Synthesis and Metabolic Fate of 4-Methylthiouridine in Bacterial tRNA

Article abstract of DOI:10.1002/cbic.202000272

Ribonucleic acid (RNA) is central to many life processes and, to fulfill its function, it has a substantial chemical variety in its building blocks. Enzymatic thiolation of uridine introduces 4-thiouridine (s⁴U) into many bacterial transfer RNAs (tRNAs), which is used as a sensor for UV radiation. A similar modified nucleoside, 2-thiocytidine, was recently found to be sulfur-methylated especially in bacteria exposed to antibiotics and simple methylating reagents. Herein, we report the synthesis of 4-methylthiouridine (ms⁴U) and confirm its presence and additional formation under stress in Escherichia coli. We used the synthetic ms⁴U for isotope dilution mass spectrometry and compared its abundance to other reported tRNA damage products. In addition, we applied sophisticated stable-isotope pulse chase studies (NAIL-MS) and showed its AlkB-independent removal in vivo. Our findings reveal the complex nature of bacterial RNA damage repair.

Full text of DOI:10.1002/cbic.202000272

Combining Chemistry and Biology
Accepted Article
Title: Synthesis and metabolic fate of 4-methylthiouridine in bacterial
tRNA
Authors: Christoph Borek, Valentin F. Reichle, and Stefanie Kellner
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: ChemBioChem 10.1002/cbic.202000272
Link to VoR: https://doi.org/10.1002/cbic.202000272
01/2020

10.1002/cbic.202000272
ChemBioChem
COMMUNICATION
Synthesis and metabolic fate of 4-methylthiouridine in bacterial
tRNA
Christoph Borek,[a]# Valentin F. Reichle,[a]# and Stefanie Kellner*[a]
[a] Dr. Christoph Borek, Valentin F. Reichle, Dr. Stefanie Kellner
Department of Chemistry, Ludwig-Maximilians-Universität München, Butenandtstr. 5-13, 81377 Munich, Germany.
[*] corresponding author: Stefanie.kellner@cup.lmu.de
[#] These authors contributed equally to this work.
Supporting information for this article is given via a link at the end of the document.
[5]
Abstract: Ribonucleic acid (RNA) is central to many life processes
and to fulfil its function it has a substantial chemical variety of its
building blocks. Enzymatic thiolation of uridine introduces 4-
thiouridine (s4U) into many bacterial transfer RNAs (tRNAs) which is
used as a sensor for UV radiation. A similar modified nucleoside, 2-
thiocytidine, was recently found to be sulfur-methylated especially in
bacteria exposed to antibiotics and simple methylating reagents.
Here, we report the synthesis of 4-methylthiouridine (ms4U) and
confirm its presence and additional formation under stress in E. coli.
We use the synthetic ms4U for isotope dilution mass spectrometry
and compare its abundance to other reported tRNA damage
products. In addition, we apply sophisticated stable isotope pulse
chase studies (NAIL-MS) and show its AlkB independent removal in
vivo. Our findings reveal the complex nature of bacterial RNA
damage repair.
as bromomethylcoumarin
or iodoacetamide [6]. The latter is
used to assess RNA transcription and stability after metabolic
RNA labeling with exogenous s4U (SLAM-Seq). Similar to
SLAM-Seq, TUC-Seq utilizes metabolically introduced s4U,
which can be chemically converted to cytidine prior to RNA
sequencing [7]). Despite its important function in bacterial tRNA
and its broad use as a metabolic label for RNA sequencing, only
little is known about its chemical reactivity inside cells.
Another sulfur decorated tRNA modification, 2-thiocytidine (s2C)
(blue in Fig. 1) has been recently found to be endogenously
[8]
methylated
and efficiently repaired, potentially through its
function as a modulator of translation [9]. A direct methylation of
s2C through electrophiles such as S-adenosylmethionine, methyl
methanesulfonate (MMS) or antibiotics (streptozotocin) was
observed. The resulting damage ms2C (Fig. 1) is substrate to
the α-ketoglutarate dependent dioxygenase AlkB and repaired
both in vitro and in vivo to restore tRNA function [8]. While s2C is
fully accessible to electrophiles in the anticodon loop of tRNA,
s4U is in tight interaction with nucleosides of the D- and T-loop
and might be less accessible to electrophiles. This raises the
question whether s4U is a target to direct methylation and if so,
how much damage forms and how bacteria react to the damage.
