Design, synthesis and biological evaluation of triaryl compounds as novel 20S proteasome inhibitors

Article abstract of DOI:10.1016/j.bmcl.2020.127508

Thirty novel triaryl compounds were designed and synthesized based on the known proteasome inhibitor PI-1840. Most of them showed significant inhibition against the β5c subunit of human 20S proteasome, and five of them exhibited IC₅₀ values at the sub-micromolar level, which were comparable to or even more potent than PI-1840. The most active two (1c and 1d) showed IC₅₀ values of 0.12 and 0.18 μM against the β5c subunit, respectively, while they displayed no obvious inhibition against the β2c, β1c and β5i subunits. Molecular docking provided informative clues for the subunit selectivity. The potent and subunit selective proteasome inhibitors identified herein represent new chemical templates for further molecular optimization.

Full text of DOI:10.1016/j.bmcl.2020.127508

Bioorganic&MedicinalChemistryLetters30(2020)127508
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis and biological evaluation of triaryl compounds as novel
20S proteasome inhibitors
⁎
Yajun Yanga, Ke Wanga, Bo Wua, Ying Yanga, Fangfang Laib, Xiaoguang Chenb, Zhiyan Xiaoa,
a Beijing Key Laboratory of Active Substance Discovery and Druggability Evaluation, Institute of Material Medica, Chinese Academy of Medical Sciences and Peking Union
Medical College, Beijing 100050, China
b The State Key Laboratory of Bioactive Substance and Function of Natural Medicines, Institute of Material Medica, Chinese Academy of Medical Sciences and Peking Union
Medical College, Beijing 100050, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Thirty novel triaryl compounds were designed and synthesized based on the known proteasome inhibitor PI-
1840. Most of them showed significant inhibition against the β5c subunit of human 20S proteasome, and five of
them exhibited IC50 values at the sub-micromolar level, which were comparable to or even more potent than PI-
1840. The most active two (1c and 1d) showed IC50 values of 0.12 and 0.18 μM against the β5c subunit,
respectively, while they displayed no obvious inhibition against the β2c, β1c and β5i subunits. Molecular
docking provided informative clues for the subunit selectivity. The potent and subunit selective proteasome
inhibitors identified herein represent new chemical templates for further molecular optimization.
Non-covalent
Proteasome inhibitors
Triaryl compounds
In 2004, the Nobel Prize in Chemistry was awarded to Aaron
Ciechanover, Avram Hershko, and Irwin Rose for their discovery of
ubiquitin mediated protein degradation by the ubiquitin–proteasome
system (UPS).1 26S proteasome is the main proteolytic component of
the UPS, a proteolytic system precisely controlling the cellular protein
homeostasis.2–4 26S proteasome is a highly conserved protease com-
plex, which is consisted of the 19S regulatory particles and the 20S
proteasome catalytic particle.5–8 Three different catalytic subunits,
namely β5, β2 and β1, are presented in 20S proteasome and are re-
sponsible for chymotrypsin-like (CT-L), trypsin-like (T-L) and caspase-
like (C-L) proteolytic activities, respectively.9,10 The N-terminal Thr1 of
the three catalytic subunits serves as the nucleophile to start the clea-
vage of corresponding peptide bonds.11 It was reported that as com-
pared to the β1 and β2 subunits, the β5 subunit plays a more important
role in the degradation of tumor suppressor proteins, and a much more
significant degradation was observed by inhibiting the β5 subunit.10
Therefore, considerable research interests have been devoted to the
development of proteasome inhibitors with CT-L inhibitory activity as
cancer therapy.12
covalent proteasome inhibitors are expected to alleviate the afore-
mentioned drawbacks related to covalent proteasome inhibitors, and
thereby have drawn more and more research interests recently.17,18 In
particular, non-covalent proteasome inhibitors with non-peptide ske-
letons are assumed to possess improved pharmacokinetic and safety
profiles as compared to those peptidic covalent inhibitors currently
available in clinical uses.19,20
PI-1833 is a non-covalent proteasome inhibitor initially discovered
via high-throughput screening.21 The oxadiazole-isopropylamide scaf-
fold presented in PI-1833 was new for proteasome inhibitors, and fur-
ther structural optimization of PI-1833 led to the discovery of PI-1840
(Fig. 1), a potent non-covalent CT-L selective inhibitor with in vivo
antitumor activity against solid tumors.21,22 Inspired by the novel
scaffold and promising pharmacological profile of PI-1840, we at-
tempted to explore the structure–activity relationship of this compound
class as non-covalent 20S proteasome inhibitors. As illustrated in Fig. 1,
the structure of PI-1840 is characterized by a unique oxadiazole-iso-
propylamide scaffold with three aryl moieties presenting in molecular
areas A, B and C. To identify more potent and subunit selective pro-
teasome inhibitors, two series of compounds were designed accord-
ingly. In series 1 (1a-1l), the oxadiazole-isopropylamide core was
maintained, while the aryl moieties in areas A and C were varied.
Virtually, chemical manipulations in area A include substitution of the
n-butyl on the phenyl and incorporation of benzoheterocyclic rings
(Ar1) to replace the phenoxy fragment. While in area C, the pyridinyl
Three 20S proteasome inhibitors, Bortezomib (Velcade), carfilzomib
(Kyprolis) and Ixazomib (Ninlaro), have been approved for the treat-
ment of relapsed or refractory multiple myeloma (MM) in 2003, 2012
and 2015, respectively.13 However, as covalent proteasome inhibitors,
they act by forming covalent bonds with Thr1 of the catalytic subunits,
which leads to poor specificity and thus severe side effects.14–16 Non-
⁎ Corresponding author.
E-mail address: xiaoz@imm.ac.cn (Z. Xiao).
https://doi.org/10.1016/j.bmcl.2020.127508
Received 25 April 2020; Received in revised form 14 August 2020; Accepted 17 August 2020
Availableonline24August2020
0960-894X/©2020PublishedbyElsevierLtd.

