Author for correspondence: Carl K.-Y. Ng Tel: +353 1 7162250 Email: email@example.com
•Sphingolipids are emerging as important mediators of cellular and developmental processes in plants, and advances in lipidomics have yielded a wealth of information on the composition of plant sphingolipidomes. Studies using Arabidopsis thaliana showed that the dihydroxy long-chain base (LCB) is desaturated at carbon position 8 (d18:1Δ8). This raised important questions on the role(s) of sphingosine (d18:1Δ4) and sphingosine-1-phosphate (d18:1Δ4-P) in plants, as these LCBs appear to be absent in A. thaliana.
•Here, we surveyed 21 species from various phylogenetic groups to ascertain the position of desaturation of the d18:1 LCB, in order to gain further insights into the prevalence of d18:1Δ4 and d18:1Δ8 in plants.
•Our results showed that d18:1Δ8 is common in gymnosperms, whereas d18:1Δ4 is widespread within nonseed land plants and the Poales, suggesting that d18:1Δ4 is evolutionarily more ancient than d18:1Δ8 in Viridiplantae. Additionally, phylogenetic analysis indicated that the sphingolipid Δ4-desaturases from Viridiplantae form a monophyletic group, with Angiosperm sequences falling into two distinct clades, the Eudicots and the Poales.
•We propose that efforts to elucidate the role(s) of d18:1Δ4 and d18:1Δ4-P should focus on genetically tractable Viridiplantae species where the d18:1 LCB is desaturated at carbon position 4.
In this study, we surveyed 21 species from various phylogenetic groups to ascertain the position of desaturation of the dihydroxy LCB (d18:1), in an effort to gain further insights into the prevalence of d18:1Δ4 and d18:1Δ8 in plants. Our results showed that d18:1Δ8 is common among the gymnosperms, whereas d18:1Δ4 may be widespread within nonseed land plants and the Poales, suggesting that d18:1Δ4 is evolutionarily more ancient than d18:1Δ8 in Viridiplantae. Additionally, phylogenetic analysis indicated that the sphingolipid Δ4-desaturase homologues from Viridiplantae form a monophyletic group, with angiosperm sequences falling into two distinct clades, the Eudicots and the Poales.
Materials and Methods
Plants and growth conditions
The various plants used for sphingolipid extraction are Physcomitrella patens, Scapania gracilis, Huperzia selago, Selaginella kraussiana, Selaginella uncinata, Osmunda regalis, Agathis regalis, Wollemia nobilis, Ginkgo biloba, Lepidozamia peroffskyana, Lepidozamia hopei, Laurus nobilis, Drimys winteri, Triticum aestivum, Lolium perenne, Hordeum vulgare, Zea mays, Malus malus, Quercus petraea, A. thaliana and Nicotiana tabacum. All plants were kindly provided by Dr Jennifer McElwain (University College Dublin, UCD) unless otherwise stated. Plants were grown in glasshouses under natural lighting, and watered daily at UCD Rosemount Environmental Research Facility unless otherwise stated. Maize (Zea mays inbred line A188) was grown in a glasshouse with appropriate daily watering (50% soil water content) as described by Virlouvet et al. (2011). The liverwort, Scapania gracilis, was collected from a north-facing stone wall in a pasture area near Faha, west of Brandon Bay, Dingle Peninsula, Ireland (Irish National Grid: Q5012) by Dr Joanne Denyer (Denyer Ecology, Ireland). S. kraussiana and S. uncinata (kindly provided by Prof. Alistair Hetherington, University of Bristol, UK), and N. tabacum (kindly provided by Dr Paul McCabe, UCD, Ireland) were grown in vitro on full-strength Murashige & Skoog medium (pH 5.8), supplemented with 1% sucrose and 0.6% plant cell-culture tested agar (Sigma) in Magenta GA-7 containers in a controlled growth room at 22°C and continuous light (blue and red LEDs at 20 μmol m−2 s−1; Philips Lighting Systems, Eindhoven, The Netherlands). P. patens ecotype ‘Gransden 2004’ (kindly provided by Dr Paul McCabe) gametophytes were grown on sterile Jiffy-7 peat pellets in Magenta GA-7 containers under controlled conditions: light intensity (50 μmol s−1 m−2), 16 h light : 8 h darkness, 75% relative humidity (RH), and 23°C in a growth chamber (Sanyo MLR-351 H; Versatile Environmental Test Chamber, Japan). A. thaliana Col-0 (kindly provided by the European Arabidopsis Stock Centre, Nottingham, UK) were grown under short-day conditions as described by Xiong et al. (2009).
