Molecular cloning and characterization of OsCDase, a ceramidase enzyme from rice

Authors

  • Mickael O. Pata,

    1. School of Biology and Environmental Science, University College Dublin, Belfield, Dublin 4, Ireland
    Search for more papers by this author
  • Bill X. Wu,

    1. Department of Biochemistry and Molecular Biology, Medical University of South Carolina, 173 Ashley Ave, Charleston, SC 29425, USA
    Search for more papers by this author
  • Jacek Bielawski,

    1. Department of Biochemistry and Molecular Biology, Medical University of South Carolina, 173 Ashley Ave, Charleston, SC 29425, USA
    Search for more papers by this author
  • Tou Cheu Xiong,

    1. School of Biology and Environmental Science, University College Dublin, Belfield, Dublin 4, Ireland
    Search for more papers by this author
    • Present address: Biochimie et Physiologie Moléculaire des Plantes-Institut de Biologie Intégrative des Plantes, UMR 0386 INRA/UMR 5004 CNRS/Montpellier SupAgro/Université Montpellier 2, Bat 7, 2 place Pierre Viala, F-34060 Montpellier Cedex 1, France.

  • Yusuf A. Hannun,

    1. Department of Biochemistry and Molecular Biology, Medical University of South Carolina, 173 Ashley Ave, Charleston, SC 29425, USA
    Search for more papers by this author
  • Carl K.-Y. Ng

    Corresponding author
    1. School of Biology and Environmental Science, University College Dublin, Belfield, Dublin 4, Ireland
      *(fax +353 1 716 1153; e-mail carl.ng@ucd.ie).
    Search for more papers by this author

*(fax +353 1 716 1153; e-mail carl.ng@ucd.ie).

Summary

Sphingolipids are a structurally diverse group of molecules based on long-chain sphingoid bases that are found in animal, fungal and plant cells. In contrast to the situation in animals and yeast, much less is known about the spectrum of sphingolipid species in plants and the roles they play in mediating cellular processes. Here, we report the cloning and characterization of a plant ceramidase from rice (Oryza sativa spp. Japonica cv. Nipponbare). Sequence analysis suggests that the rice ceramidase (OsCDase) is similar to mammalian neutral ceramidases. We demonstrate that OsCDase is a bona fide ceramidase by heterologous expression in the yeast double knockout mutant Δypc1Δydc1 that lacks the yeast ceramidases YPC1p and YDC1p. Biochemical characterization of OsCDase showed that it exhibited classical Michaelis–Menten kinetics, with optimum activity between pH 5.7 and 6.0. OsCDase activity was enhanced in the presence of Ca2+, Mg2+, Mn2+ and Zn2+, but inhibited in the presence of Fe2+. OsCDase appears to use ceramide instead of phytoceramide as a substrate. Subcellular localization showed that OsCDase is localized to the endoplasmic reticulum and Golgi, suggesting that these organelles are sites of ceramide metabolism in plants.

Introduction

Sphingolipids are ubiquitous components of eukaryotic cells, where they form key structural components of membranes, and function as important mediators of physiological responses from apoptosis to cell survival and proliferation (Dickson and Lester, 2002; Futerman and Hannun, 2004; Hannun and Obeid, 2002, 2008; Pyne and Pyne, 2000; Spiegel and Milstien, 2003). Sphingolipid metabolism has been intensely investigated in animal cells and yeast, and much is known about the various metabolic intermediates and the enzymes responsible for catalyzing their formation (Dickson and Lester, 2002; Futerman and Hannun, 2004; Hannun and Obeid, 2002, 2008; Pyne and Pyne, 2000; Spiegel and Milstien, 2003). In yeast, sphingolipid metabolites such as dihydrosphingosine and phytosphingosine have been shown to increase the ability of yeast to tolerate heat-stress conditions and grow under such conditions (Dickson et al., 1997; Jenkins et al., 1997; Mao et al., 1999). Sphingosine-1-phosphate has been shown to be important in animal cells in mediating cellular survival and proliferation, and ceramide has been implicated in mediating programmed cell death (Futerman and Hannun, 2004; Hannun and Obeid, 2002, 2008; Spiegel and Milstien, 2003). This has led to the suggestion that the dynamic balance between levels of sphingolipid metabolites within cells functions to regulate cell fate, although the situation is likely to be highly complex due to the inter-convertibility of sphingolipid metabolites (Futerman and Hannun, 2004; Hannun and Obeid, 2008). This is particularly true for ceramide and sphingosine/sphingosine-1-phosphate, because sphingosine is formed by the action of ceramidases, indicating that ceramide is a central mediator of sphingolipid-based signalling systems in cells (Hannun and Obeid, 2002, 2008).

