Involvement of Class II Phosphoinositide 3-Kinase α-Isoform in Antigen-Induced Degranulation in RBL-2H3 Cells

In this study, we present findings that suggest that PI3K-C2α, a member of the class II phosphoinositide 3-kinase (PI3K) subfamily, regulates the process of FcεRI-triggered degranulation. RBL-2H3 cells were transfected with shRNA targeting PI3K-C2α. The knockdown impaired the FcεRI-induced release of a lysosome enzyme, β-hexosaminidase, without affecting the intracellular Ca2+ mobilization. The release of mRFP-tagged neuropeptide-Y, a reporter for the regulated exocytosis, was also decreased in the PI3K-C2α-deficient cells. The release was increased significantly by the expression of the siRNA-resistant version of PI3K-C2α. In wild-type cells, FcεRI stimulation induced the formation of large vesicles, which were associated with CD63, a marker protein of secretory granules. On the vesicles, the existence of PI3K-C2α and PtdIns(3,4)P2 was observed. These results indicated that PI3K-C2α and its product PtdIns(3,4)P2 may play roles in the secretory process.


Introduction
Mast cell activation mediated by the high-affinity receptor for IgE (FceRI) is a key event in allergic inflammatory responses [1]. Cross-linking IgE-bound FceRI triggers a rapid release of granule contents, including histamine, serotonin, and proteases. Previous studies have demonstrated the role of phosphoinositide 3-kinase (PI3K) in this process. The members of the PI3K family are lipid kinases that catalyze the phosphorylation of the 3-position of the inositol ring of phosphoinositides. PI3K can be grouped into three major classes (I, II, and III) based on their primary sequences, mechanism of regulation, and substrate specificities (for a review, please see Ref. [2]). In the process of FceRI-mediated degranulation, the roles of class I subtypes, namely PI3Kd and PI3Kc, have been demonstrated [3,4].
Although less investigated than class I subtypes, recent studies have shown that the class II subtypes of PI3K are involved in a variety of cell functions [5,6]. Mammals possess three class II isoforms: PI3K-C2a, PI3K-C2b and PI3K-C2c. PI3K-C2a and C2b are widely expressed in mammalian tissues. Human PI3K-C2c showed a more restricted localization in the liver, prostate, breast and salivary glands [6,7]. A previous study has demonstrated that siRNA against the class II isoform PI3K-C2b decreases the FceRI-mediated Ca 2+ influx and degranulation of bone marrow-derived mast cells (BMMCs) [8].
Another class II isoform, PI3K-C2a, has been implicated in several vesicle trafficking pathways [9][10][11][12][13][14]. As expected from the clathrin-binding motif in its N-terminal region [9], it was demonstrated that PI3K-C2a regulates clathrin-dependent endocytosis [9,13]. Several studies have suggested the involvement of PI3K-C2a in exocytosis pathways, including translocation of glucose transporter type 4 to the plasma membrane of muscle cells, catecholamine release from adrenal chromaffin cells and insulin secretion from pancreatic b-cells [10][11][12]14,15]. However, the role of PI3K-C2a in the process of mast cell degranulation has not been reported to date. In the present study, we present results demonstrating that PI3K-C2a is involved in the exocytosis pathway of mast cells.

