Screening of cancer cell lines for SNVs as prerequisite for ASEE by Sanger sequencing

A 482 bp long region in the TERT promoter region (chr5:1,295,043–1,295,524, hg19) was investigated in order to detect heterozygous SNVs (activating mutations or SNPs) that could serve as a target for ASEE. To this end, DNA from 87 cell lines was extracted using the FlexiGene DNA Kit (QIAGEN, Venlo, Netherlands). Cell lines were selected according to availability and chances for successful transfection. Most of the cell lines screened in this study were lymphomas of B-cell origin in addition to adherent cell lines (hepatocellular carcinoma and lung adenocarcinoma). All DNA samples used in this study were authenticated using the GenePrint 10 System (Promega, Madison, Wisconsin, USA) according to the manufacturer’s instructions and cell lines were tested for mycoplasma contamination using the MycoSPY-RCR Mycoplasma Test Kit (Biontex, München, Germany) following the manufacturer’s protocol. Cell lines that were transduced or transfected were excluded from the screening. A total of 50 ng genomic DNA were used to amplify the TERT promoter region with the AmpliTaq Gold® 360 PCR Master Mix (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and primers that contained universal tags for sequencing (Additional file 1: Table S3). The PCR was performed in a Labcycler Basic 011–103 (Sensoquest, Göttingen, Germany) and the conditions applied for TERT promoter region amplification were as follows: 10 min at 95 °C, 40 cycles of 5 s at 95 °C, 30 s at 61 °C, 30 s at 72 °C and finally 5 min at 72 °C. The PCR products were purified using the AMPure XP magnetic beads (Beckman Coulter Life Sciences, Brea, California, USA) following the manufacturer’s instructions. The purified PCR products subsequently underwent a sequencing reaction using the BigDye™ Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems™, Waltham, Massachusetts, USA) according to the manufacturer’s instructions. The sequencing reactions were purified with Agencourt CleanSEQ beads (Beckman Coulter Life Sciences, Brea, California, USA) according to the manufacturer’s protocol. Sequencing was performed on a 3500xL Dx Genetic Analyzer (Applied Biosystems, Waltham, Massachusetts, USA). The peak calling was done with the Sequencing Analysis Software version 5.4 (Applied Biosystems, Waltham, Massachusetts, USA) and finally the data were visualized with the Sequence Scanner Software 2 version 2.0 (Applied Biosystems, Waltham, Massachusetts, USA). From the 87 cell lines, 77 were evaluable for both strands. Only the sequence variants that were present in both DNA strands were considered.

Plasmid propagation and validation by Sanger sequencing

In order to generate sufficient quantity of vectors for the transfection experiments, 5-alpha Competent E. coli bacteria (NEB, Ipswich, Massachusetts, USA) were transformed following the manufacturer’s guidelines. The dCas9 and DNMT expression vectors used in this study have been described before [39]. For positive clone selection after transformation, the bacteria grew in Luria broth base (LB), Miller′s modified medium (Sigma-Aldrich, St. Louis, Missouri, USA) supplemented with 100 mg/ml Ampicillin or 50 mg/ml Kanamycin (AppliChem, Darmstadt, Germany) according to the resistance cassette of each vector (Additional file 1: Table S2). Plasmid DNA was isolated using the NucleoBond Xtra Midi kit (Macherey Nagel, Düren, Germany) according to manufacturer’s instructions and measured with the Qubit dsDNA BR-Assay-Kit (Invitrogen, Waltham, Massachusetts, USA).

