|Home | About | Journals | Submit | Contact Us | Français|
Hessian fly (HF) is a biotrophic insect that interacts with wheat on a gene-for-gene basis. We profiled changes in membrane lipids in two isogenic wheat lines: a susceptible line and its backcrossed offspring containing the resistance gene H13. Our results revealed a 32 to 45% reduction in total concentrations of 129 lipid species in resistant plants during incompatible interactions within 24 h after HF attack. A smaller and delayed response was observed in susceptible plants during compatible interactions. Microarray and real-time PCR analyses of 168 lipid-metabolism related transcripts revealed that the abundance of many of these transcripts increased rapidly in resistant plants after HF attack, but did not change in susceptible plants. In association with the rapid mobilization of membrane lipids, the concentrations of some fatty acids and 12-oxo-phytodienoic acid (OPDA) increased specifically in resistant plants. Exogenous application of OPDA increased mortality of HF larvae significantly. Collectively, our data, along with previously published results, indicate that the lipids were mobilized through lipolysis, producing free fatty acids, which were likely further converted into oxylipins and other defense molecules. Our results suggest that rapid mobilization of membrane lipids constitutes an important step for wheat to defend against HF attack.
The wheat-Hessian fly (HF, Mayetiola destructor) system has become a model for studies on plant-gall midge interactions (Stuart et al., 2008). Because of its importance in agriculture, the virulence and genetics of HF have been studied extensively (Harris et al., 2003; Stuart et al., 2008). HF attack induces reprogramming of gene transcription and dramatic changes in metabolic pathways in wheat (Liu et al., 2007; Zhu et al., 2008). In susceptible wheat during compatible interactions, nutritive cells are formed at the feeding site (Harris et al., 2006), and metabolic pathways are engineered towards producing more amino acids for HF nutrition (Zhu et al., 2008). Growth of susceptible plants is inhibited after HF infestation and the infested plants die eventually after HF larvae no longer need nutrients from the host (Byers and Gallun, 1971). On the other hand, HF larvae die in plants containing an effective resistance (R) gene, and the attacked plants continue to grow normally after some initial growth deficit (Anderson and Harris, 2006). Substantial indirect evidence suggests that HF attacks wheat using salivary secretions (Byers and Gallun, 1971; Hatchett et al., 1990; Chen et al., 2008; Chen et al., 2010), which are likely to target regulatory pathways in wheat to induce changes that benefit the insect, as observed with many secreted elicitors in pathogen-plant interactions (Hueck, 1998; Bent and Mackey, 2007).
Defense in wheat is activated on a typical gene-for-gene basis in response to HF attack (Hatchett and Gallun, 1970), with resistance conditioned by major dominant R genes (Hatchett and Gallun, 1970; Liu et al., 2005). Avirulence in HF is controlled by dominant avirulence (Avr) genes (Behura et al., 2004). Since HF-wheat interaction is similar to gene-for-gene interactions observed in many pathogen-plant systems, it is likely that an Avr gene encodes a virulence factor, which is recognized either directly or indirectly by the corresponding R protein once it is secreted into host tissue during larval feeding (Keen, 1990; Bent and Mackey, 2007). The specific recognition between a pair of R and Avr proteins results in defense signaling and eventually, the launch of direct chemical defenses, resulting in antibiosis to HF larvae (Chen et al., 2009b; Harris et al., 2010).
Antibiosis in resistant plants during incompatible interactions is likely due to a combination of mechanisms, including toxic chemicals such as lectins (Subramanyam et al. 2008), reactive oxygen species (Liu et al., 2010), digestive inhibitors (Wu et al., 2008), and secondary metabolites (Liu et al., 2007). Strengthening of the cell wall by increased deposits of phenolics (Liu et al., 2007) and enhanced cuticle wax (Kosma et al., 2010) are also likely involved in antibiosis. In previous studies, we found that several genes involved in lipid metabolism were differentially regulated in plants during incompatible and compatible interactions (Liu et al., 2007). Similar results were also observed in the wheat-Russian wheat aphid interaction (Smith et al., 2010). These observations suggest that lipids may play an important role in wheat defense against HF and other insect pests.
