Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell...

26
immunology.sciencemag.org/cgi/content/full/3/25/eaas9818/DC1 Supplementary Materials for Metabolic signaling directs the reciprocal lineage decisions of αβ and γδ T cells Kai Yang, Daniel Bastardo Blanco, Xiang Chen, Pradyot Dash, Geoffrey Neale, Celeste Rosencrance, John Easton, Wenan Chen, Changde Cheng, Yogesh Dhungana, Anil KC, Walid Awad, Xi Zhi J. Guo, Paul G. Thomas, Hongbo Chi* *Corresponding author. Email: [email protected] Published 6 July 2018, Sci. Immunol. 3, eaas9818 (2018) DOI: 10.1126/sciimmunol.aas9818 The PDF file includes: Materials and Methods Fig. S1. Dynamic regulation of nutrient transporters and cell growth in developing thymocytes. Fig. S2. mTOR signaling is required for thymocyte development. Fig. S3. RAPTOR deficiency diminishes cell growth, proliferation, and nutrient transporter expression of ISP cells, but not cell survival. Fig. S4. RAPTOR controls metabolic gene expression programs in thymocytes. Fig. S5. RAPTOR deficiency affects γδ T cell development. Fig. S6. Loss of RHEB or RICTOR does not affect the development of γδ T cells. Fig. S7. SCAP and HIF1α are dispensable for T cell development. Fig. S8. MYC is implicated in αβ and γδ T cell development. Fig. S9. RAPTOR-deficient DN3 cells exhibit increased ROS but normal mitochondrial mass and membrane potential. Fig. S10. Modulation of ROS production alters fate choices of DN3 cells. Fig. S11. mTORC1 activation integrates signals from pre-TCR and NOTCH. Fig. S12. RAPTOR deficiency alters transcriptional programs and signaling and metabolic pathways. Fig. S13. Monocle pseudotime trajectory of single-cell transcriptomics data. Fig. S14. Loss of RAPTOR or MYC enhances signal strength. Fig. S15. Schematics of mTORC1 and metabolic control of redox homeostasis and signal strength in T cell lineage choices. Reference (54) Other Supplementary Material for this manuscript includes the following: (available at immunology.sciencemag.org/cgi/content/full/3/25/eaas9818/DC1) Table S1 (Microsoft Excel format). Raw data sets.

Transcript of Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell...

Page 1: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

immunology.sciencemag.org/cgi/content/full/3/25/eaas9818/DC1

Supplementary Materials for

Metabolic signaling directs the reciprocal lineage decisions of αβ and γδ T cells

Kai Yang, Daniel Bastardo Blanco, Xiang Chen, Pradyot Dash, Geoffrey Neale, Celeste Rosencrance, John Easton,

Wenan Chen, Changde Cheng, Yogesh Dhungana, Anil KC, Walid Awad, Xi Zhi J. Guo, Paul G. Thomas, Hongbo Chi*

*Corresponding author. Email: [email protected]

Published 6 July 2018, Sci. Immunol. 3, eaas9818 (2018)

DOI: 10.1126/sciimmunol.aas9818

The PDF file includes:

Materials and Methods

Fig. S1. Dynamic regulation of nutrient transporters and cell growth in developing thymocytes.

Fig. S2. mTOR signaling is required for thymocyte development.

Fig. S3. RAPTOR deficiency diminishes cell growth, proliferation, and nutrient transporter

expression of ISP cells, but not cell survival.

Fig. S4. RAPTOR controls metabolic gene expression programs in thymocytes.

Fig. S5. RAPTOR deficiency affects γδ T cell development.

Fig. S6. Loss of RHEB or RICTOR does not affect the development of γδ T cells.

Fig. S7. SCAP and HIF1α are dispensable for T cell development.

Fig. S8. MYC is implicated in αβ and γδ T cell development.

Fig. S9. RAPTOR-deficient DN3 cells exhibit increased ROS but normal mitochondrial mass

and membrane potential.

Fig. S10. Modulation of ROS production alters fate choices of DN3 cells.

Fig. S11. mTORC1 activation integrates signals from pre-TCR and NOTCH.

Fig. S12. RAPTOR deficiency alters transcriptional programs and signaling and metabolic

pathways.

Fig. S13. Monocle pseudotime trajectory of single-cell transcriptomics data.

Fig. S14. Loss of RAPTOR or MYC enhances signal strength.

Fig. S15. Schematics of mTORC1 and metabolic control of redox homeostasis and signal

strength in T cell lineage choices.

Reference (54)

Other Supplementary Material for this manuscript includes the following:

(available at immunology.sciencemag.org/cgi/content/full/3/25/eaas9818/DC1)

Table S1 (Microsoft Excel format). Raw data sets.

Page 2: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

2

Materials and Methods

Flow cytometry

For analysis of surface markers, cells were stained in PBS containing 2% (wt/vol) BSA, with anti-

CD4 (RM4-5), anti-CD8a (53-6.7), anti-TCRb (H57-597), anti-CD25 (PC61.5), anti-CD44

(1M7), anti-CD62L (MEL-14), anti-CD45.1 (A20), anti-CD45.2 (104), anti-TCRgd (eBioGL3),

anti-CD24 (M1/69), anti-CD73 (eBioTY/11.8), anti-CD5 (53-7.3), anti-CD98 (RL388), and anti-

CD71 (R17217). Intracellular Ki67 (SolA15), phospho-S6 (D57.2.2E), and phospho-4E-BP1

(236B4) were analyzed by flow cytometry according to the manufacturer's instructions. Two

antibodies (C-20, Santa Cruz Biotechnology; and D16D10, Cell Signaling Technology) were used

to measure ID3 expression by flow cytometry, with similar results. For measurement of ROS,

mitochondrial mass and potential, thymocytes were incubated in the medium containing 10 µM

CM-H2DCFDA (5-(and-6)-chloromethyl-2,7-dichlorodihydrofluorescein diacetate acetyl ester;

Life Technologies), 10 nM MitoTracker Deep Red (Life Technologies) or 20 nM TMRM

(tetramethyl rhodamine, methyl ester; ImmunoChemistry Technologies), respectively, for 30 min

at 37°C before staining surface markers. BrdU incorporation was measured in thymocytes from

mice injected with BrdU (1 mg/mouse) for 3 h according to manufacturer’s instructions (BD

Biosciences). Caspase-3 activity was measured using active caspase-3 apoptosis kit (BD

Biosciences). Flow cytometry data were acquired on LSRII or LSR Fortessa (Becton Dickinson),

and analyzed using FlowJo software (Tree Star).

Cell purification and cultures

Lineage-negative WT and Rptor–/– DN3 (CD4–CD8–TCRgd–TCRb–CD25hiCD44–) or DN3a cells

(CD4–CD8–TCRgd–TCRb–CD25hiCD44–CD28–) were sorted on a MoFlow (Beckman-Coulter) or

Page 3: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

3

Reflection (i-Cyt). DN3 cells were used for Seahorse assay and in vitro cultures with OP9 or OP9-

DL1 cells in the presence of IL-7 (0.5 ng/ml) and Flt3L (5 ng/ml) for 3 days. For metabolic

perturbation, DN3a cells were co-cultured with OP9-DL1 cells in the presence of rapamycin (10

ng/ml), DCA (10 mM), galactose (0.5 mM) and galactose oxidase (GAO, 0.3 U/ml), GSH (5 mM),

or NAC (20 mM) for 3-5 days. For T cell differentiation under a stromal cell-free culture condition

(36), sorted DN3a cells were cultured in DLL4 (2 µg/ml)-coated plates, in the presence of CXCL12

(100 ng/ml) or IL-7 (1 ng/ml) to promote aβ or gd T cell development, respectively, with or

without the indicated inhibitors for 4 days.

