HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 51, 49545-49553, December 20, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/51/49545    most recent M206687200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pitson, S. M. Articles by Wattenberg, B. W. Search for Related Content PubMed PubMed Citation Articles by Pitson, S. M. Articles by Wattenberg, B. W. Social Bookmarking                      What's this?

The Nucleotide-binding Site of Human Sphingosine Kinase 1^*

Stuart M. Pitson^§¶, Paul A. B. Moretti, Julia R. Zebol, Reza Zareie, Claudia K. Derian^**, Andrew L. Darrow^**, Jenson Qi^**, Richard J. D'Andrea^§, Christopher J. Bagley^§, Mathew A. Vadas^§§§, and Binks W. Wattenberg^§§

From the  Hanson Institute, Division of Human Immunology, Institute of Medical and Veterinary Science, Frome Road, Adelaide SA 5000, Australia, the ^§ Department of Medicine, University of Adelaide, SA 5005, Australia, and ^** Johnson & Johnson Pharmaceutical Research & Development, Spring House, Pennsylvania 19477-0776

Received for publication, July 5, 2002, and in revised form, September 16, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Sphingosine kinase catalyzes the formation of sphingosine 1-phosphate, a lipid second messenger that has been implicated in a number of agonist-driven cellular responses including mitogenesis, anti-apoptosis, and expression of inflammatory molecules. Despite the importance of sphingosine kinase, very little is known regarding its structure or mechanism of catalysis. Moreover, sphingosine kinase does not contain recognizable catalytic or substrate-binding sites, based on sequence motifs found in other kinases. Here we have elucidated the nucleotide-binding site of human sphingosine kinase 1 (hSK1) through a combination of site-directed mutagenesis and affinity labeling with the ATP analogue, FSBA. We have shown that Gly⁸² of hSK1 is involved in ATP binding since mutation of this residue to alanine resulted in an enzyme with an ~45-fold higher K_m(ATP). We have also shown that Lys¹⁰³ is important in catalysis since an alanine substitution of this residue ablates catalytic activity. Furthermore, we have shown that this residue is covalently modified by FSBA. Our data, combined with amino acid sequence comparison, suggest a motif of SGDGX_17-21K is involved in nucleotide binding in the sphingosine kinases. This motif differs in primary sequence from all previously identified nucleotide-binding sites. It does, however, share some sequence and likely structural similarity with the highly conserved glycine-rich loop, which is known to be involved in anchoring and positioning the nucleotide in the catalytic site of many protein kinases.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Sphingosine kinase catalyzes the formation of sphingosine 1-phosphate (S1P),¹ a lipid mediator that acts in mammalian cells as an intracellular second messenger, as well as a ligand for cell-surface receptors (1-3). Intracellular S1P levels increase upon activation of sphingosine kinase (4, 5) and impact on a diverse range of important regulatory pathways, including the promotion of cell proliferation and inhibition of apoptosis (6, 7). Some specific effects of intracellular S1P include mobilization of intracellular calcium, activation of ERK1/2 and phospholipase D, stimulation of DNA binding activity of NF-B and AP-1, and inhibition of JNK and caspases (reviewed in Refs. 3 and 8). S1P is also an obligatory signaling intermediate in adhesion molecule expression of vascular endothelial cells (9), indicating a role for this lipid in the inflammatory response.

Despite the importance of sphingosine kinase, very little is known regarding this enzyme structure or mechanisms of catalysis and activation. We have recently cloned human sphingosine kinase 1 (hSK1) (10). A number of other sphingosine kinases from various sources have also been identified (11-15). Sequence analysis has shown that sphingosine kinases represent a distinct family of enzymes that share no overall sequence similarity to any other known proteins, or well characterized catalytic or regulatory domains (10, 12, 13, 16, 17).

The sphingosine kinases do, however, have sequence similarity to the putative catalytic domain of diacylglycerol kinases, with hSK1 possessing 17 of 24 very highly conserved amino acids of this domain (5) (Fig. 1A). The identification of this domain as a catalytic site in diacylglycerol kinases, however, is based largely on sequence conservation and limited truncation mutagenesis rather than any strong structural or biochemical information (21, 22). Thus, very little is known of the essential substrate binding and catalytic residues of the sphingosine kinases. We have previously reported the generation of a catalytically inactive mutant of hSK1 through a single point mutation in a highly conserved glycine residue (5). While mutation of the comparable glycine in diacylglycerol kinases also ablates catalytic activity (22-25), the role this residue serves in catalysis has not been established. Recent sequence analysis has indicated some weak structural similarity may exist in this region between the sphingosine kinases, diacylglycerol kinases, and part of the nucleotide-binding site of 6-phosphofructokinase (PFK) (26). This may suggest this region is involved in nucleotide binding in sphingosine kinases and diacylglycerol kinases, although this has not been experimentally established. We have also recently shown that a further highly conserved glycine residue in hSK1, Gly¹¹³, may also be important in catalysis since mutation of this residue to alanine increases the catalytic efficiency of the enzyme, while mutation to aspartate is deleterious to catalytic activity (27).

View larger version (47K): [in this window] [in a new window]   Fig. 1.   Sequence analysis of the putative catalytic domain and nucleotide-binding motif of sphingosine kinase. A, sequence alignment of the putative catalytic domain of sphingosine kinases (SKs) and some diacylglycerol kinases (DGKs). Residues highlighted in black are those that are highly conserved within the diacylglycerol kinase putative catalytic domain family, designated by the SMART data base (18, 19). Residues highlighted in gray are also highly conserved within the sphingosine kinases. Secondary structure prediction of the aligned region of the sphingosine kinase family was performed using the Jpred server (20). H indicates -helix and E indicates -strand. B, common nucleotide-binding motifs of various kinases and other ATP- and GTP-binding proteins.

