Inhibition of delta-1 glutamate receptor current by extracellular protons

Summary

Decreases in brain pH are associated with numerous neurological and neuropsychiatric conditions yet the molecular mechanisms linking decreased brain pH with these disorders are incomplete. The ionotropic glutamate receptors (iGluRs) mediate the majority of excitatory neurotransmission in the brain and are inhibited by extracellular protons; however, the proton sensitivity of the delta-glutamate receptor subclass of iGluRs is unknown. Using whole-cell patch-clamp recordings of serotonin neurons in mouse brain slices and activating alpha 1-adrenergic receptors to induce delta 1 glutamate receptor (GluD1_R) current, we demonstrated that GluD1_R current is inhibited by physiological drops in extracellular pH. Unlike other iGluRs, protons inhibited GluD1_R current via a voltage-independent decrease in unitary current. Moreover, mice lacking GluD1_R showed impaired behavioral responses to inhalation of CO₂. Taken together, this study continues to expand on the growing body of evidence positing GluD1_R as functional ion channels and suggests that GluD1_R facilitate pH sensing in vivo.

Introduction

Small deviations (0.05 units) from neutral blood pH (7.4) have profoundly negative consequences to brain health^1,2. These deviations can arise from numerous causes including stroke, heart attack, seizures, head trauma, pneumonia and other chronic lung diseases, obstructive sleep apnea, pain, stress, and anxiety-related hyperventilation and can cause long-lasting cognitive impairment. Recently, it has been proposed that decreases in brain pH contribute to cognitive decline in aging, which are exacerbated in Alzheimer’s disease³, and neuropsychiatric symptoms in bipolar disorder and schizophrenia⁴; however, the molecular mechanisms that sense physiologically relevant shifts in pH and create long-lasting alterations in neuronal excitability is incomplete. For example, non-inactivating voltage-gated potassium channels carry persistent current but are only affected at very lower pH (<6.5⁵. Acid-sensing ion channels (ASICs) are activated by physiologically relevant drops in pH but rapidly desensitize^6-8 often carrying only a small sustained current⁹. Thus, there is a critical need to identify the primary mode by which neurons convert a change in pH to a change in the excitability of the neural network over long periods of time. Such a characterization may reveal novel, druggable targets to mitigate the negative consequences of a decrease in brain pH, irrespective of the cause.

The ionotropic glutamate receptors (iGluRs), α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPA_R), N-methyl-D-aspartate receptors (NMDA_R), and kainate receptors (kainate_R) are all sensitive to inhibition by extracellular protons, with IC₅₀ values to proton inhibition of approximately 6.2, 7.3, and 7.4, respectively^10-12. Far less is known about the delta glutamate receptors (GluD1_R and GluD2_R). The proton sensitivity of GluD_R has only been examined using the constitutively open mutant ‘Lurcher’ GluD2_R, which appear similar to NMDA_R¹³. This is largely due to the difficulty of studying the ion channel function of GluD_R as there is no known ligand that can gate these receptors in heterologous expression systems^14-17, except recent evidence to support GluD2_R gating by D-serine and GABA when inserted in artificial lipid bilayers¹⁸. However, numerous studies using acute mouse brain slices have shown that GluD_R are functional ion channels that can be modulated by G protein-coupled receptor (GPCR) activity^19-22. In 2020, we discovered that synaptic activation of G protein-coupled α1-adrenergic receptors in dorsal raphe serotonin neurons produces a long-lasting excitatory postsynaptic current via coupling to GluD1_R channels (α1-A_R-GluD1_R current)^21,23. We and others have reported that GluD1_R are open at a low-level and carry a reproducible ‘tonic’ current under basal recording conditions^21,24,25. Thus, in native brain tissue, GluD1_R carries a persistent current that alters neuronal excitability over long periods of time. Moreover, human GluD1_R variants have been linked to bipolar disorder, schizophrenia, alcohol use disorder, autism spectrum disorder, and major depressive disorder²²; however, the extent and mechanism by which GluD1_R variants contribute to these disorders is largely unknown.

Here, using whole-cell patch-clamp electrophysiology, we found that synaptic α1-A_R-GluD1_R current and tonic GluD1_R current were inhibited by extracellular protons, suggesting that GluD1_R may function in vivo as a pH chemosensor. Indeed, global deletion of GluD1_R in mice reduced the behavioral response to elevated levels of CO₂.

