CRISPR/Cas9-Induced Inactivation of the Autism-Risk Gene setd5 Leads to Social Impairments in Zebrafish

Haploinsufficiency of the SETD5 gene, encoding a SET domain-containing histone methyltransferase, has been identified as a cause of intellectual disability and Autism Spectrum Disorder (ASD). Recently, the zebrafish has emerged as a valuable model to study neurodevelopmental disorders because of its genetic tractability, robust behavioral traits and amenability to high-throughput drug screening. To model human SETD5 haploinsufficiency, we generated zebrafish setd5 mutants using the CRISPR/Cas9 technology and characterized their morphological, behavioral and molecular phenotypes. According to our observation that setd5 is expressed in adult zebrafish brain, including those areas controlling social behavior, we found that setd5 heterozygous mutants exhibit defective aggregation and coordination abilities required for shoaling interactions, as well as indifference to social stimuli. Interestingly, impairment in social interest is rescued by risperidone, an antipsychotic drug used to treat behavioral traits in ASD individuals. The molecular analysis underscored the downregulation of genes encoding proteins involved in the synaptic structure and function in the adult brain, thus suggesting that brain hypo-connectivity could be responsible for the social impairments of setd5 mutant fishes. The zebrafish setd5 mutants display ASD-like features and are a promising setd5 haploinsufficiency model for drug screening aimed at reversing the behavioral phenotypes.


Introduction
Autism spectrum disorders (ASD) are a heterogeneous group of complex neurodevelopmental syndromes affecting approximately 1% of the human population. Typical symptoms of ASD include early-onset impairments in communication, learning and social interactions, as well as restricted interests and repetitive behaviors [1]. These manifestations can be associated with a variety of other symptoms, including intellectual disability (ID), hyperactivity, motor deficits and developmental delay. Although environmental factors can be important contributing factors, recent studies have accumulated a large body of evidence indicating that mutations in an increasing number of genes can cause ASD. Most of the ASD-risk genes code for proteins involved in transcription, protein synthesis and degradation, as well as neurogenesis and synaptogenesis [2,3]. The first category is of particular interest and includes genes that code for transcription factors and chromatin remodelers, which are expected to act at a higher hierarchical level than the other classes 2 of 17 of genes, being able to regulate the expression of target genes and thus triggering the activation of specific genetic programs. Furthermore, chromatin regulators are known to play key roles in various aspects of neural development, including progenitor specification, cell-type specific differentiation, migration and generation of mature neural networks [4]. As a consequence, the impairment of chromatin remodeling results in crucial deficiencies in circuit formation and cognitive functions. Recently, converging evidence has indicated that heterozygous loss of function (LoF) mutations in the gene SETD5, encoding a SET domaincontaining histone-modifying protein, are one of the most frequent genetic causes of both ID and ASD [5][6][7][8][9][10][11][12][13]. Several studies focusing on the molecular function of SETD5 have highlighted distinct, non-mutually exclusive, chromatin-regulating activities, involving this protein in histone methylation [14,15] as well as in the interaction with the polymeraseassociated factor 1 (PAF1) complex and histone deacetylase 3 (HDAC3) complex [15][16][17][18][19]. Functional studies performed in mice indicated that Setd5 haploinsufficiency negatively affects the expression of neurodevelopmental genes, in particular those associated with synaptic functioning, learning and memory, eventually leading to an ASD-like phenotype, which includes cortical hypoconnectivity, cognitive deficit, altered social interactions and impairments in adaptive behavior [15,20].
Recently, the zebrafish (Danio rerio) has been recognized by SFARI, one of the major foundations supporting autism research, as a rapidly emerging model to study ASD [21]. Indeed, the zebrafish model offers a variety of advantages including a large number of offspring, rapid external development of its transparent embryos, the availability of classical genetic as well as reverse genetic approaches and a genome sharing over 70% of its genes with humans [22]. Furthermore, zebrafish and mammals display strong similarities in neural cell types as well as conserved signaling pathways. Finally, the availability of specific tests to assess different aspects of the complex behaviors of this social fish, makes zebrafish a powerful complementary model system to mouse, to elucidate the function of ASD risk genes and perform high-throughput drug screening.
In the present work, we used the CRISPR/Cas9 technology to generate the first zebrafish setd5 mutant line. As human patients are characterized by SETD5 haploinsufficiency, we focused our analysis on zebrafish heterozygous setd5 mutants. Interestingly, we found that these mutants display ASD-like features and therefore they can be proposed as novel system to model and study key aspects of SETD5 haploinsufficiency in humans.

