SPTLC1 variants associated with ALS produce distinct sphingolipid signatures through impaired interaction with ORMDL proteins

Research ArticleNeuroscience Open Access | 10.1172/JCI161908

Museer A. Lone,1 Mari J. Aaltonen,2,3 Aliza Zidell,4 Helio F. Pedro,4,5 Jonas A. Morales Saute,6,7 Shalett Mathew,1 Payam Mohassel,8 Carsten G. Bönnemann,8 Eric A. Shoubridge,2,3 and Thorsten Hornemann1

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Lone, M. in: JCI | PubMed | Google Scholar |

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Aaltonen, M. in: JCI | PubMed | Google Scholar |

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Zidell, A. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Pedro, H. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Morales Saute, J. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Mathew, S. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Mohassel, P. in: JCI | PubMed | Google Scholar |

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Bönnemann, C. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Shoubridge, E. in: JCI | PubMed | Google Scholar

1Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Zurich, Switzerland.

2Montreal Neurological Institute and

3Department of Human Genetics, McGill University, Montreal, Canada.

4Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack, New Jersey, USA.

5Center for Genetic and Genomic Medicine, Hackensack University Medical Center, Hackensack Meridian School of Medicine, Hackensack, New Jersey, USA.

6Medical Genetics Division and Neurology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil.

7Graduate Program in Medicine, Medical Sciences, and Internal Medicine Department, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil.

8Neuromuscular and Neurogenetic Disorders of Childhood Section, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, Maryland, USA.

Address correspondence to: Thorsten Hornemann, Institute of Clinical Chemistry, University Hospital Zurich, University of Zurich, Wagistrasse 14, Schlieren-8952, Switzerland. Email: Thorsten.hornemann@usz.ch; Phone: 0041.43.253.3101. Or to: Eric A. Shoubridge, Montreal Neurological Institute, McGill University, 3801 Rue University, Rm 676, Montreal, Quebec, Canada H3A 2B4. Email: eric.shoubridge@mcgill.ca; Phone: 1.514.398.1997.

Authorship note: MAL and MJA contributed equally to this work.

Find articles by Hornemann, T. in: JCI | PubMed | Google Scholar |

Authorship note: MAL and MJA contributed equally to this work.

Published July 28, 2022 - More info

Published in Volume 132, Issue 18 on September 15, 2022
J Clin Invest. 2022;132(18):e161908. https://doi.org/10.1172/JCI161908.
© 2022 Lone et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published July 28, 2022 - Version history
Received: May 23, 2022; Accepted: July 26, 2022 View PDF Abstract

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease that affects motor neurons. Mutations in the SPTLC1 subunit of serine palmitoyltransferase (SPT), which catalyzes the first step in the de novo synthesis of sphingolipids (SLs), cause childhood-onset ALS. SPTLC1-ALS variants map to a transmembrane domain that interacts with ORMDL proteins, negative regulators of SPT activity. We show that ORMDL binding to the holoenzyme complex is impaired in cells expressing pathogenic SPTLC1-ALS alleles, resulting in increased SL synthesis and a distinct lipid signature. C-terminal SPTLC1 variants cause peripheral hereditary sensory and autonomic neuropathy type 1 (HSAN1) due to the synthesis of 1-deoxysphingolipids (1-deoxySLs) that form when SPT metabolizes L-alanine instead of L-serine. Limiting L-serine availability in SPTLC1-ALS–expressing cells increased 1-deoxySL and shifted the SL profile from an ALS to an HSAN1-like signature. This effect was corroborated in an SPTLC1-ALS pedigree in which the index patient uniquely presented with an HSAN1 phenotype, increased 1-deoxySL levels, and an L-serine deficiency. These data demonstrate how pathogenic variants in different domains of SPTLC1 give rise to distinct clinical presentations that are nonetheless modifiable by substrate availability.

Introduction

Amyotrophic lateral sclerosis (ALS) is a progressive, neurodegenerative disease of the lower and upper motor neurons characterized by severe muscle wasting, eventually leading to paralysis and death (1, 2). Approximately 90% of the ALS cases are sporadic, frequently without a recognized heritable factor, while an increasing number of genetic causes account for about 10% of ALS cases with a familial background, and a number of those without a clear family history (3).

