Phosphatidylinositol 4,5-bisphosphate (PIP2) affects profoundly several cardiac ion channels and transporters, and studies of PIP2-sensitive currents in excised patches suggest that PIP2 can be synthesized and broken down within 30 s. To test when, and if, total phosphatidylinositol 4-phosphate (PIP) and PIP2 levels actually change in intact heart, we used a new, nonradioactive HPLC method to quantify anionic phospholipids. Total PIP and PIP2 levels (10–30 μmol/kg wet weight) do not change, or even increase, with activation of Gαq/phospholipase C (PLC)-dependent pathways by carbachol (50 μM), phenylephrine (50 μM), and endothelin-1 (0.3 μM). Adenosine (0.2 mM) and phorbol 12-myristate 13-acetate (1μM) both cause 30% reduction of PIP2 in ventricles, suggesting that diacylglycerol (DAG)-dependent mechanisms negatively regulate cardiac PIP2. PIP2, but not PIP, increases reversibly by 30% during electrical stimulation (2 Hz for 5 min) in guinea pig left atria; the increase is blocked by nickel (2 mM). Both PIP and PIP2 increase within 3 min in hypertonic solutions, roughly in proportion to osmolarity, and similar effects occur in multiple cell lines. Inhibitors of several volume-sensitive signaling mechanisms do not affect these responses, suggesting that PIP2 metabolism might be sensitive to membrane tension, per se.
- phosphatidylinositol 4,5-bisphosphate
- phorbol ester
- cardiac muscle
- G protein-coupled receptors
- phospholipase C
- cell volume
phosphatidylinositol 4,5-bisphosphate (PIP2) is the phospholipid precursor of three second messengers,d-myo-inositol 1,4,5-trisphosphate (IP3), diacylglycerol (DAG), and phosphatidylinositol 3,4,5-trisphosphate (PIP3) (66). At the same time, PIP2 serves other cellular functions. It anchors and modulates the function of numerous cell signaling proteins and cytoskeleton at the cell membrane (11, 17, 42, 65), including at least one transcription factor that is released by phospholipase C (PLC) activation (60). In addition, PIP2 metabolism is coupled to membrane trafficking, including some forms of exo- and endocytosis (7, 46). Finally, PIP2 modulates the function of phospholipases (14), receptor kinases (16, 52), and ion transporters and ion channels (25). Especially, the anchoring/recruitment functions and the modulatory functions of PIP2 beg the question as to how, and if, PIP2might be used as a cell signal. For cardiac physiology, an answer to this question seems especially important at this time, because sarcolemmal mechanisms that affect both cardiac contraction (e.g., Na+/Ca2+ exchange) and contraction frequency [e.g., G protein-coupled inwardly rectifying K+ (GIRK) channels] are strongly PIP2 dependent (27).
The minimum biochemical mechanisms involved in cardiac myocyte PIP2 metabolism (39) are summarized in Fig.1. The dominant pathway of PIP2 synthesis, as in other cells, is probably the sequential phosphorylation in the sarcolemma of phosphatidylinositol (PI) at the 4- and then the 5-positions of inositol (66). As in other cells, PIP2 is hydrolyzed by PLCs to generate IP3 and DAG, or it can be dephosphorylated to PIP and PI. DAG can be phosphorylated to generate phosphatidic acid (PA) by DAG kinases, which may be regulated by translocation to the surface membrane (32, 45), and dephosphorylation of PA can probably be an important source of DAG in heart (for example, see Ref.8). On internal membranes, PA serves as the phospholipid precursor of PI, and sarcoplasmic reticulum is a major site of PI synthesis in heart (70). As in other cells, PI in cardiac myocytes is probably transported to the surface membrane by transfer proteins (PITP) (70) as well as by constitutive membrane trafficking during the recycling of membranes to endosomes (not shown).
It seems improbable, at first, that signaling roles of PIP2might be discovered only now, after the biochemical reactions just described have been studied for many years. In fact, the potential complexities of PIP2-dependent signaling are daunting. First, PIP2 changes may occur in a highly localized fashion. In the extreme case, PIP2 might be generated and metabolized directly at its binding sites, as occurs for GTP in G protein signaling (21). Second, even if PIP2levels change, it may be difficult to differentiate effects of PIP2 changes from effects of its metabolites. Third, there are no reliable pharmacological tools to inhibit individual mechanisms involved in PIP2 metabolism. Fourth, quantitative measurements of PIP2 are more problematic than measurements of IP3. One recent innovation, which overcomes some of these problems, is to monitor the membrane association of green fluorescent protein-coupled PIP2-binding domains (57,62). However, the domains employed bind IP3with much higher affinity than they bind PIP2, so they can be used with equal validity to monitor IP3 changes in cells (30). Also, the application of these probes is limited because they must be expressed in cells.
