Binding of aminoglycoside antibiotics with modified A-site 16S rRNA construct containing non-nucleotide linkers

Article abstract of DOI:10.1016/S0960-894X(01)00760-0

The design and synthesis of synthetically modified cyclic A-site 16S rRNA construct is reported. The binding characteristics of several members of the aminoglycoside antibiotics with this novel class of synthetically modified A-site 16S rRNA constructs we

Full text of DOI:10.1016/S0960-894X(01)00760-0

Bioorganic & Medicinal Chemistry Letters 12 (2002) 365–370
Binding of Aminoglycoside Antibiotics with Modified A-site 16S
rRNA Construct Containing Non-Nucleotide Linkers
Jeffrey B.-H. Tok,* Wilson Wong and Nyla Baboolal
Department of Chemistry & Natural Sciences, York College and Graduate Center of the City University of New York,
94-20 Guy R. Brewer Blvd., Jamaica, NY 11451, USA
Received 27 August2001; accepted 9 November 2001
Abstract—The design and synthesis of synthetically modified cyclic A-site 16S rRNA construct is reported. The binding char-
acteristics of several members of the aminoglycoside antibiotics with this novel class of synthetically modified A-site 16S rRNA
constructs were subsequently investigated. # 2002 Elsevier Science Ltd. All rights reserved.
The potential in utilizing synthetic oligonucleotides as a
therapeutic approach to control gene expression has
been widely explored.1À3 Modifications in the backbone,
sugar and bases of oligonucleotides have all been well
documented to result in higher binding affinities against
their complementary target nucleic acid, and this
approach was generally referred to as the ‘antigene’
approach.4À11 In order to obtain more stable duplex
(two single-stranded nucleic acid) molecules, modifica-
tions to the hairpin loop have also been reported. For
example, instead of nucleotide-containing hairpin loop,
synthetic linkers containing either the ethylene glycol or
stilbenedicarboxamide functional moieties have been
synthesized to replace the natural hairpin loop.12À16
Such resulting constructs have been observed to form
exceptionally stable duplex structures, with a marked
increase in thermal stability (Tm) values.13,17,18 Specifi-
cally, Ma and coworkers have utilized a series of syn-
thetic linkers derived from hexaethylene glycol to
generate synthetically modified double stranded TAR
RNA construct.12,13 They have proceeded to attach the
synthetic linkers at both ends of the RNA duplex to
create cyclic miniduplexes.12 On a similar note, Pils and
Micura have also utilized linkers containing repeating
units of both tris- or tetrakis(ethylene glycol) phosphate
and 3-hydroxy-propane-1-phosphate to generate non-
nucleotide linkers as loop replacement of double helical
RNA.14 In addition, Letsinger and Wu have also inves-
tigated the utilization of stilbenedicarboxamide as a
synthetic replacement for the natural nucleotide (nt)
hairpin loop to create a mini-cyclic oligonucleotide con-
jugate.17,18 The stilbene moiety has been strategically
placed in-between two short pieces of oligonucleotide, to
potentially create a proper alignment of the stilbene
chromophore with the adjacent base pair, upon the
hybridization of the two complementary RNA strands.
We have been interested in the design of high binding
RNA decoys that could potentially be applied in con-
trolling the in vivo translation process. RNA decoys,
such as RNA aptamers, have traditionally been gener-
ated through the in vitro selection (SELEX) process.
