Previous studies have demonstrated that membrane tubule-mediated transport events in biosynthetic and endocytic routes require phospholipase A2 (PLA2) activity. Here, we show that cytosolic phospholipase A2ε (cPLA2ε, also known as PLA2G4E) is targeted to the membrane compartments of the clathrin-independent endocytic route through a C-terminal stretch of positively charged amino acids, which allows the enzyme to interact with phosphoinositide lipids [especially PI(4,5)P2] that are enriched in clathrin-independent endosomes. Ablation of cPLA2ε suppressed the formation of tubular elements that carry internalized clathrin-independent cargoes, such as MHC-I, CD147 and CD55, back to the cell surface and, therefore, caused their intracellular retention. The ability of cPLA2ε to support recycling through tubule formation relies on the catalytic activity of the enzyme, because the inactive cPLA2εS420A mutant was not able to recover either tubule growth or transport from clathrin-independent endosomes. Taken together, our findings indicate that cPLA2ε is a new important regulator of trafficking processes within the clathrin-independent endocytic and recycling route. The affinity of cPLA2ε for this pathway supports a new hypothesis that different PLA2 enzymes use selective targeting mechanisms to regulate tubule formation locally during specific trafficking steps in the secretory and/or endocytic systems.

To keep the identity and composition of the surface domains and the membrane intracellular organelles, eukaryotic cells develop transport mechanisms that allow precise delivery of the protein and lipid molecules to the specific target compartments. The membrane trafficking of newly synthesized or endocytosed molecules is executed by the membrane-bound organelles called transport carriers that emerge from a donor compartment, move through the cytosol and fuse with an acceptor compartment (Luini et al., 2005; Polishchuk et al., 2009).

Budding of transport carriers requires bending of the donor membrane, which can be achieved through several molecular mechanisms. Interaction with protein coat polymers, BAR domains, amphipathic wedges and the components of cytoskeleton machineries have been reported to deform the plain membrane domains into highly curved buds (Graham and Kozlov, 2010; McMahon and Gallop, 2005). However, induction of membrane curvature for biogenesis of transport carriers could be triggered through a local transfer and/or modification of lipids molecules, which are executed by either lipid-transfer proteins or lipid-modifying enzymes, respectively (Graham and Kozlov, 2010; Sprong et al., 2001).

Among the lipid-modifying enzymes, cytosolic phospholipase A2 (PLA2) has been hypothesized to support the budding of tubular transport carriers because these enzymes hydrolyze cylinder-shaped phospholipids into wedge-like lysophospholipids (Bechler et al., 2012; Ha et al., 2012). This change in lipid geometry favors the generation of positive membrane curvature, which is required for budding of tubular carriers (Bechler et al., 2012; Ha et al., 2012; San Pietro et al., 2009). Although PLA2 activity is required for the transport of both newly synthesized and endocytosed proteins, which PLA2 enzymes operate in regulation of specific traffic events in either biosynthetic or endocytic routes remains poorly understood. Over the past few years, two cytosolic PLA2 enzymes have been reported to regulate the formation of tubular elements that operate in different transport steps. Cytosolic Ca2+-dependent PLA2α (cPLA2α, also known as PLA2G4A) operates in intra-Golgi trafficking (San Pietro et al., 2009), whereas cytoplasmic PLA2 enzyme complex platelet-activating factor acetylhydrolase (PAFAH) Ib regulates protein export from the Golgi at the TGN level (Bechler et al., 2010) and the distribution and function of endosomes (Bechler et al., 2011). Therefore, it is likely that different PLA2 enzymes operate in various specific tubule-mediated transport steps. As a consequence, each individual compartment of either the secretory or endocytic system must acquire its own PLA2 enzyme to generate tubular transport carriers when needed.

In the endocytic system, a wide range of transport events rely on tubular membrane structures (Bonifacino and Rojas, 2006; Cullen, 2008; Polishchuk et al., 2009) and therefore might require specific PLA2 enzymes. Indeed, PLA2 activity has been demonstrated to support endosome tubulation and recycling of endocytosed material back to the cell surface (Bechler et al., 2011; de Figueiredo et al., 2001; Doody et al., 2009). Moreover, identification of several novel PLA2 isoforms revealed their association with the endo/lysosomal system (Ghosh et al., 2006a; Ohto et al., 2005). However, the role of these recently identified PLA2 enzymes in transport through well-defined endocytic and recycling pathways remains elusive.

We show here that cPLA2ε binds to the membranes of the clathrin-independent transport pathway that is involved in the uptake and recycling of specific cargoes, such as major histocompatibility complex class I (MHC-I) proteins. This association relies on the C-terminal stretch of positively charged amino acids that binds to several phosphoinositide phosphate (PIP) lipids. Silencing of cPLA2ε significantly inhibits the formation of tubular elements involved in the recycling of internalized MHC-I and therefore induces intracellular accumulation of MHC-I. The capacity of cPLA2ε to support recycling of MHC-I through tubule formation requires catalytic activity of the enzyme because the inactive cPLA2ε mutant does not rescue MHC-I transport. Taken together, our findings indicate that cPLA2ε is an important player in transport events within the clathrin-independent endocytic or recycling route and support the hypothesis that specific targeting mechanisms of PLA2 enzymes to different intracellular compartments could be needed to regulate tubule formation in diverse segments of secretory and endocytic pathways.

PLA2 activity has been reported to support the formation of membranous tubular elements operating in the endocytic system (de Figueiredo et al., 2001; Doody et al., 2009). We thus sought to identify the PLA2 isoform(s) involved in the regulation of transport events within the endocytic pathway. The superfamily of PLA2 enzymes consists of 16 groups (Dennis et al., 2011) comprising secretory and cytosolic enzymes, with the latter divided into Ca2+-sensitive (cPLA2s, or group IV PLA2s) and Ca2+-insensitive (group VI, VII and VIII PLA2) isoforms (Schaloske and Dennis, 2006). Among those that are Ca2+ sensitive, cPLA2ε has been reported to associate with the membranes of the endo/lysosomal system (Ohto et al., 2005). This affinity for the membranes of endocytic compartments prompted us to examine whether cPLA2ε might itself be involved in tubule formation and trafficking through the specific endocytic route.

cPLA2ε resides on the membranes of the clathrin-independent endocytic/recycling route

Given that the association of cPLA2ε with any specific endocytic/recycling pathway has not yet been investigated in detail, we analyzed the intracellular distribution of the cPLA2ε using a battery of the markers that label different endocytic compartments and pathways. HeLa cells transfected with cPLA2ε–GFP exhibited a clear fluorescent signal at the plasma membrane as well as along a number of intracellular tubular (Fig. 1, arrows) and vacuolar structures (Fig. 1, arrowheads). Such tubular elements decorated by cPLA2ε–GFP were surprisingly long (up to 30 µm) and resembled, from a morphological standpoint, tubular structures involved in the recycling of cargoes internalized through the clathrin-independent endocytic route driven by ARF6 GTPase (Naslavsky et al., 2003; Weigert et al., 2004). This transport route (referred to here as the clathrin-independent pathway) has been shown to operate the internalization and recycling of many clathrin-independent cargo proteins such as MHC-I, β-integrins and the GPI-anchored protein CD59, for example (Donaldson et al., 2009; Grant and Donaldson, 2009).

Fig. 1.

Intracellular localization of cPLA2ε. HeLa cells were transfected with either Myc–cPLA2ε (A,B,F,G,H) or GFP–cPLA2ε (C–E). ARF6-HA (B) and EHD1–Myc (C) were co-transfected with cPLA2ε constructs. Cells were then fixed and labeled with anti-Myc antibody to reveal cPLA2ε (A,B,F,G,H) or EHD1 (C), anti-HA antibody to label ARF6 (B) and antibodies against endogenously-expressed MHC-I (A), EEA1 (D), TfR (E), MPR (F), Lamp1 (G) and golgin-97 (H). After immunofluorescent labeling, the cells were analyzed under confocal microscope to evaluate cPLA2ε colocalization with the above markers. Arrows and arrowheads in A–H indicate cPLA2ε-positive tubules and vacuoles, respectively. (I) Morphometric analysis shows the percentage (means ± s.d., n = 30 cells) of either vacuolar or tubular cPLA2ε structures that exhibit overlap with other markers indicated. Scale bar: 5 µm.

Fig. 1.

Intracellular localization of cPLA2ε. HeLa cells were transfected with either Myc–cPLA2ε (A,B,F,G,H) or GFP–cPLA2ε (C–E). ARF6-HA (B) and EHD1–Myc (C) were co-transfected with cPLA2ε constructs. Cells were then fixed and labeled with anti-Myc antibody to reveal cPLA2ε (A,B,F,G,H) or EHD1 (C), anti-HA antibody to label ARF6 (B) and antibodies against endogenously-expressed MHC-I (A), EEA1 (D), TfR (E), MPR (F), Lamp1 (G) and golgin-97 (H). After immunofluorescent labeling, the cells were analyzed under confocal microscope to evaluate cPLA2ε colocalization with the above markers. Arrows and arrowheads in A–H indicate cPLA2ε-positive tubules and vacuoles, respectively. (I) Morphometric analysis shows the percentage (means ± s.d., n = 30 cells) of either vacuolar or tubular cPLA2ε structures that exhibit overlap with other markers indicated. Scale bar: 5 µm.

The HeLa cell line has been established as a model system to investigate transport through the clathrin-independent endocytic/recycling pathway (Donaldson et al., 2009; Grant and Donaldson, 2009). Therefore, to test the association of cPLA2ε with this trafficking route, HeLa cells expressing cPLA2ε–GFP were stained with antibodies against MHC-I. Confocal microscopy revealed substantial overlap of MHC-I and cPLA2ε at the cell surface, as well as along intracellular tubular and vacuolar elements (Fig. 1A). The same colocalization pattern of cPLA2ε was also detected for ARF6 (Fig. 1B) and other cargo following this clathrin-independent route, such as CD147 (Eyster et al., 2009). Tubular elements highlighted by cPLA2ε also contained Rab22 and EHD1 (Fig. 1C), which have both been reported to operate in the recycling of clathrin-independent cargoes (Caplan et al., 2002; Weigert et al., 2004). In contrast, markers of the other endocytic compartments and Golgi residents exhibited significantly lower overlap with cPLA2ε (Fig. 1D–I). The use of different tags (GFP, Myc or mCherry) affected neither the intracellular distribution pattern of cPLA2ε nor the extent of its overlap with each of the above markers. Notably, strong overexpression of cPLA2ε induced tubulation of clathrin-independent endocytic compartments labeled with MHC-I, whereas the overall distribution of other markers of early and late endocytic organelles remained unaffected (supplementary material Fig. S1). Thus, taken together, colocalization experiments suggest specific association of cPLA2ε with the membrane compartments of the clathrin-independent endocytic/recycling pathway.

To further prove this point, we loaded cells with different endocytic tracers that highlight several specific uptake and recycling pathways. Internalization of fluorescent anti-MHC-I antibody was used to label the clathrin-independent route (Naslavsky et al., 2003), whereas uptake of transferrin or EGF was used to reveal clathrin-dependent recycling or degradative pathways, respectively. Again, preferential targeting of cPLA2ε to the membrane compartments loaded with endocytosed MHC-I was detected (supplementary material Fig. S2A), whereas the enzyme association with endocytosed EGF and transferrin receptor was quite limited (supplementary material Fig. S2B,C). Finally, we tested whether cPLA2ε was exclusively associated with the distal (recycling) compartments of the clathrin-independent route (as revealed by colocalization with Rab22 and EHD1) or could also be detected at the early incoming clathrin-independent endocytic structures. To this end, we used expression of ARF6Q67L as a tool to arrest clathrin-independent trafficking at the level of very early clathrin-independent endosomes, which appear as large and easily detectable vacuoles in cells (Naslavsky et al., 2003). Confocal microscopy revealed cPLA2ε at the surface of such giant ARF6Q67L endosomes (supplementary material Fig. S2D; see also Fig. 3D), suggesting that the enzyme is also associated with early (incoming) membranes of the clathrin-independent pathway.

