The Arabidopsis ABHD11 Mutant Accumulates Polar Lipids in Leaves as a Consequence of Absent Acylhydrolase Activity1[OPEN]

Disruption of the ABHD11 hydrolase causes polar lipids accumulation, thereby enhancing plant growth. Alpha/beta hydrolase domain (ABHD)-containing proteins are structurally related with diverse catalytic activities. In various species, some ABHD proteins have been characterized and shown to play roles in lipid homeostasis. However, little is known about ABHD proteins in plants. Here, we characterized AT4G10030 (AtABHD11), an Arabidopsis (Arabidopsis thaliana) homolog of a human ABHD11 gene. In silico analyses of AtABHD11 revealed homology with other plant species with a conserved GXSXG lipid motif. Interestingly, Arabidopsis abhd11 mutant plants exhibited an enhanced growth rate compared with wild-type plants. Quantitative analyses of the total lipids showed that the mutant abhd11 has a high amount of phospholipid and galactolipid in Arabidopsis leaves. The overexpression of AtABHD11 in Escherichia coli led to a reduction in phospholipid levels. The bacterially expressed recombinant AtABHD11 hydrolyzed lyso(phospho)lipid and monoacylglycerol. Furthermore, using whole-genome microarray and real-time PCR analyses of abhd11 and wild-type plants, we noted the up-regulation of MGD1, -2, and -3 and DGD1. Together, these findings suggested that AtABHD11 is a lyso(phospho)lipase. The disruption of AtABHD11 caused the accumulation of the polar lipids in leaves, which in turn promoted a higher growth rate compared with wild-type plants.

Mutations in ABHD5/CGI58 cause Chanarin-Dorfman syndrome, an autosomal recessive disorder in humans (Rozenszajn et al., 1966;Dorfman et al., 1974;Chanarin et al., 1975). More recently, studies have demonstrated that human CGI-58 has lysophosphatidic acid (LPA) acyltransferase (Ghosh et al., 2008a) and lysophosphatidylglycerol (LPG) acyltransferase (Zhang et al., 2014) activities. In various cancers, high levels of ABHD11 mRNA transcripts have been reported, but the role of this enzyme in cancer metabolism is not known. ABHD11 has been identified as a possible biomarker for lung carcinoma (Wiedl et al., 2011). In vivo metabolite profiling revealed that human ABHD3 overexpressing cells exhibited elevated levels of phospholipase activity . In a similar study, ABHD6 was shown to have an enzymatic activity that hydrolyzed both monoacylglycerol (MAG) and lysophospholipid . In a rodent model, ABHD12 was reported to be a major lysophosphatidyl-Ser and MAG lipase (Blankman et al., 2013). Similarly, ABHD16A was shown to hydrolyze phosphatidyl-Ser (PS) in mammalian systems. Disruption of these two enzymes showed altered phospholipid and lysophospholipid levels and caused neuroimmunological disorders in mice (Kamat et al., 2015). ICT1, an ABHD5 homolog in yeast, was shown to encode a protein with soluble LPA acyltransferase activity that could enhance phospholipid synthesis during organic solvent stress (Ghosh et al., 2008b). Based on these studies, it is clear that ABHD proteins play a major role in maintaining lipid homeostasis in eukaryotic organisms.
In Arabidopsis (Arabidopsis thaliana), ABHD is considered to be a large family of proteins. PES1 and PES2 from Arabidopsis belong to the ABHD family of proteins, which have been shown to play a role in phytyl ester biosynthesis in chloroplasts (Lippold et al., 2012). Recombinant Arabidopsis CGI58 was demonstrated to possess LPA acyltransferase activity (Ghosh et al., 2009). Moreover, in 2010, James et al. observed the accumulation of neutral lipids in the leaves of Arabidopsis cgi58 mutants. Figure 1. In silico analyses of AtABHD11. A, Neighbor-joining tree of AtABHD11 along with known plant ABHD11 protein sequences. The maximum-likelihood bootstrap support is indicated in red color. The bootstrap threshold was set to 70% and the confidence level to over 95% as highlighted with bold lines. The scale bar indicates the number of amino acid substitutions per site. B, Sequence comparison of the AtABHD11 proteins with other closely related plant ABHD proteins. The species include G. soja, T. urartu, and G. arboreum. Residues conserved in all proteins are indicated using a conservation graph. Red underline shows the conserved lipid motifs in plant ABHD11 proteins.
Herein, we attempted to understand the role of the human ABHD11 homolog in Arabidopsis. We used both transcriptomic and lipidomic approaches to characterize AT4G10030 (AtABHD11). In silico analyses were carried out to explore the sequence similarities of AtABHD11 with other plant species, as well as within the Arabidopsis genome. Lipidomics analyses revealed that disruption of abhd11 caused a significant increase in leaf phospholipids and galactolipids compared with the wild-type levels. Microarray profiling of abhd11 mutant plants showed the altered expression of genes involved in galactolipid and phospholipid metabolic pathways. In vitro biochemical studies revealed the presence of lyso(phospho)lipase and MAG lipase activities for heterologously overexpressed recombinant AtABHD11, demonstrating that it plays an important role in maintaining lipid homeostasis in Arabidopsis leaves.

