Nonpeptidic Selective Inhibitors of the Chymotrypsin-Like (β5 i) Subunit of the Immunoproteasome

Article abstract of DOI:10.1002/anie.201600190

Elevated expression of the immunoproteasome has been associated with autoimmune diseases, inflammatory diseases, and various types of cancer. Selective inhibitors of the immunoproteasome are not only scarce, but also almost entirely restricted to peptide-based compounds. Herein, we describe nonpeptidic reversible inhibitors that selectively block the chymotrypsin-like (β5i) subunit of the human immunoproteasome in the low micromolar range. The most potent of the reversibly acting compounds were then converted into covalent, irreversible, nonpeptidic inhibitors that retained selectivity for the β5i subunit. In addition, these inhibitors discriminate between the immunoproteasome and the constitutive proteasome in cell-based assays. Along with their lack of cytotoxicity, these data point to these nonpeptidic compounds being suitable for further investigation as β5i-selective probes for possible application in noncancer diseases related to the immunoproteasome.

Full text of DOI:10.1002/anie.201600190

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201600190
German Edition: DOI: 10.1002/ange.201600190
Inhibitors
Nonpeptidic Selective Inhibitors of the Chymotrypsin-Like (b5i)
Subunit of the Immunoproteasome
Izidor Sosic , Martina Gobec , Boris Brus, Damijan Knez, Matej Zivec, Janez Konc,
+
ˇ+
ˇ
ˇ
ˇ
ˇ
ˇ
ˇ
ˇ ˇ
ˇ
Samo Lesnik, Mitja Ogrizek, Ales Obreza, Dusan Zigon, Dusanka Janezic, Irena Mlinaric-
ˇˇ
Rascan, and Stanislav Gobec*
Abstract: Elevated expression of the immunoproteasome has
been associated with autoimmune diseases, inflammatory
diseases, and various types of cancer. Selective inhibitors of
the immunoproteasome are not only scarce, but also almost
entirely restricted to peptide-based compounds. Herein, we
describe nonpeptidic reversible inhibitors that selectively block
the chymotrypsin-like (b5i) subunit of the human immuno-
proteasome in the low micromolar range. The most potent of
the reversibly acting compounds were then converted into
covalent, irreversible, nonpeptidic inhibitors that retained
selectivity for the b5i subunit. In addition, these inhibitors
discriminate between the immunoproteasome and the constit-
utive proteasome in cell-based assays. Along with their lack of
cytotoxicity, these data point to these nonpeptidic compounds
being suitable for further investigation as b5i-selective probes
for possible application in noncancer diseases related to the
immunoproteasome.
expression of the iCP-active b subunits (designated b1i, b2i,
b5i), which replace their constitutive counterparts, can be
induced under the influence of tumor necrosis factor a and
interferon-g released by T cells during acute immune and
stress responses.[6,7] Extensive research in the past decades has
resulted in the successful development of bortezomib and
carfilzomib as therapeutics for multiple myeloma and mantle-
cell lymphoma that act as b5 and b5i inhibitors of cCP and
iCP, respectively.[8,9] However, selective inhibition of the iCP
is expected to cause fewer adverse effects, since its expression
is induced only during the course of disease processes.[10,11]
Current studies also indicate that b5i-selective compounds
such as [PR-957 (ONX-0914)] provide more therapeutic
potential than inhibition of the other two active sites of the
iCP, especially for autoimmune and inflammatory dis-
eases.[12,13] From a structural perspective, the majority of
currently available inhibitors of the iCP have a peptidic
backbone and are thus prone to poor metabolic stability and
low bioavailability. Aside from improvements to the current
array of peptide-like inhibitors of the iCP, it is imperative to
develop inhibitors with different structural characteristics and
improved physicochemical properties, and consequently
improved pharmacokinetic and pharmacodynamic pro-
files.[14,15] To achieve this, options outside the classic “pro-
tease-inhibitor box” have to be considered. Various drug-
discovery approaches[16] should enable the identification of
compounds that are devoid of the peptidic backbone and can
thus avoid the current limitations of peptide-based com-
pounds. Currently, nonpeptidic inhibitors of the iCP are
scarce, with only two recent examples known.[17,18] The crystal
structure of the murine iCP (PDB ID: 3UNF) opens diverse
possibilities for advanced computer-based discovery and
design, as well as for rational optimization of selective iCP
inhibitors.[19] In the present study, we describe reversible and
irreversible nonpeptidic inhibitors of the human b5i subunit
that were discovered through a structure-based virtual-
screening approach (see the Supporting Information) using
the ZINC drug-like subset of compounds. Irreversibly acting
compounds show improved inhibition and better selectivity
properties than PR-957, the established inhibitor of b5i, along
with very low cytotoxicity.
