Visible-light-mediated oxidative demethylation of: N 6-methyl adenines

Article abstract of DOI:10.1039/c7cc05544g

We report a simple protocol that affords oxidative demethylation of N⁶-methyl groups in N⁶-methyl adenines (m⁶A). The biologically compatible photocatalyst riboflavin prompts a highly selective C-H abstraction from N⁶-methyl in adenines under the irradiation of a visible blue LED light, affording a novel and highly selective biomimetic demethylation of m⁶A and related N-methyl adenine analogues. andcopy; 2017 The Royal Society of Chemistry.

Full text of DOI:10.1039/c7cc05544g

ChemComm
View Article Online
View Journal
COMMUNICATION
Visible-light-mediated oxidative demethylation
of N6-methyl adenines†
Cite this:DOI: 10.1039/c7cc05544g
abc
Li-Jun Xie,ac Rui-Li Wang,ac Dong Wang,a Li Liuac and Liang Cheng
*
Received 18th July 2017,
Accepted 11th September 2017
DOI: 10.1039/c7cc05544g
rsc.li/chemcomm
We report a simple protocol that affords oxidative demethylation of
N6-methyl groups in N6-methyl adenines (m6A). The biologically
compatible photocatalyst riboflavin prompts a highly selective C–H
abstraction from N6-methyl in adenines under the irradiation of
a visible blue LED light, affording a novel and highly selective
biomimetic demethylation of m6A and related N-methyl adenine
analogues.
Scheme 1 Comparison of oxidative demethylation of N6-methyladenines
through biological demethylases (a), the hydroperoxide system (b) and the
novel simplified light-driven pathway (c).
N6-Methyladenosine (m6A) is the most abundant internal modi-
fication in eukaryotic mRNA, although the significant biological
roles of m6A methylation have remained largely unclear.1 Recent
studies on the distributions and mechanisms of m6A modifica-
tion have suggested that this methylation functionally modulates always lack of generality.5 Recently, Zhou et al. reported an
the eukaryotic transcriptome to influence mRNA transcription, oxidation of m6A through the bicarbonate activated peroxide
splicing, nuclear export, localization, translation, stability, and system (Scheme 1, path (b)).5a 39% of the demethylated product –
other physiological processes unknown to us.2 Fat mass-/obesity- adenosine (A) – was isolated after incubating m6A with excess H2O2
associated proteins (FTO) and AlkB homologue 5 (AlkBH5) are two for 24 hours. This finding opens up a fascinating new way for
recently discovered RNA demethylases (‘‘eraser’’) that remove the the direct removal of methyl groups from m6A; however, the
m6A modification in mRNA in the presence of Fe(II), a-KG, and comparably inferior conversion retards it further application in
dioxygen3 (Scheme 1, path (a)). In this oxidative demethylation biological processes. Thus the quest for new approaches to a
mechanism, a non-heme iron(II) and a-KG activate oxygen to form selective biocatalytic oxidation continues.
a highly reactive Fe(IV)-oxo species, which then inserts the oxygen
While it is well known that photocatalysis is a very common
into the C–H bond of m6A via a radical process, followed by principle in nature that all plants and animals depend on
tandem oxidation of the intermediates hm6A (N6-hydroxymethyl sunlight and use it by means of photoreceptors,6 we intention-
adenosine) and/or f6A (N6-formyl adenosine) to furnish the ally incorporate such photomodulation into the chemical
demethylation.4 However, owing to the necessity of formation of demethylation. In contrast to the traditional photocontrol by
unique structures that are responsible for m6A recognition for ultraviolet light as the trigger, which is not well compatible with
these proteins, the oxidative demethylation of the C–H bond in live cells due to its cytotoxicity and by causing irreversible
m6A is difficult to achieve with external chemical reagents and is nucleic acid mutations,7 we strongly expect that, by carefully
choosing diverse biologically photoreceptors, the target nucleo-
side will be specifically activated by longer wavelength lights, in
which case the reaction would be more selective, predicable
and easier to control.8 Contrary to the relatively unspecific
functionalization of nucleosides by C–H activation,9 this kind
of activation would allow a direct C–H abstraction from the
inert N6-methyl groups of m6A to form a nucleoside radical.10
Indeed, we have recently demonstrated that by cautiously
selecting the radical initiator, the unreactive C–H bonds in
a Beijing National Laboratory for Molecular Sciences (BNLMS), CAS Key Laboratory
of Molecular Recognition and Function, CAS Research/Education Center for
Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of
Sciences, Beijing 100190, China. E-mail: chengl@iccas.ac.cn
b Key Lab of Functional Molecular Engineering of Guangdong Province
(South China University of Technology), Guangzhou 510640, China
c University of Chinese Academy of Sciences, Beijing 100049, China
† Electronic supplementary information (ESI) available: Experimental procedure,
compound characterization and HPLC assays. See DOI: 10.1039/c7cc05544g
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
heterocycles can be selectively activated.11 Herein, we focus on for details). The oxidative demethylation of N6-methyl adenosine 1
the development of a visible light-driven oxidative demethylation proceeded quickly within 2 hours under blue LED light irradiation.
