Discovery of 2-((4-resorcinolyl)-5-aryl-1,2,3-triazol-1-yl)acetates as potent Hsp90 inhibitors with selectivity over TRAP1

Article abstract of DOI:10.1016/j.bmcl.2019.126809

As the most abundant heat shock protein (HSP), Hsp90 is actively involved in tumor cell growth and various responses to anti-carcinogenic stress. Hsp90 has thus emerged as a potential drug target. A structure-based drug design approach was applied to develop novel resorcinolyltriazole derivatives as Hsp90 inhibitors. Structure-activity relationships (SARs) and molecular docking were investigated to provide a rationale for binding affinity and paralog selectivity. Click chemistry between iodoethynylresorcinol and an azido derivative was used to synthesize a new family of 2-((4-resorcinolyl)-5-aryl-1,2,3-triazol-1-yl) acetates that exhibited Hsp90 binding affinities of 40–100 nM (IC₅₀). Among the synthesized molecules, the triazole alkyl acetates displayed the highest Hsp90 binding affinities. Their potency against Hsp90 was over 100-fold stronger than against TRAP1 and 1–3-fold stronger than against Grp94. In particular, compounds 18, 19, and 30 had Hsp90 inhibitory activities of ~45 nM (IC₅₀) and they displayed over 350-fold selectivity for Hsp90 over TRAP1.

Full text of DOI:10.1016/j.bmcl.2019.126809

Journal Pre-proofs
Discovery of 2-((4-Resorcinolyl)-5-aryl-1,2,3-triazol-1-yl)acetates as potent
Hsp90 inhibitors with selectivity over TRAP1
Sejin Jung, Nam Gu Yoon, Sujae Yang, Darong Kim, Won Seok Lee, Ki Bum
Hong, Changwook Lee, Byoung Heon Kang, Ji Hoon Lee, Soosung Kang
PII:
S0960-894X(19)30778-4
DOI:
https://doi.org/10.1016/j.bmcl.2019.126809
BMCL 126809
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
16 September 2019
29 October 2019
4 November 2019
Please cite this article as: Jung, S., Gu Yoon, N., Yang, S., Kim, D., Seok Lee, W., Bum Hong, K., Lee, C., Heon
Kang, B., Hoon Lee, J., Kang, S., Discovery of 2-((4-Resorcinolyl)-5-aryl-1,2,3-triazol-1-yl)acetates as potent
Hsp90 inhibitors with selectivity over TRAP1, Bioorganic & Medicinal Chemistry Letters (2019), doi: https://
doi.org/10.1016/j.bmcl.2019.126809
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Ltd.

Discovery of 2-((4-Resorcinolyl)-5-aryl-1,2,3-
triazol-1-yl)acetates as potent Hsp90 inhibitors with
selectivity over TRAP1
Sejin Junga,†, Nam Gu Yoonb,†, Sujae Yangc, Darong Kima, Won Seok Leea, Ki Bum Honga,
Changwook Leeb, Byoung Heon Kangb,*, Ji Hoon Leea,*, Soosung Kangc,d,*
aNew Drug Development Center, Daegu-Gyeongbuk Medical Innovation Foundation (DGMIF),
Daegu 41061, South Korea
bDepartment of Biological Sciences, Ulsan National Institutes of Science and Technology
(UNIST), Ulsan 44919, South Korea.
cCollege of Pharmacy and Graduate School of Pharmaceutical Sciences, Ewha Womans
University, Seoul 03760, South Korea
dWestern Seoul Center, Korea Basic Science Institute, Seoul 03760, South Korea
† These authors contributed equally.
*Corresponding Authors|
B.H. Kang: Tel. 82-52-217-2521, email kangbh@unist.ac.kr;
J.H. Lee: Tel. 82-53-790-5233, email jhlee@dgmif.re.kr;
S. Kang: Tel. 82-2-3277-6619, email sskang@ewha.ac.kr.
KEYWORDS: Hsp90, TRAP1, Grp94, inhibitor, selectivity, resorcinol, triazole.
1

ABBREVIATION; TRAP1: TNF receptor associated protein 1; Hsp90: heat shock protein 90;
Grp94: Glucose-regulated protein 94; NTD: N-terminal domain; MD: middle domain; CTD: C-
terminal domain; CADD: computer assisted drug design; TBATB: tetrabutylammonium
tribromide; DIPEA: diisopropylethylamine
ABSTRACT: As the most abundant heat shock protein (HSP), Hsp90 is actively involved in
tumor cell growth and various responses to anti-carcinogenic stress. Hsp90 has thus emerged as a
potential drug target. A structure-based drug design approach was applied to develop novel
resorcinolyltriazole derivatives as Hsp90 inhibitors. Structure-activity relationships (SARs) and
molecular docking were investigated to provide a rationale for binding affinity and paralog
selectivity. Click chemistry between iodoethynylresorcinol and an azido derivative was used to
synthesize a new family of 2-((4-resorcinolyl)-5-aryl-1,2,3-triazol-1-yl) acetates that exhibited
Hsp90 binding affinities of 40–100 nM (IC50). Among the synthesized molecules, the triazole
alkyl acetates displayed the highest Hsp90 binding affinities. Their potency against Hsp90 was
over 100-fold stronger than against TRAP1 and 1~3-fold stronger than against Grp94. In
particular, compounds 18, 19, and 30 had Hsp90 inhibitory activities of ~45 nM (IC50) and they
displayed over 350-fold selectivity for Hsp90 over TRAP1.
2

