Controlling RNA digestion by RNase H with a light-activated DNA hairpin

Article abstract of DOI:10.1002/anie.200600954

(Figure Presented) Making sense: A DNA hairpin was synthesized by attaching a 20-mer antisense oligodeoxynucleotide (asODN) to a sense ODN through a bifunctional photocleavable linker. This conjugate became much less stable upon UV photoactivation (ΔT

Full text of DOI:10.1002/anie.200600954

Angewandte
Chemie
DNA Structures
DOI: 10.1002/anie.200600954
Controlling RNA Digestion by RNase H with a
Light-Activated DNA Hairpin**
XinJing Tang and Ivan J. Dmochowski*
Many gene silencing technologies have focused on targeting
mRNA by using ribozymes, DNAzymes, and complementary
RNA or DNA oligonucleotides. It is attractive that mRNA is
accessible during transcription, processing (e.g., 5’-capping,
intro-exon splicing, polyadenylation, nuclear export), and
ribosomal binding and translation. Hybridization of an
“antisense” oligodeoxynucleotide (asODN) to
a target
mRNA inhibits translation by sterically blocking the ribo-
some and/or recruiting endogenous ribonucleases. In mam-
malian cells, mRNA/asODN duplex formation activates
RNase H-mediated hydrolysis of mRNA. Antisense ODNs
have proven effective in gene silencing[1] in many experimen-
tal systems and are being evaluated as treatments for cancer
and other diseases in human clinical trials. It was recently
reported that the stability of DNA hairpins relative to their
corresponding RNA/DNA hybrids influenced the extent of
RNA degradation by RNase H.[2] Based on this principle, we
designed a light-activated DNA hairpin to control this
enzymatic reaction (Figure 1 A).
Photoinduced generation of nucleic acid strand breaks has
been reported by several groups.[3] However, most studies
have focused on mimicking genomic lesions, whereas our
strategy is based on the photorelease of DNA for controlling
biological processes. Light-activated methods for controlling
the binding of asODNs to target mRNA would permit the
reduction of protein expression in a cell or tissue-specific
fashion and thereby limit systemic toxicity.[4] Recent efforts to
increase the specific activity of asODNs have focused on
improving delivery to cells and nuclease resistance.[5] Stability
within cells can be enhanced considerably by modifying the
backbone of the oligonucleotide, as exemplified by phosphor-
othioate,[6] peptide,[7] morpholine,[8] and “locked”[9] nucleic
acids. However, there are few reports about regulating the
activity of asODNs.[10,11]
Figure 1. A) Strategy for regulating RNA digestion by RNase H by using a light-
activated DNA hairpin. B) Sequences of the asODN-sODN conjugate and RNA
targets; structure of the heterobifunctional photocleavable linker, PC. Underlined
bases in RNA-40 are identical to RNA-20.
benzene to control RNA digestion by RNase H.[10] By
attaching five azobenzenes to a 20-mer sense DNA strand,
the binding of a complementary asODN decreased upon UV
irradiation: trans, Tm = 60.88C; cis, Tm = 42.68C. The less
conformationally extended cis isomer sterically blocked DNA
duplex formation. Based on this difference in melting
temperature, DTm = 18 K, it was possible to regulate the
hybridization of the asODN to a complementary RNA target.
However, the photoefficiency was greatly reduced by the
requirement for five azobenzenes. For DNA modified with n
moieties, each with a photoquantum efficiency of 50%, the
irradiation time required to generate the fully active molecule
should scale as 2n. In fact, prolonged UV irradiation leads to
side reactions that lower the yield of the active molecule and
cause cytotoxicity. For cell experiments, prolonged irradiation
reduces temporal as well as spatial resolution due to diffusion.
Moreover, the use of a reversible, photoisomerizable blocking
group is problematic as mRNA is an abundant and persistent
target.