setd5 Is Expressed in Zebrafish Adult Brain and during Early Embryo Development
In order to characterize the localization of setd5 mRNA in zebrafish central nervous system (CNS), we analyzed the expression of setd5 transcript on WT brain sections of adult individuals (10/12 months of age) by in situ hybridization. As shown in Figure 1a, setd5 is expressed in different areas of the telencephalon (Figure 1a, slices 1-3) (including telencephalic dorsal area, telencephalic ventral area, olfactory bulb and lateral olfactory tract), diencephalon (Figure 1a, slices 4,5) (including parvocellular preoptic nucleus, ventral hypothalamus, posterior tuberal nucleus, medial preglomerular nucleus, diffuse nucleus of the hypothalamus lower lobe, caudal area of the periventricular hypothalamus and dorsal zone of the periventricular hypothalamus), mesencephalon (Figure 1a, slices 4,5) (including longitudinal torus, optical roof, periventricular gray area of the optical roof and dorsal nucleus of the tegmentum) and of the rhombencephalon (including the cerebellar valvula, cerebellar body, cerebellar caudal lobe, LVII facial lobe and nerves of the anterior lateral line) and in the eighth cranial nerve (Figure 1a, slice 6).
We also analyzed the expression levels of the setd5 transcript during zebrafish embryo development from the zygote stage up to 72 h post-fertilization (hpf), observing relatively high expression levels already observed at zygote stage ( Figure 1b). This early appearance of setd5 RNA, also previously reported by whole mount in situ hybridization experiments [15], indicate the presence of maternal transcripts. This evidence suggests an early role played by the Setd5 protein during the first stages of embryo development.
We also analyzed the expression levels of the setd5 transcript during zebrafish embryo development from the zygote stage up to 72 h post-fertilization (hpf), observing relatively high expression levels already observed at zygote stage ( Figure 1b). This early appearance of setd5 RNA, also previously reported by whole mount in situ hybridization experiments [15], indicate the presence of maternal transcripts. This evidence suggests an early role played by the Setd5 protein during the first stages of embryo development. Figure 1. setd5 is expressed in the zebrafish adult brain and at early stages of zebrafish embryo development. (a) setd5-203 isoform localization on zebrafish WT brain sections at 10/12 months of age performing in situ hybridization technique. Representative images of a section corresponding to telencephalon (1-3), diencephalon-mesencephalon (4,5) and rhombencephalon (6), n = 3 brains analyzed. Scale bar 300 µ m. (b) Expression of setd5 isoforms in zebrafish embryos and larvae at different developmental stages, obtained by RT-qPCR analysis. Data are expressed as 2 −(ΔCt) mean ± Standard Error of the Mean (SEM), using b-actin1 as housekeeping gene. n = 3 independent experiments. Abbreviations: D, Dorsal telencephalic area; V, Ventral telencephalic area; OB, Olfactory bulb; LOT, Lateral olfactory tract; Ppa, Parvocellular preoptic nucleus; TL, Longitudinal Torus; Val, Lateral division of valvular cerebelli; TeO, Tectum opticum; PGZ, Periventricular gray zone of optic tectum; DTN, Dorsal tegmental nucleus; HV, Ventral zone of periventricular hypothalamus; PTN, Posterior tuberal nucleus; PGm, Medial progromerular nucleus; DIL, Diffuse nucleus of the inferior lobe; Hc, Caudal zone of periventricular hypothalamus; Hd, Dorsal zone of the periventricular hypothalamus; CCe, Cerebellar corpus; LCa, Caudal lobe of cerebellum; LVII, Facial lobe; ALLN, Anterior lateral line nerves; VIII, Octaval nerve; hpf, hours post fertilization. Figure 1. setd5 is expressed in the zebrafish adult brain and at early stages of zebrafish embryo development. (a) setd5-203 isoform localization on zebrafish WT brain sections at 10/12 months of age performing in situ hybridization technique. Representative images of a section corresponding to telencephalon (1-3), diencephalon-mesencephalon (4,5) and rhombencephalon (6), n = 3 brains analyzed. Scale bar 300 µm. (b) Expression of setd5 isoforms in zebrafish embryos and larvae at different developmental stages, obtained by RT-qPCR analysis. Data are expressed as 2 −(∆Ct) mean ± Standard Error of the Mean (SEM), using b-actin1 as housekeeping gene.

setd5 Knock-Out Causes a Growth Delay in Zebrafish
Since we previously demonstrated that setd5 knock-down affects zebrafish embryonic development [15], we took advantage of the CRISPR/Cas9 genome-editing technique to generate a stable mutated zebrafish model for the setd5 gene. Cas9 mRNA was injected into the zygote cell along with a guide RNA, targeting the setd5 exon 7 which is upstream of those encoding the protein functional domain SET (Figure 2a). A representative group of injected F0 embryos were processed by Melting analysis to confirm the gene editing event ( Figure S1 (Supplementary Materials)). Once they reached the adult stage, the mosaic founders able to transmit mutations to their progeny were identified by Melting analysis on heterozygous F1 generation embryos obtained after outcrossing with WT individuals. Many indel mutations were identified and we focused on a deletion of eight nucleotides: this frameshift mutation resulted in a putative LoF of setd5 gene since it determines the appearance of a premature stop codon ( Figure S2). We then confirmed the reduced expression of setd5 transcript in setd5 mutant adult brains (Figure 2b), which may lead to the activation of compensatory mechanisms of gene expression, a well-demonstrated phenomenon in zebrafish [23]. Indeed, while the expression of the setd5 paralogues setd2 and mll5 in the adult brain was not different between setd5 mutants and WT ones (Figure 2c), heterozygous mutants showed a significant increase in expression of nsd1a and nsd1b (both the isoforms), the orthologues of human NSD1 (Figure 2d), encoding a histone methyltransferase that shares the same amino acidic target, Lysine 36 of histone 3, with SETD5 [24].

