THE JOURNAL OF BIOLOGICAL CHEMISTRY
14578 –14585, 2003
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Sphingosine-1-phosphate Lyase Is Involved in the Differentiation of
F9 Embryonal Carcinoma Cells to Primitive Endoderm*
Received for publication, November 8, 2002, and in revised form, February 3, 2003
Published, JBC Papers in Press, February 12, 2003, DOI 10.1074/jbc.M211416200
Akio Kihara‡, Mika Ikeda‡, Yuki Kariya‡, Eun-Young Lee§, Yong-Moon Lee§,
and Yasuyuki Igarashi‡¶
From the ‡Department of Biomembrane and Biofunctional Chemistry, Graduate School of Pharmaceutical Sciences,
Hokkaido University, Kita 12-jo, Nishi 6-choume, Kita-ku, Sapporo 060-0812, Japan and the §College of Pharmacy,
Chungbuk National University, Chongju 361-763, South Korea
Sphingosine 1-phosphate (S1P) is a bioactive lipid
molecule that acts both extracellularly and intracellu-
larly. The SPL gene encodes a mammalian S1P lyase
that degrades S1P. Here, we have disrupted the SPL
gene in mouse F9 embryonal carcinoma cells by gene
targeting. This is the first report of gene disruption of
mammalian S1P lyase. The SPL-null cells exhibited no
S1P lyase activity, and intracellular S1P was increased
⬃2-fold, compared with wild-type cells. Treatment of F9
embryonal carcinoma cells with retinoic acid induces
differentiation to primitive endoderm (PrE). An acceler-
ation in this PrE differentiation was observed in the
SPL-null cells. This effect was apparently caused by the
accumulated S1P, since N,N-dimethylsphingosine, a S1P
synthesis inhibitor, had an inhibitory effect on the PrE
differentiation. Moreover, F9 cells stably expressing
sphingosine kinase also exhibited an acceleration in the
differentiation. Exogenous S1P had no effect on differ-
entiation, indicating that intracellular but not extracel-
lular S1P is involved. Moreover, we determined that
expression of the SPL protein is up-regulated during the
progression to PrE. We also showed that sphingosine
kinase activity is increased in PrE-differentiated cells.
These results suggest that intracellular S1P has a role in
the PrE differentiation and that SPL may be involved in
the regulation of intracellular S1P levels during this
differentiation.
Sphingosine 1-phosphate (S1P)1 is a sphingolipid metabolite
that functions as both an extracellular and intracellular sig-
naling mediator in regulating diverse biological processes such
as proliferation, differentiation, apoptosis, and cell motility
* This work was supported in part by a Grant-in-aid for Scientific
Research on Priority Areas (B) 12140201 from the Ministry of Educa-
tion, Culture, Sports, Science and Technology of Japan and by Ono
Pharmaceutical Co., Inc. The costs of publication of this article were
defrayed in part by the payment of page charges. This article must
therefore be hereby marked “advertisement” in accordance with 18
U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence should be addressed: Dept. of Biomem-
brane and Biofunctional Chemistry, Graduate School of Pharmaceuti-
cal Sciences, Hokkaido University, Kita 12-jo, Nishi 6-choume, Kita-ku,
Sapporo 060-0812, Japan. Tel.: 81-11-706-3970; Fax: 81-11-706-4986;
E-mail: yigarash@pharm.hokudai.ac.jp.
1
The abbreviations used are: S1P, sphingosine-1-phosphate; Edg,
endothelial differentiation gene; Sph, sphingosine; PLP, pyridoxal 5⬘-
phosphate; EC, embryonal carcinoma; PrE, primitive endoderm; RA,
retinoic acid; Dab2, Disabled-2; bt2cAMP, dibutyryl cyclic AMP; PE,
parietal endoderm; SPL, S1P lyase; HA, hemagglutinin; SPHK1a,
sphingosine kinase 1a; RT, reverse transcription; PBS, phosphate-buff-
ered saline; DMS, N,N-dimethylsphingosine; ERK, extracellular signal-
regulated kinase; HPLC, high-performance liquid chromatography;
contig, group of overlapping clones.
14578
Vol. 278, No. 16, Issue of April 18, pp.
Printed in U.S.A.
(1–3). Extracellular effects of S1P are mediated via members of
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
the endothelial differentiation gene (Edg)/S1P receptor family
(Edg1/S1P1, Edg3/S1P3, Edg5/S1P2, Edg6/S1P4, and Edg8/
S1P5). These receptors are coupled distinctly (via Gq-, Gi-,
G12/13-, and Rho-dependent routes) to multiple downstream
signaling pathways including those associated with adenylate
cyclase, MAP kinase, phospholipases C and D, c-Jun N-termi-
nal kinase, and nonreceptor tyrosine kinase (2, 4, 5). Intracel-
lularly, S1P has been implicated in inositol trisphosphate-in-
dependent calcium mobilization, inhibition of caspase activity,
and activation of nonreceptor tyrosine kinases (2, 3).
S1P is formed through the phosphorylation of sphingosine
(Sph), catalyzed by Sph kinase. Once formed, S1P is rapidly
cleaved by S1P lyase to hexadecenal and phosphoethanolamine
or dephosphorylated by S1P phosphohydrolase (Fig. 1). Hence,
intracellular S1P levels are determined by the balance of Sph
kinase-mediated synthesis and its degradation by S1P lyase or
S1P phosphohydrolase. Platelets, which possess high Sph ki-
nase activity and lack S1P lyase activity, accumulate S1P
abundantly (6); consequently, S1P lyase is thought to play a
central role in keeping intracellular S1P levels low. Identical to
S1P but lacking the 4,5-trans double bond, another sphingo-
lipid biosynthesis intermediate dihydrosphingosine (dihydro-
Sph) can also be phosphorylated by Sph kinase to dihy-
drosphingosine-1-phosphate (dihydro-S1P). Dihydro-S1P binds
to Edg receptors and activates them, yet does not mimic other
effects of S1P such as cell survival (7).
S1P lyase is a pyridoxal 5⬘-phosphate (PLP)-dependent en-
zyme with a conserved pyridoxal-dependent decarboxylase do-
main positioned at the middle of the protein (Fig. 2A). Recently,
S1P lyase has been identified in several organisms including
Saccharomyces cerevisiae, Dictyostelium discoideum, and Cae-
norhabditis elegans, and in mammalian cells (8 –11). Mutant
analyses demonstrated that yeast strains lacking S1P lyase
(Bst1p/Dpl1p) exhibit resistance to heat stress and unregulated
proliferation upon approaching the stationary phase (12, 13).
The disruption in the S1P lyase gene (sglA) of D. discoideum
resulted in a mutant strain with an increased viability during
the stationary phase (10). The sglA null mutant also had a
strong developmental phenotype and produced aberrant fruit-
ing bodies, with short thick stalks and no obvious apical spore
mass (10, 14).
Mouse F9 embryonal carcinoma (EC) cells are a useful model
system for studying the mechanism of endoderm differentia-
tion in mouse early embryogenesis. F9 cells can be induced to
differentiate to primitive endoderm (PrE) by the addition of
retinoic acid (RA) (15). PrE cells express several specific mark-
ers such as tissue plasminogen activator, Type IV collagen,
c-jun, cytokeratin ENDO A, and Disabled-2 (Dab2)/DOC-2 (15–
19). The subsequent addition of dibutyryl cyclic AMP
This paper is available on line at http://www.jbc.org
 Involvement of SPL in Differentiation
FIG. 1. Pathway of sphingolipid metabolism. Shown are the
pathways for de novo sphingolipid biosynthesis as well as the sphingo-
lipid-to-glycerolipid conversion. In the sphingolipid biosynthetic path-
way, ceramide is converted to sphingomyelin or one of hundreds of
glycosphingolipids. Dihydro-Sph, a sphingolipid biosynthesis interme-
diate, and Sph, a sphingomyelin metabolite, are phosphorylated to
dihydro-S1P and S1P, respectively, by Sph kinase. S1P lyase converts
dihydro-S1P and S1P to fatty aldehydes (hexadecanal and hexadecenal,
respectively) and phosphoethanolamine, all of which then enter the
glycerolipid pathways of metabolism.
