Comprehensive Synthesis of Amino Acid-Derived Thiazole Peptidomimetic Analogues to Understand the Enigmatic Drug/Substrate-Binding Site of P-Glycoprotein

Article abstract of DOI:10.1021/acs.jmedchem.7b01340

A novel set of 64 analogues based on our lead compound 1 was designed and synthesized with an initial objective of understanding the structural requirements of ligands binding to a highly perplexing substrate-binding site of P-glycoprotein (P-gp) and their effect on modulating the ATPase function of the efflux pump. Compound 1, a stimulator of P-gp ATPase activity, was transformed to ATPase inhibitory compounds 39, 53, and 109. The ATPase inhibition by these compounds was predominantly contributed by the presence of a cyclohexyl group in lieu of the 2-aminobenzophenone moiety of 1. The 4,4-difluorocyclohexyl analogues, 53 and 109, inhibited the photolabeling by [¹²⁵I]-IAAP, with IC₅₀ values of 0.1 and 0.76 μM, respectively. Selected compounds were shown to reverse paclitaxel resistance in HEK293 cells overexpressing P-gp and were selective toward P-gp over CYP3A4. Induced-fit docking highlighted a plausible binding pattern of inhibitory compounds in the putative-binding pocket of P-gp. The current study underscores the stringent requirement by P-gp to bind to chemically similar molecules.

Full text of DOI:10.1021/acs.jmedchem.7b01340

Subscriber access provided by READING UNIV
Article
Comprehensive Synthesis of Amino Acid-derived
Thiazole Peptidomimetic Analogues to Understand
Enigmatic Drug/Substrate-Binding Site of P-glycoprotein
Bhargav A. Patel, Biebele Abel, Anna Maria Barbuti, Uday Kiran
Velagapudi, Zhe-sheng Chen, Suresh V. Ambudkar, and Tanaji T Talele
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 18 Dec 2017
Downloaded from http://pubs.acs.org on December 18, 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 Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Comprehensive Synthesis of Amino Acidꢀderived
Thiazole Peptidomimetic Analogues to Understand
Enigmatic Drug/SubstrateꢀBinding Site of Pꢀ
glycoprotein
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
,§
‡
†
†
Bhargav A. Patel , Biebele Abel , Anna Maria Barbuti , Uday Kiran Velagapudi , ZheꢀSheng
†
*,‡
*,†
Chen , Suresh V. Ambudkar , Tanaji T. Talele
†
Department of Pharmaceutical Sciences, College of Pharmacy and Health Sciences, St. John's
University, Queens, NY 11439, USA
‡
Laboratory of Cell Biology, Center for Cancer Research, National Cancer Institute, National
Institutes of Health, Bethesda, MD 20892, USA
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 108
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: A novel set of 64 analogues based on our lead compound 1 was designed and
synthesized with an initial objective of understanding the structural requirements of ligands
binding to a highly perplexing substrateꢀbinding site of Pꢀgp and their effect on modulating the
ATPase function of the efflux pump. Compound 1, a stimulator of Pꢀgp ATPase activity, was
transformed to ATPase inhibitory compounds 39, 53 and 109. The ATPase inhibition by these
compounds was predominantly contributed by the presence of a cyclohexyl group in lieu of the
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
ꢀaminobenzophenone moiety of 1. The 4,4ꢀdifluorocyclohexyl analogues, 53 and 109, inhibited
125
the photolabeling by [ I]ꢀIAAP, with IC values of 0.1 and 0.76 ꢀM, respectively. Selected
5
0
compounds were shown to reverse paclitaxel resistance in HEK293 cells overexpressing Pꢀgp
and were selective toward Pꢀgp over CYP3A4. Inducedꢀfit docking highlighted a plausible
binding pattern of inhibitory compounds in the putativeꢀbinding pocket of Pꢀgp. The current
study underscores the stringent requirement by Pꢀgp to bind to chemically similar molecules.
INTRODUCTION
Failure of chemotherapy to treat various types of cancer is often linked to the phenomenon of
multidrug resistance (MDR), wherein the cancer becomes resistant to structurally and
1
functionally diverse classes of drugs. Amongst the several mechanisms responsible for MDR
phenotype, such as decline in drug uptake, alterations in cell cycle progression and drug
metabolism, augmented DNA damage repair, and decreased apoptosis, the increase in active
efflux of a drug by ATPꢀbinding cassette (ABC) transporters appears as a major contributing
2
factor for acquired resistance. Amongst the superfamily of ABC transporters which cause MDR,
3
Pꢀglycoprotein (Pꢀgp/ABCB1/MDR1) has been considered the most prevalent.
Human Pꢀgp is a 1280 amino acid long membrane protein folded into two similar halves. Each
half consist of a nucleotideꢀbinding domain (NBD) and a transmembrane domain (TMD). Each
2
ACS Paragon Plus Environment

Page 3 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
TMD can further be divided into six transmembrane (TM) helices. Pꢀgp, in normal tissues,
functions to protect cells from xenobiotic insults. However, MDR cells develop an
overexpression of Pꢀgp, which leads to the efflux of amphipathic substrate chemotherapeutic
4
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
drugs out of the cell, limiting the ability of these drugs to kill the cancer cells. Pꢀgp has a broad
substrate selectivity ranging from small molecules to macrocycles with diverse chemical
5
structures and high lipophilicity. The substrate/drugꢀbinding pocket, located in the TMDs, can
6
, 7
accommodate two to three molecules simultaneously. Moreover, the same substrate can bind
8
to multiple regions within the hydrophobic binding cavity of Pꢀgp. Powered with the energy
from ATP binding/hydrolysis process at the NBDs, Pꢀgp operates via an alternating access
mechanism for the efflux of substrates. According to this model, a conformational change occurs
upon binding of the drug to Pꢀgp in an inward open state that brings the NBDs together forming
an inward closed conformation. Drug translocation is induced during the closed conformation of
Pꢀgp, followed by the release of drug on the extracellular side where the efflux pump acquires an
outward open conformation. Subsequently, ATP hydrolysis at the NBD facilitates Pꢀgp to attain
9
its original drugꢀbinding competent inward open state conformation. Recent cryoꢀEM analyses
1
0
of the catalytic cycle of Pꢀgp supports these open and closed conformational states of Pꢀgp.
Extensive research has been documented in the development of MDR reversal compounds,
11
leading to three generations of Pꢀgp modulators over the years; however, these modulators have
proved ineffective in clinical trials due to wide range of reasons. These failures were partly
ascribed to insufficient bioavailability at tumor sites, nonꢀspecific toxicities, indiscriminate
inhibition of multiple ABC transporters, including Pꢀgp in the intestine, liver, kidney, and the
1
2ꢀ14
bloodꢀbrain barrier as well as one of the major drug metabolizing enzymes, CYP3A4.
Moreover, certain concerns have been raised regarding the relevant selection of the patient
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 108
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
population in clinical investigation of the Pꢀgp modulators. The criteria for patient selection
should consider the causal factor for MDR, including the specifically overexpressed ABC
transporter for which the modulator was developed. Additionally, single nucleotide
polymorphisms (SNPs) in these efflux pumps should be determined during patient selection
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
15
process. Accompanying conditions in patients such as inflammatory diseases may also have
1
6, 17
significant effect on the treatment outcomes.
Despite these issues, the key role played by
these efflux transporters in MDR development warrants further research for ultimate clinical
success.
Development of new classes of Pꢀgp inhibitors, devoid of the limitations associated with
previously disclosed compounds, is a significant need in oncologic settings for the treatment of
resistant cancers. Consequently, new approaches are devised to develop MDR reversal
1
8
compounds with high Pꢀgp affinity and selectivity. Using one such approach, we incorporated
chemical moieties that are often present in Pꢀgp modulators to develop a lead compound, (S)ꢀNꢀ
(2ꢀbenzoylphenyl)ꢀ2ꢀ(2ꢀmethylꢀ1ꢀ(3,4,5ꢀtrimethoxybenzamido)propyl)thiazoleꢀ4ꢀcarboxamide
(TTTꢀ28; 1), which features 3,4,5ꢀtrimethoxybenzoyl and 2ꢀaminobenzophenone moieties at the
1
9
Nꢀ and Cꢀtermini, respectively, of the (S)ꢀvalineꢀderived thiazole scaffold. Compound 1
increased intracellular paclitaxel concentration in SW620/Ad300 drugꢀresistant cell lines and
demonstrated a significant reversal of resistance to paclitaxel, doxorubicin and vincristine at a
concentration of 10 ꢀM in vitro. It also had preferential selectivity toward Pꢀgp than other tested
ABC transporters. Moreover, it has shown desirable profile in vivo as evidenced from reduced
20
tumor volume and tumor weight with no apparent sideꢀeffects in mice. Additionally,
125
biochemical assays such as [ I]ꢀIAAP and ATP hydrolysis indicated compound 1 to interact
4
ACS Paragon Plus Environment

