Structure-Based Design of Highly Selective and Potent G Protein-Coupled Receptor Kinase 2 Inhibitors Based on Paroxetine

Article abstract of DOI:10.1021/acs.jmedchem.7b00112

In heart failure, the β-adrenergic receptors (βARs) become desensitized and uncoupled from heterotrimeric G proteins. This process is initiated by G protein-coupled receptor kinases (GRKs), some of which are upregulated in the failing heart, making them desirable therapeutic targets. The selective serotonin reuptake inhibitor, paroxetine, was previously identified as a GRK2 inhibitor. Utilizing a structure-based drug design approach, we modified paroxetine to generate a small compound library. Included in this series is a highly potent and selective GRK2 inhibitor, 14as, with an IC₅₀ of 30 nM against GRK2 and greater than 230-fold selectivity over other GRKs and kinases. Furthermore, 14as showed a 100-fold improvement in cardiomyocyte contractility assays over paroxetine and a plasma concentration higher than its IC₅₀ for over 7 h. Three of these inhibitors, including 14as, were additionally crystallized in complex with GRK2 to give insights into the structural determinants of potency and selectivity of these inhibitors.

Full text of DOI:10.1021/acs.jmedchem.7b00112

Subscriber access provided by HACETTEPE UNIVERSITESI KUTUPHANESI
Article
Structure-Based Design of Highly Selective and Potent G Protein-
Coupled Receptor Kinase 2 Inhibitors Based on Paroxetine
Helen V. Waldschmidt, Kristoff T. Homan, Marilyn C. Cato, Osvaldo Cruz-Rodríguez, Alessandro Cannavo,
Michael W. Wilson, Jianliang Song, Joseph Y. Cheung, Walter J. Koch, John J.G. Tesmer, and Scott D Larsen
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00112 • Publication Date (Web): 21 Mar 2017
Downloaded from http://pubs.acs.org on March 24, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure-Based Design of Highly Selective and
Potent G Protein-Coupled Receptor Kinase 2
Inhibitors Based on Paroxetine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b,†
b
b,c
Helen V. Waldschmidt , Kristoff T. Homan , Marilyn C. Cato , Osvaldo Cruz-Rodríguez ,
d,
e
d
d
Alessandro Cannavo , Michael W. Wilson , Jianliang Song , Joseph Y. Cheung , Walter J.
d
a,b,c
a,e
Koch , John J. G. Tesmer , and Scott D. Larsen*
a
b
Department of Medicinal Chemistry, College of Pharmacy, Departments of Pharmacology and
c
d
Biological Chemistry, Life Sciences Institute, PhD Program in Chemical Biology, Center for
e
Translational Medicine, Temple University, Philadelphia, Pennsylvania, 19140, Vahlteich
Medicinal Chemistry Core, University of Michigan, Ann Arbor, Michigan, 48109
KEYWORDS: G proteinꢀcoupled receptor kinases, G proteinꢀcoupled receptors, βꢀadrenergic
receptors, protein crystallography, inhibitors, kinases, paroxetine, CCG258208, CCG215022,
CCG224406, Takeda101, Takeda103A, GSK180736A
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract. In heart failure, the βꢀadrenergic receptors (βARs) become desensitized and uncoupled
from heterotrimeric G proteins. This process is initiated by G proteinꢀcoupled receptor kinases
(GRKs), some of which are upregulated in the failing heart making them desirable therapeutic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
targets. The selective serotonin reuptake inhibitor, paroxetine, was previously identified as a
GRK2 inhibitor. Utilizing a structure based drug design approach we modified paroxetine to
generate a small compound library. Included in this series is a highly potent and selective GRK2
inhibitor, 14as, with an IC of 30 nM against GRK2 and greater than 230ꢀfold selectivity over
50
other GRKs and kinases. Furthermore, 14as showed a 100ꢀfold improvement in cardiomyocyte
contractility assays over paroxetine and a plasma concentration higher than its IC50 for over
seven hours. Three of these inhibitors, including 14as, were additionally crystallized in complex
with GRK2 to give insights into the structural determinants of potency and selectivity of these
inhibitors.
Introduction
G proteinꢀcoupled receptors (GPCRs) mediate cellular responses to many different kinds
1
of extracellular stimuli and, with over 800 members, they are the largest class of receptors in the
human genome. Given their many physiological roles, they are targeted by a large fraction of the
2
pharmaceuticals on the market as well as illicit drugs. Upon agonist binding, GPCRs recruit and
activate heterotrimeric G proteins, stimulating a variety of downstream intracellular signaling
events including the recruitment of GPCR kinases (GRKs), which regulate GPCRs via
3
, 4
phosphorylation of their cytoplasmic loops or tails. In response, βꢀarrestins bind to the
5
,
phosphorylated receptors to induce desensitization and ultimately internalization of the GPCRs.
6
ACS Paragon Plus Environment

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
There are seven known GRKs in humans, subdivided by homology into three subꢀ
families: GRK1, GRK2, and GRK4. All GRKs belong to the protein kinase A, G, and C (AGC)
family. The GRK1 subfamily consists of GRK1 and GRK7 and their expression is largely
confined to the retina. The GRK2 subfamily consists of the ubiquitously expressed GRK2, and
GRK3, which predominates in olfactory neurons. The GRK4 subfamily consists of GRK4, which
is mainly expressed in the testes and kidneys, and GRK5 and GRK6, which are ubiquitously
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
ꢀ10
expressed, although GRK6 is more prevalent in the brain.
In response to the failing heart, the sympathetic nervous system acts to increase the levels
11ꢀ13
of circulating catecholamines such as norepinephrine and epinephrine.
Upon binding of these
hormones to βꢀadrenergic receptors (βARs) in cardiomyocytes, downstream signaling is initiated
4
, 14
to improve the positive inotropic response in the heart.
In the failing heart, activation of βARs
15, 16
also leads to upregulation of GRK2 and GRK5
which in turn leads to uncoupling of the
1
1, 16, 17
βARs from G proteins,
decreased βARs at the cellular membrane, and decreased cardiac
1
3, 18, 19
output in response to hormonal stimulation.
Decreasing the levels of GRK2, in particular
through gene knockdown in the heart, has been shown to be beneficial in preventing heart failure
by renormalizing levels of catecholamines and cell surface βARs, and improving cardiac
function in mouse and swine models, making GRK2 an attractive target for heart failure
1
9ꢀ21
therapy.
Efforts to selectively target GRK2 have generated several reported inhibitor series.
22
Among these are a series of highly potent and selective compounds developed by Takeda. Two
of these inhibitors, Takeda103A (1) and Takeda101 (2) (IC = 20 nM and 30 nM, respectively
50
with >50ꢀfold selectivity over other AGC kinases, Figure 1), were crystallized with the GRK2ꢀ
23
Gβγ complex in order to ascertain their binding modes. They both stabilize an inactive
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
conformation of GRK2, and their C and D rings pack into the polyphosphate and hydrophobic
subsites of the active site, respectively, which contributes to their high potency and selectivity.
2
3
Although these compounds never advanced to clinical trials, they served as proof of principle for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
24
development of new GRK2 inhibitors in our laboratory.
The FDAꢀapproved serotoninꢀreuptake inhibitor, paroxetine hydrochloride, was
2
5
identified as a modest GRK2 inhibitor with an IC of 1.4 µM (Figure 1). Further in vivo
5
0
investigation in a mouse heart failure model revealed that paroxetine improves cardiac function
postꢀmyocardial infarction and renormalizes the levels of catecholamines and βꢀadrenergic
26
receptor density in the heart, effects that persist for up to two weeks postꢀtreatment. Because
paroxetine is only a modestly potent and selective inhibitor of GRK2, but also an FDAꢀapproved
drug with excellent ADMET properties, it was an attractive lead compound for further
optimization as a GRK2 selective agent.
In parallel, we identified another, more potent GRK2 inhibitor, GSK180736A (3) with an
2
7
IC of 0.8 ꢁM (Figure 1), a compound that is similar to paroxetine in both size and shape.
5
0
This compound was originally developed as a Rhoꢀassociated coiledꢀcoil containing kinase 1
2
8
(ROCK1) inhibitor (IC = 100 nM), but exhibited limited bioavailabilty. Through a hybridꢀ
5
0
based drug design approach utilizing compounds 3 and 1 and their respective GRK2 crystal
structures, we generated a small library of potent and selective GRK2 inhibitors. Some of these
compounds such as 4 (CCG224406, Figure 1) were demonstrated to enhance myocardial
24
contractility in mouse cardiomyocytes at a 20ꢀfold lower concentration than paroxetine.
Comparison of the binding poses of paroxetine and 3 in the active site of GRK2 (Figure
2A) shows that both compounds form interactions with the adenosine, ribose, and polyphosphate
subsites, but neither occupies the hydrophobic subsite like the Takeda compounds (Figure 2B),
ACS Paragon Plus Environment

