Repairing faulty genes by aminoglycosides: Development of new derivatives of geneticin (G418) with enhanced suppression of diseases-causing nonsense mutations

Article abstract of DOI:10.1016/j.bmc.2010.03.060

New pseudo-di- and pseudo-trisaccharide derivatives of the aminoglycoside drug G418 were designed, synthesized and their ability to readthrough nonsense mutations was examined in both in vitro and ex vivo systems, along with the toxicity tests. Two novel

Full text of DOI:10.1016/j.bmc.2010.03.060

Bioorganic & Medicinal Chemistry 18 (2010) 3735–3746
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Repairing faulty genes by aminoglycosides: Development of new derivatives
of geneticin (G418) with enhanced suppression of diseases-causing
nonsense mutations
*
Igor Nudelman , Dana Glikin , Boris Smolkin, Mariana Hainrichson, Valery Belakhov, Timor Baasov
The Edith and Joseph Fischer Enzyme Inhibitors Laboratory, Schulich Faculty of Chemistry, Technion-Israel Institute of Technology, Haifa 32000, Israel
a r t i c l e i n f o
a b s t r a c t
Article history:
New pseudo-di- and pseudo-trisaccharide derivatives of the aminoglycoside drug G418 were designed,
synthesized and their ability to readthrough nonsense mutations was examined in both in vitro and
ex vivo systems, along with the toxicity tests. Two novel lead structures, NB74 and NB84, exhibiting sig-
nificantly reduced cell toxicity and superior readthrough efficiency than those of gentamicin, were dis-
covered. The superiority of new leads was demonstrated in six different nonsense DNA-constructs
underling the genetic diseases cystic fibrosis, Duchenne muscular dystrophy, Usher syndrome and Hurler
syndrome.
Received 8 February 2010
Accepted 25 March 2010
Available online 27 March 2010
Dedicated to ‘Tetrahedron Young
Investigator 2010: Peter Seeberger’
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Aminoglycosides
Genetic diseases
Nonsense mutations
Translational therapy
Cystic fibrosis
1. Introduction
strated the ability of certain types of aminoglycosides to induce
mammalian ribosome to readthrough disease-causing PTC muta-
The treatment of genetic disorders is one of the biggest chal-
lenges lying ahead of modern medicine.1–3 There are more than
1800 inherited human diseases caused by nonsense mutations,
that is, alterations in the genetic code that prematurely stop the
translation process and lead to the production of truncated non-
functional proteins. These false stop codons are also known as pre-
mature termination codons (PTCs) and, depending on the disorder,
account for 5–70% of cases of genetic diseases, including cystic
fibrosis (CF), Duchenne muscular dystrophy (DMD), Usher syn-
drome (USH), Hurler syndrome (HS) and numerous types of cancer.
For many of those diseases there is presently no effective treat-
ment and only treatment widely used is symptomatic. While major
advancements have been made in gene therapy, it is still far from
achieving clinical success. However, other potential methods for
treating single gene related diseases have also emerged during last
decades. One such approach is the suppression of pathogenic non-
sense mutations through inducing translational readthrough of the
in-frame PTCs.2,3 Aminoglycoside antibiotics were the first drugs
that gave promising results in this respect: numerous in vitro
and in vivo experiments including clinical trials have demon-
tions and partially restore full-length functional proteins.1,3
The fact that aminoglycosides (namely paromomycin and G418,
Fig. 1) could suppress PTC mutations in cultured mammalian cells
was first demonstrated by Burke and Mogg in 1985, who also
pointed out the therapeutic potential of these drugs in treatment
of genetic disorders.4 The first genetic disease examined was CF
that is caused by mutations in the cystic fibrosis transmembrane
conductance regulator (CFTR) protein; nonsense mutations found
in the CFTR gene could be suppressed by aminoglycosides G418
and gentamicin (Fig. 1), as measured by the appearance of full-
length, functional CFTR protein in bronchial epithelial cell lines.5,6
Further experiments with gentamicin, both in mouse models of CF7
and DMD8 and in clinical trials with patients of these two dis-
eases9,10 carrying nonsense mutations in CFTR and Dystrophin
genes, respectively, evidently demonstrated partial production of
missing proteins. Other genetic disorders for which the therapeutic
potential of aminoglycosides was tested in in vitro systems, cul-
tured cell lines, or animal models including cancer,11 Rett syn-
drome,12 Hurler syndrome,13 nephrogenic diabetes insipidus,14
nephropathic cystinosis,15 retinitis pigmentosa,16 ataxia-telangiec-
tasia17 and more.
It should be noted that aminoglycoside-induced production of
full-size protein in the presence of a PTC mutation, even if only
in low amount, may be functionally significant. This is especially
* Corresponding author. Tel.: +972 4 829 2590; fax: +972 4 829 5703.
E-mail address: chtimor@tx.technion.ac.il (T. Baasov).
Equal contribution.
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.03.060

3736
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
readthrough is largely obscure.23,24 Clearly, the challenges in this
6'
Me
O
R2
6'
HO
HO
R1
I
I
emerging field should be at deciphering the intimate structural
and functional determinants of aminoglycosides as readthrough
and toxicity inducers, in order to develop new structures with im-
proved suppression activity and reduced toxicity.
O
HO
II
II
NH2
3
NH2
H2N
H2N
3
O
O
1
1
NH2
NH2
HO
HO
6
6
O
O
Towards these ends, we have recently reported on the develop-
ment of compound 1 (also named NB30, Fig. 1)25 and compound 2
(also named NB54, Fig.1)26 as novel pseudo-trisaccharide deriva-
tives of the clinical aminoglycoside paromomycin. Whereas the
first-generation lead, compound 1, exhibited significantly reduced
cytotoxicity in comparison to gentamicin and paromomycin,27 and
promoted dose-dependent suppression of nonsense mutations of
the PCDH15 gene, one of the underlying causes of type 1 Usher syn-
drome (USH1),27 its suppression potency was significantly lower
relative to that of gentamicin and paromomycin. In attempts to
further improve the suppression efficiency and reduce the toxicity
of 1, we have developed the compound 2 as a second-generation
lead structure.26 Compound 2 exhibited significantly reduced cell,
cochlear and acute toxicities, and has substantially higher stop-co-
don readthrough potency in both in vitro and ex vivo studies, than
those of gentamicin and paromomycin.26 Continuing systematic
fine-tuning of the developed leads, in the present work we report
on the design, synthesis and comparative evaluation of the third-
generation lead structures 7 and 8 (also named NB74 and NB84,
respectively, Fig. 1). They differ from the previous leads (1 and 2)
by the presence of the (R)-60-methyl group on the glucosamine ring
(ring I) and show significantly reduced cell toxicity while in paral-
lel exhibit substantially higher stop-codon readthrough potency in
both in vitro and ex vivo studies, than those of gentamicin. The ob-
served comparative data also indicate that the introduction of (R)-
60-methyl group either on compound 1 (to give compound 7) and
on compound 2 (to give compound 8) greatly enhances the stop-
codon readthrough potency of the resulted structure (compounds
7 and 8, respectively), while its effect on cell toxicity of the resulted
compounds in comparison to those of the parent structures was
negligible.