Page 5 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with the TMDs of Pꢀgp at the drugꢀbinding site with a potent stimulation of ATPase catalytic
1
9
activity.
One of the approaches to determine the direct interaction of test compounds with Pꢀgp is to
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
measure their effect on basal ATPase activity. Pꢀgp utilizes ATP hydrolysis as a mechanism to
21, 22
energize its efflux function.
The membrane vesicles prepared from cells expressing Pꢀgp
exhibit basal ATPase activity, which may have resulted from either endogenous substrates or
2
3, 24
uncoupled ATPase activity.
Compounds which are Pꢀgp substrates and likely to be
transported usually are stimulators of basal Pꢀgp ATPase activity. Alternatively, these
compounds could competitively inhibit transport of other drug substrates. Inhibitors of basal and
substrateꢀstimulated ATPase activity tend to be incompetent for transport by Pꢀgp, yet capable of
averting the transport of other substrate drugs. Moreover, inhibitors of ATP hydrolysis by Pꢀgp
are shown to act for a longer duration of time in vitro than their stimulatory counterparts to
2
5ꢀ27
reverse the multiꢀdrug resistance.
Compounds lacking interactions with Pꢀgp are devoid of
both stimulatory and inhibitory effects on basal ATPase activity. The affinity and/or the
reversing ability of compounds, within the same class of ATPase modulators, can be better
correlated with the concentrations required for half maximal stimulation (EC ) or inhibition
50
2
3
(
IC ) of the Pꢀgp ATPase activity than their extent of stimulation or inhibition. Most Pꢀgp
50
modulators stimulate ATPase activity, except for a few clinically tested inhibitors such as
28, 29
tariquidar, elacridar and zosuquidar.
These clinical candidates potently inhibited ATPase
activity, which is an indication of their high binding affinity toward Pꢀgp. In addition, siteꢀ
directed mutagenesis experiments have shown that the loss of ATPase inhibitory activity for
these clinical candidates also resulted in loss of their ability to reverse efflux of a fluorescent
substrate (NBDꢀcyclosporine A) by Pꢀgp. This correlation implies the significance of ATPase
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 108
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
29
inhibition for efficient reversal of efflux function. The absence of high resolution human Pꢀgp
crystal structure and known promiscuity of the drugꢀbinding site i.e., the presence of multiple
ligandꢀbinding pockets, poses significant difficulties in elucidating common chemical
pharmacophore/groups for efficient binding. With this backdrop, we pursued a twoꢀpronged
approach that involved: 1) transformation of an ATPase stimulator 1 into a compound with
inhibitory action on ATPase activity and 2) probing the transport substrateꢀbinding sites of Pꢀgp
through a multitude of analogue syntheses. Implementation of these approaches led to medicinal
chemistry optimization program that entailed modifications at the: 1) Cꢀterminus, 2) Nꢀterminus
and 3) central region (αꢀcarbon and carboxamide groups) of the lead compound 1 (Figure 1).
Published data on Pꢀgp modulators has firmly established how the substrateꢀbinding region of
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, 30
Pꢀgp distinguishes various structural classes of compounds on the basis of varied affinity.
However, the present study reveals that moderate structural changes, including strictly classical
and nonꢀclassical isosteric modifications in a particular series of compounds, can impact Pꢀgp
function to produce no effect, and induce stimulation or inhibition of ATPase activity. Therefore,
structureꢀactivity relationship (SAR) investigations pertaining to our previously published lead
stimulator (1) of ATPase function of Pꢀgp led to the comprehensive synthesis of 64 target
compounds with varying impact on ATPase activity. This report highlights the rigid requirement
of Pꢀgp relevant to binding to chemical analogues belonging to the same chemical series.
Moreover, this study also provides an insight for surmounting the Pꢀgp efflux problem associated
with the substrate behavior of many drugs and clinical development candidates by conducting
subtle structural modifications in the lead candidate structures. According to February 2012 FDA
guidance for industry, all investigational drugs should be tested in vitro to determine whether
they are substrates of ABC transporters. Consequently, our SAR study may provide tools for
6
ACS Paragon Plus Environment

Page 7 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
converting a Pꢀgp substrate structure into a compound with no affinity toward Pꢀgp. Further
inducedꢀfit docking of representative ATPase inhibitory target compounds was conducted within
the drug/substrateꢀbinding pocket of homology modeled structure of human Pꢀgp to analyze the
observed SAR trends.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
RESULTS AND DISCUSSION
Chemistry. Altogether, a set of 64 (S)ꢀamino acidꢀderived thiazole target compounds were
synthesized to assess their interactions with Pꢀgp using ATPase activity assay. The SAR study
was instigated based on our lead compound 1, which is a (S)ꢀvaline thiazoleꢀderived Pꢀgp
3
1
19
ATPase stimulator. Compounds 2 and 6 required for the synthesis of modifications of 1 were
synthesized as per our previous reports. Scheme 1 shows the synthesis of target compounds 5
and 7. The swapped analogue 5 was obtained over three steps: i) coupling of the acid 2 with the
3
,4,5ꢀtrimethoxyaniline using HOBt, HCTU and DIEA; ii) Boc deprotection of the coupled
product 3 using TFA and iii) subsequent coupling of generated free amine 4 with the 2ꢀ
benzoylbenzoic acid. Compound 7 was synthesized by coupling the acid 6 with the 3ꢀ
aminobenzophenone in the presence of HOBt, HCTU and DIEA.
Scheme 2 represents the synthesis of key intermediates required for coupling reactions to
obtain various Nꢀ and Cꢀterminal substitutions of (S)ꢀvalineꢀderived thiazole scaffold.
Anthranilic acid was converted to an amide 8 by first activating the carboxyl group with the
32
thionyl chloride and subsequent treatment with aniline. 2ꢀNitroaniline in pyridine was stirred
with the benzoyl chloride and subsequent palladiumꢀcatalyzed hydrogenation of the nitro group
33
gave the reverse amide 9. In the past, we observed low yields for coupling reaction with the 2ꢀ
aminobenzophenone, which may have resulted from the electron withdrawing nature of the
2
0
benzoyl group. To solve this problem, we reduced the carbonyl group to produce corresponding
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 108
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
benzhydrol intermediates, 10 and 11, with desired nucleophilicity of the 2ꢀamino group to
facilitate amide bond formation during the coupling reactions, using previously reported
2
0
procedure. Next, 3ꢀaminobenzoic acid was protected by reaction with Bocꢀanhydride in the
3
4
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
presence of triethylamine to furnish 12. The amino substituted benzoic acids were converted to
corresponding azido substituted benzoic acids (13ꢀ15) using diazotization reaction conditions
3
5
followed by sodium azide treatment. The 1ꢀbenzylpiperidinꢀ4ꢀamine, 17, was prepared over
two steps: i) benzylation of the piperidine ring nitrogen of a tertꢀbutyl piperidinꢀ4ꢀylꢀcarbamate
36
to provide 16 and ii) Boc deprotection.
The synthesis of isosteric analogues of benzophenone is depicted in Scheme 3. The
trimethoxybenzoylꢀcontaining acid 6 was reacted with several synthesized as well as
commercially available amines in the presence of HOBt, HCTU and DIEA to obtain final test
compounds 18ꢀ21.
The benzophenone analogues with Nꢀ and Cꢀtermini substitutions were synthesized according
to Scheme 4. Acid 2 was reacted with amino benzhydrol derivatives 10 and 11 to provide
coupled products 22 and 23, respectively, which were then subjected to DessꢀMartin periodinane
oxidation to produce 24 and 25 and subsequent Boc removal provided 26 and 27. Amine 27 was
reacted with 3,4,5ꢀtrimethoxybenzoyl chloride furnishing the 4’ꢀbromobenzophenone derivative
2
8. The 4’ꢀbromo was then converted to 4’ꢀazido derivative 29 using sodium azide in the
3
7
presence of proline and copper iodide. Compound 31 was prepared by coupling 26 with acid
intermediate 12 and subsequent Boc group removal. The Nꢀterminal azide compounds 32ꢀ34 and
3
5 were synthesized starting from amines 26 and 27, respectively, and reacting with appropriate
azide intermediates 13ꢀ15.
8
ACS Paragon Plus Environment