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
leaving room to increase potency and selectivity through the addition of favorable moieties to
their fluorophenyl rings. In the adenine subsite, the benzodioxole ring of paroxetine makes
analogous hydrogen bonds to the backbone of atoms in the kinase domain hinge as the indazole
ring of 3, with the exception that one of the paroxetine hydrogen bonds is a CH–O hydrogen
bond. The Bꢀrings of the two compounds both occupy the ribose subsite, where paroxetine
makes a hydrogen bond with the carbonyl of Ala321 that is not seen with 3. The fluorophenyl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5, 27
rings of both molecules pack underneath the Pꢀloop, superimposing nearly perfectly.
on these largely overlapping binding modes, we hypothesized that the structure activity
Based
relationships from our 3 hybrid campaign would be translatable, providing guidance for novel
extensions to the paroxetine scaffold. Furthermore, because paroxetine is an FDAꢀapproved drug
we anticipated that the analogous paroxetineꢀseries of compounds would exhibit superior
pharmacokinetic properties. Paroxetine has nearly twoꢀfold lower polar surface area, a lower
molecular weight, and fewer hydrogen bond donors and acceptors than 3, physical properties
2
9
that are all predictive of better absorption and permeability (Figure 1).
We report herein the successful ligation of amide moieties to the fluorophenyl ring of
paroxetine and generation of significantly more potent inhibitors of GRK2 (up to 50ꢀfold) with
high selectivity over GRK subfamilies and other AGC kinases (up to 230 fold over GRK1 and 5
as well as ROCK1 and PKA). Three representative compounds were crystallized revealing a new
hydrogen bond with the amide linker and new polar contacts formed with the extensions in the
hydrophobic subsite. Finally, we demonstrate that one of our optimized inhibitors (14as)
enhances cardiomyocyte contractility with approximately 100-fold greater potency than
paroxetine.
Results
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Chemistry. Initial hybrid analogues 14aa – 14ay were synthesized as described in Scheme 1.
Synthesis commenced with commercially available ((3S,4R)ꢀ4ꢀ(4ꢀfluorophenyl)ꢀ1ꢀ
methylpiperidinꢀ3ꢀyl)methanol (5). Silyl protection of 5 using tertꢀbutyl dimethyl silyl (TBS)
chloride followed by Nꢀdemethylation with phenyl chloroformate (PCF) gave the phenyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0
carbamate 6. Base hydrolysis provided the free amine, which was then protected to give tꢀbutyl
carbamate 7. Secꢀbutyllithium mediated aryl lithiation of the position ortho to the fluorine,
3
1
followed by trapping with carbon dioxide, yielded acid 8a. The acid was then methylated using
trimethylsilyldiazomethane to give 9a. TBS deprotection of the alcohol followed by mesylation
and nucleophilic displacement with sesamol afforded 11a. The methyl ester was then hydrolyzed
under basic conditions to give the carboxylic acid 12a. Coupling of the free acid to various
amines and final Boc deprotection yielded the carboxamides 14aa ꢀ ay.
Synthesis of homologated analogues 14ba – 14bd began with secꢀbutyllithiumꢀmediated
ortho lithiation of intermediate 7 followed by formylation with N, Nꢀdimethylformamide to yield
3
1
the carboxaldehyde. Subsequent reduction with sodium borohydride gave the alcohol 8b.
Mesylation of the benzylic alcohol with methanesulfonic anhydride followed by nucleophilic
substitution with sodium cyanide produced 9b. TBS deprotection of the remaining alcohol
followed by mesylation and nucleophilic substitution with sesamol gave 11b. Base hydrolysis of
the nitrile afforded carboxylic acid 12b. The acid was then coupled to various amines to give the
carboxamides 13ba - bd which were then Bocꢀdeprotected to provide the desired homologated
compounds 14ba - bd.
Structure Activity Relationships. All new compounds were assayed for in vitro activity against
bovine GRK1, 2 and 5, as well as PKA and ROCK1 (Table 1). We envisioned that incorporation
of an amide linker between the paroxetine scaffold and new hydrophobic subsiteꢀbinding
ACS Paragon Plus Environment

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substituents would allow for the formation of a hydrogen bond to the backbone nitrogen of
2
3, 24
Phe202 in the Pꢀloop, as observed in our compound 3 derived series.
However, simple
methyl amide (14aa) was twoꢀfold less potent than paroxetine against GRK2, suggesting that any
hydrogen bond that forms does not overcome desolvation and/or entropic penalties. Increasing
the size of the amide substituent from a methyl to a benzyl (14ab) was tolerated but not any more
potent than paroxetine alone. Lengthening the amide substituent to phenethyl 14ac resulted in a
slight loss in potency against GRK2, indicating that more lipophilicity on its own is not
beneficial to improving potency. Moving to the 2,6–difluorobenzylamide (14ad), mimicking the
Dꢀring of compound 1 and adding bulk, had no change in potency versus GRK2 and resulted in a
decrease in selectivity against GRK5. Additionally, the added lipophilicity gave a modest
increase in ROCK1 inhibition (34% at 10 µM). Further lengthening the substituent to the 2,6–
difluorophenethylamide (14ae) resulted in an even further decrease in selectivity for GRK2 over
GRK5 as it exhibited a 10ꢀfold increase in potency for GRK5 and a 2ꢀfold decrease in potency
for GRK2. Moving to a hybrid of compound 2 we investigated the addition of a trifluoromethyl
benzyl (14af) which resulted in a loss in potency for all three GRKs, suggesting a limit to the
size of the substituent that could be tolerated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We next tested structure activity relationships generated by our previously reported
compound 3ꢀbased series, which revealed that bulkier substituents in the hydrophobic subsite of
GRK2 can achieve both increased potency for GRK2 as well as increased selectivity over the
24
other GRKs and related AGC kinases like ROCK and PKA. Three representative substituents
from this endeavor are 2,6–dimethylbenzylamide, 14ag, 2,6–dichlorobenzylamide, 14ah, and
2,6–dimethoxybenzylamide, 14ai. Although the added bulk improved selectivity against
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ROCK1, all compounds displayed reduced selectivity for GRK2 over GRK1 and 5, and no
improvement in potency, suggesting inadequate engagement of the hydrophobic subsite.
Less bulky but still lipophilic amide substituents were then investigated. In our previous
SAR campaign, the most potent compound against GRK2 (IC = 60 nM) placed a 2ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
2
4
methoxybenzylamide substituent in the hydrophobic subsite. Surprisingly, translating the 2ꢀ
methoxybenzylamide onto the paroxetine scaffold resulted in our least potent GRK2 inhibitor,
1
4aj (IC = 40 ꢁM). Interestingly, homologating the 2ꢀmethoxybenzylamide by adding a
50
methylene group between the amide and fluorophenyl ring (14ba) brought back some GRK2
potency (IC = 2.4 ꢁM) but also picked up affinity for GRK5. These results, in addition to those
5
0
above, revealed that although the binding poses and structures of 3 and paroxetine are similar,
the SAR of their extended scaffolds are poorly translatable. One possible explanation is that the
overall conformation of the GRK2 kinase domain in complex with compound 3 is more closed
than in the GRK2·paroxetine complex due to different interactions formed by their hingeꢀbinding
moieties. This results in small but significant structural differences in the polyphosphate and
hydrophobic subsites and the relative orientation of the small and large lobes that form the active
2
5, 32
site.
As bulky and lipophilic groups were not tolerated, we turned our attention to smaller
more polar groups including a series of regioisomeric pyridyl methylamides. The 2ꢀpyridine,
14ak, afforded a 2ꢀfold increase in potency for GRK2 along with a 5ꢀfold increase in GRK5
potency. Thus, for the first time we were able to modestly improve GRK potency. The
corresponding 3ꢀ and 4ꢀpyridyl analogs (14al and 14am) were less potent, suggesting that the
added potency of 14ak was due to the position of the nitrogen rather than the aromatic ring
alone. Lengthening the carbon linker between the amide and the pyridines (14an, 14ao, 14ap) or
ACS Paragon Plus Environment

