Aptamer-based proximity labeling guides covalent RNA modification

Article abstract of DOI:10.1039/d1cc00786f

We describe the development of a proximity-induced bio-orthogonal inverse electron demand Diels-Alder reaction that exploits the high-affinity interaction between a dienophile-modified RhoBAST aptamer and its tetramethyl rhodamine methyltetrazine substrate. We applied this concept for covalent RNA labeling in proof-of-principle experiments.

Full text of DOI:10.1039/d1cc00786f

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Aptamer-based proximity labeling guides covalent
RNA modification†
Cite this: Chem. Commun., 2021,
7, 3480
5
a
a
a
ab
Daniel Englert, Regina Matveeva, Murat Sunbul, * Richard Wombacher* and
a
Received 16th February 2021,
Accepted 4th March 2021
Andres Jaschke
*
¨
DOI: 10.1039/d1cc00786f
rsc.li/chemcomm
7
8,9
We describe the development of a proximity-induced bio-orthogonal synthesis. The former approach repurposes methyl transferases
10
inverse electron demand Diels–Alder reaction that exploits the high- or poly(A) polymerases, which modify the mRNA cap structure or
affinity interaction between a dienophile-modified RhoBAST aptamer the 3 -end, respectively. However, these methods are unspecific and
0
and its tetramethyl rhodamine methyltetrazine substrate. We applied this label all RNAs in a sample, while sequence-specific labeling of an
1
1,12
concept for covalent RNA labeling in proof-of-principle experiments.
ROI has been accomplished using tRNA-modifying enzymes.
The latter strategy uses the incorporation of the modified nucleo-
13,14
The site-specific modification of biomolecules, e.g. with fluores- tides during enzymatic synthesis.
Applying this technique,
cent dyes or crosslinkers, is an important yet exceedingly complex various dienophiles were introduced into RNA and subsequently
task. The limited number of different functional groups present in labeled via inverse electron demand Diels–Alder (IEDDA) reaction
15,16
a given macromolecule represents a serious hurdle for achieving in vitro.
An important advantage of enzymatic incorporation is
single-site selectivity. In the field of protein modification, one of its potential in vivo application in cells in the context of metabolic
the successful solutions to this problem utilizes proximity- labeling. For this purpose, several different modified nucleoside
1–3
17
18
19
enhanced reactivity. In these approaches, a high-affinity non- analogues bearing ethynyl, azido or vinyl groups were synthe-
covalent interaction (e.g., between an enzyme and its substrate) is sized and applied in mammalian cells. However, since the modified
used to position certain reactive groups of the two binding nucleoside triphosphate (NTP) is incorporated stochastically, it is
partners in close proximity to each other, thereby causing an impossible to control the site of modification, rendering this
increased reaction rate at these positions but not at others and method unsuitable for specific labeling of a ROI.
allowing specific labeling. Here we transfer this concept to
covalent RNA modification, using a high-affinity aptamer–dye ubiquitous incorporation of ‘‘clickable’’ nucleotides into an aptamer
interaction as the guiding principle. tag and a subsequent proximity-induced selective bio-orthogonal
Aside from a few ribozyme-based methods that directly reaction. We use our recently developed fluorescent light-up apta-
Here, we report a solution to this problem by combining the
4
–6
attach the label,
the majority of currently used covalent mer RhoBAST and its fluorogenic target tetramethyl rhodamine
20
RNA labeling techniques are based on bio-orthogonal click dinitroaniline conjugate (TMR-DN) as a starting point and exploit
reactions and require the initial introduction of a chemical the high binding affinity and selectivity of the aptamer towards the
handle to the RNA of interest (ROI) in the first step. The 5-carboxy tetramethyl rhodamine (TMR) moiety to establish a
introduced functional group is subsequently derivatized via proximity-induced reaction. Owing to the high selectivity of the
the corresponding click reaction. Based on the time point of IEDDA reaction in vitro and in cellular environment, we decided to
incorporation of the reactive handle, these RNA labeling techniques enzymatically incorporate dienophile-modified nucleotides into the
can be classified into two categories: postsynthetic labeling strate- aptamer and use a TMR-methyltetrazine substrate (TMR-Tz) as a
21
gies and the direct incorporation of modified nucleotides during reactive diene (Fig. 1). Binding of TMR-Tz to the modified
RhoBAST aptamer should lead to a close proximity of diene and
a
dienophile and consequently to an enhanced reactivity in the IEDDA
reaction.
Institute of Pharmacy and Molecular Biotechnology, Heidelberg University,
Im Neuenheimer Feld 364, Heidelberg 69120, Germany.