For example, it was recently demonstrated that the
techniques of both chemical modification and in vitro
selection can be jointly applied to enable RNA decoys
that exhibit high binding affinities towards the Rev
peptide of HIV-1.19
In this report, we have applied the synthetic non-
nucleotide hairpin linkers in ethylene glycol (L1 and L2)
or stilbenedicarboxamide (L3) onto the duplex region of
the 27 nt A-site 16S rRNA to create synthetically mod-
ified cyclic A-site 16S rRNA constructs (Fig. 1). This
model RNA substrate is designed specifically to enable
higher aminoglycoside binding through additional sta-
bilization of the duplex at both its ends. The RNA
duplex is joined at both ends to form a dumbbell-
shaped construct with the two ends of the duplex cova-
lently linked through the same synthetic linkers. It was
generally hypothesized that, through reducing the num-
ber of nucleotides that are necessary for hairpin loop
formation in the oligonucleotide, the potential in form-
*Corresponding author. Tel.: +1-718-262-2656; fax: +1-718-262-
2652; e-mail: tok@york.cuny.edu
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00760-0

366
J. B.-H. Tok et al. / Bioorg. Med. Chem. Lett. 12 (2002) 365–370
ing multiple secondary structures could be minimized.
In addition, the synthetic linkers have been demon-
strated to exhibit nuclease-resistant properties,
improved cellular uptake properties, offer specific
manipulation of nucleic acid properties, and also reduce
the synthetic cost of the nucleic acid construct.20À23
glass (CPG) support(Glenn Research, Setrling, VA,
USA) and the derivatized linkers L1, L2 and L3 accord-
ing to previously published procedures,13 and purified
using RP-HPLC [Waters Spherisorb ODS-1 (C18) Col-
umns].24 The cyclic RNA construct B was obtained
from Integrated DNA Technologies, Inc. (Coralville,
IA, USA), and showed the appropriate oligomeric
mass-spectrum data.
Synthetic linkers in both the ethylene glycol (L1 and L2)
and stilbenedicarboxamide (L3) moieties were prepared
according to standard published procedures to render
them for standard nucleotide coupling (synthetic
approach is shown in Fig. 2).13,16 It is noted that the
polyethylene glycol linker L1 (derived from triethylene
glycol) is much shorter than linker L2 as itcontains only
three ethylene groups (total of seven atoms), whereas
the hexaethylene glycol L2 (derived from hexaethylene
glycol) contains six ethylene groups (total of 17 atoms).
The stilbenedicarboxamide (L3) was calculated to have
a very similar length to the linker L2 as itconatins 20
atoms in its linker. The synthetically modified RNA
constructs A–D were then synthesized via the standard
phosphoamidite chemistry (Millipore Expedite Oligo-
nucleotide Synthesizer) using both the controlled pore
The A-site decoding region of the 16S rRNA has been
the primary site in which the aminoglycoside antibiotics
bind.25À27 Upon binding, it interferes with the protein
synthesis process through the induction of codon mis-
reading. The A-site 16S rRNA region was chosen as the
subject of study as it had been well demonstrated to be
able to be truncated to a much shortened 27 nt RNA
construct, without any major resulting effect on the
binding properties of the aminoglycoside anti-
biotics.28,29 Briefly, aminoglycoside antibiotics (Fig. 3)
are among the few known existing classes of natural
RNA-binding molecules. Aminoglycosides are low
molecular weight molecules capable of interacting with
a wide range of RNAs,30À32 also known to be able to
Figure 1. Secondary structures of the linear and cyclic A-site 16S rRNA constructs utilized in this study.

