Involvement of Ca2+ in Signaling Mechanisms Mediating Muscarinic Inhibition of M Currents in Sympathetic Neurons

Acetylcholine can excite neurons by suppressing M-type (KCNQ) potassium channels. This effect is mediated by M1 muscarinic receptors coupled to the Gq protein. Although PIP2 depletion and PKC activation have been strongly suggested to contribute to muscarinic inhibition of M currents (IM), direct evidence is lacking. We investigated the mechanism involved in muscarinic inhibition of IM with Ca2+ measurement and electrophysiological studies in both neuronal (rat sympathetic neurons) and heterologous (HEK cells expressing KCNQ2/KCNQ3) preparations. We found that muscarinic inhibition of IM was not blocked either by PIP2 or by calphostin C, a PKC inhibitor. We then examined whether muscarinic inhibition of IM uses multiple signaling pathways by blocking both PIP2 depletion and PKC activation. This maneuver, however, did not block muscarinic inhibition of IM. Additionally, muscarinic inhibition of IM was not prevented either by sequestering of G-protein βγ subunits from Gα-transducin or anti-Gβγ antibody or by preventing intracellular trafficking of channel proteins with blebbistatin, a class-II myosin inhibitor. Finally, we re-examined the role of Ca2+ signals in muscarinic inhibition of IM. Ca2+ measurements showed that muscarinic stimulation increased intracellular Ca2+ and was comparable to the Ca2+ mobilizing effect of bradykinin. Accordingly, 20-mM of BAPTA significantly suppressed muscarinic inhibition of IM. In contrast, muscarinic inhibition of IM was completely insensitive to 20-mM EGTA. Taken together, these data suggest a role of Ca2+ signaling in muscarinic modulation of IM. The differential effects of EGTA and BAPTA imply that Ca2+ microdomains or spatially local Ca2+ signals contribute to inhibition of IM. Supplementary Information The online version contains supplementary material available at 10.1007/s10571-022-01303-7.


Introduction
Neural M-type (KCNQ/Kv7) K + channels are low threshold voltage-gated K + channels that play a crucial role in regulating excitability of neurons (Delmas and Brown 2005;Cooper and Jan 2003). Breakdown of M channels by loss of function mutations or pharmacological inhibitors leads to neuronal hyperexcitability (Schroeder et al. 1998;Delmas and Brown 2005). Indeed, linopirdine, an M channel blocker, has acute cognition-enhancing properties in some 1 3 animal experiments (Aiken et al. 1996), whereas M channel openers, such as retigabine, are currently being developed as novel antiepileptic drugs (Cooper and Jan 2003).
M currents (I M ) were originally described in sympathetic neurons (Brown and Adams 1980;Constanti and Brown 1981) and named due to their inhibition by muscarinic acetylcholine receptors. However, the signal transduction mechanism underlying muscarinic inhibition of I M remains uncertain. Despite the fact that muscarinic inhibition of I M was shown to be mediated by G q -protein activation (Caulfield et al. 1994;Haley et al. 2000), typical downstream pathways of G q such as PKC or IP 3 -mediated Ca 2+ signals were not involved in muscarinic inhibition of I M (Beech et al. 1991;Bosma and Hille 1989;Cruzblanca et al. 1998;del Rio et al. 1999). It was shown that neuronal I M is encoded by KCNQ family genes (Wang et al. 1998) and activities of KCNQ channels depend on phosphatidylinositol 4,5-bisphosphate (PIP 2 ) in the plasma membrane Suh et al. 2006;Zhang et al. 2003;Kim et al. 2016). Recent progress elucidated that PIP 2 modulate KCNQ2/KCNQ3 channel opening by interacting synergistically with a minimum of four cytoplasmic domains (Hernandez et al. 2008a;Choveau et al. 2018). These led to a hypothesis that depletion of PIP 2 by muscarinic stimulation is responsible for inhibition of I M (Delmas and Brown 2005;Suh and Hille 2002).
