Introduction

The two subtypes of phospholipase C-related but catalytically inactive protein (PRIP-1/2) were first identified as novel inositol 1,4,5-triphosphate (IP3) binding proteins [21, 22]. Subsequently, they were revealed to be involved in the membrane trafficking of GABAA receptors (GABAARs) [23]. In PRIP-1/2 double-knockout (PRIP-DKO) mice, GABAA currents exhibited a reduced diazepam sensitivity due to a reduced expression of GABAARs composed of γ2 subunits as a consequence of the lack of the binding between PRIP-1/2 and GABAAR-associated protein [23], in the hippocampal [30] and the cerebellar granule neurons [31]. On the other hand, PRIP-1 and PRIP-2 were also revealed to be involved in the regulation of intracellular Ca2+ stores; IP3-induced Ca2+ release (IICR) was impaired in cultured cortical neurons from PRIP-1 knockout (KO) mice [9] while store-operated Ca2+ entry (SOCE) was enhanced in hematopoietic B cells in PRIP-2 KO mice [40]. However, it has not been investigated whether or not the deletion of PRIP-1/2 influences the kinetic properties of GABAA currents through the altered regulation of intracellular Ca2+ stores, in spite of the fact that there is a large body of evidence that GABAAR-mediated currents can be regulated by the intracellular concentration of Ca2+ ([Ca2+]i) [12, 39].

Ca2+/calmodulin-dependent protein kinase II enhances GABAAR-mediated currents in spinal dorsal horn neurons [44] and in cerebellar granule neurons [11], while the activation of protein phosphatase 2B (calcineurin) by Ca2+ induces a suppression of GABAAR-mediated currents in hippocampal neurons [28] and a slowdown of the rate of GABA unbinding from GABAARs [19]. An increase in [Ca2+]i through the activation of various types of Ca2+ channels, including voltage-dependent Ca2+ channels and NMDA, IP3, and ryanodine receptor channels, causes a suppression of GABAAR-mediated currents [5, 15, 32, 42], through the acceleration of desensitization [33]. Considering the altered regulation of intracellular Ca2+ stores in PRIP-1 KO and PRIP-2 KO mice [9, 40] together with the finding that layer 2/3 pyramidal cells (PCs) in the rat neocortex express high levels of calcineurin [25, 26], it can be hypothesized that the kinetic properties of GABAAR-mediated currents differ between layer 3 (L3) PCs in wild-type (WT) and PRIP-DKO mice due to putative anomalies in intracellular Ca2+ dynamics in PRIP-DKO PCs.

Therefore, in the present study, we investigated the kinetic properties of GABAAR currents (IGABA) evoked by puff application of GABA and its modulation by manipulation of intracellular Ca2+ and Ca2+ signaling, in L3 PCs of the barrel cortex in PRIP-DKO mice. Our results suggest that the deletion of PRIP-1/2 results in the enhancement of the desensitization and resensitization of GABAARs by the upregulation of Ca2+-dependent activity of calcineurin through the potentiated Ca2+-induced Ca2+ release (CICR) and SOCE.

Methods

All experiments were performed in accordance with the guidelines for the care and use of laboratory animals approved by the animal ethics committee of the Osaka University Graduate School of Dentistry.

Slice preparation

PRIP-1/2 double-knockout (PRIP-DKO) mice [30] and WT C57BL/6J mice of both sexes at 15–21 days old were used. PRIP-DKO mice used in this study were back-crossed against the C57BL/6J background [30]. They were anesthetized with ketamine and diethylether, and the brain was quickly removed from the skull and immersed in ice-cold modified artificial cerebrospinal fluid (M-ACSF) composed of 210 sucrose, 2.5 KCl, 2.5 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, 0.5 CaCl2, and 50 d-glucose in mM. The rostral part of the brain block including the barrel cortex was cut at 15° tilted caudally from the coronal plane. With a microslicer (Super ZERO-1, Dosaka EM, Kyoto, Japan), sections of 300–350 μm thickness including the whisker barrel were cut parallel to the plane. Slices were incubated at 32 °C for 30 min in 50 % M-ACSF and 50 % normal ACSF (N-ACSF; pH 7.3) composed of 126 NaCl, 3 KCl, 1 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, 2 CaCl2, and 10 d-glucose in mM. The slices were then placed in N-ACSF at room temperature (20–24 °C). N-ACSF was continuously gassed with a mixture of 95 % O2–5 % CO2.

Whole-cell recording

Using Axopatch 200B (Molecular Devices, Foster City, CA, USA), all whole-cell recordings were made from visually identified PCs in L3 of the barrel cortex under the microscope equipped with Nomarski optics (BX-51WI DIC; Olympus, Tokyo, Japan). The patch pipettes had a DC resistance of 4–5 MΩ when filled with the standard internal solution composed of 130 Cs-gluconate, 10 CsCl, 2 MgCl2, 2 ATP-Na2, 0.4 GTP-Na3, 10 HEPES, and 5 EGTA, pH 7.3 adjusted with CsOH in mM. The Cl− equilibrium potential (E Cl) was calculated to be −57 mV. Voltage-clamp recordings were made at a holding potential of 0 mV (Figs. 1, 2, 4, 5, 6, and 7). In the experiment to investigate the reversal potentials of IGABA and its tail-current (tail-I) (Fig. 3), rundown of IGABA that would affect the current–voltage (I–V) relationship was prevented by substituting 2 mM MgCl2 and 2 mM ATP-Na2 with 5 mM Mg-ATP and increasing [EGTA]i from 5 to 10 mM [13]. To maintain E Cl at −57 mV, 130 mM Cs-gluconate was decreased by 4 mM by substituting with equimolar CsCl. The membrane potential values were corrected for the liquid junction potential (10 mV) between the standard internal solution (negative) and the extracellular solution. Using a pressure-pulsed microinjector (4–6 psi; PV830, World Precision Instruments, Sarasota, FL, USA), a puff of 200 μM GABA or 200 μM muscimol for 2 s was repeatedly applied to PCs every minute through a glass pipette, the tip of which was placed 25–50 μm away from their somata. Whole-cell recordings of IGABA were started 10 min after establishing the whole-cell configuration and clamping at 0 mV.

Fig. 1
figure 1

IGABA evoked by GABA puff applications in WT and PRIP-DKO PCs. A, B Sample traces of IGABA responses evoked in WT and PRIP-DKO PCs dialyzed with 5 mM EGTA, respectively, at 0 mV by 2-s puff application of GABA with puff-pressures of 4 and 6 psi (continuous and dotted lines, respectively). a The peak amplitude. b The amplitude at the offset of puff application. c The peak amplitude after the offset of puff application. §The half-duration of the tail-I. #The half-desensitization (half-Ds) time measured at the amplitude of [(a + b)/2]. C The relationship between the desensitization degree [Ds = (a – b)/a] of the IGABA and half-duration of the tail-I (§) induced by a puff with 4 psi. † P < 0.01. D The relationship between the half-Ds time of the IGABA (#) and half-duration of the tail-I (§) induced by a puff with 4 psi. † P < 0.01

Full size image
Fig. 2
figure 2

Effects of GABAAR antagonists on IGABA and IMus in PRIP-DKO PCs. A a Superimposed sample traces of IGABA evoked by 200 μM GABA puff application obtained before and 1, 3, and 11 min after application of 200 μM picrotoxin (PTX) and 5 min after subsequent co-application of PTX and 20 μM bicuculline (BIC) in PRIP-DKO PCs dialyzed with 5 mM EGTA. B a Superimposed sample traces of muscimol-induced currents (IMus) evoked by 200 μM muscimol puff application obtained before and 1, 3, and 10 min after application of 50 μM BIC and 5 min after subsequent co-application of BIC and 100 μM PTX in PRIP-DKO PCs dialyzed with 5 mM EGTA. The mean Ds, half-Ds time, Rs, and half-duration of the tail-I of IMus were 0.21 ± 0.11, 0.90 ± 0.18 s, 0.08 ± 0.03, and 4.84 ± 0.98 s, respectively (n = 5). A b, B b Plotting of the normalized amplitudes of IGABA (A b) or IMus (B b) following application of PTX or BIC, which were measured at the three time points: the peak time of the control IGABA (diamond), the offset of puff application (square), and the time point after the half of the half-duration of the control tail-I from the puff offset (circle). The amplitudes at the peak time of IGABA and IMus and at the 1/2 half-duration were significantly smaller (*P < 0.05 and *P < 0.05, respectively) than those at the puff offset at the latest 2 min after application of PTX and BIC, respectively

