Loss of β4-spectrin impairs Nav channel clustering at the heminode and temporal fidelity of presynaptic spikes in developing auditory brain

Beta-4 (β4)-spectrin, encoded by the gene Sptbn4, is a cytoskeleton protein found at nodes and the axon initial segments (AIS). Sptbn4 mutations are associated with myopathy, neuropathy, and auditory deficits in humans. Related to auditory dysfunction, however, the expression and roles of β4-spectrin at axon segments along the myelinated axon in the developing auditory brain are not well explored. We found during postnatal development, β4-spectrin is critical for voltage-gated sodium channel (Nav) clustering at the heminode along the nerve terminal, but not for the formation of nodal and AIS structures in the auditory brainstem. Presynaptic terminal recordings in Sptbn4geo mice, β4-spectrin null mice, showed an elevated threshold of action potential and increased failures during action potential train at high-frequency. Sptbn4geo mice exhibited a slower central conduction and showed no startle responses, but had normal cochlear function. Taken together, the lack of β4-spectrin impairs Nav clustering at the heminode along the nerve terminal and the temporal fidelity and reliability of presynaptic spikes, leading to central auditory processing deficits during postnatal development.


Results
β4-spectrin expression and Na v channel clustering at the heminode occur at a later time point than at the AIS and nodes in the auditory brainstem during postnatal development. The myelinated axon of globular bushy cell in the cochlear nucleus, referred as the calyx axon, crosses over the midline and projects contralaterally to the MNTB. We examined β4-spectrin expression in axonal domains (e.g. AIS, nodes, and the heminode) in the MNTB from mouse auditory brainstem (P25). MNTB neurons were immunostained with MAP2 and the calyx axons were anterograde-traced with tetramethylrhodamine dextran staining 15 , Fig. 1A). β4-spectrin was distinctly expressed in the AIS of MAP2 positive neurons, nodes of Ran- www.nature.com/scientificreports/ vier, and the heminode next to the calyx terminal, but not in the presynaptic terminal in the MNTB (Supplemental Fig. 1). β4-spectrin expression has been linked to Na v channel expression at the AIS and nodes in the sciatic nerve, spinal cord, hippocampus, cerebellum and cerebral cortex 8,11 . Similarly, in the auditory brainstem, β4-spectrin was well co-localized with Na v channels at the heminode near the calyx terminal (at P25, Fig. 1B). β4-spectrin expression was also associated with ankyrin-G (AnkG), which is required for β4-spectrin localization at nodes 16 . In the MNTB, β4-spectrin was well co-localized with AnkG at the AIS and nodes as early as the first week of postnatal age (P9, Supplementary Fig. 1). Notably, in immature calyx terminals at P8-10 when AnkG immunoreactivity was not distinctly detected at the heminode, Na v channels at the heminode were sparsely expressed, not forming the Na v channel cluster 6 . We examined whether β4-spectrin expression at the heminode is associated with Na v channel clustering in immature calyx axon during postnatal development (at P10). A clustering region with relatively high-intensity staining for Na v channels was detected at nodes (normalized fluorescence intensity > 0.60) as described in a previous study 6 . In young mice brainstem (at P9) the clusters of AnkG, β4-spectrin, and Na v channels were distinctly formed at nodes. In contrast, the expression of β4-spectrin and AnkG were very weak and dispersed along the heminode in the immature calyx. Accordingly, Na v channels were not formed as a cluster at the heminode yet (Fig. 1C, Supplementary Fig. 1). Thus, during postnatal development, β4-spectrin expression and clustering at the heminode along the nerve terminal occurs at a later timepoint than at the AIS or nodes in the auditory brainstem. β4-spectrin is important for Na v channel clustering at the nerve terminal during postnatal development. To determine whether β4-spectrin is a key molecule for structural refinement of the nerve terminal with Na v channel clustering during postnatal development, we examined how loss of β4-spectrin impacts the heminode, utilizing the β4-spectrin null mice (Sptbn4 geo ). These mice were generated by gene-trap mutagenesis, where expression of the β4-spectrin protein was disrupted by ROSAβgeo* insertion 8,10 . β4-spectrin is encoded by the Sptbn4 locus on chromosome 19q13.2 3 . In Sptbn4 geo mice, there was no detectable β4-spectrin expression in the MNTB located in the mouse auditory brainstem. MAP2-positive principal neurons and the calyx axon projections to the MNTB were morphologically not altered (at P25, Fig. 2A).