The intracellular distribution of cPLA2ε was further investigated using immunoelectron microscopy. To better understand the identity of the structures containing cPLA2ε, cells were loaded with anti-MHC-I antibody to label membranes of the clathrin-independent pathway. Double immunogold staining revealed both cPLA2ε and MHC-I along the plasma membrane (Fig. 2A) and in the endosomes that contained few (<9) intraluminal vesicles (Fig. 2B) and therefore could be qualified as ‘early endosomes’ on the basis of their ultrastructural features (Saftig and Klumperman, 2009). Association of cPLA2ε with the early clathrin-independent endocytic compartment was further confirmed because the enzyme was detected at the surface of the early clathrin-independent endosomes that accumulated in the cells upon overexpression of ARF6Q67L (Fig. 2C). We also found that the subpopulation of cPLA2ε-positive structures contained EEA1 (Fig. 2D), an early endosomal marker, which binds to the surface of clathrin-independent endosomes during their maturation and shortly before their convergence with the clathrin-dependent endocytic pathway (Naslavsky et al., 2003). In addition to typical vacuolar endocytic structures, cPLA2ε decorated 50- to 200-nm-thick tubular elements that contained internalized MHC-I (Fig. 2E) and resembled, from the ultrastructural standpoint, clathrin-independent recycling tubules described previously (Caplan et al., 2002). cPLA2ε labeling was quite specific, because the late endosomes with numerous intraluminal vesicles (Fig. 2F, arrowheads) and the Golgi membranes (Fig. 2D, asterisk), exhibited little or no cPLA2ε at their surface (see also quantification in Fig. 2G). Such a distribution of the cPLA2ε labeling was consistent with poor colocalization of the enzyme with either lysosomal or Golgi markers (see Fig. 1) in immunofluorescence, and indicates a specific association of cPLA2ε with incoming and recycling compartments of the clathrin-independent transport route.

Fig. 2.

Immuno-EM labeling reveals cPLA2ε association with compartments of the clathrin-independent pathway. HeLa cells were transfected with either Myc–cPLA2ε (A,B,E,F) or GFP–cPLA2ε (C,D). ARF6Q67L-HA (C) was co-transfected with cPLA2ε. Cells shown in A,B,E were also loaded with Alexa-Fluor-488-conjugated antibody against MHC-I for 1 hour to label the membranes of the clathrin-independent transport route. Then cells were fixed, frozen and sectioned as described in the Materials and Methods. Sections were labeled with either anti-Myc (A,B,E,F) or anti-GFP (C,D) antibody to reveal cPLA2ε, with anti-HA antibody to detect ARF6Q67L-HA (C), anti-Alexa-Fluor-488 antibody to reveal internalized MHC-I (A,B,E) and with anti-EEA1 to label endogenous EEA1 (D). Arrows and arrowheads indicate cPLA2ε and MHC-I, respectively, at the plasma membrane (A) and within the same early endosome (B). In C, cPLA2ε (arrows) and ARF6Q67L (arrowheads) were detected at the surface of the same endocytic structures. In D, arrows indicate cPLA2ε associated with structures decorated by EEA1 (arrowheads) while the membranes of Golgi stack (asterisk) lack cPLA2ε labeling. (E) Tubular structure with internalized MHC-I (arrowheads) exhibited cPLA2ε (arrow) at its surface. (F) cPLA2ε signal is associated with early endosomes (arrows) that contained few intraluminal vesicles whereas late endosome/lysosome-like structures remained unlabeled (arrowheads). (G). Morphometric analysis of cells labeled for both cPLA2ε and internalized MHC-I shows the percentage of cPLA2ε-associated gold particles in different compartments. MHC-I in parentheses indicates MHC-I-positive organelles. Scale bar: 120 nm (A–C), 180 nm (D–F).

Fig. 2.

Immuno-EM labeling reveals cPLA2ε association with compartments of the clathrin-independent pathway. HeLa cells were transfected with either Myc–cPLA2ε (A,B,E,F) or GFP–cPLA2ε (C,D). ARF6Q67L-HA (C) was co-transfected with cPLA2ε. Cells shown in A,B,E were also loaded with Alexa-Fluor-488-conjugated antibody against MHC-I for 1 hour to label the membranes of the clathrin-independent transport route. Then cells were fixed, frozen and sectioned as described in the Materials and Methods. Sections were labeled with either anti-Myc (A,B,E,F) or anti-GFP (C,D) antibody to reveal cPLA2ε, with anti-HA antibody to detect ARF6Q67L-HA (C), anti-Alexa-Fluor-488 antibody to reveal internalized MHC-I (A,B,E) and with anti-EEA1 to label endogenous EEA1 (D). Arrows and arrowheads indicate cPLA2ε and MHC-I, respectively, at the plasma membrane (A) and within the same early endosome (B). In C, cPLA2ε (arrows) and ARF6Q67L (arrowheads) were detected at the surface of the same endocytic structures. In D, arrows indicate cPLA2ε associated with structures decorated by EEA1 (arrowheads) while the membranes of Golgi stack (asterisk) lack cPLA2ε labeling. (E) Tubular structure with internalized MHC-I (arrowheads) exhibited cPLA2ε (arrow) at its surface. (F) cPLA2ε signal is associated with early endosomes (arrows) that contained few intraluminal vesicles whereas late endosome/lysosome-like structures remained unlabeled (arrowheads). (G). Morphometric analysis of cells labeled for both cPLA2ε and internalized MHC-I shows the percentage of cPLA2ε-associated gold particles in different compartments. MHC-I in parentheses indicates MHC-I-positive organelles. Scale bar: 120 nm (A–C), 180 nm (D–F).

The C-terminal cluster of basic amino acids defines targeting of cPLA2ε to the clathrin-independent endocytic/recycling pathway

Most of the enzymes from the cPLA2 family are thought to rely on the N-terminal Ca2+-binding (C2) domain for the membrane binding that occurs upon an increase of Ca2+ in the cytosol (Ghosh et al., 2006b). The specific properties of the given C2 domain have been reported to define the preference of its own protein (comprising cPLA2 enzymes) for specific membrane compartments (Evans et al., 2004). In this context, we wanted to check to what extent the Ca2+ and C2 domain contribute to cPLA2ε specificity for the clathrin-independent pathway. HeLa cells expressing cPLA2ε were treated with the Ca2+ chelator BAPTA to deplete cytosolic Ca2+. In parallel, Ca2+-sensitive cPLA2α was used as a control to check the effectiveness of BAPTA treatment.

Confocal microscopy revealed that a substantial pool of cPLA2ε could be detected at the plasma membrane and at intracellular structures in untreated cells (Fig. 3A). Incubation with BAPTA did not significantly change the intracellular pattern of cPLA2ε (Fig. 3A). This behavior of cPLA2ε upon depletion of Ca2+ was strikingly different from that of cPLA2α that was displaced completely from the Golgi into the cytosol after the addition of BAPTA (Fig. 3A). These observations argued against the essential role of the Ca2+ and C2 domain of cPLA2ε in the recruitment of this enzyme to clathrin-independent membranes.

Fig. 3.

cPLA2ε and cPLA2α show different sensitivity to a variation of cytosolic Ca2+ concentration. (A) HeLa cells were transfected with GFP–cPLA2α or GFP–cPLA2ε cDNAs (as indicated) and fixed directly (left column) or 30 minutes after addition of 20 µM Ca2+ chelator BAPTA (right column). Ca2+ reduction induced detachment of cPLA2α from the Golgi membranes into the cytosol but did not affect localization of cPLA2ε, which remains membrane associated. (B) Schematic representation of cPLA2ε protein and its C-terminal cluster of basic amino acids. The analysis of mouse cPLA2ε sequence revealed the presence of a polybasic (PB) cluster of four positively charged residues (red letters) at the end of C-terminal part of the enzyme. (C) HeLa cells expressing Myc–cPLA2εΔPB, GFP–cPLA2α and Myc–cPLA2ε cDNAs (as indicated on single panels) were fixed directly (left column) or 10 minutes after incubation with 10 µM Ca2+ ionophore ionomycin (right column) and stained with anti-Myc to reveal cPLA2εΔPB or cPLA2ε. cPLA2εΔPB exhibited a mostly cytosolic pattern even when ionomycin was added to the cells. In contrast, ionomycin significantly improved association of cPLA2α with perinuclear Golgi and ER membranes. Conversely, full-length cPLA2ε was displaced by ionomycin from the plasma membrane and intracellular tubular-vacuolar structures into the cytosol. (D) HeLa cells were co-transfected with either Myc–cPLA2ε and ARF6Q67L (top row) or Myc–cPLA2εΔPB and ARF6Q67L (bottom row). The cells were fixed, permeabilized and labeled with anti-ARF6 and anti-Myc antibodies. Overexpression of ARF6Q67L induced formation of large vacuoles (arrows) where cPLA2ε (top row) but not cPLA2εΔPB (bottom row) was recruited. (E) Quantification revealed a decrease in colocalization (means ± s.d., n = 30 cells) of cPLA2εΔPB with ARF6Q67L when compared with full-length cPLA2ε. Scale bar: 8 µm (A,C), 4 µm (D).

Fig. 3.

cPLA2ε and cPLA2α show different sensitivity to a variation of cytosolic Ca2+ concentration. (A) HeLa cells were transfected with GFP–cPLA2α or GFP–cPLA2ε cDNAs (as indicated) and fixed directly (left column) or 30 minutes after addition of 20 µM Ca2+ chelator BAPTA (right column). Ca2+ reduction induced detachment of cPLA2α from the Golgi membranes into the cytosol but did not affect localization of cPLA2ε, which remains membrane associated. (B) Schematic representation of cPLA2ε protein and its C-terminal cluster of basic amino acids. The analysis of mouse cPLA2ε sequence revealed the presence of a polybasic (PB) cluster of four positively charged residues (red letters) at the end of C-terminal part of the enzyme. (C) HeLa cells expressing Myc–cPLA2εΔPB, GFP–cPLA2α and Myc–cPLA2ε cDNAs (as indicated on single panels) were fixed directly (left column) or 10 minutes after incubation with 10 µM Ca2+ ionophore ionomycin (right column) and stained with anti-Myc to reveal cPLA2εΔPB or cPLA2ε. cPLA2εΔPB exhibited a mostly cytosolic pattern even when ionomycin was added to the cells. In contrast, ionomycin significantly improved association of cPLA2α with perinuclear Golgi and ER membranes. Conversely, full-length cPLA2ε was displaced by ionomycin from the plasma membrane and intracellular tubular-vacuolar structures into the cytosol. (D) HeLa cells were co-transfected with either Myc–cPLA2ε and ARF6Q67L (top row) or Myc–cPLA2εΔPB and ARF6Q67L (bottom row). The cells were fixed, permeabilized and labeled with anti-ARF6 and anti-Myc antibodies. Overexpression of ARF6Q67L induced formation of large vacuoles (arrows) where cPLA2ε (top row) but not cPLA2εΔPB (bottom row) was recruited. (E) Quantification revealed a decrease in colocalization (means ± s.d., n = 30 cells) of cPLA2εΔPB with ARF6Q67L when compared with full-length cPLA2ε. Scale bar: 8 µm (A,C), 4 µm (D).

This prompted us to further investigate whether any other specific targeting motif could be found in cPLA2ε. Sequence analysis of cPLA2ε revealed a stretch of basic amino acids (865KKRMR870) located almost at the end of its C-terminus (see Fig. 3B). Interestingly, such polybasic (PB) clusters have been reported in many proteins to allow plasma membrane binding through interaction with PI(4,5)P2 and/or PI(3,4,5)P3 (Heo et al., 2006). Thus, a DNA construct encoding truncated cPLA2ε (hereafter referred to as cPLA2εΔPB), which lacks the last 10 C-terminal amino acids comprising the basic cluster was prepared and its intracellular distribution was investigated. Deletion of the PB domain resulted in displacement of the protein from the membranes (Fig. 3C), indicating that this motif is the main determinant that defines cPLA2ε-specific targeting towards clathrin-independent compartments.

At this point we reasoned that regardless of the dominant role of the PB cluster in cPLA2ε targeting, the C2 domain could still be used by the enzyme for membrane binding under certain conditions (when the PB domain is inactive, for example). To test whether this is the case, cells expressing cPLA2εΔPB were treated with ionomycin, an ionophore that increases cytosolic Ca2+. Despite the presence of the C2 domain, cPLA2εΔPB did not exhibit any significant shift from the cytosol towards the plasma membrane or any intracellular membranes (Fig. 3C), whereas in control experiments, the effectiveness of ionomycin treatment was confirmed by a significant recruitment of Ca2+-sensitive cPLA2α to the Golgi and ER membranes and its reduction in the cytosol (Fig. 3C). Interestingly, addition of the ionomycin induced partial loss of the full-length cPLA2ε from the plasma membrane and some clathrin-independent membranes comprising recycling tubules (Fig. 3C). Such behavior of cPLA2ε was not expected for the protein with the C2 domain, but would have been likely for the protein with the PB motif. Given that the PB domain uses PI(4,5)P2 as the main ligand (Heo et al., 2006), the hydrolysis of this lipid by PLC upon addition of ionomycin (Heo et al., 2006) is likely to reduce binding of cPLA2ε to the membrane.

