Cell lines and mice
HT-2, HEK293T, EL4, E.G7-OVA and NALM6 cell lines were purchased from ATCC. YUMM1.7 cell line was kindly provided by M.B. and was reported previously93. All cell lines were regularly tested for mycoplasma and confirmed to be negative. OT-I TCR transgenic mice (OT-I mice; Stock # 003831) and Constitutive Cas9-expressing mice (Cas9 mice; Stock # 026179) both with C57BL/6 background were purchased from Jackson lab. Cas9/OT-I mice were generated by crossing Cas9 and OT-I mice with genotyping following Jackson Lab protocols. 7-8-week-old female C57BL/6 N mice purchased from Envigo (Stock # 044) were used as E.G7-OVA tumor and adoptive T cell transfer recipients. Mice were co-housed in a specific-pathogen-free barrier facility, euthanized by CO2, and used in experiments under procedures approved by the Yale University Animal Care and Use Committee.
Plasmids
The human Brie genome-wide CRISPR knockout pooled library in the pLentiCRISPRv2 one vector system (co-expressing spCas9 and sgRNA), with four sgRNAs per gene, was obtained from Addgene (Addgene # 73632, a gift from David Root and John Doench54) and prepared in the Yale Cancer Center Functional Genomics core). pMSCV-U6sgRNA(BbsI)-PGKpuro2ABFP was a gift from Sarah Teichmann (Addgene plasmid # 102796; http://n2t.net/addgene:102796; RRID:Addgene_102796). PGKpuro2ABFP was replaced by EF1a-core promoter-mScarlet gblock (IDT) to generate pMSCV-guide-EF1a-mScarlet vector compatible with Bio-Rad S3e sorting. For single cell perturb-seq, the original gRNA scaffold sequence was further replaced by a new scaffold sequence with 10x genomics compatible capture sequence 1 incorporated into the stem-loop to generate pMSCV-scguide-EF1a-mScarlet vector. Customized sub-library for bulk in vivo screening, with non-targeting control sgRNA sequences published previously and top ranked 6 sgRNA sequences per gene designed by CRISPick, was synthesized by Genscript and cloned into pMSCV-guide-EF1a-mScarlet. Customized sub-library for single cell in vivo screening, with 4 non-targeting control sgRNA sequences and top ranked 3 sgRNA sequences per gene selected from the above bulk sub-library, was synthesized by IDT and cloned into pMSCV-scguide-EF1a-mScarlet by NEB stable competent cells. The normal distribution and integrity of sub-libraries were confirmed by illumine next-generation-sequencing. Top two ranked murine Cd8a and Cul5 sgRNAs (Supplementary Data 7) were designed by CRISPick, synthesized by IDT and cloned into pMSCV-guide-EF1a-mScarlet or pMSCV-scguide-EF1a-mScarlet vector by NEB stable competent cells. pSLCAR-CD19-BBz was a gift from Scott McComb (Addgene plasmid # 135992; http://n2t.net/addgene:135992; RRID:Addgene_135992). CUL5-HA overexpression retroviral plasmid was derived from in house MIGR-IRES-GFP vector. Mouse Cul5 open-reading frame with HA tag at its C terminal was amplified by PCR. cDNA of TCR-activated mouse primary CD8+ T cells was used as PCR template. Primers were designed to add XhoI site at 5’ and HA sequence plus EcoRI site at 3’ of the PCR product (Supplementary Data 7). Backbone vector and PCR product were double digested by XhoI and EcoRI, and ligated by T4 ligation to get MIGR-CUL5-HA-IRES-GFP plasmid. Plasmids containing Myc-CUL5, HA-PCMTD2, GFP-CD247, GFP-CD3d, GFP-CD3e, GFP-CD3g and GFP-IL2RB overexpression plasmids were generated for transient expression in 293 T cells.