AtABHD11 and Its Plant Homologs
To identify homologs of ABHD-containing proteins in Arabidopsis and other plant species, the protein sequence of AT4G10030 (AtABHD11) was retrieved from The Arabidopsis Information Resource database. Homology searches in Arabidopsis and other plant species revealed that 24 Arabidopsis proteins were homologous. However, ABHD11 proteins from other plant species, such as Triticum urartu, Table I. Primers used in this study  Black boxes indicate the exons, and the triangle indicates the site of T-DNA insertion in the AtABHD11 genome. The inset shows PCR amplification using genomic DNA from 3-week-old Columbia-0 and abhd11 plants. B, Real-time PCR analysis of the Arabidopsis ABHD11 mRNA expression levels in various tissues of Columbia-0. Graph shows the expression pattern of AtABHD11 as compared to rosette leaf stage. C, Real-time PCR analysis of AtABHD11 relative expression in 3-week-old leaves. Each value represents the mean 6 SD of three biological replicates. Actin was used as an internal control to determine the quality and quantity of cDNAs. D, Phenotype of the abhd11 mutant at different growth stages as compared to Columbia-0 (wild-type). abhd11, salk_090222c. Gossypium arboreum, and Gly soja, showed the greatest homology with AtABHD11 (Fig. 1A). In 2009, Ghosh et al. demonstrated three different conserved motifs of Arabidopsis ABHD-containing proteins. A sequence comparison of AtABHD11 with closely related plant species showed commonality in the amino acid residues with a conserved lipase (GXSXG) motif. Additionally, these proteins also showed a conserved acyltransferase (HXXXXD) motif (Fig. 1B).

Identification of a Homozygous abhd11 Mutant and Phenotype Analysis
We investigated all of the SALK T-DNA insertion mutant lines of Arabidopsis ABHD11 that were available from the Arabidopsis Biological Resource Center. PCR was performed with genomic DNA extracted from Columbia-0 (wild-type) and SALK mutants of ABHD11. Homozygous lines were screened using T-DNA border-specific primers and the genespecific primers as given in Table I. Among eight Salk lines, the presence of a T-DNA insertion was confirmed only in SALK_090222c (abhd11) by PCR ( Fig. 2A, inset). The site of the DNA insertion was indicated by a triangle in a genomic DNA map. Black boxes and lines represent coding and noncoding regions of the AtABHD11 genome, respectively ( Fig. 2A). In silico analysis by Genevestigator showed that the expression of ABHD11 was elevated right before bolting stage. Further validation was carried out by real-time PCR; the analysis revealed that the expression of ABHD11 was indeed high in the rosette leaves as compared to other stages in the wild type (Fig. 2B). Hence, the rosette stage was chosen to carry out the further experiments.
To check the ABHD11 mRNA expression level in the mutant, quantitative real-time PCR was performed. Figure 2C shows the expression of AtABHD11 was more than 3-fold reduced compared with the wild type (Fig. 2C). Wild-type and abhd11 plants were grown on soil supplemented with one-half Murashige and Skoog medium under optimal conditions. The growth pattern of wild-type and abhd11 plants was analyzed at 3 weeks ( Fig. 2D, inset), 6 weeks ( Fig. 2D, left), and 10 weeks (Fig. 2D, right). The abhd11 plants showed an early bolting and higher rate of growth as compared with wild-type plants. The consistency of changes in the growth patterns was observed in plants from different batches. However, the size of the leaves in abhd11 remained the same as the control plants.  [MGDG], and digalactosyldiacylglycerol [DGDG]) were analyzed from the leaves of both wild-type and abhd11 tissues. Triacylglycerol (TAG) was analyzed from the seeds and leaves of wild-type and abhd11 tissues. Although abhd11 disruption caused an overall increase in the phospholipid content, a significant increase was observed only in PC, followed by PE, PG, and PI, whereas PA and PS did not show any significant changes (Fig.  3A). Lipid molecular species identification of PC, PE, PG, and PI showed an increased amount of unsaturated fatty acids with 16:1, 18:2, and 18:3 (Fig. 3B). Similarly, galactolipids, such as DGDG and MGDG, also showed a significant increase in abhd11 mutant plants (Fig. 4A). The analyses of abhd11 plants for DGDG and MGDG showed an elevation in the levels of (34:6)MGDG and (36:6)MGDG ( Fig. 4B) and (36:6)DGDG and (34:3)DGDG ( Fig. 4C) compared with wild-type plants. Neutral loss scans for various acyl fatty acids of leaves and seeds TAG did not reveal any significant changes in the abhd11 mutant ( Fig. 4D and 4E). In addition, TLC analysis for phospholipids of abhd11 mutant seeds had no profound changes when compared to the wild type. Together, lipidomics data for abhd11 suggest that AtABHD11 may play a role in phospholipid metabolism and could possibly channel phospholipids toward galactolipid generation.

