Akt Kinase Phosphorylation of Inositol 1,4,5-Trisphosphate Receptors*

A consensus RXRXX(S/T) substrate motif for Akt kinase is conserved in the C-terminal tail of all three inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) isoforms. We have shown that IP3R can be phosphorylated by Akt kinase in vitro and in vivo. Endogenous IP3Rs in Chinese hamster ovary T-cells were phosphorylated in response to Akt activation by insulin. LnCAP cells, a prostate cancer cell line with constitutively active Akt kinase, also showed a constitutive phosphorylation of endogenous type I IP3Rs. In all cases, the IP3R phosphorylation was diminished by the addition of LY294002, an inhibitor of phosphatidylinositol 3-kinase. Mutation of IP3R serine 2681 in the Akt substrate motif to alanine (S2681A) or glutamate (S2681E) prevented IP3R phosphorylation in COS cells transfected with constitutively active Akt kinase. Analysis of the Ca2+ flux properties of these IP3R mutants expressed in COS cell microsomes or in DT40 triple knock-out (TKO) cells did not reveal any modification of channel function. However, staurosporine-induced caspase-3 activation in DT40 TKO cells stably expressing the S2681A mutant was markedly enhanced when compared with wild-type or S2681E IP3Rs. We conclude that IP3 receptors are in vivo substrates for Akt kinase and that phosphorylation of the IP3R may provide one mechanism to restrain the apoptotic effects of calcium.

A consensus RXRXX(S/T) substrate motif for Akt kinase is conserved in the C-terminal tail of all three inositol 1,4,5-trisphosphate (IP 3 ) receptor (IP 3 R) isoforms. We have shown that IP 3 R can be phosphorylated by Akt kinase in vitro and in vivo. Endogenous IP 3 Rs in Chinese hamster ovary T-cells were phosphorylated in response to Akt activation by insulin. LnCAP cells, a prostate cancer cell line with constitutively active Akt kinase, also showed a constitutive phosphorylation of endogenous type I IP 3 Rs. In all cases, the IP 3 R phosphorylation was diminished by the addition of LY294002, an inhibitor of phosphatidylinositol 3-kinase. Mutation of IP 3 R serine 2681 in the Akt substrate motif to alanine (S2681A) or glutamate (S2681E) prevented IP 3

R phosphorylation in COS cells transfected with constitutively active Akt kinase. Analysis of the Ca 2؉ flux properties of these IP 3 R mutants expressed in COS cell microsomes or in DT40 triple knock-out (TKO) cells did not reveal any modification of channel function. However, staurosporine-induced caspase-3 activation in DT40 TKO cells stably expressing the S2681A mutant was markedly enhanced when compared with wild-type or S2681E
IP 3 Rs. We conclude that IP 3 receptors are in vivo substrates for Akt kinase and that phosphorylation of the IP 3 R may provide one mechanism to restrain the apoptotic effects of calcium.
Inositol 1,4,5-trisphosphate receptors (IP 3 R) 2 are a family of intracellular release channels that play an essential role in evoking Ca 2ϩ signals triggered by the occupation of numerous types of cell-surface receptors that are coupled to enhanced inositol-lipid turnover (1). Three different IP 3 R isoforms have been identified, and most cells appear to express multiple isoforms (2,3). IP 3 receptors have been shown to be substrates for several different protein kinases in vivo and/or in vitro. These include protein kinase A (4,5), protein kinase G (6, 7), protein kinase C (8), CaM kinase-II (8), Fyn tyrosine kinase (9,10) and cdk1/CyB (11). By far the best-characterized of these effects are those of protein kinases A and G. Two serine residues have been identified as potential sites of phosphorylation in the type I IP 3 R isoform (12). Mutagenesis studies in a DT40 IP 3 R expression system suggest that both protein kinases A and G act to enhance IP 3 -mediated Ca 2ϩ release (13). However, the preferred sites of phosphorylation and functional effects on IP 3 R channel function may differ in different cell types and between alternatively spliced variants (6,(13)(14)(15). The regulatory role of the other protein kinases phosphorylating IP 3 Rs are poorly understood.
