Discovery, biological evaluation, and structure-activity relationship of amidine based sphingosine kinase inhibitors

Article abstract of DOI:10.1021/jm901860h

Sphingosine 1-phosphate (S1P), a potent phospholipid growth and trophic factor, is synthesized in vivo by two sphingosine kinases. Thus these kinases have been proposed as important drug targets for treatment of hyperproliferative diseases and inflammation. We report here a new class of amidine-based sphingosine analogues that are competitive inhibitors of sphingosine kinases exhibiting varying degrees of enzyme selectivity. These inhibitors display K_I values in the submicromolar range for both sphingosine kinases and, in cultured vascular smooth muscle cells, decrease S1P levels and initiate growth arrest.

Full text of DOI:10.1021/jm901860h

2766 J. Med. Chem. 2010, 53, 2766–2778
DOI: 10.1021/jm901860h
Discovery, Biological Evaluation, and Structure-Activity Relationship of Amidine Based Sphingosine
Kinase Inhibitors
Thomas P. Mathews,† Andrew J. Kennedy,† Yugesh Kharel,‡ Perry C. Kennedy,‡ Oana Nicoara,§, Manjula Sunkara,^
Andrew J. Morris,^ Brian R. Wamhoff,§, Kevin R. Lynch,‡ and Timothy L. Macdonald*,†,‡
†University of Virginia, Department of Chemistry, P.O. Box 400319, McCormick Road, Charlottesville, Virginia 22904, ‡University of Virginia,
Department of Pharmacology, 1340 Jefferson Park Avenue, Charlottesville, Virginia 22908, §University of Virginia, Department of Medicine,
Cardiovascular Division, The Robert M. Berne Cardiovascular Research Center, Charlottesville, Virginia 22908, and ^University of Kentucky,
Department of Internal Medicine, Lexington, Kentucky 40506
Received September 23, 2009
Sphingosine 1-phosphate (S1P), a potent phospholipid growthand trophic factor, is synthesizedin vivo by
two sphingosine kinases. Thus these kinases have been proposed as important drug targets for treatment
of hyperproliferative diseases and inflammation. We report here a new class of amidine-based sphingosine
analogues that are competitive inhibitors of sphingosine kinases exhibiting varying degrees of enzyme
selectivity. These inhibitors display KI values in the submicromolar range for both sphingosine kinases
and, in cultured vascular smooth muscle cells, decrease S1P levels and initiate growth arrest.
Introduction
manipulation of SphK2 might also be important in preventing
neoplastic disease progression.12 Studies such as these demon-
strate the importance of both sphingosine kinase types in
disease. However, the study of dual SphK inhibitors is under-
represented in the chemical literature. A recent report in this
journal documented the effectiveness of a dual inhibitor in
U937 cells and its potential as a future mode of therapy.13
We sought to add to this growing pool of inhibitors by
synthesizing a novel class of dual sphingosine kinase inhibi-
tors. Turning to examples from the chemical literature, we
noticed that conventional methods of kinase inhibition in-
volve the use of adenosine analogues to target the ATP biding
site. Although this strategy has been successful,14 the ATP
binding site can be similar across a wide array of kinases and
such inhibitors are often burdened by a lack of selectivity and
off-target effects. In terms of sphingosine kinases, the amino
acid sequence of the ATP binding domain of SphK1 and 2 is
conserved across a number of diacylglycerol (DAG) kinase
family members, rendering the established ATP-targeting
strategy particularly problematic.
While characterizing previously reported SphK substrates,
we discovered a class of sphingosine-like dual SphK1/2
inhibitors. Herein we document the effectiveness of targeting
the sphingosine-binding domain of the sphingosine kinases by
creating a series of amidine-based, SphK1/2 inhibitors includ-
ing molecules with affinity constants of less than 1 μM. We
also demonstrate an SAR strategy that was effective in
improving potency and selectivity between SphK1 and 2.
These inhibitors are effective in depressing S1P levels in
cultured cells and initiate growth arrest in proliferating
smooth muscle cells.
Sphingosine kinases 1 and 2 (SphK1 and 2)a catalyze the
phosphorylation of D-erythro sphingosine to form sphingo-
sine 1-phosphate (S1P).1 The sphingosine kinases control, in
large part, the equilibrium between the survival factor, S1P,
and its pro-apoptotic metabolic precursor, ceramide.2 S1P has
been demonstrated to be a potent agonist at five membrane-
bound G-proteincoupledreceptors, knownasS1P1-5,3 whose
roles in physiologic and pathophysiologic states are currently
under investigation. The most well understood receptor sub-
type, S1P1, is now recognized as the receptor responsible for
the antiapoptotic properties of S1P and is also implicated in
the control of lymphocyte trafficking.4 Because of their
regulation of S1P production, SphKs1 and 2 have been
proposed to be important small molecule drug targets.5
SphK null mice, small interfering RNAs, and small mole-
cule inhibitors have provided insight into the physiologic
importance of these enzymes. Sphk1-/- and Sphk2-/- mice
develop normally, while the double null genotype is embryo-
nic lethal at midgestation.6 While these data suggest a com-
pensatory mechanism for SphK1 and 2, their unequal dis-
tribution in cellular compartments,7 SphK1’s high degree of
inducibility,8 and SphK2’s pro-apoptotic BH3 domain9,10
have led researchers to view the SphK isoforms as unequal
with respect to their roles in hyperproliferative disease states.
In support of this long-standing hypothesis, Spiegel and
co-workers demonstrated the effectiveness of a recently deve-
loped SphK1 selective inhibitor in the treatment of an animal
model of leukemia.11 However, another report indicated that
*To whom correspondence should be addressed. Phone: 434-924-
7718. Fax: 434-982-2302. E-mail: tlm@virginia.edu.
aAbbreviations: S1P, sphingosine 1-phosphate; SphK1, sphingosine
kinase type 1; SphK2, sphingosine kinase type 2; DAG, diacylglycerol;
SAR, structure-activity relationship; S1Pn (n = 1 - 5), sphingosine
1-phosphate receptor subtype.
Results
Initial Inhibitor Design. The initial design of substrate-
based SphK1/2 inhibitors required an understanding of
r
pubs.acs.org/jmc
Published on Web 03/05/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2767
previously evaluated SphK substrates. We and others have
documented that the immunomodulatory investigational
drug, fingolimod (FTY720), is inactive until phosphorylated
by Sphk2.15-17 FTY720-P is a potent agonist at the S1P
receptors, most prominently the S1P1 receptor.13 FTY720
has been shown to be efficacious in clinical trials of remitting
relapsing multiple sclerosis.18 We have examined the recep-
tor selectivity and metabolism of a number of classes of
FTY720 analogues.19-22 Despite continuing interest in their
role as cell-signaling entities, the SAR associated with sphin-
gosine analogues as SphK substrates (particularly substrates
of SphK1) has remained largely undefined. However, a
recent study from our laboratories involving the design
and evaluation of a series of heterocyclic amino alcohols as
SphK substrates provided insight into the structural require-
ments necessary for phosphorylation. Although synthetic
analogues of sphingosine phosphorylated by both SphKs
are rare, we previously reported (R)-2-amino-2-(5-(4-octyl-
phenyl)-1,2,4-oxadiazol-3-yl)propan-1-ol (VPC45129) as an
amino-alcohol substrate possessing such activity.23 We pos-
tulated that if the hydroxyl portion of this substrate were
deleted, creating the (S)-deoxy-5-phenyl-1,2,4-oxadiazole
(1), an efficient substrate-based dual inhibitor of the SphKs
could be created (Figure 1).
Synthesis of 1. Envisioning the stereochemistry of the
amine to be readily available from L-alanine, the synthesis
of 1 commenced the N-Cbz protection of the amino acid
(Scheme 1). Aqueous ammonium hydroxide in methanol
were used to covert the methyl ester to the corresponding
amide, although a reaction time of 72 h was necessary for
moderate yields. The primary amide was then converted to
the nitrile 2 through a trifluoroacetic anhydride mediated
dehydration. Hydroxylamine hydrochloride and triethyl-
amine in refluxing ethanol then provided the amide-oxime 3.
In our previously cited study on heterocyclic amino alcohols
as SphK substrates, we showed that amide-oximes such as 3
are excellent substrates for PyBOP mediated couplings, pro-
viding an efficient route to precursors of phenyl-1,2,4-oxadia-
zole systems.23 Utilizing this methodology, 4 was synthesized,
which upon heating in DMF in the presence of molecular
sieves, afforded 5 in more than 80% yield.
The final deprotection of 5 could be achieved using
standard palladium hydrogenolysis conditions, yielding 1
as the free amine. Surprisingly, after 15 h under these
conditions, several products were present by TLC analysis.
Isolation of the major product proved to be more difficult
than anticipated. Thus, only small amounts of material were
recovered from the deprotection of 5 (Scheme 2).
Reductive Cleavage of 1,2,4-Oxadiazole Yielding a Pri-
mary Amidine. On the basis of analytical LCMS spectra,
the mass of 1 was inconsistent with that of the desired
product, indicating a reduction beyond that of the N-Cbz
group had taken place. After a search through the chemical
literature, an example was found showing catalytic palla-
dium hydrogenolysis conditions, similar to those used in the
deprotection of 5, catalyzed the N-O bond cleavage of
phenyl-substituted 1,2,4-oxadiazoles.24 Applying this model
Figure 1. R enantiomer of VPC45129 and its deoxy counterpart, 1.
Scheme 1. Synthesis of the 5-Phenyl-1,2,4-oxadiazole 1
Scheme 2. Reductive Ring-Opening and Rearrangement of 1

2768 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Scheme 3. Synthesis of the Primary Amidinium Chloride 9
Mathews et al.
Table 1. 9 Activity at SphK1 and SphK2
to our system, we hypothesized that a reduction of 1 would
lead to a rapid tautomerization to the acyl-amidine 7. After
bond isomerization to the E isomer, 8, a rearrangement
yielding 9 was possible.
Primary Amidine Synthesis Verifies Rearrangement. To
verify the rearrangement, the primary amidine 9 was inde-
pendently synthesized via a separate synthetic sequence
for comparison of LCMS and NMR spectra (Scheme 3).
Because amidines can be easily generated from a nitrile
precursor,25,26 synthesis of the desired product would be
quite similar to that of 1. Stirring the protected alanine
methyl ester in ammonium hydroxide for 72 h proved an
inefficient means of generating the amide precursor neces-
sary in the synthesis of 2. To circumvent this process, the
N-Boc protection of L-alaninamide hydrochloride initiated
the synthesis of the primary amidine.
