Cell culture
HEK293T (ATCC® CRL-3216™) and A549 (ATCC® CCL-185™) cells were cultured in high glucose Dulbecco’s modified Eagle’s medium (Hyclone, SH30022.01), supplemented with 10% fetal bovine serum (FBS, Hyclone, SH30406.05) and 2 mM GlutaMAX (Gibco, 35050061) at 37 °C with 5% CO2. H1 hESCs (WiCell Research Institute) and the TRME hESCs cell line35, were grown in Matrigel (BD, 354277)-coated six-well plates in E8 medium (Stem Cells Technology, 05940) at 37 °C with 5% CO2. Both cells were authenticated and tested for the absence of mycoplasma contamination using Myco-Blue Mycoplasma Detector (Vazyme).
DE differentiation of hESCs
Differentiation of hESCs into DE cells was adopted from the previous study with minor modifications26. In brief, undifferentiated hESCs were dissociated into single-cell suspension by Accutase (Gibco, A1110501) and reseeded onto Matrigel-coated 24-well plates at a density of 1 × 105 cells/per well in E8 medium containing 10 μM Y-27632 (Selleck, S1049). When reached 80% confluency, DE differentiation was initiated by switching to the differentiation medium DMEM/F-12 (Gibco, 11330032) supplemented with 50 U ml−1 Penicillin–Streptomycin (Gibco, 15070063), chemically defined lipid concentrate (1:100, Gibco, 11905031), 10.7 μg ml−1 holo-Transferrin human (Sigma, T0665), 71 μg ml−1 l-ascorbic acid (Sigma, A8960), 14 ng ml−1 sodium selenite (Sigma, S5261), and 20 ng ml−1 Activin A (PeproTech, 12014E) and cultured for 3 days. CHIR99021 (3 μM, Selleck, S2924) was added to the medium for the first 24 h of differentiation and removed thereafter.
Construction of plasmid DNA
Lenti-FNLS-BE3-P2A-Puro-U6-sgRNA was generated using the ClonExpress II One Step Cloning Kit (Vazyme, C112), by combining the PCR-amplified FNLS-BE3 segment from pLenti-FNLS-P2A-Puro (Addgene, #110841) and the Age I/BamH I-digested LentiCRISPR v2 (Addgene, #52961) backbone. Lenti-FNLS-ABE7.10(AW)-P2A-Puro-U6-sgRNA was generated using the ClonExpress MultiS One Step Cloning Kit (Vazyme, C113) by combining the synthesized ecTadA(E59A)-ecTadA*(V106W) fragment, PCR-amplified codon-optimized Cas9n segment from Lenti-FNLS-P2A-Puro (Addgene, #110841), and the Age I/BamH I digested LentiCRISPR v2 (Addgene, #52961) backbone. U6-sgRNA fragment free FNLS-ABE7.10(AW) expression plasmid (Lenti-FNLS-ABE7.10(AW)-P2A-Puro) was generated by ligating the fragment of Kpn I/EcoR I-digested Lenti-FNLS-ABE7.10(AW)-P2A-Puro-U6-sgRNA using T4 DNA Ligase (New England Biolabs, M0202). LentiGuide-BSD-dTomato was generated by combining the PCR-amplified blasticidin S deaminase segment from pgRNA-CKB (Addgene, #73501) and the EcoR I/Xba I-digested LentiGuide-Hygro-dTomato (Addgene, #99376) backbone.
For sgHEK4, sgMYC-1−4, sgKRAS-1−4, and sgTP53-3, Lenti-FNLS-BE3-P2A-Puro-U6-sgRNA was used as a backbone, and Lenti-FNLS-ABE7.10(AW)-P2A-Puro-U6-sgRNA was used for sgMETTL3-1, sgEEF2-2, sgNEAT1, sgSOX2, sgADM, and sgSDHAF1. For sgYTHDF2, pLKO.1-blast was used as a backbone. All sgRNA/shRNA-inserted plasmids were constructed following the standard protocol of Target Guide Sequence Cloning Protocol from Dr. Feng Zhang’s laboratory (Havard University).