Here we have begun to examine the substrate-binding sites of sphingosine kinase in the hope of better understanding the catalytic mechanism of the enzyme. We have identified key residues in the nucleotide-binding site in hSK1 through the use of site-directed mutagenesis and affinity labeling with the ATP analogue FSBA.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- D-erythro-Sphingosine was purchased from Biomol Research Laboratories Inc. (Plymouth Meeting, PA). ATP, GTP, FSBA, and 8-N₃ATP were from Sigma. [-³²P]ATP and [-³²P]GTP were purchased from Geneworks (Adelaide, Australia), and sequencing grade modified trypsin from Promega (Annandale, Australia).

Construction of hSK1 Mutants-- hSK1 cDNA (GenBank^TM accession number AF200328) was FLAG epitope-tagged at the 3'-end and subcloned into pALTER site-directed mutagenesis vector (Promega Corp.), as previously described (5). Single-stranded DNA was prepared and used as template for oligonucleotide-directed mutagenesis as detailed in the manufacturer's protocol. The mutagenic oligonucleotides used to generate the point mutant constructs were as follows: for hSK1^G26A, 5'-GAACCCGCGGGGCGCCAAGGGCAA-3'; hSK1^G26D, 5'-GCTGAACCCCCGGGGCGACAAGGGCAA-3'; hSK1^K27A, 5'-CGCGGCGGCGCCGGCAAGGCC-3'; hSK1^K29A, 5'-GGCAAGGGCGCCGCCTTGCAG-3'; hSK1^S79A, 5'-GTGGTCATGGCCGGCGACGGGCTG-3'; hSK1^S79D, 5'-GTGGTCATGGATGGAGACGGCCTGATGCAC-3'; hSK1^G80A, 5'-TCATGTCTGCAGACGGGCT-3'; hSK1^G80D, 5'-TCATGTCTGACGACGGCCTGATGCAC-3'; hSK1^G82A, 5'-GTCTGGAGATGCATTGATGCACG-3'; hSK1^K103A, 5'-GCCATCCAGGCCCCCCTGTGT-3'; hSK1^K103R, 5'-GCCATCCAGCGGCCGCTGTGTAGC-3'; hSK1^G111A, 5'-AGCCTCCCTGCAGCCTCTGGCAA-3'; hSK1^G111D, 5'-TCCCAGCAGACTCTGGCAA-3' (substituted nucleotides underlined). The mutants were sequenced to verify incorporation of the desired modification and the cDNA subsequently subcloned into pcDNA3 (Invitrogen, San Diego, CA) for transient transfection into HEK293T cells.

Deletion mutagenesis to generate hSK1^17-36, hSK1^72-96, hSK1^107-119, hSK1^165-198, hSK1^338-344, hSK1^344-384, and hSK1^368-384 was performed using PCR amplification from FLAG-tagged hSK1 in pGEM4Z (10). The oligonucleotides used to generate these mutant constructs were as follows: for hSK1^17-36, 5'-GCGGCAGGGCCGCGGGA-3' and 5'-CACGTGCAGCCCCTTTTGG-3'; hSK1^72-96, 5'-GCGGCCCAGCTCCTCCGA-3' and 5'-TGGGAGACCGCCATCCAGA-3'; hSK1^107-119, 5'-GCTACACAGGGGCTTCTGGA-3' and 5'-TTGAACCATTATGCTGGCTATGA-3'; hSK1^165-198, 5'GAAGA-GGCGCAGCCCCGAA-3' and 5'-CGTCTGGCAGCCTTGCGCA-3'; hSK1^338-344, 5'-CATACCTTTCCCATCCTTGGGC-3' and 5'-ATGGTTAGCGAGGCCGTGCA-3'; hSK1^344-384, 5'-CAATTCCCCATCCACTGCAAAC-3' and 5'-GACTACAAGGACGACGATGA-3'; hSK1^368-384, 5'-CTCCACGCAACCGCTGAC-3' and 5'-GACTACAAGGACGACGATGA-3'. The PCR products were kinased, self-ligated, and then transformed into E. coli. The mutants were sequenced to verify incorporation of the desired deletions and the cDNA subsequently subcloned into pcDNA3 (Invitrogen, San Diego, CA) for transient transfection into HEK293T cells.

Cell Culture and Transfection-- Human embryonic kidney cells (HEK293T, ATCC CRL-1573) were cultured and transfected using the calcium phosphate precipitation method as described previously (5). Cells were harvested and lysed by sonication (2 watts for 30 s at 4 °C) in lysis buffer containing 50 mM Tris/HCl (pH 7.4), 10% glycerol, 0.05% Triton X-100, 150 mM NaCl, 1 mM dithiothreitol, 2 mM Na₃VO₄, 10 mM NaF, 1 mM EDTA, and protease inhibitors (Complete^TM; Roche Molecular Biochemicals). Protein concentrations in cell homogenates were determined with either the Coomassie Brilliant Blue (Sigma) or bicinchoninic acid (Pierce) reagents using bovine serum albumin as standard.

Baculovirus Generation and Purification of Recombinant hSK1-- The hSK1 cDNA was subcloned upstream of a C-terminal His₆ tag in the baculovirus transplacement vector pFastBac1 (Invitrogen). Recombinant baculovirus expressing hSK1-His was generated in accordance with the manufacturer's recommendations. Western blot analyses of infected cell lysates, with an anti-His monoclonal antibody (Amersham Biosciences), was initially used to confirm expression.

Sf9 cells (1-2 × 10⁶ cells/ml) were infected with the recombinant baculovirus (1-5 MOI) for 72 h. Infected cells were harvested and washed with phosphate-buffered saline and all subsequent steps were carried out at 4 °C. Cells were resuspended in 10 volumes of Buffer A, containing 50 mM Na₂HPO₄ (pH 8.0), 300 mM NaCl, 10% glycerol, 1 mM -mercaptoethanol, 1 mM sodium orthovanadate, 1 mM NaF, and protease inhibitors (Complete^TM). Cells were sonicated for 10 s and the homogenate mixed with Tween 20 to a final concentration of 0.5%. Additional NaCl was added to a final concentration of 600 mM and after incubation on ice for 30 min, the homogenate was centrifuged for 30 min at 15,000 × g. The resulting supernatant was incubated with 0.5 ml of Buffer A-equilibrated Ni-nitrilotriacetic acid (NTA) (Qiagen, Valencia, CA) slurry per 500 ml of original Sf9 cell culture for 1 h at 4 °C with continuous mixing. Following centrifugation at 1000 × g, the Ni-NTA was then packed into columns and washed with 5 volumes of Buffer A. The columns were subsequently washed with 20 volumes of Buffer A containing 25 mM imidazole, and eluted with 2-3 volumes of 200 mM imidazole in 50 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 10% glycerol.

