Evaluating the Contribution of Transition-State Destabilization to Changes in the Residence Time of Triazole-Based InhA Inhibitors

Article abstract of DOI:10.1021/jacs.6b11148

A critical goal of lead compound selection and optimization is to maximize target engagement while minimizing off-target binding. Since target engagement is a function of both the thermodynamics and kinetics of drug-target interactions, it follows that th

Full text of DOI:10.1021/jacs.6b11148

Subscriber access provided by UNIV TORONTO
Article
Evaluating the contribution of transition state destabilization to
changes in the residence time of triazole-based InhA inhibitors
Lauren A. Spagnuolo, Sandra Eltschkner, Weixuan Yu, Fereidoon Daryaee, Shabnam Davoodi,
Susan E. Knudson, Eleanor K. H. Allen, Jonathan Merino, Annica Pschibul, Ben Moree, Neil
Thivalapill, James J. Truglio, Joshua Salafsky, Richard A. Slayden, Caroline Kisker, and Peter J Tonge
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b11148 • Publication Date (Web): 02 Feb 2017
Downloaded from http://pubs.acs.org on February 5, 2017
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 the American Chemical Society 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 48
Journal of the American Chemical Society
1
2
3
4
5
6
Evaluating the contribution of transition state destabilization to changes in the
residence time of triazoleꢀbased InhA inhibitors
7
8
9
Lauren A. Spagnuolo,†,‡ Sandra Eltschkner,§,‡ Weixuan Yu,†,‡ Fereidoon Daryaee,† Shabnam
Davoodi,† Susan E. Knudson,¥ Eleanor K. H. Allen,† Jonathan Merino,† Annica Pschibul,§
Ben Moree,€ Neil Thivalapill,∏ James J. Truglio,∏ Joshua Salafsky,€ Richard A. Slayden, ¥*
Caroline Kisker,§,* and Peter J. Tonge†,*
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
Author Affiliations:
†Institute of Chemical Biology and Drug Discovery, Department of Chemistry, Stony Brook
University, Stony Brook, NY, 11794ꢀ3400
§Rudolf Virchow Center for Experimental Biomedicine, Institute for Structural Biology,
University of Würzburg, 97080 Würzburg, Germany
¥Department of Microbiology, Immunology and Pathology, Colorado State University, Fort
Collins, CO 80523ꢀ2025
€Biodesy, Inc., 384 Oyster Point Boulevard, South San Francisco, CA 94080
∏Great Neck South High School, 341 Lakeville Rd, Great Neck, NY 11020
‡ These authors contributed equally to this work
*Corresponding authors:
peter.tonge@stonybrook.edu,
caroline.kisker@virchow.uniꢀwuerzburg.de,
richard.slayden@colostate.edu
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 48
1
2
3
4
Abstract
5
6
7
8
A critical goal of lead compound selection and optimization is to maximize target
engagement whilst minimizing offꢀtarget binding. Since target engagement is a function of
both the thermodynamics and kinetics of drugꢀtarget interactions, it follows that the structures
of both the ground states and transition states on the binding reaction coordinate are needed
to rationally modulate the lifetime of the drugꢀtarget complex. Previously, we predicted the
structure of the rateꢀlimiting transition state that controlled the timeꢀdependent inhibition of
the enoylꢀACP reductase InhA. This led to the discovery of a triazoleꢀcontaining diphenyl
ether with an increased residence time on InhA due to transition state destabilization rather
than ground state stabilization. In the present work, we have evaluated the inhibition of InhA
by 14 triazoleꢀbased diphenyl ethers and used a combination of enzyme kinetics and Xꢀray
crystallography to generate a structureꢀkinetic relationship (SKR) for timeꢀdependent
binding. We show that the triazole motif slows the rate of formation for the final drugꢀtarget
complex by up to three orders of magnitude. In addition, we identify a novel inhibitor with a
residence time on InhA of 220 min which is 3.5ꢀfold longer than that of the INHꢀNAD adduct
formed by the tuberculosis drug, isoniazid. This study provides a clear example in which the
lifetime of the drugꢀtarget complex is controlled by interactions in the transition state for
inhibitor binding rather than the ground state of the enzymeꢀinhibitor complex, and
demonstrates the important role that onꢀrates can play in drugꢀtarget residence time.
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
2
ACS Paragon Plus Environment