Cellular expression of molecular chaperones known as heat shock proteins (HSPs) increases in
response to various stressors. HSPs perform important cellular functions, such as protein
assembly, secretion, transportation, translocation, and degradation.[1] They also regulate gene
expression via direct interaction with transcription factors.[2] HSPs are classified by size into the
Hsp110, Hsp90, Hsp70, Hsp60, Hsp40, and Hsp27 families. HSPs are usually ATP-dependent
chaperones. However, small HSPs, such as Hsp27, are not ATP-dependent.[3]
Among the various cytosolic HSPs, Hsp90 is the most abundant in both eukaryotic and
prokaryotic cells. It is rapidly deployed in response to various types of stress and affords basic
protection against detrimental protein misfolding and aggregation.[4] Since Hsp90 is over-
expressed to protect the functionality of numerous oncoproteins,[5],[6] it is actively involved in
multiple tumorigenic pathways. These include cell proliferation, survival, angiogenesis, invasion,
and metastasis.[7],[8],[9] Furthermore, Hsp90 chaperone machinery has been associated with
neurodegenerative disease, cystic fibrosis, psychiatric illness, and infectious disease.[10] For
these reasons, Hsp90 has become an important therapeutic target for drug development2, and
several Hsp90 inhibitors have been developed over the past decade as therapeutic cancer
drugs.[11],[12] Currently available Hsp90 inhibitors[13] are broadly categorized as
geldanamycin analogues, resorcinol derivatives, 2-aminobenzamides[14], or purine-type
molecules based on their primary scaffolds (Figure 1). 17-AAG[15] (1) is a representative
geldanamycin[16] derivative.[17],[18],[19] AUY-922[20] (2) and ganetespib[21] (3) are the most
effective resorcinol derivatives[22] at the cellular level. The purine derivatives BIIB021[23] (4)
and PU-H71[24] (5) have either been shelved or are undergoing clinical trials. Although
approximately 70 clinical trials have been performed or are underway with Hsp90 inhibitors as
single agent or in combination with other chemotherapeutic cancer drugs
3

(https://clinicaltrials.gov/), the observed anticancer activity of Hsp90 inhibitors has been modest.
The clinical trials were analyzed in a recent study,[25] which indicated that insufficient target
validation and non-specific targeting effects were common issues. Feasible approaches were
suggested to improve target validation and the determination of off-target effects during drug
development. Further development of Hsp90 inhibitors, improvements in patient stratification in
clinical trials, and alternative dosing strategies have also been suggested to fully realize the
therapeutic potential of Hsp90 inhibitors.[26],[27] Each of these developments prompted us to
explore additional Hsp90 molecules.
Geldanamycin derivative
Resocinol derivatives
Pyrine derivatives
O
H
H
NH
O
N
O
OH
N
HN
N
N
OH
Cl
N
O
O
N
N
N
N
N
N
N
H
O
OH
N
S
I
OH
H2N
N
N
MeO
MeO
NH2
N
N
O
O
OH
OMe
O
O
O
NH2
3
5
PU-H71 ( )
Ganetespib ( )
1
2
AUY922 ( )
17-AAG ( )
4
BIIB021 ( )
Figure 1. Representative Hsp90 inhibitors.
The Hsp90 family includes Hsp90α/β, TRAP1, and Grp94 paralogs.[28] TRAP1 is found in
mitochondria[29] and Grp94 is found in the endoplasmic reticulum.[30] Hsp90α/β proteins are
cytoplasmic chaperones, and each paralog in the Hsp90 family is associated with its own subset
of client proteins. Thus, increasing attention is being given to paralog-specific inhibition[31],[32]
to understand the individual pharmacology[33] of each chaperone and the cellular pathways of
oncologic protein stabilization associated with it, as each chaperone affects the activities of over
200 client proteins. [34] However, the discovery of Hsp90 paralog-specific inhibitors is
4

