Whole exome sequencing and database research

Exome capture was performed according to the Illumina Nextera Rapid Capture Enrichment library preparation protocol (individuals IV.1, III.4 and III.5) using 50 ng of genomic DNA. Paired-end sequencing of the libraries was performed with a NextSeq500 sequencer and the v2 reagent kit (Illumina, San Diego, California, USA). Sequences were mapped to the human genome reference (NCBI build 37/hg19 version) using the Burrows-Wheeler Aligner. Aligned reads ranged between 85,736,827 and 99,719,268. The mean coverage was ≥ 50 with 91.2%. 99% of the exome were covered at least 10x. A total of 678,344–961,783 variants per sample were called and analyzed using GensearchNGS software (PhenoSystems SA, Braine le Chateau, Belgium). Variants with a coverage of ≤ 10, a Phred-scaled quality of ≤ 15, a frequency of ≤ 15, and a MAF of ≥ 2% were neglected. Six control samples from healthy individuals were used for filtering out platform artefacts. Alamut Visual (Interactive Biosoftware, Rouen, France) software including prediction tools like SIFT, MutationTaster, PolyPhen-2, CADD-, and REVEL-Score was used for variant prioritization. Potential effects of a variant on pre-mRNA splicing were evaluated by SpliceAI, SpliceSiteFinder-like, MaxEntScan, NNSPLICE, GeneSplicer, Human Splicing Finder, ESEfinder, RESCUE-ESE, and EX-SKIP. Population databases like gnomAD v4.0.0, and GME revealed whether a variant has been previously found. Protein expression, structure, and functional aspects were investigated with UniProt and The Human Protein Atlas. Information on mouse and zebrafish models was retrieved from the MGI and ZFIN database, respectively.

Sanger sequencing

MSGN1 exon one was amplified by a touchdown PCR program using primers in the flanking introns (forward: 5´-GGTGGACTACAATATGTTAGCTTTCC-3´ and reverse: 5´-TAGACAGGTGGCAGGTAATTCC-3´). A clean-up step with ExoSAP-IT (Applied Biosystems, Foster City, USA) was followed by the sequencing reaction using the BigDye Terminator Cycle Sequencing Kit v1.1 (Applied Biosystems, Waltham, USA). Sequencing was conducted on a 3130XL capillary sequencer (Applied Biosystems, Waltham, USA) and data analysis was performed with Gensearch (PhenoSystems SA, Braine le Chateau, Belgium).

Whole genome sequencing

Short read genome sequencing was performed to exclude structural variants and non-coding variants in known skeletal dysplasia genes. In brief, genomic DNA was isolated from peripheral blood and sequenced at 30 × coverage using the Illumina TruSeq PCR-free protocol at the West German Genome Center (WGGC). Reads were aligned to the human genome build GRCh37/hg19 using BWA-MEM 0.7.17. The VarFish software was used for filtering and interpretation of variants including SVs (caller: Delly v.0.8.1) according to an in-house SOP [10].

In vitro experiments and intracellular MSGN1 localization measurements

Transfections were performed with 2.5 × 105 HEK 293T cells per well in 12-well plates. HEK 293T cell were initially grown on glass cover slides and subsequently transfected with two different amounts of DNA plasmids containing either the coding sequence of wild-type MSGN1 (WT) or of the MSGN1 p.(Arg125Leu) variant coupled to a FLAG-tag under control of the CMV promoter. 0.5 µg and 1.0 µg plasmid DNA per 100 µl total transfection volume was used with 2 to 3 µl FuGene HD transfection agent per well (Promega, Madison, USA, product-nr. E2311). 48 h after transfection, cells were paraformaldehyde fixed and immunofluorescence staining of Flag-tag was performed by standard protocols for cell HEK 293T cell cultures (primary antibody: monoclonal DYKDDDDK Tag Recombinant, Thermo Fisher Scientific, Waltham, USA, product-nr. 701629, RRID: AB_2532497, 1:250 dilution (2.5 µg/ml); secondary antibody: goat anti-rabbit-Alexa 594, Invitrogen/Thermo Fischer Scientific, product-nr. A-11012, RRID: AB_2534079, 1:1000 dilution (1 µg/ml)) and DAPI (Sigma-Aldrich, St. Louis, USA; product-nr. D9542, 1:5000 dilution (4 µg/ml)). Corresponding controls were performed by transfection of an CMV:GFP-Tag plasmid (positive control), by including non-transfected control cells, by omitting primary or secondary antibody incubations (negative controls). Images were taken with a Keyence BZ-X800 fluorescence microscope (Keyence, Osaka, Japan).

