A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements

Abstract

We report a method for studying global DNA methylation based on using bisulfite treatment of DNA and simultaneous PCR of multiple DNA repetitive elements, such as Alu elements and long interspersed nucleotide elements (LINE). The PCR product, which represents a pool of approximately 15 000 genomic loci, could be used for direct sequencing, selective restriction digestion or pyrosequencing, in order to quantitate DNA methylation. By restriction digestion or pyrosequencing, the assay was reproducible with a standard deviation of only 2% between assays. Using this method we found that almost two-thirds of the CpG methylation sites in Alu elements are mutated, but of the remaining methylation target sites, 87% were methylated. Due to the heavy methylation of repetitive elements, this assay was especially useful in detecting decreases in DNA methylation, and this assay was validated by examining cell lines treated with the methylation inhibitor 5-aza-2'deoxycytidine (DAC), where we found a 1-16% decrease in Alu element and 18-60% LINE methylation within 3 days of treatment. This method can be used as a surrogate marker of genome-wide methylation changes. In addition, it is less labor intensive and requires less DNA than previous methods of assessing global DNA methylation.

Figure 1

Direct DNA sequencing of bisulfite repetitive element PCR of Alu elements. (A) Schematic of possible fates of Alu element CpG methylation sites. Due to the mutation of CpG sites via spontaneous deamination of 5-methylcytosine to T during evolution, a CpG site can be changed into a TpG or CpA dinucleotide. Neither TpG nor CpA are targets for methylation. Following bisulfite treatment, a methylated CpG will remain CpG. However, an unmethylated CpG will give rise to TpG, which is indistinguishable from a deamination mutation of the forward strand. Thus, three possibilities arise for our sequencing data: CpG that represents a methylated CpG site, TpG that represents either an unmethylated CpG site or a mutation of the forward strand, and TpA that represents a mutation of the reverse strand followed by conversion of the unmethylated C to T by bisulfite. (B) Sequencing data of bisulfite repetitive element PCR of Alu elements. Genomic DNA was isolated from peripheral human blood and bisulfite treated. Alu element PCR was performed and the PCR product was cloned, and 15 clones were sequenced. There were 12 potential CpG methylation sites per clone for a total of 180 potential methylation sites sequenced. Black circles represent methylated CpG sites (66/180 = 36.7%). ‘T’ represents TpG sites that were either unmethylated or mutated (53/180 = 24.4%). ‘A’ represents TpA sites that were mutated (41/180 = 22.8%). ‘X’ represents other mutations (20/180 = 11.1%). The CpG sites used for pyrosequencing (‘PS’) and COBRA (‘MboI’) are indicated.

Figure 1

Direct DNA sequencing of bisulfite repetitive element PCR of Alu elements. (A) Schematic of possible fates of Alu element CpG methylation sites. Due to the mutation of CpG sites via spontaneous deamination of 5-methylcytosine to T during evolution, a CpG site can be changed into a TpG or CpA dinucleotide. Neither TpG nor CpA are targets for methylation. Following bisulfite treatment, a methylated CpG will remain CpG. However, an unmethylated CpG will give rise to TpG, which is indistinguishable from a deamination mutation of the forward strand. Thus, three possibilities arise for our sequencing data: CpG that represents a methylated CpG site, TpG that represents either an unmethylated CpG site or a mutation of the forward strand, and TpA that represents a mutation of the reverse strand followed by conversion of the unmethylated C to T by bisulfite. (B) Sequencing data of bisulfite repetitive element PCR of Alu elements. Genomic DNA was isolated from peripheral human blood and bisulfite treated. Alu element PCR was performed and the PCR product was cloned, and 15 clones were sequenced. There were 12 potential CpG methylation sites per clone for a total of 180 potential methylation sites sequenced. Black circles represent methylated CpG sites (66/180 = 36.7%). ‘T’ represents TpG sites that were either unmethylated or mutated (53/180 = 24.4%). ‘A’ represents TpA sites that were mutated (41/180 = 22.8%). ‘X’ represents other mutations (20/180 = 11.1%). The CpG sites used for pyrosequencing (‘PS’) and COBRA (‘MboI’) are indicated.

Figure 2

Calculation of the number of Alu elements being assessed by bisulfite Alu PCR. Competitive PCR was performed using a fixed amount of bisulfite-treated genomic DNA and a plasmid containing a cloned Alu element fragment with an internal duplication that gives a larger PCR product. A fixed amount of bisulfite-treated genomic DNA was mixed with serial dilutions of the plasmid which gave two PCR products, one from the genomic DNA and a larger one from the plasmid. The PCR products were quantitated using a capillary electrophoresis system, the Agilent 2100 Bioanalyzer (Agilent Technologies) and plotted above. From this experiment, it is approximated that 0.5 ng of plasmid gives equivalent PCR product to 50 ng of bisulfite-treated genomic DNA.

Figure 3

Quantitation of DNA methylation in cell lines treated with DAC using the bisulfite repetitive element PCR technique. Hct116, RKO and SW48 cell lines were treated with DAC. Genomic DNA was isolated from DAC-treated (+) and -untreated (–) controls. The genomic DNA was treated with sodium bisulfite and a non-specific PCR was performed, which amplified a pool of Alu or LINE-1 repetitive elements. The PCR product was then digested with MboI (Alu) or HinfI (LINE-1), which only cuts repetitive elements that were originally methylated. The Alu PCR assays a single methylation site and therefore the MboI digestion will cut the PCR product from 152 to 125 and 27 bp (data not shown). The LINE-1 PCR assays the methylation of two sites and therefore HinfI can generate five possible digestion products of 285, 247, 166, 128 and 38 bp (data not shown). The digested PCR product was separated by polyacrylamide gel electrophoresis and stained with ethidium bromide. The lower cut bands represent methylated repetitive elements. The upper band represents unmethylated repetitive elements or repetitive elements in which the restriction site has been mutated. The PCR bands were quantitated and the amount of methylation is shown below each gel lane. Similar results were obtained for both the Alu element and LINE-1 assays.

Figure 4

Quantitation of DNA methylation using bisulfite repetitive element PCR and pyrosequencing. (A) SssI methylase-treated DNA, (B) untreated Hct116 cell line DNA or (C) DAC-treated Hct116 cell line DNA was bisulfite treated and PCR of Alu repetitve elements was performed. The PCR product was purified and methylation was quantitated using the PSQ HS 96 Pyrosequencing System (Pyrosequencing, Inc.). The pyrogram quantitates C for methylated and T for unmethylated or mutated DNA. The shaded regions represent three tandem CpG sites quantitated in Alu elements, and the percent methylation at each site is shown above the peaks. The average methylation of the three sites is calculated on the left for each pyrogram. The maximum absolute methylation of 23.2% is calculated by SssI methylase-treated DNA (A), Hct116 cells have 20.2% methylation, and this methylation decreases to 14.5% after DAC treatment of Hct116 cells.