1572
A. Hovhannisyan et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1571–1580
The CP particles of c20S and i20S form a similar cylindrical
structure, each comprising four stacked heptameric rings ( 1–7
b1–7 b1–7 1–7). The different proteolytic activities of c20S or
investigated by in silico docking. Previous studies from our group
have suggested a covalent mechanism of inhibition associated with
a
a
a unique mode of binding of cerpegin-derived c-lactones to the bo-
i20S are confined to the two identical inner b rings. In a single b
ring, only the b1, b2 and b5 subunits possess a proteolytic activity,
the so-called caspase-like (C-L or PA), trypsin-like (T-L) and chy-
motrypsin-like (CT-L) activity, respectively.3–5,13–17 Whatever the
nature of the catalytic particle (c20S or i20S), each type of proteo-
lytic site possesses a threonine residue at position 1 capable of
nucleophilic attack on the substrates. However, specificity of the
respective sites (b1, b2 and b5) can be distinguished in vitro by
placing acidic, basic or hydrophobic residues at the C-terminal
end of fluorogenic substrates. Despite small variations in the amino
acid sequences of the respective catalytic subunits of i20S, b1i
(LMP2), b2i (MECL-1) and b5i (LMP7), the same substrates can be
used as those for c20S to evaluate the respective activities. b5i sub-
units have been recently reported as key regulators of cytokine
production with selective inhibition by the epoxyketone PR-957
and therapeutic applications for experimental arthritis.18 A recent
X-ray study7 of this compound bound to the mouse i20S or c20S
crystal structures showed the absence of conformational changes
in the S1 specificity pocket of b5i site, explaining the selectivity
of PR-957 for b5i. Site-specific proteasome inhibitors are not con-
fined to b5 sites since b1 ones have been shown to be the specific
targets of IPSI-00119 and LMP2-ek19 in the case of i20S. Both b1c
and b1i sites are inhibited by peptido-epoxyketones YU-102,20
NC-00121 and LU-102.22 As regards b2 sites, NC-022 is a specific
inhibitor of T-L activity.23
vine b1c active site.1,2
The general structures of the molecules synthesized and evalu-
ated in this study are presented in Figure 1(B–E) whereas reference
compounds are presented in Figure 1A. The b-lactone of SalA
(Fig. 1A, top) is attacked by the nucleophilic Thr1 and this leads
to formation of a highly stable acyl-enzyme.24 The
c-lactone ring
of cerpegin (Fig. 1A, bottom) is also susceptible to nucleophilic at-
tack by Thr1 and we have recently reported kinetic results showing
covalent proteasome inhibition by a cerpegin derivative.2 Further-
more, cerpegin derivatives substituted at N5 or at N5 and C1
showed selectivity towards the post-acid activity (IC50 2–
10
been designed to retain the
l
M).1,2 Based on these results, new series of molecules have
c
-lactone of the fused furopyridine ring
of cerpegin. C4- and C1-substituted furo[3,4-c]pyridine-3-one
derivatives with main variations of the functional groups at C4 po-
sition (labeled R in Fig. 1B) were synthesized. These R groups were
mainly chosen on the basis of our previous results where they
appeared as efficient N5 substituents.1,2
Some other
c-lactones not fused to a pyridine ring were also
designed and prepared (Fig. 1C, 19–22), together with two pyri-
dine-fused d-lactones (Fig. 1C, 23 and 24) as model compounds
to evaluate the effects of the lactone environment.
The role of carbonyl-derived chemical groups other than
c-lactones, placed in various heterocyclic environments (fused
furo- and thienopyrimidine-4-ones), was also investigated
Cell cycle progression and apoptosis induction are the cellular
functions best known for their susceptibility to proteasome inhibi-
tion. The dipeptide boronate bortezomib (or VelcadeÒ, initially
PS341) was the only one that came to therapeutic step among
the first generation of inhibitors. Successful treatments with this
molecule were reported in 2003 on multiple myeloma and then
in 2006 on relapsed mantle cell lymphoma. Because of drastic side
effects and the appearance of resistant cells, a second generation of
inhibitors has been developed. Two other peptide-boronates, CEP-
18770 and MLN9708, the tetrapeptide epoxyketone carfilzomib
and the b-lactone salinosporamide A, have been the subject of
intensive clinical studies.24 Carfilzomib (Kyprolis or PR-171) has
been recently approved by FDA for the treatment of multiple
myeloma.25 Salinosporamide A (SalA, NPI 0052 or marizomib) be-
longs to a large family of natural compounds (omuralide, b-lactosin
A, reviewed in Ref. 6) and is a highly potent inhibitor of c20S with a
b-lactone working as a pharmacophore. SalA has been retained in
this study as a reference b-lactone24 as well as cerpegin derivatives
(Fig. 1D and E).