For cardiac muscle, still another complexity is that phosphatidylinositide signaling appears to be altered by isolation of cardiac myocytes. In intact heart, total PIP2 levels amount to 10–30 μmol/kg wet weight (47, 72), and IP3 levels are in this same range (120–300 pmol/mg protein or 12–30 μmol/kg wet weight; see references listed in Ref. 72). PIP2 and IP3 levels both plummet in response to cell isolation, and they remain depressed in cultured adult myocytes (71, 72). An abundance of studies, in which neonatal and adult myocyte cultures were used, demonstrate increases of inositol phosphates, including IP3, with activation of Gαq-coupled receptor pathways (72). However, basal IP3 levels are so high in intact heart that it is difficult to demonstrate a rise of IP3 with activation of Gαq-coupled receptors. In rat ventricle, a rise of IP3 can be demonstrated with strong α-adrenergic stimulation by phenylephrine (23) but not in response to the physiological agonist norepinephrine (50). In rat atrium, generation of IP3 is not obviously increased during the norepinephrine response, and some results suggest that inositol phosphates are being generated from sources besides PIP2 (71, 73).
In short, it is still not clear whether PIP2 changes significantly during physiologically relevant cell signaling changes in cardiac muscle. In tissue slices prepared from canine atrial muscle, PIP2 levels fall by 20–25% within a few seconds during strong muscarinic receptor stimulation, and they return to baseline within 30 s with continued agonist application, accompanied by a Ca2+-dependent increase of PI synthesis (55). In superfused right ventricular strips from rat, strong α-adrenergic stimulation results in a modest decrease of PIP2 with a smaller rebound in a few minutes; during a rapid train of electrical stimuli, IP3 levels rise by about 30% whereas PIP2 levels do not change (53). In addition to these reports, it has been suggested that PIP2 levels may increase during β-adrenergic stimulation in intact cardiac tissue (13, 33).
Given this paucity of information, it seemed important to reexamine effects of major cardioactive hormones on PIP2 in heart. Also, it seemed important to begin to examine the influence of other physiologically relevant processes on cardiac PIP2. At the same time, however, it seemed very unattractive to perform such studies with conventional isotope flux techniques, especially with the use of intact cardiac tissue, because the measurements require long labeling times and high specific activities that are simply not practical for repeated measurements in such tissue. The alternative, then, is to perform some type of mass measurements of phospholipids that do not require radiolabeling, and the application of a mass spectrometric method is the obvious first choice. To date, however, mass spectrometric methods have not proved useful to sensitively and quantitatively detect the more negatively charged, low-abundance phospholipids, such as PIP2. Therefore, we developed and applied a new HPLC method that is highly quantitative above moderately sensitive detection limits in the range of 30 pmol (47). In brief, the total masses of anionic phospholipids from tissue samples are quantified by suppressed conductivity measurements after deacylation of the phospholipids and separation of the head groups by anion exchange HPLC. Besides examining G protein-coupled hormone responses in intact heart, we have identified substantial effects of contraction frequency, osmolarity, and DAG analogs on cardiac PIP2, and we supplement these measurements with comparative measurements in standard cell lines.
Cardiac myocytes and giant patch recording.
Guinea pigs (250–600 g) were killed by intraperitoneal injection of 10 mg/kg pentobarbital sodium. Hearts were removed after loss of reflexes and perfused in retrograde fashion (2.5 ml · g−1 · min−1) at 37°C with a solution containing 140 mM NaCl, 5 mM KCl, 10 mM HEPES, 0.5 mM Na2HPO4, 1.2 mM CaCl2, 0.5 mM MgCl2, 2 mM taurine, and 15 mM glucose, adjusted to pH 7.4 with NaOH and saturated with oxygen. Cardiac myocyte preparation and storage and Na/Ca2+ exchange current recording were carried out as described previously (10).
Left atria of young (<400 g) guinea pigs were harvested before preparation of ventricular myocytes. The atria were mounted in a water-jacketed (35°C) gas-lift superfusion chamber that recirculates bath solutions rapidly directly past the muscle. The perfusion solution described was used as bath solution, the atria were electrically stimulated at 2× threshold (1 ms), and contractions were recorded isometrically.
Measurements of anionic phospholipids.
Anionic phospholipids were quantified from quick-frozen heart tissue samples and from cell cultures as described previously (47). Briefly, after retrograde perfusion of guinea pig hearts began, the atria were dissected away and the left atria were cut in two roughly equal pieces. The atrial halves were then maintained in two open, temperated (35°C) water baths that recirculated oxygenated perfusion solution directly past the muscles. Before being frozen, atrial halves were held at their edge with fine tweezers and touched to a punctate platinum electrode in the solution path to allow electrical stimulation at 2× threshhold with 1-ms duration. After completion of a protocol, the tissue sample was frozen within 0.3 s in metal clamps (3 × 3× 0.5 cm) that were precooled in liquid nitrogen. Ventricular tissue biopsies of 70–140 mg were quickly cut from the retrograde-perfused guinea pig ventricles (37°C) and frozen within 1 s by being pressed (or smashed) between metal blocks or clamps precooled in liquid nitrogen. In preliminary experiments, we verified that multiple samples could be taken from one heart with very little variability or time-dependent changes. Both PIP and PIP2levels tended to decrease with perfusion time, and in six hearts the average decrease measured in consecutive samples was 15 ± 5% over 45 min. Duplicate phospholipid determinations were made from each ventricular biopsy, and the results were averaged. The atrial samples allowed only one determination per tissue sample.
