Regulation of Kv2.1 channel inactivation by phosphatidylinositol 4,5-bisphosphate

Phosphatidylinositol 4,5-bisphosphate (PIP2) is a membrane phospholipid that regulates the function of multiple ion channels, including some members of the voltage-gated potassium (Kv) channel superfamily. The PIP2 sensitivity of Kv channels is well established for all five members of the Kv7 family and for Kv1.2 channels; however, regulation of other Kv channels by PIP2 remains unclear. Here, we investigate the effects of PIP2 on Kv2.1 channels by applying exogenous PIP2 to the cytoplasmic face of excised membrane patches, activating muscarinic receptors (M1R), or depleting endogenous PIP2 using a rapamycin-translocated 5-phosphatase (FKBP-Inp54p). Exogenous PIP2 rescued Kv2.1 channels from rundown and partially prevented the shift in the voltage-dependence of inactivation observed in inside-out patch recordings. Native PIP2 depletion by the recruitment of FKBP-Insp54P or M1R activation in whole-cell experiments, induced a shift in the voltage-dependence of inactivation, an acceleration of the closed-state inactivation, and a delayed recovery of channels from inactivation. No significant effects were observed on the activation mechanism by any of these treatments. Our data can be modeled by a 13-state allosteric model that takes into account that PIP2 depletion facilitates inactivation of Kv2.1. We propose that PIP2 regulates Kv2.1 channels by interfering with the inactivation mechanism.

SCIeNTIFIC RepoRts | (2018) 8:1769 | DOI: 10.1038/s41598-018-20280-w of PIP 2 micelles to the cytoplasmic side of membrane patches abolished the fast (N-type) inactivation of these channels 10 . However, Kruse et al. 13 using whole-cell recordings and enzymatic methods to deplete native PIP 2 in intact cells, did not report PIP 2 sensitivity for several Kv channels, including channels previously reported to be PIP 2 sensitive (Kv1.1/Kβ1.1, Kv1.4 and Kv3.4). Kv2.1 is another ion channel that was first reported to be PIP 2 sensitive. Hilgemann et al. 17 reported that Kv2.1 channels were rescued from rundown in excised inside-out patches by applying exogenous PIP 2 . However, experiments using intact cells failed to demonstrate the PIP 2 sensitivity of Kv2.1 13 . The reasons why some Kv channels (e.g. Kv2.1) show PIP 2 sensitivity in excised patches, but not in intact cells still remains poorly understood and deserves further investigation.
In this work, we explored the PIP 2 regulation of Kv2.1 channels expressed in HEK293 cells by employing different strategies to manipulate PIP 2 levels in excised patches and intact cells. In excised inside-out patches, PIP 2 recovered the rundown of Kv2.1 channels and prevented the shift in the voltage dependence of inactivation. In whole-cell experiments, native PIP 2 depletion by membrane translocatable enzymes or M1R activation facilitated the inactivation of Kv2.1 channels. Our data and quantitative analysis suggest that PIP 2 depletion increases the transitions from closed-and open-states to the inactivated-states.

