Quantitative LC–MS Provides No Evidence for m6dA or m4dC in the Genome of Mouse Embryonic Stem Cells and Tissues

Article abstract of DOI:10.1002/anie.201700424

Until recently, it was believed that the genomes of higher organisms contain, in addition to the four canonical DNA bases, only 5-methyl-dC (m⁵dC) as a modified base to control epigenetic processes. In recent years, this view has changed dramatically with the discovery of 5-hydroxymethyl-dC (hmdC), 5-formyl-dC (fdC), and 5-carboxy-dC (cadC) in DNA from stem cells and brain tissue. N⁶-methyldeoxyadenosine (m⁶dA) is the most recent base reported to be present in the genome of various eukaryotic organisms. This base, together with N⁴-methyldeoxycytidine (m⁴dC), was first reported to be a component of bacterial genomes. In this work, we investigated the levels and distribution of these potentially epigenetically relevant DNA bases by using a novel ultrasensitive UHPLC–MS method. We further report quantitative data for m⁵dC, hmdC, fdC, and cadC, but we were unable to detect either m⁴dC or m⁶dA in DNA isolated from mouse embryonic stem cells or brain and liver tissue, which calls into question their epigenetic relevance.

Full text of DOI:10.1002/anie.201700424

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201700424
German Edition: DOI: 10.1002/ange.201700424
Epigenetics
Quantitative LC–MS Provides No Evidence for m6dA or m4dC in the
Genome of Mouse Embryonic Stem Cells and Tissues
Sarah Schiffers, Charlotte Ebert, Renꢀ Rahimoff, Olesea Kosmatchev, Jessica Steinbacher,
Alexandra-Viola Bohne, Fabio Spada, Stylianos Michalakis, Jçrg Nickelsen, Markus Mꢁller, and
Thomas Carell*
Abstract: Until recently, it was believed that the genomes of
higher organisms contain, in addition to the four canonical
DNA bases, only 5-methyl-dC (m5dC) as a modified base to
control epigenetic processes. In recent years, this view has
changed dramatically with the discovery of 5-hydroxymethyl-
dC (hmdC), 5-formyl-dC (fdC), and 5-carboxy-dC (cadC) in
DNA from stem cells and brain tissue. N6-methyldeoxyadeno-
sine (m6dA) is the most recent base reported to be present in the
genome of various eukaryotic organisms. This base, together
with N4-methyldeoxycytidine (m4dC), was first reported to be
a component of bacterial genomes. In this work, we inves-
tigated the levels and distribution of these potentially epigeneti-
cally relevant DNA bases by using a novel ultrasensitive
UHPLC–MS method. We further report quantitative data for
m5dC, hmdC, fdC, and cadC, but we were unable to detect
either m4dC or m6dA in DNA isolated from mouse embryonic
Figure 1. Depiction of the four canonical DNA bases and the epige-
stem cells or brain and liver tissue, which calls into question
their epigenetic relevance.
netic DNA marks m5dC, hmdC, fdC, and cadC, as well as the bases
m6dA and m4dC together with the synthesized isotopologues.
T
he genetic material of living organisms is constructed from
the four canonical nucleobases dA, dC, dG, and dT, which
establish the sequence information that, in multicellular
organisms, is stored in the nucleus of every cell (Figure 1).
In addition to the canonical bases, the methylated dC base 5-
methyldeoxycytidine (m5dC) is frequently found.[1] The
presence or absence of this base in specific promoter seg-
ments determines whether the gene is actively transcribed or
silenced.[1] The cell-type-specific distribution of m5dC thus
determines the identity of a given cell. Recently, 5-hydrox-
ymethyldeoxycytidine (hmdC) was found as a sixth base of
the genetic system[2,3] and in 2011, 5-formyldeoxycytidine
(fdC)[4,5] and 5-carboxydeoxycytidine (cadC)[5,6] were also
discovered, particularly in DNA isolated from stem cells, but
also in brain DNA. It is currently believed that fdC and cadC
are intermediates in an active demethylation process that
allows cells to change the methylation pattern and hence the
activity state of specific genes.[7,8] For fdC, separate epigenetic
functions are also envisaged.[9]
While the genomes of bacteria are known to also contain
N4-methyldeoxycytidine (m4dC)[10] and N6-methyldeoxyade-
nosine (m6dA),[11] attempts to detect these bases in the DNA
of higher organisms have failed until recently.[12–15] m6dA has
now been found in algae (0.4 mol% m6dA/A),[12] fruit flies
(0.001%-0.07% m6dA/A),[14] and C. elegans (0.01%-0.4%
m6dA/A),[13] and its presence has even been reported in the
DNA of vertebrates (0.00009% in X. laevis[16] and 0.00019–
0.003% of dA in murine cells and tissue[17]). These discov-
eries, especially concerning the DNA of vertebrates, have
spurred a worldwide research interest in unraveling the
function of these new bases in human genomic DNA.[18–20]
In this study, we developed an ultrasensitive triple
quadrupole mass spectrometry (QQQ-MS) method, which
in combination with ultra-high-pressure chromatography
(UHPLC) enables m4dC and m6dA to be searched for and
quantified in parallel to the more established new epigenetic
DNA marks m5dC, hmdC, fdC and cadC.
