Utilizing a structure-based docking approach to develop potent G protein-coupled receptor kinase (GRK) 2 and 5 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2018.03.082

G protein-coupled receptor (GPCR) kinases (GRKs) regulate the desensitization and internalization of GPCRs. Two of these, GRK2 and GRK5, are upregulated in heart failure and are promising targets for heart failure treatment. Although there have been several reports of potent and selective inhibitors of GRK2 there are few for GRK5. Herein, we describe a ligand docking approach utilizing the crystal structures of the GRK2–Gβγ·GSK180736A and GRK5·CCG215022 complexes to search for amide substituents predicted to confer GRK2 and/or GRK5 potency and selectivity. From this campaign, we successfully generated two new potent GRK5 inhibitors, although neither exhibited selectivity over GRK2.

Full text of DOI:10.1016/j.bmcl.2018.03.082

Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Utilizing a structure-based docking approach to develop potent
G protein-coupled receptor kinase (GRK) 2 and 5 inhibitors
Helen V. Waldschmidt a,c, Renee Bouley b, Paul D. Kirchhoff c, Pil Lee c, John J.G. Tesmer a,b,d
,
Scott D. Larsen a,c,
⇑
a Department of Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor, MI, United States
b Department of Pharmacology and the Life Sciences Institute, University of Michigan, Ann Arbor, MI, United States
c Vahlteich Medicinal Chemistry Core, College of Pharmacy, University of Michigan, Ann Arbor, MI, United States
d Department of Biological Sciences, Purdue University, West Lafayette, IN, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
G protein-coupled receptor (GPCR) kinases (GRKs) regulate the desensitization and internalization of
GPCRs. Two of these, GRK2 and GRK5, are upregulated in heart failure and are promising targets for heart
failure treatment. Although there have been several reports of potent and selective inhibitors of GRK2
there are few for GRK5. Herein, we describe a ligand docking approach utilizing the crystal structures
Received 11 January 2018
Revised 7 March 2018
Accepted 28 March 2018
Available online xxxx
of the GRK2–Gb
c
ÁGSK180736A and GRK5ÁCCG215022 complexes to search for amide substituents pre-
dicted to confer GRK2 and/or GRK5 potency and selectivity. From this campaign, we successfully gener-
ated two new potent GRK5 inhibitors, although neither exhibited selectivity over GRK2.
Ó 2018 Elsevier Ltd. All rights reserved.
Keywords:
GRKs
Kinases
Docking
GPCRs
Crystallography
G protein-coupled receptor (GPCR) kinases (GRKs) regulate the
largest class of membrane receptors in the human genome via
phosphorylation of receptor C-terminal tails or cytoplasmic loops,
and are also implicated in several disease states.1,2 During heart
failure, levels of GRK2 and GRK5 are elevated in many tissues.3–5
In the heart, this upregulation leads to increased desensitization
and uncoupling of the GPCRs located in the heart such as the b-
adrenergic and angiotensin II receptors, which regulate contractil-
ity and blood flow to the body, respectively.6,7 Knockdown of either
GRK2 or GRK5 in mice subjected to transverse aortic constriction
showed cardio-protective effects.8,9
GRK2 and GRK5 also affect non-GPCR pathways that further
mediate stress responses in the heart.10–13 GRK2 influences cardiac
glucose uptake leading to abnormal cardiac metabolism when
upregulated, altering the growth of new cardiomyocytes.10
Uniquely, GRK5 is the only GRK known to be targeted to the cell
nuclei of cardiomyocytes14,15 where it acts as a histone deacetylase
(HDAC) kinase.11 Phosphorylation of HDAC5 leads to increased
expression of myocyte enhancer factor-2 which regulates the
stress response in hypertrophy.16,17 Due to these GPCR-indepen-
dent roles, GRK2 and GRK5 represent promising targets that offer
unique therapeutic outcomes that cannot be attained by current
heart failure treatments that directly target GPCRs or angioten-
sin-converting enzyme.
GRK2 and GRK5 have also been implicated in other medical
conditions. Elevated levels of cytosolic GRK2/5 have been impli-
cated in Alzheimer’s and Parkinson’s disease.18–20 GRK5 has addi-
tionally been shown to regulate tumor growth progression in
several cancers.20–23 Also of considerable interest are the roles of
GRK2 and GRK5 in cell growth and insulin levels leading to dia-
betes.24,25 Thus, chemical probes targeting either GRK2 or GRK5
(or potentially both), would be useful as tools to investigate the
roles of GRK2 and GRK5 in cells and human disease.