Intracellular localization of MSGN1 proteins was quantified by analysis of single z-level immunofluorescence images acquired by laser scanning confocal microscopy (Nikon A1 + , Nikon NIS-Elements software, Nikon Instruments, Tokyo, Japan) via ImageJ/Fiji software ( https://fiji.sc/) [32]. In contrast to analysis by classical fluorescence microscope techniques, usage of LSM enables high resolution optical sectioning of fluorescence signals in a narrow single z-level at height of the nucleus and removes out-of-focus fluorescence signals outside of the focal plane in conjunction with strict pinhole settings. Images were taken at 600-fold magnification, in two channels (MSGN1-tag: 561 nm laser excitation, 610–620 nm emission; DAPI in nucleus: 405 nm laser excitation, 470–480 nm emission), at a resolution of 1024 × 1024 pixel corresponding to a 205 × 205 µm area. To circumvent light scattering, smallest pinhole settings per laser channel were chosen. For each transfection group, three non-overlapping images were acquired, subsequent 10 cells per image were randomly selected and fluorescence intensity values were measured in ImageJ/Fiji software by straight line histograms (line length: 10 µm (50 pixel), line thickness: 2 µm (10 pixels), mean value per pixel position in length were measured at 50 positions per line and channel). Line positions were adjusted to start in the cytoplasm (low DAPI intensity signal) and end in the nucleus (high DAPI intensity signal), with line midpoint corresponding to nucleus border (position #25). Cytoplasmic values were defined as position #15, 2.5 µm away from nucleolus border. Nuclear values were defined as position #35, 2.5 µm away from nucleolus border. Background values were defined as position #15 and #35 values measured outside of cell bodies at five positions per image.

Zebrafish animal maintenance and lines

Laboratory zebrafish embryos (Danio rerio) of the AB/TU and AB/AB wild-type strain (ZDB-GENO-010924-10; ZDB-GENO-960809-7) and transgenic Tg(EPV.Tp1-Mmu.Hbb:Venus-Mmu.Odc1) (ZDB-TGCONSTRCT-120419-4, tp1:VenusPEST, [26]) were maintained as previously described under standard aquatic conditions at an average of 24 °C water temperature [1, 44]. Embryos were staged by morphological characteristics according to Kimmel et al. [14]. “hpf” and “dpf” indicate embryonic development in hours/days-post fertilization at 28.5 °C incubation temperature, respectively. All procedures involving experimental animals were performed in compliance with local animal welfare laws, guidelines, and policies. All presented experiments have been performed in zebrafish embryos and larvae younger than 5 dpf, before free-swimming and independent feeding, and thus are not regulated as animal experiments in Germany under current legislation.

Mosaic transient-transgenic msgn1 overexpression and in vivo visualization

Mosaic transient-transgenic overexpression was performed by microinjection of a previously published plasmid containing a zebrafish msgn1 promoter fragment driving 2A-coupled expression of mCherry fluorescence protein and wild-type zebrafish msgn1 [46]. Injection of this plasmid results in prominent and severe tail malformations due to gain of Msgn1 function in mesodermal progenitor cells. Injection of msgn1 injection solution (DNA plasmid was diluted in water to a final concentration of 25 ng/µl and admixed with 0.05% Phenol-red (pH 7.0) for visualization of injection solution) into only one cell of early blastula zebrafish embryos (4-cell stage/1.0 hpf up to 16-cell stage/1.5 hpf) resulted in a fraction of msgn1 overexpressing cells within developing embryos. In addition to zebrafish wild-type msgn1 (msgn1:mCherry-2A-msgn1), plasmids containing either a zebrafish msgn1 p.(Arg72Leu) variant (msgn1:mCherry-2A-msgn1 (p.R71L)), or a human MSGN1 p.(Arg125Leu) variant (msgn1:mCherry-2A-MSGN1 (p.R125L)) were injected in a similar way into zebrafish embryos and investigated.

Injected embryos were in vivo analyzed under a Fluorescence Stereomicroscope (Leica S8 APO equipped with Leica GFP and DSR filter sets (filter nr. 10447408 and 10447412), Leica Miscrosystems, Wetzlar, Germany) at different time points during embryonic and early larval development (approximately between 16 and 48 hpf) for cells showing msgn1 overexpression by mCherry fluorescence and for resulting developmental consequences. Injection was performed in embryos of a tp1:VenusPEST transgenic line [26], which enables in vivo visualization of Notch signaling due to expression of a short-half-life version of the fluorescence protein Venus under the tp1 promoter element. The tp1 element contains 12 EBV terminal protein 1 (TP1) gene promoter fragments for endogenous Notch (NICCD) and RBPJ/CB1/Su(H) co-factor binding. Detailed microscopic investigation was performed with Zeiss Imager A1 (in situ hybridizations, Carl Zeiss AG, Jena, Germany) or a Nikon A1 + Laser scanning confocal microscope (in vivo, Nikon Corporation, Tokyo, Japan). For detailed microscopic investigations embryos were short time fixed in 4% paraformaldehyde/PBS for 30 min and were mounted in Mowiol. Nuclei were stained by DAPI incubation (1 µg/mL in PBST; 15 min) before mounting. Images were analyzed with ImageJ/Fiji (https://fiji.sc/) and arranged with CorelDraw Graphics Suite (Alludo, Canada) software.