Figure 2. SAR regions of 9.
Scheme 4. Synthesis of Head Group Analogues 12 and 13
Deprotection and coupling of 10 proved challenging as
alkylation of the nitrile by the tert-butyl cation generated in
the deprotection was a consistent problem. Limiting the
formation of the alkylated-nitrile side product was achieved
by closely monitoring reaction times of the trifluoroacetic
acid mediated deprotection of 10 and immediately coeva-
porating excess solvent with diethyl ether. Coupling of the
TFA salt of 10 to 4-octylbenzoic acid proceeded in relatively
low yields to form the amide 11. Limited amide formation
was most likely the result of steric hindrance about the
R-carbon of the amine.
As mentioned, an existing literature procedure for the
formation of amidines from nitriles made amidine products
very accessible. However, slow formation of the transient
imidate precursor made mild heating necessary to ensure
moderate product formation. After addition of ammonium
chloride, the amidinium salt 9 was recovered as a white solid
on precipitation with ethyl acetate.
With this in mind, our lead compound was divided into four
distinct regions for SAR evaluation (Figure 2).
Head Group and Linker Region Analogues. The first region
studied was the headgroup of 9. To determine if the unique
chemistry of the amidine was critical to the function of 9, the
moiety was replaced with a carboxylic acid to yield 12 and a
primary amide 13 (Scheme 4).
A small set of analogues was envisioned to probe the steric
and stereochemical requirements for potent SphK inhibition
by amidine-based inhibitors. To better understand the steric
requirements for sphingosine kinase inhibition, 14 was con-
structed through a PyBOP mediated amide coupling and
converted to the corresponding amidine, 15 (Scheme 5).
Inversion of the stereocenter in the linker region of 9 was
easily achieved by starting with D-alanine. Isobutylchloro-
formate (IBCF) followed by ammonia in methanol was
found to be a facile route for the conversion of the carboxylic
acid to the primary amide en route to the nitrile 16.27
The target-directed synthesis of amidine 9, verified the
cleavage and rearrangement of the 5-phenyl-1,2,4-oxadia-
zole based on NMR, LCMS, and HRMS analysis. When this
compound was evaluated at the SphKs (Table 1), it was
found to be a moderately potent dual inhibitor, displaying KI
values in the midmicromolar range.
Structure-Activity Relationship of 9. Realizing the poten-
tial of a dual SphK inhibitor, we focused on optimizing 9.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2769
Scheme 5. Synthesis of Linker Region Analogue 15
Scheme 6. Synthesis of Linker Region Analogue 18
Scheme 7. Synthesis of Linker Region Analogue 20
Scheme 8. (a) Synthesis of Tail Region Analogues 22, 23, 24, and 25; (b) Synthesis of Inverse Amide Tail Region Analogues
27, 28, 29, 30
Subsequent deprotection and coupling afforded the benza-
mide 17, which was easily converted to the amidine resulting
in the R stereoisomer of 9, 18 (Scheme 6).
cyclopropyl derivatives with differing tail lengths were eval-
uated. Amide coupling conditions in the synthesis of these
derivatives were identical to that of 20 (Scheme 8).
Finally, to increase steric bulk in the linker region, a
cyclopropyl ring was installed in place of the methyl group
on 9. Synthesis of this derivative was achieved through the
conversion of 4-octylbenzoic acid to the corresponding
acid chloride, which was immediately coupled to 1-amino-
1-cyclopropanecarbonitrile hydrochloride to provide the
benzamide 19. Despite increased activation of the acyl
chloride, amide formation in the presence of the cyclopropyl
ring was extremely sluggish and yields of 30-40% were
common. On the use of standard conditions, 20 was isolated
as a white solid (Scheme 7).
Finally, the amide region of 9 was inverted to determine
the importance of this portion of the lead molecule. Starting
with an aniline derivative and 1-cyano-1-cyclopropane, Py-
BOP was used to synthesize the corresponding aryl-amide.
Amide coupling reactions to form the reversed amide cyclo-
propyl scaffold were significantly improved over the pre-
viously described amide series, most likely due to a decrease
in steric bulk about the nucleophilic amine. After isolation of
the amide, standard conditions yielded the corresponding
amidines. (Scheme 8a).
Direct in Vitro Evaluation of Sphingosine Kinase Inhibitors
at SphK 1 and 2. We determined the KI values of each inhi-
bitor at recombinant SphK1 and SphK2 (see Experimental
Tail Region and Amide Region Analogues. To determine
the optimal tail length for analogues of 9, a number of

2770 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Mathews et al.
Section for details). As documented in the Table, the KI
values ranged from 0.2 to greater than 100 μM (i.e., no
inhibition observed when inhibitor was included at 50 μM).
The inhibitors appeared to be competitive, that is, inclusion
alkyl amidine, Distamycin A, was reported to have a pKa of
11.6 in water and DMSO,31 we postulated that a highly basic
amidines was critical to SphK inhibition. To test this hypothe-
sis, a carboxylic acid (12) and primary amide (13) derivatives
were synthesized. Both are similar in size and shape to the
amidine but negatively charged or neutral at a physiological
pH. After evaluation at the SphKs, there was no inhibition
observed in either case at concentrations up to 100 μM. We
concluded that the unique electrostatic properties associated
with amidines in an aqueous environment are critical to the
function of 9 and its ilk.
of the inhibitor increased the KM, but did not lower the Vmax
,
values for D-erythro-sphingosine. Data are displayed in
tabular format to show the experimental KI values for the
amidine-based inhibitors at SphK1 and SphK2. We also
show the experimental KI values as a ratio to the experi-
mental KM of sphingosine to indicate inhibitor potency and
selectivity at the sphingosine kinases (Table 2).
In the initial design of 1, we considered a previous study of
the SphKs30 that identified aspartic acid residues in the
substrate-binding domain that are critical to recognition of
the basic amine in sphingosine. As a result, there is a high
degree of stereoselectivity associated with substrates of the
sphingosine kinases including analogues of the pro-drug
FTY720, which are prominent SphK2 substrates.16 Interest-
ingly, D-(þ)-threo-sphingosine, wherein the amine displays
stereochemistry opposite that of the natural substrate, has
been found to be a weak inhibitor of the SphKs,32 Indeed, in
our own study of heterocyclic amino-alcohol substrates, R-
enantiomers of the imidazole analogues were found to be the
most active kinase substrates, reinforcing this concept of
steroselective kinase recognition.23 Presuming the same stereo-
chemical preference existed for the other racemic amino-
alcohols studied, this relative stereochemistry influenced the
stereochemistry of our initial target 1.
After the unexpected palladium catalyzed reductive ring-
opening and rearrangement of 1 to yield the primary amidine,
9, we directed our efforts to the elucidation of the stereochemi-
cal preference of kinase inhibition. Evaluation of the enantio-
mer, 18, at both kinases revealed that inversion of stereo-
chemistry in the linker region significantly influenced the
activity profile, building in selectivity at SphK2 over SphK1.
Although previously demonstrated to be important in neo-
plastic disease,12 SphK2 has been described as being pro-
apoptotic. However, a lack of potent inhibitors of this SphK
isoform prevents further understanding of this enzyme in such
disease models. The unique activity associated with 18 makes
this ligand a candidate for future scaffolds of SphK2 selective
inhibitors to elucidate the role of this kinase in disease.
Understanding that inversion of stereochemistry affected
kinase selectivity, we next directed our attention to determin-
ing how the deletion of stereochemistry affected the potency
and selectivity of our inhibitors. Such an approach would
invariably impact the steric bulk in the linker region of this
class of SphK inhibitors. The achiral 15 was devoid of steric
bulk in the linker region and displayed a substantial loss
of kinase inhibition at SphK1, although weak inhibition at
SphK2 was retained. Eliminating chirality and increasing
steric bulk in the case of the 20 resulted in enhanced potency
atSphK1withnochange in SphK2, thus improving activity as
a dual inhibitor. After establishing the influence of sterics in
the linker region on SphK inhibition, this scaffold was carried
forward in the analysis of tail region SAR.
To test for specificity, the two most potent inhibitors (23,
28) were tested at several recombinant diacylglycerol kinases
(types γ, δ1, ζ) and no inhibition was observed at concentra-
tions up to 100 μM (not shown). Many sphingosine analo-
gues, such as DMS, have been shown to act as inhibitors of
members of the protein kinase C family of enzymes. Thus, we
evaluated the same two inhibitors at PKCR and observed no
inhibition at concentrations up to 10 μM.
In Vitro Evaluation of SphK Inhibitors in Smooth Muscle
Cells. SphK1 has been described as an immediate-early
response gene in many cell types. In vascular smooth muscle
cells, activation of the S1P1 receptor by S1P can result in
smooth muscle cell proliferation and growth. Indeed, pre-
vious studies with compounds developed by our group to
target S1P receptors documented that selective inhibition of
the S1P1 receptor prevented S1P-induced vascular smooth
muscle cell proliferation/growth in vitro and in vivo.28 In
parts A and B of Figure 3, respectively, we show that serum
(10% FBS) induces vascular smooth muscle cell Sphk1
mRNA and SphK1 activity while SphK2 mRNA and acti-
vity was unchanged. Serum stimulation also increases
smooth muscle cell proliferation and growth. We tested the
hypothesis that serum-induced smooth muscle cell prolifera-
tion was regulated in part by Sphk1. Cells were pretreated
with 9 and the enantiomer 18 for 30 min and then stimulated
with 10% serum for 24 h. 9 blocked serum-induced changes
in S1P intracellular concentration (Figure 3C). 9, not 18, also
blocked smooth muscle cell proliferation (EC50 ∼10 μM,
Figure 3D). As documented in Figure 3E, 9 was not cytotoxic
to the cells on visual microscopic inspection. A cell viability
assay was performed; this revealed that 9 did not induce
cytotoxicity/death (Figure 3F).
Discussion
The sphingosine kinases control, in part, the cellular equi-
librium of the pro-survival lipid S1P and its pro-apoptotic
precursor ceramide, and the ratio of these two contrasting
metabolites has been proposed to be critical to the prolifera-
tion, survival, and death of cells.29 This hypothesis predicts a
role for the SphKs in hyperproliferative disease models. The
highly inducible nature of SphK1 mRNA and enzyme activity
has focused most attention on this SphK type,8 but a recent
literature report indicates that SphK2 is the key kinase iso-
form in macrophage polarization to an anticancer pheno-
type.12 Such contrasting data suggests that dual and selective
SphK inhibitors are needed to better characterize the relative
role of these lipid kinase types.