There are several reported GRK2 and GRK5 inhibitors (Fig. 1).
Compound 1o has a GRK2 IC50 of 460 nM but is ꢀ10-fold more
potent for GRK5 (IC50 of 59 nM).26 Thus it is one of the most potent
GRK5 inhibitors reported to date and one of the few that exhibits
some selectivity for GRK5 over GRK2. Limiting its usefulness is
its ability to inhibit tyrosine kinases (IC50 for c-Met, 8 nM and
IC50 for anaplastic lymphoma kinase, 0.3 l
M).26,27 Previously, our
lab identified compound GSK180736A as a GRK2-selective inhibi-
tor. Further development of this scaffold led us to CCG215022,
which potently inhibits both GRK2 and 5 (Fig. 1).12,13
The parent compound GSK180736A was co-crystallized with
GRK2–Gb
co-crystallized with GRK5 (PDB entry 4WNK), allowing the use of
⇑
Corresponding author at: Department of Medicinal Chemistry, College of
Pharmacy, University of Michigan, Ann Arbor, MI, United States.
c
(PDB entry 4PNK)28 whereas CCG215022 was
E-mail address: sdlarsen@umich.edu (S.D. Larsen).
https://doi.org/10.1016/j.bmcl.2018.03.082
0960-894X/Ó 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

2
H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Fig. 3. Virtual library of amides generated from GSK180736A carboxylic acid
homologs.
Fig. 1. Reported small molecule GRK inhibitors. GSK180736A and CCG215022 were
previously reported by our lab using a radioactivity assay. Compound 1o was
reported by Cho et. al. and its IC50 values were determined using a time-resolved
fluorescence resonance energy transfer assay. Due to different assay conditions the
values may not be directly comparable.
diversity of appendages in our larger ethylene linked scaffold
which already had a molecular weight of 437 g/mol.
A pharmacophore model was generated based on the ligand ori-
entations in the GRK2–Gb
cÁGSK180736A and GRK5ÁCCG215022
crystal structures. The model restricted the docked ligands to be
in an orientation similar to what was seen in the respective crystal
structures. As represented in Fig. 4, the center of the indazole and
the fluorophenyl rings were constrained by a spherical volume
indicated by the green circles in Fig. 4 (radii: 1.8 Å for the indazole
rings and 2.5 Å for the fluorophenyl). The two nitrogens of the
indazole that form hydrogen bonds with the hinge of the kinase
domain were constrained to a spherical volume with a radii of
1.8 Å (purple and cyan circles in Fig. 4). The goal was to allow these
motifs to move within a constrained volume of the active site such
that new amide substituents would necessarily be projected into
the hydrophobic subsite where they could pick up additional inter-
actions to increase potency of the molecules.
structure-guided drug design to develop potent GRK5 inhibitors.29
The two compounds bind similarly in the kinase active sites of
GRK2 and GRK5, respectively (Fig. 2). The indazole forms two
hydrogen bonds to backbone atoms in the hinge of the kinase,
the dihydropyrimidinone sits in the ribose subsite, and the fluo-
rophenyl ring packs under the P-loop. The amide-linked pyridine
extension of CCG215022 forms additional hydrogen bonds via its
amide with the P-loop and via the pyridine nitrogen with Lys220
and Asp329.28,29 We hypothesized that we could use these two
crystal structures and molecular modeling to design new com-
pounds in silico with improved selectivity and potency for either
GRK2 or GRK5.