Page 9 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 5 illustrate the synthesis of alicyclic analogues 36ꢀ40. The acid 6 was coupled with
appropriate amines in the presence of HOBt, HCTU and DIEA coupling reagents.
The cyclohexyl and related analogues were synthesized according to Scheme 6. First,
commercially available Bocꢀprotected amines were coupled with the acid 6 to obtain compounds
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
1, 45, 47, and 49 followed by Boc deprotection to yield free amine derivatives 42, 46, 48, and
0. Compounds 43, 44, and 51ꢀ67 were prepared by reacting the acid 6 with various synthesized
5
and commercially available amines.
Variedly substituted Nꢀterminus analogues were prepared according to Scheme 7. The acid 2
was coupled with the 4,4ꢀdifluorocyclohexylamine and then Bocꢀdeprotected to the amine 77.
This amine was then coupled with commercially available carboxylic acids or acid chlorides
providing compounds 69ꢀ76 and 78ꢀ80.
The thioamide derivatives were synthesized by reacting 39 or 53 independently, with
Lawesson’s reagent as shown in Scheme 8. Both diꢀ and monoꢀthioamide derivatives were
1
obtained and characterized by HꢀNMR and LCMS analyses. Compounds 82 and 84 were
designated as thiazoleꢀ4ꢀthioamides based on the ‘NH’ chemical shift of the thioamide.
1
Comparison of HꢀNMR spectrums of 53 and 77 allowed us to assign cyclohexyl amide peak as
a doublet at ~8 ppm, and hence, the trimethoxybenzoyl amide ‘NH’ can be assigned as doublet at
~9 ppm (Supporting Information Figure S1). For monoꢀthioamide 84, a doublet at 8.9 ppm
suggests preservation of trimethoxybenzoyl amide oxygen and the absence of a doublet at ~8
ppm with appearance of a doublet at 10 ppm indicated occurrence of a thioamide conversion
near the 4,4ꢀdifluorocyclohexyl ring. Such chemical shift trend is also applicable to compounds
1
3
9 and 82. Moreover, HꢀNMR spectrums of 81 and 83 showed chemical shifts of both the
doublets at around ~9.8 and ~10.6 ppm.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 108
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
Valine residue in 53 was replaced with several amino acids as shown in Scheme 9. The
thiazole rings were constructed starting from the amino protected (S)ꢀamino acids by first
3
8
converting these to respective amides 85ꢀ89, further to thioamides 90ꢀ94 and subsequently, to
3
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
2
3
4
5
6
7
8
9
0
bisꢀprotected thiazoles 95ꢀ99. The thiazoles 95ꢀ99 were synthesized employing a procedure
39
where calcium carbonate was used to neutralize in situ generated hydrobromic acid. This
procedure was opted against cumbersome three step method used previously for (S)ꢀvalineꢀ
derived thiazole construction to improve yields, simplify purification process and achieve rapid
reaction output. However, the calcium carbonate method led to partial racemization at the chiral
center for thiazoles 96ꢀ99. This was inferred based on the smaller optical rotation values for
4
0
compounds 96 and 97 as compared to those reported by Bredenkamp and colleagues. The
racemization at the αꢀcarbon may have occurred via acid catalyzed imineꢀenamine transition
41ꢀ43
states during the final aromatization step.
Deprotection using TFA for thiazoles 95ꢀ98 or 5%
4
4
piperidine in DMF for Fmocꢀproline derivative 99 yielded respective amino derivatives.
Further commercially available ethyl 2ꢀaminoꢀthiazoleꢀ4ꢀcarboxylate and the synthesized amines
were each reacted with 3,4,5ꢀtrimethoxybenzoyl chloride to furnish 100 and 102ꢀ106. Respective
ethyl esters were then hydrolyzed and coupled with the 4,4ꢀdifluorocyclohexylamine using
amide coupling reagents (HCTU, HOBt and DIEA) to provide target compounds 101 and 107ꢀ
1
11.
StructureꢀATPase Activity Relationship. A novel set of 64 target compounds was
synthesized and tested in an ATPase assay (at different concentrations: 0.05 µM, 0.5 µM, and 2.5
µM) using native membranes (High Five insect cells) that express human Pꢀgp to determine their
affinity toward Pꢀgp. The biological readouts for these compounds are reported as percentage
(stimulation/inhibition) of Pꢀgp basal ATPase activity. In the following SAR discussion,
10
ACS Paragon Plus Environment

Page 11 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
percentage of Pꢀgp’s ATPase activity was reported as stimulation of the inherent/basal activity
unless otherwise specified. The ATPase activity can be interpreted as: 1) compounds stimulating
basal ATPase activity are considered as substrates; 2) no effect on basal ATPase activity means
potentially no interaction with Pꢀgp and 3) inhibition of basal ATPase activity indicates
inhibition of Pꢀgp catalytic activity and subsequent inhibition of its efflux function.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To analyze the SAR trends, we categorized target compounds into three clusters: 1)
modifications at the Cꢀterminus and 2) the Nꢀterminus of the (S)ꢀvalineꢀderived thiazole scaffold,
and 3) the αꢀcarbon of the valine amino acid, including bioisosteric replacement of the Cꢀ and Nꢀ
terminus carboxamide with the thiocarboxamide group. A majority of the synthesized
compounds showed stimulation of ATPase activity, which suggests these compounds to be
potential Pꢀgp substrates. The comparisons between ATPase stimulatory compounds were
described with respect to binding affinity. This description is based on the percentage stimulation
at the lowest identical concentration tested for the analogues being discussed i.e. higher the
stimulation (at the same concentration), higher the binding affinity. Some of the target
compounds inhibited ATPase activity, which indicated improved affinity toward Pꢀgp and
inhibition of its efflux function. Certain analogues exhibited neither stimulation nor inhibition of
ATPase activity suggesting the loss of their affinity toward Pꢀgp. In this work, we used highꢀ
affinity Pꢀgp inhibitor, zosuquidar, as a positive control.
1
9
Based on our previous molecular modeling studies, we sought to improve the binding
efficiency of our lead compound 1 through a complementary fit in the proposed Pꢀgp
substrate/drugꢀbinding pocket. Consequently, our initial efforts were mainly focused on
performing structural modifications on the existing Nꢀ and Cꢀterminal substituents of compound
1 and subsequently, to analyze their activity based on modulation of the basal ATPase activity of
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 108
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
Pꢀgp (Table 1). Firstly, compound 5 was synthesized with the same 3,4,5ꢀtrimethoxyphenyl and
benzophenone groups, as in lead compound 1, except their positions are switched from the Nꢀ to
the Cꢀterminus and viceꢀversa. This switch led to the loss of affinity to Pꢀgp as evidenced from
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
no activity at 0.05
.05 ꢁM, 83% at 0.5 ꢁM and 88% at 2.5 ꢁM) of basal ATPase activity. However, compound 5
displayed a moderate (46%) stimulation of basal ATPase activity at 0.5 M concentration, which
ꢀM compared to that of compound 1, which had high stimulation (54% at
0
ꢀ
indicated its substrate nature toward Pꢀgp. Based on this result, the next round of SAR efforts
retained the positions of the aryl moieties similar to those seen in compound 1. Prior to varying
the aryl substitutions, we decided to find an optimal positional isomer of the benzophenone
attachment to the Cꢀterminus. Since the 2ꢀ and 4ꢀaminobenzophenone modifications at the Cꢀ
1
9
terminus were previously assessed, we synthesized the only remaining 3ꢀaminobenzophenone
regioisomer 7, which showed a weak stimulation of ATPase activity (13% at 0.05 M). This data
ꢀ
suggests the key contribution of an optimally positioned Cꢀterminal 2ꢀbenzoyl moiety for affinity
toward Pꢀgp. Therefore, we retained an orthoꢀsubstituted arrangement for the benzophenone
moiety in the next array of structural modifications on the Cꢀterminus. Subsequently, we
explored the influence of classical and nonꢀclassical isosteric replacements of the carbonyl group
present in Cꢀterminal 2ꢀaminobenzophenone moiety. The replacement of the carbonyl bridge of
the benzophenone moiety with an amide (18, 21% at 0.05
M), ether (20, NA at 0.05 M) and methylene (21, 15% at 0.05
toward Pꢀgp compared to 1. The stimulation at concentration of 0.05
indicated substrate type interactions, albeit with low binding affinity as compared to compound
. Thereafter, a set of seven compounds was prepared representing modifications at the aryl
ꢀ
M), reverse amide (19, NA at 0.05
M) resulted in loss of affinity
M for these compounds
ꢀ
ꢀ
ꢀ
ꢀ
1
moieties of both Nꢀ and Cꢀtermini of compound 1. These compounds (28, 149% at 0.5 ꢀM; 29,
12
ACS Paragon Plus Environment

