Discovery of a Small Side Cavity in Sphingosine Kinase 2 that Enhances Inhibitor Potency and Selectivity

Article abstract of DOI:10.1021/acs.jmedchem.9b01508

The sphingosine-1-phosphate (S1P) signaling pathway is an attractive drug target due to its involvement in immune cell chemotaxis and vascular integrity. The formation of S1P is catalyzed by sphingosine kinase 1 or 2 (SphK1 or SphK2) from sphingosine (Sph

Full text of DOI:10.1021/acs.jmedchem.9b01508

Subscriber access provided by BIU Pharmacie | Faculté de Pharmacie, Université Paris V
Article
Discovery of a Small Side Cavity in Sphingosine Kinase
that Enhances Inhibitor Potency and Selectivity
2
Christopher D. Sibley, Emily A Morris, Yugesh Kharel, Anne M.
Brown, David R. Bevan, Kevin R Lynch, and Webster L Santos
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01508 • Publication Date (Web): 02 Jan 2020
Downloaded from pubs.acs.org on January 2, 2020
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 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 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.
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
Discovery of a Small Side Cavity in Sphingosine
Kinase 2 that Enhances Inhibitor Potency and
Selectivity
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
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
Christopher D. Sibley, Emily A. Morris, Yugesh Kharel, Anne M. Brown, David R. Bevan, §,ƒ
†
†
‡
§, ƒ
‡
†,ƒ,*
Kevin R. Lynch, and Webster L. Santos
†
Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061, United States
§
Department of Biochemistry, Virginia Tech, Blacksburg, Virginia 24061, United States
ƒ
Virginia Tech Center for Drug Discovery, Virginia Tech, Blacksburg, Virginia 24061, United
States
‡
Department of Pharmacology, University of Virginia, Charlottesville, Virginia 22908, United
States
Keywords: Sphingosine Kinase, SphK2, Sphingosine, Sphingosine 1-Phosphate, S1P, Inhibitor,
SAR, Molecular Docking
ABSTRACT: The sphingosine-1-phosphate (S1P) signaling pathway is an attractive drug target
due to its involvement in immune cell chemotaxis and vascular integrity. The formation of S1P is
catalyzed by sphingosine kinase 1 or 2 (SphK1 or SphK2) from sphingosine (Sph) and ATP.
Inhibition of SphK1 and 2 to attenuate levels of S1P has been reported to be efficacious in animal
1
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
models of diseases such as cancer, sickle cell disease and renal fibrosis. While inhibitors of both
SphKs have been reported, improvements in potency and selectivity are still needed. Towards that
end, we performed a structure-activity relationship profiling of 8 (SLM6031434) and discovered
a heretofore unrecognized side cavity that increased inhibitor potency toward SphK2. Interrogating
this region revealed that relatively small hydrophobic moieties are preferred with 10 being the
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
most potent SphK2 selective inhibitor (K
vivo activity.
i
= 89 nM, 73-fold SphK2 selective) with validated in
INTRODUCTION
Over the last two decades, sphingolipid signaling has emerged as an attractive candidate for drug
development due to its involvement in a wide range of cellular processes. One sphingolipid in
particular, sphingosine 1-phosphate (S1P), has garnered attention due to its implication in
pathologic processes including cancer,1–4 sickle cell disease,5,6 renal fibrosis,7–9 and
autoimmunity.10,11 S1P circulates at low micromolar concentrations in the blood of mammals and
perhaps all vertebrates.12 S1P is produced solely via ATP-dependent phosphorylation of
sphingosine (Sph) by sphingosine kinases (SphK1 and SphK2) (Figure 1). The biologic roles of
both enzymes are, to a certain extent, redundant in that mice lacking either isoform are viable,
fertile and phenotypically unremarkable. However, elimination of both SphK1 and 2 is
embryonically lethal at about day E13.5.13,14 While both isoforms catalyze the conversion of Sph
to S1P, the two enzymes differ in cellular localization and degree of substrate selectivity. SphK1
is predominantly located in the cytosol while SphK2 is found in the nucleus, mitochondria, and
endoplasmic reticulum. The association of S1P with a wide array of cellular processes and diseases
is, in part, due to its acting as both an intra- and extra-cellular signaling molecule. Extracellularly,
S1P is a high affinity ligand for five G-protein coupled receptors (S1P1-5) resulting in a downstream
2
ACS Paragon Plus Environment

