Mol. Cells, Vol. 11, No. 3, pp. 303-311
Molecules
and
Cells
KSMCB 2001
In Vitro Selection of RNA against Kanamycin B
Miyun Kwon, Sung-Min Chun, Sunjoo Jeong1, and Jaehoon Yu*
Life Sciences Division, Korea Institute of Science and Technology, Seoul 130-650, Korea;
1
Department of Molecular Biology, College of Natural Sciences, Dankook University, Seoul 140-714, Korea.
(Received November 30, 2000; Accepted February 23, 2001)
Aminoglycosides are well-known antibiotics that function by interacting with ribosomal RNA in bacteria. In
order to understand the molecular details between
RNA and the drug, RNA aptamer was selected against
kanamycin B. After 12 cycles of selection, RNA was
cloned and sequenced. Among 9 clones, sequences of
three clones were identical, suggesting the selected
RNA was enriched. Among the cloned RNA molecules,
the triplicated RNA was the maximum binding RNA.
It showed a 180 nM affinity (KD) to the cognate aminoglycoside, as measured by a surface plasmon resonance, and a competition assay using a fluorescence
anisotropy technique. The affinity of the maximum
binding RNA to a similar aminoglycoside, tobramycin,
was much stronger than 12 nM of KD. The binding site
of the aminoglycoside in the maximum binding RNA
was a stem loop located at the end of the 5’ region. A
stem loop structural motif, found in this study, was
similar to those previously reported, even though the
sequences of the RNA were totally different from the
known sequences of the aminoglycoside binding site of
other aptamers. The present study suggests that the
aminoglycoside-binding region in RNA does not have a
sequence specificity, but has a shape-specific bulged
stem loop, even though it has a nanomolar affinity.
Keywords: Aminoglycoside; BIAcore; Kanamycin; RNA;
SELEX.
Introduction
Targeting RNA for drug development is not a new concept. An antisense drug approach has been intensively
investigated for the last decade, leading to clinical trials.
* To whom correspondence should be addressed.
Tel: 82-2-958-5157; Fax: 82-2-958-5189
E-mail: jhoonyu@kist.re.kr
Specific sequenced oligo-nucleotide drugs were prepared
and administered to bind to the target RNA by WatsonCrick interactions in a single stranded region of RNA.
Recently, however, high field NMR studies on the structure of RNA molecules, such as Group 1 intron of hammerhead ribozyme (Stage et al., 1995; Von Ahsen et al.,
1991), HIV-1 mRNA Rev responsive element (Zapp et al.,
1993), and a trans-activation response element revealed
new structured motives in a three-dimensional manner
(Hendrix et al., 1997; Mei et al., 1997). These advances in
determining structures, and elucidating their functions,
led the pharmaceutical industry to regard RNA as new
therapeutic targets, which are conceptually different from
antisense RNA targets (Ecker and Griffey, 1999). Aminoglycosides are well-known antibiotics functioning as new
conceptual drugs. The drugs are binding to the decoding
region of bacterial 16S ribosomal RNA in a threedimensional manner, prohibiting protein synthesis
(Cundliffe, 1990; Noller, 1991). The main binding forces
are the electrostatic interaction between positively
charged amino groups in aminoglycosides, and negatively
charged phosphate groups in RNA backbone at physiological conditions (Moazed and Noller, 1991; Reinhart and
Shield, 1980; Tanaka, 1982). Since aminoglycosides are
generally strong binders to any type of RNA motives, the
drugs could be plausible candidates for any new RNA
target. However, the drugs showed relatively poor selectivity, because the main forces of RNA-aminoglycoside
recognition came from the nonspecific electrostatic interactions.
In vitro selection of RNA against a target molecule was
usually made with iterative cycles, known as systematic
evolution of ligands by exponential enrichment (SELEX),
or in vitro selection (Ellington and Szostak, 1990; Joyce,
Abbreviations:
CRP,
tetramethylrhodamine
labeled
paromomycin; NOE, nuclear overhauser effect; NMR, nuclear
magnetic resonance; PCR, polymerase chain reaction; RNA,
ribonucleic acid; Ru, resonance unit; SPR, surface plasmon
304
RNA against Kanamycin B
resonance.
1989; Joyce and Orgel, 1989; Schmidt, 1999). This method has permitted the identification of unique high affinity
RNA (aptamer) from a large population of a random RNA
library against a ligand(s) of interest (Eaton, 1997; Hermann and Patel, 2000; Lato et al., 1995; Lorsch and
Szostak, 1996). Efforts to select specific RNA sequences
against certain aminoglycosides have been reported for
understanding the rules underlying RNA-aminoglycoside
recognition, and for developing sequence-specific RNA
targeting therapeutics (Osbome and Ellington, 1997).
RNA aptamers against lividomycin showed identical sequences to those in rRNA of Hemophilus influenzae and
the parasite Leishmania (Lato and Ellington, 1996). This
suggests that the sequences could be new RNA target sites
to control those pathogens. Aptamer against neomycin
was also reported to have a certain stem loop structure
with wobble non-Watson-Crick base-pairings (Wallis et
al., 1995). The most successful in vitro selection of RNA
was achieved against tobramycin (Wang and Rando,
1995). The tightest aptamer showed a nanomolar affinity
to tobramycin with a 1,000 fold discriminated specificity
to similar aminoglycosides.
