Redesign of aminoglycosides for treatment of human genetic diseases caused by premature stop mutations

Article abstract of DOI:10.1016/j.bmcl.2006.09.013

A series of new derivatives of the clinically used aminoglycoside antibiotic paromomycin were designed, synthesized, and their ability to read-through premature stop codon mutations was examined in both in vitro translation system and ex vivo mammalian cultured cells. One of these structures, a pseudo-trisaccharide derivative, showed notably higher stop codon read-through activity in cultured cells compared to those of paromomycin and gentamicin.

Full text of DOI:10.1016/j.bmcl.2006.09.013

Bioorganic & Medicinal Chemistry Letters 16 (2006) 6310–6315
Redesign of aminoglycosides for treatment of human genetic
diseases caused by premature stop mutations
Igor Nudelman,a, Annie Rebibo-Sabbah,b, Dalia Shallom-Shezifi,a
Mariana Hainrichson,a Ido Stahl,b Tamar Ben-Yosefb,* and Timor Baasova,*
aThe Edith and Joseph Fischer Enzyme Inhibitors Laboratory, Department of Chemistry,
Institute of Catalysis Science and Technology, Technion—Israel Institute of Technology, Haifa 32000, Israel
bDepartment of Genetics and The Rappaport Family Institute for Research in the Medical Sciences,
Faculty of Medicine, Technion—Israel Institute of Technology, Haifa 31096, Israel
Received 9 August 2006; revised 31 August 2006; accepted 6 September 2006
Available online 25 September 2006
Abstract—A series of new derivatives of the clinically used aminoglycoside antibiotic paromomycin were designed, synthesized, and
their ability to read-through premature stop codon mutations was examined in both in vitro translation system and ex vivo mam-
malian cultured cells. One of these structures, a pseudo-trisaccharide derivative, showed notably higher stop codon read-through
activity in cultured cells compared to those of paromomycin and gentamicin.
ꢀ 2006 Elsevier Ltd. All rights reserved.
A large number of human genetic disorders result from
nonsense mutations, single point alterations in the
DNA, where one of the three stop codons (TAA,
TAG or TGA) replaces an amino acid-coding codon,
leading to premature termination of the translation
and eventually to truncated, nonfunctional proteins.
Currently, hundreds of such mutations are known,1
and for many of those diseases there is presently no
effective treatment.
ity was shown to generate full-length functional proteins
in several genetic disorders.3 Furthermore, clinical trials
with cystic fibrosis patients clearly showed that amino-
glycosides can suppress premature stop mutations in
affected patients.4 However, unfortunately, this great
excitement which continues over the last two decades
was largely hampered because of the following reasons.
First and foremost, aminoglycosides are highly toxic to
mammals (nephrotoxicity and ototoxicity), and the use
of subtoxic doses in clinical trials resulted with the re-
duced read-through efficiency probably insufficient for
Aminoglycosides are highly potent, broad-spectrum
antibiotics that exert their antibacterial therapeutic ef-
fect by selectively binding to the decoding aminoacyl site
(A-site) of the bacterial 16S rRNA, and interfering with
translational fidelity during protein synthesis.2 Interest-
ingly, in the last several years, numerous experiments
performed either in vitro in mammalian translation sys-
tems, cultured cell lines, or animal models confirmed the
ability of certain types of aminoglycoside antibiotics
(Fig. 1) to induce mammalian ribosomes to read-
through stop codon mutations via insertion of a random
amino acid by a nearcognate tRNA. This unique activ-
R2
R1
6'
I
R3
O
R3
H2N2'
II
OH
6'
NH2
I
1'
5'
3
O
HO
1 NH2
O
HO
II
HO
6
H2N 2'
NH2
O
3
NH2
O
1
O
O
6
HO
HO
OH
O
5"
III
NH
OH
III
NH2
O
OH
O
R1
R2
R3
OH
H2N
CH3 NHCH3
H
Gentamicin C1
Gentamicin C2
OH
IV
Keywords: Aminoglycosides; Cystic fibrosis; Drug design; Stop muta-
tion suppression; Usher syndrome.
Paromomycin
CH3 NH2
H
H
H
NH2
Gentamicin C1A
*
Corresponding authors. Tel.: +972 4 8295228; fax: +972 4 8295225
(T.B.-Y.); tel.: +972 4 8292590; fax: +972 4 8295703 (T.B.); e-mail
addresses: benyosef@tx.technion.ac.il; chtimor@tx.technion.ac.il
These authors contributed equally to this work.
