Highly potent and stable capped siRNAs with picomolar activity for RNA interference

Article abstract of DOI:10.1002/anie.201301122

Put a cap on it: Hairpin-shaped RNAs and dumbbell-shaped RNAs were prepared using a thiol-maleimino Michael addition and exhibited good serum and thermal stability. These capped structures were shown to be cleaved by Dicer and RNA interference (RNAi) experiments showed that RhpRNA (see picture, top right) was highly efficient at RNAi with an IC₅₀ value of 6 pM. Copyright

Full text of DOI:10.1002/anie.201301122

Angewandte
Chemie
DOI: 10.1002/anie.201301122
siRNA Modification
Highly Potent and Stable Capped siRNAs with Picomolar Activity for
RNA Interference**
Lv Wei, Liqiang Cao, and Zhen Xi*
Small interfering RNAs (siRNAs) can be used to silence gene
expression in a sequence-specific manner through a process
known as RNA interference (RNAi).[1] Many efforts have
been devoted to the application of siRNAs for biological
techniques and also as therapeutic agents.[2] Currently,
synthetic siRNAs as potential therapeutic molecultes could
be applied for clinical purposes. However, serum instability,
off-target effects, immunostimulatory activity, and poor
delivery into cells are the main barriers to this application.[3]
The potential of siRNAs to become a new therapeutic
approach has stimulated much interest in chemical modifica-
tions to optimize the biological properties of these siRNAs.[4]
Improvements in the serum stability and potency of siRNAs
can decrease the dosage required, which would be beneficial
for mitigating off-target effects and could also modulate
immunostimulatory activity and improve drug delivery.
Double-stranded RNAs with different strand lengths or
shapes, compared to the canonical 21 nucleotide (nt) siRNA
duplexes, have been heavily investigated, including dumbbell-
shaped RNAs.[5] Herein, chemical mimics of hairpin-shaped
and dumbbell-shaped RNAs (Figure 1) were designed and
constructed using a biocompatible thiol–maleimido Michael
addition[6] based on a dimeric cross-linker 1,2-bis(maleimido)-
ethane (BME; Figure 2a). One or both ends of the double-
stranded RNA could easily be capped using BME. The native
21 nt siRNA duplex used in our study was siFL867–885
Figure 2. a) Chemical formulas of the Thiol-Modifier C6 S-S and BME.
DMTr=dimethoxytrityl. b) Sequences 1–7 of native si867, cross-linked
RNAs (LhpRNA, RhpRNA, dbRNA), and control linear dsRNAs
(LdsRNA, RdsRNA and dbdsRNA).
(si867), which can effectively suppress the expression of
a firefly luciferase reporter gene.[7] Modified thiol-containing
Figure 1. Scheme of the formation of cross-linked RNA through
a thiol–maleimido Michael addition.
RNAs were synthesized using a commercial thiol-modifier C6
S-S (Figure 2a) following a standard phosphoramidite elon-
gation cycle for coupling of commercial nucleoside phosphor-
amidites.[8] Based on the sequence of si867, three cross-linked
RNAs (LhpRNA, RhpRNA, and dbRNA) were designed to
have four base-pairs in addition to the Dicer cleavage site
next to the thiol-modifier C6 S-S (Figure 2b, 1–4). Three
corresponding linear dsRNAs (LdsRNA, RdsRNA, and
dbdsRNA) were employed as a control (Figure 2b, 5–7).
To construct single-end cross-linked RNAs, both left
hairpin-shaped RNAs (LhpRNA) and right hairpin-shaped
RNAs (RhpRNA) were closed at their designated ends of the
modified dsRNAs using thiol–maleimido Michael addition
(Supporting Information, Figure S1). A newly formed band
that migrated more slowly than the single-stranded RNA was
found by denaturing polyacrylamide gel electrophoresis
(PAGE) analysis (Figure S3a). ESI-TOF MS was conducted
[*] L. Wei,[+] L. Cao,[+] Prof. Dr. Z. Xi
State Key Laboratory of Elemento-Organic Chemistry
Department of Chemical Biology
Nankai University, Tianjin 300071 (China)
E-mail: zhenxi@nankai.edu.cn
[+] These authors contributed equally to this work.
[**] This work was financially supported by the Ministry of Science and
Technology of China (2010CB126102, 2011BAE06B05,
2008DFA30770) and the National Natural Science Foundation of
China (20932005, 20872067). We would like to thank one of the
referees for pointing out the importance of the persistence of
siRNAs in the cytosol on the effectiveness of the RNAi.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301122.
Angew. Chem. Int. Ed. 2013, 52, 6501 –6503
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6501

.