In WT mice, the initially broad expression of Na v channels at the heminode along the calyx terminal coalesces into a cluster during the second week of postnatal age 6 . However, Sptbn4 geo mice did not show this structural refinement of Na v channel clustering at the heminode. The calyx terminal from Sptbn4 geo mice displayed a significantly broader expression of Na v channel along the heminode compared to the clustered Na v expression observed at nodes (Fig. 2B). The Na v channel cluster size at the heminode was 20.3 ± 1.9 µm in Sptbn4 geo mice (n = 42 cells from 6 mice) and 8.0 ± 1.0 µm in WT mice (n = 44 cells from 13 mice, p < 0.001, Mann-Whitney U test) without changes in the heminode length, defined as the distance between the last paranode immunostained with Caspr and the calyx terminal (12.7 ± 1.1 µm in Sptbn4 geo mice n = 33 from 6 mice vs 11.5 ± 1.4 µm in WT mice, n = 21 from 6 mice, P24-P34, p = 0.47, Mann-Whitney U test, Fig. 2C). The size of nodes, flanked by Caspr-labelled paranodes, was measured by the length of Na v cluster. The loss of β4-spectrin had no effect on the length and width of nodal Na v cluster in Sptbn4 geo mice (2.0 ± 0.14 µm, n = 38 from 3 Sptbn4 geo mice vs 1.7 ± 0.09 µm, n = 33 from 4 WT mice, p = 0.19, Mann-Whitney U test, Fig. 2C). The node length was not dependent on node diameter, determined by the width of Na v cluster width. The slope of node length against node diameter was − 0.02 ± 0.27, n = 38 from in 3 Sptbn4 geo mice and 0.13 ± 0.16, n = 33 from 4 WT mice (p = 0.64, linear regression test, data not shown). Furthermore, we evaluated myelin thickness and axon diameter of axons in the MNTB, which include the calyx axons from the cochlear nucleus and MNTB axons using transmission electron microscopy (TEM). To evaluate myelin thickness, the g-ratio was measured as the inner radius of the axon bundle divided by the outer radius, therefore an increase in myelin thickness would be indicated by a lower g-ratio. There was no significant difference between the g-ratios of WT and Sptbn4 geo axon bundles (0.83 ± 0.002 g-ratio, n = 440 axon bundles in 3 WT mice, 0.83 ± 0.002 g-ratio, n = 636 axon bundles in 4 Sptbn4 geo mice, at P25, p = 0.11, Mann-Whitney U test, Fig. 2D, E) nor in the inner diameter length of the corresponding axon bundles (1.57 ± 0.10 µm in WT, 1.36 ± 0.06 µm in Sptbn4 geo , p = 0.91, Mann-Whitney U test, Fig. 2D, E). These results demonstrate that the loss of β4-spectrin has no significant impact to myelin thickness or axonal diameter in the auditory brainstem. Overall, these results indicate that β4-spectrin is required for the clustering of Na v channels at the heminode, but is not essential for the formation of nodes in myelinated axons in the auditory brainstem.