Gene expression profiling by microarray analysis

RNA samples from freshly isolated ISP (WT, n = 4; Rptor–/–, n = 3), DN3a, or gd T cells (WT, n

= 3; Rptor–/–, n = 3) were analyzed with the Affymetrix Mouse Gene 2.0 ST GeneChip array. Gene

set enrichment analysis (GSEA) within hallmark gene sets (MSigDB) or canonical pathways

(MSigDB c2.CP gene sets for canonical pathways) was performed as described (22). Differentially

expressed transcripts were identified by ANOVA (Partek Genomics Suite v6.5) and the Benjamini-

Hochberg method was used to estimate the false discovery rate (FDR) as described previously

(51). Lists of differentially expressed genes by > 0.5 log2 difference or more were analyzed using

Ingenuity Pathway Analysis (Qiagen, Inc.) for upstream transcription regulators.

RNA and immunoblot analysis

Real-time PCR analysis was performed with primers and probe sets from Applied Biosystems as

described (51). Immunoblots were performed as described (54) using the following antibodies:

EGR1 (44D5), RAPTOR (24C12), p-ERK (9101), ID3 (D16D10), p-p65 (all from Cell Signaling

Page 4: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

4

Technology), β-ACTIN (AC-15; Sigma); ATP5A, UQCRC2, MTCO1, SDHB, and NDUFB8

(using total OXPHOS antibody cocktail from Abcam).

Metabolic assays

OCR and ECAR were measured in XF media (non-buffered DMEM containing 5 mM glucose, 2

mM L-glutamine and 1 mM sodium pyruvate), under basal conditions and in response to 1 µM

oligomycin, 2 µM fluoro-carbonyl cyanide phenylhydrazone (FCCP) and 1 µM Rotenone using

the XF-24 Extracellular Flux Analyzer (Seahorse Bioscience). For glycolytic rate, the various

subsets of thymocytes were cultured in Click’s medium containing 1 µCi [3-3H]-glucose for 6 h.

Glycolytic flux was determined by measuring the detritiation of [3-3H]-glucose, as we previously

described (28).

Statistical analysis for biological experiments

P values were calculated by two-tailed Mann-Whitney test, two-tailed unpaired Student’s t test, or

ANOVA using GraphPad Prism, unless otherwise noted. P value of < 0.05 was considered

significant. All error bars represent the SEM.

Single-cell multiplex RT-PCR and sequencing of TCRγ chains

Thymocytes from WT or Rptor–/– mice were harvested and resuspended. After depletion of CD4+

cells using CD4 (L3T4) MicroBeads (Miltenyi Biotec), the remaining cells were stained with anti-

CD4, anti-CD8, anti-NK1.1, anti-TCRb, anti-TCRgd, anti-Ter119, anti-B220 and anti-Gr1 in 200

µl of sorting buffer (PBS containing 0.1% BSA, Fraction V (Gibco)) on ice, in dark for 20 min.

Following staining, cells were washed three times and resuspended in sorting buffer (containing

Page 5: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

5

200 units of RNAsin/ml (Promega)) and filtered through 40 µM cell strainer prior to sorting. γδ T

cells (TCRgd+CD4–CD8–NK1.1–TCRb–B220–Ter119–Gr1–) were sorted singly into each well of a

384-well plate (Eppendorf) using an iCyt Synergy cell sorter (Sony) with the following parameters:

Multi-drop sort OFF, Multi-drop exclude OFF, Division 10, Center sort %: 90. At least one column

of the 384-well plates was left unsorted to use as negative controls for PCR. After sorting, the

plates were sealed using plate sealer film (MicroAmp, Applied Biosystems) and centrifuged at 500

g for 3 min and stored at -80°C until further amplification. Unbiased amplification of TCRγ chain

at the single cell level was carried out as per method described before (24). Briefly, the cDNA

synthesis was performed directly from single cells without any RNA extraction step using the

SuperScript VILO cDNA synthesis kit (Thermo Scientific) in a volume of 1 µl of reaction mix

(0.2 µl of 5´ VILO reaction mix, 0.1 µl of 10´ SuperScript® Enzyme mix, 0.1 µl of 1% Triton X-

100 (Sigma)). Following reverse transcription, a multiplex nested PCR was carried out using a Taq

polymerase based PCR kit (Qiagen). The protocol for the first and second round PCR (except for

the primer components), sequencing, and data analysis were performed as previously described

(24, 25). The primers for the mouse TCRγ chain were used as below.

Name External Internal

TRGV1-3 For GCAGCTGGAGCAAACTG CTGAATTATCGGTCACCAG

TRGV4 For CAAATATCCTGTAAAGTTTTCATC GTTTAGAGTTTCTATTATATGTCCTTGCAAC

TRGV5 For GATATCTCAGGATCAGCTCTCC TACCCGAAGACCAAACAAGAC

TRGV6 For TCACCTCTGGGGTCATATG AGAGGAAAGGAAATACGGC

TRGV7 For CAACTTGGAAGAAAGAATAATGTC CACCAAGCTAGAGGGGTC

TRGC Rev CTTTTCTTTCCAATACACCC TCDGGAAAGAACTTTTCAAGG

Single-cell RNA sequencing (scRNA-seq)

Page 6: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

6

Lineage-negative thymocytes (Ter119–Gr1–CD19–NK1.1–CD4–CD8–TCRb–TCRgd–CD44–;

representative of DN3-DN4 cells) were sorted into culture medium, spun down, and re-suspended

in 100 µl of 1´ PBS + 0.04% BSA. The cells were counted and examined for viability using a

Luna Dual Florescence Cell Counter (Logos Biosystems). Single-cell suspensions were loaded

onto the Chromium Controller according to their respective cell counts to generate 6000 single-

cell GEMs (gel beads in emulsion) per sample. Each sample was loaded into a separate channel.

Libraries were prepared using the Chromium Single Cell 3’ v2 Library and Gel Bead Kit (10x

Genomics). The cDNA content of each sample after cDNA amplification of 12 cycles was

quantified and quality checked using a High-Sensitivity DNA chip with a 2100 Bioanalyzer

(Agilent Technologies) to determine the number of PCR amplification cycles to yield sufficient

library for sequencing. After library quantification and quality check by DNA 1000 chip (Agilent

Technologies), samples were diluted to 3.5 nM for loading onto the HiSeq 4000 (Illumina) with a

2 ´ 75 Paired-end kit using the following read length: 26 bp Read1, 8 bp i7 Index, and 98 bp

Read2. An average of 400,000,000 reads per sample was obtained (~approximately 80,000 reads

per cell).

Analysis for scRNA-seq data

Alignment, barcode assignment, and UMI counting

The Cell Ranger 1.3 Single-Cell Software Suite (10x Genomics) was implemented to process the

raw sequencing data from the Illumina HiSeq run. This pipeline performed de-multiplexing,

alignment (mm10), and barcode processing to generate gene-cell matrices used for downstream

analysis. Specifically, data for 2 WT and 3 Rptor–/– samples were combined, and UMIs mapped to

genes encoding ribosomal proteins were removed. Cells with low (bottom 2.5%, potentially dead

Page 7: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

7

cells with broken membrane) or high (top 2.5%, potentially two or more cells in a single droplet)

UMI counts were filtered. On average, 4,123 mRNA molecules (UMIs) were captured in a cell

(median: 3,448, range: 1,703–9,999), representing 2,110 genes (median: 1,968, range: 59–3,916).

