Characterization of Aminobenzylphenols as Protein Disulfide Isomerase Inhibitors in Glioblastoma Cell Lines

Article abstract of DOI:10.1021/acs.jmedchem.0c00728

Disulfide bond formation is a critical post-translational modification of newly synthesized polypeptides in the oxidizing environment of the endoplasmic reticulum and is mediated by protein disulfide isomerase (PDIA1). In this study, we report a series of α-aminobenzylphenol analogues as potent PDI inhibitors. The lead compound, AS15, is a covalent nanomolar inhibitor of PDI, and the combination of AS15 analogues with glutathione synthesis inhibitor buthionine sulfoximine (BSO) leads to synergistic cell growth inhibition. Using nascent RNA sequencing, we show that an AS15 analogue triggers the unfolded protein response in glioblastoma cells. A BODIPY-labeled analogue binds proteins including PDIA1, suggesting that the compounds are cell-permeable and reach the intended target. Taken together, these findings demonstrate an extensive biochemical characterization of a novel series of highly potent reactive small molecules that covalently bind to PDI.

Full text of DOI:10.1021/acs.jmedchem.0c00728

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Characterization of aminobenzylphenols as protein
disulfide isomerase inhibitors in glioblastoma cell lines
Andrea Shergalis, Ding Xue, Fatma Z Gharbia, Hannah Driks, Binita
Shrestha, Amina Tanweer, Kirin Cromer, Mats Ljungman, and Nouri Neamati
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00728 • Publication Date (Web): 24 Aug 2020
Downloaded from pubs.acs.org on August 26, 2020
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 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 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.
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 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Characterization of aminobenzylphenols as protein disulfide
isomerase inhibitors in glioblastoma cell lines
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
†
,#
†,#
†
†
†
Andrea Shergalis , Ding Xue , Fatma Z. Gharbia , Hannah Driks , Binita Shrestha ,
†
†
‡
†
Amina Tanweer , Kirin Cromer , Mats Ljungman , Nouri Neamati *
†Department of Medicinal Chemistry, College of Pharmacy, Rogel Cancer Center, University of
Michigan, Ann Arbor, Michigan 48109, United States
‡Department of Radiation Oncology, University of Michigan Medical School and Rogel Cancer
Center, School of Public Health, Ann Arbor, Michigan 48109, United States
KEYWORDS. Brain cancer, unfolded protein response, endoplasmic reticulum stress,
cysteine, glutathione, protein disulfide isomerase
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 80
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
Disulfide bond formation is a critical post−translational modification of newly synthesized
polypeptides in the oxidizing environment of the endoplasmic reticulum and is mediated by protein
disulfide isomerase (PDIA1). In this study, we report a series of α-aminobenzylphenol analogues
as potent PDI inhibitors. The lead compound, AS15, is a covalent nanomolar inhibitor of PDI and
the combination of AS15 analogues with glutathione synthesis inhibitor BSO leads to synergistic
cell growth inhibition. Using a nascent RNA sequencing we show that an AS15 analogue triggers
the unfolded protein response in glioblastoma cells. A BODIPY−labeled analogue binds proteins
including PDIA1, suggesting the compounds are cell-permeable and reach the intended target.
Taken together, these findings demonstrate an extensive biochemical characterization of a novel
series of highly potent reactive small molecules that covalently bind to PDI.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
1
Glioblastoma (GBM) is the leading primary brain cancer diagnosed in adults. GBM
is an aggressive, high−grade glioma with a low 5.6% five−year survival rate. Current
standard−of−care – radiation and temozolomide chemotherapy – prolong survival by a
few months, but even after remission, the cancer recurs. Thus, innovative therapies are
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
2
needed to effectively treat GBM. GBM cells survive by upregulating the protein folding
oxidoreductase protein disulfide isomerase (PDIA1, also known as PDI). PDI reduces,
oxidizes, and isomerizes disulfide bonds in nascent polypeptides and other substrates via two
3
catalytic CGHC active sites that sit 15−30 Å apart in two homogenous domains. PDI is
overexpressed in several cancers to meet the increased demands in protein synthesis.4-6
The a and a′ domains of PDI are connected by b and b′ domains, which share identity with
the a and a′ domains, but do not contain the CGHC active sites. As evidenced by the pK values,
a
the N−terminal cysteine in each a and a′ domain active site (Cys53 and Cys397) is stabilized in
the thiolate form, while the C−terminal cysteine thiolate is destabilized. This allows the
nucleophilic N−terminal cysteines to attack substrates and form mixed disulfides. The C−terminal
cysteines more selectively react with the N−terminal cysteines, mediated in the a domain by the
7
pK of Cys56 that is lowered by the local environment, containing a conserved Arg120. This
a
reaction generates a reduced substrate and disulfide in PDI.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The thiolate form of the redox active N−terminal cysteines reacts with electrophilic
compounds. Thus, many electrophilic compounds have been identified as covalent PDI inhibitors,
8
9
10
11
including PACMA31 , KSC−34 , 3,4-methylenedioxy-β-nitrostyrene (MNS) , and 16F16
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Figure 1). PACMA31, an irreversible inhibitor of PDI, demonstrated in vivo efficacy in a mouse
1
2
model of ovarian cancer. Furthermore, PACMA31 synergized with the multi−kinase inhibitor
sorafenib in a mouse model of hepatocellular carcinoma.13 One of the main challenges of
characterizing the many PDI inhibitors identified to date, recently emphasized by Foster, et al., is
1
4
competition with endogenous levels of glutathione. Active site PDI inhibitors, especially reactive
electrophiles such as PACMA31, may compete with glutathione for binding the cysteine thiols of
PDI. This strategy is further confounded by the fact that several PDI family members share similar
active site motifs, and although the CGHC active site is one of the most reactive of the thioredoxin
superfamily, electrophiles that bind this site may also bind to other CxxC−containing PDI family
members. One strategy to mitigate off−target toxicity of cysteine−reactive active site inhibitors is
to develop more selective inhibitors.15
Et
N
O
N
Me
O
OH
S
O
OH
OH
H
N
O
N
OEt
HO
O
O
MeO
O
OH
OH
Cl
O
HO
O
OH
HO
OMe
PACMA31
Bepristat 1a
Isoquercetin
O
O
O
S
N
N
N
H3C
N
NO2
O
N
S
S
N
N
N
N
O
O
S
LOC14
35G8
CCF642
CN
H
OH
N
N
N
HO
H
N
N
O
H
H
HN
Cl
N
H
HO
O
BAP2
KSC-34
Estradiol
4
ACS Paragon Plus Environment