E-mail: jaeschke@uni-hd.de, msunbul@uni-heidelberg.de
wombacher@uni-heidelberg.de
We decided to introduce the required dienophile moiety on
the C -position of uridine triphosphate (UTP), since a wide
b
5
Department of Chemical Biology, Max Planck Institute for Medical Research,
Jahnstrasse 29, Heidelberg 69120, Germany
5
variety of UTP derivatives with C -modifications are known to
be accepted by T7 RNA polymerase (T7 RNAP) and incorporated
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
1
3,15,22
d1cc00786f
into RNA.
Overall, we chose a set of four small and stable
3
480
|
Chem. Commun., 2021, 57, 3480–3483
This journal is © The Royal Society of Chemistry 2021

View Article Online
Communication
ChemComm
Fig. 1 Proximity-induced covalent RNA labeling. (A) Schematic illustration of the proximity-induced IEDDA reaction using a dienophile-modified
RhoBAST aptamer and TMR-Tz substrate. Chemical structures of TMR-Tz (B) and the synthesized dienophile-substituted uridine triphosphates for
enzymatic incorporation (C). Novel modified triphosphate derivatives (modUTP) are highlighted with a yellow box.
dienophiles with low-to-medium intrinsic reactivity (Fig. 1C). functionalization. In the final step, the different dienophile moieties
The reactivity of the dienophiles has to be low enough not to were introduced by amide coupling using the corresponding organic
cause any unspecific reaction but high enough to be accelerated acids as activated NHS-esters yielding the desired modified uridine
upon induced proximity to result in a specific labeling. There- triphosphates taUTP, cpUTP and norUTP.
1
5
fore, we prepared a vinyl-substituted vUTP derivative and
Next, we tested the enzymatic incorporation of the modified
three novel modified UTP derivatives bearing dienophile moieties UTPs in an in vitro transcription reaction using T7 RNAP, fully
with increasing ring strain and thus increasing reactivity in the replacing UTP by the modified UTPs in the reaction mixture. As
IEDDA reaction (terminal alkene o cyclopentene o norbo- initial test, we transcribed the RhoBAST aptamer (55 nt), which
21,23
rnene).
All derivatives were synthesized starting from uridine, contains 7 uridines, from a double-stranded DNA template. The
which in the first step was transferred to 5-iodo-uridine (1) by obtained transcription mixtures were analyzed by denaturing
oxidative iodination (Scheme 1). Stille cross-coupling with tributyl polyacrylamide gel electrophoresis (PAGE). For all nucleotides,
(vinyl)tin provided the vinyl-substituted uridine derivative vU, which successful incorporation was observed and full-length RNA
was converted to its nucleoside triphosphate equivalent vUTP using products were detected. As anticipated, the modified RNAs
15
a standard phosphorylation procedure. The novel modified showed a slightly lower mobility in PAGE, compared to the
uridine triphosphates taUTP, cpUTP and norUTP were prepared unmodified transcription product (Fig. S1, ESI†). The transcrip-
in 5 steps starting from 1 utilizing Sonogashira coupling with tion yield was determined by quantifying the ethidium
N-trifluoroacetyl propargyl amine. In order to increase the flexibility bromide-stained full-length RNA product band. Reactions with
of the attached linker, the alkyne was hydrogenated using Adams’ the modified analogs vUTP, taUTP and cpUTP showed excellent
catalyst to yield 3. In the next step, the modified nucleoside 3 was transcription yields (480%) relative to the control reaction
phosphorylated and deprotected to afford the precursor aminoUTP, using UTP. Even with the sterically demanding norbornene
bearing
a
primary amino group that enables late-stage derivative norUTP, transcription yields 450% were achieved.
The enzymatic incorporation of the modified UTPs was confirmed
by LC-MS analysis of enzymatically digested PAGE-purified transcrip-
tion products (Fig. S2, ESI†).
Subsequently, we tested if the commercial TMR-Tz substrate
binds to the unmodified RhoBAST aptamer and if the intro-
duced uridine modifications inhibit binding. As negative con-
trol, we designed an inactive version of the RhoBAST aptamer,
also containing 7 uridines, by mutating essential and conserved
2
0
regions (Fig. S3A, ESI†). As expected, inactive RhoBAST did
not increase the fluorescence of our previously developed turn-
2
4
on probe TMR-DN, confirming that inactive RhoBAST lost its
affinity to TMR derivatives (Fig. S3B, ESI†). After incubation
with TMR-Tz, we observed comparable fluorescence enhance-
ment (B20-fold) for the unmodified RhoBAST aptamer and the
modified aptamers with incorporated vU, taU or cpU (Fig. S3C,
ESI†). Only the norbornene-modified version displayed a notably
lower fluorescence enhancement (B12-fold turn-on). Regardless of
the modification, the inactive control RNAs showed no significant
Scheme 1 Synthesis of novel modified uridine triphosphates (modUTP).
