Epilepsy, Behavioral Abnormalities, and Physiological Comorbidities in Syntaxin-Binding Protein 1 (STXBP1) Mutant Zebrafish

Mutations in the synaptic machinery gene syntaxin-binding protein 1, STXBP1 (also known as MUNC18-1), are linked to childhood epilepsies and other neurodevelopmental disorders. Zebrafish STXBP1 homologs (stxbp1a and stxbp1b) have highly conserved sequence and are prominently expressed in the larval zebrafish brain. To understand the functions of stxbp1a and stxbp1b, we generated loss-of-function mutations using CRISPR/Cas9 gene editing and studied brain electrical activity, behavior, development, heart physiology, metabolism, and survival in larval zebrafish. Homozygous stxbp1a mutants exhibited a profound lack of movement, low electrical brain activity, low heart rate, decreased glucose and mitochondrial metabolism, and early fatality compared to controls. On the other hand, homozygous stxbp1b mutants had spontaneous electrographic seizures, and reduced locomotor activity response to a movement-inducing “dark-flash” visual stimulus, despite showing normal metabolism, heart rate, survival, and baseline locomotor activity. Our findings in these newly generated mutant lines of zebrafish suggest that zebrafish recapitulate clinical phenotypes associated with human syntaxin-binding protein 1 mutations.


Introduction
Mutations in the human syntaxin-binding protein 1 (STXBP1) gene are associated with a range of clinical outcomes. STXBP1 mutations were first identified in children with early infantile epileptic encephalopathy with burst suppression (EIEE; also known as Ohtahara Syndrome) [1,2] and subsequently found in patients diagnosed with other forms of epileptic encephalopathy including infantile spasms [3,4], Lennox-Gastaut Syndrome [4], and Dravet Syndrome [5]. These epilepsies are primarily pediatric, catastrophic, pharmacoresistant, and associated with intellectual disability. Additionally, STXBP1 mutations are sometimes associated with non-syndromic intellectual disability without epilepsy [6], or with ataxia or dyskinesia [7] that can 14 for stxbp1a, and a 800-bp region spanning from exon 13 to 3'UTR in exon 20 for stxbp1b. The T3 promoter sequence (aattaaccctcactaaaggg) was added to either the reverse primer (for the antisense probe) or the forward primer (for the sense control). Primer Sequences: stxbp1a-Fprobe: CTGACGCCTTCCAGAGTTTC, stxbp1a-Rprobe: CCTCAGCATCTGTTCCCATT; stxbp1b-Fprobe; AGTACCAGGGCACAGTGGAC, stxbp1b-Rprobe; AACAGTGAGGTGGGGCTAT G. Digoxigenin-labeled riboprobes for stxbp1a were transcribed in vitro using standard reagents including DIG RNA labeling mixes (Roche Molecular Biochemicals), Fermentas T3 RNA Polymerase, RNAse out, and were then treated with DNAse and purified using GE Healthcare illustra ProbeQuant G-50 Micro Columns.
ISH was performed following standard protocols. Larvae were fixed in 4% paraformaldehyde (pH 7.0) in phosphate-buffered saline (PBS) at 4°C overnight. Following fixation, larvae were rinsed 4X for 5 minutes in PBS with 0.1% Tween 20 (PBSTw). Larvae were then rinsed with 100% methanol for 5 minutes, transferred to fresh methanol, and subsequently stored in 100% methanol at -20°C until needed. Before experimentation, brains and larvae were rehydrated in PBS, and brains were dissected using forceps. Brains were then processed and stained in wells of a 24-well plate in incubation baskets (Intavis AG). Brains were rehydrated through a series of 5-minute washes in 75%, 50%, 25% methanol in PBS with 0.1% Tween (PBS/Tw). Brains were then washed twice for 5 minutes in PBS/Tw, treated with proteinase K for 5 to 10 minutes, and rinsed 4 times with PBS/Tw. Brains were fixed again with 4% PFA for 10 minutes at room temperature, washes 4 X5 min in PBST. Prehybridization (1 hour) and hybridization (overnight) were carried out at 58°C, in a hybridization buffer containing 50% Formamide, 5X SSC buffer, 5mg/mL yeast torula RNA, and 0.1% Tween. Posthybridization washes were carried out at 58°C: twice for 30 minutes in 50% Formamide, 2X SSC w/Tw, once for 15 minutes in 2X SSC w/Tw, and twice for 30 minutes in 0.2X SSC w/Tw. Tissues were then blocked for 1 hr in blocking buffer containing 100mM maleic acid, pH 7.2, 150mM NaCl, bovine serum albumen (BSA), sheep serum, and 0.1% Tween 20. Alkaline-phosphatase-conjugated anti-DIG Fab fragments (Roche) were then added to the blocking solution to a final dilution of 1:4000. Brains were incubated overnight at 4°C, then washed 4 times for at least 10 minutes each in Maleate Buffer and signal was revealed with NBT/BCIP (Roche), kept in the dark, at room temperature or 37°C. To stop the staining reaction, tissue was washed 5 times for 5 minutes in PBS w/Tw.
To identify the expression patterns of genes, embryos and 7 dpf larvae brains were visualized on a Nikon Eclipse Ni microscope and photographs taken with a color digital camera (Nikon DS-U3) controlled by Nikon NIS-Elements software (version 4.00.07). Stained sections were viewed in brightfield. For images of stained 7 dpf larvae brains, Nikon Elements software was used to create focused images from stacks of multiple images.