Full size image
Fig. 3
figure 3

I–V relationships of IGABA in WT and PRIP-DKO PCs. A a, B a, C a Sample traces of IGABA induced by puff applications of GABA at different holding potentials in WT and PRIP-DKO PCs dialyzed with 10 mM EGTA (A a and B a, respectively) and in PRIP-DKO PCs dialyzed with 10 mM BAPTA (C a). A b, B b, C b The mean I–V relationships for IGABA evoked in WT and PRIP-DKO PCs dialyzed with 10 mM EGTA (A b and B b, respectively) and in PRIP-DKO PCs dialyzed with 10 mM BAPTA (C b), measured at three time points: the time point of peak IGABA (triangles), the offset of puff application (squares), and 1.5 s after the offset of puff application (circles, A b and C b) or the time point of peak tail-Is (circles, B b). Respective amplitudes of IGABA measured at the three time points were normalized to the peak amplitude of IGABA evoked at 0 mV. There were no significant (*P > 0.06, n = 5) differences in the mean reversal potentials between the IGABA at the three time points in the PRIP-DKO PCs dialyzed with BAPTA (−55 ± 3, −53 ± 1, and −52 ± 2 mV). The residual amplitudes of outward IGABA evoked at holding potentials between −50 and 0 mV were calculated from the linear regression of the peak amplitudes of the inward IGABA evoked at holding potentials between −80 and −60 mV. The correlation coefficients of the residuals of the amplitudes of IGABA recorded with 10 mM EGTA and BAPTA were −0.97 ± 0.01 and 0.98 ± 0.02, respectively, which indicate significant inward and outward rectification, respectively ($ P < 0.01 and $ P < 0.01).

Full size image

Ca2+ imaging using fura-2

To measure [Ca2+]i changes in L3 PCs of the barrel cortex, slice preparations were incubated in N-ASCF containing 10 μM fura-2 AM (Dojindo Laboratories, Kumamoto, Japan) and 0.01 % Cremophor EL (Sigma-Aldrich, St. Louis, MO, USA) for 60 min at room temperature. Following incubation, the preparations were rinsed with dye-free N-ASCF for 30 min. The recording chamber was mounted on the stage of a fluorescence microscope (BX50WI-FL/DIC; Olympus) equipped with a high-speed cooled digital CCD camera (C6790-80; Hamamatsu Photonics, Hamamatsu, Japan), a high-speed switching light source (C7773; Hamamatsu Photonics), and filter units (U-MWU; DM400/BP330–385/BA420; Olympus). Preparations were observed under a water-immersion objective (LumiPlanFl × 40/0.80 or × 60/0.90; Olympus) and alternately illuminated at 340 or 380 nm. The fluorescence of 510 nm was captured by the CCD camera at a frame interval of 2 s with an exposure time of 10 ms and stored for offline analysis by an image-processing software (AquaCosmos; Hamamatsu Photonics). Ca2+ transients represent changes in the ratio (F 340/F 380) of the fluorescence at 510 nm excited by 340 nm to that excited by 380 nm as a measure of relative changes in [Ca2+]i. We made the ratiometric measurement only on the visually identified L3 PCs in slice preparations under Nomarski optics (Fig. 8Aa). In the respective slices, clear Ca2+ images were simultaneously captured in two to four cells (Fig. 8Ba or Da), which displayed similar Ca2+ transients when regions of interest were set on these cells (Fig. 8Bb, Db). A set of parameters were measured in these Ca2+ transients, among which a most typical one that had a set of standard parameter values was selected for pooled data analysis to examine statistical differences: n stands for the number of these typical cells as well as slices.

Drug application

GABAAR antagonists, 20–50 μM (+)-bicuculline or 100–200 μM picrotoxin, were bath applied. Fenvalerate (a calcineurin inhibitor) and caffeine (a ryanodine receptor agonist) were bath applied at 40 μM and 20 mM, respectively. Ruthenium red (an inhibitor of ryanodine receptor) and calcineurin were added to the internal solution at 100 μM and 0.2 U/ml, respectively. These chemicals except for calcineurin were obtained from Sigma-Aldrich. Calcineurin was obtained from Promega (Madison, WI, USA).

Simulation of IGABA

Simulation of IGABA was performed based on the kinetic model (Fig. 10A [18]). Mathematical calculations were done by using Mathematica 9 (Wolfram Research, Champaign, IL, USA). To simulate the GABA puff responses, the rate constants except for fast desensitization rate (d 2) and GABA-unbinding rate (k off) were set as follows (in s−1): k on = 15 μM−1, r 1 = 0.35, d 1 = 6, β 1 = 200, α 1 = 1,100, r 2 = 50, β 2 = 2,500, α 2 = 142, q = 1 × 10−8 μM−1, and P = 1. In the present simulation, d 2 was assumed to be dependent on the GABA concentration ([GABA]) as described in “Discussion.” The k off values in WT and PRIP-DKO GABAARs were set as 90 and 30 s−1, respectively, to simulate the GABA puff responses (Fig. 10C, D). We assumed that the maximum concentration of GABA applied by a puff through a pipette is reduced from 200 to 50 μM on the surface of the recorded cell and that the 2-s GABA puff pulse is attenuated with a time constant of 0.1 − 0.3 s (Fig. 10C, D, top).

Statistical analysis

Numerical data were expressed as the mean ± SD. The statistical significance was assessed using unpaired (†) or paired (‡) Student’s t test, ANOVA with Fisher’s protected least significant difference post hoc test (*), and single Pearson correlation coefficient ($). P < 0.05 was considered statistically significant.

Results

Enhancement of IGABA desensitization and subsequent emergence of a hump-like tail-I

First, we analyzed GABA currents (IGABA) evoked in L3 PCs by puff application of GABA to investigate whether and how the deletion of PRIP-1/2 affects kinetic properties of GABAARs. The IGABA in PRIP-DKO PCs displayed a more prominent desensitization compared to WT PCs (Fig. 1A, B). Surprisingly, PRIP-DKO PCs displayed a hump-like tail current (tail-I) at the puff offset (Fig. 1B) as if GABA acted as a partial antagonist [1, 2, 36].

The peak amplitude (a) and the amplitude at the puff offset (b) of the IGABA were measured in 13 WT and 12 PRIP-DKO PCs (Fig. 1A, B), while the peak amplitude of the hump-like tail-I following the puff offset (c) was measured in the 12 PRIP-DKO PCs (Fig. 1B). The mean value of [(a − b)/a] as a measure of the desensitization degree (Ds) was significantly († P < 0.01) larger in the PRIP-DKO PCs (0.43 ± 0.06) compared to the WT (0.19 ± 0.09) (Fig. 1C). This larger Ds in the PRIP-DKO PCs was accompanied by a longer half-duration, which is the duration at half amplitude of the tail-I at the puff offset (Fig. 1A, B, §). The mean value of the half-duration was significantly († P < 0.01) larger in the PRIP-DKO PCs (4.3 ± 0.5 s) compared to the WT (1.2 ± 0.4 s) (Fig. 1C), due to the presence of the hump-like tail-I. Furthermore, the mean duration of IGABA measured at amplitude of [(a + b)/2], i.e., the mean half-desensitization (half-Ds) time (Fig. 1A, B, #), was significantly († P < 0.01) shorter in the PRIP-DKO PCs (0.45 ± 0.10 s) compared to the WT (0.97 ± 0.10 s) (Fig. 1D). Similarly, this shorter half-Ds time in the PRIP-DKO PCs was accompanied by the longer half-duration of the tail-I. These results clearly indicate that the desensitization of IGABA was stronger in PRIP-DKO PCs than in WT PCs and suggest that in PRIP-DKO PCs the prominent hump-like tail-I is induced as a rebound response from the stronger desensitization of IGABA, consistent with the previously proposed idea that the desensitization of GABAARs is mechanistically related to the deactivation [18]. The hump-like tail-I amplitude (c − b) was normalized to the peak amplitude (a) of IGABA to obtain [(c − b)/a] as a measure of resensitization degree (Rs). Rs was 0.18 ± 0.05 in the PRIP-DKO PCs (Fig. 1B).

