Phenoxypropanolamine derivatives as selective inhibitors of the 20S proteasome β1 and β5 subunits

Article abstract of DOI:10.1016/j.bmcl.2017.10.055

New series of thiophene-containing phenoxypropanolamines were synthesized and evaluated for their potency to inhibit the three proteolytic activities of the mammalian 20S proteasome. Noticeable inhibition of both ChT-L and PA activities was obtained with

Full text of DOI:10.1016/j.bmcl.2017.10.055

Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Phenoxypropanolamine derivatives as selective inhibitors of the 20S
proteasome b1 and b5 subunits
Anna A. Hovhannisyan b,e, The Hien Pham a,e, Dominique Bouvier d,e, Xiao Tan a, SiAmmar Touhar a,
Gevorg G. Mkryan c, Ashot M. Dallakyan c, Chahrazade El Amri a, Gagik S. Melikyan b,
Michèle Reboud-Ravaux a, Michelle Bouvier-Durand a,
⇑
a Sorbonne Universités, UPMC Univ Paris 06, IBPS, CNRS UMR 8256, Biological Adaptation and Ageing (B2A), Integrated Cellular Ageing and Inflammation, 7 Quai St Bernard,
75005 Paris, France
b Department of Organic Chemistry, Yerevan State University, 1 Alex Manoogian Str., 0025 Yerevan, Armenia
c The Scientific Technological Centre of Organic and Pharmaceutical Chemistry HAS A.L.Mnjoyan Institute of Fine Organic Chemistry, 26 Azatutyan Str., Yerevan 0014, Armenia
d Sorbonne Universités, UPMC Univ Paris 06, Atelier de BioInformatique, ISYEB, UMR 7205 CNRS MNHN UPMC EPHE, Museum national d’Histoire naturelle, CP 50, 45 rue Buffon,
75005 Paris Cedex 05, France
a r t i c l e i n f o
a b s t r a c t
Article history:
New series of thiophene-containing phenoxypropanolamines were synthesized and evaluated for their
potency to inhibit the three proteolytic activities of the mammalian 20S proteasome. Noticeable inhibi-
tion of both ChT-L and PA activities was obtained with three compounds: one with unsubstituted phe-
noxypropanolamine group (7) and the two others with a p-Cl-substituted group (4 and 9). For three
other compounds (3, 8 and 10), ChT-L activity alone was significantly inhibited. In silico docking per-
formed on the b5 and b1 subunits bearing the respective ChT-L and PA catalytic sites showed features
common to poses associated with active compounds. These features may constitute a selectivity criterion
for structure-guided inhibitor design.
Received 13 July 2017
Revised 20 October 2017
Accepted 22 October 2017
Available online 23 October 2017
Keywords:
Phenoxypropanolamine
Thiophene
Ó 2017 Elsevier Ltd. All rights reserved.
20S proteasome
ChT-L, T-L and PA activities
Selective inhibition
In silico structure-based docking
Inhibitors of the constitutive 20S proteasome core particle
(CP)1–8 have been validated as anticancer drugs, since the develop-
ment of the peptide boronate bortezomib for the treatment of mul-
tiple myeloma (MM)9 and mantle cell lymphoma (MCL).10 The
peptide epoxyketone carfilzomib11 and ixazomib,12 a bortezomib
derivative, have been recently FDA-approved to overcome MM.
New peptidomimetic inhibitors derived from bortezomib13 and
carfilzomib (reviewed in14) are now under development. Smaller
organic CP inhibitors have also emerged as promising therapeutic
tools for the treatment of cancer diseases (reviewed in6–8,14,15).
Among these inhibitors are the salinosporamide A (NPI0052) b-
lactone,16,17 which is currently under clinical trials, several furopy-
ridine derivatives of cerpegin,18–20 a hydroxyurea derivative,21
tamoxifen-related ridaifens (RID-A and RID-F)22 and a 1,2,4-oxadi-
azole23 (Fig. 1). In our search for new CP organic inhibitors, we
retained the thiophene moiety due to the potency of natural thio-
phene derivatives, such as the fungal junipal,24 to display high
anti-proliferative activity25 We hypothesized that thiophene
derivatives could target CP activity based on rare studies reporting
CP inhibition by thiophene-containing molecules: the 1,3,4-oxathi-
azole-2-ones GL5 and HT117126 and the hydronaphtoquinone PI-
818227 (Fig. 1). The phenoxypropanolamine moiety was also con-
sidered for our design of these new small inhibitors, on the basis
of the efficiency of this fragment in CP inhibition when bound to
the furopyridine scaffold of cerpegin (Fig. 1 Refs. 18,20).
