Phytosphingosine-phosphate is a signal for AtMPK6 activation and Arabidopsis response to chilling


  • Christelle Dutilleul,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Ghouziel Benhassaine-Kesri,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Chantal Demandre,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Nathalie Rézé,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Alban Launay,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Sandra Pelletier,

    1. UMR INRA 1165-CNRS 8114-UEVE, Unité de Recherche en Génomique Végétale (URGV), 2, rue Gaston Crémieux, CP5708, F-91057 Evry Cedex, France
    Search for more papers by this author
  • Jean-Pierre Renou,

    1. UMR INRA 1165-CNRS 8114-UEVE, Unité de Recherche en Génomique Végétale (URGV), 2, rue Gaston Crémieux, CP5708, F-91057 Evry Cedex, France
    Search for more papers by this author
  • Alain Zachowski,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Emmanuel Baudouin,

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author
  • Isabelle Guillas

    1. UPMC Univ Paris 06, UR 5, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France and CNRS, EAC 7180, Laboratoire de Physiologie Cellulaire et Moléculaire des Plantes, F-75252, Paris, France
    Search for more papers by this author

Author for correspondence:
Isabelle Guillas
Tel: +33(0) 1 44 27 59 18


  • Long-chain bases (LCBs) are pleiotropic sphingolipidic signals in eukaryotes. We investigated the source and function of phytosphingosine-1-phosphate (PHS-P), a phospho-LCB rapidly and transiently formed in Arabidopsis thaliana on chilling.
  • PHS-P was analysed by thin-layer chromatography following in vivo metabolic radiolabelling. Pharmacological and genetic approaches were used to identify the sphingosine kinase isoforms involved in cold-responsive PHS-P synthesis. Gene expression, mitogen-activated protein kinase activation and growth phenotypes of three LCB kinase mutants (lcbk1, sphk1 and lcbk2) were studied following cold exposure.
  • Chilling provoked the rapid and transient formation of PHS-P in Arabidopsis cultured cells and plantlets. Cold-evoked PHS-P synthesis was reduced by LCB kinase inhibitors and abolished in the LCB kinase lcbk2 mutant, but not in lcbk1 and sphk1 mutants. lcbk2 presented a constitutive AtMPK6 activation at 22°C. AtMPK6 activation was also triggered by PHS-P treatment independently of PHS/PHS-P balance. lcbk2 mutants grew comparably with wild-type plants at 22 and 4°C, but exhibited a higher root growth at 12°C, correlated with an altered expression of the cold-responsive DELLA gene RGL3.
  • Together, our data indicate a function for LCBK2 in planta. Furthermore, they connect PHS-P formation with plant response to cold, expanding the field of LCB signalling in plants.


Plant exposure to low temperatures above zero (chilling temperatures) triggers a complex metabolic and structural remodelling that restrains the plant growth and development rate under these nonfreezing temperatures (chilling), but also allows some species to survive under freezing temperatures (cold acclimation process). An important component of this remodelling is the regulation of hundreds of genes, such as COR (for COld-Responsive) and LTI (for Low-Temperature Induced) (reviewed in Ruelland et al., 2009). Recent reports have also indicated a key function for DELLA factors in growth restriction and developmental control under environmental adversity, for example low temperature (Achard et al., 2008b; Li et al., 2011).

Understanding the mechanisms linking cold perception to the set-up of cold adaptive responses has long been a challenging issue. It is now agreed that cold transduction relies on a complex signalling network that triggers and controls cellular modifications and metabolic rearrangement. Although the identity and function of the actors of this network are far from clear, signalling lipids have emerged as important transducers of the cold signal (Ruelland et al., 2002; Vergnolle et al., 2005; Williams et al., 2005). As observed for other stresses, the phospholipid derivative phosphatidic acid (PtdOH) acts as a mediator of the plant response to cold. A rapid increase in PtdOH production has been observed on chilling in Arabidopsis thaliana (Ruelland et al., 2002). In this context, the origin of PtdOH is complex, as both phospholipase C (PLC), which generates inositol triphosphate and diacylglycerol, subsequently phosphorylated into PtdOH by diacylglycerol kinases (DGKs), and phospholipase D (PLD) activities are involved. PLD- and PLC/DGK-evoked PtdOH participates in distinct signalling cascades and regulates different, although overlapping, cold-responsive gene clusters (Vergnolle et al., 2005). In addition, PtdOH precursors and synthesis intermediates, that is phosphatidylinositol phosphate, phosphatidylinositol bisphosphate and inositol triphosphate, may also participate in cold signalling, which suggests that an intricate network of phospholipid-derived signals operates during the early cold stress response (Williams et al., 2005).

