Cell culture and reagents
Wild type HEK293 cells (HEK293T) were obtained from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China) and detected to be negative for mycoplasma contamination using the Myco-Blue mycoplasma detector (Vazyme; Nanjing, Jiangsu, China). Cells were cultured in high glucose DMEM supplemented with 10% FBS, incubated at 37ºC with 5% CO2 in a humidified cell incubator (Thermo Fisher Scientific; OH, USA). The plasmid pX330 carrying CRISPR/Cas9 system was kindly provided by Dr. Feng Zhang (MIT) [19]. Competent cells of the E. coli strains DH5α were purchased from Microgene (Shanghai, China). All media and supplements were purchased from Gibco (Thermo Fisher Scientific; Waltham, MA, USA). Cell growth and viability were monitored with a cell counter (Countstar; Shanghai, China).
SgRNA design and DNMT3A disruptive vector construction
Two sgRNAs targeting exon 19 of DNMT3A (GeneBank ID 806904736) were designed using the web tool provided by Dr. Zhang’s lab (http://crispr.mit.edu) as shown in Fig. 1. To construct the sgRNA plasmids, single strand primers were designed and synthesized as sgRNA1-forward: 5’-CACCGCATGATGCGCGGCCCAAGG-3’, sgRNA1-reverse 5’-AAACCCTTGGGCCGCGCATCATGC-3’, sgRNA2-forward 5’-CACCGCTCACTAATGGCTTCTACCT-3’ and sgRNA1-reverse 5’-AAACAGGTAGAAGCCATTAGTGAGC-3’. Each pair of primers were annealed to generate double-stranded cDNA, phosphorylated by T4 polynucleotide kinase at the 5’ ends (NEB, Ipswich, MA) at 37°C for 30 min, and further ligated into BbsI digested pX330 plasmids by T4 DNA ligase (Takara; Kusatsu, Shiga, Japan). The ligate was transformed to DH5α competent cells for culture overnight. Then the grown clones were selected for sequencing to get the right constructed plasmids pX330-sgRNA1 and pX330-sgRNA2.
Transfection of HEK293 cells
HEK293 cells were seeded at 2×105 cells/well into 12-well plate one day prior to transfection. When reached 70-80% confluence, the cells were co-transfected with pX330-sgRNA1 and pX330-sgRNA2 at a molar ratio of 1:1, since it was reported that double sgRNAs could result in higher editing efficiency than single one [20]. The transfection was performed using Lipofectamine 2000 reagent (Invitrogen, CA, USA) according to manufacturer’s instructions.
DNMT3A knockout clones selection
HEK293 cell pool transfected with pX330-sgRNAs were seeded into 96-well plates at the density of 0.5 cell per well for limiting dilution. After about ten days’ incubation, the plates were examined for single cell clones under microscope. When grew to about 80% confluent in the well, the clones would be detached for subpopulation and the genomic DNA was extracted with QuickExtract DNA extraction solution (Epicenter; MD, USA) for PCR verification, using primers HEK293-DNMT3A-For (5’-GTACCATCCTGTCCCCTCCAC-3’) and HEK293-DNMT3A-Rev (5’-GGCTCAGGGTTAAACGGGGA-3’), which can amplify a 798 bp fragment for HEK293 wild-type cells. By sequencing the amplified fragments, the clone with disrupted DNMT3A was selected and designated to be DNMT3A KO cell line.
DNMT3A knockout cells proliferation curve
DNMT3A KO and WT cells were cultured and seeded at 3×104 cells/well into 12-well plate. Cells were counted every 24 h for consecutive 6 days. And cell proliferation curves were compared between the two cell lines.
