A family of six genes encoding acyl-CoA-binding proteins (ACBPs), ACBP1–ACBP6, has been characterized in Arabidopsis thaliana. In this study, we demonstrate that ACBP1 promotes abscisic acid (ABA) signaling during germination and seedling development. ACBP1 was induced by ABA, and transgenic Arabidopsis ACBP1-over-expressors showed increased sensitivity to ABA during germination and seedling development, whereas the acbp1 mutant showed decreased ABA sensitivity during these processes. Subsequent RNA assays showed that ACBP1 over-production in 12-day-old seedlings up-regulated the expression of PHOSPHOLIPASE Dα1 (PLDα1) and three ABA/stress-responsive genes: ABA-RESPONSIVE ELEMENT BINDING PROTEIN1 (AREB1), RESPONSE TO DESICCATION29A (RD29A) and bHLH-TRANSCRIPTION FACTOR MYC2 (MYC2). The expression of AREB1 and PLDα1 was suppressed in the acbp1 mutant in comparison with the wild type following ABA treatment. PLDα1 has been reported to promote ABA signal transduction by producing phosphatidic acid, an important lipid messenger in ABA signaling. Using lipid profiling, seeds and 12-day-old seedlings of ACBP1-over-expressing lines were shown to accumulate more phosphatidic acid after ABA treatment, in contrast to lower phosphatidic acid in the acbp1 mutant. Bimolecular fluorescence complementation assays indicated that ACBP1 interacts with PLDα1 at the plasma membrane. Their interaction was further confirmed by yeast two-hybrid analysis. As recombinant ACBP1 binds phosphatidic acid and phosphatidylcholine, ACBP1 probably promotes PLDα1 action. Taken together, these results suggest that ACBP1 participates in ABA-mediated seed germination and seedling development.
Abscisic acid (ABA) is a phytohormone that is known to control physiological processes including seed dormancy and germination, vegetative development and reproduction (Finkelstein et al., 2002; Fujii and Zhu, 2009). In addition to its role in regulating plant growth and development, ABA also plays important roles in sensing and adaptation to abiotic and biotic stresses (Yamaguchi-Shinozaki and Shinozaki, 2006; Hirayama and Shinozaki, 2007; Ton et al., 2009). Given that these stresses severely affect plant productivity and agriculture, the ABA response has been investigated, and regulators in ABA signal transduction, including G proteins, protein kinases/phosphatases, NAD(P)H oxidases and several transcriptional factors, have been identified (Hirayama and Shinozaki, 2007; Cutler et al., 2010; Osakabe et al., 2010).
The most recent breakthroughs in ABA signaling include reports on a family of START domain proteins, REGULATORY COMPONENTS OF ABA RECEPTOR/PYRABACTIN RESISTANCE1/PYRABACTIN RESISTANCE1-LIKE (RCAR/PYR/PYL), as ABA receptors (Ma et al., 2009; Park et al., 2009; Santiago et al., 2009; Nishimura et al., 2010). These receptors bind to type 2C protein phosphatases (PP2Cs) in the presence of ABA, and facilitate the formation of RCAR/PYR/PYL–ABA–PP2C complexes to inhibit PP2C activities. This inhibition of PP2Cs releases SNF1-related protein kinase 2 proteins (SnRK2s), allowing them to phosphorylate and thereby regulate their downstream targets, the ABA-responsive element binding factors (Sheard and Zheng, 2009; Cutler et al., 2010; Raghavendra et al., 2010).
In addition to ABA receptors, various second messengers that contribute to ABA signaling include Ca2+, phosphatidic acid (PA), diacyglycerol, reactive oxygen species, phosphoinositides and cyclic adenosine 5′-diphosphate ribose (Wu et al., 1997; Jacob et al., 1999; Lemtiri-Chlieh et al., 2000; Pei et al., 2000; Zhang et al., 2004, 2009; Peters et al., 2010). Among these second messengers, PA has been reported to inhibit the activity of ABSCISIC ACID INSENSITIVE1 (ABI1), a major negative regulator of ABA responses that belongs to the PP2C family (Gosti et al., 1999; Zhang et al., 2004; Mishra et al., 2006). In Arabidopsis, a major source of PA derived from hydrolysis of membrane lipids by phospholipase D proteins, such as PLDα1 (Li et al., 2009). A previous study has shown that recombinant Arabidopsis ACYL-COA-BINDING PROTEIN1 (ACBP1) binds PA in vitro (Du et al., 2010a), raising the possibility of ACBP1 participation in ABA signaling.
The first characterized ACBP, a neuropeptide isolated from rat brain, was named diazepam-binding inhibitor because it was found to competitively inhibit benzodiazepine binding (Guidotti et al., 1983). Similar small (10-kDa) cytosolic ACBPs have been identified in eukaryotes and some prokaryotes (Burton et al., 2005). Subsequently, larger ACBPs were identified in eukaryotic species (Leung et al., 2004; Xiao and Chye, 2009, 2011a; Fan et al., 2010; Elle et al., 2011). ACBPs bind acyl-CoA esters, play an important role in the maintenance and transport of acyl-CoA esters, and participate in lipid biosynthesis, membrane biogenesis and regulation of gene expression and enzyme activities, as well as in development and stress responses, in Caenorhabditis elegans (Elle et al., 2011), mammals (Fan et al., 2010) and plants (Xiao and Chye, 2009, 2011a; Yurchenko and Weselake, 2011). A family of six genes encoding ACBPs (ACBP1–ACBP6) has been identified in Arabidopsis (Xiao and Chye, 2009, 2011a). These ACBPs vary in size from 10 to 73 kDa, and participate in lipid metabolism, plant development and stress responses (Xiao and Chye, 2009, 2011a). The smallest 10-kDa ACBP6 is induced by cold, and its over-expression enhances freezing tolerance in transgenic Arabidopsis (Chen et al., 2008). Extracellular-localized ACBP3 regulates autophagy-mediated leaf senescence (Leung et al., 2006; Xiao and Chye, 2010; Xiao et al., 2010) and functions in plant defense against biotrophic pathogens (Xiao and Chye, 2011b). A mutant lacking the large cytosolic ACBP4 displayed a decrease in membrane lipids (Xiao et al., 2008b).
