Sigma-1 Receptor Is Critical for Mitochondrial Activity and Unfolded Protein Response in Larval Zebrafish

The sigma-1 receptor (S1R) is a highly conserved transmembrane protein highly enriched in mitochondria-associated endoplasmic reticulum (ER) membranes, where it interacts with several partners involved in ER-mitochondria Ca2+ transfer, activation of the ER stress pathways, and mitochondria function. We characterized a new S1R deficient zebrafish line and analyzed the impact of S1R deficiency on visual, auditory and locomotor functions. The s1r+25/+25 mutant line showed impairments in visual and locomotor functions compared to s1rWT. The locomotion of the s1r+25/+25 larvae, at 5 days post fertilization, was increased in the light and dark phases of the visual motor response. No deficit was observed in acoustic startle response. A critical role of S1R was shown in ER stress pathways and mitochondrial activity. Using qPCR to analyze the unfolded protein response genes, we observed that loss of S1R led to decreased levels of IRE1 and PERK-related effectors and increased over-expression of most of the effectors after a tunicamycin challenge. Finally, S1R deficiency led to alterations in mitochondria bioenergetics with decreased in basal, ATP-linked and non-mitochondrial respiration and following tunicamycin challenge. In conclusion, this new zebrafish model confirmed the importance of S1R activity on ER-mitochondria communication. It will be a useful tool to further analyze the physiopathological roles of S1R.


Results
2.1. Characterization of the s1r +25/+25 Mutant Zebrafish Line S1R was firstly described as a protein with two putative transmembrane domains and two hydrophobic domains [1,40,41]. Its N-ter and C-ter ends, the latter possessing chaperone activity, were predicted to be expressed in the ER lumen. Nevertheless, other groups have suggested a one transmembrane domain with a N-ter luminal and C-ter cytoplasmic localization [42]. Mavlyutov et al. [43] and Zhemkov et al. [9] proposed a one transmembrane domain with a N-ter cytosolic and C-ter luminal localization. As of today, no consensus topology of the protein has been clearly established. The sequence drawn in red shows the amino acids representing new, erroneous, residues predicted to occur after the reading frame shift, followed by a premature stop codon in the C-ter region ( Figure 1A). Using CRISPR/Cas9-mediated mutagenesis, Rennekamp et al. [39] generated a zebrafish line with a 25 bp insertion in exon 2 of the sigmar1 gene, leading to a shift of the reading frame. The mutation was created just upstream of the gene sequence that encodes the most conserved part of the S1R protein. PCR confirmed that the s1r mutants had additional 25 bp in the coding region, thus showing a single band of 237 nucleotides compared to control larvae 212 nucleotides band ( Figure 1B). Mutation was predicted to result in loss of protein function due to the frameshift and the introduction of an early stop codon.
Disruption of both gene and protein expressions were confirmed by quantitative PCR and Western blot analyses. First, qPCR analysis showed that larvae carrying the mutation had a decreased sigmar1 expression level compared to control larvae, confirming that the mutation induced mRNA decay ( Figure 1C). Second, Western blotting ( Figure 1D) and its quantification ( Figure 1E) revealed a drastic decrease of S1R expression in s1r +25/+25 fish compared to controls, validating this zebrafish line as a S1R KO model. pared to control larvae 212 nucleotides band ( Figure 1B). Mutation was predicted to result in loss of protein function due to the frameshift and the introduction of an early stop codon.
Disruption of both gene and protein expressions were confirmed by quantitative PCR and Western blot analyses. First, qPCR analysis showed that larvae carrying the mutation had a decreased sigmar1 expression level compared to control larvae, confirming that the mutation induced mRNA decay ( Figure 1C). Second, Western blotting ( Figure 1D) and its quantification ( Figure 1E) revealed a drastic decrease of S1R expression in s1r +25/+25 fish compared to controls, validating this zebrafish line as a S1R KO model.
Mutant homozygous larvae did not show any overt morphological abnormalities in terms of body size, eye diameter, ear area and anterior and posterior otoliths (Supplementary Figure S1). They grew and became viable and fertile adults. These observations incidentally confirmed that no gross morphological alterations would likely compromise the subsequent behavioral analyses. Figure 1. Characterization of the CRISPR/cas9-induced s1r mutant zebrafish line. (A) Amino acids sequence of the S1R protein with or without mutation. New inserted amino acids are depicted in red. (B) A targeted fragment was amplified by PCR from the genomic DNA of adult zebrafish tail and analyzed after migration on agarose gel. (C) Relative sigmar1 mRNA level revealed by qPCR in s1r +25/+25 and s1r WT zebrafish larvae. Expression of sigmar1 was normalized using the zef1 reference gene. (D) Representative Western blot of whole lysates from wild-type (s1r WT ) or homozygous s1r +25/+25 mutant zebrafish. (E) Relative S1R protein level in the whole larvae at 5 dpf, normalized using total protein expression in the samples (Stain-Free ™ ). Data are shown as the mean ± SEM from n = 12 per group in (C) and n = 4 in (E). * p < 0.05, *** p < 0.001 vs. s1r WT controls; Man n-Whitney's test.

