A CRISPR-Cas9 mutation in sox9b long intergenic noncoding RNA (slincR) affects zebrafish development, behavior, and regeneration

Abstract

The role of long noncoding RNAs (lncRNAs) regulators of toxicological responses to environmental chemicals is gaining prominence. Previously, our laboratory discovered an lncRNA, sox9b long intergenic noncoding RNA (slincR), that is activated by multiple ligands of aryl hydrocarbon receptor (AHR). Within this study, we designed a CRISPR-Cas9-mediated slincR zebrafish mutant line to better understand its biological function in presence or absence of a model AHR ligand, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The slincR^osu3 line contains an 18 bp insertion within the slincR sequence that changes its predicted mRNA secondary structure. Toxicological profiling showed that slincR^osu3 is equally or more sensitive to TCDD for morphological and behavioral phenotypes. Embryonic mRNA-sequencing showed differential responses of 499 or 908 genes in slincR^osu3 in absence or presence of TCDD Specifically, unexposed slincR^osu3 embryos showed disruptions in metabolic pathways, suggesting an endogenous role for slincR. slincR^osu3 embryos also had repressed mRNA levels of sox9b—a transcription factor that slincR is known to negatively regulate. Hence, we studied cartilage development and regenerative capacity—both processes partially regulated by sox9b. Cartilage development was disrupted in slincR^osu3 embryos both in presence and absence of TCDD. slincR^osu3 embryos also displayed a lack of regenerative capacity of amputated tail fins, accompanied by a lack of cell proliferation. In summary, using a novel slincR mutant line, we show that a mutation in slincR can have widespread impacts on gene expression and structural development endogenously and limited, but significant impacts in presence of AHR induction that further highlights its importance in the developmental process.

The role of long noncoding RNAs (lncRNAs) in health effects is increasingly gaining prominence in the field of toxicology. lncRNAs are <200 nucleotide-long transcripts without any functional open reading frames (ORFs) and therefore, do not have the capability of translation into functional proteins. However, they can regulate expression of other genes or can interact with chromatins, proteins, and other RNA molecules, playing diverse roles in various physiological processes, including development (Statello et al., 2021). The role of several lncRNAs has been implicated in diseases such as cancer, cardiovascular diseases, and neurological disorders (Lekka and Hall, 2018). Experimental evidence has shown that exposure to a wide variety of environmental chemicals, such as metals (Cd, As), dioxins, polycyclic aromatic hydrocarbons (PAHs), phthalates, and phenols can dysregulate lncRNAs (Miguel et al., 2020). However, studies in toxicology solely focusing on lncRNAs are limited. Given their importance in biological processes combined with a relative lack of knowledge about their function and interactome, there is a need for better exploration of interactions between toxicant exposure and lncRNA function in toxicological studies.

Using zebrafish as a model, our group identified a lncRNA, sox9b long intergenic noncoding RNA (slincR) that is activated by multiple ligands (eg, 2,3,7,8-tetrachlorodibenzo-p-dioxin [TCDD], PAHs) of the aryl hydrocarbon receptor (AHR) homologue AHR2—a conserved ligand-dependent transcription factor that regulates developmental and immunological processes (Garcia et al., 2017, 2018). Ligand-mediated activation or inhibition of AHR can lead to adverse health effects either through biotransformation of the ligand into reactive and toxic intermediates, and through disruption of downstream targets (including slincR) or signaling mechanisms important for biological function (Garcia et al., 2017). slincR lies 38 kb upstream of the promoter region of sox9b—a developmental gene that is inhibited by TCDD resulting in embryonic jaw and cartilage malformations, and reduced tissue regenerative capacity. Our previous work showed that slincR is enriched in the 5′UTR of sox9b (Garcia et al., 2018). Using a morpholino-based approach, we showed that slincR knockdown increased sox9b levels endogenously and in the presence of the AHR2 ligand TCDD (Garcia et al., 2017, 2018). Furthermore, slincR knockdown increased mRNA levels of several sox9b target genes (Garcia et al., 2017), suggesting that slincR regulates the sox9b interactome. While we have uncovered some of the functional roles of slincR, several knowledge gaps about its sox9b dependent and independent functions remain.

Here, we developed a CRISPR-Cas9 generated slincR mutant line (slincR^osu3) to investigate how a slincR mutation can alter its functions and modulate dioxin-dependent or independent toxicity at developmental life stages. We used several readouts that are representative of AHR2 induction—morphology, behavior, gene expression, cartilage development, and tissue regeneration—to better understand the function of slincR in zebrafish development.

Materials and methods

 

Zebrafish husbandry

Tropical 5D (wildtype/WT) zebrafish lines, as well as a newly generated slincR mutant line (slincR^osu3) were reared according to Institutional Animal Care and Use Committee (IACUC) protocols at the Sinnhuber Aquatic Research Laboratory, Oregon State University. The approval numbers associated with this manuscript are IACUC-2021-0227 and IACUC-2020-0136. Adult animals were raised in a recirculating water system (28 ± 1°C) with a 14-h:10-h light-dark schedule. Adult wild-type fish were kept at a density of 6–8 fish per liter. Spawning and embryo collection were conducted as described in previously (Westerfield, 2007). During sample collection and after conclusion of experiments, larval zebrafish were euthanized with Tricane (3-amino benzoic acid ethylester; MS-222) at concentrations exceeding 150–200 mg/l.

Generation of slincR^osu3 line: guide design and synthesis, microinjections, and founder screens

Target guide selection

We designed guide RNAs to target shared sequence in exon 1 in both our dominant transcript NCBI LOC110366352 (Garcia et al., 2017) and alternate Ensemble transcript ENSDART00000154894.2. We followed recommendations for target guide design, assembly, and RNA guide synthesis protocols described previously (Varshney and Burgess, 2016) with some modifications to the guide assembly as described previously in (Shankar et al., 2022). To choose high quality target guide sequences we used CRISPRscan tracks (Moreno-Mateos et al., 2015) on the UCSC genome browser http://www.crisprscan.org/. This program finds potential 23-bp guide sequences (single-guide RNAs [sgRNAs]) located along the gene of interest that contain both a 3′ NGG PAM (protospacer adjacent motif) required for Cas9 double strand DNA cutting, and a 5′ GG sequence start site for efficient T7 polymerase RNA synthesis. We chose a guide targeting the reverse strand that had a predicted high activity score of 58 and zero off targets.

