Design and synthesis of paromomycin-related heterocycle-substituted aminoglycoside mimetics based on a mass spectrometry RNA-binding assay

Article abstract of DOI:10.1002/anie.200351354

A good starting point for the design of potent bactericidal compounds is the aminoglycoside paromomycin (1). The A-ring (shown in red) was replaced with a range of heterocycles. An ESIMS-based RNA-binding assay was used to screen the binding affinities of

Full text of DOI:10.1002/anie.200351354

Angewandte
Chemie
measurements of the ratio of free and bound RNA target.
ESIMS was also used to determine the location of ligand
binding on the RNA and to determine the specificity for 16S
and 18S rRNA.[4,5] Paromomycin exhibited an excellent bind-
ing affinity and specificity for 16S rRNA in this assay.[6]
Therefore, we used MS to carry out structure–activity
relationship (SAR) studies to determine how the structure
of the aminoglycosides affects their binding activities. Based
on these results, paromomycin mimetics were designed and
synthesized as new bactericidal compounds.
It has been estimated that over a half of all therapeutic
agents consist of heterocyclic compounds. The heterocyclic
ring system in many cases comprises the very core of the
active moiety or pharmacophore. Therefore, heterocycle-
substituted aminoglycoside mimetics may be ideal therapeu-
tic agents.
To determine which rings of paromomycin are important
in its 16S rRNA-binding activity, paromomycin (1), 5-(b-d-
neobisamine)-2-deoxystreptamine (2), and 6’-hydroxyribos-
tamycin (3) were used as standards[7] to compare their binding
affinities in an MS-based RNA-binding assay. In the ESIMS
RNA-binding assay, compound 2 exhibited a better binding
activity than 3, which indicated that the D ring of paromo-
mycin may be more important than its A ring for the
16S rRNA-binding activity and specificity.[8] Based on this
observation, paromomycin derivatives in which the A ring
was replaced with a range of heterocycles were chosen as
synthetic targets (Scheme 1).
Owing to the complex nature of the target compounds, it
is impractical to prepare them individually for biological
activity screening.[9] However, by using MS to study the SAR,
we could first screen the 4-heterocycle-substituted 2-deoxy-
streptamine derivatives, determine which motif has better
binding properties, and then design and prepare the target
Aminoglycoside Mimetics
Design and Synthesis of Paromomycin-Related
Heterocycle-Substituted Aminoglycoside
Mimetics Based on a Mass Spectrometry RNA-
Binding Assay**
Yili Ding,* Steven A. Hofstadler, Eric E. Swayze,
Lisa Risen, and Richard H. Griffey
The family of aminoglycosides, which includes neomycin,
paromomycin (1), lividomycin, kanamycin, and gentamicin, is
a very potent group of bactericidal compounds that bind to
the RNA of the small ribosomal subunit.[1] This bactericidal
action is mediated by binding of the compound to the
bacterial RNA in a way that leads to misreading of the genetic
code.[2] The decoding region of the 16S ribosomal RNA
contains a smaller subdomain (16S) that folds and retains the
key structural features of the full-length RNA.[3] This “A-site”
subdomain binds aminoglycosides at the same location as the
intact rRNA.
ESIMS was used to determine solution-phase dissociation
constants of RNA–ligand complexes based on gas-phase
[*]Dr. Y. Ding
Ribapharm
3300 Hyland Avenue, Costa Mesa, CA 92626 (USA)
Fax: (+1)714-641-7222
E-mail: yding@icnpharm.com
Dr. S. A. Hofstadler, Dr. E. E. Swayze, L. Risen, Dr. R. H. Griffey
Ibis Therapeutics, Isis Pharmaceuticals, Inc.
2292 Faraday Avenue, Carlsbad, CA 92008-7208 (USA)
[**]We thank Mr. Patrick Wheeler for NMR spectroscopic assistance.
We are grateful to the DARPA for support of this research work
through grant N65236-99-1-5419.
Scheme 1. Structures of paromomycin (1), 5-(b-d-neobisamine)-2-deoxystrept-
amine (2), 6’-hydroxyribostamycin (3), and the 4-heterocycle-substituted
substituted paromomycin mimetics.
Angew. Chem. Int. Ed. 2003, 42, 3409 – 3412
DOI: 10.1002/anie.200351354
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3409