Three types of proteolytic activities are associated to the 20S CP
and located on b1, b2 and b5 subunits in each of the two inner CP
b-rings.28 Each of these catalytic sites is characterized by N-termi-
nal threonine 1 (T1) which triggers nucleophilic attack on the sub-
strates. They differ by their specific cleavage potentiality of the
peptidic substrates and accordingly are classified as caspase-like
(or PA for post-acid), trypsin-like (T-L) and chymotrypsin-like
(ChT-L) enzymes (associated with the respective b1, b2 and b5 sub-
units). They can cleave synthetic peptidic substrates after acidic,
basic or hydrophobic residues, respectively.29
In the present work, fifteen thiophene-containing phe-
noxypropanolamines were synthezised and assayed in vitro for
their inhibitory effect on the three types of proteolytic activities.
⇑
Corresponding author.
E-mail address: mbouvier@snv.jussieu.fr (M. Bouvier-Durand).
These authors contributed equally to this work.
e
https://doi.org/10.1016/j.bmcl.2017.10.055
0960-894X/Ó 2017 Elsevier Ltd. All rights reserved.

A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
5173
Fig. 1. Structures of small organic inhibitors of the 20S proteasome. Top and median parts illustrate a panel of some thiophene-free important molecules (salinosporamide
A,16 N5-furopyridine derivatives of cerpegin18 and C4-furopyridine derivatives,20 a hydroxyurea derivative,21 tamoxifen related ridaifens (RIDs)22 and an 1,2,4-oxadiazole
derivative23). Bottom part illustrates the rare cases of thiophene-containing molecules: the oxathiazole-2-ones GL5 and HT117126 and the hydronaphthoquinone PI-8182.27
Their specificity for the CP was also investigated using two other
cellular proteases, calpain I and cathepsin B. Preliminary investiga-
tions on the inhibitory effects on cell division were also performed
since key proteins involved in the cell cycle machinery are among
the CP cellular substrates.3,8 When potent inhibitors were detected,
a docking approach was conducted in order to get insights on the
possible binding sites of these new small molecules.
The general structure of the different thiophene-containing
phenoxypropanolamine series designed here (Fig. 2A–C) is com-
prised of two domains: a molecular head with a cyclopentyl, a
cyclohexyl or a tetrahydropyrane ring substituted at the C2 posi-
tion of the thiophene ring and a phenoxypropanolamine tail, which
is in turn either unsubstituted (R0 = H for 2, 7 and 12) or diversely
substituted (R0 = p-CH3 for 3, 8 and 13; R0 = p-Cl for 4, 9 and 14; R0
= o-OCH3 for 5, 10 and 15).
taining phenoxypropanolamine series (1–5 in series I, 6–10 in ser-
ies II and 11–15 in series III) a detailed synthesis protocol, yields
and melting points as well as 1H NMR and IR data are reported
in the Supplementary section. 13C NMR and GCMS data are
reported for compounds 4 and 9.
Bioassays were first conducted in vitro with a CP fraction iso-
lated from rabbit erythrocytes and performed according to.18 In
short, measurements of the ChT-L, T-L and PA activities were made
using the respective fluorogenic substrates Suc-LLVY-AMC, Boc-
LRR-AMC and Z-LLE-bNA (Bachem France). Fluorescence was mea-
sured using black 96-well microplates and a BMG Fluostar micro-
plate reader. The inhibitory activity of compounds was expressed
as IC50 (inhibitor concentrations giving 50% inhibition). IC50 values
are summarized in Table 1. In the absence of the phe-
noxypropanolamine tail (1, 6, 11), the thiophene head alone did
not give rise to any inhibition, whatever the structure of the head
For the synthesis of new thiophene series substituted at C2 posi-
tion, 2-thiopheneacetonitrile was chosen as initial compound. It
was treated with dihalides: 1,4-dibromobutane, 1,5-dibromopen-
tane or bis (2-chloroethyl) ether and transformed into the corre-
sponding cyclopentane, cyclohexane and tetrahydropyran
derivatives I(a–c) by a method analogous to the one developed
by Tsinker et al.30 (Scheme 1, steps i, i0 and i00, respectively). Nitrile
group reduction of the obtained compounds with LiAlH4 led to
amines 1, 6 and 11 (Scheme 1, steps ii).