Lentiviral and retroviral production
Lentiviruses with genome-wide murine CRISPR library or CAR-CD19 were produced in HEK293T cells transfected with library vector, pMD2.G (addgene, #12259) and psPAX2 (addgene, #12260). Retroviruses with sub-libraries or individual non-targeting or target gene sgRNAs were produced in HEK293T cells transfected with sublibrary/guide vector and pCL-Eco (addgene, #12371). Viral soups from 24-hour and 48-hour post-transfection were harvested, combined, filtered, aliquoted and saved in −80 °C freezer until use. The titers of lentiviruses and retroviruses of each batch were determined by the transduction of HEK293T and NIH3T3 cells respectively.
In vitro genome-wide CRISPR KO screening
HT-2 cells were transduced with the lentiviral CRISPR/Cas9 library at MOI of 0.3 with over 200x coverage per sgRNA. About 25% transduction efficiency was confirmed by three-day puromycin selection to make sure one type of sgRNA per cell as the majority (Supplementary Fig. 1a). After 3 days of selection with 0.6 µg/ml puromycin, 2×107 transduced HT-2 cells were saved as input, 2×107 were cultured for 21 days either in 100 ng/ml IL2 medium (IL2) or in IL2 medium containing 120 pg/ml TGFβ1 (IL2 + TGF). For each condition, genomic DNA from 2×107 cells was extracted by DNeasy blood & tissue kit (Qiagen, Cat.69504) and sgRNA cassettes were PCR amplified for illumina NGS54.
In vivo bulk CRISPR KO screening
3×106 EG.7-OVA cells were s.c. inoculated into 7-8 weeks old C56BL/6 N female mice. Tumor sizes were monitored every three days by caliper with volume calculation as v = d2xD/2. When tumors reached 0.5 cm3, untouched CD8+ T cells were isolated from the spleens of male and female Cas9/OT-I mice by CD8+ T cell isolation kit (MACS, Cat.130-104-075), and immediately stimulated by anti-mouse CD3/28 beads (Thermo Scientific, Cat.11456D) at 1:1 ratio for 24 h, followed by retroviral sub-library transduction at MOI of 0.2. Transduced CD8+ T cells were further expanded in IL2/7/15 medium for 3 days to ensure genome-editing completed. GFP/mScarlet double positive cells were then sorted on a Bio-Rad S3e sorter. 2×106 cells were saved as input and 2×106 sorted T cells per mouse were i.v. transferred into a total of four sub-lethally irradiated (4 Gy) EG.7-OVA tumor-bearing mice. 12 days post T cell transfer, GFP+ T cells were isolated and enriched from tumors, spleens and tumor-draining lymph nodes of recipient mice by sorting. Genomic DNA from sorted cells were extracted by Qiagen DNeasy blood & tissue kit. sgRNA cassettes were PCR amplified for illumina NGS.
PCR of sgRNA cassettes for NGS
For all PCR products, a stagger P5-read1 forward primer mixture was used to increase NGS reading diversity, and different index-included P7-read2 reverse primers were used for PCR of individual replicate samples before pooled NGS. Primer sequences are shown in Supplementary Data 7. PCR reaction was set up according to NGS protocol of NEBNext UltraTM II Q5 Master Mix (NEB, Cat.M0544S). PCR products were purified by AMPure XP beads (Beckman Coulter, Cat. A63880) at 1:1 volume ratio. Purified PCR products were quantified by TapeStation (D1000 ScreenTape assay) before loading for NGS.
In vivo single cell CRISPR KO screening
Tumor inoculation and sub-library transduced T cell transfer followed the same procedures of in vivo bulk CRISPR KO screening. 7 days post T cell transfer, live transferred tumor-infiltrating CD8+ T cells as CD8a+GFP+ T cells were sorted and washed once in cold 1x PBS containing 0.04% BSA prior to resuspending in cold 1x PBS containing 0.04% BSA at 1×106 cells/ml concentration. Cells were then processed by Yale Center for Genome Analysis to generate single-cell RNA library and sgRNA library separately according to the instruction of 10x genomics 3’ V2 single cell RNA sequencing kit with sgRNA feature. Libraries were sequenced by illumina NGS.