Communications
compounds. A concise synthetic strategy for these derivatives
compound 5. Treatment of 5 with excess SO2Cl2 in CH2Cl2
gave 6, which was used as the core starting material for the
parallel synthesis.
Compound 6 was coupled with different nucleophilic
reagents to afford the corresponding 4-heterocycle-substi-
tuted 2-deoxystreptamine derivatives 7–23 in satisfactory
is summarized in Scheme 2. The hydroxy group of 5,6-di-O-
acetylated diazido-2-deoxystreptamine (4), which was
obtained from the selective enzymatic deacetylation of tri-
O-acetyl diazido-2-deoxystreptamine,[10] was protected as an
MTM ether through a Pummerer rearrangement[11] to give
Scheme 2. Synthesis of 2-deoxystreptamine derivatives with heterocyclic substituents at C4: a) DMSO (30 equiv), Ac2O (15 equiv), AcOH
(5 equiv), 24 h, 78%; b) SO2Cl2 (6 equiv), CH2Cl2, room temperature, 30 min, 100%; c) AH–QH (1.5 equiv), NaH (1.7 equiv), CH3CN/DMF, room
temperature, 2 h; then 6 (1 equiv), 76–97%; d) 1) NaOMe (0.1 equiv), MeOH, room temperature, 4 h; 2) Me3P/THF (8 equiv), H2O/THF (1:1),
10 h, 56–75%; e) 1) benzyl 2,2,2-trichloroacetimidate (4 equiv), TfOH (2 equiv), CH2Cl2, room temperature, 5 h; 2) NaOM/MeOH (0.1m), 4 h;
3) Me3P/THF (8 equiv), H2O, room temperature, 10 h, 62%; f) 1) NaOMe/MeOH (0.5m), 10 h; 2) 2,2-dimethoxypropane, TsOH, room tempera-
ture, 10 h; 3) K2CO3 (5 equiv), MeI (5 equiv), CH3COCH3/H2O (5:1), 758C, 15 h, 23%; g) BrCH2CO2Me (2.0 equiv), NaH (2.2 equiv), CH3CN,
room temperature, 4 h, 75%; h) 1) CH3CO2H/H2O (80%), 758C, 3 h; 2) NaOH/MeOH/H2O, room temperature, 5 h; 3) 4-nitro-1,2-phenylenedi-
amine (4 equiv), HCl/H2O (4n), 958C, 4 h; 4) Me3P/THF (8 equiv), H2O/THF (1:1), 10 h, 39%; i) NaH (1.3 equiv), RBnBr (1.2 equiv), room tem-
perature, 4 h, 75–85%; j) 1) CH3CO2H/H2O (80%), 758C, 3 h; 2) Me3P/THF (8 equiv), H2O/THF, room temperature, 10 h, 55–65%. DMF=N,N-
dimethylformamide, Tf=trifluoromethanesulfonyl, Ts=p-toluenesulfonic acid.
3410
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 3409 – 3412