challenging. This is because most of the developed inhibitors target the ATP binding site, which
is highly conserved throughout the Hsp90 family. The majority of reported Hsp90 inhibitors are
thus pan-Hsp90 inhibitors,[35] including PU-H71 and ganetespib. On the other hand, recent co-
crystallographic studies have shown that some molecules have adapted to binding with specific
paralogs. They have different binding affinities and exhibit characteristic binding poses with
TRAP1,Error! Bookmark not defined. Grp94,[36] and Hsp90.[37] These binding modes have
provided novel insights for the design of paralog-selective molecules.
Hsp90 is a homodimer that contains three identical domains. These are the N-terminal domain
(NTD), the middle domain (MD), and the C-terminal domain (CTD). ATP binding on the NTD
results in conformational changes via ATP lid and N-strap restructuring, which induce transient
N-terminal domain dimerization.[38],[39] These structural changes are essential for ATP
hydrolysis and Hsp90 target binding for the chaperone cycle.Error! Bookmark not defined. In
structure-based isoform-selective drug development using computer-assisted drug design
(CADD), it is crucial to choose the appropriate three-dimensional (3D) conformation from
among several target protein sub-structures. This is because the drug binding site on its NTD
changes when the shape of the entire Hsp90 protein undergoes a conformational change. While
analyzing the published X-ray co-crystal structures of Hsp90 with various ligands, we found that
about 300 of the ligands complexed with Hsp90 were currently in the Protein Data Bank.
Interestingly, approximately 30% of the ligands were resorcinol derivatives, and all of the
resorcinol co-structures formed in the Hsp90 NTD segment with an open ATP lid (Figure 2).
5

Figure 2. ATP-lid open state (a) and closed state (b). 2D-skematic view of AUY-922 binding at
Hsp90 with open ATP-lid (c).
We previously reported that preferential binding to Asn171, which corresponds to the Asn95
residue of Hsp90, is necessary to induce TRAP1-specific modification.Error! Bookmark not
defined. While the study provided the rationale for TRAP1-selective inhibition and described the
binding mode of purine-based inhibitors in detail, paralog-specific binding of resorcinol
scaffolds is less understood. Based on the 3D Hsp90 structures, we questioned whether the
Asn95 residue of Hsp90 could be utilized for Hsp90-specific inhibition when resorcinol scaffolds
were applied. We focused on a resorcinolyltriazole[40] scaffold with various acetate substituents
on the third nitrogen atom of its triazole ring.
Scheme 1. Preparation of the common intermediates 13a-b, 14a-b, and 15a-ba
6

c for 8b
O
R1
R1
O
O
O
O
d
b for 8a
e
a
Br
O
O
6
O
7
O
R1
9a,b
8a iPr
8b Et
R1
R1
R1
R1
MeO
O
O
I
O
g, h
f
i
R2
N
N
I
N
OMe
O
SiMe3
O
I
O
O
10a,b
R2: OMe OEt NH2
13a,b 14a,b 15a,b
11a,b
12a,b
aReagents and condition: a) Methyltriphenylphosphonium bromide, n-BuLi, 0 °C, THF, 16 h,
42%; b) Pd/C, H2, EtOH, 17 h, 86-92% ; c) Pd/C, EtOH, HCl, 5 h, 94%; d) TBATB, CHCl3, 2 h,
59-82%; e) tBuLi, I2, Et2O, -78 °C to rt, 1 h, 80-82%; f) Ethynyltrimethylsilane, Pd(PPh3)2Cl2,
CuI, TEA, Toluene, 45 °C, 17 h, 51-61% ; g) 1N NaOH, MeOH-CHCl3, 2 h, rt, 89-92%; h) I2,
morpholine, Toluene, 45 °C, 24 h, 37-42%; i) alkyl 2-azidoacetate or 2-azidoacetamide, CuI,
TEA, THF, 15 h, 47%.
Common intermediates for 5-iodo-1,2,3-triazoles (13a-b, 14a-b, 15a-b) were prepared from
dimethoxyacetophenone primarily via copper-catalyzed click chemistry[41] (Scheme 1). The
Wittig reaction between the carbonyl group on dimethoxyacetophenone (6) and t-BuLi-treated
methyltriphenylphosphonium bromide at 0 °C following hydrogenation over Pd/C enabled the
introduction of an isopropyl group (R1) to 8a. An ethyl group was obtained directly on R1 (8b)
following reduction over Pd/C in acidic conditions. Bromine was introduced to the resorcinol
scaffold (9a-b) using tetrabutylammonium tribromide. The bromine atom was then replaced with
iodine via treatment with t-BuLi and I2. An acetylene group was introduced at the iodine position
via the Sonogashira coupling reaction using trimethylsilylacetylene and Pd(PPh3)2Cl2 catalyst.
Prior to forming the triazole backbone, iodine was introduced to the terminal acetylene (12 a-b)
to enable subsequent coupling with an aromatic ring. Cu-catalyzed [3+2] cycloaddition of 12a-b
7