To address these questions, we report here the synthesis of the
suggested damage product ms4U (4).
RNA and especially tRNA have complex structures to fulfil their
important functions inside the organism. This is possible through
the vast chemical variety of building blocks found in RNA. To
this day over 170 modifications to either ribose or nucleobase
have been reported [1]. One group of unique tRNA modifications
is the enzymatic thiolation. In bacteria, thiolation of uridine (4-
thiouridine, s4U) is commonly found at position 8 of most tRNAs
(red in Figure 1). s4U is target of ultraviolet light [2] which leads to
The synthesis of ms4U (4) was first attempted via the formation
of the fully acetylated corresponding 4-triazolic precursor which
was meant to react with sodium thiomethanolate [10] to form the
a
reduced growth of UV exposed bacteria and as a
consequence safes bacteria from photomutagenic effects [3]. In
addition, s4U-hypomodified tRNAs were found to be targeted by
the RNA degradosome which leads to a reduced abundance of
a subset of bacterial tRNAs [4]. Due to its sulfur decoration, s4U
is a nucleophile, which can be coupled with electrophiles such
Figure 1. 3D structure of tRNA indicating the positions of enzymatically
thiolated nucleobases in bacteria. Red: 4-thiouridine (3, s4U) found at position
8 and the suggested structure of its methylated derivative 4-methylthiouridine
(4, ms4U). Blue: 2-thiocytidine at position 32 and its reported derivative 2-
methylthiocytidine.
Scheme 1. Synthesis of the key compound 4: a) Ac2O, I2, RT, 45 min; b) P2S5,
pyridine, reflux, 4 h; c) NH4OH conc., reflux, 2 h; d) MeI, EtOH 50 %, RT, 1 h.
1
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000272
ChemBioChem
COMMUNICATION
standard and tRNA from MMS exposed E. coli grown in stable
isotope labeled medium clearly showed a) perfect co-elution and
b) the expected numbers of carbon, nitrogen and sulfur atoms in
native ms4U (Fig. 2a). In a next step, we confirmed the origin of
the methyl group attached to the sulfur following our established
methylome discrimination assay [13]. For this purpose, we grew E.
coli in CD3-methionine supplemented medium, which leads to a
CD3-label of all enzymatically placed methyl groups. After
exposing E. coli to MMS, we found a high intensity signal for
CH3-methylated ms4U
desired nucleoside. We encountered several problems in the
key step due to partial deprotection of the ribose moiety, which
led to further problems with the purification. Therefore, we
decided to form ribose-protected 4-thiouridine (2) separately with
subsequent methylation adopting
a
procedure for the
corresponding 2'-chlorine riboside [11]. The complete reaction is
shown in Scheme 1. The initial peracetylation in neat acetic
anhydride with catalytic amounts of iodine is a fast and reliable
method to protect sugars in general which provided conversion
of uridine to compound (1) in high yields. The subsequent
formation of the 4-thiouridinic compound (2) by thiolation with
phosphorus pentasulfide yielded 72 %. It should be noted, that a
crystallization from ethanol, as described for the chlorinated
compound could not be observed. The deprotection was
conducted by refluxing in concentrated aqueous ammonia
solution and s4U (3) was received presumably in quantitative
yield but was used as crude product in the next step. Of note,
the more common method under Zemplén conditions [12] was not
capable of deprotecting compound (2). In a final step, the thio
group was selectively methylated by iodomethane providing
ms4U (4) in a moderate overall yield of 40 % over four steps.
and only a minor signal for CD3-methylated ms4U (Fig. 2b). We
thus prove the direct methylation of s4U through the electrophile
MMS in bacterial tRNA in vivo.
We were next interested to quantify the extent of ms4U formation
in unstressed and MMS treated tRNA. For this purpose, a stable
isotope labeled internal standard (SILIS) of ms4U was produced
by metabolic isotope labeling of E. coli. To increase the yield of
stable isotope labeled ms4U, MMS was added for 60 minutes to
the culture medium and the RNA was harvested and processed
as previously described [14]. The combination of the synthesized
ms4U and metabolically produced ms4U-SILIS allowed accurate
quantification of ms4U and other modified ribonucleosides in
bacterial tRNA (Fig. 3). For normalization, we plotted the
number of modified nucleosides per 106 canonical
ribonucleosides (rN).