Quantitative real-time PCR (qPCR)
Gene expression levels of stxbp1a and stxbp1b mRNA were examined using RNA pooled from 10 wild-type sibling larvae. Total RNA was extracted using Trizol 1 Reagent (Invitrogen, Carlsbad, CA), treated with DNAse (Ambion/Applied Biosystems, Austin, TX) and quantified with NanoDrop™ ND-1000 spectrophotometer (Thermo Scientific). Reverse-transcription reactions were performed using SuperScript™III First-Strand Synthesis System (Invitrogen) with a mix of oligo(dT) 20 and random hexamers. The cDNA templates were diluted 1:2 with DEPC sterile water. The qPCR reactions were performed using SybrGreen 1 fluorescent master mix on an ABI Prism 1 7700 Sequence Detection System using ABI PRISM SDS v9.1 software (Applied Biosystems). Zebrafish stxbp1a cDNA was amplified using the following primers, which were designed using Primer Express v3.0 (Applied Biosystems): Stxbp1a_qF1: TTGCTGGATGCCAAT GTCA; Stxbp1a_qR1: TCCGTGATCCCGTTCTTGAG. The following primers were used to amplify β-actin cDNA (GenBank Accession # FJ915059) for data normalization: β-actin-F: CATCCATCGTTCACAGGAAGTG; β-actin-R: TGGTCGTTCGTTTGAATCTCAT. Samples were run in triplicate and reactions contained 1× SYBR green master mix, 10 μM of each primer, and RNAse free water for a final volume of 10 μl. Samples without reverse transcriptase were run for each reaction as negative controls. Cycling parameters were as follows: 50°C × 2min, 95°-C × 10min, then 40 cycles of the following 95°C × 15s, 60°C × 1min. For each sample a dissociation step was performed at 95°C × 15s, 60°C × 20s, and 95°C × 15s at the end of the amplification phase to check for the presence of primer dimers or non-specific products. For both genes, qPCR efficiencies were assessed by mean of 4-fold serial dilutions of pooled cDNA. Serially diluted cDNAs were used to construct standard curves and estimates of efficiencies, slope of the curves and the correlation coefficient. Triplicate quantification values (CT; cycle threshold) were analyzed using qCalculator software (programmed by Ralf Gilsbach) which estimates qPCR efficiency E = 10^(-1/slope) and the relative gene expression between samples after normalization basing on both the Comparative ΔΔCT [36] and the Efficiency Based [37] methods.

