Received: 17 May 2018 | Accepted: 12 September 2018 DOI: 10.1002/jcb.27818

RESEARCH ARTICLE

Translocation and activation of 1 by ‐1‐phosphate

Shohei Nishino1 | Hisahiro Yamashita1 | Mizuki Tamori1 | Masato Mashimo2 | Kazuyuki Yamagata1,3 | Hiroyuki Nakamura1 | Toshihiko Murayama1

1Laboratory of Chemical Pharmacology, Graduate School of Pharmaceutical Abstract Sciences, Chiba University, Chiba, Japan Sphingosine (SphKs) and (CerK) phosphorylate sphingosine 2Laboratory of Pharmacology, Faculty of to sphingosine‐1‐phosphate (S1P) and ceramide to ceramide‐1‐phosphate (C1P), Pharmaceutical Sciences, Doshisha respectively. S1P and C1P are bioactive that regulate cell fate/function and Women’s College of Liberal Arts, Kyoto, Japan human health/diseases. The translocation and activity of SphK1 are regulated by its 225 3Laboratory of International Scholars in of Ser and by anionic lipids such as phosphatidic acid and Pharmaceuticals in Systems Biology, phosphatidylserine. However, the roles of another anionic C1P on SphK1 Graduate School of Pharmaceutical Sciences, Chiba University, Chiba, Japan functions have not yet been elucidated, thus, we here investigated the regulation of SphK1 by CerK/C1P. C1P concentration dependently bound with and activated Correspondence recombinant human SphK1. The inhibition of CerK reduced the phorbol Hiroyuki Nakamura, PhD, Laboratory of ‐ ‐ ‐ Chemical Pharmacology, Graduate School 12 myristate 13 acetate induced translocation of SphK1 to the plasma membrane of Pharmaceutical Sciences, Chiba (PM) and activation of the in membrane fractions of cells. A treatment with ‐ ‐ ‐ University, Inohana 1 8 1, Chuo ku, C1P translocated wild‐type SphK1, but not the SphK1‐S225A mutant, to the PM Chiba 260‐8675, Japan. Email: [email protected] without affecting phosphorylation signaling. A cationic RxRH sequence is proposed

to be a C1P‐binding motif in α‐type cytosolic and tumor necrosis factor α‐converting enzyme. The of four cationic amino acids to Ala in the 56‐RRNHAR‐61 domain in SphK1 reduced the phorbol 12‐myristate 13‐acetate‐ and C1P‐induced translocation of SphK1 to the PM, however, the capacity of C1P to bind with and activate SphK1 was not affected by this mutation. In conclusion, C1P modulates SphK1 functions by interacting with multiple sites in SphK1.

KEYWORDS cationic amino acids, ceramide kinase, ceramide‐1‐phosphate, negatively charged lipids, 1

1 | INTRODUCTION directly metabolized by the following : ceramide kinase (CerK) producing ceramide‐1‐phosphate (C1P), cer- Ceramide and its metabolites play crucial roles in physiolo- amidase producing sphingosine, and then sphingosine‐ gical and pathophysiological responses.1,2 Ceramide is 1‐phosphate (S1P) via /2 (SphK1/2),

Abbreviations: C1P, ceramide‐1‐phosphate; CerK, ceramide kinase; cPLA2α, α‐type cytosolic phospholipase A2; ERK1/2, extracellular signal‐ regulated kinase 1 and 2; GFP, green fluorescent protein; GST, glutathione S‐; HA, hemagglutinin; HEK, human embryonic kidney; NBD, 4‐nitrobezo‐2‐oxa‐1,3‐diazole; NVP‐231, (N‐[2‐(benzoylamino)‐6‐benzothiazolyl]‐ricyclo[3.3.1.13,7] decane‐1‐carboxamide; PA, phosphatidic acid; PIP2, PI 4,5‐bisphosphate; PI, phosphatidylinositol; PM, plasma membrane; PMA, phorbol 12‐myristate 13‐acetate; PS, phosphatidylserine; S1P, sphingosine‐1‐phosphate; SphK1, sphingosine kinase 1; TNF‐α, tumor necrosis factor α.

