RNA Ligation for Mono and Dually Labeled RNAs

Article abstract of DOI:10.1002/chem.202101909

Labeled RNAs are invaluable probes for investigation of RNA function and localization. However, mRNA labeling remains challenging. Here, we developed an improved method for 3′-end labeling of in vitro transcribed RNAs. We synthesized novel adenosine 3′,5′

Full text of DOI:10.1002/chem.202101909

Accepted Article
Accepted Article
Title: RNA Ligation for Mono and Dually Labeled RNAs
Authors: Anaïs Depaix, Agnieszka Mlynarska-Cieslak, Marcin
Warminski, Pawel J Sikorski, Jacek Jemielity, and Joanna
Kowalska
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: Chem. Eur. J. 10.1002/chem.202101909
Link to VoR: https://doi.org/10.1002/chem.202101909
01/2020

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
RNA Ligation for Mono and Dually Labeled RNAs
Anaïs Depaix,[a] Agnieszka Mlynarska-Cieslak,[a] Marcin Warminski,[a] Pawel J. Sikorski,[b] Jacek
Jemielity,[b] and Joanna Kowalska*[a]
[a] Dr. A. Depaix, M. Sc. A. Mlynarska-Cieslak, Dr. M. Warminski, Dr. J. Kowalska
Division of Biophysics, Institute of Experimental Physics, Faculty of Physics, University of Warsaw,
Pasteura 5, 02-093 Warsaw, Poland
E-mail: jkowalska@fuw.edu.pl, Twitter handle: @JoannaK_UW
[b] Dr. P. J. Sikorski, Prof. Dr. J. Jemielity
Centre of New Technologies
University of Warsaw
Banacha 2c, 02-097 Warsaw, Poland
Abstract: Labeled RNAs are invaluable probes for investigation of
RNA function and localization. However, mRNA labeling remains
challenging. Here, we developed an improved method for 3′-end
labeling of in vitro transcribed RNAs. We synthesized novel
adenosine 3′,5′-bisphosphate analogs modified at the N6 or C2
position of adenosine with an azide-containing linker, fluorescent
label, or biotin and assessed these constructs as substrates for RNA
labeling directly by T4 ligase or via postenzymatic strain-promoted
alkyne-azide cycloaddition (SPAAC). All analogs were substrates for
T4 RNA ligase. Analogs containing bulky fluorescent labels or biotin
showed better overall labeling yields than postenzymatic SPAAC.
We successfully labeled uncapped RNAs, NAD-capped RNAs, and
5′-fluorescently labeled m7Gp3Am-capped mRNAs. The obtained
highly homogenous dually labeled mRNA was translationally active
and enabled fluorescence-based monitoring of decapping. This
method will facilitate the use of various functionalized mRNA-based
probes.
chemoenzymatic methods.[9] Among currently available
enzymatic methods, two main approaches have been used. The
first is cotranscriptional labeling, in which a labeled nucleotide is
incorporated into RNA during an in vitro transcription reaction
catalyzed by DNA-dependent RNA polymerase. The second is
post-transcriptional labeling, typically performed using an RNA
ligase
or
template-independent
polymerase,[9f]
which
incorporates a labeled nucleotide at the 3′-end of in vitro
transcribed (IVT) RNA.[10] Ribozymatic posttranscriptional
approaches for internal, sequence-specific RNA labeling have
also been developed.[11]
One of the first developed RNA labeling strategies uses T4
RNA ligase 1, an enzyme known to form a phosphodiester bond
between an (oligo)nucleotide containing a 5′-phosphate group
and an oligonucleotide possessing a free 3′-hydroxyl group.[10; 12]
The minimum substrate that have been attached to mRNA using
this method is radioactively labeled cytidine or adenosine 5′,3′-
bisphosphate (pCp or pAp, respectively). Different pNp
derivatives have been later developed for the purpose of RNA
labeling, including fluorescently tagged pCp analogs,
photocaged nucleotides and isotopically labeled pAp (13C,15N-
pAp).;[13] Chemoenzymatic methods for RNA labeling include
post-transcriptional modification of prefunctionalized RNA using
reactions such as esterification,[14] amide condensation, inverse-
electron demand Diels-Alder cycloaddition, or azide-alkyne
cycloaddition (AAC), which is particularly convenient for
Introduction
RNAs play crucial roles in living organisms and are a promising
class of modern therapeutics. Single- or dual-color fluorescently
labeled or biologically tagged RNA probes have paved the way
for studying the intracellular fates and biological roles of RNAs
using advanced microscopic and analytical methods.[1]
Fluorescently labeled oligoribonucleotides and RNAs are
valuable probes for studying specific ligand or protein binding
and for investigating RNA dynamics, function, and localization.[2]
These molecules have been used for many applications,
including deciphering the biological roles of RNAs, studying viral
introducing various labels in a highly efficient and biocompatible
15]
manner.[9e;
Although the copper-catalyzed version of AAC
(CuAAC) has been extensively used to label RNA, copper ions
are cytotoxic, making this method incompatible with in vivo
imaging.[16] This inconvenience has been circumvented by the
development of dibenzocyclooctylamine derivatives that are
