The rice acyl-CoA-binding protein gene family: phylogeny, expression and functional analysis



This article is corrected by:

  1. Errata: Corrigendum Volume 190, Issue 3, 807, Article first published online: 4 March 2011

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


  • Acyl-CoA-binding proteins (ACBPs) show conservation in an acyl-CoA-binding domain (ACB domain) which binds acyl-CoA esters. Previous studies on plant ACBPs focused on eudicots, Arabidopsis and Brassica. Here, we report on the phylogeny and characterization of the ACBP family from the monocot Oryza sativa (rice).
  • Phylogenetic analyses were conducted using 16 plant genomes. Expression profiles of rice ACBPs under normal growth, as well as biotic and abiotic stress conditions, were examined by quantitative real-time reverse-transcription polymerase chain reactions. In vitro acyl-CoA-binding assays were conducted using recombinant (His)6-tagged ACBPs.
  • The ACBP family diversified as land plants evolved. Classes I and IV show lineage-specific gene expansion. Classes II and III are closely related phylogenetically. As in the eudicot Arabidopsis, six genes (designated OsACBP1 to OsACBP6) encode rice ACBPs, but their distribution into various classes differed from Arabidopsis. Rice ACBP mRNAs showed ubiquitous expression and OsACBP4, OsACBP5 and OsACBP6 were stress-responsive. All recombinant rice ACBPs bind [14C]linolenoyl-CoA besides having specific substrates.
  • Phylogeny, gene expression and biochemical analyses suggest that paralogues within and across classes are not redundant proteins. In addition to performing conserved basal functions, multidomain rice ACBPs appear to be associated with stress responses.


In higher plants, two groups of proteins that bind lipids have been characterized: the lipid-transfer proteins (Kader, 1996) and the acyl-CoA-binding proteins (ACBPs) (Leung et al., 2004; Xiao & Chye, 2009). The ACBPs bind long-chain acyl-CoA (LCACoA) esters with high specificity and affinity (Rasmussen et al., 1993; Chye, 1998; Chye et al., 2000; Knudsen et al., 2000; Leung et al., 2004, 2006; Burton et al., 2005). As a highly specialized acyl-CoA pool former and transporter, mammalian ACBPs together with LCACoAs regulate lipid metabolism and gene expression (Færgeman & Knudsen, 1997; Knudsen et al., 2000).

The highly-conserved prototype, cytosolic 10-kDa ACBP ranging between 82 and 92 amino acid residues, occurs in animals, plants, fungi and protists, but not prokaryotes and Archaea except for several plant and animal pathogens (Burton et al., 2005). An overall phylogenetic analysis of the 10-kDa prototype ACBP from metazoans and non-metazoans show lineage-specific duplication and purifying selection during evolution (Burton et al., 2005). Prototype ACBPs show conservation in structure and function spanning 1500 million yr (Færgeman et al., 2007). In addition to 10-kDa ACBPs, large multifunctional proteins containing a predicted acyl-CoA-binding domain (ACB domain) have been identified in many eukaryotes. By contrast, the evolutionary relationships of these larger ACBPs have not been investigated.

Plant 10-kDa ACBPs were first identified from Brassica napus (Hills et al., 1994) and Arabidopsis thaliana (Engeseth et al., 1996). Recombinant Arabidopsis ACBP6 (AtACBP6), the homologue of the well-characterized mammalian 10-kDa ACBP, binds acyl-CoA esters as well as saturated (16:0 and 18:0) and unsaturated (18:1 and 18:2) phosphatidylcholine (PC) (Engeseth et al., 1996; Chen et al., 2008; Xiao et al., 2008a). AtACBP6 overexpression confers freezing tolerance by upregulating the expression of phospholipase Dδ (PLDδ), which generates phosphatidic acid (PA) from phospholipids. Hence, AtACBP6 through its PC-binding ability may participate in the intracellular binding and trafficking of PC during plant lipid metabolism (Chen et al., 2008). The overexpression of B. napus 10-kDa ACBP in Arabidopsis culminated in alternated fatty acid composition in developing seeds, possibly by provision of greater acyl-CoA-binding capacity, which changes distribution of 18:1-CoA between elongation and incorporation into PC (Yurchenko et al., 2009).

Other than the 10-kDa AtACBP6, five larger forms of AtACBPs, ranging from 37.5 to 73.1 kDa, occur in Arabidopsis (Leung et al., 2004; Xiao & Chye, 2009). We have reported on the characterization of the first eukaryotic ACBP family using these Arabidopsis ACBPs (Leung et al., 2004; Xiao & Chye, 2009). AtACBP1 and AtACBP2 are subcellularly localized to the endoplasmic reticulum (ER) and plasma membrane (Chye et al., 1999; Li & Chye, 2003), while AtACBP3 is membrane-associated and apoplast-targeted (Leung et al., 2006; Xiao et al., 2010). In contrast, the remaining three AtACBPs (AtACBP4, AtACBP5 and AtACBP6) are cytosolic proteins (Chen et al., 2008; Xiao et al., 2008b). Both recombinant AtACBP1 and AtACBP2 bind linoleoyl-CoA and linolenoyl-CoA esters, the precursors of phospholipid membrane repair, and Arabidopsis overexpressing AtACBP1 or AtACBP2 displayed an enhanced tolerance to heavy metals, suggesting their role in phospholipid membrane repair following heavy metal stress (Gao et al., 2008; Xiao et al., 2008a). AtACBP1 and AtACBP2 were immunolocalized to the embryos of developing seeds (Chye et al., 1999; Chen et al., 2010) and are essential in early embryo development because the acbp1/acbp2 double mutant was observed to be embryo lethal (Chen et al., 2010). Our recent findings reveal that recombinant AtACBP3 binds PC, phosphatidylethanolamine (PE) and acyl-CoA esters in vitro, and AtACBP3 overexpression accelerated both dark-induced and age-dependent leaf senescence, demonstrating that AtACBP3 regulates leaf senescence by modulating phospholipid metabolism and promoting the degradation of the autophagy-related protein ATG8 (Xiao et al., 2010). Recombinant AtACBP4 and AtACBP5 preferentially bind oleoyl-CoA esters in the cytosol, suggesting that these ACBPs are potential candidates for acyl-CoA transport from the chloroplasts to the ER (Xiao et al., 2008b).

