Abstract
GABAA receptors (GABAARs) mediate the majority of fast inhibitory transmission throughout the brain. Although it is widely known that pore-forming subunits critically determine receptor function, it is unclear whether their single-channel properties are modulated by GABAAR-associated transmembrane proteins. We previously identified Shisa7 as a GABAAR auxiliary subunit that modulates the trafficking, pharmacology, and deactivation properties of these receptors. However, whether Shisa7 also regulates GABAAR single-channel properties has yet to be determined. Here, we performed single-channel recordings of α2β3γ2L GABAARs cotransfected with Shisa7 in HEK293T cells and found that while Shisa7 does not change channel slope conductance, it reduced the frequency of receptor openings. Importantly, Shisa7 modulates GABAAR gating by decreasing the duration and open probability within bursts. Through kinetic analysis of individual dwell time components, activation modeling, and macroscopic simulations, we demonstrate that Shisa7 accelerates GABAAR deactivation by governing the time spent between close and open states during gating. Together, our data provide a mechanistic basis for how Shisa7 controls GABAAR gating and reveal for the first time that GABAAR single-channel properties can be modulated by an auxiliary subunit. These findings shed light on processes that shape the temporal dynamics of GABAergic transmission.
SIGNIFICANCE STATEMENT Although GABAA receptor (GABAAR) single-channel properties are largely determined by pore-forming subunits, it remains unknown whether they are also controlled by GABAAR-associated transmembrane proteins. Here, we show that Shisa7, a recently identified GABAAR auxiliary subunit, modulates GABAAR activation by altering single-channel burst kinetics. These results reveal that Shisa7 primarily decreases the duration and open probability of receptor burst activity during gating, leading to accelerated GABAAR deactivation. These experiments are the first to assess the gating properties of GABAARs in the presence of an auxiliary subunit and provides a kinetic basis for how Shisa7 modifies temporal attributes of GABAergic transmission at the single-channel level.
Introduction
GABAA receptors (GABAARs) are pentameric ligand-gated ion channels assembled from 19 different pore-forming subunits and are critical for providing fast inhibitory transmission throughout the brain (Mody and Pearce, 2004; Farrant and Nusser, 2005; Olsen and Sieghart, 2009). Following their assembly, these receptors are trafficked to the cell surface and are distributed to either synaptic or extrasynaptic regions, and contribute to phasic and tonic GABAergic inhibition, respectively (Farrant and Nusser, 2005; Martenson et al., 2017; Lorenz-Guertin and Jacob, 2018). GABAARs harbor distinct binding sites and serve as targets for several endogenous ligands and drug classes, including benzodiazepines, barbiturates, neurosteroids, general anesthetics, and alcohol (Olsen, 2018; Scott and Aricescu, 2019; Castellano et al., 2021; Kim and Hibbs, 2021). Therefore, insights into mechanisms that regulate GABAAR activity are fundamental toward understanding their functional role at inhibitory synapses.
Native receptor complexes are composed of pore-forming subunits together with auxiliary subunits that alter channel expression and function (Yan and Tomita, 2012; Maher et al., 2017; Han et al., 2021). Several GABAAR-associated transmembrane accessory proteins, which include Lhfpl4 (also known as GARLH4), Clptm1, and Shisa7, recently have been discovered (Davenport et al., 2017; Yamasaki et al., 2017; Ge et al., 2018; Wu et al., 2018; Han et al., 2019; Wu et al., 2021). Shisa7 is a single-pass transmembrane protein widely expressed in several brain regions including the cortex, striatum, thalamus, central amygdala, and hippocampus. We have recently identified Shisa7 as a GABAAR auxiliary subunit (Han et al., 2019), although it could also interact with AMPARs (Farrow et al., 2015; Schmitz et al., 2017). Specifically, Shisa7 interacts with both synaptic (α1, α2, γ2) and extrasynaptic (α5) GABAAR subunits and regulates both phasic and tonic GABAergic inhibition in hippocampal pyramidal neurons (Han et al., 2019; Wu et al., 2021, 2022). Shisa7 also modulates the trafficking, pharmacology, and kinetic properties of GABAARs (Han et al., 2019; Wu et al., 2021). In particular, Shisa7 accelerates the deactivation of two major synaptic GABAAR subtypes, α1β2γ2 and α2β3γ2, and deletion of Shisa7 slows the decay of IPSCs in CA1 hippocampal neurons (Han et al., 2019). Paired recordings between hippocampal cholecystokinin (CCK)-expressing interneurons and CA1 pyramidal neurons showed slower decay of CCK–CA1 unitary IPSCs (Han et al., 2019). While Shisa7 appears to play a critical role toward modulating the temporal dynamics of GABAergic transmission, the underlying mechanisms for how Shisa7 accelerates GABAAR deactivation are currently unknown.
