Inhibition of the Inositol Trisphosphate Receptor of Mouse Eggs and A7r5 Cells by KN-93 via a Mechanism Unrelated to Ca2+/Calmodulin-dependent Protein Kinase II Antagonism*

KN-93, a Ca2+/calmodulin-dependent protein kinase II (CaMKII) inhibitor, concentration-dependently and reversibly inhibited inositol 1,4,5-trisphosphate receptor (IP3R)-mediated [Ca2+] i signaling in mouse eggs and permeabilized A7r5 smooth muscle cells, two cell types predominantly expressing type-1 IP3R (IP3R-1). KN-92, an inactive analog, was ineffective. The inhibitory action of KN-93 on Ca2+ signaling depended neither on effects on IP3 metabolism nor on the filling grade of Ca2+stores, suggesting a direct action on the IP3R. Inhibition was independent of CaMKII, since in identical conditions other CaMKII inhibitors (KN-62, peptide 281–309, and autocamtide-related inhibitory peptide) were ineffective and since CaMKII activation was precluded in permeabilized cells. Moreover, KN-93 was most effective in the absence of Ca2+. Analysis of Ca2+ release in A7r5 cells at varying [IP3], of IP3R-1 degradation in eggs, and of [3H]IP3 binding in Sf9 microsomes all indicated that KN-93 did not affect IP3binding. Comparison of the inhibition of Ca2+ release and of [3H]IP3 binding by KN-93 and calmodulin (CaM), either separately or combined, was compatible with a specific interaction of KN-93 with a CaM-binding site on IP3R-1. This was also consistent with the much smaller effect of KN-93 in permeabilized 16HBE14o− cells that predominantly express type 3 IP3R, which lacks the high affinity CaM-binding site. These findings indicate that KN-93 inhibits IP3R-1 directly and may therefore be a useful tool in the study of IP3R functional regulation.

Ca 2ϩ acts as a ubiquitous second messenger that mediates a wide array of cellular functions, including muscle contraction, neurotransmitter release, and egg activation (reviewed in Ref. 1). The spatial and temporal dynamics of Ca 2ϩ release are complex and may range from localized, brief "puffs" of Ca 2ϩ to regenerative oscillations of the global cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] i ) that may last from several minutes to several hours. The nature of the Ca 2ϩ response elicited by a specific cell stimulus is highly regulated; thus, the specificity of the cellular response to a stimulus may be dictated by the precise dynamics of the Ca 2ϩ signal (2,3).
The complexity of Ca 2ϩ signaling mandates that the regulation of Ca 2ϩ release dynamics must be very intricate and precise. The phosphoinositide pathway is a major intracellular signaling pathway that is involved in the regulation of Ca 2ϩ homeostasis (1). The hydrolysis of phosphatidylinositol 4,5bisphosphate by a phospholipase C (PLC) 1 leads to the formation of diacylglycerol and inositol 1,4,5-trisphosphate (IP 3 ). IP 3 then induces Ca 2ϩ release from the endoplasmic reticulum by binding to and activating the IP 3 receptor (IP 3 R), which acts as a ligand-gated Ca 2ϩ channel. Three isoforms of the IP 3 R, IP 3 R-1, IP 3 R-2, and IP 3 R-3, have been characterized thus far (4). The convergence of multiple regulatory pathways at the IP 3 R is likely to directly influence the characteristic properties of various Ca 2ϩ signals (5). For example, Ca 2ϩ itself is known to regulate its own release through the IP 3 R, giving rise to a bell-shaped relationship between [Ca 2ϩ ] i and IP 3 -induced Ca 2ϩ release, whereby low [Ca 2ϩ ] i potentiates Ca 2ϩ release and high [Ca 2ϩ ] i inhibits it. Ca 2ϩ may influence IP 3 R function by directly binding to the receptor, resulting in conformational changes to the receptor, or it may act through Ca 2ϩ -binding proteins such as calmodulin (CaM) (reviewed in Refs. 6 and 7). Three CaM-binding sites have been described on the IP 3 R-1: a low affinity site near the N terminus of the IP 3 R-1 that essentially binds Ca 2ϩ -free CaM (apo-CaM) (8,9), a high affinity site in the central portion of the regulatory domain that mainly binds Ca 2ϩ /CaM (9,10), and a third site that only appears after splicing out of S2 (i.e. in peripheral tissues) and is antagonized by cAMP-dependent phosphorylation of IP 3 R-1 (11). The consequences of CaM binding to IP 3 R-1 are inhibitory; apo-CaM inhibits IP 3 binding to its receptor (8,12), whereas the binding of Ca 2ϩ /CaM inhibits Ca 2ϩ release through the channel (13,14). The relation, however, between the nature of the CaMbinding sites and the functional effects of CaM remains to be established.