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
signaling cascade leading to cellular migration and survival.3,15,16 In contrast, intracellular S1P is
reported to interact with pathways influencing cellular functions ranging from cell arrest and
1
7
apoptosis to proliferation and survival. In one such example, Laviad et al. reported findings of
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
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
intracellular S1P acting as a noncompetitive inhibitor towards the pro-apoptotic protein ceramide
1
8
synthase 2 (CerS2), thus conferring pro-survival effects. Conversely, studies reported by Chipuk
et al. demonstrated that mitochondrial S1P can interact with Bcl-2 homologous antagonist killer
(BAK) protein, resulting in an increase in mitochondrial membrane permeabilization. Subsequent
elevated permeabilization leads to the release of the pro-apoptotic messenger cytochrome c into
19
the cytosol inducing apoptosis. From a pharmaceutical viewpoint, S1P signaling and synthesis
represent interesting targets to manipulate and evoke a broad range of cellular responses.
Figure 1. The Sph/S1P equilibrium. Interconversion of Sph to S1P is catalyzed by SphK while the
reverse is achieved by S1P phosphatase.
Recently, our laboratories as well as Pyne et al. disclosed findings on topographical differences
between the two isoform ligand binding pockets.20,21 The studies indicate that compared to SphK1,
the Sph binding pocket of SphK2 extends deeper into the enzyme while the mid-section is narrower
due to a isoleucine to valine substitution. Such differences afford the potential for ligand
selectivity. A curious difference between the two isoforms is the change of blood S1P levels in
response to isoform inhibition or genetic deficiency. Specifically, inhibition of SphK1 in mice or
rats reduces blood S1P levels by >50%, while inhibition of SphK2 increases blood S1P
concentrations by roughly 3-fold.22–24 The rise in S1P in response to SphK2 deficiency has been
3
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
ascribed to a cascade wherein Sph, which is formed at the hepatocyte surface by dephosphorylation
of blood S1P, is captured by SphK2-mediated phosphorylation (Kharel et al., submitted for
publication). Thus isoform-selective inhibition provides a method to either raise or lower blood
S1P levels.
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
Since 2010, the S1P pathway has become a validated, druggable target with the FDA approval
of compound 1 (fingolimod), a Sph analogue, for the treatment of relapsing remitting multiple
sclerosis (rrMS) (Figure 2). Mechanistically, 1 is a prodrug that is phosphorylated by SphK2.
Phosphorylated 1 ((S)-phospho-fingolimod) is then transported outside of cells via the S1P
25
exporter, SPNS2 , to act as a functional antagonist at the S1P1/3/4/5 receptors, triggering S1P
receptor internalization and degradation. Specifically, the absence of the cell surface S1P1 receptor
hinders the ability of naïve and central memory T lymphocytes to follow the lymph S1P
concentration gradient out of secondary lymphoid tissues into the lymph and thence to the blood,
ultimately resulting in therapeutic immunosuppression for patients with rrMS. More recently, the
FDA granted marketing approval for 2 (siponimod) for the treatment of secondary progressive
multiple sclerosis.26
Figure 2. Chemical structures of Fingolimod and Siponimod.
In addition to acting on the S1P receptors, targeting other portions of the S1P pathway such as
SphK are also promising. To date, there are examples of SphK1 selective as well as SphK1/2 dual
inhibitors. Compounds 3a (PF-543), 3b (Amgen 23), and 4 (SLP7111228) are three of the most
4
ACS Paragon Plus Environment

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potent SphK1 selective inhibitors reported and have served as valuable molecular tools to probe
the S1P pathway for better understanding (Figure 3).27-29 In an animal model for sickle cell disease
(SCD), mice treated with 3a exhibited reduced blood and plasma S1P levels resulting in decreased
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
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
erythrocyte sickling. In a separate study, rats dosed with 10 mg/kg of 4 displayed decreased blood
S1P levels by about 80%. Further evaluation of 4 exhibited favorable in vivo stability, low
nanomolar potency, and high (>200 fold) selectivity for SphK1. As is the case for SphK1 selective
inhibitors, progress has been made to develop SphK1/2 dual inhibitors. A strategy employed by
Pitson et al. utilizes both the substrate and ATP binding sites to afford compound 5 (MP-A08)
3
0
with measured human SphK1 and SphK2 K values of 27  3 µM and 6.9  0.8 µM, respectively.
i
Another dual inhibitor is compound 6 (27a) developed by Schnute and coworkers at Pfizer (Figure
).31 Inhibitor 6 exhibits an IC50 of <1.7 nM for both kinase isoforms. Further computational
3
docking revealed the phenyl ether moiety in 6 potentially allows for favorable rotation inside the
enzyme binding pocket. In this manner, 6 avoids the deep hydrophobic pocket found in the Sph
active site and instead participates in strong π−hydrophobic or π−π interactions between the phenyl
substituent and F259.
5
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
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
Figure 3. Select structures of SphK1, dual SphK1/2, and SphK2 inhibitors.
In contrast, highly potent (<100 nM) and selective (>100-fold) SphK2 inhibitors are scarce, thus
prompting our interest in expanding the known chemical toolkit. Two early generation compounds
designed by our group are 7 (SLR080811) and 8 (SLM6031434) (Figure 3).23,22 Both of these
guanidine-based inhibitors possess moderate selectivity, low micromolar SphK2 potency, and
ability to lower cellular S1P levels in vitro with measured mouse SphK2 K values of 1.3 µM and
i
0.4 µM, respectively.22 Additionally, compound 8 treatment in mice was successful in
recapitulating the inverse relationship observed with SphK2 inhibition or genetic deficiency and
an increase in blood S1P concentration. Herein we report the development, synthesis, and
biological evaluation of, to the best of our knowledge, the most potent SphK2 selective inhibitor
with in vivo activity described to date.
RESULTS AND DISCUSSION
6
ACS Paragon Plus Environment