CH
OH OH
Geneticin (G-418)
3
Figure 1. Aminoglycosides with stop codon read-through activity.
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.013

I. Nudelman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6310–6315
6311
successful therapy. Second, unlike recent insights into
our understanding of how aminoglycosides might
induce deleterious misreading of the genetic code in
prokaryote cells,5 the molecular mechanism of amino-
glycoside-induced nonsense mutation suppression in
mammalian cells remains to be established. Third, to
date, nearly all suppression experiments have been per-
formed with clinical, commercially available aminogly-
cosides,6 and no efforts have been made to optimize
their activity as stop codon read-through inducers.
Clearly, a systematic search for new structures with im-
proved termination suppression activity and lower toxic-
ity is required to extrapolate the approach to the point
where it can actually help patients.
the plain ribose (structures 2, 4, and 6) or 5-amino ribose
(structures 3, 5, and 7) attached at C5, C6, and C30 of 1,
respectively. The rational in selecting the ribose as a
third sugar ring in 4 and 6 was to retain the identity
of this sugar as in the parent paromomycin, and in par-
allel to explore another hitherto unknown areas of the
mammalian A-site rRNA with possible location of
new modes of binding. Since at physiological pH amino-
glycosides are highly charged and their interaction with
rRNA is mainly determined by electrostatic interaction,8
we reasoned that by adding additional aminosugar to
the paromamine 1 moiety, superior binding to mamma-
lian rRNA and probably better suppression activity will
result. Therefore, we selected 5-amino ribose as a third
sugar ring and prepared the new generation of pseudo-
trisaccharides 3, 5, and 7 with the expectation that they
will possibly function better as read-through inducers
than the parallel structures containing the plain ribose
ring (structures 2, 4, and 6). Similar arguments, the di-
rect addition of an extra amino group to the parom-
amine 1 moiety and its dimerization, served as a basis
for the preparation of compounds 8 and 9, respectively.
Enhanced RNA binding by using dimerized aminogly-
cosides9 and amino-aminoglycosides10 supports these
designs.
For this purpose, as an initial trial, we have modified
paromomycin and prepared a series of derivatives, 1–9
(Fig. 2). In selecting paromomycin as the modification
target we have taken into consideration the following
points. First, although to date there are not enough data
to answer the question why some aminoglycosides in-
duce termination suppression, while others do not, from
the available data, it turns out that aminoglycosides
with a C60 hydroxyl group on ring I (such as G-418
and paromomycin, Fig. 1) are generally more effective
than those with the amine functionality at the same po-
sition.6 Second, paromomycin is the least toxic among
the aminoglycosides that show considerably high sup-
pression activity.7 Based on these data we reasoned that
by dissecting paromomycin structure (via selectively
removing one or two sugar rings) we could identify a
minimal structural motif with significant suppression
activity, which then can be used as a scaffold for the con-
struction of diverse structures with improved termina-
tion suppression and probably with lower toxicity.
All the designed structures 1–9 were synthesized accord-
ing to the general strategy (Fig. 2) that involves direct
Lewis acid promoted cleavage of paromomycin into
the pseudo-disaccharide 1, which is then used as a com-
mon starting material for the preparation of all the
designed structures. For the construction of pseudo-
trisaccharides 2–7, we employed the appropriately pro-
tected three different paromamine acceptors, 11–13,
which selectively expose C5, C6, and C30 hydroxyl
groups of the paromamine moiety, respectively, for gly-
cosidation reactions. These acceptor molecules were
readily accessible from paromamine 1 as illustrated in
Scheme 1. Simultaneous conversion of all the amino
groups of 1 into the corresponding azides was done
by treatment with TfN3 to afford 10. Regioselective
Structures 2–7 preserve rings I and II of paromomycin
(paromamine 1 moiety) as the minimal structural motif
of paromomycin that binds to the mammalian ribosome
and has significant suppression activity (vide infra). The
extended sugar ring (ring III) in each structure is either
Paromomycin
OH
6'
OH
5'
I
O
OH
6'
HO
HO
6'
5'
1'
O I
I
II
NH2
HO
H2N 2'
R
5'
O
NH2
1
6
HO
HO
O 3'
1'
3
II
NH2
NH22'
1'
O
5"
O
O
II
NH2
NH2
1
H2N 2'
NH2
1
6
3
O
3
O
R
5
III
OH
O
5"
HO
6
HO
OH
III
HO OH
OH
6: R=OH
7: R=NH2
Paromamine 1
2: R=OH
3: R=NH2
HO OH
OH
6'
5'
I
O
HO
HO
1'
OH
H2N 2'
II
6'
NH2
1
OH
HO
5'
3
6'
6'
O
HO
NH2
O
5'
5'
O
O
1'
HO
OH
3'
HO
R
6
NH2
1
O
1'
O
NH2
1'
O
3
3N' H
H2N3'
H2N
NH2
1
H2N
1
NH2
O
3
O
5"
3
4: R=OH
5: R=NH2
NH2
H2N
2 O
HO
HO
O
6
OH
III
6
OH
6
OH
HO
9
8
HO OH
Figure 2. Structures of paromamine 1 and its synthetic derivatives 2–9.