setd5 Knock-Out Causes a Growth Delay in Zebrafish
Since we previously demonstrated that setd5 knock-down affects zebrafish embryonic development [15], we took advantage of the CRISPR/Cas9 genome-editing technique to generate a stable mutated zebrafish model for the setd5 gene. Cas9 mRNA was injected into the zygote cell along with a guide RNA, targeting the setd5 exon 7 which is upstream of those encoding the protein functional domain SET (Figure 2a). A representative group of injected F0 embryos were processed by Melting analysis to confirm the gene editing event ( Figure S1 (Supplementary Materials)). Once they reached the adult stage, the mosaic founders able to transmit mutations to their progeny were identified by Melting analysis on heterozygous F1 generation embryos obtained after outcrossing with WT individuals. Many indel mutations were identified and we focused on a deletion of eight nucleotides: this frameshift mutation resulted in a putative LoF of setd5 gene since it determines the appearance of a premature stop codon ( Figure S2). We then confirmed the reduced expression of setd5 transcript in setd5 mutant adult brains (Figure 2b), which may lead to the activation of compensatory mechanisms of gene expression, a well-demonstrated phenomenon in zebrafish [23]. Indeed, while the expression of the setd5 paralogues setd2 and mll5 in the adult brain was not different between setd5 mutants and WT ones (Figure 2c), heterozygous mutants showed a significant increase in expression of nsd1a and nsd1b (both the isoforms), the orthologues of human NSD1 (Figure 2d), encoding a histone methyltransferase that shares the same amino acidic target, Lysine 36 of histone 3, with SETD5 [24].  As shown in Figure 3a,b, body length and body weight were significantly reduced in mutant setd5+/− zebrafish adults when compared to setd5+/+ individuals. A morphometric analysis of explanted brains indicated that in mutant fishes the length of the telencephalon normalized to the length of the entire brain is significantly increased compared to wild type fishes (Figure 3c,d). As the volume of the entire brain is not different between mutants and wild type fishes ( Figure S4), these data suggest that setd5+/− fishes display an alteration of the relative morphological proportions of the telencephalon compared to other brain districts.
As shown in Figure 3a,b, body length and body weight were significantly reduced in mutant setd5+/− zebrafish adults when compared to setd5+/+ individuals. A morphometric analysis of explanted brains indicated that in mutant fishes the length of the telencephalon normalized to the length of the entire brain is significantly increased compared to wild type fishes (Figure 3c,d). As the volume of the entire brain is not different between mutants and wild type fishes ( Figure S4), these data suggest that setd5+/− fishes display an alteration of the relative morphological proportions of the telencephalon compared to other brain districts.   (TeO) length and width, Cerebellum (Ce) length and width, Diencephalon (Die) length and width in adult brain, normalized to the entire brain length, from 10-month-old setd5+/+ zebrafish adults. Absolute values are included in Figure S5. (d) n = number of adult brains analyzed. Data are expressed as mean ± SEM. Statistical analysis was performed by Student's t-test. * p < 0.05; ** p < 0.01.

setd5 Mutant Embryos Show Growth Delay, Microphthalmia and Deficits in Locomotor Behavior
We also performed morphometric analysis on setd5 mutant embryos at 2 days postfertilization, to check if the altered phenotype observed in knock-down embryos, characterized by microcephaly, reduced eye size, cerebral and pericardial edema [15], could also be observed in stable setd5 mutants. Mutant embryos showed a reduced body length ( Figure 4a) and eye area ( Figure 4b) compared to setd5+/+ ones, although neither cerebral nor pericardial edemas were detected in mutant embryos. In addition, the ratio between the eye diameter and the body length was decreased in mutant embryos compared to setd5+/+ ones (Figure 4c), which is indicative of microphthalmia. To look for potential microcephaly during an early phase of brain development, we analyzed the brains of 24 hpf embryos. However, we did not find significant differences between setd5+/and setd5+/+ embryonic brain areas ( Figure S4a). Furthermore, in the same brains no difference was detected in the level of phospho-histone H3 and HuC/D used as markers of proliferation and neuronal differentiation, respectively ( Figure S4b,c). be observed in stable setd5 mutants. Mutant embryos showed a reduced body length (Figure 4a) and eye area ( Figure 4b) compared to setd5+/+ ones, although neither cerebral nor pericardial edemas were detected in mutant embryos. In addition, the ratio between the eye diameter and the body length was decreased in mutant embryos compared to setd5+/+ ones (Figure 4c), which is indicative of microphthalmia. To look for potential microcephaly during an early phase of brain development, we analyzed the brains of 24 hpf embryos. However, we did not find significant differences between setd5+/-and setd5+/+ embryonic brain areas ( Figure S4a). Furthermore, in the same brains no difference was detected in the level of phospho-histone H3 and HuC/D used as markers of proliferation and neuronal differentiation, respectively ( Figure S4b,c).  We tested the locomotor activity of setd5+/− larvae at 6 dpf and observed a decrease in total distance swum ( Figure 4d) and speed (Figure 4e) in setd5+/− larvae compared to setd5+/+ ones and a similar result was obtained for setd5 knock-down embryos. To extend our analysis of locomotor behavior to setd5+/− fishes at the adult stage, we performed an open field test to evaluate the ability of the fish to explore a novel environment. As observed in setd5+/− larvae, adult setd5+/− fish showed a significant decrease in total distance swum ( Figure 5a) associated with reduced speed (Figure 5b), compared to setd5+/+ counterparts.
We tested the locomotor activity of setd5+/− larvae at 6 dpf and observed a decrease in total distance swum ( Figure 4d) and speed (Figure 4e) in setd5+/− larvae compared to setd5+/+ ones and a similar result was obtained for setd5 knock-down embryos. To extend our analysis of locomotor behavior to setd5+/− fishes at the adult stage, we performed an open field test to evaluate the ability of the fish to explore a novel environment. As observed in setd5+/− larvae, adult setd5+/− fish showed a significant decrease in total distance swum ( Figure 5a) associated with reduced speed (Figure 5b), compared to setd5+/+ counterparts.

