RIM-Binding Proteins Are Required for Normal Sound-Encoding at Afferent Inner Hair Cell Synapses

The afferent synapses between inner hair cells (IHC) and spiral ganglion neurons are specialized to faithfully encode sound with sub-millisecond precision over prolonged periods of time. Here, we studied the role of Rab3 interacting molecule-binding proteins (RIM-BP) 1 and 2 – multidomain proteins of the active zone known to directly interact with RIMs, Bassoon and CaV1.3 – in IHC presynaptic function and hearing. Recordings of auditory brainstem responses and otoacoustic emissions revealed that genetic disruption of RIM-BPs 1 and 2 in mice (RIM-BP1/2–/–) causes a synaptopathic hearing impairment exceeding that found in mice lacking RIM-BP2 (RIM-BP2–/–). Patch-clamp recordings from RIM-BP1/2–/– IHCs indicated a subtle impairment of exocytosis from the readily releasable pool of synaptic vesicles that had not been observed in RIM-BP2–/– IHCs. In contrast, the reduction of Ca2+-influx and sustained exocytosis was similar to that in RIMBP2–/– IHCs. We conclude that both RIM-BPs are required for normal sound encoding at the IHC synapse, whereby RIM-BP2 seems to take the leading role.


INTRODUCTION
The ribbon-type active zones (AZ) of inner hair cells (IHCs) are molecularly specialized to ensure temporally precise encoding of incoming sound stimuli into high-frequency firing of postsynaptic spiral ganglion neurons (SGNs). In mature mouse IHCs, neurotransmitter release from synaptic vesicles (SV) is controlled by the tight spatial coupling of Ca V 1.3 voltage-gated L-type Ca 2+channels and the SV release machinery (Brandt et al., 2005;Wong et al., 2014;Pangrsic et al., 2015). A small, defined SV pool with fast release kinetics, referred to as the readily releasable pool (RRP), is much less sensitive to the slow-binding Ca 2+ -buffer EGTA than to the fast-binding Ca 2+ -buffer BAPTA, emphasizing the Ca 2+ -nanodomain-like control of SV release (Moser and Beutner, 2000;Brandt et al., 2005;Pangrsic et al., 2015). There is also evidence for Ca 2+ nanodomain-like coupling in hair cells of other species, such as in frog auditory hair cells (Graydon et al., 2011), in rat IHCs (Goutman and Glowatzki, 2007) and IHCs of the low frequency apical cochlea of the gerbil, while a looser Ca 2+ microdomain-like coupling was reported for the high-frequency basal IHCs (Johnson et al., 2017) and immature mouse IHCs (Wong et al., 2014).
Aside from the sophisticated Ca 2+ -channel complex (reviewed in Pangrsic et al., 2018), the release of SVs in auditory IHCs also requires a finely coordinated, complex presynaptic protein network including bassoon, piccolo, Rab3-interacting molecule (RIM) 2α/2ß/3γ and RIM-binding protein (RIM-BP), organizing the precise AZ topography of Ca 2+ -channels and SV release sites (reviewed in Moser et al., 2019). RIM-BPs seem to take a central role in a large presynaptic multiprotein-complex involving voltage-gated Ca 2+ -channels, RIMs and bassoon (Hibino et al., 2002;Kaeser et al., 2011;Davydova et al., 2014;Acuna et al., 2015;Ortner et al., 2020). Specifically, the RIM-BPs' two C-terminal SH3 domains interact with proline-rich motifs in the C-terminus of the Ca V α 1D subunit and RIMs (Wang et al., 2000;Hibino et al., 2002;Kaeser et al., 2011;Ortner et al., 2020;Petzoldt et al., 2020), while the N-terminal SH3 domain interacts with the proline-rich motif of bassoon (Davydova et al., 2014). Morphological and functional studies from mouse IHCs suggest that these interactions are likely also applicable to IHC ribbon-type AZs. On the one hand, super-resolution microscopy studies on immunolabeled IHCs showed a specific stripe-like arrangement of the above mentioned interaction partners bassoon (Wong et al., 2014;Neef et al., 2018), RIM2 (Jung et al., 2015), RIM-BP2 (Krinner et al., 2017), and Ca V 1.3 Ca 2+ -channels (Frank et al., 2010;Neef et al., 2018) at IHC AZs. On the other hand, the individual genetic deletion of bassoon Frank et al., 2010;Neef et al., 2018), RIM2α and -ß (Jung et al., 2015) and RIM-BP2 (Krinner et al., 2017) significantly reduced the number of IHC presynaptic Ca V 1.3 Ca 2+ -channels. While these data show that all three proteins