Page 5 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 1. Previously reported PDI inhibitors studied in the context of ovarian cancer (PACMA31),
Huntington’s disease (LOC14), brain cancer (BAP2 and 35G8), thrombosis (Bepristat 1a and
isoquercetin), and multiple myeloma (CCF642). KSC−34 is an a−site selective probe and estradiol is an
endogenous ligand of PDIA1.
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
In addition to the increased PDI expression, GBM cells upregulate the antioxidant
defense system and are increasingly dependent on glutathione as the tumor grows.16
Glutathione is one of the most abundant molecules in the cell, with concentrations of reduced
glutathione (GSH) estimated to reach up to 10 mM in cellular compartments. The tripeptide
composed of glutamic acid, glycine, and cysteine is a key redox buffer and antioxidant molecule
involved in many cellular processes including reactive oxygen species removal, signal
1
7
transduction, and protein synthesis. Although the total glutathione concentration is similar in the
cytosol and ER, the ratio of reduced to oxidized glutathione dictates its role in each compartment.
In the cytosol, glutathione is present mainly in its reduced form, with a ratio of GSH:GSSG of
~50,000:1; the ER is a more oxidizing compartment, with an estimated GSH:GSSG ratio of less
1
8
than 7:1. Because of this environment, PDIA1 is present in its disulfide form, poised to accept
electrons from the reduced thiols in nascent polypeptides. The oxidizing environment of the ER
aids in protein folding and disulfide bond formation.
The initial goal of this study was to characterize a new class of PDI inhibitors in the context
of GBM. Via a medium−throughput biochemical screen, we identified a series of α-
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 80
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
aminobenzylphenols as inhibitors of PDI that possibly target the active site cysteines of PDI via a
retro−Michael addition reaction. Similar compounds have been reported for multiple targets,
1
9
20-23
24
25
including HDACs , MIF tautomerase
, STAT3/5 , and frataxin , among others. This finding
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
suggested that these inhibitors may not be selective for a single protein in the cells. However, we
synthesized a BODIPY−labeled analogue that did not bind any of the previously published targets.
Binding was enhanced when the cells were pre−treated with BSO, suggesting that the efficacy
may be lowered by intracellular glutathione. Additionally, potency in cells increased with BSO
treatment. Thus, the compounds are cell−permeable, and cytotoxicity is enhanced with glutathione
depletion. In all, this study represents the importance of validating in−cell target engagement early
in the drug discovery process and provides extensive characterization of a class of thiol−reactive
small molecules.
RESULTS
Chemistry. Most of the compounds in the present study were synthesized via a well−established
Betti reaction employing a phenolic compound, an aldehyde, and an amine in a one−pot fashion.26,
2
7
Reactions were either heated in toluene at 100 °C or in EtOH under microwave irradiation at
1
20 °C for 30 min depending on the electronic nature of the phenolic starting material. Typically,
simple phenols, hydroxy naphthalenes, and hydroxy quinolones can be used in such reactions,
whereas the reactivity of other hydroxy heterocyclic structures remains uncertain. In our study,
depending on the substitution pattern and electronic properties of the starting materials, three
different types of products were generated (Scheme 1). For starting materials represented by 3,
Type I compounds were obtained unambiguously, while with 4a/4b, electrophilic replacement
could occur at two distinct ortho positions of the hydroxy, leading to either Type II or Type III
6
ACS Paragon Plus Environment