of m6A with riboflavin (vitamin B2) (Scheme 1, path (c)), which is a A high yield (86%) of the demethylated product was achieved by
biological redox cofactor12 that is essential for human and animal simple chromatography, and upon analysis the product was
health because of its crucial role in many natural processes, such revealed to be identical with authentic adenine. It should be noted
as DNA repair by photolyase.13 To the best of our knowledge, this that traditional photo-oxidative N-demethylation methods that were
is the first example of chemical modulation of RNA epigenetics applicable for heterocycles, such as alkaloids,14 are not applicable
by implementing an energy-efficient and green process through a in this situation and no demethylated product was observed in all
photo-induced oxidative demethylation.
cases (entries 7 and 8). Selectfluor was chosen owing to its superior
As a starting point, the demethylation of m6A N6-methyl safety profile and lower cost, while other oxidants such as TBHP
adenosine 1 was investigated (Table 1). Before we initiated this (entry 10) or H2O2 (entry 11) were proven to be ineffective for this
research, this transformation was only achieved with modest transformation.15 Water and light were identified as having a
conversions using stoichiometric loadings of metal or inorganic crucial impact on the demethylation as no product was observed
oxidants by using protected nucleosides.5b–d A screening of photo- when the photoreaction was conducted in dry acetonitrile or under
catalysts (2a–2g), oxidants, solvents and light wavelengths quickly darkness (entries 12 and 13).
identified 10 mol% of riboflavin with selectfluor in aqueous
By extension, the chemical demethylation of other methylated
acetonitrile to be the optimal conditions (entries 1–8; see the ESI† nucleosides/bases was also explored under the optimized reac-
tion conditions (Table 2). Again, N6,6-dimethyl adenosine (m6,6A,
4), deoxy-N6-methyl adenosine (dm6A, 5) and N6-methyl adenine
(6) furnished the excepted demethylated nucleosides/bases with
high yields. In contrast, other RNA epigenetic patterns, such as
the methyl group at the 1-nitrogen of the adenine ring (m1A, 7),
5-carbon of the cytosine ring (m5C, 8) or at the 20-OH of the ribose
(Am, 9) remained inert or partially decomposed, suggesting a
particularly selective recognition and activation of the N6-methyl
C–H bond with this riboflavin-based photocatalysis. To explore
the potential of this demethylation with other biological sub-
strates, the standard procedure was applied to various endo-
genous and exogenous biomolecules, such as alkaloid (sparteine),
natural and unnatural amino acids (L-histidine, L-tryptophan,
L-phenylalanine and L-phenylglycine), nucleosides (guanosine,
inosine and deoxy-thymidine) and oligosaccharides (ribose and
xylose). All of these substrates remained intact under the same
Table 1 Riboflavin-catalyzed demethylation of N6-methyl adenosine 1a
Table
2
Riboflavin-catalyzed demethylation of other epigenetic
nucleosidesa
Entry
Deviation from above
Yield (%)
1
2
3
None
No photoreceptor
2b
86
NRb
79
4
5
2c
2d
45
62
6
7
8
9
10
11
12
13
2e
2f
2g
52
NRb
NRb
NRb
NRb
NRb
NRb
NRb
No selectfluor
TBHP
H2O2
CH3CN (anhydrous)
No light
a
Unless otherwise specified, all reactions were carried out using
nucleoside/base (0.1 mmol), riboflavin 2a (0.01 mmol, 3.8 mg) and
a
Unless otherwise specified, all reactions were carried out using 1 selectfluor (0.22 mmol, 78 mg) in the solvent of acetonitrile and water
(0.1 mmol, 28 mg), 2a (0.01 mmol, 3.8 mg) and selectfluor (0.22 mmol, (1 : 1, 2 mL total) at RT (37 1C) under 1 atm of argon (balloon). Isolated
b
c
78 mg) in the solvent of CH3CN and H2O (1 : 1, 2 mL total) at RT (37 1C) yield is given. Reaction time: 7 hours. 17% of mono-demethylated
b
d
e
for 2 hours under 1 atm of argon (balloon). Isolated yield is given. NR: product m6A 1 was isolated. Reaction time: 3 hours. Reaction time:
no reaction.