5396 | © 2018 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/jcb J Cell Biochem. 2019;120:5396-5408. NISHINO ET AL. | 5397 synthase producing sphingomyelin, and (an inhibitor of CerK) and N‐[7‐(4‐nitrobezo‐2‐oxa‐1, glucosylceramide synthase producing glucosylceramide. Re- 3‐diazole)] (NBD)‐labeled‐6‐aminocaproyl‐D‐sphingosine cent studies revealed that C1P and S1P are bioactive lipids (NBD‐ceramide) were from Cayman (Ann Arbor, MI). PS, that function extracellularly and intracellularly.1-3 S1P is PA, PI, phosphatidylcholine, phosphatidylethanolamine, generated by SphK1 and SphK2. Mice that are null for C1P (C16‐ form), D‐erythro‐sphingosine, and SphK1 or SphK2 developed normally, whereas double‐ D‐erythro‐sphingosine‐NBD (NBD‐sphingosine) were from knock‐out mice showed disturbed neurogenesis, impaired Avanti Polar Lipids (Alabaster, AL) or Matreya LLC angiogenesis, and embryonic lethality.4 These findings (Pleasant Gap, PA). Sphingomyelin and phorbol 12‐myristate suggest that SphK1 and SphK2 have overlapping and/or 13‐acetate (PMA) were from Sigma (St Louis, MO) and complementary functions. However, SphK1 is generally Merck (Darmstadt, Germany), respectively. Other chemicals assumed to stimulate cell growth and survival, whereas are from Sigma and Wako (Tokyo, Japan). SphK2 suppresses growth and enhances .5,6 Although SphK2 predominantly localizes and functions in 2.2 Plasmids the nucleus and/or endoplasmic reticulum in cells, its | regulation appears to be complex.6,7 SphK1 is mainly The pMXs‐U6‐Puro retroviral vector and pCMV‐VSV‐G localized in the and translocated to the plasma envelope vector were gifts from Dr. Kitamura (University membrane (PM) depending on stimuli such as receptor of Tokyo, Tokyo, Japan). The knockdown of CerK was agonists and growth factors. When measured in cell performed with short hairpin RNA (shRNA) to silence homogenates, an agonist stimulation resulted in an increase CerK (shCerK; target sequence GGACAAGGCAAGCG in activity of only 2‐ to 3‐fold, whereas maximal activity was GATAT). A shRNA recognizing luciferase (shLuc; a14‐fold increase.8 The localization of SphK1 to the PM is target sequence GGCTATGAAGAGATACGCC) was used beingrecognizedasakeyinitssignalingcapacity;the as a control. The nucleotides for shRNA were annealed phosphorylation of SphK1 markedly affects the structural and subcloned into the BamHI/EcoRI sites of the conformation of SphK1, and also regulates its enzymatic pMXs‐U6‐Puro retroviral vector. To construct hemagglu- activity by inducing its intracellular translocation.6,8-10 tinin (HA)‐tagged CerK (CerK‐HA), complementary DNA Furthermore, various proteins, such as and (cDNA) encoding human CerK (a gift from Dr. Kohama, /integrin‐binding protein 1, have been reported to Daiichi‐Sankyo, Co., Ltd., Tokyo, Japan) was amplified by regulate the translocation of SphK1 to the PM.9,11-13 chain reaction (PCR) using the forward primer Lipid‐protein interactions are important for the regulation 5′‐TAAAAGATCTCGGAGATGGGGGCGA‐3′ and reverse of protein localization and activity and may modulate primer 5′‐TATAGGAATTCCGCTGTGTGAGTCTGGC‐3′, cellular responses including and gene regula- and this amplification was cloned into the tion. Anionic lipids, such as phosphatidylserine (PS), pHA‐N1 vector at BglII/EcoRI sites. The pHA‐N1 vector phosphatidic acid (PA), and phosphatidylinositols (PIs) was created by replacing EGFP with HA in the pEGFP‐N1 including PIP2 (PI 4,5‐bisphosphate), play critical roles in vector. To construct green fluorescent protein (GFP)‐ these interactions.14,15 Anionic lipids, such as PS, PA, and PI tagged wild‐type (WT) SphK1 (SphK1‐GFP), cDNA 4‐phosphate, were shown to bind with and activate SphK1 in encoding human SphK1 was amplified by PCR using the vitro.10,16-19 C1P, an anionic lipid, was found to bind directly forward primer 5′‐TAAAGATCTCGAGACCATGGATC with many enzymes and proteins, such as α‐type cytosolic CAGCGGGC‐3′ and reverse primer 5′‐TATCTGCAG 20,21 phospholipase A2 (cPLA2α,groupIVAPLA2), tumor TAAGGGCTCTTCTGG‐3′, and this amplification product necrosis factor α (TNF‐α)‐converting enzyme,22 and annexin was cloned into the pEGFP‐N1 vector at BglII/PstI sites. To a2‐p11 proteins regulating cell invasion.23 Similar to PS and construct glutathione S‐transferase (GST)‐tagged PA, C1P directly bound with the lipid kinase, diacylglycerol WT‐SphK1 (GST‐SphK1), cDNA encoding human SphK1 kinase γ.24 However, the roles of C1P and CerK on SphK1 was amplified by PCR using the forward primer functions have not yet been elucidated. In the current study, 5′‐TATAGAATTCATGGATCCAGCGGGCGGCCC‐3′ and we examined the roles of CerK/C1P in SphK activity and the reverse primer 5′‐TATAGTCGACTAAGGGCTCTTCTG translocation of SphK1 to the PM in cells. G‐3′, and this amplification product was cloned into the pGEX‐6P‐1 vector at EcoRI/SalI sites. The Ser → Ala mutant of SphK1 (SphK1‐S225A‐GFP) was created by 2 MATERIALS AND METHODS | PCR using SphK1‐GFP as a template with 5′‐AAGA CACCTGCGGCGCCCGTTGTG‐3′ (sense) and 5′‐CA 2.1 Materials | CAACGGGCGCCGCAGGTGTCTT‐3′ (antisense). The

PIP2 (1,2‐dipalmitoyl form), NVP‐231 (N‐[2‐(benzoylamino)‐ 56‐RRNHAR‐61 → 56‐AANAAA‐61 mutants of SphK1 6‐benzothiazolyl]‐ricyclo[3.3.1.13,7] decane‐1‐carboxamide (SphK1‐4A‐GFP and GST‐SphK1‐4A) were created by 5398 | NISHINO ET AL.

PCR using SphK1‐GFP and GST‐SphK1 as templates, Confocal images were obtained using a Fluoview FV500 respectively, with 5′‐CTTCTTTATCCACGCCGCGGC confocal laser scanning microscope with a ×40 1.00 NA GAAGAACCAGCGTG‐3′ (sense) and 5′‐CACGCTGGTTC objective (Olympus, Tokyo, Japan). The translocation of TTCGCCGCGGCGTGGATAAAGAAG‐3′ (antisense). SphK1‐GFP to the PM was evaluated by the relative increase in fluorescence intensity at the cell periphery, as described previously25 with modifications. After the subtraction of 2.3 Cells and transfection | background fluorescence, fluorescence values in two areas Human embryonic kidney (HEK) 293T and Plat‐GP having the same dimension, one area containing the cell packaging cell lines (a gift from Dr. Kitamura) were periphery as the PM and the other being the neighboring cultured in Dulbecco modified Eagle medium contain- cytosol, were measured. The cell periphery showing greater ing 10% fetal bovine serum. Plat‐GP cells were trans- (bright) values than the cytoplasmic area was assumed to fected with retroviral vectors and the pCMV‐VSV‐G reflect the translocation of SphK1‐GFP to the PM. Cells in envelope vector using a Lipofectamine reagent accord- which a quarter and more of the whole cell periphery had ing to the manufacturer’s instructions (Invitrogen, greater fluorescence (brighter) value were defined as SphK1‐ Carlsbad, CA). Retrovirus‐containing supernatants were translocated cells, and the number of SphK1‐translocated collected 48 hours after transfection. To establish a cells was counted. The clear staining of the PM, mainly at stable CerK‐knockdown cell line (shCerK) and control cell‐cell contacts, was observed in HEK293 cells under cell line (shLuc), HEK293T cells were cultured in resting conditions, as previously reported.13 Thus, cell‐cell retrovirus‐containing supernatants with 8 μg/mL poly- contact areas were omitted from measurements. Under our brene for 8 hours and selected in 1 μg/mL puromycin. assay conditions, 30% to 50% of transfected cells were SphK1‐ The knockdown of CerK was confirmed by measuring translocated cells without the PMA treatment possibly the formation of NBD‐C1P in NBD‐ceramide–treated because of the cultivation with serum. At least 60 transfected cells. The formation of NBD‐C1P in shCerK cells for cells (20 cells per view and three views per experiment) were 30 minutes was less than 40% that in control shLuc cells. counted, and three or four independent experiments were In the transient transfection of SphK‐GFP or CerK‐HA, performed. cells were grown on 25‐mm glass coverslips and transfected with 1 μg of plasmid DNA using a Lipofec- 2.6 Lipid‐protein overlay assay tamine reagent. | The assay was performed as described previously26 with minor modifications. Briefly, tested lipids were dissolved in 2.4 Purification of the recombinant | chloroform and spotted onto a Hybond C membrane GST‐SphK1 protein (Amersham). After drying, the washed and blocked mem- GST‐tagged SphK1 were expressed in the Escherichia coli brane was exposed at 4°C overnight to recombinant GST‐ BL21 (DE3) strain upon incubation with 0.1 mM isopropyl SphK1 proteins or preparations from HEK293T cells 1‐thio‐β‐D‐galactoside and 4% ethanol at 30°C for 4 hours. expressing SphK1‐GFP. The washed membranes were After centrifugation (6000 rpm, 15 minutes), the pellet was treated with an anti‐GST antibody (27457701V; GE Health resuspended in phosphate‐buffered saline containing 10% Care, Buckinghamshire, UK) or anti‐GFP antibody (4B10; glycerol, 1 mM dithiothreitol, 1% Triton X‐100, and 0.5 mM Cell Signaling, Danvers, MA), and immune‐reactive spots phenylmethylsulfonyl fluoride (PMSF). The Triton X‐100‐ were visualized using secondary antibodies. Purified recom- insoluble fraction was lysed with Empigen lysis buffer binant GST and preparations from cells expressing GFP did (50 mM 4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid not bind to any of the lipids. [HEPES], pH 7.4, 10% glycerol, 3% Empigen BB, 4 mM ethylenediaminetetraacetic acid [EDTA], 100 mM sodium 2.7 Assays for SphK and CerK fluoride [NaF], 50 μg/mL aprotinin, 100 μM leupeptin, and | activities 0.5 mM PMSF). GST‐SphK were collected on glutathione sepharose beads from Empigen lysates and eluted with SphK activity in recombinant SphK1 and in the preparations elution buffer. from HEK293T cells expressing SphK1‐HA was measured using NBD‐sphingosineasasubstrateaspreviously described27 with minor modifications. In the measurement 2.5 Immunofluorescence | of recombinant SphK1 activity, a reaction solution (100 μlof HEK293T cells were cultured on 25‐mm glass coverslips and the buffer A solution; 20 mM Tris‐HCl, pH 7.4, 150 mM KCl, transfected as described above. Cells were fixed in 4% 2 mM [ATP], and 1 mM MgCl2) paraformaldehyde at room temperature for 20 minutes. containing 0.02% Triton X‐100 and 60 μM lipids (20 μM NISHINO ET AL. | 5399

