Mutation Research 764–765 (2014) 18–35
Contents lists available at ScienceDirect
Mutation Research/Genetic Toxicology and
Environmental Mutagenesis
journal homepage: www.elsevier.com/locate/gentox
Community address: www.elsevier.com/locate/mutres
TET enzymatic oxidation of 5-methylcytosine,
5-hydroxymethylcytosine and 5-formylcytosine
Jean Cadet a,b,∗ , J. Richard Wagner b,∗
a
b
Direction des Sciences de la Matière, Institut Nanosciences et Cryogénie, CEA/Grenoble, 38054 Grenoble, France
Département de médecine nucléaire et radiobiologie, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, Québec JIH 5N4, Canada
a r t i c l e
i n f o
Article history:
Received 1 September 2013
Accepted 4 September 2013
Available online 14 September 2013
Keywords:
Epigenetic marks
Radical oxidation reactions
DNA glycosylase-mediated repair
Hydrolytic and enzymatic deamination
Biological role
a b s t r a c t
5-Methylcytosine and methylated histones have been considered for a long time as stable epigenetic
marks of chromatin involved in gene regulation. This concept has been recently revisited with the detection of large amounts of 5-hydroxymethylcytosine, now considered as the sixth DNA base, in mouse
embryonic stem cells, Purkinje neurons and brain tissues. The dioxygenases that belong to the ten
eleven translocation (TET) oxygenase family have been shown to initiate the formation of this methyl
oxidation product of 5-methylcytosine that is also generated although far less efficiently by radical reactions involving hydroxyl radical and one-electron oxidants. It was found as additional striking data
that iterative TET-mediated oxidation of 5-hydroxymethylcytosine gives rise to 5-formylcytosine and
5-carboxylcytosine. This survey focuses on chemical and biochemical aspects of the enzymatic oxidation
reactions of 5-methylcytosine that are likely to be involved in active demethylation pathways through
the implication of enzymatic deamination of 5-methylcytosine oxidation products and/or several base
excision repair enzymes. The high biological relevance of the latter modified bases explains why major
efforts have been devoted to the design of a broad range of assays aimed at measuring globally or at the
single base resolution, 5-hydroxymethylcytosine and the two other oxidation products in the DNA of
cells and tissues. Another critical issue that is addressed in this review article deals with the assessment
of the possible role of 5-methylcytosine oxidation products, when present in elevated amounts in cellular DNA, in terms of mutagenesis and interference with key cellular enzymes including DNA and RNA
polymerases.
© 2013 Elsevier B.V. All rights reserved.
Abbreviations: ␣-KG, ␣-ketoglutarate; AML, acute myeloid leukemia; APOBEC, apolipoprotein B mRNA-editing enzyme complex; ARP, aldehyde reactive probe;
ˇ-GT, ˇ-glucosyl transferase; 5-caC, 5-carboxylcytosine; 5-cadCyd, 5-carboxyl-2 -deoxycytidine; C, cytosine; C5-MTases, cytosine-5 methyltransferases; DNMT, DIP,
DNA immunoprecipitation; DNA, methyltransferase; DSB4, double stranded ␤-helix; CAN, ceric(IV) ammonium nitrate; CO3 • − , carbonate anion radical; CE-LIF,
capillary electrophoresis with laser-induced fluorescence; CpGIs, CpG islands; DKO, double-knockout; dUrd, 2 -deoxyuridine; ESI-MS, electrospray ionization-mass spectrometry; 5-fC, 5-formylcytosine; 5-fdCyd, 5-formyl-2 -deoxycytidine; 5-fU, 5-formyluracil; FTO, obesity-associated protein; 5-glu-5-hmC, glucosylated derivative of
5-hydroxymethylcytosine; HEK, human embryonic kidney; hMLH1, human mut L homolog 1; HPLC-ESI-MS/MS, high-performance liquid chromatography-electrospray
ionization-tandem mass spectrometry; 5-hmC, 5-hydroxymethylcytosine; 5-hmdCyd, 5-(hydroxymethyl)-2 -deoxycytidine; 5-hpmdCyd 5-methylcytosine, 5-mC, 5hydroxymethyluracil, 5-hmU; 5-hpdCyd, 5-(hydroperoxymethyl)-2 -deoxycytidine, 5-iodo-2 -deoxycytidine, 5-IdCyd; hNTH1, human nth endonuclease III-like 1; hOGG1,
human oxoguanine DNA glycosylase 1; hUNG2, human uracil DNA glycosylase 1; IDH, isocitrate dehydrogenase; LCK, leukemia-associated protein with CXXC domain; MBD4,
methyl-binding domain protein 4; MQ, 2-methyl-1,4-naphthoquinone; MGMT, O6 -methylguanine-DNA-methyltransferase; MRM, multiple reaction monitoring; Na2 S2 O8 ,
sodium peroxosulfate; NP, neuronal progenitor; • NO, nitric oxide; 1 O2 , singlet oxygen; O2 • − , superoxide anion radical; • OH, hydroxyl radical; ODN, oligodeoxynucleotide;
oxBS-Seq, oxidative bisulfite sequencing; Pol II, polymerase II; ROS, reactive oxygen species; RRBS, reduced representative bisulfite sequence; SAM or AdoMet, S-adenosylL-methionine; SMUG1, single-strand specific monofunctional DNA glycosylase I; SMRT, single-molecule real-time DNA sequencing; T7H, TBDMS, tert-butyldimethylsilyl;
T7H, thymine 7-hydroxylase; TDG, thymine DNA glycosylase; TET, Ten eleven translocation; UDP-glucose, uridine 5 -diphospho-d-glucose.
∗ Corresponding authors.
E-mail addresses: jean.cadet@cea.fr (J. Cadet), richard.wagner@usherbrooke.ca (J.R. Wagner).
1383-5718/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.mrgentox.2013.09.001
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
1. Introduction
DNA methylation and histone modifications constitute well
documented inheritable changes in chromatin that participate
in the epigenetic fate of cells and repression of intragenic transcription [1–4]. DNA methylation is mediated in mammals by
cytosine-5 methyltransferases (C5-MTases) [5,6] and five enzymes
of the DNA methyltransferase (DNMT) family [7–9] that allow
the transfer of a methyl group from the ubiquitous co-factor
S-adenosyl-L-methionine (SAM or AdoMet) to the carbon 5 of cytosine, preferentially in the context of CpG motifs while CpG islands
(CGIs) are typically unmethylated [5–9]. C5 Methylation of cytosine has also been shown to occur at CpA sites in oocytes [10].
This gives rise to 5-methylcytosine (5-mC), identified as the fifth
nucleobase [11,12] that was discovered more than 60 years ago
[13,14] and considered until recently as a crucial and relatively stable epigenetic mark [15,16]. 5-mC has been found to impact several
essential biological pathways including among others regulation
of gene expression, maintenance of chromatin structure, parental
imprinting, X-chromosome inactivation, control of cell development and disease pathogenesis [17–21]. 5-Hydroxymethylcytosine
(5-hmC), an oxidation product of 5-mC that can be generated by
either hydroxyl radical (• OH) or one-electron oxidants [22] was
initially found in T-even bacteriophage [23,24] and mammalian
cells [25] using either mild acid hydrolysis or enzymatic digestion of DNA combined with paper chromatography before being
almost completely ignored after unsuccessful attempts to detect
this oxidized base in vertebrate DNA [26]. However the concept of
5-mC being a stable and static epigenetic modification has been
questioned after the rediscovery in 2009 of 5-hmC found in substantial amounts in genomic DNA of mouse embryonic stem cells
[27] and in mouse Purkinje neurons and brain [28] by two independent groups. In addition, ten eleven translocation 1 (TET1), one of
three enzymes of the TET family, was shown to be involved in the
oxidation of 5-mC into 5-hmC [26], now defined as the sixth nucleobase [29–31]. These major findings in the field of epigenetics have
provided a strong impetus toward numerous and extensive studies
aimed at better assessing the ability of TET enzymes to oxidize 5mC in various cells and tissues and determining the role of 5-hmC
in the epigenome. Another major issue that rapidly emerged was
the possible implication of 5-hmC and further oxidation products
including 5-formylcytosine (5-fC) and 5-carboxylcytosine (5-caC)
(for recent review articles, see [32–34]) as short-lived intermediates of active removal pathways of 5-mC and as a more dynamic
epigenetic regulation with respect to the well documented passive demethylation process involving DNA replication [35–37]. The
available data reported during the last three years on the formation of 5-mC oxidation products and the proposed demethylation
mechanisms, two major fast evolving research topics, are critically reviewed with emphasis on chemical and biochemical aspects.
This review article starts by a critical survey of the main methods
involving chromatographic, enzymatic and immuno-based analytical techniques that have been developed for globally measuring
5-mC oxidation products in the DNA of isolated cells and tissues.
They are described and evaluated in terms of accuracy and sensitivity since the reliability of some of the reported assays have been
questioned [38]. Similar to unprecedented efforts applied to measure oxidatively generated damage to DNA, several methods have
been designed for mapping 5-hmC at the nucleotide level in order
to gain further insights into the putative implication of the latter
oxidized pyrimidine base in active demethylation pathways. This
review is completed by presentation of the main available data on
the depletion of the levels of 5-hmC in cancer tissues that could
be used as a potential prognostic biomarker. In addition, recently
available information is provided on the biological effects of 5-mC
oxidation products including their low mutation potential in vitro.
19
It was recently shown that N6 -methyladenine, an alkylated base
present in substantial amounts in messenger RNA but not in the
DNA of mammals, is also subjected to demethylation, a dynamic
reversible process likely to be implicated in cellular regulation [39].
This involves, as for the demethylation of 5-mC by TET proteins
[40], enzymes that belong to the iron (II)- and ␣-ketoglutarate (␣KG) dependent dioxygenase family proteins. Thus, the fat mass
and obesity-associated protein FTO [39,41] and ALKBH5 [42] in
mammals are able to demethylate RNA both in vitro and in vivo.
Oxidative demethylation is expected to give rise to unstable N6 hydroxymethyadenine intermediate exhibiting an acetal structure
that is highly susceptible to hydrolysis leading to restitution of initial adenine as previously observed for other ALKB enzymes that are
involved in the repair of N1 -alkyladenine [42–44]. These reactions
will not be discussed further since emphasis in the present survey
is placed on the oxidation products of 5-mC and their biochemical
features.
2. Analysis of oxidation products of 5-methylcytosine
Major efforts have been devoted during the last 3 years toward
the development of methods aimed at measuring methyl oxidation products of 5-mC in cells and tissue after the rediscovery of
5-hmC in ES cells and neuronal tissues that was achieved using poor
resolution and semi-quantitative restriction enzyme based assays.
More accurate and quantitative analytical approaches including
high-performance liquid chromatography (HPLC) based methods
have become available whose benefits and limitations are critically
reviewed here. In addition a strong impetus has been given to the
design of several sequencing techniques aimed at specifically mapping 5-mC and the three oxidation products at the nucleotide level
for gaining more insight into the functional role of the latter epigenic marks. A comparison of the different methods is presented in
Table 1.
2.1. Thin-layer chromatography analysis of [32 P]-labelled 5-mC
oxidation products
Two restriction enzyme based methods that included sequence
specific digestion of DNA followed by a 32 P-phosphorylation
step before thin-layer chromatography (TLC) analysis of [32 P]radiolabeled nucleoside monophosphates were used for the
discovery of 5-hmC in mouse ES and neuronal cells [27,28]. Application of an adaptation of the so-called nearest-neighbor analysis of
nucleoside X at XpG sites [45] required incubation of DNA with Fok
1 restriction enzyme and detection of the novel 3 -monophosphate
of 5-hmC by 2D-TLC analysis that was later confirmed by ESIMS analysis [28]. The second assay leading to the isolation of
5-(hydroxymethyl)-2 -deoxycytidine 5-monophosphate and characterization by ESI-MS analysis involved DNA digestion by Msps1
restriction enzyme [27]. In subsequent studies, either Taq␣ 1 [46] or
EcoNI [47] was used as the restriction enzyme to reveal in cellular
DNA 5-fC and 5-caC that were shown to prevent Taq␣ 1-mediated
DNA cleavage. Global measurement of the three 5-mC oxidation
products provided by the latter Taq␣ 1 assay in CpG contexts constitutes a major advantage although it is counterbalanced by a lack
of accurate quantification.
2.2. Enzymatic derivatization 5-hmC based assays
Another global enzymatic assay aimed at measuring 5-hmC
sites in DNA is based on glucosylation of the 5-hydroxymethyl
group, which is achieved by T4 phage ˇ-glucosyl transferase
(ˇ-GT) using uridine 5 -diphospho-d-glucose (UDP-glucose) as
the co-factor [48]. The incorporation of [3 H]-labeled glucose into
20
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
Table 1
Comparison of analytical methods for measuring oxidation products of 5-mC in cellular DNA.
Method
References
Description
Advantages
Disadvantages
[32 P]-Labelling
[27,28,46,47]
Requires simple equipment
Glucosylation
[48,49,52,53]
Semi-quantitative measurement
and poor characterization
Autooxidation
5-hmU is also a substrate
Immunoassay
[55–61,63–81]
LC-MS/MS (Fig. 2)
[46,47,72,85–93,121,236]
Labeling involves the use of
restriction enzymes - TLC analysis
Label with glucose transferase:
[3 H]-glucose; dot blot with
enzymatic detection
ELISA, immuno dot blot;
Immuno-fluorescence,
histochemistry
Requires enzymatic digestion of
DNA
CAP-Fl
[95]
Improves specificity of detection
Easy to perform
Capillary electrophoresis with
fluorescence detection
Sequencing-quantitative mapping of 5-mC and 5-hmC with single-nucleotide resolution
[107]
Glucosylation + specific cleavage
Aba-seq
[110] (Yu)
Oxidation by TET; glu-C not
TAB-seq
oxidized
oxBS-seq
[110]
KRuO4 oxidation of 5-hmC to 5-fC
[103,112–115]
Glucosylation + periodate or click
GLIB and others
chemistry
SMRT
[112]
Sanger sequencing based on
differences in polymerization
DNA by ˇ-GT expressed in bacteria was shown to be quantitative, thus allowing accurate measurement of 5-hmC in various
mouse tissues with yields that decrease in the following order:
cortex > hippocampus > cerebellum > kidney > lung > liver > testis >
spleen > thymus [48]. An optimized version of the assay that
also provides partial characterization of recombinant ˇ-GT and
insights into biochemical parameters including kinetic aspects
of the glucosylation reaction is now available [49]. This constitutes a robust and relatively easy method although it requires
dedicated equipment for safely measuring radiolabeled DNA. The
selectivity of the measurement, however, is not absolute since
5-hydroxymethyluracil, a radical oxidation product of thymine
that may be generated in relatively low yield upon exposure to
oxidizing agents [50], is also a substrate for the ˇ-GT reaction
[51]. As an extension of the latter assay based on ˇ-GT-mediated
glucosylation of 5-hmC, labeling of the 5-hydroxyl group of the
base was achieved using a 6-azide substituted glucose as part of
a bioorthogonal strategy [52,53]. In a subsequent step, biotin is
attached by copper-free click chemistry that may allow detection,
affinity purification and DNA sequencing. Thus, the distribution of
5-hmC was determined in the genome of cell lines and mouse tissues with the highest yield in cerebellum DNA. This was achieved
using a dot-blot assay involving horseradish peroxidase [52].
Another enzymatic based assay has been proposed to label the
hydroxymethyl group of 5-hmC [53,54]. DNA cystosine-5 methyltransferases (C5-MTases) have been shown to allow direct coupling
of several thiols and selenols to the hydroxyl group of 5-hmC
yielding the corresponding 5-chalcogenomethyl derivatives. The
resulting modified 5-hmC 2 -deoxyribonucleoside and related 5 phosphomonester derivatives were separated by HPLC and TLC
analysis, respectively [53]. This provides through the ligation of
a biotin residue to the C5 attached functional group a possibility
of analysing all CpG sequences for 5-hmC content when M.SssI
methyltransferase is used as the coupling enzyme. However, the
overall strategy remains to be validated [51].
2.3. Immunoassays
Several antibodies, mostly polyclonal, have been raised against
5-hmC and to a lesser extent against 5-fC and 5-caC, and they are
in most cases commercially available. The rabbit polyclonal (Active
Motif) has been successfully used in immuno-dot blot assays
High sensitivity;
0.1-10 ␮g DNA
Sub-picomole
High sensitivity;
0.45 amole for 1 ␮g DNA
Lack of specificity
Cross-reactivity with C and 5-mC
Lack of quantification
Requires expensive equipment
Lack of specificty
Good recovery and sensitivity
Good recovery and sensitivity
Ambiguity at site of cleavage
Depends on enzyme activity
High conversion efficiency
High recovery; diversity of affinity
probes and labels
Great potential
Possibility of side reactions
Possibility of side reactions
Requires extensive 5-hmC
enrichment; expensive
for detecting 5-hmC in various cellular DNA and animal tissues
[55–58]. In agreement with previous observations, the distribution of 5-hmC in mouse tissue was tissue-specific being elevated in
brain, moderate in breast, liver and testis while very low in colon,
spleen and blood [58]. Another commercially developed polyclonal
antibody has been used for monitoring the distribution of 5-hmC
in several human tissues [59]. The level of 5-hmC in colon was
reported to be slightly lower than that in brain [59]. A different
affinity purified rabbit polyclonal antibody, anti 5-hmC, was specific as inferred from enzyme linked immunosorbent (ELISA) and
immunoprecipitation assay measurements [60]. The measurement
of 5-hmC in genomic DNA using either immuno dot-blot techniques or ELISA appeared to be informative [61] although being
only semi-quantitative due to the lack of calibration. The specificity of these antibodies may be questioned as it is the case for
most of immunoassays against oxidized nucleobases and nucleosides [62], which are generally far less antigenic than bulky DNA
modifications such as cyclobutane pyrimidine photoproducts or
benzo[a]pyrene adducts to guanine. Interestingly antibodies raised
against cytosine methylenesuIphate generated by the reaction of
5-hmC with bisulfite are more sensitive and less density dependent in terms of quantitative analysis than 5-hmC antibodies [63].
It was shown that mouse monoclonal anti 5-hmC (Eurogenentec)
cannot immunoprecipitate a 76-mer oligonucleotide bearing one
unique 5-hmC residue, whereas by comparison, anti-5-mC antibody shows a much higher selectivity toward 5-mC in the same
76-mer [64]. The latter 5-hmC antibody was also found to be
inappropriate for immunoprecipitation experiments with a longer
DNA probe (949 bp) [55]. Rabbit polyclonal (Active Motif) shows
a higher ability to precipitate 5-hmC containing DNA at three
TET1 bound targets than the rat monoclonal antibody (Diagenode)
[65].
Other relevant applications of mostly polyclonal antibody
(Active Motif) deal with immunofluorescence detection of 5-hmC
in mouse tissues [30], ES cells [57,66], zygotes [56,67–69], oocytes
[68,69], primordial germ cells [70–72] and neuronal cells [73,74].
Immunostaining techniques were also used to show a significant
decrease of 5-hmC in malignant glioma [75], melanoma [76,77] and
carcinoma of the breast, colon and prostate [78].
The search for 5-fC and 5-caC in genomic DNA using immunoassays has received far less attention so far than that for 5-hmC. Two
rabbit polyclonal antibodies against 5-fC and 5-caC have been prepared and their specificity was high when checked by dot-blot
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
analysis [79]. However, as for 5-hmC, the 5-fC antibody was not
found to be suitable for immunoprecipation experiments since
it shows only 1.6-fold discrimination between DNA probes containing one 5-fC residue to those with none [80]. However, both
antibodies directed against 5-fC and 5-caC were used for immunostaining experiments with cells. The formation of 5-fC and 5-caC in
the paternal pronucleus of mouse zygotes was found to occur concomitantly with a decrease in the 5-mC content [75]. Furthermore,
both 5-hmC iterative oxidation products exhibited replicationdependent dilution during preimplantation development [79]. It
was also shown again on the basis of immunofluorescence measurements that the levels of 5-fC and 5-caC remained relatively
stable during germ cell reprogramming [71]. The specificity of
immunodetection of 5-caC in mouse zygotes [71,79] and follicular cells [81] that appear to be present in very low amounts is open
to debate due to the likely occurrence of cross-reactivity with 5-mC.
