Design and synthesis of potent inhibitors of the mono(ADP-ribosyl)transferase, PARP14

Article abstract of DOI:10.1016/j.bmcl.2017.04.089

A series of (Z)-4-(3-carbamoylphenylamino)-4-oxobut-2-enyl amides were synthesized and tested for their ability to inhibit the mono-(ADP-ribosyl)transferase, PARP14 (a.k.a. BAL-2; ARTD-8). Two synthetic routes were established for this series and several compounds were identified as sub-micromolar inhibitors of PARP14, the most potent of which was compound 4t, IC₅₀?=?160?nM. Furthermore, profiling other members of this series identified compounds with >20-fold selectivity over PARP5a/TNKS1, and modest selectivity over PARP10, a closely related mono-(ADP-ribosyl)transferase.

Full text of DOI:10.1016/j.bmcl.2017.04.089

Accepted Manuscript
Design and synthesis of potent inhibitors of the mono(ADP-ribosyl)transferase,
PARP14
Kristen Upton, Matthew Meyers, Ann-Gerd Thorsell, Tobias Karlberg, Jacob
Holechek, Robert Lease, Garrett Schey, Emily Wolf, Adrianna Lucente, Herwig
Schüler, Dana Ferraris
PII:
S0960-894X(17)30473-0
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.04.089
BMCL 24942
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
28 February 2017
22 April 2017
27 April 2017
Please cite this article as: Upton, K., Meyers, M., Thorsell, A-G., Karlberg, T., Holechek, J., Lease, R., Schey, G.,
Wolf, E., Lucente, A., Schüler, H., Ferraris, D., Design and synthesis of potent inhibitors of the mono(ADP-
ribosyl)transferase, PARP14, Bioorganic & Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/10.1016/
j.bmcl.2017.04.089
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.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Design and synthesis of potent inhibitors of the mono-
(ADP-ribosyl)transferase, PARP14
Leave this area blank for abstract info.
Kristen Uptona, Matthew Meyersa, Ann-Gerd Thorsellb, Tobias Karlbergb, Jacob Holecheka, Robert
Leasea, Garrett Scheya, Emily Wolfc, Adrianna Lucentec, Herwig Schulerb, Dana Ferrarisa
O
O
NH2
NH2
O
NH
O
O
NH
O
N
OH
N
Cl
PARP14 IC50 = 160nM

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.els evier.com
Design and synthesis of potent inhibitors of the mono(ADP-ribosyl)transferase,
PARP14
Kristen Uptona, Matthew Meyersa, Ann-Gerd Thorsellb, Tobias Karlbergb,Jacob Holecheka, Robert Leasea,
Garrett Scheya, Emily Wolfc, Adrianna Lucentec, Herwig Schülerb, Dana Ferrarisa
aMcDaniel College, 2 College Hill, Westminster, MD 21157
bKarolinska Institutet, Department of Medical Biochemistry and Biophysics, Scheeles väg 2, S-17177 Stockholm, Sweden
cStevenson University, 1525 Greenspring Valley Rd, Stevenson, MD 21153
ARTICLE INFO
ABSTRACT
Article history:
Received
A series of (Z)-4-(3-carbamoylphenylamino)-4-oxobut-2-enyl amides were synthesized and
tested for their ability to inhibit the mono-(ADP-ribosyl)transferase, PARP14 (a.k.a. BAL-2;
ARTD-8). Two synthetic routes were established for this series and several compounds were
identified as sub-micromolar inhibitors of PARP14, the most potent of which was compound 4t,
IC50 = 160nM. Furthermore, profiling other members of this series identified compounds with
>20 fold selectivity over PARP5a/TNKS1, and modest selectivity over PARP10, a closely
related mono-(ADP-ribosyl)transferase.
Revised
Accepted
Available online
2009 Elsevier Ltd. All rights reserved
.