NBD‐sphingosine, 0‐20 μM C1P, and 20‐40 μM phosphati- antibodies (Cell Signaling). Immunoreactive bands were dylcholine) was prepared, and the lipids were vortexed visualized using a chemiluminescent reagent, the clarify vigorously for 2 minutes and then sonicated for 10 minutes in Western ECL (Bio‐Rad, Hercules, CA). Results a water bath. The enzyme solution (10 μl of the buffer A were analyzed using the image analyzer ChemiDoc XRS containing 200‐500 ng of the recombinant GST‐SphK1 Plus (Bio‐Rad). The intensity of chemiluminescence was protein) was added to a reaction solution, and incubated at measured using the ImageJ Software (NIH). 37°C for 30 minutes. The final concentrations in the assay mixture (200 μl) were 10 μMNBD‐sphingosine and 0.01% 2.9 Statistical analysis Triton X‐100. NBD‐sphingosine metabolites including NBD‐ | S1P were extracted by the addition of chloroform/methanol, Quantitative data for cells expressing SphK1‐GFP or andwereanalyzedonaTLCSilicaGel‐60 plate (#105724; mutants on the PM are means ± SD, and data are from Merck, Darmstadt, Germany). In the measurement of three or four independent experiments. Data for the cellular SphK activity, HEK293T cells expressing SphK1‐ phosphorylation of ERK1/2 are expressed as fold changes HA were homogenized with buffer (20 mM Tris‐HCl, pH 7.4, from the control, and are the means ± SD of three

1mM EDTA, 15mM NaF, 5mM Na3VO4,andprotease independent experiments. inhibitors). After centrifugation (1000g,4°C,10minutes),the supernatant was separated into cytosolic and membrane 3 RESULTS fractions by a second centrifugation (20 000g,4°C,30min- | utes). A reaction solution (B: 180 μl, 50 mM HEPES, pH 7.4, 3.1 Direct binding of C1P to 15 mM MgCl ,0.5%TritonX‐100, 1 mM ATP, 10 mM NaF, | 2 recombinant SphK1 and resulting and 1 μMNBD‐sphingosine) was prepared. Equal amounts activation of the enzyme in vitro of proteins in the cellular fractions (20 μl) were added to solution B, and incubated at 37°C for 30 minutes. Under our C1P is classified as a negatively charged and/or anionic conditions, the formation of NBD‐S1P was linear for at least lipid. PA and PS, anionic lipids, have been shown to bind 60 minutes, and the amounts of NBD‐S1P that formed were directly with human SphK1.10,17,18 Thus, we examined less than 0.5% of NBD‐sphingosine in the assay mixture. whether C1P directly binds with and activates SphK1 CerK activity in intact cells was examined using NBD‐ using recombinant human SphK1. In the lipid‐protein ceramide as a substrate as described previously26,28 with overlay assay, GST‐SphK1 bound with C1P in a dose‐ minor modifications. Briefly, cells were incubated with dependent manner (Figure 1A), although binding to C1P 10 μMNBD‐ceramide at 37°C for 30 minutes in modified was less than that to other anionic lipids, such as PA and Hank balanced salt solution buffer (20 mM HEPES, pH 7.4, PS. An incubation of GST‐SphK1 with the substrate NBD‐