Recently, dominant de novo missense and deletion mutations in SPTLC1 were associated with childhood-onset ALS (46). SPTLC1 and SPTLC2 are essential subunits of the enzyme serine palmitoyltransferase (SPT), which catalyzes the first and rate-limiting step in the de novo synthesis of sphingolipids (SLs) (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/JCI161908DS1). SPT typically conjugates palmitoyl-CoA with L-serine in a pyridoxal 5-phosphate–dependent reaction but it can also metabolize L-alanine and glycine under certain conditions, forming an atypical class of toxic 1-deoxysphingolipids (1-deoxySLs) that cannot be metabolized to complex SLs or degraded by canonical SL catabolism (7). Variants in the cytoplasmic domains of SPTLC1 and SPTLC2 result in pathologically increased 1-deoxySL levels, causing hereditary sensory and autonomic neuropathy type 1 (HSAN1) (8, 9), an autosomal dominant axonopathy characterized by a progressive sensory loss with variable autonomic involvement (1013). Although the majority of reported SPTLC1-HSAN1 variants are associated with sensory symptoms, a significant motor involvement was reported for the 2 variants, SPTLC1-S331F and SPTLC1-S331Y (14). The SPT enzyme complex resides in the endoplasmic reticulum (ER) membrane and at the ER-mitochondria contact sites (15). SPT is typically composed of 2 SPTLC1-SPTLC2 dimers that interact with the accessory subunits ssSPTa/b and the regulatory ORMDL proteins ORMDL1, -2, and -3, which are paralogous and functionally redundant proteins (16, 17). ORMDLs interact with the N-terminal transmembrane domain (TMD) of SPTLC1 and act as lipid sensors negatively regulating SPT activity (1820). All reported SPTLC1-ALS missense variants reside within this TMD (Figure 1A and Supplemental Figure 1B) and one causes an in-frame splice skip of exon 2 that results in a form of SPTLC1 protein completely lacking this TMD (4). Our previous results showed that SPTLC1-ALS variants caused an unregulated synthesis of SLs that did not respond to increasing concentrations of ORMDL3 in an in vitro assay, suggesting that the pathogenic variants could prevent the association of ORMDL proteins with the holoenzyme complex.

Localization and membrane association of SPTLC1 and pathogenic variants.Figure 1

Localization and membrane association of SPTLC1 and pathogenic variants. (A) Schematic of SPTLC1 displaying individual protein domains and positions of ALS and HSAN1 pathogenic variants. TMD, transmembrane domain. (B) Confocal images of SPTLC1 localization. WT-SPTLC1FLAG and variants were transiently expressed in COS-7 cells and visualized using an anti-FLAG antibody. ER-mCherry serves as an ER marker. Scale bars: 10 μm. (C and D) Expression of SPTLC1 variants in SPTLC1-KO cells. WT-SPTLC1FLAG and variants were integrated into Flp-In T-REx 293 SPTLC1-KO cells and expressed by addition of tetracycline. Whole-cell lysates were analyzed by SDS-PAGE and immunoblotting with anti-SPTLC1 and anti-SPTLC2 antibodies (C), and SPTLC2 levels were quantified (D). SPTLC2 signals were normalized to the β-actin signal. Mean ± SD, n = 4 independent replicates, unpaired 2-sided Welch’s t test. **P < 0.01, ***P < 0.001. (E) Analysis of membrane association of SPTLC1 variants. Cell lysates from Flp-In T-Rex 293 control cells and SPTLC1-KO cells expressing WT-SPTLC1FLAG and variants were centrifuged to separate the membrane pellet and cytosolic supernatant. Equal amounts of total (T), pellet (P), and supernatant (S) were analyzed by SDS-PAGE and immunoblotted for VAPB as a membrane protein control and UBB as a cytosolic protein control. See complete unedited blots for C and E in the supplemental material.

Here we have expanded the analysis of SPTLC1 variants, tested whether the variants impair the association of ORMDLs with the SPT complex, defined distinct lipid signatures associated with the expression of SPTLC1-ALS and -HSAN1 alleles, and investigated the influence of altered substrate availability on the lipid signatures and clinical phenotypes caused by mutations in different domains of SPTLC1.