The serine/threonine protein kinase Akt/protein kinase B is the cellular homologue of the viral oncogene v-Akt and is activated by various growth factors and cytokines. The membrane translocation and initial activation of the enzyme is dependent on the production of 3-phosphorylated inositol lipids catalyzed by PI 3-kinase. The activated Akt kinase then phosphorylates a number of key substrates involved in the stimulation of intermediary metabolism and promotion of cell survival, proliferation, and growth (reviewed in Refs. 16 -18). An examination of the sequence of all three IP 3 R isoforms indicates the presence of a consensus RXRXX(S/T) sequence for phosphorylation by Akt kinase (19). In the present study, we have shown for the first time that IP 3 Rs are substrates for activated Akt kinase in vivo and have investigated several possible functional consequences of this phosphorylation on Ca 2ϩ signaling.

EXPERIMENTAL PROCEDURES
Reagents-Insulin, Triton X-100 and okadaic acid were obtained from Sigma. Pfu polymerase was from Stratagene (La Jolla, CA). Trans-IT-LT-1 cationic lipid transfection reagent was obtained from Pan Vera Corp. (Madison, WI). Protogel-stabilized acrylamide solution was from National Diagnostic (Atlanta, GA). Horseradish peroxidase-conjugated goat anti-rabbit antibody was purchased from Amersham Biosciences. Enhanced chemiluminescent substrate was obtained from Pierce (Rockford, IL). 45 Ca 2ϩ was from PerkinElmer Life Sciences. Mouse antichicken IgM (clone-M4) was from Southern Biotech (Birmingham, AL).
Cells and Culture Conditions-The Chinese hamster ovary T-cell line (CHO-T), which stably expresses the human insulin receptor, was a kind gift from Dr. Bryan Wolf (Childrens Hospital University of Pennsylvania, Philadelphia, PA). This cell line was cultured in F-12 (Ham's) medium (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin (complete F-12), and incubated at 37°C in a 5% CO 2 , 95% air-humidified incubator. The prostrate cancer cell line LnCap was obtained from the American Type Culture Collection (ATCC) (Manassas, VA) and was passaged in RPMI 1640 medium with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. The DT40 cell line containing targeted deletions of IP 3 R isoforms was a kind gift of Dr. T. Kurosaki (Kansai Medical University, Moriguchi, Japan) and was cultured as described previously (13). COS-7 cells were purchased from the ATCC and were cultured as described previously (20).
DNA Constructs and Transfection-The neuronal SII(ϩ)/SI(Ϫ) type I IP 3 R construct, encoding a Kozac sequence and subcloned into pcDNA 3.1, has been described previously (20). Mutants S2681A and S2681E were made using the QuikChange site-directed mutagenesis kit (Stratagene). The forward primers were designed following the manufacturer's recommendations as follows: 5Ј-AGAGCCATGGCCCTG-GTC-3Ј for the S2681A mutant and ATGAGAGCCATGGAACTG-GTCAGCAGT-3Ј for the S2681E mutant. (codon change is shown in bold). The reverse primer was the complementary sequence of the forward primer. PCR cycling conditions were performed according to the manufacturer's instructions, and the products were confirmed by automated sequencing (Applied Biosystems Model 377; Nucleic Acid Facility, Thomas Jefferson University, Philadelphia, PA).
The cDNA encoding the human isoform of SERCA-2b in pcDNA 3.1 was a gift from Dr. J. Lytton (University of Calgary, Alberta, Canada) and Dr. David H. MacLennan (University of Toronto, Ontario, Canada). The construct encoding the membrane-targeted, constitutively active, human Akt kinase harboring a myristoylation sequence was kindly given by Dr. J. Q. Cheng (University of South Florida School of Medicine). A soluble version of a constitutively active bovine Akt kinase mutant (S473D/T308D) was kindly given by Drs. Stephanie Dimmeler and Judith Haendeler (University of Frankfurt, Frankfurt, Germany). The C-terminal 191 amino acids of the type I IP 3 R were expressed as a GST fusion protein after PCR amplification and cloning into the pGEX-2K vector (Amersham Biosciences). Expression and purification of the fusion protein was carried out as described previously (21).