The virtual compounds were then docked into ligand-free crys-
tal structures of GRK2–Gbc and GRK5 using a rigid model of the
Enumeration of virtual compounds and docking of those com-
pounds was conducted using the computational chemistry package
MOE.30 Our campaign began with an extensive virtual screen using
a library of commercially available amines from Sigma Aldrich. Vir-
tual compounds (amides B, Fig. 3) were enumerated based on the
GSK180736A template bearing homologous carboxylic acids off
the fluorophenyl ring (A) (Fig. 3). The three acid scaffolds A were
combined with primary and secondary amines (R2NH) with molec-
ular weights less than 215 g/mol. Initially nearly 15,000 amine
structures were retrieved. After removal of amines containing
expected reactive or mutagenic chemical motifs by MOE, the num-
ber of structures dropped to just over 11,000. The resulting amide
compounds B were then filtered by a molecular-weight cut off of
550 to give 9183 virtual unique amide-linked structures. We chose
a slightly higher molecular weight cut-off of 550 to allow for more
protein and a flexible ligand model that obeyed the constraints
of the pharmacophore model. The results were sorted by highest
docking scores (S), where a lower number correlates with tighter
binding, and then further divided into three groups. The first group
contained compounds predicted to be both potent and selective for
GRK2 (S score for GRK2 2 units lower than that of GRK5). The sec-
ond group contained compounds that were predicted to be both
potent and selective for GRK5 (S score for GRK5 2 units lower than
that of GRK2). The third group encompasses compounds that were
predicted to be equipotent for both GRK2 and GRK5 (having S
scores within 2 units of each other) and that had good docking
scores. After filtering out any diamines, carboxylic acids, or other
reactive, unstable, or toxic motifs that were missed in the
A
B
Phe197
P-loop
P-loop
Lys215
αC
αC
Hinge
Hinge
Asp272
Asp329
Thr264
Met266
Met274
Ala316
Ala321
Fig. 2. Comparison of the A) GRK2–Gb
with yellow carbons, CCG215022 is drawn with salmon carbons, and H-bonds are shown as dark grey dashed lines.
cÁGSK180736A (4PNK) and B) GRK5ÁCCG215022 (4WNK) crystal structures utilized in the docking campaign. GSK180736A is drawn
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
3
automated filter previously run in MOE, the forty highest scoring
virtual amides were synthesized and tested for kinase activity to
evaluate the predictive value of this docking approach.
New amides were prepared as previously described for direct
amide-linked appendages (n = 0)13 (Scheme 1). Treatment of
5-aminoindazole
1
with 2,2,6-trimethyl-4H-1,3-dioxin-4-one
resulted in the acetylated compound 2.31 Through a Biginelli
cyclization, catalyzed by ytterbium triflate, acids 3–5 (synthesis
of 3 and 4 are previously reported13 and synthesis of 5 is shown
below) were condensed with urea and acetamide 2 to yield the
dihydropyrimidones 6–8.32 Derivatives were then introduced via
amidation resulting in the final compounds 9–48.
The propanoic acid intermediate 5 was synthesized as shown in
Scheme 2 beginning with TBS protection of (4-fluorophenyl)
methanol 49 to give intermediate 50. Directed lithiation followed
by formylation gave the aldehyde 51, which was then deprotected
to furnish alcohol 52. Condensation with malonic acid followed by
hydrogenation of the resulting alkene gave acid 54. Parikh-Doering
oxidation of the benzylic alcohol cleanly provided aldehyde 5.
Biochemical evaluation of these compounds was performed
using phosphorylation assays against GRK1, 2, and 5, represent-
ing the three major GRK subfamilies, as previously described.33
Fig. 4. Pharmacophore model used for molecular docking. GRK2–Gb
cÁGSK180736A
crystal structure. The green circles represent regions in which the aromatic rings
were confined. The cyan and purple circles represent the hydrogen bonds made
between the indazole and the backbone of the hinge in the GRK2 and 5 kinase
domains. The nitrogens of the indazole were confined to these spheres. The pink
oval shows the position where virtual amides were appended and allowed to freely
extend into the hydrophobic subsite.
Scheme 1. Preparation of new amides. Reagents and conditions: a) 2,2,6-trimethyl-4H-1,3-dioxin-4-one, CH3CN, 100 °C, 16 h, 69% b) 3, 4, or 5, and Yb(OTf)3, urea, CH3CN,
100 °C, 4 h, 63–84% c) HATU, DIEA, R2NH, DMF, 16 h, 5%–85%.
Scheme 2. Synthesis of propanoic acid intermediate 5. a) TBSCl, imidazole, DIEA, DCM, 12 h, 74%, b) sBuLi, TMEDA, THF, À78 °C then DMF, rt, 3 h, 77%, c) TBAF, THF, 6 h, 74%,
d) Malonic acid, pyridine, piperidine, 100 °C, 4 h, 80%, e) 10% Pd/C, H2, EtOAc, 24 h, 91%, f) SO3-pyridine, DMSO, Et3N, 20 min, 79%.
