Phospholipase Dε and Phosphatidic Acid Enhance Arabidopsis Growth

Summary

The activation of phospholipase D (PLD) produces phosphatidic acid (PA), a new lipid messenger implicated in cell growth and proliferation, but direct evidence for PLD and PA promotion of growth at an organismal level is lacking. Here we characterized a new PLD, PLDε, and show that PLDε plays a role in promoting Arabidopsis growth. PLDε is mainly associated with the plasma membrane and is the most permissive of all PLDs tested in activity requirements. Knockout (KO) of PLDε decreases, whereas overexpression (OE) of PLDε enhances root growth and biomass accumulation. The level of PA was higher in OE, but lower in KO than in wild-type plants, and suppression of PLD-mediated PA formation by alcohol alleviated the growth-promoting effect of PLDε. OE and KO of PLDε had the opposite effect on lateral root elongation in response to nitrogen (N). Increased expression of PLDε also promoted root hair elongation and primary root growth at severe N deprivation. The results suggest that PLDε and PA promote organismal growth and play a role in N response. The lipid signaling process may play a role in translating the membrane sensing of nutrient status to increasing plant growth and biomass production.

Introduction

Membrane lipid hydrolysis by phospholipases (PL), such as PLD, PLC, and PLA, produces different classes of lipid mediators in growth response to nutrient and stress cues (Foster and Xu, 2003; Wang, 2004; Huang and Frohman, 2007). Activation of one or more of these reactions often constitutes an early, critical step in many signaling cascades. Phosphatidic acid (PA) has emerged as a new class of lipid messengers (Wang 2004; Testerink and Munnik, 2005; Wang et al., 2006; Carman and Henry, 2007). PA has been shown to regulate proteins important in cellular regulation, including G proteins, protein kinases, phosphatases, and transcriptional factors (Fang et al., 2001; Mishra et al., 2006; Zhao et al., 2007; Carman and Henry, 2007). PLD is a major family of enzymes that generate signaling PA and play a critical role in regulating the location and timing of PA production (Wang et al., 2006). PLD and PA have been associated with mammalian cell growth, proliferation, and survival signaling (Foster and Xu, 2003; Fang et al., 2001; 2003). In plants, PLD and PA have been implicated in promoting cell elongation in pollen and root hairs (Potocky et al., 2003; Anthony et al., 2004) as well as in plant growth in response to phosphorus deficiency and hyperosmotic stress (Cruz-Ramirez et al., 2006; Li et al., 2006; Hong et al., 2008). However, the role of PLD and PA in promoting whole organismal growth under normal growth conditions has not been documented in any system.

PLD is a heterogeneous family. The biochemical properties, domain structures, and genomic organization of PLDs are more diverse in higher plants than in other organisms (Wang et al., 2006). The Arabidopsis genome contains 12 PLDs, α (3), β (2), γ (3), δ, ε, and ζ (2), ten of which have the Ca²⁺/phospholipid-binding C2 structural fold, whereas two PLDζs contain phosphoinositide-interacting PH (pleckstrin homology) and PX (Phox homology) domains (Qin and Wang, 2002). The presence of the different regulatory motifs provides insights into the different modes of activation and functions of PLDs. Distinguishable biochemical properties and physiological functions have been reported for several PLDs. These include a role of PLDα1 in signaling abscisic acid regulation of stomatal movement, PLDα3 in hyperosmotic tolerance, PLDβ in defense response, PLDδ in H₂O₂ response and freezing tolerance, and PLDζs in root development in response to phosphate starvation and auxin (Zhang et al., 2003; 2004; Li et al., 2004; Mishra et al., 2006; Cruz-Ramirez et al., 2006; Li et al., 2006; Li and Xue, 2007; Hong et al., 2008). But the function of the other PLDs remains unknown.

PLDε encodes a protein that is distinctively different from the other 11 PLDs in Arabidopsis. It has the C2 structural fold, but contains no acidic residues involved in Ca²⁺ binding in the C2 domain (Qin and Wang, 2002). Phylogenetic analysis of the Arabidopsis and rice PLD families suggests that PLDε is most closely related to the PX/PH-PLDζs. Here, we characterize the properties and function of PLDε and show that PLDε is more permissive than other characterized PLDs in activity requirements, and that genetic manipulation of PLDε alters Arabidopsis root growth and biomass accumulation. Further analysis of the PLD-altered plants indicates that PLDε and its derived PA are involved in N signaling, and that the lipid signaling may play a role in connecting the membrane sensing of nutrient status to translational regulation of growth.