O
O
HO
HO
Me
Me
III
III
NH
NH
OH
OH
Me
Me
Geneticin (G418)
R1
R2
Gentamicin C1
Gentamicin C2
Gentamicin C1A
CH3 NHCH3
CH3 NH2
I
6'
HO
HO
NH2
NH2
O
H
HO
II
NH2
H2N
HO
3
O
NH2
OH
O
5''
5
AHB =
O
OH
III
OH
Paromomycin
OH NH2
O
O
O
OH
H2N
IV
6'
6'
HO
HO
HO
O
O
HO
HO
HO
NH2
NH2
3
H2N
H2N
H2N
3
O
O
1
1
NHR
NHR
O
HO
5''
5
O
OH
OH
3: R=H
4:
1
2
:
(NB30) R=H
R=AHB
:
HO OH
(NB54) R=AHB
Me
O
Me
6'
6'
HO
HO
HO
HO
HO
O
HO
NH2
NH2
3
H2N
H2N
H2N
3
O
O
1
1
NHR
NHR
O
HO
5
5
5''
O
OH
OH
2. Results
7
:
(NB74) R=H
5: R=H
6: R=AHB
8
:
(NB84) R=AHB
HO OH
2.1. Design hypotheses and synthesis
Figure 1. Chemical structures of a series of natural and synthetic 2-deoxystrepta-
mine-derived (ring II) aminoglycosides, including compounds 1–8 that were
investigated in this study.
To further improve both the premature stop codon suppression
activity and toxicity profiles of 1 and 2, we sought to explore the
effect of selective installation of the (R)-60-methyl group on these
structures to yield compounds 7 and 8, respectively, because of
the following reasons. First, although to date there are not enough
structural and mechanistic data available to answer the question
why some aminoglycosides induce termination suppression in
mammalian systems while others do not, from the existing data
it is generally accepted that aminoglycosides with a C60 hydroxyl
group on ring I (such as paromomycin and G418, Fig. 1) are more
effective than those with the amine functionally at the same posi-
tion.28–30 Second, among all aminoglycosides tested to date for
readthrough activity,31,32 G418 is the only one that has (R)-60-
methyl group on its glucosamine moiety (ring I) with the second-
ary alcohol at 60 position. Furthermore, G418 is also the strongest
readthrough inducer among all aminoglycosides tested to date5,6,31
and no detailed SAR study on its derivatives has yet been per-
formed. Third, despite its potent readthrough activity, the use of
G418 as a therapeutic agent is not possible because it is lethal even
at very low concentrations.33 However, it is not known what struc-
tural element(s) of G418 contribute to its elevated readthrough
activity and what to its high cytotoxicity. Separation of such ele-
ments, if possible, should be of high importance for the develop-
ment of new derivatives with potent suppression activity and
relevant for recessive disorders, where protein expression is essen-
tially absent, like CF, DMD, HS, ataxia-telangiectasia and others,
and even 1% of normal protein function may restore a near-normal
or clinically less severe phenotype.3,9,18 Consequently, the greatest
promise of aminoglycosides as potential pharmacogenetic agents
is primarily for the treatment of recessive disorders caused by non-
sense mutations. Recent studies have indicated that aminoglyco-
sides induced readthrough may be
a promising therapy in
autosomal-dominant disorders as well.19
Despite this great excitement and promise that continues over
the last two decades, severe side-effects of aminoglycosides
including high toxicity to mammals along with the reduced read-
through efficiency at subtoxic doses, have limited their clinical
benefit for suppression therapy.20 In addition, to date, the clinical
application of aminoglycosides is limited to their use as antibiotics,
and almost no efforts have been made to optimize their activity as
stop codon readthrough inducers. Furthermore, while the recent X-
ray crystal structures of the bacterial ribosome in complex with
aminoglycosides shed light on the mechanism of aminoglycosides
action as antibiotics,21,22 no crystal structure of human ribosome is
presently available and the mechanism of aminoglycoside-induced

I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
3737
reduced toxicity. For example, very recent study on gentamicin
demonstrated that the inversion of an absolute configuration at a
single carbon atom, from (S)-60-gentamicin C2 to (R)-60-gentamicin
C2, significantly reduces cell toxicity and apparent nephrotoxicity
of the (R)-diastereomer in comparison to that of (S)-diastereomer,
as determined in cell culture and animal studies, while the bacte-
ricidal efficacy is not affected.34 Based on these accumulative data,
we anticipated that the exploration of the impact of (R)-60-methyl
group could lead to new derivatives of aminoglycosides with im-
proved readthrough activity and probably with reduced toxicity.
Initially, the concept was tested on pseudo-disaccharides; we
synthesized compounds 3–6 (Fig. 1) and tested their comparative
in vitro readthrough efficiency. Compounds 3 and 4 were synthe-
sized as previously reported by us.25,26 As a starting material for
the synthesis of 5 and 6 we used G418 (Scheme 1). Treatment of
G418 with anhydrous HCl (AcCl in MeOH) gave a highly regioselec-
tive hydrolysis between the rings II and III to afford compound 5 in
75% isolated yield as its hydrochloride salt. We would like to note
that the isolated material was found inadequate for biological
tests, because the presence of some minute impurities including
the starting material G418. Since these impurities could not com-
pletely taken off either by repeated re-crystallization or by flash
column chromatography, we decided to perform two additional
chemical steps: simultaneous conversion of all the amino groups
into the corresponding azides (TfN3, CuSO4) to afford the com-
pound 9, which after Staudinger reaction generated the desired
product 5 as a homogeneous material. For the synthesis of the
pseudo-disaccharide 6, however, the hydrolysis product of G418
(shown by dashed arrow in Scheme 1) was directly treated with
Zn(OAc)2 and Boc2O to afford the selectively N-1-Boc-protected
derivative (60% isolated yield), which after azidation (TfN3, CuSO4)
gave the diazido derivative 10 in 88% isolated yield. Compound 10
was then converted to the corresponding N-1-AHB derivative 11 in
three successive steps in the overall 70% yield for three steps. Thus,
treatment of 10 with TFA gave the corresponding N-1 amine which
after treatment with (S)-2-hydroxy-4-azidobutyric acid (HOBT,
DCC, Et3N) and incubation of the crude reaction mixture with
methylamine afforded the selectively N-1-acylated product 11.
The Staudinger reaction then generated the desired product 6.
The synthesis of the pseudo-trisaccharides 7 and 8 was accom-
plished from the corresponding protected pseudo-disaccharide
intermediates 9 and 11, respectively, by using essentially the same
chemical transformations as illustrated in Scheme 2. Regioselec-
tive acetylation of 9 and 11 with Ac2O at low temperature gave
the corresponding C5 acceptors 12 and 13 in 60% and 70% isolated
yields, respectively. Our earlier studies with the paromamine
(compound 3)25 and N-1-AHB-paromamine (compound 4)26 scaf-
folds have demonstrated that the corresponding C5 acceptors are
highly unreactive in glycosylation reactions and only trichloroace-
timidate donors gave satisfactory results. Taking this argument
into account, for the glycosylation of 12 and 13 we used the tri-
chloroacetimidate donor 14,25 which furnished the desired pseu-
do-trisaccharides 15 and 16 in 77% and 91% isolated yields,
respectively, exclusively as beta-anomers at the newly generated
glycosidic linkage. Two sequential deprotection steps: treatment
with methylamine to remove all the ester protections, and the
Staudinger reaction to convert azides to the corresponding amines,
then gave the target compounds 7 and 8 in 80% and 72% isolated
yields, respectively, for two steps. The structures of all new com-
pounds (5–8) were confirmed by a combination of various 1D
and 2D NMR techniques, including 2D 1H–13C HMQC and HMBC,
2D COSY, and 1D selective TOCSY experiments, along with mass
spectral analysis.
