Structure-activity relationships of spectinamide antituberculosis agents: A dissection of ribosomal inhibition and native efflux avoidance contributions

Article abstract of DOI:10.1021/acsinfecdis.6b00158

Spectinamides are a novel class of antitubercular agents with the potential to treat drug-resistant tuberculosis infections. Their antitubercular activity is derived from both ribosomal affinity and their ability to overcome intrinsic efflux mediated by the Mycobacterium tuberculosis Rv1258c efflux pump. This study explores the structure-activity relationships through analysis of 50 targeted spectinamides. Compounds are evaluated for ribosomal translational inhibition, MIC activity in Rv1258c efflux pump deficient and wild type tuberculosis strains, and efficacy in an acute model of tuberculosis infection. The results of this study show a narrow structure-activity relationship, consistent with a tight ribosome-binding pocket and strict structural requirements to overcome native efflux. Rationalization of ribosomal inhibition data using molecular dynamics simulations showed stable complex formation for halogenated spectinamides consistent with the long post antibiotic effects observed. The lead spectinamides identified in this study demonstrated potent MIC activity against MDR and XDR tuberculosis and had desirable antitubercular class specific features including low protein binding, low microsomal metabolism, no cytotoxicity, and significant reductions in bacterial burdens in the lungs of mice infected with M. tuberculosis. The structure-activity relationships detailed here emphasize the need to examine efflux-mediated resistance in the design of antituberculosis drugs and demonstrate that it is possible to overcome intrinsic efflux with synthetic modification. The ability to understand the structure requirements for this class has produced a variety of new substituted spectinamides, which may provide useful alternative candidates and promote the further development of this class.

Full text of DOI:10.1021/acsinfecdis.6b00158

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
The Structure-Activity Relationships of Spectinamide
Antituberculosis agents; a Dissection of Ribosomal
Inhibition and Native Efflux Avoidance Contributions
Jiuyu Liu, David F. Bruhn, Robin B. Lee, Zhong Zheng, Tanja Janusic, Dimitri Scherbakov,
Michael S. Scherman, Helena I.M. Boshoff, Sourav Das, * Raskesh, Samanthi L. Waidyarachchi,
Tiffany A. Brewer, Begona Gracia, Lei Yang, John C. Bollinger, Gregory T. Robertson,
Bernd Meibohm, Anne J. Lenaerts, Jose Ainsa, Erik C. Böttger, and Richard E Lee
ACS Infect. Dis., Just Accepted Manuscript • DOI: 10.1021/acsinfecdis.6b00158 • Publication Date (Web): 09 Oct 2016
Downloaded from http://pubs.acs.org on October 11, 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.
ACS Infectious Diseases 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
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
The Structure-Activity Relationships of Spectinamide Antituberculosis agents; a Dissection
of Ribosomal Inhibition and Native Efflux Avoidance Contributions
1
1
1
1
2
Jiuyu Liu, David F. Bruhn, Robin. B. Lee, Zhong Zheng, Tanja Janusic, Dimitri
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
3
4
1
1
Scherbakov, Michael S. Scherman, Helena I. Boshoff, Sourav Das, Rakesh, Samanthi L.
1
1, 5
6
1
1
Waidyarachchi, Tiffany A. Brewer, Begoña Gracia, Lei Yang, John Bollinger, Gregory T.
3
5
3
6
2
1*
Robertson, Bernd Meibohm, Anne J. Lenaerts, Jose Ainsa, Erik C. Böttger, Richard E. Lee
1
. Department of Chemical Biology and Therapeutics, St. Jude Children's Research Hospital,
62 Danny Thomas Place, MS#1000, Memphis, Tennessee, 38105, USA
2
2
. Institut für Medizinische Mikrobiologie, Nationales Zentrum für Mykobakterien, Universität
Zürich, Rämistrasse 71, Gloriastrasse 30/32, CHꢀ8006 Zürich, Switzerland
3
. Mycobacterial Research Laboratories, Department of Microbiology, Colorado State
University, 1682 Campus Delivery, Fort Collins, Colorado, 80523, USA
4
. Tuberculosis Research Section, Laboratory of Clinical Infectious Diseases, National Institute
for Allergy and Infectious Disease, National Institutes of Health, 33 North drive, Bethesda,
Maryland, 20814, USA
5
. Department of Pharmaceutical Sciences, University of Tennessee Health Science Center, 881
Madison Avenue, Memphis, TN 38163, USA
6. Departamento de Microbiología, Medicina Preventiva y Salud Pública, and BIFI, Universidad
de Zaragoza, 50009ꢀZaragoza, and CIBER Enfermedades Respiratorias (CIBERES), Spain
Correspondence should be addressed to R.E.L. (Richard.Lee@StJude.org)
1
ACS Paragon Plus Environment

ACS Infectious Diseases
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
ABSTRACT
Spectinamides are a novel class of antitubercular agents with the potential to treat drug resistant
tuberculosis infections. Their antitubercular activity is derived from both ribosomal affinity and
their ability to overcome intrinsic efflux mediated by the Mycobacterium tuberculosis Rv1258c
efflux pump. In this study we explore the structure activity relationships through analysis of 50
targeted spectinamides. Compounds are evaluated for ribosomal translational inhibition, MIC
activity in Rv1258c efflux pump deficient and wild type tuberculosis strains, and efficacy in an
acute model of tuberculosis infection. The results of this study show a narrow structureꢀactivity
relationship, consistent with a tight ribosome binding pocket and strict structural requirements to
overcome native efflux. Rationalization of ribosomal inhibition data using molecular dynamics
simulations showed stable complex formation for halogenated spectinamides consistent with the
long post antibiotic effects observed. The lead spectinamides identified in this study
demonstrated potent MIC activity against MDR and XDR tuberculosis and had desirable
antitubercular class specific features including: low protein binding, low microsomal
metabolism, no cytotoxicity, and significant reductions in bacterial burdens in the lungs of mice
infected with M. tuberculosis. The structure activity relationships detailed here emphasize the
need to examine effluxꢀmediated resistance in the design of antituberculosis drugs and
demonstrate that it is possible to overcome intrinsic efflux with synthetic modification. The
ability to understand the structure requirements for this class has produced a variety of new
substituted spectinamides, which may provide useful alternative candidates and promote the
further development of this class.
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
2
ACS Paragon Plus Environment