are important for Ca 2+ -channel clustering, none of the mentioned mutants exhibit a complete loss of synaptic Ca V 1.3 Ca 2+channels or exocytosis, suggesting partially overlapping and compensatory function between these presynaptic AZ proteins to ensure normal presynaptic IHC function. In the retina, for example, RIMs were found to be upregulated upon loss of RIM-BP1/2 (Luo et al., 2017), while nonetheless a significant reduction of synaptic Ca 2+ -channels was observed at rod bipolar cell ribbon-type AZs (Luo et al., 2017). The spectrum of effects of RIM-BP deletion on presynaptic function ranges across synapses. In Drosophila melanogaster neuromuscular junctions (NMJ), genetic disruption of DRPB (RIM-BP orthologue) causes a severe impairment of Ca 2+ -channel clustering Müller et al., 2015), while at conventional synapses of the mammalian CNS or Caenorhabditis elegans synapses, the number of P/Q-or N-type Ca 2+ -channels was not affected by the loss of RIM-BPs (Acuna et al., 2015;Grauel et al., 2016;Kushibiki et al., 2019). Genetic disruption of DRPB in D. melanogaster NMJ further affected the structural AZ integrity and functional coupling between Ca 2+ -channels and SVs, resulting in a drastically reduced SV release probability Müller et al., 2015). Such looser SV-Ca 2+ -channel coupling upon RIM-BPdisruption was also observed in conventional CNS synapses and retinal ribbon synapses (Acuna et al., 2015;Grauel et al., 2016;Luo et al., 2017), whereas the tight nanodomain-like coupling remained unaltered in IHC ribbon synapses at least after recovery of the RRP from depletion (Krinner et al., 2017).
RIM-BPs also contribute to the efficient replenishment of readily releasable SVs. In D. melanogaster NMJ, the N-terminal deletion of DRBP lead to impaired SV recruitment to release sites, mediated via DRBP -Bruchpilot (BRP)/ELKS/CAST interaction (Petzoldt et al., 2020). Impaired SV replenishment was also reported for calyx of Held (Acuna et al., 2015) and ribbon synapses (Krinner et al., 2017;Luo et al., 2017). For IHC AZs, this has been suggested to reflect a role of RIM-BP2 in registering new-coming SVs in nanoscale proximity of Ca 2+ channels (Krinner et al., 2017). Recently, a role of RIM-BP in SV priming via interaction with Munc13-1 has been suggested (Brockmann et al., 2020). As priming of SVs in IHCs seems to operate without Munc13 and CAPS priming proteins (Vogl et al., 2015), other protein interactions such as the one with CAST/ELKS (Petzoldt et al., 2020) remain to be studied for a potential involvement in SV replenishment. In support of the relevance of this interaction at mammalian synapses, we note that RIM-BP expression levels are linked to CAST/ELKS abundance. In hippocampal neurons, ELKS and RIM deletion caused reduced protein levels of RIM-BP2 (Wang et al., 2016), while ELKS was found to be upregulated in RIM-BP1/2 deficient retinae (Luo et al., 2017).
In vivo experiments on RIM-BP2 knockout mice (RIM-BP2 −/− ) revealed a mild synaptopathic hearing impairment (Krinner et al., 2017). Likewise, the genetic deletion of the above mentioned AZ proteins and RIM-BP interaction partners bassoon  and RIM2α (Jung et al., 2015) caused a significant, yet limited, elevation of hearing thresholds. Hence, the auditory system phenotype suggests overlapping and compensatory function of these AZ proteins, which is in agreement with the cell physiology. Yet, the consequences of RIM-BP2 loss-of-function might have been attenuated by the presence of other RIM-BP variants, such as RIM-BP1 or -3. Hence, we tested the presence and potential role of RIM-BP1 in cochlear function by comparing RIM-BP1/2 doubleknockout mouse line (RIM-BP1/2 −/− ) (Grauel et al., 2016), to the previously studied RIM-BP2 −/− mice (Krinner et al., 2017). We employed expression analysis, electrophysiology and systems physiology and, indeed, found a synaptopathic hearing impairment in RIM-BP1/2 −/− mice that exceeds that of RIM-BP2 −/− mice. In IHC physiology, additional deletion of RIM-BP1 caused a subtle impairment of RRP exocytosis not found in RIM-BP2 −/− IHCs, suggesting that both RIM-BPs (-1 and -2) are required for normal hearing and sound encoding at the IHC ribbon synapse.

