The Identification of Phosphatidylinositol 3,5-bisphosphate in T-lymphocytes and Its Regulation by Interleukin-2*

In recent times 3-phosphoinositides have emerged as important regulators of cell metabolism, survival, and proliferation. During the last year, the phospholipid phosphatidylinositol 3,5-bisphosphate (PtdIns3,5P2) was identified in yeast, fibroblasts, SV40-transformed kidney (COS-7) cells, and platelets. The discovery of this novel phospholipid has increased the complexity of the metabolism relating to the generation of biologically active inositol-containing lipids. We describe here the identification of PtdIns3,5P2 in the CTLL-2 mouse T-lymphocyte cell line using two in vivo radiolabeling protocols. Treatment of the cells with UV radiation led to an increase in the cellular content of PtdIns3,5P2. In contrast, preincubation of the cells with wortmannin or treatment with hypertonic medium (high concentration sorbitol) led to the opposite effect. Herein we demonstrate that interleukin-2 (IL-2), the growth factor required for CTLL-2 cell proliferation, was able to increase the level of PtdIns3,5P2 with similar kinetics to that of the formation of phosphatidylinositol 3,4-bisphosphate (PtdIns3,4P2). An increase in this novel 3-phosphorylated lipid in response to IL-2 seems to be a general property of this cytokine because a similar result was obtained when the pre-B cell line BaF/3 expressing the high affinity IL-2 receptor was used. Using a constitutively active regulatory subunit of type I phosphatidylinositol 3-kinase and cells expressing a deletion of the serine-rich domain of the IL-2 receptor β chain, which is required for IL-2-stimulated type I phosphatidylinositol 3-kinase activation, we demonstrate that IL-2-induced generation of PtdIns3,5P2 is related to the activation of this enzyme. The results show for the first time the identification of PtdIns3,5P2 in both T- and B-lymphocytes and indicate its positive regulation by the mitogen IL-2.

Phosphoinositides represent a small (less than 5%) proportion of the total cell phospholipids, yet they play crucial roles in the regulation of cell metabolism through their involvement in intracellular signaling mechanisms. In the last decade there has been a large increase in the number of reports describing the phosphorylation on the 3Ј-hydroxyl of inositol giving rise to a small family of 3-phosphorylated phosphoinositides consisting of phosphatidylinositol 3-phosphate (PtdIns3P), 1 PtdIns3,4,P 2 , and phosphatidylinositol 3,4,5-trisphosphate (PtdIns3,4,5P 3 ). These 3-phosphorylated phosphoinositides are not substrates for activated phospholipase C isoenzymes, but they have biological activity per se as they are able to activate downstream effectors such as PKB/Akt and modulate several metabolic processes (1)(2)(3). Currently there are three families of mammalian PI3-K enzymes (type I, II, and III), the former two consisting of multiple isoforms (4). The substrate specificities of the PI3-Ks show some differences. Type I PI3-Ks preferentially phosphorylate PtdIns4,5P 2 (over PtdIns and phosphatidylinositol 4-phosphate (PtdIns4P)) in vivo (5), whereas the type II PI3-Ks prefer PtdIns and PtdIns4P as substrates, and PtdIns4,5P 2 is poorly used (6,7). Type III PI3-K is only able to phosphorylate PtdIns (8). Additionally, other PtdIns kinases involved in the regulation of the levels of 3-phosphorylated phosphoinositides have been characterized; these include isoforms of the PtdIns4P 5-kinase family which are able to synthesize the 3-phosphoinositides PtdIns3,4P 2 and PtdIns3,4,5P 3 (9) and the unusual phospholipids PtdIns3,5P 2 and phosphatidylinositol 5-phosphate (10). One of the latest developments in the phosphoinositide field has been the identification in vivo of the novel phospholipid PtdIns3,5P 2 in yeast, fibroblasts, COS-7 cells, and platelets (11)(12)(13). PtdIns3,5P 2 has been proposed to arise from PtdIns3P via a PtdIns3P 5-kinase (11,12), which, in the case of mammalian cells, is a wortmannin-sensitive process (11). Studies with yeast and COS-7 cells have indicated that PtdIns3,5P 2 might be an intracellular signaling molecule involved in controlling responses to stress (12).