This will have to await further independent measurements with for
example reliable and quantitative HPLC-MS/MS.
2.4. HPLC-MS and HPLC-MS/MS methods
HPLC coupled with electrospray ionization tandem mass
spectrometry (HPLC-ESI-MS/MS) is a robust and highly accurate
analytical technique particularly when the measurements are
performed according to multiple reaction monitoring (MRM)
mode of detection with quantitative isotope dilution. Analysis
of the methyl oxidation products of 5-mC (5-hmC, 5-fC, 5-caC)
and T (5-hmU and 5-fU) can be achieved in a single run by DNA
extraction from cells, DNA digestion to 2 -deoxyribonucleosides
and LC-MS/MS analysis (see Fig. 2 for a typical example). This
is considered as the gold standard method [82] for monitoring
traces of modified bases formed in cellular DNA such as oxidation
products [62] and photo-induced lesions [83]. It may be added
that MS3 detection is necessary when the level of modifications
is below the fmol range, a situation that is present for 5-CaC and
DNA lesions generated by radical oxidation pathways [50]. This is
quite different for measuring 5-hmC in ES cells and brain tissues
in which the level is several residues per 102 guanines. In sharp
contrast, the level of 8-oxo-7,8-dihydroguanine, one of the most
frequent oxidatively generated lesions in cellular DNA that is in
most cases at least three to four orders of magnitude lower [84].
This explains why HPLC with only one quadrupole as the detector
was sufficient to monitor the formation of 5-hmC in several mouse
tissues and ES cells [29,30]. However, the sensitivity of the latter
assay was too low to detect the presence of 5-fC and 5-caC, which
are formed in much lower amounts than 5-hmC. Therefore, the use
of HPLC-MS/MS with isotopically labeled internal standards has
allowed for the discovery of 5-fC in the DNA of ES cells [85]. Almost
immediately this received confirmation from HPLC-ESI-MS/MS
analysis that also shows the presence of 5-caC in genomic DNA of
ES cells and mouse tissues [46]. It was also concomitantly reported,
again on the basis of HPLC-ESI-MS/MS measurements, that 5-caC
accumulates in ES cells upon depletion of thymine DNA glycosylase
repair enzyme [47]. As a highly positive trend, there is a significant
increase in the use of HPLC-ESI-MS/MS methods [72,86–88] for
measuring 5-hmC together with 5-mC. Using this accurate analytical tool, a decrease in the level of 5-hmC was observed in the DNA
from myeloproliferative neoplastic patients [89] and human brain
tumors [90]. Other measurements have shown that the amounts
of 5-hmC increase in embryonic mouse brain during neuronal
differentiation [91]. In addition the global amounts of 5-hmC have
been quantitatively determined in 8 types of mouse spermatogenic
cells using the HPLC-MS/MS analytical tool [92]. Levels of 5-hmC
were measured in yeast (0.0004-0.3447 5-hmC/C) with a gain
in sensitivity (compared to HPLC-MS/MS) by direct infusion on
MS/MS after derivatisation with T4 phage B-glucosyltransferase
21
and prepurification of polar 5-hmC on a hydrophilic NH2 -silica
SPE column [93]. The levels of 5-fC assessed by HPLC-MS/MS were
found to be within the range 0.02 to 0.002% of guanines in mouse ES
cells [80], which are 1 to 2 orders of magnitude lower than those of
5-hmC [29]. Measurements of 5-hmC together with 5-fC and 5-caC
were carried out in the DNA of mouse ES cells as part of a comprehensive study aimed at identifying dynamic readers of 5-mC
oxidation products [90]. A capillary hydrophilic-interaction liquid
chromatography (cHLIC) method that includes a trapping column
and an ESI-quadrupole time-of-flight MS detector was designed
for the measurement of 5-mC and 5-hmC in genomic DNA from
human hepatocellular carcinoma tissues. The limit of detection of
5-mC and 5-hmC was 0.06 and 0.19 fmol, respectively [94].
2.5. Capillary electrophoresis with laser-induced fluorescence
detection of 5-hmC
A novel highly sensitive method for the measurement of
5-hmC has been designed with a limit of detection close to
0.45 amol for 1 ␮g of DNA [95]. The separation and detection are achieved by capillary electrophoresis with laser-induced
fluorescence (CE-LIF). This requires derivatization of either
the 3 - or 5 -phosphoester of 5-hmdUrd with the fluorescent
dye, 4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3propionyl ethylene diimine hydrochloride (BODIPY FL EDA). The
method has been applied for the measurement of 5-hmC in human
tissue and human cancer cell lines [95].
2.6. Sequencing methods
There are well-established protocols for genome wide methylation profiling [4,37,96–101]. Basically, the protocols involve
enrichment of methylated DNA using antibodies or methyl binding
proteins together with chemical or biochemical methods to discriminate between methylated and non-methylated DNA bases at
specific sites during sequencing. In the case of cytosine and 5-mC,
the two bases are distinguished by reaction with bisulfite, which
induces the deamination of cytosine but not 5-mC. The site of methylation is then obtained by examining changes in the sequence
induced by cytosine deamination. With new generation sequencing
techniques, several hundred million fragments can be sequenced
giving a good site-specific representation of 5-mC (and 5-hmC) in
the genome. Here, we will focus on progress that has been made in
the past couple of years on the development of assays that map
5-hmC in the genome. The basic method using bisulfite cannot
be used because both 5-mC and 5-hmC are resistant to deamination under conditions that deaminate cytosine. A number of other
methods have been developed to discriminate between 5-mC and
5-hmC: 1) recognition of 5-hmC or derivatives by specific antibodies, 2) conversion of 5-hmC to glucosylated 5-hmC (5-glu-5-hmC)
by ␤-glucosyl transferase, and 3) the oxidation of 5-hmC to 5-fC
and 5-caC by TET and chemical agents.
Antibody based assays have recently been applied to enrich DNA
fragments containing 5-hmC by DNA immunoprecipitation (DIP)
[57,60,75,102]. Fragments are precipitated in single stranded form
to favor interactions with 5-hmC. Although the method is relatively
easy and rapid, it suffers from the general lack of discrimination
with DNA fragments containing 5-mC and the inability to discern
the exact position of 5-hmC. The methylation peaks are defined in
a region according to the number of consecutive positive probes
(e.g., 4) and the length of DNA fragments (e.g., 350 bp). In addition,
the method favors densely populated CpG sites and fragments with
CA and CT repeats co-precipitate as artifacts. 5-hmC-DIP has also
been performed with antibodies against a derivative of 5-hmC, 5methylenesulfonate, which is the main product from the reaction of
bisulfite with 5-hmC [103,104]. In comparison to 5-hmC antibodies,
22
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
the use of 5-methylenesulfonate antibodies reduces background,
prevents CA and CT enrichment, and does not depend as much on
5-hmC density. Robertson et al. [105,106] identified a protein in
kinetoplasmids that specifically binds to ␤-5-glu-5-hmC, known
as J binding protein 1, and proposed a method to pull down glucosylated DNA fragments using JBP1-coated magnetic beads. As with
other assays with antibody enrichment, however, there is a bias
toward the precipitation of DNA fragments containing a high density of 5-hmC and thus these fragments may be overrepresented.
Several assays to map the position of 5-hmC in the genome are
based on the specific glucosylation of 5-hmC with UDP-6-glucose
and ␤-glucosyltransferase. One method referred to as Aba-seq uses
a restriction enzyme (AbaSI) to specifically cleave DNA fragments
containing glucosylated 5-hmC [107]. More recently 20 related proteins considered as PyuRtsII homologues have been shown to have
interesting potential for mapping the hydroxymethylone [108].
AbaSI cleaves in a narrow range of distances from 5-glu-5-hmC,
which leads to some ambiguity about the position of cleavage [104].
Nevertheless, this method showed an unprecedented number of
putative 5-hmC marks (> 12 M) in the genome of mouse ES cells
when coupled to high-throughput parallel sequencing. In comparison, the TET-assisted bisulfite sequencing (TAB-Seq) method
involves initial glucosylation of 5-hmC sites followed by oxidation
of 5-mC using purified TET enzymes [109]. Site-specific identification of C, 5-mC and 5-hmC is accomplished by subsequent
bisulfite treatment and sequencing. The conversion of 5-hmC to
5-glu-5-hmC protects 5-hmC from oxidation by TET. Thus, 5-hmC
is detected as C whereas C and 5-mC undergo deamination and are
detected as T in sequence analysis.
5-hmC is more sensitive to oxidation than 5-mC. Based on this
difference, Booth et al. [110] developed an assay to specifically
measure 5-hmC by oxidizing 5-hmC to 5-fC in DNA fragments with
KRuO4 referred to as oxidative bisulfite sequencing (oxBS-Seq).
The method reads out the absolute abundance of 5-hmC in DNA
fragments with single-base resolution using a library of reduced
representation bisulfite sequences (RRBS) including 800 representative CpG islands in mouse ES cells. The resulting fragments
containing 5-fC were sensitive to treatment with bisulfite giving
a change in sequence from C to T during DNA sequencing. RRBS
and oxRRBSs were run in parallel to discriminate between 5-mC
and 5-hmC. An alternative method was recently proposed for the
specific sequencing of 5-caC at single base resolution [111]. This
method involves the coupling of a primary amine containing a
probe and affinity label with the carboxyl group of 5-caC, resulting
in an amide that is resistant to bisulfite treatment. Two other
strategies to identify 5-hmC in the genome involve the attachment
of probes onto the glucose moiety of 5-glu-5-hmC. In one method,
a modified UDP-6-glucose molecule containing an azide group
is incorporated into 5-hmC for subsequent modification by click
chemistry [112–114]. Several substituents may be incorporated
to improve the purification (biotin) or the detection (fluorescent markers) of fragments containing 5-hmC [53]. This method
is compatible with highly sensitive single-molecule real-time
sequencing [112], which determines the position of glycosylated
and further modified 5-hmC residues by pauses in polymerization
and significant differences in the kinetics of polymerization with
DNA fragments containing 5-glu-5-hmC and bulky derivatives
of 5-glu-5-hmC. Interestingly, this method reveals prominent
5-hmC peaks on opposite strands of CpG dinucleotides, suggesting
a bias for TET oxidation of hemihydroxymethylated contexts.
Lastly, another recently proposed method involves glucosylation,
periodate oxidation and biotinylation (GLIB) [103,115]. It is based
on oxidation of the glucose moiety of 5-glu-5-hmC by sodium periodate, which gives an aldehyde product that can subsequently be
labeled with biotin using commercially available aldehyde reactive
probe (ARP). The authors report nearly quantitative precipitation
of DNA fragments that contain a single 5-hmC. A disadvantage is
that the method produces a high background due to secondary
oxidation of DNA by sodium periodate. Mild NaBH4 reduction of
the formyl group of 5-fC that allows its quantitative conversion
into 5-hmC has been used to map the 5-fC epigenetic mark in
mouse ESCs [116] using the 5-hmC-selective chemical labeling
strategy involving the coupling of an azide-modified glucose [52].
This requires the initial blocking of the enzymatically generated
5-hmC by ˇ-GT catalyzed coupling with unmodified glucose.
Although several methods including bisulfite-mediated thiol
substitution and subsequent coupling of biotin together with
nanopore analysis [117] have been proposed, the site-specific analysis of 5-hmC is still in development and no comparison has been
carried out to choose the most appropriate method for absolute and
sensitive sequencing purposes. The main challenges today lie in
the low percent of precipitation of DNA fragments, the difficulty to
sequence high density regions, and the quantitative transformation
of 5-hmC by biochemical and chemical agents.
3. Oxidation reactions of 5-mC
Two main sources of oxidation reactions of the methyl group
of 5-mC have been identified so far. Evidence for the presence of
5-hmC in the DNA of several organs of rats, mice and frogs, likely
as the result of enzymatic reactions, was provided more than 4
decades ago [25]. This receives further confirmation with the recent
discovery of TET enzymes in the hydroxylation of the methyl group
of 5-mC (Fig. 1). Comprehensive mechanisms are now available for
the oxidation reactions of 5-mC by • OH and one-electron oxidants
also giving rise among other oxidized bases to 5-hmC as further
detailed below.
3.1. TET-mediated oxidation reactions of 5-methylcytosine and
related oxidation products
There are three possible 5-mC oxidation products, which include
5-hmC, 5-formylcytosine (5-fC) and 5-carboxylcytosine (5-caC)
that have all been shown to be generated in cellular DNA by TET
proteins also referred to as 5-methylcytosine oxygenases.
3.1.1. 5-Hydroxymethylcytosine
The presence of an unknown nucleotide that showed
similar chromatographic properties to those of authentic 5(hydroxymethyl)-2 -deoxycytidine
5 -monophosphate
[118]
in the DNA of mouse cerebellum tissues was revealed using
two-dimensional (2D) TLC as part of the nearest-neighbour
analysis protocol [28]. This received confirmation from the
unambiguous characterization by electrospray ionization mass
spectrometry (ESI-MS) of a HPLC fraction of enzymatically digested
2 -deoxyribonucleosides as 5-(hydroxymethyl)-2 -deoxycytidine
(5-hmdCyd). The yield of 5-hmdCyd in Purkinje neuron and granule cells was estimated to be 0.6% and 0.2% respectively of total
2 -deoxyribonucleosides [28]. An inverse correlation was noted
between the levels of 5-hmCpG and 5mCdG in various organs.
Concomitantly, it was shown that overexpression of TET1 in
embryonic kidney HEK 293 cells leads to a decrease in 5-mC levels
while a new nucleotide was isolated by TLC after digestion by MspI
restriction endonuclease and [32 P]-post-labeling [27]. The new
compound was identified as the 5 -phosphate ester of 5-hmdCyd
by HPLC-ESI-tandem mass (MS/MS) analysis [27] after comparison
of the fragmentation pattern with that of the authentic sample
isolated from unglucosylated T4 DNA [14]. The presence of 5-hmC
in nuclear DNA was abolished when cells were expressed with
mutant TET1-CD, thus confirming the ability of TET1 to convert
5-mC into 5-hmC. Using the TLC analytical technique, 5-hmC was
found to be present in the DNA of mouse embryonic stem (ES)
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
23
SAM-DNMTs
DNMTs,MBD2
cytosine
NH2
N
DNA O
O
TDG
N
TDG
MBD4
SMUG1
pG DNA
CH3
HN
O
CH2OH
NH2
CO2H
N
O
N
5-hmU
NH2
TETs
CHO
N
O
5-caC
N
5-fC
N
O
O
N
CH3
N
thymine
O
TDG
NH2
O
DNA O
HN
TDG
AID/APOBEC
N
O
O
O
5-mC
O
pG DNA
AID/
APOBEC
TETs
NH2
TETs
CH2OH
N
O
N
5-hmC
Fig. 1. Pathways of cytosine methylation and 5-methylcytosine demethylation. C is methylated to 5-mC by DNA methyltransferases (DNMTs) using S-adenosyl-L-methionine
(SAM) as a methyl donor. Several pathways have been proposed to explain active demethylation of 5-mC (1-4): 1) single-enzyme removal of the methyl group by DNMTs
or DNA methyl binding domain protein 2 (MBD2); 2) deamination of 5-mC to thymine by activation induced deaminase (AID) or apolipoprotein B mRNA-editing enzyme
complex (APOBEC) followed by removal of thymine by thymine DNA glycosylase (TDG); 3) oxidation of 5-mC to 5-hydroxymethylcytosine (5-hmC) by ten eleven translocation
proteins (TETs) followed by deamination by AID/APOBEC and removal of 5-hydroxymethyluracil (5-hmU) by TDG, methyl binding DNA domain protein 4 (MBD4) or singlestrand-specific monofunctional uracil DNA glycosylase 1 (SMUG1); 4) iterative oxidation of 5-mC to further oxidation products, 5-fC or 5-caC, followed by removal of the latter
modifications by TDG. In addition, there may be several as yet not-well characterized proteins in mammalian cells that may catalyze the removal of different intermediates
in 5-mC demethylation.
cells but was not detectable in human T cells and mouse dendritic
cells [27]. Subsequent studies benefiting from the advent of accurate and quantitative HPLC-ESI-MS/MS [29,85] confirmed that the
highest levels of 5-hmC are present in the brain and ES cells of mice
and humans as critically reviewed in a recent survey [100]. On the
basis of immunostaining detection, it was also shown that genomic
5-hmC is present at high levels in brain and also in bone marrow
of mice [66]. Using the same analytical approach, information
was gained on the lineage-specific distribution of 5-hmC during
mammalian development of murine and human cells. Thus, 5-hmC
immunostaining that correlates with pluripotency was observed
in embryonic stem cells and lost upon differentiation [66]. Several next-generation sequencing methods aimed at specifically
mapping 5-hmC at the single nucleotide resolution in genes have
become recently available (for recent comprehensive reviews, see
[37,99–101]). As a striking result, one may quote that the levels of 5hmC in human genes when mapped using a quantitative qPCR assay
are primarily associated with the tissue type while gene expression
has a lower modulating effect [58]. The genome-wide distribution
of 5-hmC was assessed using a suitable selective chemical labeling
based assay showing that 5-hmC levels in specific gene bodies associated with neurodegenerative diseases increase with age in mouse
cerebellum [52]. The frequencies of 5-hmC were found to be more
elevated in synaptic genes than other genes in human and mice
brains [119]. In addition changes in the levels of 5-hmC were noted
at exon-intron boundaries in the frontal cortex of both human
and mouse brains. These observations are suggestive of a simultaneous influence of tissue specific distribution of 5-hmC on both
transcription and splicing [119]. Another key piece of information
was gained from application of the oxidative bisulfite sequencing
(oxBS-Seq) method that allowed mapping 5-hmC in CpGIs [110].
The presence of 5-hmC was only observed at CGIs with a preferential accumulation at intragenic, low CpC-density CGIs that were
suggested to play a role in the epigenetic reprogramming of ES
cells [110].
Following the initial discovery concerning the major role played
by TET1 in the conversion of 5-mC into 5-hmC in genomic DNA of ES
cells [27], numerous studies have been performed aimed at delineating the implication of the three TET enzymes (TET1, TET2 and
TET3). It was found that TET1 is required to maintain high levels of
5-hmC in mouse genomic regions close to CpG proximal promoters
[65].
