Introduction

The dorsal division of the mammalian cochlear nucleus (DCN) is a cerebellum-like structure in which principal cells integrate multimodal activity with primary auditory afferents, and these converging signals are thought to contribute to sound localization and sensitivity to sounds of interest (May, 2000; Oertel and Young, 2004). Cartwheel cells (CWC) are cerebellar Purkinje cell homologs that potently control this convergence by gating the multimodal signals to postsynaptic fusiform principal cells. CWCs also receive inputs from various neuromodulatory systems, including noradrenaline (NA) (Trussell, 2019). We showed previously that agonists of α2 noradrenergic receptors simultaneously halts spontaneous spike activity and enhances the strength of inhibitory signals to fusiform cells (Kuo and Trussell, 2011). The mechanism for this paradoxical action depends on the effect of spontaneous activity on synaptic strength: by reducing ongoing presynaptic firing, recovery from synaptic depression ensues and allows CWCs to mediate stronger postsynaptic signals in the fusiform cells. A related mechanism was later identified in the action of oxytocin in hippocampal interneurons (Owen et al., 2013; Tirko et al., 2018). However, the mechanism by which α2 noradrenergic receptors controlled such activity remained obscure. The most obvious candidate ion channel that could mediate such inhibition is the GIRK channel (G -protein-gated inwardly rectifying K⁺ channel) (Luscher et al., 1997; Arima et al., 1998; Li and van den Pol, 2005; Philippart and Khaliq, 2018), however attempts in our laboratory to implicate this channel failed.

The Na⁺ leak channel NALCN functions as a complex of 4 proteins: NALCN, FAM151a, UNC79 and UNC80 (Lu et al., 2010; Ren, 2011; Cochet-Bissuel et al., 2014; Kschonsak et al., 2020; Xie et al., 2020). This channel complex (here termed simply NALCN) contributes to the depolarizing drive for spontaneous firing in a wide variety of neurons, and functions in sensory, motor and circadian pathways (Nash et al., 2002; Lu et al., 2007; Lu and Feng, 2011; Xie et al., 2013; Lutas et al., 2016; Shi et al., 2016). Accordingly, mutations in the NALCN subunit or its associated proteins are linked to a variety of human diseases (Al-Sayed et al., 2013; Cochet-Bissuel et al., 2014; Chong et al., 2015; Bramswig et al., 2018). A hallmark of this channel is its inhibition by extracellular Ca²⁺; reduction of extracellular Ca²⁺ enhances the current, permitting a ready assessment of NALCN’s presence in neurons (Lu et al., 2010; Ren, 2011; Philippart and Khaliq, 2018; Chua et al., 2020). Recently it was shown that neurons of the substantia nigra express NALCN, and that both dopamine and GABA_B receptors strongly downregulated NALCN activity in this region (Lutas et al., 2016; Philippart and Khaliq, 2018). As global knockouts of NALCN die at birth (Lu et al., 2007), we generated a knockout mouse line specific to glycinergic neurons in order to test the role of this channel in auditory interneurons. We found that CWCs expressed functional NALCN, and its loss in NALCN knockouts led to cessation of spontaneous firing. Outward currents and inhibition of firing mediated by NA was eliminated in the knockouts; since NALCN generates inward excitatory currents, NA must act by inhibition of NALCN. Similar results were obtained for the GABA_B agonist baclofen. Thus, NALCN serves multiple roles in auditory function, setting the pace of neuronal firing while also mediating electrical inhibition by multiple neuromodulators and enhancement of synaptic strength.