Page 13 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
48% at 0.5
ꢀ
M; 31, 308% at 0.5
ꢀM; 32, 142% at 0.5 ꢀM; 33, 122% at 0.5 ꢀM; 34, 135% at 0.5
ꢀ
M and 35, 99% at 0.5
ꢀM) showed high stimulation of ATPase activity. Collectively, this data
indicated the importance of the carbonyl group, in addition to the sharp angular shape produced
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
by an orthoꢀbenzophenone moiety, for potent ATPase stimulation. The azido analogues, 29 and
3
2ꢀ35 were primarily synthesized as potential photoprobes to map the residues in the binding
pocket at the TMDs of Pꢀgp. To our disappointment, we could not obtain any covalently bound
Pꢀgp adduct for compounds 1 (contains benzophenone as a photoprobe), and 29 and 32ꢀ35
(contains azide as a photoprobe). This may have resulted from a significant reduction in the
affinity of these compounds for Pꢀgp and/or degradation of the compounds upon exposure to UV
light, which is required for photo crosslinking. All the compounds synthesized up to this point
were either regioisomers, isosteres or modifications of the Nꢀ and Cꢀterminal aryl substituents
and showed ATPase stimulatory effect. Therefore, we may consider these analogues as potential
Pꢀgp substrates and, hence, by a mechanism similar to the resistant cancer drugs, might
eventually undergo expulsion by the efflux pump. Based on the above data, we hypothesize that
the respective aryl moieties at both Nꢀ and Cꢀtermini are involved in interactions within the
hydrophobic substrateꢀbinding site of Pꢀgp, which causes adaptation of the Pꢀgp conformation
that results in stimulation of basal ATPase activity (vide infra). This observation prompted us to
explore nonꢀaromatic groups with a goal to obtain Pꢀgp ATPase inhibitors. Toward this goal,
first we decided to determine the impact of replacing the Cꢀterminal benzophenone moiety with
various alicyclic ring systems. Compounds 36ꢀ40 with saturated aliphatic cycles of increasing
steric bulk and lipophilicity were synthesized. The ATPase assay showed either no activity or
low stimulation of basal ATPase activity for four of the five compounds (36, NA at 0.5
ꢀM; 37,
2
5% at 0.5 M and NA at 2.5 M; 38, NA at 0.5 M; and 40, NA at 0.5 M), an indication of a
ꢀ
ꢀ
ꢀ
ꢀ
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 108
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
complete loss of affinity toward Pꢀgp. Encouragingly, compound 39, a Cꢀterminal cyclohexyl
analogue, showed 20% inhibition of basal ATPase activity at 0.05 ꢀM concentration. Thus, the
first Pꢀgp ATPase inhibitor of the series was realized. It is interesting to find that an increase in
the threeꢀdimensionality with reduction in steric bulk in compound 39 as compared to lead 1
resulted in increased affinity toward Pꢀgp with an inhibition of the catalytic process. Pꢀgp
ATPase inhibitory activity, although weak, of compound 39 with insertion of a cyclohexyl ring
in lieu of the 2ꢀaminobenzophenone moiety in compound 1 prompted us to investigate additional
saturated ring systems at the Cꢀterminus.
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
Since smaller (cyclopropyl, cyclobutyl, and cyclopentyl) or larger (adamantyl) ring size than
the cyclohexyl ring proved detrimental, we focused our attention on inserting six membered
rings at the Cꢀterminus of the (S)ꢀvalineꢀderived thiazole scaffold. In our previous study, we
found that hydrogen bonding interactions in the drugꢀinteracting site of Pꢀgp played a critical
2
9
role in offering high affinity to clinical candidates. With this notion, we synthesized hydrogen
bond donorꢀ and/or acceptor groupꢀcontaining cyclohexyl analogues to maximize interaction
probability in the drug/substrateꢀbinding pocket of Pꢀgp (Table 2). A tertꢀbutyl carbamateꢀ
protected piperidineꢀ4ꢀyl analogue, 41, exhibited a stimulatory ATPase activity as evidenced
from 48% stimulation at 2.5
ꢀM. Isosteric replacement of the cyclohexyl ring with the piperidine
ring (42, 15% at 2.5 M) and pyran ring (43, NA at 2.5
ꢀ
ꢀM) showed complete loss of their
ability to inhibit ATPase activity. Additional set of compounds, wherein we replaced the
cyclohexyl amine moiety of 39 with various saturated sixꢀmembered ring systems such as 4ꢀ
piperidone (44, NA at 2.5
48, NA at 2.5 M), 4ꢀaminocyclohexyl methylamine (50, NA at 2.5
,1ꢀdioxide (51, NA at 2.5 M), had no inhibitory effect on the ATPase activity. This data has
ꢀM), piperidinꢀ4ꢀmethylamine (46, NA at 2.5
ꢀM), 4ꢀaminopiperidine
(
ꢀ
ꢀM) and 4ꢀaminothiopyranꢀ
1
ꢀ
14
ACS Paragon Plus Environment

