Phosphorylation of type-1 inositol 1,4,5-trisphosphate receptors by cyclic nucleotide-dependent protein kinases: a mutational analysis of the functionally important sites in the S2+ and S2- splice variants.

Inositol 1,4,5-trisphosphate receptors (InsP3R) are the major route of intracellular calcium release in eukaryotic cells and as such are pivotal for stimulation of Ca2+-dependent effectors important for numerous physiological processes. Modulation of this release has important consequences for defining the particular spatio-temporal characteristics of Ca2+ signals. In this study, regulation of Ca2+ release by phosphorylation of type-1 InsP3R (InsP3R-1) by cAMP (PKA)- and cGMP (PKG)-dependent protein kinases was investigated in the two major splice variants of InsP3R-1. InsP3R-1 was expressed in DT-40 cells devoid of endogenous InsP3R. In cells expressing the neuronal, S2+ splice variant of the InsP3R-1, Ca2+ release was markedly enhanced when either PKA or PKG was activated. The sites of phosphorylation were investigated by mutation of serine residues present in two canonical phosphorylation sites present in the protein. Potentiated Ca2+ release was abolished when serine 1755 was mutated to alanine (S1755A) but was unaffected by a similar mutation of serine 1589 (S1589A). These data demonstrate that Ser-1755 is the functionally important residue for phosphoregulation by PKA and PKG in the neuronal variant of the InsP3R-1. Activation of PKA also resulted in potentiated Ca2+ release in cells expressing the non-neuronal, S2- splice variant of the InsP3R-1. However, the PKA-induced potentiation was still evident in S1589A or S1755A InsP3R-1 mutants. The effect was abolished in the double (S1589A/S1755A) mutant, indicating both sites are phosphorylated and contribute to the functional effect. Activation of PKG had no effect on Ca2+ release in cells expressing the S2- variant of InsP3R-1. Collectively, these data indicate that phosphoregulation of InsP3R-1 has dramatic effects on Ca2+ release and defines the molecular sites phosphorylated in the major variants expressed in neuronal and peripheral tissues.

Inositol 1,4,5-trisphosphate receptors are ubiquitous ligandgated ion channels localized to the endoplasmic reticulum, which function to couple activation of cell surface receptors to intracellular Ca 2ϩ release (1). Genes have been identified that encode three distinct proteins of molecular mass ϳ300 kDa and have been named the type 1, 2, and 3 InsP 3 Rs 1 (2)(3)(4)(5). Notably, the type-1 receptor gene is alternatively spliced to yield additional variants of the receptor, which have specific tissue distribution (6,7). The functional channel is a tetramer, consisting of a binding site for InsP 3 in the N terminus of each subunit (8,9), and a single calcium-permeable pore, formed from six transmembrane-spanning helices located toward the C terminus of each subunit (10,11). Between these regions is a cytoplasmic loop of ϳ1600 amino acids, which is termed the regulatory and coupling domain. In this region, sites are located that are thought to be important in modulating the Ca 2ϩ release properties of the channel (12).
Most studies of PKA-dependent phosphorylation have been performed on the S2 ϩ neuronal type-I InsP 3 R, the so called "long form" of the receptor. In this variant of the InsP 3 R-1, serine residues at Ser-1589 and Ser-1755 are phosphorylated by PKA (7,34,35), with Ser-1755 being more heavily phosphorylated (34). In contrast, little consensus exists as to the effect of PKG; in situ experiments in cerebellar slices reported Ser-1589 to be preferentially phosphorylated by PKG (35), whereas a study by Lincoln and colleagues (36) indicated that purified InsP 3 R-1 protein from cerebellum was phosphorylated preferentially on Ser-1755.
Although phosphorylation of the S2 ϩ InsP 3 R-1 was originally reported to result in decreased Ca 2ϩ release, more recent studies have indicated that phosphorylation results in enhanced release (23)(24)(25)(26). The functional effects of phosphorylation at each site have not been evaluated, presumably because of the lack of a suitable null background to perform mutational analysis.
Alternative splicing of the type-1 receptor gene results in the S2 Ϫ variant of the type-1 InsP 3 R where 40 amino acids (amino acids 1693-1732) are excised between the two phosphorylation sites (6,7). This protein is predominantly expressed in peripheral tissues and, interestingly, has been reported to be exclusively phosphorylated on Ser-1589 by PKA (7) and on Ser-1755 by PKG (36). Studies of the functional effects of phosphorylation of the peripheral form have suggested that, in contrast to the neuronal form of the receptor, phosphorylation of the "short form" of the InsP 3 R-1 results in attenuated Ca 2ϩ release (21,22). Thus, the possibility exists that differences in both the sites of phosphorylation and therefore the functional effects of phosphorylation are defined by the particular splice variants expressed in particular tissues.
