Phospholipase C-gamma mediates the hydrolysis of phosphatidylinositol, but not of phosphatidylinositol 4,5-bisphoshate, in carbamylcholine-stimulated islets of langerhans.

In pancreatic islets the activation of phospholipase C (PLC) by the muscarinic receptor agonist carbamyolcholine (carbachol) results in the hydrolysis of both phosphatidylinositol 4,5-bisphosphate (PtdInsP(2)) and phosphatidylinositol (PtdIns). Here we tested the hypothesis that PtdIns hydrolysis is mediated by PLCgamma1, which is known to be regulated by activation of tyrosine kinases and PtdIns 3-kinase. PtdIns breakdown was more sensitive than that of PtdInsP(2) to the tyrosine kinase inhibitor, genistein. Conversely, the tyrosine phosphatase inhibitor, vanadate, alone promoted PtdIns hydrolysis and acted non-additively with carbachol. Vanadate did not stimulate PtdInsP(2) breakdown. Carbachol also stimulated a rapid (maximal at 1-2 min) tyrosine phosphorylation of several islet proteins, although not of PLCgamma1 itself. Two structurally unrelated inhibitors of PtdIns 3-kinase, wortmannin and LY294002, more effectively attenuated the hyrolysis of PtdIns compared with PtdInsP(2). Adenovirally mediated overexpression of PLCgamma1 significantly increased carbachol-stimulated PtdIns hydrolysis without affecting that of PtdInsP(2). Conversely overexpression of PLCbeta1 up-regulated the PtdInsP(2), but not PtdIns, response. These results indicate that the hydrolysis of PtdIns and PtdInsP(2) are independently regulated in pancreatic islets and that PLCgamma1 selectively mediates the breakdown of PtdIns. The activation mechanism of PLCgamma involves tyrosine phosphorylation (but not of PLCgamma directly) and PtdIns 3-kinase. Our findings point to a novel bifurcation of signaling pathways downstream of muscarinic receptors and suggest that hydrolysis of PtdIns and PtdInsP(2) might serve different physiological ends.

tion of the two second messengers Ins(1,4,5)P 3 , which mobilizes Ca 2ϩ from intracellular stores, and DAG, which activates protein kinase C (1)(2)(3)(4). These pathways are well described in the regulation of insulin secretion following occupation of m3 and m1 muscarinic receptors on the surface of pancreatic ␤Ϫcells (5)(6)(7)(8). However we have shown that in addition to the classical pathway of PtdInsP 2 hydrolysis, exposure of pancreatic islets to the muscarinic receptor agonist carbachol also promotes the hydrolysis of PtdIns (9,10). The latter predominates quantitatively over, and is a precursor of, PtdInsP 2 . Although hydrolysis of either phosphoinositide species results in generation of DAG, only PtdInsP 2 gives rise to a Ca 2ϩ signal via Ins(1,4,5)P 3 production. Therefore the two hydrolytic events may have different functional consequences. The inositol phosphate, corresponding to Ins(1,4,5)P 3 , that is derived from PtdIns, is Ins(1)P 1 . This has no biological function but can be used to quantify PtdIns hydrolysis, at least under the conditions verified in our previous studies (9).
