Expression analysis of slc13a5 zebrafish paralogs (5a and 5b) and generation of zebrafish slc13a5 mutants using CRISPR/Cas9 technique.

(A-F) In situ hybridization for 5a and 5b expression at 25 hpf, 3 dpf and 5 dpf using antisense RNA probes. Insets show sections of 3 dpf and 5 dpf larvae post in situ hybridization for 5a and 5b expression. 5a and 5b are expressed in all the brain regions during early development and their expression appears to get slightly stronger in the MB at later stages. (G) Section of 3 dpf WTs; α-Slc13a5, α-HuC/D. Insets of dashed box in G showing example of a cell expressing Slc13a5, HuC/D and merged signal. MB neurons express Slc13a5 (Slc13a5⁺/HuC/D⁺). (H) Section of 3 dpf WTs; α-Slc13a5, α-Gfap. Insets of dashed box in H showing an example of a cell expressing Slc13a5, Gfap and merged signal. A small population of MB astrocytes express Slc13a5 (Slc13a5⁺/Gfap⁺). (I) Section of 3 dpf WTs; α-Slc13a5, α-vGlut1. Insets of dashed box in I showing example of a cell expressing Slc13a5, vGlut1 and merged signal. MB excitatory neurons express Slc13a5 (Slc13a5⁺/vGlut1⁺). (J) Section of 3 dpf WTs; α-Slc13a5, α-Gad67. Insets of dashed box in J showing example of a cell expressing Slc13a5, Gad67 and merged signal. MB inhibitory neurons express Slc13a5 (Slc13a5⁺/Gad67⁺). (K) Percentage of Slc13a5⁺ neurons (HuC/D⁺) and astrocytes (Gfap⁺) in the MB of 3 dpf WT. There are significantly more Slc13a5⁺/HuC/D⁺ cells compared to Slc13a5⁺/Gfap⁺ cells. n = 3 for each group. (L) Nucleotide sequences of 5a in zebrafish WT and 5a^−/−. Yellow highlighted regions indicate 11-nucleotide deletion due to mutation in 5a. (M) Nucleotide sequences of 5b in zebrafish WT and 5b^−/−. Yellow highlighted regions indicate one-nucleotide substitution followed by 13-nucleotide insertion due to mutation in 5b. (N) Amino acid sequence of 5a in human orthologue, zebrafish WT and 5a^−/−. Asterisks indicate conserved amino acids between human and zebrafish. Yellow highlighted regions indicate affected amino acids and premature stop due to mutation in 5a. (O) Amino acid sequence of 5b in human orthologue, zebrafish WT and 5b^−/−. Asterisks indicate conserved amino acids between human and zebrafish. Yellow highlighted regions indicate affected amino acids and premature stop due to mutation in 5b. (P-S) WT and slc13a5 mutant larvae at 5 dpf. No morphological differences were observed. WT, n = 5; 5a^−/−, n = 5; 5b^−/−, n = 5; 5a^−/−;5b^−/−, n = 5. (T-W) Quantification of FB, MB and HB size at 5 dpf by measuring brain width. MB size is significantly reduced in slc13a5 mutants. FB and HB size is unchanged. WT, n = 5; 5a^−/−, n = 6; 5b^−/−, n = 6; 5a^−/−;5b^−/−, n = 7 for FB size measurement. WT, n = 10; 5a^−/−, n = 8; 5b^−/−, n = 13; 5a^−/−;5b^−/−, n = 11 for MB size measurement. WT, n = 5; 5a^−/−, n = 6; 5b^−/−, n = 6; 5a^−/−;5b^−/−, n = 8 for HB size measurement. Colored lines with arrows on both ends show how the width measurements were performed. (X) Survival curve of WTs and slc13a5 mutants. Very few slc13a5 mutants survive until 30 dpf. WT, n = 78; 5a^−/−, n = 40; 5b^−/−, n = 28; 5a^−/−;5b^−/−, n = 92. Data are Mean ±  S.D., ns: no significant changes observed, **P ≤  0.01, ***P ≤  0.001, ****P ≤  0.0001- Unpaired t test. hpf, hours post fertilization; dpf, days post fertilization; WT, wild-type siblings; 5a, slc13a5a; 5b, slc13a5b; FB, forebrain; MB, midbrain; HB, hindbrain. The data underlying this figure can be found in S1 Data.

Analysis of startle response and circadian disturbances in slc13a5 mutants, along with their neuron population and neuronal apoptosis assessment.

