Hypertonic Stress Increases Phosphatidylinositol 4,5-Bisphosphate Levels by Activating PIP5KIβ*

Hyperosmotic stress increases phosphoinositide levels, reorganizes the actin cytoskeleton, and induces multiple acute and adaptive physiological responses. Here we showed that phosphatidylinositol 4,5-bisphosphate (PIP2) level increased rapidly in HeLa cells during hypertonic treatment. Depletion of the human type I phosphatidylinositol 4-phosphate 5-kinase β isoform (PIP5KIβ) by RNA interference impaired both the PIP2 and actin cytoskeletal responses. PIP5KIβ was recruited to membranes and was activated by hypertonic stress through Ser/Thr dephosphorylation. Calyculin A, a protein phosphatase 1 inhibitor, blocked the hypertonicity-induced PIP5KIβ dephosphorylation/activation as well as PIP2 increase in cells. Urea, which raises osmolarity without inducing cell shrinkage, did not promote dephosphorylation nor increase PIP2 levels. Disruption or stabilization of the actin cytoskeleton, or inhibition of the Rho kinase, did not block the PIP2 increase nor PIP5KIβ dephosphorylation. Therefore, PIP5KIβ is dephosphorylated in a volume-dependent manner by a calyculin A-sensitive protein phosphatase, which is activated upstream of actin remodeling and independently of Rho kinase activation. Our results establish a cause-and-effect relation between PIP5KIβ dephosphorylation, lipid kinase activation, and PIP2 increase in cells. This PIP2 increase can orchestrate multiple downstream responses, including the reorganization of the actin cytoskeleton.

All cells experience fluctuations in osmolarity. Unicellular organisms and plants continuously confront osmotic challenges in their environment. In higher animals, the kidney and the gastrointestinal system are routinely exposed to severe osmotic fluctuation, while the majority of cells in other organs are protected from large tonicity changes. Nevertheless, these other organs are also confronted with transient osmolarity variations due to changes in the transmembrane transport of solutes or shifts in the balance between low molecular weight pre-cursors and their macromolecular products. Recently, there has been a renewed interest in understanding the mechanism of hypertonic response in the clinical arena (1), due to the discovery that treatments using hypertonic resuscitation in experimental models of trauma, hemorrhagic shock, sepsis, and burn injury are more beneficial than conventional isotonic resuscitation (2,3). While the fundamental mechanism for such protection is not completely understood, the actin cytoskeleton, which is reorganized during hypertonic stress, has been implicated (3,4).
Actin remodeling as well as many of the other hyperosmotic responses are evolutionary conserved. These include large shifts in phosphoinositide metabolism, activation of the mitogen-activated protein and tyrosine kinase pathways, volume regulation and the reprogramming of gene transcription (3,5). Phosphatidylinositol 4-phosphate (PI4P) 4 and phosphatidylinositol 4,5-bisphosphate (PIP 2 ) levels increase dramatically in mammalian cardiac muscle and tissue culture cells that were exposed to hypertonic sucrose or NaCl (6). Other phosphoinositides, including phosphatidylinositol 3,5-bisphosphate (7,8), phosphatidylinositol 3,4-bisphosphate, and phosphatidylinositol 3,4,5-trisphosphate are increased in some other types of cells as well (9).
Hyperosmotic stress acutely induces cell shrinkage, which is subsequently corrected by volume regulation. The response cascade can be classified into four mechanistic components: volume sensing, signal generation, signal transduction, and effector activation (10). Given that PIP 2 is a multifunctional regulator and adaptor, the PIP 2 increase may coordinate many of the responses to hypertonic stress. Nevertheless, the role of PIP 2 in the response hierarchy, and the mechanism for PIP 2 increase, have yet to be elucidated. PIP 2 levels are maintained by a dynamic balance between PIP 2 synthesis and degradation. PIP 2 is synthesized primarily by the type I PIP5Ks, which phosphorylate PI4P on the D5 position of the inositol ring. PIP 2 is degraded primarily by phosphoinositide phosphatases and by PI-phospholipase Cs. Three major PIP5KI isoforms, called ␣, ␤, and ␥, have been identified (11). Additionally, the ␥ isoform is differentially spliced to generate several variants (12,13).