ACYL-COA-BINDING PROTEIN1 is a member of this Arabidopsis ACBP family and has a highly conserved homolog, ACBP2; each protein has an N-terminal hydrophobic transmembrane domain and C-terminal ankyrin repeats (Chye, 1998; Chye et al., 1999, 2000; Xiao and Chye, 2011a). Previous studies have confirmed that ACBP1 and ACBP2 are subcellularly localized to the plasma membrane and endoplasmic reticulum (Chye, 1998; Chye et al., 1999; Li and Chye, 2003, 2004). They are also localized to the embryos of developing seeds (Chye et al., 1999; Chen et al., 2010), and are essential in lipid metabolism during embryogenesis because the acbp1 acbp2 double mutant, but not the single mutants, is embryo-lethal (Chen et al., 2010). ACBP1 and ACBP2 are both induced by heavy metal stress, and their over-expression in Arabidopsis enhanced tolerance to lead [Pb(II)] and cadmium [Cd(II)], respectively (Xiao et al., 2008a; Gao et al., 2009). Their ankyrin repeats mediate protein–protein interactions (Li and Chye, 2004; Gao et al., 2009, 2010; Xiao and Chye, 2011a). ACBP2 interacted with FARNESYLATED PROTEIN6 (AtFP6) and LYSOPHOSPHOLIPASE2 (lysoPL2) at the plasma membrane in co-localization analyses using DsRed-tagged ACBP2 and GFP-tagged AtAFP6 or lysoPL2 (Gao et al., 2009, 2010). ACBP2 also interacts with an ethylene response factor, RELATED TO APETALA2.3 (RAP2.3) (Li and Chye, 2004), while both ACBP1 and ACBP2 interact with another transcription factor, RAP2.12, and have been suggested to play roles in oxygen sensing (Licausi et al., 2011). Alteration in ACBP1 (but not in ACBP2) expression is known to affect freezing tolerance (Du et al., 2010b), suggesting that, although ACBP1 and ACBP2 share redundant functions, each displays distinct roles in response to environmental stresses.
In this study, ACBP1pro::GUS was observed to be expressed in embryos, germinating seedlings, roots, floral parts (sepals and pistil) and trichomes. ACBP1 over-production promoted ABA signaling during germination and seedling development by regulation of ABA-related genes, in contrast to the acbp1 mutant, which displayed decreased sensitivity to ABA during these processes. This study suggests that Arabidopsis ACBP1 participates in ABA-mediated seed germination and seedling development.
Over-expression of ACBP1 up-regulates the expression of PLDα1 in 12-day-old seedlings
During freezing stress, expression of PLDα1, a major phospholipase that generates PA, is known to be up-regulated relative to the wild type in 5-week-old rosettes of ACBP1-over-expressing (ACBP1-OX) lines (Du et al., 2010a), in which ACBP1 is expressed from the CaMV 35S promoter (Xiao et al., 2008a). As PLDα1 plays a positive role in ABA signaling, we examined the expression of PLDα1 in the acbp1 mutant and ACBP1-OX lines upon ABA treatment by quantitative real-time RT-PCR. The results revealed that, after ABA treatment, over-expression of ACBP1 (confirmed for ACBP1-OX lines by Western blotting, Xiao et al., 2008a) led to enhanced PLDα1 expression in comparison with the wild type (Figure 1), whereas its expression in the acbp1 mutant was lower than in the wild type (Figure S1).
Over-expression of ACBP1 results in PA accumulation, and rACBP1 binds PA in vitro
We compared the lipid profiles of 5-week-old wild type Arabidopsis rosettes, grown under normal conditions, with those of ACBP1-OX2 and the acbp1 mutant (Figure 2a,b). ACBP1-OX2 showed accumulation of some PA molecular species (34:3 PA, 34:2 PA, 36:6 PA, 36:5 PA and 36:3 PA) as well as total PA, but showed a decrease in total phosphatidylcholine (PC) (Figure 2a,b). As PA is known to regulate ABA signaling (Zhang et al., 2004; Mishra et al., 2006), we investigated whether ACBP1 is involved in the ABA response. It has been reported that recombinant ACBP1 (rACBP1) binds PA and PC in filter-binding assays (Chen et al., 2010; Du et al., 2010a). Here, lipid pull-down assays were performed to confirm PA and PC binding for ACBP1. Purified recombinant ACBP1 (with ACBP2 as a control) was tested with PA and PC beads, and the results show that rACBP1 binds both PA and PC while rACBP2 binds PC only (Figure 2c), consistent with the results of filter-binding assays (Figure S2).
ACBP1 expression is induced by ABA and drought treatment in 12-day-old seedlings
The expression of ACBP1 was examined by quantitative RT-PCR in 12-day-old seedlings before and after ABA treatment. Significant induction of ACBP1 was observed 1, 3, 6 and 24 h after ABA treatment (Figure 2d). As ABA is a well-known key factor regulating drought tolerance (Schroeder et al., 2001; Yamaguchi-Shinozaki and Shinozaki, 2006), ACBP1 expression in 12-day-old seedlings was then tested in response to drought treatment. In quantitative RT-PCR analyses, ACBP1 was induced only at 3 and 6 h following drought treatment, but rapidly declined to the untreated level at 24 h (Figure 2e). The induction of ACBP1 by ABA and to a lesser extent by drought treatment suggests a potential role for ACBP1 in these processes.
