Deoxystreptamine dimers bind to RNA hairpin loops

Article abstract of DOI:10.1021/ja048936l

RNA is known to have multiple roles in critical cellular functions. Thus, there is great potential for RNA-binding small molecules as both therapeutic agents and cellular probes. Unfortunately, the multiple secondary structures that RNA can adopt have cau

Full text of DOI:10.1021/ja048936l

Published on Web 07/13/2004
Deoxystreptamine Dimers Bind to RNA Hairpin Loops
Xianjun Liu, Jason R. Thomas, and Paul J. Hergenrother*
Department of Chemistry, Roger Adams Laboratory, UniVersity of Illinois, Urbana, Illinois 61801
Received February 25, 2004; E-mail: hergenro@uiuc.edu
Scheme 1
The functions of RNA in the cell are vast and varied. The original
discovery of mRNA accorded RNA the important if somewhat
pedestrian role as the intermediary between DNA and protein.
However, over the ensuing decades it has become apparent that
RNA is more than a transient repository of genetic information
but instead has pivotal roles in cell division, gene expression, and
catalysis. Indeed, regulation of a variety of processes with small
untranslated RNAs has emerged as a general theme in biology,1
and it is now known that the protein synthesis machinery, the
ribosome, is a ribozyme.2
reduction with PMe3/NaOH furnished the desired deoxystreptamine
dimers 4b-14b (Scheme 1).
Accordingly, there has been considerable interest in the identi-
fication of small molecules that bind tightly to RNA.3 RNA-binding
natural products such as aminoglycosides and macrolides are widely
used as antibiotics, and RNA-binding small molecules have potential
against a variety of biological targets.4 However, in contrast to
DNA,5 no general paradigm currently exists by which sequence-
specific RNA-binding small molecules can be designed. Efforts
toward this goal are complicated by the variety of secondary
structures that RNA can adopt and by the fact that compounds
specifically recognizing these RNA secondary structures are lacking.
One successful approach has been the use of aminoglycosides to
bind RNA from a variety of sources. In general, aminoglycosides
recognize secondary structural elements within RNA in which the
usual base pairing has been disrupted such as bulges, internal loops,
and stem junctions.6 For instance, the aminoglycoside neomycin
will bind tightly to bulge regions of unrelated RNA sequences from
the 16S ribosome, HIV TAR, HIV RRE, and the Group I intron
with affinities in the low micromolar range.7 However, despite their
promiscuity for these types of RNA secondary structures, ami-
noglycosides do not typically bind to RNA hairpin loops,6e which
are a major RNA secondary structural motif.8 Indeed, there are no
general RNA hairpin loop-binding compounds. Herein we report
the results of a systematic study designed to identify general RNA
hairpin loop-binding small molecules.
Deoxystreptamine (1) has been reported to bind weakly to two
base units within a disrupted RNA helix.9 Solution studies carried
out with 1 have shown that it will bind to 5′-3′ two-base steps
(including GU, UG, and GG), albeit with a very low affinity (>1
mM).9 We reasoned that linkage of two such units by an appropriate
tether could lead to compounds with the requisite flexibility and
functionality to bind RNA hairpin loops. Thus, we have tethered
together two molecules of 1 with various linkers to create
deoxystreptamine dimers, and we show that these molecules have
the capacity to bind to a variety of RNA hairpin loops that are not
recognized by aminoglycosides.
To assess the ability of these deoxystreptamine dimers to bind
hairpin RNA, we utilized RNAs with tetra-, hexa-, hepta-, and
octaloops.11 All these RNAs had 5′G-U3′ steps on either side of
the loop and a common five base pair stem consisting of alternating
G-C and C-G bases (Table 1). To detect binding, all of the RNA
sequences were synthesized with a 3′-fluorescein label such that
binding could be monitored by a change in the fluorescence signal.12
The results from these binding experiments are displayed in Table
1, and a representative graph comparing a dimer of 1 and an
aminoglycoside is displayed in Figure 1. As expected, deoxy-
streptamine itself and all aminoglycosides tested had very high
dissociation constants (Kd > 1 mM).13 However, several of the
deoxystreptamine dimers had significant affinities for the RNA
hairpin loops. For dimers containing both the aromatic and aliphatic
linkers, binding was favorable for longer tether lengths. In addition,
compounds containing the aromatic linkers generally showed tighter
binding over their aliphatic counterparts.
