Sphingolipid desaturase DEGS1 is essential for mitochondria-associated membrane integrity

Abstract

Sphingolipids function as membrane constituents and signaling molecules, with crucial roles in human diseases, from neurodevelopmental disorders to cancer, best exemplified in the inborn errors of sphingolipid metabolism in lysosomes. The dihydroceramide desaturase Δ4-dihydroceramide desaturase 1 (DEGS1) acts in the last step of a sector of the sphingolipid pathway, de novo ceramide biosynthesis. Defects in DEGS1 cause the recently described hypomyelinating leukodystrophy-18 (HLD18) (OMIM #618404). Here, we reveal that DEGS1 is a mitochondria-associated endoplasmic reticulum membrane-resident (MAM-resident) enzyme, refining previous reports locating DEGS1 at the endoplasmic reticulum only. Using patient fibroblasts, multiomics, and enzymatic assays, we show that DEGS1 deficiency disrupts the main core functions of the MAM: (a) mitochondrial dynamics, with a hyperfused mitochondrial network associated with decreased activation of dynamin-related protein 1; (b) cholesterol metabolism, with impaired sterol O-acyltransferase activity and decreased cholesteryl esters; (c) phospholipid metabolism, with increased phosphatidic acid and phosphatidylserine and decreased phosphatidylethanolamine; and (d) biogenesis of lipid droplets, with increased size and numbers. Moreover, we detected increased mitochondrial superoxide species production in fibroblasts and mitochondrial respiration impairment in patient muscle biopsy tissues. Our findings shed light on the pathophysiology of HLD18 and broaden our understanding of the role of sphingolipid metabolism in MAM function.

Keywords: Bioenergetics; Demyelinating disorders; Lipid rafts; Metabolism; Neuroscience.

Figure 1. Mitochondrial damage in vivo.

(A) Muscle histopathology. Sudan black and Oil Red O staining for neutral lipids, H&E staining for fiber size, and modified Gömöri trichrome staining for the endomysial connective tissue of the quadriceps muscle biopsy from Pat. 20 and a child who served as a control (CTL). (B) Immunofluorescence for complex I and VDAC1 markers in the quadriceps muscle biopsy from Pat. 20 and a child who served as a control. (C) Skeletal muscle histochemistry. COX and SDH staining of quadriceps muscle biopsies from Pat. 9 and Pat. 20 and a child who served as a control. (D) TEM of the quadriceps muscle biopsy in the longitudinal plane from Pat. 9 and a child who served as a control. Mitochondria and LDs are highlighted in orange and pink, respectively.

Figure 2. Mitochondrial aberrant morphology in DEGS1 patient fibroblasts.

Mitochondrial (A) area and (C) crests morphology assessment by TEM and their quantification (B and D), respectively. DEGS1 patient (n = 3) and control (n = 3) fibroblasts. Data are presented as box-and-whisker plots (median, interquartile interval, minimum, maximum). *P < 0.05, 2-tailed Student’s t test.

Figure 3. Mitochondrial dynamics and bioenergetics impairment in DEGS1patient fibroblasts.

(A) Analysis of the mitochondrial morphology by MitoTracker (MitoT) staining in 4-hour live-cell imaging; disconnected mitochondria are each shown in different colors. Mitochondrial (B) area and (C) sphericity quantifications. (D) Quantification of numbers of disconnected mitochondria over time. DEGS1 patient (n = 6) and control (n = 5) fibroblasts. (E) Western blot analysis of fusion and fission proteins and (F) their quantification. DEGS1 patient (n = 5) and control (n = 5) fibroblasts. (G) Quantification of ΔΨm by measuring TMRE intensity. FCCP was used as a mitochondrial oxidative phosphorylation uncoupler. DEGS1 patient (n = 6) and control (n = 5) fibroblasts. (H) Quantification of intracellular (DHE) and mitochondrial (MitoSOX) superoxide species production levels. The complex III inhibitor antimycin A (200 μM) was used for treatment for 1 hour as a positive control for ROS production. DEGS1 patient (n = 5) and control (n = 4) fibroblasts. All experiments were done in triplicate. Data are represented as box-and-whisker plots (median, interquartile interval, minimum, maximum). *P < 0.05; **P < 0.01; ***P < 0.001, Wilcoxon’s test.

Figure 4. DEGS1 localizes at the MAM.

Immunofluorescence analysis in control fibroblasts (n = 5). Colocalization between DEGS1 (in green) and ERLIN2, ACSL4, DRP1, and MFN2 (MAM-resident proteins), MitoTracker (mitochondria maker), calnexin (ER marker), and GM140 (Golgi marker) (all in red). Colocalization area (coloc) is shown in white.

