The embryos of zebrafish were obtained from Tg(cldnb:lynGFP)^zf106 line, Tg(brn3c:mGFP)^s356t line, and AB wild type. Spawns were maintained in embryo medium at 28.5°C according to the standard recipe. Hours/days postfertilization (hpf/dpf) was used to take a record of marking embryos’ different developmental stages. Embryos lived in E3 water which add 0.03‰ 1-phenyl-2-thiourea (PTU; Sigma–Aldrich, St Louis, MO, United States) at the stage of 14 hpf, in the result of avoiding pigment formation. The Institutional Animal Care and Use Committee of Fudan University approved all the animal experiments.

Two endogenous blocking strategies were used for inhibiting the expression of Kif15 (Kif15-MO), and the precise sequences were shown as follows:

A total final concentration of 0.3 mM was injected into the yolk sacs of zebrafish embryos at the one- to two-cell stage for Kif15 knockdown. As a control, the parallel embryos were injected with the standard control-MO (Con-MO), sequenced in 5′-CCTCTTACCTCAGTTACAATTTATA-3′ to avoid errors due to injections.

The live zebrafish were anesthetized in 0.02% MS-222 (ethyl 3-aminobenzoate methanesulfonate; Sigma–Aldrich, Inc., Saint Louis, MO, United States) for 3 min and then placed on the glass slide with E3 water under the corresponding microscope. The images were taken with a stereomicroscope or an inverted fluorescence microscope.

After treatment, larvae were moved into wells of a 48-well plate (1 fish/well) to monitor their swimming behavior. Zebrafish behavior was monitored by a digital video tracking system (DVTS, Noldus Information Technology, United States). Each larva was allowed to habituate to the test environment of the system for 30 min before the start of the data acquisition. Notably, 10 acoustic/vibrational stimuli (Danio Vision intensity setting 6) with a 20-s interstimulus interval were set and applied. The total swimming distance of the larvae that responded to this stimulus was recorded. Swimming behavior studies were repeated at least three times.

Zebrafish embryos at 48 hpf after removal of yolk sac were immersed in precooled radioimmunoprecipitation assay (RIPA) lysis buffer plus cocktail and phenylmethanesulfonyl fluoride (PMSF) for schizolysis. To know the exact concentrations of the extracted proteins, we chose a BCA Protein Kit (Beyotime Institute Biotechnology, China) with the supernatant of the RIPA lysate after dissolving the samples. After electrophoresis on SDS–PAGE, the separated proteins were transferred onto poly(vinylidene fluoride) (PVDF) membranes (Immobilon-P; Millipore, Bedford, United States). Before incubation with the primary antibodies, the membranes were blocked with 5% skim milk for 2 h at room temperature. Anti-Kif15 (1:400 dilution, Proteintech, Shanghai, China) and anti-glyceraldehyde 3-phosphate dehydrogenase (anti-GAPDH; 1: 2000) were incubated with the membranes for a whole night at 4°C. Corresponding proteins were combined with Goat Anti-Mouse/Anti-Rabbit immunoglobulin G [IgG; (H + L)] HRP (1:2000 dilution, Abways Technology, Shanghai, China) after being incubated for 2 h. Using an ECL Kit (Millipore, United States), such protein bands could be colored and detected by an Azure C280 imager. The membranes need to be washed thoroughly with Tris-buffered saline with Tween 20 (TBST) before each incubation step. All experiments were strictly from three repeats. GAPDH served as the internal reference.

GraphPad Prism (version, 8.0) supports whole statistical analyses. Comparisons between every two groups were demonstrated with a t-test (two-tailed), while multiple comparisons were illustrated using one-way ANOVA. Statistics were all displayed as mean ± standard error of the mean (SEM), with p-value < 0.05 identified as statistical significance.

The role of Kif15 during inner ear development was investigated by the knockdown of Kif15 through the injection of Kif15 splicing MO to the zebrafish embryos at one- or two-cell stages. We used the Western blotting analysis to test the efficacy of Kif15 downregulation, and we detected a substantial transcript reduction of Kif15 in the zebrafish larvae after Kif15-MO injection compared to the controls injected with Con-MO (Figures 1A,B). We chose 48, 72, and 96 hpf in chronology to observe the morphology changes in the inner ear of zebrafish after Kif15-MO injection (Figures 1C–H). At 72 and 96 hpf, the essential structures of three pairs of semicircular canals (i.e., anterior, posterior, and lateral) were clearly discernible as labeled in Figure 1D, including anterior (a-), posterior (p-), and ventral (v-) protrusions and bulges. However, the normal structures of semicircular canals disappeared and were replaced by short and fused protuberances in Kif15-MO morphants (Figures 1F–H). To verify the role of Kif15 on zebrafish semicircular canal development, we performed a rescue experiment by the co-injection of Kif15 mRNA together with Kif15 splicing MO, and we found a successful rescue in the phenotype of otic vesicle (Figures 1I–K). Our observation implied that the inhibition of Kif15 by specific MO completely ruined the normal development of semicircular canals in the inner ear of zebrafish.