I M is negatively modulated by other G q -coupled neurotransmitter receptors, such as bradykinin B 2 receptors, and purinergic P2Y receptors (Bofill-Cardona et al. 2000;Jones et al. 1995). The mechanism involved in bradykinin-induced inhibition of I M has been particularly well characterized, and it has been shown that calmodulin activation induced by IP 3 -mediated intracellular Ca 2+ signals is responsible for I M inhibition Gamper and Shapiro 2003). At present, it appears to be generally accepted that I M is regulated by three modulatory pathways: PIP 2 depletion for muscarinic inhibition and second messengers, such as Ca 2+ /CaM for bradykinin-induced inhibition (Hernandez et al. 2008b), protein kinase C (Lee et al. 2010;Kosenko et al. 2012), and post-translational modification (Kim et al. 2016;Qi et al. 2014). However, despite many studies investigating the PIP 2 depletion hypothesis for the mechanism of muscarinic inhibition of I M (Ford et al. 2003;Suh and Hille 2002;Suh et al. 2006;Winks et al. 2005;Zhang et al. 2003), direct evidence is still lacking. In addition, since PIP 2 mobility determines the spatial and temporal profiles of PIP 2 depletion, the idea that PIP 2 depletion acts as the M 1 muscarinic signal in neurons requires that PIP 2 diffusion be slow (Cho et al. 2005). Yet, in general, PIP 2 diffuses rapidly in neurons (van Rheenen and Jalink 2002;Cho et al. 2005), making this hypothesis unlikely. The PIP 2 depletion hypothesis was supported by a result demonstrating that increasing PIP 2 concentration of sympathetic neurons by overexpressing the synthetic enzyme phosphoinositide-5 (PI-5)-kinase can block muscarinic inhibition of I M . Kruse and Whitten recently developed a model that describes altering M 1 R surface density and PI-5-kinase activity regulating the excitability of rat SCG neurons (Kruse and Whitten 2021). However, it is not certain whether increased resting PIP 2 levels caused by overexpression of PI-5-kinase can inhibit PIP 2 depletion induced by receptor-mediated activation of PLC (Winks et al. 2005). Given that PIP 2 can modulate a variety of cellular processes, including cortical actin organization, membrane ruffling, vesicle trafficking, and gene expression (Di Paolo and De Camilli 2006), the possibility that normal signaling mechanisms involved in muscarinic inhibition of I M are altered in excess PIP 2 cannot be excluded. To prove whether receptor-mediated inhibition is mediated by PIP 2 depletion, it is important to demonstrate that the inhibition is rescued by applying normal concentrations of exogenous PIP 2 . Indeed, inhibition of GIRK channels by phenylephrine, endothelin, and PGF2α in cardiac myocytes, which were shown to be mediated via PIP 2 depletion, are completely abolished by internally perfused 10-μM PIP 2 (Cho et al. 2005(Cho et al. , 2001. However, muscarinic inhibition of KCNQ channels was only partially reduced even by 500-μM PIP 2 (Robbins et al. 2006), which can be considered inconsistent with the PIP 2 depletion hypothesis. In the present study, we investigated the mechanism involved in muscarinic inhibition of I M by examining various possible signaling components in rat superior cervical ganglion (SCG) sympathetic neurons and HEK cells expressing KCNQ channels.

Ethics Approval
This study was reviewed and carried out in accordance with the Institutional Animal Care and Use Committee (IACUC) at Sungkyunkwan University School of Medicine (SUSM). SUSM is an Association for Assessment and Accreditation of Laboratory Animal Care International accredited facility and abides by the Institute of Laboratory Animal Resources Guide. The animals were maintained in standard environmental conditions (25 ± 2 °C; 12/12-h dark/light cycle), were given ad libitum access to water and food, and were housed under veterinary supervision at the Laboratory Animal Research Center, SUSM. Sprague-Dawley rats were purchased from Orient Bio Inc. (Sungnam, South Korea). The authors understand the ethical principles under which The Journal of Physiology operates and the experiments comply with the animal ethics checklist described in Grundy (Grundy 2015).

3
Cell Culture, Transfection, and SCG Isolation SCG neurons were cultured from 3-to 4-week-old male rats as previously described (Gamper et al. 2004). HEK293 cells were handled as previously described (Cho et al. 2006). Transfections were made using Lipofectamine 2000 reagents (Invitrogen, Carlsbad, CA) and green fluorescent protein (GFP) was used as a reporter. Plasmids encoding human KCNQ2 (GenBank accession number AF110020) and rat KCNQ3 (GenBank accession number AF091247) were kindly provided by Mark Shapiro (University of Texas Health Science Center, San Antonio, TX).

Electrophysiological Recordings
Current measurements were made with the whole-cell patch-clamp technique. Voltage clamp was performed with an EPC-8 amplifier (HEKA Instruments, Lambrecht, Germany). Filtered signals (1-2 kHz) from a patch-clamp amplifier were fed into an AD/DA converter (PCI-MIO-16E-4, National Instruments, Austin, TX), digitized at 5 kHz and stored digitally in later analysis. Electrodes were pulled from borosilicate capillaries (World Precision Instruments, Inc., Sarasota, FL) using a pipette puller (PP-83, Narishige, Tokyo) and positioned precisely with a micromanipulator (MP 225, Sutter Instrument Company, Novato, CA). The pipettes had a resistance of 2-3 MΩ when filled with pipette solution. Data were not corrected for liquid junction potential (−9 mV). For dialysis, we waited > 5 min before starting the experiment. The perfusion system was a homemade 100-µl perfusion chamber through which solution flowed continuously at 5 ml/min. All recordings were carried out at room temperature (22-24 °C).