Recombinant AtABHD11 Alters Phospholipids in Escherichia coli
To test our hypothesis of an association of AtABHD11 in phospholipid metabolism, AtABHD11 was overexpressed in BL21 (DE3) cells, and the protein was purified from the soluble fraction using a Ni 2+ -NTA column from an overexpressed pRSET-A vector and AtABHD11 (Fig.  5A). Recombinant AtABHD11 protein was confirmed by immunoblot analysis using an anti(His) 6 monoclonal antibody (Fig. 5B). Overexpression of AtABHD11 and the vector control in the presence of [ 14 C]acetate led to a reduction in PE and cardiolipin, major phospholipids of E. coli, but the identity of other radiolabeled spots are unknown (Fig. 5C). Further, PE and cardiolipin spots from five individual experiments confirmed that the incorporation of radiolabel was 2-fold lower in corresponding phospholipids when AtABHD11 was overexpressed (Fig. 5C, inset). Figure 6. In vitro acylhydrolase activity of recombinant AtABHD11. A, The phospholipase assay was conducted with total lipids extracted from the yeast cells as a substrate (2.09 3 10 5 dpm/tube) with 1 mg of the recombinant, purified AtABHD11 as the enzyme source. The eluted fraction of vector background was used as a control. After a 30 min reaction, the extracted lipids were separated on a two-dimensional TLC. B, Phospholipase activity of AtABHD11. Inset shows the representative TLC of PS lipase activity with the heat-inactivated proteins from purified vector and AtABHD11 as control. C, Lysophospholipase activity of AtABHD11. Both phospholipase and lysophospholipase activities were carried out with purified proteins from recombinant AtABHD11 and empty vector, and the assays were carried out for 30 min at 30˚C. Assay contained 50 mM NBD-fluorescent fatty acid-labeled lyso(phospho)lipids substrates in the presence of 10 mM CaCl 2 . Assay reactions were stopped, and lipids were extracted and separated on a silica-TLC plate using petroleum ether:diethyl ether:acetic acid (70:30:1, v/v) as the solvent system. D, The assay was performed with 50 mM [1-14 C]MAG (0.025 mCi/tube). Lipids were extracted and then resolved on a silica-TLC plate using the solvent system described above. 2ENZ, The assay was performed without enzyme; VEC, purified protein from overexpressed vector used as an enzyme; GENE, purified protein from overexpressed ABHD11 used as an enzyme. Graphs represent the mean (6SD) of three independent experiments.

