Supplemental data

Supplemental Figures 1-9

Supplemental Videos 1-4

Supplemental Tables 1-4

Supplemental Methods

1 + + A 5kb-3’UTR Qki-5 B GSTpi/DAPI/ GSTpi/DAPI/ GSTpi/DAPI/ GSTpi Qki C Ctrl iCKO Qki-5 Qki-6 Qki-7 /GSTpi+ (%) 0 50 100150 6kb-3’UTR Qki-6 5 - 7kb-3’UTR Qki-7 **** Ctrl Qki

7kbB-3’UTR

Qki-7b 6 - ****

1 2 3 4 5 6 7a 7b 7c 8 Qki Qkloxp 7 - ATG loxP loxP T2 **** Plp-CreER Tamoxifen iCKO Qki Ctrl iCKO Ctrl iCKO D 8 E F G Ctrl iCKO **** **** Female Male Ctrl iCKO P < 0.0001 6 ** **** 30 *** 30 4 100 **** 25 25 2 50

20 20 (%)

Clinical score 0 15 15 -2 0 Overall survival 1 2 3 4 5 Body weight (g) 10 Body weight (g) 10 0 20 40 60 Weeks post tam administration Ctrl iCKO Ctrl iCKO Days post tam administration H Ctrl iCKO I Ctrl iCKO **** *** *** 100 **** 1.0 100 1.0 75 0.8 75 0.8 0.6 0.6

50 ratio 50 ratio - 0.4 - 0.4 g g 25 0.2 25 0.2 Spinal cord

Corpus callosum 0 0.0 0 0.0 % Myelinated axons Ctrl iCKO Ctrl iCKO % Myelinated axons Ctrl iCKO Ctrl iCKO J K O P Ctrl m) P I 2 ns ns N 150 µ 1.2 m * * µ

Ctrl * AnkG 100

/ 0.8

50 0.4 Caspr iCKO

P # Axons/100 0 0.0

Ctrl iCKO Axonal diameter ( Ctrl iCKO N * I Q GFAP/DAPI

iCKO 1 wpi 2 wpi 5 wpi Relative GFAP level 0 1 2 3 4

wpi ns Ctrl

L Ctrl M N **** 1

2.5 **** ) 5 iCKO

2.0 ms 4 1.5 3 wpi *** 2 1 mV 1.0 2 5 ms 0.5 1 wpi **** Velocity (m/s) iCKO CAP CAP time ( 0.0 0 5 Stimulus Ctrl iCKO Ctrl iCKO Ctrl iCKO R S IBA1/DAPI CD3/DAPI 1 wpi 2 wpi 5 wpi IBA1+ cells/mm2 1 wpi 2 wpi 5 wpi CD3+ cells/mm2 0 200 400 600 0 20 40 60 wpi ns wpi ns Ctrl Ctrl 1 1 wpi * wpi ns 2 2 wpi **** wpi **** iCKO iCKO 5 5 Ctrl iCKO Ctrl iCKO Supplemental Figure 1. Qki is essential for maintaining myelin integrity. (A) Schema depicting the splicing isoforms of Qk and the generation of oligodendrocyte-specific Qk-iCKO mice. 3’UTR, 3’ untranslated region. (B) Representative images and quantification of immunofluorescent staining of GSTpi–Qki-5, GSTpi–Qki-6, and GSTpi–Qki-7 in the corpus callosum tissues of Qk-iCKO mice and controls 5 wpi (n = 4 mice/group). Scale bars, 10 µm. (C) Representative images showing the severe hind limb weakness and paresis of Qk-iCKO mice 4 wpi. (D) The clinical scores of Qk-iCKO mice (n = 18) and controls 2 (n = 34) are shown. Tam, tamoxifen. (E) Body weights of Qk-iCKO mice and controls 4 wpi (n = 18 female mice and n = 14 male mice in the control group. n = 11 female mice and n = 8 male mice in the Qk-iCKO group). (F) The Kaplan-Meier overall survival curves (log-rank test) of Qk-iCKO mice (n = 18) and controls (n = 34) are shown. (G-I) Representative electron micrographs of cross sections of the optic nerves (G), corpus callosum tissues (H), and spinal cords (I) of Qk-iCKO mice and controls 5 wpi (n = 4 mice/group). Scale bars, 2 µm (G); 500 nm (H and I). (J) Representative electron micrographs of longitudinal sections of the optic nerves from Qk-iCKO mice and controls 5 wpi. The node (N), paranode (P, *), internode (I), and axoglial junction (arrowhead) are indicated. Scale bars, 500 nm. (K) Representative images of immunofluorescent staining of Caspr and AnkG in the optic nerves of Qk-iCKO mice and controls 5 wpi. Scale bars, 1 µm. (L-N) Representative CAP traces (L) and quantification of conduction velocities (M) and CAP time (N) recorded from Qk-iCKO mice and controls 5 wpi. n = 4 mice/group, n = 8 optic nerves/group. (O and P) Quantification of axon density (O) and axonal diameter (P) in the experimental mice in G (n = 4 mice/group). (Q-S) Representative images and quantification of immunofluorescent staining of GFAP (Q), IBA1 (R), and CD3 (S) in the corpus callosum tissues of Qk-iCKO mice and controls 1, 2, and 5 wpi (n = 4 mice/group). Scale bars, 50 µm. Data are mean s.d.; student’s t test (B, E, H, I, and M-P); two-way ANOVA with Holm-Sidak’s multiple comparisons test (D and Q-S). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.