setd5 Mutant Adults Show a Tight Shoal
Zebrafish exhibit shoaling behavior, the complex interaction of animals moving together in coordinated movements with polarized direction. Shoal cohesion is usually stable and maintains a relatively high baseline level in adult zebrafish. The shoaling test was used to assess overall social behaviors in groups of four individuals. Compared to WT, setd5+/− adults tended to swim closer together, in tighter shoals as demonstrated by the reduced Nearest Neighbor Distance (NND; Figure 5c) and a decreased Inter-Individual

setd5 Mutant Adults Show a Tight Shoal
Zebrafish exhibit shoaling behavior, the complex interaction of animals moving together in coordinated movements with polarized direction. Shoal cohesion is usually stable and maintains a relatively high baseline level in adult zebrafish. The shoaling test was used to assess overall social behaviors in groups of four individuals. Compared to WT, setd5+/− adults tended to swim closer together, in tighter shoals as demonstrated by the reduced Nearest Neighbor Distance (NND; Figure 5c) and a decreased Inter-Individual Distance (IID; Figure 5d). Moreover, setd5+/− shoals presented a disrupted polarization compared to setd5+/+ fish group (Figure 5e) and swam with a reduced speed (Figure 5f), demonstrating an alteration of social behavior generally associated to an anxiety condition.

setd5 Mutant Adults Display Perturbed Social Interaction, Ameliorated by the Antipsychotic Drug Risperidone
Sociality is an essential component of zebrafish behavior that is easily observable and further increases the value of this species in modelling brain disorders involving deficits in social behavior, including ASD [25]. We performed a social preference test (Figure 6a) to evaluate the reaction of adult zebrafish individuals to a social stimulus as well as to a social novelty [25]. As expected [26], during round 1, the sedt5+/+ fishes spent significantly more time close to the stimulus fish in zone 1, compared to zone 2 (empty zone) which is far from the social stimulus. On the contrary, there was no significant difference between the time spent in zone 1 and in zone 2 by setd5 mutant fishes, indicating an impairment of social interest (Figure 6b). Once a novel social stimulus, represented by another fish, was added in round 2, the sedt5+/+ fishes started to spend an equal time in zone 2, which is close to the additional fish, and in zone 1 which is nearby the first social stimulus. In this second round, setd5+/− fish displayed the same behavior observed during round 1, spending the same amount of time in zone 1 and zone 2 (Figure 6b). We next repeated the social preference test exposing zebrafish sedt5+/+ and setd5 mutant fishes to risperidone, an antipsychotic drug commonly used to treat behavioral traits in ASD patients. In the case of sedt5+/+ fishes, in round 1 we obtained similar results as those seen in the absence of the drug. Interestingly, in the setd5 mutant fishes we observed a significant increase in the time spent in zone 1 compared to zone 2, indicating a rescue of the social interest by risperidone treatment. During round 2, we did not observe a significant difference between the time spent in zone 1 and zone 2, in either WT or setd5+/− fishes, as observed in the untreated fishes (Figure 6c).
To confirm that setd5 haploinsufficiency affects the social interest of adult fishes, we performed a social interaction test, tracking the nose point, beyond center point and tail base, of the individuals (Figure 6d). As shown in Figure 6e, the setd5 mutant zebrafish showed a significant increase in the distance between the nose point of the tested fish and the nose point of the stimulus, which was abolished by the treatment with risperidone (Figure 6f). treated fishes (Figure 6c).
To confirm that setd5 haploinsufficiency affects the social interest of adult fishes, we performed a social interaction test, tracking the nose point, beyond center point and tail base, of the individuals (Figure 6d). As shown in Figure 6e, the setd5 mutant zebrafish showed a significant increase in the distance between the nose point of the tested fish and the nose point of the stimulus, which was abolished by the treatment with risperidone ( Figure 6f).