Keywords:
PARP; PARP14; ARTD-8; mono(ADP-ribosyl)
transferase
Several members of the Poly-(ADP-Ribose) Polymerase
(PARP) family of proteins are attractive targets for oncology-
based drug discovery.1 Much of the research over the past three
decades has focused on the inhibition of ADP-ribosyl
rapid generation of a series of analogs; 3) Compound 1 has a low
molecular weight (MW = 234) allowing for some growth while
still maintaining parameters necessary for a chemical probe;10 4)
A recent paper described small molecule PARP14 inhibitors with
some selectivity over PARP1.11 These inhibitors bridge the
nicotinamide and the adenosine subsites. Derivatization of the
carboxylate of compound 1 provides an ideal direction to reach
the adenosine subsite and possibly the donor loop. Both of these
regions have residues unique to PARP14 and could lead to
selective PARP14 inhibitors. In the following letter we describe
the synthesis and activity of a series analogs derived from 1. A
series of (Z)-4-(3-carbamoyl-phenylamino)-4-oxobut-2-enyl
amides was synthesized and compounds 4l and 4t represent the
most potent derivatives against PARP14 in the literature. While
these inhibitors do not demonstrate appreciable selectivity over
PARP-1, compounds 4l and 4r display ~20 fold selectivity
against PARP-5a/5b, and modest selectivity over PARP10. This
polymerases PARP-1/2 and tankyrases (PARP 5a/b). However,
most other members of the PARP superfamily are mono-ADP-
ribosyltransferases, or are catalytically inactive.2 Research is
branching out toward these PARP family members, but potent
and selective tool compounds still do not exist for many of these
mono-(ADP-ribosyl)transferases.3 The pharmacological effect of
specifically inhibiting these other members of the PARP family
is unknown, thus specific inhibitors or probes must be designed
in order to delineate the effects of this inhibition. One of the
members of this family is PARP14 (a.k.a. BAL-2; ARTD-8), a
mono-(ADP-ribosyl)transferase. This protein has been
demonstrated to have a critical function in the progression of
multiple myeloma4 and roles in transcription control5 and cellular
metabolism.6 Because of its intriguing therapeutic prospects;
PARP14 has been identified as a potential drug discovery target.7
Two recent studies focused on identifying small molecule
inhibitors of PARP14.8,9 These studies identified a low
micromolar inhibitor of PARP14 (Z)-4-(3-
carbamoylphenylamino)-4-oxobut-2-enoic acid (IC50 = 7-20 µM,
Compound 1, Figure 1) and provided some initial SAR around
this compound. This compound represents an ideal starting point
for optimization for several reasons: 1) X-ray co-crystal data
exists so that the binding mode between 1 and PARP14 is known
and iterative structure based design can be used to optimize
compounds; 2) Compound 1 is readily synthesized allowing for

represents a viable starting point toward PARP14 selective
inhibitors.
formation of these byproducts, another route was established as
shown in Scheme 2. Piperidine 7 was acetylated with maleic
anhydride 3 in THF to afford the intermediate maleic amide 8 in
good yield. This amide was coupled with 3-amino benzamide
using standard HATU coupling conditions to afford the desired
amide 4b.
Our strategy for developing potent PARP14 inhibitor
derivatives consisted of the synthesis of amides from the 4-
oxobut-2-enoic acid group. Early attempts at this line of
optimization met with little success9 as shown by compound 4a
(IC50 = >20µM). Simply replacing the pyrrolidine group with a
slightly larger six membered ring such as a piperidine (4b, IC50
7.7µM) or piperazine (4c, IC50 = 11µM) led to two compounds
with potency comparable to compound 1 (Table 1). The next
logical step towards optimization of this series was to incorporate
substituents around each of these rings.
=
Scheme 1. Reagents and conditions: (a) THF, rt, 83%;8 (b)
DMF, DIEA, HNR1R2, HATU or TBTU, 10-97%.
Table 1. Inhibition of PARP14 by 4-oxobut-2-enyl amide
derivatives*
Initial SAR on this series9 confirmed several important
features of the binding mode: 1) the primary amide on the A-ring
(blue, Figure 1) forms three hydrogen bonds in the NAD+ binding
Compd
Structure
PARP14 IC50, µM
PARP1 IC50, µM
(pIC50 SD, M)
(pIC50 SD, M)
1
7.69
4.49
Scheme 2. Reagents and conditions: (a) THF, 68%; (b)
DMF, DIEA, HATU, 45%.
site of PARP14. Thus any derivatization around this amide or
replacement of the amide resulted in loss of potency; 2) the
double bond of the 4-oxo-2-butenoic acid is in a relatively small
pocket (yellow box, Figure 1), thus, methylation or replacement
of this moiety with large groups or rings resulted in loss of
activity; 3) A hydrogen bond exists between Asn1705 and the
amide oxygen of the 4-oxobut-2-enoic acid (orange circle, Figure
1), so removal of this amide will most likely result in decreased
potency; 4) A relatively large pocket (adenosine binding pocket)
exists adjacent to the carboxylate providing an obvious direction
for optimization and potential selectivity.11 Many of these
interactions are representative of most NAD+ competitive PARP-
1 inhibitors.12 This is due to the fact that the PARP family was
initially categorized based on homology within the catalytic
domain. Despite this homology in the nicotinamide binding site,
there are significant differences between the adenosine binding
site and the donor loop of PARP14 and that of e.g. PARP1.