6312
I. Nudelman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6310–6315
OAc
OH
OAc
6'
6'
5'
b
a
5' O
O
AcO
AcO
O
HO
HO
1
AcO
AcO
1'
1'
N 2'
N
N
3
N
N 2'
3
1
3
3
3
3
N
3
3
b
a
N
O
N
3
1
O
11
N
3
3
O
2-3
O
HO
10
6
6
HO
R
5
OAc
OH
5
OAc
O
5"
11
c
16a: R=OBz
O
16b: R=N
3
BzO OBz
d,e,f
O
O
O
PMP
O
HO
O
3' N
O
N
BzO
3
3
PMP
3' N
N
N
3
O
3
O
O
3
O
N
O
3
BzO
O
3' N
HO
N
3
6
OH
3
12
13
N
O
3
d,b
c
HO
6
13
12
4-5
O
R
CCl
R
R
3
O
STol
O
5"
O
5"
O
5"
17a: R=OBz
17b: R=N
3
NH
BzO OBz
O
BzO OBz
BzO OBz
O
O
14a: R=OBz
14b: R=N
15a: R=OBz
15b: R=N
3
3
d,b
O
a
3' N
R
6-7
N
3
O
3
5"
Scheme 1. Reagents and conditions: (a) TfN3, Et3N, CuSO4, in
CH2Cl2/MeOH/H2O 3:10:3, (90%); (b) Ac2O (4.2 equiv), pyridine,
ꢀ6 ꢁC, (65%); (c) cyclohexanone dimethyl ketal, CSA, DMF, 110 ꢁC,
(67%); (d) BzCl, pyridine; (e) TFA/H2O 5:3, THF, 40 ꢁC, (89% for the
two steps); (f) anisaldehyde-dimethylacetal, CSA, DMF, 50 ꢁC, (84%).
Tf, trifluoromethanesulfonyl; CSA, camphor sulfonic acid; DMF,
dimethylformamide; Bz, benzoyl; TFA, trifluoroacetic acid.
N
3
O
O
O
BzO OBz
18a: R=OBz
18b: R=N
3
Scheme 2. Reagents and conditions: (a) 15a or 15b BF3–Et2O (cat),
˚
CH2Cl2, 4 A molecular sieves, 11 ! 16a (85%), 11 ! 16b (71%),
12 ! 18a (95%), 12 ! 18b (93%); (b) i—MeNH2 (33% soln in EtOH),
ii—PMe3 (1 M in THF), NaOH 0.1 M, THF, rt; 16a ! 2 (84%),
16b ! 3 (91%), 17a ! 4 (for two steps 75%), 17b ! 5 (44%), 18a ! 6
acetylation of 10 with acetic anhydride at low tempera-
ture gave acceptor 11.11 In another pathway, treatment
of 10 with cyclohexanone dimethyl ketal gave the second
acceptor, 12, in which all functional groups, except the
C30-OH, are protected. Benzoylation of 12 was followed
by acid hydrolysis and benzylidene acetal formation
steps to afford the third acceptor 13.
˚
(84%),18b ! 7 (75%); (c) 14a or 14b NIS, TfOH (cat), CH2Cl2, 4 A
molecular sieves,13 ! 17a (68%), 13 ! 17b (76%); (d) AcOH/H2O
6:1,THF 50 ꢁC for 17a, TFA/H2O 3:2, THF, 60 ꢁC for 17b (52%),
AcOH/H2O 10:3, 1,4-dioxane, 70 ꢁC for 18a (75%), TFA/H2O 5:1,
THF, 50 ꢁC for 18b (82%). PMP, p-methoxyphenyl, NIS,
N-iodosuccinimide.