The PLC family is comprised of four major groups (PLCs ␤, ␥, ␦, and ⑀). Members of the PLC␤ family are activated by Gproteins in response to occupation of GPCRs (3,4). The recently discovered protein kinase C⑀ is a component of ras signaling pathways (11)(12)(13). PLC␦ is the most sensitive of the PLC family to Ca 2ϩ , and a rise in intracellular Ca 2ϩ is thought to underly its activation in vivo (14,15). The PLC␥ family members (PLC␥ 1 and 2) are classically activated downstream of receptor tyrosine kinases, but they also couple to muscarinic and other GPCRs in a manner secondary to stimulation of cytosolic tyrosine kinases (3,4,16,17). Although PLC␥ is a substrate for many tyrosine kinases there are indications, based both on activity assays (18 -20) and experiments using whole cells (21,22), that tyrosine phosphorylation of PLC␥ itself is not always necessary for its activation. Indeed there is recent evidence that phosphatidylinositol 3,4,5-trisphosphate, generated by the enzyme PtdIns 3-kinase, binds to the pleckstrin homology domain of PLC␥ and thereby stimulates its activity (23)(24)(25)(26). The precise mechanism underlying this mode of activation is obscure, but it does not require tyrosine phosphorylation of PLC␥. This PLC is also of interest, because compared with the PLC␤ and ␦ family members, which are relatively selective for PtdInsP 2 , PLC␥ is also capable of efficiently using PtdIns as a substrate, at least in vitro (27).
We have previously provided limited evidence that hydrolysis of PtdIns and PtdInsP 2 are independently regulated and hence potentially mediated by different PLCs. Thus, in pancreatic ␤-cells, a rise in intracellular Ca 2ϩ levels is sufficient to cause breakdown of PtdInsP 2 but not of PtdIns (10). Our current aim was to characterize further the mechanism underlying PtdIns hydrolysis, with particular reference to a potential role for PLC␥. Our results show for the first time that PLC␥ mediates PtdIns hydrolysis in vivo and that the underlying mechanism appears to involve activation of PtdIns 3-kinase and alterations in the tyrosine phosphorylation of islet cell proteins although not of PLC␥ itself.

EXPERIMENTAL PROCEDURES
Materials-The sources of all materials were as previously described (9, 10) except for tissue culture media which was from ICN Biomedicals (Seven Hills, NSW, Australia), radiochemicals from Amersham Pharmacia Biotech, and scintillant from Canberra Packard (Gladesville, NSW, Australia). All other biochemicals and specialized reagents were from Sigma except genistein, daidzein, LY294002, and wortmannin, which were purchased from Biomol (Plymouth Meeting, PA). Immunologicals were obtained from the following sources: anti-PLC␥1 and mixed monoclonal and rabbit antisera, Upstate Biotechnology, Lake Placid, NY; anti-PLC␤1, Santa Cruz Biotechnology, Santa Cruz, CA; RC20 anti-phosphotyrosine and anti-PLC␥1, Transduction Laboratories, San Diego, CA; goat anti-mouse IgG-Sepharose 4B conjugate and protein A-Sepharose 4B conjugate, Zymed Laboratories Inc., South San Francisco, CA.
Islet Isolation and Culture-Pancreatic islets were isolated following ductal infusion of collagenase into the exocrine pancreas, purified on a Histopaque 1077 gradient, and then handpicked under a binocular microscope (9). They were maintained in tissue culture for 48 -72 h in Medium 199 containing 10% fetal calf serum, 14 mM NaHCO 3 , 11.1 mM glucose, 500 IU/ml penicillin, 100 g/ml streptomycin, and 50 g/ml gentamycin. For the inositol phosphate studies, 10 Ci/ml [2-3 H]inositol was present throughout the culture period.
Generation of Adenovirus and Islet Infection-The two PLCs were present in slightly different adenoviral vectors necessitating the use of different control viruses. Rat PLC␥1 (28) was subcloned into pShuttle-CMV and recombined with pAdEasy-1 as described (29). Control virus was generated by recombination of pShuttle-CMV and pAdEasy-1. Virus-encoding rat PLC␤1 (30) was a generous gift from Dr. C. B. Newgard (Dallas, TX). Here the appropriate control virus, not containing any transgene, was based on the pJM17 adenoviral vector (31) and was provided by Dr. L. Carpenter. After islets were prelabeled in tissue culture for 44 h as described above, the [ 3 H]inositol-containing medium was removed and retained. Groups of 400 -600 islets were then incubated in 140 l of inositol-labeled culture media with 5% polyethylene glycol, to which 20 l of concentrated adenovirus stock was added (2-5 ϫ 10 7 plaque-forming units/l stock). 2 Islets were incubated with virus for 1 h at 37°C with gentle agitation every 15 min. Infected islets were washed eight times with fresh media without labeled inositol before two washes in reserved inositol-labeled culture media, and recombinant protein was allowed to express during a further 3-h culture in reserved inositol-labeled culture media.