Despite the large cadre of potential regulatory mechanisms, nothing is known about the regulation of the PIP5KIs by hypertonic stress. In this paper, we have identified the hypertonic stress regulated PIP5KI and examined the relation between its activation by Ser/Thr dephosphorylation to the hypertonicityinduced increase in PIP 2 production and actin remodeling.
Hypertonic Treatments-HeLa cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. 250 mM sucrose or 100 mM or 150 mM NaCl was added to the isotonic growth medium during hypertonic stimulation.
TLC-Cells were labeled for 4 h in phosphate-free Dulbecco's modified Eagle's medium and 40 Ci/ml [ 32 P]PO 4 and were then exposed to sucrose or NaCl. Samples were processed for TLC as described previously (23).
High Performance Liquid Chromatography (HPLC)-Lipids were extracted, deacylated, and analyzed on anion exchange HPLC columns. Negatively charged glycerol head groups were eluted with a NaOH gradient and detected on-line by suppressed conductivity (6,31). Individual peaks were identified with glycerophosphoryl inositol standards. Peak assignment was validated by spiking some cell samples with purified phospholipids as standards.
RNA Interference-The small interfering RNA (siRNA) sequences targeting the three human PIP5KI isoforms individually are as described previously (33,34). Firefly luciferase siRNA (nucleotides 695-715) was used as a negative control.
In Vitro PIP5KI Kinase Assay-Lipid kinase activity was measured by phosphorylation of PI4P using [␥-32 P]ATP as a phosphate donor. Sepharose G beads containing immunoprecipitated epitope-tagged PIP5KI were suspended in a solution containing 25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl 2 , 10 mM EDTA, 0.1 mM EGTA, 0.4% Nonidet P-40, 10% glycerol, 1 mM dithiothreitol, 1 mM sodium vanadate, 70 M PI4P (Biomol Inc.) and 35 M phosphatidylserine (Avanti). Kinase assay was initiated by adding [␥-32 P]ATP (1 Ci/50 l reaction with a final concentration of 0.2 mM ATP, PerkinElmer Life Sciences), and reaction proceeded at room temperature for 15 min. The reactions were stopped by adding CHCl 3 :MeOH:HCl, and lipids were extracted as described above and separated by TLC. Radioactivity associated with the PIP band that comigrates with a PI4P standard was quantitated by PhosphorImager analysis, and the amount of PIP5KI protein in the equivalent immunoprecipitate was determined by Western blotting. Kinase activity was normalized to the amount of immunoprecipitated protein and expressed as percent of the activity without sucrose stimulation. Under the conditions of our experiments, the kinase activity was linear from 5 to 30 min.
Membrane Fractionation-Cell homogenates were centrifuged at 100,000 ϫ g to obtain the cytosol and membrane fractions (33). Organelles and organelle membranes were separated according to their density by a multistep centrifugation procedure (36). Samples obtained by centrifugation sequentially at 19,000 ϫ g, 41,000 ϫ g (low speed), and 180,000 ϫ g (high speed) were analyzed. Plasma membrane enriched sample was obtained by floatation on a sucrose cushion. The pelleted fractions were adjusted to 100 l, and equal volumes of membrane fractions were analyzed by Western blotting with anti-HA antibody.
Isolation of the Triton X-100-insoluble Actin Cytoskeleton-Cells were lysed in Triton X-100 and centrifuged sequentially at low speed (15,900 ϫ g) to collect cross-linked actin stress fibers and networks (low speed pellet) and at high speed (336,000 ϫ g) to sediment long filaments that were not cross-linked (high speed pellet).

