A single base pair substitution in zebrafish distinguishes between innate and acute startle behavior regulation

Abstract

Behavioral thresholds define the lowest stimulus intensities sufficient to elicit a behavioral response. Establishment of baseline behavioral thresholds during development is critical for proper responses throughout the animal’s life. Despite the relevance of such innate thresholds, the molecular mechanisms critical to establishing behavioral thresholds during development are not well understood. The acoustic startle response is a conserved behavior whose threshold is established during development yet is subsequently acutely regulated. We have previously identified a zebrafish mutant line (escapist) that displays a decreased baseline or innate acoustic startle threshold. Here, we identify a single base pair substitution on Chromosome 25 located within the coding sequence of the synaptotagmin 7a (syt7a) gene that is tightly linked to the escapist acoustic hypersensitivity phenotype. By generating animals in which we deleted the syt7a open reading frame, and subsequent complementation testing with the escapist line, we demonstrate that loss of syt7a function is not the cause of the escapist behavioral phenotype. Nonetheless, escapist mutants provide a powerful tool to decipher the overlap between acute and developmental regulation of behavioral thresholds. Extensive behavioral analyses reveal that in escapist mutants the establishment of the innate acoustic startle threshold is impaired, while regulation of its acute threshold remains intact. Moreover, our behavioral analyses reveal a deficit in baseline responses to visual stimuli, but not in the acute regulation of responses to visual stimuli. Together, this work eliminates loss of syt7a as causative for the escapist phenotype and suggests that mechanisms that regulate the establishment of behavioral thresholds in escapist larvae can operate independently from those regulating acute threshold regulation.

Introduction

Behavioral thresholds are critical for organisms to respond to relevant environmental cues or threats, while ignoring irrelevant environmental stimuli [1–3]. A prime example of this is the acoustic startle response, a defensive and highly conserved behavior across vertebrates that exhibits a quantifiable behavioral threshold [4]. Across species, organisms establish innate baseline thresholds which are shaped during development [5–7]. Baseline startle thresholds can subsequently be modulated acutely through sensory stimulation (e.g. through habituation or sensitization), yet provided with enough time will return to the baseline threshold. Differences in baseline startle thresholds across genetic strains as well as evidence of generational inheritance strongly suggest the existence of genetic components that establish baseline thresholds [8, 9]. Our current knowledge regarding the molecular and circuit mechanisms underlying the regulation of the startle threshold largely stems from a wealth of studies focusing on acute regulation, such as during short term habituation and pre-pulse inhibition [10–13]. Dysregulation of both acute modifications and developmental mechanisms have been associated with several neuropsychiatric and neurodevelopmental disorders, such as in schizophrenia, autism spectrum disorders and anxiety disorders [14–17]. While there has been significant progress in understanding the mechanisms that acutely regulate the startle response, the molecular and circuit mechanisms underlying the establishment of the innate threshold of the acoustic startle response are only partially understood.

Zebrafish are a powerful system to identify molecular and circuit mechanisms that regulate behavioral thresholds. By 5 days post fertilization (5dpf), zebrafish larvae exhibit a wide variety of well characterized behaviors, including the acoustic startle response [7, 18]. Additionally, forward genetic screens have identified genes that regulate behavioral thresholds to acoustic and visual stimuli, emphasizing the power of this system to dissect genetic and molecular mechanisms of disease-relevant behaviors [12, 19–21]. Finally, cell types that mediate the initiation of the startle response have been identified, providing a coherent framework to explore mechanisms that distinguish between innate and acutely regulated thresholds [22, 23].

Here we examine the process of establishing behavioral thresholds using escapist, a zebrafish mutant line identified in a forward genetic screen based on its hypersensitivity to acoustic stimuli [21]. Using RNA sequencing we uncovered a single candidate gene, synaptotagmin 7a, linked to the escapist acoustic hypersensitive phenotype. Although tightly linked to the mutant phenotype, thorough behavioral analysis in combination with molecular genetic testing strongly suggests that loss of synaptotagmin 7a does not cause the escapist acoustic hypersensitive phenotype. However, through our extensive behavioral analysis, we uncovered an additional behavioral phenotype in escapist mutants. Specifically, we find that mutants exhibit reduced motor responsiveness to dark flashes, a stereotyped behavioral response to sudden darkness. Finally, we find that unlike the innate threshold, acute threshold modulation of visual and acoustic stimuli in escapist mutants is unaffected, consistent with the idea that mechanisms that regulate the establishment of innate thresholds can operate independently from those regulating acute thresholds.