(
a) I
c) H
2
, HNO
, PtO
N; (e) NH
3
, CHCl
, MeOH; (d) POCl
2 4 7
(aq); (f) NHS-ester, Na B O
3
; (b) TFA-propagyl amine, Pd(PPh
, proton sponge, TMP, (n-Bu
buffer, DMF.
3
)
4
, CuI, NEt
3
, DMF;
(
2
2
3
NH) H P O ,
3 2 2 2 7
n-Bu
3
3
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3480–3483 | 3481

View Article Online
ChemComm
Communication
turn-on. Furthermore, non-covalent complex formation between enhancement over the inactive norU aptamer. Comparing the
TMR-Tz and active aptamer was indicated by a bathochromic shift active to the inactive vU-RhoBAST aptamer, a more than 30-fold
of the excitation and emission spectra of TMR-Tz (Fig. S4, ESI†). To higher IEDDA rate for the proximity-induced reaction (Fig. S8,
further characterize the binding properties of the dienophile- ESI†) and a labeling yield of 10% after 24 h (Fig. S9, ESI†) could be
modified RhoBAST aptamers, we determined their dissociation observed. Based on these results, we focused on the vU-modified
constants in complex with TMR-Tz (Fig. S5, ESI†). Introduction of RNAs for further experiments. Investigation of the initial rate of
C
5
-modified uridines caused about one order of magnitude loss in the IEDDA reaction as a function of the TMR-Tz substrate
binding affinity compared to the unmodified RhoBAST aptamer concentration revealed for active vU-RhoBAST a saturation-type
= 5 nM). We observed a trend between the sterical demand curve reminiscent of the Michaelis–Menten plots of enzyme-
(K
D
of the modification and the binding affinity: the aptamer with catalyzed reactions, indicating the formation of an aptamer–
the smallest vU modification exhibited the strongest affinity substrate complex at equilibrium, followed by the rate-
(KD = 44 nM) while the one with the bulky norU modification determining chemical step, the IEDDA reaction (Fig. S10, ESI†).
showed the weakest affinity (K = 79 nM). In contrast, inactive vU-RhoBAST showed a linear concentration-
D
We next examined if the RNA-appended dienophiles would rate profile, as expected for a second-order reaction without
still react in an IEDDA reaction. For this purpose, active and proximity effects (Fig. S11, ESI†). This difference in the reaction
inactive RhoBAST RNAs (1 mM) were incubated with an excess kinetics means that at high TMR-Tz concentrations the highest
of TMR-Tz (5 mM) for 20 hours. PAGE gel analysis of the labeling rates but the lowest selectivities are achieved. Conversely, low
reactions revealed fluorescent bands with reduced mobility for substrate concentrations allow highly specific modification, albeit
all modified RNAs independent of the RNA sequence (Fig. S6, at a low rate. Mutation analysis indicated uridine U39 as the major
ESI†), confirming the IEDDA-reactivity of the modified RNAs. In site of the IEDDA reaction (Fig. S12, ESI†).
contrast, no fluorescent reaction product was detected for the
unmodified RhoBAST aptamer.
To further support the conclusion that the increased IEDDA
rate of active vU-RhoBAST RNA is based on induced proximity,
We then investigated if binding of the TMR-Tz substrate to we performed an experiment in which the formation of the
the dienophile-modified aptamers induces an increase of the RhoBAST*TMR-Tz complex was challenged by a competitor. For
reaction rate. Thus, we compared the reaction kinetics of the this purpose, we used TMR, a known high-affinity binder of
2
0,24
active and inactive RhoBAST aptamers using an equimolar RhoBAST, in 1000-fold excess over TMR-Tz.