Acylhydrolase Activity of AtABHD11
It was clear from the earlier experiment that AtABHD11 had a potential role in phospholipid metabolism. Accordingly, a significant reduction in phospholipid levels was observed when AtABHD11 was expressed heterologously. The presence of hydrolytic activity was analyzed using Ni 2+ -NTA purified AtABHD11 protein. In vitro assays were performed with total lipids as substrate, extracted from [ 14 C]acetate-labeled wild-type yeast cells. As shown in Figure 6A, two-dimensional TLC analysis confirmed that AtABHD11 protein had the capacity to hydrolyze phospholipid more efficiently than the protein purified from the vector control. This experiment also showed the broad substrate specificity of the phospholipase activity. To validate further, the phospholipase activity of AtABHD11 was monitored with various individual lysophospholipids (LPA, lysophosphatidylcholine [LPC], lysophosphatidylethanolamine [LPE], LPG, lysophosphatidylinositol, and lysophosphatidylserine), phospholipids (PA, PC, PE, PG, PI, and PS), and nonpolar (MAG, diacylglycerol [DAG], and TAG) lipid substrates. All the assays were performed by incubating 2 mg of purified recombinant AtABHD11 with 50 mM of various substrates for 30 min at 30°C. The high sensitive fluorescent-based method was used to estimate the released free fatty acid (FFA) Figure 7. Quantitative real-time analysis of lipid metabolism genes. A, Heat map shows the up-and downregulated genes in abhd11 mutant leaves. B, The graph shows the relative mRNA expression levels of several important genes involved in de novo lipid biosynthesis in abhd11 mutant leaves. Dotted line represents the base value 1 for the wild type (WT). C, GUS assay for the transiently expressed vector and ABHD11 in the wild type and abhd11. Untransformed wild-type and abhd11 leaves shown as control. D, mRNA levels of chloroplast lipid genes in AtABHD11 overexpression and complementation. E and F, Effect on phospholipid (E) and galactolipid contents (F) upon overexpression of AtABHD11. Vec, Empty vector overexpressed; ABHD11, overexpression of AtABHD11. Graphs represent the mean (6SD) of five independent experiments.
from various substrates in a 96-well plate (Savinainen et al., 2014). The cellular background activity (both vector and AtABHD11 fraction) was negligible, and the enzyme activity was calculated after normalizing with the background activity. The phospholipase activity of AtABHD11 revealed that the maximum activity was observed with PS followed by PG and PE (Fig. 6B). The lysophospholipase activity was significantly higher with LPA followed by LPE and LPG (Fig. 6C). Further, a similar assay was validated with the available NBDfluorescent fatty acid-labeled lyso(phospholipid) substrates, which confirmed the release of FFA from the corresponding substrates. Representative figure for fluorescent substrate assay for PS lipase activity was shown (Fig. 6B, inset). Acylhydrolase activity with nonpolar lipid substrates showed that the enzyme also had MAG lipase activity (Fig. 6D). Data are presented as the means of three independent experiments. Collectively, these experiments suggested that Arabidopsis ABHD11 is indeed an acylhydrolase with broader substrate specificity.

Altered Expression Profile of Lipid Genes in abhd11
The mutant and overexpression analyses confirmed the contribution of AtABHD11 to phospholipid metabolism. As shown in previous studies of the de novo pathway, excess phospholipids could be channeled to the synthesis of TAG. However, elevated levels of total phospholipid in the abhd11 mutant did not appear to influence the formation of TAG. Instead, total galactolipid levels were found to be elevated in the abhd11 mutant. We used the leaf microarray data of abhd11 mutant and wild-type tissues to assess the turnover of the excess phospholipid that accumulated in mutant leaves. As shown in the heat map ( Fig. 7A; Supplemental Data S1), abhd11 mutant had 3080 and 2417 genes that were up-and down-regulated, respectively, by more than 2fold compared with wild-type. Validation of the mRNA expression levels of lipid genes confirmed a more than 10-fold up-regulation of the MGD1, MGD2, MGD3, and DGD1 (approximately 8-fold) genes. Analysis of the genes involved in TAG biosynthesis showed that they were slightly up-regulated (Fig. 7B).
To validate the earlier experiment, AtABHD11 fulllength gene was transiently overexpressed in both wild-type and abhd11 mutant plant leaves of 3 weeks old. Cauliflower mosaic virus 35s-driven AtABHD11 expression was confirmed by GUS assay. Untransformed wild-type and abhd11 leaves were used as control (Fig. 7C). Further, real-time PCR confirmed that both overexpressed and complemented leaves significantly lowered the transcript levels of MGD1, MGD2, MGD3, and DGD2, but DGD1 showed no significant effect as compared to vector control (Fig. 7D). In addition, polar lipid contents were analyzed in overexpressed vector and AtABHD11 leaf samples of the wild type and abhd11. The results indicated that the phospholipids (Fig.  7E) and galactolipids (Fig. 7F) were significantly lowered upon overexpression of AtABHD11. Together, these experiments revealed that AtABHD11 alters the expression of chloroplast lipid genes and polar lipid content in plant leaves.

