Functional Consequences of Phosphomimetic Mutations at Key cAMP-dependent Protein Kinase Phosphorylation Sites in the Type 1 Inositol 1,4,5-Trisphosphate Receptor*

Regulation of Ca 2 (cid:1) release through inositol 1,4,5-trisphosphate receptors (InsP 3 R) has important conse- quences for defining the particular spatio-temporal properties of intracellular Ca 2 (cid:1) signals. In this study, regulation of Ca 2 (cid:1) release by phosphorylation of type 1 InsP 3 R (InsP 3 R-1) was investigated by constructing “phosphomimetic” charge mutations in the functionally important phosphorylation sites of both the S2 (cid:1) and S2 (cid:2) InsP 3 R-1 splice variants. Ca 2 (cid:1) release was investigated following expression in Dt-40 3ko cells devoid of endogenous InsP 3 R. In cells expressing either the S1755E S2 (cid:1) or S1589E/S1755E S2 (cid:2) InsP 3 R-1, InsP 3 -in-duced Ca 2 (cid:1) release was markedly enhanced compared with nonphosphorylatable S2 (cid:1) S1755A and S2 (cid:2) S1589A/ S1755A mutants. Ca 2 (cid:1) release through the S2

Inositol 1,4,5-trisphosphate receptors are intracellular ion channels that function to couple the activation of cell surface receptors for neurotransmitters, hormones, and growth factors to the initiation of intracellular Ca 2ϩ release (1). Three genes have been cloned that encode distinct proteins of a molecular mass of ϳ300 kDa, named the type 1 (InsP 3 R-1), 1 type 2 (InsP 3 R-2), and type 3 (InsP 3 R-3) InsP 3 Rs (2)(3)(4). In addition, multiple receptor proteins with distinct tissue distributions are produced by alternate splicing of the type 1 receptor gene (5,6). Most cells express multiple isoforms of InsP 3 R (7). Furthermore, the expression level and complement of receptors differ in individual tissues, and this together with regulation of the activity of the channel is thought to be a major determinant of the rich diversity of Ca 2ϩ signaling events observed in cells (7,8).
The functional channel is formed co-translationally by the tetrameric association of four individual receptor subunits (9,10). Each subunit has a binding site for InsP 3 toward the N terminus formed by a cluster of positively charged amino acids thought to coordinate the negatively charged phosphate groups of InsP 3 (11,12). The C terminus of each subunit is postulated to span intracellular membranes six times and forms a single cation-selective pore (13,14). In addition, this region signals retention of the protein to the endoplasmic reticulum (15,16). Although the InsP 3 -binding pocket and channel pore are highly conserved between InsP 3 R family members, the intervening sequence between the binding region and pore is more divergent and consists of the so-called "regulatory and coupling" or "modulatory" domain. This region, consisting of ϳ1600 amino acids, is thought to be important in modulating the Ca 2ϩ release properties of the InsP 3 R. Indeed, Ca 2ϩ release through the InsP 3 R is markedly influenced by many factors, most importantly by Ca 2ϩ itself (17). InsP 3 R activity is also influenced through interaction with numerous factors such as proteins, adenine nucleotides, and phosphorylation and in particular by cyclic nucleotide-dependent kinases (8).
Two protein kinase A (PKA) consensus sites (RRXS) at Ser-1589 and Ser-1755 are present in the InsP 3 R-1 (5,18), and the most recent studies suggest that phosphorylation of these sites results in a marked enhancement of Ca 2ϩ release (19 -22). Most interestingly, these sites are conserved through evolution from Drosophila to humans in InsP 3 R-1, but corresponding sites are not present in either the InsP 3 R-2 or InsP 3 R-3. It should be noted, however, that other regions, which are presently not defined, appear to function as PKA phosphorylation sites in InsP 3 R-2 and InsP 3 R-3. Several reports have demonstrated biochemically that both Ser-1589 and Ser-1755 can be phosphorylated in InsP 3 R-1 (18,23,24); however, these studies have not provided a consensus as to which sites are important physiologically in tissue that expresses either of the two major splice variants of InsP 3 R-1.
A recent study (19) from our laboratory has elucidated the particular sites that are functionally important for the phosphoregulation of the two major splice variants of the InsP 3 R-1. These studies were performed by using mutational analysis by substituting alanine for the serine residues present at individual putative phosphorylation sites. Although phosphorylation of both neuronal (S2ϩ) and peripheral (S2Ϫ) forms of InsP 3 R-1 by PKA resulted in enhanced Ca 2ϩ release, mutational analysis indicated that only phosphorylation of Ser-1755 was functionally important in the neuronal S2ϩ InsP 3 R-1. In contrast, both Ser-1589 and Ser-1755 appeared to be phosphorylated and significant in the peripheral S2Ϫ form of InsP 3 R-1. In addition, although the S2ϩ form of the receptor was subject to direct phosphoregulation by cGMP-dependent protein kinase (PKG), the S2Ϫ form was not influenced by activation of PKG. These data represent one of the few major differences reported for the regulation of the two major splice variants of the InsP 3 R-1.