137mM NaCl, 5.4mM KCl, 0.8mM MgSO4,0.34mM sphingosine formed NBD‐S1P, and its formation was Na2HPO4, 5.6 mM glucose, 0.44 mM KH2PO4,4.2mM markedly enhanced by the coaddition of C1P to an assay NaHCO3,1.9mMCaCl2, and 0.1% albumin). Ceramide mixture (Figure 1B). When HEK293T cell lysates metabolites including NBD‐C1P were analyzed on a TLC containing SphK1‐GFP were used, the similar binding Silica Gel‐60 plate. For quantitative analyses of NBD‐S1P and of SphK1‐GFP to C1P was confirmed (Figure 1C). In the NBD‐C1P, various amounts (0‐20 pmol) of standard NBD‐ assay using cell lysates, the binding of SphK1‐GFP to PS sphingosine and NBD‐ceramide were spotted in the upper and PIP2 was markedly less than that of recombinant area of the plate after separation by TLC. The fluorescence GST‐SphK1. intensity of NBD‐sphingosine and NBD‐ceramide was linear to 20 pmol the reagent. Fluorescence was detected with 3.2 Effect of CerK inhibition on LAS1000‐Plus (Fujifilm, Tokyo, Japan; 470‐nm excitation | PMA‐induced translocation of SphK1 and 515‐nm emission). to the PM in HEK293T cells First, we examined the translocation of endogenous 2.8 Western blot analysis | SphK1 in membrane preparations in native HEK293T The following antibodies (1:1000‐1:3000) were used: an cells before and after the PMA treatment by Western anti‐HA‐antibody (3F10; Roche, Mannheim, Germany); blot analysis. Although we examined the effects of anti–phospho‐Thr202/Thr204‐extracellular‐signal‐regulated several commercially available anti‐SphK1 antibodies, kinase (ERK) 1/2 antibody (#4370; Cell Signaling); anti‐ we could not determine the levels of SphK1 probably ERK1 and anti‐ERK2 antibodies (C16 and C14, respec- because of the low levels of SphK1 and/or many tively; Santa Cruz Biotechnology, Dallas, TX); and nonspecific bands. Next, we examined the transloca- anti‐rabbit and anti‐mouse IgG horseradish peroxidase tion of SphK1‐GFP by the immunofluorescence 5400 | NISHINO ET AL. analysis. The treatment of HEK293T cells expressing significantly weaker than that in control shLuc cells SphK1‐GFP with 100 nM PMA, an activator of (Figure 2B). In cells treated with NVP‐231 and in phosphorylation signaling mediated by shCerK cells, the formation of NBD‐C1P, which C, for 10 minutes increased the percentages of cells reflected the activity of CerK, was markedly inhibited. expressing SphK1‐GFP on the PM. In cells cultured with 200 nM NVP‐231, a selective inhibitor of CerK, 3.3 Effects of CerK inhibition and for 16 hours, the PMA‐induced translocation of | PMA on SphK activity in membrane SphK1‐GFP to the PM was significantly decreased preparations from cells expressing (Figure 2A). In shCerK cells, the PMA response was SphK1‐HA We examined the formation of NBD‐S1P from NBD‐ (A) sphingosine with and without PMA and C1P. PMA and C1P did not markedly affect the formation of NBD‐S1P in intact cells or total homogenate fractions possibly because endogenous SphK1/2, which are located in various cellular compartments, may mask their effects (data not shown). SphK activities in the membrane and cytosolic preparations were examined separately. The two types of cells, shCerK, and shLuc cells, were transiently transfected with the SphK1‐HA vector, and then further treated with PMA for 15 minutes. After cell homogeniza- tion, SphK activities in the membrane and cytosolic preparations were measured (Figures 2C and 2D). The formation of NBD‐S1P in membrane preparations from PMA‐treated shLuc cells was significantly greater than in those from PMA‐free cells. In membrane preparations (B) from shCerK cells, the formation of NBD‐S1P was not enhanced by the PMA treatment. In cytosolic prepara- tions from shLuc cells and shCerK cells, the PMA treatment did not affect the formation of NBD‐S1P. The

FIGURE 1 Ceramide‐1‐phosphate (C1P) directly binds with and activates recombinant sphingosine kinase 1 (SphK1) in vitro. A, The indicated , , and solvent (no lipid) were spotted onto a Hybond C membrane, and the lipid‐protein overlay assay was performed using recombinant

human GST‐SphK1. A total of 10 nmol of the lipids, except for PIP2 (1 nmol), was used in the upper panel, and the indicated amounts of C1P were used in the lower panel. B, Recombinant GST‐SphK1 (200 ng protein/tube) was incubated with 10 μM NBD‐sphingosine for 30 minutes in the presence of the indicated concentrations of C1P, and the formation of NBD‐S1P was examined in vitro. C, Cytosolic preparations from HEK293T cells expressing SphK1‐GFP were used for the lipid‐protein overlay assay. A total of 10 nmol of

(C) the lipids, except for PS (100 nmol) and PIP2 (5 nmol), were used. Typical images are representative of three independent experiments, and data in the lower panel in (B) are means ± SD. Cer, ceramide; GFP, green fluorescent protein; GST, glutathione S‐transferase; NBD, 4‐nitrobezo‐2‐oxa‐1,3‐diazole; PA, phosphatidic acid; PC, phosphatidylcholine; PE,

phosphatidylethanolamine; PI, phosphatidylinositol; PIP2,PI4, 5‐bisphosphate; Sph, sphingosine; SM, sphingomyelin NISHINO ET AL. | 5401

(A) (B)

(C) (D)