For the TRME assay, crRNAs with target m6A site at the 3rd base were designed. Then, full-length DR together with the U6 promoter was PCR-amplified from the pC016-LwCas13a plasmid backbone (Addgene, #91906) and cloned into the pSLQ1371 vector using the ClonExpress II One Step Cloning Kit (Vazyme, C112).
For SDHAF1 and ADM overexpression, coding sequence (CDS) of SDHAF1 and ADM were PCR amplified from open reading frames (ORF) plasmid purchase from Vigenebio (China), and cloned into the pLVX-TetOne-puro vector using the ClonExpress II One Step Cloning Kit.
sgRNAs, shRNAs, or crRNAs sequences used in this study were provided in Supplementary Table 1.
Lentiviral production and transduction
Lentivirus was packaged by co-transfection of HEK293T cells with 12 μg of lentiviral vector, 3 μg of pMD2.G (Addgene #12259), and 9 μg of psPAX2 (Addgene #12260) using Lipofectamine 2000 reagent (Invitrogen, 11668019). Lentivirus-containing media was harvested and filtered with a 0.45 µm PVDF filter (Millipore). Cells were transduced with the virus in the presence of 8 μg ml−1 polybrene. 48 h later, infected cells were selected with 1 µg ml−1 puromycin (Selleck, S7417) or 10 μg ml−1 blasticidin (Selleck, S7419).
Design and construction of the sgRNA library
The flanking sequence (30 nucleotides upstream and 30 nucleotides downstream of m6A-CLIP-seq1 identified single-nucleotide m6A sites) was extracted from the genome sequence according to the coordinate (GRCh37) of m6A loci for targetable analysis. Then, for each m6A site, we searched all possible sgRNAs with m6A sites in the editing window by sliding the editing window for every single nucleotide. To construct the sgRNA library, pooled oligonucleotides containing coding sequences of sgRNA and adapter were synthesized and cloned into the LentiGuide-BSD-dTomato vector by GENEWIZ, Inc. Lentiviral particles of the sgRNA library were produced, concentrated, and titered by GENEWIZ, Inc.
Base-editing screening
For FNLS-BE3 screening, A549 FNLS-BE3 cells were generated by transduced with the pLenti-FNLS-P2A-Puro lentivirus, and infected cells were selected by 1 μg ml−1 puromycin for 5 days. Then A549 FNLS-BE3 cells were infected by the sgRNA library lentiviral particles with a low MOI of 0.3 with the presence of 8 μg ml−1 of polybrene. 48 h after transduction, infected cells were selected by 10 μg ml−1 blasticidin for 5 days. Then, 5 × 106 cells were collected to measure the frequency of each sgRNA in the initial pool (referred to as day 0). The rest of the cells were continually cultured and passaged. 5 × 106 cells were collected on days 10, 20, and 30, respectively.
For FNLS-ABE7.10(AW) screening, H1 FNLS-ABE7.10(AW) hESCs were generated by transduced with the pLenti-FNLS-ABE7.10(AW)-P2A-Puro lentivirus and infected cells were selected by 1 μg ml−1 puromycin for 5 days. Then the whole population of H1 FNLS-ABE7.10(AW) hESCs was infected with the sgRNA library lentiviral particles with a low MOI of 0.3 with the presence of 8 μg ml−1 of polybrene. 48 h after transduction, infected cells were selected by 10 μg ml−1 blasticidin. After 5 days of selection, 5 × 106 cells were collected to measure the frequency of each sgRNA in the initial pool (referred to as day 0), and 5 × 106 cells were reseeded onto two Matrigel-coated 24-well plates for DE differentiation. 3 days later, differentiated DE cultures were stained with an APC-conjugated anti-CXCR4 antibody (Invitrogen, 17-9999-42) and ~6% cells with the lowest or highest CXCR4 expression were collected by FACS, respectively, according to the relative number of transduced sgRNAs vs. the number of cells (3–4 × 106 cells per sample). Three independent FNLS-ABE7.10(AW) screenings were performed from sgRNA library virus infection to FACS, and six sorted cell samples together with the one sample before DE induction were sequenced separately for subsequent data analyses.
