Identification of a Novel Protein Complex Containing ASIC1a and GABAA Receptors and Their Interregulation

Acid-sensing ion channels (ASICs) belong to the family of the epithelial sodium channel/degenerin (ENaC/DEG) and are activated by extracellular protons. They are widely distributed within both the central and peripheral nervous systems. ASICs were modified by the activation of γ-aminobutyric acid receptors (GABAA), a ligand-gated chloride channels, in hippocampal neurons. In contrast, the activity of GABAA receptors were also modulated by extracellular pH. However so far, the mechanisms underlying this intermodulation remain obscure. We hypothesized that these two receptors-GABAA receptors and ASICs channels might form a novel protein complex and functionally interact with each other. In the study reported here, we found that ASICs were modified by the activation of GABAA receptors either in HEK293 cells following transient co-transfection of GABAA and ASIC1a or in primary cultured dorsal root ganglia (DRG) neurons. Conversely, activation of ASIC1a also modifies the GABAA receptor-channel kinetics. Immunoassays showed that both GABAA and ASIC1a proteins were co-immunoprecipitated mutually either in HEK293 cells co-transfected with GABAA and ASIC1a or in primary cultured DRG neurons. Our results indicate that putative GABAA and ASIC1a channels functionally interact with each other, possibly via an inter-molecular association by forming a novel protein complex.


Introduction
Acid-sensing ion channels (ASICs) belong to the family of the epithelial sodium channel/degenerin (ENaC/DEG) and are activated by extracellular protons [1]. They are widely distributed within both the central and peripheral nervous systems [2]. The activation of ASICs by protons induces sodium and/or calcium influx, giving rise to depolarization and evoking action potentials in neurons [3].Acid-sensing ion channels(ASICs) are associated with various physiological and pathophysiological functions including regulation of synaptic plasticity [4], perception of pain [5], ischemic death of neurons [6] and the termination of seizures [7]. ASICs were modified by the activation of c-aminobutyric acid receptors (GABA A ), a ligand-gated chloride channels, in hippocampal neurons [8]. In contrast, the activity of GABA A receptors were also modulated by extracellular pH [9][10][11]. However, the mechanisms underlying this intermodulation remain .unclear. Megan et. al. identified the a b subunit TM2 residue mediating proton modulation of GABA A receptors [12,13]. Huang et. al. reported that external protons regulated GABA A receptor function by direct or allosteric interaction with the GABA binding site [14]. But whether there was a direct binding site for proton within the GABA A receptor was so far unknown. We hypothesized that GABA A receptors and ASICs channels might form a novel protein complex and functionally interact with each other. In the study reported here, we found that ASICs were modified by the activation of GABA A receptors either in HEK293 cells following transient co-transfection of GABA A and ASIC1a or in primary cultured dorsal root ganglia (DRG) neurons. Conversely, activation of ASIC1a also modifies the current kinetics of GABA A current. Immunoassays showed that both GABA A and ASIC1a proteins were co-immunoprecipitated mutually either in HEK293 cells following transient co-transfection of GABA A and ASIC1a or in primary cultured DRG neurons. Our results indicate that putative GABA A and ASIC1a channels functionally interact with each other, possibly via an inter-molecular association by forming a novel protein complex. ASIC1a is specifically located in DRG neurons and function as a pain sensor, thus the interaction of GABA A and ASIC1a may contribute to pain sensation.

Activation of GABA A receptors inhibits ASIC1a currents in HEK293 cells
We used a whole-cell voltage-clamp configuration to record ASIC currents in HEK293 cells co-transfected with GABA A receptor subunits (a 1 and b 2 ) and ASIC1a in response to repeated application of a pH 6 solution. The peak amplitude of whole-cell ASIC currents (evoked with pH 6 solution) in HEK293 cells was stable, averaging 2.5160.37 nA (n = 38).Under our recording conditions the responses to GABA (at 100 mM) were small relative to ASICs currents (230619 pA, n = 27) due to the small driving force on chloride at 260 mV( Figure 1A). Application of GABA reversibly inhibited ASIC currents ( Figure 1A), which was largely abolished by application of a GABA A receptors antagonist (either bicuculline or picrotoxin) ( Figure 1B). To further confirm this phenomenon, we investigated whether GABA affected ASIC1a currents in HEK293 cells transfected with ASIC1a cDNA only. The result showed that GABA had no any effect on ASIC currents ( Figure 1C). To clarify whether this inhibition is pH-dependent, we tested the effect of GABA on ASIC currents evoked by lowered pH (#3.5). In general, the current evoked with pH 3.5 solution comprised of fast transient component and followed sustained component. Our results show that activation of GABA A receptors also attenuated the peak current amplitude but enhanced the sustained current evoked with pH 3.5 solution, such effect was eliminated when GABA A R was blocked or HEK293 cells was transfected with ASIC1a cDNA only ( Figure 2). These results suggested that activation of GABA A receptors strongly regulates ASIC1a currents.