Page 7 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds. We found that regioselectivity was strongly related to the electronic properties of the
bicyclic heterocycles: electron−efficient compounds tend to generate Type II compounds, and
electron−deficient compounds generate Type III 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
Scheme 1. Synthesis of the compounds with modified Betti reaction.a
R3
CHO
OH
OH
a or b
H
N
R1
+
+
R3
N
R1
R2
R2
1
2
3
5a-r
Type I compounds
R3
CHO
HO
a or b
R1
H
N
N
+
+
R3
R1
R2
R2
HO
1
2
4a
6a-g
Type II compounds
R
CHO
HO
H
N
a or b
R1
N
R2
HO
+
+
R3
R1
R2
1
2
4b
7a-l
Type III compounds
aReagents and conditions: a) Toluene, 100 °C; b) EtOH, 120 °C microwave or conventional
heating.
Compound 8 was prepared by direct methylation of 5g with dimethyl sulfate (Scheme 2).
Compound 13, with a cyclohexane in place of the piperidine moiety, was synthesized by a
four−step strategy (Scheme 3). Benzo[d][1,3]dioxol-5-ol (9) was reacted with benzoyl chloride by
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 80
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
an ortho−selective electrophilic attack facilitated by a highly coordinating metal phenolate
intermediate generated in situ.28, 29 The diaryl ketone 10 was reacted with cyclohexylmagnesium
bromide to give tertiary alcohol 11. A TsOH−catalyzed elimination afforded 12, and 12 was
subjected to catalytic hydrogenation to give 13.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Synthesis of compound 8.a
F
F
OH
O
a
O
O
N
N
5g
8
a
Reagents and conditions: a) Me SO , K CO , acetone.
2
4
2
3
Scheme 3. Synthesis of compound 13.a
HO
O
O
a
b
HO
c
O
O
O
O
O
HO
HO
9
10
11
d
O
O
O
O
HO
HO
12
13
a
Reagents and conditions: a) i, MeI, Mg, Et O, ii, benzoyl chloride; b) cyclohexylmagnesium
2
bromide, THF, rt−50 °C; c) p-TsOH, toluene, rt; d) Pd/C, H , EtOH, 60 °C.
2
8
ACS Paragon Plus Environment

Page 9 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lead Compound AS15 Is a Nanomolar Inhibitor of PDI. A screening campaign of 1000 highly
diverse compounds from the National Cancer Institute at 40 μM in the PDI reductase assay
afforded almost 200 compounds that inhibited 50% of PDI activity (Figure 2A). Those compounds
were screened for cytotoxicity in U−87 MG and MIA PaCa−2 cell lines, and compounds with >
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
0% inhibition of cell growth at 30 μM were subjected to dose response assays in both the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and PDI reductase assays. AS15
and CD343 emerged as lead scaffolds with IC values in the PDI reductase assay of 300 ± 90 nM
5
0
and 170 ± 50 nM, respectively (Figure 2B−C). When a residue in the b′ domain of PDI important
for substrate binding, Histidine 256, was mutated to an alanine residue, both AS15 and CD343
retained their activity (Figure S1A). This result suggested AS15 and CD343 activity was not
dependent on the substrate−binding domain. Thus, we focused on determining whether inhibition
was based on the interaction between AS15 and CD343 and the active−site cysteines. AS15 and
CD343 inhibited activity of PDIA2 and, to a lesser extent, PDIA3 (Figure S1B). The lead
compounds decreased viability of U−87 MG cells with IC values of 18.3 ± 9.2 μM for AS15 and
5
0
1
0.6 ± 0.7 μM for CD343 (Figure S1C). Interaction with oxidized PDI was further probed with
the thermal shift assay; however, AS15 and CD343 did not stabilize PDI to thermal degradation,
similar to PACMA31 (Figure 2D). These initial results suggested that the compounds inhibit PDI
as thiol-reactive compounds and not via the substrate−binding domain like estradiol or BAP2.30
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. (a) Screening approach leading to the discovery of AS15 and CD343. (b) Structures of AS15 and
CD343, and IC50 values calculated in the PDI reductase assay. (c) Dose−response curves of PACMA31,
AS15, and CD343 in the PDI reductase assay. (d) Shifts in the melting temperatures of PACMA31, AS15,
and CD343 in the thermal shift assay. SAR: structure−activity relationship.
10
ACS Paragon Plus Environment