Inositol Phosphate Studies-Groups of 50 islets were preincubated at 37°C for 15 min in 500 l of a modified KRB medium containing 5 mM NaHCO 3 , 2.8 mM glucose, 1 mM CaCl 2 , 0.5% BSA, and 10 mM Hepes, pH 7.4. Prewarmed stimulating solutions were added as doubly concentrated 500-l stocks in the same medium. Reactions were terminated after 1 min by the addition of 110 l of 100% ice-cold trichloroacetic acid, and the incubation tubes immediately vortexed and placed at 4°C for a minimum of 30 min. Inositol phosphates were extracted and analyzed by HPLC using a Partisphere PAC column (12.5 cm) and ammonium phosphate gradient as previously described (10). Radioactive peaks were quantified with a Radiomatic Flo-1␤ (Series A-100) on-line detector, using Ultima Flo AP scintillant. Counts were internally corrected by monitoring for recovery of the total free inositol in each sample.
Carbachol-stimulated Tyrosine Phosphorylation-Incubations and stimulations were as for inositol phosphate studies except that groups of islets were incubated in modified KRB without BSA. Reactions were terminated by immediate removal of the incubation medium and addition of 50 l of ice-cold lysis buffer (50 mM Tris-HCl, pH 7.6, 5 mM EGTA, 5 mM EDTA, 1% Triton X-100, 2 mM phenylmethylsulfonylflouride, 5 g/ml aprotinin, 2 M pepstatin A, 2 mM benzamidine, and 1 mM sodium orthovanadate) and transferred to ice. Islets were sonicated using a Branson Sonifier 250 (20 pulses at 10% duty cycle; setting 1) and mixed with 50 l of 2ϫ Laemmli sample buffer, boiled for 5 min, and centrifuged for 5 min prior to carrying out SDS polyacrylamide gel electrophoresis on 10% gels. Proteins were then transferred to a nitro-cellulose membrane at 500 mA for 2 h and blocked for 1 h in a solution of 20 mM Tris-HCl, 137 mM NaCl, and 0.1% Tween 20, pH 7.5, containing 1% BSA for 1 h. Tyrosine-phosphorylated proteins were detected using the anti-phosphotyrosine antibody RC-20 for 1 h and visualized directly by enhanced chemiluminescence.
PLC Expression-Islet extracts were prepared from 50 islets by addition of 100 l of 1ϫ Laemmli sample buffer, boiled for 5 min, and centrifuged for 5 min prior to carrying out SDS polyacrylamide gel electrophoresis on 10% gels. The proteins were then transferred to a nitrocellulose membrane at 500 mA for 2 h and blocked for 1 h in a solution of 20 mM Tris-HCl, 137 mM NaCl, and 0.1% Tween 20, pH 7.5, containing 5% milk powder. The membranes were then incubated for 2 h with an anti-PLC antibody before incubating for 1 h with a horseradish peroxidase-conjugated secondary antibody and detection of immunoreactive proteins using the enhanced chemiluminescence system.