The normal external solution for HEK293 cell and rat SCG neuron recording was as follows (in mM): 143 NaCl, 5.4 KCl, 5 HEPES, 0.5 NaH 2 PO 4 , 11.1 glucose, 0.5 MgCl 2 , 1.8 CaCl 2 , and pH 7.4 adjusted with NaOH. The pipette solution was as follows (in mM): 126 KMeSO 4 , 14 KCl, 10 HEPES, and 3 MgCl 2 (pH 7.24 adjusted with KOH). 500 nM of TTX (Tocris, St. Louis, MO, Cat. #1069) and 0.1 μM of CdCl 2 were included in external solution to block Na + and Ca 2+ currents. [Ca 2+ ] I , the concentration of intracellular free Ca 2+ , was measured by loading cells with 2-μM Fura 2-acetoxymethyl ester (Fura 2-AM) (Molecular Probes, Eugene, OR) for 30 min in the recording medium. Changes in Ca 2+ were estimated from the ratio of Fura 2 fluorescence (at 500 nm) with alternate excitation at wavelength of 340 and 380 nm, generated by a Lambda DG-4 monochromator (Sutter Instrument Company). Fura 2 ratio data (F340/380), [Ca 2+ ] i , were measured at a frequency of 2 Hz. Intensity of illumination was minimized to reduce bleaching of Fura 2, while maintaining adequate levels of signal. Fura 2 images for digital analysis were generated using a system based on an Olympus IX71 microscope coupled to an image intensifier and CCD camera and were analyzed with MetaFluor software (MDS Analytical Technologies, Sunnyvale, CA).

Statistical Analysis
Results in the text and the figures are presented as mean ± S.E.M. (n = number of cells tested). Statistical analyses were performed using Student's t test. P-values of < 0.05 were considered statistically significant.

Characterization of I M and Muscarinic Modulation of M Channels in SCG Neurons
I M was originally characterized as outward K + currents suppressed by muscarinic stimulation (Brown and Adams 1980) and specific blockers to inhibit I M , such as linoperdine and XE991, were developed later (Aiken et al. 1996;Cooper and Jan 2003). To investigate the mechanism of muscarinic modulation of I M in rat SCG neurons, we first compared the current component inhibited by oxotremorine-M (oxo-M), a muscarinic agonist, and that inhibited by XE991. Here we used the whole-cell patch-clamp mode, a method suitable for preserving intracellular microdomains, investigating localized signaling, and directly modulating the signaling molecules driving muscarinic inhibition of M channels. To prevent the potential "run-down" of I M in the whole-cell configuration, after the pipette ruptured the membrane, we allowed I M to stabilize before tracking muscarinic inhibition.
Step depolarizing pulses were applied in 10-mV steps from a holding potential of −60 mV. In Fig. 1A, representative current traces obtained in the control condition and after applying oxo-M (10 μM) or XE991 (50 μM) are shown. We obtained the current component inhibited by oxo-M (oxo-M-sensitive) and that inhibited by XE991 (XE991 sensitive) by subtracting current traces in the presence of oxo-M or XE991, respectively, from the control current traces (  outward currents at + 20 mV and 10-μM oxo-M inhibits I M by 79.90 ± 13.62% (at + 20 mV, n = 7).
In Fig. 1C, we compared deactivation tail currents (I tail ) recorded upon repolarization (indicated as square boxes in Figs. 1A and B) and plotted the amplitude of I tail against step depolarization voltage. To avoid contamination of capacitive currents and fast components that may not be attributable to I M , we measured the amplitude of I tail by measuring the average current level between 30 and 50 ms after repolarization and subtracting the current level at steady state. The result shows the similarity between I tail of oxo-M-sensitive currents and XE-991-sensitive currents.
Furthermore, at voltage ranges up to −20 mV, I tail of oxo-M-sensitive currents and XE-991-sensitive currents are almost equivalent to I tail of control currents, indicating that I tail corresponding to −20-mV depolarization is mostly attributable to I M . In the present study, therefore, we regarded I tail as an indication of I M .
