The GxSxG motif of Arabidopsis monoacylglycerol lipase (MAGL6 and MAGL8) is essential for their enzyme activities

Abstract

Monoacylglycerol lipase (MAGL) catalyzes the hydrolysis of monoacylglycerols (MAG) to free fatty acids and glycerol, which is the last step of triacylglycerol breakdown. Among sixteen members, Arabidopsis thaliana MAGL6 (AtMAGL6) and AtMAGL8 showed strong lipase activities, but several AtMAGLs including AtMAGL16 displayed very weak activities (Kim et al. in Plant. J 85:758–771, 2016). To understand the internal factors that influence Arabidopsis MAGL activities, this study investigated the significance of ‘GxSxS motif,’ which is conserved in MAGLs. First, we observed that the presence of a serine protease inhibitor, phenylmethylsulfonyl fluoride, decreased the enzyme activity of AtMAGL6 and AtMAGL8 by IC₅₀ values of 2.30 and 2.35, respectively. Computational modeling showed that amino acid changes of the GxSxG motif in AtMAGL6 and AtMAGL8 altered the nucleophilic elbow structure, which is the active site of MAGLs. Mutating the GxSxG motif in the recombinant maltose binding protein (MBP):AtMAGL6 and MBP:AtMAGL8 proteins to SxSxG, GxAxG, and GxSxS motifs completely demolished the activities of the mutant MAGLs. In contrast, no significant differences were observed between the activities of AtMAGL16 wild type form harboring the SxSxG motif, and mutant AtMAGL16 containing the GxSxG motif. These results revealed that the glycine and serine residues of the GxSxG motif are essential for AtMAGL6 and AtMAGL8 enzyme activities, and that AtMAGL16 may not be involved in the hydrolysis of lipid substrates.

Introduction

Triacylglycerol (TAG), a storage oil, is found in seeds, fruit mesocarp, pollen grains, and senescing leaves (Huang 1992; Kaup et al. 2002; Kim et al. 2002). During seed and pollen maturation, TAG synthesized on the endoplasmic reticulum (ER) accumulates in the oil body, surrounded by a single membrane through blebbing of the outer membrane of the ER (Huang 1996; Li-Beisson et al. 2013). When seeds and pollen grains germinate, the stored TAG is mobilized to supply carbon and energy sources for growing tissues (Graham 2008). In Arabidopsis thaliana, the initial step of TAG breakdown that produces diacylglycerol (DAG) and a free fatty acid from TAG is mediated by TAG lipases, SUGAR-DEPENDENT1 (SDP1) or SDP1-LIKE (SDP1L) (Eastmond 2006; Kelly et al. 2011; Theodoulou and Eastmond 2012). DAG lipase, which catalyzes the hydrolysis of DAG to MAG and free fatty acid, has not yet been identified; but the molecular and biochemical characterizations of MAGLs that produce the glycerol and free fatty acid from MAG were recently reported (Kim et al. 2016). Fourteen recombinant MAGL proteins in-frame with maltoase binding protein (MBP) were purified as a soluble form, when 16 MAGL genes, which are present in Arabidopsis acyl-lipid metabolism (http://aralip.plantbiology.msu.edu), were expressed in Escherichia coli. Upon verifying lipase activities using a non-esterified fatty acid (NEFA)-HR (2) colorimetric kit, AtMAGL6 and AtMAGL8 exhibited nearly homologous activity to that of human (Homo sapiens) MAGL (HsMAGL). Conversely, the rest of the recombinant MAGL proteins exhibited relatively low lipase activities (Kim et al. 2016). Therefore, the internal factors that are able to affect Arabidopsis MAGL activities should be further investigated.