Adding to the complexity of lipid signalling associated with cold transduction, we recently observed that specific modifications of phosphosphingolipid species formation occurred in chilled culture cells (Cantrel et al., 2011). Sphingolipids are major components of eukaryotic membranes and are synthesized in a multiple step process, starting with the condensation of serine with palmitoyl-CoA, giving rise to the long-chain base (LCB) ketodihydrosphingosine. The different LCBs found in plants directly derive from ketodihydroxysphingosine by Δ4 desaturation, Δ8 desaturation or both, and/or C4 hydroxylation (Pata et al., 2010). The subsequent acylation of LCBs on their amide group by a long-chain fatty acid gives rise to ceramides which are metabolized into complex membrane sphingolipids through the addition of a specific polar head group. Interestingly, LCBs not only act as primary intermediates of membrane sphingolipid synthesis, but also as potent bioactive signals in eukaryotic cells. They participate in the control of fundamental cellular processes, such as cell death, proliferation and differentiation, in mammals (Chalfant & Spiegel, 2005; Zheng et al., 2006; Fyrst & Saba, 2010). In this context, a key issue of LCB bioactivity is their reversible phosphorylation/dephosphorylation by specific LCB kinases and phosphatases. Indeed, the relative ratio between phosphorylated vs unphosphorylated LCBs determines the cell fate and orientates the cell destiny towards proliferation or death (Cuvillier, 2002; Hait et al., 2006). More recently, a function for LCBs in cell signalling has been evidenced in plants. Pioneering studies on the control of stomatal movements indicated that phosphorylated LCBs (LCB-P) participate in the regulation of guard cell turgor by abscisic acid (ABA) (Ng et al., 2001; Coursol et al., 2003). Treatment with sphingosine-1-phosphate (SPH-P), the major bioactive LCB-P in mammalian cells, or 4-hydroxy-sphinganine-phosphate (PHS-P) triggered stomatal closure; conversely, the competitive inhibition of LCB kinase activity led to a dramatic decrease in stomatal closure induced by ABA treatment (Ng et al., 2001; Coursol et al., 2003). Coursol et al. (2003) further showed that ABA itself induced the rapid and transient stimulation of LCB kinase activity. Finally, these studies provided some clues on the possible roles for LCB-P in signal transduction. On the one hand, exogenous treatment with SPH-P led to variations in [Ca2+]cyt (Ng et al., 2001). On the other, stomatal closure triggered by SPH-P and PHS-P required the functionality of the heterotrimeric G-protein α-subunit GPA1 (Coursol et al., 2003, 2005). Further investigations using genetic approaches identified several molecular actors involved in LCB/LCB-P metabolism during the plant response to ABA. Worrall et al. (2008) showed that the LCB kinase SPHK1 was required for appropriate stomatal responses to ABA, but also for the inhibition of seed germination by ABA. Moreover, the LCB-P phosphatase SPPASE and the LCB-P lyase DPL1, which degrades LCB-P into C16 fatty aldehydes and phosphoethanolamine, also participate in the regulation of ABA and drought responses (Nishikawa et al., 2008; Worrall et al., 2008). In addition to participating in the ABA response, LCBs have recently been shown to be involved in the mediation of cell death. Shi et al. (2007) reported that a range of LCBs are potent inducers of cell death in Arabidopsis, and that cell death can be counteracted by a co-treatment with the corresponding LCB-P. In tobacco cells, enhancing the endogenous LCB content also induced cell death by an apoptosis-like mechanism dependent on [Ca2+]cyt and [Ca2+]nuc variations (Lachaud et al., 2010). In addition to calcium, the mitogen-activated protein (MAP) kinase AtMPK6 has recently been shown to be a transducer of LCB-induced cell death (Saucedo-Garcia et al., 2011). Finally, endogenous variations in LCB content may participate in cell death induction, as suggested by the elevation of phytosphingosine (PHS) content observed in Arabidopsis leaves infected with a Pseudomonas syringae strain triggering the hypersensitive response (Peer et al., 2010).

In the present study, we report that cold exposure leads to a transient production of PHS-P in both Arabidopsis suspension cells and seedlings. The production of cold-induced PHS-P requires the presence of the putative LCB kinase LCBK2, but not of two other LCB kinase isoforms, that is SPHK1 and LCBK1. Moreover, lcbk2 mutants present an aberrant regulation of the cold-responsive MAP kinase AtMPK6, which is correlated with the antagonistic effect of PHS-P vs PHS on AtMPK6 regulation. Interestingly, although none of the lcbk mutants presented a distinctive phenotype when grown at 22 or 4°C, lcbk2 root growth was more weakly reduced at moderate low temperature (12°C) than that of any other plant line tested, suggesting that modifications of PHS-P formation modify the plant response to intermediate cold.

Materials and Methods

Cell and plant cultures, and cold treatment

Experiments were performed using Arabidopsis thaliana (L.) Heynh. ecotype Columbia (Col-0) as cultured cells or 14-d-old seedlings. Arabidopsis seeds were sterilized, stratified for 2 d at 4°C and sown on basic half-strength Murashige & Skoog (MS) medium (M0221; Sigma-Aldrich, Lyon, France) (10 g l−1 sucrose, pH 5.7, 8 g l−1 agar). Plates were then placed in a growth chamber under continuous illumination (230 μmol m−2 s−1) at 22°C and 56% humidity. Cell cultures were grown as described previously (Ruelland et al., 2002).

For the experiments, 7 ml of cell suspension (5 d after subculture, representing c. 1 g FW) were transferred to 50-ml flasks and agitated overnight at 22°C on an orbital shaker under continuous illumination. Cold shock was applied by transferring the flasks to a cooled water bath (4°C) under agitation. Alternatively, 14-d-old plantlets (c. 50 mg FW) were transferred to 50-ml flasks containing 3 ml of half-strength MS medium and agitated overnight at 22°C on an orbital shaker under continuous illumination. Cold shock was applied as for cultured cells.


Protein A-Sepharose, o-phthaldialdehyde (OPA) and antibodies (anti-AtMPK4 (A6979) and anti-AtMPK6 (A7104)) were purchased from Sigma-Aldrich. Myelin basic protein (MBP), N,N-dimethylsphingosine (DMS), dihydrosphingosine (DHS), dihydrosphingosine phosphate (DHS-P), PHS, SPH-P and threo-dihydrosphingosine (TSP) were purchased from Enzo Life Science (Covalab, Villeurbanne, France). d-Erythro-sphingosine (SPH) was purchased from Matreya (Pleasant Gap, PA, USA). PHS-P was purchased from Avanti Polar Lipids (Alabaster, AL, USA).

Phosphorylated sphingolipid analysis

Metabolic labelling of cultured cells with [33P]-orthophosphate was performed as described in Ruelland et al. (2002), except that [33P]-orthophosphate was added to cultures 15 min before cold treatment. For pharmacological treatments, inhibitors and commercial LCBs were vacuum dried and dissolved in an appropriate volume of dimethylsulfoxide (DMSO) in order not to exceed a final concentration of 0.1% DMSO in the cell culture. They were added 30 min before cold treatment, DMSO being added in the control experiments. Sphingolipids were extracted as follows. At the indicated time, 0.7 ml of 5% Trichloroacetic acid was added to cell suspensions to stop labelling. Cells were washed twice with cold water and harvested by centrifugation. Lipids were subsequently extracted from the cell pellets by incubation at 65°C for 15 min, successively in methanol, choroform/methanol (1 : 1, v/v) and chloroform. Pooled extracts were dried under nitrogen and resuspended in methanol. To hydrolyse glycerophospholipids, total lipid extracts were incubated for 1 h at 60°C in 3 M NH4OH/methanol (1 : 1, v/v) and sphingolipids were purified by phase extraction in acetic acid/water/chloroform (7 : 3 : 3, v/v). Nitrogen-dried sphingolipids were dissolved in methanol. The incorporated radioactivity was estimated by liquid scintillation counting. Sphingolipids were developed by thin-layer chromatography (TLC) on Silica 60 plates (Merck, Darmstadt, Germany) using chloroform/acetone/methanol/acetic acid/ water (10 : 4 : 3 : 2 : 1; v/v) as solvent system (SS1) (Coursol et al., 2005). Radiolabelled sphingolipids were revealed and quantified using a Storm PhosphorImager (Molecular Dynamics, Bondoufle, France).