(bt2cAMP) induces further differentiation of PrE cells to pari-
etal endoderm (PE) (20). Differentiation to PE induces expres-
sion of thrombomodulin (21), and the levels of tissue type
plasminogen activator, Type IV collagen, and laminin are
increased (16, 20).
To investigate the role of mouse S1P lyase (SPL) as well as
that of intracellular S1P in the differentiation processes, we
have generated SPL knockout F9 cells by homologous recom-
bination. The SPL⫺/⫺ cells possess no S1P lyase activity and
accumulate intracellular S1P and dihydro-S1P. RA-induced
PrE differentiation of the SPL⫺/⫺cells was accelerated com-
pared with the wild-type cells. Moreover, expression of SPL as
well as Sph kinase activity was up-regulated by RA treatment.
These results suggest that SPL and intracellular S1P play roles
in PrE differentiation.
EXPERIMENTAL PROCEDURES
Cell Culture—Mouse F9 EC cells were grown in Dulbecco’s modified
Eagle’s medium (D6429; Sigma) containing 10% fetal calf serum sup-
plemented with 100 units/ml penicillin and 100 ␮g/ml streptomycin in
0.1% gelatin-coated dishes. For differentiation experiments, 1 ␮M all-
trans-RA (Sigma) and 250 ␮M bt2cAMP (Sigma) were added to the
medium.
Plasmids—The SPL genomic fragments were obtained by PCR using
F9 genomic DNA as follows. The SPL genomic regions corresponding to
exons 7–9 (2.1 kb), exons 9 –11 (2.4 kb), exons 11–12 (0.6 kb), and exons
12–14 (1.8 kb) were amplified using the primers (5⬘-GGCGTTAGAGA-
AGGGGATCAAAACTCC-3⬘ and 5⬘-GCATAGCTGTGTTCCTGGAGAT-
GGC-3⬘, 5⬘-GCCATCTCCAGGAACACAGCTATGC-3⬘ and 5⬘-CCAGGC-
CGTGAGCCAGCTATGCTTGG-3⬘, 5⬘-CCAAGCATAGCTGGCTCACG-
GCCTGG-3⬘ and 5⬘-CGGTAAATGTCAAAATCGTTGGATCCC-3⬘, and
5⬘-GGGATCCAACGATTTTGACATTTACCG-3⬘ and 5⬘-CCTGTCAATG-
GTTGCCTGGGCCATGCC-3⬘, respectively), and these were then
connected. The targeting vectors (targeting vector Neo and targeting
vector Puro) were constructed by replacing a 105-bp fragment
corresponding to the middle of exon 9 to the ScaI site in intron 9 with
loxP site-flanked neomycin-resistant gene and puromycin-resistant
gene, respectively (Fig. 2A). The internal ribosome entry site of the
encephalomyocarditis virus, which permits the translation of two open
reading frames from one mRNA (22), was inserted upstream of the loxP
site-flanked puromycin-resistant gene in the targeting vector Puro.
pcDNA3-HA1, a derivative of pcDNA3 (Invitrogen), was constructed
to create an N-terminal hemagglutinin (HA)-tagged gene. pCE-puro, a
derivative of pCI-neo (Promega), was designed for protein expression
14579
under the control of the human elongation factor 1␣ promoter. For this
construction, the cytomegalovirus immediate early promoter of pCI-neo
was first replaced by the human elongation factor 1␣ promoter from
pCE-FL (a gift from S. Sugano; Tokyo University) by N. Mizushima
(National Institute for Basic Biology, Okazaki, Japan) to generate pCE-
neo. Then the neomycin-resistant gene was replaced by the puromycin-
resistant gene from pPGKpurobpA to generate pCE-puro.
The SPL cDNA was amplified by PCR using primers 5⬘-AGATCTC-
CCGGAACCGACCTCCTCAAGC-3⬘ and 5⬘-GTGTGCAGTCTGTTCCA-
AACGCC-3⬘ and F9 total cDNA as a template. The amplified fragments
were cloned into pGEM-T Easy (Promega) to generate pGEM-SPL
plasmid. The pcDNA3-HA-SPL plasmid, which encodes N-terminally
HA-tagged SPL, was constructed by cloning the 1.8 kb of the BglII-NotI
fragment of pGEM-SPL into the BamHI-NotI site of pcDNA3-HA1.
pCE-puro HA-SPL was then constructed by cloning of the HA-SPL
region of pcDNA3-HA-SPL into the pCE-puro plasmid.
The pcDNA3-HA-SPHK1a, which encodes N-terminally HA-tagged
mouse Sph kinase 1a (SPHK1a), was constructed by cloning of the 1.2
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
kb of the BamHI-EcoRI fragment of pcDNA3-FLAG-SPHK1a (23) into
the BamHI-EcoRI site of pcDNA3-HA1. pCE-puro HA-SPHK1a was
then constructed by cloning of the HA-SPHK1a region of pcDNA3
HA-SPHK1a into the pCE-puro plasmid.
Production of SPL⫺/⫺ F9 Cells and Stable Transformants—The
linearized targeting vector Neo (1 ␮g) was transfected into 4 ⫻ 105 F9
cells using LipofectAMINETM 2000 reagent (Invitrogen). Cells were
subjected to G418 selection at 900 ␮g/ml for 1 week. Homologous
recombination was examined by PCR amplification of genomic DNA
using primer A (5⬘-GTGACTTCTGGGGGAACGGAAAGC-3⬘), located
in exon 7 but outside the genomic sequences present in the targeting
vector, and primer B (5⬘-ATCGGAATTCCTCGAGTCTAGAGCG-3⬘),
located upstream of the loxP site (Fig. 2A). For isolation of F9 SPL⫺/⫺
cells, the heterozygous clone (F9-1) was transfected with the linear-
ized targeting vector Puro (1 ␮g). Cells were then cultured in the
presence of 0.5 ␮g/ml puromycin for 8 days. Genomic DNAs were
prepared from resistant clones, and homologous recombination was
examined by PCR using primer C (5⬘-GCCATCTCCAGGAACACAGC-
TATGC-3⬘), located in exon 9, and primer D (5⬘-CCAGGCCGTGAGC-
CAGCTATGCTTGG-3⬘), located in exon 11. One of the puromycin-
resistant clones, F9-2, exhibited an SPL⫺/⫺ genotype. F9-4 (SPL⫹/⫹,
neomycin- and puromycin-resistant) cells were obtained in the course
of this targeting procedure.
To obtain stable transformants of the HA-SPHK1a gene, the pCE-
puro HA-SPHK1a plasmid (1 ␮g) was transfected into 4 ⫻ 105 F9 cells
using LipofectAMINETM 2000 reagent. Cells were subjected to puromy-
cin selection at 0.5 ␮g/ml for 1 week. One of the stable transformants,
F9-9, expressed the highest level of HA-SPHK1a among the isolated
clones and was used for further analyses.
Reverse Transcription (RT) PCR—F9 total RNA, isolated using Trizol
reagent (Invitrogen), was converted to cDNA using oligo(dT) primer and
ProSTARTM first strand RT-PCR kit (Stratagene). The SPL cDNA was
amplified by PCR using primer E (5⬘-CCCGGAACCGACCTCCTCAAG-
CTGAAGG-3⬘), primer F (5⬘-GTGTGCAGTCTGTTCCAAACGCC-3⬘),
and F9 total cDNA as a template.