Hypertonic Stress Increased the PIP 2 Level in Cells-When
HeLa cells were exposed to 250 mM sucrose for 10 min, the [ 32 P]PIP 2 level increased to a greater extent than [ 32 P]PI4P (Fig.  1A and Table 1). Likewise, addition of 100 mM NaCl to normal culture medium also increased PIP 2 levels to a much higher extent than PI4P (Table 1). Therefore, hypertonic stress induced by these two stimuli preferentially increases PIP 2 level.
The increase in phosphoinositide levels was evident within 2 min of sucrose stimulation and reached a maximum by 10 min (Fig. 1A). While there was a subsequent slow decline, PIP 2 was still above prestimulation levels at 20 min. The rapid increase in [ 32 P]PIP 2 and [ 32 P]PI4P suggests that the change in phosphoinositide turnover/biosynthesis is an early response to hypertonic stimulation.
HPLC, which measures the amount of each lipid and can separate the different types of phosphoinositides better than the TLC method (31), confirmed that there was a 2-fold increase in PIP 2 ( Fig. 1B and Table 1). Unexpectedly, there was almost no change in PI4P mass, despite a modest increase in its labeling by 32 P (Fig. 1A). No new peak corresponding to either PI3P, PI(3,4)P 2 , or PI(3,4,5)P 3 was detected (Fig. 1B). The response profile of the HeLA cell establishes that PIP 2 is increased selectively. This increase is most likely due to the direct activation of PIP5KIs and/or inactivation of PIP 2 phos-phatsases and is unlikely to be due to increased availability of the PI4P substrate.
The sites of PIP 2 increase were identified by immunofluorescence microscopy. As shown previously (34), the PIP 2 antibody stained small punctae that lined the plasma membrane (PM pool) and a perinuclear region (internal organelle pool) (Fig.  1C). Hypertonic treatment increased anti-PIP 2 staining intensity, especially in large punctae that line the plasma membrane. Some of these punctae were at the tips of retraction fibers that were formed when the cell shrunk (see Fig. 2B).
PIP5KI␤ Depletion Blunted the Hypertonicity-induced PIP 2 and Actin Responses-siRNA oligonucleotides were used to identify the PIP5KI isoform that is primarily responsible for the hypertonic stress-induced PIP 2 increase. As reported previously, PIP5KI␤ depletion by RNAi (33,34) decreased the basal [ 32 P]PIP 2 level to the greatest extent compared with depletion of the other PIP5KI isoforms ( Fig. 2A). PIP5KI␤ completely blocked the hyprtonicity-induced PIP 2 increase, while depletion of the other PIP5KIs had much less effect. These results showed that PIP5KI␤ accounts for most of the hypertonicityinduced PIP 2 increase. We therefore focused on its behavior for the remainder of this paper.
HeLa cells normally have long actin stress fibers and cortical actin filaments. After hypertonic stimulation, the stress fibers became thicker, and retraction fibers were formed at the cell periphery as the cell shrunk (Fig. 2B). An increase in stress fibers and polymerized actin was confirmed biochemically by isolating the Triton X-100-insoluble cytoskeleton (Fig. 2C). The amount of actin in the Triton-insoluble low speed pellet (23), which contains cross-linked actin filaments such as stress fibers, and the Tritoninsoluble high speed pellet, which contains long actin filaments that are not crosslinked sufficiently to be sedimented by centrifugation are increased, while actin in the high speed supernatant (representing actin monomers and small oligomers) decreased.  Therefore, hypertonicity promotes actin polymerization and cross-linking into stress fibers and/or networks. PIP5KI␤ depletion dramatically changed the cell shape and decreased the amount of stress fibers (Fig. 2B). Although hypertonic stress shrunk the PIP5KI␤ RNAi-treated cells, it did not promote stress fiber formation. Since PIP5KI␤ depletion blocked the actin polymerization/reorganization response, we conclude that PIP5KI␤ has a major role in orchestrating the hypertonic stress fiber response.
