Covalent docking modelling-based discovery of tripeptidyl epoxyketone proteasome inhibitors composed of aliphatic-heterocycles

Article abstract of DOI:10.1016/j.ejmech.2018.12.064

The potential of specific proteasome inhibitors to act as anti-cancer agents has attracted intensive investigations. The proteasome can be covalently inhibited by epoxyketone derivatives via a two-step reaction. Several computational approaches have been developed to mimic the covalent binding event. Compound 1 composed of a six-membered heterocyclic ring was designed by using covalent docking. With a possible different binding mode from the clinical compound Carfilzomib, it occupied the S5 pocket of 20S proteasome and showed favorable inhibitory activity. Subsequently optimization and evaluation were taken place. Among these compounds, 11h demonstrated extraordinary in vitro inhibitory activity and selectivity, and good in vivo proteasome inhibitory activity, a favorable pharmacokinetic profile and xenograft tumor inhibition. The possible binding pattern of compound 11h against proteasome was further fully explored via calculations, providing a theoretical basis for finding potent proteasome inhibitors.

Full text of DOI:10.1016/j.ejmech.2018.12.064

Accepted Manuscript
Covalent docking modelling-based discovery of tripeptidyl epoxyketone proteasome
inhibitors composed of aliphatic-heterocycles
Xiao-Wu Dong, Jian-Kang Zhang, Lei Xu, Jin-Xin Che, Gang Cheng, Xiao-Bei Hu, Li
Sheng, An-Hui Gao, Jia Li, Tao Liu, Yong-Zhou Hu, Yu-Bo Zhou
PII:
S0223-5234(18)31110-3
DOI:
https://doi.org/10.1016/j.ejmech.2018.12.064
EJMECH 10998
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 25 October 2018
Revised Date: 22 December 2018
Accepted Date: 25 December 2018
Please cite this article as: X.-W. Dong, J.-K. Zhang, L. Xu, J.-X. Che, G. Cheng, X.-B. Hu, L. Sheng, A.-
H. Gao, J. Li, T. Liu, Y.-Z. Hu, Y.-B. Zhou, Covalent docking modelling-based discovery of tripeptidyl
epoxyketone proteasome inhibitors composed of aliphatic-heterocycles, European Journal of Medicinal
Chemistry (2019), doi: https://doi.org/10.1016/j.ejmech.2018.12.064.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Covalent Docking Modelling-based Discovery of Tripeptidyl
Epoxyketone Proteasome Inhibitors Composed of Aliphatic-
Heterocycles
Xiao-Wu Dong,a,‡ Jian-Kang Zhang,a,d,‡ Lei Xu,b,c,‡ Jin-Xin Che,a Gang Cheng,e Xiao-Bei Hu,b,c
Li Sheng,b,c An-Hui Gao,b,c Jia Li,b,c Tao Liu,*,a,b Yong-Zhou Hu,*, a and Yu-Bo Zhou*,b,c
a ZJU-ENS Joint Laboratory of Medicinal Chemistry, Zhejiang Province Key Laboratory of Anti-
Cancer Drug Research, Hangzhou Institute of Innovative Medicine, College of Pharmaceutical
Sciences, Zhejiang University, Hangzhou, 310058, China
b
National Center for Drug Screening, State Key Laboratory of Drug Research, Shanghai
Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203, China
c University of Chinese Academy of Sciences, No. 19A Yuquan Road, Beijing 100049, PR China
d School of Medicine, Zhejiang University City College, Hangzhou 310015, China
e
College of Pharmaceutical Science Zhejiang Chinese Medical University, Hangzhou, 311402,
China
Corresponding Author
*For T. Liu: E-mail: lt601@zju.edu.cn
*For Y. Hu.: E-mail: huyz@zju.edu.cn
*For Y. Zhou: E-mail: ybzhou@simm.ac.cn
‡ X.D., J.Z. and L.X. contributed equally to this work.
1

ACCEPTED MANUSCRIPT
KEYWORDS: Proteasome inhibitors; tripeptide; covalent docking; interaction mode
Abbreviations: HOBt, 1-hydroxybenzotriazole; EDCI, 1-ethyl-3-(3'-dimethylaminopropyl)
carbodiimide hydrochloride; DIPEA, N,N-diisopropylethylamine; DCM, dichloromethane; THF,
tetrahydro-furan; DMF, N,N-dimethylformamide; (Boc)2O, di-tert-butyl decarbonate; TFA,
trifluoroacetic acid; AUC, Area under curve; MRT, Mean Residence Time; Cl, Clearance; Vss,
Steady-state apparent volume of distribution; RMSD, Root Mean Square Deviation.
Abstract
The potential of specific proteasome inhibitors to act as anti-cancer agents has attracted intensive
investigations. The proteasome can be covalently inhibited by epoxyketone derivatives via a
two-step reaction. Several computational approaches have been developed to mimic the covalent
binding event. Compound 1 composed of a six-membered heterocyclic ring was designed by
using covalent docking. With a possible different binding mode from the clinical compound
Carfilzomib, it occupied the S5 pocket of 20S proteasome and showed favorable inhibitory
activity. Subsequently optimization and evaluation were taken place. Among these compounds,
11h demonstrated extraordinary in vitro inhibitory activity and selectivity, and good in vivo
proteasome inhibitory activity, a favorable pharmacokinetic profile and xenograft tumor
inhibition. The possible binding pattern of compound 11h against proteasome was further fully
explored via calculations, providing a theoretical basis for finding potent proteasome inhibitors.
2