The acceptors 11–13 were separately subjected to glycos-
idation reactions with two sets of glycosyl donors, 14a–b
and 15a–b,12 to furnish the designed protected deriva-
tives 16–18 in 68–95% yields (Scheme 2). These protect-
ed compounds were then subjected to either two step or
three step deprotection to furnish the final pseudo-tri-
saccharide products 2–7 with excellent purity and isolat-
ed yields.
O
O
O
a,b
c,d
3'
12
N
8
N
3
3
3
N
N
O
3
O
O
19
O
5'
O
O
c,d
1 O
e
1'
O
6
9
3'N
O
12
Ring I in the pseudo-disaccharide 8 has D-allo configu-
ration and was constructed from the paromamine 1 by
selectively inverting the configuration at C30 (Scheme
3). Triflation of the C30-OH in 12 was followed by nucle-
ophilic displacement with azide to afford the corre-
N
3
O
3
3
N
3
3
N
O
3
N
3
N
3
3' O
O
O
O
O1
6
1'
O
5'
20
sponding
cis-diazide
19. Hydrolysis
of
the
Scheme 3. Reagents and conditions: (a) Tf2O, pyridine, (92%); (b)
NaN3, DMF, HMPA, (72%) (c) for 19 AcOH/H2O 8:1, 1,4-dioxane,
75 ꢁC, (60%), for 20 TFA/H2O 5:6, THF, 60 ꢁC, (90%); (d) i—MeNH2
(33% soln in EtOH), ii—PMe3 (1 M in THF), NaOH 0.1 M, THF, rt;
cyclohexylidene ketals with aqueous acetic acid, fol-
lowed by the Staudinger reaction, provided the designed
pseudo-disaccharide 8. Treatment of the same acceptor
12 with CH2Br2 in the presence of NaH gave the pro-
tected dimer 20, which after the similar deprotection
as in the case of 19, afforded the designed dimer 9.
The structures of 1–9 were confirmed by a combination
of various techniques, including HMQC, HMBC, 2D
COSY, and 1D TOCSY NMR, along with mass spectral
analysis. For the biological assays, all compounds (1–9)
were used in their sulfate salt forms, obtained by titra-
tion of the free base forms with diluted sulfuric acid.
˚
19 ! 8 (76%), 20 ! 9 (81%), (e) CH2Br2, NaH, DMF/HMPA 2:1, 4 A
molecular sieves, (82%). HMPA, hexamethylphosphoramide.
Initially, derivatives 1–9 were tested for their ability to
suppress a nonsense mutation in vitro, using a reporter
construct carrying an R3X nonsense mutation (a prema-
ture UGA C stop codon) of the PCDH15 gene (Table 1,
Fig. 3).13,14 Each compound was assayed at several

I. Nudelman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6310–6315
6313
Table 1. Maximal in vitro suppression and translation levels of the R3X mutation, along with the MIC values measured for the designed structures
1–9
Compounda
Concn (lg mLꢀ1
)
Supp. levelb (%)
Trans. levelb (%)
MICc (lg mLꢀ1
)
E. coli
B. Subtilis
Paromomycin
Gentamicin
Paromamine (1)
40
30
80
80
80
80
80
160
80
80
80
49
49
6
4
40 13
12
4
8
<0.5
128
64
40
74 15
9
6.2 0.2
1.3 0.1
512
256
>512
256
>512
>512
192
192
96
2
3
4
5
6
7
8
9
100 10
21
1.5 0.1
3
72
74
75
82
6
7
9
8
48
96
4.4
<1
2.9
2.0
<1
2
192
>512
48
2
2
98 10
71
22
7
2
48
48
a All the designed structures tested (1–9), along with paromomycin and gentamicin, were in their sulfate salt forms.
b The determination of the suppression and translation levels were performed as in the footnote of Figure 3. The results are averages of at least three
independent experiments.
c The MIC values were determined using the double-microdilution method, with two different starting concentrations of the tested compounds (384
and 512 lg/mL). The bacterial strains used were Escherichia coli (ATCC 25922), and Bacillus Subtilis (ATCC 6633). All the experiments were
performed in triplicates and analogous results were obtained in two to four different experiments.
different concentrations and the concentrations at which
maximal suppression levels were observed are given in
Table 1.