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

4
H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
The compounds ranged dramatically in terms of both potency
and selectivity from the sub-mM to high mM range. The results
of the top forty ranked compounds in terms of their activity
against GRKs 1, 2, and 5 as well as their GRK2 and GRK5 docking
scores (S) are summarized in Tables 1–3. Of the forty hybrid
compounds that were docked, synthesized, and tested, 23
(57.5%) showed the predicted selectivity and potency. In evaluat-
ing these compounds, we divided them into those that were pre-
dicted to be GRK2 selective and/or potent (Table 1), those that
were predicted to be GRK5 selective and/or potent (Table 2),
and those that were predicted to be similarly potent for both
GRK2 and GRK5 (Table 3).
linked analogues (n = 0) all bound with sub-mM potency except
for the N-methyl analog (14) which contains a tertiary amide and
thus loses its ability to accept a hydrogen bond from the P-loop.
The direct amide-linked compounds all retained selectivity for
GRK2 over GRK5 (averaging 22-fold) with the least selective analog
being the imidazo pyridine 16 (3.5-fold). Although the imidazo
substituent is bulky, it may retain low micromolar potency
(GRK5 IC50 = 1.9 mM) because of similarities of its substituent nitro-
gen to that of CCG215022. Of the other two core scaffolds (n = 1
and n = 2), all but the 5-methyl isoindoline 9 showed reduced
potency against GRK2 compared to the parent GSK180736A
compound.
Compounds predicted to be GRK2 selective and/or potent are
ordered in Table 1 from the predicted lowest to highest binding
affinity based on their docking scores (S). As a whole, the amide
substituents in this group were the bulkiest and most conforma-
tionally restricted of the three sets, consistent with the previously
observed structure-activity relationship (SAR) trend we reported
wherein the size and shape of the hydrophobic subsite heavily
influences GRK selectivity.13 GRK2 has a larger and shallower sub-
site pocket in comparison to GRK5, allowing bulkier D-ring sub-
stituents to selectively bind GRK2.13 Overall, the direct amide-
There were only five compounds that were predicted to be
GRK5 selective and/or potent, which may reflect the fact that the
smaller hydrophobic pocket in GRK5 makes it more difficult to
accommodate a broad array of D-ring substituents (sorted by
GRK5 S score in Table 2). Four of the five compounds had an n =
1 linker, and one had an n = 2 linker. All were, however, selective
for GRK2 over GRK5 (10 – 28-fold) with the most GRK2 potent
compound being the cyclopropylpiperidine analog 27 (GRK2 IC50
=
0.68 mM) and the most GRK5 potent compounds being 27 and
the imidazole analog 29 (both with an IC50 of 19 mM).
Table 1
GRK activity of indazole-dihydropyrimidinone based compounds predicted to be GRK2 selective and potent.
Compound
R
n
1
GRK2
(IC50 mM)
GRK1
(IC50 mM)
GRK5
(IC50 mM)
GRK2 S
GRK5 S
9
0.20 0.06
>100
6.3 1.5
À7.9
À2.2
10
11
2
2
3.8
4.7
1
3
>100
>100
>100
>100
À8.2
À8.3
ND
À5.1
12
13
1
1
2.1
2
>100
>100
6.1 2.7
>100
À8.4
À8.4
À5.9
À4.3
1.9 0.6
14
0
14
2
>100
>100
À8.6
À6.6
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
5
Table 1 (continued)
Compound
R
n
2
GRK2
(IC50 mM)
GRK1
(IC50 mM)
GRK5
(IC50 mM)
GRK2 S
GRK5 S
15
16
1.8 0.4
>100
>100
À8.6
À2.7
0
0.54 0.06
17
3
1.9 0.6
À8.7
ND
17
18
19
1
0
2
1.9 0.3
0.40 0.1
29 30
>100
27 13
2.8 0.6
>100
À8.8
À8.9
À9.0
ND
16 4.4
>100
À5.9
À6.0
20
0
0.64 0.08
>100
18 12
À9.1
À6.3
21
22
1
0
1.9 0.2
>100
>100
38 13
À9.1
À9.2
À6.2
À5.0
0.50 0.08
21
6
23
0
0.34 0.03
>100
9.5 0.18
À9.2
ND
24
25
0
1
44 12
>100
>100
>100
À9.5
À9.6
À7.5
À7.4
1.5 0.4
12
2
*All IC50 measurements are an average of three separate experiments run in duplicate. Errors shown represent standard error of the mean. S is the docking score generated by
MOE. ND stands for ‘‘not docked” as MOE was unable to fit the motif into the pocket according to the restrictions given.