Results
We first examined the expression of PI3K-C2a and PI3K-C2b mRNA in RBL-2H3 cells. Reverse transcriptase-PCR with specific primers showed that PI3K-C2a and PI3K-C2b mRNA is expressed in RBL-2H3 cells ( Figure S1). Because PI3K-C2b has been reported to regulate the FceRI-induced Ca 2+ influx and degranulation in BMMCs [8], we examined whether PI3K-C2a plays any role in the cells. To this end, we prepared RBL-2H3 cells expressing shRNA against PI3K-C2a. Two lines of cells that produce shRNA against the different sequences (seq1 or seq2) of PI3K-C2a were prepared. In the seq1-and seq2-targeted cells, the levels of PI3K-C2a mRNA were 37% and 27%, respectively, of the level observed in the control vector-transfected cells (Figure 1A). The PI3K-C2b mRNA was unaffected by the shRNA ( Figure 1A). The protein levels of PI3K-C2a in the seq1-and seq2-targeted cells, as determined by western blotting with a specific antibody, were 20% and 9.9%, respectively, of the levels observed in the control cells ( Figure 1B). The protein levels of PI3K-C2b were not significantly affected by the shRNAs ( Figure 1B).
The effect of PI3K-C2a knockdown on the FceRI-triggered release of a lysosomal enzyme, namely b-hexosaminidase, was examined ( Figure 2A). The b-hexosaminidase release was decreased significantly in both PI3K-C2a-knockdown cells. The total granule content of b-hexosaminidase was unchanged by the shRNA transfection ( Figure 2B). The results suggested that PI3K-C2a is required for efficient degranulation via FceRI. When RBL-2H3 cells were treated with calcium ionophore and phorbol ester simultaneously, a significant amount of b-hexosaminidase was released. This response was, however, unaffected by PI3K-C2a knockdown ( Figure 2C). The pan-PI3K inhibitor wortmannin, which inhibits PI3K-C2a with an IC 50 value of 420 nM [16], efficiently decreased the FceRI-triggered degranulation at 1 mM but did not alter the calcium ionophore/phorbol ester-induced response ( Figure 2D).
Neuropeptide Y (NPY) is a genuine reporter for the regulated exocytosis of mast cells [17]. In the experiments shown in Figure 3A, NPY-mRFP and EGFP-PI3K-C2a were transfected into RBL-2H3 cells. Upon stimulation, the mRFP fluorescence of the control cells decreased gradually, indicating the release of NPY from the cells. The NPY release from the PI3K-C2a-knockdown cells (shPI3K-C2a cells) was markedly slowed. The NPY release was then examined in the cells transfected with the shRNAresistant PI3K-C2a construct. The overexpression of PI3K-C2a significantly increased the NPY-mRFP release from the control and shPI3K-C2a cells ( Figure 3B), confirming the role of PI3K-C2a as a positive regulator of degranulation. A slight decrease in the mRFP fluorescence in the unstimulated cells ( Figure 3B) may be due to the spontaneous release of NPY and fluorescence bleaching.
The increase in intracellular Ca 2+ upon FceRI stimulation is one of the key events triggering degranulation [18]. A recent study has shown that PI3K-C2b regulates the FceRI-stimulated activation of the KCa3.1 channel and degranulation in BMMCs [8]. Thus, we investigated whether the down-regulation of PI3K-C2a influences the intracellular Ca 2+ mobilization following FceRI stimulation. Ca 2+ imaging analysis revealed a rapid increase in intracellular Ca 2+ upon FceRI stimulation in the control and PI3K-C2a-knockdown cells, with no difference between these cell lines ( Figure 4). The above results suggested that PI3K-C2a regulates the FceRI-triggered degranulation by a mechanism different from that induced by PI3K-C2b.