setd5 Mutation Affects the mRNA Expression Levels of Proteins Involved in Neurotransmission
Since neurotransmission is the basis of neuronal communication and is critical for normal brain development, behavior, learning and memory, we investigated different neurotransmitter pathways and proteins involved in neuronal activity in brain samples from adult setd5 mutants. In particular, we evaluated the mRNA expression levels of gad2, gad1a, gad1b (GABAergic system), dbh, dopa decarboxylase, dat, tyrosine hydroxylase 1 and 2 (catecholaminergic system), tph1a, tph1b, tph2, serta and sertb (serotoninergic system), histidine decarboxylase 1 (histaminergic system), vesicular monoamine transporter vmat2, the differentiated neuronal marker elavl3, the transcripts encoding for the synaptic proteins dyrk1aa and dyrk1ab, PSD95 and synapsin 1, synaptophysin a, synaptophysin b, mecp2, nrxn1a, nrxn1b, shank3a and shank3b. In particular, we investigated a possible relationship between the expression level of these transcripts and the levels of WT setd5 mRNA, evaluated using specific primers that do not amplify mutant setd5 transcript ( Figure S5). Although a comparison between brain samples from sedt5+/+ and setd5+/− fishes did not show any significant difference in the expression level for most of the analyzed genes ( Figure S6), we found that setd5 mutant brains specifically express reduced mRNA levels of homer1, encoding a component of the post-synaptic complex [27] (Figure 7a) and dyrk1aa (Figure 7b), which encodes for a kinase involved in various cellular processes including pluripotency maintenance, synapsis function, neuronal differentiation and whose orthologue in humans is located in the Down Syndrome critical region (DSCR) on chromosome 21 [28]. Interestingly, there is a positive correspondence between WT setd5 transcript levels and the expression of gad1a (Figure 7c

Discussion
In this study we have generated the first zebrafish setd5 mutant line using CRISPR/Cas9 technology. We focused on heterozygous mutants that recapitulate human SETD5 haploinsufficiency and characterized their morphological, molecular and behavioral phenotypes.

