Methylated DNA Immunoprecipitation Sequencing (MeDIP-seq) of 5-Hydroxymethylcytosine-Independent 5mC Enrichment in Human Glioblastoma U87-MG DNA

Two sources (U87-MG, normal human astrocyte gDNA), biological triplicate (n=3). Each MeDIP reaction uses 0.5-1 µg sonicated DNA with a matched input (no-antibody) control retained from the same fragmented pool. Methylated and unmethylated spike-in control DNAs (defined sequences) added to every IP for QC. Validation panel of 6 loci assayed in parallel by bisulfite pyrosequencing. Operators blinded to sample identity during qPCR validation via coded tubes.

BSL-1 for purified human cell-line DNA. The work uses no live BSL-2 cells if only DNA is handled; if culturing U87-MG, treat as BSL-2 (human-derived tumor line) with biosafety cabinet and appropriate waste. Ethanol is flammable. Sonication generates aerosols — keep tubes sealed and clean the Covaris bath. Wear gloves, lab coat, eye protection. Dispose of antibody/bead reagents and proteinase K digests as biohazardous liquid waste.

Input control: matched non-immunoprecipitated fragmented DNA from the same pool, processed in parallel for normalization. Spike-in positive control: in-vitro methylated control DNA — must enrich >25-fold. Spike-in negative control: unmethylated control DNA — must show no enrichment (near input level). No-antibody (beads-only) IP to assess nonspecific bead binding. Biological reference: normal astrocyte DNA as the non-tumor methylation baseline.

qPCR methylated/unmethylated spike-in enrichment >25-fold confirms a successful IP. Genome-wide enrichment concentrated in CpG-dense regions; tumor sample showing gained enrichment at normally-unmethylated CpG-island promoters of tumor suppressors. MeDIP-vs-pyrosequencing concordance Spearman ρ > 0.7. Library insert size 200-350 bp. Input shows broad, near-uniform coverage; IP shows peaked CpG-island-biased coverage.

• Genomic DNA from U87-MG and normal human astrocytes, 1 µg per reaction
• Anti-5-methylcytosine mouse mAb (clone 33D3, Diagenode C15200081 or Active Motif 39649), 1 µg per IP
• Protein A/G magnetic beads (Thermo 88803)
• IP buffer: 10 mM sodium phosphate pH 7.0, 140 mM NaCl, 0.05% Triton X-100
• Methylated and unmethylated control DNA spike-ins (Diagenode MeDIP kit C02010021)
• Proteinase K (20 mg/mL); DNA purification columns (Qiagen MinElute)
• KAPA HyperPrep library kit; AMPure XP beads
• Bisulfite pyrosequencing reagents (PyroMark Q48) for orthogonal validation
• TE buffer, 80% ethanol, nuclease-free water

To perform affinity-based capture of methylated DNA fragments from fragmented genomic DNA using a monoclonal anti-5-methylcytosine antibody, generating region-level (not single-base) methylation enrichment maps. MeDIP is antibody-density-dependent, so it preferentially captures regions with multiple methylated CpGs, making it well-suited to detecting aberrant CpG-island hypermethylation in cancer at lower sequencing cost than WGBS.

1. Sonicate 1 µg gDNA to 150-300 bp (Covaris; verify on TapeStation); reserve 10% as input.
2. Denature DNA 95°C 10 min, snap-cool on ice 10 min (5mC antibody requires single-stranded DNA).
3. Add methylated/unmethylated spike-ins; bring to 500 µL in IP buffer; add 1 µg anti-5mC antibody; rotate 4°C 2 h.
4. Add 20 µL pre-blocked protein A/G beads; rotate 4°C 2 h.
5. Magnet-wash beads 3x with cold IP buffer (5 min each, rotation).
6. Elute by digestion: 250 µL digestion buffer + Proteinase K, 55°C 3 h shaking.
7. Purify IP and input DNA on MinElute columns; elute in 20 µL.
8. qPCR-validate enrichment: methylated spike-in should show >25-fold enrichment over unmethylated spike-in.
9. Prepare sequencing libraries from IP and input (KAPA HyperPrep, 10-12 cycles); sequence single-end 50 bp to ~30M reads.

Independent variable: cell source (glioblastoma U87-MG vs normal astrocyte). Dependent variables: anti-5mC enrichment ratio (IP/input) per region and per validated locus. Controlled variables: DNA input mass (1 µg), fragment size (150-300 bp), antibody amount (1 µg), denaturation step, spike-in amounts, and number of washes.

We hypothesize that U87-MG glioblastoma DNA will show strong anti-5mC enrichment over a defined panel of tumor-suppressor CpG-island promoters (e.g., MGMT, CDKN2A/p16, PTEN) relative to a normal human astrocyte reference, and that MeDIP enrichment at these loci will correlate (Spearman ρ > 0.7) with locus-specific bisulfite-pyrosequencing methylation percentages on the same samples.

Trim and align reads to hg38 with BWA-MEM; remove duplicates with Picard. Call MeDIP peaks/windows with MEDIPS (R), which models CpG density to correct the well-known MeDIP bias toward CpG-rich regions and computes relative methylation scores (rms) and absolute methylation (ams). Generate IP/input log2 ratio bigWigs with deepTools. Identify differentially methylated windows between tumor and normal; annotate to CpG islands and promoters with ChIPseeker.

Low spike-in enrichment (<5-fold): ensure DNA was fully denatured and snap-cooled (5mC antibody binds only ssDNA) and verify antibody lot. High background/poor IP/input contrast: increase wash stringency or reduce antibody; pre-block beads with BSA. Over-fragmentation (<150 bp): reduce sonication intensity/time. CpG-density bias dominating peaks: always apply MEDIPS CpG-density correction before calling DMRs. Poor pyroseq concordance: confirm both assays used the same sonicated input pool and the same loci coordinates.

Differential methylation across genomic windows tested with MEDIPS (edgeR-based negative binomial GLM, n=3 per group); Benjamini-Hochberg FDR q<0.05 and |log2 ratio|>1. Correlate MeDIP signal with pyrosequencing % methylation across the 6-locus panel by Spearman's ρ (report p and 95% CI via bootstrap). Power: triplicate MeDIP-seq at 30M reads detects 2-fold window changes at >0.8 power for CpG-dense windows.