CRISPR screens unveil signal hubs for nutrient licensing of T cell immunity

Abstract

Nutrients are emerging regulators of adaptive immunity¹. Selective nutrients interplay with immunological signals to activate mechanistic target of rapamycin complex 1 (mTORC1), a key driver of cell metabolism^2-4, but how these environmental signals are integrated for immune regulation remains unclear. Here we use genome-wide CRISPR screening combined with protein-protein interaction networks to identify regulatory modules that mediate immune receptor- and nutrient-dependent signalling to mTORC1 in mouse regulatory T (T_reg) cells. SEC31A is identified to promote mTORC1 activation by interacting with the GATOR2 component SEC13 to protect it from SKP1-dependent proteasomal degradation. Accordingly, loss of SEC31A impairs T cell priming and T_reg suppressive function in mice. In addition, the SWI/SNF complex restricts expression of the amino acid sensor CASTOR1, thereby enhancing mTORC1 activation. Moreover, we reveal that the CCDC101-associated SAGA complex is a potent inhibitor of mTORC1, which limits the expression of glucose and amino acid transporters and maintains T cell quiescence in vivo. Specific deletion of Ccdc101 in mouse T_reg cells results in uncontrolled inflammation but improved antitumour immunity. Collectively, our results establish epigenetic and post-translational mechanisms that underpin how nutrient transporters, sensors and transducers interplay with immune signals for three-tiered regulation of mTORC1 activity and identify their pivotal roles in licensing T cell immunity and immune tolerance.

Extended Data Figure 1 (to Figure 1).. Two rounds of pooled CRISPR screening to identify novel regulators of nutrient and mTORC1 signaling.

(a) Flow cytometry analysis of IFN-γ, IL-4, IL-17A or Foxp3 expression in cells cultured in T_H0-, T_H1-, T_H2-, T_H17- or induced T_reg-polarizing condition (n = 3 samples each group). (b) Flow cytometry analysis of p-S6 in T_H1 and T_reg cells with TCR stimulation for 0 or 1 h (n = 3 samples each group). (c) Induced T_reg cells were stimulated with α-CD3/CD28 antibodies in the presence or absence of amino acids (AA) or glucose for 3 h followed by flow cytometry analysis and quantification of p-S6 level [based on mean fluorescence intensity (MFI)] (n = 3 samples each group). (d) Induced T_reg cells were labelled with CellTrace Violet (CTV) and stimulated with α-CD3/CD28 antibodies in the presence or absence of AA or glucose for 3 d, followed by flow cytometry analysis of CTV dilution (n = 3 samples each group). (e) Gating strategy used for sorting cells with the ≥ 10% highest (p-S6^hi) and ≤ 10% lowest (p-S6^lo) levels after stimulation with 0.25 or 4 μg/ml of α-CD3 antibody for 3 h (n = 2 samples each group). Mean ± s.e.m. (c). ***P < 0.001; one-way ANOVA (c). Data are representative of two (b–d) or three (a) experiments.

Extended Data Figure 2 (to Figure 1).. Validation of individual candidate mTORC1 regulators.