Establishing mutant lines
CRISPR/Cas9 mutations were generated in wild-type (TL strain) zebrafish using published techniques [38,39]. Sequence-specific sgRNA template plasmids were generated for each target site by modifying DR274 (Addgene Plasmid #42250). Plasmid Dr274 was modified to contain gene-specific sequences selected using ZiFit software [40]. In order to avoid off-target genomic mutagenesis effects, which can occur at sites that closely resemble the target site, we selected target sites that have a minimum of three mismatches with every other site in the genome.
Fertilized 1-2 cell stage zebrafish eggs were injected with an injection mix containing approximately 300ng/μl Cas9 mRNA and 15 ng/μl sgRNA. After injected eggs were incubated for one day, some were harvested to check for mutagenesis at the target site. DNA was extracted from pools of 10 injected embryos and uninjected controls, gDNA including the target site was amplified, and the target site was checked via Sanger sequencing. Multiple sequencing peaks were confirmed to be present at the sgRNA target site before proceeding.
Other Cas9/sgRNA-injected embryos were raised to adulthood. These F0-generation potential mutants were crossed, and DNA was extracted from pooled F1 embryos for PCR and sequencing, as performed on the injected embryos. To obtain stable lines with known mutations, F1 embryos were raised to adulthood and outcrossed to wild-type (WIK strain) zebrafish. For adult fish of each line, genomic DNA was extracted from tail tissue, amplified by PCR, cloned into TOPO pcr2.1 vector, and sequenced. Mutations were identified and multiple individuals from each line were sequenced to confirm that each individual carried the same mutation.
The newly generated mutant alleles are designated stxbp1a s3000 and stxbp1b s3001 using the University of California San Francisco (UCSF) "s" designation in accordance with the Zebrafish Information Network (ZFIN) guidelines.

Genotyping
For genotyping, we extracted genomic DNA from whole larvae using the Zebrafish Quick Genotyping DNA Preparation Kit (Bioland Scientific). We amplified stxbp1a gDNA using the following primers: stxbp1a-F: CACACACTTACAGCAGGAATGAGTGG, stxbp1a-R: ATTCAGAC TTCAACTGTACATGTATTGTG. These primers amplify a 275-bp region including the stxbp1a mutation site. The mutant allele was then detected by digesting the amplicon with BsaHI, for which the restriction site is absent in the mutant, and electrophoresis to separate the digested samples on a 1% agarose gel.
We amplified stxbp1b gDNA using the following primers stxbp1b-F: ATCTGCGTAGAAAG CTGAGCTTCATAG, stxbp1b-R: GTCAATGAAAATGGCACTAACTCCACACG. The mutant allele was then detected by digesting the amplicon with BsiHKAI, for which the restriction site is absent in the mutant, and electrophoresis to separate the digested samples on a 1% agarose gel.

Morphological Phenotyping
Larvae were photographed using a SteREO Discovery.V8 microscope (Zeiss). Standard lengths (distance from the anterior tip of head to the base of the caudal fin) were measured manually using DanioScope software (Noldus, version 1.0.109).

Behavioral Phenotyping
For locomotion tracking, single zebrafish larvae were placed in individual wells of a 96-well flat-bottomed Falcon culture dish (BD Biosciences). Each well contained approximately 200 μl of embryo medium. Behavior was monitored at room temperature (21-22°C) using a DanioVision system and EthoVision XT 8.0 locomotion tracking software (Noldus Information). Prior to conducting experiments, larvae still in their chorions at day 3 (homozygous mutants) had their chorions removed using forceps. 5 dpf larvae were allowed to acclimate to the tracking arena for 3 to 4 hours, and then 24 hours of continuous behavioral data were recorded beginning at 4:00PM. The light:dark cycle continued as usual: lights-off occurred at 11:00PM and lights-on at 9:00AM. At 4:00PM, when 24 hour recordings completed, larvae responses to sudden changes in light intensity were tested. 6 dpf larvae were exposed to 10 seconds of darkness (0% intensity light) followed by 10 seconds of 100% intensity light.

Heart Rate Measurements
Petri dishes containing 3 dpf larvae were removed from the incubator and allowed at least 10 minutes to acclimate to the light and temperature on the microscope stage. 30-second videos of individual unanesthetized fish were taken using 6.3x zoom magnification with a SteREO Discovery.V8 microscope (Zeiss) and an Axiocam ICm1 camera with ZEN 2012 software (Zeiss, version 1.1.2.0) at a frame rate of 26 fps. Videos were imported into DanioScope software (Noldus, version 1.0.109) as a compressed AVI file, where a subset of the beating heart was outlined and heart rate was calculated.