In rice (Oryza sativa), a model monocot species, a 10-kDa ACBP was initially identified as a major protein in phloem sap, indicating its potential in acyl-CoA transfer from companion cells to sieve elements and in long-distance transport of acyl-CoA esters (Suzui et al., 2006). Given the newly established roles of the ACBP family in the eudicot A. thaliana (Xiao & Chye, 2009; Napier & Haslam, 2010) and the lack of corresponding knowledge in monocots, we have initiated investigations of the ACBP family in rice. Evidence that AtACBPs can be used to confer tolerance to abiotic and biotic stresses in transgenic Arabidopsis (Chen et al., 2008; Gao et al., 2008, 2010; Xiao et al., 2008a, 2010) makes them attractive for applications in agriculture. However, an understanding of endogenous ACBPs in rice should first be sought before ACBPs are used in genetic manipulations to enhance stress tolerance in rice, an important staple crop in Asia. Here, the rice ACBP family was investigated by examining their expression profiles and acyl-CoA-binding abilities using recombinant (His)6-tagged proteins. We also report on the evolutionary relationship of plant ACBPs by analysis of amino acid sequences containing ACB domains among different model plant genomes, ranging from green algae, mosses, gymnosperms, eudicots (including Arabidopsis) and monocots (including rice).

Materials and Methods

Identification of plant ACBPs and alignment construction

Plant protein sequences consisting of an ACB domain were obtained by detailed blast searches of nonredundant protein sequences from the NCBI (National Center for Biotechnology Information,, PLAZA ( and Phytozome v5.0 (, using known Arabidopsis AtACBP1, AtACBP3, AtACBP4 or AtACBP6 sequences as independent probes. Available complete or partial genome sequences selected for the phylogenetic analysis include those from model organisms representing chlorophytes (Chlamydomonas reinhardtii (Merchant et al., 2007) and Ostreococcus lucimarinus (Palenik et al., 2007)), bryophyte (moss (Physcomitrella patens; Rensing et al., 2008)), lycophyte (spikemoss (Selaginella moellendorffii; phytozome v5.0)), gymnosperms (sitka spruce (Picea sitchensis; NCBI)), eudicots (cottonwood (Populus trichocarpa; Tuskan et al., 2006), Arabidopsis (A. thaliana; Arabidopsis Genome Initiative, 2000), grape (Vitis vinifera; Jaillon et al., 2007), cucumber (Cucumis sativus; Huang et al., 2009), papaya (Carica papaya; Ming et al., 2008), castor bean (Ricinus communis; Phytozome v5.0), soy bean (Glycine max; Phytozome v5.0) and barrel medic (Medicago truncatula; Phytozome v5.0)), and monocots (sorghum (Sorghum bicolor; Paterson et al., 2009), rice (O. sativa; Goff et al., 2002) and maize (Zea mays; Phytozome v5.0)). Eighty-four protein sequences with ACB domains, predicted using motif database Pfam (, were included in the data set. Incomplete and redundant sequences were removed from the final set based on gene annotation information from PLAZA and Phytozome. The gene identifier and accession numbers of sequences used in this study are listed in the Supporting Information Table S1. Six protein sequences from the non-plant taxa representing different ACBP classes were chosen as outgroups in the phylogenetic analyses: the nematode (Caenorhabditis elegans) homologues of AtACBP1 (NP_499817), AtACBP3 (NP_499531) and AtACBP6 (NP_491412), as well as the acyl-coenzyme A binding domain containing two proteins (ACBD2) from human (Homo sapiens NP_006108), mouse (Mus musculus NP_035998) and nematode (NP_496330) that each contains an ACB domain and an enoyl-CoA hydratase/isomerase (ECH) domain (Fan et al., 2010). A total of 90 full-length amino acid sequences were initially aligned using BLOSUM 30 matrix in clustal x ver. 2.0.12 (Larkin et al., 2007). An open gap penalty of 10 and an extend gap penalty of 0.1 were used in pairwise alignments, and an extend gap penalty of 0.2 was applied in multiple alignment with a delay divergent setting of 30%. Owing to the variable and poorly aligned sites at the N- and C-terminal ends, conserved domains in plant ACBPs were aligned and optimized manually using bioedit ver. 7.0.5 (Hall, 1999) and se-al ver. 2.0a11 (Rambaut, 1996).

Phylogenetic analysis

A phylogenetic tree based on the alignment of plant ACBP conserved domains was generated using neighbor-joining (NJ) methods as implemented in paup* ver. 4.0b10 (Swofford, 2003). Maximum likelihood (ML) analysis was also conducted using the LG amino acid substitution matrix (Le & Gascuel, 2008) as determined by prottest ver. 2.4 (Abascal et al., 2005). The ML tree was generated using phyml ver. 3.0 (Guindon & Gascuel, 2003) with bootstrap values calculated from 100 replicates. The phylogenetic tree was rooted using the nematode (C. elegans) ACBD2.

Plant materials and growth conditions

Rice (O. sativa cv Zhonghua 11) seeds were germinated in the dark for 3 d at 25°C and grown in pots under 14-h light (c. 29°C)/10-h dark (c. 25°C) cycles. Leaf, stem and root samples were collected 49 d after germination (Koller et al., 2002). Seed samples were collected at R4 (anthesis), R5 (milk) and R6 (soft dough stages) of rice flower to grain development, where R1 to R10 represent the 10 growth stages in reproductive development (Counce et al., 2000). For stress treatments, rice seeds were germinated in distilled water in the dark for 3 d at 25°C. After 7-d growth in distilled water, seedlings were transferred to a Petri dish containing cotton soaked in nutrient solution and incubated in a growth chamber (14-h light : 10-h dark cycles at 28°C). The nutrient solution consisted of nitrogen (N) (NH4NO3 at 20 mg l−1), phosphorus (P) (KH2PO4 at 0.05 mg l−1), potassium (K) (KCl at 10 mg l−1), magnesium (Mg) (MgCl2 at 1 mg l−1), iron (Fe) (FeNaEDTA at 1 mg l−1) and Hoagland trace elements according to Tang et al. (1989).

Stress treatments

Two-wk-old seedlings were subjected to abiotic or biotic stress treatments. For cold stress, seedlings were placed at 4°C and sampled at 0, 3, 6, 12 and 24 h after treatment (Islam et al., 2009). For drought stress, whole seedlings were exposed in the air without water supply and sampled at 0, 3, 6, 12 and 24 h (Islam et al., 2009). The roots of seedlings were submerged in 200 mM NaCl solution for salt stress and sampled at 0, 6, 12, 24 and 48 h after treatment (Islam et al., 2009). For wounding treatment, leaves of seedlings were scratched using a needle. Wounded leaves were subsequently maintained on water-saturated filter paper and sampled at 0, 0.5, 2 and 24 h after treatment (Nakashima et al., 2007). In rice blast fungus (Magnaporthe grisea) infection, leaves were sprayed with a spore suspension at a concentration of 5 × 105 conidia ml−1 (Nakashima et al., 2007). Each treatment consisted of three replicates.