Results

Identification of candidate genes underlying the escapist acoustic startle hypersensitivity phenotype

By 5dpf, zebrafish larvae respond to sudden and intense acoustic stimuli with a stereotypic short latency turn known as the startle response [24]. The probability of the all-or-none startle response increases as the acoustic stimulus intensity increases [21]. We previously performed a forward genetic screen and identified five zebrafish mutant lines that exhibit hypersensitivity of the acoustic startle response [21]. To compare the strength of the hypersensitivity phenotype across wildtype and mutant animals we exposed larvae to various stimulus intensities and plotted the larvae’s probability to perform a short latency startle response (SLC) based on stimulus intensity. We define the area under the resultant curve as the sensitivity index [21]. Specifically, for the escapist mutant line, we established crosses between escapist heterozygous carriers and divided the offspring larvae into two groups based on their sensitivity indices: putative mutants (top 25%) and putative siblings (bottom 75%). When compared to the top 25% most responsive larvae obtained from wildtype incrosses, putative escapist larvae have significantly increased sensitivity (Fig 1A and 1B). Importantly, the bottom 75% (least responsive) escapist larvae are within range of the responsiveness of all wildtype larvae, consistent with recessive inheritance of the hypersensitivity phenotype (Fig 1A and 1C).

Fig 1. RNA sequencing linkage analysis identifies a single base pair change on Chromosome 25 that is tightly linked to the escapist hypersensitive phenotype.

(A) Acoustic sensitivity curve for escapist and wildtype (WIK) 5-day old larval zebrafish. X-axis represents stimulus intensity (dBu) and Y-axis represents percent short latency (SLC) startles. Larvae were placed in the following groups: the top 25% responsive larvae from the escapist line (escapist- Top 25%, n = 8 larvae), the bottom 75% responsive larvae from the escapist line (escapist -Bottom 75%, n = 23 larvae), the top 25% responsive larvae from the wildtype WIK line (Wildtype- Top 25%, n = 8 larvae), the bottom 75% responsive larvae from the wildtype WIK line (Wildtype- Bottom 75%, n = 22 larvae). (B) The area under the sensitivity curve was computed to generate the sensitivity index for each individual larva represented in Fig 1A. Sensitivity index for the top 25% responders from the escapist line (red circles) and wildtype WIK lines (black squares). Unpaired t-test for escapist Top 25% vs wildtype Top 25%: p = 0.0001. (C) Sensitivity index for the bottom 75% escapist responders (red circles) and all responders from the wildtype WIK lines (black squares). Unpaired t-test for escapist Bottom 75% vs wildtype All responders: p = 0.3380 (ns). (D) Schematic of linked region on Chromosome 25 identified through RNA Sequencing linkage analysis using zv9 as a reference genome. Analysis of linked mutations identified one single nucleotide base pair change (chr25:4531251bp T->C or p404) as highly linked to the escapist hypersensitive phenotype. Sequence shown is of the reverse strand. (E) Acoustic sensitivity curve for 5dpf larvae from crosses of escapist carriers genotyped for the p404 lesion. Homozygous wildtype (+/+, n = 39 larvae) curve shown in black, heterozygous (+/p404, n = 50 larvae) shown in blue, homozygous mutant (p404/ p404, n = 36 larvae) shown in red. (F) Sensitivity index for escapist line sensitivity curves shown in 1E using p404 for genotyping. Kruskal-Wallis test performed with Dunn’s test for multiple comparisons. WT vs. Mut: p<0.0001; Het vs Mut: p<0.0001; WT vs Het: p = 0.8723.

To identify the gene causative for the escapist behavioral phenotype we took an unbiased, genome-wide approach to identify single nucleotide polymorphisms (SNPs) genetically linked to the escapist hypersensitive phenotype. For this, we pooled a group of phenotypically mutant escapist larvae (pool size = 50 larvae) and a separate group of phenotypically wildtype siblings (pool size = 50 larvae) derived from the same genetic cross and performed RNA sequencing analysis [21]. This analysis identified Chromosome 25 to be linked to the hypersensitivity phenotype [21]. To identify potentially causal mutations on Chromosome 25, we reasoned that the mutation causing the mutant phenotype would be unique to the escapist line. Therefore, we searched for single base pair mutations present within the escapist line but absent in ten other wildtype or mutant pools on which we previously performed RNA sequencing. We prioritized single base pair mutations with read percentages close to the read percentages predicted for recessive inheritance. Since the phenotype is inherited recessively (Fig 1A–1C), the causal mutation should be found in 100% of mutant reads and approximately 33% of sibling reads. Furthermore, we prioritized single base pair mutations predicted to result in missense or nonsense mutations in protein coding regions of genes. Based on these criteria, we identified a single nucleotide polymorphism (SNP) within the synaptotagmin7a (syt7a) gene as the sole candidate satisfying all our criteria (Fig 1D). Mutant reads of this SNP comprised 39% (9/23) of the escapist wildtype sibling pool and 93% (27/29) of the escapist mutant pool. Finally, the SNP was not observed in any other line we had previously sequenced (0/371 reads from ten mutant, sibling, and wildtype pools). To independently confirm that the SNP (referred to as p404) is linked to the escapist phenotype, we performed the acoustic startle sensitivity assay on subsequent generations of the escapist line and genotyped for p404 after behavioral testing. When compared to larvae that are genotypically heterozygous or wildtype, p404 homozygous larvae exhibit increased acoustic startle sensitivity (Fig 1E and 1F). Together, these data strongly suggest that the SNP within syt7a (p404) is tightly linked to the escapist hypersensitive phenotype.