Discussion
In this study we have generated the first zebrafish setd5 mutant line using CRISPR/Cas9 technology. We focused on heterozygous mutants that recapitulate human SETD5 haploinsufficiency and characterized their morphological, molecular and behavioral phenotypes.
In zebrafish, setd5 is expressed in the developing neural system since its initial specification [15] and its expression persists in many areas of the zebrafish adult brain including areas corresponding to ASD-affected regions such as the lateral zone of the dorsal telencephalic area, which is considered homologous to the mammalian hippocampus and preoptic area [29]. To model SETD5 haploinsufficiency in humans, we generated setd5 mutant zebrafish lines that carry specific deletions leading to the creation of a premature stop codon upstream of exons coding for the SET domain. The significant reduction in total setd5 mRNA observed in the brain of heterozygous mutants compared to sedt5+/+ brain samples suggests that the mutant setd5 transcripts may be subjected to nonsense-mediated decay, as previously demonstrated in a de novo setd5 mutation isolated in a patient with early-onset epileptic encephalopathy [30]. Interestingly, gene expression analysis aimed at looking for potential compensatory mechanisms in the brains of setd5 mutants did not detect differences in the expression level of the setd5-related genes mll5 and setd2, while there was a significant upregulation of nsd1a and nsd1b expression. Interestingly, human NSD1 protein displays mono-and di-methylation activity on the H3K36 residue, which is in turn a target of SETD5 leading to a three-methylated form of the amino acid. This might suggest that enhanced nsd1a and nsd1b expression may partially compensate for the effects of setd5 haploinsufficiency in zebrafish heterozygous mutants.
In addition, we observed that setd5 heterozygous embryos show a reduced body length compared to setd5+/+ ones, while the ratio between eye size and body length indicates that they would be affected by microphthalmia. This is very similar to the morphological phenotype described for zebrafish embryos injected with setd5 morpholino and heterozygous mouse Setd5 mutants [15,16,31]. It also represents a common clinical feature of individuals affected by human 3p25.3 microdeletion syndrome, in which one copy of SETD5 is entirely deleted [9], further supporting the idea that the zebrafish setd5 mutants carry a LoF mutation. Nevertheless, although the adult zebrafish mutants are characterized by a reduced body length and weight, they display a normal eye-head/total length ratio, suggesting a possible rescue of the microphthalmia phenotype which may be linked to the upregulation of nsd1a and nsd1b expression. We also found that, differently from embryos injected with setd5 morpholino [15], zebrafish setd5 mutants do not display microcephaly, neither at embryonic stages, nor during adulthood ( Figures S3 and S4), although we noted an alteration of relative morphological proportions of the telencephalon compared to other brain districts. The differences between setd5 knockdown embryos and setd5+/− embryos may reside in the fact that setd5 morpholino can impair the translation of both maternal and zygotic setd5 transcripts. Accordingly, the effects in setd5 knockdown embryos are possibly stronger than those observed in setd5 heterozygous mutants, which carry one wild type allele and a second allele in which the mutation more specifically affects the zygotic component of setd5 mRNA. In future experiments, it will be interesting to address the role of the setd5 maternal transcript through the generation of maternal-zygotic mutants.
The behavioral studies conducted in this work indicate that our zebrafish setd5 mutant model may effectively replicate behavioral aspects typically altered in individuals heterozygous for SETD5 mutations. An initial analysis in the open field test indicated that setd5 heterozygous larvae and adults display a significant reduction in the distance moved and velocity when compared to setd5+/+ ones. This observation is in line with a significant deficit in motor abilities described for SETD5 haploinsufficient individuals, although this phenotype has not been previously observed in Setd5+/− mice [9,13,15,16]. Similarly, previous studies have demonstrated that the zebrafish inactivation of other genes that play an important role in ASD and ID, such as mecp2 [32] and shank3b [33], also results in reduced locomotor activity. In a shoaling test, aimed at evaluating the robust zebrafish behavior of aggregating and coordinately adapting to each other's movements, we observed that setd5 heterozygous adults exhibit tight shoaling, an altered behavior that is typically related to a condition of anxiety [34]. This is in keeping with the observation that ASD patients carrying SETD5 mutations, as well as Setd5+/− mice, are characterized by increased levels of anxiety [13,20].
This behavior in the shoaling test is not typical of all fish models for ASD. For instance, dyrk1aa fish mutants display a loss of social cohesion [35]. However, mutations in other autism risk genes, such as immp2l and adra1aa, also lead to fish that exhibit tighter shoaling, indicating a convergent social phenotype, at least for a subset of zebrafish ASD models [36]. Furthermore, our data on setd5 behavior, showing that heterozygous fish display a low swimming speed together with a tight shoaling and a low polarization, complement previous observations indicating that high speed correlates with large inter-individual spacing and group polarization [36]. Additional defects in social interactions were further uncovered by the social preference test. In particular, we found that setd5+/− adults appeared to be indifferent to a social stimulus, represented by a new fish introduced during the test. When compared to control fishes, this lack of social interest displayed by mutant fishes is shown by both the reduced time spent in the proximity of the stimulus and the increased distance kept between the tested fish and the stimulus. A similar phenotype is observed in Setd5+/− mice, as well as in other models of ASD, such as dyrk1aa zebrafish mutants [34] and Synapsin knockout mice [37], both of which display abnormalities in sociability. Interestingly, we found that the altered social behavior, observed for setd5 zebrafish mutants in the social preference test, is significantly rescued following treatment with the antipsychotic drug risperidone, which is used to treat irritability in ASD patients. These data suggest that zebrafish setd5 heterozygous mutants may be a valid model for drug screening and to study the molecular mechanisms of action underlying the effects of specific compounds in setd5 haploinsufficiency conditions.
Considering the described activity of Setd5 in controlling gene transcription, we focused on a comparative expression analysis of selected markers for neuronal pathways and synaptic components. Synaptic structure and functionality are the bases of neuronal communication and are critical for normal brain development and functionality, influencing behavior, learning and memory [38]. Although we observed a positive correspondence between the expression of markers of different neurotransmitter pathways and setd5, indicating that setd5 haploinsufficiency affects neuronal signaling, a more consistent effect was found on the expression of genes involved in synaptic structure. Indeed, we found that heterozygous setd5 adult brains show reduced levels of mRNAs encoding for Homer1b, a postsynaptic scaffold protein involved in synaptic plasticity [39] and for Dyrk1aa, a kinase involved in several cellular processes, including synapse function [40]. Moreover, a significant positive correspondence was also observed between the expression of setd5 and Synaptophysin A, a protein of the presynaptic compartment, and shank3b, encoding a synaptic scaffolding protein that interacts with Homer [41]. No effect was observed in the expression of the postsynaptic scaffolding protein PSD95, indicating a specificity of action for setd5. Overall, these data suggest that Setd5 may profoundly influence synaptic structure and function. In keeping with the social phenotype of setd5 mutants, dyrk1aa [35] and shank3b [33] knockout zebrafish, as well as Synapsin knockout mice [42] display altered social behaviors, while Homer 1 knockout mice show behavioral abnormalities related to ASD and Schizophrenia [43]. It is interesting to note that many of the risk genes that have been linked to ID and ASD disorders encode synaptic scaffolding proteins and changes in the expression of any of these proteins significantly affects synaptic strength or number, as well as neuronal connectivity in the brain [44]. The reduced expression of pre-and postsynaptic markers found in both zebrafish and mouse Setd5 mutants [15,16], together with a decrease in the dendritic spine number observed in Setd5+/− mice suggests that brain hypoconnectivity could represent the main alteration caused by setd5 haploinsufficiency that is responsible for the social impairments described. Perspective molecular studies will analyze the possible conservation in zebrafish of SETD5-dependent transcription fidelity during elongation [15]. Indeed, we cannot exclude a possible alteration of splicing processes in zebrafish setd5 mutants.

Zebrafish Care
Zebrafish adults were housed in tanks at a constant temperature of 28 • C on a 14 h light/10 h dark cycle. Zebrafish embryos were obtained by natural mating and maintained at 28 • C in E3 zebrafish medium as previously described [45].