Quantitative real-time polymerase chain reactions (qRT-PCR)

Total RNA was isolated from samples using TRIzol reagent (Cat No. 15596-018; Invitrogen), and 5 μg was reverse-transcribed using the Superscript First-strand Synthesis System (Cat No. 12371-019; Invitrogen) according to the manufacturer’s instructions. The iCycler (Bio-Rad) and KAPA SYBR FAST qPCR kit (KK4607) were used in qRT-PCR with rice Actin1 (GenBank accession number X16280, Xiang et al., 2008) as an internal control; specific primers for OsACBP are listed in Table S2. Conditions for qRT-PCR were: 3 min at 95°C followed by 40 cycles of 95°C for 15 s, 60°C for 25 s and 81°C for 15 s. The specificity of PCR products was checked using a melting curve programme (61–95°C with fluorescence read every 0.5°C). Individual PCR efficiency was checked using LingerPCR (Ramakers et al., 2003). The relative expression of a specific gene was determined as previously described (Ruijter et al., 2009). The relative expression of OsACBPs was calculated from three independent qRT-PCR experiments.

Construction of plasmids

Polymerase chain reaction fragments comprising full-length cDNAs of OsACBP1 (0.28-kb), OsACBP2 (0.28-kb) and OsACBP3 (0.47-kb), as well as partial cDNAs of OsACBP4 (0.91-kb containing amino acids 32 to 336), OsACBP5 (1.5-kb containing amino acids 32 to 569) and OsACBP6 (1.3-kb containing amino acids 33 to 461) were generated by reverse-transcription polymerase chain reactions (RT-PCR) using the following primer pairs respectively: ML1064/ML1065, ML1062/ML1063, ML1057/ML1058, ML1060/ML1061, ML1053/ML1056, and ML1050/ML1051 (Table S3). The PCR products were cloned into vector pGEM-T Easy (Promega) to generate plasmids pAT500, pAT499, pAT496, pAT498, pAT502 and pAT495, respectively, and were confirmed by DNA sequence analysis. Each BamHI–EcoRI fragment of pAT500, pAT499, pAT496 and pAT502, and the BamHI–BamHI fragment of pAT498 and BglII–EcoRI fragment of pAT495 were cloned in-frame to the (His)6-tag in vector pRSET A (Invitrogen) to yield plasmids pAT504, pAT503, pAT544, pAT551, pAT543 and pAT501, respectively. The calculated molecular masses of (His)6-tagged OsACBP are: (His)6-OsACBP1 (14.2 kDa), (His)6-OsACBP2 (14.3 kDa), (His)6-OsACBP3 (21.8 kDa), (His)6-OsACBP4 (36.7 kDa), (His)6-OsACBP5 (57.8 kDa) and (His)6-OsACBP6 (49.7 kDa).

Expression and purification of recombinant (His)6-tagged OsACBPs

(His)6-tagged OsACBP expressing plasmids were introduced into Escherichia coli BL21 (DE3)Star pLysS (Invitrogen) and cultured to OD600nm = 0.4. Cells were induced with optimized concentration of isopropyl-β-d-thiogalactopyranoside (IPTG) and induction time: (His)6-OsACBP1 and (His)6-OsACBP2, 0.1 mM IPTG at 37°C for 4 h; (His)6-OsACBP3, (His)6-OsACBP4 and (His)6-OsACBP5, 0.1 mM IPTG at 28°C for 5 h; and (His)6-OsACBP6, 0.5 mM IPTG at 28°C for 3 h. Cells were harvested and soluble protein and inclusion proteins were treated following procedures described by Invitrogen. Each protein sample (10 μl) was analysed by electrophoresis on 10% ((His)6-OsACBP4, (His)6-OsACBP5 and (His)6-OsACBP6), 12% ((His)6-OsACBP3) or 15% ((His)6-OsACBP1 and (His)6-OsACBP2) sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) followed by staining with Coomassie Brilliant Blue. In Western blot analysis, proteins were electrophoretically transferred to ECL membranes (Amersham) from SDS-PAGE using the Trans-Blot cell (Bio-Rad). Western blot analysis was performed as described by Sambrook et al. (1989). The Penta-His HRP Conjugate (Qiagen) was used to detect (His)6-tagged proteins.

(His)6-tagged OsACBP1, OsACBP2, OsACBP4, OsACBP5 were optimally expressed in the soluble fraction, and (His)6-tagged OsACBP3, OsACBP6 as inclusion bodies. Batch extractions of (His)6-tagged OsACBP1, OsACBP2, OsACBP4 and OsACBP5 were carried out according to Chye (1998) and extractions of (His)6-tagged OsACBP3 and OsACBP6 were carried out under denaturing conditions following Xiao et al. (2008a). Each protein was purified using an affinity column of Ni-NTA Agarose (Qiagen) according to the supplier’s instructions. (His)6-OsACBP1, (His)6-OsACBP2, (His)6-OsACBP4 and (His)6-OsACBP5 were eluted using native elution buffer (50 mM NaH2PO4 pH 8.0, 300 mM NaCl, 250 mM Imidazole). (His)6-OsACBP3 and (His)6-OsACBP6 were eluted in Buffer D (8 M urea, 0.1 M Na2HPO4, 0.01 M Tris-Cl pH 5.9, 5% glycerol) and Buffer E (8 M urea, 0.1 M Na2HPO4, 0.01 M Tris-Cl pH 4.5, 5% glycerol) followed by refolding in 50 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) sodium salt, 200 mM NaCl, 2 mM MgCl2, 5 mM EDTA, 10% glycerol, 0.005% (v : v) Tween-20, pH 7.9 at 4°C (Xiao et al., 2008a). After dialysis for 24 h in double-distilled water, proteins were harvested and protein concentrations were determined according to Bradford (1976).