T
he eukaryotic 20S proteasome (the core particle, CP)
represents the heart of the ubiquitin-proteasome system that
is responsible for maintaining protein homeostasis and
regulation for a variety of cellular processes, ranging from
antigen processing and signal transduction, to cell differ-
entiation and apoptosis.[1,2] The 20S proteasome has three
enzymatically active subunits, namely the b1 (caspase-like),
b2 (trypsin-like), and b5 (chymotrypsin-like) subunits, which
all have distinct substrate specificities.[3] The constitutive
proteasome (cCP) is expressed in all eukaryotic cells, whereas
the immunoproteasome (iCP) is predominantly found in cells
of hematopoietic origin.[4,5] In non-hematopoietic cells,
[+]
[*] Dr. I. Sosic, Dr. M. Gobec,[+] Dr. B. Brus, D. Knez, Dr. M. Zivec,
ˇ
ˇ
ˇ
ˇˇ
Prof. Dr. A. Obreza, Prof. Dr. I. Mlinaric-Rascan, Prof. Dr. S. Gobec
Faculty of Pharmacy, University of Ljubljana
ˇ
ˇ
Askerceva 7, 1000 Ljubljana (Slovenia)
E-mail: stanislav.gobec@ffa.uni-lj.si
ˇ
Dr. J. Konc, S. Lesnik, M. Ogrizek
National Institute of Chemistry
Hajdrihova 19, 1000 Ljubljana (Slovenia)
ˇ
Dr. D. Zigon
ˇ
Department of Environmental Sciences, Jozef Stefan Institute
Jamova 39, 1000 Ljubljana (Slovenia)
In vitro evaluation of the best virtual screening hits
(Table S1 in the Supporting Information) revealed a substi-
tuted psoralene derivative (Figure 1a, virtual screening hit) as
the most promising compound, since it showed good inhib-
ition of the b5i subunit at 50 mm (residual activity = 0.5%).
This molecule was resynthesized by using a very efficient and
straightforward synthesis (Scheme S1 in the Supporting
ˇ ˇ
Prof. Dr. D. Janezic
Faculty of Mathematics
Natural Sciences and Information Technologies
University of Primorska, Glagoljaska 8, 6000 Koper (Slovenia)
ˇ
[+] These authors contributed equally to this work.
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/anie.201600190.
Angew. Chem. Int. Ed. 2016, 55, 5745 –5748
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5745

Angewandte
Communications
Chemie
To assess the selectivity of these psoralene-based inhib-
itors of b5i, compounds 1–5 (Figure 1a), which demonstrated
reasonable potency against the b5i subunit, were selected.
These compounds showed preferential inhibition of b5i over
the b2i and b1i subunits of human iCP, as well as over all three
subunits of human cCP (Table S3). Compound 3 was the most
selective; it showed more than 100-fold greater potency for
the inhibition of b5i (Ki = 1.6 mm) than b5c (Ki = 172.2 mm)
and negligible to modest activity against the b1 and b2
subunits of each of these CPs (Table S3).
To refine the potency of reversibly acting compounds,
subsequent medicinal chemistry efforts were based on the
design of covalent inhibitors. Sustained inhibition by covalent
inhibitors can only be achieved when a compound first binds
noncovalently to the target, followed by positioning of the
electrophilic “warhead” near to the desired nucleophilic
amino acid residue of the protein.[20,21] Since compound 3
already showed inhibitory activity through noncovalent
interactions with the active site of b5i, incorporation of the
correct “warhead” should improve the inhibitory potency.