IL-2 is the growth factor responsible for the proliferation of T-cells. Its signal transduction mechanisms have been studied in depth revealing that multiple signal transduction cascades are initiated after just seconds of receptor ligation (14). IL-2 does not cause the activation of the classical PtdIns cycle (15)(16)(17), but instead is a strong activator of the PI3-K pathway (18 -20), which has been shown to be crucial for cell prolifera-tion (21) and activation of the transcription factor E2F via PKB/Akt (22). IL-2 stimulation of T-cells induces the rapid elevation of PtdIns3,4P 2 and PtdIns3,4,5P 3 , but to this date, no analysis of PtdIns3,5P 2 either in resting or proliferating lymphocytes has been reported. The present study describes the presence of PtdIns3,5P 2 in both T-and B-lymphocytes. We also report here for the first time that mitogenic stimulation by IL-2 increases the level of this novel PtdInsP 2 isoform. Cell Culture, Radiolabeling, and Treatments-CTLL-2 cells were maintained and grown for experiments according to previously published work (23,24). When necessary, the CTLL-2 cells were transfected according to the protocol described by Jimenez et al. (24). Murine IL-3-dependent BaF/3 cells expressing the full-length human IL-2 receptor ␤ chain (Baf␣/␤) or a deletion mutant of the IL-2 receptor ␤ chain lacking the serine-rich domain (amino acids 267-322) (Baf␣/␤SD1) were generated as described previously (21). The cells were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM glutamine, 50 M 2-mercaptoethanol, 100 units/ml penicillin, 100 g/ml streptomycin buffered to a pH of 7.2 with 10 mM HEPES, with either 50 units/ml recombinant IL-2 or 10% WEHI.3-conditioned medium as a source of IL-3, respectively. When cells reached a density of approximately 10 6 /ml, they were washed twice in phosphate-free RPMI 1640 medium supplemented with 2 mM glutamine, 50 M 2-mercaptoethanol, 100 units/ml penicillin, and 100 g/ml streptomycin buffered to a pH of 7.2 with 10 mM HEPES before reincubating during a starving period of 6 h in the presence of 0.1% bovine serum albumin. For in vivo radiolabeling, 600 Ci/ml [ 32 P]orthophosphate was included during the last 2 h of the starving period. After this period, the cells were challenged with control medium, medium containing recombinant IL-2, or other stimuli for varying times. For in vivo metabolic radiolabeling with myo-2-[ 3 H]inositol, CTLL-2 cells were grown for a period of 6 days in RPMI 1640 medium that contained a reduced concentration (5 M) of unlabeled myo-inositol supplemented with 10% v/v 1000-Da molecular weight-cut off fetal calf serum, 2 mM MgCl 2 , and 50 units/ml IL-2. myo-2-[ 3 H]inositol was included in the medium during days 2-6 at a concentration of 10 Ci/ml. At the end of cell treatments, the cells were pelleted, washed once with ice-cold phosphate-buffered saline, and immediately frozen on dry ice. Approximately 15 ϫ 10 6 cells were used for each incubation/extraction, yielding a total incorporation of [ 32 P] radioactivity into phospholipids of approximately 4.5 ϫ 10 6 cpm and 5 ϫ 10 5 cpm in the case of [ 3 H].

Materials
Analysis of Polyphosphoinositides by HPLC-In vivo production of 3-phosphorylated lipids was determined essentially as described previously (39). Briefly, phospholipids were extracted using CHCl 3 /MeOH/ HCl in the presence of 5 mM EDTA, 5 mM tetrabutylammonium hydrogen sulfate, and 10 g of a mixture of unlabeled PtdIns, PtdIns4P, and PtdIns4,5P 2 to act as cold carriers. The isolated lower organic phases were washed once with freshly prepared artificial acidified aqueous methanolic top phase containing 5 mM EDTA and 5 mM tetrabutylammonium hydrogen sulfate. Washed organic phases were transferred into a screw-capped glass vial and dried under a gentle stream of nitrogen gas. 500 l of freshly prepared methylamine reagent (1-butanol, methanol, 25% aqueous methylamine, 11.5/45.7/42.8, v/v) containing 1 mM EDTA was added to the glass vials before heating at 53°C for 1 h. After cooling to room temperature, the contents of the vials were transferred to an Eppendorf tube and vacuum-dried at room temperature. After resuspending the mixture in 500 l of H 2 O, fatty acids and any undeacylated lipids were removed by washing twice with 500 l of the freshly prepared solvent mixture consisting of 1-butanol, petroleum ether, ethyl formate (20:4:1, v/v). More than 95% of