ACCEPTED MANUSCRIPT
1. Introduction
The ubiquitin-proteasome pathway is responsible for approximately 80-90% of eukaryotic
intracellular protein degradation1-4. In particular, accumulating evidences illustrate that various
regulatory proteins associated with cell growth, differentiation, and apoptosis in cancer cells are
controlled by the proteasome3, 4. Accordingly, proteasome inhibitors have gained extensive
interests as anti-cancer treatments5-7. To date, numerous peptidyl proteasome inhibitors, which
can be divided into boric acid, epoxyketone, vinyl sulfone and aldehyde derivatives, have been
identified and developed 5, 8. In general, the warhead of these compounds can be attacked by a
threonine residual of the proteasome, forming reversible/irreversible covalent bond8. To date,
several compounds (including MG-132, Bortezomib, Carfilzomib and Marizomib, Figure 1)
have been extensively evaluated as anti-cancer agents. However, adverse effects and drug
resistance have been frequently observed in clinical practice, including peripheral neuropathy,
9, 10
cardiovascular effects
and p-glycoprotein mediated drug resistance in certain cell lines11.
Thus, there remains a strong demand for the development of novel proteasome inhibitors with
enhanced potency, efficacy, fewer side effects and reduced drug resistance.
Figure 1. Structures of proteasome inhibitors
In eukaryotic proteasomes, three β-type subunits (β1, β2 and β5) of the 20S proteasome
contain proteolytically active centers12, 13, amongst which the β5 subunit has been shown to be
the rate-limiting proteolytic step. Crystal structures of bortezomib- and carfilzomib-bound
proteasomes revealed that both compounds bind to the β5 subunit, which is the primary target for
3

ACCEPTED MANUSCRIPT
proteasome inhibition14. Although the structure of the β5 subunit is known, structure-based drug
design for the development of new covalent proteasome inhibitors has not been performed in
detail, as calculating the binding energy for an irreversibly bound protein complex remains a
challenging issue.
Recently, computational approaches have been developed to mimic the covalent binding
event, resulting in several covalent molecular docking programs15, 16. These can be implemented
during computer-aided drug design to describe covalent interactions between inhibitors and
biological targets. Using this method, we can successfully predict the binding energy between a
nucleophilic receptor and electrophilic ligand. Blake et al.17, Schroeder et al.18 and Taunton et
al.19 performed pioneering studies in which a covalent docking method was applied during
virtual drug screening to discover new irreversible inhibitors, revealing a new strategy for
covalent inhibitors drug design. For the proteasome, Zhu et al.20 and Zhang et al.21 demonstrated
the binding pockets and key residues of the 20S proteasome that can be targeted for, providing a
potential way to design more potent inhibitors of the 20S proteasome. However, the binding
mode between the 20S proteasome and peptide epoxyketone (EK) was not fully explored, despite
disclosure of the crystal structure of carfilzomib-bound to the proteasome. It was reported that
the EK forms two covalent bonds with the proteasome via a two-step reaction22, 23. As shown in
Figure 2, the hydroxyl group of a Thr1 residue in the 20S proteasome attacks the ketone (O-C
bond formation) of EK fragment, leading to the formation of a Hemiketal intermediate. The
amino group then attacks the epoxy group from either side (N-C bond formation), generating a
stable six-membered 1,4-morpholine ring or seven-membered 1,4-oxazepane ring22, 23. It would
be of interest to perform covalent docking studies between the 20S proteasome and EK-
containing proteasome inhibitors, this can provide guidance for altering the skeleton and/or
4

ACCEPTED MANUSCRIPT
sidechains of EK-containing proteasome inhibitors.
Figure 2. Covalent bond forming process of epoxyketone to proteasome (“B” means base)
Based on the crystal structure of the carfilzomib-bound proteasome complex (PDB: 4R67),
and as part of our continuous studies on molecular docking and the discovery of proteasome
inhibitors.5, 24-31 In this study, we designed EK derivatives as proteasome inhibitors based on the
usage of covalent docking approaches. This led to the identification of a shortened tripeptide
skeleton and more flexible terminal moiety that may occupy a new pocket that differs from the
binding mode of Carfilzomib (Figure 3(A)). We fully explored the structure-activity
relationships of these new tripeptide derivatives, in vitro and in vivo proteasome inhibitory
activities. Furthermore, the drug profiles and in vivo anticancer activity of a promising candidate
(compound 11h) was extensively evaluated, which was found to be more potent than
carfilzomib. Finally, we fully explored the covalent bond forming process between 11h and the
proteasome, resulting in a stable 11h-bound proteasome complex linked by a seven-membered
1,4-oxazepane ring.
2. Results and discussion
2.1 Covalent docking-based rational design of tripeptidyl epoxyketone derivatives as
proteasome inhibitors
According to the crystal structure of the human 20S proteasome in complex with carfilzomib
(PDB: 4R67) and through the analysis of different binding pockets22, the interaction mode
5

ACCEPTED MANUSCRIPT
indicated that the hydrophobic P4 phenyl group occupied the S4 pocket, the L-leucine residue P3
pointed into the S3 pocket, the S2 pocket was able to accommodate the P2 phenylalanine moiety
and the P1 leucyl group of carfilzomib was oriented towards the S1 pocket, which maintained a
high binding affinity of carfilzomib to proteasome besides other interaction forces including
hydrogen and covalent bond. Whilst the N-tail morpholinyl of carfilzomib was forced to turn
towards the β sheet S3 due to steric effects caused by the side chains of β3Glu106 and β6Tyr107,
which lead to the lack of occupancy of the hydrophilic S5 pocket formed by Lys136, Tyr107,
Ser142 and Tyr2522, as shown in Figure 3(A). We reasoned that shortening the tetrapeptide of
carfilzomib to a tripeptide through removal of the P4 L-homophenylalanine moiety and
introducing a more flexible terminal moiety to occupy the S5 pocket would represent an effective
approach for developing new inhibitors. As shown in Figure 3(B), the N-tail of the tripeptide
was substituted with the Linker and R groups which were selected as heterocyclic alkyl or aryl
rings for the purpose of extending the N-tail to the S5 pockets and improving the molecular
pharmacokinetic properties. To validate our preliminary assumptions, the piperazine was chosen
as the linker because of its flexible and linear characteristics to facilitate the N-tail reaching S5
pocket, and the six-membered pyrazine was employed as the N-tail to occupying the S5
hydrophilic pocket, then compound 1 was therefore generated to prove this hypothesis (Figure
3(C)). As the first step (O-C bond) is rate-limiting for dual covalent bond formation, and also for
the convenience of calculation, we only assessed O-C bond formation at the initial stage. The
result predicted that compound 1 formed hydrogen bonds with Tyr21, Ala49 and Ser142. In
addition, the P1 leucyl group of compound 1 oriented to the S0 pocket formed by Thr21 and
Tyr269, the P2 phenyl pointed into the S1 pocket, the P3 leucyl group interacted with the S4
pocket, and the S5 pocket was occupied by the terminal piperazinyl moiety, as shown in Figure
6