Page 15 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
prompted us to revisit the cyclohexyl ring pattern as in compound 39 with the intent of bringing
in substitutions such as fluoro, methyl and keto at the 4ꢀposition of the cyclohexyl ring. These
efforts led to synthesis of cyclohexanꢀ4ꢀone analogue 52 (NA at 2.5
derivative 53 (34% inhibition at 2.5 M) and 4,4ꢀdimethylcyclohexyl analogue 54 (34% at 2.5
M). This piece of SAR resulted in identification of 53 with ATPase inhibitory potential that
ꢀM), 4,4ꢀdifluorocyclohexyl
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
ꢀ
was higher than that of the first ATPase inhibitory compound, 39, of the series. It is worth
mentioning that Pꢀgp can discriminate between the 4,4ꢀdifluorocyclohexyl (53) versus 4,4ꢀ
dimethylcyclohexyl (54) isosteric analogues for binding preference and inhibition versus
stimulation of the basal ATPase activity. The bioactivities displayed by the alicyclic ring
analogues imply that the cyclohexane portion of the molecule may fit favorably in the substrateꢀ
binding pocket of Pꢀgp with nonꢀpolar residues around it. To probe the steric requirements of the
substrateꢀbinding pocket of Pꢀgp in a cyclohexyl interacting region, we prepared a small set of
compounds with varying steric bulk such as dicyclohexyl amine (55, 134% at 2.5
piperidinyl piperidine (56, 47% at 2.5 M), Nꢀbenzyl piperidineꢀ4ꢀamine (57, 154% at 2.5
pyrimidinyl piperidineꢀ4ꢀamine (58, 83% at 2.5
59, 123% at 2.5 M and 60, 31% at 2.5
ꢀ
M),
M),
ꢀ
ꢀ
ꢀ
M), and two enantiomers of tetralane amines
(
ꢀ
ꢀ
M). The difference in the activities of
diastereoisomers, 59 and 60, suggested chiral discrimination by the Cꢀterminus fragment binding
region of Pꢀgp. Moderate to high stimulation of ATPase activity by these analogues suggested no
tolerance for bulky substituents around the cyclohexyl binding site with respect to inhibition of
ATPase activity, which reinforces the existence of a sterically forbidden region around the Cꢀ
terminal cyclohexyl binding pocket (vide infra).
The 4,4ꢀdifluorocyclohexyl being established as the optimal moiety at the Cꢀterminus for
ATPase inhibitory effect, we proceeded to probe optimal binding conformation of the fluoro
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 108
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
containing rings within the Pꢀgp drugꢀbinding pocket. Working toward this hypothesis, several
analogues containing the fluoro groups on various ring systems were linked to the Cꢀterminus of
the (S)ꢀvalineꢀderived thiazole scaffold as shown in Table 3. Compounds (61, 56% at 2.5
ꢀM
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
and 62, 76% at 2.5 M) with the piperidine rings, had moderate to high stimulation of ATPase
ꢀ
activity. This clearly explains the importance of the conformational disposition of the cyclohexyl
moiety in 53 for inhibitory action on ATPase activity. Two analogues, representing a
conformationally flexible extension at the cyclohexane ring, were synthesized. Compound 63
with the methylene insertion between the 4,4ꢀdifluorocyclohexyl ring and an amide ‘NH’
inhibited ATPase activity by 27% at 0.5
ꢀM but stimulated by 21% at 2.5
ꢀM. Conversely, the
4
ꢀtrifluoromethyl cyclohexyl analogue (64, 22% inhibition at 2.5
ꢀM) weakly inhibited ATPase
activity. The inhibition efficiency of the compound decreases when the 4,4ꢀdifluorocyclohexyl or
the fluoro groups are extended by one linker atom. To determine whether ATPase inhibition can
be improved by retaining the fluoro groups while eliminating saturated ring structure, we
synthesized two analogues containing fluorinated aromatic rings (65, 244% at 2.5
ꢀM and 66,
1
76% at 2.5 M) and a flexible nonꢀcyclic 4,4,4ꢀtrifluorobutyl analogue (67, 77% at 2.5 ꢀ
ꢀ
M).
These analogues have the same number of bonds between the fluorine and the amide group, as in
compound 53, but with different orientations. All of these compounds stimulated basal ATPase
activity. The vector for the fluorine atoms in various fluorinated analogs is clearly distinct from
that present in 53, which demonstrates the contribution of an optimal axial and equatorial
arrangement of the fluorines in 53 for Pꢀgp affinity. Compounds 61, 62, 65, 66, and 67
substantiates the importance of a cyclohexyl ring whereas compounds 63 and 64 suggest the
requirement for precise positioning of the fluorine atoms in the substrateꢀbinding site of Pꢀgp for
inhibition of basal ATPase activity. It is evident from the above SAR that any deviation from the
16
ACS Paragon Plus Environment

Page 17 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
,4ꢀdifluorocyclohexyl group results in a loss of ATPase inhibitory activity. Therefore, in the
next round of SAR, we retained the 4,4ꢀdifluorocyclohexyl ring structure on the Cꢀterminus and
conducted structural modifications at the Nꢀterminus.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Toward this goal, we have synthesized 12 compounds as shown in Table 4. At the outset,
various unsaturated and saturated ring systems were installed in lieu of the trimethoxyphenyl
moiety. The aromatic heterocycles such as 4ꢀthiazolyl (69, 50% at 2.5
4% at 2.5 M) produced a switch in the ATPase activity from the inhibition to a stimulation.
The 5ꢀpyrimidinyl analogue, 70 (NA at 2.5 M), failed to interact with Pꢀgp. Amongst the
saturated ring analogues, pyran 72 (NA at 2.5 M) proved inactive and the 4,4ꢀ
difluorocyclohexyl analogue, 73 (16% inhibition at 2.5 M), showed a weak inhibition of
ꢀM) and 3ꢀquinolinyl (71,
8
ꢀ
ꢀ
ꢀ
ꢀ
ATPase activity. These analogues substantiated the importance of a 3,4,5ꢀtrimethoxybenzoyl
moiety for its affinity enhancing interactions at the substrateꢀbinding pocket of Pꢀgp (vide infra).
This result prompted us to probe further SAR around the 3,4,5ꢀtrimethoxybenzoyl moiety by
synthesizing analogues that feature sequential removal of the methoxy groups in compound 53 to
obtain analogues 74 (12% at 2.5
ꢀM), 75 (36% at 2.5 ꢀM), 76 (11% at 2.5
ꢀM) and complete
elimination of 3,4,5ꢀtrimethoxybenzoyl moiety as in 77 (NA at 2.5
ꢀ
M). These compounds
produced no appreciable effect on the ATPase activity. It can be concluded that a 3,4,5ꢀ
trimethoxybenzoyl moiety at the Nꢀterminus along with the 4,4,ꢀdifluorocyclohexyl substituent at
the Cꢀterminus are indispensable for ATPase inhibition. However, it remains to be determined
whether methoxy groups can be replaced with other hydrogen bond acceptor groups. In pursuit
of this investigation, we synthesized a 3,4,5ꢀtrifluorobenzoylꢀcontaining compound 78 (23%
inhibition at 2.5
ꢀM), which expectedly retained the ATPase inhibitory effect; however, not to
the extent produced by 53. Hence, it is anticipated that the methoxy oxygen atoms may be
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 108
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
involved in hydrogen bonding interactions within the drugꢀbinding pocket of Pꢀgp (vide infra).
To probe the influence of an altered positioning of the hydrogen bond forming methoxy groups,
we decided to separate the 3,4,5ꢀtrimethoxyphenyl ring and the carboxamide group by a flexible
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
methylene linker to produce compound 79 (70% at 2.5
carbonyl linker (sterically restricted vicinal dicarbonyl functionality) to obtain 80 (176% at 2.5
M); both had shown stimulation of the ATPase activity. It is anticipated that these extended
ꢀM) and by a transꢀfavored rigid
ꢀ
analogues may have undergone steric clash within the substrateꢀbinding pocket of Pꢀgp, and
thus, led to loss of the inhibitory interactions. Based on this data, we believe that the Nꢀterminus
aromatic ring in compound 53 provide requisite hydrophobicity for recognition by Pꢀgp as well
as proper orientation of the hydrogen bond forming methoxy groups for ATPase inhibitory
effect. As a result, hereafter, we retained the 3,4,5ꢀtrimethoxybenzoyl and the 4,4ꢀ
difluorocyclohexyl moieties at the Nꢀ and Cꢀtermini of the (S)ꢀvalineꢀderived thiazole scaffold,
respectively.
Subsequent investigations sought to identify the impact of structural modifications at the
central region of compound 53, including the carboxamide to thiocarboxamide isosteric
replacements (Table 5). In anticipation of obtaining additional ATPase inhibitory compounds,
we replaced the carboxamide group at both Nꢀ and Cꢀterminus with isosteric thiocarboxamide.
This structural change would also provide analogues that are stable to hydrolytic cleavage.
Similarly, replacement of hydrophobic valine side chain with analogous proteinogenic amino
acids is anticipated to produce additional inhibitors of ATPase activity. Toward these objectives,
we synthesized isosteric thioamide analogues 81ꢀ84, which were anticipated to behave similar to
the carboxamide analogues 39 and 53. Unexpectedly, the dithioamide analogues (81, 92% at 2.5
ꢀM and 83, 64% at 2.5 ꢀM) stimulated basal ATPase activity whereas monothioamide 82 (NA at
18
ACS Paragon Plus Environment