Guide template assembly and sgRNA synthesis

Oligos and universal primers were purchased from Integrated DNA Technologies (Coralville, Iowa). List of oligos is included within Supplementary Table 1. For slincR exon 1, the 56-bp top strand oligo included a 5′ 17-bp T7 polymerase recognition sequence, followed by 19-bp slincR Exon 1 sequence (without PAM sequence) and ended with 20-bp overlapping sequence complimentary to the 3′ end of the bottom strand Constant Oligo. For constant or universal oligo, the 80-bp sequence overlaps the top strand oligo 1 at the 3′ end and is the tracrRNA (trans-activating CRISPR RNA) sequence required for Cas9 protein recognition. Amplification of the template was performed using the KOD Hot Start DNA Polymerase Kit (Millipore Sigma) according to the manufacturer’s recommendation. Final concentrations for the gene-specific and constant oligos were 0.008 μM each, forward and reverse universal primers were 0.2 μM each. Amplification was done with the following PCR conditions: denaturation at 95°C for 2 min, followed by 35 cycles of amplification (95°C for 20 s, 62°C for 10 s, and 72°C for 10 s), and a final extension at 72°C for 5 min. The HiScribe T7 Quick High Yield RNA Synthesis Kit (New England Biolabs) with 2 µl of amplified DNA guide template was used as input, for in vitro transcription according to manufacturer’s instructions and included the DNase treatment step. The sgRNA was precipitated using ethanol/sodium acetate (pH 5.2), checked for quality on a 1.2% agarose Lonza Flashgel (Lonza, Rockland, Maine), and RNA concentration was determined using a BioTek SynergyMix microplate reader with the Gen5 Take3 module. sgRNA was stored at −80°C.

Cas9 mRNA synthesis

The Qiagen QIAprep Miniprep kit was used to purify the zebrafish codon optimized plasmid pT3TS-nCas9n (a gift from Wenbiao Chen; Addgene Plasmid No. 46757). Purified plasmid was digested with XbaI (New England Biolabs, R0145S) restriction enzyme, purified by ammonium acetate precipitation, and used as template for in vitro transcription with the Ambion mMessage mMachine T3 Transcription Kit (ThermoFisher, AM1348). The Cas9 mRNA was purified again by precipitation, run on a FlashGel and quantified with BioTek plate reader, and diluted for single use aliquots at 500 ng or 1000 ng/µl and stored at −80°C.

Embryo microinjections, founder mutation screen, and propagation of the slincR^osu3 line

CRISPR-CALC (https://research.nhgri.nih.gov/CRISPRz/downloads/CRISPR_CALC.xlsx, accessed 6/24/22) was used to calculate the injection mixture components to deliver 300 pg each of the sgRNA and Cas9 mRNA in a 1.4 nl injection. WT embryos were injected into the yolk stream of staged early one cell developed embryos. Approximately 200 embryos were injected, screened for fertilization and normal development. Genomic DNA (gDNA) was extracted from a subset of 16 injected and 1 uninjected control embryos at 5 days postfertilization (dpf). Forty microlitres of 50 mM NaOH were added to wells with embryos, heated at −95°C for 15 min, cooled, and vortexed to disrupt tissue, and 8 µl of 1M TRIS pH 8 added and mixed, then 120 µl water added to dilute mix. PCR-based screening for slincR was performed with gDNA extracts using a KOD Hotstart Polymerase kit following manufactures recommendations. A 25 µl PCR reaction was run that included a 2 µl gDNA template along with primers (P1/P2, Supplementary Table 1) designed to flank guide cutting site and produce a 272-bp product, with 35 PCR cycles with annealing temperature of 60°C. Successful cutting was estimated by observation of shift of band sizes on a gel. Initial screen indicated 15/16 fish had induced indels and the remainder of the injected fish were grown to adulthood. Germline transmission and indel types were screened from 10 embryos from 1 male outcrossed male F0 by PCR and all 10 had multiple bands. Three samples representative of gel band pattern differences were sent for Sanger sequencing at OSU’s Center for Quantitative Life Sciences (CQLS) to find potential mutations. Poly Peak Parser (http://yosttools.genetics.utah.edu/PolyPeakParser/) was used to unravel chromatograms with wildtype and alternate alleles. Our 3 samples had either a −1 bp, −4 bp or + 18 bp mutation. The large 18 bp insertion was the most promising mutation for potential impact on normal linc RNA function. The male was outcrossed again and a clutch of 60 was grown to adulthood. Adults were fin clipped for gDNA extraction and sequenced to find all + 18-bp heterozygous carriers and incrossed to establish homozygous slincR^osu3 mutant line through genotyping.

Prediction of slincR mRNA secondary structure and sox9b binding

The secondary structure of slincR mRNA sequences in WT and slincR^osu3 embryos were predicted by LinearFold (Huang et al., 2019), a computationally efficient method to align and fold sequences RNA sequences. To explore binding potential between sox9b (Ensembl) and slincR mRNA sequences for WT and slincR^osu3 embryos, sequences were imported into LinearCoPartition (Zhang et al., 2022) for prediction of RNA-RNA interactions and stability. Both algorithms predict structures using a nearest-neighbor thermodynamic model with the free energy parameters (Mathews et al., 2004).

Based on predicted structures from LinearCoPartition, the binding energy between sox9b and slincR mRNA sequences was further computed. The binding energy is the free energy change of intermolecular base pairs, which is defined as $\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathrm{binding}}=\mathrm{\Delta }{G}_{\mathit{sox}9b-\mathit{slincR}}-\mathrm{\Delta }{G}_{\mathit{sox}9b}-\mathrm{\Delta }{G}_{\mathit{slincR}}$, where $\mathrm{\Delta }{G}_{\mathit{sox}9b-\mathit{slincR}}$ is the free energy change of both intra- and intermolecular base pairs, and $\mathrm{\Delta }{G}_{\mathit{sox}9b}$/$\mathrm{\Delta }{G}_{\mathit{slincR}}$ is free energy change of intramolecular base pairs only existing in sox9b/slincR sequence.

TCDD exposures

Shield-stage (approximately 6 hpf) embryos, were exposed to TCDD (SUPELCO Solutions Within; Bellfonte, Pennsylvania; CAS number 1746-01-6), with 0.1% DMSO in embryo medium as the vehicle. For each experiment, embryos used were derived from multiple spawning clutches from either our mass spawning facility or group spawning tanks; these embryos from multiple clutches were mixed to provide the pool of embryos used for experimentation. Prior to making of the working solutions, the TCDD stock solutions (1000×) were sonicated for 40 min and vortexed well. Exposures lasted 1 h in 20 ml glass vials (10 embryos/ml) with gentle rocking and under darkness. Vials were also gently inverted every 15 min to ensure proper mixing. After the exposure, embryos were rinsed 3 times with embryo medium and then transferred to 96-well plates for further assays.