3 ANOVA followed by Bonferroni’s post hoc test test hoc post Bonferroni’s by followed ANOVA and of quantification and images Representative of cords spinal of quantification and images Representative of tissues callosum mice/group) Qki ( injection tamoxifen after weeks5 controls and of staining immunofluorescent by cells dead of Quantification Supplemental A H G iCKO Ctrl iCKO Ctrl Clv–casp-3+ cells/mm2 controls 10 and 0 2 4 6 8 Optic nerve Optic Ctrl PDGFR GSTpi Olig2 ns iCKO . ( 1, 2, and 5 and 2, 1, a G - Figure 2. 2. Figure Qk Qki ) Representative) images andquantification of - iCKOmice Qk B intheoptic nerves and spinal cords of + 2

- Clv–casp-3 cells/mm 0 2 4 6 iCKOmice wpi Spinal cord Spinal Qki Ctrl Qki Olig2 ns ( iCKO n / Qki regulates myelin homeostasis independently of oligodendrocytedeath.homeostasismyelinindependently of regulates = 4 mice/group). Scale bars, 50 bars, Scale mice/group). 4 = and / DAPI PDGFR controls and C Qki ASPA Qki ASPA ASPA controls a / - + Qki ASPA + + + 5 weeks after tamoxifen injection ( injection tamoxifen after weeks 5 / immunofluorescent immunofluorescent Optic nerve Optic DAPI 0 Ctrl +

n + + cells/mm 5 weeks after tamoxifen injection tamoxifen after weeks 5 Olig2 Olig2 ( 1,000 Qki PDGFR Qki PDGFR PDGFR C = 4 mice/group) 4 = Qki- Qki+ Olig2+ Olig2 -

- + 0 I iCKO ). ** ).

**** 2,000 Ctrl **** + ns PDGFR cells/mm 1,000 2 a a a + + +

P 0 2,000 Qk Ctrl < 0.01; **** 0.01; < iCKO ns Brain a **** D

40 **** + Qki ASPA Qki ASPA ASPA - 2 ** immunofluorescent Clv iCKOmice and controls 5weeks aftertamoxifen injection ( cells/mm µ 4 - + m. Data are mean mean are Data m. staining of staining of staining 80 I . ( ASPA iCKO – iCKO Ctrl + + + ns casp ns C 0 Spinal cord Spinal - Ctrl 2 F + cells/mm ) Quantification) of - 400 P 1 3 in the optic nerves and spinal cords of of cords spinal and nerves optic the in 3 wpi < 0.0001; ns, not significant not ns, 0.0001; < PDGFR PDGFR iCKO ****

**** 800 ns Qki PDGFR Qki PDGFR PDGFR 2 n - + = 4 mice/group). Scale bars, 50 bars, Scale mice/group). 4 = staining of Olig2and of staining ( PDGFR n PDGFR a a a a a E = 4 mice/group). Scale bars, 50 bars, Scale mice/group). 4 = s.d. Qki GSTpi Qki GSTpi GSTpi + + + and and

0 nerve Optic - + 2 Ctrl a ; student’s; GSTpi ns a wpi 40 / + + + Brdu Brdu Qki + immunofluorescent Optic nerve Optic

cells/mm 0 80 Ctrl / + DAPI

inthebrains, optic nerves, and 500 iCKO

120 cells/mm inthebrains of ns ns 1,000 2 t iCKO 1,500 test( **** **** ns 2 . 5 Qki wpi ( A A Qki PDGFR Qki PDGFR PDGFR and and inthecorpus F Qki GSTpi Qki GSTpi GSTpi - + staining of ASPAof staining Qk - + PDGFR B B µ a a a GSTpi Qk + + + + + + ); ); ) PDGFR m. ( m. - 5 wpi 2 wpi 1 wpi / PDGFR iCKOmice

0 cord Spinal 0 Spinal cord Spinal one - 10 µ iCKOmice Ctrl Ctrl Ctrl ns + I a m. ( m. ) cells/mm +

- 40 400 30 a cells/mm way way a + ns ns Brdu + H n (%) iCKO iCKO 50 80 iCKO 800 ) = 4 = ** **** **** ns ns ns + 2 2 - QD-9/DAPI Ctrl iCKO 150 ** level 100 dMBP 50

0 Corpus callosum Relative Ctrl iCKO ** 60 level 40 dMBP 20 Optic nerve 0 Relative Ctrl iCKO *** 15 level 10 dMBP 5

Spinal cord 0 Relative Ctrl iCKO Supplemental Figure 3. Loss of Qki in mature oligodendrocytes leads to accumulation of degenerating myelin. Representative images and quantification of immunofluorescent staining of the QD-9 antibody in the corpus callosum tissues, optic nerves, and spinal cords of Qk-iCKO mice and controls 5 weeks post tamoxifen injection (n = 4 mice/group). Scale bars, 50 µm. Data are mean ± s.d.; student’s t test. **P < 0.01; ***P < 0.001.

5 A Tight junction Neuroactive ligand-receptor interaction Nicotine addiction Retrograde endocannabinoid signaling Morphine addiction 8-week-old mice Proximal tubule bicarbonate reclamation Taste transduction Day 0 Tam administration Biosynthesis of unsaturated fatty acids Day 1 Tam administration HTLV-I infection Endocrine and other factor-regulated calcium reabsorption Nitrogen metabolism Basal cell carcinoma Day 4 Fatty acid metabolism PI3K-Akt signaling pathway GABAergic synapse Mineral absorption Cell adhesion molecules (CAMs) PPAR signaling pathway Caffeine metabolism Serotonergic synapse 0 1 2 3 -log10 (p-value) Lysosome Hematopoietic cell lineage Phagosome Chemokine signaling pathway Natural killer cell mediated cytotoxicity Corpus callosum tissues Cell adhesion molecules (CAMs) Legionellosis Tuberculosis Staphylococcus aureus infection Cytokine-cytokine receptor interaction Toll-like receptor signaling pathway Rheumatoid arthritis Complement and coagulation cascades NF-kappa B signaling pathway Pertussis Regulation of actin cytoskeleton Leishmaniasis RNA-seq Fc gamma R-mediated phagocytosis Toxoplasmosis B cell receptor signaling pathway 0 5 10 15 -log10 (p-value) B Differentiated OLs C Neural stem cells Biosynthesis of unsaturated fatty acids ECM-receptor interaction Fatty acid metabolism Fatty acid degradation Retinol metabolism Morphine addiction PPAR signaling pathway Fatty acid metabolism Neuroactive ligand-receptor interaction Biosynthesis of unsaturated fatty acids Fatty acid degradation PPAR signaling pathway Drug metabolism Beta-alanine metabolism Protein digestion and absorption Glycosphingolipid biosynthesis Starch and sucrose metabolism Valine, leucine and isoleucine degradation Metabolism of xenobiotics by cytochrome P450 GABAergic synapse Peroxisome Ascorbate and aldarate metabolism Fatty acid elongation Metabolic pathways Metabolic pathways Circadian entrainment Malaria Neuroactive ligand-receptor interaction Chemical carcinogenesis Salivary secretion Gastric acid secretion Rap1 signaling pathway Carbon metabolism Sphingolipid metabolism Axon guidance Pentose and glucuronate interconversions ECM-receptor interaction Glutathione metabolism Glutathione metabolism Focal adhesion 0 2 4 6 8 0 2 4 6 -log10 (p-value) -log10 (p-value) D Fatty acid metabolism Acsbg1 Scd2 Scd1 Acadl Hacd2 Cpt1a Scd4 Elovl5 Scd3 Acaa2 Hadhb Acadm Acsl6 Acox1 Fads2 Fads1 Aldh3a2 Aldh7a1 Hadha Hadh Ppt1 Echs1 Acox3 Hacd4 Eci1 Acads Acot1 Eci2 Adh5 Acot2 Acadvl Acsl3 Acaa1a Elovl2 Hsd17b12 Cpt2 Mecr Acaa1b Cpt1c Acsl4 Acsl5 Acot7 1