2.2. Comparative in vitro stop codon suppression tests to eval-
uate the impact of (R)-60-methyl group on pseudo-disaccharides
(compounds 3–6) and pseudo-trisaccharides (compounds 1, 2, 7
and 8)
Initially, the influence of the (R)-60-methyl group on read-
through potential was evaluated on pseudo-disaccharide struc-
tures by using a dual luciferase reporter assay system.26,32,35
Briefly, the p2luc plasmid containing a TGA C nonsense mutation
was transcribed and translated using the TNT reticulocyte lysate
quick coupled transcription/translation system in the presence of
varying concentrations of compounds 3–6, and the stop codon sup-
pression efficiency was calculated as previously reported35
(Fig. 2A).
Me
O
Me
O
HO
HO
HO
HO
HO
1) AcCl, MeOH (75%)
2) TfN3, CuSO4 (90%)
HO
PMe3, NaOH
(70%)
NH2
OH
N3
OH
H2N
N3
G418
O
O
NH2
N3
HO
HO
9
5
1) TFA, CH2Cl2
Me
O
O
HO
HO
HO
2)
N3
HO
N3
OH
OH
HOBT, DCC
N3
1) Zn(OAc)2, Boc2O (60%)
2) TfN3, CuSO4 (88%)
O
NHBoc
HO
HO
3) MeNH2, EtOH
10
(70% for 3 steps)
Me
O
Me
O
HO
HO
HO
HO
HO
OH
N3
OH
NH2
H
OH
PMe3, NaOH
(65%)
N3
H
N
H2N
O
O
N
HO
N3
HO
NH2
OH
O
O
11
6
Scheme 1.

3738
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
N3
O
CCl3
NH
Me
O
Me
O
O
HO
HO
AcO
AcO
HO
AcO
N3
OH
N3
N3
Ac2O, py
12
BzO OBz 14
N3
O
O
R
R
HO
HO
(60% for
)
BF3-Et2O, CH2Cl2
-30 C (77% for 15)
OAc
(70% for 13)
°
9: R=N3
11: R=
12: R=N3
13: R=
O
16
)
(91% for
O
N3
N3
N
H
N
H
OH
OAc
Me
O
AcO
AcO
AcO
N3
R
OAc
15:
R=N3
N3
N3
1) MeNH2, EtOH
O
7-8
O
O
16: R=
N3
2) PMe3, THF/NaOH
O
N
H
OAc
(80% for 7 for 2 steps)
(72% for 8 for 2 steps)
BzO OBz
Scheme 2.
(R)-60-methyl group is substantially higher versus that of N-1-AHB
group; comparison of the gaps between 3?5 and 4?6 for the net
effect of (R)-60-methyl group, and the gaps between 3?4 and 5?6
for the net effect of N-1-AHB group. The pseudo-disaccharide 6
that contains both N-1-AHB and (R)-60-methyl groups shows the
highest suppression levels at all the concentrations tested.
Whether the structure 6 is reached by 3?4?6 transition or by
3?5?6 transition, gradual increase of readthrough activity is ob-
served, indicating about the additive influence of the two impor-
tant structural elements, N-1-AHB and (R)-60-methyl groups, on
the observed readthrough activity. Nevertheless, the observed sig-
nificantly higher suppression levels of the pseudo-disaccharides 5
and 6 compared to those of 3 and 4, suggested that 5 and 6 can
serve as improved scaffolds for the generation of new derivatives.
Our earlier structure–activity studies25,26 demonstrated that the
preservation of the pseudo-trisaccharide core structure of paromo-
mycin (rings I–III) in compounds 1 and 2 is important for their effi-
cient readthrough activity. Therefore, initially we focused on the
construction and tests of the new pseudo-trisaccharides, com-
pounds 7 and 8 that contain the 5-amino ribose as the ring III at-
tached at position 5 of the new scaffolds 5 and 6, respectively.
Comparative in vitro suppression tests of the pseudo-trisaccha-
rides 1, 2, 7, and 8 were performed under the same experimental
conditions as for the corresponding pseudo-disaccharides (com-
pounds 3–6, Fig. 2A), using p2luc plasmid containing a TGA C non-
sense mutation, and the observed data are shown in Figure 2B.
Very similar trend in regards to the relative influence of the (R)-
60-methyl and N-1-AHB groups on the suppression efficiency was
observed. Thus, the impact of (R)-60-methyl group is substantially
higher versus that of N-1-AHB group; at all concentrations tested,
the gaps 1?7 and 2?8 (representing the differences only by the
(R)-60-methyl group) are significantly larger than those 1?2 and
7?8 (representing the differences only by the N-1-AHB group).
While these observations were indeed very encouraging, we
wanted to take them with certain precautions. This because the
efficiency of aminoglycosides-induced readthrough is highly
dependent on: (i) the identity of stop codon (UGA > UAG > UAA),
(ii) the identity of the first nucleotide immediately downstream
from the stop codon (C > U > A P G) and (iii) the local sequence
context around the stop codon.31,32 Therefore, in attempts to pro-
vide broader understanding on structure–activity relationship of
the designed structures, we used a variety of constructs containing
different sequence contexts around premature stop codons derived
Figure 2. Levels of in vitro suppression of a TGA C nonsense mutation by (A)
pseudo-disaccharide compounds 3 (h), 4 (j), 5 (s) and 6 (d), and (B) pseudo-
trisaccharide compounds
1 (h), 2 (j), 7 (s) and 8 (d). The p2luc plasmid
containing TGA C nonsense mutation in a polylinker between Renilla and Firefly
luciferase genes32 was transcribed and translated using the TNT reticulocyte lysate
quick coupled transcription/translation system (Promega™). The luciferase activity
was determined by using the Dual luciferase reporter assay system (Promega™)
and the stop codon suppression efficiency was calculated as described previously.35
The results are averages at least 2–3 independent experiments.
As seen from the data in Fig. 2A, installation of either N-1-AHB
moiety (compound 4) or (R)-60-methyl group (compound 5) on the
paromamine scaffold (compound 3) dramatically increases its
in vitro readthrough activity. In addition, as depicted in the figure
with dashed arrows, at all the concentrations tested, the impact of

I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
3739
from the PCDH15, CFTR, Dystrophin and IDUA genes that underlie
USH1, CF, DMD and HS, respectively. The prevalent nonsense
mutations of these diseases that we chose were: R3X and R245X
for USH1, G542X and W1282X for CF, R3381X for DMD, and
Q70X for HS. DNA fragments containing the nonsense mutation
or the corresponding wild type codon, in their natural context,
were cloned in frame between the Renilla and the Firefly luciferase
genes of the p2luc vector as described previously by us.26 The
resulting six nonsense mutation-carrying plasmids, were tran-
scribed and translated in the presence of varying concentrations
of compounds 1, 2, 7, 8 and gentamicin, and the stop codon sup-
pression efficiency was calculated as previously reported26,35
(Fig. 3).