In contrast to the situation in animal cells and in yeast, much less is known about the roles of sphingolipid metabolites in regulating cellular processes in plants. Plant membranes are made up of significant amounts of sphingolipids, and it has been estimated that 7–26% of plant membrane lipids are sphingolipids (Lynch, 1993; Schmid and Ohlrogge, 1996). The first information on plant sphingolipids was mainly obtained from analyzing the lipid composition of seeds (Carter et al., 1958; Ito et al., 1985). There are indications that sphingolipids may be important cellular mediators in plants. It was reported that glucosylceramides (cerebrosides A and C) from wheat grain could induce fruiting in Schizophyllum commune, a fungus that causes wood decay (Kawai, 1989; Kawai et al., 1986), and the same glucosylceramides from the pathogenic rice fungus Magnaporthe grisea are elicitors of hypersensitive cell death and phytolexin accumulation in rice plants (Koga et al., 1998). More recently, it was reported that sphingosine-1-phosphate is a calcium-mobilizing messenger that is active in the responses of stomatal guard cells to the drought hormone abscisic acid (Ng et al., 2001), and that the effects of sphingosine-1-phosphate are dependent on heterotrimeric G-proteins (Coursol et al., 2003). Progress in the understanding of sphingolipid metabolism in plants has been aided by the characterization of Arabidopsis mutants affected in sphingolipid metabolism (Chen et al., 2006; Dietrich et al., 2008; Shi et al., 2007; Tsegaye et al., 2007), and the development of mass spectrometric methods for the analysis of plant sphingolipids (Markham et al., 2006; Markham and Jaworski, 2007; Teng et al., 2008). Mutant analyses in Arabidopsis have revealed important roles for sphingolipid metabolites in mediating sensitivity towards the mycotoxin fumonisin B1 (Tsegaye et al., 2007), and sphingolipid metabolites have been shown to be important regulators of gametophytic and sporophytic development (Dietrich et al., 2008; Imamura et al., 2007; Teng et al., 2008), thereby affecting fertility.

Ceramide has been shown to induce programmed cell death in Arabidopsis cells (Liang et al., 2003; Townley et al., 2005), suggesting that ceramide may be an important regulator of plant development. In order to further elucidate the role of ceramide in plants, it is important to understand the way in which ceramide is metabolized in plant cells. Ceramidases are enzymes responsible for metabolizing ceramide to sphingosine by cleaving the N-acyl linkage between the sphingoid base and the fatty acid. There are three types of ceramidases, and they have been classified as acid, neutral and alkaline according to their pH optima for activity (Mao and Obeid, 2002).

Here we report the cloning of a plant ceramidase from rice. We demonstrate that OsCDase is a bona fide ceramidase that appears to use ceramide instead of phytoceramide as a substrate. We also used a lipidomic approach to show that OsCDase has reverse ceramidase activity, resulting in the formation of phytoceramides with very long chain fatty acids. Subcellular localization showed that the DsRed2–OsCDase protein fusion is targeted to the endoplasmic reticulum and Golgi, suggesting that these organelles are sites of sphingolipid metabolism in plants.