Page 3 of 48
Journal of the American Chemical Society
1
2
3
4
Introduction
5
6
7
8
Drugꢀtarget interactions often occur under conditions where the concentration of the
drug or target is not constant, and thus both the thermodynamics and kinetics of drug binding
are required to fully account for timeꢀdependent changes in target occupancy in the human
body.1ꢀ4 However, often only equilibrium parameters such as IC50 values are used for
selecting and optimizing drug candidates, neglecting the potential contribution that kinetic
selectivity can make to the therapeutic index. This is important since the rate of drugꢀtarget
dissociation can occur on the same time scale as clearance of the drug from the body, and
thus even small changes in residence time can have a dramatic effect on designing dosing
regimens that widen the therapeutic window.5,6 Consequently, the structural and mechanistic
factors that control the lifetime of a drugꢀtarget complex must be fully understood to deploy
the power of drugꢀtarget kinetics in selecting and optimizing drug leads. Whilst there is
growing realization that drugꢀtarget binding kinetics can play a major role in improving the
therapeutic window, several barriers exist including the lack of extensive structureꢀkinetic
relationships (SKR) to guide the development of compounds with altered drugꢀtarget
residence times, and insufficient knowledge of the molecular factors that control the lifetime
of the drug target complex.
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
InhA, the FabI enoylꢀACP reductase from Mycobacterium tuberculosis, is inhibited
by the tuberculosis drug isoniazid through the formation of an INHꢀNAD adduct which is a
slowꢀonset inhibitor of InhA.7,8,9 Since resistance to isoniazid primarily arises from
mutations in KatG,10,11 the enzyme responsible for activating isoniazid, there has been a
concerted effort to develop direct inhibitors of InhA that do not require activation by KatG.12ꢀ
21
In the course of our InhA inhibitor discovery program, we have identified a series of
compounds that inhibit the enzyme through a twoꢀstep slowꢀonset induced fit mechanism in
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 48
1
2
3
4
5
6
which the rapid formation of the initial EI complex is followed by a slow step leading to the
final EI* complex (Scheme 1).22
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Twoꢀstep inducedꢀfit inhibition mechanism.
Previously, we used a combination of molecular dynamics simulations and Xꢀray
crystallography to map the pathway between the EI and EI* states in InhA23 This study used
the Xꢀray structure of InhA in complex with a rapid reversible inhibitor (PT155) to define the
conformation of EI, and the structure of InhA bound to several slowꢀonset inhibitors to define
the conformation of EI*. The MD simulations revealed that the transition state for
interconversion of the EI and EI* states involved a steric clash between the side chains of
residues in helix 6 and helix 7 which comprise the substrate binding loop (SBL) in InhA.23
We evaluated our model for slowꢀonset inhibition by introducing mutations that reduced the
steric clash and decreased the lifetime of the corresponding drugꢀtarget complex. We then
designed a triazoleꢀbased inhibitor, PT163, that reꢀintroduced the steric clash through its
bulky side chain resulting in a longer residence time on the mutant InhA.24 While PT163
also had longer residence time on wildꢀtype InhA, the change in the off rate (k6 in Scheme 1)
was not accompanied by an alteration in Ki*app, the dissociation constant for the final EI*
complex, suggesting a specific effect on the rate limiting transition state on the binding
reaction coordinate.24
Since triazoles are formed by reactions of alkynes with azides, click chemistry offered
a strategy to interrogate the SKR for InhA inhibition whilst providing a platform for
4
ACS Paragon Plus Environment

Page 5 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
improving the drugꢀlike properties of the diphenyl etherꢀbased scaffold. We performed a full
kinetic characterization of 14 triazoleꢀbased diphenyl ethers and 7 alkylꢀbased diphenyl
ethers and compared the SKR between the two sets of compounds to elucidate the role that
onꢀrates play in modulating the stability of the enzymeꢀinhibitor complex. Progress curve
analysis was used to determine values for Kiapp, the apparent dissociation constant for the
initial EI complex (Kiapp = k4/k3), Ki*app, the apparent dissociation constant for the final EI*
complex, and k6, the rate constant for conversion of EI* to EI. These parameters were then
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
used to calculate k5, the rate constant for conversion of EI to EI* using Ki*app
Kiapp(k6/(k6+k5)) and the overall bimolecular rate constant kon,overall for formation of EI* from
=
E+I using kon,overall = (k5+k6)/Kiapp based on the assumption that k6
formation of active enzyme from EI*.
≈ koff, the rate constant for
To further dissect the contribution that the onꢀrate plays in modulating the residence
time of inhibitors on InhA, we used a second harmonic generation (SHG) method to monitor
the kinetics of drugꢀtarget complex formation for three compounds that were representative
of the range of binding kinetics observed in compounds populating the EI and EI* states.
Taken together, these studies add fine detail to our understanding of the molecular drivers
that control the lifetime of InhAꢀinhibitor complexes, laying the foundation for the rational
design of compounds with increased residence time through both ground state stabilization
and transition state destabilization.
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 48
1
2
3
4
Results and Discussion
5
6
7
8
Generation of a click chemistry library to interrogate structure-kinetic relationships
The triazole is a privileged pharmacophore, found in the scaffolds of a number of
drug discovery programs in addition to structures of marketed antibacterial compounds.25,26
The triazole group is normally introduced using click chemistry, and thus triazoleꢀbased
compounds are often the product of analoging efforts aimed at generating structureꢀactivity
relationship (SAR) data.25,27 To generate robust SKR data, we used click chemistry to
synthesize 14 novel analogs of the triazole diphenyl ether PT163 (Figure 1). The synthetic
route and synthetic procedures/characterization data for all these compounds can be found in
the methods section and in the supporting information, respectively.
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
areas modulating tR
O
X
OH
A
X
B
X = - CH3
- CN
2'
O
O
Cu-catalyzed
N
N
N
N
- Cl
R
R
+
N3
click chemistry
5
OH
OH
OH
O
O
O
N
N
N
N
N
N
N
N
PT501
OH
PT510
O
PT163
O
CN
Cl
OH
CN
CN
CN
OH
OH
O
N
N
N
N
N
N
N
N
N
N
N
PT506
OH
PT509
O
PT503
O
OH
O
N
N
N
N
N
N
N
PT504
O
PT507
OH
Cl
OH
PT502
N
N
O
N
N
N
Br
phenyl-substituted,
hydrophobic
N
PT513
OH
PT508
CN
OH
flexible, hydrophobic
O
O
N
N
N
N
N
O
N
PT512
O
PT500
OH
NH
rigid, polar
N
N
N
PT514
O
OH
CN
N
N
rigid, bulky,
hydrophobic
N
PT511
6
ACS Paragon Plus Environment