Acyl-CoA-binding assays using purified (His)6-tagged OsACBPs

In vitro binding assays (Leung et al., 2004) with [14C]linoleoyl-CoA, [14C]linolenoyl-CoA, [14C]oleoyl-CoA, and [14C]palmitoyl-CoA (American Radiolabelled Chemicals, St. Louis, MO, USA) were carried out using Lipidex-1000 (Perkin-Elmer, Boston, MA, USA). Assays were performed in triplicate, with blanks, at each concentration of acyl-CoA. Purified recombinant OsACBPs were diluted with binding buffer (10 mM potassium phosphate buffer, pH 7.4) to a concentration of 0.8 μM. Each protein, at a concentration of 0.2 μM, was mixed with [14C]palmitoyl-CoA, [14C]oleoyl-CoA, [14C]linoleoyl-CoA or [14C]linolenoyl-CoA at final acyl-CoA concentrations ranging from 0.2 to 2 μM. After incubation for 30 min at 37°C, the mixtures were chilled on ice for 10 min. Ice-cold 50% slurry of Lipidex-1000 in binding buffer (400 μl) was added and the mixture was kept on ice for 10 min. Samples were centrifuged at 12 000 g for 10 min at 4°C and a 200 μl aliquot of the supernatant was taken for analysis of radioactivity counts using LS 6500 liquid scintillation counter (Beckman, Fullerton, CA, USA). The ratio for the number of moles of bound acyl-CoA molecules to the number of moles of recombinant proteins was calculated using the equation of Motulsky & Neubig (1997). The dissociation constant Kd values were determined from a Scatchard plot.


Identification and characterization of predicted plant ACBPs

To identify the plant ACBP families, blastp searches were initially conducted using NCBI nonredundant protein sequences (nr) database, Phytozome and PLAZA with Arabidopsis AtACBP1, AtACBP3, AtACBP4 and AtACBP6 amino acid sequences as probes. Finally, 84 plant amino acid sequences containing an ACB domain were targeted in 16 plant species representing different branches on the evolutionary tree for green plants. Ostreococcus lucimarinus and C. reinhardtii are unicellular aquatic green algae belonging to the Chlorophyta, and are members of lineages that diverged before the evolution of land plants. The moss, P. patens, and the lycophyte, S. moellendorffii, are representatives of two early-divergent lineages of land plants: mosses diverged before the origin of vascular plants and lycophytes before the origin of seed plants. Sitka spruce (P. sitchensis) is included as a representative gymnosperm, while the remaining species are angiosperms, including monocots and eudicots. Targeted ACBP proteins range in size from 66 to 682 amino acids. The modular nature of the ACB domain is supported by the presence of multidomain protein families with different domain architecture. In addition to the ACB domain, some proteins contain ankyrin (ANK) repeats and kelch motifs (Table S4). Most proteins have N-terminal ACB domains accompanied by highly heterogeneous C-terminal portions. Based on the alignment of the ACB domain (c. 85 amino acids), 29 residues were found to be conserved across all ACBPs. Six functional conserved residues (Y/F, Y, Y, K, K, Y/F) in bovine ACBP (Kragelund et al., 1999) are also conserved in plant ACBPs. However, there were variations at other positions in this domain with some showing specificity to their own subgroup.

Four classes of plant ACBPs are identified and their structures are illustrated in Table S4: Class I, small ACBPs; Class II, ANK-ACBPs; Class III, large ACBPs; and Class IV, Kelch-ACBPs. Class I proteins consist of an average of 100 amino acid residues and are present in almost every plant species examined (Burton et al., 2005). This class is very similar to first-characterized mammalian ACBPs in its size and structure. Classes II and IV are multi-domain proteins containing C-terminal ankyrin repeats and kelch motifs, respectively, which enable ACBPs to interact with partner proteins (Li & Chye, 2004; Gao et al., 2008, 2010; Li et al., 2008) (Table S4). These semi-conserved regions are not present in other subgroups, suggesting that they are associated with a specific function. Class III (large ACBPs) contains an average of 450 amino acid sequences, each with a C-terminal ACB domain and the majority are accompanied by a signal peptide or transmembrane domain at the N-terminus that dictates subcellular localisation. Proteins of Classes II, III and IV are about three- to six-fold larger than those of Class I (Table S4). Based on blastp searches within NCBI, Classes I-III are not unique to the plant kingdom as they also occur in vertebrates, nematodes and fungi, whereas nonplant Class IV-like proteins show limited (< 40%) identity at the ACB domain to, but retain the kelch motifs of, land plant Class IV. Interestingly, a group of ACBP protein sequences of intermediate length (109–208 amino acids) occurs only in green algae, moss and gymnosperm, and was categorized as Class 0. Although Class 0 has similar domain architecture to Class III ACBPs in higher plants, the former has relatively shorter protein length (109 and 208 amino acids in O. lucimarinus, 141 in P. patens and 163 in P. sitchensis) than that of Class III (215–682 amino acids).

Class I ACBPs are conserved from unicellular algae to higher plants. However, either one or more Class II, III and IV ACBPs could be absent in early-divergent plant lineages (Table 1): coexistence of Class I, II, III and IV ACBPs becomes stable after the divergence of gymnosperms and angiosperms. The apparent absence of Class III in barrel medic may result from the incomplete genome sequences available (Table 1). The presence of multiple classes of ACBPs in early-divergent plant lineages as well as in higher plants, and observations of conservation of four classes among higher plants, indicate that ACBP functions in lipid metabolism and other physiological functions existed before the evolution of land plants (Table 1). These results suggest that land plants retain multifold ACBP functions across lineages.

Table 1. Predicted copy number of plant acyl-CoA-binding protein (ACBP) classes having different domains
 Class IClass IIClass IIIClass IVClass 0Total
  1. aActual copy number may exceed this because genome sequence is still incomplete.

Green alga (Chlamydomonas reinhardtii)101103
Green alga (Ostreococcus lucimarinus)100124
Moss (Physcomitrella patens)230218
Spikemoss (Selaginella moellendorffii)1a2a0a1a0a4a
Sitka spruce (Picea sitchensis)1a1a0a1a1a4a
Arabidopsis (Arabidopsis thaliana)121206
Papaya (Carica papaya)111104
Cottonwood (Populus trichocarpa)222107
Grape (Vitis vinifera)212106
Soy bean (Glycine max)2a2a4a3a0a11a
Barrel medic (Medicago truncatula)1a1a0a1a0a3a
Castor bean (Ricinus communis)1a1a1a1a0a4a
Cucumber (Cucumis sativus)111104
Rice (Oryza sativa)311106
Sorghum (Sorghum bicolor)211105
Maize (Zea mays)2a1a1a1a0a5a

As the green algae sampled possess single genes encoding Classes I and IV, and the moss (P. patens) has multiple Class I and Class IV members, we suggest that duplication events within Classes I and IV occurred before the origin of land plants. The monocots examined (rice, sorghum and maize) appear to have two or three members in Class I and only single members in Classes II, III and IV (Table 1). This implies that gene duplication may have been limited to Class I in rice and sorghum of which complete genome sequences are available. By contrast, eudicots have one or more Class I accompanied by several members in Classes II, III and IV. The soybean ACBP family, comprising 11 members, emerged as the largest among all species examined (Table 1).