Activation of ASIC1a modifies the current kinetics of GABA A current
Activation of ASIC1a reversibly altered the overall shape of GABA A currents in HEK293 cells co-transfected with GABA A receptor subunits (a 1 and b 2 ) and ASIC1a. Activation of ASIC1a had multiple effects on the GABA A currents, not only was the peak amplitude of the ASIC current enhanced, but also the kinetics of the GABA A currents were altered. Although activation of ASIC1a did not change the rise time (10%-90%) for the GABA A currents, the time for desensitization or deactivation of GABA A currents were markedly decreased when the pH of the extracelluar solution was decreased from 7.4 to 6. After washout, the time for desensitization and deactivation was totally recovered ( Figure 3A, n = 12). To exclude the direct role of proton on GABA A currents, we transfected HEK293 cells with GABA A receptor subunits only and did not obtain any current response to pH 6 solution although the peak amplitude of GABA A currents was also altered ( Figure 3B, n = 12).These data indicate that the functions of GABA A receptors are modified by ASIC1a.

Co-immunoprecipitation of ASIC1a and GABA A proteins in transfected HEK293 cells and primary cultured neurons
To investigate the underlying mechanisms of interregulation of ASIC1a and GABA A proteins, we transiently co-transfected ASIC1a and GABA A R in HEK293 cells. Due to endogenous expression of ASIC1a in HEK293 cells, we transfected ASIC1a with HA tag. In Co-IP experiments, anti-HA magnetic beads are used for the immunoprecipitation of specific HA-tagged proteins expressed in HEK293 cells. Our results showed that GABA A specifically co-precipitated with ASIC1a only in cells co-transfected with ASIC1a and GABA A , which was confirmed by reversed Co-IP using antibodies to GABA A R b 2 (Figure 4 A). ASIC1a endogenously expressed in HEK-293 cells [15]. Indeed, in our studies, we found that endogenous ASIC1a also co-precipitated with GABA A R in HEK293 cells transfect with GABA A R a 1 b 2 subunits. It is well known that DRG neurons expressed both ASIC1a and GABA A R. To examine the possible co-expression of endogenous ASIC1a and GABA A R, primary rat DRG neurons were incubated with specific anti-GABA A R and anti-ASIC1a antibodies. The merged image indicates that ASIC1a and GABA A R are co-segregated with each other (Figure 4 B1). To further investigate a possible association between ASIC1a and GABA A R proteins, GABA A R was immunoprecipitated from rat DRG lysates with a polyclonal anti-GABA A R b 2/3 antibody. The immunoprecipitated samples were probed with ASIC1a antibody. Conversely, the total DRG lysates was precipitated with ASIC1a antibody and then probed with GABA A R b 2/3 antibody (Figure 4 with GABA largerly abolished the GABA-induced inhibition of ASICs. C, GABA had no effect on ASIC1a currents in HEK293 cells transfected with cDNA of ASIC1a only. D, statistic graph shows relative ASIC currents that were affected by GABA but reversed by antagonists of GABA A receptors. n = 6, ***, p,0.001, T-test, before vs. after drug; ###, p, 0.001, one-way ANOVA, GABA plus GABA antagonists vs. GABA alone. doi:10.1371/journal.pone.0099735.g001 B2). Taken together, our results showed that ASIC1a and GABA A proteins co-immunoprecipitated each other.

Interregulation of GABA A receptors and ASIC1a in DRG neurons
Both GABA A receptors and ASIC1a channels colocalized in rat DRG neurons. To examine the interaction of endogenous ASIC1a and GABA A receptors in primary cultured rat DRG neurons, we  and ASIC1a, which can totally abolished by co-application of picrotoxin (100 mM) with GABA. ASIC1a current traces were superimposed to the right (inset) (B). C, GABA had no effect on ASIC1a currents in HEK293 cells transfected with ASIC1a cDNA only. ASIC1a current traces were superimposed to the right (inset). doi:10.1371/journal.pone.0099735.g002 Figure 3. Activation of ASIC1a modifies the current kinetics of GABA A current. A, The representative current traces recorded from HEK293 cells co-transfected with GABA A receptor subunits (a 1 and b 2 ) and ASIC1a. ASIC1a activated by pH 6 reversibly altered the overall shape of GABA A currents. GABA A current traces were superimposed to the right. Red arrow indicates the current activated by pH 6 solution. B, pH 6 solution changed the peak current amplitude but not the shape of GABA A currents in HEK293 cells transfected with cDNA of GABA A receptor subunits only. C (cotransfected with both plasmids) and D (transfected with cDNA of GABA A receptor subunits), bar graph showing the summarized data of rise time