Realizing the tail region of our lead compound was sig-
nificantly smaller than that of the natural substrate, sphingo-
sine, we postulated that increasing linear size in this region of
the molecule would optimize kinase inhibition. The amidine
23, having a tail region 12 carbon atoms with the cyclopropyl
head group, displayed nanomolar potencies at both kinases.
Comparing the KI/KM ratios of 23 at SphK1 and 2 shows only
slight selectivity toward SphK1. The amidines 24 and 25
Perhaps the most distinct structural aspect of 9 was the
amidine head group. We immediately wished to understand
the role this unique functional group played in sphingosine
kinase inhibition. Realizing that basic amines are critical for
substrate recognition by the SphKs,30 and that a well-known

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2771
Table 2. The Binding Constants for D-erythro-Sphingosine Were Measured in the Presence (K0M) or Absence (KM) of a Fixed Concentration (0.1-50
μM) of a VPC Compound (I)a
a KI values were calculated using the equation: KI = [I]/(K0M/KM - 1).
demonstrated the limits to alkyl chain length, as these analo-
gueshaving 14and 16 carbonatomsdisplayeda steepdecrease
in inhibitory activity. Such a result is mostlikely due to the size
of the hydrophobic pocket in the substrate-binding domain of
the kinases.
Inversion of the benzamide region of the molecule provided
the most compelling results in terms of kinase selectivity.
Overall, a similar trend of alkyl group length and potency
was observed with the inverted benzamides. Potency at the
SphKs was increased at a length of 12 carbons but decreased

2772 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Mathews et al.
Figure 3. Inhibition of Sphk1 with 9 prevents serum-induced smooth muscle cell proliferation. (A) Quantitative real-time PCR analysis of
Sphk1 and Sphk2 mRNA normalized to 18S mRNA following 4 h of serum (10% FBS) stimulation (N = 3). (B) Sphk1 and Sphk2 activity
([32P]-S1P) following 0, 3, 6, and 12 h of 10% FBS stimulation (N = 3). (C) Changes in intracellular S1P concentration (C17 S1P) measured by
LC/MS following 12 h of 10% FBS stimulation (N = 2). (D) Dose-response of 18 and 9 on smooth muscle cell incorporation of BrdU in
response to 24 h 10% FBS stimulation (N = 6). (E) Light microscopy images (20ꢀ magnification) in control, 10% FBS-treated and 10% FBS þ
9 (10 μM). (F) Cytotoxic response to 9 (0.001-100 μM) and 20 μM cyclosporine A (CSA) as measured by the Promega MultiTox-Fluor assay
(N = 3). * denotes P < 0.05, N denotes biological replicate; mean ( standard deviation.
with 14 and 16 carbons. Most intriguing was the selectivity
this series displayed to SphK1. The amidine 28, the 12-carbon
analogue, displayedadramatic degreeof selectivityforSphK1
with potencies of 0.3 μM at SphK1 and 6 μM at SphK2.
Comparing the KI/KM ratios of 28 at SphK1 and 2 showed
that the analogue displayed 40-fold selectivity for SphK1 over
SphK2. Much like 18, 28 displayed unforeseen selectivity that
could serve as a template for future classes of selective,
amidine-based SphK inhibitors. Such compounds, if active
in vivo, could be useful in testing hypotheses of the relative
roles of SphK types in hyperproliferative diseases.
cellular levels of sphingolipids, a group that exists in a myriad
of forms, are typically considered to be approximately an
order of magnitude higher.34 Not surprisingly, the observed
KM of the sphingosine kinases for sphingosine is approxi-
mately 10 μM and 5 μM for SphK135 and SphK2,36 respec-
tively. Consequently, previously described selective SphK
inhibitors exhibiting seemingly poor KI values have been
effective at lowering cellular S1P production and prolifera-
tion. While such inhibitors have proven quite effective at
elucidating the role of the sphingosine kinases in various
disease states, the relatively high dosages required for such
inhibitors renders their therapeutic index unfavorable. At the
outset of this study, we hoped to identify a new class of
Plasma levels of S1P are reported as being in the high
nanomolar range, typically between 200 and 800 nM.33 While

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2773
sphingosine kinase inhibitors that would be effective both in
cells and in vivo with a therapeutic index that would eliminate
the likelihood of drug toxicity. While the amidine-containing
compounds identified in this study have improved potency
over reported sphingosine kinase inhibitors, they are largely
charged at physiologic pH, leading us to consider whether
such molecules could penetrate cells. Evaluation of our lead
compound 9 at vascular smooth muscle cells has demon-
strated that amidine-based sphingosine kinase inhibitors are
effective at lowering overall S1P levels in these cells, correlat-
ing with a decrease in cell proliferation. Currently, efforts are
underway to better characterize the more potent amidine-
based inhibitors in a number of disease models.
elevated temperature. When the reaction appeared complete by
TLC, the mixture was allowed to cool to ambient temperature
and the solution was evaporated to dryness. The crude material
was next taken up in chloroform and filtered through a fine
fritted funnel. The eluent was collected and thoroughly dried
under vacuum. Finally, the reaction mixture was taken up in
ethyl acetate and filtered again through a fine fritted funnel; the
product was recovered as a white or off-white crystalline solid.
General Procedure B: PyBOP Mediated Couplings of Amines,
Anilines, and Amide Oximes to Carboxylic Acids. To a suspen-
sion of an amine (0.43 mmol) in anhydrous DCM (4.3 mL) was
added the acid (0.43 mmol) and PyBOP (0.43 mmol). Finally,
diisopropylethylamine (1.72 mmol) was added and the reaction
was allowed to stir for 15 h. At this time, the reaction mixture
was evaporated to dryness and reconstituted in 100 mL of
EtOAc. The solution was extracted with four 10 mL portions
of 1N HCl followed by one 10 mL portion of brine. The organic
layer was dried with MgSO4 and evaporated to dryness. The
crude organic material was purified by flash chromatography.
General Procedure C: Conversion of Carboxylic Acids to Acyl
Chlorides. A carboxylic acid (0.28 mmol) and a catalytic amount
of dimethylformamide were dissolved in dichloromethane
(2.8 mL), and the solution was cooled to 0 ꢀC. Oxalyl chloride
(0.84 mmol) was added dropwise, and the reaction mixture was
warmed to room temperature. After stirring for 2 h, the solution
was evaporated to dryness; the crude material was taken up in
three 1 mL portions of diethyl ether and immediately evapo-
rated. The oil was dried under vacuum for an additional 30 min
and carried on for alkylation by an amine.
General Procedure D: Alkylation of Acyl Chlorides by Amines.
To a stirring solution of a primary amine (1.68 mmol) in DCM
(8.5 mL) was added an acyl chloride (2.184 mmol) as a solution
in an additional volume of DCM (8.5 mL). The mixture was
then cooled to 0 ꢀC, and diisopropylethyl amine (5.04 mmol)
was added dropwise. The reaction mixture was allowed to warm
to ambient temperature slowly and stirred for an additional
15 h. At this time, the reaction mixture was evaporated to dryness
and reconstituted in ethyl acetate (150 mL). The solution was
extracted with three 15 mL portions of 1N HCl followed by one
15 mL portion of brine. The organic layer was then dried with
MgSO4, filtered through a fritted funnel, and dried to an oil. The
crude mixture was purified via flash chromatography.
General Procedure E: Conversion of a Primary Amide to a
Nitrile. A primary amide (9.3 mmol) and triethylamine (18.9
mmol) was taken up in 93 mL of anhydrous THF and cooled to
0 ꢀC. Upon cooling, trifluoroacetic anhydride (11.16 mmol) was
added dropwise to the reaction mixture; after 10 min of stirring
at this low temperature, the reaction mixture was allowed to
warm to ambient temperature and evaporated to dryness. The
crude organic mixture was taken up in 250 mL of EtOAc and
extracted with four 15 mL portions of 1N HCl followed by one
10 mL portion of brine. The organic layer was then dried with
MgSO4 and evaporated to an oil.
General Procedure F: Trifluoroacetic Acid Mediated Depro-
tection of N-Boc Protected Amines. To a stirring solution of the
N-Boc protected amino-nitrile (1.2 mmol) in anhydrous DCM
(12 mL) was added TFA, dropwise (12 mL) at room tempera-
ture. After 15 min, the reaction mixture was evaporated to
dryness. Three 5 mL portions of diethyl ether were then added
and immediately evaporated to yield the trifluoroacetate salt as
an off-white solid. The salts were dried under high vacuum for a
minimum of 1 h before being carried forth.
Conclusions
At the outset of this study, we sought to synthesize a deoxy
5-phenyl-1,2,4-oxadiazole derivative of a previously described
SphK1 and SphK2 dual substrate. After an unexpected
palladium catalyzed ring-opening and rearrangement, a new
amidine-based sphingosine kinase inhibitor, 9, was developed
that displayed moderate potencies as an inhibitor of SphK1
and SphK2. Demonstration of the effectiveness of a moderate
inhibitor in a smooth muscle cell based hyperproliferative
disease model illustrates the potential of this new class of
sphingosine kinase inhibitors to answer long-standing ques-
tions on the pharmacology of the sphingosine kinases. We
have identified a potent dual SphK inhibitor (23), a potent
selective SphK1 inhibitor (28), and a moderately potent
SphK2 selective inhibitor (18). Most importantly, we have
begun to identify the pharmacophore associated with ami-
dine-based sphingosine kinase inhibitors that we hope to
exploit in future SphK inhibitors active in vivo.
Experimental Section
General Synthetic Materials and Methods. All nonaqueous
reactions were carried out in oven or flame-dried glassware
under an argon or nitrogen atmosphere with dry solvents and
magnetic stirring, unless otherwise stated. The argon and nitro-
gen were dried by passing through a tube of Drierite. Anhydrous
diethyl ether (Et2O), toluene, dichloromethane (CH2Cl2),
methanol (MeOH), tetrahydrofuran (THF), and N,N-dimethyl-
formamide (DMF) were purchased from Aldrich or VMR
Chemicals and used as received. THF was dried over activated
˚
molecular sieves (4 A) prior to use. All other reagents were
purchased from Acros chemicals and Aldrich chemicals. Except
as indicated otherwise, reactions were monitored by thin layer
chromatography (TLC) using 0.25 mm Whatman precoated
silica gel plates. Flash chromatography was performed with the
indicated solvents and Dynamic Adsorbents silica gel (particle
size 0.023-0.040 mm). Proton (1H) and carbon (13C) NMR
spectra were recorded on a Varian UnityInova 500/51 or Varian
UnityInova 300/54 at 300 K unless otherwise noted. Chemical
shifts are reported in ppm (δ) values relative to the solvent as
follows: CDCl3 (δ 7.24 for proton and δ 77.0 for carbon NMR),
DMSO-d6 (δ 2.50 for proton and δ 39.5 for carbon NMR),
CD3OD (δ 3.31 for proton and δ 47.6 for carbon NMR). All
high-resolution mass spectrometry was carried out by the High-
Resolution Mass Spectrometry Facility in the School of Che-
mical Sciences at the University of Illinois Urbana;Cham-
pagne (Urbana, IL).