Page 7 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Figure 1. Triazoleꢀbased diphenyl ethers. Circled areas on the diphenyl ether scaffold
highlight the positions that modulate residence time. The Aꢀring and Bꢀring of the diphenyl
ether scaffold are indicated with the 5ꢀ and 2’ꢀ positions labeled. Categories of the different
Aꢀring sideꢀchains are represented in the boxes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The inhibition of InhA by the diphenyl ethers was analyzed by a substrateꢀbased assay
and found to exhibit slowꢀonset inhibition characteristic of many previous enoylꢀACP
reductase inhibitors.18 Details of the global fitting analysis are described in the methods
section. Briefly, the progress curve data were globally fit to the equations which describe a
twoꢀstep inducedꢀfit mechanism (Scheme 1; see Methods Section), providing the
thermodynamic (Kiapp and Ki*app) and kinetic (k6) constants for enzyme inhibition (Table 1).
The values for Kiapp, Ki*app and k6 were then used to obtain the rate constant for conversion of
EI to EI* (k5) and the overall kon for formation of EI* (kon,overall). In a separate experiment the
value of koff was also determined for selected inhibitors using a direct dissociation method for
comparison with the values of k6 obtained by global fitting of the progress curve data.28
These studies revealed that the triazoleꢀdiphenyl ethers had overall association rate constants
(kon,overall) that were 1ꢀ3 orders of magnitude slower than the association rates of the alkylꢀ
diphenyl ethers. In addition, whereas values for k5, the rate constant for conversion of EI to
EI* varied by ~100ꢀfold cross all compounds, Kiapp, the dissociation constant for the initial EI
complex varied by ~4,000 fold, indicating that variation in inhibitor structure had a more
profound impact on the barrier leading to formation of EI than on the barrier between EI and
EI*.
The triazole series can be separated into three groups based on the nature of the 2’ꢀ Bꢀ
ring substituent, which is a methyl (ꢀCH3), chloro (ꢀCl) or cyano (ꢀCN) group (Figure 1). To
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 48
1
2
3
4
5
6
7
8
9
form a basis for comparing the effect of modulating the structure of the triazole analogs on
enzyme inhibition, kinetic data are also included in Table 1 for hexylꢀdiphenyl ethers with ꢀ
CH3 (PT70), ꢀCl (PT91) and ꢀCN (PT119) substituents at the Bꢀring 2’ꢀposition as well as
the fluoro (PT113) and bromo (PT92) analogs. Below we compare and contrast the impact
of replacing the alkyl group with the triazoleꢀbased substituents and rationalize the structureꢀ
kinetics trends within the triazoleꢀdiphenyl ether series.
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
InhA as an ideal model system for probing SKRs
The biologically active unit of InhA is a tetramer. Each monomer consists of a central
βꢀsheet on top of which the cofactor is bound and which is surrounded by three
each side. The substrateꢀbinding pocket is shielded by the substrateꢀbinding loop (SBL)
comprised of ꢀhelices 6 and 7 and the loop that connects these helices. The SBL has to
αꢀhelices on
α
open and close during the process of substrateꢀbinding in each catalytic cycle. Upon inhibitor
binding a similar behavior can be observed, and depending on the nature of the inhibitor,
different states of SBL closure are achieved.
InhA undergoes an isomerization event from an open (EI) to a closed (EI*) state upon
binding to slowꢀonset inhibitors. The relative populations of these 2 conformations is driven
by the residence time and depends on the stability of the final complex in addition to the
energy barrier at the transition state between the 2 states. Figure 2 compares the closed SBL
structure of InhA bound to PT91, representing EI*, with a more open SBL structure in which
InhA is bound to the C16ꢀNAC substrate analog.29 The conformation adopted in the closed
state does not lie on the reaction coordinate for substrate turnover since residues from
6 occupy part of the active site where the substrate would normally bind.23
αꢀhelix
8
ACS Paragon Plus Environment

Page 9 of 48
Journal of the American Chemical Society
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. Conformations of the SubstrateꢀBinding Loop (SBL).
The SBL, which includes αꢀhelix 6, the interconnecting loop, and αꢀhelix 7, adopts a closed
conformation in the structure of PT91 bound to InhA (shown in orange, PDB: 4OYR)
relative to that of the C16ꢀNAC substrate analogue, which binds in a more open conformation
(shown in red, PDB: 1BVR).
Crystal structures for 6 of the 14 novel triazoleꢀbased diphenyl ether inhibitors bound
to InhA have been solved. Triazoleꢀbased diphenyl ethers also occupy the closed
conformation and before developing SKRs, it was important to understand similarities and
differences between the alkylꢀ and triazoleꢀbased diphenyl ether compounds. The inhibitors
differ in their substitution patterns, containing either a ꢀCH3, ꢀCl or ꢀCN group at their 2'ꢀ
position, combined with either a 5ꢀhexyl or a 5ꢀtriazole substituent to which a cyclic acyl
chain is attached. The consequences of the 2’ꢀsubstituents will be discussed below. 5ꢀHexyl
substituted diphenyl ethers are generally involved in multiple hydrophobic interactions within
the substrateꢀbinding pocket. While the hexyl substituents form vanꢀderꢀWaals contacts with
F149, M199, I202 or V203, and L218, the protein–inhibitor interactions of the 5ꢀtriazoleꢀ
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 48
1
2
3
4
5
6
7
8
9
substituted compounds are more extensive and more diverse in nature. The triazole ring is in
close proximity to the aromatic rings of Y158 and F149, and also to M199 and I202 or V203,
depending on the conformational state of αꢀhelix 6. In addition, in most cases a hydrogen
bond is formed to either E219 or Q214. The size of the 5ꢀcycloalkyl ring results in a large
interaction surface within the substrateꢀbinding pocket of the protein, which is reflected by
numerous hydrophobic interactions with the residues of αꢀhelix 6 (I202 or V203) and αꢀhelix
7 (I215, L217, L218) in the SBL.
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
To interrogate structureꢀkinetics relationships, it was important to compare the
relative positions of αꢀhelix 6 and αꢀhelix 7 for a variety of InhAꢀbound conformations to
address how open or how closed the SBL is in the presence of a given compound. Figure 3
summarizes the data for the structures of the triazoleꢀdiphenyl ethers and compares the
conformational coordinates to other InhAꢀbound structures by presenting the location of all
crystal structures on the same conformational plot in which the SBL is defined by a shear and
a step torsion angle (see details in Supporting information).
Figure 3. Conformational space populated by the SBL in different InhAꢀligand
structures. The torsion angles δ1 and δ2 are used to define the conformational space
accessed by the SBL. Torsion angles have been taken from the Xꢀray structures of InhA in
10
ACS Paragon Plus Environment