Discussion
In the study reported here, activation of GABA A receptors strongly attenuates the peak current amplitude of ASIC currents in transfected HEK 293 cells. Conversely, Activation of ASIC1a modifies the current kinetics of GABA A current. These modifications included enhancement of the peak amplitude of GABA A current and slowing of channel kinetics. Similar effects were observed in primary cultured DRG neurons. Furthermore, ASIC1a is co-segregated with GABA A proteins in either transfected HEK 293 cells or rat DRG neurons, a finding verified by the immunoblotting assays. Our overall conclusion is that ASIC1a and GABA A interregulated each other through a conformation-dependent protein-protein interaction.

Interregulation of ASIC1a and GABA A receptors through conformation-dependent protein-protein interaction
Modification of ASIC1a by GABA A receptors occurred rapidly, and when the activation of GABA A chloride channels were blocked by pharmacological blockade or genetic loss, the modifications were eliminated or largely reduced. The ASIC of activation (10-90%) (i), desensitization time constant (ii) and deactivation time (iii) of GABA A currents in the presence of pH 7.4 or pH 6. ***, paired t-test, p,0.001, pH6 group vs. control group, n = 12. doi:10.1371/journal.pone.0099735.g003 currents also recovered rapidly (Figure1). This study suggests that ASIC1a current is modifed directly by activation of GABA A receptors. In 2011, Cheng et. al. found that ASICs were reversibly inhibited by activation of GABA receptors in murine hippocampal neurons and such inhibition of ASICs required opening of the chloride channels. Therefore these authors speculate that a conformation-dependent interaction might occur between GABA receptors and ASICs [16]. Our present studies further demonstrated that these two receptors form a novel protein complex by performing the Co-IP experiments. These two receptors physically couple together and interact with each other that may depend on their conformation change. However, so far we can not exclude the possibility that an intermediate protein or regulator might participate in this receptor-receptor interaction.
Furthermore, in our present studies, we also demonstrated that pH 6 extracelluar solution largely decreased the time for desensitization or deactivation of GABA A currents in HEK 293 cells co-transfected with ASIC1a and GABA A , but not in HEK293 cells transfected with GABA A cDNA only, suggesting that these two receptors combine with each other and regulate each other. External protons regulate GABA A receptor function by direct or allosteric interaction with the GABA binding site [14]. So far, we can not exclude the possibility that GABA A receptor may have a proton binding site. In summary, upon binding of GABA, the GABA A receptors undergo a conformational change and then modify current kinetics of ASIC1a by allosteric interaction with the proton binding site. Conversely, when ASIC1a is activated, the ASIC1a channels undergo a conformational change, then such change is converted to GABA A receptors. Our results indicate that putative GABA A receptors and ASIC1a channels phyically couple and functionally interact with each other, possibly via an intermolecular association.
Physiological and pathophysiological implications of the interaction between ASIC 1a and GABA A receptors of all ASICs, ASIC1a appears to play a prominent role in determining current amplitude and also affects the kinetics of H +gated current [4,[17][18][19]. In CNS neurons, ASIC1a has been shown to be involved in synaptic plasticity, learning and memory [4,20], and in acidosis-mediated, glutamate-independent neuronal injury [6,21]. ASIC1a is expressed throughout the brain, with prominent expression in areas that receive rich synaptic input [17,20,22,23]. Moreover, ASIC1a have a higher expression in GABAergic interneurons than that in the principal neurons [7,24]. Given that the GABA A receptors are the predominant inhibitory ionotropic receptors in the CNS, the interaction between ASIC1a and GABA A receptors may occur at numerous locations and could be involved in a number of brain functions. Furthermore, ASIC1a is specifically located in DRG neurons and functions as a pain sensor, thus the interaction of GABA A and ASIC1a may contribute to pain sensation.

Transfection of HEK293
HEK 293 cells were cultured in DMEM (HyClone) supplemented with 10% fetal bovine serum (HyClone), 1% penicillin/ streptomycin at 37uC in a 5% CO 2 incubator. Cells were transfected with pcDNA3.0 constructors encoding ASIC1a and/ or GABA A R a 1 b 2 , using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's instructions. All recordings were made 24 to 48 h after transfection in the GFP-positive cells.

