Indole-2-carboxamide-based MmpL3 Inhibitors Show Exceptional Antitubercular Activity in an Animal Model of Tuberculosis Infection

Article abstract of DOI:10.1021/acs.jmedchem.6b00415

Our team had previously identified certain indolecarboxamides that represented a new chemical scaffold that showed promising anti-TB activity at both an in vitro and in vivo level. Based on mutational analysis using bacteria found resistant to one of these indolecarboxamides, we identified the trehalose monomycolate transporter MmpL3 as the likely target of these compounds. In the present work, we now further elaborate on the SAR of these compounds, which has led in turn to the identification of a new analog, 4,6-difluoro-N-((1R,2R,3R,5S)-2,6,6-trimethylbicyclo[3.1.1]heptan-3-yl)-1H-indole-2-carboxamide (26), that shows excellent activity against drug-sensitive (MIC = 0.012 μM; SI ≥ 16000), multidrug-resistant (MDR), and extensively drug-resistant (XDR) Mycobacterium tuberculosis strains, has superior ADMET properties, and shows excellent activity in the TB aerosol lung infection model. Compound 26 is also shown to work in synergy with rifampin. Because of these properties, we believe that indolecarboxamide 26 is a possible candidate for advancement to human clinical trials.

Full text of DOI:10.1021/acs.jmedchem.6b00415

Subscriber access provided by Weizmann Institute of Science
Article
Indole-2-carboxamide-Based MmpL3 Inhibitors Show Exceptional
Antitubercular Activity in an Animal Model of Tuberculosis Infection
Jozef Stec, Oluseye K Onajole, Shichun Lun, Haidan Guo, Benjamin
Merenbloom, Giulio Vistoli, William R. Bishai, and Alan P. Kozikowski
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00415 • Publication Date (Web): 08 Jun 2016
Downloaded from http://pubs.acs.org on June 8, 2016
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 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Indole-2-carboxamide-Based MmpL3 Inhibitors Show Exceptional Antitubercular Activity in an
Animal Model of Tuberculosis Infection
d,e,#
a,f,#
b,#
b
b
Jozef Stec, Oluseye K. Onajole, Shichun Lun, Haidan Guo, Benjamin Merenbloom, Giulio
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
g
b,c*
a*
Vistoli, William R. Bishai, and Alan P. Kozikowski.
a
Drug Discovery Program, Department of Medicinal Chemistry and Pharmacognosy, College of
Pharmacy, University of Illinois at Chicago, 833 S. Wood St., Chicago, IL 60612, USA. Phone: +1ꢀ312ꢀ
996ꢀ7577; Fax: +1ꢀ312ꢀ996ꢀ7107; Eꢀmail: kozikowa@uic.edu
b
Center for Tuberculosis Research, Department of Medicine, Division of Infectious Disease, Johns
Hopkins School of Medicine, 1550 Orleans Street, Baltimore, MD 21231ꢀ1044, USA. Eꢀmail:
wbishai1@jhmi.edu
c
Howard Hughes Medical Institute, 4000 Jones Bridge Road, Chevy Chase, Maryland, 20815ꢀ6789, USA
d
Department of Pharmaceutical Sciences, College of Pharmacy, Chicago State University, 9501 S. King
Drive, Chicago, IL 60628, USA
e
Department of Pharmaceutical Sciences, College of Pharmacy, Marshall B. Ketchum University,
2
575 Yorba Linda Boulevard, Fullerton, CA 92831, USA
f
Department of Biological, Chemical and Physical Sciences, Roosevelt University, 425 S. Wabash
Avenue, Chicago, IL 60605, USA
g
Department of Pharmaceutical Sciences, Università degli Studi di Milano, via Mangiagalli 25, Iꢀ20133,
Milan, Italy
Abstract
Our team had previously identified certain indolecarboxamides that represented a new chemical scaffold
that showed promising antiꢀTB activity both at an in vitro and in vivo level. Based on mutational analysis
using bacteria found resistant to one of these indolecarboxamides we identified the trehalose
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
monomycolate transporter MmpL3 as the likely target of these compounds. In the present work we now
further elaborate on the SAR of these compounds which has led in turn to the identification of a new
analog,
4,6ꢀdifluoroꢀNꢀ((1R,2R,3R,5S)ꢀ2,6,6ꢀtrimethylbicyclo[3.1.1]heptanꢀ3ꢀyl)ꢀ1Hꢀindoleꢀ2ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
carboxamide (26), that shows excellent activity against drugꢀsensitive (MIC = 0.012 µM; SI ≥16000),
multidrugꢀresistant (MDR), and extensively drugꢀresistant (XDR) Mtb strains; has superior ADMET
properties, and that shows excellent activity in the TB aerosol lung infection model. Compound 26 is
also shown to work in synergy with rifampin. Because of these properties, we believe that
indolecarboxamide 26 is a possible candidate for advancement to human clinical trials.
Keywords: tuberculosis, indoleꢀ2ꢀcarboxamide, drug resistance, MmpL3 transporter, homology
modeling, aerosol infection model
Introduction
Tuberculosis (TB) is an airborne infectious disease caused by Mycobacterium tuberculosis (Mtb) that
predominantly affects the lungs. Statistics show an inexorable increase in global health problems and
socioeconomic burden brought about by TB worldwide. The regions consisting of SouthꢀEast Asia, the
Western Pacific, and Africa are the most impacted, and in 2014 accounted for nearly 86% out of the
1
estimated 9.6 million new TB cases. Given the fact that TB predominantly occurs among men (over 60%
of TB cases and deaths due to TB) during their economically most productive years, the economic loss in
2
productivity due to TB is predicted to be 1ꢀ3 trillion dollars over the next 10 years. Major challenges that
prevent the eradication of TB include the inappropriate use or overuse of prescribed medications,
inadequate diagnosis and failure to adhere to the required drug regimen, as well as financial factors that
may preclude access to the required medicines. Moreover, the improper use of TB antibiotics has led to
the emergence of multidrugꢀresistant (MDR) and extensively drugꢀresistant (XDR) TB. An estimated
ACS Paragon Plus Environment