Results
PIP 2 regulates Kv2.1 channels activity in excised inside-out patches. Patch excision leads to a gradual loss of PIP 2 from the plasma membrane, and alterations of channel function that accompany patch excision can reflect PIP 2 dependence. For instance, rundown and changes in the voltage-dependence of Kv1.2 channels in inside-out patches can be reduced by manipulations that slow the loss of PIP 2 and can be reversed by application of exogenous PIP 2 to the intracellular face of the patch 11 . To test PIP 2 dependence on Kv2.1 channels, we first recorded Kv2.1 currents in excised inside-out patches. Following patch excision into symmetrical K + , Kv2.1 currents activated by a test pulse to + 60 mV from a holding potential of −80 mV, successively decreased in amplitude (rundown) to ~20% of the initial current ( Fig. 1A-C). Bath application of exogenous PIP 2 (10 μM) to the excised inside-out patch significantly reversed the current rundown to ~76% of the initial current ( Fig. 1A-C), while doubling the PIP 2 concentration did not further reverse the current rundown ( Supplementary Fig. S1). Similarly, application of 2 mM Mg-ATP to induce the phosphorylation of PIP by the lipid kinases associated to the patch, partially recovered the Kv2.1 currents from rundown ( Supplementary Fig. S2). Additionally, the use of a phosphatase inhibitor-containing solution (FVPP 18 ), to limit the known PIP 2 dephosphorylation in insideout patches slowed the rundown of Kv2.1 currents; the rundown rate (τ rundown ) in normal external solution was 105.7 ± 9.2 s compared to 298.4 ± 39.8 s in FVPP (p < 0.01, Supplementary Fig. S3). A striking observation in inside-out patches was that recordings of Kv2.1 currents using a more negative holding potential (−120 mV) did not show a significant rundown ( Fig. 2A,B). However, after switching the holding potential to −80 mV, Kv2.1 currents quickly randown, but could be recovered by returning the holding potential back to −120 mV ( Fig. 2A,B). It is known that voltage-dependent gating can be modulated in inside-out patches 11,12,19 . We hypothesized that in inside-out patches the voltage-dependence of inactivation could be shifted to hyperpolarizing potentials and consequently, at our holding potential of −80 mV Kv2.1 currents could be inactivating, as this channel shows preferential closed-state inactivation 20 . To test this hypothesis, we compared the inactivation curves of Kv2.1 channels recorded in the whole-cell and inside-out configurations. Under whole-cell conditions, the V 1/2 of inactivation was −28.3 ± 1.4 mV (k = 5.6 ± 0.3 mV), whereas it was −76.9 ± 1.6 mV (k = 5.3 ± 0.3 mV) in the insideout patch configuration (Fig. 2C). Thus in the inside-out patch mode, the inactivation curve was largely shifted to hyperpolarized potentials (ΔV 1/2 = 48.6 mV). From the curve in the inside-out patch mode it can be seen that there is significant inactivation at −80 mV, in contrast to when the patch is held at −120 mV. Thus, it is likely that the rundown observed in the inside-out patch recordings at a holding potential of −80 mV was due to an accumulation of inactivation of Kv2.1 channels, as rundown was not observed at a holding potential of −120 mV, where inactivation does not take place. Because PIP 2 could recover the current from rundown, we tested if this effect was related to the inactivation of Kv2.1. As anticipated, application of PIP 2 partially prevented the hyperpolarizing shift in the inactivation curve of Kv2.1 channels (V 1/2 = −57.8 ± 1.3 mV, k = 10.9 ± 1.0 mV, Fig. 2C). This partial effect of PIP 2 suggests that there are additional factors contributing to the difference between the wholecell and inside-out patch inactivation curves. Altogether, these results suggest that the Kv2.1 rundown observed in inside-out patches is due to an accumulation of inactivation at −80 mV (the holding potential), as the voltage dependence of inactivation is shifted to hyperpolarized potentials. The recovery of Kv2.1 from rundown induced by PIP 2 is in turn due to a shift of the inactivation curve towards more depolarized potentials.
Effects of native PIP 2 depletion on Kv2.1 channels. Several ion channels reported to be PIP 2 dependent by means of their responses to exogenous PIP 2 application in excised inside-out patches, were shown to be insensitive to PIP 2 depletion in intact cells 13 . A number of reasons have been proposed to explain these disparities; among them is the incorporation of large amounts of PIP 2 to the inner leaflet of the plasma membrane (in inside-out patches), generating unphysiological protein conformations 14 . Therefore, we turned to whole-cell recordings to study modulation of Kv2.1 channels following native PIP 2 depletion in intact cells. For this purpose, we expressed Kv2.1 channels together with a rapamycin-translocatable lipid 5-phosphatase (FKBP-Inp54p) and the membrane anchor LDR-CFP. The FKBP and LDR domains dimerize when rapamycin is added to cells, leading to the dephosphorylation of PIP 2 on the 5′ position to PI4P 21 . In all experiments, we used 1 μM rapamycin. We also expressed Kv2.1 channels together with the M1 muscarinic receptor (M1R). M1R is a Gq-coupled receptor that activates phospholipase C (PLC) leading to PIP 2 depletion 22 . In all experiments, we used 10 μM acetylcholine (ACh) to activate the M1R.