Page 19 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
0
.5
ꢀ
M) lost the affinity. We postulate that change of oxygen to sulfur atom may have resulted in
loss or weakening of hydrogen bonding ability of the compounds (81ꢀ83) indicating their
contribution in the interactions in the substrateꢀbinding pocket of Pꢀgp. Nevertheless, 84 (26%
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibition at 2.5 ꢀM) with one thioamide retained the inhibitory potency comparable to that of
carboxamide 53 presumably due to favorable 4,4ꢀdifluoro substitution. Surprising activity profile
of the thioamide compounds demonstrates how Pꢀgp discriminates between molecular changes
as small as amide to thioamide.
Subsequently, we determined the role of the αꢀcarbon and its side chain toward ATPase
activity. Compound 101, lacking the αꢀcarbon, stimulated basal ATPase activity by 60% at 2.5
ꢀ
M and glycine analogue, 107 (13% at 2.5
interaction with Pꢀgp. Moreover, replacement of the isopropyl with an isobutyl group as in
leucine analogue, 108 (68% at 2.5 M), led to a switch from inhibition to the stimulation of
ꢀM), devoid of the isopropyl side chain, had a weak
ꢀ
ATPase activity. We presume that compound 101 may have lost some of the ATPase inhibitory
interactions due to the truncation by one bond length and a lack of side chain. Further the valine
side chain may be responsible for proper orientation of the molecule in the drug/substrateꢀ
binding pocket of Pꢀgp as evidenced from the lack of affinity of 107 toward Pꢀgp. To further
probe the isopropyl interacting region topology of Pꢀgp, we replaced the isopropyl side chain
with the benzyl and weakly acidic 4ꢀhydroxybenzyl side chains yielding ATPase inhibitory
compounds 109 (TTTꢀ150; 43% inhibition at 2.5 ꢀM) and 110 (23% inhibition at 2.5 ꢀM). The
preference for the aromatic amino acid residues, as in 109 and 110, may have resulted from
favorable piꢀpi interactions in the isopropyl binding region of Pꢀgp for ATPase inhibitory activity
(vide infra). Analogue 110 showed slightly less ATPase inhibition as compared to 109, an
observation hinted toward unfavorable contribution of the weakly acidic and polar 4ꢀhydroxyl
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 108
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
group. To explore the impact of replacing flexible valine by a rigid amino acid such as proline,
we prepared (S)ꢀprolineꢀderived thiazole analogue, 111 (21% at 2.5 ꢀM), which lost the affinity
toward Pꢀgp. It may be noted that these αꢀcarbon modified analogues (108ꢀ111) were partially
racemized. In general, the side chain fragment of these compounds play an important role in
interactions at the drugꢀbinding region of Pꢀgp.
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
Herein, we summarize the key SAR trends. Despite large, flexible and tolerant substrate/drugꢀ
binding region of Pꢀgp, the SAR data on the (S)ꢀamino acidꢀderived thiazole compounds in this
report suggests a unique ability of Pꢀgp to distinguish amongst essentially considered as isosteric
structures in medicinal chemistry. Alicyclic analogues 39, 42, and 43 showed preference for the
cyclohexyl ring over isosteric piperidine and pyran rings. Only the difluoro substitution at the 4ꢀ
position of the cyclohexyl ring (53), amongst different hydrogen bonding groups (42, 51 and 52),
improves inhibitory efficiency. Moreover, structural changes with either removal or addition of
one atom resulted in a switch from ATPase inhibition to stimulation as depicted by analogues 53
versus 61, 79, and 108. Unanticipated effects of the thioamides (81, 82 and 83) demonstrated
subtle ability of Pꢀgp to recognize specific groups. Moreover, glycine analogue 107 indicates the
significance of hydrophobic contacts in the substrateꢀbinding site of Pꢀgp. Additionally,
positioning of the trimethoxybenzoyl moiety, commonly present in several Pꢀgp modulators, is
critical at the Nꢀterminus and is involved in key hydrogen bonding interactions (vide infra).
Introduction of an aryl ring at the Cꢀterminus of the thiazole scaffold resulted in stimulation of
ATPase activity and only all carbon six membered saturated ring resulted in inhibitory activity.
45
Furthermore, a variety of molecular descriptors, reviewed by Didziapetris et. al., were
calculated based on 2D and 3D structures of the compounds (Table S1). Albeit weak correlation,
atom based logP (AlogP/octanolꢀwater partition coefficient) analysis indicates that compounds
20
ACS Paragon Plus Environment

Page 21 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with AlogP ≥3 exhibits higher affinity (moderate stimulation or inhibition of ATPase) with few
exceptions (38, 40, 42, 48, 58, 73, 74, 76, 82, and 101) paralleling similar reports by others and
1
9, 46ꢀ49
our group.
All compounds except 77 have molecular weight >400 Da. Majority of the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
target compounds are well within the acceptable range of Lipinski’s RuleꢀofꢀFive. The values for
the calculated descriptors spans to a similar range for ATPase stimulatory, inhibitory and no
affinity compounds. No apparent discrimination was found for these properties against the
biological behavior of the compounds. It is intriguing that such similar compounds render
different effects on interaction with Pꢀgp. Finally, we examined, in silico, our target compounds
against a reference set to identify any substructure commonly found in Pan Assay Interference
5
0
Compounds (PAINS). None of the target compounds, with the exception of five azideꢀ
containing compounds 29, and 32ꢀ35, were identified as potential PAINS in our analysis. These
azideꢀcontaining compounds were primarily synthesized for photoaffinity labeling experiments.
125
[ I]ꢀIAAP Photolabeling Assay. To determine whether the ATPase inhibitors are interacting
with Pꢀgp at its substrateꢀbinding site, we conducted binding competition assay of compounds 53
125
125
and 109 against [ I]ꢀIAAP. The UVꢀlightꢀinduced covalent labeling of [ I]ꢀIAAP to Pꢀgp was
inhibited by both compounds 53 and 109 in a concentrationꢀdependent fashion (Figure 2).
Compound 53 inhibited a maximum of 90% labeling at >5 ꢁM concentrations (Figure 2A). The
125
[
I]ꢀIAAP labeling IC for compound 53 was 0.1 ꢁM, which is improved compared to lead
50
1
9
compound 1 (0.72 ꢁM). However, the IC for compound 109 was 0.76 ꢁM, which is close to
5
0
that of compound 1 (Figure 2B). These results and ATPase activity data show that compound 53
inhibited basal ATPase activity by binding to TMDs of human Pꢀgp with higher affinity than
compound 1.
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 108
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
MDR Reversal Assay. MDR reversal assay was carried out using representative analogues (1,
1, 39, 53, 60, and 109) to determine the translation from biochemical assay results (ATPase and
3
125
[ I]ꢀIAAP) to clinical implication. Compounds were selected based on their effect on ATPase
activity: potent stimulators (1 and 31), inhibitors (39, 53 and 109), and moderate stimulator (60).
We determined the effect on cytotoxicity to the parental cell line (HEK 293ꢀpcDNA 3.1) and
ABCB1 (Pꢀgp) transfected cell line (HEKꢀABCB1) upon treatment of each compound in
combination with paclitaxel. Paclitaxel was used as the ABCB1 substrate chemotherapy drug.
Verapamil (10 µM) and zosuquidar (0.25 µM) were used as positive reversal compounds against
the HEKꢀABCB1 cell line (Table 6). The nonꢀtoxic test concentration of 10 µM for reversal
compounds was selected because the preliminary cytotoxicity data showed at least 80% cell
survival in both parental and resistant cell lines when treated alone (data not shown).
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 cytotoxicity results demonstrate that the ABCB1ꢀexpressing cells exhibit 28.9ꢀfold
resistance (FR) to paclitaxel, with an IC of 2.054 µM, as compared to the parental HEK 293ꢀ
5
0
pcDNA 3.1 cells, which showed an IC of 0.071 µM when treated with paclitaxel alone. Preꢀ
5
0
treatment with test compounds significantly reduced the IC50 value of ABCB1 substrate,
paclitaxel, in the HEKꢀABCB1 resistant cells. While the effect of the reversal compounds in
combination with paclitaxel showed a negligible toxic effect in the parental HEK 293ꢀpcDNA
3
.1 cells, they proved to diminish the resistance significantly in the resistant, Pꢀgp
overexpressing cells. The relative resistance of the HEKꢀABCB1 cells reduced from 28.9ꢀfold to
0.6, 0.6, 0.6, 2.9 and 1.7ꢀfold in the presence of compounds 1, 39, 53, 60, and 109 (10 µM),
respectively. These compounds performed comparably to our positive controls at their respective
concentrations: verapamil, a firstꢀgeneration Pꢀgp inhibitor and known ATPase stimulator (10
µM); and zosuquidar, a third generation potent Pꢀgp ATPase inhibitor (0.25 µM). The lowered
22
ACS Paragon Plus Environment

