A modified dinucleotide for site-specific RNA-labelling by transcription priming and click chemistry

Article abstract of DOI:10.1039/c3cc46132g

An improved strategy for RNA labelling using an alkyne-carrying dinucleotide is reported. This involves near-quantitative priming by phage RNA-polymerases followed by conjugation of different labels using click chemistry. Moreover, these transcripts bear a ligation compatible 5′-end, and thus through ligation the terminal label can be transformed to an internal one. The Royal Society of Chemistry.

Full text of DOI:10.1039/c3cc46132g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
A modified dinucleotide for site-specific
RNA-labelling by transcription priming and
click chemistry†
Cite this: Chem. Commun., 2014,
50, 1313
Received 11th August 2013,
Accepted 16th October 2013
Ayan Samanta,‡ Andre Krause‡ and Andres Jaschke*
´
¨
DOI: 10.1039/c3cc46132g
www.rsc.org/chemcomm
An improved strategy for RNA labelling using an alkyne-carrying To this end, the copper-catalysed azide–alkyne cycloaddition
dinucleotide is reported. This involves near-quantitative priming by (CuAAC) seemed appropriate as a conjugation strategy due to
phage RNA-polymerases followed by conjugation of different labels the bioorthogonal nature of the reaction and the small size of
using click chemistry. Moreover, these transcripts bear a ligation the reactive groups.4
compatible 50-end, and thus through ligation the terminal label can
Based on this idea, we report here the synthesis and enzy-
matic incorporation of a ‘‘clickable’’ dinucleotide (OdUpG,
where O represents octadiynyl) (Fig. 1) for labelling RNA at
be transformed to an internal one.
Modified RNAs have proven to be indispensable in modern life- the 50-end employing T3, SP6, or T7 RNAPs under the control of
science research. While solid-phase RNA synthesis allows for the their cognate class III promoters. This initiator has been found
site-selective introduction of a variety of chemical modifications into to be incorporated by all three polymerases in a near quantita-
short RNAs, the same task poses a major challenge for large RNAs.1 tive manner in several instances. Additionally the applicability
Large unmodified RNAs can be conveniently prepared by in vitro of this labelling strategy has been demonstrated by preparing
run-off transcription using various bacteriophage RNA polymerases various RNA-conjugates from a single transcript as well as by
(RNAPs). Due to the relaxed substrate tolerance of these polymerases labelling long RNAs. Finally, the ligation compatibility of the
at the initiation step it is possible to selectively modify the 50-end of primed transcript has been tested. Although the transcription-
a transcript using modified guanosine or adenosine analogues that priming strategy is primarily a RNA 50-labelling technique, this
are functionalised through the 50-monophosphate.2 These analo- initiator renders it applicable to site-specific, internal modifi-
gues can only be incorporated at the start of transcription because cation of RNA after ligation (Fig. 1).
they lack triphosphate – hence termed as ‘‘initiator nucleotides’’.
The dinucleotide initiator was synthesised following the
However, this approach suffers from two major drawbacks: (1) it standard phosphoramidite coupling strategy (ESI†) with an overall
requires the de novo synthesis of the complete initiator molecule yield of 27%. Although all four nucleosides can be used for attach-
and optimisation for its enzymatic incorporation for each different ment of the alkyne functionality, uridine seemed appropriate as a
initiator, and (2) all currently known initiators generate RNA first attempt due to the commercial availability of the building
transcripts whose 50-ends are blocked and cannot be ligated further. block and lack of a nucleobase protection strategy during phos-
Moreover, as the transcriptional activities of different RNAPs vary phoramidite coupling. Moreover, the deoxy-sugar was chosen for
substantially,3 it is beneficial to have a single optimised labelling- its higher coupling efficiency compared to the ribo-analogue
protocol that is compatible with all commonly used polymerases. during phosphoramidite coupling. The choice of guanosine as a
Therefore it is desirable to have an initiator nucleotide that can be priming nucleotide was inevitable due to the widespread popularity
incorporated into transcripts by a polymerase of choice under of phage class-III promoters in molecular biology.