Recently, Kv2.1 channels were reported to be insensitive to PIP 2 depletion by several experimental approaches 13 . In this work, the amplitude of Kv2.1 currents was barely affected by PIP 2 depletion; however, no effects on other Kv2.1 biophysical parameters were studied. In our hands, we observed that the recruitment of the 5-phosphatase (FKBP-Inp54p) to the plasma membrane by rapamycin had small but significant effects on the amplitude of Kv2.1 currents (Fig. 3A). The peak-and end-pulse currents were inhibited 13.7 ± 3.1% (p < 0.01) and 27.4 ± 4.5% (p < 0.01; Fig. 3B), respectively. The larger inhibition at the end of the test pulse is likely due to an acceleration of inactivation (τ control = 1.2 ± 0.2 s and τ rapamycin = 0.77 ± 0.16 s). We also expressed Kv2.1 channels together with M1R. Upon activation of M1R, Kv2.1 peak-and end-pulse currents were inhibited 23.7 ± 5.3% (p < 0.01) and 42.1 ± 7.2% (p < 0.01; Fig. 3C,D). This larger reduction upon M1R activation could be related to loss of PIP 2 and to effects of products generated downstream of PIP 2 hydrolysis induced by PLC activation, as was previously suggested 13 . The inactivation rate was also accelerated after M1R activation (τ control = 0.94 ± 0.16 s and τ M1R = 0.68 ± 0.07 s).
From these results, we concluded that PIP 2 depletion in intact cells shows smaller effects on the ion conduction of Kv2.1 channels than in excised patches. Thus, we next turned to study the effects of PIP 2 depletion on the voltage-dependent gating of Kv2.1 channels.
The voltage-dependence of activation of Kv2.1 channels is not altered by PIP 2 depletion. To examine the effect of PIP 2 depletion on the voltage-dependence of activation of Kv2.1 channels, we determined the conductance-voltage relations (G-V curves) recorded from cells studied in the whole-cell mode before (control) and after PIP 2 depletion induced by the application of rapamycin (Fig. 4). Figure 4A,B shows representative currents traces before (control, A) and after application of rapamycin (B), recorded with the protocol shown in the inset. The midpoint voltage of activation (V 1/2 ) was 13.8 ± 0.9 mV (k = 15.8 ± 0.6 mV) under control conditions, and was not significantly modified after rapamycin application (V 1/2 = 11.5 ± 1.5 mV, k = 17.3 ± 0.9 mV, p > 0.05) (Fig. 4C). Similarly, activation of M1R did not induce a significant change in the V 1/2 of activation ( Fig. 4D, Supplementary Fig. S4). The V 1/2 was 15.2 ± 1 mV (k = 16.9 ± 0.9 mV) under control conditions and 12.8 ± 2.2 mV (k = 17.9 ± 1.3 mV) after M1R activation (p > 0.05). channels in excised inside-out patches that could be prevented (at least partially) by exogenous PIP 2 (Fig. 2C). Therefore, we looked next for changes in the V 1/2 of inactivation after PIP 2 depletion in intact cells. Figure 5A,B shows representative current traces before (control, A) and after application of rapamycin (B), recorded with the protocol shown in the inset. After PIP 2 depletion by rapamycin application, the V 1/2 of inactivation was shifted toward more negative voltages relative to the control (Fig. 5C). The V 1/2 in the control was −28.3 ± 1.4 mV (k = 5.6 ± 0.3 mV), while it was −41.3 ± 2.0 mV (k = 5.6 ± 0.2 mV) following rapamycin application (i.e. a 13 mV leftward-shift, p < 0.001). Activation of M1R also resulted in a ~8 mV shift of the inactivation curve to hyperpolarized potentials (Fig. 5D, Supplementary Fig. S5). The V 1/2 in the control was −24.3 ± 0.5 mV (k = 6.3 ± 0.3 mV), while it was −31.9 ± 1.3 mV (k = 6.6 ± 0.4 mV) following M1R activation (p < 0.001). These results also demonstrate modulation of the inactivation gating of Kv2.1 channels by PIP 2 in whole cells.