Y. Yang, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127508
Fig. 1. Molecular design based on PI-1840.
Scheme 1. Synthetic route for 1a-1l.
Reagents and conditions: (a) Na2CO3, hy-
droxylamine hydrochloride, H2O, 70 °C,
14 h, 88%-98.8%; (b) chloroacetyl chloride,
acetone, rt, 30 min, 77%-95%; (c) toluene,
reflux, 2 h, 83%-98%; (d) isopropylamine,
K2CO3, CH3CN, reflux, 30 min, 88%-91%; (e)
1) Ar1(CH2)nCOOH, (COCl)2, toluene, reflux,
3 h; 2) Et3N, THF, rt, 15 min, 70%-91%.
was changed to other aromatic rings (Ar3). The structural variations in
areas A and C were designed to extend the SAR presented previously, 21
and moieties with different bulky hindrance and number of heteroa-
toms were investigated. Besides areas A and C, the oxadiazole ring in
area B was also altered to phenyl or pyridinyl (Ar2) in series 2 (2a-2r) to
further expand the chemical diversity of this compound class.
10 were further reacted with acyl chlorides to give target compounds
2a-2r. The structures of compounds 1a-1l and 2a-2r are listed in
Table 1. Similar to PI-1840 derivatives,21 atropisomers of compounds
1a-1l and 2a-2r were also observed in 1H NMR spectroscopy, which
was attributed to the hindered rotation of the amide bond. Compound
1c was taken as an example and subjected to variable-temperature 1
H
Two different synthetic routes were adopted to obtain the target
compounds. As shown in Scheme 1, compounds 1a-1l were synthesized
from commercially available nitriles 3.21 Briefly, the hydroxyamidines
4 were prepared by reacting the nitriles 3 with hydroxylamine hydro-
chloride, and were subsequently treated with chloroacetyl chloride to
yield compounds 5. Cyclization of compounds 5 in reflux toluene af-
forded oxadiazoles 6. The latter reacted with isopropylamine in reflux
acetonitrile to give compounds 7, which were further acylated to pro-
vide target compounds 1a-1l. Alternatively, compounds 2a-2r were
obtained by following the synthetic route shown in Scheme 2. Sub-
stituted23 benzyl chlorides 8 were reacted with isopropylamine in re-
flux acetonitrile to provide compounds 9, which were subsequently
reacted with aromatic boric acids to yield compounds 10. Compounds
NMR. The signals for the minor component in the spectrum measured at
25 °C almost disappeared in that measured at 80 °C, which confirmed
the existence of atropisomers.
All the synthesized compounds were evaluated for their inhibitory
activities against the CT-L activity of human 20S proteasome and PI-
1840 was used as a reference compound. As shown in Table 1, most of
the compounds showed apparent inhibition against the CT-L activity of
human 20S proteasome, and thirteen of them exhibited IC50 values at
micromolar or sub-micromolar level. The five most active compounds
(1b, 1c, 1d, 1i and 1j) were comparable to or even more potent than
the template PI-1840.
The oxadiazole-isopropylamide core and the phenyl-containing
moiety in area A were initially maintained, and only the terminal
Scheme 2. Synthetic route for compounds 2a-2r. Reagents and conditions: (a) isopropylamine, K2CO3, CH3CN, reflux, 30 min; (b) Ar3B(OH)2, Pd(PPh3)4, Na2CO3 aq,
toluene, ethanol, reflux under nitrogen protection, 85%-93% for a and b; (c) 1) Ar1(CH2)nCOOH, (COCl)2, toluene, reflux, 3 h; 2) Et3N, THF, rt, 15 min, 64%-97%.
2