2 2 WT 3 score

1 - - Z / - 2 2 - Qk 3 Supplemental Figure 4. Transcriptional regulation of involved in fatty acid desaturation and elongation by Qki. (A) Schema depicting the workflow for sample collection and RNA sequencing in the corpus callosum tissues of Qk-iCKO mice and controls (left). n = 4 mice/group. The top 20 enriched pathways are shown as bar graphs (right). (B) Bar graph showing the top 20 enriched pathways in WT and Qki-depleted differentiated oligodendrocytes (n = 3 cell lines/group). (C) Bar graph showing the top 20 enriched pathways in WT and Qki-depleted NSCs (n = 4 cell lines/group). Blue and red represent pathways whose activities were decreased or increased, respectively. (D) Heat map representing the differentially expressed genes involved in fatty acid metabolism in differentiated oligodendrocytes. 6 A SFAs MUFAs PUFAs 12:0 14:0 15:0 16:0 17:0 18:0 20:0 22:0 24:0 26:0 14:1 16:1 18:1 20:1 22:1 24:1 26:1 18:2 18:3 18:4 20:2 20:3 20:4 20:5 22:2 22:4 22:5 22:6 1 2

2.5 3 Ctrl 4 5 score

- 1 Z 2

2.5 3 - iCKO 4 5 * *** ns ** B 50 C 50 D 20 F 0.09 * G 120 ** H 12 /g) /g) /g) /g) /g) /g) 40 40 15 mol mol mol mol 0.06 mol 80 8 µ mol µ µ

30 30 µ µ µ 10 20 20 0.03 40 4 10 10 5 SFAs ( LCFAs ( PUFAs ( MUFAs ( MCFAs ( 0 0 0 0.00 0 VLCFAs ( 0 Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO E 40

Ctrl ** ****

30 iCKO ***

20 *** **** * * * *** 10 **** %) ***

mol 1.0 ( * 0.8 ** 0.6 **** ** 0.4 *** ** *** ** * **

Fatty acid compositions 0.2 0.0 FA FA 12:0 FA 14:0 FA 15:0 FA 16:0 FA 17:0 FA 18:0 FA 20:0 FA 22:0 FA 24:0 FA 26:0 FA 14:1 FA 16:1 FA 18:1 FA 20:1 FA 22:1 FA 24:1 FA 26:1 FA 18:2 FA 18:3 FA 18:4 FA 20:2 FA 20:3 FA 20:4 FA 20:5 FA 22:2 FA 22:4 FA 22:5 FA 22:6 SFAs MUFAs PUFAs I Fatty acid desaturation Fatty acid elongation J Elovl5 Elovl2 Hsd17b12 Elovl5 Elovl5 Hsd17b12 Hsd17b12 Hacd2 20:3 * Hacd3 Hacd2 Hacd2 15 0.8 *** Hacd4 Hacd3 Hacd3 0.6 Fads2 Hacd4 Fads1 Hacd4 10 18:3 18:4 20:4 20:5 22:5 22:6 v3 0.4 5 0.2 22:5 / 20:5 22:4 / 20:4 0 0.0

Exogenous 18:2 18:3 20:3 20:4 22:4 22:5 v6 Ctrl iCKO Ctrl iCKO Fads2 Elovl5 Fads1 Elovl2 Elovl5 Hsd17b12 Elovl5 Hsd17b12 Hacd2 Hsd17b12 log2 (iCKO / Ctrl) 20:2 Hacd2 Hacd3 Hacd2 Hacd3 Hacd4 Hacd3 -2 2 Hacd4 Hacd4 K ND HFD ND HFD Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO

L ND HFD Ctrl iCKO Ctrl iCKO Fluoromyelin

7 Supplemental Figure 5. Disturbed fatty acid desaturation and elongation by depletion of Qki in mature oligodendrocytes. (A) Heat map plotting the differences in the concentrations of each fatty acid species as measured by mass spectrometry in the spinal cords of Qk-iCKO mice and controls 5 weeks after tamoxifen injection (n = 5 mice/group). The green boxes highlight the saturated and monounsaturated LCFAs and VLCFAs. (B-D) Quantification of the concentrations of SFAs (B), MUFAs (C), and PUFAs (D) in the samples in A (n = 5 mice/group). (E) Quantification of the compositions of fatty acids in the samples in A (n = 5 mice/group). (F-H) Quantification of the concentrations of MCFAs (F), LCFAs (G), and VLCFAs (H) in the samples in A (n = 5 mice/group). (I) Schema showing the desaturating and elongating reactions of the PUFAs with the reduction or accumulation of specific fatty acid molecules (from lipidomic data shown in A) and the downregulated corresponding (from RNA-seq data shown in Figure 3D) in Qk-iCKO mice relative to controls. The reduction in fatty acid desaturases is shown in light blue, and the reduction in enzymes of the fatty acid elongation cycle is shown in dark blue. The scale factor indicates a log2-fold change in the fatty acid concentration between Qk-iCKO mice and controls. (J) Quantification of the product/substrate ratios of representative fatty acid-elongating reactions in the samples in A (n = 5 mice/group). (K) Representative images showing the severe abnormal hind limb weakness (second panel from the left, red arrow) and paresis (sixth panel from the left, red arrow) of Qk-iCKO mice fed an ND or HFD for 5 weeks after tamoxifen injection. (L) Representative images of staining of FluoroMyelin in the caudoputamen tissues of the experimental mice in K. Images were taken from the regions of the red-boxed areas in the schema (left panel). Scale bars, 50 µm. Data are mean ± s.d.; student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.

