Morphological, behavioral and molecular studies on CRISPR/Cas9-mediated knockout of the NOMO1 gene in zebrafish


 Background: Multiple clinical genome-wide analysis identified that chromosome 16p13.11 is a hotspot associated with neuropsychiatric disorders such as autism, schizophrenia and epilepsy. Nodal modulator 1 (NOMO1), located on human chromosome 16p13.11, was considered as a candidate gene with neuropsychiatric disorders. However, it is unknown whether the nomo1 deficiency causes neurological abnormalities, and the molecular mechanisms and pathogenesis of the NOMO1 gene remain unclear. To study the effects of nomo1 deficiency on brain development and neuropsychiatric system, a nomo1 knockout zebrafish model was established.Methods: We developed a viable vertebrate model of nomo1 loss-of-function using CRISPR/Cas9 technology and characterized nomo1 mutant zebrafish. Phenotypic and functional studies of developing nomo1 mutant zebrafish, including morphological measurements, behavioral assays, and functional mechanistic analyses, were performed.Results: Morphological differences in the phenotype of nomo1-/- zebrafish gradually became less noticeable during development, however, the enlarged interstitial spaces in midbrain and hindbrain were detected in nomo1 mutant zebrafish. Meanwhile, the nomo1 deficiency caused the change of expression levels in neurotransmitters of γ-aminobutyrate, glutamate and serotonin. Interestingly, the nomo1 loss-of-function zebrafish model exhibited social defects and repetitive behaviors in juvenile, which represented autism-like behaviors. The transcriptome analysis showed different gene expression patterns in mutant zebrafish at the genetic level. Further results revealed that the neuroactive drug PTZ recovered the decreased locomotor phenotype in larval mutant zebrafish.Conclusions: In this study, we established a nomo1 vertebrate animal model using CRISPR/Cas9 gene editing approach. The loss-of-function of nomo1 displayed autism-like behaviors and altered levels of the γ-aminobutyrate, glutamate and serotonin in zebrafish, which provide evidence that nomo1 as a candidate gene for autism. The versatility of zebrafish model is contributed to studying NOMO1-related disorders and conducting drug screening in future.Limitations: Further studies are needed to determine whether an intervention with a neuroactive drug in nomo1-/- zebrafish to alter the behavioral phenotype is applicable to the behavior of human patients.


Background
Clinical genetic evidence supports the viewpoint that chromosome 16p13.11 is a hotspot associated with various neuropsychiatric disorders, such as epilepsy, autism spectrum disorder (ASD), schizophrenia, and attention de cit hyperactivity disorder (ADHD) [1][2][3][4]. Nodal modulator 1 (NOMO1), a negative regulator of the nodal signaling pathway, is located on chromosome 16p13. 11. In 2015, Tassano et al. described a 13year-old Italian male patient with epilepsy, mental retardation, developmental disorders, and dysmorphic features who had paternally inherited interstitial deleted copy number variations (CNVs) in 16p13.11 using a comparative genomic hybridization (CGH) analysis [5]. In 2016, Brownstein et al. reported evidence of a patient with psychosis who carried duplicated CNVs in 16p13.11 [6]. From the aboved research, we discovered that the NOMO1 gene was included in relevant CNVs. Clinical studies suggest a potential role for NOMO1 in neuropsychiatric disorders; however, the contributing factors explaining how NOMO1 de ciency results in neuropsychiatric disorders remain unclear.
Nodals are essential for formation of the neuroectoderm and mesoderm during the development of early embryos. NOMO1, TMEM147 and nicalin form protein complexes that inhibit the nodal signaling pathway during the early development of zebra sh [7][8][9]. NOMO1 is a candidate gene associated with glioma, early-onset colorectal cancer and facial asymmetry [10][11][12]. In 2018, Cao et al. constructed a nomo1 knockout (KO) zebra sh model that exhibited hypoplasia and dysmorphism symptoms with a phenotype similar to chondrodysplasia in humans [13]. By determining the expression pattern of nomo1 in zebra sh using whole-mount in situ hybridization (WISH), the nomo1 gene was shown to be expressed at high levels in the anterior mesendoderm and endoderm during early embryonic development and was abundant in the brain of larval zebra sh [9,13]. However, none of these studies present evidence to directly link nomo1 de ciency with the nervous system in an animal model.
Abnormalities in the nervous system can lead to different neuropsychiatric disorders. Neuropsychiatric disorders are complex developmental diseases that seriously affect the health of affected individuals and reduce the life quality of patients. Patients with neuropsychiatric disorders usually present various behavioral phenotypes. For example, ADHD patients exhibit characterization of inattention, hyperactivity and increased impulsivity [14]. ASD is characterized by various subtypes of social de cits and the presence of repetitive stereotypic behaviors and restricted interests [15]. In the eld of neuropsychiatry, ASD is one of the widely concerned diseases, which have estimated to impact more than 1% of the population [16]. The genetic factors play an important role in the pathogenesis of ASD [17]. The diagnoses of ASD are mainly depends on the clinical characteristics, however, comorbidities are common throughout patients' lives [18,19]. The biological basis of ASD is not fully understood. Therefore, animal model is a powerful tool to study the pathogenesis of ASD.
Zebra sh have been accepted and applied to functional mechanistic studies of neuropsychiatric disorders as vertebrate model organisms. Researchers have established detection and analytical methods to measure the characteristic behavioral phenotypes, including locomotion, thigmotaxis, social behavior and aggressive behavior, in zebra sh [20,21]. The CRISPR/Cas9 technique is widely used for gene editing, and a number of transgenic zebra sh models have been developed [13,[22][23][24]. Notably, zebra sh models exhibit ASD-like phenotypes similar to those observed in human diseases. For example, dyrk1aa KO zebra sh exhibit social behavior impairments that reproduce human phenotypes of ASD in a vertebrate animal model [25]. Knockout of the shank3b gene via CRISPR/Cas9 resulted in autism-like behaviors and an enlarged ventricles size in zebra sh, which is retrospect the Phelan-McDermid syndrome (PMS) patients that frequently reported in human [22].
We established an in vivo KO model using zebra sh to understand the functional mechanisms underlying nomo1 de ciency and neuropsychiatric disorders. In the present study, nomo1 de ciency in zebra sh causes abnormal development of central nervous system. Meanwhile, the nomo1 mutant zebra sh exhibit social defects and repetitive behaviors at 2 months postfertilization (mpf), which represented autism-like behaviors. Additionally, we reported how nomo1 de ciency conspicuously in uences brain development and neurotransmitter metabolism in zebra sh. These results represent a comprehensive study of the mechanistic link between nomo1 de ciency and ASD-like phenotypes and highlight the critical role of nomo1 in brain development of zebra sh.

