This is the classical paper that is cited when a bisulfite modification of methylated DNA is performed. Here is an excerpt:
M.Frommer et al. 1992 Proc Natl Acad Sci USA 89:1827-1831Genomic sequencing protocols, which have been developed to ascertain the methylation status of selected regions within genes, utilize the Maxam and Gilbert chemical cleavage reactions [...] but have two major drawbacks: [...] First, 5-methylcytosine is identified by the lack of a band in all tracks [...]; Second, the sequence obtained represents a population average [...]. To address these problems, we have developed a genomic sequencing method that provides positive identification and localization of 5-methylcytosine in genomic DNA. The method is based on sodium bisulfite-mediated conversion of cytosine to uracil in single-stranded DNA, followed by PCR amplification of the resultant modified DNA. Exact methylation maps of single DNA strands from individual genomic DNA molecules can readily be established, where the position of each 5-methylcytosine is given by a clear positive band on a sequencing gel.
Methods: Sodium Bisulfite Modification. The bisulfite reaction was carried out on 10 ug of DNA. DNA was linearized with [...(a)] restriction enzyme [...], or sheared [...], alkali denatured, neutralized, and precipitated. Denatured DNA was incubated in a total vol of 1.2 ml with freshly prepared 3.1 M sodium bisulfite /0.5 mM hydroquinone, pH 5.0, at 50°C for 16 or 40 hr [...], followed by successive dialyses at 4°C [...] to remove unreacted bisulfite [...]
Results: Positive Identification of 5-Methylcytosine Following Treatment with Sodium Bisulfite. It has been demonstrated that, in single-stranded DNA, sodium bisulfite preferentially deaminates cytosine residues to uracil, compared with a very slow rate of deamination of 5-methylcytosine to thymine [...]. To use this difference in bisulfite reactivity for genomic sequencing of 5-methylcytosine residues, total genomic DNA is fully denatured and treated with sodium bisulfite under conditions such that cytosine is converted stoichiometrically to uracil, but 5-methylcytosine remains nonreactive. We have used either mechanically sheared or restriction enzyme cleaved DNA, which is then denatured by alkali prior to treatment with bisulfite. The second part of the procedure involves PCR amplification of any region of interest in the bisulfite-reacted DNA to yield a fragment in which all uracil (formerly cytosine) and thymine residues have been amplified as thymine and only 5-methylcytosine residues have been amplified as cytosine. Because the bisulfite reaction yields products in which opposite DNA strands are no longer complementary, it is possible to design two pairs of PCR primers such that each pair is specific for only one of the bisulfite-reacted DNA strands [...].Discussion: [...] We have found no consistent lack of reactivity of any unmethylated cytosine residue in the test substrates. Therefore, false positives or overestimation of the extent of a partially methylated sequence does not present a problem in the analysis. It should be possible to use the bisulfite method to obtain a good estimate of exact methylation levels. A small proportion of 5-methylcytosine residues in single-stranded DNA may be deaminated to thymine by the bisulfite treatment, so that the exact proportion of 5-methylcytosine in the original DNA may be slightly greater than the measured proportion of nonreactive sites after bisulfite treatment and PCR.[...] Major advantages of the bisulfite method are (i) the positive display of 5-methylcytosine and (ii) the capacity to generate sequence data for individual DNA strands from single DNA molecules [...].