syt7a mutant alleles complement the escapist mutant allele

synaptotagmin 7a is a member of the synaptotagmin gene family encoding calcium sensing proteins involved in vesicle release and replenishment [25]. In addition, synaptotagmin 7 proteins have also been shown to mediate asynchronous vesicle release [26, 27]. Structurally, synaptotagmin proteins are composed of a transmembrane region, a linker region, and two calcium domains (C2A and C2B). In zebrafish, exon 2 of syt7a encodes a transmembrane domain and the last four exons encode two calcium binding domains, C2A and C2B (Fig 2A) [28]. In other vertebrates, the synaptotagmin 7 genes encodes multiple splicing isoforms, with at least three isoforms present in mouse and in humans [29]. Two alternatively spliced syt7a isoforms are predicted in zebrafish and expression of both has been confirmed within the zebrafish retina [28]. We named the isoforms syt7aα and syt7aβ, consistent with the nomenclature of syt7a isoforms present in mice (Fig 2B). The difference between the two protein isoforms lies in the length of the linker region connecting the transmembrane region to the calcium binding domains, with syt7aβ having a longer linker domain. Alignment of the syt7aβ splicing isoforms reveals that the linker region affected by p404 is highly conserved across vertebrate species (Fig 2C), with the p404 SNP located within exon 5 of syt7a resulting in the conversion of a highly conserved tyrosine to a histidine (Fig 2C). The syt7aα isoform is predicted to be unaffected by p404.

Fig 2. CRISPR-Cas9 generated syt7a mutant alleles complement the escapist p404 mutation.

(A) Schematic of syt7a gene. Boxes indicate exons. Exon 2 (dark blue) encodes the transmembrane region of the Syt7a protein. Exons 4–7 are alternatively spliced to produce linker domains of varying lengths. The last four exons encode for the two calcium binding domains, C2A and C2B. Star represents the location of the p404 lesion within the syt7a gene. Arrows indicate regions that were targeted with CRISPR guides for generation of independent mutant alleles. (B) Schematic illustrating the protein structures of two syt7a splicing isoforms: syt7aα and syt7aβ as well as the protein structure resulting from the p404 base pair change within the escapist line. The red line indicates where the missense mutation is located within the syt7aβ isoform. (C) Clustal Omega alignment of the Syt7β protein isoform across different vertebrates. Asterisks (*) represent amino acids that are fully conserved across vertebrates, whereas colons (:) represent amino acids that are similar. Zebrafish^p404 results in an amino acid change of a highly conserved Tyrosine (Y) to a Histidine (H) in the syt7aβ isoform. (D) Depiction of predicted protein structures from CRISPR-Cas9-generated syt7a mutant lines. syt7a^67bpinsert (p431) and syt7a^34bpdel (p432) encode premature stops resulting in a predicted truncated protein of both wildtype isoforms of syt7a. The entire syt7a locus is deleted in syt7a^wldel (p433), which is predicted to produce no protein in the resulting mutant line. (E) Sensitivity curve for both siblings (n = 15 larvae, black line) and p432 homozygous mutant larvae (n = 6 larvae, red line). X-axis represents the stimulus intensity presented to larvae; Y-axis represents the average percent startles performed by larvae in each group. (F) The area under the curve calculated from the sensitivity curve of both siblings and p432 homozygous mutant individual larvae and plotted as the sensitivity index (AU). Mann-Whitney test was used to compare siblings to p432 mutants (p = 0.9257). (G) Sensitivity Index for complementation testing between escapist line and syt7a^67bpinsert (p431) allele. Homozygous WT (+/+, n = 9 larvae), +/p431 (n = 10 larvae), +/p404 (n = 22 larvae), p404 / p431 (n = 9 larvae). Groups were compared by using a one-way ANOVA with Tukey’s test for multiple comparisons. (H) Sensitivity Index for complementation testing between escapist line and syt7a^34bpdel (p432) allele. Homozygous WT (+/+, n = 14 larvae), +/p432 (n = 16 larvae), +/p404 (n = 8 larvae), p404 /p432 (n = 17 larvae) Groups were compared by using a one-way ANOVA with Tukey’s test for multiple comparisons. (I) Sensitivity Index for complementation testing between escapist line and syt7a^wldel (p433) allele. Homozygous WT (+/+, n = 8 larvae), +/p433 (n = 8 larvae), +/ p404 (n = 4 larvae), p404 /p433 (n = 8 larvae). Groups were compared by using a Kruskal-Wallis test with Dunn’s test for multiple comparisons.