Upon agonist binding, GABAARs transition between several conformational states that lead to channel opening. Single-channel recordings in various neuronal preparations (Macdonald et al., 1989a, b; Twyman et al., 1990; Brickley et al., 1999; Lorez et al., 2000) and in heterologous cells expressing GABAAR subtypes (Lema and Auerbach, 2006; Keramidas and Harrison, 2008, 2010; Dixon et al., 2014) have provided much detail regarding the importance of pore-forming subunits toward regulating GABAAR conductance and gating. However, the emergence of recently identified GABAAR transmembrane accessory proteins brings into question whether they can also modulate the biophysical properties of GABAARs.
In this study, we sought to investigate whether Shisa7 modulates GABAAR single-channel properties for α2β3γ2 GABAARs, a major synaptic subtype found in hippocampal and striatal regions (Goetz et al., 2007; Nguyen and Nicoll, 2018). We found that while Shisa7 does not change GABAAR slope conductance (IGABA), it reduced the frequency of channel openings. Single-channel analyses revealed that Shisa7 alters the duration, open probability (Po), and mean close and open times within GABAAR bursts by changing the time, but not the proportion, spent within specific close and open states during gating. Furthermore, Shisa7 reduced gating efficacy by increasing occupancy of the final close state preceding channel openings. Last, simulating GABA currents demonstrate that Shisa7 accelerates GABAAR deactivation through alterations in single-channel burst duration and Po. Together our findings show that Shisa7 is a GABAAR auxiliary subunit capable of modulating single-channel gating properties and highlight the importance of understanding how auxiliary subunits can control the biophysical characteristics of native GABAARs.
Materials and Methods
Cell lines.
HEK293T cells (catalog # CRL-3216, ATCC) were maintained with culture media containing 1% penicillin-streptomycin (Thermo Fisher Scientific), 10% FBS (Thermo Fisher Scientific) in DMEM (Thermo Fisher Scientific) in a humidified incubator at 37°C with 5% CO2.
Plasmids.
Human α2 (pcDNA3.1-Zeo+; NM_000807.3) γ2L (pcDNA3.1-Zeo+; NM_198903.2), and Shisa7 (pIRES2-EGFP; NM_001145176.2) plasmids were generated by GenScript Biotech. pIRES2-EGFP vector was a gift from Roger Nicoll's laboratory at the University of California, San Francisco. Human β3 (pcDNA3.1-Zeo+; NM_000814.6) was a gift from Joseph Lynch's laboratory at The University of Queensland (Brisbane, QLD, Australia).
Cell transfections.
HEK293T cells were seeded onto eight 12 mm coverslips that were plated on a 6 cm dish and transfected with human α2, β3, and γ2L GABAAR subunits with either pIRES2-EGFP or pIRES2-Shisa7-EGFP at a ratio of 1:1:3:3 (total cDNA, 2 µg), respectively. Plasmids were mixed in 94 µl of 37°C-warmed DMEM without FBS and penicillin-streptomycin. Six microliters of FuGENE HD Transfection Reagent (Promega) was added and vortexed to obtain a 3:1 reagent-to-DNA ratio. After 15 min, to allow transfection reagent–DNA complex formation, the mixture was added dropwise to the cells.