Methods
Zebra sh breeding and the generation of nomo1 mutant zebra sh Wild-type (WT) zebra sh of the Tuebingen (TU) strain were provided by the zebra sh facility of the Translational Medical Center for Development and Disease, Shanghai Key Laboratory of Birth Defect, Institute of Pediatrics, Children's Hospital of Fudan University. The zebra sh were raised in a circulating water system with a water temperature of 28.5°C and 14 h of light and 10 h of darkness per day (8:00-22:00, light). Zebra sh breeding, feeding and spawning were conducted strictly in accordance with the Zebra sh Book (http://z n.org/zf_info/zfbook/zfbk.html).
The detailed CRISPR/Cas9-mediated editing method was performed using standard procedures [26,27]. A synthetic speci c guide RNA (sgRNA) and Cas9 mRNA (concentrations of 30 ng/μL and 300 ng/μL, respectively) in a total volume of 3 nL were coinjected into every WT zebra sh embryo at the single-cell stage. Genomic DNA was extracted, and genotyping samples were screened for the mutation frequency by comparison with WT zebra sh samples. The primer sequences used for genotyping are shown in Table S1. The mutant chimeric zebra sh were mated with the TU strain to purify the background and obtain nomo1 +/− zebra sh. Male and female nomo1 +/− zebra sh were crossed to acquire nomo1 +/+ , nomo1 +/− , and nomo -/littermates. We collected multiple batches of littermates produced by nomo1 +/− zebra sh for the phenotypic analysis to obtain a su cient number of embryos.

RT-qPCR
Total RNAs were extracted from embryos, heads and the brain tissues of zebra sh at different developmental stages using TRIzol reagent (Ambion, USA). Genomic DNA was removed by DNase I, and total RNA (1 μg) was reverse transcribed using a PrimeScript cDNA Synthesis Kit (TaKaRa, Japan). RT-qPCR was conducted with a LightCycler ® 480 apparatus (Roche, Germany) and SuperReal PreMix Plus (Tiangen, China) according to the manufacturers' instructions. The fold changes in RNA levels were calculated using the ΔΔCt method. The RT-qPCR primer sequences are listed in Table S1.

WISH
The targeted DNA was cloned into the pGEM-T Easy vector, and probes were synthesized using a linearized plasmid through in vitro transcription with the DIG-RNA labeling Kit (Roche, Austria). The related primers of synthetic probes are shown in Table S1. Embryos of WT and mutant zebra sh were collected at different stages (12 hpf, 24 hpf and 48 hpf) and xed with 4% paraformaldehyde at 4°C overnight. WISH was performed as previously described [23], and images were captured and processed using a Leica 6000 microscope.

Drugs
Pentylenetetrazole (PTZ) (Sigma-Aldrich; P6500, St. Louis, MO) was dissolved in ultrapure water to prepare a 32 mM stock solution that was frozen at -80°C. The PTZ working solution was diluted to the appropriate concentration with system water prior to the experiments.
Behavioral assays in mutant zebra sh Locomotion and thigmotaxis tests Behavioral assays of larval zebra sh were performed at 28.5°C in 24-well plates (Fig. 2G), and the inner diameter of each well was 18 mm, providing the larvae su cient space to swim. The 24-well plates were then placed in a Zebrabox (ViewPoint Life Sciences, Lissieu, Calvados, Lower Normandy Region, France) that recorded videos tracking the larval zebra sh. The experimental procedure consisted of 55 min of continuous illumination with light at an intensity of 100 lx and two 10-min light-dark transition cycles for a total time of 75 min to elicit a photomotor response (PMR) (Fig. 2A). The experiment examined both spontaneous movement and changes due to lighting transitions. The data were quanti ed with ZebraLab software (ViewPoint Behavior Technology, France), the video rate was set to 25 frames per second (fps), and the frames were pooled into 1-min time bins. The threshold was set to 29, a suitable level to accurately detect the trajectory of larval zebra sh in motion.
Zebra sh at 15 days postfertilization (dpf), 30 dpf and 2 mpf swam freely in the open eld at 28.5°C, and the experimental time was 30 min. Behavioral recording began after an adaptation period (1-2 min) when the zebra sh acclimated to the environment. Zebra sh at 15 dpf were examined in a 9-cm diameter dish since they were a smaller size (Fig. 3A). Zebra sh at 30 dpf and 2 mpf were examined in a novel tank (inner dimensions, 30 × 30 × 20 cm) (Fig. 3C). The collected data were exported using ZebraLab software.