Results

PLDε is associated with membranes and active under many reaction conditions

PLDε was expressed in all Arabidopsis tissues examined, and the PLDε level was highest in roots and low in leaves (Figure 1a). The relative level of PLDε expression in tissues was much lower than that of PLDα1 and also different from that of PLDα1 (Figure 1a). Expression data from Genevestigator also indicated that the level of PLDε expression was much lower than that of PLDα1 except in pollen, where the expression of PLDα1 was low and that of PLDε was much higher than in other tissues. To determine whether PLDε encodes a functional enzyme, PLDε was fused with a hemagglutinin (HA)-tag at its C-terminus and expressed in Arabidopsis plants under the control of the cauliflower mosaic virus 35S promoter. PLDε was then isolated from plants by immuno-affinity chromatography. The isolated PLDε-HA was detected by the HA antibody, but not by an antibody raised against PLDα1 (Figure 1b). As a control, proteins from Arabidopsis leaves transformed with an empty HA-vector were isolated using the same immuno-affinity procedure. No protein band in the vector control was detected by either HA or PLDα antibodies (Figure 1b). These results indicate that PLDε is purified without apparent contamination from the common PLDαs.

Figure 1. Expression and biochemical properties of PLDε.

(a) Expression of PLDε and PLDα1 in Arabidopsis tissues as determined by real time PCR. The expression levels were normalized in comparison to that of UBQ10.

(b) Immunoblotting of PLDε isolated by immunoaffinity chromatography from plants transformed with 35S::PLDε-HA or an empty vector. Proteins were separated on 8% SDS-PAGE, blotted with anti-HA or anti-PLDα antibodies, and made visible by staining with alkaline phosphatase.

(c) PLDε reaction conditions. The purified PLDε-HA was used for PLD activity assay under PLDα1, β, δ, and ζ reaction assay conditions. Values are means ± SD (n = 3).

(d) PLDε activity toward different phospholipids, NBD-PC, -PE, -PG, or -PS, under PLDα1 reaction conditions. Values are means ± SD (n = 3).

Vector refers to a control in which proteins from Arabidopsis leaves transformed with an empty HA-vector were isolated using the same immunoaffinity procedure used to isolate PLDε-HA.

The purified PLDε was assayed under the reaction conditions that were defined previously for PLDα1, β, δ, and ζ, which display distinguishable requirements for Ca²⁺, polyphosphoinositides (PPI), or fatty acids, and substrate selectivity (Wang, 2004; Wang et al., 2006). PA production was proportional to the amount of enzyme and the reaction time. PLDε was active under the PLDα1 reaction condition that included 50 mM Ca²⁺, SDS, and single-lipid class vesicles, while none of the other previous characterized PLDβ, γ, δ, or ζ displayed activity under the PLDα1 condition (Qin and Wang, 2002) (Figure 1c). With μM Ca²⁺, PLDε required oleic acid for activity, a requirement similar to that of PLDδ (Figure 1c). PLDε also displayed some activity under PLDβ and γ conditions, which included μM Ca²⁺ and the lipids, PIP₂ and PE (Figure 1c). PLDε hydrolyzed the common membrane phospholipids, PC, PE, and PG, and had low activity on PS when the enzyme was assayed with single-class lipid vesicles (Figure 1d). The results indicate that PLDε is active under a broad range of reaction conditions.

PLDε was present in the microsomal, but not in soluble fractions (Figure 2a). When the membrane fraction was separated into plasma and intracellular membrane fractions, most PLDε was associated with the plasma membrane (Figure 2a). To verify the subcellular localization, PLDε was fused with yellow fluorescence protein (YFP) at the C terminus and transiently expressed in tobacco leaves. YFP alone was detected in the nucleus and cytoplasm as expected (Figure 2b, left) whereas PLDε-YFP fluorescence was detected along the plasma membrane (Figure 2b, right).

Figure 2. Subcellular localization of PLDε.

(a) Subcellular fractionation of PLDε. S, Soluble; M, microsomal fractions, PM, plasma membrane; IM, intracellular membrane. 30 μg/lane was loaded for soluble proteins, and 5 μg/lane with membrane fractions.

(b) Subcellular localization of PLDε-YFP using 35S_P:YFP:rbcS_T as a control (left) and 35S_P:AtPLDe:YFP:rbcS_T (right). The constructs were transiently expressed in tobacco leaves by infiltration. Bar indicates 50 μm.

Overexpression of PLDε increases plant growth

The PLDε-overexpressing (OE) plants, together with a homozygous knockout (KO) mutant pldε-1 (Figure 3a,b), were used to investigate the physiological functions of PLDε. More than 20 independent PLDε-overexpressing lines were obtained, and the production of PLDε-HA was confirmed by immunoblotting using HA-antibody (Figure 3c). KO mutant was isolated from the SALK_023603 line in Arabidopsis Columbia ecotype. The mutant pldε-1 contained a T-DNA insertion in the second exon, 631 bp downstream of the start codon (Figure 3a). The mutation resulted in loss of PLDε expression as indicated by the absence of detectable PLDε transcript (Figure 3b). The mutant allele co-segregated with kanamycin resistance and susceptibility in a 3:1 ratio. When seeds were germinated in hyperosmotic growth media containing NaCl, sorbitol, or 5–8% PEG, pldε-1 seedlings grew slower and had shorter roots whereas OE seedlings grew faster and had longer roots than WT seedlings (Figure 3d–g). The primary root lengths in KO were only 16%, 70%, and 60% of WT in 100 mM sorbitol, 5% PEG, and 50 mM NaCl, respectively (Figure 3g). When seedlings were germinated in non-stress media and then transferred to hyperosmotic media, KO pldε-1 seedlings accumulated only 50% or 65% of dry matter of WT in the presence of 50 mM NaCl or 100 mM sorbitol, respectively (Figure 3h). Genetic complementation of the KO mutant with the PLDε gene restored the mutant phenotype to that of WT, confirming that the growth alterations result from the loss of PLDε (Figure 3f–h).