Phospholipids were extracted from pulverized tissue in 1 ml of cold (less than −20°C) chloroform:methanol (MeOH):10 N HCl (20:40:1). Chloroform (0.3 ml) and water (0.5 ml) were added, phases were separated, and the lower phase was removed and washed with 1 ml of MeOH:2 mM oxalic acid (1:0.9). After being centrifuged at 1,000g for 10 min, the washed chloroform phase was dried under a stream of nitrogen, and phospholipids were deacylated with methylamine. The products were then resuspended in 0.5 ml of water and extracted twice with an equal volume of n-butanol:petroleum ether:ethyl formate (20:4:1) to remove fatty acids. The aqueous phase was then dried in a SpeedVac and resuspended in 33 μl of water (47). The anionic head groups were separated by anion exchange HPLC by using NaOH gradients from 2 to 80 mM in four stages, and head groups were detected by suppressed conductivity measurements. The HPLC system and columns employed were all from Dionex (Sunnyvale, CA). A DX 500 HPLC system, equipped with a GP50 gradient pump and an ASRS-ultra self-regenerating suppressor, was employed with an Ionpac AS-11-HC 2-mm column and an AG11-HC 2-mm guard column. An AS50 autosampler was used for sample injection.
COS, HEK-293, HeLa, and M1 (34) cell cultures were grown to confluence in 10-cm dishes with the use of standard culture medium for each cell type, and phospholipids were extracted for measurement of anionic phospholipids as described previously (47).
Experimental results presented are the averages of three or more tissue samples, in most cases four to seven. All error bars represent standard errors (SE) of the mean. Results are presented as mole percent of total anionic phospholipid (i.e., percent of total anionic phospholipids calculated on a mole basis). This quantification is less variable than wet (frozen) weight, because the weight determinations include variable amounts of water. Statistical significance was determined by Student's unpaired t-test and by Student'st-test for paired observations when comparing results from the same heart.
Determinants of PIP2 in cardiac sarcolemma: impressions from PIP2-sensitive currents.
To understand how cardiac PIP2 is regulated, we must ultimately be able to coordinate measurements of PIP2 in intact cells with the regulation of PIP2-sensitive transport mechanisms and with biochemical studies of PIP2metabolism in isolated membranes. As orientation, we present in Fig.2 typical current measurements in inside-out cardiac sarcolemmal patches that suggest simple kinetic and mechanistic conclusions about the generation and breakdown of PIP2. The activation of Na+/Ca2+exchange current by cytoplasmic ATP, described in Fig. 2, evidently tracks the generation of PIP2, which binds to and activates the exchanger by suppressing an autoinhibitory inactivation process (24, 26). For clarity, the presumed changes of PIP2 are depicted qualitatively below the current records in Fig. 2. The results presented are typical outward exchange currents (i.e., Ca2+ influx exchanger mode; 37°C). As shown in Fig. 2 A, the exchange current was initially activated by substituting 60 mM Cs+ on the cytoplasmic side for 60 mM Na+; 2 mM Ca2+ was present in the pipette. The current decayed by about 80% over a few seconds, which reflects the function of the exchanger's inactivation domain, and the current would remain stable at this low level without application of ATP. With application of 2 mM ATP, the current increased within 30 s to a magnitude that is somewhat greater than the peak current obtained upon application of cytoplasmic Na+, and the current usually remained stable at this activated level for a few minutes after removal of ATP when the cytoplasmic free Ca2+level was as low as it is here (0.5 μM). During that time period, two recombinant PH-domains were applied and removed from the cytoplasmic side at concentrations of 2 μM. The first domain is the GRP1 PH domain that binds PIP3 with much higher affinity than it binds PIP2 (35). It was without significant effect at this concentration. The second domain is the PLC-δ PH domain (15). It inhibited the exchange current within seconds to the baseline steady-state level recorded before application of ATP, and the inhibitory effect of the domain could be washed out quickly, consistent with a relatively low PIP2affinity. The PLC-PH domain has a dissociation constant for PIP2 of about 1 μM . Thus it is reasonable that, at a concentration of 2 μM, the domain can mask the stimulatory effect of ATP for the most part. PIP2 antibodies (31) inhibit the Na+/Ca2+ exchange current to a similar extent as the PLC-PH domain, but the effects of antibodies do not wash out, indicative of much tighter binding to PIP2(data not shown). Also, high concentrations of wortmannin (2 μM) have no evident effect on the stimulation of Na+/Ca2+ exchange current, or K+currents, by ATP in cardiac patches. This is consistent with the idea that the major PI kinase in the patch is a wortmannin-insensitive, type 2 PI 4-kinase (2).