Metabolic Measurements
Extracellular acidification rate and oxygen consumption rate, representing glycolysis and oxidative phosphorylation respectively, were measured simultaneously using an extracellular flux analyzer (XF24, Seahorse Biosciences). Prior to metabolic experiments, chorions were removed from homozygous mutant larvae using forceps. Single 5 dpf zebrafish larvae (n = 10 per group) were placed in individual wells of a 24-well islet plate and a fine screen mesh was placed to maintain the larvae in place. Rates were calculated as the mean of 3 measurements, each lasting 3 minutes, which were taken with 5 minutes between measurements.

Survival
Larval zebrafish were maintained as described above and counted daily. After day 12 they were collected and genotyped. Data were plotted using Microsoft Excel.

Electrophysiology
Electrophysiological recordings were carried out according to previously described methods [41,42]. To obtain stable physiological recordings, zebrafish larvae were briefly anesthetized in tricaine, paralyzed in α-bungarotoxin and immobilized in 1.2% low-melting temperature agarose in zebrafish egg water. Larvae were embedded so that the dorsal aspect of the brain was accessible for electrode placement. Embedded larvae were bathed in egg water and visualized using an Olympus BX50 microscope (Olympus America Inc., Center Valley, PA). Under direct visual guidance, a glass microelectrode (b1.2 mm tip diameter, 2-7 MO) was placed in the forebrain along the midline where most neuronal somas are located. Electrodes were filled with 2 M NaCl and electrical activity was recorded using a Patch Clamp PC-505B amplifier (Warner Instruments, Hamden, CT). Voltage records were low-pass filtered at 2 kHz (−3 dB, 8-pole Bessel), Notch filtered at 60 Hz, digitized at 10 kHz using a Digidata 1300 A/D interface, and stored on a PC computer running Axoscope software (Molecular Devices, Sunnyvale, CA). Electrophysiological recordings were coded and analyzed post hoc using Clampfit software (Molecular Devices).

Gene Expression
We investigated the spatial expression of stxbp1a and stxbp1b in developing zebrafish using whole-mount colorimetric in situ hybridization (ISH). At 2 and 3 days post-fertilization (dpf), stxbp1a expression was prominent throughout the anterior-posterior axis in all major central nervous system (CNS) structures including telencephalon, optic tectum (TeO), cerebellum (CeP), medulla oblongata (MO) and spinal cord (Fig 1A and 1B). In contrast, the distribution of stxbp1b at 2 dpf and 3 dpf was more restricted with expression limited to the olfactory bulb (OB), the right habenula (Ha), and the outer regions of the retina (Fig 1F and 1G). In isolated 7 dpf brain, both stxbp1a and stxbp1b were also prominent in telencephalon, TeO, CeP, and MO ( Fig 1C and 1H); isolated 7 dpf whole brains probed with sense control probes showed no staining (Fig 1D and 1I).
The temporal expression of stxbp1a and stxbp1b were investigated in developing zebrafish using quantitative real-time PCR. In RNA samples extracted from pooled wild-type larvae, both genes were expressed at 12 hours post-fertilization (hpf). By 1 dpf, stxbp1a RNA levels relative to β-actin increased 5-fold and continued a steady increase through 7 dpf (Fig 1E). In contrast, stxbp1b RNA levels increased less than 2-fold by 1 dpf but continued increasing rapidly until 2 dpf and beyond (Fig 1J).