Figure 3. PLDε and PA enhance growth under hyperosmotic stress.

(a) T-DNA insertion site in PLDε gene. Boxes denote exons and lines introns.

(b) PLDε transcript in pldε-1 and WT plants by RT-PCR. UBQ10 was used as a control. WT, wild type; pldε-1, PLDε-knockout.

(c) Immunoblotting of PLDε in 35S::PLDε-HA plants. Proteins extracted from leaves of empty-vector transformed and 35S::PLDε-HA-transformed plants were separated by SDS-PAGE, followed by immunoblotting with HA antibodies. OE, PLDε-overexpression.

(d–f) Growth phenotypes of PLDε-altered seedlings under no NaCl (d), 50 mM NaCl (e), and 100 mM sorbitol (f). COM, PLDε-knockout complementation.

(g) Primary root growth of PLDε-altered and WT seedlings in response to osmotic conditions. Seeds of PLDε-altered and WT were germinated in MS (control) or MS containing 100 mM sorbitol, 5% PEG, or 50 mM NaCl. The primary root length was measured at day 10 after seeds were sown. Values are means ± SD (n = 10) from one representative of three independent experiments. Each genotype contained 30 seedlings.

(h) The effect of PLDε on biomass. Four day-old seedlings were transferred to growth media containing 0 (control), 100 mM sorbitol, or 50 mM NaCl without or with 0.15% 1-butanol or 2-butanol for three weeks, and were harvested for dry weight determination. Values are means ± SD (n = 10) from one representative from three independent experiments. Each genotype was represented by 30 seedlings.

In the process of analyzing PLDε-altered plants, we noted that PLDε-OE plants grew larger than WT plants on agar plates without stress treatment. The growth difference was also observed with PLDε-altered plants in soil (Figure 4a–c). The fresh and dry weights of rosettes of PLDε-OE were 192% and 212% of WT, respectively, after five weeks of growth under well-fertilized conditions (Figure 4c). The increase in biomass resulted from increases in leaf size (Figure 4d) and leaf number (Figure 4e). PLDε-OE also had approximately 25% higher seed yield than WT (Figure 4f). Enhanced growth in PLDε-OE plants was also observed under less well-fertilized conditions (Figure 4c). The cell size of OE leaves was larger, whereas that of KO was smaller than WT (Figure 4g). The magnitude of increase in cell size was smaller than that in leaf area (Figure 4d), implying that cell number also increased in PLDε-OE plants. When OE, KO, and WT plants were planted in the same tray, KO plants (Figure 4b, marked by arrows) were out-competed by WT and OE and were much smaller in size than OE and WT plants.

Figure 4. Overexpressing PLDε enhances Arabidopsis growth in soil. (a) Rosettes of WT and PLDε-overexpressing (OE) plants grown in soil.

(b) Growth competition of PLDε-altered and WT plants. White arrows mark KO plants.

(c) Rosette dry weight of five week-old plants as affected by fertilizer levels. Values are means ± SD (n = 30).

(d,e) Leaf size and leaf number of five week-old plants grown in well-fertilized soil. Values are means ± SD (n = 40).

(f) Seed yield. Values are means ± SD (n = 10) from one of three independent experiments.

(g) Cell sizes of expanded leaves from five week-old plants. Values are means ± SD (n = 60). * Indicates significant difference at P < 0.05, as compared to WT based on Student’s t-test.

PLDε enhances root growth and biomass accumulation in response to nitrogen

Because of the growth difference, we investigated the effect of PLDε-alterations on plant response to macronutrients, nitrogen (N), phosphorus (P), potassium (K), and sulfur (S) with defined nutrient composition and concentrations. Seedlings were grown in agarose-plates with the individual macronutrients reduced 10 and 100 times from the MS basal media. At 0.6 mM and 6 mM N (NO₃⁻:NH₄⁺ = 2:1), the elongation of lateral roots was significantly higher in PLDε-OE, but lower in PLDε-KO than WT seedlings (Figure 5a,b). When N was maintained at 6 mM and P, S, and K were reduced 100 times to 13 μM, 15 μM, and 213 μM, respectively, WT and KO plants displayed similar rates of lateral root elongation. Under the P and K-limiting conditions, PLDε-OE roots elongated faster than WT, but the difference was not as great as that under N-deprived conditions (Figure 5a). When seedlings were grown on media containing nitrate only, the elongation rate was also greater in OE and smaller in KO than in WT seedlings (Figure 5c). The difference was greater in seedlings grown in low (0.6 and 6 mM) than in high (60 mM) levels of nitrate (Figure 5c). High levels of nitrate are known to have an inhibitory effect on lateral root development (Zhang et al., 1999; Walch-Liu et al., 2006). Genetic complementation of the KO mutant with the PLDε gene and its own promoter restored the mutant phenotype to that of WT, confirming that the growth alterations result from the loss of PLDε (Figure 5c).

