Stealth Fluorescence Labeling for Live Microscopy Imaging of mRNA Delivery

Article abstract of DOI:10.1021/jacs.1c00014

Methods for tracking RNA inside living cells without perturbing their natural interactions and functions are critical within biology and, in particular, to facilitate studies of therapeutic RNA delivery. We present a stealth labeling approach that can efficiently, and with high fidelity, generate RNA transcripts, through enzymatic incorporation of the triphosphate of tCO, a fluorescent tricyclic cytosine analogue. We demonstrate this by incorporation of tCO in up to 100% of the natural cytosine positions of a 1.2 kb mRNA encoding for the histone H2B fused to GFP (H2B:GFP). Spectroscopic characterization of this mRNA shows that the incorporation rate of tCO is similar to cytosine, which allows for efficient labeling and controlled tuning of labeling ratios for different applications. Using live cell confocal microscopy and flow cytometry, we show that the tCO-labeled mRNA is efficiently translated into H2B:GFP inside human cells. Hence, we not only develop the use of fluorescent base analogue labeling of nucleic acids in live-cell microscopy but also, importantly, show that the resulting transcript is translated into the correct protein. Moreover, the spectral properties of our transcripts and their translation product allow for their straightforward, simultaneous visualization in live cells. Finally, we find that chemically transfected tCO-labeled RNA, unlike a state-of-the-art fluorescently labeled RNA, gives rise to expression of a similar amount of protein as its natural counterpart, hence representing a methodology for studying natural, unperturbed processing of mRNA used in RNA therapeutics and in vaccines, like the ones developed against SARS-CoV-2.

Full text of DOI:10.1021/jacs.1c00014

pubs.acs.org/JACS
Article
Stealth Fluorescence Labeling for Live Microscopy Imaging of mRNA
Delivery
Tom Baladi, Jesper R. Nilsson, Audrey Gallud, Emanuele Celauro, Cécile Gasse, Fabienne Levi-Acobas,
Ivo Sarac, Marcel R. Hollenstein, Anders Dahlén, Elin K. Esbjörner,* and L. Marcus Wilhelmsson*
Cite This: J. Am. Chem. Soc. 2021, 143, 5413−5424
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Methods for tracking RNA inside living cells without
perturbing their natural interactions and functions are critical within
biology and, in particular, to facilitate studies of therapeutic RNA
delivery. We present a stealth labeling approach that can efficiently,
and with high fidelity, generate RNA transcripts, through enzymatic
incorporation of the triphosphate of tCO, a fluorescent tricyclic
cytosine analogue. We demonstrate this by incorporation of tCO in
up to 100% of the natural cytosine positions of a 1.2 kb mRNA
encoding for the histone H2B fused to GFP (H2B:GFP).
Spectroscopic characterization of this mRNA shows that the
incorporation rate of tCO is similar to cytosine, which allows for
efficient labeling and controlled tuning of labeling ratios for different
applications. Using live cell confocal microscopy and flow cytometry,
we show that the tCO-labeled mRNA is efficiently translated into H2B:GFP inside human cells. Hence, we not only develop the use
of fluorescent base analogue labeling of nucleic acids in live-cell microscopy but also, importantly, show that the resulting transcript
is translated into the correct protein. Moreover, the spectral properties of our transcripts and their translation product allow for their
straightforward, simultaneous visualization in live cells. Finally, we find that chemically transfected tCO-labeled RNA, unlike a state-
of-the-art fluorescently labeled RNA, gives rise to expression of a similar amount of protein as its natural counterpart, hence
representing a methodology for studying natural, unperturbed processing of mRNA used in RNA therapeutics and in vaccines, like
the ones developed against SARS-CoV-2.
and quantified, but they generally rely on the use of heavily
modified externally labeled oligonucleotides with chemical and
physical properties that are significantly different from their
natural counterparts. Cyanine dyes such as Cy3 and Cy5
conjugated, for instance, via strain-promoted cycloaddition
linkers remain the most common labeling choice,6,7 even
though their bulkiness and hydrophobicity significantly impede
both transcription and translation8 (has also been noted by
TriLink Biotechnologies, one of the main manufacturers of
externally labeled mRNA) of mRNA. Development of more
universally applicable RNA labeling schemes and probes that
are minimally perturbing to RNA’s functions and compatible
with live-cell fluorescence imaging is therefore needed and may
become as crucial for the RNA field as the discovery and
INTRODUCTION
■
RNA is a key molecule of life and a main active player of the
central dogma of molecular biology. RNA is also a crucial
regulator of gene expression via for instance micro (mi)- and
small interfering (si)-RNA and through its intrinsic catalytic
activity; RNA therefore plays a fundamental role in biology. It
has, for these reasons, emerged as a highly promising and
versatile drug modality with potential to modify cellular
function at the translational level, opening up entirely new
avenues to address previously undruggable targets.1 Increased
molecular and mechanistic knowledge of biological processes
that involve RNA is therefore important and requires, in many
instances, new methodological tools. In the context of RNA
therapeutics, cellular delivery remains a major challenge, and
better understanding of cellular uptake, endosomal release, and
cytosolic delivery of RNAs is needed to unleash their full
potential.2−4 A major problem in this regard relates to the
challenge of directly visualizing RNA molecules as they are
taken up, processed, and subsequently released into the
cytosol.5
Received: January 1, 2021
Published: April 2, 2021
Recent advances in RNA imaging have generated a broad
spectrum of tools and probes by which RNA can be analyzed
© 2021 The Authors. Published by
American Chemical Society
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424
5413

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Schematic outline of the study detailing the synthesis of the main building block of the study, the fluorescent tricyclic cytosine analogue
tCO triphosphate (7) and its use. Blue pathway: End-labeling of ssRNA sequences using a terminal deoxynucleotidyl transferase (TdT). Green
pathway: Cell-free transcription into H2B:GFP mRNA, transfection into human cells, and live-cell imaging of both the GFP (translation product)
and tCO (mRNA). Yellow pathway: Cell-free transcription into calmodulin-like3 RNA followed by cell-free translation into the CALML3 protein.
Color code: tCO in red, GFP in green, native nucleosides in black or beige.
understanding of the green fluorescent protein (GFP) has been
for proteins.9−11
or gold nanoparticles,17 before delivery. Another emerging
strategy is to fuse the RNA of interest to an aptamer sequence
that binds fluorescent dyes in situ,18 but this can also adversely
influence translation efficiency and RNA−protein interac-
tions.19 To reduce the risks of interfering with enzymatic
processes, chimeric mRNAs have been developed with
multiple stem-loop structures downstream of the STOP
codon, which serve as binding sites for fluorescent fusion
proteins.20−22 While versatile, all these labeling strategies
impact profoundly on the physicochemical properties and
molecular weights of RNAs and therefore risk significantly
perturbing their natural function, spatiotemporal distribution,
and transport.
A major challenge in RNA imaging is to develop new
methods and probes that are functional in living cells.12 While
methods reliant on cell fixation such as fluorescence in situ
hybridization (FISH) can visualize and quantify endogenous
mRNA interactions with astounding specificity and single-
molecule resolution,13 they fall short with respect to capturing
the spatiotemporal and conformational dynamics that are
important to RNA function. The development of FRET-pair-
functionalized antisense oligonucleotides (ASOs)14,15 is highly
interesting in this regard, but requires binding of at least two
ASOs to the mRNA target, which may sterically block protein-
binding sites, hinder mRNA translation, or induce tertiary
structure formation; in practice the method has also mainly
been used on fixed cells. To visualize RNA in live cells, probes
are often conjugated to a vehicle, such as a DNA nanocages16
Additional less perturbing strategies for labeling mRNA have
therefore been developed, relying for example on click
chemistry approaches with azido-functionalized 5′-cap ana-
logues,23 3′-polyA tails,24 or nucleotides25 enabling in-cell
5414
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
post-transcriptional labeling of mRNA. Similarly, Ziemniak et
al. have produced a range of fluorescent 5′-cap analogues that
are compatible with both transcription and translation.26
However, the resulting mRNA products only carry one single
label and have therefore not been possible to visualize inside
cells, limiting the applicability of this method.
triphosphorylation of nucleosides,42−45 but no generic method
exists to effectively synthesize and purify ribonucleoside
triphosphates in the high yields that are required for practical
use in biochemical applications.