Phosphorylation of the IP 3 R provides another means of functional regulation. Phosphorylation of the IP 3 R by protein kinases A, C, and G as well as by Ca 2ϩ /CaM-dependent protein kinase II (CaMKII) and Src family tyrosine kinases has been demonstrated (reviewed in Ref. 4), but the functional consequences of these phosphorylations are poorly understood. Cyclic phosphorylation and dephosphorylation of the IP 3 R, particularly by CaMKII, have been proposed as mechanisms by which complex Ca 2ϩ -signaling patterns, such as repetitive [Ca 2ϩ ] i oscillations, may be regulated (15,16). In any case, modifications of the IP 3 R by Ca 2ϩ , whether through accessory proteins such as CaM or via direct binding to the IP 3 R, contribute to the stringent control of various Ca 2ϩ release patterns (6). Thus, further characterization of the intricate regulatory mechanisms of IP 3 R function is essential to advance our understanding of Ca 2ϩ signaling.
We now report that KN-93, a pharmacological inhibitor of CaMKII (17), inhibits IP 3 -induced Ca 2ϩ release in both permeabilized A7r5 smooth muscle cells, which respond to IP 3 with a single [Ca 2ϩ ] i rise, and in mouse metaphase II (MII) stage eggs, which exhibit repetitive [Ca 2ϩ ] i oscillations in response to stimulation by IP 3 . Both of these cell types express predominantly the IP 3 R-1 subtype (18 -20), and the Ca 2ϩ release properties of A7r5 cells mainly reflect the characteristics of IP 3 R-1 (21,22). Importantly, our data suggest that the inhibition of IP 3 -induced Ca 2ϩ release by KN-93 is independent of the compound's ability to antagonize CaMKII activity in both cell systems. Furthermore, we show that the site of inhibition occurs at the IP 3 R itself. Thus, KN-93 may prove to be a useful tool in the study of IP 3 R function and regulation. In addition, our study suggests that caution must be exerted in the use of KN-93 as a CaMKII antagonist, given the propensity of the compound to block the Ca 2ϩ signal needed for eventual CaMKII activation.

EXPERIMENTAL PROCEDURES
Mouse Egg Recovery and Culture-Female CD-1 mice were superovulated by sequential injection of 5 IU of equine chorionic gonadotropin (Sigma) followed 48 h later by injection of 5 IU of human chorionic gonadotropin (Sigma). MII stage eggs were recovered from the oviducts of stimulated mice 12-15 h following human chorionic gonadotropin injection into a HEPES-buffered solution (TL-Hepes; Sigma) supplemented with 10% heat-treated fetal calf serum (Invitrogen). Eggs were freed of their cumulus cells by a 5-min wash in bovine testes hyaluronidase (Sigma), and washed eggs were cultured in 50 Porcine Sperm Factor (pSF) Preparation-Cytosolic pSF was prepared from boar semen as described (23). Briefly, the semen was washed, the sperm suspension was sonicated, and the lysate was ultracentrifuged. The supernatant was concentrated using ultrafiltration membranes (Centriprep-30; Amicon, Beverly, MA) to final concentrations of 20 -30 g/l protein. Protein extracts were then mixed with ammonium sulfate at 50% final saturation for 30 min at 4°C, the precipitates were collected by centrifugation (10,000 ϫ g, 15 min, 4°C), and the pellets were stored at Ϫ20°C. Upon use, the pellets were resuspended in injection buffer (75 mM KCl, 20 mM HEPES, pH 7.0), washed in the same buffer to remove the ammonium sulfate, and concentrated with ultrafiltration membranes. Samples were aliquoted and stored at Ϫ80°C.
Expression and Purification of CaM-A pAED4 expression vector containing human CaM (24) was transformed into BL21 Escherichia coli, grown to midlog phase, and induced with 0.3 mM isopropyl-1-thio-␤-D-galactopyranoside for 2 h at 37°C. The cells were lysed by three rapid cycles of freeze/thawing between liquid nitrogen and 37°C in a buffer containing 50 mM Tris-HCl, pH 7.4, 2 mM EDTA, 1 mM dithiothreitol (DTT), 0.8 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride, 10 M leupeptin, 1 M pepstatin, and 75 nM aprotinin. The lysate was cleared by ultracentrifugation at 140,000 ϫ g for 30 min at 4°C and NaCl and CaCl 2 were added to a final concentration of 500 and 50 mM, respectively. The lysate was subsequently heated to 70°C, immediately cooled on ice, and centrifuged at 140,000 ϫ g for 30 min at 4°C. The supernatant was applied to a phenyl-Sepharose column in the presence of Ca 2ϩ (50 mM Tris-HCl, pH 7.4, 5 mM CaCl 2 , 1 mM DTT, 0.1 M NaCl). The column was subsequently washed with low salt (50 mM Tris-HCl, pH 7.4, 0.1 mM CaCl 2 , 1 mM DTT) and high salt buffer (50 mM Tris-HCl, pH 7.4, 0.1 mM CaCl 2 , 1 mM DTT, 0.5 M NaCl). Finally, CaM was eluted with 50 mM Tris-HCl, pH 7.4, 5 mM EGTA, 1 mM DTT. The purity of CaM was verified by SDS-PAGE, and the protein concentration was determined (25).