Social and repetitive behavior tests
The individual social behavior (social preference behavior) and group social behavior (shoaling behavior) of juvenile zebra sh (2 mpf) were assessed at 28.5°C. A single zebra sh was placed on one side of a standard mating tank (inner dimensions, 21 × 10 × 7.5 cm), and another six WT zebra sh were placed on the other side of the mating tank and separated from the single zebra sh by a transparent plastic plate to examine social preference behavior (Fig. 4A). Region 1 was regarded as a social area, whereas Region 2 was regarded as a nonsocial area, and the experiment lasted for 30 min. Behavioral recording began after an adaptation period (1-2 min) when the zebra sh acclimated to the tank. The behavior of the zebra sh was quanti ed as a distribution of distances or regions adjacent to the group. The ratio of the time the zebra sh stayed in the social area and the distance spent away from the social area directly re ects the social activity of a single juvenile zebra sh.
For the shoaling test, six WT zebra sh (or six nomo1 -/zebra sh) were acclimated to a novel tank (inner dimensions, 30 × 30 × 20 cm) (Fig. 4F). A camera recorded the trajectory of the experimental zebra sh over 30 min, and the adaptation period was 1-2 min. The indicator of inter-individual distance was used to assess the average distance between each zebra sh in the shoal [22,28].
We observed different types of repetitive behavior (back-and-forth motions, stereotypic movement and large circular movement) in the nomo1 -/zebra sh (Fig. 4H) when we examined the spontaneous movement of juvenile zebra sh. All the repetitive behavior tests began after an adaptation period (1-2 min) when the zebra sh acclimated to the tank. Back-and-forth motion was de ned as moving one time on one edge or adjacent edges of the tank and returning to the origin. Stereotypic movement referred to repeated movement of the zebra sh in a small area, where the maximum movement distance from the beginning to the end was less than 30 mm and continuous swimming time was greater than 5 s. Large circular movement referred to the swimming of the zebra sh in a counterclockwise or clockwise circle along the edges of the tank. After the stereotypic movements were de ned, the data were obtained objectively by a computer program within the experimental period.

Kin recognition test
The mating tank, the speci cations of which were the same as the tank used in the social preference test, was divided into three compartments using transit plates to examine kin preference behavior. Three zebra sh of the same strain and same age were placed on one side of the tank, and three red zebra sh of different strains were placed on the other side of the tank. The juvenile WT and mutant zebra sh were placed in the middle area of the tank (Fig. S3A). Region A was regarded as the kin preference area, whereas Region B was regarded as the non-kin preference area. Behavioral recording began after an adaptation period (10 min) when the zebra sh adapted to the tank. Videos were recorded for 30 min. The time spent/distance moved ratio for the zebra sh that stayed in the Region A was used to measure the ability of kin recognition in zebra sh.

Preparation of para nized sections and HE staining
The brain tissues of adult zebra sh (3 mpf) were completely removed under a microscope and immersed in 4% paraformaldehyde for 24 h. Tissues were dehydrated and transparentized with the following conditions and protocol: 70% ethanol, 30°C, 30 min; 95% ethanol, 30°C, 10 min; 95% ethanol, 30°C, 10 min; 100% ethanol (I), 30°C, 10 min; 100% ethanol (II), 30°C, 10 min; xylene (I), 30°C, 30 min; and xylene (II), 30°C, 30 min. The tissues were waxed and embedded using a para n embedding station (Leica EG1150H) for 3 h at 65°C. Then, a microtome (Leica RM2235) was used to produce continuous slices at a thickness of 4 μm. The slices were oated in 40°C warm water to atten their surfaces, and then baked in a 60°C oven. HE staining was performed by staining with hematoxylin and eosin for 5 min each. Photographs were taken with a Leica 205C microscope.
Targeted metabolomics analysis of neurotransmitter Juvenile zebra sh were frozen in liquid nitrogen and placed on ice. Under the microscope, the brains were directly removed with a surgical blade, and 10 brains were placed in an Eppendorf (EP) tube. Selective/multiple reaction monitoring assays (SRM/MRM), which is based on liquid chromatographytandem mass spectrometry (LC-MS/MS), has been used to simultaneously detect neurotransmitters in animals. LC-MS/MS can determine absolute quantitative amounts of target metabolites with strong speci city and high sensitivity and accuracy [29,30].
Brain tissue samples were added to 1 mL of methanol/acetonitrile/ultrapure water (2:2:2, v/v/v), vortexed and ultrasonicated. After an incubation at -20°C for 30 min, the precipitated proteins were centrifuged at 14,000 rcf for 4 min at 4°C. The supernatant was removed and dried in vacuo. For spectrometric detection, 100 μL of an acetonitrile/water solution (1:1, v/v) were used to reconstitute the pellet, which was centrifuged at 14,000 rcf for 4 min at 4°C, after which the supernatant was collected for analysis. Samples were separated using an Agilent 1290 In nity LC Ultra Performance LC System. Mass spectrometry was conducted in negative ion mode using a 5500 QTRAP mass spectrometer (AB SCIEX) for analysis. Data on stability and repeatability were evaluated using cluster and statistical analyses.