ACBP1 is expressed in embryos and trichomes
ACYL-COA-BINDING PROTEIN1 has been previously shown to accumulate in developing embryos by immunogold labeling analysis using ACBP1-specific antibodies (Chye et al., 1999). The spatial expression of ACBP1 was investigated in this study by generating transgenic Arabidopsis lines expressing a ACBP1pro::GUS fusion. GUS staining showed ACBP1 expression in embryos at the heart-to-torpedo transition stage (Figure 3a) and an increase from the bent-cotyledon (Figure 3b) to maturation stages (Figure 3c). At the maturation stage, ACBP1 was strongly expressed throughout the whole embryo (Figure 3c).
In addition to its expression during embryo development, 3-week-old transgenic Arabidopsis showed GUS staining in various organs. ACBP1pro::GUS was expressed in trichomes of rosette leaves (Figure 3d–f) and in lateral root primordia and vascular bundles (Figure 3g). In flowers (Figure 3h), ACBP1pro::GUS was expressed in sepals to some extent (Figure 3i), and to a greater extent in the stigma, style and ovary (Figure 3j,k). No expression was observed in the petals (Figure 3i) or stamens (Figure 3j). Tests using fluorescent GUS substrate further confirmed GUS expression in the leaf trichomes of 12-day-old seedlings (Figure 3l,m).
Microarray data (Zimmermann et al., 2004; https://www.genevestigator.ethz.ch) demonstrated that ACBP1 is highly expressed during early seedling development (1–5.9 days post-germination). Decreased expression occurs in the subsequent vegetative stages (6–44.9 days post-germination) as well as the seed production stages (45–50 days post-germination). Histochemical and quantitative RT-PCR analyses were performed to further examine ACBP1 expression during seed germination and early seedling development. ACBP1pro::GUS lines showed GUS staining throughout the embryo during germination (days 0 and 1) and during the first two days post-germination (days 2 and 3; Figure S3a). As seedling establishment progressed, GUS expression decreased from day 3, although staining was still observed in the cotyledon, hypocotyl, root tip and vascular bundles of the seedling (Figure S3a). ABA suppressed embryo growth, and ACBP1pro::GUS expression was retained in ABA-treated embryos, but the GUS signal was reduced upon seedling growth in the absence of ABA (Figure S3a). Furthermore, quantitative RT-PCR analyses produced similar results showing that ACBP1 expression may be partially rescued by ABA treatment (Figure S3b). Taken together, GUS staining driven by ACBP1 expression suggests participation of ACBP1 in seed germination and early seedling development.
ACBP1 over-expression increases ABA sensitivity during germination and seedling development, but the acbp1 mutant shows decreased ABA sensitivity
Because ACBP1 is induced by ABA, a key plant hormone regulating seed dormancy and germination, freshly harvested seeds were used in germination assays to assess seed dormancy amongst the wild type, acbp1 mutant and ACBP1-OX lines (Figure 4a). The results showed that acbp1 mutant seeds were less dormant than the wild type, while ACBP1-OX lines were more dormant (Figure 4a). The germination of dry after-ripened seeds was tested in the presence and absence of ABA treatment (Figure 4b–e). For dry after-ripened seeds, no significant differences were noted (Figure 4b) between wild type and transgenic seeds germinated on Murashige and Skoog (MS) medium (Murashige and Skoog, 1962). However, in the presence of exogenous ABA (0.75, 1.5 and 2 μm), ACBP1-OX seeds were more sensitive to ABA inhibition in comparison to the wild type (Figure 4c–e). In contrast, acbp1 mutant seeds were more resistant to inhibition by ABA (1.5 and 2 μm) and completed germination faster than the wild type (Figure 4d,e).
Abscisic acid is also known to regulate seedling development and root growth in addition to seed dormancy and germination (Finkelstein et al., 2002; Fujii and Zhu, 2009). The role of ACBP1 in root development was investigated by growing 4-day-old seedlings on MS medium followed by transfer to ABA-containing MS medium for a further 7 days of growth (Figure 5a–d). The results revealed that roots of ACBP1-OX plants were significantly shorter than those of the wild type on ABA treatment (Figure 5b–d). However, it was observed that root growth of the acbp1 mutant was not as affected by ABA (150 μm) in comparison to the wild type (Figure 5c,d). These results demonstrate that ACBP1 affects ABA signaling during seed dormancy, germination and seedling development.
Given that ABA plays a crucial role in sensing and adaptation to abiotic stresses, such as drought and high salinity (Shinozaki and Yamaguchi-Shinozaki, 2007), and that ACBP1 is drought-inducible, the acbp1 mutant and ACBP1-OX transgenic plants (OX-1 and OX-2) were subjected to dehydration and high-salinity stresses. However, the ACBP1-OX lines and the acbp1 mutant did not show any significant differences from the wild type in these treatments (Figures S4 and S5). Furthermore, while reactive oxygen species (ROS) are known to be produced in guard cells exposed to ABA treatment (Kwak et al., 2003), investigations of ROS production in guard cells of 5-week-old wild type, acbp1 mutant and ACBP1-OX plants using the fluorescent dye dichlorofluorescin showed no significant differences amongst the various genotypes after ABA treatment (Figure S6).