The morphology of otoliths in control and Kif15-MO-injected embryos was detected with a white optical microscope in the bright field at 48 hpf (Figures 2A–E). Kif15 splicing MO-injected embryos exhibited various abnormalities containing increased number of otoliths (35.5%), decreased number of otoliths (4.3%), and two abnormal otoliths (15.4%) (Figures 2B–E,G). Among them, the two abnormal otolith deformities involved situations such as the abnormal size of two otoliths (Figure 2C), the abnormal arranged position of two otoliths (Figure 2D), and the combined disordered form. The proportion of normal otolith phenotype in the Kif15-MO group is only 44.9% in contrast to 87.5% in controls (Figure 2G). On the contrary, there is almost no abnormality in the number of otoliths in the Con-MO injection group (Figure 2G). We also detected successful rescue in the phenotype of otoliths after co-injecting Kif15 mRNA and Kif15 splicing MO compared to that in Kif15-MO morphants (Figures 2F,G).

Using Tg(brn3c:mGFP)^s356t line, the HCs that insert into the otoliths were identifiable and labeled by green fluorescence, named utricular (anterior to left) and saccular (posterior to right) HCs, respectively (Figures 3A,D,G). After Kif15 knockdown with splicing MO, two smaller sizes of otoliths were examined, and the numbers of both utricular and saccular HCs in otic vesicles were reduced significantly compared to the controls (Figures 3B,E,H). However, the decreased number of HCs in both utricle and saccule after Kif15 knocking down could be completely rescued by co-injection with Kif15 mRNA (Figures 3C,F,I). The quantitative analysis further showed that the number of utricular HCs in Kif15 morphants was significantly decreased at 48 hpf compared to that in Con-MO-injected larvae, while a remarkably increased number of utricular HCs were found in Kif15-MO + Kif15 mRNA group (Figure 3J). Similarly, HCs in saccule also severely reduced in Kif15-MO-injected morphants but markedly increased in Kif15-MO + Kif15 mRNA co-injected embryos (Figure 3J).

Then, the function of Kif15 during the development of the PLL system was investigated. The transgenic zebrafish Tg(cldnb:lynGFP)^zf106 was used to visualize NMs in green fluorescence (Haas and Gilmour, 2006). The NMs manifested in severely reduced number and disordered arrangement along the trunk of zebrafish after Kif15-MO injection at 48 hpf, a time point when PLL primordium finishes migration (Figures 4A–C). Co-injecting Kif15 mRNA and splicing MO notably increased the number of NMs in comparison with that in Kif15-MO morphants (Figure 4D). The quantitative analysis showed that the number of PLL NMs in Kif15 morphants was significantly decreased compared to that displayed in Con-MO-injected larvae at 48 hpf (Figure 4E). The number of PLL NMs in the Kif15-MO + Kif15 mRNA group was remarkably larger than that in Kif15-MO morphants but a little smaller than that in Con-MO embryos, indicating a successful but not complete rescue in the number of NMs by Kif15 mRNA (Figure 4E). The diagrammatic sketch demonstrating changes in the number and arrangement of NMs along the trunk and tail of zebrafish was presented after injection with Kif15-MO (Figure 4F). The result provided substantial evidence that Kif15 is critical to zebrafish in auditory system development.

According to the brief review of the literature in the background that sensory organs of zebrafish including otoliths, semicircular canals, and PLL NMs can detect sound and water movement stimuli and move in directionality, we wondered if the locomotor behavior of zebrafish was disturbed in Kif15-MO embryos. The locomotion assay was conducted to detect the locomotor behavior of zebrafish after Kif15 inhibition with splicing MO since behavioral change is a comprehensive indicator to show the impact of environmental factors on the organisms. The locomotor traces were recorded during the light-dark photoperiod stimulation with a 20-s interstimulus interval, and we detected a severe reduction in locomotor traces compared to that in the controls (Figure 5A). However, a significant increase in locomotor traces was found after performing the rescue experiment with co-injected Kif15 mRNA compared to that in Kif15-MO morphants (Figure 5A). Data analysis on the distance moved from the original point in different groups showed that the Kif15-MO injection caused a significant reduction in the distance of locomotion in comparison with the parallel controls. There was also a significant difference in distance of movement between Kif15-MO + Kif15 mRNA and Con-MO groups, indicating a partial rescue effect of Kif15 mRNA in the locomotor behavior (Figure 5B). The results were indicative that behavioral dysfunction also existed in addition to morphological deformity after Kif15 inhibition.