(a) Flow cytometry analysis and quantification of p-S6 level [based on mean fluorescence intensity (MFI)] in naïve or activated WT and Depdc5-deficient CD4⁺Foxp3⁻ T cells (n = 4 samples each group). Naïve CD4⁺ T cells among freshly isolated splenocytes from WT and Cd4^CreDepdc5^fl/fl mice were gated (indicated as TCR 0 h), or naïve CD4⁺ T cells were sorted and stimulated with α-CD3/CD28 antibodies overnight for flow cytometry analysis of p-S6 level. (b) Quantification of relative p-S6 level in induced T_reg cells transduced with sgNTC, sgSec13, sgMios, sgSeh1l, or sgWdr24 (all Ametrine⁺) were stimulated with TCR for 3 h (n = 3 samples each group). (c) Diagram of dual-color co-culture system to examine cell-intrinsic effects of deletion of a candidate gene on TCR-induced p-S6, cell size and CD71 expression. Specifically, Cas9⁺ cells transduced with sgNTC (mCherry⁺ or GFP⁺; ‘spike’) were mixed with those transduced with those targeting a specific gene (Ametrine⁺), and stimulated with α-CD3 antibody for 3 h (for p-S6) or with α-CD3/CD28 antibodies for 20 h (for cell size and CD71). (d) Validation of dual-color co-culture system by using two sgNTC-expressing vectors with different fluorophores. Cells transduced with sgNTC (GFP⁺; ‘spike’) were mixed with those transduced with sgNTC (Ametrine⁺), and stimulated with α-CD3 antibody for 3 h to examine p-S6 (see phos-flow staining), or stimulated with α-CD3/CD28 antibodies for 20 h (n = 3 samples each group) to measure cell size and CD71 expression (see surface staining). (e) Heatmap summary of log₂ FC (p-S6^hi/p-S6^lo) for individually validated candidate genes (63 positive and 21 negative regulators) including positive (Rheb, Rptor, Lamtor3, Rraga and Mtor) and negative (Cd5, Nprl3, Nprl2 and Tsc1) control genes (2 sgRNAs for each candidate). Specifically, Cas9-expressing CD4⁺ T cells transduced with sgRNA for target genes (Ametrine⁺) or non-targeting control sgRNA (sgNTC) (mCherry⁺; ‘spike’) were mixed and differentiated into induced T_reg cells. These cells were then stimulated with α-CD3 antibody for 3 h (n = 3 samples each group). Relative p-S6 level (normalized to ‘spike’) was analyzed by flow cytometry. (f) Analysis of protein–protein interaction (PPI) networks of high-confidence regulators. Specifically, 286 positive and 60 negative high-confidence hits were integrated with the composite PPI databases that encompass STRING, BioPlex and InWeb_IM databases for the inference of functional modules. Red and blue circles represent genes whose deletion represses and promotes mTORC1 activity, respectively. Mean ± s.e.m. (a, b). *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA (a, b). Data are representative of one (f) or three (d, e), or pooled from two (a) or three (b) experiments.

Extended Data Figure 3 (to Figure 1).. SWI/SNF complex represses expression of nutrient sensor Castor1 to support mTORC1 activation.

(a) Quantification (normalized to ‘spike’) of relative p-S6 level, cell size (FSC-A) and CD71 expression in induced T_reg cells transduced with indicated sgRNAs followed by stimulation with α-CD3 antibody for 3 h to measure p-S6 level, or with α-CD3/CD28 antibodies for 20 h to measure cell size (FSC-A) and CD71 expression by flow cytometry (n = 3 samples each group). (b) Imaging analysis and quantification of lysosome-associated mTOR [based on mean fluorescence intensity (MFI)] in sgNTC- or sgSmarcb1-transduced cells that were stimulated with α-CD3 antibody for 3 h, starved of amino acids (AA), and refed AA for 20 min (n > 230 cells per condition). (c) Volcano plots of expression levels of transcripts, including Castor1, in sgNTC or sgSmarcb1 (both Ametrine⁺)-transduced cells that were stimulated with CR for 3 h (n = 4 samples each group). (d) Castor1 mRNA expression in sgNTC- or sgSmarcb1 (both Ametrine⁺)-transduced cells were stimulated with α-CD3 antibody for 3 h or with α-CD3/CD28 antibodies for 20 h (n = 3 samples per group). (e) sgNTC or sgSmarcb1 (both Ametrine⁺)-transduced cells were left unstimulated (indicated by 0 h) or stimulated with α-CD3 antibody for 3 h or α-CD3/CD28 antibodies for 20 h. Immunoblot analysis and quantification of relative Castor1 expression (n = 3 samples each group). (f) Immunoblot analysis and quantification of relative p-S6K1 and p-S6 levels in cells transduced with empty vector or vector expressing Castor1, followed by stimulation with α-CD3/CD28 antibodies for 2 d (n = 3 samples each group). Mean ± s.e.m. (a, b, d–f). **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (f); one-way ANOVA (a, b, d, e). Data are representative of one (c) or two (b), or pooled from two (d, e) or three (a, f) experiments.

Extended Data Figure 4 (to Figure 2).. Sec31a is required for mTORC1 activation.