Page 11 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
complex with the triazole and alkyl diphenyl ether inhibitors as well as NADH (PDB: 2AQ8)
and C16ꢀNAc (PDB: 1BVR). The lower right quadrant of the plot is populated by structures
with a more open conformation of αꢀhelix 6, whereas the upper left quadrant contains
structures with a more closed conformation of αꢀhelices 6 and 7. Rapid reversible inhibitors
populate the EI state (open conformation) and slowꢀonset inhibitors populate the EI* state
(closed conformation).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The SBL in complexes of InhA with the rapidꢀreversible inhibitor PT155, as well as
with NADH or the C16ꢀNAC substrate analog adopts a more open conformation and these
structures cluster towards the lower right quadrant of the conformational plot. In contrast,
torsion angles for the SBL in complexes formed by the slowꢀbinding compounds including
both the triazoleꢀ and alkylꢀ diphenyl ethers, cluster towards the upper left quadrant of the
plot since in these structures the SBL adopts a more closed conformation. In an attempt to
quantify the degree of conformational change as well as the rates associated with the binding
of inhibitors that occupy different positions on the conformational plot shown in Figure 3, the
rapidꢀreversible inhibitor PT155, as well as a slowꢀbinding alkylꢀ (PT119), and triazoleꢀ
diphenyl ether (PT504) were selected for additional analysis using the conformationally
sensitive technique secondꢀharmonic generation (SHG).30ꢀ34 SHG is a nonꢀlinear optical
technique that is highly sensitive to the orientation of secondꢀharmonic active molecules
tethered to a surface. In the experiments performed here we have labeled the protein with
SHGꢀactive dye and tethered the protein to a supported lipid bilayer surface on glass. The
magnitude of the SHG signal intensity depends on the angle of the dye relative to the surface
to which the protein is bound. Upon compound binding, changes in the angular orientation of
the dye result in a change in the SHG intensity. The magnitude of the SHG signal change is
dependent on the overall structural conformational change upon ligand binding such that a
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 48
1
2
3
4
5
6
7
8
9
larger conformational change is expected to produce a larger signal change for a given
assay.31,32 The SHG signal is proportional to the sixth power of the cosine of θ, the angle
between the axis of the hyperpolarizability element of the dye and the surface normal, and
thus the SHG technique has high sensitivity to angular change. In cases where all the dye
molecules (and proteins) are aligned with a very narrow orientational distribution, i.e. all
pointing in the same orientation, SHG can detect an angular change of ~1° at the label site
given the signal to noise of the measurement.32
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
To enable SHG studies on InhA, an amine reactive secondꢀharmonic dye was
covalently attached to the lysine residues of InhA using standard amideꢀbond coupling
chemistry. The degree of labeling of InhA was determined spectrophotometrically to be 0.74
dyes per monomer. Mass spectrometry was used to identify the position and extent of
labeling and revealed that four lysines were modified: K57 (59%), K118 (21%), K132 (7%),
and K181 (13%). As InhA is an obligate tetramer, the mass spectrometry results indicate the
protein is heterogeneously labeled at several residues throughout the core tetrameric
structure.
After immobilizing the labeled protein on the NiꢀNTA bilayer using an Nꢀterminal
Hisꢀtag, we next tested each compound for the ability to modulate the conformation of InhA
by measuring the change in the SHG intensity upon compound addition. Each compound
was added at 10 µM final concentration and the change in SHG intensity was measured over
a 30ꢀmin time course. As can be seen in Figure 4a, the addition of compound PT155 to Hisꢀ
tagged InhA resulted in an increase of 7.2 ± 0.6% in the SHG intensity of the labeled,
surfaceꢀtethered protein. The addition of compounds PT119 and PT504 also resulted in an
increase in the SHG intensity of 13.1± 0.8% and 19.7 ± 0.6%, respectively. A control vehicle
12
ACS Paragon Plus Environment