Results

NALCN and modulation of spike generation

In order to assess the role of NALCN in modulation of spontaneous firing of CWCs, we generated an NALCN knockout mouse by crossing a floxed NALCN mouse line B6(Cg)-Nalcn^{tm1c(KOMP)Wtsi}/DrenJ with a GlyT2-cre line, thus deleting NALCN from glycinergic neurons (Methods). NALCN^−/− mice were slightly smaller than wildtype mice (weight ratio, 4:3) but had normal hearing, as assessed by auditory brainstem responses (Figure 1-figure supplemental 1), and the morphology of their CWCs was characteristic of those of wildtype mice (Figure 1-figure supplemental 2; (Bender and Trussell, 2009)). ‘Wildtype’ mice used here were either C57/BL6 or GlyT2-EGFP (Zeilhofer et al., 2005; Kuo et al., 2012; Ngodup et al., 2020) in which GFP is expressed in glycinergic neurons. CWC were recorded using cell-attached mode to monitor baseline spontaneous firing and the inhibition of firing by NA. Except as noted below, all recordings were made in physiological Ca²⁺ concentration (1.2 mM). In confirmation of previous studies in wildtype mice (Kim and Trussell, 2007; Kuo and Trussell, 2011) 71% of CWC exhibited spontaneous firing (mean firing rate, 17.74 ± 2.31 Hz) and bath application of 10 μM NA completely and reversibly eliminated such firing in every case (N = 8, Fig. 1A & B). However, in experiments on NALCN KO mice, CWC showed no spontaneous firing (Fig 1C & D; N = 6; difference from wildtype type, p < 0.0005), and thus NA had no additional effects on spontaneous firing (mean firing rate, KO = 0 ± 0 Hz, NA = 0; N = 6). Thus, NALCN likely provides a tonic inward current necessary to drive spontaneous firing in CWC. These effects were not accompanied by changes in membrane input resistance (WT: 135.34±21.53 MΩ (n=18) vs KO: 132.19±16.11 MΩ (n=15), p=0.46, t-test).

Figure 1

Although NALCN supports spontaneous firing, it was not clear whether the effect of NA is mediated through NALCN or some other channel. To explore this problem, we first examined evoked firing in CWC and the effects of NALCN and NA. CWC were recorded in current-clamp mode, and a family of negative and positive, 600-ms current pulses were delivered in 25-pA steps (Fig 2A1). In response to positive pulses, CWCs fire mixtures of simple (single) spikes and complex spikes, the latter containing Na⁺ spike clusters driven by a Ca²⁺-dependent current (Kim and Trussell, 2007). The timing of spikes during each current pulse was used to generate raster plots of firing (Fig 2A2), and the number of spikes at each current step was used to plot a frequency vs current intensity curve (Fig 2E). In the same set of cells, we then applied NA and repeated the current steps (Fig 2B & E). NA hyperpolarized the resting potential by −1.90 ± 0.24 mV and increased the current level required to initiate firing (rheobase, Fig 2E), indicating that NA reduced excitability. However, we also observed a suppression of the peak firing rate in NA; similar rundown of peak firing rate could be observed without change in rheobase as a result of whole-cell recording over this time period (see also Kim and Trussell, 2007), and thus frequency-intensity plots were normalized to peak intensity (Fig 2F). This manipulation revealed a clear shift to the right in the onset of firing (i.e., toward higher intensity) as a result of NA. Rheobase was defined as the current intensity at which spiking was 20% of maximum. Using this criterion, NA generated a significant increase in rheobase in response to current pulses (Ctrl, rheobase = 91.40 ± 11.35 pA, NA, 168.59 ± 14.89 pA, p = 0.0011 (t-test)). Control experiments established that wash-in of NA with the α2 blocker idazoxan did not lead to a change in rheobase (Ctrl, rheobase = 95.11 ± 16.86 pA, NA + idazoxan, 105.88 ± 21.89 pA, N = 9, p = 0.38 (t-test)), showing that the shift in rheobase was a genuine effect of α2 receptors (Fig 2I, J). This protocol was then applied to NALCN KO mice (Fig 2C, D, G & H). Although CWCs showed no spontaneous firing, they were able to respond to current steps with mixtures of simple and complex spikes, and peak firing rates did not differ from wildtypes (rheobase, Ctrl, 91.40 ± 11.35 pA; KO, 120.81 ± 17.53 mV, N = 11, p = 0.08 (t-test)). However, unlike wildtype mice, the NALCN KO mice showed no significant shift in rheobase with NA (KO, rheobase = 120.81 ± 17.53 pA, NA, 158.0 ± 12.16 pA, n = 11, p = 0.064 (t-test) (Fig 2H).

