Zebrafish mbnl mutants model physical and molecular phenotypes of 1 myotonic dystrophy 2

23 The muscleblind RNA binding proteins (MBNL1, MBNL2, and MBNL3) are highly


5
The muscleblind (MBNL) family of RNA binding proteins (MBNL1, MBNL2, and 4 6 MBNL3) is highly conserved in structure and function across multicellular species (Oddo The phenotype was most dramatic in skeletal muscle and heart, where mbnl1 and 2 2 7 mbnl2 exon 5 were almost entirely skipped in WT and predominantly included in 2 2 8 1 B/B ;2 A/A and 1 B/B ;2 A/A ;3 C/C mutants (Figs 3B and C and EV4B and C). The brain and 2 2 9 cornea exhibited mbnl splicing patterns similar to one another, but with lower 2 3 0 magnitudes changes than those observed in muscle (Figs 3D and E and EV4D and E). The corneal splicing phenotype suggests that zebrafish mbnl mutants may also model 2 3 2 the genetic disorder Fuchs endothelial corneal dystrophy, a subtype of which was 2 3 3 recently shown to be caused by an expanded CUG repeat that is associated with MBNL changes were observed in whole larvae and intestine (Figs 3A and F and EV4A and F). In most tissues, double and triple mutants exhibited larger mbnl1 and mbnl2 2 3 8 splicing changes than single mutants (Figs 3 and EV4). This is consistent with the idea  Strikingly, the genotypes with the strongest splicing phenotypes, 1 B/B ;2 A/A and 2 4 2 1 B/B ;2 A/A ;3 C/C , also had dramatic size and movement phenotypes (Figs 1-3 and EV4).

4 3
In mouse model and DM patient tissues, some regulated exons were much more 2 4 4 sensitive than others to changes in overall levels of free MBNL proteins (Wagner et al, 2 4 5 2016). Zebrafish also followed this pattern. For example, in skeletal muscle and heart, 2 4 6 mbnl1 mutation alone was sufficient to increase mbnl1 exon 5 inclusion, whereas 2 4 7 mutation of both mbnl1 and mbnl2 was required to increase mbnl2 exon 5 inclusion 2 4 8 (Figs 3B and C and EV4B and C). Overall, these results suggest that, as in other certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