To overcome this “hit-and-miss” aspect of triphosphate
synthesis, we developed a new synthesis scheme that requires
no preliminary protection of the 2′- and 3′-positions and that
facilitates purification of the final product (Figure 1). In our
hands, this new method allowed the synthesis of the tCO
ribonucleoside triphosphate used herein and two additional
nucleobase-modified ribonucleoside triphosphates (manu-
scripts in preparation) in equally good yields, and we envision
it could therefore become a generic and convenient route
toward any modified nucleoside triphosphate.
Fluorescent nucleobase analogues (FBAs) have emerged as
attractive labels for DNA and RNA. However, even though we
and others have significantly improved FBAs with respect to
brightness, excitation, and emission to facilitate their use in
fluorescence microscopy,27−29 significant challenges have
remained regarding development of FBAs that are sufficiently
enzyme-compatible to be effectively processed during tran-
scription and translation.30 FBAs have the advantage of being
internal fluorophores, with relatively small chemical modifica-
tions to the natural base that they replace. Furthermore, their
design enables normal base-pairing and -stacking of the target
nucleic acid. FBAs are therefore considered to be native-like
fluorescent labels and have been extensively used in vitro to
probe nucleic acid structure and behavior. We have, for
example, designed FBA interbase FRET pairs to obtain
detailed information on the structure and base orientation in
DNA31 and RNA,32,33 and others have used FBAs to study
biophysically ribosome-mediated codon:anticodon base-pair
formations.34 A handful of studies indicate that FBAs can be
incorporated into RNA via cell-free transcription, resulting in
for example ca. 800 nucleotide (nt) RNA strands with a
modified cystosine35 or short transcripts with fluorescent
isomorphic guanine36,37 and uridine.38,39 However, none of
these studies have proven that FBA transcripts can be
translated, and FBA-labeled RNAs have not yet been used in
biological applications or to visualize RNA molecules inside
living cells. This progress is needed to translate FBAs from
useful in vitro probes to functional tools for chemical and
medical biology.30
The tCO ribonucleoside triphosphate was synthesized from a
solid-supported ribonucleoside, a strategy that has never been
reported for modified nucleobases, but attempted with some
success for unmodified nucleobases with 2′-OMe backbone
protection. Our synthetic scheme relies on phosphoramidite
chemistry, which involves the use of cycloSal-phosphoramidite
5 and bis(tetrabutylammonium) dihydrogen pyrophosphate 6
(Figure 1). This approach was developed by Meier et al.46 to
achieve efficient 5′-triphosphorylation of short solid support-
bound DNA and RNA oligonucleotides in moderate to good
yields. In our hands, both Krupp’s47 and Meier’s46 solid-phase
triphosphorylation methods yielded the product, although
Meier’s gave a higher yield. To the best of our knowledge, the
cycloSal-phosphoramidite method on a solid support described
herein has never been applied to a single ribonucleoside before,
let alone to produce a modified nucleoside triphosphate.
Briefly, a long-chain alkylamino controlled-porosity glass
(CPG) support was functionalized with a succinyl moiety.
Protected ribonucleoside 1 was synthesized according to a
method by Fu
̈
chtbauer et al.32 and attached to the succinylated
support via ester bond formation (Figure 1). The resulting
ribonucleoside 4 on a solid support could be stored in the dark
at room temperature, with no degradation observed over three
months. Interestingly, intermediate 4 could also be synthesized
via succinylation of ribonucleoside 1 in solution followed by
coupling with the amino support 2. In both cases, the starting
nucleoside could be used unprotected at the 2′- and 3′-
positions, thus eluding the need for additional protection steps,
which makes our method straightforward. Subsequently,
triphosphorylation of ribonucleoside 4 was performed using
the cycloSal method. After triphosphorylation, support-bound
triphosphate was deacetylated and cleaved from the CPG
support using ammonium hydroxide/methylamine (AMA) for
2 h at room temperature. Subsequent reverse phase or ion-
exchange chromatography allowed the desired triphosphate 7
in a triphosphorylation yield of 60% and with a high UV purity
of 99%. Importantly, up to 85% of the unreacted nucleoside 1
could conveniently be recovered by precipitation from the first
reaction crude, compensating for the low loading achieved.
Considering this, our triphosphorylation method gives an
overall yield of up to 30% of the tCO ribonucleoside
triphosphate, which is higher than most solution-based
alternatives.
In this study, we demonstrate that the fluorescent tricyclic
cytosine analogue 1,3-diaza-2-oxophenoxazine (tCO; Absmax
=
369 nm; Emmax = 457 nm; ε369 = 9370 M−1 cm−1; ⟨ΦF⟩ =
0.24)32 can be enzymatically incorporated in high numbers
into RNA via end-labeling reactions as well as cell-free
transcription. We furthermore show that it is possible to
exchange all natural cytosines in a 1.2 kb long mRNA for tCO
(Figure 1) and retain translation competence both in vitro and
in human cells. We also demonstrate, for the first time, that an
FBA-labeled mRNA can be sufficiently fluorescent to be
directly visualized by confocal microscopy in a living human
cell and used to study mRNA delivery and protein translation
in a drug delivery context.
This presents a significant advance to the FBA and RNA
imaging fields and a new powerful tool to enable effective
visualization of RNA and thereby enable studies of RNA
function, trafficking, and localization in a variety of cellular
contexts, including for example drug delivery, virus processing,
and exosome biology.
RESULTS AND DISCUSSION
■
Synthesis of the Tricyclic Cytosine Analogue Tri-
phosphate. The tCO ribonucleoside triphosphate has, unlike
the corresponding nucleoside, never been synthesized, and,
hence, we needed to establish a synthetic route toward our
target molecule. Since the Yoshikawa40 and Ludwig−
Eckstein41 conditions were published in 1969 and 1989,
respectively, a plethora of methods have been proposed for the
Cell-Free Enzymatic Incorporation of tCO into RNA.
The tCO ribonucleoside triphosphate (tCOTP) was used to
produce fluorescently labeled RNA via two different enzymatic
methods.