In the present study, we have constructed charge mutations, substituting glutamate residues for the serine residues in the functionally important phosphorylation sites in both S2Ϫ and S2ϩ variants of the InsP 3 R-1. These mutations are predicted to mimic phosphorylation and to allow the assessment of the functional effects of phosphorylation of InsP 3 R-1. Most importantly, the Ca 2ϩ release properties of phosphomimetic mutations are predicted to be essentially independent of cell typespecific factors, including the expression of accessory proteins such as protein A-kinase anchoring proteins. These effects should also be unambiguously specific to InsP 3 R and thus independent of confounding PKA effects on other Ca 2ϩ -handling machinery. This latter consideration has historically plagued the functional assessment of PKA phosphorylation of InsP 3 R. These mutations have allowed us to define the relative sensitivity of Ca 2ϩ release of the phosphomimetic mutations and to confirm which sites are important in each splice variant. The present study has also addressed whether phosphorylation of individual sites is permissive or additive in each splice variant, and we have investigated the consequences of phosphorylating InsP 3 R-1 on Ca 2ϩ oscillations, the physiological pattern of Ca 2ϩ signaling in nonelectrically excitable cells.

MATERIALS AND METHODS
The acetoxymethyl esters of Fura-2 and Fluo-4 were purchased from Molecular Probes (Eugene, OR). Cell-permeable cyclic nucleotides and forskolin were purchased from Biomol (Plymouth Meeting, PA). All other chemicals were purchased from Sigma. The Dt-40 cells lacking InsP 3 R (Dt-40 3ko) were kindly provided by Dr. Kurosaki (Kansai Medical University, Japan) and were maintained as described previously (25)(26)(27).
Production of Mutations-The rat S2ϩ InsP 3 R-1 in the expression plasmid pIRES-GFP was digested with the restriction endonuclease SalI. The overhang created by digestion was blunted by using T4 polymerase. An EcoRI linker was then ligated onto the blunted ends of the construct. The entire receptor DNA was excised from the plasmid using EcoRI and ligated into the plasmid MXT-1. The region containing the S2Ϫ splice variant and potential PKA phosphorylation sites was excised from its backbone in pCDNA 3.1ϩ by RsrII and KasI and ligated into the InsP 3 R construct in MXT-1. The potential PKA phosphorylation sites Ser-1589 and Ser-1755 were mutated, individually in both splice variants and together in the S2Ϫ splice variant, to alanines or glutamates by using sequential PCR mutagenesis. The outside primers used for the mutagenesis reaction flanked the restriction sites RsrII and KasI. Following mutation, the resulting fragments were cut with RsrII and KasI and inserted into the InsP 3 R-1 backbone at the corresponding sites. The mutations were confirmed by Big Dye fluorescent sequencing. Mutated receptor DNAs were excised from MXT-1 by using EcoRI and ligated into the mammalian expression vector pGW (provided by Dr. David Yue, The Johns Hopkins University). Orientation was confirmed by using restriction enzyme digestion. Mutants were named based on the splice variant, either S2ϩ or S2Ϫ followed by the amino acid present at position 1589 and 1755. Thus, a mutation in S2ϩ InsP 3 R-1 S1755E is designated,"S2ϩ SE," and in InsP 3 R-1 S2ϩ S1755A is designated "S2ϩ SA". Similarly, an S2Ϫ InsP 3 R-1 with mutations in both S1589E and S1755E is designated "S2Ϫ EE". The numbering of residues is based on the full-length rat InsP 3 R-1.
Transfection of Dt-40 Cells-Dt-40 cells lacking all three InsP 3 receptor subtypes were transfected by using electroporation at 350 V and 950 microfarads (4-mm gap cuvette). 2 ϫ 10 7 cells were co-transfected with 25 g of the InsP 3 R-1 cDNA, 25 g of the muscarinic type 3 (m3R) receptor DNA, and 4 g of the red fluorescent protein plasmid pH-cRed1-N1 (Clontech). Cells were incubated with DNA in 500 l of Opti-MEM media (Invitrogen) on ice for 10 min. The cell/DNA mixture was electroporated, incubated on ice for 30 min, increased to 5 ml with Opti-MEM, and placed in a 5% CO 2 incubator at 39°C for 5 h. The cells were then centrifuged and resuspended in 12 ml of complete RPMI media (Invitrogen). Transfection efficiency was typically ϳ20%. Experiments were performed within 32 h of transfection.
Transfection of HEK-293 Cells and Assessment of Phosphorylation of S2Ϫ InsP 3 R-1-HEK-293 cells were plated onto 25-cm 2 culture flasks and allowed to grow to near-confluency. Cells were transfected with 5 g of each S2Ϫ InsP 3 R-1 DNA construct by using the LipofectAMINE reagent (Invitrogen) as per the manufacturer's instructions. The following day, batches of cells were treated in the presence or absence of 20 M forskolin for 10 min, aspirated from flasks, lysed, and immunoprecipitated with a polyclonal ␣-InsP 3 R-1 antibody that recognizes amino acids 2731-2749 of InsP 3 R-1. Immunoprecipitates were separated on 5% SDS gels transferred to nitrocellulose and then probed with either the ␣-InsP 3 R-1 antibody or a polyclonal antibody that recognizes the phosphorylated state of Ser-1755 (28) (␣-phospho-Ser-1755), kindly provided by Dr. S. Snyder. Blots that were probed with ␣-phospho-Ser-1755 were stripped and reprobed with the ␣-InsP 3 R-1 antisera to confirm the presence and relative quantity of the InsP 3 R-1.