Figure 2

In order to test whether the effects of NA on rheobase and evoked spiking is primarily due to the small hyperpolarization induced by NA, we mimicked this effect of NA by injecting current to CWC to hyperpolarize the membrane by the same amount as induced by NA (approximately 2 mV), and then repeated the current injection protocol (Figure 2-figure supplement 1). Under these conditions, we were able to replicate the effects of NA on CWC rheobase and evoked firing (rheobase Ctrl = 70.88 ± 7.68 pA, hyperpolarized = 125 ± 12.90 pA, N = 6, p = 0.0009). We then compared the ratio of the changes in rheobase in NA with respect to control under different conditions, also comparing to the relative effects of current injection. These comparisons confirmed that NA shifts rheobase dependent upon NALCN, and that this effect is likely caused by membrane potential hyperpolarization (Fig. 2K: Ctrl = 1.90 ± 0.11, KO = 1.25 ± 0.05, Idazoxan = 1.03 ± 0.11, hyperpolarized = 1.94 ± 0.16, p < 0.0001, non-parametric Kruskal-Wallis; post hoc Dunn’s test, Ctrl vs KO p = 0.0086, Ctrl vs idazoxan p = 0.0007, KO vs hyperpolarized p = 0.02, idazoxan vs hyperpolarized p = 0.0028).

Outward response generated by suppression of NALCN

As cessation of spontaneous activity and elevation of rheobase by NA is likely associated with an outward, inhibitory current, we performed voltage-clamp experiments to determine the magnitude of that current in wildtype and KO mice, and its pharmacological sensitivity. Using a patch-pipette fill containing high K⁺ (Methods), cells were held at −65 mV and a 5-ms puff of NA was applied near the soma using a pressure-ejection pipette. Under these conditions, a slowly rising outward current was observed that decayed over a time course of 10-15 seconds (Fig 3A). As G-protein coupled inward rectifier channels (GIRK) are known to mediate α2 receptor effects in some brain regions (Williams et al., 1985; Aghajanian and Wang, 1986; Arima et al., 1998; Li and van den Pol, 2005; Nimitvilai et al., 2017), we tested the role of GIRK channels using the GIRK blocker Ba²⁺. However, BaCl₂ (100 μM) had no significant effect on the amplitude of the NA-evoked current (Fig 3A, C) (Ctrl = 15.44 ± 1.66 pA; BaCl₂ = 12.86 ± 1.44, N = 6, p = 0.133). When this experiment was repeated in NALCN KO mice, the NA-evoked current was only 12.5% of that observed in wildtype mice (Fig 3B & C). Although the amplitude of the NA-evoked current in the knockout was only a few pA, bath application of BaCl₂ nevertheless produced a statistically significant block of this small residual current, essentially eliminating all response to NA (KO = 1.93 ± 0.83 pA, BaCl₂ = 0.56 ± 0.4 pA, n = 8, p = 0.05) (Fig. 3C). Thus, the outward response to NA was only minimally due to activation of K⁺ current but rather was mediated largely by inhibition of a tonic inward current generated by NALCN. Philippart and Khaliq (2018) observed that GABA_B receptors suppressed NALCN current in substantia nigra, and so we asked whether these receptors might have a similar action in CWCs. In K⁺ filled cells, puffs of the GABA_B agonist baclofen generated outward currents similar to those observed with NA (Fig 4A & C). However, unlike NA, subsequent wash-in of BaCl₂ blocked 26.99 ± 4.60 % of the baclofen response (Ctrl = 42.95 ± 2.58 pA, BaCl₂ = 30.90 ± 1.84 pA, N = 8, p = 0.0014) (Fig 4A, C & D), suggesting the partial involvement of GIRK channels. In NALCN KO mice, the baclofen currents were significantly smaller than in wildtype (Fig 4B&C). Moreover, BaCl₂ in the KO had a much greater effect on the baclofen current, as compared to wildtype, blocking by 82.68 ± 6.71% (Fig 4B, C & D) (KO = 12.91 ± 3.51 pA, BaCl₂ = 2.57 ± 1.06 pA, N = 7, p = 0.0089), indicating that after loss of NALCN, the remaining baclofen response was largely dependent on GIRK channels.