First, we tail-labeled short RNA oligonucleotides using the
terminal deoxynucleotidyl transferase enzyme (TdT, Figure 1,
5415
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. Enzymatic incorporation of the modified triphosphate by cell-free transcription with short templates, sequencing of reverse transcriptase
products, and TdT-mediated labeling. (a) Gel image (PAGE 10%; visualization by Midori green) of RNA products from 6 h of T7 RNA
polymerase-assisted transcription using the D1 Library as a DNA template. See Methods for buffer composition. (b) Gel image (PAGE 20%;
visualization by phosphorimager) of the products from TdT-mediated end-labeling of oligonucleotide TdT1 using tCOTP. The lowest band labeled
“n” represents the unreacted TdT1 oligonucleotide, and all above bands are different products with increasing additions of tCO to the tail. P:
control reaction in absence of polymerase. (c) Picture of UV-light-irradiated solution of purified TdT3 RNA end-labeled with tCO. (d) Sequence
logo from the cloning−sequencing protocol of reverse transcription products from the modified T12.3 RNA (top: products from a 6 h transcription
reaction with CTP; bottom: products from a 6 h transcription reaction with tCOTP). Lower (stars) and higher (arrow) frequency A to G
transversions are highlighted.
blue arrow), which catalyzes template-independent addition of
random nucleotides to 3′-overhangs in both DNA and
RNA.48,49 Starting from a 17-mer ssRNA (TdT1, see
Supplementary Table 1 for sequence) we demonstrate
successful addition of multiple adjacent tCOs (from one to
>25) on nearly all RNA primers (Figure 2b). The addition was
equally effective with Co(II), Mg(II), or Mn(II) as cofactor in
accord with the normally reported function of TdT,48
supporting a native-like enzymatic processing of tCO. TdT-
mediated tail-labeling of longer (50 nt, TdT2, Supplementary
Table 1) ssRNA sequences was also successful, and the main
products were more uniform, containing typically one or two
tCO (Supplementary Figure 1). The lower processivity found
for this longer RNA could be an effect of its length or of its 3′-
sequence but more likely due to the fact that the TdT is a
DNA polymerase that prefers deoxyribonucleotides both as
substrates and as templates.48 The fluorescence originating
from a tCO tail-labeled RNA was clearly observable by the
naked eye upon UV irradiation (Figure 2c). This proves that it
is possible to use tCO to site-specifically end-label RNA of
5416
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. Incorporation of tCO into full-length mRNA by T7 RNA polymerase assisted cell-free transcription. Denaturing agarose bleach gels
showing RNA transcripts formed at five different tCOTP/CTP ratios (0−100%). Direct visualization of tCO fluorescence (a) and after ethidium
bromide staining (b). RNA samples were heat-denatured (65 °C for 5 min, 1.5% bleach in the gel) prior to loading. (c) Same RNA transcripts
upon harsher denaturation (70 °C for 10 min, 2% bleach in the gel). The RiboRuler High Range RNA ladder was used.
various length, which represents an advantage in terms of
enabling dual labeling of RNA strands, where combinations of
tCO with other base- or backbone-modified fluorescent
markers will enable monitoring of, for example, in vivo
stability.
combination with the above-mentioned minor difference in
abortion frequency (20% or lower) found between CTP and
tCOTP, this suggests an overall lower level of misincorporation
for tCOTP than for the analogue reported by Stengel et al.
Altogether, this shows that the fluorescent tCOTP nucleotide is
well-tolerated by both TdT and T7 RNA polymerases, with the
latter only marginally increasing the incorporation error rate in
up to 50 nt long RNAs. In this sense, tCOTP behaves better
than the majority of previously modified nucleotides,35,36,50−52
including bulkier triphosphate analogues.53,54 It has been
reported that the T7 RNA polymerase binds the incoming
nucleotide substrate in an open conformation,55 which could
enable it to accommodate our modified base.
We also tested the capacity of T7 RNA polymerase to
incorporate tCOTP into short (20 nt/90 nt; Supplementary
Table 1) RNA transcripts, observing successful transcription
without premature transcription termination even upon
complete replacement of canonical CTP with tCOTP (Figure
2a, Supplementary Figure 2a, and Supplementary Figure 3).
This confirms that tCOTP is readily accepted also by the T7
polymerase as a substrate. To further assess the enzymatic
processing of tCOTP, we tested its propensity to be
misinserted opposite a single deoxyadenosine downstream of
the T7 promotor sequence of a guanosine-free DNA template
(Supplementary Table 1; DNA3, DNA4). Transcription
reactions were run in the presence of either CTP or tCOTP
and in the absence of UTP (Supplementary Figure 2b). With
the first template (DNA3), no noticeable difference was
observed in the reactivity of tCOTP and CTP, whereas a
mismatch closer to the T7 promoter sequence (DNA4)
resulted in slightly fewer aborted transcripts and hence more
full-length transcripts with tCOTP compared to native CTP.
This suggests a minor (20% based on densiometric analysis,
see Supporting Information) increase in error frequency with
tCOTP. A previous report by Stengel et al. used a structurally
related tricyclic cytosine analogue, tC, which carries a bulkier
sulfur atom instead of an oxygen at position 5 of the
pyrimidone ring, to produce RNA transcripts. Their reaction
proceeded with a considerable formation of mismatched tC−A
base pairs (a discrimination factor of 40 against was
reported).35 As a further comparison to this, we tested the
fidelity of incorporation of tCOTP by reverse transcription
(Supplementary Figure 3). After PCR amplification and A-
tailing reaction, the resulting amplicons were ligated to the
pGEM T vector and transfected into beta 2033 competent E.
coli cells, and plasmids stemming from white colonies were
subjected to Illumina sequencing. Multiple alignment analysis
revealed a mere 2-fold higher frequency of misincorporation
with tCOTP compared to CTP (28 vs 16 point mutations in
respectively 34 and 38 analyzed sequences; Supplementary
Figures 4 and 5); significant A to G transversions were
observed at position 62 (3 mutations, Figure 2d and
Supplementary Figure 6) and a lower number of transversions
at positions 24, 39, 42, 48, 51, 58 (1 mutation). In
Cell-Free Enzymatic Incorporation of tCO into Full-
Length mRNA Transcript. We next extended the in vitro
transcription to test if it was possible to produce full-length,
translationally active mRNAs with different degrees of tCO.
Transcription of long RNA containing a fluorescent base
analogue has only been reported once,35 but their ca. 800 nt
transcript was not active in reverse transcription and no
translation activity was reported, suggesting that the transcripts
were nonfunctional. Moreover, they showed that on increasing
the amounts of their fluorescent tricyclic cytosine nucleoside
triphosphate, the transcription reaction led to shorter tran-
scripts, suggesting premature termination. Apart from being a
fluorophore with superior brightness we envisioned that tCO,
with its oxygen in the middle ring instead of the considerably
bulkier sulfur (vdW radius of sulfur is 20−30% larger, resulting
in an approximately 100% larger occupied volume) for the
tricyclic cytosine used by Stengel et al., would represent a
better cytosine analogue that would minimize perturbations to
biological processes. We used a DNA template encoding for
histone protein H2B fused to GFP (H2B:GFP) to produce
mRNA via cell-free transcription. The template was codon
optimized (Supplementary Figure 7) to limit the number of C
repeats to limit self-quenching of adjacent tCO moieties56 (vide
infra). This optimization could also have the additional
positive effect that consecutive incorporations of tCOs would
be limited; previous work indeed suggested difficulties for
DNA polymerases to incorporate consecutive dtCOTPs.57
However, we also observed in vitro transcription of a
nonoptimized calmodulin mRNA (vide infra) even at 100%
tCOTP (0% CTP) despite this transcript containing three
CCCC repeats. We observed efficient transcription and tCO
incorporation with both T7 and SP6 RNA polymerases (Figure
1, first green arrow) at tCOTP/canonical CTP ratios ranging
5417
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. Spectroscopic characterization of cell-free synthesized tCO-modified RNA transcripts. Four reactions charged with different molar
fractions of tCOTP (blue: 100%, green: 75%, magenta: 50%, and red: 25%) in the total cytosine triphosphate pool (tCOTP + CTP) were
performed. The product transcripts were purified to wash out unreacted triphosphates prior to characterization. All reactions were performed as
independent duplicates, and the results are presented as mean standard deviation. (a) UV−vis absorption spectra normalized to A = 1 at the
RNA band, ca. 260 nm. Inset: tCOTP absorption normalized to A = 1 at the tCO band λmax (360 nm). (b) Plain bars: Fraction of incorporated tCO
(relative to the total amount of incorporated cytosines, i.e., tCO + C) in the transcripts. Checkered bars: Ratio of first-order reaction rate constants
for CTP vs tCOTP consumption. (c) Solid lines: UV−vis absorption spectra (normalized to A = 1 at the RNA band, ca. 260 nm) showing the tCO
band centered at 368−369 nm. Dashed lines: Emission spectra normalized to I = 1 at λmax (457 and 459 nm for the 25% and 100% transcript,
respectively). For clarity, the emission spectra for the 50% and 75% reactions were omitted. (d) Plain bars: Fluorescence quantum yields. Striped
bars: Fluorescence lifetime.