To further check whether PI(4,5)P2 is involved in cPLA2ε recruitment to the membranes we used ARF6Q67L as a tool to generate PI(4,5)P2-enriched membranes. As mentioned above, ARF6Q67L induces accumulation of giant clathrin-independent endosomes. Importantly, such ARF6Q67L-positive endosomes have been reported to contain high amounts of PI(4,5)P2 (Naslavsky et al., 2003). Confocal microscopy revealed cPLA2ε along the membranes of such ARF6Q67L vacuoles, whereas deletion of the PB domain resulted in dissociation of the enzyme from these PI(4,5)P2-enriched structures (Fig. 3D,E).

In the context of the above observations and the reported affinity of the PB domain for PI(4,5)P2 and PI(3,4,5)P3 (Heo et al., 2006), we checked the lipid-binding properties of cPLA2ε in a lipid overlay assay. To this end, recombinant GST-tagged cPLA2ε, its C-terminus (815–875 respectively) and cPLA2εΔPB were purified and spotted on PIP strips (see the Materials and Methods). In control experiments, GST alone did not interact with any lipid in the strips, whereas GST-fused probes for PI(3)P (p40-PX domain), PI(4)P (C-terminal fragment of Legionella pneumophila SidC protein) and PI(4,5)P2 (PLCδ-PH domain) bound corresponding lipids with high specificity (Fig. 4A). Furthermore, we found that both full-length cPLA2ε and its C-terminal peptide exhibited a substantial interaction with different phosphoinositides comprising PI(4,5)P2, whereas the cPLA2εΔPB did not show any significant affinity for any of the tested lipids, such as the negative GST control (Fig. 4A). Interestingly, cPLA2ε was found to interact with all PIPs at a similar magnitude, and in addition, exhibited a substantial affinity for phosphatidic acid (PA), which might be used by the enzyme as a substrate. However, the C-terminal PB domain of the cPLA2ε did not interact with PA and PI(3)P, and its binding to remaining PI mono-phosphates was significantly lower than to bis- and tris-phosphates. This suggests that the PB domain renders full-length cPLA2ε to a conformation that strengthens its interaction with PA and PI(3)P in vitro.

Fig. 4.

Association of cPLA2ε with phosphatidylinositol lipids. (A) PIP strip membranes containing spotted phosphatidylinositols and other lipids (as indicated in scheme) were incubated with 5 µg/ml of the following recombinant proteins: GST alone, GST–SidC-3C, GST–p40-PX, GST–PH-PLCδ, GST–cPLA2ε full-length, GST–cPLA2ε N-terminus (amino acids 1–63), GST–cPLA2ε C-terminus (815–875) and GST–cPLA2ε full-length fusion proteins. Then the membranes were washed and the proteins bound to the lipids were determined by immunoblotting with an anti-GST antibody. GST alone and GST–cPLA2ε N-terminus did not exhibit binding to any of the lipids. As expected, GST–SidC-3C, GST–p40-PX and GST–PH-PLCδ were bound to PI(4)P, PI(3)P and PI(4,5)P2, respectively. GST–cPLA2ε C-terminus and GST–cPLA2ε remained bound to different PIPs. GST–cPLA2ε binding to PA was also detected. (B–F) HeLa cells were co-transfected with cDNAs encoding Myc–cPLA2ε and GFP-tagged PI-binding probes PH-PLCδ (B,C), PH-CERT (D), PH-GRP1 (E) and FIVE-SARA (F). The cells were then fixed and labeled with anti-Myc antibody. cPLA2ε exhibited substantial overlap with PH-PLCδ (B, arrows) and was displaced from the membranes in cells overexpressing PH-PLCδ (C). No overlap of cPLA2ε with PH-CERT was detected (D). PH-GRP1 was revealed at the tips of few cPLA2ε-positive tubules (E, arrows). FIVE-SARA was bound only to the vacuolar cPLA2ε structures (F, arrows) whereas cPLA2ε tubules remained devoid of the FIVE-SARA signal (F). Scale bar: 4.5 µm (B–F).

Fig. 4.

Association of cPLA2ε with phosphatidylinositol lipids. (A) PIP strip membranes containing spotted phosphatidylinositols and other lipids (as indicated in scheme) were incubated with 5 µg/ml of the following recombinant proteins: GST alone, GST–SidC-3C, GST–p40-PX, GST–PH-PLCδ, GST–cPLA2ε full-length, GST–cPLA2ε N-terminus (amino acids 1–63), GST–cPLA2ε C-terminus (815–875) and GST–cPLA2ε full-length fusion proteins. Then the membranes were washed and the proteins bound to the lipids were determined by immunoblotting with an anti-GST antibody. GST alone and GST–cPLA2ε N-terminus did not exhibit binding to any of the lipids. As expected, GST–SidC-3C, GST–p40-PX and GST–PH-PLCδ were bound to PI(4)P, PI(3)P and PI(4,5)P2, respectively. GST–cPLA2ε C-terminus and GST–cPLA2ε remained bound to different PIPs. GST–cPLA2ε binding to PA was also detected. (B–F) HeLa cells were co-transfected with cDNAs encoding Myc–cPLA2ε and GFP-tagged PI-binding probes PH-PLCδ (B,C), PH-CERT (D), PH-GRP1 (E) and FIVE-SARA (F). The cells were then fixed and labeled with anti-Myc antibody. cPLA2ε exhibited substantial overlap with PH-PLCδ (B, arrows) and was displaced from the membranes in cells overexpressing PH-PLCδ (C). No overlap of cPLA2ε with PH-CERT was detected (D). PH-GRP1 was revealed at the tips of few cPLA2ε-positive tubules (E, arrows). FIVE-SARA was bound only to the vacuolar cPLA2ε structures (F, arrows) whereas cPLA2ε tubules remained devoid of the FIVE-SARA signal (F). Scale bar: 4.5 µm (B–F).

Given that in vitro cPLA2ε exhibited quite promiscuous interaction with different PIP species, we sought to investigate how it binds in vivo to membranes enriched in specific PIPs. To this end, we evaluated the intracellular colocalization of cPLA2ε with PIP-binding probes such as PH domains of PLCδ (PLCδ-PH), CERT (CERT-PH), GRP1 (GRP1-PH) and FIVE domain of SARA (SARA-FIVE) that have a high affinity for PI(4,5)P2, PI(4)P, PI(3,4,5)P3 and PI(3)P, respectively. Confocal microscopy revealed substantial colocalization of PLCδ-PH with cPLA2ε along tubular and some vacuolar structures (Fig. 4B), indicating that the enzyme efficiently binds to PI(4,5)P2-enriched membranes in vivo. Interestingly, strong overexpression of PLCδ-PH frequently induced a detachment of cPLA2ε from the membranes (Fig. 4C), indicating that there is competition of the proteins for the same PI(4,5)P2 binding site. In contrast, cPLA2ε exhibited an almost complete lack of overlap with the PI(4)P-specific CERT-PH, which was mainly distributed over the Golgi (Fig. 4D). Colocalization of the enzyme with GRP1-PH, which binds to PI(3,4,5)P3, was mainly restricted to the plasma membrane, and in some cases, the tips of the recycling tubules (Fig. 4E). Interestingly, a clear cPLA2ε signal was detected over the SARA-FYVE-positive endosomes (Fig. 4F), suggesting that the enzyme was able to bind to PI(3)P membranes. However, incubation with wortmannin [an inhibitor of the phosphatidylinositol and therefore 3-kinase PI(3)P synthesis] did not change the intracellular pattern of cPLA2ε, whereas the same treatment induced detachment of the PI(3)P-binding proteins (such as EEA1) from the endosomal membranes into the cytosol (not shown). This suggests that recruitment of cPLA2ε to the SARA-FYVE membranes did not rely exclusively on its interaction with PI(3)P and that alternative binding mechanisms might exist. Indeed, cPLA2ε did not appear to compete with SARA-FYVE for PI(3)P-binding sites because overexpression of SARA-FYVE did not displace the enzyme from the membranes (not shown). Therefore, taken together, the above experiments indicate that the specific association of cPLA2ε with the membrane organelles of the clathrin-independent pathway relies on the interaction of the C-terminal PB domain of the enzyme with PI(4,5)P2.

Depletion of cPLA2ε affects recycling through the clathrin-independent pathway

The specific targeting of cPLA2ε to the clathrin-independent endocytic/recycling route suggests that the enzyme is involved in the regulation of the transport events in this pathway. To test whether this is the case, we first used an RNAi approach. Although the anti-cPLA2ε antibody failed to label the endogenous enzyme in immunofluorescence experiments, in western blots the antibody recognized a ∼96 kDa band corresponding to the molecular mass of cPLA2ε. This band was significantly reduced in HeLa cells that were incubated with cPLA2ε-specific siRNA(s), indicating the sufficient efficiency of the silencing (Fig. 5A).

Fig. 5.

cPLA2ε silencing affects tubule formation and recycling in the clathrin-independent pathway. (A) HeLa cells were treated with either scramble or cPLA2ε-specific siRNAs for 96 hours, lysed and subjected to immunoblotting with antibodies against either cPLA2ε or GAPDH. The anti-cPLA2ε antibody recognized a ∼96 kDa band that corresponded to the molecular mass of cPLA2ε and was significantly reduced in cPLA2ε-silenced cells as also revealed by quantification (see graph). (B) HeLa cells were treated with scramble or cPLA2ε-specific siRNAs or transfected with the mouse GFP–cPLA2ε cDNA 72 hours after addition of cPLA2ε-specific siRNA (indicated as Rescue). After 96 hours of siRNA treatment, cells were fixed and labeled with anti-MHC-I antibody. cPLA2ε-silenced cells exhibited strong perinuclear MHC-I accumulations (arrowheads), which disappeared almost completely in cells overexpressing cPLA2ε (asterisks). (C–E) HeLa cells were subjected to either cPLA2ε RNAi or to overexpression of the enzyme, fixed and investigated using confocal microscopy to analyze frequency and length of MHC-I tubular structures (indicated by arrows). Significant reduction of MHC-I-containing tubules was detected in cells incubated with siRNA, whereas cPLA2ε overexpression induced strong tabulation of MHC membranes (C, arrows). The graphs in D and E show a reduction in the length of MHC-I tubules (means ± s.d., n = 30 cells) and their number per cell (means ± s.d., n = 30 cells), respectively, whereas the overexpression of the cPLA2ε induced a significant increase in length and number of tubular structures. (F) Control (left column) or cPLA2ε-silenced (right column) cells were incubated for 30 minutes at 4°C with anti-MHC-I antibody to detect surface-bound MHC-I (see Materials and Methods) or incubated at 37°C for 10 minutes and 20 minutes and subjected to acid washing to reveal the internal pool of internalized MHC-I. Control and cPLA2ε-depleted cells exhibited similar amounts of anti-MHC-I bound to the cell surface (top row) and internalized 10 minutes after activation of MHC-I uptake (middle row). In contrast, at 20 minutes, more internalized MHC-I was revealed in silenced cells, indicating a delay in MHC-I recycling. (G) Control and cPLA2ε-depleted cells were incubated at 4°C with anti MHC-I antibody, washed and subjected to CytD block to inhibit MHC-I recycling back to the plasma membrane. Then the cells were washed with acid buffer and fixed directly (G, top row). Alternatively, the cells were incubated for 30 minutes at 37°C to allow MHC-I recycling (G, bottom row) and subjected to an acid wash before fixation. Immunofluorescence analysis revealed only the intracellular MHC-I pool, which remained significantly higher in cPLA2ε-silenced cells. (H,I) Normalized fluorescent intensity of MHC-I (means ± s.d., n = 50 cells) corresponding to uptake (F) and recycling (G) experiments respectively. ***P<0.001, Student's t-test. Scale bars: 8 µm (B,F), 4 µm (C), 7 µm (G).