DISCUSSION
ABHD-containing proteins have been shown to possess a wide variety of hydrolytic activities (Ollis et al., 1992). Recently, many studies have been carried out to understand the role of lipid molecules in plant growth and development. In plants, phospholipids are extensively hydrolyzed by phospholipases, which have been shown to be highly expressed during various stress conditions (Wang, 2001;Testerink and Munnik, 2005;Li et al., 2006;Wang et al., 2006). However, our understanding of the enzymes that are involved in lipid homeostasis remain insufficient. ABHD5, a human homolog in plants, has been characterized and shown to have LPA acyltransferase activity (Ghosh et al., 2009). Moreover, the disruption of ABHD5 in Arabidopsis caused the accumulation of nonpolar lipids in vegetative tissues (James et al., 2010).
In this study, we characterized AT4G10030 (AtABHD11) in both mutant and overexpression conditions. The abhd11 mutant seeds, when grown on soil, exhibited a higher rate of plant growth compared to the wild type. To date, characterization of ABHD proteins in different organisms revealed their direct contributions to various aspects of lipid metabolism (Ghosh et al., 2008a(Ghosh et al., , 2008b(Ghosh et al., , 2009Blankman et al., 2013). To assess whether AtABHD11 also exhibits a similar role, we analyzed the lipidomics data of Arabidopsis abhd11 mutant tissues. In 1995, Ohlrogge and Browse demonstrated that PC is the major phospholipid in Arabidopsis tissues. Interestingly, we observed a significant increase in the phospholipid content as PC . PG . PE . PI in abhd11 mutant tissues. Furthermore, quantitative analysis of leaf phospholipids molecular species revealed that the amount of unsaturated fatty acid containing lipids was significantly increased in the abhd11 mutant as compared to the wild type. An increase in phospholipids could occur because of either an increase in lysophospholipid acyltransferase activity or the reduced turnover of phospholipids.
To check for differences in the expression patterns of lipid biosynthetic genes in the abhd11 mutant, leaf whole-genome microarray profiles of both wild-type and abhd11 mutant tissues were compared. The gene expression profile did not show notable changes in the expression of genes that are involved in TAG biosynthesis. Interestingly, significant up-regulation was observed in chloroplast lipid biosynthetic genes. This observation also supported our quantitative lipid profiling in which the amount of TAG remains unchanged, while the galactolipid levels were found to be higher in abhd11 mutants. Based on the above observations, we hypothesized that phospholipids are possibly channeled toward galactolipid synthesis in leaves. An increase in the galactolipids of vegetative tissues in abhd11 mutants implicated their similarity with Atcgi58 mutants (James et al., 2010). However, the TAG levels were not significantly altered in abhd11 mutant plants, as reported in the cgi58 mutant. This finding clearly demonstrates the diversity of the catalytic activity of ABHD proteins. Quantitative real-time PCR analyses further confirmed our hypothesis that the mRNA levels were more than 10-fold higher for MGD1, -2, and -3 and 8-fold higher for DGD1, whereas the other lipid genes were up-regulated to a lesser extent.
It has been reported that the highest amount of 16:1 and 16:3 could be observed in chloroplast lipids in Arabidopsis (Ohlrogge and Browse, 1995). In accord with an earlier report, the galactolipid increase was observed with MGDG 34:6(18:3/16:3) and 36:6(18:3/18:3) and with DGDG 36:6(18:3/18:3) and 34:3(18:1/18:2). Moreover, there was also an increase in PC 36:5(18:3/18:2). Similar data were observed when a plastidial phosphatidic acid phosphatase (PAH) gene was mutated in Arabidopsis, resulting in higher amounts of PC due to the upregulation of LPC acyltransferase (Wang et al., 2014). Unlike pah mutants, the abhd11 mutant did not cause marked changes in the expression of the LPCAT gene. There are two independent pathways that equally contribute to the synthesis of chloroplast lipids in Arabidopsis. The amount of PA was not altered in abhd11 mutants, and this could be because of the enhanced prokaryotic pathway in which PA serves as a precursor to synthesize MGDG. Increases in galactolipids also indicate the possibility of DAG as a precursor for the synthesis of various galactolipids. Overall, abhd11 mutant analysis suggests that its mutation caused an increase in phospholipids in Arabidopsis leaves; thus, these phospholipids were utilized efficiently by the MGD and DGD enzymes to generate more galactolipids in mutant leaves (Fig. 8).
The results obtained from mutant were verified by overexpression and complementation of the ABHD11 gene in its respective wild-type and abhd11 mutant plant leaves. Both gene expression and lipid analysis profiles correlated well with each other, which confirmed the involvement of AtABHD11 in polar lipid metabolism. To validate further, we hypothesized that ABHD11 is a possible phospholipase. Overexpression of AtABHD11 caused a significant reduction of phospholipids in E. coli cells. Interestingly, our in vitro experiments with individual substrates using purified recombinant AtABHD11 showed lipase activity toward phospholipids and lysophospholipids. However, a phospholipase assay using yeast total lipids as the substrate showed a slight variation regarding substrate preferences. When total lipid was used as a substrate, enzyme hydrolyzed PC more effectively than other lipids. This data also supports our earlier observation of PC increase in abhd11. These experiments suggested that the enzyme has broad substrate specificity and that the activity is dependent on the amount of substrate available in the system. Additionally, minimum activity was observed with MAG as a substrate.
Our observation of abhd11 and the enhanced growth of mutant plants are in accord with earlier observations that the phospholipids act as plant growth regulators. It has been shown that exogenous application of phospholipids to plants induced hormone-like changes and led to cell expansion and delayed senescence (Chapman, 1998;Laxalt and Munnik, 2002;Meijer and Munnik, 2003). This study clearly confirms that AtABHD11 acts as a lyso(phospho)lipase and that the abolishment of the activity increases leaf phospholipids and DAG content, thereby increasing the galactolipid content in leaves of abhd11 in Arabidopsis.