Digital Imaging of [Ca ϩ2 ] i -Transfected Dt-40 3ko cells were washed once in a HEPES-buffered physiological saline solution (HEPES-PSS) containing (in mM) 5.5 glucose, 137 NaCl, 0.56 MgCl 2 , 4.7 KCl, 1 Na 2 HPO 4 , 10 HEPES (pH 7.4), 1.2 CaCl 2 , and 1% w/v bovine serum albumin. Cells were then resuspended in bovine serum albumin HEPES-PSS with 1 M Fura-2 (AM), placed on a 15-mm glass coverslip in a low volume perfusion chamber, and allowed to adhere for 30 min at room temperature. Cells were perfused continuously for 10 min with HEPES-PSS before experimentation to allow complete Fura-2 de-esterification. A field of cells for each experiment was chosen that contained a wide range of transfection efficiency based upon the intensity of red fluorescence emitted when excited at 560 nm. Individual cells that had emission gray levels between 1500 and 2500 were subsequently chosen to standardize expression levels. [Ca 2ϩ ] i imaging was performed essentially as described previously by using an inverted epifluorescence Nikon microscope with a 40ϫ oil immersion objective lens (numerical aperture, 1.3) (19). Cells were excited alternately with light at 340 and 380 Ϯ 10 nm bandpass filters (Chroma, Rockingham, VT) using a monochrometer (TILL Photonics, Pleasanton, CA). Fluorescence images were captured and digitized with a digital camera driven by TILL Photonics software. Images were captured every 2 s with an exposure of 2 ms and 4 by 4 binning. 340/380 ratio images were calculated online and stored immediately to a hard disk.
Flash Photolysis-Transfected cells were simultaneously loaded with the visible wavelength indicator Fluo-4 and a cell-permeable form of caged inositol trisphosphate (ci-IP 3 /PM) for 30 min. ci-IP 3 /PM is a homologue of cm-IP 3 /PM. The 2-and 3-hydroxyls of the inositol ring are protected by an isopropylidene group in ci-IP 3 /PM and are protected by a methoxymethylene group in cm-IP 3 /PM (29). Like cm-IP 3 /PM, ci-IP 3 /PM diffuses across cell membranes, and the PM group is hydrolyzed by cellular esterases, and Ca 2ϩ release can be induced upon photouncaging as i-IP 3 is liberated from the cage and acts in a similar fashion to InsP 3 at InsP 3 R. 2 A further period of ϳ30 min was allowed for de-esterification of both dye and cage. Cells were illuminated at 488 Ϯ 10 nm and fluorescence collected through a 525 Ϯ 25-nm bandpass filter and captured using the Till Photonics imaging suite. These traces are displayed as % ⌬F/F o , where F is the recorded fluorescence, and F o is the mean of the initial 10 sequential frames. Photolytic release was performed as described previously by using a pulsed xenon arc lamp (Till Photonics). A high intensity (0.5-5 ms duration; 80 J) discharge of UV light (360 Ϯ 7.5 nm) was reflected onto the plane of focus by using a DM400 dichroic mirror and Nikon 40ϫ oil immersion objective, 1.3 NA.
Frequency Distribution of Ca 2ϩ Oscillations-Cells transfected with WT and mutant S2Ϫ isoforms were stimulated with varying concentrations of an ␣-IgM antibody (Southern Biotechnology Associates, Inc., Birmingham, AL). Infrequent Ca 2ϩ oscillations were produced presumably through B cell receptor cross-linking, activation of phospholipase C-␥, and subsequent production of InsP 3 . The frequency of Ca 2ϩ oscillations was determined by selecting individual peaks that displayed an increase in ratio units greater than 0.05 and are listed as frequency in milliHertz (number of oscillations in 1000 s).
Concentration-Response Relationships-Normalized ⌬F concentration-response relationships were fit with the following logistic Equation 1, where ⌬F is the change in fluorescence normalized to the maximal response; C is agent concentration; EC 50 is the concentration where the response is half of maximum, and Slope ⌬F is a slope factor related to the Hill coefficient. Statistical Analysis-The effects of treatment were determined by normalizing the peak change in fluorescence ratio by stimulation following forskolin or 8-Br-cGMP exposure to that of stimulation in control HEPES-PSS. Thus, pooled data represent a normalized fold increase over control for the treated trial. In all cases where statistical significance is indicated, two-tailed heteroscedastic t tests were performed. p values Ͻ 0.05 were considered to indicate statistical significance and are denoted by an asterisk in the figures.

Phosphomimetic Mutations in Functionally Important Phosphorylation Sites in Both S2ϩ and S2Ϫ InsP 3 R-1 Result in
Enhanced Ca 2ϩ Release-Phosphorylation of proteins results in the addition of net negative charge to the phosphoacceptor residue. In the case of PKA phosphoregulation, the functional effects of phosphorylation are thought to occur as the negative charge added to the serine or (less frequently) threonine residues neutralizes the positive charge of basic arginine or lysine residues present upstream in the classical RRX(S/T) consensus motif (30). This charge neutralization, in turn, is thought to result in a conformational change in the protein. A common approach employed to investigate the functional effects of phosphorylation is to construct phosphomimetic mutations whereby glutamic or aspartic acid residues are substituted at the phosphoacceptor site (31,32). The rationale for this strategy is that the negatively charged side chain of the substituted acidic amino acid will mimic, to an extent, the addition of a phosphate moiety to the protein.
