The synaptic ribbon of hair cells is a large presynaptic organelle implicated in precise synaptic transmission of auditory information to afferent neurons. Ribbon synapses are electron-dense bodies that vary in shape, size, and number across species, along the tonotopic axis, and even among active zones in individual hair cells. They are unique in their form and function such that they provide the entire acoustic signal to each afferent neuron, and do so tonically by modulating the rate of neurotransmitter release as stimulus intensity changes (von Gersdorff, 2001). Because release is graded, hair cells can vary synaptic output continuously to encode sustained, ever-present sounds like background noise without the need for action potentials. Ribbon synapses, which are also present in retinal photoreceptors and bipolar cells, are therefore responsible for the trafficking of large quantities of vesicles to and from the presynaptic terminal.
There is one synaptic ribbon per active zone anchored adjacent to clusters of voltage-gated calcium channels, and depending on the species and type of hair cell, a ribbon synapse can tether anywhere from tens to hundreds of synaptic vesicles for quick release: 20 to 30 in peripheral vestibular type II hair cells, 150 to 300 in central vestibular type II hair cells, 100 to 200 in mouse cochlear IHCs, and 400 in frog saccular hair cells. The readily releasable pool consequently corresponds to the total number of tethered vesicles across all of a hair cell’s synaptic ribbons, which can number from fewer than ten to as many as a hundred (von Gersdorff et al., 1996). The tethered vesicles only account for a small portion of the total vesicle count in a hair cell; many are docked on the presynaptic membrane away from the active zone, and many more are undocked and free-floating in the presynaptic terminal, but their cycling through the hair cell will undoubtedly be mediated by the ribbon synapses. The study of hair cell transmitter release has utilized microelectrodes, patch-clamp recordings from afferent neurons, patch-clamp measurements of presynaptic exocytic membrane capacitance changes, and fluorescence microscopy of presynaptic membrane turnover to visualize and quantify that release during exocytosis. These methods have provided insights into the mechanisms of synaptic coding of auditory and vestibular stimuli, such as stimulus secretion coupling, synaptic vesicle pool dynamics, and the kinetics of transmitter release.
The morphology of a ribbon synapse varies depending on the sensory information it encodes. In vestibular hair cells, which encode information about rotational movement and linear acceleration necessary for balance and spatial orientation, type I hair cells generally have small, spherical ribbons, regardless of their location within the sensory epithelium (Lysakowski and Goldberg, 1997); type II hair cells show a large variation in shapes and sizes of synaptic ribbons, particularly in hair cells located in the central part of the sensory epithelium, where sensitivity is highest: they can be large and hollow, spheroid, or elongated and plate-like, and can occur singly or in clusters. Like type 1 hair cell ribbon synapses, ribbons in peripheral type II hair cells tend to be small and spheroid. Of cochlear hair cells, which code auditory information, inner hair cells (IHCs) have synaptic ribbons of varying shapes, while outer hair cells (OHCs) have ribbons reminiscent of elongated rods. IHCs are the actual sensory cells that code sound information, whereas the OHCs mechanically amplify sound-induced cochlear vibrations.
The molecular anatomy of hair cell synaptic ribbons that depicts this dichotomy has been hard to identify in the past decade because of a lack of available material with which to study. Whereas there are only a few thousand coding hair cells in a mouse, there are millions of photoreceptors. Because retinal ribbon synapses have molecular similarities to hair cell ribbon synapses, as well as to conventional synapses, the molecular composition of retinal ribbon synapses has guided much of our current understanding of the molecular composition of hair cell ribbons (tom Dieck et al., 2005). Hair cell ribbon synapses are known to contain the C-terminal binding protein RIBEYE, a major structural protein hypothesized to also contribute to enzymatic activity; Rab3 and RIM, which are involved in tethering vesicles to the ribbon; Bassoon and Piccolo, which anchor the ribbon to the active zone; and the SNAREs syntaxin 1, SNAP25, and VAMP1/synaptobrevin 1, which form the ribbons’ key exocytic machinery.