Sphingosine Kinase Assays-- Sphingosine kinase activity was routinely determined using D-erythro-sphingosine and [-³²P]ATP as substrates, as described previously (10). In some cases [-³²P]GTP was used as the phosphate donor instead of [-³²P]ATP. A unit of sphingosine kinase activity is defined as the amount of enzyme required to produce 1 pmol of S1P/min. Substrate kinetics in cell extracts were determined after the extracts were desalted in NAP-5 columns (Amersham Biosciences) to remove other possible phosphate donors. The kinetic data were analyzed using Michaelis-Menten kinetics with the non-linear regression program Hyper 1.1s. In all cases relative sphingosine kinase activities and substrate kinetic values were adjusted to account for the slightly different expression levels of the various mutant proteins following quantitative Western blotting.

Calmodulin Binding Assay for Assessing Correct Folding of Sphingosine Kinase Mutants-- As detailed previously (27), the effect of mutations on gross folding of hSK1 was assessed by examining the interaction of the mutants with calcium/calmodulin (CaM), which is known to occur only with correctly folded hSK1 protein (10). Briefly, HEK293T cells overexpressing wild type or mutant hSK1 were harvested and lysed as described above. The cell lysates were then centrifuged (13000 × g, 15 min at 4 °C) to remove cell debris. Aliquots of the supernatants were added to tubes containing CaM-Sepharose 4B (Amersham Biosciences) pre-equilibrated with binding buffer composed of 50 mM Tris/HCl (pH 7.4), 5 mM CaCl₂, 200 mM NaCl, 10% (w/v) glycerol, 0.05% (w/v) Triton X-100, 1 mM dithiothreitol, 2 mM Na₃VO₄, 10 mM NaF, and protease inhibitors (Complete^TM) and incubated for 2 h at 4 °C with continuous mixing. The CaM-Sepharose 4B beads were then pelleted by centrifugation (5000 × g, 5 min at 4 °C) and washed twice with binding buffer. hSK1 bound to the beads were then resolved by SDS-PAGE and visualized by Western blotting via the FLAG epitope. Sepharose CL-4B (Amersham Biosciences) was used as a control for nonspecific binding to the Sepharose 4B beads.

Western Blotting-- SDS-PAGE was performed on cell lysates using 12% acrylamide gels. Proteins were blotted to nitrocellulose and the membranes blocked overnight at 4 °C in phosphate-buffered saline containing 5% skim milk powder and 0.1% (w/v) Triton X-100. hSK1 expression levels in cell lysates were quantitated over a dilution series of the lysates with the monoclonal M2 anti-FLAG antibody (Sigma), with the immunocomplexes detected with both horseradish peroxidase anti-mouse (Pierce) IgG using an enhanced chemiluminescence kit (ECL, Amersham Biosciences) and alkaline phosphatase anti-mouse IgG (Selinus/Amrad, Melbourne, Australia) using an ECF kit (Amersham Biosciences) and a Molecular Dynamics (Sunnyvale, CA) fluorimager.

Inhibition Studies with FSBA and 8-N₃ATP-- To test their ability to irreversibly inhibit sphingosine kinase activity FSBA and 8-N₃ATP were added to either purified recombinant hSK1 or extracts of HEK293T cells overexpressing hSK1 or hSK^K103R. FSBA was dissolved in dimethylsulfoxide (Me₂SO) and added to a final concentration of 1 mM (1% (v/v) Me₂SO) in reaction mixture containing hSK1 in 50 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM MgCl₂, and 10% glycerol. This mixture was then incubated in the presence or absence of 2 mM ATP at 22 °C for 30 min. The reaction mixture was then desalted in an NAP-5 column (Amersham Biosciences) to remove unreacted FSBA and assayed for sphingosine kinase activity. Similarly, 8-N₃ATP (100 µM) was added to purified recombinant hSK1 in 50 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM MgCl₂, and 10% glycerol in the presence or absence of 2 mM ATP. The 8-N₃ATP was then photoactivated by exposure to a 12-watt UV lamp (254 nm) for 90 s at a distance of 5 cm. The reaction mixture was then desalted, assayed for sphingosine kinase activity, and the residual activity compared with hSK1 treated in an identical manner, but without the addition of 8-N₃ATP.

Affinity Labeling of hSK1 with FSBA-- FSBA (dissolved in Me₂SO) was added to a final concentration of 1 mM (1% v/v Me₂SO) in 100 µl of reaction mixture containing 50 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM MgCl₂, 10% glycerol, and 15 µg of purified recombinant hSK1. This mixture was then incubated in the presence or absence of 20 mM ATP at 22 °C for 60 min and a further 16 h at 4 °C. The reaction was then stopped by the addition of 1 mM dithiothreitol. For analysis of the FSBA labeling of the whole hSK1 protein, this mixture was then acidified with acetic acid, purified on a Macrosphere C8 column (Alltech) using a linear (0-70%) acetonitrile gradient with 0.1% trifluoroacetic acid, and then analyzed by mass spectrometry. FSBA labeling for peptide analysis was carried out in the same manner, except that after termination of the FSBA labeling step by the addition of dithiothreitol, 0.05% (w/v) SDS was added, and the mixture then heated at 100 °C for 6 min. This step proved essential for obtaining adequate yields of the labeled peptide. After cooling, 100 ng of trypsin was added, and the mixture was incubated at 37 °C for 16 h with shaking. A further 100 ng of trypsin was then added and the mixture again incubated at 37 °C for 6 h. Dithiothreitol (0.2 mM), solid urea to a final concentration of 8 M, and 1 M guanidine hydrochloride were then added to the tryptic peptides. Insoluble material was removed by centrifugation (10,000 × g, 5 min) and the supernatant acidified with acetic acid. The peptides were then separated on a Macrosphere C8 column (Alltech) using an acetonitrile gradient (0-55%) with 0.1% trifluoroacetic acid and analyzed by mass spectrometry.