In this study we have investigated the sites of phosphorylation by PKA and PKG, functionally important for regulation of Ca 2ϩ release in the two major splice variants of the InsP 3 R-1. By expression of mutant InsP 3 R-1 in InsP 3 R null DT-40 cells (42,43), our studies reveal that phosphorylation of Ser-1755 by PKA or PKG results in markedly enhanced Ca 2ϩ release for S2 ϩ InsP 3 R-1. Notably, in S2 Ϫ InsP 3 R-expressing cells, PKG activation does not markedly alter Ca 2ϩ release, whereas PKAmediated phosphorylation of both Ser-1755 and Ser-1589 results in enhanced Ca 2ϩ release. Thus, the expression of particular InsP 3 R splice variants defines the functional consequences of phosphoregulation by cyclic nucleotide-dependent kinases.
Production of Mutations-The S1 Ϫ /S2 ϩ InsP 3 R-1 in the expression plasmid pIRES-GFP (provided by Dr. S. Joseph, Thomas Jefferson University) was digested with the restriction endonuclease SalI. The overhang created by digestion was blunted 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 S2 Ϫ splice variant and potential PKA phosphorylation sites was excised from its backbone in pCDNA 3.1ϩ (provided by Dr. S. Joseph, Thomas Jefferson University) by RsrII and KasI and ligated into the InsP 3 R construct in MXT-1. The potential PKA sites, Ser-1589 and Ser-1755, were mutated, individually in both splice variants and together in the S2 Ϫ splice variant, to alanines using sequential PCR mutagenesis. The outer oligonucleotides 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 IP 3 R-1 backbone at the corresponding sites. The mutations and lack of spurious misincorporations were confirmed by Big Dye fluorescent sequencing. Mutated receptor DNAs were excised from MXT-1 using EcoRI and ligated into the mammalian expression vector pGW (provided by Dr. David Yue, Johns Hopkins University). Orientation was confirmed by using restriction enzyme digestion.
Transfection of DT-40 Cells-DT-40 cells lacking all three InsP 3 receptor subtypes were transfected using electroporation, at 350 V and 950 microfarads. 2 ϫ 10 7 cells were co-transfected with 25 g of the InsP 3 R cDNA, 25 g of the muscarinic type 3 (M3) receptor, and 4 g of the red fluorescent protein plasmid pHcRed1-N1. Cells were incubated with DNA in 500 l of Opti-MEM media on ice for 10 min. The cell/DNA mixture was electroporated, incubated on ice for 30 min, brought up 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. Transfection efficiency was typically ϳ20%. Experiments were performed within 32 h of transfection. 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-2AM, placed on a 15-mm glass coverslip in a low volume perfusion chamber (Warner Instruments), and allowed to adhere for 30 min at room temperature. Cells were perfused continually for 10 min with HEPES-PSS before experimentation to allow Fura-2AM 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. [Ca 2ϩ ] i imaging was performed essentially as previously described (28,29,41) using an inverted epifluorescence Nikon microscope with a 40ϫ oil immersion objective lens (numerical aperture, 1.3). Cells were excited alternately with light at 340 and 380 nm (Ϯ10-nm bandpass filters, Chroma) using a monochrometer (TILL Photonics). 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 35 ms and no binning. 340/380 ratio images were calculated online and stored immediately to hard disk. Only data from cells exhibiting an increase in ratio units of less than 0.2 upon stimulation were used for further analysis.
Flash Photolysis-Transfected cells were simultaneously loaded by incubation with the visible wavelength indicator Fluo-4 and a cellpermeable form of caged inositol trisphosphate (ciIP 3 /PM) for 30 min. ciIP 3 /PM is a homologue of cmIP 3 /PM (45) The 2-and 3-hydroxyls of the inositol ring were protected by an isopropylidene group in ciIP 3 /PM and were protected by a methoxymethylene group in cmIP 3 /PM. Like cmIP 3 / PM, ciIP 3 /PM diffuses across cell membranes and induces internal calcium release upon photo-uncaging. 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 was collected through a 525 Ϯ 25-nm band pass filter and captured using the Till Photonics imaging suite. These traces are displayed as % ⌬F/F 0 , where F is the recorded fluorescence and F 0 is the mean of the initial 10 sequential frames. Photolytic release was performed as previously described (28,29,41) using a pulsed xenon arc lamp (Till Photonics). A high intensity (0.5-to 5-ms duration; 80 J) discharge of UV light (360 Ϯ 7.5 nm) was reflected onto the plane of focus using a DM400 dichroic mirror and Nikon 40ϫ oil immersion objective (numerical aperture, 1.3).