Results

α1-A_R-GluD1_R current is inhibited by extracellular protons

Whole-cell voltage-clamp (V_hold −65 mV) electrophysiological recordings were obtained from serotonin neurons in the dorsal raphe nucleus in acute mouse brain slices at 35° C using a potassium-based internal solution with NMDA_R (MK-801, 5 μM), AMPA_R and Kainate_R (NBQX, 3 μM), GABA-A_R (picrotoxin, 100 μM), and 5-HT1A_R (WAY-100635, 300 nM) antagonists in the external solution. A train of electrical stimuli (5 pulses, 60 Hz delivered to the brain slice via a monopolar stimulating electrode every 90s) was used to reliably evoke synaptic α1-A_R-GluD1_R current, as previously described^21,23,24. The pH of the extracellular (pH_e) bicarbonate-buffered artificial cerebral spinal fluid (aCSF) was changed from the standard 7.4 to 6.8 by increasing the concentration of saturated CO₂ from 5% to 20%. Application of pH_e 6.8 reduced the α1-A_R-GluD1_R current from −52±7 pA to −21±3 pA and was readily reversed (Figure 1A-C, p<0.001). To compare the rates of inhibition and recovery of the α1-A_R-GluD1_R current to the real-time change in pH_e, pH_e of the bulk solution was measured using carbon fiber fast-scan cyclic voltammetry²⁶ (scan rate 40 V/s). The inhibition of the α1-A_R-GluD1_R current was well-aligned with the decrease in pH_e, whereas the return to pH_e 7.4 preceded full recovery of the α1-A_R-GluD1_R current by ~4.5 minutes (Figure 1C). The kinetics of the α1-A_R-GluD1_R current were also changed significantly by pH_e 6.8, with the time constant of activation (τ-activation) increasing from 359±37 ms in pH_e 7.4 to 771±135 ms in pH_e 6.8 (Figure 1D-E, p=0.006). The time constant of decay (τ-decay) was inversely affected by protons, decreasing from 8±1 s in pH_e 7.4 to 6±1 s in pH_e 6.8 (Figure 1F, p=0.046). In a prior report, we demonstrated that the amplitude of the α1-A_R-GluD1_R current does not affect the rates of activation and decay²³, so it is unlikely the reduction in amplitude by pH_e 6.8 underlies the change in time course.

Figure 1. Extracellular protons inhibit the α1-A_R-GluD1_R current..

(A) Representative traces showing the α1-A_R-GluD1_R current in pH_e of 7.4 (black), 6,8 (teal), and recovery in 7.4 (gray). (B) Quantification of the α1-A_R-GluD1_R current amplitude while acidifying the bath recording solution and recovery upon returning to 7.4 pH_e (7.4 v 6.8, p<0.001, n=14; 7.4 v 7.4, p=0.379, n=10, REML with Holm-Šídák’s multiple comparisons test). (C) Time series of the α1-A_R-GluD1_R current amplitude during bath acidification and wash back to pH_e 7.4 (n=10-14). The teal line is the quantification of the real-time change in bath pH measured with FSCV (n=3). (D) Representative traces showing the activation kinetics of the α1-A_R-GluD1_R current at pH_e 7.4 (black) and 6.8 (teal). (E) Quantification of the activation kinetics of the α1-A_R-GluD1_R current at pH_e 7.4, 6.8, and recovery back in 7.4 (7.4 v 6.8, p=0.006, n=14; 7.4 v 7.4, p=0.168, n=9, REML with Holm-Šídák’s multiple comparisons test). (F) Quantification of the decay kinetics of the α1-A_R-GluD1_R current at pH_e 7.4, 6.8 and recovery back in 7.4 (7.4 v 6.8, p=0.046, n=13; 7.4 v 7.4, p=0.635, n=9, REML with Holm-Šídák’s multiple comparisons test). (G) Representative traces showing the current elicited by bath application of 30 μM noradrenaline (I-NA) at pH_e 7.4 (black) and 6.8 (teal). (H) Quantification of the I-NA amplitude at pH_e 7.4 and 6.8 (p<0.001, n=15 and n=17, respectively, Mann-Whitney test)