The catalytic triad, which refers to three amino acid residues, serine (S) in the GxSxG motif, aspartic acid (D), and histidine (H), is known to act together at the active site of hydrolase and transferase enzymes, such as proteases, acylases, and lipases (Brumlik and Buckley 1996; Polgar 2005). The amino acid residues in the GxSxG motif of mouse (Mus musculus) MAG lipase were reported to be essential for their enzyme activity via site-directed mutagenesis (Karlsson et al. 1997). Comparative analysis of protein models of human MAG lipase to those of chloroperoxidase F having substrate specificity for small ions, and dog (Canis lupus) gastric and human pancreatic lipases that utilize TAG as a specific substrate, revealed that human MAG lipase may harbor substrate specificity for other molecules than small ions and TAG, based on their protein structural differences around the nucleophilic serine residue in the GxSxG motif (Labar et al. 2010). Although 9 of 11 Arabidopsis MAGL members harbor the GxSxG motif, interestingly the levels of their MAG lipase activities were very diverse (Kim et al. 2016). In particular, AtMAGL14 and AtMAGL16, which contain the SxSxG motif instead of the GxSxG motif, also displayed very low lipase activities (Kim et al. 2016).

Therefore, to investigate the importance of the GxSxG motif in Arabidopsis MAGLs, the enzyme activity of Arabidopsis MAGLs was first measured in the presence of a serine protease inhibitor, phenylmethylsulfonyl fluoride (PMSF). Computational modeling of Arabidopsis MAGLs (AtMAGL-6, -8, and -16) and their mutated proteins with amino acid substitutions in a GxSxG motif suggested that the GxSxG motif might be important for the enzymatic activity of MAG lipases. More specifically, the GxSxG motif of AtMAGL6 and AtMAGL8 was mutated to the SxSxG, GxAxG, or GxSxS motif by PCR-based site-directed mutagenesis. In addition, the first serine residue of the SxSxG motif in AtMAGL16 was substituted by a glycine residue, thereby creating the recombinant AtMAGL16 protein with the GxSxG motif. As a result, we observed that two glycine residues and one serine residue in the GxSxG motif of AtMAGL6 and AtMAGL8 are vital for their lipase activities. Even though the AtMAGL16 has the GxSxG motif, no significant elevation in lipase enzyme activity was observed, suggesting that there might be unknown essential domains that influence the enzymatic activity of AtMAGL16, or AtMAGL16 may act on other substrates than lipid substrates, such as MAG. To the best of our knowledge, this is the first report to demonstrate that the GxSxG motif might be essential for plant MAG lipases.

Materials and methods

Site-directed mutagenesis of Arabidopsis MAGLs

We carried out a point mutation using an overlap extension PCR to substitute the amino acid residues in the GxSxG motif of AtMAGL6 and AtMAGL8 with SxSxG, GxAxG, and GxSxS motifs (Ho et al. 1989). The mutagenic primers were designed for substitution of the nucleotide sequences: glycine to serine residues, or serine to alanine or glycine residues (Table 1). According to Ho et al. (1989), three-step PCR was carried out with the forward, reverse, and mutagenic primers, and with MBP:AtMAGL6, MBP:AtMAGL8, MBP:AtMAGL16, and MBP:HsMAGL clones as a template (Kim et al. 2016): (1) the 5′-mutated PCR fragments were amplified with the forward and mutagenic reverse primers; (2) the 3′-mutated PCR fragments were amplified with the mutagenic forward and reverse primers; and (3) the mutated full-length PCR fragments were amplified with the forward and reverse primers using the obtained 5′- and 3′-mutated fragments as templates. The mutated AtMAGL6, AtMAGL8, and AtMAGL16 PCR fragments were cloned in the BamHI/PstI or BamHI/SalI enzyme sites of the pMAGL-C2 vector (New England Biolabs, Hitchin, UK).