CRISPR/Cas9-Generated stxbp1a and stxbp1b Mutations
The zebrafish Stxbp1a amino acid sequence (including stop codon) shares 87.2% pairwise identity with human STXBP1 sequence (Fig 2). Zebrafish Stxbp1b amino acid sequence is 78.5% identical to human STXBP1 (Fig 2). The two zebrafish paralogs, Stxbp1a and Stxbp1b, share 74.9% pairwise amino acid sequence identity. By injecting fertilized wild-type zebrafish eggs with a mixture of Cas9 mRNA and gene-specific short guide RNA (sgRNA) we targeted germline mutations in both stxbp1a and stxbp1b. We identified F1 zebrafish (offspring of the CRISPR/Cas9-injected fish) with mutations in stxbp1a ( Fig 2B) and outcrossed offspring of one F1 fish to generate a stable line. F2 offspring were sequenced to confirm that they shared an identical stxbp1a mutation. Cloning and sequencing revealed that the mutant allele is a 4 base-pair deletion (Zv9 Chromosome 21: 12,099,014-12,099,017). This deletion causes a frameshift in exon 8 of the primary stxbp1a transcript (Ensembl: ENSDART00000015629; GenBank: NM_001025182), and is predicted to cause a premature stop codon. This novel mutant was given the line designation stxbp1a s3000 .
Similarly, we identified F1 zebrafish (offspring of the CRISPR/Cas9-injected fish) with mutations in stxbp1b ( Fig 2C) and outcrossed offspring of one F1 fish to generate a stable line. F2 offspring were sequenced to confirm that they shared an identical stxbp1b mutation. Cloning and sequencing revealed that the mutant allele is a 12 base-pair deletion (Zv9 Chromosome 5: 30,611,750-30,611,761). This deletion causes a loss of the predicted start codon and second codon of the stxbp1b transcript (Ensembl: ENSDART00000005638; GenBank: NM_001089376), along with 6 upstream bases. This novel mutant was given the line designation stxbp1b s3001 .

Morphology
Homozygous stxbp1a s3000/s3000 mutant zebrafish have abnormal morphology (Fig 3). They fail to hatch out of the chorion by 5 dpf, yielding a laterally curved body shape (Fig 3A). When they are manually released from the chorion and allowed to develop, homozygous stxbp1a s3000/s3000 mutant larvae (n = 10) are not significantly different in length from their control wild-type or heterozygous siblings (n = 30, p = 0.0592, two-tailed t-test; Fig 3B). The stxbp1a s3000/s3000 mutant larvae are also darkly pigmented, with dispersed melanin on their heads and backs. Furthermore, stxbp1a s3000/s3000 mutants exhibit abnormal craniofacial development, with the anterior part of the head foreshortened (Fig 3C and 3D).
Conversely, homozygous stxbp1b s3001/s3001 mutant zebrafish have grossly normal morphology (Fig 4A). Homozygous stxbp1b s3001/s3001 mutant larvae (n = 10) are not significantly different in length from their control siblings (n = 30, p = 0.0592, two-tailed t-test; Fig 4B). Heterozygous mutant siblings and wild-type siblings were not significantly different from each other in length. The stxbp1b s3001/s3001 mutant larvae are also darkly pigmented, with dispersed melanin on their heads and backs (Fig 4C and 4D).

Behavior
To examine diurnal locomotor activity and responses to a known movement-inducing visual stimulus, we quantified movement of individual zebrafish larvae between 5 and 6 dpf. Neither heterozygous nor homozygous stxbp1a mutants displayed overt behavioral hallmarks of seizures, such as hyperactivity or whole-body convulsions (n = 23 WT, 45 Het, 21 Homo; Fig 5A). In normal embryo media, homozygous stxbp1a s3000/s3000 5 dpf mutant zebrafish did not move during the initial 10-minute recording window (Fig 5A and 5C). During the subsequent 24-hour light-dark period, several homozygous mutants moved small distances (Fig 5B), but at 6 dpf none moved in response to the sudden onset of darkness for 10 seconds (a "dark flash") ( Fig 5D).
Heterozygous stxbp1a s3000/+ 5 dpf larvae movement was not different from wild-type siblings during the initial 10-minute recording window or during most of the subsequent 24-hour light-dark period (Fig 5A-5C). In response to sudden darkness, however, heterozygous mutants exhibited significantly less movement than wild-type siblings (p = 0.042, two-tailed ttest; Fig 5D).

Discussion
Here we report the generation of stxbp1a and stxbp1b mutant zebrafish to explore the effects of STXBP1 mutations in neurodevelopmental disorders including early infantile epileptic encephalopathy with burst suppression (EIEE). Homozygous mutation of stxbp1a led to severe physiological and behavioral deficits including immobility, reduced heart rate and metabolism, hyperpigmentation, and early death, while heterozygous mutation of stxbp1a led to generally normal behavior with only a slight reduction in response to a startling visual stimulus. Strikingly, homozygous stxbp1b mutants showed pronounced epileptic seizures along with mild hyperpigmentation, with normal mobility, heart rate, metabolism, and gross morphology.