High-throughput developmental screening

Mortality and morphology

For developmental screening, WT and slincR^osu3 embryos were exposed to a concentration range of TCDD (0, 0.0625, 0.125, 0.25, 0.5, and 1 ng/ml) for 1 h at 6 hpf as described previously and then transferred to clean embryo media in 96-well plates (1 embryo per well). At 24 and 120 hpf, embryos (32 embryos per treatment) were run through a battery of morphological and behavioral assays. For mortality and morphological observations, 22 endpoints were recorded for each embryo (mortality, yolk sac edema, body axis, eye, snout, jaw, otic vesicle, pericardial edema, brain, somite, pectoral fin, caudal fin, circulation, pigmentation, trunk length, swim bladder, notochord distortion, and alterations in touch response) at 24 and 120 hpf (Truong et al., 2014b). Control mortality up to 20% was considered as an acceptable threshold and any experiment with >20% mortality in WT-DMSO treatments was discarded. Data for all assessments were recorded or imported into a laboratory information management system called the Zebrafish Acquisition and Analysis Program (ZAAP).

Statistics

All statistical estimations were done within ZAAP which uses the R platform (https://www.r-project.org/) as described in our previous publications (Reif et al., 2016; Truong et al., 2014a). Briefly, significant differences between control and exposed fish were computed using a 1-sided Fisher’s exact test, where adverse endpoints were tested to have a greater occurrence in exposed fish. To control for the family-wise error rate, a Bonferroni correction for multiple comparisons was applied (p < .01).

Larval photomotor response

To study behavioral responses, the larval photomotor response (LPR) assay was performed in Zebrabox (Viewpoint Behavior Technology). Briefly, WT and slincR^osu3 were exposed to 0 and 0.0625 ng/ml TCDD (nontoxic concentration based on morphology data) for 1 h at 6 hpf and embryos were transferred to 96-well plates as described previously (32 embryos per treatment). At 120 hpf, plates were placed in a Viewpoint Zebrabox system, and the video tracking protocol on the Viewpoint Zebrabox software was used to track total larval movement over 4 alternating light/dark cycles (epochs), with 3 min visible light (1000 lux) and 3 min dark (infrared light) per cycle. The data integration time was set to 6 s bins, and raw data files were processed using custom R scripts (version 4.0.3) to average the total distance traveled for each integration.

Statistics

Statistical significance was based on area under the curve analysis for the third epoch, using a Kolmogorov-Smirnov test (p < .05). Dead or malformed embryos were excluded from the analyses.

Adult behavior assays

 

WT and slincR^osu3 embryos (without any TCDD exposures) were raised to adulthood for adult behavioral assays. Briefly, 120 hpf larvae, with no apparent deformities, were transferred to 2.8 l tanks for the grow-out. At ∼3 months of age, the fish were run through a battery of behavioral assays, including startle response, predator avoidance, and schooling in an array of 8 tanks (105 mm × 105 mm × 130 mm) with an attached LCD monitor on the side and cameras on the top recording videos (20–22 fish per genotype per sex). For predator and schooling assays, the tanks are divided into 3 virtual zones—close, middle, and far relative to the monitor. For predator assay, video of a predatory fish is shown, and data interpretation is based on the percent time a fish spends in the far zone—an indicator of flight response from the predator. For schooling assay, video of other zebrafish is shown, and data interpretation is based on percent time a fish spends in the near zone—an indicator of social behavior of zebrafish. Video analysis was done using Noldus Ethovision (v 11.5). Detailed assay procedure is described in our previous work (Dasgupta et al., 2022).

Statistics

Data analysis and statistical estimations were done as previously described in (Dasgupta et al., 2022). For schooling and predator response, 3-way ANOVAs for the percent of time spent in each zone were executed with treatment, status (acclimation or video periods), and sex of the fish as variables, followed by Tukey’s post hoc tests. For startle response, a 2-way repeated measures ANOVA was executed for the average distance traveled during the 20 s after each tap, with treatment and sex as the main variables and tap as a nested, repeated measures variable. In all cases, a p-value of <.05 was considered statistically significant.

mRNA sequencing

Exposures and total RNA isolation

WT and slincR^osu3 embryos were exposed to 0 or 1 ng/ml TCDD in 0.1% DMSO at 6 hpf for 1 h, followed by washing as described previously. Transcriptomic assessments were done at 48 hpf, a time point that provides a snapshot of the developmental transcriptome in absence of any phenotypes, but largely drives phenotypic outcomes at 120 hpf. Embryos were raised until 48 hpf in clean embryo media and 4 biological replicates were created by pooling 8 embryos per replicate from individual wells. Embryos were then placed into Eppendorf Safelock Tubes and excess solution removed. 0.5 mM zirconium oxide beads were added along with 200 µl of RNAzol (Molecular Research Center, Inc) and the tubes were immediately placed into a Bullet Blender (Next Advance), using settings recommended by the manufacturer. Following homogenization at a speed of 8 for 3 min, tubes were taken out, 300 µl of RNAzol added, and the tubes were frozen at −80°C. Prior to isolation, the tubes were thawed at room temperature for 25 min, mixed by vortexing, and RNA was purified using the Direct-zol MiniPrep kit (Zymo Research) according to manufacturer’s instructions, including an optional DNase-1 digestion treatment for 15 min. RNA integrity (RIN) was assessed using an Agilent Bioanalyzer (Santa Clara, California). RNA samples with RIN values >8 were processed for library preparation and a 100-bp paired end sequencing was performed at the Beijing Genomic Institute (Beijing, China) using their DNBseq platform (https://www.bgi.com/global/sequencing-services/rna-sequencing-solutions/transcriptome-sequencing/) with a stranded library preparation.

Bioinformatics and gene ontology

Raw sequence reads were filtered based on Illumina quality scores and their quality was checked prior to processing using FastQC (version 0.11.7). Reads were trimmed using Cutadapt version 1.8.1 to remove the adapter (AGATCGGAAGAGCA) and trimmed files were aligned using STAR aligner against the zebrafish GRz11 genome assembly. Read counts per gene were then estimated using HTSeq with default options. Out of a total gene number of 32 520, genes with a sum of zero read counts across all 16 samples were then filtered out; this resulted in a total of 29 627 genes per sample. Following this, genes with read counts in the bottom 20^th percentile across all samples were filtered out; this resulted in a total of 26 347 genes remaining across samples. After this filtering, differential expression of genes (p_adj<.05) was assessed using Bioconductor’s DeSeq2 using RStudio (version 4.1), with pairwise comparison of all treatments to WT-DMSO or WT-TCDD. Data visualization was performed in GraphPad Prism version 9.0 and clustered heatmaps, based on Z-transformation, were generated using NGCHM web builder (https://build.ngchm.net/NGCHM-web-builder/). Gene Ontology assessments (Biological Process) were performed within ShinyGO 0.76 (http://bioinformatics.sdstate.edu/go/), with an FDR cutoff of 0.05. Raw files from sequencing data have been uploaded to NCBI Gene Expression Omnibus (GEO; GSE222975). Raw data from assays, gene expression analyses, primer details and phenotypic assessments are available within Supplemental files and tables, as well as Dryad repository https://doi.org/10.5061/dryad.t4b8gtj4w.

