Intragenomic Decarboxylation of 5-Carboxy-2′-deoxycytidine

Article abstract of DOI:10.1002/anie.202109995

Cellular DNA is composed of four canonical nucleosides (dA, dC, dG and T), which form two Watson–Crick base pairs. In addition, 5-methylcytosine (mdC) may be present. The methylation of dC to mdC is known to regulate transcriptional activity. Next to these five nucleosides, the genome, particularly of stem cells, contains three additional dC derivatives, which are formed by stepwise oxidation of the methyl group of mdC with the help of Tet enzymes. These are 5-hydroxymethyl-dC (hmdC), 5-formyl-dC (fdC), and 5-carboxy-dC (cadC). It is believed that fdC and cadC are converted back into dC, which establishes an epigenetic control cycle that starts with methylation of dC to mdC, followed by oxidation and removal of fdC and cadC. While fdC was shown to undergo intragenomic deformylation to give dC directly, a similar decarboxylation of cadC was postulated but not yet observed on the genomic level. By using metabolic labelling, we show here that cadC decarboxylates in several cell types, which confirms that both fdC and cadC are nucleosides that are directly converted back to dC within the genome by C?C bond cleavage.

Full text of DOI:10.1002/anie.202109995

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Intragenomic decarboxylation of 5-carboxy-2’-deoxycytidine
Authors: Ewelina Kamińska, Eva Korytiaková, Andreas Reichl, Markus
Müller, and Thomas Carell
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202109995
Link to VoR: https://doi.org/10.1002/anie.202109995

10.1002/anie.202109995
Angewandte Chemie International Edition
COMMUNICATION
Intragenomic decarboxylation of 5-carboxy-2’-deoxycytidine
Ewelina Kamińska,#[a] Eva Korytiaková,#[a] Andreas Reichl[a], Markus Müller[a] and Thomas Carell*[a]
Abstract: Cellular DNA is composed of four canonical nucleosides
(dA, dC, dG and T), which form two Watson-Crick base pairs. In
addition, 5-methylcytosine (mdC) may be present. The methylation of
dC to mdC is known to regulate transcriptional activity. Next to these
five nucleotides, the genome, particularly of stem cells, contains three
further dC derivatives, which are formed by stepwise oxidation of the
methyl group of mdC with the help of Tet enzymes. These are 5-
hydroxymethyl-dC, (hmdC), 5-formyl-dC (fdC) and finally 5-carboxy-
dC (cadC). It is believed that fdC and cadC are converted back into
dC, which establishes an epigenetic control cycle that starts with
methylation of dC to mdC, followed by oxidation and removal of fdC
and cadC. While fdC was shown to undergo intra-genomic
deformylation to give dC directly, a similar decarboxylation of cadC
was postulated but not yet observed on the genomic level. Here, we
show, using metabolic labelling in several cell types, that cadC
decarboxylates, which confirms that both fdC and cadC are
nucleosides that are directly converted back to dC within the genome
by C-C bond cleavage.