From the observed data in Figure 3 it can be seen that the effi-
cacy of readthrough is substantially different between different
constructs and compounds tested, with no obvious dependence
of readthrough effectiveness on the introduced type of modifica-
tion on aminoglycoside. Nevertheless, in all mutations tested
(except R3X, Fig. 3A) compound 8 induced the highest level of
readthrough, followed by 2, 7, gentamicin, and 1. The UGA C tetrac-
odon sequence (R3X) showed the best translational readthrough
than UGA A and UGA G, with the UAG C tetracodon least efficient,
in agreement with earlier observations.26,31,32 Inhibition of transla-
tion was monitored as the ratio of Renilla luciferase activity with
and without the presence of aminoglycosides.35 In all the con-
structs tested, the highest concentration of compound 8 resulted
in approximately 50% reduction in overall translation, whereas
the effects of 1, 2, 7, and gentamicin on translation were signifi-
cantly milder resulting in approximately 10–30% reduction in
overall translation.
2.3. Comparative stop codon suppression tests in mammalian
cell line
Cell line experiments were performed on three different dual
luciferase reporter plasmids harboring the PCDH15-R3X nonsense
mutation of USH1, the IDUA-Q70X nonsense mutation of HS, and
the CFTR-W1282X nonsense mutation of CF. These reporter con-
structs were the same as we used above for the in vitro study,
and have distinct advantage to control for differences in mRNA lev-
els between normal and nonsense-containing sequences over
those of single reporter or direct protein analysis, as previously
noted.32,35 The constructs were transfected into a human embry-
onic kidney cell line (HEK-293) and incubated with varying con-
centrations of 2, 8, G418 and gentamicin (Fig. 4). In a single case,
PCDH15-R3X nonsense mutation, in addition to 2, 8 and gentami-
cin, we also tested the compound 7 (Fig. 4A). For the clarity, the
data with G418 are omitted in panels A–C and is only shown in pa-
nel D for CFTR-W1282X nonsense mutation, in comparison to that
of compound 8. Very similar preference of G418 over that of 8 was
also observed in PCDH15-R3X and IDUA-Q70X nonsense mutations
(data not shown). In order to ensure suitable cell viability for each
of the tested compounds at the concentrations tested, we deter-
mined for each compound the half-maximal-lethal concentration
value (LC50 values) in HEK-293cells (Table 1).
In all the mutations tested, the observed efficacy of aminoglyco-
side-induced readthrough was in the order G418 > 8 > 2 > gentami-
cin. This trend for 8 and 2 in Q70X and W1282X (Fig. 4B and C) was
very similar to that observed for the suppression of the same stop
mutations in vitro (Fig. 3). However, the observed higher read-
through potency of 8 over that of 2 in R3X was opposite to that
of the same mutation in vitro. Interestingly, compound 7, while
exhibited significantly lower suppression activity than that of com-
pound 2 in R3X mutation in vitro, in cell line both showed very
similar activity. While these data may point to a better cell perme-
ability of 7 and 8 over that of 2, due to the presence of the (R)-60-
methyl group, it is clear that more experiments are needed to
extract this issue satisfactorily. Nevertheless, the observed similar
cell toxicity of the compounds 8 and 2 in HEK-293 cells (Table 1),
along with substantially elevated suppression activity of 8 over
that of 2 both in vitro and ex vivo in cultured cells, indicate that
the compound 8 may represent a more superior choice than com-
pound 2 in suppression therapy.
Furthermore, the comparative ex vivo suppression data in
Fig. 4D, shows only about twofold preference of G418 over that
of compound 8, while the cell toxicity data in Table 1 indicate that
G418 is about fivefold more toxic than 8. Additional comparative
toxicity tests, including nephrotoxicity and ototoxicity, the major
drawbacks of aminoglycosides24 are underway to validate the ob-
served benefits of the compound 8 over that of G418 and
gentamicin.
Figure 3. In vitro stop codon suppression levels induced by compounds 1 (s), 2
(d), 7 (h) 8 (j) and gentamicin (N) in a series of nonsense mutation context
constructs representing various genetic diseases (shown in parenthesis): (A) R3X
(USH1), (B) (USH1), (C) G542X (CF), (D) W1282X (CF), (E) R3381X (DMD), and (F)
Q70X (HS). DNA fragments were cloned between BamHI and SacI restriction sites of
the p2luc vector and the obtained constructs were transcribed and translated using
TNT quick coupled transcription/translation system.26 The amount of the translated
products was evaluated using the dual luciferase reporter assay system and the
suppression level was calculated as described previously.35 The results are averages
of at least three independent experiments.
2.4. Compounds 7 and 8 inhibit prokaryotic protein translation
with significantly lower potency and exhibit markedly reduced
bactericidal activity and cell toxicity than those of gentamicin
and G418
In our previous studies,23,25,26 we have shown that compounds
1 and 2 are about 10-fold weaker inhibitors of prokaryotic transla-
tion than gentamicin and paromomycin and exhibit almost no

3740
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
in vitro luciferase assay (Table 1). The measured half-maximal
inhibitory concentration (IC50) values show that the efficacy with
which both 7 and 8 inhibit the prokaryotic ribosome is significantly
lower than that of gentamicin and G418. These data are in accor-
dance with the observed antibacterial data of this set of com-
pounds (Table 1). Thus, while gentamicin and G418 exhibit
excellent antibacterial activities against both Gram-negative Esch-
erichia coli (R477-100) and Gram-positive Bacillus subtilis (ATCC-
6633), both 7 and 8 lack significant antibacterial activity. The ob-
served data with 7 and 8 is similar to that observed for 1 and 2
and further support the previously reported correlation in amino-
glycosides between prokaryotic antitranslational activity and MIC
values: decreased inhibition of prokaryotic translation is associ-
ated with the decrease in antibacterial activity.36
We have previously reported that the compounds 1 and 2 exhi-
bit significantly reduced cytotoxicity than those of gentamicin and
paromomycin, as measured in three different kidney-derived
cells.26 To assess whether the compounds 7 and 8 retain similar
properties, we determined comparative cell toxicity of 1, 2, 7, 8,
G418 and gentamicin in HEK-293 cells (Table 1). The LC50 values
obtained for 7 and 8 (25.8 and 7 mM) were similar to those for 1
and 2 (21.4 and 6.1 mM), and were 10- and 2.8-fold higher than
that of the clinical aminoglycoside gentamicin (LC50 = 2.5 mM)
Among all aminoglycosides tested the aminoglycoside antibiotic
G418, which is known as one of the most cytotoxic aminoglyco-
side,33 exhibited the lowest LC50 value (1.3 mM).
