Main

T cells require mitochondrial metabolism as they exit from the naive cell state to become activated, and as they return to being resting memory cells; however, the role of mitochondrial metabolism in the differentiation and function of effector T cells is less well-understood3,4,5. Metabolite tracing studies have revealed that, whereas activated T cells use glutamine for the anaplerosis of α-ketoglutarate, activated cells decrease the rate of pyruvate entry into the mitochondria in favour of lactate fermentation5,6. Despite the decreased utilization of glucose-derived carbon for mitochondrial metabolism, the tricarboxylic acid (TCA) cycle has previously been shown to contribute to IFNγ production by increasing cytosolic acetyl-CoA pools via mitochondrial citrate export7. Additionally, the TCA cycle can contribute to the electron transport chain (ETC) by generating NADH and succinate to fuel complex I and complex II, respectively. However, the role of the ETC in the later stages of T cell activation is poorly characterized. To test the contribution of the TCA cycle to the function of effector T cells, we treated cells cultured in TH1 conditions with the TCA-cycle inhibitor sodium fluoroacetate8. We titrated sodium fluoroacetate or the glycolysis inhibitor 2-deoxy-d-glucose (2DG; an inhibitor of TH1 cell activation, used as a positive control) at day 1 of T cell culture, and assayed cell proliferation at day 3 or the expression of the Ifng-Katushka reporter at day 5. Although 2DG was a more-potent inhibitor than sodium fluoroacetate at lower doses, both inhibitors impaired Ifng transcription (Fig. 1a) and T cell proliferation (Fig. 1b) in a dose-dependent manner, which suggests that the activity of TCA-cycle enzymes is required for optimal TH1 cell activation.

Fig. 1: The TCA cycle supports proliferation and function of T helper cells through distinct mechanisms.
figure 1

a, b, Mean divisions at day 3 (a) and Ifng-Katushka (Ifng-Kat) reporter expression after restimulation with phorbol myristate acetate (PMA) and ionomycin at day 5 (b) of CD4 T cells cultured in TH1 conditions with serially diluted 2DG (n = 3) or sodium fluoroacetate (NaFlAc) (n = 2 or 3). MFI, mean fluorescence intensity. c, d, Proliferation after overnight treatment on day 2 (c) and intracellular IFNγ protein expression after overnight treatment on day 4 (d) of wild-type CD4 T cells cultured in TH1 conditions with dimethylsulfoxide (DMSO), rotenone (rot), DMM, antimycin A (ant A), oligomycin (oligo) or BMS-303141 (n = 3). n, number of technical replicates. Representative plots and a graph summarizing the results of at least two independent experiments are shown. Mean and s.d. of replicates are presented on summarized plots and unpaired, two-tailed t-test used to determine significance. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ns, not significant.

Source Data

Full size image

To evaluate which processes downstream of the TCA cycle contribute to the role of the TCA cycle in T-helper-cell proliferation and function, we treated TH1 cells with inhibitors of the ETC overnight on day 2 (to evaluate proliferation) or overnight on day 4 (to evaluate cytokine production), and analysed cells the following day. Unlike impairing glycolysis with 2DG or the TCA cycle with sodium fluoroacetate, which resulted in a block of both proliferation and function, we observed a dichotomy in the role of the ETC in supporting each of these processes. Although the inhibition of complex II did not impair proliferation, blocking complex I and complex III resulted in a decrease in the number of divided cells; treatment with oligomycin displayed a modest but significant effect (Fig. 1c). Importantly, viability was not affected upon acute inhibition of ETC complexes (Extended Data Fig. 1a). Consistent with this observation, treatment with rotenone or antimycin A on day 2 resulted in cell-cycle arrest at the G2 or M phase, whereas treatment with dimethyl malonate (DMM) or oligomycin did not alter cell-cycle status (Extended Data Fig. 1b). Similar to cells cultured in TH1 conditions, cells cultured in TH2 or TH17 conditions displayed defects in proliferation and an altered cell cycle when treated with rotenone (Extended Data Fig. 2a, b, e, f), which suggests that complex I supports cell division regardless of the cytokine environment.