Antibodies—Anti-SPL antiserum was raised against recombinant
full-length SPL proteins expressed as hexahistidine-tagged fusion pro-
teins. Anti-Dab2, anti-HA Y-11, and anti-actin (A-2066) antibodies were
purchased from Transduction Laboratories (Lexington, KY), Santa
Cruz Biotechnologies, Inc. (Santa Cruz, CA), and Sigma, respectively.
Measurement of S1P Lyase Activity—Cells suspended in buffer A (phos-
phate-buffered saline (PBS), 1 mM EDTA, 1 mM phenylmethylsulfonyl
fluoride, 1⫻ protease inhibitor mixture (CompleteTM; Roche Molecular
Biochemicals), 1 mM dithiothreitol) were lysed by sonication and subjected
to in vitro S1P lyase assay as described previously (24). For use as a
substrate, [4,5-3H]dihydro-S1P was prepared by phosphorylation of [4,5-
3
H]dihydro-Sph (50 Ci/mmol; American Radiolabeled Chemical Inc.) us-
ing recombinant maltose-binding protein-fused mouse SPHK1a. After the
reaction, lipids were extracted by successive additions of a 3.75-fold vol-
ume of chloroform/methanol/HCl (100:200:1, v/v/v), a 1.25-fold volume of
chloroform, and a 1.25-fold volume of 1% KCl, with mixing. Phases were
then separated by centrifugation, and the organic phase was recovered,
dried, and suspended in chloroform/methanol (2:1, v/v). The labeled lipids
were resolved by TLC on Silica Gel 60 high performance TLC plates
(Merck) with 1-butanol/acetic acid/water (3:1:1, v/v/v).
Sph Kinase Assay—The Sph kinase assay was performed as de-
scribed previously (25).
Immunoblotting—Cells were washed with PBS twice, suspended in
buffer A, and sonicated. After centrifugation at 300 ⫻ g for 5 min at 4 °C,
the supernatant was treated with an equal volume of 10% (w/v) trichlo-
14580 Involvement of SPL in Differentiation

Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
FIG. 2. Targeted disruption of the SPL gene. A, structure of the SPL protein and schematic representation of the SPL targeting strategy.
TM, PDCD, and IRES, a putative transmembrane segment, a pyridoxal-dependent decarboxylase conserved domain, and an internal ribosome
entry site, respectively. Exons 7–14, the primers for genomic PCR (A–D), the loxP sites (closed triangle), the neomycin-resistant gene (Neor), the
puromycin-resistant gene (Puror), and internal ribosome entry site (IRES) are indicated. B, PCR amplification of genomic DNA. Genomic DNA
isolated from F9 (SPL⫹/⫹; lane 1), F9-1 (SPL⫹/⫺; lane 2), and F9-2 (SPL⫺/⫺; lane 3) cells was subjected to PCR using primers C and D. C, RT-PCR
analysis. The levels of SPL transcripts were monitored in F9 (lane 1) and F9-2 (lane 2) cells by RT-PCR as described under “Experimental
Procedures.” D, immunoblotting analysis. F9 cells were transfected with pCE-puro (vector; lanes 1 and 3) or pCE-puro HA-SPL (lanes 2 and 4), and
24 h later, total lysates were prepared. Fixed amounts of proteins (15 ␮g) were subjected to immunoblotting using anti-SPL antiserum (lanes 1 and
2) or anti-HA antibodies (lanes 3 and 4). Total proteins (20 ␮g) prepared from F9 (lane 5), F9-1 (lane 6), and F9-2 (lane 7) cells were separated by
SDS-PAGE and detected by immunoblotting with anti-SPL antiserum. The asterisk indicates nonspecific background. E, in vitro S1P lyase assay.
Total cell lysates (135 ␮g of protein) prepared from F9-4 (SPL⫹/⫹; lanes 1 and 2) and F9-2 (lanes 3 and 4) cells were incubated with 0.45 ␮Ci of
[3H]dihydro-S1P at 37 °C for 0.5 h (lanes 1 and 3) and 2 h (lanes 2 and 4). Lipids were extracted and separated by TLC. HD, hexadecanal. F and
G, metabolism of exogenously added [3H]Sph. F9-4 (lane 1) and F9-2 (lane 2) cells in 1 ml of culture medium were treated with 0.8 ␮Ci of [3H]Sph
plus cold Sph (total 100 pmol) at 37 °C for 1 h. Lipids were extracted and separated by TLC. In G, quantitative results for the amounts of S1P,
ceramide, and sphingomyelin in F9-2 cells (open columns) relative to those in F9-4 cells (closed columns) are shown. Values represent the mean ⫾
S.D. from three independent experiments. CER, ceramide; SM, sphingomyelin. H, measurement of intracellular amounts of S1P (closed columns)
and dihydro-S1P (open columns) in F9-2 and F9-4 cells using HPLC. Values represent the mean ⫾ S.D. from three independent experiments.

roacetic acid and incubated for 20 min at 0 °C. Protein precipitates were
washed with acetone and suspended in buffer B (50 mM Tris-HCl (pH 8.0),
1 mM EDTA, 1% SDS). After quantification of protein concentrations
using a BCA protein assay kit (Pierce), samples were diluted with equal
volumes of 2⫻ SDS sample buffer (125 mM Tris-HCl (pH 6.8), 4% SDS,
20% glycerol, 10% 2-mercaptoethanol, and a trace amount of bromphenol
blue). Proteins were separated by SDS-PAGE and transferred to Immo-
bilonTM polyvinylidene difluoride membrane (Millipore Corp.). The result-
ing membrane was incubated with primary antibody (anti-SPL anti-
serum, anti-Dab2 antibodies, or anti-HA Y-11 antibodies, each diluted
1:1000; or anti-actin antibodies, diluted 1:200) for 1 h and then with
secondary antibody (peroxidase-conjugated donkey anti-rabbit IgG
F(ab⬘)2 fragment or sheep anti-mouse IgG F(ab⬘)2 fragment, both from
Amersham Biosciences, and diluted 1:7500) for 1 h. Labeling was detected
by the ECL detection method (Amersham Biosciences).
[3-3H]S1P and [3-3H]Sph Labeling Experiments—[3-3H]S1P was pre-
pared by phosphorylation of [3-3H]Sph (23.5 Ci/mmol; PerkinElmer Life
Sciences) using recombinant maltose-binding protein-fused mouse
SPHK1a. Cells at ⬃70% confluence in six-well dishes were incubated
with 1 ml/well medium containing 0.8 ␮Ci of [3-3H]Sph plus cold Sph
(total 100 pmol of Sph) or 0.8 ␮Ci of [3-3H]S1P plus cold S1P (total 100
pmol), which had been complexed with 1 mg/ml fatty acid-free bovine
serum albumin (Sigma catalog no. A-6003), for 1 h at 37 °C. Lipids were
extracted and separated by TLC as described above.
Measurement of Intracellular S1P Level by HPLC—Intracellular
amounts of S1P and dihydro-S1P were measured by HPLC as described
previously (26).
Phalloidin Staining—F9 cells grown on coverslips were fixed with 3.7%
formaldehyde in PBS at 37 °C for 10 min, permeabilized with 0.1% Triton
X-100 in PBS, blocked with blocking solution (10 mg/ml bovine serum
albumin in PBS), and incubated at room temperature for 30 min with
Alexa FluorTM 488 phalloidin (Molecular Probes, Inc., Eugene, OR). Cells
were washed, mounted with Slow Fade Light Antifade Kit (Molecular
Probes), and observed under a fluorescence microscope (Axiophot 2 Plus;
Carl Zeiss) with a plane-APOCHROMAT lens (⫻63) (Carl Zeiss).