was performed with various 2-azidoacetates to yield the desired 1,2,3-triazole derivatives 13a-b,
14a-b, and 15a-b as key intermediates.
Scheme 2. Preparation of compounds 17a-da
O
O
S
MeO
MeO
HO
I
b, c
a
N
N
N
N
OMe
12a
I
OMe N
16
OH
N
17a
b, c, d
d
O
O
O
S
O
HO
HO
HO
NH
N
NH
N
NH
N
N
OH
N
OH
N
OH
17b
17c
17d
aReagents and condition: a) Azidomethylbenzene, CuI, DIPEA, NBS, CH2Cl2, THF, rt, 3 h, 56%;
b) ArB(OH)2, PdCl2(dppf)•CH2Cl2, K2CO3, DMF, H2O, microwave, 100 °C, 30min; c) BBr3,
CH2Cl2, 40 °C, 12 h; d) Pd/C, H2, MeOH, 60 °C, 12h, 50-70%.
Compounds 17a-d, which contained 5-isopropylresorcinol, were synthesized via the
iodotriazole (16) by the copper-catalyzed azide alkyne [3+2] cycloaddition of 12a and a benzyl
azide (Scheme 2). Palladium-catalyzed Suzuki coupling of 16 with various arylboronic acids was
performed using a microwave reactor. Benzyl and methyl deprotection using Pd/C in a hydrogen
atmosphere or an anhydrous solution of BBr3 in dichloromethane yielded the final products (17a-
d).
Synthesis of compounds 18-43 was carried out via the Suzuki coupling of 13a-b, 14a-b, or
15a-b to the appropriate arylboronic acid or its pinacol ester using Pd(ddpf)Cl2•CH2Cl2 in the
presence of Cs2CO3. The two methyl protecting groups on the resorcinol were removed to yield
8

the desired resorcinolyltriazole products by adding BBr3 at 40 °C (Scheme 3). A portion of the
ester underwent hydrolysis during methyl deprotection and provided acid derivatives 26-29.
Scheme 3. Preparation of compounds 18-43a
R1
R1
Ar
MeO
HO
a, b
I
R2
N
N
R2
N
N
N
OMe
N
OH
O
O
13a,b, 14a,b, 15 a,b
18-43
O
O
O
O
Ar:
S
O
O
N
S
S
N
H
O
O
OH
OH
O
N
N
O
aReagents and condition: a) R2B(OH)2, Pd(ddpf)Cl2 •CH2Cl2, Cs2CO3, DMF, H2O, microwave,
100 °C, 30 min; b) BBr3, CH2Cl2, 40 °C, 2 h
The synthesized compounds were tested in an in-vitro fluorescence polarization assay with
recombinant Hsp90α as described previously.Error! Bookmark not defined. As shown in Figure 3 and
Tables 1 and 2, all of the synthesized resorcinolyltriazoles inhibited Hsp90α and had IC50 values
in the sub-micromolar range. The isopropylresorcinol derivatives had higher Hsp90 binding
affinities than any of the ethylresorcinol derivatives. The alkyl acetate group on the triazole ring
appeared to be essential for highly effective Hsp90 binding. Eleven of the twelve most active
compounds (17a, 18-24, and 30-33) contained a methyl or ethyl acetate functionality on the
nitrogen atom in the triazole ring. Compounds 17a, 18-24, and 30-33 displayed Hsp90 binding
potencies that were similar to or higher than that of the positive control from the purine series,
PU-H71 (5). The potencies of compounds 17a, 18, 19, 21, and 30 were even higher than that of
9

ganetespib (3). Ganetespib is a pan-Hsp90 inhibitor that is safer and more effective than other
Hsp90 inhibitors.[42] It is worth noting that a sulfone group on the phenyl ring enhances binding
affinity (17a, 18, 19, 30, 31). The simple triazoles (17b-d), which lacked an additional
functionality on the nitrogen atom in the triazole ring, had limited Hsp90 binding affinities. The
triazolylacetic acids (26-29) and triazolylacetamides (34-35) displayed comparatively low Hsp90
binding affinities.
Figure 3. Fluorescence polarization assay with Hsp90α. Chemical inhibitors showing better
IC50 values than that of PU-H71 are indicated with blue bars. N=3, Mean ± SEM.
Table 1. Structures and inhibitory activity evaluation of synthesized compounds 17a-d, 18-35
Ar
Ar
Ar
HO
HO
HO
R2
N
N
N
N
NH
N
N
Bn
N
OH
N
OH
OH
O
17b-d
17a
18-35
Hsp90α IC50 (nM)
n=3, mean±SEM
Drug #
R1
R2
Ar
PU-H71
126.2 ± 1.62
50.45 ± 1.00
Ganetespib
O
O
S
Me
17a
17b
17c
45.70 ± 1.50
295.85 ± 13.14
168.84 ± 4.82
Ph
O
O
S
Me
O
17d
166.22 ± 3.75
O
10

O
O
O
S
S
Me
18
19
20
iPr
iPr
iPr
OMe
OMe
OMe
43.27 ± 0.89
43.12 ± 0.75
78.32 ± 0.94
Me
O
O
O
S
N
Me
H
O
21
22
23
24
25
iPr
iPr
iPr
iPr
iPr
OMe
OMe
OMe
OMe
OMe
47.85 ± 1.63
57.05 ± 2.22
90.95 ± 1.32
71.16 ± 1.53
153.47 ± 0.63
O
N
N
O
O
N
OH
OH
O
O
S
Me
26
iPr
OH
186.74 ± 4.64
Me
27
28
29
30
iPr
iPr
iPr
iPr
OH
OH
OH
OEt
173.14 ± 4.84
767.19 ± 40.77
175.59 ± 5.42
41.80 ± 0.45
S
O
O
O
O
S
N
Me
H
O
O
O
O
S
Me
Me
O
31
32
33
34
35
iPr
iPr
iPr
iPr
iPr
OEt
OEt
OEt
NH2
NH2
53.10 ± 3.54
77.01 ± 1.81
60.93 ± 3.48
135.50 ± 2.21
133.78 ± 5.23
S
O
O
O
S
N
Me
H
O
O
O
O
S
Me
O
O
Table 2. Structures and inhibitory activity evaluation of synthesized compounds 36-43
Ar
HO
R2
N
N
OH
N
O
Hsp90α IC50 (nM)
n=3, mean±SEM
Drug #
36
R1
Et
R2
R3
O
O
S
Me
OMe
214.20 ± 13.45
Me
37
38
39
Et
Et
Et
OMe
OMe
OMe
154.29 ± 9.19
183.59 ± 5.53
172.79 ± 11.27
S
O
O
O
O
S
O
O
N
Me
H
11