from 0 to 100% (full replacement), as demonstrated by agarose
bleach gel electrophoresis of the resulting transcripts (Figure
3a, Supplementary Figure 8a). The RNA transcripts appeared
as one single band, with a migration corresponding to the
expected 1247 nt mRNA product (H2B:GFP), demonstrating
that full-length mRNA was formed. The tCO-containing
mRNA bands could be directly visualized upon 302 nm
excitation using a conventional gel scanner (Figure 3a); the
increasing band intensities with increasing tCOTP/CTP
reaction ratio supported successful concentration-dependent
incorporation of tCO. Ethidium bromide staining (Figure 3a,
right) revealed similar amounts of RNA in all reactions,
suggesting that tCO incorporation does not reduce the cell-free
transcription reaction yield. Furthermore, unlike the RNA
product formation reported by Stengel et al.,35 no shorter
transcripts were observed, strengthening the conclusion that
the T7 RNA polymerase processes tCOTP correctly and
without premature abortion. Faint higher order bands were
apparent with all tCO-containing RNA transcripts (Figure 3b
and Supplementary Figure 8b) but could be effectively
removed by heat denaturation (Figure 3c). The intensity of
the higher order band appears to increase slightly with tCO
content, which could reflect the increased hydrophobicity
introduced by this modified base. Importantly, we demonstrate
that tCO can be successfully incorporated into full-length
mRNA and that this reaction proceeds effectively even when
canonical CTP is completely replaced with tCOTP, suggesting
that this base analogue is highly mimetic of its natural
counterpart.
Incorporation Efficiency and Spectroscopic Proper-
ties of tCO-Containing mRNA. We used absorption and
fluorescence spectroscopy to quantify the incorporation
efficiency of tCOTP compared to canonical CTP, following
purification of the RNA transcripts using a commercial cleanup
kit to remove unreacted tCOTP. Absorption spectra (Figure
4a) of tCO-containing transcripts show an absorption peak at
370 nm, consistent with its reported spectrum inside RNA.32
The relative magnitude of this peak compared to the RNA
absorption at 260 nm (reflecting total concentration) confirms
that the degree of tCO incorporation is concentration-
dependent and allowed us to determine the relative rate
constants for the incorporation of CTP and tCOTP (kC and
5418
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 5. Cell-free translation of calmodulin-3 (CALM3) mRNA. (a) Coomassie staining and (b) Western blot (WB) of the in vitro translation
reactions. NTC: no template control; + : kit template RNA control. The PageRuler Prestained Protein Ladder was used. (c) Quantification by WB
and densitometry analysis (mean % of control standard deviation; n = 3). One-way ANOVA with means comparison using Tukey’s post hoc test
showed no significant effect of tCO content in the mRNA transcripts on protein translation (p < 0.05).
k
tCO, respectively, see Methods for details). The ratios between
incorporation affects the relative brightness of the transcripts,
despite codon optimization, such that maximum emission is
achieved at 75%, although with relatively modest improvement
compared to 50% substitution (Supplementary Figure 9).
However, as shown below, a substitution level of 25% was
sufficient to visualize lipid-complexed tCO-labeled mRNA
(Figure 6c), indicating what we believe to be a suitable
labeling range for most cell applications. This is also
comparable to the 25% substitution level of U-positions with
a widely used commercial Cy5-mRNA delivered by TriLink,
vide infra. Importantly, due to the minor differences in
incorporation preference between tCOTP and CTP, we show
that it is possible to tune, with accuracy, the labeling density of
mRNA to optimize the relationship between brightness and
biological function (for example in translation as discussed
below).
the rate constants kC/ktCO were close to or slightly above unity
for all transcripts (0.96−1.4, Figure 4b), suggesting strongly
that T7 RNA polymerase discriminates only marginally
between CTP and tCOTP. Our observation of equal
incorporation efficiencies for tCOTP and CTP differs from
other published studies, where the corresponding incorpo-
ration ratio was found to be lower than 0.6 (and down to 0.13)
for d(tCOTP) incorporation into DNA and analogues
d(tCTP) into DNA and tCTP into RNA.35,57,58 This suggests
that tCO is a good nature-mimic with minimal perturbing
effects on the transcriptional process. Furthermore, a trivial but
noteworthy consequence of the determined rate constants is
that the tCO incorporation degree in our mRNA matches
extremely well the percentage of tCOTP added to the
transcription reaction (Figure 4b).
In quantitative fluorescence-based cell analyses, it is
important that fluorescence intensity is proportional to probe
concentration. We therefore examined the emissive behavior of
tCO in the mRNA transcripts as a function of its degree of
incorporation. We report a minor red-shift of the emission
spectrum (ca. 4 nm) with increasing tCO incorporation (Figure
4c), whereas more significant effects were observed on
fluorescence quantum yields and lifetimes (Figure 4d). The
figure shows that the quantum yield drops from 0.18 at 25%
incorporation to 0.09 at 100% incorporation, consistent with a
corresponding drop in fluorescence lifetime from 4.3 ns to 3.2
ns. We ascribe this to electronic coupling of molecular states of
the tCO fluorophore and a self-quenching effect56 due to
increased adjacency between tCOs57 (vide supra and
Supplementary Figure 9). The self-quenching at high tCO
In Vitro Translation of tCO-Labeled mRNA. We next
demonstrated that tCO-labeled mRNA is functional in
translation, providing the first evidence that this is possible
using an FBA-modified mRNA. We used a commercial
transcription−translation kit to produce calmodulin-3 under
cell-free conditions in bacterial lysates (Figure 1, yellow
arrows). mRNA, with the same tCOTP/CTP ratios as used
above (0 to 100% of tCOTP), was produced using the
calmodulin-3 DNA template plasmid provided with the kit, i.e.,
without any codon optimization. The expression of calm-
odulin-3 protein was first confirmed by SDS PAGE followed by
Coomassie staining (Figure 5a); the calmodulin-3 band
intensities for all samples with tCO-modified mRNA are within
10% of that of the 0% tCO control, suggesting that the FBA has
no significant effect on translation. The calmodulin-3
5419
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 6. Translation and visualization of tCO-labeled mRNA in human SH-SY5Y cells. The mRNA translation was monitored based on the
fluorescence intensity of the encoded H2B:GFP protein using confocal microscopy and flow cytometry. (a) mRNA translation following
electroporation; (top) confocal images (3× enlargements, scale bars: 10 μm) recorded 24 h postelectroporation, (middle) flow cytometry scatter
plots 24 h postelectroporation, and (bottom) mean cellular GFP fluorescence intensity (MFI GFP standard deviation) of all counted cells at 24,
48, and 72 h postelectroporation. (b) Representative MFI GFP histograms corresponding to the distributions in (a). (c) mRNA translation
following chemical transfection (lipofection), corresponding to the data shown in (a). (Top) Confocal images (3× enlargements, scale bars: 10
μm) recorded 48 h postchemical transfection, (middle) flow cytometry scatter plots 48 h postelectroporation, and (bottom) MFI GFP of all
counted cells at 24, 48, and 72 h postchemical transfection. (d) Snap-shot images from a confocal time-lapse experiment to monitor the
intracellular trafficking of 75% tCO-labeled mRNA (red) introduced, by chemical transfection, into cells with an overexpression of mRFP-Rab5 to
label early endosomes (orange). Resulting expression of H2B-GFP protein in the nucleus is shown in green. White arrows indicate discrete
mRNA−lipid complexes; scale bars: 10 μm. Supplementary Movie 1 shows the full time lapse. (e) mRNA translation of cyanine5-labeled (Cy5)
eGFP encoding mRNAs (TriLink) 24 h post-transfection (electroporation or chemical transfection). NL; nonlabeled. Scale bars: 10 μm. (f) Impact
of tCO or Cy5 incorporation on mRNA translation, represented as the ratio of cellular MFI GFP of the labeled mRNA relative to the cellular MFI
GFP for the corresponding nonlabeled RNA. All cell experiments were performed in three biological replicates.
