A Biocatalytic Cascade for Versatile One-Pot Modification of mRNA Starting from Methionine Analogues

Article abstract of DOI:10.1002/anie.201507577

Methyltransferases have proven useful to install functional groups site-specifically in different classes of biomolecules when analogues of their cosubstrate S-adenosyl-l-methionine (AdoMet) are available. Methyltransferases have been used to address different classes of RNA molecules selectively and site-specifically, which is indispensable for biophysical and mechanistic studies as well as labeling in the complex cellular environment. However, the AdoMet analogues are not cell-permeable, thus preventing implementation of this strategy in cells. We present a two-step enzymatic cascade for site-specific mRNA modification starting from stable methionine analogues. Our approach combines the enzymatic synthesis of AdoMet with modification of the 5′ cap by a specific RNA methyltransferase in one pot. We demonstrate that a substrate panel including alkene, alkyne, and azido functionalities can be used and further derivatized in different types of click reactions.

Full text of DOI:10.1002/anie.201507577

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201507577
German Edition: DOI: 10.1002/ange.201507577
mRNA Modification
A Biocatalytic Cascade for Versatile One-Pot Modification of mRNA
Starting from Methionine Analogues
Fabian Muttach and Andrea Rentmeister*
Dedicated to Professor Ulrich Hahn on the occasion of his 65th birthday
Abstract: Methyltransferases have proven useful to install
functional groups site-specifically in different classes of
biomolecules when analogues of their cosubstrate S-adeno-
syl-l-methionine (AdoMet) are available. Methyltransferases
have been used to address different classes of RNA molecules
selectively and site-specifically, which is indispensable for
biophysical and mechanistic studies as well as labeling in the
complex cellular environment. However, the AdoMet ana-
logues are not cell-permeable, thus preventing implementation
of this strategy in cells. We present a two-step enzymatic
cascade for site-specific mRNA modification starting from
stable methionine analogues. Our approach combines the
enzymatic synthesis of AdoMet with modification of the 5’ cap
by a specific RNA methyltransferase in one pot. We demon-
strate that a substrate panel including alkene, alkyne, and azido
functionalities can be used and further derivatized in different
types of click reactions.
mentioned approach is to deliver AdoMet analogues inside
a mammalian cell. The limited stability of AdoMet in aqueous
solution aggravates this problem.[6] Moreover, chemical syn-
thesis of AdoMet analogues requires acidic conditions and
results in a mixture of epimers at the sulfur atom, with only
the S,S epimer being accepted as a substrate of methyltrans-
ferases.[7]
These limitations of synthetic AdoMet analogues may be
addressed by switching to the biosynthetic route. In nature,
AdoMet is produced from methionine and ATP, and the
reaction is catalyzed by methionine adenosyltransferases
(MATs).[8] MATs have been used to produce enantiomeri-
cally pure AdoMet both by the isolated enzyme and by
microbial fermentation.[9] Methionine analogues have been
used in MAT-catalyzed reactions to investigate quorum
sensing,[10] alkyl randomization of secondary metabolites,[11]
and to identify methylation sites of protein arginine methyl-
transferases in chromatin modification,[12] but not for nucleic
acids.
Herein we report the enzymatic production of AdoMet
analogues coupled to an RNA methyltransferase. By using
this enzymatic cascade, we can directly and site-specifically
modify eukaryotic model mRNAs at their 5’ cap starting from
methionine analogues instead of synthetic AdoMet ana-
logues, and subsequently label the cap structure by using
different types of click reactions.
To realize this approach for a representative substrate
panel, we required both a MAT and an RNA methyltransfer-
ase with significant activities on the respective substrate
analogues. We used a variant of the RNA methyltransferase
GlaTgs2 (GlaTgs-Var) from Giardia lamblia, which is known
to transfer allyl, alkyne, azido, or benzyl moieties to the N2-
position of the mRNA 5’ cap.[4a,d,13]
S
-Adenosyl-l-methionine (AdoMet or SAM) is the second
most abundant cosubstrate used in nature after ATP, and is
the major source of methyl groups in biomolecules.[1]
Methylation is involved in fundamental biological processes,
such as bacterial host defense, epigenetic silencing, and
cellular signaling, and is found in all classes of biopolymers
as well as lipids and many small molecules.[1,2] Recent studies
on different methyltransferases revealed that these enzymes
can show remarkable substrate promiscuity and accept
synthetic AdoMet analogues.[2d,3] In several cases, the pro-
miscuous activity was significantly enhanced by single amino
acid substitutions.[4] The efficient transfer of a range of
functional groups to methyltransferase substrates allows for
subsequent highly selective conversion through various click
reactions.[5]
However, AdoMet and its analogues are not cell-perme-
able, and dedicated transporters in mammals have not been
described. Therefore, a major limitation of the above-
The coproduct, S-adenosylhomocysteine (SAH), inhibits
methyltransferases, but can be enzymatically degraded in the
same pot.[14] To produce the required AdoMet analogues
in situ, we explored the substrate promiscuity of recombi-
nantly produced human MATIIa I117A (MAT-Var), which
was reported to accept longer alkyl side chains bearing
methionine derivatives (see Figure S1 in the Supporting
Information).[12b] Since MAT shows strong product inhibi-
tion,[8,9b] we set up an enzyme-coupled reaction to directly
consume AdoMet (or its analogues) and label the mRNA
5’ cap (Scheme 1).