ACCEPTED MANUSCRIPT
3(C). This predicted binding mode was different to carfilzomib. From mapping results of
compound 1 and carfilzomib, the orientation and shape of their terminal moieties differed due to
the overall linear shape of the piperazin-1-yl (pyrazin-2-yl) methanone moiety of compound 1.
Furthermore, the binding energy of the ligand was predicted using the MMGBSA method and
found to be -55.383 kcal/mol, the docking score was -7.971, which indicated a favorable binding
mode. Compound 1 was then synthesized and biologically evaluated, and showed an IC50 value
of 91 nM. Although the inhibitory potency was 10 times lower than that of carfilzomib, the
designed skeleton of compound 1 was first reported and reliable. Thus, systematic structure
optimization, including different Linkers and R groups were performed to identify more potent
inhibitors (Figure 3(B)).
7

ACCEPTED MANUSCRIPT
Figure 3. (A) Cocrystal structure of Carfilzomib bound to the proteasome (PDB: 4R67). (B)
Designed tripeptide epoxyketone containing heterocycle derivatives. (C) Predicted mono bond
binding mode of compound 1 to the proteasome
2.2 Chemistry
As depicted in Schemes 1, target compounds 2-4 with the Phe-Leu-Leu-EK skeleton were
produced by condensing Leu-Leu-EK (for synthesize strategy see support information Scheme
S1) with compounds 38a, 39a and 45a. Compounds 1, 5, 6a-c, 7a-b, 8a-b, 9a-b, 10a-c, 11a-q
with a Leu-Phe-Leu-EK skeleton were prepared in the same manner to above (for the synthesis
strategy of 38a-b, 39a-b, 40a-c, 41a-b, 42a-b, 43a-b, 44a-c and 45a-r see support information
Scheme S2).
8

ACCEPTED MANUSCRIPT
Scheme 1. Synthetic route of tripeptide epoxyketone derivatives. (a) HOBt, EDCI, DIPEA,
DCM, rt.
2.3 In vitro proteasome inhibitory activity
Table 1. Assessment of proteasome activity in the presence of compounds with varying linkers
9

ACCEPTED MANUSCRIPT
Inhibitory rate at
100ng/mL (%)a
Compounds
Linker
R
IC50 (nM)a
NTb
pyrazin-2-yl
43.7±4.1
2
L1
L2
L3
L1
pyrazin-2-yl
pyrazin-2-yl
pyrazin-2-yl
27.0±0.7
NTb
3
4
5
-
-
40.1±2.6
75.8±8.9
pyrazin-2-yl
4-chlorophenyl
4-methoxyphenyl
4-chlorophenyl
-
-
-
-
29.8±5.6
14.5±5.4
24.1±11.9
13.7±4.3
6a
6b
6c
7a
L4
L5
4-methoxyphenyl
4-chlorophenyl
-
-
-
16.7±1.9
25.6±3.4
31.6±1.5
7b
8a
8b
4-methoxyphenyl
L6
L7
4-chlorophenyl
-
-
59.9±3.0
63.8±1.9
9a
9b
4-methoxyphenyl
pyrazin-2-yl
4-chlorophenyl
4-methoxyphenyl
pyrazin-2-yl
4-chlorophenyl
4-methoxyphenyl
-
-
-
-
-
-
-
-
-
355.2±15.1
782.6±45.8
743.1±30.8
6.5±0.6
10a
10b
L8
L3
10c
11a
4.6±0.4
11b
4.8±0.3
11c
Bortezomib
Carfilzomib
-
-
10.8±0.9
8.6±1.4
-
a. The inhibitory rate and IC50 values are an average of three independent determinations.
10

ACCEPTED MANUSCRIPT
b. NT = not tested.
The results of proteasome inhibition assay are shown in Table 1. Considering the exploration of
skeleton diversities and the hydrophobic properties of S2-3 pockets, another tripeptide skeleton
containing hydrophobic amino acid residues Phe-Leu-Leu-EK was also investigated. The results
indicated that the skeleton of Leu-Phe-Leu-EK was superior to Phe-Leu-Leu-EK through the
comparison of compound 2 with 5, compound 3 with 1, and compound 4 with 11a. The Leu-Phe-
Leu-EK skeleton was chosen for further investigation of different Linkers and R groups.
Figure 4. Comparison of Phe-Leu-Leu-EK and Leu-Phe-Leu-EK skeletons
Based on the designed structure of compound 1, different compounds containing aliphatic-
heterocycle linkers were synthesized and tested. The selected R groups were substituted with
phenyl or pyrazin-2-yl. Compounds 10a-c with an L8 Linker exhibited the least inhibitory
potency against the proteasome. Compounds 6b-c, 7a-b, 8a-b, 9a-b with linkers L4-L7 exhibited
IC50 values of 10-100 nM. Compounds with an N-tail of 4-chlorophenyl displayed better
inhibitory potency than 4-methoxyphenyl. Interestingly, when the linker was L3, compounds
11a, 11b and 11c displayed IC50 values of 6.5 nM, 4.6 nM and 4.8 nM respectively, which are
comparable to the inhibitory activity of bortezomib and carfilzomib. Therefore, L3 was chosen as
the linker for following SAR studies of the N-tail R groups.
Table 2. Assessment of proteasome activity in the presence of compounds with varying R groups
11