Docking studies revealed that the carboxylic group of the
most potent inhibitor (3) from the noncovalent series is
positioned in proximity to the catalytic Thr1 (Figure S1).
Therefore, various electrophilic “warheads” were introduced
on that portion of the molecule (Figure 2). The selection of
the threonine-selective reactive groups used to date was
summarized recently.[15] In this study, an alternate series of
electrophilic analogues with the potential to react with
catalytic Thr1 was prepared (Scheme S4) and evaluated
biochemically (Table 1).
Interestingly, the nitrile-based (31–36) and acrylamide-
based (34–40) compounds showed no improvement in b5i
inhibition. However, our structure-based approach did yield
compounds with significantly improved and irreversible
inhibition of b5i. The succinimidyl esters 29 and 30 inhibited
b5i activity with half-maximal inhibitory concentration (IC50)
values of 6.36 mm and 0.009 mm, respectively (Table 1, Fig-
ure S2). Very potent irreversible inhibition was also achieved
with the oxathiazolone-bearing compounds 42–44, with IC50
values ranging from 0.013 mm (42) to 0.20 mm (44; Table 1,
Figure 1. a) Structures of the virtual screening hit and the reversible
inhibitors of b5i. b) SAR for the psoralene-based inhibitors.
Information) and its Ki was determined to be 15 mm. Its
precursor in this synthesis (compound 3) was also assayed,
which attracted our attention because of its predicted similar
interaction profile with the active-site residues of b5i.
Compound 3 lacked the phenylglycine part of the virtual
screening hit; however, it proved to be a 10-fold more potent
inhibitor of the b5i subunit (Ki = 1.6 mm, Figure 1a). Further
structure–activity relationship (SAR) studies were enabled by
the construction of a focused library of diversely substituted
psoralenes (Table S2, Schemes S1–S3). In order not to over-
interpret the SAR data, only the most important findings are
shown in Figure 1b.
Figure 2. Structures of the designed electrophilic moieties coupled to the noncovalently acting scaffold of 3 (red).
5746
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5745 –5748

Angewandte
Communications
Chemie
Table 1: In vitro inhibitory potencies of covalently acting compounds derived from compound 3 against
all subunits of the human cCP and iCP. Final concentrations of the cCP and iCP in the assays were
0.8 nm and 0.2 nm, respectively. The IC50 values were determined only for compounds that showed
notable inhibition at 10 mm (RA <60%). For comparison, the IC50 values for inhibition of the b5i subunit
of commonly used peptidic inhibitors are presented (from Ref. [15]): bortezomib (0.004 mm),
carfilzomib (0.033 mm), PR-924 (0.022 mm).
monocytic leukemia and express
high levels of iCP, whereas HeLa
cells predominantly contain cCP.[24]
These compounds showed better
inhibition of b5 activity in
THP-1 cell lysates than in HeLa
cell lysates (Table 2, Figure S4).
Although approximately 10-fold
reductions in potency were ob-
served for compounds 30 and 42
when compared with in vitro assays
with purified human iCP (Table 1),
30 and 42 showed about 10-fold and
40-fold higher inhibition, respec-
tively, in the THP-1 cell lysates
than in the HeLa cell lysates
Compound IC50 b5i [mm] IC50 b2i
[mm]
IC50 b1i
[mm]
IC50 b5c
[mm]
IC50 b2c
[mm]
IC50 b1c
[mm]
b5c/
b5i[b]
PR-957[a]
0.015Æ0.002 0.90Æ0.08 0.27Æ0.04 0.16Æ0.03 3.36Æ0.26 6.59Æ0.86 10.6
6.36Æ0.61 21.6Æ1.4 15.2Æ1.4 6.39Æ1.90 19.3Æ5.5 10.5Æ2.5 1.0
0.009Æ0.002 7.58Æ0.12 5.69Æ0.24 0.22Æ0.02 8.96Æ0.35 6.13Æ0.27 23.9
29
30
42
43
44
0.013Æ0.004 NA
0.12Æ0.05 NA
0.20Æ0.04 NA
NA
NA
NA
1.07Æ0.26 NA
NA
NA
NA
82.6
>80
>50
NA
NA
NA
NA
[a] The IC50 values correspond well to published data.[15] [b] A high IC50 b5c/b5i ratio indicates selectivity
for b5i. NA: inactive at 10 mm.