For plasmid validation, 50 ng of extracted plasmid was used to amplify unique parts of each plasmid with the AmpliTaq Gold® 360 PCR Master Mix (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and primers that were specific for each vector (Additional file 1: Table S3). The PCR was performed in a Labcycler Basic 011–103 (Sensoquest, Göttingen, Germany) and the conditions applied were as follows: 15 min at 98 °C, 40 cycles of 30 s at 95 °C, 30 s at 61 °C, 30 s at 72 °C and finally 10 min at 72 °C. The PCR products were purified using the AMPure XP magnetic beads (Beckman Coulter Life Sciences, Brea, California, USA) following the manufacturer’s instructions and a sequencing reaction was performed as described above. Sequencing reactions were purified with Agencourt CleanSEQ beads (Beckman Coulter Life Sciences, Brea, California, USA) according to the manufacturer’s protocol and eventually sequenced using the 3500xL Dx Genetic Analyzer (Applied Biosystems, Waltham, Massachusetts, USA). The peak calling was done with the help of the Sequencing Analysis Software version 5.4 (Applied Biosystems, Waltham, Massachusetts, USA) and finally the data were visualized with the Sequence Scanner Software 2 version 2.0 (Applied Biosystems, Waltham, Massachusetts, USA).

Cell culture

Hep-G2 hepatocellular carcinoma cells and HEK293 cells were cultivated in RPMI 1640 Medium 1X (GIBCO Life Technologies, Carlsbad, California, USA) supplemented with 10% Fetal Bovine Serum (FBS, GIBCO Life Technologies, Carlsbad, California, USA) and 1% L-Analyl-L-Glutamine (Biochrom, Merck Millipore, Burlington, Massachusetts, USA). A-549 lung carcinoma cell line was cultivated in DMEM high glucose medium (GIBCO Life Technologies, Carlsbad, California, USA) supplemented with 10% FBS (GIBCO Life Technologies, Carlsbad, California, USA). Every 3–4 days we detached the cells from the flask bottom using diluted Trypsin 2,5% w/v in PBS w/o Ca2 + (Biochrom, Merck Millipore, Burlington, Massachusetts, USA). All cells were incubated at 37 °C and 5% CO2 in a Heracell™ 240i CO2 Incubator (Thermo Fisher Scientific, Waltham, Massachusetts, USA).

Co-transfection experiments and fluorescence activated cell sorting (FACS)

Transient transfection experiments were performed on Hep-G2 cells 24 h after seeding 850,000 cells/well in a 6-well plate (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The Lipofectamine 3000 reagent (Thermo Fisher Scientific, Waltham, Massachusetts, USA) was utilized following the manufacturer’s instructions. The following vectors, each expressing a different fluorescent marker, were used for the co-transfection of Hep-G2: dCas9-10X SunTag system (10,032 bp; TagBFP expressing), a DNMT3A-3L R887E mutant with reduced DNA affinity (6,200 bp; sfGFP expressing) [39] and a sgRNA that targets either the wildtype or the mutated allele at the C228T position (5,097 bp; DsRed expressing), hereafter termed C228T-wt and C228T-mut respectively (Additional file 1: Table S2). Because the NGG site was located on the opposite strand of the C228T mutation, the sgRNAs were designed to target the G allele (which corresponds to C wildtype allele) and the A allele (which corresponds to T mutated allele). Development, optimization and validation of these constructs are described elsewhere [12]. Upon expression of the dCas9-10X SunTag fused protein, this complex can recruit up to 10 active subunits of DNMT3A-3L [39]. All vectors were validated by Sanger sequencing prior to the transfection experiments. For the A-549 cell line transfection, 500,000 cells/well were seeded in four 6-well plates (Thermo Fisher Scientific, Waltham, Massachusetts, USA) 24 h prior to the transfection experiment. The same vectors as in the Hep-G2 experiments were employed with a different sgRNA that targets the alternative G allele of the SNP rs2853669, hereafter termed rs2853669-alt. Additionally, the non-cancer cell line HEK293 was used to establish transfection experiments. A total of 250,000 HEK293 cells/well were seeded in three wells (three technical replicates) of a 6-well plate (Thermo Fisher Scientific, Waltham, Massachusetts, USA) 24 h prior to the transfection experiment. The FuGENE HD transfection reagent was used for HEK293 experiments according to the manufacturer’s instructions (Promega, Madison, Wisconsin, USA). Since this cell line is homozygous for the wildtype C allele at C228T position, C228T-wt was used, thereby addressing both alleles. The cells were harvested by trypsinization 72 h post-transfection and filtered through a 35 µm cell strainer cap (FALCON, Corning, New York, USA). Cells that contained all three components (triple-positive) were isolated by FACS with a BD FACSAria™ III Cell Sorter (BD Biosciences, New Jersey, USA). The ranges of triple-positive cells sorted were: 45,000–295,000 for HEK293, 69,000–400,000 for Hep-G2 and 60,000–147,000 for A-549. These cells were later on handled for downstream analysis including DNA isolation, bisulfite conversion, library generation and sequencing.