J. B.-H. Tok et al. / Bioorg. Med. Chem. Lett. 12 (2002) 365–370
367
inhibit several catalytic RNAs (e.g., self-splicing group I
introns, RNase P and small ribozymes), and are also
very powerful inhibitors of prokaryotic cell growth.30À32
rRNA constructs in a 1:1 stoichiometry. The binding
studies were performed in an incubation buffer con-
taining 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM
CaCl2, and 20 mM HEPES (pH 7.40, room tempera-
ture). The fluorescence intensity curve was observed to
increase when a corresponding increasing concentration
of CRP was titrated with 5 nM of the 27 nt RNA con-
struct A. The fluorescence intensity curve reached
saturation at around 500 nM (Fig. 4), and upon appli-
cation of a previously described curve-fitting equation,33
the Kd of CRP can be calculated to be 0.33Æ0.02 mM, a
value which is in agreement with previously reported
observations. Interestingly, when the same experiment
was repeated with the various synthetically modified
RNA constructs B–E, a comparable binding curve was
observed in each of the three RNA constructs, in which
saturation was reached around 480–520 nM (Fig. 4;
only saturation curves for constructs B and E are
shown). These results strongly suggest that the stoi-
chiometry of binding for RNA constructs B–E is similar
with that of the 27 nt RNA construct A. Applying the
curve-fitting equation, the Kd of the interaction between
CRP and RNA constructs B–E can be calculated to be
0.31Æ0.02, 3.89Æ0.19, 0.34Æ0.03 and 0.081Æ0.005
mM, respectively. First, the data indicate that creating a
cyclic RNA construct B, through introducing a similar
To quantify the Kd between the synthetically modified
RNAs to the aminoglycoside antibiotics, a recently
developed fluorescence method that enables direct and
quantitative binding measurements between aminogly-
coside–RNA interactions was employed.33 Itshould be
noted that the Kd values obtained for the RNA–ami-
noglycoside interaction via the fluorescence method are
generally a few fold higher than values obtained via the
Surface Plasmon Resonance (SPR) approach, first
demonstrated by Wong and coworkers.34,35 To enable the
fluorescence methodology, the fluorescently conjugated
paromomycin aminoglycoside was firstemployed. Par-
omomycin possesses a primary hydroxy group and was
easily derivatized through conjugation with the rhod-
amine moiety to afford rhodamine-conjugated par-
omomycin (CRP; Fig. 3).33 CRP was subsequently utilized
in increasing concentration to bind the RNA constructs of
interest, and its changes in fluorescence intensity measured
(excitation at 550 nm, monitored at 580 nm).
In this experiment, it was hypothesized that CRP binds
to the 27 nt and all the synthetically modified A-site 16S
Figure 2. Synthetic schemes for the [A] ethylene glycol linker-derivatized phosphoamidite linkers L1 and L2, and [B] stilbenedicarboxamide linker-
derivatized phosphoamidite L3.13,16

368
J. B.-H. Tok et al. / Bioorg. Med. Chem. Lett. 12 (2002) 365–370
Figure 3. Structures of the various aminoglycoside antibiotics utilized in this study.
Figure 4. Representative fluorescence anisotropy plots of CRP (10 nM) as a function of concentration of [A] 27 nt A-site 16S rRNA construct A, [B]
cyclized A-site 16S rRNA construct B, [C] cyclized A-site 16S rRNA construct E.

J. B.-H. Tok et al. / Bioorg. Med. Chem. Lett. 12 (2002) 365–370
369
UUCG loop region onto the other end of the A-site
RNA duplex region, does notexhibitany significant
improvementin hte Kd binding affinity toward CRP.
Second, the usage of the short triethylene glycol linker
(L1) to replace the two UUCG loop-region of the A-site
construct to generate the cyclic RNA construct C dras-
tically decreases the binding affinity for CRP by nearly
12-fold, from 0.33 to 3.89 mM. This observation strongly
suggests that linker L1 does notprovide sufficientlength
to enable the structural flexibility required by the RNA
oligomers to adopt the optimized conformation for
CRP aminoglycoside binding. Itwas previously pre-
dicted that the short linker L1 would potentially lead to
distortion of the terminal heterocyclic bases, and thus
be unable to effectively maintain optimum Watson–Crick
hydrogen bonding formation.13 However, by increasing
the length of the polyethylene glycol linker chain length
(to L2) restores the binding affinity back to approxi-
mately the value exhibited by the wt 27 ntA-siet 16S
rRNA construct.