Y. Yang, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127508
Table 1
Inhibition against the CT-L activity of human 20S proteasome by compounds 1a-1 l and 2a-2r.
Compd
Ar1
n
1
Ar2
Ar3
IC50(μM)1
PI-18402
0.63
0.15
1a
1b
1
1
4.31
0.48
0.05
0.88
1c
1d
1e
1
1
1
0.12
0.18
1.61
0.04
0.05
0.16
1f
1
0
0
0
0
0
14.7
NA3
NA
0.45
1g
1h
1i
0.61
0.39
NA
0.09
0.09
1j
1k
1l
0
1
NA
O
O
N
N
N
2a
1.08
0.05
O
N
(continued on next page)
3

Y. Yang, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127508
Table 1 (continued)
Compd
Ar1
n
1
Ar2
Ar3
IC50(μM)1
2b
2.27
10.7
9.45
1.14
0.06
O
O
O
N
N
2c
2d
2e
1
1
1
0.62
0.52
0.37
O
S
O
O
N
N
2f
1
1
33.5
1.37
0.68
0.50
S
S
2g
O
O
O
2h
2i
1
1
NA
NA
O
O
N
N
N
2j
1
0
NA
O
N
N
N
2k
20.2
0.05
0.45
O
2l
0
0
0
0
0
NA
O
O
O
N
N
N
2m
2n
2o
2p
47.3
NA
O
S
61.6
3.36
0.73
0.17
O
N
O
O
S
2q
2r
0
0
NA
NA
N
N
N
N
N
1
2
The IC50 values are calculated from two independent measures.
The IC50 value for PI-1840 was 27 nM in Ref. 21. However, purified 20S rabbit proteasome was used for the measurement therein. Similar discrepancy was
previously observed by us.24
3
NA (not active) represents a percentage inhibition less than 30% under the concentration of 10 μM.
4