Figure S2). Remarkably, both 30 and 42 turned out to be two-
Table 2: Cell-based IC50 values against b5 activity in THP-1 and HeLa
fold to eight-fold more selective for b5i than PR-957 is
(Table 1). Oxathiazolones were recently described as a new
threonine-selective “warhead”.[22] Therefore, the results
obtained for compounds 42–44 represent a clear proof of
concept for this approach. Succinimidyl esters as “warheads”,
on the other hand, need to be considered with caution, since
these are predominantly used as protein labeling or cross-
linking reagents. Nonetheless, we showed that compound 30 is
not promiscuous, since it showed a reasonable preference for
b5i inhibition over b5c (23.9-fold), as well as over other iCP
and cCP subunits (Table 1).
cells and lysates.
Compound IC50
THP-1 cells
[mm]
IC50
HeLa cells
[mm]
IC50
IC50
HeLa lysates
[mm]
THP-1
lysates
[mm]
PR-957[a]
3
30
42
0.011Æ0.003 0.098Æ0.021 0.007Æ0.001 0.025Æ0.004
11.6Æ1.8
27.8Æ7.3
19.3Æ3.1
125Æ16
>100
>100
1.31Æ0.34
0.13Æ0.01
0.22Æ0.03
43.8Æ4.1
1.45Æ0.25
7.20Æ0.83
[a] IC50 values determined are in good agreement with published data.[23]
Although IC50 calculations are not the best method to
accurately characterize irreversible and/or mechanism-based
inhibitors,[21] these were used for comparisons with the
established b5i inhibitor PR-957 (Table 1). We also charac-
terized PR-957 fully by following the time course of the
hydrolysis of Suc-LLVY-AMC by iCP and cCP in the
presence of various concentrations of PR-957; the same was
done for the two most potent irreversibly acting compounds,
30 and 42 (Table S4). PR-957 (kinact/Ki = 14324mÀ1 sÀ1) and 42
(kinact/Ki = 15448mÀ1 sÀ1) showed very similar kinetic param-
eters, but the b5i selectivity was significantly better for
compound 42, with a kinact/Ki b5c/b5i ratio of 112, compared to
29 for PR-957 (Table S4). Additionally, the time-dependence
and irreversibility of b5i inhibition by PR-957, 30, and 42 were
determined by IC50 shift assay (Table S5, Figure S3) and
rapid-dilution assay (Table S6). To propose a mechanism for
the covalent interaction of succinimidyl ester 30 with the
active-site Thr1 of the b5i subunit, computational quantum
mechanics/ molecular mechanics analysis was performed (see
the Supporting Information). The simulation showed a very
similar reaction mechanism to that proposed for sulfonyl
fluorides.[23] However, although irreversible inhibition was
clearly demonstrated by enzyme kinetics experiments, we
could not substantiate this mechanism by time-resolved
intact-protein mass spectrometry.
(Table 2). On the other hand, despite being more active,
PR-957 showed only around 3-fold difference in inhibition
between these cell lysates (Table 2, Figure S4). Interestingly,
the IC50 for reversible inhibitor 3 determined on THP-1 cell
lysates (IC50 = 1.3 mm) was in perfect agreement with the IC50
on purified iCP (IC50 = 1.6 mm). Additionally, 3 showed 40-
fold selectivity for the inhibition of b5 activity in THP-1 cell
lysates over HeLa cell lysates.
The effects of 3, 30, and 42 on 26S proteasome activity in
intact THP-1 and HeLa cells were also measured by using the
Proteasome-Glo assay. These compounds retained good
selectivity for THP-1 cells (Table 2, Figure S5). For 3, the
inhibitory potency determined with purified CPs corre-
sponded well with the values from the Proteasome-Glo
assays (Figure 1a, Table 2). On the other hand, the IC50 values
for the irreversible compounds 30 and 42 were higher in the
Proteasome-Glo assays (Table 1 and Table 2). The data from
Table 2 indicate that the noncovalently acting psoralene 3 is
cell permeable and inhibits the immunoproteasome within
cells, whereas the covalently acting counterparts 30 and 42
permeate membranes to a lesser extent, but can selectively
inhibit iCP in the presence of other cytosolic components.
