Ca(2+)-activated but not G protein-mediated inositol phosphate responses in rat neonatal cardiomyocytes involve inositol 1,4, 5-trisphosphate generation.

Inositol phosphate (InsP) responses to receptor activation are assumed to involve phospholipase C cleavage of phosphatidylinositol 4,5-bisphosphate to generate Ins(1,4,5)P(3). However, in [(3)H]inositol-labeled rat neonatal cardiomyocytes (NCM) both initial and sustained [(3)H]InsP responses to alpha(1)-adrenergic receptor stimulation with norepinephrine (100 microM) were insensitive to the phosphatidylinositol 4,5-bisphosphate-binding agent neomycin (5 mM). Introduction of 300 microM unlabeled Ins(1,4, 5)P(3) into guanosine 5'-3-O-(thio)triphosphate (GTPgammaS)-stimulated, permeabilized [(3)H]inositol-labeled NCM increased [(3)H]Ins(1,4,5)P(3) slightly but did not significantly reduce levels of its metabolites [(3)H]Ins(1,4)P(2) and [(3)H]Ins(4)P, suggesting that these [(3)H]InsPs are not formed principally from [(3)H]Ins(1,4,5)P(3). In contrast, the calcium ionophore A23187 (10 microM) provoked [(3)H]InsP responses in intact NCM which were sensitive to neomycin, and elevation of free calcium in permeabilized NCM led to [(3)H]InsP responses characterized by marked increases in [(3)H]Ins(1,4,5)P(3) (2.9 +/- 0.2% of total [(3)H]InsPs after 20 min of high Ca(2+) treatment in comparison to 0. 21 +/- 0.05% of total [(3)H]InsPs accumulated after 20 min of GTPgammaS stimulation). These data provide evidence that Ins(1,4, 5)P(3) generation is not a major contributor to G protein-coupled InsP responses in NCM, but that substantial Ins(1,4,5)P(3) generation occurs under conditions of Ca(2+) overload. Thus in NCM, Ca(2+)-induced Ins(1,4,5)P(3) generation has the potential to worsen Ca(2+) overload and thereby aggravate Ca(2+)-induced electrophysiological perturbations.

vation are assumed to involve phospholipase C cleavage of phosphatidylinositol 4,5-bisphosphate to generate Ins(1,4,5)P 3  These data provide evidence that Ins(1,4,5)P 3 generation is not a major contributor to G protein-coupled InsP responses in NCM, but that substantial Ins(1,4,5)P 3 generation occurs under conditions of Ca 2؉ overload. Thus in NCM, Ca 2؉ -induced Ins(1,4,5)P 3 generation has the potential to worsen Ca 2؉ overload and thereby aggravate Ca 2؉ -induced electrophysiological perturbations.
Beat-to-beat regulation of cardiac muscle Ca 2ϩ is primarily controlled by ryanodine receptors on the sarcoplasmic reticulum. Ins(1,4,5)P 3 -mediated Ca 2ϩ responses in cardiac myocytes are relatively slow and weak and have been characterized as Ca 2ϩ oscillations (7,8) which have potential proarrhythmic activity. Intracellular application of Ins(1,4,5)P 3 causes action potential prolongation and degeneration (9) including activation of proarrhythmic sodium-calcium exchange (10). Furthermore, arrhythmias occurring during ischemia and reperfusion correlate with cardiac Ins(1,4,5)P 3 content (11)(12)(13)(14)(15)(16)(17). In the face of evidence for a pathological role of Ins(1,4,5)P 3 in the myocardium (reviewed in Ref. 18), it is important to better understand the mechanisms responsible for its generation in cardiac myocytes and also the mechanisms suppressing its generation under physiological circumstances.
Rat neonatal cardiomyocytes (NCM) represent a convenient system for the modelling of myocardial signaling. InsP responses to norepinephrine (NE) in myocytes are mediated via ␣ 1 -adrenergic receptors, coupling via G q to PLC-␤ isoforms (19,20) of which PLC-␤1 and PLC-␤3 are expressed in heart (21)(22)(23). Our initial studies indicated that NE stimulation in NCM was largely insensitive to the PtdIns(4,5)P 2 -binding agent neomycin (24), unlike findings in other cell types, implying that responses are largely independent of Ins(1,4,5)P 3 . This prompted us to examine the source of InsPs in NCM generated in response to ␣ 1 -adrenergic activation. In this study, we present the first direct evidence that G protein activation of NCM stimulates an InsP response that is largely independent of Ins(1,4,5)P 3 generation. In contrast, the InsP response to elevation of intracellular Ca 2ϩ is primarily via Ins(1,4,5)P 3 .

