RICE ACYL-COA-BINDING PROTEIN6 Affects Acyl-CoA Homeostasis and Growth in Rice

Acyl-coenzyme A (CoA) esters are important intermediates in lipid metabolism with regulatory properties. Acyl-CoA-binding proteins bind and transport acyl-CoAs to fulfill these functions. RICE ACYL-COA-BINDING PROTEIN6 (OsACBP6) is currently the only one peroxisome-localized plant ACBP that has been proposed to be involved in β-oxidation in transgenic Arabidopsis. The role of the peroxisomal ACBP (OsACBP6) in rice (Oryza sativa) was investigated. Here, we report on the function of OsACBP6 in rice. The osacbp6 mutant showed diminished growth with reduction in root meristem activity and leaf growth. Acyl-CoA profiling and lipidomic analysis revealed an increase in acyl-CoA content and a slight triacylglycerol accumulation caused by the loss of OsACBP6. Comparative transcriptomic analysis discerned the biological processes arising from the loss of OsACBP6. Reduced response to oxidative stress was represented by a decline in gene expression of a group of peroxidases and peroxidase activities. An elevation in hydrogen peroxide was observed in both roots and shoots/leaves of osacbp6. Taken together, loss of OsACBP6 not only resulted in a disruption of the acyl-CoA homeostasis but also peroxidase-dependent reactive oxygen species (ROS) homeostasis. In contrast, osacbp6-complemented transgenic rice displayed similar phenotype to the wild type rice, supporting a role for OsACBP6 in the maintenance of the acyl-CoA pool and ROS homeostasis. Furthermore, quantification of plant hormones supported the findings observed in the transcriptome and an increase in jasmonic acid level occurred in osacbp6. In summary, OsACBP6 appears to be required for the efficient utilization of acyl-CoAs. Disruption of OsACBP6 compromises growth and led to provoked defense response, suggesting a correlation of enhanced acyl-CoAs content with defense responses.


Background
The primary role of acyl-coenzyme A (CoA) esters is to act as intermediates in lipid synthesis and breakdown (Neess et al. 2015). De novo synthesized fatty acids in the chloroplasts are transferred in the form of C16:0-CoA and C18:1-CoA to the endoplasmic reticulum (ER) for phospholipid and triacylglycerol (TAG) synthesis (Li-Beisson et al. 2013). In seeds following imbibition, TAG is rapidly mobilized to support successful germination and seedling establishment via a series of catabolic pathways (Graham 2008;Li-Beisson et al. 2013). Rapid TAG turnover also occurred in vegetative tissues . First, TAG is hydrolyzed by lipases to free fatty acids, followed by activation to CoA esters (Li-Beisson et al. 2013). Subsequently, fatty acyl-CoAs are imported into the peroxisomes by the peroxisomal ATPbinding cassette (ABC) transporter protein COMATOSE (CTS)/PEROXISOME DEFECTIVE 3 (PED3)/PEROXI-SOMAL ABC TRANSPORTER 1 (PXA1), which possesses thioesterase activity that hydrolyzes fatty acyl-CoAs to free fatty acids and CoA as part of the transport process (De Marcos et al. 2013). Once inside the peroxisome, the formation of long-chain acyl-CoA esters is catalyzed by long-chain acyl-CoA synthetase (Li-Beisson et al. 2013). Following a series of enzymatic β-oxidation reactions by acyl-CoA oxidases, multifunctional proteins, and thiolases, fatty acids are degraded to acetyl-CoA for further use in gluconeogenesis and respiration (Li-Beisson et al. 2013).
The intrinsic nature of acyl-CoA is amphipathic (Neess et al. 2015). Although the cellular concentration of acyl-CoAs ranges from several μM to several hundred μM, the concentration of free unbound acyl-CoAs is unknown or could be extremely low (Neess et al. 2015). Usually, acyl-CoAs are bound to membrane lipids or proteins with high affinities, so that the concentration is strictly controlled to below the critical micelle concentration to avoid detergent effects that disrupt the membrane (Neess et al. 2015). The small (10-kD) acyl-CoA-binding proteins (ACBPs) are known to bind acyl-CoA esters (Knudsen et al. 2000). Each ACBP has a conserved acyl-CoA-binding domain that allows the ACBPs to bind acyl-CoA esters, which can then form a cellular acyl-CoA pool and protect acyl-CoAs against hydrolysis (Du et al. 2016). Considering that ACBPs bind acyl-CoAs with high affinities and broad ligand selectivity (C12-C26), ACBPs are believed to be the predominant acyl-CoA carriers/transporters (Knudsen et al. 2000;Neess et al. 2015;Lung and Chye 2016a, b;Guo et al. 2017).
In plants, ACBPs are highly conserved throughout the evolution of land plants and are divided into four classes (Classes I-IV) based on domain architecture (Meng et al. 2011). The functions of Arabidopsis ACBPs (AtACBPs) in development and stress responses have been extensively investigated (Du et al. 2016;Chye 2016b, 2019). Much evidence indicates that AtACBPs mediate multiple aspects of plant physiology by regulating lipid metabolism in Arabidopsis (Du et al. 2016;Chye 2016b, 2019). There is only a single 10-kD Class I ACBP member in Arabidopsis, namely AtACBP6, and AtACBP6 has been reported to be a cytosolic protein (Chen et al. 2008). AtACBP1 and AtACBP2 are Class II paralogues of each other and have been proposed to function in the ER in acyl-CoA pool formation and acyl-CoA/lipid trafficking between the plasma membrane and the ER during early embryogenesis . Furthermore, the overexpression of AtACBP1 in Arabidopsis resulted in the accumulation of phosphatidic acid (PA) during seed germination and seedling development following abscisic acid (ABA) treatment (Du et al. 2013). Similar PA accumulation in AtACBP1 overexpressors occurred during freezing treatment (Du et al. 2010). The interaction between AtACBP1 and phospholipase Dα1 (PLDα1), a phospholipase that generates PA, played an essential role in both studies (Du et al. 2010;Du et al. 2013). In addition, the interaction between AtACBP1 and STEROL C4-METH YL OXIDASE1 (SMO1) regulated both fatty acid and sterol homeostasis, and in turn modulated the expression of homeodomain-leucine zipper IV transcription factors (TFs) (Lung et al. 2018). The overexpression of another member, Class III AtACBP3, accelerated leaf senescence increasing the acyl-CoA pool and phosphatidylethanolamine (PE) content ). The variation in AtACBP3 level and PE composition between AtACBP3-overexpressors and AtACBP3-knockout mutants indicated that AtACBP3 binds and regulates PE in vivo . Furthermore, the recombinant proteins of the three cytosolic AtACBPs, comprising Class IV members (AtACBP4 and AtACBP5), and Class I AtACBP6, were reported to bind C16:0-CoA, C18:1-CoA, C18:2-CoA, and C18:3-CoA with different affinities (Hsiao et al. 2014a). The double mutants and the triple mutant of these three cytosolic AtACBPs were altered in acyl-CoA composition and displayed high ABA sensitivity during seed germination (Hsiao et al. 2014a). The decrease in oil bodies and decline in germination rate of acbp4acbp5acbp6 pollen suggested important roles for the cytosolic AtACBPs in pollen lipid metabolism (Hsiao et al. 2015).
Thus far, the functions of plant ACBPs have been widely elucidated in the eudicots, such as Arabidopsis and Brassica napus, although there have been many reports on the characterization of ACBP family members from other plant species, such as Oryza sativa, Agave americana, Vernicia fordii, Vitis vinifera, Helianthus annuus, Gossypium hirsutum, Jatropha curcas, and Elaeis guineensis (Du et al. 2016;Raboanatahiry et al. 2018;Liao et al. 2019;Amiruddin et al. 2020). Other than Arabidopsis and Brassica napus, the functions of ACBPs from other plant species have been mostly investigated using transgenic Arabidopsis (Takato et al. 2013;Meng et al. 2014;Panthapulakkal Narayanan et al. 2019). Previously, we have reported a peroxisome-localized rice Class IV ACBP, OsACBP6, which is currently the only plant ACBP identified to be localized at the peroxisome (Meng et al. 2014). The overexpression of OsACBP6 in the Arabidopsis β-oxidation-deficient mutant pxa1 rescued indole-3-butyric acid (IBA) sensitivity and jasmonic acid (JA) production after wound treatment, suggesting that OsACBP6 may be involved in peroxisomal β-oxidation (Meng et al. 2014). In this study to investigate the function of the peroxisomal OsACBP6 in rice, we analyzed a T-DNA insertional mutant of OsACBP6 (osacbp6) that demonstrated its reduction in root meristem activity and leaf growth. The results from comprehensive acyl-CoA and lipid profiling indicated that disruption of OsACBP6 compromised the maintenance of the acyl-CoA pool and altered lipid composition in leaves. By comparative transcriptomic analysis of the shoots/leaves and roots from various growth stages, differentially-expressed genes were identified in osacbp6. Gene ontology (GO) enrichment analysis and functional classification provided insights into the function of OsACBP6 in the oxidative stress response. Accumulation of reactive oxygen species (ROS) and elevation of the JA was evident in osacbp6.

