Synthesis, stability, and implications of phosphothioate agonists of sphingosine-1-phosphate receptors

Article abstract of DOI:10.1016/j.bmcl.2005.07.057

Phosphothioates may provide metabolic stability when compared to their phosphate counterparts, while retaining the potency and efficacy as agonists at sphingosine-1-phosphate (S1P) G-protein coupled receptors. Unlike their phosphate precursors, phosphothi

Full text of DOI:10.1016/j.bmcl.2005.07.057

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4470–4474
Synthesis, stability, and implications of phosphothioate agonists
of sphingosine-1-phosphate receptors
Frank W. Foss, Jr.,a Jeremy J. Clemens,a Michael D. Davis,b Ashley H. Snyder,c
Molly A. Zigler,a Kevin R. Lynchb,c and Timothy L. Macdonalda,*
aDepartment of Chemistry, University of Virginia, McCormick Road, PO Box 400319, Charlottesville, VA 22904, USA
bDepartment of Biochemistry and Molecular Genetic, University of Virginia, McCormick Road, PO Box 400319, Charlottesville,
VA 22904, USA
cDepartment of Pharmacology, University of Virginia, McCormick Road, PO Box 400319, Charlottesville, VA 22904, USA
Received 7 March 2005; revised 8 July 2005; accepted 12 July 2005
Available online 24 August 2005
Abstract—Phosphothioates may provide metabolic stability when compared to their phosphate counterparts, while retaining the
potency and efficacy as agonists at sphingosine-1-phosphate (S1P) G-protein coupled receptors. Unlike their phosphate precursors,
phosphothioate compounds with S1P-receptor profiles similar to that of FTY720, an emerging immunomodulator, were shown to
evoke prolonged lymphopenia in vivo. Analysis of mouse plasma concentrations for a series of related alcohol/phosphate/phospho-
thioate compounds showed the conversion of the phosphate to alcohol. These preliminary data highlight the importance of meta-
bolic regulation of S1P receptor ligands.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Known as an intracellular messenger molecule,1 sphin-
gosine-1-phosphate (S1P, Table 1) is a lysophospholipid
mediator that has been receiving increasing attention
due to its extracellular signaling at five G-protein cou-
pled receptors (S1P1–5).2 Activation of these receptors
is reportedly responsible for many of S1PÕs myriad cellu-
lar physiologies,1,3,4 including cell growth, proliferation,
survival and migration.5,6 Specifically, S1P receptors
have been implicated in controlling blood vessel
development, cardiac rate, blood pressure, and immune
regulation.7–10
cifically at the S1P1 receptor.12–14 Therefore, the discov-
ery of subtype selective S1P1 agonists may be desirable
as potential therapeutics.
We previously reported the synthesis of potent, subtype
selective S1P analogues with receptor profiles similar to
that of FTY720-P.15 However, initial in vivo studies on
these and other S1P/FTY720-P phosphate analogues,
such as compound 12 (Table 1), revealed that they
evoked lymphopenia transiently (<8 h), even when
administered intraperitoneally (data unpublished). To
account for this observation, we hypothesized that per-
haps the dynamic equilibrium that regulates sphingoli-
pids may be of consequence.
FTY720 (Fig. 1) is a novel immunomodulator that,
when activated by one or more kinases to FTY720-
phosphate (FTY720-P), acts at four of S1PÕs five recep-
tors.11 Unlike conventional immunosuppressants, this
drug evokes lymphopenia by inhibiting the egress of
lymphocytes from secondary lymphoid tissues into the
peripheral circulation. S1P1,4 were discovered to be
abundantly expressed on peripheral blood T-lympho-
cytes.12 Several studies suggest that this non-cytotoxic
lymphopenia occurs through FTY720-PÕs agonism spe-
One or more kinases are crucial for the activation of
FTY720 to FTY720-P.8,16 While sphingosine is phos-
phorylated by both human sphingosine kinases (SphK1
and SphK2), recent studies point to SphK2 as the kinase
more likely to be responsible for phosphorylation of
FTY720 to FTY720-P.17,18 Numerous lipid phosphatases
may play a role in the dephosphorylation of FTY720-P,
but, to date, no study has indicated a specific enzyme.