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
between the amide and the phenyl ring (14bb, 14bc) also reduced potency, providing evidence
for the importance of the position of the pyridyl nitrogen relative to the fluorophenyl ring. From
the overall SAR of these pyridine amides, it was clear that the ortho position of the nitrogen in
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4ak is key for potency against both GRK2 and GRK5.
To more efficiently exploit this polar contact we next sought to design molecules that
would have a stronger hydrogen bond acceptor in a similar position as the pyridine nitrogen of
4ak. The 2ꢀimidazolylmethyl amide 14aq, as well as its homolog 14ar, showed potency
comparable to 14ak. They also had increased potency for GRK5, presumably by making a
1
33
hydrogen bond to Lys220 as we observed previously in the GRK5·215022 (17) complex.
Replacing the imidazole amide with 3ꢀpyrazolylmethyl amide (14as), resulted in a dramatic
increase in potency for GRK2 (IC = 30 nM) while maintaining 230ꢀfold selectivity over GRK5
5
0
and more than 2500ꢀfold selectivity over GRK1, PKA, and ROCK1. Lengthening the amide
linker to give 3ꢀpyrazolylethyl amide, 14at, resulted in a 26ꢀfold decrease in potency for GRK2,
relative to 14as. Furthermore, methylation of either nitrogen of 14as (14av and 14aw) resulted in
4
0ꢀfold and 70ꢀfold decreases in GRK2 affinity, respectively, whereas placing a methyl group
adjacent to the two nitrogens (14au) did not diminish potency. These results strongly suggest
that both nitrogens of the pyrazole of 14as are necessary for its high GRK2 potency and are
likely making either one or two polar interactions. Simple oneꢀatom transposition of the two
pyrazole nitrogens (14ax) resulted in a 10ꢀfold loss in GRK2 potency relative to 14as, further
confirming the importance of the location of the nitrogens. Homologation of the 4ꢀpyrazole
methyl amide of 14ax (14ay, 14bd) could not rescue the loss in potency.
Crystallographic Analysis. To confirm our design strategy and to guide our analogue design
early in our campaign, we determined the 2.6 Å crystal structure of 14ak bound to the GRK2–
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Gβγ complex (Figure 3A). The complex crystallized in an unusual space group (P2) not
previously observed for this assembly, suggesting unique conformational changes in the kinase
domain. We additionally determined coꢀcrystal structures of 14bd and 14as in the space group
C2 at resolutions of 2.2 Å and 3.0 Å, respectively (Figure 3B and 3C). Due to the addition of
their various amide linked D rings, there are slight variations in the fitting of their A, B and C
rings among the molecules, although overall they bind similarly to paroxetine (PDB ID:
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
5
3
V5W). The benzodioxole A rings of the three inhibitors form analogous hinge interactions,
the piperidine B ring sits puckered in the ribose subsite where it forms a hydrogen bond with the
backbone carbonyl of Ala321, and the fluorophenyl C ring packs under the Pꢀloop in the
polyphosphate subsite. As seen in our previously reported compound 3 derived inhibitors, the
carbonyl of the amide linkers of all three inhibitors forms a hydrogen bond with the backbone
nitrogen of Gly201. Beyond the amide linker of 14bd, the electron density for the pyrazole is
relatively poor, likely because it extends out of the active site towards solvent. Because the D
rings of 14ak and 14as do occupy the hydrophobic subsite, we interpret the divergent packing of
14bd as a consequence of the extra methylene preceding its amide group, which makes the
substituent too long to pack into the hydrophobic subsite. Phe202 in the 14bd complex also
adopts a distinct rotamer from the paroxetine, 14ak, and 14as complexes, which allows it to pack
with the solventꢀextended pyrazole of 14bd. Relative to 14bd, the amide nitrogen of 14ak and
14as forms an additional hydrogen bond with the side chain of Asp335, which is also a
consequence of the fact that 14bd is homologated, moving this nitrogen out of reach.
The coꢀcrystal structure of the 14bd crystal complex was the most analogous to the
parent paroxetine complex in that it aligns nearly identically (r.m.s.d. deviation of 0.22 Å for the
Cα atoms of residues 185ꢀ271 in the small lobe). This is consistent with our SAR in that 14bd
ACS Paragon Plus Environment

Page 11 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
has only a 2ꢀfold increase in GRK2 potency relative to paroxetine but retains similar IC values
5
0
for both GRK1 and GRK5. The increase in potency for GRK2 seems unlikely to be due to the
additional hydrogen bond picked up by the amide linker because 14aa, which contains just an
amide substituent, is twoꢀfold less potent. Instead, it is likely due to additional buried surface
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
4
area, a key feature for GRK2 inhibitors. Selectivity is likely similar to that of paroxetine
because of the lack of contacts of its Dꢀring in the hydrophobic subsite.
Although the 2ꢀpyridine of analogue 14ak packs snuggly into the hydrophobic subsite of
GRK2, this seems to require displacement of the αB helix and adjoining loops away from the
active site by up to 1.3 Å, consistent with a larger r.m.s.d. of 0.76 Å for the Cα atoms in the
small lobe of the kinase compared to the paroxetine complex, as well as a unique hinge
conformation that changes the relative orientation of the small and large lobes by ~15˚ relative to
the 14ak and 14bd coꢀcrystal structures. These changes are probably responsible for the unique
crystal form that this complex adopts. Because the SAR showed that placement of the pyridine
nitrogen at the 3 or 4 position was detrimental to potency, it had been surmised that a polar
contact or hydrogen bond might exist between the nitrogen and the catalytic lysine as seen in our
3
3
formerly reported GRK5·17 inhibitor complex. Here the pyridine nitrogen is too distant (4.5 Å)
and not in the right orientation to form a hydrogen bond with the side chain of the catalytic lysine
(GRK2ꢀLys220), as does the analogous pyridine D ring in the GRK5·17 coꢀcrystal structure.
2
4
Thus it is instead modeled facing the solvent, as in our GRK2ꢀGβγ·17 structure. The loss in
GRK2 potency upon migration of the pyridine nitrogen to the 3 or 4 position is likely a
consequence of the loss of lipophilicity at those positions, which can make nonꢀpolar contacts in
the hydrophobic subsite. The increased potency exhibited by 14ak for both GRK2 and GRK5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
relative to paroxetine is therefore likely due to additional van der Waals interactions formed by
the D ring in the hydrophobic subsite.
The small lobe in the 14as coꢀcrystal structure is more similar to the paroxetine complex
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(r.m.s.d. of 0.41 Å) than the 14ak complex (r.m.s.d. of 0.86 Å). As in 14ak, the nonꢀ
homologated pyrazole moiety of 14as packs in the hydrophobic subsite but packs lower in the
site than the pyridine of 14ak. The pyrazole nitrogens are within hydrogen bond distance of the
side chains of both Glu239 and Lys220, with the closest being formed with Glu239. Therefore, it
is most likely that the significant increase in potency for GRK2 results from not only the specific
interactions picked up by the pyrazole nitrogens, but also because the pyrazole does not require a
conformational change ( pushing out the αB helix) as required for the pyridine of 14ak. The
carbon of the pyrazole adjacent to the two nitrogens packs against Leu235 in the αC helix, thus it
is not clear how 14au (methylated) retains high potency, but the observed interactions explain
why 14av and 14aw (Nꢀmethylated variants), 14ax and 14aq (pyrazole nitrogens moved around
the ring) exhibit large decreases in potency, as does lengthening the linker between the amide
and the pyrazole (14at).
Without a crystal structure of GRK5 in complex with one of our paroxetineꢀderived
compounds, the ability of these pyrazole derivatives to retain such high selectivity over GRK5
remains unclear. We have observed that the hydrophobic subsite in GRK2 can accommodate
2
4, 33, 34
bulkier Dꢀrings than GRK5.
Furthermore, of the compounds reported here only those with
nonꢀhomologated amide linkages exhibit any potency for GRK5, suggesting that the hydrogen
bond formed between the amide nitrogen and GRK2ꢀAsp335 (GRK5ꢀAsp329) is important for
potency against GRK5. The nonꢀhomolgated methyleneꢀlinked pyrazoles, 14as and 14au,
exhibited the highest potency for GRK5 (but retained >100 fold selectivity for GRK2), consistent
ACS Paragon Plus Environment

Page 13 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with the idea that hydrogen bond formation by the amide linker in combination with favorable
polar contacts is similarly important for GRK5 potency. Thus, it seems most likely that
differences in the kinase domain hinges and, consequently, the overall conformation of the
GRK2 and GRK5 kinase domains are responsible for the observed selectivity in the paroxetineꢀ
derived compounds. For example, if one overlays the small lobe of the GRK2·14as complex
with that of the GRK5·17 complex, the glycine in the DFG loop of the large lobe of GRK5
would sterically collide with the pyrazole of 14as. The preferred conformation of the GRK5
hinge may also engender less optimal interactions with the benzodioxole of paroxetine derived
inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Metabolic Stability. To guide the design and selection of analogs with favorable
pharmacokinetic (PK) properties, we evaluated the stability of selected compounds to incubation
with mouse liver microsomes (MLM) (Table 2). Our lead compound, paroxetine, had a t1/2 of 24
min in our MLM assay while 3 had a t1/2 of 20.6 min and 2 had a t of only 1.9 min. Appending
1/2
lipophilic carboxamides onto 3, giving our previously reported GRK2 inhibitors (15, 16, 4),
resulted in a substantial drop in stability with all three compounds having a t1/2 lower than 4
24
min. Conversely, addition of the same moieties onto the paroxetine scaffold (analogs 14ag,
14ah, 14ai) could be done without eroding metabolic stability and in the case of 14ag improving
upon the t1/2 of paroxetine twoꢀfold. Addition of a 2ꢀpyridine carboxamide to the scaffold of
compound 3 (17) was still detrimental but not as poorly tolerated as the lipophilic appendages.
Surprisingly, on the paroxetine template, the polar carboxamides (14ak, 14as, 14a4) exhibited
substantial metabolic instability, perhaps due to targeting of the heterocyclic nitrogens to the
heme of cytochrome P450. These conflicting results between the paroxetine and 3 scaffolds
suggest that the mechanism of metabolism differs substantially between the two series.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Contractility in Mouse Cardiomyocytes.As an underlying mechanism of heart failure is the
inability of the heart to properly contract we selected compounds 14ak and 14as in addition to
Takeda 2 for evaluation in an in vivo contractility assay. Following incubation with varying
doses of the inhibitors to give a baseline contraction (Figure 4, grey bars) mouse cardiomyocytes
were then stimulated with the βAR agonist isoproterenol. The resulting maximal increase in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
contraction was then measured (Figure 4, black bars). As inhibition of GRK2 should increase
the number of activated βARs we would expect our inhibitors to enhance the maximal increase
in comparison to a DMSO control. Previously, we showed that both paroxetine and 3 produce an
increase in contractility. The minimum dose needed for paroxetine to produce a significant
2
4, 25
response was 10 ꢁM, whereas 3 showed similar efficacy at 1 ꢁM.
the higher potency of 3 relative to paroxetine.
These results agree with
Evaluation of 14ak in the cardiomyocyte contractility assay did not show a significant
increase in contractility at 0.5 ꢁM or 1 ꢁM, consistent with its modest twoꢀfold increase in
GRK2 inhibition potency. The highly potent 14as, on the other hand, showed a significant
increase in contractility at a concentration of only 0.1 ꢁM, a 100-fold lower concentration than
paroxetine, consistent with the approximately 50ꢀfold increase in potency 14as has for GRK2
relative to paroxetine. Additionally, 14as showed a 5ꢀfold improvement over Takeda lead 2,
although both compounds had equal potency for GRK2 (30 nM), suggesting that 14as may have
better cell permeability. In comparison to our previously reported compound 4, this paroxetine
24
hybrid inhibitor also shows a 5ꢀfold improvement in efficacy. Importantly, these results suggest
that significant improvement in βARꢀstimulated contractility in mouse cardiomyocytes can be
achieved with potent GRK2ꢀselective inhibition.
ACS Paragon Plus Environment