(a) Interaction of endogenous Sec13 with Sec31a in induced T_reg cells as assessed by immunoprecipitation (IP)–immunoblot analysis. (b, c) sgNTC- or sgSec31a-transduced cells were starved of and refed amino acids (AA, b) or glucose (c) for 20 min, followed by immunoblot analysis of Sec31a, p-S6K1, p-S6 and β-Actin. Lower, quantification of relative p-S6K1 and p-S6 levels (n = 3 samples each group). (d) Cells transduced with the indicated sgRNAs (all Ametrine⁺) were mixed with sgNTC (mCherry⁺; ‘spike’)-transduced cells, and stimulated with α-CD3 antibody for 3 h to measure p-S6 level or with α-CD3/CD28 antibodies for 20 h to measure cell size (FSC-A) and CD71 expression by flow cytometry (normalized to ‘spike’) (n = 3 samples each group). (e–g) sgNTC-, sgSec31a- or sgSec13 (all Ametrine⁺)-transduced cells were co-transduced with constitutively active RagA^Q66L (CA-RagA)-expressing retrovirus (GFP⁺) or sgNprl2 (GFP⁺), followed by stimulated with TCR for 3 h to examine relative p-S6 level by flow cytometry (n = 3 samples each group) (e), p-S6K1 and p-S6 levels by immunoblot analysis (n = 4 samples each group) (f), or lysosomal localization of mTOR [based on mean fluorescence intensity (MFI)] (n > 700 cells per condition) (g). In f, two different exposures for p-S6K1 were included to account for the differential intensities between mock and CA-RagA or sgNprl2 conditions, and relative p-S6K1 level was quantified from long exposure for mock and short exposure for CA-RagA or sgNprl2 conditions (middle). Lower, quantification of p-S6 level. Mean ± s.e.m. (b–g). NS, not significant; **P < 0.01; ***P < 0.001; one-way ANOVA (b–g). Data are representative of two (e–g) or four (a), or pooled from three (b–d) experiments.

Extended Data Figure 5 (to Figure 2).. Sec31a–Sec13 axis promotes mTORC1 activation and cell proliferation in vivo.

(a) Cells transduced with sgNTC or sgSec31a (both Ametrine⁺) were labelled with CellTrace Violet (CTV) and transferred into Rag1^−/− mice. Flow cytometry analysis of CTV dilution, and quantification of percentage of proliferated (CTV^lo) cells at 7 d after transfer (n = 5 samples each group). (b) WT, Raptor- and Sec31a-null T_reg cells (all CD45.1⁺Ametrine⁺) were mixed with conventional CD4⁺ T cells (T_conv; CD45.2⁺) at a 1:4 ratio and transferred into Rag1^−/− mice. Quantification of the accumulation of conventional T cells in the spleen at 7 d after transfer (n = 5 samples each group). (c) Naïve CD4⁺ T cells were stimulated with α-CD3/CD28 antibodies for 0, 24, 48 or 72 h followed by immunoblot analysis of the indicated protein expression, and quantification of p-S6K1, Sec13, Sec31a, and Tsc2 (n = 3 samples each group). (d) sgNTC-, sgSec31a- and sgSec13-transduced cells were sorted and lysed with CHAPS buffer for immunoprecipitation (IP) with an antibody against Wdr24. The immunoprecipitated proteins were analyzed by immunoblot for Wdr24, Wdr59, Mios, Seh1l, Sec13, Sec31a, Sec23a and β-Actin. Mean ± s.e.m. (a–c). **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (a); one-way ANOVA (b). Data are representative of one (a, b) or two (c, d) experiments.

Extended Data Figure 6 (to Figure 3).. Sec31a protects Sec13 from proteasomal degradation to sustain mTORC1 activation.

(a) Sec13 mRNA expression in sgNTC- or sgSec31a-transduced induced T_reg cells (n = 3 samples each group). (b) sgNTC- or sgSec31a-transduced induced T_reg cells were sorted and treated with cycloheximide (CHX) for the indicated times. Total protein extracts of cells transduced with sgNTC (5 μg) or sgSec31a (12.5 μg; more protein was loaded to equalize basal Sec13 amount between these cells) were used for immunoblot analysis and quantification of relative Sec13 abundance (n = 4 samples each group). (c) Naïve CD4⁺ T cells were stimulated with α-CD3/CD28 antibodies for 0, 24, or 48 h and then treated with DMSO or MG132 at 48–72 h of stimulation, followed by immunoblot analysis and quantification of Sec13 and Sec31a expression (n = 3 samples each group). (d) HEK293T cells were transfected with HA-tagged Sec13 and 6× His-tagged WT-, K48R- or K63R-ubiquitin and treated with MG132 for 6 h. Ni-nitrilotriacetic acid (Ni-NTA) bead-based pulldown and immunoblot analysis of HA-Sec13. Lower, expression of indicated proteins in whole cell lysates (WCL). (e) sgNTC- or sgSec31a-transduced HEK293T cells were transfected with HA-tagged Sec13 and 6× His-tagged WT ubiquitin (His-tagged Ub), and treated with MG132 for 6 h. Left, Ni-NTA bead-based pulldown of His-tagged Ub-labelled proteins followed by immunoblot analysis for HA-Sec13. Right, immunoblot analysis of WCL for expression of endogenous Sec31a or HA-Sec13, His-tagged Ub, and β-Actin. Mean ± s.e.m. (a–c). NS, not significant; **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (a); two-way ANOVA (b); one-way ANOVA (c). Data are representative of two (d, e) or three (a), or pooled from two (b) or three (c) experiments.