In addition, evidence has been provided for implication of TET2
and TET3, the two other mammalian TET proteins, in hydroxylation
of the methyl group of 5-mC [36,51,55,120]. Knockdown of Tet1
and Tet2 genes in mouse ES cells was found to lead to a decrease
of 5-hmC with a concomitant increase in 5-mC content, thus providing support for a major role played by TET1 and TET2 proteins
in the enzymatic oxidation of 5-mC during differentiation [57]. It
was found that TET1 and TET2 double-knockout (DKO) embryonic
stem cells were depleted in 5-hmC as inferred from HPLC-MS/MS
measurements [121]. However, there were still elevated levels of
5-hmC in DKO mice which was rationalized in terms of an overexpression of TET3 enzyme as the result of TET1 and TET2 deficiency
[121]. Because TET1/2 double knock-out mice are viable, TET3
appears to be more important than TET1/2 in early development
although all three enzymes contribute to the hydroxylation of 5mC
and formation of 5-hmC [121]. Indirect evidence implicating TET2
in the conversion of 5-mC to 5-hmC in hematopoietic cells undergoing differentiation was gained from HPLC-MS/MS measurements
24
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
100
A: Natural product -tumor DNA
5-mC
5-hmC
5-fC
5-fU
5-hmU
75
Intensity (cps)
50
25
0
100
75
5
B: Isotopic standards
10
15
20
10
15
20
50
25
0
0
5
Retention time (min)
Fig. 2. HPLC-MS/MS analysis of the methyl oxidation products of 5-mC and thymine
in extracted DNA from a human brain tumour specimen. DNA was digested to its
component 2 -deoxyribonucleosides with P1 nuclease, snake venom phosphodiesterase, and alkaline phosphatase. HPLC analysis was carried out with a reversed
phase column (Hypersil GOLD, Thermo Scientific, 25 cm length, 2.1 mm internal
diameter, 5 um particle phase) using a gradient from 5 mM ammonium formate
(pH 5) to 8% acetonitrile in 24 min. The products were monitored by tandem
MS (API 3000, AB-Sciex) in positive mode using multiple reaction monitoring
(MRM) with the following parameters: 5-mC (m/z 483–242, CE = 10 eV); 5-hmC (m/z
258.0–142.0, CE= 10 eV); 5-fC (m/z 256.0–140.0; CE= 15 eV); 5-CaC (m/z 272–156,
CE= 15 eV); 5-hmU (m/z 259.1–125.0, CE= 15 eV); 5-fU (m/z 257.0–141.0, 15 eV)
[237]. Top panel: The amount of modifications in tumour DNA obtained from oncolumn injection of 0.4 ␮g of digested DNA was 3.54, 0.074, 0.023, <0.001 for 5-mC,
5-hmC, 5-fC and 5-caC, respectively, and 0.019 and 0.291 for 5-hmU and 5-fU,
respectively, in units of modifications per 100 C (or G). Bottom panel shows the isotopic internal standards: 5-mC (16 pmol, m/z + 3), 5-hmC (0.22 pmol, m/z + 3), 5-fC
(0.22 pmol, m/z + 3), 5-caC (0.22 pmol, m/z + 3), 5-hmU (0.56 pmol, m/z + 2 and 5-fU
(0.56 pmol).
of 5-hmdCyd. Thus, a significant decrease in the levels of 5-hmdCyd
was observed in granulocyte DNA of either TET2 mutant cells from
myeloproliferative neoplasma patients or TET2 depleted leukaemia
cell lines and cord blood CD34+ cells using RNAi with respect to
healthy control cells [89]. Evidence has been provided for a predominant implication of TET3 at the exclusion of the two other TET
proteins in the 5-mC enzymatic oxidation of paternal pronucleus
upon fertilization of oocytes as part of epigenetic reprogramming
of parental genome [68,122]. It was shown that modified histone
H3K9me2 through interaction with PGC7 protein prevents TET3mediated oxidation of 5-mC to occur in maternal genome [123].
The presence of 5-hmC has been detected in mitochondrial DNA
by qPCR analysis of immunoprecipated DNA [124]. The mechanism
of formation of 5-hmC remains to be established although it can
involve enzymes of the TET family and/or a hydroxymethylation
reaction of cytosine at C5 mediated by mtDNA methyltransferase
1 [5].
3.1.2. 5-Formylcytosine and 5-carboxylcytosine as other methyl
oxidation products of 5-mC
The formation of other methyl oxidation products of 5-mC
has been proposed as putative active demethylation intermediates [125,126] and also by analogy with the sequential oxidation
of thymine mediated by thymine 7-hydroxylase (T7H) in fungi
including Aspergillus nidulans, Neurospora crassa and Rhodotorula
glutinis. The latter enzymatic reaction leads to 5-carboxyluracil or
iso-orotate through transient formation of 5-formyluracil (5-fU) as
part of the pyrimidine salvage pathway [127–129]. However, a first
attempt to search for 5-fC and 5-caC in 5-hmC abundant areas of
the central nervous system of mice, including the central cortex,
brainstem, spinal cord and cerebellum, was unsuccessful [30]. This
may be explained by the use of a relatively insensitive HPLC-ESIsingle MS method that is not appropriate when the frequency of
oxidized nucleosides is lower than 1 modification per 105 normal nucleosides [46,62]. Subsequently the same research group
was able to unambiguously characterize the presence of 5-fC as
either the related free 2 -deoxyribonucleoside or upon derivatization with biotin hydroxylamine in the DNA of mouse ES cells. This
was achieved utilizing suitable and sensitive HPLC analytical assays
involving either ESI-MS/MS or MS3 detection techniques associated with quantitative isotope dilution [85]. The amount of biotin
labeled formyl-2 -deoxycytidine (5-fdCyd) was found to be about
3 residues per 104 2 -deoxyguanosines while the level decreases
during differentiation and becomes undetectable after 3 days. This
important finding received independent confirmation from two
other reports that appeared almost simultaneously [46,47].
In addition to 5-hmC, evidence was provided for the enzymatic formation of 5-fC and also 5-caC upon incubation of a
20-mer oligonucleotide containing 5-mC with any of the three
TET proteins. Characterization of the two novel 5-mC oxidation products was achieved using a suitable restriction enzyme
based assay [46]. This involved radioactive phosphorylation of
Taq1 restriction DNA fragments at the site of 5-mC followed by
enzymatic release of [32 P]-labeled nucleotides and subsequent
comparison with authentic samples that necessitated the design
of an improved 2D-TLC analysis. Taq␣ 1 cleaves double stranded
DNA at TNGA sites, where N is cytosine (C) or modified C, with
efficiencies that vary from 100% (C and 5-mC) to 60-75% (5hmC, 5-fC and 5-caC). The failure of detecting 5-fC and 5-caC as
TET oxidation products of 5-mC in previous studies involving
a similar enzymatic assay was due to the inability of MspI, the
restriction enzyme initially used, to cleave oligonucleotides at
the restriction sequence containing 5-fC and 5-caC sites. Further
support to the assignment of 5-fC and 5-caC was provided by
2D-TLC analysis of related oxime and amide derivatives formed
by reactions with O-ethylhydroxylamine hydrochloride and 1ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride,
respectively. Final proof of the structure for the two oxidized
nucleosides was provided from the characteristic fingerprints
given by MS analysis of the extracted TLC spots. It was shown
that TET2 exhibits the highest capacity to oxidize oligonucleotides
containing 5-mC. Evidence was also provided that both 5-hmC
and 5-fC were substrates for further oxidation by TET enzymes
albeit the initial rate of oxidation for 5-hmC (429 nM/min) was
much higher than that for either 5-fC or 5-caC (87 and 57 nM/min
respectively) [46]. Occurrence of iterative oxidation of 5-mC by
TET protein was also strongly suggested from studies that have
involved transfection of human embryonic kidney (HEK) 293 cells
that overexpress TET2 catalytic domain. The presence of 5-fC and
5-caC was established in genomic DNA in mouse ES cells by HPLCESI-MS/MS analysis in the multiple reaction monitoring (MRM)
mode after enzymatic digestion to 2 -deoxyribonucleosides.
It was found that the levels of 5-hmC, 5-fC and 5-caC were
1.3 × 103 , 20 and 3 per 106 cytosines respectively [46]. Lower
amounts of 5-fC were also detected in other mouse organs as
follows in decreasing order of quantitative importance: brain
cortex > spleen > pancreas > liver > heart > thymus > lung > kidney.
It may be added that the levels of 5-caC are much lower and only
barely detected in lung and brain cortex [46].
A similar restriction enzyme assay involving EcoN1 protein
that specifically digests DNA sequences at 5-caC sites was used in
association with 1D TLC analysis for monitoring the formation of
the latter 5-mC oxidation product in a synthetic duplex oligonucleotide containing 5-mC after incubation with nuclear extract
from human embryonic kidney (HEK) 293 T cells [47]. Support
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
for the formation of 5-caC was given by comparison of the TLC
behavior of the [32 P]-labeled unknown nucleotide X with that of
authentic 5-carboxyl-2 -deoxycytidine 5 -monophosphate [47].
Unambiguous characterization of 5-caC was gained from the HPLCESI-MS/MS analysis in the MRM mode of 2 -deoxyribonucleoside
X’ that was released by enzymatic digestion of the above oxidized
DNA duplex. The latter accurate analytical approach was also
used for monitoring the formation of 5-caC in the DNA of Tet2and Tet1-transfected HEK 293 cells [47]. Indirect support for
the implication of TET-mediated iterative oxidation of 5-mC in
an active demethylation pathway was provided by a significant
increase in the levels of 5-caC in mouse ES cells depleted in the
activity of thymidine DNA glycosylase (TDG) that is involved in
base excision repair of the latter oxidized base [47].
Genome-wide mapping of 5-fC in ES cells at the nucleotide
level was performed using a suitable deep sequencing method of
enriched DNA fragments [80] as an adaptation of an available protocol designed for pulling down biotinylated DNA [103]. As a striking
result it was found that the distribution of 5-fC in genes of mouse
ES cells was partly under the control of TDG implicated in the repair
of 5-fC [130]. Down-regulation of TDG was found to lead to 5fC enrichment in CpG islands concomitantly with an increase in
methylation upon differentiation [80].
An immunological detection approach based on the preparation of specific rabbit polyclonal antibodies anti 5-fC and 5-caC was
designed to search for the formation of the latter 5-mC oxidation
products during mouse preimplantation development [79]. Thus,
immunostaining measurements showed the existence of a correlation between an increase in the level of 5-fC and 5-caC in male
pronucleus and a decrease of 5-mC in zygotes [79]. In a more recent
study, again using an immunochemical detection approach, 5-caC
was observed in the nuclei of follicular cells of axolotl ovary with
a higher distribution in gene-rich euchromatic regions. However,
the detection of 5-caC in somatic cells of adult amphibians raises a
question as to the specificity of the antibody as already discussed
in section 2 when the frequency of the target oxidized base to be
detected is very low, i.e., of the order of a few molecules per 106
cytosines with the possibility of cross-reactivity!
3.1.3. Mechanisms of TET1-3 enzymatic oxidation reactions
The recent rediscovery of 5-hmC as an epigenetic mark has led,
through bioinformatical searching, to the identification of TET1
[27], and subsequently of TET2 and TET3 as the key enzymes
involved in oxidation of the methyl group of 5-mC in mammalian
cells (for recent reviews, see [16,36,51,131]). Mammalian TET13 proteins, first called leukaemia-associated protein with CXXC
domain (LCX), were initially identified as a fusion partner of the MLL
gene in an acute myeloid patient [4,131]. The proteins exhibited
homology with trypanosomal thymine 7-hydroxylase J-binding
proteins JBP1 and JBP2 that catalyze the conversion of thymine
into 5-hydroxymethyluracil and subsequent alkylation at C5 by a
putative ˇ-glucosyl transferase [132,133]. All three TET proteins
were characterized by a CD domain consisting of cysteine-rich
and double-stranded ß-helix (DSBH) regions which include the
␣–ketoglutarate (␣-KG) and Fe(II)-dependent dioxygenase activity
and a spacer of varying length according to the enzyme. In addition, TET1 and TET3 but not TET2 retain an ancestral CXXC zinc
finger domain that allows for binding to unmethylated CpG sites
[134–136]. The situation is different for TET2 where the ancestral CXXC protein also called CXX4 is encoded by IDAX, a distinct
gene. It was recently reported that IDAX expression in human
U937 monocyte cells led to down-regulation of TET2 [137]. Abundant literature is available on the non-heme Fe(II)/␣-KG-dependent
dioxygenase-mediated oxidation of methylated DNA, RNA and histones [125,138,139] that requires dioxygen (O2 ) together with ATP
and ascorbate as co-factors. The first step of the reaction involves
25
activation of O2 within a highly reactive Fe(IV)-oxo intermediate
[140,141] that involves the displacement of two water molecules
by the enzyme-substrate complex and subsequent generation of
an octahedral coordinated bidentate [138–140]. The ferryl species
thus generated is able to hydroxylate the methyl group of 5-mC via
hydrogen abstraction according to a “radical-rebound” mechanism
or less likely in the present case through direct insertion [138]. The
second oxygen atom from O2 is incorporated into ␣-KG giving rise
to succinate and CO2 [141]. However, a more definitive hydroxylation mechanism of 5-hmC, which is likely to apply to iterative
TET-mediated oxidation of 5-hmC and 5-fC as its hydrated form
[85], is awaiting further experiments.
Certain studies have shown that vitamin C (ascorbic acid, ascorbate) enhances DNA demethylation [142]. A possible mechanism
has now emerged implicating vitamin C as a co-factor of TET
enzymes. Supplementation with ascorbate 2-phosphate led to a
concomitant gain of 5-hmC and loss of 5-mC in 12 to 72 h after treatment of cultivated ES cells [143]. The effect was dose-dependent,
significantly diminished within TET knock-out cells, and not reproduced by supplementation with other antioxidants. Interestingly,
vitamin C-induced demethylation was localized in specific regions
of the genome, which included transcription start sites and highintensity CpG protomers. Additional studies have confirmed the
ability of vitamin C to enhance demethylation by interaction with
TET proteins in ES and other cell lines [144,145]. In mouse ES cells
treated with 100 uM of vitamin C over 24 h, there was a 4-fold
increase in 5-hmC accompanied with 11-fold and 20-fold increases
of 5-fC and 5-CaC. The effects of vitamin C supplementation were
extended to mice that are deficient in vitamin C synthesis, showing,
in comparison to normal mice, increases of 5-hmC and decreases
of 5-mC in various organs [145].
3.2. Radical oxidation reactions
As with all DNA components, 5-mC could be subjected in cells to
oxidation reactions mediated by endogenous and exogenous reactive oxygen species and processes including • OH and one-electron
oxidants [50,146]. • OH is generated through Fenton type reactions
involving the initial formation of unreactive superoxide anion radical (O2 •− ) as the result of one-electron reduction during oxidative
metabolism in mitochondria. Generation of carbonate anion radical
(CO3 •− ), a strong one-electron oxidant, is formed at inflammation
sites from a complex sequence of reactions [147,148]. This involves
initial generation of O2 •− and nitric oxide (• NO) that recombine to
produce peroxynitrite before reacting with CO2 or bicarbonate to
yield CO3 •− through the intermediary of nitrosoperoxycarbonate.
Ionizing radiation is able to generate • OH and to a lesser extent
abstract one electron from both the bases and 2-deoxyribose of
nuclear DNA as the result of indirect and direct effects respectively
[50]. Type I photosensitizers such as thiopurine immunosuppressors have the capability to oxidize nucleobases by one-electron
abstraction with emphasis on guanine [83,149].
3.2.1. Hydroxyl radical-mediated oxidation of 5-mC
Earlier pulse radiolysis measurements have shown that 5mC is more susceptible to reaction with • OH than cytosine and
to a lesser extent with thymine as inferred from the second
order rate constants that are elevated being close to diffusioncontrolled [150]. As also observed for thymine [151], • OH is able
to undergo two competitive reactions with 5-mC including hydrogen atom abstraction from the methyl group and addition to the
5,6-pyrimidine double bonds as confirmed by an exhaustive theoretical study [152]. Among several other degradation products,
5-hmC was isolated and characterized by gas-chromatography
associated with mass spectrometry (GC-MS) analysis upon exposure of 5-mC to either bromotrichloromethane in the presence
26
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
NH2
N3
5
2 6
O
NH2
CH3
OH
N
N
O
N
CH3
OH
NH2
+O2
N
H
O
N
NH2
CH3
OH
OO
H
N
O
C5-OH adduct
5-methylcytosine
(5-mC)
OH
N
N
O
NH2
+O2
OH
N
O
N
CH3
OO
OH
O
-e
N
O
N
NH2
H
products observed
by oxidation of
the nucleobase,
nucleoside
and DNA
OH
Hyd-5-mC
NH2
-H+
N
O
NH2
CH2
N
+O2
O
NH2
CH2OO
N
-e
N
O
CH2OH ox
N
O
N
5-hmC
NH2 O
NH2
NH2
CH2OO (H)
N
N
5-hpC
NH2
O
CH2OOH
N
radical cation
N
CH3
OH
H
OH
N
Imid-5-mC
CH3
N
N
O
NH3
N
C6-OH adduct
CH3
H2N
Gly-5-mC
NH2
CH3
N
O
CH3
OH
OH
H
CHO
N
O
N
5-fCyt
ox
minor O
OH
N
N
5-caC
Fig. 3. Free radical mediated oxidation of 5-methylcytosine. Two main free radical oxidation reactions lead to the efficient oxidation of 5-methylcytosine (5-mC): hydroxyl
radical (• OH) attack at C5 or C6 leading to the corresponding intermediate • OH adduct radicals and the one-electron oxidation of 5-mC leading to deprotonation of the
resulting radical cation and formation of intermediate methyl radical of 5-mC. These pathways represent 95% of 5-mC decomposition by • OH and 70% of decomposition by
one-electron oxidation. In addition, each pathway can merge due to H-atom abstraction by • OH to give intermediate methyl radicals of 5-mC and hydration of 5-mC radical
cations to give • OH adduct radicals of 5-mC. Intermediate • OH adduct radicals in the presence of oxygen give rise to hydroxyperoxyl radicals, which in turn, decompose to 5,6dihydroxy-5,6-dihydrocytosine (Gly-5-mC),: 4-amino-1-5-dihydro-5-methyl-2-H-imidazol-2-one (Hyd-5-mC), 1-carbamoyl-4,5-dihydroxy-5-methyl-2-oxo-imidazolidine
(Imid-5-mC). In contrast, intermediate methyl radicals of 5-mC transform via the hydroxyperoxyl radical under oxygenated conditions to 5-hydroperoxymethylcytosine
(5-hpC) and stable decomposition products of hydroxyperoxyl radicals or 5-hpC, which include mostly 5-hydroxymethylcytosine (5-hmC) and 5-formylcytosine (5-fC) with
very low amounts of secondary 5-carboxylcytosine (5-caC). For details see [22].
of benzoyl peroxide, a • OH generator system [153] or a Fenton reagent [154]. In a subsequent study, 5-fC was found to be
formed upon treatment of in DNA duplexes containing 5-mCpG
with Fe(II)nitrilotriacetate/H2 O2 /ascorbate [155]. Under similar
oxidizing conditions that involve the formation of • OH via metalmediated reduction of H2 O2 , 5-hmC in addition to 5-fC was
also found to be produced [156,157]. In a recent detailed study,
5-fdCyd, 5-hmdCyd together with the 5-(hydroperoxymethyl)-2 deoxycytidine (5-hpdCyd) precursor (Fig. 3) were isolated and
unambiguously characterized as • OH-mediated oxidation products
according to a metal-independent mechanism involving the reaction of several halogenated quinones with H2 O2 [158].
3.2.2. One-electron oxidation of 5-mC
The main one-electron oxidation reactions of the methyl
group of 5-mdCyd upon exposure to photoexcited 2-methyl-1,4naphthoquinone (MQ) have been elucidated (Fig. 3) [159]. Initial
generation of the pyrimidine radical cation via charge transfer to triplet excited MQ undergoes deprotonation giving rise to
the 5-(2 -deoxycytidyl)methyl radical that may also be produced
by • OH-mediated hydrogen abstraction from the methyl group
[22,158] or photolysis of 5-phenylthiomethyl-2 -deoxycytidine
[160,161]. Subsequently, fast addition of O2 to the latter methyl
centred radical leads to the related hydroperoxyl radical that may
be reduced, likely by O2 •− before being converted into 5-hpmdCyd
after protonation [22,159]. A competitive reaction of the peroxyl
radical, at least for the isolated nucleoside, is dismutation with
the generation through the concerted Russell mechanism of equal
amounts of 5-fdCyd and 5-hmdCyd [158] together with the release
of singlet oxygen (1 O2 ) as observed for peroxyl radicals derived
from the methyl group of thymine [162]. It was shown that the
formation of 5-fdCyd by MQ and UVA light photosensitization was
pH dependent within the 5–8 pH range with a maximum yield in
slightly acidic solutions [163]. Evidence was also provided for the
lack of a significant effect of cytosine methylation on the extent
of electron transport in DNA as measured by piperidine-induced
strand cleavage, likely due to the formation of 5-fC, upon injection
of a positive hole in duplex DNA upon UVA excitation of a tethered
anthraquinone group at the 3 -end [164].