To test whether the escapist hypersensitive phenotype is caused by loss of syt7a gene function, we used CRISPR/Cas9 genome editing to generate three independent syt7a mutant alleles (Fig 2D). We generated these syt7a CRISPR mutant alleles in the WIK wildtype background, as this is the background in which the escapist allele is maintained. Sequencing of the mutant alleles revealed that syt7a^67bpinsert (p431) and syt7a^34bpdel (p432) result in a 67bp insert and a 34bp deletion within exon 2 respectively, with both resulting in a premature stop codon and truncation of the transmembrane domain (Fig 2D and Table 1 and S1 Fig). Syt7a^wldel (p433) results in a deletion of the entire syt7a open reading frame and is predicted to be a protein null allele (Fig 2D and Table 1 and S1 Fig). All three mutations are predicted to affect both the syt7aα and syt7aβ isoforms. To determine whether the CRISPR/Cas9 -generated syt7a mutants themselves exhibit changes to their baseline threshold, we performed the sensitivity assay on p432 larvae. Compared to their wildtype siblings, homozygous p432 mutants do not exhibit a difference in their sensitivity curves or their sensitivity index (Fig 2E and 2F)

Table 1. Predicted protein sequence for CRISPR generated syt7a mutant alleles.

CRISPR generated syt7a alleles	Protein Sequence	Total Amino Acids
Wildtype syt7aβ protein Sequence	MYLNREEEYSKGSISLSVLLVSLAVT….QWHALKA*	653
p431 (syt7a67bpinsert) predicted protein sequence	MYLNREEEYSKGSISLSVLLVSWWR*	25
p432 (syt7a34bpdel) predicted protein sequence	MYLNREEEYSKGSISLSVLLVAVTVASVDGVNASWGRGINPEWSPSGLRIQEEEEGRKKPSMI*	63
p433 (syt7awldel) predicted protein sequence	No amino acids predicted to be synthesized	0

Bold letters indicate erroneous amino acids introduced into the syt7a protein sequence. The asterisks (*) indicates a stop codon

Finally, we crossed the CRISPR/Cas9 derived syt7a mutant alleles to the escapist line and assessed startle sensitivity on 5dpf larvae. In all three cases, larvae trans heterozygous for p404 and a CRISPR-Cas9-generated allele did not exhibit a higher sensitivity index compared to their wildtype or singly heterozygous siblings (Fig 2G–2I). Thus, the syt7a CRISPR mutant alleles genetically complement the p404 allele. This data further suggests that while the SNP p404 in the syt7a gene is tightly linked to the escapist phenotype, loss of syt7a function is not causative of the escapist hypersensitivity phenotype.

Zebrafish syt7a is dispensable for larval acoustic, visuomotor, and visual behavior

Across various species, including mice, birds, and zebrafish, syt7 is expressed ubiquitously and is enriched in the brain. Specifically, syt7 is expressed in auditory hair cells and the retina and moreover, syt7 knockout mice exhibit changes in auditory pre-pulse inhibition and anxiety-like behaviors, suggesting a potential role for syt7 in acoustically and/or visually-evoked behaviors [28, 30–32]. To test whether syt7a regulates auditory or visually evoked behaviors in zebrafish, we subjected syt7a^34bpdel (p432) mutants to previously validated behavioral paradigms including the visuomotor response [33], the dark flash response [34, 35], the light flash response [35], and the acoustic startle response [24, 36]. In all acoustic assays, p432 homozygous mutant larvae performed indistinguishably from their wildtype siblings (Figs 2E and 2F and 3Ai and 3Aii and S2 Fig), suggesting that both baseline and acute regulation of the acoustic startle response in syt7a mutants is intact. Furthermore, we detected no difference in baseline response or kinematic response to dark flashes (Fig 3Bi–3Biii), or for any tested parameters of the visuomotor response, the dark-flash response, and the light-flash response (S2 Fig). Taken together, our data indicates that in 6dpf larvae syt7a is dispensable for a broad range of acoustic, visual and visuomotor behaviors.

Fig 3. syt7a mutant larvae do not exhibit altered behavioral phenotype at 6 days post fertilization.

(A) Acoustic startle short term habituation assay. i) Percent startle of both siblings (n = 20 larvae, black) and p432 homozygous mutant larvae (n = 9 larvae, red) at baseline (10 acoustic stimuli with 40 sec interstimulus interval), as well as during stimuli that induce habituation (acoustic stimulus presented with 1sec interstimulus interval). Average percent startle is reported for stimulus number 1–10, 11–20, 21–30. Any larvae that had an average probability of startle of less than 60% during baseline stimuli were omitted from habituation analysis. ii) Habituation index calculated for both siblings and p432 homozygous mutant larvae. Habituation index is calculated using the following formula: % Habituation = [1- ((%Startle Stimulus 21–30)/ (%Startle Baseline))] *100. Mann-Whitney test was used to compare siblings to p432 mutants (p = 0.7871). (B) Analysis of dark flash response for sibling larvae (n = 74 larvae) and p432 homozygous mutant larvae (n = 20 larvae). i) O-bend probability, ii) average O-bend latency and iii) average distance traveled during O-bend are shown. For statistical purposes, values were normalized to sibling responses, and student’s t-test was performed to compare mutants and sibling groups with Bonferroni for multiple comparisons of 86 parameters analyzed during behavioral testing (See S2 Fig for all parameters tested).