Extended Data Figure 7 (to Figure 3).. Sec31a protects Sec13 from Skp1-mediated proteasomal degradation and supports T cell functional fitness.

(a) HA-tagged WT or the indicated lysine mutant constructs of Sec13 were transfected into HEK293T cells individually. Immunoblot analysis of HA and Hsp90. (b) HA-tagged WT or K260R mutant Sec13 was transfected into HEK293T cells together with the K48-only His-Ub followed by MG132 treatment and anti-HA immunoprecipitation (IP). Immunoblot analysis for HA, His-Ub and β-Actin. WCL, whole cell lysate. (c) Cas9-expressing CD4⁺ T cells were transduced with sgNTC or sgSec31a retrovirus (Ametrine⁺) together with WT or K260R mutant Sec13-expressing retrovirus (GFP⁺). Ametrine⁺GFP^lo cells (see gating on flow cytometry plot, upper) were stimulated with TCR for 0 or 3 h. Immunoblot analysis and quantification of relative Sec13 and p-S6 levels (n = 4 samples each group). (d) Volcano plot of proteins, including Skp1, that interact with HA-Sec13 in induced T_reg cells as identified by mass spectrometry (n = 3 samples each group). (e) Induced T_reg cells transduced with HA-tagged-Sec13- or empty vector-expressing retrovirus were lysed with CHAPS buffer followed by α-HA immunoprecipitation (IP) and immunoblot analysis of HA and Skp1. (f) Induced T_reg cells were lysed with CHAPS buffer followed by immunoprecipitation of endogenous Skp1 and immunoblot analysis of Skp1 and Sec13. (g) Interaction of endogenous Skp1 with Sec13 in sgNTC- or sgSec31a-transduced cells. (h) Naïve CD4⁺ T cells were stimulated with α-CD3/CD28 antibodies for 0, 24, 48 or 72 h. Immunoblot analysis of anti-Skp1 immunoprecipitants and WCL for Skp1, Sec13, and β-Actin (n = 2 samples per group). (i) Immunoblot analysis and quantification of relative expression of indicated proteins in sgNTC- or sgSkp1 (both Ametrine⁺)-transduced cells (n = 3 samples per group). (j) Indicated sgRNA-transduced cells were labelled with CellTrace Violet (CTV), and stimulated with α-CD3/CD28 antibodies for 72 h, followed by flow cytometry analysis and quantification of CTV dilution (n = 3 samples each group). (k) Diagram of SMARTA T cell transfer and LCMV infection system. Briefly, SMARTA–Cas9 CD4⁺ T cells (CD45.1⁺) transduced with sgRNA for candidate genes (CD45.1⁺Ametrine⁺) and mixed with sgNTC (CD45.1⁺mCherry⁺; ‘spike’)-transduced cells at a 1:1 ratio, and adoptively transferred into naïve (unchallenged; CD45.2⁺) mice that were left uninfected (see l) or challenged with LCMV infection (see Fig. 3f). (l) Quantification of the relative proportion (normalized to ‘spike’) of donor-derived (CD45.1⁺) T cells in the spleen of uninfected mice at 7 d after transfer (n =6 mice per group). (m) Cells transduced with sgNTC (Ametrine⁺), sgSec31a (Ametrine⁺) or sgSec31a-Skp1 (GFP⁺ and Ametrine⁺) were sorted and stimulated with α-CD3/CD28 antibodies for 20 h (n = 6-7 samples per group), followed by the measurement of extracellular acidification rate (ECAR). Oligo, oligomycin; FCCP, fluoro-carbonyl cyanide phenylhydrazone; Rot, rotenone. Mean ± s.e.m. (c, i, j, l, m). **P < 0.01; ***P < 0.001; one-way ANOVA (c, i, j, l, m, right); two-way ANOVA (m, left). Data are representative of one (d, j, l, m) or two (a, b, e–h), or pooled from two (c) or three (i) experiments.