Antibodies-The Ab against the C terminus of the type I IP 3 R (CT-1 Ab) has been described previously (22). The antibody was further affinity-purified using the peptide coupled to Ultralink beads as described by the manufacturer (Pierce). The type III IP 3 R monoclonal antibody was purchased from Signal Transduction Labs (Lexington, KY). Antibodies that recognize the phospho-(Ser/Thr)-Akt substrates (Akt-S Ab) and the phospho-Akt (Ser-473)-activated form of Akt kinase (p-Akt Ab) were obtained from Cell Signaling Technology (Beverly, MA). Recombinant Akt kinase was purchased from Biomol (King of Prussia, PA.).
Cell Lysates and Immunoprecipitation-Cells were serum-deprived for 24 -48 h and then treated with LY29004 and/or insulin. At the end of the treatment period, the medium was aspirated and the plates were washed twice in ice-cold phosphate-buffered saline. The cells were scraped into a solubilization buffer (SB) containing 150 mM NaCl, 50 mM Tris/HCl, pH 7.8, 1% Triton X-100 (w/v), 2 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 20 mM sodium fluoride, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and a 0.5ϫ dilution of a complete protease inhibitor mixture (Roche Applied Science). Lysates were precleared with 10 l of Staphylococcus aureus cell wall (Pansorbin, Calbiochem, San Diego, CA) and then immunoprecipitated overnight with IP 3 R or the indicated Abs. The Abs were recovered using protein A-Sepharose beads that were washed three times in SB buffer and processed on SDS-PAGE for immunoblotting.
In Vitro Phosphorylation-IP 3 R protein was immunoprecipitated from cerebellum microsomal membranes after solubilization in SB buffer using procedures described previously (22). The immunoprecipitated IP 3 R bound to protein A-Sepharose beads was resuspended in 50 l of phosphorylation buffer containing 50 mM Tris/HCl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol, 50 M ATP, and 5 nM okadaic acid (Buffer P). The samples were incubated with recombinant Akt kinase protein (0.22 g) and [␥-P 32 ]ATP (20 Ci) for 2 h at 30°C. The beads were washed five times with buffer P and then quenched in SDS-PAGE sample buffer (50 mM Tris/HCl, pH 6.8, 1% SDS, 0.01% bromphenol blue, 140 mM 2-mercaptoethanol and 5% (v/v) glycerol). The samples were run on a 5% gel, and proteins were transferred to a nitrocellulose filter. The filter was autoradiographed before immunoblotting sequentially with Akt-S Ab and then with CT-1 Ab. GST C-terminal tail protein was bound to glutathione-agarose overnight and was processed for phosphorylation as described above.
Preparation of Microsomal Vesicles and 45 Ca 2ϩ Flux Assay-Transient transfection of COS-7 cells was carried out using LT-1 transfec-tion reagent as described by the manufacturer (Mirus Corp., Madison, WI). Cell lysates or microsomal membranes were prepared 48 h posttransfection. The measurement of the channel activity of wild-type or mutant IP 3 Rs required the co-transfection of SERCA2b and was performed essentially as described previously (20) with the following modifications. Cells were scraped into a sucrose-based buffer containing 0.5 M sucrose, 20 mM Tris/HCl, pH 7.8, 1 mM dithiothreitol, and 50 M EGTA and lysed by passing through a 26.5 gauge needle four times, rather than using homogenization. The lysates were spun at 5,000 ϫ g to remove debris, and the supernatants were spun at 100,000 ϫ g. The vesicle pellets were resuspended in buffer containing 0.5 M sucrose, 20 mM Tris/HCl, pH 7.8, 1 mM dithiothreitol. The vesicles were used fresh or were stored at Ϫ80°C in 20% (v/v) glycerol for further use. 45 Ca 2ϩ flux measurements were carried out exactly as described previously (20).

Measurement of [Ca 2ϩ ] i in DT40
Cells-Wild-type or mutant IP 3 R constructs were introduced into TKO DT40 cells by electroporation. Co-transfection with the red fluorescent protein pHcRed1-N1 (Clontech, Palo Alto, CA) was used as a marker to locate the transfected cells. Methods for Fura-2 loading and digital imaging of the [Ca 2ϩ ] i transients have been described previously (13).