Page 3 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
KEYWORDS: StructureꢀActivity Relationship, spectinamide, spectinomycin, antituberculosis,
efflux pump, efficacy.
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
ACS Paragon Plus Environment

ACS Infectious Diseases
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
Introduction
Despite decades of chemotherapeutic intervention, tuberculosis infection remains a leading cause
of death worldwide. The increasing incidence of drug resistant tuberculosis infections and
reports of infections resistant to all available antibiotics highlight the striking and urgent need for
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
1
ꢀ3
new tuberculosis drugs and treatment regimes. We have recently demonstrated that synthetic
modification of spectinomycin is a robust approach to enhancing this potent ribosome inhibitor’s
antibacterial activity towards respiratory tract bacterial infections, including Mycobacterium
4
tuberculosis.
Spectinomycin is a natural product aminocyclitol produced by Streptomyces spectabilis. It is a
bacterial ribosome inhibitor that binds selectively to a unique binding site in RNA helix 34 of the
head domain of the bacterial 30S ribosomal subunit, blocking translocation and consequently
5ꢀ6
protein synthesis. Unfortunately, spectinomycin lacks antibacterial activity towards common
4
, 7
bacterial pathogens primarily due to efflux mechanisms. It is, however, an attractive scaffold
for the design of new agents since it is a potent inhibitor of bacterial ribosomes but not
mammalian ribosomes, it has an established high safety profile, and unlike other
8
aminoglycosides it does not have any reported ototoxicity or nephrotoxicity. The structure of
spectinomycin is unique amongst aminoglycoside antibiotics in which actinamine forms both
9
acetal and hemiacetal linkages with actinospectose. Previous efforts to develop spectinomycin
analogs with improved whole cell potency in the 1980’s indicated that the actinamine ring
10ꢀ12
portion of spectinomycin could not be modified without compromising potency,
but
1
3ꢀ15
modifications to the 3’ and 5’ positions of actinospectose ring (Cꢀring) were well tolerated.
Utilizing these early drug discovery studies and our modern knowledge of the molecular
6
interaction of spectinomycin with the bacterial ribosome, we have been able to generate two
4
ACS Paragon Plus Environment

Page 5 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
chemically distinct series of spectinomycins: spectinamides with narrow spectrum activity
4
against M. tuberculosis; and the aminomethyl spectinomycins active against a broader range of
1
6
respiratory tract and sexually transmitted bacterial pathogens.
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
This report examines the detailed structure activity relationship (SAR) of our antitubercular
spectinamide series. The spectinamides lack crossꢀresistance with existing tuberculosis
therapeutics, retain spectinomycin’s high selectivity index and are efficacious in vivo against
multiple mouse models of tuberculosis infection. Spectinomycin’s weak activity in M.
tuberculosis is the result of efflux by Rv1258c. Unlike spectinomycin, lead spectinamides are
not subject to efflux by Rv1258c, which is particularly notable given the upregulation of this
pump in some strains of multiple drug resistance (MDR) tuberculosis and its role in macrophage
1
7
induced drug tolerance. Herein, we report the synthesis of 41 new spectinamides and present a
comprehensive SAR study (50 analogs in total) for this series, discussing the structural
requirements for potent inhibition of bacterial protein synthesis, efflux avoidance, and
antituberculosis activity in vitro and in vivo.
Results and Discussion
Chemistry.
The spectinamides were synthesized in a rapid fourꢀstep protocol from spectinomycin using the
1
5
general procedure of Woitun (Scheme 1). Structural diversity was incorporated into the
sidechain in the key acylation step using targeted aryl acids. Optimization of the final
deprotection step was performed by either catalytic hydrogenation, using palladium on carbon in
methanolic HCl for 2 h, or in the case of hydrogenationꢀsensitive side chains by acid hydrolysis
with 48% HBr solution at room temperature for 2 h. During scale up for in vivo testing, the
synthesis protocol for many compounds can be further simplified by crystallization of
5
ACS Paragon Plus Environment

ACS Infectious Diseases
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
carboxybenzyl (Cbz) protected intermediate after the key acylation step. This removes the need
for column chromatography. The spectinamides were provided as neutral dihydrochloride salts
for biological testing unless otherwise noted.
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
Scheme 1. Reagents and conditions: (1) BnOCOCl, NaHCO , acetone/water, room temperature,
3
10 h, 90%; (2) NH NO , 2ꢀMethylpyridine borane , 10% acetic acid in methanol, room
4
3
temperature, 2 h, 40%; (3) HBTU, DIPEA, DMF, room temperature; (4) method 1, 10% Pd/C,
M HCl in MeOH; method 2, 48% HBr.
1
To verify the stereochemistry of spectinamides, the triꢀhydrobromide salt 1 (Figure 1A) was
crystalized from methanolꢀacetone and the resulting crystal structure was determined (Figure 1B,
Supplementary Table S1). These results confirmed the R stereochemistry assignment in the key
3’ position amino substitution position of the actinospectose ring of the bioactive spectinamides.
To make the 3’ꢀ(S) isomer of 1, the 3’ꢀepimeric amine was isolated after the reductive amination
step as a minor component by chromatography. This enantiomer was then taken through the
same reaction scheme as 1 to generate the corresponding 3’ epimer, 5.
6
ACS Paragon Plus Environment