Comparison of the observed cell toxicity data in Table 1 with
the readthrough activity data in Figure 3, demonstrates that the
introduction of (R)-60-methyl group either on compound 1 to give
compound 7 (1?7), or on compound 2 to give compound 8 (2?8),
not affects on the cytotoxicity (LC50 values of 21.4 and 25.8 mM for
1 and 7, and 6.1 and 7.0 mM for 2 and 8, respectively), while it
greatly increases the observed stop codon suppression activity
(1 < 7 and 2 < 8). This observation is one of the key outcomes of
the present work. Furthermore, these data in conjunction with
those observed for G418 and gentamicin indicate that the two un-
ique structural elements of G418, the secondary (R)-60-OH and the
garosamine ring (ring III, Fig. 1), contribute differently to its bio-
logical performance. The secondary (R)-60-OH most contributes
to its elevated readthrough potency (G418 shows far higher sup-
pression activity than gentamicin, Fig. 4C and D), while the garos-
amine ring donates the largest part to its high cytotoxicity (G418
shows only twofold lower LC50 than gentamicin in HEK-293 cells,
Table 1). Gentamicin consists of the same garosamine ring at the
same C6 position as G418 (Fig. 1), and essentially similar differ-
ence to that of HEK-293 cells was also observed in cytotoxicity as-
say in human fibroblast cells (3.8 mM and 2.0 mM for gentamicin
and G418, respectively, see Section 5 and Fig. 2S in Supplementary
data).
The impact of (R)-60-methyl group on the elevated readthrough
activities of 7, 8 and G418 is further supported from the observed
eukaryotic antitranslational data (Table 1). The efficacy with
Figure 4. Ex vivo suppression of the (A) PCDH15-R3X, (B) IDUA-Q70X, and (C) CFTR-
W1282X nonsense mutations by gentamicin (h), 2 (j), 7 (s), and 8 (d). The data
with compound 7 are given only for PCDH15-R3X mutation in the panel A. Panel D
which G418 (IC50 = 2.0
lM) inhibits eukaryotic translation is
greater than that of 8 (IC50 = 2.8
l
M) and 7 (IC50 = 17.0 M), a
l
represents comparative data between G418 and compd
8 for CFTR-W1282X
similar trend to that observed for readthrough activity: G418 >
8 > 7 (Fig. 4). Furthermore, the comparison of IC50 values of 7
and 8 to those of their parent structures 1 and 2 (IC50 values of
mutation. The constructs of p2luc plasmid harboring the R3X, Q70X and W1282
mutations were transfected to HEK-293 cells and addition of the tested compounds
was performed 6 h post transfection. Luciferase activity was determined using the
Dual Luciferase Reporter Assay System (Promega™). Stop codon readthrough was
calculated as described previously.35 The results are averages of at least three
independent experiments.
31 and 24 lM, respectively), reveals that 7 and 8 are 1.8-fold
and 8.6-fold more specific to the eukaryotic ribosome than their
parents 1 and 2. Finally, although gentamicin contains 60-NH2,
similar comparison between gentamicin (IC50 = 62
lM) and G418
(IC50 = 2.0 M), reveals that G418 is 31-fold more specific than
l
bactericidal activity against both Gram-negative and Gram-posi-
tive bacteria. To assess whether the compounds 7 and 8 retain sim-
ilar properties, we conducted comparative translation inhibition of
1, 2, 7, 8, G418 and gentamicin in a prokaryotic system, by using an
gentamicin. This collective data indicate that the observed impact
of (R)-60-methyl group on the elevated readthrough activities of 7,
8 and G418 is associated with their increased specificity to the
eukaryotic ribosome.

I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
3741
Table 1
Comparative cell toxicity, eukaryotic and prokaryotic translation inhibition, and antibacterial activity data of gentamicin, G418 and synthetic compounds 1, 2, 7 and 8a
b
Aminoglycoside
Cell toxicity LC50 (mM)
Translation inhibitionc
Eukaryotic system IC50 M) Prokaryotic system IC50 (nM)
62 28
2.0 0.3
Antibacterial activity MICd
(lM)
(l
E. coli R477/100
B. subtilis ATCC6633
Gentamicin
G418
Compound 1
Compound 2
Compound 7
Compound 8
2.5 0.3
1.3 0.1
21.4 3.9
6.1 0.6
25.8 2.8
7.0 0.6
9
4
6
9
790
588
680
556
<0.75
<1.25
100
70
42
70
8.9 1.5
460 50
31
24
4
1
160 20
17.0 0.6
2.8 0.3
1130 120
980 70
a
In all biological tests, all tested aminoglycosides were in their sulfate salt forms. The concentrations reported refer to that of the free amine form of each aminoglycoside.
Aminoglycosides-induced cell toxicity was measured in human embryonic kidney cells, HEK-293, and calculated as a ratio between the number of living cells in cultures
b
grown in the presence of the tested compound, versus cultures grown without compound. The half-maximal lethal concentration (LC50) values were obtained from fitting
concentration–response curves to the data of at least two independent experiments, using GraFit 5 software.
c
Prokaryotic and eukaryotic translation inhibition was quantified in coupled transcription/translation assays as previously described by us.23 The half-maximal con-
centration (IC50) values were obtained from fitting concentration response curves to the data of at least two independent experiments, using GraFit 5 software.
d
The MIC values were determined by using the double-microdilution method, with 384 lg/mL as starting concentration of each tested compound. All the experiments
were performed in duplicates and analogous results were obtained in three different experiments.
by installing it on compound 1 generated compound 2 as the sec-
3. Discussion
ond-generation lead structure.
Further efforts within the above stream, which are described in
this study lead us to the development of new lead compounds 7
and 8 (Fig. 5). Here we discovered a new pharmacophore, (R)-60-
methyl group, which we ‘borrowed’ from the natural aminoglyco-
side G418. The impact of (R)-60-methyl group on the elevated
readthrough activity was first demonstrated on pseudo-disaccha-
rides 3–6 (Fig. 2). Comparative in vitro suppression tests clearly
indicated that the installation of (R)-60-methyl group either on 3
or 4 leads to substantially increased readthrough efficiency of the
resulted structures 5 and 6 compared to those of their parent struc-
tures (3 and 4). Encouraged from the observed data we then in-
In the last several years, it was shown that besides their use as
antibiotics, aminoglycosides could have therapeutic value in the
treatment of human genetic disorders caused by nonsense muta-
tions.24,31 However, the most critical factor that largely limits the
potential of aminoglycosides for suppression therapy is their high
human toxicity.24 Despite its high toxicity, the clinical drug genta-
micin was/is frequently used for proof-of-concept experiments in
various diseases models and in clinical trials, and no systematic
study has been performed to tune aminoglycosides structures for
better readthrough activity and low toxicity. Recently, we hypoth-
esized that by separating the structural elements of aminoglyco-
sides that induce readthrough from those that affect toxicity we
might reach potent derivatives with improved readthrough activ-
ity and reduced toxicity.24–26 By keeping this line of research, we
systematically developed the lead compounds 125 and 2.26 We ini-
tially identified the pseudo-disaccharide 3 (paromamine) as a min-
imal core structure exhibiting significant readthrough activity, and
by attaching 5-amino ribose as a ring III spotted compound 1 as the
first lead structure (Fig. 5). Continuing our efforts, we then ‘bor-
rowed’ the pharmacophore AHB from the butirosin/amikacin and
stalled (R)-60-methyl group onto
1 and 2 to construct the
pseudo-trisaccharides 7 and 8, which were investigated in more
details. The results of the current study show that both 7 and 8
exhibit substantially greater suppression activity and far lower
toxicity than those of gentamicin. The evidence for the superior
readthrough efficiency of 7 and 8 over that of gentamicin was dem-
onstrated in vitro on six different DNA fragments derived from the
mutant PCDH15, CFTR, Dystrophin and IDUA genes carrying non-
sense mutations and representing the underlying causes for the ge-
netic diseases USH1, CF, DMD and HS, respectively (Fig. 3), and
ex vivo in cultured cell lines on three different DNA fragments that
model the genetic diseases USH1, CF, and HS (Fig. 4). Furthermore,
compound 8 also exhibits several-fold higher suppression activity
than that of the previous lead compound 2, while both (8 and 2)
display similar cytotoxicity in HEK-293 cells (Table 1). Essentially
the same trend was also observed in comparative study between
the compounds 7 and 1. The combined data suggests that installa-
tion of (R)-60-methyl group on aminoglycoside structure to yield
the (R)-60-secondary alcohol on ring I, significantly increases
suppression activity while has no significant influence on the cell
toxicity of the resulted derivative.