3.2.3. Other oxidation reactions of 5-mC
UVC irradiation of 5-mC in aerated aqueous solutions led to
the formation of cytosine together with 5-hmC, 5-fC and 5-caC
[165]. The formation of 5-hmC and 5-fC was rationalized in terms
of photo-ionization of the pyrimidine base with concomitant generation of the related radical cation and subsequent deprotonation
and reactions of the resulting 5-(cytosyl)methyl radical with O2
giving rise to 5-hmC and 5-fC as discussed above. Subsequent
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
photooxidation of 5-fC was proposed to lead to the formation of
5-caC that is expected to be followed by photo-induced decarboxylation with the generation of cytosine. However, the efficiency of
these photoreactions is very low to have any significant biological
relevance.
Comparison of the radical and TET-mediated oxidation reactions of 5-mC in terms of qualitative and quantitative distribution of
degradation products shows major differences despite some similarities. As a first remark, the radical oxidation of 5-mC is far less
specific than that provided by TET proteins since in addition to
methyl oxidation products, dihydroxylation of the 5,6-double bond
and several rearrangement products of the pyrimidine ring have
been isolated and identified (Fig. 3) [22,166–168]. Both 5-hmC and
5-fC can be generated by enzymatic and radical reactions involving
either • OH or one-electron oxidants. It may be noted that 5-caC
appears to be a specific oxidation product of TET activity since
secondary radical oxidation of 5-fC is an unlikely process in cells.
Another major difference concerns the frequency of 5-hmC and to
a lesser extent of 5-fC in that enzymatic oxidation by TET proteins
gives higher levels of 5-mC by at least two orders of magnitude for
5-hmC in ES and neuronal cells compared to radical processes even
under conditions of oxidative stress.
4. Synthesis of modified nucleosides and oligonucleotides
4.1. Labeled modified nucleosides
The preparation of isotopically labeled standards is critical for
the absolute quantification of 5-mC, 5-hmC, 5-fC and 5-caC in cellular DNA using sensitive assays based on MS. Several methods
have been developed to incorporate stable isotopes into modified
nucleosides.
The synthesis of labeled [CD3 ]-5-mdCyd has recently been
described in three steps from 5-iodo-2 -deoxycytidine by
palladium-catalyzed methylation with [CD3 ]-MgI of the protected
nucleoside (overall yield about 70%) [29]. Labeled [15 N3 ]-5-mdCyd
can also be synthesized from commercially available labeled
[15 N2 ]-thymidine in four steps including protection and deprotection of the sugar moiety [163,169,170]. Introduction of the
exocyclic amino group at C4 is straightforward by substitution
of triazole with labeled ammonium at C4 of 3 ,5 -diacetylated
thymidine. The overall yield of labeled 5-mdCyd on a small scale
(0.1-1 g) is about 50% (unpublished). Labeled 5-mC can also be
prepared at CpG sites in oligonucleotides and other DNA substrates
by treatment with methyltransferase in the presence of deuterated
SAM [171].
[18 O]-labeled hmdCyd within the hydroxymethyl group was
obtained by a multistep synthesis starting from 5-bromo-2 deoxyuridine giving an overall yield of about 20%. The key steps
involve the incorporation of H2 18 O by SN2 substitution of the
bromide group followed by amidation of 2 -deoxyuridine (dUrd)
to 2 -deoxycytidine (dCyd) using tert-butyldimethylsilyl (TBDMS)protected nucleoside and ammonium [29]. An alternative method
to prepare small quantities of labeled 5-hmdCyd involves MQphotosensitized oxidation (see below) although it is necessary to
purify this product from the photolysis mixture. 5-hmC can also be
prepared by the direct reaction of paraformaldehyde under basic
conditions, similar to the well-known reaction of uracil derivatives
[172–174]. However, a major drawback of the reaction with cytosine derivatives is that they undergo substantial deamination to
uracil derivatives.
One of the simplest methods to prepare isotopically labeled
5-fdCyd is by one-electron MQ-photosensitized oxidation labeled
5-mdCyd (above). The reaction gives methyl oxidation products of
5-mdCyd in an overall yield of about 70% and the reaction favors the
27
formation of 5-fdCyd at extended time of photolysis due to in situ
oxidation of 5-hmdCyd to 5-fdCyd. The optimal conditions for
the decomposition of 5-mdCyd by MQ-photosensitized oxidation
are pH 5 with 10% acetonitrile [163]. Starting with [15 N2 ]-urea, a
method for the preparation of [15 N2 ]-5-cadCyd was reported using
Pd-catalyzed CO coupling to 5-iodo-2 -deoxycytidine (5-IdCyd) followed by quenching with methanol to give the methyl ester [175].
A potential alternative method for the preparation of labeled 5formylcytidine involves the oxidation of 5 -protected (TBDPSO)
5-methylcytidine with sodium peroxosulfate (Na2 S2 O8 ) in buffer
solution, which gives the corresponding hydroxymethyl and formyl
derivatives in 27% and 20%, respectively [176]. The hydroxyl derivative can subsequently be converted into the formyl derivative by
treatment with ceric(IV) ammonium nitrate (CAN).
4.2. Oligonucleotides
Oligonucleotides (ODNs) containing 5-hmC, 5-fC and 5-caC are
prepared by insertion of the appropriate phosphoramidite building
block during standard ODN synthesis. A key step in the synthesis
of the building block involves Pd-catalyzed carbonylation under
CO atmosphere of the 5-iodo-substituted pyrimidine nucleoside.
This reaction gives high yields (>80%) of 5-fdUrd from dUrd as
well as 5-fdCyd from dCyd derivatives bearing an iodo substituent
at C5, with or without protection of the exocyclic amino group,
and with appropriate protection of the 3 and 5 hydroxyls of 2deoxyribose [177–180]. The complete synthesis of the building
block of 5-fC from 5-IdCyd involves 6 steps with an overall yield that
varies from 20-50% depending on the choice of protective groups
for the exocyclic amino and 2-deoxyribose 3,5-hydroxyls during
phosphoramidite synthesis [179,181]. It is not necessary to protect
the formyl group of the phosphoramidite building block of 5-fdCyd
for subsequent incorporation during ODN synthesis. To obtain the
corresponding 5-hydroxymethyl derivative, the formyl containing
building block is reduced to 5-hydroxymethyl by reaction with
NaBH4 and CeCl3 in methanol. The 5-hydroxymethyl group is more
reactive than the 5 hydroxyl of the 2-deoxyribose moiety and thus
it must be inactivated during phosphoramidite and ODN syntheses.
A number of protecting groups have been proposed: a cyanoethyl
group that requires strong basic conditions for removal is commercially available, TBDMS group that can be removed by NH4 F
and is compatible with mild ODN deprotection [178], and a cyclic
carbamate group that protects both exocyclic amino and hydroxymethyl and can be removed under mild conditions (0.4 M NaOH in
MeOH/H2 O (4:1) for 12 h at room temperature) [177]. On the other
hand, ODNs containing 5-fC can be reduced after ODN synthesis by
reaction with NaBH4 and CeCl3 in methanol while attached to resin,
thereby avoiding the necessary steps to protect the hydroxymethyl
group during synthesis of the building block [179]. In addition, 5hmC can be oxidized to 5-fC in single stranded ODNs by KRuO4 in
alkaline solution whereas two rounds of oxidation was necessary
for the reaction in double stranded DNA [110]. This procedure can
be applied to distinquish between 5-hmC and 5-fC because latter
modification engenders strand cleavage upon treatment with hot
piperidine [182]. The incorporation of 5-caC into ODNs can also be
carried out by a similar strategy to that described above for 5-hmC
and 5-fC. Recently, a modification of CO carbonylation was reported
that converts 5-IdCyd to 5-methoxycarbonyl-2 -deoxycytidine in
high yield (93%) [179]. The later nucleoside is then converted to 5cadCyd within ODNs during mild basic deprotection (0.1 M K2 CO3
in MeOH/H2 O overnight at 40 ◦ C). Thus, the phosphoramidite building block for the incorporation of 5-caC into ODNs was synthesized
in 4 steps with an overall yield of 57% [179].
Oligonucleotides containing 5-fC can also be prepared by
the incorporation of 5-(1,2-dihydroxyethyl)-2 -deoxycytidine
into ODNs as a precursor for post synthesis conversion to 5-fC
28
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
[183,184]. The phosphoramidite of 5-(1,2-dihydroxyethyl)2 -deoxycytidine was obtained by Pd(II)-catalyzed vinylation
of 5-iodo-2 -deoxyuridine (67%) and oxidation with OsO4/Nmethylmorpholine (77%). Standard ODN deprotection converts the
1,2-dihydroxyethyl moiety to a diol that is subsequently converted
to 5-fC by reaction with sodium periodate in a quantitative reaction.
Another general strategy is used frequently to incorporate
5-hmC, 5-fC and 5-caC into ODNs and long DNA fragments
using modified 2 -deoxyribonucleoside triphosphates and PCR.
The modified triphosphates of 5-hmdCyd, 5-fdCyd and 5-cadCyd
are commercially available. Recently, Carell and co-workers [185]
prepared the triphosphates of 5-fdCyd and 5-cadCyd and incorporated them into long DNA fragments by PCR. The synthesis
involved a novel method to prepare nucleoside triphosphates in
high yield (84-90%) by the addition of pyrophosphate to nucleoside
monophosphates using sulfonylimidazolium salts as the activating
agent [186]. A procedure was developed to circumvent a problem
associated with the incorporation of 5-hmC involving hydroxymethyl protection and then removal after polymerization.
4.3. Chemical stability of 5-mC methyl oxidation products
The C4-amino group of cytosine derivatives and the N-glycosidic
bond of pyrimidine 2 -deoxyribonucleoosides are subject to
hydrolysis under physiological conditions [187]. In contrast, spontaneous deamination of 5-mdCyd in double-stranded DNA at
neutral pH is very low experimentally even though the rate of
deamination of 5-mdCyd is two-fold faster than that for dCyd [188].
The low efficiency of deamination estimated to have a rate constant of 5.8 × 10−13 s−1 is consistent with theoretical studies [189].
Similar conclusions apply to 5-hmC, 5-fC and 5-caC components as
inferred from Density Functional Theory (DFT) calculations (Grand
et al., unpublished data), implying that hydrolytic deamination of
5-mC oxidation products is at best a minor process under physiological conditions. Relevant insights on stability of the N-glycosidic
bond of oxidation products of 5-hmdCyd were gained from a comprehensive DFT study performed at the B3LYP/6-31 + G(d) level
of theory [190]. It was found that 5-fdCyd and 5-cadCyd exhibit
lower energy barriers and higher exothermicities than 5-mdCyd
and 5-hmdCyd that is indicative of a slightly higher lability of
the N-glycosidic bond of the latter oxidized nucleosides [190]. It
is worth noting that the kinetic and thermodynamic parameters
thus obtained are in agreement with the higher susceptibility for
5-fmC and 5-caC to be preferential substrates among other 5substituted cytosine derivatives for cleavage by human thymine
DNA glycosylase [130]. Spontaneous deformylation of 5-fdCyd and
decarboxylation of 5-cadCyd that would lead to cytosine are also
unlikely processes under physiological conditions [180]. As further
support for the high stability for 5-cadCyd it was found that heating of aqueous solutions at reflux of either the free nucleoside or
5-caC-containing DNA fragments for several hours does not lead
to any detectable decarboxylation [175]. Some loss of the carboxyl
group was however observed when thiols that can react at C6 of
the pyrimidine ring by nucleophilic addition were present in the
solution [175]. The high stability of the three TET-mediated oxidation products of 5-mC clearly emphasize the crucial role played
by deaminases and base excision repair enzymes for achieving
efficient demethylation pathways involving 5-mC as discussed in
Section 5.
5. Biological features of 5-mC oxidation products
The complete pathway of demethylation of 5-mC remains
unclear today. Demethylation likely involves several steps: 1) the
oxidation of 5-mC to 5-hmC; 2) the oxidation of 5-hmC to 5-fC and
5-caC or the deamination of 5-hmC to 5-hydroxymethyluracil (5hmU); and 3) the removal of modified bases from step 2 by base
excision repair (BER) (Fig. 1). The involvement of related enzymes
and their function will be discussed below. The majority of evidence for cytosine demethylation comes from studies in which the
presence of methylation intermediates (5-hmC, 5-fC and 5-caC) is
associated with specific protein expression in cells and animals.
5.1. Demethylation of 5-mC and DNA repair
5.1.1. Passive demethylation of 5-mC oxidation products by cell
division
Evidence for passive dilution of 5-hmC in zygotes was gained
from immunostaining experiments involving mitotic chromosomes [191]. It was found that only one of the two sister
chromatids issued from sperm chromosomes was enriched with
5-hmC. This was rationalized in terms of replication-dependent
dilution of 5-hmC in the male pronucleus of zygotes during
preimplantation development. It was subsequently demonstrated
again by immunostaining that both 5-fC and 5-caC instead of
being quickly removed by TDG-mediated repair excision showed
replication-dependent dilution in the paternal DNA of zygotes during preimplantation development [79]. Lastly, it should be noted
that hemi-hydroxymethylated DNA at cytosine sites is not recognized by DNMT1, and thus, this mark is continuously diluted by cell
division [192] as also suggested by several authors.
5.1.2. Deamination of 5-hmC to 5-hmU by AID/APOBEC
The deamination of cytosine to uracil is a well-known process
in B lymphocytes that takes place by activation-induced deaminase (AID) to induce mutations in immunoglobulins loci and
promote antibody diversity [193]. This enzyme is a member of a
large family of cytosine deaminases known as apolipoprotein B
mRNA-editing enzyme complex (APOBEC). There is circumstantial evidence implicating AID/APOBEC in the overall demethylation
of 5-mC from DNA. AID knock-out leads to global hypermethylation in the genome of mammalian primordial germ cells compared
to wild type [194]. Similarly, HEK 293 cells without AID showed
increased levels of 5-hmC while cells overexpressing the protein
showed lower 5-hmC. Furthermore, the low levels of 5-hmC were
accompanied with detectable levels of 5-hmU as measured by
immunoblotting [121]. Knock-down of AID with adeno-associated
virus in vivo also appeared to induce the demethylation of certain
genes. On the other hand, the kinetics for the deamination of 5-mC
by AID/APOBEC was slow relative to non-methylated derivatives,
i.e., cytosine to uracil (10-fold less) and these enzymes preferentially act on single stranded DNA [186]. It may be added that no
detectable deamination of 5-hmC, 5-fC or 5-caC when inserted into
a 30-mer oligonucleotide was observed. Furthermore, the formation of 5-hmU, the deamination product of 5-hmC, was not detected
in embryonic stem cells as well as in cells that overexpress both
TET2 and AID using a sensitive HPLC-MS/MS method [195].
5.1.3. Removal of 5-hmU by TDG, MBD4 and SMUG1
Three human DNA glycosylases have been proposed in the
removal of 5-hmU from DNA: TDG (mismatch specific thymine DNA
glycosylase), MBD4 (methyl-binding domain protein 4), and singlestrand-specific monofunctional uracil DNA glycosylase 1 (SMUG1).
Using purified enzymes, both TDG and MBD4 were shown to
excise 5-hmU more efficiently than T from duplex oligonucleotide
with either 5-hmU or T opposite G [196,197]. This suggests that
the enzymes may have other important functions than protecting against C to T transitions at CpG dinucleotides. When purified
TDG is directly compared to purified MBD4, the maximum rate
of excision of 5-hmU opposite G was 6-fold faster by purified
TDG. Lastly, SMUG1 also excises 5-hmU with good efficiency under
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
similar conditions to those described above. An added feature of
SMUG1 is its ability to excise 5-hmU from both single and double
stranded substrates. Thus, all three enzymes may be involved in
the removal of 5-hmU from genomic DNA. Interestingly, AID and
TDG co-immunoprecipitate suggesting that deamination by AID
and removal of 5-hmU may take place in concert [198].
5.1.4. Removal of 5-fC and 5-caC by TDG
The main activity responsible for removal of 5-fC and 5-caC from
genomic DNA is likely TDG. Using well-defined duplex oligonucleotides containing 5-hmC, 5-fC and 5-caC, Maiti and Drohat [130]
reported rates of excision by human TDG (kmax; E > >S) for G-5fC (2.6) for G-5-caC (0.5) compared to that for G-T (1.8) in units
of min−1 at 37 ◦ C. Moréra et al. [197] reported similar kinetics:
kmax of 8 for G-5-hmU compared to that for G-T of 1.25 min−1 .
Thus, hTDG excises the oxidation products of 5-hmC, 5-fC and to a
lesser extent 5-caC, with good efficiency. In contrast, hTDG does
not excise 5-hmC. The ability of TDG to directly hydrolyze the
N-glycosidic bond directly depends on the electron withdrawing
nature of the C5-substituent. It was found that the asparagine-to
aspartate (N157D) mutant of mammalian TDG is able to selectively
excise 5-caC while 5-caC is poorly removed in GC pairs [199]. All
other enzymes tested in vitro displayed no or very poor activity
toward the removal of 5-hmC, 5-fC and 5-caC opposite G in duplex
DNA, including MBD4, human uracil DNA glycosylase 2 (hUNG2),
human nth endonuclease III-like 1 (hNTH1), human oxoguanine
DNA glycosylase 1 (hOGG1), human Nei-like DNA glycosylase 1
and 2 (NEIL1/2) [130]. In support for TDG as providing the main
activity in mammalian cells, knock-down of TDG leads to accumulation of 5-caC in the genome while overexpression reduces 5-caC
in the genome of ES cells [47]. Genome-wide mapping of 5-fC in ES
cells at the nucleotide level was performed using a suitable deep
sequencing method and enriched DNA fragments [71]. As a striking
result it was found that the distribution of 5-fC in genes of mouse
ES cells was partly under the control of TDG that is implicated in
the repair of the latter oxidized 5-mC base [73]. Down-regulation of
TDG was found to lead to 5-fC enrichment in CpGIs concomitantly
with methylation increase upon differentiation [71]. TDG depletion
in mouse ESCs led to a marked increase in the levels of both 5-fC
and 5-caC at specific classes of repetitive sequences including major
satellite repeats in nuclear DNA using antibody based genome-wide
analysis [200]. These observations provide support for the occurrence of dynamic methylation/demethylation processes within the
genome. Evidence was provided for TETs and TDG to be direct targets of microRNA-29 that appears to regulate two main enzymatic
activities of DNA demethylation [201].
5.1.5. Reversal of 5-mC and 5-hmC by DNMTs
DNA methyltransferases that normally methylate cytosine are
also known to undergo the reverse process involving removal of
formaldehyde as well as deamination [5,202]. In addition, purified
methyl binding domain protein 2 (MBD2) can directly demethylate DNA containing 5-mC by a reaction that releases formaldehyde
[203]. Recently, Chen and co-workers [204,205] clearly demonstrated the ability of DNMT enzymes to remove 5-mC and 5-hmC
from DNA substrates. To determine the removal of 5-mC and 5-hmC
and the formation of C, the authors reacted purified enzyme with a
DNA substrate containing either 5-mC or 5-hmC, and then cleaved
DNA using MspI restriction enzyme (which cleaves sequences containing CNGG, where N is 5-mC or 5-hmC), labeled the terminal
nucleotide with 32 P, reduced DNA to its component nucleotides,
and quantified the amount of N or C, 5-mC and 5-mC by 1D-TLC and
autoradiography. Using this assay, purified DNMT3A/B enzymes
converted 5-hmC to C in appropriate DNA substrates, indicating
that this enzyme can directly remove 5-hmC from DNA. In a followup study, hDNMT1 and DNMT3B were shown to convert 50% of
29
5-mC to C at the Msp1 cleavage sites in DNA substrates. Low levels of 5-hmC were observed similar to control samples suggesting
that the conversion does not pass through an intermediate compound. Interestingly, the reaction was reversible in that DNMT3B
tended to methylate DNA in the presence of SAM but reversed its
role to direct demethylation in the presence of oxidants such as
H2 O2 . This suggests that the redox state in cells may determine the
role of DNMT enzymes that is whether to methylate or demethylate
C. On the basis of mutant forms, the reactions appear to involve the
same catalytic domains for both methylation and demethylation.