The s1r +25/+25 Larvae Exhibited Increased Locomotor Response
As S1R has been shown to have a role in locomotion [32,44], we first assessed the impact of S1R invalidation on global locomotor activity of the mutant larvae using the visual motor response (VMR) assay. Larvae were placed in individual wells of a 96-well plate and their activity, following sudden changes in light intensity from 100% (ON) to 0% (OFF), was recorded during the whole experiment ( Figure 2A). The distance travelled during the training phase did not differ between control and mutant fish ( Figure 2B). However, significant increases in locomotion were measured for the s1r +25/+25 larvae during both the ON and OFF periods ( Figure 2C,D). Amino acids sequence of the S1R protein with or without mutation. New inserted amino acids are depicted in red. (B) A targeted fragment was amplified by PCR from the genomic DNA of adult zebrafish tail and analyzed after migration on agarose gel. (C) Relative sigmar1 mRNA level revealed by qPCR in s1r +25/+25 and s1r WT zebrafish larvae. Expression of sigmar1 was normalized using the zef1α reference gene. (D) Representative Western blot of whole lysates from wild-type (s1r WT ) or homozygous s1r +25/+25 mutant zebrafish. (E) Relative S1R protein level in the whole larvae at 5 dpf, normalized using total protein expression in the samples (Stain-Free ™ ). Data are shown as the mean ± SEM from n = 12 per group in (C) and n = 4 in (E). * p < 0.05, *** p < 0.001 vs. s1r WT controls; Man n-Whitney's test.
Mutant homozygous larvae did not show any overt morphological abnormalities in terms of body size, eye diameter, ear area and anterior and posterior otoliths (Supplementary Figure S1). They grew and became viable and fertile adults. These observations incidentally confirmed that no gross morphological alterations would likely compromise the subsequent behavioral analyses.

The s1r +25/+25 Larvae Exhibited Increased Locomotor Response
As S1R has been shown to have a role in locomotion [32,44], we first assessed the impact of S1R invalidation on global locomotor activity of the mutant larvae using the visual motor response (VMR) assay. Larvae were placed in individual wells of a 96-well plate and their activity, following sudden changes in light intensity from 100% (ON) to 0% (OFF), was recorded during the whole experiment ( Figure 2A). The distance travelled during the training phase did not differ between control and mutant fish ( Figure 2B). However, significant increases in locomotion were measured for the s1r +25/+25 larvae during both the ON and OFF periods ( Figure 2C,D).
The optokinetic response (OKR) assay allowed us to directly measure visual acuity of the larva by immobilizing them in a methylcellulose solution placed at the center of a cylinder on which black-and-white strips were rotating ( Figure 2E). Analysis of the eye movements showed no difference in the number of saccades between mutant and control larvae at 5 dpf ( Figure 2F). The touch-escape response, measuring a reflex-driven locomotor response, was also measured ( Figure 2G). The distance traveled during 5 s after the tactile stimulation was similar between mutant and control larvae ( Figure 2H). These observations, therefore, indicated that S1R KO zebrafish did present locomotor alteration in the VMR test that is unrelated to visual deficit or spontaneous hyperactivity.
The optokinetic response (OKR) assay allowed us to directly measure visual acuity of the larva by immobilizing them in a methylcellulose solution placed at the center of a cylinder on which black-and-white strips were rotating ( Figure 2E). Analysis of the eye movements showed no difference in the number of saccades between mutant and control larvae at 5 dpf ( Figure 2F). The touch-escape response, measuring a reflex-driven locomotor response, was also measured ( Figure 2G). The distance traveled during 5 s after the tactile stimulation was similar between mutant and control larvae ( Figure 2H). These observations, therefore, indicated that S1R KO zebrafish did present locomotor alteration in the VMR test that is unrelated to visual deficit or spontaneous hyperactivity. (F) Quantification of the number of saccades performed in 2 min. (G) Illustration of the touch escape test. The larva was placed in a rail filled with E3 medium, the tail was touched with a tip and the distance was measured. (H) The distance traveled was measured during 5 s, repeated three times per larva and averaged. Data show mean ± SEM from three replicas. The number of animals is indicated within the columns. * p < 0.05, ** p < 0.01; unpaired t-test.
We performed an immunohistochemical analysis of eye sections from both s1r WT and s1r +25/+25 larvae and observed no grossly observable nor specific morphological abnormalities of ganglion cells ( Figure 3A-C), and red and green cones labeled using Zpr-1 antibody ( Figure 3D,E). Interestingly, the number of rods, labeled using Rho4d2 antibody, decreased significantly in s1r +25/+25 larvae compare to controls ( Figure 3F,G). The constitutive invalidation of S1R in the s1r +25/+25 zebrafish resulted in an impairment of the retinal development, especially in rod cells.
Despite a lack of evidence for a role of S1R on auditory mechanisms, a study previously suggested that the selective S1R agonist cutamesine protected mouse hearing from (F) Quantification of the number of saccades performed in 2 min. (G) Illustration of the touch escape test. The larva was placed in a rail filled with E3 medium, the tail was touched with a tip and the distance was measured. (H) The distance traveled was measured during 5 s, repeated three times per larva and averaged. Data show mean ± SEM from three replicas. The number of animals is indicated within the columns. * p < 0.05, ** p < 0.01; unpaired t-test.
We performed an immunohistochemical analysis of eye sections from both s1r WT and s1r +25/+25 larvae and observed no grossly observable nor specific morphological abnormalities of ganglion cells ( Figure 3A-C), and red and green cones labeled using Zpr-1 antibody ( Figure 3D,E). Interestingly, the number of rods, labeled using Rho4d2 antibody, decreased significantly in s1r +25/+25 larvae compare to controls ( Figure 3F,G). The constitutive invalidation of S1R in the s1r +25/+25 zebrafish resulted in an impairment of the retinal development, especially in rod cells.
Despite a lack of evidence for a role of S1R on auditory mechanisms, a study previously suggested that the selective S1R agonist cutamesine protected mouse hearing from severe noise trauma [45]. Consequently, to rule out any potential neurosensorial deficits, we tested the hearing ability of s1r +25/+25 mutant larvae. Analysis of the quantity of movement of the larvae in the acoustic startle response (ASR) assay was measured at 5 dpf (Supplementary Figure S2A). Movement during training and baseline phases (Supplementary Figure S2B,C) increased significantly in s1r +25/+25 larvae compared to the controls. In a coherent manner but in a different assay, hyperlocomotion was observed in the ASR assay. However, S1R invalidation had no impact on quantity of movement after the different stimulations, which suggests that the invalidation of S1R did not result in an impairment of ear development (Supplementary Figure S2D). severe noise trauma [45]. Consequently, to rule out any potential neurosensorial deficits, we tested the hearing ability of s1r +25/+25 mutant larvae. Analysis of the quantity of movement of the larvae in the acoustic startle response (ASR) assay was measured at 5 dpf (Supplementary Figure S2A). Movement during training and baseline phases (Supplementary Figure S2B,C) increased significantly in s1r +25/+25 larvae compared to the controls. In a coherent manner but in a different assay, hyperlocomotion was observed in the ASR assay. However, S1R invalidation had no impact on quantity of movement after the different stimulations, which suggests that the invalidation of S1R did not result in an impairment of ear development (Supplementary Figure S2D).