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Preliminary Pharmacokinetic Study in Mice
.
Compounds 2 and 14as, were evaluated in an
abbreviated in vivo pharmacokinetic study in mice (Table 3). The compounds were
intraperitoneally injected into CDꢀ1 mice at a dose of 10 mg/kg, plasma samples were collected
at four different time points over seven hours, and drug levels in plasma were quantified. Due to
the abbreviated nature of this study, PK parameters could not be rigorously calculated.
Nevertheless, it is clear the halfꢀlife in vivo for 14as which can be estimated from the four time
points to be approximately one hour is superior to that of 2 which is too low to quantify in this
study. Of particular relevance to future in vivo efficacy studies in mice, 14as shows total plasma
drug levels after single IP administration that exceed the GRK2 IC for seven hours, although
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
we do not yet know at this time what the free drug concentration actually is More detailed
pharmacokinetic parameters and potential hERG liability due to the piperidine will need to be
assessed during further development of these compounds as potential heart failure therapeutics.
Discussion and Conclusions
Because paroxetine was identified as a modestly potent GRK2 inhibitor with good
selectivity we sought to improve potency for GRK2 while retaining selectivity. Our previous
GRK2 inhibitor series based on 3 successfully utilized a hybrid approach to develop improved
GRK2 inhibitors, and thus we hypothesized that a similar approach could be used with
paroxetine. A small library of paroxetineꢀderived GRK2 inhibitors were then synthesized, most
of which retained selectivity for GRK2. Although we discovered that the SAR of our previous
series was not translatable to the paroxetine series, we did learn that smaller heterocycles were
advantageous to increasing GRK2 potency. Addition of a 3ꢀpyrazolylmethyl amide to the
paroxetine scaffold (14as), resulted in a highly potent GRK2 inhibitor (IC = 30 nM) with high
5
0
selectivity over GRK1, GRK5, PKA, and ROCK1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Coꢀcrystal structures of three of the synthesized paroxetine derivatives were determined,
revealing additional hydrogen bonds with the added amide linker and the ability of GRK2 to
adopt a more open conformation than in the 3-based series, perhaps due to how the benzodioxole
group packs with the hinge. This probably underlies the high selectivity of these compounds. The
hydrophobic subsite exhibits some conformational flexibility in GRK2, as evidenced by the 14ak
complex compared to the other two structures, and excludes the D ring if it is too big to bind in
the pocket or is unable to overcome a desolvation penalty (both of which probably apply to
14bd).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We successfully improved contractility in mouse cardiomyocytes at concentrations as
low as 100 nM (14as), which is 100ꢀfold lower than paroxetine and 5ꢀfold lower than the Takeda
compound 2, suggesting that selective and potent inhibition of GRK2 is sufficient for cardiac
stimulation in the failing heart (i.e. does not require inhibition of GRK5). Evaluation in a short
pharmacokinetics study revealed that 14as was able to maintain plasma concentrations higher
than its IC for over seven hours, indicative of its potential as an in vivo therapeutic. 14as has
5
0
thus been selected for additional studies in murine models of heart failure in which paroxetine
2
6
has previously been evaluated. Because 14as shows such improved potency in cardiomyocyte
assays, we hypothesize that it will exhibit superior results in vivo relative to paroxetine,
confirming the utility of this series of GRK2 inhibitors as optimized leads for heart failure
therapeutics.
Experimentals
Chemistry. All reagents were used without further purification as received from commercial
1
sources unless noted otherwise. H NMR spectra were taken in DMSOꢀd6, MeOD, or CDCl at
3
room temperature on Varian Inova 400 MHz or Varian Inova 500 MHz instruments. Reported
ACS Paragon Plus Environment

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
chemical shifts for the H NMR spectra were recorded in parts per million (ppm) on the δ scale
from an internal standard of residual tetramethylsilane (0 ppm). Mass spectrometry data was
measured using a Waters Corporation Micromass LCT or Agilent6230 QꢀTOF. HPLC was used
to determine purity of biologically tested compounds on an Agilent 1100 series with an Agilent
Zorbax Eclipse Plus−C18 column. A gradient of 10ꢀ90% acetonitrile/water over 6 min followed
by 90% acetonitrile/water for 7 min was used with detection at 254 nm. All tested compounds
had purity >95%. Solvent abbreviations used: MeOH (methanol), DCM (dichloromethane),
EtOAc (ethyl acetate), hex (hexanes), DMSO (dimethylsulfoxide), DMF (dimethylformamide),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H O (water), THF (tetrahydrofuran). Reagent abbreviations used: HOBt (hydroxybenzotriazole),
2
EDC (1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide), DIEA (diisopropylethylamine), MgSO4
(magnesium sulfate), NaHCO (sodium bicarbonate), Na CO (sodium carbonate), NH Cl
3
2
3
4
(
ammonium chloride), NaCl (sodium chloride), K CO (potassium carbonate) and HCl
2 3
(
hydrogen chloride).
Phenyl(3S,4R)-3-(((tert-butyldimethylsilyl)oxy)methyl)-4-(4-fluorophenyl)piperidine-1-
carboxylate (6): To a 250 mL round bottom flask was added ((3S,4R)ꢀ4ꢀ(4ꢀfluorophenyl)ꢀ1ꢀ
methylpiperidinꢀ3ꢀyl)methanol (4.0 g, 17.91 mmol) and dichloromethane (40 mL). Tertꢀ
butyldimethylsilyl chloride (4.05 g, 26.9 mmol) and imidazole (1.22 g, 17.9 mmol) were added
to the reaction vessel producing a cloudy white mixture. Lastly, N,N-diisopropylethylamine (4.68
mL, 26.9 mmol) was added, giving a clear solution. The reaction was stirred overnight at room
temperature. Methylene chloride was used to dilute the reaction followed by washing with brine
(
2x). The organic layer was dried over MgSO , concentrated in vacuo, and purified using flash
4
chromatography with a 20%ꢀ40% EtOAc/Hexane gradient to give (3S,4R)ꢀ3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)ꢀ1ꢀmethylpiperidine (5.51 g, 16.3 mmol, 91%
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
yield). 1H NMR (500 MHz, Chloroformꢀd) δ 7.13 (dd, J = 8.5, 5.5 Hz, 2H), 6.96 (t, J = 8.7 Hz,
2H), 3.29 (dd, J = 10.3, 2.3 Hz, 1H), 3.15 (dd, J = 10.1, 6.0 Hz, 1H), 3.08 (dd, J = 7.6, 1.8 Hz,
1H), 2.92 (ddt, J = 11.2, 4.0, 2.4 Hz, 1H), 2.33 (s, 3H), 2.32 – 2.28 (m, 1H), 2.00 – 1.88 (m, 3H),
1.85 – 1.72 (m, 2H), 0.84 (s, 9H), ꢀ0.11 (d, J = 9.3 Hz, 6H). MS (ESI+) m/z: 338.2 (M+1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Phenyl (3S,4R)ꢀ3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidineꢀ1ꢀ
carboxylate (0.47 g, 1.40 mmol) and toluene (8 mL) were added to a 50mL flask and refluxed at
110ºC. Phenyl chloroformate (0.32 mL, 2.53 mmol) was then added dropwise and the reaction
was stirred another four hours at reflux. The reaction was then cooled to 60ºC and
trimethylamine (0.20 mL, 1.44 mmol) was added to quench the reaction which was stirred
another 40 minutes and then cooled to room temperature. The reaction was diluted with toluene
and brine. The layers were separated and the toluene layer was washed 2x with brine, dried over
MgSO , and concentrated. Purified the resulting clear viscous liquid with flash chromatography
4
(0 – 10% EtOAc/Hexanes) to give a clear syrup, phenyl (3S,4R)ꢀ3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidineꢀ1ꢀcarboxylate (0.467g, 75% yield).
Took forward as is without further purification.
tert-Butyl (3S,4R)-3-(((tert-butyldimethylsilyl)oxy)methyl)-4-(4-fluorophenyl)piperidine-1-
carboxylate (7). Compound 6 (6.03 g, 13.6 mmol) was added to a 250 mL round bottom flask
and dissolved in isopropanol (45 mL). Sodium hydroxide (8 N, 15mL) was then added and the
reaction was refluxed at 80ºC overnight. The reaction was diluted with water and toluene. The
layers were separated and the organic layer was washed 2x with brine, dried over MgSO , and
4
concentrated. Purified using flash chromatography (0ꢀ20% MeOH/DCM) to give as a clear oil
(3S,4R)ꢀ3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidine (2.41 g, 55%
1
yield). H NMR (DMSO – d , 400 MHz) δ: 7.24ꢀ7.19 (m, 2H), 7.10 (t, J=8 Hz, 2H), 3.21 (dd, J =
6
ACS Paragon Plus Environment