In tRNA of unstressed E. coli, we found 2.6 x 10-6 ms4U/rN,
which is less compared to the natural abundance of our recently
described modification ms2C (17 x 10-6 ms2C/rN) [8]
.
Figure 3. Absolute quantification of damage derived nucleosides found in
tRNA in control and 20 mM MMS exposed E. coli. Left: per 106 rN
(ribonucleosides) Right: per precursor in %. Abbreviations: m1A, 1-
methyladenosine; ms2C, 2-methylthiocytidine; m3C, 3-methylcytidine; ms4U, 4-
methylthiouridine and m3U, 3-methyluridine. From 3 biol. replicates. Error bars
represent standard deviation.
Figure 2. LC-MS/MS analysis of native and synthesized ms4U. a Co-injection
of synthesized ms4U (black) and digested tRNA from 13C (red), 15N (blue) and
34S (yellow) metabolically labeled E. coli cultures. b Native tRNA digests
screening for enzymatically methylated nucleosides (grey, CD3-methionine
derived) and damage derived nucleoside methylation (black). Abbreviations:
ms4U (4-methylthiouridine (4)), m5U (5-methyluridine) and Cm (2’-O-
methylcytidine). The mass transitions (precursor ion  product ion) are given
below the respective chromatograms.
After exposure to MMS, the abundance of the known tRNA
damage products is 2147 x 10-6 m1A/rN, 1543 x 10-6 ms2C/rN,
2772 x 10-6 m7G/rN, 530 x 10-6 m3C/rN, 478 x 10-6 m6A/rN and
41 x 10-6 m3U/rN (Fig. 3 and Fig. S1a). ms4U damage is with 91
x 10-6 ms4U/rN comparable to m3U damage in bacterial tRNA.
This value appears to be rather low, but if the abundance of
damage is normalized to the abundance of its respective
precursor nucleoside (e.g. m1A per A or ms4U per s4U) a
different conclusion must be drawn. With 0.5 % ms4U/s4U, ms4U
is of comparable abundance to the known damage product m1A
(1.1 % per A) (Fig. 3 and Fig. S1b). S4 in thiouracil is thus
similarly reactive towards electrophiles such as MMS as is the
N1 in adenine and the N7 in guanine. However, the S2 of
thiocytosine is the strongest nucleophile and thus 38% of all s2C
With the synthetic standard in hands, we developed a sensitive
LC-MS/MS method for detection of ms4U in tRNA from
unstressed E. coli. With this targeted analysis, we found a peak
in native tRNA which corresponds to the synthetic ms4U in terms
of retention time, precursor and product ion mass. In E. coli
exposed to the LD50 dose of methyl methanesulfonate (MMS),
the peak increased. A co-injection of the synthesized ms4U
2
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000272
ChemBioChem
COMMUNICATION
become methylated to ms2C in tRNA from MMS exposed E. coli.
Due to the importance of s2C during translation, where it negates
the wobble inosine binding to codons starting with adenine [9], its
comparable to the decrease found for ms2C (Fig. 4c). For ms2C,
we observe a slower repair in the absence of AlkB. Intriguingly,
ms4U loss is independent of AlkB. We conclude that AlkB is not
the demethylase of ms4U, which opens the way for two
hypotheses. The first revolves around a potential, undescribed
demethylase or dethiomethylase, which has ms4U-damaged
tRNA as substrate. SelU, a dethiogeranylase might be a
potential candidate for this reaction [17]). From a chemical
perspective, a direct dethiomethylation through attack of a
nucleophile such as water is also theoretically possible. In both
scenarios, ms4U would dethiomethylate to uridine, which is
again substrate for enzymatic thiolation. The re-thiolation during
the recovery phase can be monitored by analysis of [34S]
incorporation into original tRNA. Our NAIL-MS study, indicates
indeed an increased formation of [34S]-ms4U in original tRNA
from MMS stressed compared to unstressed bacteria (Fig. 4d).
This in vivo data hints towards a dethiomethylation of damaged
tRNA which results in uridine.
efficient repair by enzymatic demethylation has been reported [8]
.
s4U is found at position 8 in 60 % [1] of all bacterial tRNAs and in
addition at position 9 in tRNATyr
from E. coli. The chemical
QUA
properties of sulfur are exploited by the bacteria for oxidative
stress sensing through e.g. UV irradiation. Oxidative stress can
be triggered by UV irradiation following the iron dependent
Fenton chemistry. Therefore, s4U acts as a sensor for UV
irradiation [15], which leads to delayed growth of bacteria during
UV light exposure [16]. Given this important function of s4U, we
were wondering how cells react to tRNAs which have been
methylated and carry ms4U. For this purpose, we designed a
pulse chase study based on our NAIL-MS expertise.