High-throughput sequencing
Genomic DNA (gDNA) was extracted using the FastPure Cell/Tissue DNA Isolation Mini Kit (Vazyme, DC102) and DNA concentration was measured by Qubit using the Qubit™ dsDNA HS Assay Kit (Invitrogen). To generate sgRNA amplicons, we used a single-step PCR protocol which was adopted from the protocol published49. All the gDNA harvested from the screenings was used for PCR amplification in 50 µl PCR reactions. Each reaction consisted of 2.5 µg gDNA plus water, 25 μl NEBNext Ultra II Q5 PCR Master Mix, 1.25 μl 10 μM stagger forward primer, and 1.25 μl 10 μM barcoded reverse primer. PCR reactions were cycling as follow: initial denature 3 min at 98 °C; followed by 30 s denature at 98 °C, 10 s anneal at 63 °C, 25 s extension at 72 °C, for 23 cycles; and final extension for 2 min at 72 °C. PCR products were size-selected using VAHTS DNA Clean Beads (Vazyme) according to the manufacturer’s instructions and sequenced on a HiSeq2000 sequencer (Illumina).
High-throughput sequencing data analyses
Raw single-end reads were trimmed using Cutadapt to remove the constant flanking sequences of sgRNA sequences. Read counts of the sgRNAs were measured using the count command of MAGeCK27, the read count of each sgRNA was then normalized by the total reads mapped to all sgRNAs.
For BE3-based screening, log2-fold change (LFC) of normalized counts between day 30 and day 0 samples were calculated. The sgRNAs with absolute LFC > 1 were determined as significantly upregulated or downregulated sgRNAs.
For ABE-based screening, test command of MAGeCK27 was used to calculate the raw P values for the comparison between CXCR4− versus CXCR4+ sgRNA. 33 sgRNAs with three replicates averaged MAGeCK normalized read counts of CXCR4− or CXCR4+ samples <200 were removed for further analyses. The medium LFC (log2 (CXCR4+/CXCR4−)) of three replicates was used in the downstream analyses, which was calculated as the medium of the LFC of three individual replicates. sgRNA with p.high < 0.05 and medium LFC > 0.58 (fold change > 1.5) were used to determine the significantly enriched sgRNA in CXCR4+ population. sgRNA with p.low < 0.05 and medium LFC < −0.58 were used to determine the significantly enriched sgRNA in CXCR4− population. The locations and consequences of the sgRNA-induced mutations were predicted by VEP (Ensembl Variant Effect Predictor)24 with CDS has the highest priority across all isoforms, followed by 3′-UTR, 5′-UTR, intron, intergenic. sgRNAs predicted to induce missense mutations were filtered out in the identification of significantly enriched sgRNAs. Metascape50 was used to perform GO analysis for the genes targeted by the significant sgRNAs. Editing efficiencies of the sgRNAs were predicted by a machine learning algorithm BE-Hive51. FPKM (Fragments Per Kilobase of transcript per Million mapped reads) of H1 hESCs were directly obtained from the previous publication8. Heatmap was plotted using R package pheatmap. We used absolute medium LFC > 1 to identify the differential sgRNAs when comparing the hESCs with CXCR4+ or CXCR4− population.
The m6A peaks and gene expression of H1 hESCs were obtained directly from our previous publication8. The single-nucleotide m6A sites of m6A-CLIP-seq1 and miCLIP-seq20 were also obtained directly from the previous publications. The m6A peaks of A549 cells were identified using the published m6A-seq data52 with the method described in our previous publication8.
Flow cytometry and cell sorting
In brief, differentiated DE cultures were rinsed with DMEM/F-12 and dissociated with Accutase for 10–15 min at 37 °C. Cells were washed twice with ice-cold wash buffer (2% FBS in Dulbecco’s PBS, DPBS), resuspended in ice-cold blocking buffer (5% FBS in DPBS), and then incubated with the primary antibody CXCR4-APC for 1 h at 4 °C. Then cells were washed three times, resuspended with ice-cold wash buffer, and examined by a CytoFLEX S Flow Cytometer (Beckman Coulter) or sorted by the FACS MoFlo Astrios EQs system (Beckman Coulter). Data were analyzed by the FlowJo Software (FlowJo LCC). Cells incubated with the APC-conjugated isotype (Invitrogen, 17-4724-81) served as negative controls.