Like other ribbon synapses, hair cell synaptic ribbons lack synapsins, proteins that mobilize synaptic vesicles. Unlike photoreceptors, they also lack synaptotagmins 1 and 2, the calcium sensors of vesicle fusion; synaptophysin 1 and 2, which complex with synaptobrevin to prime the synaptic vesicles; and Munc13-1 (Uthaiahand Hudspeth, 2010), which partners with RIM to change syntaxin from its closed conformation to its open conformation in preparation of new SNARE complexes. Considered together, these modifications to the presynaptic molecular machinery suggest that hair cell synaptic ribbons function uniquely, even when compared to other ribbon synapses, in order to precisely code the temporally fine structure of acoustic stimuli: hair cells probably use synaptotagmin isoforms or other proteins altogether to sense changes in calcium concentration at the hair cell afferent synapse, and favor a mechanism for reaching the open conformation of syntaxin different from conventional synapses.
Recent research has implicated otoferlin, a C2-domain transmembrane protein defective in a recessive form of human deafness, as the mediator of late stage exocytosis, and the major Ca2+ sensor that triggers membrane fusion of synaptic vesicles at IHC ribbon synapses (Roux et al., 2006). Otoferlin localized to ribbon-associated synaptic vesicles, and its expression correlated with afferent synaptogenesis. It bound calcium and displayed calcium-dependent interactions with syntaxin1 and SNAP2, interactions that were necessary for proper neurotransmitter release, because despite normal ribbon synapse morphogenesis and Ca2+ current, if otoferlin was mutated there would be nearly no exocytosis in IHCs.
Exocytosis of synaptic vesicles in hair cells is initiated once their stereocilia are mechanically stimulated. Mechanosensitive channels like the apical potassium channels drive receptor potential generation, which further activate voltage-gated K+ and Ca2+ channels. In mammalian IHCs, the calcium channels most stimulated are CaV1.3 L-type Ca2+ channels. These voltage-gated channels are rapidly activated and deactivated, triggering synaptic exocytosis with little to no inactivation. Once released, the glutamate binds and activates AMPA receptors that are clustered in the postsynaptic density, with each IHC making synaptic contacts with 10 to 30 bipolar afferent neurons differing in spontaneous firing rates and thresholds.
The synaptic connectivity of auditory and vestibular hair cells vary, and even more precisely between ICHs and OCHs. ICHs synapse to myelinated type I spiral ganglion neurons, whereas OHCs in the apical cochlear turns form ribbon synapses with unmyelinated type II spiral ganglion neurons. Synapses of low spontaneous rate fibers, mostly located at the neural/modiolar side of an IHC, tend to have larger ribbons than those of high spontaneous rate fibers on the abneural/pillar side. The vestibular system is particularly rich in synaptic specializations featuring bouton-, dimorphic-, and calyx-type afferent endings. In the case of calyx or dimorphic afferents, each vestibular neuron receives input from multiple ribbon synapses of one hair cell; in the case of bouton or dimorphic afferents, input is received from many hair cells. Calyx-type terminals not only receive ribbon synapse input from the enclosed type I hair cells on their inner face, but can also receive this input from neighboring type II hair cells on their outer face.
Exocytosis of neurotransmitter to afferent neurons from hair cells is not strictly confined to their ribbon synapses. Experiments performed in goldfish retinal bipolar neurons showed release events occurring away from active zones (Zenisek et al., 2000). Fusion and exocytosis of vesicles distal from the ribbon structures occurred at a slower rate than at ribbon synapses, but this was likely the source of sustained release during prolonged stimulation, reflective of the lower internal Ca2+ signal that those vesicles experienced due to their distance away from calcium channels. Indeed, influx of calcium into bipolar neurons and hair cells is responsible for both exocytosis and endocytosis of neurotransmitters. By combining discrete Ca2+ uncaging with membrane capacitance measurements in mouse IHCs, one study quantized the minimum concentrations of calcium needed to spur exocytosis and endocytosis (Beutner et al., 2001). Calcium concentrations from 20 to 30 µM at release sites were required for observable exocytosis and a fast mode of endocytosis relative to that exhibited during lower calcium levels. Thus, IHC ribbon synapses appear to use two modes of endocytosis to retrieve exocytosed vesicles.
Activation of the rapid endocytic mode during periods of intense stimuli allows for sufficient vesicle recycling, which contributes to the ability of IHCs to maintain accurate signal transduction during prolonged sound stimuli. This calcium-dependent switch between rapid and slow modes of endocytosis may represent a general mechanism whereby synapses adapt the rate of membrane retrieval to the level of stimulation.