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Inhibitor Design. Previous work in our group was committed to the development of novel and
potent SphK2 selective inhibitors, leading to the discovery of 7, a second generation SphK2
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
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
inhibitor. Evaluation of 7 revealed a 9-fold selectivity for the SphK2 over SphK1 with a SphK2 K
i
value of 1.3 µM (12 µM for SphK1). Structurally, inhibitor 7 possesses an octylphenyl “tail”
attached to a 1,2,4-oxadiazole linker, followed by a guanylated pyrrolidine head group. Molecular
modeling studies conducted with similar SphK inhibitors possessing a guanidine head group
predict that the guanidine moiety in our scaffold hydrogen bonds to catalytic amino acid residue
Asp211 flanking the ATP active site when docked in the Sph binding pocket of SphK2.33,34
Structure-activity relationship (SAR) studies conducted by Congdon et al. on the scaffold of 7
established that deletion of the aryl portion of the octylbenzene tail resulted in a significantly
diminished inhibition towards SphK2, demonstrating that the aryl ring is essential for inhibitor
efficacy.21 Further SAR experiments focused on the tail portion led to the development of
inhibitors 8 and 9 (SLM6031422) (Table 1). Both analogues retain the 1,2,4-oxadiazole linker and
guanylated pyrrolidine head group but with an additional ether linkage to the octylbenzene tail. In
addition, 8 possesses a trifluoromethyl moiety at the 3-position of the internal aryl ring.
Surprisingly, the trifluoromethyl addition in compound 8 imparted improved selectivity (50-fold)
and potency (0.4 µM) for SphK2. In contrast, compound 9 exhibited reduced potency towards both
SphK1 and SphK2 compared to 7. These results highlight the significance of the trifluoromethyl
group in SphK2 selective binding.
The Sph binding site of SphKs can be divided up into three regions as defined by Worrell
et al. They are as follows: (1) the head region adjacent to the ATP binding site, (2) the hydrophobic
3
4
core region, and (3) the tail region. Molecular docking of 8 and 9 demonstrate different ligand
orientations in the Sph binding site of SphK2, giving insight into the role and influence the
7
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
trifluoromethyl group has in regard to SphK2 selectivity (Figure 4). Compound 8 docks in the Sph
binding pocket of SphK2 in a position that indicates more interactions in the tail region of the
binding site, as influenced by the hydrophobic core of the pocket, in particular, the space for the
trifluoromethyl group to sit between hydrophobic residues Phe548, Leu544, and Leu547 (Figure
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
4
A, B). This positioning and space at helix α8 can be accommodated in SphK2, whereas in docked
poses of 9 (Figure 4C, D) in SphK2, or docked poses of 8 in SphK1 (see supporting information),
this is not observed. In comparison, compound 9 is positioned in the Sph binding site of SphK2 in
the traditional “J-shape” conformation that has been observed in SphK1 structures crystallized
with various inhibitors.35,36 Additionally, interactions of 9 with residues in the head region near
the ATP binding site are observed. These interactions are at the threshold for hydrogen bonding
and strong electrostatic interactions, indicating a potentially weak and non-selective inhibitor
based on previous docking studies.33,34 The unique binding mode of 8 maximizes interactions in
the hydrophobic core and tail regions within the Sph binding site of SphK2, a phenomenon not
observed in SphK1 (see supporting information). There are small but significant differences
between the Sph binding site residues of SphK1 versus SphK2 with variations Ile174 to Val304 in
3
4
the hydrophobic core region and Phe288 to Cys533 in the tail region, respectively. With SphK2
having the smaller amino acid, valine, in the hydrophobic core of the binding site, the
trifluoromethyl group can favorably interact with the residues (Phe548, Leu544, and Leu547) that
form a side cavity within the Sph binding pocket of SphK2. This interaction in the side cavity is
not tolerated in SphK1 with the presence of the larger isoleucine residue. Additionally, the
presence of the smaller cysteine residue in the tail region of SphK2 grants access for our inhibitors
to migrate deeper towards the tail of the binding pocket which, in comparison, is blocked by
Phe288 in SphK1, resulting in restricted inhibitor docking (see supporting information).
8
ACS Paragon Plus Environment

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Further SAR experiments conducted in our group with the scaffold of 8 led to the
development of molecule 10 (SLM6071469) (Table 1). This inhibitor replaces the octyloxy
benzene moiety for a shorter, less flexible 4-trifluoromethyl benzyl tail that, when assessed in
vitro, exhibited comparable activity to 8. Molecular docking of 10 in the Sph binding site of SphK2
demonstrates a similar pose as 8 (Figure 4E, F). These results highlight the superiority of the 4-
trifluoromethylbenzyl tail with significantly reduced rotatable bonds, which allows favorable
hydrophobic interactions at key residues in the tail region of the binding site. These presumably
lead to the observed comparable inhibition towards SphK2. Further evaluation with SphK2
illustrated that 1 μM of 10 was successful in inhibiting SphK2 activity by roughly 77%, albeit with
decreased SphK2 selectivity, and was more potent than 8 at lower (0.3 μM) concentrations (Table
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
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
). In this study, we probed the importance of substituents spanning the internal phenyl ring and
the potential presence of an unexplored side cavity within the Sph binding pocket of SphK2. A
wide range of alkyl and aryl substituents were attached and examined for their effect on SphK2
inhibition.
9
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
Table 1. Inhibitory activity of select compounds represented as % inhibition of SphK1 and SphK2a
% hSphK1 % hSphK2 % hSphK2
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
Compound
Structure
inhibition
inhibition
inhibition
(1.0 μM)
(1.0 μM)
(0.3 μM)
8
6 ± 5
80 ± 2
8 ± 2
51 ± 4
3 ± 1
9
11 ± 3
20 ± 1
10
77 ± 3
65 ± 3
aSphK inhibition is presented as % control (no inhibitor added). Recombinant human SphK1 or
SphK2 were isolated from cell lysates. Enzyme inhibition was measured with 5 μM (SphK1) or
3
2
1
0 μM (SphK2) sphingosine and 250 μM γ-[ P] ATP. Compounds were assayed at 1.0 and 0.3
μM in triplicate.
1
0
ACS Paragon Plus Environment