Bisulfite treatment and targeted bisulfite sequencing (BS)

After isolation of triple-positive Hep-G2, A-549 and HEK293 cells, genomic DNA was extracted using the Quick-DNA/RNA™ Microprep Plus Kit (Zymo Research, Irvine, California, USA) following the instructions of the manual. A total of 1,000 ng of genomic DNA was used for bisulfite conversion and purification with the EpiTect Bisulfite Kit (QIAGEN, Venlo, Netherlands) according to the manufacturer’s instructions. The purified bisulfite converted DNA was eluted in a final volume of 20 μL. The following regions of the TERT promoter were screened: BS1 (chr5:1,295,112–1,295,401, hg19, 290 bp, 30 CpGs) used in all three cell lines and BS2 (chr5:1,295,290–1,295,642, hg19, 353 bp, 35 CpGs) used only in A-549 cells. The last nine CpGs of the BS1 assay overlapped with the first nine CpGs of the BS2 assay (Fig. 1). For the PCR amplification of the TERT promoter region, 1 μL of bisulfite converted DNA was set up, 12.5 μL of the PyroMark PCR mix from the PyroMark PCR Kit (QIAGEN, Venlo, Netherlands) and 10 pmol of each primer containing the overhang adapters from the 16S Metagenomic Sequencing Library Preparation protocol (Illumina, San Diego, California, USA). The PCR was conducted in a Labcycler Basic 011–103 (Sensoquest, Göttingen, Germany) with the following conditions for TERT BS1: 15 min at 98 °C, 7 cycles of 30 s at 98 °C, 30 s at 58–55 °C (dT -0,5/cycle), 30 s at 72 °C, 38 cycles of 30 s at 98 °C, 30 s at 55 °C, 45 s at 72 °C and finally 10 min at 72 °C. For TERT BS2 assay the following PCR conditions were used: 15 min at 95 °C, 35 cycles of 30 s at 94 °C, 30 s at 55 °C, 30 s at 72 °C and finally 10 min at 72 °C.

The VEGFA promoter locus (chr6:43,738,171–43,738,372, hg19, 202 bp, 12 CpGs) was selected as an off-target DNA methylation control since it has been shown that it is a sensitive region whose CpG island is easily methylated by epigenome editing systems [39]. The PCR conditions applied for VEGFA promoter amplification were as follows: 15 min at 95 °C, 45 cycles of 30 s at 94 °C, 30 s at 50 °C, 30 s at 72 °C and finally 10 min at 72 °C. The PCR products were purified with AMPure XP magnetic beads (Beckman Coulter Life Sciences, Brea, California, USA). Indexed PCR products were generated using the IDT for Illumina UD Indexes Plate Set A (Illumina, San Diego, California, USA) and the EPM Enhanced PCR Mix (Illumina, San Diego, California, USA) in a Biometra thermocycler (Jena Analytik, Jena, Germany) with the following PCR conditions: 3 min at 72 °C, 3 min at 98 °C, 9 cycles of 20 s at 98 °C, 30 s at 60 °C, 1 min at 72 °C and finally 3 min at 72 °C. Thereafter, the indexed PCR products were purified with the same magnetic beads mentioned above and 100 ng of each purified library were used to create pools for NGS. Targeted bisulfite sequencing was performed on an Illumina MiSeq sequencer (Illumina, San Diego, California, USA) using a paired-end 2 × 300 cycles protocol. The bisulfite conversion rate was calculated based on the ratio of total unmethylated C’s outside of CpG context to the sum of total methylated and unmethylated C’s outside of CpG context.