5.2. Mutagenesis
The potential mutagenicity of the three TET-mediated 5-mC
oxidation products has been assessed by a novel primer extension assay based on enzymatic incorporation of normal nucleoside
triphosphates past the oxidized 5-mC nucleoside [12]. There was
also little apparent misincorporation of non-cognate bases by several high and low fidelity polymerases including Klenow exo-, Pol
n and Pol k (< 1%) irrespective of the 5-mC oxidation product [180].
This is strongly indicative of low, if any, mutagenicity of 5-hmC,
5-fC or 5-caC. In another study, however, 5-fC was less efficient
than 5-mC for the insertion of dGMP during primer extension. The
misincorporation of dAMP occurred more frequently opposite 5-fC
than 5-mC, and dAMP and TMP were both incorporated more frequently, possibly explaining the formation of CG to TA transitions
and CG to AT transversions [184]. However, these results are in
sharp contrast with the high mutagenicity of 5-fC reported from a
double-stranded shuttle vector investigation involving transfection
of COS-7 cells with vectors containing the oxidized base at a specific
site [206]. A broad mutagenicity spectrum including 5-fC → G, 5-fC
→ A and 5-fC → T mutations was observed. However, the apparent
inconsistencies between the conclusions provided by in vitro and
in vivo experiments remain to be resolved. This could be achieved
using high-throughput determination of the mutagenic features of
base modifications by next-generation sequencing as an improved
extension of the shuttle vector approach [207].
5.3. Other biological features
The role of 5-mC and its three enzymatic oxidation products
on RNA transcription was recently assessed using mammalian and
yeast polymerase II (Pol II) [208]. The rates of nucleotide incorporation and the substrate specificity of Pol II elongation were lower
when either 5-fC or 5-caC was present in the template in place of
5-mC and 5-hmC [208].
In a recent detailed investigation, attempts were made to identify readers of the three TET-mediated oxidation products of 5-mC
in mouse ES cells, neuronal progenitor (NP) cells and adult mouse
brain [91]. This was achieved using powerful quantitative mass
spectrometry-based proteomics. As relevant observations, it was
found that Klf4 specifically binds to 5-mC in the DNA of ES cells,
whereas Uhrf2 has dedicated interactions with 5-hmC in NP cells.
Interestingly, larger amounts of proteins, including several DNA
repair glycosylases and transcription regulators were recruited by
5-fC and 5-caC than by 5-mC and 5-hmC in ES cells. This was rationalized in terms of further support for the implication of iterative
oxidation of 5-hmC and excision repair of 5-fC and 5-caC in active
demethylation processes [91].
5.4. Depleted levels of 5-hmC in cancer tissues
The epigenetic landscape of cancer cells is dramatically altered.
For example, thousands of genes are differentially methylated
in tumor compared to healthy tissue in glioblastoma multiform [209]. In general, tumors exhibit a global loss of cytosine
30
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
methylation (regional hypomethylation), which can induce chromosomal instability by the rearrangement of repetitive elements,
loss of parental allele-specific expression, and oncogene activation [210,211]. Global hypomethylation is accompanied with
regional increases in cytosine methylation (hypermethylation). In
particular, the expression of several DNA repair proteins can be
silenced by promotor hypermethylation, including MGMT (O6 methylguanine-DNA-methyltransferase), hMLH1 (human mut L
homolog 1) and several others [212]. The methylation of CpG dinucleotides in the promoter region of MGMT reduces the ability of
cancer cells to repair alkylation damage in DNA. The extent of methylation may be considered as a biomarker toward the response of
glioma patients to alkylating agents during treatment [213].
The level of 5-hmC in the DNA of cancer cells is lower than that
observed in normal cells as shown in numerous cases: myeloid
malignancies, melanoma, and carcinomas of prostate, breast, colon,
liver, lung and pancreas, using either immunohistochemistry or
more quantitative mass spectrometry assays. [75,214–218]. One
of the most dramatic changes of 5-hmC is observed in brain cancer.
For example, the level of 5-mC in normal human brain specimens
varies from 0.70 to 1.17%C depending on the region, and in general, neurons (1.45%) contain significantly more than astrocytes
(0.23%C). In comparison, the level of 5-mC in brain tumors ranges
from 0.03 to 0.24 depending on the subset and stage of disease, as
determined by mass spectrometry [219]. The loss of 5-hmC in cancer cells can be attributed to several causes: changes in the activity
of enzymes that methylate cytosine (i.e., DNMTs), that hydroxylate
5-mC (i.e., TETs), and lastly that remove 5-mC intermediates (TDG,
APOBEC). The activity of TET can also be modulated by isocitrate
dehydrogenase (IDH)1/2 enzymes that take part in the Krebs cycle
in mitochondria. There is a potential relationship between IDH1/2
and TET because IDH1/2 produces ␣-KG, which is a cofactor in TET
activity, and certain mutations of IDH1/2 can lead to an accumulation of 2-hydroxyglutarate, which is an antagonist of TET activity.
Thus, mutations in IDH1/2 should lead to a decrease in TET activity
and lower levels of 5-hmC. TET activity may also be inhibited by the
accumulation of succinate brought on by deficiency in succinate
dehydrogenase in the Krebs cycle as reported in gastrointestinal
stromal tumors [220]. Another factor that determines TET activity is the intracellular concentration of vitamin C as discussed in
section 3.1.
Changes in the methylation of cytosine by DNMTs does not
appear to have a major influence on the genomic level of 5-hmC
in DNA. For example, the level of 5-mC in the DNA of brain tumors
is fairly constant (4-5%) as well as that in hepatocellular carcinoma
on the basis of recent analyses with a large number of specimens
using HPLC-MS and HPLC-MS/MS [90,218,221]. It should be noted
though that significant reductions of 5-mC in cancer cells have
been reported previously [222,223]. Thus, it is reasonable to propose that the methylation of cytosine to 5-mC by DNMTs does
not dramatically change in tumors. Furthermore, the above studies [90,218,221] clearly indicate that 5-mC is not correlated with
5-hmC in brain tumor specimens.
There is strong evidence implicating TET activity as a cause of 5hmC depletion in cancer cells although a link with IDH1/2 is elusive
in cases. TET2 mutations are common in myeloid malignancies and
they appear to be formed before other transformational events such
as JAC2 and MPL mutations [224]. In myeloproliferative neoplasms,
the global level of 5-mC in the DNA of peripheral blood lymphocytes from healthy subjects decreased 50% in afflicted patients, and
decreased a further 25% in patients with TET2 mutations as measured by HPLC-MS/MS [89]. The low level of 5-mC correlated well
with TET2 mutations in samples from patients and TET2 knockdown in myeloid cell lines [63,225]. In acute myeloid leukemia
(AML) mutations in TET2 and IDH1/2 were mutually exclusive
leading to global hypermethylation with similar gene-specific
methylation signatures [226]. Knock-down of TET2 in progeny
blood cells displayed profound effects on myeloid differentiation
and the pathogenesis of myeloid malignancies. TET2 deficient mice
developed chronic myelomonocytic leukemia [227,228]. In contrast to myeloid neoplasms, no TET mutations have been reported in
melanoma although 10% of cases harbor mutations in IDH1/2 [229].
However, the expression of both TET1/2/3 and IDH2 is down regulated in melanoma compared to benign nevi (moles). When TET2
was overexpressed in melanoma cells, the genome-wide marks
of 5-hmC increased by 5-fold compared to mutant cells containing catalytically inactive enzyme. TET2 overexpression appeared
to restore the landscape of 5-hmC in a manner that was very similar to the normal landscape of benign nevi. These studies were
also extended to relevant animal models confirming that overexpression of either TET2 or IDH2 increases the level of 5-hmC in
vivo.
Evidence linking 5-hmC with TET or IDH activity is lacking in
brain cancer. Normally, increases of 5-hmC follow TET expression
in human brain tissue but this depends on the region and stage
of development [76]. The low level of 5-hmC in gliomas may be
related to reduced expression of TET or alternative localization of
the protein [76,77] although there was no association between
the expression of TET proteins and 5-hmC levels in brain tumor
specimens as measured by HPLC-MS/MS [218]. The role of IDH
mutations in the accumulation of 5-hmC is ambiguous. Kraus et al.
reported a correlation between 5-hmC and IDH1 mutations in diffuse and anaplastic astrocytomas but not in glioblastomas [90].
IDH1 mutations have also been associated with decreased 5-hmC
as measured by immunohistochemistry of astroglioma specimens
[230]. Although IDH mutations are associated with a decrease
in 5-hmC, there are a large number cases of glioma with IDH1
mutations and high 5-hmC, and conversely, those without mutations and a low level of 5-hmC [77]. Interestingly, the presence
of IDH1 mutations was strongly correlated with the accumulation of TET1 in the nucleus but not with loss of 5-hmC. There
is a common set of genes that are hypermutated in tumors with
IDH mutations in hepatic cholangiocarcinoma and glioblastomas
[231]. Other studies have reported a weak association between
the level of 5-hmC and IDH1 mutations [76]. In the case of
gliomas, one cannot rule out passive demethylation due to the
lack of restoring methylation marks during rapid proliferation
[232].
Despite the lack of understanding 5-hmC depletion in cancer
cells [232], the level of 5-mC has great potential as a biomarker
toward tumor growth and aggressiveness, and ultimately patient
survival. The levels of 5-hmC are lower in more advanced stages of
glioma [90]. In a cohort of 225 patients, low 5-hmC levels, low TET
expression and high APOBEC expression were associated with poor
prognosis in malignant glioma [76]. The level of 5-hmC but not that
of 5-mC was associated with tumor progression in human hepatocellular carcinoma [221]. Lastly, AML patients with low TET2 confer
a poor prognosis [233]. Thus, as observed for glioma, a low level of
5-mC in myeloid malignancies is associated with TET mutations
and down-regulation, although in the latter case, IDH1 mutations
appear to play an important role. Increases of either IDH2 or TET2
suppress tumor invasion and growth and lengthen tumor-free survival in relevant animal models of melanoma [75], suggesting that
one could treat melanoma cells by reestablishing a normal 5-hmC
landscape.
As an exception to the above trend, evidence has been recently
provided for an aberrant overexpression of TET1 in MLL-rearranged
leukemia leading to an increase in the levels of 5-hmC measured
either by dot blot immunoassay or HPLC-MS/MS [234]. This contrasts to the down-regulation of TET1 and TET2 genes in several
types of solid tumors as discussed above that were proposed to have
a gene suppressing role. The suggested oncogenic role of TET1 was
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
rationalized in terms of binding of MLL-fusion proteins to critical
target genes including HOXA9 and MEISI [A14].
6. Conclusions and perspectives
The recent discovery of 5-hmC as the first oxidation product of
5-mC mediated by TET enzymes has provided during the last three
and a half years a strong impetus to the development of numerous
and comprehensive research activities in the field of epigenetics.
The subsequent identification of 5-fC and 5-caC as TET iterative
oxidation products has further questioned the epigenetic significance of 5-mC oxidation products and their likely implication in
the active demethylation pathway. Other important issues deal
with the biological functions of 5-hmC and TET enzymes that are
considered as key epigenetic modifiers and possible guardians of
CpG islands [120]. Most of these studies have necessitated global
measurement of 5-mC oxidation products using in particular accurate HPLC-MS/MS and also the development of high-throughput,
single-nucleotide resolution methods for mapping 5-hmC in the
genome. Further improvement of these analytical methods is however required since in particular immunoprecipitation methods
suffer from a lack of specificity of available antibodies against 5hmC and bias to the detection of regions with high density. The
situation is even more critical for 5-fC and 5-caC that are present in
DNA in much lower amounts. These studies should provide more
accurate information of the importance of dynamic methylation
and demethylation pathways. It remains also to be established
among unresolved issues whether 5-caC can be directly converted
into cytosine by a putative decarboxylase activity. The notion of
hydroxymetylated DNA biomarkers [235] is emerging due to the
likely implication of 5-hmC in cancer with the possibility for the
latter oxidized base to exhibit diagnostic and prognostic potential
[236].
Acknowledgment
We thank Natural Sciences and Engineering Research Council of
Canada (NSERC) and Canadian Institutes of Health Research (CIHR)
for financial support.
References
[1] F. Fuks, DNA methylation and histone modifications: teaming up to silence
genes, Curr. Opin. Genet. Dev. 15 (2005) 490–495.
[2] M.M. Suzuki, A. Bird, DNA methylation landscapes: provocative insights from
epigenomics, Nat. Rev. Genet. 9 (2008) 465–476.
[3] T. Suganuma, J.L. Workman, Signals and combinatorial functions of histone
modifications, 2011, pp. 473-499.
[4] Y. Fu, C. He, Nucleic acid modifications with epigenetic significance, Curr.
Opin. Chem. Biol. 16 (2012) 516–524.
[5] Z. Liutkeviit, G. Lukinaviius, V. Maseviius, D. Daujotyt, S. Klimaauskas,
Cytosine-5-methyltransferases add aldehydes to DNA, Nat. Chem. Biol. 5
(2009) 400–402.
[6] C. Hobartner, Enzymatic labeling of 5-hydroxymethylcytosine in DNA,
Angew. Chem. Int. Ed. 50 (2011) 4268–4270.
[7] T.H. Bestor, The DNA methyltransferases of mammals, Hum. Mol. Genet. 9
(2000) 2395–2402.
[8] J.A. Law, S.E. Jacobsen, Establishing, maintaining and modifying DNA methylation patterns in plants and animals, Nat. Rev. Genet. 11 (2010) 204–220.
[9] G. Auclair, M. Weber, Mechanisms of DNA methylation and demethylation in
mammals, Biochimie 94 (2012) 2202–2211.
[10] Z.D. Smith, M.M. Chan, T.S. Mikkelsen, H. Gu, A. Gnirke, A. Regev, A. Meissner, A
unique regulatory phase of DNA methylation in the early mammalian embryo,
Nature 484 (2012) 339–344.
[11] M. Hergersberg, Biological aspects of cytosine methylation in eukaryotic cells,
Experientia 47 (1991) 1171–1185.
[12] T. Carell, C. Brandmayr, A. Hienzsch, M. Müller, D. Pearson, V. Reiter, I. Thoma,
P. Thumbs, M. Wagner, Structure and function of noncanonical nucleobases,
Angew. Chem. Int. Ed. 51 (2012) 7110–7131.
[13] R.D. Hotchkiss, The quantitative separation of purines, pyrimidines, and
nucleosides by paper chromatography, J. Biol. Chem. 175 (1948) 315–332.
[14] G.R. Wyatt, Recognition and estimation of 5-methylcytosine in nucleic acids,
Biochem. J. 48 (1951) 581–584.
31
[15] W. Reik, Stability and flexibility of epigenetic gene regulation in mammalian
development, Nature 447 (2007) 425–432.
[16] Y.I. Tsukada, Hydroxylation mediates chromatin demethylation, J. Biochem.
151 (2012) 229–246.
[17] E. Li, Chromatin modification and epigenetic reprogramming in mammalian
development, Nat. Rev. Genet. 3 (2002) 662–673.
[18] R. Jaenisch, A. Bird, Epigenetic regulation of gene expression: how the
genome integrates intrinsic and environmental signals, Nat. Genet. 33 (2003)
245–254.
[19] K.D. Robertson, DNA methylation and human disease, Nat. Rev. Genet. 6
(2005) 597–610.
[20] R.J. Klose, A.P. Bird, Genomic DNA methylation: the mark and its mediators,
Trends Biochem. Sci. 31 (2006) 89–97.
[21] P.A. Jones, Functions of DNA methylation: islands, start sites, gene bodies and
beyond, Nat. Rev. Genet. 13 (2012) 484–492.
[22] J.R. Wagner, J. Cadet, Oxidation reactions of cytosine DNA components by
hydroxyl radical and one-electron oxidants in aerated aqueous solutions, Acc.
Chem. Res. 43 (2010) 564–571.
[23] G.R. Wyatt, S.S. Cohen, The bases of the nucleic acids of some bacterial and
animal viruses: the occurrence of 5-hydroxymethylcytosine, Biochem. J. 55
(1953) 774–782.
[24] G.R. Wyatt, S.S. Cohen, A new pyrimidine base from bacteriophage nucleic
acids, Nature 170 (1952) 1072–1073.
[25] N.W. Penn, R. Suwalski, C. O‘Riley, K. Bojanowski, R. Yura, The presence of
5-hydroxymethylcytosine in animal deoxyribonucleic acid, Biochem. J. 126
(1972) 781–790.
[26] R.M. Kothari, V. Shankar, 5 Methylcytosine content in the vertebrate deoxyribonucleic acids: species specificity, J. Mol. Evol. 7 (1976) 325–329.
[27] M. Tahiliani, K.P. Koh, Y.H. Shen, W.A. Pastor, H. Bandukwala, Y. Brudno, S.
Agarwal, L.M. Iyer, D.R. Liu, L. Aravind, A. Rao, Conversion of 5-methylcytosine
to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1, Science 324 (2009) 930–935.
[28] S. Kriaucionis, N. Heintz, The nuclear DNA base 5-hydroxymethylcytosine
Is present in Purkinje neurons and the brain, Science 324 (2009)
929–930.
[29] M. Munzel, D. Globisch, T. Bruckl, M. Wagner, V. Welzmiller, S. Michalakis, M.
Muller, M. Biel, T. Carell, Quantification of the sixth DNA base hydroxymethylcytosine in the brain, Angew. Chem. Int. Ed. 49 (2010) 5375–5377.
[30] D. Globisch, M. Münzel, M. Müller, S. Michalakis, M. Wagner, S. Koch, T. Brückl,
M. Biel, T. Carell, Tissue distribution of 5-hydroxymethylcytosine and search
for active demethylation intermediates, PLoS One 5 (2010) e15367.
[31] C.X. Song, C. He, The hunt for 5-hydroxymethylcytosine: the sixth base, Epigenomics 3 (2011) 521–523.
[32] V.K.C. Ponnaluri, J.P. Maciejewski, M. Mukherji, A mechanistic overview of
TET-mediated 5-methylcytosine oxidation, Biochem. Biophys. Res. Commun.
436 (2013) 115–120.
[33] H. Zhao, T. Chen, Tet family of 5-methylcytosine dioxygenases in mammalian
development, J. Hum. Genet. 58 (2013) 421–427.
[34] L. Shen, Y. Zhang, 5-Hydroxymethylcytosine: generation, fate, and genomic
distribution, Curr. Opin. Cell Biol. 25 (2013) 289–296.
[35] J.A. Hackett, J.J. Zylicz, M.A. Surani, Parallel mechanisms of epigenetic reprogramming in the germline, Trends Genet. 28 (2012) 164–174.
[36] M.R. Branco, G. Ficz, W. Reik, Uncovering the role of 5-hydroxymethylcytosine
in the epigenome, Nat. Rev. Genet. 13 (2012) 7–13.
[37] S.R.M. Kinney, S. Pradhan, Ten eleven translocation enzymes and 5hydroxymethylation in mammalian development and cancer, in: A.R. Karpf
(Ed.) 2013, pp. 57-79.