Angewandte
Communications
to characterize the molecular weight of the artificially
prepared hairpin-shaped RNAs. The mass spectra for both
isolated LhpRNA and RhpRNA showed a single major peak
at 15944 Da, which matched with the calculated molecular
weight of 15945 Da (see Supporting Information). This
observation confirmed that capped RNAs could be obtained
from two modified RNA strands by Michael addition with
a good yield. Encouraged by these results, the thiol–male-
imido Michael addition was also used to construct dumbbell-
shaped RNA (dbRNA; Figure S2). To avoid intermolecular
polymerization, the optimal single-stranded RNA concentra-
tion for annealing was discovered to be below 0.5 mm, which
showed thiol–maleimido Michael addition only after dsRNA
annealing. A newly formed band appeared in the denaturing
PAGE analysis (Figure S3b), which was character-
ized as dumbbell-shaped RNA by ESI-TOF MS, with
a single major peak at 18414 Da (calculated value
was 18407 Da).
thermal stability of the duplexes. The melting temperatures
(Tm) of si867, LhpRNA, RhpRNA, and dbRNA were 65, 79,
79, and > 908C, respectively, showing that this modification is
effective at increasing the thermal stability of the siRNA
duplexes (Table S1). The circular dichroism (CD) spectra of
si867, LhpRNA, RhpRNA, and dbRNA showed similar
patterns for the cross-linked RNAs, which corresponded to
a typical A-form structure of duplex RNA (Figure S6).[10]
Next, the thiol–maleimido capped RNAs were treated
with Dicer, an enzyme that activates the RNA interference
pathways.[11] The Dicer enzyme BLOCK-iT (0.05 unitsmLÀ1)
was used to digest cross-linked RNAs and their corresponding
control duplexes. As shown in Figure 4, approximately 23–
25 nt RNAs were produced from LhpRNA, RhpRNA, and
To determine their biological stability, the cross-
linked RNAs (LhpRNA, RhpRNA, dbRNA) and
their corresponding control RNAs (LdsRNA,
RdsRNA, dbdsRNA) were incubated in 50%
normal human serum at 378C (Figure 3). The cross-
linked RNAs showed higher stability compared with
that of their linear counterpart or native siRNA
control. After 48 hours incubation, more than 35% of
the cross-linked RNAs remained full-length, while no
linear dsRNAs were detected after 24 hours incuba-
tion. Similar results were also seen when these RNAs
were incubated in 50% fetal bovine serum (Fig-
Figure 4. Analysis of the Dicer cleavage reaction of cross-linked RNAs and their
corresponding control dsRNAs. Annealed RNAs (2.5 mm) were incubated with
BLOCK-iT DICER (0.05 unitsmLÀ1). The reaction mixtures were analyzed by 15%
native PAGE and visualized with SYBR Gold staining. M indicates a double-
stranded RNA marker.
ure S4). The improved serum stability of the cross-linked
RNAs could be due to their continuous structures.[5c] Because
RNase A is one of the main nucleases responsible for siRNA
degradation in serum,[4e,9] a dosage-dependent RNase A
digestion of the cross-linked RNAs were also carried out.
After one hour of incubation with 1 mgLÀ1 RNase A at 378C,
80% of dbRNA was still full length, in contrast to only 60%
of si867 (Figure S5), which indicated the increased RNase A
resistance of the modified RNAs than that of the native
siRNAs. These results showed that a capping method for
siRNAs using thiol–maleimido cross-linking can significantly
improve the stability of siRNAs in biological environments.
The capping of siRNA through thiol-maleimido cross-
linking described herein also gave a dramatic increase in the
dbRNA by Dicer, which is in accordance with length of the
Dicer products of LdsRNA, RdsRNA, and dbdsRNA.[11a,12]
Under the same reaction conditions, more than 90% of the
cross-linked RNAs and duplex RNAs were digested into
smaller fragments after 20 hours of incubation. We concluded
that the cross-linked RNAs could be cleaved by the Dicer
enzyme to form dsRNAs in vitro.
The RNA interference activities of the cross-linked RNAs
and their control sequences were measured using a dual-
luciferase reporter system.[7] HEK293A cells were co-trans-
fected with dual-reporter plasmids, pRL-TK and pGL3-Rosa,
and also the different RNAs with various concentrations.
Expression levels of the two luciferase genes were assayed
48 hours after transfection. As shown in Figure 5, the IC50
values for modified LhpRNA, RhpRNA, and dbRNA were
131.75 Æ 40.32, 6.19 Æ 0.18, and 68.72 Æ 19.07 pm, respectively;
as controls, the IC50 values for native si867, LdsRNA,
RdsRNA, and dbdsRNA were 32.55 Æ 9.01, 29.45 Æ 13.33,
15.06 Æ 1.79, 98.23 Æ 31.59 pm, respectively. The highest sup-
pression was induced by RhpRNA, which was about 2.5-fold
more potent than that induced by RdsRNA. Compared to
reported IC50 values for a 21 nt duplex siRNA,[13] RhpRNA
with an IC50 value of about 6 pm showed the highest potency
for RNAi thus far.