Figure 3

Figure 4

Low Ca²⁺induced inward current mediated by NALCN

NALCN current is enhanced by reduction of extracellular Ca²⁺ (Lu et al., 2007; Lu et al., 2010; Ren, 2011; Philippart and Khaliq, 2018; Chua et al., 2020), and in the presence of blockers of voltage-dependent Na⁺, Ca²⁺, and K⁺ channels this effect is considered diagnostic for the presence of NALCN (Philippart and Khaliq, 2018). Recording in voltage clamp with the cocktail of intracellular and extracellular channel and receptor blockers described by Philippart and Khaliq, 2018 (see methods, and note that GIRK channels are also blocked here), an inward current of −107.61 ± 10.80 pA (N = 18) developed over several minutes upon shifting bath Ca²⁺ from 2 mM to 0.1 mM (Fig 5A & B). When NA was subsequently washed in, still in the presence of 0.1 mM Ca²⁺, the inward current was immediately reduced, falling by over half (−51.14 ± 9.44 pA, N = 18) (Fig 5A, B & F). When similar experiments were performed in NALCN KO mice there was no change in holding current upon reduction in bath Ca²⁺, and only a minimal current response to NA (−4.78 ± 4.66 pA, N = 6) (Fig 5C, D & F). In order to confirm that the reduction in inward current by NA was mediated by α2 receptors, we repeated the experiments in wildtype mice but in the presence of 1 μM idazoxan, and observed no effect on the inward current (Ctrl, 0.1 mM Ca²⁺ = −112.52 ± 12.80 pA, NA = −108.35 ± 5.80 pA, N = 5, Fig 5E). Additional experiments were made using baclofen to activate GABA_B receptors. Here, application of 10 μM baclofen almost completely eliminated the 0.1 mM -Ca²⁺ induced NALCN current (0.1 mM Ca²⁺ = −117.75 ± 19.6 pA, baclofen = −18.80 ± 14.5 pA, N =15, p = 0.00018, Fig 6A, C & E), while in knockout mice, neither 0.1 mM Ca²⁺ nor baclofen altered the holding current (0.1 mM Ca²⁺ = −14.48 ± 9.39 pA, baclofen = −15.90 ± 8.94 pA, N = 6, p = 0.45, Fig. 6B, D & E). Altogether, these experiments support that NALCN is expressed in CWC, and generates a Ca²⁺ sensitive inward ionic current which is suppressed by α2 and GABA_B receptors.

Figure 5

Figure 6

Since both NA and baclofen inhibit NALCN in CWC, we asked whether these two receptors act on independent or overlapping populations of NALCN channels. In these experiments, one set of control recordings assessed the magnitude of block by NA of low-Ca^2+-evoked inward current. Interleaved with these recordings were experiments in which baclofen was applied to reduce the NALCN current, and then NA subsequently applied in the continued presence of baclofen (Fig. 7A-C). If the two agonists act on independent populations of NALCN channels, NA should suppress inward current similar to that seen in control experiments. However, as shown in Fig 7, NA had virtually no blocking action of inward current following application of baclofen (0.1 mM Ca²⁺ = −164.28 ± 20.50 pA, baclofen = −53.33 ± 14.73 pA. NA = −46.94 ± 16.50 pA, N = 9, F = 41.28, p < 0.0001, repeated measures one-way ANOVA; post hoc Tukey’s test, 2 mM Ca²⁺ vs 0.1 mM Ca²⁺ p = 0.002, 0.1 mM Ca²⁺ vs baclofen p = 0.005, 0.1 mM Ca²⁺ vs NA p = 0.002, baclofen vs NA p = 0.84). We also compared with percentage block of 0.1 mM Ca²⁺ current in baclofen, or NA alone, compared to the block of current by NA in a background of baclofen. Baclofen completely occluded the response to subsequent application of NA (percent block: baclofen 70.61 ± 7.18, NA 56.18 ± 9.5, additional block 4.35 ± 3.25, n = 9, F _(2,24) = 23.83, p < 0.0001, one-way ANOVA; post hoc Tukey’s test, baclofen vs NA p = 0.34, baclofen vs addition p < 0.0001, NA vs addition p < 0.0001). These data suggest that α2 receptors target the same set of NALCN channels as GABA_B receptors, implying close apposition of the two receptors and the NALCN channel complex.

