The Arabidopsis acbp1acbp2 double mutant lacking acyl-CoA-binding proteins ACBP1 and ACBP2 is embryo lethal


Author for correspondence:
Mee-Len Chye
Tel: +852 22990319


  • In Arabidopsis thaliana, the amino acid sequences of membrane-associated acyl-CoA-binding proteins ACBP1 and ACBP2 are highly conserved. We have shown previously that, in developing seeds, ACBP1 accumulates in the cotyledonary cells of embryos and ACBP1 is proposed to be involved in lipid transfer. We show here by immunolocalization, using ACBP2-specific antibodies, that ACBP2 is also expressed in the embryos at various stages of seed development in Arabidopsis.
  • Phenotypic analyses of acbp1 and acbp2 single mutants revealed that knockout of either ACBP1 or ACBP2 alone did not affect their life cycle as both single mutants exhibited normal growth and development similar to the wild-type. However, the acbp1acbp2 double mutant was embryo lethal and was also defective in callus induction.
  • On lipid and acyl-CoA analyses, the siliques, but not the leaves, of the acbp1 mutant accumulated galactolipid monogalactosyldiacylglycerol and 18:0-CoA, but the levels of most polyunsaturated species of phospholipid, such as phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol and phosphatidylserine, declined.
  • As recombinant ACBP1 and ACBP2 bind unsaturated phosphatidylcholine and acyl-CoA esters in vitro, we propose that ACBP1 and ACBP2 are essential in lipid transfer during early embryogenesis in Arabidopsis.


During embryogenesis, which is an early essential step in the life cycle of higher plants, cell differentiation and division occur and a multicellular organism is produced from one single cell (Yeung & Meinke, 1993; Devic, 2008). In Arabidopsis thaliana, seed development can be divided into two major phases: embryo morphogenesis and maturation. Embryo morphogenesis is initiated by the fusion of the male and female gametes to form a zygote, after which rapid cell division occurs and the embryo pattern is soon established. Subsequently, during embryo maturation, cell expansion and the accumulation of reserves follow. Finally, the seed undergoes desiccation and dormancy sets in (Harada, 1997; Fan et al., 2008; Baud & Lepiniec, 2009).

To date, a large number of Arabidopsis genes essential in seed development (EMBs) have been identified using forward- and reverse-genetic approaches (Tzafrir et al., 2004; Meinke et al., 2008). Analyses of embryo development in these mutants have indicated that the majority display aborted phenotypes at or before the globular stage, reinforcing this to be most critical in embryo development (Devic, 2008). Some EMB genes affect lipid transport and metabolism (nine genes), or CoA transport and metabolism (four genes), implicating the essential role of lipids in plant embryo development (Tzafrir et al., 2003; Studies have revealed that the depletion of LPAAT1 causes embryo lethality in Arabidopsis, with embryo development arrested at the heart–torpedo stage (Kim & Huang, 2004). LPAAT1 encodes a plastid-localized lysophosphatidic acid acyltransferase (LPAAT) for the acylation of the sn-2 position of lysophosphatidic acid in phosphatidic acid (PA) biosynthesis, and has been deemed to be essential for heart–torpedo stage establishment in embryo development in Arabidopsis (Kim & Huang, 2004). By contrast, the knockout mutant of Arabidopsis LPAT2, lacking an endoplasmic reticulum (ER)-located LPAAT, is defective in female, but not male, gametophyte development (Kim et al., 2005). Antisense downregulation of ACCase in Brassica napus produces wrinkled seeds, accompanied by a decline in lipid content (Sellwood et al., 2000). In Arabidopsis, two genes, ACC1 and ACC2, encode multifunctional isoforms of acetyl-CoA carboxylase (ACCase) (Yanai et al., 1995), and knockout mutants of ACC1 are embryo lethal (Baud et al., 2003). These results demonstrate that processes affecting plant lipid metabolism, especially those related to very-long-chain fatty acid elongation, are important during the early stages of seed development.

The acyl-CoA-binding proteins (ACBPs) are a family of proteins that facilitate the binding of long-chain acyl-CoA esters at a conserved acyl-CoA-binding domain (Xiao & Chye, 2009). In Arabidopsis, six genes encode ACBPs, which are localized subcellularly in different compartments (Chye et al., 1999; Li & Chye, 2003; Chen et al., 2008; Xiao et al., 2008b). As they bind different acyl-CoA esters with varying affinities, they do not seem to have redundant roles in vivo (Chye, 1998; Chye et al., 2000; Leung et al., 2004, 2006; Gao et al., 2009; Xiao et al., 2009). In addition to the presence of the conserved ACBP domain, some ACBPs contain other functional domains, such as ankyrin repeats in ACBP1 and ACBP2, and kelch motifs in ACBP4 and ACBP5, which mediate interactions with protein partners (Li & Chye, 2004; Li et al., 2008; Gao et al., 2009). Three cytosolic ACBPs (ACBP4, ACBP5 and ACBP6) are known to bind phosphatidylcholine (PC) in vitro (Chen et al., 2008; Xiao et al., 2009). Immunogold labelling analysis has indicated previously that ACBP1 is localized to the plasma membrane of epidermal cells and in cotyledonary cells at various stages (heart, torpedo and cotyledon) of embryo development in developing seeds (Chye et al., 1999). It has been proposed that ACBP1 may be involved in lipid transfer originating from the ER to the plasma membrane during lipid metabolism in these seeds (Chye et al., 1999). Consistently, ACBP1–green fluorescent protein (GFP) and ACBP2–GFP fusion proteins were localized at the plasma membrane and ER in onion epidermal cells by particle bombardment, and the results were confirmed using subcellular fractionation followed by western blot analysis (Li & Chye, 2003). The N-terminal transmembrane domains, in both ACBP1 and ACBP2, are essential for their membrane association (Li & Chye, 2003). Given the importance of proteins/enzymes related to lipid metabolism in Arabidopsis seed development (Baud et al., 2003; Rylott et al., 2003), the accumulation of ACBP1 in seeds and embryos suggests possible roles in embryo development or seedling establishment. Thus, it was pertinent to investigate whether ACBP1 and/or ACBP2 affect early development by the analysis of single mutants as well as the acbp1acbp2 double mutant.