To investigate the consequences of InsP 3 R-1 phosphorylation, we analyzed in Dt-40 3ko cells the Ca 2ϩ release properties of phosphomimetic mutations in the functionally important sites in both S2Ϫ and S2ϩ InsP 3 R-1. This cell line provides the only known InsP 3 R null background (26,27). In initial experiments, a comparison was made between the sensitivity of Ca 2ϩ release by InsP 3 R-1 phosphomimetic mutations versus nonphosphorylatable alanine mutations at the sites we reported previously (19) to be relevant. Dt-40 3ko cells were transfected with DNA encoding HcRed to facilitate identification of transfected cells and either S2Ϫ S1589E/S1755E InsP 3 R-1 (S2Ϫ EE) or S2Ϫ S1589A/S1755A (S2Ϫ AA). Ca 2ϩ release was monitored following flash photolysis of ci-IP 3 , a cell-permeable form of caged InsP 3 . This experimental paradigm provides a relatively direct assessment of the effects of InsP 3 R-1 phosphorylation on the Ca 2ϩ release process. The amount of ci-IP 3 photo-released was controlled by varying the duration of the UV flash dis-charge (0.5-5 ms). No increase in [Ca 2ϩ ] i was observed following the longest UV discharge in cells either not loaded with ci-IP 3 /PM or not expressing HcRed. Fig. 1, B and C, shows traces from typical experiments in individual Dt-40 3ko cells expressing S2Ϫ EE or S2Ϫ AA (Fig. 1, B and C, respectively, and pooled data in Fig. 1F). In cells expressing S2Ϫ EE, Ca 2ϩ release as defined by a Ͼ0.05% ⌬F/F o increase in initial fluo-4 fluorescence was observed in ϳ60% of cells when exposed to UV discharge for 0.5 ms. Subsequent exposure to UV light for 1.25 ms elicited a more robust increase in [Ca 2ϩ ] i in all cells. Finally, photo-release following a 5-ms flash, in general evoked a further increase in the magnitude of Ca 2ϩ release. In contrast, an elevation of intracellular Ca 2ϩ was never observed under identical conditions following a 0.5-ms UV flash in cells expressing S2Ϫ AA. A significant Ca 2ϩ release was only observed in ϳ50% of cells following a 1.25-ms flash, and the majority of cells only responded to the longest uncaging duration, albeit with a smaller magnitude than in S2Ϫ EE-expressing cells exposed to the same stimulus.
A similar pattern of sensitivity was observed in Dt-40 3ko cells expressing S2ϩ S1755E (S2ϩ SE) and S2ϩ S1755A (S2ϩ SA) as shown in Fig. 1, D and E, respectively, and pooled data in Fig. 1G. Cells expressing S2ϩ SE responded more robustly to photolysis of ci-IP 3 than cells expressing S2ϩ SA exposed to an identical stimulus. These data indicate that serine to glutamate mutations at the functionally important phosphorylation sites in both splice variants of InsP 3 R-1 are "phosphomimetic," i.e. charge mutations mimic phosphorylation in that these constructs display an apparent increased functional sensitivity to InsP 3 .
Sensitivity of S2Ϫ InsP 3 R-1 and Phosphorylation Site Mutants-Experiments were next performed to determine the relative sensitivity of Ca 2ϩ release through the phosphomimetic mutations of the S2Ϫ InsP 3 R-1 with respect to the wild type and phosphorylation-deficient mutants. Dt-40 3ko cells were transfected with cDNAs encoding m3R, HcRed together with the InsP 3 R-1 construct of interest. Stimulation with the muscarinic agonist CCh results in robust increases in [Ca 2ϩ ] i in transfected cells through the G␣ q/11 -coupled stimulation of phospholipase C-␤ and subsequent formation of InsP 3 . The magnitude of the initial peak provides a good estimation of the extent of Ca 2ϩ release as this parameter in Dt-40 cells, like many cells, is essentially independent of Ca 2ϩ influx. Furthermore, in Dt-40 cells the [Ca 2ϩ ] i response to stimulation with CCh does not appreciably desensitize, and thus the effects of multiple concentrations of agonist can be assessed in a single cell. Individual HcRed-expressing cells were stimulated with increasing concentrations of CCh (1 nM to1 M) for 60 s followed by a 5-min wash between applications of agonist. In each case, these experiments were performed on multiple cells, expressing a narrow range of HcRed fluorescence and from multiple batches of transfected cells to minimize variation because of expression level of m3R and InsP 3 R-1. Concentration-response relationships were generated by normalizing each initial peak to the maximum response in the individual cell and subsequently averaging the pool of cells expressing a particular construct. Fig. 2A shows a typical example of this experimental procedure performed on Dt-40 3ko cells expressing S2Ϫ EE. A significant increase in Ca 2ϩ was detected in the majority of S2Ϫ EE-expressing cells following stimulation with 1 nM CCh, and the peak response occurred following stimulation between 10 and 50 nM CCh. In contrast, as shown in Fig. 2, B and C, Ca 2ϩ release following CCh stimulation in the wild type S2Ϫ InsP 3 R-1 or S2Ϫ AA was considerably less sensitive to CCh stimulation. Analysis of the pooled data indicated that the magnitude of the peak response to any of these constructs was not significantly different (maximum peak ⌬ response: S2Ϫ WT ϭ 0.9 Ϯ 0.1 ratio units; S2Ϫ AA ϭ 0.67 Ϯ0.1 ratio units, and S2Ϫ EE ϭ 1.13 Ϯ 0.1 ratio units), suggesting that the efficacy of Ca 2ϩ release was essentially unaltered. However, when an estimate of the relative sensitivity was made by fitting the normalized concentration-response relationships for each construct (Fig. 2D), CCh-induced Ca 2ϩ release in the S2Ϫ EE mutant (EC 50 4.3 Ϯ 1.2 nM CCh) was 7.5-fold more sensitive when compared with WT S2Ϫ InsP 3 R-1 (EC 50 ϭ 32.6 Ϯ 7.5 nM) and some 50-fold more sensitive than the nonphosphorylatable S2Ϫ AA mutant (EC 50 ϭ 229.5 Ϯ 14.6 nM). These data provide strong evidence that the S2Ϫ EE construct is more sensitive to stimulation by InsP 3 and present evidence that phosphorylation of the InsP 3 R-1 results in marked regulation of channel function. This profound regulation could be expected to have major consequences for calcium signaling events in peripheral tissue such as liver, testis, and smooth muscle which express the S2Ϫ InsP 3 R-1 (5,6). The observation that S2Ϫ AA is relatively less sensitive than S2Ϫ WT is consistent with the possibility that a proportion of the wild type receptor is constitutively phosphorylated in Dt-40 cells, thus contributing to the intermediate sensitivity of the wild type S2Ϫ InsP 3 R-1 relative to the phosphomimetic S2Ϫ InsP 3 R-1 receptor.