5-Formyl-dC (fdC) and 5-carboxyl-dC (cadC) are nucleosides that are
found in significant amounts in neurons and stem cells.[1] They are
formed by oxidation of 5-methyl-dC (mdC) by the action of Tet
enzymes via 5-hydroxymethyl-dC (hmdC).[2, 3] HmdC is found in these
genomes in large quantities. The initial methylation of dC to mdC is
performed by the dedicated methyltransferases Dnmt1, 3a and 3b.[4-
7]
The higher oxidized mdC derivatives, fdC and cadC are known to
be removed by the repair glycosylase Tdg, which cleaves the
glycosidic bond between the sugar and the corresponding base.[8, 9] In
result, there is formation of abasic sites processed by AP
endonuclease, and finally replaced by an unmodified dC (Fig. 1a).[10]
Figure 1. a) Depiction of the active demethylation pathways via Tdg-mediated
excision or direct deformylation and decarboxylation. b) Depiction of the
metabolically fed nucleosides 2’-F-cadC 1 and of the product nucleoside 2-’F-
dC 2 formed after decarboxylation. MS reference compound [15N]2 2’F-cadC 3
and [15N]2 2’F-dC 4.[17]
Since the discovery that methylation of dC is followed by oxidation
chemistry, it was postulated that fdC and cadC might directly
deformylate or decarboxylate to give dC. Chemically, these C-C bond
cleavage reactions have the advantage that potentially harmful abasic
site intermediates, formed during Tdg-mediated active demethylation,
will not occur.[11] While for fdC, deformylation was shown to occur in
A direct demethylation reaction in DNA by C-C bond cleavage will give
a product that is identical with the natural dC. Therefore, it is important
to incorporate a reporter nucleoside (cadC*) into the genome of the
cells, which generates a decarboxylated dC*. This product must be
detectable with high accuracy in the presence of an overwhelming
amount of natural dC. In the past, we successfully experimented with
2’fluoro-labelled nucleosides. Thus, for this study we decided to use
vivo, for cadC just a putative decarboxylation mechanism was
13]
postulated in vitro so far.[12,
It is still unknown whether
decarboxylation of cadC occurs in stem cells.[13-16] Here, we use our
previously described metabolic labelling approach to prove that cadC,
if present in the genome of stem- and somatic cells, does
decarboxylate.[17, 18]
2’F-cadC (1) as cadC* (Fig. 1b).[17] The synthesis of 1 and its
20]
triphosphate was reported by us previously.[19,
The 2’F-atom
ensures that the formed decarboxylated product 2’F-dC 2 is readily
detectable by UHPLC-MS/MS due to the m/z = +19 Da mass shift
relative to dC. The specific shift in retention time for fluorinated
compounds also allows to distinguish them from canonical dC. This
reduces an overlap of 2’F-dC with dC, which avoids ion suppression
that makes quantification of even very small quantities of product
possible. The 2’F-atom served also a second purpose. Natural cadC
is barely detectable in wildtype cells.[8] The levels, however, increase
by almost two orders of magnitude, when the base excision repair
(BER) pathway is interrupted by knocking out the TDG gene (Fig. SI-
6).[21] The 2’F substitution has the same effect. It blocks the BER
process, which leads to higher amounts of detectable, incorporated
2’F-cadC (1).[19]
[a]
B. Sc. E. Kamińska, M. Sc. E. Korytiaková, M. Sc. Andreas Reichl, Dr.
M. Müller, Prof. Dr. T. Carell
Department of Chemistry, Ludwig-Maximilians-Universität München,
Butenandtstr. 5-13, 81377 Munich, Germany
http://www.carellgroup.de
*
Corresponding author: Thomas.Carell@lmu.de
#
these authors contributed equally
Supporting information for this article is given via a link at the end of the
document.
This article is protected by copyright. All rights reserved.

10.1002/anie.202109995
Angewandte Chemie International Edition
COMMUNICATION
As a reference compound for exact quantification of 1, we synthesized
the isotopologue [15N]2 2’F-cadC 3, as an internal standard.
Quantification of 2 required the standard [15N]2 2’F-dC 4 (Fig. 1b).
tricine buffer. After the feeding, the cells were washed, and fresh
medium was applied (Fig. 2).
In the first experiment, we fed Neuro-2a cells for 30 min with the 1-TP
loaded transporter. The cells were harvested after 24 hours. We then
isolated the free nucleotide pool and investigated it regarding the
presence of 2’F-cadC 1 and 2’F-dC 2. In the cytosol, we detected to
the presence of delivered 1. In addition, we saw no 2 (Fig. 3a).
Together, the data show that 2’F-cadC 1 is a stable compound that,
based on our data, does not spontaneously decarboxylate during
delivery or under physiological conditions in cells.
A
disadvantage of using 2’F-cadC for studying natural
decarboxylation is of course the unnatural character of the nucleoside.