In vivo gene KO validation
1×106 sorted NC or Cul5 sgRNAs (Supplementary Data 7) transduced Cas9/OT-I T cells per mouse were i.v. transferred into sub-lethally irradiated (4 Gy) EG.7-OVA tumor-bearing mice in total of 4-6 mice per group when average tumor sizes reached around 1 cm3. Tumor sizes were monitored every two days post T cell transfer. CTLA4 KO was achieved by nucleofection of ribonucleoprotein (RNP) complex containing Cas9 protein (IDT, Cat.1081058) and Ctla4 sgRNA (Supplementary Data 7) into naïve CD8+ T cells before activation. At the endpoint, tumors and TDLNs were isolated to make single cell suspension for immediate flow cytometry staining or in vitro re-stimulation by PMA/ionomycin prior to flow cytometry staining. Metastasis into LN was defined as abnormally enlarged inguinal LN well separated from tumor mass.
Tissue processes and transferred CD8+ T cell isolation
EG.7-OVA tumors were resected and minced into 1-2 mm pieces, followed by 20 min digestion at 37 °C in HBSS buffer containing Mg2+, Ca2+, HEPES, 2% FBS, collagenase, DNase I with constant shaking at 60 rpm speed. Digestion was stopped by 10 mM ETDA PBS solution and pass through 70 µm strainer. Remaining tumor pieces were further smashed with 3 ml syringe plunger and washed through the strainer. For CD8+ T cell sorting, tumor single cell suspension in 1xPBS was layered on top of Ficol-plus and centrifuged at 1000 g x 20 min without braking at room temperature. Buffy coat in which live CD8 T cells were enriched was collected and washed in 1xPBS. Spleens and TDLNs were directly smashed with 3 ml syringe plunger through 70 µm strainers. Red blood cell lysis buffer was further applied for smashed splenocytes to remove red blood cells. Single cell suspensions from different tissues were stained with PE-anti-mouse CD8a. GFP+ CD8a+ cells were sorted by Bio-Rad S3e sorter.
In vitro CD8+ T cell stimulation and cancer cell killing assay
Following retroviral transduction of MACS purified Cas9/OT-I T cells as described above, GFP+mScarlet+ cells were sorted and expanded in culture medium containing 5 ng/ml hIL2 (R&D, Cat.202-IL-010), 2.5 ng/ml mIL7 (Peprotech, Cat.217-17) and 25 ng/ml mIL15 (Peprotech, Cat.500-P173) for 3-5 days. For flow cytometry detection of phosphorylated proteins, activation and effector markers, T cells at 1×106/ml were kept in culture withdraw of cytokines for 8 h before stimulated with 1 µg/ml plate-coated anti-mouse CD3e (Biolegend, Cat.100340) and 1 µg/ml soluble anti-mouse CD28 (Biolegend, Cat.102116) for different time points with or without addition of transporter block (Thermo Scientific, Cat.00-4980-03). For some experiments, 250 nM TAS4464 (Selleckchem, Cat.S8849) was added for the inhibition of neddylation. Supernatants at different time points without the addition of transporter block were collected and kept in −80 °C freezer for cytokine ELISA detection later. For cancer cell killing assay, EL4 cells pre-stained with CFSE (Thermo Scientific, Cat.C34554) and EG.7-OVA cells pre-stained with CellTrace violet (Thermo Scientific, Cat.C34557) according to instruction were 1:1 mixed and seeded in U-bottom 96-well plate with 1×105 total cells per well. T cells with different effector to target ratio were added and kept in culture overnight. Counting beads were added to calculate the absolute number of live CFSE+ EL4 cells and violet dye+ EG.7-OVA cells by flow cytometry. In some experiments, T cells and EL4/EG.7-OVA cells were co-cultured for 6 h with the addition of transporter block to detect antigen-specific cytotoxic T cell activation by flow cytometry.