Effects of GABAAR antagonists on IGABA and hump-like tail-Is in PRIP-DKO PCs

Next, to investigate whether or not the hump-like tail-I as well as IGABA observed in PRIP-DKO PCs is carried through GABAAR-coupled Cl− channels, we examined the effects of picrotoxin on IGABA and its tail-Is evoked in PRIP-DKO PCs dialyzed with 5 mM EGTA (n = 6). Following bath application of 200 μM picrotoxin, the IGABA and hump-like tail-Is were progressively suppressed (Fig. 2Aa), and the small current that remained 11 min after application of picrotoxin was completely abolished by an addition of 20 μM bicuculline in the extracellular solution containing picrotoxin (bottom trace). These observations clearly indicate that the IGABA is generated by the activity of GABAARs. It is noteworthy that at 1–2 min after application of picrotoxin, the peak IGABA was largely suppressed in parallel with its tail-I (Fig. 2A). Indeed, the analysis of the progress of IGABA blockade by picrotoxin revealed a close relationship between the peak IGABA and tail-I. Changes in the amplitudes of IGABA following application of picrotoxin (Fig. 2Aa) were measured at the three time points: the peak time of the control IGABA (triangle), the offset of puff application (square), and the time point after the half of the half-duration of the control tail-I from the puff offset (circle), namely 1/2 half-duration, and these amplitudes were normalized to their controls. As shown in Fig. 2Ab, the amplitudes at the peak time and at the 1/2 half-duration were more markedly suppressed after application of picrotoxin, compared with that at the puff offset. This differential blockade by picrotoxin between the amplitudes of IGABA at its peak and the puff offset was consistent with the effects of picrotoxin on GABAA current in frog sensory neurons [16]. At the end of GABA puff, there are many GABAAR channels in the desensitized state, to which picrotoxin preferentially binds [35, 37]. Then, the probability of blockade of activated GABAAR channels by picrotoxin would become smaller as the desensitization progresses, consequently causing less prominent blockade at the puff offset compared to the peak IGABA. Because picrotoxin markedly suppressed the tail-I in parallel with the peak IGABA that undergoes desensitization, the desensitized GABAAR channels are likely to be resensitized at the offset of GABA puff. These results would indicate that the hump-like tail-I, as well as IGABA, is generated by the activity of GABAAR channels.

To further confirm the sole involvement of GABAARs in generating IGABA and its tail-I, muscimol, a GABAAR agonist, was puff-applied in PRIP-DKO PCs dialyzed with 5 mM EGTA. Similar to the responses to GABA puff, a hump-like tail-I was induced at the offset of 200 μM muscimol puff application (Fig. 2Ba). The mean Ds in muscimol-induced currents (IMus) was significantly smaller († P < 0.01, n = 5) than that in IGABA (Fig. 1C), and the mean half-Ds time in IMus was significantly longer († P < 0.01) compared to the IGABA (Fig. 1D). Furthermore, the mean Rs in IMus was significantly smaller († P < 0.01) compared to the IGABA, although there was no significant difference († P > 0.2, n = 5) in the mean half-duration of the tail-I between IMus and IGABA. Thus, IMus displayed less apparent desensitization and resensitization compared to IGABA, in spite of a similar desensitization rate between IMus and IGABA [20]. This may be because both the deactivation rate of GABAARs and the agonist unbinding rate become slower when GABAARs are activated by muscimol, compared to the case with GABA [20]. Indeed, our simulation study revealed that simulated IGABA displayed less desensitization and resensitization after decreasing the deactivation rates (α 1 and α 2) and unbinding rate (k off) (figure not shown).

We then examined the effects of bicuculline on the IMus and its tail-Is. Following bath application of 50 μM bicuculline, the IMus and its tail-I were progressively suppressed after application of bicuculline (Fig. 2Ba), and the small current remained 10 min after application of bicuculline was completely abolished by an addition of 100 μM picrotoxin in the extracellular solution containing bicuculline (bottom trace). Similar to the case with the blockade of IGABA by picrotoxin, the analysis of the progress of IMus suppression by bicuculline revealed a close relationship between the desensitization and resensitization of IMus (Fig. 2Bb). Taken together, these results strongly suggest the hump-like tail-Is as well as IGABA are mediated by the activity of GABAAR channels.

Reversal potentials of IGABA and hump-like tail-Is

It is still possible that the generation of the hump-like tail-I is secondary to GABAAR-mediated IGABA. To rule out this possibility and to confirm that both the hump-like tail-Is and IGABA are carried by Cl− through GABAAR channels, the reversal potentials of IGABA and its tail-Is were examined. The I–V relationships obtained in WT and PRIP-DKO PCs dialyzed with 10 mM EGTA were examined at three time points: the time point of peak IGABA (Fig. 3Aa, Ba, triangles), the offset of puff application (Fig. 3Aa, Ba, squares), and 1.5 s after the offset of puff application (Fig. 3Aa, circle) or the time point of peak tail-Is (Fig. 3Ba, 1.2 ± 0.5 s after the puff offset, n = 5; circle). In WT PCs dialyzed with 10 mM EGTA (n = 5), all of the I–V relationships measured at the three time points invariably displayed an outward rectification (Fig. 3Ab), consistent with the previous report [3]. The reversal potentials of the IGABA measured at the three time points were almost the same as the Cl− equilibrium potential (E Cl = −57 mV). In contrast, in PRIP-DKO PCs dialyzed with 10 mM EGTA (n = 5), the I–V relationship measured at the time point of the peak amplitude (Fig. 3Ba, triangle) displayed a significant inward rectification ($ P < 0.01, n = 5; Fig. 3Bb, triangles), while those measured at the puff offset and the time point of peak tail-Is (Fig. 3Ba, square and circle, respectively) displayed a slight inward and no apparent rectification, respectively (Fig. 3Bb, squares and circles, respectively). There were no significant (WT, *P > 0.1; PRIP-DKO, *P > 0.4) differences in the mean reversal potentials between the IGABA at the three time points in the WT (−56 ± 1, −53 ± 3, and −54 ± 4 mV, n = 5) and PRIP-DKO PCs (−57 ± 2, −54 ± 4, and −53 ± 5 mV, n = 5). All these reversal potentials were nearly equal to the E Cl, indicating that the hump-like tail-Is as well as IGABA are carried by Cl− through GABAAR channels. This result suggests that the rebound hump-like tail-I is mediated by an unusual reopening of GABAARs induced at the offset of GABA puff application.

In PRIP-DKO PCs dialyzed with 10 mM BAPTA, the IGABA recorded at 0 mV exhibited less prominent desensitization and was followed by little hump-like tail-Is (Fig. 3Ca), compared to those dialyzed with 10 mM EGTA (Fig. 3Ba). Indeed, the Ds (0.19 ± 0.05) and half-Ds time (0.94 ± 0.04 s) in PRIP-DKO PCs dialyzed with 10 mM BAPTA (n = 5) were significantly († P < 0.05) smaller and longer than the Ds (0.28 ± 0.04) and half-Ds time (0.70 ± 0.10 s) in those dialyzed with 10 mM EGTA (n = 5), respectively. The I–V relationships at the three time points in the IGABA recorded with 10 mM BAPTA were outwardly rectifying with the reversal potentials close to the E Cl (Fig. 3Cb). In contrast to the IGABA recorded with 10 mM EGTA, the I–V relationship of the peak IGABA recorded with 10 mM BAPTA displayed a significant outward rectification ($ P < 0.01, n = 5). Then, the inward rectification of the I–V relationship of the peak IGABA in PRIP-DKO PCs recorded with 10 mM EGTA is likely to be mediated by Ca2+-dependent desensitization of IGABA that became more pronounced as the holding potential was more depolarized to activate voltage-dependent Ca2+ channels. The difference in the rectification between PRIP-DKO PCs recorded with 10 mM EGTA and BAPTA appeared to be consistent with the difference in the desensitization between those. Furthermore, both the desensitization and reactivation of IGABA appeared to be dependent on [Ca2+]i. Thus, the reactivation or resensitization of IGABA appeared to be correlated with the desensitization degree of IGABA in a Ca2+-dependent manner.

Enhancement of the desensitization and resensitization of IGABA in PRIP-DKO PCs by increases in [Ca2+]i

Because the desensitization of GABAAR is well-known to be accelerated by increases in [Ca2+]i [33], the effects of [Ca2+]i on the desensitization of IGABA and the generation of tail-I in PRIP-DKO PCs were examined using two different pipette solutions that contained 5 or 10 mM EGTA. The IGABA evoked at 0 mV in PRIP-DKO PCs dialyzed with 5 mM EGTA displayed a more prominent desensitization and was followed by a more prominent hump-like tail-I compared to those dialyzed with 10 mM EGTA (Fig. 4A, B). With the repetition of puff applications of GABA for 2 s every minute, the desensitization of IGABA evoked in PCs dialyzed with 5 mM EGTA appeared to progress and the hump-like tail-I appeared to be enhanced, as shown in the superimposed traces of the fifth and the 20th IGABA (Fig. 4A). Such changes appeared to be less clear in PCs dialyzed with 10 mM EGTA (Fig. 4B). The mean Ds in the respective averaged responses of the fifth–seventh and the 20th–22nd IGABA were significantly and consistently larger († P < 0.01) in the PCs dialyzed with 5 mM EGTA (n = 6) compared to those dialyzed with 10 mM EGTA (n = 5) (Fig. 4C). Furthermore, the mean half-Ds time in the fifth–seventh and that in the 20th–22nd averaged IGABA was also significantly and consistently shorter († P < 0.01) in the PCs dialyzed with 5 mM EGTA than in those dialyzed with 10 mM EGTA (Fig. 4D). These results clearly indicate that the desensitization of IGABA observed in PRIP-DKO PCs was accelerated depending on free [Ca2+]i.