Page 7 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
OH
B.
A.
H
N
H
H
O
O
O
3
'
HO
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
H
OH
HN
NH
O
N
Structure of 1
Crystal structure of 1
C.
Figure 1: A. Structure of 1. B. Crystal structure of 1 that confirms 3’ stereochemical assignment
of bioactive spectinamides; C. Spectinamide 1 ribosomal binding mode modeled from a 20 ns
molecular dynamics simulation. The nitrogen atom of the pyridine ring is shown to interact with
the hydroxyl group of C1192. The spectinamide side chain binding pocket is formed between
helix 31 and the RpsE interaction loop. Valine 25 of the RpsE loop is positioned to form a van
der Waals' interaction with the pyridine ring.
Design and Prior SAR observations.
Computational modeling of 1 in the M. tuberculosis ribosome revealed an unexplored pocket
adjacent to helix 31 and the RpsE loop that is accessed by the aryl spectinamide side chain
(Figure 1C). This model indicated hydrogen bonding between the 2ꢀpyridyl moiety of 1 to
hydroxyl groups of C1192 and G1193 that held the ring in position in the pocket. Our original
7
ACS Paragon Plus Environment

ACS Infectious Diseases
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
panel of 16 spectinamide analogs included compound 2, which only has a phenyl ring and lacks
the 2ꢀpyridyl functionality, as well as compounds 3 and 4, which have the 3ꢀ and 4ꢀpyridyl,
respectively. These compounds were interesting as the ribosomal IC s were slightly increased
5
0
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
(3.02ꢀ5.07 µg/ml, Table 1), but the minimum inhibitory concentrations (MICs) for these
compounds were much higher (50ꢀ200 µg/ml, Table 1), more than could be expected based on
the less dramatic change in ribosomal inhibition. When this set of compounds was tested against
M. tuberculosis with Rv1258c inactivated, the MIC of compounds 2, 3 and 4 decreased
substantially (16ꢀ32 fold), but the MIC of 1 remained the same (Table 1). This set of compounds
confirmed that the 2ꢀpyridyl introduces favorable interactions with the ribosomal active site and
indicated it is necessary for avoidance of efflux by pump Rv1258c.
Based on our model of 1 bound to the active site, we proposed three series of modifications to
build the SAR of the series: exploration of the linker to incorporate a spacer that orients the
pyridyl ring in the ribosome side chain binding site; bioisosteric replacement of the 2ꢀpyridyl
ring with different heteroaryl ring systems to enable increased structural diversity; introduction
of electron withdrawing (EW) or electron donating (ED) groups into the pyridine ring to explore
how altering the electrochemical properties of the ring affects ribosome binding and efflux.
Compounds were tested in vitro for inhibition of ribosomal translation using purified M.
smegmatis ribosomes, whole cell activity against M. tuberculosis H37Rv, and for activity against
the efflux deficient M. tuberculosis H37Rv Rv1258c knockout (KO). Compounds fell into four
categories: i) modifications that resulted in a loss of ribosomal activity; ii) modifications that
retained ribosomal inhibition and have similar whole cell activity against both H37Rv wild type
and the Rv1258c KO, indicating the modification enables the compound to avoid efflux; iii)
modifications that retained ribosomal inhibition, but have greater than 2 fold differences in
8
ACS Paragon Plus Environment

Page 9 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
whole cell activity between H37Rv and the Rv1258c KO, indicating they are subject to efflux;
iv) modifications that retained ribosomal inhibition, but resulted in decrease in activity against
H37Rv, indicating a possible reduction in cell permeability, uptake, or accumulation.
To examine the importance of the chirality of the 3’ position at the spectinamide core, the 3’ꢀ(S)ꢀ
isomer of 1 was synthesized (compound 5) and found to be inactive in both whole cell and
ribosomal assays (MIC >200 µg/ml, IC 42.23 µg/ml). This confirmed the antitubercular
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
5
0
activity of 1 remains on target and its antimicrobial activity is dependent on ribosomal inhibition.
Modification to the linker.
Four linker derivatives were synthesized incorporating amine (6), carbamate (7), urea (8), and
sulfonamide (9) linkages. In these initial studies, the phenyl aromatic ring was utilized as 2ꢀ
pyridyl precursors were not commercially available. The lack of the 2ꢀpyridyl resulted in MICs
of 200 µg/ml or greater, and this subset therefore did not contribute to the MIC SAR and were
only considered for ribosomal inhibition. Isosteric linker changes were generally not well
tolerated with respect to ribosomal inhibition, only compound urea 8 (3.77 µg/ml) had
antiribosomal activity similar to 2 (3.02 µg/ml). This compound was later remade to include the
2ꢀpyridyl ring in combination with the urea linker, resulting in compound 10. Binding to the
ribosome for 10 was marginally improved (2.24 µg/ml), although at the MIC level (12.5 µg/ml)
10 was much more active than 8 (>200 µg/ml). When 10 was tested against the Rv1258c KO, its
activity improved further (3.1 µg/ml) indicating that the constraints of the urea linker may
interfere with efflux avoidance. To conclude the linker SAR, the influence of the length of the
spacer between the 3’ꢀamide carbonyl bond to the pyridine ring in 1 was explored. Shortening of
the acetamide to a benzamide spacer (11) led to a decrease in ribosomal inhibition (7.19 µg/ml)
9
ACS Paragon Plus Environment

ACS Infectious Diseases
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
that was accompanied by a corresponding loss of antibacterial activity (MIC >200 µg/ml).
Increasing the length to a propanamide spacer (12) maintained ribosomal affinity (3.77 µg/ml),
but resulted in substantial loss of antibacterial potency (50 µg/ml). The MIC of 12 against the
Rv1258c KO improved 8 fold (6.3 µg/ml), indicating that increasing chain length increases
propensity for efflux by pump Rv1258c. These SAR results support our hypotheses that firstly,
the positioning the pyridyl ring nitrogen in the ribosomal pocket through correct spacer
orientation is key for optimum ribosomal binding enabling stabilizing hydrogen bonding
interactions; secondly, the spatial arrangement of 2ꢀpyridyl relative to the spectinamide core is
key to efflux avoidance.
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 1. Exploring the SAR of 1: the effect of pyridyl nitrogen position (1-4) and the linker (6-9
and 10-12).
H37Rv
MIC
Rv1258c
KO MIC
(µg/ml)
Efflux
Ratio
Ribosome IC50
(µg/mL)
Compound
SPC
(µg/ml)
100
6.3
16
0.36
1
2
3
1.6
50
3.1
1.6
0.5
32
32
16
ꢀ
1.01
3.02
5.07
4.75
42.23
200
200
>200
6.3
4
5
12.5
>200
(
sꢀisomer)
6
200
ND
>200
25
ꢀ
ꢀ
ND
16.62
3.77
7
8
>200
>200
>8
9
>200
12.5
>16
10.97
1
0
ACS Paragon Plus Environment