[38] F. Matarese, E. Carrillo-de Santa Pau, H.G. Stunnenberg, 5Hydroxymethylcytosine: a new kid on the epigenetic block? Mol. Syst.
Biol. 7 (2011) 562.
[39] G. Jia, Y. Fu, C. He, Reversible RNA adenosine methylation in biological regulation, Trends Genet. 29 (2013) 108–115.
[40] L.M. Iyer, M. Tahiliani, A. Rao, L. Aravind, Prediction of novel families of
enzymes involved in oxidative and other complex modifications of bases in
nucleic acids, Cell Cycle 8 (2009) 1698–1710.
[41] G. Jia, Y. Fu, X. Zhao, Q. Dai, G. Zheng, Y. Yang, C. Yi, T. Lindahl, T. Pan, Y.G.
Yang, C. He, N6-Methyladenosine in nuclear RNA is a major substrate of the
obesity-associated FTO, Nat. Chem. Biol. 7 (2011) 885–887.
[42] T. Duncan, S.C. Trewick, P. Koivisto, P.A. Bates, T. Lindahl, B. Sedgwick, Reversal
of DNA alkylation damage by two human dioxygenases, Proc. Natl. Acad. Sci.
U.S.A. 99 (2002) 16660–16665.
[43] P.A. Aas, M. Otterlei, P.Ø. Falnes, C.B. Vågbø, F. Skorpen, M. Akbari, O. Sundheim, M. Bjørås, G. Slupphaug, E. Seeberg, H.E. Krokan, Human and bacterial
oxidative demethylases repair alkylation damage in both RNA and DNA,
Nature 421 (2003) 859–863.
[44] B. Sedgwick, P.A. Bates, J. Paik, S.C. Jacobs, T. Lindahl, Repair of alkylated DNA:
recent advances, DNA Repair 6 (2007) 429–442.
[45] B.H. Ramsahoye, Nearest-neighbor analysis, Methods Mol. Biol. (Clifton, N.J.)
200 (2002) 9–15.
[46] S. Ito, L. Shen, Q. Dai, S.C. Wu, L.B. Collins, J.A. Swenberg, C. He, Y. Zhang,
Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5carboxylcytosine, Science 333 (2011) 1300–1303.
[47] Y.F. He, B.Z. Li, Z. Li, P. Liu, Y. Wang, Q. Tang, J. Ding, Y. Jia, Z. Chen, N. Li, Y.
Sun, X. Li, Q. Dai, C.X. Song, K. Zhang, C. He, G.L. Xu, Tet-mediated formation
of 5-carboxylcytosine and its excision by TDG in mammalian DNA, Science
333 (2011) 1303–1307.
32
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
[48] A. Szwagierczak, S. Bultmann, C.S. Schmidt, F. Spada, H. Leonhardt, Sensitive enzymatic quantification of 5-hydroxymethylcytosine in genomic DNA,
Nucleic Acids Res. 38 (2010) e181.
[49] J. Terragni, J. Bitinaite, Y. Zheng, S. Pradhan, Biochemical characterization of recombinant ␤-glucosyltransferase and analysis of global
5-hydroxymethylcytosine in unique genomes, Biochemistry 51 (2012)
1009–1019.
[50] J. Cadet, J.R. Wagner, DNA base damage by reactive oxygen species, oxidizing agents, and UV radiation, Cold Springs Harb. Perspect. Biol. 5 (2013)
a012559.
[51] E. Kriukiene, Z. Liutkevičiute, S. Klimašauskas, 5-Hydroxymethylcytosine –
the elusive epigenetic mark in mammalian DNA, Chem. Soc. Rev. 41 (2012)
6916–6930.
[52] C.X. Song, K.E. Szulwach, Y. Fu, Q. Dai, C. Yi, X. Li, Y. Li, C.H. Chen, W. Zhang, X.
Jian, J. Wang, L. Zhang, T.J. Looney, B. Zhang, L.A. Godley, L.M. Hicks, B.T. Lahn,
P. Jin, C. He, Selective chemical labeling reveals the genome-wide distribution
of 5-hydroxymethylcytosine, Nat. Biotechnol. 29 (2011) 68–75.
[53] C.X. Song, C. He, Bioorthogonal labeling of 5-hydroxymethylcytosine in
genomic DNA and diazirine-based DNA Photo-cross-linking probes, Acc.
Chem. Res. 44 (2011) 709–717.
[54] Z. Liutkeviciute, E. Kriukiene, I. Grigaityte, V. Masevicius, S. Klimasauskas,
Methyltransferase-directed derivatization of 5-hydroxymethylcytosine in
DNA, Angew. Chem. Int. Ed. 50 (2011) 2090–2093.
[55] S. Ito, A.C. Dalessio, O.V. Taranova, K. Hong, L.C. Sowers, Y. Zhang, Role of tet
proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass
specification, Nature 466 (2010) 1129–1133.
[56] K. Iqbal, S.G. Jin, G.P. Pfeifer, P.E. Szabo, Reprogramming of the paternal genome upon fertilization involves genome-wide oxidation of 5methylcytosine, Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 3642–3647.
[57] G. Ficz, M.R. Branco, S. Seisenberger, F. Santos, F. Krueger, T.A. Hore, C.J. Marques, S. Andrews, W. Reik, Dynamic regulation of 5-hydroxymethylcytosine
in mouse ES cells and during differentiation, Nature 473 (2011) 398–402.
[58] C.E. Nestor, R. Ottaviano, J. Reddington, D. Sproul, D. Reinhardt, D. Dunican,
E. Katz, J.M. Dixon, D.J. Harrison, R.R. Meehan, Tissue type is a major modifier
of the 5-hydroxymethylcytosine content of human genes, Genome Res. 22
(2012) 467–477.
[59] W. Li, M. Liu, Distribution of 5-hydroxymethylcytosine in different human
tissues, J. Nucleic Acids 2011 (2011) 870726.
[60] K. Williams, J. Christensen, M.T. Pedersen, J.V. Johansen, P.A.C. Cloos, J. Rappsilber, K. Helin, TET1 and hydroxymethylcytosine in transcription and DNA
methylation fidelity, Nature 473 (2011) 343–349.
[61] L.G. Acevedo, A. Sanz, M.A. Jelinek, Novel DNA binding domain-based assays
for detection of methylated and nonmethylated DNA, Epigenomics 3 (2011)
93–101.
[62] J. Cadet, T. Douki, J.L. Ravanat, J.R. Wagner, Measurement of oxidatively generated base damage to nucleic acids in cells: facts and artifacts, Bioanal. Rev.
4 (2012) 55–74.
[63] M. Ko, Y. Huang, A.M. Jankowska, U.J. Pape, M. Tahiliani, H.S. Bandukwala, J.
An, E.D. Lamperti, K.P. Koh, R. Ganetzky, X.S. Liu, L. Aravind, S. Agarwal, J.P.
Maciejewski, A. Rao, Impaired hydroxylation of 5-methylcytosine in myeloid
cancers with mutant TET2, Nature 468 (2010) 839–843.
[64] S.G. Jin, S. Kadam, G.P. Pfeifer, Examination of the specificity of
DNA methylation profiling techniques towards 5-methylcytosine and 5hydroxymethylcytosine, Nucleic Acids Res. 38 (2010) e125.
[65] H. Wu, A.C. D‘Alessio, S. Ito, Z. Wang, K. Cui, K. Zhao, Y.E. Sun, Y. Zhang,
Genome-wide analysis of 5-hydroxymethylcytosine distribution reveals its
dual function in transcriptional regulation in mouse embryonic stem cells,
Genes Dev. 25 (2011) 679–684.
[66] A. Ruzov, Y. Tsenkina, A. Serio, T. Dudnakova, J. Fletcher, Y. Bai, T. Chebotareva, S. Pells, Z. Hannoun, G. Sullivan, S. Chandran, D.C. Hay, M. Bradley, I.
Wilmut, P. De Sousa, Lineage-specific distribution of high levels of genomic
5-hydroxymethylcytosine in mammalian development, Cell Res. 21 (2011)
1332–1342.
[67] A. Inoue, S. Matoba, Y. Zhang, Transcriptional activation of transposable elements in mouse zygotes is independent of Tet3-mediated 5-methylcytosine
oxidation, Cell Res. 22 (2012) 1640–1649.
[68] T.P. Gu, F. Guo, H. Yang, H.P. Wu, G.F. Xu, W. Liu, Z.G. Xie, L. Shi, X. He, S.G.
Jin, K. Iqbal, Y.G. Shi, Z. Deng, P.E. Szabó, G.P. Pfeifer, J. Li, G.L. Xu, The role of
Tet3 DNA dioxygenase in epigenetic reprogramming by oocytes, Nature 477
(2011) 606–612.
[69] G.T. Morgan, P. Jones, M. Bellini, Association of modified cytosines and the
methylated DNA-binding protein MeCP2 with distinctive structural domains
of lampbrush chromatin, Chromosome Res. 20 (2012) 925–942.
[70] J.A. Hackett, R. Sengupta, J.J. Zylicz, K. Murakami, C. Lee, T.A. Down, M.A.
Surani, Germline DNA demethylation dynamics and imprint erasure through
5-hydroxymethylcytosine, Science 339 (2013) 448–452.
[71] S. Yamaguchi, K. Hong, R. Liu, A. Inoue, L. Shen, K. Zhang, Y. Zhang, Dynamics
of 5-methylcytosine and 5-hydroxymethylcytosine during germ cell reprogramming, Cell Res. 23 (2013) 329–339.
[72] J. Vincent, Y. Huang, P.Y. Chen, S. Feng, J. Calvopiña, K. Nee, S. Lee, T. Le, A.
Yoon, K. Faull, G. Fan, A. Rao, S. Jacobsen, M. Pellegrini, A. Clark, Stage-specific
roles for Tet1 and Tet2 in DNA demethylation in primordial germ cells, Cell
Stem Cell 12 (2013) 470–478.
[73] M. Mellén, P. Ayata, S. Dewell, S. Kriaucionis, N. Heintz, MeCP2 binds to 5hmC
enriched within active genes and accessible chromatin in the nervous system,
Cell 151 (2012) 1417–1430.
[74] M. Hahn, R. Qiu, X. Wu, A. Li, H. Zhang, J. Wang, J. Jui, S.G. Jin, Y. Jiang, G.
Pfeifer, Q. Lu, Dynamics of 5-hydroxymethylcytosine and chromatin marks in
mammalian neurogenesis, Cell Rep. 3 (2013) 291–300.
[75] C.G. Lian, Y. Xu, C. Ceol, F. Wu, A. Larson, K. Dresser, W. Xu, L. Tan, Y. Hu, Q.
Zhan, C.W. Lee, D. Hu, B.Q. Lian, S. Kleffel, Y. Yang, J. Neiswender, A.J. Khorasani,
R. Fang, C. Lezcano, L.M. Duncan, R.A. Scolyer, J.F. Thompson, H. Kakavand,
Y. Houvras, L.I. Zon, M.C. Mihm Jr., U.B. Kaiser, T. Schatton, B.A. Woda, G.F.
Murphy, Y.G. Shi, Loss of 5-hydroxymethylcytosine is an epigenetic hallmark
of melanoma, Cell 150 (2012) 1135–1146.
[76] B.A. Orr, M.C. Haffner, W.G. Nelson, S. Yegnasubramanian, C.G. Eberhart,
Decreased 5-Hydroxymethylcytosine is associated with neural progenitor
phenotype in normal brain and shorter survival in malignant glioma, PLoS
One 7 (2012) e41036.
[77] A. Waha, T. Müller, M. Gessi, L.J. Isselstein, D. Luxen, D. Freihoff, J. Freihoff, A.
Becker, M. Simon, J. Hammes, D. Denkhaus, A. Zur Mühlen, T. Pietsch, Nuclear
exclusion of TET1 is associated with loss of 5-hydroxymethylcytosine in IDH1
wild-type gliomas, Am. J. Pathol. 181 (2012) 675–683.
[78] M.C. Haffner, A. Chaux, A.K. Meeker, D.M. Esopi, J. Gerber, L.G. Pellakuru, A.
Toubaji, P. Argani, C. Iacobuzio-Donahue, W.G. Nelson, G.J. Netto, A.M. De
Marzo, S. Yegnasubramanian, Global 5-hydroxymethylcytosine content is significantly reduced in tissue stem/progenitor cell compartments and in human
cancers, Oncotarget 2 (2011) 627–637.
[79] A. Inoue, L. Shen, Q. Dai, C. He, Y. Zhang, Generation and replication-dependent
dilution of 5fC and 5caC during mouse preimplantation development, Cell Res.
21 (2011) 1670–1676.
[80] E.A. Raiber, D. Beraldi, G. Ficz, H. Burgess, M.R. Branco, P. Murat, D. Oxley,
M.J. Booth, W. Reik, S. Balasubramanian, Genome-wide distribution of 5formylcytosine in ES cells is associated with transcription and depends on
thymine DNA glycosylase, Genome Biol. 13 (2012) R69.
[81] A. Alioui, L.M. Wheldon, A. Abakir, Z. Ferjentsik, A.D. Johnson, A. Ruzov, 5Carboxylcytosine is localized to euchromatic regions in the nuclei of follicular
cells in axolotl ovary, Nucleus 3 (2012) 565–569.
[82] J. Cadet, H. Poulsen, Measurement of oxidatively generated base damage in
cellular DNA and urine, Free Radic. Biol. Med. 48 (2010) 1457–1459.
[83] J. Cadet, S. Mouret, J.L. Ravanat, T. Douki, Photoinduced damage to cellular
DNA: direct and photosensitized reactions, Photochem. Photobiol. 88 (2012)
1048–1065.
[84] J. Cadet, T. Douki, J.L. Ravanat, Oxidatively generated base damage to cellular
DNA, Free Radic. Biol. Med. 49 (2010) 9–21.
[85] T. Pfaffeneder, B. Hackner, M. Truß, M. Münzel, M. Müller, C.A. Deiml, C. Hagemeier, T. Carell, The discovery of 5-formylcytosine in embryonic stem cell
DNA, Angew. Chem. Int. Ed. 50 (2011) 7008–7012.
[86] T. Le, K.P. Kim, G. Fan, K.F. Faull, A sensitive mass spectrometry method for
simultaneous quantification of DNA methylation and hydroxymethylation
levels in biological samples, Anal. Biochem. 412 (2011) 203–209.
[87] L. Shen, Y. Zhang, Enzymatic analysis of tet proteins: Key enzymes in the
metabolism of DNA methylation, 2012, pp. 93–105.
[88] S. Liu, J. Wang, Y. Siu, C. Guerrero, Y. Zeng, D. Mitra, P.J. Brooks, D.E. Fisher,
H. Song, Y. Wang, Quantitative assessment of Tet-induced oxidation products
of 5-methylcytosine in cellular and tissue DNA, Nucleic Acids Res. 41 (2013)
6421–6429.
[89] E. Pronier, C. Almire, H. Mokrani, A. Vasanthakumar, A. Simon, B. Da Costa Reis
Monte Mor, A. Massé, J.P. Le Couédic, F. Pendino, B. Carbonne, J. Larghero,
J.L. Ravanat, N. Casadevall, O.A. Bernard, N. Droin, E. Solary, L.A. Godley,
W. Vainchenker, I. Plo, F. Delhommeau, Inhibition of TET2-mediated conversion of 5-methylcytosine to 5-hydroxymethylcytosine disturbs erythroid
and granulomonocytic differentiation of human hematopoietic progenitors,
Blood 118 (2011) 2551–2555.
[90] T.F.J. Kraus, D. Globisch, M. Wagner, S. Eigenbrod, D. Widmann, M. Münzel,
M. Müller, T. Pfaffeneder, B. Hackner, W. Feiden, U. Schüller, T. Carell, H.A.
Kretzschmar, Low values of 5-hydroxymethylcytosine (5hmC), the sixth base,
are associated with anaplasia in human brain tumors, Int. J. Cancer 131 (2012)
1577–1590.
[91] C. Spruijt, F. Gnerlich, A. Smits, T. Pfaffeneder, P.T.C. Jansen, C. Bauer, M.
Münzel, M. Wagner, M. Müller, F. Khan, H. Eberl, A. Mensinga, A. Brinkman,
K. Lephikov, U. Müller, J. Walter, R. Boelens, H. van Ingen, H. Leonhardt, T.
Carell, M. Vermeulen, Dynamic readers for 5-(hydroxy)methylcytosine and
its oxidized derivatives, Cell (2013) 1146–1159.
[92] H. Gan, L. Wen, X. Liao, T. Lin, J. Ma, J. Liu, C.-X. Song, M. Wang, C. He, C. Han, F.
Tan, Dynamics of 5-hydroxymethylcytosine during mouse spermatogenesis,
Nat. Commun. 4 (2013) 1995.
[93] Y. Tang, J.M. Chu, W. Huang, J. Xiong, X.W. Xing, X. Zhou, Y.Q. Feng, B.F. Yuan,
Hydrophilic material for the selective enrichment of 5- hydroxymethylcytosine and its LC-MS/MS detection, Anal. Chem. 85 (2013) 6129–6135.
[94] M.L. Chen, F. Shen, W. Huang, J.H. Qi, Y. Wang, Y.Q. Feng, S.M. Liu, B.F.
Yuan, Quantification of 5-methylcytosine and 5-hydroxymethylcytosine in
genomic DNA from hepatocellular carcinoma tissues by capillary hydrophilicinteraction liquid chromatography/quadrupole TOF mass spectrometry, Clin.
Chem. 59 (2013) 824–832.
[95] A.M. Krais, Y.J. Park, C. Plass, H.H. Schmeiser, Determination of genomic 5hydroxymethyl-2 -deoxycytidine in human DNA by capillary electrophoresis
with laser-induced fluorescence, Epigenetics 6 (2011) 560–565.
[96] K. Grønbæk, C. Müller-Tidow, G. Perini, S. Lehmann, M.B. Treppendahl, K.
Mills, C. Plass, B. Schlegelberger, A critical appraisal of tools available for monitoring epigenetic changes in clinical samples from patients with myeloid
malignancies, Haematologica 97 (2012) 1380–1388.
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
[97] T.N. Ndlovu, H. Denis, F. Fuks, Exposing the DNA methylome iceberg, Trends
Biochem. Sci. 36 (2011) 381–387.
[98] D. Zilberman, S. Henikoff, Genome-wide analysis of DNA methylation patterns, Development 134 (2007) 3959–3965.
[99] C.S. Ku, N. Naidoo, M. Wu, R. Soong, Studying the epigenome using next
generation sequencing, J. Med. Genet. 48 (2011) 721–730.
[100] C.X. Song, C. Yi, C. He, Mapping recently identified nucleotide variants in the
genome and transcriptome, Nat. Biotechnol. 30 (2012) 1107–1116.
[101] R.P. Nagarajan, S.D., Fouse, R.J.A., Bell, J.F. Costello, Methods for cancer
epigenome analysis, in: A.R. Karpf (Ed.), 2013, pp. 313–338.
[102] S.G. Jin, X. Wu, A.X. Li, G.P. Pfeifer, Genomic mapping of 5hydroxymethylcytosine in the human brain, Nucleic Acids Res. 39 (2011)
5015–5024.
[103] W.A. Pastor, U.J. Pape, Y. Huang, H.R. Henderson, R. Lister, M. Ko, E.M.
McLoughlin, Y. Brudno, S. Mahapatra, P. Kapranov, M. Tahiliani, G.Q. Daley,
X.S. Liu, J.R. Ecker, P.M. Milos, S. Agarwal, A. Rao, Genome-widemapping of 5hydroxymethylcytosine in embryonic stem cells, Nature 473 (2011) 394–397.
[104] Y. Huang, W.A. Pastor, J.A. Zepeda-Martínez, A. Rao, The anti-CMS technique
for genome-wide mapping of 5-hydroxymethylcytosine, Nat. Protoc. 7 (2012)
1897–1908.
