Bisulfite Sequencing

Bisulfite sequencing is the primary biochemical and DNA sequencing methodology used to determine the pattern of epigenetic DNA methylation—specifically 5-methylcytosine (5mC)—at single-base-pair resolution across a genome. Developed in 1992 by Marianne Frommer, Louise E. McDonald, Douglas S. Millar, Collis M. Watt, Peter L. Molloy, and colleagues at the CSIRO Division of Biomolecular Engineering in Sydney, Australia, bisulfite conversion solved a biochemical blind spot of molecular genetics: why standard polymerase chain reaction (PCR) and sequencing polymerases erase epigenetic methylation marks.

In canonical Watson-Crick base pairing, both unmodified cytosine (C) and methylated 5-methylcytosine (5mC) carry identical hydrogen-bonding donors and acceptors along their Watson-Crick edge (C2 carbonyl, N3 ring nitrogen, and C4 exocyclic amino group), forming three hydrogen bonds with guanine (G). Consequently, DNA polymerases used in PCR, Sanger sequencing, and short-read massively parallel sequencing cannot distinguish 5mC from C. Treating single-stranded DNA with sodium bisulfite (NaHSO₃) converts unmethylated cytosines into uracil (U) (which amplifies as thymine [T] during PCR) while leaving 5mC unmodified as C, translating an epigenetic mark into a binary C-versus-T sequence difference across the ~28 million CpG dinucleotides of the human genome.

Step 1: Reversible Nucleophilic Sulfonation at Cytosine C6

Bisulfite conversion requires denatured, single-stranded DNA (ssDNA) because double-stranded Watson-Crick base pairing sterically shields the C5=C6 double bond of cytosine from nucleophilic attack. After alkaline or thermal denaturation (95–98 °C), ssDNA is incubated with sodium bisulfite (2–5 M NaHSO₃ / Na₂S₂O₅) at acidic pH 5.0 and 50–65 °C for 2 to 16 hours.

At pH 5.0, protonation of the N3 ring nitrogen activates the pyrimidine ring of unmethylated cytosine toward nucleophilic addition. The bisulfite anion (HSO₃⁻) attacks the electrophilic C6 carbon across the C5=C6 double bond, saturating the ring to form the diastereomeric intermediate cytosine-6-sulfonate (5,6-dihydrocytosine-6-sulfonate).

Step 2: Hydrolytic Deamination at C4 and Step 3: Alkaline Desulfonation

Once the C5=C6 double bond is saturated in cytosine-6-sulfonate, the conjugated aromaticity of the pyrimidine ring is broken, increasing the susceptibility of the C4 imino/amino carbon to attack by water (H₂O). Irreversible hydrolytic deamination replaces the C4 exocyclic amino group (–NH₂) with a carbonyl oxygen (=O), releasing ammonium (NH₄⁺) and generating uracil-6-sulfonate.

Finally, in Step 3 (Alkaline Desulfonation), the reaction pH is raised to pH 9.0–10.0 (using NaOH or Tris-base buffer). Under alkaline conditions, a base abstracts the proton at C5, triggering β-elimination of the C6 sulfonate group as sulfite (SO₃²⁻), restoring the C5=C6 double bond, and yielding uracil (U) at >99.5% conversion efficiency. Why does 5-methylcytosine (5mC) resist conversion? The methyl group (–CH₃) covalently attached to carbon-5 (C5) exerts both an electron-donating inductive effect (deactivating the C5=C6 bond toward nucleophilic HSO₃⁻ attack) and steric hindrance against sulfonation and subsequent deprotonation, reducing the reaction rate by more than two orders of magnitude so 5mC remains intact as C.

Comparison of Genomic Methylation Sequencing Modalities

Whole-Genome Bisulfite Sequencing (WGBS) and RRBS

In 2008 and 2009, Ryan Lister and Joseph R. Ecker at the Salk Institute coupled sodium bisulfite conversion with high-throughput sequencing (MethylC-Seq / WGBS) to generate the first single-base-resolution methylomes of Arabidopsis thaliana and human embryonic stem cells, discovering non-CpG methylation (mCHG and mCHH) in pluripotent stem cells and brain neurons. Because bisulfite treatment converts ~70%–80% of all genomic cytosines into thymines, the resulting sequencing reads have reduced four-base sequence complexity (~20% A, ~30% G, ~50% T, <2% C on the converted strand). Three-letter bioinformatics aligners—such as Bismark (Felix Krueger and Simon Andrews, 2011) and BSMAP—perform in silico C-to-T (and G-to-A) conversions on both the read and the reference genome to map reads and calculate fractional methylation β = C / (C + T) at each CpG site.

To profile clinical cohorts at lower sequencing depth, Reduced Representation Bisulfite Sequencing (RRBS)—introduced by Alexander Meissner, Rudolf Jaenisch, and Eric Lander in 2005—digests genomic DNA with the methylation-insensitive restriction endonuclease *Msp*I, which cleaves at every 5′-C^CGG-3′ recognition motif regardless of whether the internal CpG is methylated. Size-selecting 40–220 bp restriction fragments enriches >85% of human CpG islands and gene promoters while sequencing only 1% to 3% of the genome.