Measurement of Caspase Activation in DT40 Cells-Stable cell lines expressing the S2681A and S2681E mutants were prepared by electroporation of 0.5 ml of DT40 TKO cells (10 6 cells/ml) with 40 g of linearized DNA using a Gene Pulser apparatus (Bio-Rad; 340 V and 950 microfarads). The cells were grown in 30 ml of culture medium for 24 h before being aliquoted and subjected to selection at 1.5 mg/ml geneticin. For caspase-3 assays, the staurosporine-treated cells were centrifuged (2000 ϫ g for 2 min), washed once in phosphate-buffered saline and lysed in a medium containing 20 mM Tris/Hepes, pH 7.4, 0.1% CHAPS, 5 mM dithiothreitol, and 2 mM EDTA. Lysate (10 g) was assayed for caspase-3 activity in a final volume of 200 l using 25 M DEVD-7amino-4-methylcoumarin as substrate after 30 min of incubation at 37°C.

IP 3 R Phosphorylation by Akt Kinase in Vitro-The identification of
Akt kinase substrates has been facilitated by the presence of a robust phosphorylation motif RXRXX(S/T) (19). A single motif of this kind is conserved within the C-terminal tail of all three IP 3 R isoforms and is also conserved in IP 3 Rs cloned from several different species, with the exception of Caenorhabditis elegans (Fig. 1A). To observe the phosphorylation directly, we incubated a GST fusion protein encoding the C-terminal tail of type I IP 3 R with catalytically active Akt kinase and [ 32 P]ATP. The fusion protein was then isolated with glutathione-agarose and analyzed by autoradiography after SDS-PAGE (Fig. 1B). The data show that the C-terminal tail fusion protein was phosphorylated by Akt kinase. Phosphorylation could also be detected by immunoblotting with a phosphospecific Ab that detects the consensus Akt phosphorylation sequence (see Akt-S blot in Fig. 1B). Similar experiments were carried out with the full-length type I IP 3 R immunoprecipitated from cerebellum lysates (Fig. 1C). In this case, a low level of [ 32 P] incorporation was observed even without the addition of Akt kinase, possibly reflecting the co-immunoprecipitation of endogenous kinases. The addition of Akt kinase greatly enhanced 32 P labeling of the IP 3 R. Specific phosphorylation at the Akt site was evident when the phosphorylation was detected by immunoblotting with the Akt-S Ab (Fig. 1C). IP 3 R Phosphorylation by Akt Kinase in Vivo-Occupation of insulin receptors leads to PI 3-kinase activation and the stimulation of Akt kinase activity (16). To examine IP 3 R phosphorylation in an intact cell, we used CHO-T cells, which overexpress human insulin receptors (23). Cell lysates were prepared in a buffer that contained a mixture of NaF, vanadate, pyrophosphate and okadaic acid to preserve the phosphorylation state of assayed proteins. Fig. 2A shows that the addition of 100 nM insulin promotes IP 3 R phosphorylation as detected by immunoblotting with Akt-S Ab in lysates immunoprecipitated with affinity-purified IP 3 R-1-specific Ab. In multiple experiments, the phosphorylation was evident within 5-10 min and peaked at 30 -60 min. The maximal increase was 4.6 Ϯ 0.9-fold in four separate experiments. The profile of IP 3 R changes approximately correlated with the time course of Akt kinase activation as estimated from immunoblotting the lysates with a phosphospecific Ab recognizing phosphoserine 473 of Akt kinase. Insulin-mediated IP 3 R phosphorylation could also be demonstrated when the lysates were immunoprecipitated with Akt-S Ab and then immunoblotted with IP 3 R-1-specific Ab (Fig. 2B). Because insulin activation of Akt kinase is secondary to PI 3-kinase stimulation, it would be anticipated that the effect of insulin on the IP 3 R would be sensitive to inhibition by the PI 3-kinase inhibitor LY294002. Fig. 3 shows that this is indeed the case. A small amount of IP 3 R phosphorylation can be observed in the absence of stimulation and could be due to the presence of a basal activity of Akt kinase. The data in Fig. 3 also show that the endogenous type III IP 3 R in CHO-T cells was also phosphorylated on a Akt consensus site in response to insulin stimulation.
