Connections between Sphingosine Kinase and Phospholipase D in the Abscisic Acid Signaling Pathway in Arabidopsis*

Background: Sphingosine kinase (SPHK) and phospholipase D (PLD) produce different lipid mediators involved in abscisic acid (ABA) response. Results: Ablation of SPHKs and PLDα1 attenuates ABA-induced production of LCBPs and PA. Phyto-S1P closes stomata in sphk1, sphk2, but not in pldα1, whereas PA closes stomata in all mutants. Conclusion: SPHK acts upstream of PLDα1, whereas PLDα1 promotes SPHK. Significance: The roles of lipid messengers in the ABA signaling pathway are clarified. Phosphatidic acid (PA) and phytosphingosine 1-phosphate (phyto-S1P) both are lipid messengers involved in plant response to abscisic acid (ABA). Our previous data indicate that PA binds to sphingosine kinase (SPHK) and increases its phyto-S1P-producing activity. To understand the cellular and physiological functions of the PA-SPHK interaction, we isolated Arabidopsis thaliana SPHK mutants sphk1-1 and sphk2-1 and characterized them, together with phospholipase Dα1 knock-out, pldα1, in plant response to ABA. Compared with wild-type (WT) plants, the SPHK mutants and pldα1 all displayed decreased sensitivity to ABA-promoted stomatal closure. Phyto-S1P promoted stomatal closure in sphk1-1 and sphk2-1, but not in pldα1, whereas PA promoted stomatal closure in sphk1-1, sphk2-1, and pldα1. The ABA activation of PLDα1 in leaves and protoplasts was attenuated in the SPHK mutants, and the ABA activation of SPHK was reduced in pldα1. In response to ABA, the accumulation of long-chain base phosphates was decreased in pldα1, whereas PA production was decreased in SPHK mutants, compared with WT. Collectively, these results indicate that SPHK and PLDα1 act together in ABA response and that SPHK and phyto-S1P act upstream of PLDα1 and PA in mediating the ABA response. PA is involved in the activation of SPHK, and activation of PLDα1 requires SPHK activity. The data suggest that SPHK/phyto-S1P and PLDα1A are co-dependent in amplification of response to ABA, mediating stomatal closure in Arabidopsis.

Phosphatidic acid (PA) 4 produced by phospholipase Ds (PLDs) has been identified as important lipid signaling molecules in cell growth, development, and stress responses in both plants and animals (1,2). In Arabidopsis, the level of PAs increases rapidly under various conditions, including chilling, freezing, wounding, pathogen elicitation, dehydration, salt, nutrient starvation, nodule induction, and oxidative stress (1)(2)(3)(4). PLD and PAs are involved in the response of guard cells to abscisic acid (ABA) (5)(6)(7)(8). ABA failed to induce stomatal closure in PLD␣1-deficient plants, whereas overexpression of PLD␣1 resulted in increased sensitivity to ABA (8). PLD␣1 mediates ABA signaling via PA interacting with ABI1 phosphatase 2C (7). This interaction impedes the negative function of ABI1 in ABA response and mediates ABA-promoted stomatal closure (7,9). On the other hand, PLD␣1 interacts with the GDP-bound G␣ to regulate stomatal opening (9). PLD␣1 has also been implicated in reactive oxygen species production in Arabidopsis through the regulation of NADPH oxidase activity to promote stomatal closure (8). These studies indicate that PA is an important second messenger in the regulation of multiple mediators that determine stomatal aperture in response to ABA.
ABA is an important endogenous phytohormone regulating developmental processes and stress responses in plants (10,11). In response to drought stress, ABA levels increase rapidly and initiates a network of signaling pathways in guard cells leading to stomatal closure (11). A number of intermediate components of the ABA signaling pathway have been identified by forward and reverse genetic approaches (10 -14). Recently, proteins, known as pyrabactin resistance 1, pyr1-like proteins, or regulatory components of ABA receptors have been identi-fied as ABA receptors (15)(16)(17)(18). ABA binds to the receptor pyrabactin resistance/pyr1-like protein/regulatory components of ABA receptors, resulting in inhibition of the negative regulator ABI1, allowing SNF1-related kinase 2 (SnRK2) activation, mediating downstream signaling (11). Pyrabactin resistance/pyr1-like proteins are soluble proteins present in the cytosol and nucleus (17). Other proteins that interact with ABA were reported to be localized in the plastids or on the plasma membrane (14,19,20). The role of cell membrane in ABA perception and signaling is not fully understood (22).