(A) Schematic representation of acoustic startle protocol. (B) Quantification of distance traveled during baseline and acoustic stimulation periods at 5 dpf. slc13a5 mutants are hyperactive in light without any external stimuli and are more responsive to startle, with significantly higher distance traveled during baseline and acoustic stimulation periods compared to WT. WT, n = 21; 5a^−/−, n = 8; 5b^−/−, n = 8; 5a^−/−;5b^−/−, n = 24. ‘ns’ values are not shown on the graph due to space constraints. (C) Heat maps of locomotion of larvae at 5 dpf under acoustic startle, showing that slc13a5 mutants are more active than WTs (red shows highest presence and blue shows least presence). WT, n = 4; 5a^−/−, n = 4; 5b^−/−, n = 4; 5a^−/−;5b^−/−, n = 4. (D) Quantification of distance traveled in 10 hours of darkness in night at 4 dpf, 6 hours of light in the daytime at 5 dpf, 10 hours of darkness in night at 5 dpf and 6 hours of light in the daytime at 6 dpf. slc13a5 mutants are more active than WTs in light as well as in darkness. WT, n = 10; 5a^−/−, n = 10; 5b^−/−, n = 10; 5a^−/−;5b^−/−, n = 10 for each time point. (E) 5 dpf WTs and slc13a5 mutants; α-HuC/D (green). (F) Quantification of HuC/D⁺ cells in the optic tectum at 5 dpf. slc13a5 mutants show a reduction in neuron numbers compared to WTs. WT, n = 16; 5a^−/−, n = 11; 5b^−/−, n = 15; 5a^−/−;5b^−/−, n = 18. (G) Quantification of HuC/D⁺ cells in the optic tectum at 3 dpf. slc13a5 mutants show a reduction in neuron numbers compared to WTs. WT, n = 11; 5a^−/−, n = 11; 5b^−/−, n = 10; 5a^−/−;5b^−/−, n = 11. (H) Live 5 dpf slc13a5 mutants and WTs, stained with A.O. (green). (I) Quantification of A.O.⁺ cells in the optic tectum at 5 dpf. slc13a5 mutants show an increase in A.O.⁺ cell numbers compared to WTs. WT, n = 13; 5a^−/−, n = 17; 5b^−/−, n = 11; 5a^−/−;5b^−/−, n = 18. (J) Percentage of CC3⁺/HuC/D⁺ cells (out of total HuC/D⁺ cells) in the optic tectum at 5 dpf. slc13a5 mutants show an increase in CC3⁺ neuron population compared to WTs. WT, n = 8; 5a^−/−, n = 7; 5b^−/−, n = 8; 5a^−/−;5b^−/−, n = 10. Data are Mean ±  S.D., * P ≤  0.05, **P ≤  0.01, ***P ≤  0.001, ****P ≤  0.0001- Unpaired t test. A.O., acridine orange; CC3, cleaved-Caspase3. The data underlying this figure can be found in S1 Data.

E/I imbalance, fosab expression and brain hyperexcitability analysis in slc13a5 mutants.

(A-C) qPCR analysis for relative vglut2a, gad1b and fosab mRNA expression in the heads of 5 dpf slc13a5 mutants compared to WT. WT and slc13a5 mutants, n =  3 ×  10 larvae pooled. vglut2a is upregulated and gad1b is downregulated in slc13a5 mutants indicating dysfunctional E/I balance. fosab is upregulated in slc13a5 mutants indicating brain hyperexcitability. (D, E) 5 dpf WTs and slc13a5 mutants; α-HuC/D (green), c-Fos (red). Percentage of Fosab⁺/HuC/D⁺ cells (out of total HuC/D⁺ cells) in the optic tectum at 5 dpf. slc13a5 mutants show an increase in Fosab⁺ neuron population compared to WTs. WT, n = 6; 5a^−/−;5b^−/−, n = 7. (F-I) Representative extracellular recordings obtained from optic tectum of 6 dpf WTs and slc13a5 mutants, and a pie chart of the number of slc13a5 mutants showing different patterns of extracellular field potentials. The repetitive inter-ictal like discharges (<1s duration) with above threshold (>0.2mV), high-frequency, large-amplitude spikes seen in 7.4% 5a^−/−;5b^−/− larvae are indicative of increased network hyperexcitability. 18.5% 5a^−/−;5b^−/− larvae also displayed below threshold (<0.2mV) brain activity. WT, n = 8 out of 8 with no abnormal activity. (J-M) Representative extracellular recordings obtained using a NeuroProbe placed in the optic tectum of 5 dpf WTs and 5a^−/−;5b^−/− larvae, and a pie chart of number of 5a^−/−;5b^−/− larvae showing different patterns of neuronal activity. 91.7% WTs and 100% 5a^−/−;5b^−/− larvae intermittently twitched (as indicated by red arrows in J) but the deflections were distinguishable from epileptic activities. 16.7% 5a^−/−;5b^−/− larvae showed epileptic pattern consisting of a series of short bursts, with a total duration of> 10s (as shown by red arrows in K) and 8.3% 5a^−/−;5b^−/− larvae showed repetitive twitches (>5s) at consistent frequency of ~ 1.8 Hz (as shown by red arrows in L), another indication of seizures. WT, n = 12 out of 12 with no abnormal activity. Data are Mean ±  S.D., * P ≤  0.05, **P ≤  0.01- Unpaired t test. The data underlying this figure can be found in S1 Data.