Polar lipids are major components of membranes. Eukaryotic membranes not only form mechanical barriers to separate the interior of a cell or a subcellular structure from the outside environment, but also serve as critical interfaces for interactions between cellular structures, cells, and organisms. Membranes consist of diverse lipids and lipid composition changes in response to either internal signals or outside stimuli (McMahon and Gallop, 2005). During interactions between plants and herbivores, lipid changes in cellular membranes of the attacked host could be critical for defense signaling (Laxalt and Munnik, 2002) and the launch of direct defenses (Kosma et al., 2010). Until recently, however, analysis of compositional changes in membrane lipids has been labor-intensive. Because of the technical difficulty, studies on the roles of lipids in plant defense against herbivores have been focused on analysis of individual lipid species or a subset of lipids (Seigler, 1983; Saini et al., 1994). The development of electrospray ionization mass spectrometry (ESI-MS) has made possible the detection and quantification of membrane lipids (Welti et al., 2002; Forrester et al., 2004). The sensitivity, accuracy, and reproducibility of ESI-MS make it a powerful tool for systematically profiling changes in lipids in organisms under various conditions.
The objectives of this study were: 1) To profile lipid compositions in wheat seedlings of two different genotypes, a susceptible line and a backcrossed offspring containing the R gene H13; 2) To investigate temporal changes in membrane lipids in resistant and susceptible wheat plants after HF attack; 3) To analyze changes in abundance of transcripts encoding proteins involved in lipid metabolic pathways; 4) To examine if downstream metabolites of membrane lipids play a role in wheat defense against HF. We provide compelling evidence that a significant portion of membrane lipids were mobilized for plant defense in resistant plants during incompatible interactions.
Newton (the susceptible parent) and Molly (the backcrossed offspring carrying the R gene H13, Patterson et al., 1994) were grown under normal growth conditions and profiled for lipid content at 10 to 11 days after germination (Table 1), when HF infestation took place. A total of 129 lipid species were detected in the second leaf-sheath (10 mm tissue above the base of the first leaf node) of wheat seedlings. The detected lipids included 14 apparent molecular species (as defined by total acyl carbons: total carbon-carbon double bonds) of digalactosyldiacylglyerol (DGDG), 15 species of monogalactosyldiacylglycerol (MGDG), 7 species of phosphatidylglycerol (PG), 20 species of phosphatidylcholine (PC), 20 species of phosphatidylethanolamine (PE), 10 species of phosphatidylinositol (PI), 20 species of phosphatidylserine (PS), 9 species of phosphatidic acid (PA), 5 species of lysophosphatidylglycerol (lysoPG), 5 species of lysophosphatidylcholine (lysoPC), and 4 species of lysophosphatidylethanolamine (lysoPE). PC and PE were the two major groups, representing 41–52% and 22–25% of total detected lipids, respectively. DGDG, MGDG, PG, PI, PS, and PA were also relatively abundant, each group representing 1–8% of total detected lipids. The three groups of lysophospholipids, lysoPG, lysoPC, lysoPE, were also detected, but their concentrations were very low, each group representing less than 0.3% of total detected lipids. No statistically significant difference was observed in lipid concentrations between Newton and Molly and between day-10 and day-11 samples.
In terms of acyl chains, diacyl lipids with 32, 34, 36, 38, 40, and 42 total acyl carbons were detected (Table S1). Lipids with 34 and 36 carbons were predominant, each group representing 42 to 53%, respectively. Lipids with 38 carbons were about 2%, and lipids with all other acyl chain combinations together accounted for less than 1%. Again, there were no statistically significant differences in the concentrations of lipids with specific acyl groups between the two plant genotypes, and between the two time points. A complete list of the 129 lipids and their concentrations in the second leaf-sheath is given in Table S2 (http://www.scientificsocieties.org/MPMIXtras/2012/MPMI-01-12-0022_TableS2.xls).
A preliminary analysis indicated that genes coding for lipases and lipid transfer proteins were upregulated in resistant plants within an hour following HF infestation (data not shown). Accordingly, a time course consisting of 0.5, 3, 6, 12, 24 h was designed for the assay. The total concentration of detected lipids was reduced in resistant plants during incompatible interactions between 3 and 24 h after HF attack (Figure 1). Specifically, an average of 32.7, 25.9, 44.6, and 38.9% reduction was observed in infested tissue at 3, 6, 12, and 24 h, respectively. No significant difference was observed between uninfested and infested tissues of resistant plants during incompatible interactions at 72 h after HF attack, indicating that the reduction in lipid concentration at earlier time points was restored. In comparison, total concentration of detected lipids did not change significantly in susceptible plants during compatible interactions at all time points after HF attack.
The impact of HF attack on different subgroups of lipids in resistant plants during incompatible interactions was relatively consistent, a significant reduction in concentrations of all subgroups at all time points with three exceptions (Table 2). The three exceptions were lysoPG (not affected within the tested period), PA (not affected except for the 24 h time point), and lysoPC (increased slightly at 0.5 h time point). In comparison, the impact of HF attack on different lipid subgroups in susceptible plants during compatible interactions varied from subgroup to subgroup. Many lipid subgroups either did not change or increased in concentration in susceptible plants after HF attack. For those lipid subgroups that did show a decrease in concentration, the percentage of decrease was often much smaller than that in resistant plants.
