Abstract
Zebrafish are popular models for biological discovery. For investigators of the auditory and vestibular periphery, manipulations of hair cell and synaptic mechanisms have relied on inferences from extracellular recordings of physiological activity. We now provide data showing that hair cells and supporting cells of the lateral line can be directly patch-clamped, providing the first recordings of ionic channel activity, synaptic vesicle release, and gap junctional coupling in the neuromasts of living fish. Such capabilities will allow more detailed understanding of mechano-sensation of the zebrafish.
Introduction
Zebrafish hold promise for uncovering physiological mechanisms that are difficult to realize in other model species, because of the benefits of rapid reproductive rate and reduced organismal complexity (Rinkwitz et al., 2011). Indeed, progress has been enormous with this species, even in the auditory/vestibular arena, a field that usually prides itself on understanding human mechanisms with mammalian models. For example, great strides have been made in hair cell mechano-transduction and synaptic mechanisms (Nicolson, 2005). However, an obstacle to truly capitalizing on zebrafish for hair cell mechanism discovery is the heretofore inability to obtain patch-clamp recordings from neuromast sensory hair cells. We now report that this impediment has been overcome. We show preliminary data of whole-cell recordings from cells within lateral line neuromasts of living zebrafish, demonstrating ionic channel activity and synaptic vesicle release in hair cells. Additionally, we identify voltage-dependent gap junctional coupling in neuromast supporting cells. Our approach should help investigators obtain cellular data crucial to understanding the powerful genetic manipulations already available in zebrafish.
Materials and Methods
Zebrafish of either sex ranging in age from 3 to 14 days postfertilization (dpf) were anesthetized in Tricane and mounted in a recording chamber using dental floss tie downs (see Ricci and Fettiplace, 1997). Viability was monitored by visually monitoring heart rate and blood flow. An upright Nikon Eclipse was used for viewing, and recordings were made with an Axon 200B amplifier with an Axon DD1322 digitizer. Images were enhanced with a Hamamatsu CCD camera. All recordings and image capture were made with jClamp software (Scisoft). Cells were held at −80 mV. Extracellular solution was as follows (in mm): 125 NaCl, 1.0 KCl, 2.2 MgCl2, 2.8 CaCl2, 10 HEPES, 6 d-glucose, 285 mOsm, pH 7.6. Pipette solution was (in mm): 90 CsCl, 20 TEA, 5 Na2ATP, 3.5 MgCl2, 10 HEPES, 1 EGTA, 260 mOsm, pH 7.2. Intracellular KCl (110 mm) solution lacked TEA. P/-5 protocols were made at a subtraction holding potential of −80 mV. Pipette resistance was typically 6.5 MΩ with Cs pipette solutions. Pipettes used to clean a pathway toward hair cells were ∼1–2 MΩ. For patch pipettes, 1.5-mm-thick-walled borosilicate glass was used without any coating. Capacitance measures were made with a dual sine admittance technique (Santos-Sacchi, 2004; Schnee et al., 2011a,b). Recordings were made at room temperature. Data are reported as mean ± SE.
Results
Zebrafish lateral line neuromasts are peripherally located on each side of the zebrafish. Before scale formation, the organ is accessible via micropipette. Figure 1 shows a series of Hoffman optics images at different levels through the neuromast, starting apically where the hair cell kinocilia insert into the gelatinous cupola (A,B). Each kinocilium arises from one hair cell, and can be used to count the number of hair cells in the neuromast. In this example, 17 hair cells are present. Further medial, the outlines of the supporting cell boundaries (C,D; arrows) are visible. In the center of the neuromast, the tear drop-shaped hair cells are visible (E; arrows). To prepare for patch clamping, the supporting cells at the neuromast periphery must be breeched by suction from a large-tipped pipette, whereupon entry into the neuromast is confirmed by positive pressure expanding the neuromast extracellular volume (F–H; arrows depict fluid-filled space). By maneuvering a patch pipette under positive pressure through the path previously made, a hair cell can be patched (I). This approach has been used to record from hair cells in a variety of hair cell organs (Ricci and Fettiplace, 1997; Ricci et al., 2005). We have recorded from >10 neuromasts on several fish. Gigohm seals were readily obtained. With Cs pipette solutions, average membrane capacitance (Cm) of hair cells determined from transient analysis at −80 mV at the beginning of recording was 1.6 ± 0.09 pF (n = 5). Usually, 1–2 hair cells were recorded during a day's work.
Electrode approach for patch clamping zebrafish hair cells. Zebrafish were anesthetized and fixed laterally onto a recording chamber. A–E, Images show neuromast structure from cupola/kinocilia top (A) to location of hair cell soma (E). See Results for details. F–I, Outer supporting cell ring is breeched with large suction electrode, providing a path for patch electrode to seal on hair cell. See Results for details. Scale bar (A), 8 μm. Images from 7 dpf.
