© 2000 Macmillan Magazines LtdarticlesNATURE CELL BIOLOGY | VOL 2 | AUGUST 2000 | www.nature.com/ncb507Receptor-mediated hydrolysis of plasma membrane messenger PIP2 leads to K+-current desensitization Evgeny Kobrinsky*, Tooraj Mirshahi*†, Hailin Zhang*, Taihao Jin* and Diomedes E. Logothetis*‡Departments of *Physiology & Biophysics and †Medicine, Mount Sinai School of Medicine of the New York University, New York, New York 10029, USA‡e-mail: [email protected] bisphosphate (PIP2) directly regulates functions as diverse as the organization of the cytoskeleton, vesicular transport and ion channel activity. It is not known, however, whether dynamic changes in PIP2 levels have a regulatory role of physiological importance in such functions. Here, we show in both native cardiac cells and heterologous expression systems that receptor-regulated PIP2 hydrolysis results in desensitization of a GTP-binding protein-stimulated potassium current. Two receptor-regulated pathways in the plasma membrane cross-talk at the level of these channels to modulate potassium currents. One pathway signals through the βγβγβγβγ subunits of G proteins, which bind directly to the channel. Gβγβγβγβγ subunits stabilize interactions with PIP2 and lead to persistent channel activation. The second pathway activates phospholipase C (PLC) which hydrolyses PIP2 and limits Gβγβγβγβγ-stimulated activity. Our results provide evidence that PIP2 itself is a receptor-regulated second messenger, downregulation of which accounts for a new form of desensitization.ellular excitability can be retarded by activation of G-pro-tein-gated inwardly rectifying potassium (GIRK) channels(see, for example, refs 1–3), found in atrial and pacemakercardiac cells, central neurons and endocrine cells. In the heart,acetylcholine (ACh) signals through G proteins to regulate theactivity of GIRK channels and slow heart rate. In general, M1 mus-carinic receptors (as well as M3 and M5) typically couple to theGqα/11 family to stimulate effectors such as PLC, whereas M2receptors (as well as M4) couple via the pertussis toxin (PTX)-sen-sitive Gi and Go subunits to stimulate effectors such as GIRK chan-nels (for review see ref. 4). Stimulation of GIRK currents is knownto proceed through M2 muscarinic receptors, causing activation ofPTX-sensitive G proteins that leads to dissociation of the heterot-rimeric complex into α and βγ subunits. In turn, the βγ dimersinteract directly with the GIRK channels to stimulate their activity(see, for example, refs 5–7). In response to a sustained exposure toACh (for example 100 µM), atrial cells exhibit K+ current activa-tion that reaches a peak (p1) within several hundred millisecondsand gradually decreases to a quasi-steady-state (qss) level within 1min (Fig. 1a). This characteristic reduction of K+ current in thecontinuous presence of ACh is independent of receptor desensiti-zation and has been referred to as ‘short-term’ desensitization8. Asecond brief exposure to ACh yields a response with a reducedpeak (p2), indicating lack of recovery from desensitization. Themolecular mechanisms of GIRK current desensitization have notyet been established.ResultsPLC-mediated hydrolysis of PIP2 results in K+ current desensitiza-tion in rat atrial cells. Block of the Gq pathway at the receptor, Gprotein or effector levels in an atrial myocyte greatly attenuated K+current desensitization (Fig. 1). The M1 antagonist pirenzepine (1µM) had no effect on current desensitization (Fig. 1b). In contrast,the M3 antagonist 4-DAMP mustard hydrochloride (4-DAMP; 0.5µM) largely inhibited current desensitization and yieldedresponses of comparable magnitude to that observed between thefirst and second ACh applications (Fig. 1c). This result suggestedthat the ACh-induced desensitization involved M3 and not M1muscarinic receptors in atrial cells. A control peptide, ADRK,included in the pipette solution, did not significantly affect thedesensitization characteristics of the K+ current (Fig. 1d)9. In con-trast, when we included in the patch pipette the peptide QLKK,which has been shown to block Gqα-mediated activation of PLC9,there was a greatly reduced decline in current between peak andquasi-steady-state levels, and the second response to ACh wascomparable in magnitude to the first (Fig. 1e). Moreover, treat-ment of an atrial myocyte with the PLC inhibitor U73122 (1 µM)had similar inhibitory effects on the desensitization process (Fig.1f). Summary data, comparing control conditions versus a block ofPLC with U73122, a block of Gqα with the QLKK peptide, or ablock of M3 receptors by 4-DAMP, showed significantly increasedquasi-steady-state to peak ratios (Fig. 1g). Summary data of theratio of the second to first peak current responses to ACh alsoshowed a significant increase as a result of a block by either the M3antagonist or the blockers of Gqα or PLC (Fig. 1h). These resultssuggest that the K+ current desensitization seen in native atrialmyocytes involves M3 receptor stimulation of a Gq pathway leadingto activation of PLC.Current desensitization through PLC-mediated hydrolysis of PIP2in COS-1 cells. We next turned to heterologous expression systems,where signalling through particular G-protein-coupled pathwayscan be controlled by expressing the proteins involved. We per-formed whole-cell recordings in COS-1 cells co-expressing GIRKchannel subunits and M2 receptor with or without M1 receptor (Fig.2). We used M1 receptor to activate the Gq pathway because in ourhands heterologous expression of this receptor yielded robustresponses. Application of 5 µM ACh on a cell expressing M2 recep-tor alone resulted in a macroscopic current with small changesbetween peak and quasi-steady-state levels (Fig. 2a, left panel). Asecond application of ACh yielded a similar response (Fig. 2a, rightpanel). In contrast, in cells expressing both M1 and M2 receptors,ACh elicited currents that showed a marked decline between peakand quasi-steady-state levels, suggesting K+-current desensitization(Fig. 2b, left panel). A second ACh application yielded a very smallresponse, suggesting that K+ currents had not yet recovered fromdesensitization (Fig. 2b, right panel). Blocking PLC activity withU73122 before the application of ACh largely prevented M1-recep-tor-mediated desensitization of K+ currents (Fig. 2c, left panel).Additionally, in cells treated with U73122 (1 µM) a second applica-tion