Materials
[1-14 C]Monooleoyl-rac-glycerol (MAG, 55 mCi/mmol) was obtained from American Radiolabeled Chemicals. [ 14 C]Acetate (51 mCi/mmol) was obtained from Bhabha Atomic Research Centre (Mumbai, India). All of the fluorescent phospholipid substrates and lysolipids used in the enzyme assays and the lipid standards were from Avanti Polar Lipids. Oligonucleotide primers, anti-His-tag monoclonal antibody, and all other reagents were obtained from Sigma-Aldrich. Alkaline phosphatase substrate was from Perkin-Elmer. Restriction endonucleases and Pfu polymerase were from Thermo Scientific. Columbia-0 and AT4G10030 mutant lines were obtained from the Arabidopsis Biological Resource Center.

In Silico Analysis
The primary sequence of the AtABHD11 protein was retrieved from The Arabidopsis Information Resource database. Homologs of AtABHD11 from Arabidopsis (Arabidopsis thaliana) and other plant species were retrieved from the National Center for Biotechnology Information database. Sequences were subjected to construct a neighbor-joining tree. The plant AtABHD11 homologs were further analyzed for sequence similarity by multiple sequence alignment. Both analyses were carried out using CLC workbench software.

Plant Growth Conditions
Arabidopsis Columbia-0 (wild-type) and abhd11 muatnt lines were grown vertically on either soil or half-strength Murashige and Skoog medium supplemented with 0.5% (w/v) Suc and solidified with 0.8% agar. After 2 d of stratification at 4°C, seedlings were transferred to 23°C under a 16 h day (140 mmol m 22 s 21 ) and 8 h night regime.

Characterization of abhd11 Mutant
T-DNA insertions in abhd11 mutant lines (salk_090222c, salk_109230, salk_090220, salk_090225, salk_090229, salk_109046, salk_090232, and salk_090227) were analyzed with a T-DNA left border and corresponding genespecific primers as shown in Table I. A homozygous line was identified by PCR using genomic DNA as a template from wild-type and abhd11 mutant plants. The mutation was further confirmed by quantitative real-time PCR using cDNA synthesized from an equal amount of leaf total RNA. Actin was employed as an internal control in both wild-type and abhd11 mutants.

Lipidomics Analysis of abhd11 Mutant and Wild-Type Plants
Total lipids were extracted from 100 mg of wild-type and salk_090222c (abhd11) dried seeds or eight rosette leaves. Lipids were extracted and processed according to the described protocol (Welti et al., 2002;Wanjie et al., 2005). Lipids were analyzed at the KS Lipidomic Research Center by continuous infusion into an electrospray ionization source on a triple-quadrupole mass spectrometer (API14000, Applied Biosystems). The molecular species of lipids were quantified in comparison to the two internal standards using a correction curve determined between standards (Wanjie et al., 2005). Among five replicates, three replicates were chosen and used for further Student's t test to determine statistical significance. The values are represented as nmol of signals per extracted tissue dry weights. Phospholipid, MGDG, and DGDG amounts were represented as normalized mass spectral signal, which is normalized to the 1 nmol of two internal standards of that class (Vu et al., 2012).