We then examined the effect of PI3K-C2a knockdown on the granule dynamics following FceRI stimulation. In the experiment shown in Figure 5A, RBL-2H3 cells were stimulated with antigen, fixed, and then stained with a specific antibody against CD63, a marker protein of secretory granules (also known as lysosomeassociated membrane protein 3, LAMP3). In the resting state, most of the CD63-positive granules had a diameter smaller than 1 mm, as has been reported in a previous study [19]. Upon antigen stimulation, CD63-positive large vesicles (.2.5-mm diameter) appeared in more than a half of the cell population ( Figure 5A, B) likely due to the granule-granule fusion and the granule swelling that occur during the process of exocytosis [20][21][22][23].
Previous studies have indicated that PI3K-C2a can catalyze the phosphorylation of PtdIns and PtdIns(4)P to generate PtdIns(3)P and PtdIns(3,4)P 2 , respectively [13]. To gain insight into the function of PI3K-C2a, the PH domain of Akt (Akt-PH) that recognizes PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 was transfected into RBL-2H3 cells. In the unstimulated cells, a large amount of EGFP-labeled Akt-PH was found on the plasma membrane ( Figure 5E). This signal may be due to the products of class I PI3K because it can be observed similarly in the control and shPI3K-C2a cells. Upon the FceRI stimulation, CD63-positive large vesicles associated with Akt-PH appeared in the control cells. Such vesicles were rarely observed in the shPI3K-C2a cells ( Figure 5E, G). Similar results were obtained when the tandem PH domain of TAPP1 (26TAPP1-PH) was used as a selective probe for PtdIns(3,4)P 2 ( Figure S2A, C). A similar experiment was then performed with the PtdIns(3)P probe 36FYVE EEA1 (the FYVE domain of EEA1). The FceRI stimulation tended to increase the 36FYVE EEA1 association with the CD63-positive large vesicles, but this response was observed even in the absence of PI3K-C2a ( Figure 5F, G). We next used another PtdIns(3)P probe, 26FYVE Hrs (the tandem FYVE domain of Hrs), which has been successfully used to monitor the PI3K-C2a-mediated PtdIns(3)P production at secretory vesicles in PC12 cells [14]. The experiment with 26FYVE Hrs produced similar results to that with 36FYVE EEA1 (Figure S2B, C). The C215S mutant of the 26FYVE Hrs , which does not bind PtdIns(3)P [24], did not associate with the vesicles ( Figure S2D). These results suggested that PI3K-C2a and its product PtdIns(3,4)P 2 may play some roles in the process of FceRI-mediated degranulation.
We then examined the intracellular localization of PI3K-C2a. In the experiment shown in Figure 6A, RBL-2H3 cells were transfected with EGFP-tagged PI3K-C2a and mRFP-tagged NPY and stained with the specific CD63 antibody. Before stimulation, PI3K-C2a was found on the intracellular small vesicles and ruffling membrane. After antigen stimulation, PI3K-C2a-associated CD63-positive large vesicles appeared in about 45% of the cell population ( Figure 6A, B). Nearly 40% of this cell population possessed one or more NPY-containing vesicles, whereas the others possessed empty vesicles only ( Figure 6B). It has been reported that granules, which had released their contents, are recycled (60%) or collapse to the plasma membrane (40%) in RBL-2H3 cells [22]. Therefore, the above results suggest the association of PI3K-C2a with the secretory granules.
In the experiment shown in Figure 6C, RBL-2H3 cells were transfected with EGFP-26TAPP1-PH, and the endogenous PI3K-C2a and CD63 were stained with specific antibodies. The result showed that PtdIns(3,4)P 2 locates on the large vesicles that associate with PI3K-C2a and CD63. The existence of PtdIns(3,4)P 2 on the large vesicles was confirmed when a specific antibody against PtdIns(3,4)P 2 was used for the analysis (Figure 6D). These results suggested that PI3K-C2a and its product PtdIns(3,4)P 2 play some roles in the process of FceRI-mediated degranulation.