In a pioneering example, Matsunaga et al. recently
applied the UV-mediated trans!cis isomerization of azo-
[*] X. Tang, Prof. I. J. Dmochowski
Department of Chemistry
University of Pennsylvania
231 South 34th Street, Philadelphia, PA 19104-6323 (USA)
Fax: (+1)215-898-2037
E-mail: ivandmo@sas.upenn.edu
[**] Thiswork wassupported by a Camille and Henry DreyfusNew
Faculty Award, the University of Pennsylvania Genomics Institute
(PGI) and Institute for Medicine and Engineering (IME). We thank
Jim Eberwine, Feng Gai, and Eric Meggers for access to instru-
mentation. Alan Gewirtz is acknowledged for valuable discussions.
In related studies, DNA hybridization was recently
photomodulated between 14% (blocked) and 80% (acti-
vated) by using an average of 14–16 nitrophenyl adducts per
20-mer asODN.[11] Further precedent exists for the control of
Supporting information for thisarticle isavailable on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 3523 –3526
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3523

Communications
DNA and RNA activity with light.[12] However, the use of
RNase H assays were also performed with RNA-20 and
RNA-40. After photolysis of the DNA hairpin, RNA-20 and
RNA-40 digestion increased by a factor of 1.3 ꢀ 0.2 and 4.4 ꢀ
0.4, respectively, versus a 9.0 ꢀ 1.0-fold increase for RNA-15
(Figure 3). Three units of fresh RNase H gave nearly com-
multiple photoactive groups in most cases has limited the
utility of these methods for biological investigations. Our
strategy for photoregulating DNA involves the incorporation
of a single, maximally effective blocking group.
We synthesized a light-activated DNA hairpin by cova-
lently attaching a 20-mer asODN to an ODN with 12
complementary bases through a heterobifunctional photo-
cleavable linker (PC, Figure 1). This 1-(5-(N-maleimido-
methyl)-2-nitro-phenyl)ethanol N-hydroxysuccinimide ester
was designed to react with thiol and amine functionalities on
opposite ends of the linker (Figure 1B). PC was synthesized
in four steps in a 14% yield. Its dual reactivity made it even
more versatile for bioconjugation reactions than the 1-(5-
(aminomethyl)-2-nitrophenyl)ethanol linker used previ-
ously.[13,14] Covalent attachments of the thiolated asODN to
the maleimido group and the amino-ODN to the N-hydrox-
ysuccinimide ester occurred in good yields. Purification of the
conjugated DNA hairpin was achieved by ion exchange
chromatography at 408C. This step was particularly important
because any unreacted asODN would readily form a duplex
with the target RNA and thus promote RNase H activity.
The DNA hairpin exhibited a high melting temperature,
Tm = 808C. Upon UV illumination, the carbamate linkage was
broken and the asODN/ODN duplex became much less
stable, Tm = 518C. Photocleavage was confirmed by HPLC
and MALDI-TOF mass spectrometry. This large difference in
thermal stability, DTm = 29 K, allowed the photomodulation
of RNA/asODN duplex formation, thereby regulating RNA
hydrolysis by RNase H (Figure 1A).
Figure 3. RNase H assays showing degradation of RNA-15, RNA-20,
and RNA-40 target sequences in 60 min at 378C with 1 unit of
RNase H. Error bars signify the variation from two separate trials.
The control lanes show hydrolysis of RNA hybridized to asODN.
plete digestion of RNA-15 (98%), RNA-20 (88%), and
RNA-40 (84%) in 1 h in the presence of the control asODN
or photolyzed conjugate. This compared favorably with the
consumption of only 49% of RNA-15, 70% of RNA-20, and
27% of RNA-40 by using the DNA hairpin (see the
Supporting Information). RNA-40 is more representative of
cellular mRNA, which has secondary structure that can limit
nonspecific DNA hybridization and RNase H digestion.
Within a cell, where each mRNA molecule can template the
synthesis of several thousand proteins per hour,[15] the ability
to photomodulate target mRNA between 73% and 16% of
basal levels, as was shown for RNA-40, would produce a large
effect.