Though the stimulatory effect of ATP can remain quite stable for minutes after removal of ATP when cytoplasmic free Ca2+ is low, the effect can reverse very rapidly during application of a high free cytoplasmic Ca2+ concentration. As shown in Fig.2 B, application of solution with 0.2 mM free Ca2+ for just 15 s could completely reverse the stimulatory effect, and similar effects have been reported for ATP-sensitive K+ (KATP) currents in cardiac patches (18). The immediate fall of current on application of high Ca2+ reflects a direct inhibitory effect of cytoplasmic Ca2+ on the reverse, Ca2+-influx transport mode of the exchanger; however, it was evident after removal of Ca2+ that the exchanger regulatory state had in fact been returned to the pre-ATP level. Thereafter, the current could be activated again to the previous stimulated level by reapplication of ATP, and the protocol could be repeated three to five times in a single patch with a stepwise decrease of the rate of the ATP effect (data not shown). These and previous results are all consistent with Ca2+ activating a PLC in cardiac sarcolemma with rather low Ca2+ affinity (>20 μM) (26). Relevant to biochemical results presented subsequently, we did not find that the actions of cytoplasmic Ca2+ in these experiments were markedly enhanced by activators of Gq-type G proteins [e.g., high concentrations of guanosine 5′-O-(3-thiotriphosphate) (GTPγS)].
Figure 2 C shows one more routine result that is consistent with the involvement of a powerful PLC activity in the effect of high cytoplasmic Ca2+. When experiments were performed in the continuous presence of a high cytoplasmic Ca2+concentration (>15 μM), the ability to evoke an ATP response faded rapidly and eventually was lost, consistent with the cleavage of all PI available to PI kinases in the cardiac sarcolemma. This effect could be demonstrated more clearly at a low cytoplasmic pH (6.6), because low pH favors exchanger inactivation (12, 28) and thereby smaller exchange currents. As shown in Fig. 2 C, application of ATP in the presence of 15 μM free Ca2+ (10 mM EGTA with 6.5 mM total Ca2+ at pH 6.6) stimulated the current only transiently. Over the course of 2 min, the effect of ATP faded, the current decreased to a low level, and thereafter it was impossible to stimulate the current again by application of ATP. Presumably, during the combined application of high Ca2+ and ATP, the membrane became depleted of PI and PIP as they were converted to PIP2 and then to DAG by PLC. We project from these transient responses that the total amount of PI in the membrane that is converted to PIP2 and then DAG is not more than six times greater than the amount of PIP2 present at the peak of the ATP response. From biochemical measurements, we know that total cardiac PI in guinea pig cardiac tissue is about 40 times greater than PIP2, and presumably much of that PI arises from internal membranes (47). The suggestion that the generation of PIP2 in cardiac sarcolemma can be limited by the amount of PI present in the sarcolemma is also supported by routine results in patches from mouse myocytes; application of PI vesicles to patches greatly facilitates and is often required to obtain a stimulatory effect of ATP (data not shown). In summary, routine results from studies with excised cardiac patches suggest that cardiac PIP2 levels may be strongly influenced, on the time scale of seconds to 1 min, by the activities of lipid kinases, PLCs, and PI synthesis and transfer to the surface membrane. Although powerful lipid phosphase activities have been identified in crude cardiac membranes (47), lipid phosphatase activity has not yet been identified with certainty in cardiac sarcolemma.
PIP2 in atrial muscle.
We first tested for changes of cardiac PIP and PIP2 in response to several agonists that are known to activate PLCs in heart via Gαq-coupled pathways, starting with muscarinic stimulation in atrial muscle. The acetylcholine-sensitive K+ (GIRK)-channels in atrial muscle are activated by release of βγ-subunits from Gαi upon acetylcholine binding at M2-muscarinic receptors (43). In heart, as well as other tissues, the activation of GIRKs can be dampened by the parallel activation of Gαq/PLC-coupled pathways (29, 36). Depletion of PIP2 by PLC activity is one possible explanation for this inhibition, as well as for the fade or “desensitization” of acetylcholine responses (36). Figure 3,A–C, summarizes our relevant results for superfused left guinea pig atria, in which we examined effects of strong muscarinic receptor activation by the stable agonist carbachol.
Figure 3 A shows the typical contractile response of guinea pig atrial strips, stimulated at 1 Hz, when a high concentration of carbachol (50 μM) was applied. Contraction force plummeted in the first seconds of carbachol exposure as a result of action potential shortening and thereby a reduced Ca2+ influx via Ca2+ channels (58). Contractions then increased again over the course of 10 min to a level that could match the control level (50 μM). This recovery was thought for some time to reflect desensitization of muscarinic receptors, but it was found later that the action potential remained very brief and spikelike during continued agonist application (58). Pertussis toxin blocks the negative inotropic effect of carbachol, thereby making carbachol a positive inotropic agent. The positive inotropic effect is probably mediated by M3 receptors, as shown for mouse atria (48), and it may involve prostaglandin release from endothelium (64).