Single-channel recordings.
Cell-attached, single-channel recordings were made from GFP-positive cells 24 h post-transfection at room temperature (20–22°C). Current signals were acquired using an Axopatch 200B amplifier (Molecular Devices) at a sampling rate of 100 and 10 kHz Bessel low-pass filter. The data were digitized using a Digidata 1440B digitizer using pClamp 10 (Molecular Devices). Coverslips containing transfected HEK293T cells were perfused with an external solution containing the following (in mm): 102.7 NaCl, 20 Na-gluconate, 2 KCl, 14 d-glucose, 15 sucrose, 10 HEPES, 20 TEA-Cl, 1.2 MgCl2, 2 CaCl2, at pH 7.4 with NaOH and ∼320–330 mOsm. The internal solution was similar to the external solution, with the addition of either 3 µm or 10 mm GABA (catalog #0344, Tocris Bioscience). Sylgard 184 (Dow Corning) was applied to the tip of each pipette, heated using a VT-750C Varitemp heat gun, and fire polished using a microforge (catalog #MF-830, Narishige). After obtaining cell-attached mode, a pipette potential of +100 mV was applied to visualize single-channel events. To measure GABAAR single-channel slope conductance, the pipette holding potential was changed in 25 mV increments from +100 to −100 mV, and opening amplitudes were measured at each voltage step. Borosilicate filamented glass pipettes (outer diameter, 1.5 mm; inner diameter, 0.87 mm; Sutter Instrument) were pulled using a P-97 puller (Sutter Instruments). The pipette tip resistance ranged from 10 to 25 MΩs.
Single-channel kinetic analysis.
The QuB software was provided by Gabriela Popescu's laboratory at the State University of New York at Buffalo and was used for kinetic analysis of single-channel bursts. Care was taken to analyze patches that only displayed minimal overlapping opening events (<1% of total duration analyzed) at the beginning of the recording. Patches that exhibited ∼5000–15,000 events were used for kinetic analysis. After correction for baseline drift, single-channel data were idealized into close or open events using a segmental K-means algorithm with a cutoff resolution of 70 µs. Clusters of single-channel bursts were selected by eye before being separated into lists of discrete bursts for each patch. We performed maximum likelihood (MIL) fitting to obtain critical shut time (τcrit) values that were used to identify individual bursts. The τcrit value was determined for each patch to preserve the three shortest close (C) and the three shortest open (O) states that were consistent across patches. A linear, single-gate way 3C3O model scheme that displayed the largest log likelihood (ΣLL) value for both control (Ctrl) and Shisa7 conditions was used for each patch. This particular arrangement has also been previously used to model GABAAR activation (Lema and Auerbach, 2006; Keramidas and Harrison, 2010; Dixon et al., 2014). MIL estimation was used to obtain single-channel burst kinetic properties, dwell time distributions, and the rate and equilibrium constants during gating.
Modeling and simulation of macroscopic currents.
The QuB software was used to simulate macroscopic GABA currents. Specifically, close and open states, along with the average rate constants between them were arranged in the aforementioned 3C3O activation model. GABA binding (K+1) and unbinding (K–1) rate constants were determined by globally fitting the 3C3O average rate constants obtained from saturating GABA (10 mm) to subsaturating (3 µm) data files. The rate constants within the 3C3O model were set to fixed, and two ligand-dependent, close-binding steps were appended to C2 (see Fig. 5A). After running MIL to obtain the equilibrium dissociation constant [Kd (K–1/K+1)] values from global fitting, each data file containing single-channel recordings made under saturating concentrations were opened and appended with the same two ligand-dependent, close-binding steps along with the obtained K+1 and K–1 values. To simulate the macroscopic GABA current, a simulation protocol consisting of a single 1 ms application and 10 mm GABA was implemented. The number of channels simulated was set to 1000 and the response setting was set to deterministic. The simulation data files were exported to Clampfit 11.2, and monoexponential and biexponential fits were used to calculate the mean rise and weighted tau decay times, respectively.
Experimental design and statistical analyses.
