Unraveling the Molecular Players at the Cholinergic Efferent Synapse of the Zebrafish Lateral Line

Abstract

The lateral line (LL) is a sensory system that allows fish and amphibians to detect water currents. LL responsiveness is modulated by efferent neurons that aid in distinguishing between external and self-generated stimuli, maintaining sensitivity to relevant cues. One component of the efferent system is cholinergic, the activation of which inhibits afferent activity. LL hair cells (HCs) share structural, functional, and molecular similarities with those of the cochlea, making them a popular model for studying human hearing and balance disorders. Because of these commonalities, one could propose that the receptor at the LL efferent synapse is a α9α10 nicotinic acetylcholine receptor (nAChR). However, the identities of the molecular players underlying ACh-mediated inhibition in the LL remain unknown. Surprisingly, through the analysis of single-cell expression studies and in situ hybridization, we describe that α9, but not the α10, subunits are enriched in zebrafish HCs. Moreover, the heterologous expression of zebrafish α9 subunits indicates that homomeric receptors are functional and exhibit robust ACh-gated currents blocked by α-bungarotoxin and strychnine. In addition, in vivo Ca²⁺ imaging on mechanically stimulated zebrafish LL HCs show that ACh elicits a decrease in evoked Ca²⁺ signals, regardless of HC polarity. This effect is blocked by both α-bungarotoxin and apamin, indicating coupling of ACh-mediated effects to small-conductance Ca²⁺-activated potassium (SKs) channels. Our results indicate that an α9-containing (α9*) nAChR operates at the zebrafish LL efferent synapse. Moreover, the activation of α9* nAChRs most likely leads to LL HC hyperpolarization served by SK channels.SIGNIFICANCE STATEMENT The fish lateral line (LL) mechanosensory system shares structural, functional, and molecular similarities with those of the mammalian cochlea. Thus, it has become an accessible model for studying human hearing and balance disorders. However, the molecular players serving efferent control of LL hair cell (HC) activity have not been identified. Here we demonstrate that, different from the hearing organ of vertebrate species, a nicotinic acetylcholine receptor composed only of α9 subunits operates at the LL efferent synapse. Activation of α9-containing receptors leads to LL HC hyperpolarization because of the opening of small-conductance Ca²⁺-activated potassium channels. These results will further aid in the interpretation of data obtained from LL HCs as a model for cochlear HCs.

Keywords: Xenopus oocytes; calcium imaging; efferent; lateral line; nicotinic receptor; zebrafish.

Figure 1.

ɑ9 (but not ɑ10) is expressed in zebrafish LL neuromasts and the posterior macula in the otic vesicle. A–F, Whole-mount in situ hybridization with antisense (A–D) and sense (E) ɑ9, and antisense ɑ10 (F) riboprobes. Representative lateral views, with anterior to the left and dorsal to the top, are shown. Arrow indicates the otic vesicle, and arrowheads point to selected neuromasts. B–D, Large-scale view of the otic vesicle (C) and neuromasts (B, D). C, Dotted line delimits the otic vesicle; dotted-dashed line outlines the posterior macula (pm). Scale bars: in A, E, F, 100 µm; B, 40 µm; C, 25 µm; D, 10 µm.

Figure 2.

Zebrafish recombinant ɑ9 forms homomeric and heteromeric receptors with ɑ10 with distinct biophysical properties. A, Representative responses evoked by ACh in oocytes expressing zebrafish ɑ9, ɑ10, or ɑ9ɑ10 (1:2) nAChRs. B, Concentration–response curves for zebrafish ɑ9, ɑ9ɑ10 (1:1), and ɑ9ɑ10 (1:2) nAChRs. Values are the mean ± SEM. Lines are best fit to the Hill equation. C, Top, Representative responses of zebrafish α9 and α9α10 (1:2) nAChRs to a 60 s application of ACh (one order of magnitude higher than their corresponding EC₅₀). Bottom, Desensitization rate shown as percentage of current remaining 20 s after the peak response relative to the maximum current amplitude elicited by ACh. Lines indicate the median and IQR. Symbols represent individual oocytes (n = 13 and 9, respectively, *p = 7.09e-06, U=0.0, Mann–Whitney test). D, Representative I–V curves obtained by the application of voltage ramps (−120 to +50 mV, 2 s) at the plateau response to 10 μm ACh for both zebrafish α9 and α9α10 (1:2) nAChRs. Values were normalized to the agonist response at +50 mV for each receptor.