These results showed that the consensus RNA motif for
the aminoglycoside binding is a short stem with a bulge
or a loop (Cho et al., 1998; Jiang et al., 1998; Patel et al.,
1997; Wang et al., 1996). However, no consensus sequence has been found, and the binding affinity of the
aptamers was within a moderate range (Lato et al., 1995),
except the tobramycin-specific aptamer (Wang et al.,
1996). Kanamycin B, which has an additional 3-hydroxyl
group compared with tobramycin at the A ring (Fig. 1),
was thought to be a reasonable target for RNA selection.
It might elucidate any specific and consensus sequence
that is similar to tobramycin-specific aptamers. Then the
comparison of aptamer sequences would help to understand the general rules of recognition between aminoglycoside and RNA.
Here we report the in vitro selection of the RNA against
kanamycin B. After 12 cycles of selection, characteristic
sequences were found. Most of these showed a high
nanomolar affinity to the cognate aminoglycoside using a
surface plasmon resonance and fluorescence anisotropy
techniques. The truncation and footprinting studies of the
selected sequences elucidated that the binding site of the
drug is 5′-end with a stem loop structure, which show no
sequence homology to previously reported sequences.
Materials and Methods
In vitro selection of RNA against kanamycin The procedures
were modified from known protocols (Clark, 1998; Doudna et
al., 1995; Jayasena and Gold, 1997; Mezei and Storts, 1994).
Chemicals used for this study were from Sigma (USA). Random
131 nucleotides DNA library was purchased from the Midland
Fig. 1. Structures of aminoglycosides; kanamycin B, tobramycin,
kanamycin A, neomycin, paromomycin.
Certified Company (USA). The oligonucleotide was composed of 70 random sequences flanked by 20 nucleotides of
primer regions both at the 5′ and 3′ region, and the T7 promoter
at the 5′ region.
Polymerase chain reaction Ten microliters of cDNA was added to a mixture of 10× buffer (500 mM KCl, 100 mM Tris-HCl,
pH 9.0, 1% Triton X-100) 10 µl, 2.5 mM dNTP mixture 10 µl,
25 mM MgCl2 8 µl, 2.5 µM 3′ primer 1 µl, 2.5 µM 5′ primer 1
µl, and Taq polymerase (Promega, USA; 1 unit/µl) 1 µl to make
a 100 µl solution, then amplified to double-stranded DNA by a
Perkin-Elmer Model 2400 thermal cycler. The PCR mixture was
subjected to 20 repetitions of three cycles; each cycle was 95°C,
30 s; 58°C, 30 s; 74°C, 30 s. Eight microliters from the resulting solution were used for a 1.5% agarose electrophoresis to
confirm the correct size; the rest were concentrated by ethanol
precipitation and dissolved with 20 µl of water for the next procedure.
In vitro transcription A mixture of the PCR solution (5 µl), 5×
buffer (200 mM Tris-HCl, 30 mM MgCl2, 10 mM spermidine,
50 mM NaCl at pH 7.9; 20 µl), 100 mM DL-dithiothreitol
(DTT; 20 µ1), 2.5 mM NTP mixture (20 µ1), T7 RNA polymerase (Promega, USA; 50 units/µ1; 1 µ1) and deionized water (34
µ1) was incubated at 37°C for 2 h. One µ1of RNase-free RQ1
DNase (Promega, USA; 1 unit/µ1) was then added. After incubation for 10 min at 37°C, an equal volume (100 µ1) of the PCI
mixture (phenol:chloro-form:isoamylalcohol = 25:24:1) was
added and stirred for 5 min. The resulting solution was centrifuged at 14,000 rpm for 10 min. The upper phase was transferred to a new tube and concentrated by ethanol precipitation.
The resulting RNA was purified by gel electrophoresis with 7.0
M urea containing denaturating polyacrylamide (6%) at 20 mA
for 30 min. A UV illuminated RNA band was cut and transferred
to a tube containing 500 µ1 of the elution buffer (0.5 M ammonium acetate, 1 mM EDTA, 0.2% SDS at pH 8.0) and allowed to
stand at 37°C for 4 h. The eluted RNA was transferred to a tube
and purified by phenol extraction and ethanol precipitation. The
amount of RNA was measured at 260 nm using a UV spectro-
Miyun Kwon et al.
photometer.
Selection of RNA by affinity chromatography One hundred
micrograms of the original RNA library (110-mer) were used for
the first selection cycle. For 2 and 3 cycles, 20 µg of RNA was
used, then 5 µg for the next selection cycle. RNA was diluted
with 300 µl of the binding buffer (250 mM NaCl, 1 mM MgCl2,
0.1 mM EDTA, 1 mM DTT, 20 mM Tris-HCl at pH 7.5),
denaturated for 5 min at 65°C, and allowed to cool to room
temperature. The RNA solution was transferred to a tube containing 200 µl (6 mg/ml) of ethanolamine-immobilized oxirane
acrylic beads and incubated for 30 min with vortexing. Following quick centrifugation, the solution was carefully transferred
to a tube containing 100 µl (0.14 mg/ml) of kanamycinimmobilized oxirane acrylic beads. The resulting suspension
was incubated for 30 min at room temperature. Following quick
centrifugation, the solution was removed. The beads were then
washed 8 times with 500 µl of a binding buffer. Then, 400 µl of
kanamycin B (800 µM) in the same buffer was added for affinity elution. The resulting suspension was incubated for 30 min.