Materials and Methods

Plant materials and growth conditions

The acbp2 T-DNA insertion mutant was identified from a T-DNA insertional library from the Torrey Mesa Research Institute of Syngenta ( After surface sterilization and chilling at 4°C for 2 d, seeds of Arabidopsis thaliana (L.) Heynh wild-type (WT) (ecotype Columbia), and acbp1 and acbp2 mutants, were germinated and grown on Murashige and Skoog medium (Murashige & Skoog, 1962) supplemented with 2% sucrose grown under cycles of 8 h dark (21°C) and 16 h light (23°C). Soil-grown plants were also grown under cycles of 8 h dark (21°C): 16 h light (23°C).

Immunohistochemical localization of ACBP2 using light microscopy

Immunohistochemical localization of ACBP2 using anti-ACBP2-specific antibodies (Chye et al., 2000; Li & Chye, 2003) was performed as described previously (Chye et al., 1999). Briefly, Arabidopsis siliques containing developing seeds at various stages of embryo development were fixed and embedded in paraffin following Chye et al. (1999). Blocking and antibody reactions were carried out in 1% BSA in phosphate-buffered saline. The sections were incubated for 5 min with phosphate-buffered saline and 0.1% saponin (Sigma) before incubation with phosphate-buffered saline containing 0.1% saponin, 1% BSA and 2% goat serum for 1 h at room temperature. Sections were incubated with rabbit anti-ACBP2-specific antibodies (1 : 1000, v/v) at 4°C overnight and, subsequently, with the secondary antibody, biotinylated alkaline-phosphatase-conjugated goat anti-rabbit antibodies (1 : 1000, v/v; Bio-Rad), at room temperature for 2 h. Levamisole (1 mM; Sigma) was included in the alkaline phosphatase reaction to inhibit endogenous phosphatase, following the instructions of the manufacturer (Bio-Rad).

Identification of the acbp2 mutant

The homozygous acbp2 mutant was isolated by PCR amplification using two primer pairs: first, ACBP2 gene-specific forward primer ML251 (5′-ATCGGCGTTGGTTTTTCGTTTTTGAGAAT-3′) with reverse primer ML252 (5′-TTGCCGCCAAAGTCGGTTATTTATTCGTT-3′); and second, ML205 (5′-CGTCACCCAGAGGAGTC-3′) with the T-DNA left border primer Oligo113 (O113; 5′-TAGCATCTGAATTTCATAACCAATCTCGATACAC-3′). The PCR products were separated by electrophoresis on 0.8% agarose and DNA was transferred to a nylon membrane (Hybond-N; Amersham). The blot was hybridized overnight at 42°C to a random-primed 32P-labelled full-length ACBP2 gene probe. The blot was washed in 0.1 × standard saline citrate, 0.1% sodium dodecylsulphate (SDS) at 65°C for 10 min. The position of the T-DNA insertion was confirmed by DNA sequence analysis of the resultant PCR products.

Western blot analysis

Total plant protein was extracted (Chye et al., 1999) from mature silique-bearing plants of WT Arabidopsis or the acbp2 mutant. The protein concentration was determined using the Bio-Rad Protein Assay Kit following the method of Bradford (1976). Ten micrograms of total protein were loaded per well in sodium dodecylsulphate–polyacrylamide gel electrophoresis (SDS-PAGE). The proteins were electrophoretically transferred to Hybond-C membrane (Amersham) from the SDS–PAGE gel using the Trans-Blot cell (Bio-Rad). Affinity column-purified ACBP2-specific antibodies (Chye et al., 2000; Li & Chye, 2003) were used in western blot analysis. The Amplified Alkaline Phosphatase Goat Anti-rabbit Immuno-blot Assay Kit (Bio-Rad) was used following the instructions of the manufacturer for the detection of cross-reacting bands.

Semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) analysis

Using TRIzol reagent (Invitrogen, Cat No. 15596-018), total RNA was isolated from rosettes and siliques of 7-wk-old WT, acbp1 and acbp2 plants. First-strand synthesis was carried out using the Superscript™ First-strand synthesis system (Invitrogen, Cat No. 12371-019). Gene-specific primers for RT-PCR were used as described previously (Xiao et al., 2008a) and the numbers of cycles used in amplification with each primer pair were within the linear range.

Screening of the acbp1acbp2 double mutant

The acbp1 (Xiao et al., 2008a) and acbp2 homozygous mutants were crossed and their resultant F2 population was screened for acbp1acbp2 double mutants. F2 seeds were sterilized and grown on kanamycin-containing Murashige and Skoog medium. From kanamycin-resistant (for acbp2 allele) plants, DNA was extracted and primer combinations ML179/ML209 and ML179/SLB1 (Xiao et al., 2008a), O113/ML206 (5′-TCGGGGTGGGGATGATGC-3′) and ML206/ML252 (Fig. 3a) were used in PCR to screen for the acbp1 and acbp2 alleles, respectively. As acbp1acbp2 double mutants were not obtained after > 200 F2 plants had been screened, acbp1ACBP2+/− (i.e. homozygous for acbp1 and heterozygous for acbp2) and ACBP1+/−acbp2 (i.e. heterozygous for acbp1 and homozygous for acbp2) plants were subsequently generated. The self-fertilized F3 seeds of acbp1ACBP2+/− or ACBP1+/−acbp2 plants were compared with WT by light microscopy, the percentages of aborted ovules in open siliques from WT and acbp1ACBP2+/− or ACBP1+/− acbp2 plants were calculated and their whole-mount embryos were observed. For complementation testing, transgenic line acbp1::35S-ACBP1 (cACBP1-2; Xiao et al., 2008a) was crossed to acbp1ACBP2+/− plants and the F1 progenies were used for further analysis.


Developing seeds or excised embryos were placed in Herr’s solution (Herr, 1971) composed of 85% lactic acid, chloral hydrate, phenol, clove oil and xylene (2 : 2 : 2 : 2 : 1) for 2 h to overnight. Slides were viewed using a Leitz photomicroscope employing differential interference contrast optics.