To control for amplification bias of one allele, unique molecular identifiers (UMIs) were added to the forward primer sequence. These UMIs allow quantification of the original DNA fragments among the final sequencing reads (Additional file 1: Table S3). The length of each UMI is six nucleotides, resulting in a maximum of 4,096 unique UMI per sequencing experiment. Therefore, it is expected that several UMIs are found more than once if thousands (i.e., > 4,096) of reads per allele were sequenced. In Hep-G2 and A-549 cell lines, we observed all possible UMI sequences (minimum complexity of 4,096 molecules) and we did not observe higher frequency than 50 reads per UMI. Assessment of UMIs in Hep-G2 and A-549 cells showed no overrepresentation of single UMI groups for each allele, indicating a negligible impact of clonal PCR amplification products on measured DNA methylation levels (Additional file 1: Figures S4–S9).

Targeted DNA methylation data analysis

The targeted BS data were reviewed and corrected with the fastQC [55] and cutadapt [56] tools for adapter content and sequencing quality. A read was kept for processing when its minimum length was 100 nucleotides and the minimum quality was 25. The sequencing quality values decreased near the end of the reads, as is typical for Illumina sequencing, demonstrating the anticipated accumulation of low sequencing quality scores, particularly in mate 2. A total of 120 nucleotides were automatically removed from the end of mate 2 because they did not meet the strict quality standards (quality scores ≥ 25), which are applied to assure good data quality and reliable base calls. The reads were then aligned against a gene-specific reference (Supplementary Methods) using BISMARK [57] with bowtie2 [58] and the non-directional protocol. Additionally, to account for alignment errors and enable later deconvolution of allele-specific DNA methylation rates, the allele-specific locations were N-masked. The alignments were then divided using SNPsplit [59] which uses the annotated SNPs to discriminate between the two alleles. Finally, DNA methylation calling was performed on the split alignments using BISMARK's methylation extractor function (with the no_overlap and comprehensive parameters). Only the samples with a minimum number of 500 reads after DNA methylation calling were included in further analysis. When calculating the average DNA methylation gain after ASEE, the CpG sites included in the sgRNA binding site were excluded since no DNA methylation can occur at this place. The DNA methylation gain was estimated according to the samples transfected with the scrambled sgRNA, when these were available. Alternatively, the untreated samples were used to compute the difference between treated and reference samples.

At least 40,000 raw reads were obtained for Hep-G2 after NGS and at least 21,000 passed the filters of quality control and were processed for downstream analysis of the TERT promoter region covered by BS1 (Fig. 1, Additional file 1: Table S4). At least 70,000 raw reads were acquired for A-549 samples after NGS and at least 13,000 passed the filters of quality control and were processed for downstream analysis of the TERT promoter region covered by BS1 and BS2 (Fig. 1, Additional file 1: Table S4). The DNA methylation analysis of the VEGFA locus was performed with the same workflow described above, omitting the splitting of the aligned reads into two different alleles. The DNA methylation analysis of BS1 in HEK293 was performed in the same way as the analysis for the VEGFA locus since this cell line lacked a TERT promoter sequence variant. At least 12,500 raw reads were obtained for HEK293 after NGS and at least 10,300 passed the filters of quality control and were processed for downstream analysis of the TERT promoter region covered by BS1 (Fig. 1, Additional file 1: Table S4). The average bisulfite conversion rate was ~ 99% in all analyzed Hep-G2 and HEK293 samples (Additional file 1: Table S4). The average bisulfite conversion rate of A-549 samples was ~ 99% for BS1 and VEGFA assays (Additional file 1: Table S4).

