Phospholipase A2 group IVD mediates the transacylation of glycerophospholipids and acylglycerols

Abstract

In mammalian cells, glycerolipids are mainly synthesized using acyl-CoA-dependent mechanisms. The acyl-CoA-independent transfer of fatty acids between lipids, designated as transacylation reaction, represents an additional mechanism for lipid remodeling and synthesis pathways. Here, we demonstrate that human and mouse phospholipase A2 group IVD (PLA2G4D) catalyzes transacylase reactions using both phospholipids and acylglycerols as substrates. In the presence of monoglycerol and diacylglycerol (MAG and DAG), purified PLA2G4D generates DAG and triacylglycerol, respectively. The enzyme also transfers fatty acids between phospholipids and from phospholipids to acylglycerols. Overexpression of PLA2G4D in COS7 cells enhances the incorporation of polyunsaturated fatty acids into triacylglycerol stores and induces the accumulation of lysophospholipids. In the presence of exogenously added MAG, the enzyme strongly increases cellular DAG formation, while MAG levels are decreased. PLA2G4D is not or poorly detectable in commonly used cell lines. It is expressed in keratinocytes, where it is strongly upregulated by proinflammatory cytokines. Pla2g4d-deficient mouse keratinocytes exhibit complex lipidomic changes in response to cytokine treatment, indicating that PLA2G4D is involved in the remodeling of the lipidome under inflammatory conditions. Transcriptomic analysis revealed that PLA2G4D modulates fundamental biological processes including cell proliferation, differentiation, and signaling. Together, our observations demonstrate that PLA2G4D has broad substrate specificity for fatty acid donor and acceptor lipids, allowing the acyl-CoA-independent synthesis of both phospholipids and acylglycerols. Loss-of-function studies indicate that PLA2G4D affects metabolic and signaling pathways in keratinocytes, which is associated with complex lipidomic and transcriptomic alterations.

Keywords: PLA2G4D; cPLA2δ; diacylglycerol; enzymolgy/enzyme mechanisms; glycerolipids; monoacylglycerol; phospholipases; phospholipids/metabolism; transacylase; triacylglycerol.

Fig. 1

PLA2G4D acts as phospholipase with broad head group specificity. (A) Schematic illustration of the domain architecture of the mPLA2G4 family showing C2 domains (C2), the patatin-like domain (patatin/cPLA2), and the position of the serine (S) – aspartate (D) catalytic dyad. (B, C) (Lyso)phospholipase activity of partially purified murine PLA2G4D and the mS370A mutant in the absence and presence of 1 mM CaCl₂. FFA release was determined using commercially available colorimetric kits. (D, E) (Lyso)phospholipase activity of partially purified human PLA2G4D in the absence and presence of 1 mM CaCl₂. All enzyme activity assays were carried out by incubating 1 μg of partially purified protein with 20 μl of lipid substrate (1 mM) in PBS (pH 7.4) containing 2% BSA (FA free) for 1 h at 37°C (n = 3). Data are presented as mean ± SD. Statistical comparison was performed with multiple unpaired two-tailed Student’s t test, followed by Bonferroni posthoc analysis. Statistically significant differences are shown as: ∗P < 0.05; ∗∗P < 0.01; and ∗∗∗P < 0.001. PLA2G4, phospholipase A2 group IV.

Fig. 2

PLA2G4D catalyzes the transfer of fatty acids between phospholipids. (A, C) TLC analysis of PLA2G4D-derived reaction products using sn-1-18:1 LPE and sn-1-18:1 LPC as substrates (1 mM) in the absence and presence of 1 mM CaCl₂. (B, D) Densitometric quantification of reaction products PE and PC is shown in A and C. (E–L) PLA2G4D-derived reaction products using an equimolar substrate mixture (0.5 mM each) of sn-1-16:0-sn-2-20:4 PC and sn-1-18:1 LPE, or sn-1-16:0-sn-2-20:4 PE and sn-1-18:1 LPC as donor and acceptor, respectively. The mS370A mutant was used as negative control (G, K). Lipids were analyzed by HPLC-MS. PE and PC subspecies are labeled according to their acyl-chain composition. All enzyme activity assays were carried out under conditions described in Figure 1 (n = 3). Data are presented as mean ± SD. Statistical comparison in (H, L) were performed with unpaired two-tailed Student’s t test, and in (B, D, E, F, I, J) with one-way ANOVA followed by Bonferroni posthoc analysis. Statistically significant differences are shown as: ∗P < 0.05; ∗∗P < 0.01; and ∗∗∗P < 0.001. LPC, lyso-phosphatidylcholine; LPE, lyso-phosphatidylethanolamine; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PLA2G4, phospholipase A2 group IV.