Closed-state inactivation in Kv2.1 is more favorable following PIP 2 depletion. An important pathway of inactivation in Kv2.1 channels originates from the closed states 20 . Because the V 1/2 of inactivation is altered by PIP 2 depletion, is conceivable that the inactivation of Kv2.1 channels is also modified from the closed-states. To test this hypothesis, we isolated closed-state inactivation by examining the development of inactivation at a non-activating voltage (−40 mV), before (control) and after PIP 2 depletion by rapamycin (Fig. 6). In these experiments, a pulse to + 80 mV tested the available current after the pre-pulse to −40 mV. As the pre-pulse duration increased, channels inactivated and the current gradually decreased. Figure 6A,B shows representative currents traces before (control, A) and after application of rapamycin (B), recorded with the protocol shown in the inset. The development of closed-state inactivation was significantly accelerated after PIP 2 depletion and well described, assuming an exponential decay (Fig. 6C). For instance, under control conditions, the time constant of the closed-state inactivation was 29.5 ± 6.8 s, whereas it was 13.4 ± 1.1 s following rapamycin application (p < 0.01). The closed-state inactivation was also accelerated by the activation of M1R (Fig. 6D, Supplementary  Fig. S6), the time constant of closed-state inactivation was 32.5 ± 3.9 s, whereas it was 13.6 ± 1.9 s following M1R activation (p < 0.01). The recovery of Kv2.1 channels from inactivation is delayed by PIP 2 depletion. Next, we examined the time course of the recovery from inactivation of Kv2.1 channels. Recovery from inactivation was measured using the protocol shown in the inset of Fig. 7. Figure 7A,B shows representative current traces before (control, A) and after application of rapamycin (B). The recovery kinetics of Kv2.1 current were best fit by a double exponential function that was modestly but significantly altered by PIP 2 depletion (Fig. 7C). The two recovery time constants were 0.15 ± 0.02 s and 1.07 ± 0.6 s for control, and 0.3 ± 0.03 s and 2.1 ± 0.5 s following rapamycin application (p < 0.01). Similar effects on the recovery kinetics were observed after M1R activation (Fig. 7D, Supplementary Fig. S7), The two recovery time constants were 0.18 ± 0.05 s and 1.9 ± 0.35 s for control, and 0.32 ± 0.5 s and 4.1 ± 0.5 s following M1R activation (p < 0.01).