For plantlet labelling, 53 MBq l−1 of [33P]-orthophosphate was added to each flask, 30 min before cold shock. At the indicated time, the plantlets were removed from the flask and immediately frozen in liquid nitrogen. Following grinding, sphingolipids were extracted and analysed as described above for cell labelling.

When needed, commercial LCB standards were co-separated with radiolabelled sphingolipids, revealed by spraying with 0.01% (w/v) primuline in 80% acetone and visualized under UV.

Purification of cold-evoked LCB-P

One litre of culture cells (c. 150 g FW) was divided into 7-ml fractions and submitted to radioactive labelling and cold shock as described above. After extraction, labelled sphingolipids were pooled and separated by preparative TLC developed in SS1. Radiolabelled sphingolipids from cells kept at 22°C were used as a control for all the purification steps. The cold-evoked LBC-P was scraped off, dissolved in methanol and successively developed in SS1 and then in chloroform/methanol/0.25% KCl (55/45/5, v/v). At each step, radiolabelled sphingolipids were revealed and identified using a Storm PhosphorImager (Molecular Dynamics). Finally, the cold-responsive radiolabelled LCB-P was eluted from silica with methanol and analysed by high-performance liquid chromatography (HPLC) after OPA derivatization (see the next paragraph).

HPLC analysis of free LCBs

Free LCB analysis was performed as described previously (Markham et al., 2006). For sample standardization, C20-4-SPH (d20:1) was added as an internal standard. Equivalent amounts of extracts were derivatized with OPA (Merrill et al., 1988). HPLC analyses were carried out using Beckman Coulter system Gold 128 series pumps by reverse phase HPLC on a 4.6 mm × 250 mm gemini-C18 column (Phenomenex, Le Pecq, France). Elution was performed at 0.7 ml min−1 with 30% solvent RA (5 mM potassium phosphate, pH 7) and 70% solvent RB (100% methanol) for 2 min, increasing to 80% solvent RB by 8 min, followed by an isocratic flow for 10 min, before increasing to 100% solvent RB by 15 min, with a 3-min 100% solvent RB wash, before returning to 70% solvent RB and re-equilibrating for 2 min. Fluorescence was excited at 340 nm and detected at 455 nm (RF-10A XL; Shimadzu, Champs sur Marne, France). Results were analysed and integrated using 32 Karat software (Beckmann Coulter, Inc., Fullerton, CA, USA).

MAP kinase assays

MAP kinase activity was assayed following immunoprecipitation of protein extracts using specific anti-AtMPK4 and anti-AtMPK6 antibodies (Sigma-Aldrich), followed by in vitro kinase assay as described by Teige et al. (2004). The phosphorylation of MBP was quantified by PhosphorImager analysis after sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE). For the determination of AtMPK4 and AtMPK6 abundance, the same amounts of proteins were separated on SDS-PAGE and transferred onto nitrocellulose membrane. Western blot analyses were performed using anti-AtMPK4 and anti-AtMPK6 antibodies according to the manufacturer’s instructions.

RNA isolation, semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) and quantitative RT-PCR analyses

Total RNA was extracted from nitrogen-frozen Arabidopsis cells or plantlets by grinding samples in 0.2 M Tris-HCl, pH 7.5, 0.5% (m/v) SDS, 0.25 M NaCl and 25 mM EDTA. After three steps of extraction with phenol/chloroform (1/1, v/v), total RNA was precipitated in 2 M lithium chloride, resuspended in water and precipitated once more; 2 μg of RNA were treated with DNase I (Sigma) and used for semi-quantitative RT-PCR as described in Vergnolle et al. (2005). Gene-specific primers are shown in Supporting Information Table S1. S19 was used as a control. Quantitative real-time RT-PCR was performed for five genes on samples treated at 22, 12 or 4°C. The primers for RT-PCR were selected with Primer3 (; optimal length, 21 nucleotides; optimal temperature, 60°C). The primer pairs were first tested on a dilution series of genomic DNA (5, 0.5, 0.05, 0.005 ng) to generate a standard curve and assess their PCR efficiency, which ranged between 90% and 99%. Reverse transcription (RT) was performed on 500 ng of total RNA with an oligodT primer (18-mer) and 200 U of Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) for 1 h at 42°C in 40 μl. The enzyme was then heat inactivated at 65°C and the samples were treated with RNase H. Quantitative PCRs were performed in 15 μl, with 0.1 μl of RT reaction, 900 nM final concentration of each primer pair and SYBRGreen PCR master mix 2 × (Eurogentec, Seraing, Belgium). Corresponding minus RT controls were performed with each primer pair. Conditions were as follows: 95°C, 10 min; 40 × (95°C, 15 s; 60°C, 1 min); and a dissociation step to discriminate primer dimers from the PCR product. All reactions were performed in RT duplicate with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Pleasanton, CA, USA), and data were analysed with the SDS software provided by the manufacturer. Three housekeeping genes were used to calculate an average normalization factor (AT5G60390, AT2G19740, AT5G09870) for each sample pair. The normalized discrete cosine transform (DCT) for each differentially expressed gene was calculated as follows: normalized DCT = raw DCT − normalization factor.

Statistical analysis

The experimental data reported represent at least three independent biological repeats. When appropriate, results are reported as the mean values ± standard errors. Mean comparisons were calculated by Student’s t-test with P values indicated in the legends.