Page 11 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
1
1
0
1
12.5
>200
50
3.1
>200
6.3
4
ꢀ
2.24
7.19
3.80
12
8
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
Bioisosteric replacement with different heteroaryl ring systems
Bioisosteric replacement of the pyridyl side chain with other heterocylic ring systems was
examined with the purpose of further mapping SAR, finding new potential leads and providing
alternative scaffolds in case of the discovery of unwanted secondary pharmacology associated
with the 2ꢀpyridyl side chain of 1 later in development.
The addition of a second nitrogen to the ring of 2ꢀlinked pyridines produced interesting SAR.
The most pronounced change was observed with the 1,3 pyrimidine 13, which has very poor
ribosomal activity (45.50 µg/ml, Table 2) and has no antitubercular activity, an observation that
is further explored in the molecular modeling section of this study. Pyridazine 14, pyrazine 15,
and pyrimidine 16 also had poor MIC activity (12.5ꢀ25 µg/ml), but in contrast to 13, they are
potent ribosomal inhibitors (≤0.63 µg/ml). The MIC of these compounds did not improve
against the Rv1258c KO, indicating their poor activity is not due to efflux. With the exception of
13, the mechanistic reason for these results is unclear, but we hypothesize that putting the second
nitrogen in the ring may affect uptake or accumulation within the tuberculosis cell.
The second heterocyclic systems explored were the pyridine bioisosteres: 5ꢀmember ring
thiazole (4ꢀyl) 17, aminothiazole 18 and thiazole (2ꢀyl) 19; all compounds in this set were potent
ribosomal inhibitors (<1 µg/ml), with good antitubercular MIC values (≤6.3 µg/ml). These
results are consistent with the medicinal chemistry literature where pyridyl groups are commonly
1
1
ACS Paragon Plus Environment

ACS Infectious Diseases
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
replaced by similar sized thiazole rings. Introduction of smaller 5 member rings with 2 nitrogen
atoms including imidazoles 20, 21, 23 and pyrazole 22 produced analogs with a range of MIC
activities. 5ꢀImidazoyl 20 and 2ꢀimidazoyl 21 had poor MIC values (25 and 12.5 µg/ml), despite
considerable ribosomal inhibition (1.10 and 0.56 µg/ml). These results suggest that uptake of
these compounds is limited by the increased basicity of the imidazole (pKa 7) ring over pyridyl
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
(pKa 5) which may disfavor permeation into cells or promote efflux by a separate pump. In this
subꢀseries, the requirement for a heteroaryl nitrogen adjacent to the linker to avoid efflux is
observed by comparison of the MIC active pyrazole 22 (6.3 µg/ml) and the MIC inactive
imidazole 23 (100 µg/ml), which is effluxed (8ꢀfold).
Finally, a series of phenyl fused fiveꢀsix member heteroaryl bicyclic ring systems that all retain
the key 2ꢀlinked aryl nitrogen were synthesized to determine if a larger group could be tolerated
both stereochemically and electrochemically at the back of the five member ring. The 2ꢀlinked
benzoxazole 24, benzthiazole 25, and benzimidazoles 26 / 27 all had very good MICs/ribosomal
inhibition and were not subject to efflux. Indole 28, which has an NH instead of an N adjacent to
the linker position, however was subject to a very high efflux ratio (32 fold) with an MIC of 25
µg/ml. These results indicate the importance that the key heteroaryl nitrogen does not become
sufficiently basic to be protonated at physiological pH for MIC activity. This protonation state is
likely a contributing factor when comparing the MIC activities of the contrasting matched pair of
the weakly active imidazole 21 (pKa 7.0) and strongly active benzimidazole 27 (pKa 5.6).
1
2
ACS Paragon Plus Environment

Page 13 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
Table 2. Examination of different heteroaryl ring systems
H37Rv MIC Rv1258c KO
Compound
Ribosome
Efflux Ratio
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
(µg/ml)
MIC (µg/ml)
IC (µg/mL)
5
0
1
1.6
3.1
0.5
ꢀ
1.01
13
>200
>200
45.50
0.63
2
2
5
5
12.5
2
14
25
25
1
0.62
0.61
0.65
15
1
2.5
0.5
0.5
16
17
3.1
6.3
6.3
18
6.3
1
0.62
19
20
21
22
23
6.3
25
6.3
12.5
6.3
1
2
2
1
8
0.72
1.10
0.56
0.42
1.51
12.5
6.3
6.3
100
12.5
6
.3
1.6
.3
1.6
3.1
0.8
3.1
1.6
0.8
2
2
0.61
1.09
0.26
0.26
0.77
2
4
5
2
26
6
2
27
1
28
2
5
32
1
3
ACS Paragon Plus Environment