RNase H activity was studied with RNA oligomers of
varied length, RNA-15 (15-mer), RNA-20 (20-mer), and
RNA-40 (40-mer). Digestion of each RNA target was
compared between solutions containing the control single-
stranded asODN (CCAACGTTTCGGACCGTATT) or the
DNA hairpin. Figure 2 shows hydrolysis of 45% of the total
The observed photomodulation efficiency was controlled
predominantly by the relative thermodynamic stabilities of
the conjugate, pre-and postphotolysis, and the RNA/asODN
duplexes. The melting temperatures for duplexes of asODN
with RNA-15 and RNA-20 were 63 and 708C, respectively.
For these two RNA targets, the increases in RNase H activity
upon photoactivation arose from the greater stability of the
DNA hairpin (DTm = 17, 108C) compared with the RNA/
asODN duplexes. Free-energy calculations gave DG values
for asDNA/sDNA, asDNA/RNA-15, and asDNA/RNA-20 at
378C in RNase H buffer of ꢁ12.4, ꢁ14.5, and ꢁ16.3 kcal
molꢁ1, respectively.[16] For the conjugate, DG = ꢁ17.0 kcal
molꢁ1 was calculated by determining the changes in enthalpy
and entropy upon duplex formation, DHd and DSd, respec-
tively, from the concentration dependence of the melting
equilibria by using the vanꢀt Hoff expression (see the
Supporting Information). Binding RNA-15 or RNA-20 to
the DNA hairpin was disfavored thermodynamically, DDG =
2.5 and 0.7 kcalmolꢁ1, respectively. The decrease in DDG with
increasing RNA oligomer length helps to explain the higher
background levels of RNA-20 digestion, before photolysis.
Figure 2. Denaturing PAGE (20%) analysis of RNA-15 digestion by
RNase H at 378C. Lanes1–2, asODN; Lanes3–4, conjugate
(conj.ꢁUV); Lanes5–6, UV-activated conjugate (conj. +UV).
RNA-15 target after photoactivation as compared with only
4.6% for the conjugate in a 1 h experiment with 1 unit of
enzyme. Under the same conditions, 66% RNA degradation
for the control single-stranded asODN was observed. The UV
light used in these experiments had no effect on the enzyme
or RNA stability. Control experiments confirmed that
RNase H hydrolyzed RNA only when hybridized with
asODN.
3524
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3523 –3526

Angewandte
Chemie
RNase H has been shown to promote the formation and
cleavage of RNA/DNA duplexes, even under conditions
where RNA/DNA hybridization is thermodynamically disfa-
vored.[17] After photolysis, the activated asODN/sODN
hybrid readily converted to asODN/RNA-15 and asODN/
RNA-20 (DDG = ꢁ2.1 and ꢁ3.9 kcalmolꢁ1, respectively).
hydrolysis by RNase H. This represents the most efficient
method for photomodulating this enzyme activity. Oligonu-
cleotide conjugates can be designed to photoregulate many
other processes for biotechnological and cellular applications.
In addition to thermodynamic considerations, the higher Experimental Section
The standard procedure for RNase H assay: The DNA conjugate was
background level of RNase H activity for RNA-20 was owing
to the nature of the blocking group. The 12 complementary
bases on the sense ODN were designed to block the binding
and degradation of RNA targets with the same 12-base
recognition motif. However, the eight mismatched bases of
the asODN in the DNA hairpin were able to base pair with
RNA-20 and recruit RNase H. Although RNA-20 should be
one of the most difficult sequences for this DNA hairpin to
photodiscriminate against, the photomodulation efficiency
(1.3 ꢀ 0.2) was slightly greater than background levels.
annealed in the 1 ” ribonuclease H reaction buffer (Tris-HCl (20 mm),
KCl (20 mm), MgCl2 (10 mm), ethylenediaminetetraacetic acid
(EDTA; 0.1 mm), dithiothreitol (DTT; 0.1 mm) ; pH 8.0) by heating
to 958C, then slowly cooling to 708C. The temperature was held at
708C for 10 min to melt nonspecific DNA structures, before further
slow cooling to 378C. [g-32P]-labeled RNA oligonucleotide was added
and incubated at 378C for 20 min to allow RNA/DNA duplex
formation. RNase H was added to the mixture and incubated at 378C.