Page 13 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
injection was also performed and the change in SHG intensity over the time course was
measured as ꢀ2.6 ± 1.0%. Comparison of the change in SHG intensity indicated that the
response of each compound was significantly different than the vehicle control. In addition,
the data indicate that the change in SHG intensity upon injection of each compound was
significantly different compared to the other compounds in the series. The different signal
change results suggest that PT155, PT119, and PT504 all induce distinct conformations of
InhA upon binding. While binding of PT155 is accompanied by the smallest degree of
change relative to the control, binding of PT119 and PT504 is accompanied by a larger
degree of conformational change, which agrees with the structural data in Figure 3 where
PT155 occupies the most open conformation, followed by PT119 and lastly, PT504, which is
the most closed of the three selected compounds. Therefore, the SHG measurements of the
degree of conformational change of the protein in solution are consistent with the changes
observed in the crystal structures.
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)
(b)
Figure 4. Bindingꢀinduced conformational changes in InhA detected by SHG.
(a) Magnitude of the SHG signal change induced by ligand binding to InhA 30 min after
compound addition. (b) 30ꢀminute time course for approach to final conformational change
upon inhibitor binding. SHG intensity has been normalized for the addition of each
compound. The addition of each compound occurred shortly after the initial reading of the
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 48
1
2
3
4
5
6
7
8
SHG intensity at t = 0. PT119 corresponds to the red trace, PT155 to the green trace, and
PT504 to the blue trace. The vehicle only control trace is shown in black. For all
experiments N = 4 measurements were performed. Error bars = SEM.
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
In addition to the magnitude of the conformational change, the SHG data in Figure 4b
indicate that the time course of conformational change is distinct for compound PT155
compared to compounds PT119 and PT504. The addition of PT155 results in a stepꢀlike
increase in the SHG intensity of InhA, consistent with kinetic data indicating the compound
is a rapid binding, reversible inhibitor. This result is in contrast to the time courses obtained
for compounds PT119 and PT504, which require a longer time to reach a plateau compared
to PT155. This difference in the rate of conformational change shows that the kinetics of
PT119 and PT504 binding are substantially different than PT155. Finally, we simulated the
experimental data obtained from the SHG assay using the kinetic and thermodynamic
parameters obtained from progress curve analysis for PT119 and PT504 (Table 1), which
involved estimating the levels of E, EI, and EI* using 10 µM inhibitor and 1 µM InhA. This
approach was able to reproduce the SHG data (Figures S3ꢀS4), indicating that the SHG
technique provides an alternative method for the detection of slowꢀbinding inhibitors.
14
ACS Paragon Plus Environment

Page 15 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
Table 1: Chemical and Biological Data for the Diphenyl Ether Inhibitorsa
7
8
9
c
tR
clogPb
Ki*app
(nM)
k5
kon, overall
MIC
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
Compound
Kiapp (nM)
k6 (minꢀ1)
(min);
(ꢀg/mL)
(clogD)
(minꢀ1)
(Mꢀ1 minꢀ1)
(tR direct dissociation)
OH
O
R1
5
R1
3.13
1.56
6.97 (5.99)
6.41
210 ± 34
220 ±6
29 ± 4
27 ± 1
0.23
0.18
1.2 x 106
9.6 x 105
0.036 ± 0.010
0.026 ± 0.004
30±7; (40±4)
38±5; (31±5)
PT70
CH3
F
PT113
1.56
3.13
6.95
7.1
52 ± 25
120 ± 1
5.0 ± 2.0
34 ± 1
0.12
0.28
2.5 x 106
3.3 x 106
0.013 ± 0.005
0.11 ± 0.014
90±37; (148±13)
9.0±1; (42±7)
PT91
PT92
Cl
Br
10000 ±
2000
2.5
5.90 (5.47)
40 ± 1
5.0
5.0 x 105
0.020 ± 0.0040
50±0.4; (47±4)
PT119
CN
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 48
1
2
3
4
5
6
7
8
9
c
tR
clogPb
Ki*app
(nM)
k5
kon, overall
MIC
Compound
Kiapp (nM)
k6 (minꢀ1)
(min);
(ꢀg/mL)
(clogD)
(minꢀ1)
(Mꢀ1 minꢀ1)
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
(tR direct dissociation)
OH
R1
O
N
N
R2
N
R1
R2
O
12.5
2.2
1800000
N/Ad
N/Ad
N/Ad
RRe
N/Ad
PT500
CH3
HN
17000 ±
3000
0.78
6.25
3.6 (3.25)
5.8
70 ± 14
77 ± 15
1.2
7.4 x 104
1.9 x 105
0.005 ± 0.002
0.014 ± 0.011
190±10; (206±15)
55±8
PT501
PT502
CH3
CH3
Cl
1900 ± 500
6600 ± 300
0.34
6.25
5.5
100 ± 10
68 ± 9
0.56
1.1
8.6 x 104
6.9 x 104
0.008 ± 0.001
110±2; (97±12)
PT503a
PT504
CH3
Cl
16000 ±
10000
0.156
3.62
0.0047 ± 0.0011
220±49; (625±78)
16
ACS Paragon Plus Environment

Page 17 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
c
tR
clogPb
Ki*app
(nM)
k5
kon, overall
MIC
Compound
CN
Kiapp (nM)
k6 (minꢀ1)
(min);
(ꢀg/mL)
(clogD)
(minꢀ1)
(Mꢀ1 minꢀ1)
(tR direct dissociation)
3.98
99000 ±
16000
0.313
370 ± 10
1.4
1.4 x 104
0.0052 ± 0.0002
194 ± 3
PT506
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
(2.77)
28000 ±
7000
1060 ±
200
0.625
0.625
1.25
3.59
3.55
5.5
0.28
0.50
0.91
1.0 x 104
7.1 x 104
1.9 x 104
0.0108 ± 0.0002
0.0107 ± 0.002
0.0222 ± 0.0003
92±2
93±18
45±1
PT507
PT508
PT509
CN
CN
CN
Br
7100 ± 1400
150 ± 30
50000 ±
10000
1190 ±
200
110000 ±
20000
0.625
0.625
0.625
5.5
4.5
4.2
280 ± 60
190 ± 37
350 ± 70
10.5
0.27
2.23
9.5 x 104
4.5 x 104
2.7 x 104
0.0267 ± 0.0003
0.0084 ± 0.0017
0.0094 ± 0.0018
37±1
PT510
PT511
PT512
CH3
CN
CN
6300 ± 1300
119±23
106±21
83000 ±
17000
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 48
1
2
3
4
5
6
7
8
9
c
tR
clogPb
Ki*app
(nM)
k5
kon, overall
MIC
Compound
Kiapp (nM)
k6 (minꢀ1)
(min);
(ꢀg/mL)
(clogD)
(minꢀ1)
(Mꢀ1 minꢀ1)
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
(tR direct dissociation)
45000 ±
12000
298±
41
1.56
6.25
NDf
4.7
4.9
3.99
3.46
3.20
8.9 x 104
1.7 x 104
3.9 x 104
0.0266 ± 0.0047
0.0106 ± 0.002
0.0092 ± 0.002
38±7
PT513
PT514
PT163a
CH3
CH3
CH3
200000 ±
40000
610 ±
122
94±18
81000 ±
23000
3.95
233 ± 39
113±28; (114±18)
aAll kinetic assays performed at 20 °C except for two compounds noted where assays were performed at 25 °C. Errors are standard errors from global data analysis.
bclogP values were determined using ChemDraw and clogD values were calculated using ACDLabs Suite
cValues in parentheses are the residence times determined from the 32PꢀNAD+ direct dissociation method.
dN/A – not applicable
eRR – rapid reversible
fND – not determined
18
ACS Paragon Plus Environment