8 A B C input IP: IgG IP: Qki-5 D Ctrl iCKO - + - + - + RNase A OLs Corpus callosum tissues

1 2 3 1 2 3 CytoplasmNucleoplasmChromatin Qki-5 PPARa Qki-5 PPARb PPARb b-actin Lyse cell 2,782 98 815 1,439 48 864 PPARg Lamin B2 RNase A digestion b-actin H2A co-IP E 40 GFP Qki-5WT Qki-5V157E 30 20 10 0 Relative gene expression Scd1 Scd2 Scd4 Ppt1 Cpt2 Fads1 Fads2 Elovl2 Elovl5 Acot1 Acot2 Acaa2 Hadh Acadl Acox1 Cpt1a Hadha Hadhb Acadm Acsbg1

F G 10 kb Hadha Hadhb Acads Acadm Acadl Eci1 Acox1 Aldh3a2 PPARb ChIP 30 PPRE p-value = 1e-21 input 0 30 Qki-5 0 60 seq HA ChIP - PPARb 0

PPRE p-value = 1e-61 ChIP 60 HA 0 30 Pol II 0

Acox3 Acot1 Eci2 Adh5 Acot2 Echs1 Acadvl Cpt1a Angptl4 Plin2 Slc25a20 Mlycd Klf10

30 input 0 30 Qki-5 0 60 seq - PPARb 0

ChIP 60 HA 0 30 Pol II 0 H I J K L Qki-5 Qki-5 PPARb 0.25 Qki-5 ChIP PPARb ChIP PPARb ChIP Qki-5 (8,380 peaks) 0.20 PPARb PPRE p-value = 1e-12 15.8% 13.3% 0.15 1,062 19.1% 59.6% 15.4% 65.6% 0.10

Total genes 5.7% 0.05 5.5% PPARb

TSS TSS Read count per -2,500 TSS 2,500 TSS (-1 kb to +100 bp) (1,393 peaks) Genomic region (5’ à 3’) Genomic region (5’ à 3’) million mapped reads Exon Intron Intergenic M N O Qki-5 ChIP-qPCR PPARb ChIP-qPCR WT (DMSO) WT (GW501516) **** ns Qk-/- (DMSO) Qk-/- (GW501516) **** ** WT WT 5 5 0.8 -/- 0.12 -/- Qk ** Qk 4 4 3 0.6

activity 2 0.08 3 Relative luciferase luciferase 1 0.4 * ns 2 0 0.04 Vector + % of input 0.2 * % of input ns 1 HA−Qki-5WT + 0.0 0.00 0 HA−Qki-5V157E + Relative gene expression HA-PPARb + + + NC NC Scd2 Scd2 Scd1 Scd4 Acadl PPRE-luciferase + + + Cpt1c Cpt1c Acaa2 Acadm Cpt1a 9 Supplemental Figure 6. Qki-5 interacts with PPARb to regulate the transcription of genes involved in fatty acid metabolism. (A) Immunoblots showing the expression of all homologs of PPAR in the corpus callosum tissues of Qk-iCKO mice and controls 5 weeks after tamoxifen injection. (B) Immunoblots showing the expression of Qki-5 in the subcellular fractions of differentiated oligodendrocytes. (C) Co-IP of Qki-5 and PPARb in differentiated oligodendrocytes in which the cell lysate was treated with RNase A. (D) Venn diagram depicting the overlap between the differentially expressed genes (Qk-iCKO group versus the control group) and genes with QRE in freshly isolated mouse oligodendrocytes (left) and adult corpus callosum tissues (right). (E) RT-qPCR evaluation of the expression levels of genes involved in fatty acid metabolism in Qk-/- cells with ectopically expressed Qki-5WT, Qki-5V157E, or GFP. (F) Significant enrichment of PPRE motifs in PPARb- and HA-binding events from differentiated oligodendrocytes with ectopic expression of HA-PPARb. (G) Representative ChIP-seq binding peaks of Qki-5, PPARb, HA, and Pol II on gene loci associated with fatty acid metabolism. The ChIP-seq peaks on six known PPARb target genes are indicated in dotted box. (H) Overlap of ChIP-seq peak sets of Qki-5 and PPARb in freshly isolated mouse oligodendrocytes. (I) Significant enrichment of PPRE motif in PPARb-binding events from freshly isolated mouse oligodendrocytes. (J) ChIP-seq density heat maps of Qki-5 and PPARb within 1 kb of the TSS in freshly isolated mouse oligodendrocytes. All peaks are rank ordered from high to low Qki-5 occupancy. (K) Average genome-wide occupancies of Qki- 5 and PPARb within 2.5 kb of the TSS in freshly isolated mouse oligodendrocytes. (L) Genomic global distribution of the ChIP-seq peaks of Qki-5 and PPARb in freshly isolated mouse oligodendrocytes. (M) ChIP-qPCR showing the recruitment of Qki-5 and PPARb to chromatin in WT and Qk-/- differentiated oligodendrocytes. NC, negative control. (N) RT-qPCR showing the transcriptional levels of representative genes involved in fatty acid metabolism in WT and Qk-/- differentiated oligodendrocytes treated with 100 nM GW501516 or vehicle (DMSO) for 48 hours. (O) Bar graph showing the expression of PPAR-responsive luciferase reporter in HEK293T cells with ectopically expressed Qki-5 (WT or V157E mutant) or empty vector. Data are representative of 3 independent experiments (A-C, E, and M-O). Data are mean s.d.; student’s t test (M); one- way ANOVA followed by Bonferroni’s post hoc test (O). *P < 0.05; **P < 0.01; ****P < 0.0001; ns, not significant.