The distribution and severity of individual lipids affected by HF attack were quite different between resistant and susceptible plants (Table S3 and Figure S1). More lipids had reduced concentration in resistant plants in comparison with those in susceptible plants. In addition, the effect of HF attack on lipids was evident earlier and with greater magnitude in resistant plants than that in susceptible plants. Specifically, there were seven, 31, 25, 52, and 46 individual lipids with concentration reduced in resistant plants at 0.5, 3, 6, 12, and 24 h after HF attack, respectively. Only four lipids across all time points (one at 0.5 h, two at 6 h, and one at 24 h) exhibited an increase in concentration (Figure S1) in resistant plants after HF attack. On the other hand, there were zero, zero, 11, 11, and 31 lipid species with concentration reduced in susceptible plants at 0.5, 3, 6, 12, and 24 h after HF attack, respectively. Fourteen lipids (two at 0.5 h, five at 12 h, and seven at 24 h) exhibited an increase in concentration in susceptible plants (Figure S1). Most galactolipid species (DGDG and MGDG) were affected by HF attack in plants during both incompatible and compatible interactions (Table S3, Figure S1). The number of phospholipids affected by HF attack varied, and only a few lipids in PG, PA, and PE subgroups were negatively affected.
To examine the impact of HF attack on the level of transcripts of lipid metabolism-related genes, primer pairs (Table S4) for nine wheat transcripts were designed according to representative cDNA sequences, and quantitative real-time PCR (qPCR) was carried out to determine the abundance of these transcripts. Three lipase-encoding transcripts were analyzed, including a GDSL-motif lipase (Figure 2A), an extracellular lipase (Figure 2B), and a triacylglyerol lipase gene (Figure 2C). The abundance of all three transcripts increased greatly in resistant plants during incompatible interactions at some point after HF attack. The increase of the GDSL-motif lipase transcript was most dramatic, with over 1,000 fold at 0.5 h. On the other hand, the abundance of these three transcripts either decreased, or remained the same, or slightly increased in susceptible plants during compatible interactions.
Three phospholipase gene transcripts were analyzed, including a phospholipase A2 (Figure 2D), a phospholipase C (Figure 2E), and a phospholipase D (Figure 2F). HF attack affected the abundance of these transcripts in both resistant and susceptible plants. Specifically, the abundance of phospholipase A2 transcript increased about 6-fold in resistant plants at 0.5 h after HF attack, and increased around 2-fold thereafter. The abundance of this phospholipase A2 transcript also increased around 2-fold in susceptible plants throughout the experimental time period. The impact of HF attack on the phospholipase C transcript was similar to that on the phospholipase A2 transcript, but to a more moderate degree. The abundance of the phospholipase D transcript increased slightly at the 0.5- and 6-h time points in resistant plants, and increased around 1-fold (i.e. 2-fold change) at the 6-h time point in susceptible plants.
Three transcripts each encoding a different lipid transfer protein were also analyzed (Figure 2G to 2I). The abundance of transcript AF334185 (Genbank accession number) increased 1.5- to 8-fold in resistant plants, but decreased 0.5- to 2-fold in susceptible plants (Figure 2G). The abundance of transcripts CD896324 (Figure 2H) and CN012788 (Figure 2I) did not change in resistant plants, but decreased 2- to 8-fold in susceptible plants.
To gain insight on the molecular basis for lipid reduction in resistant plants during incompatible interactions and possible products to which the mobilized lipids were converted, changes in the transcript abundance of genes potentially involved in lipid metabolism were analyzed through microarrays (Table 4 and Table S5, http://www.scientificsocieties.org/MPMIXtras/2012/MPMI-01-12-0022_TableS5.xls). A total of 168 probe sets were included in the analysis, including 43 from transcripts putatively involved in lipolysis (various lipases), 55 from transcripts likely involved in lipid transport (lipid transfer proteins), 31 from transcripts related to oxylipin synthesis, 27 from gene transcripts related to wax synthesis, five from gene transcripts encoding putative enzymes for fatty acid synthesis, and seven from gene transcripts encoding putative enzymes for fatty acid catabolism. Alteration in the abundance of individual transcripts after HF attack is shown in Table S5. Changes in different groups of transcripts at different time points are summarized in Table 3.