Fig. 4
figure 4

Effects of EGTA on the desensitization and resensitization of IGABA in PRIP-DKO PCs. A, B Sample traces of the fifth and the 20th IGABA evoked in PRIP-DKO PCs dialyzed with 5 mM EGTA (A) and dialyzed with 10 mM EGTA (B) by repeated puff application of GABA every minute. C The mean Ds in the fifth–seventh (triangles) and the 20th–22nd averaged IGABA (circles) evoked in the PRIP-DKO PCs dialyzed with 5 mM EGTA (filled symbols: 0.35 ± 0.07 and 0.46 ± 0.08, respectively, n = 6) and in those dialyzed with 10 mM EGTA (open symbols: 0.22 ± 0.04 and 0.29 ± 0.04, respectively, n = 5). † P < 0.01. D The mean half-Ds time in the fifth–seventh and the 20th–22nd averaged IGABA evoked in the PRIP-DKO PCs dialyzed with 5 mM EGTA (0.52 ± 0.07 and 0.44 ± 0.07 s, respectively, n = 6) and in those dialyzed with 10 mM EGTA (0.80 ± 0.07 and 0.73 ± 0.05 s, respectively, n = 5). † P < 0.01. E The relationship between the Ds and resensitization degree [Rs = (c − b)/a] in the fifth–seventh (triangles) and the 20th–22nd averaged IGABA (circles) evoked in the PRIP-DKO PCs dialyzed with 5 mM EGTA (Rs: 0.12 ± 0.04 and 0.20 ± 0.06, respectively, n = 6) and in those evoked in the PRIP-DKO PCs dialyzed with 10 mM EGTA (Rs: 0.03 ± 0.01 and 0.08 ± 0.01, respectively, n = 5). † P < 0.01. F The relationship between the Ds and half-duration of the tail-I in the fifth–seventh and the 20th–22nd averaged IGABA evoked in the PRIP-DKO PCs dialyzed with 5 mM EGTA (half-duration: 3.58 ± 0.46 and 4.30 ± 0.51 s, respectively, n = 6) and in those evoked in the PRIP-DKO PCs dialyzed with 10 mM EGTA (half-duration: 2.73 ± 0.42 and 3.37 ± 0.63 s, respectively, n = 5). † P < 0.02

Full size image

Because the probability of reopening or resensitization would increase as the number of desensitized GABAAR channels increases [18], the relationship between the degrees of desensitization and resensitization was examined by plotting the Rs (= (c − b)/a) against Ds. As seen in the scattergram (Fig. 4E), both the mean Ds and Rs in the fifth–seventh and the 20th–22nd averaged IGABA in the PCs dialyzed with 5 mM EGTA (filled triangle and circle) were significantly and consistently larger († P < 0.01) than in those in the PCs dialyzed with 10 mM EGTA (open triangle and circle), respectively. Thus, the larger Rs invariably accompanied the larger Ds observed in the PCs dialyzed with EGTA at the lower concentration, regardless of the trials of GABA puff applications. Similarly, the significantly larger († P < 0.02) mean half-duration of the hump-like tail-I invariably accompanied the larger Ds observed in the PCs dialyzed with EGTA at the lower concentration, regardless of the trials of GABA puff applications (Fig. 4F).

Taken together, these results clearly indicate that both the desensitization and resensitization of IGABA in PRIP-DKO PCs are dependent on [Ca2+]i. This suggests that [Ca2+]i may be higher in PRIP-DKO PCs than in WT PCs at a holding potential of 0 mV. Both the Ds and Rs increased after repeated application of GABA over 22 trials, probably due to an increase in [Ca2+]i during a long recording at 0 mV. This phenomenon was used as a control for the experiments described below.

Effects of changes in [Ca2+]o on IGABA in PRIP-DKO PCs

We next investigated the possible Ca2+-dependent relationship between the desensitization and resensitization of IGABA in PRIP-DKO PCs, by examining how the IGABA and its tail-I evoked at 0 mV in PRIP-DKO PCs dialyzed with 5 mM EGTA are altered following an increase in [Ca2+]o from 0.5 to 2 mM. As the WT PCs never displayed a hump-like tail-I regardless of EGTA concentration (figure not shown), only PRIP-DKO mice were examined in this experiment.

Whole-cell recordings were made from PRIP-DKO PCs in such slices that were pre-incubated for 30 min in the extracellular solution containing 0.5 mM Ca2+. IGABA was repeatedly evoked by seven trials of GABA puff application every minute in 0.5 mM [Ca2+]o and also repeatedly evoked by over 15 trials of GABA puff application every minute after [Ca2+]o was increased from 0.5 to 2 mM. The averaged IGABA evoked by the fifth–seventh GABA puffs at 0 mV in 0.5 mM [Ca2+]o displayed much less desensitization and was followed by a very small hump-like tail-I (Fig. 5Aa), in contrast to the fifth–seventh averaged IGABA evoked in 2 mM [Ca2+]o in the different sample groups (Fig. 4A). However, after [Ca2+]o was increased from 0.5 to 2 mM, the averaged IGABA evoked by the 13th–15th GABA puffs which correspond to the 20th–22nd ones after starting the whole-cell recording displayed a prominent desensitization and was followed by a prominent hump-like tail-I at the offset of the GABA puff (Fig. 5Aa), similar to the fifth–seventh averaged IGABA evoked at 2 mM [Ca2+]o in the different sample groups (Fig. 4A).

Fig. 5
figure 5

Effects of [Ca2+]o and fenvalerate on IGABA in PRIP-DKO PCs. A a Superimposed sample traces of the averaged IGABA evoked by the fifth–seventh GABA puffs at 0 mV in 0.5 mM [Ca2+]o and that evoked by the 13th–15th GABA puffs started after [Ca2+]o was increased from 0.5 to 2 mM in a PRIP-DKO PC dialyzed with 5 mM EGTA. A b, A c The changes in the Ds (A b) and half-Ds time (A c) in the averaged IGABA evoked by the fifth–seventh GABA puffs in 0.5 mM [Ca2+]o (open triangles) and that evoked by the 13th–15th GABA puffs in 2 mM [Ca2+]o (open circles). Gray triangles represent the respective mean values in the fifth–seventh averaged IGABA at 2 mM [Ca2+]o shown in Fig. 4C, D. ‡ P < 0.01, † P < 0.01. A d, A e The relationships between the Ds and Rs (A d) and between the Ds and half-duration of the tail-I (A e). Gray triangles represent the respective relationships in the fifth–seventh averaged IGABA at 2 mM [Ca2+]o shown in Fig. 4E, F. ‡ P < 0.01, † P < 0.01. B a Superimposed sample traces of the averaged IGABA evoked by the fifth–seventh GABA puffs in the control condition and that evoked by the 13th–15th GABA puffs started after application of 40 μM fenvalerate in a PRIP-DKO PC dialyzed with 5 mM EGTA. B b, B c The changes in the Ds (B b) and half-Ds time (B c) in the averaged IGABA evoked by the fifth–seventh GABA puffs in the control condition and that evoked by the 13th–15th GABA puffs after fenvalerate (Fen) application. Gray triangles and circles represent the mean values in the fifth–seventh and the 20th–22nd averaged IGABA at 2 mM [Ca2+]o, respectively, shown in Fig. 4C, D. ‡ P < 0.03. B d, B e The relationships between the mean Ds and Rs (B d) and between the mean Ds and half-duration of the tail-I (B e). Gray triangles and circles represent the relationships in the fifth–seventh and the 20th–22nd averaged IGABA at 2 mM [Ca2+]o, respectively, shown in Fig. 4E, F. ‡ P < 0.02

Full size image

As illustrated in Fig. 5Ab, Ac), the mean values of the Ds and half-Ds time in the averaged IGABA evoked by the fifth–seventh GABA puffs in 0.5 mM [Ca2+]o (open triangles: 0.08 ± 0.03 and 0.96 ± 0.12 s, respectively) were significantly († P < 0.01 and † P < 0.01, respectively, n = 7) different from those in the fifth–seventh averaged IGABA in 2 mM [Ca2+]o shown in Fig. 4C, D (gray triangles). However, when compared between these two averaged IGABA evoked before and after [Ca2+]o was increased from 0.5 to 2 mM, the mean Ds and half-Ds time were significantly (n = 7) increased to 0.35 ± 0.10 (‡ P < 0.01) and decreased to 0.59 ± 0.16 s (‡ P < 0.01), respectively (Fig. 5Ab, Ac), compare open triangles and circles). Consequently, the former and latter parameters in the averaged IGABA evoked in 2 mM [Ca2+]o corresponding to the 20th–22nd averaged IGABA became similar († P > 0.9 and † P > 0.6, respectively) to those in the fifth–seventh averaged IGABA in 2 mM [Ca2+]o (Fig. 5Ab, Ac), compare open circles and gray triangles). These results indicate that the desensitization of IGABA observed in PRIP-DKO PCs at 0.5 mM [Ca2+]o was weak but became prominent due to a [Ca2+]o increase to 2 mM, rather than due to the repetition of GABA puff application.