Allosteric model of the effects of PIP 2 depletion on Kv2.1 channels. To better understand the mechanism by which PIP 2 alters the function of Kv2.1 channels, we constructed a kinetic model. Figure 8A shows a 13-state model used to reproduce the effects of PIP 2 depletion induced by rapamycin on Kv2.1 channels. As previously proposed by others 19,20 , in the building of a Kv2.1 channel model we assumed a sequential activation of the four channel subunits in the closed state (C 0 , C 1 , C 2 , C 3 , and C 4 ), before the channel could reach the open state (O). Then, from either the open or closed states the channel could inactivate (I 0 , I 1 , I 2 , I 3 , I 4 , I o0 , I 5 ). Transitions between states were controlled by voltage dependent exponential rate constants. At a holding potential of −100 mV, channels dwelled predominantly in the C 0 state, and went to constitutive closed states (C 1 , C 2 , C 3 to C 4 ) as the membrane voltage depolarized until the channels reached the open state (O). At the same time, as the membrane was depolarized the Kv2.1 channel could reach the inactivated states from the closed states. The model contains 11 free parameters (Table 1) whose values need to be determined from experimental data. To obtain the values of the parameters of each rate constant, we performed best global fits using ICHMASCOT 23 . We constrained the calculation by including in the fitting procedure the following data set: (a) the time course of the potassium currents at different membrane voltages (Fig. 4A,B), (b) the time course of currents induced by long pulses at different inactivating voltages (Fig. 5A,B), (c) the time course of inactivation from closed-states (Fig. 6A,B), and (d) the time course of recovery from inactivated states (Fig. 7A,B).  we restricted the number of free parameters assuming that rate constants, k CO and k OC , which mediate the closed-open and open-closed transitions, respectively, did not change. We based this assumption on the observation that the kinetic constants of activation and deactivation as a function of membrane voltage were not statistically different between control and rapamycin conditions. Additionally, the voltage-dependence of the conductance, measured at maximum current, was equal in both conditions (Fig. 4C). On the other hand, the ability of Kv2.1 channels to reach the inactivated state when PIP 2 was depleted, from the open (Fig. 5) or the closed-state (Fig. 6), suggest that f , k CI , k OI , and k O5 are modified. At the same time, the data presented in Fig. 7 show that channel current could be recovered almost back to control levels. Because of this reason b, k IC , k IO , and k O 5 were free to compensate for the more pronounced inactivation observed in the presence of rapamycin. The modified kinetic model fits the Kv2.1 channel once PIP 2 is depleted by rapamycin. Figure 8B-E (middle panel) shows the best fits in blue on top of experimental recordings (red traces). The best-fit parameter values for the control and rapamycin conditions obtained this way are listed in Table 1.
To make sure that the model captured steady-state properties of Kv2.1 channels, we simulated the currents using IonChannelLab 24 . The sets of rate constants are listed on Table 1. The simulated currents were then used to determine the voltage-dependent activation, voltage-dependent inactivation, time course of inactivation and time course of recovery from inactivation. In Fig. 8B-E (right panel), we compared the experimentally determined (black dots = control; red dots = rapamycin) voltage-dependence of activation (B), voltage-dependence of inactivation (C), development of closed-state inactivation (D), and recovery from inactivation (E), to those predicted by the model (continuous black and red lines). In general, the model nicely predicted most of these properties. However, the development of closed-state inactivation (Fig. 8D, right panel) was faster than the experimental data; the experimental inactivation time constant was 29.5 ± 6.8 s vs 14.4 s obtained from the simulated currents. This discrepancy was larger for the rapamycin condition. The model predicts a time constant that is about ~2.7 times faster than that calculated experimentally. These discrepancies may be due to new inactivated states that the channel could occupy when PIP 2 becomes depleted. Additional studies will be necessary to resolve these differences.

Discussion
Numerous ion channels have been reported to be PIP 2 sensitive 15 . However, differences between intact cells and excised patches have been observed regarding the modulation of Kv channels by this lipid 13,14 . Kv2.1 channels were originally reported as PIP 2 sensitive, as PIP 2 was able to rescue Kv2.1 currents from rundown in excised patches 17 . However, later reports showed that native PIP 2 depletion in intact cells barely affected Kv2.1 function 13 . Here, we have further investigated the regulation of Kv2.1 channels by PIP 2 using different approaches. We found that inactivation of Kv2.1 channels is sensitive to PIP 2 levels.
First, we confirmed that Kv2.1 currents can be rescued from rundown by application of exogenous PIP 2 or Mg-ATP to excised inside-out patches (Fig. 1A-C and Supplementary Fig. S2). Additionally, we found that Kv2.1 current rundown depends on the holding potential (Fig. 2). Our data indicate that this phenomenon is related to the modulation of the inactivation of Kv2.1 in excised inside-out patches. After patch excision, the inactivation curves shift to hyperpolarized potentials (Δ48.6 mV compared to whole-cell, Fig. 2C), and hence, at −80 mV (the holding potential of most of our experiments) the channels can inactivate, leading to a gradual loss of Kv2.1 currents. This modulation of inactivation under inside-out conditions was previously reported for Kv4 channels 19 , channels which, like Kv2.1, possess preferential closed-state inactivation. The fact that PIP 2 can recover Kv2.1 currents from rundown is due to the fact that it can restore (at least partially) the changes in inactivation that occurred following patch excision (Fig. 2C).
