Targeting RNAs with tobramycin analogues

Article abstract of DOI:10.1002/anie.200460558

(Chemical equation presented). A library of 4,6-linked analogues of the drug tobramycin with various mono- or diaminosugars attached to the 6-position of the two-ring core has been prepared (see scheme; Bn = benzyl, Tol = p-tolyl). The compounds were scre

Full text of DOI:10.1002/anie.200460558

Communications
RNA Binding
Targeting RNAs with Tobramycin Analogues**
Fu-Sen Liang, Sheng-Kai Wang, Takuji Nakatani, and
Chi-Huey Wong*
There are growing interests in developing small molecules to
specifically target RNA because of the potential therapeutic
applications.[1] Aminoglycosides and macrolides, for example,
have been known to exhibit antibiotic activities by interacting
with the bacterial ribosomal RNA.[2] Recent discoveries of
other small molecules that control gene expression in living
cells by attenuating RNA activities have shed more light on
the promising potential of developing RNA-binding mole-
cules as drugs.[3] Many aminoglycosides have been developed
to target not only the bacterial ribosomal RNA 16S A-site,
but also other RNA sequences, including the regulatory
domains of HIV-1 mRNA, the oncogenic Bcr-Abl mRNA
sequence, and the group I intron.[4]
From NMR and X-ray crystallographic studies,[2,5] it is
clear that the two-ring cores (rings I and II) of both
tobramycin (1) and paromomycin (2), which are 4,5- and
4,6-linked aminoglycosides, respectively (see Figure 1), sit in
the bulges of A1408, A1492, and A1493 of the A-site and make
very similar contacts with the RNA bases and the phosphate
backbones. The surface plasmon resonance (SPR) binding
studies of the naturally occurring aminoglycosides with the
wild-type or mutant 16S A-site RNAs show that the binding
affinity and specificity vary when the compositions or the
linking positions of the additional sugar moieties change.[6]
Previous studies also suggest that both neamine and nebr-
amine are basic cores for binding to various RNA sequences
and for cell permeability.[7] It is possible that by keeping the
Figure 1. The structures of 4,6- and 4,5-linked aminoglycosides.
[*] F.-S. Liang, S.-K. Wang, Dr. T. Nakatani, Prof. C.-H. Wong
Department of Chemistry and
The Skaggs Institute of Chemical Biology
The Scripps Research Institute
10550 N. Torrey Pines Road, La Jolla, CA 92037 (USA)
Fax: (+1)858-784-2409
E-mail: wong@scripps.edu
[**] Supported by the NIH.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
6496 ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200460558
Angew. Chem. Int. Ed. 2004, 43, 6496 –6500

Angewandte
Chemie
two-ring core and changing the carbohydrate unit
attached to position 5 or 6 of the core may affect
both affinity and specificity.
Most of the aminoglycoside analogues devel-
oped to date have been modified by the attachment
of various nonsugar moieties to the original amino-
glycosides or to the two-ring core through different
linkers. Here, we designed and synthesized a new
library of 4,6-linked tobramycin analogues with
various mono- or diaminosugars attached to the 6-
position of the deoxystreptamine ring (Figure 2).
With the relatively rigid conformation of the carbo-
hydrate framework, it was hoped that new amino-
glycosides with higher binding affinities and selec-
tivities would be found.
The protected nebramine core 5 was derived
from tobramycin (1; Scheme 1). The amine groups
of tobramycin (1) were first converted into azides as
protecting groups by the diazo transfer reaction.[8]
This was followed by benzylation of the alcohol
groups to give the fully protected tobramycin
derivative 6. Cleavage of the glycosidic bond
between the nebramine core and the third ring was
catalyzed by a Lewis acid in the presence of p-
thiocresol as nucleophile.[9] The advantage of this
method, instead of HCl or copper chloride catalyzed
cleavage, is that the resulting cleaved third ring can
Figure 3. Thioglycoside building blocks as a) the first donors and b) the second donors.
Tol =p-tolyl, Troc=trichloroethyloxycarbonyl (CCl3CH2OC(O)-), Bz=benzoyl.
be recovered as a thioglycoside building block
and used in library synthesis.
A
group of protected monosaccharide
building blocks, which contained one or more
amine groups, was then selected and synthe-
sized from our thioglycoside building block
database (Figure 3). Two sets of building blocks
were chosen with significant differences in
reactivity as glycosylation donors.[10] The first
set, which includes compounds 7–17, contains
thioglycoside donors with high reactivity which
were designed by the introduction of electron-
Figure 2. Synthetic strategy of tobramycin analogues.
donating protecting groups (Figure 3a). The
second set, which includes compounds 18–29, contains
thioglycoside donors with lower reactivity, tuned by elec-
tron-withdrawing protecting groups (Figure 3b). Thio-
glycosides 7–17 were either directly coupled to the protected
nebramine core 5 or were used to assemble disaccharide
building blocks with thioglycosides 18–29 by N-iodosuccini-
mide (NIS)–TfOH promoted glycosylation (TfOH =
trifluorosulfonic acid).[11] After several unsuccessful attempts
under different glycosylation conditions, these disaccharide
building blocks were found to be relatively inactive. The
protecting groups were then changed to benzyl groups to
provide higher reactivity toward the nebramine core 5
(Scheme 2). Benzenesulfinylpiperidine (BSP) and Tf2O
were used as promoters for these glycosylation reactions
(Scheme 3).[12] Stereoisomers at the anomeric positions (a or
b linkage) were usually formed in these glycosylation
reactions and these were separated by column chromatog-
Scheme 1. Synthesis of the protected nebramine core: a) TfN3, ZnCl2
(cat.), NEt3, CH2Cl2, H2O, MeOH; b) NaH, BnBr, TBAI, DMF, 82%
(two steps); c) p-TolSH, BF3·Et2O, CH2Cl2, 5 43%, 7 47%. Tf=triflate
(CF3SO2), Bn=benzyl, TBAI=tetra-n-butylammonium iodide,
DMF=N,N-dimethylformamide.
Angew. Chem. Int. Ed. 2004, 43, 6496 –6500
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6497