Cold stress triggers the rapid and transient synthesis of PHS-P in cell cultures and plantlets

Using in vivo [33P]-orthophosphate labelling of Arabidopsis cultured cells, preliminary analysis suggested that a specific phosphosphingolipid that migrated close to commercial LCB-P standards was synthesized following cold exposure (Cantrel et al., 2011; Fig. 1a). To fully establish the identity of this phosphorylated species, we purified the cold-evoked phosphosphingolipid from the other radiolabelled sphingolipids by eluting the lipid of interest and developing it successively in three different solvent systems. The purified lipid species was derivatized with the NH2-specific reagent OPA. As shown in Fig. 1(b), a single peak corresponding to the purified species was detected that co-eluted with a commercial PHS-P standard. These data therefore provide evidence that PHS-P is formed on cold exposure.

Figure 1.

4-Hydroxy-sphinganine-phosphate (PHS-P) formation in chilled Arabidopsis cultured cells and plantlets. (a) After 15 min of labelling in the presence of [33P]-orthophosphate, cultured cells were exposed to 4°C for the indicated time or kept at 22°C. After extraction, radiolabelled sphingolipids were developed by thin-layer chromatography (TLC), together with cold authentic standards. Labelled lipids were detected by autoradiography, and quantified using a Storm PhosphorImager. Unlabelled commercial standards were visualized with primuline staining. The TLC shown is representative of four independent repeats. DHS-P, dihydrosphingosine phosphate; 4-SPH-P, Sphingosine-P. (b) Following [33P]-orthophosphate labelling and multistep purification (see Materials and Methods section), the cold-evoked phospho-long-chain base (LCB-P) was derivatized with o-phthaldialdehyde (OPA) and separated by high-performance liquid chromatography (HPLC). Fluorescence traces for purified LCB-P (black trace) and derivatized commercial PHS-P standard (grey trace) are presented. Signal intensity is indicated in arbitrary fluorescence units (A.U.). (c) Quantification of the PHS-P abundance in radiolabelled culture cells after different cold exposures, relative to the total incorporated radioactivity, expressed as per cent. (d) Fourteen-day-old plantlets were exposed to 4°C for up to 90 min or maintained at 22°C in the presence of [33P]-orthophosphate. After extraction, sphingolipids were treated as described in (a), and PHS-P was quantified relative to the total incorporated radioactivity, expressed as per cent. Asterisks indicate significant differences at < 0.05 (*) and < 0.01 (**).

To further characterize PHS-P formation on cold exposure, we compared the kinetics and levels of PHS-P synthesis in Arabidopsis cultured cells and in 14-d-old seedlings. As shown in Fig. 1(a,c), radiolabelled PHS-P was only faintly detected in unstressed cells, even after 240 min of labelling. By contrast, radiolabelled PHS-P was detected as soon as 1 min after cold exposure, reached a plateau between 3 and 30 min, and disappeared after 240 min. PHS-P represented a maximum of 28% of the radiolabelled sphingolipids after 3–30 min. No variation in PHS-P formation was observed in cells maintained at 22°C (Fig. 1a). As shown in Fig. 1(d), the kinetics of PHS-P synthesis in labelled seedlings were similar to those in cultured cells. Moreover, a maximum PHS-P level of 24% of the radiolabelled sphingolipids was reached after 3–30 min. Taken together, these data show that PHS-P synthesis is a generic, rapid and transient response of Arabidopsis to cold exposure.

Cold-triggered PHS-P synthesis is impaired in lcbk2 mutant plants

As a result of the rapidity of formation of PHS-P on cold exposure, we hypothesized that it originated from the phosphorylation of PHS by LCB kinase activity. In a first step, we analysed the effect of two LCB kinase inhibitors, DMS and TSP, on the formation of PHS-P triggered by cold in cell cultures. To minimize artefactual effects of the inhibitors, we analysed PHS-P formation after a short (30 min) treatment with inhibitors and stress duration (10 min). At this time point, PHS-P abundance had reached a steady and maximum level (Fig. 1c). As shown in Fig. 2, DMS and TSP inhibited PHS-P formation by 40% and 50%, respectively, suggesting that LCB kinase activity was indeed required for PHS-P formation. Nevertheless, the inhibition was only partial and was observed using high concentrations of inhibitors, which might generate side effects in plant cells. To confirm the implication of LCB kinase activity and to further identify the enzyme implied in cold-responsive PHS-P formation, we analysed the ability of a range of Arabidopsis LCB kinase mutant lines to synthesize PHS-P following cold exposure. We focused our interest on At2g46090 (LCBK2; this study), At4g21540 (SPHK1; Worrall et al., 2008) and At5g23450 (LCBK1; Imai & Nishiura, 2005), which have been proposed as LCB kinases in Arabidopsis (Worrall et al., 2008). As shown in Fig. 3(a), SPHK1, LCBK1 and LCBK2 genes were expressed in unstressed plantlets, and RNA abundance was not modified in seedlings exposed to cold. For subsequent investigations, we isolated homozygous mutant lines for the LCBK1 gene (one line, designated lcbk1) and LCBK2 gene (two lines, referred to as lcbk2.1 and lcbk2.2) (Fig. S1). We also used a previously described mutant line for sphk1 (Worrall et al., 2008). As shown in Fig. 3(b), the four mutant lines tested no longer expressed the mutated gene, whereas the expression of the two other LCB kinase genes was not modified. Using in vivo labelling with radioactive orthophosphate, we subsequently analysed the synthesis of PHS-P in the different mutant lines kept at 22°C or exposed to cold stress. As shown in Fig. 3(c), lcbk2 mutants exhibited a higher level (2.7- and 4.2-fold higher after 30 min for lcbk2.1 and lcbk2.2, respectively) of PHS-P synthesis compared with wild-type (WT) and lcbk1 and sphk1 mutants when kept at 22°C. However, although a dramatic increase in PHS-P synthesis was observed in WT, sphk1 and lcbk1 mutants exposed for 30 min to cold, compared with plantlets kept at 22°C (2.6, 2.1 and 3.3, respectively), PHS-P synthesis was not modified in lcbk2.1 and lcbk2.2 mutants (1.0 and 0.8, respectively) (Fig. 3c and Table S2). These data indicate that PHS-P synthesis detected in lcbk2 mutants is not regulated by cold stress. This constitutive activity resulted from the enhancement of a basal LCBK2-independent LCB kinase activity. However, PHS-P-synthesizing activity was induced in WT, sphk1 and lcbk1 plantlets on cold exposure. We therefore propose that LCBK2, but not SPHK1 and LCBK1, participates in cold-triggered PHS-P formation.