HPLC
High-performance liquid chromatography (HPLC) was also employed to examine the levels of norepinephrine, dopamine (DA), 3,4-dihydroxyphenylacetic acid (DOPAC), and serotonin (5-HT) in the brains of juvenile zebra sh. The samples were suspended in ice-cold PBS (20 μL/sample) and ground completely. Then, the samples were centrifuged at 11,900 rpm at 4°C for 10 min, and the protein content of 1 μL of the supernatant from each sample was quanti ed. The remaining supernatants were added to 2 μL of perchloric acid (0.2 N) and centrifuged again at 11,900 rpm at 4°C for 10 min. The supernatant was collected and stored at -80°C. The HPLC analysis was performed using an Agilent 1200 HPLC system (Agilent, USA) with Antec DECADE SDC electrochemical detector (Antec, the Netherlands). The expression levels of neurotransmitters were normalized to the protein content.

Transcriptomics
The total RNA was extracted from each sample of brain tissue from juvenile zebra sh using TRIzol reagent (Ambion, USA) according to the manufacturer's instructions. RNA degradation and contamination were monitored on 1% agarose gels. The RNA concentration was measured using a Qubit ® RNA Assay Kit with a Qubit ® 2.0 uorometer (Life Technologies, CA, USA), and RNA integrity was assessed using an RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Three micrograms of RNA per sample was used as the input for RNA sample preparation. Sequencing libraries were generated using a NEBNext ® Ultra TM RNA Library Prep Kit for Illumina ® (New England Biolabs, USA) according to the manufacturer's recommendations, and index codes were added to attribute sequences to each sample. PCR products were then puri ed (AMPure XP system), and library quality was assessed using the Agilent Bioanalyzer 2100 system. Sequencing fragment data detected using a high-throughput sequencer was converted from image data into sequence data (reads) containing sequence information from each sequenced fragment and its corresponding sequencing quality by CASAVA base recognition. High-quality reads were aligned to the zebra sh reference genome (GRCz11) using HISAT2 v2.0.5. Differentially expressed genes (DEGs) were identi ed using the DESeq2 R package (1.16.1), which uses statistical methods to determine differential expression from digital gene expression data using a model based on a negative binomial distribution. The resulting p-values were adjusted using the Benjamini and Hochberg approach for controlling the false discovery rate. Genes with an adjusted p-value <0.05 according to DESeq2 were de ned as DEGs. In addition, a Gene Ontology (GO) enrichment analysis of the DEGs was implemented by the clusterPro ler R package, which corrected the gene length bias. GO terms with corrected p-values <0.05 were considered signi cantly enriched in DEGs. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database is a resource used to understand high-level functions and utilities of biological systems, such as a cell, an organism and an ecosystem, and it is speci cally focused on analyzing large-scale molecular datasets generated by genome sequencing and other high-throughput experimental technologies (http://www.genome.jp/kegg/). We used the clusterPro ler R package to assess the statistically signi cant enrichment of DEGs in KEGG pathways.

Statistical analysis
The experimental data in this study were analyzed and mapped with GraphPad Prism 6.0 software.
Values are presented as the means ± SEM. Differences between two groups were analyzed using paired Student's t-test and corrected Student's t-test, and p-values <0.05 indicated a signi cant difference.

Results
The nomo1 mRNA expression level from early embryonic development to the juvenile stage and generation of nomo1 mutant zebra sh The nomo1 gene in zebra sh is homologous to the human NOMO1 gene, and they share 68% and 70% identity in cDNA and protein sequences, respectively. Cao et al. and Haffner et al. have con rmed the expression pattern of nomo1 from early embryonic development to larval stages in zebra sh using WISH [9,13]. RT-qPCR was performed to determine the expression level of nomo1 during different developmental stages. At 12 hpf-7 dpf, whole zebra sh embryo were used since they were a smaller size. At 14 dpf, heads of zebra sh were used and at 1-2 mpf, the brain tissues were studied. The expression of the nomo1 mRNA in embryonic zebra sh increased before 48 hpf and decreased at 5 dpf and 7 dpf, as determined using RT-qPCR (12 hpf-7 dpf). As development progressed, the expression of the nomo1 mRNA peaked a second time at 14 dpf, and the highest mRNA expression level was detected in brain tissues from 2 mpf zebra sh (Fig. 1F).
The sgRNA of nomo1 was designed by editing exon 7, which was located before the functional structural domain FN3 in the genomic sequence of zebra sh [31]. The sequence was 5'-GGG CTA TGA TGT CTC TGG AG-3' (Fig. 1A). Fig. 1B-D shows the process by which nomo1 homozygous zebra sh were generated using CRISPR/Cas9 technology. Genomic DNA was extracted, and the speci c PCR products were sequenced, con rming that the nomo1 sequence contained a 1-base deletion (Fig. 1E), resulting in a frameshift mutation and truncated protein 10 amino acids after the mutation (Fig. 1H). The abnormal nomo1 mRNA expression level was detected in mutant zebra sh at 48 hpf, 14 dpf and 2 mpf, as determined using RT-qPCR (Fig. 1G).