Considering this enzymatic regulation and FTY720Õs
activity in vivo, we hypothesized that our previously
synthesized compounds were deactivated by phospha-
Keywords: Sphingosine-1-phosphate; Phosphothioate; Lymphopenia.
*
Corresponding author. Tel.: +1 43 49 24 059 5; fax: +1 43 49 82 23
02; e-mail: fwf4b@virginia.edu
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.07.057

F. W. Foss, Jr. / Bioorg. Med. Chem. Lett. 15 (2005) 4470–4474
4471
H2N
H2N
OH
OH
OPO3H2
Kinase(s)
Phosphatase(s)
OH
C8H17
C8H17
FTY720
FTY720-P
Figure 1. Enzymatic control of FTY720:FTY720-P. Kinases convert
inactive FTY720 to active immunomodulator FTY720-P. Conversely,
FTY720-P is dephosphorylated by one or more phosphatases.
tase activity. Furthermore, we investigated two apparent
strategies to deal with the subject of bio-activation. Both
methods take advantage of the equilibrium mentioned
above. One approach, based on early structure activity
relationship studies of FTY720,19 pertains to the synthe-
sis of sterically hindered alcohols that contain or resem-
ble the 2-amino-1,3-propanediol moiety of FTY720.
The second tactic, also investigated by other groups,20
uses the synthesis of more biologically stable phosphate
isosteres, such as the phosphothioates displayed in this
work.
While both avenues are being examined in our laborato-
ries, this paper describes the preparation of phospho-
thioate analogues of S1P that: (i) are readily
synthesized in parallel with their phosphate counter-
parts; (ii) are comparatively subtype-selective at S1P
receptors; and (iii) markedly induce sustainable lympho-
penia in vivo.
Specifically, a series of 2-demethylenehydroxy-FTY720
analogues (Scheme 1, 5–7) was synthesized. These com-
pounds were chosen with the knowledge that alcohol 5
does not induce lymphopenia.20 Also, synthesized was
phosphothioate 13 (Scheme 2): This compound was cho-
sen to complement phosphate 12, one of our initial lead
candidates.
Initial steps in the synthesis of compounds 5–7 (Scheme
1) were adapted from the work of Durand et al.21 Com-
pound 1 was obtained by the Friedel–Crafts acylation of
commercially available materials, 1-phenyloctane in 2-
bromoacetyl bromide, with the help of AlCl3. Notably,
this reaction proceeded with higher yields in the
absence of solvent as compared to the addition of
1,2-dichloroethane. a-Bromo-ketone 1 was then con-
verted to intermediate 3, as previously described.21
Hydrolysis-decarboxylation of 3 was achieved in one
step under harsh acidic conditions. Problems related to
initial solubility, under standard aqueous conditions,
were best overcome by the addition of methanol. This
technique yielded a surprisingly small percent (ꢀ20%
1
by H NMR) of the anticipated methyl ester 4a, as well
as desired compound 4b. Conveniently, the acid/ester
mixture was carried onto a lithium aluminum hydride
reduction to provide the racemic amino alcohol 5 in
sufficient yield. After standard N-Boc protection, com-
pound 5 was efficiently transformed to a common phos-
phite intermediate (see 10, Scheme 2), with di-tert-butyl-
N,N-diisopropylphosphoramidite. This intermediate
was not isolated, but rather oxidized, in situ, by H2O2
or elemental sulfur (S8), to form the protected phosphate
or phosphothioate tri-esters, respectively. Acid mediated

4472
F. W. Foss, Jr. / Bioorg. Med. Chem. Lett. 15 (2005) 4470–4474
Scheme 1. Reagents and conditions: (a) NaOEt, EtOH, 60 ꢁC, 1 h, then N-acetamido-diethylmalonate, 60 ꢁC, 1 h, 90%; (b) Et3SiH, TiCl4, CH2Cl2,
12 h, rt, 83%; (c) 12 M HCl, MeOH, reflux, 2 h, 78%; (d) LiAlH4, THF, reflux, 10–16 h, 74%; (e) Boc2O Et3N, CH2Cl2, 0 ꢁC to rt, 4 h, 91%; (f) di-tert-
butyl-N,N-diisopropylphosphoramidite, 1H-tetrazole, CH3CN, CH2Cl2, THF, rt, 12 h, then H2O2, rt, 1 h, 36%; (g) di-tert-butyl-N,N-
diisopropylphosphoramidite, 1H-tetrazole, CH3CN, CH2Cl2, THF, rt, 12 h, then S8, rt, 1 h, 61%; (h) trifluoroacetic acid, CH2Cl2, rt, 4 h, 70%;
(i) trifluoroacetic acid, bromotrimethylsilane, thiophenol, CH2Cl2, 4 h, 0 ꢁC to rt, 95%.