In order to investigate if the 2’F-atom influences the decarboxylation
reaction, we saturated the C5-C6 bond of the cadC - and 2’F-cadC-
methylesters and studied the spontaneous decarboxylation behavior
after ester cleavage.[13] We found that both compounds decarboxylate
in a similar manner, reassuring us that the effect of the 2’F-atom on
the studied decarboxylation reaction is negligible (Fig. SI-8). When we
started to experiment with 2’F-cadC 1 in cellulo, we learned that it is
problematic to label the cellular genome with this compound. The
negatively charged 2’F-cadC was only to a small extent taken up by
cells and in addition, its intracellular conversion to the triphosphate,
as needed for incorporation into the genome, happened to be
inefficient as well. It is known, that the phosphorylation of cadC by
kinases to the corresponding nucleotides is inefficient.[22]
Next, we harvested the cells using RLT buffer (Qiagen) supplemented
with
400 M
of
2,6-di-tButyl-4-methylphenol
(BHT)
and
desferoxamine mesylate (DM) as well as β-mercaptoethanol (1:100).
The genomic DNA was isolated using a spin column kit (Zymo
Research) and digested according to our established method.[18] In
brief, the isolated DNA was incubated for 5 min at 95 °C, cooled down
on ice and incubated for 4 h at 37°C with the Degradase digestion
mixture (Zymo Research) with the addition of the isotope-labelled
internal standards. Next, the samples were diluted with 450 µL of
water and were extracted with chloroform. After lyophilization of the
aqueous phase, the digested samples were resuspended in water,
filtered and analyzed by UHPLC-MS/MS.[18]
In order to prove that this digestion method is efficient for 2’F-cadC-
containing DNA, we digested a short ssDNA containing a synthetically
embedded 2’F-cadC 1 (Fig. SI-1). Best results were obtained with the
Degradase mix (Fig. SI-1).[24] We next tested the digestion using a
147-base long dsDNA (Widom 601)[25] containing 1. Using LC-MS, we
detected all canonical DNA bases (dA, dC, dG, T) and in addition, 2’F-
cadC 1 at the expected level, confirming that efficient digestion is
possible using the Degradase method. In contrast to this, however we
noticed that when we added the same amount of a 2’F-cadC-
containing DNA strand to normal genomic DNA, we obtained an
astonishingly small signal for 2’F-cadC (1). Indeed, only 10% of the
expected signal was detected, arguing that the detectability of 2’F-
cadC (1) is strongly reduced in a complex environment (Fig. SI-2).
Due to this strong signal suppression, we therefore abstained from
exact quantification of 1, using internal standard 3 in all further
experiments.
To investigate the decarboxylation process, we isolated and digested
genomic DNA from Neuro-2a cells fed with 1-TP. In the above-
described method, indeed we clearly detected a signal for 2’F-cadC 1
in the genome of cells fed with 1-TP based on its retention time, which
was identical with the reference compound 3 and its fragmentation
pattern, which was indistinguishable from 3 (Fig. 3b). To our delight,
we also detected the decarboxylated product 2, which was not seen
in the cytosolic fraction. Compound 2 was detected at levels of 2.4x10-
Figure 2. Depiction of the experimental setup for the inefficient delivery of 1
(blue path) and the accelerated transfer of 1-TP into cells with the help of the
calixarene transporter (red path).
5
per dN, 24 h post feeding. The result shows that while the
decarboxylated product 2 does not form in the cytosol by spontaneous
decarboxylation, it is present in the genome, arguing that
decarboxylation takes place when 2’F-cadC 1 is incorporated into
genomic DNA.
Although we detected 2’F-cadC (1) in the cytosol, we were unable to
detect incorporated 1 in the genome. We had to continuously feed the
cells for 3 days with 1 to reach detectable, but unquantifiable levels of
2’F-cadC 1. After intensive experimentation with different 2’F-cadC
delivery methods, we finally succeeded with the help of a cyclodextrin
Next, we investigated if the decarboxylation signals forms in a dose
dependent manner (Fig. 3c). Feeding Neuro-2a cells with final
concentrations of 100 µM and 300 µM of the 1-TP loaded transporter
gives indeed an increasing signal for 2’F-dC 2 (Fig. 3d). We also see
an increase in the MS signal intensity of 2’F-cadC 1 (Fig. 3c), as
expected.
transporter, which is modified with
a cell-penetrating peptide
derivative (Fig. 2).[23] This transporter encapsulates nucleoside
triphosphates and allows them to be transferred across the cell
membrane. Application of this transporter was indeed successful. It
allowed us to deliver the 1-TP with only one 30 min feeding pulse in
This article is protected by copyright. All rights reserved.