Although we have demonstrated that each site can be phosphorylated, it is not definitively known whether the functional effects of phosphorylating individual sites are independent and additive or alternatively if the full effect is seen following phosphorylation of an individual site. Thus, experiments were next performed to assess the sensitivity of single phosphomimetic mutations within each phosphorylation site. Concentration-response relationships for CCh-induced Ca 2ϩ release were constructed for Dt-40 3ko cells expressing either S2Ϫ S1755E InsP 3 R-1 (S2Ϫ SE) or S2Ϫ S1589E InsP 3 R-1 (S2Ϫ ES). The data were analyzed as described previously for Fig. 2D. Fig. 3A shows the fit for the normalized concentration-response relationship for S2Ϫ SE and S2Ϫ ES and for comparison also shows the fit for S2Ϫ EE and S2Ϫ WT (Fig. 3A, dotted lines; data from Fig. 2D). Once again the maximal initial peak responses in either mutant were not significantly altered from wild type (S2Ϫ ES ϭ 0.54 Ϯ 0.1 ⌬ ratio units; S2Ϫ SE ϭ 0.78 Ϯ 0.1 ⌬ ratio units); however, the sensitivity of each of these mutants was significantly shifted, such that the EC 50 for CCh-induced Ca 2ϩ release was enhanced ϳ3-fold over the response in wild type for either mutant (EC 50 for CCh-induced release: S2Ϫ ES ϭ 12.4 Ϯ 0.5 nM; S2Ϫ SE ϭ 13.5 Ϯ 1 nM). These data are summarized in Table I.
The enhanced apparent sensitivity of individual phosphomimetic mutants was essentially equal and intermediate between the sensitivity of the S2Ϫ WT and S2Ϫ EE mutations. These data are consistent with phosphorylation of individual sites being functionally additive. To address this possibility directly, experiments were performed evaluating the effects of activating endogenous PKA in cells expressing S2Ϫ SE or S2Ϫ ES to mimic prior phosphorylation of an individual site. As shown in Fig. 3B, and reported previously, activation of PKA by incubation with forskolin results in a dramatic potentiation of Ca 2ϩ release in Dt-40 3ko cells expressing S2Ϫ WT (19). Most surprisingly, although the sensitivity to CCh was enhanced, as evidenced by the low concentration of CCh necessary to evoke threshold Ca 2ϩ release, no potentiation of Ca 2ϩ release following forskolin incubation was observed in Dt-40 3ko cells expressing either S2Ϫ ES ( Fig. 3C; pooled data in Fig. 3F), S2Ϫ SE ( Fig. 3D; pooled data in Fig. 3F), or S2Ϫ EE ( Fig. 3E; pooled data in Fig. 3F). Thus, although each phosphorylation site in S2Ϫ InsP 3 R-1 can be phosphorylated and mimicking phosphorylation of both sites leads to a receptor with enhanced sensitivity relative to phosphorylation of an individual site, phosphorylation of both sites in situ does not appear to occur and therefore is not functionally additive. Although these data may seem paradoxical, a possible explanation, consistent with all the observations is that the initial phosphorylation of either Ser-1589 or Ser-1755 leads to a conformational change in the receptor that now precludes the phosphorylation of the additional site. To test this hypothesis, experiments were performed to determine the phosphorylation state of the various mutants after raising cAMP levels. These experiments were performed in HEK-293 cells because of the high transfection efficiency and low endogenous levels of InsP 3 R-1 (18,33). HEK-293 cells were transfected with the constructs as indicated in Fig. 4, incubated in the presence or absence of forskolin for 10 min, then pelletted, and lysed. Following incubation of the lysates with ␣-InsP 3 R-1 antibody, the immune complexes were captured and separated on SDS gels, and the phosphorylation status of Ser-1755 was determined by Western blotting with an antibody that specifically recognizes this phosphorylated residue in InsP 3 R-1 (28). As shown in Fig. 4A, wild type S2Ϫ InsP 3 R-1 was robustly phosphorylated after forskolin incuba-

FIG. 2. InsP 3 sensitivity of S2؊ InsP 3 R-1 and phosphomimetic mutations. Concentration-response relationships for CCh-induced
Ca 2ϩ signals were examined in Dt-40 3ko cells expressing m3 receptors and InsP 3 R-1 constructs. A, cells transfected with S2Ϫ EE were stimulated with increasing concentrations of CCh as indicated (n ϭ 9 cells). A similar paradigm was performed for cells expressing S2Ϫ WT (n ϭ 10) (B) and for cells expressing S2Ϫ AA constructs (n ϭ 9) (C). The magnitude of the initial peak (as an indicator of Ca 2ϩ release) for each response was normalized to the maximum response in each cell. The pooled data and the fit that describes each concentration-response relationship for each construct is shown in D and illustrates that the sensitivity of Ca 2ϩ release was greatest in S2Ϫ EE followed by S2Ϫ WT and S2Ϫ AA. tion, whereas no phosphorylation was detected in wild type S2Ϫ AA or untransfected HEK-293 cells. Similarly, in Fig. 4B, a marked increase in phosphorylation could be detected in S2Ϫ InsP 3 R-1, whereas no phosphorylation could be detected in cells transfected with S2Ϫ EE, S2Ϫ SE or, most importantly, the S2Ϫ ES construct. These data are strongly supportive of the contention that the initial phosphorylation of one site precludes further phosphorylation at the additional residue as Phosphoregulation of S2 ϩ/Ϫ InsP 3 R-1 46247 suggested by the functional data and indicate that physiologically only phosphorylation of a single residue is functionally relevant in the S2Ϫ splice variant of InsP 3 R-1.