O
O
S
S
Me
40
Et
OEt
304.19 ± 32.34
Me
41
42
43
Et
Et
Et
OEt
OEt
NH2
161.45 ± 3.15
178.29 ± 3.66
469.82 ± 11.86
O
O
O
N
O
O
S
Me
Compounds 17-24 and 30-33, which were more potent than PU-H71 in the Hsp90 inhibition
assay, were further analyzed to examine their inhibitory activities against Grp94 and TRAP1. All
of the tested compounds were either moderately good or strong Grp94 inhibitors. Compounds 19
and 22 had IC50 values of ~50 nM. In contrast, the resorcinolyltriazole compounds were poor
TRAP1 inhibitors with IC50 values exceeding 5 μM. The data indicated the resorcinolyltriazoles
bound tightly to Hsp90 and Grp94, but not TRAP1.
Table 3. Hsp90α paralog selectivity of selected compounds. IC50 values (nM) were calculated
and compared.
IC50 ± SD (nM)
TRAP1
IC50 Ratio
TRAP/Hsp90α Grp94/Hsp90α
Drug #
Grp94
Hsp90α
PU-H71
126.2 ± 1.62
300.6 ± 37.56
300.6 ± 37.56
2.4
2.4
0.7
1.7
2.1
1.3
2.4
1.6
0.9
2.2
2.5
2.4
Ganetespib
50.45 ± 1.00
45.70 ± 1.50
43.27 ± 0.89
43.12 ± 0.75
78.32 ± 0.94
47.85 ± 1.63
57.05 ± 2.22
90.95 ± 1.32
71.16 ± 1.53
41.80 ± 0.45
37.7 ± 0.29
8480 ± 105.7
15974 ± 329.9
17362 ± 124.5
24515 ± 225.1
11883 ± 252.8
6983 ± 147.6
24487 ± 256.4
14147 ± 119.6
16014 ± 284.7
37.7 ± 0.29
78.01 ± 2.20
88.98 ± 0.65
54.10 ± 2.37
190.22 ± 6.70
77.96 ± 2.41
49.42 ± 0.01
197.13 ± 3.70
174.80 ± 2.88
100.46 ± 7.01
0.7
17a
18
19
20
21
22
23
24
30
185.6
369.2
402.6
313.0
248.3
122.4
269.2
199.3
383.1
12

31
32
33
53.10 ± 3.54
77.01 ± 1.81
60.93 ± 3.48
35102 ± 573.0
33173 ± 171.3
25448 ± 414.4
96.20 ± 2.27
209.28 ± 5.78
192.09 ± 6.22
661.1
430.8
417.7
1.8
2.7
3.2
The most potent compounds (17a, 18, 19, 21, 30) were tested in cytotoxic activity assays and
mode-of-action studies using HeLa and PC3 cancer cells. Compounds 17a, 21, and 30 exhibited
potent cytotoxic activity in HeLa cells, and the IC50 of 17a (~206 nM) was the lowest of the
three (Table 4). 17a, 21, and 30 also exhibited higher cytotoxicity in PC3 cells than the other
compounds with IC50 values of 230–320 nM (Table 4). Cellular Hsp90 inhibition leads to Hsp70
induction and the degradation of client proteins, such as Akt and Chk1. 17a, 21, and 30 reduced
expression of the Hsp90 client proteins Akt and Chk1, which was consistent with the MTT data.
They also increased the expression of Hsp70 to a greater extent than 18 and 19 in both the HeLa
and PC3 cell lines (Figure 4). Although 18 and 19 were good Hsp90 inhibitors in the in-vitro
binding assay, they did not reduce HER2, AKT, or Chk1 levels in the HeLa cells. This might
have been due to different intracellular permeation or distribution of 18 and 19 in the HeLa cells.
For example, they may have accumulated in cellular compartments rather than in the cytosol.
Table 4. Cytotoxic activity against cancer cells. IC50 values (nM) were calculated from MTT
assay.
HeLa
IC50 ± SD (nM)
PC3
IC50 ± SD (nM)
Drug #
PU-H71
226.5 ± 12.99
11.65 ± 2.71
206 ± 14.26
>2000
363.0 ± 37.73
9.3 ± 2.89
Ganetespib
17a
18
285.9 ± 26.52
459.7 ± 50.11
569.0 ± 48.11
234.5 ± 21.90
19
>2000
21
499.7 ± 37.05
13