[*] F. Muttach, Prof. Dr. A. Rentmeister
University of Münster, Department of Chemistry
Institute of Biochemistry
Wilhelm-Klemm-Strasse 2, 48149 Münster (Germany)
E-mail: a.rentmeister@uni-muenster.de
Prof. Dr. A. Rentmeister
Cells-in-Motion Cluster of Excellence (EXC 1003 - CiM)
University of Münster (Germany)
We first established our coupled MAT/GlaTgs system
using the natural substrate l-methionine, but also d-methio-
nine and a racemic mixture of both, and the mRNA cap
analogue m7GpppA as methyl acceptor (Figure S2). As
Supporting information (including experimental details) for this
article is available on the WWW under http://dx.doi.org/10.1002/
anie.201507577.
Angew. Chem. Int. Ed. 2016, 55, 1917 –1920
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1917

Angewandte
Communications
Chemie
The longer pentenynyl and hexenynyl
groups of 1d and 1e, respectively, were also
transferred. The conversion could be pushed to
35% and 100% at 7 mol% catalyst, respec-
tively. In contrast to Se-1c, the selenium-based
amino acid Se-1d increased the yield of 3d only
moderately compared to its sulfur-based coun-
terpart 1d. Interestingly, despite its longer side
chain, 1e gave higher yields than 1d, which has
also been observed in combination with protein
methyltransferases.[12b] Aromatic groups, how-
ever, turned out to be poor substrates. The
benzyl-bearing methionine analogue 1g gave
only 2–5% conversion, and the para-substi-
tuted 1h yielded only traces of 3h.
Scheme 1. Schematic overview of the enzymatic modification of eukaryotic mRNA
bearing a 5’ cap starting from racemic methionine analogues. An AdoMet analogue 2 is
generated in situ from 1 and ATP using MAT-Var and serves as a cosubstrate for the
RNA methyltransferase GlaTgs-Var, thereby leading to modification of the 5’ cap at
position N2. The introduced groups (R) can be further treated in different types of click
reactions to introduce affinity or fluorescent tags.
After establishing this biocatalytic cascade
for site-specific modification of the 5’ cap we
wanted to test whether the functional groups
expected, l-methionine led to the formation of m2,7GpppA,
Table 1: One-pot enzymatic modification of m7GpppA starting from S-
and Se-based methionine analogues.[a]
whereas d-methionine was not consumed but did not com-
promise the conversion, thereby allowing us to use racemic
methionine analogues in the following experiments.[15]
Next, we used the enzymatic cascade to convert methio-
nine analogues and modify m7GpppA with different func-
tional groups site-specifically at N2. We synthesized a panel of
S-derivatized homocysteines bearing different alkene, alkyne,
and azido groups (1a–f) as well as benzylic moieties (1g,h;
Table 1, and Figure S3). Furthermore, we synthesized two
analogues based on selenohomocysteine because the resulting
Se-AdoMet analogues are more efficient at side-chain trans-
fer.[6b,16]
The racemic methionine analogues were directly con-
verted in one pot containing both enzymes as well as ATP and
m7GpppA. Typically, 300 mm m7GpppA was treated with an 8-
fold excess of ATP and a 20-fold excess of racemic methionine
analogues at a catalyst loading of 3–7 mol% MAT-Var
(relative to m7GpppA). The GlaTgs-Var concentration was
typically three times higher than the MAT-Var concentration
to ensure that in situ formed AdoMet analogues are directly
converted.[4a,12a] This is important because MAT generally
shows strong product inhibition by AdoMet.[9b] In comparison
to synthetic AdoMet analogues, only one epimer is formed
(Figure S4).