Phylogenetic analyses of plant ACBPs

In order to obtain an overview of the phylogenetic relationships among plant ACBP families, both NJ and ML trees were reconstructed based on the conserved regions. The tree topologies of NJ (tree not shown) and ML (Fig. 1) are very similar: the only major discrepancies are the doubtful positions of P. patens Class II ANK-ACBPs. Class I ACBPs are basal to other Classes, although the topology lacks bootstrap support (Fig. 1). The green algae Class I ACBP homologues (i.e. C. reinhardtii and O. lucimarinus) are possibly distantly related to the remaining taxa in Class I ACBPs. The remaining taxa within Class I comprise mosses, lycophytes, gymnosperms and angiosperms, forming a relatively well-supported clade (Fig. 1: BS = 83%). The phylogenetic relationships within Class I are broadly similar to the plant kingdom phylogeny, with mosses and lycophytes basal to the seed plants. Class I ACBPs in land plants originated before the divergence of moss lineages. We note that the ACBP sequence Pt00g72560 of the eudicot P. trichocarpa may be misplaced in the phylogeny (although lacking bootstrap support) as it is not grouped within the Class I clade (Fig. 1); significantly, however, sequence Pt00g03340 of the same species is unequivocally clustered within Class I. Class 0 is a moderately supported clade with a bootstrap value of 78% and consists of green algae, moss and spruce ACBPs only. The lineages within Class 0 are highly divergent, as indicated by the long branch lengths. The protein size of Class 0 (109–208 amino acids) is intermediate between Class I (66–155 amino acids) and III (215–682 amino acids), suggesting that ACBPs may increase in size during evolution. Class III and IV ACBPs also appear to form two moderately to weakly supported clades (BS = 76% and BS = 55%, respectively), although with the exclusion of the green alga C. reinhardtii Class III ACBP; the placement of the latter notably lacked bootstrap support throughout the tree. As with the Class I green algae ACBPs, the green alga Class III ACBPs are likely to be distantly related to those in eudicots and monocots. The absence of Class III ACBPs in mosses and gymnosperms suggests that green alga Class III ACBPs and land plant Class III ACBPs are not descended from a common ancestor. Land plant Class III ACBPs originated after the divergence of gymnosperm and angiosperm Class II, which has closer relationship with gymnosperm Class II (BS = 66%). Class II ACBPs do not appear to form a monophyletic group (unequivocal conclusions are precluded because of inadequate bootstrap support), although monocot and eudicot Class II ACBPs collectively form a weakly supported clade (BS = 63%).

Figure 1.

 Phylogeny of plant acyl-CoA-binding proteins (ACBPs) based on maximum likelihood analysis using PhyML with LG substitution model, based on 90 ACBP sequences. Nematode (Caenorhabditis elegans) Class I, II and III ACBPs, and human, mouse and nematode ACBD2 were used as outgroups. The tree was rooted with the nematode ACBD2. The ACBP protein sequences were identified from proteomes of different plant species, 10 with completely sequenced genomes and six with incomplete genomes. The assigned gene identifiers and accession numbers for the genes are displayed on the tree. Bootstrap values (> 50%) are recorded at the nodes. Abbreviations of generic names are given in the Materials and Methods section.

The rice ACBP family

Six genes (Os08g06550, Os06g02490, Os03g37960, Os04g58550, Os03g14000 and Os03g61930) encode rice ACBPs, designated as OsACBP1 to OsACBP6, respectively (Fig. 2). OsACBP1 to OsACBP3 range from 91 to 155 amino acids while OsACBP4 (336 amino acids), OsACBP5 (569 amino acids) and OsACBP6 (655 amino acids) are much larger proteins. The ACB domain is located at the N-terminal segment in each, except OsACBP5 which contains a C-terminal ACB domain. According to the phylogenetic tree, OsACBP1–OsACBP3 belong to Class I, but OsACBP3 (which shares 95% identity with OsACBP1) is longer because of a 63-amino acid C-terminal extension. OsACBP4 possesses ankyrin repeats and belongs to Class II, OsACBP5 is designated Class III and kelch motif-containing OsACBP6 is a Class IV protein.

Figure 2.

 Comparison in architecture of the rice and Arabidopsis acyl-CoA-binding protein (ACBP) families. The ACB domain, ankyrin repeats, kelch motifs, putative transmembrane domain, signal peptide and putative chloroplast transit peptide are shown. Six rice ACBPs are designated as OsACBP1–OsACBP6 (protein size indicated in the brackets). Arabidopsis ACBPs are designated as AtACBP1–AtACBP6 (Chye, 1998; Chye et al., 2000; Leung et al., 2004, 2006; Xiao et al., 2008a).

Interestingly, both rice and Arabidopsis ACBP families contain only one member each in Class III, namely, OsACBP5 and AtACBP3. By contrast, one Class I (AtACBP6), two Class II (AtACBP1, AtACBP2) and two Class IV (AtACBP4, AtACBP5) members occur in Arabidopsis, and three Class I (OsACBP1, OsACBP2 and OsACBP3) and one each of Class II (OsACBP4) and Class IV (OsACBP6) are present in rice (Fig. 2). Comparison of amino acid identities between rice and Arabidopsis are as follows: Classes I (71%), II (45%), III (56%) and IV (62%).

Expression profiles of OsACBPs

Differences in expression of OsACBP genes in various organs and stages of seed development were addressed using qRT-PCR. The expression of the six OsACBP mRNAs was detected in all organs, such as seed, leaf, stem and root (Fig. 3). In comparison with germinating seeds (72 h after imbibition), all six mRNAs were relatively highly expressed in leaf, moderately expressed in root and showed very low expression in stem. However, expression varied among the six during seed development from spikelets with flowers at anthesis to milk and soft dough stages. OsACBP1 expression remained at about the same level at anthesis, milk and soft dough stages while OsACBP2 expression peaked at dough stage. By contrast, OsACBP3 and OsACBP4 peaked at anthesis. OsACBP5 was highly-expressed throughout the whole reproductive phase. For OsACBP6, expression during seed development was lower than leaf with lowest expression at the milk stage.

