Proteome analysis of zebrafish myelin. (A) One-dimensional separation of myelin biochemically purified from brains and optic nerves of adult zebrafish. Myelin was separated by 1D SDS-PAGE on Tris-glycine gradient gels, and proteins were visualized by silver staining (0.5 μg load) or colloidal Coomassie (CBB250; 5 μg load). Bands are annotated that mainly comprise known myelin proteins according to mass spectrometric identification. Arrowheads indicate bands in which no known myelin proteins were identified. (B) Myelin separated on Tris-glycine gradient gels (5 μg load) before (pre-wash) or after (post-wash) an additional step of high-pH and high-salt conditions. The indicated grid divides each CBB250-stained lane into equally sized slices, which were excised for automated tryptic digest and LC-MS analysis, thereby, respectively, identifying 890 (pre-wash) and 1274 (post-wash) proteins (Supplementary Table 1). (C) Number and relative abundance of proteins identified and quantified in purified myelin by in-solution digestion and two data independent acquisition MS modes, MS^E and UDMS^E. Note that MS^E (orange) identifies fewer proteins but provides a higher dynamic range for quantification of proteins in purified myelin compared to UDMS^E. MS^E thus facilitates quantification of highly abundant myelin proteins. ppm, parts per million. (D) Venn diagram comparing the number of proteins identified in myelin by MS^E, UDMS^E, and gel-based approaches.

Relative abundance of CNS myelin proteins in zebrafish. Pie chart visualizing the MS^E dataset (Figure 1C and Supplementary Table 1). The relative protein abundance is given in percent with relative standard deviation (RSD). Note that known myelin proteins constitute about 37% of the total myelin protein, whereas proteins so far not associated with myelin constitute the remaining 63%.

Comparison of the zebrafish myelin proteome with other datasets. (A) Scatter plot of the log₂-transformed relative abundance of proteins identified in this study by MS^E in zebrafish CNS myelin against those identified by UDMS^E. Data points representing known myelin proteins are labeled in green; other data points are in gray. The correlation coefficient was calculated for known myelin proteins (green) and for all proteins (gray) identified using both MS modes. The linear regression line is plotted for all proteins. ppm, parts per million. (B) Same as (A) but plotted against the scRNA-seq-based transcriptome of mature zebrafish oligodendrocytes (mOL) according to the normalized mean of all 19 cells in the mOL cluster according to prior assessment of oligodendrocyte lineage cells (Marisca et al., 2020). (C) Venn diagram comparing known myelin proteins identified in zebrafish CNS myelin in this study with those previously identified in CNS mouse myelin (Jahn et al., 2020). If a paralog has been reported in zebrafish, the protein name of the paralog is given. (D) Same as (A) but only known myelin proteins identified by MS^E in zebrafish myelin are plotted against the CNS myelin proteome of c57Bl/6N mice as previously assessed by the same mode (Jahn et al., 2020). Only proteins present in both datasets are plotted. (E) Same as (D) but known myelin proteins identified by MS^E in zebrafish CNS myelin are plotted against the mouse PNS myelin proteome as previously assessed by the same mode (Siems et al., 2020). (F) Same as (D) but known myelin proteins identified by MS^E in mouse CNS myelin mode (Jahn et al., 2020) are plotted against those identified by MS^E in mouse PNS myelin (Siems et al., 2020).

Expression of myelin-related mRNAs in OPCs and mature oligodendrocytes. Violin plots comparing the relative abundance of selected myelin-related transcripts in OPCs and mature oligodendrocytes according to a previously established scRNAseq dataset (Marisca et al., 2020). Data are given as normalized log-transformed mRNA abundance in transcripts per million (TPM) in OPCs (gray) versus mature oligodendrocytes (mOL) (green). Median with interquartile ranges; n = 189 OPCs (combined clusters 1–4 in Marisca et al., 2020) versus n = 19 mOL (cluster 5 in Marisca et al., 2020); ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001 by Welch’s t-test. For precise p-values, see section “Materials and Methods.” The mRNAs for oligodendrocyte transcription factor 2 (olig2), myelin regulatory factor (myrf), and SRY-box transcription factor 10 (sox10) serve as markers. ns, not significant.

Expression and subcellular localization of CD59 in the zebrafish spinal cord. (A) Gene structure of zebrafish cd59. Filled boxes indicate the open reading frame. (B) tSNE plot showing cells displaying high cd59 transcript abundance (labeled in red) by scRNAseq (Marisca et al., 2020). Cells with low or absent cd59 expression are labeled in gray. (B′) Oligodendrocyte lineage cluster layout for the tSNE plot in (B) with the 19 cells in the mature oligodendrocyte cluster (mOL; green) and the combined OPC clusters (gray). Note the overlap between high cd59 expression and the mature oligodendrocyte cluster. (C) Example micrographs of RNAscope in situ hybridization detecting cd59 transcripts in transverse spinal cord sections of Tg(olig1:nls-mApple × mbp:nlsGFP) zebrafish larvae at 5 dpf (scale bar, 10 μm). Magnified images to the right show a Cd59 + (purple) mbp:nls-EGFP (blue) mature oligodendrocyte (individual channels and merge). Seventy-two transverse sections of four animals were taken and imaged for quantification in (D,E). No Cd59 + cells were found positive for olig1:nls-mApple (green) representing OPCs. (D) Quantification of the percentage of cd59 RNAscope-positive cells that were mature oligodendrocytes (mOL) according to expression of the mbp:nlsEGFP transgene. n = 27 oligodendrocytes from 72 transverse sections taken along the entire spinal cord (anterior, mid-trunk, posterior) of four individual animals were quantified. (E) Quantification of the percentage of mbp:nlsGFP transgene positive mature oligodendrocytes (mOL) expressing cd59 transcripts to an abundance identified by RNAscope in situ hybridization. n = 28 cells from four individual animals were quantified. (F) Schematic of plasmid design for expression of a CD59 reporter construct in mature oligodendrocytes. The eyfp open reading frame was inserted after the sequence encoding the CD59 signal peptide (SP, labeled with an asterisk). (G) Example image of an individual mature oligodendrocyte at 4 dpf co-expressing membrane targeted mScarlet-CAAX and EYFP-CD59. Scale bar, 20 μm. (H) Examples of individual myelin sheaths in micrographs as in (G). Arrows point at discrete CD59 puncta, and arrowheads point to broader patches of CD59 localization. Scale bar, 10 μm. (I) Phylogeny of CD59 in an unrooted phylogenetic tree. All phylogenetic relationships are in agreement with the hypothesis that CD59 emerged at the root of vertebrates and was retained in all vertebrate groups. Two CD59 paralogs (CD59a and CD59b) exist in mice. There is no evidence of CD59 paralogs in teleost fish. (J) Immunodetection of CD59 (10-nm gold particles; black arrowheads) on cross-sectioned optic nerves of mice at postnatal day 75 (P75). Micrograph representative of 50 axon/myelin units in n = 2 biological replicates. Note that CD59 is mainly detected in compact and abaxonal myelin. Scale bar, 200 nm. (K) Immunoblot analysis of myelin biochemically purified from c57Bl/6N mouse brains at ages P18, P75, and 6 months using antibodies specific for CD59. Blot shows n = 2 biological replicates per age. CNP was detected as control.