their own right, interacting with a distinct group of protein readers and eliciting a
unique epigenetic response [20, 21]. Therefore, it is of great interest to map the
locations of 5hmC, 5fC, and 5caC along the genome.
Standard bisulfite sequencing cannot distinguish between 5mC and its oxidized
forms. Similarly to 5mC, 5hmC resists bisulfite-mediated deamination (Fig. 7)
[65, 78, 114]. Therefore, previous studies that employed bisulfite conversion to
map 5mC detected both hydroxymethylation and methylation together. In contrast,
5fC and 5caC readily undergo deamination upon treatment with bisulfite and thus
cannot be distinguished from C [10, 11, 115]. In addition, 5fC also undergoes
deformylation which is mediated by sulfonation of the C6 position of the nucleoside
to generate the corresponding alcohol. Base-catalyzed deprotonation of the alcohol
and β-elimination produce cytosine [11]. Determining the levels and positions of
epigenetic modifications in the genome is important to understanding the biological
function of these DNA marks. Therefore, novel sequencing methodologies were
needed to separately map all four epigenetic marks.
The first method for distinguishing between 5mC and 5hmC during bisulfite
sequencing utilized an additional chemical reaction on the genomic DNA. Oxidative
bisulfite sequencing (oxBS-Seq) includes DNA treatment with potassium
perruthenate (KRuO 4 ) to selectively oxidize the 5hmC to 5fC (Fig. 8) [115]. Following oxidation, the DNA is treated with bisulfite to deaminate cytosine. Booth et al.
coupled oxBS-Seq with RRBS to increase the sequencing depth at potentially
interesting CG repeats by reducing the amount of DNA that could be sequenced.
5hmC was identified primarily at CG sites located within low-density CpG islands
(CGIs) [115].
As an alternative to chemical oxidation to distinguish between 5mC and 5hmC,
enzymatic oxidation can be used (TET-assisted bisulfite sequencing, TAB-Seq)
[116]. The latter technique utilizes T4 β-glycosyl transferase and UDP-glucose to
protect the primary alcohol on 5hmC, followed by oxidation of 5mC to 5caC using
recombinant TET1 protein (Figs. 8 and 9) [117]. This is coupled with bisulfite
treatment and sequencing of the original sample as well as the TET-treated DNA
to provide data for the locations of 5mC and 5hmC. Using the technique on the
whole genome, 5hmC was found to be almost exclusively found at CG sites, and
nearly half of all 5hmC sites were present in distal regulatory elements
[116]. TAB-Seq has been applied to identify changes in 5mC and 5hmC in
intergenic regions of kidney tumors [118] and to identify sites of 4mC in bacterial
DNA [33].
Mapping 5fC along the genome was accomplished in part by modification of the
previous techniques to distinguish between 5mC and 5hmC. The first reported
method for 5fC was formylcytosine chemical-assisted bisulfite sequencing (fCABSeq) [90]. In this approach, chemical labeling of 5fC is conducted with Oethylhydroxylamine to protect 5fC from conversion during DNA treatment with
bisulfite (Figs. 8 and 9) [90, 115]. DNA sequencing is performed on paired samples
that have either undergone fCAB protection plus bisulfite treatment or standard
bisulfite treatment. Through the mapping of standard bisulfite sequencing results
and fCAB-Seq data to the genome, sites specifically modified with 5fC could be
Experimental Methodologies for Detection and Mapping of Epigenetic DNA Marks
507
unique epigenetic response [20, 21]. Therefore, it is of great interest to map the
locations of 5hmC, 5fC, and 5caC along the genome.
Standard bisulfite sequencing cannot distinguish between 5mC and its oxidized
forms. Similarly to 5mC, 5hmC resists bisulfite-mediated deamination (Fig. 7)
[65, 78, 114]. Therefore, previous studies that employed bisulfite conversion to
map 5mC detected both hydroxymethylation and methylation together. In contrast,
5fC and 5caC readily undergo deamination upon treatment with bisulfite and thus
cannot be distinguished from C [10, 11, 115]. In addition, 5fC also undergoes
deformylation which is mediated by sulfonation of the C6 position of the nucleoside
to generate the corresponding alcohol. Base-catalyzed deprotonation of the alcohol
and β-elimination produce cytosine [11]. Determining the levels and positions of
epigenetic modifications in the genome is important to understanding the biological
function of these DNA marks. Therefore, novel sequencing methodologies were
needed to separately map all four epigenetic marks.
The first method for distinguishing between 5mC and 5hmC during bisulfite
sequencing utilized an additional chemical reaction on the genomic DNA. Oxidative
bisulfite sequencing (oxBS-Seq) includes DNA treatment with potassium
perruthenate (KRuO 4 ) to selectively oxidize the 5hmC to 5fC (Fig. 8) [115]. Following oxidation, the DNA is treated with bisulfite to deaminate cytosine. Booth et al.
coupled oxBS-Seq with RRBS to increase the sequencing depth at potentially
interesting CG repeats by reducing the amount of DNA that could be sequenced.
5hmC was identified primarily at CG sites located within low-density CpG islands
(CGIs) [115].
As an alternative to chemical oxidation to distinguish between 5mC and 5hmC,
enzymatic oxidation can be used (TET-assisted bisulfite sequencing, TAB-Seq)
[116]. The latter technique utilizes T4 β-glycosyl transferase and UDP-glucose to
protect the primary alcohol on 5hmC, followed by oxidation of 5mC to 5caC using
recombinant TET1 protein (Figs. 8 and 9) [117]. This is coupled with bisulfite
treatment and sequencing of the original sample as well as the TET-treated DNA
to provide data for the locations of 5mC and 5hmC. Using the technique on the
whole genome, 5hmC was found to be almost exclusively found at CG sites, and
nearly half of all 5hmC sites were present in distal regulatory elements
[116]. TAB-Seq has been applied to identify changes in 5mC and 5hmC in
intergenic regions of kidney tumors [118] and to identify sites of 4mC in bacterial
DNA [33].
Mapping 5fC along the genome was accomplished in part by modification of the
previous techniques to distinguish between 5mC and 5hmC. The first reported
method for 5fC was formylcytosine chemical-assisted bisulfite sequencing (fCABSeq) [90]. In this approach, chemical labeling of 5fC is conducted with Oethylhydroxylamine to protect 5fC from conversion during DNA treatment with
bisulfite (Figs. 8 and 9) [90, 115]. DNA sequencing is performed on paired samples
that have either undergone fCAB protection plus bisulfite treatment or standard
bisulfite treatment. Through the mapping of standard bisulfite sequencing results
and fCAB-Seq data to the genome, sites specifically modified with 5fC could be
Experimental Methodologies for Detection and Mapping of Epigenetic DNA Marks
507