Metabolic health analysis in slc13a5 mutants.

(A) Schematic representation of how the Seahorse bioanalyzer displays mitochondrial bioenergetics being regulated by pharmacological inhibitors. (B) Quantification of basal respiration at 6 dpf. slc13a5 mutants exhibit a significant reduction in basal respiration compared to WT. WT, n = 58; slc13a5 mutants, n = 20 (individual values plotted from five cycles). (C) Quantification of ATP-linked respiration at 6 dpf. slc13a5 mutants exhibit a significant reduction in ATP-linked respiration compared to WT. WT, n = 59; slc13a5 mutants, n = 20 (individual values plotted from three cycles). (D) Quantification of total mitochondrial respiration at 6 dpf. slc13a5 mutants exhibit a significant reduction in total mitochondrial respiration compared to WT. WT, n = 59; slc13a5 mutants, n = 20 (individual values plotted from three cycles). (E) Quantification of non-mitochondrial respiration at 6 dpf. slc13a5 mutants exhibit a significant reduction in non-mitochondrial respiration compared to WT. WT, n = 59; slc13a5 mutants, n = 20 (individual values plotted from three cycles). (F) Quantification of maximum respiratory capacity at 6 dpf. slc13a5 mutants exhibit a significant increase in maximum respiratory capacity compared to WT. WT, n = 59; slc13a5 mutants, n = 20 (individual values plotted from five cycles). (G) Quantification of reserve capacity at 6 dpf. slc13a5 mutants exhibit a significant increase in reserve capacity compared to WT. WT, n = 59; slc13a5 mutants, n = 20 (individual values plotted from five cycles). (H) Quantification of proton leaks at 6 dpf. This parameter was unchanged in slc13a5 mutants compared to WT. WT, n = 32; slc13a5 mutants, n = 10 (individual values plotted from five cycles). Data are Mean ±  S.D., ns: no significant changes observed, * P ≤  0.05, ****P ≤  0.0001- Unpaired t test. The data underlying this figure can be found in S1 Data.

Transcriptomics profiling of slc13a5 mutants.

(A-B) Bulk RNA sequencing was performed on pools of zebrafish heads of 5 dpf WT, 5a^−/−, 5b^−/− and 5a^−/−;5b^−/−. WT, 5a^−/−, 5b^−/− and 5a^−/−;5b^−/−, n =  3 ×  15 heads pooled. Heatmaps were constructed for prioritized DEGs using ComplexHeatmap. Genes upregulated in 5a^−/− and 5a^−/−;5b^−/− larvae compared to WTs (padj < 0.05 and log2FC >  1.5) are involved in apoptosis, inflammation and immune response and metabolism (A). Genes upregulated in 5b^−/− larvae compared to WTs (padj < 0.05 and log2FC >  1.5) are involved in apoptosis and inflammation and immune response (A). Genes downregulated in 5a^−/− and 5a^−/−;5b^−/− larvae compared to WTs (padj < 0.05 and log2FC < −1.5) are involved in epilepsy and other neurological disorders, calcium ion transport, sleep regulation and negative regulation of apoptosis (B). Genes downregulated in 5b^−/− larvae compared to WTs (padj < 0.05 and log2FC < −1.5) are involved in neurological disorders and negatively regulate apoptosis (B). Heatmaps with a full list of DEGs are provided in S7–S10 Figs. DEGs, differentially expressed genes. The data underlying this figure can be found in S2 Data.

Assessment of calcium events, pERK levels, NMDA receptor expression and extracellular zinc levels in slc13a5 mutants.