Among 43 lipase transcripts, 22 increased in abundance in resistant plants at some points during the 3-day period after HF attack, whereas only five transcripts decreased. In contrast, only one transcript increased and as many as nine transcripts decreased in abundance in susceptible plants at some points during the same period. Transcript abundance was affected the most during the first 24 h. Among the different types of lipases, transcripts for lipase class 3 (triacylglycerol lipase) and GDSL-motif lipases were affected the most. Only a small number of phospholipase transcripts was affected (Table S5).
Among the 55 transcripts encoding putative lipid transfer proteins, 27 of them increased and 9 decreased in abundance in resistant plants after HF attack. In contrast, none of the transcripts increased and 29 decreased in susceptible plants after HF attack (Table 3). Most of the transcripts that increased in abundance in resistant plants showed a decrease in susceptible plants (Table S5). The opposite changes resulted in dramatic differences in the abundance of these transcripts between resistant and susceptible plants.
Among the 31 transcripts potentially involved in oxylipin synthesis, 23 increased and 8 decreased in abundance in resistant plants after HF attack (Table 3). One of them increased more than 600-fold (Table S5). In contrast, no oxylipin transcripts increased and 6 decreased in abundance in susceptible plants. The transcripts in this category were derived from three types of genes: ω-3 fatty acid desaturases (ωFAD), lipoxygenase (LOX), and allene oxide synthase (AOS). All these enzymes participate in oxylipin synthesis. In each type, there were transcripts with increased abundance in resistant plants (Table S5).
Among the 27 transcripts putatively involved in wax synthesis, nine increased and four decreased in abundance in resistant plants after HF attack (Table 3). In contrast, four increased and nine decreased in susceptible plants after HF attack. There are many different types of genes involved in synthesis of different types of cuticle wax (Kunst and Samuel, 2009) and only a small number of gene transcripts were included in the microarray. Among the different types, transcripts encoding fatty acyl CoA reductases were affected the most (Table S5). Fatty acyl-CoA reductases catalyze the formation of fatty alcohols, components of plant cuticle wax.
Twelve transcripts encoding fatty acid metabolic enzymes were included in the microarray, with five involved in fatty acid synthesis and seven involved in fatty acid degradation (Table 3). All five transcripts for fatty acid synthesis encoded acetyl-CoA carboxylase, an enzyme required for the irreversible carboxylation of acetyl-CoA to produce malonyl-CoA during fatty acid synthesis. The seven transcripts for fatty acid degradation encoded two types of enzymes, enoyl-CoA hydratase (an enzyme that hydrates the double bond on acyl-CoA) and fatty acid hydroxylase (an enzyme involved in fatty acid degradation). No significant changes in the abundance of all 12 transcripts were observed in resistant plants after HF attack. In contrast, the abundance of the two transcripts encoding fatty acid hydroxylases increased in susceptible plants within 24 h after HF attack (Table S5).
The change in gene expression in lipid-related pathways pointed to increased accumulation of defense-related molecules from the mobilized lipids. Since the majority (over 50%) of membrane lipids contained 18 carbon fatty acids (Table S1), we measured the concentrations of 18 carbon fatty acids in plants during both compatible and incompatible interactions (Figure 3A). The concentrations of the fatty acid 18:3 increased over 4 and 2 fold in resistant plants during incompatible interactions 24 and 72 h, respectively, after HF infestation. In comparison, the concentration of the fatty acid 18:3 decreased ~50% at 24 h and no significant change at 72 h in susceptible plants during compatible interactions. The concentrations of the fatty acid 18:2 did not change significantly in either resistant or susceptible plants. The concentrations of the fatty acid 18:1 increased about 2 fold at both 24 and 72 h after HF infestation in resistant plants, but no significant change in susceptible plants. The concentration of the fatty acid 18:0 was un-measurable under our condition (data not shown).
The concentrations of downstream metabolites OPDA and jasmonic acid (JA) were also measured (Figure 3B). The concentrations of OPDA increased more than 30 and 8 fold at 24 and 72 h, respectively, after HF infestation in resistant plants during incompatible interactions. No significant changes in OPDA concentration were observed in susceptible plants during compatible interactions. Interestingly, no significant changes in JA concentrations were observed in plants during both compatible and incompatible interactions. For comparison, we also measured the concentrations of salicylic acid (SA), another defense-related plant hormone. The amounts of SA increased more than 8 and 5 fold at 24 and 72 h, respectively, after HF infestation in resistant plants. No significant changes in SA concentration were observed in susceptible plants.
