Loss of LUC7L2 and U1 snRNP subunits shifts energy metabolism from glycolysis to OXPHOS

Abstract

Oxidative phosphorylation (OXPHOS) and glycolysis are the two major pathways for ATP production. The reliance on each varies across tissues and cell states, and can influence susceptibility to disease. At present, the full set of molecular mechanisms governing the relative expression and balance of these two pathways is unknown. Here, we focus on genes whose loss leads to an increase in OXPHOS activity. Unexpectedly, this class of genes is enriched for components of the pre-mRNA splicing machinery, in particular for subunits of the U1 snRNP. Among them, we show that LUC7L2 represses OXPHOS and promotes glycolysis by multiple mechanisms, including (1) splicing of the glycolytic enzyme PFKM to suppress glycogen synthesis, (2) splicing of the cystine/glutamate antiporter SLC7A11 (xCT) to suppress glutamate oxidation, and (3) secondary repression of mitochondrial respiratory supercomplex formation. Our results connect LUC7L2 expression and, more generally, the U1 snRNP to cellular energy metabolism.

Keywords: 7q-; LUC7; MDS; Tarui disease; cancer; ferroptosis; myelodysplastic syndrome; phosphofructokinase; spliceosome; system X(c)(−).

Figure 1:. Identification of Pre-mRNA Splicing Components as Repressors of OXPHOS. See also Figure S1 and Table S1.

(A) Overview of the main ATP-generating pathways in human cells. OXPHOS: oxidative phosphorylation. ETC: electron transport chain. TCA: tricarboxylic acid cycle. (B) Gene-level analysis of a genome-wide CRISPR/Cas9 screen in glucose and galactose. Each dot represents an expressed, non-essential gene (n = 9,189). (C) Gene ontology analysis generated using a gene list ranked by viability in galactose against GO components. (D–F) Functional validation of the screening results. Basal whole-cell oxygen consumption rates (OCR), extracellular acidification rates (ECAR) and OCR/ECAR ratios were simultaneously measured after CRISPR/Cas9-mediated gene depletion in K562 cells grown in glucose-containing media. Data are shown as mean ± SEM (n≥3 independent experiments. *P<0.05, **P<0.01, ***P<0.001, t-test relative to control (GFP) sgRNA-treated cells. NDUFB5 is a control with known role in OXPHOS.

Figure 2:. LUC7L2 Impacts Metabolic State-Dependent Cell Growth and Bioenergetics. See also Figure S2.

(A–C) Cell proliferation of LUC7L2^KO K562 cells grown in (A) glucose and (B) treated with 2-deoxyglucose (2-DG) or when glucose was replaced by galactose or (C) glucose with OXPHOS inhibitors. (D) Respiratory parameters of LUC7L2^KO cells as determined by oxygen consumption rate (OCR). (E) Basal glycolytic activity in LUC7L2^KO cells as determined by extracellular acidification rate (ECAR). (F) Relative mtDNA abundance and (G) citrate synthase activity of LUC7L2^KO cells. All data are shown as mean ±SEM (n≥3). *P< 0.05, **P<0.01, ***P<0.001, t-test relative to LUC7L2^WT cells.

Figure 3:. Metabolite Analysis in LUC7L2-Depleted Cells Reveals Crossovers at Phosphofructokinase and System X_c⁻. See also Figure S3 and Table S2–3.

(A) Intracellular levels of metabolites in LUC7L2^KO K562 cells as determined by LC-MS. (B) Extracellular levels of metabolites as determined by LC-MS analysis of the spent media from (A). Positive and negative values illustrate metabolite secretion and consumption by the cells, respectively. All data are shown as mean ±SEM (n = 5–8). *P<0.05, **P<0.01, ***P<0.001, t-test relative to LUC7L2^WT. (C) Media acidification of LUC7L2^KO K562 cells grown in glucose.

Figure 4:. LUC7L2 Encodes a U1 snRNP Subunit Involved in Pre-mRNA Splicing. See also Figure S4 and Table S4–7.

(A) Confocal microscopy of a single nucleus from a HeLa cell expressing LUC7L2-GFP and immunolabeled with antibodies to SRSF2. (B) LUC7L2-interacting proteins as determined by IP-MS (n = 2). (C) Representation of LUC7, SNRPA and SNRNP70 on the yeast U1 snRNP (PDB 5UZ5) (Li et al., 2017). (D) Proportion of eCLIP peaks mapping to splicing snRNAs in HeLa and K562 cells (eCLIP n = 2, each). (E) Representation of the genes bound by LUC7L2 at P<10⁻⁴. (F) Proportion of LUC7L2 eCLIP peaks in pre-mRNAs at P<10⁻⁴. (G) Meta-analysis of LUC7L2 binding sites across shared eCLIP peaks at P<10⁻⁴. (H) Differential gene expression in LUC7L2^KO cells (n = 3 for each cell type and each genotype) as determined by RNA deep-sequencing at FDR<10⁻⁴ and >|1.5| fold change. (I) Alternative splicing events seen in LUC7L2^KO cells as determined by rMATS at FDR<0.1 and |Δψ|>0.05 (n = 3 for each cell type and each genotype). (J) Types of alternative splicing in LUC7L2^KO cells with SE: skipped exon; MXE: mutually-exclusive exons; A5SS: alternative 5′ splice site; A3SS: alternative 3′ splice site; RI: retained intron. (K) Alternative events presenting an eCLIP peak at a 250-nucleotide distance from splicing events at P<10⁻² (in darker shade).