Sensitivity of S2ϩ InsP 3 R-1 and Phosphorylation Site Mutants-
The sensitivity of Ca 2ϩ release via the neuronal S2ϩ InsP 3 R-1 was next assessed by using a similar experimental paradigm to that used for the S2Ϫ form of the receptor. Although biochemically both Ser-1589 and Ser-1755 are equally susceptible to phosphorylation by PKA in S2ϩ InsP 3 R-1 (18), only Ser-1755 appears to be functionally relevant in terms of modulating Ca 2ϩ release (19). Fig. 5A shows the fits for the normalized concentration-response relationships for CCh-induced Ca 2ϩ release in cells expressing S2ϩ WT, phosphomimetic S2ϩ SE, and S2ϩ SA. The apparent sensitivity of CChinduced Ca 2ϩ release in cells expressing S2ϩ WT was essentially identical to S2Ϫ WT InsP 3 R-1 (EC 50 for CCh-induced Ca 2ϩ release; S2ϩ InsP 3 R-1 ϭ 32.1 ϩ 4.2 nM versus S2Ϫ 32.6 ϩ 7.5 nM). These data are in agreement with the published literature that indicates that InsP 3 binding and InsP 3 -induced calcium release is identical in the two major splice variants of the InsP 3 R-1 (34 -36) and is therefore supportive of our contention that the initial CCh-stimulated [Ca 2ϩ ] i peak is a good indicator of InsP 3 R function. In a similar fashion to the S2Ϫ InsP 3 R-1, mutation of Ser-1755 to either alanine or glutamic acid did not significantly alter the maximal initial peak upon CCh stimulation (maximum peak ⌬ response: S2ϩ WT ϭ 0.61 Ϯ 0.1 ⌬ ratio units; S2ϩ SE ϭ 0.59 Ϯ 0.1 ⌬ ratio units; S2ϩ SA ϭ 0.84 Ϯ 0.2 ⌬ ratio units) but did, however, significantly affect the apparent sensitivity of Ca 2ϩ release as shown in Fig.  5A. The S2ϩ SE mutant exhibited a similar EC 50 for CChinduced Ca 2ϩ release as the S2Ϫ EE mutation (EC 50 : S2ϩ SE ϭ 3.6 Ϯ 0.2 nM CCh; S2Ϫ EE ϭ 4.3 nM), being ϳ9-fold more sensitive than S2ϩ WT and ϳ32-fold more sensitive than the nonphosphorylatable S2ϩ SA mutation (EC 50 S2ϩ SA ϭ 116.3 Ϯ 3 nM CCh). These data are summarized in Table I.
In contrast to the S2ϩ InsP 3 R-1 S1755E construct, mutation of Ser-1589 to glutamate in the S2ϩ form of InsP 3 R-1 did not significantly affect the apparent sensitivity of Ca 2ϩ release as shown in Fig. 5B. Both the maximum peak response to CCh and the sensitivity of the receptor were very similar to that of S2ϩ WT InsP 3 R-1 (maximum peak ⌬ response ϭ 0.59 Ϯ 0.2 ⌬ ratio units; EC 50 ϭ 23.7 Ϯ 1.2 nM CCh). However, it is formally possible that phosphorylation of Ser-1589 is only functionally important following phosphorylation of Ser-1755 in S2ϩ InsP 3 R-1. To test this idea, experiments were performed activating PKA with forskolin in Dt-40 3ko cells expressing S2ϩ ES or SE mutants. Treatment with forskolin resulted in a marked potentiation of CCh-induced Ca 2ϩ release in cells expressing either S2ϩ WT or S2ϩ ES (Fig. 6, A and B, respectively, and pooled data in D) presumably as Ser-1755 was phosphorylated. Although cells expressing S2ϩ SE were more sensitive to CCh, no further potentiation was observed following forskolin incubation ( Fig. 6C and pooled data in Fig. 6D). These data indicate that it is unlikely that phosphorylation of Ser-1755 is permissive for any functional effect of phosphorylating Ser-1589 in S2ϩ InsP 3 R-1.