Microinjection and [Ca 2ϩ ] i Monitoring in MII Eggs-Microinjection techniques were performed as previously described (23). Glass micropipettes were filled with fura-2 dextran (fura-2D, dextran 10 kDa; Molecular Probes, Inc., Eugene, OR), pSF, IP 3 (Molecular Probes), adenophostin A (a generous gift from Dr. K. Tanzawa, Sankyo Co., Tokyo, Japan), peptide 281-309, or AIP, and the loaded solutions were expelled into the eggs' cytoplasm by pneumatic pressure (PLI-100 picoinjector; Harvard Apparatus, Cambridge, MA). Injection volumes were ϳ5-10 pl, resulting in final intracellular concentrations of injected compounds of 1.5-3% of the concentration in the injection pipette. [Ca 2ϩ ] i monitoring of fura-2D-loaded eggs was performed exactly as previously described (23). [Ca 2ϩ ] i monitoring was initiated 30 -45 min following fura-2D injection. Differences in basal [Ca 2ϩ ] i levels from egg to egg may be partly attributable to variability in the volume of fura-2D received by each egg. 45 Ca 2ϩ Fluxes on Permeabilized Cells-Cell culture and 45 Ca 2ϩ fluxes on permeabilized A7r5 smooth-muscle cells and 16HBE14o Ϫ bronchial epithelial cells were essentially performed as previously described (21,26). In brief, cells were permeabilized by treating them for 10 min with 20 g/ml saponin at 25°C in a medium containing 120 mM KCl, 30 mM imidazole-HCl, pH 6.8, 2 mM MgCl 2 , 1 mM ATP, and 1 mM EGTA. The nonmitochondrial Ca 2ϩ stores were loaded for 45 min in loading medium containing 120 mM KCl, 30 mM imidazole-HCl, pH 6.8, 5 mM MgCl 2 , 5 mM ATP, 0.44 mM EGTA, 10 mM NaN 3 , and 150 nM free 45 Ca 2ϩ (23 Ci/ml; Amersham Biosciences). Efflux was performed in a medium containing 120 mM KCl, 30 mM imidazole-HCl, pH 6.8, and 1 mM EGTA. Activation of the IP 3 R was routinely performed with 1 M IP 3 (Roche Diagnostics, Mannheim, Germany). Any further modifications to loading or to efflux media are indicated in the figure legends. Free [Ca 2ϩ ] was calculated by the Cabuf program (available on the Internet at ftp://ftp.cc.kuleuven.ac.be/pub/droogmans/cabuf.zip) and based on the stability constants given by Fabiato and Fabiato (27). The efflux medium was collected every 2 min for a 20-min time period. At the end of the experiment, the 45 Ca 2ϩ remaining in the stores was released by incubation with 1 ml of a 2% SDS solution for 30 min.
[ 3 H]IP 3 Binding to Microsomes of IP 3 R-1-overexpressing Sf9 Cells-The heterologous expression of IP 3 R-1 in Sf9 insect cells, the preparation of microsomes, and the [ 3 H]IP 3 binding assay were previously described in detail (12,28). The binding buffer contained 50 mM Tris-HCl, pH 7.0, 10 mM EGTA, and 10 mM ␤-mercaptoethanol. [ 3 H]IP 3 (Amersham Biosciences) was used at a concentration of 5 nM, and protein concentration amounted to 50 g of protein/sample.
Western Blots-Western blots were performed as previously described (29). For each replicate, 15 control or KN-93-treated eggs were collected 2 h after injection of adenophostin A. Time-matched uninjected controls were analogously collected. Proteins were separated in 4% SDS-polyacrylamide gels, followed by transfer to nitrocellulose membranes (Micron Separation, Westboro, MA) using a Mini Trans Blot Cell (Bio-Rad). After several washes, the membranes were incubated overnight with a rabbit polyclonal antibody raised against the C-terminal end of IP 3 R-1 (Rbt04) (30). Membranes were then incubated with a secondary antibody coupled to horseradish peroxidase and were developed using Western blot chemiluminescent reagents (PerkinElmer Life Sciences). The developed membranes were exposed to maximum sensitivity film (Eastman Kodak Co., Fisher), and quantification of IP 3 R-1 bands was performed using Adobe Photoshop (Mountain View, CA). Bands from uninjected control eggs (KN-93treated or -untreated) were taken to represent 100% intensity, and bands from adenophostin A-injected eggs were compared with their respective controls.