Results
The Rice osacbp6 Mutant Displayed Growth Retardation DNA sequence analysis of the osacbp6 mutant indicated that the T-DNA had inserted into the 11th exon of OsACBP6 as shown in Fig. 1a. The location of the T-DNA insertion in OsACBP6 was confirmed by PCR using a combination of gene-specific primers, ML2066, and ML2067, as well as T-DNA left border primers 2717LB and ML2067 (Fig. 1b). Semiquantitative RT-PCR using gene-specific primers ML1050 and ML1051 confirmed the absence of OsACBP6 expression in the mutant (Fig. 1c). The homozygous osacbp6 line was then used in subsequent experiments.
As OsACBP6 had been proposed to play a role in fatty acid β-oxidation through its acyl-CoA binding ability (Meng et al. 2014), osacbp6 was examined for the typical phenotypes related to defects in β-oxidation. When osacbp6 and the wild type were tested in seed germination and seedling establishment in water, both germinated at about the same time but the development of the roots of osacbp6 seedlings was repressed with primary root reduction around 40% ( Fig. 1d and e). When Fig. 1 Characterization of the rice osacbp6 mutant. a T-DNA insertion in OsACBP6, and primer locations used for PCR genotyping. Black and grey boxes indicate coding and untranslated regions, respectively. b PCR genotyping of osacbp6. Primer combinations ML2066/ML2067 (top gel) and 2717LB/ML2067 (bottom gel) in PCR to identify osacbp6 homozygous mutant. c Semi-quantitative RT-PCR showing the knockout of OsACBP6 in osacbp6. ACTIN1 served as a control. d Growth of wild type (WT, Oryza sativa var. japonica cv. Dongjin) and osacbp6. Images from left to right represent 5-, 7-day-old seedlings grown in water, and 21-day-old soil-grown seedling. Bars, 1 cm. e The primary root length (left) and shoot length (right) comparison between WT and osacbp6. Primary root reduction of osacbp6 is shown in the middle. Values are mean ± SD (n = 25). f Comparison of the leaf length of 21-day-old soil-grown seedlings between WT and osacbp6. Values are mean ± SD (n = 5). Asterisks indicate significant differences between WT and osacbp6 as evaluated by Student's t-tests: ***P < 0.001 osacbp6 and the wild type were grown in soil, osacbp6 root development was retarded (around 30% primary root reduction), but not as severe as previously observed in water (Fig. S1). In both situations, there was no obvious difference in the length of the shoots between osacbp6 and the wild type at the early stage of the seedling development (Fig. 1e, Fig. S1b). However, the leaf length of 21day-old osacbp6 was significantly shorter than the wild type with a reduction of around 25% ( Fig. 1d and f). Furthermore, osacbp6 grains were reduced in length but remained unchanged in width in comparison to the wild type (Fig. S2a). Lastly, given that the product of IBA β-oxidation is indole-3-acetic acid (IAA) which has growth inhibition effect, the IBA sensitivity of osacbp6 was examined to investigate whether β-oxidation process was blocked in osacbp6. Reduced primary root elongation of osacbp6 after IBA treatment indicated that osacbp6 was sensitive to growth inhibition by IBA (Fig. S2b), suggesting the IBA β-oxidation was not blocked in osacbp6. Taken together, from phenotypic observations it appears that βoxidation was not severely affected due to the loss in OsACBP6.
When the osacbp6 mutant was complemented, the three independent complemented lines (designated as osacbp6-C1, osacbp6-C2, and osacbp6-C4) displayed normal growth similar to the wild type, including the primary root length, 21-day-old leaf length, and grain length (Fig. 2, Fig. S3). Moreover, the phenotype of seedlings from hemizygous complemented line osacbp6-C4 were examined. In contrast with wild type normal growth of primary root and leaves, seedlings from hemizygous complemented line contained both wild type growth (168) and retarded growth (51) of roots and leaves with a 3:1 segregation ratio (χ 2 = 0.34). These findings confirmed that growth arrest was caused by the loss of OsACBP6.
In addition, expression information from the Rice eFP Browser (http://bar.utoronto.ca/efprice/cgi-bin/efpWeb. cgi) showed that OsACBP6 is highly expressed at seedling roots, young leaf and mature leaf (Fig. S4), suggesting OsACBP6 may have important roles in these tissues. Therefore, the root meristem activity was further investigated, 5-ethynyl-2′-deoxyuridine (EdU) staining indicated that reduced EdU labeling was observed in the root tip of osacbp6 but not in the complemented line ( Fig. 3a-d), indicating that cell proliferation in the root meristem affected by the loss of OsACBP6 could be complemented by the OsACBP6 ORF. On the other hand, observation of leaf growth showed that the emergence of the third leaf of osacbp6 was one day later than the wild type. When the third leaf of osacbp6 emerged from the leaf sheath, the third leaf length of the wild type and the complemented line was about 6 cm longer than osacbp6 (Fig. 3e). Meanwhile, the leaf elongation rate of osacbp6 was significantly lower than the wild type and the complemented line (Fig. 3f). These results indicated both root and the leaf development were affected in the osacbp6 mutant and function was restored upon complementation.