5420
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
production was further verified and quantified by Western blot
(Figure 5b). Densiometric analysis of the calmodulin-3 bands
(normalized to the intensities of the GAPDH loading controls)
of three replicate samples showed that the protein yields from
tCO-labeled mRNAs were within 80−137% of the unlabeled
control (Figure 5c); one-way Anova analysis with Tukey’s post
hoc test showed that the observed variations in sample means
were not statistically significant; p < 0.05. Still, judging from
the trends in the Coomassie-stained gel and WB data it may be
possible that the translation of the 100% tCO-labeled mRNA
could be slightly impaired. Altogether, however, the in vitro
translation data suggest that tCO, unlike other sugar-52 or base-
modified59 nucleotide analogues, behaves very similar to
canonical cytosine with respect to ribosomal processing and
translation in cell-free systems. This shows that tCO therefore
functions as a novel fluorescent mimic of cytosine in both
transcription and translation.
these effects appear affected by tCO content, suggesting that
the tCO-labeled mRNA that reaches the cytosol in a functional
form is processed as the corresponding nonlabeled mRNA.
Chemical transfection using Lipofectamine resulted in
considerably fewer GFP-positive cells (Supplementary Figure
10c), but, importantly, no tCO-dependent effects on either
transfection efficiency or mean cellular GFP fluorescence were
observed (Figure 6c). This is in agreement with the cell-free
translation experiments (Figure 5) and suggests that tC0-
labeled mRNA, if appropriately delivered, can yield as much
cellular protein product as the corresponding nonlabeled
mRNA (see Figure 6f and further below). Following chemical
transfection, the mean fluorescence intensity increases over
time (Figure 6c), reflecting a continuous Lipofectamine-
mediated and tCO-independent endocytic uptake and escape
of mRNA cargo to the cytosol.
The complexation of the tCO-labeled mRNA with Lipofect-
amine enabled its direct visualization inside cells using live cell
confocal microscopy (Figure 6c, red puncta). This represents
the first observation of FBA-labeled nucleic acids inside a living
cell and demonstrates that tCO is bright enough to be tracked
following 405 nm laser excitation and yielded sufficient
contrast over the autofluorescent cellular background that is
typical to live cells. We therefore proceeded to visualize, in real
time, both the uptake and subsequent translation of an FBA-
modified mRNA by time-lapse recordings of cells over-
expressing an early endosome biomarker (mRFP-Rab5,
orange) (Figure 6d). We observed temporal colocalization
(Supplementary Movie 1) of the tCO signal with mRFP-Rab5
proteins, highlighting the fact that the mRNA transits through
the early endosome following endocytosis. Importantly, our
results show a successful new methodology that enables not
only simultaneous spatiotemporal tracking of the uptake and
trafficking of mRNA during cellular delivery but also the
appearance of its correctly localized and folded translation
product.
Translation and Visualization of tCO-Labeled mRNA
in Human Cells. We used two different approaches
(electroporation and chemical transfection more closely
mimicking drug delivery) to introduce tCO-labeled mRNAs
into human cells and test their in-cell translatability. The
transcripts used were 5′-capped and 3′-protected by poly
adenylation (by ca. 300 nt) to avoid degradation; we found
this to be necessary also for tCO-containing mRNA, suggesting
that the FBA does not alter its stability (see further below).
We observed, by confocal fluorescence microscopy of live
SH-SY5Y cells, the expression of a correctly localized (nuclear)
and folded (fluorescent) protein product following both
electroporation and chemical transfection of the tCO-labeled
mRNAs (Figure 6a and c, Supplementary Figure 10a and b),
suggesting functional processing of the FBA-modified mRNAs
by human ribosomal machineries as well as low frequency of
misincorporations of tCO into RNA during cell-free tran-
scription as reported in Figure 2.
We thereafter used flow cytometry to quantify the
translation of protein, based on mean cellular GFP
fluorescence intensities (Figure 6, Supplementary Figure 10).
Electroporation was used to introduce mRNA directly to the
cytosol and resulted in significantly higher transfection
efficiency (i.e., number of GFP-positive cells) than chemical
transfection with Lipofectamine (Supplementary Figure 10c).
We observed that the transfection efficiency following
electroporation decreased with increasing tCO content of the
mRNA, an effect that rendered an effective lowering of the
mean fluorescence intensity of the cell populations electro-
porated with mRNA containing tCO. Although this could
indicate that tCO impedes in-cell translation, our results using
chemical transfection (see below) suggest this is rather an
effect of how the mRNA was introduced. It is possible that the
increased hydrophobicity of tCO-labeled mRNAs in combina-
tion with the high invasiveness and electric fields of
electroporation caused some tCO mRNA degradation and/or
imposed secondary structures that prevented translation. A
slightly higher propensity to adopt secondary structures in
vitro is indicated from the gel analysis in Figure 3. In any case,
the effect was not associated with toxicity (Supplementary
Figure 10g) and the fluorescence signal persisted over 72 h
postelectroporation, although during this time the GFP
fluorescence gradually decreased (Figure 6a). This is likely a
combined effect of cell division (the doubling time of SH-
SY5Y cells under our experimental conditions is approximately
24 h) and cytosolic mRNA degradation. Importantly, none of
Finally, we compared the translation of the tCO-labeled
H2B-GFP encoding mRNA to that of the nonlabeled variant
(Figure 6f) to illustrate the effect of the FBA modification on
in-cell translation (which is apparent with electroporation but
not with chemical transfection as also discussed above). We,
thereafter, did the same experiment with a commercial eGFP-
encoding Cy5-labeled mRNA (with ca. 25% of all U positions
carrying the Cy5 dye) relating its translation to that of the
corresponding nonlabeled eGFP control (Figure 6e). Despite
the low labeling density, the Cy5-labeled mRNA has only ca.
20% of the translation capacity of its corresponding eGFP-
encoding nonlabeled mRNA following both electroporation
and chemical transfection (Figure 6f). This is consistent with
previous in vitro translation studies that confirm the trans-
lation-impeding nature of the Cy5 modification.8 Although the
tCO- and Cy5-labeled mRNAs encode for different proteins,
the normalization against their respective nonlabeled controls
may still give an indication of the relative effects on translation
of the two label types. Notable, in this respect, is that
Lipofectamine-delivered tCO-labeled mRNAs appear to
function similarly to their controls, whereas the Cy5-labeled
variant is significantly translation impeded. This could indicate
an improved translatability of mRNA carrying our fluorescent
FBA.
5421
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
strategies also holds significant combined potential, as it
would enable selective and site-specific incorporation of dual
labels to allow for, for example, FRET applications. We believe
that the development reported here will benefit pharmaceutical
industry, clinical laboratories, and academic partners aiming at
furthering their understanding of uptake and endosomal escape
mechanisms and allow them to take vital steps toward new and
improved delivery strategies for next-generation nucleic acid-
based drugs as well as further development of the recently
successful mRNA-based vaccines.