Genomic DNA extraction, PCR amplification, and Sanger sequencing
Genomic DNA was isolated using FastPure Cell/Tissue DNA Isolation Mini Kit (Vazyme, DC102), and genomic regions of interest were amplified by using a 2×Phanta Max Master Mix (Vazyme, P511) according to the manufacturer’s instructions. Purified DNA was sequenced by an ABI 3730xl DNA Analyzer (Applied Biosystems) and analyzed using SnapGene (GSL Biotech LLC). Primer sequences used for target amplification were provided in Supplementary Table 2.
Plasmid and siRNA transfection
To introduce point mutations into hESCs, 1 × 106 H1 hESCs were transfected with 2 μg base editor plasmid using Lipofectamine Stem transfection reagent (USA, Invitrogen, stem00008) following the manufacturer’s instructions. 24 hours after transfection, transfected cells were selected with 1 µg ml−1 puromycin for 48 h and reseeded onto Matrigel-coated 6 cm dishes at 5 × 103 cells per dish with the presence of CloneR (StemCell, 05888) following the manufacturer’s instructions. Single cell-derived clones were picked about 7 days later, amplified in culture, and then genotyped by Sanger sequencing of the gRNA-targeted site. For siRNA knockdown experiments, 1 × 105 H1 hESCs were transfected with the siRNA oligo (50 nM final siRNA concentrations) using DharmaFECT1 transfection reagents (Dharmacon, T-2001) following the manufacturer’s instructions. Knockdown efficiencies of siRNA-targeted genes were detected by real-time quantitative PCR (RT-qPCR). siRNA sequences used in this study were provided in Supplementary Table 3.
TRME cell line construction
NKX2-5eGFP/w hESCs were dissociated into single-cell suspension by Accutase and reseeded onto Matrigel-coated 24-well plates at a density of 1 × 105 cells/per well in E8 medium containing 10 μM Y-27632 and cells were co-nucleofected with TRME editor plasmid (dCas13a-ALKnes)35 and transposase plasmid at a mass ratio of 1000:1 using the Neon® Transfection System (Thermo Fisher Scientific). 24 h after transfection, cells were treated 1 μg ml−1 doxycycline with daily media change until stable colonies appeared. Then, cells were transduced with crRNA-expressing lentiviruses, and cells with both GFP and mCherry expression were sorted by FACS and expanded for further experiments.
RNA binding protein immunoprecipitation (RIP)
EZ-Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Sigma-Aldrich, 17-701) following the manufacturer’s instructions. The anti-YTHDF2 (Proteintech, 24744-1-AP, 5 μg per sample) antibody was used for RIP. The input and IP RNA of each sample was purified and evaluated through RT-qPCR.
Western blot
Cells were lysed in RIPA buffer (Cell Signaling Technology, 9806) supplemented with 1 mM PMSF and Proteinase inhibitor (Roche, 4693132001). 30 μg protein per lane was fractionated on 6–12% SDS–PAGE and transferred to the PVDF membrane (GE Healthcare Life Sciences). Membranes were blocked in the blocking buffer (DPBS, supplemented with 5% skimmed milk, 0.1% Tween 20) for 1 h at RT. Membranes were then incubated with the primary antibodies including anti-Cas9 (1:3000, Diagenode, C15310258), and anti-β-actin (1:1000, 4A Biotech, 4ab080291) in the antibody dilution buffer (Solarbio) overnight at 4 °C. Then membranes were washed by DPBS containing 0.1% Tween-20, incubated with HRP-conjugated secondary antibody (Beyotime Biotechnology, 1:1000) in antibody dilution buffer for 1 h at RT, and visualized by the Clarity™ Western ECL Substrate (Bio-Rad).
Alkaline Phosphatase (AP) staining
AP stainings were performed by using the Alkaline Phosphatase Detection Kit (Sigma-Aldrich, SCR004) following the manufacturer’s instructions.