conditions that are optimised only once, bears a convertible residue
The incorporation of OdUpG by RNAPs was optimised at two
for further postsynthetic conjugation with a variety of different levels – (a) at varying GTP concentrations (Fig. 2A, C and E)
labels and renders the 50-end of the transcript accessible to ligation. and (b) at varying NTP concentrations (Fig. 2B, D and F). While a
large excess of initiator over GTP will result in a high incorpora-
Institute of Pharmacy and Molecular Biotechnology, Heidelberg University,
tion of the dinucleotide into the transcripts, it will hamper the
overall transcription yield since GTP is required during elonga-
tion too. The best NTP concentration was screened always with
Im Neuenheimer Feld 364, Heidelberg 69120, Germany. E-mail: jaeschke@uni-hd.de;
Fax: +49 6221 54 6430; Tel: +49 6221 54 4851
† Electronic supplementary information (ESI) available: Full experimental section.
the optimal GTP concentration determined from the previous
experiment. Since the incorporation of the dinucleotide itself
See DOI: 10.1039/c3cc46132g
‡ Contributed equally.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 1313--1316 | 1313

View Article Online
Communication
ChemComm
Fig. 1 Schematic representation of the RNA-labelling approach (A), chemical
structure of the dinucleotide initiator (B).
induced only a slight shift in the electrophoretic mobility that
was indistinguishable from the un-initiated n + 1 transcript, the
appended alkyne function was subsequently conjugated to a
biotinyl residue by CuAAC using biotin azide, thus enabling an
unambiguous determination of the labelling efficiency by
streptavidin electrophoretic mobility shift assay (Strep-EMSA).
GMP-priming was used as a control in all of these experiments
(Fig. 2, the first lane in every gel picture).
The labelling efficiencies and relative transcription yields are
summarised in Fig. 2 and 3. Based on the utilisation of the
initiator at varying GTP and NTP concentrations, the polymerases
can be categorised into two broader groups, one consisting of T7
and SP6 RNAPs and the other of T3 RNAP. The relative
transcription yield increased with increasing GTP concentra-
tions (at least for the tested concentration range) for T7 and SP6
RNAPs whereas it decreased for T3 RNAP above a certain
threshold concentration (0.4 mM) (Fig. 3A). The labelling
efficiency decreased for all three RNAPs with increasing GTP
concentrations, SP6 being affected the most while T7 the least
(Fig. 3A). Furthermore, within the tested NTP concentration
range, the transcription yield decreased only very slightly for T7
and SP6 RNAPs with increasing NTP concentrations whereas the
Fig. 2 Optimisation of transcriptional labelling by T3 (A and B), SP6 (C and D)
and T7 (E and F) RNAPs. A, C and E denote labelling at varying GTP concentra-
tions, while B, D, and F indicate the same at varying NTP concentrations. ‘a’ and
‘b’ denote streptavidin-shifted biotinylated RNA and unmodified RNA, respec-
tively. The left lane in every gel denotes transcription initiation with GMP. The
initiator nucleotide (GMP/OdUpG) was added to the transcription mixture at a
concentration of 4 mM in every experiment. The purified transcript was
subjected to CuAAC with biotin azide (compound 5), followed by incubation
with streptavidin and gel analysis.
decrease in transcription yield for T3 was found to be drastic if the transcription yield is not the major concern but the
(Fig. 3B). However, the labelling efficiencies remained largely labelling efficiency is, it is recommended to use T7 RNAP since
constant for T3 RNAP, decreased only slightly for T7 and the best labelling efficiency could be achieved with this polymerase
increased drastically for SP6 RNAP with increasing NTP concen- while maintaining a moderate transcription yield.
trations (Fig. 3B). Based on these observations it can be concluded
Although the optimal transcription protocol is always a
that if a transcription has to be performed at a relatively high NTP compromise between labelling efficiency and transcription yield,
concentration, it is desirable to use SP6 RNAP since the transcrip- the best conditions in our opinion were found to be (1) 4 mM
tion yield remains nearly constant but the labelling efficiency initiator, 0.2 mM GTP and 1 mM of the other NTPs for T3 RNAP
increases with increasing NTP concentrations. Similarly if a with a labelling efficiency of 72%, (2) 4 mM initiator, 0.8 mM GTP
transcription needs to be performed at a relatively low NTP and 4 mM of the other NTPs for SP6 RNAP with a labelling
concentration, it is recommended to use T3 RNAP since this efficiency of 82% and (3) 4 mM initiator, 0.8 mM GTP and 1 mM
polymerase resulted in high transcription yield with a moderate of the other NTPs for T7 RNAP with a labelling efficiency of 80%.