10.1002/anie.202109995
Angewandte Chemie International Edition
COMMUNICATION
incorporation of 2’F-cadC 1 due to the large ion suppression, or that
2’F-cadC 1 decarboxylates to 2’F-dC 2 more efficiently than 2’F-fdC 5
deformylates. Based on earlier chemical studies and our current
knowledge of both processes, we speculate that chemically-favoured
decarboxylation is the contributing factor.[26]
We next asked the question if the product nucleoside 2’F-dC (2) is
remethylated. To this end, we quantified the levels of 2’F-mdC (7) in
both experiments using the internal standard [D3] 2’F-mdC (8). We
clearly detected 7 in both experiments (Fig. 4a). Again, higher
amounts were detected from 2’-F-cadCTP feeding. This finding
confirms that after deformylation or decarboxylation, the formed
product 2 is remethylated, which suggests the presence of a putative
regulatory chemical cycle that starts with methylation of dC by Dnmt
enzymes to mdC, followed by oxidation of mdC to hmdC, fdC and
cadC and then deformylation and decarboxylation of fdC and cadC to
dC, which could start a new circle.[27] Interestingly, when we measured
the levels of 7 formed by methylation of 2 24 h post 1-TP feeding, we
found only 0.2% methylation. When the same measurement was
performed after 72 h, we measured 2.1%, very close to the natural
methylation levels of Neuro-2a cells of about 2.5% (Fig. SI-7). This
result shows that while an early harvest allows us to detect higher
amounts of 2’F-dC, the levels of 2’F-mdC are underestimated,
potentially because the methyltransferases do not have sufficient time
to achieve remethylation.
Figure 3. Metabolic feeding of 2’-F-cadCTP to Neuro-2a cells. Investigation of
the decarboxylation reaction in the cytosol (a) and in genomic DNA (b). Dose
response data showing increased MS signal of 2’-F-cadC 1 (c) and levels of 2’-
F-dC 2, (d) after 24h with increasing feeding concentration.
Then, we performed a time-course experiment to investigate at which
timepoint decarboxylation becomes detectable (Fig. SI-5). We
discovered that 2’F-dC 2 is detectable only after about 8 h at the
earliest, which shows that the C-C bond cleavage has a late onset
(Fig. SI-5a). As a control experiment, to further exclude that we detect
the incorporation of an impurity, we co-fed a 100 µM solution of 1-TP
with a 1% impurity of 2’F-dCTP 2-TP (Fig. SI-5b). In this experiment,
we see again genome incorporated 1 and 2’F-dC 2, but now, 2
appeared already after 30 min.[17] Together, the data show that the
detected 2’F-dC (2) is formed from 2’F-cadC (1) by C-C bond
cleavage.
In the next experiment, we analyzed deformylation of 2’F-fdC (5) and
decarboxylation of 2’F-cadC (1) side by side (Fig. 4a). We fed Neuro-
2a cells with 2’F-fdC-TP in one experiment and with 2’F-cadC-TP in
the second. We used the same amount (400 M) of both materials
and of the transporter. We then exchanged the medium and allowed
the cells to recover for 24 h. Afterwards, the cells were harvested, the
genomic DNA was isolated, digested and analyzed by UHPLC-
MS/MS using the synthetic internal standards 4 as well as [15N]2 2’F-
Figure 4. a) Comparison of deformylation and decarboxylation by feeding of 1-
triphosphate and 5-triphosphate as well as investigation of the remethylation of
the product 2. b) Decarboxylation in different cells (J1, R1, E14). c) Depiction of
2’F-fdC (5) and 2’F-mdC (7) the reference compounds [15N2] 2’F-fdC (6) and
[D3] 2’F-mdC (8) needed for exact quantification.[17]
Finally, we studied how much decarboxylation is taking place in
different cells (Fig. 4b). We investigated the process in somatic
(Neuro-2a, CHO-K1) and in mouse embryonic stem cells (J1, E14,
R1). The stem cells were investigated at the pre-implantation and
post-implantation stages, obtained by culturing the cells in either a2i
or CR media as described in the SI. The data depicted in Fig. 4b show
that upon feeding of 1-TP at 100 µM for 24 h, 2 is clearly detectable
in all cases but the levels vary between the different cell types. The
highest levels of 2 were formed in stem cells cultured under CR
conditions. These are the cells representing post-implantation
embryos, that naturally have the highest methylation levels due to
epigenetic reprogramming during cell lineage differentiation.[28] In
fdC (6) for quantification (Fig. 4c).