Figure 2 shows K currents recorded from a patched hair cell with a K-based intracellular solution. Raw currents were evoked with voltage steps (C) following a prepulse to −120 mV from a holding potential of −80 mV. The outward currents were composed of at least two components, a fast decaying one and a slower noninactivating one (A). The fast-inactivating (A-type) current was removed with a prepulse to −40 mV (B), and by subtraction (traces A minus B), the transient current was isolated (D). Figure 2E shows an average I–V curve of unsubtracted currents collected from a −80 mV holding potential, illustrating outward rectification averaging 0.25 ± 0.07 nA at +10 mV (n = 3). Slope resistance between −120 and −60 mV was ∼2 GΩ. Using Cs solutions and P/N leakage subtraction, we also identified an inward Ca2+ current (Fig. 2F). The current was small and noninactivating during voltage steps. At −10 mV the Ca current averaged 0.018 ± 0.003 nA (n = 5). Ca influx is expected to release synaptic vesicles at the basal pole of hair cells. Figure 2G illustrates the release of vesicles, measured as a membrane capacitance increase. In this case, during the course of a 3 s depolarization to −10 mV, Cm increased ∼40 fF. Open circles depict average responses. The average increase in Cm at 3 s under these conditions was 52 ± 23 fF (n = 3; initial Cm was 1.37 ± 0.15 pF). Assuming a vesicle capacitance of 38–50 aF (Schnee et al., 2005; Graydon et al., 2011), the Cm increase equates to a surface area equivalent to 1040–1370 vesicles. This release is greater than estimates of vesicles associated with synaptic ribbons (2–5 ribbons possessing ∼120 vesicles each; Obholzer et al., 2008; Trapani and Nicolson, 2011), which predicts 12–30 fF for vesicles of 50 aF size. The larger measured values indicate recruitment of non-ribbon-associated vesicles, though not nearly at the rate observed in turtle or rat (Schnee et al., 2011a,b). We also tested for transduction currents during whole-cell voltage clamp by displacing the cupola sinusoidally with a fluid jet delivered by a puff pipette (n = 4) (Fig. 3). Currents of ∼100 pA were evoked in this cell of ∼1.3 GΩ input impedance. A decrease in amplitude during the extent of stimulation resembles adaptation. It should be noted that the Tricane anesthetic that we used is expected to reduce MET function (Farris et al., 2004).
K, Ca currents and synaptic vesicle release from hair cells in living zebrafish. A, Outward K currents induced by voltage steps from prepulse level (−120 mV) to a range of voltages depicted in C. B, K currents generated following prepulse to −40 mV. D, Subtraction of traces (A minus B) reveals A-type inactivating current. E, Outward currents without leakage subtraction. Average I–V (+SE; n = 3) of last 20 ms of 100 ms current traces. F, Inward Ca current is revealed with P/-5 leakage subtraction protocol collected with Cs pipette solution. Average I–V (+SE; n = 5) of last 10 ms of 50 ms current traces. Inset, Noninactivating Ca current trace, generated in response to −30 mV step from −80 mV holding potential. Scale is 6 pA/10 ms. G, Hair cell synaptic vesicle release was determined with membrane capacitance measures. The cell was held at −80 mV and stepped to −10 mV for 3 s; vesicle release is evident from increase in Cm: δCm is ∼40 fF (dark traces). Endocytosis was not observed in our experimental time frame. Dotted lines show initial capacitance level. Open circles indicate average (±SE) responses at 1 (n = 5), 2 (n = 3), and 3 (n = 3) s following the onset of depolarization. Data were obtained at 7–12 dpf.
MET current activity during cupola deflection. A, A patch pipette (pp) voltage-clamped one hair cell in the neuromast, while a puff pipette (puff) was used to stimulate the kinocilia-embedded cupola. B, Suction phase of cupola stimulation showing kinocilia displaced closer to puff pipette. C, Receptor currents generated by sinusoidal fluid jet activity. Stimulus voltage driving the water jet is shown beneath currents. With the solutions we used, rectification of the response is expected (Farris et al., 2004). Scale bar, 6 μm. Data were obtained at dpf 7. Stimulus frequency was 2.0 Hz.