Since both OPDA and SA increased dramatically in resistant plants, we tested if exogenous application of OPDA to susceptible plant genotypes has an impact on HF larval survival. The survival rate of HF larvae decreased ~50% in plants treated with exogenous OPDA (Figure 3Ca). Similarly, application of SA also resulted in lower survival in comparison with untreated control plants. Interestingly, there was no further increase in HF larval mortality when both OPDA and JA were applied (data not shown). To examine if the decrease in larval survival was due to any effect of the exogenous chemical application on HF larval hatch or migration, the number of HF larvae in treated plants was counted 24 h after larval hatch and migration. There was no difference in the numbers of larvae that migrated successfully to the feeding site between control and SA-treated plants (Figure 3Cb). OPDA was omitted from this observation due to the high cost of the chemical.
The mobilized membrane lipids could also be converted into cuticular wax and cutin, which can result in strengthened cell wall in resistant plants during incompatible interactions (Kosma et al., 2010). Accordingly, epidermal permeability of wheat sheath at the HF-feeding site was determined by neutral red staining (Figure 3D). Wheat sheaths of uninfested plants were not permeable to the red staining dye (UC). Wheat sheaths of susceptible plants during compatible interactions were strong permeable to the dye, resulting in full red staining (CI). In comparison, wheat sheaths of resistant plants during incompatible interactions were also not permeable to the dye (II), indicating that cell wall was strengthened or damage inflicted by Hf was repaired in resistant plants following HF attack.
We profiled changes in membrane lipids in HF-feeding tissue of wheat seedlings during incompatible and compatible interactions, and examined changes in transcript abundance of genes involved in lipid metabolism. We observed a rapid reduction of membrane lipids specifically in resistant plants during incompatible interactions. Along with the reduction in membrane lipids, transcripts encoding various lipases rapidly accumulated in resistant plants (Table 3, Table S5, and Figure 2), suggesting that lipid mobilization was achieved via lipolysis. Changes in membrane lipids in plants have been observed in response to other stresses including low temperatures (Vigh et al., 1985; Cyril et al., 2002; Welti et al., 2002; Wang et al., 2006), salts (Surjus and Durand, 1996); heavy metals (Nesterov et al., 2008), herbicides (Keppler and Novacky, 1989), drought (Ferrari-Iliou et al., 1984; De Paula et al., 1990; Gigon et al., 2004), and pathogens (Keppler and Novacky, 1989; Saini et al., 1994); and lipolytic enzyme transcript levels and activities are elevated in stressed plants (El-Jafid et al., 1989; Welti et al., 2002; Gigon et al., 2004; Shah, 2005).
Previously we have observed that there was a transient, but severe down-regulation of the central metabolic pathways including glycolysis, the citric acid cycle, and the pentose phosphate pathway specifically in resistant plants during incompatible interactions (Zhu et al., 2008). The down-regulation of primary metabolism is likely for the purpose to conserve energy and nutrients for plant defense. The rapid mobilization of membrane lipids could provide necessary resources for the synthesis of defense chemicals. Since the membrane lipids in leaf-sheaths of wheat seedlings contained predominantly 18 carbon acyl chains (Table S1), the accumulation of the 18 carbon fatty acids was analyzed in resistant plants during incompatible interactions after HF attack. The analysis revealed an unbalanced accumulation of these fatty acids, with a dramatic increase in 18:3 and 18:1, no change in 18:2, and un-detected 18:0 (Figure 3A). The impact of the unbalanced accumulation of the fatty acids remains to be determined. Some of the PUFAs such as 18:1 and 18:3 are themselves defense-signaling molecules (Shah, 2005), in addition to being precursors for oxylipin synthesis. In association with the increase of 18:1 and 18:3 fatty acids, the amounts of OPDA increased greatly in resistant plants (Figure 3B). Interestingly, the amount of JA did not change in the same plants. The significance of elevated OPDA but not JA again remains to be delineated. OPDA itself does play a role in plant defense (Stintzi et al. 2001; Chandra-Shekara et al., 2007). In addition to elevated OPDA, the defense signaling molecule SA also increased greatly in resistant plants during incompatible interactions. The interactions between JA and SA pathways in plant defense is complex, including both antagonism and synergism (Mur et al. 2006). On one hand, high levels of SA inhibit endogenous production of JA (Mur et al. 2006). On the other hand, genes in both JA and SA pathways are often induced upon pest infestation (Zhu-Salzman et al., 2004). In wheat resistance against HF attack, genes in both JA and SA pathways are upregulated (Liu et al., 2007). Our data suggested that the intermediates including 18:1 and 18:3 fatty acids, and OPDA in the JA synthetic pathway may provide synergism with the SA pathway in wheat defense against HF attack. Consistent with this possibility, exogenous application of both OPDA and SA increased HF larval mortality (Figure 3C).
