ABSTRACT
Ribbon synapses of cochlear inner hair cells (IHCs) employ efficient vesicle replenishment to indefatigably encode sound. In neurons, neuroendocrine and immune cells, vesicle replenishment depends on proteins of the mammalian uncoordinated 13 (Munc13, also known as Unc13) and Ca2+-dependent activator proteins for secretion (CAPS) families, which prime vesicles for exocytosis. Here, we tested whether Munc13 and CAPS proteins also regulate exocytosis in mouse IHCs by combining immunohistochemistry with auditory systems physiology and IHC patch-clamp recordings of exocytosis in mice lacking Munc13 and CAPS isoforms. Surprisingly, we did not detect Munc13 or CAPS proteins at IHC presynaptic active zones and found normal IHC exocytosis as well as auditory brainstem responses (ABRs) in Munc13 and CAPS deletion mutants. Instead, we show that otoferlin, a C2-domain protein that is crucial for vesicular fusion and replenishment in IHCs, clusters at the plasma membrane of the presynaptic active zone. Electron tomography of otoferlin-deficient IHC synapses revealed a reduction of short tethers holding vesicles at the active zone, which might be a structural correlate of impaired vesicle priming in otoferlin-deficient IHCs. We conclude that IHCs use an unconventional priming machinery that involves otoferlin.
INTRODUCTION
The mechanisms that establish fusion competence of synaptic vesicles are classically defined as tethering, docking and priming. In this framework, vesicles are first loosely tethered to the presynaptic active zone membrane, then closely attach to the membrane upon docking and finally undergo further maturation steps to gain full fusion competence. Recent high-resolution ultrastructural work has indicated that such clear distinctions of morphological and functional preparatory steps in vesicle fusion might have been too simple. Instead, protein tethers of different lengths and numbers have been proposed to establish vesicular fusion competence (Fernández-Busnadiego et al., 2010; Fernández-Busnadiego et al., 2013; Siksou et al., 2009). In neurons, neuroendocrine, immune and airway epithelial cells, this process employs priming factors belonging to the mammalian uncoordinated 13 (Munc13, also known as Unc13) and Ca2+-dependent activator proteins for secretion (CAPS) families (Dudenhöffer-Pfeifer et al., 2013; Imig et al., 2014; Speidel et al., 2005; Zhu et al., 2008). The Munc13 protein family includes the neuronal isoforms Munc13-1, Munc13-2, Munc13-3 (also known as Unc13a, Unc13b and Unc13c, respectively) and brain-specific angiogenesis inhibitor I-associated protein 3 (Baiap3), as well as the non-neuronal Munc13-4 isoform (Unc13d), whereas CAPS1 and CAPS2 (also known as CADPS and CADPS2, respectively) constitute the CAPS protein family (Ann et al., 1997; Augustin et al., 2001; Betz et al., 2001; Brose et al., 1995; Koch et al., 2000; Shiratsuchi et al., 1998; Speidel et al., 2003). Munc13s and CAPSs are evolutionarily conserved (i.e. UNC-13 and UNC-31 in C. elegans, and dUnc13 and dCaps in Drosophila; Aravamudan et al., 1999; Renden et al., 2001; Richmond et al., 1999), and genetic deletion causes dramatic defects, ranging from severe reduction to complete loss of the readily releasable pool of synaptic vesicles (RRP) and total arrest of spontaneous and evoked neurotransmission in several cell types (Augustin et al., 1999; Jockusch et al., 2007; Liu et al., 2010; Varoqueaux et al., 2002).
Replenishment of the RRP is likely rate-limiting for tonic neurotransmitter release at ribbon synapses. Governed by receptor potentials, each inner hair cell (IHC) active zone transmits acoustic information through graded release of up to hundreds of vesicles per second. For this challenging task, IHC synapses must employ mechanisms of vesicle replenishment that involve otoferlin, a multi-C2 domain protein that is crucial for exocytosis in cochlear IHCs and vestibular hair cells (Roux et al., 2006; Dulon et al., 2009; Pangršič et al., 2012). Otoferlin is required for hearing (Roux et al., 2006; Yasunaga et al., 1999) and thought to act as a priming factor and vesicular Ca2+-sensor for release in IHCs (Johnson and Chapman, 2010; Pangršič et al., 2010; Roux et al., 2006). However, which other proteins contribute to establishing vesicular fusion competence in IHCs remains to be determined. Here, we combined functional and morphological approaches to investigate the roles of Munc13-like priming factors in IHCs. Our data indicate that the conventional Munc13- and CAPS-dependent priming machinery of central nervous system (CNS) synapses does not operate in exocytosis at IHC ribbon synapses.