Results

Cloning, phylogenetic and expression analysis of OsCDase

We used blast to search the rice genome database for sequences bearing similarities to the human neutral ceramidase gene ASAH2 (GenBank accession number NP_063946) and identified a single-copy gene on chromosome 1 designated Os01g43520. The predicted amino acid sequence of the putative rice ceramidase (OsCDase) has 42% sequence identity and 59% sequence similarity to ASAH2, suggesting that OsCDase may be a neutral ceramidase. We then used RT-PCR and 5′- and 3′-RACE PCR to obtain the full-length transcript for OsCDase. OsCDase is made up of 10 exons (Figure 1a). Interestingly, the 5′ UTR of OsCDase contains a 777 bp intron, although the Ensembl Exon report (http://www.gramene.org) predicted an intron of 255 bp. In silico analysis of the 5′-UTR of OsCDase using RegRNA (regulatory RNA motifs and elements finder; http://regrna.mbc.nctu.edu.tw/index.html) showed the presence of a 90 bp internal ribosomal entry site (IRES) element at position −32 to −122. Phylogenetic analysis of neutral ceramidases from various organisms indicates that OsCDase clusters with the putative ceramidases from Arabidopsis thaliana and the neutral ceramidase homologue of Dictyostelium discoideum (Figure 1b). Clustal W alignment of the amino acid sequence of OsCDase with ceramidases from various organisms (Figure S1) showed the presence of the highly conserved hexapeptide sequence GDVSPN within the larger conserved amidase domain NXGDVSPNXXC (Figure 1c) that is important for ceramidase activity (Galadari et al., 2006). Expression analysis indicated that steady-state levels of OsCDase transcripts are found throughout the seedling, with a higher level of expression in roots compared with shoots (Figure 1d).

Figure 1.

 Cloning, sequence analysis and expression of rice ceramidase (OsCDase) in planta.
(a) Schematic representation of the sequence organization of OsCDase. The sizes of the 5′-UTR (white), exons (black) and 3′-UTR (grey) are indicated by numbers below the boxes. The sizes of introns are shown above in italics.
(b) Phylogenetic tree of neutral ceramidases from various organisms (Os, Oryza sativa; At, Arabidopsis thaliana; Dd, Dictyostelium discoideum; Rn, Rattus norvegicus; Mm, Mus musculus; Cf, Canis familiaris; Hs, Homo sapiens; Gg, Gallus gallus; Dp, Drosophila pseudoobscura; Dm, Drosophila melanogaster) generated using T-Rex (http://www.trex.uqam.ca/).
(c) Clustal W alignment of the neutral ceramidases from various organisms showing the highly conserved hexapeptide sequence GDVSPN within the larger conserved amidase domain NXGDVSPNXXC that is important for ceramidase activity.
(d) RT-PCR expression analysis of OsCDase in shoots and roots of 2-week-old seedlings. OsEF1α (a housekeeping gene) was used as a control.

Biochemical characterization of OsCDase

We cloned the full-length OsCDase coding sequence into the yeast expression vector pYES2/CT under the control of the Gal1 promoter, and the resulting plasmid was transformed into the yeast double knockout mutant Δypc1Δydc1, which lacks the yeast ceramidases YPC1p and YDC1p (Mao et al., 2000). This is important as it allowed us to express recombinant OsCDase and minimize contamination from endogenous yeast ceramidase activity that may interfere with subsequent activity characterization of OsCDase. Expression of recombinant OsCDase was induced by culturing the transformed yeast in medium containing galactose, and cells were harvested at various time points to determine the optimal time for induction of OsCDase expression. Homogenates were obtained from harvested cells and assayed for ceramidase activity using C12-NBD-(N-[12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-D-erythro-sphingosine) ceramide as a substrate (Figure 2a). Ceramidase activity was detectable 2 h after induction with galactose, peaking at 10–12 h, and decreasing thereafter. We could not detect any ceramidase activity after induction with galactose in yeast containing the vector alone (Figure 2a). The amount of recombinant OsCDase was determined by Western blotting using anti-His antibody, and the amount of recombinant OsCDase peaked at about 12 h (Figure 2b), in agreement with optimal activity at 12 h post-induction (Figure 2a). On the basis of these data, we selected 12 h as the duration for induction of expression of OsCDase for subsequent biochemical characterization. Importantly, these results demonstrate that OsCDase has robust ceramidase activity.

Figure 2.

 Time course of induction of rice ceramidase (OsCDase) expression in the yeast double-knockout strain Δypc1Δydc1.
OsCDase was cloned into the vector pYES2/CT, and the construct was transformed into the yeast double-knockout strain Δypc1Δydc1, which lacks yeast ceramidases.
(a) Expression of OsCDase was induced by growth in galactose medium, and the culture was terminated at various time points and whole-cell lysates were assayed for ceramidase activity using d-erythro-C12-NBD-ceramide as substrate. Values represent means from two independent experiments.
(b) Western blot of whole-cell lysates (10 μg protein per lane) using anti-His antibody at various time points after growth in galactose medium.