There was a dramatic difference in the RNAi efficiency
between LhpRNA and RhpRNA at low concentrations (for
example, below 0.1 nm), which may largely be due to the
effect of the different structures of the two modifications on
the formation of the RISC (RNA-induced silencing complex).
Figure 3. Serum stability of modified RNAs and their corresponding
control dsRNAs in 50% normal human serum. The reaction mixtures
were analyzed using 15% native PAGE and visualized with SYBR Gold
staining. M indicates a double-stranded RNA marker.
6502 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 6501 –6503

Angewandte
Chemie
c) S. M. Elbashir, J. Harborth, W. Lendeckel, A. Yalcin, K.
Weber, T. Tuschl, Nature 2001, 411, 494 – 498; d) S. M. Elbashir,
W. Lendeckel, T. Tuschl, Genes Dev. 2001, 15, 188 – 200; e) G. J.
Hannon, Nature 2002, 418, 244 – 251.
[2] a) T. Tuschl, Nat. Biotechnol. 2002, 20, 446 – 448; b) M. Man-
oharan, Curr. Opin. Chem. Biol. 2004, 8, 570 – 579; c) D. H. Kim,
J. J. Rossi, Nat. Rev. Genet. 2007, 8, 173 – 184.
[3] a) J. K. Watts, G. F. Deleavey, M. J. Damha, Drug Discovery
Today 2008, 13, 842 – 855; b) A. de Fougerolles, H. Vornlocher, J.
Maraganore, J. Lieberman, Nat. Rev. Drug Discovery 2007, 6,
443 – 453.
[4] a) J. M. Layzer, A. P. McCaffrey, A. K. Tanner, Z. Huang, M. A.
Kay, B. A. Sullenger, RNA 2004, 10, 766 – 771; b) D. V. Morris-
sey, J. A. Lockridge, L. Shaw, L. Blanchard, Nat. Biotechnol.
2005, 23, 1002 – 1007; c) C. R. Allerson, N. Sioufi, R. Jarres, T. P.
Prakash, N. Naik, A. Berdeja, L. Wanders, R. H. Griffey, E. E.
Swayze, B. Bhat, J. Med. Chem. 2005, 48, 901 – 904; d) S. Choung,
Y. J. Kim, S. Kim, H. O. Park, Y. C. Choi, Biochem. Biophys. Res.
Commun. 2006, 342, 919 – 927; e) J. M. Hong, Y. Y. Huang, J. Li,
F. Yi, J. Zheng, H. A. Huang, N. Wei, Y. Q. Shan, M. R. An, H. Y.
Zhang, J. G. Ji, P. Z. Zhang, Z. Xi, Q. A. Du, Z. C. Liang, FASEB
J. 2010, 24, 4844 – 4855.
Figure 5. RNA interference with si867, LhpRNA, RhpRNA, dbRNA, and
three corresponding control dsRNAs. The luciferase activities of the
samples were measured according to the instructions of the dual-
luciferase reporter assay kit (Promega, USA). Dosage curves of the
RNA silencing effect were fitted by a Boltzmann nonlinear least-
squares regression to give the IC50 values (Figure S7). The data shown
are the meanÆstandard deviation of three independent experiments.
[5] a) D. H. Kim, M. A. Behlke, S. D. Rose, M. S. Chang, S. Choi,
J. J. Rossi, Nat. Biotechnol. 2005, 23, 222 – 226; b) X. G. Sun,
H. A. Rogoff, C. J. Li, Nat. Biotechnol. 2008, 26, 1379 – 1382;
c) N. Abe, H. Abe, Y. Ito, J. Am. Chem. Soc. 2007, 129, 15108 –
15109.
[6] a) K. Kuhn, D. J. Owen, B. Bader, A. Wittinghofer, J. Kuhlmann,
H. Waldmann, J. Am. Chem. Soc. 2001, 123, 1023 – 1035; b) M.
Vçlkert, K. Uwai, A. Tebbe, B. Popkirova, M. Wagner, J.
Kuhlmann, H. Waldmann, J. Am. Chem. Soc. 2003, 125, 12749 –
12758; c) L. Yi, H. Li, L. Sun, L. Liu, C. Zhang, Z. Xi, Angew.
Chem. 2009, 121, 4094 – 4097; Angew. Chem. Int. Ed. 2009, 48,
4034 – 4037; d) S. J. Ye, K. T. Nguyen, A. P. Boughton, C. M.