S1R Invalidation Modulated the Expression of ER Stress Genes, in Resting and Tunicamycin-Induced Pathological Conditions
S1R is a key player in the ER stress response, acting as a modulator for the induction of ER stress pathways [4,9,[46][47][48]. To investigate if S1R impacts ER stress modulation, we analyzed the mRNA expression levels of ER stress factor genes in the basal condition ( Figure 4) and after a tunicamycin challenge ( Figure 5A) in the whole zebrafish larvae. Variations are summarized in Figure 5B.
tions are summarized in Figure 5B.
In the control condition, the expression levels of the unfolded protein response (UPR) inducers, bip and hsp90b1, were unaffected in mutant larvae compared to controls. However, the expression levels of a large number of the UPR primary effectors related to the IRE1 and PERK pathways were consistently down-regulated in mutant larvae compared to controls. First, ire1 and its secondary effectors, xbp1s and xbp1us, were highly significantly down-expressed ( Figure 4A). Second, perk and its secondary effectors eif2s1 and atf4β were also significantly down-regulated. Third, atf6 was unchanged and chop mildly but significantly over-expressed ( Figure 4A). At the protein level, Western blot analyses were done using the few specific antibodies available in zebrafish and showed that the level of Bip and p-Eif2α/Eif2α were significantly decreased in s1r +25/+25 larvae compared to controls, while Chop level was unchanged ( Figure 4B,C). (C) Representative Western blots. zeif2α level and Stain-Free ™ were used as a loading control in qPCR and Western blot analyses, respectively. Data are expressed as mean ± SEM, normalized to s1r WT control level with n = 5-9 in each group. * p < 0.05, ** p < 0.01, *** p < 0.001; unpaired t-test. (C) Representative Western blots. zeif2α level and Stain-Free ™ were used as a loading control in qPCR and Western blot analyses, respectively. Data are expressed as mean ± SEM, normalized to s1r WT control level with n = 5-9 in each group. * p < 0.05, ** p < 0.01, *** p < 0.001; unpaired t-test.
In the control condition, the expression levels of the unfolded protein response (UPR) inducers, bip and hsp90b1, were unaffected in mutant larvae compared to controls. However, the expression levels of a large number of the UPR primary effectors related to the IRE1 and PERK pathways were consistently down-regulated in mutant larvae compared to controls. First, ire1 and its secondary effectors, xbp1s and xbp1us, were highly significantly down-expressed ( Figure 4A). Second, perk and its secondary effectors eif2s1 and atf4β were also significantly down-regulated. Third, atf6 was unchanged and chop mildly but significantly over-expressed ( Figure 4A). At the protein level, Western blot analyses were done using the few specific antibodies available in zebrafish and showed that the level of Bip and p-Eif2α/Eif2α were significantly decreased in s1r +25/+25 larvae compared to controls, while Chop level was unchanged ( Figure 4B,C).
These results suggested that S1R invalidation in zebrafish could lead to a deregulation of the ER stress response as it impacted the expression level of numerous effectors in basal condition particularly related to the IRE1 and PERK pathways. Tunicamycin, by provoking the accumulation of unfolded proteins, is a strong ER stress inducer and is commonly used as a cellular stressor [49]. When s1r WT or s1r +25/+25 larvae were treated at 4 dpf with 2 µg/mL of tunicamycin for 24 h, a strong ER stress was monitored ( Figure 5A). The mRNA levels of the different markers were drastically increased in mutant and control larvae. First, s1r was moderately but significantly increased after tunicamycin in s1r WT larvae and, as expected, absent in s1r +25/+25 larvae ( Figure 5A). Second, the levels of the UPR inductors bip and hsp90b1 were highly significantly increased. Third, all three UPR pathways were also mobilized with significant increases of ire1 and its effectors xbp1s and xbp1us, perk and its effectors atf4α and atf4β, and atf6 ( Figure 5A). Finally, chop level was highly significantly increased ( Figure 5A) as well.
Interestingly, the genotype had a major impact on the ER stress response to tunicamycin treatment, with increased over-expression of bip, ire1, xbp1s, xbp1us, atf4β, and atf6, in s1r +25/+25 larvae as compared to s1r WT larvae ( Figure 5A,B). These observations, decreased levels of IRE1 and PERK effectors in basal condition and increased over-expression of not only IRE1 and PERK effectors but also ATF6 in the ER stress condition, suggests that S1R invalidation altered the UPR response and regulation of ER stress. Interestingly, the alteration is coherent with a compensatory over-reaction of the pathways, resulting in a mild impact on Chop level. challenge on ER stress gene expression levels in s1r +25/+25 zebrafish larvae at 5 dpf. The selected genes were analyzed using cDNA prepared from whole zebrafish larvae and relative mRNA expression were expressed as percentage of s1r WT + 0.1% DMSO controls. The fold change from 0.1% DMSO treatment is indicated. Data are expressed in mean ± SEM of n = 5 determination in each group. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. DMSO treatment; ° p < 0.05, °° p < 0.01, °°° p < 0.001 vs. s1r WT line; two-way ANOVA followed by unpaired t-test. (B) Schematic summary of ER stress pathway alterations observed in basal condition or after tunicamycin treatment in s1r +25/+25 zebrafish line.
These results suggested that S1R invalidation in zebrafish could lead to a deregulation of the ER stress response as it impacted the expression level of numerous effectors in basal condition particularly related to the IRE1 and PERK pathways. Tunicamycin, by challenge on ER stress gene expression levels in s1r +25/+25 zebrafish larvae at 5 dpf. The selected genes were analyzed using cDNA prepared from whole zebrafish larvae and relative mRNA expression were expressed as percentage of s1r WT + 0.1% DMSO controls. The fold change from 0.1% DMSO treatment is indicated. Data are expressed in mean ± SEM of n = 5 determination in each group. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. DMSO treatment; • p < 0.05, •• p < 0.01, ••• p < 0.001 vs. s1r WT line; two-way ANOVA followed by unpaired t-test. (B) Schematic summary of ER stress pathway alterations observed in basal condition or after tunicamycin treatment in s1r +25/+25 zebrafish line.