Fig. 5.

cPLA2ε silencing affects tubule formation and recycling in the clathrin-independent pathway. (A) HeLa cells were treated with either scramble or cPLA2ε-specific siRNAs for 96 hours, lysed and subjected to immunoblotting with antibodies against either cPLA2ε or GAPDH. The anti-cPLA2ε antibody recognized a ∼96 kDa band that corresponded to the molecular mass of cPLA2ε and was significantly reduced in cPLA2ε-silenced cells as also revealed by quantification (see graph). (B) HeLa cells were treated with scramble or cPLA2ε-specific siRNAs or transfected with the mouse GFP–cPLA2ε cDNA 72 hours after addition of cPLA2ε-specific siRNA (indicated as Rescue). After 96 hours of siRNA treatment, cells were fixed and labeled with anti-MHC-I antibody. cPLA2ε-silenced cells exhibited strong perinuclear MHC-I accumulations (arrowheads), which disappeared almost completely in cells overexpressing cPLA2ε (asterisks). (C–E) HeLa cells were subjected to either cPLA2ε RNAi or to overexpression of the enzyme, fixed and investigated using confocal microscopy to analyze frequency and length of MHC-I tubular structures (indicated by arrows). Significant reduction of MHC-I-containing tubules was detected in cells incubated with siRNA, whereas cPLA2ε overexpression induced strong tabulation of MHC membranes (C, arrows). The graphs in D and E show a reduction in the length of MHC-I tubules (means ± s.d., n = 30 cells) and their number per cell (means ± s.d., n = 30 cells), respectively, whereas the overexpression of the cPLA2ε induced a significant increase in length and number of tubular structures. (F) Control (left column) or cPLA2ε-silenced (right column) cells were incubated for 30 minutes at 4°C with anti-MHC-I antibody to detect surface-bound MHC-I (see Materials and Methods) or incubated at 37°C for 10 minutes and 20 minutes and subjected to acid washing to reveal the internal pool of internalized MHC-I. Control and cPLA2ε-depleted cells exhibited similar amounts of anti-MHC-I bound to the cell surface (top row) and internalized 10 minutes after activation of MHC-I uptake (middle row). In contrast, at 20 minutes, more internalized MHC-I was revealed in silenced cells, indicating a delay in MHC-I recycling. (G) Control and cPLA2ε-depleted cells were incubated at 4°C with anti MHC-I antibody, washed and subjected to CytD block to inhibit MHC-I recycling back to the plasma membrane. Then the cells were washed with acid buffer and fixed directly (G, top row). Alternatively, the cells were incubated for 30 minutes at 37°C to allow MHC-I recycling (G, bottom row) and subjected to an acid wash before fixation. Immunofluorescence analysis revealed only the intracellular MHC-I pool, which remained significantly higher in cPLA2ε-silenced cells. (H,I) Normalized fluorescent intensity of MHC-I (means ± s.d., n = 50 cells) corresponding to uptake (F) and recycling (G) experiments respectively. ***P<0.001, Student's t-test. Scale bars: 8 µm (B,F), 4 µm (C), 7 µm (G).

To understand whether cPLA2ε depletion has an impact on the organization of clathrin-independent pathway we first stained control and silenced cells for MHC-I. Confocal microscopy revealed a significant increase in the MHC-I-associated fluorescent signal in the perinuclear area upon cPLA2ε knock down (Fig. 5B, arrowheads), indicating that the recycling of MHC-I from the perinuclear compartment back to the cell surface might be delayed upon depletion of cPLA2ε. Indeed, expression of mouse cPLA2ε (that is resistant to human cPLA2ε-specific siRNAs) in silenced cells significantly reduced MHC-I accumulation in the perinuclear compartment (Fig. 5B, asterisks), probably by rescue of MHC-I recycling back to the plasma membrane.

Recycling through the clathrin-independent route has been reported to rely on the formation of long tubular elements that serve as precursors for the formation of transport carriers directed to the surface of the cells (Grant and Donaldson, 2009). In this context, we checked whether silencing of cPLA2ε affects the formation of such long tubular elements given that PLA2 activity has been demonstrated to facilitate the formation of the membrane tubules, in both the secretory and endocytic pathways (Bechler et al., 2012; Ha et al., 2012). Confocal microscopy of the control cells frequently revealed several MHC-I tubular elements emerging from the perinuclear area (Fig. 5C, arrows). Both the length and the number of such tubular structures were reduced by knockdown of cPLA2ε (Fig. 5C–E). In contrast, overexpression of cPLA2ε (about four times above the endogenous level) induced the opposite phenotype, with a compelling increase in MHC-I tubular elements (Fig. 5C–E).

To prove the involvement of cPLA2ε in the regulation of the transport through the clathrin-independent route, MHC-I uptake and recycling assays were used. First, we determined the impact of cPLA2ε depletion on the internalization of MHC-I. Confocal microscopy revealed that the amount of endocytosed MHC-I was similar in control and silenced cells at the early time points (i.e. within 10 minutes) after activation of endocytosis (Fig. 5F). This suggests that depletion of cPLA2ε did not affect early stages of the internalization in the clathrin-independent endocytic pathway. However, we noted that with time (after 15–20 minutes), the intracellular MHC-I pool tended to increase in silenced cells (Fig. 5F), which was probably due to an impairment of the MHC-I recycling back to the cell surface.

To test whether this was the case we used cytochalasin D, which has been reported to block and therefore synchronize internalized MHC-I within the clathrin-independent recycling compartment (Weigert et al., 2004). Indeed, cytochalasin D treatment allowed similar amounts of MHC-I to accumulate over 30 minutes in control and cPLA2ε-silenced cells (Fig. 5G), indicating that depletion of cPLA2ε did not impact MHC-I internalization. Next, the drug was removed to trigger MHC-I recycling (Weigert et al., 2004). Fig. 5G shows that 30 minutes after activation of the recycling (cytochalasin D wash-out), almost all the MHC-I was transported away from the intracellular compartments in control cells. In contrast, cPLA2ε-depleted cells still exhibited quite strong MHC-I staining associated with the intracellular membranes (Fig. 5G), indicating a delay in recycling through the clathrin-independent route. The accumulation of MHC-I in silenced cells was confirmed by morphometric analysis in both uptake and recycling experiments (Fig. 5H,I). Next, the MHC-I recycling in control and silenced cells was investigated under steady-state conditions (i.e. without cytochalasin D block). Similar to when the cytochalasin D block was released, cPLA2ε ablation induced retention of MHC-I in the clathrin-independent recycling compartment (supplementary material Fig. S3).

In addition, we checked whether cPLA2ε is engaged in the regulation of the transport events within other endocytic/recycling routes. We found that cPLA2ε-depleted cells did not exhibit significant changes in the pattern of either M6PR or caveolin, suggesting that transport through endosome-to-Golgi or caveolin-dependent routes, respectively, does not require this enzyme (supplementary material Fig. S4A,B). Similarly, knockdown of cPLA2ε did not affect uptake and delivery of EGF to the late endocytic compartment (supplementary material Fig. S4C). When fluorescent transferrin (Tfn) was added to cells, the amount of internalized tracer remained almost the same in control and silenced cells at both early (10 minutes) and late (20 minutes) time intervals after the pulse (supplementary material Fig. S4D), indicating that neither uptake nor recycling of Tfn is affected by a deficit of cPLA2ε.

Thus, taken together, the above results suggest that cPLA2ε operates mainly and quite specifically in transport through the clathrin-independent route. To further support this conclusion, we investigated the association of cPLA2ε with specific cargoes – CD55, CD98 and CD147 – that use the clathrin-independent uptake and recycling pathway (Eyster et al., 2009). All three proteins exhibited a substantial overlap with the enzyme along tubular elements and vesicular structures (Fig. 6A). Next, the impact of cPLA2ε depletion on CD55 and CD147 recycling was evaluated. Control and cPLA2ε-silenced cells were uploaded with antibodies against either CD55 or CD147 in the presence of cytochalasin D to accumulate corresponding cargoes in the recycling clathrin-independent compartment. Cytochalasin D was then washed out to activate the recycling of either CD55 or CD147. Fig. 6B,C shows that in control cells, the intracellular pool of CD55 was consumed within 30 minutes after cytochalasin D washout as a result of recycling, whereas cPLA2ε-deficient cells still retained a significant amount of internalized CD55. Similarly, cPLA2ε suppression induced intracellular retention of CD147 (Fig. 6D), indicating that the enzyme is required to support recycling of the specific cargoes through the clathrin-independent route.

Fig. 6.

cPLA2ε is required for recycling of cargo proteins trafficking through the clathrin-independent endocytic pathway. (A) HeLa cells were transfected with Myc-tagged cPLA2ε, fixed and stained with antibodies against CD55, CD147 or CD98. Immunofluorescence analysis revealed substantial cPLA2ε colocalization with CD55, CD147 and CD98 both at the level of vesicular structures (arrowheads) and tubular element (arrows). (B,C) Control and cPLA2ε-silenced cells were incubated with antibody against CD55 at 4°C to allow it to bind to the cell surface. Then the cells were washed and shifted to 37°C for 30 minutes in the presence of cytochalasin D to synchronize internalized CD55 antibody in the recycling compartment. The cells were then fixed directly or incubated for an additional 30 minutes with cytochalasin-D-free medium to allow CD55 recycling to the cell surface. Before fixation the cells were subjected to acid wash to eliminate CD55 signal from the cell surface. The cells were further permeabilized and incubated with fluorescent secondary antibody to reveal an intracellular pool of the CD55. cPLA2ε silencing did not significantly affect the CD55 uptake (top row) but induced retention of CD55 upon activation of the recycling (bottom row). Quantification of normalized fluorescence (means ± s.d., n = 30 cells) was performed as shown in Fig. 5H and revealed a significantly higher amount of CD55 retained in cPLA2ε-silenced cells 30 minutes after activation of the recycling. (D) Similar recycling experiments were performed with CD147 and revealed a higher retention of internalized CD147 in cPLA2ε-deficient cells despite activation of the recycling. ** P<0.01, ***P<0.001, Student's t-test. Scale bars: 4.2 µm (A), 8 µm (B).

Fig. 6.

cPLA2ε is required for recycling of cargo proteins trafficking through the clathrin-independent endocytic pathway. (A) HeLa cells were transfected with Myc-tagged cPLA2ε, fixed and stained with antibodies against CD55, CD147 or CD98. Immunofluorescence analysis revealed substantial cPLA2ε colocalization with CD55, CD147 and CD98 both at the level of vesicular structures (arrowheads) and tubular element (arrows). (B,C) Control and cPLA2ε-silenced cells were incubated with antibody against CD55 at 4°C to allow it to bind to the cell surface. Then the cells were washed and shifted to 37°C for 30 minutes in the presence of cytochalasin D to synchronize internalized CD55 antibody in the recycling compartment. The cells were then fixed directly or incubated for an additional 30 minutes with cytochalasin-D-free medium to allow CD55 recycling to the cell surface. Before fixation the cells were subjected to acid wash to eliminate CD55 signal from the cell surface. The cells were further permeabilized and incubated with fluorescent secondary antibody to reveal an intracellular pool of the CD55. cPLA2ε silencing did not significantly affect the CD55 uptake (top row) but induced retention of CD55 upon activation of the recycling (bottom row). Quantification of normalized fluorescence (means ± s.d., n = 30 cells) was performed as shown in Fig. 5H and revealed a significantly higher amount of CD55 retained in cPLA2ε-silenced cells 30 minutes after activation of the recycling. (D) Similar recycling experiments were performed with CD147 and revealed a higher retention of internalized CD147 in cPLA2ε-deficient cells despite activation of the recycling. ** P<0.01, ***P<0.001, Student's t-test. Scale bars: 4.2 µm (A), 8 µm (B).

Catalytic activity of cPLA2ε is required to drive trafficking through the clathrin-independent recycling pathway

The recycling of the MHC-I and similar clathrin-independent endocytic cargoes back to the cell surface is mediated through tubular elements that serve as precursors for transport carriers carrying the cargo to the plasma membrane (Grant and Donaldson, 2009; Weigert et al., 2004). The formation of tubular membrane structures operating in several intracellular transport routes has been reported to require catalytic activity of PLA2 enzymes (Bechler et al., 2011; Bechler et al., 2010; San Pietro et al., 2009). It has been hypothesized that phospholipid hydrolysis by PLA2 enzymes results in an increase in wedge-shaped lysolipids that induce or support a high curvature of the tubule membrane (Bechler et al., 2012; Ha et al., 2012; Polishchuk et al., 2009). Thus we wondered whether the catalytic activity of cPLA2ε is needed for the generation of the tubular elements that support recycling of the MHC-I from the intracellular compartments back to the cell surface. Given that the specific chemical inhibitors for cPLA2ε are yet to be developed, we generated an inactive cPLA2ε protein. According to predictions (Ghosh et al., 2006b), an active site in cPLA2ε contains a nucleophilic serine (S420), which is conserved among different cPLA2 enzymes (Fig. 7A) and should have a key role in cPLA2ε catalytic activity. Thus a cPLA2εS420A mutant was created and expressed in HEK293 cells to evaluate its catalytic activity. Fig. 7B shows that the mutant exhibited a reduced capacity to hydrolyze phospholipids compared with the wild-type cPLA2ε enzyme, although both cPLA2ε and cPLA2εS420A were expressed at the same level in the cells, as evaluated by western blot (not shown). Further analysis revealed that overexpression of cPLA2ε substantially increased the amount of lysolipids in the cells, whereas cPLA2εS420A failed to stimulate lysolipid production (Fig. 7C). Thus it turns out that the catalytic activity of the mutant was significantly reduced compared with the wild-type cPLA2ε. Next, an inactive cPLA2εS420A mutant was expressed in HeLa cells and its intracellular distribution was investigated using confocal microscopy and compared with the wild-type enzyme. Cells expressing wild-type cPLA2ε exhibited multiple membrane tubular elements (Fig. 7D). In contrast, such tubular structures were significantly reduced in cPLA2εS420A-positive cells (Fig. 7D,E) and the mutant enzyme was detected along the plasma membrane as well as at the number of the endosome-like intracellular membrane compartments (Fig. 7D). Thus, the cPLA2εS420A isoform is capable of binding to the membranes (similar to the wild-type protein), but does not support tubule formation owing to the loss of catalytic activity.