While we cannot exclude the involvement of an unknown
dethiomethylase, we tested the possibility of spontaneous ms4U
dethiomethylation. For this purpose, we simulated potential
cellular environments and exposed synthesized ms4U as free
nucleoside prior to quantitative LC-MS/MS analysis (Fig. S2). A
dethiomethylation was observed after incubation with DTT
(dithiothreitol). No dethiomethylation was observed under
acidic/alkaline conditions, in growth medium or in the presence
of cysteine or BSA (bovine serum albumin as an exemplary
protein).
In summary, we describe the existence of thiomethylated s4U in
bacterial tRNA. The low abundance of ms4U indicates its
formation as a lesion through the constantly present electrophile
S-adenosylmethionine. During exposure of bacteria to
methylating agents, such as MMS, RNA is damaged and the
methylation products of canonical nucleosides (m1A, m7G, m3C,
m3U and m6A) emerge.
In addition, modified nucleosides with a pronounced nucleophilic
character, such as s2C and s4U, become methylated. As evident
from Fig. 3 (right), s2C is more prone to direct methylation
compared to s4U. This can be explained by both, the chemical
reactivity of the S2 in cytidine compared to the S4 in uridine and
its location within the tRNA. Due to the exocyclic amine in
cytidine, s2C has an increased electron density which improves
its nucleophilic character over the S4 in uridine. The uridine S4
is furthermore more prone to solvation which further decreases
its nucleophilicity. In addition to the difference in nucleophilicity,
s2C is exposed and accessible in the anticodon-loop of the tRNA,
while s4U is buried in the D-/T-loop fold.
Figure 4. a Principle of a pulse-chase NAIL-MS experiment. The bacteria are
grown in unlabeled media before and after exposure to MMS. After 1 h MMS
exposure, the media is replaced with
[ [
15N], 34S] and [CD3]-methionine
containing medium. b Formation and loss during recovery of ms4U after 20
mM MMS exposure in wildtype (wt, green) and AlkB deficient (ΔAlkB, grey)
E.coli. c Formation and loss during recovery of ms2C after 20 mM MMS
exposure in wildtype (wt, green) and AlkB deficient (ΔAlkB, grey) E.coli. d
Abundance of [34S]-ms4U after 5 hours in control (ctrl) and MMS exposed wt
and AlkB deficient bacteria. All data from 3 biol. replicates. Error bars
represent standard deviation.
The goal of this assay is to discriminate the damaged tRNAs
and exclude signals from tRNAs transcribed during recovery
from MMS stress. Thus, we can follow the metabolic fate of
ms4U/rN independently from dilution by transcription. For this
purpose, cells are grown in medium containing only [14N] and
[32S]. Consequently, the RNA is completely labeled with [14N]
and all s4U have a [32S] label (original s4U), e.g. m/z (s4U) 261.
In this medium, the bacteria are exposed to MMS (20 mM) and
s4U is converted to ms4U and e.g. A to m1A. After exposure,
MMS is removed by exchanging the medium with stable
isotopes containing medium. During the following recovery
period, newly transcribed tRNA will be [15N] labeled,
enzymatically methylated nucleosides will be [CD3] labeled and
new s4U will have a [34S] label (new s4U, m/z 265 and new m1A,
m/z 290). The experimental design is shown in Fig. 4a. Using
LC-MS/MS analysis, we detect the formation of ms4U during
MMS exposure with around 50 x 10-6 ms4U/original rN. In the
subsequent recovery period, we trace the abundance of ms4U
and normalize to the abundance of original rN. In wildtype E. coli,
we see a constant decrease of ms4U over time (Fig. 4b) which is
Our studies reveal a differential reaction of the cells towards
these RNA damages. One class of lesions is repaired through
enzymatic demethylation using an oxidative demethylation
mechanism. Namely, m1A, m3C (Fig. S3 a/b) and ms2C (Fig. 4c)
are substrate to enzymatic demethylation through AlkB. The
second class comprises lesions, which are lost from the RNA
over time, but in an AlkB independent manner (ms4U and m6A).
The third class of RNA damage comprises m7G, which is not
removed from tRNA (Fig. S3 d).