Sphingolipids are essential components of eukaryotic membranes and their metabolites also function as important regulators of many cellular processes (23,24). Phosphorylated sphingolipids, such as sphingosine 1-phosphate (S1P), are potent messengers in the regulation of a variety of processes in animals, including cell proliferation and survival (25). A number of genes involved in sphingolipid biosynthesis have been identified and characterized in Arabidopsis (26,27). These studies indicate important roles for sphingolipids in plant growth, development, and response to stresses. Phosphorylated long-chain bases (LCBP), such as S1P and phytosphingosine 1-phosphate (phyto-S1P), have been implicated in the regulation of ABA-mediated stomatal behavior through G proteins in plants (28 -31). A recent study suggests that sphingosine and S1P are absent in Arabidopsis leaves due to the lack of expression of sphingolipid ⌬4-desaturase (32). However, plants have other LCBPs, including phyto-S1P, a LCBP produced by sphingosine kinase (SPHK) (30). Phyto-S1P is implicated as a signaling molecule regulating ABA-dependent stomatal movement (30).
SPHK activity was recently established in Arabidopsis, and two genes, SPHK1 (At4g21540) and SPHK2 (At4g21534), have been cloned and characterized (30 -33). Both SPHKs were active and able to use various long-chain bases (LCBs) as substrates (31,33). SPHK activity was shown to be rapidly induced by ABA and the production of phyto-S1P was involved in promotion of stomatal closure in response to ABA (29,30). Overexpression of SPHK1 increased ABA sensitivity during stomatal closure and germination (31). However, the physiological function of SPHK2 is unknown, and the mode of regulation of SPHK activation remains elusive. We recently showed that PA interacted with SPHK1 and SPHK2 and promoted their activity in vitro (33). This study was undertaken to determine the cellular and physiological functions of the PA-SPHK interaction. The results show that PA interacts directly with SPHK in Arabidopsis and that PLD␣1 and PA act downstream of SPHK. Together, PLD␣1/PA and SPHK/phyto-S1P function in a positive feedback loop to amplify the ABA signal for stomatal closure in Arabidopsis.
Plant Growth Conditions and Treatments-Plants were grown in soil in a growth chamber with cool white light of 160 mol m Ϫ2 s Ϫ1 under 12 h light/12 h dark and 23/19°C cycles. The seed germination assay and root elongation assay were performed on agar plates containing 1 ⁄ 2 Murashige and Skoog (MS) medium supplemented with 1% sucrose. Desiccated seeds were sterilized in 70% ethanol followed by 20% bleach, rinsed three times with sterilized water, and placed on plates with or without ABA. The plates were kept at 4°C for 2 days before moving to the growth chamber under the same conditions described previously. For root elongation measurements, 4-day-old seedlings were transferred to 1 ⁄ 2 MS medium with 0 to 10 M ABA; root lengths were recorded daily.
RNA Extraction and Real-time PCR-Real-time PCR was performed as described previously (34). Briefly, total RNA was digested with RNase-free DNase I and 1 g of RNA was used for synthesis of the first-strand cDNA using an iScript cDNA synthesis kit in a total reaction volume of 20 l according to the manufacturer's instructions (Bio-Rad). The primer sequences were described previously (33). The efficiency of the cDNA synthesis was assessed by real-time PCR amplification of a control gene encoding UBQ10 (At4g05320). cDNAs were then diluted to yield similar threshold cycle (C t ) values based on the C t of the UBQ10. The level of individual gene expression was normalized to that of UBQ10 by subtracting the C t value of UBQ10 from the tested genes. PCR was performed with a MyiQ system (Bio-Rad) using SYBR Green. Each reaction contained 7.5 l of 2ϫ SYBR Green master mix reagent (Bio-Rad), 3.5 l of diluted cDNA, and 200 nM of each gene-specific primer in a final volume of 15 l. The following standard thermal profile was used for all PCRs: 95°C for 3 min; and 50 cycles of 95°C for 30s, 55°C for 30 s, and 72°C for 30 s.
Purification of SPHK from Protoplasts and Immunoprecipitation-Mesophyll protoplasts were isolated from 4-week-old Arabidopsis leaves overexpressing SPHK2 according to a procedure previously described (35). Protoplast labeling and protein extraction was performed as described previously (7). Protoplasts were labeled with 0.5 mg/ml of 16/12-NBD-PC (Avanti) for 80 min and washed two times with the protoplast W5 buffer (35) to remove unlabeled NBD-PC. NBD-PC-labeled protoplasts were treated with 50 M ABA for 0 -30 min, followed by lysis in protoplast lysis buffer (20 mM Tris-HCl, pH 7.5, 20 mM KCl, 1 mM EDTA, 10 mM DTT, 0.5% Triton X-100, 50% glycerol, 10 g/ml of antipain, 10 g/ml of leupeptin, 10 g/ml of pepstatin, 1 mM phenylmethylsulfonyl fluoride) on ice for 5 min. Spermidine (5 mM) was added to the lysate followed by centrifugation at 10,000 ϫ g for 10 min. The cellular extract was incubated with anti-FLAG beads (Sigma) at 4°C for 3 h. The beads were pelleted by centrifugation and washed three times. Washed beads were extracted with chloroform:methanol (2:1). The extracts were dried under a stream of N 2 , dissolved in chloroform, and separated by TLC (Silica Gel 60 F254; Merck, Darmstadt, Germany). NBD-PA, scraped from TLC plates, was quantified using a fluorescence spectrophotometer, by comparing fluorescence intensities to those on a standard curve constructed with known amounts of NBD-PA.
Fluorescence-based in Vivo Assay of Sphingosine Kinase Activity-Protoplasts were prepared from fully expanded leaves of 4-week-old Arabidopsis. Protoplasts were incubated in 0.1 mg/ml of NBD-sphingosine for 80 min on ice and washed briefly. Washed protoplasts were kept at room temperature for 30 min. To determine in vivo sphingosine kinase activity based on the production of NBD-sphingosine 1-phosphate (NBD-S1P), 50 M ABA was added to NBD-sphingosine-labeled protoplasts (3 ϫ 10 5 for each assay) and incubated in a glass tube at room temperature for the indicated times (0 -20 min). 800 l of chloroform:methanol:concentrated HCl (100:200:1; v/v/v) was added to extract the lipids. 250 l of chloroform and 250 l of 2 M KCl were added sequentially. The sample was vortexed and centrifuged to generate a two-phase system. The lower chloroform phase was collected into a clean glass tube. Samples were dried under nitrogen and then resuspended in 50 l of chloroform. Lipid samples were spotted onto TLC plates and separated with chloroform:acetone:methanol:acetic acid:water (10: 4:3:2: The regions corresponding to NBD-S1P and NBD-sphingosine were marked, scraped from the plates, placed in 600 l of chloroform:methanol:water (5:5:1), vortexed, and centrifuged for 5 min at 15,000 ϫ g. The fluorescence (excitation 460 nm, emission 534 nm) of the eluted lipids was measured in a fluorescence spectrophotometer.
To assay the activity of the purified SPHK1 and SPHK2 using NBD-sphingosine as substrate, 1-10 g of NBD-sphingosine was incubated in sphingosine kinase buffer (20 mM Tris, pH 7.4, 20% glycerol, 1 mM mercaptoethanol, 1 mM EDTA, and 0.25% (v/v) Triton X-100, 1 mM ATP, and 10 mM MgCl 2 ) with 10 g of SPHK1 or SPHK2 purified from Escherichia coli for 10 min at 37°C. Lipid extraction and separation by TLC was described above.
Fluorescence-based in Vivo Assay of Phospholipase D Activity-A PLD activity assay was performed according to a procedure described previously (7). Protoplasts prepared from leaves of 4-week-old plants were incubated in 0.5 mg/ml of NBD-PC for 80 min on ice. To determine PLD activity, as affected by ABA treatment at different time points in vivo, 100 M ABA was added to the NBD-PC-labeled protoplasts, and 100-l aliquots (ϳ1.5 ϫ 10 5 for each assay) were transferred to a new tube at the end of each treatment. 0.4 ml of hot isopropyl alcohol (75°C) was added, and the mixture was incubated for 10 min at 75°C to inactivate PLD. Lipids were extracted with 0.5 ml of chloroform:methanol:water (5:5:1). The phases were separated and 100 l of chloroform were added to the aqueous phase, vortexed, centrifuged at 15,000 ϫ g for 2 min, and the lower chloroform phases were pooled. Each sample was dried under nitrogen and 20 l of chloroform:methanol (95:5) were added. NBD-PC and NBD-PA were separated by TLC developed in chloroform:methanol:NH 4 OH (65:35:5) and visualized under UV illumination. The regions corresponding to NBD-PC and NBD-PA were marked and scraped from the plates. The scraped silica gel was placed in 600 l of chloroform:methanol (2:1), vortexed, and centrifuged for 5 min at 15,000 ϫ g. The eluted lipids were quantified by fluorescence spectrophotometry (excitation 460 nm, emission 534 nm).

ESI-MS/MS Analysis of Lipid Molecular
Species-Lipids were extracted and PA was analyzed by electrospray ionization tandem mass spectrometry (ESI-MS/MS) as described by Xiao et al. (36). Expanded leaves of 4 -5-week-old plants were sprayed with 100 M ABA with 0.01% Triton X-100. The leaves were excised and immersed in 3 ml of isopropyl alcohol with 0.01% butylated hydroxytoluene (preheated to 75°C) immediately after sampling. The experiment was repeated 3 times with 5 replicates of each treatment each time.
HPLC/ESI-MS/MS Analysis of LCBPs-Sample preparation and analysis of LCB(P)s was carried out according to the method described by Markham and Jaworski (37) with some modifications. Briefly, 4 -5-week-old plants were sprayed with 100 M ABA with 0.01% Triton X-100. The excised leaves were extracted 5 times with solvent H (lower phase of isopropyl alcohol/hexane/water, 55:20:25 (v/v/v)) with agitation in a 60°C water bath for 15 min. The extract was transferred to a new glass tube and the combined extract was dried under a stream of nitrogen. Further steps of sample preparation and mass spectrometry analysis were carried out as described previously (37).

Manipulations of SPHKs and Their Expression in Response to
ABA-To determine the function of SPHK1 and SPHK2 in Arabidopsis, we isolated two T-DNA insertion mutant lines for SPHK1 and SPHK2. Sphk1-1 (Salk_042034) and sphk2-1 (Salk_000250) each has a T-DNA insertion before the (SPHK1 or SPHK2) start codon (Fig. 1A). Both lines were homozygous confirmed by PCR (Fig. 1B). Plants of sphk1-1 and sphk2-1 grew and developed normally as WT under normal conditions in soil (supplemental Fig. S1). The mutant sphk2-1 displayed almost no detectable SPHK2 transcript, whereas its SPHK1 expression level was comparable with WT, as quantified by real-time PCR. In sphk1-1, the SPHK1 transcript was decreased by 81% compared with WT, whereas the transcript of SPHK2 was also comparable with WT (Fig. 1C). The expression of SPHK1 and SPHK2 was restored to the WT level in both sphk1-1 and sphk2-1 that were genetically rescued by the genomic sequence including both SPHK1 and SPHK2 (Fig. 1C). SPHK2-OE lines driven by 35 S-promoter were generated in our previous study, and the production of the introduced SPHK2 was detected by immunoblotting (33). Real-time PCR revealed that the expression level of SPHK2 was increased by 7-and 11-fold in SPHK2-OE2 and SPHK2-OE5 (Fig. 1C).
SPHK activity was shown to be quickly induced by ABA in a previous study (29). To determine whether the transcript levels of SPHK1 or SPHK2 are increased in response to ABA, we sprayed WT Arabidopsis leaves with ABA and checked the expression levels of SPHK1 and SPHK2 by real-time PCR. The transcript level of ABI1 began to increase 5 min after ABA treatment, but the transcript level of SPHK1 and SPHK2 did not change significantly (Fig. 1D). The level of ABA-induced ABI1 expression in both sphk1-1 and sphk2-1 (supplemental Fig. S2) was similar to that in WT Arabidopsis leaves (Fig. 1D). The results suggest that SPHK1 and SPHK2 are not induced at the transcriptional level by ABA and that knock-out of either SPHK1 or SPHK2 does not affect the ABA-induced expression of ABI1.

PA Interacts with SPHK and Promotes the Activity of SPHK in
Arabidopsis-Our previous study using E. coli-expressed proteins showed that PA bound to SPHK1 and SPHK2, and the interaction promoted the SPHK activity in vitro (33). To demonstrate their interaction and function in plants, we isolated protoplasts from the SPHK2-OE line, which expressed FLAGtagged SPHK2. NBD-PC-labeled protoplasts were washed and treated with 50 M ABA followed by lysis and immunoprecipitation with anti-FLAG beads. The lipid was extracted from the immunoprecipitated fraction and separated by TLC. NBD-PA was co-precipitated with SPHK2 ( Fig. 2A, inset). ABA treatment for 30 min increased the amount of NBD-PA pulled down with SPHK2 ϳ6-fold, suggesting that ABA activated PLD␣1 and increased the amount of PA interacting with SPHK2 in Arabidopsis cells (Fig. 2A).
To determine whether PA promotes SPHK activity in the cell, we developed an assay, using NBD-sphingosine-labeled protoplasts, for production of NBD-S1P in vivo. First, we used SPHK purified from E. coli to confirm that Arabidopsis SPHK could phosphorylate NBD-sphingosine. Both SPHK1 and SPHK2 phosphorylated NBD-sphingosine to NBD-S1P (Fig.  2B). We then labeled protoplasts with NBD-sphingosine followed by treatment with ABA or PA. Lipid extracts were separated by TLC and photographed under UV light (supplemental Fig. 3). ABA treatment increased SPHK activity; the highest level of NBD-S1P was produced after 2.5 min of ABA treatment  MARCH 9, 2012 • VOLUME 287 • NUMBER 11 (Fig. 2C). The level of NBD-S1P in SPHK2-OE protoplasts was 36% higher, whereas the level in sphk1-1 and sphk2-1 protoplasts was, respectively, 19 and 40% lower than WT at 2.5 min of ABA treatment (Fig. 2C).

Cross-talk between PLD and Sphingosine Kinase in Signaling
The ABA-induced activity of SPHK was also impaired in pld␣1; the level of NBD-S1P produced in pld␣1 was ϳ33% lower than that in WT. The results indicate that PLD␣1 is involved in activating SPHK in response to ABA (Fig. 2C). To determine whether the PLD product PA could stimulate SPHK in the cell, we added PA (18:1/18:1) to the protoplasts. Addition of PA increased NBD-S1P production by more than 60% in protoplasts of WT and pld␣1 at 5 min after treatment (Fig. 2D). Similar to the ABA treatment, the increased SPHK activity in the PA treatment was the highest in SPHK2-OE and lower in sphk1-1 and sphk2-1 protoplasts. However, unlike the ABA treatment, PA-treated WT and pld␣1-1 protoplasts exhibited the same magnitude and pattern of NBD-S1P increase (Fig. 2D). These data support the conclusion that SPHK is a target of PA and PLD-produced PA is involved in the SPHK activation in response to ABA.
SPHK Acts Upstream of PLD␣1 in the Signaling Pathway of the ABA-mediated Stomatal Closure-To determine the relationship of SPHK/phyto-S1P and PLD␣1/PA in the ABA signaling pathway, we measured stomatal aperture in response to phyto-S1P in SPHK and PLD␣1 mutants. Phyto-S1P produced by SPHK was shown previously to induce stomatal closure (30). We used phyto-S1P to treat epidermal peels and found that phyto-S1P caused stomatal closure in WT, sphk1-1, and sphk2-1 but not in pld␣1 or the double knock-out mutants of pld␣1sphk1-1 or pld␣1sphk2-1 (Fig. 3A). The result suggests that SPHK and phyto-S1P act upstream of PLD␣1 and PA.
We then treated the epidermal peels with PA to determine the effect of PA on stomatal closure in these mutant lines. PA (18:1/18:1) was able to cause stomatal closure in WT, pld␣1, sphk1-1, and sphk2-1 (Fig. 3B). This result is consistent with the finding (Fig. 3A) that PLD␣1 and PA act downstream of SPHKs to promote stomatal closure. To augment the finding, we added 1-butanol, which decreases PA production by PLD, to the Arabidopsis epidermal peels treated with phyto-S1P. 1-Butanol partially blocked the phyto-S1P-promoted stomatal closure in WT, sphk1-1, and sphk2-1, but had no effect on pld␣1 (Fig. 3A). The results support the notion that PLD/PA is involved in mediating the phyto-S1P signal in stomatal closure.
ABA-promoted PLD␣1 Activation Is Attenuated in SPHK Mutants-The above results indicate that both SPHK and PLD␣1 are involved in the same signaling pathway in ABApromoted stomatal closure, with SPHK and phyto-S1P acting upstream of PLD␣1. To define the effect of SPHK on PLD activity and PA production in response to ABA, we measured PA production in vivo using NBD-PC-labeled leaf protoplasts exposed to ABA or phyto-S1P. NBD-PC is hydrolyzed by PLDs and production of NBD-PA reflects the PLD activity (7). In addition, NBD-PA bound to SPHK2 ( Fig. 2A) and had a similar effect as a regular PA to induce stomatal closure in WT (supplemental Fig. S4), indicating that the presence of NBD in PA does not affect the function of PA. The production of PA increased almost 2-fold in WT in 40 min after the start of ABA treatment (Fig. 4, A and B). However, the increase in PA in both sphk1-1 and sphk2-1 was significantly smaller than that in WT. Compared with WT, after 40 min of ABA treatment, PA production in sphk1-1 and sphk2-1 was 17 and 30% lower, respectively (Fig. 4B). In pld␣1, the PA level was lower than WT and SPHK mutants, and there was no significant increase in PA (Fig.  4B), supporting the previous conclusion that PLD␣1 is the major PLD responsible for ABA-induced PA production (7).
We reasoned that if PLD␣1 acts downstream of SPHK, phyto-S1P should be able to activate PLD␣1. To test this hypothesis, we first tested whether phyto-S1P could stimulate PLD␣1 directly in vitro. Additions of different concentrations of phyto-S1P failed to increase PLD␣1 directly, indicating other cellular effectors are involved in the PLD activation by phyto-S1P (supplemental Fig. S5). We then treated the protoplasts with phyto-S1P and measured PA production in protoplasts (Fig. 4C). The production of PA was increased ϳ2-fold by phyto-S1P in WT and both SPHK mutants. PA reached the highest level after 10 min of incubation. Knock-out of PLD␣1 abolished the ABA or phyto-S1P-induced increase in PA (Fig. 4,  B and C). The response of PLD activity to phyto-S1P indicates that SPHK and phyto-S1P are involved in activation of PLD␣1 to produce PA in response to ABA.
ABA Induces Different PA Changes in WT, sphk1-1, sphk2-1, and SPHK2-OE Lines-To characterize the effect of SPHKs on PA production in response to ABA, we quantitatively profiled the changes in PA species in Arabidopsis leaves sprayed with ABA using ESI-MS/MS. Knock-out of PLD␣1 greatly reduced the PA production in response to ABA (8). The total amount of PA in sphk1-1 and sphk2-1 was not significantly different from   MARCH 9, 2012 • VOLUME 287 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8291 that of WT without ABA treatment (Fig. 5A). In WT, PA reached the highest level at 10 min after ABA treatment and then went down to the pretreatment level after 40 min (Fig. 5A). The total PA level was also increased in sphk1-1, sphk2-1, and SPHK2-OE leaves after ABA treatment (Fig. 5A). The PA level was higher than WT after ABA treatment in SPHK2-OE. However, the amount of PA was significantly lower in sphk1-1 and sphk2-1 treated by ABA for 5 and 10 min than in WT (Fig. 5A).

Cross-talk between PLD and Sphingosine Kinase in Signaling
The results indicate that decreased SPHK expression attenuates ABA-induced activation of PLD␣1, in agreement with the results for the in vivo PLD activity assay (Fig. 4B).
The change of PA species in response to ABA at 10 min was analyzed for WT, pld␣1, sphk1-1, sphk2-1, and SPHK2 (8, 38). The levels of all PA species were decreased in pld␣1 and the major overall decreases were due to decreases in 34:2 PA and 34:3 PA, two very abundant PAs in Arabidopsis leaves (Fig. 5B). In comparison, the levels of most PA species (except 36:6 and 36:5 PA) were higher in WT than in sphk1-1 and sphk2-1 after 10 min of ABA treatment (Fig. 5B). Overexpression of SPHK2 mainly resulted in higher levels of 34:2 PA and 34:3 PA compared with WT and other PA species did not change significantly (Fig. 5B). The results show that the activation of SPHK1 and SPHK2 affects levels of 34-carbon PAs more than other PAs.
LCBP Profiling Reveals Regulation of SPHK by PA-To determine the effect of PLD␣1/PA on the level of different LCBPs in Arabidopsis, LCBP species were profiled to measure LCBP changes in response to ABA. We first analyzed the LCBPs in Arabidopsis leaves from WT and mutant lines. The total content of four major LCBP species (d18:0-P, d18:1-P, t18:0-P, and t18:1-P) was comparable in WT, pld␣1, and sphk1-1 (Fig. 6A). The LCBP level in sphk2-1 was about 57% lower than that in WT, indicating that ablation of SPHK2 dramatically decreased LCBP production in Arabidopsis leaves (Fig. 6A). The total LCBP level was increased by 40% when SPHK2 was overexpressed in Arabidopsis (Fig. 6A). The lower level of total LCBP in sphk2-1 was mainly due to the decrease of t18:0-P and t18: 1-P (Fig. 6B). ABA treatment increased the LCBP content by 58% in WT leaves at 2 min after ABA treatment, but no such ABA-induced increase occurred in sphk1-1, sphk2-1, or pld␣1 (Fig. 6C).
LCBP production in pld␣1 was not induced by ABA as much as in WT and the increase was delayed until 10 min compared to WT (Fig. 6C). There is no significant increase in t18:1-P, and increases in d18:0-P and d18:1-P occurred 10 min after ABA treatments (Fig. 6D). The knock-out of PLD␣1 attenuated ABA activation of SPHKs, indicating that PLD and PA are involved in SPHK activation in response to ABA (Fig. 8).
SPHK2-KO and OE Alter Arabidopsis Sensitivity to ABA-To determine the effect of SPHK1 and SPHK2 mutations on the Arabidopsis response to ABA, we assayed ABA responses of sphk1-1 and sphk2-1 together with SPHK2-OE lines. Stomatal aperture was decreased by ABA in WT. However, sphk1-1 and sphk2-1 were less sensitive to ABA-promoted stomatal closure (Fig. 7A). Double mutants pld␣1sphk1-1 and pld␣1sphk2-1 were insensitive to ABA-caused stomatal closure like pld␣1 (Fig. 7A). Introducing a genomic sequence containing both SPHK1 and SPHK2 under their native promoters into sphk1-1 and sphk2-1 restored the stomatal response to ABA for both mutants, indicating that loss of SPHK1 and SPHK2 is responsible for the ABA response phenotype (Fig. 7A).
Knockdown of SPHK1 or SPHK2 decreased, whereas overexpression of SPHK2 increased ABA sensitivity during ABAinhibited root elongation (Fig. 7B). The root length of the two SPHK mutants was longer than that of WT under 5 or 10 M ABA. Overexpression of SPHK2 increased ABA sensitiv-FIGURE 6. Alterations of SPHKs change LCBP content and composition in Arabidopsis leaves. A, total LCBP content (mol %) in leaves from 4 -5-week-old WT, pld␣1, sphk1-1, sphk2-1, and SPHK2-OE5. B, LCBP composition in leaves from 4 -5-week-old WT, pld␣1, sphk1-1, sphk2-1, and SPHK2-OE5. C, total LCBP content in WT Arabidopsis leaves treated with ABA. 4 -5-Week-old Arabidopsis was sprayed with 100 M ABA with 0.01% Triton X-100 followed by sphingolipid extraction and MS analysis. D, LCBP composition in the leaves treated with 50 M ABA with 0.01% Triton X-100 for 0 -15 min. Data were calculated as molar percentage over the total amount of LCB (sphinganine (d18:0), 8-sphingenine (d18:1), phytosphingosine (t18:0), and 4-hydroxy-8-sphingenine (t18:1)) and LCBP (d18:0-P, d18:1-P, t18:0-P and t18:1-P). The experiment was performed twice and the results were consistent. Values are mean Ϯ S.E. for one experiment (n ϭ 5). Asterisks in B indicate that the mean value is significantly different from that of the WT at p Ͻ 0.05, based on Student's t test. Asterisks in C indicate that the mean value is significantly different from that of the 0-min ABA treatment for each Arabidopsis line at p Ͻ 0.05. Asterisks in D indicate that the mean value is significantly different from that of the 0 min ABA treatment for each Arabidopsis line at p Ͻ 0.05 based on Student's t test.
ity during ABA-inhibited root elongation as the root lengths in the OE lines were shorter than that of WT (Fig. 7B). Manipulation of SPHK1 and SPHK2 also altered ABA sensitivity during seed germination and post-germination growth. sphk1-1 and sphk2-1 germinated earlier than WT on 1 ⁄2 MS plates with different concentrations of ABA, whereas the germination of SPHK2-OE seeds was delayed and its postgermination growth was inhibited (Fig. 7, C and D). The data suggest that SPHK2 is involved in the control of three ABA responses in Arabidopsis.

DISCUSSION
SPHK1 and SPHK2 are two genes closely linked on chromosome 4 in Arabidopsis based on molecular cloning, sequence analysis, and distinguishable expression patterns (33). We isolated two T-DNA mutants, sphk1-1 and sphk2-1, for SPHK1 and SPHK2 separately. Real-time PCR indicated that the SPHK1 expression level was dramatically reduced in sphk1-1, whereas the transcript of SPHK1 was slightly induced in sphk2-1. In addition, the expression level of SPHK2 in sphk1-1 is not significantly different from that of WT. These data provide further evidence that SPHK1 and SPHK2 are two separate genes. SPHK1 was reported to have a role in two ABA signaling pathways in regulation of stomatal aperture and seed germination (31). The present study shows that both SPHK mutants display decreased sensitivity to ABA-promoted stomatal closure, ABA-inhibited root elongation and ABA-inhibited seed germination. In addition, SPHK2-OE lines were more sensitive to ABA in three ABA-mediated responses, indicating that SPHK2 is involved in ABA-mediated signaling pathways (Fig. 8).
Quantitative analysis of LCBP showed that the total LCBP level remained the same as WT in sphk1-1 but decreased about 57% in sphk2-1. The decreased LCBP content mainly came from t18:0-P and t18:1-P. There was still 43% of LCBP in sphk2-1 compared with WT, which is presumably a result of SPHK1 and other kinases including AtLCBK1 and AtCERK (39,40). These data indicate that whereas SPHK2 contributes more than SPHK1 to LCBP production in leaves, SPHK1 and SPHK2 have unique and overlapping functions in LCBP synthesis in Arabidopsis leaves. Availability of SPHK1xSPHK2 double knock-out mutants will be helpful to further determine the functions of both SPHKs. But isolating such mutants by crossing sphk1-1 and sphk2-1 has been unsuccessful because SPHK1 and SPHK2 are closely linked (33).
Our previous in vitro study showed that PA binds to SPHKs and stimulates their activity, suggesting that Arabidopsis SPHKs are molecular targets of PA (33). The present study using protoplasts provides in vivo evidence that PA binds to and stimulates SPHK. More evidence was garnered from the SPHK activity assay and quantitative profiling of LCBPs from leaves. Addition of PA promoted the production of NBD-S1P in WT protoplasts and SPHK activity was attenuated in pld␣1 when protoplasts were treated with ABA. LCBP analysis indicated that LCBP content increased by 58% in WT Arabidopsis leaves after a 2-min ABA treatment. Knock-out of PLD␣1 resulted in less than 10% increase of LCBP in response to ABA treatment, indicating PLD␣1 and PA were involved in promotion of SPHK activity in response to ABA (Fig. 8).
Phyto-S1P (t18:0-P) was capable of promoting stomatal closure (30). Phyto-S1P is one of the major LCBPs found in Arabidopsis leaves; it can serve as a signaling molecule to mediate ABA response. Our data show that ABA induced the increased production of all 4 LCBPs in Arabidopsis leaves. Whether the other three LCBPs are involved in the ABA-mediated signaling pathway needs to be determined. LCBPs have broad cellular functions in animals, and more functions of LCBPs in plants also should be explored.
In summary, the present physiological, genetic, and enzymatic analyses combined with lipid profiling clearly indicate a co-dependence between the two lipid signaling reactions, SPHK/phyto-S1P and PLD/PA (Fig. 8). PA produced by PLD␣1 interacts with SPHK and is required for SPHK activation in response to ABA. Increased phyto-S1P activates PLD␣1, leading to an increase in PA level. PA functions as a signaling molecule to regulate downstream proteins including ABI1 and NADPH oxidase in ABA-mediated stomatal closure. The ABA signal is transduced to downstream pathways and regulates ion channels, leading to stomatal closure (Fig. 8). It will be of interest in future studies to determine whether the interplay between PLD␣1/PA and SPHK/phyto-S1P is involved in other signaling and regulatory pathways in plant growth, development, and response to stresses.