Oligodendrocyte-specific transcriptomics reveals OL-enriched expression of ZDHHC9 and GOLGA7.

A: Fluorescent image of widespread EGFP expression in the brain of Mobp-EGFP mice (Left). Confocal images of EGFP and other cell-specific markers (NG2 for OPC, CC1 for OLs, and NeuN for neurons) (Right). White arrowheads indicate colocalization of EGFP+ cells and CC1-immunoreactivities. Boxed area in left panel corresponds to the region used to acquire confocal images for NG2 and CC1 in right panel. Scale bars: 500 (left) and 20 (right) µm. B: Heatmap for expression levels of previously reported cell type-specific gene clusters from the OL-specific RNA-Seq results. EGFP+ OLs were isolated from cortices of three Mobp-EGFP mice with FACS, and their RNAs were used for RNA-Seq. C, D: FPKM values of ZDHHC-protein acyltransferases (PATs) (C) and Golgin subfamily A members (D) expressed in OLs. E: Heatmap of relative expression of all PATs and Golgin subfamily A member genes in mouse OLs and other brain cells according to data from this study (left) and Zhang et al., 2014 (Ref. #28; right). F: Heatmap of relative expression of all PATs and Golgin subfamily A member genes in human OLs and other brain cells according to Zhang et al., 2016 (Ref. #29).

HA-ZDHHC9WT localizes to puncta in OL processes, but other PATs, and XLID mutant forms of ZDHHC9, are restricted to OL cell bodies.

A: Experimental timeline. B: Images of morphologically mature OLs transfected as in A to express the indicated HA-tagged PATs and immunostained with the indicated antibodies. Lower panels show enlarged images of the boxed regions in upper panels. C: Images as in lower panels of B of OLs transfected to express WT HA-ZDHHC9 or the indicated XLID mutant forms of ZDHHC9. D: Quantified data confirms that HA-ZDHHC9 WT occupies a greater percentage of the total OL area (GFP+ area) compared with other PATs, or with HA-ZDHHC9 XLID mutants. *: p=0.032 (HA-ZDHHC9WT vs HA-ZDHHC3); p=0.0053 (HA-ZDHHC9WT vs HA-ZDHHC7); p=0.004 (HA-ZDHHC9WT vs HA-ZDHHC17); p=0.0036 (HA-ZDHHC9WT vs HA-ZDHHC9-R96W); p=0.0058 (HA-ZDHHC9WT vs HA-ZDHHC9-P150S); p=0.0047 (HA-ZDHHC9WT vs HA-ZDHHC9-R148W), Data are from 3-5 cells per condition, pooled from n=3 cultures.

No detectable gross abnormality in oligodendrocyte development in Zdhhc9 KO mice.

A: Fluorescent (upper) and confocal (bottom) images of MBP immunostaining in the brain of 6-week-old WT and Zdhhc9 KO mice. MBP confocal images were taken from layers IV/V of CTX. Scale bars: 500 (upper panel) and 20 (bottom) µm. B: Fluorescent (upper) and confocal (bottom) images of EGFP in the brain of 8-week-old control (Mobp-EGFP) and Zdhhc9 KO (Mobp-EGFP; ZDHHC9y/-) male mice. EGFP confocal images were taken from the CC. Scale bars: 500 (upper panel) and 50 (bottom) µm. C: Quantification of EGFP+ OLs in CTX, CC, and spinal cord (SC) from 4-week (P28) and 8-week (P56)-old control and Zdhhc9 KO Mobp-EGFP mice (n=3 – 5 mice for each group). Two-way ANOVA and pair-wise comparison was performed with Šidák’s multiple comparison test. ns: not significant. D: Confocal images of cortical NG2+ OPCs in 8-week-old WT and Zdhhc9 KO male mice. Scale bar: 50 µm. E: Quantification of the density of NG2+ OPCs in CTX of 4-week-old WT and Zdhhc9 KO mice (n=3 per group). Unpaired Student’s t-test. F-1: Experimental scheme for OPC fate tracing. Pdgfra-CreER; RCE and Pdgfra-CreER; RCE; Zdhhc9 KO mice were administered with tamoxifen injection starting at P21 for 3 days, and mouse brains were sampled at P42. F-2: Schematic diagram of OPC fate tracing. G: Confocal images of EGFP, NG2, and ASPA in CTX of control and Zdhhc9 KO Pdgfra-CreER; RCE mice (P21+21). Yellow arrowheads: overlapping signals between EGFP and NG2; white arrows: overlapped signals between EGFP and ASPA. Scale bars: 50 µm. H: Quantification of EGFP+NG2+ OPCs and EGFP+ASPA+ OLs. I: Percentage of NG2+ OPCs and ASPA+ OLs among EGFP+ cells. Data are mean ± SEM (E, H, I). For H and I, N= 10 (control) or 7 (Zdhhc9 KO) male and female mice. Two-way ANOVA and pair-wise comparisons were performed with Šidák’s multiple comparison tests. ns: not significant.

Genetic sparse cell labeling reveals altered complexity of oligodendrocyte processes in Zdhhc9 KO mice.