Extended Data Figure 8 (to Figure 4).. SAGA complex suppresses mTORC1 activation.

(a) Quantification (normalized to ‘spike’) of relative p-S6 level, cell size (FSC-A) and CD71 expression in induced T_reg cells transduced with sgRNA for Ccdc101 or Taf6l, followed by stimulation with α-CD3 antibody for 3 h to measure p-S6 level (left) or with α-CD3/CD28 antibodies for 20 h to measure cell size (FSC-A; middle) and CD71 (right) expression by flow cytometry (n = 3 samples each group). (b) Immunoblot analysis and quantification of relative p-S6K1 and p-S6 expression in sgNTC- or sgCcdc101-transduced cells that were starved of and refed amino acids (AA) for 20 min (n = 4 samples each group). Mean ± s.e.m. (a, b). *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA (a, b). Data are representative of three (a), or pooled from three (b) experiments.

Extended Data Figure 9 (to Figure 4).. SAGA complex represses the expression of nutrient transporters and mTORC1 activation.

(a) Heatmap of differentially expressed genes in Ccdc101-null T_reg cells stimulated with α-CD3/CD28 antibodies for 0 (n = 3 samples each group) or 20 h (n = 4 samples each group). (b) Slc2a1, Slc16a10 or Slc43a1 mRNA expression in sgNTC- or sgCcdc101-transduced cells at steady state (n = 3 samples each group). (c) Immunoblot analysis and quantification of relative Glut1 expression in sgNTC- or sgCcdc101-transduced cells that were stimulated with α-CD3/CD28 antibodies for 0 or 20 h (n = 3 samples each group). (d) Flow cytometry analysis and quantification of 2-NBDG uptake in sgNTC or sgCcdc101 (both Ametrine⁺)-transduced cells were stimulated with α-CD3/CD28 antibodies for 20 h (n = 4 samples each group). (e, f) Quantification of relative p-S6 level in cells transduced with the indicated sgRNAs that were stimulated with TCR for 3 h (n = 3 samples each group). (g) Principal component analysis (PCA) of ATAC-seq for cells transduced with sgNTC (n = 4 samples) or sgCcdc101 (both Ametrine⁺) (n = 3 samples) and stimulated with α-CD3/CD28 antibodies for 20 h. (h) Motif enrichment analysis of ATAC-seq of sgNTC- and sgCcdc101-transduced cells (n = 4 samples each group). (i) Footprinting analysis of Sp3 binding in ATAC-seq. (j) Accessibility of the Sp3 locus in sgNTC- and sgCcdc101-transduced cells as identified by ATAC-seq. Highlighted peaks in the red box indicate differential accessible regions. (k) Immunoblot analysis of Sp3 expression in sgNTC- or sgCcdc101 (both Ametrine⁺)-transduced cells. Mean ± s.e.m. (b–f). ***P < 0.001; two-tailed unpaired Student’s t-test (b, d); one-way ANOVA (c, e, f). Data are representative of one (a, g–j), two (b, e, f, k), or pooled from two (c, d) experiments.

Extended Data Figure 10 (to Figure 4).. SAGA complex prevents mTORC1 hyperactivation to enforce immune homeostasis in vivo.

(a, b) Induced T_reg cells transduced with indicated sgRNAs were stimulated with α-CD3/CD28 antibodies for 20 h (n = 3 samples each group). Flow cytometry analysis and quantification of staining with active caspase-3 (a) and fixable viability dye (FVD, b) (n = 3 per group). (c) Cells transduced with indicated sgRNAs were stimulated with α-CD3/CD28 antibodies for 20 h (n = 5–7 samples per group), followed by the measurement of extracellular acidification rate (ECAR). Oligo, oligomycin; FCCP, fluoro-carbonyl cyanide phenylhydrazone; Rot, rotenone. (d) Immunoblot analysis and quantification of Ccdc101 expression in naïve CD4⁺ T cells from WT and Cd4^CreCcdc101^fl/fl mice (n = 3 mice per group). (e) Quantification of CD71 expression on naïve or activated WT and Ccdc101-deficient CD4⁺ T cells. Naïve CD4⁺ T cells among freshly isolated splenocytes from WT and Cd4^CreCcdc101^fl/fl mice (n = 4 mice per group) were gated (indicated as 0 h), or naïve CD4⁺ T cells were stimulated with α-CD3/CD28 antibodies for 20 h. (f) Flow cytometry analysis and quantification of numbers of total, double-negative (DN), double-positive (DP), CD4 single-positive (CD4SP), and CD8 single-positive (CD8SP) thymocytes from WT and Cd4^CreCcdc101^fl/fl mice (n = 4 mice each group). (g) Flow cytometry analysis and quantification of proportions and numbers of splenic CD4⁺ and CD8⁺ T cells from WT and Cd4^CreCcdc101^fl/fl mice (n = 4 mice each group). (h) Flow cytometry analysis and normalized ratio of CD122⁺ versus CD122⁻ cells among CD44^hi populations (gated on splenic CD8⁺ T cells) from indicated mice (n = 4 mice each group). Mean ± s.e.m. (a–h). NS, not significant; **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (d–h); one-way ANOVA (a, b, c, right); two-way ANOVA (c, left). Data are representative of one (c) or two (a, b), or pooled from three (d–h) experiments.