Microarray Analysis
RNA isolation from abhd11 (salk_090222c) and wild-type Arabidopsis was performed using a Qiagen RNase plant mini-kit following the manufacturer's protocol with DNAse treatment. The RNA concentration and purity were determined at an optical density ratio of 260/280 using a Nanodrop ND-1000 spectrophotometer, and the integrity of total RNA was verified on an Agilent 2100 Bioanalyzer using a RNA 6000 Nano LabChip.
The samples used for gene expression were labeled using an Agilent Quick-Amp labeling kit (p/n5190-0442). Total RNA (500 ng) was reverse transcribed at 40°C using an oligo(dT) primer tagged to a T7 polymerase promoter and converted to double-stranded cDNA. The synthesized double-stranded cDNA molecules were used as a template for cRNA generation. The cRNA was generated by in vitro transcription, and the dye Cy3 CTP (Agilent) was incorporated during this step. The cDNA synthesis and in vitro transcription steps were carried out at 40°C. Labeled cRNA was cleaned up using Qiagen RNeasy columns.
The labeled cRNA (600 ng) sample was fragmented at 60°C and hybridized onto genotypic designed custom Arabidopsis 8360K (AMADID No: 037661) arrays. Fragmentation of the labeled cRNA and hybridization were carried using a gene expression hybridization kit (Agilent Technologies). Hybridization was carried out in Agilent's Surehyb chambers at 65°C for 16 h. Hybridized slides were washed using Agilent gene expression wash buffers  and scanned using an Agilent Microarray Scanner (Part No. G2600D).
Data extraction from images was carried out using Feature Extraction software Version 11.5 (Agilent). Raw data were analyzed using GeneSpring GX software from Agilent. Significant genes that were up-and down-regulated showing 1-fold (log base 2) and greater changes within samples with respect to control. Genes were classified based on the functional category and pathways using Genotypic Biointerpreter-Biological Analysis Software.

Quantitative Real-Time PCR
Wild-type and abhd11 total RNA were isolated from leaves using a Sigma Plant Total RNA isolation kit. Next, 1 mg of total RNA was used to prepare cDNA using a Thermo Scientific RevertAid H Minus first-strand cDNA synthesis kit. Primers were designed using Primer Express software 3.0 (Applied Biosystems), and the sequences are listed in Table I. The quality and quantity of the RNA were determined using nanodrop. Further, amount of cDNA in both wild-type and abhd11 were confirmed with actin primers. Quantitative realtime PCR mix was prepared using cDNAs (1:20 dilution) with Power SYBR Green PCR Master Mix. The mRNA expression levels were analyzed in triplicate using an Applied Biosystems machine (SDS2.1). The quantitative real-time PCR data are presented as the means of three biological replicates, and the data were analyzed using Student's t test.

Transient Expression of AtABHD11
Eight Arabidopsis leaves of 3 weeks old were infected with Agrobacterium tumefaciens strain (GV3101) carrying the control cauliflower mosaic virus 35s promoter-driven PBI121 vector or PBI121-ABHD11 fusion construct. Prior to infection, bacterial cells carrying empty vector or the AtABHD11 construct were grown individually overnight in YEP medium containing rifampicin (50 mg/mL) and kanamycin (50 mg/mL). The cells were pelleted, washed, and resuspended in induction medium (5.55 mM Glc, 60.25 mM K 2 HPO 4 , 33.04 mM KH 2 PO 4 , 2.34 mM sodium citrate, 7.57 mM (NH 4 ) 2 SO 4 , 5.55 mM Fru, 43.43 mM glycerol, 10 mM MES, pH 5.6, and 100 mM acetosyringone) and grown for 5 h at 30°C. The cells were pelleted and resuspended in infiltration medium (10 mM MgSO 4 and 10 mM MES, pH 5.6, to an optical density of 0.5, followed by the addition of 200 mM acetosyringone). After 48 h of infiltration, leaves were used for the GUS assay (Ueki et al., 2008) to confirm the ABHD11 expression. Untransformed wild-type and abhd11 mutant leaves were used as controls for GUS assay. Real-time PCR was performed as mentioned earlier using cDNA prepared from equal amount of transiently overexpressed plant leaves, and the results were compared with the vector control. The overexpressed leaves were also harvested for lipid analysis as described earlier. The total lipids were run on silica-TLC plate using acetone:toluene:water (90:30:7.5, v/v) as the solvent system. The lipids were scrapped and re-extracted and analyzed by AB SCIEX 6500 Q-TOF mass spectrometer. The results were analyzed with PeakView and LipidView softwares (AB SCIEX). The amount of lipids was expressed as a percentage of total lipids per milligram tissue dry weight. Analysis were performed with both individual lipids and total lipids, and the results are presented as mean of five biological replicates.