Fig. 3

PLA2G4D acylates mono- and diacylglycerol. (A) TLC analysis of m- and hPLA2G4D-derived reaction products incubated with rac-18:1 MAG (1 mM) in the absence and presence of 1 mM CaCl₂. The mS370A mutant was used as negative control. (B, C) Densitometric quantification of m- and hPLA2G4D-derived reaction products shown in (A). Data is presented as the relative ratio of products. (D) Schematic representation of MAG hydrolysis and transacylation reactions. (E) Time-dependent formation of glycerol and FFAs in the presence of mPLA2G4D using rac-18:1 MAG (1 mM) as substrate. The release of glycerol and FFAs was quantified using commercial kits. Linear regression was used to calculate the specific MMAT activity of the enzyme, based on the difference between glycerol and FFA release (insert). The dotted line indicates the 95% confidence interval (CI), and R² shows the coefficient of determination. (F) mPLA2G4D-derived reaction products using an equimolar substrate mixture of 2–18:1 MAG and 2-20:4 MAG (0.5 mM each). Assay was performed in the absence of 1 mM CaCl₂, and lipids were analyzed by HPLC-MS. DAG subspecies are labeled according to their acyl-chain composition. (G) Relative abundance of 18:1 and 20:4 acyl chains in total DAG, calculated from data shown in (F). (H) TLC analysis of mPLA2G4D-derived reaction products incubated with rac-18:1 MAG or 2-18:1 MAG (0.5 mM each) in the absence of CaCl₂. (I) Total DAG levels calculated as sum of sn-1,2 and sn-1,3-DAG band intensities shown in (H). (J) Abundance of sn-1,2 and sn-1,3-DAG isomers, based on band intensities shown in (H). (K–P) PLA2G4D-derived reaction products using an equimolar substrate mixture (0.5 mM each) of rac-18:1 MAG and sn-1-16:0-sn-2-20:4 PE, or rac-18:1 DAG and sn-1-16:0-sn-2-20:4 PE, respectively. Assays were performed in the absence of CaCl₂ and lipids were analyzed by HPLC-MS. All enzyme activity assays were carried out under conditions described in Figure 1 (n = 3). Data are presented as mean ± SD. Statistical comparisons in (F, K, L, N, O) were performed with one-way ANOVA followed by Bonferroni posthoc analysis and in (G, I, J) with unpaired two-tailed Student’s t test. Statistically significant differences are shown as: ∗P < 0.05; ∗∗P < 0.01; and ∗∗∗P < 0.001. DAG, diacylglycerol; MAG, monoacylglycerol; PE, phosphatidylethanolamine; PLA2G4, phospholipase A2 group IV; TAG, triacylglycerol.

Fig. 4

PLA2G4D is a cytosolic enzyme that interacts with membranes in a calcium-dependent manner. (A) Confocal imaging of COS7 cells stably overexpressing His-tagged mPLA2G4D. Cells were fixed with paraformaldehyde and stained with an anti-6xHis antibody (cyan). To increase cytosolic Ca²⁺ concentrations, cells were treated with thapsigargin/calcimycin A23187 (5 μM each) or DMSO for 15 min before fixation. Maximum intensity projection shows the highest pixel intensities from stacked images indicating that the enzyme shows primarily cytosolic localization. Zoomed sections from a single image plane suggest that PLA2G4D partially colocalizes with cytoplasmic and plasma membranes. Nuclei were stained with DAPI (blue). Scale bar: 40 μm. Data are representative for two independent experiments. (B) Western blotting analysis of total (1,000 g supernatant), cytosolic (100,000 g supernatant), and membrane fractions (100,000 g pellet) of COS7 cells overexpressing His-tagged mPLA2G4D. Cell fractionation was performed in the absence and presence of 2 mM CaCl₂ and 10 mM EDTA. The plasma membrane protein Cadherin and ER protein calnexin were used as markers for the membrane fractions. β-Actin and Coomassie staining served as loading control. Data are representative for two independent experiments. (C) Western blotting analysis of total fraction (TF; 1,000 g supernatant), heavy mitochondrial fraction (HMF; 3,000 g pellet), light mitochondrial fraction (LMF; 17,000 g pellet), microsomal fraction (MICRO; 100,000 g pellet), and cytosolic fraction (CYTO; 100,000 g supernatant) of COS7 cells overexpressing His-tagged mPLA2G4D. Cell fractionation was performed in the absence and presence of 2 mM CaCl₂. Lamp1, Calnexin, pan-Cadherin, and SDHA were used as marker-proteins for the endo-lysosomal compartment, ER, plasma membrane, and mitochondria, respectively. β-Actin served as loading control. PLA2G4, phospholipase A2 group IV.