The expression level of each gene is normalized to 10,000 UMIs per cell and log transformed by

adding 0.25 to the expression matrix.

Clustering

The subpopulation structure of the whole dataset was inferred using Latent Cellular State Analysis

(LCA), a novel clustering algorithm developed in house for analyzing large-scale scRNA-seq data

(manuscript in preparation). Briefly, LCA first used singular value decomposition (SVD) to derive

latent cellular states from the expression matrix for individual cells. The number of significant

cellular states was determined using the Tracy-Widom test on eigenvalues. A modified version of

spectral clustering was performed on the significant cellular states of individual cells (cellular

states explained by total UMIs ignored) with different number of clusters (2-30). The optimal

number of clusters was manually selected from top models determined by the silhouette measure

for solutions with different number of clusters.

Data visualization

Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor

embedding (tSNE). Expression of individual genes or pathway scores were color coded (gray: not

expressed; from low to high: blue-green-yellow-red) for each cell on tSNE plots.

Generation of gene signatures and statistical analysis for over-representation

Page 8: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

8

We generated DN3a signature from the genes specifically expressed in DN3a but not in DN3b,

DN3b-DN4 or DN4 cells (92 genes with FDR < 0.05, log2 fold change > 1.5 for each comparison)

(45) (GSE15907). The pre-TCR signaling signature was generated from genes upregulated in

DN3b cells compared with DN3a cells (75 genes with FDR < 0.05, log2 fold change > 2)

(GSE15907). The gd T cell gene signature was generated from genes selectively expressed by gd

T cells (46) (121 genes upregulated in gd T cells, but not DN3b cells, as compared with DN3a

cells, with FDR < 0.05, log2 fold change > 2). For gene signatures of metabolic and signaling

pathways, the list of 2,088 genes with strong and dynamic expression in thymocyte differentiation

(45) was used to intersect with gene sets from MSigDB (hallmark mTORC1 signaling, hallmark

MYC targets v1, hallmark glycolysis, KEGG NOTCH signaling pathway, and PDGF ERK DN.V1

DN) for refined gene signatures relevant to thymocyte development. For scRNA-seq analysis,

genes from these signatures were retained if they were expressed in at least 10% of cells in any

cluster. Gene signature activity was calculated as the average value for all retained genes. Relative

over-representation of gene signatures for a specific cluster (Fig. 6D) was determined using one-

tailed Mann-Whitney test (cells in a specific cluster vs all other cells). Differential activity of a

gene signature or individual gene expression between defined clusters was determined using two-

tailed Mann-Whitney test.

Pseudo-temporal analysis

Monocle 2 (47) was used to generate the single cell pseudo-temporal trajectories with default

settings.

Statistical analysis for WT and Rptor–/– enrichment for individual clusters

Page 9: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

9

The numbers of cells in and outside a specific cluster were counted for each mouse. Logistic

regression was used to assess the association between the cluster cell counts and the independent

predictor of WT and Rptor–/– samples (Fig. 6E).

Page 10: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

10

Fig. S1

Fig. S1. Dynamic regulation of nutrient transporters and cell growth in developing thymocytes. (A) Diagram of T cell development in the thymus. Hematopoietic stem cells (HSC) derived from the bone marrow migrate to the thymus and progress through well-defined developmental stages to become ab or gd T cells. Double-negative cells (DN, without expression of CD4 and CD8) at early thymocyte development can be divided into four stages, with the DN3 stage being a crucial checkpoint for T cell lineage choices. DN3 cells with successful b-selection continue their developmental process by transitioning through the DN4, ISP (immature single-positive) and DP (double-positive) stages before becoming CD4 or CD8 single-positive cells (SP). Alternatively, DN3 cells complete gd-selection to become gd T cells, followed by induction of CD73 expression to mark fully committed gd T cells. (B and C) Relative expression of CD98 (B) and CD71 (C) (MFI in DN3 cells is set to 1) in thymocyte subsets. (D) Plot of cell size (FSC) in thymocyte subsets. Data are representative of at least five independent experiments (B to D).

D

FSC0 50K 100K 150K 200K 250K

0

20

40

60

80

100 DN3DN4ISPDP

1.19e51.3e51.34e597852

0

2

4

6

8

Rel

ativ

e C

D71

DN4

DN3a

DN3b

DP

DN3DN4ISPDP

0

2

4

6

8CB

0

1

2

3

Rel

ativ

e C

D98

0

1

2

3

DN4

DN3a

DN3b

DP

DN3DN4ISPDP

Cel

ls

HSCT lineage

commitmentDN1 DN2

DN3 (DN3a and DN3b)

gd T cells Committed gd T cells (CD73+)

DN4 ISP DP

CD4 SP

CD8 SP⍺b lineage

gd lineage

Lineage choices

A

Page 11: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

11

Fig. S2

Fig. S2. mTOR signaling is required for thymocyte development. (A and B) Analysis of RAPTOR mRNA (A) and protein (B) expression in thymocyte subsets from WT and Rptor–/– mice. (C) Frequencies of thymocyte subsets from WT and Rptor–/– mice. Each symbol represents an individual mouse (WT, n = 7; Rptor–/–, n = 6). (D) Cellularity of CD4+ and CD8+ T cells in the spleen of WT and Rptor–/– mice. Each symbol represents an individual mouse (n = 3 each group). Data are means ± SEM. ns, not significant; *P < 0.05, **P < 0.005, and ***P < 0.0001. One-way ANOVA with Tukey’s test (C), or two-tailed unpaired t test (D). Data are representative of at least three independent experiments (A to D).

A

DN3 DN4 ISP DP

Rpt

or m

RN

A WT Rptor–/–

0 0.5 1.0 1.5 2.0 B

WT

Rpt

or–/

– W

T R

ptor

–/–

DN3 DN4

b-ACTIN RAPTOR

D

0246810

Cel

l num

ber (×1

06 )

** *

0 2 4 6 8

10 WT Rptor–/–

CD4+ CD8+

Thym

ocyt

es (%

)

DN DP CD8 SP

CD4 SP

***

***

ns ***

0 20 40 60 80

100 C WT Rptor–/–

Page 12: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

12

Fig. S3

Fig. S3. RAPTOR deficiency diminishes differentiation, proliferation, and nutrient transporter expression of ISP cells, but not cell survival. (A) ISP cells from WT and Rptor–

/– mice were cultured with OP9-DL1 cells for 1 day, followed by analysis of CD4 and CD8 (left panel) and CD24 and CD8 expression on CD4–CD8+ ISP cells (right panel). (B) Caspase-3 activity in thymocyte subsets in WT and Rptor–/– mice. (C) BrdU incorporation in thymocyte subsets in WT and Rptor–/– mice pulsed with BrdU for 3 h. (D) Gating strategy for DN3a (Lin–CD4–CD8–CD25+CD44–CD27–) and DN3b (Lin–CD4–CD8–CD25+CD44–

CD27+) cells. (E) Relative phosphorylation of S6 and 4E-BP1 (MFI in DN3a cells is set to 1) in DN3a and DN3b cells from WT mice. (F and G) Plots of FSC, CD71 and CD98 (F), and icTCRb (G) in thymocyte subsets in WT and Rptor–/– mice. Data are means ± SEM. Data are representative of at least three independent experiments (A to G). Numbers indicate percentage of cells in gates.