Page 11 of 71
Journal of Medicinal Chemistry
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
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
Figure 4. Comparison of the docked poses of 8 (A, B), 9 (C, D), and 10 (E, F) in a homology
model of hSphK2 (PDB ID: 3VZB used as a starting template). Key residues in the binding pocket
are represented by grey sticks and are labeled, ATP is shown in orange and colored by element.
The SphK2 protein structure is depicted in grey cartoon. Distances between interacting atoms are
shown as dashed lines. Inhibitors are shown as stick and colored by element, with the carbon atom
indicating inhibitor (8 – blue, 9 – teal, 10 – cyan). Panels B, D, and F show the inhibitor as Van
der Waals radii in dots to represent volume occupancy of the inhibitors.
Chemistry. To investigate substitutions on the internal aryl ring of the 10 scaffold, analogues 15a-
q, and 15s were synthesized containing various alkyl and aryl moieties at the 2- and 3-positions as
shown in Scheme 1. Intermediates 11a-e were synthesized in a microwave reactor via two different
1
1
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
Suzuki-Miyaura coupling conditions using 3-bromo-4-hydroxybenzonitrile (11g) and various
boronic acids. Compounds 11f-o are available for purchase from commercial sources. Allylation
of 4-hydroxybenzonitrile (11f) with allyl bromide and potassium carbonate was performed
followed by microwave-assisted Claisen rearrangement to afford compound 11p. Subsequent
hydrogenation of the allyl group with Pd/C in the presence of tetrahydroxydiboron afforded the
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
7
propyl analog 11q. Compound 11r is available for purchase from commercial sources. With the
requisite materials in hand, derivatives 11a-r were reacted with 4-(trifluoromethyl)benzyl bromide
and potassium carbonate at 80 ºC in acetonitrile to afford intermediates 12a-r. Subsequently, a
mixture of hydroxylamine hydrochloride, triethylamine (TEA) and 12a-r were refluxed in ethanol
to generate amidoxime intermediates 13a-r. Afterwards, 13a-r were dissolved in DMF and heated
to 100 ºC in the presence of Boc-L-proline, HCTU and Hünig’s base to afford 1,2,4-oxadiazole
intermediates 14a-r. Synthesis of 14s was performed via a Sonogashira coupling reaction with 14g
and TMS-acetylene, followed by removal of the TMS group using TBAF and subsequent
hydrogenation. Lastly, Boc deprotection of compounds 14a-s was completed using trifluoroacetic
acid (TFA), which was followed by reaction with N,N’-di-Boc-1H-pyrazole-1-carboxamidine and
Hünig’s base to install the guanidine moiety. Removal of the Boc protecting groups was performed
with HCl in methanol to provide the desired analogues 15a-q, 15s, and 10 as HCl salts (see Table
2
for structures).
1
2
ACS Paragon Plus Environment

Page 13 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of Analogues 15a–q, 15s, and 10 a
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
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
a
Reagents and conditions: (a) boronic acid derivative, Cs CO , PdCl (dppf), DMF, 100 ºC,
2
3
2
microwave, 60 min, 52-57%; (b) boronic acid derivative, K PO H, Pd(OAc) P(Cy) tol/H O,
2
4
2,
3,
2
1
00 ºC, 72 h, 50%; (c) allyl bromide, K CO , CH CN, 80 ºC, 4-6 h, 92%; (d) neat, 200 ºC,
2 3 3
microwave, 20 min, 76%; (e) Pd/C, B (OH) , H O, CH Cl , rt, 18 h, 85%; (f) 4-
2
4
2
2
2
(
trifluoromethyl)benzyl bromide, K CO , CH CN, 80 ºC, 4-6 h, 48-99%; (g) NH OH•HCl, Et N,
2 3 3 2 3
EtOH, reflux, 3 h, 47-96%; (h) Boc-L-proline, DIEA, HCTU•PF , DMF, 100-110 ºC, 18 h, 27-
6
6
5%; (i) Et N, CuI, TMS-acetylene, Pd(PPh) Cl , THF, reflux, 48 h, 74%; (j) TBAF, THF, rt, 3
3 2 2
h, 100%; (k) H (g), Pd/C, EtOH, rt, 18 h, 45%. (l) TFA, CH Cl 3-12 h, 90-99%; (m) N,N’-di-
2
2
2,
1
3
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
Boc-1H-pyrazole-1-carboxamidine, DIEA, CH CN, 50 ºC, microwave, 3 h, 47-96%; (n) HCl (g),
MeOH, rt, 15 min, 90-99%.
3
Synthesis of methyl analogue 20 is outlined as shown in Scheme 2. Execution of a nucleophilic
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
aromatic substitution reaction of 4-(trifluoromethyl)benzyl alcohol with commercially available
4
-fluoro-3-methylbenzonitrile (16) and sodium hydride afforded benzonitrile intermediate 17.
Next, a mixture of 17, hydroxylamine hydrochloride, and TEA in ethanol was heated to generate
compound 18. Afterwards, HCTU coupling with Boc-L-proline, Boc deprotection, installation of
the guanidine group and final deprotection yielded the desired analogue 20 as an HCl salt.
Scheme 2. Synthesis of Analogue 20a
aReagents and conditions: (a) 4-(trifluoromethyl)benzyl alcohol, NaH, DMF, 0 ºC-rt, 18 h, 54%;
(
b) NH OH•HCl, Et N, EtOH, reflux, 3 h, 89%; (c) Boc-L-proline, DIEA, HCTU•PF , DMF, 100
2
3
6
ºC, 18 h, 36%; (d) TFA, CH Cl 12 h, 90%; (e) N,N’-di-Boc-1H-pyrazole-1-carboxamidine,
2
2,
DIEA, CH CN, 50 ºC, microwave, 3 h, 86%; (f) HCl (g), MeOH, rt, 15 min, 90%.
3
Synthesis of isopropyl analogue 27 is shown in Scheme 3. Intermediate 22 was produced
via the nucleophilic substitution of commercially available 3-acetyl-4-hydroxybenzonitrile (21)
using 4-(trifluoromethyl)benzyl bromide and potassium carbonate. Next, reduction of the acetyl
moiety was completed through
a
Wittig reaction using n-butyl lithium and
1
4
ACS Paragon Plus Environment