The concentration-dependent effects of oxo-M to inhibit I M were tested, while changes of I tail were monitored in 5-s intervals by applying a hyperpolarizing pulse to −60 (or −55) mV for 1.5 s from the holding potential of -20 mV. Extent of inhibition induced by 1-μM and 10-μM oxo-M were 42.54 ± 3.51% (n = 12) and 84.13 ± 7.34% (n = 16), respectively (Fig. 2C). Re-examination of the Current Hypothesis: PIP 2 Depletion and PKC Activation PIP 2 depletion and PKC activation have been considered to be the most likely candidates involved in muscarinic inhibition of I M (Delmas and Brown 2005). However, previous studies have found that muscarinic inhibition of I M was not fully blocked by application of exogenous PIP 2 or PKC inhibitors (Brown and Yu 2000;Ford et al. 2003;Hoshi et al. 2003;Robbins et al. 2006;Shapiro et al. 2000). We confirmed that exogenous PIP 2 (up to 200 μM) had little effect on muscarinic inhibition of I M (Fig. 2A). Consistent with the data from sympathetic neurons, in HEK293 cells expressing KCNQ2/KCNQ3 channels and M1 muscarinic receptors, application of PIP 2 did not block the effect of oxo-M on I M (Yoon 2010). To apply PIP 2 into cells, diC8-PIP 2 was added to the pipette solution, and the conventional whole-cell mode was accomplished by rupturing membranes after a giga-seal was made. We have recently shown that application of diC8-PIP 2 disrupts M channel regulation induced by the altered channel's affinity for PIP 2 (Lee et al. 2010), suggesting that exogenous PIP 2 is effectively delivered to the plasma membrane and available to M channels. We also confirmed that calphostin C, a highly specific PKC inhibitor, did not affect muscarinic inhibition of I M (Fig. 2B). Summarized data are shown in Fig. 2C. In addition, a higher dose of calphostin C (up to 10 μM) and two other PKC inhibitors, 1-μM chelerythrine and 100-nM bisindolylmaleimide I, were tested and neither affected oxo-M-induced M current modulation (SI 1).
We then hypothesized that muscarinic inhibition of I M uses multiple signaling pathways such that inhibition can only be blocked when multiple pathways are simultaneously suppressed. To test this hypothesis, both PIP 2 depletion and PKC activation were prevented with exogenous PIP 2 (20 μM) and calphostin C (1 μM). In the presence of 20-μM PIP 2 and calphostin C, however, 10-μM oxo-M was still capable of inhibiting I M ( Fig. 2B and C).

Looking for New Possibilities
Since we did not find evidence that messengers downstream of the Gα q pathway mediate muscarinic inhibition of I M , we searched for other possibilities. It is well known that the G βγ dimer mediates voltage-dependent modulation of N-type Ca 2+ channels through numerous neurotransmitters, including noradrenaline, acetylcholine, and dynorphin (Herlitze et al. 1996;Ikeda 1996;Tsien et al. 1988). We tested the involvement of the G βγ subunit in oxo-M-induced I M modulation using HEK cells transfected with G α -transducin (G α TD) together with KCNQ2, KCNQ3, and M 1 muscarinic receptors. G α TD is a G α subunit which predominantly attains the GDP-bound state when overexpressed in SCG neurons and buffers G βγ (Kammermeier and Ikeda 1999). As shown in Fig. 3A, expression of G α TD did not affect muscarinic inhibition of the KCNQ current by 1-μM oxo-M (87.84 ± 4.02% (n= 4) in controls or 89.77 ± 7.74% (n = 5) in G α TD-expressed cells, p > 0.05). In addition, scavenging G βγ subunits with intracellular anti-G βγ antibody in SCG neurons was also tested. In the presence of anti-G βγ antibody, 10-μM oxo-M-induced inhibition of M current was not different from control conditions. (Fig. 3B). These data implicate that G βγ subunits do not have a critical role in muscarinic inhibition of I M .
Recently, it was reported that receptor activation regulates channel activity by affecting the surface expression of channel proteins (Chung et al. 2009). Because KCNQ channels are known to be targets of trafficking, i.e., insertion into and removal from the plasma membrane (Etxeberria et al. 2008;Schuetz et al. 2008), we tested the possibility that KCNQ channel trafficking is involved in muscarinic inhibition of I M . However, a block of protein trafficking with 50-μM blebbistatin, a cell-permeable inhibitor of class-II myosins, did not affect muscarinic inhibition of I M (SI 2). In addition, pretreatment of a protein trafficking inhibitory cocktail including chloroquin (100 μM), monensin (100 μM), and nocodazole (20 μM) had no effect on muscarinic modulation of I M (data not shown), suggesting that KCNQ channel trafficking did not contribute to the regulation of I M by M 1 muscarinic receptors.

Revisiting Ca 2+ -Dependent Hypothesis
Recently, the role of Ca 2+ signaling in I M inhibition was well established for I M inhibition by bradykinin Gamper and Shapiro 2003) or by P2Y receptor stimulation (Zaika et al. 2007), in which CaM was shown to be responsible. However, inhibition of CaM activation using dominant-negative CaM that cannot bind Ca 2+ (Geiser et al. 1991) did not affect muscarinic inhibition of I M (Gamper and Shapiro 2003;Zaika et al. 2007), leading to the conclusion that Ca 2+ signaling is not involved in muscarinic inhibition of I M . This conclusion was supported by the observation that in SCG neurons, oxo-M did not induce Ca 2+ transients as robustly as bradykinin did (Cruzblanca et al. 1998;del Rio et al. 1999). In early experiments, however, muscarinic inhibition of I M was shown to be suppressed when intracellular Ca 2+ was heavily buffered (Beech et al. 1991). Further, Shapiro et al. (Shapiro et al. 2000) reported that oxo-Minduced Ca 2+ transients are blocked by thapsigargin. Taken together, these findings implicate Ca 2+ signaling in muscarinic receptor stimulation. Thus, we re-examined the roles of Ca 2+ signaling in muscarinic inhibition of I M inhibition.