Fig. 7.

Catalytically inactive cPLA2εS420A mutant does not support formation of MHC-I membrane tubules. (A) Alignment of group IV PLA2 family of enzymes catalytic domain was performed using UniProt Blast tool (identity is shown in red and similarity in blue) and revealed key serine 420 residue (highlighted in yellow) of the active site conserved among those isoforms. On the basis of the studies of sequence similarities, a cPLA2εS420A point mutant was generated. (B) HEK293T cells were transfected with empty pCMV–Myc vector or vectors encoding Myc–cPLA2ε and Myc–cPLA2εS420A. Then glycerophospholipid hydrolysis (quantified as percentage of total counts; means ± s.d., n = 4 experiments) was evaluated. The cells expressing cPLA2ε exhibit a significant increase in phospholipase activity, whereas transfection of cPLA2εS420A did not result in any increase in glycerophospholipid hydrolysis. (C) HEK293T cells were transfected with empty pCMV–Myc vector or cDNAs encoding either Myc–cPLA2ε or Myc–cPLA2εS420A. Then lysolipid production (means ± s.d., n = 4 experiments) was evaluated. Lysolipids were detected to increase significantly in cells expressing cPLA2ε, whereas cPLA2εS420A transfection did not enhance lysolipid generation. (D,E) HeLa cells were transfected with either Myc–cPLA2ε or Myc–cPLA2εS420A, fixed and stained with antibodies against Myc. Confocal microscopy revealed cPLA2ε to induce formation of numerous tubular elements (D, left panel, arrows), whereas only few tubular elements were detected in the cells expressing cPLA2εS420A (D, right panel, arrows). The number of tubules (means ± s.d., n = 50 cells) was significantly reduced in cells expressing cPLA2εS420A (E) (F,G) HeLa cells were transfected with Myc-tagged cPLA2ε or cPLA2εS420A and then incubated for 30 minutes at 4°C with anti-MHC-I antibody to label MHC-I at the cell surface. After incubation, cells were washed and incubated at 37°C for 30 minutes to fill the entire clathrin-independent endocytic route. After brief acid washing to remove surface-bound antibody, cells were fixed, permeabilized and labeled with an Alexa Fluor 488 secondary antibody to reveal internalized MHC-I and with anti-Myc antibody to reveal cPLA2ε or cPLA2εS420A. Confocal images show that both cPLA2ε and cPLA2εS420A are associated with MHC-I-positive vacuolar endosomes (F,G; arrowheads), whereas MHC-I tubular structures containing the enzyme were detected only in cells expressing cPLA2ε (F; arrows). ***P<0.001, Student's t-test. Scale bars: 4.2 µm (D,F,G).

Fig. 7.

Catalytically inactive cPLA2εS420A mutant does not support formation of MHC-I membrane tubules. (A) Alignment of group IV PLA2 family of enzymes catalytic domain was performed using UniProt Blast tool (identity is shown in red and similarity in blue) and revealed key serine 420 residue (highlighted in yellow) of the active site conserved among those isoforms. On the basis of the studies of sequence similarities, a cPLA2εS420A point mutant was generated. (B) HEK293T cells were transfected with empty pCMV–Myc vector or vectors encoding Myc–cPLA2ε and Myc–cPLA2εS420A. Then glycerophospholipid hydrolysis (quantified as percentage of total counts; means ± s.d., n = 4 experiments) was evaluated. The cells expressing cPLA2ε exhibit a significant increase in phospholipase activity, whereas transfection of cPLA2εS420A did not result in any increase in glycerophospholipid hydrolysis. (C) HEK293T cells were transfected with empty pCMV–Myc vector or cDNAs encoding either Myc–cPLA2ε or Myc–cPLA2εS420A. Then lysolipid production (means ± s.d., n = 4 experiments) was evaluated. Lysolipids were detected to increase significantly in cells expressing cPLA2ε, whereas cPLA2εS420A transfection did not enhance lysolipid generation. (D,E) HeLa cells were transfected with either Myc–cPLA2ε or Myc–cPLA2εS420A, fixed and stained with antibodies against Myc. Confocal microscopy revealed cPLA2ε to induce formation of numerous tubular elements (D, left panel, arrows), whereas only few tubular elements were detected in the cells expressing cPLA2εS420A (D, right panel, arrows). The number of tubules (means ± s.d., n = 50 cells) was significantly reduced in cells expressing cPLA2εS420A (E) (F,G) HeLa cells were transfected with Myc-tagged cPLA2ε or cPLA2εS420A and then incubated for 30 minutes at 4°C with anti-MHC-I antibody to label MHC-I at the cell surface. After incubation, cells were washed and incubated at 37°C for 30 minutes to fill the entire clathrin-independent endocytic route. After brief acid washing to remove surface-bound antibody, cells were fixed, permeabilized and labeled with an Alexa Fluor 488 secondary antibody to reveal internalized MHC-I and with anti-Myc antibody to reveal cPLA2ε or cPLA2εS420A. Confocal images show that both cPLA2ε and cPLA2εS420A are associated with MHC-I-positive vacuolar endosomes (F,G; arrowheads), whereas MHC-I tubular structures containing the enzyme were detected only in cells expressing cPLA2ε (F; arrows). ***P<0.001, Student's t-test. Scale bars: 4.2 µm (D,F,G).

Next, we checked to what extent targeting of cPLA2εS420A to the clathrin-independent endocytic/recycling pathway was preserved. Confocal microscopy revealed that, similar to cPLA2ε, the mutant enzyme was found in endocytic structures containing internalized MHC-I (Fig. 7F,G arrowheads), a main cargo marker of the clathrin-independent pathway. Therefore, targeting of the cPLA2ε to the clathrin-independent compartments appeared to be unaffected by the S420A mutation. However, in contrast to cPLA2ε-positive cells, where MHC-I tubular elements were detected (Fig. 7F, arrows), the pattern of MHC-I localization was altered upon cPLA2εS420A expression. Cells that expressed cPLA2εS420A exhibited a significant reduction of MHC-I tubules (Fig. 7G), indicating that activity of the enzyme is needed for their formation.

To confirm the requirement of cPLA2ε activity for recycling through the clathrin-independent pathway, we investigated whether cPLA2εS420A is capable of rescuing the delay of the MHC-I retrieval to the cell surface in cPLA2ε-silenced cells. To this end, cPLA2ε-depleted cells were transfected with either cPLA2ε or cPLA2εS420A, loaded with an anti-MHC-I antibody, and subjected to cytochalasin D block to prompt accumulation of MHC-I in the clathrin-independent endocytic compartment. Then, the cytochalasin D block was released to investigate the fate of intracellular MHC-I. Analysis of the silenced cells 30 minutes after cytochalasin D wash-out, revealed a significant amount of MHC-I that was still retained within the perinuclear compartment (Fig. 8A,B), whereas cPLA2ε overexpression reduced the intracellular pool of the MHC-I (Fig. 8C,D) by reactivation of the recycling. Unlike the wild-type enzyme, cPLA2εS420A was unable to rescue MHC-I transport, because the amount of MHC-I in cells that expressed the inactive mutant enzyme was similar to that of the untransfected silenced cells (Fig. 8E–G). Thus, taken together, the above results suggest that the catalytic activity of cPLA2ε is required to drive the formation of clathrin-independent tubular elements and to support recycling through the MHC-I route.

Fig. 8.

Catalytic activity of cPLA2ε is required to support MHC-I recycling. (A–F) HeLa cells were treated with scramble (A) and cPLA2ε-specific siRNA (B) for 96 hours. Two additional sets of siRNA-treated cells were also transfected with mouse Myc–cPLA2ε (C,D) or with its inactive mutant cPLA2εS420A (E,F). At the end of the siRNA treatment the cells were incubated for 30 minutes at 4°C with an anti MHC-I antibody to label MHC-I present at the cell surface (binding step). After incubation, the cells were washed and kept in cytochalasin D containing medium at 37°C for 30 minutes to allow internalization of MHC-I and to prevent its recycling back to the cell surface. Then the cells were subjected to acid wash to remove the antibody still present at the plasma membrane and incubated in cytochalasin D for 30 minutes to activate recycling. Before fixation, the cells were subjected to acid wash and remaining intracellular MHC-I was revealed using secondary fluorescent antibody. Control cells exhibited low levels of internalized MHC-I (A), whereas cPLA2ε-silenced cells revealed accumulation of endocytosed MHC-I in the perinuclear recycling compartment (B; arrows). Such accumulation of MHC-I (C,D; arrows) in cPLA2ε-deficient cells was in contrast to overexpression of cPLA2ε (C,D; cells outlined by dashed line). cPLA2εS420A overexpression (E,F; cells outlined by dash line) failed to rescue internalized MHC-I from retention in the perinuclear compartment (E,F; arrows). (G) Fluorescence signal associated with internalized MHC-I was quantified in the cells treated as described above and was normalized to MHC-I fluorescence in control cells. The graph shows that normalized MHC-I fluorescence (means ± s.d., n = 30 cells) increases upon cPLA2ε RNAi. cPLA2ε overexpression reduced MHC-I signal in cPLA2ε-silenced cells, whereas cPLA2εS420A mutant did not impact MHC-I increase induced by cPLA2ε ablation. ***P<0.001, Student's t-test. Scale bar: 7 µm (A–F).

Fig. 8.

Catalytic activity of cPLA2ε is required to support MHC-I recycling. (A–F) HeLa cells were treated with scramble (A) and cPLA2ε-specific siRNA (B) for 96 hours. Two additional sets of siRNA-treated cells were also transfected with mouse Myc–cPLA2ε (C,D) or with its inactive mutant cPLA2εS420A (E,F). At the end of the siRNA treatment the cells were incubated for 30 minutes at 4°C with an anti MHC-I antibody to label MHC-I present at the cell surface (binding step). After incubation, the cells were washed and kept in cytochalasin D containing medium at 37°C for 30 minutes to allow internalization of MHC-I and to prevent its recycling back to the cell surface. Then the cells were subjected to acid wash to remove the antibody still present at the plasma membrane and incubated in cytochalasin D for 30 minutes to activate recycling. Before fixation, the cells were subjected to acid wash and remaining intracellular MHC-I was revealed using secondary fluorescent antibody. Control cells exhibited low levels of internalized MHC-I (A), whereas cPLA2ε-silenced cells revealed accumulation of endocytosed MHC-I in the perinuclear recycling compartment (B; arrows). Such accumulation of MHC-I (C,D; arrows) in cPLA2ε-deficient cells was in contrast to overexpression of cPLA2ε (C,D; cells outlined by dashed line). cPLA2εS420A overexpression (E,F; cells outlined by dash line) failed to rescue internalized MHC-I from retention in the perinuclear compartment (E,F; arrows). (G) Fluorescence signal associated with internalized MHC-I was quantified in the cells treated as described above and was normalized to MHC-I fluorescence in control cells. The graph shows that normalized MHC-I fluorescence (means ± s.d., n = 30 cells) increases upon cPLA2ε RNAi. cPLA2ε overexpression reduced MHC-I signal in cPLA2ε-silenced cells, whereas cPLA2εS420A mutant did not impact MHC-I increase induced by cPLA2ε ablation. ***P<0.001, Student's t-test. Scale bar: 7 µm (A–F).

Membrane trafficking, and in particular events associated with generation of transport carriers, cause local changes in the physical properties of membrane bilayers that, in turn, are usually associated with variation in the composition of both lipids and protein. To execute this process, cells host a vast repertoire of mechanisms to allow dynamic and reversible membrane remodeling associated with transport carrier biogenesis through activity of membrane-associated proteins and lipid-modifying enzymes (Graham and Kozlov, 2010; McMahon and Gallop, 2005). Among the latter, PLA2 enzymes have emerged as important regulators of transport processes, which rely on biogenesis of membrane tubular structures (Bechler et al., 2012; Ha et al., 2012; Polishchuk et al., 2009).