Following quick centrifugation, the solution was carefully transferred to a tube containing a precooled mixture of 40 µl of 3 M
NaOAc (pH 5.2) and 1 ml of absolute ethanol. The eluted RNA
was then pre-cipitated and centrifuged at 14,000 rpm for 30 min
at 4°C. The solution phase was then removed. Next 200 µl of
70% ethanol was added to the pellet. Following quick stirring
for 1 min, the resulting solution was again centrifuged at 14,000
rpm for 5 min at 4°C. The solution phase was then removed, and
the resulting RNA was subjected to RT-PCR for the next selection cycle.
Reverse transcription Eluted RNA was diluted with 40 µl of
deionized water. To this solution, 1 µl (2.5 µM) of a 3’-primer
was added. The resulting solution was heated at 65°C for 5 min
and cooled on an iced bath. A mixture of 10 mM dNTP mixture
(6 µl), 5× buffer (250 mM Tris-HCl, 375 mM KCl, 15 mM
MgCl2, 50 mM DTT at pH 8.3; 15 µl), and Moloney murine
leukemia virus RTase (Promega, USA; 50 units/µl; 1 µl) were
added to the solution. The mixture was incubated at 37°C for 40
min. The resulting cDNA solution was allowed to stand at 95°C
for 5 min to remove RTase activity. The cDNA solution was
then cooled and used for the PCR without further purification.
Generation of 32P labeled RNA and affinity chromatography
A reaction mixture was comprised of 1 µg of cDNA, 5 µl of 5×
buffer, 5 µl of 100 mM DTT, 4 µl of 2.5 mM A,G,CTP mixture,
0.5 µl of 100 µM UTP, 2 µl of [α-2P]UTP (Amersham Pharmacia Biotech, UK; 20 mCi/ml), and 1 µl of T7 RNA polymerase
(Promega, USA; 50 units/µl) to make a total volume of 25 µl.
Conditions for the enzyme reaction and purification were the
same as those used for the cold NTP mixture. Purified RNA was
diluted with 20 µl of deionized water. Then, one microliter of
the RNA solution was diluted with 5 ml of a scintillation cocktail for β-ray counting. The total amount of RNA for the binding
assay was adjusted to 4 × 104 CPM. The hot RNA was affinity
chromatographed by the same protocols used with cold RNA.
305
All of the solutions from each fraction of washing and elution
were counted separately using a β-counter (Wang and Table 1.
Sequences of kanamycin B - specific RNA.
5'-CCATAATACGACTCACTATAGGGGAGCTCGGTACCGAATTC- N70 -AAGCTTTGCAGAGGATCCTT-3'
T7 RNApolymerase
K6
K7
K8
K9
K12
K20
K21
K31
K32
1
41
1
41
1
41
1
41
1
41
1
41
1
41
1
41
1
41
SacI
TGAGTTGGCT
CTTGACCCTG
TCGCCCTATA
GCGGTGGCCA
TCGCCCTATA
GCGGTGGCCA
TCGCCCTATA
GCGGTGGCCA
ACAAXACXXX
GCGGTGCGAT
TGAGTTGGCT
CTTGACCCTG
GGTATTGTTA
GATCGACTAC
GTGAAATACC
AGGCTAATCC
GGTATGAAGT
CTAACAGTCC
KpnI
GCGAAGGACC
GAGAAGCCTG
GGGGTGTTGA
GAACTTTTCG
GGGGTGTTGA
GAACTTTTCG
GGGGTGTTGA
GAACTTTTCG
XCXCCGGATC
GAATXGTGGT
GCGAAGGACC
GAGAAGCCTG
ACCTGAGAAT
CGTCCTGAAG
TGTTAGTTAT
AATGCTCGCT
CTGGTTCGCG
AGTGAGTCAA
EcoRI
HindIII
CGCGGACAAG
AGCGCGGGTA
GGGAAAGTGT
TTCTCATCAA
GGGAAAGTGT
TTCTCATCAA
GGGAAAGTGT
TTCTCATCAA
GCGCGACATC
CTTTGTCTGC
CGCGGACAAG
AGCGCGGGTA
TTTCTGCGGC
TAGGCTCGTA
TTGTAACGTG
TGTACTGTAC
GTGGCGTGGG
TACTTTGTTC
CGGTTTAAGG
GCGACAAGGT
GCGACAAGGT
GCGACAAGGT
CAATGGTTCC
CGGTTTAAGG
TTTAATTGGG
AAGGTAACTG
TGTAAAAAAC
BamHI
40
70
40
70
40
70
40
70
40
70
40
70
40
70
40
70
40
70
Rando, 1995). After 12 cycles of selection, and cloning of the
selected RNA against kanamycin B, several clones were obtained (Table 1). The binding constants were measured by a
surface plasmon resonance technique; most of these showed
nanomolar affinity to kanamycin.
Cloning and sequencing The selected RNAs after 12 cycles
were cloned and sequenced as a standard protocol (Clark, 1988;
Mezei and Storts 1994).
Affinity measurement using BIAcore analysis The procedures were modified from known protocols (Hendrix et al.,
1997; Kraus et al., 1998). The assay was performed at room
temperature, except as otherwise noted.