RESULTS AND DISCUSSION
Hearing is unaffected in mouse mutants lacking Munc13 or CAPS priming factors
To assess the impact of genetic disruption of Munc13 and CAPS proteins on auditory function, we recorded auditory brainstem responses (ABRs) evoked by short tone bursts and clicks in knockout (KO) mice for Munc13-1, -2, -3, -4 and Baiap3 as well as CAPS1 and CAPS2. Given that genetic deletion of Munc13-1 and CAPS1 results in perinatal lethality, we recorded ABRs from mice heterozygous for these genes. We did not observe alterations of ABR thresholds nor changes in amplitudes or latencies of the ABR wave I, reporting the compound action potential of the spiral ganglion, in any of the mutants when compared to wild-type (WT) littermates (Fig. 1; supplementary material Fig. S1). Moreover, we recorded distortion product otoacoustic emissions to evaluate outer hair cell function, but did not detect a statistically significant change for any of the mutant mouse strains suggesting intact cochlear amplification (data not shown). Therefore, disruption of Munc13 and CAPS does not seem to affect sound encoding in the cochlea. However, we note that testing the effect of complete deletion of Munc13-1 and CAPS1 in IHCs will require future experiments on conditional knockout mice, as the heterozygous state tested here might provide protein copy numbers that still support normal functionality (Augustin et al., 1999).
Hearing thresholds remain unaffected in Munc13 and CAPS deletion mutants. (A) Domain overview of Munc13-like protein isoforms highlighting the conservation of the Munc13 homology domain (MHD; adapted from Koch et al., 2000). C1, C1 domain; C2, C2 domain. Comparable ABR thresholds in (B) Munc13-1+/− and Munc13-4−/−, (C) Munc13-2−/−, (D) Munc13-3−/− and (E) CAPS1+/− CAPS2−/− mice and age-matched WT animals. ABR waveforms (I–V) in response to 80 dB click stimuli are shown for (F) CAPS1+/− CAPS2−/− and (G) Munc13-1+/− mice respectively. n values are shown on the figures.
Hearing thresholds remain unaffected in Munc13 and CAPS deletion mutants. (A) Domain overview of Munc13-like protein isoforms highlighting the conservation of the Munc13 homology domain (MHD; adapted from Koch et al., 2000). C1, C1 domain; C2, C2 domain. Comparable ABR thresholds in (B) Munc13-1+/− and Munc13-4−/−, (C) Munc13-2−/−, (D) Munc13-3−/− and (E) CAPS1+/− CAPS2−/− mice and age-matched WT animals. ABR waveforms (I–V) in response to 80 dB click stimuli are shown for (F) CAPS1+/− CAPS2−/− and (G) Munc13-1+/− mice respectively. n values are shown on the figures.