ACCEPTED MANUSCRIPT
Compounds
11d
R
IC50 (nM)a
3.4±0.4
4-fluorophenyl
4-benzoylphenyl
biphenyl-4-yl
isoxazol-3-yl
thiazol-2-yl
1,3,4-thiadiazol-2-yl
benzothiazol-2-yl
pyridin-3-yl
pyridin-4-yl
pyrimidin-2-yl
methyl
2.4±0.3
11e
9.4±1.2
11f
6.4±0.2
11g
1.5±0.2
11h
11i
4.6±0.4
3.9±0.3
11j
3.6±0.5
11k
9.5±1.2
11l
7.7±0.6
11m
232.2±10.2
484.1±33.4
92.0±9.4
1089.6±81.8
10.8±0.9
8.6±1.4
11n
morpholine-4-yl
morpholinoethyl
2-hydroxyethyl
-
11o
11p
11q
Bortezomib
Carfilzomib
-
a. The IC50 values are an average of three independent determinations.
As shown in Table 2, amongst the compounds with different R groups, a dramatic loss of
inhibitory potency was observed when R was substituted with aliphatic chains or saturated
12

ACCEPTED MANUSCRIPT
heterocyclic rings (compound 11n-11q). The inhibitory activities of compounds with aryl
heterocyclic substitutions (compounds 11g-11m) were comparable to phenyl substitution
(compounds 11d-11f). Compound 11h exhibited an IC50 value of 1.5 nM, which was 7-fold
improved potency over bortezomib and 6-fold improved potency over carfilzomib. The most
potent proteasome inhibitor identified, 11h, was selected for further biological evaluation. In
addition, covalent modelling was employed to predict the potential interaction mode between
11h and proteasome.
The MMGBSA dG binding score and docking score of the potent compound 11h were -45.837
kcal/mol and -4.681, respectively, which were favorable for binding. The predicted binding poses
of 11h in the proteasome are shown in Figure 5, which was similar to compound 1, with the N-
tail thiazol-2-yl occupying the S5 hydrophilic pocket which formed by Lys136, Tyr107, Ser142
and Tyr25. Two additional hydrogen bonds formed with Gly48 and Ala49 were observed.
Figure 5. (A) Surface view of proteasome bound to compound 11h. (B) Docking poses of
compound 11h bound to the proteasome
2.4 Selectivity evaluation of compound 11h
The selectivity of compound 11h against other types of proteasome and protease was also
evaluated. As shown in Table 3, the inhibitory activities were determined by monitoring the
decreases in hydrolysis of the fluorogenic substrate Z-LLE-AMC (for caspase-like proteasome
13

ACCEPTED MANUSCRIPT
β2 subunit), Suc-LLVY-AMC (for constitutive proteasomeβ5 subunit) and Suc-LLVY-AMC
(for immunoproteasome β5i subunit). The inhibitory activity of compound 11h against
constitutive proteasome was more potent than that of bortezomib and carfilzomib. In addition,
compound 11h also showed extraordinary selectivity to the immuno proteasome, caspase-like
proteasome subunit and other proteases such as thrombin, Matrix metalloproteinase-12
(MMP12), Dipeptidyl Peptidase-4 (DPP4), caspase3 and trypsin.
Table 3. Inhibitory activity of compound 11h against other types of proteasome and proteases
Proteasome inhibitory activity (IC50,
%Inhibition (10 µg/mL)a
nM)a
Compd.
Constitutive Immuno Caspase-like
Thrombin MMP12 DPP4 Caspase3 Trypsin
Bortezomib
Carfilzomib
10.8±0.9
8±2
128±26
1812±246
1263±321
-30.5%
-2.2%
-8.4%
2.7%
21.2%
2.8%
33.7%
1%
17.4%
8.4%
9.2%
8.6±1.4
34± 3
32±1
-1.79% 18.6%
1.5±0.2
-5.0%
20.4%
11h
a. The values are an average of three independent determinations.
2.5 Anti-proliferation activities against cancer cell lines
Based on the evaluation of proteasome inhibitory activities, compound 11h was further
assessed for its anti-proliferative activity against different cancer cell lines (Table 4). The IC50 of
compound 11h against RPMI 8226, MM.1S, MM.1R multiple myeloma cells and HCT116 colon
cancer cells were below 10 nM. Compound 11h was more potent than Carfilzomib in several cell
lines, including LP1 cholangiocarcinoma cells, H929 multiple myeloma cells, PC-3 prostatic
cancer cells and MG803 gastric cancer cells.
Table 4. Antiproliferative activities of compound 11h against different tumor cell lines
14

ACCEPTED MANUSCRIPT
Compounds (IC50, nM)
Cancer cells
Carfilzomib
2.1±0.1
11h
RPMI 8226
LP1
2.2±0.2
22.7±2.4
2.3±0.2
12.7±3.7
1.8±0.3
MM.1S
MM.1R
H929
1.8±0.2
2.2±0.3
116.1±25
9.6±1.2
51.7±11.7
9.8±0.8
HCT116
PC-3
23.9±2.8
934.0±89.5
43.3±2.9
11.6±1.2
450.3±26.8
25.9±3.6
MG803
A549
2.6 In vivo pharmacokinetic evaluation
It was important to assess the pharmacokinetic properties of the lead compounds prior to in vivo
efficacy assessments. Because of the low oral bioavailability of peptide proteasome inhibitors,
was the pharmacokinetic profile of compound 11h was evaluated by using intravenous
administration (i.v) (Table 5). The pharmacokinetic parameters including AUC, MRT, t1/2 and
Vss suggested an improved performance of compound 11h over carfilzomib.
Table 5. Pharmacokinetic parameters of intravenous administered compound 11h
Compounds
Parameters
Carfilzomib
21311
11h
AUC0-t (min*ng/ml)
33007
15