Page 19 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
Triazole-based diphenyl ethers modulate residence time by transition state
destabilization
7
8
9
Alterations in drugꢀtarget residence time occur through modulation in the
magnitude of the barrier between the ground and transition states on the binding
reaction coordinate. This can occur either by changes in the overall affinity of a drug
for a target and/or changes in the stability of the rate limiting transition state(s). An
increase in affinity will lead to an increase in residence time if stabilization of the
drugꢀtarget complex ground state is not accompanied by changes to the transition
state. However, there are examples where a change in residence time is not
accompanied by a corresponding increase in potency.35 In this case, changes in
kon,overall and k6 for a twoꢀstep inhibition mechanism can provide insight into explicit
effects on the rate limiting transition state for formation and breakdown of the final EI
complex (EI* for a twoꢀstep mechanism). Indeed, in the limiting scenario where an
increase in residence time occurs without a change in thermodynamic affinity, there
must be a corresponding decrease in the on rate for formation of the drugꢀtarget
complex. Thus, as noted by Schoop and Dey,36 a complete structureꢀkinetic analysis
must include both the on and off rates for drugꢀtarget binding. In this regard, the
contribution of kon,overall to changes in residence time and the overall affinity of the
drug for the target can be assessed using a twoꢀdimensional (2ꢀD) kinetic map. In
Figure 5a the onꢀ and offꢀ rate data are shown for both the triazole and alkylꢀdiphenyl
ethers where it can be seen that the two compound classes are clustered separately on
the 2ꢀD plot. In general, the triazole analogs have longer residence times than the
alkyl diphenyl ethers but bind either with similar potency or less tightly to the enzyme
– thus the residence times of the triazoles result from significantly slower on rates
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
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 48
1
2
3
4
5
6
(~100ꢀ1000ꢀfold slower) than the corresponding alkyl compounds, a result that can be
clearly visualized on the kinetic map.
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
25
20
15
10
5
24
20
16
12
8
(c)
(b)
(d)
ꢁꢁGTS
20
15
10
5
ꢁꢁGTS
ꢁꢁGTS
PT119
PT506
PT70
PT501
PT91
PT504
4
0
0
0
GS
ꢁꢁGGS
E + I
EI
ꢁꢁG
E + I
E + I
ꢁꢁGGS
EI*
EI
3
EI
3
-4
-5
EI*
5
EI*
-5
0
1
2
4
5
6
0
1
2
4
5
6
0
1
2
3
4
6
Reaction Coordinate
Reaction Coordinate
Reaction Coordinate
Figure 5. Kinetic map and free energy profiles for InhA inhibition
(a) 2ꢀD kinetic map of koff vs. kon with diagonal lines showing Ki*. Free energy
profiles comparing the inhibition of InhA by PT70 and PT501 (b), PT91 and PT504
(c), and PT119 and PT506 (d). This figure highlights the role of transition state
destabilization in increasing the residence time of the triazoleꢀdiphenyl ethers.
The kinetic map in Figure 5 indicates that the triazole substituents destabilize
the rateꢀlimiting transition state on the binding reaction coordinate relative to the
20
ACS Paragon Plus Environment

