DNA was extracted from peripheral blood lymphocytes collected from all individuals belonging to the family of interest. We performed customized sequencing of epilepsy-related genes ( Supplementary Table 1 ) using NextSeq500 (Illumina, San Diego, CA, USA) to identify a possible causative gene mutation. Single nucleotide variations, insertions and deletions were then examined using GATK (https://software.broadinstitute.org/gatk/). Variants annotated in ANNOVAR (http://annovar.openbioinformatics.org/en/latest/) and variants with a minor allele frequency of at least 0.05 were filtered out (the candidate normal population database included 1000 Genomes, Exome Variant Server and EXAC). PolyPhen-2, SIFT and MutationTaster and GERP++ were used to predict the damage caused by all candidate variants, and the candidate causative variants obtained from the sequencing analysis were also confirmed by Sanger sequencing.

Flag-tagged human WT KIF1A was obtained by subcloning the complementary DNA (cDNA) into the pcDNA3.1-3xFlag-T2A-EGFP plasmid using primers:

The missense mutation 1190C > A (p. Ala397Asp) was constructed by overlap PCR mutagenesis using the following primers:

The procedure used for culturing primary hippocampal neurons was previously described (Yang et al., 2019). Briefly, hippocampal neurons were prepared from C57BL/6 mouse embryos at gestational day 18. The brains were removed, and the hippocampi were dissected from the brains. The obtained tissue was digested using 3 ml of trypsin solution for 15 min. The obtained neurons were cultured in neurobasal medium supplemented with B27, L-glutamine, penicillin, and streptomycin. The neurons were plated onto glass coverslips coated with poly-D-lysine in a 37°C incubator with 5% CO₂. The neurons were transfected with KIF1A-WT or KIF1A-A397D plasmids using the calcium phosphate transfection method at 7 days in vitro (DIV7).

Whole-cell recordings were obtained from green fluorescent protein (GFP)-positive neurons transfected with KIF1A-WT or KIF1A-A397D plasmids. To explore the intrinsic excitability of the neurons, we applied a depolarizing current of 500 ms in the current clamp mode starting from −30 pA and at increments of 10 pA to induce action potentials. The rheobase was defined as the first current step that was able to induce action potential firing in a neuron. For the evaluation of synaptic transmission, we maintained the neurons at the potential of −70 mV in artificial cerebrospinal fluid (ACSF) containing 140 mM NaCl, 5 mM KCl, 1.8 mM CaCl₂, 1.2 mM MgCl₂, and 10 mM D-glucose. The miniature excitatory post-synaptic currents (mEPSCs) were recorded in the presence of 1 μM tetrodotoxin (TTX) and 0.1 mM picrotoxin (PTX), and the miniature inhibitory post-synaptic currents (mIPSCs) were recorded in the presence of 1 μM TTX, 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and 100 μM (2R)-amino-5-phosphonovaleric acid (APV). Electrophysiological data were acquired using a Multiclamp 700B amplifier and Digidata 1440A, and the data were analyzed using Mini Analysis (Synaptosoft, Leonia, NJ, USA) and Clampfit 10.3.

For western blotting, total protein was obtained from primary cultured hippocampal neurons 3 days after transfection with the KIF1A-WT or KIF1A-A397D plasmid using the calcium phosphate transfection method at DIV7. The western blot analysis was performed as described previously (Zhang et al., 2019), antibodies using for western blot including: mouse anti-Flag (Sigma), Goat anti-mouse IgG (Proteintech), rabbit anti-GAPDH (Proteintech), Goat anti-rabbit IgG (Proteintech).