First, we examined whether oxo-M could induce Ca 2+ transients in SCG neurons and whether oxo-M-induced Ca 2+ transients would be significantly smaller than 1 3 bradykinin-induced Ca 2+ transients, as previously described (Cruzblanca et al. 1998;del Rio et al. 1999). [Ca 2+ ] i was measured in SCG neurons using Fura 2-AM. As shown in Fig. 4, 10-μM oxo-M did induce Ca 2+ transients in all SCG neurons tested though there was large cell-to-cell variation. On average, oxo-M-induced Ca 2+ transients were not different from bradykinin-induced Ca 2+ transients (Fig. 4). The peak amplitude of Ca 2+ transient induced by oxo-M was similar to that induced by bradykinin. Paired t test confirmed that the effects of oxo-M on intracellular Ca 2+ concentration were not significantly different from those of bradykinin in the same neurons (n = 3, p > 0.05). These results imply that the Ca 2+ -releasing powers of oxo-M and bradykinin are not significantly different.

3
Since muscarinic receptor stimulation elicited an increase of [Ca 2+ ] i , we tested whether Ca 2+ mediates muscarinic modulation of I M using 20-mM BAPTA. When the pipette solution contained 20-mM BAPTA, oxo-M minimally affected I M (Fig. 5Aa). Superimposed deactivation tail currents showed that in a cell dialyzed with 20-mM BAPTA, I M after 5 min oxo-M treatment was almost overlapped with that in controls. To rule out the possibility that I M itself was reduced by 20-mM BAPTA, we obtained the steady-state current-voltage relationships for oxo-M-sensitive currents and XE-991-sensitive currents in BAPTA-loaded cells, using the same protocol as in Fig. 1. Figure 5Ac demonstrates the proportion of XE991-sensitive currents to total steady-state currents in 20-mM BAPTA-loaded cells was not reduced compared to that in 0.1-mM EGTA (52.40 ± 2.46% [n = 19] vs. control: 26.49 ± 1.41% [n = 7] at + 20 mV). The oxo-M-induced inhibition of I M was determined as a proportion of oxo-M-sensitive currents to XE991-sensitive currents at + 20 mV as in Fig. 1 and was 22.21 ± 3.94% (n = 19, Fig. 5Ac), which was significantly smaller than that in controls (79.90 ± 13.62% (n = 7), p < 0.01). We then tested whether the effect of BAPTA on suppressing muscarinic inhibition of I M could be attributed to inhibiting local Ca 2+ signals. To do this, we tested the effect of EGTA on muscarinic inhibition of I M . BAPTA is a fast Ca 2+ buffer that effectively buffers all types of Ca 2+ signals, whereas EGTA is a slow Ca 2+ buffer that selectively spares local Ca 2+ signals (Neher 1998). With 20-mM EGTA, muscarinic receptors were still able to inhibit M currents (Fig. 5Ba). The oxo-M-induced inhibition of I M was calculated as a proportion of oxo-M-sensitive currents to XE991-sensitive currents at + 20 mV and was 103.05 ± 11.60% (n = 5, Fig. 5Bc), which was not significantly different from that of control conditions (p > 0.05). We confirmed that proportions of I M to total outward currents in the presence of 20-mM EGTA (49.60 ± 3.48% [n = 4]) at + 20 mV) were similar to those in the presence of 20-mM BAPTA (52.40 ± 2.46% [n = 19]).
Next, we directly examined the effect of [Ca 2+ ] i on M currents in SCG neurons. We perfused cells with a bathing solution containing 10-μM ionomycin, a Ca 2+ ionophore, and 10-mM CaCl 2 to raise [Ca 2+ ] i . We found that as [Ca 2+ ] i increased, the M current amplitude declined in parallel (SI 3). This is consistent with previous data showing the sensitivity of KCNQ2/KCNQ3 channels to [Ca 2+ ] i (Delmas and Brown 2005;Gamper and Shapiro 2003;Kosenko and Hoshi 2013). Taken together, these data suggest that Ca 2+ signals play a critical role in M 1 muscarinic receptor-mediated regulation of M channels. The differential effects of EGTA and BAPTA imply that M channels are very close to Ca 2+ sources and are regulated by local Ca 2+ signals.