Loss of Munc13 or CAPS priming factors does not alter Ca2+ currents and exocytosis of IHCs
To clarify the contributions of the main Munc13 and CAPS isoforms to IHCs presynaptic function, we analyzed presynaptic Ca2+ currents and exocytosis in the respective deletion mutant mice. We used an organotypic culture approach to investigate the effect of genetic deletion of both CAPS1 and CAPS2 (hereafter CAPS1/2-DKO) or both Munc13-1 and Munc-13-2 (hereafter Munc13-1/2-DKO) on Ca2+-driven exocytosis in IHCs in vitro (Fig. 2). Cultured organs of Corti appear to mature analogously to the in vivo situation (Sobkowicz et al., 1982) and are suitable for patch-clamp recordings of presynaptic function (Nouvian et al., 2011; Reisinger et al., 2011). After a week in culture, the overall organ of Corti morphology was preserved and IHCs abundantly expressed otoferlin (Fig. 2A). When comparing IHC Ca2+ currents from CAPS1/2-DKO and Munc13-1/2-DKO with data from WT and otoferlin-knockout (Otof-KO) mice, we did not detect differences in voltage-dependence, amplitude (Fig. 2B, maximal amplitudes WT, 292±25 pA; Otof-KO, 291±31 pA, CAPS1/2-DKO: 306±30 pA, Munc13-1/2-DKO: 286±20 pA; mean±s.e.m., P>0.05 between all groups) or kinetics (Fig. 2C). Exocytosis was monitored as changes in membrane capacitance (ΔCm) in response to the maximal Ca2+ influx elicited by depolarization of varying durations (Fig. 2D–G). Although the ΔCm values were indistinguishable between WT, CAPS1/2-DKO and Munc13-1/2-DKO IHCs, those of Otof-KO exhibited dramatically reduced exocytosis, consistent with previous reports using acute preparations (Beurg et al., 2010; Roux et al., 2006). Given that exocytosis in IHCs acquires otoferlin-dependence around P4 in vivo (Beurg et al., 2010), our findings indicate a functional maturation of IHCs in organotypic culture and further validate this approach for studying IHCs of perinatally lethal mutant mice. Detailed analysis of ΔCm values for stimuli of 2–50 ms duration by exponential fitting revealed comparable RRP size and depletion kinetics in CAPS1/2-DKO and Munc13-1/2-DKO IHCs (Fig. 2H; WT, 26.8±4.2 fF; CAPS1/2-DKO, 20.9±5.4 fF; Munc13-1/2-DKO, 31.9±5.5 fF; P>0.05 between all groups), whereas the strongly reduced exocytosis of Otof-KO prohibited such analysis. Moreover, we tested presynaptic IHC function in mice lacking Munc13-4 and Baiap3, but did not detect any changes in Ca2+ currents or exocytosis (supplementary material Fig. S1B). We conclude that Munc13-like priming factors are dispensable for IHC presynaptic function.
Ca2+ currents, RRP size and sustained exocytosis are conserved in CAPS1/2-DKO and Munc13-1/2-DKO but exocytosis is strongly impaired in Otof-KO animals. (A) Maximum projection of a confocal z-stack taken from a WT organotypic culture after 7 days in vitro (DIV) immunostained for otoferlin (green) and F-actin (red); filled arrowheads indicate IHCs; clear arrowheads outer hair cells, OHC. (B) Whole-cell Ca2+ current–voltage relationships from cultured Munc13-like priming factor mutants and Otof-KO IHCs at DIV7–8. Voltage-dependence and Ca2+ current amplitudes are retained across all genotypes. (C) Mean±s.e.m. sample traces of a 50 ms depolarization to the maximum Ca2+ current potential for each genotype. (D) Exocytic ΔCm in response to varying step depolarizations and corresponding Ca2+ current integrals show comparable exocytosis in WT, CAPS1/2-DKO and Munc13-1/2-DKO IHCs. In Otof-KO, exocytosis is dramatically reduced. (E) Mean maximum Ca2+ current and corresponding ΔCm of the respective groups for a 50 ms depolarization. (F) ΔCm plotted as a function of Ca2+ current charge shows comparable release efficiency between Munc13-like mutants and WT IHCs. (G) Initial ΔCm were fitted with an exponential function to estimate RRP sizes as quantified in H. All genotypes, apart from Otof-KO, showed comparable RRP amplitudes. *P<0.05; ***P<0.001 for Otof-KO versus WT. n values are shown on the figures.