Alteration in ACBP1 expression regulates ABA signaling genes
Analyses by quantitative RT-PCR were performed on 12-day-old Arabidopsis seedlings to examine whether an alteration in ACBP1 expression affects the expression of ABA signaling genes. The results revealed that expression of AREB1, RD29A, MYC2 and PLDα1 was significantly up-regulated in ACBP1-OX lines, with or without ABA treatment (Figure 1). In contrast, expression of AREB1 and PLDα1 was suppressed in the acbp1 mutant in comparison to the wild type following ABA treatment (Figure S1). The expression of 9-CIS-EPOXYCAROTENOID DIOXYGENASE3 (NCED3), ABA-DEFICIENT2 (ABA2), RESPIRATORY BURST OXIDASE HOMOLOGUE D (AtrbohD) and AtrbohF was up-regulated in ACBP1-OX lines in the absence of ABA treatment, but showed no significant differences from the wild type after ABA treatment (Figure 1). ABI1, which encodes a key PP2C that negatively regulates ABA signaling, was suppressed in ACBP1-OX lines following treatment (Figure 1). In addition, expression of RECEPTOR-LIKE PROTEIN KINASE1 (RPK1) was similar between the wild type and ACBP1-OX lines (Figure 1). Given that ACBP1 over-expression increased the expression of two ABA biosynthesis genes, NCED3 and ABA2, and knockout of ACBP1 down-regulated ABA2 expression (Figure S1), we analyzed ABA content in seeds and leaves of Arabidopsis plants. However, no significant differences were seen in comparison to the wild type (Figure S7). When rACBP1 was tested to determine whether it binds ABA in vitro, binding assays did not reveal any binding between them (Figure S8).
ACBP1 interacts with PLDα1
It has been reported that PLDα1 promotes ABA signaling by inhibition of ABI1 activity (Zhang et al., 2004). The expression of PLDα1 was up-regulated in ACBP1-OX lines but down-regulated in the acbp1 mutant following ABA treatment (Figure 1 and Figure S1). Bimolecular fluorescence complementation (BiFC) assays were subsequently used to test whether ACBP1 interacts with PLDα1 in vivo. The nuclear basic leucine zipper (bZIP) transcription factor bZIP63 was translationally fused to the N-terminal 155-amino acid portion of YFP (NYFP) and the C-terminal 84-amino acid portion of YFP (CYFP) to generate bZIP63-NYFP and bZIP63-CYFP, respectively, which were then used as the positive controls (Walter et al., 2004). Subsequently, ACBP1 was fused to NYFP, while PLDα1 was fused to CYFP, to generate ACBP1-NYFP and PLDα1-CYFP, respectively. Transient co-expression of these two proteins in Nicotiana benthamiana (tobacco) leaf cells revealed YFP fluorescence signals at the plasma membrane (Figure 6a), indicating that ACBP1 interacts with PLDα1 in vivo. A signal was also detected in the nuclei from co-expression of bZIP63-NYFP and bZIP63-CYFP (Figure 6b), while negative control pairs (P35S::NYFP + P35S::CYFP, ACBP1− NYFP + P35S::CYFP and P35S::NYFP + PLDα1−CYFP) did not interact (Figure 6c–e). To further confirm the interaction between ACBP1 and PLDα1, yeast two-hybrid analysis was performed using the Matchmaker Gold yeast two-hybrid system (Clontech). The N-terminal transmembrane domain (amino acids 1–40) of ACBP1 was removed, and the resulting Δ40ACBP1 sequence was used for the assay. Growth and α-galactosidase activity were observed in the positive control and the Δ40ACBP1/PLDα1 interaction, but not in the negative control, confirming interaction between ACBP1 and PLDα1 (Figure 6f).
ACBP1-OX lines accumulate PA, but the acbp1 mutant does not after ABA treatment
Increasing data suggest that PA is a vital signaling messenger in plants in response to ABA (Zhang et al., 2004, 2009; Katagiri et al., 2005; Peters et al., 2010). We have previously reported that rACBP1 binds PA in vitro, and ACBP1-OX lines accumulate PA after freezing treatment (Du et al., 2010a). Here, we investigated whether alterations in ACBP1 expression affect PA content in both seeds and seedlings in response to ABA treatment. In the control group of germinating seeds (Figure 7a), PA molecular species (34:2 PA, 34:3 PA, 36:2 PA and 36:3 PA) were increased in both ACBP1-OX lines in comparison to the wild type, whereas 32:0 PA decreased in the acbp1 mutant. After ABA treatment (1 μm for 12 h), some PA molecular species (34:1 PA, 34:4 PA and 36:4 PA) accumulated in both ACBP1-OX lines in contrast to decreases in 36:3 PA and 36:4 PA in the acbp1 mutant (Figure 7b). The total PA of ABA-treated seeds was lower (0.82-fold) in the acbp1 mutant but higher (1.36-fold) in ACBP1-OX lines in comparison with the wild type (Table 1). The total PA of untreated ACBP1-OX seeds was also higher (1.67-fold) than that of the wild type (Table 1).
Table 1. Total amount of phospholipids and galactolipids in wild type, the acbp1 mutant, and ACBP1-OX (OX-1 and OX-2) seeds in response to ABA treatment The values are means ± SD (nmol mg−1 dry weight; n =4)
0 μm ABA (12 h)
1 μm ABA (12 h)
Significant differences (P <0.05) in comparison to the wild type in the same experiment are shown in bold.
Value lower than wild type in the same experiment (P <0.05).
Value higher than wild type in the same experiment (P <0.05).
In addition to PA, significant differences were also observed in other lipids amongst the wild type, acbp1 mutant and ACBP1-OX lines. In untreated seeds of the acbp1 mutant, total phosphatidylinositol (PI) and phosphatidylserine (PS) levels were higher than the wild type, whereas digalactosyldiacylglycerol (DGDG) and monogalactosyldiacylglycerol (MGDG) levels were lower (Table 1). An elevated PI content was also observed in the acbp1 mutant in comparison to the wild type following ABA treatment. In untreated but not in ABA-treated ACBP1-OX seeds, total PC increased compared with the wild type (Table 1).