qPCR for selected genes

Exposure and qPCR reactions

To confirm the expression levels of selected genes specific to the AHR2 regulatory pathway and sox9b targets, we performed qPCR analysis for 7 genes—cyp1a, slincR, sox9b, fabp2, fgfr3, adamts3, notch3, srfp2. Forward and reverse primers specific to each target (primer sequences listed in Supplementary Table 1) were synthesized from Integrated DNA Technologies (San Diego, California); these primer sets were validated and successfully used in our previous studies (Garcia et al., 2017, 2018). WT and slincR^osu3 embryos were exposed to 0 and 1 ng/ml TCDD at 6 hpf for 1 h and total RNA was extracted at 48 hpf as described previously (4 biological replicates per treatment). In a separate experiment, we also did a similar exposure, but with 0.125 ng/ml TCDD. Ten microliters of 1-step qRT-PCR reactions were set up consisting of 5 μl SYBR Green master Mix and 0.08 μl reverse transcriptase enzyme mix (Power SYBR Green RNA-to-CT 1-Step Kit; Applied Biosystems, Foster City, California), 0.2 μl each of 10 μM forward and reverse primers, and 20 ng RNA per reaction. The QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific, Waltham, Massachusetts) was used under the following cycling conditions; reverse transcription at 48°C for 30 min, denaturation and activation of SYBR polymerase at 95°C for 10 min, followed by 40 cycles of amplification (95°C for 15 s, 60°C for 1 min).

Statistics

Expression values were normalized to β-actin and analyzed using the 2^−ΔΔCt method. Unpaired Student’s tests or 2-way ANOVAs, followed by Tukey’s post hoc tests were used to estimate the statistical significance of treatments.

Alcian Blue cartilage staining

Exposure and staining

To visualize head cartilage structure, we performed Alcian Blue staining following of WT and slincR^osu3 embryos to 0 or 1 ng/ml TCDD for 1 h. Postexposure, embryos were raised in clean water and at 72 hpf, larvae were euthanized with Tricaine and fixed in 4% paraformaldehyde overnight at 4°C. Pigmentation was removed by incubating fixed larval samples for 1 h in a mixture of 3% H₂O₂/1% KOH. The cartilage was stained with 0.4% Alcian Blue 8GX (Sigma-Aldrich, St Louis, Missouri) in 70% ethanol and 80 mM MgCl₂ as described in (Garcia et al., 2018). Larvae were imaged using a Nikon SMZ 1500 Stereomicrocope (Nikon Inc, Melville, New York). Morphometric analysis (N = 10 embryos per treatment) of ventral larval pharyngeal cartilages was performed using ImageJ v1.51j8.

Statistics

Statistical significance was determined using a 2-way ANOVA (treatment and mutation as 2 factors) and Tukey’s post hoc analysis within GraphPad Prism 9 (p < .05).

Caudal fin regeneration experiments

Caudal fin amputation

To study regenerative capacity, WT and slincR^osu3 embryos were exposed to 0.1% DMSO and 1 ng/ml TCDD at 6 hpf for 1 h (25 per treatment) and incubated at 28°C in petri dishes. At 48 hpf, larvae were anaesthetized in tricaine, placed on an agar plate, and the caudal fin primordia was amputated with a surgical crystal microblade just posterior to the notochord. Larvae were then transferred to a petri dish containing chemical-free fish water to reverse the anesthesia and transferred to 96-well plates (1 embryo per well). The embryos were then imaged with a Keyence BZ-X700 microscope (Keyence Corporation of America, Itasca, Illinois) and then incubated at 28°C for 3 days. At the end of 3 days, the embryos were again imaged to measure the extent of regeneration (N = 20–24 live embryos per genotype).

Statistics

Measurements for regenerative length were performed within ImageJ and results were graphed using GraphPad Prism 9. Statistical significance was estimated using a 2-way ANOVA, followed by Tukey’s post hoc tests (p < .05).

Cell proliferation assay with bromodeoxyuridine/5-bromo-2′-deoxyuridine

BrdU staining and immunohistochemistry

To study if regenerative impacts were driven by cell proliferative capacity, a cell proliferation assay was conducted on regenerating fin tissue after pulse labeling with bromodeoxyuridine/5-bromo-2′-deoxyuridine (BrdU) (Roche Applied Science) as previously described (Zodrow and Tanguay, 2003). Following amputation at 48 hpf as described previously, WT or slincR^osu3 larvae were incubated in clean media until 24 h postamputation (hpa). Following this, larvae were incubated in a 10 mM BrdU solution in embryo media for 6 h at 28°C. The larvae were then anaesthetized in tricaine and fixed in 4% paraformaldehyde overnight. The fixed larvae were dehydrated with methanol and then stored in methanol at −20°C. Immunohistochemistry was conducted on the stored larvae by rehydrating with a graded methanol/PBST (phosphate-buffered saline [PBS] and 0.1% Tween 20) series. The larvae were then treated with proteinase K in PBST for 20 min at room temperature and then rinsed 3 times (approximately 3 min each) with PBST. The larvae were refixed in 4% paraformaldehyde for 30 min and then washed 3 times in water, followed by 2 rinses in 2N HCl and incubation in 2N HCl at room temperature for 1 h. After 3 washes with PBST, the larvae were then blocked with 1% normal goat serum in PBST for 1 h at room temperature and then incubated with anti-BrdU antibody (1:100; G3G4; Developmental Studies Hybridoma Bank, Iowa City, Iowa) overnight at 4°C. After four 30-min washes with PBST, the larvae were incubated with a secondary antibody (1:1000; Alexa 546-conjugated goat antimouse; Molecular Probes, Eugene, Oregon) for 4 h at room temperature. The larvae were then washed 4 times for 30 min in PBST and imaged using the Keyence BZ-X700 (N = 12 embryos per treatment).

Statistics

Results were graphed using GraphPad Prism 9 and statistical estimations were performed using a Student’s t test (p < .05).