Page 21 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
corresponding alkyl analogs which impacts both the rate of formation and the rate of
breakdown of the EI* complex. In Figure 5b, the binding reaction coordinates for the
Bꢀring methyl compounds PT70 and PT501 are plotted using 4 ꢀM inhibitor as a
reference value. PT501 binds to InhA with ~ 2ꢀfold lower affinity than PT70 (Ki*app
70 and 29 nM, respectively) but with a residence time ~7ꢀfold longer than that of
PT70. These changes are accompanied by a 16ꢀfold decrease in the on rate for
PT501 (7.4 x 104 Mꢀ1minꢀ1 compared to 1.2 x 106 Mꢀ1minꢀ1 for PT70) and thus the
triazole substituent has destabilized the transition state by ~ 1.7 kcal/mol relative to
that of the corresponding alkyl diphenyl ether. Comparisons of PT91 and PT504 (Bꢀ
ring chloro), as well as PT119 to PT506 (Bꢀring cyano) show similar trends (Figure
5).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Analogs with a B-ring cyano group have a reduced affinity to InhA, which is
accompanied by slower on-rates
In Figure 6 we show an expanded kinetic map for the triazole diphenyl ethers.
Except for the PT511 (Bꢀring ꢀCN)/PT514 (Bꢀring ꢀCH3) pair, which have bulky
cyclohexyl Aꢀring substituents (discussed later), replacement of a Bꢀring ꢀCH3 group
with a ꢀCN substituent generally results in a reduction in the onꢀrate. This is observed
for analogs with cyclopropyl (PT506 (ꢀCN) and PT501 (ꢀCH3)), ethyl (PT509 (ꢀCN)
and PT510 (ꢀCH3)), and cyclopentyl (PT512 (ꢀCN) and PT513 (ꢀCH3)) substituents.
This trend is also observed within the alkyl diphenyl ether series, for example
comparison of PT119 (ꢀCN) and PT70 (ꢀCH3). The slower onꢀrates of these pairs of
compounds are coupled to comparable or weaker binding affinity, as shown in the
kinetic map. This indicates modest to no change in residence time between the 2’ꢀCN
and 2’ꢀCH3 analogs with similar Aꢀring tails. Figure 6 shows that the onꢀrate for
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 48
1
2
3
4
PT506 relative to PT501 is 5ꢀfold slower and that this lateral shift is accompanied by
5
6
no change in residence time.
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. 2ꢀD kinetic map for the triazole inhibitors. Analogs with 2’ꢀCN
substituents are shown in red and those with 2’ꢀCH3 substituents are in blue. All 2’ꢀ
CN analogs are shifted left relative to the 2’ꢀCH3 derivatives except for the
PT511/PT514 pair. 2’ꢀCN analogs have slower onꢀrates and weaker binding, causing
little to no change in residence time compared to the 2’ꢀCH3 analogs.
Analysis of the structural data reveals that the 2’ꢀCN group lacks the favorable
hydrophobic interactions that the 2’ꢀCl and 2’ꢀCH3 analogs make with the enzyme
(Figure 7). Specifically, in both triazoleꢀ and alkylꢀdiphenyl ether structures, the 2’ꢀ
substituent is accommodated in a mostly hydrophobic environment formed by G96,
A198 from αꢀhelix 6 and the ribose pyrophosphate portion of the cofactor.
22
ACS Paragon Plus Environment

Page 23 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
We speculate that the orientation of the 2’ꢀsubstituents may partly be
responsible for differences in timeꢀdependent inhibition between analogs with 2’ꢀCN
substituents compared to those with 2’ꢀCl or 2'ꢀCH3 substituents. Upon binding to the
enzyme, the 2'ꢀCN group has to move past the carbonyl oxygen of G96 to form a
hydrogen bond with the backbone NH group. Since the interaction with the partially
negatively charged carbonyl oxygen is unfavorable, this could contribute to the
energy barrier that has to be overcome upon binding and it may take more time for the
2'ꢀCN substituent to be properly positioned in the binding pocket and to reach its
final, energetically lowest state. Thus, it is likely that the lack of stabilizing vanꢀderꢀ
Waals contacts between A198 and G96 coupled with the electronic repulsion and only
modest hydrogenꢀbonding to the backbone of G96 account for slower binding of the
polar 2’ꢀCN substituent in the hydrophobic pocket compared to the 2’ꢀCl and 2’ꢀCH3
analogs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A comparison of the orientation of the 2’ꢀsubstituents in different InhAꢀ
diphenyl ether complex structures reveals significant differences, particularly between
the 2’ꢀCN in contrast to the 2’ꢀCl and 2'ꢀCH3 groups. The most striking difference
can be seen in comparing the distances between the 2’ꢀsubstituents and the Cα of
A198 (Figure 7). The 2’ꢀCl substituents of PT91 and PT504 and also the 2'ꢀCH3
groups of PT501 and PT514 tend to be oriented more towards A198, thereby forming
close vanꢀderꢀWaals contacts with the Cα of A198. Considering only subunits with
fully defined SBLs, vanꢀderꢀWaals distances ranging from 3.2 Å (PT504) to 3.5 Å
(PT91), and 3.2 Å to 3.5 Å within the InhAꢁNAD+ꢁPT501 structures can be observed.
In contrast, the 2’ꢀCN groups are oriented further away from A198 and located in
closer proximity to G96, although still not close enough to be in optimal hydrogenꢀ
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 48
1
2
3
4
5
6
7
8
9
bonding distance to the backbone NH of G96: this distance ranges from 3.1 Å in the
InhAꢁNAD+ꢁPT512 to 3.6 Å for PT506. In addition, an unfavorable interaction with
the carbonyl group of G96 may also be present. Interestingly, because of the tilted
arrangement of the 2'ꢀCN group in the bound state, the carbon atom of this group
approaches the Cα of A198 leading to a reduction in the distance to NH of G96, which
may lead to an additional vanꢀderꢀWaals contact with the Cα of A198 and thereby be
beneficial for SBL stability. Our structural data therefore suggest that hydrophobic 2'ꢀ
substituents, especially the smaller 2’ꢀCl group or 2’ꢀCH3 group, are preferable in this
position since they permit a closer approach to A198 allowing the SBL to adopt a
more closed, stable state. In contrast, the more polar 2’ꢀCN substituent forms subꢀ
optimal interactions with A198, while the potentially favorable hydrogen bond to the
NH group of G96 is likely offset by unfavorable interactions with the G96 carbonyl
group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
24
ACS Paragon Plus Environment