[*] S. Schiffers, C. Ebert, R. Rahimoff, O. Kosmatchev, J. Steinbacher,
Dr. F. Spada, Dr. M. Mꢀller, Prof. Dr. T. Carell
Center for Integrated Protein Science (CiPSM) at the Department of
Chemistry, LMU Mꢀnchen
Butenandtstr. 5–13, 81377 Mꢀnchen (Germany)
E-mail: Thomas.Carell@lmu.de
Homepage: http://www.carellgroup.de
Dr. A.-V. Bohne, Prof. Dr. J. Nickelsen
Biocenter of the LMU
Dept. Biologie 1 – Botanik, Molekulare Pflanzenwissenschaften
Grosshaderner Strasse 2–4, 82152 Planegg-Martinsried (Germany)
Dr. S. Michalakis
CiPSM, Department of Pharmacy, LMU Mꢀnchen
Pharmacology for Life Sciences
Butenandtstr. 7, 81377 Mꢀnchen (Germany)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
http://dx.doi.org/10.1002/anie.201700424.
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
For the quantitative measurements, we first chemically
synthesized the two isotopologues of m6dA and m4dC shown
in Figure 1 as internal standards for the analytical method.
The prepared compounds D3-m6dA and 15N2-m4dC are three
and two mass units heavier, respectively, than the natural
bases. Despite these molecular-weight differences, they have
identical properties during the UHPL chromatography step
so that they strictly coelute with their natural counterparts,
thus allowing them to enter the mass spectrometer at exactly
the same time as the internal standards. The availability of
these isotopologues makes the method highly reliable and
strictly quantitative. The syntheses of the two compounds,
together with all analytical data, are given in the Supporting
Information.
We first benchmarked our study with an investigation of
genomic DNA isolated from the unicellular green algae
Chlamydomonas reinhardtii and the cyanobacterium Syne-
chocystis. In both cases, DNAwas isolated after cell lysis using
a standard method (see the Supporting Information). The
isolated DNA was subsequently digested with a mixture of
three commercially available digestion enzymes (Nuclease
S1, Antarctic Phosphatase, and Snake Venom Phosphodies-
terase; see the Supporting Information). We next added the
isotope-labelled standards D3-m6dA and 15N2-m4dC to the
obtained digestion mixture and performed UHPLC-QQQ
analysis. For the mass spectrometry detection, we selected
fragmentation of the glycosidic bond as the indicative and
hence recorded mass transition. This is m/z = 266.12!150.08
for m6dA and m/z = 269.14!153.10 for its isotopologue D3-
m6dA. For m4dC, we also used fragmentation of the glycosidic
bond, which gives a mass transition of m/z = 242.11!126.07
for the natural compound m4dC and m/z = 244.11!128.07 for
its isotopologue 15N2-m4dC (Figure 2A).
Figure 2. A) Fragmentation patterns of m4dC and m6dA. B–D) Quanti-
tative data of the bases m4dC and m6dA in Synechocystis (B), Chlamy-
domonas (C) and of these bases and the other epigenetic DNA marks
hmdC, fdC, cadC, and m5dC in mouse embryonic stem cells (D).
material of Chlamydomonas. This is interesting because
Synechocystis is considered a relative of the chloroplasts
present in Chlamydomonas.