Reaction of these primary amines with makeup substituted 2-
phenoxymethyloxiranes31 was realized by opening the epoxy cycle
in the presence of catalytic amounts of water, resulting in N-substi-
tuted phenoxypropanolamines 2–5 (series I), 7–10 (series II) and
12–15 (series III) (Scheme 1, steps iii). For the purpose of biological
studies, the final compounds were transferred to water-soluble
oxalates or hydrochlorides. For each of these new thiophene-con-
and the targeted catalytic site. The presence of
a phe-
noxypropanolamine tail was beneficial to get inhibition of ChT-L
and PA sites for several compounds of the cyclopentyl and cyclo-
hexyl series (I and II, Table 1). The best IC50 values of this study
(ranging from 12 to 18 mM) were obtained with compounds 3, 4
and 7 to 10 for ChT-L activity and 4, 7 and 9 for PA activity. Thus,
homologous compounds 4 and 9 (from series I and II, both with a
p-chlorophenoxypropanolamine group) and the unsubstituted
compound 7 (series II) had quite similar efficiencies on both ChT-
L and PA activities. With one exception (14), tetrahydropyrane-
containing compounds (Series III, Table 1, compounds 12, 13, 15)
failed to significantly inhibit any type of proteasomal activity.
Compound 14, with a p-chlorophenoxypropanolamine group, gave
rise to some inhibition of the ChT-L activity alone. For this latter
activity, it should be noted that, in the cyclohexyl series (II), the

5174
A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
Fig. 2. General structures of thiophene-containing phenoxypropanolamine ligands (A–C). series I (A, 2–5), series II (B, 7–10) and series III (C, 12–15) contain in the thiophene
head a cyclopentyl, a cyclohexyl or a tetrahydropyran ring respectively. The asterisk indicates the two possible conformers R and S.
R0 substituent did not play a major role. Indeed, the unsubstituted
compound 7 was as efficient as compounds 8, 9 and 10 with the
either proper proteasome inhibition or a general cellular stress
leading to death. In this direction, more efficient compounds
should be designed in order to rule out unspecific cytotoxic effects.
In the same way for rat BM, the minimal tolerated doses (MTD)
should be determined in order to rule out treatment innocuity in
animals. While preliminary, these in vivo results open the way to
highly potent proteasome inhibitors.
respective
p-methyl-,
p-chloro-
and
o-methoxy-phe-
noxypropanolamine groups. In spite of their relatively low efficien-
cies, the new thiophene-containing phenoxypropanolamines, 4, 9
and 7, present the advantage of inhibiting both ChT-L and PA.
Specificity towards CP was examined using fractions of two
other concurrent cellular proteases: calpain I and cathepsin B.18
Calpain I and cathepsin B were purified fractions from human ery-
throcytes and human hepatocytes respectively (Calbiochem, VWR
International S.A.S, France). Suc-LLVY-AMC and Z-RR-AMC were
used as their respective fluorogenic substrates. Compounds 4 and
2 were tested on these proteases and no inhibition was observed,
indicating a selectivity toward proteasome.