Extended Data Figure 11 (to Figure 5).. T_reg-specific deletion of Ccdc101 disrupts immune homeostasis and boosts antitumour response.

(a) Quantification of relative p-S6 level in splenic CD4⁺Foxp3⁺ cells from WT and Foxp3^CreCcdc101^fl/fl (~8 weeks old) mice (n = 4 mice per group). (b) Quantification of relative Foxp3 expression (gated on splenic Foxp3⁺CD4⁺ T cells) from WT and Foxp3^CreCcdc101^fl/fl mice (n = 4 mice per group). (c) Quantification of percentages of effector/memory (CD44^hiCD62L^lo) subsets in splenic CD4⁺Foxp3⁻ and CD8⁺ T cells from WT and Foxp3^CreCcdc101^fl/fl (~8 weeks old) mice (n = 4 mice each group). (d) Representative flow cytometry analysis of IL-2⁺ or IFN-γ⁺ population of splenic CD4⁺Foxp3⁻ and CD8⁺ T cells from WT and Foxp3^CreCcdc101^fl/fl (~8 weeks old) mice. (e–g) WT and Foxp3^CreCcdc101^fl/fl mice were inoculated with MC38 colon adenocarcinoma cells. T_reg cells (CD45⁺CD4⁺YFP⁺), non-T_reg immune cells (CD45⁺YFP⁻CD11b⁻) and myeloid cells (CD45⁺CD11b⁺) were isolated and sorted from tumours, and mixed at a 1:2:1 ratio for scRNA-seq analysis (2 biological replicates, pooled from 3-4 mice each, per group) at 19 d after tumour inoculation. Dot plot showing the differentially expressed marker genes for 4 subclusters of CD8⁺ T cells in the MC38 tumours (e; see Methods for details). UMAP embeddings of CD8⁺ T cells grouped by genotype (f, left) and indicated subclusters (f, right). Frequencies of the indicated subclusters were quantified for each genotype (g). T_eff-like, effector-like CD8⁺ T cells; T_ex-like, exhaustion-like CD8⁺ T cells; T_cm-like, central memory-like CD8⁺ T cells; T_em-like, effector/memory-like CD8⁺ T cells. (h) Flow cytometry analysis and quantification of the percentages of CD44^hiCD62L^lo intratumoural CD8⁺ T cells from WT and Foxp3^CreCcdc101^fl/fl mice (n ≥ 5 per group). (i) Flow cytometry analysis and quantification of IFN-γ⁺ and TNF-α⁺ cells among intratumoural CD8⁺ T cells from WT and Foxp3^CreCcdc101^fl/fl mice (n ≥ 5 per group). (j) Violin plots of scRNA-seq data depicting Icos, Tnfrsf18, Ctla4 and Ifng expression in intratumoural T_reg cells. (k) Flow cytometry analysis and quantification of ICOS, GITR and CTLA-4 for intratumoural T_reg cells from WT and Foxp3^CreCcdc101^fl/fl mice (n ≥ 5 per group). (l) Flow cytometry analysis and quantification of IFN-γ expression in intratumoural T_reg cells from WT and Foxp3^CreCcdc101^fl/fl mice (n ≥ 5 per group). Mean ± s.e.m. (a–c, h, i, k, l). *P < 0.05; **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (a–c, h, i, k, l); Two-sided Wilcoxon rank sum test in j. Data are representative of one (e–l) or two (d), or pooled from two (a–c) experiments.

Extended Data Figure 12 (to Figure 5).. Schematic of applying CRISPR screening and integrative analyses to dissect nutrient and mTORC1 signaling in primary T cells.