Loss of the tumor promoter PTEN, a negative regulator of the PI 3-kinase pathway, leads to constitutive activation of Akt kinase in several types of cancer (24). We have used LnCaP cells, a prostate carcinoma cell line devoid of functional PTEN (25), to examine the phosphorylation state of IP 3 Rs. Fig. 3B shows that the constitutive activity of the Akt kinase in these cells is accompanied by the constitutive phosphorylation of the type I IP 3 R as detected by Akt-S Ab. This basal phosphorylation was abolished by pretreatment with LY294002 (Fig. 3B, lane 2).
Effect of Akt Kinase Phosphorylation on IP 3 R Channel Function-COS-7 cells have very low levels of endogenous type I IP 3 R (20). In Fig.   4A, we have transfected the type I IP 3 R into COS-7 cells either alone or together with a constitutively active Akt kinase (myr-Akt) (26). Lysates were immunoprecipitated with type I IP 3 R Ab and blotted with Akt-S Ab. The only condition under which substantial phosphorylation was observed was when both type I IP 3 R and myr-Akt were transfected together (Fig. 4A, upper panel, lane 3). Immunoblots with IP 3 R Ab and Ser-473 phospho-Akt Ab indicated the presence of IP 3 R and active Akt in the expected lanes.

FIGURE 1. Phosphorylation of the IP 3 R in vitro.
A, the ClustalW alignment of the C-terminal RXRXX(S/T) sequence of the different IP 3 R isoforms and IP 3 Rs from several species is shown. The Gen-Bank TM accession numbers were as follows: rat type I (P29994), rat type II (P29995), rat type III (Q63269), Xenopus (BAA03304), Drosophila (P29993), and C. elegans (AJ243719). B, 20 g of the GST C-terminal tail of IP 3 R-1 (amino acids 2591-2749) was incubated for 2 h with 20 Ci [ 32 P]ATP and 0.22 g of Akt kinase (Biomol). Detailed incubation conditions were as described "Experimental Procedures." The GST-tail was recovered by the addition of glutathione-agarose, and the sample was run on a 5% gel and autoradiographed (lanes 1 and 2) or immunoblotted with Akt-S Ab (lanes 3 and 4) or Ab to the C terminus of type I IP 3 R (CT1-Ab; lanes 5 and 6). GST alone was not phosphorylated by Akt (not shown). C, the same procedure as described for B, except that the phosphorylation was carried out with type I IP 3 R immunoprecipitated from 300 g of cerebellum lysate.
To study the functional consequences of Akt kinase phosphorylation, we produced two mutant type I IP 3 Rs in which the Akt kinase serine substrate at position 2681 was mutated to either alanine or glutamate. The latter mutation was designed to mimic the effect of Akt kinase phosphorylation. The ability of the wild-type and mutant IP 3 Rs to be phosphorylated in COS-7 cells transfected with constitutively active myr-Akt is shown in Fig. 4B. Cell lysates were immunoprecipitated with type I IP 3 R Ab and probed for phosphorylation at the Akt site with Akt-S Ab. As expected, there was a marked suppression of the phosphorylation signal when the Ser-2681 site was mutated to either alanine or glutamate. The small residual signal seen with the S2681E mutant may be due to weak recognition of the glutamate residue by the phosphospecific Akt-S Ab (27).