ACS Infectious Diseases
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
Introduction of electron donating/withdrawing group into the pyridine ring
The final component of SAR explored in this study is the effect of substitutions to the 2ꢀlinked
pyridine ring. Spectinamides were synthesized containing electron donating (ED) or electron
withdrawing (EW) substituents placed around the pyridyl ring in order to determine if
modulating the electrochemistry of the ring could fine tune either the ribosomal binding or efflux
avoidance (Table 3). All the compounds generated had good ribosomal inhibition (≤1 µg/mL).
In the first subꢀseries, a methyl group was introduced into each of the four free positions around
the 2ꢀlinked pyridyl ring. The addition of the methyl group to either the 3 (29) or 6 (32)
positions was not well tolerated due to an increased susceptibility of the compounds to efflux.
The 4 (30) and the 5 (31) substituted isomers had good MIC values suggesting that these are the
best positions for further modification. The substitution at the 4 position (33-36) has significant
effects on the chemistry of the critical pyridyl nitrogen. Compounds with electron donating
groups were not favored, having poor (12.5ꢀ25 µg/ml) MIC activity and were subject to mild
efflux. It is likely that 35 equilibrates to the corresponding 4ꢀpyridone in which the pyridyl
nitrogen is protonated, which is consistent with prior SAR where the heteroaryl nitrogen is
protonated (28 and 20). The isopropyl substituted analog 33 had reasonable MIC activity (6.3
µg/ml), but was subject to mild efflux (4ꢀfold). This suggests that introducing sterically larger
groups to this position can be tolerated, which is consistent with our molecular modeling
experiments. The 4ꢀchloro substituted analog 36 was the most active in this subset, indicating
that electron withdrawing groups are favored in this position. Unfortunately, attempts to
synthesize the corresponding 4ꢀfluoro analogs failed due to chemical reactivity of 4ꢀ
fluoropyridines to nucleophilic attack, which is also a concern for the 4ꢀchloropyridine analogs.
The pharmacological implications in regard to safety and potential susceptibility to conjugative
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
1
4
ACS Paragon Plus Environment

Page 15 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
phase II metabolism with glutathione of 4ꢀchloro substituted pyridines, is an issue we have
1
8
recently evaluated in detail.
The 5ꢀposition is the most tolerant position for substitution and can be substituted with a variety
of electron withdrawing and donating groups, all with good chemical stability. The MIC activity
for all eight compounds (ED, 37ꢀ39; EW, 40 and 41; halogen, 42ꢀ44) with mono substitution at
this position was good, consistent with our modeling experiments that indicate that this position
points largely away from the ribosome. The most notable compounds in this subset are the
halogenated analogs 44, 42, 43 and hydroxyl analog 39, all with excellent MIC values and the
halogenated analogs all having long post antibiotic effects as discussed later (Table 5). The
difference in MIC between 4 (35, 25 µg/ml) and 5 (39, 1.6 µg/ml) hydroxy substitutents is
profound. Overall, the tolerability of this position to the introduction of different functional
groups especially electron donating groups provides opportunity to perform further chemistry at
the 5 position.
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
Dual substitutions to the pyridyl ring were explored (Table 4) using 4ꢀchloro substituted 36 as a
reference. Introduction of an isopropyl group at the 3ꢀposition led to a decrease in MIC activity
(45, 50 µg/ml), due in part to a small loss of ribosomal inhibition (2.3 µg/ml), but most
prominently due to increased efflux (32 fold). 4ꢀchloro and 5 position dual substitution using
methyl (47), methoxy (48), fluoro (49) produced compounds with excellent MIC and ribosomal
inhibition that were not subject to efflux. As previously noted 4ꢀchloro substituted pyridines are
potentially chemically reactive, producing a concern that they may be subject to phase II
conjugative metabolism. However, we have demonstrated that introduction of an electron
withdrawing 5ꢀfluoro group (49) can stabilize the 4ꢀchloro group, while compounds with
1
8
electron donating groups, such as 47 and 48 are more electrophilic and subject to metabolism.
1
5
ACS Paragon Plus Environment

ACS Infectious Diseases
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
The final compound in this subset is 50, a 4ꢀmethoxy and 5ꢀfluoro substituted analog, essentially
a hybrid of 34 and 42. Unfortunately, only moderate MIC activity was achieved, consistent with
the notion that an electron donating functional group in the 4ꢀposition is disfavored.
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 3. Introducing electron donating/withdrawing group(s) into pyridine ring
H37Rv
MIC
(µg/ml)
Rv1258c
KO MIC
(µg/ml)
Ribosome
IC50
(µg/mL)
Efflux
Ratio
Compound
1
1.6
3.2
0.5
1.01
29
30
25
3.1
3.1
6.3
6.3
3.1
4
0.5
1
1.00
0.47
0.54
3
1
32
25
6.3
4
0.50
3
3
6.3
12.5
25
1.6
3.1
4
4
2
1
0.36
0.39
0.44
0.54
34
35
12.5
1.6
3
6
1.6
37
6.3
3.1
1.6
3.1
3.1
1.6
3.1
0.8
6.3
6.3
3.1
1.6
0.8
3.1
1.6
0.8
1
0.5
0.5
2
0.30
0.39
0.84
0.44
0.96
0.55
0.42
0.53
3
8
39
40
41
42
43
44
4
0.5
2
1
1
6
ACS Paragon Plus Environment

Page 17 of 53
ACS Infectious Diseases
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
Table 4. Introducing dual groups into pyridine ring
Ribosome
IC50
H37Rv MIC Rv1258c KO
Compound
Efflux Ratio
(µg/ml)
MIC (µg/ml)
(µg/mL)
36
45
1.6
1.6
1.6
1
0.54
50
32
2.31
4
6
3.1
0.8
1.6
0.8
2
1
0.59
0.23
47
4
4
5
8
9
0
3.1
0.8
6.3
3.1
0.8
3.1
1
1
2
0.42
0.22
0.92
Ribosome SAR summary
The ribosomal SAR observed with respect to substitution of the aryl sidechain to the 3’ position
of the spectinomycin ring largely rests on positioning the 2ꢀpyridyl ring correctly in the binding
site. First, chirality of the ring substitution is absolute with only 3’ (R) amides active. Linker
length and type is also key in positioning the side chain ring to make optimum interactions. A
shorter linker such as 11 or alternatives such as carbamate (7) or sulfonamide (9) were
disfavored. The position of the linker to nitrogen in the pyridyl ring also had a significant effect,
with the 2ꢀlinked series being optimum. In terms of substitution to the pyridyl ring the most
dramatic loss in affinity was generated by the addition of a second nitrogen adjacent to the linker
(13). The remainder of the compounds were designed to fit within the envelope of the
1
7
ACS Paragon Plus Environment