Figure 5. Lateral root elongation of PLDε-altered seedlings responds differently to N, P, K, or S deprivation.

(a) Root elongation under different nutrient deprivation conditions: N (0.6 mM), P (0.013 mM), S (0.015 mM), or K (0.213 mM). Seedlings were germinated and grown in agarose-MS media for one week and then transferred to agarose-MS media deprived of individual nutrients. Elongation of total lateral roots of individual plants was measured between five and seven days after transfer. Values are means ± SD (n = 30).

(b) Seedling morphology of PLDε-OE, KO, and WT plants. Four day-old seedlings were transferred to MS salt containing 6 mM N for 12 days.

(c) Lateral root elongation. Root length was measured eight and nine days after transferring four day-old seedlings to MS salt with 0.6, 6, and 60 mM nitrate. Values are means ± SD (n = 60).

PLDε-OE accumulated substantially more biomass, whereas KO had less biomass than WT, when grown at various levels of N (0.6, 2, 6, 60 mM) (Figure 6a). The dry weight of KO plants on average was 20% less than that of WT, whereas that of OE plants was approximately 20, 30, and 40% higher than in WT plants at N levels of 0.6, 2, to 6 mM, respectively. PLDε-OE plants grew more and longer lateral roots, whereas KO had fewer and shorter roots than WT (Figure 6b). The difference was greater in seedlings grown in low (0.6 to 6 mM) rather than high (60 mM) levels of nitrogen. At 6 mM N, the number and length of lateral roots of OE plants were two-fold higher than those of WT and KO plants (Figure 6b). The primary root length was no different in OE, WT, and KO plants at 6 and 60 mM N, but under severely N-limited conditions (0.6 mM), roots of OE plants were about 20% longer than those of WT and KO plants (Figure 6b). When seeds were germinated directly in N-limited conditions (0.1, 0.6 mM), primary roots of PLDε-KO were shorter than WT, whereas those of PLDε-OE seedlings were longer than WT (Figure 7a,b). OE plants had more, KO plants had less, lateral roots than WT (Figure 7c). These results suggest that the basal PLDε gene function is needed for maintaining normal primary and lateral root growth, particularly under N-limiting conditions. In addition, root hairs in PLDε-OE plants were more than twice the length of those of WT plants (Figure 7d,e), whereas root hair density was no different in the genotypes. The effect on root hair length was apparent only under severe N-limiting condition (0.1 mM N), while no difference in root hair length was observed at 6 mM or 60 mM N.

Figure 6. PLDε enhances biomass production at different levels of N.

(a) Biomass accumulation. Four week-old seedlings grown on MS salt agar plates were collected. Inset: Biomass at 0.6 mM N. Values are means ± SD (n=30).

(b) Root growth in response to N availability. Root length and number were measured six days after transferring four day-old seedlings to MS salt with different N levels. To observe the effect of butanol on root growth and biomass accumulation, plants were grown on plates with or without 0.15% 1-butanol or 2-butanol for four weeks. MS salt medium with decreased P level was used to test the effect of P deprivation on growth. Values are means ± SD (n = 30).

Figure 7. Primary root growth and root hair elongation under N deprivation.

(a) Seedling morphology when seeds were germinated in MS salt containing different N concentrations, 0.6, and 0.1 mM N.

(b,c) Primary root length (10 day-old) and lateral root numbers. Values are means ± SD (n = 20) in each of two independent experiments.

(d) Root hair growth of seedlings germinated in growth media containing 0.1 mM N for seven days. Values are means ± SD (n = 60).

(e) Root hair length. Values are means ± SD (n = 20).

* Denotes significant difference at P < 0.05, as compared to WT based on Student’s t-test.

PLDε-derived PA is involved in growth promotion

To determine whether PLD-produced PA is involved in growth alteration, we transferred Arabidopsis seedlings to growth media containing 1-butanol or 2-butanol. PLD uses 1-butanol, but not 2-butanol, as a substrate to form phosphatidylalcohol at the expense of PA (Wang et al., 2006). Thus, 1-butanol treatment was expected to suppress PLD-mediated PA production without inhibiting PLD degradation of membrane lipids. 1-Butanol inhibited the number and length of lateral roots in all genotypes, but the magnitude of inhibition by 1-butanol was greater on OE plants and smaller on KO plants (Figure 6b). No significant difference was observed in the number and length of lateral roots among OE, WT, and KO plants after the 1-butanol treatment (Figure 6b). The treatment of 1-butanol also eliminated the difference in biomass accumulation among WT, OE, and KO plants grown under high salinity (Figure 3h) and 6 mM N (Figure 6b). In contrast, the control treatment with 2-butanol had no inhibitory effect at the concentration tested (Figure 3h, 6b).