Ca2+ currents, RRP size and sustained exocytosis are conserved in CAPS1/2-DKO and Munc13-1/2-DKO but exocytosis is strongly impaired in Otof-KO animals. (A) Maximum projection of a confocal z-stack taken from a WT organotypic culture after 7 days in vitro (DIV) immunostained for otoferlin (green) and F-actin (red); filled arrowheads indicate IHCs; clear arrowheads outer hair cells, OHC. (B) Whole-cell Ca2+ current–voltage relationships from cultured Munc13-like priming factor mutants and Otof-KO IHCs at DIV7–8. Voltage-dependence and Ca2+ current amplitudes are retained across all genotypes. (C) Mean±s.e.m. sample traces of a 50 ms depolarization to the maximum Ca2+ current potential for each genotype. (D) Exocytic ΔCm in response to varying step depolarizations and corresponding Ca2+ current integrals show comparable exocytosis in WT, CAPS1/2-DKO and Munc13-1/2-DKO IHCs. In Otof-KO, exocytosis is dramatically reduced. (E) Mean maximum Ca2+ current and corresponding ΔCm of the respective groups for a 50 ms depolarization. (F) ΔCm plotted as a function of Ca2+ current charge shows comparable release efficiency between Munc13-like mutants and WT IHCs. (G) Initial ΔCm were fitted with an exponential function to estimate RRP sizes as quantified in H. All genotypes, apart from Otof-KO, showed comparable RRP amplitudes. *P<0.05; ***P<0.001 for Otof-KO versus WT. n values are shown on the figures.
Cochlear IHCs apparently lack Munc13-like priming factors
Previous reports have indicated that Munc13-1 protein is absent from chicken cochlear extracts (Uthaiah and Hudspeth, 2010), however, the expression patterns of the remaining Munc13 and CAPS isoforms in the cochlea remain to be established. Therefore, we performed immunostainings with extensively tested antibodies for all Munc13 and CAPS isoforms (Cooper et al., 2012) on acutely dissected organs of Corti from hearing P16–P17 WT mice (Fig. 3). Consistent with our electrophysiological data arguing against a functional role of these proteins in presynaptic IHC function, we did not detect specific Munc13 or CAPS immunofluorescence within IHCs. Rather, immunoreactivities of all tested proteins, including Munc13-1 but not Munc13-2 were restricted to presynaptic terminals of efferent olivocochlear neurons, as evident from colocalization with the neuronal synaptic vesicle marker synapsin (isoforms 1 and 2), which served as internal positive controls in these experiments. We did not find a specific immunolabeling for Munc13-4 in organs of Corti (i.e. a labeling that was absent from knockout tissue) with currently available antibodies. Although we cannot exclude that there is Munc13-4 expression in IHCs, we conclude, based on our analysis of presynaptic function (supplementary material Fig. S1B), that this isoform plays a minor – if any – role in vesicular release from IHC active zones. In the present study, we could establish that IHCs seem to operate without Munc13s or CAPS proteins, a finding that is in line with the notion that their interacting partners neuronal soluble N-ethylmaleimide-sensitive-factor attachment receptors (SNAREs) appear to be absent from IHCs (Nouvian et al., 2011). Based on our findings, we propose that the priming machinery of IHCs is molecularly distinct from conventional neuronal active zones and likely involves the synaptic ribbon and/or bassoon (Frank et al., 2010; Snellman et al., 2011) and otoferlin (Pangršič et al., 2010).
IHCs apparently lack Munc13-like priming factors. Single confocal planes of P16–P17 WT organs of Corti immunostained for (A) Munc13-1 (B) Munc13-3, (C) CAPS1 and (D) CAPS2. Calbindin (calb.) was used as IHC marker and synapsin (isoforms 1 and 2) to counterstain olivocochlear efferent presynaptic terminals. Individual channels are presented alongside for clarity. The inset in D shows a schematic illustration of a typical IHC (blue) with afferent (white) and efferent (red) neuronal innervation showing a single synaptic complex for simplicity.
IHCs apparently lack Munc13-like priming factors. Single confocal planes of P16–P17 WT organs of Corti immunostained for (A) Munc13-1 (B) Munc13-3, (C) CAPS1 and (D) CAPS2. Calbindin (calb.) was used as IHC marker and synapsin (isoforms 1 and 2) to counterstain olivocochlear efferent presynaptic terminals. Individual channels are presented alongside for clarity. The inset in D shows a schematic illustration of a typical IHC (blue) with afferent (white) and efferent (red) neuronal innervation showing a single synaptic complex for simplicity.