30
402 ± 33.20
323.7 ± 31.28
Figure 4. Client degradation. HeLa cells (a) and PC3 cells (b) were incubated with the drugs (1
μM) as indicated for 48 hours and analyzed by western blotting.
Our docking analysis showed that resorcinol hydroxyl group on 30 participated in direct or
water-mediated hydrogen bonding with the carboxylate group on Asp93 (Figure 5a). The
resorcinol isopropyl group in 30 was directed towards the large hydrophobic pocket formed by
Met98, Leu107, Ile110, Phe138, and Val150. These are common and well established binding
configurations of resorcinol derivatives as Hsp90 inhibitors. It has also been suggested that one
of the nitrogen atoms in the triazole ring of the synthesized molecules is involved in direct
hydrogen bonding with Thr184 and a water-mediated interaction with Asp93. The carboxyl
group on the triazole acetate moiety of the synthesized compounds is a distinct structural motif
that may contribute to the stability and selectivity of Hsp90 binding. In our docking model, it
was oriented towards Lys58 and Asp102 to form a polar-polar interaction. This additional
interaction could explain the good affinities of the methyl and ethyl ester derivatives (15-24, 30-
33). Our best compounds exhibited docking poses with Grp94 that were very similar to their
14

poses with Hsp90α/β (Figure 5c). The distinguishing acetate group on the triazole ring of Hsp90
inhibitor 30 participated in a hydrogen bond with Asn162 in Grp94, and a nitrogen atom in the
triazole ring of 30 formed a direct hydrogen bond with Thr245. The similarity between the
binding poses of 30 with Hsp90α/β and Grp94 could explain the low one- to two-fold paralog
specificity of 30 over Grp94.
The Protein Data Bank currently has TRAP1 co-crystal structures for ADP (PDB: 5F5R),
AMP (PDB: 4IVG), and DN401 (PDB: 5Y3N), which is a potent TRAP1 inhibitor. The docking
scores with the DN401-TRAP1 complex (PDB: 5Y3N) and the AMP-bound structure (PDB:
4IVG) were quite poor, and their binding poses were highly non-specific. This was because the
ATP lids in these structures were located near the drug binding site, which was interrupted by the
methylsulfone group on 30. The ATP lid of the ADP-bound structure (PDB: 5F5R) was open,
and it appeared to have sufficient room to bind with 30. However, the protein structure overlay
and docking analysis showed that the hinge of the ATP lid near Leu172 aligned slightly closer to
the other helical segment (Figure 4d). Leu172 thus occupied the previously spacious
hydrophobic pocket, which was used to bind the isopropyl group on resorcinol. This shifted one
resorcinol hydroxyl group away from its original binding pocket, which prevented the formation
of a hydrogen bond between the acetate group on triazole and Asn171. The poor binding poses
of 30 with TRAP1 could explain why its specificity for Hsp90 was ~300-fold higher than it was
for TRAP1.
According to a recent study[43], selective Hsp90 isoform inhibitor induces the degradation of
select Hsp90 clients without increasing Hsp90 levels. Thus, the development of Hsp90α/β
isoform-selective inhibitors may introduce a new paradigm for the treatment of various diseases
with minimal side effects.Error! Bookmark not defined. The study also shows that structure-
15

guided design has afforded a Hsp90β-selective inhibitor to target Hsp90β-specific residues.
However, current pan-Hsp90 inhibitors, such as BIIB021, SNX2111, AUY922, and 17-DMAG,
do not differently interact with the specific residues on Hsp90β. These inhibitors exhibit <1.5-
fold selectivity for Hsp90α or Hsp90β.[44] We also examined the molecular docking of 30 with
Hsp90β to estimate whether 30 could interact with the isoform-specific residues. But, its binding
mode and Hsp90β docking score were quite similar to those of pan-Hsp90 inhibitors (Figure 5b).
Figure 5. Proposed binding modes between 30 and Hsp90α (a), Hsp90β (b), Grp94 (c), and
TRAP1 (d). PDB code: 2XAB for Hsp90α (gray); 1UYM for Hsp90β (yellow); for Grp94
(green); 5F5R for TRAP1 (purple). Glide docking or Induced Fit Docking module in
Schrodinger Software package were used to obtain docking poses.
16