The third set of compounds evaluated were those that were pre-
dicted to be potent inhibitors of both GRK2 and GRK5. These eigh-
teen compounds are shown in Table 3 sorted by their predicted
binding potentials (weakest to tightest) for either GRK2 or GRK5.
A wide variety of appendages were predicted to bind well to both
GRK2 and GRK5, including bulky lipophilic substituents, smaller
lipophilic substituents, and various heterocycles. The bulkier lipo-
philic appendages as well as the smaller lipophilic substituents (31,
32, 36, 38, 39, 41, 42, 44, 46, 47, and 48) all exhibited GRK2 selec-
tivity and very poor binding to GRK5 (no sub-mM IC50 values). The
cyclohexyl methyl ester analog (45), which exhibited some
lipophilicity via its cyclohexyl and some polarity via its methyl
ester, had the best GRK2 IC50 (0.16 mM), and was 120-fold selective
over GRK5.
All five heterocyclic compounds from this set bound to both
GRK2 and GRK5 with potencies lower than 5 mM with the excep-
tion of the pyrazole 40. Two of these heterocyclic compounds
bound nearly equipotently to GRK2 and GRK5: the oxadiazoles
37 (GRK2 IC50 = 0.16 mM, GRK5 IC50 = 0.38 mM) and 33 (GRK2
IC50 = 0.25 mM, GRK5 IC50s = 0.26 mM). Moving forward it may be
advantageous to expand upon these results by investigating a
broader range of small heterocyclic appendages.
To better understand the molecular basis for how the equipo-
tent GRK2/GRK5 inhibitor 33 exhibits improved potency against
GRK2, the crystal structure of 33 in complex with GRK2–Gbc was
determined using conditions described previously (PDB: 6C2Y).12
The inhibitor binds analogously to our previously reported
GSK180736A-based GRK2 inhibitors (Fig. 5A)13 and in accordance
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

6
H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Table 2
GRK activity of indazole-dihydropyrimidinone based compounds predicted to be GRK5 selective and potent.
Compound
R
n
1
GRK2
(IC50 mM)
GRK1
(IC50 mM)
GRK5
(IC50 mM)
GRK2 S
GRK5 S
26
1.4 0.3
>100
27
6
À7.2
À9.2
27
1
0.68 0.04
>100
19 10
À7.1
À9.4
28
29
1
1
1.6 0.09
2.0 0.3
>100
>100
25
9
À8.1
À7.7
À10.2
À10.2
19 10
30
2
4.0 0.5
>100
66
9
À8.3
À10.7
*All IC50 measurements are an average of three separate experiments run in duplicate. Errors shown represent standard error of the mean. S is the docking score generated by
MOE.
with the pharmacophore model utilized in the docking campaign.
The carbonyl of the amide linker forms an additional hydrogen
bond with the backbone nitrogen of Phe202. The oxadiazole ether
then packs back into the hydrophobic pocket with the oxygen fac-
ing the solvent exposed region of the active site.
linked compounds likely because the added length of the linkers
allowed the amide bond to project further into the hydrophobic
subsite or, if there is a clash, out into solvent. Generally, addition
of large lipophilic appendages was unfavorable and resulted in a
decrease in potency except for cases where there was a basic
nitrogen that may have contributed to improved water solubility.
Our screening results also showed that there is little apparent
correlation between the docking score (S) and the potency of inhi-
bition exhibited by the selected compounds. We thus carried out
additional docking of all forty compounds allowing both the ligand
and kinase to be flexible (Supplementary Table 2). Flexible docking
Comparison of the GRK2–Gb
c
Á33 crystal complex with the ini-
tial, best scored docking pose of 33 bound to GRK2 (Fig. 5B)
revealed, however, that the docking algorithm was unable to cor-
rectly model the oxadiazole ether appendage in the hydrophobic
subsite. This discrepancy is likely due to the fact that the protein
was kept rigid during docking. For example, the crystal structure
shows up to a 2.6 Å shift of the P-loop (at the Phe202 C
relative to the predicted docking pose, allowing the oxadiazole to
pack into a slightly larger hydrophobic pocket.