Importantly, compounds 3, 30, and 42 did not significantly
affect the viability of the THP-1 and HeLa cell lines at
concentrations up to 50 mm (Figure S6).
To determine whether reversible inhibitor 3 and its
irreversibly acting counterparts 30 and 42 can inhibit iCP in
the presence of other cytosolic components, cell lysates from
HeLa and THP-1 cells were incubated with 3, 30, and 42.
THP-1 cells were used since they are derived from acute
This lack of cytotoxicity is probably indicative of low off-
target binding. With off-target effects in mind, the structural
similarity with known intercalators (e.g., bergapten)
Angew. Chem. Int. Ed. 2016, 55, 5745 –5748
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5747

Angewandte
Communications
Chemie
[6] M. Groettrup, S. Khan, K. Schwarz, G. Schmidtke, Biochimie
2001, 83, 367 – 372.
prompted us to check possible intercalation-based interac-
tions of these molecules with DNA. Using a DNA binding
assay,[25] the randomly selected compounds 1, 3, and 30 did not
intercalate between DNA base pairs, thus excluding inter-
calation as a possible unwanted interaction (Figure S7).
In summary, a focused library of nonpeptidic reversible
b5i inhibitors was synthesized, followed by conversion into
irreversibly acting compounds. The latter retain selectivity for
b5i and provide more pronounced inhibition of b5i in THP-
1 cell lysates than in HeLa cell lysates. We also showed that
the reversible inhibitor 3 can enter human cells and inhibit the
proteasomes within them. The very marginal cytotoxic effects
of these compounds are an important feature for the future
development of nonpeptidic and b5i-selective probes to study
iCP involvement in diseases, such as inflammation or auto-
immune disorders. Of note, intensive work on peptidic
derivatives has recently culminated in very potent and b5i-
selective compounds;[26] however, we believe that the de-
scribed strategy of inhibitor design will aid in the future
discovery of structurally diverse nonpeptidic compounds and
will pave the way toward either more potent inhibitors or
selective second-generation probes that can be used to further
define the biology of the iCP and additionally uncover its
potential as a therapeutic target.
[7] T. A. Griffin, D. Nandi, M. Cruz, H. J. Fehling, L. V. Kaer, J. J.
Monaco, R. A. Colbert, J. Exp. Med. 1998, 187, 97 – 104.
[8] S. D. Demo, C. J. Kirk, M. A. Aujay, T. J. Buchholz, M. Dajee,
M. N. Ho, J. Jiang, G. J. Laidig, E. R. Lewis, F. Parlati et al.,
Cancer Res. 2007, 67, 6383 – 6391.
[9] P. G. Richardson, T. Hideshima, K. C. Anderson, Cancer Control
2003, 10, 361 – 369.
[10] F. Parlati, S. J. Lee, M. Aujay, E. Suzuki, K. Levitsky, J. B. Lorens,
D. R. Micklem, P. Ruurs, C. Sylvain, Y. Lu et al., Blood 2009,
114, 3439 – 3447.
[11] D. J. Kuhn, S. A. Hunsucker, Q. Chen, P. M. Voorhees, M.
Orlowski, R. Z. Orlowski, Blood 2009, 113, 4667 – 4676.
[12] T. Muchamuel, M. Basler, M. A. Aujay, E. Suzuki, K. W. Kalim,
C. Lauer, C. Sylvain, E. R. Ring, J. Shields, J. Jiang et al., Nat.
Med. 2009, 15, 781 – 787.
[13] M. Basler, M. Dajee, C. Moll, M. Groettrup, C. J. Kirk, J.
Immunol. 2010, 185, 634 – 641.
[14] A. F. Kisselev, M. Groettrup, Curr. Opin. Chem. Biol. 2014, 23,
16 – 22.