It has been proposed that the application of exogenous PIP 2 to the cytoplasmic face of excised patches could produce physiologically irrelevant results. Consequently, it has been suggested that the use of tools to deplete PIP 2 that maintain the cellular environment as stable as possible would produce more reliable data 13,14 . Considering these potential drawbacks, we turned to strategies to deplete native PIP 2 in whole-cell experiments, to test whether we would corroborate our results from inside-out patches.
Depletion of PIP 2 on whole-cell experiments by means of the recruitment of a lipid 5-phosphatase (FKBP-Inp54p) to the plasma membrane, or the activation of a Gq coupled receptor (M1R) with the consequent activation of phospholipase C (PLC) and PIP 2 hydrolysis, confirmed the modulation of the inactivation gating of Kv2.1 channels by PIP 2 . Kv2.1 channels exhibited their characteristic U-shaped voltage-dependence of inactivation (Fig. 5C,D). It has been suggested that U-shaped inactivation consists of a mix of both open-and closed-state inactivation with preferential inactivation from closed states throughout the voltage range 20 . PIP 2 depletion did not modify the U-shaped inactivation curve, suggesting that preferential inactivation from closed-states was conserved (Fig. 5C,D); in fact, represent closed, inactivated and opened states, respectively. Rate constants a, b and k OC depend exponentially on membrane voltage, and k CO , k IC , k CI , k IO , k OI , k O5 , and k O 5 are voltage independent rate constants (Table 1). Voltagedependent activation and inactivation from closed states are coupled by an allosteric factor f . (B-E), Simulated currents (blue) reproduced from the kinetic model shown in (A), in control conditions (left panel) and rapamycin conditions (middle panel), using values from Table 1, and by applying the corresponding voltage protocols shown in Figs 3-6 (insets). Experimentally obtained currents are shown in black (control) and red (rapamycin). Curves plotted in right panels (continuous lines) were obtained from simulated currents in control (black) and rapamycin (red) conditions. Symbols represent the experimental data.
SCIeNTIFIC RepoRts | (2018) 8:1769 | DOI:10.1038/s41598-018-20280-w the inactivation from closed-states was favored after PIP 2 depletion (Fig. 6). In agreement with a more favorable inactivation from closed-states after PIP 2 depletion, the inactivation curve was shifted toward negative voltages (Fig. 5C,D). Furthermore, the recovery of Kv2.1 channels from inactivation was delayed after PIP 2 depletion (Fig. 7), indicating that all inactivation parameters are modulated by PIP 2 . Interestingly, the changes in the inactivation gating occurred in the absence of a significant alteration of the activation mechanism (Figs 4, 8), suggesting that both mechanisms are uncoupled in Kv2.1 channels, contrary to what has been suggested for other Kv channels 25 .