The total reaction volume was 10 mL and the final concentrations of
conjugate or control (asODN), and RNA (RNA-15, RNA-20, or
RNA-40) were 0.01 mm and 2 mm, respectively. Time points were
taken at 10 and 60 min by sampling 4 mL of the reaction mixture,
adding 6 mL loading buffer (EDTA (50 mm), formamide (90%)), and
heating to 958C for 3 min to terminate the reaction. Finally, 5 mL of
the resulting solution was subject to electrophoresis on a polyacryl-
amide gel containing 7m urea. RNA imaging was performed by using
a Storm phosphorimager and quantified with IMAGEQUANT
software (Amersham Biosciences). To measure RNA degradation
after photoactivation, the annealed DNA hairpin was illuminated (Xe
lamp with monochromator, 355 nm, 36 mWcmꢁ2, 10 min) and [g-32P]-
labeled RNA oligonucleotide was then added and incubated for
20 min. RNA degradation by RNase H was determined as described
above. See the Supporting Information for further experimental
details.
The DNA hairpin exhibited the highest photomodulation
efficiency towards the RNA-15 substrate. Contributing fac-
tors were the stability of the DNA hairpin relative to the
asODN/RNA-15 duplex (DDG = ꢁ2.5 kcalmolꢁ1), proper
complementarity of the blocking strand (slightly shorter
than the target), and the structure of the RNA itself, which
limited nonspecific DNA hybridization. The RNA melting
temperatures of RNA-15, RNA-20, and RNA-40 were 518C,
548C, and 638C, respectively. The expected correlation
between increasing RNA strand length and structure was
interesting in view of the higher photomodulation efficiency
for RNA-40 than RNA-20. These targets had the same
20 bases complementary to asODN and yet RNA-40 exhib-
ited much lower background levels of hydrolysis. Evidently,
for targets such as RNA-40, it is important that the asODN
has a high degree of complementarity to compete with the
stable RNA stem-loop structure. To test this hypothesis, a
RNA 40-mer was employed with only 15 complementary
bases (CUUGUACAGAAAUACGGUCCGAAACCAAC-
CUCUGUUAUUG, underlined bases are identical to
RNA-15). No RNase H activity was observed towards this
substrate by using the photolyzed DNA hairpin, and only
13% digestion was seen with the control asODN. The results
with the four RNA substrates provide guidelines for designing
photoactive DNA hairpins against a specific RNA target.
Notably, in the earlier azobenzene example, a small DTm
of 18 K required a 10-fold excess of sense ODN relative to
asODN to limit RNA/asODN duplex formation.[10] This
strategy is not practical for most biological applications as
the sense ODN can diffuse away or become degraded inside
the cell. Furthermore, introducing an organic chromophore in
the middle of an ODN typically lowers the DNA duplex
melting temperature only by a few degrees Celsius per
blocking group.[10,11,14] We demonstrated herein that a much
larger DTm can be achieved by conjugating the sense ODN to
the asODN through a single photocleavable linker. The PC
linker plays a role in stabilizing the conjugate.
Received: March 10, 2006
Published online: April 24, 2006
Keywords: DNA structures · oligonucleotides · photoactivation ·
.
RNase H
[1] a) C. Cazenave, C. Stein, N. Loreau, N. Thuong, L. Neckers, C.
Subasinghe, C. Helene, J. Cohen, J. Toulme, Nucleic Acids Res.
1989, 17, 4255; b) M. Kozak, Proc. Natl. Acad. Sci. USA 1986, 83,
2850; c) J. B. Opalinska, B. Machalinski, J. Ratajczak, M. Z.
Ratajczak, A. M. Gewirtz, Clin. Cancer Res. 2005, 11, 4948.
[2] J. Li, B. Bourdelat-Parks, J. H. Boatright, R. M. Wartell,
Biochemistry 2003, 42, 10945.