EXPERIMENTAL PROCEDURES
Culture of Neonatal Cardiomyocytes-NCM were prepared from 1-to 3-day-old Sprague-Dawley rat pups, essentially as described previously (25). NCM were isolated by repeated trypsin digestion with gentle mechanical dispersion, pre-plated twice for 30 min each to remove nonmyocytes and left to attach for 18 h in medium 199, 10% fetal calf serum, 0.1 mM bromodeoxyuridine, 50 units/ml penicillin G, and 50 g/ml streptomycin sulfate onto uncoated dishes, at a typical seeding density of 700 cells/mm 2 . Medium was then replaced with a defined serum-free medium consisting of medium 199, 2% (w/v) bovine serum albumin, 10 g/ml human insulin, 10 g/ml bovine apo-transferrin, 0.1 mM bromodeoxyuridine, 50 units/ml penicillin G, and 50 g/ml streptomycin sulfate. Bromodeoxyuridine was omitted after 3 days and cells were labeled with [ 3 H]inositol in defined serum-free medium for 48 h prior to experiments. Wells were washed extensively with nonradioactive medium before use.
Saponin Permeabilization of NCM-Cell permeabilization was achieved by treatment with 50 g/ml saponin in an intracellular-like medium containing an ATP-generating system for 4 min at 37°C. Intracellular medium had the following constituents (in mM): KCl, 110; NaCl, 10; KH 2 PO 4 , 1; MgSO 4 , 5; EGTA, 1; Hepes, 20; ATP, 2; creatine phosphate, 10; with 20 Sigma units/ml creatine phosphokinase, 0.2% (w/v) bovine serum albumin and CaCl 2 to give a free [Ca 2ϩ ] of 100 nM (determined by reference to a Ca 2ϩ -EGTA titration curve), adjusted to pH 7.2 with NaOH. Plates were washed 4 times with intracellular medium and fresh intracellular medium was added before the commencement of [ 3 H]InsP generation protocols.
Generation and Extraction of [ 3 H]InsPs-10 mM LiCl was added 10 min before the addition of an appropriate stimulator to intact NCM in medium 199, or permeabilized NCM in intracellular medium with an ATP-generating system, and maintained during the treatment period in order to inhibit InsP breakdown by inositol polyphosphate-1-phosphatase and inositol monophosphatase (26,27). 1 M Propranolol was also added to intact NCM to block ␤-adrenergic receptors. For stimulation of intact NCM, stock solutions of 10 mM NE in 0.5 M 2-mercaptoethanol (to inhibit catecholamine oxidation) or 10 mM A23187 in dimethyl sulfoxide were prepared freshly on the experimental day and diluted into medium 199 containing LiCl and propranolol just before use. After stimulation, medium was removed from intact NCM and the reaction terminated by the addition of ice-cold 5% trichloroacetic acid, 2.5 mM disodium EDTA, 5 mM sodium phytate (sodium inositol hexakisphosphate). Permeabilized NCM reactions were stopped by the addition of an equal volume of ice-cold 10% trichloroacetic acid, 5 mM disodium-EDTA, 5 mM sodium phytate to the intracellular medium. Wells were scraped, washed once with 5% trichloroacetic acid, 2.5 mM disodium EDTA, 5 mM sodium phytate and the extracts collected. Trichloroacetic acid-insoluble material was removed by low-speed centrifugation, a 1:1 mixture of freon/tri-n-octylamine was added to remove trichloroacetic acid, and the upper aqueous phase containing [ 3 H]InsPs was prepared for high performance liquid chromatography as described previously (28).

Analysis of [ 3 H]InsP Isomers-Separation and quantitation of [ 3 H]
InsP isomers was performed using established anion-exchange high performance liquid chromatography techniques (28,29). Briefly, 3 Hlabeled compounds were eluted with a complex gradient of ammonium phosphate at pH 3.8. Retention times and integrated peak values were obtained via an on-line ␤-counter. Identification of [ 3 H]InsP isomers was made with reference to appropriate commercial standards.
Materials-Fetal calf serum was specially selected for low endotoxin and obtained from the Commonwealth Serum Laboratories, Parkville, Australia. Medium 199, Hepes, L-glutamine, bovine serum albumin, NaHCO 3 , and other materials for the preparation of cell culture solutions and media were cell culture grade, obtained from Sigma and dissolved in highly purified Milli-Q H 2 O. Human insulin was Actrapid TM from Novo Nordisk Pharmaceuticals, and bovine apo-transferrin was from Sigma. GTP␥S and Ins(1,4,5)P 3 were obtained through Sigma and Sapphire Bioscience, Australia. [ 3 H]Ins(1,4,5)P 3 (21.00 Ci/ mmol) was purchased from NEN Life Science Products Inc. and [ 3 H]inositol (18.00 Ci/mmol) was obtained from Amersham Pharmacia Biotech. Other reagents were obtained from Sigma or BDH/AnalaR and were of analytical reagent grade.
Treatment of Data-Differences between treatment groups were assessed by unpaired Student's t test or 1-way ANOVA as appropriate, and accepted as statistically significant at p Ͻ 0.05. Unless otherwise noted, results shown are from representative experiments, which were independently repeated three times.