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
methyltriphenylphosphonium bromide dissolved in THF and cooled to 0 ºC for eighteen hours to
generate compound 23. Next, refluxing 23 in ethanol for three hours in the presence of
hydroxylamine hydrochloride and TEA generated amidoxime intermediate 24. Following the
sequence of oxadiazole formation, hydrogenation, Boc deprotection, installation of guanidine and
final removal of Boc groups produced derivative 27.
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
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
Scheme 3. Synthesis of Analogue 27a
a
Reagents and conditions: (a) 4-(trifluoromethyl)benzyl bromide, K CO , CH CN, 80 ºC, 6 h,
2
3
3
9
7%; (b) n-BuLi, CH P(Ph) Br, THF, 0 ºC-rt, 18 h, 59%; (c) NH OH•HCl, Et N, EtOH, reflux, 3
3 3 2 3
h, 90%; (d) Boc-L-proline, DIEA, HCTU•PF , DMF, 100 ºC, 18 h, 67%; (e) H (g), Pd/C, EtOH,
6
2
rt, 18 h, 72%; (f) TFA, CH Cl 12 h, 90%; (g) N,N’-di-Boc-1H-pyrazole-1-carboxamidine, DIEA,
2
2,
CH CN, 50 ºC, microwave, 3 h, 62%; (h) HCl (g), MeOH, rt, 15 min, 90%.
3
Structure-Activity Relationship Studies and Biologic Evaluation of Derivatives. The goal of
this SAR was to improve SphK2 inhibition observed with 10 and probe the significance of
substitutions on the internal phenyl ring with respect to SphK2 inhibition and selectivity. A library
of compounds with various substituents spanning the 2-, 3- and 5-positions of the internal phenyl
1
5
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
ring were synthesized and assayed according to a previously described protocol.29,38 This assay,
3
2
which utilizes recombinant SphKs and -[ P]ATP, was used as a preliminary screen to identify
inhibitors suitable for further evaluation. The compounds were assayed at 0.3 M. To confirm the
advantage of substituents attached to the 3-position of the internal phenyl ring, 15f was synthesized
to represent the minimal scaffold with no substituents attached (Table 2). As expected, removal of
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
the CF
3
group resulted in little to no inhibitory activity towards SphK2. Compounds 15a-d were
moiety for a much bulkier aryl substituent to probe
designed to replace the relatively small CF
3
the size of any potential docking space, as well as potential - stacking or hydrophobic
interactions with nearby residues Phe-548 and His-556. Substitution with a phenyl ring (15a)
resulted in poor SphK2 inhibition. Likewise, decoration of the phenyl substituent (15b-c) to exploit
potential hydrophobic interactions, also led to poor inhibition of SphK2. Compound 15d was
synthesized to mimic 10, but with an additional phenyl ring in between the scaffold backbone and
CF
3
. Our rationale was that the added phenyl linker could potentially extend the CF moiety deeper
3
into the previously unexplored side cavity leading to increased ligand affinity. However, this
increase in steric bulk resulted in a significant loss of efficacy compared to 10. These results
suggest that larger substituents sterically clashed inside that portion of the active site, directing us
towards smaller moieties. Thus, halogen substituents at the 3-carbon were explored with
compounds 15g and 15i-j. The identity of the halogen seemed to be important with the most
successful being the chloro variant (15i), inhibiting normal SphK2 activity by about 33%.
Compared to inhibitor 15i, it was observed that the fluoro (15j) and bromo (15g) derivatives did
not have as large of an inhibitory impact towards SphK2. We next investigated the effect of
migrating substituents over to the 2-position of the internal phenyl ring. However, compared to 10,
both the 2-chloro (15l) and 2-CF (15m) moieties had minimal impact on SphK2 inhibition. We
3
1
6
ACS Paragon Plus Environment

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
also examined the effect of the electronics on the phenyl ring with electron withdrawing (15n) or
donating (15k) groups. While the nitro derivative 15n was roughly twice as effective as methoxy
variant 15k with inhibiting SphK2, neither was improved compared to 10. Next, the impact of 3,5-
disubstitution was determined with compound 15h. Unfortunately, addition of a second moiety on
our scaffold resulted in poor inhibition of SphK2. Taken together, our data indicate that
monosubstitution at the 3-position of the aryl ring is conducive to further structure-activity
relationship profiling. Thus, alkyl substituents were synthesized to probe the size of this potential
side cavity within the binding site. Replacing the trifluoromethyl group with a smaller methyl (20)
led to a significant decrease in potency while the slightly larger ethyl (15s) substituent had modest
SphK2 inhibitory activity. To fine-tune the size of the side cavity, we synthesized compounds
containing propyl (15q), allyl (15p), isopropyl (27), cyclopropyl (15e), and tert-butyl (15o) groups.
Collectively, these moieties had similar inhibitory activity against SphK2 along with 10.
Nonetheless, our studies indicate that there is a bona fide side binding cavity around the central
phenyl ring that can accommodate a substituent larger than methyl group, but smaller than a phenyl
ring.
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
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
7
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
Table 2. Activity of SphK2 inhibitors represented as % inhibition of the enzyme.a
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
%
SphK2
% SphK2
inhibition
(0.3 μM)
Cmpd
10
R
inhibition Cmpd
R
(0.3 μM)
65 ± 3
15n
15k
33  4
15  1
15f
2  0.2
28  1
15a
15h
20
23  2
3  0.4
42  4
1
5b
31  1.9
8  0.8
1
5c
15s
1
5d
12  1
24  3
15q
15p
60  4
59  4
1
5g
1
8
ACS Paragon Plus Environment