Expression
The more severe neurodevelopmental phenotypes correlated with mutation of stxbp1a may be related to its higher conservation and broader early CNS expression compared to stxbp1b. Due to an ancient whole-genome duplication in the ancestor of all teleost fishes, zebrafish have two copies of many single-copy human genes [43][44][45][46]. Following the teleost-specific genome duplication, the duplicated gene pairs (ohnologs) experienced a range of outcomes including nonfunctionalization, sub-functionalization, and neo-functionalization [47,48]. Sub-functionalization may happen via protein changes [49] or via regulatory element loss [50] as described by the "duplication-degeneration-complementation" model, in which ancestral expression domains are differentially lost in different genes [51]. For example, recent evidence suggests that duplication of a broadly expressed corticotropin-releasing hormone (CRH) gene led to one onholog becoming restricted to only a sub-region of the zebrafish hypothalamus [52,53]. On the other hand, some ohnologs retain similar expression domains and overlapping functions. For example, zebrafish have two similar homologs of the human SCN1A gene, scn1laa and scn1lab; both genes are expressed in the brain and homozygous loss-of-function mutation Metabolic and survival deficits caused by stxbp1a s3000 mutation. Homozygous stxbp1a s3000/s3000 mutant larvae have lower metabolism than controls (siblings) and die prematurely. (A) 5 dpf stxbp1a s3000/s3000 homozygous mutants (n = 10) had significantly lower extracellular acidification (ECAR) than siblings (n = 10; mean ± SD). * = p < 0.0001 (two-tailed t-test). (B) 5 dpf stxbp1a s3000/s3000 homozygous mutants (n = 10) had significantly lower oxygen consumption (OCR) than siblings (n = 10; mean ± SD). * = p < 0.0001 (two-tailed t-test). (C) Heart rates of stxbp1a s3000/s3000 mutant larvae (n = 15) at 3 dpf were significantly lower than heart rates of their siblings (n = 33). All measured values are plotted; mean values are indicated by horizontal lines. (D) Homozygous stxbp1a s3000/s3000 mutants die as larvae. Homozygous mutants (n = 50) and siblings (n = 50) were maintained in petri dishes without food and counted each day from 2 dpf until 10dpf. The stxbp1a s3000/s3000 mutants began dying at 6 dpf, and 98% (49) died by 10dpf. Only 2% (1) of siblings died at 10dpf.
doi:10.1371/journal.pone.0151148.g008 in scn1lab results in epilepsy and serves as a model for Dravet syndrome [34,54,55]. Understanding the parallel evolutionary histories of a given gene family in humans and zebrafish can facilitate the development of appropriate zebrafish models of human diseases.
In the case of STXBP1, early expression of stxbp1a and stxbp1b exhibit very different patterns, even though both are broadly expressed in the brain by 7 dpf. There may also be important sub-functionalization later in development due to distinct expression patterns of stxbp1a and stxbp1b at the cellular level. The restricted expression of stxbp1b in the olfactory bulb and habenula suggests important roles in the development of these regions or in early-developing behaviors that require the function of circuits that include these regions. Interestingly, the right dorsal habenula is activated by olfactory stimuli, suggesting that expression of stxbp1b in both olfactory bulb and right habenula may delineate parts of a developing neural circuit [56].