As seen in the scattergrams (Fig. 5Ad, Ae), the mean values of the Rs and half-duration in the averaged IGABA evoked by the fifth–seventh GABA puffs in 0.5 mM [Ca2+]o (open triangles: 0.04 ± 0.02 and 2.74 ± 0.62 s, respectively) were significantly († P < 0.01 and † P < 0.03, respectively, n = 7) smaller than those in the fifth–seventh averaged IGABA in 2 mM [Ca2+]o (gray triangles). However, when compared between these two averaged IGABA evoked before and after [Ca2+]o was increased from 0.5 to 2 mM, the mean Rs and half-duration were significantly (n = 7) increased to 0.11 ± 0.04 (‡ P < 0.01) and 3.72 ± 0.57 s (‡ P < 0.01), respectively (Fig. 5Ad, A e), compare open triangles and circles). Consequently, the former and latter parameters in the averaged IGABA evoked in 2 mM [Ca2+]o corresponding to the 20th–22nd averaged IGABA became similar († P > 0.9 and † P > 0.2, respectively) to those in the fifth–seventh averaged IGABA at 2 mM [Ca2+]o (Fig. 5Ad, Ae), compare open circles and gray triangles). Thus, the increase in the Ds was invariably accompanied by the increases in the Rs as well as half-duration in a manner dependent on [Ca2+]o or [Ca2+]i, consistent with the observation shown in Fig. 4E, F (compare open and filled symbols). Taken together, it can be concluded that both the desensitization and resensitization of IGABA at 0.5 mM [Ca2+]o were weak due to a low [Ca2+]o, but became prominent after increasing [Ca2+]o to 2 mM.

Effects of a calcineurin inhibitor on IGABA in PRIP-DKO PCs

Because the alterations of IGABA kinetics in PRIP-DKO PCs were mediated by high [Ca2+]i, we next investigated Ca2+-dependent signal transduction involved in the alterations of IGABA kinetics. It has been reported that the inhibition of calcineurin suppresses the desensitization of GABA currents in acutely dissociated hippocampal neurons [28] and increases the rate of GABA unbinding from GABAARs in cultured hippocampal neurons [19]. Therefore, we examined whether and how fenvalerate, an inhibitor of calcineurin, modulates IGABA in PRIP-DKO PCs dialyzed with 5 mM EGTA.

IGABA was repeatedly evoked by seven trials of GABA puff application in the control condition and also repeatedly evoked by over 15 trials of GABA puff after bath application of 40 μM fenvalerate. Following application of fenvalerate for at least 15 min, the desensitization of IGABA was alleviated, while the hump-like tail-I after the puff offset was markedly decreased, as observed in the superimposed traces of the averaged IGABA evoked by the fifth–seventh GABA puffs in the control condition and that evoked by the 13th–15th GABA puffs after fenvalerate application (Fig. 5Ba). When compared between these two averaged IGABA evoked before and after fenvalerate application, the mean Ds and half-Ds time were significantly (n = 5) decreased from 0.30 ± 0.08 to 0.13 ± 0.06 (‡ P < 0.03) and increased from 0.54 ± 0.11 to 0.85 ± 0.24 s (‡ P < 0.03), respectively (Fig. 5Bb, Bc). Both the Ds and half-Ds time in the averaged IGABA evoked by the fifth–seventh GABA puffs in the control condition were not significantly († P > 0.2 and † P > 0.6, respectively) different from those in the fifth–seventh averaged IGABA shown in Fig. 4C, D (Fig. 5Bb, Bc, compare open and gray triangles), whereas the changes in these parameters caused by fenvalerate were opposite to the changes caused by the repeated application of GABA puff (Fig. 5Bb, Bc, compare open and gray symbols). These results suggest that the desensitization of IGABA in PRIP-DKO PCs can be alleviated by calcineurin inhibition.

As seen in the scattergrams (Fig. 5Bd, Be), when compared between these two averaged IGABA evoked before and after fenvalerate application, the mean Rs and half-duration were significantly (n = 5) decreased from 0.17 ± 0.02 to 0.05 ± 0.03 (‡ P < 0.01) and from 3.93 ± 0.22 to 3.26 ± 0.25 s (‡ P < 0.02), respectively, along with the Ds. Thus, the decrease in the Ds was invariably accompanied by the decreases in the Rs as well as half-duration. These changes caused by fenvalerate were also opposite to the changes caused by the repeated application of GABA puff shown in Fig. 4E, F (Fig. 5Bd, Be, compare open and gray symbols). Taken together, these results suggest that the enhancement of desensitization of IGABA and its subsequent resensitization, at least partly due to a slowdown of GABA-unbinding rate, in PRIP-DKO PCs are mediated by the higher endogenous activity of calcineurin, which might have been brought about by the higher [Ca2+]i, compared to WT PCs.

Effects of calcineurin on IGABA in WT PCs

We next investigated whether or not calcineurin can induce a hump-like tail-I in WT PCs. In WT PCs dialyzed with 0.2 U/ml calcineurin and 5 mM EGTA (n = 6), the IGABA evoked 5 min after establishing the whole-cell configuration did not display a hump-like tail-I (Fig. 6Aa). However, 40 min after the whole-cell configuration, the normalized peak amplitude of IGABA decreased significantly (‡ P < 0.01) (Fig. 6Bb, open circles) and a hump-like tail-I emerged at the offset of the GABA puff (Fig. 6Aa). In contrast, in the WT PCs dialyzed with 5 mM EGTA alone (n = 6), the normalized peak amplitude of IGABA did not decrease markedly (‡ P > 0.1) when compared between those obtained 5 and 40 min after the whole-cell configuration (Fig. 6Bb, filled circles) and the IGABA never displayed a hump-like tail-I (Fig. 6Ba). The decrease in the normalized peak amplitude of IGABA in WT PCs dialyzed with calcineurin and EGTA may be largely due to the desensitization caused by the activity of calcineurin. However, there were no significant differences in the mean Ds (‡ P > 0.06) and half-Ds time (‡ P > 0.6) between the IGABA evoked 5 min and that evoked 40 min after the whole-cell configuration (Fig. 6Ab, Ac, open circles). This may be because GABAAR channels may had already been desensitized partly by the activity of calcineurin when the GABA puff was applied 40 min after the whole-cell configuration as reflected in the decrease in amplitude (Fig. 6Bb) and because the fraction of GABAAR channels available for further desensitization during the GABA puff might be small due to the limited activity of calcineurin brought about exogenously. However, the mean Rs and the mean half-duration were significantly (‡ P < 0.01) increased (Fig. 6Ad, Ae, open circles). Thus, the activation of calcineurin enhanced the desensitization of IGABA and slowed the GABA-unbinding rate (k off), consistent with the findings in the previous reports [19, 28]. In the WT PCs dialyzed with 5 mM EGTA alone, there were no significant differences in the mean Ds (‡ P > 0.8), the mean half-Ds time (‡ P > 0.4), and the mean half-duration (‡ P > 0.7) between the IGABA evoked 5 min and that evoked 40 min after the whole-cell configuration (Fig. 6Ab, Ac, Ae), filled circles). The mean half-duration of the tail-I was significantly († P < 0.01) longer in the WT PCs dialyzed with calcineurin and EGTA than in those dialyzed with EGTA alone when examined 40 min after the whole-cell configuration (Fig. 6Ae), reflecting the presence of hump-like tail-Is. Taken together, these results strongly suggest that calcineurin is really required to induce the hump-like tail-I.