We now provide evidence for the role of a recently identified cPLA2ε in trafficking through the clathrin-independent endocytic route. We found that cPLA2ε is specifically associated with the clathrin-independent endocytic pathway, where it regulates the recycling process through the formation of tubular elements that carry internalized clathrin-independent cargoes back to the cell surface (Fig. 9A).

Fig. 9.

Schematic illustration of cPLA2ε action in the clathrin-independent endocytic/recycling pathway. (A) The scheme illustrates trafficking of clathrin-independent (red and green circles) and clathrin-dependent (blue and yellow circles) cargo proteins through the endocytic system. cPLA2ε drives formation of tubular elements at the level of clathrin-independent recycling compartments and therefore supports recycling of clathrin-independent cargo proteins back to the cell surface (red arrow). (B) Inset 1, cPLA2ε hydrolyses the fatty acids (FA) at the sn-2 position of cylindrical phospholipids to form wedge-shaped lysophospholipids. Inset 2, this formation of wedge-shaped lysophospholipids favors generation of spontaneous membrane curvature and transformation of endosome membrane with relatively low curvature into highly curved tubular membranes, which operate in the recycling of clathrin-independent cargoes towards the plasma membrane.

Fig. 9.

Schematic illustration of cPLA2ε action in the clathrin-independent endocytic/recycling pathway. (A) The scheme illustrates trafficking of clathrin-independent (red and green circles) and clathrin-dependent (blue and yellow circles) cargo proteins through the endocytic system. cPLA2ε drives formation of tubular elements at the level of clathrin-independent recycling compartments and therefore supports recycling of clathrin-independent cargo proteins back to the cell surface (red arrow). (B) Inset 1, cPLA2ε hydrolyses the fatty acids (FA) at the sn-2 position of cylindrical phospholipids to form wedge-shaped lysophospholipids. Inset 2, this formation of wedge-shaped lysophospholipids favors generation of spontaneous membrane curvature and transformation of endosome membrane with relatively low curvature into highly curved tubular membranes, which operate in the recycling of clathrin-independent cargoes towards the plasma membrane.

Polybasic C-terminal stretch is a new targeting motif for compartmentalization of cytosolic PLA2

Our findings are similar to those of recent studies that demonstrated a nearly identical involvement of other PLA2 enzymes in membrane tabulation, which is needed for intra-Golgi transport (San Pietro et al., 2009), protein export from the TGN (Bechler et al., 2010) and recycling of endocytosed transferrin (Bechler et al., 2011). The notion that distinct PLA2 enzymes were found to support non-overlapping trafficking steps suggests that each PLA2 isoform possesses an array of targeting signals for specific compartmentalization. This, in turn, allows any given PLA2 enzyme to generate tubular transport carriers in a specific segment of either the secretory or endocytic route. How do such signals operate to ensure correct compartmentalization of the PLA2 enzymes?

Among the members of cPLA2 family, the N-terminal C2 domain is considered to be the main targeting determinant that, upon increase in cytosolic Ca2+ concentrations, promotes the enzyme binding to the correct membrane domain (Ghosh et al., 2006b). In contrast to other cPLA2 enzymes, binding of cPLA2ε to clathrin-independent endocytic membranes requires neither Ca2+ nor the C2 domain and relies on a KKRxR polybasic (PB) stretch at the C-terminus, which is capable of interacting with PI(4,5)P2 (Heo et al., 2006). We propose that this novel mechanism of Ca2+-independent targeting allows cPLA2ε to be compartmentalized differently than the other cPLA2 enzymes. Among the other members of the cPLA2 group, only cPLA2β contains a similar PB sequence at its C-terminus. However, it is not clear to what extent the PB domain contributes to localization of cPLA2β because the involvement of the C2 domain in cPLA2β membrane binding was clearly documented (Ghosh et al., 2006b). cPLA2α lacks the C-terminal PB domain and is normally targeted to the Golgi and ER membranes when cytosolic Ca2+ increases (Evans et al., 2001). Recently, a cluster of positively charged amino acids in the central part of the cPLA2α catalytic domain has been reported to determine enzyme targeting to the phagosome cup in macrophages (Casas et al., 2010). The presence of this positively-charged cluster might also explain why some cPLA2α can be detected at CD59-positive endosomes (Cai et al., 2012) involved in the clathrin-independent endocytic pathway and contain elevated amounts of PI(4,5)P2 similar to phagosomes (Botelho et al., 2000). These findings, however, are in contrast with recent data on the ability of the cPLA2α C2 domain to bind to the phagosome alone (Zizza et al., 2012). Moreover, others have demonstrated that the PB domain operates in the regulation of cPLA2α catalytic activity rather than in enzyme targeting (Tucker et al., 2009). This indicates that cPLA2ε is the only member of the family that heavily relies on the C-terminal basic patch to be targeted to the appropriate intracellular compartment.

Role of phosphoinositide lipids in targeting of cPLA2ε

We found that the C-terminal portion of cPLA2ε, which contains the PB domain as well as the full-length enzyme, binds in vitro to a number of phosphoinositides, as is expected for proteins with the PB cluster (Heo et al., 2006). Several lines of evidence, however, suggest that in vivo, cPLA2ε exhibits a higher affinity for PI(4,5)P2-enriched membranes. First, cPLA2ε was found to overlap with the PI(4,5)P2 marker PH-PLCδ along tubular and vacuolar structures of the clathrin-independent endocytic route. Second, a strong overexpression of PH-PLCδ resulted in displacement of cPLA2ε from the clathrin-independent endocytic membranes, probably because of competition for the same PI(4,5)P2 binding sites (Fig. 4C). Third, we observed significant dissociation of cPLA2ε from the membranes upon treatment with ionomycin, which stimulated PI(4,5)P2 hydrolysis by PLC (Heo et al., 2006). Finally, we found that cPLA2ε was strongly recruited to PI(4,5)P2-enriched vacuoles that form in HeLa cells upon overexpression of ARFQ67L (Naslavsky et al., 2003). Notably, the ability of cPLA2ε to bind to such vacuoles was lost after truncation of the PB motif.

Among the other PI-specific probes, only SARA-FYVE exhibits some overlap with cPLA2ε at the intracellular endocytic structures, indicating that the enzyme also interacts with PI(3)P-enriched membranes. This probably allows cPLA2ε to escort incoming clathrin-independent endosomes even when the PI(4,5)P2 to PI(3)P switch happens in their membranes as a result of the intersection with the conventional clathrin-dependent endocytic route (Naslavsky et al., 2003).

Further along the clathrin-independent pathway, cPLA2ε continues to remain associated with recycling tubular elements, where, as reported, PI(3)P is exchanged back to PI(4,5)P2 and the clathrin-independent cargoes are sorted from other endocytosed molecules (Naslavsky et al., 2003). Therefore, specific PI-binding properties support targeting of cPLA2ε to the compartments of the clathrin-independent route. Furthermore, PIPs have been shown to promote cPLA2ε activity (Ghomashchi et al., 2010) indicating that a specific PIP composition is likely to be required, not only for targeting but also for the activity of the enzyme.

Interestingly, the specific binding of the other PLA2 enzyme PAFAH-Ib to PI(4)P and PI(3)P was convincingly documented and hypothesized to play a key role in the association of PAFAH-Ib with Golgi and early endosome membranes, respectively (Bechler et al., 2011). In contrast to cPLA2ε, the precise PIP-binding portion of PAFAH-Ib remains to be identified (Bechler et al., 2011). Nonetheless, the example of cPLA2ε and PAFAH-Ib reveals the affinity for specific PIP(s) as a novel and important mechanism that contributes to differential compartmentalization of the PLA2 enzymes.

cPLA2ε specifically functions in the clathrin-independent endocytic/recycling pathway

Taking into consideration the selective association of cPLA2ε with clathrin-independent endocytic membranes, we expected the enzyme to operate in the transport processes that occur within this route. Although the protein is distributed along both incoming and recycling membranes of the clathrin-independent pathway, we found that cPLA2ε is required only for a subset of transport events along the clathrin-independent pathway. Depletion of cPLA2ε affects neither uptake of clathrin-independent-specific cargo MHC-I nor its subsequent delivery to the perinuclear recycling compartment. However, further trafficking of MHC-I was impaired in cPLA2ε-depleted cells because of a significant reduction in the tubular elements involved in MHC-I recycling back to the plasma membrane. Therefore, cPLA2ε is required specifically for tubule-mediated transport events. In contrast, the cPLA2ε apparently does not operate in the uptake and early clathrin-independent endocytic trafficking steps, where the generation of tubular membranes has not been reported (Donaldson et al., 2009; Grant and Donaldson, 2009).

Surprisingly, cPLA2α was recently detected at CD59-positive endosomes where the enzyme interacts with EHD1 and induces fission and vesiculation of CD59-positive tubular elements (Cai et al., 2012). At first glance these findings appear to be in contrast with the expected role of cytosolic PLA2 enzymes to support membrane tubulation (Bechler et al., 2011; Bechler et al., 2010; San Pietro et al., 2009; Yang et al., 2011). However, the function of cPLA2α in clathrin-independent endocytosis might require a complex interplay with EHD1 (Cai et al., 2012). Moreover, in contrast to cPLA2ε, suppression of cPLA2α seems to have very mild impact on trafficking through the clathrin-independent endocytic route (Cai et al., 2012). Therefore, a specific contribution of each of the above cPLA2 enzymes in membrane morphogenesis and trafficking along the clathrin-independent pathway should be carefully examined in future.

Notably, the inhibitory impact of cPLA2ε depletion was specific for the clathrin-independent route. Other endocytic/recycling pathways remained almost perfectly functional in cPLA2ε-deficient cells despite a number of those pathways being operated through tubular elements. For instance, tubule-mediated processes such as sorting of M6PR from the endosomes to the TGN (Bonifacino and Rojas, 2006; Cullen, 2008) or recycling of internalized Tfn to the plasma membrane (Bonifacino and Rojas, 2006; Cullen, 2008) were not seriously challenged by knockdown of cPLA2ε. This is probably due to the relatively poor association of cPLA2ε with the membranes involved in Tfn and M6PR trafficking (see Fig. 1). Thus, other PLA2 enzymes (or even different molecular mechanisms) are likely to drive the formation of tubular elements in M6PR and Tfn recycling pathways. Indeed, the biogenesis of the tubular membranes that recycle Tfn back to the cell surface has been shown to require the PLA2 activity of PAFAH-Ib (Bechler et al., 2011). Meanwhile, whether the biogenesis of M6PR tubules from the endosomes involves any PLA2 enzyme is yet to be determined.

Catalytic activity of cPLA2ε is required for tubule formation and recycling through the clathrin-independent route

The selectivity of cPLA2ε for the clathrin-independent pathway prompted us to investigate the mechanism through which the enzyme regulates biogenesis of the tubular elements involved in recycling of clathrin-independent cargoes. The elucidation of this mechanism is of great importance given that the proteins that use this transport pathway play a significant role in a number of physiological and pathological processes, such as immune surveillance and defense (MHC-I, CD55, CD59) (Ferrone et al., 2000; Gelderman et al., 2004), cell adhesion (β1-integrin) (Brown et al., 2001) and tumor cell invasion (CD147 and CD98) (Cantor and Ginsberg, 2012; Guo et al., 2000). This makes cPLA2ε an appealing target for the development of specific chemical inhibitors or activators that would regulate surface exposure and hence also regulate the specific activities of the clathrin-independent proteins.

We found that the catalytic activity of cPLA2ε is required to support formation of the membrane tubules that mediate clathrin-independent recycling, because it has also been reported for other PLA2 enzymes (Bechler et al., 2011; Bechler et al., 2010; San Pietro et al., 2009; Yang et al., 2011). Although a specific cPLA2ε inhibitor is still not available, we used an inactive cPLA2εS420A mutant to confirm this. This mutant is capable of binding to the clathrin-independent endocytic compartments, but lacks the capacity to hydrolyze phospholipids into lysolipids. As a consequence, expression of cPLA2εS420A did not induce clathrin-independent tubule growth and, therefore, failed to rescue MHC-I recycling in cPLA2ε-depleted cells. In contrast, overexpression of fully functional cPLA2ε triggered massive tubulation of clathrin-independent endocytic membranes that carry MHC-I (see Fig. 7), probably because of the strong increase in lysolipid production.