Page 11 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure−Activity Relationships Reveal AS15 Analogues Are Not Substrate−Binding
Domain Inhibitors. To further assess the mechanism of inhibition of AS15 and CD343 and
determine the structural requirements for their activity, we obtained 89 analogues from Chem Div
libraries and the NCI Developmental Therapeutics Program and tested their potency in the PDI
reductase assay (Tables 1−4). Compounds were tested for purity, and only compounds with purity
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
>
95% were used for structure−activity relationship analysis. Generally, all AS15 and CD343
analogues possess an α-aminobenzylphenol core that consists of phenolic, amino, and phenyl
moieties. While the phenolic moiety is limited to 5-hydroxybenzo[d][1,3]dioxole (AS15
analogues) or 8-hydroxyquinoline (CD343 analogues), the amino and phenyl moieties are
diversified. For compounds with the 5-hydroxybenzo[d][1,3]dioxole moiety, a variety of cyclic
amines including morpholine, piperidine, piperazine, and pyrrolidine were incorporated. Most of
the compounds inhibited PDI with IC50 values below 1 µM, and different halogens and
electron−donating groups such as chloro, fluoro, methoxy, amino, and hydroxy on the phenyl ring
were well−tolerated. Some of the compounds showed moderate cytotoxicity with IC values of
5
0
1
0−30 µM; however, correlation with PDI inhibition was not observed (Table 1). Aromatic amines
such as aminopyridine, aminopyrimidine, and imidazole were generally tolerated, and PDI
inhibition was comparable to those with saturated cyclic amines. Ureas were not tolerated and led
to a complete loss of activity, possibly because of their loss of basicity that contributed to PDI
binding (Table 2). Compounds without the bicyclic aromatic core or hydroxy group were inactive
against PDI and non−toxic (Figure S2). Furthermore, compounds without the tertiary amine were
inactive against PDI and were non−toxic, with the exception of tri−methoxy substituents on the
aromatic ring (Table S1). CD analogues possess the 8-hydroxyquinoline core along with an
aromatic amine such as 2-aminopyridine or aniline, and their ability to inhibit PDI activity was
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 80
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
comparable to the AS15 analogues, except that the 4-methyl substitution of the pyridine was not
compatible when methyl or chloro was present on the phenyl moiety. Interestingly, many of these
compounds showed stronger cytotoxicity with IC values as low as 2.1 ± 0.1 µM (Tables 3 and
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
).
Table 1. SAR of Compounds with 5-Hydroxybenzo[d][1,3]dioxole and Saturated Cyclic Amine
Moieties
R2
O
R1
O
HO
NSC
Number
PDI IC50
(μM)a
MTT IC50
compound
R
1
R
2
μM)b
(
AS15/NC014
NC016
NC133
NC107
NC108
NC110
NC161
NC134
NC115
NC117
NC141
NC018
NC015
NC022
NC020
368252
368260
381577
368248
368256
368261
364724
381579
368275
368277
667921
368267
368253
368274
368273
3,4-di-OCH
3
0.30 ± 0.09
0.64 ± 0.31
0.98 ± 0.30
< 0.2
18.3 ± 9.2
> 10
3,4,5-tri-OCH
3
3
2,4,6-tri-OCH
13.6 ± 1.7
> 30
3,4-OCH
2-OH, 3-OCH
3-OCH , 4-OH
2
O-
3
0.90 ± 0.75
0.23 ± 0.11
0.13 ± 0.06
0.33 ± 0.05
0.092 ± 0.023
0.23 ± 0.06
0.34 ± 0.09
0.96 ± 1.08
0.70 ± 0.13
0.88 ± 0.36
1.27 ± 0.34
> 10
N
3
27.1 ± 2.4
> 10
O
4-OCH
2-OCH
4-Cl
3
3
27.7 ± 2.7
> 30
4-F
> 30
2-OH
> 30
3,4-di-OCH
3, 4, 5-tri-OCH
4-OCH
4-N(CH
3
24.8 ± 1.2
> 10
3
N
3
28.3 ± 1.6
24.7 ± 2.4
3 2
)
12
ACS Paragon Plus Environment

Page 13 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
NC024
NC162
NC116
370278
368254
368276
2,4-di-OCH
3
2.59 ± 1.84
0.35 ± 0.11
0.27 ± 0.04
> 30
> 10
> 30
3,4-OCH
4-F
2
O-
N
NC120
369090
4-OH
0.68 ± 0.35
> 30
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
HN
NC025
NC026
NC027
NC028
NC122
NC123
NC124
370279
370281
370283
370285
370280
370282
370284
2,4-di-OCH
4-OCH
4-N(CH
3
0.36 ± 0.05
1.52 ± 0.02
< 0.20
> 30
22.6 ± 5.9
> 30
3
3 2
)
N
2-OH, 3-OCH
4-F
1.65 ± 0.65
< 0.20
> 10
3
> 30
3,4-OCH
2
O-
0.023 ± 0.017
0.18 ± 0.02
> 30
3,4,5-tri-OCH
3
24.2 ± 7.1
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity IC50 values were determined in U−87 MG cells.
Table 2. SAR of Compounds with 5-Hydroxy benzo[d][1,3]dioxole and Other Amine Moieties
R2
O
R1
O
HO
NSC
Number
PDI IC50
(μM)a
MTT IC50
compound
R
1
R
2
μM)b
(
NC163
NC300
NC017
NC019
NC165
NC299
368255
368281
368265
368270
368268
368279
N
3,4-OCH
3,5-di-OCH
3,4-OCH
4-OCH
2-OH, 3-OCH
3,5-di-OCH , 4-OH
2
O-
, 4-OH
O-
0.072 ± 0.017
0.17 ± 0.04
1.18 ± 0.11
1.30 ± 0.39
9.02 ± 4.61
2.20 ± 0.53
> 10
> 10
> 10
> 10
> 10
> 10
N
H
3
2
N
3
N
N
H
3
3
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 80
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
NC166
NC118
NC119
368278
368280
369087
4-F
4-F
12.72 ± 3.47
0.67 ± 0.14
0.62 ± 0.04
> 10
> 10
> 10
N
4-OCH
3
N
O
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
NC023
369678
4-OH
> 40
> 10
H N
2
N
H
NC029
NC030
NC168
371006
371007
371005
3,4-OCH
2
O-
> 40
> 40
> 30
> 10
> 10
> 10
O
3,4,5-tri-OCH
3
N
H
N
H
4-F
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity IC50 values were determined in U−87MG cells.
Table 3. SAR of compounds with 8-Hydroxyquinoline and 2-Aminopyridine Moieties
R2
OH
N
R1
PDI IC50
MTT IC50
(μM)b
compound
CD528
ID
R
1
R
2
(
μM)a
N
3,5-di-OCH
OH
3
, 4-
G856−2528
0.18 ± 0.06
7.4 ± 1.2
N
H
CD343
CD345
CD639
CD611
CD344
CD346
CD355
CD638
CD626
4896−2958
4896−3004
7706−0076
4896−3086
4896−3003
4896−3082
4896−4013
4896−3084
G856−2546
4-Cl
0.17 ± 0.05
1.89 ± 0.31
0.50 ± 0.17
0.46 ± 0.07
0.17 ± 0.12
0.73 ± 0.12
0.19 ± 0.09
2.86 ± 2.37
0.47 ± 0.17
10.6 ± 0.7
6.9 ± 1.6
9.4 ± 0.6
> 10
2,6-di-Cl
2-Cl
4-CH
4-CH CH
2,5-di-CH
2,4,6-tri-CH
3
N
2
3
9.3 ± 5.9
13.8 ± 5.9
10 ± 0.6
11.9 ± 0.4
> 10
N
H
3
3
3-OCH
2-F
3
14
ACS Paragon Plus Environment