C

BrdU

Cel

ls

DN4DN3 ISP

100

101

102

103

104

0

50

100

150

200

250

5.18

100

101

102

103

104

0

20

40

60

80

100

4.98

100

101

102

103

104

0

50

100

150

200

250

8.91

100

101

102

103

104

0

30

60

90

120

12.1

100

101

102

103

104

0

20

40

60

80

100

10.6

100

101

102

103

104

0

5

10

15

20

39

WT

Rptor–/–

DP

100

101

102

103

104

0

1000

2000

3000

0.67

100

101

102

103

104

0

2000

4000

6000

8000

10K

3.95

B

CD4

DN3 DN4 DPISP

WT

Cas

pase

-3

CD8

Rptor–/–

0 102 103 104 105

0

102

103

104

105

0.779

0 102 103 104 105

0

102

103

104

105

0.861

0 102 103 104 105

0

102

103

104

105

1.46

0 102 103 104 105

0

102

103

104

105

0.849

0 102 103 104 105

0

102

103

104

105

1.4

0 102 103 104 105

0

102

103

104

105

1.14

0 102 103 104 105

0

102

103

104

105

0.469

0 102 103 104 105

0

102

103

104

105

0.472

F

0 102

103

104

105

0

20

40

60

80

100

0 102

103

104

105

0

20

40

60

80

100

0 102

103

104

105

0

20

40

60

80

100

CD98

0 102

103

104

105

0

20

40

60

80

100

0 102

103

104

105

0

20

40

60

80

100

CD71C

ells

100

101

102

103

104

0

20

40

60

80

100

0 50K 100K 150K 200K 250K

0

20

40

60

80

100

0 50K 100K 150K 200K 250K

0

20

40

60

80

100

0 50K 100K 150K 200K 250K

0

20

40

60

80

100

FSC

DN3a DN4DN3b

0 102

103

104

105

0

20

40

60

80

100

0 102

103

104

105

0

20

40

60

80

100

0 50K 100K 150K 200K 250K

0

20

40

60

80

100

ISPWT Rptor–/–

GDN3a DN4DN3b

icTCRb

Cel

ls

ISPWT Rptor–/–

0 102 103 104 1050

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

D

CD

44

CD27

FSC

DN3b

DN3a

CD25

DN3

Lin–CD4–CD8–E

0

2

4

6

0

1

2

3DN3aDN3a

DN3b

DN3bDN3a

DN3b

Rel

ativ

e p-

S6

Rel

ativ

e p-

4E-B

P1

0

2

4

6

0

1

2

3

CD4

CD

8

Rptor–/–WT

0 102 103 104 105

0

102

103

104

105

0.581

26.2

0.894

60.7

0 102 103 104 105

0

102

103

104

105

0.957

11.8

0.811

75.7

CD24

CD

8

Rptor–/–WT

0 102 103 104 105

0

102

103

104

105

98.8

1.16

0 102 103 104 105

0

102

103

104

105

98.4

1.42

A

Page 13: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

13

Fig. S4

Fig. S4. RAPTOR controls metabolic gene expression programs in thymocytes. (A) List of the top 10 downregulated gene sets in the databases of hallmarks and canonical pathways (MSigDB) in Rptor–/– ISP cells, as identified by gene set enrichment assay (GSEA) of microarray profiles (WT, n = 4; Rptor–/–, n = 3). NES, normalized enrichment score; FDR, false discovery rate. No significantly upregulated gene sets (FDR < 0.05) were identified in these databases. (B) Analysis of Idi1, Hmgcs1 and Sqle mRNA expression in WT and Rptor–/– ISP cells. (C) List of the top 10 potential transcription regulators downregulated in Rptor–/– ISP cells as identified by Ingenuity Pathway Analysis (IPA) of microarray profiles (WT, n = 4; Rptor–/–, n = 3). Data are from one experiment (A, C), or representative of three independent experiments (B).

Hallmark gene sets NES FDR q-valHALLMARK_MTORC1_SIGNALING -2.704 <0.001HALLMARK_MYC_TARGETS_V1 -2.608 <0.001HALLMARK_MYC_TARGETS_V2 -2.476 <0.001HALLMARK_OXIDATIVE_PHOSPHORYLATION -2.256 <0.001HALLMARK_TNFA_SIGNALING_VIA_NFKB -2.231 <0.001HALLMARK_CHOLESTEROL_HOMEOSTASIS -2.174 <0.001HALLMARK_UNFOLDED_PROTEIN_RESPONSE -1.885 5.43E-04HALLMARK_UV_RESPONSE_UP -1.874 4.75E-04HALLMARK_FATTY_ACID_METABOLISM -1.801 0.001095HALLMARK_E2F_TARGETS -1.689 0.004129Canonical pathway gene sets NES FDR q-valREACTOME_M_G1_TRANSITION -2.470 <0.001REACTOME_AMYLOIDS -2.451 <0.001REACTOME_RNA_POL_I_PROMOTER_OPENING -2.409 <0.001REACTOME_INFLUENZA_LIFE_CYCLE -2.380 <0.001REACTOME_CHOLESTEROL_BIOSYNTHESIS -2.360 <0.001REACTOME_RESPIRATORY_ELECTRON_TRANSPORT_ATP_SYNTHESIS_BY_CHEMIOSMOTIC_COUPLING_AND_HEAT_PRODUCTION_BY_UNCOUPLING_PROTEINS_ -2.347 <0.001REACTOME_SYNTHESIS_OF_DNA -2.344 <0.001REACTOME_S_PHASE -2.334 <0.001REACTOME_TCA_CYCLE_AND_RESPIRATORY_ELECTRON_TRANSPORT -2.298 <0.001REACTOME_ASSEMBLY_OF_THE_PRE_REPLICATIVE_COMPLEX -2.296 <0.001

A

B

0.0

0.5

1.0

1.5

Idi1mRNA

0.0

0.5

1.0

1.5

Hmgcs1

mRNA

0.0

0.5

1.0

1.5

SqlemRNA

WTRptor–/–

0

0.5

1.0

1.5

0

0.5

1.0

1.5

0

0.5

1.0

1.5

CUpstream regulators P-valueMYC 1.23E-20FOXO1 2.79E-14CREB1 3.66E-14SREBF2 5.16E-13SREBF1 1.37E-11TRIM24 1.93E-09MYCN 4.42E-06PPARGC1B 1.71E-05HNF4A 2.11E-05ATF4 2.65E-05

Page 14: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

14

Fig. S5

ATC

Rgd

+ /TC

Rb+

0.00

0.05

0.10

0.15

0.20

00.050.10

0.15

0.20***

WT Rptor–/–

E

I

K

J

TCRgd0 102 103 104 105

0

102

103

104

105

0.0859

0 102 103 104 105

0

102

103

104

105

0.115

Cas

pase

-3

WT Rptor–/–F

CD45.2+ Rptor–/–CD45.2+ WT WT/CD45.1+ Rptor–/–/CD45.1+

CD4

TCRgd

776 656 745 785

23781 905 33965 29973

CD45.2+ WTCD45.2+ Rptor–/–

14403/31904/12727/11925

Cel

ls

WT/CD45.1+

Rptor–/–/CD45.1+

CD73

0

20

40

60

Vg1.