Changes in Acyl-CoA and Lipid Profiles in osacbp6
To gain an insight into the relationship between the loss of OsACBP6 and compromised growth, the acyl-CoA profiles in osacbp6 was determined. The absence of OsACBP6 function is expected to disrupt fatty acid βoxidation and result in the accumulation of long-chain acyl-CoAs. To test this hypothesis, a comparison was made amongst the acyl-CoA content in 21-day-old leaves of osacbp6, the wild type and the complemented line. As revealed in acyl-CoA measurements, the content of acyl-CoAs in osacbp6 was significantly higher than the wild type and complemented line (osacbp6-C4) (Fig. 4a). Amongst all tested acyl-CoA esters, C18:3-CoA was notably enriched in the leaves of 21-day-old osacbp6 in contrast to the wild type and osacbp6-C4 (Fig. 4b).
These findings indicate that loss of OsACBP6 had likely compromised the maintenance of an acyl-CoA pool, particularly in the utilization of C18:3-CoA.

Comparative Transcriptomic Analysis between osacbp6 and the Wild Type
To further analyze the potential role of OsACBP6 in growth regulation, a comparative transcriptomic analysis Values are means ± SD (n = 10). c Primary root length of WT, osacbp6, C1, C2, and C4 (upper panel) and primary root reduction of osacbp6 relative to WT, C1, C2, and C4 (lower panel). Values are means ± SD (n = 10). Asterisks indicate significant differences as evaluated by Student's t-tests: ***P < 0.001. WT, Oryza sativa var. japonica cv. Dongjin was performed between osacbp6 and the wild type at different growth stages (Dataset S2, S3, S4, S5, S6, S7). Differentially expressed genes (DEGs) in osacbp6 were associated with primary metabolic processes, including the light reaction, cell wall and lipid metabolism; and secondary metabolic processes related to terpenes, flavonoids, and phenylpropanoids. In osacbp6, genes involved in fatty acid synthesis and elongation, phospholipid synthesis, and lipid degradation appeared to be affected (Fig. S6). However, the key enzymes in β-oxidation, such as acyl-CoA oxidase, multifunctional protein, and 3ketoacyl-CoA thiolase were not differentially expressed in osacbp6 (Fig. S6). Furthermore, most of the DEGs were responsive to biotic stress (Fig. S7a). Compared with the wild type, various categories related to biotic stress were differentially expressed in osacbp6, including plant hormone signaling, PR proteins, cell wall, proteolysis, signaling, TFs, heat shock proteins, secondary metabolites. DEGs associated with defense hormone (ABA, JA, salicylic acid (SA), and ethylene) and growth hormone (auxin, gibberellin (GA), brassinosteroid (BR) and cytokinin) (Fig. S7b) and differentially expressed TFs ( Fig. S7c) are displayed in Dataset S8 and S9, respectively. Amongst differentially expressed TFs, basic helixloop-helix (bHLH) was strongly activated in the young seedlings, especially in 5-day-old roots ( Fig. S7c). In 21day-old leaves, WRKY was more activated than in the other samples (Fig. S7c).
Significantly enriched Gene Ontology (GO) categories were illustrated in Dataset S2. The GO term oxidation reduction was the most enriched in young seedlings and was common to all samples, except for the lack of a significantly enriched GO term in 7-dayold shoots. Another significantly enriched category was a response to oxidative stress that was only evident in the downregulated DEGs, and solely consisted of putative peroxidase precursor genes (class III peroxidase). Gene expression of peroxidases was verified by quantitative real-time RT-PCR (qRT-PCR). In comparison to the wild type, eight peroxidases were downregulated in 7-day-old osacbp6 roots and upregulated in the complemented line (Fig. 5a). In 7day-old shoots and 21-day-old leaves, the number of differentially expressed peroxidase was lower than 7- Values are means ± SD (n = 10). e and f Measurement of the length of the third leaf (e) and leaf elongation rate (f). Values are means ± SD (n = 5). Asterisks indicate significant differences as evaluated by Student's t-tests: **P < 0.01, ***P < 0.001. DAPI, 4′,6-Diamidino-2-Phenylindole; WT, Oryza sativa var. japonica cv. Dongjin day-old roots. In particular, LOC_Os03g13180 and LOC_Os04g59190 had decreased in osacbp6 in contrast to an increase in the complemented line in all test samples (Fig. 5a). In addition, expression information from the Rice eFP Browser (http://bar.utoronto. ca/efprice/cgi-bin/efpWeb.cgi) indicated that most of these eight peroxidases are highly expressed in roots (Fig. S8), suggesting probable significant roles in roots. Furthermore, the peroxidase activities were compared amongst osacbp6, the wild type, and the complemented line. In comparison with the wild type, the peroxidase activities in osacbp6 decreased 25% in 7-day-old roots, 17% in 7-day-old shoots, and 17% in 21-day-old leaves (Fig. 5b). As expected, the level of peroxidase activities in the complemented line remained similar to the wild type with the exception of an increase in 7-day-old shoots (Fig. 5b). Given that peroxidases are bifunctional enzymes that can oxidize substrates at the expense of H 2 O 2 but also produce ROS (Passardi et al. 2004), an abundant decline in the expression of peroxidases and reduced peroxidase activities suggest that ROS homeostasis in osacbp6 is perturbed. Taken together, a loss in OsACBP6 had resulted in transcriptional changes in the expression of genes related to redox, H 2 O 2 scavenging, and defense responses. Values are means ± SD (n = 3). H, value of osacbp6 is higher than WT; L, value of osacbp6 is lower than WT. Values of C4 are all significantly lower than osacbp6 as evaluated by Student's t-tests. Asterisks indicate significant differences as evaluated by Student's t-tests: *P < 0.05, **P < 0.01, ***P < 0.001. FW, fresh weight; WT, Oryza sativa var. japonica cv. Dongjin. DAG, Diacylglycerol; TAG, Triacylglycerol; MGDG, Monogalactosyldiacylglycerol; DGDG, Digalactosyldiacylglycerol; PC, Phosphatidylcholine; PG, Phosphatidylglycerol; PA, Phosphatidic acid; PI, Phosphatidylinositol; PE, Phosphatidylethanolamine; PS, Phosphatidylserine