β4-spectrin at the heminode is important for maintaining the reliability and temporal fidelity of presynaptic spikes at the nerve terminal. The disruption of Na v channel clustering at heminode impaired the fidelity of presynaptic action potential (AP) and increased the failure rate during a high-frequency spiking 5,17 . To determine how loss of β4-spectrin at the heminode impacts presynaptic excitability, presynaptic AP was elicited from the calyx terminal by supra-threshold current injection in whole-cell recordings (Fig. 3A, B). Our primary measures of excitability were taken from phase plots (Fig. 3B, C), plots of the membrane potential slope (dV/dt) versus the membrane potential 18 . In both WT (n = 6 mice) and Sptbn4 geo mice (n = 12 mice, at P14-16), all cells displayed a single inflection in the rising phase of the AP (Fig. 3C), indicating that APs generate at the heminode next to the presynaptic terminal. We used a primary criterion of the membrane potential at a phase plot slope of 10 mV ms -1 to estimate the threshold of presynaptic AP. The AP threshold was significantly higher in Sptbn4 geo mice than WT mice (− 43.3 ± 6.12 mV, n = 15 cells vs − 48.6 ± 3.79 mV, n = 10 cells, respectively, p = 0.012, Mann-Whitney U test) without changes in AP amplitude (81.8 ± 9.16 mV, n = 14 cells for Spt-bn4 geo vs 82.4 ± 4.96 mV, n = 9 cells for WT, p = 0.816, Fig. 3B-D). In the phase plot, the maximum (Max) of dV/ dt was lower in Sptbn4 geo mice (527 ± 127.2, n = 12 cells for Sptbn4 geo vs 703 ± 69.5, n = 6 cells for WT, p = 0.017, Mann-Whitney U test), indicating that the rise time of AP was slower in Sptbn4 geo mice than WT mice (Fig. 3E). The rheobase, depolarizing currents required to generate APs, was significantly larger in Sptbn4 geo mice com- www.nature.com/scientificreports/ pared to WT mice (100.2 ± 34.92 pA, n = 14 cells vs 58.4 ± 15.03, n = 9 cells, respectively, p = 0.004, Mann-Whitney U test, Fig. 3F). There was no difference in the resting potential (− 64.6 ± 3.82 mV, n = 17 cells for Sptbn4 geo vs − 66.9 ± 2.01 mV, n = 12 cells for WT, p = 0.06, Mann-Whitney U test, Fig. 3G). An elevated threshold, a slower rising of spike, and a higher rheobase at the calyx terminal indicates that lack of β4-spectrin reduced presynaptic excitability.
In addition, we examined how loss of β4-spectrin impacts the reliability and temporal fidelity of presynaptic terminal spikes during high-frequency stimulation (Fig. 3H). In WT mice, the calyx terminal was able to follow www.nature.com/scientificreports/ high-frequency stimulation and displayed spikes without failures at 100 Hz and 200 Hz. The calyx terminal in Sptbn4 geo mice showed a number of spike failures at 200 Hz (~ 30%) through 500 Hz (~ 80%). Although the failure rate was increased throughout the stimulus frequencies in both genotypes, the number of spikes in response to high-frequency stimulus were significantly decreased the calyx terminal in Sptbn4 geo mice (at 200 Hz, 33.1 ± 2.5 spikes/50 stimuli in Sptbn4 geo mice, n = 16 cells vs 49.3 ± 0.7 spikes in WT, n = 10 cells, p < 0.0001, twoway ANOVA). These results demonstrate that the presence of β4-spectrin at the heminode is critical for maintaining the reliability and temporal fidelity of presynaptic spikes during early postnatal development.
Functional relevance of β4-spectrin at the nerve terminal in central auditory processing. To further investigate how structural and functional alterations at the calyx axon terminal caused by β4-spectrin loss impact the auditory transmission throughout the central auditory pathway, we examined the auditory brainstem responses (ABRs) from Sptbn4 geo mice (P21-P25) in response to click and pure tone stimuli with varying sound pressure levels (90 to 20 dB SPL, Fig. 4A, B). There was a significant increase in ABR thresholds of Sptbn4 geo mice compared to WT mice when presented with click stimuli (54 ± 4.1 dB in Sptbn4 geo mice and 34 ± 3.4 dB in WT mice, n = 10 and 10, respectively, p < 0.05, two-way ANOVA with Bonferroni's test). When presented with pure tones (4,8,16, and 32 kHz), the thresholds were overall much higher in the Sptbn4 geo mice than the WT mice (n = 10 mice for each group Next, we examined peripheral and central conduction by evaluating the temporal acuity among the peaks in ABRs. Conduction was measured and determined as interpeak intervals peripherally (time between peaks I and II of ABR waves) and centrally (time between peaks II and IV). The time difference between peaks I and II of ABR waves in the WT and Sptbn4 geo mice were 0.84 ± 0.073 ms and 1.18 ± 0.082 ms, respectively, showing no significance difference (n = 10 and 9, respectively, p > 0.05, two-way ANOVA with Bonferroni's test, Fig. 4D).