Y. Yang, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127508
Table 2
When the oxadiazole ring in PI-1840 was replaced by meta- (2a) or
para-phenyl (2e), only a slight decrease in the inhibitory activity was
observed. Further substitution of the 3-pyridinyl at Ar3 with other
aromatic rings, the inhibitory activity diminished in various extent (2b-
2d and 2f). Compounds with meta- or para-phenyl introduced at Ar2
(2a-2h and 2k-2n) were generally less active than their analogs with an
oxadiazole ring at this area. The substitution of the oxadiazole ring in
PI-1840 with a pyridinyl (2i) was evidently unfavored, and the adverse
effect of a pyridinyl at Ar2 was further supported by the activity data of
compounds 2j, 2q and 2r. Interestingly, although 2e and 2g were
comparably active, the presence of a five-membered thienyl at Ar3
obviously made the compound more sensitive to the substitution of n-
butyl with phenoxyl and results in significant loss of activity (2f vs 2h),
which is consistent with data obtained for series 1.
Inhibitory activities of selected compounds against CT-L, T-L and C-L activities
of human 20S proteasome.
Compd
IC50 (μM)1
CT-L
T-L
C-L
1c
0.12
0.18
1.08
0.63
0.04
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
1d
0.05
0.05
0.15
2a
PI-1840
1
The IC50 values are calculated from two independent measures.
pyridinyl was varied (compounds 1a-1d). It looks like that the Ar3
moiety in area C is tolerable to structural variations, while heteroaro-
matic groups seem to be favored (1a vs 1b, 1c and 1d), especially those
with a hetero-atom vicinal to the oxadiazole ring (PI-1840 vs 1b, 1c and
1d). The n-butyl on the phenyl in area A of PI-1840 was then replaced
with phenoxyl group (1e), which caused a slight decrease in the in-
hibition against the CT-L activity. However, the effects of such a sub-
stitution on inhibitory activities appeared to vary with different Ar3
moieties. When Ar3 changed to furyl (1c), substitution of n-butyl with
phenoxyl (1f) resulted in a significant decline in the inhibitory activity.
When benzoheterocyclic rings were incorporated in area A (1 g-1l), the
inhibitory activities were generally weaker. Again, the discrepancy in
activity loss for compounds with different Ar3 was observed. The re-
placement of area A in PT-1840 with 2-benzofuryl only led to marginal
decrease in the inhibitory activity when five-membered furyl or thienyl
presented at Ar3 (1c, 1d vs 1i, 1j), while for compounds with naphthyl
or 2-quinolyl at Ar3, a dramatic drop in activity was detected (1a, 1b vs
1k, 1l). The different extent in activity loss for compounds with dif-
ferent Ar3 implicates the involvement of alternative binding poses,
which might be sensitive to the Ar3 substitution.
Compounds (1c, 1d and 2a) with the highest potency against CT-L
activity in each series were further evaluated for their effects on T-L and
C-L activities of human 20S proteasome. PI-1840 was again taken as a
reference compound. As illustrated in Table 2, all the four compounds
showed no significant inhibition against the T-L and CT-L activities.
The constitutive proteasome (cCP) and the immunoproteasome
(iCP) are two different types of eukaryotic proteasomes. The former is
expressed in all eukaryotic cells, whereas the latter is mainly expressed
in cells of hematopoietic origin. PI-1840 was selective for β5c over
β5i,22 and consistently, compounds 1c and 1d showed no significant
inhibition against the β5i subunit (with percentage inhibition of 7.3%
and 25.3% at 10 μM, respectively).
Molecular docking of compound lc was performed to explore the
molecular basis for the selective inhibition against different 20S pro-
teasome subunits. As shown in Fig. 2, compound lc forms hydrogen
bonds with residues Thr1 and Thr21 in the active site of the β5c sub-
unit. Additional hydrophobic interactions with Ala20, Met45-Ala49,
Gly129 and Tyr169 of the β5c subunit were also observed (Fig. 2A). For
Fig. 2. The interaction modes of compound lc with (A) β5c, (B) β2c, (C) β1c and (D) β5i subunits as proposed by molecular docking. For cCP and iCP, the structures
of 5LF3 and 6E5B were applied respectively.
5

