Identification of milky way as a hypomorphic allele of nsfa

(A) Map of the critical region. A missense mutation was identified in nsfa. The null allele of nsfa (st53) is also indicated. Protein sequence of the linker region of NSF near I209 (boxed) from various species.

(B)Expanded melanophore phenotype.

(C)Nsfa immunolabeling in spinal cord and motor neurons in WT, nsfa^I209N homozygous mutants, dnsfa^I209N/st53 compound heterozygotes, and nsfa^st53 homozygous mutants. Scale bar, 100 μm.

(D)Summary of phenotypes in the null and I209N nsfa mutants.

(E–J) Myelination and the number of nodes of Ranvier in the lateral-line nerve of mutant nsfa^I209N/st53 larvae are comparable with that seen in WT siblings. Representative images (E–G) and quantification (H–J) of MBP, AcTub, and FIGQY antibody labeling (with colabeling of the postsynaptic density by MAGUK) in WT, nsfa^I209N/st53, and nsfa^st53 larvae. Images are maximum intensity z projections from the initial segment of the posterior lateral-line nerve. Quantification data are shown as mean ± SEM; p values were determined by Student’s t tests to compare with WT group. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001. For all experiments, n ≥ 5 fish per genotype (5 dpf). All images and data are representative of 2 or 3 independent experiments. Scale bar, 10 μm.

The I209N mutation in Nsfa selectively affects hearing and vestibular function without causing paralysis

(A) Still frames from high-speed videos of larvae. nsfa^st53 homozygotes were paralyzed and unresponsive to touch, whereas nsfa^I209N homozygotic and nsfa^I209N/st53 larvae displayed robust startle reflexes.

(B) (B and C) Acoustic evoked behavioral responses of larvae exposed to a 600-Hz stimulus at the intensities indicated normalized to homozygous WT siblings. n indicates number of fish tested.

(D–K) Vestibulospinal reflexes are reduced in nsfa^I209N homozygotic and nsfa^I209N/st53 larvae. (E–G) Raw traces (blue) of the tail movements of representative WT and mutant larvae. The rolling median (orange trace) is a movement artifact. (H–K) The maximum tail angle (H and I) and the normalized integral (J and K) of sibling cohorts were quantified using ZebraZoom. n indicates number of fish tested.

Mean ± SEM and two-way ANOVA with Benjamini-Hochberg correction for each dataset were performed. *p < 0.05, **p < 0.01, ***p < 0.001.

The gross morphology of ribbon synapses is normal in nsfa^I209N/st53 mutants

(A) Image and diagrams of neuromasts depicting the hair cells, afferent neurons, and ribbon synapses.

(B) Representative images of afferent innervation (HNK-1) of hair cells in WT, nsfa^I209N/st53 mutants, and nsfa^st53 mutants. Presynaptic hair-cell ribbons are labeled with Ribeye b antibody (magenta). Note the lack of innervation of nsfa null hair cells.

(C) Ribbon synapses in WT, nsfa^I209N/st53 mutants, and nsfa^st53 mutants. A pan-MAGUK antibody was used to label the postsynaptic density of afferent terminals (green), and ribbons were visualized with anti-Ribeye b antibody (magenta).

(D) Number of ribbons per neuromast (n ≥ 12 neuromasts per genotype).

(E) Average size of each ribbon (n ≥ 12 neuromasts per genotype).

(F) Colocalization of Ribeye b and MAGUK (n ≥ 8 neuromasts per genotype).

(G) Integrated density of MAGUK immunolabel per punctum (n ≥ 12 neuromasts per genotype).

(H) Images of the synaptic vesicle marker VGlut3 in WT, nsfa^I209N/st53 mutants, and nsfa^st53 mutants (Ribeye b in magenta).

(I) Quantification of the intensity of the VGlut3 immunolabel (n ≥ 11 neuromasts per genotype).

Quantification data are shown as mean ± SEM; p values are determined by ANOVA tests to compare with WT group. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001. For all experiments, n ≥ 6 fish per genotype (5 dpf). All images and data are representative of 2 or 3 independent experiments. Scale bars, 5 μm.

Decreased phase locking to mechanical stimuli at hair-cell ribbon synapses in nsfa^I209N/st53 mutants at 60 Hz (A) Schematic of the recording paradigm. Hair cells were mechanically stimulated after establishing a loose-patch recording from the innervating afferent neuron (5 dpf). (B and C) Evoked spike rates and phase locking during a lower-frequency stimulus.

(B) Representative traces from WT (upper) and nsfa^I209N/st53 mutants (middle) during 20-Hz stimulation (bottom). Shown are 60 overlaid sweeps of spiking resulting in 546 spikes in WT and 559 spikes in nsfa^I209N/st53 mutants.

(C) Average spike latency histograms from all spikes during 60 continuous seconds of 20-Hz stimulation. WT latency values (black bars) and nsfa^I209N/st53 mutant values (gray bars) were fit by a Gaussian distribution (black and gray line, mean fraction of 20-Hz period 0.158 ± 0.002 and 0.159 ± 0.003 s, respectively).

(D and E) The timing of evoked spikes is less tightly coupled to a higher-frequency stimulus in the nsfa^I209N/st53 mutant. (D) Representative traces from WT (upper) and nsfa^I209N/st53 mutants (middle) during 60-Hz stimulation (bottom). Shown are 60 overlaid sweeps of spiking resulting in 219 spikes in WT and 217 spikes in nsfa^I209N/st53 mutants. (E) Average latency histograms from all spikes during 60 continuous seconds of 60-Hz stimulation. WT latency values (black bars) and nsfa^I209N/st53 values (gray bars) were fit by a Gaussian distribution (black and gray line, mean fraction of 60-Hz period 0.258 ± 0.003 and 0.298 ± 0.003 s, respectively). The peak of activity in mutant neurons is shifted to a later time point of the 60-Hz cycle (4.28 versus 4.95 ms) in comparison.