RESULTS
In the following experiments breakdown of PtdIns and PtdInsP 2 were assessed as increases in Ins(1)P 1 and Ins(4)P 1 levels, respectively. This approach has been previously verified on the basis of time course and inhibitor studies, as well as flux measurements (9). Changes in steady-state levels of these inositol phosphates at 1 min post-stimulation, as performed here, are not as pronounced as in end point accumulation assays, such as following prolonged stimulation in the presence of LiCl to inhibit inositol phosphate breakdown (30). For technical reasons, however, the rationale underlying measurement of PtdIns hydrolysis as an increase in Ins(1)P 1 does not necessarily hold at longer time points nor in the presence of LiCl, which precluded use of these alternative protocols (9). As a first means of addressing whether different PLCs might be involved in mediating hydrolysis of PtdIns and PtdInsP 2 , we made use of the tyrosine kinase inhibitor, genistein, which has been widely demonstrated to inhibit activation of PLC␥ (16,17). As shown in Fig. 1, genistein reduced carbachol-stimulated PtdIns hydrolysis in a dose-dependent manner, with 70% inhibition occurring at 100 M genistein. Most importantly this inhibition was significantly greater than that seen with an equivalent dose of daidzein, a genistein analogue that does not affect tyrosine kinase activity. However the effects of genistein and daidzein were identical on PtdInsP 2 hydrolysis suggesting that, in contrast to PtdIns hydrolysis, this is not specifically  (1)P 1 and Ins(4)P 1 , respectively. Inositol monophosphate isomers were separated by HPLC using an ammonium phosphate gradient and a Partisphere PAC column. For further details see "Experimental Procedures." Results are presented as percentages of the carbachol-stimulated increment in either Ins(1)P 1 or Ins(4)P 1 and are means Ϯ S.E. of eight to ten individual determinations. As calculated using Student's t test, * denotes p Ͻ 0.05, ** denotes p Ͻ 0.001 versus the corresponding value for carbachol alone, and § denotes p Ͻ 0.05 as indicated.
influenced by tyrosine kinase inhibition.
Although the results above are suggestive of a selective requirement for tyrosine kinases in the PtdIns response, the high nonspecificity of genistein complicated further analysis. Similar nonspecificity has been seen on other parameters when using a range of other tyrosine kinase inhibitors in islet cells (32). We therefore turned to the converse approach of using sodium orthovanadate, a tyrosine phosphatase inhibitor that augments tyrosine phosphorylation of several islet cell proteins (Ref. 33 and not shown). This potently stimulated PtdIns hydrolysis to ϳ75% of the maximal carbachol response (Fig. 2). The combination of carbachol plus vanadate exerted no greater effects than those because of carbachol alone. In marked contrast, vanadate alone caused only a very modest increase in PtdInsP 2 hydrolysis and actually potentiated the response to carbachol. This is probably explained by the ability of vanadate to stimulate PtdInsP 2 hydrolysis in islets secondary to Ca 2ϩ influx (34) whereas we have previously shown that Ca 2ϩ influx is insufficient to promote PtdIns hydrolysis (10). The current experiments therefore suggest that increased tyrosine phos-phorylation is sufficient to activate PtdIns hydrolysis that occurs via a mechanism that is non-additive (and thus potentially similar) to that of carbachol. Taken together the findings of the genistein and vanadate experiments implicate tyrosine phosphorylation as a component of the mechanism by which carbachol activates PtdIns hydrolysis. Because there should be no requirement for activation of tyrosine kinases in activation of PLC␤ and PLC␦, in contrast to PLC␥ (3,4), our results also suggest that hydrolysis of PtdIns and PtdInsP 2 are mediated by different PLCs and that PLC␥ is implicated in the PtdIns response.
We next analyzed the ability of carbachol to increase tyrosine phosphorylation of PLC␥ and other islet cell proteins. Looking first at islet cell lysates, carbachol increased the tyrosine phosphorylation of a number of proteins, including prominent species at ϳ30, 32, 36, 38, 64, and 125 kDa (Fig. 3). These responses were rapid in onset (maximal at 1-2 min) but remained elevated over the 30 min of stimulation. Although these experiments established that carbachol could activate tyrosine kinases, there was no marked increase in phosphorylation in substrates around the 145-kDa region, corresponding to PLC␥. To address this issue directly, PLC␥ was immunoprecipitated from lysates of 600 carbachol-treated and untreated islets. Although PLC␥ was detectable in the immunoprecipitates, its phosphotyrosine content was extremely low under basal conditions and not increased by prior carbachol stimulation (results not shown).