Recent studies have demonstrated that selectivity can be achieved
by targeting the adenosine site,11 and the same has been done for
the tankyrases.3 Thus, our strategy for optimization and potential
selectivity was to modify this series of compounds with
4a
>20a
7.7
(5.12 0.09)
4b
11.2
(4.95 0.08)
4c
*IC50 values for inhibition of PARP14 catalytic fragment by 4b
and 4c were obtained as described.3
Several 2-, 3- and 4-substituted piperidine derivatives were
synthesized and analyzed in vitro as shown by compounds 4d-n
substituents that probe this portion of the enzyme.
(Table 2). The 2-phenyl (IC50 = 1.1µM) and 2-benzyl (IC50
3.5µM) piperidines 4d and 4e demonstrated slightly better
=
The synthesis of (Z)-4-(3-carbamoylphenylamino)-4-oxobut-
2-enyl amides 4d-v is outlined in Scheme 1. The carboxylic acid
1 was synthesized in good yield according to the literature
procedure using 3-aminobenzamide 2 and maleic anhydride 3.8
Coupling of compound 1 with secondary amines using HATU or
TBTU afforded the amides 4d-v. Interestingly, minor byproducts
were observed under these conditions and characterized as the
succinimides 6d-v. These derivatives presumably arise from the
intramolecular cyclization to form the maleimide 5 followed by
Michael addition of the amines. In some cases (e.g. piperidine),
the Michael addition occurred prior to the coupling resulting in a
complicated mixture of products. In order to circumvent the
potency than the parent compound 4b. Although the 3-substituted
ethyl ester 4f was inactive, notable improvements were also seen
with the 3-phenyl analog 4g (IC50 = 0.69µM) and the 3-phenoxy
4h (IC50 = 1.4µM) piperidine analogs. While 4g was the first sub-
micromolar derivative in our hands, this analog exhibited very
little selectivity against PARP1 (IC50 = 1.6µM) and PARP5a
(IC50 = 2.7µM). Similar to the 3-substituted ester, the 4-
substituted ester derivative 4i did not demonstrate appreciable
potency against PARP14. Homologation of this ester led to a
definite improvement in activity as shown by compound 4j (IC50
= 2.4µM). The 4-phenylpiperidine derivative 4k demonstrated
low micromolar activity, (IC50 = 2.2µM), but interestingly the

addition of an electron withdrawing fluorine atom led to
compound that was approximately 10x more potent (4l, IC50
0.27µM). Addition of a carbon between the piperidine and
*IC50 values for inhibition of full length PARP1 and PARP14
and PARP5a catalytic fragments were obtained as described.3
=
The inhibitory potency of these piperazine derivatives is
detailed in Table 3. While the core piperazine derivative 4c
displayed relatively modest potency (4c, IC50 = 11µM), the t-
butoxy carbonyl derivative 4o (IC50 = 0.37µM) demonstrated a
notable improvement. Improvements in potency were also noted
by incorporation of an aryl ring onto the piperazine nitrogen.
Compounds 4p, 4r, 4s and 4t all displayed potency below 1µM.
In fact, 4t (IC50 = 160nM) is almost an order of magnitude more
potent than the parent 4c and represents one of the most potent
PARP14 inhibitors in the literature. Addition of another aromatic
ring as demonstrated by the biphenyl derivative 4q (IC50 = 4.2
µM) was not as effective as just one aryl ring. Similarly,
aromatic ring also seemed to improve potency leading to another
sub-micromolar derivative (4m, IC50 = 0.45µM). More
importantly compounds 4l and 4m demonstrated notable
improvements in selectivity over PARP5a. Compound 4l was
~30 fold selective and compound 4m was ~11 fold selective over
PARP5a. Compound 4l was also ~10 fold selective over the
mono-ADP-ribosyltransferase, PARP10. Unfortunately, the most
potent derivatives of this series were nearly all equipotent against
PARP1, but these data were encouraging and prompted the
synthesis of structurally similar N-aryl piperazines to potentially
improve the potency and selectivity.