Western blot analysis
DNMT3A KO and WT HEK293 cells of 1 ×106 were washed with PBS, lysed using 100 μL RAPA lysis buffer containing protease inhibitors cocktail (Roche; Penzberg, Germany), and separated by a 10% SDS-PAGE. After transferring onto a 0.45 μm PVDF membranes, immunoblotting was performed. For detection of DNMT3A deficiency, primary mouse monoclonal antibody against GAPDH (Sangon; Shanghai, China) and polyclonal rabbit-anti-human DNMT3A (Sangon) were used at 1:1000 dilution. For detection of MAPK and PI3K-Akt pathways, primary monoclonal antibodies against human Erk (137F5; Cell Signaling Technology; Danvers, MA, USA), phosphor-Erk (197G2; Cell Signaling Technology), JNK (D-2; Santa Cruz; Dallas, TX, USA), phosphor-JNK (G9; Cell Signaling Technology), Akt (11E7; Cell Signaling Technology), and phosphor-Akt (244F9; Cell Signaling Technology) were used. HRP-conjugated anti-mouse IgG or anti-rabbit IgG antibodies (Jackson ImmunoResearch; PA, USA) were used for secondary antibodies. Signals were detected with enhanced chemiluminescence (Millipore; MA, USA) and visualized with a gel imaging system (Tanon; Shanghai, China).
Genome-wide DNA methylation analysis by UPLC-ESI-MS/MS
Genomic DNA of cells were extracted by AxyPrep Kit (Axygen; Hangzhou, Zhejiang, China) and RNase A was added to remove RNA. Then the genomic DNA was hydrolyzed by DNase I at 37 ºC for 1 h, denatured at 100 ºC for 3 min, and immediately cooled down on ice for 10 min, then treated with Nuclease P1 at 37ºC for 16 h, followed by treatment of alkaline phosphatase at 37 ºC for 2 h. The nucleotides were stored at -20 ºC before UPLC-ESI-MS-MS detection.
Acquity UPLC (Waters,USA) coupled with Triple Quad™ 5500 mass spectrometry (Sciex,USA) was used to quantitatively analyze m5dC and dG. UPLC-ESI-MS/MS method was established to evaluate DNA methylation status of genome [21]. Reference nucleotide standards of A, G, T, C, dA, dG, dC, U and m5dC were purchased from Sigma (Sigma Aldrich, St. Louis, MO, USA) and dissolved in H2O to a final concentration of 1.0 mg/ml. UPLC and electronic spray were used to separate and detect the standards at multiple reaction monitoring (MRM) mode. The m5dC(m/z 241.9→126.3)and dG (m/z 268.1→152.3) were chosen as parent and child ion pairs for quantitative detection. The CE voltage of both m5dC and dG was 15 eV, and the DP voltage was 40 V, respectively. Standard curves of m5dC and dG were first graphed and the level of cytosine methylation was calculated as (m5dC/dG) x 100%.
RNA-seq to reveal transcriptional response to DNMT3A deficiency
Total RNA was extracted from 106 of DNMT3A KO or WT cells. Oligo(dT) magnetic beads were used to enrich mRNA. CDNA was obtained using Illumina TruseqTM RNA sample prep Kit, and pair-end sequencing (insert size = 300 bp, read length = 150 bp) was performed according to the standard protocol of Novaseq 6000 (Illumina, CA, USA). Raw sequencing reads were filtered to include only high quality reads in downstream analysis: 1) clip adapter sequence from reads, and remove reads with no insertion; 2) clip 3’ low quality bases (Phred quality < 20), and remove the whole read if there exists a single base with Phred quality < 10; 3) remove the reads that have more than 10% ambiguous bases (N); remove the reads that are shorter than 20 bp after clipping. The filtered reads were aligned to human transcriptome (build GRCh38) by TopHat [22]. PCR duplicates were marked and ignored in downstream analysis. All the data were deposited into the open-access Genome Sequence Archive (gsa.big.ac.cn) under accession no. CRA002294.
The read count data of DNMT3A KO and WT cells was analyzed by Cufflink software to identify the differential gene expression induced by DNMT3A deficiency [23]. We used FPKM (Fragments Per Kilobase of exon model per Million mapped reads) to estimate genes expression levels. False discovery rate (FDR) p values were calculated using the method proposed by Benjamini and Hochberg (1995) to correct for multiple testing. Differentially expressed genes in DNMT3A KO cells were identified by FDR p value ≤ 0.05 and absolute logarithm of fold change (log2FC) ≥ 2.
KEGG pathway analysis of differentially expressed genes
For the purpose of pathway enrichment analysis, we defined differential expression using a loose definition (FDR p value ≤ 0.05 and absolute log2FC ≥ 1). The Ensembl IDs of differentially expressed genes were analyzed by KOBAS (http://kobas.cbi.pku.edu.cn) for KEGG pathway enrichment. The pathways with FDR p value≤ 0.05 were considered significantly differentially expressed.