(S)-1-(5-(4-Octylphenyl)-1,2,4-oxadiazol-3-yl)ethanamine (1).
Compound 5 (0.43 mmol) was dissolved in wet ethanol (20 mL)
and to this solution was added approximately 200 mg of
activated palladium on carbon (30% w/w). The ambient atmo-
sphere was then replaced with hydrogen and stirred for 15 h.
After this time, the reaction mixture was filtered through celite
and concentrated to an oil. After purification by column chro-
matography (15-20% MeOH in CHCl3), 1 was recovered as an
General Procedure A: Conversion of Nitriles to Amidines. A
nitrile (0.21 mmol) was taken up in 2.1 mL of anhydrous
methanol; 42 μL of a solution of 0.5 M sodium methoxide in
methanol was immediately added. This mixture was allowed to
stir for 15 h at 43 ꢀC, at which time 0.231 mmol of ammonium
chloride was added to the reaction, while still stirring at this

2774 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Mathews et al.
off-white solid (0.06 mmol). 1H NMR (500 MHz, DMSO) δ 9.05
(bs, 3H), 8.94 (d, J = 6.2, 1H), 7.91 (d, J = 8.1, 2H), 7.28 (d, J =
7.9, 2H), 4.75-4.64 (m, 1H), 2.61 (t, J = 7.6, 2H), 1.56 (dd, J =
6.9, 14.1, 1H),1.50 (d, J = 7.2, 3H), 1.31-1.14 (m, 10H), 0.83
(t, J = 6.9, 3H). 13C NMR (126 MHz, DMSO) δ 172.94, 166.97,
146.99, 130.96, 128.53, 128.36, 47.57, 40.45, 40.28, 40.11,
39.95, 39.78, 39.61, 39.45, 35.40, 31.71, 31.18, 29.24, 29.11,
29.04, 22.52, 18.83, 14.41. LCMS and HRMS data was identi-
cal to 9.
(S)-Benzyl 1-Cyanoethylcarbamate (2). General procedure E
was used to convert 5.7 mmol of N-Cbz-alaninamide to the title
product. After flash chromatography, 4.9 mmol of product was
recovered. 1H NMR (500 MHz, CDCl3) δ 7.50-7.30 (m, 5H),
5.15 (s, 2H), 4.75-4.61 (m, 1H), 1.56 (d, J = 7.2, 3H). 13C NMR
(126 MHz, CDCl3) δ 154.87, 135.50, 128.66, 128.53, 128.32,
119.14, 67.72, 38.08, 19.55.
(S,Z)-Benzyl 1-Amino-1-(hydroxyimino)propan-2-ylcarbamate
(3). The intermediate 2 (4.7 mmol) was dissolved in anhydrous
ethanol (7.5 mL). To this solution was added triethylamine
(10.81 mmol) and hydroxylamine hydrochloride (10.34 mmol).
The reaction mixture was heated to reflux for three hours, after
which time it was concentrated to an oil and reconstituted in
dichloromethane (150 mL). The organic layer was washed with
three 10 mL portions of water and one 10 mL portion of brine.
The organic solution was dried of MgSO4 and evaporated to
dryness. Purification by flash chromatography yielded 2.3 mmol
of 3 as an off-white solid. 1H NMR (500 MHz, CD3OD) δ
7.45-7.22 (m, 5H), 5.07 (s, 2H), 4.21 (q, J = 7.0, 1H), 1.34 (d,
J = 7.1, 3H). 13C NMR (126 MHz, CD3OD) δ 156.56, 156.27,
136.78, 128.04, 127.58, 127.45, 66.13, 18.11.
(R,Z)-Benzyl 1-Amino-1-(4-octylbenzoyloxyimino)propan-2-
ylcarbamate (4). General procedure B was used to couple 4
(2.1 mmol) to 4-octylbenzoic acid. After column chromatogra-
phy, 1.6 mmol of the title product was recovered as a white solid.
1H NMR (300 MHz, CDCl3) δ 7.93 (d, J = 8.2, 2H), 7.35
(s, 5H), 7.25 (d, J = 7.2, 3H), 5.41-5.22 (m, 3H), 5.12 (s, 2H),
4.5-3.8 (m, 1H), 2.71-2.61 (m, 2H), 1.69-1.53 (m, 5H),
1.31-1.26 (m, 9H), 0.86 (t, J = 6.7, 3H). 13C NMR (126
MHz, CDCl3) δ 164.08, 159.49, 156.77, 148.73, 135.92, 129.45,
128.55, 128.30, 128.07, 126.83, 67.25, 48.22, 36.03, 31.85, 31.15,
29.42, 29.26, 29.22, 22.65, 17.66, 14.11.
(S)-tert-Butyl 1-Cyanoethylcarbamate (10). Sodium carbo-
nate (1.5 g) was dissolved in 15 mL of water, and to this solution
was added alaninamide hydrochloride (5.7 mmol). In a separate
flask, di-tert-butyl dicarbonate (11.4 mmol) was dissolved in
1,4-dioxanes (11.4 mL). This organic solution was then added to
the aqueous solution and allowed to stir for 15 h. After this time,
the reaction mixture was diluted with 25 mL of 1N HCl and
extracted with four 25 mL portions of ethyl acetate. The organic
layers were combined, washed with brine, dried over MgSO4,
and concentrated to a white solid. This solid was then subjected
to general procedure E. After flash chromatography, 3.1 mmol
of the title product was recovered. 1H NMR (500 MHz, CDCl3)
δ 4.82 (bs, 1H), 4.62 (bs, 1H), 1.54 (d, J = 7.2, 3H), 1.46 (s, 8H).
13C NMR (126 MHz, CDCl3) δ 154.10, 119.53, 37.57, 28.21,
19.62.
(S)-N-(1-Cyanoethyl)-4-octylbenzamide (11). General proce-
dure F was used to deprotect 10 (1.17 mmol). After standard
purification techniques, the off-white solid was coupled to 4-
octylbenzoic acid (1.17 mmol) using general procedure B. Flash
chromatography yielded 0.5 mmol of the title product. 1H NMR
(500 MHz, CDCl3) δ 7.70 (d, J = 8.1, 2H), 7.25 (dd, J = 4.0, 7.8,
2H), 6.56 (bs, 1H), 5.16 (p, J = 7.2, 1H), 2.64 (t, J = 6.4, 2H),
1.65 (ddd, J = 9.5, 10.2, 19.7, 5H), 1.42-1.15 (m, 10H), 0.87 (dd,
J = 5.9, 7.0, 3H). 13C NMR (126 MHz, CDCl3) δ 166.58,
148.04, 129.94, 128.80, 127.18, 119.40, 36.24, 35.85, 31.83, 31.13,
29.39, 29.21, 22.63, 19.58, 14.08.
(S)-2-(4-Octylbenzamido)propanoic Acid (12). General proce-
dure B was used to couple 0.31 mmol of L-alanine methyl ester to
4-octylbenzoic acid. After standard work up procedures, the
recovered material was dissolved in 5 mL of isopropyl alcohol.
In a separate flask, 0.62 mmol of sodium hydroxide was
dissolved in 2 mL of water and poured into isopropyl alcohol
solution. The mixture was heated to 60 ꢀC for 3 h, at which time
the reaction mixture was cooled to ambient temperature and
evaporated to dryness. The crude material was taken up in
50 mL of a 4:1 suspension of ethyl acetate to 1N HCl. The
aqueous layer was extracted three additional times with ethyl
acetate. The organic layers were combined, washed with brine,
and dried over magnesium sulfate. Evaporation of solvent
yielded 0.23 mmol of the title product as an off-white solid. 1H
NMR (300 MHz, CDCl3) δ 7.71 (d, J = 8.2 Hz, 2H), 7.31-7.16
(m, 3H), 6.80 (d, J = 6.9 Hz, 1H), 4.81-4.77 (m, 1H), 2.76-2.51
(m, 2H), 1.57 (d, J = 7.1 Hz, 5H), 1.27 (d, J = 7.3 Hz, 11H), 0.87
(t, J = 6.7 Hz, 3H). 13C NMR (75 MHz, CDCl3) δ 176.41,
168.48, 147.81, 131.10, 129.06, 127.39, 48.95, 36.08, 32.07, 30.81,
29.24, 23.00, 18.37, 14.32.
(S)-Benzyl 1-(5-(4-Octylphenyl)-1,2,4-oxadiazol-3-yl)ethyl-
carbamate (5). The intermediate 4 (1.5 mmol) was dissolved in
˚
DMF (30 mL). One g of activated 4 A molecular sieves were
added and the reaction mixture was heated to 110 ꢀC. After 6 h,
the mixture was cooled to ambient temperature and filtered
through a medium frit. After diluting with 250 mL of ethyl
acetate, the solution was extracted with eight 15 mL portions of
water and one 15 mL portion of brine. The organic layer was
collected, dried over MgSO4, and evaporated to dryness. Puri-
(S)-N-(1-Amino-1-oxopropan-2-yl)-4-octylbenzamide (13).
General procedure B was used to couple 0.42 mmol of
L-alaninamide hydrochloride to 4-octylbenzoic acid. After stan-
dard work up and purification procedures, 0.23 mmol of the title
1
1
fication by flash chromatography yielded 1.14 mmol of 5. H
product was recovered. H NMR (500 MHz, CD3OD) δ 7.79
NMR (300 MHz, CDCl3) δ 8.01 (d, J = 8.3, 2H), 7.41-7.28 (m,
7H), 5.42 (d, J = 8.6, 1H), 5.26-5.04 (m, 2H), 2.78-2.60 (m,
2H), 1.74-1.51 (m, 5H), 1.32-1.25 (m, 10H), 0.88 (t, J = 6.8,
3H). 13C NMR (126 MHz, CDCl3) δ 176.26, 171.80, 155.46,
148.69, 136.21, 129.17, 128.53, 128.14, 121.37, 67.01, 44.24,
36.08, 31.84, 31.09, 29.41, 29.22, 22.66, 20.72, 14.12.