The level of PA in plants was measured to determine whether PA production was altered by KO and OE of PLDε. Leaf PA content from soil-grown KO plants was approximately 50% lower, and in OE plants, 15% higher than that of WT (Figure 8a). To measure PA changes in roots, seedlings were grown on plates with defined nitrogen levels. The level of PA in KO roots was only 67% of WT, whereas OE was slightly higher than WT at 2 mM N (Figure 8b). The levels of major membrane lipids, including PC, PG, MGDG, and DGDG, were similar in KO, OE, and WT roots. However, the PE level was higher in KO than WT roots, but lower in OE than WT roots. The inverse changes in PA and PE in roots suggest that most PA is derived from PLDε hydrolysis of PE (Figure 8c). The results from alcohol treatments and lipid analysis indicate that PLDε is active in PA production and that PLDε-produced PA is involved in growth promotion.

Figure 8. Effect of PLDε alterations on lipid content and composition.

(a) Leaf PA content from four week-old, soil-grown plants under well-fertilized soil conditions. Values are means ± SE (n = 5) and each replicate contained six leaves from six individual plants. The identification of the bars is the same as indicated in Figure 8b and c.

(b) Lipid content in roots from seedlings grown in MS containing 2 mM N for three weeks.

(c) Lipid molecular species of PE and PA in roots from seedlings grown in MS containing 2 mM N for three weeks. Values are means ± SE (n = 4), and each replicate contained at least 20 seedlings.

Effect of PLDε alterations on nitrogen acquisition and assimilation

To determine the effect of PLDε alterations on N metabolism, we measured N uptake and metabolism under different levels of N. The expression of PLDε itself was induced two-fold when seedlings were transferred from N-rich (60 mM) to N-limited (0.6 mM) conditions (Figure 9a, upper panel). Data from Genevestigator also indicated that the expression of PLDε was increased under N-limited conditions, but deficiency in sulfur (S), potassium (K) or phosphorus (P) did not affect its expression (Figure 9b; Li et al., 2006). The expression of genes encoding two nitrate transporters, NRT1.1 and NRT2.1, which coordinate N absorption at different levels of N, were measured by real time PCR. NRT1.1 is regarded as a dual affinity nitrate transporter (Tsay et al., 1993; Huang et al., 1996), whereas NRT2.1 is a high-affinity, low-capacity nitrate transporter (Cerezo et al., 2001). The mRNA level of NRT1.1 was high, whereas that of NRT2.1 was undetectable in all genotypes grown under high N conditions. When seedlings were transferred from 60 mM N to 0.6 mM N, expression of the high-affinity NRT1.1 in OE was higher than that of WT and KO plants (Figure 9a, low panel). When seedlings were germinated and grown with limited N, the level of NRT1.1 expression was lowest in KO while the level of NRT2.1 was highest in OE seedlings (Figure 9a). When nitrate uptake was compared in the genotypes, OE seedlings exhibited 1.5-fold higher nitrate uptake than WT and KO seedlings at 6 h after transferring the seedlings from a nitrate-starved to a 2 mM nitrate medium (Figure 9c). The changes in the nitrate transporter expression and uptake suggest that OE of PLDε promotes nitrogen acquisition.

Figure 9. Changes in nitrate uptake and assimilation in PLDε-altered plants.

(a) The expression levels of PLDε and nitrate transporters NRT1.1 and NRT2.1 under different N conditions by real time PCR. Seedlings grown in 60 or 0.6 mM N for 10 days, or five day-old seedlings on 60 mM N were transferred to 0.6 mM N for five days (60→0.6 mM). Values are means ± SD (n = 3).

(b) Expression of PLDε in nitrate (N), sulfur (S), and potassium (K)-normal and deficient conditions. The data were obtained from Genvestigator.

(c) Nitrate uptake. Values are means ± SD (n = 6).

(d) Activity of enzymes in N assimilation and metabolism. NR (pmol NO₂⁻/min/mg protein), NiR (pmol NH₄⁺/min/mg protein), GS (μmol γ-glutamylhydroxamate/min/mg protein). GDH (μmol NADPH/min/mg protein). Ten day-old seedlings on nitrate were extracted for the assays. Values are means ± SD (n = 3) from one of the three independent experiments.

The conversion of nitrate to amino acids requires several enzymes. The reductions of NO₃⁻ to NO₂⁻ and then to NH₄⁺ are catalyzed sequentially by nitrate reductase (NR) and nitrite reductase (NiR). NH₄⁺ is incorporated into organic molecules by the glutamine synthetase (GS)/Glu synthase pathway (Crawford, 1995; Walch-Liu et al., 2006). The NR activity was similar in the genotypes at 0.6 and 6 mM N, but OE seedlings showed a higher NiR activity than WT and KO at 0.6 and 6 mM N (Figure 9d). Similarly, the activity of GS was higher in OE seedlings (Figure 9d). In contrast, the activity of glutamate dehydrogenase (GDH), which catalyzes oxidative deamination of glutamate, was lower in OE than WT and KO seedlings (Figure 9d). The difference in NiR, GS, and GDH was greatest between OE and WT seedlings grown in low, 0.6 mM N. The results indicate that OE of PLDε improves plant N utilization by enhancing the N assimilation pathway and by decreasing glutamate catabolism.