24 hours. NR: no reaction.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017

View Article Online
ChemComm
Communication
conditions, again demonstrating the bimolecular compatibility
of this selective demethylation process.
To investigate the oxidative process and the demethylation
mechanism, we used high-performance liquid chromatography
(HPLC) to monitor the demethylation reaction (Fig. 1). When a
50 mM aliquot of m6A (Fig. 2a) in a mixture of acetonitrile and
water (1 : 1, 2 mL total) was treated with riboflavin (10 mol%)
and selectfluor (2.2 equiv.) at RT for 30 minutes, three new peak
were observed, one of which was identical with that of authentic
adenosine (Fig. 1(a)). Liquid chromatography–mass spectro-
metry (LC-MS) was used to characterize the other two new peaks,
and it turned out that these two initially generated products were
hm6A and f6A, respectively (Fig. 1(b)). Substantial conversion
of hm6A was observed in 30 minutes (Fig. 2e), and f6A to A in
60 minutes (Fig. 1(a)). To our surprise, another intermediate
Fig. 2 Proposed mechanism for the riboflavin catalysed oxidative
demethylation of N6-methyl adenines.
ca6A (N6-carboxyl adenosine)16 was also observed from LC-MS
(Fig. 1(b)), which was suspected to be originated from further
oxidation of f6A. However, carbamic acid was chemically
unstable under the oxidative conditions and was believed to
undergo a rapid decarboxylation to afford the demethylated
product A.
Based on previous reports and these experimental results, a
putative mechanism (Fig. 2) is outlined to illustrate the reaction
route. Initially, the irradiation of riboflavin (PC = photocatalyst)
with 470 nm visible light leads to the formation of the electro-
nically excited photocatalyst (*PC), which then activated m6A 1
by selectively abstracting the C–H bond from the N6-methyl
group in m6A 1 via an initial single electron transfer (SET)
process to form a radical cationic nucleoside I10 in a similar
manner as enzymes FTO or AlkBH5. Deprotonation followed by
further SET oxidation affords an iminium salt II, primed for
nucleophilic attack by water to generate the intermediate hm6A.
The unstable hemiaminal would undergo decomposition in
an aqueous solution to yield adenosine A and formaldehyde
(Fig. 2, path (a)), which was proven by an HPLC assay for the
determination of formaldehyde in the oxidation mixture (see
the ESI† for details). Alternatively, a subsequent second-round
oxidation of the hemiaminal hm6A to a relatively stable form-
amide f6A followed by water-assisted decomposition afforded
formic acid and the demethylated nucleoside A (path (b)).
A third-round oxidation product ca6A, which seems unstable
under these physiological conditions, might also be involved in
this process. The major demethylation pathway of m6A by the
‘‘eraser’’ FTO was believed to proceed via the decomposition of
hm6A, while the subsequent hydrolysis of f6A was a minor and
Fig. 1 Analysis of riboflavin catalysed photochemical demethylation of
m6A. (a) HPLC analysis of the demethylation process at 0 min, 30 min,
60 min and 120 min, respectively. The new peaks were assigned as A, hm6A
and f6A, respectively. (b) Mass profile of the reaction mixtures at 30 min and
60 min, respectively. ca6A was observed as a potential intermediate.