 Involvement of SPL in Differentiation
RESULTS
Generation of F9 SPL⫺/⫺ Cells—To inactivate the SPL gene,
we constructed targeting vectors for homologous recombina-
tion. To obtain genomic information about the SPL gene, the
mouse genome data base MGSCV3 was searched using the
BLAST program. We found that the Mus musculus WGS
supercontig Mm10_WIFeb01_211 (accession number NW_
000027) contains the SPL gene. This sequence data revealed
that the SPL gene is located on chromosome 10, and the open
reading frame consists of 14 exons. Based on this sequence
information, we obtained a SPL genomic fragment that corre-
sponds to exons 7–14. We constructed two targeting vectors in
which exon 9 was replaced by a loxP site-flanked neomycin-
resistant gene (targeting vector Neo) or a puromycin-resistant
gene (targeting vector Puro) (Fig. 2A). We utilized a promoter
selection strategy (i.e. selection markers lacking their promot-
ers and transcribed under the control of the SPL promoter only
when properly targeted). To facilitate internal translation of
the puromycin-resistant gene, we introduced an internal ribo-
some entry site upstream of the puromycin-resistant gene in
the targeting vector Puro. The SPL protein is a PLP-dependent
enzyme and contains a conserved pyridoxal-dependent decar-
boxylase domain positioned from amino acid 167 to amino acid
452 (Fig. 2A). In a PLP-dependent enzyme, PLP exists as a
Schiff base with its aldehyde group forming an imine linkage
with the ⑀-amino group of a lysine residue. In SPL, Lys-353 is
predicted to be the Schiff base-forming residue. Since Lys-353
is encoded within exon 10, the disruption of exon 9 was ex-
pected to render the C-terminally truncated SPL completely
inactive.
To target the first allele of the SPL gene, F9 cells were
transfected with the targeting vector Neo. One of the neomycin-
resistant clones, named F9-1, showed a SPL⫹/⫺ genotype con-
firmed by PCR (data not shown) using its genomic DNA and
primers A and B shown in Fig. 2A. F9-1 cells were then trans-
fected with the targeting vector Puro, and a clone containing
the second targeted SPL allele (named F9-2) was isolated. Fig.
2B shows the result of PCR analysis using the genomic DNA,
primer C located in exon 9, and primer D located in exon 11.
Although only a 2.4-kb fragment was amplified from genomic
DNA prepared from F9 cells (Fig. 2B, lane 1), both a 2.4-kb and
a 3.8-kb DNA fragment were detected in PCR products from
F9-1 cells (Fig. 2B, lane 2). On the other hand, a 3.8-kb frag-
ment and a faint 4.8-kb fragment were amplified from genomic
DNA of F9-2 cells, but the 2.4-kb fragment was not detected
(Fig. 2B, lane 3). RT-PCR analysis demonstrated the loss of
SPL mRNA in F9-2 cells (Fig. 2C).
Next, we prepared an antibody against recombinant full-
length mouse SPL to detect the SPL protein. Fig. 2D shows the
immunoblotting analysis of the total lysates prepared from F9
cells transfected with HA-SPL cDNA. The anti-SPL antiserum,
as well as anti-HA antibodies, revealed that the HA-SPL pro-
tein migrated slightly faster (59 kDa) than the predicted mo-
lecular mass (65.0 kDa) (Fig. 2D, lanes 2 and 4). The lysates of
the vector transfectant did not produce this band (Fig. 2D,
lanes 1 and 3). The anti-SPL antiserum also detected the en-
dogenous SPL protein in the lysates of the vector- and HA-SPL
cDNA-transfected and untransfected F9 cells as a 58-kDa band
(Fig. 2D, lanes 1, 2, and 5). However, the SPL protein was
reduced in F9-1 (mSPL⫹/⫺) cells and was absent in F9-2
(mSPL⫺/⫺) cells (Fig. 2D, lanes 6 and 7). Above all, these
results confirmed the proper targeting in F9-2 cells. In the
course of the targeting experiments, we also obtained F9 cells
resistant to both neomycin and puromycin but carrying an
intact SPL gene. We used these cells (F9-4), which always
14581
exhibited the same phenotype as original F9 cells, as a wild-
type control for further analyses as indicated.
To investigate the S1P lyase activity in the SPL⫺/⫺ cells, we
performed an in vitro S1P lyase assay using total cell lysates
and [4,5-3H]dihydro-S1P. Wild-type cell lysates converted di-
hydro-S1P to hexadecanal in a time-dependent manner (Fig.
2E, lanes 1 and 2), also generating dihydro-Sph by the action of
the phosphohydrolase. In contrast, lysates from SPL⫺/⫺ cells
converted dihydro-S1P only to dihydro-Sph and displayed no
S1P lyase activity (Fig. 2E, lanes 3 and 4), indicating that SPL
is the sole S1P lyase in F9 cells.
We next examined the intracellular accumulation of S1P
using exogenously added [3-3H]Sph. After a 1-h incubation
with [3-3H]Sph, lipids were extracted and separated by TLC.
Wild-type cells accumulated only a small amount of S1P and
converted most of the Sph to ceramide and sphingomyelin (Fig.
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
2F, lane 1). The SPL⫺/⫺ cells showed an increased accumula-
tion of S1P by 3.4-fold compared with wild-type cells, whereas
conversion to ceramide and sphingomyelin in the SPL⫺/⫺ cells
was indistinguishable from that in wild-type cells (Figs. 2, F
and G). Next, we measured steady-state levels of S1P and
dihydro-S1P using HPLC. The SPL⫺/⫺ cells showed an increase
in intracellular S1P (about 2-fold) and dihydro-S1P (about 2.5-
fold) compared with the SPL⫹/⫹ cells (Fig. 2H; see also Fig. 7E).
Disrupting the SPL Gene Accelerates the Differentiation of F9
Cells—We examined the effect of SPL gene disruption on cell
growth and morphology. The growth rate of SPL⫺/⫺ cells was
only slightly reduced compared with that of wild-type cells
(data not shown). The morphology of SPL⫺/⫺ cells was indis-
tinguishable from that of wild-type cells (data not shown).
Recently, it was reported that S1P lyase has a central role in
the development of D. discoideum (10). With this in mind, we
investigated the role of SPL in the differentiation of F9 cells to
PrE and, subsequently, PE in the presence of 1 ␮M RA and 250
␮M bt2cAMP. Fig. 3A shows phase-contrast images and phal-
loidin-staining patterns of EC, PrE, and PE cells. PrE cells
manifest an enlarged and flattened morphology, whereas PE
cells exhibit rounded shapes with long cell processes (15, 20)
(Fig. 3A). By day 3 of treatment, all of the wild-type cells had
differentiated to a cell type with typical PrE morphology (Fig.
3B), whereas almost no cells (⬍1%) had differentiated to PE.
PE morphology was observed within small areas of the culture
at day 4 (5%), and about 78% of the cells had differentiated to
PE at day 5 (Fig. 3B). In contrast, SPL⫺/⫺ cells that had
differentiated to exhibit PE morphology could be detected even
at day 3 (21%) (Fig. 3B). Moreover, most of the SPL⫺/⫺ cells by
day 4 (83%) and day 5 (95%) had differentiated to PE (Fig. 3B),
further establishing that differentiation to PE was accelerated
by the disruption of the SPL gene.
It is possible that the accelerated PE differentiation observed
in the SPL⫺/⫺ cells was due to acceleration from EC cells to PrE
cells, acceleration from PrE cells to PE cells, or both. To dis-
tinguish between these possibilities, we next examined
whether differentiation from EC cells to PrE cells was acceler-
ated in the SPL⫺/⫺ cells, using the differentiation-specific
marker Dab2, a candidate tumor suppressor of breast and
ovarian tumors (27, 28). A previous study demonstrated that
PrE differentiation is accompanied by the expression of two
spliced isoforms of Dab2, p96 and p67 (19, 29). p96 binds to the
Src homology 3 domains of Grb2 and may function to modulate
Ras pathways by competing with Sos for binding to Grb2 (30).