RESULTS
Time Course of NE and A23187 Stimulation of InsP Generation-NCM were used for [ 3 H]InsP generation experiments at 5 days after initial isolation and radiolabeled with 20 Ci/ml [ 3 H]inositol 48 h before experiments. As it has been suggested that short-and long-term stimulation of PLC have different characteristics, we examined changes in individual [ 3 H]InsP isomers in detail over the first minute of PLC stimulation and compared this to a 20-min stimulation period. After a 10-min pretreatment period with 10 mM LiCl and 1 M propranolol in medium 199, PLC was activated by stimulation of ␣ 1 -adrenergic receptors with 100 M NE, or by raising intracellular Ca 2ϩ by treatment with 10 M A23187, in the continued presence of LiCl and propranolol, for the specified time periods.
In contrast, when NCM were stimulated with A23187 for either 30 s (Fig. 2b) or 20 min (Fig. 2d), 5 mM neomycin caused ϳ50% inhibition of the response. Approximately 50% inhibition by this concentration of neomycin has been reported in a number of different cell types (33,34). Addition of neomycin in the absence of agonist had no effect on unstimulated [ (35), the experiments were repeated using AlF 4 Ϫ which causes an immediate activation of G q , independent of GDP/ GTP exchange (20,36). After 1 min of AlF 4 Ϫ treatment, total [ 3 H]InsP content was not significantly increased compared with control (Fig. 3b), and there was no increase in [ 3 H]Ins(1,4,5)P 3 (Fig. 3a). 20 min of G protein stimulation with GTP␥S led to an approximate doubling of total [ 3 H]InsPs ( Fig. 3d), again without detectable increase in [ 3 H]Ins(1,4,5)P 3 (Fig. 3c).
In contrast to 1 min of G protein activation, when elevated [Ca 2ϩ ] was used to activate [ 3 H]InsP generation for 1 min, total [ 3 H]InsPs were significantly increased (Fig. 3b). This stimulation was characterized by a very marked rise in the level of [ 3 H]Ins(1,4,5)P 3 (Fig. 3a). Although total [ 3 H]InsPs in response to 20 min of 10 M Ca 2ϩ were only slightly increased above those seen with 20 min of 100 nM Ca 2ϩ (Fig. 3d), the rise in [ 3 H]Ins(1,4,5)P 3 was sustained, with the level of [ 3 H]Ins(1,4,5)P 3 at 20 min of 10 M Ca 2ϩ remaining significantly higher than that at 20 min of 100 nM Ca 2ϩ (Fig. 3c). Thus, unlike the sustained effect of GTP␥S, the [ 3 H]InsP response to elevated Ca 2ϩ involved significant increases in [ 3 H]Ins(1,4,5)P 3 .

Blockade of [ 3 H]Ins(1,4,5)P 3 Breakdown during G Proteinactivated [ 3 H]InsP Generation in Permeabilized NCM-
The low sensitivity to neomycin in NE-stimulated NCM (together with demonstrable neomycin inhibition of A23187 responses) suggested that the major proportion of InsPs detected upon G protein activation of phospholipase C do not arise from generation of Ins(1,4,5)P 3 and its subsequent breakdown. In order to investigate this further, and considering that permeabilized NCM experiments showed similar effects of G protein activation and Ca 2ϩ elevation to those observed in intact NCM, we attempted to block metabolism of any [