Discussion
PI3K-C2a has been implicated in several exocytosis pathways [10][11][12]14,15]. In the present study, we found the first demonstration of the involvement of PI3K-C2a in the antigen-induced degranulation pathway. We observed that the expression of shRNAs against PI3K-C2a significantly reduced the FceRImediated b-hexosaminidase release ( Figure 2). The release of NPY-mRFP, a reporter of the regulated exocytosis of mast cells  [17], was increased by PI3K-C2a overexpression and was impaired in the PI3K-C2a-deficient cells (Figure 3). The impaired exocytosis in PI3K-C2a-deficient cells was significantly rescued by shRNA-resistant PI3K-C2a expression (Figure 3).
A class II PI3K isoform, PI3K-C2b, has been reported to regulate the FceRI-induced activation of the KCa3.1 channel in BMMCs [8]. In T cells, this isoform is required for the TCRstimulated activation of the KCa3.1 channel and Ca 2+ influx [25]. In the former, siRNA against PI3K-C2b impairs the FceRI-induced increase in intracellular Ca 2+ and degranulation [8]. In the present study, we observed that impaired degranulation in the PI3K-C2a knockdown cells did not accompany a decrease in the Ca 2+ response (Figure 4). The result indicates that PI3K-C2a and PI3K-C2b share different roles in the FceRI-regulated degranulation pathway. The biochemical background for this functional difference should be examined.
Secretory granules in mast cells have been shown to enlarge upon antigen-stimulation due to granule-granule fusion and/or granule swelling [20][21][22][23]. In the present study, we observed the formation of CD63-positive large vesicles in the FceRI-stimulated RBL-2H3 cells ( Figure 5A and B). CD63, also known as LAMP3, is a membrane protein containing a lysosome-targeting domain [26]. In RBL-2H3 cells, CD63 co-localizes with the SNARE proteins syntaxin3 and VAMP7, both of which are involved in the fusion of secretory granules with plasma membranes [27]. In BMMCs from the CD63-knockout mice, FceRI-induced degranulation was impaired [28]. In the present study, we observed the existence of PI3K-C2a on the CD63-positive large vesicles ( Figure 6A), suggesting a link between PI3K-C2a and CD63 during FceRI-mediated degranulation. On this point, it is intriguing to note that the PMA/ionomycin-induced degranulation, which is unaffected in the shPI3K-C2a cells ( Figure 2C), is not impaired in CD63-knockout mice [28]. It has been reported that mast cells have three types of secretory granule subset [29]. The subset driven by ionomycin and PMA is reported to be distinct from the FceRI-driven one [30]. PI3K-C2a may be involved in the exocytosis of FceRI-specific granules.
Previous studies have indicated that class II PI3Ks have the in vitro ability to phosphorylate PtdIns and PtdIns(4)P to generate PtdIns(3)P and PtdIns(3,4)P 2 , respectively [16]. It has been reported that PtdIns(3)P is the sole product in the presence of Ca 2+ as the divalent cation [6,31,32]. In contrast, a recent study demonstrated that the immunopurified PI3K-C2a preferentially produces PtdIns(3,4)P 2 in the presence of both Mg 2+ and Ca 2+ [13]. Regarding the in vivo product of class II PI3Ks, previous studies have yielded conflicting data on the preference of PtdIns(3)P or PtdIns(3,4)P 2 as a product [10][11][12][13][14]. In the present study, we observed the association of Akt-PH and 26TAPP1-PH, markers of PtdIns(3,4)P 2 , with CD63-positive large vesicles, and this association was impaired markedly in the shPI3K-C2a cells ( Figure 5E, G, S2A and C). The presence of PtdIns(3,4)P 2 on the vesicles of normal cells can be confirmed by blotting with a specific antibody ( Figure 6D). The results suggest that PtdIns(3,4)P 2 plays a functional role in PI3K-C2a-induced degranulation.
We observed that PtdIns(3)P-positive large vesicles were formed in the antigen-stimulated RBL-2H3 cells ( Figure 5G, S2C). The number of these vesicles was lower than that of the PtdIns(3,4)P 2positive vesicles (,0.1360.03 and 0.4760.07 in each control cell, respectively). PI3K-C2a knockdown did not impair the formation of the PtdIns(3)P-positive vesicles ( Figure 5G, S2C). It has been reported that crosslinking FceRIs induces membrane ruffling [33], which in turn increases the formation of a macropinosome, which has a diameter of 0.5-10 mm and follows a fate similar to that of endosomes [34]. Early endosomes are reported to be rich in PtdIns(3)P. Thus, the PtdIns(3)P-positive vesicles observed in our present study are considered to be a macropinosome.
As described above, our results suggest that PI3K-C2a promotes antigen-induced degranulation through PtdIns(3,4)P 2 production. This speculation is in accordance with previous studies that suggest that PI3K-C2a specifically phosphorylates PtdIns(4)P in the process of clathrin-mediated endocytosis [13] and in insulinstimulated MIN6 cells [11]. In the former case, the role of PtdIns(3,4)P 2 to recruit SNX9 to the clathrin-coated pits has been demonstrated [13]. In contrast, several studies including those on the insulin secretion of pancreatic b-cells [15], the catecholamine release from adrenal chromaffin cells [12,14] and the glucose transporter type 4 translocation in insulin-stimulated cells [10] have indicated that PI3K-C2a regulates the later steps of exocytosis by producing PtdIns(3)P. A further study to identify the effector molecules of PtdIns(3,4)P 2 in mast cells is warranted to confirm the role of PI3K-C2a in FceRI-mediated degranulation.