Cloning and Expression of AtABHD11
Arabidopsis wild-type cDNA was used to amplify AtABHD11 using genespecific primers as given in Table I. The gene was cloned into a pRSET-A vector using BamHI and KpnI restriction sites. A positive clone was confirmed by double-digestion, and the accuracy of the sequence was confirmed by DNA sequencing using T7 forward and reverse primers. The AtABHD11 gene construct was transformed into Escherichia coli BL21 (AI3) cells. Cells were then induced with 0.5 mM isopropylthio-b-galactoside (IPTG) for 6 h at 37°C to enable gene expression. The cell pellets were resuspended in lysis buffer containing 50 mM Tris-HCl, pH 8.0, and 300 mM NaCl, 10% glycerol, 2 mM MgCl 2 , and 1 mM PMSF. Cells were disrupted by sonication and then the sample was subjected to centrifugation (5000 rpm) for 10 min followed by ultracentrifugation for 90 min (35,000 rpm) at 4°C. AtABHD11 expression was confirmed by western-blot analysis using an anti-His 6 monoclonal antibody (1:5000, v/v). Vector protein was used as a negative control. For protein purification, the supernatant containing His-tagged recombinant protein was allowed to bind to Ni 2+ -NTA agarose beads. After the collection of the flow-through, the column was washed with lysis buffer containing 20 mM imidazole. The bound protein was eluted with 250 mM imidazole in lysis buffer. The amount of purified recombinant protein was estimated using Lowry's method (Hartee, 1972). Bovine serum albumin was used as a standard to calculate the amount of protein. The protein purity was analyzed by resolving onto 12% (w/v) SDS-PAGE followed by Coomassie Brilliant Blue staining. The purified recombinant protein was subjected to dialysis overnight at 4°C, and the enzymatic assay was then performed.

In Vivo [ 14 C]Acetate Labeling of Lipids
Cells containing pRSET-A vector and pRSET-A-AtABHD11 constructs were grown in Luria-Bertani medium in the presence of 0.2 mCi/mL [ 14 C]acetate. Cells were allowed to grow until A 600 = 0.8/mL and then were induced with 0.5 mM IPTG for 6 h. Total lipids were extracted as described (Bligh and Dyer, 1959) from cells (A 600 = 20) that overexpressed the vector control or AtABHD11, and were run in a two-dimensional phospholipid solvent system (first dimension, chloroform:methanol:ammonia, 70:25:5, v/v; second dimension, chloroform: acetone:methanol:aceticc acid:water, 50:20:10:15:5, v/v). Data are presented as the means of five biological replicates, and data were analyzed using Student's t test.
In Vitro Phospholipase Activity AtABHD11 enzyme hydrolase activity was initially assessed using [ 14 C]acetate-labeled endogenous lipid substrates from wild-type yeast cells. The assay was conducted with 2 mg of purified vector and ABHD11 protein in the presence of [ 14 C]acetate-labeled total lipid substrates. After 30 min of assay at 30°C, lipids were extracted and separated using the two-dimensional phospholipid solvent system described above.

In Vitro Acylhydrolase Assay
The acylhydrolase activity of AtABHD11 was monitored with various individual lipid substrates like lysophospholipids (LPA, LPC, LPE, LPG, lysophosphatidylinositol, and lysophosphatidylserine) and phospholipids (PA, PC, PE, PG, PI, and PS). The assay was performed by incubating 2 mg of dialyzed purified recombinant AtABHD11 with assay buffer containing 50 mM of various substrate, 50 mM Tris-HCl, pH 8.0, 1 mM dithiothreitol, and 10 mM CaCl 2 . After 30 min of assay at 30°C, lipids were extracted and run on silica-TLC using petroleum ether:diethyl ether:acetic acid (70:30:1, v/v) as a solvent system. The release of FFA was determined as mentioned below. To validate the earlier experiment, the same assay was repeated with 50 mM NBD-fluorescent fatty acid-labeled lyso(phospho)lipids substrates, and fluorescent fatty acid release was identified on a TLC by scanning at 488 nm in typhoon FLA 9500.
For nonpolar lipid substrates, the acylhydrolase activity of AtABHD11 was measured by monitoring [ 14 C]fatty acid release from [1-14 C]MAG, TAG, and DAG (0.025 mCi) lipid substrates. The assay was conducted as mentioned above. Enzymatic products were monitored on a phosphor imager, corresponding spots were scraped from the TLC plate, and the radioactivity was quantified using a liquid scintillation counter. Experiments were carried out in triplicate, the release of radiolabeled FFA was counted from three biological replicates, and the data were analyzed using Student's t test.

Determination of Fatty Acid Release
The acylhydrolase activity of AtABHD11 was calculated by estimation of released FFA using a fluorescent FFA estimation kit (Cayman Chemical, Catalog No. 700310;Savinainen et al., 2014). Briefly, the enzyme assay was performed by above-mentioned standard conditions, and the released FFA was coupled via three-step enzymatic reaction to produce H 2 O 2 -dependent generation of resorufin and the fluorescence monitored (Ex 530 /Em 590 ) using Thermo Scientific Varioskan Flash Multimode Reader. The amount of FFA was calculated by normalizing with respective controls, such as background, no enzyme control, and heat-inactivated controls.