Page 19 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10.2, 3.0 Hz, 1H), 3.12 (td, J = 11.2, 10.2, 7.7 Hz, 2H), 2.98 – 2.92 (m, 1H), 2.40 (td, J = 11.4,
4.6 Hz, 1H), 2.34 (dd, J = 12.0, 10.7 Hz, 1H), 2.18 (s, 1H), 1.70 (tdt, J = 11.0, 7.2, 3.5 Hz, 1H),
1.60 – 1.47 (m, 2H), 0.8 (s, 9H), ꢀ0.13 (d, J = 6.7 Hz, 6H). MS (ESI+) m/z: 444.3 (M+1). (3S,
4R)ꢀ3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidine (1.05 g, 3.25 mmol),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
boc anhydride (0.99 g, 4.54 mmol), N,N-diisopropylethylamine (0.85 mL, 4.88 mmol) and THF
30 mL) were added to a 250 mL round bottom flask and stirred under nitrogen overnight.
(
Reaction was quenched with water and diluted with EtOAc. The organic layer was washed 2x
with brine, dried with MgSO , and concentrated to give a clear oil. The oil was further purified
4
using flash chromatography (10% EtOAc/Hexanes) to give, as a clear syrup, tertꢀbutyl (3S,4R)ꢀ
3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidineꢀ1ꢀcarboxylate (7, 1.35 g,
1
98% yield). H NMR (DMSO – d , 400 MHz) δ: 7.26ꢀ7.19 (m, 2H), 7.09 (t, J=8 Hz, 2H), 4.26
6
(d, J=8.0 Hz, 1H), 4.01 (br s, 1H), 3.46ꢀ3.02 (m, 4H), 2.69 (s, 1H), 1.81ꢀ1.43 (m, 3H), 1.39 (s,
9H), 0.80 (s, 9H), ꢀ0.13 (s, 6H); HPLC purity: 95%; MS (ESI+) m/z: 424.4 (M+1), 446.4
+
(
M+Na ).
5-((3S,4R)-1-(tert-Butoxycarbonyl)-3-(((tert-butyldimethylsilyl)oxy)methyl)piperidin-4-yl)-2-
fluorobenzoic acid (8a). N1,N1,N2,N2ꢀtetramethylethyleneꢀ1,2ꢀdiamine (0.549 g, 4.72 mmol)
was added to THF (10 mL) in a 50 mL round bottom flask cooled to ꢀ78 ⁰C under argon. Secꢀ
buytllithium (3.63 mL, 4.72 mmol) was added and the reaction mixture was stirred for thirty
minutes then (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀ
fluorophenyl)piperidineꢀ1ꢀcarboxylate (1.0 g, 2.36 mmol) in THF (10 mL) was added and the
solution was stirred at ꢀ78 ⁰C for 1 hour. Freshly broken CO was added and the mixture allowed
2
to warm to room temperature. The reaction was extracted with ether 2x, washed with NaCl 2x,
dried with MgSO and concentrated in vacuo. The resulting intermediate was used without
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
further purification.
(3S,4R)-tert-Butyl-3-(((tert-butyldimethylsilyl)oxy)methyl)-4-(4-fluoro-3-
(hydroxymethyl)phenyl)piperidine-1-carboxylate (8b). N1,N1,N2,N2ꢀtetramethylethyleneꢀ1,2ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
diamine (0.548 g, 4.72 mmol) was added to THF (10 mL) in a 50 mL round bottom flask cooled
to ꢀ78 ⁰C under nitrogen. Secꢀbuytllithium (3.63 mL, 4.72 mmol) was added and the reaction
mixture was stirred for ten minutes then (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluorophenyl)piperidineꢀ1ꢀcarboxylate (0.998 g, 2.36
mmol) in THF (10 mL) was added and the solution was stirred at ꢀ78 ⁰C for 1 hour. A solution of
anhydrous DMF (2 ml) in THF (5 mL) was added and the reaction was allowed to warm to room
temperature (4 hours). The reaction was quenched with NH Cl (10 mL), extracted with ether 2x,
4
washed with NaCl 3x, dried over MgSO , concentrated in vacuo, and purified using flash
4
chromatography with a 1%ꢀ10% EtOAc/Hexane gradient to give (3S,4R)ꢀtertꢀbutylꢀ3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀformylꢀphenyl)piperidineꢀ1ꢀcarboxylate (1.075g
1
crude, took all crude to next step, R :0.46 (10% EtOAc/Hexanes)). H NMR (DMSO – d , 400
f
6
MHz) δ: 10.20 (s, 1H), 7.71ꢀ7.26 (m, 2H), 7.35 (t, J=8.0 Hz, 1H), 4.27 (d, J=12.0 Hz, 1H), 4.05
br s, 1H), 3.55ꢀ3.10 (m, 4H), 2.63ꢀ2.55 (m, 1H), 1.83ꢀ1.52 (m, 3H), 1.57 (s, 9H), 0.82 (s, 9H), ꢀ
(
0.12 (s, 6H); HPLC purity: 95%; MS (ESI+) m/z: 452.4 (M+1). (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀformylphenyl)ꢀpiperidineꢀ1ꢀcarboxylate (crude,
1.075 g, 2.38 mmol) was added to THF (15 mL) in a 50 mL round bottom flask cooled to 0 ⁰C.
Sodium borohydride (0.0262 g, 1.19 mmol) was added and the reaction mixture was allowed to
come to room temperature and stirred 3 hours. Methanol (10 mL) and citric acid (5%, 10 mL)
were added to the reaction mixture and allowed to stir 30 minutes. The reaction was extracted
with ether/ethyl acetate 2x, washed with NaCl 2x, dried over MgSO , concentrated in vacuo, and
4
ACS Paragon Plus Environment