Further illustrating distinct roles for complex I and complex II in T-helper-cell proliferation and function, we observed that the ATP citrate lysase (ACLY) inhibitor BMS-303141 significantly decreased IFNγ production, consistent with previous work7, whereas the effect of inhibition of complex I or ATP synthase with rotenone or oligomycin, respectively, was not significant. By contrast, impairing complex II activity with DMM, or complex III activity with antimycin A, significantly reduced IFNγ production to levels below those observed with BMS-303141 (Fig. 1d). Together, these observations suggest that the TCA cycle supports TH1 function by enabling cytosolic acetyl-CoA production and by fuelling a succinate-dehydrogenase (SDH)-driven ETC. This role for the ETC was specific to the T-helper-cell cytokine culture conditions to which the cells were exposed during activation. Unlike TH1 cells, inhibiting the ETC had a minimal effect on function of TH2 effector cells; inhibition of complex I or complex III resulted in a slight, but significant, increase in IL-4 reporter activity (Extended Data Fig. 2c). By contrast, TH17 cells displayed sensitivity to inhibition of both complex I and complex II (Extended Data Fig. 2d). These data indicate that the ETC has program-specific roles in regulating the effector functions of T helper cells.

To corroborate the effects of DMM on the function of TH1 cells, we tested the capacity of three additional inhibitors of complex II—thenoyltrifluoroacetone (TTFA), 3-nitropropionic acid (3NP) and atpenin A5—to inhibit IFNγ production in TH1 cells. Each drug impaired complex II activity, as assayed by cellular succinate accumulation (Extended Data Fig. 3a). Consistent with our results for DMM treatment, TH1 cells treated with 3NP, TTFA or atpenin A5 produced significantly less IFNγ than control cells (Fig. 2a). In keeping with a role for the TCA cycle and complex II in promoting TH1 cell function, cells cultured overnight with a membrane-permeable form of succinate (diethyl succinate) produced more IFNγ (Extended Data Fig. 3b). To genetically test the requirement of complex II activity in TH1 cells, we generated a retroviral single-guide (sg)RNA expression vector (which we named MG-Guide) that is compatible with transduction of mouse T cells (Extended Data Fig. 4a, b). To validate the system, we transduced CD4 T cells with sgRNA and observed a rapid loss of protein expression when using sgRNAs that targeted Tbx21 or Il12rb1, genes that are essential for TH1-cell cytokine production; this loss led to a decrease in capacity for IFNγ production (Extended Data Fig. 4c–f, Supplementary Table 1). Transduction of TH1 cells with a sgRNA targeting Sdha, which encodes the catalytic subunit of complex II, impaired capacity for IFNγ production (Extended Data Fig. 3c). To provide further genetic evidence that complex II activity is required for the function of TH1 cells, we tested the requirement for Sdhc, which encodes an essential subunit of complex II. We cultured CD4 T cells isolated from Sdhcfl/fl TetO-cre−/+ Rosa26rtTA/+ (hereafter, Sdhc conditional knockout (cKO)) or Sdhc+/+ TetO-cre−/+ Rosa26rtTA/+ control (hereafter, wild-type) mice that had been treated in vivo with doxycycline for ten days in TH1 conditions. Unbiased mass-spectrometry analysis of metabolites in wild-type and Sdhc cKO TH1 cells revealed that Sdhc cKO cells had increased levels of cellular succinate and α-ketoglutarate, which confirms the loss of SDH activity (Extended Data Fig. 3d, e). Consistent with our drug and sgRNA studies, Sdhc cKO cells produced significantly less IFNγ at day 5 post-activation (Fig. 2b). However, Sdhc cKO TH1 cells proliferated significantly more than wild-type controls, which suggests that proliferation and effector function are processes that are uncoupled by complex II activity (Fig. 2c). To test whether processes in addition to proliferation that are involved in T-helper-cell differentiation were affected, we assayed the effect of SDH deficiency on histone acetylation. We found that Sdhc cKO cells exhibited increased H3K9 acetylation, and that DMM treatment as well as delivery of Sdha-targeting sgRNA increased H3K9 and H3K27 acetylation; this suggests that complex II antagonizes T-helper-cell differentiation by negatively regulating both proliferation and histone acetylation (Fig. 2d, Extended Data Fig. 5a–c).