We have utilized a microsomal 45 Ca 2ϩ flux assay (20) to investigate the functional effects of Akt phosphorylation on IP 3 R channel function. To deal with a homogeneous population of modified IP 3 Rs, we have employed the S2681E mutant as a phosphomimic and compared its properties to the wild-type and S2681A mutant. These data are shown in Fig. 5. The IP 3 flux response with a saturating dose of IP 3 (10 M) was not different for the three constructs and was equally sensitive to inhibition when the extravesicular Ca 2ϩ was elevated from 0.2 to 4.5 M (Fig. 5A). The dose response for added IP 3 was also not different between the three constructs (Fig. 5B). We have also evaluated the function of the three constructs after transient transfection into DT-40 cells containing targeted deletions of all three IP 3 R isoforms (28,29). The cells were co-transfected with the M3 muscarinic receptor, and Ca 2ϩ signals were recorded in response to increasing doses of carbachol (Fig.  5C). This protocol has previously been used to measure the IP 3 sensitivity of various phosphomimic mutants of the protein kinase A phosphorylation sites (30). The results indicate that there are no significant differences in the EC 50 for carbachol-induced Ca 2ϩ release between the three constructs (mean Ϯ S.E. of the indicated number of cells was 32.1 Ϯ 4.2 (29), 31.4 Ϯ 8 (9), and 34.1 Ϯ 6.3 nM (16) for wild-type, S2681A, and S2681E mutants, respectively). The overall conclusion from these experiments is that phosphorylation of IP 3 Rs at the Akt site does not directly modify its channel function.
Akt Kinase Phosphorylation of IP 3 Rs and Caspase-3 Activation-In a recent study, Assefa et al. (31) have shown that IP 3 Rs are necessary to obtain maximal caspase-3 activation in DT40 cells treated with staurosporine as an apoptotic stimulus. To determine whether Akt kinase phosphorylation of IP 3 Rs influences caspase-3 activation, we made stable DT40 cell lines expressing the S2681A and S2681E mutant type I IP 3 Rs. The expression of the mutant IP 3 Rs measured by immunoblotting is shown in Fig. 6A. The levels of type I IP 3 R in the S2681A mutant were significantly lower than in any of the other cell lines used. The reason for this is unknown but was consistently observed in several different clones. The growth characteristics of the IP 3 R mutant cell lines were not significantly different from wild-type, TKO, or a double knock-  1 and 2) or type I IP 3 R (lanes 3 and 4). The activity status of Akt was assessed separately as in A (lanes 5 and 6).  . Calcium responses of wild-type, S2681A and S2681E IP 3 R mutants. A, microsomes from COS cells transfected with the indicated IP 3 R constructs and co-transfected with SERCA2b cDNA were prepared and used for a 45 Ca 2ϩ flux assay as described under "Experimental Procedures." The amount of 45 Ca 2ϩ accumulated in oxalate-loaded microsomes over a 15-min period in the presence of 10 M IP 3 was measured and is expressed as a percentage of the control accumulation in the absence of IP 3 . The experiment was done in EGTA/HEDTA Ca 2ϩ buffers at free [Ca 2ϩ ] of 0.2 and 4.5 M calibrated with a Ca 2ϩ -sensitive microelectrode. The data shown is the mean Ϯ S.E. of five independent experiments. B, the IP 3 dose-response relationship was measured at a free [Ca 2ϩ ] of 0.2 M. The data are the mean Ϯ S.E. of three independent experiments. WT, wild-type. C, single cell Fura-2 imaging was carried out on DT-40 cells transfected with the indicated IP 3 R constructs and the M3 muscarinic receptor. The procedures for sequential stimulation by increasing doses of carbachol and analysis of the data were carried out as described previously (30). out DT40 cell line containing only the type I IP 3 R (data not shown). The ability of the Akt mutants to support caspase-3 activation is shown in Fig. 6B. In agreement with the findings of Assefa et al. (31), the TKO cell line showed the lowest staurosporine-mediated activation of caspase-3. Caspase-3 activation in the phosphomimic S2681E mutant cell line was comparable with that in the double knock-out cell line containing the wild-type type I IP 3 R. A somewhat higher level of caspase-3 activation was observed in the control DT40 cell line retaining all three IP 3 R isoforms. However, the largest caspase-3 activation was observed in the cell line containing the non-phosphorylatable S2681A mutant. This activation was almost 3.5-fold greater than that observed for the S2681E cell line. The results do not correlate with the absolute amounts of IP 3 R, because the highest and lowest caspase-3 activation are observed in cell lines containing the least IP 3 R. We also tested staurosporine-induced caspase-3 activation in COS-7 cells transiently transfected with wildtype, S2681A, and S2681E type I IP 3 Rs (Fig. 6C). As observed in DT40 cells, the maximal activation of caspase-3 was observed in cells transfected with the S2681A mutant. In this series of experiments, the expression of the S2681A and S2681E mutants as measured by immunoblotting was 69.1 Ϯ 2.0% and 68.5 Ϯ 8.7% of the wild-type IP 3 R, respectively (Fig. 4B). Thus, the differences in caspase-3 activation in COS-7 cells are unlikely to be related to differences in IP 3 R expression.