PIP5KI␤ Was Recruited to Membranes by Hypertonic Stress-Immunofluorescence was used to examine the effect of hypertonic stress on the subcellular distribution of PIP5KI␤. PIP5KI␤ was cytosolic and also associated with plasma membrane and endomembranes (Fig. 2D). Hypertonic stress increased the amount of PIP5KI␤ at the cell periphery and the retraction fibers also contained PIP5KI␤.
The increase in PIP5KI␤ membrane association was confirmed by subcellular fractionation using two different methods. Centrifugation at 100,000 ϫ g showed that under isotonic conditions, ϳ45% of the kinase was recovered in the pellet (membranes) (Fig. 3A). HA-PIP5KI␤ migrated as a doublet in both the supernatant (cytosol) and pellet, which, as will be shown later, was due to a difference in the extent of phosphorylation. Hypertonic stimulation collapsed the doublet into a single band and increased the recovery of PIP5KI␤ (70% of total) in the pellet fraction.
Multistep fractionation (36) provided additional information about the partitioning of PIP5KI␤ among different organelle fractions (Fig. 3B). Under isotonic conditions, slightly less than half of the total PIP5KI␤ was membrane associated, and of this, half was recovered in the PM enriched fraction (Fig. 3B). Sucrose stimulation decreased the percentage of PIP5KI␤ in the cytosolic fraction by 60% and almost doubled the percentage recovered in the PM and high speed pellet fraction. Therefore, the immunofluorescence and biochemical data both show that there is an increase PIP5KI␤ membrane association.
Hypertonic Stress Dephosphorylated PIP5KI␤ but Not the Other PIP5KIs-Park et al. (29) showed that PIP5KI isoforms are constitutively phosphorylated and that they can be activated by Ser/Thr dephosphorylation. Hypertonic stress col-  OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 lapsed the HA-PIP5KI␤ doublet into a single band in SDSpolyacrylamide gels, which would be consistent with dephosphorylation (Fig. 3A). Dephosphorylation was confirmed by a decrease in 32 P labeling of the upper band in the doublet. Under isotonic conditions, both HA-PIP5KI␤ bands were 32 P-labeled, and the upper band was more highly phosphorylated (Fig. 4A). The 32 P intensity of the upper band decreased dramatically within 3 min of sucrose treatment, while that of the lower band was not decreased to a similar extent. Our results suggest that PIP5KI␤ is constitutively phosphorylated at multiple sites and that a subset of these sites is preferentially dephosphorylated as an early response to hypertonic stress. The time course of dephosphorylation of the upper band paralleled the rise in PIP 2 level (Fig. 1A), lending further support to the possibility that PIP5KI␤ dephosphorylation increases PIP 2 levels in cells.
Effects of PIP5KI␤ Dephosphorylation-The effect of hypertonic stress on the lipid kinase activity of PIP5KI␤ was examined using an in vitro kinase assay. HA-PIP5KI␤ that was immunoprecipitated from hypertonically stressed cells had three times higher lipid kinase activity than those from the unstimulated control, while the activities of PIP5KI␣, -␥L, and -␥S were not changed significantly (Fig. 4C). Therefore, hypertonic stress selectively activates PIP5KI␤.
We next examined the effect of inhibiting Ser/Thr protein phosphatases on the hypertonic response. We tested the following Ser/Thr phosphatase inhibitors: caly A inhibits PP1 and PP2A; okadaic acid inhibits PP2A at 1-10 nM concentrations (IC 50 0.51 nM) and PP1 at higher concentrations (IC 50 42 nM). Cyclosporine A inhibits PP1B but not PP1A.