[3] a) A. Dussy, C. Meyer, E. Quennet, T. A. Bickle, B. Giese, A.
Marx, ChemBioChem 2002, 3, 54; b) P. Ordoukhanian, J.-S.
Taylor, Bioconjugate Chem. 2000, 11, 94; c) K. Zhang, J.-S.
Taylor, J. Am. Chem. Soc. 1999, 121, 11579; d) P. Ordoukhanian,
J.-S. Taylor, J. Am. Chem. Soc. 1995, 117, 9570; e) M. C. Pirrung,
X. Zhao, S. V. Harris, J. Org. Chem. 2001, 66, 2067.
[4] T. L. H. Jason, J. Koropatnick, R. W. Berg, Toxicol. Appl.
Pharmacol. 2004, 201, 66.
[5] M. Rubenstein, P. Tsui, P. Guinan, Drugs Future 2004, 29, 893.
[6] W. Stec, G. Zon, W. Egan, B. Stec, J. Am. Chem. Soc. 1984, 106,
6077.
[7] J. C. Hanvey, N. J. Peffer, J. E. Bisi, S. A. Thomson, R. Cadilla,
J. A. Josey, D. J. Ricca, C. F. Hassman, M. A. Bonham, K. G. Au,
S. G. Carter, D. A. Bruckenstein, A. L. Boyd, S. A. Noble, L. E.
Babiss, Science 1992, 258, 1481.
In conclusion, a DNA hairpin was designed that efficiently
regulates DNA/DNA and DNA/RNA duplex formation by
using a single photoactive group. The stability of the DNA
hairpin, proper complementarity of the blocking group, and
RNA structure were important factors affecting RNA
[8] J. Summerton, Biochim. Biophys. Acta 1999, 1489, 141.
[9] L. Wang, C. J. Yang, C. D. Medley, S. A. Benner, W. Tan, J. Am.
Chem. Soc. 2005, 127, 15664.
Angew. Chem. Int. Ed. 2006, 45, 3523 –3526
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3525

Communications
[10] D. Matsunaga, H. Asanuma, M. Komiyama, J. Am. Chem. Soc.
2004, 126, 11452.
[11] B. Ghosn, F. R. Haselton, K. R. Gee, W. T. Monroe, Photochem.
Photobiol. 2005, 81, 953.
[12] a) L. Kröck, A. Heckel, Angew. Chem. 2005, 117, 475; Angew.
Chem. Int. Ed. 2005, 44, 471; b) S. Shah, S. Rangarajan, S. H.
Friedman, Angew. Chem. 2005, 117, 1352; Angew. Chem. Int. Ed.
2005, 44, 1328; c) H. Ando, T. Furuta, R. Y. Tsien, H. Okamoto,
Nat. Genet. 2001, 28, 317; d) A. Yamazawa, X. Liang, H.
Asanuma, M. Komiyama, Angew. Chem.. 2000, 112, 2446;
Angew. Chem. Int. Ed. 2000, 39, 2356.
[13] a) X. Bai, Z. Li, S. Jockusch, N. J. Turro, J. Ju, Proc. Natl. Acad.
Sci. USA 2003, 100, 409; b) X. Tang, J. L. Richards, A. E. Peritz,
I. J. Dmochowski, Bioorg. Med. Chem. Lett. 2005, 15, 5303.
[14] X. Tang, I. J. Dmochowski, Org. Lett. 2005, 7, 279.
[15] Y. Arava, Y. Wang, J. D. Storey, C. L. Liu, P. O. Brown, Proc.
Natl. Acad. Sci. USA 2003, 100, 3889.
[16] DG was determined by using an Excel spreadsheet (http://
www.biochem.umd.edu/biochem/kahn/teach_res/thermsim.xls)
for simulating the melting curves of non-self-complementary
DNA.
[17] E. Zamaratski, P. I. Pradeepkumar, J. Chattopadhyaya, J.
Biochem. Biophys. Methods 2001, 48, 189.
3526
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3523 –3526