Discussion

PLD and PA have been implicated in growth promotion through their roles in stimulating cellular growth proponents, such as mTOR, PDK, and mitogen-activated kinases, but direct evidence for promoting growth at an organismal level is lacking (Huang and Frohman, 2007; Foster and Xu, 2003; Wang et al., 2006). The present study shows that PLDε promotes root growth and biomass accumulation. The increase comes from both a bigger cell size and greater cell numbers. Arabidopsis has 12 PLD genes, but increased expression of other PLDs, i.e. PLDα2, PLDα3, or PLDδ, does not result in the overt growth enhancement (results not shown). The analysis of PLDε biochemical properties and expression provide insights to the distinctive effect of PLDε on plant growth. PLDε is the most permissive of all the characterized PLDs in terms of reaction requirements. However, PLDε’s level of expression in vegetative tissue is much lower than that of PLDα1. PLDε lacks a Ca²⁺-C2 interaction, thus, making it less dependent on Ca²⁺ for its activation than are other PLDs. In addition, the exclusive localization in membranes allows PLDε to activate rapidly and to have access to membrane substrates without the relocation to membrane that other PLDs undergo. This could allow overexpressed PLDε to be active, resulting in production of PA that mediates plant response to stimuli.

The results indicate that the enhanced growth by PLDε is mediated via generation of the lipid messenger PA. The activity of PLDε on PA production has been demonstrated by biochemical assays of PLDε activity and by measuring altered PA levels in KO and OE plants. Lipid profiling from root tissue showed that most PA is derived from 34:3- and 34:2-PE, suggesting that unsaturated PA molecular species may be involved in sensing nutrient and osmotic cues to regulate root growth. In addition, the PA effect on growth alterations was corroborated by alcohol suppression of PA production. The rise in signaling PA is often transient (Wang et al., 2006), since PA is not metabolically stable. In addition, the location and timing of PA changes are important to PA functions. This may explain, in part, the result of direct measurements of PA in plants, which shows that while KO of PLDε significantly decreases PA content, OE has a relatively small impact on the PA increase in roots.

The potential for growth is affected greatly by an organism’s ability to sense and utilize available nutrients. Previously, PLDζs have been shown to play a role in plant response to P deficiency by promoting primary root growth and membrane lipid remodeling because KO of PLDζs impedes root growth and P deficiency-induced lipid changes (Cruz-Ramirez et al., 2006; Li et al., 2006). The present data indicate that KO of PLDε resulted in no overt effect on root growth and biomass accumulation under P-, K-, or S-deprived conditions. On the other hand, PLDε-OE roots elongated faster, whereas KO roots elongated slower than WT roots. The opposite effect of PLDε-KO and OE on root elongation occurred in plants grown under agarose- or agar-media containing NO₃⁻:NH₄⁺ (2:1) or NO₃⁻ only. The elongation of lateral roots in response to external N is regarded as a classic indicator of N signaling (Crawford, 1995; Walch-Liu et al., 2006; Hirel et al., 2007). The relationship between N availability, uptake, and root development is well elucidated, and root architecture and growth are critical to nutrient acquisition. The PLDε effect on root growth and morphology are different at different levels of N. At severe N deprivation (0.1 or 0.6 mM), PLDε promotes elongation of primary roots and root hairs, whereas no such effect was observed with sufficient N supply (6 or 60 mM). PLDε-OE plants exhibit increased expression of N transporters. Furthermore, more differences in OE, KO and WT occurred in 0.6 mM N than 6 mM N in nitrate assimilation enzymes NiR, and GS. The results indicate that at severe N deprivation, PLDε plays a role in increasing the root surface area to improve N uptake and utilization. These results are consistent with the notion that N uptake is one of the most critical N utilization activities under N-limiting conditions in a number of plant species (Crawford, 1995; Hirel et al., 2007). At sufficient N supply, PLDε promotes lateral root growth and biomass production. The protein level and N content per unit biomass are not significantly different in WT and PLDε-altered plants. These results suggest that PLDε is involved in N signaling.

N is a vital nutrient for cellular structure and metabolism, and N deficiency is a major limiting factor that adversely impacts agricultural productivity. Great progress has been made in understanding the process of plant N uptake and assimilation in recent years (Crawford, 1995; Walch-Liu et al., 2006; Hirel et al., 2007). However, much less is known about the signaling events in plant response to N availability. The present finding of the growth effect of PLDε raises intriguing questions as to whether PLD and PA-mediated signaling plays a role in connecting the sensing of external nutrient cues at membranes to translation regulation and growth alterations. PA is found to bind to PDK1. The PA-PDK1 interaction activates AGC2-1 kinase to promote root hair growth (Anthony et al., 2004). This is consistent with the present result that overexpressing PLDε promotes root hair growth. PDK1 also phosphorylates S6K (Mahfouz et al., 2006). In mammalian cells, S6K is a nodal point integrating nutrient and stress inputs through translation activity to regulate growth and nutrient usage (Wullschleger et al., 2006). Overexpression of PLD1 increases S6K activity and cell size, whereas suppression of PLD1 reduces cell size (Fang et al., 2003). PA has been shown to bind to S6K and its upstream kinase mTOR in mammalian cells. The PA-mTOR interaction occurs at the site that competes with rapamycin binding (Fang et al., 2001; Lehman et al., 2007). The effect of PA on AtTOR is unknown but plant TOR is insensitive to rapamycin and thus may not bind rapamycin (Mahfouz et al., 2006). It would be of great interest to determine whether PA interaction with multiple kinases may tether these proteins to membranes and assemble multi-protein signaling complex in growth regulation.