reactive in strain-promoted azide-alkyne cycloaddition, which, in
contrast to CuAAC, does not require any catalyst.[15b; 17]
Various methods have been employed to afford
monolabeled RNAs for versatile applications;[2e; 7a; 7c; 9f; 13b; 15c; 18]
however, there is still room for improvement and for the
development of new methods. Dually labeled RNA probes have
also recently attracted interest owing to their advanced
application in single-molecule spectroscopy and conformational
studies. Labeling of IVT mRNAs is particularly challenging
because of the large size and heterogeneity of the mRNAs and
the tendency of mRNAs to adopt complex tertiary
conformations.[9g] Moreover, mRNAs have atypical 5′ ends,
usually modified by a 7-methylguanosine cap, whereas their 3′
ends are polyadenylated. Both of these terminal structures are
crucial for the activity of mRNAs in translation; therefore, a
4]
mRNA replication pathways,[3] and detecting cancer.[2c;
Fluorescent labeling of RNA enables the use of multiple
techniques (fluorescence intensity changes, anisotropy,
fluorescence resonance energy transfer, or fluorescence
lifetime) to visualize and monitor enzymatic processes with high
sensitivity.[5] Biological tagging with molecules such as biotin
enables the application of various RNA capture techniques,
relying on the high affinity of the tag to another entity (e.g., biotin
and (strept)avidin).[6] Biotin-(strept)avidin technology-based
assays are valuable tools for protein affinity purification, pull-
down experiments, in vivo visualization, and cellular delivery,
among others.[7]
The synthesis of short fluorescent or biologically tagged
RNA probes can be achieved via full chemical synthesis,
whereas longer RNA probes require different enzymatic[8] and
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
desirable mRNA labeling method should not disrupt the
functionalities of these elements. Recently, other types of
unconventional mRNA caps with unknown biological roles have
also been discovered, and molecular tools to investigate them
are also needed.[19] To study the fates of IVT mRNAs and
mRNA-based therapeutics, it is necessary to apply site-specific,
linkers (ethylene or polyethylene glycol; Figures 1A and S1) and
used them in chemoenzymatic reactions to access RNA with
various modifications at their 3′ ends (Figure 1B). The labeling
method was first optimized for short in vitro transcribed RNAs
(35 nucleotides long) and later adapted to longer, variously
capped RNAs, including mRNAs (Figure 1C). We identified
compounds that enabled straightforward incorporation of
fluorescent labels or biotin at the 3′ ends of RNAs, providing
access to structurally diverse RNA probes as valuable research
tools, including 3′-mono- and 3′,5′-dually labeled mRNAs that
retained translational activity.
controllable, and scalable procedures that yield
a highly
homogenous product. Despite recent progress in the
development of mRNA-based vaccines and therapies, methods
for achieving this goal are limited.
Results and Discussion
Synthesis of nucleotides for RNA modification
All pAp analogs were synthesized starting from unprotected
nucleosides in a one-pot three-step reaction yielding the
corresponding 2′,3′-cyclophosphonucleoside 5′-phosphates as
the key intermediates (Scheme 1A). These intermediates were
then subjected to RNAse T2 to achieve regioselective cleavage
leading to the target 3′-phosphonucleoside 5′-phosphates. N6-
substituted pAp analogues 1a and 2a were synthesized starting
from unprotected 6-chloropurineriboside (Scheme 1A). This
latter was substituted at position 6 with an appropriate amino-
azido linker, either 8-azido-3,6-dioxaundecan-1-amine or 2-
azidoethylamine. The resulting N6-functionalized nucleosides
(6a and 7a, respectively) were converted into the corresponding
5′,3′-diphosphates using a one-pot three-step procedure. To this
end, nucleosides were treated with phosphoryl chloride (3
equiv.) to afford the corresponding nucleoside 5′-
dichlorophosphates I, as intermediates. When high-performance
liquid chromatography (HPLC) monitoring showed complete
conversion of the starting material,
phosphoryl chloride was added,
a
second portion of
yielding the
dichlorophosphonucleoside 5′-dichlorophosphate intermediates
II as two (3′ and 2′) regioisomers. The latter were precipitated
with diethyl ether and treated with heterocyclic nucleophiles to
induce cyclization of the 2′/3′-dichlorophosphate with
concomitant substitution of the chlorine atoms, followed by
hydrolysis to the desired 2′,3′-cyclophosphonucleoside 5′-
phosphate. We tested different nucleophiles and bases as
cyclization-inducing
reagents,
including
triethylamine,
diisopropylamine, imidazole, 3-amino-1,2,4-triazole, and 1,2,4-
triazole. The first two did not induce cyclization, whereas,
imidazole and triazole derivatives were effective yielding
corresponding phosphorimidazolide or -triazolide intermediates.