Scheme 2. Reagents and conditions: (a) 4-octylaniline, PyBOP, DIEA, CH2Cl2, rt, 4 h, 96%; (b) H2, 10% Pd/C, EtOH, rt, 12 h, 96%; (c) 1H-tetrazole,
di-tert-butyl-N,N-diisopropylphosphoramidite, 1:1 CH2Cl2/THF, rt, 12 h; (d) H2O2, rt, 4 h, 40% (two steps); (e) 1:1 TFA/CH2Cl2, rt, 4 h, 98%; (f) S8,
rt, 3 h, 37% (two steps); (g) thiophenol, TMSBr, 1:1 TFA/CH2Cl2, rt, 4 h, 94%; (h) Na2CO3(aq)/EtOAc.
deprotection was achieved on completion of the synthe-
sis of both compounds 6 and 7. An excess of cation scav-
engers was employed in the formation of 7 to avoid an
O- to S- transfer of a t-butyl group, which could not
be deprotected under numerous acidic conditions.
determined to be lower than 80% of S1PÕs maximum
are noted as partial agonists (PA).
Binding assay data show that phosphates 6 and 12 have
EC50 values similar to that of FTY720-P at S1P1. While
phosphothioate 7 was found to be less potent at S1P1 than
phosphate 6, phosphothioate 13 was equipotent to phos-
phate 12. Desirably, the two phosphothioates retained a
similar subtype selectivity between S1P1 and S1P3 recep-
tors when compared to their phosphate precursors.
The synthesis of compounds 11, 12, and 13 (Scheme 2)
began with, commercially available protected threonine.
A PyBOP condensation with 4-octylaniline led to amide
8. The alcohol moiety was unmasked by hydrogenolysis
to compound 9 and subsequently phosphorylated, selec-
tively oxidized, and deprotected by the methods de-
scribed above to provide both the phosphate 12 and
phosphothioate 13. The alcohol 11 was accomplished
by a similar deprotection and neutralization of interme-
diate 9.
Compounds 6, 7, 12, and 13 were tested further for their
ability to induce lymphopenia in mice (see Fig. 2). Mice
were rendered lymphopenic 4 h after the injection of any
of the compounds, but this response was markedly
diminished at 19 h post injection in mice treated with
phosphate-containing compounds. In contrast, phos-
phothioates 7 and 13 evoked lymphopenia at the 19 h
interval.
Receptor activities of S1P; FTY720-P; and compounds
5–7 and 11–13 are given as EC50 values (Table 1). The
results were determined by ligand dependent binding
of GTP[c35S] to individual S1P-receptor-G-protein com-
plexes that were overexpressed in HEK293T cells.15 Effi-
cacy levels were also determined as a percentage of S1PÕs
maximal effect at each individual receptor. Maxima
While phosphates 6 and 12 are slightly less potent than
FTY720-P, neither caused extensive lymphopenia at
19 h. This inactivity was described for compound 6;18
however, its comparatively higher potency at S1P1,

F. W. Foss, Jr. / Bioorg. Med. Chem. Lett. 15 (2005) 4470–4474
4473
compounds by LC–MS understandably catalyzed the
hydrolysis of phosphothioate 13 to phosphate 12.24
For this reason, it was difficult to quantify the concen-
tration of phosphothioate in plasma. The large amount
of detected phosphate in animals treated with phospho-
thioate 13 is likely explained by this acid-catalyzed
hydrolysis. Successfully, a lower concentration of alco-
hol was present in animals treated with phosphothioate
13 than in animals treated with phosphate 12.
Lymphopenia Assay
5
4
3
2
1
0
4 hours
19 hours
That compound 6 induces brief lymphopenia in animal
models, unlike less potent phosphothioate 7, suggests
that 6Õs equilibrium lies far to the left (dephosphoryl-
ated) in the relationship described in Figure 1. Without
knowledge of kinase or phosphatase specificity, it is pre-
mature to state whether this lack of activity is due to
alcohol 5 being a relatively poor substrate for SphK2
or 6 being a reasonably good substrate for one or more
phosphatases.