Animals
RIM-BP1/2 double-knockout mice (RIM-BP1/2 −/− ) (Grauel et al., 2016) and C57BL/6 mice of either sex were used for experiments. Previously published data from RIM-BP2 knockout mice (RIM-BP2 −/− ) and their RIM-BP2 wild-type littermates (RIM-BP2 +/+ ) were used for comparison as indicated (Krinner et al., 2017). Electrophysiology, RNAscope and immunohistochemistry experiments were performed in 2-3 week-old mice (i.e. after hearing-onset). Systems physiology was carried out in 8-10 week-old mice. All experiments complied with national animal care guidelines and were approved by the University of Göttingen board for animal welfare and the animal welfare office of the state of Lower Saxony.
RT-PCR and Single-Cell Nested RT-PCR C57BL/6 mice at the age of postnatal days (p) 14 through 16 were used to determine the general expression of RIM-BPs in the organ of Corti and in single IHCs. For RT-PCR of organs of Corti we isolated total RNA from preparations of the organ of Corti and brain (control) using TRIzolReagent (Invitrogen). Reverse transcription was performed with SuperScriptII RT according to the manufacturer's instructions using oligo (dT) primers. Sequences of RIM-BP-specific primers are listed in Table 1 and only the first PCR was run with less than 30 cycles. For single-cell PCR, individual IHCs from the apical coils of freshly dissected organs of Corti were harvested after cleaning off supporting cells at a high bath perfusion rate (3 ml/min). Ten IHCs per trial were collected to check the expression of RIM-BP1 and -2. Each individual IHC was aspirated into a glass pipette and the pipette content was transferred into first strand cDNA synthesis mix containing after the dilution: 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 5 mM MgCl 2 , 5 mM DTT, 100 units of SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, United States) and 40 units RNaseOUT Ribonuclease inhibitor (Invitrogen). Reverse transcription was performed with oligo (dT) primers according to the manufacturer's instructions. Each cDNA mix was used as a template for two subsequent PCR reactions with nested primers specific for RIM-BP1 or RIM-BP2 cDNA. Instead of cDNA, distilled water was used in the negative control reaction. Single-cell nested RT-PCR was performed three times.

Patch-Clamp Recordings From Inner Hair Cells and Analysis
Perforated patch-clamp recordings were essentially carried out as described in Moser and Beutner (2000). Apical turns of 2-3 week-old RIM-BP1/2 −/− mouse organs of Corti were freshly dissected and whole-cell Ca 2+ -current and exocytosis from IHCs were recorded at room temperature (22-24 • C). The extracellular patch-clamp solution contained (in mM): 110 NaCl, 35 TEA-Cl, 2.8 KCl, 2 CaCl 2 , 1 MgCl 2 , 10 NaOH-HEPES, 11.3 D-glucose, pH 7.3. The internal pipette solution contained (in mM): 130 Cs-gluconate, 10 TEA-Cl, 10 4-AP, 10 HEPES, 1 MgCl 2 , amphotericin B (300 µg/ml), pH 7.2. The patch-clamp used an EPC-10 amplifier and Patchmaster software (HEKA Elektronik, Lambrecht, Germany). Voltages were corrected for liquid junction potentials (14 mV) and currents were leak-corrected using a p/10 protocol. For analysis Igor Pro software (Wavemetrics, Lake Oswego) was used. For membrane capacitance (C m ) measurements, IHCs were stimulated by depolarizations to −14 mV with intervals of 60-120 s to allow for recovery of IHC exocytosis. For currentvoltage relationships (IVs) measurements, currents were evoked by 10 ms step depolarizations to various potentials from −100 to +30 mV in 5 mV increments. IVs were calculated from the currents during the last 8 ms of the step depolarization. From these, fractional activation curves were calculated by calculating the Ca 2+ conductance from the Ca , with V the command potential and V rev the reversal potential of the current obtained from the x-axis crossing of an extrapolating line fit to the currents from 6 to 26 mV. After normalizing these traces to the maximum conductance in the range of −20 to 10 mV, they were fit with a Boltzmann equation with G n,max , the maximum conductance, V the command potential, V half the voltage of half-maximal activation, and k the slope factor. We measured C m changes ( C m ) using the Lindau-Neher technique (Lindau and Neher, 1988) as previously described (Moser and Beutner, 2000). Briefly, the exocytic C m was quantified as the difference of the averaged C m 400 ms before and after the depolarization. To avoid impact of C m -transients related to conductance or gating of ion channels on C m estimation (Moser and Beutner, 2000;Neef et al., 2007) we skipped the first 100 ms of post-depolarization C m for estimating the average. Mean C m and Ca 2+ -current estimates present grand averages calculated from the mean estimates of individual IHCs, where each depolarization was repeated 2-3 times. This avoided dominance of IHCs contributing more sweeps.