FIGURE 2 Effects of ceramide kinase (CerK) inhibition on the phorbol 12‐myristate 13‐acetate (PMA)‐induced translocation of sphingosine kinase 1 (SphK1)‐green fluorescent protein (GFP) to the PM and on the increase in SphK activity in membrane preparations. A and B, HEK293T cells transiently transfected with the SphK1‐GFP vector were treated with 100 nM PMA for 10 minutes, and the translocation of SphK1‐GFP to the PM was examined using confocal microscopy. A, Cells were pretreated with vehicle and 100 nM NVP‐231, an inhibitor of CerK, for 16 hours before the PMA treatment. B, HEK293T‐shCerK (shCerK) cells and HEK293T‐shLuciferase (shLuc) cells as the control were transiently transfected with the SphK1‐GFP vector. The inhibition of 4‐nitrobezo‐2‐oxa‐1,3‐diazole (NBD)‐C1P formation by NVP‐231 (A) and shCerK (B) was shown in the insets. In (A) and (B), typical images and quantitative data were shown in the upper and lower panels, respectively. C and D HEK293T‐shCerK cells and HEK293T‐shLuciferase cells were transiently transfected with the SphK1‐HA vector, and cells were treated with 100 nM PMA for 15 minutes. After homogenization, the membrane (C) and cytosolic (D) preparations were separated by centrifugation. SphK activity was examined in vitro with 1 μM NBD‐sphingosine (NBD‐Sph) for 30 minutes. Quantitative data are expressed as fold changes from the control in shLuc cells 5402 | NISHINO ET AL. results obtained revealed that the inhibition of CerK the translocation of SphK1 to the PM without changing reduced the PMA‐induced translocation of SphK1 to the phosphorylation signaling to ERK. PM and the resulting increase in SphK activity in membrane preparations (Figure 2). 3.6 | Effect of the SphK1‐4A mutant on the translocation of SphK1 to the plasma 3.4 | A stimulatory effect of C1P on membrane translocation of SphK1 to the plasma Previous studies showed that C1P regulated the activities membrane 29,30 22 of cPLA2α and TNF‐α‐converting enzyme via an We examined the effects of the extracellular addition of C1P interaction with the cationic RxRH sequence, the C1P‐ on the translocation of SphK1‐GFP in HEK293T cells. The binding motif, in these proteins. SphK1 has a cation‐rich treatment with 1 μM C1P alone induced the translocation of domain, 56‐RRNHAR‐61, and these residues have been SphK1‐GFP to the PM after a 1‐hr treatment (Figure 3A). implicated in targeting to PS‐rich membrane domains.9 A similar experiment was performed in shCerK cells We generated the mutant SphK1‐4A in which four (Figure3B).ThePMAresponsewasweakerinshCerKcells cationic amino acids in the domain were displaced to than in control (shLuc) cells, whereas the treatment with Ala. We initially examined the activity of recombinant C1P markedly induced the translocation of SphK1‐GFP to GST‐SphK1‐4A because the domain 54‐TERR‐57 has the PM. A combination of C1P/PMA did not exert additive been reported to contribute to the ATP‐ in effects in shLuc or shCerK cells. We then investigated the SphK1.9 The formation of NBD‐S1P by recombinant GST‐ translocation of SphK1‐GFP in cells expressing CerK‐HA SphK1‐4A was similar to that in WT‐GST‐SphK1, and the (CerK‐HA cells). The expression and activity of the CerK‐HA coaddition of 1 μM C1P increased the formation of NBD‐ protein were confirmed in CerK‐HA cells (Figure 3C). The S1P in the two enzymes to a similar degree (Figure 5A). expression of CerK‐HA slightly increased the translocation of In the lipid‐protein overlay assay, the binding of SphK1‐GFP to the PM in the absence of PMA (Figure 3D). recombinant GST‐SphK1‐4A to C1P was similar to that The additional treatment of CerK‐HA cells with 100 nM of WT‐GST‐SphK1 (Figure 5B). The similar binding of PMA for 10 minutes enhanced the translocation of SphK1‐4A‐GFP to C1P, PA, and PS was confirmed in the SphK1‐GFP to the PM. The PMA‐induced translocation of assay using HEK293T cell lysates (data not shown). In SphK1 was similar in control and CerK‐HA cells. contrast, the translocation of SphK1‐4A‐GFP to the PM induced by PMA (Figure 5C) and C1P (Figure 5D) was significantly less than that in HEK293T cells expressing 3.5 Role of the Ser225 residue of SphK1 | WT‐SphK1‐GFP. on C1P‐induced translocation of SphK1 to the plasma membrane Treatment with PMA is known to induce the phosphor- 4 | DISCUSSION ylation of the Ser225 of SphK1 through the activation of ERK1/2, resulting in the PM localization and activation In the current study, we revealed four C1P‐induced of SphK1.8,10 The effects of the displacement of Ser225 to responses in SphK1: (i) the direct binding of C1P to Ala in SphK1 on the PMA response were examined. In SphK1 in vitro, (ii) a C1P‐induced increase in the activity SphK1‐S225A‐GFP cells, the percentage of cells expres- of recombinant GST‐SphK1 in vitro, (iii) the involvement sing the protein in the PM was less than 20% in the of C1P in the PMA‐induced translocation of SphK1 to the absence of PMA, and PMA did not induce the transloca- PM in cells, and (iv) the C1P‐induced translocation of tion of the protein to the PM (Figure 4A). The SphK1 to the PM in cells. The mutation of four cationic extracellular addition of 1 μM C1P caused the transloca- residues to Ala in the domain 56‐RRNHAR‐61 in SphK1 tion of SphK1 to the PM in SphK1‐WT‐GFP cells, but not reduced the C1P‐ and PMA‐induced translocation of SphK1‐S225A‐GFP cells (Figure 4B). Thus, the existence SphK1 to the PM (responses (iii) and (iv)), however, the of the Ser225 residue appeared to be essential for the mutation did not affect responses (i) and (ii). PMA‐ and C1P‐induced translocation of SphK1 to the PM. PMA induced the phosphorylation of ERK1/2 10 and 4.1 C1P‐induced interaction with and 30 minutes after the treatment, and the response was not | activation of SphK1 in vitro modified by NVP‐231 (Figure 4C). The treatment of cells with 1 μM C1P alone did not phosphorylate ERK1/2, and We showed for the first time that C1P directly bound C1P did not change PMA‐induced phosphorylation with recombinant human GST‐SphK1 (Figure 1A) and signaling to ERK1/2 (Figure 4D). Thus, C1P stimulates SphK1‐GFP in HEK293T cell homogenates (Figure 1C), NISHINO ET AL. | 5403

(A) (C)

(D)

(B)

FIGURE 3 Effects of ceramide‐1‐phosphate (C1P) supplementation and ceramide kinase (CerK) expression on the translocation of sphingosine kinase 1 (SphK1)‐green fluorescent protein (GFP) to the plasma membrane (PM). A and B Cells expressing SphK1‐GFP were pretreated with 1 μM C1P for 1 hour, and the translocation of SphK1‐GFP was examined 10 minutes after the 100 nM phorbol 12‐myristate 13‐acetate (PMA) treatment. In (A) HEK293T cells were used. Typical images and quantitative data are shown in the upper and lower panels, respectively. In (B) HEK293T‐shCerK (shCerK) cells and HEK293T‐shLuciferase (shLuc) cells were used. Quantitative data are shown. C and D Control cells and HEK293T cells expressing CerK‐ hemagglutinin (HA) were transiently transfected with the SphK1‐GFP vector. In (C), the formation of NBD‐C1P and expression of CerK‐HA in cells transfected with the CerK‐HA vector were shown. In (D), the translocation of SphK1‐GFP was examined 10 minutes after the 100 nM PMA treatment. In (C) and (D), typical images are representative of three independent experiments. Quantitative data for the translocation of SphK1 were shown in the lower panel in (D)

29,30 and that C1P activated recombinant GSH‐SphK1 activity N‐terminal lipid‐binding C2 domain of cPLA2α. C1P in vitro (Figure 1B). To date, cPLA2α and TNF‐α‐ also inhibited recombinant TNF‐α‐converting enzyme converting enzyme are the only documented proteins in activity via C1P‐binding cationic motifs in the enzyme.22 which the amino acids and/or domain(s) responsible for Thus, we constructed SphK1‐4A, in which Arg and His C1P binding have been identified. C1P bound with residues in the 56‐RRNHAR‐61 domain in SphK1 were 54 and activated cPLA2α via an interaction with the cationic displaced to Ala. The displacement of Thr , close to the motif (a RxRH sequence, Arg59/61, and His62) in the cation‐rich domain, to Ala [Ref. 10] and that of Arg185/ 5404 | NISHINO ET AL.

(A) (B)

(C) (D)

FIGURE 4 Translocation of sphingosine kinase (SphK1)‐S225A‐green fluorescent protein (GFP) to the plasma membrane (PM) and effects of phorbol 12‐myristate 13‐acetate (PMA) and ceramide‐1‐phosphate (C1P) on phosphorylation signaling. HEK293T cells were transiently transfected with vectors for wild‐type SphK1‐GFP (SphK1‐WT‐GFP) and SphK1‐S225A‐GFP. The translocation of the two SphK1 proteins were examined 10 minutes after the 100 nM PMA treatment (A) and 1 hour after the 1 μM C1P treatment (B). Typical images and quantitative data are shown in the upper and lower panels, respectively. C and D The PMA‐induced phosphorylation of extracellular signal‐regulated kinase 1 and 2 (ERK1/2) was examined in cells pretreated with 200 nM NVP‐231 for 16 hours (C) and with 1 μM C1P for 1 hour (D). In (C) and (D), cell lysates were subjected to Western blot analyses: antibodies against phosphorylated ERK1/2 (p‐ERK1/2) and ERK1/2 were used. The phosphorylation of ERK1/2 was expressed as fold changes in values 10 minutes after the PMA treatment in the control without NVP‐231/C1P