We observed the transfer of allyl, alkyne, and azido
moieties with different steric demands to m7GpppA with
moderate to excellent yields (Table 1, Figure 1, and Fig-
ure S5). The small allyl group in 1b was transferred almost
quantitatively at 3 mol% MAT-Var. The chemically more
interesting propargyl group of 1c, however, resulted in only
11% conversion, even at 7 mol% MAT-Var, most likely
because of the low stability of the in situ formed AdoMet
analogue 2c.[6b] This limitation was solved by switching to the
selenium-based methionine analogue. Se-1c drastically
increased the conversion, and 91–100% of N2-propargyl-
m7GpppA could be obtained at 3–7 mol% MAT-Var. These
results are in line with results using chemically synthesized Se-
2c.[6b]
Entry
Substr.
MAT [mol%]
Prod.
Conv. [%]
1
2
3
4
5
6
7
8
1a
1b
1b
1c
3
3
5
3
7
3
5
7
3
7
3
7
3
5
7
3
5
3
7
3
3a
3b
3b
3c
3c
3c
3c
3c
3d
3d
3d
3d
3e
3e
3e
3 f
100
95
100
4
11
91
100
100
10
35
17
47
38
77
100
12
72
2
5
1c
Se-1c
Se-1c
Se-1c
1d
9
10
11
12
13
14
15
16
17
18
19
20
1d
Se-1d
Se-1d
1e
1e
1e
1 f
1 f
1g
1g
3 f
3g
3g
3h
1h
traces
[a] Conversions at different catalyst loadings (relative to m7GpppA) are
shown. Reaction conditions: 6 mm d/l-1, 2.5 mm ATP, 300 mm
m7GpppA, MAT-Var as indicated, GlaTgs-Var 3” higher than MAT-Var,
incubation for 8–9 h at 238C.
1918
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1917 –1920

Angewandte
Communications
Chemie
3 f at 5 mol% MAT and the product was validated by
MALDI-TOF-MS (Table 1, Figure 1, and Figure S14). The
subsequent SPAAC reaction of 3 f with dibenzocyclooctyne-
sulforhodamine B (DBCO-SRB) yielded a fluorescently la-
beled 5’ cap, which showed a fluorescent band after PAGE
separation only if all the components were present (Fig-
ure S15). Again, enzymatic transfer was also successful in
eukaryotic cell lysate (Figure S15).
To ensure that our approach is suitable for labeling
eukaryotic mRNAs, we produced 24 and 106 nt long RNAs by
in vitro transcription and capping with the vaccinia capping
system[17] and subjected them to our biocatalytic MAT/
GlaTgs cascade (Figure 3 and Figure S22). Starting from 1e,
subsequent labeling with Cy5-azide resulted in specific
fluorescent modification of alkyne-modified RNA (Fig-
ure 3a, see Figure S16 for the 106 nt RNA). Biotin-azide
reacted in a similar manner (Figure S17).
Figure 1. Representative HPLC chromatograms of selected reactions
from Table 1. Neg.: negative control without 1.
appended to the N2-position are suitable handles for subse-
quent derivatization. Therefore, we treated 3c (from con-
version of Se-1c) and 3e with biotin-azide in a copper(I)-
catalyzed azide–alkyne cycloaddition (CuAAC; Figure 2a)
Figure 3. Labeling of a 24 nt 5’-capped model mRNA starting from
methionine analogues 1e or 1 f followed by CuAAC or SPAAC,
respectively. a) Gel analysis after CuAAC by SYBR staining and
fluorescence (lex =628 nm, lem =716 nm). As a control, 5’-capped
RNA that was not treated with 1e is shown. b) Gel analysis of the
labeled 24 nt model mRNA after SPAAC by SYBR staining and
fluorescence (lex =520 nm, lem =609 nm). As a control, 5’-capped
RNA that was not treated with 1 f is shown. Bromophenol blue was
used as a dye marker (lmax =590 nm). Resolution on a 15% denatur-
ing PAGE (1.5 h at 25 W).
Figure 2. Postsynthetic labeling of alkyne-functionalized m7GpppA.
a) Enzymatic introduction of different terminal alkynes followed by
CuAAC with different labels R’. b,c) HPLC and MS analysis confirming
CuAAC of 3e with biotin-azide. d) In-gel fluorescence of Cy5-4e after
separation on a 10% denaturing PAGE (1.5 h at 25 W).
Starting from 1 f, the SPAAC reaction was also established
with DBCO-SRB for the 24 nt 5’-capped RNA. The resulting
SRB-labeled RNA was analyzed by in-gel fluorescence. A
comparison with RNA staining by SYBR revealed that the
SPAAC reaction also works with 5’-capped RNA modified in
an enzymatic cascade starting from 1 f (Figure 3b). Mass
analysis of unmodified 24 nt RNA and capped 24 nt RNA is
shown in the Supporting Information (Figure S18).