5 1
Misregulation of alternative splicing was widespread in zebrafish mbnl mutants 2 5 2 To understand the genome-wide impact of mbnl mutation on alternative splicing, 2 5 3 we performed RNA-Seq analysis using RNA isolated from the skeletal muscle of adult 2 5 4 WT, 1 B/B , 2 A/A , 3 C/C , 1 B/B ;2 A/A , 1 B/B; 3 C/C , 2 A/A ;3 C/C , and 1 B/B ;2 A/A ;3 C/C zebrafish. There were 2 5 5 no significant changes in the levels of total mbnl1, mbnl2, or mbnl3 mRNAs in mutants splicing events were detected in all mbnl mutants, with misregulated cassette exons 2 5 8 outnumbering alternative 3' and 5' splice sites, retained introns, and mutually exclusive  Table EV3). The widespread splicing changes in all mutants 2 6 0 suggested that all three zebrafish mbnl proteins contributed to splicing regulation, and 2 6 1 that each mbnl mutation caused at minimum a partial loss of function. 1 B/B ;2 A/A and 2 6 2 1 B/B ;2 A/A ;3 C/C fish had the largest number of significantly misregulated splicing events 2 6 3 ( Fig 4A). Surprisingly, 2 A/A ;3 C/C mutants exhibited fewer overall dysregulated alternative 2 6 4 splicing events than 2 A/A mutants or 3 C/C mutants ( Fig 4A). In mouse models, many 2 6 5 cassette exon events were shown to be dysregulated in opposite directions in Mbnl3 2 6 6 mutants and other Mbnl mutants (Thomas et al., 2017). Perhaps opposing splicing 2 6 7 regulation could be a contributing factor to the apparent mild splicing and tank position 2 6 8 phenotypes of 2 A/A ;3 C/C mutants. We also analyzed the magnitude of the splicing change (absolute value of the 2 7 0 change in percent spliced in (ΔPSI) between WT and mutant) for the 62 cassette exon 2 7 1 events that were significantly misregulated in the majority of mbnl mutants (at least 4 of 2 7 2 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 1 3 7). The highest magnitude of splicing change across events was also observed in 2 7 3 1 B/B ;2 A/A and 1 B/B ;2 A/A ;3 C/C fish ( Fig 4B). These results suggest that zebrafish Mbnl  We next asked which genotypes were most similar in their overall splicing 2 7 7 phenotypes. All cassette exon splicing changes that were observed in any mutant 2 7 8 compared to WT were compiled into a single list, and the top 800 events with the largest 2 7 9 variability in PSI were identified. Using the gap statistic, we determined that all samples 2 8 0 could be clustered into two groups based on the splicing phenotype for those events. A 2 8 1 K-means cluster analysis with two centers was then performed, in which all WT and 2 8 2 mutant samples were analyzed for their similarity in PSI for the 800 events. The  We analyzed the frequency of the Mbnl protein binding motif, YGCY, within and 2 8 8 surrounding the cassette exons that were significantly misregulated in 1 B/B ;2 A/A and 2 8 9 1 B/B ;2 A/A ;3 C/C mutants as well as those misregulated cassette exons that were in 2 9 0 common between the two mutants. YGCY motifs were enriched in the introns 2 9 1 downstream of exons whose inclusion was activated by the presence of Mbnl proteins, 2 9 2 while YGCY motifs were enriched upstream and within the cassette exons that were repressed by Mbnl proteins (Fig 4D). This was consistent with findings in other DM 2 9 4 model systems, and suggests that zebrafish Mbnl proteins played a direct role in 2 9 5 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020.  Zebrafish mbnl mutants exhibited disease-relevant alternative splicing changes 2 9 8 Given the widespread changes in alternative splicing in mbnl mutant zebrafish, 2 9 9 we asked whether these changes were conserved with those identified in human DM 3 0 0 patients. Using publicly available datasets, we identified cassette exons whose  As anticipated, mbnl1 exon 5 and mbnl2 exon 5 appeared on both lists of 3 1 0 orthologous misregulated cassette exons, and PSI values were strikingly similar as of disease-associated exons that were misregulated in mbnl mutant model fish, some of 3 1 8 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 1 5 which were only annotated in an earlier zebrafish genome assembly (GRCz10, Table   3 1 9 EV3). For example, the decreased inclusion of alternative exons within the human 3 2 0 ATP2A1 and RYR1 genes contributes to altered calcium homeostasis in DM muscle,  alternative splicing of the zebrafish orthologs of CLCN1, whose splicing misregulation 3 3 0 contributes to myotonia in DM. Thus, the zebrafish may not be an appropriate model for 3 3 1 studying myotonia, although it can be used to study many other important disease  mutants. These fish exhibited molecular and physical phenotypes similar to those 3 3 7 observed in humans with DM and mouse models, including decreased body size, 3 3 8 impaired movement, and widespread changes in alternative splicing. We propose that zebrafish 1 B/B ;2 A/A and 1 B/B ;2 A/A ;3 F/F mutants, which exhibited the most dramatic 3 4 0 phenotypes, represent severe forms of DM1 with long CTG repeats in which most, but 3 4 1 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. These new zebrafish models, and the accompanying RNA-Seq data that we that were misregulated in all zebrafish mbnl mutants except for 2 A/A ;3 C/C mutants are exons whose mis-regulation was not conserved between humans with DM and 3 5 0 zebrafish are candidates for contributing to disease-related changes in muscle histology 3 5 1 that were not observed in the zebrafish mbnl mutants (Fig EV3). Zebrafish DM models adult fish can be generated easily to study subtle or variable phenotypes and to test 3 5 4 potential therapeutics. Additionally, transparent larval zebrafish can be used to directly 3 5 5 study disease-related phenotypes, such as altered gut motility and heart abnormalities,  All zebrafish experiments were performed with the guidance and approval of the certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 1 8 ruler using a tablet computer, finclipped, and genotyped. To ensure identical density 3 8 8 and feeding conditions, fish were compared with others from the same clutch (Fig   3  8  9 EV2A-C) or different clutches that had been raised together in the same tank (Fig 1C).
For measurement of larval fish, unfed embryos of different genotypes were raised to 7 3 9 1 days post fertilization (dpf) in separate dishes at a density of one fish per mL of embryo 3 9 2 medium, with the exception of the larvae in Fig 1E, which were grown in the same dish methane sulfonate, laid out on a microscope slide in 3% methylcellulose, and imaged 3 9 5 using a Leica M165FC microscope. An investigator that was blinded to genotype    certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 1 9 RNA (20-200 ng) was reverse transcribed with Superscript II Reverse  (Table EV1). The resulting cDNA was amplified by 28-31 cycles of PCR using 4 1 3 Taq polymerase and the forward (F) and reverse (R) primers indicated in Table EV1. Samples were separated by electrophoresis on a 6% bis-Acrylamide (19:1) gel that was  and then a cDNA library was prepared using the NEBNext Ultra RNA Library Prep Kit Analyzer NGS Analysis DNF-474 kit (Advanced Analytical) and quantitated using the  0.11.8) and aligned to the GRCz11 zebrafish genome using STAR (version 2.5.1b).