Fig. 2: Complex II uncouples differentiation and effector function of TH1 cells.
figure 2

a, Intracellular IFNγ protein expression in PMA and ionomycin-restimulated wild-type CD4 T cells cultured in TH1 conditions at day 5 after overnight treatment with DMSO, DMM (10 mM), 3NP (1 mM), TTFA (100 μM) or atpenin A5 (1 μM) (n = 3). b, c, Intracellular IFNγ protein expression (b) and proliferation of CD4 T cells (c) from doxycycline-treated Sdhc cKO or wild-type (WT) mice cultured in TH1 conditions at day 5. Data combined from 5 independent experiments: wild type, n = 13; Sdhc cKO, n = 14 biological replicates. Two-tailed t-test. d, Total cellular H3K9 acetylation (H3K9Ac) of wild-type and Sdhc cKO cells cultured in TH1 conditions at day 3 (n = 3). Two-sided t-test. e, TBET protein expression of wild-type (n = 4) and Sdhc cKO (n = 3) cells cultured in TH1 conditions at day 5. Two-sided t-test. f, DAVID Gene Ontology (GO) pathway analysis of genes that are downregulated in cKO mice compared to wild-type controls. P < 0.05. g, Heat map of gene expression from RNA-seq results for the cytokine production GO pathway. n, number of technical replicates, except where noted otherwise. Representative plots and a graph summarizing the results of at least two independent experiments are shown, except where noted otherwise. Mean and s.d. of replicates are presented on summarized plots and unpaired, two-tailed t-test used to determine significance. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Source Data

Full size image

To test the role of complex II in promoting other aspects of the functional program of TH1 cells, we evaluated TBET protein expression in Sdhc cKO and wild-type cells on day 5 after activation. Consistent with defects in IFNγ production, TH1 cells from Sdhc cKO mice had reduced levels of expression of TBET protein (Fig. 2e). To further investigate a role for complex II in supporting the functional program of TH1 cells, we performed RNA sequencing (RNA-seq) on effector TH1 cells from Sdhc cKO and wild-type mice at day 5 after activation. Consistent with a decrease in TBET expression, TH1 cells from mice deficient in complex II exhibited significantly decreased expression of genes that are key to the TH1 cell program and genes that are important during T-helper-cell activation. Notably, DAVID (Database for Annotation, Visualization and Integrated Discovery) Gene Ontology pathway analysis indicated ‘cytokine production’ and ‘regulation of lymphocyte proliferation’ as the most-dysregulated pathways (Fig. 2f, g, Extended Data Fig. 5d, e, Supplementary Table 2). These data indicate that SDH activity is a primary mechanism through which mitochondrial metabolism supports the functional programming of TH1 cells.

We next sought to investigate which aspects of mitochondrial metabolism are antagonized by SDH to constrain proliferation. The consumption of α-ketoglutarate is known to modulate the activity of mitochondrial shuttling systems that are required to maintain the cellular redox balance and the production of key cytosolic metabolites9,10,11. The malate–aspartate shuttle and mitochondrial citrate export are two such systems; they regulate the oxidation state of nicotinamide adenine dinucleotides (NAD) in the mitochondria and the transport of acetyl-CoA from the mitochondria to the cytosol, respectively. On the basis of our data that Sdhc cKO TH1 cells exhibit increased proliferation (Fig. 2c) and increased cellular α-ketoglutarate levels (Extended Data Fig. 3e), we hypothesized that these mitochondrial transport systems promote the early stages of TH1 cell proliferation.

To test the requirement of these transport systems for TH1 cell activation, we designed three sgRNAs per gene of interest and conducted individual sgRNA knockout experiments using MG-Guide, measuring IFNγ protein (Fig. 3a). We found that, compared to cells transduced with an empty MG-Guide vector, cells that express sgRNAs that target Mdh1, Mdh2, Slc25a11 or Slc1a3 produced less IFNγ protein—comparable to the levels observed with sgRNAs that target the positive-control Tbx21 gene—as did two of the three sgRNAs designed to target Got1 and Got2, which suggests that the malate–aspartate shuttle is critical during TH1 cell activation (Fig. 3b). In addition, we observed defective IFNγ production in TH1 cells that express sgRNA against Cs, Slc25a1 and Acly, which indicates that citrate synthesis and export for cytosolic acetyl-CoA production are also required (Fig. 3b).