[105] A.B. Robertson, J.A. Dahl, R. Ougland, A. Klungland, Pull-down of 5hydroxymethylcytosine DNA using JBP1-coated magnetic beads, Nat. Protoc.
7 (2012) 340–350.
[106] A.B. Robertson, J.A. Dahl, C.B. Vagbo, P. Tripathi, H.E. Krokan, A. Klungland, A novel method for the efficient and selective identification of
5-hydroxymethylcytosine in genomic DNA, Nucleic Acids Res. 39 (2011) e55.
[107] Z. Sun, J. Terragni, J.G. Borgaro, Y. Liu, L. Yu, S. Guan, H. Wang, D. Sun, X. Cheng,
Z. Zhu, S. Pradhan, Y. Zheng, High-resolution enzymatic mapping of genomic
5-hydroxymethylcytosine in mouse embryonic stem cells, Cell Rep. 3 (2013)
567–576.
[108] J.G. Borgaro, Z. Zhu, Characterization of the 5-hydroxymethylcytosinespecific DNA restriction endonucleases, Nucleic Acids Res. 41 (2013)
4198–4206.
[109] M. Yu, G.C. Hon, K.E. Szulwach, C.X. Song, L. Zhang, A. Kim, X. Li, Q. Dai,
Y. Shen, B. Park, J.H. Min, P. Jin, B. Ren, C. He, Base-resolution analysis
of 5-hydroxymethylcytosine in the mammalian genome, Cell 149 (2012)
1368–1380.
[110] M.J. Booth, M.R. Branco, G. Ficz, D. Oxley, F. Krueger, W. Reik, S.
Balasubramanian, Quantitative sequencing of 5-methylcytosine and 5hydroxymethylcytosine at single-base resolution, Science 336 (2012)
934–937.
[111] X. Lu, C.X. Song, K. Szulwach, Z. Wang, P. Weidenbacher, P. Jin, C. He, Chemical
modification-assisted bisulfite sequencing (CAB-seq) for 5-carboxylcytosine
detection in DNA, J. Am. Chem. Soc. 135 (2013) 9315–9317.
[112] C.X. Song, T.A. Clark, X.Y. Lu, A. Kislyuk, Q. Dai, S.W. Turner, C. He, J. Korlach,
Sensitive and specific single-molecule sequencing of 5- hydroxymethylcytosine, Nat. Methods 9 (2012) 75–77.
[113] K.E. Szulwach, X. Li, Y. Li, C.X. Song, H. Wu, Q. Dai, H. Irier, A.K. Upadhyay, M.
Gearing, A.I. Levey, A. Vasanthakumar, L.A. Godley, Q. Chang, X. Cheng, C. He,
P. Jin, 5-hmC-mediated epigenetic dynamics during postnatal neurodevelopment and aging, Nat. Neurosci. 14 (2011) 1607–1616.
[114] K.E. Szulwach, X. Li, Y. Li, C.X. Song, J.W. Han, S. Kim, S. Namburi, K.
Hermetz, J.J. Kim, M.K. Rudd, Y.S. Yoon, B. Ren, C. He, P. Jin, Integrating 5hydroxymethylcytosine into the epigenomic landscape of human embryonic
stem cells, PLoS Genet. 7 (2011) e1002154.
[115] W.A. Pastor, Y. Huang, H.R. Henderson, S. Agarwal, A. Rao, The GLIB technique
for genome-wide mapping of 5-hydroxymethylcytosine, Nat. Protoc. 7 (2012)
1909–1917.
[116] C.X. Song, K.E. Szulwach, Q. Dai, Y. Fu, S.Q. Mao, L. Lin, C. Street, Y. Li, M.
Poidevin, H. Wu, J. Gao, P. Liu, L. Li, G.L. Xu, P. Jin, C. He, Genome-wide profiling
of 5-formylcytosine reveals its roles in epigenetic priming, Cell 153 (2013)
678–691.
[117] W.W. Li, L. Gong, H. Bayley, Single-molecule detection of 5hydroxymethylcytosine in DNA through chemical modification and
nanopore analysis, Angew. Chem. Int. Ed. 52 (2013) 4350–4355.
[118] S. Tardy-Planechaud, J. Fujimoto, S.S. Lin, L.C. Sowers, Solid phase synthesis
and restriction endonuclease cleavage of oligodeoxynucleotides containing
5-(hydroxymethyl)-cytosine, Nucleic Acids Res. 25 (1997) 553–558.
[119] T. Khare, S. Pai, K. Koncevicius, M. Pal, E. Kriukiene, Z. Liutkeviciute, M. Irimia,
P. Jia, C. Ptak, M. Xia, R. Tice, M. Tochigi, S. Moréra, A. Nazarians, D. Belsham,
A.H.C. Wong, B.J. Blencowe, S.C. Wang, P. Kapranov, R. Kustra, V. Labrie, S.
Klimasauskas, A. Petronis, 5-hmC in the brain is abundant in synaptic genes
and shows differences at the exon-intron boundary, Nat. Struct. Mol. Biol. 19
(2012) 1037–1044.
[120] K. Williams, J. Christensen, K. Helin, DNA methylation: TET proteins-guardians
of CpG islands? EMBO Rep. 13 (2012) 28–35.
[121] M.M. Dawlaty, A. Breiling, T. Le, G. Raddatz, M.I. Barrasa, A.W. Cheng, Q. Gao,
B.E. Powell, Z. Li, M. Xu, K.F. Faull, F. Lyko, R. Jaenisch, Combined deficiency of
Tet1 and Tet2 causes epigenetic abnormalities but is compatible with postnatal Development, Dev. Cell 24 (2013) 310–323.
[122] K. Iqbal, S.G. Jin, G.P. Pfeifer, P.E. Szabó, Reprogramming of the
paternal genome upon fertilization involves genome-wide oxidation of 5-methylcytosine, Proc. Natl. Acad. Sci. U.S.A. 108 (2011)
3642–3647.
[123] P.E. Szabó, G.P. Pfeifer, H3K9me2 attracts PGC7 in the zygote to prevent Tet3mediated oxidation of 5-methylcytosine, J. Mol. Cell Biol. 4 (2012) 427–429.
33
[124] L.S. Shock, P.V. Thakkar, E.J. Peterson, R.G. Moran, S.M. Taylor, DNA methyltransferase 1, cytosine methylation, and cytosine hydroxymethylation in
mammalian mitochondria, Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 3630–3635.
[125] C. Loenarz, C.J. Schofield, Oxygenase catalyzed 5-methylcytosine hydroxylation, Chem. Biol. 16 (2009) 580–583.
[126] S.C. Wu, Y. Zhang, Active DNA demethylation: many roads lead to Rome, Nat.
Rev. Mol. Cell Biol. 11 (2010) 607–620.
[127] J.A. Smiley, M. Kundracik, D.A. Landfried, V.R. Barnes Sr., A.A. Axhemi, Genes of
the thymidine salvage pathway: thymine-7-hydroxylase from a Rhodotorula
glutinis cDNA library and iso-orotate decarboxylase from Neurospora crassa,
Biochim. Biophys. Acta 1723 (2005) 256–264.
[128] J.M. Simmons, T.A. Müller, R.P. Hausinger, FeII/␣-ketoglutarate hydroxylases
involved in nucleobase, nucleoside, nucleotide, and chromatin metabolism,
Dalton Trans. (2008) 5132–5142.
[129] J.W. Neidigh, A. Darwanto, A.A. Williams, N.R. Wall, L.C. Sowers, Cloning and
characterization of Rhodotorula glutinis thymine hydroxylase, Chem. Res.
Toxicol. 22 (2009) 885–893.
[130] A. Maiti, A.C. Drohat, Thymine DNA glycosylase can rapidly excise 5formylcytosine and 5-carboxylcytosine: Potential implications for active
demethylation of CpG sites, J. Biol. Chem. 286 (2011) 35334–35338.
[131] L. Tan, Y.G. Shi, Tet family proteins and 5-hydroxymethylcytosine in development and disease, Development 139 (2012) 1895–1902.
[132] Z. Yu, P.A. Genest, B. ter Riet, K. Sweeney, C. DiPaolo, R. Kieft, E. Christodoulou,
A. Perrakis, J.M. Simmons, R.P. Hausinger, H.G.A.M. van Luenen, D.J. Rigden, R.
Sabatini, P. Borst, The protein that binds to DNA base J in trypanosomatids has
features of a thymidine hydroxylase, Nucleic Acids Res. 35 (2007) 2107–2115.
[133] L.J. Cliffe, G. Hirsch, J. Wang, D. Ekanayake, W. Bullard, M. Hu, Y. Wang, R. Sabatini, JBP1 and JBP2 proteins are Fe2+ /2-oxoglutarate-dependent dioxygenases
regulating hydroxylation of thymidine residues in trypanosome DNA, J. Biol.
Chem. 287 (2012) 19886–19895.
[134] H. Zhang, X. Zhang, E. Clark, M. Mulcahey, S. Huang, Y.G. Shi, TET1 is a DNAbinding protein that modulates DNA methylation and gene transcription via
hydroxylation of 5-methylcytosine, Cell Res. 20 (2010) 1390–1393.
[135] J. Song, O. Rechkoblit, T.H. Bestor, D.J. Patel, Structure of DNMT1–DNA complex reveals a role for autoinhibition in maintenance DNA methylation,
Science 331 (2011) 1036–1040.
[136] C. Xu, C. Bian, R. Lam, A. Dong, J. Min, The structural basis for selective binding
of non-methylated CpG islands by the CFP1 CXXC domain, Nat. Commun. 2
(2011) 227.
[137] M. Ko, J. An, H.S. Bandukwala, L. Chavez, T. Äijö, W.A. Pastor, M.F. Segal, H.
Li, K.P. Koh, H. Lähdesmäki, P.G. Hogan, L. Aravind, A. Rao, Modulation of
TET2 expression and 5-methylcytosine oxidation by the CXXC domain protein
IDAX, Nature 497 (2013) 122–126.
[138] M.A. McDonough, C. Loenarz, R. Chowdhury, I.J. Clifton, C.J. Schofield, Structural studies on human 2-oxoglutarate dependent oxygenases, Curr. Opin.
Struct. Biol. 20 (2010) 659–672.
[139] C. Loenarz, C.J. Schofield, Physiological and biochemical aspects of hydroxylations and demethylations catalyzed by human 2-oxoglutarate oxygenases,
Trends Biochem. Sci. 36 (2011) 7–18.
[140] C. Krebs, D.G. Fujimori, C.T. Walsh, J.M. Bollinger Jr., Non-heme Fe(IV)-oxo
intermediates, Acc. Chem. Res. 40 (2007) 484–492.
[141] T.D.H. Bugg, Dioxygenase enzymes: catalytic mechanisms and chemical models, Tetrahedron 59 (2003) 7075–7101.
[142] T.L. Chung, R.M. Brena, G. Kolle, S.M. Grimmond, B.P. Berman, P.W. Laird,
M.F. Pera, E.J. Wolvetang, Vitamin C promotes widespread yet specific DNA
demethylation of the epigenome in human embryonic stem cells, Stem Cells
28 (2010) 1848–1855.
[143] K. Blaschke, K.T. Ebata, M.M. Karimi, J.A. Zepeda-Martínez, P. Goyal, S. Mahapatra, A. Tam, D.J. Laird, M. Hirst, A. Rao, M.C. Lorincz, M. Ramalho-Santos,
Vitamin C induces Tet-dependent DNA demethylation and a blastocyst-like
state in ES cells, Nature 500 (2013) 222–226.
[144] E.A. Minor, B.L. Court, J.I. Young, G. Wang, Ascorbate induces ten-eleven
translocation (Tet) methylcytosine dioxygenase-mediated generation of 5hydroxymethylcytosine, J. Biol. Chem. 288 (2013) 13669–13674.
[145] R. Yin, S.Q. Mao, B. Zhao, Z. Chong, Y. Yang, Ascorbic acid enhances Tetmediated 5-methylcytosine oxidation and promotes DNA demethylation in
mammals, J. Am. Chem. Soc. 135 (2013) 10396–10403.
[146] J. Cadet, S. Loft, R. Olinski, M.D. Evans, K. Bialkowski, J.R. Wagner, P.C. Dedon,
P. Møller, M.M. Greenberg, M.S. Cooke, Biologically relevant oxidants and terminology, classification and nomenclature of oxidatively generated damage
to nucleobases and 2-deoxyribose in nucleic acids, Free Radic. Res. 46 (2012)
367–381.
[147] Y.A. Lee, B.H. Yun, S.K. Kim, Y. Margolin, P.C. Dedon, N.E. Geacintov, V.
Shafirovich, Mechanisms of oxidation of guanine in DNA by carbonate radical
anion, a decomposition product of nitrosoperoxycarbonate, Chem. Eur. J. 13
(2007) 4571–4581.
[148] J. Cadet, J.L. Ravanat, M. TavernaPorro, H. Menoni, D. Angelov, Oxidatively
generated complex DNA damage: tandem and clustered lesions, Cancer Lett.
327 (2012) 5–15.
[149] R. Brem, P. Karran, Multiple forms of DNA damage caused by UVA photoactivation of DNA 6-thioguanine, Photochem. Photobiol. 88 (2012) 5–13.
[150] T. Masuda, H. Shinohara, M. Kondo, Reactions of hydroxyl radicals with nucleic
acid bases and the related compounds in gamma irradiated aqueous solution,
J. Radiat. Res. 16 (1975) 153–161.
[151] C. von Sonntag, Free-Radical-Induced DNA Damage and its Repair. A Chemical
Perspective, Springer-Verlag, Berlin-Heidelberg-NewYork, 2006.
34
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
[152] A. Grand, C. Morell, V. Labet, J. Cadet, L.A. Eriksson, . H atom and . OH radical
reactions with 5-methylcytosine, J. Phys. Chem. A 111 (2007) 8968–8972.
[153] G.D. Castro, C.J. Stamato, J.A. Castro, 5-Methylcytosine attack by free radicals arising from bromotrichloromethane in a model system: Structures of
reaction products, Free Radic. Biol. Med. 17 (1994) 419–428.
[154] G.D. Castro, M.I. Díaz Gómez, J.A. Castro, 5-Methylcytosine attack by hydroxyl
free radicals and during carbon tetrachloride promoted liver microsomal lipid
peroxidation: Structure of reaction products, Chem.-Biol. Interact. 99 (1996)
289–299.
[155] N. Murata-Kamiya, H. Kamiya, N. Karino, Y. Ueno, H. Kaji, A. Matsuda, H. Kasai,
Formation of 5-formyl-2 -deoxycytidine from 5-methyl-2 -deoxycytidine in
duplex DNA by Fenton-type reactions and ␥-irradiation, Nucleic Acids Res.
27 (1999) 4385–4390.
[156] A. Burdzy, K.T. Noyes, V. Valinluck, L.C. Sowers, Synthesis of stable-isotope
enriched 5-methylpyrimidines and their use as probes of base reactivity in
DNA, Nucleic Acids Res. 30 (2002) 4068–4074.
[157] H. Cao, Y. Wang, Quantification of oxidative single-base and intrastrand crosslink lesions in unmethylated and CpG-methylated DNA induced by Fentontype reagents, Nucleic Acids Res. 35 (2007) 4833–4844.
[158] J. Shao, C.-H. Huang, B. Kalyanaraman, B.-Z. Zhu, Potent methyl oxidation of
5-methyl-2 -deoxycytidine by halogenated quinoid carcinogens and hydrogen peroxide via a metal-independent mechanism, Free Radic. Biol. Med. 60
(2013) 177–182.
[159] C. Bienvenu, J.R. Wagner, J. Cadet, Photosensitized oxidation of 5-methyl2 -deoxycytidine by 2-methyl-1,4-naphthoquinone - characterization of 5(hydroperoxymethyl)-2 -deoxycytidine and stable methyl group oxidation
products, J. Am. Chem. Soc. 118 (1996) 11406–11411.
[160] Q. Zhang, Y. Wang, Independent generation of 5-(2 -deoxycytidinyl)methyl
radical and the formation of a novel cross-link lesion between 5methylcytosine and guanine, J. Am. Chem. Soc. 125 (2003) 12795–12802.
[161] Q.Q. Zhang, Y. Wang, Generation of 5-(2 -deoxycytidyl)methyl radical and
the formation of intrastrand cross-link lesions in oligodeoxyribonucleotides,
Nucleic Acids Res. 33 (2005) 1593–1603.
[162] F.M. Prado, M.C.B. Oliveira, S. Miyamoto, G.R. Martinez, M.H.G. Medeiros, G.E.
Ronsein, P. Di Mascio, Thymine hydroperoxide as a potential source of singlet
molecular oxygen in DNA, Free Radic. Biol. Med. 47 (2009) 401–409.
[163] H. Yamada, K. Tanabe, T. Ito, S.I. Nishimoto, The pH effect on the
naphthoquinone-photosensitized oxidation of 5-methylcytosine, Chem. Eur.
J. 14 (2008) 10453–10461.
[164] S. Kanvah, G.B. Schuster, One-electron oxidation of DNA: the effect of replacement of cytosine with 5-methylcytosine on long-distance radical cation
transport and reaction, J. Am. Chem. Soc. 126 (2004) 7341–7344.
[165] E. Privat, L.C. Sowers, Photochemical deamination and demethylation of 5methylcytosine, Chem. Res. Toxicol. 9 (1996) 745–750.
[166] C. Bienvenu, J. Cadet, Synthesis and kinetic study of the deamination of the
cis diastereomers of 5,5-dihydroxy-5,6-dihydro-5-methyl-2 -deoxycytidine,
J. Org. Chem. 61 (1996) 2632–2637.
[167] C. Bienvenu, C. Lebrun, J. Cadet, Photo-sensitized formation of the cis-(5R,6R)
diastereomer of 5,6-dihydroxy-5,6-dihydro-5-methyl-2 -deoxycytidine, J.
Chim. Phys. Phys. -Chim. Biol. 94 (1997) 300–305.
[168] H. Cao, Y. Jiang, Y. Wang, Kinetics of deamination and Cu(II)/H2 O2 /ascorbateinduced formation of 5-methylcytosine glycol at CpG sites in duplex DNA,
Nucleic Acids Res. 37 (2009) 6635–6643.
[169] C.J. Lafrançois, J. Fujimoto, L.C. Sowers, Synthesis and characterization of isotopically enriched pyrimidine deoxynucleoside oxidation damage products,
Chem. Res. Toxicol. 11 (1998) 75–83.
[170] K.J. Divakar, C.B. Reese, 4-(1,2,4-triazol-1-yl)- And 4-(3-nitro-1,2,4-triazol-15-tri-O-acetylarabinofuranosyl)pyrimidin-2(1H)-ones.
yl)-1-(␤-D-2,3,
Valuable intermediates in the synthesis of derivatives of 1-(␤-darabinofuranosyl)cytosine (Ara-C), J. Chem. Soc., Perkin Trans. 1 (1982)
1171–1176.
[171] J.L. Herring, D.K. Rogstad, L.C. Sowers, Enzymatic methylation of DNA in
cultured human cells studied by stable isotope incorporation and mass spectrometry, Chem. Res. Toxicol. 22 (2009) 1060–1068.
[172] R.E. Cline, R.M. Fink, K. Fink, Synthesis of 5-substituted pyrimidines via
formaldehyde addition, J. Am. Chem. Soc. 81 (1959) 2521–2527.
[173] F. Maley, Nucleotide interconversions. VI. The enzymic formation of 5methyluridylic acid, Arch. Biochem. Biophys. 96 (1962) 550–556.
[174] A.A.H. Abdel-Rahman, E.S.H. El Ashry, Efficient synthesis of 5-hydroxymethyl
pyrimidines and their nucleosides using microwave irradiation, Synlett
(2002) 2043–2044.