For immunofluorescence staining, the cultured neurons were fixed with 4% paraformaldehyde/4% sucrose in phosphate-buffered saline (PBS) for 40 min and permeabilized with 0.3% Triton X-100 in PBS for 15 min. Primary antibodies were diluted and added to the coverslip, and incubated overnight in a humidified chamber at 4°C, then coverslips were washed three times with PBS and incubated with the secondary fluorescent antibodies at room temperature for 1 h. Images were captured using a confocal laser scanning microscope (Leica, Wetzlar, Germany). The following primary antibodies were used: rabbit anti-GFP (Invitrogen), Guinea pig anti-vGLUT (synaptic systems), mouse anti-PSD-95 (Cell Signaling Technology), mouse anti-Flag (Sigma). The following secondary antibodies were used: Alexa Fluor 488-conjugated goat anti rabbit IgG (Invitrogen), Alexa Fluor 405-conjugated goat anti Guinea pig IgG (Jackson ImmunoResearch), Alexa Fluor 594-conjugated goat anti mouse IgG (Invitrogen).

For the assessment of neuronal morphology, the neurons were transfected with the KIF1A-WT or KIF1A-A397D plasmid at DIV7 and fixed at DIV10. In Sholl analysis, concentric circles were drawn around the soma every 10 μm, and the intersections with neurite branches were counted, then primary and secondary neurites in the same neurons were then counted.

For the analysis of spine density, the neurons were transfected with the KIF1A-WT or KIF1A-A397D plasmid at DIV7 and fixed at DIV16. The spine density was then quantified from two randomly selected secondary or tertiary dendrites per neuron. All these experiments were performed in a blinded manner by two observers.

Adult male and female zebrafish (Danio rerio) of the AB strain were obtained from the China Zebrafish Resource Center (http://www.zfish.cn/) and maintained at 28.5°C under a 14-h light/10-h dark cycle using standard procedures. The fertilized eggs were collected via natural spawning. The fertilized embryos were maintained in medium containing 1.5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.6, 17.4 mM NaCl, 0.21 mM KCl, 0.12 mM MgSO₄, and 0.18 mM Ca(NO3)₂ at 28.5°C. All the experiments with zebrafish were approved by the Ethics Committee of Chongqing Medical University.

WT zebrafish kif1aa cDNA was cloned into the Tol2 expression vector (a gift from Koichi Kawakami, National Institute of Genetics) using following primers:

The mutation encoding A433D in zebrafish was cloned into the kif1aa sequence via PCR site-directed mutation using primers: F: gacaaaagcccttcctactacact; R: tagtaggaagggcttttgtctatttgagtaatctgctgtatcctg. Embryos at the one-cell stage were transfected with the WT kif1aa-Tol2 or A433D kif1aa-Tol2 plasmid with transposase via cytoplasmic microinjection.

For behavior monitoring, zebrafish larvae at 5 days post-fertilization (d.p.f.) were placed in 24-well Falcon culture dishes that contained embryo medium. A charge-coupled device (CCD) camera and EthoVision 3.1 locomotion tracking software were used to monitor the swim activity of the mutant (A433D kif1aa-Tol2) and WT (WT kif1aa-Tol2) larvae, and each larva was monitored for 15 min. The swim activity was categorized by three observers who were blind to the larva phenotype as stage 0 (baseline activity), stage 1 (small increase in swim activity), and stage 2 (large increase in movement) (Hortopan et al., 2010).

For local field potential recording, zebrafish larvae at 5 d.p.f. were immobilized in low-melting temperature agarose. A glass filled with 2 M NaCl was placed in the optic tecum of zebrafish larvae, and each recording was performed for 15 min in the current clamp mode with high-pass filtering above 0.1 Hz and low-pass filtering below 1 kHz. The digital gain was 10 [MultiClamp 700B Amplifier (Axon, Sunnyvale, CA, USA) and Digidata1440A (Axon, Sunnyvale, CA, USA)]. Spontaneous events were defined as instances in which the amplitude exceeded three times the background noise (Schubert et al., 2014). Clampfit 10.3 (Molecular Devices, Sunnyvale, CA, USA) software was used for the data analyses.

The measurement data (means ± SDs) from two groups were compared using Student's t test, and the data from more than two parameters were analyzed by two-way ANOVA. Chi-square tests were used to compare the differences in the behavior and local field potential (LFP) of larvae. P < 0.05 was considered statistically significant, and the statistical analyses were performed using GraphPad Prism version 7.0 (GraphPad Software).