6
7
12
13
Vehicle
Compound
Figure 2. Lymphopenia induced by compounds 6, 7, 12, and 13.
Compounds were dissolved in 3% fatty acid free-BSA and injected ip at
8 mg/kg. Blood was drawn from the orbital sinus at 4 and 19 h;
lymphocyte counts were determined using a Hemavet blood analyzer.
Results are mean of three measurements.
A greater understanding of SphK2 and other sphingo-
sine related enzymes may allow for the development of
S1P1 agonists with increased bioavailability as phos-
phates. Reports in this area, and the synthesis of various
phosphate mimetics, are to follow.
while suspected, was not previously known. The data
suggest that potency of compounds 6 and 12, as was
hypothesized, is of less consequence when enzymatic
regulation is taken into account.
To support the hypothesis that kinases and phospha-
tases were specifically involved in regulating our
phosphate compounds, plasma concentrations of com-
pounds 11, 12, and 13 were investigated and compared
in vivo (see Fig. 3).22,23 The compounds were adminis-
tered independently by intraperitoneal injection and
blood was drawn at 4 and 24 h. The data show that 11
is present in plasma from 4 to 24 h in detectable concen-
trations, 2.3 and 0.6 lM, respectively. No phosphate is
detected in mice treated with compound 11. In compar-
ison, the phosphate was detected at concentrations
below those of the alcohol at both time points. As
hypothesized, the alcohol 11 was present in larger con-
centrations than phosphate 12 for mice treated with 12.
Acknowledgments
This work was supported by grants from the NIH
[NIGMS R01 GM067958 (to K.R.L.) and NIGMS F31
GM064101 (to M.D.D.)].
References and notes
1. Spiegel, S.; Milstien, S. FEBS Lett. 2000, 476, 55.
2. Anliker, B.; Chun, J. J. Biol. Chem. 2004, 279, 20555.
3. Hla, T.; Lee, M.-J.; Ancellin, N.; Paik, J. H.; Kluk, M. J.
Science 2001, 294, 1875.
Plasma concentrations of animals treated with 11 or 12
support the hypothesis that our phosphate analogues
undergo dephosphorylation. However, the eluent (con-
taining H2O, CH3CN, and TFA) used to analyze these
4. Osborne, N.; Stainier, D. Y. R. Annu. Rev. Physiol. 2003,
65, 23.
5. Hla, T. Prostaglandins Other Lipid Mediat. 2001, 64, 135.
6. Kluk, M. J.; Hla, T. Biochim. Biophys. Acta 2002, 1582,
72.
7. Kupperman, E.; An, S.; Osborne, N.; Waldron, S.;
Stainier, D. Y. Nature 2000, 406, 192.
Plasma Assay
8. Pyne, S.; Pyne, N. J. Biochem. J. 2000, 349, 385.
9. Forrest, M.; Sun, S.-Y.; Hajdu, J.; Bergstrom, J.; Card,
D.; Doherty, G.; Hale, J.; Keohane, C.; Meyers, C.;
Milligan, J.; Mills, S.; Nomura, N.; Rosen, H.; Rosen-
bach, M.; Shei, G.-J.; Singer, I. I.; Tian, M.; West, S.;
White, V.; Xie, J.; Proia, R. L.; Mandala, S. J. Pharmacol.
Exp. Ther. 2004, 309, 758.
10. Liu, Y.; Wada, R.; Yamashita, T.; Mi, Y.; Deng, C.-X.;
Hobson, J. P.; Rosenfeldt, H. M.; Nava, V. E.; Chae, S.-S.;
Lee, M.-J.; Liu, C. H.; Hla, T.; Spiegel, S.; Proia, R. L. J.
Clin. Invest. 2000, 106, 951.
3.5
3.0
2.5
2.0
Alcohol
1.5
Phosphate
1.0
0.5
0.0
4
24
4
24
4
24
Alcohol
Phosphate Thiophosphate
11. 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. J. Biol. Chem. 2002, 277, 21453.
Figure 3. Plasma concentrations of compounds 11, 12, and 13.