We next examined the biochemical properties of OsCDase and determined the effects of various amounts of protein on ceramidase activity at various times (Figure 3a). The results indicate that the reaction is linear over a period of 120 min (Figure 3a). The reaction catalyzed by OsCDase showed classical Michaelis–Menten kinetics (Figure 3b), and showed an apparent Km of 39.34 μm (1.23 mol%) and an apparent Vmax of 0.079 pmol−1 h−1 μg−1 protein, as determined using the Lineweaver–Burk plot (Figure 3c). OsCDase activity was enhanced in the presence of Mg2+, Ca2+, Mn2+ (2.5–7.5 mm) and Zn2+ (2.5–5 mm) but inhibited in the presence of Fe2+ (Figure 3d). Because ceramide has been shown to induce micromolar increases in cytosolic free calcium in plants (Townley et al., 2005), we examined the effects of micromolar concentrations of Ca2+ on OsCDase activity in vitro. OsCDase activity was only weakly activated by free calcium ions in the micromolar range, but we observed a significant peak activation of OsCDase activity in the presence of 5 mm Ca2+ (t test, < 0.01; Figure S2). We observed OsCDase activity over a broad pH range, with an optimum pH of 5.7–6.0 (Figure 3e).

Figure 3.

 Biochemical characterization of the rice ceramidase (OsCDase) in the yeast double-knockout strain Δypc1Δydc1.
OsCDase was cloned into the vector pYES2/CT and the construct was transformed into the yeast double-knockout strain Δypc1Δydc1, which lacks yeast ceramidases. Expression of OsCDase was induced by growth in galactose medium for 12 h.
(a) The effect of various amounts of yeast protein on ceramidase activity at various assay times.
(b) Michaelis–Menten representation for OsCDase activity towards increasing concentrations of d-erythro-C12-NBD-ceramide.
(c) Lineweaver–Burk plot of OsCDase activity.
(d) Effects of increasing concentrations of cations on OsCDase activity.
(e) pH optimum determination. The buffers used are described in Experimental procedures.
For (b)–(e), 50 μg of total yeast protein was used for ceramidase activity assay and the reaction time was 1 h. Values are means ± SD of four replicates from two independent experiments.

We also examined the substrate specificity of OsCDase by substituting d-erythro-C12-NBD-ceramide with C12-NBD-phytoceramide. Phytoceramides are the predominant ceramide species in plants, including rice (Dunn et al., 2004; Lynch and Dunn, 2004; Ohnishi et al., 1985; Sperling and Heinz, 2003). Interestingly, our results suggest that phytoceramide is not a substrate for OsCDase (Figure 4a). We also measured endogenous levels of dihydroceramide (Figure 4b) and phytoceramide (Figure 4c) in the yeast double knockout mutant Δypc1Δydc1 expressing OsCDase. Our results suggest that dihydroceramide is unlikely to be a substrate for OsCDase (Figure 4b). Additionally, measurements of endogenous levels of phytoceramide suggests that OsCDase does not use phytoceramide as a substrate (Figure 4c), an observation consistent with the lack of activity when C12-NBD-phytoceramide was used as a substrate in vitro (Figure 4a). We also measured changes in the endogenous levels of α-OH-phytoceramide and showed that OsCDase does not use α-OH-phytoceramide as a substrate (data not shown). Interestingly, we observed that endogenous levels of phytoceramide with fatty acid chain lengths of C26 and C28 were elevated in Δypc1Δydc1 when expression of OsCDase was induced by growth in galactose. This suggests that OsCDase may have reverse ceramidase activity and catalyze the formation of phytoceramides with very long chain fatty acids (Figure 4c).

Figure 4.

 Substrate utilization by rice ceramidase (OsCDase).
OsCDase was cloned into the vector pYES2/CT and the construct was transformed into the yeast double-knockout strain Δypc1Δydc1, which lacks yeast ceramidases.
(a) Expression of OsCDase in the yeast double-knockout strain Δypc1Δydc1 was induced by growth in galactose medium for 12 h, and whole-cell lysates were assayed for ceramidase activity using d-erythro-C12-NBD-ceramide (C12-NBD-Cer) or C12-NBD-phytoceramide (C12-NBD-PhytoCer) as substrates. Wild-type yeast expressing YPC1p and YDC1p was used as a control to show that the yeast ceramidases use C12-NBD-phytoceramide as a substrate.
(b, c) Endogenous levels of (b) dihydroceramide and (c) phytoceramide following induction of OsCDase expression in the yeast double-knockout strain Δypc1Δydc1. Values are means and SD of four replicates from two independent experiments.