There is now ample evidence that tyrosine phosphorylation of PLC␥ itself is not always necessary for its activation. One alternative activation mechanism is via stimulation of PtdIns 3-kinase activity and subsequent generation of phosphatidylinositol 3,4,5-trisphosphate, which binds to the pleckstrin homology domain of PLC␥ (23)(24)(25)(26). This mechanism was investigated in experiments examining the effects of well documented, structurally diverse, PtdIns 3-kinase inhibitors (35) on carbachol-stimulated phosphoinositide hydrolysis. As shown in Fig.  4, 100 nM wortmannin and 50 M LY294002 inhibited PtdIns hydrolysis by more than 40 and 70%, respectively. Corresponding effects on PtdInsP 2 hydrolysis were much less pronounced, with wortmannin exerting no significant effect and LY294002 inhibiting by less than 30%. Thus PtdIns hydrolysis exhibits a selective requirement for PtdIns 3-kinase, consistent with a role for that enzyme as an upstream regulator of PLC␥.
To examine the involvement of PLC␥ more directly we used

FIG. 2. Effect of inhibition of tyrosine phosphatases on basal and carbachol-stimulated hydrolysis of PtdIns and PtdInsP 2 .
Incubations and experimental analyses were performed as described in the legend to Fig. 1 except that where indicated 1 mM sodium orthovanadate (Van) was added 15 min prior to stimulation with 0.5 mM carbachol (Carb). Hydrolysis of PtdIns and PtdInsP 2 was quantified as Ins(1)P 1 and Ins(4)P 1 , respectively. For further details see "Experimental Procedures." Results are presented as percentages of the carbacholstimulated increment in either Ins(1)P 1 or Ins(4)P 1 and are means Ϯ S.E. of eight to ten individual determinations. As calculated using Student's t test, ** denotes p Ͻ 0.001 versus the corresponding value for carbachol alone, and § denotes p Ͻ 0.001 as indicated.

FIG. 3. Carbachol-stimulated tyrosine phosphorylation in islets.
Cultured islets were incubated for various times with 0.5 mM carbachol in BSA-free KRB. Following SDS polyacrylamide gel electrophoresis and transfer to a nitrocellulose membrane, phosphotyrosine-containing proteins were visualized via incubation with RC20 anti-phosphotyrosine antibody and enhanced chemiluminescence. For further details see "Experimental Procedures." Shown is an immunoblot representative of five separate experiments. MW, molecular mass. recombinant adenovirus to overexpress this enzyme in pancreatic islets, as demonstrated in Fig. 5A. Most importantly this virus caused a 40% increase in carbachol-stimulated PtdIns hydrolysis (Fig. 5B). In marked contrast it did not affect PtdInsP 2 hydrolysis at all. The specificity of these findings were confirmed in control experiments using a PLC␤1 adenovirus previously documented to increase PtdInsP 2 hydrolysis by ϳ50% in pancreatic islets treated for 10 min with carbachol in the presence of LiCl (30). Under the conditions of our assay (1-min stimulation, no LiCl) the agonist-stimulated PtdInsP 2 response was increased to 119.3 Ϯ 4.7% versus 100 Ϯ 2.3% in carbachol-treated, control virus-infected islets (n ϭ 12, p Ͻ 0.005), whereas PtdIns hydrolysis was not significantly affected; 106.5 Ϯ 10.2 versus 100 Ϯ 4.8 (n ϭ 12). Basal responses were not affected by the PLC␤1 virus (not shown). DISCUSSION The major finding of the current study is that carbacholstimulated PtdIns hydrolysis in pancreatic islets is mediated by PLC␥. This conclusion was based partially on the differential sensitivities of the PtdIns and PtdInsP 2 responses to genistein, vanadate, wortmannin, and LY294002, which are consistent with the known mechanisms of activation of PLC␥ (3,4). This was confirmed directly by the evidence that overexpression of PLC␥1 but not PLC␤1 resulted in a selective increase of agonist-stimulated PtdIns breakdown. Although we have not formally assessed the PLC␦ family members in this context, we consider their involvement highly unlikely. As is the case in other cell types (14,15), PLC␦ is known to be activated in ␤-cells as a direct consequence of a rise in intracellular Ca 2ϩ (36). Although this is sufficient to promote PtdInsP 2 hydrolysis in pancreatic islets (37), we have previously shown this was not the case for PtdIns breakdown (10). Our current data are also supported by earlier findings that PLC␥ efficiently utilizes both