The aforementioned C4- and C1-substituted furo[3,4-c]pyridine-
3-one derivatives (Fig. 1B) were synthesized as outlined in
Scheme 1A and B. Scheme 1A sums up steps for the synthesis of
compounds of series IV and VIII (1–4), III and VI (7–18). Com-
pounds XI (5, 6) were synthesized according to Scheme 1B. The un-
fused
c-lactone models 19, 20, 21, 22 were synthesized as
previously reported (19, 20,35,36 21,2 2237). The pyridinone-fused
d-lactones 23 and 24 were prepared as mentioned in38 for 23,
and according to Scheme 1C for 24.
The most convenient way of synthesis of nor-cerpegin (1,1-di-
methyl furo[3,4-c]pyridine-3,4-dione) (compound 1 in series IV)
was based on the use of easily accessible carbonitriles I reacting
with DMF DMA (Scheme 1A, i). Further cyclization in a AcOH/HCl
(3:1) mixture of dimethylaminovinyl derivatives II led to high
yields of product IV (1) (Scheme 1A, iii). The same steps were used
to produce compound 2.
Primary 4-amino derivatives III (7–9) were obtained in high
yields by reacting intermediates II in 20% ammonia (Scheme 1A,
ii). For the synthesis of 4-substituted amino furo pyridines VI
(10–14), the 4-chloro-derivative V obtained from compounds IV
(7) and PCl5 interaction at 130–150 °C during 15–20 min was used
possessing a c
-lactone moiety.1,2 These two categories of lactones
are covalent inhibitors and, together with peptide mimetics, they
represent a large class of molecules known to target the 20S cata-
lytic core.26–28 Non-covalent inhibitors have been more recently
developed28 including hydroxyurea (HU) compounds,29 oxadiaz-
oles,30 and peptidic derivatives.31–34
(Scheme 1A, i
amines were reacted with V under reflux in xylene (Scheme 1A,
a), whereas secondary amines were reacted in boiling DMF
v). On a first way to compounds VI (10–14), primary
v
,
The aim of the present study was to synthesize new molecules
bearing a
c
-lactone moiety on the model of cerpegin and deriva-
(Scheme 1A,
(Scheme 1A,
v
, b), and aromatic amines were reacted in AcOH
tives1,2,35 with substitutions at the C4 of the furo[3,4-c]pyridine-
3-one skeleton. Optimization of their inhibitory power towards
c20S fractions in vitro was first investigated and the most interest-
ing compounds from this approach were selected for testing their
efficiency on i20S fractions. Furthermore, calpain 1 and cathepsin
B activities were assayed in order to check the specificity of protea-
some inhibition. Whole cell assays were also conducted to measure
intracellular proteasome inhibition and to evaluate cytotoxity.
New furo- and thieno[2,3-d]pyrimidine-4-one compounds were
also synthesized and assayed for their aptitude to inhibit c20S
proteasome. For the most potent compounds identified as inhibi-
tors of c20S or i20S, the possible mechanisms of binding were
v, c). All reactants are summed up in Supplementary
Table S1. On a second way, several substituted amino derivatives
VI (15–18) with a side-chain hydroxyl group were synthesized
from the primary amine III (7) through reaction with oxiranes
(Scheme 1A, vi). The presence of a hydroxyl group on a side-chain
could lead to significant improvement of the binding affinity to
proteasome active sites, as already reported in our investigations
on cerpegin N5 derivatives.2
For comparison purposes, 4-thioxo VIII (3, 4) (Scheme 1A) and
4-methyl XI (5, 6) (Scheme 1B) derivatives were synthesized. The
4-thioxo derivatives VIII (3, 4) were obtained from the correspond-
ing thiocarboxamides VII (Scheme 1A, i0) without isolating