In summary, our cascade approach allows for the site-
specific transfer of a variety of chemical functionalities to the
mRNA cap structure starting from stable and easy to
synthesize analogues of the amino acid methionine. Alkyne-
bearing side chains of different lengths and azido moieties
were efficiently transferred to the 5’ cap as well as 5’-capped
RNA. Subsequent click reactions gave access to biotin- and
fluorophore-modified mRNAs. Since methionine analogues
can be taken up by the cell[12b, 18] and the enzymatic steps also
work selectively in eukaryotic cell lysate, we anticipate that
our approach will enable the intracellular labeling of RNA. In
combination with other RNA or DNA methyltransferases,
our approach should provide a general concept towards the
site-specific intracellular modification of nucleic acids. In light
and confirmed the formation of the expected products, biotin-
4c and biotin-4e, respectively, by HPLC and MALDI-TOF-
MS (Figure 2b,c and Figures S10 and S11). Two fluorescent
dyes, Cy5-azide and Eterneon-azide, were also treated with
3c and 3e, respectively. Gel analysis revealed that a fluores-
cent band was only obtained if both enzymes and the required
amino acid (1e or Se-1c) were present (Figure S12A,B).
Importantly, we could also perform the enzymatic cascade
reaction in eukaryotic cell lysate, which indicates that the
enzymatic cascade also works selectively in the presence of
cellular components (Figure 2d, Figure S13).
To make our approach suitable for intracellular applica-
tions, we wanted to introduce a functional group that reacts in
a truly bioorthogonal reaction. We devised methionine
analogue 1 f to transfer an azido moiety, which can be reacted
in a strain-promoted azide–alkyne cycloaddition (SPAAC).
The reaction of 1 f with m7GpppA resulted in a 72% yield of
Angew. Chem. Int. Ed. 2016, 55, 1917 –1920
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1919

Angewandte
Communications
Chemie
1056; b) D. Stummer, C. Herrmann, A. Rentmeister, Chemis-
tryOpen 2015, 4, 295 – 301; c) J. M. Holstein, D. Stummer, A.
Rentmeister, Chem. Sci. 2015, 6, 1362 – 1369.
of the recent discoveries of reversible adenosine methylation
in bacterial and eukaryotic mRNA,[19] we anticipate a broad
range of applications in the near future.
ˇ
ˇ
˙
[6] a) G. Lukinavicius, M. Tomkuviene, V. Masevicius, S. Klima-
ˇ
sauskas, ACS Chem. Biol. 2013, 8, 1134 – 1139; b) S. Willnow, M.
Martin, B. Lüscher, E. Weinhold, ChemBioChem 2012, 13,
1167 – 1173; c) C. Dalhoff, G. Lukinavicius, S. Klimasauakas, E.
Weinhold, Nat. Protoc. 2006, 1, 1879 – 1886.
Acknowledgements
[7] P. K. Chiang, R. K. Gordon, J. Tal, G. C. Zeng, B. P. Doctor, K.
Pardhasaradhi, P. P. McCann, FASEB J. 1996, 10, 471 – 480.
[8] G. D. Markham, M. A. Pajares, Cell. Mol. Life Sci. 2009, 66, 636 –
648.
[9] a) J. Chu, J. Qian, Y. Zhuang, S. Zhang, Y. Li, Appl. Microbiol.
Biotechnol. 2013, 97, 41 – 49; b) M. Kotb, N. M. Kredich, J. Biol.
Chem. 1985, 260, 3923 – 3930.
[10] S. P. Zano, P. Bhansali, A. Luniwal, R. E. Viola, Arch. Biochem.
Biophys. 2013, 536, 64 – 71.
[11] S. Singh, J. J. Zhang, T. D. Huber, M. Sunkara, K. Hurley, R. D.
Goff, G. J. Wang, W. Zhang, C. M. Liu, J. Rohr, S. G. Van Lanen,
A. J. Morris, J. S. Thorson, Angew. Chem. Int. Ed. 2014, 53,
3965 – 3969; Angew. Chem. 2014, 126, 4046 – 4050.
A.R. thanks the DFG for support by an Emmy Noether
fellowship (RE 2796/2-1) and the Fonds der Chemischen
Industrie. F.M. is supported by the DFG Priority Programme
SPP 1784 (RE 2796/3-1). We thank Dr. Daniela Stummer,
Stephan H. H. Schiefelbein, Henning Klaasen, Kathrin
Kirchhoff, and Dr. Wolfgang Dçrner for excellent technical
assistance. The mass spectrometry and NMR facilities of the
organic chemistry department are gratefully acknowledged
for analytical services.