A: Experimental flow of the sparse genetic labeling of OLs, from acquisition of confocal images of mEGFP from CTX in P56 Mobp-iCreER; mT/mG mice to 3D OL process tracing to Sholl analysis. mEGFP signal allows for detailed process morphology of individual OLs, compared to more complex ‘bulk’ MBP signal. B: Representative results from tracing of OL processes. Scale bars: 20 µm. C, D: Branch numbers (C) and process length (D) were compared between control (Mobp-iCreER; mT/mG) and Zdhhc9 KO (Mobp-iCreER; mT/mG; Zdhhc9 KO) mice. N=15 OLs (5 OLs per mouse, three mice per group). Two-way ANOVA and pair-wise comparisons were performed with Šidák’s multiple comparison tests (C and D). ns: not significant. *, p <0.05; **, p <0.01.

Abnormal oligodendrocyte processes in Zdhhc9 KO mice.

A: Confocal microscopy of mEGFP in P56 Mobp-iCreER; mT/mG mice reveals morphological abnormalities, such as non-homogenous EGFP+ cell processes and spheroid-like membrane folding (yellow arrowheads). Scale bar: 20 µm. B: 3D-reconstruction of EGFP+ OLs and DAPI+ nuclei from the images shown in (A) with the Imaris software. Scale bar: 7 µm. C: OL processes and connected OL cell body were traced. Two different cells are shown in magenta and green. Arrowheads indicate spheroid-like swelling. D, E: Quantification of process abnormalities per field (D) and per cell (E). N=15 cells from 3 mice per group. Student’s t-test. ***, p < 0.001.

Altered myelination in Zdhhc9 KO corpus callosum in vivo.

A: Electron micrographs of corpus callosum axons from P56 male mice of the indicated genotypes. Wild type axons are uniformly myelinated, but in Zdhhc9 KO some large axons are hypomyelinated (red dotted outline), while a subset of small axons appear hypermyelinated (blue dotted outline). There is also frequent dysmyelination (green asterisk). B: Quantification of images from A reveal no change in total number of axons in Zdhhc9 KO. n.s.: not significant (p=0.7117, t test, N=3 mice per genotype). C: Percentage of unmyelinated axons from mice of the indicated genotypes (N=3 mice per genotype; *;p=0.0036, t test). D: number of hypermyelinated small axons per field in corpus callosum images from mice of the indicated genotypes (N=3 mice per genotype; *; p=0.0150, t test). E: G-ratio of small diameter axons (0.1-0.5 μm) from mice of the indicated genotypes. F: Average G-ratio from E confirms hypermyelination of small diameter axons in P50 Zdhhc9 KO mice. G: as A, but from P30 mice. Deficits in myelination are already evident at this time. H: As E, but for all axons from G. I: Increased heterogeneity of myelination in Zdhhc9 KO corpus callosum at P30, represented by an increased interquartile range of G-ratios (*;p=0.0425, t test).

ZDHHC9 loss cell-autonomously impairs maturation of cultured OLs.

A: Timeline of experiment. B: Immunofluorescent images of cultured OLs, after infection with the indicated lentiviruses and fixation as in A. C: Quantified data from B reveal that Zdhhc9 knockdown reduces the percentage of GFP-expressing (virally infected) MBP+ cells i.e., mature OLs (**; p=0.0067, t-test, n=4 individual cultures per condition). D: Likewise, Zdhhc9 knockdown reduces the percentage of MBP-expressing GFP+ cells. (*; p=0.0359, t-test, n=4 individual cultures per condition). E: Left: Images of individual OLs after infection and fixation as in B. Middle column: reconstructed outline of individual OLs. Right column: Images from middle column with superimposed concentric circles for Sholl analysis. F: Sholl analysis from OLs reconstructed as in E confirms reduced morphological complexity of Zdhhc9 knockdown OLs. n=8 cells per condition. *:p<0.05; **:p<0.01; ***:p<0.001; ****:p<0.0001, individual t tests.

ZDHHC9 palmitoylates MBP in cultured cells and in vivo.

A: Western blots of palmitoyl- and total proteins from HEK293T cells transfected to express the indicated cDNAs. B: Quantified data from C confirms that co-transfection of HA-ZDHHC9 and myc-Golga7 greatly increase MBP palmitoylation. No signal for palmitoyl-MBP is seen in the absence of the key reagent hydroxylamine (NH2OH), confirming assay specificity. **:p<0.01, Kruskal-Wallis test, n=5-6 experiments per condition. C: Western blots to detect MBP in total lysates and ABE fractions from forebrain WM (CC and striatum) from mice of the indicated genotype. D: Quantified data from C confirms that Zdhhc9 loss reduces palmitoyl, but not total, levels of 17 kDa and 21.5 kDa MBP isoforms. palmitoyl:total 17.5 kDa isoform, n=15-16 per genotype; **:p=0.0078, palmitoyl:total 21 kDa isoform, n=8 per genotype,. ‘N’ number is lower for 21.5kDa MBP because in a subset of experiments this isoform was not efficiently extracted and was hence undetectable.