Immobilization of Kanamycin B on CM5 sensorchip A
carboxymethylated sensorchip (Biacore, Sweden) was equilibrated with 200 µl of 100 mM N-[2-hydroxyethyl]piperazine-N’[2-ethanesulfonic acid] (HEPES) at pH 8.0. It was then activated
by injection of 200 µl of 1:1 mixture of 400 mM EDC and 100
mM NHS in water to a flowcell. Five hundred µl of kanamycin
B (1 mM in 100 mM HEPES at pH 8.0) was injected with a
flow rate of 10 µl/min using the KINJECT command. Twenty
microliters of 1.0 M ethanolamine in the same buffer was then
injected to block the remaining activated groups on the surface
of the sensorchip. As a reference, an ethanolamine solution was
injected and immobi-lized to a different flowcell. Two hundred
of the Resonance Units (RU) were increased in the sensogram
by the previously mentioned immobilization process.
Binding affinity measurement using kanamycin-immobilized
sensor chip Prior to injection, each RNA sample in the running
buffer (10 mM HEPES, 0.1 mM EDTA, 100 mM NaCl, pH 6.8)
was heated to 70°C for 5 min and allowed to cool slowly to
306
RNA against Kanamycin B
room temperature in order to make proper 3-D structures. RNA
samples (original pool of RNA, selected pool of RNA, and 7
clones of RNA) were serially diluted with the running buffer
(250 mM NaCl, 1 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, 20
mM Tris-HCl at pH 7.5) to make 1, 0.5, 0.25, 0.12, and 0.06 µM
concentrations, respectively. All of the samples were injected in
the serial automated mode: sample injection (80 µl), regeneration (5 µl of 1 M NaCl, 50 mM NaOH) and re-equilibration
(with 100 µl of the running buffer). Each sample was injected at
a flow rate of 10 µl/min using the KINJECT command.
Determination of dissociation constants using fluorescence
anisotropy Tetramethylrhodamine labeled paromomycin (CRP)
was obtained from Professor Rando (Harvard Medical School,
USA) and used for a fluorescent probe without further purification. A fluorescence anisotropy measurement was performed on
a Perkin-Elmer LS-50B luminescence spectrometer equipped
with a thermostat at 20°C. The tracer solution was excited at
510 nm, and monitored at 550 nm. For every single point, seven
measurements were made and averaged. The measurement was
per-formed in a buffer containing 140 mM NaCl, 5 mM KCl,
1mM MgCl2, and 20 mM HEPES at pH 7.5. The equation in the
reference was used for the determination of the dissociation
constant between RNA and CRP (Kd):
A = A0 + ∆A{([RNA]0 + [CRP]0 + Kd) – ([RNA]0 + [CRP]0 +
Kd)2 – 4[RNA]0[CRP]0 1/2}/2
Where A and A0 are the fluorescence anisotropy of CRP in the
presence and absence of RNA, respectively, ∆A is the different
concentration of RNA minus the fluorescence anisotropy in the
absence of RNA. [RNA]0 and [CRP]0 are the initial concentrations of RNA and CRP, respectively.
In the competitive binding assay, the following equation was
used for the calculation of KD values:
[Aminoglycoside]0 = {KD(AΩ – A)/[Kd(A – A0) + 1]} ×
{[RNA]0 – Kd(A – A0)/(AΩ – A0) –
[CRP]0(A – A0)/(AΩ – A0)}
Where KD is the dissociation constant between RNA and the
aminoglycoside, and [Aminoglycoside]0 is the initial concentration of aminoglycosides. A, AΩ, and A0 are fluorescence anisotropy values of the sample, totally bound tracer, and totally free
tracer, respectively. Both Kd and KD were determined by a nonlinear curve fitting using the above two equations.
Synthesis of truncated RNA and measurement of binding
affinity
Synthesis of RNA Oligonucleotide DNAs were obtained from
Oligo’s, Etc., Inc. In order to make four truncated RNAs (K8-1,
K8-1-1, K8-2, and K8-3), a synthetic oligonucleotide DNA template was in vitro transcribed using T7 RNA polymerase. All of
the DNA templates were designed to contain GG at the end of
5′-end to increase the efficiency of the transcription. The sequences of four synthetic oligonucleotides are as follows:
K8-1
K8-1-1
K8-2
K8-3
5′-AATTTAATACGACTCACTATAGGGAGCTCGGTACCGAATTCTCGCCCTATAGGGGT-3′
5′-AATTTAATACGACTCACTATAGGGAGCTCGGTACCGAATTCTC-3′
5′-AATTTAATACGACTCACTATAGGGAAAGTGTGCGACAAGGTGCGGTGGCCAGAACTTTTC-3′
5′-AATTTAATACGACTCACTATAGGGTTCTCATCAAAAGCTTTGCAGAGGACC-3′
Affinity measurement The truncated RNAs were affinity measured using the fluorescence anisotropy technique as was described in the previous section.
RNA footprinting study
RNA labeling The first step for the RNA footprinting study was
the 5′-end labeling of K8 RNA. The K8 RNA was
dephosporylated by treatment of RNA with alkaline phosphatase
(New England Biolab, USA), followed by phenol extraction and
ethanol precipitation. In the second step, a mixture of 32P labeled gamma ATP, and the purified RNA, was in vitro treated
with polynucleotide kinase (New England Biolab, USA). The
resulting RNA was purified by ethanol precipitation. The precipitated RNA was then gel electrophoresed on 6% acrylamide-7 M
urea. The right sized RNA was cut from the gel and eluted with
a 0.3 M NaOAc solution followed by ethanol precipitation.