Preliminary in vivo bioassays were also conducted with com-
pound 4, one of the most efficient inhibitor in vitro, in order to test
possible effects on cell proliferation. A 24-h treatment with this
compound was applied to 5 rats per treatment (0.2 and 2 mg per
rat) and bone marrow (BM) cells from femora were collected as
indicated in.32,33 Cell division activity was measured by counting
metaphase plates prepared according to.34 At doses as low as 0.2
and 2 mg per rat, compound 4 induced 50% and 75% cell division
inhibition, respectively. In the same way, HeLa cell proliferation,
measured according to35,36 was fully inhibited following a 50-mM
treatment with compound 4 for 72 h and apoptotic figures were
systematically observed (underlined by arrows pointing to frag-
mented nuclei in supplementary Fig. S1). A lower concentration
(10 mM) did not cause cell division arrest nor apoptosis. XTT assays,
performed according to37, confirmed these results since mitochon-
drial activity was totally inhibited at 50 mM while no inhibition
was detected at 10-mM concentration. Thus, high doses of com-
pound 4 were shown to be powerful in vivo, leading to cell division
arrest in both cultured human HeLa cells and rat BM. This anti-pro-
liferative effect is dose-dependent and accompanied in HeLa cells
by an apoptotic program at high concentrations only. In the future,
work is needed to correlate this accompagnying cytotoxicity with
In silico docking calculations were performed to study the puta-
tive binding of compounds of series I and II to the ChT-L and PA
active sites. The catalytic chains from the bovine 20S proteasome
were used as targets. As previously outlined and discussed20 mam-
malian proteasomes show strong amino acid sequence identity
(>95%) in their catalytic b-type subunits, and the few non-identical
residues are external to the active sites. The bovine crystal struc-
ture (PDB ID 1IRU38) was obtained from the Protein Data Bank.39
The AutoDock Vina program40 was used for docking calculations
with default parameters. The docking space was restricted to a
20 ꢀ 20 ꢀ 20 Å box centered on the side chain oxygen of N-termi-
nal threonine of b5 (ChT-L) and b1 (PA) subunits. Visual Molecular
Dynamics (VMD)41 was used to analyse results and to prepare
molecular graphics. Computed Kd values were obtained from bind-
ing energy data according to the equation
D
D
G0 = ꢁRT ln Kd, where
G0 is the binding energy given by the Vina fitness function, R the
gas constant and T the temperature (310 K).
In compounds of the diverse series with a phenoxypropanol tail,
an asymmetric carbon (secondary alcohol) gives rise to two config-
urations R and S, which were both tested in all cases.
For compounds 1, 6 and 11, comprised of a thiophene head only,
no complexes could be obtained, which confirms the in vitro data
on the isolated CP (Table 1).
Due to their significant effect on in vitro ChT-L activity, com-
pounds 3 and 4 of series I, with a p-methyl- and a p-chloro-phe-
noxypropanolamine tail, respectively, were first tested in the b5
subunit (Fig. 3). Fig. 3A gives a general view of the ChT-L binding
site contributed by both b5 and b6 subunits. Details are given in
the figure legend.

A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
5175
Scheme 1. Synthesis of thiophene-containing phenoxypropanolamines. Reagents and conditions: (i) 2-thiopheneacetonitrile, NaOH, dihalide [(i)1,4-dibromobutan; (i0) 1,5-
dibromopentan; (i00) 2-chloroethyl ether], 50 °C, 50–60% yield; (ii) Ia-c, benzene, LiAlH4 in ether, reflux 5 h, 81% yield; (iii) 1 (or 6 or 11), different 2-phenoxymethyloxiranes,
R0-substituted on the benzene ring, 2–3 drops of water, isopropyl alcohol, reflux 20 h, 50–89% yield.
As shown in Fig. 3B (3) and C (4), the first rank poses (those with
the lowest Kd value, Kd = 13 M) always showed the ligands bound
the lower efficiencies observed in vitro for 2 and 5 (20 and 38
mM, respectively).
l
inside the active site of the enzyme. In the most frequently
observed conformation, the thiophene head was found in the S3
subsite. For the sake of simplicity, this orientation was referred
to as ‘‘b6-S3-head”, where the p-substituted tails lied in the groove
between threonine residues T1 and T21 (see dotted double arrow
in Fig. 3A). The high frequency of occurrence of this orientation
(30% of all first rank hits for compound 4) could be related to the
best in vitro IC50 values (Table 1).
Whatever the compound (3 and 4), no more than one putative
H-bond could be predicted between the ligand and the enzyme,
on the basis of a distance between donnor and acceptor heteroa-
toms shorter than 3.0 Å. This H-bond should involve the secondary
alcohol of the ligand’s tail and heteroatoms in residue T21. The
ligand amine group was seen close to the acidic group of D125
(Fig. 3B, C), which opens the possibility of Coulombic interactions.