Figure 3.

Zebrafish α9 and α9α10 nAChRs have a high Ca²⁺ contribution to the total inward current and are not modulated by extracellular Ca²⁺. A, Top, Representative responses to a near EC₅₀ concentration of ACh (α9, 10 μm; α9α10, 300 μm) in oocytes expressing zebrafish α9 and α9α10 nAChRs before (light colors) and after (solid colors) a 3 h incubation with BAPTA-AM. Bottom, Percentage of the initial peak response remaining after BAPTA-AM incubation. Lines indicate the median and IQR. Symbols represent individual oocytes (n = 14 and 8, respectively), B, ACh response amplitude as a function of extracellular Ca²⁺ concentration (top, α9; bottom, α9α10). ACh was applied at a near EC₅₀ concentration (α9, 10 μm; α9α10, 300 μm). Current amplitudes recorded at different Ca²⁺ concentrations in each oocyte were normalized to the response obtained at 1.8 mm Ca²⁺ in the same oocyte (α9, gray circles; α9α10, pink circles). V_hold, −90 mV. Bars represent mean ± SEM (α9, black bars, n = 8; α9α10, red bars, n = 5).

Figure 4.

Zebrafish α9 and α9α10 nAChRs are reversibly blocked by ɑ-Btx and strychnine. A, B, Responses to 10 μm (α9) or 300 μm (α9α10) ACh alone, in the presence of α-Btx (A) or Str (B), or after washing with control bath solution for 5 min in oocytes expressing zebrafish α9 or α9α10 nAChRs are shown. A, Oocytes were preincubated with 100 nm α-Btx for 1 min before the addition of the agonist. α-Btx inhibited ACh-elicited responses through α9 nAChRs by 94.59 ± 2.17% (n = 3) and through α9α10 nAChRs by 83.66 ± 5.50% (n = 3). B, Oocytes were preincubated with 1 μm Str for 1 min before the addition of ACh. Str inhibited ACh-evoked currents through α9 nAChRs by 95.44 ± 1.19% (n = 4) and through α9α10 nAChRs by 57.59 ± 5.1% (n = 3).

Figure 5.

Mechanical stimulation elicits a robust Ca²⁺ signal that is inhibited by isradipine A, Representative functional Ca²⁺ images of a double-transgenic neuromast expressing GcAMP7a in HCs. A, Prestimulus baseline grayscale image (ROIs are drawn around each visible hair cell; i), spatial patterns of GCaMP7a Ca²⁺ signals (ii, iii), during a 2 s mechanical stimulus in either the anterior–posterior (→) or in the posterior–anterior direction (←), are color coded according to the ΔF/F₀ heat map. B, Representative temporal curves of mechanosensitive Ca²⁺ responses (ΔF/F₀) of HCs numbered in A, normalized to the peak intensity for each cell. Shaded areas indicate the time when the neuromast was mechanically stimulated. C, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs over two trials with the same stimulation after 1 min (1° stimulus, light red; 2° stimulus, dark red). Curves are aligned to the onset of the mechanical stimulus. Bottom, Peak ΔF/F₀ for single HCs (n = 113, each in its preferred orientation) over two trials with the same stimulation 1 min apart. D, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs before (red) and after (purple) preincubation with 10 μm isradipine. Curves are aligned to the onset of the mechanical stimulus. Bottom, Preincubation with 10 μm isradipine drastically reduced peak ΔF/F₀ (n = 23, W = −258, *p = 8.726e-05, MPRBC = 0.935; Wilcoxon matched-pairs signed-rank test). Scale bar, A, 5 μm. Calibration: C, D, 1.5 s; C, D, 25% ΔF/F₀. Duration of the stimulus in C and D top, is indicated by gray lines below each trace.

Figure 6.