Results

Loss of exon 2 in SPTLC1 impairs integration into the ER membrane. SPTLC1 is an ER-localized protein with an N-terminal TMD (Figure 1A and Supplemental Figure 1B). The amino acids in the TMD could be important for the interaction with ORMDL proteins, but as they could also be essential for ER targeting and membrane integration of SPTLC1, we first set out to determine the localization and membrane association of a subset of pathogenic SPTLC1 variants in detail, namely the ALS variants Y23F, L39del, F40S41del, a variant missing the whole of exon 2 (ex2del) induced by aberrant splicing in an A20S patient (4), and the HSAN1 variants C133W and S331F (Figure 1A). C-terminally FLAG-tagged SPTLC1 was colocalized with the ER marker Sec61b-mCherry by confocal microscopy when expressed transiently in COS-7 cells, and ER localization was observed for all variants (C133W, S331F, Y23F, L39del, and F40S41del) except for the ex2del variant, which lacks the entire TMD and showed a mostly cytosolic distribution when expressed transiently (Figure 1B).

For the biochemical analysis of SPTLC1 variants, C-terminally FLAG-tagged WT SPTLC1 and the C133W, S331F, Y23F, L39del, F40S41del, and ex2del variants were integrated into Flp-In T-REx 293 SPTLC1-KO cells (Figure 1C). The stability of SPTLC2 depends on SPTLC1 (15, 21) and the diminished level of SPTLC2 in SPTLC1-KO cells (17% of control) was rescued upon reexpression of WT, C133W, S331F, and Y23F variants, while a partial rescue was observed with the L39del (77% of control), F40S41del (68% of control), and ex2del (38% of control) variants (Figure 1, C and D).

Considering that the ex2del variant partially rescues SPTLC2 protein levels, we hypothesized that a fraction of this variant might still be associated with SPTLC2 at the ER and mitochondrial membranes that could have been masked in the microscopy analysis due to a high expression of transiently transfected constructs. To further analyze the membrane association of SPTLC1 and variants, membranes and cytosol were separated by ultracentrifugation from Flp-In T-Rex 293 control cells and variant-expressing SPTLC1-KO cells. Endogenous SPTLC1 and WT-SPTLC1FLAG were found in the membrane pellet, as was the ER membrane protein VAPB, while the cytosolic protein UBB was in the supernatant (Figure 1E). All variants were predominantly detected in the membrane pellet fraction, except for the ex2del variant which was enriched in the cytosolic supernatant (Figure 1E). However, part of the ex2del variant was present in the membrane pellet, suggesting that a portion of it was still associated with membranes, likely through interaction of the SPTLC1 aminotransferase domain with SPTLC2.

In summary, neither the HSAN1- or the ALS-causing mutations in SPTLC1 impair the ER localization and membrane association of the protein, except for the TMD-lacking ex2del variant which is predominantly soluble in the cytosol and only partially associated with membranes.

ORMDLs fail to interact with SPTLC1 variants. The patient mutations in SPTLC1 could impair the interaction of SPTLC1 with ORMDLs, as structural studies have shown that amino acids in the TMD interact with one of the ORMDL-TMDs, and some amino acids close to the catalytic site, such as SPTLC1-S331, interact with the N-terminal loop of ORMDLs that can reach into the active site to occupy the substrate-binding tunnel (16). To analyze the interaction of pathogenic SPTLC1 variants with SPTLC2 and ORMDLs, FLAG-tagged SPTLC1 and variants were purified by FLAG immunoprecipitation from digitonin-solubilized membrane fractions, and input and eluate fractions were analyzed by immunoblotting. Similar amounts of SPTLC2 copurified with all variants except the ex2del variant, which showed a reduced interaction (Figure 2A). However, the level of ex2del variant was lower in the membrane fraction input, as a majority of the protein is cytosolic, and these cells also have less SPTLC2 (Figure 1, B–D). Immunoblotting with a pan ORMDL antibody that detects all isoforms (22) showed that the interaction with ORMDLs was completely abolished by the ex2del variant, which lacks the whole TMD (Figure 2A). The other variants affecting the SPTLC1 TMD, L39del and F40S41del, showed a diminished interaction with ORMDLs, while Y23F retained interaction with ORMDLs (Figure 2A). The variants with mutations in the active site, C133W and S331F, showed a reduced interaction with ORMDLs (Figure 2A).