Y. Yang, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127508
Table 3
inconsistency between activities at molecular and cellular levels sug-
gests poor cellular permeability or alternative mechanisms are in-
volved.
In vitro antitumor activity of selected compounds against MDA-MB-468.
Compd
IC50(μM)1
Compd
IC50(μM)1
1b
35.0
38.8
27.0
6.8
0.06
1j
> 40
Declaration of Competing Interest
1c
0.67
2.95
2a
2b
2 g
2p
19.3
30.2
> 40
31.9
2.16
1d
0.21
0.63
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
1e
1.22
1i
> 40
7.6
PI-1840
0.85
1
The IC50 values are calculated from two independent measures.
Acknowledgements
β2c and β1c subunits, only one hydrogen bond was detected between
compound 1c and Thr1, while interactions with the residues Asp53 (for
β2c) or Arg45 (for β1c), which were believed to be critical for effective
inhibition,25 were not observed (Fig. 2B and 2C). To fit the active site
of the β5i subunit, compound lc adopted a pose different from that in
the active site of β5c subunit. Although hydrogen bonds are monitored
between compound lc and residues Gly47 and Gln53, the crucial hy-
drogen bonds with Thr1 and Ser21 were absent (Fig. 2D). Molecular
docking revealed the interaction modes of compound 1c with different
subunits, and partially explained the β5c selectivity of compound lc.
However, the structural features of this compound class responsible for
the subunit selectivity is still ambiguous.
This research was funded by National Natural Science Foundation of
China (21672263) and CAMS Innovation Fund for Medical Sciences
(CIFMS, 2016-I2M-3-009).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.127508.
References
1. Rentsch A, Landsberg D, Brodmann T, et al. Angew. Chem. Int. Ed.
2013;52:5450–5488.
PI-1840 was previously reported to inhibit the proliferation of
human breast cancer cell line MDA-MB-468.21 To further compare their
biological activities of the target compounds and PI-1840, ten com-
pounds were tested in parallel to PI-1840 against MDA-MB-468 cells
using the SRB assay. As shown in Table 3, seven compounds showed
apparent inhibition against the cancer cells, and compound 1e ex-
hibited comparable activity to PI-1840.22
2. Finley D. Annu. Rev. Biochem. 2009;78:477–513.
3. Murata S, Yashiroda H, Tanaka K. Nat. Rev. Mol. Cell Biol. 2009;10:104–115.
4. Gallastegui N, Groll M. Trends Biochem. Sci. 2010;35:634–642.
5. Santoro AM, Cunsolo A, D'Urso A, et al. Chem. Sci. 2016;7:1286–1297.
6. Gräwert MA, Groll M. Chem. Commun. 2012;48:1364–1378.
7. Nandi D, Tahiliani P, Kumar A, et al. J. Biosci. 2006;31:137–155.
8. Thibaudeau TA, Smith DM. Pharmacol. Rev. 2019;71:170–197.
9. Borissenko L, Groll M. Chem. Rev. 2007;107:687–717.
It was observed from Tables 1 and 3 that the inhibition against CT-L
activity is inconsistent with the cytotoxicity of the selected compounds
and activities at the molecular level were generally more potent than
those at the cellular level. Such observations imply the involvement of
poor cellular permeability and/or the presence of alternative mechan-
isms.
10. Philipp MC, Craig MC. ACS Cent. Sci. 2017;3:830–838.
11. Groll M, Biochim HR. Biophys. Acta, MolCell Res. 2004;1695:33–44.
12. Genin E, Reboud-Ravaux M, Vidal J. Curr. Top Med. Chem. 2010;10:232–256.
13. Sherman DJ, Li J. Molecules. 2020;25:671.
14. Perfetti V, Palladini G, Brunetti L, et al. Eur. J. Gastroenterol. Hepatol.
2007;19:599–601.
15. Enrico O, Gabriele B, Nadia C, et al. Bri. J. Daematol. 2007;138:396–397.
16. Villarrubia B, Betlloch I, Mataix J, et al. Bri. J. Dermatol. 2007;156:784–785.
17. Marechal X, Pujol A, Richy N, et al. Eur. J. Med. Chem. 2012;52:322–327.
18. Huber EM, Groll M. Angew. Chem. Int. Ed. 2012;51:8708–8720.
19. Guedes RA, Aniceto N, Andrade MAP, et al. Int. J. Mol. Sci. 2019;20:5326.
20. Gallastegui N, Beck P, Arciniega M, et al. Angew Chem. Int. Ed. 2012;51:247–249.
21. Ozcan S, Kazi A, Marsilio F, et al. J. Med. Chem. 2013;56:3783–3805.
22. Kazi A, Ozcan S, Tecleab A, et al. J. Bio. Chem. 2014;289:11906–11915.
23. Li Z, Gelbaum C, Fisk JS, et al. J. Org. Chem. 2016;81:8520–8529.
24. Xu K, Xiao Z, Tang YB, et al. Bioorg. Med. Chem. Lett. 2012;22:2772–2774.
25. Schrader J, Henneberg F, Mata RA, et al. Science. 2016;353:594–598.
In summary, thirty compounds were designed based on the known
non-covalent 20S proteasome inhibitor PI-1840. With structural varia-
tions on different molecular areas of PI-1840, preliminary structure–-
activity relationships of this compound class as proteasome inhibitors
were drawn and novel structures with IC50 values at the sub-micro-
molar level against the CT-L activity of human 20S proteasome were
obtained. Compounds with β5c selectivity and cellular activity were
also identified, which are worth of further optimization. However, the
6