1

WT Rptor–/–

Vd6.30 102 103 104 105

0

102

103

104

105 27.3 2.33

68.8 1.560 102 103 104 105

0

102

103

104

105 40.1 3.07

55.2 1.58

WT Rptor–/–

ns

Vg1.

1+(%

)

0

20

40

60

0

20

40

60

80L ns

Vg2

Vd40 102 103 104 105

0

102

103

104

105 28.3 10.9

53.3 7.540 102 103 104 105

0

102

103

104

105 26.3 10.5

10.852.4

WT Rptor–/–

Vg2+

(%)

WT Rptor–/–0

20

40

80

60

Freq

uenc

y (%

)

Vg1 Vg2 Vg3 Vg4 Vg5 Vg6 Vg7

Thymus

TRGV1

TRGV2

TRGV3

TRGV4

TRGV5

TRGV6

TRGV70

20

40

60

80

Frequency (

%) WT

KORptor–/–

Thymus

TRGV1

TRGV2

TRGV3

TRGV4

TRGV5

TRGV6

TRGV70

20

40

60

80

Frequency (

%) WT

KOWT

Thymus

TRGV1

TRGV2

TRGV3

TRGV4

TRGV5

TRGV6

TRGV70

20

40

60

80

Frequency

(%) WT

KO

0

20

40

80

60

CD40 102 103 104 105

0

102

103

104

105

66.5

3.09

0 102 103 104 105

0

102

103

104

105

23.4

6.19

TCRgd

WT Rptor–/–

CD730 102 103 104 105

0

20

40

60

80

100

Cel

ls

16521/40331

Rptor–/–

WT

0

1

2

3

TCRgd

+(×

103 )

ns

WT Rptor–/–

CD

4+(×

104 )

0

5

10

15

20

25

WT Rptor–/–

***To

tal n

umbe

rs (×

104 )

TCRgd

+ CD

73+

(×10

3 )

0

1

2

3 ***

WT Rptor–/–0

20

40

60 *

WT Rptor–/–0

20

40

60

80

100

WT Rptor–/–

*

CD

4+(%

)

0

1

2

3

4

5

WT Rptor–/–

*

TCRgd+

(%)

0

20

40

60

80

100

WT Rptor–/–

*

TCRgd

+ CD

73+

(%)

0

20

40

60

02040

8060

100

02040

8060

100

012345

0

1

2

3

0

1

2

3

05

10

2015

25

0.0

0.5

1.0

1.5

Rpt

or m

RN

A0

0.5

1.0

1.5

B

WT Rptor–/–

C

FSC

Cel

ls

CD98CD710 200 400 600 800 1000

0

20

40

60

80

100

385/341

100 101 102 103 1040

20

40

60

80

100

82.1/44.4

100 101 102 103 1040

20

40

60

80

100

500/363D

WT Rptor–/–

Ki67

Cel

ls

100 101 102 103 1040

3

6

9

12

80.4

100 101 102 103 1040

5

10

15

20

25

54.8

Rptor–/–WT

G H

Page 15: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

15

Fig. S5. RAPTOR deficiency affects gd T cell development. (A) Ratio of TCRgd+ to TCRb+ cells in WT and Rptor–/– mice. Each symbol represents an individual mouse (n = 8 each group). (B) Analysis of Rptor mRNA in WT and Rptor–/– TCRgd+ cells (Rptor mRNA in WT is normalized to 1). (C) Plots of FSC, CD71 and CD98 in WT and Rptor–/– TCRgd+ cells, with MFI plotted above graphs. (D) Plot of Ki67 expression in WT and Rptor–/– TCRgd+ cells. Numbers indicate the frequency of the Ki67+ population. (E) Caspase-3 activity of WT and Rptor–/– TCRgd+ cells. (F) WT or Rptor–/– cells were cultured with OP9-DL1 cells, followed by examination of CD4+ and TCRgd+ T cells (left), and CD73 expression on TCRgd+ cells with MFI plotted above the graph (right). (G, H) Frequencies (G) and numbers (H) of total cells, CD4+ T cells, TCRgd+ cells, and committed gd T cells (TCRgd+CD73+) in F. (I) Left: WT (CD45.2+) or Rptor–/– (CD45.2+) DN3a cells were cultured with CD45.1+ DN3a cells in the presence of OP9-DL1 cells, followed by examination of TCRgd+ and CD4+ T cells, with cell numbers included in the graphs. Right: Expression of CD73 on TCRgd+ cells with MFI plotted above the graph. (J) Single cell analysis of diverse TCRg chains expressed in thymic TCRgd+ T cells of WT and Rptor–/– mice. The Vg usage derived from the CDR3g sequence data is plotted. (K and L) Flow cytometry of Vg1.1+Vd6.3+ (K) and Vg2+Vd4+ (L) TCRgd+ T cell subsets in the thymus. Right, frequency of Vg1.1+ or Vg2+ cells. Each symbol represents an individual mouse (WT, n = 4; Rptor–/–, n = 3). Data are means ± SEM. *P < 0.05 and ***P < 0.001. Two-tailed unpaired t test (A), paired t test (G), one-tailed unpaired t test (H), or two-tailed Mann-Whitney test (K, L). Data are representative of at least three independent experiments (A, C to I), or two independent experiments (B, J to L). Numbers indicate percentage of cells in the quadrants or gates.

Page 16: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

16

Fig. S6

Fig. S6. Loss of RHEB or RICTOR does not affect the development of gd T cells. (A) Left: Flow cytometry of thymocyte subsets of WT and Rheb–/– mice. Right: Cellularity of total thymocytes. Each symbol represents an individual mouse (n = 5 each group). (B) Flow cytometry of TCRgd+ cells in WT and Rheb–/– mice (left). Frequency (middle) and cellularity (right) of TCRgd+ cells in the thymus of WT and Rheb–/– mice (n = 3 each group). (C) Left: Flow cytometry of TCRgd+ cells in WT, Rptor–/–, Rictor–/– and Rptor–/–Rictor–/– thymocytes. Right: Frequency and cellularity of TCRgd+ cells and total thymocyte cellularity in the indicated mouse groups. Each symbol represents an individual mouse (WT, n = 7; Rptor–/–, n = 5; Rictor–/–, n = 3; Rptor–/–Rictor–/–, n = 7). (D) Flow cytometry of CD73 and CD24 expression on TCRgd+ cells in the thymus from the indicated mice. Data are means ± SEM. *P < 0.01, **P < 0.001, and ***P < 0.0001. Two-tailed unpaired t test (A, B for cellularity analyses), two-tailed Mann-Whitney test (B for frequency analysis), or one-way ANOVA with Tukey’s test (C). Data are representative of two independent experiments (D), or at least three independent experiments (A to C). Numbers indicate percentage of cells in quadrants or gates.