Morphological analysis of nomo1 -/zebra sh
We measured morphological changes in mutant zebra sh to examine the consequences of nomo1 de ciency during zebra sh development (Fig. S1). The mortality of the early mutant embryos (24 hpf) was higher than WT zebra sh (Fig. S1H). Mutant embryos exhibited the following morphological changes during development: developmental retardation (the extended rupture of membranes), tail bending (Fig. S1D, arrow), pericardial edema (Fig. S1E, arrow) and developmental malformation (Fig.  S1F). The ratios of morphological changes are shown in Fig. S1H. However, these morphological differences in the phenotype of nomo1 -/zebra sh gradually became less noticeable during development ( Fig. S1A-C). We statistically analyzed the body length of zebra sh at different developmental stages and did not observe a signi cant difference (Fig. S1G).

Nomo1 de ciency signi cantly affected the locomotion of 7 dpf zebra sh under different illumination intensities
We examined the locomotor activity of 7 dpf zebra sh and their reactions to light-dark transitions to determine whether the loss-of-function of nomo1 would modulate the behavior of larval zebra sh. After 20 min of adaptation, the locomotion and thigmotaxis behaviors of larval zebra sh were analyzed ( Fig.  2A). We recorded the trends in the swimming behaviors of WT and mutant zebra sh in 24-well plates (Fig. 2B). An analysis of the average distance moved per minute from minutes 21 to 60 (L0) under light conditions (Fig. 2C) showed that the locomotion of nomo1 -/zebra sh was signi cantly reduced compared with nomo1 +/+ zebra sh, con rming the speci city of the phenotype.
Similarly, the mutant zebra sh displayed signi cantly decreased locomotor activity during the two lightdark cycles, and WT and mutant zebra sh showed a light-sensitive reaction in every light-dark cycle ( Fig.  2D-E). Although the WT and mutant zebra sh showed photosensitivity in the two light-dark cycles, the mutant zebra sh had a more intense response to lighting changes in the rst light-dark cycle than WT zebra sh, whereas no statistically signi cant differences in photosensitivity were observed in the second light-dark cycle (Fig. 2F).
In addition, we measured the thigmotaxis of WT and mutant zebra sh under light conditions by recording the percentage of time spent/distance moved in the inner zone (Fig. 2G). The nomo1 mutant zebra sh exhibited an edge preference under continuous illumination, indicating increased thigmotaxis behavior (Fig. 2H).
The 15 dpf and 30 dpf nomo1 -/zebra sh showed increased locomotion We also analyzed the locomotion of nomo1 -/zebra sh during development. A different sized container was used for 15 dpf and 30 dpf zebra sh, and the trajectories of the zebra sh are shown in Fig. 3A and 3C. The diagram shows a trend in the swimming behavior of WT and mutant developing zebra sh during the experimental period (Fig. 3B, D, and E). The locomotor activity of 15 dpf and 30 dpf zebra sh was signi cantly enhanced. The average distance moved per minute was not signi cantly different between nomo1 -/zebra sh and WT zebra sh at 2 mpf (Fig. 3F).
Juvenile nomo1 -/zebra sh exhibited autism-like behaviors The social preference behavior (Videos S1-S2) and shoaling behavior (Videos S3-S4) were visually presented in the form of videos. The nomo1 +/+ zebra sh swam along the social area throughout the experiment, whereas the nomo1 -/zebra sh swam in a dispersed and random manner. The heat map in Fig. 4B-C was obtained by analyzing the trajectory of the zebra sh in the mating tank. The ratio of time spent/distance moved in the social area was statistically signi cantly lower for nomo1 -/zebra sh than for nomo1 +/+ zebra sh (Fig. 4D-E). We used one of the videos to analyze the trajectory of the zebra sh.
The shoaling behavior was detected to assess the social skills of WT and nomo1 -/zebra sh [32]. Generally, the group WT zebra sh swam together in an open-eld test, re ecting the social nature of the species. The nomo1 -/zebra sh showed an increased inter-individual distance with their companions than the nomo1 +/+ zebra sh, indicating that the nomo1 -/zebra sh had a weaker clustering ability (Fig. 4G).
We also used videos to display the different types of repetitive behaviors of 2 mpf mutant zebra sh, including back-and-forth motion (Video S5), stereotypic movement (Video S6) and large circular movement (Video S7). The trajectories of the repetitive behaviors are shown in Fig. 4H. The numbers of different types of repetitive behaviors were counted. The mutant zebra sh exhibited more of these behaviors than the WT zebra sh and displayed a repetitive behavior-related pattern. The back-and-forth motion was signi cantly increased in mutant zebra sh compared to WT zebra sh (Fig. 4I). However, the number of stereotypic movements and large circular movements were not signi cantly different between nomo1 -/and WT zebra sh (Fig. 4I).
The WT zebra sh (Video S8) and nomo1 -/zebra sh (Video S9) displayed similar tendencies in kin preference region and non-kin preference region. The time spent/distance moved ratio of the mutant was the same as nomo1 +/+ zebra sh (Fig. S3B-C).
Loss-of-function of nomo1 affected the midbrain and hindbrain during zebra sh development At early stages, abundant nomo1 mRNA is transcribed in the endoderm and anterior mesendoderm [9]. In our study, changes in the patterning of the endoderm marker sox17 and mesoderm marker hgg1 were detected in nomo1-de cient mutants. Loss-of-function of nomo1 resulted in a different expression pattern of hgg1 (Fig. 5A-B) (50% of cases, n=20). In our study, nomo1 mutant embryos displayed an increased number of sox17-positive cells (Fig. 5C-D, arrows) (62% of cases, n=18).