10 Plp-CreERT2; PpardL/L (Ppard-iCKO) Oligodendrocyte-specific A Tamoxifen Ppardloxp allele Ppard knockout injection Exon 3 Exon 4 Exon 5 Exon 3 Exon 5 loxP loxP 8-week old 13-week old loxP

B Ctrl iCKO DAPI / Olig2 / b PPAR C D Ctrl iCKO E Ctrl iCKO F G ** 130 ** 100 1.0 ** 120 75 0.8 110 0.6

50 ratio 100 - 0.4 g 90 25 0.2

Latency to fall (s) 80 0 0.0 Ctrl iCKO % Myelinated axons Ctrl iCKO Ctrl iCKO

H Ctrl iCKO I Ctrl iCKO ns 2 2,000 ns 2 2,000 1,500 1,500 DAPI DAPI / /

cells/mm 1,000 1,000 cells/mm + 500 + 500 Olig2 ASPA Olig2 ASPA 0 0 Ctrl iCKO Ctrl iCKO

J Ctrl iCKO Ctrl iCKO K Ctrl iCKO PLP ns Fluoromyelin 1.2

0.8

0.4 expression Relative PLP 0.0 MBP ns 1.2

0.8

0.4 expression Relative MBP 0.0 ** MAG 1.2 1.5 ns 0.8 1.0 fluoromyelin level 0.4 0.5 expression

Relative MAG 0.0

0.0 Relative Ctrl iCKO Supplemental Figure 7. Ablation of Ppard in mature myelinating oligodendrocytes leads to demyelination with reduction in myelin lipids. (A) Schema showing the PpardLoxp allele and the generation of oligodendrocyte-specific Ppard-knockout mice. (B) Representative images of immunofluorescent staining of PPARb-Olig2 in the corpus callosum tissues of Ppard-iCKO mice and controls 5 weeks after tamoxifen injection. n = 4 mice/group. Scale bars, 50 µm. (C) Latency to fall (s) off the rotarod (20 r.p.m.) for Ppard-iCKO mice and controls 5 weeks after tamoxifen injection (n = 7 mice in the control group, n = 6 mice in the Ppard-iCKO group). (D-G) Representative electron micrographs and quantification of the optic nerves of Ppard-iCKO mice and control mice 5 weeks after tamoxifen injection (n = 3 mice/group). Scale bars, 2 µm (D); 500 nm (E). (H and I), Representative images and quantification of immunofluorescent staining of ASPA and Olig2 in the corpus callosum tissues of the experimental mice in B (n = 4 mice/group). Scale bars, 50 µm. (J) Representative images and quantification of immunofluorescent staining of PLP, MBP, and MAG in the corpus callosum tissues of the experimental mice in B (n = 4 mice/group). Scale bars, 50 µm. (K) Representative images and quantification of staining of FluoroMyelin in the experimental mice in B (n = 4 mice/group). Images were taken from the regions of the red-boxed areas in the schema. Scale bars, 50 µm. Data are mean s.d.; student’s t test. **P < 0.01; ns, not significant. 11 A Veh KD3010 Veh KD3010 Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO

B Veh Bex Veh Bex Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO

C Veh KD3010 Ctrl iCKO Ctrl iCKO Fluoromyelin

D Veh Bex Ctrl iCKO Ctrl iCKO Fluoromyelin

E Ctrl iCKO Ctrl iCKO Ctrl iCKO Ctrl iCKO * * ** * 1.5 *** 1.5 1.5 1.5 ns

ns level * level level **** level **** ** ** 1.0 1.0 1.0 1.0 Scd1 Scd2 Scd4 Elovl5 0.5 0.5 0.5 0.5

0.0 0.0 0.0 0.0 Relative Relative Relative Veh KD3010 Veh KD3010 Veh KD3010 Relative Veh KD3010 Supplemental Figure 8. PPARb and RXR agonists alleviate Qki depletion-induced demyelination. (A and B) Representative images showing severe abnormal hind limb weakness (second panel from the left, red arrow) and paresis (sixth panel from the left, red arrow) in Qk-iCKO mice receiving daily oral administration of KD3010, Bex, or Veh for 5 weeks after tamoxifen injection. (C and D) Representative images of staining of FluoroMyelin in the caudoputamen tissues of Qk-iCKO mice and controls receiving daily oral administration of KD3010, Bex, or Veh for 5 weeks after tamoxifen injection. Images were obtained from the regions of the red-boxed areas in the schema (left panel). Scale bars, 50 µm. (E) Quantification of the expression level of genes involved in fatty acid metabolism in the corpus callosum tissues of Qk-iCKO mice and controls receiving KD3010 or Veh for 5 weeks after tamoxifen injection by RT-qPCR. Data are mean ± s.d.; one-way ANOVA with Fisher’s LSD test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.

12 A B ns C D 2.5 2.0 ns 2 1,200 ** 100 2.0 1.5 Active and early demyelinating 800 75 1.5

Active and post-demyelinating 1.0 cells/mm 1.0 + 400 50 0.5 Mixed active/inactive and demyelinating expression 0.5 expression Relative MBP 25 Relative MAG % of lesions Mixed active/inactive and post-demyelinating 0.0 0.0 OLIG2 0 0 ControlPPMS ControlPPMS ControlPPMS

E Fatty acid elongation PPAR signaling pathway Sphingolipid metabolism 0.50 Control 0.20 Control 0.40 Control NES: 2.15 NES: 1.29 NES: 2.10 0.25 0.10 0.20

0.00 ELOVL1ELOVL5 HACD1 0.00 SCDACSL1SCD5 0.00 UGT8 SGPP1SPTLC2 Enrichment score Enrichment score Enrichment score

0.40 MS 0.15 MS 0.30 MS NES: 1.82 NES: 0.98 NES: 1.71 0.10 0.20 0.15 0.05 0.00 0.00 0.00 ELOVL1ELOVL5 HACD1 SCD SCD5ACSL1 SPTLC2UGT8SGPP1 Enrichment score Enrichment score Enrichment score -0.10

F Normal SCDs Demyelination ELOVLs SCDs ELOVLs …... QKI-5 …… QKI-5 L L Oligodendrocyte L L Oligodendrocyte ?