Table 1 List of primers used in this study

Purification of recombinant proteins from E. coli

MBP, MBP:HsMAGL, MBP:AtMAGL6, MBP:AtMAGL6_G111S, MBP:AtMAGL6_S113A, MBP:AtMAGL6_G115S, MBP:AtMAGL8, MBP:AtMAGL8_G117S, MBP:AtMAGL8_S119A, MBP:AtMAGL8_G121S, MBP:AtMAGL16, and MBP:AtMAGL16_S139G vectors were transformed into E. coli BL21 (DE3)-RIL strains (Kim et al. 2016). At an OD of 0.6–0.7, the E. coli cells were induced with 0.5 mM isopropyl-β-D-thiogalactopyranoside. After 12 h of incubation at 37 °C, E. coli cells were harvested and resuspended in buffer (200 mM NaCl, 20 mM Tris–HCl, pH 7.4, 10 mM β-mercaptoethanol, and 1 mM EDTA). After the addition of 100 μg mL⁻¹ lysozyme (Sigma-Aldrich, St. Louis, MO, USA), resuspended cells were incubated on ice for 30 min. Incubated cells were lysed on ice with a Vibra-Cell sonicator (Sonics and Materials Inc., Newtown, CT, USA). MBP and MBP:MAGL proteins were purified using amylose resin (New England BioLabs) and electrophoresed on 12 % SDS–polyacrylamide gels. Proteins were visualized by staining the gel with 0.2 % Coomassie Brilliant Blue R-250 (Bio-Rad, Hercules, CA, USA).

Measurement of MAGL activity

The lipase assay was performed as previously described by Kim et al. (2016). The MBP:MAGL proteins were quantified using the Bradford methods (Bradford 1976). The purified MBP:MAGL proteins (0.1 μg) were incubated in 100 μl of 50 mM sodium phosphate buffer (pH 8.0) or glycine–NaOH buffer (pH 9.0), 0.2 % Triton X-100, and 150 μM MAG (M7640; Sigma-Aldrich) containing an linoleic acid at the sn-1 position for 5 min at 30 °C. The released NEFAs were measured with the commercial NEFA-HR kit (Wako Pure Chemicals, Osaka, Japan), using Synergy H1 Hybrid Readers (BioTek, Winooski, USA) at 546 nm.

Inhibition assays

The MBP, MBP:HsMAGL, and MBP:AtMAGL proteins (0.1 μg) were preincubated with a PMSF inhibitor for 10 min at 37 °C in assay buffer (50 mM sodium phosphate buffer, pH 8.0, or glycine–NaOH buffer, pH 9.0). Linoleoylglycerol (sn-1 MAG; M7640; Sigma-Aldrich) substrates were emulsified at 0.2 % Triton X-100 by sonication for 30 s with three repeats. MAG substrates (150 μM in final concentration) were added to the preincubated samples, and the reaction mixtures were incubated for an additional 5 min at 30°C. The released NEFA products were measured with the NEFA-HR colorimetric method assay kit, using Synergy H1 Hybrid Readers at 546 nm.

Prediction of 3-dimensional protein structures

The crystal structure of a human MAGL (PDB ID: 3HJU; http://www.rcsb.org/pdb/) was used as a template for the computational modeling. The modeling of AtMAGLs and their mutated proteins was performed using the Protein Homology/analogY Recognition Engine V 2.0 program (PHYRE2; http://www.sbg.bio.ic.ac.uk/phyre2). Three-dimensional (3D) protein graphic images were produced using the UCSF Chimera version 1.10.2 program (http://www.cgl.ucsf.edu/chimera).

Results and discussion

Phylogenetic relationships of MAGLs between Arabidopsis and other species

Given that there have been few studies on plant MAGLs, we compared the phylogenetic relationship between Arabidopsis and other species MAGLs. After obtaining the amino acid sequences of human, mouse, rat (Rattus norvegicus), and yeast (Saccharomyces cerevisiae) MAGL genes from NCBI (http://www.ncbi.nlm.nih.gov), we conducted multiple alignments on them with the amino acid sequences of AtMAGL6, AtMAGL8, and AtMAGL16 using ClustalW (http://www.genome.jp/tools/clustalw/) and subsequently generated their phylogenetic tree using MEGA5.2 (Tamura et al. 2011). In Fig. 1A, the catalytic triad was conserved in MAGLs from all species tested, while the GxSxG lipase motif was shown in every MAGL except AtMAGL16. In addition, the phylogenetic tree in Fig. 1B showed that yeast MAGL branched into mammalian and plant MAGLs. The amino acid sequences of mouse and rat MAGLs have approximately 92 % identity, while the former and the latter have approximately 84 and 84 % identity with human MAGL, respectively. Arabidopsis MAGL6 and MAGL8 have approximately 74 % identity with each other, whereas they showed relatively low identity (34–35 %) with AtMAGL16.