The coordinated upregulation of genes in the oxylipin synthesis pathway including genes coding for key enzymes such as ωFAD and AOS was likely the molecular basis for conversion of mobilized membrane lipids into signaling defense molecules such as 18:1 and 18:3 fatty acids and OPDA. By the same logic, part of the mobilized lipids might have also been converted into cuticular wax and cutin (Kosma et al., 2010). The abundance of several transcripts encoding enzymes potentially involved in wax and cutin synthesis increased in resistant plants (Table 3, Table S5). Increased abundance of many wax-related transcripts was also reported by another group (Kosma et al., 2010). Other evidence supporting wax synthesis came from the strong upregulation of many transcripts encoding different lipid-transfer proteins (Table S5). Analysis of amino acid sequences encoded by corresponding rice genes that shared the highest similarity with the increased wheat transcripts revealed that all except one have a secretion signal peptide even though the protein sequences themselves are very diversified (data not shown). Lipid-transfer proteins with a secretion signal are likely secreted into the cell wall and participate in wax and cutin formation (Kader, 1996). Consistent with the possibility of increased wax and cutin, the ability of HF larvae to compromise wheat sheath impermeability is neutralized in resistant plants during incompatible interactions (Figure 3D).
While the evidence is solid for the conversion of mobilized membrane lipids into defense molecules in resistant plants, other questions remain to be answered. Among them, does increased mobilization of membrane lipids affect membrane integrity? The decrease in lipid concentration appeared to be relatively similar among different classes except PA and lysophospholipids (Table 2), which are the degradation products of other phospholipids. Increases in the levels of these classes can increase membrane instability (Lee and Chan, 1977). On the other hand, considering that HF attack causes a growth deficit in resistant plants during the first few days (Anderson and Harris, 2006), the mobilized membrane lipids might be compensated by a reduced amount of total membrane in temporally non-dividing cells. Among the upregulated lipase transcripts, the majority of them encoded lipase 3 and GDSL-motif lipases. Phospholipase transcripts were less upregulated and in some cases, upregulation could only be detected by the more sensitive method qPCR (Figure 2). The strong increase in non-phospholipase transcripts suggested that other non-polar lipids such as triacylglyerides were also mobilized. The mobilization of structural components and nutrients might constitute a defense cost, resulting in growth deficit during this period. Mobilization of membrane lipids for defense observed in this study is also likely to happen in plants against some biotrophic pathogens and other herbivores. Various phospholipases are upregulated in plants in response to pathogen attack (Dhondt et al., 2000; de Torres Zabela et al., 2002; Shah, 2005). Many other lipid-metabolism related genes are also upregulated in other plant-pathogen interactions and the upregulation is often associated with increased plant resistance (Ohta et al., 1991; Kachroo et al. 2001; Maldonado et al., 2002; Oh et al., 2005; Shah, 2005; Kwon et al., 2009; Sarowar et al., 2009).
In conclusion, our data, along with published evidence, suggested that membrane lipids are rapidly mobilized to produce defense molecules that enhance wheat resistance to HF. Specifically, we think membrane lipids were mobilized by lipolysis (Figure 4). The released fatty acids from mobilized membrane lipids were then converted into defense signaling molecules such as 18:3 and 18:1 fatty acids and OPDA, and potentially into cell wall strengthening components including wax and cutin. The rapid mobilization of membrane lipids and subsequent conversion into defense molecules were likely part of wheat defense mechanism against HF attack during incompatible interactions. On the other hand, the inhibition of both lipid mobilization and lipid conversion into defense molecules might be a necessary molecular process for the susceptibility of wheat seedlings during compatible interactions with HF.
Two wheat isogenic lines, Newton and Molly, were used in this study. Newton is a susceptible cultivar whereas Molly is a backcrossed offspring containing the R gene H13 (Patterson et al., 1994). A Kansas HF population that is virulent to Newton but avirulent to Molly was used for infestation (Chen et al., 2009a). The Kansas population contains the Avr gene corresponding to the R gene H13.