Identification of position of desaturation of the predominant d18:1 LCB by Liquid Chromatography-Fluorecence Detection (LC-FLD)
Sphingolipids from leaves (and A. thaliana shoots) were extracted and hydrolysed as described by Markham et al. (2006) and Islam et al. (2011). All chemicals and solvents are from Sigma unless otherwise stated. Briefly, 50 mg of finely ground freeze-dried samples were spiked with 10 μl internal standards (250 μg ml−1 d16:1 and/or 200 μg ml−1 d20:1; Matreya Inc., Pleasant Gap, USA) before hydrolysing with 1 ml of dioxane and 1 ml of 10% (w/v) barium hydroxide octahydrate in water for 16 h at 110°C. At the end of hydrolysis, 2 ml of 2% (w/v) ammonium sulphate was added to precipitate barium ions and to reduce the occurrence of a flocculent precipitate during subsequent derivatizing. LCBs were extracted with 2 ml of diethylether and centrifuged to separate the aqueous and organic phases. The upper phase was removed to a second tube, dried under nitrogen, and derivatized with o-phthaldialdehyde as previously described (Merrill et al., 2000). The chromatographic separation of LCBs were carried out with a reversed-phase high-performance liquid chromatography (HPLC) system consisting of binary pump, autosampler, column oven (1200 RRLC; Agilent Technologies, Stuttgart, Germany) and fluorescence detector (FP-2020 Plus; Jasco, Tokyo, Japan). The analytical column was an Eclipse XBD-C18 column (4.6 mm ID × 250 mm, particle size 5 μm; Agilent Technologies) and the column oven and autosampler temperatures were maintained at 35 and 4°C, respectively. Elution was carried out at flow rate of 1.5 ml min−1 with 70% solvent A (5 mM potassium phosphate, pH 7), 30% solvent B (100% methanol) for 4 min, increasing to 85% solvent B for 10 min, maintained in this condition for up to 25 min before increasing to 100% solvent B for 30 min with a 5 min 100% solvent B wash before returning to initial conditions and re-equilibrating for 5 min. Separated derivatized LCBs were monitored using the fluorescence detector at excitation and emission wavelengths of 340 and 455 nm, respectively.
Database searches, sequence alignment and phylogenetic analysis
The identification of sphingolipid Δ4-desaturase homologues was performed using the megaBLAST and tBLASTn programs and the experimentally known A. thaliana sphingolipid Δ4-desaturase cDNA and amino acid sequences, respectively, as a query in the Nucleotide collection nr/nt (http://www.ncbi.nlm.nih.gov/nuccore), the recent release 5b.60 of the B73 maize genome assembly (http://maizesequence.org), Ensembl (http://www.ensembl.org/) and KEGG GENES (http://www.genome.jp/kegg/genes.html). In the case that multiple gene models were predicted for a single locus, preference was given to the model best supported by expressed sequence tag (EST) coverage and conservation with other sphingolipid Δ4-desaturase genes. The cDNA coding sequences were aligned using ClustalW with the BioEdit Sequence Alignment Editor v7.0.0 (Hall, 1999) and visually refined on the basis of the amino acid translated sequences. Bayesian phylogenetic analysis was carried out with MrBayes v3.1.1 (Ronquist & Huelsenbeck, 2003) using standard settings (aa = mixed) on eight conserved blocks (306-amino-acid alignment matrix) selected with Gblocks using a relaxed selection of blocks from amino acid-translated sequence alignments (Talavera & Castresana, 2007). Three chains (two heated) were run twice for 106 generations, with a sampling of the cold chain parameters every 100 generations and a burn-in of 5000 samples. Convergence was followed with potential scale reduction factor and average SD of split frequencies. A majority rule consensus tree was built using TreeView v1.6.6 (Page, 1996) with posterior probabilities of nodes above 0.85 × 100 indicated. Additionally, the online version of PhyML (http://www.atgc-montpellier.fr/phyml/; Guindon et al., 2010) was used to perform maximum likelihood reconstruction of phylogenies based on amino acid-translated sequences using the Jones–Taylor–Thornton substitution model. Five randomly BIONJ (Guindon et al., 2010) starting trees were used. The final topology was detected by a best of NNI and SPR (Guindon et al., 2010), and the branch support was evaluated by 500 bootstraps.