Unlike p96, p67 largely resides in nuclei and may function as a
transcriptional co-factor (31). We prepared total cell lysates
from the wild-type cells and the SPL⫺/⫺ cells treated with RA
and bt2cAMP for 0 – 6 days and examined the expression levels
of Dab2 (Fig. 3C). Although both p96 and p67 appeared only at
14582 Involvement of SPL in Differentiation
FIG. 3. Effects of SPL knockout on differentiation. A, morphol-
ogy of F9 EC, PrE, and PE cells. F9 EC cells were differentiated to PrE
cells and PE cells by incubation with 1 ␮M RA and by incubation with
1 ␮M RA and 250 ␮M bt2cAMP, respectively, for 5 days. Cells were then
fixed, permeabilized, and stained with phalloidin to visualize F-actin.
Left panels, phase-contrast images; right panels, phalloidin-staining. B,
kinetics of differentiation of F9-4 (SPL⫹/⫹) and F9-2 (SPL⫺/⫺) cells.
Cells at a density of 400 cells/cm2 were incubated with 1 ␮M RA and 250
␮M bt2cAMP for the indicated time periods. Cells were photographed
under a phase-contrast microscope at ⫻200 magnification. The arrows
indicate PE-differentiated cells. C, expression of Dab2 is up-regulated
in SPL⫺/⫺ cells. F9-2 (lanes 1, 3, 5, 7, 9, 11, and 13) and F9-4 (lanes 2,
4, 6, 8, 10, 12, and 14) cells were cultured in the presence of 1 ␮M RA and
250 ␮M bt2cAMP for the indicated time periods. Total proteins (10 ␮g)
were separated by SDS-PAGE, followed by immunoblotting with anti-
Dab2 antibodies or, to demonstrate uniform protein loading, anti-actin
antibodies.
day 3 in the wild-type cells (Fig. 3C, lane 8), they could both be
detected even at day 2 in the SPL⫺/⫺ cells (Fig. 3C, lane 5).
Moreover, expression of Dab2 was up-regulated in SPL⫺/⫺
cells, compared with that in the wild-type cells, at day 3 (Fig.
3C, lanes 7 and 8). We also obtained similar results when these
cells were incubated with only RA, which induces differentia-
tion of F9 EC cells to PrE cells but not further into PE cells
(data not shown). These results confirmed that the differenti-
ation of EC cells to PrE cells was accelerated in SPL⫺/⫺ cells.
Next, we investigated whether differentiation of PrE cells to
PE is also enhanced by disruption of the SPL gene. For this
purpose, we first cultured both the wild-type and the SPL⫺/⫺
cells in medium containing only RA for 5 days to allow them to
differentiate to PrE, and then bt2cAMP was added to the cul-
tures. Within 1 day after the addition of bt2cAMP, 50% of the
wild-type cells had differentiated to PE (Fig. 4, left upper pan-
el). More importantly, the SPL⫺/⫺ cells had also differentiated
to PE at a frequency similar to the wild-type cells (48%) (Fig. 4,
FIG. 4. Disruption of SPL has no effect on differentiation of
PrE cells to PE cells. F9-4 (SPL⫹/⫹; left panels) and F9-2 (SPL⫺/⫺;
right panels) cells were first differentiated to PrE cells by incubation
with 1 ␮M RA for 5 days. These cells were then treated with 1 ␮M RA
and 250 ␮M bt2cAMP for 1 day in the absence (upper panels) or presence
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
of 1.5 ␮M DMS (lower panels). Cells were photographed under a phase-
contrast microscope at ⫻200 magnification. The arrows indicate PE-
differentiated cells.
right upper panel), indicating that SPL is not involved in the
differentiation of PrE cells to PE.
Disruption of the SPL gene has two effects: the accumulation
of S1P (and dihydro-S1P) and the cessation of the sphingolipid-
to-glycerolipid pathway. To examine which of these is the cause
of the accelerated differentiation, we utilized a Sph kinase
inhibitor, N,N-dimethylsphingosine (DMS). Given that S1P ex-
hibited a positive effect on the differentiation, it would be
expected that DMS would inhibit the differentiation. In con-
trast, if one of the S1P metabolites such as hexadecanal or
phosphoethanolamine exerts a negative effect on the differen-
tiation, DMS may stimulate the differentiation in wild-type
cells and may have no effect on the differentiation of the
SPL⫺/⫺ cells. We cultured the wild-type and the SPL⫺/⫺ cells in
the presence of RA and bt2cAMP with or without DMS. Since
high concentrations of DMS can inhibit other protein kinases
(32), we used DMS at low concentrations (1–2 ␮M). Within this
range, DMS cannot inhibit Sph kinase activity completely (33),
yet even 1.5 ␮M DMS had an inhibitory effect on the differen-
tiation (Fig. 5A). Although in the absence of DMS 80% of the
SPL⫺/⫺ cells differentiated to PE by day 4, those incubated
with DMS differentiated at a very low frequency (3%) (Fig. 5A,
upper panels). Similar results were obtained from experiments
using wild-type cells. Although the wild-type cells differenti-
ated to PE efficiently by day 5 (82%) in the absence of DMS,
most still exhibited PrE morphology in the presence of DMS at
day 5 (PE cells, 1%) (Fig. 4A, lower panels). We also investi-
gated the effect of DMS on the expression of Dab2. Both the p96
and p67 forms of Dab2 were reduced by treatment with DMS in
a dose-dependent manner (Fig. 5B) In contrast, DMS had no
effect on differentiation of already PrE-differentiated wild-type
and SPL⫺/⫺ cells to PE (Fig. 4, lower panels).
The above results suggested that S1P (dihydro-S1P), but not
its metabolites, plays a role in the EC-to-PrE differentiation. To
confirm this, we investigated the effects of overproduction of
Sph kinase on the differentiation. Previous study has revealed
that overexpression of Sph kinase results in an increase in
intracellular amounts of S1P (34). We established F9 clones
(F9-9) stably expressing HA-tagged mouse SPHK1a. We
detected HA-SPHK1a in the lysates of F9-9 cells by immuno-
blotting with anti-HA antibodies (Fig. 5C, lane 2). Then we
investigated the differentiation status of the cells by immuno-
blotting with anti-Dab2 antibodies. F9-9 cells treated with RA
and bt2cAMP for 2 days expressed an increased amount of
Dab2 compared with the control F9 cells (Fig. 5C, lanes 3 and
4). Thus, PrE differentiation was accelerated in the presence of
higher levels of S1P, whether they resulted from SPL disrup-
tion or overexpression of Sph kinase.
 Involvement of SPL in Differentiation
FIG. 5. S1P plays a role in the PrE differentiation. A, F9-2
(SPL⫺/⫺; upper panels) and F9-4 (SPL⫹/⫹; lower panels) cells, each at a
density of 400 cells/cm2, were incubated with 1 ␮M RA and 250 ␮M
bt2cAMP for the indicated time periods in the absence (left panels) or
presence of 1.5 ␮M DMS (right panels). Cells were photographed under
a phase-contrast microscope at ⫻200 magnification. The arrows indi-
cate PE-differentiated cells. B, expression of Dab2 is inhibited by DMS.