Page 15 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CD613
NC272
CD354
CD361
CD350
CD377
CD341
CD373
CD348
CD349
CD352
CD362
CD601
CD363
CD594
5994−0466
1014
2-Cl, 3-OH
0.37 ± 0.03
< 0.20
> 10
> 10
−
4896−4000
5994−0131
4896−3501
7706−0074
4896−0018
7033−0321
4896−3250
4896−3254
4896−3773
5994−0331
G856−2531
5994−0397
5704−0657
2-CH
3
> 40
> 10
2,5-di-CH
3
> 40
> 10
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-F
0.14 ± 0.06
0.88 ± 0.23
> 40
> 10
2-F
13.1 ± 3.5
20.6 ± 7.8
9.2 ± 0.4
9.9 ± 1.4
7.2 ± 2.2
9.3 ± 0.9
8.9 ± 2.2
2.1 ± 0.1
3.4 ± 0.6
> 10
2-F, 6-Cl
3-F
N
0.35 ± 0.07
> 40
N
2-Cl
H
2,6-di-Cl
4-OBn
> 40
0.62 ± 0.29
5.39 ± 0.68
26.50 ± 19.15
0.53 ± 0.15
0.15 ± 0.02
4-CH
4-CF
4-NO
3-OH, 4-OCH
2
CH
3
3
2
3
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity IC50 values were determined in U−87MG cells.
Table 4. SAR of Compounds with 8-Hydroxyquinoline and Aniline Moieties
OH
R
N
N
H
NSC
Number
PDI IC50
compound
R
MTT IC50
a
(μM)
b
(μM)
NC266
NC268
NC269
NC270
NC273
1008
1010
1011
1012
1015
−
0.11 ± 0.04
0.15 ± 0.05
0.30 ± 0.05
1.48 ± 1.52
< 0.20
> 10
> 10
> 10
> 10
> 10
4-NO
2
4-COOH
2-COOH
2 5
2-COOC H
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 80
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
NC282
84087
2-OCH
3
0.39 ± 0.11
> 10
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity IC50 values were determined in U U−87MG cells.
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
Having explored the SAR around the amino and phenyl moieties, which generally had a
nominal impact on PDI inhibition, we synthesized and tested an additional 38 compounds to
diversify the phenolic moiety and probe the impact on PDI inhibition. Notably, this is also the first
exploration of the reactivity and regioselectivity of a variety of hydroxy bicyclic heterocycles in
Betti-style reactions and the biological activity of their products. Compounds were classified into
three types based on the substitution pattern on the phenolic moiety. Type I compounds have a
similar substitution pattern as CD343 (Table 5). The results indicated that 8-hydroxyquinazoline
(
5c-d), 5-hydroxyquinoxaline (5e), 4-hydroxybenzothiazole (5f), 7−hydroxybenzofuran (5g),
−hydroxy−1,4−benzodioxane (5h), 4−hydroxybenzo[d][1,3]dioxole (5i), and 5-
hydroxyisoquinoline (5j) are all well tolerated. 5h has the highest potency with an IC value of
5
5
0
2
90 ± 120 nM. The incorporation of 4-indanol (5k) led to a complete loss of the potency, indicating
a H−bond acceptor might be essential. The fact that 5l lost activity despite containing a H−bond
acceptor suggests a bicyclic structure is necessary to fulfill the steric requirement around this
moiety. Using 5g as a model compound, different substituents were introduced. All substituents
were well−tolerated, and 5p and 5q exhibited IC values below 0.2 µM. For Type II compounds
5
0
with a similar substitution pattern as AS15 (Table 6), both 5-hydroxybenzo[d][1,3]dioxole and 6-
hydroxy-1,4-benzodioxane are preferred, with 6-hydroxyindoline leading to a moderate loss of
activity (6d). Similar to 5k in the CD series, compounds 6f and 6g completely lost activity. Type
III compounds with substitution patterns different from AS15 and CD343 were also synthesized
by incorporating different phenolic moieties (Table 7). The majority of the compounds lost their
16
ACS Paragon Plus Environment