DISCUSSION
The present study challenges the assumption that InsP responses in NCM depend on the primary generation of Ins(1,4,5)P 3 , followed by its metabolism to generate a range of other InsPs. We provide evidence for generation of Ins(1,4,5)P 3 in response to elevation of intracellular Ca 2ϩ , as shown by the marked increase in Ins(1,4,5)P 3 observed during the response to elevated free Ca 2ϩ in permeabilized NCM and by the sensitivity of A23187 stimulation to the PtdIns(4,5)P 2 -binding aminoglycoside neomycin in intact NCM. In contrast to Ca 2ϩ -activated responses, ␣ 1 -adrenergic receptor stimulation leads to InsP generation which is relatively insensitive to neomycin, and direct G protein stimulation by either GTP␥S or AlF 4 Ϫ initiates InsP responses which are largely independent of Ins(1,4,5)P 3 .
Early studies in non-cardiac cell types suggested that there is a change in the preferred substrate for phospholipase C away from PtdIns(4,5)P 2 as agonist stimulation is maintained, such that Ins(1,4,5)P 3 generation predominates only in the initial phase of the response (5,6). In NCM at both the 30-s and 20-min time points, the total [ 3 H]InsP response to A23187 was inhibited by neomycin, indicating Ins(1,4,5)P 3 generation. In contrast, the response to NE was insensitive to neomycin at Similar insensitivity of NE responses to neomycin has been observed in adult rat cardiomyocytes (38) as well as normoxic preparations of adult rat heart tissue (29) but InsP responses under reperfusion conditions, during which large increases in Ins(1,4,5)P 3 are observed, are effectively blocked by aminoglycosides (13,14,17). Increases in levels of all InsP isomers are inhibited by aminoglycosides under reperfusion conditions, not merely Ins(1,4,5)P 3 . Although neomycin at high concentrations may well have cellular effects unrelated to its PtdIns(4,5)P 2binding capability, the differential sensitivity of the Ca 2ϩ and NE responses to neomycin argues against a nonspecific action in this case. Thus, these data suggest that increased cytosolic Ca 2ϩ is a stimulus for Ins(1,4,5)P 3 generation, whereas ␣ 1adrenergic receptor stimulation of PLC triggers an InsP re-sponse that is largely independent of PtdIns(4,5)P 2 hydrolysis.
As there are several G protein-coupled receptor systems present in NCM that could activate PLC, we chose to examine InsP responses to direct G protein activation in permeabilized NCM to test whether the relative incapacity of ␣ 1 -adrenergic receptor stimulation to induce sustained Ins(1,4,5)P 3 generation extended to all G protein-mediated PLC activities in NCM. Neither short-term (1 min) nor long-term (20 min) G protein stimulation was able to significantly increase [ 3 H]Ins (1,4,5) (29). Direct hydrolysis of PtdIns, which has been reported as an alternative to PtdIns(4,5)P 2 hydrolysis (4, 5, 33) would lead to increases in the levels of Ins(1)P rather than Ins (4) The InsP generation model suggested by these experiments allows for production of sn-1,2-diacylglycerol without concomitant generation of Ins(1,4,5)P 3 in response to G protein PLCmediated stimuli. Activation of conventional and novel PKC isoforms in heart has been reported under a number of conditions, especially in relation to hypertrophic growth (39 -41). Suppression of Ins(1,4,5)P 3 generation under physiological conditions in NCM, as previously suggested for adult rat heart (29), lends support to the view that increases in Ins(1,4,5)P 3 are deleterious to the heart (18). Direct intracellular application of Ins(1,4,5)P 3 causes action potential degeneration (9, 10) and Ins(1,4,5)P 3 generation has been shown to be related to the initiation of arrhythmias under ischemia and reperfusion conditions (11)(12)(13)(14)(15)(16)(17). It is therefore possible that the abbreviated InsP pathway present in cardiac myocytes protects the heart from Ins(1,4,5)P 3 generation and consequent arrhythmias, presumably arising from enhanced Ca 2ϩ release from the cardiac sarcoplasmic reticulum. If this is the case, then it is interesting that Ca 2ϩ overload itself causes Ins(1,4,5)P 3 generation, potentially initiating a positive feedback relationship between Ca 2ϩ and Ins(1,4,5)P 3 which would further aggravate any electrophysiological damage to the heart. An alternative explanation for preferential Ins(1,4)P 2 generation in the heart is that Ins(1,4)P 2 itself has a signaling function. While Ins(1,4)P 2 is generally considered to be an inactive metabolite of Ins(1,4,5)P 3 , there have been a number of studies suggesting that it has a role in cellular growth (42,43) and other metabolic pathways (44,45). Clearly, further investigation would be required to establish any such role in the heart.
The mechanisms underlying the different responses to NE and Ca 2ϩ remain to be determined. Whereas PLC-␤ isoforms respond to G protein activation (46), PLC-␦ isoforms (particularly PLC-␦1) are likely mediators of Ca 2ϩ -activated InsP responses in the absence of other upstream signaling molecules (46,47). Surprisingly, PLC-␦ has somewhat less specificity for PtdIns(4,5)P 2 relative to PtdIns(4)P than does PLC-␤ when studied in vitro using synthetic lipid micelles (2). It is, however, possible that the substrate specificities of the different PLC isoforms might be altered by particular features of the cardiac sarcolemma.
In summary, the present study demonstrates that generation of Ins(1,4,5)P 3 in NCM requires Ca 2ϩ -mediated activation of PLC and that G protein activation of PLC causes an InsP response which is largely independent of Ins(1,4,5)P 3 generation. Further studies will be required to elucidate the molecular events responsible for the selection between Ins(1,4)P 2 and Ins(1,4,5)P 3 generation in NCM.