labelling efficiency. However, under these low-NTP conditions Initiated transcripts were further characterised by high resolution
1314 | Chem. Commun., 2014, 50, 1313--1316
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
Fig. 3 Overview of the relative transcription yields and labelling efficiencies at varying GTP (A) and NTP (B) concentrations. Broken and solid lines denote
relative yields and labelling efficiencies, respectively. Relative yields were calculated with respect to the corresponding GMP-primed reactions in Fig. 2
(see ESI† for calculation details).
mass spectrometry (ESI,† Table S1). To the best of our knowl- splint (65 nucleotides) (ESI,† Table S2). Nearly 7-fold excess of
edge, this is the only study reporting comparative data on the acceptor RNA over the donor strand along with a mixture of
individual labelling efficiencies and transcription yields of T3, T4 RNA ligase 2, T4 RNA ligase 1 and T4 DNA ligase was found
SP6 and T7 RNAPs with an initiator nucleotide.
to be necessary to achieve efficient ligation (ESI,† Fig. S4). The
To demonstrate the robustness of this click-type labelling requirement of RNA ligase 1, a single strand specific ligase, is
approach, transcription priming was performed with T7 RNAP plausible due to the fact that the currently designed splint does
using the aforesaid optimised conditions from a template carrying not incorporate the added alkyne-bearing nucleoside in a
two 20-OMe substitutions (ESI,† Table S2).5 The purified primed- duplex environment, thus creating a single nucleotide bulge
transcript was then reacted by CuAAC with a variety of commercial at the ligation joint. Current efforts include the optimisation of
and self-synthesised compounds (ESI,† Fig. S2B). Attachment of the ligation protocol for improved ligation yield at stoichio-
these various tags resulted in substantial gel-shifts for the products metric ratios of the donor to acceptor strand with a perfectly
compared to the starting materials and thus enabling a straight- double stranded ligation site. However, it can be concluded, by
forward assessment of the click-conjugation efficiency (ESI,† comparing this result with the ligation results obtained from an
Fig. S2A). These data indicate near-quantitative conversion in all unmodified UpG-primed transcript, that the OdUpG initiator is
cases, irrespective of the nature of the coupling partner. This indeed tolerated by three different classes of enzymes widely
represents a major advancement in comparison to all other used in molecular biology, namely RNA polymerases, poly-
previously described initiators, as one central intermediate is nucleotide kinases and ligases. The resulting ligated RNA can
sufficient to synthesise a wide variety of different conjugates further be derivatised by CuAAC, thereby enabling site-specific,
without the need to synthesise dozens of initiator nucleotides internal modification of RNA, which considerably broadens the
and to optimise their enzymatic incorporation.
scope of this approach.
Furthermore, a glycine riboswitch (length 209 nucleotides)
The combination of this approach with a previously reported
from Bacillus subtilis and a random RNA pool (length 233 tailing-based click modification method for RNA 30-ends and
nucleotides) (ESI†) were subjected to transcriptional labelling internal positions would allow considerable freedom in the
with OdUpG by T7 RNAP under the aforesaid optimised condi- choice of the labelling positions, and also allows for sequential
tions. The resulting transcripts were conjugated with a Cy5- multiple labelling.6
azide by CuAAC (ESI,† Fig. S3A and B, respectively). Both of
Although in this study the initiator dinucleotide was self-
these experiments clearly demonstrate the applicability and synthesised, it can also be purchased from various oligonucleotide
sequence-independence of this RNA-labelling approach for synthesis companies owing to the commercial availability of the
labelling long, structured RNAs that are beyond the limits of monomers. Hence this labelling strategy is completely open to
standard solid-phase chemistry.
people lacking any expertise or equipment in organic synthesis.