These findings indicate that cPLA2ε is a missing piece of the molecular puzzle that operates in membrane-remodeling events associated with recycling in the clathrin-independent route. Indeed, the core components of this puzzle have already been uncovered. This process apparently requires ARF6-dependent activation of PI(4)P5-kinase (Donaldson et al., 2009). PI(4)P5-kinase produces PI(4,5)P2, which allows binding of cPLA2ε to clathrin-independent endosomes, where the enzyme supports biogenesis of tubular membranes through the generation of wedge-shaped lysolipids (Fig. 9B). However, PI(4,5)P2 provides an attractive binding site for PLD2 (Donaldson, 2009), which produces negatively-curved phosphatidic acid (Donaldson et al., 2009; Yang et al., 2008; Zimmerberg and Kozlov, 2006) and, therefore, favors fission of the clathrin-independent endocytic tubular elements into recycling transport carriers (Donaldson, 2009). Thus, the interplay between cPLA2ε and PLD2 activities is likely to play a key role in the regulation of transport carrier biogenesis in the clathrin-independent route. Interestingly, similar roles in tubular biogenesis and fission were demonstrated for cPLA2α and PLD2 enzymes at the Golgi membranes (San Pietro et al., 2009; Yang et al., 2008; Yang et al., 2011). Thus, it will be important to identify the proteins that interact with cPLA2ε to control its activity and to understand how such proteins are integrated in the molecular signaling network that drives transport events within the clathrin-independent endocytic/recycling route.

Recombinant DNA constructs and cDNA mutagenesis

DNA was from the following sources: GFP–cPLA2ε from T. Shimizu (University of Tokyo, Japan), GFP–Rab22a from R. Weigert (NIH, USA), Myc–EHD1 from S. Caplan (University of Nebraska, USA), ARF6 from J. Donaldson (NIH, USA) and GFP–cPLA2α C. Leslie (NJMRC, USA), GFP-tagged PH-PLCδ, CERT-PH, GRP1-PH and SARA-FIVE was from M. A. De Matteis (TIGEM, Italy). The mouse cPLA2ε was then subcloned from pEGFP-C1 into different N-terminal tagged vectors. It was first amplified by PCR using the following primers: CGGAATTCAAATGCAGTCTATTCCACACT (forward) and ACCTCGAGACTAGGAGGGACACTGGCTCCT (reverse), and cloned into EcoRI and XhoI sites of the p-Myc–CMV vector (Clontech, Mountain View, CA), pm-Cherry-CMV vector (provided by G. Burgstaller) and pGEX4T.1 (GE Healthcare Life Sciences). Human cPLA2ε N-tail (amino acids 1–63) was PCR amplified from PLA2G4E human DNA using primers: CG GAA TTC ATG CGC GCC AAG TGC ATT AC (forward) and CCG CTC GAG AGG AAC CTG CTC AGC CCT C (reverse), and cloned into EcoRI and XhoI sites of the pGEX4T.1. The pGEX4T.1- cPLA2ε C-tail (amino acids 815–875) was PCR amplified from mouse pEGFP-cPLA2ε using primers: CG AGA TCT CTG GGC CAG CTG AAC ATC T (forward) and CG GAA TTC CTA GGA GGG ACA CTG GCT (reverse), and cloned into BglII and EcoRI sites of pGEX4T.1. Myc–cPLA2εS420A and Myc–cPLA2εΔPB were prepared by site-directed mutagenesis with the QuikChange kit (Stratagene, La Jolla, CA). For the Myc–cPLA2εS420A mutant, the forward and reverse primer sequences were: ACA TCA CCG GCT TGG CAG GTG CAA CCT GG (forward) and CCA GGT TGC ACC TGC CAA GCC GGT GAT GT (reverse). Myc–cPLA2εΔPB was obtained by introducing a stop codon at position 865. The forward and reverse primers were: TTG AGA CTA GCA ATG TAG AAG AAA CGC ATG A (forward) and T CAT GCG TTT CTT CTA CAT TGC TAG TCT CAA (reverse).

RNA interference

Oligonucleotide duplexes (5′-GGG UCA UGG CUG UCC AAU U-3′ and 5′-AAU UGG ACA GCC AUG ACC C-3′) targeting human cPLA2ε were from Qiagen. siCONTROL, a universal negative control siRNA, was from Sigma. SiRNA delivery into HeLa cells was executed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Cells were incubated with siRNA for 96 hours to achieve optimal cPLA2ε suppression (evaluated by immunoblot analysis). Quantification of western blots was performed using the ECL western blotting system (Amersham Pharmacia Biotech, Little Chalfont, UK).

Cells, reagents and antibodies

HeLa cells were grown in DMEM containing 10% FCS. Cytochalasin D, latrunculin A, BAPTA-AM, and saponin were from Sigma-Aldrich. Antibodies were obtained from the following sources: anti-MHC-I, anti-CD147, anti-CD55 and anti-CD98 from Biolegend (San Diego, CA), anti-EEA1 from Novus Biologicals (Littleton, CO), anti-c-Myc and anti-transferrin receptor from Sigma-Aldrich (St. Louis, MO), anti-Lamp1, anti-giantin and anti-GST from M. A. De Matteis (TIGEM, Naples, Italy), anti-M6PR from Chemicon (Temecula, CA), anti-HA from Covance (Princeton, NJ), anti-caveolin-1, anti-golgin-97 and anti-ARF6 from Santa Cruz Biotechnology (Dallas, TX), anti-Rabbit IgG-HRP and anti-Mouse IgG-HRP from Bethyl Laboratories (Montgomery, TX), secondary antibodies conjugated with Alexa Fluor 488 or Alexa Fluor 568, transferrin–Alexa-Fluor-568 and Alexa-Fluor-488–EGF were from Invitrogen (Carlsbad, CA). Generation of anti-cPLA2ε N-tail antibody was performed according standard protocols in the protein service facility.

Protein–lipid binding assay

PIP strips (Echelon Biosciences, Salt Lake City, UT) were blocked in TBST supplemented with 3% fatty-acid-free BSA and then incubated with 500 ng/ml of three different GST-fused cPLA2ε proteins: full-length cPLA2ε, N-tail cPLA2ε, C-tail cPLA2ε. GST-fused PH-PLCδ, p40-PX and SidC were used as positive control probes for PI(4,5)P2, PI3P and PI4P respectively, whereas GST alone was used as a negative control. At the end of the incubation, the PIP strips were washed with TBST, incubated with anti-GST antibody and revealed with an anti-mouse HRP conjugate (Amersham, Piscataway, NJ). After final washing, a chemiluminescence system (GE Healthcare, Oslo, Norway) was used to detect binding of the GST fusions to the phospholipids.

Immunofluorescence and antibody internalization

For immunofluorescence analyses, the cells were prepared and examined with a Zeiss LSM-710 confocal microscope (Carl Zeiss, Gottingen, Germany) as described previously (Polishchuk et al., 2003). To assess internalization of the MHC-I antibody, cells were incubated with 1 µg/ml antibody against human MHC-I on ice and washed with PBS to remove the unbound antibody. At this point, a set of coverslips were fixed and incubated with secondary antibody to label the surface pool of MHC-I. The remaining cells were incubated in pre-warmed medium for varying times to activate endocytosis of the antibody. After internalization, the cells were washed with cold PBS, the surface-bound MHC-I antibody was removed by acid wash. Finally, the cells were fixed and internalized MHC-I was revealed after using Alexa-Fluor-488- or Alexa-Fluor-546-conjugated secondary antibodies.

To perform recycling assays, an initial binding step was performed as described above. The cells were further incubated at 37°C in the presence of 10 µM cytochalasin D to allow internalization of the surface-bound MHC-I antibody and accumulation in the recycling compartment (Weigert, 2004). The antibody that still remained bound to the cell surface was removed by acid wash. At this point, a set of coverslips were fixed, permeabilized and incubated with a secondary antibody to label the internalized anti-MHC-I IgG. Another set of coverslips was washed in PBS and then incubated at 37°C for 30 minutes with complete medium to allow recycling of MHC-I. After this period, the treated cells were incubated with a low-pH buffer to remove the surface pool of antibody, then were fixed, permeabilized and incubated with the secondary fluorescent antibody to reveal the intracellular pool of MHC-I. CD55, CD98 and CD147 recycling assays were performed as described for MHC-I, using a 15 µg/ml concentration of the corresponding antibody.

In experiments where the internalization of Alexa-Fluor-546–transferrin (25 µg/ml) or Alexa-Fluor-488–EGF (100 ng/ml) was monitored, the cells were first incubated for 30 minutes in DMEM without serum and then were further kept at 37°C for varying times. Cells were then washed with cold PBS to remove the unbound transferrin and incubated at 37°C for varying times to allow internalization of Alexa-Fluor-546–Tfn. After internalization, the cells were washed with cold PBS (for Alexa-Fluor-546–Tfn, the surface-bound transferrin was removed by acid wash) and fixed.

Image analysis

Overlap of cPLA2ε with different markers of endocytic and secretory compartments was evaluated in confocal images using the colocalization module of LSM3.2 software. ROI function was used to analyze colocalization selectively in either tubular or vacuolar structures containing cPLA2ε. The amount of MHC-I, CD55, CD147 or transferrin at the surface or inside the cell was estimated as follows: 30–50 cells/coverslip were randomly selected, imaged with the pinhole completely open and with identical acquisition parameters. Under these conditions, the amount of internalized cargo was proportional to the average fluorescence. For each channel, the average fluorescence of each individual cell was measured using Zeiss LSM 3.2 software.

Phospholipase activity

HEK293T cells, transfected with pCMV-Myc vector or vectors encoding Myc–cPLA2ε, Myc– cPLA2εS420A were disrupted by sonication in a homogenizing buffer. Centrifugation at 10,000 g for 10 minutes and 100,000 g for 1 hour were carried out for fractionation, as reported previously (Ohto, 2005). The phospholipase reaction was started by the addition of enzyme sources (400 µg of supernatant from the 100,000 g centrifugation) to substrates [30,000 cpm/sample of [3H]-glycerol-labeled total cellular lipids, prepared as reported previously (Berrie et al., 2007)] in 100 mM HEPES-NaOH, pH 7.5, 1 mg/ml BSA, 4.5 mM CaCl2 and 1 mM dithiothreitol, and the reaction mixture was incubated at 37°C for 2 hours and 30 minutes. The assay was terminated by Bligh and Dyer extraction and the radioactivity recovered in the upper phase containing deacylated lipids was evaluated by liquid scintillation. The lower organic phase was dried and separated by thin layer chromatography (TLC) and the radioactivity associated with the labeled lysolipids was quantified by gas ionization scanning (Radio Isotope Thin Layer Analyzer, Raytest, Germany).

Electron microscopy

HeLa cells were transfected with Myc-tagged cPLA2ε or co-transfected with cPLA2ε–Myc and ARF6Q67L–GFP. In a subset of experiments the cells expressing cPLA2ε–Myc were loaded with Alexa-Fluor-488-conjugated antibodies against MHC-I for 1 hour to allow completion of the entire clathrin-independent endocytic pathway. Cryosections were prepared and labeled as described previously (Polishchuk et al., 2003) and investigated using a Tecnai G2 Spirit BioTWIN electron microscope (FEI, Einhoven, The Netherlands) equipped with a Veletta CCD camera for digital image acquisition.

We would like to acknowledge everybody who provided us with antibodies and DNA; Graciana Diez-Roux and Ellen Abrams for critical reading of the manuscript; and TIGEM Advanced Microscopy and Imaging Core for microscopy support.

Author contributions

M.C., S.M., R.B., A.L. and R.S.P. conceived and designed experiments. M.C., S.M., A.V.E., G.P., S.I., M.S., A.D.P., C.I., M.E. and G.D.T. performed experiments; M.C., S.M., G.P., S.I., M.V.E., R.B., A.L. and R.S.P. analyzed data; M.C. and R.S.P. wrote the manuscript.

Funding

This work was supported by the AIRC [grant number IG 10233]; and Telethon [grant number GTF08001] to R.S.P. Part of this work was supported by a fellowship to M.C. that was provided by Banca Popolare di Lanciano e Sulmona within “uno piu uno” project and by Fondazione Negri Sud onlus.