Page 3 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
17,000 (~3.3%) of new TB cases in 2014 were identified as MDRꢀTB. Among those patients with
1
MDRꢀTB, approximately 31,000 (~ 9.7%) were found to be infected with XDRꢀTB.
The current TB drug pipeline includes several chemical entities representing various known and novel
chemical scaffolds that interfere with the function of molecular targets operating within essential
metabolic pathways such as cell envelope biosynthesis, nucleic acid and protein synthesis, and membrane
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
energy production/transport. These chemical entities are currently being evaluated at various stages
4
spanning from preclinical development to Phase III clinical trials. Of special importance is the approval
5
of bedaquiline by the U.S. Food and Drug Administration in December 2012, as well as delamanid and
6
paraꢀaminosalicylic acid by the European Medicines Agency in November 2013. All three drugs were
approved for combination therapy against pulmonary MDRꢀTB infections when other effective treatment
options are not useful due to resistance or tolerability. However, as with many known antiꢀTB drugs, the
use of bedaquiline is accompanied by various adverse effects with irregular heart rhythm being the most
7
pronounced. Thus, the need for safe, highly effective, and fast acting antiꢀTB drugs with novel modes of
action cannot be overemphasized.
The elucidation of essential molecular targets within Mtb and the elaboration of novel chemical scaffolds
possessing desirable physicochemical and ADME/Tox properties are at the forefront of the discovery and
development pipeline to identify better antitubercular therapeutics. Recently, MmpL3, a protein that
belongs to the resistance, nodulation and cell division (RND) family of membrane transporters has been
8
identified as a druggable molecular target essential for Mtb’s survival. MmpL3, encoded by the Rv0206c
gene, is a large protein anchored in the inner lipid bilayer that is comprised of 11 transmembrane helices,
9
two extracellular domains (E1, E2), and one intracellular domain (C1). MmpL3 is responsible for
translocation of mycolic acids in the form of trehalose monomycolate (TMM) from their production site
in the cytoplasm into the cell envelope. Subsequently TMM is processed in the extracytoplasmic space by
mycolyl transferases of the antigen 85 complex, which transfer the lipid portion of TMM to another TMM
10
molecule to afford trehalose dimycolate (TDM), or to arabinogalactan yielding wallꢀbound mycolates.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The released trehalose is recycled by ATPꢀbinding cassette transporters (LpqYꢀSugAꢀSugBꢀSugC), a
11
process known to be essential for the virulence of Mtb. The arabinogalactanꢀbound mycolates are long
αꢀbranched βꢀhydroxy fatty acids that make up the thick, waxy, and difficultꢀtoꢀpermeate outer coating of
12
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the bacterial envelope. Accordingly, inactivation of the MmpL3 transporter prevents the translocation of
the mycolic acids to the outer layer thus weakening the bacterial cell wall and impacting the viability and
virulence of Mtb. Additionally, other recent studies reveal that MmpL3 along with another family
13
member MmpL11 play a role in the transport of hemeꢀiron. The precise extent to which MmpL11 and
9
MmpL3ꢀmediated iron uptake contribute to TB virulence is still under study. Lastly, a more recent paper
has provided evidence that a number of the MmpL3 inhibitors may actually work by causing the
dissipation of the proton motive force that is required for the function of MmpL3 and other MmpLꢀdriven
14
processes.
In our previous work, based on phenotypic screening, we identified the indoleꢀ2ꢀcarboxamide 1 as a
1
5
highly potent scaffold against drugꢀsusceptible and drugꢀresistant Mtb strains (Figure 1). Employing a
standard medicinal chemistry approach, a number of analogs were made which lead to compound 3 that
15
was found to show superior activity against susceptible, MDR and XDR Mtb strains in comparison to 1.
Using genome sequencing methodology we also determined that compound 3 targets the MmpL3
1
6
transporter protein which has consequences for the viability of Mtb as detailed above. Herein we report
further SAR studies on these indolecarboxamides, together with in vitro pharmacokinetic (PK) analysis as
well as in vivo TB mouse efficacy studies for the best of these new indoles. These data point to compound
26 as a possible candidate for further advancement through the development pipeline.
ACS Paragon Plus Environment

Page 5 of 53
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. Hit compound 1, lead compounds 2 and 3, and strategy for additional chemical modifications of
the indoleꢀ2ꢀcarboxamide scaffold 3.
Results
Following the amide coupling protocol (Scheme 1), 40 novel indolecarboxamides and related compounds
were efficiently generated. Briefly, the substituted indoleꢀ2ꢀcarboxylic acids were reacted under an argon
atmosphere with an alkylꢀ or cycloalkylamine in the presence of 1ꢀethylꢀ3ꢀ(3ꢀ
.
dimethylaminopropyl)carbodiimide hydrochloride (EDC HCl) and hydroxybenzotriazole (HOBt) as
coupling agents and triethylamine as a base to obtain compounds 3−39 (Scheme 1). To prepare
compounds 6 and 9, the starting materials 52 and 53 were reacted with cyclooctylamine under standard
amide coupling conditions followed by catalytic hydrogenation of the nitro groups (Scheme 1). To
prepare intermediate 55, 3,5ꢀdimethoxyaniline (54) was reacted with ethyl bromoacetate to afford ethyl 2ꢀ
[(3,5ꢀdimethoxyphenyl)amino]acetate which was reacted in turn with N,Nꢀdimethylformamide dimethyl
acetal (DMFDMA ) under microwave irradiation. Next, Lewis acid (AlCl )ꢀpromoted cyclization of 55
3
afforded the cyclized indoleꢀ2ꢀcarboxylate, which after basic hydrolysis gave the corresponding acid 56.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The indole carboxylic acid 56 was reacted with cycloheptylamine and cyclooctylamine to afford
compounds 18 and 19, respectively.
1
5
4
,6ꢀDimethylꢀ1Hꢀindoleꢀ2ꢀcarboxylic acid (43) (Scheme 2) was reacted with ammonium chloride to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
afford compound 57, which was reduced to its corresponding primary amine and reacted with
1
5
cycloheptanecarbonyl chloride to afford compound 40. The amide group of compound 3 was reduced to
the corresponding amine 41. The benzothiazole analog 42 was obtained from the benzo[d]thiazolꢀ2ꢀamine
58 by reaction with cycloheptanecarbonyl chloride (Scheme 2).
a
Scheme 1. Synthesis of indoleꢀ2ꢀcarboxamides
a
.
Reagents and conditions: (a) EDC HCl, HOBt, corresponding amines, Et N, CH Cl , rt, 12–16 h; (b)
3
2
2
ethyl bromoacetate, K CO , acetone, 60 °C, overnight; (c) DMFDMA, microwave, 150 °C, 10 min; (d)
2
3
AlCl , CH Cl , 24 h, rt; (e) LiOH, EtOH, reflux, 3 h; (f) 10% Pd/C, MeOH/EtOAc (4:1), H .
3
2
2
2
a
Scheme 2. Synthesis of analogs containing indole and benzothiazole skeleton
ACS Paragon Plus Environment

