The BisPCR(2) method for targeted bisulfite sequencing

Abstract

Background: DNA methylation has emerged as an important regulator of development and disease, necessitating the design of more efficient and cost-effective methods for detecting and quantifying this epigenetic modification. Next-generation sequencing (NGS) techniques offer single base resolution of CpG methylation levels with high statistical significance, but are also high cost if performed genome-wide. Here, we describe a simplified targeted bisulfite sequencing approach in which DNA sequencing libraries are prepared following sodium bisulfite conversion and two rounds of PCR for target enrichment and sample barcoding, termed BisPCR(2).

Results: We have applied the BisPCR(2) technique to validate differential methylation at several type 2 diabetes risk loci identified in genome-wide studies of human islets. We confirmed some previous findings while not others, in addition to identifying novel differentially methylated CpGs at these genes of interest, due to the much higher depth of sequencing coverage in BisPCR(2) compared to prior array-based approaches.

Conclusion: This study presents a robust, efficient, and cost-effective technique for targeted bisulfite NGS, and illustrates its utility by reanalysis of prior findings from genome-wide studies.

Keywords: DNA methylation; Next-generation sequencing; Targeted bisulfite sequencing.

Fig. 1

Schema of BisPCR² method for targeted bisulfite sequencing. DNA sequencing libraries are prepared by bisulfite conversion of genomic DNA followed by two rounds of PCR for target enrichment (PCR#1) and subsequent sample barcoding (PCR#2). Partial adapter overhangs are added to target enrichment primers to permit simplified library preparation by PCR. PCR#1 amplicons are pooled prior to the PCR#2 reaction for each biological sample. Due to the presence of the unique barcodes, all PCR#2 amplicons can be pooled for a single next-generation sequencing run.

Fig. 2

BisPCR² DNA sequencing libraries. a Bioanalyzer gel visualizing the five amplicon fragments of a representative sample, ND1, following PCR#2. b Bioanalyzer electropherogram quantifying the amount of each fragment in ND1, illustrating the roughly equivalent amounts of all five amplicons. c Average reads per amplicon for each sample. ND non-diabetic, T2D type 2 diabetic.

Fig. 3

BisPCR² DNA methylation analysis confirms increased CpG methylation in type 2 diabetic human islets at the MEG3 locus. a Average percent CpG methylation at the MEG3 locus for five non-diabetic and five type 2 diabetic human islet samples measured by BisPCR². p values calculated by a two-tailed t test. *p < 0.05. Error bars indicate SEM. b Quantification of average percent CpG methylation by pyrosequencing using the same samples and same MEG3 PCR primer sequences as in a. Only 2 of 19 CpGs are covered in the pyrosequencing assay. Data displayed as in a.

Fig. 4

Validation of CpG loci differentially methylated in type 2 diabetic pancreatic islets by BisPCR². Average percent CpG methylation in 5 non-diabetic and 5 type 2 diabetic human islet samples at loci previously shown to be differentially methylated in type 2 diabetic human islets, including a INS, b IRS1, c CDKN1A, and d PDE7B. Black arrows indicate CpGs analyzed previously by pyrosequencing by Dayeh and colleagues [18]. p value calculated by two-tailed t tests. *p < 0.05. Error bars indicate SEM.