Arg186 to Ala [Ref. 12] in SphK1 reduced enzyme activity activity (Figure 5A). Since the side chains of Arg56 and to approximately 50% and less than 25% of WT‐SphK1, Arg57 were not directly implicated in critical polar respectively, and various reduced SphK1 contacts with bound ATP in SphK1,9 the unimpaired activity.31 Contrary to expectations, the displacement of activity of SphK1‐4A may be reasonable. GST‐SphK1‐4A Arg56/Arg57/His59/Arg61 to Ala did not affect SphK1 bound with C1P to a similar extent as WT‐GST‐SphK1 in NISHINO ET AL. | 5405

(A) (B)

(D)

(C)

FIGURE 5 Phorbol 12‐myristate 13‐acetate (PMA)‐ and ceramide‐1‐phosphate (C1P)‐induced translocation of SphK1‐4A‐GFP. A, Recombinant wild‐type GST‐SphK1 (WT) and GST‐SphK1‐4A (4A) were incubated with 2 μM4‐nitrobezo‐2‐oxa‐1,3‐diazole (NBD)‐sphingosine (NBD‐Sph) for 30 minutes with and without 1 μM C1P. The amount of glutathione S‐transferase (GST)‐proteins was 500 ng protein/tube. Quantitative data for the formation of NBD‐S1P are expressed as fold changes in the value of GST‐SphK1‐WT. B, The lipid‐protein overlay assay was performed using GST‐SphK1‐WT and GST‐SphK1‐4A in the presence of 10 nmol of C1P. C and D, HEK293T cells were transfected with vectors for SphK1‐WT‐GFP and SphK1‐4A‐GFP, and the translocation of the two proteins to the plasma membrane was examined 10 minutes after the 100 nM PMA treatment (C) and 1 hour after the 1 μM C1P treatment (D). Typical images and quantitative data are shown in the upper and lower panels, respectively. GFP, green fluorescent protein the lipid‐protein binding assay (Figure 5B), and SphK1‐ be a direct binding site of C1P or the site responsible for 4A‐GFP bound with anionic lipids including C1P and PA the C1P‐induced activation of the enzyme. There are to a similar extent as WT‐SphK1, as described in Section several cationic domains in SphK1. The hydrophobic 3. Thus, the 56‐RRNHAR‐61 domain in SphK1 may not residues in lipid‐binding loop‐1 of SphK1, which were 5406 | NISHINO ET AL. previously identified as a binding locus for a calcium/ resident time of cPLA2α to the substrate lipid by decreasing integrin‐binding protein 1,11 were further implicated in a the dissociation constant of the enzyme, therefore, C1P direct, curvature‐sensitive interaction with negatively induced the intracellular translocation of cPLA2α to charged liposomes that contain PA and/or PS.32 It was membrane fractions.29,30,36 Taking these results into con- proposed that the SphK1 membrane engagement surface sideration, C1P in the PM appears to positively regulate the might comprise the surface‐exposed hydrophobic resi- translocation (and/or attachment) of SphK1 to the PM. dues on lipid‐binding loop‐1 acting in concern with The site‐directed mutagenesis revealed that many cationic residues that engage anionic lipids.9 Also, Pyne’s residues in SphK1 regulated the localization group suggested that crucial prominently exposed re- of SphK1,9,31 but the role(s) of the 56‐RRNHAR‐61 sidues in the postulated anionic lipid interaction box in domain of SphK1 have not yet been elucidated. To the SphK1 are Lys27/Lys29/Arg186,9 and then a role of this best of our knowledge, the current study is the first report positively charged site on the enzyme’s surface on the to show modifications in the translocation of SphK1 to anionic lipid interaction was experimentally validated.33 the PM by mutagenesis in the 56‐RRNHAR‐61 domain of Moreover, SphK1 has the 183‐KYRR‐186 and 308‐KGRH‐ the enzyme (Figure 5), although the domain did not 311 domains, and the SphK1 crystallographic model appear to be a strong and/or major binding site of C1P as shows that Arg207, Arg209, and Lys332 were located discussed in Section 4.1. How does C1P regulate the adjacently.34 Thus, these cationic domains may be C1P's translocation of SphK1 to the PM in the 56‐RRNHAR‐61 binding and/or activation site(s) of SphK1. We have tried domain‐sensitive manner? The dimerization of SphK1 the binding of C1P to several truncated forms of SphK1, has been hypothesized for SphK1 membrane engage- but the experiments have been unsuccessful (data not ment.9 They proposed that Arg56 and Arg57 in the 56‐ shown). The structural plasticity and/or conformation RRNHAR‐61 domain in SphK1 might potentially be mobility of SphK1 that was critical for functions of involved in a relay mechanism that links conformational SphK19,33 may interfere with the advance of analysis. changes in the regulatory “R‐loop” for making the putative dimerization interface and/or involving the regulatory C‐terminal region of SphK1, thus these 4.2 Roles of CerK/C1P and the | residues might modulate a dynamic monomer‐dimer 56‐RRNHAR‐61 domain of SphK1 on equilibrium that is relevant to the membrane residence. the translocation of SphK1 to the PM Thus, C1P may interact with the 56‐RRNHAR‐61 domain In the current study, the treatment with PMA induced the of SphK1 and modulate the translocation of SphK1 to the translocation of SphK1‐GFP to the PM in HEK293T cells. PM by changing the monomer‐dimer equilibrium. The pharmacological (Figure 2A) and molecular (Figure 2B) Furthermore, the intracellular translocation of SphK1 is inhibition of CerK reduced the PMA‐induced translocation regulated by other proteins including calcium/integrin‐ of SphK1 to the PM, and the inhibition of CerK reduced binding protein 1.6,11 The effects of the 4A mutation of PMA‐induced increases in SphK activity in membrane SphK1 on the dimerization of SphK1 and the SphK1’s preparations (Figure 2C). A supplementation with C1P interaction with other regulatory proteins should be induced the translocation of SphK1‐GFP to the PM (Figures studied in future. 3A and B). Thus, C1P appeared to have a positive effect on the translocation of SphK1 to the PM. Where does C1P 4.3 Role of Ser225 phosphorylation on contribute to the translocation of SphK1 to the PM? C1P is | C1P‐induced translocation of SphK1 to formed in cellular organelles/membranes including the Golgi the PM complex and PM,29,30 and C1P formed in the Golgi complex is transported by the C1P transfer protein to the PM and The phosphorylation of Ser225 in SphK1 appears to be other membrane fractions.35 The effect of CerK‐HA expres- essential for the PMA‐induced translocation of SphK1 to the sion on translocation of SphK1‐GFP to the PM was observed, PM; this translocation is mediated by calcium/integrin‐ but not as strong as expected (Figure 3D). Cellular binding protein 1 and by binding to anionic lipids in the compartment(s) of C1P in cells treated with the vector for PM.6,8,11 Under our conditions in which cells were cultivated CerK may exist apart from the compartment(s) of SphK1 in the presence of serum, approximately 40% of cells showed responsible for its intracellular movement, and/or the the localization of SphK1‐GFP in the PM without stimuli, intracellular transport system of C1P may not act sufficiently while the translocation of SphK1‐S225A‐GFP was less than under our conditions. Delon et al18 proposed that the 20% of control cells (Figure 4A). The translocation of SphK1‐ generation of PA by the immune receptor‐coupled activation GFP to the PM induced by not only PMA (Figure 4A), but of phospholipase D in the PM drives SphK1 to the PM. In the also C1P (Figure 4B) was decreased in SphK1‐S225A cells. A