Interaction of SPTLC1 variants with ORMDLs.Figure 2

Interaction of SPTLC1 variants with ORMDLs. (A) Immunoblot analysis of proteins copurified with SPTLC1 variants. Membrane fractions from Flp-In T-REx 293 control cells and SPTLC1-KO cells expressing WT-SPTLC1FLAG and variants were solubilized by digitonin and subjected to FLAG immunoprecipitation. Input (5%) and eluate (IP: anti-FLAG, 40%) fractions were analyzed by SDS-PAGE and immunoblotting. HSPA5 was used as a negative control. (B and C) Analysis of the SPT complex by blue native PAGE. Membrane fractions from Flp-In T-REx 293 control (CTRL), SPTLC1-KO, and SPTLC2-KO cells (B), or SPTLC1-KO cells expressing WT-SPTLC1FLAG and variants (C) were analyzed by blue native PAGE and immunoblotted with anti-SPTLC1, anti-SPTLC2, and anti-ORMDL antibodies. See complete unedited blots for AC in the supplemental material. (D) SPT activity in WT HEK293 cells after the siRNA-mediated silencing of ORMDL expression, in the presence or absence of C6-ceramide (C6-Cer). Cells were transfected with either nontargeting scrambled control or isoform-independent ORMDL siRNA. Isotope labeling using D3-15N-L-serine was done 72 hours after transfection. De novo–formed sphingolipids (SLs) were quantified by the incorporation of isotope-labeled D3-15N-L-serine. (E) SPT activity in SPTLC1-deficient HEK293 cells expressing WT SPTLC1 and the SPTLC1-ALS variants L39del and ex2del. Cells were transfected with either scrambled or ORMDL siRNAs. Seventy-two hours after transfection, cells were labeled with D3-15N-L-serine (for 16 hours) and total de novo–formed SLs quantified by LC-MS. (F) SPT activity in WT- or ALS-variant-expressing SPTLC1-KO cells in response to the addition of C6-Cer. Cells were grown with increasing concentrations of C6-Cer. Each data point reflects a single measurement. (G) De novo SL formation in patient-derived primary fibroblasts carrying either the SPTLC1p.L39del or the F40S41del mutation. Control (WT) and mutant fibroblasts were transfected with scrambled or mutant allele–specific siRNA. SL de novo formation was measured in the presence of C6-Cer. The plot shows the total amount of de novo–formed SLs relative to untreated controls. (H and I) Formation of canonical and 1-deoxySL in SPTLC1-KO cells expressing the HSAN1 variants SPTLC1, C133W, S331F, and S331Y. The plot shows total de novo–formed SLs (H) and 1-deoxySLs (I) at the given C6-Cer concentration. SPT activity was measured by the incorporation of D3-15N-L-serine and D4-L-alanine. Mean ± SD, n = 3 independent replicates, 1-way ANOVA with Bonferroni’s adjustment for multiple comparisons. **P < 0.01; ***P < 0.001; ****P < 0.0001.

Next, we asked whether the interaction of ORMDLs with the SPT complex could be detected on a native gel where protein-protein interactions within a protein complex are preserved. The separation of digitonin-solubilized membrane fractions on a native gel, followed by immunoblotting with anti-SPTLC1, -SPTLC2, and -ORMDL antibodies, showed that ORMDLs migrated together with SPTLC1 and SPTLC2 in a complex of roughly 350–400 kDa (Figure 2B). This complex was completely absent in SPTLC1- or SPTLC2-KO cells (Figure 2B), while ORMDL proteins could still be detected on a denaturing gel, although at lower levels compared with control (Supplemental Figure 2A). Notably, the levels of ORMLDs were reduced in SPTLC1-KO (40% of control) and SPTLC2-KO (49% of control) cells (Supplemental Figure 2, B and C).