Several controls established the importance of the dimeric nature
of deoxystreptamine for RNA binding. Monomer 1 showed no
detectable binding. In addition, while compound 11b bound the
RNAs, the tetraazide analogue 11a showed no binding. Finally,
neither the linkers alone, nor a linker with a single monomer
attached (compound 15) showed significant binding to the RNA
hairpin loops. Taken together, the data indicate that dimeric versions
of deoxystreptamine are necessary for binding. In addition, the
experiments with the tetraazide 11a suggests that the binding
observed is due to direct interactions between the amines of
deoxystreptamine and the RNAs.
To assess if these newly identified RNA-small molecule
interactions were general for loops of different sequences, nine
variants of RNA heptaloop C were created, and the binding of these
RNAs to 14b was determined. All of these loops essentially have
the same affinity for 14b (∼6 µM, see Supporting Information for
details); thus, the binding appears to be independent of the primary
loop sequence. In addition, a competition experiment was performed
in which binding of 14b to two of the hairpin loops was assessed
in the presence of a large excess of the stem. In both cases, no
change in Kd was observed, indicating that the deoxystreptamine
dimer is not binding to the stem.
To construct the desired dimers, the enantiopure diacetate diazide
of deoxystreptamine was created through known protocols (2 in
Scheme 1).10 Protecting group manipulation provided alcohol 3,
the appropriately protected precursor for dimerization. Treatment
of 3 with sodium hydride, followed by addition of a variety of
dibromides gave the protected dimers 4-14. Deprotection of the
acetonide with AcOH provided tetraazides 4a-14a, and azide
This 6 µM binding affinity of 14b with loop C rivals that of
standard aminoglycoside-RNA pairs.6c To establish the precise
location of this small molecule-nucleic acid interaction, an RNA
9
9196
J. AM. CHEM. SOC. 2004, 126, 9196-9197
10.1021/ja048936l CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Table 1. Binding of Deoxystreptamine Dimers to RNA
Stem-Loopsa
Figure 2. RNase I footprint of loop C with various concentrations of 14b.
potentially be paired with binders of other RNA secondary structural
elements to obtain even higher affinities.14
In summary, we have shown that simple dimers of deoxy-
streptamine show strong binding to RNA hairpin loops, a secondary
structural element that is not generally recognized by standard RNA-
binding molecules such as aminoglycosides. Further explorations
into the relationship between loop size and linker type, as well as
applications of these molecules to RNAs of biological interest, are
ongoing and will be reported in due course.
Acknowledgment. This work was funded through the Beckman
Young Investigator program.
Supporting Information Available: Full experimental protocols
and characterization data (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) (a) Kiss, T. Cell 2002, 109, 145. (b) Ambros, V. Cell 2001, 107, 823. (c)
Carrington, J. C.; Ambros, V. Science 2003, 301, 336.
(2) (a) Nissen, P.; Hansen, J.; Ban, N.; Moore, P. B.; Steitz, T. A. Science
2000, 289, 920. (b) Ban, N.; Nissen, P.; Hansen, J.; Moore, P. B.; Steitz,
T. A. Science 2000, 289, 905.
(3) Reviews on small molecule-RNA binding: (a) Sucheck, S. J.; Wong,
C.-W. Curr. Opin. Chem. Biol. 2000, 4, 678. (b) Hermann, T. Angew.
Chem., Int. Ed. 2000, 39, 1890. (c) Michael, K.; Tor, Y. Chem. Eur. J.