C1P‐cPLA2α system, a treatment with C1P increased the combinatory treatment with PMA/C1P did not exert additive NISHINO ET AL. | 5407 effects. Thus, similar to cells treated with serum and PMA, the present report were obtained in cells expressing the phosphorylation of the Ser225 of SphK1 appeared to be recombinant constructs of SphK1 including SphK1‐GFP. essential for the translocation of SphK1 to the PM in C1P‐ Studies by using native SphK1 are required to prove the treated cells. Previous studies showed that the treatment of regulation of SphK1 by the CerK/C1P pathway. In addition, cells with C1P caused phosphorylation signaling, including the physiological and/or pathological roles of the CerK/C1P ERK activation,2,20 thus, the C1P treatment may modify pathway on SphK1 activity, including the factors responsible phosphorylation signaling. However, this possibility may be for the regulation of the pathway, warrant further study. excluded because C1P did not affect the phosphorylation of ERK1/2 with and without PMA (Figure 4C and 4D). We previously reported that the treatment of cells with 100 nM ORCID PMA did not change the formation of C1P.28 Thus, the Hiroyuki Nakamura http://orcid.org/0000-0001- treatment with PMA appeared to affect the translocation of 6328-9727 SphK1 independently, but also co‐operatively with C1P. The Toshihiko Murayama http://orcid.org/0000-0002- mutation of the Thr54 and Asn89,bothofwhichareneutral 1908-1408 amino acids, of SphK1 has been shown to reduce the PMA‐ induced translocation of SphK1 to the PM, and the phosphorylation of Ser225 has been suggested to expose REFERENCES Thr54 and Asn89 and/or other PS‐binding residues in SphK1, 10,37 1. Maceyka M, Spiegel S. metabolites in inflamma- thereby enhancing the PM attachment of the enzyme. tory disease. Nature. 2014;510:58‐67. 225 Similarly, the phosphorylation of the Ser of SphK1 may 2. Gomez‐Muñoz A, Presa N, Gomez‐Larrauri A, Rivera IG, affect the conformation of SphK1, thereby enhancing the Trueba M, Ordoñez M. Control of inflammatory responses by interaction between the 56‐RRNHAR‐61 domain and C1P. ceramide, sphingosine 1‐phosphate and ceramide 1‐phosphate. Prog Lipid Res. 2016;61:51‐62. 3. Fu P, Ebenezer DL, Ha AW, Suryadevara V, Harijith A, 4.4 | Problems to be solved Natarajan V. Nuclear lipid mediators: role of nuclear sphingo- lipids and sphingosine‐1‐phosphate signaling in epigenetic In PMA‐ and C1P‐treated cells, the phosphorylation of the 225 regulation of inflammation and gene expression. J Cell Ser of SphK1 appeared to be essential for the translocation Biochem. 2018;119:6337‐6353. of SphK1 to the PM. However, previous studies demon- 4. Mizugishi K, Yamashita T, Olivera A, Miller GF, Spiegel S, strated the translocation of SphK1 to the PM or PS‐ Proia RL. Essential role for sphingosine kinases in neural and containing lipid rafts in a Ser225 phosphorylation‐indepen- vascular development. Mol Cell Biol. 2005;25:11113‐11121. dent manner; the activation of the M3 muscarinic receptor13 5. Liu H, Toman RE, Goparaju SK, et al. Sphingosine kinase type ‐ and deletion of the C terminus (367‐384) of SphK1.38 In the 2 is a putative BH3 only protein that induces apoptosis. J Biol ‐ current study, cotreatments with PMA/C1P exerted signifi- Chem. 2003;278:40330 40336. 6. Wattenberg BW, Pitson SM, Raben DM. The sphingosine and cant stimulatory effects on the translocation of SphK1‐ superfamily of signaling kinases: localization ‐ S225A GFP to the PM: the value was 41.4 ± 6.8% (P <0.05, as a key to signaling function. JLipidRes. 2006;47:1128‐1139. n = 4), although the treatment with PMA and C1P alone did 7. Igarashi N, Okada T, Hayashi S, Fujita T, Jahangeer S, not cause the translocation of SphK1‐S225A‐GFP to the PM Nakamura S. Sphingosine kinase 2 is a nuclear protein and (Figure 4A and 4B). It is reported that a partial and/or large inhibits DNA synthesis. J Biol Chem. 2003;278:46832‐46839. population of SphK1 molecules expressed in cells were 8. Pitson SM, Moretti PA, Zebol JR, et al. Activation of ‐ phosphorylated at Ser225.8,10,39 Given a possible dimerization sphingosine kinase 1 by ERK1/2 mediated phosphorylation. EMBO J. 2003;22:5491‐5500. of SphK1, endogenously phosphorylated SphK1 may affect ‐ ‐ ‐ 9. Adams DR, Pyne S, Pyne NJ. Sphingosine kinases: emerging the C1P induced translocation of SphK1 GFP and SphK1 structure‐function insights. Trends Biochem Sci. 2016;41:395‐409. ‐ 225 S225A GFP in cells. The mechanism(s) underlying the Ser 10. Stahelin RV, Hwang JH, Kim JH, et al. The mechanism of phosphorylation–dependent and ‐independent translocation membrane targeting of human sphingosine kinase 1. J Biol of SphK1 to the PM need to be elucidated in more detail. The Chem. 2005;280:43030‐43038. CerK/C1P and SphK/S1P pathways co‐operatively regulate 11. Jarman KE, Moretti PAB, Zebol JR, Pitson SM. Translocation of sphingosine kinase 1 to the plasma membrane is mediated by eicosanoid synthesis through the direct activation of cPLA2α ‐ ‐ and upregulation of cyclooxygenase‐2, respectively,40,41 and calcium and integrin binding protein 1. J Biol Chem. 2010;285:483‐492. these two pathways alone or in collaboration regulate 1-3 12. Sutherland CM, Moretti PAB, Hewitt NM, Bagley CJ, Vadas pathological responses including inflammation. The pre- MA, Pitson SM. The calmodulin‐binding site of sphingosine sent results suggest the novel regulation of the SphK1/S1P kinase and its role in agonist‐dependent translocation of pathway by the CerK/C1P pathway. Many previous studies sphingosine kinase 1 to the plasma membrane. J Biol Chem. concerning characteristics of SphK18,10,13,18 and our results in 2006;281:11693‐11701. 5408 | NISHINO ET AL.