The position of desaturation of the d18:1 LCBs from various Viridiplantae species was determined by HPLC analyses of their o-phthaldialdehyde-derivatized LCBs in an effort to gain further insights into the prevalence of d18:1Δ4 and d18:1Δ8 in plants. Fig. 1(a) shows the HPLC separation of o-phthaldialdehyde-derivatized LCB standards, d16:1, t18:0, d18:1Δ4, d18:0 and d20:1. Fig. 1(b) shows the HPLC separation of o-phthaldialdehyde-derivatized LCBs from shoots of A. thaliana spiked with d16:1 and d18:1Δ4. The results are consistent with previous observations that d18:1Δ4 is absent in shoots of A. thaliana (Markham et al., 2006; Markham & Jaworski, 2007; Michaelson et al., 2009; Islam et al., 2011). Although d18:1Δ4 and d18:1Δ8 have the same molecular mass, the retention time of their o-phthaldialdehyde derivatives are sufficiently different to allow us to distinguish between them (Fig. 1b). Fig. 1(c,d) shows representative chromatograms from Agathis australis (Kauri) and maize, showing the desaturation of the d18:1 LCB at carbon positions 8 and 4, respectively.
Table 1 lists the various Viridiplantae species examined and the position of desaturation of the d18:1 LCB. The results showed that the d18:1 dihydroxy LCB in the nonseed land plants, moss (P. patens), liverwort (S. gracilis), clubmoss (H. selago), spikemosses (S. kraussiana, S. uncinata) and fern (O. regalis), is desaturated at carbon position 4 (Table 1). The widespread occurrence of d18:1 LCB desaturated at carbon position 4 among the nonseed land plants suggests that Δ4-desaturation may be evolutionarily more ancient than Δ8-desaturation.
Table 1. Position of desaturation of the predominant dihydroxy long-chain base (LCB) in various Viridiplantae species.
Analyses of the Poales, exemplified by wheat (T. aestivum), barley (H. vulgare), perennial ryegrass (L. perenne) and maize (Z. mays), also showed desaturation of the d18:1 dihydroxy LCB at carbon position 4 in these species (Table 1). Islam et al. (2011) also reported desaturation of the d18:1 dihydroxy LCB at carbon position 4 in the model temperate grass, Brachypodium distachyon Bd21, and the results from this study showing desaturation of the d18:1 dihydroxy LCB at carbon position 4 in other members of the Poales suggest the potential usefulness of B. distachyon Bd21 as the model species for elucidating the role(s) of d18:1Δ4 in the Poales.
Interestingly, we observed that the d18:1 dihydroxy LCB in the extant conifers, A. australis and W. nobilis, and the cycads, L. peroffskyana and L. hopei, are desaturated at carbon position 8 (Table 1), suggesting that d18:1Δ8 may be the predominant d18:1 dihydroxy LCB in the gymnosperms. However, we caution against drawing conclusions based on the analyses of four gymnosperms. We were unable to conclusively determine the position of desaturation of the d18:1 LCB from G. biloba as the amounts of d18:1 dihydroxy LCBs were too low to allow conclusive determination of the position of desaturation, although good recovery of the externally supplied d20:1 dihydroxy LCB standard was observed (data not shown). We propose that future studies should focus on the analyses of more species of monocots, conifers, cycads and Gnetales in order to gain further insights into prevalence of d18:1Δ8 dihydroxy LCBs in monocots and gymnosperms.