Statistical analyses were performed using GraphPad Prism 9.0. All experiments were repeated at least three times. Shapiro–Wilk tests were used to assess normal distributions before between-group statistical comparisons. Two-tailed Mann–Whitney U tests were performed to assess for significant differences between control and Shisa7 conditions.
Results
Shisa7 does not change GABAAR single-channel slope conductance
We previously reported that Shisa7 increases GABA-evoked, whole-cell currents for several GABAAR subtypes (Han et al., 2019; Wu et al., 2021). Although this effect has been attributed to enhanced surface receptor trafficking (Han et al., 2019; Wu et al., 2021), it is unknown whether Shisa7 can also increase GABAAR unitary conductance. We performed cell-attached, single-channel recordings for one of the major synaptic GABAAR subtypes, α2β3γ2L, in HEK293T cells in the presence or absence of Shisa7 to determine whether Shisa7 changes IGABA. A subsaturating concentration of 3 µm GABA was included in the patch pipette, which elicits GABAAR single-channel opening and closing events (Fig. 1A). We find that α2β3γ2L GABAARs exhibit a slope conductance of 25 ± 0.5 pS (Fig. 1B,C), similar to previous observations of other heteropentameric GABAARs subtypes (Lema and Auerbach, 2006; Dixon et al., 2014). In the presence of Shisa7, the slope conductance is 23 ± 0.3 pS, which is not significantly different compared with Ctrl (n = 9; Shisa7, n = 9; U = 32, p = 0.49, Mann–Whitney U test, two tailed; Fig. 1B,C). Thus, although Shisa7 regulates GABAAR trafficking, pharmacology, and channel kinetics, it does not affect the single-channel slope conductance of α2β3γ2L receptors.
Shisa7 does not change GABAAR slope conductance. A, Representative traces of cell-attached recordings of α2β3γ2L GABAARs cotransfected with either GFP (Ctrl) or Shisa7 under different patch potentials (Vp). 3 μM GABA was included intrapipette and the amplitude of single-channel openings were recorded at different membrane potentials ranging from +100 to −100 mV. B, Plot depicting the current–voltage relationship between α2β3γ2L cotransfected with either Ctrl (n = 9) and Shisa7 (n = 9). C, Bar graph depicting the α2β3γ2L GABAAR slope conductance between Ctrl and Shisa7 (p = 0.49, Mann–Whitney U test, two tailed). Error bars indicate the SEM.
Shisa7 reduces the frequency of GABAAR single-channel opening events
Under subsaturating GABA concentrations, GABAAR single-channel opening events present as either brief, isolated openings or bursts of openings (Lema and Auerbach, 2006; Mortensen and Smart, 2007; Dixon et al., 2014). To examine whether Shisa7 alters single-channel opening frequency, we quantified the number of openings from both event types using cell-attached recordings from cells exposed to 3 µm GABA at a holding potential of +100 mV. In α2β3γ2L GABAARs, both control and Shisa7 display single-channel currents with an amplitude of ∼2 pA at this holding potential (Fig. 2A). We observe that Shisa7 strongly reduces the frequency of all single-channel opening events by ∼65% (Ctrl, n = 8; Shisa7, n = 10; U = 12, p = 0.0117, Mann–Whitney U test, two tailed; Fig. 2B). Further analysis of these events shows that Shisa7 reduces both isolated (Ctrl, n = 8; Shisa7, n = 10; U = 14, p = 0.0205, Mann–Whitney U test, two tailed; Fig. 2C) and burst opening (Ctrl, n = 8; Shisa7, n = 10; U = 5, p = 0.0009, Mann–Whitney U test, two tailed; Fig. 2D) events by 64% and 69%, respectively. Together, these results demonstrate that Shisa7 reduces the frequency of single-channel openings in α2β3γ2L GABAARs under subsaturating conditions.