Results

slincR ^{osu 3} has an 18 bp insertion, with an altered predicted secondary structure and predicted sox9b binding

We previously determined that the dominant zebrafish slincR transcript is 466 bases long, polyadenylated, and contains 2 exons. Alternate transcripts likely exist and have not been studied. Upon gene editing, the homozygous slincR mutant line (slincR^osu3) contains an 18 bp insertion within Exon 1 of slincR sequence (Figure 1A); the insertion was validated over multiple generations using Sanger sequencing of gDNA (Supplementary Figure 1). Structural analysis predicted that the secondary mRNA structure is altered in slincR mutants (Figure 1B). We also conducted the RNA-RNA interaction predictions between sox9b and slincR. The Ensembl sox9b mRNA has a 181 base-long 5′UTR; our previous study showed that slincR is enriched within 5′UTR of sox9b, and subsequently, sox9b is repressed upon slincR activation (Garcia et al., 2018). Our current analysis showed strong binding potential of slincR to sox9b sequence, specifically within its 5′UTR. Interestingly, binding energy of the mutated slincR to sox9b was lower than unmutated slincR, suggesting a higher binding potential (Figure 1C). Consistent with this, nucleotide-nucleotide binding potential between sox9b and slincR in slincR^osu3 embryos were higher compared with WT, with more binding within the 5′UTR and upstream promoter areas. Overall, this suggests that within slincR^osu3 fish, slincR potentially binds more effectively to sox9b compared with WT.

Figure 1.

slincR mutation and secondary structure predictions. A, slincR mRNA sequence (466 nucleotides), with exon 1 highlighted in yellow. 18 bp insertion in slincR^osu3 is indicated within exon1. P1F and P1R are diagnostic primers for genomic DNA, with P1R being located within the intron. P2F and P2R serve as both diagnostic and qPCR primers for cDNA. Primer details are included in Supplementary Table 1. B, Secondary mRNA structure (using LinearFold) of WT and slincR^osu3 embryos, with 18 bp insertion site indicated. Green: Stems (canonical helices); Red: Multiloops (junctions); Yellow: Interior Loops; Blue: Hairpin loops; Orange: 5′ and 3′ unpaired region. C, RNA-RNA interactions between slincR and sox9b using LinearCoPartition. Cyan backbone: slincR; Gray backbone: 1000 bases upstream of sox9b 5′UTR; Black backbone: sox9b 5′UTR; Orange backbone: sox9b ORF. Gray arch: intramolecular base pairs; Blue arch: intermolecular base pairs. Insertion sites and insertions are indicated on the backbone.

slincR mutation alters bioactivity in presence of TCDD

To study the bioactivity of slincR^osu3 embryos in the presence and absence of TCDD, we performed high throughput developmental screening on embryos exposed to a concentration range (0–1 ng/ml) of TCDD at 6 hpf for 1 h. Raw data are presented in Supplementary Table 2. Figure 2A presents results from our morphological screening, consisting of 18 endpoints. slincR^osu3 did not display any morphological deficits in absence of TCDD. For TCDD exposures, the lowest effect concentration (LEC) for WT and slincR^osu3 were 0.25 and 0.125 ng/ml, respectively, demonstrating that slincR^osu3 embryos were more sensitive to TCDD.

Figure 2.

High throughput screening phenotypes. A, Mortality and 20 morphological effects of WT and slincR^osu3 embryos exposed to a range of TCDD concentrations for 1 h at 6 hpf. Observations were performed at 24 hpf (marked as ‘24’) or 120 hpf. Numbers within cells represent percent abnormal values. B–D, Larval photomotor responses (LPR) in different treatments at 120 hpf. L1–3 and D1–3 represent light and dark epochs, respectively. (B) and (C) compares TCDD (0.0625 ng/ml) to DMSO treatments for each genotype; (D) compares DMSO (unexposed) treatments between the 2 genotypes. N = 32 for mortality/morphology; N = 23–31 for LPR. Statistical significance was based on area under the curve analysis for the third epoch, using a Kolmogorov-Smirnov test (p < .05). Detailed statistical outputs are presented in Supplementary Table 4.

slincR mutation alters behavioral phenotypes at developmental and adult stages

To study the effects of slincR mutation on behavior during development, we performed an LPR assay at a nontoxic TCDD concentration (0.0625 ng/ml). Raw data and detailed statistical output are presented in Supplementary Tables 3 and 4. Our results showed that both WT and slincR^osu3 larvae displayed subtle but significant hypoactivity in response to TCDD (Figs. 2B–D). Furthermore, slincR^osu3 larvae displayed hypoactivity in the absence of TCDD (vehicle only), suggesting an endogenous mutation-specific effect in the absence of slincR induction.

To determine if behavioral effects due to the slincR mutation persisted into adulthood and later-life health effects, we raised unexposed WT and slincR^osu3 embryos into adulthood and performed a battery of adult behavioral assays. Raw data and detailed statistical output are presented in Supplementary Tables 5–8. Male slincR^osu3 fish spent less time in the far zone and more time in close zone than the negative control fish in the predator and schooling assays, respectively (p < .01) (Figs. 3A and 3B). Mutation-specific effects were also seen in startle tap response assay (p = .01), but no overall impacts of sex were seen (Figure 3C). Overall, these adult behavior data suggest that mutation in slincR can elicit behavioral deficits and can be sex-specific.

Figure 3.

Adult behavior assays. WT and slincR^osu3 embryos were reared to approximately 3 months age and a battery of adult behavior assays were performed (20–22 fish per genotype per sex). A, Percent time spent in far zone in predator assay. B, Percent time spent in close zone in schooling assay. For both these assays, *Represents statistical significance following 2-way ANOVAs (p < .001). C, Habituation following startle response. Statistical estimations were done using 2-way ANOVAs with taps as nested factors. Each circle represents an individual fish. Full details of statistical output are presented in Supplementary Tables 5–8.

