2
J. Holechek et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
proliferation, a selective inhibitor would provide drug discovery
groups with an optimizable lead for an anti-cancer therapeutic;
3) several clinical PARP1 inhibitors also inhibit PARP10 and the
effects of this polypharmacology cannot be thoroughly examined
until selective inhibitors are discovered.3
in more potent derivatives of PARP10 and PARP14 with selectivity
over all other members of the PARP family. This manuscript outli-
nes the design, synthesis and evaluation of a series of diaryl ethers
as selective inhibitors of mono-(ADP-ribosyl)transferases PARP10
and PARP14. Two lines of optimization led to a PARP10 selective
A
recent study disclosed an interesting chemotype with
compound 8b with aqueous solubility of 42 lM at pH 7.4 and rea-
sonable metabolic stability (16% after 2 h). In addition, two sub-
remarkable selectivity towards PARP10.10 This compound
(OUL35, 1, Fig. 1) displayed sub-micromolar potency towards
micromolar PARP14 inhibitors were discovered, compounds 8k
PARP10 (IC50 = 0.329
lM) and more importantly > 300 fold selec-
(PARP14 IC50 = 0.78 lM) and 8m (PARP14 IC50 = 0.7 lM). Both of
tivity towards the poly(ADP-ribosyl)ating members of the PARP
family (PARP1, PARP2, PARP5a and PARP5b). This compound repre-
sents a good scaffold for selective inhibitors of PARP10 and poten-
tially other mono-(ADP-ribosyl)transferases for several reasons: 1)
Compound 1 possesses a low molecular weight (MW = 256) leav-
ing room for the inevitable increase in MW during optimization;
2) While unoptimized, 1 possesses some selectivity towards
these derivatives were selective over PARP1. Crystallography stud-
ies unearthed a binding mode that takes advantage of the same
hydrophobic pocket near the d-loop of these two enzymes as we
noted in the 3-amino benzamide series.11 Taken together, these
data make these compounds useful probes for these two enzymes
to further understand their biological function as potential drug
targets.
mono-(ADP-ribosyl)transferases (PARP10 IC50 = 0.329
IC50 = 23.4 M, PARP15 IC50 = 4.17 M) versus poly(ADP-ribosyl
polymerases (PARP1/2, 5a/b IC50 > 100
M);10 3) The diaryl ether
l
M, PARP14
The synthesis of the diaryl ether scaffold and amide derivatives
7b-c and 8a-r is outlined in Scheme 1. In the first step, a nucle-
ophilic aromatic substitution was conducted between methyl 3-
and 4-hydroxybenzoates 3a-b and 4-bromobenzonitrile 4 in mod-
erate to good yield affording the esters 5a and 5b. Hydrolysis of the
esters and nitrile was conducted in the same step affording the car-
boxylate amides 6a-b. Coupling of various amines to these car-
boxylic acids afforded the 3- and 4-substituted amide derivatives
7b-c and 8a-r.
As mentioned above, the first logical direction to explore was 3-
and 4-substituted diaryl amides with a similar structure to those
from the 3-amino benzamide series 2a-b (Fig. 2). This strategy
could presumably lead to compounds that would have the follow-
ing characteristics: 1) beneficial interactions with the side pocket
of the d-loop as seen in our previous publication,11 thus maintain-
ing or improving the PARP10 and PARP14 potency over compound
1; 2) minimal inhibitory potency against the polymerases PARP1/2
and 5a/b, thus improving the selectivity towards mono-(ADP-ribo-
syl) transferases; 3) improved physicochemical properties (e.g.
cLogP) making these derivatives amenable to validation studies
in cellular systems. Tables 1 and 2 outline the results from these
structure activity studies.
l
l
l
scaffold can be readily synthesized making this series amenable
to rapid synthetic diversification; 4) binding models indicate that
modification of one of the benzamide moieties will allow for sub-
stituents that probe the D-loop region, an area of the enzyme
bridging the NAD + binding pocket that has low sequence homol-
ogy amongst the overall closely related mARTs. Therefore, opti-
mization of 1 might potentially lead to further improvements in
selectivity.
Another recent study from our labs11 described a series of 3-
amino benzamide derivatives with several analogs possessing
sub-micromolar potency against PARP14. One member of this ser-
ies (2a, Fig. 1), displayed an IC50 value of 160 nM and good physic-
ochemical properties (cLogD = 2.2 at pH 7) providing a basis for a
small molecule chemical probe for PARP14. While 2a demon-
strated ꢀ6-fold selectivity over PARP5a, it was relatively non-
selective against PARP1. This is not surprising given that 3-
aminobenzamide moiety is a common pharmacophore for PARP1
inhibitors.12 Taken together, the 3-aminobenzamide series and
the diaryl ether series demonstrated an interesting overlap and a
potential pharmacophore for mono-(ADP-ribosyl)transferases as
shown in Fig. 2. The similarities between compounds 1 and 2a
include the following: 1) Both compounds have an electron rich
benzamide moiety which most likely binds the nicotinamide bind-
ing site of these mARTs through a series of hydrogen bonds;3,10 2)
both compounds have an oxygen atom positioned near this benza-
mide ring (orange circle); 3) both compounds incorporate another
planar structure near this oxygen (yellow box).
While the chemical space around the 3-amino benzamide series
has been established, similar structure activity studies around the
diaryl ether series have not been conducted. Because the aryl
amide moiety (green box, Fig. 2) was shown to improve the
PARP14 inhibitory potency dramatically,11 a logical direction for
structure activity studies is through the synthesis of 3- and 4-sub-
stituted diaryl amides. We hypothesized incorporation of an aryl
piperidine or piperazine onto the diaryl ether scaffold would result
Under our assay conditions, the potency of compound 1 against
PARP10 was 0.64
the trend noted in the original report.10 Very similar inhibition data
were seen with the 3-substituted isomer 8a over PARP10 (IC50
0.51 M). From our previous studies, we noted that aryl piperidi-
lM and IC50 = 10 lM against PARP14, similar to
=
l
nes, particularly fluoro- and chloro-substituted aryl piperidines
improved the potency and bound to a unique side pocket near
the D-loop. For this reason, 4-fluorophenyl piperidine amides 7b
and 8b and 3-fluorophenyl piperidines 7c and 8c were synthesized
and tested. An immediate trend was noted in that the 3-substi-
tuted amides (8b and 8c) were more potent against both PARP10
and PARP14 than the 4-substituted isomers (7b and 7c). For this
reason, the rest of our derivatization efforts focused on 3-substi-
tuted amides. In addition, both of these derivatives were ꢀ5–10-
fold more selective towards PARP10 than PARP14 (Table 1).
Scheme 1. aReagents and conditions: (a) DMSO, 130 °C, 76% yield for 5a, 45% yield
for 5b; (b) H2O2, NaOH, MeOH, 89% yield for 6a, 75% yield for 6b; (c) HNR1R2, HATU,
DIEA, DMF, 18–95% yield.
Figure 2. Proposed pharmacophore for mono-(ADP-ribosyl) transferase inhibitors.