Page 7 of 53
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
a
.
Reagents and conditions: (a) EDC HCl, HOBt, NH Cl, Et N, CH Cl , rt, 12–16 h; (b) LiAlH , THF,
4
3
2
2
4
.
reflux, overnight; (c) EDC HCl, HOBt, cycloheptanecarboxylic acid, Et N, CH Cl , rt, 12–16 h; (d)
3
2
2
cycloheptanecarbonyl chloride, CH Cl , rt, overnight.
2
2
Discussion and Conclusions
Structure-Activity Relationships
In general, the biological activity of the analogs was found to depend upon the polarity of the substituents
introduced, and thus the more lipophilic compounds were found to be the most active. Compounds
possessing monoꢀsubstitution such as a methyl group at position 4 of the indole moiety (4, 5) retained
antitubercular activity, with compound 5 bearing the Nꢀcyclooctyl group being over 8ꢀfold more active
than its closest analog 4 bearing the Nꢀcycloheptyl group. On the other hand, the more polar 4ꢀ
aminoindole derivative 6 showed inferior activity. A similar trend in activity was observed for
compounds 7ꢀ9 possessing methyl or amino substitution at position 5 of the indole ring. However, these
compounds were consistently less potent than their corresponding counterparts 4ꢀ6 which further
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
confirms the lipophilicityꢀdriven activity of these analogs but also indicates that position 4 is preferred for
functionalization. In the case of position 6, mixed results were obtained as compound 11 bearing the more
lipophilic bromo group was only slightly more active than its close analog 10 possessing the 6ꢀmethyl
group. On the other hand, compound 12 bearing an Nꢀadamantyl group was 2.6ꢀ and 2ꢀfold more active
than compounds 10 and 11, respectively, although has a ClogP comparable to that of 10, which is lower
than that of analog 11. In line with the activity of the adamantyl derivative, we were also interested in
evaluating bicyclic substituents. Our attention was drawn to the pinene skeleton as this structure is known
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
17
for its broad antimicrobial activity. It was pleasing to find that compound 13 bearing the Nꢀ
(
1R,2R,3R,5S)ꢀ(ꢀ)ꢀisopinocampheyl group is 4ꢀfold more active than compound 12 and 9ꢀfold more active
than analog 11. Again, compound 13 is the most lipophilic molecule among compounds 10ꢀ13 supporting
the requirement of high lipophilicity for bioactivity. To complete the investigation of the effect of monoꢀ
substitution of the indole ring on activity, a methyl group was installed at position 7 in compounds 14 and
15. As observed earlier for compound pairs 4/5 and 7/8, compound 15 bearing the Nꢀcyclooctyl group
was more active than its counterpart 14 possessing the Nꢀcycloheptyl substituent.
Next, we examined the effect of diꢀsubstitution with methyl and methoxy groups on the indole ring. As
expected, the Nꢀcyclooctyl indoles 17 and 19 were more active than their Nꢀcycloheptyl counterparts (16
and 18). Surprisingly, the 4,6ꢀdimethoxy compounds (18 and 19) had almost the same activity as the 5,7ꢀ
dimethyl analogs (16 and 17), although the latter pair has higher ClogP values than compounds 18 and 19,
respectively. Knowing that biological activity of this series of compounds is lipophilicityꢀdriven and that
the methyl groups on the indole core present a potential metabolic liability (susceptible to metabolic
1
6
oxidation), we replaced the 4,6ꢀdimethyl substituents with more lipophilic and at the same time
18
metabolismꢀresistant 4,6ꢀdihaloꢀgroups. Compound 20 containing a 4,6ꢀdichloroꢀindole scaffold was the
first to be prepared, and this analog showed similar activity as the lead compound 3. Derivative 21
bearing the Nꢀadamantyl group was equally potent as compound 20, although its ClogP value is lower
than that of 20. Another structurally close analog 22 bearing the Nꢀexoꢀ2ꢀnorbornyl substituent showed a
ACS Paragon Plus Environment

Page 9 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
greater than 30ꢀfold drop in activity when compared to 20. This result is most likely associated with the
substantial difference in lipophilicity observed between the two molecules. Introduction of the Nꢀ(R)ꢀ(+)ꢀ
bornyl group provided compound 23 which was 9ꢀfold more potent than 22 but approximately 4ꢀfold less
active than 20. Interesting results were obtained with compounds 24ꢀ27 which bear an isopinocampheyl
moiety on the amide nitrogen atom. The 4,6ꢀdichloroindole nucleus with the Nꢀ(1R,2R,3R,5S)ꢀ(ꢀ)ꢀ
isopinocampheyl group 24 proved to be 2ꢀfold more active than compound 23 but also was approximately
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
2
ꢀfold less active than compound 20 having the Nꢀcyclooctyl group. Of note is the fact that compound 24
is more lipophilic than 20 and has the same ClogP value as molecule 23. On the other hand, the Nꢀ
1S,2S,3S,5R)ꢀ(+)ꢀisomer 25 turned out to be 2ꢀ, 4ꢀ, and 7ꢀfold less potent compared to indoles 23, 24,
(
and 20, respectively. Given the fact that all three compounds 23ꢀ25 have the same ClogP values, it is
apparent that the stereochemical orientation of the amide nitrogen substituent plays some role in the
biological activity of this series of compounds. Even more pronounced results were obtained for the pair
of enantiomers 26 and 27 possessing the 4,6ꢀdifluoro indole scaffold. In this case, the dextrorotatory
isomer 27 was nearly 16ꢀfold less potent than the levorotatory isomer 26 while the latter showed about the
same potency as 20 and about a 2ꢀfold improvement in activity compared to 24. This is a noteworthy
result as the lipophilicity of compound 26 is lower that of analogs 20 and 24, yet it shows an in vitro
antitubercular potency comparable to that of 20. Based on their satisfactory in vitro activity, all three
compounds, i.e. 20, 24, and 26 were selected for further in vivo studies in the serum inhibitory titer assay
(
SIT).
Appendage of a simple nꢀheptyl or nꢀoctyl aliphatic chain on the amide nitrogen yielded the highly
lipophilic compounds 28 and 29, but surprisingly these were nearly inactive. These findings clearly
support the importance of a more compact, αꢀbranched cycloaliphatic group on the amide nitrogen in
order to observe high activity. Compounds 30 and 31 containing the Nꢀgeranyl moiety provided mixed
results with compound 30 possessing inferior activity and compound 31 being inactive. Compounds 30
and 31 depict once again the apparent impact of lipophilicity on activity. The βꢀbranched Nꢀsubstituent in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compound 32 also led to loss of activity while a similar alphaꢀbranched Nꢀsubstituent gave the reasonably
active compound 33 and the modestly active compound 34. This result clearly supports the importance of
a lipophilic branched cycloaliphatic group for the bioactivity as it is likely to be more resistant to
metabolism than the nonꢀbranched substituents. Compound 35 with a 5,6ꢀfused amide substituent retained
activity. On the contrary, compound 36 possessing the same Nꢀsubstituent as 35 turned out to be inactive
most likely due to a drastic decrease in lipophilicity. The spiro compounds 37 and 38 provided the highest
in vitro antitubercular activity identified in this set of compounds. Given their high potency, these
compounds were also selected for in vivo SIT assay. The azabicylic compound 39 bearing a tertiary amine
showed a significant decrease in activity. Furthermore, compound 40 featuring a reversed amide linker,
and the benzothiazole derivative 42 were inactive against a susceptible strain of Mtb. Reduction of the
amide group of 3 to obtain the amine 41 resulted in a 24ꢀfold drop in TB activity thus highlighting the
importance of the carboxamide linker for antiꢀTB activity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Overall, the SAR for this set of compounds is complex with several factors being decisive for high
antitubercular activity. The indole nucleus possessing 4,6ꢀdisubstitution with lipophilic groups is crucial
for activity. Halogen substituents such as F or Cl are preferred over methyl groups, which may constitute
a metabolic liability. Carbocyclic groups with an alphaꢀbranched carbon on the Nꢀindolecarboxamide
atom yielded the best compounds. The dependence of activity on the absolute configuration of this
substituent suggests that it may be involved in specific interactions with the target, rather than just
enhancing the compound’s lipophilicity.
Table 1. MIC and ClogP values of indoleꢀ2ꢀcarboxamides analogs.
b
b
Analog
X
R
a
ClogP
Analog
X
R
a
ClogP
MIC
MIC
ACS Paragon Plus Environment