slincR mutation affects global gene- and pathway-level responses

To understand the dynamic patterns of gene expression changes following slincR mutation in the presence or absence of 1 ng/ml TCDD, we performed mRNA-sequencing on 48 hpf embryos, followed by Gene Ontology (GO) assessments. Raw data from DeSeq2 results and Gene Ontology results are presented in Supplementary Tables 9–13. For global gene expression assessments, all treatments were compared with wildtype controls (WT-DMSO). Our data showed disruption of 143, 499, and 908 mRNAs in WT-TCDD, slincR^osu3-DMSO, and slincR^osu3-TCDD treatments, respectively (Figure 4A). There were 14 common dysregulated genes between WT-TCDD and slincR^osu3-DMSO treatments (regions i+ii), and 472 unique genes that were only disrupted in slincR^osu3-TCDD treatment (region iii). We then generated clustered heatmaps of samples; as expected, when clustered by WT-TCDD differentially expressed genes (DEGs), the TCDD treatments clustered separately from DMSO and when clustered by slincR^osu3-DMSO DEGs, the mutants clustered separately from WT (Supplementary Figs. 2 and 3). A broader look at overall mRNA dysregulation patterns (Figure 4B) showed that while WT-TCDD largely showed gene inductions (117/143; approximately 81%), slincR^osu3-DMSO and slincR^osu3-TCDD showed a mix of inductions and reductions in mRNA levels (215/499; 43% induction for slincR^osu3-DMSO and 532/908 or 58% induction for slincR^osu3-TCDD), suggesting a slate of deleterious effects due to slincR mutation. GO analyses showed that WT-TCDD treatments disrupted pathways associated with xenobiotic response, developmental patterning (cell development, migration, differentiation, morphogenesis) as well as cartilage and vasculature development (Figure 4C). For slincR^osu3-DMSO, the most significantly disrupted pathways encompassed regulation of metabolic and immune responses (Figure 4D). For slincR^osu3-TCDD, disrupted pathways included subsets of pathways disrupted both in WT and TCDD exposed animals (xenobiotic response, developmental patterning) and slincR^osu3-DMSO (metabolic and immune pathways), indicating that responses were likely additive or synergistic with TCDD exposure and slincR mutation (Figure 4E). Surprisingly, the 472 genes unique to slincR^osu3-TCDD treatment (region iii of Figure 4A) did not collectively map to any significantly disrupted pathways, although individual genes were associated with stress response, ion homeostasis, and tissue development pathways (Supplementary Table 14).

Figure 4.

mRNA sequencing data. WT or slincR^osu3 embryos were exposed to 1 ng/ml TCDD for 1 h at 6 hpf. Sequencing was done at 48 hpf, with 4 pooled replicates per treatment. A, Venn diagram representing differentially expressed genes for each treatment compared with WT-DMSO. B, Cumulative heatmap of all increased and reduced mRNA levels across the treatments at p_adj<.05. Responses noted were trimodal: blue—increased, red—decreased, and white—no differential expression. List of differentially expressed genes is included within Supplementary Tables 9 and 10. C–E, Gene ontology assessments (Top 10 pathways for Biological Processes) as assessed using ShinyGO. Full list included within Supplementary Tables 11–13.

AHR2- and sox9b-associated genes are differentially expressed following slincR mutation

Comparison of all treatments to WT-DMSO

Based on the transcriptomic data, we sought to study the mutation-dependent differential expression of selected sox9b and AHR2 target genes (including slincR) (Figure 5A; sample clustering in Supplementary Figure 4). mRNA levels of slincR were reduced in the slincR^osu3-DMSO treatment and increased in both WT-TCDD and slincR^osu3-TCDD treatments. Among other genes, levels of cyp1a, cyp1b1, cyp1c2, ahrrb, ahrra, wfikkn1, and ahr2 were increased in both WT-TCDD and slincR^osu3-TCDD but were not affected in slincR^osu3-DMSO treatment. Interestingly, mRNA levels of cyp1c1 were increased in all treatments, including unexposed slincR^osu3-DMSO. In addition to cyp1c1 and slincR, 12 other genes were disrupted both in WT-TCDD and slincR^osu3-DMSO treatments and 11 of these genes showed same directionality for both treatments. Neurodevelopmental genes her4.2 and her4.3 were the only genes in this subset with decreased mRNA levels in all treatments.

Figure 5.

Expression profiles of selected genes based on mRNA sequencing or qPCR. WT or slincR^osu3 embryos were exposed to 1 ng/ml TCDD for 1 h at 6 hpf. Sequencing and qPCRs were done at 48 hpf, with 4 pooled replicates per treatment. A, Heatmap representing log₂ fold changes for common AHR target genes, as well as genes that were commonly dysregulated in WT-TCDD and slincR^osu3-DMSO treatments, based on mRNA sequencing data, p_adj<.05. White cells indicate no differential expression. B, qPCR-based comparisons of endogenous expressions (WT-DMSO vs slincR^osu3-DMSO) for cyp1a, slincR, sox9b, and sox9b target genes adamts3, fgfr3, fabp2, notch3, and srfp2. Each dot represents a biological replicate. *Represents statistical significance using a 2-tailed t test (p < .05).

Our sequencing data did not show any differential expression of sox9b in WT-TCDD or slincR^osu3-DMSO treatments. This may be due to potential variability in libraries, because sox9b fold changes, even in response to TCDD, are small. Therefore, we used qPCR to further validate expression of sox9b, along with slincR, cyp1a, and a battery of sox9b downstream target genes studied in Garcia et al. (2017). When compared with WT-DMSO, slincR^osu3-DMSO treatment showed reduced mRNA levels of slincR, sox9b, adamts3, fgfr3, and fabp2 (Figure 5B). For 1 ng/ml TCDD exposures, slincR and cyp1a expression data for qPCR were consistent with sequencing data, with induced slincR and cyp1a expression both in WT-TCDD and slincR^osu3-TCDD treatments (Figs. 6A and 6B). sox9b levels were reduced in all treatments, with an 85% reduction in slincR^osu3-DMSO, confirming a deleterious effect of the mutation on sox9b mRNA (Figure 6C). Consistent with this, mRNA levels of all downstream targets of sox9b were reduced in WT-TCDD and slincR^osu3-TCDD treatments (Figs. 6D–H).

Figure 6.

qPCR for selected genes at 1 ng/ml TCDD. WT or slincR^osu3 embryos were exposed to 0.1% DMSO or 1 ng/ml TCDD for 1 h at 6 hpf. qPCRs were done at 48 hpf. Each dot represents a biological replicate, N = 4 pooled replicates per treatment. Different letters on each bar indicates statistically significant different using a 2-way ANOVA followed by Tukey’s post hoc test (p < .05).

Comparison to WT-TCDD (1 ng/ml)

Leveraging the qPCR data, we also compared slincR^osu3-TCDD exposures to WT-TCDD exposures to assess if there was an interaction between 1 ng/ml TCDD and the slincR mutation. For slincR, induction was inhibited in slincR^osu3-TCDD treatment (Figure 6B). For sox9b and its targets, we saw mixed responses. While sox9b, fabp2, and adamts3 levels in slincR^osu3-TCDD treatments were reduced to basal slincR^osu3-DMSO levels or lower (Figs. 6C–E), fgfr3, notch3, and srfp2 levels remained unchanged (Figs. 6F–H).