In all cases, stereochemistry induced only very small variations in
computed Kd values. Several different poses with higher energy
were observed for less efficient compounds (e.g. 2 and 5) which
are illustrated in supplementary Fig. S2. This is concordant with
Docking results in the PA active site are shown in Fig. 4 and
supplementary Fig. S3. Fig. 4A gives a general view of the PA
binding site contributed by both b1 and b2 subunits and details
are presented in the figure legend. The first rank poses showed
the ligands inside the active site of the enzyme, in two possible
conformations, thereafter referred to as ‘‘S10-head” (Fig. 4B, illus-
trated with compound 3) and ‘‘b2-S3-head” (Fig. 4C, illustrated
with compound 4).
In the S10-head conformation (Fig. 4B, 3), the thiophene head
group occurred at the entrance of the S10 subsite channel42 extend-
ing from residue T1 in b1 up to residue Y30 in the neighbouring b7
subunit. The ligand’s tail was directed towards the S3 subsite
where it could form a characteristic T-shaped p-p interaction with
Y114 in subunit b2 (Fig. 4B). A similar interaction with Y114 could
be observed with most types of tails. Only did the p-chloro substi-
tution in compound 4 (series I) and compound 9 (series II) cause
the ligand to adopt the b2-S3-head conformation (Fig. 4C), in
which the thiophene ring of the head interacted with Y114 at S3
and the tail entered the S10 channel. Computation of Kd values gave

5176
A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
Table 1
Inhibition of the CT-L, T-L or PA activities of the 20S proteasome by thiophene-containing phenoxypropanolamines 1–15 (series I, II, III). Inhibitions are expressed as IC50 values
(mM) or, in case of low inhibitions, as (%) inhibition values at 100 M (italics). ni = no inhibition.
l
Compound
Series I, II III
IC50 (mM) or (%) inhibition at 100 mM
ChT-L
T-L
PA
Series I
R
1
2
3
4
5
H
(26%)
ni
ni
CH2CHOHCH2OC6H5
19.3 0.7
15.4 0.4
15.8 0.7
37.2 0.9
(41%)
(38%)
(42%)
(48%)
66.5 3.5
24.2 1.1
18 0.8
36 0.4
CH2CHOHCH2OC6H4-p-CH3
CH2CHOHCH2OC6H4-p-Cl
CH2CHOHCH2OC6H4-o-OCH3
Series II
R
6
H
ni
ni
ni
7
8
9
10
CH2CHOHCH2OC6H5
CH2CHOHCH2OC6H4-p-CH3
CH2CHOHCH2OC6H4-p-Cl
CH2CHOHCH2OC6H4-o-OCH3
14.9 0.2
15.5 0.6
18 0.8
17.1 0.1
ni
(36%)
ni
18.1 0.7
30
2
12.8 0.8
22 1.2
(40%)
Series III
R
11
12
13
14
15
H
ni
ni
ni
(36%)
ni
ni
ni
CH2CHOHCH2OC6H5
CH2CHOHCH2OC6H4-p-CH3
CH2CHOHCH2OC6H4-p-Cl
CH2CHOHCH2OC6H4-o-OCH3
(31%)
80 0.2
40 19
ni
(39%)
(49%)
(53%)
ni
Fig. 3. In silico docking of series I compounds 3 and 4 in the ChT-L active site of the bovine 20S proteasome (b5 subunit). General view of the active site (A) without any ligand
(the rectangular frame marks limits of the field shown in B and C). Subunit b5 bears the catalytic residue T1 and several other residues including T21 and G47 lining a deep
groove where previous studies have localized the binding of furopyridine derivatives (double dotted arrow).18–20 The binding site is comprised of several subsites (S1, S10, S3
. . ..) for different parts of the polypeptide chain substrate. Subsite S1, the entry of which is labeled with a white arrow is hidden behind G47 and correspond to the binding site
of the amino side chain upstream the scissile peptide bond (specificity pocket). Subsite S10 is a secondary binding site for the substrate peptide chain. Subsite S3, including
residue D125, belongs to b6 subunit. Compounds 3 (B) and 4 (C) were in b6-S3-head orientation. In this most frequently observed orientation, the p-substituted tails lied in
the groove formed between T1 and T21. Protein chains are shown as solvent accessible surfaces (light-grey for b5 and dark-yellow for b6). The residues and atoms relevant to
the enzyme activity are labeled and colored according to CPK convention. Compounds are displayed as stick models colored by atom type.