the radioactivity routinely partitioned into the aqueous phase, indicating almost complete phospholipid deacylation. The aqueous phase was vacuum-dried before storage at Ϫ70°C. Separation of all the deacylated phospholipids was performed by HPLC employing a Partisphere SAX column (4.6 ϫ 235 mm, 5 m, Whatman) with a gradient of 0 -1 M ammonium phosphate, pH 3.75, over 120 min. The gradient consisted of 0 -10 min 100% for pump A, 10 -70 min linear rise to 25% for pump B, 70 -120 min steep linear rise to 100% for pump B. The pump and column washout was from 120 to 130 min with 100% for pump A (H 2 O). Radiolabeled deacylated phospholipids were detected by on-line radiochemical monitoring (Beckman Instruments, Inc., Fullerton, CA and EG & G Berthold, Bad Wildbad, Germany). Peak-associated radioactivity was expressed as % of total radioactivity detected to eliminate any inter-sample variation. 3-Phosphorylated phosphoinositide HPLC standards ([ 32 P]GroPIns3P, [ 32 P]GroPIns3,4P 2 , and [ 32 P]GroPIns3,4,5P 3 ) were prepared by the action of PI3-K (immunoprecipitated from 10 8 CTLL-2 cells using an anti-p85 antibody) on the substrates PtdIns, PtdIns4P, and PtdIns4,5P 2 in the presence of [␥ 32 P]ATP followed by their deacylation as described previously (24). In addition, [
It has been reported that treatment of eukaryotic and prokaryotic cells with agents that induce stress result in the modulation of the levels of PtdIns3,5P 2 (12). To study the regulation of PtdIns3,5P 2 in CTLL-2 cells, we treated the cells with various defined agents to determine their effects on the level of PtdIns3,5P 2 , PtdIns3P, PtdIns3,4P 2 , and PtdIns3,4,5P 3 . Cells that were treated with UV radiation for 20 min increased their PtdIns3,5P 2 by approximately 75% above control (Fig. 2). Furthermore, UV radiation caused an approximately 2-fold increase in the radioactivity associated with PtdIns3P, a small (15%) decrease in the level of PtdIns3,4P 2 , and a large decrease (80%) in PtdIns3,4,5P 3 compared with control-treated cells (Fig. 3). In response to hypertonic shock (high concentration sorbitol), a decrease (to approximately 65% of control) in the radiolabeling associated with PtdIns3,5P 2 was observed (Fig.   2), which was in agreement with that seen in COS-7 cells (12). Fig. 3 also indicates that the action of sorbitol on the levels of PtdIns3P, PtdIns3,4P 2 , and PtdIns3,4,5P 3 was slight, as only slight (10 -30%) decreases in their levels were observed. In addition to agents that induce stress, the level of PtdIns3,5P 2 has also been shown to be sensitive to the PI3-K inhibitor wortmannin (11). To further characterize our cell system and to determine whether PtdIns3,5P 2 in CTLL-2 cells showed similar wortmannin sensitivity, we treated the cells with the PI3-K inhibitor. The radioactivity associated with PtdIns3,5P 2 fell to approximately 25% of control (Fig. 2). In addition, the levels of PtdIns3P, PtdIns3,4P 2 , and PtdIns3,4,5P 3 were affected by treatment with wortmannin. The former two phospholipids suffered moderate decreases (between 20 and 50%), whereas PtdIns3,4,5P 3 was reduced by more than 90% (Fig. 3).
In the light of the discovery of PtdIns3,5P 2 (11,12), it has now become important to completely separate all three isomers (PtdIns3,5P 2 , PtdIns3,4P 2 , and PtdIns4,5P 2 ) of PtdInsP 2 to determine their cellular levels before and after cell stimulation. For that reason we decided to investigate whether IL-2 affected the level of PtdIns3,5P 2 in CTLL-2 cells by using our SAX-HPLC system capable of high resolution PtdInsP 2 isomer separation. Fig. 4 shows that IL-2 stimulated an early and transient increase in the PtdIns3,4,5P 3 content of the cells (approximately 75% above control). This finding is in contrast to the slow rise in the amount of PtdIns3,4P 2 (approximately a 2-fold increase above control at 20 min). As is the case for other mitogens, IL-2 caused no consistent perturbation of the amount of PtdIns3P in CTLL-2 cells at all the time points examined (data not shown). Our results agree with those reported by Remillard and co-workers (18). In response to IL-2, there was a gradual accumulation of PtdIns3,5P 2 (approximately 75% above control at 20 min). This increase in PtdIns3,5P 2 displayed kinetics similar to those in the accumulation of PtdIns3,4P 2 . The increase in PtdIns3,4,P 2 at 20 min was approximately 50% higher than for PtdIns3,5P 2 .