CRP and 400 nM of RNA construct A. The fluores-
cence of the complex was observed to quenched gradu-
ally in a saturable fashion, and using the previously
described curve-fitting equation, the Kd of neomycin B
can be calculated. The Kd value was calculated to be
0.41Æ0.03 mM, a value which is in agreement with the
previous observations utilizing the same fluorescence
techniques.33,36
Using the same experimental procedures and condi-
tions, neomycin B was individually titrated against the
various complexes of CRP (10 nM) and RNA con-
structs B–E (400 nM). Extrapolating the observed
values in each of the four complexes afforded us Kd
values of 0.38Æ0.03, 4.34Æ0.24, 0.47Æ0.03 and
0.091Æ0.008 mM for neomycin against RNA constructs
B–E, respectively (data summarized in Table 1B). The
results obtained again reconfirmed the previous obser-
vations that: (i) cyclizing the A-site 16S rRNA construct
A to afford construct B does notexhibitany significant
improvement in binding affinities towards the neomycin
B aminoglycoside; and (ii) cyclic RNA constructs C and
D, which comprised of both short and long ethylene
glycol linkers, do not enhance the stability of the
resulting RNA constructs as demonstrated by the Kd
binding values towards neomycin B. Instead, RNA
construct C, comprised of short triethylene glycol linker
L1, again demonstrated an overall 11-fold decrease in its
Kd values towards neomycin B when compared with
construct A. Again, the obtained Kd data for construct
D showed a comparable Kd binding value when com-
pared to the 27 nt 16S rRNA construct A. Likewise, an
enhancementin Kd values is observed only with the
cyclic RNA construct E. Taken together, these results
strongly suggest that the stilbenedicarboxamide linker
affords a much stabler resulting RNA construct when
compared to either the natural nucleotides or the poly-
ethylene glycol linkers L2 and L3, thus translating into a
much tighter binding with the neomycin B aminoglyco-
side.
Interestingly, the only visible improvement in the Kd
binding affinity of CRP is observed towards cyclic RNA
construct E, which consists of stilbenedicarboxamide
linkers at both ends of the duplex RNA. The measured
Kd of 0.081 mM of CRP interaction with cyclic RNA
construct E is approximately 4-fold higher compared to
that of CRP–RNA construct A interaction, for which
the Kd is measured to be 0.33 mM (data summarized in
Table 1A).
To further examine the binding characteristics of other
aminoglycosides towards the cyclized RNA constructs
B–E, a competition experiment was performed. Again,
it was hypothesized that 1:1 stoichiometry binding exists
between the various examined aminoglycosides (namely
neomycin B, paromomycin, tobramycin, kanamycin B
and streptomycin) with the RNA constructs A–E. In t he
competition experiment, we are interested in monitoring
the changes in fluorescence anisotropy values through
the displacement of RNA-bound CRP molecules when
an increasing concentration of the various amino-
glycoside ligands was titrated to the CRP–RNA
complex.
The same experimental approach was subsequently
applied to the remaining aminoglycosides in par-
omomycin, tobramycin, kanamycin B and streptomycin.
Their binding affinities were examined towards RNA
constructs B–D. The Kd values obtained for these ami-
noglycosides all showed a similar trend in their affinities
towards the various cyclic RNA constructs studied, with
neomycin B being the tightest, while streptomycin
To ensure that this method of approach is appropriate
for this study, an increasing concentration of the well-
studied neomycin B (0–50 mM) was added to the CRP
construct A complex first, which comprised of 10 nM of
Table 1. (A) Summary of the Kd (mM) values of the CRP tracer molecule against RNA constructs A–E; (B) summary of Kd (mM) of the various
tested aminoglycoside molecules against RNA constructs A–E
Molecule
RNA Construct A
0.33Æ0.02
RNA Construct B
0.31Æ0.02
RNA Construct C
3.89Æ0.19
RNA Construct D
0.34Æ0.03
RNA Construct E
0.081Æ0.005
(A)
CRP
(B)
Neomycin B
Paromomycin
Tobramycin
Kanamycin B
Streptomycin
0.41Æ0.03
1.64Æ0.12
2.07Æ0.24
2.95Æ0.30
n.b.
0.38Æ0.03
1.58Æ0.11
2.04Æ0.22
2.90Æ0.18
n.b.
4.34Æ0.24
5.33Æ0.37
n.b.a
0.47Æ0.03
1.70Æ0.09
2.12Æ0.23
2.75Æ0.25
n.b.