KN-93 Inhibits pSF-induced [Ca 2ϩ ] i Oscillations-
The initial goal of this study was to evaluate the involvement of CaMKII in mouse egg activation in response to injection of pSF, a proteinaceous factor isolated from boar semen that induces [Ca 2ϩ ] i oscillations in mouse MII eggs by stimulating the phosphoinositide pathway (23,(31)(32)(33). We initially found that 25 M KN-93 completely blocked pSF-induced meiotic resumption, which is an early step in the egg activation program (see Table  I). Since [Ca 2ϩ ] i oscillations are responsible for inducing egg activation and resumption of meiosis, we also monitored the Ca 2ϩ responses in KN-93-treated eggs. We were surprised to find that 25 M KN-93 also completely suppressed pSF-induced [Ca 2ϩ ] i oscillations. As shown in Fig Inhibitory Effects of KN-93 Are Downstream of IP 3 Production-There is evidence indicating that KN-93 and other CaMKII antagonists may abrogate IP 3 production by acting at the level of PLC (34,35). Therefore, it is possible that our observed lack of oscillations in MII eggs in the presence of KN-93 reflects an inhibition of IP 3 production. Since direct measurements of IP 3 production are not yet possible in mammalian eggs, we directly assayed Ca 2ϩ release in MII eggs in response to IP 3 injection in order to determine whether KN-93 affects IP 3 production. Fig To further characterize the inhibitory mechanism of KN-93, we tested the effect of the compound on IP 3 -induced Ca 2ϩ release in permeabilized A7r5 smooth muscle cells, a somatic cell system in which the properties of the IP 3 R-1 have been studied extensively (e.g. see Refs. 13,21,26). In permeabilized A7r5 cells, the addition of 1 M IP 3 , a subsaturating concentration in this system, to the efflux medium induced a more than 3-fold increase in the fractional Ca 2ϩ loss from the nonmitochondrial internal stores, whereas the additional inclusion of KN-93 in the efflux medium reduced this IP 3 -induced Ca 2ϩ release in a concentration-dependent way (Fig. 3A). A closer examination of the concentration dependence of KN-93 inhibition is presented in Fig. 3B. Under those conditions (1 M IP 3 , no Ca 2ϩ ), KN-93 inhibited IP 3 -induced Ca 2ϩ release half-maximally at a concentration of 17 M. Increasing the IP 3 concentration (see further) or the Ca 2ϩ concentration (Fig. 3C), how- ] i oscillations Peptide inhibitors (peptide 281-309 and AIP) were introduced into eggs by microinjection; concentrations reported are approximate intracellular concentrations following microinjection. Eggs were treated with the other inhibitors for a 30-min preincubation. Egg activation was judged based on extrusion of the second polar body, an indicator of meiosis resumption. "Normal" indicates that at least 50% of eggs extruded the second polar body by 1.5 h post-pSF injection; "delayed" indicates that at least 50% of eggs required greater than 2 h post-pSF injection to exhibit second PB extrusion; "inhibited" indicates that second PB extrusion was never seen. "[Ca 2ϩ ] i oscillations" refers to eggs that exhibited Ն3 [Ca 2ϩ ] i spikes in 10 min following the initial high amplitude rise. Statistically significant differences compared with untreated controls are indicated by an asterisk ( 2 , p Ͻ 0.05).  (37)) had any effect on the IP 3 -induced Ca 2ϩ release in A7r5 cells up to concentrations as high as 50 M (Fig. 3B). Ca 2ϩ release are therefore not unique to a single cell type and may represent a general effect of this compound.