(A-C) Quantification of the amplitude and frequency of calcium events (ΔF/F > 2) and representative single neuron calcium traces in the MB at 3 dpf. 5a^−/−;5b^−/− larvae show a significant increase in calcium events compared to WTs. WT, n = 5; 5a^−/−;5b^−/−, n = 6. (D) Whole-brain activity MAP-map depicting significant changes in pERK signal, calculated using Mann-Whitney U statistic Z score. The significance threshold was set based on an FDR whereby 0.05% of control pixels is set as significant. Voxels exhibiting significantly higher intensity values of pERK are denoted in green and those exhibiting significantly lower pERK intensity values are depicted in magenta, in the slc13a5 mutants compared to WTs at 5 dpf. slc13a5 mutants show an enhanced neural activity via increased pERK levels in different regions of the brain (green) compared to WT. The lower pERK intensity shown in the eyes (magenta) could be a background autofluorescence captured while imaging. WT, n = 9; 5a^−/−;5b^−/−, n = 8. (E) Section of 3 dpf MB region of WTs; α-Slc13a5, α-Grin1. (E’, E”, E’’’) Higher magnification of dashed boxes in E. White arrowheads point to Slc13a5⁺/Grin1⁺ cells in MB. (F, F’, F”, F’’’) Slc13a5 expression shown in the brain section from E. (G, G’, G”, G’’’) Grin1 expression shown in the brain section for E. (H, I) 5 dpf WTs and 5a^−/−;5b^−/− larvae; Grin1 (green), NeuN (red). (J, K) Quantification of the fluorescence intensity (CTCF) of Grin1 (α-Grin1) in the neurons (NeuN⁺) of optic tectum. 5a^−/−;5b^−/− larvae show a significant increase in Grin1 expression in the neurons of optic tectum compared to WTs at 3 dpf and 5 dpf. WT, n = 4; 5a^−/−;5b^−/−, n = 4 at 3 dpf. WT, n = 4; 5a^−/−;5b^−/−, n = 6 at 5 dpf. (L) 50 hpf larvae stained with CellMask Orange and Palm-ZP1. (M, M’, N, N’, Q) Quantification of the fluorescence intensity (CTCF) ratios of Palm-ZP1:CellMask Orange in the plasma membrane. 5a^−/−;5b^−/− larvae show a significant decrease in Palm-ZP1:CellMask Orange CTCF ratio in the plasma membrane compared to WTs. Untreated WT, n = 4; Untreated 5a^−/−;5b^−/−, n = 4. White arrows indicate the plasma membrane regions used for CTCF measurements. (O, O’, P, P’, Q) Quantification of the fluorescence intensity (CTCF) ratios of Palm-ZP1:CellMask Orange in the plasma membrane. ZnCl₂-treated WTs show a significant increase in Palm-ZP1:CellMask Orange CTCF ratio in the plasma membrane compared to untreated WTs. Untreated WT, n = 4; ZnCl₂-treated WT, n = 5. Similarly, ZnCl₂-treated 5a^−/−;5b^−/− larvae show a significant increase in Palm-ZP1:CellMask Orange CTCF ratio in the plasma membrane compared to untreated 5a^−/−;5b^−/− larvae. Untreated 5a^−/−;5b^−/−, n = 4; ZnCl₂-treated 5a^−/−;5b^−/−, n = 4. White arrows indicate the plasma membrane regions used for CTCF measurements. (R) Untreated WTs (i.e., in E3 water) and DMSO-treated WTs (vehicle for Palm-ZP1) showed no significant changes in Palm-ZP1:CellMask Orange CTCF ratios in the plasma membrane. Untreated WTs, n = 4; DMSO-treated WTs, n = 3. Data are Mean ±  S.E.M., ns: no significant changes observed, * P ≤  0.05, **P ≤  0.01, ***P ≤  0.001- Unpaired t test. FDR, false discovery rate. CTCF, corrected total cell fluorescence. The data underlying this figure can be found in S1 Data.

Effect of NMDA receptor signaling suppression and zinc treatment on slc13a5 mutants.