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
M. A. Nobrega, K. L. Han, Q. Cui and C. He, Nat. Commun., 2013,
4, 1798; (d) G. Zheng, J. A. Dahl, Y. Niu, P. Fedorcsak, C. M. Huang,
C. J. Li, C. B. Vågbø, Y. Shi, W. L. Wang, S. H. Song, Z. Lu, R. P.
Bosmans, Q. Dai, Y. J. Hao, X. Yang, W. M. Zhao, W. M. Tong,
X. J. Wang, F. Bogdan, K. Furu, Y. Fu, G. Jia, X. Zhao, J. Liu,
H. E. Krokan, A. Klungland, Y. G. Yang and C. He, Mol. Cell, 2013,
49, 18.
inefficient demethylation pathway due to the slow reorganiza-
tion step of the enzyme with the nucleic acid strand.17 However,
the fact that the substantial generation and hydrolysis of f6A
was observable from HPLC clearly proved the second mecha-
nism proceeding by two rounds of oxidation followed by amide
hydrolysis (path (c)).
4 F. Ye, L. Zhang, L. Jun, M. Zheng, H. Jiang and C. Luo, MedChem-
Comm, 2014, 5, 1797.
In conclusion, we have developed a novel light-driven system
for the effective oxidative demethylation of m6A and its
N6-methyl analogues. Riboflavin was selected as a biologically
compatible photoreceptor to selectively abstract hydrogen from
the N6-methyl group in adenine. Photo-induced SET and hydrogen
abstraction between the m6A and excited riboflavin afforded an
iminium intermediate, which later underwent hydration and
oxidation to afford the intermediates hm6A, f6A and ca6A, followed
by decomposition to give the demethylated product. This new
platform replying on riboflavin achieves an inert C–H bond
functionalization in epigenetic nucleosides, which is not amenable
to direct derivation. Further extension of this biorthogonal chem-
istry for diverse functionalization of nucleosides and oligonucleic
acids is currently investigated in our lab and will be reported in
due course.
This work was supported by the National Natural Science
Foundation of China (21778057, 21502201, 21372224, 21420102003,
and 21232008), the National Key R&D Program of China
(2017YFA0208100 and 2016YFA0602900), Beijing Natural Science
Foundation (2162049), the Young Elite Scientist Sponsorship
Program by CAST (2015QNRC001), Shandong Province Indepen-
dent Innovation and Achievements Transformation Special Fund
(2014XGC06001), the Open Fund of the Key Laboratory of Func-
tional Molecular Engineering of Guangdong Province (2016kf08)
(South China University of Technology) and Chinese Academy of
Sciences. Insightful comments from Prof. Xinjing Tang (Peking
University), Prof. Jiangyun Wang (Institute of Biophysics,
Chinese Academy of Sciences) and Prof. Jianwei Sun (Hong Kong
University of Science and Technology) are greatly appreciated.
5 (a) J. Wu, H. Xiao, T. Wang, T. Hong, B. Fu, D. Bai, Z. He, S. Peng,
X. Xing, J. Hu, P. Gao and X. Zhou, Chem. Sci., 2015, 6, 3013;
(b) Y. Kitade, R. Nakanishi, M. Sako, K. Hirota and Y. Maki, Chem.
Pharm. Bull., 1991, 39, 1902; (c) T. Kato, S. Ogawa and I. Ito,
Tetrahedron Lett., 1981, 22, 3205; (d) T. Itaya, K. Ogawa,
H. Matsumoto and T. Watanabe, Chem. Pharm. Bull., 1980, 28, 2819.
6 (a) A. Aukauloo and W. Leibl, Adv. Bot. Res., 2016, 79, 63; (b) J. A.
´
´
Macia-Agullo, A. Corma and H. Garcia, Chemistry, 2015, 21, 10940;
(c) D. M. Schultz and T. P. Yoon, Science, 2014, 343, 985.
7 For examples, see (a) R. Brem, M. Guven and P. Karran, Free Radical
Biol. Med., 2017, 107, 101; (b) A. Conconi and B. Bell, Nature, 2017,
545, 165; (c) A. F. El-Yazbi and G. R. Loppnow, Trends Anal. Chem.,
2014, 61, 83; (d) J.-R. Meunier, A. Sarasin and L. Marrot, Photochem.
Photobiol., 2002, 75, 437.
8 (a) S. Roslin and L. R. Odell, Eur. J. Org. Chem., 2017, 1993;
(b) O. Boubertakh and J. P. Goddard, Eur. J. Org. Chem., 2017,
2072; (c) X. K. Cheng, X. G. Hu and Z. Lu, Chin. J. Org. Chem.,
2017, 37, 251.