Figure 3.

 Quantitative real-time PCR analysis of OsACBP expression under normal growth conditions. Leaf, stem and root samples were collected 49 d after germination. Seed samples were collected at anthesis, milk and soft dough stages. The expression levels were normalized to the expression level of 72-h germinating seeds. (a) OsACBP1, (b) OsACBP2, (c) OsACBP3, (d) OsACBP4, (e) OsACBP5 and (f) OsACBP6. H, value > 72-h germinating seed (< 0.05); L, value < 72-h germinating seed (< 0.05). Bars represent SE (= 3).

The roles of OsACBPs in responses to abiotic and biotic stress were examined using qRT-PCR. OsACBPs differed in response to different treatments, including high salinity, drought, cold, wounding and rice blast fungus infection (Fig. 4). In high salinity and drought treatments, OsACBP1, OsACBP2 and OsACBP3 mRNAs were not affected (Fig. 4a,b). OsACBP4 and OsACBP5 mRNAs peaked at 12 h after salt treatment then remained at relatively high levels. OsACBP4 was rapidly induced by drought. OsACBP1, OsACBP2, OsACBP3, OsACBP4 and OsACBP5 mRNAs were suppressed within 12 h after cold treatment and then recovered to near normal levels at 24 h. Cold treatment had no apparent influence on OsACBP6 expression up to 12 h, and lower expression ensued at 24 h (Fig. 4c). Wounding treatment suppressed the expression of all OsACBPs except OsACBP5 and OsACBP6 (Fig. 4d) which were rapidly induced, peaking at 0.5 h followed by a decrease in expression to levels lower than untreated. Rice blast fungus (M. grisea) infection induced the expression of OsACBP5 and suppressed the expression of the other genes, especially OsACBP1, OsACBP2, OsACBP3 and OsACBP4 (Fig. 4e). These results were found to correlate well with microarray data (

Figure 4.

 Quantitative real-time PCR analysis of OsACBPs expression under stress treatments: (a) high salinity, (b) drought, (c) cold, (d) wounding and (e) infection with rice blast fungus (open squares, mock; closed squares, 5 dpi, days post inoculation). The expression levels were normalized to that of Actin1. H, value > at 0 h (< 0.05); L, value < at 0 h (< 0.05). Bars represent SE (= 3).

Analysis of the 5′-flanking regions (1.5 kb upstream of the putative transcription start site) of OsACBPs using the PLACE database ( and PlantCARE ( revealed the presence of many putative stress response-related cis-elements including DRE/CRT (dehydration-responsive element/C-repeat), ABRE (cis-acting element involved in the abscisic acid responsiveness), MBS (MYB-binding site involved in drought-inducibility), and CGTCA motifs (cis-acting regulatory element involved in the methyl jasmonate (MeJA)-responsiveness). In comparison, the number of putative ABRE and MeJA responsive elements exceeded those associated with other forms of stress or plant hormone responses. The presence of DRE/CRT elements was specific to OsACBP4, OsACBP5 and OsACBP6.

(His)6-tagged OsACBPs bind acyl-CoA esters in vitro

All six (His)6-tagged OsACBPs recombinant proteins were expressed and purified in E. coli. Use of the (His)6-tagged expression vector pRSET A (Invitrogen) enabled easy purification of the recombinant proteins from crude bacterial extracts using Ni-NTA agarose (Qiagen). As advised by the supplier, the transmembrane domain or signal peptide were deleted from OsACBP4, OsACBP5 and OsACBP6 to prevent occurrence of possible toxic effects on host cells arising from association or incorporation the recombinant protein into vital membrane systems.

(His)6-OsACBP1, (His)6-OsACBP2, (His)6-OsACBP4 and (His)6-OsACBP5 were detected in the soluble fractions of bacterial extracts, while (His)6-OsACBP3 and (His)6-OsACBP6 were in inclusion bodies. On Coomassie Brilliant Blue-staining, the molecular masses were as predicted: (His)6-OsACBP1, 14.2 kDa; (His)6-OsACBP2, 14.3 kDa; (His)6-OsACBP3, 21.8 kDa; and (His)6-OsACBP6, 49.7kDa. However, (His)6-OsACBP4 (calculated molecular mass 36.7 kDa; apparent size 45 kDa) and (His)6-OsACBP5 (calculated molecular mass 57.8 kDa; apparent size c. 98 kDa) appeared larger than predicted on SDS-PAGE (Fig. 5). On Western blot analysis, all recombinant (His)6-tagged OsACBP proteins crossreacted with the Penta-His HRP Conjugate (Qiagen) (data not shown) indicating that (His)6-OsACBP4 and (His)6-OsACBP5 had migrated anomalously in SDS-PAGE gels. The isoelectric point (pI) values of (His)6-OsACBP4 and (His)6-OsACBP5 are 4.6 and 4.2, respectively. The low mobility of these two proteins on SDS-PAGE gel may have been caused by the relatively large negative charge at neutral pH (Matagne et al., 1991) (−28 for (His)6-OsACBP4 and –70 for (His)6-OsACBP5).

Figure 5.

 Purification of (His)6-OsACBPs recombinant protein. (a) 15% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS--PAGE) gel shows the 14.2-kDa (His)6-OsACBP1 protein purified from Escherichia coli soluble fraction 4 h after isopropyl β-d-thiogalactoside (IPTG) induction; (b) 15% SDS-PAGE gel shows the 14.3-kDa (His)6-OsACBP2 protein purified from E. coli soluble fraction 4 h after IPTG induction; (c) 12% SDS-PAGE gel shows the 21.8-kDa (His)6-OsACBP3 protein purified from E. coli inclusion bodies 5 h after IPTG induction; (d) 10% SDS-PAGE gel shows the 45-kDa (His)6-OsACBP4 protein purified from E. coli soluble fraction 5 h after IPTG induction; (e) 10% SDS-PAGE gel shows the 98-kDa (His)6-OsACBP5 protein purified from E. coli soluble fraction 5 h after IPTG induction; (f) 10% SDS-PAGE gel shows the 49.7-kDa (His)6-OsACBP6 protein purified from E. coli inclusion bodies 3 h after IPTG induction. M, Marker. For (a), (b), (d) and (e), lane 1, flow-through fraction; lanes 2–9, washing fraction with native washing buffer; lanes 10–13, eluted fraction with native elution buffer. For (c) and (f), lane 1, flow-through fraction; lane 2, washing fraction at pH 6.3; lanes 3–8, eluted fractions with buffer D at pH 5.9; lanes 9–14, eluted fractions with buffer E at pH 4.5.