Page 11 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
(
µM)
(µM)
4
,6ꢀ
7
8
3
0.013
0.93
0.11
20
5.59
4.54
5.10
3.44
4.54
5.10
3.44
5.10
5.56
5.07
6.28
24
25
26
27
28
29
30
31
32
33
34
4,6ꢀdichloro
4,6ꢀdichloro
4,6ꢀdifluoro
4,6ꢀdifluoro
4,6ꢀdichloro
4,6ꢀdichloro
4,6ꢀdichloro
4,6ꢀdifluoro
4,6ꢀdichloro
4,6ꢀdichloro
4,6ꢀdifluoro
0.021
6.88
6.88
5.74
5.74
6.19
6.72
6.75
5.61
6.98
5.97
4.83
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
dimethyl
4ꢀmethyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
5
0.086
0.012
0.19
>300
>300
11
4ꢀmethyl
4ꢀamino
5ꢀmethyl
5ꢀmethyl
5ꢀamino
6ꢀmethyl
6ꢀBr
6
7
7.4
8
0.88
80
9
10
0.11
0.09
0.042
0.010
>192
175
11
12
13
6ꢀBr
5.9
6ꢀBr
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
14
7ꢀmethyl
7ꢀmethyl
30
3.5
4.54
5.10
5.04
5.59
4.01
4.57
6.16
5.67
5.32
35
36
4,6ꢀdichloro
4,6ꢀdifluoro
4,6ꢀdichloro
4,6ꢀdifluoro
4,6ꢀdichloro
0.72
102
5.00
3.86
7.04
5.90
3.94
4.53
5.65
4.47
15
16
17
18
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
,7ꢀ
dimethyl
,7ꢀ
3.5
37
0.005
0.003
10.9
54
5
0.21
3.2
38
dimethyl
4,6ꢀ
39
dimethoxy
4
,6ꢀ
dimethoxy
,6ꢀ
4,6ꢀ
19
0.19
0.011
0.011
0.39
40
dimethyl
4,6ꢀ
4
2
0
1
41
0.31
>300
0.29
dichloro
4,6ꢀ
dimethyl
5ꢀtrifluoroꢀ
methoxy
2
42
dichloro
4
,6ꢀ
dichloro
,6ꢀ
c
22
INH
4
2
3
0.043
6.88
dichloro
a
The lowest concentration of drug leading to at least a 90% reduction of bacterial growth signal by the
microplate alamarBlue assay (MABA). MIC values are reported as an average of three individual
b
c
measurements; Calculated using ChemBioDraw Ultra 13.0; INH = Isoniazid.
ACS Paragon Plus Environment

Page 13 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In vivo Bioavailability and in vitro Activity of Selected Indole-2-carboxamides against Clinical
Isolates of Mtb.
The most promising compounds in this series, namely 20, 21, 24, 26, 37, 38, and the comparator
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
19
compound 2 reported by Novartis (prepared inꢀhouse) were further evaluated for their in vivo
2
0
bioavailability in the serum inhibition titration (SIT) assay (Figure 2). As depicted in Figure 2,
compound 24 showed much higher bioavailability and thus in vivo efficacy when compared to 2 and 20,
but it was less effective than the reference drug isoniazid. On the other hand, compound 26 showed
excellent in vivo efficacy that was comparable to that of isoniazid when tested at a dose of 100 mg/kg.
Compound 26 also showed significant improvement in bioavailability when compared to the lead
15
compound 3 (results not shown). The other compounds selected for SIT analysis, namely 21, 37, and 38
turned out to be inactive (results not shown) in this assay and thus there were dropped from further
studies.
Based on their activity in the SIT screening assays, compounds 20, 24, and 26 were next tested against a
21
panel of clinical isolates of Mtb originally obtained from pulmonary TB patients (Table 2). Among
those strains, one was drugꢀsensitive (DS) (V4207), two were MDR Mtb (KZN494, V2475), and two
were XDR Mtb (TF274, R506). Compound toxicity against mammalian cells was tested on Vero cells and
a selectivity index (SI) subsequently calculated (Table 2). All three compounds retained their high activity
or showed only a moderate (up to 4ꢀ and 8ꢀfold for 26 and 24, respectively) decrease in activity against
the tested clinical isolates of Mtb. Of note is the fact that compounds 20 and 26 showed a 2ꢀfold increase
in activity against the XDR Mtb strains R506 and TF274, respectively. The high activity of these
compounds against MDR and XDR TB indicates not only their potential to treat drugꢀresistant Mtb
strains but also suggests the likelihood that these compounds will not exhibit crossꢀresistance with
currently used medications. Additionally, all the three compounds exhibited high IC values against Vero
5
0
cells, thus resulting in high SI indices, and thus indicating their potential lack of toxicity toward
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mammalian cells. We also tested 26 against the TSK54R mutant which we had generated in our initial
studies and used in the DNA deep sequencing efforts to identify MmpL3 as the target of the
indolecarboxamides. The MIC was 0.0625ꢀ0.125 ꢀg/mL, which corresponds to a 16ꢀ32ꢀfold loss of
activity against the mutant indicating MmpL3 as the potential target. Based on its excellent in vitro
activity, superior SI, and high efficacy in the SIT assay, compound 26 was selected for further
development.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Activity [MIC (µM)] of indoleꢀ2ꢀcarboxamides against selected clinical isolates of Mtb and
Vero cells [IC (µM)].
50
d
Strain
Phenotype DS
20
0.011
H37Rv
V4207 KZN494 V2475
TF274 R506
IC50
SI
a
a
b
b
DS
MDR
MDR
0.011
0.171
XDR
XDR
0.006
0.021
Vero Cells
23.5ꢀ47
43.8
c
0.011
0.043
NA
0.011
0.021
2136ꢀ4272
2085
2
4
0.021
0.012
0.043
26
0.023
0.012
0.047
0.006
0.012
≥192
≥16000
a
b
Resistance to isoniazid and rifampin; Resistance to isoniazid, rifampin, levofloxacin, ofloxacin, and
c
d
kanamycin; Data not available due to growth defect; Selectivity Index (SI) = IC
/ MIC
0 (Vero)
5
Preliminary ADMET Studies on Compound 26.
Based on the encouraging biological results described above, we conducted some ADMET studies on
compound 26 to assess its drugꢀlike properties (Table 3). When incubated with human, mouse, rat or dog
liver hepatocytes, at least 50% of compound 26 remains unchanged after 2 hours incubation using a
starting concentration of 1 ꢀM and the human hepatocytes. More substantial degradation was found using
hepatocytes from the other species. Using liver microsomes, almost 70% of the compound remained
unchanged after 1 hour using a starting concentration of 3 ꢀM.
ACS Paragon Plus Environment