Systems Physiology: Auditory Brainstem Responses and Distortion Product Otoacoustic Emissions
Auditory brainstem responses (ABR) and distortion product otoacoustic emissions (DPOAE) were performed as described in Jing et al. (2013) and Strenzke et al. (2016). In this study, 8-10 week-old RIM-BP1/2 −/− mice were anesthetized with ketamine (125 mg/kg) and xylazine (2.5 mg/kg) i.p. and the core body temperature was maintained constant at 37 • C using a heat blanket (Hugo Sachs Elektronik-Harvard Apparatus). A TDT II System run by BioSig software (Tucker Davis Technologies) was used for stimulus generation, presentation, and data acquisition. With a JBL 2402 speaker, tone bursts (4/6/8/12/16/24/32 kHz, 10 ms plateau, 1 ms cos2 rise/fall) or clicks of 0.03 ms were presented ipsilaterally in the free field at 40 Hz (tone bursts) or 20 and 100 Hz (clicks). The difference potential between vertex and mastoid subdermal needles was amplified 50,000 times, filtered (400-4,000 Hz) and sampled at a rate of 50 kHz for 20 ms, 1,300 times, to obtain two mean ABR traces for each sound intensity. Hearing thresholds were determined with 10 dB precision as the lowest stimulus intensity that evoked a reproducible response waveform in both traces by visual inspection by two independent observers. For DPOAE, continuous primary tones [frequency f2 = 1.2 * f1, intensity l2 = l1 − 10 decibel (dB) Sound pressure levels (SPL)] were delivered through the MF1 speaker system (Tucker Davis Technologies) and a custom-made probe containing a MKE-2 microphone (Sennheiser). The microphone signal was amplified (DMX 6Fire, Terratec) and the DPOAE amplitude at 2 * f2-f1 was analyzed by fast Fourier transformation using custom-written Matlab software (Mathworks). SPL are provided in dB SPL root mean square (RMS) (tonal stimuli) or dB SPL peak equivalent (clicks).

Statistical Data Analysis
For statistical data analysis, Igor Pro software (Wavemetrics) and Origin software (Originlab) were used. Normality of distribution was tested with the Jarque-Bera test and variances were compared with the F-test. Unpaired, two-tailed Wilcoxon rank test (Mann-Whitney U test) was used to compare non-normal data or data with unequal variances, else Student's t-test was employed. For patch-clamp capacitance, ABR and DPOAE data, a one-way ANOVA was used for multiple comparisons followed by post hoc Tukey's test. Data are presented as mean ± SEM. Data from RIM-BP1/2 double-knockout mice (RIM-BP1/2 −/− ) are presented in magenta, data from RIM-BP2 knockout mice (RIM-BP2 −/− ) are presented in green, and data from their wildtype littermates (RIM-BP2 +/+ ) are presented in black.

