Sub-cellular metabolomics investigation of mitochondrial CI deficiency in cytosol & mitochondria of Ndufs4 mice

Abstract

Mitochondrial Complex I (CI) is regarded as the foremost entry point for electrons that fuel oxidative phosphorylation. CI deficiency (CID) is the most common mitochondrial disease (MD) in humans and Leigh syndrome (LS), an MD characterized by progressive neuromuscular degeneration, is a debilitating result thereof. The complex clinical and biochemical LS manifestations lead to difficult diagnosis and a lack of effective treatment strategies, thus model systems such as whole-body Ndufs4 knockout (KO) mice have greatly contributed to our understanding of the underlying pathogenesis. Through tissue-level metabolic characterization of the neural and muscular involvement at the forefront of the Ndufs4 KO phenotype, several critical metabolic pathways were shown to contribute to the observed MD. Although organelle-specific regulation of the involved enzymes and metabolites is well supported by literature, no compartment-specific metabolic studies have been conducted in this model to date. The aim of this study was to adapt existing cell fractioning procedures for subsequent mitochondrial and cytosolic metabolomics in Ndufs4 KO whole brain and skeletal muscle to address this research gap. Hypothesis-generating metabolomics methods utilizing proton nuclear magnetic resonance spectroscopy (1H-NMR), gas chromatography time-of-flight mass spectrometry (GC-TOF-MS) and liquid chromatography tandem mass spectrometry (LC-MS/MS) were applied at various stages. We firstly present a novel methodology for mitochondrial and cytosolic purification isolation via magnetically activated cell sorting (MACS) and a novel filtration method. Complementary enzymatic, proteomic and respirometric assays proved the superior mitochondrial yield and purity of MACS over a standard centrifugal method, while mitochondrial integrity was not affected. The cytosolic fraction was accordingly shown to be sufficiently organelle free for metabolic comparison. Concentration curves of metabolite standards showed that the buffer environment does not critically detriment measurement precision, linearity or accuracy. Finally, a pilot application of this methodology to wildtype (WT) mouse brain fractions delivered 60–70 viable metabolites with sufficient accuracy and identity confidence for further in vivo metabolomics study. In application of the developed workflow to a statistically viable cohort of Ndufs4 KO and WT mice (n = 8 each), LS-related metabolic compartmentalization was shown. Untargeted 1H-NMR and targeted LC-MS/MS amino acid and acylcarnitine profiling delivered 40–45 viable metabolites with excellent measurement precision and confident identification. The significantly altered brain metabolites showed increases in key bioenergetic intermediates- with greater

effect at the mitochondrial level. These features corroborated previous studies, implicating the role of neural redox (NADH, FADH2) dysregulation and anabolic perturbations underlying mitochondrially-initiated integrated stress responses. Compartmentalized decreases in muscle bioenergetic intermediates emphasized the role of mitochondrial lactate and creatine metabolism, while providing evidence for disturbed proteostasis due to prolonged energy deficit. Additionally, a first report of sub-cellular calcium (Ca2+) and magnesium (Mg2+) measurement via 1H-NMR also revealed Ca2+ accumulation specific to the neural mitochondria, contrasted by bicompartmental Mg2+ depletion in the muscle fractions. In summary, this study provides a simple, cost-effective pipeline for sub-cellular metabolomics in mammalian disease models. In application, this methodology provides proof of mitochondrial/cytosolic compartmentalization in known pathological Ndufs4 KO metabolic pathways. These results should pave the way for better understanding of mitochondrial metabolism in MD, towards improved diagnostic and therapeutic strategies.