The family of interest had six affected patients over three generations, which suggests autosomal dominant inheritance ( Figure 1A ). The clinical characteristics of the epilepsy patients in the family are described in Table 1 . All the affected individuals presented with generalized tonic-clonic seizures without mental retardation, and three of these patients suffered from diabetes mellitus. The seizure onset age of patient III2 was 12 years, and electroencephalogram (EEG) showed generalized sharp waves that were prevalent bilaterally ( Figure 1B ). She started treatment with levetiracetam, which resulted in good control of her seizures. The seizure onset ages of the other patients, namely, II2, II3, II5, and II9, were 18, 16, 21, and 24 years, respectively. None of these individuals had received standard anti-epileptic therapy. Patient II9 presented an abnormal brain CT scan due to ischemic stroke, and the CT scans of the brains of the other patients were normal.

We first performed customized sequencing of epilepsy-related genes of two epilepsy patients (III2 and II3) and one unaffected individual (II4) and identified a possible causal variant (c.1190C > A, p. Ala397Asp) on the neck-coiled coil of KIF1A in heterozygosis. The frequency of this mutation in the Genome Aggregation Database (http://gnomad.broadinstitute.org) is 0.000008133. Sanger sequencing of this gene in all members of this family was then performed to assess the segregation of KIF1A mutations. This rare mutation was found in all affected family members and was not detected in the unaffected family members ( Figure 1C ). In addition, the mutation was presumed to be damaging by PolyPhen-2 and MutationTaster.

The location of the identified mutation in KIF1A is strongly conservation between all species [( Figure 1D ) www.ncbi.nlm.nih.gov/homologene]. As shown in Figure 2 , many point mutations in KIF1A have been identified to date, and these mutations have been associated with various neuropathies.

Prior to using the primary cultured neurons in subsequent experiments, the expression of the wild-type and mutant KIF1A proteins in the neurons was confirmed by western blot and immunofluorescence. The western blot results suggest that the mutant and wild-type proteins were expressed equally in the primary cultured neurons ( Supplementary Figures S1A, B ). Also, we revealed the wild type protein and mutant protein expressed equally in soma and nerites ( Supplementary Figures S1C, D ).

In general, increased neuronal firing is due to intrinsic excitability or altered synaptic transmission. Therefore, we tested the intrinsic excitability of the neurons through whole-cell patch-clamp recordings of the primary cultured neurons. The primary cultured neurons were transfected with the indicated plasmid at DIV7, and the whole-cell recording was performed at DIV14. We first checked the resting membrane potential of each neuron and found no difference between the WT and mutant groups ( Figure 3C ). A depolarizing current of 500 ms was then applied in the current clamp mode starting from −30 pA and at increments of 10 pA to induce neuronal firming ( Figure 3A ). We analyzed the injected current intensity between the two groups and found no difference ( Figure 3D ). The action potential recordings revealed that the number of action potentials induced by the injected currents was unaffected ( Figures 3B, E ). In conclusion, mutant KIF1A did not affect the intrinsic excitability of neurons.

We did not find any difference in intrinsic excitability between the WT and mutant groups and therefore speculate that mutant KIF1A might disrupt synaptic transmission. To investigate this hypothesis, we obtained whole-cell patch-clamp recordings of mEPSCs and mIPSCs in neurons at DIV14 that had been transfected with the plasmid at DIV7. Compared with the expression of WT KIF1A, the expression of mutant KIF1A resulted in a significant increase in the frequency of mEPSCs in neurons, whereas the amplitude of mEPSCs did not show a significantly difference between neurons expressing WT and those expressing mutant KIF1A ( Figures 4A–C ). The mIPSC analysis revealed that the frequency and amplitude did not differ between the two neuron groups ( Figures 4D–F ). Taken together, these results indicate that the expression of mutant KIF1A leads to an enhanced excitatory synaptic transmission.