Fig. 3: The malate–aspartate shuttle and mitochondrial citrate export are required for histone acetylation and proliferation in differentiating TH1 cells.
figure 3

a, Schematic of the malate–aspartate shuttle and mitochondrial citrate export. aKG, α-ketoglutarate; Asp, aspartate; Cit, citrate; Glu, glutamate; Mal, malate; OAA, oxaloacetate. b, Intracellular IFNγ protein expression in Cas9-expressing CD4 T cells, transduced with sgRNAs targeting the indicated enzymes and transporters, cultured in TH1 conditions after restimulation at day 5. Graphs show individual sgRNAs for each gene as well as the average for all three sgRNAs (n = 2 or 3 biological replicates). c, d, Total cellular H3K9 acetylation at day 4 of Cas9-expressing CD4 T cells transduced with sgRNAs against the indicated enzymes and transporters, in the absence or presence of 5 nM or 20 nM exogenous acetate added 1 day after transduction, cultured in TH1 conditions (n = 3 technical replicates). e, f, Heat map summarizing downregulated genes determined by RNA-seq for cells expressing Slc25a1-targeting sgRNA (e) or Slc25a11-targeting sgRNA (f). P < 0.05. EV, empty vector; KO, knockout. Representative plots and a graph summarizing the results of at least two independent experiments are shown. Mean and s.d. of replicates are presented on summarized plots and unpaired, two-sided t-test used to determine significance. *P < 0.05, **P < 0.01, ***P < 0.001.

Source Data

Full size image

Previous reports have suggested that ACLY activity is required for TH1-cell histone acetylation, and the ETC has previously been shown to support epigenetic remodelling7,12. To test the role of both shuttle systems during TH1-cell epigenetic remodelling, we evaluated total cellular H3K9 and H3K27 acetylation. We found that impairing Acly, Slc25a1, Mdh1, Slc25a11 and Slc1a3 results in decreased H3K9 acetylation, and that acetate supplementation could compensate for these defects (Fig. 3c, d). By contrast, H3K27 acetylation was largely unaffected by targeting these genes (with the exception of Slc25a1); however, the addition of acetate resulted in increased H3K27 acetylation regardless of the condition (Extended Data Fig. 6a). This effect of acetate on histone acetylation is largely explained by an increase in total H3 content, whereas the effect of the sgRNA on acetylation is only partially explained by changes in total histone mass (Extended Data Fig. 6b–d).

To evaluate the transcriptional effects of deficiency in the malate–aspartate shuttle, we performed RNA-seq at day 5 after activation on TH1 cells that express sgRNA against Slc25a1 or Slc25a11. Consistent with a role for the shuttles in promoting TH1 cell differentiation, we observed decreased expression of genes with known roles in T cell activation and TH1 cell programming. Targeting either of the transporters led to impaired expression of Il2rb, whereas loss of Slc25a1 affected key T-cell-activation genes (such as Nfatc1, Rela and Mapk3) and disruption of Slc25a11 resulted in the loss in expression of genes including Tbx21, Nfatc3, Ccnd2 and Myc (Fig. 3e, f, Extended Data Fig. 6e, f, Supplementary Tables 3, 4).

Given the importance of Il2rb, Myc and Ccnd2 in T-helper-cell division, we next evaluated the role of the shuttles in regulating T-helper-cell proliferation. To test this, we evaluated cell division in cells cultured in TH1 conditions that express sgRNAs targeting Acly, Slc25a1, Mdh1, Slc25a11 or Slc1a3. Relative to controls, targeting any of these genes resulted in modestly—but significantly—decreased proliferation (Extended Data Fig. 7). Collectively, these data demonstrate that the malate–aspartate shuttle and mitochondrial citrate export are required for TH1 cell proliferation and transcriptional remodelling.