Page 23 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
test concentration of zosuquidar was necessary due to the potency of the compound and because
it was cytotoxic at 5 µM concentration. Competitive inhibition of the Pꢀgp can be reversible or
irreversible throughout the 72 h incubation period. We envision the ATPase stimulators may
reverse MDR through competition with the substrate paclitaxel while the ATPase inhibitors
prevent the paclitaxel efflux by nonꢀcompetitive inhibition. Compound 31 has a weak reversal
effect at 10 µM (foldꢀresistance of 17.8) suggesting it might be effluxed out at a rapid rate.
Overall, compounds 1, 39, 53, 60, and 109 were found to overcome Pꢀgpꢀmediated resistance to
paclitaxel.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
One of the drawbacks, with previous generations of Pꢀgp modulators, including our lead
20
compound 1 (CYP3A4 IC = 8.2 µM), was concomitant inhibition of CYP3A4. Therefore, we
50
conducted in vitro CYP3A4 inhibition assay with representative compounds as per our prior
2
0
report (Table 7). Pꢀgp ATPase inhibitors (39, 53, and 109) did not inhibit CYP3A4 up to 25
µM whereas compound 31, which is an ATPase stimulator and benzophenoneꢀcontaining
analogue has CYP3A4 inhibition of 44% at 25 ꢀM and 83% at 50 ꢀM. Hence, replacement of
the Cꢀterminus benzophenone moiety of compound 1 with a cyclohexyl group offered selectivity
toward Pꢀgp over CYP3A4.
Molecular Modeling. Based on the SAR on Pꢀgp ATPase activity of the synthesized
analogues, we hypothesize that inhibitory compounds occupy a sterically demanding cavity in
the drugꢀbinding site and are involved in hydrogen bonding interactions with specific residues.
To rationalize this hypothesis, we carried out inducedꢀfit docking (a method that allow both
protein side chains and ligand flexibility) of compounds 53 and 109 (Figure 3) within the
transmembrane domain of homology modeled structure of human Pꢀgp derived from the crystal
5
1
structure of mouse Pꢀgp (PDB ID: 4Q9H). We chose mouse Pꢀgp because it is the only
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 108
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
available mammalian crystal structure with high resolution and has 87% identical sequence to
that of human Pꢀgp. Figures 3A and 3B illustrates docking pose for compound 53. The docking
data showed hydrogen bonding interaction of the 4ꢀmethoxy oxygen atom with the side chain of
Gln990 (H COꢀꢀꢀꢀH NꢀGln990). The amide group, connecting the trimethoxyphenyl and the Cα
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
2
of the valine, entered into a hydrogen bonding interaction with the phenolic hydroxyl of Tyr307
C=OꢀꢀꢀꢀHOꢀTyr307). The 4,4ꢀdifluorocyclohexyl moiety occupies a space that is closely
(
surrounded by residues Phe335, Phe336, Leu339, Ile340, and Phe343. Thus, any modification of
the 4,4ꢀdifluorocyclohexyl moiety at the Cꢀterminus will result in steric hindrance and a loss of
inhibitory activity. The isopropyl group of a valine residue of the compound may be involved in
hydrophobic interactions with the residues, Phe728 and Ala729, in the transmembrane region of
Pꢀgp.
Binding model of (S)ꢀisomer of 109 within the substrateꢀbinding region of Pꢀgp indicates the
formation of three hydrogen bonds (4ꢀH COꢀꢀꢀꢀH NꢀGln838; 3ꢀH COꢀꢀꢀꢀH NꢀGln990 and the Nꢀ
3
2
3
2
terminusꢀNHꢀꢀꢀꢀOHꢀTyr307) as shown in Figures 3C and 3D. Moreover, the side chain benzyl is
stabilized by an edgeꢀtoꢀface (Tꢀshape) interaction with the side chain of Tyr310. The 4,4ꢀ
difluorocyclohexyl moiety is surrounded by residues Phe335, Phe336, Leu339, Ile340, and
Phe343 similar to that observed for compound 53. The surface representation of bound (S)ꢀ
isomer of 109 also shows its binding at the sterically demanding binding site especially at the
cyclohexyl region (Figure 3E).
Since, both cisꢀ and transꢀforms for (S)ꢀproline are reported to exist in small peptides, we
52
performed ligand alignment using both isomers. Superposition of docked conformations of
compound 53 and conformationally restricted cisꢀ and transꢀproline isomers (111) indicated a
substantial movement of the proline analogue as compared to the valine analogue 53 (Figure 3F
24
ACS Paragon Plus Environment

Page 25 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and 3G). The trimethoxybenzoyl portion of the molecule shifts away from the position that was
discerned from the valineꢀderived analogue 53. Therefore, it can be concluded that the
conformationally restricted prolineꢀderived analogue, 111, presented interactions in the manner
that were detrimental to its binding affinity toward Pꢀgp. This may have resulted from the loss of
key hydrogen bonding interaction with the side chains of Tyr307 and Asn842 and/or Gln990 in
case of proline analogue.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
It may be noted that docking experiments cannot necessarily identify the correct binding
pattern of the ligand and particularly in this case due to a large and flexible binding pocket of Pꢀ
gp. However, predicted binding models may facilitate future optimization efforts. The interaction
pattern based on the inducedꢀfit docking poses as well as the SAR data together suggests that the
trimethoxybenzoyl group is surrounded by polar side chains of Gln725, Asn842, Gln838, and
Gln990, which may facilitate proper orientation of the molecule in this region. Since the side
chain benzyl group of 109 is involved in Tꢀshape aromaticꢀaromatic interaction, it could be
replaced with heteroaromatic ringꢀcontaining unnatural amino acids to strengthen an edgeꢀtoꢀ
face interaction with the side chain of Tyr310. Furthermore, the two amide nitrogen atoms can be
linked together to form macrocyclic analogues by ring closing metathesis. Such macrocyclic
analogues would bind within the substrateꢀbinding region of Pꢀgp in an entropically favorable
manner compared to their nonꢀcyclic counterparts. This will form the basis for our future SAR
studies.
CONCLUSIONS
A series of substituted (S)ꢀamino acidꢀderived thiazole analogues were synthesized to probe
the substrateꢀbinding site within the TMDs of Pꢀgp and their effect on basal ATPase activity was
determined. Starting from the lead compound 1, a potent stimulator of ATPase activity, few
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 108
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
inhibitors (39, 53, 78, 84, 109 and 110) of ATPase activity were identified. The inhibitory
activity toward ATPase was largely contributed by the cyclohexane ring. Moreover, replacement
of the cyclohexane moiety with piperidine (61 and 62), aromatic (65 and 66) or open chain alkyl
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
(67) groups resulted in a switch from inhibition to stimulation of ATPase activity. The
trimethoxy group at the Nꢀterminus was critical for preserving inhibitory function and is
involved in hydrogen bonding interactions as evidenced from 74, 76 and 78. Compounds 53, 61,
7
9, and 108 show that variation by one bond length results in significant shift in affinity toward
125
Pꢀgp. Moreover, photolabeling competition data with [ I]ꢀIAAP and compounds 53 and 109
provided evidence for interaction of these compounds with the TMDs of Pꢀgp. MDR reversal
assay data showed that ATPase stimulatory compound 31 as well as ATPase inhibitors 39, 53
and 109 act as MDR reversal compounds, presumably with different mechanisms. To our
delight, these ATPase inhibitors did not inhibit CYP3A4 unlike starting lead compound 1 and
previous generations of Pꢀgp modulators. The substrateꢀbinding pocket of Pꢀgp is thought to be
very flexible in nature and recognizes various size and type of chemical structures. The effect of
the synthesized series, in this report, reinforces this fact; however, it is intriguing to discover that
an extension or truncation by only one methylene group switches the inhibition to the stimulation
of ATPase activity.
EXPERIMENTAL
Chemical Synthesis. Materials and Instrumentation. Chemicals were purchased from Aldrich
Chemical Co. (Milwaukee, WI), AK scientific (Union City, CA), A2Z chemicals (Irvine, CA),
Oakwood Products (West Columbia, SC), TCI America (Portland, OR) and Alfa Aesar (Ward
Hill, MA), and were used as received. All chemicals were confirmed for uniformity by thin layer
chromatography (TLC) with silica gel as the adsorbent layer (250 microns) on aluminum backed
26
ACS Paragon Plus Environment