Shisa7 reduces the frequency of GABAAR single-channel opening events. A, Representative traces of cell-attached recordings of α2β3γ2L GABAARs cotransfected with either Ctrl or Shisa7 in the presence of 3 μm GABA intrapipette and held at +100 mV. B–D, Bar graphs depicting the frequency of all GABAAR opening events (B), including isolated openings (C) and burst openings (D). Ctrl, n = 8; Shisa7, n = 10; Mann–Whitney U test, two tailed; *p < 0.05; ***p < 0.005. Error bars indicate the SEM.
Shisa7 decreases GABAAR burst duration and Po
We have also previously shown that Shisa7 can accelerate GABAAR deactivation kinetics (Han et al., 2019). However, a mechanistic basis for this phenomenon has not been identified. At the microscopic level, it has been demonstrated that GABAAR deactivation kinetics are largely controlled by the duration and Po of receptor burst events (Jones and Westbrook, 1995; Dixon et al., 2014). To determine whether Shisa7 regulates single-channel burst kinetics, we analyzed burst activity elicited by 10 mm GABA in cell-attached configuration. Given that two GABA binding sites reside on GABAARs (Baumann et al., 2003; Masiulis et al., 2019), full occupancy of these sites using a saturating GABA concentration enables us to negate the effects of agonist binding during the gating process (Lema and Auerbach, 2006; Mortensen and Smart, 2007). Thus, under these conditions, single-channel activity exhibits as quiescent periods that represent receptor desensitization and burst events that reflect fully bound receptors. After idealizing single-channel burst events into close or open states, we observe at least three modes of burst activity based on Po (Fig. 3A), as described in previous studies (Lema and Auerbach, 2006; Dixon et al., 2014; Brodzki et al., 2020) in both control and Shisa7 conditions. As we were interested in understanding how burst kinetics determine GABAAR deactivation, our analyses pooled all observed modes of bursts together. We find that Shisa7 decreases the activation of α2β3γ2L GABAARs by reducing both the burst duration (Ctrl, n = 10; Shisa7, n = 12; U = 18, p = 0.0044, Mann–Whitney U test, two tailed) and Po (Ctrl, n = 10; Shisa7, n = 12; U = 5, p < 0.0001, Mann–Whitney U test, two tailed; Fig. 3B). Furthermore, Shisa7 increases the burst mean close time (MCT; Ctrl, n = 10; Shisa7, n = 12; U = 28, p = 0.0358, Mann–Whitney U test, two tailed) while also reducing burst mean open time (MOT; Ctrl, n = 10; Shisa7, n = 12; U = 17, p = 0.0034, Mann–Whitney U test, two tailed; Fig. 3B). Together, these findings demonstrate that in addition to decreasing opening frequency, Shisa7 can alter GABAAR single-channel burst properties.
Shisa7 decreases GABAAR burst duration and Po. A, Representative traces of single-channel clusters of bursts elicited by 10 mm GABA in α2β3γ2L GABAAR cotransfected with either Ctrl or Shisa7 (top), with the blue bar indicating representative bursts from both conditions (bottom). B, Bar graphs depicting mean duration, mean Po, MCT, and MOT of single-channel bursts (Ctrl, n = 10; Shisa7, n = 12; Mann–Whitney U test, two tailed; *p = 0.05, **p = 0.005, ****p < 0.0001). Error bars indicate the SEM.
Shisa7 alters close and open dwell time components during bursts
Several linear models of GABAAR activation during bursts have been explored previously (Lema and Auerbach, 2006; Keramidas and Harrison, 2010; Dixon et al., 2014) using different schematic arrangements of the close and open states. To further understand how Shisa7 modulates GABAAR single-channel burst kinetics, we used a 3C3O activation model scheme (see Fig. 5A) that displayed the best ΣLL value for each patch between control and Shisa7 conditions. This scheme has been used previously to model GABAAR activation during bursts (Lema and Auerbach, 2006; Keramidas and Harrison, 2010; Dixon et al., 2014).
