In humans, DNA methylation is the most stable epigenetic mark [1], and it occurs
at C5 position of cytosine (5mC), mainly in a CpG dinucleotide context. The CpG
dinucleotides are mainly located in the CpG islands (occurring at ca 60% of all gene
promoters), in repeated sequences and in CpG island shores [2]. If promoter CpG
islands are methylated, the corresponding gene is repressed due to a poor recognition
by transcription factors and recruitment of proteins involved in chromatin
remodelling such as methyl DNA-binding proteins (MBPs) [3].
Failure in maintaining DNA methylation and establishment of new DNA methylation patterns are associated with under- or overexpression of the affected genes,
ultimately leading to inflammation, cancer and other diseases. DNA methylation is
catalysed by DNA methyltransferases (DNMT) that mediate the transfer of a methyl
group from the S-adenosyl-L-methionine (SAM or AdoMet, 1) to position 5 of
cytosine in DNA [4] (Figs. 1 and 2a).
The ten-eleven translocator (TET) enzymes catalyse the 2-oxoglutarate (2OG, 3)dependent oxidation of 5mC in a cascade of iterative steps to give 5-hydroxymethyl
cytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxycytosine (5caC) (Figs. 1 and 2).
The discovery of TET oxidation of 5mC to 5hmC in 2009 garnered interest as an
important epigenetic regulator [5, 6]. Since the identification of 5hmC in human cells,
other DNA modifications (5fC, 5caC as well as 5-hydroxymethyluracil (5hmU)) have
been identified and found to be catalysed by TET1–3 [7–11].
Methylation at position 5 of cytosine is a reversible modification, and there are
four different demethylation pathways that could lead to the reformation of cytosine
identified to date (Fig. 1). First, the 5mC mark is not maintained during DNA
replication, thus 5mC is diluted by subsequent rounds of cell division, a process
known as passive DNA demethylation. Second, 5mC can be iteratively modified to
provide 5hmC, 5fC and 5caC, which then can be diluted in cell division, a combination of active modification followed by passive dilution (AM-PD). The third
process is an active pathway via a combination of thymidine-DNA glycosylase
(TDG) and base excision repair (BER) enzymes. As part of DNA repair mechanism,
TDG recognises a thymidine guanosine pair mismatch and deaminates the thymidine
base leaving an a-basic sugar. TDG likewise performs this excision efficiently on
5fC and 5caC [12] followed by DNA repair using the base excision repair (BER)
mechanism that introduces C. Finally, recent studies by Iwan et al. demonstrated, by
labelling the 5fC sugar and base independently, that 5fC can be directly converted by
the human cell to C without the change in base or sugar, suggesting direct
deformylation [13]. The protein(s) associated with this process of 5fC, however,
have not been identified to date. In plants, a direct 5mC to C pathway is present,
following a similar TDG-BER mechanism, where repressor of silencing 1 (ROS1)
recognises 5mC [14].
Many questions arise around these enzymes and their roles in regulating cytosine
modifications, which are very dynamic but are specific in the location and context
(tissues, diseases, etc.). How are they precisely regulated in a concerted manner?
Why was this mechanism of DNA demethylation selected in mammals? Their role in
diseases, such as cancer, neuronal diseases, inflammation and infection, is well
established, but the mechanisms involved are still to be fully understood.
258
R. Belle et al.
at C5 position of cytosine (5mC), mainly in a CpG dinucleotide context. The CpG
dinucleotides are mainly located in the CpG islands (occurring at ca 60% of all gene
promoters), in repeated sequences and in CpG island shores [2]. If promoter CpG
islands are methylated, the corresponding gene is repressed due to a poor recognition
by transcription factors and recruitment of proteins involved in chromatin
remodelling such as methyl DNA-binding proteins (MBPs) [3].
Failure in maintaining DNA methylation and establishment of new DNA methylation patterns are associated with under- or overexpression of the affected genes,
ultimately leading to inflammation, cancer and other diseases. DNA methylation is
catalysed by DNA methyltransferases (DNMT) that mediate the transfer of a methyl
group from the S-adenosyl-L-methionine (SAM or AdoMet, 1) to position 5 of
cytosine in DNA [4] (Figs. 1 and 2a).
The ten-eleven translocator (TET) enzymes catalyse the 2-oxoglutarate (2OG, 3)dependent oxidation of 5mC in a cascade of iterative steps to give 5-hydroxymethyl
cytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxycytosine (5caC) (Figs. 1 and 2).
The discovery of TET oxidation of 5mC to 5hmC in 2009 garnered interest as an
important epigenetic regulator [5, 6]. Since the identification of 5hmC in human cells,
other DNA modifications (5fC, 5caC as well as 5-hydroxymethyluracil (5hmU)) have
been identified and found to be catalysed by TET1–3 [7–11].
Methylation at position 5 of cytosine is a reversible modification, and there are
four different demethylation pathways that could lead to the reformation of cytosine
identified to date (Fig. 1). First, the 5mC mark is not maintained during DNA
replication, thus 5mC is diluted by subsequent rounds of cell division, a process
known as passive DNA demethylation. Second, 5mC can be iteratively modified to
provide 5hmC, 5fC and 5caC, which then can be diluted in cell division, a combination of active modification followed by passive dilution (AM-PD). The third
process is an active pathway via a combination of thymidine-DNA glycosylase
(TDG) and base excision repair (BER) enzymes. As part of DNA repair mechanism,
TDG recognises a thymidine guanosine pair mismatch and deaminates the thymidine
base leaving an a-basic sugar. TDG likewise performs this excision efficiently on
5fC and 5caC [12] followed by DNA repair using the base excision repair (BER)
mechanism that introduces C. Finally, recent studies by Iwan et al. demonstrated, by
labelling the 5fC sugar and base independently, that 5fC can be directly converted by
the human cell to C without the change in base or sugar, suggesting direct
deformylation [13]. The protein(s) associated with this process of 5fC, however,
have not been identified to date. In plants, a direct 5mC to C pathway is present,
following a similar TDG-BER mechanism, where repressor of silencing 1 (ROS1)
recognises 5mC [14].
Many questions arise around these enzymes and their roles in regulating cytosine
modifications, which are very dynamic but are specific in the location and context
(tissues, diseases, etc.). How are they precisely regulated in a concerted manner?
Why was this mechanism of DNA demethylation selected in mammals? Their role in
diseases, such as cancer, neuronal diseases, inflammation and infection, is well
established, but the mechanisms involved are still to be fully understood.
258
R. Belle et al.