Morphology
Like Stxbp1 homozygous mutant mice, stxbp1a s3000/s3000 and stxbp1b s3001/s3001 larvae are normal in size [12]. The dark pigmentation of both stxbp1a s3000/s3000 larvae and stxbp1b s3001/s3001 larvae likely reflect defects in the pathway that controls pigment granule aggregation and allows adaptive matching of pigmented area to environmental light conditions. Zebrafish retina and Normal metabolic function in stxbp1b s3001 mutants. Homozygous stxbp1b s30010/s3001 mutant larvae have unaffected metabolism compared to controls (siblings). (A) 5 dpf stxbp1b s30010/s3001 homozygous mutants (n = 11) did not differ from siblings (n = 52) in extracellular acidification (ECAR) (twotailed t-test; mean ± SD). (B) 5 dpf stxbp1b s30010/s3001 homozygous mutants (n = 11) did not differ from siblings (n = 52) in oxygen consumption (OCR) (twotailed t-test; mean ± SD). (C) Heart rates of stxbp1b s30010/s3001 mutant larvae (n = 16) at 3 dpf were not significantly different from heart rates of their siblings (n = 34). All measured values are plotted; mean values are indicated by horizontal lines. (D) All stxbp1b s30010/s3001 mutant larvae tested (n = 8) survived until 10 dpf; 12.5% of their control siblings (n = 48) died by 10dpf. pineal gland relay information to the hypothalamus that causes release of hormones which then control dispersion or aggregation of skin melanosomes [57][58][59][60] Dispersion of melanosomes in both stxbp1a s3000/s3000 larvae and stxbp1b s3001/s3001 larvae thus suggests that visual or neuroendocrine processes are disrupted even in the more behaviorally and metabolically normal stxbp1b s3001/s3001 mutants. Dispersed melanin is also seen in other mutants with impaired visual functions [61,62], and in the epileptic scn1lab s552 homozygous mutant larvae [34,55]. Although Stxbp1 is necessary for normal release of hormones from mouse neuroendocrine cells [63], specific neuroendocrine defects have not been documented in children harboring mutations in STXBP1.
During normal zebrafish postembryonic development, the head transitions from a more square shape with a ventrally located mouth to a more triangular shape with a more anterior and dorsally positioned mouth [64]. The shortened head phenotype of stxbp1a s3000/s3000 larvae likely reflects a problem in early neurodevelopment and may be the result of loss of neurons or Forebrain field potential recordings from homozygous stxbp1a s3000/s3000 and stxbp1b s3001/s3001 mutant larvae compared to a control recording from an electrode in 1.2% low-melting point agarose or age-matched wild-type zebrafish; approximately 1 min of gap-free recordings are shown. (B) Magnified view of the yellow highlighted region in the above trace from a stxbp1b s3001/s3001 mutant larva. (C) Mean duration of all spontaneous events (greater than 50 msec in duration) was significantly greater in stxbp1b s3001/s3001 mutant larvae compared to stxbp1a s3000/s3000 larvae, WT or agar; *p < 0.001 One-way ANOVA on ranks with a Dunn's multiple comparison test. (D) Frequency of spontaneous events was also significantly greater in stxbp1b s3001/s3001 mutant larvae compared to WT or agarose; *p < 0.001 One-way ANOVA on ranks with a Dunn's multiple comparison test.
doi:10.1371/journal.pone.0151148.g010 a reduced progenitor pool. Another possibility is that stxbp1a mutation could disrupt normal development of neural crest cells, which give rise to both melanophores and craniofacial structures [65].

Behavior
The almost complete immobility of stxbp1a homozygous mutant larvae is similar to the complete immobility of Stxbp1 mutant mice [12]. We found that treatment of stxbp1a s3000/s3000 larvae with pentylenetetrazole (PTZ), a GABA antagonist [66], fails to induce seizures. PTZ is a potent convulsant in wild-type zebrafish, but our results suggest that it is not sufficient to generate coordinated neural activation when synaptic vesicle fusion is disrupted.
By 5 dpf, wild-type zebrafish reliably sleep during the dark phase [67,68] and are more active during the light phase. This diurnal rhythm in behavior was severely disrupted in stxbp1a mutants, but not in stxbp1b mutants. In humans, sleep disruptions are associated with both epileptic encephalopathies (e.g. Lennox Gastaut syndrome) [69] and autism [70].
In our dark-induced movement assay, heterozygous 6 dpf stxbp1a s3000/+ larvae also spent significantly less time moving compared to wild-type siblings. A darkness-induced "O-bend" response is part of the normal larval zebrafish behavioral repertoire by 5 dpf [71], and can be reliably elicited by 10 s intervals of darkness [72]. Unlike the C-bend startle response induced by acoustic or mechanosensory startling stimuli, the O-bend is characterized by higher turn angle and longer latency, and does not require the Mauthner neurons [71]. The neural circuit underlying this behavioral response to a dark flash has not been characterized in detail, but is known to undergo protein-dependent habituation [73]. Heterozygous mutation of stxbp1a, while not significantly impairing the larvae's ability or inclination to move in general, may disrupt their ability to respond normally to arousing stimuli. Combined with the dispersed melanosomes that we found even in light conditions in homozygous stxbp1a and stxbp1b mutants, this decreased responsiveness to dark stimuli suggests the possibility that vision may be impaired in these mutants.