DISCUSSION
Akt kinase has established roles in suppressing apoptosis, promoting growth and proliferation, and mediating many of the effects of insulin (reviewed in Refs. 16 -18). The main conclusion of the present study is that IP 3 Rs can be added to the list of potential targets of this key enzyme. We have demonstrated that this phosphorylation occurs in vitro and in vivo at a consensus Akt kinase site located in the cytosol-exposed C-terminal tail of the receptor. The generation of 3-phosphorylated phosphoinositides by PI 3-kinase is thought to initiate the activation of Akt as a result of recruitment of the enzyme to the plasma membrane. This translocation utilizes the pleckstrin homology domain in the N-terminal region of Akt. Subsequent phosphorylation of serine 473 and threonine 308 leads to complete activation of the enzyme, which can then migrate to the cytosol, nucleus, mitochondria, or Golgi apparatus to phosphorylate its various substrates. Although the exact amount of IP 3 R phosphorylated in vivo has not been quantitated in the present study, we can roughly estimate this from immunoprecipitation of the phosphorylated IP 3 R with Akt-S Ab. The results indicate that only a small fraction (Ͻ5%) of the total IP 3 R pool is phosphorylated in COS-7 cells transfected with IP 3 R and myr-Akt (data not shown). We were also unable to detect co-immunoprecipitation of endogenous or overexpressed IP 3 Rs with myr-Akt (data not shown). The bulk of IP 3 Rs are in the ER membrane, although smaller pools have been reported in the plasma membrane, nucleus, and the Golgi apparatus (32). There have been very few studies of Akt kinase localization in the ER (33). As far as we are aware, protein tyrosine phosphatase 1B is the sole ER substrate of Akt kinase reported in the literature (34). The finding that growth factors can increase ER PtdInsP 3 levels (35) raises the possibility that Akt kinase may be activated directly on ER membranes. Thus phosphorylation of the IP 3 R could be confined to a specific subcellular pool that contains activated Akt kinase (e.g. plasma membrane, nuclei, or Golgi apparatus) or could reflect a small pool of activated Akt kinase in ER membranes. Interestingly, a constitutively active Akt kinase construct (S437D/T308D), which is active without membrane targeting (36), does not phosphorylate the IP 3 R (data not shown). Further work is required to clarify the exact relationship between Akt kinase and IP 3 R localization.
A second key conclusion from the present study is that phosphorylation of the IP 3 R does not modify IP 3 -mediated Ca 2ϩ channel function. This conclusion was based on experiments utilizing IP 3 Rs that were mutated at the 2681 position to either alanine or glutamate. Neither of the mutants showed significantly different behavior from wild-type receptors in either COS-7 cells or after transient transfection into IP 3 R TKO DT40 cells. This conclusion is in agreement with the finding that the dominant negative form of Akt kinase or LY294002 does not affect bradykinin-mediated intracellular Ca 2ϩ release in porcine endothelial cells (37). LY294002, which suppressed the Akt kinase activation mediated by carbachol in CHO cells or IgM in DT40 cells, also had no effect on the ability of these agonists to mobilize intracellular Ca 2ϩ as measured with Fura-2 (data not shown).
