Simplified MethylRAD Sequencing to Detect Changes in DNA Methylation at Enhancer Elements in Differentiating Embryonic Stem Cells

Debapriya Saha, Allison B. Norvil, Nadia A. Lanman, Humaira GowherEpigenomes 2020 · DOI

TL;DR

This paper introduces a simplified and more robust MethylRAD sequencing method to accurately detect DNA methylation changes in challenging, low-CpG content regulatory regions like enhancers. They demonstrate its utility in tracking epigenetic shifts during embryonic stem cell differentiation and its sensitivity to LSD1 inhibition, highlighting its potential for diagnostic applications.

What Problem Does This Paper Address?

Understanding how DNA methylation changes during development and disease is crucial, especially in regulatory elements like enhancers. These regions often have low CpG content, making them difficult to study with traditional methods like whole-genome bisulfite sequencing (WGBS). WGBS is also expensive and generates massive datasets, much of which might be redundant for studying specific methylation changes. Existing reduced-representation methods (like RRBS) and immunoprecipitation-based techniques (like MeDIP) have their own limitations, including resolution, bias, and technical complexity. The authors aim to provide a more accessible, sensitive, and accurate method to specifically profile methylation changes in these low-CpG regulatory regions, particularly during embryonic stem cell (ESC) differentiation, where such changes are critical for developmental processes.

Methods

The core of this work is the development of an improved MethylRAD sequencing technique. MethylRAD relies on the methylation-dependent restriction enzyme FspEI, which cuts DNA at methylated cytosines. The key innovation here is a simplified experimental workflow:

  • Streamlined Fragment Isolation: Instead of a tedious gel excision of specific fragments after library preparation (as in conventional MethylRAD), they directly excise the 30-35 bp FspEI-digested fragments from an agarose gel before library preparation. This makes the process more straightforward and reproducible.
  • Bioinformatic Analysis Tool: They also developed an associated bioinformatic pipeline for analyzing the generated data.

This improved method allows for the detection of methylation at FspEI recognition sites (CmC, specifically CCGG and CCWGG). They also employed methylation-dependent qPCR (MD-qPCR) for validation of specific enhancer regions.

Datasets

The study utilized embryonic stem cells (ESCs) for their experiments. Specifically, they analyzed:

  • Undifferentiated ESCs.
  • ESCs differentiated for 7 days (untreated).
  • ESCs differentiated for 7 days in the presence of an LSD1 inhibitor (pargyline).

They also compared their MethylRAD data to existing bulk ESCs RRBS data from a previous study (Guo et al., 2013) to assess the fidelity and coverage of their method.

Key Insights and Ideas

  • MethylRAD's Sensitivity for Low-CpG Regions: The improved MethylRAD method is highly effective at detecting methylation in CpG-poor regions, such as enhancers, which are often missed or poorly represented by other techniques like RRBS.
  • LSD1's Crucial Role in Enhancer Methylation: The study strongly implicates the histone demethylase LSD1 in regulating DNA methylation at pluripotency gene enhancers (PpGes) during ESC differentiation. Inhibition of LSD1 significantly impairs the gain of DNA methylation in these enhancers.
  • CpG Content Dictates Sensitivity to LSD1 Inhibition: Enhancers with intermediate or low CpG content are much more sensitive to LSD1 inhibition regarding DNA methylation changes compared to high-CpG content enhancers. This suggests a differential regulatory mechanism based on CpG density.
  • Simplified MethylRAD is Robust: The streamlined protocol makes MethylRAD a more accessible and reproducible technique for researchers, while maintaining high sensitivity and accuracy.

Key Results

  • Increased CpG Site Detection: The improved MethylRAD method identified a 32% increase in unique CpG sites compared to RRBS data.
  • Enhancer Enrichment: MethylRAD showed a 10-fold increase in detecting LSD1-bound enhancers and a 2-fold increase in ESC enhancers compared to RRBS.
  • LSD1 Inhibition Impact: In the presence of the LSD1 inhibitor pargyline, a significant reduction (from ~27% to ~6.3%) was observed in the number of enhancers that gained DNA methylation post-differentiation.
  • CpG Content Correlation: Approximately 40% of high-CpG enhancers gained methylation post-differentiation, while only 25% of low/intermediate-CpG enhancers did. However, the methylation gain in low/intermediate-CpG enhancers was reduced to ~5% with LSD1 inhibition, compared to a reduction to ~20% in high-CpG enhancers.
  • Method Validation: The MethylRAD data showed good overlap with RRBS data, and MD-qPCR confirmed the findings for specific enhancers.

Limitations and Future Work

  • CpA/CpT Methylation Measurement: While MethylRAD can detect non-canonical cytosine methylation (CpA and CpT), accurately measuring changes in these is challenging due to their lower abundance compared to CpG methylation. Increasing sequencing depth could improve this.
  • "Background" Methylation: The study acknowledges that MethylRAD primarily detects methylated DNA. While low read counts can be used as a proxy for unmethylated regions, this is an indirect measure.
  • E-P Loop Regulation: The authors suggest a potential role for enhancer-promoter (E-P) loops in regulating promoter methylation, which could be experimentally tested in future work.
  • Spurious Methylation: The study noted a small fraction of PpGes showing decreased methylation post-differentiation, which increased with pargyline treatment. This could be due to spurious methylation or random differentiation events, warranting further investigation.
  • Diagnostic Potential: The authors emphasize the method's suitability for diagnostic purposes, particularly for identifying differentially methylated regions in CG-poor regions relevant to disease and development. Future work could focus on applying this method to various disease contexts.

Summary updated

Abstract

Differential DNA methylation is characteristic of gene regulatory regions, such as enhancers, which mostly constitute low or intermediate CpG content in their DNA sequence. Consequently, quantification of changes in DNA methylation at these sites is challenging. Given that DNA methylation across most of the mammalian genome is maintained, the use of genome-wide bisulfite sequencing to measure fractional changes in DNA methylation at specific sites is an overexertion which is both expensive and cumbersome. Here, we developed a MethylRAD technique with an improved experimental plan and bioinformatic analysis tool to examine regional DNA methylation changes in embryonic stem cells (ESCs) during differentiation. The transcriptional silencing of pluripotency genes (PpGs) during ESC differentiation is accompanied by PpG enhancer (PpGe) silencing mediated by the demethylation of H3K4me1 by LSD1. Our MethylRAD data show that in the presence of LSD1 inhibitor, a significant fraction of LSD1-bound PpGe fails to gain DNA methylation. We further show that this effect is mostly observed in PpGes with low/intermediate CpG content. Underscoring the sensitivity and accuracy of MethylRAD sequencing, our study demonstrates that this method can detect small changes in DNA methylation in regulatory regions, including those with low/intermediate CpG content, thus asserting its use as a method of choice for diagnostic purposes.
BibTeX
@inproceedings{saha2020simplified,
  author = {Debapriya Saha and Allison B. Norvil and Nadia A. Lanman and Humaira Gowher},
  title = {{Simplified MethylRAD Sequencing to Detect Changes in DNA Methylation at Enhancer Elements in Differentiating Embryonic Stem Cells}},
  booktitle = {Epigenomes},
  year = {2020},
  doi = {10.3390/epigenomes4040024},
}
Figure