a net advantage to the b2-S3-head orientation taken by compound
from distances between donnor and acceptor heteroatoms shorter
than 3.0 Å. These two H-bonds could involve both the secondary
alcohol and the amine of the ligand’s tail and heteroatoms in resi-
due T21 and/or G47. The stereochemistry of the compounds had no
significant effect.
4 (Kd = 11
3 (Kd = 40
l
l
M) compared to the S10-head orientation of compound
M). The former orientation corresponded to the best
in vitro IC50 values for PA (Table 1, compounds 4 and 9). An average
number of two putative ligand-enzyme H-bonds was deduced

A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
5177
Fig. 4. In silico docking of series I compounds 3 and 4 in the PA active site of the bovine 20S proteasome. General view of the active site (A) without any ligand (the rectangular
frame delineates the field shown in B and C). The structure of b1 is very close to that of b5, with the catalytic residue T1 and a groove lined by residues T21 and G47 (double
dotted arrow). Subsite S3 belongs to the b2 subunit and includes residue Y114 supposed to play important parts in the binding of furopyridines derivatives.18–20 Compound
3 (B) was in S10-head orientation and the ligand tail was directed towards Y114 in the S3 subsite belonging to b2 subunit. Compound 4 (C) was in b2-S3-head orientation and
the p-Cl phenoxypropanolamine tail was seen to enter the S10 channel. It is noticeable that the homologous compound 9 in series II also adopted this b2-S3-head orientation
as did compound 4 (not shown). Protein chains are shown as solvent accessible surfaces (b1 subunit in light grey and neighbouring b2 subunit in green). The residues and
atoms relevant to the enzyme activity are labeled and colored according to CPK convention. Compounds are displayed as stick model colored by atom type.
To assess the effects of the thiophene-associated saturated ring
in the ligand’s head, compounds 2 and 7 of series I and II, with the
same unsubstituted phenoxypropanolamine tail, were docked.
Briefly, the poses obtained with homologs bearing a pentane (2)
or a hexane (7) ring were almost superimposable (illustrated for
PA in supplementary Fig. S3A, B). Computed Kd values were very
close and did not allow us to discriminate between the two struc-
tures although in vitro results gave a slight advantage to the hexyl
structure of series II. Ligands bearing a tetrahydropyrane ring were
not tested in silico due to their absence of inhibitory effect in vitro.
We have never observed direct contact between the ligands and
subsite S1, which is the canonical site of specificity for amino acid
side chains of peptidic substrates, (reviewed in3). It is now
accepted that other subsites, including S10, could also serve as
binding sites for non peptidic, small organic molecules.24,19,43 The
present study confirms and extends previous observations that
the S3 subsite may also be involved in the binding of small
ligands.18–20
In conclusion, some of the thiophene-containing phe-
noxypropanolamine derivatives studied here were moderately effi-
cient and specific CP inhibitors. Three compounds (4, 7 and 9) were
inhibitors of both ChT-L and PA activities when considering IC50
values between 12 and 18 mM. In the same range of values, three
molecules (3, 8 and 10) worked with the same efficiency on ChT-
L only but became mild inhibitors of PA activity above this range.
The phenoxypropanolamine moiety was already used in C4-furopy-
ridine-3-one derivatives (Fig. 1) which yielded mild inhibition of
PA activity (IC50 = 48 mM).18 Acting on both ChT-L and PA activities,
thiophene derivatives 4, 7 and 9 are therefore less selective but
more efficient inhibitors. The thiophene moiety was present in
oxathiazole derivatives GL5 and HT1171 (Fig. 1) studied on
Mycobacterium tuberculosis proteasome26 showing an IC50 in the
range of 10 mM in cell lysates. The authors showed that the oxathi-
azole moiety of HT1171 combines irreversibly with the N-terminal
T1 and crystallographic studies have focused on this reaction and
did not provide any information on thiophene interaction in the
active site.