1998, 4, 2091. (d) Chow, C. S.; Bogdan, F. M. Chem. ReV. 1997, 97,
1489.
a All Kd values are in µM. b Higher concentrations were insoluble.
(4) Zaman, G. J. R.; Michiels, P. J. A.; van Boeckel, C. A. A. Drug DiscoVery
Today 2003, 8, 297.
(5) Dervan, P. B.; Edelson, B. S. Curr. Opin. Struct. Biol. 2003, 13, 284.
(6) (a) Russell, R. J. M.; Murray, J. B.; Lentzen, G.; Haddad, J.; Mobashery,
S. J. Am. Chem. Soc. 2003, 125, 3410. (b) Jin, E.; Katritch, V.; Olson,
W. K.; Kharatisvili, M.; Abagyan, R.; Pilch, D. J. Mol. Biol. 2000, 298,
95. (c) Ryu, D. H.; Litovchick, A.; Rando, R. R. Biochemistry 2002, 41,
10499. (d) Schroeder, R.; Waldsich, C.; Wank, H. EMBO J. 2000, 19, 1.
(e) Tok, J. B.-H.; Cho, J.; Rando, R. R. Biochemistry 1999, 38, 199-
206.
(7) Wilson, D. W.; Li, K. Curr. Med. Chem. 2000, 7, 73.
(8) Varani, G. Annu. ReV. Biophys. Biomol. Struct. 1995, 24, 379.
(9) Yoshizawa, S.; Fourmy, D.; Eason, R. G.; Puglisi, J. D. Biochemistry
2002, 41, 6263.
(10) Greenberg, W. A.; Priestley, E. S.; Sears, P. S.; Alper, P. B.; Rosenbohm,
C.; Hendrix, M.; Hung, S. C.; Wong, C. H. J. Am. Chem. Soc. 1999, 121,
6527.
(11) Stable RNA stem-loops were designed using mfold (Mathews, D. H.;
Sabrina, J.; Zuker, M.; Turner, D. H. J. Mol. Biol. 1999, 288, 911). The
corresponding pentaloop showed multiple structures in mfold and was
thus not utilized in this study.
(12) For use of fluorescently labeled RNA to determine Kd values with small
molecules, see: (a) Llano-Sotelo, B.; Chow, C. S. Bioorg. Med. Chem.
Lett. 1999, 9, 213. (b) Haddad, J.; Kotra, L. P.; Llano-Sotelo, B.; Kim,
C.; Azucena, E. F.; Liu, M.; Vakulenko, S. B.; Chow, C. S.; Mobashery,
S. J. Am. Chem. Soc. 2002, 124, 3229.
(13) Absence of aminoglycoside binding was also confirmed by RNase I
footprinting.
Figure 1. Binding of paromomycin and 14b with loop D.
footprinting experiment was conducted. Loop C was 5′ radiolabled
with γ-32P-ATP and incubated with various concentrations of 14b.
The gel in Figure 2 clearly reveals a strong protection from RNase
I digestion in the loop region. Analysis of the G10 band of this gel
via densitometry provides a Kd of 11 µM for the 14b-loop C
interaction, consistent with the value obtained from fluorescence.
The data thus indicate that deoxystreptamine dimers can bind
tightly to RNA hairpin loops of a variety of sizes. Although
deoxystreptamine itself is a very weak (Kd > 1 mM) binder to
distorted regions of RNA secondary structure,9 it is now apparent
that spacing two deoxystreptamine units with an appropriate linker
can give rise to much tighter RNA binding. Such compounds can
(14) For examples of such an approach, see: (a) Lee, J.; Kwon, M.; Lee, K.
H.; Jeong, S.; Hyun, S.; Shin, K. J.; Yu, J. J. Am. Chem. Soc. 2004, 126,
1956. (b) Krishnamurthy, M.; Gooch, B. D.; Beal, P. A. Org. Lett. 2004,
6, 63.
JA048936L
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9197