Figure 3, B–D, shows our relevant measurements of PIP2 and PIP from paired left atria halves. As shown in Fig. 3 B, application of a high carbachol concentration (50 μM) for 1 min during 1-Hz electrical stimulation resulted in no detectable change of PIP2, whereas PIP tended to increase. As shown in Fig. 3 C, both PIP and PIP2 tended to increase during application of carbachol (50 μM) for 10 min with 1-Hz stimulation. Although these changes were just below the level of significance (P = 0.04), the total phosphoinositide levels (i.e., PIP + PIP2) were significantly increased at 10 min (P < 0.05). In addition to these experiments, we tested whether depletion of PIP2 during carbachol application might be more pronounced in quiescent atria. Again, PIP2 did not change significantly during either short- or long-term application of carbachol. Also, we tested for effects of strong α-adrenergic stimulation on PIP and PIP2 levels in atria. We detected no significant changes of either PIP or PIP2 with application of 50 μM phenylephrine for 10 min (data not shown).
The results just described suggest that PLC activation by Gαq-coupled hormones does not result in significant depletion of global PIP2 in atrium. As shown in Fig.3 D, we also tested whether PIP and PIP2 levels in atrium are affected by the PLC inhibitor U-73122, which blocks some, but not all, Gαq responses coupled to PLCs in heart (4, 9, 41). At a concentration of 20 μM, U-73122 had no significant effect on PIP2 or PIP levels in left guinea pig atria after incubation for 20 min. The lack of effect at this high concentration further suggests that PLC activity may not be a major determinant of basal PIP2 levels in guinea pig atrium.
PIP2 in ventricle.
Figure 4 describes effects of activating four G protein-coupled receptor systems in beating, arterially perfused guinea pig ventricles. As shown in Fig. 4 A, stimulation of α-adrenergic receptors by 50 μM phenylephrine for 10 min had no effect on PIP2 and caused a significant (24%) increase of PIP. Similarly, as shown in Fig. 4 B, stimulation of endothelin receptors by 0.5 μM endothelin-1 for 10 min had no significant effect on PIP2 or PIP, although PIP tended to increase. Adenosine activates multiple receptors in most cardiac tissues and is thought to be a major mediator of protein kinase C (PKC) activation in cardiac ischemia (8). Because adenosine has strong negative chronotropic effects, hearts were paced electrically via punctate platinum electrodes at 2 Hz inserted into the ventricular wall. As shown in Fig. 4 C, application of 0.2 mM adenosine for 10 min caused significant (22%) depletion of PIP2 without a significant change of PIP, and this is the only significant agonist-induced PIP2 depletion that we have identified to date. Figure 4 D shows the responses obtained for strong activation of cardiac β-receptors by 0.5 μM isoproteronol for 10 min, which, on the basis of visual inspection, induced large increases of contraction magnitude and frequency. In contrast to a previous report (33), PIP2 did not change in guinea pig ventricles, whereas PIP increased significantly by 19%. Similar experiments with guinea pig atria generated very similar results (data not shown).
DAG-dependent mechanisms can cause PIP2 depletion in heart.
The results presented up to now demonstrate that stimulation of several PLC-coupled receptor systems does not result in global PIP2depletion, at least not when agonists are applied one at a time. Two explanations are possible. Hormone-activated PLC activity may not be active enough to deplete PIP2 significantly, or else compensatory mechanisms ensure that PIP2 does not decrease in intact cells. A role for compensatory mechanisms is supported by the observation that PIP levels increase with several of the agonists employed, and one simple feedback possibility is that PKC activation by DAG results in the activation of lipid kinases. Evidence for such a mechanism has been described previously in multiple cell types (5, 22, 48). Therefore, we tested for effects of strong PKC activators on PIP2 in cardiac tissues and, for comparison, in cell cultures. As shown in Fig.5 A, 12-min treatment with the phorbol ester, phorbol 12-myristate 13-acetate (PMA; 1 μM), caused depletion, not enhancement, of PIP2 by >30%. Both PIP and PI (not shown) increased on average, but the changes were not significant. In another group of experiments in ventricles, with the use of a lower PMA concentration (0.5 μM), PIP2significantly decreased by 14%, whereas PIP significantly increased by 9% (data not shown).
As shown in Fig. 5, B and C, treatment of different cell lines with PMA caused different effects. Significant PIP2 depletion (35%) occured in M1 cells, but PIP2 increased significantly in HEK-293 cells (0.5 μM PMA). Using 1-oleoyl-2-acetyl-sn-glycerol (OAG; 2 μM) as the DAG surrogate, we also obtained a decrease of PIP2 in M1 cells but an increase of PIP2 in HEK-293 cells (data not shown). It is unlikely that the decrease of PIP2 with these agents could be due to PLC activation, because PLCs are inhibited, in general, by PKC-mediated phosphorylation (59, 75). In the case of M1 cells, PI increased significantly when PIP2decreased (data not shown) in the presence of OAG, consistent with the idea that lipid phosphatases might be activated during the response to PKC activators.
PIP2 is activity dependent in heart.