We aimed to select a reagent that enables hydrolysis of all P-N
bonds at neutral and mildly acidic pH (7–3) because at lower pH
values the hydrolysis of the P-N bond was accompanied by
undesired opening of the 2’,3’-cyclophophate back to 2’- and 3’-
phosphates. 1,2,4-Triazole was the optimal cyclization-inducing
agent, as the resulting 5′-phosphorotriazolide (III) was
hydrolyzed within 30 min at pH 5, affording compounds 6b and
7b without any observable formation of by-products resulting
from the 2′,3′-cyclophosphate opening. Recently, an alternative
method providing direct access to 2’,3’-cyclophophonucleoside
5’-phosphates from nucleosides has been reported.[20] Finally,
6b and 7b were selectively converted to the corresponding 3′-
phosphates (1a and 2a) by RNAse T2 cleavage (Figure S2A-D,
Figure 1. A. Adenosine 5′,3′-bisphosphate analogs designed and used in the
ligation step. B. General strategy for generation of 3’-labeled RNA. C.
Derivatives used in IVT as first transcribed nucleotides affording uncapped
(incorporating ATP at the 5’ end) and capped RNAs (incorporating NADpG 4
or FAM-L2N-m7Gp3AmpG 5).
In this study, we revisited the long-known strategy for RNA
3′-end labeling with T4 RNA ligase (LT4) to develop a widely
applicable method suitable for RNA, including mRNA and
labeling. We sought to design new reagents that could serve as
efficient substrates for LT4 applicable to copper-free
functionalization, bio-tagging, and fluorescent labeling of RNA,
including mRNA. Because mRNAs contain polyA tails at the 3’
end, we focused on the development of pAp derivatives. To that
end, we obtained a set of pAp analogs substituted at different
positions (C2 or N6 adenine) with different azide-containing
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
Scheme 1. A. Representative pAp synthesis for N6 derivatives 1a and 2a. B. Representative click reaction with DIBAC-SCy5 yielding compound 1b The
.
structure of SCy5 is shown in Figure 1.
~70% yield) with an overall yield from unprotected nucleoside of
4-7%.
(sulfoCy5, SCy5) pAp analog (2b), which enabled
straightforward visualization of the labeled product after gel
electrophoresis. The labeling efficiency was tested for 35 nt IVT
RNA (at 1 µM) and different excess amounts (from 1- to 200-
fold) of 2b in the presence of 1 U/µL LT4. The RNAs were
purified using a commercially available kit and analyzed by high-
resolution polyacrylamide gel electrophoresis (HRPAGE, 15%
polyacrylamide in the presence of 7 M urea and 1× TBE buffer)
to assess the efficiency of the ligation step. The gels were first
scanned for Cy5 fluorescence (to visualize labeled RNA),
stained with ethidium bromide (EtBr), and scanned for EtBr
fluorescence (to visualize total RNA). The analysis revealed that
the unlabeled RNA substrate was heterogeneous and, in
addition to the expected 35 nt main product, contained
transcripts 1 nt shorter or longer, which were visible on the gel
below and above the most intense band, respectively (Figure 2,
ref). Under the tested conditions, the RNA 3′-end labeling
efficiency was highly dependent on the concentration of 2b
(Figure 2, lanes 1–7, and Figure S3); that is, the labeling
efficiency gradually improved as the concentration of 2b
increased from 1 to 100 µM. Surprisingly, when using a higher
excess (200 equiv., Figure S3), the ligation efficiency was not
improved. To assess the lower limit of RNA concentration, we
explored the applicability of these conditions to the labeling of
RNA at lower concentrations. As shown in Figure 2, lanes 8–11,
the conditions were suitable for 0.5 µM RNA, although the
efficiency of the ligation step was slightly lower, and using 0.1
µM RNA resulted in an even lower efficiency (Figure S3),
The synthesis of the pAp derivative modified at position C2 (3a)
was attempted in a similar way starting from 2-fluoroadenosine
and included aromatic nucleophilic substitution, followed by
exhaustive phosphorylation, 2′,3′-cyclization, and C2-substituted
adenosine was not susceptible to hydrolysis by RNAse T2, even
at much higher concentrations than initially tested. Therefore,
the synthesis was completed by chemical hydrolysis of 2′,3′-
cyclophosphate under acidic conditions, followed by separation
of the resulting 2′- and 3′-isomers by reverse phase (RP)-HPLC
(Figure S2E, F, 8% yield), with an overall 0.3% yield from 2-
fluoroadenosine. Finally, all azido-functionalized derivatives (1a–
3a) were labeled in a strain-promoted azide-alkyne cycloaddition
reaction with DIBAC-functionalized Cy5 or biotin (Scheme 1B,
DIBAC-SCy5) to give the resulting labeled pAp derivatives (1b,
2b, 2c, and 3b; Figure S1) with good yields.