Caly A increased the intensity of the 32 P label in the upper band of the PIP5KI␤ doublet under isotonic conditions and blocked the sucrose-induced dephosphorylation (Fig. 5A). Okadaic acid had no effect at 10 nM (data not shown) but did increase HA-PIP5KI␤ basal phosphorylation and blocked dephosphorylation at 100 nM. The differential effects of caly A and okadaic acid on PIP5KI␤ dephosphorylation suggest that the PP1 phosphatases promote PIP5KI␤ dephosphorylation during hypertonic stress. Cyclosporine A did not block dephosphorylation (data not shown), ruling out a PP2B involvement.
Caly A was used to evaluate the relationship between the hyperonicity-induced PIP5KI␤ dephosphorylation and lipid kinase activation. We found that caly A decreased basal PIP5KI␤ activity by 63% and blocked PIP5KI␤ activation by sucrose (Fig. 5B). We also used caly A to determine whether PIP5KI␤ dephosphorylation is a primary trigger for the hypertonic PIP 2 response. Caly A dampened the PIP 2 response (Fig.  5C), and this effect was specific for PIP 2 , because PI4P increased normally. It is curious though that caly A had minimal effect on the PIP 2 level of the cell under isotonic condition (Fig. 5C), even though it inhibited PIP5KI␤ in vitro (Fig. 5B). It is possible that PIP 2 did not decrease in the calyculin A-treated cells because of compensatory changes that restore the ambient isotonic PIP 2 level. However, these compensations are not able to raise PIP 2 to a sufficiently high level to compensate for the lack of PIP5KI␤ activation during hypertonic stress. Taken together, the series of experiments establish that there is a cause and effect relationship between hypertonicity-induced PIP5KI␤ dephosphorylation, lipid kinase activation, and PIP 2 increase in cells.
Effects of PIP5KI␤ Dephosphorylation on Its Steady State Membrane Association-Since hypertonicity induces PIP5KI␤ dephosphorylation and also promotes its recruitment to membranes (Figs. 2-4), we examined the possibility that the more dephosphorylated PIP5KI␤ is preferentially membrane associated. However, the ratio of the two bands in the PIP5KI␤ doublet in the 100,000 ϫ g supernatant and pellet fractions (Fig. 3A) were similar. Therefore, the more phosphorylated and less phosphorylated PIP5KI␤ associate with membranes to a similar extent under the steady state isotonic conditions used here. We conclude that the increase in membrane association during hypertonic stress cannot be simply attributed to dephosphorylation.
The Relationship between the Hypertonicity-induced PIP 2 Response, Volume Change, and Actin Remodeling-Hypertonic NaCl or sucrose induces cell shrinkage and reorganization of the actin cytoskeleton. In contrast, 200 mM urea, which is cell permeant and therefore increases osmolarity without inducing cell shrinkage, did not increase PIP 2 nor PIP5KI␤ dephospho-rylation (Fig. 6A). Therefore, the PIP 2 and dephosphorylation responses are both dependent on volume changes.
Cell shrinkage deforms the acin cytoskeleton and imposes mechanical tension on the integrins (37). Since the actin cytoskeleton can potentially act as a volume sensor and PIP5KIs are activated by integrin signaling (4,15,38,39), we investigated the possibility that PIP5KI␤ dephosphorylation depends on signals transmitted by the actin cytoskeleton. Latrunculin A, which depolymerizes actin filaments, and jasplakinolide, which stabilizes actin filaments, were used to interfere with the cytoskeletal response. We found that neither blocked the PIP 2 increase nor PIP5KI␤ dephosphorylation (Fig. 6B). Therefore, FIGURE 4. Hypertonic stress promotes the dephosphorylation of PIP5KI␤ but not the other PIP5Ks. A, time course of PIP5KI␤ dephosphorylation. HeLa cells overexpressing HA-PIP5KI␤ were labeled with 32 P and incubated with isotonic or hypertonic (250 mM sucrose) medium for the periods indicated. HA-PIP5KI␤ was immunoprecipitated and analyzed by PhosphorImager analysis and Western blotting. Duplicate samples are shown. B, phosphorylation of PIP5KI isoforms. Experimental conditions were similar to that described above. C, PIP5KI activity. PIP5KIs were immunoprecipitated with anti-epitope antibodies and used for in vitro lipid kinase assays and for Western blotting. The amount of 32 P incorporated into PIP 2 was determined by PhosphorImager analysis after TLC. Kinase activity was normalized against the amount of immunoprecipitated PIP5KIs (based on Western blots; data not shown) and the specific activity ( 32 P/protein) of sucrose-treated samples was expressed as percent of an equivalent sample from cells not exposed to hypertonic stress. Data were mean Ϯ S.E. of three different experiments, each done in duplicate.