Third generation (compounds 7 and 8)
(R)-6'-methyl
e
M
OH
O
6'
Paromamine
NH2
HO
HO
H2N
O
NH2
1
HO
5
OH
Earlier studies have shown that aminoglycosides with a 60-OH
group on ring I (such as G418 and paromomycin, Fig. 1) are more
effective than those with an amine at the same position.31,37 One
of the key differences between the prokaryotic and eukaryotic
A-site is the nucleotide in the 1408 position: an adenine in the pro-
karyotes and a guanine in the eukaryotes. An A1408G mutation in
various engineered bacteria leads to enhanced resistance towards
aminoglycosides, but much higher levels of resistance are observed
towards aminoglycosides with 60-NH2 than towards those with 60-
OH,29,30 in agreement with the suppression results obtained in the
mammalian system. The crystal structures of the bacterial A-site in
complex with different aminoglycosides show that indeed the 60
functional groups form key H-bonds with the A1408 base, and
OH
H2N
NH2
5''
O
O
AHB
Second generation
First generation
(compound 1)
(compound 2)
HO OH
Figure 5. Chronological development of lead compounds on the paromamine (3)
scaffold by systematic optimization of structure–activity–toxicity relationship as
follow: first-generation compound 1, by attaching 5-amino ribose at C5 position;25
second-generation compound 2, by further attaching AHB group at N-1 position of
the compound 1;26 third-generation compounds 7 and 8, by installing (R)-60-Me on
1 and 2, respectively (current work).

3742
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
models of the A1408G mutations based on these structures suggest
that this mutation would prevent such interaction completely in
60-NH2 molecules.30,38 A close check of these earlier data29,30 re-
vealed that although only paromomycin and G418 conferred sig-
nificantly low resistance to the bacteria harboring A1408G
mutation, there was noteworthy difference in the level of resis-
tance between the paromomycin and G418: the relative resis-
tances observed in different studies were 1–8-fold for G418 and
16–80-fold for paromomycin.29,30,39 Based on this difference, we
hypothesized that the presence of the secondary (R)-60-OH group
in G418 versus that of primary 60-OH in paromomycin could, in
principle, contribute to the observed significantly reduced resis-
tance of G418 versus that of paromomycin.40 This hypothesis is
validated by our current data broadly showing that the compounds
having secondary (R)-60-OH group suppress premature stop codons
more efficiently than their parent structures with the primary 60-
OH. The observed higher specificity to the eukaryotic ribosome,
of the compounds containing (R)-60-OH group than their parent
structures with the primary 60-OH, further supports this validation
(Table 1). Finally, the results of this study also suggest that the
structural element that most likely contributes to the extremely
high readthrough activity of G418 is the (R)-60-OH group, while
the presence of C6-linked garosamine ring (ring III) primarily do-
nates to its elevated cytotoxicity.
HOD (d = 4.63) with D2O as the solvent. 13C NMR spectra were re-
corded on a Bruker Avance™ 500 spectrometer at 125.8 MHz, and
the chemical shifts reported (in ppm) relative to the residual sol-
vent signal for CDCl3 (d = 77.00), or to external sodium 2,2-di-
methyl-2-silapentane sulfonate (d = 0.0) for D2O as the solvent.
Mass spectra analysis were obtained either on a Bruker Daltonix
Apex 3 mass spectrometer under electron spray ionization (ESI),
or by a TSQ-70B mass spectrometer (Finnigan Mat). Reactions were
monitored by TLC on Silica Gel 60 F254 (0.25 mm, Merck), and spots
were visualized by charring with a yellow solution containing
(NH4)Mo7O24ꢀ4H2O (120 g) and (NH4)2Ce(NO3)6 (5 g) in 10%
H2SO4 (800 mL). Flash column chromatography was performed
on Silica Gel 60 (70–230 mesh). All reactions were carried out un-
der an argon atmosphere with anhydrous solvents, unless other-
wise noted. G418 (geneticin), paromomycin and gentamicin were
purchased from Sigma. All other chemicals and biochemicals, un-
less otherwise stated, were obtained from commercial sources. In
all biological tests, all the tested aminoglycosides were in their sul-
fate salt forms [Mw (g/mol) of the sulfate salts were as follow:
compd 3 (paromamine)—408.1, compd 4—546.9, compd 5—437.1,
compd 6—533.1, G418—692.7; gentamicin—653.2; compd
1
(NB30)—563.0; compd 2 (NB54)—652.8; compd 7 (NB74)—564.3;
compd 8 (NB84)—685.9].