Callus induction

Callus induction was carried out according to Liu et al. (2004). Aborted embryos of 4-d after fertilization (4-DAF) acbp1ACBP2+/− plants were excised under the microscope, placed onto callus-inducing medium containing 2,4-dichlorophenoxyacetic acid (2,4-D; 0.5 mg l−1), indoleacetic acid (2 mg l−1) and N6-(2-isopentenyl)adenine (0.5 mg l−1), and cultured for 3 wk at 22°C in the dark. Embryos from WT, acbp1 and acbp2 plants at similar developmental stages (DSs) were excised and grown on callus-inducing medium under similar conditions.

Lipid and acyl-CoA profiling

Total lipid extraction was carried out according to Welti et al. (2002) and lipids were dissolved in chloroform for analysis. The profiles of membrane lipids were measured by an automated electrospray ionization tandem mass spectrometer (Devaiah et al., 2006).

Acyl-CoA profiling was carried out as described previously (Larson & Graham, 2001; Larson et al., 2002). Briefly, acyl-CoAs from frozen samples of 7-wk-old rosettes and siliques were extracted, dried under vacuum at 40°C and subsequently reacted with 50 μl of buffered chloroacetaldehyde reagent to form fluorescent acyl-etheno-CoA derivatives. Acyl-CoA standards purchased from Avanti (Avanti Polar Lipids, Inc., Alabaster, AL, USA) were similarly treated. The derived standards and acyl-etheno-CoA samples were separated and quantified by reversed-phased high-performance liquid chromatography on a LUNA phenyl-hexyl column (Phenomenex, Torrance, CA, USA; 150 × 2.0 mm, 5 μm) together with a 4 × 2 mm phenyl-propyl guard column. The solvent system was identical to the longer gradient conditions reported by Larson & Graham (2001).

Lipid binding assay

(His)6-ACBP1 and (His)6-ACBP2 were expressed in the soluble fraction and inclusion bodies, respectively, of Escherichia coli extracts and were each purified through an affinity column of Ni-NTA Agarose (Qiagen) as described (Chye, 1998; Chye et al., 2000). Binding of (His)6-ACBP1 and (His)6-ACBP2 to various lipids on filters was carried out according to Chen et al. (2008).

To further confirm the interactions of (His)6-ACBP1 and (His)6-ACBP2 with PC, the Lipidex 1000 competition assay was carried out to determine whether PC was capable of competing with [14C]linoleoyl-CoA (American Radiolabelled Chemicals, St. Louis, MO, USA, in binding (His)6-ACBP1 and (His)6-ACBP2 (Rosendal et al., 1993; Gao et al., 2009). In the Lipidex 1000 binding assay, the incubation medium contains both ACBP1/ACBP2 and radiolabelled acyl-CoA. Subsequently, unbound radiolabelled acyl-CoA was removed from the incubation medium by Lipidex 1000. The bound radiolabelled acyl-CoA remains in the supernatant and was detected by the measurement of radioactivity counts. In Lipidex 1000 competition assays, PC liposome was further added to the incubation medium containing both ACBP1/ACBP2 and radiolabelled acyl-CoA. If PC liposome successfully competes with acyl-CoA in binding to ACBP1/ACBP2, bound radiolabelled acyl-CoA will decrease and a decline in radioactivity count will result.

PC liposome was prepared according to Sano et al. (1998). Different concentrations of PC liposome (0–5 μM) were mixed with 0.8 μM [14C]linoleoyl-CoA and 0.2 μM (His)6-ACBP1 or (His)6-ACBP2. Each mixture was incubated for 30 min at 37°C, and 400 μl of ice-cold 50% slurry of Lipidex 1000 (PerkinElmer, Waltham, MA, USA, and binding buffer were added. Samples were centrifuged at 12 000 g for 5 min at 4°C, and a 200-μl aliquot of the supernatant was taken for the analysis of radioactivity counts using an LS 6500 liquid scintillation counter (Beckman, Brea, CA, USA). Assays were performed in triplicate, with blanks, at each concentration of PC liposome.


Expression of ACBP1 and ACBP2 mRNAs during Arabidopsis development

Our previous studies have indicated that ACBP1 (Chye, 1998) and ACBP2 (Chye et al., 2000; Gao et al., 2009) mRNAs are expressed in all plant organs, whereas ACBP1 protein accumulates predominantly in developing seeds (Chye, 1998), and ACBP2 in flowers and siliques (Kojima et al., 2007). To further understand the dynamic expression patterns of ACBP1 and ACBP2 during Arabidopsis development, data on their expression at various stages in development were retrieved from the microarray database Gene Chronologer (Zimmermann et al., 2004; Expression of both ACBP1 and ACBP2 was relatively high in early DS1 (refers to seedlings aged 1–5.9-d post-germination). Expression decreased from stages DS2 to DS8, spanning day 6 to day 44.9 (Fig. 1a). However, at DS9 (representing the seed set DS), both genes again showed high expression (Fig. 1a). Further investigation of ACBP1 and ACBP2 expression during seed development was performed by analyses of their expression at different seed stages (SSs) using the e-FP Browser (Winter et al., 2007; database. Results suggested that the expression of ACBP1 and ACBP2 increased from SS5 (walking-stick stage of development) to SS7 (late cotyledonary stage) or SS8 (green cotyledonary stage) and was relatively high in dry seeds (SS9) (Fig. 1b).

Figure 1.

 Microarray data showing the expression patterns of ACBP1 (open bars) and ACBP2 (closed bars) mRNAs. (a) Expression of ACBP1 and ACBP2 at various developmental stages (DSs). DS1, Arabidopsis seedlings of 1–5.9 d; DS2, seedlings of 6–13.9 d; DS3, seedlings of 14–17.9 d; DS4, plants of 18–20.9 d; DS5, plants of 21–24.9 d; DS6, plants of 25–28.9 d; DS7, plants of 29–35.9 d; DS8, plants of 36–44.9 d; DS9, seeds of 45–50 d. (b) Expression of ACBP1 and ACBP2 at various seed stages (SSs). SS1, globular; SS2, early heart; SS3, late heart; SS4, torpedo; SS5, walking-stick; SS6, early curled cotyledons; SS7, late cotyledons; SS8, green cotyledons; SS9, dry seeds.