[15] E. M. Huber, M. Groll, Angew. Chem. Int. Ed. 2012, 51, 8708 –
8720; Angew. Chem. 2012, 124, 8838 – 8850.
[16] G. Sliwoski, S. Kothiwale, J. Meiler, E. W. Lowe, Pharmacol. Rev.
2014, 66, 334 – 395.
[17] V. Kasam, N.-R. Lee, K.-B. Kim, C.-G. Zhan, Bioorg. Med.
Chem. Lett. 2014, 24, 3614 – 3617.
[18] P. Beck, M. Reboud-Ravaux, M. Groll, Angew. Chem. Int. Ed.
2015, 54, 11275 – 11278; Angew. Chem. 2015, 127, 11428 – 11431.
[19] E. M. Huber, M. Basler, R. Schwab, W. Heinemeyer, C. J. Kirk,
M. Groettrup, M. Groll, Cell 2012, 148, 727 – 738.
[20] A. S. Kalgutkar, D. K. Dalvie, Expert Opin. Drug Discovery
2012, 7, 561 – 581.
[21] J. Singh, R. C. Petter, T. A. Baillie, A. Whitty, Nat. Rev. Drug
Discovery 2011, 10, 307 – 317.
[22] H. Fan, N. G. Angelo, J. D. Warren, C. F. Nathan, G. Lin, ACS
Med. Chem. Lett. 2014, 5, 405 – 410.
Acknowledgements
This work was supported by the Slovenian Research Agency.
We thank OpenEye Scientific Software, Inc (Santa Fe, NM,
USA; http://www.eyesopen.com) for free academic licenses
of their software. We also thank Chris Berrie for critical
reading of the manuscript.
[23] C. Dubiella, H. Cui, M. Gersch, A. J. Brouwer, S. A. Sieber, A.
Krüger, R. M. J. Liskamp, M. Groll, Angew. Chem. Int. Ed. 2014,
53, 11969 – 11973; Angew. Chem. 2014, 126, 12163 – 12167.
[24] D. Niewerth, G. J. L. Kaspers, Y. G. Assaraf, J. van Meerloo, C. J.
Kirk, J. Anderl, J. L. Blank, P. M. van de Ven, S. Zweegman, G.
Jansen et al., J. Hematol. Oncol. 2014, 7, 7.
[25] V. Pierroz, T. Joshi, A. Leonidova, C. Mari, J. Schur, I. Ott, L.
Spiccia, S. Ferrari, G. Gasser, J. Am. Chem. Soc. 2012, 134,
20376 – 20387.
Keywords: drug design · immunoproteasome · inhibitors ·
irreversible inhibition · small-molecule inhibitors
How to cite: Angew. Chem. Int. Ed. 2016, 55, 5745–5748
Angew. Chem. 2016, 128, 5839–5842
[1] A. L. Goldberg, Biochem. Soc. Trans. 2007, 35, 12 – 17.
[2] N. Gallastegui, M. Groll, Trends Biochem. Sci. 2010, 35, 634 –
642.
[26] G. de Bruin, E. M. Huber, B.-T. Xin, E. J. van Rooden, K. Al-
Ayed, K.-B. Kim, A. F. Kisselev, C. Driessen, M. van der Stelt,
G. A. van der Marel et al., J. Med. Chem. 2014, 57, 6197 – 6209.
[3] T. P. Dick, A. K. Nussbaum, M. Deeg, W. Heinemeyer, M. Groll,
´
M. Schirle, W. Keilholz, S. Stevanovic, D. H. Wolf, R. Huber
et al., J. Biol. Chem. 1998, 273, 25637 – 25646.
[4] S. Murata, K. Sasaki, T. Kishimoto, S. Niwa, H. Hayashi, Y.
Takahama, K. Tanaka, Science 2007, 316, 1349 – 1353.
[5] M. Groettrup, C. J. Kirk, M. Basler, Nat. Rev. Immunol. 2010, 10,
73 – 78.
Received: January 7, 2016
Revised: February 12, 2016
Published online: April 1, 2016
5748
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5745 –5748