Next, we tested the utility of the synthesized compounds in
LT4-mediated RNA labeling in two different variants. The first
variant involved direct labeling of RNA at the 3′ end using a
fluorescently labeled or biotinylated pAp analog (Figure 1B,
pathway A). The second variant was a one-pot two-step
labelingprocedure, wherein an azido pAp derivative was ligated
to the RNA 3′ end followed by copper-free click chemistry
(SPAAC reaction with DIBAC-SCy5 ) with the resulting RNA-N3
(Figure 1B, pathway B). In the initial experiments, we aimed to
determine the excess pAp analog required for effective ligation
to the RNA 3′ end. To this end, we used a fluorescently labeled
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
suggesting that the RNA concentration should be maintained
higher than 0.5 µM during the labeling procedure.
Overall, SCy5 labeling was more efficient when performed
directly via one-step ligation, than via the two-step procedure.
The latter, albeit slightly more complex, may deliver superior
results in cases in which the label is not accepted by LT4, owing
to its structure or electrostatic properties, or if RNA is to be
labeled in biological mixtures, e.g., for pull-down experiments.
Indeed, the RNA functionalized with derivative 1a containing an
azidoethylene linker could be useful for experiments requiring
labeling of RNA by SPAAC in various biological settings.
The abovementioned labeling reactions were performed on
uncapped (5′-triphosphorylated) RNAs. However, endogenous
RNA molecules are often modified at their 5′ ends by different
canonical and noncanonical cap structures. Therefore, we next
tested whether the labeling procedure was compatible with 5′
capped RNAs. As the first model RNA, we chose NAD-linked
RNA. NAD was recently discovered as a tag of the 5′ ends of
some RNAs in bacteria and eukaryotes.[21] The biological role of
this noncanonical cap is not yet well understood, and developing
NAD-RNA probes is crucial to elucidating its function. Therefore,
we aimed to synthesize NAD-capped RNA modified at the 3′ end
with either a fluorescent label or biotin to provide NAD-RNA-
based fluorescent or affinity probes, respectively. NAD-RNA was
obtained by IVT using NADpG trinucleotide (4), as previously
reported (Scheme S1, Figure S4).[22] The resulting NAD-RNA35
was subjected to reactions with compounds 1b, 2b, 2c, 3b, 1a,
2a, and 3a following SPAAC and analyzed by HRPAGE (Figure
4). The two biotinylated regioisomers of 2c were separated
during HPLC purification (named 2c-I and 2c-II according to the
elution order) and were therefore used separately in the ligation
experiment.
Next, we compared the labeling using compound 2b with
similar procedures using other pAp analogs (1b and 3b) and
with a two-step procedure consisting of labeling RNA with 1a,
2a, and 3a, followed by attachment of sulfo-cyanine 5 (SCy5)
using copper-free click chemistry (SPAAC; Figure 3). The
conditions optimized earlier for 2b were used to compare the
different labeling variants. All azido derivatives were
incorporated into RNA (Figure 3A, lanes 1–3); however, the
structure of the linker unexpectedly influenced the incorporation
of the labeled derivative. Ligation of the N6-substituted labeled
pAp containing a shorter linker (1b; Figure 3A, lane 4) was less
efficient than that with longer linker (2b; Figure 3A, lane 5).
Moreover, we also found that the compound substituted at the
C2 position (3b) afforded a better labeling yield than the
corresponding N6-derivative (2b). RNA-N3 obtained by ligation
with compounds 1a, 2a, or 3a were purified by commercial kit
and subjected to SPAAC with DIBAC-SCy5. The efficiency of
the click reaction depended on the linker length; in this case, the
reaction was most efficient for the derivative with the shorter
linker
Figure 2. Optimization of LT4-mediated RNA (35 nt) labeling with compound
2b. HRPAGE (15% PAA, 1× TBE) of the resulting RNAs (100 ng) after ligation
with pAp 2b on uncapped RNA35
.
(1a; Figure 3A, lane 7 versus lanes 8 and 9). Furthermore,
because the C2-substituted compound 3b was most efficiently
incorporated at the RNA 3′ end, we tested whether the
compound was more suitable for labeling RNA at lower
concentrations (Figure 3B). Even under the most restrictive
conditions, that is, 0.1 µM RNA and 10 µM 3b (100-fold excess,
lane 1), 3b was incorporated with high efficiency (~70%, Figure
3B, lane 1), which was more than 2-fold higher than that of
compound 2b (~30%, Figure S3). These results suggested that
C2-modified pAp analogs, despite being synthetically more
challenging, may be analogs of choice for labeling procedures
performed at low RNA concentrations (e.g., if the amount of
RNA material is limited).
Figure 3. A. HRPAGE (15% PAA, 7 M urea, 1× TBE)-labeling of uncapped
RNA35, direct ligation versus ligation followed by SPAAC. B. HRPAGE (15%
PAA, 7 M urea, 1× TBE)-labeling of uncapped RNA35 using pAp 3b.