The enhanced growth by PLDε also occurs under hyperosmotic stress imposed by high salinity and water deficiency (Figure 3). Activation of PLD has been found under various hyperosmotic conditions (Wang, 2004; Testerink and Munnik, 2005). In nature, simple hyperosmotic stress does not occur because of the interplay between water availability and nutrient transport. It is known that water deficiency impedes nutrient uptake and drought induces N deficiency (Heckathorn et al., 1997; Foyer et al., 1998). Thus, the altered growth in PLDε-OE and KO under hyperosmotic stress could be related to the positive role of PLDε and PA in nutrient signaling. These results raise the exciting possibility that the PLD and PA-based membrane lipid signaling act as a key integrator of multiple plant stresses for optimal cell growth in response to drought and nutrient stresses that calls for further investigation.

Experimental procedures

Isolation of PLDε knockout and HA-tagged PLDs expressed in plants

A PLDε T-DNA insertion mutant was identified from SALK_023603 of the Salk Arabidopsis T-DNA knockout collection from the Ohio State University Arabidopsis Biological Resource Center. The homozygous T-DNA insertion mutant pldε-1 was isolated by PCR-based screening using PLDε-specific primers and a T-DNA left border primer. The loss of transcription of PLDε was confirmed by reverse transcription PCR using PLDε specific primers. To overexpress PLDε in Arabidopsis plants, a DNA fragment for the PLDε gene was amplified from wild type Columbia ecotype Arabidopsis genomic DNA using gene-specific primers. PLDε was fused with a HA-tag to the 3′-end and then cloned into a binary vector pKYLX7 under the control of the 35S promoter. All primers used in this study are listed in Supplemental Table 1 online.

Plant growth and treatments

Surface-sterilized seeds were germinated in MS salt agar or agarose (to remove potential nutrient contaminants) plates for 4 days and then transferred to modified MS salt agar plates containing 0.1, 0.6, 2, 6, and 60 mM N (NO₃⁻: NH₄⁺; 2:1) or NO₃⁻ only. KCl was added to compensate for the lower than normal K⁺ concentration in the media with reduced KNO₃ levels (Martin et al., 2002). Seedlings were grown on plates in a vertical orientation in a growth chamber with a 16-h light/8-h dark cycle, at 23/21°C, under cool fluorescent white light (200 μmol m⁻² s⁻¹). Alternatively, seeds were directly germinated on agar plates and grown without a transfer step. For osmotic stress experiments, surface-sterilized seeds were germinated or 4 day-old seedlings that were transferred to MS salt growth media containing 50, 100 mM NaCl, 5%, 8% PEG, or 100 mM sorbitol. For soil grown plant experiments, plants were grown in growth chambers with 12-h light/12-h dark, 23/21°C, 50% humidity, 200 μmol m⁻²s⁻¹ of light intensity and watered with fertilizer (Scotts 15-5-15 Cal-Mag, 200 ppm nitrogen) once per week (well-fertilized conditions) or only once per life cycle (less well-fertilized). The fertilizer contained 15% total nitrogen (1.2% ammonia, 11.75% nitrate, 2.05% urea), 5% available phosphate, and 15% soluble potassium.

RNA extraction and real-time PCR

Total RNA was extracted from leaves or seedlings using a CTAB method. DNA was removed from RNA by digestion with RNase-free DNaseI (Li et al., 2006). RNA without DNA contamination was used as a template for reverse transcription PCR for the synthesis of cDNA using an iScript kit (Bio-Rad). Quantitative real-time PCR was performed with a MyiQ sequence detection system (Bio-Rad) by monitoring SYBR green fluorescent labeling of double stranded DNA synthesis as described previously (Li et al., 2006). The efficiency of the cDNA synthesis was assessed by real-time PCR amplification of a control gene encoding UBQ10 (At4g05320) and the UBQ10 gene C_t value was 20 ± 0.5. Only cDNA preparations that yielded similar C_t values for the control genes were used for determination of PLD gene expression. The level of PLD expression was normalized to that of UBQ10 by subtracting the C_t value of UBQ10 from the C_t value of PLD genes (Li et al., 2006). Expression levels of genes were normalized by comparison to the UBQ10 gene.