CUL5/PCMTD2 single and double KO in primary mouse CD8+ T cells
MACS-purified CD8+ T cells from spleens of WT mice were cultured in 5 ng/ml mIL7 overnight. Then PCMTD2 knockout of CD8+ T cells was achieved by CRISPR-Cas9. In brief, pre-designed crRNAs (IDT) for mouse CUL5 and PCMTD2 (Supplementary Data 7) or non-targeting control was annealed with tracrRNA (IDT, Cat.1072532) to form guide RNA, which then was mixed with Cas9 protein to generate ribonucleoprotein (RNP) complex. 1×106 CD8+ T cells were re-suspended in P3 buffer (Lonza) containing RNP complex and enhancer DNA (IDT, Cat.1075915) and under electroporation with CM-137 program of 4D-nucleofactor X Unit (Lonza). Electroporated T cells recovered overnight in IL7 medium were then activated by anti-mouse CD3/CD28 beads (at 1:1 ratio for two days and expanded in culture medium containing 5 ng/ml hIL2, 2.5 ng/ml mIL7 and 25 ng/ml mIL15 for 2 days before experiments.
RT-qPCR
RNA was extracted from cells by RNeasy plus mini kit (Qiagen, Cat.74034) and quantified by nanodrop. cDNA was then synthesized by iScript Reverse Transcription Supermix (Bio-Rad, Cat.1708841) for RT-qPCR from Bio-Rad. qPCR reaction was set up by using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Cat.1725270) and run on CFX96 Real-Time PCR Detection System (Bio-Rad). Primers of individual genes were selected from PrimerBank as shown in Supplementary Data 6 and synthesized by IDT. The specificity of qPCR reaction was confirmed by melting curve. Relative quantification of individual genes was normalized to Actb.
ELISA
Supernatants of in vitro stimulated CD8+ T cells without secretion blocking were harvested at 16-hour post stimulation, and stored in the −80 °C freezer before test. Mouse IFNg ELISA (Thermo Scientific, Cat.KMC4021) were performed according to the instruction. In brief, capture antibodies were coated in 96-well plate overnight in 4 °C. Cytokine standards and sample supernatants were then added, followed by the addition of detection antibodies and secondary antibody-HRP conjugation. After extensive washes between each step, TMB was added for 20 min at room temperature in dark. After stopping the reaction, absorbance at 450 nm was detected. IFNg concentrations were then calculated based on the standard curves.
Flow cytometry
Cells were washed once in 1x cold PBS and resuspended in 50 µl 1xPBS containing anti-CD16/32 (BD, Cat.553142) and fixable aqua live/dead dye (Thermo Scientific, Cat.L34957) for 10 min on ice. Without wash, 50 µl surface antibody/tetramer mixture in 1xPBS was added for 15 min on ice, followed by 1x cold PBS wash twice. For intracellular cytokine, CTLA4, CUL5 and pERK1/2 staining, cells were fixed in 4% PFA for 20 min at room temperature and washed in 1x permeable buffer (Thermo Scientific, Cat.00-833356). 50 µl intracellular antibody mixture in 1x permeable buffer was added for 30 min at room temperature, followed by 1x permeable buffer twice. 50 µl anti-rabbit-PE or -Alexa Fluor 350 secondary antibody in 1x permeable buffer was added for 30 min at room temperature, followed by 1x permeable buffer twice. For intracellular pSTAT5 staining, cells were fixed in 4% PFA for 20 min at room temperature and washed in 1x PBS, followed by permeabilization in pre-colded methanol for 20 min on ice. After 1xPBS wash, 50 µl anti-pSTAT5-APC in 1xPBS was added for 30 min at room temperature, followed by 1xPBS wash twice. After final wash, cells were resuspended in 1x PBS and detected by BD LSR II. Antibodies used for flow cytometry are as follows:
anti-mouse: Biolegend: CD3-Pacific Blue (Cat.155611), CD8a-PE-Cy7 (Cat.100721), CD62L-BV605 (Cat.104437), CD11c-Pacific blue (Cat.117321), NK1.1-APC (Cat.108709), CD19-PE-Cy7 (Cat.115519), CD122-APC (Cat.105911), CD127-APC-Cy7 (Cat.135039), GZMB-APC (Cat.372203), IFNg-APC-Cy7 (Cat.505849), TNF-PE-Cy7 (Cat.506323), IL2-PB (Cat.503820), CD25-APC-Cy7 (Cat.101917), CD5-PB (Cat.100641), ICOS-APC (Cat.107711), PD1-PE-Cy7 (Cat.135215), CTLA4-APC (Cat.106309), CD62L-PE-Cy7 (Cat.104417), Vb5-PB (Cat.139515), Va2-PE (Cat.127807); Thermo Scientific: CD137-PB (Cat.48-1371-82), F4/80-APC(Cat.17-4801-82), CD4-BUV395(Cat.363-0042-80), Foxp3-PE-CY7(Cat.25-5773-80); BD: CD8-BUV395(Cat.563786), CD107a-APC (Cat.560646), pSTAT5-Alexa 647 (Cat.562076), anti-human (Biolegend): GZMB-APC (Cat.372203), IFNg-PE (Cat.502508), CTLA4-APC (Cat.369611), anti-rabbit: IgG-PE (Thermo Scientific, Cat. P-2771MP), IgG-Alexa Fluor 350 (Thermo Scientific, Cat.A-11069), anti-pERK1/2 (Cell Signaling Technology, Cat. 9101), Rabbit polyclonal IgG anti-CUL5 (Thermo Scientific, Cat.A302-173A); SIINFEKL-H-2K(b) tetramer-BV421 (NIH Tetramer Core Facility, Cat.53995)
Proteomics analysis and co-IP for mass spectrometry
For proteomics measurements, 1×106 cytokine expanded CD8+ T cells with or without CUL5 KO were either immediately harvested as T0 or continuously cultured in medium without cytokines for 8 h. Then cytokine deprived CD8+ T cells were harvested either immediately as T8 or after 16 h stimulation with 1 µg/ml plate-coated anti-CD3 and 1 µg/ml soluble anti-CD28 as T16. Harvested cells were washed in cold PBS for 3 times, snap frozen in liquid nitrogen and stored in −80 °C freezer. Cell pellets were lysed by the lysis buffer 10 M urea containing the cOmplete™ protease inhibitor cocktail (Roche, #11697498001). The cell tubes were ultrasonicated at 4 °C for two cycle (1 min per cycle) using a VialTweeter device (Hielscher-Ultrasound Technology)94,95, and then centrifuged at 20,000 x g for 1 h to remove the insoluble material. For the supernatant protein mixture, the reduction and alkylation were conducted with 10 mM Dithiothreitol (DTT) for 1 h at 56 °C and then 20 mM iodoacetamide (IAA) in dark for 45 min at room temperature. The samples were diluted by 100 mM NH4HCO3 and digested with trypsin (Promega) at ratio of 1:20 (w/w) overnight at 37 °C. The digested peptides purification was performed on C18 column (MarocoSpin Columns, NEST Group INC) and 1 µg of the purified peptides was injected for mass spectrometry analysis.
For co-IP in triplicate, 1×107 CD8+ T cells per replicate transduced with MIGR-IRES-GFP control vector or MIGR-CUL5-HA-IRES-GFP CUL5-HA overexpression vector were stimulated in 1ug/ml anti-CD3 coated plate plus 1ug/ml soluble anti-CD28 for 12 h. 10 µM MG-132 was added for the last 8 h before harvesting cells. Harvested cells were washed in cold PBS for 3 times and immediately resuspended in IP lysis buffer (Thermo Scientific, Cat.87787) with proteinase/phosphatase inhibitor cocktail on ice for 10 min. After a high-speed centrifuge, supernatant was transferred into a new tube, followed by overnight incubation with anti-HA antibody (Biolegend, Cat.901515). Protein G dynabeads (Thermo Scientific, Cat.10003D) were then added with rotation at room temperature for 1 hour. Dynabeads containing bound protein complex were sequentially washed with cold IP lysis buffer without proteinase/phosphatase inhibitor cocktail 4x, and wash buffer without detergent 2x on magnetic stand. Protein complex was eluted twice by elution buffer containing 0.5 M ammonium hydroxide and 0.5 mM EDTA. Combined elutes were snap frozen in liquid nitrogen and stored in −80 °C freezer. The elution was dried with SpeedVac, and then resolved with 50 μL 6 M urea for reduction and alkylation. Other proteomic sample preparation steps were identical with the above cell sample protocol.