RNA isolation and RT-PCR
The total RNA from RBL-2H3 cells was isolated with Sepasol-RNA I Super G (Nacalai tesque, Kyoto, Japan), and first-strand cDNA was synthesized by reverse transcription using random

Degranulation assay
Degranulation was determined as the release of the granule marker b-hexosaminidase. RBL-2H3 cells cultured on a 96-well plate were sensitized with anti-DNP IgE (100 ng/ml) for 1 h in culture medium. IgE-sensitized RBL-2H3 cells were washed twice with PIPES buffer (119 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 0.4 mM MgCl 2 , 5.6 mM glucose, 1 mg/ml BSA, and 25 mM PIPES, pH 7.6) and then incubated at 37uC for 10 min. After addition of the indicated concentrations of DNP-BSA or vehicle, the reaction plate was maintained at 37uC for 10 min. For receptor-independent stimulation, unsensitized cells were incubated in PIPES buffer with PMA (30 nM) for 10 min and then stimulated with A23187 (1 mM) for 10 min. The cell pellets were solubilized in PIPES buffer containing 1% Triton X-100. The bhexosaminidase activities of the supernatants and the solubilized pellets were measured by incubating with 4-nitrophenyl N-acetylbeta-D-glucosaminide (Sigma) in sodium citrate (pH 4.5) for 1 h at 37uC. Then, 0.5 M Tris(hydroxymethyl)aminomethane was used to stop the reaction, and the absorbance was read at 405 nm. Degranulation is expressed as a percentage of b-hexosaminidase activity in the supernatant divided by the total (supernatant plus pellet) activity.

Measurement of changes in intracellular Ca 2+
IgE-sensitized RBL-2H3 cells in multi-well, glass-bottom dishes (Matsunami Glass) were incubated with Fluo-8 dye (AAT Bioquest Inc., Sunnyvale, CA, USA) in RPMI 1640 medium supplemented with 25 mM HEPES and 0.1% FBS at 25uC for 20 min. The cells were washed twice with PIPES buffer, stimulated with 1 mM DNP-BSA, and then observed under a fluorescence microscope (BZ-9000; Keyence, Tokyo, Japan) at 30uC using a GFP-BP filter. Fluorescence images were collected every 20 s, and the fluorescence intensities (F) of the individual cells were quantified with a BZ-II analyzer (Keyence). The data are shown as DF/F 0 , where F 0 is the fluorescence intensity before stimulation and DF is the difference between F and F 0 . The data were obtained from three separate experiments (24 cells were monitored in total).

Transfection
The transfection of plasmid DNA was conducted using the Neon transfection system (Life Technologies Co., Carlsbad, CA, USA) according to the manufacturer's instructions. In brief, RBL-2H3 cells (10 5 cells) were washed with PBS and resuspended in 10 mL of R buffer containing 1 mg of plasmid DNA. The resuspended cells were then transferred into a gold tip and electroporated by two pulses at 1200 V for 20 ms followed by incubation in growth media without antibiotics for 48 h.

Immunocytochemistry
IgE-sensitized RBL-2H3 cells in glass-bottom dishes were stimulated with 1 mg/mL DNP-BSA for 3 min. After washing with PBS, the cells were fixed with PBS containing 4% formaldehyde for 15 min. The cells were permeabilized with PBS containing 0.3% Triton X-100 and 0.5% BSA for 60 min, incubated with anti-CD63 (1:500 dilution) or anti-PtdIns(3,4)P 2 (1:200 dilution) and/or anti-PI3K-C2a antibody (1:250 dilution) at 4uC overnight and then incubated with Alexa 488-or Alexa 647labeled (Fab9) 2 fragment of goat anti-mouse IgG (1:1,000 dilution) and/or Alexa 555-labeled (Fab9) 2 fragment of goat anti-rabbit IgG (1:1,000 dilution) for 2 h at room temperature. For the cells transfected with NPY-mRFP, the permeabilization was done with PBS containing 0.05% saponin and 3% BSA for 10 min. Microscopic analysis was performed using the Keyence BZ-9000 with CFI Plan Apo VC60xH lens (Keyence, Osaka, Japan). To obtain improved optical resolution along the z-axis, z-stacks were captured at 1-mm steps over a Z-axis distance of 3 mm, and individual planes as well as the sum projection of entire stacks were compared.

Release of Neuropeptide-Y
RBL-2H3 cells expressing neuropeptide Y (NPY)-mRFP were transfected with either EGFP or shRNA-resistant EGFP-tagged PI3K-C2a. The cells in glass-bottom dishes were sensitized with IgE and then stimulated with 1 mg/mL DNP-BSA. The intensity of the red fluorescence in the cells showing EGFP expression was monitored under a fluorescence microscope using a Texas Red filter.

Statistical analysis
Statistical significance was determined using unpaired, two-tail distribution, student's t-test. Data indicated with one asterisk or two asterisks have values of *P,0.05 or **P,0.01, respectively. Figure S1 mRNA expression of class II PI3K in RBL-2H3 cells. PCR using RBL-2H3 cDNA as the template was performed with primers specific for PI3K-C2a or PI3K-C2b.