RIM-BP1 Is Expressed in the Mouse Organ of Corti
Previous studies showed a localization of RIM-BP2 to ribbontype AZs of IHCs and presynaptic terminals of efferent lateral olivocochlear neurons (Krinner et al., 2017;Ortner et al., 2020). Such RIM-BP2 immunofluorescence was abolished in RIM-BP2-deficient IHCs demonstrating a specific labeling by the RIM-BP2 antibody. Here, we tested the hypothesis of a comparable expression pattern of RIM-BP1 in the mouse organ of Corti. For that we performed immunolabeling and confocal microscopy of mouse IHCs with two different triple antibody stainings for RIM-BP1, CtBP2/RIBEYE (marking the presynaptic ribbon), and either the presynaptic density marker bassoon (Supplementary Figure 1A) or synapsin 1/2 (Supplementary Figure 1B), marking the conventional presynaptic terminals of efferent lateral olivocochlear neurons (Safieddine and Wenthold, 1999). We found colocalizing immunofluorescence of RIM-BP1, CtBP2/RIBEYE and bassoon (Supplementary Figure 1A) as well as of RIM-BP1 with synapsin1/2 (Supplementary Figure 1B). Labeling for RIM-BP1, however, was also present in the RIM-BP1/2-deficient IHCs, questioning the antibody specificity. In an attempt to further clarify this issue, we investigated the expression of RIM-BP1 in IHCs on the mRNA level with two different approaches. We performed (i) the RNAscope mRNA detection assay using fluorescent RNA probes with consecutive immunohistochemistry (Supplementary Figure 2; Salehi et al., 2018) and (ii) nested RT-PCR from 2-week-old C57Bl/6 mouse organs of Corti and individual IHCs (Supplementary Figure 3). In the nested RT-PCR, both RIM-BP1 and -2 were detected in brain tissue and the organ of Corti (Supplementary Figures  3A,B). However, while RIM-BP2 mRNA was detected in individual IHCs (Supplementary Figure 3B), RIM-BP1 was not (Supplementary Figure 3A). This finding is consistent with our results from the RNAscope assay using fluorescent RNA probes targeting RIM-BP1 (Supplementary Figure 2D) and
In order to address the relevance of RIM-BP1 for IHC exocytosis, we measured the exocytic membrane capacitance changes ( C m ) in response to voltage-gated Ca 2+ -influx triggered by step depolarizations to -14 mV of varying length (Figure 2A). It is thought that short IHC depolarization durations up to 20 ms ( Figure 2C) primarily trigger exocytic release of the RRP (Moser and Beutner, 2000), while longer IHC depolarization durations (>20 ms) (Figure 2B) also probe the sustained phase of SV release that involves SV replenishment to the RRP and subsequent SV fusion (Moser and Beutner, 2000;Schnee et al., 2005;Goutman and Glowatzki, 2007;Meyer et al., 2009;Neef et al., 2009). Sustained exocytosis was not significantly different in RIM-BP1/2 −/− IHCs when compared to control RIM-BP2 +/+ IHCs or RIM-BP2-deficient IHCs (Figure 2B, p-values are summarized in Table 3). However, RRP exocytosis, probed by short depolarizations (≤20 ms, Figure 2C) was mildly but significantly reduced in RIM-BP1/2-deficient IHCs (p 5ms = 0.003, p 10ms = 0.04; one-way ANOVA and post hoc Tukey's test, Table 3), not seen in RIM-BP2-deficient IHCs (Krinner et al., 2017). The effect was largest for 5 ms step depolarizations, where also a significant difference between the RIM-BP1/2-deficient and RIM-BP2-deficient IHCs was detected (p 5ms = 0.03; one-way ANOVA and post hoc Tukey's test) ( Figure 2C and Table 3). The reasons for not finding a significant reduction in sustained exocytosis in RIM-BP1/2 −/− IHCs that we previously found in RIM-BP2 −/− IHCs remain unclear. One might speculate that SVs not released by short depolarizations in RIM-BP1/2 −/− IHCs get recruited later, partially masking the SV replenishment deficit reported for RIM-BP2 −/− IHCs. Moreover, we note that cell-tocell variability is high for exocytosis in response to longer stimuli, which might also contribute to this discrepancy. In conclusion, our data indicate that both RIM-BPs contribute to SV exocytosis at the IHC ribbon synapse. Among RIM-BPs at the IHC ribbon synapse, RIM-BP1 seems to be required for exocytosis of the RRP, while RIM-BP2 takes a prevailing role in clustering Ca 2+channels at the IHC AZ and enabling efficient SV replenishment (Krinner et al., 2017).