0.091Æ0.008
0.40Æ0.03
0.68Æ0.07
0.91Æ0.10
n.b.
n.b.
n.b.
an.b., no observable binding.

370
J. B.-H. Tok et al. / Bioorg. Med. Chem. Lett. 12 (2002) 365–370
showed no visible binding properties (data summarized
in Table 1B). Importantly, among all the cyclic RNA
constructs studied, only the cyclic RNA construct E,
which contained the stilbenedicarboxamide linker at
both ends, again demonstrated the tightest overall
binding characteristics towards the aminoglycosides
(data summarized in Table 1).
6. Wagner, R. W.; Matteucci, M. D.; Lewis, J. G.; Gutierrez,
A. J.; Moulds, C.; Froehler, B. C. Science 1993, 260, 1510.
7. Wagner, R. W. Nature 1994, 372, 333.
8. Wagner, R. W.; Matteucci, M. D.; Grant, D.; Huang, T.;
Froehler, B. C. Nature Biotech. 1996, 14, 840.
9. Stein, C. A. Chem. Biol. 1996, 3, 319.
10. Yamakawa, H.; Abe, T.; Saito, T.; Takai, K.; Yamamoto,
N.; Takaku, H. Bioorg. Med. Chem. 1998, 6, 1025.
11. Riordan, M. L.; Martin, J. C. Nature 1991, 350, 442.
12. Ma, M. Y.-X.; McCallum, K.; Climie, S. C.; Kuperman,
R.; Lin, W. C.; Sumner-Smith, M.; Barnett, R. W. Nucleic
Acid Res. 1993, 21, 2585.
13. Ma, M. Y.-X.; Reid, L. S.; Climie, S. C.; Lin, W. C.;
Kuperman, R.; Sumner-Smith, M.; Barnett, R. W. Biochem-
istry 1993, 32, 1751.
14. Pils, W.; Micura, R. Nucleic Acid Res. 2000, 28, 1859.
15. Letsinger, R. L.; Wu, T. J. Am. Chem. Soc. 1994, 116, 811.
16. Letsinger, R. L.; Wu, T. J. Am. Chem. Soc. 1995, 117,
7323.
17. Herrlein, M. K.; Nelson, J. S.; Letsinger, R. L. J. Am.
Chem. Soc. 1995, 117, 10151.
18. Herrlein, M. K.; Letsinger, R. L. Angew. Chem., Int. Ed.
Engl. 1997, 36, 599.
19. Nelson, J. S.; Giver, L.; Ellington, A. D.; Letsinger, R. L.
Biochemistry 1996, 35, 5339.
20. Bensenler, F.; Fu, D.; Ludwig, J.; McLaughlin, L. W. J.
Am. Chem. Soc. 1993, 115, 8483.
21. Wiederholt, K.; McLaughlin, L. W. Nucleic Acid Res.
1997, 27, 2487.
22. Wu, X.; Pitsch, S. Nucleic Acid Res. 1998, 21, 4315.
23. Steiz, O. Angew. Chem., Int. Ed. Engl. 1999, 111, 3466.
24. Coupling yields in using linkers L1–L3 to yield RNA con-
structs C, D and E were satisfactory (93, 90, 91%, respec-
tively), in which the oligomers were isolated as trityl-on
derivatives by RP-HPLC, and detritylated with 80% acetic
acid. The concentrations of the oligomers were quantified by
OD260 measurements.
25. Woodcock, J.; Moazed, D.; Cannon, M.; Davies, J.; Nol-
ler, H. F. The EMBO J. 1991, 10, 3099.
26. von Ahsen, U.; Noller, H. F. Science 1995, 267, 234.
27. von Ashen, U.; Davies, J.; Schroeder, R. J. Mol. Biol.
1992, 226, 935.
28. Fourmy, D.; Recht, M. I.; Blanchard, S. C.; Puglisi, J. D.
Science 1996, 274, 1367.
29. Fourmy, D.; Yoshizawa, S.; Puglisi, J. D. J. Mol. Biol.
1998, 277, 333.