To investigate the biochemical mechanism that might explain these observations, we performed mass-spectrometry analysis of T cells transduced with guides targeting either Slc25a1 or Slc25a11 sgRNA. As expected, we found that disrupting citrate transport results in decreased levels of cellular acetyl-CoA (Extended Data Fig. 8a–c). Unexpectedly, targeting Slc25a11 resulted in a decreased cellular NADH/NAD+ ratio, which suggests that the activity of complex I is a primary mechanism by which cellular NADH/NAD+ is regulated in activated TH1 cells (Fig. 4a, Extended Data Fig. 8d, e). Moreover, targeting either shuttle system resulted in diminished levels of intermediates of the pentose phosphate pathway and of N-carbamoyl-l-aspartate, an essential precursor molecule for nucleotide synthesis (Fig. 4a, Extended Data Figs. 8b, c, 9a, b). Consistent with a role for the shuttling systems in providing mitochondrial NADH for the ETC, Seahorse analysis demonstrated that rates of basal and maximal oxygen consumption were impaired upon expression of sgRNAs targeting either Mdh1, Slc25a11 or Slc1a3 (Fig. 4b). This was not substantially compensated for by increased glycolysis, as the extracellular acidification rate was minimally affected (Fig. 4b).

Fig. 4: The malate–aspartate shuttle promotes complex I activity, which is required for aspartate synthesis and T-helper-cell proliferation.
figure 4

a, Cellular NADH/NAD+ ratio and N-carbamoyl-l-aspartate measured by liquid chromatography–mass spectrometry analysis in Cas9-expressing CD4 T cells transduced with sgRNA targeting Scl25a11, and cultured in TH1 conditions as described in Methods (n = 2 biological replicates, n = 2 technical replicates). AU, arbitrary units. b, Baseline oxygen consumption rate (OCR), maximal OCR and baseline extracellular acidification rate (ECAR) of Cas9-expressing CD4 T cells transduced with sgRNAs targeting the indicated enzymes and transporters, cultured in TH1 conditions at day 4 (n = 3 biological replicates). c, d, Cellular NADH/NAD+ and ATP/AMP ratios (c) and aspartate and N-carbamoyl-l-aspartate (d) measured by liquid chromatography–mass spectrometry analysis in wild-type CD4 T cells cultured in TH1 conditions, and treated with DMSO or rotenone for 4 h on day 4 (n = 3 technical replicates). e, Proliferation measured at day 3 of wild-type CD4 T cells cultured in TH1 conditions and treated on day 2 with DMSO (clear and grey bar) or rotenone (blue bars) ± 20 mM aspartate (n = 3 technical replicates). Representative plots and a graph summarizing the results of at least two independent experiments are shown. Mean and s.d. are presented on summarized plots and unpaired, two-sided t-test used to determine significance. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Source Data

Full size image

Having observed that complex I supports early T-helper-cell proliferation and that the malate–aspartate shuttle fuels complex I (Fig. 1c), we next sought to examine the biochemical mechanism by which complex I promotes proliferation by performing mass-spectrometric analysis on rotenone-treated cells. As expected, inhibiting complex I increased the NADH/NAD+ ratio and decreased the ATP/AMP ratio (Fig. 4c, Extended Data Fig. 9a, b). Rotenone treatment also led to decreased pools of cellular aspartate and N-carbamoyl-l-aspartate in these cells, similar to previous observations in cancer-cell lines13,14 (Fig. 4d). To test whether this aspartate synthesis deficiency contributed to the proliferative defects of rotenone-treated cells, we supplemented rotenone-treated cells with aspartate and evaluated cell division and the cell cycle. Aspartate supplementation resulted in a significant recovery of cell proliferation, and a partial release from the arrest at the G2 or M phase following rotenone treatment (Fig. 4e, Extended Data Fig. 9c). These data demonstrate that the regulation of complex I by mitochondrial shuttling systems determines the cellular redox balance and the cytosolic aspartate availability that is required for T cell proliferation.

Using approaches that combine network-level genetic interrogation of metabolic pathways, pharmacology, transcriptomics and metabolomics, we demonstrate how TH1 cells meet the distinct metabolic demands of differentiation and function during the course of activation. To generate the substrates needed for proliferation and epigenetic remodelling, early activated T helper cells fuel complex I through the malate–aspartate shuttle and mitochondrial citrate export. Unlike the carbon-neutral malate–aspartate shuttle (which exchanges malate for α-ketoglutarate), complex II moves carbon forward in the TCA cycle; this restricts processes that support differentiation and promotes the late-stage effector function of TH1 cells, which permits cells to exit the cell cycle and adopt their terminal program (Extended Data Fig. 10). These findings illustrate how differentiation and terminal effector function—previously understood to be concordantly regulated by signal transduction—are controlled by distinct metabolic modules, which elucidates how cell programming is governed by parallel transcriptional and biochemical networks.