Subcellular localization of OsCDase

In silico analysis of the amino acid sequence of OsCDase using TMHMM (http://www.cbs.dtu.dk/services/TMHMM-2.0) indicated that OsCDase is a transmembrane protein. Additionally, we used SLP-Local (subcellular location predictor based on local features of amino acid sequence; http://sunflower.kuicr.kyoto-u.ac.jp/~smatsuda/slplocal.html) to analyse the OsCDase sequence, and OsCDase was predicted to be localized to the secretory pathway. In order to determine the subcellular localization of OsCDase in planta, we cloned the full-length OsCDase coding sequence into the expression vector pGDR (Goodin et al., 2000) containing the red fluorescent protein DsRed2 as the reporter, driven by the constitutive CaMV 35S promoter, for transient expression in onion epidermal cells following biolistic transformation. To determine the subcellular localization of DsRed2–OsCDase, we co-transformed onion epidermal cells with green fluorescent protein (GFP) targeted to the endoplasmic reticulum (ER-GFP) or Golgi (ST-GFP). We observed co-localization of DsRed2–OsCDase with ER–GFP (Figure 5a) and ST-GFP (Figure 5b), indicating that OsCDase is localized to the ER and Golgi, and that these organelles are sites of ceramide metabolism in planta.

Figure 5.

 Subcellular localization of OsCDase fused to the red fluorescent protein (DsRed2) following biolistic transformation of onion epidermal cells.
OsCDase was cloned into the vector pGDR and the plasmid was used for biolistic transformation of onion epidermal cells. ER-GFP (ER-targeted/retained green fluorescent protein) and ST-GFP (golgi-targeted green fluorescent protein) were used for co-localization studies. Co-localization is visualized in yellow (and indicated by arrowheads) in the ‘Merge’ panels. Localization of the nucleus is shown in the transmitted light micrograph. Images were acquired using confocal laser scanning microscopy and are representative of 30–40 transformed cells. Scale bar = 50 μm.

Discussion

We have cloned a ceramidase from rice (OsCDase), and sequence analysis showed that it has similarities to neutral ceramidases from a variety of organisms (Figure 1 and Figure S1). Interestingly, we observed the presence of an intron of 777 bp within the 5′-UTR of OsCDase. This is in contrast to the predicted intron within the 5′-UTR of 255 bp. Introns in plant genes are characterized by a strong nucleotide bias towards T proximal to the AG intron acceptor site, and there is an A/T bias throughout the intron relative to the adjacent exon. It has been proposed that such nucleotide biases are required for efficient intron recognition and splicing by the spliceosome (Lorkovic et al., 2000). However, the situation is less clear for introns within the 5′-UTR of plant genes as there appears to be no nucleotide bias to distinguish introns from exon sequences (Chung et al., 2006). The lack of a nucleotide bias for introns in 5′-UTRs appears to be the case in human genes also (Eden and Brunak, 2004). This may account for the discrepancy between the predicted size of the intron and the results we obtained using 5′-RACE PCR. The significance of the long intron within the 5′-UTR of OsCDase is unclear, although evidence to date suggests that the presence of 5′-UTR introns may enhance gene expression and that the length of the intron may also influence the level of gene expression in plants (Chung et al., 2006; Rose, 2004; Rose and Beliakoff, 2000).

In silico analysis of the 5′-UTR of OsCDase showed the presence of a 90 bp IRES element (position −32 to −122). This is interesting as IRES elements have been found in the 5′-UTRs of mRNAs of proteins involved in cellular proliferation and apoptosis (Holcik et al., 2000). IRES elements were first discovered in piconavirus mRNAs, where they function to initiate translation of uncapped viral mRNAs (Pelletier and Sonenberg, 1998). It has been proposed that IRES-dependent (and m7G cap-independent) translation of mRNA may be an important regulatory point for eukaryotic cells, allowing fine tuning of their responses to stress through IRES-dependent translation of survival or pro-apoptotic factors (Holcik et al., 2000). Given the observations that ceramide can initiate programmed cell death in plant cells (Liang et al., 2003; Townley et al., 2005), the identification of an IRES element in the 5′-UTR of OsCDase is interesting, and suggests that OsCDase may be subject to translational regulation as part of the programmed cell death machinery in plant cells.