With two rounds of genome-wide and focused CRISPR screenings, we identify 346 high-confidence mTORC1 signaling factors, including many novel activators and inhibitors, as well as known regulators (identified in other systems) that have not been studied in primary T cells. Notably, using analysis of protein–protein interaction (PPI) networks and unbiased functional and proteomic approaches, we further establish the epigenetic and posttranslational mechanisms underpinning the three-tier regulatory modules of nutrient signaling, composed of nutrient transporters (e.g. via affecting expression of Glut1 and other transporters by SAGA complex), sensors (e.g. via epigenetic regulation of Castor1 expression by SWI/SNF complex) and transducers (e.g. via shaping GATOR2 complex stability by Sec31a; regulating Sec13 ubiquitination at lysine 260), which transmit immunological and nutrient cues to mTORC1 signaling for proper regulation of T cell activity in vivo and in vitro.

Figure 1.. Genome-wide CRISPR screening uncovers mTORC1 regulatory networks in T_reg cells.

(a) Genome-wide CRISPR screening approach. (b, c) Fold change (FC)/FC plot for primary (b) or secondary CRISPR screening (n = 6 sgRNAs per gene) (c). Positive (red) and negative (green) regulators. (d) Functional enrichment plot of genes validated by secondary CRISPR screening. (e) Immunoblot analysis and quantification of relative p-S6K1 and p-S6 in TCR-stimulated sgNTC- and sgSmarcb1-transduced cells after amino acid (AA) stimulation (n = 3 samples each group). (f) Gene expression profiles in sgNTC- or sgSmarcb1-transduced cells at 20 h of TCR stimulation (n = 4 samples each group). (g) Cells expressing indicated sgRNAs were stimulated with TCR for 3 h. Quantification of relative p-S6 level and immunoblot analysis of Smarcb1 and Castor1 expression (n = 3 samples each group). Mean ± s.e.m. (e, g). ***P < 0.001; one-way ANOVA (e, g). Data are representative of one (b–d, f) or two (g), or pooled from three (e) experiments.

Figure 2.. Sec31a is crucial for nutrient and GATOR2-dependent mTORC1 activation and the abundance of Sec13.

(a) Interaction of endogenous Sec31a with indicated proteins in induced T_reg cells. (b) Images and quantification of lysosome-associated mTOR [calculated as mean fluorescence intensity (MFI)] for sgNTC- and sgSec31a-transduced cells after amino acid (AA) starvation and refeeding (n ≥ 700 cells per condition). (c) Naïve SMARTA–Cas9 T cells (Ametrine⁺CD45.1⁺) from sgNTC or sgSec31a-transduced ‘retrogenic’ mice were co-transferred with ‘spike’ (Ametrine⁻CD45.1⁺) naïve SMARTA cells (1:1 ratio) into naïve (CD45.2⁺) mice, followed by LCMV infection. Flow cytometry analysis and quantification of relative percentage of indicated cells (n = 5 mice per group). (d) Immunoblot analysis and quantification of indicated protein expression in indicated sgRNA-transduced cells (n = 3 samples each group). (e) sgNTC- or sgSec31a (both Ametrine⁺)-transduced cells were co-transduced with Sec13-overexpressing or empty vector (both GFP⁺) retrovirus. Immunoblot analysis of indicated proteins, and quantification of relative p-S6 level (n = 3 samples each group). Mean ± s.e.m. (b–e). NS, not significant; **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (c) or one-way ANOVA (b, d, e). Data are representative of two (b, c) or five (a), or pooled from three (d, e) experiments.

Figure 3.. Sec31a protects Sec13 from Skp1-mediated proteasomal degradation.