S1R Invalidation Altered Mitochondrial Bioenergetics
S1R is enriched in MAMs [1] and facilitates Ca 2+ flux from the ER to the mitochondria [1,3]. Recent studies have shown that an alteration in the cross-talk between these two organelles affects the bioenergetics of the cells [46,[50][51][52]. Moreover, S1R activity directly impacts mitochondrial oxidative respiration [53]. We hypothesized that the absence of S1R would disturb the mitochondrial activity in the mutant larvae. We therefore measured the oxygen consumption rate (OCR) before and after addition of complex inhibitors or decoupling agents of the respiratory chain, namely oligomycin, FCCP, antimycin A and rotenone, and analyzed several parameters of the mitochondrial respiration ( Figure 6A). Basal respiratory rate, ATP production-related and non-mitochondrial respiration ( Figure 6B,C,F) were significantly decreased in s1r +25/+25 mutant larvae at 5 dpf, compared to control larvae, while the maximal respiration rate and proton leak remained unchanged ( Figure 6D,E). These results confirmed in the zebrafish line a specific role of S1R on bioenergetics and particularly ATP production-related mitochondrial oxidative respiration. When s1r WT or s1r +25/+25 larvae were treated at 4 dpf with 2 μg/mL of tunicamycin for 24 h, mitochondrial metabolism was affected ( Figure 7A). Basal respiration, ATP production and proton leak were highly decreased in s1r WT ( Figure 7B,C,E), showing an impact of ER stress on mitochondrial respiration. Non-mitochondrial respiration was increased while maximal respiratory capacity was unchanged by tunicamycin in s1r WT (Figure 7D,F). Interestingly, the ER stressor failed to decrease basal respiration and ATP productionrelated respiration in s1r +25/+25 larvae below the level measured in V-treated controls (Figure 7B,C). Maximal respiration, proton leak and non-mitochondrial respiration levels were equally affected in s1r WT and s1r +25/+25 lines ( Figure 7D-F), showing a S1R-dependent impact of ER stress on mitochondrial energetics. When s1r WT or s1r +25/+25 larvae were treated at 4 dpf with 2 µg/mL of tunicamycin for 24 h, mitochondrial metabolism was affected ( Figure 7A). Basal respiration, ATP production and proton leak were highly decreased in s1r WT ( Figure 7B,C,E), showing an impact of ER stress on mitochondrial respiration. Non-mitochondrial respiration was increased while maximal respiratory capacity was unchanged by tunicamycin in s1r WT (Figure 7D,F). Interestingly, the ER stressor failed to decrease basal respiration and ATP productionrelated respiration in s1r +25/+25 larvae below the level measured in V-treated controls ( Figure 7B,C). Maximal respiration, proton leak and non-mitochondrial respiration levels were equally affected in s1r WT and s1r +25/+25 lines ( Figure 7D-F), showing a S1R-dependent impact of ER stress on mitochondrial energetics.