Immunolocalization of ACBP2 in Arabidopsis embryos using light microscopy

As immunolocalization using anti-ACBP1 polyclonal antibodies has revealed that ACBP1 protein accumulates in the cotyledonary cells of embryos in developing seeds (Chye et al., 1999), we were interested in investigating whether ACBP2 accumulates in seeds as well. Longitudinal sections of siliques at various stages of development were immunolocalized using anti-ACBP2-specific antibodies (Chye et al., 2000; Li & Chye, 2003) followed by light microscopy. The results revealed that embryos at the globular (Fig. 2a), heart (Fig. 2b), heart–torpedo transition (Fig. 2c) and mature cotyledonary (Fig. 2d) stages were immunostained and expressed ACBP2. By contrast, similar stage embryos using preimmune serum from the same rabbit as a negative control did not show any staining (Fig. 2e–h).

Figure 2.

 Immunolocalization of the acyl-CoA-binding protein ACBP2 using anti-ACBP2 antibodies in developing Arabidopsis seeds. (a–d) Longitudinal sections of seeds at globular (a, indicated by arrow), heart (b), heart–torpedo transition (c) and cotyledon (d) stages showing embryos stained with anti-ACBP2 primary antibody followed by biotinylated goat anti-rabbit antibody and reacted with streptavidin-biotinylated alkaline phosphatase. (e–h) Longitudinal sections of seeds at globular (e, indicated by arrow), heart (f), heart–torpedo transition (g) and cotyledon (h) stages showing embryos stained with preimmune serum followed by biotinylated goat anti-rabbit antibody and reacted with streptavidin-biotinylated alkaline phosphatase. Bars, 50 μm.

Characterization of an acbp2 knockout mutant

In this study, an acbp2 T-DNA insertional mutant was isolated from CS19943 (a T-DNA mutagenized Arabidopsis seed pool generated by Thomas Jack; ecotype Col-6 (Thomas Jack laboratory, Dartmouth College, New Hampshire, USA)) using a combination of gene-specific and T-DNA border-specific primers (Winkler et al., 1998). The location of the T-DNA insertion in ACBP2 was confirmed by PCR using ACBP2 gene-specific primers ML251 and ML252, as well as gene-specific forward primer ML205 and reverse primer Oligo113 (O113) that maps to the T-DNA left border (Fig. 3a,b). By further DNA sequencing analysis, the insertion site was mapped to intron 5 on ACBP2 (Fig. 3a). Western blot analysis using ACBP2-specific antibodies (Fig. 3c) and RT-PCR analysis using gene-specific primers ML206/ML192 (Fig. 3d) confirmed the mutant to be a functional knockout line. As we have characterized previously a T-DNA insertional knockout mutant of ACBP1 (Xiao et al., 2008a), we were able to compare the expression of ACBP1 in the acbp2 mutant and that of ACBP2 in the acbp1 mutant. Semi-quantitative RT-PCR results showed that, in rosettes, the expression of ACBP1 in the acbp2 mutant and the expression of ACBP2 in the acbp1 mutant were upregulated in comparison with WT (Fig. 3d). Consistent with microarray data (Fig. 1), the expression levels of ACBP1 and ACBP2 in WT siliques were significantly higher than those in rosettes, whereas the upregulation of ACBP1 in the acbp2 mutant, and vice versa, was not evident (Fig. 3d).

Figure 3.

 Characterization of acbp1 and acbp2 knockout mutants. (a) T-DNA insertion in ACBP2, and the primer locations used for PCR genotyping of the acbp2 allele. (b) Southern blot analysis of PCR products amplified from acbp2 mutant DNA (lanes 1 and 2) or wild-type (WT) Arabidopsis DNA (lanes 3 and 4) using primer pairs ML205 and O113 (lanes 1 and 3) or ML251 and ML252 (lanes 2 and 4). The gel was blotted onto Hybond-N and hybridized to 32P-labelled ACBP2 cDNA. (c) Western blot analysis of total protein from the silique-bearing acbp2 mutant (lane 1) or WT Arabidopsis (lane 2) using ACBP2-specific antibodies. Ten micrograms of total proteins were loaded onto each lane on a 10% sodium dodecylsulphate–polyacrylamide gel. Asterisk indicates the position of a nonspecific immunoreacting band. (d) Semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) analysis of the expression of ACBP1 and ACBP2 in WT, acbp1 and acbp2 mutants. Total RNA was extracted from rosettes and siliques of 7-wk-old WT, acbp1 and acbp2 plants. The 18S transcript was used as a control.

The growth and development of the acbp1 and acbp2 mutants were compared. At early (2-wk-old seedling) and late (5-wk-old plants) development, both mutants were similar to WT under normal growth conditions (Supporting Information Fig. S1a,b). To gain a better insight into the functions of ACBP1 and ACBP2 in seed development, embryos from the acbp1 and acbp2 mutants were carefully analysed. Siliques of the acbp1 and acbp2 mutants produced normal seeds, similar to WT (Fig. S1c). Whole-mount examination of embryo development in WT, acbp1 and acbp2 siliques (Fig. S1d) indicated that acbp1 and acbp2 embryos reached the heart stage at 4 DAF and the bent cotyledonary stage at 6 DAF. These results suggest that the acbp1 and acbp2 single mutants displayed no obvious morphological changes in embryonic and seed development and were indistinguishable from WT.

The acbp1acbp2 double mutant is embryo lethal

As ACBP1 and ACBP2 are highly conserved (82% identity) and are both expressed in embryos, we next investigated whether they function redundantly in lipid metabolism during seed development. To test this possibility, a combination of acbp1 and acbp2 mutations would be necessary. Hence, we crossed the two single acbp1 and acbp2 mutants, generated their F1 progeny, and subsequently screened their F2 populations for acbp1acbp2 double mutants. However, after genotyping > 200 F2 plants, acbp1acbp2 double mutants were not encountered. We were only able to identify genotypes of either acbp1ACBP2+/− (i.e. homozygous for acbp1 and heterozygous for acbp2; lane 1 in Fig. 4a) or ACBP1+/−acbp2 (i.e. heterozygous for acbp1 and homozygous for acbp2; lanes 6, 9 and 11 in Fig. 4a).

Figure 4.