Bechler
M. E.
,
Doody
A. M.
,
Racoosin
E.
,
Lin
L.
,
Lee
K. H.
,
Brown
W. J.
(
2010
).
The phospholipase complex PAFAH Ib regulates the functional organization of the Golgi complex.
J. Cell Biol.
190
,
45
53
.
Bechler
M. E.
,
Doody
A. M.
,
Ha
K. D.
,
Judson
B. L.
,
Chen
I.
,
Brown
W. J.
(
2011
).
The phospholipase A2 enzyme complex PAFAH Ib mediates endosomal membrane tubule formation and trafficking.
Mol. Biol. Cell
22
,
2348
2359
.
Bechler
M. E.
,
de Figueiredo
P.
,
Brown
W. J.
(
2012
).
A PLA1-2 punch regulates the Golgi complex.
Trends Cell Biol.
22
,
116
124
.
Berrie
C. P.
,
Iurisci
C.
,
Piccolo
E.
,
Bagnati
R.
,
Corda
D.
(
2007
).
Analysis of phosphoinositides and their aqueous metabolites.
Methods Enzymol.
434
,
187
232
.
Bonifacino
J. S.
,
Rojas
R.
(
2006
).
Retrograde transport from endosomes to the trans-Golgi network.
Nat. Rev. Mol. Cell Biol.
7
,
568
579
.
Botelho
R. J.
,
Teruel
M.
,
Dierckman
R.
,
Anderson
R.
,
Wells
A.
,
York
J. D.
,
Meyer
T.
,
Grinstein
S.
(
2000
).
Localized biphasic changes in phosphatidylinositol-4,5-bisphosphate at sites of phagocytosis.
J. Cell Biol.
151
,
1353
1368
.
Brown
F. D.
,
Rozelle
A. L.
,
Yin
H. L.
,
Balla
T.
,
Donaldson
J. G.
(
2001
).
Phosphatidylinositol 4,5-bisphosphate and Arf6-regulated membrane traffic.
J. Cell Biol.
154
,
1007
1018
.
Cai
B.
,
Caplan
S.
,
Naslavsky
N.
(
2012
).
cPLA2α and EHD1 interact and regulate the vesiculation of cholesterol-rich, GPI-anchored, protein-containing endosomes.
Mol. Biol. Cell
23
,
1874
1888
.
Cantor
J. M.
,
Ginsberg
M. H.
(
2012
).
CD98 at the crossroads of adaptive immunity and cancer.
J. Cell Sci.
125
,
1373
1382
.
Caplan
S.
,
Naslavsky
N.
,
Hartnell
L. M.
,
Lodge
R.
,
Polishchuk
R. S.
,
Donaldson
J. G.
,
Bonifacino
J. S.
(
2002
).
A tubular EHD1-containing compartment involved in the recycling of major histocompatibility complex class I molecules to the plasma membrane.
EMBO J.
21
,
2557
2567
.
Casas
J.
,
Valdearcos
M.
,
Pindado
J.
,
Balsinde
J.
,
Balboa
M. A.
(
2010
).
The cationic cluster of group IVA phospholipase A2 (Lys488/Lys541/Lys543/Lys544) is involved in translocation of the enzyme to phagosomes in human macrophages.
J. Lipid Res.
51
,
388
399
.
Cullen
P. J.
(
2008
).
Endosomal sorting and signalling: an emerging role for sorting nexins.
Nat. Rev. Mol. Cell Biol.
9
,
574
582
.
de Figueiredo
P.
,
Doody
A.
,
Polizotto
R. S.
,
Drecktrah
D.
,
Wood
S.
,
Banta
M.
,
Strang
M. S.
,
Brown
W. J.
(
2001
).
Inhibition of transferrin recycling and endosome tubulation by phospholipase A2 antagonists.
J. Biol. Chem.
276
,
47361
47370
.
Dennis
E. A.
,
Cao
J.
,
Hsu
Y. H.
,
Magrioti
V.
,
Kokotos
G.
(
2011
).
Phospholipase A2 enzymes: physical structure, biological function, disease implication, chemical inhibition, and therapeutic intervention.
Chem. Rev.
111
,
6130
6185
.
Donaldson
J. G.
(
2009
).
Phospholipase D in endocytosis and endosomal recycling pathways.
Biochim. Biophys. Acta
1791
,
845
849
.
Donaldson
J. G.
,
Porat-Shliom
N.
,
Cohen
L. A.
(
2009
).
Clathrin-independent endocytosis: a unique platform for cell signaling and PM remodeling.
Cell. Signal.
21
,
1
6
.
Doody
A. M.
,
Antosh
A. L.
,
Brown
W. J.
(
2009
).
Cytoplasmic phospholipase A2 antagonists inhibit multiple endocytic membrane trafficking pathways.
Biochem. Biophys. Res. Commun.
388
,
695
699
.
Evans
J. H.
,
Spencer
D. M.
,
Zweifach
A.
,
Leslie
C. C.
(
2001
).
Intracellular calcium signals regulating cytosolic phospholipase A2 translocation to internal membranes.
J. Biol. Chem.
276
,
30150
30160
.
Evans
J. H.
,
Gerber
S. H.
,
Murray
D.
,
Leslie
C. C.
(
2004
).
The calcium binding loops of the cytosolic phospholipase A2 C2 domain specify targeting to Golgi and ER in live cells.
Mol. Biol. Cell
15
,
371
383
.
Eyster
C. A.
,
Higginson
J. D.
,
Huebner
R.
,
Porat-Shliom
N.
,
Weigert
R.
,
Wu
W. W.
,
Shen
R. F.
,
Donaldson
J. G.
(
2009
).
Discovery of new cargo proteins that enter cells through clathrin-independent endocytosis.
Traffic
10
,
590
599
.
Ferrone
S.
,
Finerty
J. F.
,
Jaffee
E. M.
,
Nabel
G. J.
(
2000
).
How much longer will tumour cells fool the immune system?
Immunol. Today
21
,
70
72
.
Gelderman
K. A.
,
Tomlinson
S.
,
Ross
G. D.
,
Gorter
A.
(
2004
).
Complement function in mAb-mediated cancer immunotherapy.
Trends Immunol.
25
,
158
164
.
Ghomashchi
F.
,
Naika
G. S.
,
Bollinger
J. G.
,
Aloulou
A.
,
Lehr
M.
,
Leslie
C. C.
,
Gelb
M. H.
(
2010
).
Interfacial kinetic and binding properties of mammalian group IVB phospholipase A2 (cPLA2beta) and comparison with the other cPLA2 isoforms.
J. Biol. Chem.
285
,
36100
36111
.
Ghosh
M.
,
Loper
R.
,
Gelb
M. H.
,
Leslie
C. C.
(
2006a
).
Identification of the expressed form of human cytosolic phospholipase A2beta (cPLA2beta): cPLA2beta3 is a novel variant localized to mitochondria and early endosomes.
J. Biol. Chem.
281
,
16615
16624
.
Ghosh
M.
,
Tucker
D. E.
,
Burchett
S. A.
,
Leslie
C. C.
(
2006b
).
Properties of the Group IV phospholipase A2 family.
Prog. Lipid Res.
45
,
487
510
.
Graham
T. R.
,
Kozlov
M. M.
(
2010
).
Interplay of proteins and lipids in generating membrane curvature.
Curr. Opin. Cell Biol.
22
,
430
436
.
Grant
B. D.
,
Donaldson
J. G.
(
2009
).
Pathways and mechanisms of endocytic recycling.
Nat. Rev. Mol. Cell Biol.
10
,
597
608
.
Guo
H.
,
Li
R.
,
Zucker
S.
,
Toole
B. P.
(
2000
).
EMMPRIN (CD147), an inducer of matrix metalloproteinase synthesis, also binds interstitial collagenase to the tumor cell surface.
Cancer Res.
60
,
888
891
.
Ha
K. D.
,
Clarke
B. A.
,
Brown
W. J.
(
2012
).
Regulation of the Golgi complex by phospholipid remodeling enzymes.
Biochim. Biophys. Acta
1821
,
1078
1088
.
Heo
W. D.
,
Inoue
T.
,
Park
W. S.
,
Kim
M. L.
,
Park
B. O.
,
Wandless
T. J.
,
Meyer
T.
(
2006
).
PI(3,4,5)P3 and PI(4,5)P2 lipids target proteins with polybasic clusters to the plasma membrane.
Science
314
,
1458
1461
.
Luini
A.
,
Ragnini-Wilson
A.
,
Polishchuck
R. S.
,
De Matteis
M. A.
(
2005
).
Large pleiomorphic traffic intermediates in the secretory pathway.
Curr. Opin. Cell Biol.
17
,
353
361
.
McMahon
H. T.
,
Gallop
J. L.
(
2005
).
Membrane curvature and mechanisms of dynamic cell membrane remodelling.
Nature
438
,
590
596
.
Naslavsky
N.
,
Weigert
R.
,
Donaldson
J. G.
(
2003
).
Convergence of non-clathrin- and clathrin-derived endosomes involves Arf6 inactivation and changes in phosphoinositides.
Mol. Biol. Cell
14
,
417
431
.
Ohto
T.
,
Uozumi
N.
,
Hirabayashi
T.
,
Shimizu
T.
(
2005
).
Identification of novel cytosolic phospholipase A(2)s, murine cPLA(2)delta, epsilon, and zeta, which form a gene cluster with cPLA(2)beta.
J. Biol. Chem.
280
,
24576
24583
.
Polishchuk
E. V.
,
Di Pentima
A.
,
Luini
A.
,
Polishchuk
R. S.
(
2003
).
Mechanism of constitutive export from the golgi: bulk flow via the formation, protrusion, and en bloc cleavage of large trans-golgi network tubular domains.
Mol. Biol. Cell
14
,
4470
4485
.
Polishchuk
R. S.
,
Capestrano
M.
,
Polishchuk
E. V.
(
2009
).
Shaping tubular carriers for intracellular membrane transport.
FEBS Lett.
583
,
3847
3856
.
Saftig
P.
,
Klumperman
J.
(
2009
).
Lysosome biogenesis and lysosomal membrane proteins: trafficking meets function.
Nat. Rev. Mol. Cell Biol.
10
,
623
635
.
San Pietro
E.
,
Capestrano
M.
,
Polishchuk
E. V.
,
DiPentima
A.
,
Trucco
A.
,
Zizza
P.
,
Mariggiò
S.
,
Pulvirenti
T.
,
Sallese
M.
,
Tete
S.
 et al. (
2009
).
Group IV phospholipase A(2)alpha controls the formation of inter-cisternal continuities involved in intra-Golgi transport.
PLoS Biol.
7
,
e1000194
.
Schaloske
R. H.
,
Dennis
E. A.
(
2006
).
The phospholipase A2 superfamily and its group numbering system.
Biochim. Biophys. Acta
1761
,
1246
1259
.
Sprong
H.
,
van der Sluijs
P.
,
van Meer
G.
(
2001
).
How proteins move lipids and lipids move proteins.
Nat. Rev. Mol. Cell Biol.
2
,
504
513
.
Tucker
D. E.
,
Ghosh
M.
,
Ghomashchi
F.
,
Loper
R.
,
Suram
S.
,
John
B. S.
,
Girotti
M.
,
Bollinger
J. G.
,
Gelb
M. H.
,
Leslie
C. C.
(
2009
).
Role of phosphorylation and basic residues in the catalytic domain of cytosolic phospholipase A2alpha in regulating interfacial kinetics and binding and cellular function.
J. Biol. Chem.
284
,
9596
9611
.
Weigert
R.
,
Yeung
A. C.
,
Li
J.
,
Donaldson
J. G.
(
2004
).
Rab22a regulates the recycling of membrane proteins internalized independently of clathrin.
Mol. Biol. Cell
15
,
3758
3770
.
Yang
J. S.
,
Gad
H.
,
Lee
S. Y.
,
Mironov
A.
,
Zhang
L.
,
Beznoussenko
G. V.
,
Valente
C.
,
Turacchio
G.
,
Bonsra
A. N.
,
Du
G.
 et al. (
2008
).
A role for phosphatidic acid in COPI vesicle fission yields insights into Golgi maintenance.
Nat. Cell Biol.
10
,
1146
1153
.
Yang
J. S.
,
Valente
C.
,
Polishchuk
R. S.
,
Turacchio
G.
,
Layre
E.
,
Moody
D. B.
,
Leslie
C. C.
,
Gelb
M. H.
,
Brown
W. J.
,
Corda
D.
 et al. (
2011
).
COPI acts in both vesicular and tubular transport.
Nat. Cell Biol.
13
,
996
1003
.
Zimmerberg
J.
,
Kozlov
M. M.
(
2006
).
How proteins produce cellular membrane curvature.
Nat. Rev. Mol. Cell Biol.
7
,
9
19
.
Zizza
P.
,
Iurisci
C.
,
Bonazzi
M.
,
Cossart
P.
,
Leslie
C. C.
,
Corda
D.
,
Mariggiò
S.
(
2012
).
Phospholipase A2IVα regulates phagocytosis independent of its enzymatic activity.
J. Biol. Chem.
287
,
16849
16859
.

Competing interests

The authors declare no competing interests.

Supplementary information