Discussion
In the present study we characterized a novel zebrafish line carrying a mutated S1R protein. The mutation resulted in the loss of the protein C-terminal tail, known to be carrying the activity [1]. The s1r +25/+25 zebrafish line is therefore constitutively devoid of the S1R chaperone activity that allows a pleiotropic modulatory action on several intracellular pathways and interorganelle dialogs [4,49]. Our first observation was that the s1r +25/+25 zebrafish developed from the larval to the adult stage with no grossly observable morphologic or functional defects. Moreover, the fish reproduced well under homozygous breeding, confirming an absence of major neuroendocrine or physiological alteration. A similar observation was made in murine KO models. Both the Sigmar1 tm1Lmon mouse line developed by Esteve [44] and the Sigmar1 Gt(IRESBetageo)33Lex line developed by Lexicon [54,55] did not show developmental defects and reproduced normally. The availability of a S1R KO zebrafish line will be of major use for future research, particularly since: (i) zebrafish embryos are nearly transparent which allows easy examination of the development of internal structures; (ii) the model is particularly prone to in vivo imaging and electrophysiological approaches; and (iii) eggs are fertilized and developed outside the mother's body, which makes zebrafish an ideal model organism for studying early development [56]. Indeed, S1R may play a major yet underexplored role during early development of many organs, including the brain. For instance, S1R was shown to regulate the formation of