KN-93 Does Not Affect Ca 2ϩ Store Loading-Based on the previous findings, only two possible mechanisms of KN-93mediated inhibition remain; the compound may significantly reduce the Ca 2ϩ level in IP 3 -sensitive stores, or it may directly inhibit Ca 2ϩ release through the IP 3 R. We tested the first of these hypotheses in A7r5 cells by assessing the Ca 2ϩ content of the internal stores either after performing store loading in the absence or presence of the various KN-compounds at concentrations up to 50 M or by measuring the effects of those same compounds on the passive Ca 2ϩ leak. ATP-driven Ca 2ϩ uptake in the nonmitochondrial stores amounted in control conditions to 1.6 Ϯ 0.2 nmol of Ca 2ϩ /10 6 cells (n ϭ 12) taken up in 45 min. None of the compounds inhibited Ca 2ϩ uptake (Ca 2ϩ uptake in the presence of 50 M KN-93, KN-92, and KN-62 amounted to 111 Ϯ 4, 113 Ϯ 3, and 110 Ϯ 12%, respectively, of the control uptake; three independent experiments, each performed in triplicate). Furthermore, none of the compounds affected the passive Ca 2ϩ leak from the internal stores (data not shown). Thus, it can be concluded that internal stores are filled to equal capacity in KN-93-treated and untreated A7r5 cells. We obtained similar results in mouse MII eggs, since neither ampli-tude nor duration of Ca 2ϩ release induced by the Ca 2ϩ chelator ionomycin was affected by KN-93 (data not shown). Further, basal [Ca 2ϩ ] i did not increase upon exposure of eggs to 25 M KN-93, indicating that the compound does not induce Ca 2ϩ leakage and subsequent depletion from internal stores. However, the regulation of Ca 2ϩ store dynamics in eggs is more complicated due to the oscillatory pattern of Ca 2ϩ release that occurs after stimulation. Since the occurrence of persistent [Ca 2ϩ ] i oscillations depends on the timely replenishment of the Ca 2ϩ stores and is, therefore, ultimately dependent on Ca 2ϩ influx into the cell, we had to ascertain that KN-93 was not inhibiting the [Ca 2ϩ ] i oscillations primarily by blocking the so-called capacitative Ca 2ϩ entry mechanism. The mechanisms underlying capacitative Ca 2ϩ entry in eggs are very poorly understood, so direct measurements of the process are difficult. We therefore performed several indirect experiments to examine whether KN-93 could affect [Ca 2ϩ ] i oscillations indirectly (i.e. by interfering with Ca 2ϩ entry). We first reasoned that if inhibition of [Ca 2ϩ ] i oscillations in eggs by KN-93 is due solely to antagonism of capacitative Ca 2ϩ entry, then eggs held in Ca 2ϩ -free medium should also be unable to mount oscillations under our conditions. However, we found that injection of 10 M adenophostin A in eggs held in nominally Ca 2ϩ -free me- dium induced [Ca 2ϩ ] i oscillations ( Fig. 4A; n ϭ 6/7), which is consistent with previous reports (38). Furthermore, the ability of a second injection of adenophostin A to induce Ca 2ϩ release in the same KN-93-treated egg indicates that the filling grade of the Ca 2ϩ stores after the initial [Ca 2ϩ ] i spike is still sufficiently high to support further Ca 2ϩ release events ( Fig. 4B; n ϭ 8/8). Last, we found that Ca 2ϩ entry, as evaluated by fura-2 bleaching by Mn 2ϩ (39,40), was not significantly antagonized by KN-93 (data not shown). We therefore conclude that KN-93 does not significantly alter Ca 2ϩ store content and Ca 2ϩ store refilling in A7r5 cells or MII eggs and, thus, that the main mechanism of inhibition by KN-93 in those cells must involve the IP 3 R itself.
Inhibition of IP 3 -induced Ca 2ϩ Release Is Independent of CaMKII Inhibition-The absence of an inhibitory effect of KN-62 on IP 3 -induced Ca 2ϩ release in permeabilized A7r5 cells (Fig. 3B), the absence of ATP, Mg 2ϩ , and CaM in our assay conditions (Fig. 3), and the fact that the inhibitory effect of KN-93 is reversed at higher Ca 2ϩ concentrations (Fig. 3C), do not point to an effect of KN-93 on CaMKII activity in those cells. To analyze the possibility that KN-93 abrogates, at least partially, IP 3 R function in eggs by interfering with a regulatory effect of CaMKII on the IP 3 R, we assessed the effects of a panel of other commercially available CaMKII inhibitors on pSFinduced Ca 2ϩ release in eggs (Table I). As observed in A7r5 smooth muscle cells, KN-62 had no effect on pSF-induced [Ca 2ϩ ] i oscillations or on egg activation. We also tested two highly potent and specific CaMKII inhibitory peptides. Neither peptide 281-309, which inhibits CaMKII by mimicking the autoinhibitory domain (41), nor AIP, which is a nonphosphorylatable, competitive substrate for autophosphorylation of CaMKII (42), inhibited [Ca 2ϩ ] i oscillations or egg activation. Finally, inhibition of Ca 2ϩ /CaM, the upstream agonist of CaMKII, with W-7 (43) also failed to inhibit pSF-induced [Ca 2ϩ ] i oscillations. W-7 did, however, significantly delay egg activation, as evidenced by a delay in extrusion of the second polar body in eggs treated with the compound, an effect that has been previously documented (44). This delay cannot necessarily be attributed to a CaMKII effect, since Ca 2ϩ /CaM activates a wide range of target molecules. We also found that KN-93 (25 M) did not significantly reduce the basal CaMKII activity in MII eggs compared with untreated or KN-92-treated controls (data not shown). This is important, since, if KN-93 were to inhibit IP 3 -induced Ca 2ϩ release by reducing CaMKII activity, then the compound would be expected to lower the basal level of CaMKII activity prior to the induction of Ca 2ϩ release. Thus, the inability of KN-93 to reduce basal CaMKII levels in eggs, together with the failure of CaMKII inhibitors aside from KN-93 to antagonize Ca 2ϩ release in both eggs and A7r5 cells, leads to the conclusion that KN-93 inhibits IP 3induced Ca 2ϩ release through a mechanism that does not involve inhibition of CaMKII activity.