Discussion
We have demonstrated that muscarinic inhibition of I M in rat SCG neurons is significantly suppressed by 20-mM BAPTA in pipette solutions (Fig. 5). In fact, similar findings were observed in earlier studies (Beech et al. 1991;Kirkwood et al. 1991). However, Ca 2+ was not regarded as a candidate for couple muscarinic receptor activation to M channel inhibition, because oxo-M failed to increase the intracellular Ca 2+ concentration in SCG neurons (Beech et al. 1991;del Rio et al. 1999). The current understanding of the distinction between muscarinic receptor-induced inhibition of I M and bradykinin-induced inhibition of I M relies, at least in part, on the inability of oxo-M to induce Ca 2+ increase (Delmas and Brown 2005). However, some controversial reports stated that oxo-M induces Ca 2+ increase in acutely dissociated rat SCG neurons (del Rio et al. 1999;Cruzblanca et al. 1998) and tsA-201 cells . Furthermore, oxo-M was invariably shown to induce Ca 2+ mobilization in other parts of the brain (Irving and Collingridge 1998;Seymour-Laurent and Barish 1995), and oxo-M-induced effects, such as GIRK channel inhibition, were shown to be mediated  (Sohn et al. 2007). Other reports exploring the PIP 2 depletion hypothesis showed that M current could be inhibited by muscarinic receptors in neurons overexpressing the neuronal calcium sensor, NCS-1 (Winks et al. 2005). NCS-1 up-regulates PI-4-kinase, which catalyzes the first step in PIP 2 synthesis from PI and requires Ca 2+ to be activated (Koizumi et al. 2002;Rajebhosale et al. 2003;Zhao et al. 2001;Taverna et al. 2002;Burgoyne and Weiss 2001). In neurons overexpressing NCS-1, Winks et al. reported that M current inhibition by oxo-M was conserved, while the inhibitory effect of bradykinin was reduced (Winks et al. 2005). They claimed that oxo-M does not evoke enough Ca 2+ release to activate NCS-1 for replenishing PIP 2 hydrolyzed by muscarinic stimulation. However, others reported similar basal Ca 2+ levels, ~ 100 nM, in PC12 cells and SCG neurons, which are enough to activate NCS-1 in PC12 cells (Cruzblanca et al. 1998;Meijer et al. 2014;Koizumi et al. 2002;Rajebhosale et al. 2003;Taverna et al. 2002). In this case, under basal conditions during muscarinic receptor activation, NCS-1 could compensate for PIP 2 consumption by PLC. If PIP 2 depletion hypothesis is true, in cells overexpressing NCS-1, oxo-M is unlikely to inhibit M current because NCS-1 replenishes PIP 2 but supporting our hypothesis that PIP 2 depletion is not the primary mechanism.
In the present study, we re-examined whether the inability of oxo-M to induce Ca 2+ mobilization is a consistent finding. Surprisingly, we found that muscarinic receptor stimulation by oxo-M is capable of increasing intracellular Ca 2+ concentration in rat sympathetic neurons. We noted that the cell-to-cell variation of the amplitude of Ca 2+ transients was quite large; thus, in some cells, Ca 2+ transients induced by oxo-M were small, which was similar to a previous report . However, this was also the case for B2R-induced Ca 2+ transients (Fig. 4). Thus, we conclude that the current understanding of dual modulatory pathways for M channel regulation based on different Ca 2+ mobilization powers, the ability to evoke changes of [Ca 2+ ] i , needs to be re-examined. We do not know the reason for the discrepancy of Ca 2+ mobilization power of oxo-M among different studies, but differences in receptor density Falkenburger et al. 2013;Kruse and Whitten 2021), signaling microdomains that modify the efficacy of IP 3 to open its receptor (Zaika et al. 2011), or experimental conditions may be involved. Previous studies used voltageclamped neurons with Fura 2 in pipette solutions and held the membrane potential near or below −60 mV during Ca 2+ measurements (Beech et al. 1991;del Rio et al. 1999;Falkenburger et al. 2013;Dickson et al. 2013), whereas we used intact sympathetic neurons with Fura 2-AM. If Ca 2+ entry through the voltage-dependent ion channels contributed to the oxo-M-induced [Ca 2+ ] i rises, this pathway might have been blocked by clamping the membrane potential hyperpolarized to −60 mV or below. The fact that removal of external Ca 2+ prevented the rise of [Ca 2+ ] i by muscarinic receptor stimulation may support this possibility (Foucart et al. 1995).