Page 19 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
5i
33  2
23  2
27
67  5
52  6
1
5j
15e
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
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
5l
40  4
35  3
15o
66  7
1
5m
aSphK inhibition is presented as % control (no inhibitor added). Recombinant human SphK2
was isolated from a cell lysate. Enzyme activity was measured with 10 μM (SphK2), sphingosine
and 250 μM γ-[ P] ATP. Compounds were assayed at 0.3 μM in triplicate.
32
Considering our observations summarized in Table 2, select inhibitors that displayed moderate
(
>42%) or greater SphK2 inhibitory activity were carried forward for further evaluation. Thus,
inhibitors were subjected to an in vitro assay that utilizes the budding yeast Saccharomyces
3
9
cerevisiae as a platform for assessing inhibitors of human SphK1 and 2 (hSphK1 and 2). In short,
this assay takes advantage of the observed toxicity of excessive levels of phosphorylated long
chain bases, such as S1P, towards S. cerevisiae. Rescue of a modified yeast cell line (KYA1)
designed to harbor plasmids expressing either hSphK1 or 2 in the presence of select inhibitors can
be accomplished in a dose dependent fashion by measuring the growth of yeast culture. In this
manner, dose-response curves with our lead compound 10 demonstrated hSphK2 inhibition with
an EC50 of 59 nM, while also displaying a 16-fold selectivity for hSphK2 over hSphK1 (Figure 6).
As shown in Table 3, the n-propyl (15q) and isopropyl (27) derivatives were the most potent
inhibitors with measured EC50 values of 37 nM and 32 nM respectively. Interestingly, the
trifluoromethyl (10), allyl (15p), cyclopropyl (15e) and tert-butyl (15o) analogs were slightly less
1
9
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
6
4
2
0
6
A
B
2
7
27
15s
15e
15s
15e
4
2
0
1
1
1
1
5q
15q
0
10
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
5o
5p
1
5o
5p
1
-
9
-8
-7
-6
-9
-8
-7
-6
log [cmpd]
log [cmpd]
Figure 6. Rescue of growth of yeast strain KYA1 expressing hSphK1 (A) or hSphK2 (B) in the
presence of inhibitor over 24 hours.
potent with EC50 values ranging from 50 nM to 84 nM, while the ethyl derivative (15s) was the
least potent (EC50 = 146 nM). From our experience, there are some nuances in using the yeast cells
as a secondary screening tool. For example, inhibitor concentrations higher than 3 M result in
cell death; fortunately, our compounds are effective at much lower concentrations. Furthermore,
guanidine containing compounds accumulate inside the yeast cells at varying degrees, thus,
generating apparent improved efficacy. Nonetheless, this assay affords a rapid and simple process
for validating SphK inhibition in vitro.
23
The inhibitory constant (K
shown in Table 4, evaluation of the propyl derivative (15q) displayed good SphK2 inhibition with
a K of 201 nM whereas the tert-butyl (15o), isopropyl (27), and cyclopropyl (15e) substituents
exhibited similar K ’s ranging from 197-186 nM. Interestingly, the trifluoromethyl analogue (10)
proved to be the most potent SphK2 inhibitor in the series with a K of 89 nM. This result is
i
) was determined for the most potent derivatives in this library. As
i
i
i
consistent with our molecular modeling studies that indicate a side cavity within the Sph binding
pocket comprised of Phe548, Leu544, and Leu547. That is, the side cavity is relatively small,
2
0
ACS Paragon Plus Environment

Page 21 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. EC50 Values of Select SphK2 Inhibitors. a
hSphK1 EC50
nM)
hSphK2 EC50
(nM)
Compound
Structure
(
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
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
27
>550  85
>1000  160
>800  100
32  6
146  15
50  11
1
5s
1
5e
15q
10
>800  105
37  10
>950  65
>500 45
59  12
73  9
1
5o
1
5p
>1000  135
84  11
2
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
a
EC50 values of various compounds were calculated from the growth curves of yeast strain KYA1
encoding either hSphK1 or hSphK2 in the presence of inhibitor over 24 hours. For detailed assay
conditions, see the experimental section.
hydrophobic, and can best accommodate a trifluoromethyl group due to its optimal size and
hydrophobic nature. Furthermore, the selectivity of the compounds between SphK2 vs. SphK1
follows the trend of increasing size leading to poor SphK2 selectivity attributing to the fact that
SphK2 has an overall smaller binding pocket than SphK1. Specifically, SphK2 selectivity
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
decreases from CF > cyclopropyl ~ isopropyl > propyl > tert-butyl. One of the key residues that
3
lines the SphK2 substrate binding site is Val304, which is Ile174 in SphK1. It is predicted that
Ile174 of SphK1 causes a steric clash with the trifluoromethyl group forcing the molecule to shift
upwards and away from the ATP binding site, and as a result, the inhibitor migrates away from
the side cavity (see supporting information). However, the variation of Ile174 → Val304 in SphK2
does not cause this steric clash, thus allowing the molecule to slide deeper into the binding site
permitting the trifluoromethyl group to lock into the side cavity leading to improved SphK2
selectivity. Likewise, when the phenyl substituent is a tert-butyl group, the steric clash with Ile174
in SphK1 becomes even greater, sending the inhibitor towards the head of the binding pocket.
However, unlike 10, the increased steric clash orients the molecule towards the ATP binding site,
leading to marginal SphK1 inhibition (see supporting information).
To assess inhibition of SphK2 in vivo, we administered C57BL/6 male mice with a single
1
0 mg/kg intraperitoneal (i.p.) dose of 10 and measured blood S1P levels. S1P level has proven
to be an excellent pharmacodynamic marker for SphK2 inhibition.22-24 Thus, blood S1P levels
were measured at the indicated time points up to 8 hours post administration of 10 via LC-MS/MS.
As shown in Figure 7, the levels of 10 in blood peaked at ~8 M, which increased blood S1P levels
by nearly two-fold after 4 hours and remained elevated for up to 6 hours. The elevation of S1P
2
2
ACS Paragon Plus Environment

Page 23 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
levels indicates blockade of SphK2 activity in vivo and recapitulate the inverse relationship
between blood S1P levels and SphK2 inhibition. Taken together, our investigations suggest that
10 not only inhibits SphK2 in vitro using recombinant protein and in yeast, but also in vivo using
C57BL/6 mice.
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
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
Table 4. K
Values of Select SphK2 Inhibitors.a
i
hSphK1 K
nM)
i
hSphK2 K
i
hSphK2
selectivity
Cmpd
Structure
(
(nM)
1
5q
9600  2100
201  11
48
1
5o
6100  1700
3500  800
5300  1700
6500  1500
197  76
192  39
186  12
89  8
31
18
28
73
2
7
15e
1
0
aInhibitory constants for recombinant enzymes were obtained by kinetic analysis of S1P
production using variable concentration of sphingosine and a fixed concentration of ATP in the
2
3
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
presence or absence of compounds. Selectivity for each compound was determined by dividing
the K
m
SphK2 by the K of SphK1.
m
12
10
8
1
0
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
S1P
**
*
6
*
4
2
0
0
1
2
3
4
5
6
7
8
Time (h)
Figure 7. Blood concentrations of S1P and 10 in C57BL/6 male mice after IP administration with
10 mg/kg of 10. The level of significance is indicated for S1P concentrations (*P < 0.05 and **P
< 0.01) using an unpaired t-test (compared to vehicle injected control).
CONCLUSIONS
Modulating S1P signaling at the level of S1P receptors represents a clinically validated approach
for managing relapsing remitting multiple sclerosis. This fact encourages investigations as to
whether other nodes in the S1P signaling cascades are also drug targets. Inhibition of the S1P
generative enzymes SphK1 and 2 can facilitate the decrease or increase in blood S1P levels,
respectively, which may have the potential for treating other diseases such as sickle cell disease,
renal fibrosis and cancer. As a result of SphK1 and its inhibitors being a subject of intense
investigation, potent and SphK1 selective compounds have been developed. However, potent and
SphK2 selective inhibitors are lacking. In this report, we discovered a heretofore undisclosed side
cavity within the Sph binding site. Interrogating this cavity with various functional groups revealed
that relatively small hydrophobic moieties (i.e., trifluoromethyl) are best accommodated. Indeed,
characterization of inhibitor 10 revealed the most potent SphK2 selective inhibitor to date (K 89
i
2
4
ACS Paragon Plus Environment