Mass Spectrometry-- Whole-protein analysis and preliminary peptide analyses were performed on a PE/Sciex API-100 electrospray-ionization mass spectrometer (PE Biosystems, Melbourne, Australia) operating in positive-ion mode. Spectra of the proteins' ion series were converted to a true-mass scale using the instrument's software. Data files corresponding to peptide maps of hSK1 were interrogated by Sherpa software (University of Washington) for the presence of ions corresponding to FSBA-derivatized peptides. The derivatized adenosylbenzoylsulfonyl-peptides were further characterized using an electrospray-ionization quadrupole/time-of-flight mass spectrometer (Q-TOF2, Micromass, Manchester, UK). The ion-series spectra for the analyzed peptides were converted to a true-mass scale using the instrument's software.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Deletion Mutagenesis of hSK1-- In an attempt to more clearly define the essential domains required for catalysis and regulation of hSK1 we generated and analyzed a series of deletion mutants of this protein. Sequence alignments of the sphingosine kinase 1 and 2 isoforms from human (10, 13), mouse (12, 13), and rat (15), together with sphingosine kinases from Saccharomyces cerevisiae (11), Arabidopsis thaliana (14), and putative sphingosine kinases from Schizosaccharomyces pombe and Caenorhabditis elegans have indicated five highly conserved regions within these proteins (10, 12, 13). These conserved regions span residues 17-36, 72-96, 107-119, 165-198, and 338-344 of hSK1. Therefore, we generated deletion mutants of hSK1 where these five highly conserved regions were individually removed (hSK1^17-36, hSK1^72-96, hSK1^107-119, hSK1^165-198, and hSK1^338-344). In addition we generated and analyzed two truncation hSK1 mutants lacking the C-terminal 41 and 17 amino acids of this protein (hSK1^344-384 and hSK1^368-384, respectively).

These hSK1 mutants were expressed in HEK293T cells and the cell lysates assayed for sphingosine kinase activity. Overexpression of wild type hSK1 routinely results in an elevation of sphingosine kinase activity in cell extracts of 1000-fold over extracts of cells transfected with a control plasmid. While protein expression levels of the hSK1 mutants with internal deletions were similar to that of the wild type hSK1 (Fig. 2), in no case did the extracts derived from mutant transfected cells exhibit sphingosine kinase activity greater than that of the endogenous levels present in HEK293T cells (data not shown). To assess the effect of these mutations on gross folding of hSK1 we exploited the known interaction of hSK1 with calmodulin (10, 26). We have previously observed that this binding is sensitive to the folded state of hSK1 (10). This is demonstrated, for example, by the observation that when expressed in Escherichia coli, only catalytically active, but not inactive recombinant hSK1 is bound by immobilized calmodulin (10). This analysis showed that, unlike wild type hSK1, these hSK1 mutants could not bind specifically to calmodulin-Sepharose (Fig. 2). Since the mutagenesis did not alter any of the putative calmodulin-binding motifs in hSK1 this suggests that all of these mutations yielded misfolded hSK1 protein. Therefore, the lack of catalytic activity in these mutants probably reflects their inability to fold correctly, rather than necessarily indicating that the deleted regions are required for catalysis. Similar lack of activity was also observed for the hSK1^344-384 truncation mutant lacking the C-terminal 41 amino acids (data not shown). The protein expression level of this mutant was comparable to wild type hSK1, but again, the lack of calmodulin binding suggested that this truncation yielded misfolded hSK1 protein (Fig. 2). In contrast, the hSK1^368-384 truncation mutant lacking the C-terminal 17 amino acids generated a correctly folded protein (Fig. 2) that possessed unaltered catalytic activity compared with wild type hSK1 (Fig. 3).

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Calmodulin binding analysis to assess correct folding of hSK1 mutants. The selective binding of the hSK1 mutants to calmodulin (CaM)-Sepharose was used as an indicator of correct protein folding. Extracts from HEK293T cells expressing the various hSK1 mutants (Load) were incubated with CaM-Sepharose, the beads washed and subjected to SDS-PAGE. Bound hSK1 proteins were then visualized by Western blotting via their FLAG epitope. Binding to Sepharose CL-4B was used as a control to account for any nonspecific binding to the Sepharose 4B beads.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Site-directed mutagenesis of hSK1. HEK293T cells were transfected with pcDNA3 plasmids containing cDNAs of wild type hSK1 or the various hSK1 mutants, or empty vector alone (pcDNA3). The cells were harvested 24 h after transfection and analyzed for sphingosine kinase activity and expression levels of the FLAG-tagged hSK1 constructs. Sphingosine kinase activities are expressed relative to the activity of wild type hSK1 and are corrected for slight differences in their expression levels (lower panel). All data shown are means ± S.D. of triplicate determinations from three independent experiments.