Because nomo1 is important for the development of the early mesendoderm, we examined various key neural development-related genes in WT and mutant zebra sh. The expression levels of the genes (HuC, neurog1, islet1, egr2b and foxb1a) in nomo1 mutant zebra sh were not detectably different compared with the levels in WT zebra sh at 24 hpf (Fig. 5E), whereas the mRNA expression levels of the genes decreased signi cantly in the zebra sh at 48 hpf (Fig. 5F). We used RT-qPCR to detect the differences in expression levels in zebra sh, and then a semi-quantitative analysis was performed using WISH to verify the expression pattern. The nomo1 -/zebra sh exhibited signi cant differences in brain development compared with the WT, which was characterized by an absence of expression in midbrain and hindbrain at 48 hpf (Fig. 5G-L, arrows).
We performed HE staining of brain slices from adult zebra sh to determine whether the abnormalities in the brains of larval zebra sh persist throughout development ( Fig. 5M-P). HE staining showed that the nomo1-de cient zebra sh had an abnormal brain structure during development. The changes were mainly manifested as loose and fragile brain tissue, and the midbrain and hindbrain regions of mutant zebra sh exhibited enlarged interstitial spaces. However, staining for forebrain-related markers, such as fezf2, and dopaminergic neurons did not reveal differences between WT and nomo1 -/larvae, indicating that the nomo1 de ciency did not substantially alter forebrain and dopaminergic neuron development (Fig. S2). Based on the aforementioned microscopic morphology at the overall cellular level, we concluded that nomo1 regulates early neural developmental patterning in vivo, strongly in uencing the development of the midbrain and hindbrain in zebra sh.
Expression levels of neurotransmitters and metabolites were altered in the brains of juvenile nomo1 -/zebra sh We examined neurotransmitters and metabolites in juvenile zebra sh brains to reveal the mechanism underlying the autism-like behaviors. Here, we focused on the following 13 neurotransmitters and their metabolites using the LC-MS/MS method: r-amino-butyric acid, levodopa, DA, epinephrine, 5-HIAA, serotonin, 3-methoxytyramine, acetylcholine, histamine, normetanephrine, tyramine, glutamate, and glutamine. Fig. 6A shows the cluster relationships of each metabolite between nomo1 +/+ and nomo1 -/zebra sh, and the tree structure at the left shows the clustering relationship of each group. The results of the hierarchical clustering analysis showed increased levels of ten neurotransmitters and their metabolites in nomo1 -/zebra sh compared with nomo1 +/+ zebra sh. Detailed descriptions of the quantities of the different metabolites are shown in Fig. 6B, and the expression of γ-aminobutyric acid, levodopa, epinephrine, serotonin, 3-methoxytyramine, histamine, normetanephrine, tyramine, glutamate, and glutamine was signi cantly higher in nomo1 -/zebra sh than in nomo1 +/+ zebra sh.
Another method for detecting neurotransmitters, HPLC analysis, produced the same trend. In nomo1 -/zebra sh, the levels of serotonin and norepinephrine were obviously increased (Fig. 6C). Although the levels of dopamine and its metabolite DOPAC showed an increasing trend, the difference was not statistically signi cant (Fig. 6C).
The transcriptome analysis showed different gene expression patterns in the WT and mutant zebra sh The juvenile zebra sh were selected for the transcriptome sequencing analysis to further study the functional mechanism between ASD-like phenotypes and nomo1 de ciency at the genetic level. A volcano map visually shows the distributions of DEGs for each of the comparisons. Two hundred ninetytwo genes were downregulated, 254 genes were upregulated, and the expression of 546 genes was changed at the transcriptional level in mutant zebra sh (Fig. 7A). The altered DEGs are listed and their pvalues are shown in Table S2. We performed GO annotation and KEGG pathway analyses to understand the functions and roles of DEGs. The GO annotation classi cation chart revealed the functions of DEGs in biological processes, cellular components and molecular functions (Fig. 7B). Through the classi cation and statistical analysis of KEGG signaling pathways, nomo1 was shown to participate in the biological processes of metabolism, cellular function, intestinal immunity, and pathogen infection (Fig. 7C).
A functional analysis of the DEGs revealed that nos2a and hbba1 were upregulated, whereas rnps1, ass1, slc39a8, slc43a2b, cfap100, slc15a2 and entpd8 were downregulated in mutant zebra sh compared to their expression in WT zebra sh. RT-qPCR was used to verify the mRNA expression levels of DEGs and con rmed that the difference in gene expression was consistent with the sequencing results (Fig. 7D).
The decreased locomotion of 7 dpf zebra sh was recovered by the neuroactive drug PTZ We applied the neuroactive drug PTZ, an antagonist of the γ-aminobutyrate (GABA) receptor, and measured the locomotor activity of larval zebra sh to explore the effects of drugs that alter neurological disease-related phenotypes. Here, 8 mM PTZ was utilized. The experiment lasted for 60 min and was conducted under continuous illumination. The zebra sh were maintained in a 24-well plate at a room temperature of 28°C room. The movement trajectories of zebra sh (nomo1 +/+ , nomo1 -/-, nomo1 +/+ +PTZ, and nomo1 -/-+PTZ) in the experimental period are shown in Fig. 2I. Upon exposure to PTZ, WT zebra sh exhibited increased locomotion, similar to our previous study [20]. The locomotor activity of mutant zebra sh treated with PTZ increased signi cantly compared with nomo1 -/zebra sh, and a statistically signi cant difference in locomotor activity was observed under continuous illumination conditions (Fig.  2J).