ACCEPTED MANUSCRIPT
AUC0-∞ (min*ng/ml)
33212
6.09
21425
3.57
MRT (min)
t1/2 (min)
10.59
150.5
0.916
5.18
CL (L/min/kg)
Vss (L/kg)
233.4
0.834
2.7 In vivo proteasome inhibition assay
Based on these results, compound 11h was selected for the in vivo evaluation of the inhibitory
activity against proteasome. BALB/c mice (n= 3 per dose group) received i.v. administration at
varied doses, then the blood samples were collected and the proteasome activity was measured.
The results were shown in Figure 6(A). After 1h of intravenous administration, both compound
11h and Carfilzomib inhibited the proteasome to comparable levels. However, compound 11h
displayed longer and more sustained inhibition than carfilzomib after 24 and 96 h post-
administration, indicating that the compound may function through irreversible binding.
Furthermore, it is demonstrated that the inhibitory activity can be maintained even after 100
times dilution of the compounds (Figure 6B), revealing that the irreversible binding of 11h to
proteasome was not affected.
Figure 6. (A) In vivo proteasome inhibitory activity evaluation of Carfilzomib and 11h. (B) The
16

ACCEPTED MANUSCRIPT
proteasome inhibitory activity assessment after dilution of Carfilzomib and 11h. The values are
an average of three independent experiments. The values were an average of three independent
determinations.
2.8 In vivo efficacy of xenograft tumor inhibition
Based on the biological and pharmacokinetic evaluation, further in vivo efficacy test of xenograft
tumor inhibition was performed. The multiple myeloma RPMI 8226 NOD/SCID tumor was
implanted into mice and compound 11h was administered i.v for 18 days (Figure 7). The tumor
inhibition rates were 54.41% (4 mg/kg i.v.) and 65.61% (5 mg/kg i.v.) respectively, indicating a
significant effect (P<0.01) through comparison to the 0.5% CMC-Na control group. The tumor
inhibition rate was comparable to carfilzomib (4mg/kg i.v., 58.10%), and better than bortezomib
(1 mg/kg i.v., 34.78%). The body weight of the mice was unaffected by compound 11h, while the
body weight of mice declined and two mice died in Carfilzomib group, indicating that compound
11h has a better safety profile. These studies confirmed that the RPMI 8226 NOD/SCID
xenograft tumor can be effectively controlled by compound 11h.
17

ACCEPTED MANUSCRIPT
Figure 7. (A) The influences of compounds on relative tumor volume. (B) The influences of
compounds on mice body weight. (C) Images of xenograft tumor after dosing for 18 days.
2.9 Dual covalent bond forming calculation
In order to further investigate and validate the binding mode of compound 11h, the two steps
covalent bond forming reactions were fully predicted. If we input the entire compound-protein
complex, the computer would suffer high complexity and burden for calculation. With the aim of
assessing whether compound 11h will form dual covalent bond with proteasome under the mono
bond binding circumstance and providing a reference binding mode, we decided to simplified the
compound 11h-proteasome mono bond complex to simplify the calculation. The complex was
then inputted into the calculation program and an 1,4-oxazepane ring was generated (Figure 8A),
the dynamic progress of dual covalent bond forming process was also recorded as cartoon (see
18

ACCEPTED MANUSCRIPT
Support File, open with the opensource Jmol software).
After the dual covalent bond forming process was completed, the remaining part of 11h-
proteasome complex was re-added to the dual covalent bond complex and dynamic optimization
was performed. As shown in Figure 8(B), the RMSD value of 11h-proteasome complex was
calculated from 0 to 2000 ps trajectory, and reached equilibration conditions and swings around
an average value after 400 ps simulation time. Based on the stable conformation obtained from
MD simulation, the interactions involved in the complex were analyzed. As shown in Figure
8(C) & (D), the P1 leucyl group hydrophobic interacted with Ala46 and Gly47 residues, the P2
phenyl pointed to S0 pocket, the P3 leucyl group interacted with S4 pocket and S5 pocket was
occupied by N-tail thiazol-2-yl. Beside the covalent bond, compound 11h also formed two H-
bonds with Thr21 and Gly47. Therefore, we assumed that compound 11h may display a different
binding pattern compared to that of carfilzomib.
19

ACCEPTED MANUSCRIPT
Figure 8. (A) The output of 1,4-oxazepane ring formation. (B) The RMSD value during the 2000
ps simulation time. (C) Surface view of proteasome bound to compound 11h. (D) Binding pattern
of compound 11h.
3. Conclusion
Through the investigation of the interaction mode of carfilzomib with proteasome, compound 1
composed of a six-membered heterocyclic ring was designed on the basis of covalent docking,
further biological evaluation demonstrated a promising proteasome inhibitory potency. After
initial exploration, a series of tripeptidyl EK derivatives were synthesized and tested. Compound
11h was selected for further evaluation on selectivity, anti-tumor proliferation, in vivo
proteasome inhibition, pharmacokinetics and xenograft tumor inhibition, showing improved
biological properties compared to carfilzomib or bortezomib. In addition, the covalent bond
20