ACS Infectious Diseases
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
spectinamide side chain binding site formed adjacent to helix 34 and below RpsE interaction
loop, details of which are discussed in greater detail in the subsequent molecular modeling
section.
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
Efflux SAR summary
In regard to the key element of SAR – efflux avoidance, critical to the success of this series is the
presence of a nonꢀprotonated heterocyclic nitrogen immediately adjacent to linker to the
spectinomycin core. Within these tables there are several examples (1 vs 2, 8 vs 10, 22 vs 23)
that if the key nitrogen is replaced, MIC activity is adversely affected in wild type M.
tuberculosis but the MIC activity is retained in the tuberculosis strain in which the Rv1258c
efflux pump has been deleted. The need for nonꢀprotonated heterocyclic nitrogen (N not NH) to
avoid efflux is also clearly demonstrated in the 27 and 28 pair.
Within this requirement, different heteroaryl rings are well tolerated providing opportunities to
expand the series and manipulate potential offꢀtarget pharmacologic effects. Steric and
electronic substitutions to the heteroaryl (pyridyl) ring alter the propensity of compounds in this
series to avoid efflux. Electron donating substitutions to the 4ꢀposition of the pyridyl ring
increase the propensity for efflux and decrease MIC activity. Conversely, introduction of 4ꢀ
position of electron withdrawing substitutions appears favorable with the caveat of introduction
of potential chemical instability. Modifications to the 3 and 6 positions appear sterically
unfavorable as evidenced by 29, 32, and 45, potentially indicating that there is a biochemical
recognition element to efflux avoidance. Substitutions to the pyridyl 5 position are broadly
tolerated in respect to the efflux SAR and may be coupled with 4ꢀposition substitutions to
produce potent dual substituted compounds such as 49. To ensure no MIC biases were
accidentally introduced in the creation of Rv1258c KO strain, a complemented strain in which
1
8
ACS Paragon Plus Environment

Page 19 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
the Rv1258c pump was reintroduced to the KO strain was used as a control. All MICs of the
compounds in this series against the complemented strain were within one well of the H37Rv
wild type parental strain indicating that the efflux effects noted in this manuscript were specific
to the Rv1258c pump.
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
Modeling
To understand the structureꢀactivity relationship with respect to ribosome affinity, key
compounds were docked into the active site and then subjected to a 20 ns molecular dynamics
simulation. These experiments were directed at understanding how small modifications on the
aryl side chain can affect binding affinity, such as observed with the improvement in activity
with side chains from benzyl (2) to pyridyl (1) and to 4ꢀchloropyridyl (36). In the case of 2, the
benzene ring remained significantly more dynamic than 1 whose pyridyl nitrogen provided a
possibility for a hydrogen bond that stabilized the ring in the simulations and increased the
predicted binding affinity by 0.7 kcal/mol. With two possible hydrogen bonding partners
(C1192, G1193) (Figure 2A and C), the pyridine ring however did not adopt a single pose in the
20 ns production run. Compound 13 (IC =45.5 µM) with two nitrogens bound 39 fold weaker
5
0
than 1 with a calculated relative binding free energy of 2.21 kcal/mol. (Table S2). The
decomposed energy terms suggested that both the electrostatic and nonꢀpolar component of the
relative ꢀGbind was less favorable than reference compound 1. To rationalize this, we calculated
the DonorꢀAcceptor (DꢀA) distances between the two nitrogen atoms on the ring and
o
G1193/C1192. Using a DꢀA cutꢀoff distance of 4.0 Å and a cutꢀoff angle of 60 for HꢀDꢀA,
hydrogen bond with C1192 was present only 11% of the time, and the interaction with G1193
1
9
ACS Paragon Plus Environment

ACS Infectious Diseases
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
was even weaker. Thus with the addition of the second nitrogen, the pyrimidine did not establish
a second favorable interaction. A likely explanation is that the two hydrogen bonds competed
with each other in 13 resulting in an overall weaker hydrogen bond and disfavored interaction
with the lipophilic RpsE residues at the back of the aryl binding pocket.
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
Adding a chlorine to the pyridine ring (36) allowed an additional favorable interaction between
the halogen and amide backbone of F30, and in conjunction with hydrogen bonding of the
pyridyl nitrogen and G1193, locked the ring movement in the simulation (Figure 2 B and D).
This increased binding affinity by a predicted 0.5 kcal/mol. It is tempting to speculate that this is
the reason that the halogenated pyridines such as 36 possess long post antibiotic effects as
discussed further in the next section, though experimentally, residence time determinations are
hard to perform on such a complex macromolecular system as the bacterial ribosome.
2
0
ACS Paragon Plus Environment

Page 21 of 53
ACS Infectious Diseases
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 2. Snapshots taken from the last 20ns MD trajectory for 1 (A) and 36 (B). The two
possible hydrogen bonding interactions made by 1 are shown in magenta and teal (A).The
AcceptorꢀDonor (AꢀD) distances (in Å) between the pyridine Nitrogen and G1193 and C1192
plotted for 1 (C) and 36 (D).
Post antibiotic effect.
While MIC is an important parameter used in the development of antibacterial compounds, it
observes antimicrobial activity with static drug exposure unrepresentative of the oscillating
exposures encountered by a pathogen in vivo. Metabolism and clearance of compounds in vivo
2
1
ACS Paragon Plus Environment