In vitro binding assays with acyl-CoA esters were carried out to investigate the potential of rice ACBPs in lipid transport. Using Lipidex assays, the binding of (His)6-tagged OsACBPs to acyl-CoA esters, [14C]palmitoyl-CoA, [14C]oleoyl-CoA, [14C]linoleoyl-CoA and [14C]linolenoyl-CoA, which represent saturated, monounsaturated and polyunsaturated fatty acids, respectively, were tested. Palmitoyl-CoA, oleoyl-CoA, linoleoyl-CoA and linolenoyl-CoA were selected because this set of fatty acids had also been used in the characterization of Arabidopsis and Brassica ACBPs (Leung et al., 2004, 2006; Gao et al., 2008; Yurchenko et al., 2009). Fig. 6 indicates that recombinant (His)6-tagged OsACBPs preferentially bound [14C]linolenoyl-CoA and that recombinant (His)6-OsACBP6 showed highest affinity with a Kd value of 0.8 μM (Fig. 6 and Table 2). Only (His)6-OsACBP1 could bind [14C]oleoyl-CoA (Kd value of 0.41 μM). (His)6-OsACBP4 displayed good binding affinity to [14C]linoleoyl-CoA (Kd value of 0.36 μM). Except (His)6-tagged OsACBP2 and (His)6-tagged OsACBP3, the other recombinant proteins were able to each bind two or three acyl-CoA esters. The Kd values of (His)6-tagged OsACBPs for various acyl-CoA esters, as calculated from Scatchard plots, are listed in Table 2.

Figure 6.

 Lipidex assays of (His)6-OsACBPs incubated with [14C]palmitoyl-CoA, [14C]oleoyl-CoA, [14C]linoleoyl-CoA or [14C]linolenoyl-CoA. (His)6-OsACBPs (at a final concentration of 0.2 μM) was incubated with [14C]palmitoyl-CoA, [14C]oleoyl-CoA, [14C]linoleoyl-CoA or [14C]linolenoyl-CoA at final acyl-CoA concentrations ranging from 0.2 to 2 μM. The solutions were mixed with Lipidex-1000. Aliquots (200 μl) of the supernatant were taken for analysis of radioactivity. Assays were performed in triplicate, with blanks, at each concentration of acyl-CoA. Bars represent SE (= 3). (a) [14C]palmitoyl-CoA; (b) [14C]oleoyl-CoA; (c) [14C]linoleoyl-CoA; (d) [14C]linolenoyl-CoA.

Table 2. Dissociation constant (KdμM) values for different acyl-CoA esters
Recombinant protein16:018:018:118:2
  1. 16:0, [14C]palmitoyl-CoA; 18:0, [14C]oleoyl-CoA.

  2. 18:1, [14C]linoleoyl-CoA; 18:2, [14C]linolenoyl-CoA.



The ACBP family is conserved in plants

Photosynthetic organisms diversified from fungi and animals approx. 1500 million yr ago (Yoon et al., 2004). Despite such an ancient divergence age, prototype (10 kDa) ACBP shows conservation across various species (Burton et al., 2005). The high evolutionary conservation and ubiquitous expression of prototype ACBPs suggest its potential involvement in basal cellular functions (Færgeman et al., 2007). Previous investigations (including those on phylogeny) focused on the prototype ACBP from animals and yeast (Burton et al., 2005), and reports on the phylogeny of multidomain ACBPs are lacking.

In this study, the phylogenetic relationships among different classes of ACBPs within the plant kingdom indicate that Class I small ACBPs may be ancestral. Although green algae (C. reinhardtii and O. lucimarinus) have Class I, III and IV ACBPs, the respective clades lack bootstrap support (and in the case of C. reinhardtii Class III is not clustered within the main Class III clade), suggesting that the green algae ACBPs are only distantly related to those of the other plant taxa. Both Class I small ACBPs and larger multidomain ACBPs (Class IV Kelch-ACBPs) are conserved from unicellular algae to flowering plants. Previous phylogenetic analysis of plant prototype ACBPs suggest that multiple paralogues were derived from lineage-specific duplication events following the divergence of different plant species (Burton et al., 2005). In this study, prototype ACBPs within a species failed to group into one common branch because duplicated genes must have diverged considerably such that nearly no sequence similarity could be identified (Zhang, 2003).

Class II (ANK-ACBPs) is another multidomain ACBP. Ankyrin repeats have been reported to mediate protein–protein interactions in combination with other functional modules in many multidomain plant proteins. In Arabidopsis, partner modules of ankyrin repeats include the ACB domain, zinc finger, potassium channel proteins and the ATPase-associated domain (Zhang et al., 1992; Sedgwick & Smerdon, 1999; Becerra et al., 2004). AtACBP2 has been demonstrated to interact with an ethylene-responsive element-binding protein (AtEBP), heavy-metal-binding farnesylated protein (AtFP6) and lysophospholipase 2 via its ankyrin repeats (Li & Chye, 2004; Gao et al., 2008, 2010). The obvious absence of Class II in green algae (O. lucimarinus and C. reinhardtii) suggests that Class II evolved as a new protein arising from domain combination (Yang & Bourne, 2009). As the protein domain is an independent evolutionary unit, duplication and recombination of domains are widespread in genome evolution (Vogel et al., 2004; Yang & Bourne, 2009). The number of domain combinations found in eukaryotic protein sets in animals, fungi, land plants and the amoeba Dictyostelium discoideum is large (Itoh et al., 2007). Ancestral domains, especially in plants, fungi and protists, are often used as partner domains. Combinations arising show varying functions that have evolved from common ancestors (Itoh et al., 2007). How-ever, there was inadequate bootstrap support to determine the precise origin of Class II. One possibility is that the amino acid sequence of the Class II ACB domain region had diverged considerably from its putative ancestor, and thus phylogenetic signals were lost.