0

50

100

150

200

0.0

0.2

0.4

0.6

0

1

2

3

4

TCRgd

+(%

)

WT Rptor–/– Rptor–/–Rictor–/–Rictor–/–

100 101 102 103 104100

101

102

103

104

0.437

100 101 102 103 104100

101

102

103

104

1.77

100 101 102 103 104100

101

102

103

104

0.354

100 101 102 103 104100

101

102

103

104

3.59

WTRptor–/–

Rictor–/–

Rptor–/–

Rictor–/–

0

1

2

3

4WT

Raptor ko

Rictor ko

DKo

**

***ns

TCRgd

TCRb

C

0 102 103 104 105

0

102

103

104

105

0.237

0 102 103 104 105

0

102

103

104

105

0.433

WT Rheb–/–

TCRb

TCRgd

0.0

0.2

0.4

0.6

TCRgd

+(%

)

WT Rheb–/–

ns

0

0.2

0.4

0.6

Thym

ocyt

es (�

106 )

0

50

100

150

200

0 102 103 104 105

0

102

103

104

105 2.19 88.8

1.9 5.370 102 103 104 105

0

102

103

104

105 2.65 85

3.92 6.29

CD4

CD

8

WT Rheb–/–

A B

0

50

100

150

200

WT Rheb–/–

ns

0.0

0.2

0.4

0.6

0.8

0

0.2

0.4

0.6

0.8

TCRgd

+(�

106 )

WT Rheb–/–

ns

D

TCRgd

+(×

106 ) *

ns

0 102 103 104 105

0

102

103

104

105 72.2 19

6.981.790 102 103 104 105

0

102

103

104

105 72.1 18.4

7.711.78

WT Rictor–/–

0 102 103 104 105

0

102

103

104

105 63.5 32.5

2.571.47

Rptor–/–Rictor–/–

0 102 103 104 105

0

102

103

104

105 52.1 37.5

8.042.36

Rptor–/–

CD73

CD

24

Thym

ocyt

es (×

106 )

***ns

*****

Page 17: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

17

Fig. S7

Fig. S7. SCAP and HIF1a are dispensable for T cell development. (A to D) Analysis of thymocyte development of WT and Scap–/– mice. (A) Flow cytometry (left) of thymocyte subsets in WT and Scap–/– mice. Frequencies (middle) and cellularity (right) of thymocyte subsets. Each symbol represents an individual mouse (n = 3 each group). (B) Flow cytometry of DN1-4 subsets (gated on lineage-negative DN cells) of WT and Scap–/– thymocytes. (C) Flow cytometry of ISP cells in WT and Scap–/– mice. (D) Flow cytometry of TCRgd+ cells in WT and Scap–/– thymocytes. (E and F) Flow cytometry of thymocyte subsets (E), and TCRgd+ cells (F) in the thymus of WT and Hif1a–/– mice. Data are means ± SEM. One-way ANOVA with Tukey’s test (A). Data are representative of two independent experiments (A to F). Numbers indicate percentage of cells in quadrants or gates.

A

100 101 102 103 104100

101

102

103

104

3.3281.3

1.89 11.3100 101 102 103 104

100

101

102

103

104

2.3985.4

8.271.76

CD4

CD

8Scap–/–WT

ns

0

15

30100

150

200

Thym

ocyt

es(�

106 )

DN DP CD8SP

CD4SP

ns nsns

WTScap–/–

ns

01530

100150200

C

0 102 103 104 105

0

102

103

104

105

91.2

8.68

CD8

TCRb

Scap–/–WT

0 102 103 104 105

0

102

103

104

105

94

5.77

B

100 101 102 103 104100

101

102

103

104

0.585 1.55

37 60.8100 101 102 103 104

100

101

102

103

104

0.526 1.17

36.4 61.9

CD25

CD

44

Scap–/–WTD

100 101 102 103 104100

101

102

103

104

0.167

100 101 102 103 104100

101

102

103

104

0.159

TCRb

Scap–/–WT

TCRgd

100 101 102 103 104100

101

102

103

104

1.3 89.3

2.26 5.68100 101 102 103 104

100

101

102

103

104

1.42 87.3

2.13 7.31

CD4

CD

8

Hif1a–/–WTF

100 101 102 103 104100

101

102

103

104

0.157

100 101 102 103 104100

101

102

103

104

0.17

TCRgd

TCRb

Hif1a–/–WTE

0510156080100

Thym

ocyt

es (%

)

DN DP CD8SP

CD4SP

ns

ns

ns

ns

05

101560

10080

Page 18: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

18

Fig. S8

Fig. S8. MYC is implicated in ab and gd T cell development. (A) Left: Flow cytometry of thymocyte subsets in WT and Myc–/– mice. Right: Cellularity of total thymocytes. Each symbol represents an individual mouse (n = 4 each group). (B) Flow cytometry of DN1-4 subsets (gated on lineage-negative DN thymocytes) in WT and Myc–/– mice. (C) Cellularity of DN subsets in WT and Myc–/– thymocytes. Each symbol represents an individual mouse (n = 3 each group). (D) Plots of icTCRb in the indicated WT and Myc–/– DN subsets. (E) Cellularity of TCRgd+ cells in the thymus of WT and Myc–/– mice. Each symbol represents an individual mouse (n = 4 each group). Data are means ± SEM. *P < 0.05, and **P < 0.005. Two-tailed unpaired t test (A, C, E). Data are representative of at least three independent experiments (A to E). Numbers indicate percentage of cells in quadrants or gates.

B

CD

44

0 102 103 104 105

0

102

103

104

105 1.6 4.72

29.9 63.90 102 103 104 105

0

102

103

104

105 0.0655 1.03

10.2 88.7

CD25

WT Myc–/–A

100 101 102 103 104100

101

102

103

104

1.989.4

5.51.45

100 101 102 103 104100

101

102

103

104

1.6118.8

48.910.3

CD

8Myc–/–WT

CD40

100

200

300

WTThym

ocyt

es(�

106 ) **

Myc–/–

DC

Cel

ls0 102 103 104 105

0

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

DN3a DN3b

icTCRb

WT Myc–/– E

TCRgd

+(�

106 )

0.0

0.2

0.4

0.6

0.8

WT

ns

00.20.40.60.8

Myc–/–0.00

0.08

0.160.2

5.2

10.2

DN

sub

sets

(�10

6 )

00.080.16

5.20.2

DN1 DN2 DN3 DN4

nsns

*ns

WTMyc–/–

10.2

Page 19: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

19

Fig. S9

Fig. S9. RAPTOR-deficient DN3 cells exhibit increased ROS but normal mitochondrial mass and membrane potential. (A) The expression of proteins involved in OXPHOS in DN3 and DN4 cells of WT and Rptor–/– mice. (B) Measurement of glycolytic activity in WT and Rptor–/– DN3 cells. DPM: disintegrations per minute. (C) Relative MFI of ROS in thymocyte subsets of WT and Rptor–/– mice (MFI of WT DN3 cells is set to 1). (D and E) Left: Plots of MitoTracker (D) and TMRM (E) in thymocyte subsets in WT and Rptor–/– mice. Right: Relative MFI of MitoTracker (Mito) (D) and TMRM (E) in thymocyte subsets of WT and Rptor–/– mice (MFI of WT DN3 cells is set to 1). Data are means ± SEM. *P < 0.0001. Two-way ANOVA with Bonferroni’s test (C to E). Data are representative of two independent experiments (A), or representative/combination of two independent experiments (B to E).