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 represents a normalized -fold increase over control for the treated trial. Two tailed heteroscedastic t tests with p values Ͻ 0.05 were considered to have statistical significance.

PKA Phosphorylation of Ser-1755 in S2 ϩ InsP 3 R-1 Results in
Enhanced Ca 2ϩ Release-A majority of studies have reported that PKA activation results in enhanced InsP 3 -induced Ca 2ϩ release from the neuronal form of the InsP 3 R-1 (23,24,26). To confirm these earlier findings, experiments were performed in DT-40 3ko cells transfected with S2 ϩ InsP 3 R-1. Because to our knowledge there are no reports of G␣ q -coupled receptors expressed in DT-40 cells, initial experiments were performed eliciting [Ca 2ϩ ] i changes by stimulating the endogenous B cell receptor with ␣-IgM antibody. This resulted in somewhat irregular Ca 2ϩ oscillations, which were not reversible when the antibody was removed making paired analysis of any effects of raising cyclic nucleotides difficult to interpret (data not shown). Thus, in all further experiments, DT-40 3ko cells were cotransfected with the M3 receptor and HcRed, to facilitate iden-tification of transfected cells (inset, Fig. 1). Stimulation of M3 receptors with the muscarinic agonist carbachol (CCh), provided a convenient means of stimulating [Ca 2ϩ ] i changes, presumably through G␣ q -induced activation of phospholipase C and production of InsP 3 . Stimulation with a low concentration of CCh (25-50 nM) resulted in an increase in [Ca 2ϩ ] i , which returned to baseline when the agonist was removed. Following 10-min incubation with 20 M forskolin to maximally raise cAMP, the initial peak in [Ca 2ϩ ] i elicited by identical CCh treatment was markedly potentiated ( Fig. 1A and pooled data in Fig. 1D). After washout of forskolin a subsequent stimulation resulted in a [Ca 2ϩ ] i change comparable to the initial exposure. The potentiation was most marked when threshold elevations in [Ca 2ϩ ] i were evoked by the initial exposure to CCh, and therefore only cells in which the initial CCh treatment evoked a ⌬340/380 ratio of Ͻ0.2 ratio units were included for analysis. These data are consistent with previous reports that phosphorylation of the S2 ϩ InsP 3 R results in enhanced Ca 2ϩ release by increasing the sensitivity of the receptor to InsP 3 (23,26). [Ca 2ϩ ] i changes were never evoked by CCh treatment in cells transfected with only M3 cDNA in the absence of InsP 3 R or likewise InsP 3 R with no M3 cDNA. Similarly, cells exhibiting no HcRed fluorescence seldom responded to CCh treatment (black trace, Fig. 1A).
Phosphorylation of both Ser-1589 and Ser-1755 have been reported following PKA activation (34). To ascertain whether one or both of these serine residues is important for phosphoregulation of the S2 ϩ variant of the InsP 3 R-1, individual point mutations were constructed where these serine residues were mutated to alanine (S1589A and S1755A). A similar potentiation of the initial peak of the CCh-induced [Ca 2ϩ ] i elevation was observed following forskolin treatment in cells expressing S1589A S2 ϩ InsP 3 R (Fig. 1, B and D), however, mutation of S1755A resulted in the complete abrogation of any potentiation upon forskolin treatment (Fig. 1, C and D). These data clearly support the assertion that phosphorylation of Ser-1755 is the important event underlying enhanced Ca 2ϩ release through the neuronal InsP 3 R-1 following PKA activation. It follows, therefore, that phosphorylation of Ser-1589 after 20 M forskolin treatment (which could reasonably be expected to result in the maximal generation of cAMP) is either not occurring to a significant extent or, perhaps more likely, because phosphorylation of this site has been demonstrated at this level of stimulation (7,34), is functionally not important in modulating Ca 2ϩ release.