Table 2. Inhibition data by piperidinyl-amide derivatives*
incorporation of a heteroaryl group onto the piperazine (4v, IC50
= 2.8 µM), was also detrimental to inhibitory potency as was the
addition of an electron donating methoxy group on the aromatic
ring (4u, IC50 = 3.4 µM). The succinimide byproduct, 6r, of one
of the most potent aryl piperazine analogs, 4r, was completely
inactive (PARP14 IC50 >20µM). Unfortunately, the most potent
derivatives in this sub-series (i.e. 4p, 4r, 4s and 4t) displayed a
relative lack of selectivity against PARP1, but these analogs still
maintained reasonable selectivity over PARP5a (4p = 6.7 fold, 4r
= 23 fold, 4s = 3.9 fold, 4t = 5.6 fold) and PARP10 (4r = 7.8
fold, 4s = 6.1 fold). We also measured the inhibition of several
other mono-ADP-ribosyltransferases by 4s (IC50 values:
#
-NR1R2
PARP14
IC50 µM
PARP1 IC50
µM
PARP5a
IC50 µM
PARP12, 8.2 µM; PARP15, 0.52 µM; PARP16, 0.56 µM).
(pIC50 SD,
M)
(pIC50 SD,
M)
(pIC50 SD,
M)
1.1
4d
4e
(5.97 0.14)
3.5
(5.45 0.07)
21% @ 5
µM
4f
0.69
(6.16 0.09)
1.6
(5.81 0.3)
2.7
(6.24 0.1)
4g
Figure 2. Crystal structure of 4s bound to the catalytic domain of
PARP14. Protein sidechains discussed in the text are indicated.
Protein Data Bank entry 5NQE.
1.4
4h
4i
90%@5 µM
(5.99 0.13)
A crystal structure of the catalytic domain of PARP14 in
complex with 4s can illustrate what we believe to be the general
binding mode of the more potent derivatives of the scaffold
examined here (Figure 2). As expected, the aminobenzamide
moiety is bound in the nicotinamide pocket in a conformation
very similar to the vast majority of PARP inhibitors, which
occupy that site. The oxybutenyl linker is within hydrogen
bonding distance of the Tyr1721 hydroxyl. However, the
flexibility of the linker allows the distal fluorinated aryl ring of
4s to reside in a hydrophobic environment created by the
Leu1755, Val1773, and Val1784 side chains as well as Leu1782,
the small hydrophobic residue that replaces the catalytic
glutamate of PARP1. Our binding data indicate that the
28% @ 5
µM
2.4
4j
4k
4l
(6.39 0.15)
2.2
(5.65 0.08)
0.27
0.27
8.0
(6.56 0.09)
(6.56 0.1)
(5.1 0.09)
0.45
0.23
4.8
4m
4n
piperazine ring has a conformation that is slightly better suited to
accommodate these hydrophobic interactions, as compared to the
piperidine. Together, these results suggest that a number of
chemistry strategies remain to achieve more potent inhibition of
PARP14: The hydrophobic interaction site of 4s is delimited by
an aromatic amino acid (Tyr1714), with π-π stacking/interactions
(6.35 0.07)
(6.63 0.06)
(5.32 0.07)
0.65
(6.17 0.07)

0.37
6.0
potential, as well as by His1750 and His1753. Furthermore, the
D-loop is apparently flexible and can adopt to a number of ligand
induced conformations.3,8,9,11 Since this loop is relatively variable
among PARP family members, engagement of a non-conserved
D-loop side chain remains a promising strategy toward a potent
and selective mono-(ADP-ribosyl)transferase inhibitor.