Overall, the finding of ms4U as a natural and stress induced
lesion in bacterial tRNA confirms the importance of tRNA
modifications during stress response.
3
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000272
ChemBioChem
COMMUNICATION
Acknowledgements
V.R. and S.K. are grateful for funding from the Fonds der
chemischen Industrie. The project is financially supported by the
Deutsche Forschungsgemeinschaft [Projektnummer 325871075,
SFB 1309]. S.K. and C.B. thank Mark Helm for lab and fume
hood access. S.K. thanks Thomas Carell and his group for
instrument time and advice.
Keywords: tRNA • modified nucleoside • RNA damage • NAIL-
MS • epitranscriptome
[1]
P. Boccaletto, M. A. Machnicka, E. Purta, P. Piatkowski, B.
Baginski, T. K. Wirecki, V. de Crecy-Lagard, R. Ross, P. A.
Limbach, A. Kotter, M. Helm, J. M. Bujnicki, Nucleic Acids
Res 2018, 46, D303-D307.
[2]
G. Thomas, A. Favre, Biochem Biophys Res Commun
1975, 66, 1454-1461.
[3]
[4]
A. Caldeira de Araujo, A. Favre, EMBO J 1986, 5, 175-179.
S. Kimura, M. K. Waldor, Proc Natl Acad Sci U S A 2019,
116, 1394-1403.
[5]
[6]
K. Schmid, M. Adobes-Vidal, M. Helm, Bioconjug Chem
2017, 28, 1123-1134.
V. A. Herzog, B. Reichholf, T. Neumann, P. Rescheneder,
P. Bhat, T. R. Burkard, W. Wlotzka, A. von Haeseler, J.
Zuber, S. L. Ameres, Nat Methods 2017, 14, 1198-1204.
A. Lusser, C. Gasser, L. Trixl, P. Piatti, I. Delazer, D.
Rieder, J. Bashin, C. Riml, T. Amort, R. Micura, Methods
Mol Biol 2020, 2062, 191-211.
[7]
[8]
[9]
V. F. Reichle, D. P. Petrov, V. Weber, K. Jung, S. Kellner,
Nat Commun 2019, 10, 5600.
S. Vangaveti, W. A. Cantara, J. L. Spears, H. DeMirci, F.
V. t. Murphy, S. V. Ranganathan, K. L. Sarachan, P. F.
Agris, J Mol Biol 2020, 432, 913-929.
[10]
S. R. Shaver, W. Pendergast, S. M. Siddiqi, B. R. Yerxa, D.
K. Croom, R. W. Dougherty, M. K. James, A. N. Jones, J.
L. Rideout, Nucleosides and Nucleotides 1997, 16, 1099-
1102.
[11]
[12]
I. L. Doerr, J. J. Fox, J Org Chem 1967, 32, 1462-1471.
G. Zemplén, A. Gerecs, I. Hadácsy, Berichte der
deutschen chemischen Gesellschaft (A and B Series)
1936, 69, 1827-1829.
[13]
[14]
V. F. Reichle, V. Weber, S. Kellner, Chembiochem 2018,
19, 2575-2583.
K. Borland, J. Diesend, T. Ito-Kureha, V. Heissmeyer, C.
Hammann, A. H. Buck, S. Michalakis, S. Kellner, Genes
(Basel) 2019, 10.
[15]
[16]
[17]
A. Favre, A. M. Michelson, M. Yaniv, J Mol Biol 1971, 58,
367-379.
T. V. Ramabhadran, T. Fossum, J. Jagger, J Bacteriol
1976, 128, 671-672.
M. Sierant, G. Leszczynska, K. Sadowska, P. Komar, E.
Radzikowska-Cieciura, E. Sochacka, B. Nawrot, FEBS
Lett 2018, 592, 2248-2258.
4
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000272
ChemBioChem
COMMUNICATION
Entry for the Table of Contents
Through combination of synthetic and analytical chemistry, the structure of the modified
nucleoside 4-methylthiouridine was confirmed in bacterial tRNA. The modification is of low
abundance and forms additionally after bacterial stress induced by methylating agents
which argues towards its nature as a natural tRNA lesion. By sophisticated pulse chase
analysis we show its repair through a dethiomethylation mechanism in vivo.
Institute and/or researcher Twitter usernames: @KellnerLab
5
This article is protected by copyright. All rights reserved.