Fig. 1

Amino acid sequence alignment and phylogenetic relationship of MAGLs from Arabidopsis and other species. The deduced amino acid sequences of seven MAGL genes were aligned by CLUSTALX version 2.1. The accession numbers of the aligned MAG lipases in Arabidopsis database (http://www.arabidopsis.org) or GenBank (http://www.ncbi.nlm.nih.gov/) are as follows: AtMAGL6; At2g39400, AtMAGL8; At2g39420, AtMAGL16; At5g19290, human (Homo sapiens) MAGL; AAH06230.1, mouse (Mus musculus) MAGL; CAC69874.1, rat (Rattus norvegicus) MAGL; ALL87453.1 and yeast (Saccharomyces cerevisiae) MAGL; CAA81932.1. The conserved and identical amino acid residues are shaded in gray and black, respectively. (A) The GxSxG and SxSxG motifs are shown in red box. The amino acid residues of the catalytic triad, Ser, Asp, and His, are marked by asterisks. (B) The phylogenetic tree was generated from alignments of deduced amino acid sequences of MAGLs from Arabidopsis and other species, using the maximum likelihood method based on the WAG model, and gamma distributed with invariant sites (I) in the MEGA program (version 5.2; Tamura et al. 2011). The bootstrap value percentages of 500 replicates are shown at the branching points (Felsenstein 1985). The scale bar represents the distance unit between sequence pairs

Computational modeling of Arabidopsis MAGLs and their mutated proteins

To investigate the tertiary structure of the narrow region surrounding the GxSxG motif in Arabidopsis MAGLs, computational modeling of Arabidopsis MAGL and their mutated proteins was carried out using HsMAGL protein structure (Labar et al. 2010) as a template and the PHYRE2 program. The tertiary protein structure images were visualized by the UCSF Chimera program. Figure 2A and B show the ribbon structures of HsMAGL and AtMAGL8 proteins harboring the GxSxG motif and Asp (D) and His (H) residues. Through a cross-section of 3-dimensional AtMAGL6 and AtMAGL8 structures, a U-shaped nucleophilic elbow structure around the nucleophilic serine residue in the GxSxG motif is highlighted (Fig. 2C, G). When the GxSxG motif in AtMAGL6 and AtMAGL8 proteins was replaced with the SxSxG, GxAxG, or GxSxS motif, the shape of the nucleophilic elbow around the nucleophilic serine changed to a wider W shape (Fig. 2D–F, H–J). In addition, when the SxSxG motif of AtMAGL16 was substituted with the GxSxG motif, the size of the wide nucleophilic elbow reduced, but did not completely recover to the U-shaped nucleophilic elbow structure (Fig. 2K, L). Therefore, the computational modeling of Arabidopsis MAGLs and their mutated proteins also indicates that the GxSxG motif might exert effects on the enzymatic activity of Arabidopsis MAGLs.

Fig. 2

Comparison of three-dimensional protein models between Arabidopsis MAGLs and their mutated forms. (A, B) Three-dimensional protein models of (A) HsMAGL, and (B) AtMAGL proteins, which are represented by the catalytic triad, the Ser (S) in the GxSxG motif, His (H), and Asp (D). (C–L) Transverse sections for environments near the nucleophilic elbow serine residue in the catalytic triad of Arabidopsis MAGLs and their mutated proteins. The U- or W-shaped lines along the nucleophilic serine residues are shown in red. (C) AtMAGL8 having the GxSxG motif, and mutated AtMAGL8 containing (D) G117S, (E) S119A, or (F) G121S in the GxSxG motif. (G) AtMAGL6 harboring the GxSxG motif, and mutated AtMAGL6 having (H) G111S, (I) S113A, or (J) G115S in the GxSxG motif. (K) AtMAGL16 having the SxSxG motif, and (L) mutated AtMAGL16 containing S139G in the SxSxG motif