preservation of the pseudo-trisaccharide core structure
of the parent paromomycin (rings I-III) in 3 is important
for efficient read-through activity. The suppression data
of 2–7 show that in the series of pseudo-trisaccharides,
an increased number of amino groups in each pair leads
to improved read-through activity, in agreement with
the reports about enhanced RNA binding of aminogly-
coside derivatives with extra amino groups.9,10 Howev-
er, the data obtained with 8 and 9 indicate that merely
increasing the number of amino groups on the parom-
amine scaffold does not always lead to an increase in
read-through activity. Nevertheless, the observed 13-
fold higher suppression level of derivative 3 compared
to that of the corresponding ribose derivative 2, and
over 3-fold better activity compared to that of parom-
amine 1, suggest that the presence of C500-NH2 in 3 is
responsible for its elevated read-through activity.
As seen from the data in Table 1, removal of either one
ring (ring IV) or two rings (rings III and IV) in paromo-
mycin dramatically decreases its maximal in vitro
read-through activity from 49% suppression to 1.3%
(compound 2) and 6.2% (compound 1), respectively.
The substantially higher suppression level of 1 (6.2%)
compared to that of the pseudo-trisaccharide fragment
2 (1.3%) implies that paromamine 1 represents the min-
imal structural motif of paromomycin that has signifi-
cant suppression activity. Connection of the plain
ribose ring to the paromamine scaffold either at C6
(compound 4) or C30 position (compound 6), along with
the addition of one extra amine (8) or paromamine
dimerization (9), gave lower suppression levels than that
of paromamine itself.
To further evaluate the read-through potential of
compound 3, its activity was assayed in cultured
mammalian cells using a dual luciferase reporter sys-
tem,16 and compared to the activities of paromomycin
and gentamicin (Fig. 4). At all concentrations tested,
the activity of 3 is superior to those of paromomycin
and gentamicin. In addition, gentamicin (Fig. 1),
which is currently the only clinically relevant amino-
glycoside shown to have the ability to suppress non-
sense mutations in patients, is less efficient than
either paromomycin or 3. The improved read-through
efficiency of compound 3 relative to those of paromo-
mycin and gentamicin in the whole-cell system (in
contrast to the in vitro results) indicates that the effec-
tive concentration of compound 3 inside the cell and
near the ribosome is higher than the effective concen-
trations of paromomycin and gentamicin. This could
be an outcome of better penetration and improved up-
take of 3 by the cells, resulting from its smaller size
compared to paromomycin and gentamicin. In addi-
tion, since compound 3 has one less positive charge
(+4) than either paromomycin (+5) or gentamicin
The most important results, however, were observed
when instead of plain ribose, the 5-amino ribose was
connected at different positions. The observed suppres-
sion levels of all these derivatives (compounds 3, 5,
and 7) are higher than the parallel structures containing
plain ribose ring at the same location (2, 4, and 6,
respectively). Compound 3, in which the ribosamine is
attached to C5 of the paromamine structure, shows
the highest suppression level among the new derivatives.
The maximal suppression level of 3 (21%) was obtained
at 80 lgmLꢀ1, a concentration which is much higher
than the concentrations required for the maximal activ-
ities of both paromomycin and gentamicin (Table 1,
Fig. 3). Nevertheless, the fact that compound 3 still re-
tained translation level of 72% at 80 lg mLꢀ1 is also a
very important result, as discussed below.
In the series of pseudo-trisaccharides 3, 5, and 7, a par-
ticular influence of the position of the ribosamine ring is
observed: C5 (3) ꢁ C6 (5) > C30 (7), suggesting that the

6314
I. Nudelman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6310–6315
Comp. 3
20 40 80 100
A
0
μg/mL
35 kDa
25 kDa
25
20
15
10
5
100
75
50
25
0
0
0
20
40
80 100
[Comp. ]60(μg/mL)
3
Figure 4. Ex vivo suppression of a nonsense mutation by compound 3,
paromomycin, and gentamicin. The p2Luc plasmid containing a TGA
C nonsense mutation in a polylinker between the renilla luciferase and
firefly luciferase genes16 was transfected into COS-7 cells and addition
of tested compounds was performed after 15 h. Luciferase activity was
determined following 24 h of incubation, using the Dual Luciferase
Reporter Assay System (Promega). Stop codon read-through was
calculated as described previously.16 The results are averages of at least
three independent experiments.