Compounds were dissolved in 3% fatty acid free-BSA and injected ip
at 8 mg/kg. Blood was drawn from the orbital sinus at 4 and 24 h;
plasma concentrations were determined using a LCQ LC–MS. Results
are means of five measurements.
12. Graler, M. H.; Goetzl, E. J. FASEB 2004, 18, 551.

4474
F. W. Foss, Jr. / Bioorg. Med. Chem. Lett. 15 (2005) 4470–4474
13. Matloubian, M.; Lo, C. G.; Cinamon, G.; Lesneski, M. J.;
Xu, Y.; Brinkmann, V.; Allende, M. L.; Proia, R. L.;
Cyster, J. G. Nature 2004, 427, 360.
14. Sanna, M. G.; Liao, J.; Jo, E.; Alfonso, C.; Ahn, M.-Y.;
Peterson, M. S.; Webb, B.; Lefebvre, S.; Chun, J.; Gray,
N.; Rosen, H. J. Biol. Chem. 2004, 279, 13839.
15. Clemens, J. J.; Davis, M. D.; Lynch, K. R.; Macdonald, T.
L. Bioorg. Med. Chem. Lett. 2003, 13, 3401.
16. Rosen, H.; Sanna, G.; Alfonso, C. Immunol. Rev. 2003,
195, 160.
17. Paugh, S. W.; Payne, S. G.; Barbour, S. E.; Milstien, S.;
Spiegel, S. FEBS Lett. 2003, 544, 189.
18. Billich, A.; Bornancin, F.; Devay, P.; Mechtcheriakova, D.;
Urtz, N.; Baumruker, T. J. Biol. Chem. 2003, 278, 47408.
19. Kiuchi, M.; Adachi, K.; Kohara, T.; Minoguchi, M.;
Hanano, T.; Aoki, Y.; Mishina, T.; Arita, M.; Nakao, N.;
Ohtsuki, M.; Hoshino, Y.; Teshima, K.; Chiba, K.;
Sasaki, S.; Fujita, T. J. Med. Chem. 2000, 43, 2946.
20. Hale, J. J.; Neway, W.; Mills, S. G.; Hajdu, R.; Keohane,
C. A.; Rosenbach, M.; Milligan, J.; Shei, G-J.; Chrebet,
G.; Bergstrom, H.; Card, D.; Koo, G. C.; Koprak, S. L.;
Jackson, J. J.; Rosen, H.; Mandala, S. Bioorg. Med. Chem.
Lett. 2004, 14, 3351.
with detection by electrospray ionization mass spec-
trometry (LCQ-Classic, ThermoFinnigan), similar to a
previously published procedure.23 In summary, plasma
samples were purified using a Strata X solid-phase
extraction cartridge (Phenomenex) that had been pre-
equilibrated with methanol and ultrapure water (Barn-
sted). The plasma samples and internal standards were
loaded and washed with ultrapure water. The analytes
of interest were eluted with 50:50 methanol/acetonitrile
under vacuum and concentrated by lyophilization
(Savant). Partially purified samples were reconstituted
in 30 lL acetonitrile, injected into a Phenomenex phen-
yl-hexyl column (5 lm, 50· 2.00 mm) using a Water
2695 HPLC system, and separated using a gradient
method of 0.02% trifluoroacetic acid in water and
acetonitrile. Initial composition consisted of 0.02%
trifluoroacetic acid in water and the concentration of
acetonitrile was increased to 100% over 25 min. Samples
were analyzed in positive ion mode, with cations
monitored by their appropriate mass to charge ratios.
23. Allende, M. L.; Sasaki, T.; Kawai, H.; Olivera, A.; Mi, Y.;
van Echten-Dekert, G.; Hajdu, R.; Rosenbach, M.;
Keohane, C. A.; Mandala, S.; Spiegel, S.; Proia, R. L.
J. Biol. Chem. 2004, 279, 52487.
21. Durand, P.; Peralba, P.; Sierra, F.; Renaut, P. Synthesis
2000, 4, 505.
22. Plasma levels of compounds 11, 12, and 13 were
determined by high performance liquid chromatography
24. This conversion was confirmed by the comparison of MS
data obtained by direct injection of 13 with the MS data
obtained post-LC elution of the pure compound.