Resolving 5-Hydroxymethylcytosine (oxBS/TAB-Seq) and Enzymatic Methyl-seq (EM-seq)

Classical sodium bisulfite sequencing exhibits two primary biochemical limitations:

1. Inability to Distinguish 5mC from 5hmC: Ten-Eleven Translocation (TET1/2/3) dioxygenases oxidize 5mC into 5-hydroxymethylcytosine (5hmC), an active demethylation intermediate and stable epigenetic mark enriched in brain Purkinje neurons and embryonic stem cells. In sodium bisulfite, 5hmC reacts with HSO₃⁻ to form cytosine-5-methylenesulfonate (CMS), which resists deamination and is read as C—making 5hmC indistinguishable from 5mC. Oxidative Bisulfite Sequencing (oxBS-Seq) (Shankar Balasubramanian and Wolf Reik, 2012) uses potassium perruthenate (KRuO₄) to selectively oxidize 5hmC into 5-formylcytosine (5fC), which bisulfite converts to U, whereas TET-Assisted Bisulfite Sequencing (TAB-Seq) (Chuan He, 2012) protects 5hmC with β-glucosyltransferase (5gmC) while recombinant TET1 oxidizes all 5mC to 5caC (which deaminates to U), reading only 5hmC as C.
2. Acid-Induced Depurination and DNA Fragmentation: Incubating DNA at pH 5.0 and 65 °C in molar bisulfite triggers protonation of purine rings (adenine and guanine), causing glycosidic bond cleavage (depurination) and phosphodiester backbone scission that degrades 85% to 95% of input DNA. To eliminate chemical degradation, Enzymatic Methyl-seq (EM-seq) uses recombinant TET2 and T4 β-glucosyltransferase to oxidize and glucosylate 5mC and 5hmC (protecting them from deamination), followed by the cytidine deaminase APOBEC3A to enzymatically deaminate unmodified cytosines into uracil at neutral pH—preserving long fragments and picogram quantities of cell-free circulating tumor DNA (ctDNA) in liquid biopsies.

Why are bisulfite-treated DNA strands no longer complementary after sodium bisulfite conversion?

In native double-stranded DNA, a top strand 5′-A-mC-G-C-T-3′ is paired with its reverse-complement bottom strand 3′-T-G-mC-G-A-5′. Sodium bisulfite converts unmethylated cytosines into uracil independently on each denatured single strand, while leaving guanine bases unaltered. After conversion, the top strand becomes 5′-A-mC-G-U-T-3′, whereas the bottom strand becomes 3′-T-G-mC-G-A-5′ (if its single C was methylated) or 3′-T-G-U-G-A-5′. Because G no longer pairs with the newly generated U residues at formerly unmethylated sites, the two converted strands are no longer complementary, requiring strand-specific PCR primers or post-bisulfite adapter tagging (PBAT) and four-strand reference alignment (OT, OB, CTOT, CTOB) in Bismark.

How do bisulfite sequencing experiments measure the non-conversion false-positive rate using unmethylated bacteriophage lambda (λ) DNA spike-ins?

Because any unmethylated cytosine that fails to react with sodium bisulfite (due to incomplete thermal denaturation or secondary hairpin structure) remains a C and is falsely called as methylated 5mC, every WGBS or EM-seq library includes a 0.1% to 0.5% mass spike-in of unmethylated bacteriophage lambda (λ) genomic DNA (48,502 bp, isolated from dam⁻ / dcm⁻ E. coli). By measuring the percentage of cytosines that remain read as C across the λ phage control genome (or across non-CpG CHH contexts in mammalian somatic tissues), analysts directly quantify the bisulfite conversion efficiency, which must exceed 99.5% (<0.5% false-positive non-conversion rate).

How is bisulfite or enzymatic methylation sequencing of cell-free DNA (cfDNA) used for multi-cancer early detection (liquid biopsy)?

Apoptotic and necrotic cells shed 160–170 bp nucleosome-protected fragments of cell-free DNA (cfDNA) into peripheral blood plasma. Unlike somatic driver gene mutations—which vary from patient to patient and do not indicate which anatomical organ shed the DNA—epigenetic CpG island methylation patterns are both tissue-specific and recurrently altered early in carcinogenesis (global hypomethylation coupled with focal promoter CpG island hypermethylation). By applying targeted bisulfite or enzymatic methylation sequencing across thousands of differentially methylated regions (DMRs) in plasma cfDNA, machine-learning classifiers simultaneously detect a circulating tumor signal and identify the Tissue of Origin (TOO) (e.g., pancreas, ovary, liver, or lung) from a single blood draw.

1. Frommer, M., McDonald, L. E., Millar, D. S., Collis, C. M., Watt, F., Grigg, G. W., Molloy, P. L., & Paul, C. L. (1992). "A Genomic Sequencing Protocol That Yields a Positive Display of 5-Methylcytosine Residues in Individual DNA Strands." Proceedings of the National Academy of Sciences, 89(5), 1827–1831. https://doi.org/10.1073/pnas.89.5.1827
2. Lister, R., Pelizzola, M., Dowen, R. H., Hawkins, R. D., Hon, G., Tonti-Filippini, J., Nery, J. R., Lee, L., Ye, Z., Ngo, Q. M., et al. (2009). "Human DNA Methylomes at Base Resolution Show Widespread Epigenomic Differences." Nature, 462(7271), 315–322. https://doi.org/10.1038/nature08514
3. Booth, M. J., Branco, M. R., Ficz, G., Oxley, D., Krueger, F., Reik, W., & Balasubramanian, S. (2012). "Quantitative Sequencing of 5-Methylcytosine and 5-Hydroxymethylcytosine at Single-Base Resolution." Science, 336(6083), 934–937. https://doi.org/10.1126/science.1220671
4. Vaisvila, R., Ponnaluri, V. K. C., Sun, Z., Langhorst, B. W., Saleh, L., Guan, S., Dai, N., Campbell, M. A., Sexton, B. S., Marks, K., et al. (2021). "Enzymatic Methyl Sequencing Detects DNA Methylation at Single-Base Resolution from Picograms of DNA." Genome Research, 31(7), 1280–1289. https://doi.org/10.1101/gr.266551.120