Site-directed Mutagenesis of hSK1-- Since the deletion mutagenesis approach was not helpful in defining the catalytic domain of hSK1, we undertook a more selective site-directed mutagenesis of hSK1. Residues highly conserved between the diacylglycerol kinase putative catalytic domain and other sphingosine kinases (Fig. 1A) were selected for mutation in hSK1. Surprisingly little is known of the structure or essential catalytic residues of diacylglycerol kinases (22). The similarity in residues 16-153 of hSK1 to the putative diacylglycerol kinase catalytic domain, however, provided an opportunity to define the nucleotide-binding site of hSK1 since, presumably, this similarity arises primarily from the ability of both enzymes to utilize ATP. Conserved glycines within glycine-rich regions of hSK1 were highly represented in the residues selected for mutagenesis since this is a characteristic often associated with loop regions coordinating nucleotide binding in many protein kinases (28, 29) and other ATP- and GTP-binding proteins (30, 31) (Fig. 1B). Sequence analysis has shown that there are three such conserved regions rich in glycine residues in hSK1 (Fig. 1A). Our previous studies have already suggested at least two of these regions are important in catalysis since Asp mutations at Gly⁸² and Gly¹¹³ had deleterious effects on the catalytic activity of hSK1 (5, 27). Thus, to extend these studies we mutated Gly²⁶, Gly⁸⁰, and Gly¹¹¹ to study their importance in hSK1. Initially we generated disruptive Gly  Asp mutations. This increased the size of the residue and introduced a negative charge, and therefore was expected to be maximally disruptive to the short-range interactions of the glycine residues. We then also generated more conservative Gly  Ala mutations at these three residues, as well as Gly⁸², to test whether subtle changes in the structure at these sites would affect catalytic activity.

Expression of hSK1^G26D, hSK1^G80D, hSK1^G111D, and wild type hSK1 in HEK293T cells, followed by Western blotting showed that all four proteins were present in similar mass levels (Fig. 3). Importantly, calmodulin binding assays also indicated that, unlike the deletion mutants, the hSK1 point mutants were correctly folded since they bound specifically to calmodulin-Sepharose (Fig. 2). Analysis of the sphingosine kinase activities of these proteins in the cell lysates, however, showed that in all cases the Gly  Asp mutations dramatically decreased the catalytic activity of these mutants (Fig. 3). In fact, hSK1^G26D and hSK1^G111D possessed less than 1.2% of the activity of wild type hSK1, while hSK1^G80D was catalytically inactive. These results suggest that all three glycine-rich pockets may be important in the action of hSK1, and are consistent with the previously reported catalytically inactive hSK1^G82D mutant (5), and hSK1^G113D mutant, which possessed less than 2% of the activity of wild type hSK1 (27).

To probe the role of these residues further, we examined the effect of the more conservative Gly  Ala mutations by analyzing the hSK1^G26A, hSK1^G80A, and hSK1^G111A mutants, as well as the hSK1^G82A mutant to complement our earlier studies (5). Again, expression of these mutants in HEK293T cells showed that all were expressed to similar mass levels, with the mutations having no apparent effect on gross folding of the proteins (Fig. 2). Although mutation of glycine to alanine is relatively subtle in terms of structure and charge, this mutation at either Gly²⁶, Gly⁸⁰, Gly⁸², or Gly¹¹¹ of hSK1 had a profound effect on the catalytic activity of this enzyme. In fact, hSK1^G26A, hSK1^G82A, and hSK1^G111A possessed only 4, 5.5, and 1.7%, respectively, of the sphingosine kinase activity of wild type hSK1 under the standard assay conditions, while hSK1^G80A was catalytically inactive (Fig. 3). Although activity of hSK1^G26A, hSK1^G82A, and hSK1^G111A was substantially reduced, there was sufficient sphingosine kinase activity above the endogenous levels to determine the substrate kinetics of these modified hSK1s. This was then performed to gain an insight into the effect of these mutations on substrate affinities (Table I). Substrate kinetics for both ATP and GTP were analyzed since wild type hSK1 was found to be able to utilize both these nucleotides as phosphate donors, although with considerably higher K_m(GTP) (Table I). Remarkably, the hSK1^G82A mutant had an ~45-fold higher K_m(ATP), and 3-fold higher K_m(GTP), compared with wild type hSK1. This suggests that Gly⁸² is involved in nucleotide binding and is consistent with the lack of catalytic activity of the hSK1^G82D mutant (5). In contrast, the hSK1^G26A mutant had a very similar K_m(ATP) to wild type hSK1. This suggests that although the Gly  Ala or Asp mutation at this residue is deleterious to catalytic activity, Gly²⁶ may not be directly involved in nucleotide binding. Both mutants had similar K_{m(sphingosine)} values as wild type hSK1, suggesting the binding pocket for this substrate was unaltered by the mutagenesis.

Since the data suggested the importance of the region encompassing Gly⁸⁰ to Gly⁸² in nucleotide binding, we performed further mutagenesis on Ser⁷⁹, another highly conserved residue in this region. We again generated a disruptive Ser  Asp mutation, and a more conservative Ser  Ala mutation. As for the glycine mutations, the disruptive Ser  Asp mutation ablated the catalytic activity of hSK1 (Fig. 3), while leaving the expression levels and apparent gross folding of the protein unaffected (Fig. 2). The more conservative alanine mutation, however, had only a minor effect on catalytic activity. Substrate kinetic analysis of the hSK1^S79A mutant also revealed only a minor increase (1.5-fold) in the K_m(ATP) compared with wild type hSK1 (Table I). The effect on GTP utilization, however, was more marked. The hSK1^S79A mutant exhibited a 3-fold increase in the K_m(GTP) and a 7.5-fold decrease in the relative specificity constant (V_max/K_m) for this nucleotide. This suggests the involvement of Ser⁷⁹ in nucleotide binding and provides further evidence for the importance of this region in this function.

Basic (generally lysine) residues are usually required for direct interaction with the negatively charged nucleotide phosphate groups. For this reason, Lys²⁷ and Lys²⁹ in hSK1 have been previously proposed to be involved in nucleotide binding due to the concentration of positive charge (12, 14, 16, 17, 32). Therefore, we generated alanine mutants at these residues to test this hypothesis. Expression of the hSK1^K27A and hSK1^K29A mutants in HEK293T cells, however, revealed that these mutations had only minor effects on catalytic activity. hSK1^K27A and hSK1^K29A possessed 60 and 85%, respectively, of the catalytic activity of wild type hSK1 (Fig. 3), and also had comparable substrate kinetics for ATP and sphingosine as wild type hSK1 (Table I). Therefore, Lys²⁷ and Lys²⁹ do not appear to have a major role in catalysis in hSK1.