Fig. 6
figure 6

Effects of calcineurin on IGABA in WT PCs. A a Superimposed sample traces of IGABA recorded 5 and 40 min after the whole-cell configuration in a WT PC dialyzed with 0.2 U/ml calcineurin and 5 mM EGTA. A b The mean Ds in IGABA recorded 5 and 40 min after the whole-cell configuration in WT PCs dialyzed with 5 mM EGTA alone (filled circles: 0.13 ± 0.03 and 0.13 ± 0.04, respectively, n = 6) and in those dialyzed with calcineurin and EGTA (open circles: 0.15 ± 0.05 and 0.23 ± 0.12, respectively, n = 6). A c The mean half-Ds time in IGABA recorded 5 and 40 min after the whole-cell configuration in WT PCs dialyzed with EGTA alone (filled circles: 1.03 ± 0.11 and 1.07 ± 0.06 s, respectively, n = 6) and in those dialyzed with calcineurin and EGTA (open circles: 1.06 ± 0.12 and 1.03 ± 0.10 s, respectively, n = 6). A d The change in the Rs in the IGABA recorded 5 and 40 min after the whole-cell configuration in WT PCs dialyzed with calcineurin and EGTA (0.001 ± 0.002 and 0.07 ± 0.04, respectively, n = 6). ‡ P < 0.01. A e The mean half-duration in IGABA recorded 5 and 40 min after the whole-cell configuration in WT PCs dialyzed with EGTA alone (filled circles: 1.49 ± 0.30 and 1.47 ± 0.30 s, respectively, n = 6) and in those dialyzed with calcineurin and EGTA (open circles: 1.34 ± 0.20 and 2.15 ± 0.33 s, respectively, n = 6). ‡ P < 0.01, † P < 0.01. B a Superimposed sample traces of IGABA recorded 5 and 40 min after the whole-cell configuration in a WT PC dialyzed with 5 mM EGTA alone. B b Normalized peak amplitudes of IGABA recorded 5 and 40 min after the whole-cell configuration in WT PCs dialyzed with calcineurin and EGTA (open circles) and in those dialyzed with EGTA alone (filled circles). ‡ P < 0.01, † P < 0.05

Full size image

Effects of ruthenium red on IGABA in PRIP-DKO PCs

The enhanced desensitization and resensitization of IGABA in PRIP-DKO PCs are likely to be mediated by the higher [Ca2+]i. It is possible that the increases in [Ca2+]i are brought about either by the potentiated CICR due to the impairment of IICR in the shared Ca2+ store as shown in PRIP-1 KO mice [9] and/or by the enhanced SOCE as demonstrated in PRIP-2 KO mice [40]. Therefore, we next investigated the possible involvements of CICR and the subsequent SOCE in inducing the desensitization and resensitization of IGABA, by examining whether and how IGABA evoked in PRIP-DKO PCs can be modulated by the intracellular application of ruthenium red that blocks CICR [10, 29]. With the repetition of GABA puff applications for 2 s every minute, the desensitization of IGABA evoked in PRIP-DKO PCs dialyzed with 100 μM ruthenium red and 5 mM EGTA was progressively alleviated while the hump-like tail-I after the puff offset was progressively suppressed, as observed in the superimposed traces of the fifth and the 22nd IGABA (Fig. 7A).

Fig. 7
figure 7

Enhanced desensitization and resensitization of IGABA in PRIP-DKO PCs are induced by CICR. A Sample traces of the fifth and the 22nd IGABA in PRIP-DKO PCs dialyzed with 100 μM ruthenium red and 5 mM EGTA. B, C The changes in the Ds (B) and half-Ds time (C) in the fifth–seventh (open triangles) and the 20th–22nd averaged IGABA (open circles). Gray triangles and circles represent the mean values in the fifth–seventh and the 20th–22nd averaged IGABA at 2 mM [Ca2+]o, respectively, shown in Fig. 4C, D. ‡ P < 0.02. D, E The relationships between the Ds and Rs (D) and between the Ds and half-duration (E) in the fifth–seventh (open triangles) and the 20th–22nd averaged IGABA (open circles). Gray triangles and circles represent the relationships in the fifth–seventh and the 20th–22nd averaged IGABA at 2 mM [Ca2+]o, respectively, shown in Fig. 4E, F. ‡ P < 0.02

Full size image

The mean Ds in the 20th–22nd averaged IGABA (0.17 ± 0.04) was significantly (‡ P < 0.02, n = 5) smaller than that in the fifth–seventh averaged IGABA (0.27 ± 0.05) (Fig. 7B), and the mean half-Ds time was significantly (‡ P < 0.01, n = 5) longer in the 20th–22nd averaged IGABA (0.65 ± 0.05 s) compared to the fifth–seventh averaged IGABA (0.50 ± 0.07 s) (Fig. 7C). Both the Ds and half-Ds time in the control condition were not significantly († P > 0.06 and † P > 0.6, respectively) different from those in the fifth–seventh averaged IGABA shown in Fig. 4C, D (Fig. 7B, C, compare open and gray triangles) whereas the changes in these parameters following intracellular diffusion of ruthenium red were opposite to the changes caused by the repeated application of GABA puff (Fig. 7B, C, compare open and gray symbols). These results clearly indicate that the desensitization of IGABA in PRIP-DKO PCs was markedly alleviated in the presence of intracellular ruthenium red. As a consequence of the alleviation of IGABA desensitization by ruthenium red, the Rs was significantly (‡ P < 0.02, n = 5) decreased from 0.13 ± 0.04 to 0.06 ± 0.03 along with the Ds following the repetition of puff applications of GABA from the fifth–seventh trials to the 20th–22nd trials (Fig. 7D). Thus, the decrease in Rs invariably accompanied the decrease in the Ds following intracellular diffusion of ruthenium red. Similarly, a significant shortening of the half-duration of the hump-like tail-I in the 20th–22nd averaged IGABA (3.57 ± 0.74 s) from that in the fifth–seventh averaged IGABA (4.02 ± 0.71 s) (Fig. 7E, ‡ P < 0.01, n = 5) accompanied the decrease in the Ds. These changes caused by ruthenium red were also opposite to the changes caused by the repeated application of GABA puff shown in Fig. 4E, F (Fig. 7D, E, compare open and gray symbols). These results suggest that the enhanced desensitization and resensitization of IGABA in PRIP-DKO PCs are largely induced by increases in [Ca2+]i through CICR and/or the subsequent SOCE. Although ruthenium red blocks not only CICR but also various Ca2+ channels such as TRP channels [43] and voltage-gate Ca2+ channels [6], ruthenium red at least blocks CICR. This is because the blockade of Ca2+ channels may result in the blockade of CICR. Provided that CICR is involved in the enhanced desensitization and the subsequent resensitization of IGABA observed in PRIP-DKO PCs but not in WT PCs, CICR should be more potent in PRIP-DKO PCs than in WT PCs. In the next experiment, we examined whether or not CICR and the subsequent SOCE differ between WT and PRIP-DKO PCs using a Ca2+-imaging method.

Apparent enhancement of CICR in PRIP-DKO PCs

Changes in [Ca2+]i were measured in visually identified L3 PCs in slice preparations incubated with 10 μM fura-2 AM (Fig. 8A). Usually, Ca2+ images were captured simultaneously in two to four cells in the same slices (Fig. 8Ba, Da). These cells displayed similar Ca2+ transients in response to a brief (3 min) application of high-K+ (20 mM) solution or a brief (3 min) application of 20 mM caffeine followed by a brief (3 min) application of the mixed solution of 20 mM K+ and 20 mM caffeine (the combined application of the caffeine and the mixed solutions). In a slice preparation obtained from PRIP-DKO mice, the two PCs in the same slice (Fig. 8Ba, single and double arrowheads) displayed similar large Ca2+ transients in response to the brief application of the high-K+ solution, while the combined application of the caffeine and the mixed solutions 5 min after washout of the preceding high-K+ solution induced only small increases in [Ca2+]i (Fig. 8Bb). A similar observation has been made in nine of 10 PRIP-DKO slices examined. In contrast, in the remaining one of the 10 slices, the brief application of high-K+ solution induced small Ca2+ transients while the subsequent combined application of the caffeine and the mixed solutions induced large Ca2+ transients simultaneously in the three PCs as represented by the three superimposed traces of similar Ca2+ transients (Fig. 8Bc). In a slice preparation obtained from WT mice, the brief application of the high-K+ solution induced similar large Ca2+ transients while the subsequent combined application of the caffeine and the mixed solutions induced only small increases in [Ca2+]i, in three PCs as represented by the three superimposed traces of Ca2+ transients (Fig. 8C). A similar observation has been made in eight of eight WT slices examined. The amplitude of Ca2+ transients was measured as a difference in the F 340/F 380 ratio from the baseline level. There was no significant († P > 0.9) difference in the mean peak amplitude of Ca2+ transients induced by the brief application of high-K+ solution between the PRIP-DKO (1.12 ± 0.10, n = 9) and WT PCs (1.11 ± 0.18, n = 8).