We also analyzed lipids from 12-day-old seedlings of an untreated control (Figure 8a) and two ABA-treated groups (100 μm; 1 or 3 h) (Figure 8b,c). After 1 h of 100 μm ABA treatment, 32:0 PA and 34:6 PA were significantly elevated in both ACBP1-OX lines, while other PA molecular species were slightly increased in comparison to the wild type (Figure 8b). After 3 h of ABA treatment, more significant differences were observed between the wild type and ACBP1-OX lines in terms of PA molecular species (34:2 PA, 36:3 PA, 36:4 PA, 36:5 PA and 36:6 PA). In contrast, the acbp1 mutant showed a significant decrease in 32:0 PA and 36:2 PA without ABA treatment, and the levels of PA molecular species (32:0 PA, 34:4 PA and 34:6 PA) were lower than those in the wild type after 3 h of ABA treatment (Figure 8c). Correspondingly, total PA content was elevated in the ACBP1-OX lines in comparison with the wild type after ABA treatment (3 h), but decreased in the acbp1 mutant (Table 2). There were also some significant differences in other lipids amongst the wild type, acbp1 mutant and ACBP1-OX lines after treatment with ABA (100 μm; Table 2). In the acbp1 mutant, total PS (after 1 h treatment) and levels of glycerolipids, DGDG and MGDG (after 3 h treatment) were lower than in the wild type (Table 2). However, a decrease in PS (after 1 h treatment) was also observed in both ACBP1-OX lines, and no significant differences were detected in DGDG and MGDG (after 3 h treatment) between the wild type and ACBP1-OX lines (Table 2).
Table 2. Total amount of phospholipids and galactolipids in 12-day-old wild type, the acbp1 mutant, and ACBP1-OX (OX-1 and OX-2) seedlings in response to ABA treatment
0 μm ABA, 1 h
100 μm ABA, 1 h
100 μm ABA, 3 h
The values are means ± SD (nmol mg−1 dry weight; n =5). Significant differences (P <0.05) in comparison to the wild type in the same experiment are shown in bold.
Value lower than wild type in the same experiment (P <0.05).
Value higher than wild type in the same experiment (P <0.05).
Phosphatidic acid is an important signaling messenger in response to both biotic and abiotic stresses in plants (Testerink and Munnik, 2005; Li et al., 2009). Of the six recombinant His6-tagged ACBPs tested for phospholipid binding using filter-binding assays, all rACBPs bound PC but only rACBP1 bound PA (Figure 2c and Figure S2) (Chen et al., 2008; Xiao et al., 2009, 2010; Du et al., 2010a). A previous study (Du et al., 2010a) indicated that over-expression of ACBP1 increases the expression of PLDα1, which encodes a major phospholipase that generates PA, and PA is known to accumulate upon freezing stress (Welti et al., 2002; Li et al., 2008). It has also been reported that ABA stimulates PLDα1 activity and enhances PA production in leaf protoplasts (Zhang et al., 2004). As a member of the PP2C family, ABI1 works as a principal negative regulator of ABA responses, as demonstrated by studies using mutants of ABI1 (Gosti et al., 1999; Merlot et al., 2001). Suppression of PLDα1 results in ABA insensitivity, whereas PLDα1-over-expressing Arabidopsis lines shows enhanced ABA sensitivity (Sang et al., 2001; Zhang et al., 2004). PLDα1-derived PA has been suggested to directly bind to ABI1, resulting in suppression of the activity of ABI1, which is tethered to the plasma membrane, limiting its access to nuclear factors (Zhang et al., 2004; Li et al., 2009). However, it is unclear how PLDα1-derived PA is retained in the vicinity of the plasma membrane. Given that both ACBP1 and PLDα1 are localized to the plasma membrane, and rACBP1 binds PA in vitro (Chye et al., 1999; Fan et al., 1999; Li and Chye, 2003; Du et al., 2010a), it is possible that ACBP1 may maintain membrane association of PAs released by PLDα1 at the plasma membrane. Previous studies have shown that ACBP1 is localized at the plasma membrane and ER by Western blot analysis (Chye, 1998), immunoelectron microscopy (Chye et al., 1999) and confocal microscopy using ACBP1–GFP fusions (Li and Chye, 2003). PLDα1 is also located at the plasma membrane and intracellular membranes as verified by Western blot analysis (Fan et al., 1999). Western blot analysis has also shown that ACBP1 accumulates in siliques, roots, stems and seeds, with highest accumulation in the siliques and seeds (Chye, 1998). While membrane PLDα1 is expressed in siliques, roots, stems, flowers, young leaves and seedlings, the highest accumulation occurred in siliques (Fan et al., 1999). Also, the highest activity of membrane-associated PLDα1 was detected in the siliques (Fan et al., 1999), consistent with its protein accumulation profile. Furthermore, ACBP1 interacts with PLDα1 as shown in this study, and rACBP1 is known to also bind PC (Chen et al., 2010). Given that rACBP1 is the only ACBP of six Arabidopsis ACBPs that binds both PA and PC, it is possible that ACBP1 facilitates PC/PA exchange and promotes PLDα1 action. Previously, Gao et al. (2010) showed that: (i) ACBP2 interacts with the enzyme lysoPL2, (ii) rACBP2 binds lysoPC, the substrate of lysoPL2, and (iii) this enhances lysoPL2 function in oxidative stress. Also, in vitro studies have demonstrated that the Brassica napus 10-kDa ACBP positively affects the activity of Arabidopsis LYSOPHOSPHATIDYLCHOLINE ACYLTRANSFERASE for transfer of the acyl group from PC into acyl-CoA (Yurchenko et al., 2009).