Comparison to WT-TCDD (0.125 ng/ml)

Since TCDD elicited differential morphological responses in WT versus slincR^osu3 embryos at 0.125 ng/ml, we studied the relative slincR, sox9b, or cyp1a expressions in WT and slincR^osu3 embryos exposed to 0 or 0.125 ng/ml TCDD. Our data showed similar expression patterns as for 1 ng/ml TCDD, with slincR and sox9b repression and no impacts on cyp1a (Supplementary Figure 5).

slincR mutation alters jaw cartilage development

Since TCDD disrupts normal development of jaw cartilage through a sox9b-dependent mechanism, we developmentally exposed WT and slincR^osu3 embryos to TCDD for 1 h at 6 hpf and evaluated cartilage development patterns at 72 hpf. Our assessments focused on head/jaw cartilages (Figure 7A) where we measured lower jaw length (LJL); intercranial distance (ICD); ceratohyal cartilage length (CCL); and estimated ratios of LJL and CCL with ICD (based on Hu et al., 2020), in order to examine changes in relative positions (Figs. 7B–F). While we saw no significant differences in LJL/ICD ratios (Figure 7G), CCL/ICD ratios for all treatments and mutants were reduced compared with WT-DMSO (Figure 7H). However, there were no differences between WT-TCDD, slincR^osu3-DMSO, and slincR^osu3-TCDD treatments, suggesting that slincR mutation elicits TCDD-like changes.

Figure 7.

Alcian Blue staining for jaw cartilage. WT or slincR^osu3 embryos were exposed to 0.1% DMSO or 1 ng/ml TCDD for 1 h at 6 hpf. A, Staining in entire fish at 72 hpf. B, Head area with representative distance parameters: lower jaw length (LJL); intercranial distance (ICD); ceratohyal cartilage length (CCL). C–F, Representative images from each treatment, with locations of various structures: ep—ethmoid plate, m—Meckel’s cartilage, ch—ceratohyal cartilage, cb—ceratobranchial. G, H, LJL/ICD and CCL/ICD ratios. Statistical significance was assessed using a 2-way ANOVA followed by Tukey’s post hoc test (p < .05). ****p < .001, ns=not significant. N = 10 per treatment.

slincR mutation inhibits caudal fin regeneration

Since previous studies have shown that sox9b repression plays a significant role in TCDD-induced inhibition of tissue regeneration (Andreasen et al., 2006), we amputated caudal fins of 48 hpf-old WT and slincR^osu3 embryos in the presence or absence of TCDD and evaluated their regenerative capacity by 120 hpf (3 dpa or days postamputation). Our data showed significant inhibition (approximately 50% mean) of regenerative capacity across all treatments compared with WT-DMSO (Figs. 8A and 8C) and none of the treatments were different compared with each other. To understand if such inhibition was caused by lack of cell proliferation in the regenerative plane, we immunostained unexposed WT and slincR^osu3 embryos with an anti-BrdU antibody and showed that cell proliferation was significantly inhibited (approximately 50% mean) within the regenerative area in slincR^osu3 embryos (Figs. 8B and 8D).

Figure 8.

Regenerative capacity following caudal fin amputations. WT or slincR^osu3 embryos were exposed to 0.1% DMSO or 1 ng/ml TCDD for 1 h at 6 hpf. Their caudal fins were surgically amputated at 48 hpf (0 days postamputation or dpa), followed by measurement of regenerative capacity at 3 dpa or cell proliferation at 1 dpa. A, Regenerative capacity at 3 dpa. N = 20–24 live embryos per treatment. Statistical comparisons were performed using a 2-way ANOVA followed by a Tukey’s post hoc test (p < .05). B, Cell proliferation within WT-DMSO and slincR^osu3-DMSO treatments using BrdU immunostain at 1 dpa. N = 12 embryos per treatment. Statistical comparisons were performed using an unpaired Student’s t test (p < .05). C, D, Representative images from amputated and BrDU-stained WT-DMSO and slincR^osu3 fish. Dotted lines indicate plane of amputation. Arrow indicates proliferating cells.

Discussion

 

slincR was first identified in our lab as an AHR2 ligand-inducible lncRNA from a transcriptomic screen of PAHs (Goodale et al., 2013). Upon exposure to xenobiotic AHR2 ligands, slincR mRNA levels are overexpressed consistently in concordance cyp1a induction (Garcia et al., 2018). We have previously shown that, when induced by AHR2 ligands such as TCDD, slincR binds to the 5′UTR of sox9b and represses its expression (Garcia et al., 2018). Therefore, on AHR induction, slincR potentially plays a crucial role in modulating the sox9b interactome and sox9b-dependent developmental processes such as cartilage development. However, not much is known about the function of this lncRNA. Furthermore, the antagonistic role between slincR and sox9b is not representative to all AHR2 ligands; in fact, many PAHs that induce slincR do not display a sox9b repression (Garcia et al., 2018). We have also identified possible orthologs of slincR within rodents and humans (Garcia et al., 2018), indicating that slincR may have an important role in the human response to xenobiotic ligands. Therefore, we generated a slincR mutant line to better understand how slincR regulates development, both endogenously and following AHR2 induction.

Effects of TCDD exposure on WT embryos

Consistent with previous studies, our developmental TCDD exposures resulted in morphological phenotypes (edema, bent body axis, abnormal developmental behavior, and cartilage deformities) at 5 dpf (Figure 2A) that are associated with AHR2 induction (Shankar et al., 2020). TCDD toxicity was preceded by induction of AHR2 pathway genes cyp1a, cyp1b1, ahrra, ahrrb and, importantly, slincR (Figure 5A). Although not evident in mRNA-seq, qPCR confirmed that TCDD repressed sox9b mRNA levels (Figure 6C). These data enabled us to use TCDD as a model AHR2 inducer and leverage the phenotypic and genetic responses as a benchmark to understand how slincR regulates development and developmental toxicity.

Effects of slincR mutation in absence of TCDD: endogenous consequences of mutation

Our slincR^osu3 mutant line contains an 18 bp insertion within the exon 1 of slincR, which is predicted to significantly change the slincR secondary structure (Figure 1). mRNA seq and qPCR confirmed that basal slincR mRNA levels were reduced in the slincR^osu3 embryos. qPCR showed an 85% decrease in sox9b mRNA levels (as well as reduced mRNA levels of a battery of sox9b downstream targets) which was similar to, or lower than slincR mRNA levels in WT-TCDD treated embryos (Figs. 4B and 5). These were contrary to our findings with slincR morphants where the expression of these genes was increased with slincR knockdown (Garcia et al., 2017). Therefore, despite slincR inhibition in both, our slincR^osu3 line is not functionally identical to a slincR morphant used in our previous studies (Garcia et al., 2017, 2018) and may be a gain-of-function mutation with respect to sox9b patterns. Intermolecular binding predictions of slincR and sox9b secondary mRNA structures revealed that potential binding of slincR to the sox9b 5′UTR may be stronger in slincR^osu3 (Figure 1C). These observations suggest: (1) the effect of potentially higher slincR-sox9b binding potentially overrides the effect of decreased slincR mRNA levels in slincR^osu3 and leads to a stronger repression of sox9b and (2) the repression of sox9b due to the slincR mutation is regulated by pathways independent of direct slincR-sox9b binding. Therefore, the interaction between slincR and sox9b is potentially not exclusively antagonistic as previously assumed.