HTG transcriptome analysis

For functional readout of triple-positive cells, RNA was extracted using the Quick-DNA/RNA™ Microprep Plus Kit (Zymo Research, Irvine, California, USA) according to the manufacturer’s protocol. HTG Transcriptome Panel (2 × 8) assay which covers the vast majority of the human mRNA transcripts including isoforms with 19,616 probes (HTG Molecular Diagnostics, Inc., Tuscon, Arizona, USA) required 70 ng of extracted RNA. After target protection, 4 μL was taken from each sample for library preparation (addition of adapters and molecular barcodes) with the HTG EdgeSeq (Illumina) Tag Pack (HTG Molecular Diagnostics, Inc., Tuscon, Arizona, USA) and the OneTaq® Hot Start 2X Master Mix in GC Buffer (NEB, Ipswich, Massachusetts, USA). The indexing PCR was performed in a Labcycler Basic 011–103 (Sensoquest, Göttingen, Germany) with the following PCR conditions: 4 min at 95 °C, 19 cycles of 15 s at 95 °C, 45 s at 56 °C, 45 s at 68 °C and finally 10 min at 68 °C. After library purification with AMPure XP magnetic beads (Beckman Coulter Life Sciences, Brea, California, USA) according to HTG instructions, the purified libraries were quantified using the KAPA Library Quant Kit (Illumina) Universal qPCR mix (Roche, Basel, Switzerland) and the LightCycler 480 II (Roche, Basel, Switzerland). The libraries were subsequently sequenced with a NextSeq sequencer (Illumina, San Diego, California, USA) using the Illumina NextSeq 500/550 High output v2.5 Reagent Kit (75 cycles) (Illumina, San Diego, California, USA). At least 19,647,280 raw reads were obtained for each sample (Additional file 1: Table S5). Quality control was done using the HTG EdgeSeq Reveal Software (HTG Molecular Diagnostics, Inc., Tuscon, Arizona, USA). Mean and standard deviation of Log2 transformed CPM values were plotted to show gene expression for TERT and VEGFA. All CPM values used in this study are based on at least three independent experiments and can be found in Additional file 1: Table S5.

Zombie NIR™ fixable viability experiments on Hep-G2 and A-549 cell lines and flow cytometry analysis

In order to assess the effect of ASEE on cell viability, Hep-G2 were co-transfected using all available TERT sgRNAs (C228T-mut, C228T-wt and rs2853669-alt) and the scrambled sgRNA. Cell viability experiments were performed using the Zombie NIR™ Fixable Viability Kit (BioLegend, San Diego, California, USA) following the manufacturer’s instructions. The cells were subsequently analyzed 48, 72 and 96 h post-transfection with the BD LSRFortessa™ Flow Cytometer (BD Biosciences, New Jersey, USA) and the BD FACSDiva™ Software (BD Biosciences, New Jersey, USA). The percentage of dead cells within the triple-positive, the single-positive and triple-negative population was calculated for each sample based on the Zombie staining fluorescence. For the A-549 cell line, co-transfection and cell viability experiments were performed as mentioned above with rs2853669-alt and scrambled sgRNAs. The percentage of dead cells in the triple-positive and triple-negative population was evaluated within each sample by calculating the average and standard deviation based on three technical replicates.

Statistical analysis

For comparison of DNA methylation between each CpG site among the different samples, a 2-tailed T test for samples with same variance was performed and the P values were Bonferroni corrected. When the q value was > 1, it was considered as 1 for visualization purposes in the respective figures. Corrected P values (q value) lower than 0.05 were considered statistically significant and were shown in the respective figures. All comparisons conducted in this study were done by using the scrambled sgRNA transfected samples as reference when these were available. Otherwise, untreated samples were taken as reference. For comparison between time points in the viability experiments, a 2-tailed T test for samples with same variance was performed.