The samples were measured by data-independent acquisition (DIA) mass spectrometry method as described previously96,97,98. The Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo Scientific) instrument coupled to a nanoelectrospray ion source (NanoFlex, Thermo Scientific) and EASY-nLC 1200 systems (Thermo Scientific, San Jose, CA). A 120-min gradient was used for the data acquisition at the flow rate at 300 nL/min with the temperature controlled at 60 °C using a column oven (PRSO-V1, Sonation GmbH, Biberach, Germany). Each DIA-MS cycle consisted of one MS1 scan and 33 MS2 scans of variable isolated windows with 1 m/z overlapping between windows. The MS1 scan range was 350 – 1650 m/z and the MS1 resolution was 120,000 at m/z 200. The MS1 full scan AGC target value was set to be 2E6 and the maximum injection time was 100 ms. The MS2 resolution was set to 30,000 at m/z 200 with the MS2 scan range 200 – 1800 m/z and the normalized HCD collision energy was 28%. The MS2 AGC was set to be 1.5E6 and the maximum injection time was 50 ms. The default peptide charge state was set to 2. Both MS1 and MS2 spectra were recorded in profile mode. DIA-MS data analysis was performed using Spectronaut v1599,100,101 with “directDIA” by searching against the mouse SwissProt protein database. The oxidation at methionine was set as variable modification, whereas carbamidomethylation at cysteine was set as fixed modification. Both peptide and protein FDR cutoffs (Qvalue) were controlled below 1% by Spectronaut and the resulting quantitative data matrix were exported from Spectronaut. The sample-wise normalization was disabled for IP-MS experiments. All the other settings in Spectronaut were kept as Default.
Tandem Ubiquitin Binding Entities (TUBES) IP for mass spectrometry
CUL5 KO and NC mouse CD8+ T cells were stimulated with flask coated anti-CD3 and soluble anti-CD28 for 16 h with 10 µM MG132 added in the last 8 h before harvesting. Then TUBES (LifeSensors, Cat. UM-0501M-1000) was used to pull down ubiquitinated proteins based on the instruction of the kit that later were measured by mass spectrometry as above.
Human T cell culture and manipulation
CD8+ T cells were purified from human PBMCs by untouched human CD8+ T cell isolation kit (MACS, Cat.130-094-156) and cultured at 1×106/ml in RPMI1640 medium containing 10% FBS, 5 ng/ml hIL7 (R&D, Cat.207-IL-005) and 50 ng/ml hIL15 (R&D, Cat.247-IL-005) overnight. Then CUL5 knockout of CD8+ T cells was achieved by CRISPR-Cas9. In brief, pre-designed crRNA (IDT) for human CUL5 (Supplementary Data 7) or non-targeting control was annealed with tracrRNA (IDT, Cat.1072532) to form guide RNA, which then was mixed with Cas9 protein to generate ribonucleoprotein (RNP) complex. 2×106 CD8+ T cells were re-suspended in P3 buffer (Lonza) containing RNP complex and enhancer DNA (IDT, Cat.1075915) and under electroporation with EO-115 program of 4D-nucleofactor X Unit (Lonza). Electroporated T cells recovered overnight in IL7 and Il15 medium were then activated by anti-human CD3/CD28 beads (Thermo Scientific, Cat.11131D) at 1:1 ratio for two days. For CAR-T generation, one day post beads activation, T cells were transduced with lentiviruses containing CAR-CD19 by spinfection supplied with 8 µg/ml polybrene and 5 ng/ml hIL2. Activation beads were magnetically removed two days later and T cells at 5×105/ml concentration were expanded in culture medium containing 5 ng/ml hIL2, 5 ng/ml hIL7 and 50 ng/ml hIL15. Fresh medium with cytokines was replaced every two days to keep T cell concentration lower than 2×106/ml.