DRG Cell Isolation and Culture
All the animal experiments were approved by the Medical Ethics Committee of Shandong University (number ECAES-DUSM 2012029). Adult Wistar rats were euthanized by cervical dislocation and the entire spinal columns were removed. Bilateral DRGs were collected and washed twice with L-15 medium (Gibco, Gaithersburg, MD). They were then incubated in 10 ml L-15 medium containing 10 mg collagenase type 1 (Sigma, St. Louis, MO) and 0.25 ml 0.25% Trypsin (HyClone, Thermo scientific, USA) at 37uC for 50 min. DRGs were removed from the enzyme solution, centrifuged for 5 min at 1,000 revolutions/min, washed twice with L-15 medium, and transferred to 2 ml L-15 medium containing 10% FBS. The ganglia were triturated with a suction pipe for 3-min and then centrifuged for 50 seconds at 1,000 revolutions/min. Supernatants were placed into 35 mm diameter Petri dishes. The cells were then cultured at 37uC in a 5% CO2 incubator (Thermo Forma, Hamilton, NJ, USA). In our study, we chose freshly isolated neurons from rat DRGs in the range of 15-30 mm diameter to test the effect of GABA on acid-evoked currents or pH on GABA-induced current.

Electrophysiological Recordings
Whole-cell voltage-clamp and current-clamp recordings were performed at room temperature (22-25uC) using a computer amplifier (Multiclamp 700B; Axon, New York, NY, USA) and a Digidata (1440A; Axon). Patch pipettes were filled with intracellular solution including (in mM): KCl 140, MgCl 2 2.5, HEPES 10, EGTA 11 and Na 2 ATP 5 with pH adjusted to 7.2 using KOH. Cells were bathed in extracellular saline containing (in mM): NaCl 150, KCl 5, CaCl 2 2.5, MgCl 2 2, HEPES 10, D-glucose 10 with pH adjusted to 7.4 using NaOH. The resistance of the recording pipettes was in the range of 5-8 MV. The series resistance was compensated for 70-80% after establishing a whole-cell configuration. The membrane potential was held at 260 mV throughout the recordings unless otherwise specified. Current-clamp recordings were obtained by switching to current-clamp mode after a stable whole-cell configuration was formed in the voltage-clamp mode. In this experiment, only cells with a stable resting membrane potential (less than 250 mV) were used. Signals were filtered at 4 kHz and then digitized at 10 kHz. The data were analyzed with the pCLAMP 10 acquisition software (Axon Instruments, CA, USA).

Co-immunoprecipitation (Co-IP)
For Co-IP in human embryonic kidney 293 (HEK293) cells, Lipofectamine 2000 (Invitrogen) was used to transiently transfect ASIC1a with HA tag and/or GABA A R a 1 b 2 following the manufacturer's instructions. Cell protein was purified for ASIC1a and GABA A R expression 36 h post-transfection. HEK293 cells were lysed in 1 ml of lysis buffer (1% Triton X-100, 50 mM Tris buffer pH 7.5) with a freshly added protein inhibitor mixture tablet. For Co-IP, the protein complexes were immunoprecipitated by anti-HA beads agarose (Sigma). After incubating, the beads were pelleted and washed three times in lysis buffer (1% Triton X-100, 50 mM Tris buffer pH 7.5), and the samples were loaded and run in 8% SDS-PAGE gel. The precipitates were transferred to a PVDF membrane and immunoblotted by anti-GABA A R b 2 antibody (Millipore). The blots were developed by the enhanced chemiluminescence kit. Rat DRG or HEK293 cells were lysed in lysis buffer (1% Triton X-100, 50 mM Tris buffer pH 7.5) and a mixture of protease inhibitors. Anti-ASIC1a antibody or anti-GABA A R b 2/3 antibody was immunoprecipitated with protein A/ G-agarose beads (Santa Cruz Biotechnology). The protein complexes were immunoprecipitated with antibody cross-linked protein A/G-agarose beads. After incubating the lysates with cross-linked antibody, the beads were pelleted and washed three times in lysis buffer (1% Triton X-100, 50 mM Tris buffer pH 7.5), and the samples were loaded and run in 8% SDS-PAGE gel. The precipitates were transferred to a PVDF membrane and immunoblotted by anti-GABA A R b 2/3 antibody (Millipore) or anti-ASIC1 antibody (Sigma). The blots were developed by the enhanced chemiluminescence kit.

Drug Application
All drugs were purchased from Sigma-Aldrich Corp. (St. Louis, MO, USA). Except picrotoxin (dissolved in DMSO), all drugs were initially made up as stock solutions in distilled water and subsequently diluted in the external solution of the cells at a maximum of 1:1000 to achieve their final working concentrations.

Data Analysis
Data were expressed as mean 6 SEM and compared statistically using paired t tests by Sigma Plot 10.0. A p,0.05 was required for the results to be considered statistically significant.