All azido-modified pAp analogs (1a, 2a, and 3a) were
efficiently incorporated at the 3′-end of NAD-RNA (Figure 4,
lanes 1–3), and incorporation of the SCy5-labeled derivatives
was found to be dependent on the linker length and the position
of the substitution, similar to uncapped RNAs (Figure 4, lanes 4–
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
6). We also found that the nature of the label (SCy5 or biotin) did
not influence the efficacy of the ligation (Figure 4, lanes 10 and
11 versus lane 5). The SPAAC reactions on the resulting azido-
substituted NAD-RNAs were again the most efficient for RNA
derivatives with shorter linkers (Figure 4, lane 7 versus lanes 8
and 9). In agreement with the findings for uncapped RNA,
labeling was more efficient in the case of direct ligation using a
labeled adenosine analog and most efficient when using the C2-
substituted derivative 3b (Figure S5). Overall, the results were
similar to those for uncapped RNA, indicating that the labeling
procedure using pAp derivatives was compatible with RNAs
modified at the 5′ end.
good yield and high purity. The separation was efficient enough
to allow for the recovery of unreacted 5′-labeled RNA.
Dually labeled RNAs are promising biological tools because
they can be visualized using two different fluorescence
channels. Such probes could benefit several intensively
investigated research areas, such as RNA decapping,
intracellular transport of different types of RNA, RNA
localization, storage, and conformational dynamics.
Because the 3′-end labeling procedure was compatible with
5′ capped RNAs, we envisaged that this approach could also be
combined with mRNA 5′-end labeling to afford dually labeled
RNAs. To demonstrate this, we first performed labeling on a 35-
nt RNA substrate carrying a fluorescently labeled m7G cap. m7G
is a canonical cap present at the 5′ end of eukaryotic mRNAs.
m7G-capped mRNAs can be efficiently prepared by IVT in the
presence of trinucleotide cap analogs, such as m7GpppAmpG.[23]
Here, we designed a fluorescent analog of this trinucleotide to
obtain a 5′-capped and labeled RNA substrate, which could be
converted into a dually labeled 5′-capped RNA probe by ligation
with a fluorescent pAp analog.
Figure 5. A. HPLC profile of dually labeled RNA35 purification; B. HRPAGE
analysis (15% PAA, 7 M urea, 1× TBE). 1: RNA35 after ligation with 2b and
before HPLC purification. 2: First collected fraction (retention time, 5.7 min;
recovered substrate). 3: Second collected fraction (retention time, 8.8 min;
purified dually labeled RNA35).
Therefore, we next applied our methodology, i.e., in vitro
cotranscriptional 5′-capping/labeling followed by 3′-end post-
transcriptional ligation/labeling, to mRNA. Using the appropriate
DNA plasmid as a template, the same fluorescently labeled
trinucleotide 5 and T7 polymerase was used to obtain FAM-L2N-
m7Gp3Am-mRNA, coding for Gaussia luciferase.
In contrast to RNA35, the maximum concentration of labeled
mRNA was 0.1 µM, and HPLC analysis of IVT mRNA showed
that the capping was not full (signals partially overlapped).
Therefore, we purified the monolabeled mRNA by HPLC to
remove the main portion of uncapped mRNA in order to perform
labeling on high-purity monolabeled mRNA. mRNA labeling
performed in the presence of either 500- or 1000-fold excess 2b
gave similarly high conversion into dually labeled mRNA (86%
yield from triplicate experiments). Remarkably, HPLC purification
allowed the separation of mono- and dually labeled mRNAs in a
very efficient way, and the two mRNA isomers, presumably
corresponding to the two regioisomers of 4,5-disubstituted
triazole in the 2b derivative, were also separated (Figure 6A-B).
To demonstrate the potential of this dually labeled RNA for
mRNA research, we first performed an mRNA decapping assay
using human decapping protein 2 (hDcp2). Typically, RNA
decapping can be precisely monitored for short RNAs, for which
capped and uncapped RNAs can be separated during gel
electrophoresis and independently quantified. For longer RNAs,
such separation is not possible, and other methods need to be
employed; however, these methods are not fully quantitative.[23b;
Figure 4. HRPAGE (15% PAA, 7 M urea, 1× TBE) of NAD-RNAs. Direct
labeling/ligation versus ligation followed by SPAAC. The ligation/SPAAC
efficiency was calculated based on a densitometric analysis (CLIQ software) of
EtBr stained gel, as average of 2 or 3 replicates.
Trinucleotide 5 (FAM-L2N-m7GpppAmpG, Figure 1C, Scheme
S2) was incorporated into the RNA by IVT with T7 polymerase
(Figure S4).[24] The resulting m7G-RNA35 was directly labeled
with the fluorescent pAp analog 2b. The labeling reactions were
visualized by HRPAGE and RP-HPLC (Figure 5). The RNA
analyzed by PAGE was visualized for both Cy5 fluorescence
and FAM fluorescence before EtBr staining. We found that LT4-
mediated ligation was very efficient, yielding the desired dually
labeled RNA as the major product. HPLC analysis revealed that
unlabeled, monolabeled, and dually labeled RNAs were easily
separated, allowing us to isolate the dually labeled RNA with
25]
Therefore, we tested whether the formation of uncapped
mRNAs catalyzed by hDcp2 can be directly visualized using
dually labelled mRNA. As shown in Figure 6C, almost complete
decapping occurred within 30 min, as demonstrated by the
decrease in fluorescein signal over time as a result of 5′-cap
loss, whereas the signal from the SCy5 label (attached at the 3′
end of mRNA) corresponding to total RNA remained constant.