CONCLUSION
■
In this study we demonstrate that tCO, a fluorescent cytosine
base analogue with moderate chemical modification, in a
remarkable way takes the role of natural cytosine in a number
of biochemical processes. This FBA is correctly recognized by
several enzymatic machineries, including RNA polymerases,
reverse transcriptase, and bacterial as well as human ribosomes.
We have also developed a robust, generic, and affordable
synthesis method to produce the triphosphate variants of
unnatural nucleotides, presenting the first successful synthesis
of the tCOTP nucleoside triphosphate. This chemical develop-
ment is critical to enable the use of tCO and other FBAs in
biochemical and biological applications.
ASSOCIATED CONTENT
■
sı
* Supporting Information
We demonstrate that tCO can be incorporated into short and
long sequences of RNA via both end-labeling and cell-free
transcription, which enables versatile introduction of FBAs in
both flanking and functional segments of different kinds of
RNA. Importantly, we have managed to incorporate tCO in up
to 100% of natural cytosine positions of a full-length 1.2 kb
mRNA, without any premature abortions. Analysis of the
transcription products demonstrated that tCO is incorporated
into RNA virtually as efficiently as native CTP and therefore
constitutes a true nature-mimicking fluorescent modification in
this respect.
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c00014.
Supporting Movie 1 shows the internalization of the
75% tCO-labeled mRNA and its expression over 55 h in
Huh7 cells (1 s = 1.1 h) (AVI)
Characterization of tCOTP, all experimental procedures,
sequence alignments, additional electrophoresis gels,
confocal images, and flow cytometry data (PDF)
AUTHOR INFORMATION
■
Astoundingly, we also found that tCO-labeled mRNA is
translated into its correctly folded and localized protein
product, both in vitro and in live cells. This has not been
shown for FBA-modified mRNA before; in fact previous
published work using the cytosine analogue tC failed to
generate translation-competent mRNA even starting from a
considerably shorter (0.8 kB) template. In addition, both in
vitro translation experiments and chemical transfection of cells
indicate that tCO-labeled mRNAs can translate equally well as
the corresponding nonlabeled control, suggesting that this type
of internal label can indeed be less disrupting to mRNA
function compared to a Cy5-label. This points to the flexibility
and versatility of the herein discovered tCO-labeling method,
which offers opportunity to tune labeling density to optimize
brightness and distribution of fluorophores along the mRNA
sequence. Moreover, when using Cy5-labeled mRNA to study
delivery, the modus in the field is to mix in a fraction of the
labeled mRNA with the corresponding unlabeled mRNA, a
procedure that increases the risk of incorrectly reported rates
and levels of delivery and translation even further, something
that now can be avoided with our nature-mimicking labeling
technique.
Finally, we present the first example of an FBA-labeled
nucleic acid that can be directly visualized in live cells, showing
that tCO’s brightness and absorption at 405 nm are sufficient to
overcome previous limitations with FBA probes in biological
applications.30 Moreover, we demonstrate how this conven-
iently allows for spatiotemporal monitoring of uptake,
trafficking, and organelle colocalization of chemically trans-
fected mRNA in a live cell model with simultaneous detection
of its translation into H2B:GFP protein.
Corresponding Authors
L. Marcus Wilhelmsson − Department of Chemistry and
Chemical Engineering, Chemistry and Biochemistry,
Chalmers University of Technology, SE-41296 Gothenburg,
Sweden; orcid.org/0000-0002-2193-6639;
Email: marcus.wilhelmsson@chalmers.se
Elin K. Esbjörner − Department of Biology and Biological
Engineering, Chemical Biology, Chalmers University of
Technology, SE-41296 Gothenburg, Sweden; orcid.org/
0000-0002-1253-6342; Email: eline@chalmers.se
Authors
Tom Baladi − Department of Chemistry and Chemical
Engineering, Chemistry and Biochemistry, Chalmers
University of Technology, SE-41296 Gothenburg, Sweden;
Medicinal Chemistry, Research and Early Development,
Cardiovascular, Renal and Metabolism, BioPharmaceuticals
R&D, AstraZeneca, Gothenburg, Sweden; orcid.org/
0000-0003-3421-9417
Jesper R. Nilsson − Department of Chemistry and Chemical
Engineering, Chemistry and Biochemistry, Chalmers
University of Technology, SE-41296 Gothenburg, Sweden
Audrey Gallud − Department of Biology and Biological
Engineering, Chemical Biology, Chalmers University of
Technology, SE-41296 Gothenburg, Sweden
Emanuele Celauro − Department of Biology and Biological
Engineering, Chemical Biology, Chalmers University of
Technology, SE-41296 Gothenburg, Sweden
Cécile Gasse − Génomique Métabolique, Genoscope, Institut
Franco̧ is Jacob, CEA, CNRS, Univ Evry, Université Paris-
Saclay, 91057 Evry, France
We envision that our straightforward approach for
introducing nonperturbing fluorescent labels into RNA will
be an excellent addition to existing imaging tools, applicable
for elucidating trafficking mechanisms such as endosomal
escape and exosome formation, both of which are of
fundamental importance for pharmaceutical development.
Applying both TdT end-labeling and T7 transcription
Fabienne Levi-Acobas − Department of Structural Biology
and Chemistry, Laboratory for Bioorganic Chemistry of
Nucleic Acids, CNRS UMR3523, Institut Pasteur, 75724
Paris CEDEX 15, France
Ivo Sarac − Department of Structural Biology and Chemistry,
Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS
UMR3523, Institut Pasteur, 75724 Paris CEDEX 15, France
5422
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(11) Wittrup, A.; Ai, A.; Liu, X.; Hamar, P.; Trifonova, R.; Charisse,
K.; Manoharan, M.; Kirchhausen, T.; Lieberman, J. Visualizing lipid-
formulated siRNA release from endosomes and target gene knock-
down. Nat. Biotechnol. 2015, 33 (8), 870−6.
Marcel R. Hollenstein − Department of Structural Biology
and Chemistry, Laboratory for Bioorganic Chemistry of
Nucleic Acids, CNRS UMR3523, Institut Pasteur, 75724
Paris CEDEX 15, France; orcid.org/0000-0003-0263-
9206
Anders Dahlén − Medicinal Chemistry, Research and Early
Development, Cardiovascular, Renal and Metabolism,
BioPharmaceuticals R&D, AstraZeneca, Gothenburg, Sweden
(12) Li, Y.; Ke, K.; Spitale, R. C. Biochemical Methods To Image
and Analyze RNA Localization: From One to Many. Biochemistry
2019, 58 (5), 379−386.
(13) Burke, K. S.; Antilla, K. A.; Tirrell, D. A. A Fluorescence in Situ
Hybridization Method To Quantify mRNA Translation by Visualizing
Ribosome−mRNA Interactions in Single Cells. ACS Cent. Sci. 2017, 3
(5), 425−433.
(14) Wadsworth, G. M.; Parikh, R. Y.; Kim, H. D.; Choy, J. S.
mRNA detection in budding yeast with single fluorophores. Nucleic
Acids Res. 2017, 45 (15), e141−e141.
(15) Gaspar, I.; Hövelmann, F.; Chamiolo, J.; Ephrussi, A.; Seitz, O.
Quantitative mRNA Imaging with Dual Channel qFIT Probes to
Monitor Distribution and Degree of Hybridization. ACS Chem. Biol.
2018, 13 (3), 742−749.