Page 17 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity, suggesting an unfavorable binding mode. The exceptions included compounds containing
a 5-hydroxyindole or 5-hydroxyindazole, with the latter leading to comparable inhibition of PDI
at sub−micromolar IC values for all of its analogues. It is possible that in this case the -NH group
5
0
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
of indole/indazole forms additional interactions with PDI as a H−bond donor. Most of the
synthesized α-aminobenzylphenols were not cytotoxic at 30 µM, which was similar to the trend
observed with the NC and CD series (Tables 5−7).
Table 5. SAR of Type I Compounds
R
OH
R2
N
R1
PDI IC50
MTT
R -N-R
compound
5a
R
phenol moiety
1
2
μM)a
b
(
% inhibition at 30 μM)
(
OH
N
N
O
0
0
.39 ± 0.03
.69 ± 0.13
2,3-OMe-Ph
74 (IC50: 33 μM)
54 (IC50: 37 μM)
OH
N
N
N
5
b
4-F-Ph
4-F-Ph
4-F-Ph
4-F-Ph
4-F-Ph
4-F-Ph
OH
N
4.86 ± 2.04
1.20 ± 0.51
32
44
0
5
c
N
OH
N
N
5
d
HN
N
OH
N
N
N
N
0
0
0
.66 ± 0.13
.57 ± 0.25
.56 ± 0.13
5
e
N
OH
N
0
5f
S
OH
O
0
5g
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 80
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
OH
O
N
0
2
.29 ± 0.12
0
5
h
4-F-Ph
O
OH
N
N
N
N
N
N
N
N
N
N
O
.83 ± 1.26
0
0
5
i
4-F-Ph
4-F-Ph
O
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
OH
OH
OH
0.61 ± 0.20
> 10
5
j
N
28
0
5
k
4-F-Ph
N
>
30
5l
4-F-Ph
OH
O
O
1
1
0
.23 ± 0.23
.07 ± 0.01
.41 ± 0.11
< 0.2
0
5m
4-F-Ph
OH
OH
OH
OH
OH
O
O
O
O
O
0
5
n
2-OH,4-F-Ph
5
5
o
1
5
5
p
q
N
O
< 0.2
9
F
O
F
0
.30 ± 0.03
5
5r
3
4-CF -Ph
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity determined in U−87 MG cells.
18
ACS Paragon Plus Environment

Page 19 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 6. SAR of Type II Compounds
R
R1
N
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
R2
HO
PDI IC50
(μM)a
MTT
R -N-R
compound
6a
R
phenol moiety
1
2
b
(
% inhibition at 30 μM)
O
N
O
6
6 (IC50: 32 μM)
2,3-OMe
1.06 ± 0.31
2.9 ± 2.4
O
HO
O
N
N
7c
1
6
b
H
4-F
O
HO
HO
HO
O
1
0
6
c
0.67 ± 0.29
O
H
N
N
6
d
4-F
7.01 ± 1.39
0.51 ± 0.13
> 30
O
O
N
O
F
F
4
0
6
e
2,3-OMe
4-F
HO
N
N
13
0
6
f
HO
6
g
4-F
> 100
HO
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
c
at least three independent experiments. MTT cytotoxicity determined in U−87 MG cells. MTT cytotoxicity determined in
A−172 cells.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 80
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
Table 7. SAR of Type III Compounds
R
N
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
MTT
PDI
compound
R
phenol moiety
a
(% inhibition at 30
IC (μM)
50
b
μM)
N
1
0
7
a
4-F
> 30
HO
N
S
> 100
1
7
b
4-F
4-F
HO
N
N
> 100
0
0
7
c
HO
N
10.20 ± 2.88
7
d
4-F
HO
N
>
100
0
0
7
e
4-F
4-F
HO
NH
1
.81 ± 0.06
7
f
HO
N
NH
0.35 ± 0.14
0.71 ± 0.10
87
0
7
g
4-F
HO
N
NH
7
h
4-CF
3
HO
N
NH
0
.52 ± 0.36
0
7
i
4-NO
2
HO
N
NH
N
S
0.34 ± 0.26
0
7
j
HO
20
ACS Paragon Plus Environment