PtdInsP 2 and PtdIns as substrates in activity assays in vitro, whereas the PLC␤ and ␦ and families show a much higher degree of specificity for PtdInsP 2 (20). However our results are the first to demonstrate directly that PLC␥ acts on PtdIns in a more physiological setting using intact cells. This might seem surprising, but in practice PtdIns hydrolysis would be difficult to detect unless specifically addressed. There is one earlier study of epidermal growth factor-stimulated phosphoinositide turnover in A-431 cells, now known to involve PLC␥, demonstrating a transient rise in inositol trisphosphate but a persistent increase in total inositol monophosphate (38). Although the latter was interpreted as evidence of PtdIns hydrolysis, it is not clear that this increase actually represented Ins(1)P 1 derived from PtdIns. Nevertheless it would now be of interest to re-examine the A-431 cells more closely for direct evidence of PtdIns hydrolysis, using the rigorous analytical criteria we established previously for pancreatic islets (9).
A corollary of our major finding is that PLC␥ appeared not to play a role in carbachol-stimulated PtdInsP 2 hydrolysis in islet cells. This is consistent with studies demonstrating that PtdInsP 2 is targeted in pancreatic islets and ␤-cell lines by PLC␤1, PLC␤3, and PLC␦1 (30,36). However it is clearly inconsistent with a host of earlier data from other cell types demonstrating that PLC␥ does hydrolyze PtdInsP 2 (3,4). The most obvious explanation for this apparent contradiction is tissue differences, especially because the great majority of previous studies were undertaken with transformed cell lines, instead of the terminally differentiated primary tissue used here. An alternative explanation is that under conditions following occupation of GPCRs, where both PLCs are activated, PLC␤ might compete more effectively than PLC␥ for access to PtdInsP 2 . On the other hand receptor tyrosine kinases, which would not be expected to couple to PLC␤, could stimulate PtdInsP 2 hydrolysis through the activation of PLC␥ alone. This proposal would need to be supported by direct evidence but is attractive in that it might help to explain the necessity for PLC␥ activation, which is presumably not to stimulate PtdInsP 2 hydrolysis, because this would be effectively mediated by PLC␤. Another potential explanation is that the substrate preference of PLC␥ might be conditioned by its activation mechanism. For example PtdInsP 2 is likely to be . Isolated islets were cultured as described in the legend to Fig. 1 except that recombinant adenovirus encoding PLC␥1 or control virus (Shuttle) was added for 1 h, 3 h prior to conducting the experiment. A, lysates from islets infected with control adenovirus or PLC␥1 adenovirus were immunoblotted with ant-PLC␥1 antisera. B, following a 15-min incubation, islets were stimulated or not for 1 min with 0.5 mM carbachol (Carb), and PtdIns and PtdInsP 2 were quantified as Ins(1)P 1 and Ins(4)P 1 , respectively. For further details see "Experimental Procedures." Results are presented as percentages of the carbachol-stimulated increment in either Ins(1)P 1 or Ins(4)P 1 and are means Ϯ S.E. of thirteen to fourteen individual determinations. As calculated using Student's t test,* denotes p Ͻ 0.005 versus carbachol control. sequestered in vivo by binding proteins such as profilin, which have much lower affinities for PtdIns (39). In principle, therefore, PtdIns should be easier for PLC␥ to access than PtdInsP 2 , and indeed there is some evidence that tyrosine phosphorylation of PLC␥ is necessary for displacement of profilin (40). Consistent with this, our findings would suggest that PLC␥ does not need to be tyrosine-phosphorylated to hydrolyze Ptd-Ins, although there was at least a partial requirement for activation of an upstream tyrosine kinase. In addition, PtdIns hydrolysis was dependent on the activation of PtdIns 3-kinase. Although a well documented route for stimulation of PLC␥, this has only been addressed previously in the context of PtdInsP 2 hydrolysis (23)(24)(25)(26). Likewise there is very little information on whether other potential mechanisms for activation of PLC␥, such generation of phosphatidic acid (19) or arachidonic acid (20), might be also be involved in regulation of PtdIns hydrolysis.