.. PPARb RXRa PPARb RXRa ...... SCDs ELOVLs SCDs ELOVLs

desaturation elongation Myelin desaturation elongation Myelin ......

saturated fatty acids unsaturated fatty acids L L ligands myelin proteins glycosylation

Supplemental Figure 9. Myelin lipid metabolism is preferentially reduced in MS lesions. (A) Percentage of MS lesions with different inflammatory and demyelinating activities. (B-D) Quantification of immunofluorescent staining of MBP (B), MAG (C), and OLIG2 (D) in the frontal lobes of 5 neurological disease-free human brains (control) and 30 lesions from 6 patients with PPMS. (E) Comparison of the GSEA signatures of lipid metabolism pathways between the MS lesions and the controls. (F) Schematic of the essential role of Qki-PPARb-RXRa complex-regulated fatty acid desaturation and elongation in mature myelin maintenance. Data are mean ± s.d.; student’s t test. **P < 0.01; ns, not significant.

13 Supplemental Video 1. Severe abnormal hind limb weakness and paresis in Qk-iCKO mice.

8-week-old Qk-iCKO mice and control mice were injected with tamoxifen and the severity of abnormal hind limb weakness and paresis of these animals was recorded 5 weeks after injection.

Supplemental Video 2. High-fat diet alleviates Qki-depletion–induced neurological deficits.

8-week-old tamoxifen-injected Qk-iCKO mice and control mice were fed either a normal diet

(ND) or the high-fat diet (HFD). The severity of abnormal hind limb weakness and paresis of these animals was recorded 5 weeks after treatment.

Supplemental Video 3. PPARb agonist alleviates Qki-depletion–induced neurological deficits. 8-week-old tamoxifen-injected Qk-iCKO mice and control mice were orally administered 50 mg/kg/day KD3010 or vehicle (Veh). The severity of abnormal hind limb weakness and paresis of these animals was recorded 5 weeks after treatment.

Supplemental Video 4. RXR agonist alleviates Qki-depletion–induced neurological deficits.

8-week-old tamoxifen-injected Qk-iCKO mice and control mice were orally administered 100 mg/kg/day bexarotene (Bex) or vehicle. The severity of abnormal hind limb weakness and paresis of these animals was recorded 5 weeks after treatment.

Supplemental Table 1 (separate Excel file). Lipidomic analysis of the concentrations of lipids in the spinal cords of Qk-iCKO mice and control mice. Tamoxifen was administered to

8-week-old Qk-iCKO mice and control mice (5 in each group), and their spinal cords were collected 5 weeks later for further lipidomic analysis.

14

Supplemental Table 2 (separate Excel file). Lipidomic analysis of the compositions of lipids in the spinal cords of Qk-iCKO mice and control mice. Tamoxifen was administered to 8- week-old Qk-iCKO mice and control mice (5 in each group), and their spinal cords were collected 5 weeks later for further lipidomic analysis.

15 Supplemental Table 3. Human brain sample information.

RMMSCA Tissue type Age PMIB Sex Cause of Disease Disease-modifying # at death duration therapies / death immunomodulat- ory therapy 408 Control 74 48 hours M unknown N/A N/A 482 Control 46 N/A F unknown N/A N/A 487 Control 59 19 hours M unknown N/A N/A 489 Control 73 N/A F unknown N/A N/A 503 Control 55 N/A F unknown N/A N/A Primary progressive 436 60 10 hours M unknown unknown unknown multiple sclerosis Primary progressive 438 75 20 hours F unknown 36 years none multiple sclerosis Primary slow 449 progressive multiple 74 7 hours F unknown unknown none sclerosis Primary slow Respiratory 495 progressive multiple 78 20 hours F 33 years none failure sclerosis Primary progressive 507 69 0 hour F unknown 46 years Natalizumab multiple sclerosis symptoms Primary slow 43 years; 534 progressive multiple 71 43 hours F unknown unknown diagnosed sclerosis 23 years

ARocky Mountain Multiple Sclerosis Center

Bpostmortem interval

Supplemental Table 3. Human brain sample information. The patients’ age at the time of death, sex, postmortem interval, pathological diagnosis, cause of death, disease duration, and disease-modifying therapies are shown.