escapist p404 larvae exhibit baseline changes to visual stimuli

Finally, we asked whether behavioral deficits in escapist mutants were limited to the developmental, innate startle threshold, or whether escapist mutants also exhibited deficits in acute threshold regulation. Using p404 as a genotyping marker for escapist mutants, we utilized an array of behavioral assays including previously validated acoustic and visual behavioral paradigms to test for the presence of additional behavioral phenotypes in escapist mutants [37]. Consistent with our previous findings, escapist larvae homozygous for p404 exhibit acoustic startle hypersensitivity as indicated by an increased sensitivity index (Fig 1E and Fig1F and S2 Fig), yet display no difference in acute regulation of the startle response as assayed by pre-pulse inhibition (S2 Fig) and habituation (Fig 4Ai and Fig 4Aii). While p404 homozygous mutant escapist larvae are not significantly different after Bonferroni correction in dark flash parameters such as in the probability of response to the dark flash or average latency in response (Fig 4Bi and Fig 4Bii), we did observe significant decreases in the average distance traveled and in the average displacement during dark flash responses (Fig 4Biii and S2 Fig). Given that escapist mutants exhibit wildtype-like kinematic parameters of the acoustic startle response [21], this suggests differential roles for escapist in regulating baseline responses to acoustic compared to visual behaviors. Additionally, we do not observe differences in dark flash habituation in p404 homozygous mutant escapist larvae (S2 Fig), suggesting that visual habituation remains intact. Overall, detailed behavioral characterization reveals that rather than exhibiting a global increase in responsiveness, escapist p404 larvae exhibit an increased response probability selectively to acoustic stimuli. Importantly, changes were only observed in baseline responsiveness but not in assays that measure acute threshold regulation, underscoring the selectiveness of the behavioral phenotypes observed in escapist p404 larvae, and hence the specificity of the process disrupted in these animals.

Fig 4. escapist p404 larvae exhibit increased acoustic startle sensitivity and decreased visual responses at 6 days post fertilization.

(A) Acoustic startle short term habituation assay: i) Average probability of startle for siblings (n = 29 larvae, black) and p404 homozygous mutant larvae (n = 23 larvae, red) in escapist line at baseline (10 acoustic stimuli with 40 sec interstimulus interval), as well as during stimuli that induce habituation (acoustic stimulus presented with 1sec interstimulus interval). Average probability of startle is grouped by stimulus number 1–10, 11–20, 21–30. Any larvae that had an average probability of startle of less than 0.6 during baseline stimuli were omitted from habituation analysis. ii) Habituation index calculated for both siblings and p404 homozygous mutant larvae. Habituation index is calculated using the following formula: % Habituation = [1- ((%Startle Stimulus 21–30)/ (%Startle Baseline))] *100. (B) Dark flash response for sibling larvae (n = 95 larvae) and p404 homozygous mutant larvae (n = 34 larvae). i) O-bend probability, ii) average distance traveled and iii) average distance traveled are shown. For statistical purposes, values were normalized to sibling responses, and student’s t-test was performed to compare mutants and sibling groups with Bonferroni used for multiple comparisons of 86 parameters analyzed during behavioral testing (See S2 Fig for all parameters tested). P-values that are significant after analysis and statistical correction are indicated with asterisks (***).

Discussion

The genetic regulators that underlie establishment of innate behavioral thresholds and acute regulation of behavioral thresholds are still not fully understood. Additionally, how the genetic, molecular, and circuit mechanisms overlap and/or diverge is an outstanding question in the field of genetics and behavior. From a forward genetic screen we had previously identified five mutant lines that display hypersensitivity to acoustic stimuli. We previously showed that one of these mutant lines harbors a mutation in the cytoskeletal regulator cyfip2 [21]. Here we focus on a second mutant line from the genetic screen, the escapist line. Through RNA sequencing analysis, we identified a SNP (p404) that is tightly linked to the escapist mutant phenotype. p404 is located within the syt7a gene, a zebrafish homolog of the synaptotagmin 7 (syt7) gene. Through genetic complementation, we determined that loss of syt7a is likely not causative of the escapist hypersensitive phenotype. However, by using p404 as a tool to select for escapist mutants, we identify that escapist larvae exhibit baseline changes to auditory and visual behaviors, but not in acute regulation of these behaviors.