Because the only difference between the experiment in Fig. 1  (A and C) is a conservative point mutation in the InsP 3 R-1, we have assumed that the effects observed on the peak Ca 2ϩ signal following forskolin treatment are predominately the result of alteration of Ca 2ϩ release. It is, however, formally possible that phosphorylation by PKA of other signaling molecules, such as the muscarinic receptor, phospholipase C, or Ca 2ϩ clearance machinery could contribute to the observed response. Experiments were therefore performed utilizing ciIP 3 /PM, a cell-permeable form of caged InsP 3 (45) to more directly induce Ca 2ϩ release in cells transfected with wild-type S2 ϩ InsP 3 R-1. Using a brief flash of UV light (ϳ0.5 ms, indicated by the arrow in Fig.  2) small elevations in [Ca 2ϩ ] i could be evoked. Subsequent exposure to UV light never produced an increase larger than that initially evoked, however longer flashes of UV light (5 ms) could evoke larger peak increases ( Fig. 2A; max uncage, indicated by the large arrowhead). No effect of UV light was observed in cells either not loaded with cage or alternatively not expressing InsP 3 R. In contrast, when cells were incubated with 10 M forskolin for 5 min prior to a second identical exposure to UV light, a marked increase in the initial Ca 2ϩ peak was evoked ( Fig. 2B and pooled data in Fig. 2C). This potentiation of InsP 3 -induced release was of similar magnitude to that seen for CCh-treated cells exposed to forskolin and supports the notion that the predominant effect of forskolin treatment is to regulate Ca 2ϩ release through phosphorylation of InsP 3 R.
PKA showing the normalized -fold increase in initial [Ca 2ϩ ] i peak over control resulting from forskolin treatment of cells expressing S2 ϩ InsP 3 R-1 and serine to alanine mutants. Forskolin treatment resulted in CCh responses in Wild type and S1589A S2 ϩ InsP 3 R-1 cells being significantly different from S1755A S2 ϩ InsP 3 R-1-expressing cells. The schematic inset depicts the S2 ϩ InsP 3 R-1 regulatory and coupling domain. The black-shaded region represents the S2 splice region. The functionally important phosphorylation of Ser-1755 is indicated by a red circle.
of InsP 3 R-1 is predominantly expressed in peripheral tissues and in fetal brain during neuronal development (6,7). The S2 Ϫ InsP 3 R-1 has been reported to be phosphorylated by PKA in a number of tissues, including platelets, vas deferens, smooth muscle, and hepatocytes (7,30,46). In contrast to the neuronal form of the InsP 3 R-1, studies performed with S2 Ϫ InsP 3 R-1 purified from vas deferens and smooth muscle have demonstrated that the receptor is almost exclusively phosphorylated on Ser-1589 (7). Reports have also suggested a different functional outcome as a result of PKA activation, because the majority of studies, for example in megakaryocytes, have indicated that phosphorylation results in inhibition of Ca 2ϩ release (22). An important caveat relevant to the interpretation of this functional data is that, in contrast to InsP 3 R-1 in cerebellum (Ͼ99% S2 ϩ InsP 3 R-1), in peripheral tissues multiple InsP 3 R types are invariably expressed to varying degrees (47). Thus, unequivocally attributing an effect to an individual homotetrameric receptor may be problematic.
In the next series of experiments we therefore employed similar experimental paradigms to assess the effect of PKAinduced phosphorylation on a homogeneous population of S2 Ϫ InsP 3 R-1. In DT-40 3ko cells expressing wild-type S2 Ϫ InsP 3 R-1, incubation with 20 M forskolin for 10 min resulted in a marked potentiation of the initial CCh-induced Ca 2ϩ peak (Fig. 3A, pooled data in Fig. 3E) in a similar fashion to that demonstrated for the S2 ϩ InsP 3 R-1. In contrast to the S2 ϩ variant of the InsP 3 R-1, neither single mutation in amino acids corresponding to Ser-1589 or Ser-1755 in S2 ϩ InsP 3 R-1 resulted in a loss of this enhanced Ca 2ϩ signal (Fig. 3, C-E).
These data indicate that both serine residues can be phosphorylated in this form of the receptor and furthermore is consistent with the observation that the peripheral InsP 3 R-1 is more readily phosphorylated by PKA than the neuronal form (7). However, no potentiation was observed in cells transfected with a double mutant where both serines were mutated to alanine (S1589A/S1755A S2 Ϫ InsP 3 R-1), confirming that no additional functionally important phosphorylation sites are present in the S2 Ϫ InsP 3 R-1 (Fig. 3, D and E).