Metabolism, heart rate, and survival
Human mutations in STXBP1 have been linked to deficits in mitochondrial respiratory chain complex IV deficits in liver [9] and impaired oxidative phosphorylation in skeletal muscle [10]. Yet cellular metabolism has not, to our knowledge, been previously investigated in animal models of STXBP1 deficiency. The decreased mitochondrial respiration that we observed in stxbp1a s3000/s3000 mutant zebrafish larvae could be due directly to decreased activity of respiratory chain enzymes in neurons. Decreased glycolytic rates may be decreased to match the decreased energetic needs of the brain. Because we measured glycolysis and mitochondrial respiration at the whole-organism level, there could also be indirect effects of impaired brain development on metabolism in other tissues, for example by neuroendocrine or autonomic signaling.
As would be expected, we did not observe any gross abnormalities in heart development. We observed a small but significant decrease in mean heart rate of homozygous stxbp1a mutants, which could be potentially explained by several types of heart regulation, including intrinsic, neural, and hormonal effects [74,75]. Lack of physical activity and reduced metabolism in homozygous stxbp1a mutants could contribute to this difference.
The early death we observed in stxbp1a s3000/s3000 homozygous mutant zebrafish corresponds to the known phenotype of STXBP1 homozygous mutant mice [12]. The homozygous mutant mice die shortly after birth, likely due to paralysis and failure to breathe. Larval zebrafish do not rely on circulating hemoglobin for oxygenation, as their tissues receive sufficient oxygen via diffusion and their heart rates are not sensitive to hemoglobin function [76]. Therefore, the early death phenotype we observe is unlikely to be due to insufficient tissue oxygenation. Because the stxbp1a s3000/s3000 homozygous mutants die before their siblings, we can exclude the possibility that they simply die from starvation. The early death phenotype can likely be explained by the increasing necessity of synaptic release for even basic autonomic survival functions during the first 10 days of life [61,62].

Modeling human epilepsies in zebrafish
The diversity of consequences of STXBP1 deleterious mutations seen clinically is reflected by the different phenotypes we observed in our two zebrafish mutant lines. Most remarkably, we found evidence during electrophysiology monitoring that homozygous stxbp1b mutant larvae are epileptic, with frequent spontaneous abnormal electrical events indicative of seizures. Our results point to stxbp1b mutant zebrafish as a promising model for EIEE, also known as Ohtahara syndrome [77], and represent a novel CRISPR-Cas9 generated mutant zebrafish line modeling a human form of epilepsy. The electrical phenotype we observe in stxbp1b mutant larvae resemble the "severe and continuous epileptic EEG abnormality" described by Ohtahara and colleagues in human patients, known as a suppression-burst pattern.
Previous efforts modeling epilepsy in zebrafish have largely relied on existing mutations generated randomly, which are not available for every gene, or on morpholino knockdown, which can generate transient effects that are often not replicable in mutant zebrafish [78]. Because we wanted to create stable and consistent loss of function, we chose instead to generate specific germline mutations in stxbp1a and stxbp1b. The mutant phenotypes that we observe likely reflect near-total loss of function of the stxbp1a (frameshift and premature stop codon) and stxbp1b (loss of start codon), respectively. Patients' loss-of-function mutations in STXBP1 include nonsense, missensene, frameshift, and splicing site mutations, suggesting that loss of STXBP1 mutation is causal in EIEE [1,4,79] Recent developments in CRISPR/Cas9 genome editing hold tremendous promise for replicating specific patient-derived mutations in zebrafish in this emerging era of precision medicine [80]. Given the conserved structure and function of many zebrafish homologs of human epilepsy genes, a personalized approach to editing genes in zebrafish could yield fundamental insights as well as new therapies. Furthermore, many factors influencing disease development, including genetic interaction of STXBP1 with other loci, as well as epigenetic, developmental, and sex differences, can be mechanistically studied in model organisms including zebrafish.