Page 27 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
plates (Agela Technologies). Reactions were monitored by TLC, and visualized using Ultraviolet
(UV) light at 254 nm. Melting points were recorded on a Stuart Melting Point apparatus (model ꢀ
TM
1
13
SMP20) and are uncorrected. Bruker 400 Ultrashield spectrometer ( H at 400 MHz and C at
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
00 MHz) equipped with a zꢀaxis gradient probe was used to record NMR experiments. H NMR
13
and C NMR chemical shifts were reported downfield from tetramethylsilane (TMS, internal
1
standard) in parts per million (δ ppm). The H NMR data are presented as follows: chemical shift
(multiplicity {s (singlet), bs (broad singlet), d (doublet), t (triplet), dd (doublet of doublets), m
13
(
multiplet) and oct (octet)}, number of protons, coupling constant). The C NMR (proton
decoupled, fluorine coupled) data are presented as follows: chemical shift (multiplicity {d
(doublet), t (triplet)}). Column chromatography purifications were performed with silica gel (40ꢀ
6
3 µm) obtained from Silicycle Inc. (Quebec City, Canada) and flash chromatography was
performed using Reveleris® X2 flash chromatography system (BÜCHI Corporation, New
Castle, DE). Preparative TLC was performed using Silica Gel GF 1000 µm 20x20cm glass
backed plates from Analtech (Miles Scientific, Newark, DE). Target compounds purity analysis
was accomplished utilizing Agilent 1100 HPLC system with an autosampler (Agilent, Santa
Clara, CA) eluting a Cꢀ18 reverse phase column (Agilent Eclipse plus C18, 3.5 ꢀm, 4.6 x 100
mm) with an isocratic mobile phase flow (1.0 mL/min), and samples monitored under UV light
at 254 nm. All target compounds were established to be ≥95% pure (major peak area/total
combined area of peaks). Mass analyses were carried out on Agilent 1260 infinity series liquid
chromatography (LC) system (C18 column, Agilent InfinityLab poroshell 120, ECꢀC18, 2.7
ꢀm,
4
.6 x 50 mm) connected with Agilent 6120 quadrupole mass spectrometer (MS). The elemental
analyses (C, H, and N) were performed by Atlantic Microlabs, Inc., (Norcross, GA) and the
observed values were within ±0.4% of calculated values.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 108
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
31 19
20
Synthesis. Compounds 2,
6
and 10 were synthesized according to the procedures as
described in our previous reports.
Method A: Typical Procedure for Peptide Coupling Reactions. A homogenous
solution/suspension of carboxylic acid (1 eq) in either dry N,Nꢀdimethylacetamide (DMA; 10
mL) or dry dichloromethane (DCM; 15 mL), under nitrogen atmosphere, was brought to 0°C and
then added with diisopropylethylamine (DIEA; 1.5 eq), HCTU (1.5 eq), and HOBt (1.5 eq). The
resulting mixture was stirred for 10ꢀ15 min at 0°C before the addition of cold
solution/suspension of amine (1.2 eq) in appropriate solvent (DMA or DCM; ~1ꢀ2 mL). It was
then stirred overnight at rt, after which the resulting solution was evaporated. Ethyl acetate was
added to a crude product mass and washed with 1N aqueous potassium bisulfate (KHSO4)
solution. The acidic aqueous layer was then extracted with 2X ethyl acetate. Combined ethyl
acetate extracts were partitioned with saturated aqueous solution of sodium carbonate (Na CO ).
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
3
Ethyl acetate extract was collected, dried over anhydrous magnesium sulfate (MgSO ) and
4
evaporated. The remaining mass was purified by chromatography (preparative TLC or flash
chromatography) using nꢀhexanes/ethyl acetate as mobile phase to obtain coupled product as a
colorless oil. This oil was triturated with ethyl acetate and nꢀhexanes to provide a pure solid.
Method B: General Procedure for the NꢀBoc Deprotection. NꢀBoc protected amine (1 eq)
dissolved in DCM (20 mL) was slowly added to trifluoroacetic acid (TFA; 10 eq) at 0°C. A
resulting reaction mixture was kept at 0°C for 10 min and then it was stirred at rt for an
appropriate period (4ꢀ12 h). The volatiles were evaporated to leave viscous mass, which was
dissolved in water. It was partitioned with ethyl acetate and the acidic aqueous layer was
collected. The water layer was carefully neutralized using saturated aqueous solution of Na CO3
2
28
ACS Paragon Plus Environment

Page 29 of 108
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and then extracted with ethyl acetate. The ethyl acetate extract was passed over anhydrous
MgSO and evaporated to obtain free amine, which was used without further purification.
4
Method C: Procedure for the Synthesis of Carboxamides by MixedꢀAnhydride Method.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Appropriate carboxylic acid (1 eq) in anhydrous THF (50 mL) under nitrogen was cooled to ꢀ
2
0°C. Then, isobutyl chloroformate (1.2 eq) and Nꢀmethylmorpholine (NMM; 1.2 eq) were
added to the reaction mixture and stirred at ꢀ20°C for 1ꢀ2 h. To this suspension was added excess
25 mL) of aqueous ammonia (30%), which produced a clear biphasic solution that was allowed
(
to stir for an additional 30 min ꢀ 2 h at rt. After concentrating the reaction mixture, it was
extracted with 2X ethyl acetate. The combined ethyl acetate extract was partitioned with aqueous
1
N KHSO , collected, dried over anhydrous MgSO and evaporated to provide desired
4 4
carboxamide as a solid.
Method D: General Procedure for the Synthesis of Thioamides. To a solution of
carboxamide (1 eq) in anhydrous THF (40 mL) was added Lawesson’s reagent (0.57 eq) under
nitrogen atmosphere. The suspension was stirred for 8 h at rt. At this time, entire solid material in
the reaction mixture dissolved to produce a clear yellow solution. Subsequently, saturated
aqueous NaHCO solution was slowly added to the reaction flask until effervescence ceases.
3
Ethyl acetate (15 mL) was added to the quenched mixture and stirred for an additional hour. The
TLC analysis showed two spots with near R values of which the upper faint spot disappeared
f
from the organic layer after extraction workup. The ethyl acetate extract was dried (anhydrous
MgSO ), concentrated and then purified through column chromatography using nꢀhexanes/ethyl
4
acetate as eluent giving pure thioamide as a thick oil.
Method E: General Procedure for the Construction of Thiazole Scaffold. The thioamide (1
eq) in 40 mL anhydrous ethanol under inert condition was cooled using a sodium chloride/iceꢀ
29
ACS Paragon Plus Environment