Discussion
Although some functional roles of nomo1 have been described in previous studies, the effects of neuropsychiatric system caused by nomo1 de ciency remain unclear. In this study, we reported the morphological, behavioral and functional mechanism of nomo1 mutant zebra sh at different developmental stages. Previous nomo1 KO zebra sh model that was associated with bone formation and cartilage development [13]. Authors observed early-stage bone developmental defects in larvae before 7 dpf, which might be one of reasons for the behavioral disorders in larval and infant zebra sh.
Behavioral features are in uenced by many factors, particularly in context of the neuropsychiatric system and motor ability. Although the change of locomotor activity is di cult to estimate from the nervous system or bone developmental defects, the social-level results make sense, emphasizing the importance of nomo1 in neuropsychiatric systems. The locomotion of juvenile mutant zebra sh did not change compared with WT zebra sh; however, the speci c movement pattern did change. Based on the abundance of nomo1 throughout the brain, we consider that the overall ber morphology of the brain might have changed due to the mutation, subsequently in uencing the nervous system to affect the behavioral phenotype. Thus, nomo1 loss-of-function exerted an important effect on the neuropsychiatric system in zebra sh.
Our study provides several important ndings. We highlight the essential role of nomo1 in ASD-like phenotypes during development. Behavioral assays of nomo1 mutant zebra sh in the juvenile period presented social defects and repetitive behaviors, which is similar with autism-like behaviors. Similar phenotypes have been observed in other transgenic mutant zebra sh models of shank3b [22], dyrk1a [25] and fmr1 [33]. Here, we analysed our results by comparing them with ndings from studies that used zebra sh to investigate the function of genes that are strongly associated with ASD (Table S3).
Combined with clinical research, our results show the functional de ciency of nomo1 as an underlying disease mechanism for autism. Autism is one of the neuropsychiatric disorders, and comorbidities are common in patients [18,34]. Developing nomo1 mutant zebra sh exhibited decreased/increased locomotion, speculating the mutant zebra sh not only displayed autism-like behavior but also related to ADHD-or epilepsy-like phenotypes. However, further studies are warranted to determine whether the behavioral phenotype observed in animal models is directly associated with human diseases. Neuropsychiatric disorders exhibit substantial heterogeneity. Different gene mutations may lead to a speci c disease, and a single gene mutation can cause comorbidities in individuals with ASD and other neuropsychiatric diseases. Therefore, studies of the functions of different genes, different signaling pathways and different mutation models are warranted to determine the pathogenesis of the diseases.
Moreover, additional behavioral measurements, including comorbidities such as seizures and cognitive impairment, are needed to fully assess neuropsychiatric disease-like behavioral phenotypes using validate animal model.
Based on our results, the loss-of-function of nomo1 in uenced the development of the midbrain and hindbrain. Haffner et al. indicated that transiently decreased expression of nomo1 affects the development of the mesendoderm and hatching glands [9]. Speci cally, our study revealed abnormal expression levels of mesendoderm and endodermal markers in steadily inherited embryonic mutant zebra sh. The potential neural induction properties of the mesendoderm indicate that the mesendoderm is necessary for brain development [35,36]. Meanwhile, we detected abnormalities in brain developmental markers during the early neural development of mutant larvae accompanied by loosening of the tissue structure in the midbrain and hindbrain of adult mutant zebra sh. Patients with ASD exhibit a reduced brain parenchyma in magnetic resonance imaging (MRI) [37]. Thus, our ndings also provided a logical consistency between structural abnormalities in the brain and behavioral phenotypes related to ASD. Moreover, researchers have indicated that nomo1 affects body axis development [9], which might be the cause of the increased mortality and varying degrees of developmental malformations observed in larval zebra sh in this study. Notably, the delayed neurodevelopment in terms of the appearance of the phenotype and body length in mutant zebra sh became less noticeable during development. The mechanism underlying this effect is currently unknown, but nomo1 can affect morphological development at speci c stage to exert a more pronounced effect on the nervous system across developmental stages.
Behavioral phenotypes are associated with a complex pathogenesis, which is typically characterized by alterations in the activity of neurotransmitters. Genetic association studies have veri ed the key role of DA in the etiology of ADHD [38,39]. In our study, targeted metabolomics showed a signi cant increase in the levels of levodopa, a precursor of DA. However, the level of the metabolite DA had not changed in mutant zebra sh compared to WT zebra sh, as determined using HPLC. This result is consistent with the changes in dopamine neurons observed in larvae zebra sh using WISH. We speculated that other precursors of DA collectively modulated the level of DA. Interestingly, we detected alterations in γaminobutyrate (GABA), glutamate and serotonin levels when analyzing neurotransmitter metabolism.
Moreover, 5-HT was shown to accumulate at a high rate using both of these methods. Evidence supports a novel and important link between schizophrenia, GABA and glutamate alterations [40]. In addition, several ndings suggest a correlation between glutamate and GABA in subjects with ASD, which also indicated that the neurotransmitter contributed to brain development [41,42]. Abnormalities in the neurotransmitter system have been observed in children or animal models with neuropsychiatric disorders, involving mainly glutamatergic, GABAergic and other neuronal populations [43,44]. Clinical studies have documented elevated levels of 5-HT in the peripheral blood of children with ASD [45], and some researchers have reported differences in the 5-HT distribution in different regions of the brains of patients with ASD [46]. Therefore, we concluded that the nomo1 de ciency displayed a strong correlation with brain metabolism of GABA, glutamate and 5-HT in zebra sh, which represent a vital decision-making aspect of autism-like behavior. However, the mechanism of how neurotransmitters affect behavioral changes is unclear. In future studies, metabolism mechanism and pathway of neurotransmitters in brain involved in nomo1 KO zebra sh model will be examined.
Transcriptome sequencing analyses have identi ed multiple DEGs associated with neuropsychiatric disorders. For example, the association of RNPS1 and NOS2 with neurodevelopmental disorders in patients has been veri ed [47,48]. Additionally, researchers have presented evidences strongly suggesting and association between SLC39A8 and schizophrenia [49]. We also identi ed other DEGs with abnormal expression levels in mutant zebra sh (ass1, hbba1, slc43a2b, cfap100, entpd8, and slc15a2). Some of these genes are novel, and their functions have not been completely elucidated in the neuropsychiatric system. However, the DEGs may provide important clues regarding the neuropsychiatric disorders observed due to a nomo1 de ciency at the genetic level. Although our study was limited, the transcriptome results provided clues for studying the interaction network between genes. Moreover, it will provide new directions for studies of the signaling pathways associated with DEGs correlated with nomo1 de ciency with neuropsychiatric system. The analysis of neurotransmitter metabolism revealed that nomo1 is associated with GABA signaling in zebra sh. GABA is an inhibitory neurotransmitter of the central nervous system, and PTZ is an antagonist of the GABA receptor. Therefore, we attempted to reverse part of the behavioral phenotypes in the nomo1 mutant zebra sh model using PTZ. PTZ increased the spontaneous movement of larval zebra sh under light conditions [20]. In addition, Mussulini et al. found that an increase in the PTZ concentration increased body movement and seizure-like behavior, which were observed following treatment with 15 mM PTZ in adult zebra sh [50]. In the present study, we inferred that mutant zebra sh remained sensitive to neuroactive drugs and that PTZ partially improved the locomotor activity of larval mutant zebra sh. The ASD include a range of characteristic behavioral phenotypes, whereas the exploration of therapeutic effects of targeted small-molecule compounds obviously require further research. This study provides a powerful functional basis for the future exploration of drug sensitivity and intervention in nomo1 mutant zebra sh and may provide additional insights into research performed on this topic in the future.
Taken together, the data presented here conclusively indicate that nomo1 de ciency causes abnormal development of the nervous system and autism-like behaviors in zebra sh. The nomo1 zebra sh model described in this study provides an important foundation for neuroactive drug screens in precision medicine in the future.