0.9
4p
4q
4r
4s
(6.44 0.15)
0.8
(5.2 0.08)
9.8
(6.05 0.11)
4.2
(5.38 0.2)
0.41
(6.09 0.09)
0.51
(5.01 0.08)
9.4
Table 3. Inhibition data by piperazinyl-amide derivatives*
(6.39 0.1)
(6.3 0.11)
0.56
(5.03 0.06)
2.7
0.70
(6.15 0.07)
(6.25 0.04
(5.57 0.28)
0.22
(6.65 0.04)
0.9
(6.01 0.13)
0.16
(6.21 0.16)
4t
3.4
(5.47 0.11)
4u
4v
#
-NR1R2
PARP14
IC50 µM
PARP1
IC50 (µM)
PARP5a
IC50 (µM)
2.8
(5.55 0.1)
(pIC50 SD,
M)
(pIC50 SD,
M)
(pIC50 SD,
M)
*IC50 values for inhibition of full length PARP1 and PARP14
and PARP5a catalytic fragments were obtained as described.3
0.37
(6.43 0.09)
4o
2. Hottinger, M. O.; Hassa, P. O.; Lüscher, B.; Schüler, H.; Koch-
Nolte, F.; Trends Biochem. Sci., 2010, 35, 208-219.
3. Thorsell, A.-G.; Ekblad, T.; Karlberg, T.; Löw, M.; Pinto, A.F.;
Trésaugues, L.; Moche, M.; Cohen, M.S.; Schüler H. J. Med.
Chem. 2017, 60, 1262-1271.
In conclusion we outline the synthesis and activity of several
of the most potent inhibitors of PARP14 described. Despite the
lack of selectivity against PARP-1 several analogs in this series
display >10 fold selectivity over PARP5a and >5 fold selectivity
over closely related PARP10 (4r = 7.8 fold, 4s = 6.1 fold). Future
work will focus on achieving selectivity over PARP-1 by
improving interactions between the amide side chains and the D-
loop of PARP14.
4. Barbarulo A.; Iansante V.; Chaidos A.; Naresh K.; Rahemtulla A.;
Franzoso G.; Karadimitris A.; Haskard D. O.; Papa S.; Bubici C.,
Oncogene 2013, 32, 4231-4242.
5. Mehrotra P.; Riley R.; Patel F.; Li L.; Voss S.; Goenka S., J. Biol.
Chem. 2011, 286, 1767-1776.
6. Iansante V.; Choy P. M.; Fung S. W.; Liu Y.; Chai J.-G.; Dyson
J.; Del Rio A.; D’Santos C.; Williams R.; Chokshi S.; Anders R.
A.; Bubici C., Nature Comm. 2015, 6:7882 | DOI:
10.1038/ncomms8882.
Acknowledgments
7. Camicia R.; Winkler H. C.; Hassa P. O., Mol. Cancer 2015, 14, 1-
62.
Special appreciation goes to the Johns Hopkins University Drug
Discovery group, McDaniel College Chemistry Department,
McDaniel College Student Faculty Research Fund, Student
Research and Creativity Grant, and Dr. Garry Smith at the Fox
Chase Cancer Center for generous support. H.S. received funding
from the Swedish Cancer Society (2014/716), the Swedish
Research Council (2015-04603), and the IngaBritt and Arne
Lundbergs Research Foundation (no. 403). We thank the Protein
Science Facility at Karolinska Institutet/SciLifeLab and the
beamline staff at the Diamond Light Source (Didcot, UK) for
excellent support. Synchrotron data collection was supported by
the European Community’s Seventh Framework Programme
(FP7) under BioStruct-X (no. 283570).
8. Andersson, C. D.; Karlberg, T.; Ekblad, T.; Lindgren, A. E. G.;
Thorsell, A.-G.; Spjut, S.; Uciechowska, U.; Niemiec, M. S.;
Wittung-Stafshede, P.; Weigelt, J.; Elofsson, M.; Schüler, H.;
Linusson, A., J. Med. Chem. 2012, 55, 7706-7718.
9. Ekblad, T.; Lindgren, A. E. G.; Andersson, C. D.; Caraballo, R.;
Thorsell, A.-G.; Karlberg, T.; Spjut, S.; Linusson, A.; Schüler, H.;
Elofsson, M. Eur. J. Med. Chem. 2015, 95, 546-551.
10. Frye, S. V., Nature Chem. Biol. 2010, 6, 159-161.
11. Peng, B.; Thorsell, A.-G.; Karlberg, T.; Schüler, H.; Yao, S. Q.,
Angew. Chem. Int. Ed. 2016, 55, 1-7.
12. Ferraris, D. J. Med. Chem. 2010, 53, 4561-4584.
Supplementary data
Supplementary data contains experimental procedures,
enzyme purification and enzymatic assay details. These data can
be found online.
References and notes
1. Curtin, N. J., Sharma, R. A. eds. PARP Inhibitors for Cancer
Therapy, Springer International Press, Switzerland, 2015.