Decreasing extracellular pH can reduce neurotransmitter release via inhibition of voltage-gated calcium channels²⁷. To determine whether extracellular protons inhibited the α1-A_R-GluD1_R via presynaptic inhibition of noradrenaline release, we bypassed the presynaptic element and bath applied noradrenaline onto the brain slice (30 μM, in the presence of 300 nM idazoxan²⁸). In pH_e 7.4, noradrenaline produced an average inward current of −55±4 pA^21,23 (I-NA, Figure 1G-H). Decreasing pH_e to 6.8 reduced the average amplitude of I-NA by ~50% to −30±4 pA (Figure 1G-H, p<0.001), indicating that proton-mediated inhibition of the α1-A_R-GluD1_R current is not due to presynaptic mechanisms. Overall, these results show that extracellular acidification by −0.6 units from neutral pH inhibits the α1-A_R-GluD1_R current via a postsynaptic mechanism.

Protons inhibit the α1-A_R-GluD1_R current by reducing unitary current

With the exception of GluN1/GluN3A NMDA_R and GluK2/GluK4 Kainate receptors, iGluRs are inhibited by extracellular protons in a voltage-independent manner that largely reduces the open probability without altering channel conductance^10-13,29,30 (recent evidence suggests a slight decrease in calcium permeability for NMDA_R³¹). To determine how protons inhibit α1-A_R-GluD1_R current, we examined the current-voltage relationship generated by voltage ramps (−120 mV to +30 mV, 1 mV/10 ms) at the peak of the α1-A_R-GluD1_R current in pH_e 7.4 and 6.8 (Figure 2A). Consistent with our prior report²¹ the current reversed polarity at −32.0±3.2 mV in pHe 7.4. In pH_e 6.8, the current was reduced across the voltage range, without shifting the reversal potential (−29.8±3.1 mV, Figure 2B, p=0.732). The inhibitory effect of protons, similar to the other iGluRs, was not voltage-dependent (Figure 2C, p=0.814), suggesting action at a site distal to the channel pore.

Figure 2. Protons inhibit α1-A_R-GluD1_R current by reducing unitary current.

(A) Background-subtracted voltage ramps (−120 mV to +30 mV, 1 mV/10 ms) collected at the peak of the α1-A_R-GluD1_R current at pH_e 7.4 (gray, n=9) and 6.8 (teal, n=9). The shaded area represents the S.E.M. (B) Quantification of the reversal potential of the α1-A_R-GluD1_R current at pH 7.4 and 6.8 (p=0.734, n=9, Wilcoxon matched-pairs signed rank test). (C) Quantification of the voltage-dependence of proton-mediated inhibition. For each holding potential, the current observed ay pH 6.8 was divided by the current seen at pH 7.4 and converted to a percentage (p=0.814, n=9, Friedman test). (D) Representative traces of the peak noise of the α1-A_R-GluD1_R current at pH_e 7.4 (black) and 6.8 (teal). Noise was measured in a 700 ms window at the peak of the α1-A_R-GluD1_R current. (E) Stationary fluctuation analysis measured at the peak of the α1-A_R-GluD1_R current at pH_e 7.4 (black, n=16) and 6.8 (teal, n=16). The data was fit with a linear regression, suggesting open probability is small, and the slope of the regression line giving estimates of unitary current (p=0.012, ANCOVA). (F) Quantification of the single-channel conductance of GluD1_R. Single-channel conductance was calculated by estimating unitary current for each cell and then converting to single-channel conductance with the corresponding reversal potential obtained in B (7.4 v 6.8, p=0.003, n=16, Wilcoxon matched-pairs signed rank test).

A feature of the α1-A_R-GluD1_R current is a significant increase in membrane noise variance (σ²)²¹ (Figure 2D). Using between-cell stationary fluctuation analysis (with the assumption that the macroscopic current and σ² arise from identical, independent channels opening and closing), single-channel properties of GluD1_R were estimated^32,33. The α1-A_R-GluD1_R current σ²-amplitude relationship in pH_e 7.4 and 6.8 was well-fit by a linear regression^21,24 (Figure 2E), suggesting that open probability is small (<<0.5) and the slope of the regression line estimates unitary current. In pH_e 7.4, unitary current was ~−1.4 pA, which was reduced by ~70% in pH_e 6.8 to ~−0.4 pA (Figure 2E, p = 0.012). Accordingly, single-channel conductance was reduced by ~40% from 41.1±3.8 pS in pH_e 7.4 to 26.7±4.8 pS in pH_e 6.8 (Figure 2F, p = 0.003). Thus, protons inhibit the α1-A_R-GluD1_R current by reducing unitary current in a voltage-independent manner.