Analysis of the three close components reveals that Shisa7 increases the time constants of the two shortest close states without affecting the proportion spent in any of the three close states (Ctrl, n = 10; Shisa7, n = 12; Mann–Whitney U tests; Fig. 4A,C,D). In contrast, Shisa7 decreases the time constants of the two longest open states without also altering open state proportions (Ctrl, n = 10; Shisa7, n = 12; Mann–Whitney U tests; Fig. 4B,E,F). Given that the time constants of the two shortest close states and the two longest open states operate within microsecond and millisecond timescales, respectively, the decrease in burst Po by Shisa7 is most likely driven by the reduction in the duration of long open states, with a smaller contribution also arising from brief close states. Together, these results suggest that Shisa7 changes GABAAR burst properties by altering the dwell time constants operating in particular close and open states during the gating process.
Shisa7 alters close and open dwell time components during bursts. A, Close dwell time histograms of Ctrl (top) and Shisa7 (bottom) fitted with three exponential functions (black dotted lines) and the composite density function (black solid line). B, Open dwell time histograms of Ctrl (top) and Shisa7 (bottom) fitted with three exponential functions (red dotted lines) and the composite density function (red solid line). C, Mean time constants of the three close components. D, Percentage of the three close components. E, Mean constants of the three open components. F, Percentage of the three open components. Ctrl, n = 10; Shisa7, n = 12; Mann–Whitney U test, two tailed; *p < 0.05, **p < 0.005, ***p < 0.0005. Error bars indicate the SEM.
Kinetic modeling of GABAAR bursts indicates that Shisa7 reduces the efficacy of channel opening
Next, we determined the forward and reverse rate constants between state transitions for each patch using the aforementioned 3C3O scheme arrangement (Fig. 5A, Table 1). These were used to obtain equilibrium constants of state transitions within bursts. We find that compared with control, Shisa7 reduces E1 (Ctrl, n = 10; Shisa7, n = 12; U = 23, p = 0.138, Mann–Whitney U test, two tailed) and E3 (Ctrl, n = 10; Shisa7, n = 12; U = 10, p = 0.0004, Mann–Whitney U test, two tailed), which represent transitions from the final close state directly proceeding channel opening states (Fig. 5B, Table 2). We do not find significant differences between control and Shisa7 in the preactivation step, ϕ, a short-lived, close–close state transition that ultimately progresses toward channel activation (Gielen et al., 2012; Kaczor et al., 2021; Dixon et al., 2014; Gielen and Corringer, 2018, Lape et al., 2008), or in Σ, a close–close transition (Lema and Auerbach, 2006) that leads the channel farther away from opening (Ctrl, n = 10; Shisa7, n = 12; p > 0.05, Mann–Whitney U tests, two tailed; Fig. 5B) Furthermore, there are no significant differences in E2 (Ctrl, n = 10; Shisa7, n = 12; U = 43, p = 0.2829, Mann–Whitney U test, two tailed; Fig. 5B), representing the occupancy of the shortest open state and is consistent with no changes in time or the proportion spent in the presence of Shisa7 for this state (Fig. 4E,F). Together, these findings demonstrate that Shisa7 shortens α2β3γ2L GABAAR burst duration by reducing the efficacy of direct transitions from close to open states.
Kinetic modeling and simulation of GABAAR bursts reveal that Shisa7 increases occupancy of the final close state to accelerate receptor deactivation. A, A linear, single-gateway model consisting of three close states (C1–3; black) and three open states (O1–3; red) was selected to model and simulate α2β3γ2L GABAAR burst activation. Two GABA-binding close steps (purple) were added to estimate Kd values for macroscopic current simulations. Cu, Unliganded close state; Cm, monoliganded close state. Green arrow depicts GABA binding. Forward and reverse rate constants are shown (blue). B, Bar graphs depicting the mean equilibrium constants between Ctrl and Shisa7. C, A 1 ms stimulation pulse of 10 mm GABA (top) was used to simulate macroscopic GABAAR currents (bottom) using rate constants from Ctrl or Shisa7. D, Bar graphs depicting representative simulated rise time (left) and weighted deactivation (right) between Ctrl and Shisa7. E, A working model depicting Shisa7-dependent regulation of macroscopic GABA currents. Ctrl, n = 10; Shisa7, n = 12; Mann–Whitney U test, two tailed; *p < 0.05, ***p < 0.0005, ****p < 0.0001. Error bars indicate the SEM. N, Number of GABAARs present at the surface.