 Genotyping and initial characterization of the Arabidopsis acbp1acbp2 double mutant. (a, b) Specificity of the primer combinations (right) used in PCR to screen the acbp1acbp2 double mutant from the kanamycin-resistant (for acbp2 allele) F2 population. ML179 and ML209 are ACBP1 gene-specific primers and SLB1 is the T-DNA-specific primer used for genotyping the acbp1 allele (Xiao et al., 2008a). No acbp1acbp2 double mutants were obtained from over 200 F2 plants screened, and so acbp1ACBP2+/− (i.e. homozygous for acbp1 and heterozygous for acbp2; lane 1 in (a)) and ACBP1+/− acbp2 (i.e. heterozygous for acbp1 and homozygous for acbp2; lanes 6, 9 and 11 in (b)) were subsequently generated. (c) Comparison of self-fertilized F3 seeds of acbp1ACBP2+/− and ACBP1+/−acbp2 plants with wild-type (WT) by light microscopy (top). Ovules in open siliques from acbp1ACBP2+/− plants showing aborted ovules (indicated by arrows). Bars, 1 mm.

The morphology of seeds from the self-pollinated acbp1ACBP2+/− and ACBP1+/−acbp2 plants was further examined. As shown in Fig. 4(c), in contrast with WT siliques, which contained almost all normal seeds that were large and green, the siliques of acbp1ACBP2+/− and ACBP1+/−acbp2 plants contained both normal seeds which were large and green, as well as a proportion of aborted seeds (Fig. 4c). The percentage of aborted seeds in acbp1ACBP2+/− and ACBP1+/−acbp2 plants of c. 35% (Table 1) correlated well with a 3 : 1 segregation ratio. However, when the ACBP1 full-length cDNA was introduced into an acbp1ACBP2+/− background by crossing acbp1ACBP2+/− with cACBP1-2 (previously verified in Xiao et al. (2008a) to be a complemented line for the acbp1 mutant), only 5.7% of seeds were aborted (Table 1). This indicates that acbp1acbp2 embryos could be rescued by the introduction of ACBP1 cDNA. In addition, by taking advantage of the kanamycin-resistant phenotype conferred by the T-DNA in the acbp2 mutant, the ratio of resistant to sensitive progenies from both acbp1ACBP2+/− and ACBP1+/−acbp2 was c. 2 : 1 when grown on Murashige and Skoog medium containing kanamycin (Table 2). These results confirm that a combination of acbp1 and acbp2 mutations affects embryo development.

Table 1.   Segregation and complementation analyses of Arabidopsis acbp1ACBP2+/− plants
 Normal seedsAborted seedsχ2 (hypothesis)
  1. Seeds were analysed from 10 siliques of single plants in each genotype. The genotypes of each plant were confirmed by PCR before analysis. Chi-squared values were calculated for the hypothesized segregation ratio and showed no significant deviation (> 0.05) from the hypothesized ratio.

ACBP1+/−acbp22911020.25 (3:1)
acbp1ACBP2+/−252900.39 (3:1)
acbp1ACBP2+/−×cACBP1368210.35 (15:1)
Table 2.   Ratio of kanamycin resistant to sensitive (Kanr : Kans) for the progeny from self-pollinated Arabidopsis ACBP1+/−acbp2 and acbp1ACBP2+/− plants
 KanrKansχ2 (hypothesis)
  1. Chi-squared values are given for the hypothesized segregation ratio and showed no significant deviation (> 0.05) from the hypothesized ratio.

ACBP1+/−acbp23161660.32 (2:1)
acbp1ACBP2+/−3781800.24 (2:1)

To further study the cause of embryo lethality in the acbp1acbp2 double mutant, embryo development at various stages in acbp1ACBP2+/− and WT plants was investigated. In the siliques of WT, most embryos reached the globular (Fig. 5g) or heart (Fig. 5h,i) stage by 3 DAF. However, significant variation in embryo development was observed in seeds derived from a single silique of the acbp1ACBP2+/− line. As shown in Fig. 5, in addition to ‘normal’ seeds in acbp1ACBP2+/− siliques which develop to globular (Fig. 5d) and heart (Fig. 5e,f) stages, there were also aborted ovules arrested at zygote (Fig. 5a), two-cell (Fig. 5d) or eight-cell (Fig. 5c) stage in the acbp1ACBP2+/− siliques. Further analysis revealed that 161 of 225 (72%) embryos in acbp1ACBP2+/− siliques reached globular and heart stages at 3 DAF, whereas the remaining 64 embryos (28%) were aborted early at either zygote, two-cell or eight-cell stage (Fig. 5j). The percentage of aborted embryos in the siliques of acbp1ACBP2+/− plants would be in agreement with an expected lethality in the acbp1acbp2 double mutation. In summary, these results demonstrate that ACBP1 and ACBP2 are essential for embryo and seed development, and embryos of the acbp1acbp2 double mutant were aborted at early embryo development.

Figure 5.

 Combination effect of acbp1 and acbp2 mutations on Arabidopsis embryo development. Embryo development of 3-DAF (days after fertilization) ovules of acbp1ACBP2+/− (a–f) and wild-type (WT,g–i) plants. At this stage, most of the WT embryos have reached globular (g) or heart stage (h, i); in the acbp1ACBP2+/− siliques, aborted ovules were arrested at the zygote (a), two-cell (b) or eight-cell (c) stage. Normal seeds in the same acbp1ACBP2+/− siliques develop to globular (d) and heart (e, f) stages. Arrows in (a–c) indicate arrested embryos. A, axis; BCD, basal cell descendant; C, cotyledon; LCD, lens cell descendant; S, suspensor. Bars, 25 μm. (j) Percentage of different embryo stages (early stage (including zygote, four-cell or eight-cell), closed bars; globular stage, open bars; heart stage, grey bars) in acbp1ACBP2± (28%, 50% and 22%, respectively) and WT (3%, 73% and 24%, respectively) plants. Asterisk represents 3% of early stage embryos in WT.

Aborted embryos from acbp1ACBP2+/− are defective in callus induction

To further investigate the nature of the embryo aberration in acbp1ACBP2+/− plants, callus induction was performed. To this end, the aborted embryos from 4-DAF acbp1ACBP2+/− plants were excised and subsequently cultured on a callus induction medium. As controls, embryos from WT, acbp1 and acbp2 at a similar DS were also excised and cultured. As shown in Fig. S2, WT, acbp1 and acbp2 embryos formed callus tissue after a 3-wk induction period, whereas the aborted embryos, which should be acbp1acbp2 double mutants, failed to form calli.