Figure 7

Noradrenergic effect not mediated by control of cAMP levels

The α2 receptor is a Gi/o GPCR and thus it is possible that inhibition of adenylyl cyclase and thus reduction of cAMP levels might mediate the action of NA on NALCN. To test this idea, CWCs were recorded in the same solutions used in the previous section, in order to block K⁺ channels and other non-NALCN contributions to current. Then NA (100 μM) was puffed onto the recorded neuron, which suppresses NALCN and generates a net outward current. After recording control responses, 1 mM 8-Br-cAMP was washed into the bath, and NA responses again recorded. The average response amplitudes were 23.9±3.5 pA in control and 27.8±5.7 pA in 8-Br-cAMP (n=7 cells, p=0.1147, paired t-test). On the assumption that 8-Br-cAMP would effectively saturate cAMP actions in the neurons, the absence of an effect on the NA response suggests that inhibition cAMP production does not mediate NA action on NALCN, pointing to the possibility of direct effect of G-proteins on the channel.

Noradrenergic enhancement of glycine release

CWC mediate synaptic inhibition through the release of glycine onto fusiform principal cells as well as onto other CWC in the DCN (Mancilla and Manis, 2009; Roberts and Trussell, 2010; Kuo and Trussell, 2011; Apostolides and Trussell, 2014). NA has been shown to enhance feedforward inhibition by suppressing background spiking activity (Kuo and Trussell, 2011; Lu and Trussell, 2016). We tested more directly the effect of NA on synaptic inhibition and the role of NALCN by electrically evoking IPSCs while recording from CWC. Bipolar stimulation electrodes were positioned in the DCN molecular layer in order to stimulate CWC axons. The resulting glycinergic IPSCs were significantly enhanced upon bath application of 10 µM NA (Ctrl = −1.37 ± 0.31 nA, NA = −1.82 ± 0.43 nA, % change = 32.99 % (increase), N = 8, Fig 8A & B). This average increase ranged widely from 2% to 81%, suggesting some cells received glycinergic input from CWC that were not spontaneously active and therefore not affected by NA. The experiment was repeated on NALCN KO mice, there was no significant increase in IPSC amplitude (KO = −2.17± 0.54 nA, NA = −2.08 ± 0.55 nA, % change = 3.77% (decrease), N = 10, p = 0.0011; Fig 8 B & C). Thus, NA-mediated increase in the strength of inhibition (Kuo and Trussell, 2011) is likely mediated by NALCN.

Figure 8

Discussion

Mechanism of adrenergic modulation

The primary effect of noradrenergic modulation of the CWC is to enhance the strength of inhibitory synaptic transmission (Kuo and Trussell, 2011). Modulation by NA acts simultaneously on all the CWC neurons synapses by impacting spike generation and thereby the degree of frequency-dependent synaptic depression. We show here that NALCN is the ion channel that mediates this effect on firing. α2 receptors are known to mediate inhibition of firing through enhancement of K⁺ channel activity, and GIRK channels play a prominent role in these effects (Williams et al., 1985; Aghajanian and Wang, 1986; Arima et al., 1998; Li and van den Pol, 2005). However, in CWC, almost none of the effect of NA was blocked by the K⁺ channel blocker Ba²⁺. Indeed, since in NALCN KOs the NA-induced effects on spiking, outward currents or block of inward currents were all eliminated, it is almost certain that NALCN and not K⁺ channels play the key role in noradrenergic modulation in CWC. A caveat is that we only examined ionic currents and voltage changes measurable with a somatic electrode. NA effects in distal dendrites may have been undetectable at the soma.