Keywords: 5’ cap · click chemistry · labeling · RNA ·
SAM synthetase
[12] a) R. Wang, W. Zheng, M. Luo, Anal. Biochem. 2014, 450, 11 –
19; b) R. Wang, K. Islam, Y. Liu, W. Zheng, H. Tang, N. Lailler,
G. Blum, H. Deng, M. Luo, J. Am. Chem. Soc. 2013, 135, 1048 –
1056.
How to cite: Angew. Chem. Int. Ed. 2016, 55, 1917–1920
Angew. Chem. 2016, 128, 1951–1954
[13] J. M. Holstein, D. Schulz, A. Rentmeister, Chem. Commun.
2014, 50, 4478 – 4481.
[14] D. Schulz, A. Rentmeister, RNA Biol. 2012, 9, 577 – 586.
[15] C. T. Holloway, R. C. Greene, C. H. Su, J. Bacteriol. 1970, 104,
734 – 747.
[16] a) M. Tomkuviene˙, B. Clouet-d’Orval, I. Cerniauskas, E. Wein-
hold, S. Klimasauskas, Nucleic Acids Res. 2012, 40, 6765 – 6773;
b) I. R. Bothwell, M. K. Luo, Org. Lett. 2014, 16, 3056 – 3059.
[17] S. Shuman, J. Biol. Chem. 1990, 265, 11960 – 11966.
[18] J. Xie, P. G. Schultz, Nat. Rev. Mol. Cell Biol. 2006, 7, 775 – 782.
[19] a) X. Deng, K. Chen, G.-Z. Luo, X. Weng, Q. Ji, T. Zhou, C. He,
Nucleic Acids Res. 2015, 43, 6557 – 6567; b) D. Dominissini, S.
Moshitch-Moshkovitz, S. Schwartz, M. Salmon-Divon, L. Ungar,
S. Osenberg, K. Cesarkas, J. Jacob-Hirsch, N. Amariglio, M.
Kupiec, R. Sorek, G. Rechavi, Nature 2012, 485, 201 – 206;
c) K. D. Meyer, Y. Saletore, P. Zumbo, O. Elemento, C. E.
Mason, S. R. Jaffrey, Cell 2012, 149, 1635 – 1646.
[1] M. Fontecave, M. Atta, E. Mulliez, Trends Biochem. Sci. 2004,
29, 243 – 249.
[2] a) L. A. Wessjohann, J. Keim, B. Weigel, M. Dippe, Curr. Opin.
Chem. Biol. 2013, 17, 229 – 235; b) K. K. Biggar, S. S. C. Li, Nat.
Rev. Mol. Cell Biol. 2015, 16, 5 – 17; c) A. Jeltsch, ChemBioChem
2002, 3, 274 – 293; d) Y. Motorin, J. Burhenne, R. Teimer, K.
Koynov, S. Willnow, E. Weinhold, M. Helm, Nucleic Acids Res.
2011, 39, 1943 – 1952; e) E. Staudacher, Biol. Chem. 2012, 393,
675 – 685; f) D. E. Vance, Biochim. Biophysi. Acta Biomembr.
2014, 1838, 1477 – 1487.
[3] C. Dalhoff, G. Lukinavicius, S. Klimasauskas, E. Weinhold, Nat.
Chem. Biol. 2006, 2, 31 – 32.
[4] a) J. M. Holstein, D. Stummer, A. Rentmeister, Protein Eng.
Des. Sel. 2015, 28, 179 – 186; b) K. Islam, W. Zheng, H. Yu, H.
Deng, M. Luo, ACS Chem. Biol. 2011, 6, 679 – 684; c) R. Wang,
W. Zheng, H. Yu, H. Deng, M. Luo, J. Am. Chem. Soc. 2011, 133,
7648 – 7651; d) D. Schulz, J. M. Holstein, A. Rentmeister,
Angew. Chem. Int. Ed. 2013, 52, 7874 – 7878; Angew. Chem.
2013, 125, 8028 – 8032.
Received: August 13, 2015
[5] a) R. Wang, K. Islam, Y. Liu, W. Zheng, H. Tang, N. Lailler, G.
Blum, H. Deng, M. Luo, J. Am. Chem. Soc. 2013, 135, 1048 –
Revised: October 22, 2015
Published online: December 22, 2015
1920 www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1917 –1920