The thiophene-containing hydronaphtoquinone derivative
PI8182 (Fig. 1) was described as a ChT-L inhibitor (IC50 = 3 mM)
but it was further demonstrated that a thioglycolic acid side chain
on the hydronaphtoquinone double ring was critical to the inhibi-
tory activity on ChT-L27 and thiophene replacement by hydropho-
bic aryl moieties yielded much more active compounds.
In the light of these reports and of the results reported here, the
phenoxypropanolamine tail seems to play more determining parts
in proteasome inhibition than the thiophene head. Indeed, the tail
bears chemical groups able to associate the ligand to the active site,
which is not the case of the thiophene head.
Docking experiments revealed common features of those poses
associated with active compounds. The most striking feature is the
placement of the phenoxypropanolamine tail in the groove of the
active site where it is predicted to form H-bonds with residues lin-
ing this groove. In addition, most of the compounds efficient in
ChT-L and PA inhibition adopted binding modes with the thio-
phene head close to the S3 subsite (S3 head). In b1, the most fre-
quent conformation (S10 head) characterized milder inhibitors (3,
5, 8 and 10). It is the p-Cl substituent in compounds 4 and 9 that
made them adopt the more active b2-S3 head conformation, prob-
ably for steric and/or electronic reasons.
Acknowledgements
The work was supported by the Yerevan State University (A. A.
H., G. S. M.), by the Scientific Technological Center of Organic and
Pharmaceutical Chemistry in Yerevan (G. G. M., A. M. D.), by the
Vietnamian Ministery for Education (T. H. P.), by the Pierre et Marie
Curie University, UPMC, Paris 6, Integrated cellular Ageing and
Inflammation, 7 Quai St Bernard 75005, Paris (T. H. P., X. T., S. A.
T., C. E. A., M. R. R., M. B. D.) and Atelier de Bioinformatique, ISYEB,

5178
A.A. Hovhannisyan et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 5172–5178
21. Gallastegui N, Beck P, Arciniega M, Huber R, Hillebrand S, Groll M. Angew Chem
Int Ed. 2012;51:247.
22. Hasegawa M, Yasuda Y, Tanaka M, et al. Eur J Med Chem. 2014;71:290.
23. Maréchal X, Genin E, Qin L, et al. Curr Med Chem. 2013;20:2351.
24. Birkinshaw JH, Chaplen P. Biochem J. 1955;60:255.
UMR 7205, CNRS MNHN UPMC EPHE, CP 50, 45 Rue Buffon, 75005,
Paris (DB).
Supplementary data
25. Crew AP, US7 Patent, N 101, 893, B2, Sept. 5. 2006.
26. Lin G, Li D, de Carvalho LPS, et al. Nature. 2009;461:621.
27. Ge Y, Kazi A, Marsilio F, et al. J Med Chem. 1978;2012:55.
28. Groll M, Ditzel L, Löwe J, et al. Nature. 1997;386463.
29. Groll M, Heinemeyer W, Jäger S, et al. Proc Natl Acad Sci USA. 1999;96:10976.
30. Tsinker MG, Mnjoyan AL, Akopian EN. Arm Chim J. 1969;22:314.
31. Petrov V, Stephenson OJ. Pharm Pharmocol. 1953;6:359.
32. Antiproliferative assays in rat bone marrow cells. Wistar adult rats (average
Supplementary data are associated to the main text in a
separate file including: data for chemistry introducing the new
synthesized compounds with elemental analysis and NMR data. 3
figures (S1: for proliferation and apoptosis assays, and S2, S3: for
docking data) and their respective legends. Supplementary data
associated with this article can be found, in the online version, at
https://doi.org/10.1016/j.bmcl.2017.10.055.
weight 180 g) received compound
4 via intraperitoneal injections at the
following doses: 0.2 and 2 mg per rat (5 rats were injected per treatment and a
control group did not receive injections). Bone marrow cells were collected
from femora 24 h after the intraperitoneal injection according to33. Metaphasic
plates were prepared, allowing mitotic events to be counted according to34
33. Aghekyan GG, Mkryan TG, Gukasyan AS, et al. Chim J Arm. 2010;3:398.
34. Ford EHR, Woollam DHM. Exp Cell Res. 1963;32:320.
.