Finally, we observed cells that had larger than the average 1.6 pF capacitance of hair cells, which we identify as supporting cells (n = 4). Auditory support cells are known to be joined into a syncytium and display large capacitances (Santos-Sacchi, 1991). Figure 4 illustrates supporting cell recordings where voltage was ramped to voltage extremes of ∼±90 mV. Capacitance, in this case calculated from transient current response to step voltages, varied according to holding potential (Fig. 4A). In fact, there is the expected reciprocal relationship between input resistance and capacitance as gap junctions uncouple and recouple during depolarization and hyperpolarization, respectively (Fig. 4B). Auditory supporting cells possess voltage-dependent gap junctions (Zhao and Santos-Sacchi, 2000) and exhibit this type of behavior, as well.
Input capacitance and resistance of supporting cell syncytium. A, B, An individual supporting cell was patch-clamped (A) and the membrane voltage modulated with a slow ramp to ∼±90 mV (B). Note slow changes in input capacitance corresponding to membrane voltage, indicative of voltage-dependent gap junctional coupling found in the organ of Corti. Input capacitance indicates that several cells contribute to membrane surface area. Dotted lines indicate zero levels. Data were obtained at 12 dpf. C, In another cell, the reciprocal relationship between input resistance and capacitance as gap junctions uncouple and recouple during depolarization and hyperpolarization, respectively, is shown. Data were obtained at 3dpf.
Discussion
The zebrafish is a powerful model for auditory/vestibular research. Here we provide data demonstrating the feasibility of patch recording hair cells and supporting cells in lateral line neuromasts from the living zebrafish. The approach is similar to the one we have used to record from a variety of auditory/vestibular end organs (Ricci et al., 2005; Schnee et al., 2011a), and should allow direct assessment of hair cell function in the multitude of existing genetic manipulations available in the zebrafish.
Zebrafish hair cells possess outward K currents, and inward Ca currents similarly found in other auditory/vestibular-like organs. For example, the component K currents presented here are similar to those in the frog vestibular hair cell, where depolarizing prepulse was found to dissociate A-type and delayed rectifier-type contributions (Norris et al., 1992). Indeed, the basolateral electrical properties measured in the zebrafish hair cells are quite similar to those measured in bird (Lang and Correia, 1989; Ricci and Correia, 1999), frog (Norris et al., 1992; Masetto et al., 1994), turtle (Brichta et al., 2002), and mammal (Eatock and Hutzler, 1992). Though the zebrafish hair cell Ca current is small, its properties are similar to those of CaV(1.3) identified in zebrafish (Sidi et al., 2004) and its magnitude is similar to that of other vestibular hair cell types including those of frog (Prigioni et al., 1992; Martini et al., 2000), chick (Ohmori, 1984), and rat (Bao et al., 2003). Very recently, dissociated inner ear hair cells of the zebrafish were successfully recorded, and the reported K current was similar to that reported here (Einarsson et al., 2012).
The calcium current elicited was sufficient to evoke substantial transmitter release, indicative of vesicle replenishment from pools remote to the synaptic ribbon. The apparent absence of a superlinear vesicular release component that is found in auditory hair cells of turtle, mouse, and rat (Schnee et al., 2011a,b), and which may be associated with a release of intracellular stored Ca, might underscore differences in functional requirements for vesicle recruitment mechanisms in zebrafish hair cells. Perhaps this difference relates to auditory versus vestibular hair cell properties, although calcium-induced calcium release has been observed in frog vestibular hair cells (Lelli et al., 2003); alternatively, the very small volume of zebrafish hair cells may not require elaborate vesicle recruitment mechanisms, or the small volume may make a kinetic separation between release components more difficult to observe. Finally, in contrast to observations in some other species (Schnee et al., 2005; Graydon et al., 2011), the substantial vesicle release arising from such small Ca currents may indicate efficient synaptic mechanisms. Of course, the preliminary data we report here require further investigation. Nevertheless, one of the main benefits of recordings in the zebrafish hair cell will be to capitalize on the wealth of tools and genetic manipulations available for the ribbon synapse (Nicolson, 2005; Francis et al., 2011) in a preparation that we now show to be readily accessible.
Aside from direct investigations on sensory hair cell mechanisms, the neuromast provides an opportunity to understand supporting cell roles in sensory transduction. In the organ of Corti, electrical coupling has been proposed to sink K away from active hair cells, thereby preserving indefatigable hair cell function (Santos-Sacchi, 1985, 2000). Our observation of similar voltage-dependent gap junctional coupling in zebrafish neuromasts indicates a corresponding role in this simple organ, with the preparation offering model benefits similar to those for hair cell study.
Footnotes
This work was supported by NIH Grants EY021195 to D.Z., DC009913 to A.J.R. and J.S.-S., P30 44992 to A.J.R., and DC000273 to J.S.-S.
- Correspondence should be addressed to Joseph Santos-Sacchi, Surgery (Otolaryngology), Yale University School of Medicine, BML 224, 333 Cedar Street, New Haven, CT 06510. joseph.santos-sacchi{at}yale.edu