Tonic GluD1_R current is inhibited by extracellular protons

In cell lines and brain slices, GluD1_R and GluD2_R are open and carry tonic cation current^21,24,25,34. While the origin of tonic GluD1_R current remains unknown, it is not affected by arresting GTP/GDP exchange nor elevating free Gβγsubunits²⁴, allowing for measurements of GluD1_R channel activity separate from G protein-mediated GluD1_R currents. Tonic GluD1_R current was revealed by the open-channel pore blocker 1-napthyl acetyl spermine (NASPM, 100 μM, Figure 3A), which produced an apparent outward current at pH_e 7.4 and 6.8 (Figures 3A and 3B). When compared with measurements in pH_e 7.4, the magnitude of tonic GluD1_R current in pH_e 6.8 was reduced by ~50% from −32.6 pA to −17.6 pA in pH_e 6.8 (Figures 3A and 3C). A feature of the tonic GluD1_R current is a significant decrease in membrane noise variance (σ²) upon NASPM application^21,24. However, unitary current has not been determined previously. In control controls (pHe 7.4), tonic unitary current was ~−0.25 pA, significantly less than unitary current of α1-A_R-GluD1_R current (p<0.001, n=10 for both groups, data not illustrated). There was no significant effect of protons on the tonic unitary current (7.4: −0.22±0.09 pA, n=10; 6.8: −0.28±0.19 pA, n=10; data not illustrated).

Figure 3. Tonic GluD1_R current is inhibited by extracellular protons.

(A) Representative traces showing tonic GluD1r current as revealed by application of 100 μM NASPM in pH_e 7.4 (top) and 6.8 (bottom). (B) Quantification of whole-cell current immediately preceding (control) and at the peak of the NASPM current (NSP) at pH_e of 7.4 (black, p=0.002, n=10, Wilcoxon matched-pairs signed rank test) and 6.8 (teal, p=0.020, n=10, Wilcoxon matched-pairs signed rank test). (C) Quantification of the tonic GluD1_R current peak amplitude in pH_e 7.4 (gray) and 6.8 (teal, p=0.015, n=10, Mann-Whitney test). (D) Representative traces showing the α1-A_R-GluD1_R current inhibition by a submaximal (10 μM) application of NASPM in pH_e 7.4 (black, left) and 6.8 (teal, right). (E) Concentration-response curve of the α1-A_R-GluD1_R current to varying concentrations of NASPM at pH_e 7.4 (black) and 6.8 (teal), demonstrating equivalent steady-state potency and efficacy (IC₅₀ 7.4: 5.35 μM, 6.8: 4.27 μM, p=0.544).

In principle, a decrease in tonic GluD1_R current could be explained by a decrease in the ability of NASPM to block GluD1_R channels. Thus, we generated a concentration response curve to determine the efficacy and potency of NASPM to block the α1-A_R-GluD1_R current (Figures 3D and 3E). NASPM (100 μM) nearly eliminated the α1-A_R-GluD1_R current in both pH_e 7.4 and pH_e 6.8, demonstrating equivalent efficacy (Figure 3E). In pH_e 7.4 and pH_e 6.8, the concentration of NASPM required to achieve 50% block (IC₅₀) was not statistically different; 5.35 μM and 4.23 μM, respectively (Figures 3D and 3E). Taken together, the results show that extracellular protons inhibit tonic GluD1_R current, suggestive of direct inhibition of GluD1_R conductance.

Mice lacking GluD1_R have reduced sensitivity to elevated levels of CO₂

To determine whether GluD1_R functions as a pH chemosensor in vivo, wild type or global GluD1_R knockout mice (GluD1_R-KO) inhaled elevated levels of CO₂. In mice, inhalation of 10% CO₂ reduces brain pH from 7.15 to 6.97³⁵. In brief, an acrylic induction chamber was equipped with an inlet to deliver room air or CO₂, an outlet for CO₂ scavenging, and a CO₂ sensor. Mice habituated to the chamber for 15 minutes for two consecutive days before being exposed to either air or 8.5% CO₂ on the third day (Figure 4A). Air or CO₂ was administered to the chamber for 10 minutes starting after the mouse was placed inside and the chamber sealed. When compared to wild type mice, GluD1_R-KO mice froze significantly less over the 15-minute exposure to CO₂ (Figures 4B and 4D). As a metric of the behavioral sensitivity to CO₂, the CO₂ level when each mouse had spent a cumulative time freezing of 1.5 mins (10% of the trial) was recorded. On average, wild type mice froze for 10% of the trial at 4.8% CO₂ (Figure 4C). In contrast, GluD1_R-KO mice required 6.1% CO₂ (Figure 4C, p=0.040). Overall, these results suggest that mice lacking GluD1_R are less sensitive to elevated levels of CO₂, positing GluD1_R as an in vivo pH-sensor.