ACh inhibits mechanically evoked Ca²⁺ signals, and this inhibition is heterogeneous and independent of HC polarity A, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs over two trials with the same stimulation 1 min apart (1° stimulus, light red; 2° stimulus, dark red). Curves are aligned to the onset of the mechanical stimulus. Bottom, Peak ΔF/F₀ for single HCs (n = 113) over two trials with the same stimulation after 1 min. B, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs before (red) and after (blue) the application of 1 mm ACh. Curves are aligned to the onset of the mechanical stimulus. Bottom, ACh application reduces mechanosensitive Ca²⁺ responses (n = 114, W = −3493, *p = 7.89e-07, MPRBC = 0.532, Wilcoxon matched-pairs signed-rank test). C, ACh-mediated reduction in mechanically evoked Ca²⁺ signals is reversed after a 1 min wash with extracellular imaging solution (n = 37; Friedman test: F = 18.54, p = 9.418e-05; Dunn's multiple comparisons test: Extra vs ACh, *p = 0.000705; Extra vs Wash, p = 0.608054). D, Basal Ca²⁺ levels show no significant differences before and during the application of 1 mm ACh (n = 45 cells, t = 0.7816, df = 44, p = 0.4386, two-tailed paired t test). E, ACh reduces mechanosensitive Ca²⁺ responses in HCs of opposing polarity (Ant-Post: n = 62, W = −1227, *p = 1.698e-05, MPRBC = 0.628; Post-Ant: n = 52, W = −582, *p = 0.008, MPRBC = 0.422; Wilcoxon matched-pairs signed-rank test). F, HCs of opposing polarity exhibit no significant differences between their ACh-mediated relative change in peak ΔF/F₀ (U = 1560, p = 0.7687, Mann–Whitney test). G, Distribution of II calculated as (ΔF/F₀extra – ΔF/F₀ACh)/ΔF/F₀extra) for ACh-treated HC. Inset, Distribution of change index [calculated as (ΔF/F₀stim1 – ΔF/F₀stim2)/ΔF/F₀stim1] for two successive mechanical stimuli under control conditions. The distribution of the change index is centered around 0. A reduced number of cells (<10%) exhibit large negative change index values that occur when the fluorescence signal is greater during the 2° stimulus, suggesting that these might be outliers. Lines inside violin plots in A, B and E indicate the median and IQR. Calibration: A, B, 1.5 s; A, B, 25% ΔF/F₀. Duration of the stimulus in A and B, top, is indicated by gray lines below each trace.

Figure 7.

ACh-mediated inhibition of evoked Ca²⁺ signals is blocked by α-Btx and apamin. A, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs, before (red) and after (green) the application of 10 μm α-Btx. Bottom, Mechanosensitive Ca²⁺ signals show no significant difference before and after 10 μm α-Btx treatment (Extra: median ΔF/F₀ = 0.509; IQR, 0.252–1.134; vs α-Btx: median ΔF/F₀ = 0.534; IQR, 0.331–1.325; n = 25, W = −45, p = 0.5449, MPRBC = 0.138). B, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs, after the application of 10 μm α-Btx (green) and after the coapplication of 1 mm ACh and 10 μm α-Btx (blue). Bottom, When coapplied with 10 μm α-Btx, ACh-mediated inhibition is blocked (n = 25, W = −87, p = 0.2521, MPRBC = 0.268). C, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four hair cells, before (red) and after (orange) the application of 10 μm apamin. Bottom, Mechanosensitive Ca²⁺ signals show no significant difference before and after 10 μm apamin treatment (Extra: median ΔF/F₀ = 0.599; IQR, 0.243–1.216; versus Apa: median ΔF/F₀ = 0.567; IQR, 0.191–1.040; n = 41, W = −91, p = 0.5554, MPRBC = 0.106). D, Top, Representative temporal ΔF/F₀ curves of mechanosensitive Ca²⁺ responses of four HCs, after the application of 10 μm apamin (orange) and after the coapplication of 1 mm ACh and 10 μm apamin (blue). Bottom, ACh-mediated inhibition is blocked by 10 μm apamin (n = 60, W = −322, p = 0.2359, MPRBC = 0.1759). A Wilcoxon matched-pairs signed-rank test was used in all cases. Calibration: A–D, 1.5 s; A–D, 25% ΔF/F₀. Curves in A–D are aligned to the onset of the mechanical stimulus. The duration of the stimulus is indicated by gray lines below each trace.

Figure 8.

Schematics of the cholinergic LL efferent synapse. LL HCs are innervated by afferent (red) and cholinergic efferent (green) fibers. Evidence for efferent cholinergic fibers contacting afferent neurons (dashed light green) is still missing. The net effect of LL efferent cholinergic activity is to hyperpolarize HCs. This is mediated by the activation of an α9* nAChR with high Ca²⁺ permeability. Subsequent activation of Ca²⁺-dependent potassium SK channels drives HC hyperpolarization. Postsynaptic cisterns (PCs) opposed to efferent terminals (Dow et al., 2018) have been proposed to participate in Ca²⁺ compartmentalization and/or Ca²⁺-induced Ca²⁺ release mechanisms.