SIRT5-mediated lysine desuccinylation impacts diverse metabolic pathways

Abstract

Protein function is regulated by diverse posttranslational modifications. The mitochondrial sirtuin SIRT5 removes malonyl and succinyl moieties from target lysines. The spectrum of protein substrates subject to these modifications is unknown. We report systematic profiling of the mammalian succinylome, identifying 2,565 succinylation sites on 779 proteins. Most of these do not overlap with acetylation sites, suggesting differential regulation of succinylation and acetylation. Our analysis reveals potential impacts of lysine succinylation on enzymes involved in mitochondrial metabolism; e.g., amino acid degradation, the tricarboxylic acid cycle (TCA) cycle, and fatty acid metabolism. Lysine succinylation is also present on cytosolic and nuclear proteins; indeed, we show that a substantial fraction of SIRT5 is extramitochondrial. SIRT5 represses biochemical activity of, and cellular respiration through, two protein complexes identified in our analysis, pyruvate dehydrogenase complex and succinate dehydrogenase. Our data reveal widespread roles for lysine succinylation in regulating metabolism and potentially other cellular functions.

Figure 1. Profiling Lys Succinylation Proteome in Mouse Cells

(A) Schematic representation of experimental workflow for the identification of Lys succinylation substrates in Sirt5 KO mouse liver (left) and SILAC quantification of Lys succinylation in WT and Sirt5 KO MEFs (right) (see also Figure S1). (B) Venn diagram showing the total numbers of Lys succinylation sites identified in MEFs and in mouse liver tissue and their overlap (see also Figure S2). (C) Pie chart showing the distribution of the number of Ksucc site identifications per protein. (D) Venn diagram showing overlap between quantifiable Lys succinylation sites and Lys acetylation sites in MEFs.

Figure 2. Quantitative Analysis of Lys Succinylation Proteome

(A) Histogram showing the ratio distribution of quantifiable Lys succinylation sites between Sirt5 KO and WT MEFs (see also Figure S3). (B) Scatter plot showing the quantification of Lys succinylation sites in relation to peptide intensities. (C) Full MS and MS/MS spectra for the identification and quantification of K498 succinylation on Pkm. b and y ions indicate peptide backbone fragment ions containing the N and C terminal, respectively. ++ indicates doubly charged ions, and * indicates loss-of-amine ions. Succinylated Lys is underlined and colored in red.

Figure 3. Absolute Stoichiometries of Lys Succinylation Sites in Sirt5 KO and WT MEF Cells

(A) Comparison of stoichiometries of Ksucc sites in WT (red dots) and Sirt5 KO MEF cells (blue dots). (B) Comparison of distributions of Ksucc stoichiometries between WT (pink bars) and Sirt5 KO ells (red bars). (C) List of representative Ksucc sites with significant changes in absolute stoichiometries between WT and Sirt5 KO cells.

Figure 4. Characterization of Lys Succinylation Proteome

(A) Icelogo representation shows flanking sequence preferences for all Lys succinylation sites (see also Figure S4). (B) Venn diagram showing cellular compartment distribution of Ksucc proteins. (C) Bar graphs showing representative ontology annotations enriched with Ksucc proteome (see also Figure S5). (D) Table of all KEGG pathways involving succinyl-CoA in mice showing the enrichment of each pathway in Ksucc proteins (see also Table S3).

Figure 5. Subcellular Localization of Human and Mouse SIRT5

(A and B) Subcellular fractionation of control and Sirt5 KD HEK 293T cells (A) and WT and Sirt5 KO mouse livers (B). Fractionation controls were SDHA and SOD2 (mitochondria), HSP90 (cytosol), and histone H3 (nucleus). MT, mitochondria; N, nucleus. (C) Confocal microscopy analysis of SIRT5 localization in control and Sirt5 KD cells along with MitoTracker Red (MT).

Figure 6. Interaction Network and Protein Complex Analysis of Lys Succinylation Proteome

(A–E) Protein-protein interaction network on the basis of the STRING database (v8.3) (A) and visualized in Cytoscape and representative protein complexes enriched in Lys succinylation proteome: Frataxin complex (B), HMGB1-HMGB2-HSC70-ERP60-GAPDH complex (C), PDC (D), and SDH-mABC1-PIC-ANT-ATPase complex (E). Each node is color coded to represent the Log2 of the median SILAC KO/WT ratio for each protein. The size of the node corresponds to the relative number of Ksucc sites identified on each protein.

Figure 7. SIRT5 Suppresses the Activities of Pyruvate Dehydrogenase Complex and Succinate Dehydrogenase

(A) Quantitative Ksucc changes on PDC subunits between Sirt5 KO and control MEFs (see also Tables S1A and S5). (B) SIRT5 desuccinylates PDC subunits in vitro. Left, ponceau S; Right, IB: Ksucc. (C) SIRT5 suppresses PDC activity in vitro. PDC activity in reactions in (B). Data are mean ± SEM of triplicate measurements; p value designates comparison via a paired two-tailed t test. (D) SIRT5 immunoblot in Sirt5 KD cell lines and control. LC, loading control. (E) PDC activity in Sirt5 KD cells and controls ± 5 mM DCA. Data summarize four independent experiments, each done in quadruplicate. Error bars designate SEM; ***p < 0.001 by two-way ANOVA. (F) Quantitative Ksucc changes on SDHA between SIRT5 KO and control MEFs (see also Tables S1A and S6 and Figure S7). (G) Increased SDH activity in Sirt5 KD cells. Data are mean ± SEM of triplicate measurements; ***p < 0.001 by an unpaired two-tailed t test. (H) Representative oxygen consumption in Sirt5 KD and control cell lines. Additions are as follow: cells; Pyr/Mal, pyruvate + malate; F-Pyr, fluoropyruvate; S/R, succinate + rotenone; AA, antimycin A; A/T, ascorbate + TMPD; KCN. (I) Quantification of (H). Succ, succinate + rotenone; Asc/TMPD, ascorbate + TMPD. Data are mean ± SEM of triplicate measurements; *p < 0.05 by an unpaired two-tailed t test. (J) Increased respiration in the presence of pyruvate or succinate but not ascorbate + TMPD in SIRT5-deficient mitochondria. Top, plotted data; bottom, raw data (blue or turquoise, Sirt5 KO; red or pink, WT control). OCR, oxygen consumption rate. Data are representative of five independent experiments performed in quadruplicate on a littermate WT/KO pair. Statistical comparisons were performed with an unpaired two-tailed t test; error bars designate SEM.