(16) He, L.; Lu, D.-Q.; Liang, H.; Xie, S.; Luo, C.; Hu, M.; Xu, L.;
Zhang, X.; Tan, W. Fluorescence Resonance Energy Transfer-Based
DNA Tetrahedron Nanotweezer for Highly Reliable Detection of
Tumor-Related mRNA in Living Cells. ACS Nano 2017, 11 (4),
4060−4066.
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c00014
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare the following competing financial
interest(s): A.G., A.D., E.C., and T.B. are employees of
AstraZeneca and may hold shares in the company.
(17) Wang, B.; Chen, Z.; Ren, D.; You, Z. A novel dual energy
transfer probe for intracellular mRNA detection with high robustness
and specificity. Sens. Actuators, B 2019, 279, 342−350.
(18) Chen, X.; Zhang, D.; Su, N.; Bao, B.; Xie, X.; Zuo, F.; Yang, L.;
Wang, H.; Jiang, L.; Lin, Q.; Fang, M.; Li, N.; Hua, X.; Chen, Z.; Bao,
C.; Xu, J.; Du, W.; Zhang, L.; Zhao, Y.; Zhu, L.; Loscalzo, J.; Yang, Y.
Visualizing RNA dynamics in live cells with bright and stable
fluorescent RNAs. Nat. Biotechnol. 2019, 37 (11), 1287−1293.
(19) Urbanek, M. O.; Galka-Marciniak, P.; Olejniczak, M.;
Krzyzosiak, W. J. RNA imaging in living cells - methods and
applications. RNA Biol. 2014, 11 (8), 1083−1095.
ACKNOWLEDGMENTS
■
We gratefully acknowledge Prof. Tom Brown at the University
of Oxford for fruitful discussions about the triphosphate
synthesis and Linda Thunberg at AstraZeneca Gothenburg for
help with purification of the triphosphate. This work was
conducted as part of the FoRmulaEx research center for
nucleotide delivery and with associated financial support to
E.K.E. and L.M.W. from the Swedish Foundation for Strategic
Research (SSF, Grant No. IRC15-0065). F.L.-A., I.S., and
M.H. acknowledge funding from Institut Pasteur.
(20) Bertrand, E.; Chartrand, P.; Schaefer, M.; Shenoy, S. M.; Singer,
R. H.; Long, R. M. Localization of ASH1 mRNA Particles in Living
Yeast. Mol. Cell 1998, 2 (4), 437−445.
REFERENCES
■
(21) Yan, X.; Hoek, T. A.; Vale, R. D.; Tanenbaum, M. E. Dynamics
of Translation of Single mRNA Molecules In Vivo. Cell 2016, 165
(4), 976−989.
(22) Wang, C.; Han, B.; Zhou, R.; Zhuang, X. Real-Time Imaging of
Translation on Single mRNA Transcripts in Live Cells. Cell 2016, 165
(4), 990−1001.
(1) Crooke, S. T.; Witztum, J. L.; Bennett, C. F.; Baker, B. F. RNA-
Targeted Therapeutics. Cell Metab. 2018, 27 (4), 714−739.
(2) Dowdy, S. F. Overcoming cellular barriers for RNA therapeutics.
Nat. Biotechnol. 2017, 35 (3), 222−229.
(3) Crooke, S. T.; Wang, S.; Vickers, T. A.; Shen, W.; Liang, X.-h.
Cellular uptake and trafficking of antisense oligonucleotides. Nat.
Biotechnol. 2017, 35, 230.
(23) Mamot, A.; Sikorski, P. J.; Warminski, M.; Kowalska, J.;
Jemielity, J. Azido-Functionalized 5′ Cap Analogues for the
Preparation of Translationally Active mRNAs Suitable for Fluorescent
Labeling in Living Cells. Angew. Chem., Int. Ed. 2017, 56 (49),
15628−15632.
(4) Pei, D.; Buyanova, M. Overcoming Endosomal Entrapment in
Drug Delivery. Bioconjugate Chem. 2019, 30 (2), 273−283.
(5) Gilleron, J.; Querbes, W.; Zeigerer, A.; Borodovsky, A.; Marsico,
G.; Schubert, U.; Manygoats, K.; Seifert, S.; Andree, C.; Stöter, M.;
Epstein-Barash, H.; Zhang, L.; Koteliansky, V.; Fitzgerald, K.; Fava,
E.; Bickle, M.; Kalaidzidis, Y.; Akinc, A.; Maier, M.; Zerial, M. Image-
based analysis of lipid nanoparticle−mediated siRNA delivery,
intracellular trafficking and endosomal escape. Nat. Biotechnol. 2013,
31 (7), 638−646.
(24) Anhäuser, L.; Huwel, S.; Zobel, T.; Rentmeister, A. Multiple
̈
covalent fluorescence labeling of eukaryotic mRNA at the poly(A) tail
enhances translation and can be performed in living cells. Nucleic
Acids Res. 2019, 47 (7), e42−e42.
(25) Croce, S.; Serdjukow, S.; Carell, T.; Frischmuth, T. Chemo-
enzymatic Preparation of Functional Click-Labeled Messenger RNA.
ChemBioChem 2020, 21 (11), 1641−1646.
(6) Armitage, B. A. In Heterocyclic Polymethine Dyes: Synthesis,
Properties and Applications; Strekowski, L., Ed.; Springer: Berlin,
Heidelberg, 2008; pp 11−29.
(26) Ziemniak, M.; Szabelski, M.; Lukaszewicz, M.; Nowicka, A.;
Darzynkiewicz, E.; Rhoads, R. E.; Wieczorek, Z.; Jemielity, J.
Synthesis and evaluation of fluorescent cap analogues for mRNA
labelling. RSC Adv. 2013, 3 (43), 20943−20958.
(7) Zhang, Y.; Kleiner, R. E. A Metabolic Engineering Approach to
Incorporate Modified Pyrimidine Nucleosides into Cellular RNA. J.
Am. Chem. Soc. 2019, 141, 3347−3351.
(27) Xu, W.; Chan, K. M.; Kool, E. T. Fluorescent nucleobases as
tools for studying DNA and RNA. Nat. Chem. 2017, 9, 1043.
(28) Sinkeldam, R. W.; Greco, N. J.; Tor, Y. Fluorescent Analogs of
Biomolecular Building Blocks: Design, Properties, and Applications.
Chem. Rev. 2010, 110 (5), 2579−2619.
(29) Wilhelmsson, L. M. Fluorescent nucleic acid base analogues. Q.
Rev. Biophys. 2010, 43 (2), 159−183.
(30) Xu, W.; Chan, K. M.; Kool, E. T. Fluorescent nucleobases as
tools for studying DNA and RNA. Nat. Chem. 2017, 9 (11), 1043−
1055.
(8) Custer, T. C.; Walter, N. G. In vitro labeling strategies for in
cellulo fluorescence microscopy of single ribonucleoprotein machines.
Protein Sci. 2017, 26 (7), 1363−1379.
(9) Shimomura, O.; Johnson, F. H.; Saiga, Y. Extraction, Purification
and Properties of Aequorin, a Bioluminescent Protein from the
Luminous Hydromedusan, Aequorea. J. Cell. Comp. Physiol. 1962, 59
(3), 223−239.
(10) Jung, G. In Fluorescent Analogues of Biomolecular Building
Blocks: Design and Applications; Wilhelmsson, L. M.; Tor, Y., Eds.;
Wiley, 2016; pp 55−90.
5423
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(31) Börjesson, K.; Preus, S.; El-Sagheer, A. H.; Brown, T.;
Albinsson, B.; Wilhelmsson, L. M. Nucleic Acid Base Analog FRET-
Pair Facilitating Detailed Structural Measurements in Nucleic Acid
Containing Systems. J. Am. Chem. Soc. 2009, 131 (12), 4288−4293.