ACS Infectious Diseases
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
results in a transient period of time in which a compound must exert an effect sufficient to kill
and suppress bacterial reꢀgrowth until the next dose is administered. For injectable compounds
1
9
with relatively short in vivo halfꢀlives, such as the spectinomycin, we anticipated this is could
be an important driver of in vivo efficacy. As a way to prioritize compounds for later resource
and time intensive in vivo efficacy experiments, we performed post antibiotic effect (PAE)
experiments in vitro using M. bovis BCG as a safer tuberculosis surrogate (Table 5). Within the
limited number of compounds tested, a clear SAR was observed; unsubstituted pyridyl and
thiazole analogs, 1 and 17, exhibited moderate PAEs (23ꢀ27 hrs), whereas analogs with
halogenated substitutions, 47,49, 49 and 44, exerted a lengthy PAE that exceeded 100 hours.
The PAE of the hydroxyl analog 39 (20 hours) was lower that the corresponding halogens.
These results are consistent with the previous molecular dynamics experiments, which predict
enhanced ribosomal residency times of the halogenated spectinamides.
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
In vitro ADME profiling
Select lead compounds were subject to further pharmacological profiling, including in vitro
testing of cytotoxicity, microsomal stability, protein binding; all key factors when evaluating
potential antibiotic substances for further advancement. Consistent with our previously published
studies on 44, 36 and 1, all the new spectinamides tested in these experiments exhibited no
apparent cytotoxicity to HepG2 cells (all >400 µg/mL). The stability to both human and mouse
microsomal metabolism was also excellent for most compounds in the series. All compounds in
the series had a human microsomal halfꢀlife ≥6.3 h, except for 36 and 38, indicative of low phase
I hepatic metabolic potential for the series. As we have noted previously, 47 and 36 are subject
2
2
ACS Paragon Plus Environment

Page 23 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
18
to some phase II conjugative metabolism. Protein serum binding was low for all compounds
with the lowest levels for 39 (35%) and highest for 49 (56%), which is consistent with increasing
lipophilicity of the pyridyl substitutions. The negligible toxicity towards mammalian cells, low
propensity for phase I metabolism, low protein binding and high solubility are shown here to be
a clear and desirable general properties of this antibiotic class.
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 5. In vitro ADME, PAE and cytotoxicity testing of selected series leads.
Microsomal
Protein
stability
MIC
Post
Cytotoxicity
(HepG2)
binding (%)
t1/2(h)
#
(H Rv)
Antibiotic
37
^
(
µg/mL)
Effect (h)
(µg/mL)
human mouse human mouse
SPC*
1*
100
1.6
1.6
0.8
3.1
0.8
ꢀ
2.5
>10
2.6
>10 51.6 52.6
>10 62.1 36.9
>400
>400
>400
>400
>400
>400
23 (8)
66 (29)
3
6*
4*
1.5
>50
5.6
56.4 39.8
4
133 (26) 21.2
ꢀ
29.8
17*
27 (13)
18.5
>50
52.4 41.6
49
116 (19)
>50 56.3 39.3
2
3
ACS Paragon Plus Environment

ACS Infectious Diseases
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
4
6
3.1
0.8
1.6
3.1
1.6
3.1
3.1
ꢀ
>50
>50
7.4
>50 46.6 36.4
>50 43.9 42.7
13.4 35.1 30.7
>400
>400
>400
>400
>400
>400
>400
47
102 (15)
20 (18)
56 (12)
137
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
39
38
27
3.4
4.8
51.4 57.3
>50
>50 26.3 12.1
41
61 (17)
72 (10)
22.4 27.0 51.4 58.4
48
6.3
5.5
61.6 72.9
^
Results of PostꢀAntibiotic Effect experiments are presented (in hours) as the average of 2ꢀ3 biologically
independent replicates, with standard deviation indicated in parenthesis using M. bovis BCG. * previously
reported in ref 4,
In vivo efficacy
To acertain if there was any notable difference in the in vivo efficacy or tolerability in relation to
the selection of the spectinamide side chain, key spectinamides were tested in a mouse model of
acute tuberculosis infection (Figure 3 AꢀC) and for their maximum tolerated dose (Table S3).
Data shown here results from two new separate trials (Figure 3B and 3C, Table S4) with the
spectinamides being dosed at 200mg/kg twice daily (BID) for 9 days unless otherwise noted.
These data can be compared with our previous trial with 1, 42, 36, 44 performed in identical
2
4
ACS Paragon Plus Environment

Page 25 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
manner (Figure 3A). However, it must be noted that comparative significance can only be
studied within individual trials as bacterial loads often differ between trials.
In the first new trial, three representative heteroaryl compounds with low MICs and excellent
ribosomal inhibition (thiazole 17, benzothiazole 25 and benzimidazole 27) were chosen to
determine if the pyridyl group could be replaced without penalty, and three dualꢀsubstituted 4ꢀ
chloro analogs with low MICs and excellent ribosomal inhibition (49, 47, and 46) were chosen to
see if dual substitution was advantageous. From this group, benzothiazole 25 was excluded due
to mild toxicity noted when dosage exceeded 100 mg/kg during preliminary maximum tolerated
dose testing. Both 17 and 27 reduced the bacterial load in mice with dosing at 200 mg/kg BID 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
1.6 logs. Treatment with 3ꢀ methyl substituted 46 produced the least reduction in bacterial loads,
while 5ꢀ substituted analogs 49 and 47 both exerted more than 2 logs of bacterial killing. Perhaps
the most notable result from this trial was that the efficacy of 49 was comparable to the clinically
used antitubercular, streptomycin, which was adminstered using the same dosing strategy.
In the second new efficacy trial, compounds 41, 39, 38, 48 were tested to explore the SAR of
nonꢀhalogen electron donating groups in the 5ꢀposition. Compound 41, the 5ꢀ trifluoromethoxy
substituted analog, exerted lowest antibacterial activity (1.1 log reduction in lungs). Two of the
most potent analogs in terms of MIC, 39 and 38, yielded 1.5 and 1.6 logs of killing. 48 (the 4ꢀ
chloro 5ꢀmethoxyl analog) yielded a 1.6 log reduction in pulmonary load, similar to that of
streptomycin in this trial (l.6 log CFU reduction). These results suggest that switching the 5ꢀ
substitution between electron withdrawing (42, Figure 3A) and electron donating (39 and 38)
substitutions have little effect on the in vivo efficacy in this model. It should also be noted that
the BID dosing schedule used in this trial might negate the effects of the prominent differences in
PAE between compounds with this substitution which may be more likely to be observed in long
2
5
ACS Paragon Plus Environment