The fluorescein signal was stable in the absence of hDcp2
(Figure S6), proving that this dually labeled mRNA was a
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
Figure 6. A. HPLC profile of dually labelled mRNA purification. B. Agarose electrophoresis (1.5%, TBE). 1: uncapped mRNA, 2: FAM-m7Gp3Am-mRNA
(substrate), 3: FAM-m7Gp3Am-mRNA after HPLC (retention time, 9.9 min), 4 and 6: two different batches of FAM-m7Gp3Am-mRNA-SCy5 isomer 1 (retention time,
17.3 min), 5 and 7: two different batches of FAM-m7Gp3Am-mRNA-SCy5 isomer 2 (retention time, 17.7 min), 8: Riboruler high range Thermofisher. C. Decapping
assay monitored by agarose electrophoresis (1.5%, TBE). D. Relative fluorescence intensity as a function of time from 3 replicates of mRNA decapping assay; E.
Translation efficiency of unlabeled, monolabeled, or dually labeled mRNA in HeLa cells, the data represent mean values from 6 replicates ± SEM. .
valuable tool for visualizing mRNA and for directly monitoring
and quantifying decapping (Figure 6D).
protocol may serve as excellent probes for studying RNA fate
and function in vitro and in vivo. Notably, the RNA35 probes
developed here have been recently employed to study the
activation of decapping in in vitro reconstituted biomolecular
condensates.[26] The full length, translationally active mRNAs
carrying two different labels at the 5′ and 3′ ends, respectively,
are better model molecules to monitor the decapping process in
vitro and are also potentially applicable to more advanced
experiments in cell extracts or living cells directed towards
finding links among mRNA localization, mRNA turnover,
and protein expression and for following the fate of mRNA-
based therapeutics.
Finally, we asked whether the dually labeled mRNA was suitable
for experiments in living cells. We speculated that for this
purpose, the mRNA would have to retain its basic biological
function, i.e., undergo translation despite the 5′ and 3′ end
modifications. Therefore, we studied the expression of dually
labeled mRNAs in cultured HeLa cells. As references, we used
monolabeled 5′-capped mRNA and uncapped mRNA (the latter
of which was expected to be translationally inactive). As shown
in Figure 6E, the dually labeled mRNA was translationally active,
although the protein expression level was lower than that of
monolabeled mRNA. The effect of decreased expression may
result either from the presence of a bulky label at the 3′ end or
the presence of 3′-phosphate at the RNA terminus. Further
experiments are required to clarify this. However, the expression
was significantly higher than that of uncapped mRNA, confirming
the utility of our mRNA probes for mRNA research.
Acknowledgements
Financial support from the National Science Centre
(2015/18/E/ST5/00555 to J.K. and 2019/33/B/ST4/01843 to J.J.)
is gratefully acknowledged.
Conclusion
Keywords: RNA labeling • 3’-phosphoadenosine 5’-phosphate •
decapping • mRNA probe • T4 ligase •
In conclusion, we developed a straightforward chemo-enzymatic
method to provide monolabeled or dually labeled RNA in an
efficient and rapid manner based on LT4 and novel pAp
analogs. The compounds enabled either direct or two-step
labeling of RNA at the 3′ end and were also compatible with
cotranscriptional 5′ capping and labeling. Ligation could be
applied to short RNAs and long RNAs, including full-length
mRNAs. Labeled and biotinylated RNAs obtained using our
[1]
[2]
a) Y. Su, M. C. Hammond, Curr. Opin. Biotechnol. 2020, 63, 157-166;
b) S. Ray, J. R. Widom, N. G. Walter, Chem Rev 2018, 118, 4120-
4155.
a) S. Pitchiaya, L. A. Heinicke, T. C. Custer, N. G. Walter, Chem. Rev.
2014, 114, 3224-3265; b) G. Bao, W. J. Rhee, A. Tsourkas, Annu. Rev.
Biomed. Eng. 2009, 11, 25-47; c) Y. Xia, R. Zhang, Z. Wang, J. Tian, X.
Chen, Chem. Soc. Rev. 2017, 46, 2824-2843; d) Z. Wang, W. Liu, C.
Fan, N. Chen, Methods 2019, 161, 16-23; e) A. Mamot, P. J. Sikorski,
M. Warminski, J. Kowalska, J. Jemielity, Angew. Chem. Int. Ed. Engl.
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
2017, 56, 15628-15632; Angew. Chem. 2017, 129, 15834-15838; f) B.
Juskowiak, Anal. Bioanal. Chem. 2011, 399, 3157-3176.
A. Mor, A. White, K. Zhang, M. Thompson, M. Esparza, R. Munoz-
Moreno, K. Koide, K. W. Lynch, A. Garcia-Sastre, B. M. Fontoura, Nat.