Plasmid vector cloning and mutagenesis

In vitro HEK 293T transfection experiments were performed with CMV promoter driven human tagged MSGN1 (Origene, Rockville, USA, product-nr. RC225212, CDS: NM_001105569). Patient MSGN1 variant was introduced into the plasmid by site-directed mutagenesis (Q5 Site-Directed Mutagenesis Kit, New England Biolabs/NEB, Ipswich, USA, product-nr. E0554S) and was validated via Sanger sequencing.

Zebrafish injection experiments were performed with msgn1 promoter driven zebrafish msgn1 plasmid (sk-tol2-msgn1:mCherry-2A-msgn1) [46]. Human MSGN1 c.374G > T, p.(Arg125Leu) and zebrafish msgn1 c.211AG > CT p.(Arg71Leu) missense variants were introduced by site-directed mutagenesis in subcloned coding sequences without start codons (Q5 Site-Directed Mutagenesis and PCR cloning Kit, New England Biolabs/NEB, product-nr. E0554S and E1202S) and were validated via Sanger sequencing. Subsequently, the newly established variant coding sequences were used to replace zebrafish msgn1 (WT) in sk-tol2-msgn1:mCherry-2A-msgn1 by restriction site cloning. Primer sequences and used plasmids are listed in Additional file 8: Table S2.

Generation of mutant msgn1 Arg71Leu mRNA and overexpression in zebrafish

A cDNA library was extracted from embryos and wild-type msgn1 was amplified with PCR using primers listed in Additional file 8: Table S2. Mutant msgn1 p.(Arg71Leu) was created using overlap extension primers. The wild-type msgn1 and mutant msgn1 p.(Arg71Leu) cDNA were cloned into separate pCS2 + vectors between BamHI and EcoRI restriction sites. In vitro transcription was used to create and isolate mRNA (mMESSAGE mMACHINE™ SP6 Transcription Kit, Invitrogen/Thermo Fisher Scientific, Cat# AM1340). Wild-type embryos were either injected with 200 pg wild-type msgn1 mRNA (58 embryos) or msgn1 p.(Arg71Leu) mRNA (54 embryos) at one-cell stage and compared with untreated ones (49 embryos) in three independent experiments. The embryos were fixed at 8-somite stage in 4% paraformaldehyde before in situ hybridization.

Zebrafish embryo RNA in situ hybridization

RNA in situ hybridization was performed according to standard protocols [39]. RNA probes were synthesized from cloned partial mRNA sequences of target genes using the DIG or FLU RNA Labeling Kit (Roche, Basel, Switzerland, product-nr. 11685619910 and 11175025910). All detected expressions patterns with newly established RNA probes were in accordance with previously published and ZFIN database patterns. Used in situ probes were targeted against: tbxta (ZDB-GENE-980526-437; synonyms: T/ta/brachyury/no tail/ntl), msgn1 (ZDB-GENE-030722-1; synonyms: mespo), bmp2a (ZDB-GENE-980526-388), tbx6 (ZDB-GENE-020416-5; synonyms: tbx6r/fss/fused somites/tbx24), tbx16 (ZDB-GENE-990615-5; synonyms: spt/spadetail). In-situ experiments were performed two times independently and included n ≥ 10 embryos per sample and condition. Detailed information and primers used for probe cloning are listed in Additional file 8: Table S2.

Image and statistical analysis of RNA in situ hybridization embryos

The embryos were flat mounted via dissection of the yolk sac and imaged under a Nikon SMZ1500 stereomicroscope (HR Plan Apo 1X WD 54), Nikon DS-Ri1 digital camera with reflected light at 23 °C room temperature. FIJI software (ImageJ 1.54f) [32] was used to assess intensity of tbxta staining in the tailbud using a standardized circular region (120 µm in diameter, ROI seen in Fig. 4A). Images were first inverted, then the mean intensity of anterior tissue background (17 µm in diameter) was subtracted from the mean intensity of tailbud (ROI). Then each intensity was normalized to the average intensity of uninjected embryos in an entire experiment.

We used unpaired two-tailed Kruskal–Wallis nonparametric test without equal standard deviation assumption in Fig. 4D. The statistical tests and distribution calculations (median and quartiles, confidence intervals) were performed in GraphPad Prism 9.5.0 software.