Twenty germinated wheat seeds were planted in each 10-cm-diameter pot filled with PRO-MIX ‘BX’ potting mix (Hummert Inc., Earth City, MO) in a growth chamber programmed at 20:18°C (L:D) with a photoperiod of 14:10 (L:D) h. When wheat seedlings reached 1.5 leaf stage (stage 11 on Zadoks scale) (Zadoks et al. 1974), the plants were infested with one HF female per plant for 6 h when the egg density reached around ~25 eggs per plant. The females were then removed from the plants to prevent them from laying more eggs. In this way, the larval hatching time is relatively uniform. The initial HF larval attack time was determined as following: Three days after eggs being laid, a few plants were cut at the base (just above the node where roots begin). The cut plants were put into a 50 ml plastic tube with 5 ml water. The tube with the plants were put in the same growth chamber and monitored for larval hatch every 10 minutes. The time when about half of the expected neonates (~10 per plant) from the cut plants entered into the water was taken as the HF larval initial attack time. Each treatment was biologically repeated four times for lipid profiling but three times for qPCR. Each sample was collected from 20 individual plants in a pot. For each treatment, control plants were grown in the same growth chamber under the identical conditions. Plant samples were collected by obtaining a 10-mm long section of the second leaf sheath after removing HF larvae. Samples for RNA extraction were frozen in liquid nitrogen (N2) immediately after sampling and stored at −80°C until RNA extraction. Samples for lipid extraction were immediately soaked in 3 ml preheated (75°C) isopropanol with 0.01% butylated hydroxytoluene (BHT) in a 50 ml tube with a Teflon-lined screw cap for 15 min to inactivate phospholipase D. Each tube with samples was then added with 1.5 ml chloroform and 0.6 ml water followed by a brief vortex. The samples were then stored in a −20°C freezer.
Lipids were extracted from tissue samples following the method described by Welti et al. (2002). Briefly, lipids were extracted from the frozen wheat tissues by shaking at room temperature for 1 h. The lipid extract was then transferred to a glass tube with Teflon-lined screw-cap, and 4 ml chloroform/methanol (2:1 v/v) with 0.01% BHT were added to the remaining tissues. The sample tissues were again shaken at room temperature for 30 min. The extract from the second extraction was combined with the first extract. This extraction procedure was repeated five times, at which time the wheat tissues were completely white. The combined extracts were washed once with 1 ml of 1 M KCl and once with 2 ml of water. After washing, tubes with lipid samples were filled with nitrogen to allow the solvent to evaporate. Lipid extract was then dissolved in 1 ml of chloroform for further analysis. The remaining plant tissues after lipid extraction were dried at 105°C overnight. The dried tissues were weighed and recorded as “dry mass” for each sample.
To profile lipid species and determine lipid concentration in individual samples, an automated ESI-tandem mass spectrometry approach was followed as described by Welti et al. (2002) with modifications as described in Xiao et al. (2010). The apparent lipid molecular species were quantified in comparison to internal standards of the same lipid class. Analysis of free fatty acids, OPDA, JA, and SA was conducted as described by Zhu et al. (2010).
The Affymetrix GeneChip® Wheat Genome Array (Affymetrix Inc., Santa Clara, CA) was used for wheat microarray studies. RNA was isolated from frozen tissue with TRIzol reagent (Invitrogen, Carlsbad, CA), according to the manufacturer’s instruction. Generation of cRNA, fluorescent labeling, hybridization, scanning and quantification of hybridized Affymetrix wheat arrays was performed according to the Affymetrix GeneChip® Expression Analysis Technical Manual (Affymetrix) in the KSU Gene Expression Facility (Liu et al., 2007).
Three biological repeats were conducted for each treatment. Microarray data were initially processed using GCRMA (Wu et al. 2004), an improved version of robust multiarray averaging. Probe sets that are potentially involved in lipid metabolism were analyzed with the empirical Bayes method, followed by Benjamini and Hochberg (1995) multiple comparison-adjustment method to control false discovery rate. Microarray data were deposited to the data base of National Center for Biotechnology Information (NCBI) with accession number GSE34445 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE34445).
DNase-treated total RNA was used for cDNA synthesis using random hexamers with a superscript reverse transcriptase kit (BD Biosciences, San Jose, CA). Samples were then treated with RNase H (Invitrogen). cDNA was quantified on a Nanodrop ND-1000 (NanoDrop Technologies Inc., Wilmington, DE) spectrophotometer and samples were diluted to 15 ng/µl to ensure equal amounts of cDNA template for quantification of mRNA abundance.
qPCR was performed with iQ SYBR Green Supermix on a iCycler real time detection system (Bio-Rad, Hercules, CA). Each reaction was carried out with 2 µl of a 1/40 (v/v) dilution of the first cDNA strand, 0.5 µM of each primer in a total volume of 25 µl. The cycling conditions were: 95°C for 5 min followed by 45 cycles of denaturation at 95°C for 20 s, annealing and extension at 62°C to 64.5°C, depending on the primer set, for 45 s. At the end of the cycles, PCR amplification specificity was verified by obtaining a dissociation curve, derived by cooling the denatured samples to 55°C and raising the temperature 0.5°C for 10 s for each cycle, for a total of 80 cycles until reaching 95°C. The PCR products were analyzed on 1.5% agarose gels, and subsequently purified and sequenced to confirm faithful amplification. Primers were designed using the Beacon Designer (v2.0) software from Biosoft International (Palo Alto, CA) and the primer sequences were listed in Table 1. Plasmid DNA containing the corresponding insert was used to generate a calibration standard curve, where Cycle Threshold (Ct) values are plotted to serve as standard concentrations. The transcript concentration for each sample was calculated based on the standard concentrations. A negative control without template was always included for each primer set. Template concentrations in different samples were normalized by the ribosomal protein L21 (RPL21).