These results indicate that peripheral conduction was not altered in Sptbn4 geo mice, which is consistent with the previous work showing a normal functioning periphery in quivering mice 19,20 . For the central conduction, Sptbn4 geo mice had significantly greater interpeak intervals of 3.2 ± 0.29 ms (n = 9 mice) versus 1.9 ± 0.067 ms in WT mice (n = 10 mice, p < 0.0001, two-way ANOVA with Bonferroni's test). The prolonged interval indicates that central conduction was slowed down and major signal transmission mechanisms were impaired in the central auditory pathway in Sptbn4 geo mice. The results were comparable with those observed in patients lacking β4-spectrin 3 . In addition, we analyzed signaling strength in the auditory pathway by quantifying amplitudes of individual ABR waves as the µV difference between the positive peak and the preceding negative peak. The ABRs from Sptbn4 geo mice displayed significantly reduced amplitudes of waves II and III, which arise from neuronal activity in the cochlear nucleus (CN) and the superior olivary complex (SOC) (p < 0.0001 and p < 0.05, respectively, two-way ANOVA, Fig. 4E-F). Reduced ABR amplitude indicates that neuronal activity or synaptic synchrony in the CN and SOC are impaired in Sptbn4 geo mice. These results are consistent with the disruption of heminodal Na v channel clustering and loss of the temporal fidelity of presynaptic spikes found in these mice, which causes these central hearing deficits.
To determine how the disruption in temporal fidelity of auditory impulses and transmission caused by Sptbn4 loss influences central auditory processing, we tested the ability of mice to startle to loud noise stimuli using the acoustic startle response (ASR) test. Intriguingly, Sptbn4 geo mice lacking β4-spectrin were unable to elicit a startle response to any of the noise stimuli presented, whereas WT mice clearly showed a startle response with the threshold at 105 dB (p < 0.001, two-way ANOVA with Bonferroni's test, Fig. 4G). Using 120 dB startle noise as the measurement for peak startle amplitude, there was a significant difference in the magnitude of the startle in Sptbn4 geo mice compared to WT mice (60.8 ± 5.6 mV at 120 dB, n = 7 Sptbn4 geo mice vs 911.3 ± 131.7 mV at 120 dB, n = 10 WT mice, p < 0.0001, Mann Whitney U test, Fig. 4G). To examine whether mouse body weight may influence no ASRs in Sptbn4 geo mice, the weights of the mice undergoing ASR testing were measured. There was no significant difference (8.3 ± 0.7 g in Sptbn4 geo mice, 7.9 ± 0.3 g in WT mice, p = 0.56, unpaired t-test, Fig. 4H). To confirm that cochlear function is normal and that cochlear deficits are not responsible for the lack of a response to startle in Sptbn4 geo mice, we performed the in vivo distortion product otoacoustic emissions (DPOAE) test, which measures outer hair cell function. There was no significant difference in the distortion products of pure tones presented at 8, 12, 16 or 32 kHz in Sptbn4 geo mice compared to WT mice (39.5 ± 8.6 dB SPL at 8 dB, 40.6 ± 7.5 dB SPL at 12 dB, 40.3 ± 6.6 dB SPL at 16 dB, 19.7 ± 2.9 dB SPL at 32 dB in Sptbn4 geo mice, n = 7 mice, 49.4 ± 4.5 dB SPL at 8 dB, 51.9 ± 3.5 dB SPL at 12 dB, 45.3 ± 4.6 dB SPL at 16 dB, 26.2 ± 3.3 dB SPL at 32 dB in WT mice, n = 9 mice, p = 0.87, two-way ANOVA, Fig. 4I), indicating that the sensory output of outer hair cell function in Sptbn4 geo mice is normal. Taken together, the results suggest Sptbn4 geo mice have relatively normal cochlear function, but have delayed and reduced central auditory transmission, which results in central auditory processing deficits.