In fact, it was also observed in earlier studies that high concentrations of Ca 2+ chelator (20-mM BAPTA) can suppress muscarinic inhibition of I M (Beech et al. 1991;Kirkwood et al. 1991). However, when 10-mM Ca 2+ was added to 20-mM BAPTA-containing pipette solutions to increase free Ca 2+ concentration from 12 to 143 nM, oxo-M was still capable of inhibiting I M (Beech et al. 1991;Cruzblanca et al. 1998). These results were interpreted to mean that the effect of BAPTA to inhibit a muscarinic effect on I M was attributable to lowering of resting [Ca 2+ ] i rather than inhibiting the rise of [Ca 2+ ] i by muscarinic stimulation (Beech et al. 1991). This interpretation was based on the assumption that adding 10-mM Ca 2+ to 20-mM BAPTA-containing solution increases free Ca 2+ concentration without affecting Ca 2+ buffering power during the event of rising [Ca 2+ ]. That is not the case, but Ca 2+ buffering power is compromised because BAPTA is occupied by the added Ca 2+ , resulting in reduction of free BAPTA that is capable of buffering Ca 2+ rise. Therefore, these results and our results may present evidence that Ca 2+ signals that mediate muscarinic inhibition of I M are only suppressed by a very high concentration of BAPTA. The requirement of high concentrations of BAPTA may imply two possibilities: muscarinic inhibition of I M somehow requires minimum resting Ca 2+ , as was suggested previously (Beech et al. 1991), or muscarinic inhibition of I M is mediated by local Ca 2+ signals that can be suppressed by high concentrations of fast Ca 2+ buffer, BAPTA. To distinguish these two possibilities, we tested 20-mM EGTA, which has similar equilibrium affinities for binding Ca 2+ with BAPTA, but its Ca 2+ -binding rate is too slow to suppress dynamic Ca 2+ rises. Since it is expected that resting Ca 2+ concentrations should be lowered to the same extent by EGTA as it is by BAPTA, differences in the ability of BAPTA and EGTA to block Ca 2+ -triggered responses can be regarded to represent the involvement of local Ca 2+ signals. Indeed, we found that muscarinic inhibition of I M was potently blocked by BAPTA but completely insensitive to EGTA (Fig. 5), implying that the spatial proximity between the Ca 2+ source of oxo-M-induced Ca 2+ rise and Ca 2+ -sensitive target proteins that mediate I M inhibition is so close that the coupling can only be blocked by very high concentrations of fast Ca 2+ buffer. We calculated the length constant of Ca 2+ microdomain in the presence of either BAPTA or EGTA, λ, which represents the mean distance that a Ca 2+ ion diffuses before it is captured by a buffer molecule (Fig. 6). According to equation (7) by Neher (Neher 1998), λ = √(D Ca /k on [B] o (k on and [B] o representing the rate constant of Ca 2+ binding to the buffer and the concentration of free buffer, respectively). We referred 1 3 diffusion coefficient of free Ca 2+ ion (223 μm 2 /s) as reported by Allbritton et al. (Allbritton et al. 1992), and the values for k on and the rate constant of Ca 2+ binding to the buffer, to the study by Naraghi (Naraghi 1997), in which k on for BAPTA and EGTA is 4.5 × 10 8 /M.s and 2.7 × 10 6 /M.s, respectively. λ is calculated to be about 5 nm and 60 nm in the presence 1 3 of 20-mM BAPTA and 20-mM EGTA, respectively. Thus, our results imply that M channels are localized in the vicinity of the muscarinic receptor sensitive Ca 2+ source within 60 nm, but away from this source at more than 5 nm. To prove this model, it is necessary to identify the Ca 2+ source and Ca 2+ -sensitive target proteins involved in I M inhibition. Further studies will be required to clarify this issue.
In addition, our results showed that muscarinic inhibition of M channel was not attenuated by blocking PIP 2 depletion with supply of PIP 2 (up to 200 μM), or by blocking PKC activation with a specific blocker, calphostin C, or by blocking both PIP 2 and PKC pathways. As other potential signaling pathways may have been involved, we examined whether G βγ subunits of G protein or trafficking of channel proteins were involved in muscarinic inhibition of I M . The results showed that neither seemed to have a critical role.