To investigate the formation of the SPT complex upon expression of the pathogenic variants, isolated membrane fractions were analyzed on a native gel. Expression of WT-SPTLC1FLAG in SPTLC1-KO cells rescued the formation of the SPT complex, as SPTLC1, SPTLC2, and ORMDLs could be detected in a complex, although at a slightly higher molecular weight than in control due to the epitope tag (Figure 2C). In SPTLC1-KO cells expressing the S331F, L39del, F40S41del, and ex2del variants, the full-size SPT complex was disassembled, as ORMDLs were absent from the complex (Figure 2C), while ORMDLs could still be detected in these samples on a denaturing gel (Supplemental Figure 2D). Instead, the S331F, L39del, and F40S41del SPTLC1 variants formed slightly lower molecular weight complexes with SPTLC2 that were devoid of ORMDLs, and the ex2del variant only allowed the assembly of an approximately 140 kDa SPTLC1-SPTLC2 complex, suggesting that loss of exon 2 also interferes with oligomerization of SPTLC1-SPTLC2 heterodimers (Figure 2C). SPTLC1, SPTLC2, and ORMDLs could still be detected in the full SPT complex with the Y23F and C133W variants, although C133W showed lower levels of this complex (Figure 2C). We observed a disassembly of the SPT complex also in S331Y and L39del patient fibroblasts, with reduced ORMDLs in the SPT complex and a shift in SPTLC2, although the phenotype was not as dramatic as upon expression of variants in a KO background, likely due to patient cells having 1 WT allele (Supplemental Figure 2, E and F).

In conclusion, coimmunoprecipitation and native PAGE analyses show that pathogenic SPTLC1 variants S331F, L39del, F40S41del, and ex2del lose interaction with ORMDLs, leading to a shift in the size of the SPTLC1-SPTLC2 complex. The C133W variant also showed a reduced interaction with ORMDLs, while Y23F supported the interaction and holoenzyme complex assembly.

SPTLC1 variants show impaired regulation by ORMDLs. We next investigated the enzymatic properties of pathogenic SPTLC1 variants by analyzing de novo SL synthesis in HEK293 SPTLC1-KO cells expressing the pathogenic SPTLC1 variants. Cellular SPT activity was analyzed with a metabolic labeling assay in which cells were grown in serine- and alanine-deficient medium supplemented with stable isotope–labeled D3-15N-L-serine (to label canonical SL) and D4-L-alanine (to label 1-deoxySL). For quantification, total SLs were extracted and quantified by high-resolution mass spectrometry. The interaction of ORMDLs with SPT inhibits SPT activity, as knockdown (Figure 2D) (23) or knockout (22) of all ORMDLs leads to increased SL synthesis. In contrast, while the expression of SPTLC1 variants L39del and ex2del in SPTLC1-KO cells led to an overall increase in SL synthesis compared with WT-SPTLC1–expressing cells, silencing ORMDLs had little or no effect on the activity of the L39del and ex2del variant (Figure 2E), suggesting that increased SL synthesis by the variants is primarily caused by dysfunctional feedback inhibition rather than by an increased activity of the holoenzyme.

ORMDLs play a role in homeostatic feedback regulation of SPT activity by ceramides (24), as addition of cell-permeant C6-ceramide (C6-Cer) inhibits SL synthesis, while the inhibitory effect is blunted in the absence of ORMDLs (Figure 2D) (23). In WT-SPTLC1–expressing cells, de novo SL synthesis was gradually reduced and ultimately suppressed with increasing concentrations of C6-Cer (Figure 2F). In contrast, SPTLC1 variants affecting the TMD showed a reduced response to inhibition of SL synthesis by C6-Cer, with the strongest effect seen for the L38R, F40S41del, and ex2del variants. An attenuated C6-Cer response was also observed in fibroblasts derived from L39del and F40S41del patients, which could be restored after knockdown of the mutant mRNA transcripts (Figure 2G) using siRNAs that specifically silence the expression of the respective mutant alleles, as shown previously (4). Additionally, the SL synthesis by variants S331F and S331Y was resistant to C6-Cer treatment, whereas the response in the C133W-expressing cells was similar to WT (Figure 2H). C6-Cer–induced inhibition was also seen for synthesis of 1-deoxySL in C133W- but not in S3331Y- and S331F-expressing cells (Figure 2I).