Synaptotagmin 7 plays roles in vesicle release and fusion [38], long term potentiation, and synaptic facilitation [39, 40]. Syt7 knock out mice exhibit several behavioral deficits, including a deficit in auditory pre pulse inhibition [32]. In syt7a mutant zebrafish larvae, we failed to detect any changes in auditory or visual behaviors. The lack of behavioral changes within the syt7a mutant larvae is unlikely to be a result of differences in genetic background between escapist and the syt7a mutant alleles. The syt7a CRISPR mutant alleles were generated within a WIK wildtype zebrafish background, the same wildtype background in which escapist is maintained, thus minimizing the probability that the genetic background is preventing detection of a change in behavior in syt7a mutants. Although genetic background may not explain why we do not see a phenotype within the syt7a mutants, it is possible that deficits are instead masked by genetic redundancy or compensation within the syt7a mutants. In fact, two syt7 paralogs, syt7a and syt7b, have been identified in zebrafish. Morpholinos targeting syt7b have effects on asynchronous release of vesicles in motor neurons [27]. Yet, whether syt7b can partially compensate for the loss of syt7a remains unclear. In drosophila loss of syt7 has a dosage specific effect on synaptic vesicle release, consistent with the idea that expression levels of synaptotagmin 7 are important for its function [41]. Thus, future studies involving the analysis of double mutant combinations of the synaptotagmin family members will enable further exploration of the role of syt7 in regulating zebrafish behavior.

Through genetic complementation and behavioral testing, we determined that the escapist acoustic hypersensitivity phenotype is not due to loss of syt7a. It is possible that p404 does not result in a loss of syt7a function. For example, rare, recessive antimorphic mutant alleles can result in multiple copies of the mutant allele interfering with the function of related genes and proteins [42]. Knockdown of the mutant syt7a protein in escapist p404 mutants could determine if the escapist phenotype is caused by a recessive antimorph. Alternatively, it is also possible that syt7a is not the causative gene. For example, SNPs in potentially causative genes that are expressed at low levels at 5–6 dpf, the developmental age at which we collected our larval samples for RNA sequencing, would not have been included within our results. Alternatively, mutations resulting in nonsense mediated decay would reduce expression of the causative gene in escapist mutants and preclude detection through our pipeline. Future work analyzing gene expression changes between escapist mutants and siblings could identify the causative gene. Furthermore, analysis of gene expression changes in escapist mutants could provide insight into molecular pathways affected, establishing an entry point to study the molecular mechanisms involved in establishing behavioral thresholds.

By utilizing p404 as a genetic marker for escapist mutant larvae, we determined that the gene causing these behavioral changes in escapist is involved in a) establishing the baseline threshold for the acoustic startle response and b) involved in regulating kinematic features of baseline responses to dark flash stimuli. The gene causing these phenotypes in escapist may establish the acoustic startle threshold in wildtype larvae through cell types within the acoustic startle circuit. The Mauthner cells are reticulospinal neurons necessary and sufficient to activate contralateral motor neurons and initiate the acoustic startle response [43]. Regulatory neurons which synapse onto the Mauthner cells, such as the feedforward excitatory spiral fiber neurons, are important for the establishment of the acoustic startle threshold [44]. In cyfip2 mutants, which also exhibit decreased acoustic startle threshold, spiral fiber activity and recruitment is increased during normally subthreshold acoustic stimuli [21]. Calcium imaging on spiral fiber neurons at normally subthreshold acoustic levels in escapist larvae would determine if escapist larvae exhibit a similar increased activity in spiral fiber neurons. Alternatively, calcium imaging or activity mapping of other cell types within the startle circuit, such as auditory hair cells which provide acoustic input to the Mauthner cells, or feedforward glycinergic neurons which synapse onto the Mauthner, would allow identification of cell types that may be affected in the escapist line. Determining the cell types affected in the escapist line will help provide insight into the cellular mechanisms involved in establishing the acoustic startle threshold.

The escapist line, along with the five other mutant lines that exhibit a lowered acoustic startle threshold, still exhibit intact acute regulation of the acoustic startle threshold [21]. The existence of mutants that solely affects establishment of the startle response highlights that there is a subset of genetic regulators that are necessary for establishment but dispensable for acute regulation. There are additionally zebrafish mutant lines that exhibit a defect in both acute regulation (habituation) and in establishment of the innate startle threshold [12]. Genetic mapping has identified the genes affected in several of these mutants, including the pregnancy associated plasma protein pappaa, the palmitoyltransferase hip14, as well as calcium voltage-gated channel subunit cacna2d3 [12, 20, 45]. While these genetic mutants exhibit both changes to establishment and acute regulation of the startle threshold, whether they regulate both processes through common or independent mechanisms remains unclear. Future work characterizing the cell types and molecular mechanisms involved in the phenotypes across these mutants will further clarify the convergent and divergent mechanisms underlying maintenance and regulation of the startle response.

Our work has identified and characterized a useful genetic marker for identifying acoustically hypersensitive mutants within the escapist zebrafish line. Additionally, by utilizing this marker we have identified that escapist mutants show changes to establishment of acoustic and visual behaviors, while acute regulation of these behaviors in escapist mutants remain intact. Together, our findings support the idea that molecular and circuit mechanisms that regulate establishment of the acoustic startle response and dark flash responses can be independent of the mechanisms that acutely regulate these behaviors. Overall, our work lays the groundwork for determining the molecular and circuit pathways involved in establishment of baseline thresholds and furthers our understanding of the differences between acute regulation and innate behavioral thresholds.