Otoferlin is enriched in IHC active zone membranes and regulates vesicular tether formation
Given that IHC exocytosis seems to operate without the classical priming factors, we aimed to further investigate the mechanism by which otoferlin facilitates vesicle replenishment. Our confocal data revealed an enrichment of otoferlin at the active zone membrane (Fig. 4A), a localization analogous to Munc13 and CAPS priming factors at neuronal active zones. In these experiments, we employed an antibody that recognizes an intraluminal C-terminal epitope of otoferlin and observed otoferlin clustering in the active zone membrane at the base of the ribbon that could also be found with immunogold labelings (data not shown). Otoferlin enrichment at the release site might indicate an involvement in vesicle tethering, which has been implicated as an ultrastructural correlate of vesicular fusion competence (Fernández-Busnadiego et al., 2010; Frank et al., 2010). Therefore, we used electron tomography of IHC synapses to investigate vesicular tethers at active zones of WT and otoferlin-deficient IHCs, instantaneously frozen following stimulation with high K+ to analyze ongoing synaptic activity (Fig. 4B–G). We focused our analysis on filamentous tethers connecting membrane-proximal vesicles to the presynaptic density and active zone membrane, as this population has been proposed to represent readily-releasable vesicles (Fernández-Busnadiego et al., 2010; Frank et al., 2010). Interestingly, we failed to detect tethers <5 nm, thought to comprise assembled SNARE complexes (Fernández-Busnadiego et al., 2010), consistent with a previous report suggesting that IHC ribbons operate without neuronal SNAREs (Nouvian et al., 2011). As previously described (Roux et al., 2006), there was no difference in the number of membrane-proximal and overall ribbon-associated vesicles (data not shown). However, we detected a statistically significant reduction, of roughly two-thirds, in the fraction of tethers shorter than 10 nm, resulting in significantly increased average tether lengths at Otof-KO active zones (Fig. 4E–H). It is tempting to speculate that the loss of short tethers is directly related to the exocytosis deficit, analogous to the reduction of short tethers in synaptosomes following treatment with clostridial neurotoxins or hypertonic sucrose (Fernández-Busnadiego et al., 2010). Whether otoferlin is a tether constituent and, if so, homophilic interactions of otoferlin molecules localized to vesicular and plasma membranes are involved, remains to be investigated in future studies. Additionally, it will be crucial to identify and characterize interaction partners of otoferlin that might contribute to establishing vesicular fusion competence.
Otoferlin is enriched in IHC active zone membranes and plays a role in vesicular tethering. (A) Immunostaining for otoferlin and the synaptic ribbon marker CtBP2, showing apparent apposition of the two proteins at the active zone (arrowheads). Asterisks indicate IHC nuclei. The boxed area is presented as a confocal z-stack of the indicated ribbon, illustrating the clustering of otoferlin in a plane that is segregated from those containing the CtBP2-labeled ribbon. (B) Representative tomographic sections of (Bi) WT and (Bii) Otof-KO ribbon synapses after stimulation with high K+ to evoke membrane turnover with corresponding rendered models. Blue, active zone membrane; magenta, presynaptic density; yellow, membrane-associated synaptic vesicles (SVs). (C) Representative images of vesicles connected to the plasma membrane (PM) through a filamentous tether (clear arrowhead). Despite the comparable numbers of tethered synaptic vesicles in both genotypes (D), the reduction of short tethers leads to a highly significant shift towards longer tethers in Otof-KO samples (E), and an increased mean tether length (F). (G) Normalized histograms (5 nm binned) and (H) cumulative probability density plots of tether lengths raw data. *P<0.05; ***P<0.001 for Otof-KO versus WT. n values are shown on the figures.
Otoferlin is enriched in IHC active zone membranes and plays a role in vesicular tethering. (A) Immunostaining for otoferlin and the synaptic ribbon marker CtBP2, showing apparent apposition of the two proteins at the active zone (arrowheads). Asterisks indicate IHC nuclei. The boxed area is presented as a confocal z-stack of the indicated ribbon, illustrating the clustering of otoferlin in a plane that is segregated from those containing the CtBP2-labeled ribbon. (B) Representative tomographic sections of (Bi) WT and (Bii) Otof-KO ribbon synapses after stimulation with high K+ to evoke membrane turnover with corresponding rendered models. Blue, active zone membrane; magenta, presynaptic density; yellow, membrane-associated synaptic vesicles (SVs). (C) Representative images of vesicles connected to the plasma membrane (PM) through a filamentous tether (clear arrowhead). Despite the comparable numbers of tethered synaptic vesicles in both genotypes (D), the reduction of short tethers leads to a highly significant shift towards longer tethers in Otof-KO samples (E), and an increased mean tether length (F). (G) Normalized histograms (5 nm binned) and (H) cumulative probability density plots of tether lengths raw data. *P<0.05; ***P<0.001 for Otof-KO versus WT. n values are shown on the figures.