Lipid and acyl-CoA profile changes in the acbp1 and acbp2 mutants

To investigate whether the altered expression of ACBP1 or ACBP2 in the mutants affects plant lipid metabolism, the lipid composition in both rosettes and siliques of 7-wk-old WT (Col), acbp1 and acbp2 mutants was analysed using electrospray ionization tandem mass spectrometry (ESI MS/MS; Welti et al., 2002; Devaiah et al., 2006). From lipid profiling data, we observed that WT plants showed different patterns of lipid profiles in rosettes and siliques (Table 3). Specifically, the total contents of the galactolipids digalactosyldiacylglycerol (DGDG) and monogalactosyldiacylglycerol (MGDG) were significantly lower in siliques, whereas those of the phospholipids, such as PC, phosphatidylethanolamine (PE) and phosphatidylinositol (PI), as well as lysoPC and lysoPE, were significantly higher in siliques than in rosettes (Table 3). There were few differences between the rosettes of WT and the single acbp1 and acbp2 knockout mutants with regard to the total amounts of all lipid species analysed, including DGDG, MGDG, phosphatidylglycerol (PG), PC, PE, PI, phosphatidylserine (PS), PA, lysoPC, lysoPE and lysoPG (Table 3). However, in comparison with WT siliques, significantly higher levels of MGDG (< 0.01) and significantly lower amounts of the phospholipids PC, PI and PS (< 0.05 or < 0.01) were observed in the acbp1 mutant, but not in the acbp2 mutant (Table 3). Further, analyses of the lipid profiles of WT and acbp1 mutant siliques revealed that galactolipid species, such as 34:6- and 34:3-MGDG/DGDG and 36:6-MGDG, were increased significantly in the acbp1 mutant, whereas 36:4-, 38:5- and 38:4-MGDG/DGDG, 38:6-MGDG and 36:5-DGDG were decreased significantly in comparison with WT (Fig. 6). By contrast, most of the polyunsaturated species of the phospholipids, such as 34:2-, 36:4-, 36:3-, 38:5-, 38:4-, 38:3-, 40:5-, 40:4- and 40:3-PC, 36:4-, 38:5-, 38:4-, 38:3- and 40:3-PE, and 34:2-, 36:5-, 36:4-, 36:3- and 36:2-PI, were significantly lower than in WT, except for the species of 34:3- and 36:6-PC, which were higher in the acbp1 mutant siliques (Fig. 6).

Table 3.   Lipid profiles in rosettes and siliques of 7-wk-old wild-type (WT) (Col), acbp1 and acbp2 Arabidopsis mutants grown at 16 h light (23°C) : 8 h dark (21°C)
Lipid classRosettesSiliques
  1. DGDG, digalactosyldiacylglycerol; MGDG, monogalactosyldiacylglycerol; PG, phosphatidylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; PA, phosphatidic acid.

  2. Values are the means ± SD (% of total polar glycerolipids analysed; = 3). Significant differences between siliques and rosettes of WT and between acbp1 mutant and WT siliques are shown in bold (*, < 0.05; **, < 0.01).

  3. aValue increased in WT siliques in comparison with WT rosettes in the same experiment.

  4. bValue decreased in WT siliques in comparison with WT rosettes in the same experiment.

  5. cValue higher in siliques of the acbp1 mutant in comparison with siliques of WT in the same experiment.

  6. dValue lower in siliques of the acbp1 mutant in comparison with siliques of WT in the same experiment.

DGDG13.52 ± 0.6513.84 ± 0.4714.43 ± 0.5311.46 ± 0.19**b11.33 ± 0.3812.01 ± 0.14
MGDG66.84 ± 1.4068.29 ± 0.1466.71 ± 0.7147.57 ± 0.49**b52.91 ± 1.39**c48.46 ± 1.02
PG5.55 ± 0.065.59 ± 0.485.55 ± 0.666.23 ± 0.476.55 ± 0.445.47 ± 0.25
PC7.68 ± 0.487.73 ± 0.057.79 ± 0.3722.16 ± 0.71**a19.35 ± 1.47*d20.83 ± 0.83
PE1.50 ± 0.241.30 ± 0.111.42 ± 0.114.18 ± 0.25**a3.77 ± 0.354.39 ± 0.09
PI2.78 ± 0.182.77 ± 0.063.48 ± 0.287.25 ± 0.58**a5.15 ± 0.19**d7.70 ± 0.03
PS1.82 ± 2.830.18 ± 0.010.18 ± 0.040.60 ± 0.040.41 ± 0.04**d0.62 ± 0.04
PA0.25 ± 0.050.24 ± 0.040.36 ± 0.320.36 ± 0.100.35 ± 0.060.28 ± 0.06
LysoPC0.014 ± 0.0020.013 ± 0.0010.019 ± 0.0040.060 ± 0.007**a0.061 ± 0.0130.078 ± 0.011
LysoPE0.046 ± 0.0110.041 ± 0.0080.059 ± 0.0190.124 ± 0.012**a0.108 ± 0.0170.142 ± 0.002
LysoPG0.005 ± 0.0010.008 ± 0.0060.005 ± 0.0050.005 ± 0.0040.014 ± 0.0110.017 ± 0.026
Figure 6.

 Membrane lipid content (% of total glycerolipids analysed) in siliques of 7-wk-old wild-type (WT) Arabidopsis and acbp1 mutant. H, value of acbp1 mutant higher than WT; L, value of mutant lower than WT (*, P < 0.05; **, < 0.01). DGDG, digalactosyldiacylglycerol; MGDG, monogalactosyldiacylglycerol; PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidyl-ethanolamine; PG, phosphatidylglycerol; PI, phosphatidylinositol; PS, phosphatidylserine. Values are the means ± SD (= 3). Open bars, WT; closed bars, acbp1.

Analyses of the acyl-CoA content of rosettes from both acbp1 and acbp2 single knockout mutants did not reveal any differences in acyl-CoA profile in comparison with WT (Fig. 7a). In siliques, only 18:0-CoA in the acbp1 mutant was higher than in WT (Fig. 7b). The acyl-CoA profiles of rosettes and siliques from the acbp1ACBP2+/− plants resembled those of the acbp1 mutant (Fig. 7).

Figure 7.