We next modified the reported UHPLC-QQQ method[10]
for the simultaneous quantification of m4dC and m6dA,
together with the other epigenetically relevant bases m5dC,
hmdC, fdC, and cadC. To this end, the UHPLC gradient was
fine-tuned to enable full separation of all six compounds.
Finally, we measured precise calibration curves for all of the
compounds (see Figures S1 and S2 in the Supporting Infor-
mation). This subsequently allowed exact quantification of all
of the discussed epigenetic DNA marks in a given sample
(Figure 2B–D).
With these positive results in hand, we extended our study
to mouse embryonic stem cells (ESCs; wt J1, Figure 2D).
m6dA in particular was recently reported to occur as an
epigenetically relevant DNA mark in mouse ESCs (mESC
cell line wt TT2).[17] When performing the measurements, we
turned the mass spectrometer to maximum sensitivity. But
even in this mode, we were unable to detect a signal for m6dA
within the detection limits of our system Table S2. In contrast,
the other epigenetically relevant bases hmdC, fdC, cadC, and
even the oxidative lesion 8-oxodG, which we also quantified
in parallel, were clearly detectable. The 8-oxodG level was
4.8 ꢀ 10ꢀ5 per dN. The rare and difficult to detect cadC was
clearly seen even at levels of only 9.0 ꢀ 10ꢀ8 per dN. For m6dA,
in contrast, a signal did not appear. We also re-measured wt
TT2 cells as described and still did not detect m6dA over
background levels (see Figure S4).[17] To obtain unequivocal
proof that m6dA is not present in stem cells, we added 13CD3-
methionine to the mESC culture. Methionine provides the
methyl group for the biosynthesis of m6dA. With 13CD3-
methionine, this would lead to an m/z-shift of + 4. We tuned
the mass spectrometer to the new m/z-transition and again
were unable to see any signal for 13CD3-m6dA (Figure S8).
We subsequently turned our attention to adult mouse
tissue and analysed DNA isolated from liver and whole brain
Since m4dC and m6dA are well known in bacteria, we first
analysed the cyanobacterium Synechocystis (PCC6803), and
we indeed found both bases (Figure 2B). The base m6dA was
detected at a level of 8.4 ꢀ 10ꢀ3 per dN and for m4dC we
measured a value of 5.9 ꢀ 10ꢀ3 per dN. The constitutional
isomer m5dC and all other dC-derived epigenetic DNA marks
were detectable, but were not quantified in this experiment.
Next, we analyzed two different strains of Chlamydomo-
nas reinhardtii (CC-3491 and wt 7d +), in which m6dA has just
recently been discovered,[12] and the levels of m6dA were
determined to be 8.4 ꢀ 10ꢀ4 per dN for CC-3491 and 6.9 ꢀ 10ꢀ4
per dN for wt 7d + (Figure 2C). This corresponds to about
3000 m6dA bases per Chlamydomonas genome (genome size
1.2 ꢀ 108), which at 0.7% of the dA is a relatively high
number. In both strains, m4dC was not detected, thus showing
that this base is unlikely to be a component of the genetic
2
www.angewandte.org
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
(Figure S3) using our UHPLC-QQQ method. Figure 3 shows
The m6dA data are highly interesting (Figure 3 right
the data obtained from mouse liver. The middle column
shows the data we obtained for hmdC. The already reported
D2-15N2-hmdC standard elutes at a retention time of 2.25 min
column). While the D3-m6dA standard was clearly detectable
at a retention time of 10.1 min, the unlabelled m6dA provided
a very weak signal. We then performed a control experiment
to determine the limit of detection
and investigated the digestion so-
lution alone, which contains all of
the commercial enzymes but no
isolated DNA (red chromato-
gram). A weak signal for m6dA
was again detectable at a retention
time of 10.1 min. After subtracting
this background signal (red) from
the
measured
chromatogram
(blue), we obtained the black line
showing that the original signal at
10.1 min is purely caused by back-
ground derived from the enzyme
mixture. Here it is important to
note that most of these proteins
are recombinant proteins obtained
from bacterial expression systems
and bacterial DNA contains plenty
of m6dA. To support the evidence
that m6dA is not present in verte-
brate DNA, we performed the
same experiment with HeLa cells
and also observed no signal for
m6dA (Figure 3B).