Page 15 of 53
Journal of Medicinal Chemistry
Table 3. ADMET data for indole-2-carboxamide 26.
Assay 26
Hepatocytes, % remaining at 2 hrs, 52 (human), 0.7 (rat), 15 (mouse), 18 (dog)
starting concentration 1
Liver microsomes, % remaining after 69.5 (human, 60 mins)
hr, starting concentration 3
1
2
3
4
5
6
7
8
9
ꢀ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
1
ꢀM
CYP inhibition (% inhibition)
CYP2C9
ꢀ5.2
36.6
ꢀ6.7
CYP2D6
CY3A4
Cacoꢀ2 permeability
AꢀB (cm/s)
ꢀ
6
6
0.19 x 10
0.25 x 10
>30
ꢀ
BꢀA (cm/s)
hERG IC (ꢀM)
50
As shown in Table 3, in the presence of compound 26 none of the tested CYP isoforms (CYP2D6,
CYP2C19, and CYP3A4) showed more than 40% inhibition. Low inhibition of hERG (IC > 30 ꢀM)
5
0
suggests low potential risk for blocking the cardiac calcium channel and causing QTꢀprolongation. As the
xenobioticꢀsensing transcription factors that include the pregnane X receptor (PXR), aryl hydrocarbon
receptor (AhR), and the constitutive androstane receptor (CAR) upregulate the expression of drugꢀ
metabolizing genes (CYP3A4, CYP1A2 and CYP2B6, respectively), the potential of 26 to cause
transactivation of the human PXR, CAR, and AhR receptors was assessed. At low concentrations of the
compound tested (0.1ꢀ1 ꢁg/mL), no significant induction of the respective reporter genes coupled to these
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
nuclear receptors was observed. Only at higher concentrations (3 and 8 ꢁg/mL) was 26 able to upregulate
the induction of metabolizing genes by these receptors. However, the levels of induction produced by 26
were still lower than those of the positive controls, i.e. rifampin for PXR, CITCO for CAR, and 3ꢀ
methylcholanthrene for AhR. We also carried out the bacterial reverse mutation assay in Salmonella
typhimurium tester strains TA98, TA100, TA1535, and TA1537 and Escherichia coli tester strain WP2
uvrA for compound 26 to evaluate its intrinsic mutagenic activity, and for the most part these results were
negative. Only in the case of the TA98 tester strain with the S9 liver fraction present was an increase in
the ratio of the revertant colony count for the drug compared to the DMSO solvent control observed. This
ratio was 2.5 with 500 ꢀg/plate of drug. Even at 5000 ꢀg/plate of the 26 this ratio was only 2.0 whereas
this ratio was 173 for the positive control 2ꢀaminoanthracene. Thus we believe the genotoxicity risk of
this compound to be negligible although further studies to detect chromosomal aberrations should be
done.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Pharmacokinetic comparisons of 26 and 2 using LC-MS. Next, we carried out comparative
pharmacokinetic studies on 26 and the comparator compound 2 using traditional liquid chromatographyꢀ
mass spectrometry (LCꢀMS) analysis (Table 4). A single dose of 100 mg/kg was administered orally to
female Balb/c mice. As shown in the table below, 26 showed significantly better PK parameters than 2 as
reflected by higher AUClast (area under the curve up to the last measurable concentration) and higher
AUMClast (area under the first moment curve to the last measurable concentration; this is a measure of
AUC which emphasizes mean residence time).
Table 4. Pharmacokinetic comparisons of compounds 2 and 26.
Compound
TMAX
(h)
CMAX
AUClast
AUMClast
MRTlast
(h)
2
(ꢁg/unit)
(hꢀꢁg/unit)*
(h ꢀꢁg/unit)*
ACS Paragon Plus Environment

Page 17 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
6
Plasma
Lung
4
4
2
2
1.71
3.50
0.67
1.30
18.9
51.0
5.62
18.1
137
433
36.4
150
7.28
8.49
6.47
8.30
2
Plasma
Lung
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
*
unit = mL for plasma samples; unit = gram for lung samples; AUMC = area under the first
moment curve; MRT = mean residence time.
Mouse efficacy comparisons of 26 and 2. Lastly, we completed in vivo efficacy studies for 26 and 2
using two different mouse infection and monotherapy models. In order to enable headꢀtoꢀhead
comparisons, we deliberately used 0.5% carboxymethylcellulose (CMC) for both of the test compounds
although we are aware that more advanced formulations are now available. A minimum effective dose
(
MED) model was conducted using Balb/c mice that had been aerosolꢀinfected and treated from day 1 at
2.5, 25, 50 and 100 mg/kg for both compounds. While both compounds were protective, 26 showed
1
more obvious doseꢀdependent effects than 2 on both CFU enumerations in the lungs (Figure 3) and gross
pathology such as lung weight and spleen weight. For comparison, at 100 mg/kg dosage level, there was a
CFU reduction of 1.71ꢀ and 2.12ꢀlog in the lungs after treatment with 26 for 2 weeks and 4 weeks,
respectively. For 2, the CFU reduction was 0.45ꢀ and 1.08ꢀlog for the same dosage level and treatment
durations, respectively (p<0.0001). A minimum bactericidal dose (MBD) model was also carried out.
After 4 weeks of treatment at 200 mg/kg, 26 achieved 0.6ꢀlog CFU reduction while 2 left the CFU count
unchanged compared with that of initiation (p=0.0002). All treatment levels (25, 50, 100 and 200 mg/kg)
of both indolecarboxamides protected mice from death. In these studies, we used the universal
formulation, which is 0.5% carboxymethylcellulose in water. The universal formulation used in this study
differs from the optimized customized proprietary formulation used by Novartis, which in turn might
have contributed to the observed difference in the efficacy between compound 26 and 2 in mice model.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. In vivo efficacy comparisons of compounds 26 and 2 in a mouse infection
3
model. After aerosol Mtb infection with 10 CFU, mice were treated orally with 12.5, 25,
50, and 100 mg/kg of 26 or 2 daily. EMB was at 100 mg/kg/d. Lung CFU counts were
determined on days 1, 14, and 28 (5 mice/grp).
In vitro and in vivo Synergy between Indole-2-carboxamide 26 and Rifampin.
Previous studies suggested that mutation in the mmpL3 gene confers increased susceptibility of Mtb to
22
rifampin . Since indolecarboxamides target a molecular channel, we hypothesized that the antibacterial
effect of dual exposure to indolecarboxamide and rifampin might meet the criteria for true synergy rather
than being two independent, additive effects. To address this question, we applied the Bliss independence
model which evaluates multiplicative effects in growth measures and has been used extensively in
2
3
evaluating pharmacologic drugꢀdrug interactions. As shown in Table 5, Bliss synergy was observed in
multiple subꢀMIC concentration combinations as highlighted. For instance, 0.5×MIC of 26 combined
with 0.0313×MIC of rifampin achieved 99.2% inhibition (MIC cutoff = 90%). Similarly, 0.25× or
0
.0313×MIC of 26 plus 0.25× or 0.5×MIC of rifampin also achieved MIC level of inhibition (99.4% or
9
0.5% respectively, Table 5). Rifampin is a key component of multidrug TB therapy and is considered
ACS Paragon Plus Environment