Page 21 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
purified using flash chromatography (80 mL silica gel) of 20% EtOAc/Hexane to give (3S,4R)ꢀ
tertꢀbutyl 3ꢀ(((tertꢀbutyldiꢀ methylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀ
(
hydroxymethyl)phenyl)piperidineꢀ1ꢀcarboxylate (0.627 g, R :0.22 (20% EtOAc/Hexanes), 59%
f
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
yield from 2). H NMR (DMSO – d , 400 MHz) δ: 7.30 (d, J=8.0 Hz, 1H) 7.15ꢀ7.08 (m, 1H)
6
7.05 (t, J=8.0 Hz, 1H), 5.24 (s, 1H), 4.52 (s, 2H), 4.30 (d, J=12.0 Hz, 1H), 4.05 (br s, 1H), 3.35
(
s, 1H), 3.24 (s, 1H), 3.14 (t, J=8.0 Hz, 1H), 2.72 (s, 2H), 1.82ꢀ1.46 (m, 3H), 1.42 (s, 9H), 0.83
+
(
s, 9H), 0.09 (s, 6H); HPLC purity: 95%; MS (ESI+) m/z: 454.4 (M+1), 476.4 (M+Na ).
(
3S,4R)-tert-Butyl 3-(((tert-butyldimethylsilyl)oxy)methyl)-4-(4-fluoro-3-
(methoxycarbonyl)phenyl) piperidine-1-carboxylate (9a). To a solution of 5ꢀ((3S,4R)ꢀ1ꢀ(tertꢀ
butoxycarbonyl)ꢀ3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)piperidinꢀ4ꢀyl)ꢀ2ꢀfluorobenzoic acid
(
1.08 g, 2.31 mmol) in 20% (V/V) Methanol/toluene (30 mL) was added drop wise TMSꢀ
diazomethane (1.27 mL, 2.54 mmol). Gas evolved and reaction is slightly yellow. The reaction
was stirred for 30 minutes and was then quenched with acetic acid until gas no longer evolved
and the yellow color was gone. Then the reaction was concentrated in vacuo. Dissolved in ethyl
acetate and ether and washed 2x with brine, dried with MgSO and concentrated. The crude
4
product was purified using flash chromatography with a 20%ꢀ30% EtOAc/Hexane gradient to
give (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀ
(
methoxycarbonyl)phenyl) piperidineꢀ1ꢀcarboxylate (0.96 g crude, 1.99 mmol, 86% yield over
1
two steps). H NMR (400 MHz, DMSOꢀd ) δ 7.69 (dd, J = 7.2, 2.3 Hz, 1H), 7.56 (dq, J = 7.2,
6
2.3 Hz, 1H), 7.29 (dd, J = 10.8, 8.5 Hz, 1H), 4.25 (d, J = 13.0 Hz, 1H), 4.03 (d, J = 8.2 Hz, 1H),
3
.84 (s, 3H), 3.22 (s, 1H), 3.14 (dd, J = 10.5, 7.4 Hz, 1H), 2.72 (s, 1H), 2.58 (s, 2H), 1.76 (s, 1H),
.67 (d, J = 12.5 Hz, 1H), 1.63 – 1.48 (m, 1H), 1.41 (s, 9H), 0.82 (s, 9H), ꢀ0.11 (d, J = 4.4 Hz,
1
6H). MS (ESI+) m/z: 482.3 (M+1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
3S,4R)-tert-Butyl-3-(((tert-butyldimethylsilyl)oxy)methyl)-4-(3-(cyanomethyl)-4-
fluorophenyl)piperidine-1-carboxylate (9b). (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀ(hydroxymethyl)phenyl)piperidineꢀ1ꢀcarboxylate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(0.344 g, 0.759 mmol) was added to anhydrous dichloromethane (8 mL) in a 50 mL round
bottom flask cooled to 0 ⁰C under nitrogen. Diisopropylethylamine (0.265 mL, 1.517 mmol) was
added to the reaction mixture followed by methanesulfonic anhydride and the reaction mixture
was stirred for 0.5 hours. The reaction mixture was washed 1x with distilled water, dried over
MgSO , and concentrated in vacuo. Anhydrous dimethyl sulfoxide (8 mL) was added to the 50
4
mL flask followed by sodium cyanide (0.112 g, 2.276 mmol) and stirred overnight at room
temperature under nitrogen. The reaction was diluted with ether and washed with NaCl (6x). The
aqueous layers were then washed 2x with ether. The organic layers were combined, dried over
MgSO , concentrated in vacuo, and purified using flash chromatography (30 mL silica gel) of
4
10ꢀ20% gradient of EtOAc/Hexane to give (3S,4R)ꢀtertꢀbutylꢀ3ꢀ(((tertꢀ
butyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(3ꢀ(cyanomethyl)ꢀ4ꢀfluorophenyl)piperidineꢀ1ꢀcarboxylate
1
(0.210g, R :0.51 (20% EtOAc/Hexanes), 60% yield). H NMR (CDCl , 400 MHz) δ: 7.25ꢀ7.18
f
3
(m, 1H) 7.13 (s, 1H) 7.03 (t, J=9.0 Hz, 1H), 4.37ꢀ4.16 (m, 2H), 3.74 (s, 2H), 3.37ꢀ3.23 (m, 1H),
3
.19ꢀ3.08 (m, 1H), 2.68 (br s, 2H), 1.82ꢀ1.56 (m, 4H), 1.48 (s, 9H), 1.24 (s, 9H), ꢀ0.07 (s, 6H);
+
HPLC purity: 99%; MS (ESI+) m/z: 463.1 (M+1), 485.1 (M+Na ).
(
3S,4R)-tert-Butyl 4-(4-fluoro-3-(methoxycarbonyl)phenyl)-3-(hydroxymethyl)piperidine-1-
carboxylate (10a). To a solution of (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ
4ꢀfluoroꢀ3ꢀ(methoxycarbonyl)phenyl) piperidineꢀ1ꢀcarboxylate (1.68 g, 3.49 mmol) in THF (50
(
mL) was added acetic acid (0.60 mL, 10.46 mmol) followed by tetrabutylammonium fluoride
o
(1.0 M, 10.46 mL, 10.46 mmol). The resulting mixture was stirred overnight at 60 C. Diluted
ACS Paragon Plus Environment

Page 23 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with ethyl acetate/ether and treated with saturated aq. NH Cl. The layers were separated and the
4
organic layer was washed with NaCl (2x), dried with MgSO , and concentrated in vacuo.
4
Purified using flash chromatography with a 10%ꢀ30% EtOAc/Hexane gradient to give (3S,4R)ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tertꢀbutyl 4ꢀ(4ꢀfluoroꢀ3ꢀ(methoxycarbonyl)phenyl)ꢀ3ꢀ(hydroxymethyl)piperidineꢀ1ꢀcarboxylate
1
as a clear oil (1.29 g, 3.51 mmol, 100% yield). H NMR (400 MHz, Chloroformꢀd) δ 7.75 (dd, J
=
=
6.9, 2.4 Hz, 1H), 7.34 (ddd, J = 8.4, 4.5, 2.4 Hz, 1H), 7.08 (dd, J = 10.5, 8.5 Hz, 1H), 4.36 (d, J
13.0 Hz, 1H), 4.21 (s, 1H), 3.92 (s, 3H), 3.43 (dt, J = 11.1, 3.8 Hz, 1H), 3.25 (dt, J = 11.3, 6.0
Hz, 1H), 2.85 – 2.68 (m, 2H), 2.61 (s, 1H), 1.83 (dq, J = 7.1, 3.9, 3.5 Hz, 1H), 1.77 (d, J = 13.1
Hz, 1H), 1.67 (qd, J = 12.3, 4.2 Hz, 1H), 1.48 (s, 9H), 1.20 (s, 1H). MS (ESI+) m/z: 368.2
(M+1), 390.2 (M+Na+).
(3S,4R)-tert-Butyl-4-(3-(cyanomethyl)-4-fluorophenyl)-3-(hydroxymethyl)piperidine-1-
carboxylate (10b). (3S,4R)ꢀtertꢀbutyl 3ꢀ(((tertꢀbutyldimethylsilyl)oxy)methyl)ꢀ4ꢀ(3ꢀ
cyanomethyl)ꢀ4ꢀfluoropheny l)piperidineꢀ1ꢀcarboxylate (0.501 g, 1.081 mmol) and tetraꢀnꢀbutyl
(
ammonium fluoride (1.30mL, 1.30 mmol) were added to THF (10 mL) in a 50 mL round bottom
flask at room temperature and allowed to stir overnight. The next day another equivalent of tetraꢀ
nꢀbutylammonium fluoride (1.30 mL, 1.30 mmol) was added to the reaction mixture and stirred
for 3.0 hours. The reaction mixture was quenched with NH Cl, extracted with ether/ethyl acetate
4
1x and washed with NaCl (2x). The organic layers were combined, dried over MgSO4,
concentrated in vacuo, and purified using flash chromatography (25mL silica gel) of 30ꢀ50%
gradient of EtOAc/Hexane to give (3S,4R)ꢀtertꢀbutylꢀ4ꢀ(3ꢀ(cyanomethyl)ꢀ4ꢀfluorophenyl)ꢀ3ꢀ
(
hydroxymethyl) piperidine ꢀ1ꢀcarboxylate (0.240g, R :0.33 (30% EtOAc/Hexanes), 64% yield).
f
1
H NMR (CDCl , 400 MHz) δ: 7.24 (dd, J =2.4 Hz, J =7.2 Hz, 1H) 7.18ꢀ7.12 (m, 1H) 7.03 (t,
3
1
2
J=9.0 Hz, 1H), 4.41ꢀ4.32 (m, 2H), 4.19 (br s, 1H), 3.74 (s, 1H), 3.42 (dd, J =3.0 Hz, J =11.0 Hz,
1
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
H), 3.24 (q, J=6.4 Hz, 1H), 2.87ꢀ2.52 (m, 3H), 1.87ꢀ1.57 (m, 4H), 1.48 (s, 9H) ; HPLC purity:
+
89%; MS (ESI+) m/z: 349.3 (M+1), 371.3 (M+Na ).
(3S,4R)-tert-Butyl 3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(4-fluoro-3-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
(methoxycarbonyl)phenyl) piperidine-1-carboxylate (11a). To a 0 C solution of (3S,4R)ꢀtertꢀ
butyl 4ꢀ(4ꢀfluoroꢀ3ꢀ(methoxycarbonyl) phenyl)ꢀ3ꢀ(hydroxymethyl)piperidineꢀ1ꢀcarboxylate
(0.687 g, 1.87 mmol) in 25 mL DCM was added diisopropylethylamine (0.980 mL, 5.61 mmol)
followed by methanesulfonylchloride (0.651 g, 3.74 mmol). The reaction was allowed to warm
to room temperature and stirred overnight. The resulting product was washed with water (1x),
then brine (1x), dried over MgSO , and concentrated in vacuo to give a clear oil (0.833 g crude).
4
o
The product was used in subsequent reactions without further purification. To a 0 C solution of
benzo[d][1,3]dioxolꢀ5ꢀol (0.543 g, 3.93 mmol) in 5 mL DMF was added 60% sodium hydride in
mineral oil (0.164g, 4.11 mmol). The solution turned light pink and was stirred for five minutes.
To the reaction was added (3S,4R)ꢀtertꢀbutyl 4ꢀ(4ꢀfluoroꢀ3ꢀ(methoxycarbonyl)phenyl)ꢀ3ꢀ
(((methylsulfonyl)oxy)methyl)piperidineꢀ1ꢀcarboxylate (0.833 g, 1.87 mmol) in 5 mL DMF. The
o
reaction was heated to 70 C for 1 hour then cooled to room temperature and treated with sat.
NH Cl solution. Extracted with ethyl acetate/ether (2x), washed with brine, dried with MgSO ,
4
4
and concentrated in vacuo. The crude product was purified by flash chromatography using a
gradient of 5% MeOH/DCM to give as a clear oil (3S,4R)ꢀtertꢀbutyl 3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀ
yloxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀ(methoxycarbonyl)phenyl)piperidineꢀ1ꢀcarboxylate (0.452 g, 0.927
1
mmol, 50% yield over two steps). H NMR (400 MHz, Methanolꢀd ) δ 7.76 (dd, J = 6.9, 2.4 Hz,
4
1
H), 7.49 (ddd, J = 8.5, 4.6, 2.5 Hz, 1H), 7.14 (dd, J = 10.7, 8.5 Hz, 1H), 6.61 (d, J = 8.5 Hz,
H), 6.35 (d, J = 2.5 Hz, 1H), 6.16 (dd, J = 8.5, 2.5 Hz, 1H), 5.85 (q, J = 1.2 Hz, 2H), 4.40 (d, J
1
= 13.3 Hz, 1H), 4.22 (d, J = 3.3 Hz, 1H), 3.87 (s, 3H), 3.61 (dd, J = 9.8, 2.8 Hz, 1H), 3.50 (dd, J
ACS Paragon Plus Environment