Both RIM-BP1 and 2 Are Required for Normal Hearing
Finally, we investigated whether additional genetic disruption of RIM-BP1 would cause a synaptopathic hearing impairment exceeding that found in mice lacking RIM-BP2 alone (Krinner et al., 2017). For that, we recorded ABRs from RIM-BP1/2 −/− mice. ABR waves reflect the compound neural action potential firing along the auditory pathway initiated at the first auditory synapse between sensory IHCs and SGNs. Hearing thresholds ( Figure 3A) were determined as the lowest stimulus intensity that evoked a reproducible response ABR waveform. Compared to the hearing thresholds of RIM-BP2 +/+ control mice, we found a significant threshold elevation by on average 20 dB SPL (SD ± 4.0 dB SPL) for all recorded frequencies in RIM-BP1/2 −/− mice (p 4kHz = 0.0007, p 6kHz = 0.01, p 8kHz = 0.0001, p 12kHz < 0.0001, p 16kHz < 0.0001, p 24kHz = 0.002, p 32kHz = 0.003; one-way ANOVA and post hoc Tukey's multiple comparison test). Interestingly, there was a significant threshold elevation also compared to RIM-BP2 −/− mice (p 12kHz = 0.003, p 16kHz = 0.004, p 24kHz = 0.03, p 32kHz = 0.02; one-way ANOVA and post-hoc Tukey's multiple comparison test). Along the same lines, the ABR wave amplitudes ( Figure 3B and Table 4) were strongly reduced in RIM-BP1/2 −/− mice compared to RIM-BP2 +/+ during 20 Hz click stimulation (wave I: p = 0.0006, wave III: p = 0.001, wave IV: p = 0.005; one-way ANOVA and post-hoc Tukey's multiple comparison test). When we compared the ABRs of RIM-BP1/2 −/− mice to RIM-BP2 −/− mice at 20 Hz click stimulation, we found a significant ABR wave amplitude decrease in wave IV (p = 0.03; one-way ANOVA and post hoc Tukey's multiple comparison test), which likely reflects additional deficits in synaptic transmission in the auditory brainstem. In addition, the ABR wave I amplitude difference between RIM-BP1/2 −/− mice and RIM-BP2 −/− mice reached statistical significance when increasing the rate click stimulation to 100 Hz (p = 0.03; one-way ANOVA and post hoc Tukey's multiple comparison test; Table 4). Since ABR wave I reflects synchronized firing of the SGNs induced by synaptic transmission from sensory IHCs, these data suggest a role of RIM-BP1 either in the sensory IHCs themselves or in the SGNs, downstream of the IHC ribbon synapse. To exclude that the hearing threshold elevation of RIM-BP1/2deficient mice is caused by impaired outer hair cell-mediated cochlear amplification, we also recorded DPOAEs (Figure 3C). We found normal DPOAE amplitudes in RIM-BP1/2 −/− mice (for all F2 intensities: p > 0.05; one-way ANOVA and post hoc Tukey's multiple comparison test), indicating preserved outer hair cell function, which is keeping with our previous findings in RIM-BP2-deficient mice (Krinner et al., 2017).

DISCUSSION
Here we probed for a role of RIM-BP1 in synaptic sound encoding at the IHC-SGN synapse. Building on our prior analysis of mice lacking RIM-BP2 (Krinner et al., 2017), we now studied the effects of additional deletion of RIM-BP1. RIM-BP1/2 −/− mice showed an aggravation of the auditory synaptopathy (Moser and Starr, 2016) phenotype, with further elevation of the hearing threshold and stronger reduction of ABR wave amplitudes than found in RIM-BP2 −/− mice, despite normal cochlear amplification. Our recent work suggests that, indeed, RIM-BP1 and 2 both play a role in mammalian hearing with a leading role of RIM-BP2 and partial functional redundancy of RIM-BP1 and 2. Scrutinizing the requirement for the presence of both RIM-BPs indicated a role for high frequency synaptic transmission in vivo, where the effect of additional RIM-BP1 deletion on ABR thresholds and amplitudes was enhanced at higher stimulation frequencies.
Mechanistically, such a hearing impairment could be caused by either a deficit in presynaptic transmitter release from sensory IHCs or a functional defect of the postsynaptic SGNs. On the presynaptic side, a reduction of IHC transmitter release could be caused indirectly via impaired Ca 2+ -influx and/or directly via immediate effects on SV release. We found that at IHC ribbon synapses, additional deletion of RIM-BP1 did not further alter voltage-gated Ca 2+ -influx beyond the reduction found in RIM-BP2 deficient IHCs. In contrast to our data on IHCs, Compared to RIM-BP2-deficient IHCs, a difference in exocytic C m of RIM-BP1/2-deficient IHCs was detected for 5 ms IHC depolarization duration (indicated by asterisk and bracket, p = 0.03). (B,C) Data information: Data represent IHC grand averages, mean ± SEM; one-way ANOVA and for multiple comparisons, post-hoc Tukey's test, p-values are summarized in Table 3. Significance levels: n.s. p ≥ 0.05, *p < 0.05, **p < 0.01; n = number of IHCs; age of mice: p14-p16. Data of RIM-BP2-deficient (RIM-BP2 −/− , n = 10, green) and control (RIM-BP2 +/+ , n = 12, black) IHCs were adapted from Krinner et al. (2017).
combined genetic deletion of RIM-BP1 and -2 in rod bipolar cell AZs, reduced the number of Ca 2+ -channels more strongly than the deletion of the individual RIM-BPs (Luo et al., 2017).