Microbiol. 2016, 1, 16069.
a) A. Saady, M. Wojtyniak, E. Varon, V. Bottner, N. Kinor, Y. Shav-Tal,
C. Ducho, B. Fischer, Bioconjugate Chem. 2020, 31, 1188-1198; b) R.
Verbeke, I. Lentacker, L. Wayteck, K. Breckpot, M. Van Bockstal, B.
Descamps, C. Vanhove, S. C. De Smedt, H. Dewitte, J. Controlled
Release 2017, 266, 287-300.
a) J. L. Kirschman, S. Bhosle, D. Vanover, E. L. Blanchard, K. H.
Loomis, C. Zurla, K. Murray, B. C. Lam, P. J. Santangelo, Nucleic Acids
Res. 2017, 45, e113; b) N. Muthmann, K. Hartstock, A. Rentmeister,
Wiley Int. Rev.: RNA 2020, 11, e1561.
[24] T. M. Coleman, G. Wang, F. Huang, Nucleic Acids Res. 2004, 32, e14.
[25] a) M. Warminski, P. J. Sikorski, Z. Warminska, M. Lukaszewicz, A.
Kropiwnicka, J. Zuberek, E. Darzynkiewicz, J. Kowalska, J. Jemielity,
Bioconjugate Chem. 2017, 28, 1978-1992; b) F. Picard-Jean, C. Brand,
M. Tremblay-Letourneau, A. Allaire, M. C. Beaudoin, S. Boudreault, C.
Duval, J. Rainville-Sirois, F. Robert, J. Pelletier, B. J. Geiss, M.
Bisaillon, PLoS One 2018, 13, e0193804; c) M. V. Deshmukh, B. N.
Jones, D. U. Quang-Dang, J. Flinders, S. N. Floor, C. Kim, J. Jemielity,
M. Kalek, E. Darzynkiewicz, J. D. Gross, Mol Cell 2008, 29, 324-336.
[26] R. W. Tibble, A. Depaix, J. Kowalska, J. Jemielity, J. D. Gross, Nat.
Chem. Biol. 2021.
[3]
[4]
[5]
[6]
[7]
E. P. Diamandis, T. K. Christopoulos, Clin. Chem. 1991, 37, 625-636.
a) J. T. Rodgers, P. Patel, J. L. Hennes, S. L. Bolognia, D. P. Mascotti,
Anal. Biochem. 2000, 277, 254-259; b) M. Wilchek, E. A. Bayer, O.
Livnah, Immunol. Lett. 2006, 103, 27-32; c) H. A. Vincent, J. O. Phillips,
C. A. Henderson, A. J. Roberts, C. M. Stone, C. E. Mardle, L. E. Butt,
D. M. Gowers, A. R. Pickford, A. J. Callaghan, PLoS One 2013, 8,
e79142; d) S. Bednarek, V. Madan, P. J. Sikorski, R. Bartenschlager, J.
Kowalska, J. Jemielity, Philos. Trans. R. Soc., B 2018, 373; e) P. C.
Trippier, Chemmedchem 2013, 8, 190-203.
[8]
[9]
E. Paredes, M. Evans, S. R. Das, Methods 2011, 54, 251-259.
a) J. T. George, S. G. Srivatsan, Methods 2017, 120, 28-38; b) K. Krell,
D. Harijan, D. Ganz, L. Doll, H. A. Wagenknecht, Bioconjugate Chem.
2020, 31, 990-1011; c) O. A. Krasheninina, J. Thaler, M. D. Erlacher, R.
Micura, Angew. Chem. Int. Ed. Engl. 2021; Angew. Chem. 2021, 133,
7046-7050; d) L. Anhauser, A. Rentmeister, Curr. Opin. Biotechnol.
2017, 48, 69-76; e) K. J. Westerich, K. S. Chandrasekaran, T. Gross-
Thebing, N. Kueck, E. Raz, A. Rentmeister, Chem. Sci. 2020, 11, 3089-
3095; f) L. Anhauser, S. Huwel, T. Zobel, A. Rentmeister, Nucleic Acids
Res. 2019, 47, e42; g) M. Zhao, F. D. Steffen, R. Borner, M. F.
Schaffer, R. K. O. Sigel, E. Freisinger, Nucleic Acids Res. 2018, 46,
e13; h) S. Croce, S. Serdjukow, T. Carell, T. Frischmuth,
ChemBioChem. 2020, 21, 1641-1646.
[10] O. C. Uhlenbeck, R. I. Gumport, in The enzymes, (Ed.: P. D. Boyer),
New York: Academic Press, 1982, p. 31.
[11] a) M. Ghaem Maghami, S. Dey, A. K. Lenz, C. Hobartner, Angew.
Chem. Int. Ed. Engl. 2020, 59, 9335-9339; Angew. Chem. 2020, 132,
9421 –9425; b) M. Ghaem Maghami, C. P. M. Scheitl, C. Hobartner, J.
Am. Chem. Soc. 2019, 141, 19546-19549; c) A. K. Sharma, J. J. Plant,
A. E. Rangel, K. N. Meek, A. J. Anamisis, J. Hollien, J. M. Heemstra,
ACS Chem. Biol. 2014, 9, 1680-1684.