Our docking campaign also allowed us to obtain further insight
into the SAR of GSK180736A derivatives. The most favorable and
best predicted analogues for GRK activity were the direct
amide-linked moieties, likely due to the fact that the two crystal
structures used as receptors for the docking campaign were crys-
tallized with ligands that also have a direct amide-linker. Tertiary
amides were better tolerated in the methylene- and ethylene-
a
carbon)
was able to better reproduce the crystal structure of GRK2–Gb
c
Á33;
however, this second set of flexible docking scores still did not cor-
relate well with the IC50 values. This likely reflects the fact that
kinases are known to be highly flexible and it is difficult to predict
the movement of the surrounding elements of the GRK active site.
Additionally, as is well accepted in the computational chemistry
community, docking scores are notoriously poor at quantitatively
predicting binding affinities. Rather docking is useful for qualita-
tive binding pose generation and compound prioritization for
synthesis.
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
7
Table 3
GRK activity of indazole – dihydropyrimidinone based compounds predicted to be potent against both GRK2 and GRK5.
Compound
R
n
1
GRK2 (IC50 mM)
GRK1 (IC50 mM)
GRK5 (IC50 mM)
GRK2 S
GRK5 S
31
3.5 0.7
>100
>100
À8.0
À9.7
32
33
0
0
4.6
1
>100
>100
À8.1
À9.3
À9.4
0.25 0.07
30 20
0.26 0.2
À10.9
34
35
1
1
0.92 0.1
>100
3.5 0.8
1.7 0.3
À9.6
À11.0
À9.3
0.46 0.08
26 10
À10.3
36
0
1.2 0.7
>100
71 20
À10.5
À9.6
37
38
0
1
0.16 0.07
2.4 0.5
3.0 0.4
>100
0.38 0.2
À9.7
À9.8
À11.0
À11.0
13
7
39
40
0
0
4.1 2.5
1.1 0.1
>100
>100
77 30
À9.8
À11.2
À9.8
44
8
À10.4
41
42
2
1
12
5
>100
>100
>100
À10.5
À10.0
À9.8
1.5 0.40
16 6.1
2.4 0.4
À11.0
43
44
1
1
0.43 0.2
0.83 0.3
>100
>100
À10.0
À10.1
À11.0
À11.0
9.3
19
4
45
0
0.16 0.2
83 20
3
À10.4
À10.1
(continued on next page)
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

8
H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Table 3 (continued)
Compound
R
n
1
GRK2 (IC50 mM)
GRK1 (IC50 mM)
GRK5 (IC50 mM)
21
GRK2 S
GRK5 S
46
0.75 0.05
>100
4
À10.6
À10.6
47
0
4.1 0.9
>100
>100
À10.7
À11.2
48
0
29 20
>100
>100
À10.8
À11.0
*All IC50 measurements are an average of three separate experiments run in duplicate. Errors shown represent standard error of the mean. S is the docking score generated by
MOE.
bulk of binding energy. This challenge is further evident in the fact
that of the forty initial most potent compounds predicted by dock-
ing/scoring, only five were predicted to bind selectively to GRK5
while seventeen were predicted to bind selectively to GRK2. Mov-
ing forward with this particular series of compounds, it seems
promising to further investigate the oxadiazole, isoxazole, and
pyrazole compounds with direct amide-linkages to further
improve GRK5 potency, perhaps in parallel with altering features
of the core of the GSK180736A scaffold.
Declaration of interest
None.
Acknowledgements
This work was supported by NIH grants HL071818, and
HL122416 (to J.T.), and American Heart Association grant
15PRE22730028 (to H.V.W.). J.T. and S.D.L. were supported by
grants from the Center for Discovery of New Medicine, University
of Michigan. Use of the Advanced Photon Source was supported
by the U. S. Department of Energy, Office of Science, Office of Basic
Energy Sciences, under Contract No. DE-AC02-06CH11357, and use
of LS-CAT Sector 21 was supported by the Michigan Economic
Development Corporation and Michigan Technology Tri-Corridor
Grant [085P1000817].
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmcl.2018.03.082.