Discussion
In the present study we characterized a novel zebrafish line carrying a mutated S1R protein. The mutation resulted in the loss of the protein C-terminal tail, known to be carrying the activity [1]. The s1r +25/+25 zebrafish line is therefore constitutively devoid of the S1R chaperone activity that allows a pleiotropic modulatory action on several intracellular pathways and interorganelle dialogs [4,49]. Our first observation was that the s1r +25/+25 zebrafish developed from the larval to the adult stage with no grossly observable morphologic or functional defects. Moreover, the fish reproduced well under homozygous breeding, confirming an absence of major neuroendocrine or physiological alteration. A similar observation was made in murine KO models. Both the Sigmar1 tm1Lmon mouse line developed by Esteve [44] and the Sigmar1 Gt(IRESBetageo)33Lex line developed by Lexicon [54,55] did not show developmental defects and reproduced normally. The availability of a S1R KO zebrafish line will be of major use for future research, particularly since: (i) zebrafish embryos are nearly transparent which allows easy examination of the development of internal structures; (ii) the model is particularly prone to in vivo imaging and electrophysiological approaches; and (iii) eggs are fertilized and developed outside the mother's body, which makes zebrafish an ideal model organism for studying early development [56]. Indeed, S1R may play a major yet underexplored role during early development of many organs, including the brain. For instance, S1R was shown to regulate the formation of dendritic spines in hippocampal neurons and the knockdown of S1R by siRNAs provoked a deficit in the formation of dendritic spines [57,58]. Moreover, S1R regulates proper tau phosphorylation and axon extension that are primary events required for neuroplasticity and neurodevelopment [59]. The use of the s1r +25/+25 zebrafish line will now foster in vivo studies on the developmental role of S1R in an integrated vertebrate organism.
The behavioral analyses of the s1r +25/+25 larvae first revealed a hyperlocomotor response in the VMR test that appeared likely unrelated to visual or other neurosensorial modification. This phenotype trait is not shared by S1R KO mouse lines, since WT or KO mice failed to show basal hyperlocomotion compared to WT littermates, particularly when tested from 8 to 48-weeks of age and in different open-field or maze tests [44,55,56,60]. Locomotor behaviors, including walking in mice or swimming in zebrafish, are generated by networks of neurons known as central pattern generators located in the spinal cord of vertebrates. Morphological and physiological considerations on locomotor behaviors in zebrafish have been recently remarkably reviewed by Berg et al. [61]. In particular, it appears that central pattern generators involve: (i) motoneurons that activate muscle; (ii) ipsilateral excitatory interneurons that provide excitatory drive; (iii) inhibitory commissural interneurons that ensure left-right alternation; and (iv) ipsilateral inhibitory interneurons that contribute to burst termination [61]. S1R is highly expressed in motoneurons [62], and S1R activity has been shown to protect motoneurons from insults or neurodegeneration [15,[63][64][65]. Furthermore, S1R is expressed in interneurons [66] and alteration of S1R activity directly impacts interneurons activity. For instance, in a mouse model of ALS shown to provoke a loss of S1R in lumbar motoneurons [63], choline acetyltransferase expression accumulated in the soma of motoneurons with a drastic diminution of efferences toward Renshaw interneurons [63]. It is therefore not surprising that a constitutive loss of S1R activity during zebrafish development rapidly led to locomotor alterations in larva.
Interestingly, the observation that the touch-escape response was unaltered in the s1r +25/+25 larva, as compared to s1r WT control larva, suggested that the hyperlocomotor response was restricted to the spinal swim neuronal network, but did not affect more complex integrated networks. This was unexpected, as the touch-escape behavior was one of the first behavioral responses that allowed monitoring in vivo in zebrafish of the Ca 2+ rise associated with single action potentials in the dendrites, soma and nucleus [67]. A massive activation of the motoneuron pool, and a differential activation of populations of hindbrain neurons, were elicited by the escape response and, as S1R activity has been shown to be an important physiological modulator of intracellular Ca 2+ mobilization [3], a behavioral impact could be measured. However, the touch-escape behavior involves a relay of sensory information to populations of reticulospinal neurons, including the command-like Mauthner neuron, which in turn excites target neurons in the spinal cord. This connectivity implements a dynamic behavioral response that transiently interrupts ongoing swimming, directing the fish away from the threat [61]. Considering the major ethological importance of the behavior, it is likely that the neuronal network adapted to the developmental absence of the S1R intracellular modulatory effect in either motoneurons or interneurons.
Since S1R has been shown to be expressed in RGCs, PRs, RPE cells and surrounding the soma of cells in the inner nuclear layer [68], we analyzed the effect of the loss of S1R in the presence of RGCs and PRs. Notably, in s1r +25/+25 larva, a significant decrease of the number of RGCs and in the size of the rods outer segments was observed. This data confirms the important role of S1R in retinal physiology. Nevertheless, our VMR experiment failed to detect any deficit, suggesting that the loss of retinal cells is not sufficient to induce physiological alteration. In order to determine if lack of S1R leads to retinal deficit, a more precise investigation measuring ERG [69] in our model would be useful. S1R activity is crucial at MAMs, where it simultaneously modulates the induction of ER stress pathways by releasing Bip [1] and chaperoning IRE1 [5], and (ii) mitochondrial homeostasis, as previously detailed. Koshenov et al. [46] recently provided an elegant demonstration in human neuroblastoma SH-SY5Y cells that ER stress is a physiological signal promoting S1R-induced increase in mitochondrial ATP production. The authors propose that S1R activity orchestrates Ca 2+ leak from the ER to promote mitochondrial bioenergetics and maintain a balanced metabolism of reactive oxygen species during early stress event. The latter was induced by a limited 2-h exposure of the cells to Tunicamycin [46]. Indeed, we observed a marked deregulation of ER stress response in s1r +25/+25 zebrafish line, with very decreased expression of the genes associated with the IRE1 and PERK pathways, namely, ire1 and perk, but also of their effectors xbp1s/us and eif2s1/atf4β, respectively. These decreased levels were related to increased responses to the Tunicamycin challenge. Indeed, exaggerated increases following the stressor treatment were observed for ire1, xbp1s/us, atf4β and even for atf6, suggesting that the ATF6 pathway is mobilized to compensate for IRE1 and PERK. S1R activity has repeatedly been shown to be related to IRE1 and PERK pathways in physiopathological conditions [5,70,71]. These deregulations are coherent with a major role of S1R in the ER stress response, and more generally in the regulation between the ER stress response and mitochondrial bioenergetics. The s1r +25/+25 zebrafish line will, therefore, be a valuable tool for the in vivo validation of these mechanistical hypotheses, and to better apprehend the crucial role of S1R in ER-mitochondrial communication as suggested by recent studies [46,53,71]. Finally, we demonstrated that S1R is able to modulate the expression level of BiP without affecting its mRNA. This interesting observation suggests that S1R controls the expression level of BiP at a post-translational level, highlighting a novel mode of regulation of the first orchestrator of the UPR, BiP. Other experiments are needed to decipher how the lack of S1R leads to the BiP protein deregulation and its fundamental role in the control of the UPR in basal and stressed conditions.
Analyses of the mitochondrial oxidative respiration of s1r +25/+25 zebrafish line showed significant decreases, particularly in basal respiration capacity and in the respiration linked to the production of ATP. Basal respiration is the respiration used to meet the endogenous ATP demand of the cell and drive the proton leak pathway. ATP-linked respiration estimates the respiration that is used to drive mitochondrial ATP synthesis. The latter is the sum of the ATP utilization, ATP synthesis and substrate supply and oxidation. By facilitating Ca 2+ transfer into the mitochondria, S1R activity contributes to the proper functioning of the tricarboxylic acid (TCA) cycle through facilitation of pyruvate dehydrogenase activity. The TCA cycle provides NADH cofactor to the complexes of the respiratory chain. In turn, S1R activity was shown to facilitate oxidative respiration in basal physiological conditions in brain cells [47] as well as in cardiac cells [68]. S1R activity particularly increased complex I activity, which is a NADH-dependent ubiquinone oxidoreductase, the activity of which is also Ca 2+ -dependent and directly modulated by mitochondrial calcium uniporter (MCU) activity [53,70,72]. Therefore, in zebrafish, the data confirmed that S1R activity is not a requisite to mitochondria physiology but appears necessary, as S1R invalidation resulted in a −26% decrease in basal respiration and −34% in the respiration related to the formation of ATP. Induction of ER stress by Tunicamycin resulted in significant alterations of the different parameters related to oxidative respiration, as previously observed [46,73,74] that were progressive and observed at shorter exposure times (2 h, 4 h, data not shown). Interestingly, loss of S1R abolished the decrease of basal and ATP-linked OCR after Tunicamycin treatment, suggesting that S1R was necessary for the adaptation of mitochondrial physiology to ER stress. However, how S1R interferes with the modulation of basal and ATP-linked respiration requires further investigation. Our data confirmed that S1R activity plays an important role in the physiopathological regulation of mitochondrial bioenergetics [46,53] and confirmed the potentialities of the zebrafish model for the in vivo studies of these aspects.