We then determined our limit
of detection for m6dA to be 3.5 ꢀ
10ꢀ7 per dN, which corresponds to
170 m6dA bases per murine
genome. This is
a very small
number and demonstrates the
excellent sensitivity of our
method. It shows that the maxi-
mum number of m6dA that could
be theoretically present and would
not be detectable by our method is
less than 170 m6dA bases per
genome, which led us to conclude
that m6dA is likely not epigeneti-
cally relevant but rather formed as
a DNA lesion, perhaps by mis-
guided methyltransferases. Spiking
tests with synthetic nucleoside and
Figure 3. A) Chromatograms of the mass signal of mouse liver DNA. UHPLC-QQQ data obtained for
m5dC and m4dC (left), hmdC (middle), and m6dA (right, blue line) and their corresponding isotopic
standards are shown. Additionally, for m6dA, the chromatogram for the digest blank is shown (red
line) and a computed baseline (black line), which was determined by subtracting the digest blank
from the sample. B) A chromatogram of the mass signal from UHPLC-QQQ data obtained for m6dA
in DNA from HeLa cells.
and shows the expected fragmentation of the glycosidic bond,
providing the fragmentation signature m/z = 262.12!146.07,
which allowed assignment of the signal. The naturally
occurring hmdC is detected at exactly the same retention
time with a mass transition of m/z = 258.11!142.06, thus
unequivocally demonstrating the presence of hmdC in murine
liver DNA. Regarding the different monomethylated dC
compounds m5dC and m4dC (left column), the epigenetic
DNA mark m5dC is clearly detected at a retention time of
3.2 min, but for m4dC with a retention time of 3.5 min, there is
obviously no signal present.
DNA from Chlamydomonas nevertheless confirmed the
sensitivity of the method, since the input amount equalled
the found amount (see Figure S6).
To find a potential source for m6dA in mESC DNA that
could explain previous sequencing data,[17] we thought that
m6dA-containing bacterial DNA that gets degraded could
provide the m6dA nucleoside, which then might get incorpo-
rated into mESC DNA. This is indeed a possibility. When we
added the m6dA nucleoside to a mESCs culture, we indeed
saw incorporation of some m6dA into the genome (Fig-
ure S7).
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
[7] R. M. Kohli, Y. Zhang, Nature 2013, 502, 472 – 479.
We also cannot fully exclude the possibility that the
presence of few m6dA bases, at levels below our detection
limit, could have a biological function. In addition, it is
possible that at certain stages of organismal development,
certain methyltransferases are activated that may induce high
m6dA levels at specific time points that may have escaped our
detection.[18] Our data, however, show clearly that the
maximum possible levels of m6dA in the analysed organisms
and mESCs under normal conditions are far lower than so far
believed.
[8] S. Schiesser, T. Pfaffeneder, K. Sadeghian, B. Hackner, B.
Steigenberger, A. S. Schrçder, J. Steinbacher, G. Kashiwazaki,
G. Hçfner, K. T. Wanner, C. Ochsenfeld, T. Carell, J. Am. Chem.
Soc. 2013, 135, 14593 – 14599.
[9] M. Su, A. Kirchner, S. Stazzoni, M. Mꢁller, M. Wagner, A.
Schrçder, T. Carell, Angew. Chem. Int. Ed. 2016, 55, 11797 –
11800; Angew. Chem. 2016, 128, 11974 – 11978.
[10] M. Yu, L. Ji, D. A. Neumann, D. H. Chung, J. Groom, J.
Westpheling, C. He, R. J. Schmitz, Nucleic Acids Res. 2015, 43,
e148.
[11] M. Ehrlich, M. A. Gama-Sosa, L. H. Carreira, L. G. Ljungdahl,
K. C. Kuo, C. W. Gehrke, Nucleic Acids Res. 1985, 13, 1399 –
1412.
Acknowledgements
[12] Y. Fu, G. Z. Luo, K. Chen, X. Deng, M. Yu, D. Han, Z. Hao, J.
Liu, X. Lu, L. C. Dore, X. Weng, Q. Ji, L. Mets, C. He, Cell 2015,
161, 879 – 892.
[13] E. L. Greer, M. A. Blanco, L. Gu, E. Sendinc, J. Liu, D.
Aristizabal-Corrales, C. H. Hsu, L. Aravind, C. He, Y. Shi, Cell
2015, 161, 868 – 878.