0.0

0.5

1.0

1.5WTKO

DN3 DN4 ISP DP

ns ns

ns

ns

WTRptor–/–

0

0.5

1.0

1.5

Rela

tive

Mito

E

0 102 103 104 1050

20

40

60

80

100

TMRM

Cells

WT Rptor–/–

0.0

0.5

1.0

1.5WTRaptor ko

Rela

tive

TMRM

DN3 DN4 ISP DP

ns ns

nsns

0

0.5

1.0

1.5 WTRptor–/–

WT

Rptor

–/–

WT

Rptor

–/–

DN3 DN4

ATP5AUQCRC2

MTCO1SDHB

NDUFB8b-ACTIN

A

0

1

2

3WTRaptor ko

DN3 DN4 ISP DP

*

Rela

tive

ROS

ns ns ns

0

1

2

3 WTRptor–/–

C

D

Cells

WT Rptor–/–

MitoTracker0 102 103 104 105

0

20

40

60

80

100

0

100

200

300

3 H-H

2O (D

PM)

0

100

200

300

WTRptor–/–

B

Page 20: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

20

Fig. S10

Fig. S10. Modulation of ROS production alters fate choices of DN3 cells. (A) Left: WT DN3a cells were co-cultured with OP9-DL1 cells in the absence or presence of Gal/GAO, followed by analysis of TCRgd and CD4 expression. Right: Frequencies and numbers of CD4+ and TCRgd+ cells (n = 5 each group). (B) Top: WT DN3a cells were co-cultured with OP9-DL1 cells with or without DCA treatment, followed by analysis of TCRgd and CD4 expression. Bottom: Fold change of the ratio of TCRgd+ to CD4+ cells (WT without DCA treatment is set to 1, n = 3 each group). (C) Analysis of ROS production in WT DN3a cells co-cultured with OP9-DL1 cells in the absence or presence of DCA for 3 days, with MFI plotted in graphs. (D) Left: WT DN3a cells were cultured with plate-bound NOTCH ligand DLL4 in the presence of CXCL12 and treated with the indicated inhibitors, followed by examination of CD4 and TCRgd expression. Right: Frequency of TCRgd+

cells. (E) WT DN3a cells were co-cultured with OP9-DL1 cells in the absence or presence of NAC or GSH for 5 days, and analyzed for cellularity of total cells in the culture (n = 3 each group). (F) WT DN3a cells were cultured with plate-bound NOTCH ligand DLL4 in the presence of IL-7 and treated with either NAC or GSH, followed by examination of TCRgd+ frequency (left; n = 4 each group) and cellularity (right; n = 3 each group). (G) MFI of CD73 on TCRgd+ cells in F. (H) WT and Rptor–/– DN3a cells were co-cultured with OP9-DL1 cells in the absence or presence of NAC for 3 days, followed by analysis of live cells (7-AAD– cells). Data are means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.0001. Two-tailed Mann-Whitney test (A for frequency analysis), two-tailed unpaired t test (A for cellularity analysis), one-tailed unpaired t test (B), or one-way ANOVA with Tukey’s test (E to G). Data are representative of at least three independent experiments (A to H). Numbers indicate percentage of cells in gates.

0

20

40

60

80

100A B

0 102 103 104 1050

1000

2000

3000

4000

5000

0 102 103 104 1050

100

200

300

ROS

Mock DCA

Cells

6526 14146

C

HROS OP9-DL1 cell survival

Mock NAC0

50

100

150

Live

cel

ls (%

)

WT

KO

Mock NAC

Live

cel

ls (%

)

WTRptor–/–

0.0

0.5

1.0

1.5 nsns

Tota

l cel

ls (×

106 )

0

10

20

30Mock Gal/GAO

0 102 103 104 105

0

102

103

104

105

77.4

2.15

0 102 103 104 105

0

102

103

104

105

0.624

20.3

0

5

10

15

20

E

0 102 103 104 105

0

102

103

104

105

50.6

3.88

0 102 103 104 105

0

102

103

104

105

24.9

8.04

0 102 103 104 105

0

102

103

104

105

3.5

17.7

CD4

TCRgd

Mock DCA Gal/GAOD

TCRgd

+(%

)

Mock

NACGSH

Mock

DCAGal/

GAO

CD4

TCRgd

TCRgd

+(%

)

Mock Gal/GAO

**

0

1

2

3

4

5F

Mock

NACGSH

TCRgd

+(×

103 ) **

* G

0

1

2

3

Mock

NACGSH

CD73

MFI

(×10

4 ) ***

CD4+

(%)

MockGal/GAO

**

0 102 103 104 105

0

102

103

104

105

46.8

3.49

0 102 103 104 105

0

102

103

104

105

2.38

6.43

DCAMock

TCRgd

CD4

0

5

10

15

20

TCRgd+

(%)

Mock

NACGSH

***

0

2

4

6

8

10

TCR

gd (x

103)

TCRgd

+(×

103 )

MockGal/GAO

*

CD4+

(×10

4 )

0

5

10

15

MockGal/GAO

***

Fold

cha

nge

ofTC

Rgd

+ /CD4

+

Mock DCA0

10

20

30

40 **

02040

8060

100

012

43

5

0

5

10

15

0

10

20

30

0

5

10

15

20

0

5

10

15

20

0

10

20

40

30

024

86

10

0

50

100

150

0

1

2

3

0

0.5

1.0

1.5

Page 21: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

21

Fig. S11

Fig. S11. mTORC1 activation integrates signals from pre-TCR and NOTCH. (A) Plots of p-S6 in DN3 cells from WT and Rag1–/– mice, with MFI plotted in graphs. (B) Phosphorylation of S6 and 4E-BP1 in WT and Rag1–/– DN3a cells co-cultured with OP9-DL1 cells in the absence of IL-7 for 2 days (WT, n = 4; Rag1–/–, n = 5). (C and D) DN3a cells from WT or Bcl2-Tg mice were cultured with OP9 or OP9-DL1 cells in the absence of IL-7 for 2 days, followed by examination of phosphorylation of S6 and 4E-BP1 (C) and Annexin V and 7-AAD (D), with MFI plotted above graphs. (E) WT DN3 cells were co-cultured with OP9-DL1 cells in the absence or presence of rapamycin (10 ng/ml) for 3 days, followed by analysis of CD4 and TCRgd. (F) WT DN3 cells were co-cultured with OP9 or OP9-DL1 cells in the absence or presence of IL-7 for 2 days, followed by analysis of p-S6, with MFI plotted in graphs. Data are means ± SEM. **P < 0.005 and ***P < 0.001. Two-tailed unpaired t test (B). Data are representative of three (A, B, E, F), or two (C, D) independent experiments. Numbers indicate percentage of cells in gates.