Our data demonstrate that PIP 2 depletion is not the 'principal' contributor to the inhibition of I M by muscarinic signaling, in contrast to the PIP 2 depletion hypothesis for muscarinic action. Clearly, KCNQ channels are PIP 2 sensitive: they are directly activated by exogenous PIP 2 . However, two distinct questions must be answered to evaluate the role of PIP 2 signals. The first is whether PIP 2 functions as a regulator of channel activity and a second, distinct question is whether PIP 2 signals play a role in M channel modulation by muscarinic receptors. The PIP 2 depletion hypothesis is based on a number of experiments focusing on the first question. For example, selective depletion of PIP 2 using an engineered chemical dimerization system almost completely suppressed the current, whereas PIP 2 directly applied to excised patches augmented the current (Suh et al. 2006;Zhang et al. 2003). These data, however, cannot rule out the possibility that during muscarinic signaling, PIP 2 depletion plays no role in decreasing current amplitudes. To determine the role of PIP 2 in receptor-mediated regulation of ion channels, specific experiments testing the role of PIP 2 in receptorinduced channel modulation are needed. To do this, we and others applied exogenous PIP 2 into neurons during G q PCRs signaling. This experimental protocol was used to confirm the role of PIP 2 depletion in G q PCR-mediated GIRK channel regulation (Cho et al. 2005(Cho et al. , 2001Meyer et al. 2001). As opposed to PIP 2 -mediated GIRK channel regulation, muscarinic inhibition of KCNQ channels was not suppressed by PIP 2 (up to 200 μM, Fig. 2) and only partially reduced even by 500-μM PIP 2 (Robbins et al. 2006). These results underscore the notion that PIP 2 depletion might play only a small role in muscarinic inhibition of I M . Another piece of evidence in favor of the PIP 2 model is that the dynamics and extent of channel inhibition by muscarinic receptors showed a close correlation to the translocation of the PH-GFP construct that provided a measure of overall PIP 2 hydrolysis (Hernandez et al. 2008b;Horowitz et al. 2005). However, the dynamics of PIP 2 hydrolysis does not necessarily correlate to the dynamics of PIP 2 depletion, especially in the proximity of a given ion channel. Since no method is available to measure changes in local concentrations of PIP 2 , we developed a two-dimensional diffusion model to estimate them. Using this model, we found that the key to determining the spatial and temporal profiles of PIP 2 depletion is PIP 2 mobility (Cho et al. 2005), in that a profound PIP 2 depletion restricted to the microdomain adjacent to PLC occurs only when PIP 2 mobility is low. However, in general, diffusion constants for PIP 2 in neurons are not low (~ 0.5-2 μm 2 /s, (van Rheenen and Jalink 2002;Cho et al. 2005)) and there is no evidence that the physical properties of PIP 2 in SCG neurons are dissimilar from those of PIP 2 in other neurons. When PIP 2 is diffusible, the PIP 2 depletion induced by PLC activation is readily attenuated by diffusion, so that the resulting changes become slower and smaller (Cho et al. 2005;Cui et al. 2010). This is in contrast to the rapid and profound I M inhibition observed during muscarinic stimulation. Support for the PIP 2 model also comes from the fact that the recovery from inhibition requires cytoplasmic ATP and is delayed by inhibition of the PI-4-kinase, which replenishes PIP 2 (Ford et al. 2003;Suh and Hille 2002). However, although recovery from inhibition requires resynthesis of PIP 2 , whether the process of inhibition itself results directly from PIP 2 breakdown is unclear. It must be noted that for GIRK channels in HEK cells, the recovery from receptor-induced inhibition depends primarily on PIP 2 regeneration but receptorinduced inhibition results primarily from the change in the channel's affinity for PIP 2 by other signaling molecules . Furthermore, as mentioned earlier, although muscarinic inhibition of I M is reduced by overexpressing the synthetic enzyme PI-5-kinase (Winks et al. 2005), whether muscarinic receptor signaling no longer functions in cells with increased resting PIP 2 level caused by overexpression of PI-5-Kinase needs to be tested and resolved.
Reportedly, Ca 2+ elevation does not require as much G q PCR stimulation or local receptor density as does PIP 2 depletion or M channel inhibition (Falkenburger et al. Thus, Ca 2+ signaling sensitizes SCG neurons to M 1 R activation, allowing I M inhibition even when M 1 R is at low receptor density or under minimal muscarinic acetylcholine stimulation.
The results of our study do not exclude or support the possibility that PIP 2 depletion and Ca 2+ elevation might occur simultaneously in response to muscarinic stimulation; and this could enhance the inhibitory effect. However, we do show that use of exogenous Ca 2+ buffer to inhibit Ca 2+ rises could block muscarinic inhibition, whereas exogenously restoring PIP 2 levels could not, suggesting Ca 2+ elevation, not PIP 2 depletion, is the principal driver of muscarinic inhibition of M currents.
Author Contributions JYY and WKH conceived the experiments. JYY designed and executed the experiments. JYY analyzed the data. JYY and WKH interpreted the data. JYY wrote the manuscript. JYY and WKH revised the manuscript. All authors approve the final version of the manuscript for publication and agree to be accountable for all aspects of the work. All listed authors meet the requirements for authorship, and all those who qualify for authorship are listed.
Funding No funding was received.

Data Availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflict of interest
The authors declare that they have no conflict of interest.
Ethical Approval This study was reviewed and carried out in accordance with the Institutional Animal Care and Use Committee (IACUC) at Sungkyunkwan University School of Medicine (SUSM).

Consent to publish Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.

Fig. 6
Model for Ca 2+ microdomains for muscarinic modulation of I M in SCG neurons. The length constant of Ca 2+ microdomains associated with individual M1 muscarinic receptors (M1) is calculated to be ~ 5 nm and 60 nm in the presence of 20-mM BAPTA (left) and 20-mM EGTA (right), respectively. Ca 2+ rises in response to muscarinic stimulation can lead to weakening channel-PIP 2 interaction, resulting in the decrease of M currents 1 3 1 3