KN-93 Does Not Affect IP 3 Binding to the IP 3 R-Since the inhibition of IP 3 -induced Ca 2ϩ release by KN-93 was counteracted at increasing IP 3 concentrations, we first determined whether KN-93 acts competitively or noncompetitively with respect to IP 3 binding to the IP 3 R. A Lineweaver-Burk plot of the fractional IP 3 -induced Ca 2ϩ release measured in permeabilized A7r5 cells versus the IP 3 concentration shows that the presence of KN-93 (25 M) did not alter the K d for IP 3 compared with untreated control conditions (Fig. 5A). This provides evidence that KN-93 did not interfere with the IP 3 -binding site of the IP 3 R. Moreover, direct measurements of [ 3 H]IP 3 binding to the IP 3 R-1 in Sf9-microsomes also showed that KN-93 (50 M) does not reduce IP 3 binding (Table II). Finally, we analyzed in eggs the effects of KN-93 on IP 3 binding to the IP 3 R-1 in vivo. In mouse MII eggs as well as in several other systems that have been studied, it is known that the binding of IP 3 to the IP 3 R-1 induces degradation of the IP 3 R-1 (29,(45)(46)(47). Since this IP 3 R degradation absolutely depends on the binding of its ligand (48), analysis of IP 3 R-1 levels following agonist stimulation can be used as a measure for the in vivo binding of IP 3 to its receptor. As shown in Fig. 5B, injection of 10 M adenophostin A in untreated MII eggs caused a significant reduction in the amount of IP 3 R-1 protein detectable by Western blot compared with uninjected controls. KN-93 treatment did not significantly alter this degradation of the IP 3 R in response to adenophostin A injection. Quantification of three blots from independent experiments indicates that both untreated and KN-93-treated eggs exhibited a reduction in IP 3 R-1 levels of ϳ40% in response to adenophostin A injection compared with uninjected controls (Fig. 5C). These results verify that KN-93 does not hinder the binding of IP 3 to the IP 3 R-1 in vitro or in vivo.
KN-93 Directly Affects IP 3 R-1 Function-Given that KN-93 does not interfere with IP 3 binding, a distinct possibility is that KN-93 interferes with the function of the IP 3 R-1 Ca 2ϩ release channel, preventing the passage of Ca 2ϩ ions from the endoplasmic reticulum. Such a mechanism has been postulated for the inhibition of IP 3 R-1 function by the binding of Ca 2ϩ /CaM to the receptor (9,13,14). Since KN-93 inhibits CaMKII by competitively binding to the Ca 2ϩ /CaM-binding site of the enzyme, a tantalizing hypothesis is that KN-93 may bind to a Ca 2ϩ / CaM-binding site on the IP 3 R-1 and, like Ca 2ϩ /CaM itself, inhibit the function of the Ca 2ϩ release channel.
First, we explored the possibility that KN-93 is acting through the Ca 2ϩ /CaM-binding site of the IP 3 R-1 by analyzing the combined effects of KN-93 and of Ca 2ϩ /CaM on IP 3 -induced Ca 2ϩ release in permeabilized A7r5 cells. If KN-93 and Ca 2ϩ / CaM act through different sites, it can be expected that the effect of the two inhibitory factors together should be additive. However, whereas the inclusion of either KN-93 (25 M) or CaM (3 M) each caused a significant reduction in IP 3 -induced Ca 2ϩ release, the combination of both antagonists did not cause a larger inhibition than that caused by KN-93 alone (Fig. 6). Similar results were obtained when using other concentrations of KN-93 or of CaM (data not shown).
