol. 32, pp. 248. c) Choo, Q.ꢀL.;
Kuo, G.; Weiner, A. J.; Overby, L. R.; Bradley, D. W.; Houghton, M., Isolation of
a cDNA clone derived from a bloodꢀborne nonꢀA, nonꢀB viral hepatitis genom,
Science 1989, 244, 359; d) Lavanchy, D., The global burden of hepatitis C, Liver
Int. 2009, 29, 74; c) O’Brien, C.; Godofsky, E.; RodriguezꢀTorres, M.; Afdhal,
N.; Pappas, S. C.; Pockros, P.; Lawitz, E.; Bzowej, N.; Rustgi, V.; Sulkowski, M.;
Sherman, K.; Jacobson, I.; Chao, G.; Knox, S.; Pietropaolo, K.; Brown, N. A.,
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 25
1
2
3
4
5
6
7
8
9
Randomized trial of valopicitabine (NM283), alone or with PEGꢀInterferon, vs.
retreatment with PEGꢀinterferon plus ribavirin (PEGIFN/RBV) in hepatitis C
patients with previous nonꢀresponse to PEGIFN/RBV: First interim results,
Hepatology 2005, 42, 234A; e) Huang, Z.; Murray, M. G.; Secrist, J. A., III,
Recent development of therapeutics for chronic HCV infection, Antiviral Res.
2006, 71, 351.
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
(6) Saad, J.S.; Miller J.; Tai J.; Kim A.; Ghanam, R.H.; Summers M.F., Structural
basis for targeting HIVꢀ1 Gag proteins to the plasma membrane for virus
assembly, Proc. Natl. Acad. Sci. 2006, 103, 30, 11364ꢀ11369.
(7) Massiah, M.; Starich, M.R.; Paschall, C.M.; Summers, M. F.; Christensen, A.M.;
Sundquist, W. I., Threeꢀdimensional structure of the human immunodeficiency
virus type 1 matrix protein, J. Mol. Biol. 1994, 244, 198ꢀ223.
(8) McKercher, G.; Beaulieu, P.L.; Lamarre, D.; LaPlante, S.R.; LeFebvre, S.;
Pellerin, C.; Thauvette, L.; Kukolj, G., Specific inhibitors of HCV polymerase
identified using an NS5B with lower affinity for template/primer substrate,
Nucleic Acids Research, 2004, 32, 422.
(9) LaPlante, S. R.; Gillard, J. R.; Jakalian, A.; Aubry, N.; Coulombe, R.; Brochu, C.;
Tsantrizos, Y.S.; Poirier, M.; Kukolj, G.; Beaulieu, P.L., Importance of ligand
bioactive conformation in the discovery of potent indoleꢀdiamide inhibitors of the
hepatitis C virus NS5B, Journal of the American Chemical Society , 2010, 132,
15204ꢀ15212.
(10) Massiah, M.A.; Worthylake, D.; Christensen, A.M.; Sundquist, W.I.; Hill,
C.P.; Summers, M.F., Comparison of the NMR and Xꢀray structures of the HIVꢀ1
24
ACS Paragon Plus Environment

Page 25 of 25
Journal of Medicinal Chemistry
1
2
3
4
5
6
matrix protein: evidence for conformational changes during viral assembly,
Protein Science 1996, 5, 12, 2391ꢀ2398.
7
8
9
(11) Stammers, T.; Rancourt, J.; Coulombe, R.; Jakalian, A.; LaPlante, S.;
Thavonekham, B.; Fazal, G.; Goulet, S.; Wernic, D.; Duplessis, M.; Garneau, M.;
Kukolj, G.; Beaulieu, P.L., Anthranilic AcidꢀBased Thumb Pocket 2 HCV NS5B
Polymerase Inhibitors with SubꢀMicromolar Potency in the CellꢀBased Replicon
Assay, Bioorg. Med. Chem. Lett., 2013, in press. Also see patents WO
2008/019477 and WO 2007/87717.
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 OF CONTENTS GRAPHIC
25
ACS Paragon Plus Environment