In order to determine the biochemical characteristics of OsCDase, we expressed recombinant OsCDase in the yeast double knockout mutant Δypc1Δydc1, which lacks the yeast ceramidases YPC1p and YDC1p (Mao et al., 2000). Our results showed that OsCDase is indeed a bona fide ceramidase and shows typical Michealis–Menten kinetics. OsCDase activity was enhanced in the presence of Ca2+, Mg2+, Mn2+ and Zn2+, but inhibited in the presence of Fe2+ (Figure 3d). We showed that OsCDase activity was significantly activated in the presence of 5 mm Ca2+ (Figure S2), whereas Lynch (2000) reported a 100% increase in ceramidase activity in the presence of 1 mm Ca2+. We also tested the effects of micromolar concentrations of Ca2+ on OsCDase activity and showed that OsCDase activity was only weakly activated by micromolar concentrations of Ca2+ (Figure S2). It is unlikely that the activity of the ER/Golgi-localized OsCDase is influenced by changes in cytosolic-free calcium concentrations, and our observation of significant activation of OsCDase activity in the presence of 5 mm Ca2+ is consistent with the ER being an important intracellular calcium store. The divalent ion-dependent enhancement of ceramidase activity is in agreement with studies showing that Ca2+, Mg2+ and Mn2+ enhance the activity of the human neutral ceramidase (Galadari et al., 2006). Interestingly, while Zn2+ (2.5–5 mm) appeared to enhance the activity of OsCDase, the same concentrations inhibited the activity of the human neutral ceramidase (Galadari et al., 2006). Additionally, the activity of the Dictyostelium discoideum neutral ceramidase homologue did not appear to be affected by the presence of Ca2+, Mg2+, Mn2+ or Zn2+ (Monjusho et al., 2003). The observation that OsCDase has a pH optimum of 5.7–6.0 is interesting as it bears sequence similarities to neutral ceramidases. The observation of an acidic pH dependency of enzymes of the neutral ceramidase family has been reported in Dictyostelium discoideum, in which it was demonstrated that the neutral ceramidase homologue had a pH optimum of 3 (Monjusho et al., 2003). Lynch (2000) also reported that ceramidase activity from plant membrane fractions showed an acidic pH dependency of between 5.2 and 5.6. This is in agreement with our observation that OsCDase showed an acidic pH optimum of between pH 5.7 and 6.0. The slight acidic pH dependency of OsCDase may explain the observation that it clusters phylogenetically with the ceramidase from Dictyostelium discoideum (Figure 1b). Further work on putative members of the neutral ceramidase family from other plants, for example the three candidates from Arabidopsis, will provide greater insights into the acidic pH dependency of the neutral ceramidase family.

Analysis of substrate utilization by OsCDase revealed a substrate preference for ceramide but not phytoceramide (Figure 4a). This is interesting as phytoceramide is a major ceramide species in fungi and plants (Dunn et al., 2004; Lynch and Dunn, 2004; Ohnishi et al., 1985; Sperling and Heinz, 2003). The substrate preference of sphingolipid-metabolizing enzymes from plants for metabolites of low abundance is not without precedence. For example, Coursol et al. (2003, 2005) showed that sphingosine kinase from Arabidopsis has a substrate preference for sphingosine, which is of low abundance compared to dihydrosphingosine and phytosphingosine, both of which are predominant sphingolipid metabolites in plants. Analysis of yeast ceramide species following expression of OsCDase in the yeast double knockout mutant Δypc1Δydc1 also showed that both dihydroceramide and phytoceramide are unlikely to be substrates for OsCDase (Figure 4b,c). Reverse ceramidase activity has been reported in members of the neutral ceramidase family (El Bawab et al., 2001; Kita et al., 2000; Okino et al., 1998; Wu et al., 2007). Analyses of yeast sphingolipids following induction of OsCDase expression in the double knockout mutant Δypc1Δydc1 showed elevated levels of phytoceramide with fatty acid chain lengths of C26 and C28 (Figure 4c). This observation suggests that OsCDase may show reverse ceramidase activity, leading to elevated levels of these two phytoceramide species.