Communications
macrolides were first screened at 1 mm against the
above RNA molecules. By using tobramycin (1) and
neomycin B (3) as standards, compounds with high
binding signals were selected for the determination of
dissociation constants (Kd).[17] The Kd values of
selected compounds from each RNA screening were
determined for all RNA sequences in this study to
examine the binding specificity (Table 1).
From the results of the binding assay, several
molecules displayed an affinity in the nanomolar range
to specific RNA sequences. Whereas all of the selected
compounds showed some limited affinity toward the
RNAs tested, several compounds exhibited higher
affinity to specific RNAs. Tobramycin (1) showed a
higher affinity to the HIV frameshift signal (0.64 mm,
2–100-fold selectivity); neomycin B (3) was more
specific to the bacterial 16S A-site and the human
tyrosine sulfotransferase mRNA (0.2 mm and 0.3 mm,
8–14-fold); sisomycin (4) was selective for the HCV
IRES IIId domain (0.26 mm, 5 to > 100-fold); 33
displayed a relatively high affinity to the HCV IRES
IIId (0.7 mm, 1.6–5-fold); 34 also showed a high affinity
to the HCV IRES IIId (0.25 mm, 1.7–38-fold); 35 was
selective for the HIV frameshift signal (0.32 mm, 5.6–
Scheme 2. Representative synthesis of disaccharide building blocks. a) NIS, TfOH,
CH2Cl2, mol. sieves (4 Š), À458C; b) NaOMe/MeOH; c) NaH, BnBr, TBAI, DMF, 54%
(three steps); d) NIS, TfOH, CH2Cl2, mol. sieves (4 Š), À458C; e) AcOH, H2O, 808C;
f) NaH, BnBr, TBAI, DMF, 41% (three steps). NIS=N-iodosuccinimide.
Scheme 3. Synthesis of tobramycin analogues. a) BSP, Tf2O, CH2Cl2, mol.
sieves (4 Š), À458C, 32–78%; b) Raney Ni, N2H4, EtOH, dioxane; c) HCl
(0.1n), H2, Pd(OH)2 (20%), MeOH/H2O, 19–65% (two steps).
BSP=benzenesulfinylpiperidine.
raphy (silica gel). The number of different protecting
groups used was kept to a minimum, with the use of only
azido and benzyl groups to simplify the deprotection
steps afterward.
Deprotection was completed by first, the reduction
of the azido groups with Raney Nickel and anhydrous
hydrazine and second, palladium hydroxide catalyzed
hydrogenation under acidic conditions to remove the
benzyl groups (Scheme 3), to give the tobramycin
analogues 32–92 (Figure 4 and Supporting Informa-
tion). An attempt to remove both groups in one step by
hydrogenation in the presence of different palladium
catalysts only gave mixtures of incompletely depro-
tected products probably owing to the poisoning of the
catalysts by amines.[13]
The SPR assay was then used to study the binding
affinity and specificity of these tobramycin analogues
with a number of short RNA sequences (24 to 48 bps),
which were identified from sequence-conserved and
functionally important regions of several disease-
related bacterial, viral, or human RNAs, such as the
bacterial ribosomal 16S A-site, E. coli. transglycosidase
mRNA, Hepatitis C virus (HCV) internal ribosome
entry site (IRES) RNAs,[14] HIV frameshift signal,[15]
HIV protease mRNA, human oncogenic Bcr-Abl
mRNA,[16] and human tyrosine sulfotransferase
Figure 4. Selected tobramycin analogues (with R groups as shown) which
were active in the SPR studies (structures of other synthesized analogues can
be found in the Supporting Information).
mRNA (Figure 5). The synthetic compound library as
well as commercially available aminoglycosides and
6498
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 6496 –6500