References
Fig. 5. Comparison of the GRK2–Gb
predicted pose. A) The crystal complex is shown with 33 in dark green. Hydrogen
bonds are shown as black dashes and the 3 omit map is shown as a magenta
cÁ33 (PDB: 6C2Y) crystal structure and its
1. Suo WZ, Li L. Dysfunction of G Protein-Coupled Receptor Kinases in Alzheimer’’s
Disease. Sci World J. 2010;10.
2. Guccione M, Ettari R, Taliani S, Da Settimo F, Zappalà M, Grasso S. G-protein-
r
mesh. B) The docked (cyan) and the crystal (green) poses overlaid.
coupled receptor kinase
2 (GRK2) inhibitors: current trends and future
perspectives. J Med Chem. 2016;59:9277–9294.
3. Rockman HA, Koch WJ, Lefkowitz RJ. Seven-transmembrane-spanning
receptors and heart function. Nature. 2002;415:206–212.
4. Huang ZM, Gold JI, Koch WJ. G protein-coupled receptor kinases in normal and
failing myocardium. Front Biosci Landmark. 2011;16:3047–3060.
5. Anastasios L, Giuseppe R, Walter JK. GRK2 inhibition in heart failure: something
old, something new. Curr Pharm Des. 2012;18:186–191.
6. Eschenhagen T. Beta-adrenergic signaling in heart failure – adapt or die. Nat
Med. 2008;14:485–487.
7. Hullmann J, Traynham CJ, Coleman RC, Koch WJ. The expanding GRK
interactome: Implications in cardiovascular disease and potential for
therapeutic development. Pharmacol Res. 2016;110:52–64.
8. Gold JI, Gao E, Shang X, Premont RT, Koch WJ. Determining the absolute
requirement of G protein-coupled receptor kinase 5 for pathological cardiac
hypertrophy: short communication. Circ Res. 2012;111:1048–1053.
9. Rockman HA, Choi DJ, Rahman NU, Akhter SA, Lefkowitz RJ, Koch WJ. Receptor-
specific in vivo desensitization by the G protein-coupled receptor kinase-5 in
transgenic mice. Proc Natl Acad Sci. 1996;93:9954–9959.
Nonetheless, our in silico screening was successful at identifying
two new compounds that showed high potency against both GRK2
and GRK5 (33 and 37). Inhibiting GRK2 and GRK5 simultaneously
may prove advantageous in treating heart failure, as it would inhi-
bit multiple processes implicated in this disease via both GPCR
related and unrelated mechanisms.4,8 However, we were unable
to discover any compounds that displayed high GRK5 selectivity.
This may be due to the fact that the lead GSK180736A compound
already has intrinsic GRK2 selectivity over GRK5 (>130-fold) and
that GRK5 has a smaller, more restrictive hydrophobic subsite.
Thus, we had the difficult task of trying to build in potency for
GRK5 while simultaneously attempting to reduce potency for
GRK2 without altering the core of the scaffold responsible for the
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082

H.V. Waldschmidt et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
9
10. Ciccarelli M, Chuprun JK, Rengo G, et al. G protein-coupled receptor kinase 2
activity impairs cardiac glucose uptake and promotes insulin resistance after
myocardial ischemia. Circulation. 2011;123. 1953-U245.
11. Martini JS, Raake P, Vinge LE, et al. Uncovering G protein-coupled receptor
kinase-5 as a histone deacetylase kinase in the nucleus of cardiomyocytes. Proc
Natl Acad Sci. 2008;105:12457–12462.
22. Geras-Raaka E, Arvanitakis L, Bais C, Cesarman E, Mesri EA, Gershengorn MC.
Inhibition of constitutive signaling of kaposi’s sarcoma–associated herpesvirus
G protein-coupled receptor by protein kinases in mammalian cells in culture. J
Exp Med. 1998;187:801–806.
23. Kim JI, Chakraborty P, Wang Z, Daaka Y. G-protein coupled receptor kinase 5
regulates prostate tumor growth. J Urol. 2012;187:322–329.
12. Homan KT, Larimore KM, Elkins JM, Szklarz M, Knapp S, Tesmer JJG.
24. Wang F-L, Tang L-Q, Wei W. The connection between GRKs and various
signaling pathways involved in diabetic nephropathy. Mol Biol Rep.
2012;39:7717–7726.
25. Li H, Gan W, Lu L, et al. A genome-wide association study identifies GRK5 and
RASGRP1as Type 2 diabetes loci in chinese hans. Diabetes. 2013;62:291–298.