Lys¹⁰³ resides 21 residues C-terminal of the glycine-rich region around residues 79-82 in hSK1. This placement is somewhat reminiscent of the nucleotide-binding loop motif of many protein kinases (GXGXXGX_14-23K) (28, 29). In protein kinases the glycine-rich pocket forms a flexible loop for positioning of the ATP and solvent exclusion, and the lysine is involved in anchoring of the - or -phosphate of this nucleotide (28). Therefore, we generated a Lys¹⁰³  Ala mutation in hSK1. This was expected to be maximally disruptive to the short-range interactions of the lysine since it decreased the size of the residue and removed the positive charge. We also generated the more conservative Lys¹⁰³  Arg mutation, where the positive charge of the residue was retained, to test whether subtle changes in the structure at this site would affect catalytic activity. Expression of hSK^K103A in HEK293T cells, indeed, showed that this Lys  Ala mutation ablates catalytic activity (Fig. 3), while not affecting gross folding of the protein (Fig. 2). hSK1^K103R, with the more conservative Lys  Arg mutation, however, displayed substantial catalytic activity. Substrate kinetic analysis of the hSK1^K103R mutant demonstrated that it possessed a V_max of 64% of wild type hSK1, and showed similar K_m(ATP), K_m(GTP), and K_{m(sphingosine)} values to that of the wild type enzyme (Table I).

Inhibition of hSK1 Activity by Affinity Labeling ATP Analogues-- To complement the mutagenic approach for elucidation of the nucleotide-binding site of hSK1 we wished to also map this site by covalent modification with an affinity labeling ATP analogue. Both FSBA and 8-N₃ATP have been used extensively for this purpose with other nucleotide-binding enzymes (33-41). Therefore, we initially examined the ability of these two reagents to irreversibly inhibit the catalytic activity of hSK1 as a measure of their ability to bind to the nucleotide-binding site of this protein. Both 1 mM of FSBA and 50 µM photoactivated 8-N₃ATP irreversibly inhibited the sphingosine kinase activity of purified recombinant hSK1 (Fig. 4) by over 50% under the conditions used. Importantly, this inhibition of hSK1 activity by FSBA was prevented by the presence of 10 mM ATP (Fig. 4), suggesting that FSBA was specifically acting at the nucleotide-binding site. In contrast, inhibition of hSK1 by 8-N₃ATP was not affected by the presence of a 200-fold molar excess of ATP, indicating substantial nonspecific photoaffinity labeling of hSK1 remote from the nucleotide binding. Thus, FSBA was used for all further affinity labeling experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Inhibition of hSK1 activity by FSBA and 8-N3ATP. Purified recombinant hSK1 was incubated with 1 mM FSBA (A) or 100 µM photoactivated 8-N3ATP (B) in the presence (open bars) or absence (filled bars) of 2 mM ATP. Residual sphingosine kinase activity was then assayed and compared with that remaining in Me2SO or photoactivation controls for FSBA and 8-N3ATP treatment, respectively. All data shown are means ± S.D. of triplicate determinations from three independent experiments.

Mapping of the Nucleotide-binding Site of hSK1 by Covalent Modification with FSBA-- Modification of purified recombinant hSK1 by FSBA was monitored by mass spectrometry. Analysis of the unmodified recombinant hSK1 revealed a protein with a molecular mass of 43,626 Da. This molecular mass is consistent with proteolytic cleavage of the N-terminal methionine and subsequent acetylation of Asp² on the C-terminally His-tagged hSK1. Treatment of hSK1 with FSBA under the conditions used to inhibit activity resulted in the appearance of a protein peak of molecular mass of 44,059 Da, precisely the mass increase of 433 Da predicted from the addition of a sulfonylbenzoyladenosine group. This analysis showed that ~50% of hSK1 was labeled by FSBA (Fig. 5), consistent with the observed 50% inhibition of catalytic activity by this reagent (Fig. 4) under the same conditions. Only one modification per hSK1 molecule could be detected, with this labeling being specific for the nucleotide-binding site since it was abolished by the presence of a 20-fold molar excess of ATP (Fig. 5, FSBA + ATP).

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Mass spectrometric analysis of the covalent modification of hSK1 by FSBA. The intact purified recombinant hSK1 was analyzed by mass spectrometry following incubating with either Me2SO (Control), 1 mM FSBA in Me2SO (FSBA) or 1 mM FSBA in Me2SO with 20 mM ATP (FSBA + ATP).

Identification of the modified amino acid was accomplished by mass spectrometric analysis of trypsin-digested, FSBA-labeled hSK1. A mass increase of 433 Da, due to modification by FSBA, was observed in two peptides (Fig. 6). One peptide, showing a mass shift from 8591 to 9024 Da, corresponds to residues 66-144, and a second, showing a mass shift from 8747 to 9180 Da, corresponds to residues 66-145 of hSK1. These two peptides clearly arose from incomplete digestion of hSK1 at the dibasic bond occurring between residues 144 and 145. Therefore, the modification by FSBA occurs in the region spanning residues 66-144 of hSK1. Under the conditions used, only lysine or tyrosine residues are modified by FSBA (34). Thus, the only residues in the identified peptide that may have been labeled are Lys¹⁰³, Tyr¹²³, or Tyr¹²⁶.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Mass spectrometric analysis of tryptic peptides released from hSK1 following covalent modification by FSBA. Tryptic peptides released following incubation of the purified recombinant hSK1 either Me2SO (Control), or 1 mM FSBA in Me2SO (FSBA) were analyzed by mass spectrometry.