Figure 5:. Role of LUC7L2-Mediated PFKM and SLC7A11 Alternative Splicing in Energy Metabolism. See also Figure S5.

(A) Representation of PFKM exons 10–13, LUC7L2 binding sites as determined by eCLIP, antisense oligonucleotides (ASO) targeting sites and the expected transcripts. Ψ: percent spliced in reported by rMATS in K562 cells. E: exon. A negative Δψ value indicates exon skipping. PTC: premature termination codons. (B) RT-PCR (top) and immunoblot (bottom) of LUC7L2^KO K562 cells (left) or HAP1 cells treated for 48h with ASO targeting the 5′SS of PFKM exon 12 (right). (C) Relative ECAR (n = 3–5) and (D) glycogen in LUC7L2^KO K562 cells expressing control cDNAs (GFP) or PFKM cDNA (n = 2–4). (E) Representation of SLC7A11 exons 6–10 as in (A). (F) RT-PCR of LUC7L2^KO K562 cells with primers amplifying transcripts corresponding to SLC7A11 exons 6–12. SE: skipped exon. (G) Immunoblot on LUC7L2^KO K562 cells with antibodies to SLC7A11 and ACTIN. (H) Cell viability of LUC7L2^Rescue (corresponds to LUC72L2^KO expressing LUC7L2 cDNA) and LUC7L2^KO HAP1 cells grown for 24h in galactose relative to glucose (n = 3). SAS: 500μM sulfasalazine. (I) RT-PCR (top) and immunoblot (bottom) of HAP1 cells treated for 48h with ASOs targeting the 5′SS of exon 7 and/or exon 9 of SLC7A11. (J) Media glutamate (n = 4), (K) representative seahorse trace (shown as mean ±SD), and (L) viability in galactose of HAP1 cells treated for 48h with the indicated ASOs (n = 3). All data are shown as mean ±SEM (unless otherwise stated) with * P<0.05, ** P<0.01, ***P<0.001, t-test relative to control.

Figure 6:. Proteomic Analysis of LUC7L2^KO and Galactose-Grown Cells Reveals Secondary Complexes I+III+IV Accumulation. See also Figure S6 and Table S8.

(A) Volcano plots and immunoblots of OXPHOS protein expression in LUC7L2^KO and (B) in galactose-grown K562 cells. > indicates a protein not shown but reported in Table S8. (C) Blue-Native PAGE on a mitochondria-rich fraction isolated from LUC7L2^KO K562 cells and stained with coomassie or (D) immunoblotted with the indicated antibodies. Parallel blots in which the same lysate was loaded were used to avoid antibodies cross-reactivity. SCs: supercomplexes. CI-V: complexes I to V. (E) Immunoblot on K562 cells expressing Cas9 and treated with sgRNAs targeting glycolytic enzymes with the indicated antibodies. (F) Model of the secondary regulation of the respiratory chain by LUC7L2 and galactose.

Figure 7:. Pre-mRNA Splicing and Partial Redundancy Within the LUC7 Family. See also Figure S7.

(A) Phylogenetic tree of the LUC7 protein family. LUC7 is from S. cerevisiae. ZnF: Zinc-finger domain. S/R-rich: Serine and arginine-rich domain. (B) Quantitative PCR detecting LUC7 family transcripts in LUC7L2^KO K562 cells (n = 3). (C) Representation of LUC7L exons 1–3, LUC7L2 binding sites as determined by eCLIP, and the expected transcripts. Ψ: percent spliced in reported by rMATS in K562 cells. I: intron. E: exon. (D) RT-PCR amplifying LUC7L exon 1 to exon 2. Arrowheads: retained entities in LUC7L. (E) Immunoblot of LUC7 proteins in cell lines expressing Cas9 and sgRNAs targeting the indicated genes using the indicated antibodies and (F) number of cells after 4 days of growth in glucose-containing media. (G) Oxygen consumption analysis of LUC7L2^KO K562 cells expressing cDNAs of LUC7 family members. All data are shown as mean ±SEM with * P<0.05, ** P<0.01, ***P<0.001, t-test relative to control.