Limitations
As shown in the present study, the loss-of-function of nomo1 elicits autism-like behaviors; however, more evidence on molecular pathogenic mechanisms of nomo1 de ciency is needed. Moreover, the effects of an intervention with a neuroactive drug on the behavioral phenotype and functional mechanism of mutant zebra sh requires additional detailed and extensive evidence.

Conclusions
In this study, we generated a nomo1 loss-of-function zebra sh model that presented autism-like phenotypes, particularly social defects and repetitive behaviors in juvenile zebra sh. The enlarged interstitial spaces of midbrain and hindbrain in nomo1 -/zebra sh suggest the functional mechanism of nomo1 de ciency in brain development. The increased neurotransmitter levels of γ-aminobutyrate, glutamate and serotonin in nomo1 -/zebra sh and transcriptome analysis at the genetic level also provide further evidence supporting the potential role of nomo1 de ciency resulted in autism-like phenotypes. These results indicate a functional de ciency of nomo1 as a candidate gene for autism. The nomo1 mutant zebra sh provide a valuable model for future studies of the molecular pathogenesis of nomo1-related disorders and drug screening.

Declarations Acknowledgments
We thank all members of the Qiang Li laboratory for their support, and we thank the Translational Medical Center for Development and Disease of Children's Hospital of Fudan University in China. We thank Dr. Ning Guo for the advice on behavioral analyses and Chenwen Zhu for technical discussions related to the experiments.

Funding
This study was supported by grants from the National Natural Science Foundation of China (NSFC, no. 81771632 and no. 81271509) to Qiang Li. Xiuyun Liu is supported by a grant from the NSFC (no. 81601329).

Availability of data and materials
The datasets analyzed in the current study are available from the corresponding author upon reasonable request.

Authors' contributions
This study was designed by QL, XW and FL. FL, JL, TTL and JJ performed the experiments. FL wrote the manuscript, with extensive editing by QL. YLZ, XYL and YXJ designed the CRISPR/Cas9-mediated mutagenesis of nomo1 loci in zebra sh and provided assistance with the identi cation of homozygous animals. QZ provided technical assistance with transcriptomics and HPLC analyses. All authors read and approved the nal manuscript.  Differences in the expression of neurotransmitters and metabolites between the brain tissues of WT and nomo1-/-juvenile zebra sh. Thirteen neurotransmitters and metabolites were analyzed using the LC-MS/MS, and four neurotransmitters were analyzed using HPLC. (A) The clustering analysis of nomo1+/+ and nomo1-/-zebra sh using SRM/MRM revealed that neurotransmitters and metabolites tended to be upregulated in mutant zebra sh. (B) Statistical analysis of the SRM/MRM data. The vertical axis denotes the normalized levels of neurotransmitters and metabolites (N=3×8 groups each of WT and mutant zebra sh). (C) Statistical analysis of the HPLC data. The vertical axis denotes the levels of 4 neurotransmitters and metabolites in zebra sh brain tissues (N=8×7 for each group). Data are presented as the means±SEM, *p < 0.05, **p < 0.01, and ***p < 0.001.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.