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nM, 73-fold SphK2 selective). Molecular modeling studies suggest that the trifluoromethyl group
fits securely into the side cavity generated by residues Phe548, Leu544, and Leu547 imparting
increased potency towards SphK2, whereas the Val304 to Ile174 variation in SphK1 causes a steric
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
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
clash that forces the CF group out of the side cavity. Collectively, these phenomena increase
3
SphK2 inhibitor potency and selectivity. Overall, our investigations provide a platform for
developing SphK inhibitors that may lead to compounds useful for treating diseases where
sphingosine mediated signaling events have gone awry.
EXPERIMENTAL
Sphingosine Kinase Biological Assays. The percent inhibition of SphK2 from synthesized
compounds was carried forward using a previously described protocol.29,38 Recombinant human
SphK2 (10 μM) was isolated from a cell lysate and incubated with (0.3 μM) or without compound,
32
sphingosine, and 250 μM γ-[ P] ATP via scintillation counting. Compounds were assayed in
triplicate.
The growth of recombinant yeast (Saccharomyces cerevisiae) cells encoding either
3
9
hSphK1 or hSphK2 was performed according to our previously described protocol. Briefly, yeast
strain KYA1 harboring a plasmid encoding either hSphK1 or hSphK2 was selected, maintained
and grown in SC-URA media with 2% glucose overnight at 30 ºC. Following overnight growth
media was diluted 1:100 into SC-URA media supplemented with 2% galactose and various
concentrations of test inhibitor. After another incubation period of 24 - 48 hours at 30 ºC, cellular
growth was quantified by measuring absorbance at 600 nm.
In vivo mouse studies were conducted as follows. Groups of 8 to 10 week old male
C57BL/6 mice were injected intraperitoneally (i.p.) with either vehicle or 10 mg/kg
2
5
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
SLM6071469 (2% solution of hydroxypropyl-β-cyclodextrin, Cargill Cavitron 82004). After
0
.5, 1, 2, 4, 6, and 8 hours of injection, whole blood was collected from the retro-orbital sinus
22
and 10 L of blood was processed for LC/MS analysis as described previously. The use of
mice for these studies was approved by the University of Virginia’s School of Medicine Animal
Care and Use Committee prior to experimentation.
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
Molecular Docking. To visualize and rationalize differences in efficacy of 8, 9, 10, and 15o
molecular docking was utilized to predict position and ligand-protein interactions of the inhibitors
2+
in the sphingosine binding pocket of SphK2. The SphK2 model, with ATP and Mg bound, was
generated with Molecular Operating Environment (MOE) and energy minimized as previously
described.33,34 In order to draw, display, and characterize chemical structures, substructures, and
reactions for preparation in docking programs, Marvin 17.3.13, 2017 was used (ChemAxon;
http://www.chemaxon.com) to create structure files that were cleaned in 3D . AutoDock Tools was
40
used to prepare the protein and ligand files. In order to perform the docking, AutoDock Vina was
41
utilized to dock the inhibitors to SphK2, with 9 poses being created for each inhibitor. The grid
3
box was set to 20 X 20 X 28 Å with grid spacing of 1.000 Å. To include all known key residues
in the sphingosine binding cavity, the grid box was positioned at the approximate center of the
ligand-binding cavity based on the position of sphingosine in the crystal structure (PDB ID:
3
VZB).33,35 The lowest energy docked pose for each molecule was then used to compare potential
differences in ligand docking position inside the SphK2 active site.
General Material and Synthetic Procedures. Where indicated, reactions utilizing a microwave
reactor were performed with a Discover SP microwave synthesizer (CEM Corporation). All
2
6
ACS Paragon Plus Environment