Several intriguing aspects of NA action are revealed in this work when considered together with our previous study (Kuo and Trussell, 2011). First, the action NA on feedforward inhibition is apparently quite specific to NALCN. No effect of NA was observed on parallel fiber EPSCs in CWC or CWC IPSCs in fusiform cells recorded during paired voltage-clamp recordings (Kuo and Trussell, 2011). This suggests that the mechanism for control of inhibitory strength is optimized to act globally on the CWC’s synapses by controlling spontaneous firing. Presumably, α2 receptors are localized in such a way to affect membrane potential and rheobase at the soma and/or at the axon initial segment, although localization studies will be required to address this uncertainty. Despite the DCN’s dense innervation from noradrenergic fibers (Kromer and Moore, 1980; Klepper and Herbert, 1991), we do not know the morphology of noradrenergic synapses in the DCN. However, if NA release is through “volume transmission”, it may be that noradrenergic modulation acts on groups of CWCs and all their synapses. Second, although CWC have GIRK channels, and α2 activate GIRK channels in some other neuronal cell types (Arima et al., 1998; Paladini and Williams, 2004; Li and van den Pol, 2005), the CWC has selected for a mechanism in which an inhibitory action is associated with a decrease in a tonic membrane conductance. The effect on membrane potential of this decrease in conductance is small, ∼2 mV. Yet, we show that that shift alone is sufficient to profoundly inhibit spontaneous firing and thereby increase the strength of inhibitory synapses. Interestingly, an inverse mechanism of modulation was described in hippocampus, in which oxytocin increases spontaneous firing of interneurons and thus decreases inhibitory synaptic strength (Owen et al., 2013). In this case oxytocin acts by decrease of a resting inhibitory conductance mediated by Kv7 channels (Tirko et al., 2018). Thus, modulation by inhibitory or excitatory conductance decrease provides a means for global control of inhibitory synaptic strength in the CNS.

Convergence of GABAergic and adrenergic systems

Unlike α2 receptors, the actions of GABA_B receptors were more complex, as GABA_B receptors both inhibited NALCN and activated GIRK channels, as previously observed in substantia nigra neurons (Philippart and Khaliq, 2018). Comparisons of the magnitude of NA action in control recordings and in a background of baclofen indicated that the two modulators acted on a common population of NALCN channels. While activation of α2 receptors may lead to inhibition of adenylyl cyclase (Dohlman, et al. 1987), we did not observe an effect of 8-Br-cAMP on NA responses, arguing against a role for cAMP. If gating of the channels proceeds directly by membrane-delimited actions of G-proteins, it is likely that a pool of GABA_B and α2 receptors are in relatively close apposition to one another and to NALCN. However, unlike GABA_B receptors, α2 receptors did not appreciably activate GIRK channels. Altogether, these results suggest two distinct pools of GABA_B receptors in the somatodendritic membrane: one associated with GIRK and one with NALCN. In addition, a third pool is localized in CWC nerve terminals. There, baclofen suppresses release of action potential-independent exocytosis (Apostolides and Trussell, 2013), suggesting modulation by GABA_B receptors of Ca²⁺ channels or exocytic proteins. Regarding the somatodendritic populations, immunolocalization with electron microscopy indicated that GABAB1 subunits are concentrated at the base of dendritic spines of CWC, rather than associated with GABAergic synapses (Lujan et al., 2004), with the suggestion that they respond to “spillover” of GABA. It remains unclear which ion channel this particular dendritic GABA_B population modulates.

Auditory processing

The DCN is the site of diverse forms of neuromodulation mediated by G-protein coupled receptors (Trussell and Oertel, 2018; Trussell, 2019), in which transmitters act at presynaptic membrane to regulate exocytosis (Tang and Trussell, 2015, 2017), at the axon initial segment to control spike pattern (Bender et al., 2010), at somatodendritic membrane to boost firing (Tang and Trussell, 2017), or at dendritic spines to control postsynaptic efficacy (He et al., 2014). By restricting α2 modulation to the spike-generation mechanism and not presynaptic release zones, the transmitter is able to enhance IPSC amplitude while simultaneously reducing background inhibitory “noise”, thus enhancing the salience of a given interneuron’s impact. CWC are activated by parallel fibers conveying multisensory input to DCN fusiform cells. If, consistent with other studies (Berridge and Waterhouse, 2003), NA is released during brain states associated with wakefulness, this enhanced signal-to-noise mediated by NA would serve to more precisely sculpt the output of DCN based on such multisensory signals. Moreover, since individual CWC-fusiform synaptic pairs apparently receive excitatory input from different populations of parallel fibers (Roberts and Trussell, 2010), such enhanced inhibition might serve a more precise computational role that a standard feedforward inhibitory mechanism. NALCN therefore appears to play a potent role in selective sensory filtering during heightened states of vigilance.