[12] A. G. Bruce, O. C. Uhlenbeck, Nucleic Acids Res. 1978, 5, 3665-3677.
[13] a) R. Cosstick, L. W. McLaughlin, F. Eckstein, Nucleic Acids Res. 1984,
12, 1791-1810; b) S. Keyhani, T. Goldau, A. Blumler, A. Heckel, H.
Schwalbe, Angew. Chem. Int. Ed. Engl. 2018, 57, 12017-12021;
Angew. Chem. 2018, 130, 12193-12197.
[14] a) J. H. Ren, D. J. Goss, Nucleic Acids Res. 1996, 24, 3629-3634; b)
M. Ziemniak, M. Szabelski, M. Lukaszewicz, A. Nowicka, E.
Darzynkiewicz, R. E. Rhoads, Z. Wieczorek, J. Jemielity, RSC Adv.
2013, 3.
[15] a) S. Nainar, M. Kubota, C. McNitt, C. Tran, V. V. Popik, R. C. Spitale,
J. Am. Chem. Soc. 2017, 139, 8090-8093; b) N. J. Agard, J. M. Baskin,
J. A. Prescher, A. Lo, C. R. Bertozzi, ACS Chem. Biol. 2006, 1, 644-
648; c) J. M. Holstein, L. Anhauser, A. Rentmeister, Angew. Chem. Int.
Ed. Engl. 2016, 55, 10899-10903; Angew. Chem. 2016, 128, 11059-
11063; d) U. Pradere, J. Hall, Bioconjugate Chem. 2016, 27, 681-687;
e) M. Jezowska, D. Honcharenko, A. Ghidini, R. Stromberg, M.
Honcharenko, Bioconjugate Chem. 2016, 27, 2620-2628.
[16] D. Ganz, D. Harijan, H.-A. Wagenknecht, RSC Chem. Biol. 2020, 1, 86-
97.
[17] a) N. J. Agard, J. A. Prescher, C. R. Bertozzi, J. Am. Chem. Soc. 2004,
126, 15046-15047; b) M. F. Debets, S. S. van Berkel, S. Schoffelen, F.
P. Rutjes, J. C. van Hest, F. L. van Delft, Chem. Commun. 2010, 46,
97-99.
[18] A. R. Kore, I. Charles, L. Yang, S. Kuersten, Nucleosides Nucleotides
Nucleic Acids 2009, 28, 292-302.
[19] C. Julius, Y. Yuzenkova, Wiley Int. Rev. RNA 2019, 10, e1512.
[20] T. M. Haas, P. Ebensperger, V. B. Eisenbeis, C. Nopper, T. Durr, N.
Jork, N. Steck, C. Jessen-Trefzer, H. J. Jessen, Chem. Commun. 2019,
55, 5339-5342.
[21] a) Y. G. Chen, W. E. Kowtoniuk, I. Agarwal, Y. Shen, D. R. Liu, Nat.
Chem. Biol. 2009, 5, 879-881; b) D. J. Luciano, J. G. Belasco, Trends
Biochem. Sci. 2015, 40, 245-247; c) H. Cahova, M. L. Winz, K. Hofer,
G. Nubel, A. Jaschke, Nature 2015, 519, 374-377; d) X. Jiao, S. K.
Doamekpor, J. G. Bird, B. E. Nickels, L. Tong, R. P. Hart, M. Kiledjian,
Cell 2017, 168, 1015-1027.e1010.
[22] A. Mlynarska-Cieslak, A. Depaix, E. Grudzien-Nogalska, P. J. Sikorski,
M. Warminski, M. Kiledjian, J. Jemielity, J. Kowalska, Org. Lett. 2018,
20, 7650-7655.
[23] a) M. Ishikawa, R. Murai, H. Hagiwara, T. Hoshino, K. Suyama, Nucleic
Acids Symp. Ser. (Oxf.) 2009, 129-130; b) P. J. Sikorski, M. Warminski,
D. Kubacka, T. Ratajczak, D. Nowis, J. Kowalska, J. Jemielity, Nucleic
Acids Res. 2020, 48, 1607-1626.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202101909
Chemistry - A European Journal
RESEARCH ARTICLE
Entry for the Table of Contents
Insert graphic for Table of Contents here.
Labelling of long RNAs, protein-coding RNAs in particular, still poses a challenge. We developed new adenosine 3’,5’-bisphophate
analogs as reagents for RNA 3’-end labelling by T4 RNA ligase. The compounds were efficient T4 substrates yielding 3’-mono-
labelled RNAs with various 5’-ends (triphosphate, NAD). Combination of 3’- and 5’-end labelling yielded dually labeled m7G-capped
RNAs that enabled quantitative analysis of decapping and were translationally active. The proposed method enables the creation of
versatile probes for investigation of RNA function and localization.
Institute and/or researcher Twitter usernames: @JoannaK98008579
8
This article is protected by copyright. All rights reserved.