Generation of setd5 Mutant Zebrafish
For the CRISPR/Cas9 gene editing experiment, each zebrafish 1-cell zygote was injected in the cell with 2 nL of solution containing~12.5 ng/µL of gRNA (20 nucleotide sequence complementary to the target: GGGAGACACGAATTCGGCAA),~300 ng/µL of Cas9 mRNA [46] and 0.5% of Phenol Red as a tracer. Successful gene editing in mosaic embryos at 2 days post-fertilization (dpf) was confirmed by Melting analysis, performed using SensiFAST™ HRM Kit (Bioline) according to the manufacturer's instructions in the Corbett Rotor-Gene 6000 machine (Qiagen, Hilden, Germany). To identify the individuals capable of transmitting setd5 gene mutations through their germline, F0 generation adults were outcrossed with wild type individuals and the eventual heterozygous condition of F1 progeny was screened by Melting analysis, followed by mutation identification by Sanger sequencing. Presumptive off-targets were analyzed using Cas-OFFinder (CRISPR Rgen Tool, http://www.rgenome.net/cas-offinder/ (accessed on 26 July 2020)) and the modification of genes including a similar RNA guide target up to 4 mismatches was excluded by Sanger sequencing. Primers used are listed in Tables S2-S4.

Morphological Analysis
Images of zebrafish embryos at 48 hpf were obtained by the stereomicroscope Nikon SMZ1500n with a digital camera CoolSNAP-cf. Adult male fishes (10-months old) of F1 generation were anesthetized, weighted and then photographed by a camera. Eye area and body length were calculated by ImageJ software (RRID:SCR_003070).

Adult Brain Dissection
Each adult zebrafish male (10-12 months of age) was sacrificed, kept on ice, photographed to measure body length, weighed and then placed under a stereomicroscope to dissect the brain. The head was isolated by cutting with a sterile scalpel at the level of the anterior fins. Soft tissues were removed from the ventral side of the skull and eyes with surgical forceps. The skull was then opened, and the brain was transferred into TRIzol ® reagent (Invitrogen, Waltham MA, USA) or quickly photographed to perform brain measurements and then fixed for 6 h in 4% paraformaldehyde (PFA) at 4 • C, successively cryoprotected in 30% sucrose in phosphate buffer saline (PBS) overnight (O/N), sectioned using a cryostat (12 micron-thick sections) and finally collected onto polarized slides (SuperFrost ® Plus; MenzelGläser, Braunschweig, Germany).

Extraction of Total RNA and RT-qPCR
Total RNA was extracted and purified using RNAeasy Plus Mini (Qiagen) according to the manufacturer's instructions. The concentration and purity of total RNA was determined by NanoDrop™. First strand cDNA was synthesized using QuantiTect Reverse Transcription Kit (Qiagen) according to the manufacturer's instructions.
RNA expression levels were evaluated by quantitative reverse transcription-polymerase chain reaction (RT-qPCR) using the SYBR Green method (SensiMix SYBR kit; Meridian, London, UK), following the manufacturer's protocol. Real time PCR and relative quantification of each gene expression was performed essentially as previously described [50]. Primers used are listed in Table S1. The transcript level of examined genes was normalized to b-actin1 mRNA level according to standard procedures.

In Situ Hybridization on Frozen Tissue Sections
Brain samples of 10-months old WT adult male brains were fixed with 4% PFA in PBS for 6 h at 4 • C, then sectioned coronally (12 µm-thick) and collected on polarized slides. In situ hybridization on frozen tissue sections was performed as previously described [51], with some modifications. The setd5 probe was diluted in hybridization buffer at 50 ng/mL and denatured at 85 • C. Cryosections were thawed and washed in PBS and incubated with the setd5 probe at 65 • C O/N. Then, the slides were washed at 65 • C in a solution containing SSC 1X, 50% formamide, 0.1% Tween 20, while the final washing step were conducted in MABT (MAB, 0.1% Tween 20) at RT. After 1 h-long equilibration in blocking solution including MABT, 2% blocking solution (Roche, Basilea, Switzerland) and 20% lamb serum at room temperature (RT), slides were then incubated with anti-DIG Fab fragment conjugated with alkaline phosphatase (Roche, diluted 1:2500 in blocking solution) in a wet chamber at 4 • C O/N. After brief incubation in a buffer to inhibit endogenous alkaline phosphatase (100 mM Tris-HCL pH 9.5, 50 mM MgCl2, 100 mM NaCl, 2 mM levamisole and 0.1% Tween 20), slides were stained in BM purple staining solution (Roche) and placed in the dark at RT. The reaction was stopped by washes in PBST and samples were mounted with Aqua-Poly/Mount. After the staining procedure, images were acquired using stereomicroscope Nikon SMZ1500.