Gel Electrophoresis The labeled RNA (2 × 106 cpm) was then
incubated in the presence (1 µM–100 µM) and absence of tobramycin for 10 min, followed by a reaction with RNase T1
(New England Biolab, USA) for 8 min at room temperature. The
reaction was quenched by the addition of 3 M NaOAc (1/10
volume), then purified with ethanol precipitation. Gel electrophoresis was performed for 2 h on a 15% acrylamide-7 M urea
gel. An alkaline hydrolysis ladder and RNase T1 sequencing
ladder were included on the gels to permit identification of individual bands.
Results and Discussion
Selection of RNA In order to select kanamycin B-specific
RNA, kanamycin B was immobilized to Sepharose beads
using amine functionality in a mild basic condition (Yu et
al., 1999). The quantity of immobilized kanamycin B was
measured by a ninhydrin test. A purchased DNA library
was multiplicated by a polymerase chain reaction to make
at least 3.5 × 1014 of diversity for the first selection cycle.
The initial RNA library was filtered through ethanolamine-immobilized beads to remove non-specific binders.
Then, the resulting RNA was incubated with kanamycin
B-immobilized beads. Twenty column volumes of washing with a binding buffer removed weak binders. Then the
remaining RNA was affinity eluted with 2.0 mM of kan-
Miyun Kwon et al.
307
amycin B. The eluted RNA was purified by ethanol
precipitation followed by reverse-transcription and ampli-
RNA. The result suggested that the selected RNA was
affinity maturated, even though it was composed of several
Table 2. Binding affinities of various RNA to kanamycin B
using BIAcore.
Table 3. Binding affinities of K8 RNA with a variety of aminoglycosides using fluorescence anisotropy.
RNA
Random RNA Pool
Selected RNA Pool
K8 RNA
ka (1/Ms)
kd (1/s)
KD (M)
7.78 × 102
8.79 × 104
6.45 × 104
2.37 × 10-4
3.84 × 10-4
5.65 × 10-5
3.05 × 10-7
4.37 × 10-9
8.76 × 10-10
fied by PCR for the next cycle.
After 12 cycles of selection, the eluted RNA reached
about a 8.0% level of the applied RNA using 32Plabeled RNA. For comparison, the eluted RNA from the
initial RNA library was only a 0.4% level of the applied
RNA at the same elution condition. Since further selection
cycles of RNA did not increase the quantity of eluted
RNA, 12-cycled RNA was cloned. Nine clones of the
RNA were obtained and sequenced, as shown in Table 1.
Three sequences were identical, suggesting the selected
RNA was partially enriched. There was, however, no noticeable sequence homology between other cloned RNAs
except triplicated sequences. A homology search was carried out on clone K8 and other clones using the BLAST
program. K8 RNA showed no interesting sequence homology to any pathogenic target.
Affinity measurement of RNA with aminoglycosides
Initially, a surface plasmon resonance technique was used
for the affinity measurement of a variety of RNAs to the
cognate aminoglycoside. The technique had several advantages over conventional biological binding assays: the
association and dissociation constants could be measured
by the real time kinetics between analytes and ligands in
solution, and in the immobilized matrix, respectively. The
procedure for the immobilization of ligand was nondestructive and safe, because it needed no radioisotope
labeling or fluorescent tagging. In order to measure affinities, the initial immobilization of kanamycin B to a sensor
chip was carried out (described in the experimental section). To the immobilized aminoglycoside, solutions of
RNA were then eluted as analytes for the affinity measurement. We also tried to measure affinities with a reverse
method, using immobilized RNA and aminoglycoside
solution as an analyte; but, the RU in the reverse method
was less sensitive than the initial method, due to the small
change of the aminoglycoside binding.
The binding affinity of the RNA was measured by
changing at least 5 different concentrations of RNA as an
analyte in solution with immobilized kanamycin B for the
accurate measuring. The selected affinity constants were
tabulated in Table 2. First of all, the selected pools of
RNA showed as much as 100 times the strong binding
constant in comparison to that of the original pool of
Aminoglycoside
Tobramycin
Kanamycin B
Kanamycin A
Neomycin
Paromomycin
KD (nM)
Error boundary (R value)
11.6
181
4400
1100
1500
3.7 (R = 0.997)
55 (R = 0.995)
700 (R = 0.995)
310 (R = 0.996)
330 (R = 0.990)
eral clones. The binding affinities of individual cloned
RNA were also measured by the same method. The maximum binding RNA was the triplicated RNA, suggesting
that the enrichment of RNA was almost done. The binding
affinities of other clones were almost the same as that of
the selected RNA pool. Further binding studies concentrated on the maximum binding RNA (K8 RNA).
The affinity of the K8 RNA to the cognate molecule was
confirmed by a competition assay using a fluorescent anisotropy and CRP as a fluorescence probe. The binding
affinity (Table 3) of K8 RNA to kanamycin B was measured as 180 nM of KD, which is observed in the same order of magnitude by use of the SPR technique. The advantage of this competition assay is the ability to compare
the RNA binding affinities to those of similar aminoglycosides, such as kanamycin A, tobramycin, neomycin,
paromomycin, and streptomycin (Fig. 1). As shown in
Table 3, the RNA binding affinities to neomycin or
paromomycin were weak, suggesting that the binding
might be specific to kanamycin type aminoglycosides.