In addition to G. biloba, we were also unable to conclusively determine the position of desaturation of the d18:1 dihydroxy LCB from D. winteri and Q. petraea, although we were able to determine that the d18:1 dihydroxy LCB from apple is desaturated at carbon position 8 (Table 1). In the case of D. winteri, the extraction protocol used resulted in poor recovery of endogenous LCBs and externally supplied d20:1 dihydroxy LCB standard following extraction. Although good recovery of the externally supplied d20:1 dihydroxy LCB standard was observed in the case of Q. petraea, the amounts of d18:1 dihydroxy LCBs were too low to allow us to conclusively determine the position of desaturation.
As d18:1Δ4 formation is likely to be catalysed by a sphingolipid Δ4-desaturase, we have identified, on the basis of sequence identity to the experimentally known A. thaliana sphingolipid Δ4-desaturase (Michaelson et al., 2009), 22 putative sphingolipid Δ4-desaturases from Viridiplantae, including the published sphingolipid Δ4-desaturases from tomato (L. esculentum) and rice (Oryza sativa japonica) (Supporting Information, Table S1; Ternes et al., 2002; Hashimoto et al., 2008). Closer inspection of their amino acid sequences revealed the presence of a strongly conserved first His box motif HELSH in the N-terminal region in all sequences belonging to the Viridiplantae. All members of the sphingolipid Δ4-desaturase family possess the first His box motif HE(D)xS(T)H, which is known with two other conserved His motifs to be essential for function (Shanklin et al., 1994; Hashimoto et al., 2008). To assess the diversity of the sphingolipid Δ4-desaturase family beyond the three published plant species, a phylogenetic tree was constructed that included the published plant sphingolipid Δ4-desaturases and 80 related genes from representatives of different groups of Euglenozoae, Amoebozoae, fungi, animals, Rhodhophytae and Viridiplantae (Fig. 2). Maximum likelihood and Bayesian reconstruction methods were congruent and showed that the Viridiplantae sequences form a monophyletic group, with Angiosperm sequences falling into two distinct clades corresponding to the Poales and the Eudicots (Fig. 2). Interestingly, sphingolipid Δ4-desaturases from Poales are not predicted to have any transmembrane spanning domains (Table S1). This prediction is widespread within nonseed land plants (P. patens) and Eudicots (Populus trichocarpa; L. esculentum and Vitis vinifera #22), some of which show desaturation of the d18:1 dihydroxy LCB at carbon position 4 (Fig. 2 and Table S1). A larger sample of HPLC and sequence data from Viridiplantae might help to further elucidate the potential link between the absence of transmembrane domain and the presence of d18:1Δ4.
Collectively, our results suggest that d18:1Δ4 may be widespread within the Viridiplantae kingdom, and that efforts to elucidate the potential role(s) of d18:1Δ4 and d18:1Δ4-P should focus on genetically tractable species where the d18:1 dihydroxy LCB is desaturated at carbon position 4. It is envisaged that the cloning and characterization of the putative sphingolipid Δ4-desaturase genes from nonseed land plants and Poales are likely to yield insights into the potential role(s) of d18:1Δ4 and d18:1Δ4-P in plants.
This work is supported by Science Foundation Ireland (SFI) Research Frontiers Programme and Equipment Grants (06/SFI/RFP/GEN034, 06/SFI/RFP/GEN034ES, 08/SFI/ RFP/EOB1087) to C.K-Y.N, and by a grant from the ANR ‘Biotechnologies et bioressources’ program (AMAIZING) to S.C. We thank Brian Fagan, Bredagh Moran, Frances Downey and Suhas Shinde for assistance with plant growth, and Catherine Damerval for critical review of the phylogenetic analysis.