We thank the Deutsche Forschungsgemeinschaft for financial
support via SFB1032 (TP-A5), SFB749 (TP-A4), SPP1784,
CA275 and the Excellence Cluster CiPSM.
[14] G. Zhang, H. Huang, D. Liu, Y. Cheng, X. Liu, W. Zhang, R. Yin,
D. Zhang, P. Zhang, J. Liu, C. Li, B. Liu, Y. Luo, Y. Zhu, N.
Zhang, S. He, C. He, H. Wang, D. Chen, Cell 2015, 161, 893 – 906.
[15] S. Hattman, C. Kenny, L. Berger, K. Pratt, J. Bacteriol. 1978, 135,
1156 – 1157.
Conflict of interest
The authors declare no conflict of interest.
[16] M. J. Koziol, C. R. Bradshaw, G. E. Allen, A. S. Costa, C. Frezza,
J. B. Gurdon, Nat. Struct. Mol. Biol. 2016, 23, 24 – 30.
[17] T. P. Wu, T. Wang, M. G. Seetin, Y. Lai, S. Zhu, K. Lin, Y. Liu,
S. D. Byrum, S. G. Mackintosh, M. Zhong, A. Tackett, G. Wang,
L. S. Hon, G. Fang, J. A. Swenberg, A. Z. Xiao, Nature 2016, 532,
329 – 333.
Keywords: DNA · epigenetics · methyldeoxyadenosine ·
methyldeoxycytidine · mass spectrometry
[18] J. Liu, Y. Zhu, G. Z. Luo, X. Wang, Y. Yue, X. Wang, X. Zong, K.
Chen, H. Yin, Y. Fu, D. Han, Y. Wang, D. Chen, C. He, Nat.
Commun. 2016, 7, 13052.
[19] W. Huang, J. Xiong, Y. Yang, S.-M. Liu, B.-F. Yuan, Y.-Q. Feng,
RSC Adv. 2015, 5, 64046 – 64054.
[1] J. A. Law, S. E. Jacobsen, Nat. Rev. Genet. 2010, 11, 204 – 220.
[2] M. Tahiliani, K. P. Koh, Y. Shen, W. A. Pastor, H. Bandukwala,
Y. Brudno, S. Agarwal, L. M. Iyer, D. R. Liu, L. Aravind, A.
Rao, Science 2009, 324, 930 – 935.
[3] S. Kriaucionis, N. Heintz, Science 2009, 324, 929 – 930.
[4] T. Pfaffeneder, B. Hackner, M. Truss, M. Mꢁnzel, M. Mꢁller,
C. A. Deiml, C. Hagemeier, T. Carell, Angew. Chem. Int. Ed.
2011, 50, 7008 – 7012; Angew. Chem. 2011, 123, 7146 – 7150.
[5] S. Ito, L. Shen, Q. Dai, S. C. Wu, L. B. Collins, J. A. Swenberg, C.
He, Y. Zhang, Science 2011, 333, 1300 – 1303.
[20] G. Z. Luo, F. Wang, X. Weng, K. Chen, Z. Hao, M. Yu, X. Deng,
J. Liu, C. He, Nat. Commun. 2016, 7, 11301.
[6] Y. F. He, B. Z. Li, Z. Li, P. Liu, Y. Wang, Q. Tang, J. Ding, Y. Jia,
Z. Chen, L. Li, Y. Sun, X. Li, Q. Dai, C. X. Song, K. Zhang, C.
He, G. L. Xu, Science 2011, 333, 1303 – 1307.
Manuscript received: January 13, 2017
Revised: February 15, 2017
Final Article published: && &&, &&&&
4
www.angewandte.org
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Epigenetics
S. Schiffers, C. Ebert, R. Rahimoff,
O. Kosmatchev, J. Steinbacher,
A.-V. Bohne, F. Spada, S. Michalakis,
J. Nickelsen, M. Mꢁller,
T. Carell*
&&&& — &&&&
Quantitative LC–MS Provides No
Evidence for m6dA or m4dC in the
Genome of Mouse Embryonic Stem Cells
and Tissues
Previous results challenged: Highly sen-
sitive mass spectrometry reveals that
m6dA is most likely not an epigenetic
base in the mouse genome. It appears
that in vertebrates, the levels of this
modified base, and that of N4-methyl-
deoxycytidine, are much lower than pre-
viously thought.
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!