As expected, the
concentration (1 mM) of TMR-Tz (Fig. 2A and Fig. S7, ESI†). addition of TMR to the proximity-induced reaction resulted in a
For the reaction of the inactive RhoBAST samples, the expected dramatic drop of reaction product to roughly the level observed
reactivity trend was observed: the RNAs modified with vinyl (vU) for the inactive mutant (Fig. 2B). Addition of TMR to the IEDDA
and the terminal alkene (taU) displayed the lowest IEDDA rate, reaction of the inactive RhoBAST mutant, on the other hand,
whereas RNAs with incorporated cyclic dienophiles (cpU and had no influence on the reaction yield. The IEDDA reaction of
norU) reacted faster. Overall, the reaction rate increased with vU-RhoBAST was additionally performed with 1 mM of tetrazine-
ring strain (vU B taU o cpU { norU). Importantly, a different modified fluorescein (FAM-Tz) and cyanine (Cy5-Tz) dyes that
2
4
reactivity pattern was observed for the modified active RhoBAST are not expected to bind RhoBAST (Fig. S13A, ESI†). Only the
aptamers. The active taU and cpU aptamers showed IEDDA rates combination of vU-RhoBAST aptamer and TMR-Tz resulted in a
similar to their inactive counterparts, implying the absence of significantly increased product formation of active vs. inactive
proximity effect. In contrast, remarkably higher rates were mea- aptamer, further substantiating the importance of aptamer–
sured for the active vU and norU RhoBAST aptamers, compared to substrate complex formation. Using higher concentrations
their inactive mutants, indicating a proximity-induced reactivity (100 mM) of FAM-Tz and Cy5-Tz resulted in unspecific multiple
enhancement caused by binding of TMR-Tz to the dienophile- labeling of vU-modified RNAs (Fig. S13B, ESI†).
modified RhoBAST aptamers. The active norU RhoBAST aptamer
showed the highest reaction rate, but only
To demonstrate the specificity of our proximity-based
a
two-fold IEDDA system, we treated a mixture of different in vitro tran-
scribed vU-modified RNAs with TMR-Tz. RhoBAST (55 nt), an
inactive RhoBAST mutant with an elongated stem (67 nt), and
an unrelated 39 mer (1 mM each) were mixed and treated with
TMR-Tz (1 mM). Analysis of the reaction products showed that
only the RhoBAST RNA was significantly labeled with TMR-Tz
and hence confirmed the expected selectivity of the proximity-
induced reaction (Fig. 3A). As a control, all modified RNAs were
labeled unspecifically using high concentrations of Cy5-Tz
(5 mM). Encouraged by these in vitro results, we aimed to apply
our proximity-induced covalent IEDDA approach to RNAs tran-
scribed inside living cells by combining it with metabolic
labeling. vU is known to be taken up from the medium,
metabolically converted to vUTP, and incorporated into cellular
RNAs, although the incorporation efficiency appears to be
Fig. 2 IEDDA reaction of modified active and inactive RhoBAST aptamers.
A) Kinetics of IEDDA reactions with equimolar concentrations of RNA and
(
TMR-Tz (1 mM) at 37 1C. (B) Yield of the IEDDA reaction (37 1C, 20 h) of
vU-modified RhoBAST RNAs (1 mM) with TMR-Tz (1 mM) in the presence or
absence of excess competitor TMR (1 mM).
3
482
|
Chem. Commun., 2021, 57, 3480–3483
This journal is © The Royal Society of Chemistry 2021

View Article Online
Communication
ChemComm
Fig. 3 Covalent labeling of vU-RhoBAST in RNA mixtures. (A) Denaturing PAGE gel (20%) analysis of individual and one-pot labeling of vU-modified active
55 nt) and inactive (elongated version, 67 nt) RhoBAST and vU-modified 39 mer (1 mM each) with TMR-Tz (1 mM) and Cy5-Tz (5 mM). (B) PAGE gel (10%) analysis
of labeling reactions of total RNA isolated from E. coli transcribing active and inactive RhoBAST aptamers. Bacteria were grown in the presence or absence of
(
ꢀ1
vU (2 mM). IEDDA reactions were performed using 50 mM TMR-Tz and 4 mg ml total RNA at 37 1C for 20 h (10 mg sample per lane).
1
9
rather low. Here, we cloned active and inactive RhoBAST
aptamers into a pET vector and transformed them into E. coli
BL21 (DE3). Then, the E. coli strains were grown in the presence
and absence of vU (2 mM), and total RNA was isolated. After
refolding of the RNA, we incubated the samples with TMR-Tz
and performed PAGE analysis using in-gel fluorescence and
EtBr staining (Fig. 3B). For total RNA samples extracted from
E. coli growing in the absence of vU (negative control), no
fluorescent signal was observed, whereas in the isolated RNA
samples from E. coli growing in presence of vU, only one distinct
band was detected for the sample containing the active RhoBAST
aptamers. This fluorescent band showed the same mobility on the
2 S. S. Gallagher, J. E. Sable, M. P. Sheetz and V. W. Cornish,
ACS Chem. Biol., 2009, 4, 547–556.
3
M. Sunbul, L. Nacheva and A. J ¨a schke, Bioconjugate Chem., 2015, 26,
466–1469.