In summary, the data presented indicate that cyclic
RNA comprising of stilbenedicarboxamide linker at
both ends to create a dumbbell-shaped construct is
approximately four-fold more effective in the general
binding of aminoglycosides when compared to its hair-
pin-shaped RNA counterpart. The binding character-
istics of rhodamine-derivatized paromomycin tracer
(CRP) ligand are used to bind the various cyclic RNA
constructs, and its complex is subsequently utilized to
afford Kd values of five more other aminoglycosides in
paromomycin, neomycin B, tobramycin, kanamycin B
and streptomycin. The data obtained are consistent with
previous observations,33 which also have indicated neo-
mycin B demonstrated the tightest overall binding affi-
nities to the A-site 16S rRNA molecular targets. The
application of the stilbenedicarboxamide linker to
replace the UUCG RNA loop in the A-site 16S rRNA
construct had been shown to be both useful and feasible
in further stabilizing the duplex region. It is hypothe-
sized that the extra stabilization conferred to the overall
RNA construct is through the p–p stacking interactions
between the stilbenedicarboxamide moiety and the
adjacent base pairs of the nucleotides at the end of the
RNA duplex region.16,37,38 Through molecular modeling
(MM2 minimization using Hyperchem V5.01a), it was
indeed demonstrated by Lewis and coworkers that the
stilbenedicarboxamide moiety does adopt a coplanar
structure with the adjacent guanine and cytosine base
pairs.37 Alht ough a modestfour-fold enhancementin Kd
binding affinity was obtained, nevertheless these results
have generated a possible approach towards our effort in
designing and synthesizing RNA decoys that are capable
of better regulating gene expression at the translation level.
30. Sucheck, S. J.; Wong, C.-H. Curr. Opin. Chem. Biol. 2000,
4, 678.
31. Schroeder, R.; Waldsich, C.; Wank, H. The EMBO J.
2000, 19, 1.
32. Tor, Y.; Hermann, T.; Westhof, E. Chem. Biol. 1998, 5,
R277.
33. Wang, Y.; Hamasaki, K.; Rando, R. R. Biochemistry
1997, 36, 768.
Acknowledgements
J. B.-H. Tok gratefully acknowledges York College, the
McNair Scholars Program and RF-CUNY for a Sum-
mer ’01 Faculty Award. W. Wong and N. Baboolal are
both McNair undergraduate scholars.
34. Hendrix, M.; Priestley, E. S.; Joyce, G. F.; Wong, C.-H. J.
Am. Chem. Soc. 1997, 119, 3641.
35. Park, W. K. C.; Aner, M.; Jaksche, H.; Wong, C.-H. J.
Am. Chem. Soc. 1996, 118, 10150.
References and Notes
1. Agrawal, S. Antisense Therapeutics; Humana Press:
Totowa, NJ, 1996.
36. Tok, J. B.-H.; Huffman, G. R. Bioorg. Med. Chem. Lett.
2000, 10, 1593.
2. Crooke, S. T.; LeBleu, B. Antisense Research and Applica-
tions; CRC Press: Boca Raton, FL, 1993.
3. Wickstrom, E. Prospects for Antisense Nucleic Acid Ther-
apy of Cancer and AIDS; Wiley-Liss: New York, 1991.
4. Nielsen, P. E. Curr. Opin. Mol. Ther. 2000, 2, 282.
5. Micklefield, J. Curr. Med. Chem. 2001, 8, 1157.
37. Lewis, F. D.; Wu, T.; Liu, X.; Letsinger, R. L.; Greenfield,
S. R.; Miller, S. E.; Wasielewski, M. R. J. Am. Chem. Soc.
2000, 122, 2889.
38. Lewis, F. D.; Liu, X.; Wu, Y.; Miller, S. E.; Wasielewski,
M. R.; Letsinger, R. L.; Sanishvili, R.; Joachimiak, A.; Ter-
eshko, V.; Elgi, M. J. Am. Chem. Soc. 1999, 121, 9905.