DNA Deaminase-Based Hydroxymethyl–Seq: Nondestructive and Base-Resolution

3 scientists at a bench working

The most common epigenetic mark throughout the genome is 5-methylcytosine (5mC). Additionally, 5-hydroxymethylcytosine (5hmC) is also known to play a role in various processes, including embryonic stem cell development and tumorigenesis. There is also an enrichment of 5hmC in human neurons. Accurately quantifying and determining the genomic location of these two different marks is therefore crucial to our understanding of molecular biology and disease.

Standard bisulfite sequencing (BS-seq) has been a reliable method for detecting un-methylated and methylated cytosine residues in the genome. However, BS-seq cannot distinguish between different epigenetic marks. This is because un-methylated cytosine residues undergo deamination (C → U) while the methylated cytosines remain unconverted after bisulfite treatment.

To address this limitation, there are a few available options that rely on localizing different cytosine variants based upon their chemical reactivity. Oxidative bisulfite sequencing (oxBS-seq) selectively oxidizes 5hmC to form 5-formylcytosine (5fC). Following bisulfite treatment, 5fC is converted to uracil while 5mC remains as cytosine. Therefore, oxBS-seq provides an accurate readout for 5mC. However, the results from both oxBS-seq and BS-seq must be compared so that 5hmC sites can be inferred.

Other bisulfite-coupled approaches include tet-assisted bisulfite sequencing (TAB-seq). This approach utilizes ten-eleven translocation (TET) proteins, which have a naturally occurring involvement in 5mC demethylation. In the case of TAB-seq, 5hmC conversion is blocked and TET oxidizes 5mC to 5-carboxylcytosine (5caC), which is then converted to uracil after bisulfite treatment.

Although these approaches improve the ability to distinguish epigenetic marks in the genome, they both rely on bisulfite conversion. This makes studying epigenetic changes in rare cell types more challenging. Because bisulfite chemical treatment requires a high temperature and different pH conditions to perform, it can significantly degrade certain targets. When the starting amount of DNA is limited, genomic coverage will be less and additional methods may have to be used to help enrich these regions. In turn, some conflicting 5hmC results have been produced.

A reliable method for detecting epigenetic modifications at high resolution that is also less destructive would therefore be beneficial for studying currently inaccessible genomes. To address these challenges, a research group from the Perelman School of Medicine at the University of Pennsylvania reported on an improved method for detecting 5hmC status in Nature Biotechnology. They built upon their previous work where they studied the AID/APOBEC protein family. These proteins are known to perform a diversity of physiological functions, including immune defense; where genetic material from pathogens is targeted and then mutated.

What is relevant to epigenetics is that their study highlighted the APOBEC3A (A3A) family member, which selectively deaminates cytosines and 5mC, but does not modify 5hmC. The researchers quantified this reaction and found that compared to other cytosines, there was a 5,000-fold reduction in the A3A deamination rate of 5hmC. This gave them the idea to exploit A3A’s differential reactivity in order to map 5hmC bases in the genome.

In the group’s most recent study, they developed APOBEC-Coupled Epigenetic Sequencing (ACE-Seq). In the optimized ACE-seq protocol, A3A enzymatically converts cytosine and 5mc to uracil, whereas 5hmC remains as cytosine.

“We’re hopeful that this method offers the ability to decode epigenetic marks on DNA from small and transient populations of cells that have previously been difficult to study,” said co-senior author Rahul Kohli, MD, PhD, an assistant professor of Biochemistry and Biophysics, and Medicine. Emily Schutsky, Kohli’s graduate student, is the first author of the study.

Using bacteriophage DNA, the University of Pennsylvania researchers first showed that ACE-seq cytosine conversion rate was similar to bisulfite. They also concluded that the ACE-seq non-conversion rate of 5hmC outperformed current methods. Moreover, results from different cell lines were compared to previous TAB-seq studies. The 5hmC base-resolution map for mouse excitatory neurons produced by ACE-seq was highly correlated with TAB-seq whole brain cortex. Overall, ACE-seq was highly reliable in distinguishing 5mC and 5hmC.                                                                                

In the same study, the authors showcased how ACE-seq could be applied to help answer questions surrounding the genomic distribution of 5hmC. BS-seq and ACE-seq base-resolution profiles for excitatory neurons were utilized together so that 5hmC/5mC signals were detected. They found that 5hmC was predominantly associated with CG dinucleotides, which confirms previous results by Tab-seq.

“We were able to show that sites along the genome that appear to be modified are in fact very different in terms of the distribution of these two marks,” Kohli said. “This finding suggests important and distinctive biological roles for the two marks on the genome.”

Most importantly, there is a key difference that sets ACE-seq apart from other bisulfite-coupled methods. This includes the initial concentration of genomic DNA for ACE-seq, which was 1,000-fold less than TAB-seq. Using mouse embryonic stem cells; the researchers quantified the amount of short versus long DNA fragment sizes produced from both ACE-seq and bisulfite. In comparison to bisulfite treatment, ACE-seq had a much higher frequency of longer DNA fragments (~1,000 base pairs in length) that remained intact.

“This technological advance paves the way to better understand complex biological processes such as how the nervous system develops or how a tumor progresses,” said co-senior author Hao Wu, PhD, an assistant professor of Genetics. 

Studying small traces of DNA is therefore possible because the enzymatic reaction for cytosine conversion used in ACE-seq does not harm DNA. This method could also benefit certain diseases such as cancer, where detecting small amounts of free-floating DNA released by tumors into the bloodstream can help with earlier diagnoses.

Sources cited

Booth MJ, Ost TWB, Beraldi D, et al. (2013). Oxidative bisulfite sequencing of 5-methylcytosine and 5-hydroxymethylcytosine. Nature Protocols, 8(10):1841-51.

Schutsky EK, DeNizio J, Hu P., et al. (2018). Nondestructive, base-resolution sequencing of 5-hydroxymethylcytosince using a DNA deaminase. Nature Biotechnology, 36:1083-90.

Schutsky E.K., Nabel C.S., Davis A.K.F., et al. (2017). APOBEC3A efficiently deaminates methylated, but not TET-oxidized, cytosine bases in DNA. Nucleic Acids Research, 45(13): 7655–65.

University of Pennsylvania School of Medicine. “There’s a better way to decipher DNA’s epigenetic code to identify disease: Study improves decades-old method.” 8 October 2018.

Related Articles


If you like reading our articles…

Join our e-newsletter! Stay up-to-date with our weekly posts on epigenetics and health, nutrition, exercise, and more.