Page 19 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
particularly important for clearance of M. tuberculosis persisters and to achieving a relapseꢀfree cure in
sixꢀmonths. However, rifampin is a potent inducer of both cytochrome Pꢀ450 oxidative enzymes and the
24
Pꢀglycoprotein transport system. Because of these interactions, dose adjustments are necessary when
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
rifampin is used with numerous drugs including warfarin, contraceptives, cyclosporine, quinidine sulfate,
2
4b, 25
and verapamil, to name a few.
Importantly, rifampin also interacts with HIVꢀrelated protease
inhibitors and significantly reduces the concentration of many standardꢀdose protease inhibitors,
26
compromising HIV treatment efficacy. Our findings of synergy between 26 and rifampin are important
because it raises the possibility that indolecarboxamideꢀcontaining treatment regimens with lower
rifampin doses could be efficacious. Reduced rifampin exposures could be especially important for
patients with HIV/TB coꢀinfection who require antiꢀretroviral therapy.
Table 5. In vitro drug combination of indole-2-carboxamide 26 and rifampin.
Compound 26
0.0313 0.0625
0.125
0.25
0.5
1.0
2.0
4.0
4.0
2.0
1.0
0.0
0
.5
0.1
0.1
0.1
0.0
0.0
0.1
0.4
0.1
0.0
0.0
0.1
0.6
0.8
0.1
0.0
0.1
0.6
1.0
1.0
1.0
Rif
0.25
0.1
0.1
0.0
0.0
0
.125
.0625
.0313
0
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Note: Checkerboard synergy study showing the calculated effect of the combination of 26 combined and
rifampin. Bliss independence modeling was applied to wells containing < 1×MIC concentration of either
compound. Synergy (ꢂfaxy >0) was observed in multiple combinations and highlighted.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Most interestingly, in an acute mouse infection and chemotherapy model, treatment with a combination of
26 and rifampin achieved synergistic effects, as evaluated using the Bliss synergy model (Figure 4). This
implies a great potential for indolecarboxamide 26 and analogs to be used in the standard regimen for
shortening the duration of TB treatment.
Figure 4. Bliss synergy model using rifampin and 26. Compound
2
6 synergizes with rifampin in an acute mouse infection and chemotherapy
model.
ACS Paragon Plus Environment

Page 21 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Molecular Docking of Compound 26 and Related Indole-2-carboxamides
In order to gain some understanding of how the indolecarboxamides interact with the MmpL3 transporter,
we constructed a possible model of the transporter using methods provided in the supplemental section.
As shown in Figure 5, the overall folding of the MmpL3 transporter is characterized by a complex TM
bundle, including 12 TM helices, and two large periplasmic domains: the first domain (D1) is located
between TM1 and TM2 and the second (D2) between TM7 and TM8. Compared to the selected template
(MexB transporter, see Table S1 in SI), MmpL3 conserves the two periplasmic domains which define the
soꢀcalled pore domain and are characterized by β−α−β motifs, while it lacks the soꢀcalled docking
domains (DN and DC) which are usually involved in contacting modulating proteins. The overall folding
can be divided into two halves, each composed of six TM helices and one periplasmic domain. The two
halves are connected by an intracellular helical domain (D3). Figure 5 also shows the bound compound 3
thus revealing the location of the putative binding pocket which is on the cytoplasmic side of the TM
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
27
bundle mostly lined by TM4 and TM10.
Moreover, since the considered template (MexB transporter) includes an asymmetric homotrimer where
each subunit adopts a different conformation which may represent snapshots of the transport process, it is
worth noting that the monomer utilized as the template is the soꢀcalled binding monomer (monomer B)
since the nꢀdodecylꢀDꢀmaltoside (DDM) ligand is bound in the cavity defined by its periplasmic domains
and the channel is open to the periplasm from which it can capture the substrates. Although the
considered pocket for the inhibitors does not correspond to that lined by the periplasmic domains, one
may conjecture that the generated homology model (for details see Table S1 in SI) would be in a
conformation suitable to investigate the binding of the transported substrates. Notably, the last discarded
MmpL3 model (model #5, Table S1 in SI) was generated by using the soꢀcalled extrusion monomer
(
monomer C) of its template, which shows a wider TM bundle to allow the substrate release. This may
explain why the pocket of model #5 is clearly wider and suggests that this cavity, albeit less functional,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
can be seen as an alternative binding conformation that the transporter can assume during the transport
cycle.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Cartoon structure of the MmpL3 homology model as colored by N (blue) to C (red) sequence.
The bound inhibitor 3 is also depicted as the greyꢀcolored CPK model.
Figure 6A shows the main interactions stabilizing the putative complex between the MmpL3 model and
3
, which can be schematized as follows. The first subpocket accommodates the indole ring which elicits
an extended charge transfer interaction with Arg259 reinforced by π−π stacking with Phe644 plus a set of
hydrophobic contacts with surrounding apolar residues such as Ile256, Val278, Ala281, Val285, and
Pro330. Moreover the indole NH group stabilizes a Hꢀbond with Ser325. A central subpocket stabilizes a
rich set of Hꢀbonds with the ligand’s amide moiety: the carbonyl oxygen atom contacts Tyr252, Ser288,
and Ser325, while the amide NH group interacts with Asp640 which in turn establishes an ionꢀpair with
ACS Paragon Plus Environment

Page 23 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the aboveꢀcited Arg259. Notably, Tyr252 is highly conserved among the MmpL isoforms and is adjacent
to another residue (Gly253), the mutation of which was found to decrease the inhibition activity of known
MmpL3ꢀtargeting compounds such as SQ109 and AU1235. Again, Arg259 and Asp640 are supposed to
1
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
be involved in the electrochemical proton gradient that provides energy to the transporter. The last
hydrophobic subpocket harbors the cyclooctyl ring which approaches a set of alkyl side chains including
Ile248, Ile291, Ile292, and Val643.
Even though the considered MIC values are clearly influenced by lipophilicity as evidenced by Table 1
and as discussed above, docking results are in line with the reported activity values and allow inꢀdepth
rationalization of the obtained SARs. Thus, the above mentioned apolar residues surrounding the indole
ring can easily explain the negative influence of polar substituents as seen in the amino derivatives (e.g.,
6
) and indeed the computed complexes confirm that the amino group is never involved in clear polar
contacts but at most bumps against Phe644. Again, the disfavored status of position 5 (see, for example,
compound 8) on the indole ring can be explained by considering that the substituent approaches the less
flexible backbone atoms of Thr277 and Val288. Halogen substituents clearly make a positive contribution
to binding due to their lipophilicity although the computed complexes do not show clear halogen bonds
involving these substituents (see Figure 6B). The relevance of the Hꢀbond involving the indole NH group
is witnessed by the inactivity of the benzothiazole derivative (42).
The rich set of Hꢀbonds stabilized by the amide moiety easily explains why even subtle modifications in
that moiety have marked detrimental effects. Notably, the amide moiety determines the overall pose of
the bound inhibitor, and indeed the computed complex for 40 shows a flipped pose in which the
cycloheptyl ring takes the place of the indole moiety to allow the amide interactions to be properly
elicited. Finally, the role of the Nꢀlinked moiety appears to be completely governed by its lipophilicity, a
finding easily explainable when considering the aboveꢀmentioned set of surrounding apolar residues.
Figure 6C shows the putative complex for 26 and reveals that the isopinocampheyl moiety is conveniently
arranged within the apolar subpocket with all the methyl substituents suitably approaching alkyl side
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chains, which also provides the basis for the observed stereoselectivity of these derivatives. The inactivity
of the noncyclic derivatives (28-31) may have three explanations since (a) linear substituents are less able
to contact the surrounding apolar residues, (b) their flexibility may play a disfavoring entropic role, and
(c) they can bump against Phe635 which restricts this subpocket. By contrast, even extended cyclic
moieties (as seen in the spiro compounds 37 and 38) are able to maximize their hydrophobic contacts
while remaining reasonably far from Phe365.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The performed MD simulations are unsuitable for exploring large protein conformational shifts induced
by the bound inhibitors but can be conveniently utilized to analyze the relevance (and stability) of each
monitored interaction. The dynamic profile of these interactions as assessed by their distance average and
standard deviation (see Table S2 in SI) allows for some interesting considerations. All Hꢀbonds show
similar profiles which are suggestive of an overall marked stability of such interactions. Indeed, the
distance profile of these Hꢀbonds shows a roughly constant trend, apart from the contact involving
Asp640 which reveals two distinct and alternating distance clusters around 2.7 Å and 4.5 Å, respectively.
Notably, the best correlations (see Figure S1 in SI) are obtained by the frames where this contact assumes
the larger distance values probably as such an overall arrangement allows a proper accommodation even
of the bulkiest inhibitors. Finally, the standard deviation values, which can be seen as descriptors of the
interaction stability, reveals that the charge transfer interaction and the hydrophobic contacts possess the
highest stability, thus confirming the pivotal role of such interactions in stabilizing the simulated
complexes.
ACS Paragon Plus Environment