Materials and methods

Ethics statement

All animal protocols were approved by the University of Pennsylvania Institutional Animal Care and Use Committee (IACUC).

Zebrafish husbandry

Larvae were raised at 28–29 degrees Celsius on a 14-hour light: 10-hour dark cycle in E3 media.

RNA sequencing analysis

RNA sequencing data for the escapist allele collected in Marsden et al, 2018 was further analyzed to identify candidate mutations for the escapist mutant line in the linked region Chromosome 25 [21]. Given that the forward genetic screen was performed using ENU mutagenesis, we focused on single base pair mutations within Chromosome 25 for potential candidate mutations. Causative mutations are expected to be recessive and therefore are expected to comprise close to 100% of reads within the escapist mutant pool and approximately 33% in the sibling pool. Priority was given to mutations that were predicted to have deleterious changes to coding sequences of genes, such as nonsense, missense mutations, or changes to splicing donor or acceptor sites.

Generation of CRISPR derived syt7a alleles

Guides targeting syt7a were designed by using CHOP CHOP and IDT software. Guides 2 and 3 were designed to target exon 2, which encodes the transmembrane domain of syt7a protein, while guides 1 and 4 were designed to target the 5’ and 3’ UTR (Table 2). WIK wildtype embryos were co-injected with either guides 2 and 3 to target the transmembrane domain, or with guides 1 and 4 for a whole locus deletion. Injected G0 embryos were raised to adulthood. G0 fish were then outcrossed, F1 larvae were collected and screened to identify mutant carriers. F1 mutant larvae were raised to adulthood, outcrossed and F2 larvae grown to generate stable mutant lines. F2 carriers from each stable line were then incrossed and gDNA was collected from F3 homozygous mutants. F3 gDNA was then amplified using primers flanking the CRISPR targeted regions, TOPO cloned, and Sanger sequenced to confirm each mutation. The genotyping primer sequences used to sequence these syt7a mutant alleles can be found in Table 3.

Table 2. CRISPR guide sequences.

Guide Number	syt7a target	Sequence
1	5’ UTR	TGCTCGCGCGTAATTCTGGGAGG
2	Exon 2	CGTGCTGCTGGTGTCGTTGGCGG
3	Exon 2	CGTTGGCGGTAACTGTGTGTGGG
4	3’ UTR	GCTGCCAGGCGGCACAACACGGG

Table 3. Primer sequences for genotyping.

Mutation	Protocol	Forward/Reverse Primer Sequence	Product from Wildtype Sequence	Product from Mutant Sequence
escapist p404	KASP	Sequences Proprietary- LGC Genomics	N/A	N/A
p431 (syt7a 67bpinsert )	PCR	CTCCATCTCTTTGAGCGTGC/ CAGCTTGCGTTGACACCA	104 bp	171 bp
p432 (syt7a34bpdel)	PCR	CTCCATCTCTTTGAGCGTGC/ CAGCTTGCGTTGACACCA	104 bp	70bp
p433 (syt7a wldel )	PCR	GCGGGGCAATGTCTCAGAG/ GCAGTGCTCCAATCAGGGTTC	N/A	469bp

Behavioral analysis

Zebrafish larvae were behaviorally tested between 4–6 dpf.

Acoustic startle hypersensitivity assay

Larvae were presented with 6 different stimulus intensities, each presented 5–10 times with an interstimulus interval of 40 seconds. Larval responses to the stimuli were recorded using a high-speed camera (Photron Fastcam MiniUX) at 1000 frames per second. Videos were then tracked using FLOTE v2.0 and Batchan to analyze behavior of each individual larvae as previously described [24]. The average percent of escape responses performed by larvae was recorded at each stimulus intensity. The area under each curve was then calculated using PRISM software to obtain the sensitivity index.

Acoustic startle habituation assay

Larvae were presented with 40 acoustic stimuli total. During the pre-habituation phase, larvae were presented with 10 stimuli with a 40 second interstimulus interval. During the habituation phase, larvae were presented with 30 stimuli with a 1 second interstimulus interval. Stimuli were grouped into 4 different bins: baseline response, habituation stimuli 1–10, habituation stimuli 11–20, habituation stimuli 21–30. The average percent of larvae responding with escape responses were recorded during each habituation bin. Larvae that did not respond with over 80% startles at baseline were omitted from further analysis. Habituation index is calculated using the following formula: % Habituation = [1- (%Startle Stimulus 21–30)/(%Startle Baseline)] *100.

Complementation testing

Independently generated syt7a heterozygous F2 fish were crossed to escapist carriers heterozygous for the p404 single base pair mutation. Larvae were raised to 5dpf and tested on the acoustic hypersensitive assay. Larvae were collected after testing and genotyped for the independently generated syt7a mutant allele and for p404 (Table 3).