16 Supplemental Table 4. Primer sequences.

Primer Name Primer sequence Comments Scd1-Forward CCTCTTCGGGATTTTCTACTACAT RT-qPCR Scd1-Reverse TCCAGTTTTCCGCCCTTCTC RT-qPCR Scd2-Forward AAGATGATCTATATGACCCCACCT RT-qPCR Scd2-Reverse CACGTCATTCTGGAACGCCA RT-qPCR Scd4-Forward TGTCTGACCTGAAAGCCGAG RT-qPCR Scd4-Reverse GTGGAAGCCCTCGCCTAAAG RT-qPCR Fads1-Forward GTGATCGACCGGAAGGTGTA RT-qPCR Fads1-Reverse AGCGCTTTATTCTTGGTGGGT RT-qPCR Fads2-Forward CATCGACCGCAAGGTCTACA RT-qPCR Fads2-Reverse TCTGAGAGCTTTTGCCACGG RT-qPCR Elovl2-Forward TACCCTGGACAGCGCATCG RT-qPCR Elovl2-Reverse AGAGGATGAGCTCCACCAGCA RT-qPCR Elovl5-Forward ATGGAACATTTCGATGCGTCA RT-qPCR Elovl5-Reverse GTCCCAGCCATACAATGAGTAAG RT-qPCR Acot1-Forward ATACCCCCTGTGACTATCCTGA RT-qPCR Acot1-Reverse CAAACACTCACTACCCAACTGT RT-qPCR Acot2-Forward CTCTTCCTGCCCCCAGAACC RT-qPCR Acot2-Reverse CAGCCCAATTCCAGGTCCTT RT-qPCR Acaa2-Forward CTGCTACGAGGTGTGTTCATC RT-qPCR Acaa2-Reverse AGCTCTGCATGACATTGCCC RT-qPCR Hadh-Forward TCGTGAACCGACTCTTGGTG RT-qPCR Hadh-Reverse ATTTCATGCCACCCGTCCAA RT-qPCR Hadha-Forward TGCATTTGCCGCAGCTTTAC RT-qPCR Hadha-Reverse GTTGGCCCAGATTTCGTTCA RT-qPCR Hadhb-Forward ACTACATCAAAATGGGCTCTCAG RT-qPCR Hadhb-Reverse AGCAGAAATGGAATGCGGACC RT-qPCR Ppt1-Forward ACCGCTGGTGATCTGGCAT RT-qPCR Ppt1-Reverse TCTCCACATCCTCCATCATGTTCT RT-qPCR Acadl-Forward TCTTTTCCTCGGAGCATGACA RT-qPCR Acadl-Reverse GACCTCTCTACTCACTTCTCCAG RT-qPCR Acadm-Forward AGGGTTTAGTTTTGAGTTGACGG RT-qPCR Acadm-Reverse CCCCGCTTTTGTCATATTCCG RT-qPCR Acsbg1-Forward ATGCCACGCGGTTCTGAAG RT-qPCR Acsbg1-Reverse GAGCTGGTTTGCGAGTTGTCT RT-qPCR Acox1-Forward TAACTTCCTCACTCGAAGCCA RT-qPCR Acox1-Reverse AGTTCCATGACCCATCTCTGTC RT-qPCR Cpt1a-Forward CTCCGCCTGAGCCATGAAG RT-qPCR Cpt1a-Reverse CACCAGTGATGATGCCATTCT RT-qPCR Cpt2-Forward CAGCACAGCATCGTACCCA RT-qPCR Cpt2-Reverse TCCCAATGCCGTTCTCAAAAT RT-qPCR Actb-Forward CCACCATGTACCCAGGCATT RT-qPCR Actb-Reverse CCGATCCACACAGAGTACTT RT-qPCR

17 Continued mC-Scd2-Pmt-Forward ATGGAGTTGCAAGTGGTCGG ChIP-qPCR mC-Scd2-Pmt-Reverse GCTTCAGCGCTTTCTCTGGA ChIP-qPCR mC-Cpt1c-Pmt-Forward CATCGAGGGATGCTGGTGAG ChIP-qPCR mC-Cpt1c-Pmt-Reverse GCGAGTAGGGCTTCTCCATC ChIP-qPCR mC-NC-Forward ATGCCTAACTTCCAGTTCCAGG ChIP-qPCR mC-NC-Reverse AGCTTAGAGCAGAAAGCTGGT ChIP-qPCR

Supplemental Table 4. Primer sequences. A complete list of the sequences of the primers used in this study is shown.

18 Supplemental Methods

RNA isolation and real-time qPCR. RNA was extracted from the corpus callosum tissues of the mouse brains, freshly isolated mouse oligodendrocytes, NSCs, and differentiated oligodendrocytes using an RNeasy Mini Kit, according to the manufacturer’s instructions

(Qiagen). Two micrograms of RNA were reverse transcribed to cDNA using SuperScript III

First-Strand Synthesis SuperMix (Thermo Fisher Scientific), and real-time qPCR was performed using iTaq Universal SYBR Green Supermix (Bio-rad) in an Applied Biosystems

7500 Fast Real-Time PCR system. The relative transcription levels of the genes of interest were normalized against Actb and calculated using the ΔΔCT method. A complete list of the primer sequences used in this study can be found in Supplemental Table 4.

GSEA in human tissue samples. Raw gene expression profiling files were downloaded from the

Gene Expression Omnibus under accession number GSE38010 (40). The latest version of custom CDF files (v22) (59, 60) was used to map probes from the Affymetrix

U133 Plus 2.0 Array to the Ensemble transcript database, resulting in robust multi-array average- normalized and log-transformed gene expression values (61). A single-sample GSEA was performed on the expression profiles of 4 cases (CAP1, CAP2, CP1, and CP2) and 2 controls

(control 1 and control 2) for 8 pathways (fatty acid metabolism [KEGG: 01212], fatty acid biosynthesis [KEGG: 00061], biosynthesis of unsaturated fatty acids [KEGG: 01040], fatty acid elongation [KEGG: 00062], fatty acid degradation [KEGG: 00071], PPAR signaling pathway

[KEGG: 03320], sphingolipid metabolism [KEGG: 00600], and glycerophospholipid metabolism

[KEGG: 00564]). Normalized enrichment scores were calculated for representation and further

19 analyses. GSEA (v2.2.4) was used to generate the representative enrichment plots for pathways per sample (58).

Genome-wide computational analysis for RNA motifs of Qki. A computational screening of the mouse reference genome (mm10 version) with the QRE motif (UACUAAC) was performed on

MotifMap (http://motifmap.ics.uci.edu) (62).

Immunoprecipitation and immunoblotting. Differentiated oligodendrocytes were crosslinked by

1% formaldehyde for 10 minutes and lysed in NP-40 buffer (50 mM Tris-HCl [pH 7.5], 150 mM

NaCl, 5 mM EDTA, and 0.3% NP-40) for 30 minutes at 4°C. The whole-cell lysate was sheared using a Bioruptor Pico sonication device (Diagenode) for 30 cycles (30 seconds on, 30 seconds off) and centrifuged at 13,000 g at 4°C for 15 minutes. The supernatant was incubated with antibodies against Qki-5 (immunizing rabbit with a short synthetic peptide

[CGAVATKVRRHDMRVHPYQRIVTADRAATGN]), HA (ab18181; Abcam), PPARb (PA1-823A;

Thermo Fisher Scientific), or normal immunoglobulin G overnight at 4°C. Magnetic beads with recombinant protein G (Thermo Fisher Scientific) were added to the immunoprecipitation system for another 2 hours at 4°C. The immunoprecipitates were washed extensively with NP-40 buffer

3 times, and the bound proteins were boiled with SDS-PAGE sample buffer at 95°C for 20 minutes before further immunoblotting.