Figure 2.

Effect of long-chain base (LCB) kinase inhibitors on cold-triggered 4-hydroxy-sphinganine-phosphate (PHS-P) formation in Arabidopsis cultured cells. The relative abundance of [33P]-PHS-P was determined after 10 min of cold shock in cultured cells treated with the indicated concentrations of N,N-dimethylsphingosine (DMS) or threo-dihydrosphingosine (TSP). The thin-layer chromatography (TLC) shown is representative of four repeats. Data are the mean ± SE of the four independent repeats. DMSO, dimethylsulfoxide. Asterisks indicate significant differences at < 0.05 (*).

Figure 3.

Characterization of Arabidopsis long-chain base (LCB) kinase T-DNA insertion mutants. (a) The expression of LCBK1, LCBK2 and SPHK1 was analysed by semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) on 14-d-old wild-type (WT) seedlings exposed to 22 or 4°C for 1 h. S19 was used as a standard. A representative experiment from three independent repeats is shown. (b) The expression of LCBK1, LCBK2 and SPHK1 was analysed by semi-quantitative RT-PCR on 14-d-old seedlings from lcbk1, sphk1 and two lcbk2 (lcbk2.1 and lcbk2.2) mutant lines. S19 was used as a standard. A representative experiment from three independent repeats is shown. (c) Relative 4-hydroxy-sphinganine-phosphate (PHS-P) formation in 14-d-old plantlets from WT and lcbk1, sphk1 and lcbk2 mutants. In vivo labelling with radioactive orthophosphate was performed for 15 min at 22°C (0 min/22°C) followed by an additional 30 min at 22°C (30 min/22°C) or at 4°C (30 min/4°C). After extraction, sphingolipids were developed by thin-layer chromatography (TLC) and radiolabelled sphingolipids were detected and quantified using a Storm PhosphorImager. PHS-P formation was normalized by reference to PHS-P formation in WT plants at the onset of chilling (0 min/22°C). Data are the mean ± SE of at least five independent repeats. Different letters (a, b or c) indicate statistical differences at P < 0.01. Detailed values are presented in Supporting Information Table S2.

PHS-P participates in cold-responsive MAP kinase regulation

An early response of plants to cold stress is the transient activation of MAP kinase activity which is implicated in the onset of the plant adaptive response. In Arabidopsis, AtMPK4 and AtMPK6 are rapidly activated on cold shock (Ichimura et al., 2000). We therefore investigated a possible function for PHS-P in regulating AtMPK4 and AtMPK6 (Teige et al., 2004). Indeed, the ability of sphingolipids to activate MAP kinase pathways in plants has been reported (Lieberherr et al., 2005); Saucedo-Garcia et al., 2011). We compared the activation of AtMPK4 and AtMPK6 on cold exposure in WT and lcbk2 plantlets. As shown in Fig. 4, a 30-min cold treatment led to the activation of AtMPK4 and AtMPK6 in WT seedlings. A similar response was observed for AtMPK4 in both lcbk2 mutant lines, indicating that the perturbation of PHS-P formation in lcbk2 had no effect on AtMPK4 regulation. By contrast, lcbk2 mutants presented a constitutively high AtMPK6 activity which was not enhanced further following cold exposure. As lcbk2 mutants also presented a constitutively high formation of PHS-P, we hypothesized that PHS-P could be implicated in the regulation of AtMPK6. To investigate this point, we submitted Arabidopsis cultured cells to exogenous treatments with a range of phosphorylated and unphosphorylated LCBs. When treatments were performed at 22°C, AtMPK6 was strongly activated by PHS-P, therefore strengthening the functional link hypothesized between the PHS-P level and AtMPK6 regulation in lcbk2 mutants (Fig. 5a). In these conditions, sphingosine-4-phosphate (4-Sphingosine-P) was also a potent activator of AtMPK6. However, AtMPK6 activity was not affected by unphosphorylated LCB or DHS-P, suggesting a specific effect for PHS-P and 4-SPH-P in our conditions. This was further supported on comparing the effect of similar treatments on AtMPK4 activity. Indeed, AtMPK4 was activated only slightly and by a large range of both phosphorylated and unphosphorylated LCBs (Fig. 5a). As shown in Fig. 5(b), the activation of AtMPK6 by PHS-P was observed for concentrations ranging from 6 to 60 μM, but not at higher concentrations. To further investigate the regulation of AtMPK6 by PHS-P, we performed similar experiments in cold-stressed cells (Fig. 5c). Strikingly, although PHS-P hardly modified AtMPK6 activity, PHS treatment strongly impaired AtMPK6 activation. Only DHS, but to a lesser extent, was also found to inhibit AtMPK6 activation. No significant effect was observed on AtMPK4 activity (Fig. 5c). Compared with the activation of AtMPK6 by PHS-P, the inhibition by PHS was dose dependent up to 95 μM (Fig. 5d). These experiments indicate that PHS-P, and also its unphosphorylated species PHS, are potent regulators of AtMPK6 activation. To determine whether AtMPK6 regulation by PHS-P at 22°C was dependent on modifications of PHS/PHS-P balance or was caused by PHS-P per se, we treated cultured cells with PHS or PHS-P alone, or with the two LCBs simultaneously. To prevent PHS to PHS-P conversion, LCB(P) treatments were performed in the presence of DMS to block LCB kinase activity. As observed by HPLC analyses of LCBs following treatment, PHS and PHS-P application led to the accumulation of the corresponding LCB in the cells, and showed the absence of detectable PHS-P neo-synthesis when PHS and DMS were added simultaneously (Fig. S2). As shown in Fig. 5(e), PHS-P induced AtMPK6 activity in treated cells, both in the presence and absence of PHS, indicating that the effect of PHS-P is not achieved through a modification of PHS/PHS-P balance, but by PHS-P per se. Taken together, these data support the idea that the constitutive activation of AtMPK6 observed in lcbk2 mutants is linked to constitutively high PHS-P formation in these mutants.

Figure 4.