Fig. 8
figure 8

Comparison of CICR between WT and PRIP-DKO PCs. A a, A b Photomicrographs showing a Nomarski (A a) and fura-2 fluorescence image (A b, F 340) of the same two L3 PRIP-DKO PCs indicated with arrowheads. B a Sample F 340 and F 380 images of the two PCs (indicated with single and double arrowheads) in a PRIP-DKO slice obtained before (1) and at the peak time of CICR (2), which are indicated with 1 and 2 in the Ca2+ transients representing the F 340/F 380 ratio (B b). B b Ca2+ transients showing CICR in response to application of the high-K+ (20 mM) solution in the two PCs in a PRIP-DKO slice, which are indicated with single and double arrowheads in (B a). No CICR in response to the subsequent combined application of the caffeine and the mixed solutions. B c Small responses to the high-K+ solution followed by CICR induced in response to the subsequent combined application of the caffeine and the mixed solutions in three PCs in a PRIP-DKO slice. C CICR induced in response to the high-K+ solution followed by small responses to the subsequent combined application of the caffeine and the mixed solutions in three PCs in a WT slice. D a Sample F 340 and F 380 images of the three PCs (indicated with single and double arrowheads and arrow) in a PRIP-DKO slice obtained before (1) and at the peak time of CICR (2), which are indicated with 1 and 2 in the Ca2+ transients representing the F 340/F 380 ratio (D b). D b Ca2+ transients showing CICR in response to the combined persistent application of the caffeine and the mixed solutions in the three PCs in a PRIP-DKO slice, which are indicated with single and double arrowheads and arrow in D a. An asterisk indicates a small Ca2+ transient observed during the 20-mM caffeine solution alone period. E CICR induced in response to the combined persistent application of the caffeine and the mixed solutions in four PCs in a WT slice

Full size image

Provided that the first application of the high-K+ solution induced CICR, the subsequent combined application of the caffeine and the mixed solutions may have failed to induce CICR due to the preceding depletion of Ca2+ stores. Thus, a large Ca2+ transient is likely to be induced by CICR. Because the sole application of the high-K+ solution failed to induce CICR although in one of 10 PRIP-DKO slices, the high-K+ solution was applied in the presence of the caffeine solution to securely induce CICR. Indeed, the persistent application of the high-K+ solution that was preceded by 3 min by the persistent application of 20 mM caffeine invariably induced large Ca2+ transients. In a slice preparation obtained from PRIP-DKO mice, the three PCs in the same slice (Fig. 8Da, single and double arrowheads and arrow) displayed similar large Ca2+ transients in response to the combined persistent application the caffeine and the mixed solutions, which decayed to a plateau level (Fig. 8Db). Similar observation was made in WT PCs (Fig. 8E). Such a plateau of [Ca2+]i caused by the combined persistent application of the caffeine and the mixed solutions was maintained at least for 10 min examined (Fig. 8Db, De). The mean peak and plateau amplitudes of Ca2+ transients induced by the combined persistent application in the PRIP-DKO PCs (1.19 ± 0.09 and 0.32 ± 0.02, respectively, n = 6) were significantly († P < 0.03 and † P < 0.02, respectively) larger than those in the WT PCs (1.00 ± 0.13 and 0.22 ± 0.05, respectively, n = 5). However, 20 mM caffeine alone neither induced CICR nor affected the baseline level of [Ca2+]i in PCs in 11 of 17 PRIP-DKO slices examined and in 13 of 18 WT slices examined while it induced only a small Ca2+ transient in the remaining six PRIP-DKO and five WT slices, similar to a small [Ca2+]i increase observed during the caffeine alone period in the combined application of the caffeine and the mixed solutions (Fig. 8Db, *). These effects of caffeine were consistent with the previous report in cultured hippocampal neurons [38]. Because CICR is known to be followed by SOCE [7, 34], these results suggest that Ca2+ entry occurred following the depletion of Ca2+ stores by CICR [34].

Enhanced SOCE in PRIP-DKO PCs

In PRIP-DKO PCs, whether SOCE occurred following CICR or not was directly examined by reducing [Ca2+]o. The CICR was induced twice consecutively by the first and second applications of the mixed solution of 20 mM caffeine and 20 mM K+ (caffeine/high-K+ solution), which were separated by 35–40 min (Fig. 9Aa). Immediately after the first and second CICR, the caffeine/high-K+ solution was changed to the Ca2+-free extracellular solution and the standard extracellular solution containing 2 mM Ca2+, respectively. Compared to the decrease in the F 340/F 380 ratio observed during the application of standard extracellular solution containing 2 mM Ca2+ following the second CICR, the F 340/F 380 ratio appeared to decrease more steeply during the application of the Ca2+-free extracellular solution following the first CICR and increased immediately after [Ca2+]o was increased from 0 to 2 mM (Fig. 9Aa). The decay time course of the F 340/F 380 ratio was fitted with a double exponential curve and was compared between the responses obtained in the two extracellular solutions. The fast decay time constant (τ1) was significantly smaller (‡ P < 0.01, n = 6) in the responses to the change to 0 mM [Ca2+]o (38 ± 13 s) compared to those to 2 mM [Ca2+]o (127 ± 30 s) whereas there was no significant (‡ P > 0.6, n = 6) difference in the amplitude of the F 340/F 380 ratio (0.22 ± 0.07 and 0.21 ± 0.07, respectively) that decreases with the τ1 (Fig. 9Ab). These observations would clearly indicate that CICR was followed by SOCE.

Fig. 9
figure 9

Enhancement of SOCE in PRIP-DKO PCs. A a A rapid decay of the F 340/F 380 ratio following the first CICR after [Ca2+]o was decreased to zero and a slow decay following the second CICR when [Ca2+]o was maintained at 2 mM. The dotted gray curves represent double exponential curves that describe the decay of the F 340/F 380 ratio. A b An acceleration of τ 1 without changes in the initial amplitude following a decrease in [Ca2+]o from 2 (circle) to 0 mM (triangle). ‡ P < 0.01. B a, B b Two superimposed traces of Ca2+ transients obtained from two PCs in a WT slice (B a) and those of Ca2+ transients obtained from two PCs in a PRIP-DKO slice (B b) in response to the combined application of the caffeine and the mixed solutions which were washed out immediately after CICR. B c Superimposed traces of Ca2+ transients shown in B a and B b. After washout of the combined application of the caffeine and the mixed solutions, the F 340/F 380 ratio decayed differently between PCs in WT and PRIP-DKO slices. B d Superimposed scaled traces of Ca2+ transients recorded from WT and PRIP-DKO PCs, indicating a slower decay from the peak in the PRIP-DKO PC compared to the WT. Double-headed arrows represent the half-decay times from the peak of CICR. C The mean half-decay times from the peak CICR induced in response to the brief application of the high-K+ solution (brief K+), the combined persistent application of the caffeine and the mixed solutions (persistent caffeine/K+) and the combined application of the caffeine and the mixed solutions (brief caffeine/K+) in PCs in WT (filled columns) and PRIP-DKO (open columns) slices. † P < 0.02

Full size image

Next, we examined whether the SOCE following the potentiated CICR in PRIP-DKO PCs is more potent compared to the WT. When Ca2+ transients were induced by the combined application of the caffeine and the mixed solutions, it decayed differently between PCs in PRIP-DKO and WT slices (Fig. 9Ba–c) as reflected in the half-decay time (Fig. 9Bd, compare the respective double-headed arrows), to a low [Ca2+]i level from which it further decayed with a similar slow time course (Fig. 9Ba–c). The mean half-decay time from the peak CICR was significantly longer in the PRIP-DKO PCs than in the WT PCs, irrespective of the responses to the brief application of the high-K+ solution (WT 82 ± 18 s, n = 8; PRIP-DKO 117 ± 24 s, n = 9; † P < 0.01) or those to the combined persistent application of the caffeine and the mixed solutions (WT 82 ± 14 s, n = 6; PRIP-DKO 109 ± 9 s, n = 5; † P < 0.02) or those to the combined application of the caffeine and the mixed solutions (WT 76 ± 18 s, n = 6; PRIP-DKO 106 ± 14 s, n = 6; † P < 0.02) (Fig. 9C). Because SOCE occurs following CICR [34] as demonstrated in Fig. 8D, these results strongly suggest that both the CICR and the subsequent SOCE are more potent in PRIP-DKO PCs than in WT PCs.