Our results demonstrate that, following ABA treatment of 12-day-old seedlings, expression of PLDα1 is enhanced in ACBP1-OX lines but repressed in the acbp1 mutant, and expression of ABI1 is repressed in ACBP1-OX lines but increased slightly in the acbp1 mutant. In addition, lipid analysis results showed that ACBP1-OX lines accumulated PA molecular species in comparison with the wild type in response to ABA treatment of 12-day-old seedlings. This increase of PA in ACBP1-OX lines may facilitate inhibition of ABI1 activity as more PA is available to bind ABI1. In contrast, the acbp1 mutant has a lower PA content (0.9-fold) than the wild type after ABA treatment. PA also promotes ABA signaling during seed germination (Katagiri et al., 2005). In this study, we report that germinating seeds of ACBP1-OX lines have a higher PA content than the wild type with or without ABA treatment, while the acbp1 mutant seeds have a lower PA content than the wild type. Similar PA changes were observed in 5-week-old rosettes of the acbp1 mutant and ACBP1-OX lines when they were subjected to freezing treatment: the total PA content was lower (0.5-fold) in the acbp1 mutant but higher (mean 1.7-fold) in ACBP1-OX lines (Du et al., 2010a). These PA changes in response to freezing treatment were also associated with PLDα1 expression (suppressed in the acbp1 mutant but elevated in 5-week-old rosettes of ACBP1-OX lines; Du et al., 2010a), consistent with the changes in PLDα1 expression following ABA treatment observed in this study. However, it should be noted that, perhaps due to differences in developmental stages, 5-week-old rosettes did not display up-regulation of PLDα1 in ACBP1-OX lines in the absence of freezing treatment (Du et al., 2010a), but PLDα1 up-regulation was observed in the untreated 12-day-old seedlings in this study. Given that PLDα1 expression is closely related to ACBP1 expression in ABA and freezing treatment, the interaction between ACBP1 and PLDα1 was confirmed here by BiFC assays using fluorescence-tagged derivatives, and this interaction was found to occur at the plasma membrane.
Previous studies have shown that ABI1 is a negative regulator of ABA responses during seed germination and vegetative growth (Nishimura et al., 2010). ABI1 activity also negatively regulates seed dormancy and drought adaptive responses (Gosti et al., 1999; Merlot et al., 2001; Murata et al., 2001). PLDα1-derived PA promotes ABA responses by interacting with ABI1 (Zhang et al., 2004; Mishra et al., 2006). PLDα1-derived PA also stimulates the activities of AtrbohD and AtrbohF (Zhang et al., 2009). These two proteins are known to function in ABA signaling during seed germination and root growth because an atrbohD atrbohF double mutant is less sensitive to ABA inhibition during these processes (Kwak et al., 2003). They are also vital for ROS production in leaves in response to pathogen infection (Torres et al., 2002) and in guard cells following ABA treatment (Kwak et al., 2003). However, our results showed that, while alteration in ACBP1 expression affected seed dormancy and responsiveness to ABA during germination and seedling development, it did not affect drought tolerance in Arabidopsis. Expression of ACBP1 was induced transiently at 3 and 6 h after drought treatment, but declined to a level similar to the untreated control at 24 h. In contrast, ABA-induced ACBP1 expression was detected at 1, 3, 6 and 24 h. Furthermore, ACBP1pro::GUS expression was not detected in the guard cells in histochemical GUS assays, ACBP1-OX lines showed no phenotypic changes upon drought treatment, and guard cell ROS amounts did not vary among genotypes, indicating that ACBP1 may not function in drought tolerance.
ACYL-COA-BINDING PROTEIN1 over-expression up-regulated the expression of AREB1 and MYC2 encoding, two transcriptional activators in ABA signaling (Abe et al., 2003; Fujita et al., 2005; Yoshida et al., 2010). Over-expression of ACBP1 also increases the expression of RD29A, an ABA-responsive gene (Yamaguchi-Shinozaki and Shinozaki, 1994, 2006). AREB1 has been identified to be a putative interactor of ACBP1 by yeast two-hybrid screening (Tse, 2005), and the interaction between ACBP1 and AREB1 was confirmed by a 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) filter assay (Tse, 2005). Furthermore, given the recent findings of an interaction between ACBP1/ACBP2 and the transcription factor RAP2.12 (Licausi et al., 2011), which is consistent with our earlier observation of ACBP2 interaction with another RAP (RAP2.3) at the plasma membrane via its ankyrin repeats (Li and Chye, 2004), it is feasible that ACBP1 may work together with ABA-responsive transcription factors in ABA signaling.
The over-expression of ACBP1 not only promotes ABA signaling, but has been reported to enhance Pb(II) tolerance in Arabidopsis (Xiao et al., 2008a). Although rACBP1 has been observed to bind Pb(II) (Xiao et al., 2008a), and ACBP1-OX plants accumulate Pb(II) in their shoots, the mechanism of ACBP1 function in trafficking and sequestration of Pb(II) is yet unclear. In this study, we used GUS assays to demonstrate that ACBP1 is expressed in trichomes. It has been reported that plants grown on industrially contaminated soil (containing a high concentration of heavy metals) are enriched in trichomes (Singh and Sinha, 2004; Gupta and Sinha, 2009), suggesting the significance of trichomes in sequestering heavy metals. As the bases of plant trichomes are known to accumulate Pb(II) (Martell, 1974; Lei et al., 2008), ACBP1 expression may be related to this potential function in leaf trichomes.
In summary, this study shows that ACBP1 promotes ABA signaling in germination as well as in seedling development. Fluorescence-tagged ACBP1 (ACBP1–NYFP) interacted with PLDα1–CYFP at the plasma membrane. The elevation of PLDα1 expression in ACBP1-OX lines and its repression in the acbp1 mutant in response to ABA treatment, as indicated by quantitative RT-PCR, provides an explanation for the PA increase in ACBP1-OX lines and the corresponding decrease in the acbp1 mutant. The accumulation of PA in ACBP1-OX lines probably inhibits ABI1 activity and promotes ABA responses (Figure 9). Also, ACBP1 over-expression enhanced the expression of some ABA-responsive genes (AREB1, MYC2 and RD29A), thereby promoting ABA signaling (Figure 9).