Subcellular fractionation, PLDε-HA purification and PLD activity assays

Proteins were extracted from leaves of four-week-old plants using chilled buffer A (Qin et al., 1997), followed by centrifugation at 6,000 g for 10 min. The supernatant was centrifuged at 100,000 g for 60 min, and the resultant supernatant and pellet were referred to as soluble and microsomal fractions, respectively. The microsomal fraction was separated into the plasma membrane and intracellular membrane fractions by two-phase partitioning as described previously. Marker enzymes for the plasma membrane and intracellular membranes are ATPase and cytochrome c reductase, respectively (Fan et al., 1999). PLDε was purified from PLDε-OE Arabidopsis leaves using HA-antibody affinity chromatography. Briefly, protein extracts were incubated with HA monoclonal antibody (1:500) at 4°C for 2 h. Protein A agarose was added to the mixture and incubated with agitation for 2 h at 4°C. Beads were pelleted and washed four times with a wash buffer containing 0.5% Triton X-100. The purified PLDs were assayed for activity conditions previously defined for PLDα1, β, δ, and ζ1 (Qin et al., 1997; Pappan et al., 1998; Wang and Wang, 2001; Qin and Wang, 2002;). To assay substrate usage, fluorescent NBD-PC, -PE, -PG, and -PS were used, and each reaction contained PLDε-HA isolated from 1 mg leaf proteins under the PLDα1 reaction condition (50 mM Ca²⁺, 0.5 mM SDS, and 2 mM lipids). The resultant lipids were separated on TLC plates and quantified by fluorescence spectrophotometry (excitation at 460 nm, emission at 534 nm) (Pappan et al., 1998).

Subcellular localization of PLDε-YFP

Arabidopsis PLDε cDNA was cloned by PCR amplification from an Arabidopsis leaf cDNA RACE pool. The amplified PCR product was directly cloned into the p35S_FAST/YFP which was derived from p35S_FAST (Ge et al., 2005) by introducing eYFP (Sigma). Agroinfiltration procedure for the transient protein expression in tobacco leaves was performed according to the method described by Voinnet et al. (2003). In brief, Agrobacterium tumefaciens strains C58C1 carrying binary constructs were grown to stationary phase at 28°C in YEP medium. Bacterial cells were collected by centrifugation at 6000g for 15 min at room temperature and resuspended in 10 mM MES (pH 5.7), 10 mM MgCl₂ and 150 mg/mL of acetosyringone. For coinfiltration, Agrobacterium cultures carrying different constructs were mixed at equal ratio and left for 3 h at room temperature before infiltration. Leaves of 3-week-old Nicotiana benthamiana plants were infiltrated with the bacterial cultures through abaxial air spaces. In all experiments, Agrobacterium C58C1 carrying the 35S:p19 construct (Voinnet et al., 2003) was coinfiltrated to achieve maximum level of protein expression. The YFP fluorescence was observed under microscope (Zeiss LSM 510 confocal/multiphoton microscope).

Immunoblotting of PLD

Proteins were extracted from leaves or seedlings as previously described (Qin et al., 1997). Homogenates were centrifuged at 6000 g for 10 min. For PLD-HA detection, supernatant proteins (30 μg/lane) were separated by 8% (w/v) SDS-PAGE gel, followed by transfer to a PVDF membrane. Membranes were blotted with anti-HA antibody (1:1000) and then incubated with a secondary antibody as described (Zhang et al., 2004).

Leaf cell size and lipid analysis

Leaf discs (0.5 cm diameter) were taken from the middle of fully expanded leaves from 5 week-old plants, and fixed in ethanol:glacial acetic acid (3:1, v/v) for 30 min. Leaf discs were transferred sequentially to 75%, 50%, 25%, and 0% ethanol (v/v) for 15 min each. Cell sizes were measured under a microscope using the IMAGEPRO software (Media Cybernetics, Silver Spring, MD). Lipids were extracted and analyzed by ESI-MS/MS, and levels of PA, PC, PE, PG, PI, and PS molecular species were added to determine lipid content for each head-group class as previously described (Devaiah et al., 2006).

Nitrate uptake and assimilation enzymes

Nitrate uptake was assayed according to the method by Doddema and Telkamp (1979). To measure the N-metabolizing enzymes, 10 day-old seedlings were homogenized with chilled extraction buffer (100 mM K₂HPO₄, pH 7.5, 1 mM DTT, 1 mM EDTA and 10 mM cysteine), followed by centrifugation at 10,000 g at 4°C for 15 min. The supernatant was used for assaying the activities of NR (Wilkinson and Crawford, 1991), NiR (Takahashi et al., 2001), GS (Rhodes et al., 1975), and GDH (Turano et al., 1996), according to the methods previously described.

Abbreviations footnote

ABA

abscisic acid

DAG

diacylglycerol

NBD-PC

1-oleoyl, 2-12[(7-nitro-2-1,3-benzoxadiazol-4-yl) amino]dodecanoyl phosphatidylcholine

PA

phosphatidic acid

PC

phosphatidylcholine

PE

phosphatidylethanolamine

PtdBut

phosphatidylbutanol

PLD

phospholipase D

PI

phosphatidylinositol

PS

phosphatidylserine