Discussion
We found that β4-spectrin expression and clustering at the nerve terminal occurs at a later time point compared with those at the AIS or at nodes in the auditory brainstem during postnatal development. The location of β4-spectrin is temporally associated with the structural refinement of the nerve terminal around hearing onset (P13) consistent with β4-spectrin being critical for Na v channel clustering at the heminode. Although the loss of β4-spectrin severely impairs Na v channel clustering at the heminode, Na v channel clustering at nodes and   www.nature.com/scientificreports/ synchronized neuronal activities, delayed conduction in auditory brainstem circuitry, and impaired central auditory processing. In Sptbn4 geo mice, loss of β4-spectrin disrupted Na v channel clustering, specifically at the heminode without alterations in nodal and AIS structure, leading to elevation of AP threshold and increased failures during AP train. The results indicate that heminodal structure is critical for tuning of presynaptic spikes during postnatal development of the auditory brainstem. The correlation between the disruption of Na v channel clustering at heminode and impaired temporal fidelity of presynaptic spikes was previously observed in other neuropathological models such as dys-/demyelination models 5,17 . Here, we showed that the auditory brainstem of Sptbn4 geo mice has normal myelination, indicating loss of temporal fidelity of auditory transmission are likely due to disruption of the heminodal structure and presynaptic failures rather than an alteration in myelination. Notably, Na v channel expression at nodes and the AIS in the auditory brainstem were not altered in Sptbn4 geo mice (at P25), supporting the previous finding that β4-spectrin is not essential for the targeting and localization of Na v channels at nodes or AIS formation 16 . However, adult mice without β4-spectrin (> 3 months old Sptbn4 geo mice) showed a loss of AnkG and Na v channels from the AIS, and the disruption of nodal structure 8,10 . Thus, β4-spectrin is essential for maintaining membrane structure and the proper molecular organization of nodes, the AIS, and the axonal cytoskeleton 7,21 . During brain development, nodes and the AIS form properly in Sptbn4 geo mice, but in the adulthood or during aging, the lack of β4-spectrin results in the destabilization of the nodal and AIS structures 7,8 . Here, we found β4-spectrin is essential for Na v channel clustering at the heminode rather than at nodes and AIS in the auditory brainstem during early development. The structural and functional alterations in the nerve terminal caused by loss of β4-spectrin critically impacts central auditory processing. The result suggests that the spatial and temporal expression of β4-spectrin at axon segments during development are important to determine functional relevance in different brain regions.
Central auditory processing disorders (CAPD) manifest as difficulties in certain auditory tasks such as sound localization, temporal integration, discrimination, and auditory performance with competing acoustic signals (e.g. listening in noise), without significant loss in hearing sensitivity 22 . CAPD is frequently observed in more complex systemic disorders including specific language impairment, dyslexia, attention deficits/hyperactivity disorders, and autism spectrum disorder 4 . Speech-evoked ABR onset-waves in children with CAPD had significantly longer latencies with larger variability than in their age-matched controls 23 . Alterations in temporal features of sound, which are represented by first spike latencies and their variability, impact auditory temporal processing 4 . Thus, temporal precision is a critical element for central auditory processing. However, the cellular mechanisms underlying CAPD remain unknown, due to the limited availability of animal models with CAPD.
Utilizing in vivo auditory function tests in Sptbn4 geo mice, we observed that there was no alteration in cochlear function with normal DPOAEs, significant impairments in ABRs, and complete loss of ASRs in Sptbn4 geo mice during postnatal development. Those in vivo tests evaluated the outer hair cell function, neuronal activity and conduction along the auditory processing pathway, and its sensorimotor gating ability, respectively. The results indicate that Sptbn4 geo mice may have impairments in sound localization and auditory processing, although they have a normal cochlear function. Notably, phenotypes in Sptbn4 geo mice are focused on central deficits in the auditory brainstem, although the peripheral auditory nerve deficits can be involved. We found that temporal fidelity of APs at the nerve terminal was impaired by dispersion and failures of spikes in Sptbn4 geo mice. The loss of temporal accuracy of APs in central processing disorders could be caused by myelin loss or deficits in myelination. For example, the structural alteration of the axon heminode and the loss of the fidelity of presynaptic spikes has been demonstrated in dys-/demyelinating axons 17 . In adult mice (> 9-month-old), loss of nodal β1-and β4-spectrins altered axon diameter and myelin thickness in the peripheral nervous system, indicating an axon injury response and motor dysfunction 13 . However, ultrastructural analysis using TEM showed that loss of β4-spectrin did not affect axonal integrity and myelin thickness in young Sptbn4 geo mice (~ 1 month old, Fig. 2). Thus, in juvenile mice, the loss of temporal fidelity at the nerve terminal critically contributes to central conduction and processing deficits in the auditory brainstem independent of myelin loss or axonal degeneration. The results suggest that alterations in Na v channel clustering at the heminode without the disruption of nodal structure, axon integrity, and myelin can significantly affect auditory processing during development.