Page 25 of 53
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 6. Main interactions stabilizing the putative MmpL3 complex with compound 3 (3A), 21 (3B),
and 26 (3C).
The present work serves to underscore the importance of conducting a more comprehensive SAR program
in order to identify the best candidate drug for advancement to clinical studies. We have assembled a
total of over 80 indolecarboxamides to date, and we believe that the compound 26 identified in the present
work represents a potential clinical candidate. As in our earlier work, it is likely that this compound acts
as an inhibitor of the mycolic acid transporter MmpL3, although this activity may indirectly stem from its
initial action on the proton motive force which leads in turn to inhibition of MmpL3. A putative
homology model for the MmpL3 transporter has also been generated, and while this model can only be
deemed approximate at this stage, it is useful in providing further possible insights into the SAR of the
indolecarboxamides. In any event, compound 26 shows superior pharmaceutical properties, for by and
large the compound displays acceptable ADMET parameters as well as in vivo activity. Additionally, the
findings of synergy between 26 and rifampin are notable as this raises the possibility that
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
indolecarboxamideꢀcontaining treatment regimens with lower rifampin doses could be efficacious. Thus,
in comparison to the Novartis compound 2 (NITDꢀ304) that is under development by the TB Global
Alliance, we believe that 26 may in fact represent a better candidate, as it is more effective in reducing the
CFU count using the TB aerosol lung infection model.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
1
. Chemistry
General information. The following carboxylic acids: 4ꢀmethylꢀ1Hꢀindoleꢀ2ꢀcarboxylic acid, 5ꢀ
methylꢀ1Hꢀindoleꢀ2ꢀcarboxylic acid, 6ꢀmethylꢀ1Hꢀindoleꢀ2ꢀcarboxylic acid, 7ꢀmethylꢀ1Hꢀ
indoleꢀ2ꢀcarboxylic acid, 4,6ꢀdichloroꢀ1Hꢀindoleꢀ2ꢀcarboxylic acid, and 6ꢀbromoꢀ1Hꢀindoleꢀ2ꢀ
carboxylic acid, were purchased from SigmaꢀAldrich, ChemꢀImpex, Combiꢀblocks, and Ark
Pharma. Anhydrous dichloromethane (CH Cl ) was obtained by distillation over calcium
2
2
1
13
hydride. H NMR and C NMR spectra were recorded on Bruker or JEOL spectrometers at 400
and 100 MHz, respectively. NMR spectra were reprocessed by ACDNMR Processor Academic
Edition. Standard abbreviations indicating multiplicity were used as follows: s = singlet, d =
doublet, dd = doublet of doublets, t = triplet, q = quadruplet, m = multiplet, and br = broad. HRꢀ
MS experiments were performed on QꢀTOFꢀ2TM (Micromass) and ITꢀTOF (Shimadzu)
instruments. TLC was performed on Merck 60 F254 silica gel plates. Flash chromatography was
®
®
performed using a CombiFlash Rf system with RediSep columns or manually using Merck
silica gel (40–60 mesh). Final compounds were purified by preparative HPLC unless otherwise
stated. Preparative HPLC employed an ACE 5ꢀAQ (21.2 mm × 150 mm) column, with detection
at 254 and 280 nm on a Shimadzu SCLꢀ10A VP detector, flow rate = 17.0 mL/min. Method 1:
50–100% CH OH/H O in 30 min; 100% CH OH for 5 min; 100–50% CH OH/H O in 4 min.
3
2
3
3
2
ACS Paragon Plus Environment