Increases in protein phosphotyrosine content in response to carbachol or other GPCRs have not been demonstrated previously using pancreatic islets. In one earlier study (43) carbachol increased tyrosine phosphorylation of a single, unidentified protein of 125 kDa in a transformed insulin-secreting cell line, whereas it had been reported previously that 15-min muscarinic stimulation of pancreatic islets was without effect (33). These results are clarified and extended by our findings of an early (maximal at 1-2 min) incorporation of phosphate into tyrosine residues of a number of proteins, including one around 125 kDa. Tyrosine phosphorylation can therefore be added to the other known responses of pancreatic islets to muscarinic receptor occupation. In addition to the hydrolysis of both PtdIns and PtdInsP 2 (5-10), these include activation of protein kinase C (44,45), mobilization of Ca 2ϩ from intracellular stores (8), Na ϩ -dependent membrane depolarization (46), and both gating and subsequent inhibition of voltage-dependent Ca 2ϩ influx (47). Indeed it is possible that tyrosine phosphorylation might lie upstream of some or other of these events.
In conclusion we have mechanistically dissociated the hydrolysis of PtdIns from that of PtdInsP 2 in pancreatic ␤-cells on the basis of requirements for tyrosine phosphorylation, PtdIns 3-kinase, and involvement of different PLC isoforms. The demonstration that PLC␥1 targets PtdIns exclusively, whereas other PLCs acts on PtdInsP 2 , points to a hitherto unsuspected diversification of signaling pathways downstream of muscarinic receptors. This is apparent from the consideration that although PtdInsP 2 breakdown releases the two intracellular messengers Ins(1,4,5)P 3 and DAG, which regulate Ca 2ϩ mobilization and protein kinase C activation, respectively, the only product of PtdIns hydrolysis possessing known biological function is DAG. Thus different protein kinase cascades of differing Ca 2ϩ dependences could be initiated from the hydrolysis of each phospholipid species. Alternatively, a DAG/protein kinase C readout emanating from PtdIns might lend itself to regulation in a different manner than that of the PtdInsP 2 pathway. There is already some evidence for this in that PtdIns hydrolysis seems to predominate over longer stimulation and is therefore perhaps less sensitive to down-regulation than is PtdInsP 2 regulation (10). This might also be consistent with reports of biphasic DAG generation in response to stimulation of a variety of tissues (41,42,48). In addition, hydrolysis of PtdIns, but not of PtdInsP 2 , is directly influenced by the prevailing membrane potential (10) and might therefore serve as a means of integrating signals arising from different external stimuli. Direct experimental support for these proposals will await a better understanding of the mechanisms underlying the two processes, especially of the tyrosine kinases and their substrates that control PtdIns hydrolysis. However this would be more conveniently addressed in a model system that is more amenable to molecular intervention and biochemical analysis than are isolated pancreatic islets.