ACS Infectious Diseases
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
20
term efficacy trials of chronic infection using once daily dosing 5 days a week. Previous
pharmacokinetic testing on 1, 42, 36, and 44, demonstrated that compounds in this class share
4
similar profiles with short halfꢀlives (0.44ꢀ1.0h) dictated by renal clearance. This is driven by
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 spectinomycin core rather than the spectinamide sidechain, with exposures similar to
1
6
aminoglycoside antibiotics.
The results of the new efficacy trials were similar to our prior studies (Figure 3A). All new
analogs tested (Figure 3B and 4C) demonstrated statistically significant in vivo efficacy in
reducing bacterial burden in lungs of gammaꢀinterferon KO mice when compared to saline
controls (P < 0.001), showing more than one log CFU reduction versus the control after nine
1
0
consecutive days of treatment. All compounds with the exception of 41 (100 mg/kg) were well
tollerated at a dosing of 200 mg/kg (Table S3). The most active compounds generally were
those with the lowest MIC values. The stand out compounds were 49, with potency equivalent
to that of streptomycin when administered at the same dose (P > 0.05), and 39, the compound
that was tolerated the best, with a maximum tolerated dose (MTD) >400 mg/kg in mice. These
results show that there are many bioiosteric ring choices for the spectinamide side chain, which
may be useful should adverse events be found associated with specific side chains.
2
6
ACS Paragon Plus Environment

Page 27 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
A.
3
.5
2
B.
C.
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
1.5
1
0.5
0
-
0.5
-1
Figure 3. Efficacy of spectinamides in a gamma interferon KO mouse model of acute
4
tuberculosis infection. Results are from three separate experiments (A) , (B) and (C) and show
the reduction in lung CFU counts in comparison to PBS injected control. The reduction of
bacterial load in lungs of infected mice was calculated as the difference between the vehicle
carrier injected control group and groups treated for 9 days with streptomycin or experimental
compounds. Groups of 4ꢀ7 mice were administered treatment doses at 200 mg/Kg twice daily,
execpt 41 at 100 mg/Kg. Error bars indicate the SEM.
Activity against drug resistant clinical M. tuberculosis isolates
To further confirm 49 and 39 as viable candidates, MIC testing was performed against a panel of
2
9 clinical isolates, including 18 MDR and 5 extensively drug resistant isolates (Table S5). The
MIC , the lowest concentration of the antibiotic at which 90% of the isolates were inhibited, and
9
0
MIC range in this panel for 49 was 0.78 (0.39ꢀ1.56) and 39 was 3.13 (1.56ꢀ3.13), both
impressively narrow. These results indicate a low potential of cross resistance to exisiting drugs,
similar to our current lead 44.
2
7
ACS Paragon Plus Environment

ACS Infectious Diseases
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
CONCLUSION
The structure activity relationships detailed in this manuscript provides further insight into the
important role efflux plays in intrinsic antibiotic resistance in tuberculosis. The ability to
understand the structural rules for the spectinamide class has produced a variety of substituted 2ꢀ
Nꢀheteroaryl linked spectinamides with proven in vivo efficacy. Having a 2ꢀNꢀheteroaryl motif
hydrogen bond donor is essential for evading Rv1258c mediated efflux modulating activity.
Computational studies show modifications to the 2 linked pyridyl ring help lock orientation of
the aryl ring in a desirable position in the ribosome, corresponding to the long PAE in 4 and 5
halogenated analogs. These key computational and SAR findings should enable further rounds of
inhibitor design. Based on IC , potency, and efficacy data, 49 (MIC potency, metabolic
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
0
stability and long post antibiotic effect) and 39 (MIC potency, high tolerability) could be
potential antituberculosis drug candidates for further preꢀclinical evaluation providing
alternatives therapeutic to current lead 44 for the development of this class of antitubercular
agents. Studies are currently ongoing to examine the mechanism of uptake of the spectinamide
series, synergy mechanisms with other antituberculosis agents such as rifampin and
pyrazinamide, and the ability of these agents to partner with other tuberculosis experimental
therapeutics to shorten tuberculosis treatment times for MDRꢀtuberculosis.
2
8
ACS Paragon Plus Environment

Page 29 of 53
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
Spectinomycin dihydrochloride pentahydrate was purchased from Waterstone Technology
catalog number 81249, CAS number 22189ꢀ32ꢀ8, 95% in purity). All solvents used for
(
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
chromatography and liquid chromatography were purchased from Aldrich. Flash column
chromatography silica cartridges were obtained from Biotage Inc. Reactions were monitored by
thinꢀlayer chromatography (TLC) on preꢀcoated Merch 60 F254 silica gel plates and visualized
using UV light (254 nm). A Biotage FLASH column chromatography system was used to purify
1
mixtures. H NMR spectra were recorded on a Varian INOVAꢀ500 spectrometer or on a Bruker
400 MHz NMR spectrometer. Chemical shifts (δ) are reported in parts per million relative to the
residual solvent peak or internal standard (tetramethylsilane), and coupling constants (J) are
reported in hertz (Hz). High resolution mass spectra were recorded on a Waters Xevo G2 QTOF
LC–MS using ESI. Purity of the products was confirmed by UPLC/MS (the Waters Acquity).
Elemental analysis was tested by Atlantic Microlab Inc. Xꢀray was analyzed on Bruker D8
diffractometer.
Synthesis of 1, 3-Bis (N-benzyloxycarbonyl)-spectinomycin:
The title compound was synthesized according to the procedure described in the Journal of
2
1
Organic Chemistry.
Synthesis of 1, 3-Bis (N-benzyloxycarbonyl)-3’-dihyro-3’-deoxy-3’(R)-aminospectinomycin:
1, 3ꢀBis (Nꢀbenzyloxycarbonyl)ꢀspectinomycin (6.6g, 11mmol) and NH NO (8.8g, 110mmol) in
4 3
100mL of 10% acetic acidꢀmethanol were stirred to clear solution at room temperature. 2ꢀ
Methylpyridine borane (0.82g, 7.7mmol) was added to the reaction mixture in one portion. The
2
9
ACS Paragon Plus Environment