sequenced an average of 1,000 times, which allows estimation of 5fC levels to
Æ0.1%. Sequencing of DNA from mouse embryonic stem cells showed that 5fC
levels were increased at poised enhancer sequences, while 5mC was also reduced at
those same enhancers, indicating 5mC oxidation by TET could function in epigenetic priming [90]. Another chemical method for identification of 5fC is reduced
bisulfite sequencing (redBS-Seq). The redBS-Seq method employs sodium
cyanoborohydride to reduce 5fC to 5hmC in genomic DNA (Figs. 8 and 9)
[119]. The DNA is then treated with bisulfite to deaminate cytosine, and both
redBS and standard bisulfite-treated DNA samples are sequenced. Cytosine
formylation levels are determined as in fCAB-Seq above. Using this technique,
5fC was shown to be present in CpG sites asymmetrically with 5hmC, where one
strand contains 5fC and the opposite strand 5hmC [119].
An additional method for determination of oxidized forms of 5mC by sequencing
is modification-assisted bisulfite sequencing (MAB-Seq) (Figs. 8 and 9)
[120]. Detection of 5caC and 5fC is simultaneously possible by treatment of the
DNA with bacterial methylase M. SssI, which methylates all unmodified CG dinucleotides to give 5mCG. Following methylation to protect all unmodified cytosine in
a CG context, bisulfite treatment of DNA will lead to deamination of 5fC and 5caC.
This results in sequencing output that can identify the locations of 5fC and 5caC
together. MAB-Seq is reliant on quantitative conversion of C to 5mC by M. SssI and
is limited to detection of 5fC and 5caC only in the CG context. MAB-Seq was
further developed to be specific for 5caC (caMAB-Seq) (Figs. 8 and 9), which takes
advantage of the reduction of 5fC by sodium borohydride prior to performing the
steps of MAB-Seq [120]. Using MAB-Seq and caMAB-Seq, 5fC and 5caC have
been mapped at single-base resolution to regions with higher chromatin accessibility
and therefore TET accessibility [121]. The sequencing profiles obtained via
MAB-Seq and caMAB-Seq indicate that there might be preferential locations for
TET activity, either through chromatin accessibility or sequence-dependent removal
of 5fC and 5caC by TDG [120].
The first report for mapping 4mC along the genome employed a modification of
the TAB-Seq methodology [33]. This is possible because 4mC undergoes deamination at a rate between C and 5mC [33]. To allow for 4mC detection, the bisulfite
conditions were optimized to ensure the retention of 4mC in the genome while
deaminating TET-generated 5caC. Using this methodology, 4mC was mapped in
C. kristjanssonii, a bacterium with similar levels of 5mC and 4mC [33]. This
methodology, 4mC-TAB-Seq, is well suited for identification of 4mC in genomes
which have high concentrations of 4mC at particular sites, especially in bacteria
which utilize 4mC as part of their restriction-modification systems [31]. Application
of this methodology to mammalian cells and tissues is limited by the incomplete
(40–55%) retention of 4mC after bisulfite conversion, making accurate identification
of sites possessing less than complete methylation nearly impossible [33].
Mapping N6MedA can also be accomplished using restriction enzyme-based
sequencing (6mA-RE-Seq) [122]. This sequencing technique takes advantage of
previous information gained from antibody enrichment methods, which found that
N6MedA was found in CATG and CATC motifs in the C. reinhardtii genome, a
510
C. L. Seiler et al.
Æ0.1%. Sequencing of DNA from mouse embryonic stem cells showed that 5fC
levels were increased at poised enhancer sequences, while 5mC was also reduced at
those same enhancers, indicating 5mC oxidation by TET could function in epigenetic priming [90]. Another chemical method for identification of 5fC is reduced
bisulfite sequencing (redBS-Seq). The redBS-Seq method employs sodium
cyanoborohydride to reduce 5fC to 5hmC in genomic DNA (Figs. 8 and 9)
[119]. The DNA is then treated with bisulfite to deaminate cytosine, and both
redBS and standard bisulfite-treated DNA samples are sequenced. Cytosine
formylation levels are determined as in fCAB-Seq above. Using this technique,
5fC was shown to be present in CpG sites asymmetrically with 5hmC, where one
strand contains 5fC and the opposite strand 5hmC [119].
An additional method for determination of oxidized forms of 5mC by sequencing
is modification-assisted bisulfite sequencing (MAB-Seq) (Figs. 8 and 9)
[120]. Detection of 5caC and 5fC is simultaneously possible by treatment of the
DNA with bacterial methylase M. SssI, which methylates all unmodified CG dinucleotides to give 5mCG. Following methylation to protect all unmodified cytosine in
a CG context, bisulfite treatment of DNA will lead to deamination of 5fC and 5caC.
This results in sequencing output that can identify the locations of 5fC and 5caC
together. MAB-Seq is reliant on quantitative conversion of C to 5mC by M. SssI and
is limited to detection of 5fC and 5caC only in the CG context. MAB-Seq was
further developed to be specific for 5caC (caMAB-Seq) (Figs. 8 and 9), which takes
advantage of the reduction of 5fC by sodium borohydride prior to performing the
steps of MAB-Seq [120]. Using MAB-Seq and caMAB-Seq, 5fC and 5caC have
been mapped at single-base resolution to regions with higher chromatin accessibility
and therefore TET accessibility [121]. The sequencing profiles obtained via
MAB-Seq and caMAB-Seq indicate that there might be preferential locations for
TET activity, either through chromatin accessibility or sequence-dependent removal
of 5fC and 5caC by TDG [120].
The first report for mapping 4mC along the genome employed a modification of
the TAB-Seq methodology [33]. This is possible because 4mC undergoes deamination at a rate between C and 5mC [33]. To allow for 4mC detection, the bisulfite
conditions were optimized to ensure the retention of 4mC in the genome while
deaminating TET-generated 5caC. Using this methodology, 4mC was mapped in
C. kristjanssonii, a bacterium with similar levels of 5mC and 4mC [33]. This
methodology, 4mC-TAB-Seq, is well suited for identification of 4mC in genomes
which have high concentrations of 4mC at particular sites, especially in bacteria
which utilize 4mC as part of their restriction-modification systems [31]. Application
of this methodology to mammalian cells and tissues is limited by the incomplete
(40–55%) retention of 4mC after bisulfite conversion, making accurate identification
of sites possessing less than complete methylation nearly impossible [33].
Mapping N6MedA can also be accomplished using restriction enzyme-based
sequencing (6mA-RE-Seq) [122]. This sequencing technique takes advantage of
previous information gained from antibody enrichment methods, which found that
N6MedA was found in CATG and CATC motifs in the C. reinhardtii genome, a
510
C. L. Seiler et al.