To demonstrate that the mitogenic activity of IL-2 correlated with the generation of PtdIns3,5P 2 , we determined the accumulation of this novel lipid in Baf/3 cells expressing either the wild-type human ␤ chain of the IL-2 receptor or a deletion of the serine-rich domain within the human ␤ chain of the IL-2 receptor, which abolishes IL-2-stimulated PI3-K activity and cell proliferation. In Baf␣/␤ cells, IL-2 caused an approximately 2-fold increase in the radioactivity associated with both PtdIns3,5P 2 and PtdIns3,4P 2 after 20 min of incubation, which was completely inhibited by wortmannin (Fig. 5). These IL-2stimulated increases in PtdIns3,5P 2 and PtdIns3,4P 2 were not observed in Baf␣/␤SD1 cells (Fig. 5). Both types of cells responded to UV radiation in a manner similar to that seen in CTLL-2 cells by increasing the radioactivity found in the PtdIns3,5P 2 fraction (approximately 2-fold) and decreasing by approximately 30% that found in PtdIns3,4P 2 (Fig. 5). So far, the results suggested that IL-2-activated type I PI3-K activity was a prerequisite for an increase in the cellular PtdIns3,5P 2 content. To consolidate these findings, we chose to take advantage of the recently described constitutively active regulatory subunit of type I PI3-K p65 (24). In vivo metabolic radiolabeling and HPLC analysis of p65-transfected cells indicated virtually no increases in PtdIns3P compared with wild-type cells, despite the increase in the levels of both PtdIns3,4P 2 and PtdIns3,4,5P 3 by approximately 50 -60% (Fig. 6), which was in agreement with previous results (24). When the level of PtdIns3,5P 2 was analyzed in CTLL-2 cells expressing p65, a dramatic increase in the level of this lipid, more than 3-fold compared with empty vector-transfected cells, was observed (Fig. 7). When both cell types were stimulated with IL-2 for 20 min, the PtdIns3,5P 2 content of empty vector-transfected cells and p65 vector-transfected cells was increased by approximately 80 and 110%, respectively (Fig. 7). DISCUSSION We have demonstrated for the first time the existence of an agonist (IL-2)-stimulated pathway leading to the accumulation of PtdIns3,5P 2 in T-and B-lymphocytes. In addition to lymphocytes, PtdIns3,5P 2 has been identified (11)(12)(13) and speculated to exist (25)(26)(27) in both yeast and other mammalian cells.
The mechanism for the formation of PtdIns3,5P 2 appears to be through the enzymatic conversion of PtdIns3P by a 5-kinase. Evidence for this hypothesis has come from in vivo radiolabel-ing procedures indicating that the last phosphate group attached to the PtdIns3,5P 2 is the group that has the highest specific activity (11,12). Furthermore, this activity is, at least in mouse fibroblasts, sensitive to wortmannin (11). The responsible enzyme, Fab1p, has recently been described in yeast (28,29). To date no information is available concerning its mammalian counterpart. Treatment of cells with UV radiation is considered to be a stress signal that immediately causes the activation of various enzymes including the c-Jun amino-terminal protein kinase (JNK) cascade (30). Because UV irradiation of cells is anti-mitogenic, it is consistent with the observation in our cells of a decrease in phospholipids considered to be mitogenic (PtdIns3,4P 2 and PtdIns3,4,5P 2 ). Our results indicated that the enzyme activated in response to UV radiation was a PI3-K specific for the phosphorylation of PtdIns, most likely a type III PI3-K. The rise in PtdIns3P was far greater than the rise in PtdIns3,5P 2 , which suggests that only a small proportion of the PtdIns3P is used for PtdIns3,5P 2 synthesis, leaving the majority for other roles that could include regulation of endocytosis, membrane trafficking, and protein sorting (31,32). Another possible role of the PtdIns3P generated in response to UV radiation could be that of the activation of the JNK cascade. In this regard, JNK activation has been shown to be mediated by a PI3-K exhibiting wortmannin/LY294002 sensitivity (33).