Angewandte
Chemie
monosaccharide or disaccharide moiety to form analogues of
tobramycin. The SPR studies of the interactions of these
analogues with certain RNA sequences show that new
aminoglycosides can be created to target RNA in
a
sequence-selective manner. Workis in progress to further
study the interactions of selected compounds and RNAs to
understand the origin of selectivity and to design better RNA-
binding molecules as inhibitors of translational processes in
cell-based systems.
Received: May 5, 2004
Revised: June 15, 2004
Keywords: carbohydrates · drug design · glycosylation ·
.
RNA recognition · surface plasmon resonance
[1] a) D. Ecker, R. H. Griffey, Drug Discovery Today 1999, 4, 420;
b) S. J. Sucheck, C.-H. Wong, Curr. Opin. Chem. Biol. 2000, 4,
678; c) T. Hermann, Angew. Chem. 2000, 112, 1962; Angew.
Chem. Int. Ed. 2000, 39, 1890; d) J. Gallego, G. Varani, Acc.
Chem. Res. 2001, 34, 836.
[2] a) N. Ban, P. Nissen, J. Hansen, P. B. Moore, T. A. Steitz, Science
2000, 289, 905; b) D. Moazed, H. F. Noller, Nature 1987, 327, 389;
c) F. Schlünzen, R. Zarivach, J. Harms, A. Bashan, A. Tocilj, R.
Albrecht, A. Yonath, F. Franceschi, Nature 2001, 413, 814; d) D.
Fourmy, M. I. Recht, S. C. Blan-
Figure 5. RNA sequences used for the SPR binding assay. All RNA
sequences were 5’-biotin-labeled and were heated to 808C for 2 min
and allowed to cool down slowly to refold into the most stable confor-
mations before immobilization onto a streptavidin-coated sensor chip.
SPR experiments were performed according to the procedure reported
previously.[17]
Table 1: Kd values [mm] for selected compounds screened against different RNAs.[a,b]
chard, J. D. Puglisi, Science 1996,
274, 1367; e) Q. Vicens, E. Westhof,
ChemBioChem 2003, 4, 1018; f) Y.
Tor, ChemBioChem 2003, 4, 998.
[3] a) M. Howard, R. A. Frizzell, D. M.
Bedwell, Nat. Med. 1996, 2, 467;
b) G. Werstuck, M. R. Green, Sci-
ence 1998, 282, 296; c) W. Winkler,
A. Nahvi, R. R. Breaker, Nature
2002, 419, 952.
[4] a) M. L. Zapp, S. Stern, M. R.
Green, Cell 1993, 74, 969; b) S.
Wang, P. W. Huber, M. Cui, A. W.
Czarnik, H.-Y Mei, Biochemistry
1998, 37, 5549; c) S. J. Sucheck,
W. A. Greenberg, T. J. Tolbert, C.-
H. Wong, Angew. Chem. 2000, 112,
1122; Angew. Chem. Int. Ed. 2000,
39, 1080; d) U. von Ahsen, J.
Davies, R. Schroeder, Nature
1991, 353, 368.
Compound
16S-
AS
EcTG
HCV2b
HCV3d
HIV-
FSS
HIV-
PAS
hBcr-
Abl
hTSulT
1
3
4
2.1
0.2
1.3
5.9
1.7
4.3
3
1.8
3.8
6.6
2.6
4.6
3.6
2.3
3.7
19
2.8
12
2.7
2.9
4.3
12
2.2
2.9
0.26
6.8
0.7
0.25
2.7
–
2.3
2.3
4
3.3
0.53
2
0.64
1.6
2.7
4.7
1.2
3.4
0.32
1
0.71
4.6
1.7
2
14
2
1.3
2
10
3.2
3.5
0.75
3.8
3.2
0.7
5
65
0.3
37
15
19
1.1
0.83
6.8
5.1
3
5
4
14
1.2
7
32
33
34
35
36
37
38
39
40
41
42
43
11
25
1.8
0.62
4.9
5.3
14
1
0.42
8
3
1.6
9.4
8.1
–
52
14
9.6
13
9.4
3.9
5.1
0.42
1.8
9.4
0.28
12
2.6
3.8
3.5
3.1
9.5
4
8.5
6.4
3.5
14
0.67
2.5
1.7
7.6
1.6
4.5
1
6
[a] RNA abbreviations used: 16S-AS: Bacterial 16S A-site; EcTG: E. coli. transglycosidase mRNA;
HCV2b: HCV IRES domain IIb; HCV3d: HCV IRES domain IIId; HIV-FSS: HIV frameshift signal; HIV-
PAS: HIV protease active site mRNA; Bcr-Abl: human oncogenic Bcr-Abl mRNA; hTSulT: human
tyrosine sulfotransferase mRNA; [b] Errors of Kd values range from Æ5% to Æ29%.
[5] a) D. Fourmy, S. M. I. Recht, J. D.
Puglisi, J. Mol. Biol. 1998, 277, 347;
b) S. Yoshizawa, D. Fourmy, J. D.
Puglisi, EMBO J. 1998, 17, 6437;
c) A. P. Carter, W. M. J. Clemons,
D. E. Brodersen, R. J. Morgan-Warren, B. T. Wimberly, V.
Ramakrishnan, Nature 2000, 407, 340; d) Q. Vicens, E. Westhof,
Structure 2001, 9, 647; e) Q. Vicens, E. Westhof, Chem. Biol.
2002, 9, 747.
25-fold); 37 bound better to the HCV IRES IIb domain
(0.28 mm, 2.5 to > 100-fold); 39 selectively bound the E. coli.
transglycosidase mRNA (0.42 mm, 4–23-fold); 41 exhibited a
high affinity to the HCV IRES IIId domain and the E. coli.
transglycosidase mRNA (0.53 mm and 0.67 mm, 2–16-fold).
The charges of the aminoglycosides tested range from + 5 to
+ 7. However, upon comparing the Kd values and the number
of charges, no direct correlations were found.
[6] C.-H. Wong, M. Hendrix, E. S. Priestley, W. A. Greenberg,
Chem. Biol. 1998, 5, 397.
[7] a) S. R. Kirk, N. W. Luedtke, Y. Tor, J. Am. Chem. Soc. 2000, 122,
980; b) S. J. Sucheck, A. L. Wong, K. M. Koeller, D. D. Boehr,
K.-a. Draker, P. S. Sears, G. D. Wright, C.-H. Wong, J. Am.
Chem. Soc. 2000, 122, 5230; c) J. Haddad, L. P. Kotra, B. Llano-
Sotelo, C. Kim, E. F. Azucena, Jr., M. Liu, S. B. Vakulenko, C. S.
Chow, S. Mobashery, J. Am. Chem. Soc. 2002, 124, 3229; d) S.
In conclusion, we have developed an efficient method to
replace a monosaccharide unit of tobramycin with another
Angew. Chem. Int. Ed. 2004, 43, 6496 –6500
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 6499