Fig. 5

PLA2G4D overexpression alters cellular acylglycerol and phospholipid profiles. (A, B) Volcano plots showing changes in glycerophospholipids in m- and hPLA2G4D-expressing COS7 cells compared to control cells expressing the inactive mS370A mutant. The x-axis shows the log₂ (fold change) of the indicated phospholipid subspecies. The y-axis shows the -log₁₀(P value). Lipids were analyzed by HPLC-MS 48 h posttransfection. (C, D) Total LPC and LPE levels calculated based on the sum of signals from all detected subspecies. (E, F) Lipid subspecies of LPC and LPE in COS7 cells expressing m-, hPLA2G4D, or the mS370A mutant. (G, H) Volcano plots showing changes of MAG, DAG, and TAG subspecies in m- and hPLA2G4D-expressing cells compared to control cells expressing the mS370A mutant. (I–K) Total MAG, DAG, and TAG levels calculated based on the sum of signals from all detected subspecies. Data are shown as mean ± SD (n = 5) and are representative of three independent experiments. Statistical comparison in (A, B, E, F, G, H) was performed with multiple unpaired two-tailed Student’s t test followed by Bonferroni posthoc analysis. Large-sized spots in volcano plots represent lipids with an adjusted P value < 0.05. Statistical comparisons in (C, D, I, J, K) were performed with one-way ANOVA followed by Bonferroni posthoc analysis. Statistically significant differences are shown as: ∗, #P < 0.05; ∗∗, ###P < 0.01; and ∗∗∗, ###P < 0.001. DAG, diacylglycerol; LPC, lyso-phosphatidylcholine; LPE, lyso-phosphatidylethanolamine; MAG, monoacylglycerol; PLA2G4, phospholipase A2 group IV; TAG, triacylglycerol.

Fig. 6

PLA2G4D enhances the incorporation of MAG from extracellular sources into cellular acylglycerols independently of DGAT. (A) Representative TLC of acylglycerol levels of control (empty vector transfection) and PLA2G4D-expressing COS7 cells loaded with 300 μM rac-18:1 MAG for 210 min in the absence and presence of DGAT1 and DGAT2 inhibitors (5 μM each). (B, C) Densitometric quantification of acylglycerol bands in rac-18:1 MAG loaded cells in the absence and presence of DGAT inhibitors (shown in A). Experiments were performed in triplicates (n = 3) and data are presented as mean ± SD and are representative for two independent experiments. Statistical comparison in all graphs was performed with an unpaired two-tailed Student’s t test. Statistically significant differences are shown as: ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001. DGAT, diacylglycerol acyltransferasel; FC, free cholestero; MAG, monoacylglycerol; PLA2G4, phospholipase A2 group IV.

Fig. 7

PLA2G4D mRNA expression in mouse tissues and keratinocytes. (A) Pla2g4d mRNA expression normalized to 36b4 housekeeping (HK) gene in murine tissues. Ct values for Pla2g4d are shown above each bar. Tissues were analyzed in 2–4 biological replicates. (B) Pla2g4d mRNA expression normalized to 36b4 HK gene in primary keratinocytes and the human HaCaT keratinocyte cell line (n = 3). Primary cells were isolated from the epidermis of two-day-old neonatal mice and cultured for four days in the presence of low CaCl₂ concentrations (0.06 mM). Keratinocyte differentiation was induced by increasing CaCl₂ to 0.2 mM for 24 h. Before RNA isolation, primary and HaCaT keratinocytes were cultured in absence and the presence of IL17A and/or TNFα (20 ng/ml each) for 24 h. Ct values above 30 were considered as no or negligible expression for Pla2g4d (nd). Data are shown as mean ± SD. Statistical comparisons in (B) were performed with one-way ANOVA followed by Bonferroni posthoc analysis. Statistically significant differences are shown as: ∗P < 0.05; ∗∗P < 0.01; and ∗∗∗P < 0.001. BAT, brown adipose tissue; CM, cardiac muscle; iWAT, inguinal white adipose tissue; gWAT, gonadal white adipose tissue; SM, skeletal muscle; SI, small intestine; PLA2G4, phospholipase A2 group IV.