ROS Homeostasis Is Perturbed in osacbp6
To examine whether the ROS levels were altered in osacbp6, superoxide anion (O 2 •-) and H 2 O 2 contents were examined. The blue and brown coloration in 21day-old osacbp6 leaf blades indicated the accumulation of O 2 •and H 2 O 2 , respectively (Fig. 5c). Furthermore, H 2 O 2 was quantified in osacbp6, the wild type, and the complemented line (Fig. 5d). The H 2 O 2 levels in the root and shoot of 7-day-old osacbp6 seedling and the leaves of 21-day-old osacbp6 were significantly higher than the wild type and complemented line. Meanwhile, there was no significant difference between the wild type and complemented line suggesting the restoration of OsACBP6 function in the latter.
As OsACBP6 was localized at the peroxisomes (Meng et al. 2014), and H 2 O 2 overaccumulated in osacbp6, peroxisomal ROS scavenging was next explored. Given that plant peroxisomes employ two major enzymatic antioxidant systems, catalase (CAT) and ascorbate peroxidase (APX)/ monodehydroascorbate reductase (MDAR) system, to detoxify the H 2 O 2 generated by β-oxidation (Eastmond 2007), the expression of APX, MDAR, and CAT was examined in osacbp6, the wild type and complemented line by qRT-PCR (Fig. 6). In comparison to the wild type, there was no significant difference in the expression of the eight APX genes in osacbp6, including peroxisomal localized OsAPX3 and OsAPX4 (Kaur and Hu 2011). The expression of five MDAR genes was examined. Except LOC_Os08g05570, the rest four MDAR genes are all peroxisomal localized (Kaur and Hu 2011). Two of the five MDAR genes was enhanced in osacbp6. In particular, LOC_Os08g44340 was induced in osacbp6 but repressed in the complemented line. Also, the expression of one of the three CAT genes was slightly elevated. Hence, despite H 2 O 2 accumulation in osacbp6, the peroxisomal ROS scavenging system did not seem to be activated. Moreover, measurements of APX, MDAR and CAT activities further support that there was no significant difference in peroxisomal ROS scavenging system between osacbp6 and the wild type (Fig. 6b). Above findings suggested that the distribution of H 2 O 2 may not be confined to the peroxisome.