13. ter Braak M, Danneberg K, Lichte K, et al. Gαq‐mediated phospholipase A2 via a novel site in the C2 domain. JBiol plasma membrane translocation of sphingosine kinase‐1 and Chem. 2007;282:20467‐20474. cross‐activation of S1P receptors. Biochim Biophys Acta. 30. Ward KE, Bhardwaj N, Vora M, Chalfant CE, Lu H, Stahelin 2009;1791:357‐370. RV. The molecular basis of ceramide‐1‐phosphate recognition 14. Stace C, Ktistakis N. Phosphatidic acid‐ and phosphatidylser- by C2 domains. J Lipid Res. 2013;54:636‐648. ine‐binding proteins. Biochim Biophys Acta. 2006;1761:913‐926. 31. Baker DL, Pham TCT, Sparks MA. Structure and catalytic 15. Stahelin RV, Scott JL, Frick CT. Cellular and molecular function of sphingosine kinases: analysis by site‐directed muta- interactions of phosphoinositides and peripheral proteins. genesis and . Biochim Biophys Acta. 2013;1831: Chem Phys Lipids. 2014;182:3‐18. 139‐146. 16. Olivera A, Rosenthal J, Spiegel S. Effect of acidic phospholipids 32. Shen H, Giordano F, Wu Y, et al. Coupling between on sphingosine kinase. J Cell Biochem. 1996;60:529‐537. endocytosis and sphingosine kinase 1 recruitment. Nat Cell 17. Pitson SM, D’andrea RJ, Vandeleur L, et al. Human sphingo- Biol. 2014;16:652‐662. sine kinase: purification, molecular cloning and characteriza- 33. Pulkoski‐Gross MJ, Jenkins ML, Truman JP, et al. An intrinsic tion of the native and recombinant enzymes. Biochem J. lipid‐binding interface controls sphingosine kinase 1 function. J 2000;350:429‐441. Lipid Res. 2018;59:462‐474. 18. Delon C, Manifava M, Wood E, et al. Sphingosine kinase 1 is an 34. Wang Z, Min X, Xiao SH, et al. Molecular basis of sphingosine intracellular effector of phosphatidic acid. J Biol Chem. kinase 1 substrate recognition and . Structure. 2013;21: 2004;279:44763‐44774. 798‐809. 19. Don AS, Rosen H. A lipid binding domain in sphingosine 35. Simanshu DK, Kamlekar RK, Wijesinghe DS, et al. Non‐ kinase 2. Biochem Biophys Res Commun. 2009;380:87‐92. vesicular trafficking by a ceramide‐1‐phosphate transfer protein 20. Nakamura H, Hirabayashi T, Shimizu M, Murayama T. Ceramide‐ regulates eicosanoids. Nature. 2013;500:463‐467. 1‐phosphate activates cytosolic phospholipase A2α directly and by 36. Subramanian P, Vora M, Gentile LB, Stahelin RV, Chalfant CE.

PKC pathway. Biochem Pharmacol. 2006;71:850‐857. Anionic lipids activate group IVA cytosolic phospholipase A2 21. Pettus BJ, Bielawska A, Subramanian P, et al. Ceramide via distinct and separate mechanisms. J Lipid Res. 2007;48:

1‐phosphate is a direct activator of cytosolic phospholipase A2. 2701‐2708. J Biol Chem. 2004;279:11320‐11326. 37. Hengst JA, Guilford JM, Fox TE, Wang X, Conroy EJ, Yun JK. 22. Lamour NF, Wijesinghe DS, Mietla JA, Ward KE, Stahelin Sphingosine kinase 1 localized to the plasma membrane lipid RV, Chalfant CE. Ceramide kinase regulates the production raft microdomain overcomes serum deprivation induced of tumor necrosis factor α (TNFα) via inhibition of growth inhibition. Arch Biochem Biophys. 2009;492:62‐73. TNFα‐converting enzyme. J Biol Chem. 2011;286:42808‐42817. 38. Hengst JA, Guilford JM, Conroy EJ, Wang X, Yun JK. 23. Hankins JL, Ward KE, Linton SS, et al. Ceramide 1‐phosphate Enhancement of sphingosine kinase 1 catalytic activity by mediates endothelial cell invasion via the aneexin a2‐p11 deletion of 21 amino acids from the COOH‐terminus. Arch heterotetrameric protein complex. J Biol Chem. 2013;288: Biochem Biophys. 2010;494:23‐31. 19726‐19738. 39. Barr RK, Lynn HE, Moretti PAB, Khew‐Goodall Y, Pitson SM. 24. Takeshita E, Kume A, Maeda Y, Sakai H, Sakane F. Deactivation of sphingosine kinase 1 by protein phosphatase Diacylglycerol kinase γ is a novel anionic binding 2A. J Biol Chem. 2008;283:34994‐35002. protein with a selective binding preference. Biochem Biophys 40. Nakamura H, Murayama T. The role of sphingolipids in Res Commun. 2014;444:617‐621. . J Pharmacol Sci. 2014; 25. Zhao C, Du G, Skowronek K, Frohman MA, Bar‐Sagi D. 124:307‐312. Phospholipase D2‐generated phosphatidic acid couples EGFR 41. Chalfant CE, Spiegel S. Sphingosine 1‐phosphate and ceramide stimulation to Ras activation by Sos. Nature Cell Biol. 1‐phosphate: expanding roles in cell signaling. J Cell Sci. 2007;9:706‐712. 2005;118:4605‐4612. 26. Matsuzaki W, Takahashi H, Nakamura H, Murayama T. Effects of glycerophospholipids on ceramide kinase activity: cardioli- pin‐affected cellular formation of ceramide‐1‐phosphate. Biol Pharm Bull. 2016;39:1708‐1717. 27. Billich A, Ettmayer P. Fluorescence‐based assay and sphingo- How to cite this article: Nishino S, Yamashita H, sine kinases. Anal Biochem. 2004;326:114‐119. Tamori M, et al. Translocation and activation of 28. Tada E, Toyomura K, Nakamura H, et al. Activation of sphingosine kinase 1 by ceramide‐1‐phosphate. and ceramide kinase by vanadate via a tyrosine J Cell Biochem. 2019;120:5396‐5408. kinase‐mediated pathway. J Pharmacol Sci. 2010;114:420‐432. https://doi.org/10.1002/jcb.27818 29. Stahelin RV, Subramanian P, Vora M, Cho W, Chalfant CE. Cearamide‐1‐phosphate binds group IVA cytosolic