Indeed, binding affinities of K8 RNA to tobramycin were
measured to be 12 nM, discriminating by factors of 100−
1,000 to neomycin type molecules. The discrimination of
the RNA to tobramycin and kanamycin B is critical, suggesting that the A ring of the aminoglycoside is very important in RNA binding. The sharp decrease of binding
affinity to kanamycin A also suggested that one of the
most important interactions is derived from the 2-amino
group in the A ring. Probably, the lack of this interaction
in kanamycin A might require a totally different binding
mode from that of kanamycin B or tobramycin, affording
a low binding affinity. Hydroxyl groups in the A ring also
contributed to make a strong binder. The 4-Hydroxyl position seemed to be more important than the 3-position,
considering that tobramycin is a tighter binder to K8 RNA
than is the cognate molecule. It seemed that the 4hydroxyl group in kanamycin B resulted in the
intramolecular hydrogen bonding with the 3-hydroxyl
group, while losing the specific hydrogen bonding donor
to the RNA molecule. The 4-hydroxyl group was available for making such specific hydrogen bonding in tobramycin. However, a structural study is needed for elucida-
308
RNA against Kanamycin B
tion of the detailed molecular recognition of the aminoglycosides and RNA.
radiodiagram with the 5′-end labeled RNA, however,
showed dramatic different ladders (Fig. 3) in the presence
Table 4. Relative binding affinities of truncated RNAs with tobramycin and neomycin using fluorescence anisotropy.
Binding affinities to aminoglycoside
RNA
Tobramycin
Neomycin
+++
++
++
+
−
+
−
−
−
−
K8 (110 mer)
K8-1 (35 mer)
K8-1-1 (22 mer)
K8-2 (40 mer)
K8-3 (35 mer)
+++; <10 nM, ++; 1.0 µM – 10 nM, +; 1.0 – 10 µM, −; > 10 µM.
1
2
3 4 5
6
Fig. 2. Secondary structure of K8 RNA.
Truncation and footprinting study Secondary structures
of the triplicated sequenced RNA was predicted and shown
in Fig. 2 using the program Mulfold (Mathew et al., 1999).
By this secondary structure, four truncated RNA ules
were designed; first 1-22 (K8-1-1); second 1-35 (K8-1);
third 42-79 (K8-2); final 82-110 (K8-3). The four RNA
molecules were made by an in vitro transcription reaction
from the corresponding double stranded template RNA
with a T7 promoter region at 5′-ends. Then, the relative
binding affinities of these truncated RNAs was carried out
utilizing a fluorescence anisotropy technique. Competitive binding affinities of the truncated RNA molecules to
tobramycin and neomycin were tabulated in Table 4. The
result showed that the binding affinities of the K8-1-1 and
K8-1 RNA molecules to tobramycin were reduced a little
in comparison with the full-length K8 RNA. However,
affinities of the K8-2 and K8-3 RNA molecules were
sharply reduced, suggesting that the binding site of the
aminoglycoside is at the 5′-end in the full-length RNA.
The shortest truncated K8-1-1 RNA that is composed of
22 nucleotides was long enough to accommodate three
rings of aminoglycosides. On the other hand, neomycin
was not bound to any truncated RNA molecules, suggesting that the affinity of the K8-1 and K8-1-1 RNA molecules is specific for tobramycin.
In order to confirm the result of the truncation study,
and to define the exact binding region of tobramycin in
K8 RNA, footprinting of K8 RNA was carried out. Initially, K8 RNA was 32P labeled both by the 5′- and 3′-ends.
As indicated by the truncation study, the binding site of
the aminoglycosides was at the end of the 5′-region.
Therefore, the autoradiodiagram with the 3′-end labeled
RNA showed identical ladders, regardless of tobramycin
in the RNA solution (data not shown). The auto-
Fig. 3. Tobramycin binding to K8 RNA. Autoradiodiagram of a
15% denatured polyacrylamide gel showing a tobramycin binding site of 5′-32P labeled K8 RNA. Lane 1, Intact K8 RNA; lane
2, alkaline hydrolysis; lane 3, G-specific RNase T1 sequencing
Miyun Kwon et al.
reaction; lane 4, without tobramycin; lane 5, 1.0 µM tobramycin; lane 6, 10 µM tobramycin; lane 7, 100 µM tobramycin.
and absence of tobramycin. As shown in the
autoradiodiagram, the ladder of the 5′-region RNA disappeared in the presence of as low as 1.0 µM of a concentration of tobramycin This is due to the binding of the
drug to the 5′-end region of the RNA molecule. The result
suggested that a stem loop from the 5′-end of K8 RNA is
a distinctive binding site of the aminoglycoside. The
footprinting result by a RNase T1 reaction was well correlated with the result of the truncation study.
General motives of RNA for aminoglycoside binding
The results of the truncation and footprinting studies suggest that the binding site of the aminoglycoside is composed of a short stem with a bulge and a loop at the end of
the 5′-region. As previously reported, the NMR structural
study of the aminoglycoside-RNA complex of tobramycin-specific aptamer suggested that a short stem loop is
the most consensus-binding motif, regardless of aminoglycosides as ligands (Cho et al., 1998; Jiang et al., 1997;
Wallis et al., 1995). The structures of RNA motives in
various aminoglycosides specific aptamers are summarized in Fig. 4. As shown in the figure, there are several
distinctive characters in the stem loop, which has at least
22-nt long sequences. The aptamer against kanamycin B
in this study has the shortest 22-nt long sequences.