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
=
–
9.9, 6.6 Hz, 1H), 2.82 (td, J = 11.7, 3.8 Hz, 3H), 2.11 – 1.97 (m, 1H), 1.86 – 1.77 (m, 1H), 1.76
1.64 (m, 1H), 1.49 (s, 9H). MS (ESI+) m/z: 488.2 (M+1).
(3S,4R)-tert-Butyl 3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(3-(cyanomethyl)-4-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
fluorophenyl)piperidine-1-carboxylate (11b). (3S,4R)ꢀtertꢀbutylꢀ4ꢀ(3ꢀ(cyanomethyl)ꢀ4ꢀ
fluorophenyl)ꢀ3ꢀ(hydroxymethyl)piperidineꢀ1ꢀcarboxylate was added to a dry 25 mL round
o
bottom flask, with 5 mL anhydrous dichloromethane cooled to 0 C followed by DIEA (0.346
mL, 1.979 mmol) and lastly, methanesulfonic anhydride (0.230 g, 1.319 mmol) and stirred for 2
hours. The reaction mixture was washed 1x with NaCl, dried of MgSO , and concentrated in
4
vacuo. To a separate 25 mL round bottom flask sodium hydride (0.058 g, 1.451 mmol) was
o
added with DMF under nitrogen at 0 C followed by sesamol (0.1913 g, 1.385 mmol) in 2 mL
DMF and stirred for 20 minutes. The dried, mesylated (3S,4R)ꢀtertꢀbutylꢀ4ꢀ(3ꢀ(cyanomethyl)ꢀ4ꢀ
fluorophenyl)ꢀ3ꢀ(hydroxymethyl)piperidineꢀ1ꢀcarboxylate was then added with 2 mL DMF to
o
the reaction vessel with the activated sesamol, heated to 70 C, and stirred for 10 minutes. The
reaction was quenched with NH Cl, extracted with EtOAc/ether 2x, washed with brine 3x, dried
4
over MgSO and purified using flash chromatography with a 0% to 80% gradient of
4
EtOAc/Hexanes to give (3S,4R)ꢀtertꢀbutyl 3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ4ꢀ(3ꢀ
(cyanomethyl)ꢀ4ꢀfluorophenyl) piperidineꢀ1ꢀcarboxylate (0.138 g, Rf: 0.40 (40%
1
EtOAc/Hexanes), 46% yield). H NMR (CDCl , 400 MHz) δ: 7.25ꢀ7.21 (m, 1H), 7.16ꢀ7.12 (m,
3
1
H), 7.02 (t, J=8.8 Hz, 1H), 6.63 (d, J=8.8, 1H), 6.35 (d, J=2.4, 1H), 6.14 (dd, J =2.4 Hz, J =8.4
1 2
Hz, 1H), 5.88 (s, 2H), 4.43 (broad s, 1H), 4.25 (broad s, 1H), 3.72 (s, 2H), 3.62ꢀ3.59 (m, 1H),
3.46ꢀ3.42 (m, 1H), 2.90ꢀ2.67 (m, 3H), 2.10ꢀ1.60 (m, 3H), 1.50 (s, 9H) ; HPLC purity: 98%; MS
+
(
ESI+) m/z: 469.1 (M+1), 491.1 (M+Na ).
5-((3S,4R)-3-((Benzo[d][1,3]dioxol-5-yloxy)methyl)-1-(tert-butoxycarbonyl)piperidin-4-yl)-2-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
fluorobenzoic acid (12a). To a round bottom flask equipped with a stir bar was added (3S,4R)ꢀ
tertꢀbutyl 3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ4ꢀ(4ꢀfluoroꢀ3ꢀ
(methoxycarbonyl)phenyl)piperidineꢀ1ꢀcarboxylate (0.297g, 0.608 mmol), 1M NaOH (1.83 mL,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
.825 mmol), H O (5 mL), and Methanol (12 mL). The reaction was stirred overnight at room
2
temperature. Ether was added to the reaction and the resulting layers were separated. To the
aqueous layer, 10% citric acid was added to give a pH of 4. The aqueous layer was then
extracted 2x with ethyl acetate. The ethyl acetate layers were then combined and washed 1x with
NaCl, dried with MgSO , and concentrated to give, with no further purification, as an amorphous
4
solid 5ꢀ((3S,4R)ꢀ3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ1ꢀ(tertꢀbutoxycarbonyl)piperidinꢀ4ꢀ
1
yl)ꢀ2ꢀfluorobenzoic acid (0.226g, 0.477 mmol, 79% yield). H NMR (400 MHz, Methanolꢀd ) δ
4
7.77 (dd, J = 7.0, 2.4 Hz, 1H), 7.47 (ddd, J = 8.6, 4.5, 2.5 Hz, 1H), 7.13 (dd, J = 10.7, 8.5 Hz,
1
H), 6.61 (d, J = 8.5 Hz, 1H), 6.36 (d, J = 2.4 Hz, 1H), 6.16 (dd, J = 8.5, 2.5 Hz, 1H), 5.84 (s,
H), 4.41 (d, J = 13.4 Hz, 1H), 4.20 (d, J = 13.3 Hz, 1H), 3.61 (dd, J = 9.8, 3.0 Hz, 1H), 3.52
2
(
dd, J = 9.9, 6.7 Hz, 1H), 2.87 (s, 1H), 2.81 (td, J = 11.7, 3.9 Hz, 2H), 2.05 (ddd, J = 15.0, 7.1,
.0 Hz, 1H), 1.81 (dd, J = 13.5, 3.7 Hz, 1H), 1.70 (qd, J = 12.7, 4.6 Hz, 1H), 1.49 (s, 9H). MS
(ESI+) m/z: 474.2 (M+1).
4
2-(5-((3S,4R)-3-((Benzo[d][1,3]dioxol-5-yloxy)methyl)-1-(tert-butoxycarbonyl) piperidin-4-yl)-
2-fluorophenyl)acetic acid (12b). (3S,4R)ꢀtertꢀbutyl 3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ4ꢀ
(3ꢀ(cyanomethyl)ꢀ4ꢀfluorophenyl) piperidineꢀ1ꢀcarboxylate (0.459 g) in EtOH (3.0 mL) was
added to a 25 mL pressure vessel followed by 50% NaOH (0.26 mL) and heated to 98 ⁰C
overnight. The following day further excess 50% NaOH (0.10 mL) was added and the reaction
was again stirred overnight. The reaction mixture was cooled to room temperature and diluted
with H O then 2N HCl was added to reach pH 2.0. The resulting suspension was extracted with
2
ACS Paragon Plus Environment