26. Cho SY, Lee BH, Jung H, et al. Design and synthesis of novel 3-(benzo[d]oxazol-
2-yl)-5-(1-(piperidin-4-yl)-1H-pyrazol-4-yl)pyridin-2-amine derivatives as
selective G-protein-coupled receptor kinase-2 and -5 inhibitors. Bioorg Med
Chem Lett. 2013;23:6711–6716.
Identification and Structure-Function Analysis of Subfamily Selective
G
Protein-Coupled Receptor Kinase Inhibitors. ACS Chem Biol. 2014;10:310–319.
13. Waldschmidt HV, Homan KT, Cruz-Rodríguez O, et al. Structure-based design,
synthesis, and biological evaluation of highly selective and potent G protein-
coupled receptor kinase 2 inhibitors. J Med Chem. 2016;59:3793–3807.
14. Johnson LR, Scott MGH, Pitcher JA.
G protein-coupled receptor kinase 5
contains DNA-binding nuclear localization sequence. Mol Cell Biol.
a
2004;24:10169–10179.
15. Yi YP, Gerdes AM, Li FQ. Myocyte redistribution of GRK2 and GRK5 in
hypertensive, heart-failure-prone rats. Hypertension. 2002;39:1058–1063.
16. McKinsey TA, Olson EN. Toward transcriptional therapies for the failing heart:
chemical screens to modulate genes. J Clin Invest. 2005;115:538–546.
17. Chang SR, McKinsey TA, Zhang CL, Richardson JA, Hill JA, Olson EN. Histone
deacetylases 5 and 9 govern responsiveness of the heart to a subset of stress
signals and play redundant roles in heart development. Mol Cell Biol.
2004;24:8467–8476.
18. Obrenovich ME, Palacios HH, Gasimov E, Leszek J, Aliev G. The GRK2
overexpression is a primary hallmark of mitochondrial lesions during early
alzheimer disease. Cardiovasc Psychiatry Neurol. 2009;2009:327360.
19. Er B, Vv G, Jn J, Jl B, Ev G. Arrestins and two receptor kinases are upregulated in
Parkinson’s disease with dementia. Neurobiol Aging. 2008;29:379–396.
20. Wu C-C, Tsai F-M, Shyu R-Y, Tsai Y-M, Wang C-H, Jiang S-Y. G protein-coupled
27. Lee J, Han S-Y, Jung H, et al. Synthesis and structure–activity relationship of
aminopyridines with substituted benzoxazoles as c-Met kinase inhibitors.
Bioorg Med Chem Lett. 2012;22:4044–4048.
28. Homan KT, Wu E, Wilson MW, Singh P, Larsen SD, Tesmer JJG. Structural and
functional analysis of
G protein-coupled receptor kinase inhibition by
paroxetine and a rationally designed analog. Mol Pharmacol. 2014;85:237–248.
29. Homan KT, Waldschmidt HV, Glukhova A, et al. Crystal structure of G protein-
coupled receptor kinase 5 in complex with a rationally designed inhibitor. J Biol
Chem. 2015.
30. Molecular Operating Environment (MOE), Chemical Computing Group ULC:
1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2017.
31. Sridharan V, Ruiz M, Menéndez JC. Mild and high-yielding synthesis of b-keto
esters and b-ketoamides. Synthesis. 2010;2010:1053–1057.
32. Drewry DHE, Brian; Goodman, Krista B, Green, Darren, Victor, Steven, Jung,
David, Kendall; Lee, Denis; Stavenger, Robert A, Wad, Sjoerd, Nocolaas
Chemical Compounds. 29 December 2004, 2004.
33. Homan KT, Tesmer JJG. Molecular basis for small molecule inhibition of G
protein-coupled receptor kinases. ACS Chem Biol. 2015;10:246–256.
receptor kinase
5 mediates Tazarotene-induced gene 1-induced growth
suppression of human colon cancer cells. BMC Cancer. 2011;11. 175-175.
21. Métayé T, Menet E, Guilhot Jl, Kraimps J-L. Expression and activity of g protein-
coupled receptor kinases in differentiated thyroid carcinoma. J Clin Endocrinol
Metab. 2002;87:3279–3286.
Please cite this article in press as: Waldschmidt H.V., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.082