CaM is also a partial inhibitor of [ 3 H]IP 3 binding to IP 3 R-1 (8, 12). Although it is not yet known how the effects of CaM on IP 3 binding relate to its effect on channel function, it was important to verify whether KN-93 could interfere with the ability of CaM to inhibit IP 3 binding. The need to use neutral pH for observing the inhibitory effect of CaM on IP 3 binding (8), combined with the quite low density of IP 3 -binding sites in A7r5 cells, forced us to shift for this type of analysis to Sf9 cells heterologously expressing IP 3 R-1 (12,28). KN-93 (50 M) by itself did not affect IP 3 binding to Sf9 microsomes, but when added together with a supramaximal concentration of CaM (20 M), it completely reversed the inhibitory effect of the latter (Table II).
Finally, we investigated whether KN-93 could affect IP 3induced Ca 2ϩ release in permeabilized 16HBE14o Ϫ cells, which predominantly express IP 3 R-3 (49). The significance of these experiments lies in the fact that the IP 3 R-3 lacks the high affinity Ca 2ϩ /CaM-binding site present in the modulatory region of IP 3 R-1 (10). KN-93 was, however, markedly less effective at inhibiting IP 3 -induced Ca 2ϩ release in 16HBE14o Ϫ cells compared with A7r5 cells (only 22% inhibition at 50 M KN-93, compared with 67% inhibition in A7r5 cells under the same conditions), whereas KN-92 and KN-62 were virtually ineffective (data not shown).

DISCUSSION
Although KN-93 reportedly acts as a specific inhibitor of CaMKII activity (17), several other effects of the compound, unrelated to CaMKII inhibition, have been reported, including blockade of voltage-dependent K ϩ channels in vascular myocytes (50) and interference with Ca 2ϩ influx in pancreatic ␤-cells (51). The results of our current study provide further evidence of additional effects of KN-93 when the compound is used to probe CaMKII-related cellular functions; however, our results also give rise to the exciting possibility that KN-93 may be a valuable tool for the study of IP 3 R-related cellular functions. Through a series of experiments that probed critical regulatory mechanisms of IP 3 -mediated Ca 2ϩ signaling, we have demonstrated that KN-93 antagonizes IP 3 -induced Ca 2ϩ release in both mouse MII eggs and permeabilized A7r5 smooth muscle cells. This inhibition involves direct modulation of IP 3 R-1 function via a mechanism that is not related to a putative regulation of the IP 3 R-1 by CaMKII activity. Furthermore, we directly evaluated several functional properties of the IP 3 R-1 and found that KN-93 does not prevent IP 3 binding through competitive or allosteric inhibition and that the capac- CaMKII has been shown to phosphorylate the IP 3 R both in vitro and in vivo (52,53); however, the functional significance of this phosphorylation is not clear. Zhu et al. (16) and Matifat et al. (53) have reported that CaMKII activity is inhibitory with respect to IP 3 -induced Ca 2ϩ release in HeLa cells and Xenopus oocytes, respectively, and may represent one of the mechanisms by which Ca 2ϩ release is terminated at elevated [Ca 2ϩ ] i . In contrast, the results obtained by Zhang et al. (15) in permeabilized 3T6 fibroblasts imply that CaMKII activity may increase the sensitivity of the IP 3 R to IP 3 , thus potentiating Ca 2ϩ release during the initiation of a [Ca 2ϩ ] i rise, although the kinase in question could not be directly identified as CaMKII. A stimulatory function for CaMKII could explain the severe abrogation of Ca 2ϩ release observed in MII eggs and A7r5 cells in response to KN-93 treatment; however, this is unlikely, due to the inability of other CaMKII inhibitors to mimic the results obtained with KN-93. We can further discount a role of CaMKII in the results obtained in A7r5 cells, since the plasma membrane permeabilization of these cells should remove most, if not all, kinases and other cytosolic proteins, including CaM, from the system. Furthermore, since Ca 2ϩ efflux was measured in the absence of added ATP, Mg 2ϩ , Ca 2ϩ , or CaM, kinases in general and CaMKII in particular are not likely to have been activated. Finally, KN-93 is maximally effective at inhibiting IP 3 -induced Ca 2ϩ release in the absence of Ca 2ϩ , whereas the involvement of CaMKII would imply that the inhibition should increase at higher [Ca 2ϩ ] i . For these reasons, we can conclude that in our experimental conditions, KN-93 inhibition of IP 3induced Ca 2ϩ release is not dependent on CaMKII.