In these experiments the degree of potentiation appeared similar when comparing wild-type to either S1589A or S1755A S2 Ϫ InsP 3 R-1; expression of each construct revealed an increase of ϳ3-fold in CCh-induced Ca 2ϩ release in the presence of forskolin. These data suggest that, if the assumption is made that phosphorylation of each particular site occurs independently, phosphorylation of individual sites appears not to result in an additive effect on Ca 2ϩ release. A potential exists, however, that the dye used in these experiments (Fura-2; k d ϳ 150 A similar experimental paradigm as described in Fig. 1 was utilized to assess the consequences and functionally important phosphorylation sites in S2 Ϫ InsP 3 R-1. A, treatment of wild type S2 Ϫ InsP 3 R-1 with forskolin resulted in enhanced CCh-stimulated Ca 2ϩ release with respect to a control CCh stimulation. B, a similar potentiation was observed with S1589A S2 Ϫ InsP 3 R-1-expressing cells. C, a similar enhanced [Ca 2ϩ ] i signal was observed following forskolin treatment in S1755A S2 Ϫ InsP 3 R-1-expressing cells. D, no effect of forskolin treatment was observed in double mutant S1589A/S1755A S2 Ϫ InsP 3 R-1-expressing cells. E, pooled data for the number of cells indicated for each construct. The filled bars indicate data for the particular construct obtained using the low affinity Ca 2ϩ indicator Fura-2FF. Normalized -fold increase is only significantly altered in the double mutant. The schematic inset depicts the functionally important phosphorylation of Ser-1589 and Ser-1755, indicated by red circles. nM) is saturated with Ca 2ϩ upon CCh exposure in the presence of forskolin, thereby masking any additive effect of phosphorylating both sites. For this reason, similar experiments were performed using the lower affinity Ca 2ϩ indicator Fura-2-FF (k d ϳ 10 M). Although the degree of potentiation was somewhat greater in cells expressing wild-type S2 Ϫ InsP 3 R-1 (ϳ4.5fold), there was no significant difference in the extent of enhancement when comparing wild-type to S1589A-and S1755Aexpressing cells (Fig 3E, filled bars). These data indicate that, although it is possible that Fura-2 measurements may indeed underestimate the degree of PKA-induced enhancement of Ca 2ϩ release resulting from PKA phosphorylation, our data suggest that phosphorylation of individual sites appears not to be functionally additive.
To assess if the phosphorylation of a particular site was favored in S2 Ϫ InsP 3 R-1, we compared the minimal concentration of forskolin sufficient to enhance CCh-induced Ca 2ϩ release in wild-type versus S1589A and S1755A mutants. In cells transfected with wild-type S2 Ϫ InsP 3 R-1, 100 nM, 500 nM, or 1 M forskolin failed to enhance the subsequent CCh-induced Ca 2ϩ release (three experiments, Ͼ5 cells for each condition). Incubation with 5 M forskolin resulted in a 1.26 Ϯ 0.21-fold potentiation of Ca 2ϩ release (n ϭ 6). At this threshold concentration of forskolin a similar degree of potentiation was observed in both S1589A (1.36 Ϯ 25-fold; n ϭ 8) and S1755A (1.37 Ϯ 0.26-fold; n ϭ 7) S2 Ϫ InsP 3 R-1. Thus, functionally, it appears that a particular site is not obviously subject to preferential phosphoregulation by PKA.
These data demonstrate that, in contrast to the S2 ϩ variant, both Ser-1589 and Ser-1755 are functionally important phosphorylation substrates in the S2 Ϫ InsP 3 R. Our data do not provide any definitive explanation to account for these differences; however, it would appear that excision of the 40 amino acids in the S2 Ϫ form of the InsP 3 R-1 either alters the structure of the receptor allowing access to the kinase or perhaps allows the interaction with an accessory protein necessary to confer this effect. The only known structural difference between the splice variants is the insertion of an adenine nucleotide binding site in the S2 Ϫ InsP 3 R-1. It remains to be determined if insertion of this motif per se is an important determinant of the differential susceptibility of the splice variants to phosphorylation by PKA.