Plant materials and growth conditions
The acbp1 mutant and ACBP1-OX lines (OX-1 and OX-2) were derivatives of Arabidopsis thaliana ecotype Columbia and have been characterized by Xiao et al. (2008a). All seeds used for determining the degree of dormancy and germination percentages on ABA or water were grown on the same shelf in the same growth room at the same time and harvested on the same day. Seeds were tested immediately (freshly harvested) or held at 23–25°C for 60 days (dry after-ripened) until testing. Arabidopsis seeds were surface-sterilized, and then grown on MS medium supplemented with 0.8% w/v agar and 2% w/v sucrose under 16 h light (23°C)/8 h dark (21°C) cycles. Plants were also grown in soil under similar environmental conditions.
TRIzol reagent (Invitrogen, www.invitrogen.com) was used for the extraction of RNA from 12-day-old Arabidopsis seedlings subjected to treatment with 100 μm ABA [(±)-ABA; Sigma-Aldrich, www.sigmaaldrich.com] or drought treatment (on dry Whatman No. 1 filter paper in a chamber under 45 ± 3% relative humidity) for various durations (0, 1, 3, 6 and 24 h) under continuous white light. First-strand cDNA was synthesized using the Superscript First-Strand Synthesis System (Invitrogen), and quantitative RT-PCR was performed using FastStart Universal SYBR Green Master Mix (Roche, www.roche.com) on a StepOne Plus Real-Time PCR System (Applied Biosystems, www.appliedbiosystems.com). Relative gene expression was normalized to the expression of ACTIN2. Changes in mRNA were analyzed based on three independent experiments as previously described (Schmittgen and Livak, 2008). The gene-specific primers used in quantitative RT-PCR are listed in Table S1.
Seeds were incubated on MS medium or MS medium supplemented with 1 μm ABA at 4°C for 3 days and then moved to a 22°C chamber for 12 h as described by Katagiri et al. (2005). Subsequently, total seed lipid was extracted according to the protocol provided by Kansas Lipidomics Research Center (http://www.k-state.edu/lipid/lipidomics/AT-seed-extraction). In brief, seeds were weighed and then incubated in 1 ml isopropanol with 0.01% butylated hydroxytoluene at 75°C for 15 min. The seeds were cooled on ice and thoroughly homogenized using a Dounce-type homogenizer (Wheaton, http://wheaton.com/). Samples were extracted using 2 ml chloroform/methanol (1:1 v/v) and three times using 2 ml chloroform/water (1:1 v/v). The combined extracts were washed (vigorous shaking at 4°C) with 0.5 ml of 1 M KCl, followed by 1 ml water, and then evaporated under nitrogen. The samples were dissolved in chloroform for analysis.
Total lipid of seedlings and rosettes was extracted as described by Welti et al. (2002). Twelve-day-old seedlings were treated with 100 μm ABA (for 1 or 3 h) or 0 μm ABA (control; for 1 h) as described by Peters et al. (2010). Seedlings and 5-week-old rosettes were harvested and transferred immediately to 3 ml isopropanol with 0.01% butylated hydroxytoluene before incubation at 75°C for 15 min. Subsequently, 1.5 ml chloroform and 0.6 ml water were added, followed by agitation for 1 h, after which the extract was removed. The tissue was re-extracted with chloroform/methanol/butylated hydroxytoluene (2:1:0.01 v/v/w) four or five times, until the tissue appeared white. The dry weight was obtained by incubation of the tissue overnight at 105°C. The combined extracts were subjected to washing with 1 ml of 1 M KCl, followed by 2 ml water. Finally, the combined solvents were evaporated under nitrogen, and the samples were dissolved in chloroform for analysis. Lipids were analyzed by electrospray ionization triple quadrupole mass spectrometry as described by Xiao et al. (2010).
Lipid pull-down assays
His-tagged ACBP1 and ACBP2 were expressed and purified as described previously (Chye, 1998; Chye et al., 2000). Protein–lipid pull-down assays were performed using PA and PC beads (Echelon Biosciences Inc, www.echelon-inc.com) according to the manufacturer's instructions. Briefly, 10 μg of purified proteins (diluted in 50 μl binding buffer, 10 mM HEPES, 150 mM NaCl, 0.25% Igepal, Sigma-Aldrich, http://www.sigmaaldrich.com pH 7.4) were incubated in 50 μl lipid beads for 3 h at 4°C with gentle shaking. Samples were centrifuged for 1 min at 80 g 4°C, and the supernatant was collected. Pellets were then washed three times with 1 ml binding buffer, and bound proteins were eluted with Laemmli sample buffer (4% SDS, 20% glycerol, 10% 2-mercaptoethanol, 0.02% bromphenol blue, 0.125 M Tris/HCl, pH 6.8) by heating at 95°C for 5 min. Proteins were subsequently analyzed by SDS–PAGE and Coomassie Brilliant Blue staining.
Generation of the ACBP1pro::GUS construct and GUS assays
The ACBP1pro::GUS fusion construct consists of a 1.8-kb 5′ region flanking ACBP1 that was amplified by primer pair ML801/ML802. The primer sequences are provided in Table S2. The 1.8-kb ACBP1pro fragment was cloned into pGEM-T Easy vector (Promega, www.promega.com) to yield plasmid pAT350, from which a 1.8-kb BamHI–SmaI fragment was subsequently cloned into similar sites on binary vector pBI101.3 (Clontech, www.clontech.com) to generate pAT352. DNA sequence analysis was used to verify the PCR fragment. The construct pAT352 was used in transformation of Arabidopsis (ecotype Columbia) by Agrobacterium tumefaciens and floral dip (Clough and Bent, 1998).