The same study also showed a significant compensatory elevation of RIM protein levels in RIM-BP1/2-deficient retinae (Luo et al., 2017). Other studies pointed toward redundant functions of RIMs and RIM-BPs in promoting the synaptic abundance of Ca 2+ -channels: The combined deletion of RIMs and RIM-BPs reduced Ca 2+ -channel abundance (Kushibiki et al., 2019) or Ca 2+ -triggered release (Acuna et al., 2016) more drastically than would have been expected from the effects observed upon the deletion of individual genes. Since the impact of RIMs (Jung et al., 2015) and RIM-BP2 (Krinner et al., 2017) on synaptic Ca V 1.3 Ca 2+ -channel abundance in IHCs has been shown, one could argue, that either RIM-BP1 has only a minor role on synaptic Ca 2+ -channel abundance or that its effect was masked by an increased abundance of other AZ proteins like RIMs due to compensatory upregulation upon the combined loss of both RIM-BP1 and 2. Such compensatory upregulation of other multidomain proteins of the AZ should be tested in future studies that might also test the possibility of an upregulated expression of RIM-BP1 in RIM-BP2-deficient IHCs.
Then, the aggravation of the auditory synaptopathy observed in RIM-BP1/2 −/− mice might point to a direct involvement of RIM-BP1 in SV exocytosis. Indeed, upon RIM-BP1/2 deletion, we found a subtle but significant reduction of synchronous SV exocytosis triggered by brief step-depolarizations not found in RIM-BP2 −/− mice ( Figure 2B) likely resulting in less synchronized SGN activation and, hence, smaller ABR wave I amplitude in RIM-BP1/2 −/− mice ( Figure 3B). A correspondence of impaired RRP exocytosis and reduced wave amplitude has previously been reported for bassoondeficient IHCs Buran et al., 2010). However even though exocytosis is nearly completely abolished during in vitro capacitance measurements with 5 ms step depolarization, the ABR wave I recorded in vivo did not show such a dramatic amplitude reduction. We consider several possible and non-exclusive explanations for the discrepancy of the in vivo and in vitro data. (i) There is greater experimental variability at the level of in vitro patch-clamp recordings from individual IHCs containing just a dozen of synapses with stimulation repeated only 2-3 times per IHC, while ABR in response to 80 dB click reflect sound encoding of thousands of IHC-SGN synapse averaged over 1,000 trials. (ii) The age differs between the recordings of the in vitro patch-clamp C m measurements (2-3 week-old mice) and the in vivo ABR measurements (8-10 week-old mice) which could influence synaptic transmission due to synapse maturation beyond 2-3 weeks. Morphological studies pointed out that cochlear thresholds and the subdivisions of SGNs according to their spontaneous firing rate are mature only after 28 days, which is especially critical for in vivo studies (Liberman and Liberman, 2016). (iii) The in vitro patch-clamp C m measurements were done at room temperature, whereas ABR recordings were done at 37 • C (mammalian body temperature). In vitro experiments done in frog and mammalian auditory hair cells showed, that at higher temperatures, the Ca 2+ -current activation kinetics were accelerated and amplitudes increased, which reduces synaptic delay of glutamate release and supports synchronization of Summary of IHC grand average data from perforated patch-clamp recordings of Ca 2+ -currents (I Ca ) from RIM-BP1/2 −/− , RIM-BP2 +/+ and RIM-BP2 −/− IHCs (Figure 1). Whole-cell I Ca was analyzed regarding amplitude, voltage of half-maximal activation (V half ) and voltage-dependence of Ca 2+ -channel gating (slope factor k representing voltage-sensitivity of Ca 2+ -influx). Data represent IHC grand averages, mean ± SEM; n = number of IHCs; p-values and statistical test are depicted for each dataset, significance level: n.s. p ≥ 0.05, *p < 0.05. Data of RIM-BP2-deficient (RIM-BP2 −/− ) and control (RIM-BP2 +/+ ) IHCs were adapted from Krinner et al. (2017).