Zebrafish and Maintenance
The present study followed the recommendations of the ARRRIVE guidelines [75] and the European Union Directive 2010/63. The zebrafish s1r +25/+25 line was a gift from Drs Jing-Ruey Joanna Yeh and Randall T. Peterson (Cardiovascular Research Center, Massachusetts General Hospital, Charlestown, MA, USA) [39]. Briefly, the fish were generated using CRISPR-cas9, creating a 25bp insertion into exon 2 of the sigmar1 gene. Adult zebrafish were bred and maintained under standard conditions in an automated fish tank system (ZebTEC, Tecniplast, West Chester, PA, USA) at 28 • C, pH 7, conductivity around 500 mS and with a 14 h:10 h light:dark cycle. Eggs were obtained by natural spawning and maintained in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl 2 , 0.33 mM MgSO 4 , 0.05% methylene blue) at 28 • C. Unfertilized embryos were discarded and the water was changed only the first day. Larvae were euthanized by an extended immersion in an ice bath for biochemical experiments. Each experimental procedure was carried out in triplicate and larvae originated from three different crosses.

Genotyping
For isolated homozygous s1r WT and s1r +25/+25 zebrafish, genomic DNA was extracted from the tail fins at 2 months of age and the tissue was lysed in 50 mM NaOH at 95 • C for 1 h. A 237 bp region of sigmar1 gene, encompassing the 25 bp insertion, was amplified by PCR using GoTaq Green Master Mix (Promega, Madison, WI, USA). The primers were as follows: sigmar1, 5 -ATAGGTCAGGATCATGAGCAGG-3 (forward); 5 -TTATGACCTGAATGTCCACCGG-3 (reverse). The DNA fragments were separated by electrophoresis on a 3% agarose gel and the genotype analyzed. The control amplicon should give one band of 237 bp while the mutant should give one band of 212 bp. The mutation was confirmed by Sanger sequencing.

Chemical Treatment
To induce ER stress, larvae were incubated at 4 dpf for 24 h with 2 µg/mL of Tunicamycin (sc-3606, Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted in 0.1% DMSO. Control larvae were treated with 0.1% DMSO diluted in E3 medium.

RT-PCR and Quantitative Real-Time PCR (qPCR)
Total RNA from 20 pooled whole s1r WT or s1r +25/+25 larvae at 5 dpf were extracted using a Nucleospin ® RNA kit (Macherey-Nagel, Hoerdt, France) according to the manufacturer's instructions. Concentration and purity were evaluated using the Agilent RNA 6000 Nano ® Kit (Agilent Technologies, Santa Clara, CA, USA). RNA samples (1 µg/µL) were denatured 5 min at 70 • C and reverse transcribed into cDNA, 1 h at 37 • C, using M-MLV reverse transcriptase (Promega). Primer sequences are detailed in the Supplementary Table S1. Control reactions were performed with sterile water to determine signal background and DNA contamination. The standard curve of each gene was confirmed to be in a linear range, while zef1α gene was used as a reference.

Visual Motor Response (VMR) Assay
The VMR assay quantifies the locomotor activity of the zebrafish larvae to light intensity changes. The protocol was previously described (Crouzier et al., 2021). In brief, 5 dpf larvae were transferred in a 96-well plate (#7701-1651, Whatman, Maidstone, UK) with 300 µL E3 medium and placed in a Zebrabox ® (ViewPoint, Lissieu, France). The locomotor behavior was monitored using an automated video-tracking device equipped with an infra-red (IR) camera under IR light illumination (Zebralab ® , Viewpoint). The light protocol was as follow: 30 min of acclimatization in the dark (0% light intensity), then two cycles of 10-min duration light ON and 10-min duration light OFF periods. The brightness changes were immediate (<<1 s). The activity was measured in mm/s. For each larva, the mean (OFF-ON) values were calculated to account for inter and intragroup variability in locomotor response.