Figure 4: Mice lacking GluD1_R display a reduced sensitivity to elevated levels of CO₂.

(A) Cartoon schematic detailing the behavioral paradigm. Air was administered on days 1 and 2. Either CO₂ or air was administered on day 3. (B) Time series of cumulative time freezing (right axis) for GluD1-WT (blue) and GluD1-KO (orange) mice during CO₂ exposure (gray line, left axis). CO₂ was administered for 10 of the 15-minutes trial before being turned off, resulting in a consistent steady state value of ~8.5% (n=24). (C) Quantification of the percentage of CO₂ in the testing chamber that elicited 10% of the total time spent freezing for GluD1-WT (blue) and GluD1-KO (orange) mice (p=0.040, n=12, Welch’s t-test). (D) Quantification of the cumulative time spent freezing for GluD1-WT (blue) and GluD1-KO (orange) mice (WT, p<0.001, n=12; KO, p<0.001, n=11; WT v KO in CO₂, p=0.002, n=11-12, 2-way ANOVA with Fischer’s LSD test).

Discussion

Mechanism of proton inhibition of GluD1_R channels

AMPA_R, NMDA_R, and Kainate_R are sensitive to protons with IC₅₀ values of ~6.2, 7.3, and 7.4, respectively^10-13,36. Proton sensitivity of the constitutively open GluD2_R A654T variant (Lurcher) suggested that native GluD_R may also be inhibited by extracellular protons¹³. Here, we found that extracellular protons in a physiologically relevant range inhibit native GluD1_R current significantly, albeit with lower potency when compared to NMDA_R current. The inhibition was not dependent on presynaptic mechanisms, and the magnitude of both synaptically evoked α1-A_R-GluD1_R current and tonic GluD1_R current were inhibited to comparable degrees, suggesting direct inhibitory action on GluD1_R.

While there is subtype-specificity in proton-sensitivity of the iGluRs, the mechanism of proton inhibition is largely the same. Extracellular protons exert their inhibitory effects by reducing the probability of a channel opening without altering channel conductance^10-12,29. Here the proton-mediated inhibition of GluD1_R current was likely due to a decrease in the unitary current, without dramatically changing the relative permeability of cations. One important limitation is that since the σ²-amplitude relationship is well-fit with a linear regression, instead of a parabolic relationship^32,37, a decrease in the probability of opening could be reflected as a decrease in unitary current. Since the probability of opening shapes the time course of activation of GluN1/GluN2B NMDA_R current³⁸ and pH_e 6.8 increased the τ-activation of the α1-A_R-GluD1_R current, protons may also be reducing the probability of opening of GluD1_R. Single-channel recordings of GluD1_R current would be ideal to solve this problem; however, technical limitations current prevent single-channel GluD1_R currents to be recorded in native cells^14,15,18.

For the GluN1/GluN2A NMDA_R, protons inhibit current by destabilizing the extracellular domain into a ‘splayed’ conformation that hinders gating by glutamate^39,40. The structure of the rat GluD1_R shows a unique, non-swapped domain architecture between the ligand binding domains and amino terminal domains which is thought to lead to ‘floppy’ extracellular arms of GluD1_R that resemble that of the protonated GluN1/GluN2A NMDA_R⁴¹. These key differences in the rigidity of the structure may underlie the lower potency of protons to inhibit GluD1_R current, which posits that protons inhibit GluD1_R current by hindering gating by a yet-to-be identified cognate ligand. Alternatively, protons may be stabilizing GluD1_R in a sub-conducting state. It is well-established that NMDA_R and AMPA_R channels have multiple conductance states^42-47, and recent work has shown sub-conductance states of human GluD2_R channels when expressed in artificial lipid bilayers¹⁸. Here, we found that tonic GluD1_R current arises from channels in a lower conductance state relative to a higher conductance state following α1-A_R activation, and this tonic unitary current is not sensitive to extracellular protons. One possibility is that extracellular protons ‘trap’ GluD1_R in a sub-conducting state and prevent α1-A_R activation to drive opening to higher conductance levels. Future work resolving the structure of GluD1_R under acidic conditions will provide insight into the precise mechanism.