Mello, Z. Chen, Langmuir 2010, 26, 6471 – 6477; e) C. A. Hong,
S. H. Lee, J. S. Kim, J. W. Park, K. H. Bae, H. Mok, T. G. Park, H.
Lee, J. Am. Chem. Soc. 2011, 133, 13914 – 13917.
[7] Q. Z. Zhang, C. H. Zhang, Z. Xi, Cell Res. 2008, 18, 1077 – 1079.
[8] N. Usman, K. K. Ogilvie, M. Y. Jiang, R. J. Cedergren, J. Am.
Chem. Soc. 1987, 109, 7845 – 7854.
[9] a) J. J. Turner, S. W. Jones, S. A. Moschos, M. A. Lindsay, M. J.
Gait, Mol. Biosyst. 2007, 3, 43 – 50; b) J. Haupenthal, C. Baehr, S.
Kiermayer, S. Zeuzem, A. Piiper, Biochem. Pharmacol. 2006, 71,
702 – 710.
3’-modified siRNAs are typically better tolerated by the
RNAi enzymes,[14a] while the 5’-phosphate of siRNA, which is
anchored in the MID pocket of Ago2, is more important in
RISC.[14b] In our study, RhpRNA (with a 3’-overhang on the
antisense strand) was much more potent than almost all
reported siRNAs, with its IC50 value of 6 pm, while LhpRNA
and dbRNA exhibited worse RNAi potency compared with
that of the native siRNA. Note that RhpRNA has been
cleaved by Dicer slower than LhpRNA, so in addition to the
factors discussed above, the difference in potency may also be
explained by a longer persistence of RhpRNA in the cytosol
and thus a longer RNAi effect. These results further indicate
that chemical modification of the 3’-end of siRNAs can be
used to improve RNAi efficiency and increase RNA stability.
In conclusion, we have demonstrated herein that the
capping of siRNAs using a thiol–maleimido cross-linking
method resulted in an improvement in both the serum
stability and the thermal stability of siRNAs. The capped
RNAs can still be recognized and digested by Dicer. RNA
interference experiments showed that RhpRNA was a highly
efficient modification with an IC50 value of 6 pm, more potent
than any reported siRNAs. Our results should provide an
efficient method to improve siRNA serum stability and
potency for RNAi, and further improvements in chemical
modifications of siRNAs will provide better siRNA-based
biological methods and therapeutic agents.
[10] V. I. Ivanov, L. Minchenk, A. Schyolki, A. Poletaye, Biopolymers
1973, 12, 89 – 110.
[11] a) I. J. MacRae, K. H. Zhou, F. Li, A. Repic, A. N. Brooks, W. Z.
Cande, P. D. Adams, J. A. Doudna, Science 2006, 311, 195 – 198;
b) D. Siolas, C. Lerner, J. Burchard, W. Ge, P. S. Linsley, P. J.
Paddison, G. J. Hannon, M. A. Cleary, Nat. Biotechnol. 2005, 23,
227 – 231; c) A. Vermeulen, L. Behlen, A. Reynolds, A. Wolfson,
W. S. Marshall, J. Karpilow, A. Khvorova, RNA 2005, 11, 674 –
682.
[12] J. E. Park, I. Heo, Y. Tian, D. K. Simanshu, H. Chang, D. Jee,
D. J. Patel, V. N. Kim, Nature 2011, 475, 201 – 207.
Received: February 7, 2013
Published online: May 7, 2013
[13] a) R. K. K. Far, G. Sczakiel, Nucleic Acids Res. 2003, 31, 4417 –
4424; b) M. Overhoff, M. Alken, R. K. K. Far, M. Lemaitre, B.
Lebleu, G. Sczakiel, I. Robbins, J. Mol. Biol. 2005, 348, 871 – 881;
c) T. A. Vickers, W. F. Lima, J. G. Nichols, S. T. Crooke, Nucleic
Acids Res. 2007, 35, 6598 – 6610.
[14] a) J. Willibald, J. Harder, K. Sparrer, K. K. Conzelmann, T.
Carell, J. Am. Chem. Soc. 2012, 134, 12330 – 12333; b) Y. Wang,
S. Juranek, H. Li, G. Sheng, G. S. Wardle, T. Tuschl, D. J. Patel,
Nature 2009, 461, 754 – 761.
Keywords: dumbbell RNA · hairpin RNA · Michael addition ·
.
RNA interference · siRNA
[1] a) A. Fire, S. Q. Xu, M. K. Montgomery, S. A. Kostas, S. E.
Driver, C. C. Mello, Nature 1998, 391, 806 – 811; b) P. D. Zamore,
T. Tuschl, P. A. Sharp, D. P. Bartel, Cell 2000, 101, 25 – 33;
Angew. Chem. Int. Ed. 2013, 52, 6501 –6503
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6503