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dichloromethane 3x, washed with NaCl 2x, dried with MgSO , and concentrated in vacuo to give
4
2
2
4
ꢀ(5ꢀ((3S,4R)ꢀ3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ1ꢀ(tertꢀbutoxycarbonyl)piperidinꢀ4ꢀyl)ꢀ
1
ꢀfluorophenyl)acetic acid without further purification (0.372 g, 78% yield). H NMR (CDCl ,
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
00 MHz) δ: 7.11ꢀ7.04 (m, 2H), 7.00 (t, J=9.0 Hz, 1H), 6.20 (d, J=8.8, 1H), 6.34 (2, 1H), 6.14 (d,
J=8.4 Hz, 1H), 5.88 (s, 2H), 4.42 (broad s, 1H), 4.23 (broad s, 1H), 3.66 (s, 2H), 3.64ꢀ3.58 (m,
1H), 3.48ꢀ3.42 (m, 1H), 2.88ꢀ2.72 (m, 2H), 2.71ꢀ2.60 (m, 1H), 2.04ꢀ1.92 (m, 1H), 1.85ꢀ1.63 (m,
+
3
H), 1.49 (s, 9H) ; HPLC purity: 87%; MS (ESI+) m/z: 488.2 (M+1), 410.2 (M+Na ); (ESIꢀ)
m/z: 486.2 (Mꢀ1).
tert-Butyl (3S,4R)-3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(4-fluoro-3-
(methylcarbamoyl)phenyl)piperidine-1-carboxylate. (13aa). 5ꢀ((3S,4R)ꢀ3ꢀ((benzo[d][1,3] dioxolꢀ
5ꢀyloxy)methyl)ꢀ1ꢀ(tertꢀbutoxycarbonyl)piperidinꢀ4ꢀyl)ꢀ2ꢀfluorobenzoic acid (0.08 g, 0.169
mmol), 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide (0.073 g, 0.338 mmol), N,Nꢀ
diisopropylethylamine (0.131 mL, 0.676 mmol), hydroxybenzotriazole (0.051g, 0.338 mmol),
and 2M methylamine in THF (0.169 mL, 0.338 mmol) were added to THF (5 mL) in a 15 mL
round bottom flask and stirred overnight at room temperature. The resulting solution was diluted
with ethyl acetate and saturated sodium bicarbonate, and then the layers were separated. The
organic layer was then washed with brine (2x), dried with MgSO , and concentrated in vacuo.
4
The crude product was purified using flash chromatography using 30% ꢀ 90% EtOAc/Hexanes to
give as an amorphous solid tertꢀbutyl (3S,4R)ꢀ3ꢀ((benzo[d][1,3]dioxolꢀ5ꢀyloxy)methyl)ꢀ4ꢀ(4ꢀ
fluoroꢀ3ꢀ(methylcarbamoyl)phenyl)piperidineꢀ1ꢀcarboxylate (0.070 g, 0.144 mmol, 85% yield).
1
H NMR (400 MHz, Chloroformꢀd) δ 7.94 (dd, J = 7.6, 2.4 Hz, 1H), 7.30 – 7.22 (m, 1H), 7.04
(dd, J = 11.8, 8.4 Hz, 1H), 6.78 – 6.65 (m, 1H), 6.62 (d, J = 8.4 Hz, 1H), 6.33 (t, J = 1.8 Hz, 1H),
6.12 (dt, J = 8.6, 1.8 Hz, 1H), 5.88 (d, J = 1.3 Hz, 2H), 4.44 (s, 1H), 4.24 (s, 2H), 3.58 (dd, J =
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
9
.4, 2.8 Hz, 1H), 3.44 (dd, J = 9.4, 6.7 Hz, 1H), 3.03 (d, J = 4.0 Hz, 3H), 2.76 (q, J = 12.5, 11.8
Hz, 3H), 2.15 – 2.05 (m, 1H), 1.74 (dq, J = 17.3, 13.5, 13.0 Hz, 2H), 1.50 (d, J = 1.3 Hz, 9H).
MS (ESI+) m/z: 486.8 (M+1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
tert-Butyl (3S,4R)-3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(3-(benzylcarbamoyl)-4-
fluorophenyl)piperidine-1-carboxylate (13ab). Compound 13ab was synthesized as described for
intermediate 13aa from intermediate 12a replacing methylamine with benzylamine (59% yield).
1
H NMR (500 MHz, Chloroformꢀd) δ 7.97 (dd, J = 7.5, 2.5 Hz, 1H), 7.38 – 7.34 (m, 3H), 7.29
(ddt, J = 8.7, 6.4, 3.4 Hz, 2H), 7.03 (ddd, J = 14.9, 10.0, 4.2 Hz, 2H), 6.62 (d, J = 8.4 Hz, 1H),
6
.34 (d, J = 2.5 Hz, 1H), 6.13 (dd, J = 8.5, 2.5 Hz, 1H), 5.87 (s, 2H), 4.68 (d, J = 5.6 Hz, 2H),
.45 (s, 1H), 4.25 (s, 1H), 3.59 (dd, J = 9.4, 2.8 Hz, 1H), 3.45 (dd, J = 9.5, 6.5 Hz, 1H), 2.77 (dt,
4
J = 23.6, 11.9 Hz, 3H), 2.10 (ddq, J = 11.5, 7.7, 4.0 Hz, 1H), 1.80 (d, J = 13.2 Hz, 1H), 1.76 –
.68 (m, 1H), 1.50 (s, 9H). MS (ESI+) m/z: 562.7 (M+1).
tert-Butyl (3S,4R)-3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(4-fluoro-3-
phenethylcarbamoyl)phenyl)piperidine-1-carboxylate (13ac). Compound 13ac was synthesized
1
(
as described for intermediate 13aa from intermediate 12a replacing methylamine with 2ꢀ
1
phenylethylamine (77% yield). H NMR (500 MHz, Chloroformꢀd) δ 7.93 (dd, J = 7.5, 2.4 Hz,
1
H), 7.37 – 7.30 (m, 2H), 7.29 – 7.23 (m, 4H), 7.01 (dd, J = 11.6, 8.4 Hz, 1H), 6.80 – 6.71 (m,
H), 6.63 (d, J = 8.4 Hz, 1H), 6.35 (d, J = 2.5 Hz, 1H), 6.14 (dd, J = 8.5, 2.5 Hz, 1H), 5.88 (s,
1
2H), 4.46 (s, 1H), 4.25 (s, 1H), 3.74 (q, J = 5.9 Hz, 2H), 3.59 (dd, J = 9.5, 2.9 Hz, 1H), 3.45 (dd,
J = 9.5, 6.6 Hz, 1H), 2.94 (t, J = 7.0 Hz, 2H), 2.87 – 2.69 (m, 3H), 2.10 (tdt, J = 10.4, 7.0, 3.0
Hz, 1H), 1.83 – 1.67 (m, 2H), 1.51 (s, 9H). MS (ESI+) m/z: 576.8 (M+1).
(3S,4R)-tert-Butyl -3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(3-((2,6-
difluorobenzyl)carbamoyl)-4-fluorophenyl)piperidine-1-carboxylate. (13ad) Compound 13ad
ACS Paragon Plus Environment

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was synthesized as described for intermediate 13aa from intermediate 12a replacing
1
methylamine with 2,6ꢀdifluorobenzylamine (24% yield). H NMR (CDCl , 400 MHz) δ 7.95 (dd,
3
J = 7.5, 2.4 Hz, 1H), 7.35 – 7.20 (m, 2H), 7.16ꢀ7.13 (m, 1H), 7.01 (dd, J = 11.8, 8.4 Hz, 1H),
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
.94 – 6.75 (m, 2H), 6.60 (d, J = 8.5 Hz, 1H), 6.32 (d, J = 2.4 Hz, 1H), 6.11 (dd, J = 8.5, 2.5 Hz,
1H), 5.87 (s, 2H), 4.46 (br s, 1H), 4.19 (br s, 1H), 3.56 (dd, J = 9.5, 2.8 Hz, 1H), 3.42 (dd, J =
.4, 6.6 Hz, 1H), 2.80 – 2.69 (m, 3H), 2.07 (m, 2H), 1.74ꢀ 1.69 (m, 3H), 1.49 (s, 9H). MS (ESI+)
9
m/z: 598.9 (M+1).
(3S,4R)-tert-Butyl -3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(3-((2,6-
difluorophenethyl)carbamoyl)-4-fluorophenyl)piperidine-1-carboxylate (13ae). Compound 13ae
was synthesized as described for intermediate 13aa from intermediate 12a replacing
1
methylamine with 2ꢀ(2,6ꢀdifluorophenyl)ethanꢀ1ꢀamine (87% yield). H NMR (CDCl , 400
3
MHz) δ 7.90 (dd, J = 7.5, 2.4 Hz, 1H), 7.24 – 7.07 (m, 2H), 7.00 (dd, J = 11.6, 8.4 Hz, 1H), 6.95
–
6.65 (m, 4H), 6.60 (d, J = 8.5 Hz, 1H), 6.32 (d, J = 2.4 Hz, 1H), 6.11 (dd, J = 8.5, 2.5 Hz, 1H),
.86 (s, 2H), 4.41 (br s, 1H), 4.22 (br s, 1H), 3.70 (q, J = 6.9 Hz, 1H), 3.56 (m, 1H), 3.49 – 3.28
5
(m, 1H), 3.02 (t, J = 6.8 Hz, 2H), 2.79 – 2.68 (m, 2H), 2.10ꢀ2.04 (m, 1H), 1.87 – 1.65 (m, 2H),
1.48 (s, 9H). MS (ESI+) m/z: 612.0 (M+1).
tert-Butyl (3S,4R)-3-((benzo[d][1,3]dioxol-5-yloxy)methyl)-4-(4-fluoro-3-((2-
(trifluoromethyl)benzyl)carbamoyl)phenyl)piperidine-1-carboxylate (13af). Compound 13af was
synthesized as described for intermediate 13aa from intermediate 12a replacing methylamine
1
with (2ꢀ(trifluoromethyl)phenyl)methanamine (28% yield). Yellow oil. H NMR (400 MHz,
Chloroformꢀd) δ 7.91 (m, 1H), 7.65 (t, J = 6.2 Hz, 2H), 7.52 (t, J = 7.6 Hz, 1H), 7.38 (t, J = 7.7
Hz, 1H), 7.25ꢀ7.23 (m, 3H), 7.12 – 6.92 (m, 2H), 6.20 (br s, 1H), 5.87 (m, 2H), 5.09ꢀ4.96 (m,
ACS Paragon Plus Environment