(d, J = 8.5 Hz 2H), 7.28 (d, J = 8.2 Hz, 2H), 4.56 (q, J = 7.2 Hz,
1H), 2.67 (t, J = 7.5 Hz, 2H), 1.66-1.60 (m, 2H), 1.47 (d, J = 7.2
Hz, 3H), 1.37-1.19 (m, 10H), 0.89 (t, J = 7.0 Hz, 3H). 13C
NMR (126 MHz, CD3OD) δ 176.59, 168.54, 147.05, 131.15,
128.13, 127.16, 49.26, 35.32, 31.59, 28.97, 22.29, 16.88, 13.00.
N-(Cyanomethyl)-4-octylbenzamide (14). General procedure
B was used to couple aminoacetonitrile bisulfate (0.42 mmol) to
4-octylbenzoic acid. After flash chromatography, 0.38 mmol of
the title product was recovered. 1H NMR (300 MHz, CDCl3) δ
7.70 (d, J = 8.3, 2H), 7.25 (d, J = 7.6, 2H), 6.78 (t, J = 5.5, 1H),
4.37 (d, J = 5.8, 2H), 2.64 (t, J = 9, 2H), 1.62 (dt, J = 7.4, 22.1,
2H), 1.3-1.26 (m, 10H), 0.87 (t, J = 6.7, 2H). 13C NMR (75
MHz, CDCl3) δ 174.80, 148.82, 130.03, 129.08, 127.46, 117.30,
77.67, 77.25, 76.83, 36.09, 32.06, 31.35, 29.62, 29.44, 28.19,
22.86, 14.31.
(S)-N-(1-Amino-1-iminopropan-2-yl)-4-octylbenzamide Hydro-
chloride (9). General procedure A was used to convert 0.24 mmol
of 12 to the title product. After previously described recrystalli-
zation techniques, 0.078 mmol of a white powdery product was
1
recovered. H NMR (500 MHz, DMSO) δ 8.91 (bs, 1H), 8.66
(d, J = 5.2, 1H), 7.86 (d, J = 8.1, 2H), 7.29 (d, J = 8.1, 3H), 7.17
(bs, 1H), 4.65-4.47 (m, 1H), 2.61 (t, J = 7.6, 2H), 1.56 (dd, J =
6.9, 14.1, 2H), 1.47 (d, J = 7.2, 3H), 1.32-1.14 (m, 10H), 0.83
(t, J = 6.8, 3H). 13C NMR (126 MHz, DMSO) δ 172.75, 167.26,
147.07, 130.91, 128.54, 128.35, 47.91, 35.39, 31.70, 31.21, 29.23,
29.02, 22.52, 18.64, 14.41. LCMS: tR = 8.4 min; m/z = 304, 607
(dimer). HRMS m/z cald for C18H30N3O (M þ H), 304.2389;
found, 304.2384.
N-(2-Amino-2-iminoethyl)-4-octylbenzamide Hhydrochloride
(15). General procedure A was used to convert 0.38 mmol of
14 to the title product. Upon isolation with standard purifica-
tion techniques, 0.16 mmol of the title product was recovered as

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2775
a white solid. 1H NMR (500 MHz, DMSO) δ 8.94 (s, 1H), 7.82
(d, J = 8.0, 4H), 7.29 (d, J = 7.9, 2H), 4.15 (d, J = 5.2, 2H), 2.61
(t, J = 7.4, 2H), 1.68-1.47 (m, 2H), 1.25-1.21 (m, J = 10H),
0.83 (t, J = 6.4, 3H). 13C NMR (126 MHz, DMSO) δ 168.79,
167.52, 147.03, 130.99, 128.61, 128.17, 35.39, 31.70, 31.17, 29.23,
29.10, 29.02, 22.51, 14.40. LCMS: tR = 6.9 min; m/z = 290.
HRMS m/z calculated for C17H28N3O (M þ H), 290.2232;
found, 290.2232.
(R)-tert-Butyl 1-Cyanoethylcarbamate (16). To a stirring so-
lution of N-Boc-D-alanine (7.9 mmol) in dichloromethane
(16 mL) was added triethylamine (35.55 mmol). This solution
was cooled to -25 ꢀC, and isobutyl chloroformate (15.8 mmol)
was added dropwise. The reaction was stirred at this low
temperature for 45 min, at which time ammonia (23.7 mmol)
as a 7N solution in methanol was added and the reaction
mixture and was allowed to warm to ambient temperature and
stirred for an additional 15 h. After this time, the reaction
mixture was concentrated to an oil by rotary evaporation and
reconstituted in ethyl acetate (200 mL). The organic layer was
washed with four 20 mL portions of 1N HCl and one 15 mL
portion of brine and dried over MgSO4. The solvent was
evaporated, and the intermittent product was dried under
vacuum. General procedure E was then used to convert the
primary amide to a nitrile. Four mmol of the title product was
recovered as a white powdery solid. 1H and 13C spectral data was
identical to that of 10.
solid. 1H NMR (500 MHz, CDCl3) δ 7.70 (dd, J = 3.2, 8.1, 2H),
7.21 (d, J = 8.0, 2H), 2.66-2.58 (m, 2H), 1.59 (d, J = 5.6, 4H),
1.36 (q, J = 5.8, 2H), 1.32-1.20 (m, 15H), 0.87 (t, J = 6.9, 3H).
13C NMR (126 MHz, CDCl3) δ 168.16, 148.04, 129.96, 128.74,
127.32, 120.30, 35.86, 31.88, 31.15, 29.59, 29.45, 29.31, 29.23,
22.67, 20.88, 16.91, 14.12.
N-(1-Cyanocyclopropyl)-4-dodecylbenzamide (21; C12). Gene-
ral procedure C was used to convert 4-dodecylbenzoic acid
(0.52 mmol) to the corresponding acyl chloride. After standard
work up procedures, the intermediate was immediately sub-
jected to conditions described in general procedure D in the
presence of 1-amino-1-cyclopropanecarbonitrile hydrochloride
(0.4 mmol). After flash chromatography, 0.15 mmol of the title
product was recovered. 1H NMR (500 MHz, CDCl3) δ 7.68 (d,
J = 7.1, 2H), 7.23 (d, J = 7.5, 2H), 6.82 (s, 1H), 2.63 (t, J = 7.7,
2H), 1.65-1.57 (m, 4H), 1.35 (t, J = 7.0, 2H), 1.29-1.25 (m,
17H), 0.88 (t, J = 6.5, 3H). 13C NMR (126 MHz, CDCl3) δ
167.81, 148.06, 130.01, 128.75, 127.19, 120.07, 35.85, 31.89, 31.11,
29.61, 29.53, 29.42, 29.32, 29.20, 22.66, 20.88, 16.97, 14.08.
N-(1-Cyanocyclopropyl)-4-tetradecylbenzamide (21; C14).
Tetradecanoic acid (0.61 mmol) was subjected to the methods
described in general procedure C. After standard work up
procedures were used, the intermediate acyl chloride was
coupled to 1-amino-1-cyclopropanecarbonitrile hydrochloride
(0.47 mmol) using general procedure D. After flash chromato-
1
graphy, 0.28 mmol of the title product was isolated. H NMR
(R)-N-(1-Cyanoethyl)-4-octylbenzamide (17). General proce-
dure F was used to deprotect 16 (1.7 mmol). General procedure
B was then used to couple the resulting amine to 4-octylbenzoic
acid. After standard purification techniques, the title product
(0.85 mmol) was recovered. 1H and 13C spectral data were
identical to that of 11.
(300 MHz, CDCl3) δ 7.68 (d, J = 8.2, 2H), 7.23 (d, J = 8.2, 2H),
6.85 (s, 1H), 2.63 (t, J = 9, 2H), 1.64-1.55 (m, 4H), 1.35 (dd, J =
5.8, 8.5, 2H), 1.25 (s, 20H), 0.87 (t, J = 6.6, 3H). 13C NMR
(75 MHz, CDCl3) δ 168.05, 148.30, 130.19, 128.98, 127.43,
120.32, 36.09, 32.15, 31.38, 29.88, 29.68, 29.58, 29.45, 22.92,
21.10, 17.20, 14.36.
(R)-N-(1-Amino-1-iminopropan-2-yl)-4-octylbenzamide Hydro-
chloride (18). General procedure A was used to convert 17 (0.38
mmol) to the title product. After standard purification techni-
N-(1-Cyanocyclopropyl)-4-hexadecylbenzamide (21; C16).
General procedure C was used to convert hexadecanoic acid
(0.58 mmol) to the corresponding acyl chloride. After standard
work up techniques, the intermediate was coupled to 1-amino-1-
cyclopropanecarbonitrile hydrochloride using the methods de-
scribed in general procedure D. Upon flash chromatography
ques, 18 (0.15 mmol) was recovered as a white solid. 1H and 13
C
spectral data were identical to that of 9. LCMS: tR = 8.4 min;
m/z = 304. HRMS m/z calculated for C18H30N3O (M þ H),
304.2389; found, 304.2382.
1
0.41 mmol of the title product was recovered. H NMR (500
N-(1-Cyanocyclopropyl)-4-octylbenzamide (19). General pro-
cedure C was used to convert 4-octylbenzoic acid (0.53 mmol) to
the corresponding acyl chloride. After standard workup proce-
dures, the acyl-chloride was coupled to 1-amino-1-cyclopropyl-
carbonitrile hydrochloride (0.41 mmol) using general procedure
D. After purification by flash chromatography, the title product
(0.21 mmol) was recovered. 1H NMR (500 MHz, CDCl3) δ 7.69
(d, J = 8.3, 2H), 7.21 (d, J = 8.4, 2H), 7.12 (s, 1H), 2.68-2.58
(m, 2H), 1.59 (dd, J = 5.8, 8.3, 4H), 1.35 (dd, J = 5.9, 8.3, 2H),
1.32-1.20 (m, 10H), 0.87 (t, J = 7.0, 3H). 13C NMR (126 MHz,
CDCl3) δ 168.06, 148.02, 129.96, 128.73, 127.27, 120.23, 35.85,
31.83, 31.13, 29.39, 29.21, 22.63, 20.88, 16.91, 14.08.