The presence of a free 50-OH in the OdUpG primed transcript Moreover, since this labelling-protocol is compatible with all
renders the transcript to be amenable to splinted ligation with commonly used RNAPs in molecular biology, this approach can
another RNA strand. Thus the 50-terminal modification will be directly be implemented on one0s RNA of interest without any
transformed to an internal one as part of a ligation fragment. This further cloning or PCR manipulation of the available plasmids. In
would however require the non-natural 50-nucleoside to be accepted combination with the well-established protocols for enzymatic
as a substrate by both a polynucleotide kinase and a ligase. To test in vitro RNA synthesis, we expect that the initiator dinucleotide
this hypothesis, an OdUpG-primed transcript (25 nucleotides) was described in this study will help to overcome the present short-
subjected to enzymatic 50-end phosphorylation using T4 poly- comings of chemical RNA synthesis, thereby allowing for the
nucleotide kinase (PNK), followed by ligation to an acceptor introduction of variable modifications at the 50-terminus or at
RNA (40 nucleotides) in the presence of a complementary DNA internal positions of RNA by click chemistry.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 1313--1316 | 1315

View Article Online
Communication
ChemComm
3 I. D. Pokrovskaya and V. V. Gurevich, Anal. Biochem., 1994, 220,
420–423.
This work was supported by the DFG Grant #Ja 794/3. The
authors acknowledge S. Ameta, S. Suhm and A. Hertl.
4 A. H. El-Sagheer and T. Brown, Chem. Soc. Rev., 2010, 39, 1388–1405;
A. H. El-Sagheer and T. Brown, Proc. Natl. Acad. Sci. U. S. A., 2010, 107,
15329–15334; J. Gierlich, G. A. Burley, P. M. Gramlich, D. M.
Hammond and T. Carell, Org. Lett., 2006, 8, 3639–3642; J. Gierlich,
K. Gutsmiedl, P. M. Gramlich, A. Schmidt, G. A. Burley and T. Carell,
Chemistry, 2007, 13, 9486–9494; V. Hong, N. F. Steinmetz,
M. Manchester and M. G. Finn, Bioconjugate Chem., 2010, 21,
1912–1916; T. Ishizuka, M. Kimoto, A. Sato and I. Hirao, Chem.
Commun., 2012, 48, 10835–10837; C. Y. Jao and A. Salic, Proc. Natl.
Acad. Sci. U. S. A., 2008, 105, 15779–15784; Y. Motorin, J. Burhenne,
R. Teimer, K. Koynov, S. Willnow, E. Weinhold and M. Helm, Nucleic
Acids Res., 2011, 39, 1943–1952; J. Qiu, A. H. El-Sagheer and T. Brown,
Chem. Commun., 2013, 49, 6959–6961; H. Rao, A. A. Sawant,
A. A. Tanpure and S. G. Srivatsan, Chem. Commun., 2012, 48, 498–500.
Notes and references
¨
1 H. Lonnberg, Bioconjugate Chem., 2009, 20, 1065–1094; E. Paredes,
M. Evans and S. R. Das, Methods, 2011, 54, 251–259.
2 S. Fusz, S. G. Srivatsan, D. Ackermann and M. Famulok, J. Org. Chem.,
2008, 73, 5069–5077; F. Huang, J. He, Y. Zhang and Y. Guo, Nat.
Protocols, 2008, 3, 1848–1861; F. Huang and Y. Shi, Bioorg. Med.
Chem. Lett., 2010, 20, 6254–6257; N. Li, C. Yu and F. Huang, Nucleic
Acids Res., 2005, 33, e37; E. Paredes and S. R. Das, ChemBioChem, 2011,
12, 125–131; C. Pitulle, R. G. Kleineidam, B. Sproat and G. Krupp, Gene,
¨
1992, 112, 101–105; J. Schoch, S. Ameta and A. Jaschke, Chem. Commun.,
2011, 47, 12536–12537; R. Fiammengo, K. Musilek and A. Jaschke, J. Am.
¨
¨
Chem. Soc., 2005, 127, 9271–9276; B. Seelig and A. Jaschke, Bioconjugate 5 C. Kao, S. Rudisser and M. Zheng, Methods, 2001, 23, 201–205.
¨
Chem., 1999, 10, 371–378; J. Wolf, V. Dombos, B. Appel and S. Mu¨ller, 6 M. L. Winz, A. Samanta, D. Benzinger and A. Jaschke, Nucleic Acids
Org. Biomol. Chem., 2008, 6, 899–907.
Res., 2012, 40, e78.
1316 | Chem. Commun., 2014, 50, 1313--1316
This journal is ©The Royal Society of Chemistry 2014