F9-2 (lanes 1– 4) and F9-4 (lanes 5– 8) cells at concentrations of 12,000
cells/cm2 and 3,000 cells/cm2, respectively, were treated with 1 ␮M RA,
250 ␮M bt2cAMP, and 0 ␮M (lanes 1 and 5), 1 ␮M (lanes 2 and 6), 1.5 ␮M
(lanes 3 and 7), or 2 ␮M DMS (lanes 4 and 8) and cultured for the
indicated time periods. Total proteins (15 ␮g) were separated by SDS-
PAGE, followed by immunoblotting with anti-Dab2 antibodies. C, PrE
differentiation is accelerated by overexpression of Sph kinase. F9 (lanes
1 and 3) and F9-9 (HA-SPHK1a stable transformants; lanes 2 and 4)
cells, each at a concentration of 12,000 cells/cm2, were incubated with 1
␮M RA and 250 ␮M bt2cAMP for 2 days, and lysates were prepared.
Total proteins (15 ␮g) were separated by SDS-PAGE, followed by im-
munoblotting with anti-HA (lanes 1 and 2) or anti-Dab2 (lanes 3 and 4)
antibodies.
S1P can act both extracellularly, via Edg/S1P family recep-
tors, and intracellularly. Therefore, the possibility that S1P
accumulated in the SPL⫺/⫺ cells is released into the medium
and is acting extracellularly in an autocrine or paracrine fash-
ion cannot be excluded. To distinguish whether S1P stimulates
the differentiation intracellularly or extracellularly, we inves-
tigated the effects of exogenously added S1P or dihydro-S1P on
the F9 differentiation. The wild-type and SPL⫺/⫺ cells were
incubated with S1P or dihydro-S1P at 0.1 or 1 ␮M for 2 days.
Then total cell lysates were prepared and subjected to immu-
noblotting using anti-Dab2 antibodies. As shown in Fig. 6A,
neither S1P nor dihydro-S1P treatment enhanced the expres-
sion of Dab2 in either cells; rather, these compounds were
slightly inhibitory at 1 ␮M. Moreover, the appearance of PE
morphology was also unchanged by S1P or dihydro-S1P (data
not shown). Thus, exogenous S1P or dihydro-S1P did not in-
duce differentiation, indicating that the lipids act intracellu-
larly to stimulate the differentiation.
To examine whether extracellular S1P can be imported into
the cell, thereby causing an accumulation of intracellular S1P,
we next performed a [3H]S1P labeling experiment. Although
[3H]Sph was rapidly imported into the cells and converted to
ceramide and sphingomyelin (Fig. 6B, lane 1), the labeling by
[3H]S1P was very inefficient, and S1P was not detected (Fig.
6B, lane 2). A longer autoradiography exposure (Fig. 6B, lane 3)
14583
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
FIG. 6. Effects of exogenous S1P and dihydro-S1P on differen-
tiation. A, F9-4 (SPL⫹/⫹; lanes 1–5) and F9-2 (SPL⫺/⫺; lanes 6 –10)
cells, each at a concentration of 12,000 cells/cm2, were treated with 1 ␮M
RA, 250 ␮M bt2cAMP, and 0 ␮M S1P (lanes 1 and 6), 0.1 ␮M S1P (lanes
2 and 7), 1 ␮M S1P (lanes 3 and 8), 0.1 ␮M dihydro-S1P (lanes 4 and 9),
or 1 ␮M dihydro-S1P (lanes 5 and 10) and incubated for 2 days. Total
proteins (15 ␮g) were separated by SDS-PAGE and detected by immu-
noblotting with anti-Dab2 antibodies. B, F9 cells in 1 ml of culture
medium were treated with 100 pmol of Sph containing 0.8 ␮Ci of
[3H]Sph (lane 1) or with 100 pmol of S1P containing 0.8 ␮Ci of [3H]
S1P (lanes 2 and 3) at 37 °C for 1 h. Lipids were extracted, sepa-
rated by TLC, and visualized by autoradiography. Lane 3 represents a
longer autoradiography exposure of lane 2. CER, ceramide; SM,
sphingomyelin.
revealed a pattern of metabolism for S1P similar to that for
Sph.
Up-regulation of SPL and Sph Kinase Activities during PrE
Differentiation of F9 Cells—We next examined whether SPL
expression is regulated during differentiation. F9 cells were
cultured in the presence of RA and bt2cAMP for 0 – 6 days, and
SPL protein levels were measured in the cell lysates by immu-
noblotting. We found that SPL increased in a time-dependent
manner and reached maximal level at 5 days (Fig. 7A). Cells
incubated with RA alone for 5 days (PrE cells) accumulated the
same amount of SPL compared with cells incubated with RA
and bt2cAMP together (PE cells) (Fig. 7B), indicating that SPL
expression is up-regulated during differentiation of EC cells to
PrE. Consistent with these observed increases in SPL levels, an
in vitro S1P lyase assay using total cell lysates further demon-
strated that the S1P lyase activity of PrE cells was about 4-fold
higher than that of EC cells (data not shown).
Although intracellular S1P levels were only about 2-fold
higher in the SPL⫺/⫺ cells than in wild-type cells (Fig. 2H),
significant acceleration was observed in the differentiation of
the SPL⫺/⫺ cells (Fig. 3). This observation led us to consider the
possibility that differences in intracellular S1P levels between
the wild-type and SPL⫺/⫺ cells may be even greater after
differentiation. To investigate this possibility, we first meas-
ured Sph kinase activity in F9 EC and RA-induced PrE cells.
We found that PrE cells possessed 2.7-fold higher Sph kinase
activity than EC cells (Fig. 7C). A similar increase in Sph
kinase activity was observed in the SPL⫺/⫺ cells (Fig. 7C).
Thus, Sph kinase activity is also up-regulated during PrE
differentiation. Next, we investigated intracellular S1P accu-
mulation using [3H]Sph labeling. The wild-type and SPL⫺/⫺
cells were incubated with [3H]Sph for 1 h, and labeled mem-
brane was separated by TLC. The PrE-differentiated wild-type
cells accumulated only low amounts of S1P, nearly the same
levels as in the undifferentiated cells (Fig. 7D, lanes 1 and 2). In
contrast, the S1P levels increased 3.9-fold during the differen-
14584 Involvement of SPL in Differentiation
FIG. 7. Expression of the SPL protein and Sph kinase activity
are up-regulated during differentiation to PrE cells. A, kinetics of
intracellular SPL accumulation during differentiation. F9-4 (SPL⫹/⫹)
cells were cultured in the presence of 1 ␮M RA and 250 ␮M bt2cAMP for
0 (lane 1), 1 (lane 2), 2 (lane 3), 3 (lane 4), 4 (lane 5), 5 (lane 6), and 6
(lane 7) days. Total proteins (10 ␮g) were separated by SDS-PAGE,
followed by immunoblotting with anti-SPL antiserum or, to demon-
strate uniform protein loading, anti-actin antibodies. B, expression
levels of the SPL protein in different stages of differentiation. F9 cells
were incubated with no agents (lane 1) or with 1 ␮M RA (lane 2) or 1 ␮M
RA and 250 ␮M bt2cAMP (lane 3) for 5 days. Total proteins (15 ␮g) were
separated by SDS-PAGE and detected by immunoblotting with anti-
SPL antiserum. C, Sph kinase assay. F9-4 (SPL⫹/⫹) and F9-2 (SPL⫺/⫺)
cells were differentiated to PrE cells by incubation with 1 ␮M RA for 6
days. Total cell lysates were prepared from undifferentiated (EC) and
PrE-differentiated cells, and 50 ␮g of protein were subjected to an in
vitro Sph kinase assay using 50 ␮M D-erythro-Sph, [␥-32P]ATP, and 1
mM cold ATP for 15 min at 37 °C. Lipids were separated by TLC, and
radioactivities associated with S1P were quantified using a Phosphor-
Imager BAS2000 (Fuji Film). Values are illustrated relative to the Sph
kinase activity associated with F9-4 EC cells (0.625 ⫾ 0.076 pmol/mg/
min) and represent the mean ⫾ S.D. from three independent experi-
ments. SK, Sph kinase. D, metabolism of exogenous Sph in EC and PrE
cells. F9-4 (lanes 1 and 2) and F9-2 (lanes 3 and 4) cells incubated with
mock agent (lanes 1 and 3) or 1 ␮M RA for 6 days were treated with 0.8
␮Ci of [3H]Sph plus cold Sph (total 100 pmol) at 37 °C for 1 h. Lipids
were extracted and separated by TLC. CER, ceramide; SM, sphingomy-
elin. E, measurement of intracellular amounts of S1P using HPLC. F9-4
and F9-2 cells incubated with mock agent (closed columns) or 1 ␮M RA
(open columns) for 6 days were subjected to HPLC analysis. Values
represent the mean ⫾ S.D. from three independent experiments.
tiation of the SPL⫺/⫺ cells to PrE cells (Fig. 7D, lanes 3 and 4).