Immunofluorescence
For immunofluorescence, cells were fixed in 4% paraformaldehyde (Solarbio) for 30 min at RT and washed with 0.3 M glycine in DPBS. Cells were then blocked and permeabilized in the permeabilization buffer (DPBS supplemented with 8% donkey serum, 8% goat serum, and 0.3% Triton X-100) at RT for 1 h. Cells were then stained with the primary antibody in primary antibody statin buffer (DPBS supplemented with 1% BSA, 1% goat serum, and 0.25% Triton-X) at 4 °C overnight. After washing with DPBS containing 0.1% BSA and 0.1% Triton X-100, cells were stained with fluorescent secondary antibody in secondary antibody buffer (DPBS supplemented with 0.05% Triton X-100 and 1% BSA) at RT for 1 h and analyzed using the Operetta CLS system (Perkin Elmer) in the same settings. Primary antibodies included SOX17 (1:200, R&D, AF1924), FOXA2 (1:200, Cell Signaling Technology, 8186S), SOX2 (1:200, Abcam, ab79351), NANOG (1:200, Cell Signaling Technology, 3580S), OCT4 (1:200, Santa Cruz, sc-5279), SOX1 (1:200, Cell Signaling Technology, 4194S), Brachury (1:100, R&D, AF2085), Ki67 (1:100, BD Biosciences, 550609). Secondary antibodies used were Alexa488-conjugated donkey anti-rabbit (1:800, Invitrogen, A21206), Alexa488-conjugated goat anti-mouse (1:800, Invitrogen, A21121), Alexa647-conjugated goat anti-mouse (1:800, Invitrogen, A21242), Alexa555-conjugated goat anti-mouse (1:800, Invitrogen, A21127), Alexa594-conjugated donkey anti-goat (1:800, Invitrogen, A11058).
SELECT for detection of m6A
1 μg total RNA from the control group or expression level normalized RNA from the experimental group was mixed with 40 nM up primer, 40 nM down primer, 5 μM dNTP, 1.7 μl 10× CutSmart buffer (New England Biolabs, B7204), DEPC H2O to the final volume 17 μl. The following temperature annealed the mixture of RNA and primers: 1 min at 90 °C, 1 min at 80 °C, 1 min at 70 °C, 1 min at 60 °C, 1 min at 50 °C and 6 min at 40 °C. Subsequently, a 3 μl of enzyme mixture containing 0.01 U Bst 2.0 DNA polymerase (New England Biolabs, M0537), 0.5 U SplintR ligase (New England Biolabs, M0375), and 10 nmol ATP was added in the annealing products. The final mixture was incubated at 40 °C for 20 min, heat inactivation at 80 °C for 20 min and stored at 4 °C. qPCR was then performed in QuantStudioTM 7 Flex Real-Time PCR System (Applied Biosystems) using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711). Relative SELECT products between the experimental group and control group were calculated using the 2−∆∆Ct method. Primers used in the SELECT assays were provided in Supplementary Table 4.
RNA stability assay and RT-qPCR
For RNA stability assays, cells were treated with transcription inhibitor actinomycin D (Sigma, A9415) at 5 μg ml−1. RNA samples were collected at various time points and analyzed by RT-qPCR. For RT-qPCR, total RNA was extracted using the FastPure Cell/Tissue Total RNA Isolation Kit (Vazyme, RC101). 1 μg of DNA-free total RNA was then reverse-transcribed using HiScript II Q RT SuperMix for qPCR (Vazyme, R223). RT-qPCR was carried out using the ChamQ Universal SYBR qPCR Master Mix (Vazyme, Q711) and performed in a QuantStudioTM 7 Flex Real-Time PCR System (Applied Biosystems). 18S and GAPDH were used as the reference gene in RNA stability assay and gene expression assay, respectively. Relative fold-change was calculated using the 2−∆∆Ct method. The RT-qPCR primer sequences used in this study were provided in Supplementary Table 5.
Statistics and reproducibility
Graphs and statistical analyses were carried out using GraphPad Prism 8.0 (GraphPad Software Inc.). Statistical significance of differences was estimated by two-tailed unpaired Student’s t-test for two groups comparisons and one-way ANOVA with Tukey’s post hoc test for multiple groups comparisons. A P value of <0.05 was determined as statistically significant. Data were presented as means ± standard deviation (SD) quantified from at least three biological repeats unless otherwise stated. The immunoblot (Supplementary Fig. 1b, f) and immunostaining (Supplementary Figs. 1c, g, 3, 8b, c, 9b–d) experiments were performed at least three independent times with similar results.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