ACCEPTED MANUSCRIPT
forming process of compound 11h with proteasome was fully investigated. Further dynamic
optimization disclosed the binding mode of 11h to the proteasome, which was possibly a
different interaction mode from that of Carfilzomib. Compound 11h is currently in pre-clinical
development in our laboratory.
4. Materials and Methods
4.1 Chemistry
General experimental information
All reagents and solvents were used as purchased from commercial sources. Chromatography
was performed using silica gel (200-300 mesh). All reactions were monitored by TLC, using
1
silica gel plates with fluorescence F254 and UV light visualization. H NMR and 13C NMR
spectra were obtained on a Bruker Advance III 500 with use of CDCl3, CD3OD, (CD3)2CO or
DMSO-d6 as solvent. Chemical shifts are referenced to the residual solvent peak and reported in
ppm (d scale) and all coupling constant (J) values are given in Hz. The following multiplicity
abbreviations are used: (s) singlet, (d) doublet, (t) triplet, (q) quartet, (m) multiplet, and (br)
broad. ESI-MS data were recorded on Esquire-LC-00075. Melting point data were recorded on
Buchi B-540.
General procedure A for preparation of amide compounds (14a, 15a, 16a-c, 30a-b, 33a-b, 1-
5, 6a-c, 7a-b, 8a-b, 9a-b, 10a-c, 11a-q).
To a solution of the acid (1.2 mmol) in DCM was added HOBt (1.2 mmol) and EDCI (1.8
mmol). After stirring at room temperature for 30 min, the resulting mixture was added amine
hydrochloride (1 mmol) and N,N-Diisopropylethylamine (2 mmol). The reaction mixture was
further stirred at room temperature for 3 h and quenched with saturated sodium bicarbonate. The
organic layer was collected, washed with saturated brine, dried over anhydrous Na2SO4 and
concentrated to afford crude product. The crude product was purified by silica gel
chromatography.
General procedure B for preparation of urea compounds (17a-b, 18a-b, 19a-b, 31a-b, 32a-c,
33a-b, 34a-b, 35a-b, 36a-c, 37a-r).
21

ACCEPTED MANUSCRIPT
To a solution of the primary amine (10 mmol) in DCM (30 mL) was added triphosgene and
triethylamine (3 mL) slowly. The reaction mixture was stirred at room temperature for 3 h and
concentrated. The residue was then dissolved in 20 mL DCM, to the solution was added
secondary amine (10 mmol). After stirring at room temperature for 2 h, the reaction mixture was
washed with saturated sodium bicarbonate (20 mL×1), dried over anhydrous Na2SO4 and
concentrated. The residue was purified by silica gel chromatography.
General procedure C for preparation of amide compounds (20a-c, 21a-q).
To a 50 mL three-necked flask was added piperidine-carboxylic acid (12 mmol) and anhydrous
DCM (25 mL), followed by the addition of pyridine(30mmol) and SOCl2 (14 mmol) under N2
atmosphere. After stirred at R.T. for 30min, the mixture was added primary amine and
triethylamine dissolved in DCM (15mL) dropwise. The reaction was further stirred at RT for 6 h
and diluted with brine (30 mL). The organic layer was collected, and the aqueous layer was
extracted with DCM. The combined organic layers were dried over anhydrous Na2SO4and
concentrated. The residue was purified by silica gel chromatography.
General procedure D for preparation of acid compounds (22a, 25a-b, 38a-b, 39a-b, 40a-c,
42a-b, 43a-b, 44a-c, 45a-r).
To a solution of ester (5 mmol) in acetone was added LiOH (0.5 N, 20 mL) dissolved in water.
The reaction mixture was stirred at room temperature for 0.5 h. After the acetone was removed
under reduced pressure, the aqueous solution was acidized to pH =2–3 with 1 N HCl, then
extracted with ethyl acetate (20 mL×3). The combined organic layer was dried over Na2SO4 and
concentrated to afford white solid. The product can be used directly in next step without further
purification.
General procedure E for de-Boc protection (23a, 24a-c, 26a-b, 27a-b, 28a-c, 29a-q)
To a solution of N-Boc-protected compound (5 mmol) in DCM (20 mL) was added CF3COOH
(5 mL). The resulting mixture was stirred at room temperature for 1h. The solvent was removed
under reduced pressure to afford colorless oil. The crude product can be directly used in the next
step without further purification.
4.2 Docking calculation
22

ACCEPTED MANUSCRIPT
Mono covalent bond formation docking via Maestro 11.1.011
Proteasome protein (PDB: 4R67) was prepared for covalent docking calculations using the
Protein Preparation Wizard utility. Molecules were prepared using the LigPrep module. Both the
energy of protein and the ligand structures were minimized using the OPLS3 force field. The
CovDock protocol was used to predict the pose of the covalently bound in the active site. The
reactive residue was set as Thr1. A thorough pose prediction docking mode was used. The
MMGBSA dG binding score was recorded. Other parameters were maintained at the default
configuration.
Dual covalent bond formation docking via MOPAC
With the prepared complex, we applied MOPAC package32 to simulate the process of epoxide
ring-opening reaction. First hydrogen atoms were added to the system model by running the
keyword of ‘ADD-H’, next the transition state for the epoxide ring-opening reaction has been
successfully located by using these keywords “the mozyme ts charge=0”.33 Afterwards, to
verified the transition state, we run following keywords “mozyme force oldgeo charge=0” and
“mozyme irc=1* oldgeo x-priority=0.2 charge=0” and obtained Intrinsic Reaction Coordinates
for this reaction. The resulting trajectory (.xyz filetype) was visualized by Jmol (an open-source
Java viewer for chemical structures in 3D. http://www.jmol.org/).
Dynamic optimization via Discovery Studio 2.5
The docked structures of compound 11h complexed within the active pocket of proteasome were
used as the initial structures for MD calculations. A CHARMm force field was applied to the
complex and the resulting system was subjected to double-fold minimization (10,000 cycles of
steepest descent minimization and 10,000 cycles of conjugate gradient minimization). The
system was heated from 50 K to 300 K over a period of 2 ps and subsequently equilibrated for
200 ps. Starting from the last frame of the equilibration, a production simulation was performed
for 2000 ps using the NPT ensemble under a constant temperature of 300 K and pressure of 1
atm. Other parameters of MD simulation were maintained at the default Discovery Studio
configuration.
4.3 Biological assays
23