FIGURE 5. Effects of caly A.
HeLa cells overexpressing HA-PIP5KI␤ were exposed to 5 nM caly A for 15 min prior to sucrose stimulation. A, phosphorylation. Cells were 32 P-labeled, and HA-PIP5KI␤ was immunoprecipitated with anti-HA antibody. B, in vitro lipid kinase assay using immunoprecipitated HA-PIP5KI␤. 32 P incorporation into PI4P was normalized against the amount of immunoprecipitated kinase (determined by parallel Western blots; data not shown) and expressed as a percentage of control. Values are mean Ϯ S.E. of three independent experiments, each done in duplicate. C, the phosphoinositide response in cells. The amounts of 32 P incorporated into the bands comigrating with the standard PI4P and PIP 2 on TLC were plotted in arbitrary units (mean Ϯ S.E. of three independent experiments).
PIP5KI␤ dephosphorylation and PIP 2 increase do not directly depend on cytoskeletal remodeling.
Under isotonic conditions, the small GTPase Rho and its downstream effector Rho kinase stimulate stress fiber formation and promote PIP5KI targeting and activation (15,22,40). Since Rho is activated by hypertonic stress (41,42), we examined the possibility that Rho kinase activation promotes PIP5KI␤ dephosphorylation. This is not the case because the Rho kinase inhibitor, Y27843, did not block PIP5KI␤ dephosphorylation (Fig. 6C) even though it inhibited actin stress fiber formation (data not shown).

DISCUSSION
In this paper, we report that hypertonic stress activates and targets PIP5KI␤ to promote high levels of site-specific PIP 2 generation. The increase in PIP 2 levels is a direct result of PIP5KI␤ activation by Ser/Thr dephosphorylation.
The involvement of PIP5KI␤ was established by RNA intereference. The lack of response in cells depleted of PIP5KI␤ was not due to nonspecific cell death, because PIP5KI␤ RNAi has surprisingly little effect on another signaling pathway (34). Furthermore, although the other PIP5KIs are not major players in the hypertonic response, they have important roles that have been identified by RNAi depletion (34) and by gene knock-out (43). Our results support the growing realization that the PIP5KIs are functionally specialized, and by extension, the PIP 2 pools of the cell may be functionally and perhaps even physically segregated.
The phosphorylation status of PIP5KI␤ is likely to be maintained by a balance between protein kinases and phosphatases. Theroretically, the hypertonicity-induced PIP5KI␤ dephosphorylation could be achieved either by inhibiting a PIP5KI␤ protein kinase or activating its phosphatase. However, although hypertonicity activates many kinases (5,44), there are relatively few examples of hypertonicity-inactivated kinases. Therefore, our current working hypothesis is that PIP5KI␤ is dephosphorylated primarily by activation of a caly A-sensitive PP1. The involvement of PP1 is supported by the ability of caly A to inhibit PIP5KI␤ dephosphorylation as well as activation, and also to block PIP 2 increase in cells. Our conclusion is further supported by the finding that PP1 dephosphorylates and activates PIP5KI␤ in vitro (29).