Effect of PKG Activation on S2ϩ InsP 3 R-1 and Phosphomimetic Mutants-We have reported previously that PKG activation results in direct phosphoregulation of only the neuronal S2ϩ form of InsP 3 R-1 (19). In addition, PKG regulation of Ca 2ϩ release only occurs by phosphorylation of Ser-1755. In contrast to these data, a recent report (18) has shown that Ser-1589 and not Ser-1755 is phosphorylated upon activation of PKG. It should be noted that these experiments were performed by using the mouse S1ϩ/S2ϩ splice variant of InsP 3 R-1, whereas our functional experiments utilized rat S1Ϫ/S2ϩ InsP 3 R-1. The species and splice variant differences aside, the reported differences are difficult to reconcile with data demonstrating that mimicking phosphorylation of Ser-1589 in S2ϩ InsP 3 R-1 does not influence Ca 2ϩ release (Figs. 5 and 6). Despite these differences, we performed experiments to define further the phosphorylation sites important for regulation of Ca 2ϩ release by PKG. Fig. 7A shows a representative trace from cells express-  ing S2ϩ WT InsP 3 R-1 thus illustrating the marked potentiation of Ca 2ϩ release upon specific activation of PKG by 8-Br-cGMP. A similar striking potentiation of Ca 2ϩ release was also observed in cells expressing S2ϩ ES ( Fig. 7B and pooled data Fig. 7D), again presumably as Ser-1755 is phosphorylated following incubation with 8-Br-cGMP. In contrast, in cells expressing S2ϩ SE no effect on CCh-induced Ca 2ϩ release following activation of PKG was observed ( Fig. 7C and pooled data in Fig. 7D). These data provide evidence that it is unlikely that phosphorylation of Ser-1589 by PKG either in isolation or following phosphorylation of Ser-1755 plays a role in modulation Ca 2ϩ release through S2ϩ InsP 3 R-1. However, we cannot rule out that phosphorylation of this residue impacts other processes, such as protein interactions or clustering important for InsP 3 R-1 function.
The specific pattern of phosphorylation occurring upon stimulation of PKA or PKG could be cell type-specific as, for example, in the case of PKA as a result of the targeting through protein A-kinase anchoring proteins, as has been demonstrated recently (37) for InsP 3 R. Alternatively, cell-specific effects could conceivably occur through restricted access of the kinase to its substrate. Notwithstanding the general importance of kinase targeting for efficient, localized phosphorylation, the phosphomimetic constructs used in this study reveal the intrinsic, functionally important sites in a manner independent of the particular cellular context because any targeting step is circumvented. These sites and the functional consequences of phosphoregulation are thus likely a general property of the InsP 3 R-1. In addition, the particular sites in S2ϩ InsP 3 R-1 are entirely consistent with our earlier study (19) of the sites functionally important in enhanced Ca 2ϩ release following PKA or PKG phosphoregulation, and this reinforces the view that phosphorylation of S2ϩ InsP 3 R-1 Ser-1589 has no significant role, at least in terms of Ca 2ϩ release.
Effect of InsP 3 R Phosphorylation on Ca 2ϩ Signaling Events-In many nonelectrically excitable cell types, the physiological mode of Ca 2ϩ signaling is through the generation of Ca 2ϩ oscillations (1,38). Moreover, it is a generally held view that the spatial and temporal properties of Ca 2ϩ oscillations make an important contribution to defining the fidelity and specificity of Ca 2ϩ signaling. In many current models addressing the mechanism underlying Ca 2ϩ oscillations, a key feature is the regulation of Ca 2ϩ release through InsP 3 R. We therefore next designed experiments to assess the consequences of InsP 3 R-1 phosphorylation (expressed in isolation) on the initiation and generation of Ca 2ϩ oscillations. In preliminary ex- FIG. 5. Functional sensitivity of S2؉ InsP 3 R-1. Concentration-response relationships were generated exactly as described previously. A, the normalized relationships for S2ϩ WT (n ϭ 11), S2ϩ SE (n ϭ 6), and S2ϩ SA (n ϭ 5) are shown. The fits illustrate the increased sensitivity of the S2ϩ SE relative to S2ϩ WT and S2ϩ SA constructs. B, the normalized concentration-response relationship for S2ϩ ES is shown with S2ϩ SE and S2ϩ WT (dotted lines for comparison). These data indicate that the sensitivity of S2ϩ ES is similar to S2ϩ WT.
periments we failed to initiate Ca 2ϩ oscillations with reproducible characteristics with CCh in m3R-transfected cells (data not shown). Therefore, we chose to stimulate cells with an ␣-IgM antibody and to initiate Ca 2ϩ oscillations through activation of the endogenous B cell receptor, phospholipase C-␥ activation, and the formation of InsP 3 . Dt-40 3ko cells expressing either S2Ϫ WT or S2Ϫ EE to mimic PKA phosphorylation of the InsP 3 R-1 were stimulated with various concentrations of ␣-IgM antibody as illustrated by the selection of representative traces in Fig. 8. Stimulation with 250 ng/ml ␣-IgM proved to be a threshold concentration in S2Ϫ WT-expressing cells. This degree of stimulation generally resulted in a single small increase in [Ca 2ϩ ] i after a long latency (Fig. 8A, left panel, and pooled data in Fig. 8, C-E). An identical stimulus in cells expressing S2Ϫ EE, in contrast, resulted in repetitive Ca 2ϩ transients following a much shorter latency, consistent with the increased apparent sensitivity of the S2Ϫ EE constructs (Fig. 8A, right panel, and pooled data Fig. 8, C-E). Most interestingly, stimulation of S2Ϫ WT-expressing cells with 500 ng/ml ␣-IgM an-FIG. 6. Effects of PKA phosphorylation on S2؉ InsP 3 R-1 single glutamate substitution constructs. Threshold CCh-stimulated Ca 2ϩ release is markedly potentiated by activation of PKA following forskolin treatment in S2ϩ WT-expressing cells (A) or S2ϩ ES (B). C, in contrast, no potentiation of threshold CCh-stimulated Ca 2ϩ release is observed in cells expressing S2ϩ SE. D, pooled data illustrate that S2ϩ ES has properties identical to S2ϩ WT, and thus phosphorylation of this residue is unlikely to impact Ca 2ϩ release. Numbers in parentheses indicate the number of analyzed cells.