Figure 5. DEGS1 is detected in MAM-isolated fractions.

(A) Western blot analysis of all recovered fractions during MAM collection from human brain white matter of control cases, adults (n = 2) and children (n = 2). Note that DEGS1 is more abundant in the MAM fraction than in the ER fraction (25 μg of protein per lane). (B) Representation of DEGS1 enrichment normalized to the total amount of protein in each fraction. Data are represented as box-and-whisker plots (median, interquartile interval, minimum, maximum). Homog, homogenate; CM, crude mitochondria; M+MAM, mitochondria attached to MAM.

Figure 6. DEGS1 impairment leads to MAM disruption.

(A) Schematic representation of PL synthesis and trafficking and cholesterol esterification at the MAM. (B) PL synthesis and trafficking assay. ³H-PS/³H-PE lipid ratio reflecting incorporation after 12 hours of ³H-serine into ³H-PS and ³H-PE in human fibroblasts from patients with DEGS1 mutations (n = 3) and control individuals (n = 3). (C) SOAT1 activity assay. ³H-CE lipid levels reflecting the incorporation of ³H-cholesterol into ³H-CE after 6 hours in human fibroblasts from patients with DEGS1 mutations (n = 4) and control individuals (n = 5). (D and E) DhCer/cer, DhSM/SM, and (F and G) DhHexCer/HexCer lipid levels in the total and MAM fractions from fibroblasts of patients with DEGS1 mutations (n = 3–5) and control individuals (n = 4–5). (H) Distance between the ER and the mitochondria (M) in fibroblasts using TEM and its (I) quantification. DEGS1 patient (n = 3) and control (n = 3) fibroblasts. Data are represented as box-and-whisker plots (median, interquartile interval, minimum, maximum). *P < 0.05; **P < 0.01, 2-tailed Student’s t test.

Figure 7. Neutral lipids and PA accumulation and their regulation.

(A) Representative microscopy images of Oil Red O, an esterified cholesterol and LD marker, counterstained with hematoxylin as a nuclear marker, and its (B) Feret’s diameter and (C) number/cell quantification. DEGS1 patient (n = 5) and control (n = 5) fibroblasts. (D) PA and glyceride levels. DEGS1 patient (n = 4) and control (n = 4) fibroblasts. mRNA levels of (E) genes encoding key enzymes involved in the LD synthesis pathway DGKA, DGAT1, and DGAT2, and (F) the lipogenic gene master regulators SREBF1a, SREBF1c, and SREBF2. DEGS1 patient (n = 4–5) and control (n = 5) fibroblasts. (G) mRNA levels of SREBF target genes involved in cholesterol synthesis: HMGCS1, HMGCR, MVD, and SQLE. DEGS1 patient (n = 4) and control (n = 4) fibroblasts. (H) CE/FC lipid ratio and (I) PS, PE, and PC lipid levels in human fibroblasts from patients with DEGS1 mutations (n = 4–5) and control individuals (n = 4–5). All experiments were done in triplicate. Data are presented as box-and-whisker plots (median, interquartile interval, minimum, maximum). *P < 0.05; **P < 0.01; ***P < 0.001, 2-tailed Student’s t test.

Figure 8. Schematic representation of the pivotal role of DEGS1 in the regulation of MAM and mitochondria function.

As a consequence of DEGS1 loss, an imbalance in SL metabolism (DhCer, Cer, DhSM, SM, DhHexCer, HexCer) occurs (Figure 6, D–G). Since SLs are essential components of membranes, mitochondrial and ER-MAM appear physically disrupted and thus functionally impaired. MAMs cannot be properly formed, since the distance between mitochondria and the ER is greater than in controls (Figure 6, H and I). Moreover, 2 hallmarks of MAM function, PL transport/synthesis and CE synthesis, are decreased (Figure 6, B and C). The MAM components DRP1 and DGAT2 (42, 59), have their function affected leading to (a) decreased mitochondrial fission (Figure 3, A–F) and (b) increased size and numbers of LDs (Figure 7, A–C) due to DGAT2’s main role in LD formation (68). Mitochondria appear larger in size and hyperfused (Figure 1D, Figure 2, and Figure 3, A–D), with OXPHOS impairment (Figure 1C, Table 1, and Supplemental Table 2), decreased membrane potential (Figure 3G), and augmented production of superoxide anion levels (Figure 3H).