1
4 M. Ghaem Maghami, C. P. M. Scheitl and C. H o¨ bartner, J. Am. Chem.
Soc., 2019, 141, 19546–19549.
R. I. McDonald, J. P. Guilinger, S. Mukherji, E. A. Curtis, W. I. Lee
and D. R. Liu, Nat. Chem. Biol., 2014, 10, 1049.
6 A. K. Sharma, J. J. Plant, A. E. Rangel, K. N. Meek, A. J. Anamisis,
5
J. Hollien and J. M. Heemstra, ACS Chem. Biol., 2014, 9,
1
680–1684.
D. Schulz and A. Rentmeister, ChemBioChem, 2014, 15, 2342–2347.
8 J. M. Holstein, D. Schulz and A. Rentmeister, Chem. Commun., 2014,
0, 4478–4481.
J. M. Holstein, D. Stummer and A. Rentmeister, Chem. Sci., 2015, 6,
362–1369.
7
5
9
1
PAGE gel as the in vivo transcribed and ethidium bromide stained 10 M.-L. Winz, A. Samanta, D. Benzinger and A. J ¨a schke, Nucleic Acids
Res., 2012, 40, e78.
1 F. Li, J. Dong, X. Hu, W. Gong, J. Li, J. Shen, H. Tian and J. Wang,
Angew. Chem., Int. Ed., 2015, 54, 4597–4602.
12 S. C. Alexander, K. N. Busby, C. M. Cole, C. Y. Zhou and
N. K. Devaraj, J. Am. Chem. Soc., 2015, 137, 12756–12759.
3 J. D. Vaught, T. Dewey and B. E. Eaton, J. Am. Chem. Soc., 2004, 126,
RhoBAST. In contrast, no signal was detected for the inactive
RhoBAST sample, demonstrating the specific labeling of RhoBAST
in in vivo transcribed complex RNA mixtures.
In conclusion, we described the first application of in vitro
selected RNA aptamers to increase the reaction rate of a biortho-
1
1
11231–11237.
gonal IEDDA reaction by proximity effects. We could clearly 14 N. Milisavljevi ˇc , P. Perl ´ı kov ´a , R. Pohl and M. Hocek, Org. Biomol.
Chem., 2018, 16, 5800–5807.
demonstrate that the over 30-fold reactivity increase is due to the
specific interaction between aptamer and target. This novel princi-
1
5 J. T. George and S. G. Srivatsan, Bioconjugate Chem., 2017, 28,
529–1536.
ple enabled us to specifically label an in vitro or in vivo transcribed 16 P. Asare-Okai, E. Agustin, D. Fabris and M. Royzen, Chem. Commun.,
1
2
014, 50, 7844–7847.
7 C. Y. Jao and A. Salic, Proc. Natl. Acad. Sci. U. S. A., 2008, 105,
5779–15784.
ROI in a mixture of modified RNAs. However, for wide-spread
practical application as a covalent labeling tool the reaction rate
must be increased significantly without sacrificing selectivity.
Open Access funding provided by the Max Planck Society.
1
1
18 S. Nainar, S. Beasley, M. Fazio, M. Kubota, N. Dai, I. R. Corr ˆe a Jr and
R. C. Spitale, ChemBioChem, 2016, 17, 2149–2152.
9 M. Kubota, S. Nainar, S. M. Parker, W. England, F. Furche and
R. C. Spitale, ACS Chem. Biol., 2019, 14, 1698–1707.
1
2
0 M. Sunbul, J. Lackner, A. Martin, D. Englert, B. Hacene, F. Gr u¨ n,
K. Nienhaus, G. U. Nienhaus and A. Ja
DOI: 10.1038/s41587-020-00794-3.
¨
schke, Nat. Biotechnol., 2021,
Conflicts of interest
2
2
2
1 B. Oliveira, Z. Guo and G. Bernardes, Chem. Soc. Rev., 2017, 46,
895–4950.
2 M. Flamme, L. K. McKenzie, I. Sarac and M. Hollenstein,
Methods, 2019, 161, 64–82.
3 P. Werther, J. S. M o¨ hler and R. Wombacher, Chem. – Eur. J., 2017,
23, 18216–18224.
There are no conflicts to declare.
4
Notes and references
1
S. Tsukiji, M. Miyagawa, Y. Takaoka, T. Tamura and I. Hamachi,
Nat. Chem. Biol., 2009, 5, 341–343.
24 M. Sunbul and A. J ¨a schke, Nucleic Acids Res., 2018, 46, e110.
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3480–3483 | 3483