What then is the functional role of IP 3 R phosphorylation by Akt kinase? Studies in DT40 and Jurkat cells have established that genetic ablation of IP 3 R isoforms induces resistance to apoptotic stimuli (9,29). In bladder cancer cell lines, the small interfering RNA-induced suppression of type I IP 3 R prevented apoptosis mediated by cisplatin (39). It has been generally assumed that these findings reflect a requirement for intracellular Ca 2ϩ release for one or more steps in the apoptotic cascade. . Activation of caspase-3 in stable DT40 cell lines expressing the S2681A and S2681E mutants. A, lysates from stable cell lines expressing the S2681A and S2681E mutants were compared with the type I IP 3 R levels expressed in DT40 double knock-out (DKO) cells that express only the type I IP 3 R. 100 g of lysate protein was loaded on the gels and immunoblotting was done with type I IP 3 R Ab. B, the indicated cell lines were treated with 1 M staurosporine (STS) for 3 and 6 h, and then aliquots of cells were centrifuged and lysates were assayed for caspase-3 activity as described under "Experimental Procedures." C, COS-7 cells were transfected at ϳ70% confluence with wild-type, S2681A, and S2681E IP 3 R constructs. At 48 h post-transfection, the cells were treated with staurosporine and assayed for caspase-3 as described for DT40 cells. The data are the mean Ϯ S.E. of four experiments. WT, wild-type; S-A, S2681A; S-E, S2681E.
A component of this cascade that shows such a Ca 2ϩ requirement is the activation of caspase-3. Thus caspase-9 and -3 are not activated by various stimuli in Jurkat cells deficient in type I IP 3 R (40). Assefa et al. (31). have shown that restoration of type I IP 3 R into DT40 TKO cells restores a normal caspase-3 activation and apoptotic response to stimulation by staurosporine or IgM. Interestingly, they also showed that the same effect could be observed with an IP 3 R mutant that was functionally inactive as an IP 3 -sensitive channel as a result of the deletion of the N-terminal 225 amino acids (31). Type-I IP 3 Rs have been shown to be caspase-3 substrates with cleavage generating a 95-kDa C-terminal membrane fragment (41,42). This "channel-only" fragment has been reported to be constitutively open and to induce leakage of Ca 2ϩ from the ER when transiently expressed in COS cells (43). Stable expression of the 95-kDa fragment in DT40 TKO cells was also able to support caspase-3 activation and apoptosis, whereas a mutant IP 3 R that could not be cleaved by caspase-3 was inactive in these assays (31). These studies emphasize that the role of the type I IP 3 R in apoptosis may be independent of its function in IP 3 -mediated Ca 2ϩ release.
The results in Fig. 6 are in agreement with the findings of others that IP 3 Rs play a role in sustaining caspase-3 activation (31,40). We find that the non-phosphorylatable S2681A mutant is more effective in this regard than wild-type IP 3 Rs or the S2681E mutant. These data show that one functional effect of Akt phosphorylation of IP 3 Rs is to suppress the activation of downstream caspases. The exact mechanism by which IP 3 Rs regulate the activation of caspase-3 is unknown. For agents that activate the intrinsic apoptotic pathway, caspase-3 activation occurs as a secondary consequence of cytochrome c release from the mitochondria, a process that is known to be sensitized by Ca 2ϩ (reviewed in Refs, 44 -46). A number of recent studies have highlighted the importance of the IP 3 R in mediating transfer of Ca 2ϩ from the ER to the mitochondria in several models of apoptosis (47,48). The caspase-cleaved IP 3 R, rather than the full-length form, seems particularly important for sustaining apoptosis in DT40 cells (31). Our findings raise the possibility that Akt phosphorylation of the IP 3 R may suppress Ca 2ϩ leakage through the channel-only 95-kDa caspase-3 cleavage fragment or inhibit in some other manner the IP 3 R-mediated transfer of Ca 2ϩ from the ER to the mitochondria.
Other possible IP 3 -independent functional effects of IP 3 R phosphorylation can be envisaged. More than 25 proteins have been found to interact with IP 3 Rs, leading to the suggestion that the receptor may function in ER membranes to assemble and integrate signaling systems (49). Cytochrome c and Bcl-2 both bind to the IP 3 R C-terminal tail containing the Akt phosphorylation site and regulate channel function (38,47). However, IP 3 R interaction with cytochrome c (data not shown) or Bcl-2 3 was not modified by Akt phosphorylation. Altered interactions of the phosphorylated IP 3 R with other target proteins cannot presently be excluded.