Previous studies have shown that KIF1A is a neuron-specific expressed protein (Okada et al., 1995) and have demonstrated that knockout of KIF1A in mice results in death within 24 h after birth (Yonekawa et al., 1998). We thus hypothesize that KIF1A plays an important role in neuronal development and that mutant KIF1A might be involved in neuronal development. To test this hypothesis, we transfected primary cultured neurons with a plasmid encoding either WT or mutant KIF1A at DIV7 and fixed these cells at DIV10. The resulting neuronal branching was analyzed through Sholl analysis, and surprisingly, no difference in neuronal branching was found between the WT and mutant neurons ( Figures 5A, B ). We subsequently measured the number of primary and secondary neurites and found that mutant KIF1A did not affect the number of neurites ( Figures 5C, D ). In brief, mutant KIF1A does not affect neuronal development.

KIF1A is essential for synaptogenesis in the hippocampus (Kondo et al., 2012). In our study, the expression of mutant KIF1A resulted in a significant increase in the frequency of mEPSCs in neurons compared with that observed with WT KIF1A. The increased frequency of mEPSCs might be due to increased presynaptic vesicle release probability, and the increased number of excitatory synapses per neuron is also responsible for the observed increase in the frequency of mEPSCs. Dendritic spines contain the majority of the excitatory synapses of hippocampal pyramidal neurons, and changes in the spine density lead to alterations in the mEPSC frequency and are associated with aberrations in synaptic plasticity. To determine whether the mutant KIF1A affects the dendritic spines, we transfected neurons with a plasmid construct encoding the WT or mutant KIF1A at DIV7 and fixed these neurons at DIV16. Consistent with our hypothesis, the neurons transfected with mutant KIF1A exhibited a significantly higher density of dendritic spines and vGLUT compared with the neurons transfected with WT KIF1A ( Figures 6A–D ). To further verify the excitatory synapse formation, vGLUT-positive and PSD-95 clusters were examined using double immunofluorescence staining ( Figures 6E, F ), our results suggest that mutant KIF1A is responsible for the observed increase in the excitatory synaptic density.

We subsequently investigated the functional consequence of mutant kif1a in vivo by establishing a zebrafish model. Zebrafish have two kif1a homologue genes termed kif1aa and kif1ab that encode two protein isoforms, Kif1aa and Kif1ab, which show 85 and 80% identity to human KIF1A, respectively. The mutant alanine in the affected patients corresponds to A433 in Kif1aa and A410 in Kif1ab. Both of these genes have the same function in neurons, and kif1aa, which is also located on chromosome 2, and Kif1aa share higher (85%) similarity with human KIF1A. Therefore, we selected zebrafish kif1aa for our zebrafish experiment.

To establish the zebrafish model, we cloned mutant and WT kif1aa into the tol2 expression vector (WT kif1aa-Tol2 or A433D kif1aa-Tol2) and evaluated the expression of Kif1aa in zebrafish larvae using GFP. Mutant (A433D kif1aa-Tol2) or WT kif1aa (WT kif1aa-Tol2) was introduced into embryos at the one-cell stage by cytoplasmic microinjection ( Figure 7A ). Three days later, neither WT nor mutant zebrafish exhibited gross dysmorphologies ( Figure 7B ). Normal-looking zebrafish larvae displaying a touch response at 5 d.p.f. were selected for behavior monitoring. These zebrafish larvae were placed in a well of a 24-well falcon plate, and the motors of the freely swimming fish were observed using a stereomicroscope. The results showed that 67.2% of the WT larvae displayed normal behavior that could be characterized as S0 (baseline activity), and only 11.5% of the WT larvae exhibited abnormal behaviors that could be characterized S2 (large increase in movement). The analysis of the mutant larvae revealed that 40.6% displayed excessive motor activity that could be characterized as S2, and the percentage of larvae showing baseline activity (S0) was 39.0%, which was significantly lower than that found for the WT larvae ( Figure 7C ). To determine whether the missense mutation in Kif1aa resulted in excessive brain electrophysical activity, we obtained local field potential recordings from the zebrafish tectum at 5 d.p.f. and observed spontaneous epileptiform activity (polyspike discharges) in 42.2% of mutant larvae and 9.8% of WT larvae (this difference was significant) ( Figures 7D, E ).