Neutral ceramidases have been shown to be localized to the plasma membrane, endosome-like organelle, mitochondria and endoplasmic reticulum/Golgi compartments (El Bawab et al., 2000; Mitsutake et al., 2001; Yoshimura et al., 2004). Additionally, neutral ceramidases may also be secreted (Romiti et al., 2000). We showed that OsCDase is localized to the endoplasmic reticulum and Golgi in planta, suggesting that the ER/Golgi compartments are sites of ceramide metabolism in plants. This observation is consistent with in silico prediction of OsCDase as a transmembrane protein localized to the secretory pathway.

In conclusion, we have cloned a bona fide plant ceramidase from rice and showed that it displays classical Michaelis–Menten kinetics with an optimum activity ranging from pH 5.7 to 6.0. OsCDase activity was enhanced in the presence of Ca2+, Mg2+, Mn2+ and Zn2+, but inhibited in the presence of Fe2+. Expression analysis indicated that steady-state levels of OsCDase transcripts are found throughout the seedling, with a higher level of expression in roots compared to shoots. Subcellular localization showed that OsCDase is localized to the ER/Golgi compartments, suggesting that these organelles are sites of sphingolipid metabolism in planta. Given the observations that ceramide is an important sphingolipid metabolite involved in the regulation of programmed cell death in plants (Liang et al., 2003; Townley et al., 2005), our characterization of the rice ceramidase is an important step in understanding how ceramide is metabolized and provides a basis for further investigations into how sphingolipid metabolism is co-ordinated with developmental processes in plants.

Experimental procedures

Plant growth

Rice (Oryza sativa ssp japonica cv. Nipponbare) seeds were germinated in full-strength Murashige and Skoog medium (Sigma, http://www.sigmaaldrich.com) at pH 5.8 in a plant growth cabinet (Microclima 1750, Snijders, http://www.snijders-scientific.nl) under the following conditions: photosynthetic photon flux density, 1000 μmol m−2 sec−1, 12 h light/12 h darkness, with a day temperature of 28°C and a night temperature of 24°C under 85% relative humidity.

cDNA cloning and determination of 5′- and 3′-UTRs

Total RNA was isolated from 2-week-old seedlings using the RNeasy® Mini Kit (Qiagen, http://www.qiagen.com) according to the manufacturer’s instructions. Total RNA (500 ng) that had been DNase-treated (Turbo DNA-free™ kit, Ambion, http://www.ambion.com) was used for cDNA synthesis using the Accuscript™ high-fidelity RT-PCR system (Stratagene, http://www.stratagene.com). Aliquots (3 μl) of the product were used for PCR amplification using PfuUltra high-fidelity DNA polymerase (Stratagene) and the forward primer 5′-ATGGAGGCTTCATCTTGGTTGTG-3′ and reverse primer 5′-ACGCACCGCGAAAGCACGAGA-3′.

Cloning of the 5′- and 3′-UTRs was achieved using a SMART-RACE cDNA amplification kit (Clontech, http://www.clontech.com) according to manufacturer’s instructions with the following gene-specific primers: for 5′-RACE, 5′-GAAGCTCCCCATACCGACCAG-3′ and for 3′-RACE, 5′-CAAGGCCCGCAAGATTGAGTTC-3′.

The full-length coding sequence and 5′- and 3′-UTRs were cloned into the pGEM T-Easy vector (Promega, http://www.promega.com/) before sequencing (Eurofins MWG GmbH, http://www.the-mwg.com).

Expression of OsCDase in yeast

The full-length OsCDase coding sequence was subcloned into the yeast expression vector pYES2/CT (Invitrogen, http://www.invitrogen.com) and transformed into the yeast double knockout mutant Δypc1Δydc1, which lacks the yeast ceramidases YPC1p and YDC1p (Mao et al., 2000), using the lithium acetate method as previously described (Gietz and Woods, 2002). Transformed yeast cultures were grown in synthetic minimal medium enriched in dextrose and complemented by a selective uracil-deficient drop-out supplement (Clontech), and cell lysates were isolated as previously described (Galadari et al., 2006).