Materials and Methods

Animals

All procedures were approved by the Oregon Health and Science University’s Institutional Animal Care and Use Committee. C57BL/6J, GlyT2-EGFP mice (Zeilhofer et al., 2005; Ngodup et al., 2020), and NALCN conditional knock mice of either sex, postnatal days (P) 17–40 were used in this study. GlyT2-GFP mice were backcrossed into the C57BL/6J and maintained as heterozygous. As global knockout of NALCN is lethal, we generated a glycinergic neuron-specific knock out by crossing NALCN^flox/flox mice (B6(Cg)-Nalcn^{tm1c(KOMP)Wtsi}/DrenJ) with GlyT2-Cre mice (Tg(Slc6a5-cre)1Uze), resulting in NALCN^flox;GlyT2-Cre offspring. The F1 litters were back crossed with NALCN^flox/flox mice to generate NALCN^flox/flox;GlyT2-Cre that lack NALCN expression in glycinergic cells. NALCN conditional knock out mice were smaller than age matched WT litters but had normal hearing as established by auditory brainstem response.

Brain-slice preparation

Animals were anesthetized with isoflurane and decapitated. The brain was quickly removed and placed into ice-cold sucrose cutting solution. Sucrose solution contained (in mM) 76 NaCl, 26 NaHCO₃, 75 sucrose, 1.25 NaH₂PO₄, 2.5 KCl, 25 glucose, 7 MgCl₂, and 0.5 CaCl₂, bubbled with 95% O₂: 5% CO₂ (pH 7.8, 305 mOsm). Coronal slices containing DCN were cut at 210 µm in ice-cold sucrose solution on a vibratome ((VT1200S; Leica Microsystems, Wetzlar, Germany) or (7000smz-2; Campden Instruments, Loughborough, UK)). Slices were transferred into standard artificial cerebrospinal fluid (ACSF) containing (in mM) 125 NaCl, 20 NaHCO₃, 1.2 KH₂PO₄, 3 HEPES, 2.1 KCl, 20 glucose, 1 MgCl₂, 1.2 CaCl₂, 2 Na-pyruvate, and 0.4 Na L-ascorbate, bubbled with 95% O₂:5% CO₂ (pH 7.4, 300–310 mOsm). Slices were recovered at 34°C for 40 min and were maintained at room temperature until recording.

Electrophysiology

Slices were transferred to a recording chamber and perfused with standard ACSF at 3 ml/min and maintained at 31-34°C with an in-line heater (TC-324B; Warner Instruments, Hamden, CT). Cells were viewed using an upright microscope (BX51WI; Olympus, Tokyo, Japan) with a 60X objective, equipped with an infrared custom-made Dodt contrast, CCD camera (Retiga 2000R; QImaging, Surrey, Canada), and fluorescence optics. All recordings were collected from CWCs of the DCN. CWCs were targeted by their location in the molecular and fusiform cell layers of DCN, and by their round soma (Kim and Trussell, 2007). Identification was then confirmed by their distinctive firing pattern (simple or complex spikes) (Golding and Oertel, 1997). In slices from GlyT2-EGFP, glycinergic cells in the DCN were identified by their GFP expression. In some experiments, 0.1% biocytin (B1592; Thermo Fisher Scientific, Waltham, MA) was added to the pipette solution for post-hoc identification of CWCs. Recording pipettes were pulled from 1.5 mm OD, 0.84 mm ID borosilicate glass (1B150-F; World Precision Instruments, Sarasota, FL) to a resistance of 2–4 MΩ using a horizontal puller (P-97 or P-1000; Sutter Instruments, Novato, CA). In most experiments, internal recording solution contained in (mM) 113 K gluconate, 2.75 MgCl₂, 1.75 MgSO₄, 0.1 EGTA, 14 Tris₂-phosphocreatine, 4 Na₂-ATP, 0.3 Tris-GTP, 9 HEPES with pH adjusted to 7.25 with KOH, mOsm adjusted to 290 with sucrose (E_Cl, −84 mV). For voltage clamp to isolate NALCN current, internal solution contained in (mM) 87 CsMeSO₃, 18 CsCl, 5 CsF, 10 TEA-Cl, 10 HEPES, 5 EGTA, 5 Mg-ATP, 0.3 Na₂-GTP, 13 di-Na phosphocreatine, 2 QX-314 (pH 7.25, 295 mOsm). For a few voltage-clamp experiments, we used an internal solution containing (in mM) 103 CsCl, 10 TEA-Cl, 2.75 MgCl₂, 9 HEPES, 0.1 EGTA, 0.3 Tris-GTP, 14 Tris₂-phosphocreatine, 4 Na₂-ATP, 3.5 QX-314 (pH adjusted to 7.2 with CsOH). Puff application of agonists and antagonists was delivered through a picospritzer (Picospritzer III; Toohey Company, Fairfield, NJ), at 7-10 psi, with borosilicate glass capillaries. NA or Baclofen applications were at 100 µM and 5-10 ms in duration. The puff pipette was placed around 100 µm from the soma of the recorded cell to avoid mechanical disturbance.