Cold-responsive mitogen-activated protein (MAP) kinase activities in Arabidopsis lcbk2 mutants. Fourteen-day-old wild-type (WT) and lcbk2 mutants were either maintained at 22°C or transferred for 1 h at 4°C. AtMPK4 and AtMPK6 activities were assayed following immunoprecipitation with specific antibodies using myelin basic protein (MBP) as substrate. Phosphorylated MBP was detected by PhosphorImaging. WB shows the abundance of AtMPKs detected by Western blot using specific antibodies. Figures show representative experiments from three independent repeats. WB, western blot.

Figure 5.

Effect of exogenous treatments with long-chain bases (LCBs) on AtMPK4 and AtMPK6 activities. (a) Cultured cells were incubated for 30 min at 22°C with 20 μM of the indicated LCB or LCB-P. AtMPK4 and AtMPK6 activities were assayed following immunoprecipitation with specific antibodies using myelin basic protein (MBP) as substrate. Phosphorylated MBP was detected by PhosphorImaging. WB shows the abundance of AtMPKs detected by Western blot using specific antibodies. Figures show representative experiments from three independent repeats. DHS, dihydrosphingosine; DHS-P, dihydrosphingosine phosphate; 4-SPH, sphingosine; 4-SPH-P, sphingosine-Phosphate. (b) Cells were incubated for 30 min at 22°C with the indicated concentrations of 4-hydroxy-sphinganine-phosphate (PHS-P), and AtMPK6 activity was determined as in (a). (c) Cells were incubated for 30 min at 4°C with 20 μM of the indicated LCB or LCB-P, and AtMPK4 and AtMPK6 activities were determined as in (a). (d) Cells were incubated for 30 min at 4°C with the indicated concentrations of phytosphingosine (PHS), and AtMPK6 activity was determined as in (a). (e) AtMPK6 activity was determined in cells incubated at 22°C for 30 min in the presence of 90 mM N,N-dimethylsphingosine (DMS), together with 20 mM PHS and/or PHS-P. Dimethylsulfoxide (DMSO) treatment was used as a control. Data shown are representative of at least three repeats. WB, Western Blot.

Characterization of lcbk2 mutant responses to cold

To obtain further insight into the consequences of the impairment of cold-evoked PHS-P formation for the cold response of lcbk2 mutants, we first compared the expression of cold-regulated genes in WT and mutant plants. Among the different genes tested, we observed that the expression of the NITRATE-INDUCED 3 (NOI3) gene on cold exposure was fully impaired in the lcbk2 mutant (Fig. 6). We have previously identified NOI3 in a microarray screen for genes regulated by LCB kinase inhibitors (C. Dutilleul & I. Guillas, unpublished). The lack of expression of NOI3 in cold-treated lcbk2 mutants therefore suggests the existence of cold-responsive genes, the expression of which is regulated by sphingolipid signalling during cold stress.

Figure 6.

Cold-responsive gene expression in Arabidopsis lcbk2 mutants. Fourteen-day-old wild-type (WT) and lcbk2 mutants were maintained at 22°C or transferred to 4°C for 1 h. Gene expression of NITRATE-INDUCED 3 (NOI3) was analysed using quantitative reverse transcription-polymerase chain reaction (RT-PCR) as described in the Materials and Methods section.

The phenotypic response of lcbk2 mutants on cold exposure was compared with that of WT and other LCB kinase mutants. No modification of germination, foliar growth or bolting was observed between the different lines grown at 22, 12 or 4°C (data not shown). As shown in Fig. 7, root growth was similar in the mutant lines tested relative to WT, when experiments were performed at 4°C. By contrast, although WT, sphk1 and lcbk1 mutants showed similar root growth at 12°C, lcbk2 lines exhibited a significantly higher root growth (Fig. 7). This unexpected phenotype prompted us to investigate the molecular responses occurring in WT and lcbk2 at this particular temperature. At 12°C, PHS-P formation was 2.3-fold higher in lcbk2 than in WT plantlets (Table S3). However, AtMPK6, which was constitutively activated at 22°C in lcbk2 plantlets, was activated to comparable levels in WT and lcbk2 plantlets exposed to 12°C (Fig. S3). As observed at 22 and 4°C, ATMPK4 activity was similar in WT and lcbk2 plantlets exposed to 12°C (Fig. S3a). Moreover, the NOI3 gene was not expressed in plantlets exposed to 12°C (Fig. S3b). As no difference was observed for AtMPK6 activation and NOI3 expression between WT and lcbk2 seedlings, we conclude that AtMPK6 and NOI3 probably do not participate in the enhanced root growth phenotype observed in lcbk2 plantlets at 12°C.

Figure 7.

Primary root growth of long-chain base (LCB) kinase mutants exposed to cold. Three-day-old seedlings were either maintained at 22°C or transferred to 12 or 4°C. Primary root growth was analysed over 7 d (22°C), 14 d (12°C) and 42 d (4°C). The inset represents a magnification of root growth at 4°C. The results are expressed as daily growth in millimetres and represent the mean ± SE (n = 40–120). Asterisks indicate the statistical difference at < 0.0001. WT, wild-type.

Previous reports have indicated that DELLA regulators are key factors linking environmental stress responses and modulation of plant development (Achard et al., 2008b). In particular, DELLA regulators participate in cold-dependent growth repression (Achard et al., 2008a). We therefore analysed the expression of the five DELLA genes in WT and lcbk2 plants exposed to 12 and 4°C. As shown in Fig. 8(a), although the expression of RGA, GAI, RGL1 and RGL2 was not affected by cold treatment, RGL3 transcripts accumulated in a temperature-dependent manner in cold-treated WT plants, which is consistent with previous reports (Achard et al., 2008a; Li et al., 2011). Interestingly, the cold-responsive accumulation of RGL3 transcripts was not observed in lcbk2 mutants exposed to 12°C. By contrast, RGL3 expression was not affected by lcbk2 mutation when plants were exposed to 4°C. A similar pattern of gene response was observed for the AtHB12 gene, a putative target for DELLA factors in stress responses (Achard et al., 2008b). To further establish the link between RGL3 and AtHB12 regulation and PHS-P, we analysed the effect of exogenous treatments with PHS or PHS-P on RGL3 and AtHB12 expression. As shown in Fig. 8(b), both RGL3 and AtHB12 were regulated by PHS and PHS-P treatment. PHS-P application enhanced RGL3 transcript abundance, but decreased AtHB12 expression in treated cells. In addition, the abundance of AtHB12 transcripts was enhanced by PHS (Fig. 8b). Taken together, these data strongly suggest that the impairment of RGL3 and AtHB12 activation, which is observed in lcbk2 mutants at 12°C, is associated with PHS-P, and participates in the differences in root growth observed in lcbk2 plantlets exposed to mild cold.