Our results show for the first time that PIP5KI␤ depletion decreases the amount of actin filaments under isotonic conditions and blocks hypertonicity-induced actin polymerization/reorganization. These results place PIP5KI␤ activation and PIP 2 increase upstream of actin remodeling. This placement is further supported by additional lines of evidence. Urea, which increases osmolarity without causing cell shrinkage, does not promote PIP5KI␤ dephosphorylation nor increase PIP 2 . Since urea also does not promote actin assembly (38), these events are inter-related. Our results with actin poisons and Rho kinase inhibitor clearly establish that PIP5KI␤ dephosphorylation is not dependent on mechanical transduction by the actin cytoskeleton.
Rho is activated by hypertonic stress, and Rho recruits PIP5KI␤ to the plasma membrane under isotonic conditions (14,15). However, we find that PIP5KI␤ is dephosphoryled normally in the presence of the Rho kinase inhibitor, and dephosphorylated PIP5KI␤ is not preferentially recruited to membranes. We propose that the current result can be explained by hypothesizing that PIP5KI␤ is subject to regulation at multiple levels. Exposure to hypertonicity immediately triggers PIP5KI␤ dephosphorylated by a volume sensing, Rho kinase-independent protein phosphatase. In addition, Rho kinase is also activated either independently or downstream of PIP5KI␤ dephosphorylation. The latter possibility is suggested by the finding that PIP 2 activates a Rho GEF in yeast (45). Rho and Rho kinase can then stimulate PIP5KI␤ further by direct binding and targeting. Thus, PIP5KI␤ activation by dephosphorylation is an apical signal, which can be further modulated or propagated by other downstream regulators and crosstalk at multiple levels. The actin cytoskeleton may be remodeled initially by the PIP 2 generated through PIP5KI␤ activation and then further downstream by Rho activation.
It is hypothesized that a stronger cortical actin network tempers the inflammatory cascade during traumatic injury (4,46) by blocking leukocyte exocytosis (4,46,47). In addition, the reorganized cytoskeleton reinforces endothelial cell:cell and cell:matrix adhesion, to minimize the monolayer disruption that exacerbates the injury response (48). The type of actin remodeling appears to be cell-specific. For example, hypertonicity induces cortical actin assembly in CHO cells through an Arp2/3:cortactin pathway that is activated by Rac and Cdc42, but it induces stress fiber assembly in macrophages and epithelial cells by activating Rho and Rho kinase (47). We find that hypertonicity increases actin assembly as well as actin filament cross-linking into stress fiber in HeLa cells, suggesting that PIP 2 activates regulatory proteins that favor actin nucleation and assembly and inhibits those that promote filament depolymerization or severing (49). These results are consistent with our previous finding that overexpression of PIP5KI␤ induces NWASP:arp2/3 dependent actin polymerization (50) and actin stress fiber formation in CV1 cells (23).
Our study suggests that the increase in PIP 2 levels may explain how hypertonic resuscitation protects against the inappropriate inflammatory responses in burn and trauma patients. This model is supported by the finding that knock-out mice that do not express the mouse equivalent of human PIP5KI␤ exhibit enhanced passive cutaneous and systemic anaphylaxis (32). Like the PIP5KI␤ RNAi HeLa cells, the mast cells of these null animals have decreased actin filaments, and they are hyperactive in degranulation and cytokine production. We predict that these mice will be more susceptible to burn induced complications, and hypertonic resuscitation might be less effective in protecting against these injuries because of the lack of a PIP 2 response.
In conclusion, PIP5KI␤ is the major source of the PIP 2 generated during hypertonic stress, and this PIP 2 is necessary for the hypertonicity-induced reorganization of the actin cytoskeleton. This study provides mechanistic understanding of how hypertonicity induces PIP5KI␤ activation, increases PIP 2 , and reorganizes the actin cytoskeleton. PIP5KI␤ is dephosphorylated in response to volume changes upstream of cytoskeletal reorganization, and it generates an apical signal that has a central role in regulating the actin cytoskeleton and most likely a host of other hypertonic responses.