The proliferation of T-lymphocytes is a process primarily controlled by IL-2. It activates various signal transduction pathways including the ras-MAPK (mitogen-activated protein kinase), PI3-K/Akt, and Jak/STAT cascades (14). IL-2 is a strong stimulator of the PI3-K pathway, and it is essential for IL-2-driven cell proliferation. IL-2-stimulated PI3-K has been relatively well characterized and has involved the use of in vivo metabolic radiolabeling experiments to demonstrate increases in PtdIns3,4P 2 and PtdIns3,4,5P 3 (18 -20). However, to date no analysis of PtdIns3,5P 2 generation in lymphocytes in response to IL-2 has been performed. IL-2 stimulation of CTLL-2 cells also increases PtdIns3,5P 2 levels, although the mechanism responsible for the elevation of this novel lipid is not completely clear. Results from our experiments using Baf␣/␤ cells and Baf␣/␤SD1 cells have helped to clarify this point. Within the ␤ chain of the IL-2 receptor is a serine-rich cytoplasmic domain that is responsible for the association of Jak1 (34). Deletion of this domain abolishes PI3-K activation following IL-2 binding (21,35), probably because of the role of Jak1 in p85 association (36). In Baf␣/␤SD1 cells, no such domain is present, and no increase in the PtdIns3,5P 2 content of the cells was found in response to IL-2 stimulation. This finding suggests that type I PI3-K activity is required for IL-2-stimulated PtdIns3,5P 2 generation. As a consequence of irradiation with UV light, both cells types were able to elevate the cellular level of PtdIns3,5P 2 to approximately the same level, thereby demonstrating that their response to a non-receptor-mediated event was identical and fully independent of the type of IL-2 receptor ␤ chain that they expressed. We have previously demonstrated that CTLL-2 cells stably transfected with a mutant form of the regulatory subunit of PI3-K, p65, induces constitutive activation of PI3-K (24). We have used this model herein to determine its effects on the accumulation of PtdIns3,5P 2 . In these cells the PtdIns3,5P 2 content was elevated with respect to empty vector-transfected cells. In addition, IL-2 stimulation of these cells induces an approximately two-fold elevation of the level of PtdIns3,5P 2 , as it does in wild-type CTLL-2 cells. It can be concluded from these experiments that in CTLL-2 cells expressing a constitutively active allele of PI3-K, there is a constitutive elevation of the PtdIns3,5P 2 level. This observation coincides with recent work by Klippel and co-workers (27) who have demonstrated that an inducible constitutively activated PI3-K was able to increase the levels of PtdIns3,4P 2 and PtdIns3,4,5P 3 and that of a peak in the HPLC profile putatively identified as PtdIns3,5P 2 . Interestingly, IL-2 stimulation of CTLL-2 and Baf␣/␤ cells as well as control-treated p65-transfected CTLL-2 cells results in the generation of PtdIns3,5P 2 without increases in PtdIns3P. Although the increase in PtdIns3,5P 2 provoked by IL-2 was approximately the same as seen in response to UV irradiation, the apparent lack of PtdIns3P production suggested that the pathway for the formation of PtdIns3,5P 2 did not totally overlap that initiated by UV radiation, despite the fact that IL-2 stimulates JNK activation (37). Generation of PtdIns3P is usually attributable to type III PI3-K activity, which has been described as having constitutive activity in both resting and actively growing cells (4). It has been generally accepted that PtdIns is not a good substrate for type I PI3-K in vivo. These observations are based on the fact that no accumulation of PtdIns3P is detected following mitogen stimulation. However, when p65 is transiently expressed in COS-7 cells, its PI3-K activity is extremely high, which forces the in vivo elevation of PtdIns3P, PtdIns3,4P 2 , and PtdIns3,4,5P 3 levels. 2 In this situation the degradation of PtdIns3P is delayed as compared with p65 stably transfected cells. This observation would suggest that the PtdIns3P generated in vivo following type I PI3-K activation, either by IL-2 receptor ligation or p65 cell transfection, is removed rapidly by one or more of the following enzymes: a PtdIns3P 3-phosphatase, a PtdIns3P 4-kinase, or a PtdIns3P 5-kinase. It could be envisaged that in response to IL-2, and for that matter other ligands, no net increases in PtdIns3P are observed because of its rapid removal via one or more of the three enzymes mentioned above.
PtdIns3,5P 2 represents a new member of the family of 3-phosphoinositides, which are known to be essential for the regulation of numerous cellular activities. However, to date PtdIns3P and PtdIns3,5P 2 have not been assigned definitive roles. Observations of the involvement of phosphoinositides during protein translocation and membrane reorganization or protein trafficking events are the possibilities so far proposed (31,32,38). PtdIns3,5P 2 may be a strong candidate for a true second messenger. To substantiate this claim, work focusing on its generation through a PtdIns3P 5-kinase, its subcellular site of synthesis, its possible relocalization, and its interaction with proteins, which may or may not contain pleckstrin homology domains (which bind polyphosphoinositides), will lead the way to understanding why cells have a requirement for yet another isomer of PtdInsP 2 , particularly in the case of mitogenic signals such as that delivered by IL-2 in T-lymphocytes.