(F) Spike rate comparison between WT and nsfa^I209N/st53 mutants at both 20 Hz (WT, 12 ± 3 spikes/s, n = 4; nsfa^I209N/st53, 13 ± 3 spikes/s, n = 4, p = 0.80) and 60 Hz (WT, 19 ± 3 spikes/s, n = 7; nsfa^I209N/st53, 16 ± 1 spikes/s, n = 7, p = 0.32). Center lines represent the mean.

(G) Vector strength (r) of the coupling between stimulus and response (phase locking) between WT and nsfa^I209N/st53 mutants at 20 Hz (WT, r = 0.96 ± 0.02, n = 4;

nsfa^I209N/st53 mutants, r = 0.96 ± 0.02, n = 4) and 60 Hz (from C: WT, r = 0.91 ± 0.01, n = 6; nsfa^I209N/st53 mutants, r = 0.81 ± 0.02, n = 7, p = 0.001). p values are determined by unpaired Student’s two-tailed t test. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001.

Aberrant timing and decreased recovery of spontaneous activity at hair-cell ribbon synapses in nsfa^I209N/st53 mutants

(A) Representative traces of spontaneous afferent activity in WT (green and orange traces) and nsfa^I209N/st53 mutants (brown and blue traces).

(B and C) Mean ISI (B) and ISI CV (C) for spontaneous activity for WT (n = 8 cells) and nsfa^I209N/st53 mutants (n = 4 cells). Mean ± SEM are shown; p values were determined by unpaired Student’s two-tailed t tests; *p < 0.05, **p < 0.01, ***p < 0.001.

(D–G) Recurrence plots reveal regularity of the timing of consecutive spikes (spike time n plotted versus spike time n – 1) for the four cells shown in (A). Note the greater spread and clustering in three quadrants for the mutant panels, indicating a more irregular, bursting pattern in mutant neurons.

(H–K) Recovery of spontaneous activity is delayed in nsfa^I209N/st53 mutants. (H) Representative traces show the return of spontaneous activity after a prolonged stimulation of hair cells at 60 Hz. (I and J) Corresponding recurrence plots for the first 900 recovered spontaneous spikes for the recordings shown in (H).

(K) Recovery of spontaneous spike rate as shown by cumulative spike number over time following cessation of 90 s of 60-Hz stimulation. Shading represents error. n indicates cell number.

I209N NSF-αSNAP form complexes with either ternary or binary neuronal SNARE complexes

(A) Cryo-EM structure of NSF 20S complex (PDB: 6MDM).

(B) N-D1 linker region. I209 is shown as a red stick model.

(C) Size-exclusion chromatography (SEC) profile of t20S consisting of I209N NSF, ternary (syntaxin-1A-SNAP-25-synaptobrevin-2) SNARE complex (tSNARE), and αSNAP mixed in 1:5:25 M ratios. Predicted molecular weights for the SEC are indicated above the dotted line.

(D) SEC of b20S consisting of I209N NSF, binary (syntaxin-1A-SNAP-25) SNARE complex (bSNARE), and αSNAP mixed in 1:5:25 M ratios.

(E) SDS-PAGE gel of fractions from SEC of t20S. The elution volumes are indicated above the gel.

(F) SDS-PAGE of protein ladder and gel fractions from SEC of b20S.

(G) SDS-PAGE gel of protein ladder, NSF stock, tSNARE, bSNARE, αSNAP, t20S filtered through a 100k MWCO concentrator, and b20S filtered through a 100k MWCO concentrator.

SEC peaks are typical of a 20S complex preparation with WT NSF.⁵ The sizes of the protein species are as follows: I209N NSF, 82.8 kDa monomeric and 496.8 kDa hexameric; αSNAP, 33.3 kDa; ternary SNARE complex, 65.2 kDa; binary SNARE complex, 54.4 kDa.

Kinetic traces of I209N NSF show diminished ability to disassemble ternary SNARE complex but retention of binary SNARE complex disassembly activity

(A) Schematic of NSF-mediated disassembly. NSF (purple), αSNAP (beige), and SNARE helices are depicted.

(B) NSF-mediated disassembly of ternary complex. [NSF] = 42.2 nM.

(C) NSF-mediated disassembly of binary complex. [NSF] = 42.2 nM.

(D) Initial rate of NSF-mediated disassembly of ternary complex with NSF. [NSF] = 42.2 nM.

(E) Initial rate of NSF-mediated disassembly of binary complex with NSF. [NSF] = 42.2 nM.

(F) NSF-mediated disassembly of ternary complex. [NSF] = 8.3 nM.

(G) NSF-mediated disassembly of binary complex. [NSF] = 8.3 nM.

(H) Initial rate of NSF-mediated disassembly of ternary complex with NSF. [NSF] = 8.3 nM.

(I) Initial rate of NSF-mediated disassembly of binary complex with NSF. [NSF] = 8.3 nM.

All experiments are n = 3. Error bars in (B), (C), (F), and (G) are SEM of the three independent experiments. Error bars in (D), (E), (H), and (I) are SE from Python package SciPy linear regression. p values in (D), (E), (H), and (I) are derived from one-way ANOVA. ns, not significant; *p < 0.05, **p < 0.01.