To understand the modulation of Kv2.1 inactivation after PIP 2 depletion, we incorporated our experimental results into a kinetic model (Fig. 8). Quantitatively, there is strong voltage-dependence between closed-state transitions (Table 1) , k ctrl and ΔG k rap , , k rap are energy barriers and rate constants in the control and rapamycin conditions, respectively, that the increment of the rate constants toward inactivated states, once PIP 2 is dephosphorylated, decreases the energy barrier ( . RT 0 3 for k k / CI ctrl CI rap : : and RT 1 0 . for k k / OI ctrl IO rap : : ), hence transitions toward inactivated states are more likely. For example, at −40 mV, in control conditions, after visiting the closed states, the channel reaches the C 4 state with 0.01 of occupation probability after a 6-s depolarization (Fig. 8C). In contrast, the C 4 state has 0.03 of occupation probability when rapamycin was applied. However, the occupation probability of I 4 changed from 0.07 to 0.26 by applying rapamycin. High probability of occupancy of the I 4 state is responsible for the time acceleration from the closed state (Fig. 8D). By closing the channel from this conformational state, by applying −100 mV, and reopening with + 80 mV, Kv2.1 has 0.79 of probability of opening in the control condition, but only 0.53 in the rapamycin condition: the final result is the change of the I-V inactivation curve to negative voltages (Fig. 8C, right panel). By applying + 40 mV, a membrane voltage that favors the opening of Kv2.1, the probability of occupation of the O, I O0 , I 5, states is 0.25, 0.06 and 0.25, respectively, under control conditions. In the presence of rapamycin these states showed a 0.18, 0.12 and 0.31 probability of occupancy. In summary, PIP 2 depletion increases the transitions from open-and closed-states to the inactivated states, with the C 4 to I 4 being the most favorable following PIP 2 depletion.
What are the physiological implications of our findings? Kv2.1 channels are expressed in a wide variety of cells, such as central neurons 26 , cardiomyocytes 27 , and pancreatic β cells 28 , contributing to a wide range of physiological processes. For example, Kv2.1 is involved in the glucose-stimulated insulin secretion (GSIS) in pancreatic β-cells 29 . Opening of Kv2.1 channels repolarizes β-cell action potentials, limiting Ca 2+ entry and, thus, insulin secretion 30,31 . Interestingly, the covalent attachment of small ubiquitin-like modifiers proteins to Kv2.1 channels (SUMOylation) modulates pancreatic β-cell excitability. SUMOylation modulates the inactivation gating of Kv2.1, which resulted in the widening of β-cell action potential and a decrease in firing frequency 4,5 , leading to an increase in the glucose-dependent insulin secretion. In this context, the modulation of Kv2.1 channels inactivation by PIP 2 could result in similar effects. For instance, PIP 2 levels in β-cells are dynamically regulated 32 . An important mechanism of such regulation is the stimulation of muscarinic receptors by acetylcholine, which is the major parasympathetic neurotransmitter released within pancreatic islets during food intake 33 . Thus, muscarinic signaling could be influencing the functioning of Kv2.1 channels, influencing β-cell excitability.
In summary, our data indicate that PIP 2 modulates the inactivation gating of Kv2.1 channels, which could have an influence on the excitability of Kv2.1 expressing cells.

Data analysis.
Patch-clamp data were processed using Clampfit 10 (Molecular Devices) and analyzed in Origin 8.6 (OriginLab Corp. Northampton, MA, USA). Conductance (G) was calculated using the following equation: where I p is the peak current amplitude at the test potential V, and V rev is the potassium reversal potential. Normalized conductance was then plotted against V to determine the voltage dependence of activation. The activation curves were then analyzed using the Boltzmann equation: where V represents the test potential, V 1/2 is the potential at which the conductance was half-maximally activated, and K is the slope. The voltage dependence of Kv2.1 channel steady-state inactivation was determined using a three-step protocol. From a holding potential of −120 mV, a 100 ms depolarizing step to + 80 mV was applied (P 1 ), after a brief repolarization to the holding potential, a 6-s conditioning pulse to potentials between −120 and + 80 mV was applied (P 2 ) followed by a final pulse to + 80 mV (P 3 ). The normalized current was calculated dividing the P 3 / P 1 current and plotted versus the conditioning potential (P 2 ). The data were fitted with the Boltzmann equation: Where V is the conditional potential, V 1/2 is the potential at which the conductance was half-inactivated, and K is the slope. Data are presented as mean ± S.E.M. (n = number of cells, recorded from at least 5 different experiments). Statistical comparisons were made using paired or unpaired Student´s t-test where applicable. Statistical significance was set at p < 0.05. Data availability statement. The datasets generated during and/or analysed during the current study are available from the corresponding authors upon request.