Since our mutagenesis had already shown Lys¹⁰³ to be important in catalysis, we predicted that this residue was the target of FSBA modification. To confirm this we performed FSBA inhibition experiments on the hSK1^K103R mutant described earlier. Since FSBA is unreactive with arginine residues we predicted that if Lys¹⁰³ was the reactive residue in wild type hSK1, then the Lys¹⁰³  Arg mutant should be refractive to FSBA inhibition. Following reaction of FSBA with hSK1^K103R and subsequent removal of the unreacted affinity label, we could, indeed, detect no effect of FSBA on the catalytic activity of this mutant (Fig. 7). This confirms Lys¹⁰³ as the target of FSBA modification in hSK1.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   FSBA does not inhibit hSK1K103R. Residual sphingosine kinase activity was measured following incubation of wild type hSK1 or hSK1K103R with either Me2SO (Control), 1 mM FSBA in Me2SO (FSBA), or 1 mM FSBA in Me2SO with 2 mM ATP (FSBA + ATP). All data shown are means ± S.D. of triplicate determinations from three independent experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The sphingosine kinases comprise a novel class of enzymes that show very little sequence similarity to other known proteins. With the exception of the poorly characterized diacylglycerol kinase putative catalytic domain, sphingosine kinases have no other recognizable domains or functional sequence motifs. Although we have recently generated inactive and hyperactive sphingosine kinase mutants (5, 27) virtually nothing was known of the substrate-binding sites or essential catalytic residues in this enzyme. In the current study we have identified in hSK1, through site-directed mutagenesis and affinity labeling with FSBA, a novel nucleotide-binding motif that differs in primary sequence from all previously identified nucleotide-binding sites.

Our initial deletion mutagenesis of hSK1 did not yield correctly folded hSK1 mutants when expressed in HEK293T cells. Since the five internal regions deleted in hSK1 are highly conserved between all known sphingosine kinases (10, 12) it is perhaps not unexpected that these regions are important to the structural integrity of the protein. More surprising, however, was that truncation of the C-terminal 41 amino acids also affected the gross folding of hSK1 since this region is considerably more divergent in sequence between the known sphingosine kinases (10). The basis for this instability is not presently clear, but suggests that there are important structural interactions between all segments of the sequence. This is consistent with its instability in tissue extracts (10, 42) and is probably due to the hydrophobic nature of the protein, which may make it prone to poor solubility and aggregation. In contrast, truncation of the C-terminal 17 amino acids yielded a protein with unaltered catalytic activity. This is consistent with the observations that this proline-rich region appears to reside, at least in part, on the surface of the protein since it contains a demonstrated protein-docking site for TRAF2 (43), and antibodies raised against this region recognize the intact, folded protein (44).

Site-directed mutagenesis of conserved residues within the hSK1 sequence with similarity to the diacylglycerol kinase catalytic domain has provided insight into the residues critical for catalysis. Interestingly, most of the conserved residues in this domain family are mainly centered around three highly conserved glycine-rich pockets (Fig. 1A). This suggested that one or more of these regions may be involved in nucleotide binding since glycine-rich regions are commonly involved in such a role in other enzymes (Fig. 1B). Indeed, the site-directed mutagenesis indicated that all three regions were important for catalytic activity since conservative Gly  Ala mutations in each region were deleterious to activity. Only the region centered around Ser⁷⁹ to Gly⁸², however, appeared to be involved in nucleotide binding. This was most clearly demonstrated by the much higher K_m(ATP) and K_m(GTP) of the Gly⁸²  Ala mutant and the complete abolition of catalytic activity by the Gly⁸⁰  Ala mutation. Furthermore, the Ser⁷⁹  Ala mutant, while having a comparable V_max to wild type hSK1, has lower nucleotide affinity, particularly for GTP. In contrast, mutations in the other glycine rich regions, including Gly¹¹³ (27), while affecting catalytic efficiency, did not alter any measurable binding affinity for ATP.

We have also shown that Lys²⁷ and Lys²⁹, in the first glycine-rich region in hSK1, are not involved in nucleotide binding or catalysis. This is notable since these residues have been previously proposed by several groups to comprise a likely ATP-binding site in several sphingosine kinases (12, 14, 16, 17, 32). Clearly, this is not the case. This conclusion is consistent with the absence of these residues in human and mouse sphingosine kinase 2 isoforms (13). Instead, we have revealed an essential role for Lys¹⁰³ in the catalytic activity of hSK1 since mutation of this residue to alanine resulted in an inactive mutant. Affinity labeling of hSK1 with the ATP analogue FSBA confirmed the location of Lys¹⁰³ in the ATP-binding pocket. Although, in the extended conformation, the reactive sulfonyl fluoride moiety of FSBA is located in a position analogous to the -phosphate of ATP (34), in protein kinases this reagent usually modifies the lysine involved in anchoring of the - or -phosphate of this nucleotide (33, 35-37). In protein kinases this lysine resides around 14 to 23 residues C-terminal to the glycine-rich loop that forms the ATP-binding pocket. Interestingly, this is analogous to the position of Lys¹⁰³ located 21 residues C-terminal of the glycine-rich region of hSK1, which we have already identified as part of the nucleotide-binding region. This would suggest a similarity in the nucleotide-binding site of sphingosine kinases and protein kinases. Our affinity labeling studies with FSBA, and by analogy to protein kinases, indicate that Lys¹⁰³ in hSK1 is probably involved in anchoring and orienting the nucleotide by interacting with its - and -phosphates (29). In virtually all protein kinases (29) and phosphatidylinositol phosphate (PIP) kinases (45, 46), conservative mutation to arginine of the corresponding lysine ablates catalytic activity. However, suprisingly in hSK1, the Lys¹⁰³  Arg mutation results in an enzyme with considerable catalytic activity. Substrate kinetic analysis revealed the V_max of the hSK^K103R mutant was reduced by a third compared with the wild type hSK1, but possessed an unaltered K_m(ATP) (Table I). Although this is unusual, almost identical results have been recently described with an arginine mutation at the corresponding invariant lysine of atypical protein kinase C (47), with this mutant enzyme also having unaltered binding affinity for ATP and a marginally reduced V_max. The only other report of similar results has been with the cyclin-dependent kinase-activating kinase (48), although this enzyme appears to have a markedly divergent nucleotide-binding pocket. The results with hSK1 would suggest that the Lys¹⁰³ contacts a nucleotide phosphate, but does not contribute significantly to binding energy. Instead, the function of this residue may be to orientate the nucleotide to enable efficient phosphotransfer by the catalytic residues.