Intrinsic JA Accumulated in osacbp6
Arising from an observation in differentially expressed plant hormone-related genes, JA biosynthesis was investigated next to address if it was compromised in osacbp6. Because JA is produced through β-oxidation in the peroxisomes (Hu et al. 2012) and the overexpression of OsACBP6 in the Arabidopsis pxa1 mutant had rescued JA production after wound treatment (Meng et al. 2014). The intrinsic JA level was significantly elevated in 21-day-old osacbp6 leaves (5.64 ng g − 1 ), five times higher than the wild type (1.08 ng g − 1 ) (Fig. 7a). Other plant hormones (SA, IAA and GA 3 ) were also enhanced in osacbp6 over the wild type (Fig. 7a). Therefore, measurements of hormone levels indicated that JA and SA production was not blocked in osacbp6. Subsequently, the expression of genes related to JA biosynthesis and the JA signaling pathway in osacbp6 were examined. JA biosynthesis includes the release of α-linolenic acid (α-LeA) from the lipid membrane by phospholipases, oxygenation of α-LeA by lipoxygenase or α-dioxygenases (DOXs) followed by steps catalyzed by allene oxide synthase (AOS) and allene oxide cyclase (AOC) to yield the JA precursor, cis-12-oxo-phytodienoic acid (OPDA). OPDA is then reduced by OPDA reductase (OPR) and subjected to β-oxidation to generate JA (Wasternack and Feussner 2018). When the expression of phospholipases D (OsPLDα4) and lipase EXTRA GLUME 1 (EG1) that function in releasing α-LeA (Qi et al. 2011;Cai et al. 2014), OsDOX, OsAOS2, OsAOC, OsOPR3 (Koeduka et al. 2005;Mei et al. 2006;Guo et al. 2014) and transcription factor OsMYC2 that positively regulates the JA-responsive genes (Uji et al. 2016), were . Asterisks indicate significant differences as evaluated by Student's ttests: *P < 0.05, **P < 0.01, ***P < 0.001. APX, ascorbate peroxidase; MDAR, monodehydroascorbate reductase; CAT, catalase; WT, Oryza sativa var. japonica cv. Dongjin examined, only OsDOX and OsAOS2 expression increased in osacbp6, while others remained a similar level to the wild type (Fig. 7b), consistent with transcriptomic data. Given that OsDOX is activated by H 2 O 2 (Zhu et al. 2013), elevated JA level in osacbp6 would be a consequence of repressed H 2 O 2 scavenging in osacbp6.

Discussion
The level and availability of acyl-CoA is strictly controlled by proteins with binding abilities to meet the cellular requirement (Neess et al. 2015). Acyl-CoAs are sequestered and directed toward specific metabolic pathways (Neess et al. 2015). It has been proposed that OsACBP6 is implicated in the formation and regulation of the peroxisomal acyl-CoA pool and donation of acyl-CoA for β-oxidation (Meng et al. 2014). This study revealed that the disruption of peroxisomal matrixlocalized OsACBP6 in rice resulted in growth retardation in both roots and leaves as well as smaller grain size, resembling the phenotype of Arabidopsis and barley β-oxidation dysfunction mutants deficient in the peroxisomal ABC transporter (Zolman et al. 2001;Mendiondo et al. 2014), although the IBA sensitivity remained in osacbp6. The above findings suggest that osacbp6 was not severely affected in β-oxidation function. However, increased acyl-CoA levels (specifically C18:3-CoA) and some TAG molecular species content observed in osacbp6 over the wild type suggest it bears similarity to the Arabidopsis mutants defective in acyl-CoA import (peroxisomal ABC transporter pxa1/cts/ped3) and activation (peroxisomal long-chain acyl-CoA synthetase lac-s6lacs7) and core enzymes of fatty acid β-oxidation (mfp2, kat2) (Germain et al. 2001;Fulda et al. 2004;Rylott et al. 2006;Graham 2008). The functions of plant peroxisomal ABC transporters are highly conserved between eudicot Arabidopsis and monocot barley (Zolman Fig. 7 Alteration of plant hormone levels in the rice osacbp6 mutant. a Measurement of plant hormone content in 21-day-old leaves. b Relative expression of genes related to JA biosynthesis and its signaling pathway. Values are means ± SD (n = 3). Asterisks indicate significant differences between the wild type (WT) and osacbp6 as evaluated by Student's t-tests: *P < 0.05, **P < 0.01, ***P < 0.001. WT, Oryza sativa var. japonica cv. Dongjin. IAA, indole-3-acetic acid; ABA, abscisic acid; SA, salicylic acid; JA, jasmonic acid; GA, gibberellin; PLD, phospholipase D; EG1, EXTRA GLUME 1; DOX, α-dioxygenases; AOS, allene oxide synthase; AOC, allene oxide cyclase; OPR, OPDA reductase et al. 2001; Mendiondo et al. 2014). Although the role of the rice peroxisomal ABC transporter is yet unclear, two orthologues of Arabidopsis PXA1 co-exists (Kaur and Hu 2011). Localization of OsACBP6 in the peroxisomal matrix suggests that OsACBP6 would not be able to transfer acyl-CoAs directly from the cytosol into the peroxisome, although OsACBP6 does have a putative N-terminal transmembrane domain. Taken together with findings that recombinant OsACBP6 binds C18:2-CoA and C18:3-CoA with high binding affinities in vitro (Meng et al. 2011), the increased acyl-CoA levels in osacbp6 and their corresponding decline in the complemented line, support a role for OsACBP6 in the peroxisome most likely associated in the maintenance of a peroxisomal acyl-CoA pool, especially for C18:3-CoA, and to facilitate efficient utilization of acyl-CoAs imported by PXA1 orthologues for subsequent β-oxidation. Loss of OsACBP6 might disrupt the utilization of C18:3-CoA in the following β-oxidation and induce a feedback inhibition to the import of C18:3-CoA into the peroxisomes which might lead to the accumulation of C18:3-CoA in the cytosol. On the other hand, DEGrelated to fatty acids synthesis, phospholipids biosynthesis, and TAG biosynthesis were limited in osacbp6, suggesting loss of OsACBP6 may not increase TAG content. However, it has been proposed that acyl-CoA utilization has a feedback influence on the rate of TAG lipolysis, and the mechanism of acyl-CoA-mediated inhibition of TAG breakdown is not clear (Graham 2008;Quettier and Eastmond 2009). Therefore, in osacbp6, the relationship between acyl-CoA accumulation and an increase in some TAG molecular species content awaits to be further elucidated.
Arabidopsis mutants defective in β-oxidation-related genes commonly display failure in early seedling establishment due to a block in TAG turnover (Graham 2008). The phenotypic difference between the wild type and these mutants is indistinguishable at the vegetative phase but some of them show abnormal inflorescence later (Richmond and Bleecker 1999;Fulda et al. 2004;Rylott et al. 2006;Wiszniewski et al. 2014). In contrast, as starch in the endosperm is the major energy supply for monocots post-germinative growth, the seedling establishment does not seem to be affected by a deficiency in fatty acid β-oxidation (Mendiondo et al. 2014;Xu et al. 2017). However, distinguishable phenotypes emerge at the vegetative phase. A rice aim1 mutant appeared as a short root mutant with reduced meristem activity due to blocked SA biosynthesis and reduced ROS level . By repressing the expression of redox and ROS-scavenging genes and maintaining a proper level of ROS, SA maintains root meristem activity . In this study, the root meristem activity of osacbp6 also declined, but unlike aim1, its H 2 O 2 level in root was elevated. It has been demonstrated that Arabidopsis bHLH TF UPBEAT1 (UPB1) regulates the balance of root tip cell proliferation and differentiation by controlling the expression of a set of peroxidases and therefore the balance of O 2 •and H 2 O 2 (Tsukagoshi et al. 2010). Ectopic expression of UPB1 represses the peroxidases and results in H 2 O 2 accumulation and decreased meristem size in the root tip (Tsukagoshi et al. 2010). In osacbp6 root, reduction in root meristem activity, repression in gene expression of a set of peroxidases, decrease in peroxidase activities, and accumulation of H 2 O 2 occurred, which were restored in the complemented lines. These findings are partially consistent with ROS-controlled root growth regulated by UPB1 (Tsukagoshi et al. 2010). Unfortunately, as a UPB1 orthologue has not been identified in the rice genome using a BLAST search in NCBI (National Center for Biotechnology Information, http://blast.ncbi.nlm.nih.gov/Blast.cgi), the TF responsible for such regulation remains elusive.
The regulatory mechanisms for the growth of leaf and root are known to be different (Lu et al. 2014). In Arabidopsis, leaf growth and final size are controlled by MYBlike TF KUODA1 (KUA1)-regulated peroxidase activities which in turn maintain ROS homeostasis (Lu et al. 2014). Although leaf and root growth are controlled by different TFs, the direct targets of these TFs are peroxidases and by regulation in the expression of peroxidases, ROS homeostasis is maintained (Tsukagoshi et al. 2010;Lu et al. 2014). In osacbp6, compromised leaf development was represented by slow elongation and accompanied by a high H 2 O 2 level. Consistent with root, leaf peroxidases of osacbp6 were also differentially expressed and activities were reduced. In addition, given that O 2