Humans and mice with pathogenic Sptbn4 variants have impaired motor function, neuropathy, as well as auditory deficits [1][2][3] . Surprisingly, Sptbn4 geo mice completely lost the ability of acoustic startle response. In addition to central auditory deficits, motor dysfunctions may play a role in the inability of acoustic startle response for the Sptbn4 geo mice. In a previous study, the Sptbn4 geo mice appeared normal besides fine tremors and a progressively worsening hindlimb paralysis defined by hindlimb contraction when lifting the mice by the tail was observed beginning at 2-3 months of age 8 . However, we did not observe the clenching of the hindlimb in Sptbn4 geo mice at P22-P23, when ASR testing was performed (data not shown). The possibility of motor deficits playing a role in the lack of a startle reflex observed in the Sptbn4 geo mice is unlikely, but more likely due to auditory processing deficits. Severe motor dysfunction in Sptbn4 geo observed in adulthood might be caused by complex deficits including nodal disruption, myelin loss, and axonal degeneration 13 . Here we demonstrate that the lack of β4-spectrin critically impaired central auditory processing during postnatal development, prior to severe neurological phenotypes in the adult including motor dysfunction.

Materials and methods
Animals. Sptbn4 geo mice (β4-spectrin null mice), which were previously characterized in several groups 8,10 , were used. Sptbn4 geo mice were provided by Dr. Bhat's laboratory (UTHSCSA). Identification was based on genotyping, with + / + individuals used as WT mice. Pups of either sex were used for presynaptic recordings (at P14-16) and for in vivo ABR, ASR, and DPOAEs test and immunostaining (P22-P34). Experiments were in accordance with animal welfare laws, complied with ARRIVE guidelines and protocols approved by the Univer- In vivo auditory brainstem response (ABR) test. ABR recordings were performed as described (Kim et al., 2013). Briefly, mice were anesthetized with 3.5% isoflurane and maintained with 2.5% isoflurane during recording (1 l/min O 2 flow rate). ABR recordings were performed in a sound attenuation chamber (Med Associates, Albans, VT). Subdermal needle electrodes (Rochester Electro-Medical, Lutz, FL) were placed on the top of the head, ipsilateral mastoid, and contralateral mastoid as the active, reference, and ground electrode, respectively. The signal differences in the ABRs between the vertex and the mastoid electrodes were amplified and filtered (100-5000 Hz). Acoustic stimuli were generated by an Auditory Evoked Potentials Workstation (Tucker-Davis Technologies [TDT], Alachua, FL). Closed-field click stimuli were presented to the left ear. The signals consisted of a series of amplitude-modulated square waves (0.1 ms duration, 16/s) through TDT Multi-Field Magnetic Speakers. The sound stimuli were delivered through a 10-cm plastic tube (Tygon; 3.2-mm outer diameter) at a repeat rate of 16/s. Sound intensities ranged from 90 to 20 dB, with 5-dB decrements, and responses to 512 sweeps were averaged.
Acoustic startle response test. Mice of either sex between ages P22-P23 were put in a Plexiglas holding cylinder located in sound-attenuated chamber using the SR-LAB startle response system (San Diego Instruments, San Diego, CA). Weights of mice were recorded to account for sex or size differences. Sound levels from the chamber speakers were calibrated with a digital sound level meter (Part Number MS-M80A, Mengshen). Each trial consisted of an initial acclimation period of 5 min of background level noise (at 65 dB) followed by 5 rounds of randomized noise stimuli playing for 20 ms with 10 s of background between each noise stimuli at 65 dB (background), 95 dB, 105 dB, 115 dB, and 120 dB. Stimuli was produced by a digital signal processingcontrolled system amplified and emitted by a loudspeaker. Startle responses were measured inside the soundattenuated chamber by a movement-sensitive piezo-accelerometer platform. The maximum was used as the startle amplitude at each tone with millivolts (mV) as the unit of measurement and averaged per mouse.