 Acyl-CoA content in rosettes (a) and siliques (b) of 7-wk-old wild-type (WT) Arabidopsis and acbp1, acbp2 and acbp1ACBP2+/− mutants. H, value of acbp1 mutant and acbp1ACBP2+/− plants higher than WT (*, P < 0.05). Values are the means ± SD (= 3). Open bars, WT; closed bars, acbp1; dark grey bars, acbp2; light grey bars, acbp1ACBP2+/−.

His-tagged ACBP1 and ACBP2 recombinant proteins bind to unsaturated PC and acyl-CoA esters in vitro

Previously, we have used in vitro filter-binding assays to show that recombinant ACBP4, ACBP5 and ACBP6 bind PC, and not PE, PI, PA and lysoPC (Chen et al., 2008; Xiao et al., 2009). To explore whether recombinant ACBP1 and ACBP2 bind to phospholipids, purified ACBP1 and ACBP2 recombinant proteins (Chye, 1998; Chye et al., 2000; Gao et al., 2009) were similarly tested with various concentrations of PC, PE, PI and PS, as well as MGDG. These phospholipids were chosen because their contents had been altered in the siliques of the acbp1 mutant in comparison with WT (Table 3, Fig. 6). The results indicated that (His)6-ACBP1 and (His)6-ACBP2 bind PC, but not MGDG, in a dose-dependent manner (Fig. 8a). They also do not bind PE, PI and PS (data not shown). As the PC used in Fig. 8(a) was 1,2-diacyl-sn-glycero-3-phosphatidylcholine, consisting of 33% 16:0, 13% 18:0, 31% 18:1 and 15% 18:2 fatty acids, the binding of several fatty acid species of PC to (His)6-ACBP1 and (His)6-ACBP2 was tested subsequently. The results showed that both (His)6-ACBP1 and (His)6-ACBP2 bind the unsaturated species of PC tested (18:1-PC and 18:2-PC) but, unlike (His)6-ACBP4 (Xiao et al., 2009), (His)6-ACBP5 (Xiao et al., 2009) and (His)6-ACBP6 (Chen et al., 2008), they do not bind saturated species of PC (16:0-PC and 18:0-PC) (Fig. 8b).

Figure 8.

 Interaction of His-tagged ACBP1 and ACBP2 recombinant proteins with phosphatidylcholine (PC). (a) (His)6-ACBP1 and (His)6-ACBP2 bind to PC on filters. Various concentrations (0, 6, 12.5, 25.0 and 32.5 μM) of PC and monogalactosyldiacylglycerol (MGDG) were spotted onto nitrocellulose and incubated with 1 μg ml−1 of purified (His)6-ACBP1 or (His)6-ACBP2 proteins. (His)6-ACBP–PC binding was detected by immunoblotting with horseradish peroxidase-conjugated anti-penta-His antibodies. (b) Effect of PC acyl species on (His)6-ACBP–PC binding. Fifty micromoles of lipid (PC, 16:0-PC, 18:0-PC, 18:1-PC or 18:2-PC) spotted onto nitrocellulose were incubated with 1 μg ml−1 of purified (His)6-ACBP1 or (His)6-ACBP2 protein. (His)6-ACBP–lipid binding was detected by immunoblotting with horseradish peroxidase-conjugated anti-penta-His antibodies. PC, 1,2-diacyl-sn-glycero-3-phosphocholine; 16:0-PC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine; 18:0-PC, 1,2-distearoyl-sn-glycero-3-phosphocholine; 18:1-PC, 1,2-dioleoyl-sn-glycero-3-phosphocholine; 18:2-PC, 1,2-dilinoleoyl-sn-glycero-3-phosphocholine. (c) Displacement of [14C]linoleoyl-CoA by PC liposome in Lipidex competition assays. Different concentrations of PC liposome (0, 0.4, 0.8, 1.5, 3.0, 5.0 μM) were incubated with 0.8 μM [14C]linoleoyl-CoA and 0.2 μM (His)6-ACBP1 or (His)6-ACBP2. The mixtures were mixed with Lipidex-1000. Aliquots (200 μl) of the supernatant were taken for the analysis of radioactivity. Assays were performed in triplicate, with blanks, at each concentration of PC liposome. The bound acyl-CoAs in the presence of PC liposome (0–5 μM) were expressed relative to the value obtained from a reaction containing 0 μM PC liposome (100%). Squares, [ACBP1]; diamonds, [ACBP2]. Bars represent SD (= 3).

Lipidex assays have shown previously that (His)6-ACBP1 and (His)6-ACBP2 bind [14C]18:2-CoA and [14C]18:3-CoA (Gao et al., 2009). As they also bind to 18:2-PC, Lipidex competition assays were used to determine whether 18:2-PC liposome can compete with [14C]18:2-CoA in binding (His)6-ACBP1 and (His)6-ACBP2. The binding of [14C]18:2-CoA to both (His)6-ACBP1 and (His)6-ACBP2 (Fig. 8c) was observed to decrease in the presence of 18:2-PC liposome, implying that binding was displaced by the addition of 18:2-PC liposome.


Given that both recombinant ACBP1 and ACBP2 proteins have been shown recently to bind in vitro to linoleoyl-CoA and linolenoyl-CoA esters, the precursors in phospholipid membrane repair, we have suggested that they may share a role in the repair of the phospholipid bilayer membrane following heavy metal stress (Xiao et al., 2008a; Gao et al., 2009). We have further demonstrated that Arabidopsis ACBP1 (Chye et al., 1999) and ACBP2 (this study in Figs 1, 2) accumulate in the embryos of developing seeds, and probably play essential roles in embryogenesis, possibly in membrane biogenesis. Northern blot and western blot analyses showed that ACBP1 and ACBP2 mRNAs and proteins are not seed specific (Chye, 1998; Chye et al., 2000; Gao et al., 2009), and this supports the dual functions of these two ACBPs in both plant development and the plant stress response.