ACCEPTED MANUSCRIPT
20S Constitutive and Immuno Proteasomes inhibition assay
The human constitutive proteasome was given by Dr. Jiang-ping Wu (Notre-Dame Hospital,
Montreal, QC, Canada), The inhibition of enzyme activity was determined by monitoring the
decrease in hydrolysis of the fluorogenic substrate Z-LLE-AMC for caspase-like(β2) and Suc-
LLVY-AMC for Chymotrypsin-like(β5). The immunoproteasome purchased by Boston
Biochem. The inhibition of enzyme activity was determined by monitoring the decrease in
hydrolysis of the fluorogenic substrate Suc-LLVY-AMC for CT-L(β5i). The reaction under the
following buffer conditions: 20 mM Tris,pH 7.5. Inhibitors at a different concentration were
incubated with constitutive proteasome or immunoproteasome proteasome for 15 min prior to
substrate addition with a final concentration of 50 µM. The reaction was monitored using an
excitation wavelength of 355 nm and an emission wavelength of 460 nm using Envision
microplate reader (PerkinElmer). The dose response of inhibition test was carried out in
duplicate. And the IC50 data was calculated using the software GraphPad Prism 5, and the
equation “sigmoidal dose-response (variable slope)” was chosen for curve fitting.
General procedure for protease inhibition assay
To evaluate the selectivity of 11h, a single concentration (10 µg/mL) was tested using our panel
including 5 kinases and 5 proteases. The 11h inhibition for JAK2、Aurora A、PKC-ε and IKK-
β were performed using commercially available homogeneous time-resolved fluorescence assay
kits (Cisbio International, Bagnols/Cèze, France). The 11h inhibition for Drak2 was tested using
ADP-Glo Kinase Assay (Promega, Madison, Wis). The 11h inhibition for Trypsin was evaluated
according to the experimental procedure of Sigma Aldrich. In enzymatic Assay of MMP12,
Enzyme activity was measured using light absorption detection. MMP12 hydrolyzed the
substrate Ac-Pro-Leu-Gly-[(S)-2-mercapto-4-methyl-pentanoyl]-Leu-Gly-Oet and then the
product was monitored by light absorption at 412 nm by microplate reader. The change in
intensity is calculated and the initial velocity of the reaction is calculated. Caspase-3 was assayed
with a substrate (AC-DEVD-AMC). The fluorescence generated was monitored at wavelengths
355nm(excitation) and 460nm(emission) using Envision (PerkinElmer). To measure the activity
of DPP4, a continuous fluorometric assay was employed using Ala-Pro-AMC, which is cleaved
by the enzyme to release the fluorescent aminomethylcoumarin (AMC). the AMC was monitored
using an excitation wavelength of 355 nm and an emission wavelength of 460 nm using Envision
24

ACCEPTED MANUSCRIPT
microplate reader (PerkinElmer). The kinetics of thrombin inhibition was determined from the
hydrolysis reaction of a specific Thrombin Substrate Ⅲ (CALBIOCHEM). The hydrolysis rate
was monitored using an excitation wavelength of 355 nm and an emission wavelength of 460 nm
using Envision microplate reader (PerkinElmer). The above enzyme activity experiments all use
three duplicates.
Cell culture
The MM.1S, MM.1R, RPMI 8226, LP1, H929, HCT116, A549 cell lines were obtained from the
American Type Culture Collection (ATCC) (Manassas, VA, USA), PC-3 and MG803 cell lines
were obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences.
Cells were maintained at 37 °C in a humidified atmosphere containing 5% CO2, cultured in
RPMI-1640 media supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and
100 µg/mL streptomycin from Invitrogen (Grand Island, NY, USA). , Except HCT116
in McCoy‘s 5a Medium Modified, and A549 in F12 media supplemented with 10% fetal bovine
serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin.
General procedure for tumour cell anti-proliferation assay
An 80 µL aliquot of cells was seeded into 96-well plates, and then treated with 20 µL of 0.2%
DMSO or varying concentrations of tested compounds for 72 h. Cell viability was measured
using the CellTiter 96® AQueous Non-Radioactive Cell Proliferation Assay (MTS; Promega,
Madison, WI). First, 20 µL of the combined MTS/PMS solution was pipetted into each well of
the 96-well plates and then incubated for 2-4 h at 37 °C in a humidified, 5% CO2 atmosphere.
The optical density was determined at 490 nm (background subtraction at 690 nm) using a
SpectraMax 340 microplate reader (Molecular Devices, Sunnyvale, CA, USA). The growth
inhibitory ratio was calculated as follows: Growth inhibitory ratio = (Acontrol - Asample)/Acontrol. IC50
values were derived from a nonlinear regression model (curvefit) based on a sigmoidal dose
response curve (variable slope) and computed using GraphPad Prism version 5.02, GraphPad
Software.
In vivo pharmacokinetic evaluation
25