Co-IP-Western Blot
For co-IP experiments carried out in HEK293T cells, cells were transfected with the indicated constructs for 24 h and MG132 (20 µM) was added to the culture medium 4hrs before collection. Cells were then lysed in a cell lysis buffer (50 mM HEPES, 150 mm NaCl, 1% Triton X-100, 0.5% NP40, 2 mM MgCl2, 2 mM EGTA, 10% Glycerol, protease inhibitor mixture (Roche), and phosphatase inhibitor mixture (Roche)). After centrifugation, the supernatants were immunoprecipitated by the indicated antibodies and then analyzed by western blotting. Antibodies used for the Co-IP experiments are anti-HA (Biolegend, 901513), anti-Myc (Proteintech, 16286-1-AP), and anti-GFP (Santa Cruz, sc-9996).
For co-IP of endogenous proteins in human primary CD8 cells, NC and CUL5 KO CD8+ T cells were prepared as described above and cultured with MG132 (20 µM) for 4hrs before collection. Immunoprecipitation was carried out as described above with anti-CUL5 antibody (Thermo Fisher A302-173A), followed by Western analysis using the same anti-CUL5 antibody and an -anti-PCMTD2 antibody (Thermo Fisher PA5-69557).
Ubiquitination assay
The transfected cells were treated with MG132 (20 µM) for a duration of 6-8 h and subsequently collected in a denaturing buffer (6 M guanidine-HCl, 0.1 M Na2HPO4/NaH2PO4 pH 7.5, 10 mM imidazole). The lysates were then incubated with Ni-nitrilotriacetic acid (Ni-NTA) agarose beads for 3 h. This was followed by four washes with the denaturing buffer and two additional washes with a low-salt buffer (25 mM Tris-HCl at pH 6.8, 20 mM imidazole). The beads were then eluted by boiling in the SDS sample buffer in the presence of 200 mM imidazole. After centrifugation, the supernatants were subjected to analysis via western blotting.
Western Blotting
WT mouse naïve CD8+ T cells were activation by anti-CD3/CD28 beads at 1:1 ratio for 0 or 48 h. Human CD8+ T cells were activated by anti-CD3/CD28 beads at 1:1 ratio for two days and NC or CUL5 KO ones were further expanded in 5 ng/ml hIL2, 5 ng/ml hIL7 and 50 ng/ml hIL15. Cells were then pelleted and washed twice with cold 1xPBS, followed by direct lysis in 2x SDS loading buffer (Thermo scientific, Cat.NP0007). Rabbit polyclonal IgG anti-CUL5 (Thermo Scientific, Cat.A302-173A) was used to detect mouse or human CUL5 expression. Rabbit polyclonal IgG anti-PCMTD2 (Thermo Scientific, Cat. PA5-69557) was used to detect mouse PCMTD2 expression. Mouse monoclonal IgG anti-GAPDH (Proteintech, Cat.60004-1-Ig), rabbit monoclonal IgG anti-Hsp90 (Proteintech, Cat. 13171-1-AP) and rabbit monoclonal IgG anti-Histone H3 (Cell Signaling Technology, Cat. 4499) were used as internal controls.
Human CD8+ T cell activation in vitro
Cytokine expanded CD8+ T cells were stimulated in 1 µg/ml plate-coated anti-CD3 and 1 µg/ml soluble anti-CD28. CAR-T cells were stimulated with NALM B cells (ATCC, Cat.CRL-3273) at 1:2 ratio. Stimulated CD8+ T cells or CAR-T cells were cultured for 6 h with the addition of transporter inhibitor prior to flow cytometry staining.