Distortion product otoacoustic emissions (DPOAE) test.
Mice were anesthetized with 3.5% isoflurane and maintained with 2.5% isoflurane during recording. DPOAE recordings were performed in a sound attenuation chamber (Med Associates, Albans, VT). Acoustic stimuli were generated by an Auditory Evoked Potentials Workstation (Tucker-Davis Technologies [TDT], Alachua, FLO). The ER-10B + recording microphone (Etymotic Research, Elk Grove Village, IL) with ear-tip was inserted into the ear canal. The sound stimuli were delivered through two TDT Multi-Field Magnetic Speakers connected to the recording microphone by 10-cm coupling tubes (TDT; 3.2 mm outer diameter). Pure tones were presented at 20% frequency separation between f1 and f2 at 8, 12, 16, and 32 kHz. Sound intensities ranged from 80 to 20 dB, with 10-dB decrements, and responses to 512 sweeps were averaged. Distortion products were calculated as 2f1-f2 minus the noise floor that were detected by the recording microphone and amplified by RZ6 processor (TDT).
Electrophysiology. Slices were perfused with normal aCSF at 2 ml/min and visualized using an infrared differential interference contrast microscope (AxoExaminer, Zeiss, Oberkochen Germany) with a 63 × waterimmersion objective and a CMOS camera (ORCA-Flash2. 8 www.nature.com/scientificreports/ needle, which were then incubated in normal aCSF bubbled with carbogen at 37 °C for 30 min. All slices were fixed with 4% (w/v) paraformaldehyde in PBS for 10 min. Free-floating sections were blocked in 3% goat serum and 0.3% Triton X-100 in PBS for 30 min. Slices were incubated with the primary antibody overnight at room temperature. The following primary antibodies were used: mouse anti-sodium channel (PanNa; 1:400; Sigma), rabbit anti-calretinin (CR; 1:200; Invitrogen), guinea pig anti-Caspr (Caspr; 1:200; from Dr. Bhat's laboratory, UTHSCSA), rabbit anti-β4-spectrin (β4; 1:200; from Dr. Bhat's laboratory, UTHSCSA), mouse anti-ankyrinG (AnkG; 1:200; Neuromab). Antibody labeling was visualized by incubation with appropriate Alexa dye-conjugated secondary antibodies (1:500; Invitrogen) for 2 h at room temperature. Stained slices were viewed with laser lines at 488 nm, 568 nm, and 647 nm using a 40x/1.40 or 63 × /1.40 oil-immersion objective on a confocal laser-scanning microscope (LSM-710; Zeiss). Stack images were acquired at a digital size of 1024 × 1024 pixels with optical section separation (z-interval) of 0.5 µm and were later crop to the relevant part of the field without changing the resolution. The heminode length was measured from the point at the outer edge of the principal neuron at the neck of the calyx terminal to the beginning of detectable Caspr at the heminode. The Na v channel cluster size at the heminode was measured as the length from beginning to end of the Na v cluster expression, and the paranodal length and width was measured from Na v channel expression mentioned previously 6 . The confocal image stacks were analyzed using ImageJ software.
Transmission electron microscopy. The brain was removed and immersed in ice-cold low-calcium aCSF (mentioned previously in slice preparation) and the MNTB from the brainstem was dissected at 200 µm-thick section using a microtome (VT1200s, Leica), then fixed in a 4% Formaldehyde-1% Glutaraldehyde solution and stored at 4 °C. Further processing was performed by the UTHSCSA Electron Microscopy Lab as previously described 25,26 . Axon bundle images were analyzed at a final magnification of 8000X, with the longest length of the axon measured as the inner diameter and the inner radius divided by the outer radius as the g-ratio.

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).