Figure 8.

Effects of lcbk2 mutation and phytosphingosine/4-hydroxy-sphinganine-phosphate (PHS/PHS-P) treatment on DELLA and AtHB12 gene expression. Gene expression was analysed using semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) as described in the Materials and Methods section. Data are representative of at least three repeats. (a) Fourteen-day-old wild-type (WT) and lcbk2 seedlings were maintained at 22°C or transferred to 12 or 4°C for 4 h (b) Cultured cells were treated for 4 h at 22°C with dimethylsulfoxide (DMSO) or with 20 μM of Phytosphingosine/Phytosphingosine-Phosphate (PHS/PHS-P).


In the present study, we investigated the variation and possible function of PHS-P formation during the early stages of Arabidopsis response to low temperature. Using in vivo metabolic labelling, we showed that PHS-P is rapidly and transiently synthesized in both cultured cells and plantlets. These characteristics are consistent with the involvement of PHS-P in cell signalling triggered by low temperature. The kinetics of PHS-P formation make it probable that PHS-P is derived from PHS phosphorylation by LCB kinase activity. At the onset of this work, two bona fide LCB kinases had been identified in Arabidopsis: SPHK1 (At4g21540; Worrall et al., 2008) and LCBK1 (At5g23450; Imai & Nishiura, 2005). In addition, a third isoform (At2g46090), designated LCBK2, and which presented sequence similarities with SPHK1 and LCBK1, was also proposed as a putative LCB kinase (Coursol et al., 2005; Worrall et al., 2008). Our metabolic labelling data indicated that LCBK1 and SPHK1 were not required for the formation of PHS-P evoked on cold exposure in vivo. By contrast, cold-triggered PHS-P formation was dramatically impaired in lcbk2 mutants. This constitutes the first evidence for a function for LCBK2 in sphingolipid metabolism. Surprisingly, the formation of PHS-P was enhanced in unstressed lcbk2 mutants, suggesting that LCBK2 depletion provokes a deregulation of PHS/PHS-P metabolism, leading to an overproduction of PHS-P via the activity of the remaining LCB kinases. However, it suggests that, during these experiments, LCBK2 becomes active only after cold exposure. This observation could explain why no LCB kinase activity has been measured so far on recombinant LCBK2 (Worrall et al., 2008). Post-translational regulations of mammalian LCB kinases, leading to modifications of protein activity, location and/or stability, and therefore impacting on their biological function, have been reported (reviewed in Snider et al., 2010). LCBK2 might therefore constitute an example of a stress-activated LCB kinase in plants, and offers the opportunity for further investigations on LCB kinase regulation.

In addition to displaying enhanced PHS-P formation, lcbk2 mutants presented a high constitutive activity of the MAP kinase AtMPK6. By contrast, AtMPK4 was not affected. AtMPK6 has recently been shown to be a transducer in the LCB signalling pathway, leading to cell death in Arabidopsis (Saucedo-Garcia et al., 2011). Together with AtMPK4, AtMPK6 is transiently activated on cold exposure and participates in Arabidopsis cold tolerance (Ichimura et al., 2000; Teige et al., 2004). Contrary to AtMPK4, which was activated in WT and lcbk2 mutants only after cold exposure, the activation of AtMPK6 was not further increased after cold stress. The deregulation of AtMPK6, observed in the lcbk2 mutant, might be a consequence of the modification of PHS-P synthesis in these mutants. Indeed, exogenous treatments of cell cultures with PHS-P mimicked cold-regulated AtMPK6 activation and did not affect AtMPK4. In contrast with the observations of Saucedo-Garcia et al. (2011), unphosphorylated LCBs did not trigger AtMPK6 activation in our conditions, which might reflect differences in the biological material used and/or in the treatment procedure. Although it did not affect AtMPK6 activity in unstressed cells, PHS treatment inhibited cold-evoked AtMPK6 activation. Interestingly, the antagonistic effect of PHS and PHS-P on AtMPK6 activity was not dependent on the relative PHS/PHS-P balance, but, rather, relied on the signalling properties of the molecules per se. We observed that other LCB/LCB-P also modified AtMPK6 activation. Together with the data of Saucedo-Garcia et al. (2011), our results therefore suggest that AtMPK6 could constitute a common transducer of LCB signals in different stress contexts. So far, the consequences of the constitutive activation of AtMPK6 in lcbk2 mutants remain unclear. As proposed recently, it might help to speed up or prime the adaptive stress response (Beckers et al., 2009). Further investigations are required to address this hypothesis and to analyse the mechanisms linking PHS-P, AtMPK6 activation and the control of cold-responsive processes.

We did not observe significant differences in the growth capacity and development of lcbk2 mutants when exposed to 4°C, compared with the other lines tested. Nevertheless, the expression of specific cold-responsive genes was affected, as exemplified for the nitrate-responsive gene NOI3. Surprisingly, the impairment of NOI3 expression observed in lcbk2 mutants exposed to 4°C was also observed in cold-treated cells challenged with PHS and PHS-P (Fig. S4). This result is contrary to what might be expected from the NOI3 expression pattern in lcbk2 mutants. This discrepancy might reflect the differences between endogenously generated and exogenously applied molecules in terms of concentrations, subcellular localization and/or tissue responsiveness (in our case, cultured cell vs plantlet responsiveness). Moreover, as reported in tobacco cells treated with DHS, exogenous PHS/PHS-P could modify the level of a range of LCB/LCB-P, and possibly other signalling sphingolipids, such as ceramide (Lachaud et al., 2010). Although the differences observed between lcbk2 mutants and PHS-P treatments are not understood, our data suggest that the defect in NOI3 expression in cold-treated lcbk2 plants could be related to the perturbation of PHS-P formation in this mutant. This effect is probably indirect, as PHS/PHS-P is not sufficient to induce NOI3 expression in unstressed cells. Contrary to other cold-responsive genes analysed in our study, NOI3 was only expressed after exposure to 4°C. Moreover, an effect of exogenously applied PHS/PHS-P was observed only when NOI3 expression was induced by low temperature, suggesting a complex interplay between cold and LCB signalling leading to NOI3 regulation. NOI3 is located at the plasma membrane and belongs to a family of 12 homologues (designated NOI1–12) including RIN4, a key component of Arabidopsis resistance towards Pseudomonas syringae (Kim et al., 2005; Takemoto & Jones, 2005). So far, the role of the other NOI proteins remains unknown. In this context, the regulation of the NOI3 gene during the plant response to low temperature suggests that the function of NOI proteins is not restricted to plant–pathogen interactions and might affect the plant response to a larger range of environmental cues.