Rate constants for α2β3γ2L GABAAR bursts
Equilibrium constants for α2β3γ2L GABAAR bursts
Model-based simulations reveal that Shisa7 accelerates synaptic current deactivation through changes in burst kinetics
Finally, we sought to simulate macroscopic synaptic currents using our kinetic model with the rate constants obtained from single-channel bursts during gating. We performed global fitting of single-channel recordings obtained under subsaturating GABA concentrations to obtain GABA binding (K+1) and unbinding (K–1) rate constants (Table 1). This allowed us to estimate the Kd values for the control and Shisa7 conditions, which yielded relatively similar Kd values of 38.2 and 51.6 µm, respectively (Table 2).
We appended two ligand-dependent binding steps to our 3C3O activation model (Fig. 5A) and the K+1 and K–1 values obtained from the global fitting were added to each saturating data file. We used a simulation protocol consisting of a 1 ms application of 10 mm GABA (Fig. 5C), which allowed us to determine the time constants for the rise and deactivation time of simulated macroscopic current by fitting monoexponential and biexponential functions, respectively. We find that although Shisa7 does not change the rise time (Ctrl, n = 10; Shisa7, n = 12; U = 53, p = 0.6744, Mann–Whitney U test, two tailed) of the macroscopic current, it decreases the weighted deactivation time of the simulated GABA current (Ctrl, n = 10; Shisa7, n = 12; U = 0, p = 0.000003, Mann–Whitney U test, two tailed; Fig. 5D). Altogether, these findings demonstrate that changes to receptor burst gating kinetics mediated by Shisa7 are sufficient to recapitulate accelerated GABAAR deactivation, as previously shown (Han et al., 2019).
Discussion
Channel gating of ionotropic receptors plays a critical role in determining the strength and duration of synaptic transmission (Lester et al., 1990; Keramidas and Harrison, 2010; Dixon et al., 2014). In AMPA-type glutamate receptors, the regulation of channel properties by transmembrane auxiliary subunits has been extensively investigated. Indeed, several auxiliary subunits such as the TARP (transmembrane AMPAR regulatory protein) and cornichon family have been shown to modulate AMPAR single-channel properties (Jackson and Nicoll, 2011; Coombs and Cull-Candy, 2021). In contrast, it remains unknown whether GABAAR auxiliary subunits could directly modulate GABAAR function at the single-channel level. As IPSC kinetic profiles have been shown to play an important role in dynamic temporal processes (Xie and Manis, 2014), it is important to understand the factors that influence GABAAR gating.
We have recently shown that Shisa7, a single-pass transmembrane protein, interacts directly with GABAAR subunits and regulates both phasic and tonic GABAergic inhibition (Han et al., 2019; Wu et al., 2021). In addition to controlling GABAAR surface abundance, Shisa7 speeds up the deactivation kinetics of synaptic GABAAR subtypes such as α1β2γ2 and α2β3γ2, and the loss of Shisa7 slows the decay of IPSCs in hippocampal neurons (Han et al., 2019). Interestingly, Shisa7 also enhances benzodiazepine potentiation of submaximal GABA responses, and global deletion of Shisa7 reduces the anxiolytic and sedative properties of diazepam in vivo (Han et al., 2019). While these findings demonstrate the importance of Shisa7 toward GABAAR function, it remains unclear whether GABAAR single-channel properties are also modulated by Shisa7.