Page 27 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Method 2: 25–100% CH OH/H O in 30 min; 100% CH OH for 5 min; 100–25% CH OH/H O in
3
2
3
3
2
4
min. Method 3: 15ꢀ100% CH OH/H O in 30 min; 100% CH OH for 5 min; 100–15%
3 2 3
CH OH/H O in 4 min. Both solvents contained 0.05 vol % of trifluoroacetic acid (TFA). Purities
3
2
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
of final compounds were established by analytical HPLC, which was carried out using an
Agilent 1100 HPLC system with a Synergi 4 ꢁm HydroꢀRP 80A column, on a variable
wavelength detector G1314A. Method 1: flow rate = 1.4 mL/min; gradient elution over 20
minutes, from 30% MeOHꢀH O to 100% MeOH with 0.05% TFA. Method 2: flow rate = 1.4
2
mL/min; gradient elution over 20 minutes, from 50% MeOHꢀH O to 70% MeOH with 0.05%
2
TFA. The purity of all tested compounds was
above (see supporting information).
≥95% as determined by the method described
General procedure for the synthesis of 4-42.
To a solution of the appropriate carboxylic acid (1 equiv) in anhydrous dichloromethane
CH Cl ) (4 mL/mmol) at room temperature were added anhydrous hydroxybenzotriazole
(
2
2
(HOBt, 1 equiv) and 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide hydrochloride (EDC·HCl,
equiv) under an argon atmosphere. After stirring for 10 min, the appropriate substituted amine
1 equiv) and triethylamine (1.5 equiv) were added, and the reaction mixture was stirred at room
temperature until disappearance of the starting material (usually 12 to 16 h). After this time water
2 mL) was added, and the mixture was extracted with EtOAc (3×10 mL). The organic layers
were separated, washed with brine, dried over anhydrous Na SO , filtered, and concentrated
1
(
(
2
4
under reduced pressure. The residue was purified by flash chromatography on SiO using 0ꢀ40%
2
EtOAcꢀHexane gradient (unless specified differently) to obtain the indolecarboxamides in yields
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ranging from 34 to 95%, which were further purified by preparative HPLC unless otherwise
stated.
1
N-Cycloheptyl-4-methyl-1H-indole-2-carboxamide (4). Yield 84% (white powder). H NMR
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(400 MHz, d ꢁDMSO) δ 11.5 (s, 1H), 8.24 (d, J = 8.0 Hz, 1H), 7.25 (d, J = 8.4 Hz, 1H), 7.22 (s,
6
1
1
1
H), 7.05 (dd, J = 7.2, 8.0 Hz, 1H), 6.83 (d, J = 7.2 Hz, 1H), 4.01 (m, 1H), 2.48 (s, 3H), 1.91ꢀ
1
3
.87 (m, 2H), 1.71ꢀ1.54 (m, 10H); C NMR (100 MHz, d ꢀDMSO) δ 159.9, 136.1, 131.4, 130.2,
6
27.3, 123.2, 119.5, 109.8, 101.2, 49.9, 34.4 (2C), 27.9 (2C), 23.8 (2C), 18.5. HRMS (ESI) m/z
+
calcd for C H N O ([M+H] ) 271.1805; found: 271.1808.
1
7
23
2
1
N-Cyclooctyl-4-methyl-1H-indole-2-carboxamide (5). Yield 74% (white powder). H NMR
(
400 MHz, d ꢁDMSO) δ 11.5 (s, 1H), 8.22 (d, J = 8.0 Hz, 1H), 7.25 (d, J = 9.2 Hz, 1H), 7.22 (s,
6
1
6
1
H), 7.05 (t, J = 7.6 Hz, 1H), 6.83 (d, J = 7.2 Hz, 1H), 4.05 (m, 1H), 2.49 (s, 3H), 1.82ꢀ1.67 (m,
13
H), 1.61ꢀ1.52 (m, 8H); C NMR (100 MHz, d ꢀDMSO) δ 159.9, 136.1, 131.4, 130.2, 127.3,
6
23.2, 119.5, 109.8, 101.2, 48.9, 31.6 (2C), 26.9 (2C), 25.0, 23.4 (2C), 18.5. HRMS (ESI) m/z
+
calcd for C H N O ([M+H] ) 285.1961; found: 285.1967.
1
8
25
2
1
4-Amino-N-cyclooctyl-1H-indole-2-carboxamide (6). Yield 78 % (white powder). H NMR
(
400 MHz, CD OD) δ 7.45 (d, J = 8.4 Hz, 1H), 7.26 (m, 1H), 7.21 (s, 1H), 7.00 (d, J = 7.2 Hz,
3
1
3
1
H), 4.18 (m, 1H), 1.94ꢀ1.64 (m, 14H); C NMR (100 MHz, CD OD) δ 162.2, 139.4, 134.0,
3
125.4, 125.3, 122.3, 113.3, 112.4, 100.9, 51.4, 33.7 (2C), 28.3 (2C), 27.0, 25.3 (2C). HRMS
+
(
ESI) m/z calcd for C H N O ([M+H] ) 286.1914; found: 286.1918.
1
7
24
3
1
N-Cycloheptyl-5-methyl-1H-indole-2-carboxamide (7). Yield 78% (offꢀwhite powder). H
NMR (400 MHz, d ꢁDMSO) δ 11.4 (s, 1H), 8.18 (d, J = 7.6 Hz, 1H), 7.36 (s, 1H), 7.33 (d, J =
6
8.4 Hz, 1H), 7.07 (s, 1H), 7.00 (d, J = 8 Hz, 1H), 4.00 (m, 1H), 2.36 (s, 3H), 1.88 (m, 2H), 1.66ꢀ
13
1
.43 (m, 10H); C NMR (100 MHz, d ꢀDMSO) δ 159.9, 134.7, 132.0, 128.1, 127.3, 124.9,
6
ACS Paragon Plus Environment

Page 29 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
20.6, 111.9, 102.0, 50.0, 34.4 (2C), 27.8 (2C), 23.9 (2C), 21.1. HRMS (ESI) m/z calcd for
+
C H N O ([M+H] ) 271.1805; found: 271.1805.
17 23
2
1
N-Cyclooctyl-5-methyl-1H-indole-2-carboxamide (8). Yield 74% (offꢀwhite powder). 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
NMR (400 MHz, d ꢁDMSO) δ 11.4 (s, 1H), 8.16 (d, J = 7.6 Hz, 1H), 7.36 (s, 1H), 7.32 (d, J =
6
8
6
1
.4 Hz, 1H), 7.07 (s, 1H), 7.00 (d, J = 8.4 Hz, 1H), 4.04 (m, 1H), 2.36 (s, 3H), 1.81ꢀ1.68 (m,
1
3
H), 1.60ꢀ1.53 (m, 8H); C NMR (100 MHz, d ꢀDMSO) δ 159.9, 134.7, 132.0, 128.1, 127.3,
6
24.9, 120.6, 111.9, 102.0, 48.9, 31.6 (2C), 26.9 (2C), 25.0, 23.5 (2C), 21.1. HRMS (ESI) m/z
+
calcd for C H N O ([M+H] ) 285.1961; found: 285.1958.
1
8
25
2
1
5
-Amino-N-cyclooctyl-1H-indole-2-carboxamide (9). Yield 71 % (white powder). H NMR
(
(
400 MHz, d ꢁDMSO) δ 11.79 (s, 1H), 9.85 (s, 2H), 8.31 (d, J = 8.0 Hz, 1H), 7.57 (s, 1H), 7.52
6
1
3
d, J = 8.8 Hz, 1H), 7.22 (s, 1H), 7.14 (d, J = 8.4 Hz, 1H), 4.04 (m, 1H), 1.81ꢀ1.51 (m, 14H); C
NMR (100 MHz, d ꢀDMSO) δ 159.4, 135.1, 133.9, 127.0, 118.0, 114.8, 133.4, 102.6, 49.0, 31.6
6
+
(
2C), 26.9 (2C), 25.0, 23.4 (2C). HRMS (ESI) calcd for C H N O ([M+H] ) m/z 286.1914;
17 24
3
found: 286.1912
1
N-Cyclooctyl-6-methyl-1H-indole-2-carboxamide (10). Yield 78% (white powder). H NMR
(
400 MHz, d ꢁDMSO) δ 11.31 (s, 1H), 8.12 (d, J = 8.4 Hz, 1H), 7.47 (d, J = 8.4 Hz, 1H), 7.20 (s,
6
1
3
1
H), 7.09 (s, 1H), 6.86 (d, J = 8.4 Hz, 1H), 4.03 (m, 1H), 2.38 (s, 3H), 1.77ꢀ1.51 (m, 14H); C
NMR (100 MHz, d ꢀDMSO) δ 159.9, 136.7, 132.3, 131.5, 125.0, 121.6, 121.0, 111.7, 102.4,
6
+
4
8.8, 31.6 (2C), 26.9 (2C), 25.0, 23.4 (2C), 21.5. HRMS (ESI) calcd for C H N O ([M+H] )
18
25
2
m/z 285.1961; found: 285.1959.
6-Bromo-N-cyclooctyl-1H-indole-2-carboxamide (11). The title compound was obtained by
following the general procedure. The crude material was purified by flash column
chromatography on SiO with a hexaneꢀEtOAc gradient as the eluent followed by
2
ACS Paragon Plus Environment