We detected clearly the fed starting material 2’F-fdC 5 and again only
just traces (due to ion suppression and/or low incorporation) of 2’F-
cadC 1. Despite this, the amount of the C-C bond cleavage product
2’F-dC 2 was higher when we fed 1-TP compared to 5-TP (Fig. 4a).
The exact quantification of 2 in both experiments, using the
isotopically labelled material 4, clearly showed more 2’F-dC 2 derived
from decarboxylation than deformylation. This interesting result
shows that we either dramatically underestimate the genomic
This article is protected by copyright. All rights reserved.

10.1002/anie.202109995
Angewandte Chemie International Edition
COMMUNICATION
[16]
[17]
Y. Feng, N.-B. Xie, W.-B. Tao, J.-H. Ding, X.-J. You, C.-J. Ma, X.
Zhang, C. Yi, X. Zhou, B.-F. Yuan, Y.-Q. Feng, CCS Chemistry
2021, 3, 994-1008.
K. Iwan, R. Rahimoff, A. Kirchner, F. Spada, A. S. Schröder, O.
Kosmatchev, S. Ferizaj, J. Steinbacher, E. Parsa, M. Müller, T.
Carell, Nat. Chem. Biol. 2018, 14, 72-78.
F. R. Traube, S. Schiffers, K. Iwan, S. Kellner, F. Spada, M. Müller,
T. Carell, Nat. Protoc. 2019, 14, 283-312.
A. S. Schröder, O. Kotljarova, E. Parsa, K. Iwan, N. Raddaoui, T.
Carell, Org. Lett. 2016, 18, 4368-4371.
A. S. Schröder, E. Parsa, K. Iwan, M. Wallner, S. Serdjukow, T.
Carell, Chem. Commun. 2016, 52, 14361-14364.
R. Rahimoff, O. Kosmatchev, A. Kirchner, T. Pfaffeneder, F.
Spada, V. Brantl, M. Müller, T. Carell, J. Am. Chem. Soc. 2017,
139, 10359-10364.
M. Zauri, G. Berridge, M.-L. Thézénas, K. M. Pugh, R. Goldin, B.
M. Kessler, S. Kriaucionis, Nature 2015, 524, 114-118.
Z. Zawada, A. Tatar, P. Mocilac, M. Buděšínský, T. Kraus, Angew.
Chem. Int. Ed. 2018, 57, 9891-9895.
general, stem cells show an about ten times higher decarboxylation
activity than somatic cells, which underpins the potential epigenetic
importance of the process.
In summary, the presented data show that next to deformylation of
fdC, we also need to consider decarboxylation of cadC as a
mechanism for active demethylation. All further efforts now need to be
concentrated at finding the cellular entities or circumstances that
enable these C-C bond cleavage reactions. Although the here
reported data clearly point to the existence of decarboxylation, we
need to emphasize that cells are complex systems and we feed an
unnatural compound. Without clear identification of the biological
entity responsible for the process, we cannot completely rule out that
unknown processes other than intragenomic decarboxylation are
responsible for the measured data. During the review process of this
manuscript Feng and coworkers showed an incorporation of the F-
carboxylcytidine as a nucleoside and interestingly managed to detect
the decarboxylation of cadC to dC much earlier on.[29]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
F. Yuan, Y. Bi, J.-Y. Zhang, Y.-L. Zhou, X.-X. Zhang, C.-X. Song,
RSC Adv. 2019, 9, 29010-29014.