Phosphorylation of S2 ϩ InsP 3 R-1 by PKG-Similar experiments were performed to assess the effects on Ca 2ϩ release of phosphorylating InsP 3 R-1 with PKG. Cells transfected with S2 ϩ InsP 3 R-1 together with M3 receptor and HcRed were stimulated with low threshold concentrations of CCh (25-50 nM) followed by a 10-min treatment with 10 M 8-Br-cGMP, a specific activator of PKG (48). Subsequent re-stimulation with an identical concentration of CCh revealed a marked potentiation of the [Ca 2ϩ ] i change. In a similar fashion to PKA activation, this was manifested as an increase in the CCh-induced initial peak ( Fig. 4A and pooled data in Fig. 4E). Moreover, these data are again consistent with phosphorylation increasing the sensitivity of the InsP 3 R to InsP 3 , because sub-threshold increases in [Ca 2ϩ ] i were readily potentiated to substantial increases in [Ca 2ϩ ] i (Fig. 4, A, B, and E).
Next, the effect of phosphorylation of Ser-1589 and Ser-1755 in S2 ϩ InsP 3 R-1 mutants was assessed. In cells transfected with S1589A a similar potentiation by 8-Br-cGMP was observed (Fig. 4, B and E), whereas no enhanced [Ca 2ϩ ] i signal was observed in cells transfected with S1755A (Fig. 4, C and E), suggesting strongly that phosphorylation of Ser-1755 by PKG is the important event underlying this potentiation of Ca 2ϩ signaling. Although PKG and PKA consensus motifs are similar (RRXS), these data are consistent with reports that Ser-1755 is phosphorylated by PKG and furthermore that phosphorylation by PKG is enhanced by the presence of an aromatic amino acid, four amino acids downstream from the phosphory-  Fig. 1 was performed to assess the effects and site(s) of phosphorylation by PKG on S2 ϩ InsP 3 R-1. A, treatment with 10 M 8-Br-cGMP to specifically activate PKG results in a marked potentiation of CCh-evoked Ca 2ϩ release when compared with control CCh stimulation in the absence of PKG activation. B, a similar potentiation of Ca 2ϩ release following PKG activation was observed in cells expressing S1589A S2 ϩ InsP 3 R-1. C, PKG activation does not enhance Ca 2ϩ release by S1755A S2 ϩ InsP 3 R-1. D, PKG does not inhibit CCh-induced Ca 2ϩ release by S1755A S2 ϩ InsP 3 R-1. E, pooled data for the number of cells indicated for each construct. Normalized -fold increase by S1755A S2 ϩ InsP 3 R-1 was significantly different from both wild-type and S1589A S2 ϩ InsP 3 R-1. The schematic inset depicts the functionally important phosphorylation of Ser-1755, indicated by a red circle. latable serine as is the case for Ser-1755 in InsP 3 R-1 (49). Indeed the substantial degree of potentiation may reflect the favorable nature of this site for phosphorylation by PKG. Although phosphorylation of Ser-1755 appears to be responsible for the potentiated signal, our data thus far do not exclude the possibility that the response of wild-type S2 ϩ InsP 3 R-1 is a result of a net phosphorylation of both serines by PKG, with phosphorylation of Ser-1589 actually resulting in inhibited release. Thus, the effect of PKG activation on larger [Ca 2ϩ ] i responses to CCh stimulation was tested in cells transfected with S1589A InsP 3 R-1. Using this paradigm, no effect of 10 M 8-Br-cGMP was observed (Fig. 4D), confirming that either Ser-1589 is not phosphorylated or has no functional consequence in S2 ϩ InsP 3 R-1.
Although the most likely consequence of 8-Br-cGMP treatment is to activate PKG, it is possible that this compound could lead to cAMP accumulation and activation of PKA indirectly by inhibiting cAMP phosphodiesterase (50). Additionally, a somewhat less likely scenario is that 8-Br-cGMP might bind to and activate PKA directly. Experiments were therefore performed to confirm that the potentiation of Ca 2ϩ signaling observed with 8-Br-cGMP treatment was not the result of PKA activation. Cells transfected with S2 ϩ InsP 3 R-1 were preincubated for 30 min with the cell-permeable PKA inhibitor, myristoylated PKI. This treatment completely abolished any potentia-tion of the CCh-induced [Ca 2ϩ ] i elevation after forskolin treatment ( Fig. 5A and pooled data in Fig. 5C). In contrast, similar treatment with PKI did not alter the potentiation induced by 8-Br-cGMP treatment (Fig. 5, B and C); this treatment still resulted in a ϳ20-fold potentiation of the CCh-stimulated [Ca 2ϩ ] i signal (compare Fig. 4A and Fig. 5B). These data clearly indicate that treatment with forskolin and 8-Br-cGMP results in the selective activation of PKA or PKG, respectively.
