SLC25A48 controls mitochondrial choline import and metabolism

Abstract

Choline is an essential nutrient for the biosynthesis of phospholipids, neurotransmitters, and one-carbon metabolism with a critical step being its import into mitochondria. However, the underlying mechanisms and biological significance remain poorly understood. Here, we report that SLC25A48, a previously uncharacterized mitochondrial inner-membrane carrier protein, controls mitochondrial choline transport and the synthesis of choline-derived methyl donors. We found that SLC25A48 was required for brown fat thermogenesis, mitochondrial respiration, and mitochondrial membrane integrity. Choline uptake into the mitochondrial matrix via SLC25A48 facilitated the synthesis of betaine and purine nucleotides, whereas loss of SLC25A48 resulted in increased production of mitochondrial reactive oxygen species and imbalanced mitochondrial lipids. Notably, human cells carrying a single nucleotide polymorphism on the SLC25A48 gene and cancer cells lacking SLC25A48 exhibited decreased mitochondrial choline import, increased oxidative stress, and impaired cell proliferation. Together, this study demonstrates that SLC25A48 regulates mitochondrial choline catabolism, bioenergetics, and cell survival.

Keywords: bioenergetics; brown adipose tissue; cancer metabolism; choline; mitochondria; purine nucleotides.

Figure 1.. SLC25A48 is required for BAT thermogenesis and mitochondrial membrane integrity in vivo

(A) SLC25A family protein abundance from BAT mitochondria of wild-type male mice fed a standard diet or high-fat diet for 8 weeks. n = 5 per group. (B) Representative immunofluorescent image of brown adipocytes expressing SLC25A48-FLAG. A plasma membrane marker (WGA), mitochondrial marker (TOM20), and nuclear marker (DAPI) were shown. Scale bar, 20 μm. (C) Relative SLC25A48 mRNA levels in indicated tissues from male control and SLC25A48-KO mice. n = 3 per group. (D) Rectal temperature of mice during cold tolerance test. n = 3 per group. (E) BAT mitochondrial structure of male control mice and SLC25A48-KO mice at 12 weeks old fed standard diet. Right: quantification of mitochondrial size and cristae density (n = 3 per group). Representative image. Scale bar, 0.5 μm. (F) BAT mitochondrial respiration. n = 3 per group. (G) H₂O₂ production in isolated BAT mitochondria from male control and SLC25A48-KO mice. n = 4 control, n = 5 SLC25A48-KO. (H) Lipidomic analysis of BAT mitochondria from male wild-type and SLC25A48-KO mice. n = 8 per group. (I) Serum levels of malondialdehyde (MDA) in male control and SLC25A48-KO mic on a standard diet. n = 9 per group. (J) Serum metabolomics from male control and SLC25A48-KO mice on a standard diet. n = 8 per group. Statistics: unpaired t test (A and G–J); two-way ANOVA with Holm-Šídák’s multiple comparisons test (C); two-way ANOVA with Šídák’s multiple comparisons test (D–F).

Figure 2.. The cell-autonomous role of SLC25A48 in the regulation of choline metabolism

(A) Cellular respiration of SLC25A48-KO and rescue (SLC25A48-KO^{SLC25A48-FLAG}) brown adipocytes. n = 10 per group. (B) H₂O₂ production in isolated mitochondria of brown adipocytes. n = 3 per group. (C) Top 50 genes co-evolved with human SLC25A48 (left) and mouse Slc25a48 (right) across species ranked by Z score. (D) Schematic of mitochondrial contribution to one-carbon metabolites. Choline is imported into the mitochondrial matrix to be metabolized by CHDH and ALDH7A1 to form betaine. Methyl-groups from betaine are transferred to the methionine cycle or the folate cycle through production to formate. Formate is used at two positions (C2 and C8) in purine nucleotide synthesis. (E) Whole-cell metabolomics of SLC25A48-KO and SLC25A48-rescue brown adipocytes. n = 8 per group. (F) Pathway analysis of metabolites down- and upregulated in (E). (G) Mitochondrial metabolomics from brown adipocytes. n = 8 per group. Data represented as Z score. (H) ¹³C₂-choline tracing to detect purine metabolites. SLC25A48-KO and rescue cells were incubated with ¹³C₂-choline (1 mM) for 24 h. Values relative to SLC25A48-KO. n = 4 per group. Statistics: unpaired t test (A, B, E, G, and H).