Acoustic Startle Response (ASR) Assay
The ASR assay quantifies the locomotor activity of zebrafish larvae to a sound stimulus. The ASR test was also performed in the ZebraBox ® . Experimental conditions were similar as those used for the VMR. The protocol consisted first in acclimating larvae during 30 min in silence (35 dB ambient), followed by a 1-s stimulation with a white noise at 90 dB repeated three times with an intertrial time interval of 5 min. The amount of movement was measured for each larva. To normalize the values, baseline activity, i.e., the activity recorded during 2 min before each stimulation, was subtracted from the post stimulation activity levels.

Optokinetic Response (OKR) Assay
The zebrafish larvae were immersed per group of four in a Petri dish (35 mm diameter) containing 2.5% methylcellulose (#9004-65-3, Sigma Aldrich, St. Louis, MO, USA). Larvae were placed dorsal up and in a X shape, to avoid touching and interfering with each other. All measurements were done between 2:00 and 6:00 p.m. The room temperature was 28 • C and the light was off. Visual system performance of larval zebrafish was assessed using a videotracking device (VisioBox ® , ViewPoint). Forty-six mm wide black-and-white strips were projected at 2 rpm for 1 min clockwise and then 1 min counterclockwise. Eye saccades were recorded for 2 min at 25 frames/s using a FL3-U3-32S2M 1/2.8-inch monochrome IR camera (FLEA3, FLIR). The number of eye saccades was recorded (PHIVisualize software) and data expressed as number of saccades per 2 min.

Touch Response
Larvae were individually placed at one extremity of a rail (18 cm × 0.4 cm) filled with 200 µL of E3 medium. The room temperature was 28 • C and the light was on. The larva tail was briefly touched with a tip and the escape distance measured during 5 s. The same procedure was repeated three times with an intertrial time interval of 1 min to reduce stress. Data was expressed as the averaged distance swam per larva.

Morphological Analyses
After prolonged exposure to ice, larvae were completely immobilized in a 2.5% methylcellulose solution contained in a petri dish. Larvae size and eye area were measured with ImageJ software v1.46 (NIH, Bethesda, MD, USA) on an image taken with a stereo microscope (Olympus, Tokyo, Japan), at ×3.2 magnification. Ear and otoliths area were measured at ×6.3 magnification.

Cell Counting
Cone cells, immunolabeled with Zpr-1 antibody were quantified individually and the total area of rod outer segments, immunolabeled with Rho4d2 antibody, was evaluated. Both measures were normalized to the length of the associated retina. The number of apoptotic cells was assessed by directly counting cells labeled with the anti-cleaved caspase-3 antibody per entire retina section. Ganglion cells, highlighted by DAPI counterstaining, were counted in three specific regions of the retina, consistent from one sample to another. The total number of ganglion cells from all regions was averaged per larva. The thickness of the ganglion cells layer was also quantified for each distinct region of the retina.

Seahorse XF Cell Mito Stress Test
The oxygen consumption rate (OCR) of 5 dpf larvae was measured with a Seahorse XFe24 extracellular flux analyzer (Agilent), according to [76]. The Seahorse temperature space was maintained at 28 • C. The larvae were placed individually in a well of a Seahorse XFe24 spheroid microplate, in 500 µL of E3 medium. A grid was placed on the larvae to maintain them at the bottom of the wells throughout the experiment. Two wells were kept empty per experiment and were considered as the "blank" condition. At 4 dpf, the larvae were treated for 24 h with tunicamycin (2 µg/mL) diluted in E3 medium and placed at 5 dpf in the measurement plate without treatment. Control larvae were treated with 0.1% DMSO. Measurements of total zebrafish OCR were performed according to the manufacturer's instructions. Four basal analysis cycles were recorded, then five cycles recorded after administration of oligomycin (25 µM), five cycles after administration of carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) (8 µM) and nine cycles after rotenone + antimycin A (15 µM). Calculations of the different OCR parameters (nonmitochondrial respiration; basal respiration; maximal respiration; proton leak; ATP production) were done according to the Seahorse XF Cell Mito Stress Test Kit user guide (Agilent).

Statistical Analyses
Data were expressed as mean ± SEM. Statistical significance between groups was determined by a two-way ANOVA, with zebrafish genotype and treatment as independent factors, and/or an unpaired Student's t-test (for parametric data) or a Mann-Whitney's test (for nonparametric data). All two-way ANOVA results are detailed in the Supplementary  Table S2. The levels of statistical significance considered were p < 0.05, p < 0.01 and p < 0.001. Statistical analyses were performed using the GraphPad Prism v7.0 software.

Conflicts of Interest:
The authors declare that the work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.