N-(1-Carbamimidoylcyclopropyl)-4-octylbenzamide Hydro-
chloride (20). General procedure A was used to convert 15
(0.17 mmol) to the corresponding amidine. After standard
workup and purification techniques 20 (0.04 mmol) was reco-
vered as a white solid. 1H NMR (500 MHz, CD3OD) δ 7.82 (d,
J = 8.3, 2H), 7.30 (d, J = 8.2, 2H), 2.76-2.62 (m, 2H), 1.75 (dd,
J = 5.9, 8.5, 2H), 1.64 (dd, J = 7.1, 14.3, 2H), 1.56 (dd, J = 6.1,
8.4, 2H), 1.29 (dd, J = 7.1, 15.0, 11H), 0.89 (t, J = 7.0, 3H). 13C
NMR (126 MHz, CD3OD) δ 172.52, 169.94, 147.79, 130.21,
128.19, 127.47, 35.32, 32.59, 31.58, 30.98, 29.11, 28.97, 28.82,
22.29, 18.16, 12.99. LCMS: tR = 8.5 min; m/z = 316. HRMS
m/z calcd for C19H30N3O (M þ H), 316.2389; found, 316.2381.
N-(1-Cyanocyclopropyl)-4-decylbenzamide (21; C10). Gene-
ral procedure C was used to convert 2.184 mmol of 4-decylben-
zoic acid to the corresponding acyl chloride. After work up, this
intermediate was immediately subjected to methods described in
general procedure D in the presence of 1-amino-1-cyclopropa-
necarbonitrile hydrochloride (1.68 mmol). After flash chroma-
tography, 1 mmol of the title product was recovered as a white
MHz, CDCl3) δ 7.68 (d, J = 8.2, 2H), 7.23 (d, J = 8.1, 2H), 6.81
(s, 1H), 2.63 (t, J = 7.7, 3H), 1.64-1.58 (m, 4H), 1.35 (q, J = 4.9,
2H), 1.33-1.21 (m, 27H), 0.88 (t, J = 6.9, 3H). 13C NMR (126
MHz, CDCl3) δ 167.79, 148.06, 130.01, 128.75, 127.19, 120.07,
97.26, 35.84, 31.90, 31.11, 29.66, 29.63, 29.54, 29.43, 29.33,
29.20, 22.67, 20.88, 16.96, 14.08.
N-(1-Carbamimidoylcyclopropyl)-4-decylbenzamide Hydro-
chloride (22). General procedure A was used to convert 21;
C10 (0.12 mmol) to the corresponding amidine. After standard
recrystallization techniques, 0.026 mmol of the title product was
1
recovered. H NMR (500 MHz, CD3OD) δ 7.81 (d, J = 6.1,
2H), 7.30 (d, J = 6.3, 2H), 2.67 (t, J = 6.7, 2H), 1.75 (s, 2H), 1.63
(s, 2H), 1.56 (s, 2H), 1.29 (d, J = 20.9, 14H), 0.89 (dd, J = 4.5,
7.0, 3H). 13C NMR (126 MHz, CD3OD) δ 172.51, 169.93,
147.78, 130.20, 128.34, 128.04, 127.48, 35.32, 32.59, 31.63,
30.97, 29.27, 28.81, 28.66, 22.30, 18.15, 12.98. LCMS: tR =
10.34; m/z = 344. HRMS m/z calcd for C21H34N3O (M þ H),
344.2702; found, 344.2698.
N-(1-Carbamimidoylcyclopropyl)-4-dodecylbenzamide Hydro-
chloride (23). General procedure A was used to convert 21;
C12 (0.15 mmol) to the corresponding amidine. Upon pre-
viously described recrystallization procedures 0.06 mmol of
1
the title product was recovered. H NMR (500 MHz, DMSO)
δ 9.08 (s, 1H), 8.66 (bs, 4H), 7.81 (d, J = 6.7, 2H), 7.27 (d, J =
6.9, 2H), 2.60 (s, 2H), 1.65 (s, 2H), 1.54 (s, 2H), 1.37 (s, 2H), 1.21
(s, 19H), 0.83 (s, 3H). 13C NMR (126 MHz, DMSO) δ 172.22,
167.99, 147.01, 131.21, 128.45, 128.32, 35.38, 33.09, 31.74,
31.23, 29.45, 29.27, 29.16, 28.98, 22.54, 18.42, 14.42. LCMS:
t
R = 10.85 min; m/z = 372. HRMS m/z calcd for C23H38N3O
(M þ H), 372.3015; found, 302.3005.

2776 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Mathews et al.
N-(1-Carbamimidoylcyclopropyl)-4-tetradecylbenzamide Hydro-
chloride (24). General procedure A was used to convert 0.26
mmol of 21; C14 to the corresponding amidine. After standard
recrystallization techniques, 0.04 mmol of the title product was
recovered as a white solid. 1H NMR (500 MHz, DMSO) δ 9.04
(s, 1H), 8.61 (bs, 4H), 7.81 (d, J = 7.8, 2H), 7.27 (d, J = 5.8, 2H),
2.60 (t, J = 7.1, 2H), 1.65 (s, 2H), 1.54 (d, J = 5.5, 2H), 1.37
(s, 2H), 1.24-1.21 (m, 21H), 0.83 (t, J = 6.5, 3H). 13C NMR
(126 MHz, DMSO) δ 172.23, 168.02, 147.03, 131.24, 128.45,
128.31, 127.93, 35.38, 33.12, 31.72, 31.20, 29.44, 29.25, 29.13,
28.98, 22.52, 18.40, 14.39. LCMS: tR = 12.3 min; m/z = 400.
HRMS m/z calcd C25H42N3O (M þ H), 400.3328; found,
400.3322.
1-Carbamimidoyl-N-(4-decylphenyl)cyclopropanecarboxamide
Hydrochloride (27). General procedure A was used to convert
0.99 mmol of 26; C10 to the corresponding amidine. After
standard recrystallization techniques, 0.51 mmol of the title
1
product was recovered as a white solid. H NMR (500 MHz,
DMSO) δ 9.69 (s, 1H), 8.96 (s, 4H), 7.63-7.31 (m, 2H),
7.17-7.04 (m, 2H), 2.66-2.30 (m, 2H), 1.66-1.34 (m, 6H),
1.19 (dd, J = 30.7, 9.2 Hz, 14H), 0.90-0.76 (m, 3H). 13C NMR
(126 MHz, DMSO) δ 168.27, 166.44, 138.41, 136.53, 128.64,
128.58, 121.33, 121.28, 34.98, 31.73, 31.44, 29.98, 29.45, 29.30,
29.13, 29.02, 22.54, 15.12, 14.41. LCMS: tR = 9.8 min; m/z =
344. HRMS m/z calcd C21H34N3O (M þ H), 344.2702; found,
344.2690.
1-Carbamimidoyl-N-(4-dodecylphenyl)cyclopropanecarboxamide
Hydrochloride (28). General procedure A was used to convert
0.99 mmol of 26; C12 to the corresponding amidine. After
standard recrystallization techniques, 0.45 mmol of the title
N-(1-Carbamimidoylcyclopropyl)-4-hexadecylbenzamide Hy-
drochloride (25). General procedure A was used to convert 0.18
mmol of 21; C16 to the corresponding amidine. However,
instead of methanol, ethanol was used to solubilize the starting
material. After standard recrystallization techniques, 0.04 mmol
of the title product was recovered as a white solid. 1H NMR (500
MHz, DMSO) δ 9.06 (s, 1H), 8.65 (bs, 4H), 7.81 (d, J = 7.5, 2H),
7.27 (d, J = 7.5, 2H), 2.67-2.56 (m, 2H), 1.65 (s, 2H), 1.54 (d,
J = 4.0, 2H), 1.37 (s, 2H), 1.21 (bs, 25H), 0.84 (t, J = 5.4, 3H).
13C NMR (126 MHz, DMSO) δ 172.16, 168.01, 147.02, 131.21,
128.45, 128.31, 35.38, 33.11, 31.74, 31.24, 29.46, 29.27, 29.15,
28.99, 22.54, 18.41, 14.42. LCMS: tR = 12.4; m/z = 428. HRMS
m/z calcd C27H46N3O (M þ H), 428.3641; found, 428.3642.
1-Cyano-N-(4-decylphenyl)cyclopropanecarboxamide (26; C10).
General procedure B was used to couple 4-decylaniline (1.00
mmol) to 1-cyanocyclopropanecarboxylic acid. After column
chromatography, 0.99 mmol of the title product was recovered
as a white solid. 1H NMR (300 MHz, CDCl3) δ 8.16 (s, 1H), 7.41
(d, J = 8.4 Hz, 2H), 7.14 (d, J = 8.4 Hz, 2H), 2.73-2.40 (m, 2H),
1.77 (dd, J = 8.2, 4.4 Hz, 2H), 1.68-1.51 (m, 4H), 1.29 (d, J =
10.1 Hz, 14H), 0.90 (t, J = 6.6 Hz, 3H). 13C NMR (75 MHz,
CDCl3) δ 163.54, 140.27, 134.81, 129.13, 120.87, 120.27, 35.63,
32.16, 31.69, 29.86, 29.75, 29.60, 29.49, 22.95, 18.44, 14.39,
14.32.
1
product was recovered as a white solid. H NMR (300 MHz,
DMSO) δ 9.57 (s, 1H), 9.16 (s, 2H), 8.91 (s, 2H), 7.45 (d, J = 8.1
Hz, 2H), 7.10 (d, J = 8.3 Hz, 2H), 2.48 (m, 2H), 1.52 (m, 6H),
1.21 (s, 18H), 0.83 (s, 3H). 13C NMR (75 MHz, DMSO) δ
168.33, 166.76, 138.72, 136.69, 128.88, 121.61, 35.21, 31.97,
31.67, 30.17, 29.69, 29.53, 29.39, 29.26, 22.77, 15.42, 14.65.
LCMS: tR = 11.5 min; m/z = 372. HRMS m/z calcd
C23H38N3O (M þ H), 372.3015; found, 372.3011.
1-Carbamimidoyl-N-(4-tetradecylphenyl)cyclopropanecarbox-
amide Hydrochloride (29). General procedure A was used to
convert 0.99 mmol of 26; C14 to the corresponding amidine.
After standard recrystallization techniques, 0.37 mmol of the
title product was recovered as a white solid. 1H NMR (300 MHz,
DMSO) δ 9.57 (s, 1H), 9.16 (s, 2H), 8.91 (s, 2H), 7.45 (d, J = 8.1
Hz, 2H), 7.10 (d, J = 8.3 Hz, 2H), 2.48 (s, 2H), 1.52 (dd, J =
40.8, 22.8 Hz, 6H), 1.21 (s, 22H), 0.83 (s, 3H). 13C NMR (75
MHz, DMSO) δ 168.33, 166.76, 138.72, 136.69, 128.88, 121.61,
35.21, 31.97, 31.67, 30.17, 29.69, 29.53, 29.39, 29.26, 22.77,
15.42, 14.65. tR = 12.5 min; m/z = 400. HRMS m/z calcd
C25H42N3O (M þ H), 400.3328; found, 400.3334.