Many enzymes/proteins involved in plant lipid metabolism, such as acyl-CoA oxidases (ACXs), CoA biosynthetic enzymes (HAL3s), LPAAT and ACCase, are essential for either early embryogenesis or seedling development in Arabidopsis (Baud et al., 2003; Rylott et al., 2003; Kim & Huang, 2004; Kim et al., 2005; Rubio et al., 2006). More interestingly, impairment of CoA biosynthesis in the Arabidopsis hal3a-1halb double mutant causes embryo lethality at the early/mid-globular stage which precedes triacylglycerol (TAG) deposition, suggesting that CoA is crucial in early embryo development (Rubio et al., 2006). In addition, complete breakdown of short-chain acyl-CoA oxidase activity in the Arabidopsis acx3-1acx4-1 double mutant arrests embryo development at a very early stage (Rylott et al., 2003). Possible explanations for the embryo defects observed in the acx3-1acx4-1 double mutant include the accumulation of acyl-CoA or short-chain fatty acids to toxic levels, and the failure of CoA transfer to the acyl-CoA pool or to the production of fatty acid- or lipid-based signalling molecules (Rylott et al., 2003). The redundant functions of ACBP1 and ACBP2 in seed development share the characteristics of these previously reported proteins because embryo lethality in the acbp1acbp2 double mutant resembles that in hal3a-1halb and acx3-1acx4-1. Observations of early embryo arrest and lack of callus induction indicate that the development of the acbp1acbp2 embryo is arrested during embryo morphogenesis, which precedes fatty acid accumulation and lipid storage.

During early seed development, one of the major plant membrane lipids, PC, serves as a main substrate for desaturation in the ER and plays a central role in the assembly of unsaturated TAG (Browse, 1997). An acyl-CoA:lysophosphatidylcholine acyltransferase catalyses the exchange of polyunsaturated fatty acyl chains on sn-2 of PC with linoleoyl-CoA (18:2-CoA) and linolenoyl-CoA (18:3-CoA) in the acyl-CoA pool. Subsequently, glycerol-3-phosphate acyltransferase (GPAT) and lysophophatidic acid acyltransferase (LPAAT) are involved in the formation of polyunsaturated diacylglycerol (DAG) and acyl-CoA:sn-1,2-diacylglycerol acyltransferase in the acylation of the sn-3 position of DAG to produce TAG (Browse, 1997). Previous evidence has suggested that the acyl-CoA preference of mitochondrial GPAT in young rat liver is regulated by the presence of an ACBP bound to acyl-CoA (Kannan et al., 2003). The 10-kDa Brassica napus ACBP, which was the first ACBP identified from plants, is highly expressed in embryos, and has also been demonstrated to activate GPAT activity in vitro (Hills et al., 1994; Brown et al., 1998). A recent study has revealed that the overexpression of BnACBP in Arabidopsis seeds increases polyunsaturated (18:2 and 18:3) and decreases saturated and monounsaturated (16:0, 18:0 and 20:1) fatty acids (Yurchenko et al., 2009). Furthermore, recombinant BnACBP is important for the activity of lysophosphatidylcholine acyltransferase in the exchange of the acyl group between the acyl-CoA pool and PC (Yurchenko et al., 2009). Given that only one 10-kDa ACBP has been isolated from Brassica so far (Hills et al., 1994; Yurchenko et al., 2009), whereas a family of six ACBPs exists in Arabidopsis (Xiao & Chye, 2009), it is possible that the larger ACBP1 and ACBP2, which are membrane associated and accumulate in embryos, would play a more significant role in this event in Arabidopsis, rather than the 10-kDa ACBP6.

Furthermore, comparison of the lipid profiles in rosettes and siliques of WT Arabidopsis has shown that extra-plastidial phospholipids (PC, PE, PI and PS) accumulate significantly in siliques (Table 3), consistent with a previous report (Devaiah et al., 2006), and confirm the crucial role of these phospholipids in embryo development. Our biochemical data have also revealed that siliques (but not rosettes) of the acbp1 mutant show higher levels of galactolipid MGDG and lower levels of polyunsaturated PC, PE, PI and PS species in comparison with WT, demonstrating that the knockout of one of these two genes affects the phospholipid composition in siliques, although these changes per se may not effectively hinder embryo development. A lack of significant changes in phospholipids in siliques of the acbp2 mutant indicates that ACBP1 may play a more prominent role in this event than its redundant homologue ACBP2. Consistently, acyl-CoA profiling revealed that the level of 18:0-CoA was higher in siliques of the acbp1 mutant (but not in the acbp2 mutant), suggesting that, similar to BnACBP (Yurchenko et al., 2009), ACBP1 may play a role in the exchange of acyl-CoAs, especially 18:0-CoA, affecting the phospholipid content in siliques. This exchange was blocked to a certain extent in the acbp1 mutant and an accumulation of acyl-CoA in acbp1 siliques corresponded to a decline in phospholipids. In addition, the upregulation of ACBP1 in the acbp2 mutant, and vice versa, occurred only in rosettes (but not siliques), indicating that these two ACBPs may share both redundant and distinct cellular functions during plant development. Moreover, in vitro evidence has shown that recombinant ACBP1 and ACBP2 bind unsaturated PC (18:1-PC and 18:2-PC) and acyl-CoAs (18:2-CoA and 18:3-CoA), further supporting their roles in phospholipid metabolism. These results also reinforce ACBP1 and ACBP2 function in phospholipid membrane biogenesis (Chye, 1998; Xiao et al., 2008a; Gao et al., 2009), which would be crucial in early embryo development. Taken together, we propose that the combination of acbp1 and acbp2 mutations probably abolishes the formation of an acyl-CoA pool in the ER or disrupts acyl-CoA/lipid trafficking between the ER and the plasma membrane during early embryogenesis.


We thank Professor R. Welti and Ms M. Roth (Kansas Lipidomics Research Center, KS, USA) for lipid profiling, the Arabidopsis Information Resource (TAIR) for the provision of the acbp2 mutant seed pools and Professor S. F. Chen (University of Hong Kong) for the provision of high-performance liquid chromatography for acyl-CoA analysis. This work was supported by the University of Hong Kong (CRCG grant 10207211, studentships to QFC and JYM, and postdoctoral fellowship to SX) and the University Grants Committee of the Hong Kong Special Administrative Region, China (Project No. AoE/B-07/99). Lipid profiling was performed at Kansas Lipidomics Research Center, where method development and instrument acquisition were supported by the National Science Foundation (EPS 0236913, MCB 0455318, DBI 0521587), Kansas Technology Enterprise Corporation, Kansas IDeA Network of Biomedical Research Excellence (INBRE) of National Institute of Health (P20RR16475) and Kansas State University, KS, USA.