•-
can be converted to H 2 O 2 spontaneously or by enzymatic activities (Apel and Hirt 2004), the elevated level of O 2 •in osacbp6 also might be an alternative source of H 2 O 2 accumulation. Due to the complexity of peroxidases in promoting or restricting plant growth (Passardi et al. 2004), the regulation mechanism of the reduced peroxidase activities and accumulation of ROS in retarded leaf development of osacbp6 were elusive. They might contribute to abnormal cell proliferation or differentiation and ultimately reduce leaf development. Taken together, from the findings on root and leaf, retarded growth in osacbp6 mostly possible resulted from a disruption of ROS homeostasis. On the other hand, as a higher level of H 2 O 2 occurred throughout osacbp6 development, it appears that stress response is constitutively activated in osacbp6. Diminished growth of osacbp6 is consistent with studies on growth-defense tradeoffs, in which plant growth is inhibited by the activation of defense responses (Guo et al. 2018). Meanwhile, the examination of various plant hormones further supports this hypothesis. Alteration in the plant hormone levels was noted in osacbp6, especially an increase in JA content. The core jasmonate signaling pathway functions in the perception of various developmental and stress-related cues and mediation of transcriptional responses . One of the output responses of the jasmonate signaling pathway is developmental responses, including growth inhibition, defensive structures, and effects on reproduction and fertility . As the first step of JA biosynthesis, plants employ different lipases to liberate α-LeA from chloroplast membrane lipids, including DONGLE (DGL), DEFECTIVE IN ANTHER DEHISCENCE1 (DAD1), and phospholipase A (PLA) PLA-Iγ1 in Arabidopsis, PLDα4 and α5 and EG1 in rice (Ellinger et al. 2010;Qi et al. 2011;Cai et al. 2014). In contrast to an induction triggered in JA biosynthesis (Qi et al. 2011), PLDα4 and α5 showed no change in osacbp6. Despite that EG1 is mainly required for development in rice (Cai et al. 2014), there was no change for EG1 in osacbp6. Limited alteration in expression of these lipases is consistent with lack in an obvious loss of chloroplast membrane lipids in comparison to the wild type, suggesting α-LeA in osacbp6 is sufficient to produce enough JA to confer the balance between growth and defense. Amongst enzymes involved in JA biosynthesis, only DOX and AOS were induced in osacbp6. Given that DOX is activated by H 2 O 2 (Zhu et al. 2013), enhanced of JA biosynthesis in osacbp6 under normal growth conditions appears to be a perception of the elevated H 2 O 2 level and in turn adjusted the corresponding transcriptional responses to diminish growth. Elevated JA level in osacbp6 indicated that JA production from the β-oxidation of JA precursor, OPDA, was not blocked due to the loss of OsACBP6. As we discussed above, OsACBP6 would not be responsible for direct import of substrate into the peroxisomes. Therefore, in combination with the changes in the expression of genes related to JA biosynthesis, elevated JA level is most likely resulted from the secondary effects of the loss of OsACBP6.
Acyl-CoA esters are not only important intermediates in lipid metabolism but also play a role in regulating multiple cellular processes (Schmidt et al. 2018;Lung and Chye 2019). Recently, direct evidence showed that acyl-CoA esters function in the cellular signaling pathway in Arabidopsis (Schmidt et al. 2018). An increased C18:1-CoA level promoted the dissociation of TF RELA TED TO APETALA 2.12 (RAP2.12) with its partner AtACBP1, which determined subsequent hypoxiaspecific gene expression (Schmidt et al. 2018). Meanwhile, C16:0-CoA, C18:0-CoA, and C18:1-CoA triggered distinct transcriptional responses (Schmidt et al. 2018). Our findings that elevated acyl-CoA content in osacbp6 was accompanied by alterations in lipid composition as well as ROS and plant hormone levels further support the critical roles of acyl-CoA esters in the signaling network in both dicots and monocots. Although OsACBP6 has two cytosolic orthologues (AtACBP4 and AtACBP5) in Arabidopsis, a mere single mutation in any made them indistinguishable from the wild type (Hsiao et al. 2015). Even in the acbp4acbp5 double mutant, the phenotypic discrepancy was limited because another cytosolic AtACBP6 co-exists (Hsiao et al. 2015). Based on our results, a mutation in OsACBP6 led to phenotypic and physiological changes in rice. Furthermore, the correlation of increased acyl-CoA pool size with the loss of OsACBP6 also suggests a possible role of overexpression of OsACBP6 in the wild type in facilitating acyl-CoAs utilization.