In this study, a structure-based drug design approach was employed to develop a novel
paralog-specific series of resorcinolyltriazole Hsp90 inhibitors. This approach was useful for
exploring specific interactions between molecules and proteins, even with subtle differences
between the residues in the drug binding site of the Hsp90 family. Click chemistry between
iodoethynylresorcinol and an azide afforded nearly 30 new 4,5-diaryl-substituted 1,2,3-triazoles.
All of the synthesized molecules exhibited sub-micromolar Hsp90 binding affinities. Among
them, members of the methyl and ethyl acetate-bearing triazole series displayed the highest
Hsp90 binding affinities. Their selectivity for Hsp90 was over 100 times higher than their
selectivity for the Hsp90 isoform TRAP1 and one to three times higher than their selectivity for
Grp94. The IC50 values of the Hsp90 inhibitors 18, 19, and 30 were ~45 nM, and they displayed
>350-fold selectivity for Hsp90 over TRAP1. These Hsp90 inhibitors therefore have potential for
further development as antitumor agents.
Acknowledgements
Funding: S. Kang was supported by Korea Basic Science Institute (C39532) and National
Research Foundation of Korea grant funded by MSIT (NRF-2019R1A2C2004142 and
2018R1A5A2025286). B.H. Kang was supported by National Research Foundation of Korea
(NRF) grant funded by MSIT (NRF-2018R1A5A1024340) and Korea Drug Development Fund
(KDDF) funded by MSIT, MOTIE, and MOHW (KDDF-201512-02, Republic of Korea).
Appendix A. Supplementary Data
Supplementary data associated with this article, including experimental conditions and spectral
data can be found online at DOI:
17

References
References
[1]
Whitesell L, Lindquist SL. Nat. Rev. Cancer. 2005; 5: 761–772.
[10] Bieb MM, Buchner J. Cold Spring Harb Perspect Biol. 2019; pii: a034017.
[11] Kamal A, Thao L, Sensintaffar J, Zhang L, Boehm MF, Fritz LC, Burrows FJ. Nature
2003; 425: 407- 410.
[12] Trepel J, Mollapour M, Giaccone G, Neckers L. Nature Reviews Cancer 2010; 10: 537-
549.
[13] Neckers L, Workman P. Clinical Cancer Research 2012; 18: 64-76.
[14] Huang KH, Veal JM, Fadden RP, Rice JW, Eaves J, Strachan J-P, Barabasz AF, Foley
BE, Barta TE, Ma W, Silinski MA, Hu M, Partridge JM, Scott A, DuBois LG, Freed T, Steed
PM, Ommen AJ, Smith ED, Hughes PF, Woodward AR, Hanson GJ, McCall WS, Markworth
CJ, Hinkley L, Jenks M, Geng L, Lewis M, Otto J, Pronk B, Verleysen K, Hall SE. J. Med.
Chem. 2009; 52: 4288–4305.
[15] Waza M, Adachi H, Katsuno M, Minamiyama M, Sang C, Tanaka F, Inukai A, Doyu M,
Sobue G. Nature Medicine 2005; 11: 1088–1095.
[16] Stebbins CE, Russo AA, Schneider C, Rosen N, Hartl U, Pavletich NP. Cell 1997; 89:
239-250.
[17] Roe SM, Prodromou C, O'Brien R, Ladbury JE, Piper PW, Pearl LH. J. Med. Chem.
1999; 42: 260-266.
[18] Garcia-Carbonero R, Carnero A, Paz-Ares L. Lancet Oncol. 2013; 14: e358–e369.
[19] Sydor JR, Normant E, Pien CS, Porter JR, Ge J, Grenier L et al. Proc Natl Acad Sci USA.
2006; 103: 17408–17413.
[2]
Lianos GD, Alexiou GA, Mangano A, Mangano A, Rausei S, Boni L, Dionigi G, Roukos
DH. Cancer Lett. 2015; 360: 114–118.
[20] Jensen MR, Schoepfer J, Radimerski T, Massey A, Guy CT, Brueggen J, Quadt C,
Buckler A, Cozens R, Drysdale MJ, Garcia-Echeverria C, Chène P. Breast Cancer Res. 2008;
10: R33.
[21] Ying W, Du Z, Sun L, Foley KP, Proia DA, Blackman RK, Zhou D, Inoue T, Tatsuta N,
Sang J, Ye S, Acquaviva J, Ogawa LS, Wada Y, Barsoum J, Koya, K. Mol. Cancer Ther.2012;
11: 475-484.
18