Brain tissues were homogenized in NP-40 buffer and lysed in the same manner as described above. Protein concentrations were measured using a DC protein assay (Bio-rad). We then subjected 20 to 40 µg of protein to a standard immunoblotting protocol. The following antibodies were used according to their manufacturer’s directions: anti-PLP (ab105784), anti-

20 PPARa (ab24509), and anti-HA (ab18181) were from Abcam; anti-PPARg (#2435), anti-RXRa

(#3085), and anti-Lamin B2 (#13823) were from Cell Signaling Technology; anti-MOG (sc-

73330) and anti-histone H2A (sc-10807) were from Santa Cruz Biotechnology; anti-PPARb

(PA1-823A) was from Thermo Fisher Scientific; anti-RXRb (NBP1-30897) was from Novus

Biologicals; RXRg (CPA2042) was from Cohesion Biosciences; anti-MBP (SMI-94R) was from

Covance; anti–Qki-5 (IHC-00574) was from Bethyl Laboratories; and anti–b-actin (A5441) was from Sigma-Aldrich.

RNase A treatment. The differentiated oligodendrocytes were washed once with PBS, washed again with ASE buffer (20 mM Tris-HCl [pH 7.5], 0.5 mM EGTA, and 5 mM MgCl2), and permeabilized with ASE buffer containing 0.1% Triton X-100 at room temperature for 5 minutes. The cells were treated with 100 µg/mL RNase A (DNase- and protease-free; Thermo

Fisher Scientific) at room temperature for 20 minutes to digest RNA in the living cells (63).

Cells were washed with PBS and subjected to crosslinking, sonication, and co-IP experiments, as described above. Alternatively, the differentiated oligodendrocytes were lysed in NP-40 buffer at

4°C for 30 minutes, followed by digestion with 100 µg/mL RNase A at room temperature for 20 minutes. The cell lysates were then used for co-IP.

ChIP-seq and ChIP-qPCR. ChIP assays were carried out as described previously (64). The sheared chromatin from differentiated oligodendrocytes with ectopic expression of HA-PPARb

(with lentiviral plasmid pInducer20) or freshly isolated mouse oligodendrocytes was incubated with anti–Qki-5 (immunizing rabbit with a short synthetic peptide

[CGAVATKVRRHDMRVHPYQRIVTADRAATGN]), anti-PPARb (PA1-823A; Thermo Fisher

21 Scientific), anti-HA (#3724S; Cell Signaling Technology), or anti-Pol II (ab817; Abcam) antibodies in High Salt Buffer (50 mM HEPES-NaOH [pH 7.5], 300 mM NaCl, 1 mM EDTA,

1% Triton, 0.1% SDS, and 0.1% sodium deoxycholate with freshly added protease inhibitor) overnight at 4°C. The protein-DNA complexes were immobilized on prewashed protein A/G agarose beads (Smart-lifesciences). After several washes, the bound fractions were eluted and reverse crosslinked in the elution buffer (50 mM Tris-HCl [pH 8.0], 10 mM EDTA, and 1%

SDS) and digested with RNase A and proteinase K at 65°C for 4 hours, DNA samples were then purified using a PCR purification kit (Qiagen). Libraries were generated by VAHTS Universal

DNA Library Prep Kit (Vazyme) and sequenced by Illumina HiSeq X Ten or NovaSeq 6000

(Jiangxi Haplox Clinical Lab Cen, Ltd). The FASTQ data were trimmed by trim_galore

(v0.4.4_dev) and mapped to the mouse genome (mm10 version) using bowtie (v1.2.2) (65), then peaks were identified by homer (v4.10.1) (66) with the following steps: makeTagDirectory and findPeaks Sample_tag -style factor -size auto -minDist default -i Input_tag -fdr 0.001. The top

1,000 peaks of PPARb and HA from differentiated oligodendrocytes and the top 300 peaks of

PPARb from freshly isolated mouse oligodendrocytes were used for motif search. Motif was called by homer with findMotifsGenome.pl -size 300 -len 14,15,16 -mis 3. Signal plots and

Heatmaps were generated with ngsplot (67). ChIP-qPCR was performed using iTaq Universal

SYBR Green Supermix, as described above. A complete list of the sequences of the primers used in this study can be found in Supplemental Table 4.

References

59. Dai M, et al. Evolving gene/transcript definitions significantly alter the interpretation of

GeneChip data. Nucleic Acids Res. 2005;33(20):e175.

22 60. Sandberg R, Larsson O. Improved precision and accuracy for microarrays using updated

probe set definitions. BMC bioinformatics. 2007;8(1):48.

61. Bengtsson H, Ray A, Spellman P, Speed TP. A single-sample method for normalizing and

combining full-resolution copy numbers from multiple platforms, labs and analysis

methods. Bioinformatics. 2009;25(7):861-867.

62. Liu Y, Sun S, Bredy T, Wood M, Spitale RC, Baldi P. MotifMap-RNA: a genome-wide

map of RBP binding sites. Bioinformatics. 2017;33(13):2029-2031.

63. Chamousset D, et al. RRP1B targets PP1 to mammalian cell nucleoli and is associated with

Pre-60S ribosomal subunits. Mol Biol Cell. 2010;21(23):4212-4226.

64. Lan F, et al. A histone H3 lysine 27 demethylase regulates animal posterior development.

Nature. 2007;449(7163):689-694.

65. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment

of short DNA sequences to the human genome. Genome Biol. 2009;10(3):R25.

66. Heinz S, et al. Simple combinations of lineage-determining transcription factors prime cis-

regulatory elements required for macrophage and B cell identities. Mol Cell.

2010;38(4):576-589.

67. Shen L, Shao N, Liu X, Nestler E. ngs.plot: Quick mining and visualization of next-

generation sequencing data by integrating genomic databases. BMC Genomics.

2014;15:284.

23