The ability of KN-93 to abrogate Ca 2ϩ release evoked by the direct addition of IP 3 or adenophostin A to MII eggs and A7r5 cells indicates that the compound exerts a significant effect on IP 3 R-1 function. However, in the case of MII eggs injected with pSF, an additional effect of KN-93 on IP 3 production cannot be ruled out. It has recently been shown that KN-93 inhibited, by ϳ2-fold, phosphoinositide turnover stimulated by G␣ q -activated PLC␤ 1 or PLC␤ 3 in somatic cells, although PLC activity still remained elevated above basal levels (35). This study concluded that this inhibitory effect was not due to the demonstrated ability of CaMKII to phosphorylate PLC and may be due to an unknown side effect of KN-93. Due to the small size of mammalian eggs and an inability to simultaneously stimulate a large number of eggs, measurements of phosphoinositide turnover cannot currently be performed in mouse MII eggs. However, since it is known that Ca 2ϩ can stimulate PLC activity (54), it is possible that KN-93 may have an indirect, negative effect on PLC due to its ability to abrogate IP 3 -induced Ca 2ϩ release.
The activity of the IP 3 R can be modulated either by compounds acting on the IP 3 -binding site or else by compounds acting on the opening of the Ca 2ϩ release channel. Analysis of the kinetics of KN-93 inhibition indicate a noncompetitive mechanism, since KN-93 does not alter the K d for IP 3 . Direct measurement of [ 3 H]IP 3 binding to the IP 3 R-1 and analysis of IP 3 R-1 down-regulation in eggs indicated that KN-93 did not abrogate IP 3 binding to the IP 3 R-1. Therefore, it is likely that the inhibition imposed by KN-93 occurs downstream of IP 3 binding. Since CaM is an important modulator of both IP 3 binding (8,12) and IP 3 R function (9,13,14) and since KN-93 is known to interact with the CaM-binding site of at least CaMKII (17), we hypothesized that KN-93 might inhibit the IP 3 R by interacting with one or more of its CaM-binding sites. Up to now, three different CaM-binding sites have been described for the IP 3 R-1 (9 -11), but the functional significance of each is still unclear. The Ca 2ϩ -independent CaM-binding site present in the N terminus of the receptor (9) might be responsible for the inhibitory effects on IP 3 binding (8, 12, 55), but it is unclear whether the high affinity Ca 2ϩ /CaM-binding site located in the modulatory region is involved in the regulation of the channel function of IP 3 R-1 (56, 57) as originally thought. Moreover, IP 3 -induced Ca 2ϩ release in cells expressing predominantly IP 3 R-3 is also inhibited by CaM (9,58), although a characteristic structural feature of IP 3 R-3 is the absence of the high affinity Ca 2ϩ /CaM-binding site conserved in IP 3 R-1 and IP 3 R-2 (10). Significantly, we found that KN-93 is far less effective at inhibiting IP 3 -induced Ca 2ϩ release in permeabilized 16HBE14o Ϫ cells, which predominantly, but not exclusively, express the IP 3 R-3 (49). Independently of the physiological site of action of Ca 2ϩ /CaM, a mechanism whereby KN-93 inhibits IP 3 R-1 function by occupying the high affinity Ca 2ϩ / CaM binding site is therefore consistent with the lack of inhibition in 16HBE14o Ϫ cells and with our finding that inhibition of IP 3 R-1 occurs downstream of IP 3 binding. Moreover, we evaluated the effect of exogenous Ca 2ϩ /CaM on the ability of KN-93 to inhibit IP 3 -induced Ca 2ϩ release to determine whether the two antagonists act synergistically. Since no addi-  tive effect of both inhibitors was observed, our results are at least consistent with the notion that both antagonists act through the same binding site. Alternatively, the possibility exists that the interaction of either CaM or KN-93 induces a large conformational change of the receptor, precluding the binding of the other. Additionally, the fact that KN-93 can abrogate the inhibitory effect of CaM on IP 3 binding suggests that it can interact in a direct or indirect way with the CaMbinding site located in the N-terminal region of the receptor. The relation between this CaM-binding site, which is essentially Ca 2ϩ -independent, and the inhibitory effect on IP 3 R function, which is essentially Ca 2ϩ -dependent, is however not yet known.
In conclusion, our results indicate that the widely used CaMKII inhibitor KN-93 potently inhibits IP 3 -mediated Ca 2ϩ signals by interacting directly with the IP 3 R. This effect may be mediated by its interaction with one or several of the CaMbinding sites on the receptor. This is relevant, since, to date, there is no known highly specific and reversible inhibitor of the IP 3 R available. KN-93 can therefore be a helpful tool in the structure-function analysis of the IP 3 R. Moreover, it may be possible that related compounds could be developed with higher specificity for the regulatory CaM-binding site on IP 3 R-1.