W. J. Lowary PT, J Mol. Biol. 1998, 276, 19-42.
E. Korytiaková, E. Kamińska, M. Müller, T. Carell, Angew. Chem.
Int. Ed. 2021, 60, 16869-16873.
[25]
[26]
[27]
[28]
A. Parry, S. Rulands, W. Reik, Nat. Rev. Genet. 2021, 22, 59-66.
S. Takahashi, Kobayashi, S., Hiratani, I., CMLS 2018, 75, 1191-
1203.
Acknowledgements
[29]
Y. Feng, J.-J. Chen, N.-B. Xie, J.-H. Ding, X.-J. You, W.-B. Tao, X.
Zhang, C. Yi, X. Zhou, B.-F. Yuan, Y.-Q. Feng, Chem. Sci. 2021.
We thank the Deutsche Forschungsgemeinschaft (DFG, German
Research Foundation) for financial support via SFB1309 (PID
325871075) and SFB1361 (PID 393547839). This project has
received funding from the European Research Council (ERC) under
the European Union's Horizon 2020 research and innovation
programme (grant agreement n° EPiR 741912). Additional funding
was provided by the Volkswagen Foundation (EvoRib) and the DFG
priority program SPP1784 (PID 277203618).
Keywords: epigenetics; carboxylcytidine; decarboxylation; C-C
bond cleavage; active demethylation
[1]
T. Carell, C. Brandmayr, A. Hienzsch, M. Müller, D. Pearson, V.
Reiter, I. Thoma, P. Thumbs, M. Wagner, Angew. Chem. Int. Ed.
2012, 51, 7110-7131.
[2]
[3]
S. Kriaucionis, N. Heintz, Science 2009, 324, 929-930.
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.
[4]
[5]
Z. Li, H. Dai, S. N. Martos, B. Xu, Y. Gao, T. Li, G. Zhu, D. E.
Schones, Z. Wang, Genome Biol. 2015, 16.
C.-C. Chen, K.-Y. Wang, C.-K. J. Shen, J. Biol. Chem 2013, 288,
9084-9091.
[6]
[7]
[8]
C. Luo, P. Hajkova, J. R. Ecker, Science 2018, 361, 1336-1340.
M. Okano, D. W. Bell, D. A. Haber, E. Li, Cell 1999, 99, 247-257.
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.
[9]
L. S. Pidugu, Q. Dai, S. S. Malik, E. Pozharski, A. C. Drohat, J.
Am. Chem. Soc. 2019, 141, 18851-18861.
[10]
[11]
[12]
[13]
H. E. Krokan, M. Bjoras, Cold Spring Harb. Perspect. Biol. 2013, 5,
a012583-a012583.
A. J. Wossidlo M, Sebastiano V, Lepikhov K, Boiani M, Reinhardt
R, Schöler H, Walter J., EMBO J. 2010, 29, 1877-1888.
S. Schiesser, B. Hackner, T. Pfaffeneder, M. Müller, C. Hagemeier,
M. Truss, T. Carell, Angew. Chem. Int. Ed. 2012, 51, 6516-6520.
S. Schiesser, T. Pfaffeneder, K. Sadeghian, B. Hackner, B.
Steigenberger, A. S. Schröder, J. Steinbacher, G. Kashiwazaki, G.
Höfner, K. T. Wanner, C. Ochsenfeld, T. Carell, J. Am. Chem. Soc.
2013, 135, 14593-14599.
[14]
[15]
G. Song, G. Wang, X. Luo, Y. Cheng, Q. Song, J. Wan, C. Moore,
H. Song, P. Jin, J. Qian, H. Zhu, Nat. Commun. 2021, 12, 795.
Z. Liutkeviciute, E. Kriukiene, J. Licyte, M. Rudyte, G.
Urbanaviciute, S. Klimasauskas, J. Am. Chem. Soc. 2014, 136,
5884-5887.
This article is protected by copyright. All rights reserved.