Broad behavioral assay at larval stages

Broad behavioral assay at 6dpf larval stage was performed as described in Campbell et al, 2023 [37]. Briefly, 6 days post fertilization larvae were placed in individual wells in a 10x10 well acrylic testing arena. The following behavioral tests were performed in the following order: Visual Motor Response (Light ON followed by Light OFF), Light Flash, Dark Flash, Acoustic Startle Response Behaviors. Approximately 50 larvae from each cross were tested per trial. Larvae were genotyped after behavior was performed. A total of 86 parameters were tested across the various behavioral paradigms. For statistical purposes, values were normalized to sibling responses, and student’s t-test was performed to compare mutants and sibling groups with Bonferroni used for multiple comparisons of 86 parameters analyzed during behavioral testing.

Visuomotor response. Larvae were placed in the testing arena and allowed to acclimate for 30 minutes before testing. Larvae were then recorded for 8 minutes in light (Light ON), and 8 minutes in the dark (Light OFF). Video was recorded at 20 frames per second.

Light Flash: Larvae were presented with 15 500ms white light stimuli after previously being in dark conditions. Each stimulus was separated by an interstimulus interval of 30 sec. Responses were recorded at 500 frames per second.

Dark Flash Baseline: Larvae were presented with 15 1sec dark flash stimuli after previously being in light conditions for 5 minutes. Each stimulus was separated by 30 sec interstimulus intervals. Responses were recorded at 500 frames per second.

Dark Flash Habituation: Larvae were presented with 4 blocks of 14 1 sec dark flash stimuli after Dark Flash baseline stimuli. Dark flash habituation stimuli were separated by 10 second interstimulus intervals. The first and third block of dark flash habituation stimuli were unrecorded, while the second and fourth block of dark flash habituation were recorded at 500 frames per second.

Acoustic Behaviors: Larvae were presented with the following stimuli (recorded at 500 frames per second): Pre-pulse Inhibition: low intensity pulse followed by high intensity pulse with inter-stimulus interval of 50ms, High intensity stimuli (10 times), Acoustic habituation: High intensity acoustic stimuli delivered 30 times with 1 sec ISI.

Statistical analysis

Graph generation and statistical analysis was performed using a combination of Python as well as Graphpad Prism (www.graphpad.com). The D’Agostino and Pearson test was used to assess normality. If data was not normally distributed, the Mann-Whitney or Kruskall-Wallis test with Dunn’s test for multiple corrections was used to compare groups. If data was normally distributed, the students t-test or one-way ANOVA with Tukey’s test for multiple corrections was used to compare groups. Statistical analysis of data from the broad behavioral assay was performed as described in the behavioral methods.

Supporting information

S1 Fig. Sanger sequences for CRISPR generated syt7a mutant alleles.

Chromatographs for A) p431 (syt7a^67bpinsert), B) p432 (syt7a^34bpdel), and C) p433 (syt7a^wldel) aligned to syt7a wildtype sequence. The wildtype syt7a sequence is indicated with lowercase bold letters. The mutated sequence is indicated with uppercase letters. The mutations of p431 and p432 are within exon 2 of the syt7a gene. The mutations for p433 are within the 5’UTR and 3’UTR of the syt7a and excise the entire coding syt7a sequence between these locations.

(TIF)

S2 Fig. Comprehensive behavioral analysis of escapist (p404) and syt7a^34bpdel (p432) at 6 days post fertilization.

a) Heat map summary of the results from our broad behavioral assay for escapist (p404) line and syt7a^34bpdel (p432) line. Different behavioral assays are designated by block: VMR Light ON (light gray box), VMR Light OFF (dark gray block), Light Flash (yellow block), Dark Flash (green block), and Acoustic Startle Response (red block). Each box within a block represents a different parameter tested during each assay. Responses from homozygous mutants were normalized to responses from their respective siblings (WT and Hets combined). The colors of the heatmap represent the Z-score, or the difference between the normalized average response from homozygous mutants compared to the average responses of their respective siblings in standard deviations, with red signifying mutants had a larger response than siblings and blue signifying mutants had a smaller response than siblings. A student t-test was used with a Bonferroni correction to correct for the multiple comparisons performed in the broad behavioral assay. Parameters that had a significant p-value after the Bonferroni correction are indicated with the pound sign (#) within the box.

(TIF)

S1 Data

(XLSX)

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

“This work was supported by grants to M.G (NIH National Institute of Neurological Disorder and Stroke: R01NS118921, R01NS097914 and National Eye Institute R01EY024861), E.A.O. (NIH National Institute on Deafness and other Communication Disorders: 5T32DC016903), P.D.C. (NIH National Institute of Mental Health: T32MH019112 and R25MH119043) and J.C.N. (NIH National Institute of Neurological Disorder and Stroke: K99NS111736 and NINDS R00NS111736). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.”.