9 V. Gayakhe, Y. S. Sanghvi, I. J. S. Fairlamb and A. R. Kapdi, Chem.
Commun., 2015, 51, 11944.
10 (a) M. Sako, T. Makino, Y. Kitade, K. Hirota and Y. Maki, J. Chem.
Soc., Perkin Trans. 1, 1992, 1801; (b) A. J. S. C. Vieira and S. Steenken,
J. Am. Chem. Soc., 1990, 112, 6986; (c) T. Endo and J. Zemlicka, J. Org.
Chem., 1979, 44, 3652.
11 (a) H.-Y. Huang, L. Cheng, J.-J. Liu, D. Wang, L. Liu and C.-J. Li,
J. Org. Chem., 2017, 82, 2656; (b) D.-L. Kong, L. Cheng, T. Yue, H.-R.
Wu, W.-C. Feng, D. Wang and L. Liu, J. Org. Chem., 2016, 81, 5337;
(c) H.-R. Wu, L. Cheng, D.-L. Kong, H.-Y. Huang, C.-L. Gu, L. Liu,
D. Wang and C.-J. Li, Org. Lett., 2016, 18, 1382; (d) D.-L. Kong,
L. Cheng, H.-R. Wu, Y. Li, D. Wang and L. Liu, Org. Biomol. Chem.,
2016, 14, 2210.
12 (a) H. Lida, Y. Imada and S.-I. Murahashi, Org. Biomol. Chem., 2015,
13, 7599; (b) R. Cibulka, Eur. J. Org. Chem., 2015, 915; (c) M. Insinska-
Rak and M. Sikorski, Chem. – Eur. J., 2014, 20, 15280; (d) G. de
Gonzalo and M. W. Fraaije, ChemCatChem, 2013, 5, 403.
13 (a) I. M. Wijaya, T. Iwata, E. D. Getzoff and H. Kandori, J. Am. Chem.
Soc., 2016, 138, 4368; (b) H. Bohr, K. J. Jalkanen and F. B. Malik,
Mod. Phys. Lett. B, 2005, 19, 473; (c) R. Epple, E. U. Wallenborn and
T. Carell, J. Am. Chem. Soc., 1997, 119, 7440.
14 (a) G. B. Kok and P. J. Scammells, Bioorg. Med. Chem. Lett., 2010,
20, 4499; (b) J. A. Ripper, E. R. T. Tiekink and P. J. Scammells, Bioorg.
Med. Chem. Lett., 2001, 11, 443; (c) J. Santamaria, M. T. Kaddachi
and J. Rigaudy, Tetrahedron Lett., 1990, 31, 4735; (d) J. Santamaria,
R. Ouchabane and J. Rigaudy, Tetrahedron Lett., 1989, 30, 3977;
(e) J. Santamaria, R. Ouchabane and J. Rigaudy, Tetrahedron Lett.,
1989, 30, 2927; ( f ) S.-I. Murahashi, T. Naota and K. Yonemura,
J. Am. Chem. Soc., 1988, 110, 8256.
Conflicts of interest
There are no conflicts to declare.
Notes and references
1 (a) J. Song and C. Yi, ACS Chem. Biol., 2017, 12, 316; (b) L. Liu and
T. Pan, Nat. Struct. Mol. Biol., 2016, 23, 98; (c) T. Pan, Trends Biochem. 15 (a) A. Hauser and R. Bohlmann, Synlett, 2016, 1870; (b) M. H.
Sci., 2013, 38, 204.
Daniels and J. Hubbs, Tetrahedron Lett., 2011, 52, 3543.
16 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 and
G. L. Xu, Science, 2011, 333, 1303.
2 Y. Yue, J. Liu and C. He, Genes Dev., 2015, 29, 2343.
3 (a) L. Shen, C. X. Song, C. He and Y. Zhang, Annu. Rev. Biochem.,
2014, 83, 585; (b) G. Zheng, J. A. Dahl, Y. Niu, Y. Fu, A. Klungland,
Y. G. Yang and C. He, RNA Biol., 2013, 10, 915; (c) Y. Fu, G. Jia, 17 B. Wang, Z. Cao, D. A. Sharon and S. Shaik, ACS Catal., 2015,
X. Pang, R. N. Wang, X. Wang, C. J. Li, S. Smemo, Q. Dai, K. A. Bailey,
5, 7077.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017