In summary, our study provides evidence for an unconventional molecular priming machinery at cochlear IHC ribbon synapses. Using deletion mutants for Munc13 and CAPS priming factors, we show for the first time that (1) hearing is intact in the absence of various Munc13-like priming factors, (2) IHC presynaptic exocytosis is normal in Munc13-1/2-DKO and CAPS1/2-DKO mice whose release at glutamatergic CNS synapses is strongly impaired or completely abolished, and (3) Munc13 and CAPS proteins appear to be absent from IHC presynaptic active zones. Instead (4), otoferlin appears to be required to form short tethers (<10 nm) between vesicles and active zone membranes, providing a candidate mechanism for the highly efficient replenishment of synaptic vesicles in IHCs.
MATERIALS AND METHODS
Animals, organotypic culture and ABRs
All animal experiments conformed to national animal care guidelines and were approved by the animal welfare office of Lower Saxony. Wild-type C57Bl/6 (WT), CAPS1/2-DKO (Jockusch et al., 2007; Speidel et al., 2005), Munc13-1/2-DKO (Augustin et al., 1999; Varoqueaux et al., 2002) and otoferlin-KO (Otof-KO; Reisinger et al., 2011) mice of either sex were used to prepare organotypic cultures of the organ of Corti as described previously (Nouvian et al., 2011; Reisinger et al., 2011). Briefly, organs of Corti were dissected from embryonic day 18–19 mutant or WT mice of either sex in HEPES-buffered HBSS supplemented with 250 ng/ml fungizone and 10 µg/ml Penicillin G (Sigma-Aldrich), then placed on CellTakTM-coated coverslips (BD Biosciences) and incubated in DMEM/F12 with 5% FBS for 7–8 days prior to electrophysiological characterization.
Munc13-1+/− (Augustin et al., 1999), Munc13-2-KO (Varoqueaux et al., 2002), Munc13-3-KO (Augustin et al., 2001) and CAPS1+/−-CAPS2-KO (Jockusch et al., 2007) as well as Munc13-4-KO (Jinx mice; Jackson Laboratories) and Baiap3-KO (Wojcik et al., 2013) were used alongside littermates or age-matched (3–10 weeks old) C57Bl/6 WT mice for testing ABR thresholds as described previously (Neef et al., 2009).
Patch-clamp of IHCs
Perforated-patch recordings were performed on IHCs from either acutely dissected P14 mice (Munc13-4-KO; Baiap3-KO) or IHCs from cultured apical organs of Corti (Munc13-1/2-DKO; CAPS1/2-DKO; Otof-KO) using an extracellular solution composed of (in mM): NaCl (103), KCl (2.8), MgCl2 (1), HEPES (10), TEA-Cl (35), D-glucose (11.2) and CaCl2 (2 for acute preparations, 10 for cultures); ∼300 mOsm/l, pH 7.2 and a Cs-based intracellular solution containing (in mM): Cs-gluconate (130), TEA-Cl (10), 4-amino-pyridine (10), MgCl2 (0.05) and HEPES (10); ∼290 mOsm/l, pH 7.1 freshly supplemented with amphotericin B (250 µg/ml). We employed an elevated [Ca2+] in order to enhance Ca2+ currents and RRP release kinetics. All experiments were done at 22–25°C using an EPC10 amplifier with PatchMaster software (HEKA electronics). Capacitance measurements were performed using the Lindau-Neher technique (Lindau and Neher, 1988; Moser and Beutner, 2000). Currents were leak-subtracted with a p/10 protocol (Armstrong and Bezanilla, 1974); IHCs with leak currents exceeding −50 pA at the holding potential of −87 mV were discarded from analysis. Liquid-junction potentials were calculated with the Igor Pro LJP calculator and corrected offline.