Inhibition of enzyme activities of AtMAGL6 and 8 in the presence of PMSF, which is a serine protease inhibitor

To understand the role of the GxSxG motif in the activities of Arabidopsis MAGLs, we investigated the effect of a serine protease inhibitor, PMSF, on the lipase activities of AtMAGL6 and AtMAGL8 (Muccioli et al. 2008). Figure 3A shows that PMSF is able to produce an irreversible MAGL–PMSF adduct and hydrofluoric acid (HF), by specifically binding to the hydroxyl group of the serine residue in the active site of the serine protease, thereby inhibiting its enzymatic activity (Han et al. 2012). MBP, MBP:AtMAGL6, MBP:AtMAGL8, and MBP:HsMAGL proteins were expressed in E. coli and purified using amylose resin. Each purified protein was incubated at different concentrations of PMSF and further incubated in the reaction buffer with MAG substrates. Finally, the amount of non-esterified fatty acid (NEFA) products was measured using the NEFA assay kit. As a result, the half maximal inhibitory concentration (IC₅₀) values for MBP:HsMAGL, MBP:AtMAGL6, and MBP:AtMAGL8 proteins in response to PMSF were calculated to be 3.30, 2.30, and 2.35, respectively, indicating that Arabidopsis MAGLs have a similar inhibition rate to that of human MAGL at 10 times higher PMSF concentration (Fig. 3B). These results suggest that the serine residue present in the active site of Arabidopsis and human MAGLs is important for MAGL lipase activity.

Fig. 3

Effect of a serine protease inhibitor, PMSF on the activities of maltose-binding protein (MBP) and the recombinant MBP:HsMAGL, MBP:AtMAGL6, and MBP:AtMAGL8 proteins. (A) A proposed mechanism for covalent inactivation of a MAGL protein by a PMSF inhibitor. HF Hydrofluoric acid. (B) Dose-dependent inhibition of MAGLs by a PMSF inhibitor. MBP, MBP:HsMAGL, MBP:AtMAGL6, and MBP:AtMAGL8 proteins (0.1 μg) were preincubated with a PMSF inhibitor (10⁻⁷–10⁻¹ mM) for 10 min at 25 °C in lipase assay buffer (Kim et al. 2016). Following preincubation, emulsified MAG substrates containing an 18:2 fatty acid at the sn-1 position were added, and incubated for an additional 5 min at 30 °C. The values for MAGL activity are an average of three independent experiments ± standard errors

With respect to the activity of recombinant HsMAGL proteins, the IC₅₀ values in response to PMSF were 3.30 and 3.20, when 7-HCA (7-hydroxycoumarinyl arachidonate) and 4-NPA (4-nitrophenylacetate) substrates, respectively, were used (Muccioli et al. 2008; Savinainen et al. 2010). Rat MAG lipase had an IC₅₀ value of 3.81 in response to PMSF, when 2-arachidonoylglycerol (2-AG) was used as a substrate (Saario et al. 2004). The present study also confirmed that even though a MAG substrate was utilized, the IC₅₀ value of MBP:HsMAGL in response to PMSF was similar to the existing values.