Phosphorylation of S2 Ϫ InsP 3 R-1 by PKG-PKA or PKG phosphorylation of the S2 Ϫ InsP 3 R-1 in megakaryocytes and smooth muscle cells has been suggested to inhibit Ca 2ϩ release (21,36,37). This observation is difficult to reconcile with our data for phosphorylation of S2 Ϫ InsP 3 R-1 by PKA, because phosphorylation of either Ser-1589 or Ser-1755 resulted in potentiated release. Nevertheless, we next performed similar experiments to elucidate the effect of PKG phosphorylation of S2 Ϫ InsP 3 R-1. Interestingly, no effect of 10 M 8-Br-cGMP treatment was observed in cells transfected with either wildtype ( Fig. 6A and pooled data in Fig. 6D), S1589A S2 Ϫ InsP 3 R-1 (Fig. 6, B and D) or S1755A S2 Ϫ InsP 3 R-1 (Fig. 6, C and D). However, in the same batches of cells transfected with wildtype S2 Ϫ InsP 3 R-1, treatment with forskolin resulted in the expected ϳ4-fold increase in the initial peak (n ϭ 5 cells, data not shown). These data also reinforce the contention that 8-Br-cGMP specifically activates PKG without altering PKA activity; the logic being that if this was not the case, activation of  Fig. 1 was utilized to assess the effects of PKG phosphorylation of S2 Ϫ InsP 3 R-1. A, PKG activation has no effect on CCh-evoked Ca 2ϩ release by wild-type S2 Ϫ InsP 3 R-1. B, similarly no effect was observed in S1589A S2 Ϫ InsP 3 R-1-expressing cells. C, no effect was observed in S1755A S2 Ϫ InsP 3 R-1-expressing cells. D, pooled data for the number of cells indicated for each construct. The schematic depicts absence of phosphorylation by PKG.
PKA by 8-Br-cGMP in S2 Ϫ InsP 3 R-1-expressing cells would be expected to result in data similar to PKA activation shown in Fig. 3.
Although it can not be ruled out, a somewhat unlikely possibility explaining these data is that an accessory protein normally required specifically for PKG phosphorylation of the S2 Ϫ form is absent from the Dt-40 cell line. An alternate explanation is that changes in the tertiary structure of the S2 Ϫ InsP 3 R-1 in relation to the long form may result in the functional differences between the splice variants by now hindering the interaction with an accessory protein that normally interacts with the S2 ϩ InsP 3 R-1. In a similar vein, it is possible that deletion of the 40-amino acid insert between Ser-1589 and Ser-1755 in the short form simply defines access to the phosphorylation sites by PKG by again altering the structure of the protein in relation to the S2 ϩ form. These ideas notwithstanding, our data demonstrate that PKA phosphorylation of either Ser-1589 or Ser-1755 results in enhanced Ca 2ϩ release in the S2 Ϫ variant of the receptor (Fig. 3, A-C). Thus, it seems likely that PKG is simply not capable of directly phosphorylating this receptor. These data, although somewhat surprising, may be consistent with a recent report, which suggests that the inhibitory effect of phosphorylation by PKG on InsP 3 R in smooth muscle are actually the result of phosphorylation of an accessory protein termed IRAG (51,52).
In conclusion, using mutational analysis in a null InsP 3 R background, this study has elucidated the serine residues in InsP 3 R-1 functionally important for modulating Ca 2ϩ release by cyclic nucleotide-dependent protein kinases. An important finding of this study is that, although phosphorylation of either Ser-1589 or Ser-1755 can result in markedly enhanced Ca 2ϩ release, the particular splice variant of InsP 3 R-1 expressed dictates which sites are susceptible to phosphorylation. Potentiation of Ca 2ϩ release through InsP 3 R-1 phosphorylation thus provides a powerful means of enhancing and amplifying Ca 2ϩsignaling events when multiple signaling pathways are activated. Additionally, our evidence suggests that PKG appears not to directly regulate S2 Ϫ InsP 3 R-1. Therefore, this splice variation may also define which kinase is capable of phosphorylating the receptor at these sites and thus the specificity of functional response.