5.2. Synthetic part
4. Summary and conclusions
5.2.1. 60-(R)-Methyl-1,20,3-triazido-paromamine (9)
The titled compound was prepared by direct acid promoted
cleavage of G418. To a stirred anhydrous methanol (50 mL) was
added acetyl chloride (25 mL) over 10 min at 0 °C. After about
15 min stirring, the commercial G418 sulfate (5.0 g, 7.20 mmol)
was added and the reaction was heated to reflux. Propagation of
the reaction was monitored by TLC [CH2Cl2/MeOH/H2O/MeNH2
(33% solution in EtOH), 10:15:6:15], which indicated completion
after 16 h. The reaction mixture was evaporated under vacuum
to dryness and then purified by flash chromatography. The column
was washed with the mixture of CH2Cl2/MeOH (1:1, 200 mL), fol-
lowed by MeOH (500 mL). The product was eluted with the mix-
ture of 5% NH4OH in MeOH/H2O [(25% NH4OH solution in
H2O):(15% H2O):(80% MeOH)] to yield compound 5 as a powder
(1.8 gr, 75%). NMR (500 MHz, D2O, pD 3.5): ‘Ring I’: dH 1.10 (d,
3H, J = 6.5 Hz, CH3-6), 3.34 (dd, 1H, J1 = J2 = 10.0 Hz, H-4), 3.36
(dd, 1H, J1 = 4.0, J2 = 10.5 Hz, H-2), 3.80 (dd, 1H, J1 = 9.0,
J2 = 10.0 Hz, H-3), 3.78–3.82 (m, 1H, H-5), 4.11–4.15 (m, 1H, H-6),
5.50 (d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH 1.79 (ddd, 1H,
J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.41 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-
2eq), 3.21–3.26 (m, 1H, H-1), 3.48–3.52 (m, 1H, H-3), 3.51 (dd,
1H, J1 = J2 = 9.5 Hz, H-6), 3.57 (dd, 1H, J1 = J2 = 9.0 Hz, H-5), 3.77
(dd, 1H, J1 = J2 = 9.5 Hz, H-4). 13C (NMR 125 MHz, D2O): dC 16.2
(CH3-60), 30.0 (C-2), 50.9, 51.4, 55.9, 67.0, 71.2, 71.9, 74.0, 76.2,
77.0, 83.3, 99.3 (C-10); MALDI TOFMS calcd for C13H27N3O7
([M+H]+) m/e 338.2; measured m/e 338.2)
We would like to note that the product from the above step
contained about 1% of the starting G418 along with other impuri-
ties (ꢁ1%, originally exist in the starting material). These impuri-
ties could not be completely removed by repeated column
purifications. Therefore, the product from the above step (1.8 g,
5.30 mmol) was converted to the corresponding 1,3,20-triazido
derivative by following the published procedure,44 using Tf2O
(8.0 mL, 48.0 mmol) and NaN3 (3.13 g, 48.0 mol). The reaction pro-
gress was monitored by TLC (EtOAc/MeOH 95:5), which indicated
completion after 8 h. The resulted crude was purified by flash chro-
matography (EtOAc/hexane 95:5) to yield compound 9 (1.98 g, 90%
yield). NMR (500 MHz, MeOD): ‘Ring I’: dH 1.24 (d, 3H, J = 6.0 Hz,
CH3-6), 3.07 (dd, 1H, J1 = 4.0, J2 = 10.0 Hz, H-2), 3.36 (dd, 1H,
J1 = 8.5, J2 = 9.5, H-4), 3.92 (dd, 1H, J1 = J2 = 9.5 Hz, H-3), 3.92–
In summary, the newly developed aminoglycosides compounds
7 and 8 exhibit both appreciably higher suppression efficiency and
lower toxicity than those of gentamicin. This achievement has sev-
eral significant implications. First, the results of this study further
validates our design hypothesis that by separating the structural
elements of aminoglycosides that induce readthrough from those
that affect toxicity we can reach potent derivatives with improved
readthrough activity and reduced toxicity.24–26 Second, the ob-
served continued inability of our previous leads, 125 and 2,26 along
with current 7 and 8, to show significant antibacterial activity in
conjunction with their decreased prokaryotic ribosome specificity,
are striking and remain to be further investigated. Of greatest con-
cern is whether these data are linked or not to the observed rela-
tively reduced cell toxicity potential of 1, 2, 7 and 8. This view is
supported by the fact that the mammalian mitochondrial protein
synthesis machinery is very similar to the prokaryotic machinery41
and that the aminoglycoside-induced toxicity may, at least in part,
be connected to drug-mediated dysfunction of the mitochondrial
ribosome.42,43 Third, the observed high impact of the (R)-60-sec-
ondary alcohol on the elevated readthrough activity and specificity
to the eukaryotic ribosome, along with no significant influence on
the cell toxicity, support the feasibility of testing this novel phar-
macophore for the generation of novel structures that may act as
drug for the treatment of genetic diseases. Finally, the observed
relatively low toxicity and high degree of potency of the new gen-
eration structures 7 and 8 in targeting all six different nonsense
constructs underlying USH1, CF, DMD and HS, support the feasibil-
ity of testing these novel aminoglycosides in treating these dis-
eases in animal and human subjects.
5. Experimental
5.1. General techniques
NMR spectra (including 1H, 13C, DEPT, 2D COSY, 1D TOCSY,
HMQC, HMBC) were routinely recorded on a Bruker Avance™
500 spectrometer, and chemical shifts reported (in ppm) are rela-
tive to internal Me4Si (d = 0.0) with CDCl3 as the solvent, and to

I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
3743
3.95 (m, 1H, 1H, H-5), 4.01–4.04 (m, 1H, H-6), 5.72 (d, 1H,
J = 4.0 Hz, H-1); ‘Ring II’: dH 1.37 (ddd, 1H, J1 = J2 = J3 = 12.5 Hz, H-
2ax), 2.22 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-2eq), 3.24 (dd, 1H,
J1 = J2 = 9.5 Hz, H-6), 3.38–3.43 (m, 1H, H-1), 3.48–3.52 (m, 1H, H-
3), 3.38–3.43 (m, 1H, H-5), 3.47–3.53 (m, 1H, H-4). 13C (NMR
125 MHz, MeOD): dC 16.8 (CH3-60), 32.2 (C-2), 59.9, 60.6, 63.5,
68.1, 71.2, 73.0, 74.1, 76.8, 76.8, 78.8, 97.5 (C-10); ESI calcd for
C13H20N9O7 ([MꢂH]ꢂ) m/e 414.2; measured m/e 414.1).
J1 = J2 = J3 = 12.5 Hz, H-2ax), 1.89–1.96 (m, 1H, H-2eq), 2.75–2.83
(m, 1H, H-3), 3.18 (dd, 1H, J1 = J2 = 9.0 Hz, H-4), 3.22 (dd, 1H,
J1 = J2 = 9.0 Hz, H-6), 3.32–3.39 (m, 1H, H-1), 3.47 (dd, 1H,
J1 = J2 = 9.0 Hz, H-5); the additional peak in the spectrum was iden-
tified as follow: dH 1.34 (s, 9H, Boc). 13C NMR (125 MHz, D2O): dC
16.7 (CH3-60), 29.4 (Boc, 3C), 36.2 (C-2), 51.0, 57.0, 67.9, 71.6,
72.9, 75.5, 76.4, 76.5, 77.6, 88.7, 102.6 (C-10), 159.7 (Boc, CO). MAL-
DI TOFMS calcd for C18H35N3O9 ([M+H]+ m/e 438.2; measured m/e
438.2).
5.2.2. 60-(R)-Methyl-paromamine (5)
The product from the above step (5.5 g, 12.60 mmol) was con-
verted to the corresponding 3,20-diazido derivative by following
the published procedure,44 using Tf2O (20 mL, 120 mmol) and
NaN3 (10 g, 153 mmol). The reaction progress was monitored by
TLC (EtOAc/MeOH 9:1), which indicated completion after 8 h. The
crude product was purified by flash chromatography (EtOAc/hex-
ane 85:15) to yield the title compound 10 (5.4 g, 88% yield). 1H
NMR (500 MHz, CDCl3): ‘Ring I’: dH 1.03 (d, 3H, J = 6.5 Hz, CH3-6),
2.99 (dd, 1H, J1 = 4.0, J2 = 10.0 Hz, H-2), 3.20 (dd, 1H, J1 = 9.0,
J2 = 9.5, H-4), 3.59–3.62 (m, 1H, H-5), 3.72 (dd, 1H, J1 = J2 = 9.5 Hz,
H-3), 3.76–3.80 (m, 1H, H-6), 5.33 (d, 1H, J = 4.0 Hz, H-1); ‘Ring
II’: dH 1.18 (ddd, 1H, J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.08–2.16 (m,
H-2eq), 2.98 (dd, 1H, J1 = J2 = 9.5 Hz, H-6), 3.20–3.28 (m, 2H, H-1,
H-3), 3.20 (dd, 1H, J1 = J2 = 9.5 Hz, H-4), 3.31 (dd, 1H,
J1 = J2 = 9.0 Hz, H-5); the additional peaks in the spectrum were
identified as follow: dH 1.23 (s, 9H, Boc). 13C NMR (125 MHz,
CDCl3): dC 17.4 (CH3-60), 27.7 (Boc, 3C), 32.9 (C-2), 48.8 (C-1),
59.4, 62.9, 68.6, 71.1, 73.0, 73.2, 74.2, 76.6, 79.9, 97.3 (C-10),
159.3 (Boc, CO). MALDI TOFMS calcd for C18H32N7O9 ([M+H]+) m/
e 490.2; measured m/e 490.3).