Conclusions
In conclusion, acyl-CoA profiling in this study revealed the role of ACBPs in acyl-CoA utilization is conserved in both Arabidopsis and rice. However, it is worth mentioning that growth retardation in osacbp6 was associated with a series of changes including changes in acyl-CoA pool size, lipid metabolism, ROS content, and plant hormone levels as well as other transcriptional response, indicating the correlation of enhanced acyl-CoAs content with defense responses.

Plant Materials and Growth Conditions
T-DNA line PFG_1B-14,906 (designated as osacbp6) was obtained from Postech PFG T-DNA, South Korea (Ryu et al. 2004). To prepare the samples for transcriptomic analysis, seeds of wild-type rice (Oryza sativa var. japonica cv. Dongjin) and osacbp6 were soaked in water and germinated in dark at 25°C for 2 days. To collect the roots and shoots, germinated seeds grown in water were placed in the growth chamber under a 10-h-light (30°C) / 14-h-dark (28°C) cycle. Roots and shoots were sampled separately at 5 and 7 days after imbibition. To collect the leaf blade, germinated seeds were grown in soil under the same growth condition as described above. The leaf blades were collected at 21 days after imbibition. For IBA treatment, germinated seeds were hydroponically cultured in Hoagland solution supplemented with 10 mg L − 1 IBA for 7 days. Cell proliferation in the root meristem was visualized by the incorporation of EdU into DNA at S-phase (Kotogany et al. 2010). EdU staining was performed using an EdU kit (C10310, Apollo 488, Ribobio) according to Li et al. (2015). Briefly, roots of 5-day-old rice seedlings (10 seedlings for each genotype) were immersed in a 50 μM EdU solution for 2 h. After fixation in 4% paraformaldehyde for 30 min, roots were incubated with freshly-prepared Apollo solution for 30 min, followed by 4′,6-Diamidino-2-Phenylindole (DAPI, 1 mg mL − 1 ) staining for 10 min and fluorescence microscopy. Images were analyzed using ImageJ software (Schneider et al. 2012). In order to observe leaf elongation, the length of the third leaf was measured daily with a ruler using the top edge of the pot as a reference (Pettko-Szandtner et al. 2015). Measurements were carried out on five plants starting from the emergence of the third leaf for 6 days.

Generation of osacbp6-Complemented Transgenic Rice
An OsACBP6 putative promoter region containing 2.0kb upstream of the transcription start site was amplified using PCR with primer pair Promoter-F (5′-CAGTCGTCTCACAACTGCCACTGTAAATTTGT CTA-3′ with the BsmBI underlined) and Promoter-R (5′-CAGTCGTCTCAGCTTCGCGGGCGATGAATGA-3′ with BsmBI underlined). The 1.97-kb open reading frame (ORF) of OsACBP6 was amplified by RT-PCR using primer pair ORF-F (5′-CAGTCGTCTCAAAGCA TGGCGAGCTCCGGACTCGC-3′ with BsmBI underlined) and ORF-R (5′-CAGTCGTCTCATACATCAAG ACTCGGACTTATCAG-3′ with BsmBI underlined). To construct the OsACBP6 promoter driving OsACBP6 cDNA plasmid (pOsACBP6::OsACBP6) for complementation, the PCR products of promoter and ORF were digested by BsmBI, and the binary vector pBWA(V)Hccdb-Tnos was digested by Eco31I. The promoter and the ORF regions were ligated to the vector pBWA(V)Hccdb-Tnos using T4 ligase. The pOsACBP6::OsACBP6 construct was then introduced into osacbp6 by Agrobacterium-mediated transformation. Positive transgenic rice lines were confirmed by PCR and hygromycin screening, and the T3 generation was used to observe changes in the phenotype.

Acyl-CoA and Lipid Profiling
For acyl-CoA measurement, acyl-CoA esters were extracted from freshly harvested leaves from at least five individual 21-day-old (100 mg, three replicates) soil-grown plants according to Woldegiorgis et al. (1985) and Larson and Graham (2001). For lipid profiling, leaves from at least five individual 21-day-old soil-grown plants were freshly collected for analysis (60 mg, three replicates). Plant materials were inactivated with hot isopropanol, and total lipid extraction was performed as previously described (Welti et al. 2002).

Transcriptomic Analysis and DEG Identification
Total RNA from 5-day-old roots, 5-day-old shoots, 7day-old roots, 7-day-old shoots and 21-day-old leaves, respectively, was extracted using the RNeasy Plant Mini Kit (Qiagen). Transcriptome analysis was conducted at Annoroad Gene Technology Corporation (Beijing, China). The data from three replicates were collected and analyzed. Genes whose expression is increased by over 100% or decreased by over 50% with adjusted Pvalue ≤0.05 were considered as DEGs.

Quantitative Real-Time RT-PCR
Total RNA from 7-day-old roots, 7-day-old shoots and 21-day-old leaves, respectively, was extracted using the RNeasy Plant Mini Kit (Qiagen) and reverse transcribed by ReverTra Ace® qPCR RT Master Mix with gDNA Remover (Toyobo). qRT-PCR was performed on Light-Cycler Roche 480 (Roche) with LightCycler® 480 SYBR Green I Master (Roche) under the following conditions: 95°C for 10 min, then 40 cycles of 95°C for 10 s and 60°C for 20 s. Primer sequences were listed in Dataset S10. ACTIN1 and UBIQUITIN1 were used as reference genes. The relative expression level was calculated using the 2 -ΔΔCt method (Schmittgen and Livak 2008).