SV release (Nouvian, 2007;Chen and von Gersdorff, 2019). In addition, the authors found a temperature-dependent increase in synchronous exocytosis, which was not only due to the accelerated Ca 2+ -current, but likely due to a temperaturedependent increase in efficiency of Ca 2+ influx triggered SV exocytosis (Nouvian, 2007). These differences could explain the discrepancy between the strongly reduced response to 5 ms depolarization and the lesser reduction of ABR wave I amplitude. Thus, it would be interesting to perform the patch-clamp recordings at physiological temperature. (iv) The peak preceding wave I in Figure 3B is likely the summating potential (SP), thought to reflect the synchronous depolarization of IHCs (Whitfield and Ross, 1965;Dallos et al., 1972). That the SP is unaffected by loss of RIM-BP1 and/or RIM-BP2 indicated that the mechanotransduction machinery is unaffected, supporting a conclusion that the reduction in wave I amplitude is not due to a deficit upstream of the ribbon synapse, but rather reflects reduced activation of SGNs or desynchronization of SGN activation. Indeed, the link of RRP exocytosis in IHCs to synchronous SGN activation and ABR wave I amplitude has been established Wittig and Parsons, 2008;Buran et al., 2010;Rutherford et al., 2012;Li et al., 2014).  As for other synapses, RIM-BPs might be involved in a tight Ca 2+ -channel-SV coupling Acuna et al., 2015;Grauel et al., 2016;Luo et al., 2017). While tight Ca 2+channel-SV coupling seems preserved in RIM-BP2 −/− IHCs at least after RRP recovery from depletion (Krinner et al., 2017), deletion of both RIM-BP1 and 2 might impair this coupling and, consequently, synchronous SV release ( Figure 2C). Alternatively, subpools of the RRP might exist (Goutman and Glowatzki, 2007) and be differently dependent on RIM-BP1 and 2. Future experiments, e.g., using paired recordings (Goutman and Glowatzki, 2007) or dual color imaging of presynaptic Ca 2+ signals and glutamate release (Özçete and Moser, 2020) should address these possibilities for individual IHC synapses. The better preserved sustained exocytosis in RIM-BP1/2 −/− doubleknockout IHCs compared to RIM-BP2 −/− single knockout IHCs might then simply reflect a partial masking of impaired SV replenishment (Krinner et al., 2017) by the protracted release of the RRP in the absence of both RIM-BPs.
RRP in IHCs due to the loss of RIM-BP1. Such transsynaptic coupling of IHCs and SGNs was described for the AMPA receptor -PSD-95 complex, which is required for the correct spatial alignment of Ca V 1.3 Ca 2+ -channels across the synaptic cleft (Fell et al., 2016).

AUTHOR'S NOTE
RIM-binding proteins are multidomain proteins of the presynaptic active zone that interact with Ca 2+ channels and other proteins forming release sites for synaptic vesicles.
In the present study, we probed for a function of RIM-BP1, another RIM-BP family member. We found that disruption of RIM-BPs 1 and 2 in mice causes a synaptopathic hearing impairment exceeding that of mice lacking only RIM-BP2. While the reduction of Ca 2+ -influx seemed comparable between IHCs of both mutant genotypes, deletion of both RIM-BPs caused an additional impairment of synchronous exocytosis of the readily releasable pool of synaptic vesicles. Hence, while RIM-BP2 appears to be the dominant isoform, both RIM-BPs are required for normal sound encoding at the inner hair cell ribbon synapse.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by University of Göttingen Board for Animal Welfare and Animal Welfare Office of the State of Lower Saxony. Written informed consent was obtained from the owners for the participation of their animals in this study.

AUTHOR CONTRIBUTIONS
SK and TM designed the study. SK performed immunohistochemistry, RNAscope and confocal microscopy, and patch-clamp capacitance measurements. FP performed RT-PCR experiments. DB and CV contributed to the RNAscope experiments. TM and SK prepared the manuscript with contributions of all authors.

FUNDING
This work was supported by grants of the German Research Foundation through the collaborative research centers 889 (project A02 to TM) and 1286 (project B05 to TM) and the Multiscale Bioimaging Cluster of Excellence (MBExC to TM). CV is a Creutzfeldt Fellow of the Elisabeth and Helmut Uhl Foundation. In addition, this research is supported by Fondation Pour l'Audition (FPA RD-2020-10).