Our previous study on the eudicot Arabidopsis showed that Class II is highly expressed in seeds and is critical to seed development (Chen et al., 2010). However, in this study, the expression of monocot rice Class II did not mimic its Arabidopsis homologues, suggesting that the function of Class II diverged with the split between monocots and eudicots. Interestingly, rice Class III is highly expressed throughout seed development, but this was not the case for Arabidopsis Class III (Xiao et al., 2010), indicating that the function of Class III had also diverged following this split. There seems to be functional complementarity between Class II and Class III, which correlates well with their phylogenetic positions. Hence, our phylogenetic analysis helps us better understand the development, mechanism and function of the plant ACBP family.

Comparison of rice and Arabidopsis ACBPs in binding acyl-CoA esters

As an ancient functional domain, c. 30% of residues among the four classes are conserved in the ACB domain and six residues are essential in acyl-CoA-binding (Kragelund et al., 1999). However, other residues conserved within each group vary among the four classes. Possible reasons for such divergence would include a changing environment and the consequence of domain combination that affects individual domain sequence, structure and function (Zhang, 2003; Yang & Bourne, 2009). Variation in the ACB domain may have culminated in a variety of acyl-CoA-binding preferences among classes. In Arabidopsis, variations in the acyl-CoA-binding abilities of ACBPs seem to be correlated with diversified biochemical functions (Leung et al., 2004, 2006; Li & Chye, 2004; Chen et al., 2008; Gao et al., 2008, 2010; Li et al., 2008; Xiao et al., 2008a,b, 2010). In comparison with the Arabidopsis ACBP family, OsACBP1 (homologue of the Class I prototype 10-kDa ACBP) showed a similar preference for oleoyl-CoA esters. Class II homologues, OsACBP4 and AtACBP2, both displayed binding to linoleoyl-CoA and linolenoyl-CoA. However, Class IV homologue OsACBP6 did not resemble AtACBP4 and AtACBP5 in binding oleoyl-CoA. Differences also exist in Class III between OsACBP5 and AtACBP3. Recombinant AtACBP3 binds linoleoyl-CoA and linolenoyl-CoA, while recombinant OsACBP5 could only bind linolenoyl-CoA. Hence, this study reveals that although the general function of OsACBPs in binding acyl-CoA esters are retained, the preferred choice of acyl-CoAs bound need not be identical in each Arabidopsis homologue.

Response of the rice ACBP family to biotic and abiotic stresses

Given that the divergence of Class I, II, III and IV lineages is correlated with the divergence of gymnosperms and angiosperms, and that only ankyrin repeats or kelch motifs comprise the additional domains in ACBPs, we suggest that the functions of each subgroup are essential and distinct. According to theoretical population genetics, most new genes are generated by gene duplication and subfunctionalization can stabilize the maintenance of duplicated genes (Zhang, 2003). The existence plus the divergent expression of paralogues within each ACBP class suggest that subfunctionalization has been achieved. These findings are consistent with studies on the Arabidopsis ACBP family which reveal that each class seems to have unique biological functions (Xiao & Chye, 2009). Members of Arabidopsis ACBPs are involved in different pathways in response to biotic and abiotic stresses. The ability to bind linoleoyl-CoA and linolenoyl-CoA, the precursors of phospholipid repair following heavy metal stress, has perhaps resulted in AtACBP1- or AtACBP2-overexpressors being conferred enhanced Pb(II) and Cd(II) tolerance (Gao et al., 2008; Xiao et al., 2008a). Furthermore, recombinant AtACBP3 binds PC, PE, linoleoyl-CoA and linolenoyl-CoA, and functions in the regulation of leaf senescence by modulating membrane phospholipid metabolism (Xiao et al., 2010). The potential interaction of AtACBP4 and its partner AtEBP in defense responses is supported by observations that they are both induced by the ethylene precursor (1-aminocyclopropane-1-carboxylic acid), MeJA and Botrytis cinerea infection (Li et al., 2008). The overexpression of AtACBP6 and depletion of AtACBP1 affect phospholipid metabolism and alter cold tolerance (Chen et al., 2008; Du et al., 2010).

Given the different phylogenetic positions of eudicots and monocots, we have assumed that the function of the rice ACBP family may not necessarily be identical to that in Arabidopsis. As lipids are essential mediators in response to stresses, including chilling, freezing, wounding, pathogens, drought and salt (Wang, 2004), we investigated the expression profiles of the rice ACBP family under stress treatments (salt, drought, cold, wounding treatment and pathogen infection). In contrast to the cold-inducible expression of Arabidopsis Class I AtACBP6 (Chen et al., 2008), the expression of Class I OsACBP1, OsACBP2 and OsACBP3 were not induced by any form of stress tested in this study. The expression of Class II OsACBP4 and Class III OsACBP5 were both induced by drought and salt treatment, and may have potential use in agriculture. Given the close link between the signaling pathways in drought and high-salinity responses in rice (Rabbani et al., 2003), the induction of OsACBP4 and OsACBP5 mRNA is possibly regulated by the same mechanism. Furthermore, the OsACBP5 transcript was induced by rice blast disease infection and Class IV OsACBP6 by wounding. Our observations on the differential expression patterns between the rice and Arabidopsis classes upon stress treatments suggest that rice ACBPs are not mere duplicates of their Arabidopsis homologues.

The presence of putative cis-elements at the 5′-flanking regions of genes encoding OsACBPs further supports their potential roles in stress. The dehydration responsive element-binding protein (DREB)/C-repeat binding factor (CBF) family recognizes DRE/CRT and is involved in drought or salinity stress responses, while ABRE is recognized by ABA-responsive transcription factors (AREB/ABF) and plays a role in ABA signaling (Hirayama & Shinozaki, 2010). Stress-related cis-elements, including DRE/CRT and ABRE, evident in the 5′-flanking regions of OsACBP4, OsACBP5 and OsACBP6 further indicate that the large rice ACBPs can possibly function in stress tolerance. In view of this, we conclude that ACBP multigene families occur in both eudicots and monocots, and these families have expanded during the evolution of plants. The response of certain members of rice ACBPs to biotic and abiotic stresses indicates they have adopted novel functions during evolution and warrant further investigation.


We thank C. Lo and W.K. Yip (The University of Hong Kong) for provision of rice (O. sativa cv Zhonghua 11) seeds and the LS 6500 liquid scintillation counter, respectively. H.M. Lam (The Chinese University of Hong Kong) kindly provided us with rice blast fungus. We also thank S. Xiao for suggestions and discussion on the manuscript. This work was supported by the University of Hong Kong (CRCG Grant 10400274 and postgraduate studentship awarded to WM).