Paromomycin
B
μg/mL
35 kDa
0
5
10 20 30 40
25 kDa
100
80
60
40
20
0
50
40
30
20
10
antibacterial activity against both strains, the synthetic
derivatives did not retain the antibiotic character of nat-
ural aminoglycosides. The antibacterial activities of all
new compounds, including compound 3 which had rela-
tively high read-through activity, are markedly lower
than that of the parent paromomycin.
0
0
10
20
30
40
[Paromomycin] (μg/mL)
SP6
TAA TAG
β-Gal
rat PGM
Cpromoter
The origin of aminoglycoside toxicity is still controver-
sial and probably results from a combination of different
factors/mechanisms.15 Yet, several encouraging results
obtained with compound 3 suggest that it might exhibit
reduced toxicity relative to that of the parent paromo-
mycin or gentamicin compounds. The data in Table 1
and Figure 3 show that besides the significant read-
through activity of compound 3, it also retained much
higher translation level (ꢂ72% at 80 lgmLꢀ1) than
either gentamicin or paromomycin (ꢂ40% at 30–40 lg
mLꢀ1). Another factor thought to be one of the causes
for the cytotoxicity of aminoglycosides is their binding
to mitochondrial 12S rRNA A-site (whose sequence is
very close to the bacterial A-site).8 Thus, the observed
inability of 3 to show significant antibacterial activity
could also hint about its lower toxicity. Nevertheless,
only cell cytotoxicity assay could determine whether
these results truly indicate about lower toxicity of com-
pound 3.
... ATG TTT TGA CAG TTT TAT CTC TGG ACA...
Stop codon
35 kDa Read-through
25 kDa No Read-through
Translation
products
Figure 3. Levels of in vitro suppression of a nonsense mutation (•) and
overall translation (ꢃ) by compound 3 (A) and by paromomycin (B).
Panel C describes the reporter system used for the in vitro translation
assays.
A plasmid-based reporter construct contains a TGA C
nonsense mutation between a 25-kDa polypeptide encoding open
reading frame (ORF) and a 10-kDa polypeptide encoding ORF, such
that efficient translation termination at the stop codon results in the
production of a 25-kDa polypeptide, while suppression of the nonsense
mutation by the aminoglycoside allows the synthesis of a 35-kDa
protein.6 The nonsense mutation context was derived from PCDH15
cDNA. The reporter plasmid was in vitro transcribed and translated in
the presence of [35S]methionine, and the reaction products were
separated by SDS–PAGE and quantified using PhosphorImager
analysis. The suppression level was calculated as the relative propor-
tion of the 35-kDa product out of total protein (the sum of 35-kDa and
25-kDa), and the translation level was calculated as the relative
proportion of the total protein at each drug concentration out of the
total protein without drug. The results are averages of at least three
independent experiments.
In summary, this study provides a new direction for
the development of novel aminoglycoside-based small
molecules that selectively target mammalian cells by
means of optimizing the efficiency of aminoglycoside-
induced suppression of premature stop mutations; this
progress may offer promise for the treatment of many
genetic diseases. Thus, although the ‘ideal readthrough
inducer molecule’ is yet to be identified, the results
introduced in this study indicate that this is an achiev-
able goal. Direct cytotoxicity and cell permeability as-
says of 3, along with the design and synthesis of
‘second generation derivatives’ of paromomycin, are
currently underway.
(+5), it is possible that 3 has less nonspecific binding
to other cell components than those of paromomycin
and gentamicin.
Finally, the designed structures 1–9 were investigated as
antibacterial agents against both Gram-negative (Esche-
richia coli) and Gram-positive (Bacillus subtilis) bacteria
(Table 1). From the MIC values, it can be seen that
whereas gentamicin and paromomycin exhibit excellent

I. Nudelman et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6310–6315
6315
8. Vicens, Q.; Westhof, E. Chembiochem. 2003, 4,
1018.
Acknowledgments
9. Michael, K.; Wang, H.; Tor, Y. Bioorg. Med. Chem. 1999,
7, 1361.
10. Wang, H.; Tor, Y. Angew. Chem., Int. Ed. 1998, 37, 109.
11. Ding, Y.; Swayze, E. E.; Hofstadler, S. A.; Griffey, R. H.
Tetrahedron Lett. 2000, 41, 4049.