A non-Watson-Crick base pairing(s) resulted in widening the groove to accommodate relatively large molecules,
such as aminoglycosides. A stem found in this work also
has two G-U non-Watson-Crick base-parings. The G-U
wobble base pairs were also found in aptamers against
neomycin (Wallis et al., 1995), suggesting that wobble GU base-pairing is essential for aminoglycosides binding.
Bulges are also observed as general and essential motif
for the aminoglycoside binding. A major role of the
bulge(s) is to widen the groove of the stems to accommodate larger molecules (Jiang et al., 1997). There were two
small symmetrical bulges in K8 RNA in this study in
comparison with the usual unsymmetrical bulges found in
most of the aminoglycoside-specific RNA. Another additional role of the bulge base might provide extrainteraction with an alicyclic ring (B ring) in aminoglycosides. This interaction was closely examined by NMR
studies, and showed that a single bulged base, positioned
as a flap, closed the groove (Jiang et al., 1997). Removal
of a small bulge showed a reduction of binding affinity in
tobramycin aptamers (Cho et al., 11998). Neomycin
aptamer also showed a short unsymmetrical bulge, which
showed direct interactions with the drug by footprinting
assay (Wallis et al., 1995).
Finally, the loop flanked by the stem seems to be essential for the binding. A mutational study of the tobramycin
specific RNA showed that the sequences of the loop did
affect the strong binding (Cho et al., 1998). The foot
309
printing result of kanamycin B aptamer in this study confirmed that a loop structure is an essential element for the
A
B
C
D
E
Fig. 4. Comparison of a secondary structure of RNA against
various aminoglycosides. A. aptamer against tobramycin (Cho
et al., 1998). B. aptamer against tobramycin (Jiang et al., 1997).
C and D. aptamers against neomycin (Wallis et al., 1995). E.
aptamer against kanamycin B (this work).
aminoglycoside binding, even though the loop is composed of only 6 bases. NMR studies of the tobramycin
specific loop also suggested that adjacent bases along the
stem are directly conjugated with aminoglycosides by
tracing NOEs of imino protons. This suggests that the
adjacent bases are interacting with the aminoglycoside
(Jiang et al., 1997); however, no sequence homology was
observed in the bases of the loops among aminoglycoside
aptamers.
Our experimental data, and other previous results, led
us to conclude that there is no consensus sequence in the
aminoglycoside binding region, even though shape complementarities and electrostatic interactions accommodated strong binding between aminoglycoside and the RNA
aptamers. Instead, there were spatial stem-loop structural
motives in aminoglycoside-specific RNA. Many RNA
sequences, then, would fit these general recognition motives. However, as shown in the conserved domain in 16S
ribosomal RNA, the drugs must differentiate one base
change in 27 nt RNA sequences between human and E.
coli (Fourmy et al., 1996). This criterion requires that
further refining of existing aminoglycosides is necessary
for more specific drugs. Addition of another functional
310
RNA against Kanamycin B
group(s) to the aminoglycoside backbone, for example,
could increase the specificity to target RNA sequences by
addition of extra interactions (Kirk et al., 2000). Efforts to
increase those specificities between the drug and RNA are
in progress.
In summary, the experiments reported are the selection
of RNA sequences against kanamycin B. The selected
sequences showed a stem loop structure of the aminoglycosides binding region with 12 nM affinity. The aminoglycoside binding region is the 5′-end; this is proven by
truncation and a RNA footprinting study.
Acknowledgments This work is financially supported by the
21st Century Frontier Genome Research Project. JY is a beneficiary of a fund from KIST (V00443). We thank Mr. Joon-Young
Jeon for his cordial help in a SPR study.
References
Cho, J., Hamasaki, K., and Rando, R. R. (1998) The binding site
of a specific aminoglycoside binding RNA molecule. Biochemistry 37, 4985−4992.
Clark, J. M. (1988) Novel non-templated nucleotide addition
reactions catalyzed by prokaryotic and eukaryotic DNA polymerases. Nucleic Acids Res. 16, 9677−9686.
Cundliffe, E. (1990) Recognition site for antibiotics within
rRNA. In the ribosome: structure, function, and evolution,
Hill W. E. (ed.), pp. 479−490, American Society for Microbiology, Washington, D. C.
Doudna, J. A., Cech, T. R., and Sullenger, B. A. (1995) Selection of an RNA molecule that mimics a major autoantigenic
epitope of human insulin receptor. Proc. Natl. Acad. Sci.
USA 92, 2355−2359.
Eaton, B. E. (1997) The joys of in vitro selection: chemically
dressing oligonucleotides to satiate protein targets. Curr.
Opin. Chem. Biol. 1, 10−16.
Ecker, D. J. and Griffey, R. H. (1999) RNA as a small-molecule
drug target: doubling the value of genomics. Drug Dev. Today 4, 420−429.
Ellington, A. D. and Szostak, J. W. (1990) In vitro selection of
RNA molecules that bind specific ligands. Nature 346, 818−
822.
Fourmy, D., Recht, M. I., Blanchard, S. C., and Puglisi, J. D.
(1996) Structure of the A site of Escherichia coli 16S ribosomal RNA complexed with an aminoglycoside antibiotic.
Science 274, 1367−1371.
Hendrix, M., Priestley, E. S., Joyce, G. F., and Wong, C.-H.
(1997) Direct observation of aminoglycoside-RNA interactions by surface plasmon resonance. J. Am. Chem. Soc. 119,
3641−3648.