Health implications

Acidic shifts in brain pH can arise from cardiovascular, pulmonary, neurological, and neuropsychiatric disorders and can cause cognitive impairment. In this study, we demonstrate that long-lasting excitatory current carried by GluD1_R channels is reduced by extracellular protons. In vivo, noradrenergic activation of α1-A_R drives serotonin neurons to fire action potentials⁴⁸ and release serotonin^49,50 via GluD1_R conductance²¹. Thus, a decrease in GluD1_R current in serotonin neurons, in principle, would promote a decrease in serotonin neuronal excitability with a decrease in brain pH. Reduced excitability of serotonin neurons and subsequent serotonin release is expected to alter mood and affect, sleep/wake control, and thermoregulation, which may contribute to the symptoms of depression, anxiety, bipolar disorder, and schizophrenia. In humans, variations in the gene encoding GluD1_R (GRID1) have been linked to many of the same disorders^22,51 although how these genetic variations impact the ion channel function of GluD1_R is not yet known. In the present study, global GluD1_R knock-out mice displayed reduced behavioral sensitivity to elevated levels of CO₂. Impaired response to CO₂ may contribute to mortality due to sleep apnea, sudden unexpected death in epilepsy, and sudden infant death syndrome. Conversely, enhanced sensitivity to CO₂, for example in individuals with bipolar disorder⁵², elicits aversive panic-like states. This study lays the foundation to further investigate the role of GluD1_R in pH chemosensing and the consequences of proton-mediated channel inhibition in neural network excitability and health following acidic insult.

Overall, these results continue to expand on the growing literature supporting ion channel function of with similar biophysical properties as other iGluRs. It also begins to contextualize the physiological role GluD1_R play in mediating behavioral responses to changes in brain pH.

Methods

Animals

All experiments were conducted in accordance with the University of Iowa with the approval of the University of Iowa Institutional Animal Care and Use Committee. All electrophysiology experiments used male and female C57BL/6J obtained from The Jackson Laboratory (>2 months old, #000664). All behavioral assays used male and female GluD1^+/+ and GluD1^−/− mice (mixed 129/SvEv and C57BL/J strain) originally generated by Dr. Jian Zuo (St. Jude Children’s Research Hospital). Mice were group-housed on a 12/12 light/dark cycle.

Acute brain slice preparation and electrophysiological recordings

Brain slices were made, and electrophysiological recordings conducted, as previously described. In brief, mice were anesthetized with isoflurane and euthanized via rapid decapitation. Brains were quickly removed and placed in warmed modified Krebs’ buffer (bubbled with 5/95% CO₂/O₂) containing (in mM): 126 NaCl, 2.5 KCl, 1.2 MgCl₂, 1.2 CaCl₂, 1.2 NaH₂PO₄, 21.5 NaHCO₃, 11 D-glucose and 5 μM MK-801 to reduce excitotoxic cell death during slice collection. Coronal slices (240 μm) were collected with a vibrating microtome (Leica) and incubated in the modified Krebs’ buffer at 28 °C for at least 30 minutes before recording.