Page 21 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
N
NH
0
.29 ± 0.13
.83 ± 0.29
0
0
7
k
N
HO
HO
N
NH
N
N
0
7
l
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
aInhibition of PDI was assessed by PDI reductase assay. IC50 values are indicated as the mean ± SD (standard deviation) of
b
at least three independent experiments. MTT cytotoxicity determined in U−87 MG cells.
AS15 Analogues Covalently Bind to PDI. We addressed the importance of the phenolic hydroxy
and amine moieties of the series with two compounds; 8 lost activity with the replacement of the
phenolic hydroxy (5g) with a methoxy (Figure 3A). Similarly, replacing the piperidine in 6b with
a cyclohexane abolished its PDI inhibition (13) (Figure 3B). Both compounds demonstrated the
importance of the α-aminobenzylphenol core to the potency of this series of compounds.
Due to the nature of disulfide bonds, we hypothesized that the AS15 analogues may be
attacked by the nucleophilic cysteine thiols in the active sites of reduced PDI, as is the case for
other electrophilic PDI inhibitors. Upon a reaction that exposes the Michael acceptor, the free thiol
on PDI could react and form a covalent adduct (Figure 3C). This type of retro Michael addition
1
9
to protein thiols has been observed with hydroxyquinolines like CD343. A zinc−dependent
19
mechanism opened a quinone methide for selective reaction with HDAC5 and HDAC9 , another
3
1
quinone methide intermediate was found to react with protein thiols over forming DNA adducts ,
and co−crystallization confirmed pyridinylmethyl quinoline fragment binding to MIF tautomerase
2
0
via a proline residue. In the case of the pyridinylmethyl quinoline fragment, the compound bound
via a retro Michael addition reaction that formed the quinone methide intermediate. This
intermediate was primed to undergo the aza−Michael addition to covalently link to a proline
residue in MIF tautomerase.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 80
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
To confirm AS15 analogues covalently label PDI, we incubated 10 μM PDI with 100 μM
DTT and 100 μM 6a or 6f and monitored adduct formation with quadrupole time−of−flight mass
spectrometry (QTOF). We found that the fragment matching the proposed mechanism of action
was apparent rapidly after the compound was added (Figure 3D). 6a bound reduced PDI at three
sites. Recombinant PDI contains six cysteines, four of which form disulfide bonds when oxidized.
When 6a was incubated in a mixture of 1:1:1 PDI:GSTO1:GRP78, the mass of PDI increased by
two equivalents of the fragment, demonstrating that 6a can bind PDI in the presence of competing
proteins (Figure 3E, Figure S3A). An inactive analogue of 6a without the aromatic benzoxole
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
(6f) did not demonstrate covalent binding to PDI under the same conditions (Figure S3B).
To assess the covalent binding nature, we measured the kinact/K of the lead compounds in
I
the PDI reductase assay with reduced PDI. For covalent inhibitors, the kinact/K is the ratio of the
I
observed rate of inactivation after a reversible reaction to form a protein−inhibitor (P−I) complex
with all the protein molecules (kinact) to the concentration of inhibitor required to reach half of the
maximum rate of covalent bond formation (K ). The kinetics of covalent PDI inhibitors 16F16,
I
PACMA31, and AS15 were measured by assessing activity in the PDI reductase assay at
3
−1 −1
incubation times from 5 to 60 minutes. AS15 inhibited PDI with a kinact/K of 2.6 × 10 M s
I
2
−1 −1
2
(
Figure 3F). kinact/K for PACMA31 was 2.0 × 10 M s and k /K for 16F16 was 1.7 × 10
I inact I
−
1 −1
M s . Thus, AS15 was more efficient at inhibiting PDI than both covalent inhibitors PACMA31
and 16F16. Furthermore, a gel−based competition assay with the fluorescent probe of PACMA31,
PACMA57, confirmed AS15 analogues could compete with PACMA57 to bind PDI (Figure
12
3
G). Because PACMA57 binds PDI at either Cys397 or Cys400, this suggests that AS15 can
also bind one of these cysteines. Inactive analogue 6f did not compete with PACMA57 for binding
to PDI.
22
ACS Paragon Plus Environment

Page 23 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Structure−activity relationship indicates that compounds are covalent PDI inhibitors. (a)
Comparison of compounds with free hydroxy or methoxy on the bicyclic moiety. (b) Comparison of
compounds with piperidine or cyclohexane moiety. (c) Proposed mechanism of inhibition via retro−Michael
addition reaction. (d) 6a covalently binds to PDI. 100 μM 6a was incubated with 10 μM PDI and 100 μM
DTT for 30 min prior to injection. (e) 6a binds to PDI preferentially over GSTO1. 100 μM 6a was incubated
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 80
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
with a mixture of 10 μM PDI, 10 μM GSTO1, and 10 μM GRP78 as a non-specific binding control for 30 min
prior to injection. (Close-up of GRP78 trace in Supplementary Figure S3.) (f) Concentration− and
time−dependent PDI inhibition curves for kinact/K determination of PACMA31, 16F16, and AS15. Activity
I
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 measured using the PDI reductase assay. Absorbance was monitored over time at various
concentrations and preincubation times with indicated compounds. (g) Gel−based competition with
recombinant PDI, PDI inhibitors or inactive analogue 6f at 100 μM, and PACMA57.
To address whether the AS15 analogues targeted PDI irreversibly, like PACMA31, we
performed washout experiments with the PDI reductase assay. AS15 (50 μM) was incubated with
4
0 μM PDI for 3 hours at room temperature. After 3 hours, the PDI−AS15 complex was diluted
1
00−fold into reaction buffer, the reaction was incubated another hour at 37 °C, and insulin was
added as a substrate to initiate the reaction. PACMA31 (an irreversible inhibitor) at 1 and 100 μM
Figure 4A) and BAP2 (a reversible inhibitor) at 0.5 and 50 μM (Figure 4B) were used as controls.
(
We found that both AS15 and CD343 did not maintain the characteristics of the high concentration
of inhibitor after dilution to the low concentration (Figures 4C−D). However, when we tested
enzyme activity recovery with compound concentrations 10 times that of the protein to mimic the
mass spectrometry conditions, AS15 behave like the expected irreversible inhibitor (Figure S4).
CD343 and 6a were too active at this ratio. We further tested whether the compounds were binding
in the b′ domain of oxidized PDI with the ANS (anilinonaphthalene sulfonic acid) spectral scan.32
ANS is a dye that fluoresces upon binding hydrophobic pockets and specifically targets the b′
domain in PDI. B′ domain−selective inhibitors of PDI such as estradiol and bepristat 1a compete
3
2
with ANS. AS15 and CD343 did not lower the fluorescence of ANS (Figure 4E). Our combined
24
ACS Paragon Plus Environment