Hermann, T. and Patel, D. J. (2000) Adaptive recognition by
nucleoic acid aptamer. Science 287, 820−825.
Jayasena, V. K. and Gold, L. (1997) In vitro selection of selfcleaving RNAs with a low pH optimum. Proc. Natl. Acad.
Sci. USA 94, 10612−10617.
Jiang, L., Suri, A. K., Fiala, R., and Patel, D. J. (1997) Saccha-
ride-RNA recognition in an aminoglycoside antibiotic-RNA
aptamer complex. Chem. Biol. 4, 35−50.
Joyce, G. F. (1989) RNA evolution and the origins of life. Nature 338, 217−224.
Joyce, G. F. and Orgel, L. E. (1989) The RNA World, Gesteland,
R. F. and Atkins, J. F. (eds.), pp. 1−14, Cold Spring Harbor
Laboratory Press, Cold Spring Harbor.
Kirk, S. R., Luedtke, N. W., and Tor, Y. (2000) Neomycinacrydine conjugate: a potent inhibitor of Rev-Rre binding. J.
Am. Chem. Soc. 122, 980−981.
Kraus, E., James, W., and Barclay, A. N. (1998) Cutting edge:
novel RNA ligands able to bind CD4 antigen and inhibit
CD4+ T lymphocyte function. J. Immunol. 160, 5209−5212.
Lato, S. M. and Ellington, A. D. (1996) Screening chemical
libraries for nucleic-acid-binding drugs by in vitro selection:
a test case with lividomycin. Mol. Divers. 2, 103−110.
Lato, S. M., Boles, A. R., and Ellington, A. D. (1995) In vitro
selection of RNA lectins: using combinatorial chemistry to
interpret ribozyme evolution. Chem. Biol. 2, 291−303.
Lorsch, J. R. and Szostak, J. W. (1996) Chance and necessity in
the selection of nucleic acid catalysts. Acid. Chem. Res. 29,
103−110.
Mathews, D. H., Burkard, M. E., Freier, S. M., Wyatt, J. R., and
Turner, D. H. (1999) Predicting oligonucleotide affinity to
nucleic acid targets. RNA 5, 1458−1469.
Mei, H. Y., Galan, A. A., Halim, N. S., Mack, D. P., Moreland,
D. W., Sanders, K. B., Truong, H. N., and Czarnik, A. W.
(1997) Discovery of selective, small-molecule inhibitors of
RNA complexes-I. The Tat protein/TAR RNA complexes required for HIV-1 transcription. Bioorg. Med. Chem. Lett. 5,
1173−1184.
Mezei, L. M. and Storts, D. R. (1994) PCR Technology: current
innovations. pp. 21−28, CRC Press, Boca Raton, FL.
Moazed, D. and Noller, H. F. (1987) Interactions of antibiotics
with functional sites in 16S ribosomal RNA. Nature 327,
389−394.
Noller, H. F. (1991) Ribosomal RNA and translation. Ann. Rev.
Biochem. 60, 191−227.
Osbome, S. E. and Ellington, A. D. (1997) Nucleic acid selection and the challenge of combinatorial chemistry. Chem.
Rev. 97, 349−370.
Patel, D. J., Suri, A. K., Jiang, F., Jiang, L., Fan, P., Kumar, R.
A., and Nonin, S. J. (1997) Structure, recognition, and adaptive binding in RNA aptamer complexes. Mol. Biol. 271,
645−664.
Rinehart, K. L. Jr. and Shield, L. S. (1980) Aminoglycoside
antibiotics: an introduction. In Aminocyclitol antibiotics.
Rinehart, K. L. and Suami, T. (eds.), pp. 1−11, American
Chemical Society, Washington D.C.
Schmidt, F. J. (1999) Ribozymes-why so many, why so few?
Mol. Cells 9, 459−463.
Stage, T. K., Hertel, K. J., and Uhlenbech, O. C. (1995) Inhibition of hammerhead ribozyme by neomycin. RNA 1, 95−101.
Tanaka, N. (1982) Mechanism of action of aminoglycoside antibiotics. Handbook of Experimental Pharmacology, Vol. 62,
pp. 221−240, Springer-Verlag, New York.
Von Ahsen, U., Davies, J., and Schroeder, R. (1991) Antibiotic
inhibition of group 1 intron ribozyme function. Nature 353,
368−370.
Wallis, M. G., von Ahsen, U., Schroeder, R., and Famulok, M.
Miyun Kwon et al.
(1995) A novel RNA motif for neomycin recognition. Chem
Biol. 2, 543−552.
Wang, Y. and Rando, R. R. (1995) Specific binding of aminoglycoside antibiotics to RNA. Chem. Biol. 2, 281−290.
Wang, Y., Killian, J., Hamasaki, K., and Rando, R. R. (1996)
RNA molecules that specifically and stoichiometrically bind
aminoglycoside antibiotics with high affinities. Biochemistry
35, 12338−12346.
Yu, J., Chun, S. M., Park, H., and Jeong, S. (1999) Picomolar
311
scale determination carbohydrates covalently immobilized on
activated beads using hydroxyl functionality. J. Biochem.
Mol. Biol. 32, 98−102.
Zapp, M. L., Stern, S., and Green, M. R. (1993) Small molecules
that selectively block RNA binding of HIV-1 Rev protein inhibit Rev function and viral production. Cell 74, 969−978.