When compared with WT embryos, slincR^osu3 mutants were morphologically normal during development, but displayed behavioral abnormalities during both developmental and adult stages (Figs. 2 and 3). Specifically, during development, behavioral (LPR) phenotypes mimicked WT-TCDD responses (hypoactive). Behavioral phenotypes during development may be driven by a combination of overall fish health and neurodevelopmental effects. mRNA sequencing showed extensive gene-level disruption of slincR^osu3 mutants that aligned to metabolic and immunological pathways (Figure 4D); these can have effects on developmental health that impact behavior. Additionally, both WT-TCDD and slincR^osu3 treatments showed reduced mRNA levels of her4.2 and her4.3—the only transcripts that were repressed (instead of induced) in both treatments (Figure 5A). The her4 genes are downstream targets of notch signaling pathways that are enriched in the developing nervous system and regulate crucial aspects of neurogenesis, such as maintenance of neural progenitor fate as well as neural stem cell fate (Schmidt et al., 2013). However, their role in driving behavioral phenotypes is not clear and further functional studies need to be performed to better understand their role in the AHR pathway and the slincR interactome. We also saw sex-specific effects in the various readouts of adult behavioral assays. These effects may be driven by metabolic or immunological dysregulation as suggested by the gene expression data; however, the mode of action needs extensive evaluation. Overall, the data suggests that the slincR mutation significantly affected aspects of sensory system development.

In addition to behavior, slincR^osu3 mutants also showed differential responses during cartilage development and tissue regeneration. Previous studies have shown that sox9b knockdown causes TCDD-like adverse effects in head/jaw cartilage development that is partially rescued by sox9b mRNA injection (Xiong et al., 2008). Consistent with this, our current work showed that slincR mutation causes TCDD-like effects in CCL/ICD ratios (Figure 7H). This suggests at least a partial regulatory role of slincR in cartilage development. Zebrafish can regenerate their caudal fin tissues following amputation in a process that involves formation of wound epithelium, blastema formation, and the distal to proximal propagation of cell proliferation (Mathew et al., 2009). Our previous work has shown that sox9b mutants disrupt the caudal fin regeneration process (Mathew et al., 2008). We also demonstrated that TCDD exposures inhibit this regenerative process and that it is partially modulated by sox9b repression through an interaction between the AHR and Wnt pathways (Mathew et al., 2009; Shankar et al., 2020), and is accompanied by lack of cell proliferation within the regenerative plane (Mathew et al., 2009). In the current study, unexposed slincR^osu3 embryos displayed a similar TCDD-like phenotype, with lack of tissue regeneration accompanied by a lack of cell proliferation (Figure 8). However, unlike previous studies, we did not see any disruption of Wnt signaling in presence or absence of TCDD; perhaps due to our transcriptomic profiling in whole embryos instead of amputated fins where Wnt signaling may be enriched (Mathew et al., 2009). Collectively, these data showed that the basal slincR-sox9b co-repression in slincR^osu3 embryos plays a crucial role in driving jaw, cartilage, and regeneration phenotypes that were consistent with previous work. Since endogenous slincR levels are also regulated by AHR2 (Garcia et al., 2017), the current study further confirmed the importance of an endogenous AHR-slincR-sox9b axis and its role in developmental events.

Effects of slincR mutation in presence of TCDD: mutation effects following AHR induction

Based on previous evidence, our initial hypothesis was that a slincR mutation would limit its function, would increase sox9b levels and mitigate or rescue TCDD-induced phenotypes. Our data partially invalidated this hypothesis and showed that, compared with WT, slincR^osu3 embryos were equally or more sensitive to TCDD. As expected, the slincR mutation did not affect the TCDD-mediated induction of AHR2 target genes cyp1a, ahrra, ahrrb, wfikkn1, foxq1a, because slincR likely lies downstream of these genes. sox9b and its targets remained repressed in the slincR^osu3-TCDD treated animals, mimicking the response patterns in WT-TCDD or slincR^osu3-DMSO treated animals (Figure 6). Interestingly, slincR^osu3 embryos were more sensitive to the morphological effects of TCDD than WT, with a higher percentage of abnormal slincR^osu3 embryos at 0.125 ng/ml TCDD (Figure 2A). This was not sox9b-dependent because global sox9b repression was identical to slincR^osu3-DMSO and may have been driven by other pathways that were impacted by the slincR mutation. However, tissue-level interactions between sox9b and slincR expression still need to be determined. Both cartilage morphometrics and regenerative capacity in slincR^osu3-TCDD treated animals remained comparable to WT-TCDD animals (Figs. 7 and 8). Overall, based on these results, the role of slincR in driving these phenotypes following TCDD exposure and AHR2 induction needs further evaluation.

Conclusion

The absence of protein-coding capacity for lncRNAs makes it challenging to study them, because their regulatory role can be multifaceted, but the established molecular readouts to directly determine their function are limited. We used CRISPR-Cas9 mutations in slincR to study how a mutation in this gene and corresponding changes in sox9b expression affects developmental processes, endogenously and upon xenobiotic AHR activation (Figure 9). sox9b is a transcription factor that has a plethora of developmental roles and any structural reorganization of the slincR RNA could result in downstream consequences for developmental health. We suggest that the mutant slincR binds more strongly to the sox9b 5′UTR repressing it even further than does the WT slincR; however, confirmation of this is still needed. The slincR mutation resulted in endogenous deleterious effects on the transcriptome, specifically upon metabolic and immunological pathways, and lead to abnormal cartilage development and fin regeneration—both of which are likely driven by sox9b repression. Finally, we showed that slincR induction by TCDD resulted in mixed effects on TCDD-dependent phenotypes, e, g., a greater proclivity for slincR mutants to develop morphological abnormalities, but limited impacts on other phenotypes. TCDD targets several different pathways through AHR2 signaling; many of them may be independent of slincR which may explain the limited slincR-TCDD interaction on phenotypes. While further studies are required to characterize the molecular targets of this slincR mutation, this work upended our previous simplistic assumption that a slincR mutation would increase sox9b expression and rescue TCDD phenotypes at a whole embryonic level and opened several more lines of investigation that will have to be pursued to fully understand the developmental role of this important lncRNA.

Figure 9.

Summary of findings. Blue and red cells indicate over or underexpression for genes. Gray cells indicate presence of abnormal effects. White cells indicate no expression changes or no effects. All comparisons to WT-DMSO.

 

Dryad Digital Repository DOI: https://doi.org/10.5061/dryad.t4b8gtj4w

Supplementary data

Supplementary data are available at Toxicological Sciences online.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.