(A-C) Quantification of the amplitude and frequency of calcium events (ΔF/F > 2) and representative single neuron calcium traces in the MB at 3 dpf. Memantine treatment and grin1a;grin1b MO injections significantly reduced the amplitude and frequency of calcium events in 5a^−/−;5b^−/− larvae compared to vehicle-treated or control MO-injected 5a^−/−;5b^−/− larvae. Vehicle-treated/control MO-injected WT larvae, n = 5; vehicle-treated/control MO-injected 5a^−/−;5b^−/− larvae, n = 5; memantine-treated 5a^−/−;5b^−/− larvae, n = 6; grin1a;grin1b MO-injected 5a^−/−;5b^−/− larvae, n = 6. (D-F) Quantification of the amplitude and frequency of calcium events (ΔF/F > 2) and representative single neuron calcium traces in the MB at 3 dpf. ZnCl₂ treatment significantly reduced the amplitude and frequency of calcium events in 5a^−/−;5b^−/− larvae compared to vehicle-treated 5a^−/−;5b^−/− larvae. Vehicle-treated WT larvae, n = 4; vehicle-treated 5a^−/−;5b^−/− larvae, n = 9; ZnCl₂-treated 5a^−/−;5b^−/− larvae, n = 6. (G) Quantification of basal respiration before and after treatment with memantine and MK801 at 6 dpf. Memantine and MK801 treatments rescue compromised basal respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 6; vehicle-treated 5a^−/−;5b^−/−, n = 6; memantine-treated 5a^−/−;5b^−/−, n = 12; MK801-treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from five cycles). (H) Quantification of ATP-linked respiration before and after treatment with memantine and MK801 at 6 dpf. Memantine and MK801 treatments rescue compromised ATP-linked respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 5; vehicle-treated 5a^−/−;5b^−/−, n = 9; memantine-treated 5a^−/−;5b^−/−, n = 10; MK801-treated 5a^−/−;5b^−/−, n = 5 (individual values plotted from three cycles). (I) Quantification of total mitochondrial respiration before and after treatment with memantine and MK801 at 6 dpf. Memantine and MK801 treatments rescue compromised total mitochondrial respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 10; vehicle-treated 5a^−/−;5b^−/−, n = 10; memantine-treated 5a^−/−;5b^−/−, n = 6; MK801-treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from three cycles). (J) Quantification of non-mitochondrial respiration before and after treatment with memantine and MK801 at 6 dpf. Memantine and MK801 treatments rescue compromised non-mitochondrial respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 8; vehicle-treated 5a^−/−;5b^−/−, n = 5; memantine-treated 5a^−/−;5b^−/−, n = 6; MK801-treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from three cycles). (K) Quantification of basal respiration before and after treatment with ZnCl₂ at 6 dpf. ZnCl₂ treatment rescues compromised basal respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 5; vehicle-treated 5a^−/−;5b^−/−, n = 4; ZnCl₂-treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from five cycles). (L) Quantification of ATP-linked respiration before and after treatment with ZnCl₂ at 6 dpf. ZnCl₂ treatment rescues compromised ATP-linked respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 5; vehicle-treated 5a^−/−;5b^−/−, n = 4; ZnCl₂-treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from three cycles). (M) Quantification of total mitochondrial respiration before and after treatment with ZnCl₂ at 6 dpf. ZnCl₂ treatment rescues compromised total mitochondrial respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 5; vehicle-treated 5a^−/−;5b^−/−, n = 4; ZnCl₂ -treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from three cycles). (N) Quantification of non-mitochondrial respiration before and after treatment with ZnCl₂ at 6 dpf. ZnCl₂ treatment rescues compromised non-mitochondrial respiration in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 5; vehicle-treated 5a^−/−;5b^−/−, n = 4; ZnCl₂ -treated 5a^−/−;5b^−/−, n = 6 (individual values plotted from three cycles). (O) Quantification of total distance traveled in 10 min in acoustic startle before and after treatment with memantine at 5 dpf. Memantine treatment rescues impaired startle response in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 16; vehicle-treated 5a^−/−;5b^−/−, n = 58; memantine-treated 5a^−/−;5b^−/−, n = 43. (P) Quantification of total distance traveled in 10 min in acoustic startle before and after treatment with ZnCl₂ at 5 dpf. ZnCl₂ treatment rescues impaired startle response in 5a^−/−;5b^−/− larvae. Vehicle-treated WT, n = 9; vehicle-treated 5a^−/−;5b^−/−, n = 17; memantine-treated 5a^−/−;5b^−/−, n = 10. Data are Mean ±  S.E.M. and Mean ±  S.D., * P ≤  0.05, **P ≤  0.01, ^***P ≤  0.001, ****P ≤  0.0001- Unpaired t test. MO, morpholino. The data underlying this figure can be found in S1 Data.

Model of relationship between slc13a5 mutations, NMDA receptor signaling and zinc chelation.

Unlike in WT larvae, slc13a5 mutations in zebrafish presumably cause an increase in extracellular citrate levels that lead to zinc chelation, thereby preventing its inhibitory effect on NMDA receptors and ultimately leading to an excessive influx of calcium to the neurons, causing epilepsy. Blocking NMDA receptors and/or treatment with zinc prevents this enhanced influx of calcium into the neurons, thereby rescuing epilepsy-related phenotypes in the slc13a5 mutants.