Human CAR-T cell in vitro killing
T cells expressing CAR-CD19 were sorted based on GFP reporter. Sorted CAR-T cells were co-cultured with CellTrace violet stained NALM B cells at different effector to target (E:T) ratio overnight. Violet+ live NALM B cells were detected by flow cytometry with counting beads added to calculate absolute number. Killing % =(live B cells in B cell culture alone – live B cells in co-culture)/live B cells in B cell culture alone. For chronic and repeated killing assay, CAR-T cells were stimulated every two to three days in 6-well plate coated with 1ug/ml anti-CD3 before harvesting for overnight killing setup. Remaining CAR-T cells were transferred into new 6-well plate for next round repeated activation and killing.
Bioinformatics analysis
Bulk CRISPR KO screening analysis
Initial quality check was performed using the FastQC program and sequence adapters were trimmed using the Cutadapt tool. Genome-scale and enriched sub-library scale CRISPR/Cas9 KO screening performed using MAGeCK version 0.5.9.2102. MAGeCK uses a maximum likelihood-based estimation (MLE) to measure Z-scores for each gene from the log2 fold changes of each sgRNAs in a robust approach. We rank our sgRNAs/genes based on robust ranking aggregation (RRA) and p-values. For bulk in vivo screening, four replicates in each group (Tumor, Spleen and Tumor-draining lymph node and Input) expression were averaged and positive RRA scores were used for the identification of top-ranked genes.
In vivo single-cell CRISPR KO screening analysis
Seurat 4.0 was used to process single-cell sequencing data. In the quality control (QC) analysis, poor quality cells with nFeature_RNA < 600, nCount_RNA < 1200, log10(Gene Per UMI) < 0.8 and Mitochondrial gene percentage over 10% were excluded. Genes with zero count numbers were removed. In addition, TCR genes were removed to prevent clustering bias caused by the contribution of variable V(D)J transcripts in major variable components. Cell cycle markers from Tirosh et al., 2015103 is loaded with Seurat and cell-cycle scores for G2M, S or G1 phase were quantified and assigned to each cell as the metadata. In addition, the sgRNA data was added into corresponding Seurat objects as the metadata. The cells with more than 1 sgRNA or without sgRNA were excluded. The feature-barcode matrix was first normalized and scaled with default settings in Seurat. Then the 3000 top variable genes were identified, which served as the input to principal component analysis (PCA) for dimensionality reduction. We applied Louvain algorithm for clustering and retained the 20 leading principal components as an input for further visualization. The UMAP embedding was used to visualize the single cells on a two-dimensional space with a perplexity of 100. Cell clusters were annotated with hall marker genes. Differentially expressed gene (DEG) analysis among different cell groups were based on the non-parametric Wilcoxon rank sum test with logfc.threshold=0.25 and min.pct =0.1. Functional analyses of sgRNA perturbations were evaluated by DEG analysis of sgRNA specific groups. Enrichr104,105 was then used to perform signaling pathway enrichment analysis with The Molecular Signatures Pathway Database (MsigDB_Hallmark_2020) to discover enriched biological pathways based on the list of identified DEGs.
Proteomics data statistical analysis
The two-sided Student’s t-test was used to find the differentially abundant proteins, Protein groups with p < 0.01 and a fold-change > 1.8 were reported as significant in total protein MS while p < 0.05 and a fold-change >1.5 in co-IP MS. The R software was used for the data virtualization with the packages ggplot2 (boxplot), scatterplot (volcano plots), pheatmap (correlation), corrplot (correlation), factoextra (PCA). The GO enrichment analysis was performed by Metascape106 with default Express Analysis. All scripts for bioinformatics analysis will be available upon request.
Statistics and reproducibility
Animals were grouped unblinded, but investigators were blinded for most of the qualification experiments. No data were excluded from analysis. Minimal group sizes for tumor progression studies were determined by using power calculations with the DSS Researcher’s Tookit with an α of 0.05 and power of 0.8. Statistical methods are described in the figure legends and statistical analysis was done with GraphPad Prism V10. Replicates used in statistical analysis are biological replicates.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.