Angewandte
Chemie
yields.[12] Deprotection and reduction of 7–23 gave the
derivatives 24–40, respectively. Direct alkylation of 4 with
benzyl 2,2,2-trichloroacetimidate followed by deprotection
gave the 4-benzyl-2-deoxystreptamine derivative 41. Deace-
tylation of compound 5 with methanolic sodium methoxide
was followed by protection of the free hydroxy groups as an
isopropylidene. Subsequent cleavage of the MTM ether group
afforded the product 42. Alkylation of compound 42 with
BrCH2CO2Me, o-BrBnBr, p-BrBnBr, and o-NO2BnBr under
basic conditions yielded 4-alkylated 2-deoxystreptamine
derivatives 43–46. Treatment of 43 with NaOMe/MeOH/
H2O followed by condensation with 3-nitrophenylenediamine
(Philips reaction)[13] and subsequent deprotection gave the 4-
benzimidazolyl-2-deoxystreptamine derivative 47. Deprotec-
tion of 44–46 provided derivatives 48–50, respectively.
The ESIMS RNA-binding assay was then used to evaluate
the binding affinities of compounds 24–40 and 47–50 for a 27-
mer RNA representing the 16S A site. The compounds were
screened against the 16S A site at equal concentrations in
separate experiments. [(7-Trifluoromethyl)-4-quinolinyl]sul-
fanyl-2-deoxystreptamine (32) showed better binding affinity
(68 mm) than other 4-heterocycle-substituted 2-deoxystrept-
amine derivatives.[14]
Based on this result, the paromomycin mimetic in which
the A ring is replaced with heterocycle I was selected as the
target. Its concise synthesis is summarized in Scheme 3.
Treatment of 15 with methanolic sodium methoxide gave the
product 51. The glycosylation donor 52 was obtained from the
acidic hydrolysis of neomycin B in three steps,[7] which is a
more simple and effective method than the total synthesis
approach. The glycosylation reaction between 51 and C,D-
ring donor 52 was performed in CH2Cl2 in the presence of
TMSOTf. After deacetylation, the desired product 53 was
isolated as a minor product and 54 was isolated as the major
product. Reduction of the azido groups of 53 with Me3P/
NaOH/THF/H2O gave the final target 55. By using a similar
method, 54 was converted into 56.[15]
Scheme 3. Synthesis of 55 and 56: a) NaOMe/MeOH (0.5m), room tem-
perature, 10 h, 85%; b) 1) 51 (1 equiv), 52 (2 equiv), molecular sieves
(4 ꢀ), TMSOTf (1 equiv), 4 h; 2) NaOMe/MeOH (0.5m), 2 h, 10% (53),
31% (54); c) Me3P/THF (8 equiv), THF/H2O (2:1), 10 h, 46%; d) Me3P/
THF (8 equiv), THF/H2O (2:1), 10 h, 52%. TMS=trimethylsilyl.
The ESIMS RNA-binding assay was used to evaluate the
binding activity of 55. This aminoglycoside mimetic exhibited
good RNA-binding activity (Kd < 1 mm).[16] Its Kd value is
higher than that of paromomycin (110 nm), but lower than
those of apramycin (2 mm), bekanamycin (2 mm), and tobra-
mycin (2 mm).[4–6]
Keywords: aminoglycosides · antibiotics · heterocycles ·
mass spectrometry · synthesis design
.
Compound 55 was tested in a coupled bacterial tran-
scription/translation assay in which the ability of a compound
to inhibit either the transcription of a DNA template into
mRNA or the subsequent translation of this mRNA into
functional luciferase protein is evaluated. Compound 55
inhibited this coupled assay with an IC50 of 2 mm, and was
shown to have a minimum inhibitory concentration (MIC) of
3 mm against a gram-negative E. coli strain (ATCC 25922).
In summary, we used MS to study SAR and were able to
design heterocyclic aminoglycoside mimetics more efficiently.
A concise synthesis route was used to prepare heterocyclic
paromomycin mimetics from neomycin B. The use of this
strategy to synthesize more heterocyclic aminoglycoside
mimetics will be reported in due course.
[1] T. K. Stage, K. J. Hertel, O. C. Uhlenbeck, RNA 1995, 1, 95.
[2] B. D. Davis, L. Chen, P. C. Tai, Proc. Natl. Acad. Sci. USA 1986,
83, 6164.
[3] P. Purohit, S. Stern, Nature 1994, 370, 659.
[4] S. A. Hofstadler, K. A. Sannes-Lowery, S. T. Crooke, D. J. Ecker,
H. Sasmor, S. Manalili, R. H. Griffey, Anal. Chem. 1999, 71,
3436.
[5] K. A. Sannes-Lowery, R. H. Griffey, S. A. Hofstadler, Anal.
Biochem. 2000, 2, 264.
[6] R. H. Griffey, S. A. Hofstadler, K. A. Sannes-Lowery, D. J.
Ecker, S. T. Crooke, Proc. Natl. Acad. Sci. USA 1999, 96, 10129.
[7] Y. L. Ding, E. E. Swayze, S. A. Hofstadler, R. H. Griffey,
Tetrahedron Lett. 2000, 41, 4049.
[8] The MS RNA-binding data were obtained from the assay
described in reference [4]; paromomycin (1): Kd = 0.11 mm, 2:
Kd = 80 mm, 3: Kd > 80 mm.
Received: March 7, 2003 [Z51354]
[9] T. Usui, S. Umezawa, Carbohydr. Res. 1988, 174, 133.
Angew. Chem. Int. Ed. 2003, 42, 3409 – 3412
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3411

Communications
[10] W. A. Greenberg, E. S. Priestley, P. S. Sears, P. B. Alper, C.
Rosenbohm, M. Hendrix, S. C. Hung, C. H. Wong, J. Am. Chem.
Soc. 1999, 121, 6818.
[11] E. J. Corey, P. H. Hua, B. C. Pan, S. P. Seitz, J. Am. Chem. Soc.
1982, 104, 6818.
[12] Y. L. Ding, S. A. Hofstadler, E. E. Swayze, R. H. Griffey, Org.
Lett. 2001, 3, 1621.
[13] M. A. Phillips, J. Chem. Soc. 1930, 1409.
[14] The MS RNA-binding data were obtained from the assay
described in reference [4], and neamine was used as the standard
compound. Some of the MS RNA-binding data were reported in
reference [12].
[15] Selected data: 53: 1H NMR (400 MHz, CD3OD): d = 5.69 (d,
1H, J = 12.4 Hz), 5.55 (d, 1H, J = 12.4 Hz), 5.16 (d, 1H, J =
1
2.0 Hz), 5.12 ppm (d, 1H, J = 5.2 Hz); 54: H NMR (400 MHz,
CD3OD): d = 5.82 (d, 1H, J = 12.4 Hz), 5.49 (d, 1H, J = 12.4 Hz),
5.37 (d, 1H, J = 2.0 Hz), 5.04 ppm (d, 1H, J = 1.6 Hz); 55.
1H NMR (400 MHz, D2O): d = 5.67 (d, 1H, J = 12.4 Hz), 5.45 (d,
1
1H, J = 10.8 Hz), 5.16 (s, 1H), 5.07 ppm (s, 1H); 56: H NMR
(400 MHz, D2O): d = 5.51(d, 1H, J = 12.0 Hz), 5.31 (d, 1H, J =
12.4 Hz), 5.20 (s, 1H), 4.72 ppm (d, 1H, J = 7.6 Hz).
[16] The MS RNA-binding data were obtained from the assay
described in reference [4], and paromomycin was used as the
standard compound.
3412
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 3409 – 3412