References
1. Kisselev AF, Goldberg AL. Chem Biol. 2001;8:739.
2. Adams J, Kauffman M. Cancer Invest. 2004;22:304.
3. Borissenko L, Groll M. Chem Rev. 2007;107:687.
4. Kim KB, Crews CM. J Med Chem. 2008;51:2600.
35. Cytotoxicity and antiproliferative effects upon HeLa cells. Cytotoxicity was tested
on HeLa cells, a human cervix carcinoma cell line (kindly provided by Dr Suzie
Lefèbvre from Cell Biology Department of the Jacques Monod Institute,
University Paris-Diderot, France). Cells were cultured in microplates as
already described36 in 100 mL DMEM medium supplemented with 10% fetal
calf serum and penicillin and streptomycin (100 units/mL and 100 mg/mL
respectively). A treatment with compound 4, was applied to the cells (5ꢂ103
cells/well) at two concentrations (10 and 50 mM) for 18, 48 and 72 h.
5. Genin E, Reboud-Ravaux M, Vidal J. Cur Top Med Chem. 2010;10:232.
6. Kisselev AF, van der Linden WA, Overkleeft HS. Chem Biol. 2012;19:99.
7. Gräwert MA, Groll M. Chem Comm (Camb). 2012;48:1364.
8. Rentsch A, Landsberg D, Brodmann T, Bülow L, Girbig A-K, Kalesse M. Angew
Chem Int Ed Engl. 2013;52:5450.
9. Kane RC, Farrell AT, Sridhara R, Pazdur R. Clin Cancer Res. 2006;12:2955.
10. Kane RC, Dagher R, Farrell A, et al. Clin Cancer Res. 2007;13:5291.
11. Lawasut P, Chauhan D, Laubach J, et al. Curr Hematol Malig Rep. 2012;7:258.
12. Shirley M. Drugs. 1935;2015:75.
13. Scarbaci K, Troiano V, Ettari R, et al. ChemMedChem. 1801;2014:9.
14. Teicher BA, Tomaszewski JE. Biochem Pharmacol. 2015;96:1.
15. Beck P, Dubiella C, Groll M. Biol Chem. 2012;393:1101.
16. Macherla VR, Mitchell SS, Manam RR, et al. J Med Chem. 2005;48:3684.
17. Chauhan D, Catley L, Li G, et al. Cancer Cell. 2005;8:407.
18. Pham TH, Hovhannisyan A, Bouvier D, et al. Bioorg Med Chem Lett.
2012;22:3822.
Antiproliferative effects of compound
4 were directly observed in the
microplates under inverted microscope and micrographs were acquired for
countings. Cytotoxicity, in terms of decreased mitochondrial activities, was
analysed using XTT assays according to the protocol already described in37
36. Koroleva ON, Pham TH, Bouvier D, et al. Biochimie. 2015;108:94.
37. Basse N, Montes M, Maréchal X, et al. J Med Chem. 2010;53:509.
.
38. Unno M, Mizushima T, Morimoto Y, et al. Structure (London, England: 1993).
2002;10:609.
39. Berman HM, Westbrook J, Feng Z, et al. Nucleic Acids Res. 2000;28:235.
40. Trott O, Olson AJ. J Comput Chem. 2010;31:455.
41. Humphrey W, Dalke A, Schulten KJ. Mol Graphics. 1996;14:33.
42. Groll M, Larionov OV, Huber R, de Meijere A. Proc Natl Acad Sci USA.
2006;103:4576.
19. Hovhannisyan A, Pham TH, Bouvier D, et al. Bioorg Med Chem Lett.
2013;23:2696.
20. Hovhannisyan A, Pham TH, Bouvier D, et al. Bioorg Med Chem Lett.
2014;24:1571.
43. Beck P, Reboud-Ravaux M, Groll M. Angew Chem Int Ed. 2015;54:11275.