1-Carbamimidoyl-N-(4-hexadecylphenyl)cyclopropanecarbox-
amide Hydrochloride (30). General procedure A was used to
convert 0.99 mmol of 26; C16 to the corresponding amidine.
However, instead of methanol, ethanol was used to solubilize
the starting material. After standard recrystallization techni-
ques, 0.27 mmol of the title product was recovered as a white
solid. 1H NMR (500 MHz, DMSO) δ 9.92 (s, 2H), 9.82-9.58 (s,
1H), 9.06 (s, 2H), 7.47 (s, 2H), 7.09 (s, 2H), 2.49 (s, 2H), 1.46 (d,
J = 62.7 Hz, 6H), 1.21 (s, 26H), 0.83 (s, 3H). 13C NMR (75
MHz, DMSO) δ 168.32, 166.75, 138.71, 136.69, 128.87, 121.59,
35.21, 31.97, 31.69, 30.17, 29.70, 29.54, 29.39, 29.28, 22.78,
15.42, 14.66. tR = 13.2; m/z = 428. HRMS m/z calcd
C27H46N3O (M þ H), 428.3641; found, 428.3639.
1-Cyano-N-(4-dodecylphenyl)cyclopropanecarboxamide (26;
C12). General procedure B was used to couple 4-dodecylaniline
(1.00 mmol) to 1-cyanocyclopropanecarboxylic acid. After col-
umn chromatography, 0.99 mmol of the title product was
recovered as a white solid. 1H NMR (300 MHz, CDCl3) δ
8.18 (s, 1H), 7.41 (d, J = 8.3 Hz, 2H), 7.14 (d, J = 8.3 Hz,
2H), 2.57 (t, J = 7.6 Hz, 2H), 1.83-1.67 (m, 2H), 1.67-1.45
(m, 4H), 1.28 (s, 18H), 0.90 (t, J = 6.4 Hz, 3H). 13C NMR
(75 MHz, CDCl3) δ 163.53, 140.24, 134.84, 129.12, 120.88,
120.26, 35.63, 32.19, 31.70, 29.93, 29.77, 29.63, 29.51, 22.96,
18.44, 14.40, 14.31.
1-Cyano-N-(4-tetradecylphenyl)cyclopropanecarboxamide (26;
C14). General procedure B was used to couple 4-tetradecylani-
line (1.00 mmol) to 1-cyanocyclopropanecarboxylic acid. After
column chromatography, 0.99 mmol of the title product was
recovered as a white solid. 1H NMR (300 MHz, CDCl3) δ
8.24-8.00 (m, 1H), 7.40 (d, J = 8.4 Hz, 2H), 7.15 (d, J = 8.4
Hz, 2H), 2.76-2.38 (m, 2H), 1.77 (dd, J = 8.2, 4.4 Hz, 2H),
Sphingosine Kinase Assay. Human SphK1 and mouse SphK2
cDNAs were used to generate mutant baculoviruses that en-
coded these proteins. Infection of Sf9 insect cells with the viruses
for 72 h resulted in >1000-fold increases in SphK activity in
10000g supernatant fluid from homogenized cell pellets. The
enzyme assay conditions were exactly as described,16 except
infected Sf9 cell extract containing 2-3 μg protein was used as a
source of enzyme.
1.67-1.50 (m, 4H), 1.27 (s, 22H), 0.89 (t, J = 6.6 Hz, 3H). 13
C
NMR (75 MHz, CDCl3) δ 163.50, 140.32, 134.74, 129.16,
120.81, 120.28, 35.63, 32.18, 31.70, 29.93, 29.76, 29.63, 29.50,
22.95, 18.45, 14.39.
Vascular Smooth Muscle Cell Bioassays. Rat aortic vascular
smooth muscle cells were cultured in vitro as previously de-
scribed.28 Total mRNA was extracted using Trizol (Invitrogen),
cDNA was synthesized using the iScript cDNA synthesis kit
(BioRad), and quantitative real-time polymerase chain reaction
(PCR) was used to measure Sphk1, Sphk2, and 18S mRNA
expression as previously described.27 Primer sequences: Sphk1
F: GTGCCATCCAGAAACCCCTA, R: AGTCCACGTCA-
GCAACGAAA; Sphk2 F: TGGGAAGGCATTGTCACTGT,
R: AAGAAGCGAGCAGTTGAGCA; 18S F: CGGCTAC-
CACATCCAAGGAA, R: AGCTGGAATTACCGCGGC.
Proliferation was assessed by BrdU incorporation (10 μM)
1-Cyano-N-(4-hexadecylphenyl)cyclopropanecarboxamide (26;
C16). General procedure B was used to couple 4-hexadecylani-
line (1.00 mmol) to 1-cyanocyclopropanecarboxylic acid. After
column chromatography, 0.99 mmol of the title product was
recovered as a white solid. 1H NMR (300 MHz, CDCl3) δ
8.29-7.98 (m, 1H), 7.43 (t, J = 14.0 Hz, 2H), 7.14 (d, J = 8.4
Hz, 2H), 2.78-2.44 (m, 2H), 1.77 (dd, J = 8.2, 4.4 Hz, 2H),
1.67-1.50 (m, 4H), 1.27 (s, 26H), 0.89 (t, J = 6.7 Hz, 3H). 13
C
NMR (75 MHz, CDCl3) δ 163.50, 140.29, 134.78, 129.14,
120.83, 120.27, 35.63, 32.19, 31.71, 29.95, 29.77, 29.64, 29.51,
22.96, 18.44, 14.40.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2777
References
based on the Roche BrdU immuno-absorbance assay. Cytotoxic
response was determined by the Promega MultiTox-Fluor
Assay.
(1) Hait, N. C.; Oskeritzian, C. A.; Paugh, S. W.; Milstien, S.; Spiegel,
S. Sphingosine kinases, sphingosine 1-phosphate, apoptosis and
diseases. Biochim. Biophys. Acta 2006, 1758, 2016–2026.
Quantification of S1P by LC/MS. S1P and DHS1P were
quantified as their bisacetylated derivatives (BisAcS1P and
BisAcDHS1P) using adaptations of previously reported meth-
ods.37 Derivatization eliminates sample carryover that is a noted
problem with these analytes. In brief, cells were washed in PBS
and then scraped into 2 mL of ice cold MeOH, which was
transferred to glass tubes containing 1 mL CHCl3, 0.5 mL 0.1 M
HCl, and 50 pmol of C17 S1P. After vortexing, an additional 1
mL of CHCl3 and 1.3 mL 0.1 M HCl were added to the tubes
containing the cell suspension. The tubes were vortexed a second
time and centrifuged. The lipid-containing organic phase was
separated into a second 4 mL glass vial. The organic layer was
evaporated to dryness and the material was acetylated with
acetic anhydride in pyridine. After a second evaporation, the
resulting material was dissolved in MeOH for analysis. Samples
were separated on an Aglient Technologies Zorbax Eclipse
XDB-C8 column, 4.6 mm ꢀ 150 mm, 5 mm particle size
using methanol/water/HCOOH, 79/20/0.5, v/v, with 5 mM
NH4COOH as solvent A and methanol/acetonitrile/HCOOH,
59/40/0.5, v/v, with 5 mM NH4COOH as solvent B. Analytes
were detected by electrospray ionization tandem mass spectro-
metry using an ABI 4000 Q-Trap instrument operated in
positive ionization mode with optimized declustering potentials,
collision energies, and exit potentials. The following m/z transi-
tions were monitored: BisAcC17S1P 448/388; BisACS1P, 462/
402; BisAcDHS1P, 464/404. Calibration curves were generated
using BisAcS1P BisAcDHS1P and BisAcDHS1P that were
prepared from S1P, DHS1P, and C17S1P obtained from Avanti
Polar Lipids and quantified independently by phosphorus
analysis.
(2) Spiegel, S.; Milstien, S. Sphingosine-1-phosphate: and enigmatic
signaling lipid. Nat. Rev. Mol. Cell Biol. 2003, 4, 397–407.
(3) Lynch, K. R. Lysophospholipid Receptor Nomenclature. Biochim.
Biophys. Acta 2002, 1582, 70–71.
(4) Rosen, H.; Sanna, M.; Germana, C.; Stuart, M.; Gonzalez-
Cabrera, P. J. Tipping the gatekeeper: S1P regulation of endothe-
lial barrier function. Trends Immunol. 2007, 28, 102–107.
(5) French, K. J.; Upson, J. J.; Keller, S. N.; Zhuang, Y.; Yun, J. K.;
Smith, C. D. Discovery and evaluation of inhibitors of human
sphingosine kinase. Cancer Res. 2003, 63, 5962–5969.
(6) Mizugishi, K.; Yamashita, T.; Olivera, A.; Miller, G. F.; Spiegel, S.;
Proia, R. L. Essential role for sphingosine kinases in neural and
vascular development. Mol. Cell. Biol. 2005, 25, 11113–11121.
(7) Okada, T.; Ding, G.; Sonoda, H.; Kajimoto, T.; Haga, Y.;
Khosrowbeygi, A.; Gao, S.; Miwa, N.; Jahangeer, S.; Nakamura,
S.-I. Involvement of N-terminal-extended form of sphingosine
kinase 2 in serum-dependent regulation of cell proliferation and
apoptosis. J. Biol. Chem. 2005, 280, 36318–36325.
(8) Johnson, K. R.; Becker, K. P.; Facchinetti, M. M.; Hannun, Y. A.;
Obeid, L. M. PKC-dependent activation of sphingosine kinase 1
and translocation to the plasma membrane. Extracellular release
of sphingosine-1-phosphate induced by phorbol 12-myristate 13-
acetate (PMA). J. Biol. Chem. 2002, 277, 35257–35262.
(9) Maceyka, M.; Sankala, H.; Hait, N. C.; Le Stunff, H.; Liu, H.;
Toman, R.; Collier, C.; Zhang, M.; Satin, L. S.; Merrill, A. H.;
Milstien, S.; Spiegel, S. SphK1 and SphK2, sphingosine kinase
isoenzymes with opposing functions in sphingolipid metabolism.
J. Biol. Chem. 2005, 280, 37118–37129.
(10) Liu, H.; Toman, R. E.; Goparaju, S. K.; Maceyka, M.; Nava, V. E.;
Sankala, H.; Payne, S. G.; Bektas, M.; Ishii, I.; Chun, J.; Milstien,
S.; Spiegel, S. Sphingosine kinase type 2 is a putative BH3-only
protein that induces apoptosis. J. Biol. Chem. 2003, 278, 40330–
40336.