Next, we also investigated intracellular S1P levels using
HPLC. Again, the S1P levels were unchanged in the wild-type
cells during the differentiation, whereas a significant increase
in the S1P levels was observed in the SPL⫺/⫺ cells (Fig. 7E).
These results indicated that the intracellular amounts of S1P
are regulated by not only synthesis but also degradation during
PrE differentiation.
DISCUSSION
In this study, we constructed a SPL-deficient F9 cell line.
Until now, disruptants of the S1P lyase gene were generated
only in S. cerevisiae and D. discoideum (10, 12–14). Thus, this
is the first report of gene disruption of mammalian S1P lyase.
The SPL⫺/⫺ F9 cells constructed in this study had no S1P lyase
activity (Fig. 2E), indicating that the SPL protein is the sole
S1P lyase, at least in F9 cells and, most likely, in mammalian
cells. Since S1P is stored in high concentrations in platelets,
which lack S1P lyase activity but have S1P phosphohydrolase
activity, it is likely that S1P lyase has a central role in clearing
intracellular S1P. A previous study demonstrated that a 500-
fold overproduction of the Sph kinase SPHK1a resulted in only
a 4 – 8-fold increase in S1P levels (34). One possibility for these
moderate increases is that degradation by SPL keeps intracel-
lular S1P levels low. Consistent with this theory, a 2.6-fold
increase in Sph kinase activity during PrE differentiation re-
sulted in a 1.7-fold increase in S1P levels in the SPL⫺/⫺ cells
but not in the wild-type cells (Fig. 7E).
Disruption of the sglA gene, which encodes S1P lyase in
D. discoideum, affects multiple stages throughout develop-
ment, including the ability to form migrating slugs, the control
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
of cell type-specific gene expression, and terminal spore differ-
entiation (10). Here we demonstrated that the mammalian S1P
lyase, SPL, is also involved in the differentiation of F9 cells.
Disruption of the SPL gene resulted in accelerated differenti-
ation to PrE cells (Fig. 3). Expression of the PrE-specific
marker Dab2 was observed earlier and was enhanced in the
SPL⫺/⫺ cells (Fig. 3C). This situation is similar to that observed
in sglA mutant cells, which proceed through early development
slightly faster than wild-type cells (10). Moreover, expression of
contact site A, which is expressed at the onset of development,
occurred slightly earlier and was extended longer in the mu-
tant (10). Thus, it seems that the function of S1P lyase in early
development is conserved.
There are two plausible accounts for the effect of SPL knock-
out on promoting differentiation. Since S1P lyase catalyzes the
conversion of S1P (dihydro-S1P) to fatty aldehyde and phos-
phoethanolamine, both of which are used as glycerolipid pre-
cursors, inactivation of S1P lyase leads to both the accumula-
tion of S1P and the loss of its products (i.e. the shutdown of the
cross-talk between sphingolipids and glycerolipids). To exam-
ine which of these is the cause of the accelerated differentia-
tion, we utilized an inhibitor of Sph kinase (DMS), which di-
minishes both intracellular S1P and its degradation products.
Treatment of cells with low concentrations of DMS efficiently
inhibited the PrE differentiation (Fig. 5, A and B). Moreover,
overexpression of Sph kinase also resulted in an acceleration of
the differentiation (Fig. 5C). These results support the conclu-
sion that accumulation of S1P rather than shutdown of the
sphingolipid-to-glycerolipid pathway is responsible for the ef-
fect of SPL disruption.
S1P can act extracellularly by binding to members of the
Edg/S1P family of G protein-coupled receptors. Our RT-PCR
results showed that F9 EC cells express Edg1/S1P1, Edg5/S1P2,
and Edg6/S1P4 but not Edg3/S1P3.2 A previous study demon-
strated that Edg5/S1P2 is down-regulated during RA- and
bt2cAMP-induced differentiation (35). The authors proposed a
role for Edg5/S1P2 in maintaining the stem cell phenotype of
undifferentiated F9 cells. Our results presented here revealed
that exogenously added S1P or dihydro-S1P does not induce the
differentiation but rather slightly inhibits it. Therefore, we
suppose that S1P signaling mediated by Edg receptors may be
inhibitory to the differentiation. We also demonstrated that
exogenously added Sph did not accelerate the differentiation
either (data not shown). As shown in Figs. 2F and 6B, conver-
sion of exogenous Sph to S1P is only ⬍1%. Most Sph was
converted to other sphingolipids. Taking into account the fact
that other sphingolipids such as ceramide and Sph are also
signaling molecules, the effects of exogenous Sph cannot be
2
A. Kihara, C. Ogawa, and Y. Igarashi, unpublished results.
 Involvement of SPL in Differentiation
attributed only to the effects of S1P. Our results using 3H-
labeled S1P demonstrated that exogenously added S1P did not 8.
result in an increase in intracellular amounts of S1P (Fig. 6B).
9.
The pattern of metabolism of S1P was similar to that of Sph, 10.
and the signals of metabolites of S1P were weaker than those of
Sph (Fig. 6B). Therefore, we suppose that F9 can import S1P 11.
only through conversion to Sph by the phosphatase on the cell 12.
surface.
13.
There is abundant evidence that S1P can also function in-
tracellularly in calcium homeostasis, inhibition of apoptosis, 14.
and cell growth (36 –38). The level of S1P is very low in a
15.
steady-state system and is rapidly increased by the activation 16.
of Sph kinase induced by numerous external stimuli including
platelet-derived growth factor (39), tumor necrosis factor ␣ 17.
18.
(40), nerve growth factor (41), activation of protein kinase C
(42), and cross-linking of the immunoglobulin receptors Fc⑀RI 19.
20.
(43) and Fc␥R1 (44). Thus, S1P appears to function as a second 21.
messenger. However, molecular mechanisms of intracellular
S1P actions are largely unknown. Therefore, it is also unclear 22.
how the accumulated S1P causes the accelerated differentia- 23.
tion of F9 cells. Previous studies demonstrated that RA-in-
duced differentiation of F9 cells to PrE cells causes activation of 24.
Ras and a decline in Gi␣2 levels, both of which lead to activation 25.
of ERK (21, 45). Moreover, transfection of the constitutively
26.
active form of mitogen-activated protein kinase kinase, MEK,
induced PrE differentiation of F9 cells without RA treatment 27.
(21). Although an ERK-independent pathway in RA-induced
28.
PrE differentiation has been reported (21), antisense disrup-
tion of p42 ERK abolished PrE differentiation (45). Thus, ERK 29.
seems to play an important role in RA-induced progression of
F9 EC cells to PrE. We investigated the activation of ERK in 30.
wild-type and SPL⫺/⫺ cells treated with RA and found no 31.
difference between them (data not shown), indicating that ERK 32.
is not involved in the promotional effect of S1P on differentia-
tion. Therefore, further work will be needed to determine the 33.
precise molecular mechanism of the role of S1P in this differ-
entiation as well as any intracellular S1P signaling pathway. 34.