OsCDase enzyme assays and biochemical characterization

Enzyme activity was measured using d-erythro-C12-NBD-ceramide (Avanti Polar Lipids, http://www.avantilipids.com) or C12-NBD-phytoceramide (Lipidomics Core Facility, Medical University of South Carolina, Charleston, SC) at a final concentration of 100 μm (3.13 mol%) in a 100 mm phosphate buffer, pH 5.7, containing 0.2% Triton X-100 and 5 mm MgCl2 in a total volume of 100 μl. To determine the pH optima, the substrate was dissolved in the following buffers: 100 mm acetate buffer, pH 3.91–5.33; 100 mm phosphate buffer, pH 5.04–6.66; 100 mm Tris buffer, pH 7.1–8.24. The enzymatic reaction was continued for 1 h at 37°C, and the reaction was terminated by addition of chloroform/methanol (1:1, v/v). The mixture was left at room temperature for 5 min, after which it was centrifuged for 5 min at 5500 rpm (Accuspin microR, Fisher, http://www.fishersci.com), and the organic phase was removed and dried using a SpeedVac evaporator (Savant, http://www.gmi-inc.com). Lipids were dissolved in 15 μl of chloroform/methanol (1:1, v/v), and spotted on to TLC plates (Partisil LK6D; Whatman, http://www.whatman.com). The C12-NBD-fatty acid products were separated from the substrate by developing the plate in a solvent system comprising chloroform/methanol/25% ammonium hydroxide (60:20:0.5, by volume). The TLC plate was scanned using a phosphorimager (Storm 860; GE Healthcare, http://www.gehealthcare.com) in blue fluorescence mode. The NBD-fatty acid products were identified by comparison with C12-NBD-fatty acid standards, and were quantified using ImageQuant™ software (GE Healthcare, http://www.gehealthcare.com). The specific activity was determined using standard curves with known concentrations of d-erythro-C12-NBD-ceramide.

Analysis of ceramides by mass spectrometry

Accumulated ceramides were analyzed by MS using reverse-phase HPLC coupled to an electrospray-triple quadrupole mass spectrometer, operating in positive ionization, multiple reaction monitoring (MRM) mode. Mass separations were performed using a ThermoFinnigan TSQ 7000 mass spectrometer (Thermo Scientific, http://www.thermo.com) as previously described (Bielawski et al., 2006).

Subcellular localization of OsCDase

The full-length OsCDase coding sequence was cloned downstream of the coding sequence for the red fluorescent protein DsRed2 under the control of the constitutive 35S CaMV promoter in the vector pGDR (kindly provided by Dr Michael Goodin, University of Kentucky, Lexington, KY; Goodin et al., 2000). The vector was sequenced to ensure that OsCDase was cloned in-frame with the DsRed2 reporter gene. Biolistic transformation of onion (Allium cepa) epidermal cells was performed using 1 μg of plasmid DNA precipitated onto 100 μg of 1 μm gold particles using CaCl2 and spermidine. DNA-coated gold particles were washed with ethanol and resuspended in ethanol for transformation of onion epidermal cells using a biolistic particle delivery system (Bio-Rad PDS-1000/He, http://www.bio-rad.com) with a pressure of 7.58 MPa. After overnight incubation, localization was observed using confocal laser scanning microscopy with a Leica TCS-SP2 confocal microscope (Leica, http://www.leica-microsystems.com). Transmitted light micrographs were obtained using a transmitted light detector. Co-localization was performed with green fluorescent protein targeted to the ER (plasmid pGFP:C-domain, kindly provided by Dr Wendy Boss, North Carolina State University, Raleigh, NC) (Persson et al., 2002) or Golgi (plasmid pVKH18-EN6::ST-GFP, kindly provided by Dr Chris Hawes, Oxford Brookes University, Oxford, UK; Saint-Jore et al., 2002).

Acknowledgements

This work is funded by a Science Foundation Ireland grant BR/04/B0581 (to C.K.-Y.N.), a National Institutes of Health (NIH) grant CA87584 (to Y.A.H.), a Irish Research Council for Science, Engineering and Technology Embark Postdoctoral Fellowship (to T.C.X.), and part of the research was conducted in a facility constructed with support from NIH Grant C06 RR018823 from the Extramural Research Facilities Program of the National Center for Research Resources.

The GenBank accession number for the nucleotide sequence of OsCDase is EU422991.

Ancillary