Effect of amino acid substitutions on Arabidopsis MAG lipase activity

The computational modeling of Arabidopsis MAGLs and the inhibition of enzymatic activity of Arabidopsis MAGLs in response to PMSF strongly prompted us to examine the significance of each amino acid residue in the GxSxG motif of Arabidopsis MAGLs in their enzymatic activity. Thus, the GxSxG motif of AtMAGL6 and 8 was replaced with the SxSxG, GxAxG, or GxSxS motif. In addition, the first serine residue in the SxSxG motif of AtMAGL16 was substituted with a glycine residue to investigate if very low lipase activities of AtMAGL16 are caused by the presence of the SxSxG motif instead of the GxSxG motif (Fig. 4). After the transformation of all recombinant vectors in E. coli, the induced proteins were purified and electrophoresed on 12 % SDS-PAGE. Approximately 43 kDa of MBP and approximately 72 kDa of MBP:AtMAGL6, MBP:AtMAGL6_G111S, MBP:AtMAGL6_S113A, MBP:AtMAGL6_G115S, MBP:AtMAGL8, MBP:AtMAGL8_G117S, MBP: AtMAGL8_S119A, MBP: AtMAGL8_G121S, MBP:AtMAGL16, and MBP: AtMAGL16_S139G were identified (Fig. 5). Lipase activities of the purified proteins were measured using the NEFA assay kit, when the MAG substrates containing 18:2 fatty acids at the sn-1 position were supplemented.

Fig. 4

Schematic diagrams of expression vectors harboring the recombinant MBP:MAGLs and their mutated proteins

Fig. 5

SDS-polyacrylamide gel electrophoresis of the purified MBP, MBP:MAGLs, and mutated MBP:MAGLs. MBP and MBP:MAGLs were purified from E. coli, electrophoresed on a 12 % SDS-PAGE gel, and stained with Coomassie blue R-250. MBP and MBP:MAGLs are indicated by black arrowheads. M Molecular weight standard [Fermentas, kilodalton (kDa)]; MBP ~43 kDa; 1 MBP:AtMAGL6; 2 MBP:AtMAGL6_G111S; 3 MBP:AtMAGL6_S113A; 4 MBP:AtMAGL6_G115S; 5 MBP:AtMAGL8; 6 MBP:AtMAGL8_G117S; 7 MBP: AtMAGL8_S119A; 8 MBP: AtMAGL8_G121S; 9 MBP:AtMAGL16; 10 MBP: AtMAGL16_S139G

In agreement with the findings of Kim et al. (2016), the lipase activities of AtMAGL6 and AtMAGL8 were observed to be 21.8 and 19.3 μmol mg⁻¹ min⁻¹, respectively, but no lipase activity was observed in six types of mutated proteins (Table 2). This result indicates that two glycine residues and a serine residue in the GxSxG motif present in AtMAGL6 and AtMAGL8 are integral to the lipase activity. Although in the case of AtMAGL16 the SxSxG motif was changed into the GxSxG, no significant changes in the lipase activity were observed (Table 2), suggesting that AtMAGL16 may be an enzyme that degrades other substrates, not MAG.

Table 2 Effect of amino acid substitutions on MAG lipase activity

To date, little is known about alterations in enzyme activities by a point mutation of the amino acid residues in the GxSxG motif present in the MAG lipases. However, the essential role of GxSxG motif existing in various types of lipases has been reported (Kurat et al. 2006; Rabin and Hauser 2005; Wada et al. 2009). As evidenced, the complete deletion of the GxSxG motif in mouse phospholipase A₂ (PLA₂) almost eliminated its enzymatic activity (Wada et al. 2009). Also, mutating the serine residue in the GxSxG motif of Pseudomonas aeruginosa patatin-like phospholipase and yeast triglyceride lipase 4 with an alanine residue almost demolished their lipase activities (Rabin and Hauser 2005; Kurat et al. 2006).

In conclusion, very few studies have been reported about plant MAG lipases, because their functions could not be inferred from amino acid sequence similarity. In the current study, we revealed that glycine residues, as well as a serine residue within the GxSxG motif in Arabidopsis MAGL6 and MAGL8, are critical for MAG lipase activity. Although AtMAGL16 was mutated to contain the GxSxG motif in its active site, its MAG lipase activity was not significantly increased, suggesting that AtMAGL16 may not be a MAG lipase. Taken together, this study provides information about the essential motif of plant MAGLs, which among the genes involved in plant lipid metabolism have been least studied (McGlew et al. 2015).