Communications
Hanessian, M. Tremblay, E. E. Swayze, Tetrahedron 2003, 59,
983.
[8] P. T. Nyffeler, C.-H. Liang, K. M. Koeller, C.-H. Wong, J. Am.
Chem. Soc. 2002, 124, 10773.
[9] B. Wu, J. Yang, Y. He, E. E. Swayze, Org. Lett. 2002, 4, 3455.
[10] T. K. Ritter, K.-K. T. Mong, H. Liu, T. Nakatani, C.-H. Wong,
Angew. Chem. 2003, 115, 4805; Angew. Chem. Int. Ed. 2003, 42,
4657.
[11] a) P. Konradsson, U. E. Udodong, B. Fraser-Reid, Tetrahedron
Lett. 1990, 31, 4313; b) Z. Zhang, I. R. Ollmann, X.-S. Ye, R.
Wischnat, T. Baasov, C.-H. Wong, J. Am. Chem. Soc. 1999, 121,
734.
[12] a) D. Crich, M. Smith, J. Am. Chem. Soc. 2001, 123, 9015; b) K.-
K. T. Mong, H.-K. Lee, S. G. Durꢀn, C.-H. Wong, Proc. Natl.
Acad. Sci. USA 2003, 100, 797.
[13] B. P. Czech, R. A. Bartsch, J. Org. Chem. 1984, 49, 4076.
[14] a) P. J. Lukavsky, G. A. Otto, A. M. Lancaster, P. Sarnow, J. D.
Puglisi, Nat. Struct. Biol. 2000, 7, 1105; b) C. M. T. Spahn, J. S.
Kieft, R. A. Grassucci, P. A. Penczek, K. Zhou, J. A. Doudna, J.
Frank, Science 2001, 291, 1959.
[15] T. Jacks, M. D. Power, F. R. Masiarz, P. A. Luciw, P. J. Barr, H. E.
Varmus, Nature 1988, 331, 280.
[16] E. Shtivelman, B. Lifshitz, R. P. Gale, E. Canaani, Nature 1985,
315, 550.
[17] C.-H. Wong, F.-S. Liang, Methods Enzymol. 2003, 362, 340.
6500
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 6496 –6500