Lys¹⁰³ is highly conserved among the sphingosine kinases, occuring between 17 and 21 amino acids C-terminal of the conserved glycine-rich region. In murine sphingosine kinase 2 (13) and the putative C. elegans sphingosine kinase, however, an arginine appears to substitute for this lysine. Our data indicate that this may have only minor effects on catalytic activity. Surprisingly, the putative rice (Oryza sativa) sphingosine kinase does not possess a lysine or arginine in this region. Neither do some diacylglycerol kinases (22). This suggests that although these enzymes have sequence similarity to hSK1, they must have alternative mechanisms to compensate for the lack of this positively charged residue in its proposed role of orienting the nucleotide.

Our data, combined with amino acid sequence comparison, suggest a motif of SGDGX_17-21K is involved in nucleotide binding in the sphingosine kinases. While the glycines in this motif appear invariant, the serine is somewhat more variable since glycines occur in this position in a small number of sphingosine kinases (Fig. 1A) and almost all diacylglycerol kinases (22). This, together with our finding that hSK1 activity was not substantially affected by an alanine substitution in this position may indicate that small residues are tolerated in this position.

While this motif is unique, it has some similarities to the nucleotide-binding loop of protein kinases (GXGXXGX_14-23K) (28, 29), with the two glycines of the sphingosine kinase motif possibly corresponding to the first two glycines of the protein kinase motif. This protein kinase glycine-rich loop is very highly conserved, with the first and second glycines essentially invariant. The third glycine is less conserved, occurring in 85% of all protein kinases, but is always a small residue (28, 29). Notably, in the sphingosine kinases this third glycine is not present, with this position occupied by bulky residues, usually histidine (Fig. 1A). The mammalian diacylglycerol kinases, however, do usually have a glycine in this position (22). Although not always included in the protein kinase nucleotide-binding motif, there is also an essentially invariant valine two residues C-terminal to the glycine-rich loop in protein kinases (i.e. GXGXXGXV) (28) that is important in catalysis through its hydrophobic interactions with the adenosine base (28, 29, 49). This residue, corresponding to Val⁸⁷ in hSK1, is also highly conserved in the sphingosine kinases and diacylglycerol kinases and provides further evidence for similarity in their nucleotide-binding sites to that of the protein kinases.

By analogy to the protein kinases the SGDG motif in sphingosine kinases is probably involved in forming a flexible loop between two -strands that creates the nucleotide-binding pocket (28). Such a conformation is consistent with our secondary structure predictions from the primary sequence of hSK1 (Fig. 1A). In protein kinases this glycine-rich loop provides the conformational flexibility to anchor and position the nucleotide, and also exclude water from the binding pocket. This latter role appears important to ensure low ATPase activity (50). While the sphingosine kinase nucleotide-binding region is similar to that of protein kinases, it is clearly different since it lacks the third conserved glycine of the protein kinase motif. Since the glycine-rich -sheet contacts the whole nucleotide (28) it is possible that this difference may be one reason for the different nucleotide specificity of these two enzyme classes, with hSK1 capable of utilizing both ATP and GTP as phosphate donors (Table I), while GTP is generally not tolerated by the protein kinases. The sphingosine kinase nucleotide-binding region is also somewhat reminiscent of the nucleotide-binding loop motif of PIP kinases (GXSGSX₁₂K), which can also utilize both ATP and GTP (45, 46). Furthermore, structural elucidation of PIP kinase II has, indeed, shown that its nucleotide-binding loop differs slightly in conformation from that of protein kinases (45). Interestingly, a recent study has suggested some structural similarity may exist between the SGDG region in the sphingosine kinases and a GGDG motif in PFK (26), which forms part of the ATP-binding site of this enzyme (51). Since, in PFK, the aspartate in this region chelates the Mg²⁺ bridging the - and -phosphates of ATP, it is tempting to speculate a similar role for the comparable aspartate in the sphingosine kinases. Structural modeling of hSK1 based on its published sequence alignment to PFK (not shown), however, places Lys¹⁰³ in a site remote from the nucleotide-binding pocket. Since we have unambiguously shown in the current study that this lysine resides in the nucleotide-binding site there is some question over the extent of the proposed structural similarity of the sphingosine kinases to PFK.

In this study we have shown, through a combination of site-directed mutagenesis and affinity labeling with the ATP analogue FSBA, the involvement in nucleotide binding of Lys¹⁰³ and the glycine-rich region centered around Ser⁷⁹ to Gly⁸² in hSK1. While this nucleotide-binding loop has similarities to that of the protein kinases, hSK1 does not possess recognizable protein kinase HRDLK or DFG motifs that are involved in the phosphotransfer reaction, and interaction with the Mg²⁺ bridging the - and -phosphates of ATP, respectively (29). Neither does it appear to have the equivalent motifs described in PIP kinases (46) and phosphoinositide 3-kinases (52). Clearly further biochemical analysis and direct structural determination will be required to fully elucidate this novel catalytic site.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The 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: Hanson Institute, Div. of Human Immunology, Inst. of Medical and Veterinary Science, Frome Rd., Adelaide SA 5000, Australia. Tel.: 618-8222-3480; Fax: 618-8232-4092; E-mail: stuart.pitson@imvs.sa.gov.au.

Supported by a Georgina Dowling Medical Research Associateship from the University of Adelaide.

Supported by an H. M. Lloyd Senior Research Fellowship in Oncology from the University of Adelaide.

^§§ Supported by grants from the National Health and Medical Research Council of Australia.

Published, JBC Papers in Press, October 18, 2002, DOI 10.1074/jbc.M206687200

	ABBREVIATIONS

The abbreviations used are: S1P, sphingosine 1-phosphate; hSK1, human sphingosine kinase 1; FSBA, 5'-p-fluorosulfonylbenzoyladenosine; 8-N₃ATP, 8-azidoadenosine 5'-triphosphate; Me₂SO, dimethyl sulfoxide; CaM, calmodulin; DGK, diacylglycerol kinase; PFK, 6-phosphofructokinase; MOI, multiplicity of infection.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES