Modifications of Sphingolipid Content Affect Tolerance to Hemibiotrophic and Necrotrophic Pathogens by Modulating Plant Defense Responses in Arabidopsis  

Abstract

Sphingolipids are emerging as second messengers in programmed cell death and plant defense mechanisms. However, their role in plant defense is far from being understood, especially against necrotrophic pathogens. Sphingolipidomics and plant defense responses during pathogenic infection were evaluated in the mutant of long-chain base phosphate (LCB-P) lyase, encoded by the dihydrosphingosine-1-phosphate lyase1 (AtDPL1) gene and regulating long-chain base/LCB-P homeostasis. Atdpl1 mutants exhibit tolerance to the necrotrophic fungus Botrytis cinerea but susceptibility to the hemibiotrophic bacterium Pseudomonas syringae pv tomato (Pst). Here, a direct comparison of sphingolipid profiles in Arabidopsis (Arabidopsis thaliana) during infection with pathogens differing in lifestyles is described. In contrast to long-chain bases (dihydrosphingosine [d18:0] and 4,8-sphingadienine [d18:2]), hydroxyceramide and LCB-P (phytosphingosine-1-phosphate [t18:0-P] and 4-hydroxy-8-sphingenine-1-phosphate [t18:1-P]) levels are higher in Atdpl1-1 than in wild-type plants in response to B. cinerea. Following Pst infection, t18:0-P accumulates more strongly in Atdpl1-1 than in wild-type plants. Moreover, d18:0 and t18:0-P appear as key players in Pst- and B. cinerea-induced cell death and reactive oxygen species accumulation. Salicylic acid levels are similar in both types of plants, independent of the pathogen. In addition, salicylic acid-dependent gene expression is similar in both types of B. cinerea-infected plants but is repressed in Atdpl1-1 after treatment with Pst. Infection with both pathogens triggers higher jasmonic acid, jasmonoyl-isoleucine accumulation, and jasmonic acid-dependent gene expression in Atdpl1-1 mutants. Our results demonstrate that sphingolipids play an important role in plant defense, especially toward necrotrophic pathogens, and highlight a novel connection between the jasmonate signaling pathway, cell death, and sphingolipids.

Plants have evolved a complex array of defenses when attacked by microbial pathogens. The success of plant resistance first relies on the capacity of the plant to recognize its invader. Among early events, a transient production of reactive oxygen species (ROS), known as the oxidative burst, is characteristic of successful pathogen recognition (Torres, 2010). Perception of pathogen attack then initiates a large array of immune responses, including modification of cell walls, as well as the production of antimicrobial proteins and metabolites like pathogenesis-related (PR) proteins and phytoalexins, respectively (Schwessinger and Ronald, 2012). The plant hormones salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) are key players in the signaling networks involved in plant resistance (Bari and Jones, 2009; Tsuda and Katagiri, 2010; Robert-Seilaniantz et al., 2011). Interactions between these signal molecules allow the plant to activate and/or modulate an appropriate array of defense responses, depending on the pathogen lifestyle, necrotroph or biotroph (Glazebrook, 2005; Koornneef and Pieterse, 2008). Whereas SA is considered essential for resistance to (hemi)biotrophic pathogens, it is assumed that JA and ET signaling pathways are important for resistance to necrotrophic pathogens in Arabidopsis (Arabidopsis thaliana; Thomma et al., 2001; Glazebrook, 2005). A successful innate immune response often includes the so-called hypersensitive response (HR), a form of rapid programmed cell death (PCD) occurring in a limited area at the site of infection. This suicide of infected cells is thought to limit the spread of biotrophic pathogens, including viruses, bacteria, fungi, and oomycetes (Mur et al., 2008).

During the past decade, significant progress has been made in our understanding of the cellular function of plant sphingolipids. Besides being structural components of cell membranes, sphingolipids are bioactive metabolites that regulate important cellular processes such as cell survival and PCD, occurring during either plant development or plant defense (Dunn et al., 2004; Berkey et al., 2012; Markham et al., 2013). The first evidence of the role of sphingolipids in these processes came from the use of the fungal toxins fumonisin B1 (FB1) and AAL, produced by the necrotrophic agent Alternaria alternata f. sp. lycopersici. These toxins are structural sphingosine (d18:1) analogs and function as ceramide synthase inhibitors. They triggered PCD when exogenously applied to plants. Mutant strains in which the production of such toxins is abrogated failed to infect the host plant, implying that toxin accumulation is required for pathogenicity and that the induction of plant PCD could be considered a virulence tool used by necrotrophic pathogens (Berkey et al., 2012). Moreover, several studies revealed that ceramides (Cers) and long-chain bases (LCBs) are also potent inducers of PCD in plants. For example, exogenously applied Cers and LCBs (d18:0, d18:1, or t18:0) induced PCD either in cell suspension cultures (Liang et al., 2003; Lachaud et al., 2010, 2011; Alden et al., 2011) or in whole seedlings (Shi et al., 2007; Takahashi et al., 2009; Saucedo-García et al., 2011). AAL- and FB1-induced PCD seemed to be due to the accumulation of free sphingoid bases (dihydrosphingosine [d18:0] and phytosphingosine [t18:0]; Abbas et al., 1994; Brandwagt et al., 2000; Shi et al., 2007). Spontaneous cell death in lag one homolog1 or l-myoinositol1-phosphate synthase mutant could be due to trihydroxy-LCB and/or Cer accumulation (Donahue et al., 2010; Ternes et al., 2011). Deciphering of Cer participation in the induction of HR and associated PCD also came from studies on accelerated cell death5 (acd5) and enhancing resistance to powdery mildew8 (RPW8)-mediated hypersensitive response (erh1) mutants, which displayed overaccumulation of Cers. These mutants exhibited spontaneous cell death and resistance to biotrophic pathogens, which seemed to be linked with SA and PR protein accumulation (Liang et al., 2003; Wang et al., 2008).

Altogether, these data provide evidence of a link between PCD, defense, and sphingolipid metabolism. However, the fatty acid hydroxylase1/2 (atfah1/atfah2) double mutant that accumulates SA and Cers was more tolerant to the obligate biotrophic fungus Golovinomyces cichoracearum but did not display a PCD-like phenotype, suggesting that Cers alone are not involved in the induction of PCD (König et al., 2012). Moreover, Saucedo-García et al. (2011) postulated that dihydroxy-LCBs, but not trihydroxy-LCBs, might be primary mediators for LCB-induced PCD. The sphingoid base hydroxylase sbh1  /sbh2 double mutant completely lacking trihydroxy-LCBs showed enhanced expression of PCD marker genes (Chen et al., 2008). On the contrary, increase in t18:0 was specifically sustained in plant interaction with the avirulent Pseudomonas syringae pv tomato (Pst) strain and correlated with a strong PCD induction in leaves (Peer et al., 2010). Thus, the nature of sphingolipids able to induce PCD is still under debate and may evolve depending on plants and their environment. The phosphorylated form of LCBs (LCB-Ps) could abrogate PCD induced by LCBs, Cers, or heat stress in a dose-dependent manner (Shi et al., 2007; Alden et al., 2011). Furthermore, blocking the conversion of LCBs to LCB-Ps by using specific inhibitors induced PCD in cell suspension culture (Alden et al., 2011). Recently, overexpression of rice (Oryza sativa) LCB kinase in transgenic tobacco (Nicotiana tabacum) plants reduced PCD after treatment with FB1 (Zhang et al., 2013). Genetic mutation on LCB-P lyase encoded by the AtDPL1 gene, modifying the LCB-LCB-P ratio, could impact PCD levels after treatment with FB1 (Tsegaye et al., 2007). Altogether, these data point to the existence of a rheostat between LCBs and their phosphorylated forms that controls plant cell fate toward cell death or survival.

Data on plant sphingolipid functions are still fragmentary. Only a few reports have described interconnections between sphingolipids, cell death, and plant defense responses, almost exclusively in response to (hemi)biotrophic pathogens. Knowledge about such relations in response to necrotrophic pathogens is still in its infancy (Rivas-San Vicente et al., 2013; Bi et al., 2014). In this report, the link between sphingolipids, cell death, and plant defense has been explored in response to Botrytis cinerea infection and in comparison with Pst infection. For this purpose, Atdpl1 mutant plants, disturbed in LCB/LCB-P accumulation without displaying any phenotype under standard growth conditions (Tsegaye et al., 2007), have been analyzed after pathogen infection. Our results revealed that modification of sphingolipid contents not only impacted plant tolerance to hemibiotrophs but also greatly affected resistance to necrotrophs. Whereas the SA signaling pathway is globally repressed in Atdpl1-1 compared with wild-type plants, the JA signaling pathway is significantly enhanced. Cell death and ROS accumulation are markedly modified in Atdpl1-1 mutant plants. We further demonstrated that phytosphingosine-1-phosphate (t18:0-P) and d18:0 are key players in pathogen-induced cell death and ROS generation. Here, we thus established a link between JA signaling, PCD, and sphingolipid metabolism.

RESULTS

Necrotrophic and Hemibiotrophic Infection Differently Affect the Atdpl1 Mutant Plant Response

In order to assess the role of sphingolipids in plant immune responses to necrotrophic and hemibiotrophic pathogens, we used the Atdpl1 mutant, which is affected in the LCB/LCB-P rheostat by accumulating t18:1-P (Tsegaye et al., 2007). Whereas Atdpl1 shows no developmental phenotype compared with wild-type plants under standard conditions, it exhibits a higher sensitivity to FB1 (Tsegaye et al., 2007). B. cinerea and Pst have been widely used to decipher defense mechanisms to necrotrophic and hemibiotrophic pathogens in Arabidopsis (Glazebrook, 2005). To get some information about the susceptibility of the Atdpl1 mutant to B. cinerea or Pst (either virulent [Pst DC3000] or avirulent [Pst AvrRPM1] strain), three independent Atdpl1 mutant lines were challenged with these pathogens. The three Atdpl1 mutant lines displayed similar responses upon pathogen challenge (Fig. 1). In B. cinerea-infected wild-type plants, disease symptoms, showing chlorosis and necrosis, increased more rapidly than in B. cinerea-infected Atdpl1 plants (Fig. 1A). On the contrary, symptoms developed in response to Pst infection were more pronounced in mutant plants than in wild-type plants (Fig. 1A). The lesion diameters were scored 48 and 60 h after drop inoculation with B. cinerea and classified into size categories (Fig. 1B). Interestingly, Atdpl1 plants did not display necrotic lesions of the largest size, whereas wild-type plants showed 10% of these lesions 48 h post inoculation (hpi). Only 2% of the largest lesions were observed in Atdpl1 plants compared with 12% for wild-type plants 60 hpi. Furthermore, Atdpl1 mutants displayed a greater percentage of small necrotic lesions than wild-type plants. Atdpl1 lines displayed approximately 45% and 65% of small lesions, whereas wild-type plants showed only 17% and 24% of small lesions 48 and 60 hpi, respectively. Consequently, fewer lesions of medium size were observed in Atdpl1 lines than in wild-type plants (Fig. 1B).

Figure 1.

Atdpl1 mutants are more tolerant to B. cinerea but more susceptible to Pst than the wild type. B. cinerea conidia suspension was deposited by using drop inoculation (A and B) or spray inoculation (E) on leaves of wild-type (WT) and Atdpl1 mutant plants. Pst solution was infiltrated into wild-type and Atdpl1 mutant leaves (A, C, and D). A, Photographs represent disease symptoms observed 60 or 72 h after infection by the fungus or Pst, respectively. B, Symptoms due to B. cinerea infection were scored by defining three lesion diameter (d; in mm) classes. Statistical differences of the mean lesion diameters between wild-type and Atdpl1 plants were calculated with a Kruskal-Wallis test: **, P < 0.01; and ***, P < 0.005. C and D, Bacterial growth of virulent Pst strain DC3000 (C) and avirulent Pst strain AvrRPM1 (D) at 0, 6, 24, 48, and 54 hpi. E, B. cinerea and Pst growth was quantified by qRT-PCR 3 and 48 h after pathogen infection in leaves of wild-type and Atdpl1 mutant plants. Asterisks indicate significant differences between wild-type and Atdpl1 samples according to Student’s t test: ***, P < 0.005. Results are representative of three independent experiments.

The average lesion diameter in the Atdpl1 mutant was significantly lower than that in wild-type plants (**, P < 0.01 and ***, P < 0.005; Fig. 1B). Plants were also infiltrated with Pst DC3000 or Pst AvrRPM1 at 10⁷ colony-forming units (cfu) mL⁻¹, and bacterial populations were evaluated 0, 6, 24, 30, 48, and 54 hpi. As already described, avirulent strain growth was less important compared with the virulent strain in wild-type plants (Fig. 1, C and D). Interestingly, infection with both bacterial strains revealed an increased susceptibility of Atdpl1 plants, allowing about 10-fold more bacterial growth as compared with wild-type plants (Fig. 1, C and D). These results were also correlated by fungal and bacterial population quantification in infected leaves by quantitative reverse transcription (qRT)-PCR (Fig. 1E). Interestingly, the AtDPL1 expression profile was similar after either B. cinerea or Pst infection (Supplemental Fig. S1). Until 12 hpi, no AtDPL1 transcript accumulation could be observed. AtDPL1 expression increased significantly between 12 and 24 hpi and rose continuously until the later stages of infection. Symptoms due to either B. cinerea invasion or infection with the virulent or avirulent strain of Pst visually appeared between 24 and 30 hpi (data not shown) and thus are delayed slightly compared with AtDPL1 expression. Deregulation of photosynthesis is considered a tool for evaluating the first sign of pathogen infection (Berger et al., 2007; Bolton, 2009). Repression of the RbcS gene (encoding the small subunit of ribulose-1,5-bisphosphate carboxylase) after pathogen infection occurred at the same time (B. cinerea) or slightly earlier (Pst) compared with AtDPL1 expression and symptom appearance (Supplemental Fig. S1), suggesting that an immediate consequence of pathogen perception includes the induction of AtDPL1 gene expression. Collectively, these data indicate that lack of AtDPL1 activity in mutant plants significantly delays the development of lesions triggered by B. cinerea infection but renders plants more susceptible to Pst infection.

Sphingolipid Profiles in Wild-Type and Atdpl1-1 Plants Are Affected Differently upon Pathogen Infection

To determine whether changes in the levels of certain sphingolipids are responsible for the delayed development of B. cinerea infection in the Atdpl1 mutant, sphingolipid profiles were analyzed. The main sphingolipid species in Arabidopsis, LCBs and LCB-Ps (Fig. 2), glycosylinositol phosphoceramides (GIPCs; Fig. 3), Cers (Fig. 4), hydroxyceramides (hCers; Fig. 5), and glucosylceramides (GlcCers; Supplemental Fig. S2), were first quantified in both the wild type and the Atdpl1-1 mutant at 0 hpi (Supplemental Fig. S3). In wild-type and Atdpl1-1 plants, LCB/LCB-P basal levels were almost in the same range as those already described (Tsegaye et al., 2007; Supplemental Fig. S3). As described previously, the only significant alteration in sphingolipid basal levels observed in the Atdpl1-1 mutant compared with the wild type under typical growth conditions was an increase in one specific LCB-P (4-hydroxy-8-sphingenine-1-phosphate [t18:1-P]; Tsegaye et al., 2007; Supplemental Fig. S3). Next, we investigated the influence of B. cinerea infection on the sphingolipid profile in wild-type plants. B. cinerea infection triggered LCB accumulation (from 6× for 4,8-sphingadienine [d18:2] to 20× for d18:0; Fig. 2A) but also a moderate increase in sphingosine-1-phosphate (d18:1-P) and t18:1-P amount (4× and 2.5×, respectively) compared with mock-inoculated wild-type plants (Fig. 2E). The amount of total GIPCs and, more precisely, saturated α-hydroxylated very-long-chain fatty acid (VLCFA)-containing GIPCs (C24 and C26; Fig. 3, A and C) was significantly lower after B. cinerea infection than in mock-treated plants (200 and 300 nmol g⁻¹ dry weight, respectively; Supplemental Fig. S4). Moreover, d18:0-, d18:1-, and t18:1-GIPC levels were also reduced after B. cinerea infection (Fig. 3, A and C). The amount of total Cers is 4 times higher in B. cinerea- than in mock-inoculated wild-type plants (84 versus 21 nmol g⁻¹ dry weight; Supplemental Fig. S4). Most Cer molecules were affected by the presence of B. cinerea (Fig. 4, A and C). Finally, the level of total hCers was not modified (Supplemental Fig. S4); however, significant accumulation of saturated α-hydroxylated C16-, C18-, and C26-containing hCers and d18:0-hCer was observed after challenge with B. cinerea (Fig. 5, A and C). No change could be noticed in GlcCer levels (Supplemental Figs. S2 and S4).

Figure 2.

Free LCB and LCB-P accumulation after challenge with pathogen. Leaves of wild-type (WT) or Atdpl1-1 mutant plants were sprayed with B. cinerea spore suspension (Bc) or potato dextrose broth (PDB; Control; A, B, E, and F) or infiltrated with Pst DC3000, Pst AvrRPM1, or MgCl₂ (Control; C, D, G, and H). Quantifications of LCBs (A–D) and LCB-Ps (E–H) were performed 48 hpi. Asterisks on wild-type bars indicate significant differences between the pathogen-treated wild-type sample and the control sample, and asterisks on Atdpl1-1 bars indicate significant differences between the pathogen-treated wild-type sample and the pathogen-treated Atdpl1-1 sample according to Student’s t test: *, P < 0.05; **, P < 0.01; and ***, P < 0.005. Results are means of four to five independent biological experiments ± sd. Notice the different scale of LCB-P levels between wild-type and Atdpl1-1 plants. DW, Dry weight.

Figure 3.

GIPC contents after B. cinerea or Pst infection. Leaves of wild-type (WT; left) or Atdpl1-1 mutant (right) plants were sprayed with PDB (Control; A and B) or B. cinerea spore suspension (Bc; C and D) or infiltrated with MgCl₂ (Control; E and F), Pst DC3000 (G and H), or Pst AvrRPM1 (I and J). Quantifications were performed 48 hpi. Asterisks on wild-type bars indicate significant differences between the pathogen-treated wild-type sample and the control sample, and asterisks on Atdpl1-1 bars indicate significant differences in total species between the pathogen-treated wild-type sample and the pathogen-treated Atdpl1-1 sample according to Student’s t test: *, P < 0.05; **, P < 0.01; and ***, P < 0.005. Asterisks have only been considered for the total species displaying the same fatty acid or hydroxylation/unsaturation degree. Results are means of four to five independent biological experiments ± sd. DW, Dry weight.

Figure 4.

Cer species produced by wild-type and Atdpl1-1 mutant plants upon pathogen infection. Leaves of wild-type (WT; left) or Atdpl1-1 mutant (right) plants were sprayed with PDB (Control; A and B) or B. cinerea spore suspension (Bc; C and D) or infiltrated with MgCl₂ (Control; E and F), Pst DC3000 (G and H), or Pst AvrRPM1 (I and J). Quantifications were performed 48 hpi. Asterisks on wild-type bars indicate significant differences between the pathogen-treated wild-type sample and the control sample, and asterisks on Atdpl1-1 bars indicate significant differences between the pathogen-treated wild-type sample and the pathogen-treated Atdpl1-1 sample according to Student’s t test: *, P < 0.05; **, P < 0.01; and ***, P < 0.005. Asterisks have only been considered for the total species displaying the same fatty acid or hydroxylation/unsaturation degree. Results are means of four to five independent biological experiments ± sd. DW, Dry weight.

Figure 5.

hCer species produced by wild-type and Atdpl1-1 mutant plants upon pathogen infection. Leaves of wild-type (WT; left) or Atdpl1-1 mutant (right) plants were sprayed with PDB (Control; A and B) or B. cinerea spore suspension (Bc; C and D) or infiltrated with MgCl₂ (Control; E and F), Pst DC3000 (G and H), or Pst AvrRPM1 (I and J). Quantifications were performed 48 hpi. Asterisks on wild-type bars indicate significant differences between the pathogen-treated sample and the control sample, and asterisks on Atdpl1-1 bars indicate significant differences between the pathogen-treated wild-type sample and the pathogen-treated Atdpl1-1 sample according to Student’s t test: *, P < 0.05; **, P < 0.01; and ***, P < 0.005. Asterisks have only been considered for the total species displaying the same fatty acid or hydroxylation/unsaturation degree. Results are means of four to five independent biological experiments ± sd. DW, Dry weight.

To better understand the role of sphingolipids in plant resistance to the necrotrophic fungus, a comparison between sphingolipid profiles in B. cinerea-infected Atdpl1-1 mutant and wild-type plants was then performed. With respect to the LCB/LCB-P pool, wild-type plants contained more LCBs (Supplemental Fig. S4), especially d18:0 and d18:2 (Fig. 2, A and B), whereas the Atdpl1-1 mutant accumulated more LCB-Ps (Supplemental Fig. S4), especially t18:0-P and t18:1-P (9- and 18-fold, respectively), when compared with wild-type plants (Fig. 2, E and F). The amount of total GIPCs and, more precisely, saturated α-hydroxylated VLCFA-containing GIPCs (C22, C24, and C26; Fig. 3, C and D) was significantly higher in Atdpl1-1 mutant than in wild-type plants after B. cinerea infection (370 versus 220 nmol g⁻¹ dry weight, respectively; Supplemental Fig. S4). Total Cer amount was similar in both types of plants (Fig. 4, C and D; Supplemental Fig. S4), but B. cinerea infection triggered an increase in hCer contents, especially saturated and monounsaturated VLCFA-containing hCers (Fig. 5, C and D), in Atdpl1-1 mutant compared with wild-type plants (75 versus 27 nmol g⁻¹ dry weight, respectively; Supplemental Fig. S4). Moreover, trihydroxy-hCers also accumulated three times in the mutant compared with wild-type plants in response to the fungus (Fig. 5, C and D). No significant change was observed in total GlcCer amount (Supplemental Figs. S2 and S4).

In order to compare sphingolipid profiles in response to a hemibiotrophic pathogen, analyses were performed 48 h after wild-type plant inoculation with avirulent or virulent strains of Pst. Our data confirmed previous results showing that sphingolipid increase was more sustained during the incompatible than the compatible interaction (Peer et al., 2010). Increase in t18:0 was observed in response to both types of bacteria, but infection with only Pst AvrRPM1 triggered a significant decrease of d18:1 (Fig. 2C). After infection with Pst AvrRPM1, an increase in d18:2-P, t18:0-P, and t18:1-P was observed, whereas only t18:0-P level was increased in response to Pst DC3000 (Fig. 2G). GIPC levels were also not significantly modified in response to both types of bacteria (Fig. 3, E, G, and I; Supplemental Fig. S4). Total contents of d18:0-, d18:1-, t18:0-, and t18:1-Cers were increased after infection with Pst AvrRPM1 (Fig. 4, E and I). Only an increase in trihydroxy-Cers could be noticed in response to Pst DC3000 (Fig. 4, E and G). Moreover, the t18:0-Cer level was higher in the case of the incompatible interaction than in the case of the compatible one (40 versus 24 nmol g⁻¹ dry weight, respectively; Supplemental Fig. S4). C16-, C24-, and C26-Cers also accumulated in response to both strains of Pst (Fig. 4, E, G, and I), and only C16-Cer accumulation was more pronounced in the case of interaction with Pst AvrRPM1 compared with Pst DC3000 (45 versus 18 nmol g⁻¹ dry weight, respectively; Fig. 4, E, G, and I). Total contents of d18:0-hCers were increased in response to Pst (Fig. 5, E, G, and I). t18:0-hCers accumulated after challenge with the virulent strain and t18:1-hCers after challenge with the avirulent strain (Fig. 5, E, G, and I). Similar to B. cinerea infection, no regulation of GlcCer content could be noticed (Supplemental Figs. S2 and S4). Comparison of sphingolipid profiles between Pst-infected wild-type and Atdpl1-1 mutant plants revealed an increase in d18:0 (1.5×) in Atdpl1-1 plants certainly due to infiltration, since it was also observed in control plants. An increase in t18:0-P level (5×) was detected in Atdpl1-1 mutant plants compared with the wild type only in response to the avirulent strain (Fig. 2H). No significant regulation of GIPC, Cer, hCer, or GlcCer pools was observed in response to either the virulent or avirulent strain (Figs. 3–5; Supplemental Fig. S2).

Changes in Sphingolipid Profiles Affect Pathogen-Induced Cell Death

Recently, several reports have revealed that some sphingolipids are important players in HR and associated PCD (Berkey et al., 2012; Markham et al., 2013). HR is an effective strategy of plants to protect themselves against (hemi)biotrophic microorganisms (Coll et al., 2011). In contrast, PCD processes promote the spread of necrotrophic pathogens such as B. cinerea (Govrin and Levine, 2000; Govrin et al., 2006). Thus, changes in sphingolipid profiles and differences in tolerance upon B. cinerea or Pst infection prompted us to examine the cell death response upon pathogen attack. We thus measured electrolyte leakage to detect changes in loss of ions caused by plasma membrane damage characteristic of plant cell death (Dellagi et al., 1998; Kawasaki et al., 2005). Ion leakage measured after the inoculation of Atdpl1-1 plants with B. cinerea or Pst was reduced compared with wild-type plants (Fig. 6). These results suggested that modification in sphingolipid content could play a role in modulating cell death processes in response to pathogen infection.

Figure 6.

Electrolyte leakage in Atdpl1-1 mutants after pathogen inoculation. Conductivity (μS cm⁻¹) is shown for solution containing leaf discs from either the wild type (WT) or the Atdpl1-1 mutant inoculated with B. cinerea (Bc) or PDB (Control) solution (A) or Pst DC3000, Pst AvrRPM1, or 10 mm MgCl₂ (B). Each value represents the mean ± sd of three replicates per experiment. The experiment was repeated three times with similar results.

Expression levels of PCD marker genes, such as the flavin-containing monooxygenase FMO and the senescence-associated genes SAG12 and SAG13 (Brodersen et al., 2002), were also evaluated in order to verify if cell death responses are modified in Atdpl1-1 mutant plants (Fig. 7). FMO and SAG13 were induced in both types of plants with increasing infection spread of B. cinerea. Interestingly, these inductions occurred earlier and stronger in the wild type (between 12 and 24 hpi) than in the Atdpl1-1 mutant (between 24 and 30 hpi; Fig. 7, A and C). SAG12 was only induced 48 hpi in both the wild type and the Atdpl1-1 mutant, and similar to SAG13 and FMO, its expression was stronger in the wild type (10,000×) than in the Atdpl1-1 mutant (2,000×; Fig. 7E).

Figure 7.

Time course of PCD marker gene expression after B. cinerea or Pst infection. Leaves were sprayed with B. cinerea spore suspension (Bc) or PDB (Control; A, C, and E) or infiltrated with a bacterial solution (Pst DC3000 or Pst AvrRPM1) or MgCl₂ (Control; B, D, and F). The mean values ± sd from one representative experiment are shown. qRT-PCR of FMO (A and B), SAG13 (C and D), and SAG12 (E and F) expression was performed in wild type (WT) and Atdpl1-1 mutant plants with five biological replicates with comparable results.

As expected in the case of Pst infection, SAG13 and FMO gene expression were induced earlier and stronger during the incompatible interaction than during the compatible interaction (Fig. 7, B and D). Wild-type and mutant plants displayed similar expression profiles with both types of bacteria; however, induction was less pronounced in Atdpl1-1 mutant plants. Similar to B. cinerea infection, SAG12 transcript accumulation occurred only at the later stages of infection (Fig. 7F). It is noteworthy that induction of these PCD marker genes followed a pattern similar to AtDPL1 gene expression in wild-type plants in response to either B. cinerea or Pst infection (Supplemental Fig. S1, A and B).

SAG12 is only expressed in senescent tissues. In contrast, SAG13 and FMO are expressed in different PCD processes (Lohman et al., 1994; Brodersen et al., 2002). Collectively, our data suggest that the induction of SAG13 and FMO after either B. cinerea or Pst infection could result from an HR-like PCD, whereas a senescence program is activated later. This could also explain the tolerance of Atdpl1 mutant plants toward B. cinerea and their higher susceptibility toward Pst.

Modification of Sphingolipid Contents Affect ROS Production in Response to Pathogen Infection

Transient production of ROS is a hallmark of successful pathogen recognition (Torres, 2010). To investigate whether sphingolipid content perturbation in Atdpl1-1 plants affected pathogen recognition, we compared ROS production in the mutant versus wild-type plants. Wild-type plants displayed a transient oxidative burst, peaking around 300 min (B. cinerea) or 40 min (Pst) after inoculation (Fig. 8). This transient burst was significantly induced by 2.5 times in B. cinerea-infected Atdpl1-1 plants compared with wild-type plants (Fig. 8A). On the contrary, ROS levels were significantly reduced in Pst-infected Atdpl1-1 plants compared with Pst-infected wild-type plants (Fig. 8, B and C). Our results thus demonstrated that signaling events linked to pathogen recognition are affected by sphingolipid perturbation in Atdpl1-1 plants.

Figure 8.

Transient ROS production in response to pathogen infection in wild-type and Atdpl1-1 mutant plants. The time course of ROS production in wild type (WT) and Atdpl1-1 mutant plants is shown in response to B. cinerea (Bc; A), Pst DC3000 (B), or Pst AvrRPM1 (C) infection. Leaf discs were immersed in a solution containing either 10⁵ spores mL⁻¹  B. cinerea or 10⁸  cfu mL⁻¹  Pst. Error bars represent se from 12 biological repetitions. Three independent experiments were performed with similar results. RLUs, Relative light units.

Exogenous t18:0-P and d18:0 Differently Modify Pathogen-Induced Cell Death and ROS Production

Major changes in LCB-P contents in B. cinerea-inoculated Atdpl1-1 mutant plants are an increase in t18:0-P levels and a decrease in d18:0 amounts (Fig. 2). We thus tested the ability of these sphingolipids to modulate pathogen-induced cell death (Fig. 9) and ROS production (Fig. 10). Our data showed that exogenous t18:0-P or d18:0 alone did not affect cell death or ROS production, a finding consistent with data obtained by Coursol et al. (2015). In t18:0-P-treated wild-type plants, symptoms and ion leakage triggered by B. cinerea or Pst infection were reduced significantly (Fig. 9, A, C, and E). Exogenously applied d18:0 did not modify disease symptoms and electrolyte leakage in wild type-infected plants by B. cinerea and slightly reduced electrolyte leakage triggered by the virulent Pst strain (Fig. 9, B, D, and F). Interestingly, disease symptoms and electrolyte leakage were strongly reduced when wild-type plants were coinfiltrated with d18:0 and Pst AvrRPM1 (Fig. 9, B and F).

Figure 9.

Exogenous effects of t18:0-P and d18:0 on electrolyte leakage in response to pathogen infection in wild-type plants. A and B, B. cinerea conidia suspension was deposited on leaves of wild-type and Atdpl1-1 mutant plants 15 min after infiltration of either t18-0-P or d18:0 solution. Pst and either t18-0-P or d18:0 solution were coinfiltrated into wild-type and Atdpl1-1 leaves. Photographs represent symptoms observed 60 or 72 h after infection by the fungus or Pst, respectively. C to F, Conductivity (μS cm⁻¹) of solution containing t18:0-P- or d18:0-infiltrated leaf discs from the wild type inoculated by spraying B. cinerea (Bc) or PDB (Control) solution (C and D) or by infiltration of Pst DC3000, Pst AvrRPM1, or 10 mm MgCl₂ (E and F). Each value represents the mean ± sd of three replicates per experiment. The experiment was repeated three times with similar results.

Figure 10.

Exogenous effects of t18:0-P and d18:0 on ROS production in response to pathogen infection in wild-type plants. The time course of ROS production in t18:0-1-P- or d18:0-treated wild-type plants is shown in response to B. cinerea (Bc; A and D), Pst DC3000 (B and E), or Pst AvrRPM1 (C and F) infection. Leaf discs were immersed in a solution containing 100 µm t18:0-1-P or d18:0 and either 10⁵ spores mL⁻¹  B. cinerea or 10⁸  cfu mL⁻¹  Pst. Error bars represent se from 12 biological repetitions. Three independent experiments were performed with similar results. RLUs, Relative light units.

Whereas the addition of t18:0-P increased and delayed ROS production upon challenge with B. cinerea, it reduced the Pst-induced oxidative burst (Fig. 10, A–C). d18:0 had no significant effect on ROS accumulation triggered by B. cinerea (Fig. 10D). However, it dramatically reduced the Pst-induced oxidative burst (Fig. 10, E and F). These data indicate that exogenously applied t18:0-P and d18:0 modify signaling events and cell death triggered by infection with these two pathogens.

SA and ET/JA Signaling Pathways Are Modified in Atdpl1-1 Mutant Plants after Pathogen Challenge

Disruption of sphingolipid contents between wild-type and Atdpl1 plants could result in the differential activation of defense responses after pathogen infection. PR1 and PR5 are well-known SA-dependent defense marker genes, NONEXPRESSED PATHOGEN RELATED1 (NPR1) was shown to be a key regulator of SA-mediated suppression of JA signaling (Spoel et al., 2003), and DEFENSIN (PDF1.2), CHITINASE (CHIT), and ETHYLENE RESPONSE FACTOR1 (ERF1) expression is regulated by JA and ET whereas VEGETATIVE STORAGE PROTEIN1 (VSP1) and JASMONATE ZIM-DOMAIN8 (JAZ8) are mostly responsive to JA (Glazebrook, 2005; Pieterse et al., 2009). First, the expression pattern of these defense genes was monitored in wild-type and Atdpl1-1 mutant plants. No significant difference in the expression of these defense genes was detected in wild-type and Atdpl1-1 mutant plants grown under standard conditions (Figs. 11 and 12). These results indicated that inactivation of the gene encoding LCB-P lyase itself did not result in any defense response changes in plants. The expression levels of defense-related genes in Atdpl1-1 mutant plants were then compared with wild-type plants in response to B. cinerea infection (Fig. 11). Whereas PR1, PR5, NPR1, and VSP1 expression showed similar induction levels in both genotypes, the expression of PDF1.2, CHIT, ERF1, and JAZ8 was markedly enhanced in Atdpl1-1 mutant compared with wild-type plants. At 48 hpi, there was a 12-fold increase for PDF1.2 and a 2-fold increase for CHIT, ERF1, and JAZ8 compared with wild-type plants (Fig. 11). Since JA-responsive genes were up-regulated in the Atdpl1-1 mutant, the expression of three genes encoding key enzymes in JA biosynthesis, LIPOXYGENASE2 (LOX2), ALLENE OXIDE CYCLASE2 (AOC2), and 12-OXOPHYTODIENOIC ACID REDUCTASE3 (OPR3) as well as JAR1, encoding the enzyme that converts JA to the jasmonoyl-isoleucine (JA-Ile) conjugate (Staswick and Tiryaki, 2004), was also followed. The results showed that LOX2 and AOC2 were significantly up-regulated up to 24 hpi in the Atdpl1-1 mutant but transcripts returned to a level comparable to the wild type thereafter (Fig. 11). In contrast, the expression of OPR3 was similar in both genotypes. JAR1 expression was not affected by the fungus inoculation (Fig. 11). These results indicated that both JA synthesis and signaling pathways were enhanced in Atdpl1-1 mutant plants.

Figure 11.

Expression levels of JA and SA pathway-associated genes in wild-type (WT) and Atdpl1-1 mutant plants during B. cinerea (Bc) infection. Results are expressed as the fold increase in transcript level compared with the untreated control (0 h), referred to as the 1× expression level. Values shown are means ± sd of duplicate data from one representative experiment among five independent repetitions.

Figure 12.

Expression levels of JA and SA pathway-associated genes in wild-type (WT) and Atdpl1-1 mutant plants during Pst infection. Results are expressed as the fold increase in transcript level compared with the untreated control (0 h), referred to as the 1× expression level. Values shown are means ± sd of duplicate data from one representative experiment among five independent repetitions.

When infected with Pst, wild-type plants displayed a strong induction of PR1 expression, and, as expected, this induction was more pronounced (4-fold at 48 hpi) in the case of the incompatible interaction (Fig. 12). Surprisingly, a significant repression of this gene was observed 30 hpi in the Atdpl1-1 mutant compared with wild-type plants (6× for Pst DC3000 and 4× for Pst AvrRPM1), but the level of PR1 expression was still higher in incompatible compared with compatible interactions. Accumulation of PR5 transcripts was also slightly more important in wild-type plants, but expression levels were more important in the case of the compatible interaction (Fig. 12). Under Pst attack, NPR1 was slightly induced, but no difference between the wild type and the Atdpl1-1 mutant was observed. CHIT expression was also more induced in response to Pst AvrRPM1 (90×) than Pst DC3000 (20×) in wild-type plants, and this induction profile was similar in Atdpl1-1 mutant plants (Fig. 12). As already described, inoculation with the bacterial pathogen (Pst DC3000 or Pst AvrRPM1) led to a dramatic repression of PDF1.2 expression, either in wild-type or in Atdpl1-1 mutant plants. In contrast, ERF1 and JAZ8 were induced during Pst infection, and VSP1 expression was slightly induced when challenged by Pst DC3000 but repressed after Pst AvrRPM1 infection (Fig. 12). Expression of these three genes was markedly enhanced in the Atdpl1-1 mutant compared with wild-type plants. At the end of the time course, VSP1, ERF1, and JAZ8 mRNA levels were 2-, 3-, and 6-fold higher in Atdpl1-1 than in wild-type plants after infection with either virulent or avirulent strains, respectively. Similar to B. cinerea infection, JAR1 expression was not affected by inoculation with Pst. Regarding genes involved in the JA biosynthetic pathway, LOX2 was repressed, AOC2 was not induced during Pst challenge, and OPR3 was induced slightly, but no difference between the two genotypes was observed (Fig. 12). These data suggested that only the JA signaling pathway is positively affected in mutant plants upon challenge with Pst.

To get further information on Atdpl1-1 mutant defense responses, some defense-related phytohormones were also quantified (Fig. 13). No change in phytohormone basal levels was observed between wild-type and Atdpl1-1 mutant plants (Fig. 13). This implied that Atdpl1-1 mutant plants, in contrast to other mutants with modified sphingolipid contents, do not display high constitutive SA amounts (Greenberg et al., 2000; Wang et al., 2008; Ternes et al., 2011; König et al., 2012). Following pathogen attack, all phytohormone levels were enhanced. SA accumulation was essentially unchanged in the mutant compared with wild-type plants, whatever the pathogen considered. Interestingly, levels of JA and its biologically active conjugate, JA-Ile, were 2 to 3 times higher in the Atdpl1-1 mutant compared with wild-type plants after B. cinerea or Pst infection, respectively. However, no difference in JA levels between virulent and avirulent interaction was noticed, but JA-Ile accumulation was slightly higher in the case of the avirulent interaction in Atdpl1-1 plants. Together, our data suggest that the JA-dependent signaling pathway is preferentially activated in the Atdpl1-1 mutant in response to pathogen infection.

Figure 13.

Analysis of phytohormone accumulation in stressed wild-type and Atdpl1-1 mutant plants. JA, JA-Ile, and SA accumulation is shown in wild-type (WT) and Atdpl1-1 mutant plants 0 or 30 h following B. cinerea (A) or Pst DC3000 or Pst AvrRPM1 (B) infection. Asterisks indicate significant differences between wild-type and Atdpl1-1 samples according to Student’s t test: *, P < 0.05; **, P < 0.01; and ***, P < 0.005. Values shown are means ± sd from one representative experiment among five independent repetitions. FW, Fresh weight.

DISCUSSION

Only a few articles have described a connection between sphingolipid content, PCD, and defense reactions during biotic stress (Berkey et al., 2012). Furthermore, most of them focused on responses against (hemi)biotrophic pathogens, the role of sphingolipid in plant defense against necrotrophs being largely unsolved (Rivas-San Vicente et al., 2013; Bi et al., 2014). Moreover, nearly all studies revealed basal sphingolipid levels, and data of sphingolipid contents during pathogen infection were often not available (Peer et al., 2010; Bi et al., 2014). This work describes a comparison of sphingolipid content during hemibiotrophic and necrotrophic infection. In this study, we investigated the consequences of the disruption of the sphingolipid profiles on plant immunity responses such as cell death, ROS production, and signaling of plant defense responses during pathogen infection.

Interplays between Sphingolipids and PCD

As in animal systems, new emerging evidence showed that bioactive sphingolipids play a critical role as modulators of plant PCD (Berkey et al., 2012; Saucedo-García et al., 2015). Here, sphingolipid content analyses showed that infection by B. cinerea or Pst triggered the accumulation of some species known to act in favor of cell survival (LCB-Ps and hCers) or cell death (LCBs and Cers). Interestingly, the Atdpl1-1 mutant displayed higher levels of d18:0 in response to infiltration (Fig. 2). Moreover, this LCB reduced Pst-induced cell death and symptoms, especially in the case of the incompatible interaction (Fig. 9). Since the HR often contributes to resistance to (hemi)biotrophic pathogens, our results suggested that a modification in d18:0 levels could impact plant cell death and, thus, resistance to such pathogens. Recently, Coursol et al. (2015) showed that the addition of d18:0 had no significant effect on the viability of cryptogein-treated cells, indicating that distinct mechanisms of regulation are involved in cell death of cell cultures or plant tissue or after treatment by an elicitor or a pathogen. Necrotrophs are pathogens that derive nutrients from dead or dying cells. PCD, including HR, can be beneficial to this kind of pathogen and, thus, could facilitate their infection and spread of disease (Govrin and Levine, 2000; Mayer et al., 2001; Govrin et al., 2006). Plants that are less potent to activate HR or with reduced cell death present enhanced tolerance to B. cinerea infection and vice versa (Govrin and Levine, 2000; van Baarlen et al., 2007). Similarly, antiapoptotic genes conferred resistance to necrotrophic fungi in transgenic plants (Dickman et al., 2001; El Oirdi and Bouarab, 2007). A general pattern established that infection of Arabidopsis by B. cinerea is promoted by and requires an active cell death program in the host (van Kan, 2006), and resistance against this fungus depends on the balance between cell death and survival (van Baarlen et al., 2007).

Interestingly, the Cer-accumulating acd5 mutant or Cer-infiltrated plants were more susceptible to several Botrytis spp. (van Baarlen et al., 2004, 2007). Moreover, myriocin, a potent inhibitor of SERINE PALMITOYLTRANSFERASE (SPT), the first enzyme of sphingolipid biosynthesis, had a death-antagonistic effect during the Botrytis  elliptica-Lilium interaction (van Baarlen et al., 2004). This suggests that sphingolipid metabolism is involved in cell death triggered by Botrytis spp. Cell death activation could thus be disturbed in Atdpl1 plants, leading to a higher susceptibility toward (hemi)biotrophs and higher tolerance toward necrotrophs. In this work, B. cinerea infection triggered Cer and LCB accumulation in wild-type plants. It is thus possible that the necrotrophic fungus promoted plant PCD-inducing factors (e.g. sphingolipids) in order to facilitate its penetration and spread inside plant cells. However, exogenous d18:0 did not modify ion leakage in the presence of B. cinerea, suggesting that this LCB alone is not involved in such a mechanism. Sphingolipid analysis revealed that B. cinerea-infected Atdpl1-1 plants accumulated more VLCFA-hCers and t18:0-P and t18:1-P but fewer Cers and LCBs compared with wild-type plants. Interestingly, our data showed that exogenous t18:0-P reduced B. cinerea- and Pst-induced cell death (Fig. 9). Thus, t18:0-P appears to be essential to modulate plant cell death and, thus, plant resistance in response to pathogen infection. Moreover, it was recently demonstrated that AtFAH1 or 2-hydroxy VLCFAs, and thereby VLCFA-hCers, were key factors in BAX INHIBITOR1-mediated cell death suppression (Nagano et al., 2012). These results confirmed that sphingolipids play an important role in plant defense responses and that the plant is able to adjust its response by regulating a dynamic balance between cell death (e.g. HR)- or cell survival-related sphingolipids. However, in contrast to infected Atdpl1 mutant, the fah1/fah2 double mutant presented a reduced amount of hCers and elevated levels of Cers and LCBs but showed no lesion phenotype (König et al., 2012). Thus, it seems that the connection between sphingolipids and PCD is regulated by a fine-tuned process and, thus, could be more complex than expected. Other parameters, such as defense signaling pathways, could be involved in such a mechanism.

Interconnections between Sphingolipids and Defense Mechanisms

Sphingolipids (e.g. LCBs and Cers) participate in the induction and/or control of plant cell death. Moreover, plant cell death processes, such as HR, are also associated with plant defense or disease. It is thus conceivable that some sphingolipids play key roles in plant innate immunity. Recent studies brought to light interconnections between sphingolipids and defense mechanisms. Resistance to biotrophic pathogens often required ROS production (Torres et al., 2002). Consistent with this, Pst-infected Atdpl1-1 mutant displayed a reduced accumulation of ROS and was more sensitive to the bacterial attack. In addition, Atdpl1-1 mutant accumulated more d18:0 in response to infiltration (Fig. 2), and d18:0 strongly reduced ROS production upon challenge with this bacterium (Fig. 10). B. cinerea-infected Atdpl1-1 plants displayed a higher production of ROS (Fig. 8). Several studies demonstrated that resistance against B. cinerea (and other necrotrophs) is accompanied by the generation of ROS, and mutants impaired in ROS production failed to resist the necrotrophic pathogen (Contreras-Cornejo et al., 2011; Kraepiel et al., 2011; L’Haridon et al., 2011; Rasul et al., 2012; Savatin et al., 2014; Zhang et al., 2014). It has been shown that LCBs but not LCB-Ps alone are able to induce ROS production (Peer et al., 2011). In this study, exogenously applied t18:0-P increased B. cinerea-induced ROS generation (Fig. 10). Accordingly, cryptogein-induced ROS accumulation is enhanced by a pretreatment with some LCB-Ps, especially t18:0-P (Coursol et al., 2015). This suggests that sphingolipids may interact differently with ROS production depending on the presence or not of an elicitor or pathogen. Interestingly, the similarity of ROS accumulation upon infection between Atdpl1-1 plants and t18:0-P- or d18:0-treated wild-type plants indicated that t18:0-P and d18:0 could have key roles in pathogen perception and, thus, in plant resistance toward hemibiotrophic and necrotrophic pathogens.

Several lines of evidence showed that plants disrupted in sphingolipid metabolism often displayed a spontaneous enhanced SA pathway (Greenberg et al., 2000; Brodersen et al., 2002; Wang et al., 2008; Ternes et al., 2011; König et al., 2012; Mortimer et al., 2013; Rivas-San Vicente et al., 2013; Wu et al., 2015). Recently, it was shown that SA and its analog benzothiadiazole affect sphingolipid metabolism (Shi et al., 2015), including AtDPL1 gene expression (Wang et al., 2006). Since activation of the SA-dependent pathway is effective against biotrophic and hemibiotrophic pathogens, it has been postulated that sphingolipids played a key role in defense against such pathogens in an SA-dependent pathway (Sánchez-Rangel et al., 2015). However, whereas acd5, erh1, and the double mutant fah1/fah2 exhibited enhanced resistance to powdery mildew, they displayed a similar phenotype to wild-type plants upon infection with P. syringae pv maculicola or Verticillium longisporum (Wang et al., 2008; König et al., 2012). This suggests that SA, sphingolipid-triggered cell death, and plant resistance could be independent regarding the plant/pathogen pair. Unfortunately, only basal levels of sphingolipid were described, and no sphingolipid quantification during pathogen infection is available, making difficult a direct link between sphingolipid metabolism and plant defense. In this work, infection with either necrotrophic or hemibiotrophic pathogen induced the production of all quantified phytohormones. It has been reported that several pathogens, including B. cinerea and Pst, activated both SA and JA accumulation (Zimmerli et al., 2001; Govrin and Levine, 2002; Schmelz et al., 2003; Spoel et al., 2003; Block et al., 2005; Glazebrook, 2005; Veronese et al., 2006), and cross talk is thus used by the plant to adjust its response in favor of the most effective pathway. Interestingly, SPT-silenced tobacco plants displayed higher basal SA levels and were more susceptible to A. alternata infection. However, no information concerning SA, JA, or sphingolipid levels in response to infection is available, especially as transgenic plants still displayed residual NbLCB2 gene expression (Rivas-San Vicente et al., 2013). In Arabidopsis, the acd5 mutant displayed constitutive high SA levels and expression of the PR1 gene. This mutant was also more susceptible to B. cinerea and contained higher Cer levels but reduced apoplastic ROS and PR1 and CHIT transcript accumulation upon infection (Greenberg et al., 2000; Bi et al., 2014). Consistent with this, Atdpl1-1 mutant plants displayed similar Cer levels and PR1 expression, higher apoplastic ROS accumulation, and CHIT up-regulation in response to infection but was more resistant to the necrotrophic fungus.

In Arabidopsis, it is now well admitted that SA has an antagonistic effect on JA signaling and reciprocally (Bostock, 2005; Glazebrook, 2005; Spoel et al., 2007; Thaler et al., 2012; Derksen et al., 2013). In tomato (Solanum lycopersicum), B. cinerea produces an exopolysaccharide that activates the SA pathway, which, through NPR1, antagonizes the JA signaling pathway, thereby allowing the fungus to enhance its disease (El Oirdi et al., 2011). Moreover, NPR1 needs to be activated by SA (Cao et al., 1998; Spoel et al., 2003). Here, SA accumulated in wild-type plants and NPR1 was also stimulated upon infection with B. cinerea. However, the SA signaling pathway was similar in Atdpl1-1 plants. Moreover, JA biosynthetic and signaling pathways were enhanced in the Atdpl1-1 mutant in response to B. cinerea inoculation. In the Atdpl1-1 mutant, it thus seems that perturbation in sphingolipid metabolism rendered either SA unable to activate NPR1 or NPR1 unable to antagonize JA accumulation. Thus, our results highlighted that disturbance of sphingolipid metabolism could impact not only the cell death program but also the JA signaling pathway, leading to plant tolerance toward necrotrophic pathogens such as B. cinerea. In that case, the relationship between sphingolipids and JA could be either indirect, implying that changes in sphingolipids operate in the cross talk between SA and JA pathways but in an NPR1-independent manner, or direct, as some key genes involved in JA biosynthesis are up-regulated in Atdpl1-1 plants. Similar to B. cinerea, a virulent strain of Pst, via its toxin coronatine, exerts its virulence by stimulating the JA signaling pathway in order to inhibit the SA signaling pathway and, thus, facilitate its growth and development (Zhao et al., 2003; Brooks et al., 2005; Laurie-Berry et al., 2006; Uppalapati et al., 2007; Geng et al., 2012; Zheng et al., 2012; Xin and He, 2013). The dramatic reduction in the expression of the SA-dependent marker gene PR1 in Atdpl1-1 mutant plants could thus be explained by the overaccumulation of jasmonates in these plants. Whereas the VSP1 and JAZ8 expression profile correlated the JA and JA-Ile accumulation profile in response to infection with virulent or avirulent Pst, PDF1.2 and CHIT expression did not. PDF1.2 and CHIT require both JA and ET signaling pathways but also the function of MITOGEN-ACTIVATED PROTEIN KINASE4 (MPK4), as JA-treated mpk4 mutants fail to express PDF1.2 (Petersen et al., 2000). A discrepancy between PDF1.2 expression and JA accumulation has also been observed during the induced systemic resistance triggered by Pseudomonas  fluorescens, which is regulated through the JA signaling pathway (van Wees et al., 1999). This suggested that a component in the JA or ET signaling pathway might be deficient/nonfunctional in Atdpl1-1 mutant plants in response to Pst infection or that defense against Pst in the Atdpl1-1 mutant might be regulated through a pathway that does not include PDF1.2 or CHIT. Collectively, our results suggested that AtDPL1 could be a negative and/or a positive regulator of the JA- and SA-regulated defense pathway, respectively. Whereas a relationship between SA signaling and sphingolipids has often been described (Sánchez-Rangel et al., 2015), our results highlight, to our knowledge for the first time, that sphingolipids could also play a key role in the JA signaling pathway.

In conclusion, we propose a model in which plant cells of the Atdpl1 mutant select the most appropriate response to defend themselves against pathogen attack by acting on sphingolipid metabolism in order to modulate the cell death/survival balance, in close cooperation with the JA and/or SA signaling pathways (Fig. 14). Whereas SA involvement in PCD is well known, the relationship between JA and cell death is less understood. Plants treated with coronatine, which shares structural similarities with JA-Ile and functional similarities with JA, develop chlorosis (Bender et al., 1999; Overmyer et al., 2003). Coronatine-deficient mutants of Pst DC3000 are reduced in disease-associated necrosis and chlorosis (Brooks et al., 2004, 2005). It has been reported that JA is also essential in FB1- and AAL-induced cell death (Asai et al., 2000; Zhang et al., 2011). Interestingly, the Atdpl1 mutant is more sensitive to FB1 treatment (Tsegaye et al., 2007). Thus, sphingolipid metabolism seemed to be intimately connected to defense processes to regulate plant responses to biotic stresses. In Arabidopsis, MPK6, which is involved in the plant defense response (Ren et al., 2008; Beckers et al., 2009), has recently been described as an important contributor to the LCB-mediated PCD (Saucedo-García et al., 2011). However, the deciphering of the precise pathway leading to sphingolipid-induced cell death is far from being totally elucidated. Further identification of target genes and their functions will provide new insights into how sphingolipids could be linked to cell death and defense processes.

Figure 14.

Schematic overview of interconnections between sphingolipid metabolism, cell death, and defense signaling pathways in Atdpl1 mutant plants upon pathogen attack. Upon disruption of the AtDPL1 gene, infected plants accumulate some LCB-P, hCer, and GIPC species, thus reducing cell death. In the Atdpl1 mutant, sphingolipid metabolism may also indirectly modulate cell death through its tight connection (double-headed dashed arrow) as a positive and/or negative regulator to jasmonate and/or SA signaling pathways, respectively. Reduced cell death and high levels of jasmonates could thus explain that Atdpl1 mutant plants are more tolerant to B. cinerea but more susceptible to Pst. T-bars indicate inhibition; single-headed arrows indicate activation; double-headed arrows indicate unknown regulatory mechanisms. Ald, Aldehyde; Ethan-P, phosphoethanolamine; LCBK, LCB kinase; LCB-P Pase, LCB-P phosphatase; LOH, LAG ONE HOMOLOG; SPHK1, SPHINGOSINE KINASE1.

MATERIALS AND METHODS

Chemicals

Phytosphingosine-1-phosphate (t18:0-P) and dihydrosphingosine (d18:0) were purchased from Avanti Polar Lipids. Stock solutions were prepared in ethanol:dimethyl sulfoxide (2:1, v/v; t18:0-P) or ethanol (d18:0) and dissolved to a final concentration of 100 µm. Luminol and horseradish peroxidase were obtained from Sigma-Aldrich.

Plant Material and Growth Conditions

Seeds of the Arabidopsis (Arabidopsis thaliana) SALK lines 020151 (referred to as Atdpl1-1), 093662 (Atdpl1-2), and 078119 (Atdpl1-3) containing a transfer DNA (T-DNA) insertion in the At1g27980 locus were obtained from the Nottingham Arabidopsis Stock Centre (http://arabidopsis.info). The SALK_020151 mutant was chosen for the performed experiments because it exhibits the same phenotype as the other Atdpl1 mutants but displays a complete lack of mRNA and a higher LCB/LCB-P accumulation in response to FB1 treatment (Tsegaye et al., 2007). Mutant and wild-type (Columbia-0) plants were grown and maintained under 12-h-light/12-h-dark conditions (150 μmol m⁻² s⁻¹, 20°C, and 60% humidity) for 35 d.

Isolation of the T-DNA Insertion Mutant and Genotype Characterization

The mutants SALK_020151, SALK_093662 (Tsegaye et al., 2007), and SALK_078119 were isolated according to the published procedure in SIGnAL (Alonso et al., 2003). The genotype of the knockout mutant line was analyzed by PCR using primers specific for the AtDPL1 gene (forward, 5′-AGAAAGGCCTCAAAGCTTGTC-3′; and reverse, 5′-TGCCAAATAGCATCATTCCTC-3′) and a primer specific for the T-DNA (LB1a, 5′-TGGTTCACGTAGTGGGCCATCG-3′).

Sphingolipidomic Analysis

Sphingolipid extraction from 10 to 20 mg of lyophilized tissue and profiling by liquid chromatography-electrospray ionization-tandem mass spectrometry were performed as described (Markham and Jaworski, 2007) with modifications, using the Shimadzu Prominence ultra-HPLC system and a 4000QTRAP mass spectrometer (AB Sciex). Sphingolipids were separated on a 100-mm Dionex Acclaim C18 column. Data analysis was performed using Analyst 1.6 and Multiquant 2.1 software (AB Sciex). Four to five biologically independent repeats were performed, and a minimum of three technical replicates were run from each sample.

RNA Extraction and Real-Time qRT-PCR

Isolation of total RNA and real-time PCR were performed as described by Le Hénanff et al. (2013). Gene-specific primers are described in Supplemental Table S1. For each experiment, PCR was performed in duplicate, and at least three independent experiments were analyzed. Transcript levels were normalized against those of the ACTIN gene as an internal control. Fold induction compared with mock-treated sample was calculated using the ƊƊCt method (Ct GI [unknown sample] – Ct GI [reference sample]) – (Ct actin [unknown sample] – Ct actin [reference sample]). GI is the gene of interest.

Pathogen Growth and Inoculation

Botrytis cinerea strain B05.10 was grown on solid tomato (Solanum lycopersicum) medium (25% [v/v] tomato juice and 2.5% [w/v] agar) during 21 d at 22°C. Collected conidia were resuspended in PDB supplemented by 0.02% (v/v) Silwet L-77 to a final density of 10⁵ conidia mL⁻¹. After incubation for 3 h at 22°C and 150 rpm, germinated spores were used for plant inoculation by spraying the upper face of the leaves. Control inoculations were performed with 0.02% PDB Silwet L-77.

The bacterial leaf pathogen Pseudomonas syringae pv tomato strain DC3000 or Pst AvrRPM1 was cultured overnight at 28°C in liquid King’s B medium, supplemented with rifampicin (50 µg mL⁻¹) and kanamycin (50 µg mL⁻¹). Subsequently, bacterial cells were collected by centrifugation and resuspended in 10 mm MgCl₂ to a final density of 10⁷  cfu mL⁻¹ (optical density = 0.01). The bacterial solutions were thus infiltrated from the abaxial side into leaf using a 1-mL syringe without a needle. Control inoculations were performed with 10 mm MgCl₂.

Leaves were collected from 0 to 48 hpi, frozen in liquid nitrogen, and stored at −80°C until use.

Pathogen Assay in Planta

B. cinerea infections were performed as described previously (Le Hénanff et al., 2013). Plants were placed in translucent boxes under high humidity at 150 µE m⁻² s⁻¹. Five or six leaves per plant were drop inoculated with 5 µL of the conidia suspension adjusted at 10⁵ conidia mL⁻¹ in PDB. Lesion diameters were measured 48 and 60 hpi. Forty to 60 leaves were inoculated per treatment and per genotype, and experiments were independently repeated four times.

Bacterial infections were performed as described previously (Sanchez et al., 2012). Briefly, eight foliar discs from four leaves were excised using a cork borer and ground in 1 mL of 10 mm MgCl₂ with a plastic pestle. Appropriate dilutions were plated on King’s B medium with appropriate antibiotics, and bacterial colonies were counted. Data are reported as means and sd of the log (cfu cm⁻²) of three replicates. Growth assays were performed four times with similar results.

Electrolyte Leakage

Ten minutes after bacteria injection (Torres et al., 2002) or 20 h after B. cinerea infection (Govrin and Levine, 2002), 9-mm-diameter leaf discs were collected from the infected area and washed extensively with water for 50 min, and then eight discs were placed in a tube with 15 mL of fresh water. To test the sphingolipid effect on ion leakage, pathogen inoculum was supplemented or not with 100 µm t18:0-P or d18:0. Conductivity measurements (three to four replicates for each treatment) were then conducted over time using a B-771 LaquaTwin (Horiba) conductivity meter.

ROS Production

Measurements of ROS production were performed as described previously (Smith and Heese, 2014). Briefly, single leaf disc halves were placed in wells of a 96-well plate containing 150 μL of distilled water and then incubated overnight at room temperature. Just before ROS quantification, distilled water was replaced by 150 μL of an elicitation solution containing 20 μg mL⁻¹ horseradish peroxidase and 0.2 μm luminol. For tests involving bacteria, Pst was added to the elicitation solution to a final bacterial concentration of 10⁸  cfu mL⁻¹. For tests involving B. cinerea, germinated spores were added to the elicitation solution to a final density of 10⁵ conidia mL⁻¹. For tests involving sphingolipids, 100 µm t18:0-P or d18:0 was added concomitantly with bacterium or fungus to the elicitation solution.

Phytohormone Analysis

Phytohormones were quantified using ultra-HPLC-tandem mass spectrometry according to Glauser et al. (2014).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Time course of AtDPL1 and RbcS expression after B. cinerea or Pst infection.

• Supplemental Figure S2. Glucosylceramide contents after B. cinerea or Pst infection.

• Supplemental Figure S3. Total content in major sphingolipid classes in WT and Atdpl1-1 mutant plants before infection with B. cinerea or Pst.

• Supplemental Figure S4. Total content in major sphingolipid classes in wild-type and Atdpl1-1 mutant plants after infection with B. cinerea or Pst.

• Supplemental Table S1. Gene-specific primers used in real-time reverse-transcription PCR.

ACKNOWLEDGMENTS

We thank Gaetan Glauser and Neil Villard (Neuchâtel Platform of Analytical Chemistry, University of Neuchâtel) for excellent technical assistance in phytohormone quantification.

Glossary

 
• ROS

  reactive oxygen species

 
• SA

  salicylic acid

 
• JA

  jasmonic acid

 
• ET

  ethylene

 
• HR

  hypersensitive response

 
• PCD

  programmed cell death

 
• FB1

  fumonisin B1

 
• AAL

  toxin produced by Alternaria alternata f. sp. lycopersici

 
• Cer

  ceramide

 
• LCB

  long-chain base

 
• LCB-P

  long-chain base phosphate

 
• hpi

  hours post inoculation

 
• cfu

  colony-forming units

 
• qRT

  quantitative reverse transcription

 
• GIPC

  glycosylinositol phosphoceramide

 
• hCer

  hydroxyceramide

 
• GlcCer

  glucosylceramide

 
• VLCFA

  very-long-chain fatty acid

 
• JA-Ile

  jasmonoyl-isoleucine

 
• T-DNA

  transfer DNA

 
• PDB

  potato dextrose broth

Author notes

1

This work was supported by the Region Champagne-Ardenne (grant no. A2101–03).

*

Address correspondence to [email protected].

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: Sandrine Dhondt-Cordelier ([email protected]).

[OPEN]

Articles can be viewed without a subscription.

S.D., C.C., F.B., and S.D.-C. designed the research; M.M.-R., D.L.B., J.M., and S.D.-C. performed the experiments; M.M.-R., D.L.B., J.M., and S.D.-C. analyzed the data; M.M.-R., D.L.B., J.M., S.D., C.C., F.B., and S.D.-C. wrote the article.

www.plantphysiol.org/cgi/doi/10.1104/pp.15.01126