Cell-attached (voltage-clamp) recordings were made using normal extracellular solution. Whole-cell patch-clamp recordings were made using a Multiclamp 700B amplifier and pCLAMP 10 software (Molecular Devices; Sunnyvale, CA). Signals were digitized at 20–40 kHz and filtered at 10 kHz by Digidata 1440A (Molecular Devices). In voltage clamp, cells were held at −65 mV, with access resistance 5-30 MΩ compensated to 40– 60%. In current clamp with control solutions, the resting membrane potential was maintained at −60 to −70 mV with bias current. To isolate NALCN currents, synaptic blockers, NBQX (10 µM), MK-801 (10 µM), SR-95531 (10 µM) or picrotoxin (100 µM), strychnine (0.5 µM), apamin (100 nM), and BaCl₂ (200 µM) were added to the bath solution. In those experiments, shifts in Ca²⁺ from 2 mM to 0.1 mM were accompanied by a shift in Mg²⁺ from 1 mM to 3 mM. To record evoked IPSCs, CWCs were stimulated with brief voltage pulses (100 µs) using a stimulus isolation unit (Iso-Flex; A.M.P.I, Jerusalem, Israel) via a glass microelectrode placed in the molecular layer of the DCN.

Pharmacology

All drugs in the slice experiments were bath applied. Receptor antagonists used in this study included: NBQX (AMPA receptors; Sigma-Aldrich, St. Louis, MO), MK-801 (NMDA receptors; Sigma-Aldrich), SR-95531 (GABA_AR; Tocris Bioscience, Bristol, UK), strychnine (glycine receptors; Sigma-Aldrich).

ABRs

Auditory brainstem responses (ABR) were acquired from C57B6/J and NALCN KO mice between P40-50. Mice were anesthetized with a dose of 80 mg/kg ketamine:16 mg/kg xylazine. ABRs were recorded differentially with electrodes at the vertex and pinna with ground at the base of the tail. Tone pips (5 msec, 0.5 msec rise/fall with 4-msec steady-state plateau) were generated digitally with a PXI-4461 card (National Instruments, Austin, TX), amplified (SA1; Tucker-Davis Technologies, Alachua, FL) and delivered using a compact, close-field sound system consisting of two CUI earphones and an electric microphone (FG-23329-P07; Knowles, Itasca, IL) coupled to a probe tube (Buran, 2015; Hancock K et al., 2015; Buran et al., 2020). Tone levels were incremented in 5-dB steps from 0 to 90-dB SPL. ABRs were recorded in a sound-proof booth using a Grass P511 amplifier and digitized using a National Instruments PXI 4661 card. ABR threshold was obtained for each animal. ABR peaks and thresholds were identified by visual inspection of the waveforms in a custom-written analysis program (Buran, 2015). Body temperature was maintained between 36 and 37°C using a homeothermic blanket (Harvard Apparatus, Cambridge, MA).

Experimental design and statistical analyses

Electrophysiological data were analyzed using pClamp 10.4 software (Molecular Devices), Axograph, or IGOR Pro v6.3 or v8 (WaveMetrics, Lake Oswego, OR) and NeuroMatic (Rothman and Silver, 2018).Figures were made using IGOR Pro, Affinity Designer, and Adobe Illustrator. Statistics were performed in IGOR Pro, Axograph, Python, Microsoft Excel, or Prism 7 (GraphPad, San Diego, CA). Error bars are represented as mean ± SEM unless otherwise stated.

Acknowledgements

This work was supported by NIH grants R35NS116798 and DC004450 to L.O.T. and JSPS KAKENHI Grant Number 22K06838 to T.I. We thank Drs. John Williams and members of the Trussell Lab for comments.