Page 25 of 48
Journal of the American Chemical Society
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 7. Interaction of the 2’ꢀsubstituent with the enzyme
The 2'ꢀsubstituents interact with G96 and A198 in the InhA substrateꢀbinding pocket.
(a) PT504, (b) PT501, (c) PT514, (d) PT506, (e) PT512, (f) PT511. The –Cl group
and –CH3 groups of PT501, PT504 and PT514 form vanꢀderꢀWaals interactions with
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 48
1
2
3
4
5
6
7
8
9
A198. The ꢀCN group in PT506, PT511 and PT512 lacks favorable contacts with
A198 and could form a weak hydrogen bond with the backbone NH of G96 but which
also could be accompanied by unfavorable interaction with the backbone carbonyl of
G96.
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
Structural features unique to PT514 suggest a steric limit is reached in the active
site
While the 2’ꢀCH3 analogs have faster onꢀrates and cluster to the right of the 2ꢀ
D kinetic map compared with the 2’ꢀCN analogs with the same Aꢀrings, PT514
appears to be an outlier (Figure 6). PT514 has a 4ꢀfold slower onꢀrate compared with
either PT504 or PT501 (6.9 x 104 Mꢀ1minꢀ1 vs. 7.4 x 104 Mꢀ1minꢀ1). As shown in
Figure 8, the triazole ring can hydrogen bond with either E219 or Q214. Below we
discuss the positions of surrounding residues on both αꢀhelix 6 and αꢀhelix 7 in an
attempt to elucidate the differences in the molecular interactions among the triazoleꢀ
based inhibitors.
26
ACS Paragon Plus Environment

Page 27 of 48
Journal of the American Chemical Society
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 8. Interaction of αꢀhelix 7 with the 5ꢀsubstituent
(a) Hydrogenꢀbonding interaction of PT504 (orange) and PT506 (magenta) with
E219 and the inwardꢀrotated I215; (b) Hydrogenꢀbonding interaction of PT501
(ruby), PT511 (cyan), PT512 (blue) and PT514 (green) with Q214 and I215 pointing
out of the hydrophobic pocket; (c) Replacement of the hydrogenꢀbond with Q214 by
inwardꢀrotated I215; PT501 (ruby), PT511 (cyan), PT512 (blue) compared to the
PT91ꢀconformation (yellow).
In the structures of PT504 and PT506, the sideꢀchain of I215 is oriented
toward the inhibitor and an additional hydrogen bond is formed between the triazole
moiety and E219 (Figure 8). This buried state of I215 seems to be entropically
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 48
1
2
3
4
5
6
7
8
9
favored, as this hydrophobic residue is shielded from solvent. The inwardꢀrotation of
I215 in combination with the hydrogenꢀbond formed to E219 might have a positive
synergistic effect on the residence times of these inhibitors and seems to represent the
energetically most favorable SBL conformation among the InhAꢀinhibitor complexes
discussed here. However, this phenomenon only seems to be possible for the
inhibitors PT504 and PT506, which possess a more flexible and less bulky
cyclopropane substituent combined with either a 2'ꢀCl or ꢀCN group. The residual
space in the substrateꢀbinding pocket upon binding PT504 or PT506, presumably
creates a higher degree of rotational freedom for αꢀhelix 7 allowing the side chain of
I215 to be accommodated and the side chain of E219 to reach into the substrateꢀ
binding pocket and form a hydrogenꢀbond.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In contrast, the more bulky substituent in PT501 fills the binding pocket and
makes a simultaneous inwardꢀrotation of I215 and hydrogenꢀbond formation with
E219 impossible. This leads to a change of the hydrogenꢀbond donor from E219 to
Q214 for larger inhibitors, including PT501 as well as PT511, PT512 and PT514
(Figure 8). Close examination of each monomeric unit in the crystal structures
reveals that αꢀhelix 7 in the InhA complexes formed with PT501, PT511 and PT512
can also adopt a conformation similar to complexes formed by the alkyl diphenyl
ethers in which I215 is oriented into the substrateꢀbinding pocket but in which there is
no hydrogenꢀbonding interaction with either E219 or Q214 (Figure 8c).
We propose that to reach a conformational state where I215 is oriented
towards the hydrophobic pocket, a rotation of αꢀhelix 7 may have to take place, where
Q214 is moved upwards with respect to the substrateꢀbinding pocket (Figure 9).
28
ACS Paragon Plus Environment

Page 29 of 48
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
During this rotation, Q214 would have to pass the open space between the 5ꢀ
cycloalkyl substituent and either V203 or V207, depending on the extent to which αꢀ
helix 6 has closed. Consequently, the inward orientation of I215 is only possible
within certain limits defined by the bulkiness of the 2'ꢀ and 5ꢀsubstituents. In contrast
to the PT501, PT511 and PT512 InhAꢀinhibitor complexes, the structure of PT514
bound to InhA does not display the “in”ꢀstate of I215 in any of the subunits,
suggesting that the combination of a 2'ꢀCH3 group with a 5ꢀcyclohexane ring exceeds
the ‘steric limit’ that would otherwise permit αꢀhelix 7 to rotate.
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 9. Proposed mechanism for αꢀhelix 7 closure that prevents inward
rotation of I215 upon binding of inhibitors that reach a “steric demand” limit.
(a) View onto αꢀhelix 7 (α7) along the rotation axis (depicted by a black dot); (b)
View onto α7 perpendicular to the rotation axis; From the initial state (residues
shown in red) Q214 has to pass the open space between the inhibitor's 5ꢀcycloalkyl
moiety and either V203 (b, purple, representing a more open conformation of α6) or
L207 (b, magenta, representing a more closed conformation of α6) to reach the final
state (residues shown in orange) where I215 is oriented into the hydrophobic pocket.
The course of rotation is depicted by an arrow, an exemplary inhibitor is shown in
blue.
29
ACS Paragon Plus Environment