Discussion

In the present study, we showed that the deletion of PRIP-1/2 accelerated the desensitization of IGABA but caused a hump-like tail-I at the offset of the GABA puff (Figs. 1, 4, and 5). The enhancement of desensitization and resensitization of GABAARs in PRIP-DKO PCs was mediated by the enhanced activity of Ca2+-dependent calcineurin through the potentiated CICR and SOCE. To the best of our knowledge, there has been no report on such hump-like tail-Is that are induced by GABA itself, although similar tail-Is were observed following the removal of anesthetic drugs such as propofol [36] and pentobarbital [1, 2].

[Ca2+]i dependence of IGABA desensitization and resensitization and their abolishment by a calcineurin inhibitor

It was clearly demonstrated that both the acceleration of IGABA desensitization and the generation of the hump-like tail-I were brought about by increases in [Ca2+]i, as revealed by changes in [Ca2+]o from 0.5 to 2 mM (Fig. 5A) and by changes in [EGTA]i from 10 to 5 mM (Fig. 4). Consistent with the idea that desensitization is mechanistically related to the deactivation in GABAARs [18], the progress of desensitization of IGABA was invariably accompanied by the enhancement of the hump-like tail-I (Figs. 4 and 5A). These results suggest that the deletion of PRIP-1/2 results in an enhancement of the desensitization and resensitization of GABAARs through increases in [Ca2+]i. The involvement of CICR and the following SOCE in both the desensitization of IGABA and the generation of the hump-like tail-I in PRIP-DKO PCs was also demonstrated by the intracellular application of ruthenium red (Fig. 7).

It has been demonstrated that the desensitization of GABAARs was enhanced by increases in [Ca2+]i [15, 33] and that a calcineurin inhibitor, cyclosporin A-cyclophilin A complex, suppressed the desensitization of GABA currents in acutely dissociated hippocampal neurons [28]. It has also been reported that the inhibition of calcineurin increased the rate of GABA unbinding from GABAARs [19]. Consistent with these previous studies, the bath application of fenvalerate alleviated the desensitization of IGABA and markedly decreased the hump-like tail-I in the present study (Fig. 5B). Thus, it is likely that the hump-like tail-I in PRIP-DKO PCs was generated as a result of an acceleration of desensitization of IGABA coupled with a slowdown of the GABA unbinding, which was mediated by Ca2+-dependent activation of calcineurin. In support of this idea, it was observed that an injection of calcineurin in WT PCs induced a hump-like tail-I (Fig. 6). Because CICR and the following SOCE were more potent in PRIP-DKO PCs than in WT PCs (Figs. 8 and 9), these results strongly suggest that the enhancement of desensitization and resensitization of GABAARs in PRIP-DKO PCs was largely mediated by the upregulation of Ca2+-dependent activity of calcineurin due to the potentiation of CICR followed by SOCE.

A possible mechanism for the higher [Ca2+]i in PRIP-DKO PCs

There is evidence that ryanodine and IP3 receptors share a common functional Ca2+ pool in cultured rat cerebellar granule [17], acutely dissociated rat hippocampal [14], rat cerebellar Purkinje [24], and rat neocortical neurons [27]. As the affinity of PRIP-1 for IP3 is higher than that of PRIP-2 [41], the hydrolysis of IP3 is suppressed more by PRIP-1 compared to PRIP-2. Therefore, the impairment of IICR was largely caused by the deletion of PRIP-1 in cultured cortical neurons [9]. Then, CICR may have been potentiated due to the impairment of IICR in the shared Ca2+ pool in PRIP-1 KO mice. On the other hand, it has also been reported that the SOCE was enhanced in hematopoietic B cells in PRIP-2 KO mice [40]. The SOCE that is regulated by the filling state of endoplasmic reticulum (ER) Ca2+ stores is an important Ca2+ influx pathway in non-excitable and excitable cells [7]. The Ca2+ sensor stromal interacting molecules (STIMs) localized in the ER sense the Ca2+ concentration of ER and activate Ca2+ release-activated Ca2+ (CRAC) channel upon the depletion of internal Ca2+ following IICR or CICR by binding to the CRAC protein 1 (CRACM1 or Orai1) [7]. This has also been demonstrated in cultured cortical neurons [8]. In view of these findings, the deletion of PRIP-2 may cause the higher activity of STIMs and Orai1 and thereby increase the activity of SOCE. Taken together, it is possible that the higher [Ca2+]i in PRIP-DKO PCs is mediated by the enhancement of CICR and the following enhanced SOCE through the deletion of PRIP-1 and PRIP-2, respectively. However, further studies are apparently required to clarify how the deletion of PRIP-2 affects the activity of STIMs and Orai1 in L3 PCs of the barrel cortex.

A possible kinetic mechanism underlying the generation of the hump-like tail-I

We simulated the IGABA followed by a hump-like tail-I using a model previously proposed (Fig. 10A [18]) to examine whether the possible increase in the fast desensitization rate (d 2) and the possible decrease in the unbinding rate (k off) can lead to a generation of the hump-like tail-I at the offset of the GABA puff. It has been reported that after GABAARs had been desensitized by application of GABA, GABA binding was greatly increased in the desensitized GABAARs compared to the non-desensitized GABAARs, and the binding increase was dependent on the concentration of the pre-applied GABA as was the case with the degree of desensitization of GABA currents [4]. Then, when the probability of being in the desensitized state (D fast) for GABAARs was increased by increasing GABA concentration ([GABA]) or during the 2-s puff application of GABA, D fast would be recruited more compared to Open2. Thus, it is likely that the d 2, but not β 2, would increase in a manner dependent on [GABA] (Fig. 10B). Because Bound2 that is bifurcated into Open2 and D fast increases in a manner dependent on [GABA] in the present model, we incorporated the idea in the present model by defining d 2 as follows:

$$ {d}_2=\frac{d_{\max }}{1+{\left(\frac{K_h}{\left[\mathrm{GABA}\right]}\right)}^n} $$

where d max is the maximum desensitization rate, K h is the [GABA] that yields the half maximum desensitization rate, and n is the Hill coefficient. The [GABA] dependence of d 2 can be enhanced by an increase in n and a decrease in K h (Fig. 10B). We assumed that calcineurin increased d 2 by increasing the [GABA] dependency of the desensitization rate and the d 2 and k off were changed between the simulated WT and PRIP-DKO PCs. These changes were comparable to those caused by activation of calcineurin reported previously [19, 28].

Fig. 10
figure 10

A kinetic model for a hump-like tail-I. A A kinetic model of GABAARs representing mono- and double-liganded states, each providing access to open and desensitized states. B Hypothesized GABA concentration ([GABA])-dependent changes of d 2 in WT and PRIP-DKO GABAARs. C, D Top: presumed [GABA] changes during puff application of GABA. Two hundred micromolars GABA applied by a rectangular pressure pulse was assumed to be diluted four times in the extracellular solution, and the onset and offset of the puff application were assumed to be attenuated with a time constant raging between 0.1 and 0.3 s. Bottom: superimposed traces of the simulated IGABA under the condition that the attenuation time constant is 0.3 and 0.1 s (solid and interrupted traces, respectively) in simulated WT (C) and PRIP-DKO PCs (D)

Full size image

When GABA puff was applied by a rectangular pressure pulse through a puff pipette containing 200 μM GABA, the ejected GABA was assumed to be diluted four times in the extracellular solution, and the onset and offset of the puff application were assumed to be attenuated with a time constant raging between 0.1 and 0.3 s. In the simulated WT PC where k off = 90, d max = 3,600, K h = 2,000 μM, and n = 1, IGABA was induced without a hump-like tail-I in response to 2-s GABA puff at 50 μM (Fig. 10C). In contrast, in the simulated PRIP-DKO PC where k off = 30 and d max were the same as those in WT, K h = 200 μM, and n = 1.5, IGABA displayed a prominent desensitization and was followed by a prominent hump-like tail-I (Fig. 10D). Thus, by assuming a slowdown of k off and an acceleration of d 2 through an increase in n and a decrease in K h, a hump-like tail-I could be induced. At the offset of GABA puff, a sharp decrease in d 2 to a level smaller than the fast resensitization rate constant (r 2) occurred to subsequently induce a hump-like tail-I. Decreases in the decay time constant at the offset of GABA puff pulse from 0.3 to 0.1 s decreased the half-duration of the hump-like tail-I, leaving its amplitude or Rs almost unchanged (Fig. 10D). Only PRIP-DKO PCs but not WT PCs displayed hump-like tail-Is in response to the same GABA puff that may have decayed slowly. These observations clearly indicate that the generation of the hump-like tail-I reflects kinetic differences between GABAARs in WT and PRIP-DKO PCs. Taken together, it was suggested that a higher calcineurin activity in PRIP-DKO L3 PCs might have caused a slowdown of k off and an acceleration of d 2 through the modulation of its GABA concentration dependency, leading to a generation of hump-like tail-Is in PRIP-DKO PCs.