Immunofluorescence and Quantitative Analysis
Immunofluorescence analysis was performed on 24 hpf embryos. In brief, embryos were fixed in 4% PFA for 1 h at RT, cryoprotected in 30% sucrose in phosphate buffer saline (PBS) O/N, successively sectioned using a cryostat (12 µm-thick sections) and collected onto polarized slides. Sections have been washed and successively incubated O/N at 4 • C in a mixture containing both a rabbit polyclonal anti-pHH3 (1:400, cat. No. SC-8656-R, Santa Cruz Biotechnology, Dallas TX, USA) and a mouse monoclonal anti-HuC/D (1:100, cat. No. 16A11; ThermoFisher, Waltham MA, USA) primary antibodies previously used in zebrafish [52,53], diluted in 0.3% Triton X-100 (Merck, Darmstadt, Germany) in PBS. After washes, sections were incubated for 4 h at RT in a solution containing both Oregon Green 488 anti-rabbit (cat. No. 011038; Molecular Probes, Eugene OR, USA) and Alexa Fluor 594 anti-mouse (cat. No. A11032; Invitrogen) secondary antibodies, both diluted 1:500 in 0.3% Triton X-100 (Merck) in PBS. Successively, sections were washed and counterstained for 10 min at RT with 3 µg/mL Hoechst 33258 (Sigma-Aldrich) and finally slides were mounted with Aqua-Polymount (Polysciences Incorporated, Warrington, PA, USA). Adult brains' cryosections (12 µm-thick) were washed, subject only to Hoechst 33258 counterstaining and mounted with Aqua-Polymount (Polysciences Incorporated). Fluorescence images of sectioned 24 hpf embryonic and adult brains were acquired with the microscope Nikon Eclipse Ti at 40× magnification, connected to the digital camera Nikon DS-Ri2, and equipped with the software NIS-Elements AR 5.11.03 (Nikon Corporation, Tokyo, Japan). Measurements were performed on CNS, excluding eyes and ventricles, by the ImageJ software.

Behavioral Analyses
All the behavioral tests were carried out in a quiet room, with a temperature between 27 and 28 • C, using system water. Adult male fishes (around 1-year age) and larvae at 6 dpf were recorded using a high-speed infrared camera (set to 30 frames per second) using Point Grey Fly Cap 2 software [54] or Noldus Media Recorder (Noldus, Wageningen, the Netherlands). All behavioral tests were conducted between 10 a.m. and 5 p.m. Behavioral recordings started after an acclimation period (1 h) to habituate fishes to a new environment.
For the open-field test, 6 dpf larvae were placed, one by one, into a 100 mm Petri dish and allowed to habituate for 1 min and then recorded for 5 min. The open field apparatus for adults consisted of a standard clear plastic fish tank that measured 20 × 22 × 37.5 cm, filled with 10 L of system water. Adult zebrafish were allowed to freely swim inside the tank, and videos were recorded for 5 min without habituating the fish to this setup. Video recordings were analyzed by Ethovision software (Noldus; RRID:SCR_000441) to calculate the total distance swum, the speed and the time spent moving or not moving.
The shoaling test was performed in a standard clear plastic fish tank that measured 20 × 22 × 37.5 cm filled with 10 L system water. A group of 4 fishes was placed in the novel tank and recorded for 10 min without adaptation. Video recordings were analyzed by Zebralab software (ViewPoint, Civrieux, France) to calculate the mean speed of the shoal, shoal polarization, Inter-Individual Distance (IID) and Nearest Neighbor Distance (NND) [31,32].
The social preference test [26] consisted of a plexiglass tank divided into 5 cells. The cells were separated by a transparent divider and water exchange was ensured by the presence of small holes. In the first round, the tested fish was placed into the center of the tank and recorded for 5 min in the presence of one male WT stimulus. In the second round, a new WT male stimulus was additionally introduced, and the experimental fish was recorded again for 5 min. The time spent the different tank zones was quantified using the video tracking software Ethovision software (Noldus; RRID:SCR_000441).
The setup of the social interaction test consisted of a mating tank, to separate the tested fish from the stimulus by a transparent divider. In the first round, the tested fish was placed into the left side of the tank and recorded for 5 min. In the second round, a WT male stimulus was introduced on the right side of the tank and the two fishes were recorded for 5 min. The distances between the nose-point of subjects were quantified by Ethovision software (Noldus; RRID:SCR_000441), using Multiple Body Point Module and the Social Interaction Module.
The social preference test and social interaction test were repeated after treatment with risperidone (Merck). At the onset of the experiment, the stock solution in dimethyl sulfoxide DMSO was diluted in system water to the final concentration of 170 µg/L [26]. Fishes were exposed to risperidone for 15 min prior to behavioral analysis.

Statistical Analysis
After verification of the normal distribution of the data, statistical analysis was performed with one-way analysis of variance (ANOVA) followed by the appropriate post hoc test or Student's t-test, using the software GraphPad PRISM version 6.0 (RRID:SCR_002798). The same software was used to conduct linear regression analysis for gene expression studies. Value of p < 0.05 was considered significant.

Conclusions
Our study underlines the evolutionary conservation of SETD5 activity providing further evidence of its role in regulating molecular, morphological and behavioral aspects underlying the SETD5 haploinsufficiency phenotype. This work also highlights the effects of Setd5 haploinsufficiency on other ASD-risk genes in zebrafish, suggesting potential convergent molecular mechanisms that may be commonly dysregulated in different groups of ASD patients. The heterozygous setd5 mutant that we generated displays changes in social behaviors that are endophenotypes for autism and dysregulated expression of genes encoding crucial synaptic proteins, therefore representing a new model for SETD5 haploinsufficiency. The rescue of the altered social interactions by risperidone is a promising indication of the suitability of this model for drug screening aimed at reversing more specifically the behavioral phenotypes, thus contributing to the development of future therapeutic treatments.  Data Availability Statement: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.