(11) Paugh, S. W.; Paugh, B. S.; Rahmani, M.; Kapitonov, D.;
Almenara, J. A.; Kordula, T.; Milstien, S.; Adams, J. K.; Zipkin,
R. E.; Grant, S.; Spiegel, S. A selective sphingosine kinase 1
inhibitor integrates multiple molecular therapeutic targets in
human leukemia. Blood 2008, 112, 1382–1391.
Liquid Chromatography and Mass Spectrometry for Evalua-
tion of Chemical Purity. All compound subjected to biological
evaluation were determined to be >95% pure by LCMS evalua-
tion. High performance liquid chromatography-mass spectro-
metry (LCMS) was carried out on a Waters 2695 separations
module and a Finnigan LCQ series mass spectrometer. All
compounds were evaluated for purity using a Thompson Instru-
ment Company Advantage C18 column. Columns were outfitted
(12) Weigert, A.; Schiffmann, S.; Sekar, D.; Ley, S.; Menrad, H.;
€
Werno, C.; Grosch, S.; Geisslinger, G.; Brune, B. Sphingosine
kinase 2 deficient tumor xenografts show impaired growth and fail
to polarize macrophages towards and anti-inflammatory pheno-
type. Int. J. Cancer 2009, PMID: 19618460.
˚
with 5 μm beads with a 60 A pore size; columns were 250 mm in
(13) Wong, L; Tan, S. S. L.; Lam, Y.; Melendez, A. J. Synthesis and
evaluation of sphingosine analogues as inhibitors of sphingosine
kinases. J. Med. Chem. 2009, 52, 3618–3626.
(14) Goldstein, D. M.; Gray, N. S.; Zarrinkar, P. P. High-throughput
kinase profiling as a platform for drug discovery. Nat. Rev. Drug
Discovery 2008, 7, 391–397.
(15) Mandala, S.; Hajdu, R.; Bergstrom, J.; Quackenbush, E.; Xie, J.;
Milligan, J.; Thornton, R.; Shei, G.-J.; Card, D.; Keohane, C.;
Rosenbach, M.; Hale, J.; Lynch, C. L.; Rupprecht, K.; Parsons,
W.; Rosen, H. Alteration of lymphocyte trafficking by sphingo-
sine-1-phosphate receptor agonists. Science 2002, 296, 346–349.
(16) Brinkmann, V.; Davis, M. D.; Heise, C. E.; Albert, R.; Cottens, S.;
Hof, R.; Bruns, C.; Prieschl, E.; Baumruker, T.; Hiestand, P.;
Foster, C. A.; Zollinger, M.; Lynch, K. R. The immune modulator
FTY720 targets sphingosine 1-phosphate receptors. J. Biol. Chem.
2002, 277, 21453–21457.
length and 4.6 mm in diameter. Mobile phase A consisted of
HPLC grade H2O and 0.01% TFA; mobile phase B consisted of
MeCN and 0.01% TFA. LCMS identification and purity utilized
a binary gradient starting with 90% A and 10% B and linearly
increasing to 100% B over the course of 10 min, followed by an
isocratic flow of 100% B for an additional 10 min. A flow rate of
1 mL/min was maintained throughout the HPLC method. The
purity of all products was determined by integration of the total
ion count (TIC) spectra and integration of the ultraviolet (UV)
spectra at 214 nm; a Waters 486 tunable absorbance detector was
used to collect all UV data. Retention times are abbreviated as tR;
mass to charge ratios are abbreviated as m/z.
(17) Kharel, Y.; Lee, S.; Snyder, A. H.; Sheasley-O’Neill, S. L.; Morris,
M. A.; Setiady, Y.; Zhu, R.; Zigler, M. A.; Burcin, T. L.; Ley, K.;
Tung, K. S. K.; Engelhard, V. H.; Macdonald, T. L.; Pearson-
White, S.; Lynch, K. R. Sphingosine kinase 2 is required for
modulation of lymphocyte traffic by FTY720. J. Biol. Chem.
2005, 280, 36865–36872.
(18) Dev, K. K.; Mullershausen, F.; Mattes, H.; Kuhn, R. R.; Bilbe, G.;
Hoyer, D.; Mir, A. Brain sphingosine-1-phosphate receptors:
implication for FTY720 in the treatment of multiple sclerosis.
Pharmacol. Ther. 2008, 117, 77–93.
(19) Clemens, J. J.; Davis, M. D.; Lynch, K. R.; Macdonald, T. L.
Synthesis of para-Alkyl Aryl Amide Analogues of Sphingosine-1-
Phosphate: Discovery of Potent S1P Receptor Agonists. Bioorg.
Med. Chem. Lett. 2003, 13, 3401–3404.
Acknowledgment. This work was supported by grants
from the National Institutes of Health (R01 GM067958 to
K.R.L. and T.L.M., T32 GM008715 to P.C.K.; HL081682 to
B.R.W.; GM50388 to A.J.M. and American Heart Associa-
tion SDG to B.R.W. We also thank Dr. Pam Schoppee-Bortz,
Ph.D., Cardiovascular Division, MSTP-candidate John
Hogan, predoctoral candidate Monica Lee, ME, Department
of Biomedical Engineering, and Amanda M. Gellett, Depart-
ment of Pharmacology, University of Virginia, for their
technical expertise.
(20) Clemens, J. J.; Davis, M. D.; Lynch, K. R.; Macdonald, T. L.
Synthesis of 4(5)-phenylimidazole-based analogues of Sphingo-
sine-1-Phosphate and FTY720: Discovery of Potent S1P1 Receptor
Agonists. Bioorg. Med. Chem. Lett. 2005, 15, 3568–3572.
Supporting Information Available: Additional 1H and 13C
NMR data for 9, 13, 20, and 22. This material is available free of
charge via the Internet at http://pubs.acs.org.

2778 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Mathews et al.
(21) Foss, F. W.; Snyder, A. H.; Davis, M. D.; Rouse, M.; Okusa, M.
D.; Lynch, K. R.; Macdonald, T. L. Synthesis and biological
evaluation of R-aminophosphonates as potent, subtype-selective
sphingosine 1-phosphate receptor agonists and antagonists.
Bioorg. Med. Chem. 2007, 15, 663–677.
(22) Zhu, R.; Snyder, A. H.; Kharel, K.; Schaffter, L.; Sun, Q.;
Kennedy, P. C.; Lynch, K. R.; Macdonald, T. L. Asymmetric
Synthesis of Conformationally Constrained Fingolimod Analo-
gues-Discovery of an Orally Active Sphingosine 1-Phosphate
Receptor Type-1 Agonist and Receptor Type-3 Antagonist.
J. Med. Chem. 2007, 50, 6428–6435.
(23) Foss, F. W.; Mathews, T. P.; Kharel, Y.; Kennedy, P. C.; Snyder,
A. H.; Davis, M. D.; Lynch, K. R.; Macdonald, T. L. Synthesis and
biological evaluation of sphingosine kinase substrates as sphingo-
sine 1-phosphate receptor prodrugs. Bioorg. Med. Chem. 2009, 17,
6123–6136.
(24) Tyrkov, A; Usova, A. Reduction of Substituted 5-(Nitromethyl)-3-
phenyl-1, 2, 4-oxadiazoles. Russ. J. Org. Chem. 2004, 40, 286–287.
(25) Kitagawa, H.; Ozawa, T.; Takahata, S.; Iida, M.; Saito, J.;
Yamada, M. Phenylimidazole derivatives of 4-pyridone as dual
inhibitors of bacterial enoyl-acyl carrier protein reductases FabI
and FabK. J. Med. Chem. 2007, 50, 4710–4720.
smooth muscle cell phenotype. Arterioscler., Thromb., Vasc. Biol.
2008, 28, 1454–1461.
(29) Wymann, M. P.; Schneiter, R. Lipid signalling in disease. Nat. Rev.
Mol. Cell Biol. 2008, 9, 162–176.
(30) Yokota, S.; Taniguchi, Y.; Kihara, A.; Mitsutake, S.; Igarashi, Y.
Asp177 in C4 domain of mouse sphingosine kinase 1a is important
for the sphingosine recognition. FEBS Lett. 2004, 578, 106–110.
(31) Koppel, I.; Koppel, J.; Leito, I.; Grehn, L. Basicity of 3-amino-
propionamidine derivatives in water and dimethyl sulphoxide.
J. Phys. Org. Chem. 1996, 9, 265–268.
(32) Buehrer, B. M.; Bell, R. M. Inhibitionof sphingosine kinase in vitro
and in platelets. J. Biol. Chem. 1992, 267, 3154–3159.
(33) Lynch, K. R.; Macdonald, T. M. Sphingosine 1-phosphate chemi-
cal biology. Biochim. Biophys. Acta 2008, 1781, 508–512.
(34) Hannun, Y. A.; Obeid, L. M. Principles of bioactive lipid signal-
ling: lessons from sphingolipids. Nat. Rev. Mol. Cell Biol. 2008, 9,
139–150.
(35) Olivera, A.; Kohama, T.; Tu, Z.; Milstien, S.; Spiegel, S. Purifica-
tion and characterization of rat kidney sphingosine kinase. J. Biol.
Chem. 1998, 273, 12576–12583.
(36) Liu, H.; Sugiura, M.; Nava, V. E.; Edsall, L. C.; Kono, K.; Poulton,
S.; Milstien, S.; Kohama, T.; Spiegel, S. Molecular cloning
and functional characterization of a novel mammalian sphingo-
sine kinase type 2 isoform. J. Biol. Chem. 2000, 275, 19513–
19520.
(26) Schaefer, F. C.; Peters, G. A. Base-Catalyzed Reaction of Nitriles
with Alcohols. A Convenient Route to Imidates and Amidine Salts.
J. Org. Chem. 1961, 26, 412–418.
(27) Tang, F.-Y.; Qu, L.-Q.; Xu, Y.; Ma, R.-J.; Chen, S.-H.; Li, G.
Practical Synthesis of Structurally Important Spirodiamine Tem-
plates. Synth. Commun. 2007, 37, 3793–3799.
(37) Berdyshev, E. V.; Gorshkova, I. A.; Garcia, J. G. N.; Natarajan,
V.; Hubbard, W. C. Quantitative analysis of sphingoid base-
1-phosphates as bisacetylated derivatives by liquid chromatogra-
phy-tandem mass spectrometry. Anal. Biochem. 2005, 339, 129–
136.
(28) Wamhoff, B. R.; Lynch, K. R.; Macdonald, T. L.; Owens, G. K.
Sphingosine-1-phosphate receptor subtypes differentially regulate