After testing hormone-coupled PLC-activation, we tested other interventions that might affect cardiac PIP2 by other pathways, starting with changes of electrical pacing. As shown in Fig.6, a decrease of contractile activity caused by decreasing contraction frequency led to significant reduction of total PIP2 in several isolated heart preparations. Figure 6, A–D, shows results for isolated, paired left guinea pig atria, which are completely quiescent in the absence of electrical stimulation. The results in Fig. 6 A are for atrial halves that were initially stimulated at 0.1 Hz for 15 min. Thereafter, one atrial half was quick-frozen, and the second half was stimulated at 2.0 Hz for 5 min before freezing. PIP2 levels were increased by 26% at 2.0 Hz (P < 0.01), whereas PIP levels (not shown) did not significantly change. Figure6 B shows results from a reverse protocol. First, the atrial halves were left unstimulated for 15 min, and then both halves were stimulated at 2.0 Hz for 5 min, one half was quick-frozen, and electrical stimulation to the other half was terminated for 5 min before freezing. PIP2 was decreased by 29% (P < 0.01) in the quiescent atria compared with the electrically stimulated atrial halves. Thus the effect of electrical stimulation is reversible on the time scale of 5 min. Again, PIP did not change significantly (data not shown).
Next, we tested whether the increase of PIP2 with electrical stimulation is a Ca2+-dependent process. In separate sets of experiments, atria were incubated with nifedipine (1 μM), verapamil (1 μM), or ryanodine (5 μM ) for 10 min. Thereafter, electrical stimulation did not result in a significant increase of PIP2, although the average values still increased. As shown in Fig. 6 C, preincubation with 2 mM Ni2+ to block Ca2+ influx by all mechanisms completely abolished the effect of electrical stimulation. Also, we tested whether the effect of electrical stimulation could be enhanced by increasing Ca2+ transients. To do so, we incubated atrial halves with isoproterenol (0.5 μM) with electrical stimulation at 0.1 Hz. Contractions were visibly greatly increased by isoproterenol, and atria remained quiescent in the absence of electrical stimulation. After 10 min, one half was stimulated at 0.1 Hz for an additional 5 min, and the second half was stimulated at 2 Hz for 5 min (Fig. 6 D). PIP2 was significantly decreased by 15%, not increased, at the higher frequency. Similarly, we found that treatment of atria and ventricles with ouabain at a concentration that increases contraction without any increase of resting tension or arrhythmias (0.5 μM) (56) caused significant PIP2 depletion by 13% (data not shown). Thus multiple results suggest that large Ca2+ transients favor PIP2 depletion in heart.
Figure 6, E and F, demonstrates effects of cardiac excitation-contraction coupling activity on PIP2content in guinea pig ventricles and in Xenopus frog ventricles, respectively. For arterially perfused guinea pig hearts, we compared samples from ventricles paced at 2 Hz via a platinum electrode with samples taken 8 min later after stopping pacing and arresting the hearts pharmacologically via combined Ca2+ channel and Na+ channel blockade (0.3 μM verapamil + 3 μM flecanaid) (38). PIP2 in the arrested hearts (Fig. 6 E, 0 Hz) was decreased by 22% (P < 0.05) compared with control hearts. To determine effects of electrical activity in frog ventricles, we prepared frog ventricle halves and maintained them rapidly superfused by the same procedures described for atria. These experiments were performed at room temperature (23°C) with a 30% diluted perfusion solution. The frog ventricle halves were initially stimulated at 0.5 Hz for 10 min, and then one half was stimulated for another 10 min while the other half was left quiescent for 10 min, and the samples were quick-frozen. As shown in Fig.6 F, PIP2 was decreased by 27% in the quiescent ventricle halves compared with those stimulated at 0.5 Hz (P < 0.05).
The increases of PIP2 with contractile activity, described here, are clearly consistent with results of others, showing that PIP2 synthesis in cardiac membranes can be stimulated by the presence of micromolar free Ca2+ concentrations (3). Therefore, we tested in excised patches from guinea pigs for an effect of cytoplasmic Ca2+ on the ATP-dependent stimulation of Na+/Ca2+ exchange current and K+ currents, presumed to reflect PIP2generation. The time course and the extent of current stimulation by ATP were not obviously affected by changing cytoplasmic free Ca2+ over the range from 0 to 5 μM (data not shown). Thus it seems likely that the stimuatory effects of Ca2+ on PIP2 require factors that can be washed out of giant excised patches.
PIP and PIP2 levels are highly sensitive to cell volume changes.
We examined effects of cell volume changes on PIP and PIP2for three reasons. First, cell volume changes affect numerous regulatory processes by changing the concentrations of cell constituents, by mechanically perturbing cytoskeleton, and by affecting membrane tension. Resolution of changes to such a nonspecific intervention may in the long run provide paradigms to probe specific regulatory mechanisms. Second, some of the ion channels and transporters that are affected by cell volume changes are candidates for regulation by PIP2. Both Na+/H+exchange (20) and Na+/Ca2+exchange (74) are PIP2 sensitive, and both are activated by shrinkage. Third, the molecular mechanisms by which volume changes affect ion transport are still not well established.
To test for effects of modest osmolarity changes in the hypotonic range, we compared PIP and PIP2 levels in guinea pig ventricles perfused with a solution containing reduced total NaCl (100 mM) vs. the same solution with 100 mM added sucrose. Six ventricles were perfused first with the hyposmotic solution for 10 min, and then a tissue sample was taken, the ventricle was perfused for 5 min with the sucrose-containing solution, and a second sample was taken. In six other ventricles, the order of the experiment was reversed. Regardless of the order, PIP2 and PIP were increased by about 20% in the solution with higher osmolarity, and Fig.7 A presents the pooled results.