Immunohistochemistry and confocal microscopy
Immunohistochemistry was performed on acutely dissected apical organs of Corti whole-mount preparations as described previously (Neef et al., 2009). Specimens were imaged using Leica SP2 or SP5 laser-scanning confocal microscopes with a 1.4 NA 63× or 100× oil-immersion objective, respectively. Images were thresholded for background subtraction and deconvolved using ImageJ (Schneider et al., 2012).
Antibodies
The following primary antibodies (supplementary material Fig. S1A) were used for immunolabellings: polyclonal rabbit anti-Munc13-1, anti-bMunc13-2 (brain-specific Munc13-2 isoform), anti-ubMunc13-2 (ubiquitous Munc13-2 isoform), anti-Munc13-3 antibodies (Cooper et al., 2012; Varoqueaux et al., 2005), anti-CAPS1, anti-CAPS2 (Synaptic Systems), guinea pig antibody against both synapsin 1 and 2 (Synaptic Systems), mouse anti-calbindin D28k (Swant) and mouse or rabbit anti-otoferlin antibodies (Abcam/Synaptic Systems), detected by species-specific secondary anti-IgG antibodies conjugated to AlexaFluor® 488, 568 and 647 (Life technologies). Phalloidin–AlexaFluor®568 (Life technologies) was used to label actin-containing stereocilia.
Electron microscopy and tomographic reconstruction
For high-pressure freezing, explanted apical-turn organs of Corti from three WT and two Otof-KO mice (all P14) were placed in 200-µm aluminium type A specimen carriers filled with stimulation solution (Pangršič et al., 2010). Type B lids were dipped in hexadecene (Sigma-Aldrich) and the specimens were frozen with a HPM100 (Leica). Freeze substitution was performed in an AFS2 (Leica) as described previously (Rostaing et al., 2006). For electron-tomography, 250-nm sections (Ultracut E ultramicrotome; Leica) were collected on formvar-coated 100 copper mesh grids and post-stained with 4% uranyl-acetate and Reynold's lead-citrate. Subsequently, 10-nm gold particles were applied to the grid. Single or double tilt series were acquired at a JEOL JEM 2100 transmission electron microscope at 200 kV at tilt angles ranging from −62–−55° to +55–+62° with 1° increments using Serial-EM software. Tomograms were generated using the IMOD package etomo and rendered using 3dmod (http://bio3d.colorado.edu/imod/). Tether lengths were analyzed in 3D from tomograms using ImageJ by determination of starting (x1y1z1) and ending (x2y2z2) coordinates in virtual sections and calculating the lengths with the following formula: √[(x2−x1)2+(y2−y1)2+(z2−z1)2]. Synaptic vesicles connected to the presynaptic density and active zone membrane were considered in the analysis.
Statistical analysis
Data are presented as mean±s.e.m. One-way ANOVA with post-hoc Tukey were performed on normally distributed data of multiple groups (assessed by Kolmogorov–Smirnov test); P≤0.05 was accepted as statistically significant. For comparison of fractional data (i.e. tether length), a χ2-test was used.
Acknowledgements
We thank Jens Rettig for providing Munc13-4-KO mice. We thank Christian Rüdiger, Stefan Thom, Sandra Gerke and Christiane Senger-Freitag for expert technical assistance. Moreover, we would like to express our gratitude to Nicola Strenzke and Tina Pangršič-Vilfan for supporting the study and critically reading the manuscript.
Author contributions
C.V., C.W. and T.M. designed the study. C.V. performed electrophysiology. C.V. and B.H.C. performed immunohistochemistry, C.W. performed EM tomography. E.R., S.M.W., K.R. and N.B. performed mouse mutagenesis and provided discussion and input into the manuscript preparation. C.V., J.N. and C.W. analyzed the data. C.V., C.W. and T.M. prepared the manuscript.
Funding
This work was supported by an intramural grant of the University Medical Center Goettingen to C.V. and grants of the German Research Foundation (DFG) through the Collaborative Research Center SFB 889 [projects A2 to T.M, A4 to E.R., A7 to C.W. and B1 to J.R. and N.B.].
References
Competing interests
The authors declare no competing or financial interests.