In summary, the lack of interaction of SPTLC1-ALS variants with the ORMDLs results in dysfunctional feedback regulation and increased SPT activity, leading to higher SL levels in cells.

Lipid signatures of motor and sensory neuropathy. To define and compare the distinct alterations in SL classes and species induced in SPTLC1-associated ALS and HSAN1 disease conditions, we analyzed SL profiles in variant-expressing cells and patient-derived plasma and fibroblasts.

First, we expressed the variants in HEK293 SPTLC1-KO cells and quantified the labeled de novo–synthesized SL species, including ceramides, sphingomyelins (SMs), and 1-deoxyceramides. Compared with WT-SPTLC1–expressing cells, SPTLC1-ALS variants affecting the TMD (A20S, Y23F, L38R, L39del, F40S41del, ex2del) showed an increased formation of canonical SLs, including those with saturated (d18:0), mono- (d18:1), and di-unsaturated (d18:2) sphingoid bases (Figure 3, A–F). The highest levels were measured for the ex2del variant, followed by the L38R, L39del, and F40S41del variants, while the A20S and Y23F variants showed modest accumulations. The HSAN1 variant C133W did not increase synthesis of canonical SLs but in contrast induced the formation of 1-deoxySLs (m18:0, m18:1), which were not formed by the TMD variants (Figure 3, G and H). The S331F and S331Y variants gave rise to a combined biochemical phenotype with increased synthesis of both canonical SLs and 1-deoxySLs.

De novo sphingolipid (SL) synthesis in variant-expressing cells.Figure 3

De novo sphingolipid (SL) synthesis in variant-expressing cells. (AF) De novo formation of ceramides (Cer, AC), sphingomyelins (SM, DF), and 1-deoxyceramides (G and H) in HEK293 SPTLC1-KO cells expressing WT-SPTLC1 and variants. Cells were probed for de novo SPT activity by a stable isotope labeling assay that incorporates D3-15N-L-serine in SL and D4-L-alanine in 1-deoxySL, inducing a mass shift (+3 Da) in the respective de novo–formed SLs and absolute levels of each SL species were measured relative to an internal lipid standard. Data are represented as mean ± SD, n = 3 independent replicates, 1-way ANOVA with Bonferroni’s adjustment for multiple comparison. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

A comparison of the plasma SL profile between the ALS patients carrying the Y23F, L39del, and F40S41del variants and unrelated HSAN1 patients carrying the C133W variant revealed similar changes in SL profiles, with increased SL levels in ALS patient plasma and higher 1-deoxyceramide levels in the HSAN1 patients (Supplemental Figure 3, A–D). There were, however, variations in the SL classes and species, as ceramides, SMs, and hexosyl-ceramides were differently affected in different patients, and the relative increases were more prominent for ceramides with saturated (d18:0) sphingoid bases compared with unsaturated (d18:1 or d18:2) backbones (Supplemental Figure 3, A–D).

To characterize the SL pattern associated with SPTLC1-ALS and -HSAN1 variants, we compared the profile of de novo–formed SL species between HEK293 SPTLC1-KO cells expressing the pathogenic SPTLC1 variants. A heatmap cluster analysis showed significant differences in the de novo–formed SL profiles (Figure 4A). The HSAN1 variants C133W, S331F, and S331Y clustered due to the increase in 1-deoxySL species, while the ALS TMD variants clustered due to the relative enrichment in canonical SL species. The S331 variants showed a mixed pattern that partly resembled the TMD variants due to increased synthesis of canonical SLs, with the S331F variant showing more similarity with the HSAN1 C133W variant, and S331Y with the ALS TMD variants (Figure 4A). The ex2del variant showed the strongest relative accumulation of canonical SL species, followed by the L38R, F40S41del, and L39del variants.