Fig. 8

Pla2g4d-deficient keratinocytes show complex lipidomic changes. (A) Principal component analysis of untargeted HPLC-MS lipidomics data from primary keratinocytes isolated from two-day-old WT (n = 5) and Pla2g4d-deficient (KO; n = 8) neonatal mice. Cells were cultured for four days under low CaCl₂ concentrations (0.06 mM) and subsequently stimulated with IL17A and TNFα (20 ng/ml each) for 24 h. (B–D) Total lipid class levels of basal and IL17A/TNFα treated WT and KO keratinocytes. Changes in lipid classes were calculated based on the sum of all analyzed subspecies. Based on signal intensities, classes are grouped in high, medium, and low abundance. (E) Heat maps displaying the z-score of PE, PPE, EtherPE, EtherPC, HexCER, CER, and SM subspecies were labeled by the number of C-atoms in fatty acids and the total number of double bonds. The z-score was calculated as number of SDs above or below from the mean of the dataset. Statistical comparisons were performed with multiple unpaired two-tailed Student’s t test with correction for multiple comparisons using the false discovery rate (FDR). Statistically significant differences are shown as: ∗, #P < 0.05; ∗∗, ##P < 0.01; and ∗∗∗, ###P < 0.001. CER, ceramide; CL, cardiolipin; LNAPE, lyso-N-acyl-PE; HexCER, hexosylceramide; PE, phosphatidylethanolamine; PLA2G4, phospholipase A2 group IV; PPE, plasmalogen-PE; SM, sphingomyelin.

Fig. 9

RNA-sequencing of primary WT and Pla2g4d-deficient (KO) keratinocytes. Primary keratinocytes were isolated from two-day-old WT (n = 5) and KO (n = 8) neonatal mice and cultured for four days under low CaCl₂ concentrations (0.06 mM). Subsequently, WT and KO cells were stimulated with IL17A and TNFα (20 ng/ml each) for 24 h. (A) “GO biological process” gene set enrichment analysis (GSEA) comparing KO and WT cells in the absence of cytokines (basal). The top 20 activated and 7 suppressed gene sets are shown. The “gene ratio” (x-axis) represents the ratio of core enriched genes (genes before or after the point at which the running enrichment score reaches its maximum or minimum; see B–D and H–J) to the total number of genes in the gene set. The “count” (dot size) reflects the total number of genes in the gene set and the color coding highlights the FDR-adjusted P value of the gene set. (B–D) Gene set enrichment plots of GO molecular function gene sets “Cytokine activity” and “chemokine activity,” as well as the GO biological process gene set “lipid catabolic process” comparing KO and WT cells in the absence of IL17A and TNFα (basal). The plots display the GSEA statistics, including the “normalized enrichment score (NES)” and FDR-adjusted P value. (E, F) Selected core enriched genes from gene sets shown in (B–D). Differential gene expression between KO and WT cells is visualized in a heat map, displaying the z-score (indicates number of SDs a data point is from the mean of the dataset). Additionally, the fold change (KO/WT) and FDR-adjusted P value are shown. (G) “GO biological process” GSEA comparing KO and WT cells in the presence of IL17A and TNFα. The top 20 activated and top 20 suppressed gene sets are shown. (H–J) Gene set enrichment plots of GO biological process gene sets “nuclear division,” “keratinocyte differentiation,” and “lipid catabolic process” comparing KO and WT cells in the presence of IL17A and TNFα. (K, L) Selected core enriched genes from gene sets shown in (I, J). Differential expression between KO and WT cells is visualized in a heat map, displaying the z-score. Additionally, the fold change (KO/WT) and FDR-adjusted P value are shown. Statistically significant differences are shown as: ∗P < 0.1; ∗∗P < 0.01; and ∗∗∗P < 0.001. FDR, false discovery rate; PLA2G4, phospholipase A2 group IV.