(52) Aurup, H.; Siebert, A.; Benseler, F.; Williams, D.; Eckstein, F.
Translation of 2′-modified mRNA in vitro and in vivo. Nucleic Acids
Res. 1994, 22 (23), 4963−4968.
(53) Smith, C. C.; Hollenstein, M.; Leumann, C. J. The synthesis
and application of a diazirine-modified uridine analogue for
investigating RNA−protein interactions. RSC Adv. 2014, 4 (89),
48228−48235.
(54) Domnick, C.; Eggert, F.; Kath-Schorr, S. Site-specific enzymatic
introduction of a norbornene modified unnatural base into RNA and
application in post-transcriptional labeling. Chem. Commun. 2015, 51
(39), 8253−8256.
(55) Temiakov, D.; Patlan, V.; Anikin, M.; McAllister, W. T.;
Yokoyama, S.; Vassylyev, D. G. Structural Basis for Substrate
Selection by T7 RNA Polymerase. Cell 2004, 116 (3), 381−391.
(56) Lakowicz, J. R. In Principles of Fluorescence Spectroscopy, 3rd ed.;
Springer, 2006; pp 277−330.
(32) Fuchtbauer, A. F.; Preus, S.; Börjesson, K.; McPhee, S. A.;
̈
Lilley, D. M. J.; Wilhelmsson, L. M. Fluorescent RNA cytosine
analogue − an internal probe for detailed structure and dynamics
investigations. Sci. Rep. 2017, 7 (1), 2393.
(33) Fuchtbauer, A. F.; Wranne, M. S.; Bood, M.; Weis, E.; Pfeiffer,
̈
P.; Nilsson, J. R.; Dahlén, A.; Grøtli, M.; Wilhelmsson, L. M. Interbase
FRET in RNA: from A to Z. Nucleic Acids Res. 2019, 47 (19), 9990−
9997.
(34) Liu, W.; Shin, D.; Ng, M.; Sanbonmatsu, K. Y.; Tor, Y.;
Cooperman, B. S. Stringent Nucleotide Recognition by the Ribosome
at the Middle Codon Position. Molecules 2017, 22 (9), 1427.
(35) Stengel, G.; Urban, M.; Purse, B. W.; Kuchta, R. D.
Incorporation of the Fluorescent Ribonucleotide Analogue tCTP by
T7 RNA Polymerase. Anal. Chem. 2010, 82 (3), 1082−1089.
(36) McCoy, L. S.; Shin, D.; Tor, Y. Isomorphic Emissive GTP
Surrogate Facilitates Initiation and Elongation of in Vitro Tran-
scription Reactions. J. Am. Chem. Soc. 2014, 136 (43), 15176−15184.
(37) Li, Y.; Fin, A.; McCoy, L.; Tor, Y. Polymerase-Mediated Site-
Specific Incorporation of a Synthetic Fluorescent Isomorphic G
Surrogate into RNA. Angew. Chem., Int. Ed. 2017, 56 (5), 1303−1307.
(38) Tanpure, A. A.; Srivatsan, S. G. A Microenvironment-Sensitive
Fluorescent Pyrimidine Ribonucleoside Analogue: Synthesis, Enzy-
matic Incorporation, and Fluorescence Detection of a DNA Abasic
Site. Chem. - Eur. J. 2011, 17 (45), 12820−12827.
(57) Stengel, G.; Urban, M.; Purse, B. W.; Kuchta, R. D. High
Density Labeling of Polymerase Chain Reaction Products with the
Fluorescent Base Analogue tCo. Anal. Chem. 2009, 81 (21), 9079−
9085.
(58) Stengel, G.; Purse, B. W.; Wilhelmsson, L. M.; Urban, M.;
Kuchta, R. D. Ambivalent Incorporation of the Fluorescent Cytosine
Analogues tC and tCo by Human DNA Polymerase α and Klenow
Fragment. Biochemistry 2009, 48 (31), 7547−7555.
̀
(59) Noyon, C.; Roumeguere, T.; Delporte, C.; Dufour, D.; Cortese,
̀
M.; Desmet, J.-M.; Lelubre, C.; Rousseau, A.; Poelvoorde, P.; Neve, J.;
Vanhamme, L.; Boudjeltia, K. Z.; Van Antwerpen, P. The presence of
modified nucleosides in extracellular fluids leads to the specific
incorporation of 5-chlorocytidine into RNA and modulates the
transcription and translation. Mol. Cell. Biochem. 2017, 429 (1), 59−
71.
(39) Manna, S.; Srivatsan, S. G. Synthesis and Enzymatic
Incorporation of a Responsive Ribonucleoside Probe That Enables
Quantitative Detection of Metallo-Base Pairs. Org. Lett. 2019, 21
(12), 4646−4650.
(40) Yoshikawa, M.; Kato, T.; Takenishi, T. Studies of
Phosphorylation. III. Selective Phosphorylation of Unprotected
Nucleosides. Bull. Chem. Soc. Jpn. 1969, 42 (12), 3505−3508.
(41) Ludwig, J.; Eckstein, F. Rapid and efficient synthesis of
nucleoside 5′-0-(1-thiotriphosphates), 5′-triphosphates and 2′,3′-
cyclophosphorothioates using 2-chloro-4H-1,3,2-benzodioxaphos-
phorin-4-one. J. Org. Chem. 1989, 54 (3), 631−635.
(42) Roy, B.; Depaix, A.; Périgaud, C.; Peyrottes, S. Recent Trends
in Nucleotide Synthesis. Chem. Rev. 2016, 116 (14), 7854−7897.
(43) Burgess, K.; Cook, D. Syntheses of Nucleoside Triphosphates.
Chem. Rev. 2000, 100 (6), 2047−2060.
(44) Flamme, M.; McKenzie, L. K.; Sarac, I.; Hollenstein, M.
Chemical methods for the modification of RNA. Methods 2019, 161,
64−82.
(45) Hocek, M. Enzymatic Synthesis of Base-Functionalized Nucleic
Acids for Sensing, Cross-linking, and Modulation of Protein−DNA
Binding and Transcription. Acc. Chem. Res. 2019, 52 (6), 1730−1737.
(46) Sarac, I.; Meier, C. Efficient Automated Solid-Phase Synthesis
of DNA and RNA 5′-Triphosphates. Chem. - Eur. J. 2015, 21 (46),
16421−16426.
(47) Gaur, R. K.; Sproat, B. S.; Krupp, G. Novel solid phase
synthesis of 2′-o-methylribonucleoside 5′-triphosphates and their α-
thio analogues. Tetrahedron Lett. 1992, 33 (23), 3301−3304.
(48) Sarac, I.; Hollenstein, M. Terminal Deoxynucleotidyl Trans-
ferase in the Synthesis and Modification of Nucleic Acids.
ChemBioChem 2019, 20 (7), 860−871.
(49) Motea, E. A.; Berdis, A. J. Terminal deoxynucleotidyl
transferase: The story of a misguided DNA polymerase. Biochim.
Biophys. Acta, Proteins Proteomics 2010, 1804 (5), 1151−1166.
(50) Srivatsan, S. G.; Tor, Y. Fluorescent Pyrimidine Ribonucleo-
tide: Synthesis, Enzymatic Incorporation, and Utilization. J. Am.
Chem. Soc. 2007, 129 (7), 2044−2053.
(51) Vaught, J. D.; Dewey, T.; Eaton, B. E. T7 RNA Polymerase
Transcription with 5-Position Modified UTP Derivatives. J. Am.
Chem. Soc. 2004, 126 (36), 11231−11237.
5424
https://doi.org/10.1021/jacs.1c00014
J. Am. Chem. Soc. 2021, 143, 5413−5424