[175] S. Schiesser, B. Hackner, T. Pfaffeneder, M. Müller, C. Hagemeier, M. Truss,
T. Carell, Mechanism and stem-cell activity of 5-carboxycytosine decarboxylation determined by isotope tracing, Angew. Chem. Int. Ed. 51 (2012)
6516–6520.
[176] A.A.H. Abdel Rahman, T. Wada, K. Saigo, Facile methods for the synthesis of
5-formylcytidine, Tetrahedron Lett. 42 (2001) 1061–1063.
[177] M. Munzel, D. Globisch, C. Trindler, T. Carell, Efficient synthesis of 5hydroxymethylcytosine containing DNA, Org. Lett. 12 (2010) 5671–5673.
[178] Q. Dai, C.X. Song, T. Pan, C. He, Syntheses of two 5-hydroxymethyl2 -deoxycytidine phosphoramidites with TBDMS as the 5-hydroxymethyl
protecting group and their incorporation into DNA, J. Org. Chem. 76 (2011)
4182–4188.
[179] Q. Dai, C. He, Syntheses of 5-formyl- and 5-carboxyl-dc containing DNA oligos
as potential oxidation products of 5-hydroxymethylcytosine in DNA, Org. Lett.
13 (2011) 3446–3449.
[180] M. Münzel, U. Lischke, D. Stathis, T. Pfaffeneder, F.A. Gnerlich, C.A. Deiml,
S.C. Koch, K. Karaghiosoff, T. Carell, Improved synthesis and mutagenicity of
oligonucleotides containing 5-hydroxymethylcytosine, 5-formylcytosine and
5-carboxylcytosine, Chem. Eur. J. 17 (2011) 13782–13788.
[181] H. Hashimoto, S. Hong, A.S. Bhagwat, X. Zhang, X. Cheng, Excision of
5-hydroxymethyluracil and 5-carboxylcytosine by the thymine DNA glycosylase domain: Its structural basis and implications for active DNA
demethylation, Nucleic Acids Res. 40 (2012) 10203–10214.
[182] W. Mao, J. Hu, T. Hong, X. Xing, S. Wang, X. Chen, X. Zhou, A convenient method
for selective detection of 5-hydroxymethylcytosine and 5-formylcytosine
sites in DNA sequences, Org. Biomol. Chem. 11 (2013) 3568–3572.
[183] H. Sugiyama, S. Matsuda, K. Kino, Q.M. Zhang, S. Yonei, I. Saito, New synthetic
method of 5-formyluracil-containing oligonucleotides and their melting
behavior, Tetrahedron Lett. 37 (1996) 9067–9070.
[184] N. Karino, Y. Ueno, A. Matsuda, Synthesis and properties of oligonucleotides
containing 5-formyl-2 -deoxycytidine: In vitro DNA polymerase reactions on
DNA templates containing 5-formyl-2 -deoxycytidine, Nucleic Acids Res. 29
(2001) 2456–2463.
[185] B. Steigenberger, S. Schiesser, B. Hackner, C. Brandmayr, S.K. Laube, J. Steinbacher, T. Pfaffeneder, T. Carell, Synthesis of 5-hydroxymethyl-, 5-formyl-,
and 5-carboxycytidine- triphosphates and their incorporation into oligonucleotides by polymerase chain reaction, Org. Lett. 15 (2013) 366–369.
[186] S. Mohamady, A. Desoky, S.D. Taylor, Sulfonyl imidazolium salts as reagents
for the rapid and efficient synthesis of nucleoside polyphosphates and their
conjugates, Org. Lett. 14 (2012) 402–405.
[187] C.S. Nabel, S.A. Manning, R.M. Kohli, The curious chemical biology of cytosine:
Deamination, methylation,and oxidation as modulators of genomic potential,
ACS Chem. Biol. 7 (2012) 20–30.
[188] J.C. Shen, W.M. Rideout, P.A. Jones, The rate of hydrolytic deamination of
5-methylcytosine in double-stranded DNA, Nucleic Acids Res. 22 (1994)
972–976.
[189] V. Labet, C. Morell, J. Cadet, L.A. Eriksson, A. Grand, Hydrolytic deamination of
5-methylcytosine in protic medium-A theoretical study, J. Phys. Chem. A 113
(2009) 2524–2533.
[190] R.T. Williams, Y. Wang, A density functional theory study on the kinetics
and thermodynamics of N-glycosidic bond cleavage in 5-substituted 2 deoxycytidines, Biochemistry 51 (2012) 6458–6462.
[191] A. Inoue, Y. Zhang, Replication-dependent loss of 5-hydroxymethylcytosine
in mouse preimplantation embryos, Science 334 (2011) 194.
[192] V. Valinluck, L.C. Sowers, Endogenous cytosine damage products alter the site
selectivity of human DNA maintenance methyltransferase DNMT1, Cancer
Res. 67 (2007) 946–950.
[193] K.M. Schmitz, S.K. Petersen-Mahrt, AIDing the immune system-DIAbolic in
cancer, Semin. Immunol. 24 (2012) 241–245.
[194] C. Popp, W. Dean, S.H. Feng, S.J. Cokus, S. Andrews, M. Pellegrini, S.E. Jacobsen,
W. Reik, Genome-wide erasure of DNA methylation in mouse primordial germ
cells is affected by AID deficiency, Nature 463 (2010) 1101–U1126.
[195] C.S. Nabel, H. Jia, Y. Ye, L. Shen, H.L. Goldschmidt, J.T. Stivers, Y. Zhang, R.M.
Kohli, AID/APOBEC deaminases disfavor modified cytosines implicated in
DNA demethylation, Nat. Chem. Biol. 8 (2012) 751–758.
[196] H. Hashimoto, X. Zhang, X. Cheng, Excision of thymine and 5hydroxymethyluracil by the MBD4 DNA glycosylase domain: Structural basis
and implications for active DNA demethylation, Nucleic Acids Res. 40 (2012)
8276–8284.
[197] S. Moréra, I. Grin, A. Vigouroux, S. Couvé, V. Henriot, M. Saparbaev, A.A.
Ishchenko, Biochemical and structural characterization of the glycosylase
domain of MBD4 bound to thymine and 5-hydroxymethyuracil-containing
DNA, Nucleic Acids Res. 40 (2012) 9917–9926.
[198] S. Cortellino, J. Xu, M. Sannai, R. Moore, E. Caretti, A. Cigliano, M. Le Coz,
K. Devarajan, A. Wessels, D. Soprano, L.K. Abramowitz, M.S. Bartolomei, F.
Rambow, M.R. Bassi, T. Bruno, M. Fanciulli, C. Renner, A.J. Klein-Szanto, Y. Matsumoto, D. Kobi, I. Davidson, C. Alberti, L. Larue, A. Bellacosa, Thymine DNA
glycosylase is essential for active DNA demethylation by linked deaminationbase excision repair, Cell 146 (2011) 67–79.
[199] H. Hashimoto, X. Zhang, X. Cheng, Selective excision of 5-carboxylcytosine by
a thymine DNA glycosylase mutant, J. Mol. Biol. 425 (2013) 971–976.
[200] L. Shen, H. Wu, D. Diep, S. Yamaguchi, A.C. D‘Alessio, H.L. Fung, K. Zhang,
Y. Zhang, Genome-wide analysis reveals TET- and TDG-dependent 5methylcytosine oxidation dynamics, Cell 153 (2013) 692–706.
[201] P. Zhang, B. Huang, X. Xu, W.C. Sessa, Ten-eleven translocation (Tet) and
thymine DNA glycosylase (TDG), components of the demethylation pathway,
are direct targets of miRNA-29a, Biochem, Biophys. Res. Commun. 437 (2013)
368–373.
[202] R. Métivier, R. Gallais, C. Tiffoche, C. Le Péron, R.Z. Jurkowska, R.P. Carmouche,
D. Ibberson, P. Barath, F. Demay, G. Reid, V. Benes, A. Jeltsch, F. Gannon, G. Salbert, Cyclical DNA methylation of a transcriptionally active promoter, Nature
452 (2008) 45–50.
[203] S. Hamm, G. Just, N. Lacoste, N. Moitessier, M. Szyf, O. Mamer, On the mechanism of demethylation of 5-methylcytosine in DNA, Bioorg. Med. Chem. Lett.
18 (2008) 1046–1049.
[204] C.C. Chen, K.Y. Wang, C.K.J. Shen, The mammalian de novo DNA methyltransferases DNMT3A and DNMT3B are also DNA 5-hydroxymethylcytosine
dehydroxymethylases, J. Biol. Chem. 287 (2012) 33116–33121.
[205] C.C. Chen, K.Y. Wang, C.K.J. Shen, DNA 5-methylcytosine demethylation activities of the mammalian DNA methyltransferases, J. Biol. Chem. 288 (2013)
9084–9091.
J. Cadet, J.R. Wagner / Mutation Research 764–765 (2014) 18–35
[206] H. Kamiya, H. Tsuchiya, N. Karino, Y. Ueno, A. Matsuda, H. Harashima, Mutagenicity of 5-formylcytosine, an oxidation product of 5-methylcytosine, in
DNA in mammalian cells, J. Biochem. 132 (2002) 551–555.
[207] B. Yuan, J. Wang, H. Cao, R. Sun, Y. Wang, High-throughput analysis of
the mutagenic and cytotoxic properties of DNA lesions by next-generation
sequencing, Nucleic Acids Res. 39 (2011) 5945–5954.
[208] M.W. Kellinger, C.X. Song, J. Chong, X.Y. Lu, C. He, D. Wang, 5-Formylcytosine
and 5-carboxylcytosine reduce the rate and substrate specificity of RNA polymerase II transcription, Nat. Struct. Mol. Biol. 19 (2012) 831–833.
[209] G. Foltz, J.G. Yoon, H. Lee, T.C. Ryken, Z. Sibenaller, M. Ehrich, L. Hood, A.
Madan, DNA methyltransferase-mediated transcriptional silencing in malignant glioma: a combined whole-genome microarray and promoter array
analysis, Oncogene 28 (2009) 2667–2677.
[210] F. Gaudet, J.G. Hodgson, A. Eden, L. Jackson-Grusby, J. Dausman, J.W. Gray, H.
Leonhardt, R. Jaenisch, Induction of tumors in mice by genomic hypomethylation, Science 300 (2003) 489–492.
[211] M. Ehrlich, DNA hypomethylation in cancer cells, Epigenomics 1 (2009)
239–259.
[212] M.E. Hegi, D. Sciuscio, A. Murat, M. Levivier, R. Stupp, Epigenetic deregulation of DNA repair and its potential for therapy, Clin. Cancer. Res. 15 (2009)
5026–5031.
[213] M. Weller, R. Stupp, G. Reifenberger, A.A. Brandes, M.J. van den Bent, W.
Wick, M.E. Hegi, MGMT promoter methylation in malignant gliomas: ready
for personalized medicine? Nat. Rev. Neurol. 6 (2010) 39–51.
[214] C. Liu, L. Liu, X. Chen, J. Shen, J. Shan, Y. Xu, Z. Yang, L. Wu, F. Xia, P. Bie, Y. Cui,
X.W. Bian, C. Qian, Decrease of 5-hydroxymethylcytosine is associated with
progression of hepatocellular carcinoma through downregulation of TET1,
PLoS One 8 (2013) e62828.
[215] T. Gambichler, M. Sand, M. Skrygan, Loss of 5-hydroxymethylcytosine and
ten-eleven translocation 2 protein expression in malignant melanoma,
Melanoma Res. 23 (2013) 218–220.
[216] H. Yang, Y. Liu, F. Bai, J.Y. Zhang, S.H. Ma, J. Liu, Z.D. Xu, H.G. Zhu, Z.Q. Ling,
D. Ye, K.L. Guan, Y. Xiong, Tumor development is associated with decrease
of TET gene expression and 5-methylcytosine hydroxylation, Oncogene 32
(2013) 663–669.
[217] Y. Kudo, K. Tateishi, K. Yamamoto, S. Yamamoto, Y. Asaoka, H. Ijichi, G.
Nagae, H. Yoshida, H. Aburatani, K. Koike, Loss of 5-hydroxymethylcytosine is
accompanied with malignant cellular transformation, Cancer Sci. 103 (2012)
670–676.
[218] Q. Yang, K. Wu, M. Ji, W. Jin, N. He, B. Shi, P. Hou, Decreased 5-hydroxymethylcytosine (5-hmC) is an independent poor pronostic factor in gastric
cancer patients, J. Biomed. Nanotechnol. 9 (2013) 1607–1616.
[219] S.G. Jin, Y. Jiang, R. Qiu, T.A. Rauch, Y. Wang, G. Schackert, D. Krex, Q. Lu,
G.P. Pfeifer, 5-Hydroxymethylcytosine is strongly depleted in human cancers
but its levels do not correlate with IDH1 mutations, Cancer Res. 71 (2011)
7360–7365.
[220] E.F. Mason, J.L. Hornick, Succinate dehydrogenase deficiency is associated
with decreased 5-hydroxymethylcytosine production in gastrointestinal
stromal tumors: implications for mechanisms of tumorigenesis, Mod. Pathol.
(2013), http://dx.doi.org/10.1038/modpathol.2013.86 [Epub ahead of print].
[221] M.L. Chen, F. Shen, Quantification of 5-methylcytosine and 5hydroxymethylcytosine in genomic DNA from hepatocellular carcinoma
tissues by capillary hydrophilic-interaction liquid chromatography/quadrupole time-of-flight mass spectrometry, Clin. Chem. 59 (2013)
824–832.
[222] B. Cadieux, T.T. Ching, S.R. VandenBerg, J.F. Costello, Genome-wide hypomethylation in human glioblastomas associated with specific copy number
alteration, methylenetetrahydrofolate reductase allele status, and increased
proliferation, Cancer Res. 66 (2006) 8469–8476.
[223] E. Hervouet, E. Debien, L. Campion, J. Charbord, J. Menanteau, F.M. Vallette,
P.F. Cartron, Folate supplementation limits the aggressiveness of glioma via
the remethylation of DNA repeats element and genes governing apoptosis
and proliferation, Clin. Cancer. Res. 15 (2009) 3519–3529.
35
[224] F. Mohr, K. Döhner, C. Buske, V.P.S. Rawat, T.E.T Genes, New players in DNA
demethylation and important determinants for stemness, Exp. Hematol. 39
(2011) 272–281.
[225] S.M.C. Langemeijer, M.G. Aslanyan, J.H. Jansen, TET proteins in malignant
hematopoiesis, Cell Cycle 8 (2009) 4044–4048.
[226] M.E. Figueroa, O. Abdel-Wahab, C. Lu, P.S. Ward, J. Patel, A. Shih, Y. Li, N.
Bhagwat, A. Vasanthakumar, H.F. Fernandez, M.S. Tallman, Z. Sun, K. Wolniak,
J.K. Peeters, W. Liu, S.E. Choe, V.R. Fantin, E. Paietta, B. Löwenberg, J.D. Licht,
L.A. Godley, R. Delwel, P.J.M. Valk, C.B. Thompson, R.L. Levine, A. Melnick,
Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype,
disrupt TET2 function, and impair hematopoietic differentiation, Cancer Cell
18 (2010) 553–567.
[227] K. Moran-Crusio, L. Reavie, A. Shih, O. Abdel-Wahab, D. Ndiaye-Lobry, C. Lobry,
M. Figueroa, A. Vasanthakumar, J. Patel, X. Zhao, F. Perna, S. Pandey, J. Madzo,
C. Song, Q. Dai, C. He, S. Ibrahim, M. Beran, J. Zavadil, S. Nimer, A. Melnick,
L. Godley, I. Aifantis, R. Levine, Tet2 loss leads to increased hematopoietic
stem cell self-renewal and myeloid transformation, Cancer Cell 20 (2011)
11–24.
[228] C. Quivoron, L. Couronné, V. Della Valle, C. Lopez, I. Plo, O. Wagner-Ballon,
M. Do Cruzeiro, F. Delhommeau, B. Arnulf, M.H. Stern, L. Godley, P. Opolon,
H. Tilly, E. Solary, Y. Duffourd, P. Dessen, H. Merle-Beral, F. Nguyen-Khac, M.
Fontenay, W. Vainchenker, C. Bastard, T. Mercher, O. Bernard, TET2 inactivation results in pleiotropic hematopoietic abnormalities in mouse and is
a recurrent event during human lymphomagenesis, Cancer Cell 20 (2011)
25–38.
[229] T. Shibata, A. Kokubu, M. Miyamoto, Y. Sasajima, N. Yamazaki, Mutant IDH1
confers an in vivo growth in a melanoma cell line with BRAF mutation, Am. J.
Pathol. 178 (2011) 1395–1402.
[230] Y. Liu, W. Jiang, J. Liu, S. Zhao, J. Xiong, Y. Mao, Y. Wang, IDH1 mutations inhibit
multiple ␣-ketoglutarate-dependent dioxygenase activities in astroglioma, J.
Neurooncol. 109 (2012) 253–260.
[231] P. Wang, Q. Dong, C. Zhang, P.F. Kuan, Y. Liu, W.R. Jeck, J.B. Andersen, W. Jiang,
G.L. Savich, T.X. Tan, J.T. Auman, J.M. Hoskins, A.D. Misher, C.D. Moser, S.M.
Yourstone, J.W. Kim, K. Cibulskis, G. Getz, H.V. Hunt, S.S. Thorgeirsson, L.R.
Roberts, D. Ye, K.L. Guan, Y. Xiong, L.X. Qin, D.Y. Chiang, Mutations in isocitrate
dehydrogenase 1 and 2 occur frequently in intrahepatic cholangiocarcinomas
and share hypermethylation targets with glioblastomas, Oncogene 32 (2013)
3091–3100.
[232] G.P. Pfeifer, S. Kadam, S.G. Jin, 5-hydroxymethylcytosine and its potential
roles in development and cancer, Epigenetics Chromatin 6 (2013) 10.
[233] J.P. Patel, M. Gönen, M.E. Figueroa, H. Fernandez, Z. Sun, J. Racevskis,
P. Van Vlierberghe, I. Dolgalev, S. Thomas, O. Aminova, K. Huberman,
J. Cheng, A. Viale, N.D. Socci, A. Heguy, A. Cherry, G. Vance, R.R. Higgins, R.P. Ketterling, R.E. Gallagher, M. Litzow, M.R.M. Van Den Brink,
H.M. Lazarus, J.M. Rowe, S. Luger, A. Ferrando, E. Paietta, M.S. Tallman, A.
Melnick, O. Abdel-Wahab, R.L. Levine, Prognostic relevance of integrated
genetic profiling in acute myeloid leukemia, N. Engl. J. Med. 366 (2012)
1079–1089.
[234] H. Huang, X. Jiang, Z. Li, Y. Li, C.X. Song, C. He, M. Sun, P. Chen, S. Gurbuxani, J. Wang, G.M. Hong, A.G. Elkahloun, S. Arnovitz, K. Szulwach, L. Lin, C.
Street, M. Wunderlich, M. Dawlaty, M.B. Neilly, R. Jaenisch, F.C. Yang, J.C.
Mulloy, P. Jin, P.P. Liu, J.D. Rowley, M. Xu, J. Chen, TET1 plays an essential
oncogenic role in MLL-rearranged leukemia, Proc. Natl. Acad. Sci. U.S.A. 110
(2013) 11994–11999.
[235] E. Olkhov-Mitsel, B. Bapat, Strategies for discovery and validation of methylated and hydroxymethylated DNA biomarkers, Cancer Med. 1 (2012)
237–260.
[236] S. Fukushige, A. Horii, DNA methylation in cancer: a gene silencing mechanism
and the clinical potential of its biomarkers, Tohoku J. Exp. Med. 229 (2013)
173–185.
[237] F. Samson-Thibault, G.S. Madugundu, S. Gao, J. Cadet, J.R. Wagner, Profiling cytosine oxidation in DNA by LC-MS/MS, Chem. Res. Toxicol. 25 (2012)
1902–1911.