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
solvents used were dried using a PureSolv solvent system and all chemical reagents were
purchased from commercial sources without further purification. Aluminum-backed silica gel, 200
μm, F254 plates were utilized for all thin-layer chromatography (TLC) experiments. A Combiflash
Rf purification system with flash grade silica gel, 40−63 μm was used for all column
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
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
chromatography. All reported H NMR spectra were measured at 400 MHz and the complementary
1
3
1
C NMR frequency was 101 MHz. Reported H NMR chemical shifts are in ppm with the solvent
resonance as an internal standard (CDCl
3
: 7.26 ppm; CD
3
OD: 3.31 ppm; (CD
3
2
) CO: 2.05 ppm).
13
Reported C NMR chemical shifts are in ppm with the solvent resonance as the internal standard
(CDCl
3
: 77.16 ppm; CD
3
OD: 49.00 ppm; (CD
3
2
) CO: 206.26 ppm). High-resolution mass
spectrometry (HRMS) data was obtained with an LC-MS time-of flight mass spectrometer by
electrospray ionization (ESI). All compounds tested in biological assays possess ≥95% purity, as
determined by HPLC analyses. HPLC was performed using a Thermo Electron TSQ triple
quadrupole mass spectrometer fitted with an ESI source.
General Procedure A: Suzuki-Miyaura Coupling of Aryl Bromides with Phenyl Boronic
Acids. 3-bromo-4-hydroxybenzonitrile (11g) (1.0 equiv) was dissolved in DMF (0.2 M solution)
and put under a nitrogen gas atmosphere. The appropriate phenyl boronic acid (3.0 equiv) was then
added followed by Cs
2
CO
3
(2.0 equiv). The resulting mixture was degassed for 15 min by bubbling
(0.03 equiv) was added and the resulting mixture was
N
2
through solution. Next, Pd(dppf)Cl
2
heated in a microwave reactor at 100 ºC for 60 min. Next, the solution was partitioned between
EtOAc and LiBr aqueous solution. Using additional EtOAc, the aqueous LiBr solution was washed
three times and the combined organic layers were dried over Na
2
SO , filtered, and concentrated
4
via vacuum. The resulting concentrate was then purified by silica gel chromatography to yield a
2
7
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
mixture of the major product detected by TLC as well as a minor impurity. Isolated fractions were
then carried forward crude without further characterization.
General Procedure B: Williamson Ether Synthesis of Hydroxybenzonitrile Derivatives.
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
K
2
CO
3
(4 equiv) was added to a solution of the CH CN (0.2 M solution) and the appropriate
3
hydroxybenzonitrile derivative (1.0 equiv). 4-(trifluoromethyl)benzyl bromide (1.2 equiv) was
then added followed by flushing the reaction container with nitrogen gas. The reaction mixture
was refluxed at 80 ºC for 4 – 6 hours. Afterward, the reaction was concentrated via vacuum and
then purified by silica gel chromatography to yield the desired product.
General Procedure C: Synthesis of Amide-Oxime Analogs. To a solution of ethanol (0.2 M)
and the appropriate benzonitrile (1.0 equiv), hydroxylamine hydrochloride (2.0 equiv) was added
followed by TEA (3.0 equiv). The reaction mixture was heated to 85 ºC for 4 hours. The resulting
mixture was then concentrated via vacuum and then purified by silica gel chromatography to yield
the desired product.
General Procedure D: Coupling of Amide-Oxime Analogs with Boc-L-Proline. The
appropriate amide-oxime derivative (1.0 equiv) was dissolved in DMF (0.2 M solution) and put
under an N
2
blanket. Boc-L-Proline (1.2 equiv) followed by Hünig’s base (1.8 equiv) and HCTU
(1.2 equiv) were then added and stirred vigorously at 100 ºC for 18 hours. The reaction progress
was monitored by TLC. Subsequently, the resulting solution was partitioned between EtOAc and
LiBr aqueous solution. Using additional EtOAc, the aqueous LiBr solution was washed three times
and the combined organic layers were dried over Na
2
SO , filtered, and concentrated via vacuum.
4
The resulting concentrate was purified by silica gel chromatography.
General Procedure E: Boc Deprotection of Pyrrolidine Intermediates With TFA,
Guanylation of Deprotected Amines and Subsequent Boc Deprotection of Guanyl Groups
2
8
ACS Paragon Plus Environment

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with HCl (g). To a solution of dichloromethane (0.2 M) and the appropriate Boc protected
pyrrolidine (1.0 equiv), 1.0 M TFA (10 equiv) was added and allowed to stir at room temperature
for 3 – 12 hours. The reaction progress was monitored by TLC. Next, the dichloromethane was
removed via vacuum and replaced with diethyl ether to promote TFA salt precipitation. Once the
desired product was precipitated, the diethyl ether was removed via vacuum and the resulting Boc
deprotected product was carried forward crude. Thereafter, the resulting Boc deprotected product
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
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
was dissolved in CH
3
CN (0.2 M) and put under an N atmosphere. To the reaction mixture,
2
Hünig’s base (10 equiv) followed by N,N’-di-Boc-1H-pyrazole-1-carboxamidine (1.2 equiv) were
added and the reaction mixture was heated to 50 ºC inside a microwave reactor for 3 hours. The
reaction mixture was then concentrated via vacuum, and the resulting concentrate was then
purified by silica gel chromatography to afford bis-Boc-protected intermediates. Following silica
gel chromatography, bis-Boc-protected intermediates were then dissolved in methanol (0.2 M) and
bubbled with HCl gas for 2 – 5 minutes to yield the desired Boc deprotected moiety. Lastly, the
organic mixture was concentrated via vacuum and triturated with diethyl ether to afford the
corresponding analogue as an HCl salt.
CHARACTERIZATION
(S)-2-(3-(3-(trifluoromethyl)-4-((4-(trifluoromethyl)benzyl)oxy)phenyl)-1,2,4-oxadiazol-5-
yl)pyrrolidine-1-carboximidamide (10). Synthesized by General Procedure E: 58 mg, 59%,
1
white solid. H NMR (400 MHz, CD
3
OD)  8.28 – 8.26 (m, 2H), 7.70 (q, J = 8.3 Hz, 4H), 7.44 (d,
J = 9.2 Hz, 1H), 5.45 (dd, J = 8.0, 1.8 Hz, 1H), 5.42 (s, 2H), 3.78 (t, J = 9.4 Hz, 1H), 3.62 (q, J =
13
9
.6 Hz, 1H), 2.61 – 2.44 (m, 2H), 2.28 – 2.19 (m, 1H), 2.13 – 2.02 (m, 1H); C NMR (101 MHz,
CD
OD) δ 179.28, 168.45, 159.92, 157.11, 141.91, 134.01, 131.26 (q, 2JCF = 32.5 Hz), 128.54,
3
127.25 (q, 3JCF = 5.0 Hz), 126.53 (q, 3JCF = 3.6 Hz), 125.53 (q, 1JCF = 272.9 Hz), 124.63 (q, 1JCF
=
2
9
ACS Paragon Plus Environment