(a) Immunoblot analysis of indicated protein expression in sgNTC- or sgSec31a-transduced cells treated with MG132 or DMSO for 9 h. Lower, quantification of relative Sec13 expression (n = 3 samples each group). (b) Control (sgNTC)- or Sec31a-null HEK293T expressing HA-tagged WT or K260R mutant Sec13 were treated with cycloheximide (CHX) for indicated times. Immunoblot analysis of HA and Hsp90 and quantification of relative Sec13 abundance (n = 3 samples each group). (c) T cells expressing HA-tagged WT or K260R mutant Sec13 were stimulated with α-CD3/CD28 for 48 or 72 h (with MG132 treatment for the last 6 h), followed by immunoprecipitation with anti-HA antibody. Immunoblot analysis for HA and β-Actin. (d) Indicated sgRNA-transduced cells were mixed with sgNTC (mCherry⁺; ‘spike’)-transduced cells, and stimulated with TCR for 3 h. Quantification of relative p-S6 level (n = 3 samples each group). (e) Immunoblot analysis of indicated protein expression and quantification of relative Sec13 abundance in indicated sgRNA-transduced cells (n = 4 samples each group). (f) Quantification of relative proportion of donor-derived (CD45.1⁺) T cells in the spleen of LCMV-infected mice (CD45.2⁺) at 7 d post-infection. See also Extended Data Fig. 7k (n = 10 mice per group). Mean ± s.e.m. (a, b, d–f). NS, not significant; ***P < 0.001; one-way ANOVA (a, d–f); two-way ANOVA (b). Data are representative of two (b–d) or three (e), or pooled from two (f) or three (a) experiments.

Figure 4.. SAGA complex suppresses nutrient transporter expression and mTORC1 activation.

(a) Immunoblot analysis and quantification of relative p-S6K1 and p-S6 in sgNTC- or sgCcdc101-transduced cells after glucose stimulation (n = 4 samples each group). (b) Slc2a1, Slc16a10 or Slc43a1 mRNA expression in sgNTC- and sgCcdc101-transduced cells after α-CD3/CD28 stimulation for 20 h (n = 3 samples each group). (c) Indicated sgRNA-transduced cells were stimulated with TCR for 3 h to quantify relative p-S6 level. Lower, immunoblot analysis of Ccdc101 and Glut1 expression (n = 3 samples each group). (d) sgNTC- or sgCcdc101 (both Ametrine⁺)-transduced cells were co-transduced with sgRraga or sgSec13 (both GFP⁺), mixed with sgNTC (mCherry⁺; ‘spike’)-transduced cells, and stimulated with TCR for 3 h to examine relative p-S6 level (n = 3 samples each group). (e) Immunoblot analysis of indicated protein expression in induced T_reg cells transduced with indicated retrovirus. (f) Quantification of relative p-S6 level in freshly-isolated naïve (0 h) or α-CD3/CD28-stimulated (20 h) WT and Ccdc101-deficient CD4⁺ T cells (n = 4 mice per group). (g) Flow cytometry analysis and quantification of frequencies of effector/memory (CD44^hiCD62L^lo) subsets in splenic CD4⁺Foxp3⁻ and CD8⁺ T cells from indicated mice (n = 4 mice per group). (h, i) Quantification of relative Foxp3 expression (h) or p-S6 levels (i) in splenic CD4⁺Foxp3⁺ T_reg cells from indicated mice (n = 4 mice per group). Mean ± s.e.m. (a–d, f–i). NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (b, f–i) or one-way ANOVA (a, c, d). Data are representative of two (b–e), or pooled from two (a) or three (f–i) experiments.

Figure 5.. Steady state and tumour challenge phenotypes of Foxp3^CreCcdc101^fl/fl mice.

(a) Quantification of IL-2⁺ or IFN-γ⁺ populations of splenic CD4⁺Foxp3⁻ and CD8⁺ T cells from indicated mice (~8 weeks old) (n = 4 mice each group). (b) H&E staining of indicated tissues from indicated mice (> 4 months old) (n = 5 mice each group). Arrows indicate various inflammatory features. (c–f) Indicated mice were inoculated with MC38 colon adenocarcinoma cells. (c) Tumour size and end point tumour weight (n ≥ 5 mice per group). (d) Intratumoural T_reg cells, non-T_reg immune cells, and myeloid cells were sorted at 19 d after tumour challenge, followed by scRNA-seq analysis (see Methods) (n = 2 replicates). UMAP embeddings of CD45⁺ immune cells grouped by genotype (left) and indicated immune cell subclusters (middle). Right, frequencies of indicated immune cell subclusters. (e, f) Quantification of the percentage of tumour-infiltrated CD8⁺ T (e) or T_reg cells (f, left), and Foxp3 mean fluorescence intensity (MFI; gated on CD4⁺Foxp3⁺ cells; f, right) at 19 d after tumour challenge (n ≥ 5 mice per group). Mean ± s.e.m. (a, c, e, f); **P < 0.01; ***P < 0.001; two-tailed unpaired Student’s t-test (a, c, lower; e, f); two-way ANOVA (c, upper). Data are representative of one (b, d–f) or two (a, c) experiments.