Methods

T cell assays and sgRNA delivery

CD4 T cells were isolated from constitutive Cas9-expressing (Cas9tg) B6 mice15, stimulated with anti-CD3 and anti-CD28 coated beads (Miltenyi T Cell Activation/Expansion Kit, mouse), and cultured in assay-determined TH1 conditions (5 ng/ml IL-2, 2 ng/ml IL-12 and 10 μg/ml anti-IL-4). On day 1 post-activation, T cells were transduced with MG-Guide retrovirus using spin transduction at 1,200g for 90 min at 37 °C. IFNγ cytokine was measured by adding brefeldin A, 1 h after the addition of PMA (20 ng/ml) and ionomycin (20 ng/ml); 4 h after restimulation, cells were fixed, stained with anti-CD4 (Biolegend), anti-GFP (Millipore) and anti-IFNγ (Biolegend), and analysed by flow cytometry. To assay for Ifng-Katushka, IL-4–GFP, and IL-17–GFP expression, T cells from Ifng-Katushka16, 4GET (Jackson Labs, 004190) and IL-17–GFP (Jackson Labs, 018472) reporter mice were activated with PMA and ionomycin for four hours, stained with anti-CD4 and then analysed by flow cytometry for reporter activity in GFP+ cells. Cell division was measured by labelling cells with CellTrace Violet (Thermo) before activation, and evaluated for proliferation at day 3 after activation; where indicated, inhibitors and metabolites were added to the medium overnight on day 2 after activation. Cell-cycle status was determined by intracellular flow cytometry analysis of Ki67 and DAPI, at day 3 after activation; where indicated, inhibitors and metabolites were added to the medium overnight on day 2 after activation. Mitochondrial reactive oxygen species was measured by flow cytometry in CD4 T cells by staining cells with MitoSOX Red mitochondrial superoxide indicator (Thermo) and anti-CD4 for 30 min at 37 °C in the presence of the indicated inhibitors. For all experiments using inhibitors or metabolite supplementation, the following doses were used: 1 μM rotenone (Sigma), 10 mM DMM (Sigma), 1 mM 3NP (Sigma), 100 μM TTFA (Sigma), 1 μM atpenin A5 (Cayman Chemical), 1 μM antimycin A (Sigma), 1 μM oligomycin (Sigma), 5mM diethyl succinate (Sigma) or 20mM aspartate. All mice required for this study were housed and maintained under specific-pathogen-free conditions in the animal facility of the Yale University School of Medicine, and all corresponding animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Yale University. This study was conducted in compliance with all relevant ethical regulations. All cells used for experimentation were collected from male and female mice at 6–8 weeks of age.

MG-Guide vector generation, sgRNA cloning and retroviral production

MG-Guide was generated by removing the IRES element from MIGR1 (Addgene) by EcoRI and NotI digestion, and adding the human U6 promoter and SV40 promoter from pMKO-GFP (Addgene) by infusion assembly (Clonetech). To add the sgRNA cloning site, the vector was digested with AgeI and EcoRI and combined by infusion assembly with an IDT Gene Block containing two BbsI restriction sites upstream of a scaffold RNA sequence and a U6 stop. To clone individual sgRNAs, MG-Guide was digested with BbsI and pairs of oligonucleotides (Sigma) with complimentary overhangs were annealed and ligated into the vector. For retroviral production, 1 μg of MG-Guide plasmid and 0.5 μg of EcoHelper plasmid were transfected into 5 × 105 HEK293T cells (source ATCC, identity unconfirmed, not tested for mycoplasma) in a 6-well plate using X-tremeGENE 9 DNA Transfection Reagent (Roche) overnight. The medium was then replaced, and virus was collected 24 h later. Isolated CD4 T cells (1 × 106) were stimulated overnight, and spin-transduced in the viral preparation with 1 μg/ml polybrene at 1,200g for 90 min at 37 °C.