In histochemical GUS assays (Liu et al., 2004), samples were immersed and vacuum-infiltrated in a GUS staining solution (100 mM sodium phosphate buffer, 0.1% Triton X-100, 2 mM potassium ferricyanide, 2 mM potassium ferrocyanide, 1 mg ml−1 5-bromo-4-chloro-3-indolyl-β-d-glucuronide, pH 7.0) for 30 min, followed by incubation and observation over a period ranging from 3 to 16 h at 37°C. Samples were then cleared in 70% ethanol for photography. For fluorescent microscopy, 12-day-old seedlings were incubated for 1 h in a 0.1% Silwet L-77 (LEHLE SEEDS, http://www.arabidopsis.com/) solution supplemented with 50 mM ImaGene Green (Invitrogen). Following a brief wash in 0.1% Silwet L-77, samples were photographed using a confocal laser scanning microscope (Zeiss LSM 510 META, www.zeiss.com).
To examine seed dormancy in the wild type, acbp1 mutant and ACBP1-OX lines, over 100 freshly harvested seeds were placed on Whatman No. 1 filter papers moistened with sterile water in Petri dishes under continuous light at 21°C as described by Liu et al. (2009). The germination of these seeds was examined daily using a microscope (Leica MZ6, http://www.leica.com/). For germination assays, over 200 dry after-ripened seeds (per experiment) were sown on MS medium in the presence or absence of (±)-ABA (0.75, 1.5 or 2 μm; Sigma-Aldrich). Seeds were first chilled at 4°C for 2 days, and then germinated and grown under 16 h light (23°C)/8 h dark (21°C) cycles, and germination was recorded daily. To examine their development post-germination, seeds were germinated on MS medium for 4 days, after which seedlings were transferred to fresh MS medium or MS medium supplemented with 100 or 150 μm ABA for a further 7 days growth, followed by photography and root length measurements.
Experiments on BiFC interaction
BiFC assays were performed by transient transfection of N. benthamiana (tobacco) leaves (Walter et al., 2004; Vlad et al., 2009). The coding region of ACBP1 was amplified using primers ML1320 and ML1321 (Table S3), and cloned into pGEM-T Easy vector (Promega) to generate plasmid pAT610. This plasmid was then digested with XbaI and XhoI to yield a fragment containing the ACBP1 coding region, which was cloned into similar sites on vector pSPYNE-35S (Walter et al., 2004) to generate pSPYNE-35S-ACBP1. Similar strategies were used to amplify PLDα1 using primers ML1374 and ML1375 (Table S3). An XbaI–BamHI fragment containing the PLDα1 coding region was cloned into similar sites of vector pSPYCE-35S (Walter et al., 2004) to generate pSPYCE-35S-PLDα1. Agrobacterium tumefaciens LBA4404 derivatives containing fluorescence-tagged constructs were incubated with shaking at 28°C in Luria–Bertani (LB) broth for approximately 16 h. Subsequently, leaves of 4-week-old tobacco were used for infiltration as described previously (Yang et al., 2000). Epidermal cell layers of agro-infiltrated leaves were examined for fluorescence under a Zeiss LSM 510 META microscope 2 days after infiltration.
Yeast two-hybrid analysis
The yeast two-hybrid analysis was performed using the Matchmaker Gold yeast two-hybrid system (Clontech) according to the manufacturer's instructions. Yeast cells (Y2HGold strain) co-transformed with pGBKT7-53 and pGADT7-T were used as the positive control, and cells co-transformed with pGBKT7-Lam and pGADT7-T were used as the negative control. A 1.1-kb EcoRI–EcoRI fragment of ACBP1 cDNA encoding a truncated ACBP1 lacking its transmembrane domain (by deletion of amino acids 1-40) was digested from pAT61 (Chye, 1998) and cloned into the EcoRI site of the bait vector pGBKT7 to generate plasmid pAT248 (Δ40ACBP1). The full-length PLDα1 cDNA was amplified using primers ML1953 and ML1375 (Table S3). A ClaI–BamHI fragment containing the coding region of PLDα1 was cloned into similar sites of the prey vector pGADT7. Co-transformed yeast cells were first selected on SD medium/–Leu/–Trp double drop-out agar plates at 30°C for 3 days. Subsequently, one colony from yeast transformants containing each pair of constructs was diluted in 30 μl of 0.9% w/v NaCl, and a 1/10th dilution was spotted on SD/–His/–Leu/–Trp plates (TDO) and SD/–His/–Leu/–Trp plates containing 5-bromo-4-chloro-3-indolyl-alpha-D-galactopyranoside (X-α-Gal)/aureobasidin A (TDO/X/A). The plates were then incubated at 30°C for 3 days before photography.
The Arabidopsis Genome Initiative or GenBank accession numbers for the sequences referred to in this paper are shown in parentheses: ACBP1 (At5 g53470; NM_124726), ACTIN2 (At3G18780; NM_112764), ABI1 (At4 g26080; NM_118741), PLDα1 (At3 g15730; NM_112443), AtrbohD (At5G47910; NM_124165), AtrbohF (At1G64060; NM_105079), AREB1 (At1G45249; NM_179446), MYC2 (At1G32640; NM_102998), NCED3 (At3G14440; NM_112304), ABA2 (At1G52340; NM_104113), RD29A (At5G52310; NM_124610) and RPK1 (At1G69270; NM_105594).
We thank M. Roth and R. Welti (Kansas Lipidomics Research Center, KS) for lipid profiling, J. Kudla (University of Münster, Germany) for pSPYNE-35S and pSPYCE-35S, J.A. Tanner (University of Hong Kong) for ITC analysis, and C. Lo and Y-G. Du (University of Hong Kong) for determination of ABA content. This work was supported by the Hong Kong Research Grants Council (HKU765511M), University of Hong Kong studentships (Z-Y.D., M-X.C. and Q-F.C.) and a postdoctoral fellowship (S.X.) and by the Wilson and Amelia Wong Endowment Fund.