Although no phenotypic differences could be observed between lcbk2, WT and the other lcbk mutants when grown at 22 or 4°C, lcbk2 mutants retained a higher root growth capacity under moderate temperature decrease (i.e. 12°C) compared with WT and the other LCB kinase mutants. At 12°C, PHS-P formation was not stimulated in WT plants, and lcbk2 mutants therefore retained a 2.3-fold higher rate of PHS-P formation compared with WT (Table S3). Interestingly, we found that the expression of DELLA factors was also specifically deregulated in lcbk2 plantlets exposed to 12°C. Indeed, we observed that the transcriptional activation of RGL3, which was the only DELLA gene undergoing transcriptional activation at 12°C in our experiments, was impaired in lcbk2 mutants. The five DELLA proteins, designated RGA, GAI and RGL1–3, are major repressors of plant growth and development, the action of which is counteracted by gibberellins (Achard & Genschik, 2009). Under adverse conditions, they participate in plant growth inhibition, therefore optimizing the chance for plant survival. Of particular interest, the accumulation of RGL3 transcripts has been correlated directly with the repression of plant growth in response to temperature decrease (Achard et al., 2008a). The lack of induction of RGL3 at 12°C in lcbk2 mutants could therefore account for the higher root growth observed at this temperature. Interestingly, the accumulation of RGL3 transcripts was not affected in lcbk2 mutants at 4°C, which correlates with the absence of the root growth phenotype at this temperature. The mechanisms leading to the specific impairment of RGL3 expression at 12°C, but not at lower temperature, remain unknown. As the intensity of the plant response to cold is quantitatively correlated with the amplitude of the temperature decrease (Usadel et al., 2008), it is conceivable that the mechanisms leading to the impairment of RGL3 expression in lcbk2 at 12°C are overwhelmed at lower temperatures. The altered expression of RGL3 in lcbk2 is probably related to altered PHS-P metabolism. This is further supported by the observation that RGL3 transcripts accumulated in cultured cells challenged with PHS-P. These data suggest that RGL3 expression is sensitive to LCBs in vivo and might constitute a target for LCB signalling during the cold response. We also observed that the expression of AtHB12, a homeobox gene that functions as a negative regulator of growth (Olsson et al., 2004), was impaired at 12°C in lcbk2 mutants. Interestingly, AtHB12 has recently been reported to be among the 126 genes commonly deregulated in a DELLA quadruple mutant challenged with various stresses (Achard et al., 2008b). Moreover, athb12 mutants present a decreased inhibition of root growth by ABA (Olsson et al., 2004). As for RGL3, AtHB12 expression was also affected in cultured cells challenged with either PHS or PHS-P. Nevertheless, it presented an opposite profile of regulation compared with RGL3, as it was inhibited by PHS-P and stimulated by PHS treatment. Additional evidence is therefore required to fully establish the relationship between RGL3 expression and AtHB12 regulation following cold stress and LCB/LCB-P treatment. As discussed for NOI3, although AtHB12 appears to be an LCB-regulated gene, our data also reveal different effects of endogenously evoked and exogenously applied LCB/LCB-P. It should be noted that exogenous treatment of cells with PHS/PHS-P has been performed at 22°C and has shown the ability of these molecules to regulate AtHB12 in unstressed material, which differs significantly from the physiological situation of lcbk2 plantlets exposed to low temperature. Finally, the different responses of NOI3, AtHB12 and RGL3 reported in our study further illustrate the difficulties encountered in unravelling the link between gene regulation and a specific LCB/LCB-P species, because of the numerous metabolic interconnections of sphingolipids (Cowart et al., 2010; Lucki & Sewer, 2011). Although the mechanisms by which PHS/PHS-P affect their expression is currently unknown, the deregulation of RGL3 and AtHB12 at 12°C in lcbk2 plantlets probably accounts for the root growth phenotype observed in this mutant, and is most probably a consequence of altered PHS-P formation.

As reported, the alteration in PHS-P formation in either lcbk2 mutants or cell cultures challenged with commercial LCBs had multiple effects on the plant response to cold, including cold-responsive gene expression, AtMPK6 activation and plant growth. Unexpectedly, although lcbk2 mutants were indeed impaired in cold-triggered PHS-P formation, they also presented a constitutively higher rate of PHS-P synthesis. The existence of such compensatory mechanisms is likely to dampen the impact of LCB kinase mutations, and might further explain the absence of phenotypes under severe stress, as proposed by Worrall et al. (2008) and suggested in this study. Indeed, most of the mutants impaired in LCB-P metabolism only showed faint phenotypes, if any, under the different stressing conditions tested previously (Nishikawa et al., 2008; Worrall et al., 2008; Michaelson et al., 2009). The recent identification of a new LCB kinase, designated SPHK2 (At4g21534), which shares strong molecular and biochemical similarities with SPHK1, further illustrates the existence of multiple LCB kinases possibly acting in a redundant way (Guo et al., 2011). In this sense, mutants in multiple LCB kinases will become a valuable tool to further investigate the function of LCB signalling in the plant stress response.


This work was supported by the Ministère de la Recherche et de l’Enseignement Supérieur, by the Centre National de la Recherche Scientifique (CNRS) and by the Université Pierre et Marie Curie. C.D. was a recipient of a post-doctoral fellowship from CNRS (2006–2008). We thank R. Atanassova for technical assistance.