ACCEPTED MANUSCRIPT
This study was performed in strict accordance with the laboratory animal management
regulations and under the rules and principles of the international guide for biomedical research
in experimental animals (State Scientific and Technological Commission Publication No. 8-27
Rev. 2017) and was approved by Shanghai Model Organisms Center,Inc (Shanghai, China).
Male ICR mice were intravenous administered of 5 mg/kg 11h (dose vehicle: 1% DMSO, 10%
Solutol, 10% EtOH, 75% Saline;). Blood samples were collected at 2, 5, 10, 15, 30 and 60 min
after dosing. The whole blood was centrifuged to yield plasma samples for analysis by LC-
MS/MS.
In vivo proteasome inhibition assay
This study was performed in strict accordance with the laboratory animal management
regulations and under the rules and principles of the international guide for biomedical research
in experimental animals (State Scientific and Technological Commission Publication No. 8-27
Rev. 2017) and was approved by Shanghai Model Organisms Center,Inc (Shanghai, China).
BALB/c mice (n = 3 per dose group) received i.v. administration of varied doses in 0.5% CMC-
Na. At selected time (1 h, 24 h and 96 h) points after administration, blood samples were
collected and processed to cell lysates, and proteasome activity was measured using LLVY-
AMC as a substrate, as the 20S proteasome chymotrypsin-like inhibition assay described.
Classical bulk dilution method
Bulk dilution is a convenient way to evaluate the reversibility of inhibition. In brief, 100-fold
over the concentration required for the 20S Constitutive Proteasomes activity assay and 10-fold
IC50 concentration of 11h were mixed. After 30 minutes, this mixture was diluted 100-fold with
reaction buffer. Then, the substrate was added to initiate the reaction and the kinetic of enzyme
was the same as 20S Constitutive assay system.
In vivo efficacy of xenograft tumor inhibition
CB-17 SCID mice (female,17-23 g) were obtained from Beijing Vital River Laboratory Animal
Technology Co., Ltd. (Shanghai, People，s Republic of China). The experiments were conducted
in full compliance with the Guide for Care and Use of Laboratory Animals and approved by the
Shanghai Model Organisms Center,Inc. 100 µL MM.1S cells(1×107 cells/mouse) were delivered
26

ACCEPTED MANUSCRIPT
by subcutaneous injection into the right flank of the test animal. Tumor volume was estimated
using the standard formula:(length×width2)/2. Once tumors reached a group mean of 100 to 300
mm3, animals were randomized to the following treatment groups (n = 6 per group): Bortezomib
(1mg/kg, i.v. twice weekly for three weeks on days 1 and 4), Carfilzomib (4mg/kg, i.v. twice
weekly for three weeks on days 1 and 2), 11h (4mg/kg, 5mg/kg, i.v. twice weekly for three
weeks on days 1 and 2) and control group (n = 12, i.v. twice weekly for three weeks on days 1
and 4). Mice were monitored daily and tumor volumes were measured twice weekly. The study
lasted for 21 days. After the final treatment, mice were sacrificed and tumor were excised and
weighed.
Supporting Information
1
13
Additional synthetic procedures, compounds information, H NMR and C NMR spectra are
available in supplementary materials.
References
1. Etlinger, J. D.; Goldberg, A. L. Soluble Atp-Dependent Proteolytic System Responsible for
Degradation of Abnormal Proteins in Reticulocytes. P. Natl. Acad. Sci. USA 1977, 74, 54-58.
2. Lee, D. H.; Goldberg, A. L. Proteasome inhibitors: valuable new tools for cell biologists.
Trends Cell Biol. 1998, 8, 397-403.
3. Naujokat, C.; Hoffmann, S. Role and function of the 26S proteasome in proliferation and
apoptosis. Lab. Invest. 2002, 82, 965-980.
4. Adams, J. The proteasome: A suitable antineoplastic target. Nat. Rev. Cancer 2004, 4, 349-
360.
5. Zhang, J. K.; Wu, P.; Hu, Y. Z. Clinical and Marketed Proteasome Inhibitors for Cancer
Treatment. Curr. Med. Chem. 2013, 20, 2537-2551.
6. Qin, J. Z.; Ziffra, J.; Stennett, L.; Bodner, B.; Bonish, B. K.; Chaturvedi, V.; Bennett, F.;
27

ACCEPTED MANUSCRIPT
Pollock, P. M.; Trent, J. M.; Hendrix, M. J. C.; Rizzo, P.; Miele, L.; Nickoloff, B. J.
Proteasome inhibitors trigger NOXA-mediated apoptosis in melanoma and myeloma cells.
Cancer Res. 2005, 65, 6282-6293.
7. Hussain, A. R.; Ahmed, M.; Ahmed, S. O.; Al-Thari, S.; Khan, A. S.; Razack, S.; Platanias,
L. C.; Al-Kuraya, K. S.; Uddin, S. Proteasome inhibitor MG-132 mediated expression of
p27Kip1 via S-phase kinase protein 2 degradation induces cell cycle coupled apoptosis in
primary effusion lymphoma cells. Leukemia Lymphoma 2009, 50, 1204-1213.
8. Borissenko, L.; Groll, M. 20S proteasome and its inhibitors: Crystallographic knowledge for
drug development. Chem. Rev. 2007, 107, 687-717.
9. Chen, D.; Frezza, M.; Schmitt, S.; Kanwar, J.; Dou, Q. P. Bortezomib as the First
Proteasome Inhibitor Anticancer Drug: Current Status and Future Perspectives. Curr. Cancer
Drug Tar. 2011, 11, 239-253.
10. Richardson, P. G.; Briemberg, H.; Jagannath, S.; Wen, P. Y.; Barlogie, B.; Berenson, J.;
Singhal, S.; Siegel, D. S.; Irwin, D.; Schuster, M.; Srkalovic, G.; Alexanian, R.; Rajkumar, S.
V.; Limentani, S.; Alsina, M.; Orlowski, R. Z.; Najarian, K.; Esseltine, D.; Anderson, K. C.;
Amato, A. A. Frequency, characteristics, and reversibility of peripheral neuropathy during
treatment of advanced multiple myeloma with bortezomib. J. Clin. Oncol. 2006, 24, 3113-
3120.
11. Verbrugge, S. E.; Assaraf, Y. G.; Dijkmans, B. A. C.; Scheffer, G. L.; Al, M.; den Uyl, D.;
Oerlemans, R.; Chan, E. T.; Kirk, C. J.; Peters, G. J.; Van der Heijden, J. W.; de Gruijl, T. D.;
Scheper, R. J.; Jansen, G. Inactivating PSMB5 Mutations and P-Glycoprotein (Multidrug
Resistance-Associated Protein/ATP-Binding Cassette B1) Mediate Resistance to Proteasome
Inhibitors: Ex Vivo Efficacy of (Immuno)Proteasome Inhibitors in Mononuclear Blood Cells
28