Three mM solutions of OPDA or SA in 0.02% Silwet and 0.5% ethanol were sprayed on wheat seedlings with a small hand sprayer. First spray was carried out 24 h before HF larval hatch. The spray was continued once every 24 h for 5 consecutive days. The numbers of surviving larvae were counted three weeks after larval hatch. To determine if the first spray affected larval hatch and migration, a set of control and treated plants were dissected and neonate larvae that reached the feeding site were counted and compared 24 h after larval hatch.
Neutral red stain (Sigma-Aldrich, St. Louis, MO, USA) was used to determine epidermal permeability of cells as described by Kosma et al (2010). Briefly, plants were dissected 2 days after the initial HF larval attack. After peeling off the first leaf-sheath, the HF larval feeding site of the second leaf-sheath was stained with 0.1% neutral red stain for 10 min, followed by five times washing with water. Uninfested plants were dissected and stained as negative controls in the same way. After staining, plant tissues were examined under a fluorescent microscope (Zeiss Axioplan-2) and photographed with a Nikon Coolpix 4500 Digital camera.
For comparison of total lipids, analysis of variance was first conducted, followed by the Tukey's Honestly Significant Difference (HSD) test to identify differences significant between uninfested and infested samples for each genotype at each time point. These tests ensured that the family-wise type I error for all comparisons was controlled at the desired significance level.
For comparison of lipid subgroups, the empirical Bayes procedure was used to determine if a change in percentage was significant, followed by Bonferroni correction for multiple adjustments. Hierarchical linear models were applied to evaluate how the concentration of each lipid subgroup varied between uninfested and infested samples at each time point. P-values from a moderated F-statistic were multiplied by the number of lipid subgroups as well as the number of time points (truncated at 1).
For comparison of individual lipid species, the empirical Bayes method of Smyth (2004) with moderated t-statistics was used to compare data from uninfested and infested wheat seedlings of each plant genotype at each time point. Bonferroni correction was applied to control family-wise error rate for each lipid species.
For qPCR data, statistical significance was analyzed by ANOVA using the PROC#MIXED procedure of SAS (SAS institute Inc., SAS/STAT User’s Guide, Version 9.13). The ANOVA analysis model included treatment, time-points and their interaction. Data from three biological replicates (each replicate assayed two times in independent qPCR experiments) were combined and included as a random effect in the analysis. Fold-change calculations were performed by comparing mRNA abundance of selected genes in infested plants with that of the same genes in uninfested controls. Fold-change was considered statistically significant if the P-value was ≤0.05. The number of fold increase or decrease was defined as the number of fold change minus 1, for example, a 2 fold change was considered 1 fold increase or decrease.
Authors thank Drs. C. Michael Smith and Robert Bowden for reviewing an earlier version of the manuscript, and Mary Roth, Richard Jeannotte, and Thilani Samarakoon for technical assistance. This paper is contribution no. 11-070-J from the Kansas Agricultural Experiment Station. This research is a joint effort from the United States Department of Agriculture-Agriculture Research Service at Manhattan, and Fayetteville State University. This research was partially supported by a grant (P20 MD001089) from NIH, NCMHD, Department of Health and Human Services. The lipid analyses described in this work were performed at the Kansas Lipidomics Research Center, where instrument acquisition and method development was supported by NSF (EPS 0236913, MCB 0920663 and 0455318, DBI 0521587), Kansas Technology Enterprise Corporation, Kansas INBRE (NIH P20RR16475), and Kansas State University.
*Mention of commercial or proprietary product does not constitute endorsement by the USDA.
Table S1. Composition of acyl chain in detected lipids.
Table S2. Lipid species and concentrations in control and HF attacked plants. (http://www.scientificsocieties.org/MPMIXtras/2012/MPMI-01-12-0022_TableS2.xls)
Table S3. Lipid species with reduced concentrations in resistant plants.
Table S4. Primer pairs used for quantitative real-time PCR analysis.
Table S5. Changes in transcript levels of genes potentially involved in lipid metabolism. (http://www.scientificsocieties.org/MPMIXtras/2012/MPMI-01-12-0022_TableS5.xls)
Figure S1. Impact of HF attack on individual lipid species.