RNA-seq analysis

Raw reads from RNA-seq were aligned to the mouse genome mm10 with STAR 2.7.017, and gene-expression levels were measured by HTSeq 0.11.118. Subsequently, differential expression analysis between different groups was performed with DESeq219.

Seahorse analysis

Analysis was performed on cells at day 3, day 4 and day 5 after activation. Cells were washed three times in complete Seahorse medium (Seahorse Bioscience) with 10 mM glucose, 1 mM sodium pyruvate and 2 mM glutamine. Cells were plated at 4 × 104 cells per well in a 96-well Seahorse assay plate, pretreated with poly-d-lysine. Cells were equilibrated to 37 °C for 30 min before assay. OCR (pmoles/min) and ECAR (mpH/min) were measured as indicated upon cell treatment with oligomycin (0.5 mM), FCCP (0.2 mM), rotenone (1 μM), DMM (10 mM) and antimycin A (1 μM), according to the manufacturer’s instructions.

Metabolome extraction

Cells were seeded at 1 × 106 cells/ml and incubated for 4 h in complete RPMI containing dialysed FBS medium. Cells were then transferred to 1.5-ml tubes and pelleted (1 min, 6,000g, at room temperature). Medium was removed by aspiration and the cells were washed once with 500 μl of PBS. Metabolome extraction was performed by the addition of 50 μl of ice cold solvent (40:40:20 acetonitrile:methanol:water + 0.5% formic acid). After a 5-min incubation on ice, acid was neutralized by the addition of NH4HCO3. After centrifugation (15 min, 16,000g, at 4 °C), the clean supernatant was transferred to a clean tube, frozen on dry ice and kept at -80 °C until liquid chromatography–mass spectrometry (LC–MS) analysis20.

Succinate quantification

Wild-type CD4 T cells (1 × 106) were activated under TH1 culture conditions. After 4 days, cells were replated into fresh medium and cultured with DMSO, 10 mM DMM, 1 mM 3NP, 100 μM TTFA or 1 μM atpenin A5 for 6 h. Cells were then collected, processed and analysed using the Succinate Assay Kit (Abcam) according to the manufacturer’s protocol.

LC–MS analysis

Cell extracts were analysed using a quadrupole–orbitrap mass spectrometer (Q Exactive, Thermo Fisher Scientific) coupled to hydrophilic interaction chromatography via electrospray ionization. Liquid chromatography separation was on a XBridge BEH Amide column (2.1 mm × 150 mm, 2.5-μm particle size; Waters) using a gradient of solvent A (20 mM ammonium acetate, 20 mM ammonium hydroxide in 95:5 water:acetonitrile, pH 9.45) and solvent B (acetonitrile). Flow rate was 150 μl/min, column temperature was 25 °C, autosampler temperature was 5 °C and injection volume was 10 μl. The liquid chromatography gradient was: 0 min, 90% B; 2 min, 85% B; 3 min, 75% B; 7 min, 75% B; 8 min, 70% B; 9 min, 70% B; 10 min, 50% B; 12 min, 50% B; 13 min, 25% B; 14 min, 25% B; 16 min, 0% B; 21 min, 0% B; 22 min, 90% B; 25 min, 90% B. Autosampler temperature was 5 °C and injection volume was 10 μl. The mass spectrometer was operated in negative-ion mode to scan from m/z 70 to 1,000 at 1 Hz and a resolving power of 140,00021. Data were analysed using the MAVEN software22.

Statistical analysis

Experiments were conducted with technical and biological replicates at an appropriate sample size, as estimated by our prior experience. No statistical methods were used to predetermine sample size. No methods of randomization and no blinding were applied. All data were replicated independently at least once as indicated in the figure legends, and all attempts to reproduce experimental data were successful. For all bar graphs, mean + s.d. are shown. All statistical analysis was performed using GraphPad Prism 7 (or more recent versions). P values < 0.05 were considered significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; P values > 0.05 were considered as non-significant. FlowJo 8.0 (or more recent versions) (Treestar) was used to analyse flow cytometry data. All sample sizes and statistical tests used are detailed in each figure legend.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this paper.