Figure 7 B shows the larger effects of doubling osmolarity from an isoosmotic solution containing 140 mM NaCl to one with 250 mM added sucrose. Multiple tissue samples were taken from each ventricle to test for reversal of the sucrose effects in a single protocol. Both PIP2 and PIP increased by about 50% within 3 min upon perfusion with hypertonic solution, and both decreased by about 50% within 3 min after a return to isotonic solution. With a second application of the hypertonic solution, both PIP and PIP2doubled, demonstrating that the effects of osmolarity are rapidly reversible.
Figure 8 A shows effects of exposing M1, HeLa, COS, and HEK-293 cells to standard culture medium containing an additional 250 mM sucrose for 10 min. The PIP2 and PIP levels were normalized, as a percentage, to the phospholipid amounts in control cells (100%). In Hela and COS cells, PIP2 doubled in the hypertonic solution, whereas PIP increased by about 40%. In M1 cells, the relative increase of PIP was smaller, and in HEK-293 cells, the increase was still larger. As shown in Fig. 8 B for HEK-293 cells, the addition of 120 mM excess NaCl to the culture medium caused a 2.4-fold increase of PIP2 and a 1.6-fold increase of PIP within 10 min. Thus cell shrinkage, not the presence of sugar, is probably the initiating cause of the changes of PIP and PIP2.
Given these large and robust changes of PIP and PIP2, we tested whether inhibitors of various signaling enzymes might affect the responses to hypertonic solution. Each inhibitor was applied to HEK-293 or COS cell cultures for at least 10 min, cells were harvested before and after addition of sucrose, and the sucrose response was compared with and without the inhibitor in at least two dishes of cells. Compounds tested included wortmannin (5 μM) to inhibit type 3 PI 4-kinase activity and D3 lipid kinases, cyclosporin (10 μM) to inhibit calcineurin, genistein (0.2 mM) to inhibit tyrosine kinases, nocodazole (20 μM) to disrupt microtubules, PD-98059 (10 μM) to inhibit p38 kinase, chelerythrin ( 1 μM ) to inhibit PKCs, Y-27632 (3 μM) to inhibit Rho kinases, and β-methyl-cyclodextrin (2 mM) to remove cholesterol. Of the compounds tested, only wortmanin and cyclodextrin had significant effects on basal PIP and PIP levels, namely, a 23 and 30% decrease, respectively. However, the absolute increases of PIP and PIP2, induced by sucrose, were not affected significantly by any inhibitor tested.
In this study we have characterized total PIP and PIP2levels in intact cardiac tissue in relation to several physiologically relevant interventions, including activation of G protein-coupled receptor systems, electrical pacing, and changes of cell volume. The measurements have two definite limitations. First, cardiac tissue contains multiple cell types, although myocytes certainly make up the majority of tissue volume. Second, PIP2-dependent signaling may take place in a localized fashion within cells so that measurements of global phosphoinositides will miss important signaling events entirely. Nevertheless, the measurements provide important clues about factors that may regulate cardiac PIP2. Changes of PIP2 on the order of 30% (Figs. 5 A,6 A, and 7) are sufficiently large to affect some PIP2-sensitive mechanisms. For example, the high PIP2 concentrations of intact heart, compared with isolated myocytes (47), will presumably support KATP channel opening in the presence of relatively high ATP concentrations (61). A 30% increase of PIP2with increase of cardiac frequency, as might occur during vigorous exercise, would presumably cause the opening of a small fraction of KATP channels as a physiological regulatory mechanism, independent of metabolic inhibition. Furthermore, it can be assumed that local changes of PIP2 are larger than the global changes actually measured.
Control of cardiac PIP2.
Cardiac myocytes, like most cells, contain multiple PLC types that are modulated by multiple cell signaling pathways, including not only Gαq-coupled receptors but also tyrosine kinases, probably trafficking mechanisms, and Ca2+ itself (67). Three agents that with good certainty activate Gα q /PLC-coupled pathways—carbachol, phenylephrine, and endothelin-1—do not cause PIP2depletion in guinea pig atria or ventricles. Of course, we might have missed large local changes in these measurements, and depletion may be masked by biochemical mechanisms that favor resynthesis of PIP2 (for example, see Refs. 54 and 63). As already noted in connection with Fig. 2, our failure to resolve Gαq/PLC-mediated PIP2 depletion is consistent with our failure, over several years, to find evidence for Gαq-mediated inhibition of PIP2-dependent currents in cardiac membrane patches. Whether other hormones, such as insulin, might activate more powerful PLC systems, or whether we have failed to properly activate Gαq-coupled PLC's in patches, remains to be seen. Patch recordings described here (Fig. 2) do indeed favor the idea that the Ca2+-dependent activation of one or more PLCs can deplete PIP2 in the cardiac sarcolemma very rapidly, and our results for electrical stimulation in the presence of isoproterenol (Fig. 6 D) confirm that PIP2 can decrease during trains of action potentials when Ca2+ transients are of large magnitude (53). If the patch results are relevant to the intact cells, then the Ca2+ affinity of the PLC involved is so low that it could only sense Ca2+ transients close to Ca2+release sites.