This research was supported by the Israel Science
Foundation founded by the Israel Academy of Sciences
and Humanities (Grant No. 766/04), by the German-
Israeli Foundation for Scientific Research
&
Development (Grant no. 1-2104-1432.2/2004 to
T.B.Y), by the Mizutani Foundation for Glycoscience
(Grant No. 060012, T.B. and T.B.Y.), and in part by
the B. E. Lusting Fund for Cystic Fibrosis Research at
the Technion. D. Shallom-Shezifi is supported in part
at the Technion by a Neaman Fellowship. The authors
thank David M. Bedwell (University of Alabama at
Birmingham) for providing the pDB650 plasmid and
John F. Atkins (University of Utah) for providing the
p2luc plasmid.
12. Fridman, M.; Belakhov, V.; Lee, L. V.; Liang, F. S.;
Wong, C. H.; Baasov, T. Angew. Chem., Int. Ed. Engl.
2005, 44, 447.
13. Mutations in the PCDH15 gene (which encodes proto-
cardherin 15) cause type 1 Usher syndrome (USH1), which
is characterized by profound prelingual hearing loss,
vestibular areflexia, and prepubertal onset of retinitis
pigmentosa (RP) (Petit, C. Annu. Rev. Genomics Hum.
Genet. 2001, 2, 271). In humans, four different PCDH15
USH1-causing nonsense mutations (R3X, R245X, R643X,
and R929X) have been reported. Interestingly, while the
above nonsense mutations of PCDH15 cause USH1,
certain missense mutations in the same gene cause only
nonsyndromic deafness, which is not associated with RP.
Such observations suggest that partial or low level activity
of the protein encoded by this gene may be sufficient for
normal retinal function, making it a suitable candidate for
read-through therapy.
14. Suppression of nonsense mutations by compounds 1–9
was tested in vitro using a reporter plasmid harboring the
R3X mutation of the PCDH15 gene (Ahmed, Z. M.;
Riazuddin, S.; Bernstein, S. L.; Ahmed, Z.; Khan, S.;
Griffith, A. J.; Morell, R. J.; Friedman, T. B.; Riazuddin,
S.; Wilcox, E. R. Am. J. Hum. Genet. 2001, 69, 25). To
create this plasmid, the oligonucleotides GATCCATGT
TTTGACAGTTTTATCTCTGGACA and AGCTTGTC
CAGAGATAAAACTGTCAAAACATG were annealed
to each other, and inserted into the BamHI and HindIII
sites of plasmid pDB650.6
References and notes
1. Atkinson, J.; Martin, R. Nucleic Acids Res. 1994, 22, 1327.
2. Magnet, S.; Blanchard, J. S. Chem. Rev. 2005, 105, 477.
3. (a) Keeling, K. M.; Bedwell, D. M. Curr. Pharmacoge-
nomics 2005, 3, 259; (b) Kerem, E. Curr. Opin. Pulm. Med.
2004, 10, 547.
4. Wilschanski, M.; Yahav, Y.; Yaacov, Y.; Blau, H.;
Bentur, L.; Rivlin, J.; Aviram, M.; Bdolah-Abram, T.;
Bebok, Z.; Shushi, L.; Kerem, B.; Kerem, E. N. Engl. J.
Med. 2003, 349, 1433.
5. (a) Carter, A. P.; Clemons, W. M.; Brodersen, D. E.;
Morgan-Warren, R. J.; Wimberly, B. T.; Ramakrishnan, V.
Nature 2000, 407, 340; (b) Francois, B.; Russell, R. J.;
Murray, J. B.; Aboul-ela, F.; Masquida, B.; Vicens, Q.;
Westhof, E. Nucleic Acids Res. 2005, 33, 5677; (c) Kaul, M.;
Barbieri, C. M.; Pilch, D. S. J. Mol. Biol. 2005, 346, 119.
6. Manuvakhova, M.; Keeling, K.; Bedwell, D. M. RNA
2000, 6, 1044.
15. (a) Forge, A.; Schacht, J. Audiol. Neurootol. 2000, 5, 3; (b)
Mingeot-Leclercq, M. P.; Tulkens, P. M. Antimicrob.
Agents Chemother. 1999, 43, 1003.
7. Fujisawa, K.; Hoshiya, T.; Kawaguchi, H. J. Antibiot.
(Tokyo) 1974, 27, 677.
16. Grentzmann, G.; Ingram, J. A.; Kelly, P. J.; Gesteland, R.
F.; Atkins, J. F. RNA 1998, 4, 479.