Figure 3.. SLC25A48 is required for mitochondrial choline import and metabolism

(A) Choline uptake into SLC25A48-KO and SLC25A48-rescued HEK293T cells. n = 4 per group. (B) Choline uptake in purified mitochondria by incubating with ³H-choline (5 nM) in the presence of non-labeled choline at 10 μM or 1 mM. n = 3 per group. (C) Schematic of d9-choline tracing experiment in (D) and (E). Isolated mitochondria from SLC25A48-KO and SLC25A48-rescue cells were incubated with d9-choline (M + 9) in a dose- or time-dependent manner. The mitochondria and the experimental buffer were separated through centrifugation and analyzed by LC-MS for d9-betaine (M + 9). (D) Dose-dependent betaine synthesis in mitochondria and experimental buffer. n = 3 per group per concentration. (E) Time-dependent betaine synthesis in mitochondria and experimental buffer with d9-choline at 20 μM. n = 3 per group per concentration. Statistic: unpaired t test. (F) Schematic of betaine production experiment in (G). Mitochondria were purified via differential centrifugation. The isolated mitochondria and cytoplasmic compartments were separately incubated with 20 μM d9-choline and analyzed by LC-MS to detect d9-betaine (M + 9). (G) Left: time course of betaine production from d9-choline in the isolated mitochondria and cytoplasmic compartments. Values relative to total M + 9 betaine at 5 min in SLC25A48 rescued mitochondria. Right: d9-betaine levels at 30 min of incubation. In the cytoplasmic compartment, d9-betaine only met the detection limit in 2 of 3 samples in both groups. n = 3 per group per time point. Statistics: two-way repeated-measures ANOVA with Šídák’s multiple comparisons test (A); two-way ANOVA with Tukey’s multiple comparisons test (B and G); two-way ANOVA with Šídák’s multiple comparisons test (D).

Figure 4.. Mitochondrial choline catabolism via SLC25A48 regulates cell growth

(A) Manhattan plot of GWAS for SNPs associated with circulating choline levels in 6,136 participants. (B) Association of SLC25A48 SNP rs200164783 for changes in 1,391 metabolites. (C) DNA sequencing of wild-type control and homozygous knockin (SNP-KI) of SLC25A48 SNP rs20016478 (A>G) in HEK293T cell. (D) cDNA sequencing from SLC25A48 mRNA in SNP-KI cells. Right: mRNA expression of SLC25A48 exon 5 region in wild-type control and SNP-KI cells by qPCR. n = 2 for wild-type and 3 for SNP-KI. (E) Choline uptake in the isolated mitochondria from wild-type control and SNP-KI cells. n = 3 per group. (F) H₂O₂ production in isolated mitochondria. n = 6 per group. (G) Indicated hexosylceramide levels. Values are normalized to the total hexosylceramide contents of control cells. n = 8 per group. (H) Cell growth of wild-type control and SNP-KI cells in DMEM containing 10% FBS. n = 3 per group. (I) Cell proliferation by EdU incorporation assays in (H). n = 3 per group. (J) Cell viability of indicated cancer cell lines by CRISPR-Cas9-mediated SLC25A48 deletion. n = 3 per group. (K) Cell-cycle analysis of SKOV3 ovarian cancer cells 16 h after transfection with control or SLC25A48-KO plasmid. n = 3 per group. (L) Mitochondrial H₂O₂ production in live cells in the presence or absence of betaine at 10 mM for 6 h. n = 32 for wild-type control cells for basal and betaine, 32 for SNP-KI basal, and 34 for betaine. Statistics: unpaired t test (D–G and I); two-way repeated-measures ANOVA with Šídák’s multiple comparisons test (H); two-way ANOVA with Šídák’s multiple comparisons test (J–L).