The depleting effect of Ca2+ during trains of action potentials is obviously counteracted normally by a mechanism that increases PIP2 in a Ca2+-dependent fashion in guinea pig heart and frog ventricle (Fig. 6). Because PIP2increases without an increase of PIP, the results are consistent with a primary stimulatory effect of Ca2+ on PIP kinase activity, as has already been proposed from biochemical studies using cardiac membranes (3). This stimulatory effect of Ca2+transients is evidently different from the stimulatory effects of α-adrenergic stimulation and cell shrinkage, in which PIP increases (Fig. 3). It remains to be determined what role, if any, the Ca2+ dependence of PIP2 synthesis might play in regulation of cardiac excitation-contraction coupling.
Effects of DAG analogs.
The suggestion that PKCs might activate the synthesis of PIP and PIP2 in close association with the activation of Gαq/PLC-coupled pathways was made several years ago (5, 22), and our results do not contradict this possibility. One possible site of action is the phosphorylation of PI-transfer proteins by PKCs (68). However, for PMA application in heart and in the M1 kidney cell line, depletion of PIP2, rather than enhancement, can be of substantial magnitude. Possibly, this PIP2-depleting effect supports the inhibition of inward rectifier K+ channels that is often observed with PKC activation (19, 29, 40, 44, 69) and that in some cases is known to involve channel phosphorylation (40). Because both PIP and PI tend to increase when PIP2 decreases in the presence of DAG analogs, the results are consistent with an activation of lipid phosphatases. DAG-dependent mechanisms affect profoundly both insertion and retrieval processes at the surface membrane, and in yeast some lipid phosphatases are regulated by trafficking mechanisms (49). Thus we speculate that a DAG-dependent trafficking mechanism may regulate cardiac lipid phosphatases.
In this connection, it seems noteworthy that, from several receptor agonists tested, only adenosine causes significant PIP2depletion in heart (Fig. 4). In cardiac myocytes from most species, adenosine appears to induce long-term activation of several PKC isoforms, and this activation may reflect phosopholipase D activation with generation of DAG by dephoshphorylation of PA (8). Because PA is an activator of the type 1 PIP kinases (1), DAG kinase and PA phosphatases might play key roles in the overall regulation of PIP2. Obviously, much further work is required to test these possibilities.
Volume sensitivity of PIP and PIP2.
In the absence of a rigid cell wall, all cells must regulate their cell volume in the long-term, and numerous regulatory mechanisms are likely involved. However, no pathway is firmly established from sensor to effector at this time. Changes of PIP and PIP2 with changes of cell volume (Figs. 7 and 8) are large, robust, rapidly reversible, and highly reproducible responses of cardiac muscle and several cell lines. Also, in plants, PIP2 increases during hyperosmotic stress (51). The responses are not blocked by inhibiting a number of signaling mechanisms that are known to be changed during cell shrinkage, including the disruption of microtubules. Possibly, therefore, the responses rely on a relatively direct mechanism, and one speculation is that the insertion of membrane that can occur with cell swelling (37) brings important regulatory enzymes to the surface membrane. Because both PIP and PIP2 change with osmolarity changes, insertion of lipid phosphatases in hypotonic solutions, as well as their retrieval in hypertonic solutions, would account for the results. Regardless of the molecular mechanism, our failure to pharmacologically influence the effects of shrinkage are reminiscent of attempts to pharmacologically influence the activation of Na+/H+ exchange activity by cell shrinkage (20). The stimulation of cardiac Na+/Ca2+ exchange by cell shrinkage also does not appear to involve either protein kinases or microtubules (74). It seems reasonable, therefore, to suggest that the shrinkage-induced increase of PIP2 may activate both of these transport systems in hypertonic medium.
In summary, whole tissue measurements of PIP and PIP2demonstrate that Gαq activation favors the synthesis of PIP, and they give no evidence for PIP2 depletion as a normal hormone-mediated response, except in the case of adenosine. Cardiac activity, induced by electrical pacing, usually results in an accumulation of PIP2 by a Ca2+-dependent mechanism that counteracts Ca2+-activated PIP2degradation. DAG-dependent processes can cause substantial PIP2 depletion in heart. The fact that global PIP2 can change by 20–100% with changes of cardiac activity and cell volume suggests that larger changes are taking place locally and that such signals might indeed be used by cardiac cells to regulate PIP2-dependent processes. It seems increasingly important to elucidate how cardiac lipid kinases and phosphatases are regulated and with what consequences for cardiac function.
We thank Drs. Dong M. Kang, Helen Yin, and Joe Albanesi (University of Texas Southwestern) for helpful discussions and encouragement.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-515323 (to D. W. Hilgemann).
Address for reprint requests and other correspondence: D. W. Hilgemann, Dept. of Physiology, Univ. of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75390-9040 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published April 18, 2002;10.1152/ajpcell.00486.2001
- Copyright © 2002 the American Physiological Society