[22] Woodhead AJ, Angove H, Carr MG, Chessari G, Congreve M, Coyle JE, Cosme J,
Graham B, Day PJ, Downham R, Fazal L, Feltell R, Figueroa E, Frederickson M, Lewis J,
McMenamin R, Murray CW, O’Brien MA, Parra L, Patel S, Phillips T, Rees DC, Rich S, Smith
D-M, Trewartha G, Vinkovic M, Williams B, Woolford AJ-A. J. Med. Chem.2010; 53: 5956-
5969.
[23] Lundgren K, Zhang H, Brekken J, Huser N, Powell RE, Timple N, Busch DJ, Neely L,
Sensintaffar JL, Yang Y-C, McKenzie A, Friedman J, Scannevin R, Kamal A, Hong K,
Kasibhatla SR, Boehm MF, Burrows FJ. Mol. Cancer Ther. 2009; 8: 921-929
[24] Caldas-Lopes E, Cerchietti L, Ahn JH, Clement CC, Robles AI, Rodina A, Moulick K,
Taldone T, Gozman A, Guo Y, Wu N, de Stanchina E, White J, Gross SS, Ma Y, Varticovski L,
Melnick A, Chiosis G. Proc. Natl. Acad. Sci. 2009; 106: 8368-8373
[25] Neckers L, Blagg B, Haystead T, Trepel JB, Whitesell L, Picard D. Cell Stress
Chaperones. 2018; 23: 467-482
[26] Jaeger AM, Whitesell L. Annu. Rev. Cancer Biol. 2019; 3: 275-297.
[27] Schopf FH, Biebl MM, Buchner J. Nat. Rev. Mol. Cell Biol. 2017; 18: 345–360.
[28] Johnson JL. Biochim. Biophys. Acta - Mol. Cell Res. 2012; 1823: 607–613.
[29] Song HY, Dunbar JD, Zhang YX, Guo D, Donner DB. J. Biol. Chem. 1995; 270: 3574-
3578.
[3]
Jego G, Hazoumé A, Seigneuric R, Garrido C. Cancer Lett. 2013; 332: 275–285.
[30] Zhu G, Lee AS. J. Cell. Physiol. 2014; 230: 1413-1420.
[31] Park HK, Jeong H, Ko E, Lee G, Lee JE, Lee SK, Lee AJ, Im JY, Hu S, Kim SH, Lee JH,
Lee C, Kang S, Kang BH. J. Med. Chem. 2017; 60: 7569–7578.
[32] Patel HJ, Patel PD, Ochiana SO, Yan P, Sun W, Patel MR, Shah SK, Tramentozzi E,
Brooks J, Bolaender A, Shrestha L, Stephani R, Finotti P, Leifer C, Li Z, Gewirth DT, Taldone
T, Chiosis G. J. Med. Chem. 2015; 58: 3922-3943.
[33] Patel PD, Yan P, Seidler PM, Patel HJ, Sun W, Yang C, Que NS, Taldone T, Finotti P,
Stephani RA, Gewirth DT, Chiosis G. Nat. Chem. Biol. 2013; 9: 677–684.
[34] Zhao R, Davey M, Hsu YC, Kaplanek P, Tong A, Parsons AB, Krogan N, Cagney G,
Mai D, Greenblatt J, Boone C, Emili A, Houry WA. Cell. 2005; 120: 715–727.
[35] Taldone T, Patel PD, Patel M, Patel HJ, Evans CE, Rodina A, Ochiana S, Shah SK,
Uddin M, Gewirth D, Chiosis G. J. Med. Chem. 2013; 56: 6803–6818.
[36] Jiang F, Guo AP, Xu JC, You QD, Xu XL. J. Med. Chem. 2018; 61: 9513-9533.
19

[37] Ernst JT, Neubert T, Liu M, Sperry S, Zuccola H, Turnbull A, Fleck B, Kargo W, Woody
L, Chiang P, Tran D, Chen W, Snyder P, Alcacio T, Nezami A, Reynolds J, Alvi K, Goulet L,
Stamos D. J. Med. Chem. 2014; 57: 3382−3400.
[38] Harris SF, Shiau AK, Agard DA. Structure. 2004; 12: 1087–1097.
[39] Schulze A, Beliu G, Helmerich DA, Schubert J, Pearl LH, Prodromou C, Neuweiler H.
Nat. Chem. Biol. 2016; 12: 628–635.
[4]
Wegele H, Müller L, Buchner J. Rev. Physiol. Biochem. Pharmacol. 2004; 1–44.
[40] Taddei M, Ferrini S, Giannotti L, Corsi M, Manetti F, Giannini G, Vesci L, Milazzo FM,
Alloatti D, Guglielmi MB, Castorina M, Cervoni ML, Barbarino M, Foderà R, Carollo V, Pisano
C, Armaroli S, Cabri W. J. Med. Chem. 2014; 376: 2258-2274.
[41] Kolb HC, Finn MG, Sharpless KB. Angew. Chem. Int. Ed. 2001; 40: 2004–2021.
[42] Proia DA, Bates RC. Cancer Research 2014; 74: 1294-1300.
[43] Khandelwal A, Kent CN, Balch M, Peng S, Mishra SJ, Deng J, Day VW, Liu W,
Subramanian C, Cohen M, Holzbeierlein JM, Matts R, Blagg BSJ. Nature Communications
2018; 9: 425.
[44] Ernst JT, Liu M, Zuccola H, Neubert T, Beaumont K, Turnbull A, Kallel, A. Vought B,
Stamos D. Bioorg. Med. Chem. Lett Bioorg. Med. Chem. Lett. 2—14: 24; 204–208.
[5]
[6]
[7]
Chatterjee S, Burns TF. Int. J. Mol. Sci. 2017; 18: 1978.
Garcia-Carbonero R, Carnero A, Paz-Ares L. Lancet Oncol. 2013; 14: e358–e369
Joly AL, Wettstein G, Mignot G, Ghiringhelli F, Garrido C. J. Innate Immun. 2010; 2:
238–247.
[8]
[9]
Hassan S, Biswas MHU, Zhang C, Du C, Balaji KC. Oncogene. 2009; 28: 4386–4396.
Azad AA, Zoubeidi A, Gleave ME, Chi KN, Nat. Rev. Urol. 2015; 12: 26–36.
20

21

22