3 6
The alternative splicing events were then analyzed using rMATS (version 4.0.2), SRA, study # SRP158284. The data were processed as described above and then 4 4 7 compared to the zebrafish data for orthologous mis-splicing events. Orthologous exons 4 4 8 were found using a custom python script and using tblastx to confirm. Exons were 4 4 9 counted as orthologous if the genes from which they were transcribed were at least 4 5 0 75% conserved between species, the exon was in the same place in the transcript  with non-regulated exons)) was plotted in heatmap form. RT-qPCR was used to measure mbnl RNA levels in zebrafish tissues. Total 4 6 8 RNA was prepared from 5 dpf larvae and adult tissues using the same procedure as for  certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 2 2 calculate relative mRNA levels. Five biological samples of each genotype or tissue type 4 7 7 were run in triplicate and data were normalized to the expression of the housekeeping 4 7 8 gene eef1a1l1 (Vanhauwaert et al, 2014). Primers used for RT-qPCR are shown in 4 7 9 Table EV1. protein expression), whole 5 dpf zebrafish larvae, adult skeletal muscle, heart, brain, consisting of a custom glass aquarium measuring 18 cm deep X 25.4 cm long X 7.6 cm 4 9 5 wide (Fig 2A). The fish were monitored with a Logitech camera during a 5-minute time 4 9 6 frame to characterize basic exploratory swimming behavior. The raw tracking data was 4 9 7 analyzed with custom software (DaniOPEN, https://github.com/stednitzs/daniopen) to 4 9 8 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint 2 3 measure swim distance and distance from the bottom of the tank (Stednitz et al, 2018).

9 9
Following the test period, the fish were returned to their home tanks. were taken using a Leica DMLB microscope.  We thank the staff of the University of Oregon Zebrafish Facility for maintaining fish and  Research reported in this publication was supported by the National Institute of General award numbers 1P50GM098911 and 1P01GM125576 to K.G. and 5R01GM121862 to 5 2 1 certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that they have no conflict of interest. certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Comai L, Reddy S (2016) Muscleblind-like 3 deficit results in a spectrum of age- certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was not this version posted April 17, 2020. ; https://doi.org/10.1101/665380 doi: bioRxiv preprint  one-way ANOVA with Tukey's multiple comparisons test. Data bars that do not share a 8 4 7 letter above them are significantly different from one another.  certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.