Bisulfite DNA analysis and quantitative PCR verification of DNMT3A regulated genes
DNMT3A is responsible for the de novo methylation of multiple genes, and its mutation can lead to demethylation of promoter CpG and thus elevate gene expression at the transcript level, which further up-regulate or down-regulate related downstream genes indirectly. Therefore, from the gene pool which transcript level was interfered by DNMT3A knockout as determined by RNA-seq, we selected 3 representative genes to verify by bisulfite DNA analysis as well as quantitative PCR: RUNX1, IQGAP3, and DNMT3B. RUNX1 is known to be regulated by DNMT3A in hematopoietic carcinogenesis [24]. IQGAP3 is a scaffolding protein that is involved in cancer cells proliferation, and with no correlation with DNA methyltransferases reported before [25]. All 3 genes were hot studied in malignancy development and helpful to understand the functions of DNMT3A.
DNA methylation status of selected genes were analyzed by bisulfite sequencing PCR (BSP). Genomic DNA was extracted with an Axygen Genomic DNA Miniprep Kit (San Francisco, CA, USA), and 0.5 μg of DNA was modified through bisulfite treatment using a Bisuldream® — Methylation Universal kit (Miozyme; Shanghai, China). Bisulfite-PCR of the genes promoter regions (Table S1) was performed using the following specific primers: RUNX1 forward: 5’- TTTTTAGGTTTTAAAATATTTGTGAGTTGT-3’, RUNX1 reverse: 5’- CACCTACCCTCCCCCAAACTATAC-3’, IAGAP3 forward: 5’- GTAGAAAAGGAGTTTGGAAGGAATAAGA-3’, IQGAP3 reverse: 5’- ACTCACAAACTACCCAACCTAAACC-3’, and DNMT3B forward 5’- TTAAAGTAGGATGATAGGTAGGGGTAT-3’, DNMT3B reverse: 5’- CCCTAAAAAATCAAAAACCCTAAAC-3’. The amplified fragments were inserted into pMD19-T vectors (Takara; Tokyo, Japan), and 10-15 clones for each gene were selected for sequencing. The results were analyzed by a web-based quantification tool for methylation analysis (http://quma.cdb.riken.jp).
To detect the transcription levels of the above selected three genes, the DNMT3A KO and WT HEK293 cells were cultured and RNA samples were extracted using Direct-zol RNA kit (Zymo Research; Irvine, CA, USA). Then cDNA was synthesized according to the protocol of the RT-PCR kit (Takara; Kusatsu, Shiga, Japan) and used as templates for quantitative PCR. The primers were designed using Primer Primier 5.0 (Premier Biosoft; Palo Alto, CA, USA) according to published sequences (NCBI Accession number: D43967 for RUNX1, AB105103 for IQGAP3, AF156487 for DNMT3B and M33197 for GAPDH). The following sequences for primers were synthesized (Sangon Biotech; Shanghai, China) as RUNX1 forward: 5' – TCTCTTCCTCTATCTTCCA– 3’, RUNX1 reverse: 5'–GGTATGTGCTATCTGCTTA–3’; IQGAP3 forward: 5'–GACCACTACCTAACTCAG–3’, IQGAP3 reverse 5'–GCATCATCAACAACTTCTA–3’; DNMT3B forward: 5’- GGCAAGTTCTCCGAGGTCTCTG-3’, DNMT3B reverse: 5’-TGGTACATGGCTTTTCGATAGGA-3’; and GAPDH forward: 5'–CTCTGGTAAAGTGGATATTGT–3’, GAPDH reverse: 5'– GGTGGAATCATATTGGAACA–3’). The real-time PCR procedures were performed with 25 μL PCR reaction systems including 12.5 μL qPCR Mix (Toyobo; Osaka, Japan), 0.4 μM of each primer, and 1 μL template cDNA by thermocycler (StepOnePlus; ThermoFisher, USA). The delta-delta threshold cycle (CT) method was used to calculate relative copy numbers of targeted genes related to housekeeping gene GAPDH.