Previous work characterizing the single-channel properties of GABAARs has mainly focused on differences between distinct receptor subtypes (Macdonald et al., 1989a, b; Twyman et al., 1990; Lema and Auerbach, 2006; Barberis et al., 2007; Keramidas and Harrison, 2008, 2010; Dixon et al., 2014), critical domains involved in gating (Kash et al., 2003; Szczot et al., 2014; Janve et al., 2016; Hernandez et al., 2017; Kisiel et al., 2018; Brodzki et al., 2020; Nors et al., 2021), and the impact of disease-related mutations in pore-forming subunits on receptor function (Janve et al., 2016; Hernandez and Macdonald, 2019). However, it has yet to be explored whether GABAAR-associated, transmembrane accessory proteins could also alter GABAAR single-channel properties.
Here, we demonstrate that while Shisa7 did not change GABAAR single-channel slope conductance, it decreased the frequency of single-channel openings. More importantly, Shisa7 modulated GABAAR single-channel kinetics by decreasing both the duration and open probability of bursts. Shisa7 also increased and decreased the burst MCT and MOT, respectively. Further kinetic analyses of individual close and open dwell components within bursts demonstrates that Shisa7 changes the time, but not the proportion, of certain states during gating. Determining the equilibrium constants between state transitions during gating suggests that Shisa7 decreases the ability to keep GABAARs open by increasing the occupancy of the final close state preceding openings, without changes to any direct close–close or open–open state transitions. Since estimation of the equilibrium dissociation constant revealed relatively similar Kd values between GABAARs in the presence or absence of Shisa7, this indicates that the main effect of Shisa7 on channel activity is largely due to its effects on receptor gating following activation. Moreover, simulating macroscopic currents demonstrates that Shisa7 accelerates GABAAR deactivation, consistent with our recent report (Han et al., 2019).
Based on our data, we propose a working model describing Shisa7-dependent regulation of GABAARs in response to GABA activation (Fig. 5E). GABAARs exhibit larger macroscopic whole-cell currents in the presence of Shisa7. This effect is primarily because of enhanced surface trafficking of GABAARs (Han et al., 2019; Wu et al., 2021), as Shisa7 does not appear to alter IGABA. However, by shortening the duration of the two longest open states during bursts and by increasing the occupancy of the final closed state, Shisa7 decreases the Po and duration of single-channel bursts, which lead to accelerated GABAAR deactivation kinetics (Han et al., 2019). These properties would result in GABAAR-mediated currents that are both larger and faster during phasic inhibition in the presence of Shisa7 (Han et al., 2019).
It is worth mentioning that some GABAARs examined in isolation, such as α1-containing and α2-containing receptors, typically exhibit a much slower deactivation in heterologous cells compared with IPSCs mediated by these receptors in hippocampal pyramidal neurons (Picton and Fisher, 2007; Han et al., 2019). Our findings that Shisa7 can accelerate GABAAR deactivation through decreasing the receptor burst duration and Po, and that slower decay kinetics of IPSCs were previously observed in Shisa7 KO hippocampal neurons (Han et al., 2019) indicate a key physiological role for Shisa7 in regulating the temporal profile of fast inhibitory transmission. Currently, it remains unknown whether Shisa7 would regulate single-channel gating of other types of GABAARs in a manner similar to that we report here. The role of Shisa7 in the regulation of IPSC kinetics in other types of neurons where Shisa7 is expressed is also unclear. It will therefore be important in the future to systematically examine the function of Shisa7 in modulating GABAAR kinetics in a subunit- and neuron-specific manner. In summary, our data provide a mechanistic understanding for Shisa7-dependent regulation of GABAAR function at the single-channel level and shed new light on processes that control GABAAR gating and inhibitory transmission.
Footnotes
This work was supported by the National Institutes of Health/National Institute of Neurological Disorders and Stroke Intramural Research Program (to W.L.). We thank Dr. Joseph Lynch for providing us with the human β3 GABAAR subunit. We also thank Drs. Gabriela Popescu and Gary Iacobucci for providing the QuB software. In addition, we thank Drs. Stefano Vicini, Marek Brodzki, Gary Iacobucci, and Robert Pearce for helpful discussions regarding kinetic analysis of single-channel data and simulation of macroscopic currents.
The authors declare no competing financial interests.
- Correspondence should be addressed to Wei Lu at luw4{at}mail.nih.gov