The SPL⫺/⫺ cells constructed here can be used as a new tool in 35.
investigating the role of intracellular S1P. 36.
37.
Acknowledgments—We thank N. Mizushima for the pCE-neo plas-
mid and useful advice, S. Sugano for the pCE-FL plasmid, A. Wada (this
38.
laboratory) for pcDNA3-HA1 and pcDNA3-HA-SPHK1a plasmids, and
Chie Ogawa (this laboratory) for technical support. 39.
40.
REFERENCES
1. Igarashi, Y. (1997) J. Biochem. (Tokyo) 122, 1080 –1087
2. Pyne, S., and Pyne, N. J. (2000) Biochem. J. 349, 385– 402 41.
3. Spiegel, S., and Milstien, S. (2000) FEBS Lett. 476, 55–57 42.
4. Spiegel, S., and Milstien, S. (2000) Biochim. Biophys. Acta 1484, 107–116
5. Hla, T. (2001) Prostaglandins Other Lipid Mediators 64, 135–142 43.
6. Yatomi, Y., Igarashi, Y., Yang, L., Hisano, N., Qi, R., Asazuma, N., Satoh, K., 44.
Ozaki, Y., and Kume, S. (1997) J. Biochem. (Tokyo) 121, 969 –973
7. Van Brocklyn, J. R., Lee, M. J., Menzeleev, R., Olivera, A., Edsall, L., Cuvillier, 45.
14585
O., Thomas, D. M., Coopman, P. J. P., Thangada, S., Liu, C. H., Hla, T., and
Spiegel, S. (1998) J. Cell Biol. 142, 229 –240
Saba, J. D., Nara, F., Bielawska, A., Garrett, S., and Hannun, Y. A. (1997)
J. Biol. Chem. 272, 26087–26090
Zhou, J., and Saba, J. D. (1998) Biochem. Biophys. Res. Commun. 242, 502–507
Li, G., Foote, C., Alexander, S., and Alexander, H. (2001) Development 128,
3473–3483
Van Veldhoven, P. P., Gijsbers, S., Mannaerts, G. P., Vermeesch, J. R., and
Brys, V. (2000) Biochim. Biophys. Acta 1487, 128 –134
Skrzypek, M. S., Nagiec, M. M., Lester, R. L., and Dickson, R. C. (1999) J.
Bacteriol. 181, 1134 –1140
Gottlieb, D., Heideman, W., and Saba, J. D. (1999) Mol. Cell. Biol. Res.
Commun. 1, 66 –71
Li, G., Alexander, H., Schneider, N., and Alexander, S. (2000) Microbiology
146, 2219 –2227
Strickland, S., and Mahdavi, V. (1978) Cell 15, 393– 403
Rickles, R. J., Darrow, A. L., and Strickland, S. (1988) J. Biol. Chem. 263,
1563–1569
Bjersing, J. L., Brorsson, A., and Heby, O. (1997) J. Cell. Biochem. 67, 378 –385
Duprey, P., Morello, D., Vasseur, M., Babinet, C., Condamine, H., Brûlet, P.,
and Jacob, F. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 8535– 8539
Cho, S. Y., Cho, S. Y., Lee, S. H., and Park, S. S. (1999) Mol. Cells 9, 179 –184
Downloaded from www.jbc.org at University of Connecticut Health Center Library on April 13, 2008
Strickland, S., Smith, K. K., and Marotti, K. R. (1980) Cell 21, 347–355
Verheijen, M. H., Wolthuis, R. M., Bos, J. L., and Defize, L. H. (1999) J. Biol.
Chem. 274, 1487–1494
Jackson, R. J., Howell, M. T., and Kaminski, A. (1990) Trends Biochem. Sci. 15,
477– 483
Murate, T., Banno, Y., T-Koizumi, K., Watanabe, K., Mori, N., Wada, A.,
Igarashi, Y., Takagi, A., Kojima, T., Asano, H., Akao, Y., Yoshida, S., Saito,
H., and Nozawa, Y. (2001) J. Histochem. Cytochem. 49, 845– 855
Van Veldhoven, P. P. (2000) Methods Enzymol. 311, 244 –254
Olivera, A., Kohama, T., Tu, Z., Milstien, S., and Spiegel, S. (1998) J. Biol.
Chem. 273, 12576 –12583
Min, J. K., Yoo, H. S., Lee, E. Y., Lee, W. J., and Lee, Y. M. (2002) Anal.
Biochem. 303, 167–175
Mok, S. C., Wong, K. K., Chan, R. K., Lau, C. C., Tsao, S. W., Knapp, R. C., and
Berkowitz, R. S. (1994) Gynecol. Oncol. 52, 247–252
Tseng, C. P., Ely, B. D., Li, Y., Pong, R. C., and Hsieh, J. T. (1998) Endocri-
nology 139, 3542–3553
Smith, E. R., Capo-chichi, C. D., He, J., Smedberg, J. L., Yang, D. H., Prowse,
A. H., Godwin, A. K., Hamilton, T. C., and Xu, X. X. (2001) J. Biol. Chem.
276, 47303– 47310
Xu, X. X., Yi, T., Tang, B., and Lambeth, J. D. (1998) Oncogene 16, 1561–1569
Cho, S. Y., Jeon, J. W., Lee, S. H., and Park, S. S. (2000) Biochem. J. 352,
645– 650
Igarashi, Y., Hakomori, S., Toyokuni, T., Dean, B., Fujita, S., Sugimoto, M.,
Ogawa, T., El-Ghendy, K., and Racker, E. (1989) Biochemistry 28,
6796 – 6800
Yatomi, Y., Ruan, F., Megidish, T., Toyokuni, T., Hakomori, S., and Igarashi,
Y. (1996) Biochemistry 35, 626 – 633
Olivera, A., Kohama, T., Edsall, L., Nava, V., Cuvillier, O., Poulton, S., and
Spiegel, S. (1999) J. Cell Biol. 147, 545–558
Li, Y., MacLennan, A. J., and Rogers, M. B. (1998) Exp. Cell Res. 239, 320 –325
Mattie, M., Brooker, G., and Spiegel, S. (1994) J. Biol. Chem. 269, 3181–3188
Meyer zu Heringdorf, D., Lass, H., Alemany, R., Laser, K. T., Neumann, E.,
Zhang, C., Schmidt, M., Rauen, U., Jakobs, K. H., and van Koppen, C. J.
(1998) EMBO J. 17, 2830 –2837
Cuvillier, O., Pirianov, G., Kleuser, B., Vanek, P. G., Coso, O. A., Gutkind, S.,
and Spiegel, S. (1996) Nature 381, 800 – 803
Olivera, A., and Spiegel, S. (1993) Nature 365, 557–560
Xia, P., Gamble, J. R., Rye, K. A., Wang, L., Hii, C. S., Cockerill, P., Khew-
Goodall, Y., Bert, A. G., Barter, P. J., and Vadas, M. A. (1998) Proc. Natl.
Acad. Sci. U. S. A. 95, 14196 –14201
Edsall, L. C., Pirianov, G. G., and Spiegel, S. (1997) J. Neurosci. 17, 6952– 6960
Mazurek, N., Megidish, T., Hakomori, S., and Igarashi, Y. (1994) Biochem.
Biophys. Res. Commun. 198, 1–9
Choi, O. H., Kim, J. H., and Kinet, J. P. (1996) Nature 380, 634 – 636
Melendez, A., Floto, R. A., Gillooly, D. J., Harnett, M. M., and Allen, J. M.
(1998) J. Biol. Chem. 273, 9393–9402
Gao, P., and Malbon, C. C. (1996) J. Biol. Chem. 271, 30692–30698