Brain slice electrophysiological recordings were made from serotonin neurons in the dorsal raphe nucleus, as previously described^21,23,24,28. Slices were continuously perfused with modified Krebs’ buffer maintained at 35 °C with an in-line heating element and temperature controller (Warner Instruments). Recordings were made with MultiClamp 700B amplifiers, Digidata 1440A and 1550B converters, and ClampEx software (Molecular Devices, ver. 10.7 and 11.3) with borosilicate glass electrodes (World Precision Instruments and Warner Instruments) wrapped in parafilm to minimize pipette capacitance. Pipette resistances were 3.0-4.5 MΩ when filled with an internal solution containing (in mM): 104.56 K-methysulfate, 3.73 KCl, 5.3 NaCl, 4.06 MgCl₂, 4.06 CaCl₂, 7.07 HEPES (K), 3.25 BAPTA (4K), 0.26 GTP-Na salt, 4.87 ATP-Na salt, 4.59 creatine phosphate-Na salt, adjusted to ~275 mOsm and maintained at pH 7.2. For voltage-ramp experiments, a cesium-based internal solution was used containing (in mM): 120 CeMeS, 5.3 NaCl, 7.07 HEPES, 3.25 BAPTA (4K), 0.26 Na-GTP salt, 4.87 Na-ATP salt, 4.59 Na-creatine phosphate, 4.06 MgCl₂, 4.06 CaCl₂, adjusted to ~275 mOsm and maintained at pH 7.2. Series resistance was monitored throughout the recordings. All reported voltages are corrected for an −8 mV liquid junction potential between the internal and external solutions. α1-A_R-excitatory postsynaptic currents (α1-A_R-GluD1_R current) (V_hold = −65 mV) were evoked every 90-s by trains of current (5 pulses, 0.5 ms, 60 Hz) to the brain slice via a borosilicate glass monopolar stimulating electrode filled with modified Krebs’ buffer. The stimulating electrode was placed within 200 μm of the recorded neuron. All drugs were applied by bath application, and pH was reduced by bubbling the modified Krebs’ buffer in 20/80% CO₂/O₂ gas mixture for at least 30 minutes at 35 °C prior to application. All recordings were made in the presence of GluN (MK-801, 5 μM), GluA/GluK (NBQX, 3 μM), GABA_A (picrotoxin, 100 μM), and 5-HT1A (WAY-100635, 300 nM) receptor antagonists to isolate the α1-A_R-EPSC.

Stationary Fluctuation Analysis

Stationary fluctuation analysis was conducted as previously described^21,24. In brief, a single recording in pH_e 7.4 and 6.8 were selected for each cell, and we subtracted the baseline variance from the variance at the peak of the α1-A_R-GluD1_R current for each recording and measured the amplitude of the α1-A_R-GluD1_R current. The σ²-I relationship was fit with a simple linear regression, and the slopes were compared with an ANCOVA test. To calculate single-channel conductance (γ), we used the following relationship:

$$\gamma =\frac{i}{\left(V_{\text{hoid }}-E_{\text{rev }}\right)}$$

Where i is unitary current (obtained by dividing σ²/I for each cell, see Bean et al., 1990), V_hold is the holding voltage (−65 mV), and E_rev is the reversal potential of the α1-A_R-GluD1_R current obtained from the voltage ramp experiments. Note, while E_rev was not statistically different, single-channel conductance was calculated with the corresponding E_rev estimate obtained in pH_e 7.4 and 6.8.

Behavioral assays

For all behavioral experiments, mice were placed in a rabbit CO₂ induction chamber (18” L x 12” W x 12” H; EZ-179; EZ Systems) fit with a CO₂ sensor (USB-DXC220t-CAL; Dracal Technologies). Mice were habituated to the chamber for 15 minutes on subsequent days before testing on day 3. During habituation and air exposure, room air was circulated through the chamber using an aquarium pump. To expose mice to elevated levels of CO₂, the aquarium pump line was clipped off and a line connecting a gas cylinder (20% CO₂, 21% O₂, N₂ balance) was opened. The CO₂ gas mix was pumped into the induction chamber for 10 minutes at a rate of 5.0 L/min resulting in a maximum level of CO₂ in the chamber of ~8.5%. Higher flow rates resulted in an audible hissing noise that resulted in mice freezing before any appreciable increase in CO₂ was observed. All experiments were recorded using a camera mounted above the CO₂ chamber. Total time spent freezing was scored manually from these videos using the free Behavioral Observation Research Interactive Software (BORIS), with the scoring investigator blind to the sex, genotype, and condition.

Experimental design and statistical analysis

All data were analyzed using ClampFit 10.7 and 11.1. Data are presented as representative current traces, or as means ± S.E.M, with each point representative of a unique cell. ‘N’ denotes the number of cells for all electrophysiological experiments. All data sets are analyzed with nonparametric statistical tests (Wilcoxon matched-pairs signed rank test for two-group within analyses, Mann-Whitney tests for two-group between analyses, Kruskal-Wallis test for three or more between-group analyses, Friedman test - or mixed-effects analysis if conditions had unequal ‘N’s’ - for three or more within-group analyses, or 2-way ANOVA). All datasets were checked for sex differences, and statistically significant sex differences are reported. An alpha value of 0.05 was used to determine statistical significance, and exact p-values are reported unless p<0.001 or >0.999. All statistical analyses were performed in GraphPad Prism (GraphPad Software, Inc.).