Page 25 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
results from the thermal shift assay, washout experiment, and ANS spectral scan demonstrated that
AS15 and CD343 are likely not substrate−binding domain inhibitors like estradiol and BAP2, but
may be covalent inhibitors of PDI.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Recovery of PDI activity upon treatment with PACMA31 (a), BAP2 (b), AS15 (c), or CD343 (d).
For the diluted samples (purple curves), 40 μM PDI was incubated with 100 μM PACMA31, 50 μM BAP2,
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 80
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
5
0 μM AS15, or 50 μM CD343 for 3 h at rt before dilution. (f) ANS spectral scan with 5 μM PDI and 100 μM
PACMA31, BAP2, AS15, or CD343.
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
AS15 Analogues Compete with Glutathione. Because the AS15 analogues seemed highly
reactive toward nucleophilic attack, we hypothesized that glutathione may also react with the
compounds. Glutathione is present in high concentrations in the cytoplasm and is an important
33
redox regulator in the ER. The oxidizing environment of the ER is maintained by the ratio of
reduced to oxidized glutathione, which is lower than the ratio in the cytoplasm. Incubating the
compounds with N−acetyl cysteine (NAC) before adding them to the PDI reductase assay rendered
the compounds inactive (Figure 5A). Although both the AS15 series and PACMA31 are
irreversible PDI inhibitors, the differing characteristics in the NAC competition assay, as well as
differing kinetics, suggest that the compounds may have different mechanisms of action. AS15
analogues were also inactivated with competing glutathione at physiological concentrations (5
mM) in the PDI reductase assay (Figure 5B). The compounds retained activity when incubated
with PDI before adding GSH or NAC, providing further evidence of potential thiol reactivity
(Figure S5).
To assess trends in the glutathione sensitivity of the AS15 series, we screened all the
analogues in the PDI reductase assay in the presence of 5 mM GSH. The high concentrations of
glutathione in the cytoplasm may inactivate the compounds in vivo; thus, the in vitro IC value
5
0
1
4
may not be a reliable indicator of in vivo activity. Although the PDI reductase assay is performed
at a relatively high concentration of DTT (500 μM), we added 5 mM glutathione to mimic a more
physiological environment (Table S2). This issue is particularly relevant because the analogues
26
ACS Paragon Plus Environment

Page 27 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
behave as reversible thiol adducts. Substrate−binding domain inhibitors isoquercetin and BAP2
remained active in the presence of high GSH concentrations. Most of the AS15 analogues were
inactive in the presence of competing GSH. These results provided further support that the AS15
analogues may act via addition to PDI. We found two AS15 analogues that maintained potency in
the presence of competing glutathione: 14 and 5d (Figure 5C).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. N-Acetyl cysteine and glutathione inactivate AS15 analogues. (a) N-Acetyl cysteine (NAC)
competition in PDI reductase assay. Compounds were incubated with NAC prior to addition to a final
concentration of 10 μM in the PDI reductase assay with PDIA1. (b) Glutathione competition in PDI reductase
assay. GSH was added to the PDI reaction buffer prior to 10 μM compound addition. (c) Top two AS15
analogues that are least sensitive to competition with 5 mM GSH in the PDI reductase assay.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 80
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
After our GSH screen, we retested the purity of 14 and 5d and observed that 5d
spontaneously formed a dimer after long−term storage. As a result, we re-purified the monomeric
and dimeric forms of 5d and tested each form in the PDI reductase assay in the presence of
glutathione (Figure 6A). The dimeric form of 5d was less sensitive to glutathione competition
than the monomer (Figure 6B). Furthermore, we found that both the monomer and dimer of 5d
bound to PDI (Figure 6C). Incubation with the a′c domain gave a species with two fragments 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
d monomer bound (Figure S6A). When we incubated the monomer and dimer of 5d with a C53S
mutant of PDI, we observed one and two species bound, respectively (Figure S6B).
3
4
Although the PDI disulfides are 500−fold more reactive than glutathione , we
hypothesized that glutathione may be inactivating the AS15 analogues and contributing to lower
their potency in cells. Pretreatment of GBM cells with BSO for 24 hours before adding the
compounds increased potency in the colony formation assay (Figure 6D−H, Figure S7). While
the monomer of 5d was more sensitive to BSO addition, the 5d dimer was more potent and its
potency was not dependent on BSO addition. These results support the hypothesis that glutathione
depletion sensitizes GBM cells to PDI inhibition.
28
ACS Paragon Plus Environment

Page 29 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Dimerized analogue is less sensitive to GSH treatment. (a) Structure of 5d dimerization. (b) Percent
inhibition of 5d monomer and dimer in the PDI reductase assay in competition with 5 mM glutathione. (c) Protein
mass spectrometry confirms fragment of 5d binds to PDI. Toxicity of PACMA31 (d), AS15 (e), CD343 (f), 5d
29
ACS Paragon Plus Environment