0 102 103 104 105

0

102

103

104

105

4.19

54.5

0 102 103 104 105

0

102

103

104

105

8.78

34

CD4

TCRgd

Mock

Rapamycin

CA

0 102 103 104 1050

20

40

60

80 34473447

0 102 103 104 1050

100

200

300

400

10071007

p-S6

Cells

WT Rag1–/–

E OP9 OP9-DL1

IL-7 0 102 103 104 1050

20

40

60

80

100

1027

0 102 103 104 1050

20

40

60

80

100 1190

0 102 103 104 1050

10

20

30

40

50

2368

0 102 103 104 1050

20

40

60

80 2556

p-S6

Cells

-

+

FOP9 OP9-DL1

0 102 103 104 105

0

102

103

104

105

5.97

0 102 103 104 105

0

102

103

104

105

1.69

0 102 103 104 105

0

102

103

104

105

3.35

0 102 103 104 105

0

102

103

104

105

3.13

Annexin V

7-AA

D

WT

Bcl2-Tg

OP9OP9-DL1

Bcl2-Tg

p-4E-BP10 102 103 104 105

0

20

40

60

80

100

1014/2564

p-S60 102 103 104 105

0

20

40

60

80

100

982/6457

WT

0 102 103 104 1050

20

40

60

80

100

981/2424

0 102 103 104 1050

20

40

60

80

100

742/6741

p-4E-BP1p-S6

D

B

0

1

2

3

4

0

5

10

15

p-S6

MFI

(×10

3 )

p-4E

-BP1

MFI

(×10

3 )

WT

Rag1 ko

WTWT

Rag1 koRag1–/–

*****

Page 22: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

22

Fig. S12

D

GmTORC1 signaling

tSNE-1

tSNE

-2

High

Low

Glycolysis

tSNE-1

tSNE

-2

-20 -10 10 200

-20

-10

1020

0

MYC targets

-20 -10 10 200

-20

-10

1020

0

tSNE-1-20 -10 10 200

tSNE

-2-2

0-1

010

200

A B

NES = 1.57FDR = 0.027

NES = 1.88FDR = 0.003

NES = 1.77FDR = 0.004

C

p-p65 b-ACTIN

0

Rptor–/–WT

45 75 0 45 75Time (s)

NOTCH signaling

tSNE-1-20 -10 10 200

-20

-10

1020

0tS

NE-2

High

Low

F

Cluster1 5 8 2 4 7 3 6

***

Il2ra

Ptcra

***

***

E

***Notch1

Notch3

Dtx1

1 5 8 2 4 7 3 6Cluster

Dtx3l

***

***

***

tSNE-1

Pre-TCR signaling

-20 -10 10 200

-20

-10

1020

0tS

NE-2

ERK signaling

tSNE-1

tSNE

-2-2

0-1

010

200

-20 -10 10 200

High

Low

H

1 5 8 2 4 7 3 6Cluster

Slc7a5

Shmt2***

***

Ldha

Tpi1***

***

Page 23: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

23

Fig. S12. RAPTOR deficiency alters transcriptional programs and signaling and metabolic pathways. (A and B) GSEA reveals the significant enrichment of the upregulated pathways in RAPTOR-deficient DN3a cells (A) and gd T cells (B). (C) Immunoblot analysis of p-p65 (S276) and b-ACTIN in WT and RAPTOR-deficient gd T cells stimulated with anti-gdTCR for the indicated time points. (D) Violin plots of Il2ra, Ptcra and additional NOTCH target genes among the 8 clusters. A violin plot combines the box plot and the local density estimation into a single display. The black bars and thin lines within the violin plots indicate the interquartile range (1st quantile – 3rd quantile) and the entire range of the data (up to 1.5 fold of interquartile range from 1st/3rd quantile), respectively, and the white dots in the center indicate the median values. (E) tSNE (top) and violin (bottom) plots of the signature of NOTCH signaling among the 8 clusters. (F) tSNE plots of the signatures of pre-TCR and ERK signaling among the 8 clusters. (G) tSNE plots of gene signatures of mTORC1 signaling, MYC targets and glycolysis among the 8 clusters. (H) Violin plots of metabolic genes among the 8 clusters. Data are means ± SEM. ***P < 0.001. Two-tailed Mann-Whitney test (D, E, and H). Data are from one experiment (A, B (WT, n = 3; Rptor–/–, n = 3); D to H (WT, n = 2; Rptor–/–, n = 3)), or representative of two (C) independent experiments.

Page 24: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

24

Fig. S13

Fig. S13. Monocle ‘pseudotime’ trajectory of single cell transcriptomics data. (A) Top left panel: pseudotime trajectory of scRNA-seq data colorized by pseudotime (blue indicates early time, and red indicates late time). The remaining 8 panels are the cell projections for each individual cluster. (B) Projection of DN3a gene signature onto pseudotime trajectory. (C) Projection of gd T cell gene signature onto pseudotime trajectory. (D) Pseudotime trajectory of cellular density colorized by genotypes: WT (blue) and Rptor–/– (red).

Late (Cluster 4)

Cluster 2

Early (Clusters 1, 5, 8)

A

D

C

BPseudotime Cluster 1 Cluster 2

Cluster 3 Cluster 4 Cluster 5

Cluster 6 Cluster 7 Cluster 8

DN3a signature

Component 1

Com

pone

nt 2

gd T cell signature

Component 1

Com

pone

nt 2

WT Rptor–/–

Component 1

Com

pone

nt 2

Com

pone

nt 2

Component 1

High

Low

High

Low

High

Low

Page 25: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

25

Fig. S14

Fig. S14. Loss of RAPTOR or MYC enhances signal strength. (A and B) Plots of p-ERK (A) and expression of ID3 (B) in WT and Myc–/– DN3 cells, with MFI plotted in graphs. (C and D) Comparison of ROS production (C) and CD5 expression (D) between DN3 cell subsets from WT and Myc–/– mice. (E) WT DN3a cells were co-cultured with OP9-DL1 cells in the absence or presence of DCA (10 mM) for 3 days, followed by analysis of ID3 expression, with MFI plotted in graphs. (F and G) Expression of CD73 on TCRgd+ cells from co-culture of WT DN3a cells with OP9-DL1 cells upon DCA (F) or Gal/GAO treatment (G), with MFI plotted in graphs. Data are representative of at least three independent experiments (A to G).

ACe

lls

p-ERK

WT Myc–/–

0 102 103 104 1050

30

60

90

120

503

0 102 103 104 1050

50

100

150

200

250 1078503 1078

E

0 102 103 104 1050

100

200

300 22827

ID3

Cells

Mock DCA

0 102 103 104 1050

20

40

60

80

3344922827 33449

F

0 102 103 104 1050

5

10

15

206393

0 102 103 104 1050

5

10

15 18064

Mock DCA

CD73

Cells

180646393

G

CD730 102 103 104 105

0

5

10

15

203846

0 102 103 104 1050

20

40

60

35610

Mock Gal/GAO

Cells

3846 35610

D

WTMyc–/–

DN3a DN3b

0 102 103 104 1050

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

CD5Ce

lls

ID3

Cells

0 102 103 104 1050

50

100

150

2001094

0 102 103 104 1050

200

400

6001866

Myc–/–WT

18681094

B

C

ROS0 102 103 104 105

0

20

40

60

80

100

0 102 103 104 1050

20

40

60

80

100

Cells

DN3a DN3b 2181/4279 5552/8781WTMyc–/–

Page 26: Supplementary Materials for - Science · 2018-07-02 · Data visualization Underneath cell variations were visualized in a 2D projection by t-distributed stochastic neighbor embedding

26

Fig. S15

Fig. S15. Schematics of mTORC1 and metabolic control of redox homeostasis and signal strength in T cell lineage choices. mTORC1, activated by TCR and NOTCH signaling, intersects with MYC in a feed-forward manner to orchestrate cell metabolism in developing thymocytes. Dynamic metabolic remodeling, including proper engagement of glycolysis and OXPHOS, contributes to anabolism and cell expansion. However, dysregulated metabolism results in aberrant ROS production and altered redox balance, which in turn impinges upon the signal strength, thereby altering lineage choices of ab and gd T cells.

RAPTOR/mTORC1

MYC

Metabolic remodeling(Glycolysis & OXPHOS)

IL-7

NOTCH (DN3a) TCR

(DN3b)

⍺b T cells gd T cellsLineage choices

AnabolismSignal strength

ROS