Sphingolipid (SL) signatures in variant-expressing cells.Figure 4

Sphingolipid (SL) signatures in variant-expressing cells. (A) Heatmap cluster analysis of de novo–formed SL species from HEK293 SPTLC1-KO cells expressing WT SPTLC1, SPTLC1-ALS, or SPTLC1-HSAN1 variants. Absolute levels of each SL species were measured relative to an internal lipid standard. Shown is a plot of the log10-transformed data with Euclidean distance measure. SB, sphingoid base. (B) Volcano plot comparing de novo–formed SL species in SPTLC1-KO cells expressing the SPTLC1-HSAN1 (C133W) and SPTLC1-ALS (ex2del) variants. MetaboAnalyst Suite 5.0 was used for comparison of species profiles as heatmaps and volcano plot. n = 3 independent biological replicates, significance (P) and fold change (FC) are represented as dotted lines.

We next explored the changes in SL species in detail by volcano plot comparison of SL enrichment in cells expressing the HSAN1 variant C133W relative to those expressing the ALS variant ex2del. In addition to the prominent enrichment of de novo–formed 1-deoxyceramide species in C133W-expressing cells, in ex2del-expressing cells there was an enrichment of distinct ceramide species with uncommon N-acyl chains (Figure 4B). The most abundant ceramides in HEK293 cells carry fatty acids C16, C24:0, and C24:1, while ceramides with C18:0, C20:0, C22:0, and C22:1 acyl chains, which have very low abundance in WT-expressing cells, accumulated in cells expressing ex2del (Figure 4B) and also in cells expressing other TMD and S331 variants (Supplemental Figure 4, A and B). The intermediate chain length ceramide species form only a small fraction of the total SLs in WT-SPTLC1–expressing cells, but an increased fraction in ex2del-expressing cells relative to total species, despite the overall higher SL amount in mutant-expressing cells (Supplemental Figure 4C). The relative increase in the intermediate chain length C18, C20:0, and C22 species in mutant-expressing cells was accompanied by a decrease in the relative proportion of abundant C16:0 and C24:0 species (Supplemental Figure 4C). An increase in the intermediate acyl chain–containing ceramides (C18 to C22) was also seen in primary fibroblasts isolated from L39del and F40S41del patients (Supplemental Figure 5, A and B) in which targeted siRNA–mediated knockdown of the mutant alleles normalized the acyl chain profiles (Supplemental Figure 5C). Lastly, the plasma species profiles in patients carrying Y23F, L39del, and F40S41del showed similar changes in ceramide species (Supplemental Figure 6).

In conclusion, the lipid signature for ALS-associated variants is characterized by increased SL synthesis and accumulation of uncommon acyl chain length ceramide species, whereas in HSAN1 the signature is characterized by increased deoxySL synthesis.

Serine availability modulates the clinical presentation of SPTLC1 variants. SL analysis showed clear differences between the SPTLC1 variants when expressed in cells; however, the correlation of variants with disease presentation is not always straightforward, and the L39del and S331 variants have been reported to cause both sensory and motor phenotypes (4, 25), suggesting that additional factors influence disease outcome.

L-Serine and L-alanine availability modulates the synthesis of canonical and 1-deoxy lipids, respectively, as 1-deoxySL formation is induced under conditions of L-serine restriction (26, 27). We therefore tested the effect of reduced serine availability on SL profiles of SPTLC1 variant–expressing cells. WT-SPTLC1–, ex2del-, and L39del-expressing cells were cultured at decreasing L-serine and constant L-alanine concentrations. In the absence of L-serine, we observed a reduction in total SL synthesis (Figure 5A) and an increase in 1-deoxySL synthesis in both WT- and variant-expressing cells (Figure 5B). However, the extent of 1-deoxySL formation with reduced serine conditions was higher in variant-expressing cells than in those expressing WT-SPTLC1. Interestingly, the L39del variant showed a significantly higher 1-deoxySL accumulation than the ex2del under these conditions (Figure 5B). Heatmap cluster analysis showed a gradual shift of SL enrichment profiles from the ALS lipid signature toward the HSAN1 signature, which is distinguishable for the L39del mutant–expressing cells (Figure 5C).

Sphingolipid (SL) signatures shift upon amino acid availability.

留言 (0)

沒有登入
gif