Detection of ROS and Quantitative Measurement of Hydrogen Peroxide (H 2 O 2 )
Nitroblue tetrazolium (NBT) and 3,3′-diaminobenzidine tetrahydrochloride (DAB) staining were conducted as described by Fukao et al. (2011). Each leaf blade from a 21-day-old rice plant grown in soil was immersed in a solution containing 0.5 mg mL − 1 NBT (dissolved in 10 mM potassium phosphate buffer, pH 7.6) and 1 mg mL − 1 DAB (dissolved in 50 mM Tris-Acetate buffer, pH 5.0). After incubation under vacuum for 5 min, samples were kept in the dark at 25°C for 3 h for NBT staining and 24 h for DAB staining. Subsequently, plant materials were boiled in 95% (v/v) ethanol for 20 min to remove chlorophyll and rehydrated in 40% (v/v) glycerol. Each experiment was repeated on at least five plants and representative images are displayed.
H 2 O 2 content was measured using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen). Briefly, roots and shoots of 7-day-old seedlings and 21day-old leaf blades from at least three individual plants were freshly collected (50 mg, three replicates) and ground to a fine powder in liquid nitrogen followed by the immediate addition of 1 mL ice-cold 20 mM sodium phosphate buffer (pH 6.5). The supernatant was collected by centrifugation (10,000 g, 4°C, 10 min) followed by analysis according to the manufacturer's instructions.

Measurement of Antioxidative Enzyme Activity
Peroxidase activity was measured according to Chen et al. (2013). Briefly, roots and shoots of 7-day-old seedlings and 21-day-old leaf blades from at least six individual plants were freshly collected and ground in 50 mM sodium phosphate buffer (pH 7.8) supplemented with 1% polyvinylpyrrolidone and β-mercaptoethanol (1:5, m/ v) on ice. The supernatant was collected after centrifugation (13,000 g, 4°C, 15 min) and used for peroxidase activity assay. The reaction mixture (2 mL) composed of 50 mM sodium acetate buffer (pH 5.6), 5.4 mM guaiacol, 15 mM H 2 O 2 , and 100 μL of enzyme extract. The absorbance at 470 nm increase in 1 min was recorded using Cary 60 UV-Vis (Agilent). Peroxidase activity was determined by the increase in the absorbance at 470 nm per minute per mg protein. The enzyme extract from 21day-old leaf blades was also used for determination of CAT, APX and MDAR activity. CAT and APX activity were measured according to Wu and Yang (2016). The reaction mixture for CAT activity determination composed of composed of 50 mM sodium phosphate buffer (pH 7.8) 1.5 mL, ddH 2 O 1 mL, enzyme extract 200 μL and 300 μL 0.1 M H 2 O 2 . The reaction mixture for APX activity determination composed of 50 mM sodium phosphate buffer (pH 7.8) 1.25 mL, 5 mM ascorbic acid (AsA) 200 μL, enzyme extract 50 μL and 500 μL 0.1 mM H 2 O 2 . While MDAR activity was measured according to Hossain and Asada (1985). The reaction mixture composed of 50 mM sodium phosphate buffer (pH 7.8) 450 μL, 0.1 mM AsA 500 μL, 0.1 mM NADH 800 μL, enzyme extract 100 μL and 0.14 unit of ascorbate oxidase. CAT, APX and MDAR activity were determined by the decrease in the absorbance at 240 nm, 290 nm and 340 nm, respectively, per minute per mg protein.

Quantification of Plant Hormones
Plant hormone extraction and quantification were conducted according to Pan et al. (2010) with minor modifications. Briefly, leaf tissue was ground into fine powder in liquid nitrogen. Plant hormones were extracted from 1 g powder with 10 mL isopropanol/H 2 O/concentrated HCl (2:1:0.002, v/v/v) extraction buffer supplemented with 8 μL internal standards solution containing 1 μg mL − 1 of each internal standard compound (d 5 -IAA, d 6 -ABA, d 6 -SA, H 2 JA, and d 2 -GA 4 ). After shaking for 30 min at 4°C, 20 mL dichloromethane was added to the extraction buffer followed by further shaking for 30 min at 4°C. Samples were then centrifuged at 13000 g for 5 min and the lower phase was transferred, concentrated to dryness with nitrogen flow in the dark and resolved with 400 μL methanol (0.1% formic acid). Hormones were quantified using HPLC-MS/MS (SCIEX 6500 QTRAP) equipped with Poroshell 120 SB-C18 column (Agilent).
Additional file 1: Fig. S1 Growth of the rice osacbp6 mutant and the wild type (WT) in soil. a Images representing 5-and 7-day-old seedlings grown in soil. Bars, 1 cm. b Shoot and primary root length comparison between WT and osacbp6 (left) and primary root reduction of osacbp6 (right) Values are means ± SD (n = 13). Asterisks indicate significant differences between WT and osacbp6 as evaluated by Student's t-tests: ***P < 0.001. WT, Oryza sativa var. japonica cv. Dongjin.
Additional file 2: Fig. S2 Grain size and indole-3-butyric acid (IBA) sensitivity of osacbp6. a The grain length of osacbp6 is shorter than WT. Values are means ± SD (n = 30). Asterisks indicate significant differences between WT and osacbp6 as evaluated by Student's t-tests: ***P < 0.001. b osacbp6 is sensitive to the growth inhibition effect of IBA. Values are means ± SD (n = 5). Asterisks indicate significant differences between mock and IBA treatment as evaluated by Student's t-tests: ***P < 0.001. WT, Oryza sativa var. japonica cv. Dongjin.
Additional file 8: Fig. S8 Expression of peroxidases from the Rice eFP Browser. Absolute expression levels in the seedling root, young leaf, and mature leaf were retrieved from the Rice eFP Browser (http://bar. utoronto.ca/efprice/cgi-bin/efpWeb.cgi).