FIG. 7. Effect of PKG activation in single phosphomimetic constructs. Threshold CCh-stimulated Ca 2ϩ release is markedly potentiated by activation of PKG following 8-Br-cGMP treatment in S2ϩ WT (A) or S2ϩ ES-expressing cells (B). C, in contrast. no potentiation of threshold CCh-stimulated Ca 2ϩ release is observed in S2ϩ SE-expressing cells. D, pooled data illustrate that S2ϩ ES has properties identical to S2ϩ WT, and thus phosphorylation of this residue following activation of PKG is unlikely to impact Ca 2ϩ release. Numbers in parentheses indicate the number of analyzed cells. tibody resulted in Ca 2ϩ oscillations with similar frequency to cells expressing S2Ϫ EE ( Fig. 8B and pooled data C-E). However, the latency before the initiation of an increase in Ca 2ϩ was significantly shorter in S2Ϫ EE-expressing cells. In addition, the magnitude of the initial transient was also significantly larger in S2Ϫ EE-expressing cells when compared with WT. Stimulation with 1 g/ml ␣-IgM antibody resulted in transients that were indistinguishable in terms of latency, frequency, or initial peak magnitude in S2Ϫ EE-or S2Ϫ WTexpressing cells (pooled data Fig. 8, C-E). Thus, PKA-mediated phosphorylation, by increasing the sensitivity of the InsP 3 R-1 to InsP 3 , may define the threshold at which cells begin to oscillate but does not markedly influence the temporal properties of Ca 2ϩ oscillations when initiated. This latter observation presumably reflects the fact that the frequency of oscillations is primarily defined by mechanisms such as Ca 2ϩ feedback (39) rather than the absolute sensitivity of the InsP 3 R-1 to InsP 3 within a defined range. These data are largely consistent with data from hepatocytes where PKA activation resulted in cells exhibiting a lower threshold for activation by InsP 3 infusion (40) or agonist activation (41).
These data using phosphomimetic mutations of InsP 3 R-1 splice variants are in broad agreement with a number of studies that have reported increased sensitivity of InsP 3 R-1 activity following phosphoregulation by PKA (19 -22, 41, 42). The most important consequence of this increased sensitivity appears to be in defining the threshold where a cell will respond to a stimulus. Given the almost ubiquitous expression of various forms of InsP 3 R-1, this is likely a generally important phenomenon. A number of mechanisms are plausible to explain the increased sensitivity of InsP 3 -induced Ca 2ϩ release. These include modulation of InsP 3 binding, an idea supported by measurements of InsP 3 R binding in hepatocytes (42) (presumably S2Ϫ InsP 3 R-1 and InsP 3 R-2). In these studies the apparent affinity of InsP 3 binding was enhanced ϳ2-fold at resting [Ca 2ϩ ], and the [Ca 2ϩ ] necessary for half-maximal stimulation of InsP 3 binding was reduced. In contrast, a study of recombinant S1Ϫ S2ϩ InsP 3 R-1 expressed in SF9 cells and reconsti- In contrast, multiple Ca 2ϩ transients were elicited in S2Ϫ EE-expressing cells following a shorter latency. B, cells were stimulated with 500 ng/ml ␣-IgM antibody. Cells transfected with either S2Ϫ WT or S2Ϫ EE exhibited multiple Ca 2ϩ transients of similar frequency, although the initial peak was generally larger and latency shorter in S2Ϫ EE-expressing cells. C, the pooled data from cells stimulated with various concentrations of ␣-IgM for frequency of oscillations is shown; D, is shown for latency; E, is shown for the magnitude of the initial peak. Numbers in parentheses indicate the number of analyzed cells. tuted into lipid bilayers has reported a similar increase in InsP 3 R sensitivity to InsP 3 but that the bell-shaped Ca 2ϩ sensitivity of channel opening is not altered following phosphorylation of the InsP 3 R-1 (20). These findings suggest that modulation of the Ca 2ϩ sensitivity of channel activity is unlikely to account for the increased apparent sensitivity of the receptor, at least in this form of the InsP 3 R-1. A number of alternative mechanisms for altering InsP 3 R-1 sensitivity are conceivable. For example, phosphorylation of the receptor could alter the gating of the channel directly. In addition, phosphorylation might secondarily modulate the receptor by regulation of the association of regulatory factors such as proteins or adenine nucleotides (43)(44)(45). Indeed, a precedent for this type of regulation exists because PKA phosphorylation has been shown to alter the association of calmodulin with the S2Ϫ form of InsP 3 R-1 (36). It is envisioned that further studies utilizing phosphomimetic mutations of InsP 3 R-1 will greatly facilitate elucidating the mechanism responsible for the increased sensitivity of InsP 3 -induced Ca 2ϩ release via InsP 3 R-1.