Compound 9 (56 mg, 0.14 mmol) was dissolved in a mixture of
THF (3.0 mL) and aqueous NaOH (1 mM, 3.0 mL). This mixture was
stirred at room temperature for 10 min, after which PMe3 (1 M
solution in THF, 1.1 mL, 1.10 mmol) was added. The reaction pro-
gress was monitored by TLC [CH2Cl2/MeOH/H2O/MeNH2 (33% solu-
tion in EtOH), 10:15:6:15], which indicated completion after 1 h.
The reaction mixture was purified by flash chromatography on a
short column of silica gel. The column was washed with the follow-
ing solvents: THF (100 mL), CH2Cl2 (100 mL), EtOH (50 mL), and
MeOH (100 mL). The product was then eluted with the mixture
of 20% MeNH2 solution (33% solution in EtOH) in 80% MeOH. Frac-
tions containing the product were combined and evaporated under
vacuum. The residue was re-dissolved in a small volume of water
and evaporated again (2–3 repeats) to afford the free amine form
of compound 5. The analytically pure product was obtained by
passing the above product through a short column of Amberlite
þ
CG50 (NH4 form). First, the column was washed with water, then
the product was eluted with a mixture of 10% NH4OH in water to
yield compound 5 (32.0 mg, 70%). For the storage and biological
tests, compound 5 was converted to its sulfate salt form as follow.
The free base form was dissolved in water, the pH was adjusted to
6.7 with H2SO4 (0.1 N) and lyophilized to afford the sulfate salt of 5
as a white foamy solid. 1H NMR (500 MHz, D2O, pD 3.5) ‘Ring I’: dH
1.10 (d, 3H, J = 6.5 Hz, CH3-6), 3.34 (dd, 1H, J1 = J2 = 10.0 Hz, H-4),
3.36 (dd, 1H, J1 = 4.0, J2 = 10.5 Hz, H-2), 3.80 (dd, 1H, J1 = 9.0,
J2 = 10.0 Hz, H-3), 3.78–3.82 (m, 1H, H-5), 4.11–4.15 (m, 1H,
H-6), 5.50 (d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH 1.79 (ddd, 1H,
J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.41 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-
2eq), 3.21–3.26 (m, 1H, H-1), 3.48–3.52 (m, 1H, H-3), 3.51 (dd,
1H, J1 = J2 = 9.5 Hz, H-6), 3.57 (dd, 1H, J1 = J2 = 9.0 Hz, H-5), 3.77
(dd, 1H, J1 = J2 = 9.5 Hz, H-4). 13C (NMR 125 MHz, D2O): dC 16.2
(CH3-60), 30.0 (C-2), 50.9, 51.4, 55.9, 67.0, 71.2, 71.9, 74.0, 76.2,
77.0, 83.3, 99.3 (C-10); MALDI TOFMS calcd for C13H27N3O7
([M+H]+) m/e 338.2; measured m/e 338.2).
5.2.4. 20,3-Diazido-1-N-[(S)-4-azido-2-hydroxybutanoyl]-60-(R)-
methyl-paromamine (11)
Compound 10 (1.3 g, 2.76 mmol) was dissolved in a mixture of
trifluoroacetic acid (12 mL) and dichloromethane (30 mL). The
reaction progress was monitored by TLC (CH2Cl2/MeOH 8:2),
which indicated completion after 1 h. The reaction mixture was
concentrated to dryness under reduced pressure. The resulting
crude was dissolved in a mixture of Et3N (10 mL) and DMF
(10 mL) and cooled to ꢂ30 °C. In a separate flask, (S)-2-hyroxy-
4-azidobutyric acid (2.0 g, 15.78 mmol) was dissolved in anhy-
drous DMF (30 mL) and cooled to 0 °C. To the cold solution, DCC
(3.6 g, 17.23 mmol) and HOBT (2.4 g, 17.48 mmol) were added
and the resulted mixture was stirred at 0 °C for about 1 h. This
mixture was carefully added by syringe to the cold solution of
the amine at ꢂ30 °C. The reaction was stirred at ꢂ30 °C for 1 h
and then allowed to warm to room temperature for an additional
1 h. The mixture was then treated with a solution of MeNH2 (33%
solution in EtOH, 30 mL) and the reaction progress was monitored
by TLC (CH2Cl2/MeOH 8:2). After completion of the reaction
(about 8 h) the mixture was concentrated and purified by flash
chromatography (MeOH/CH2Cl2 1:9) to provide the titled com-
pound 11 (1.0 g, 70% yield). 1H NMR (500 MHz, MeOD): ‘Ring I’:
dH 1.24 (d, 3H, J = 5.0 Hz, CH3-6), 3.20 (dd, 1H, J1 = 4.5,
J2 = 10.0 Hz, H-2), 3.42 (dd, 1H, J = 9.5 Hz, H-4), 3.81–3.85 (m,
1H, H-5), 3.94 (dd, 1H, J1 = J2 = 10.0 Hz, H-3), 3.95–3.99 (m, 1H,
H-6), 5.53 (d, 1H, J = 3.0 Hz, H-1); ‘Ring II’: dH 1.45 (ddd, 1H,
J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.24–2.29 (m, 1H, H-2eq), 3.34 (dd,
1H, J1 = J2 = 9.5 Hz, H-6), 3.41 (dd, 1H, J1 = J2 = 9.5 Hz, H-5), 3.42–
3.48 (m, 1H, H-3), 3.74 (dd, 1H, J1 = J2 = 9.5 Hz, H-4), 3.73–3.78
(m, 1H, H-1); the additional peaks in the spectrum were identified
as follows: d 1.81–1.89 (m, 1H), 2.05–2.10 (m, 1H), 3.42–3.46 (m,
2H), 4.14 (dd, 1H, J1 = 4.0, J2 = 10.5 Hz). 13C (NMR 125 MHz,
MeOD): dC 17.6 (CH3-60), 32.3 (C-2), 32.8, 47.2, 50.5, 50.7, 59.4,
68.7, 68.8, 71.1, 73.0, 73.4, 73.8, 76.5, 80.0, 97.3 (C-10), 174.7
(CO). MALDI TOFMS calcd for C17H29N10O9 ([M+H]+) m/e 516.2;
measured m/e 516.2).
5.2.3. 20,3-Diazido-1-N-[(tert-butoxy)carbonyl]-60-(R)-methyl-
paromamine (10)
Zn(OAc)2 (6.0 g, 27.00 mmol) was added to a stirred solution of
compound 5 in its free base form (7.0 g, 21.00 mmol) in H2O
(30 mL) and DMF (150 mL), and the mixture was stirred for 12 h
at room temperature.
A solution of di-tert-butyldicarbonate
(6.0 g, 27.00 mmol) in DMF (20 mL) was added to the reaction mix-
ture over a time period of 30 min and the mixture was stirred for
an additional 24 h. The reaction progress was monitored by TLC
[CH2Cl2/MeOH/H2O/MeNH2 (33% solution in EtOH), 10:15:6:15].
The reaction mixture was diluted with MeOH (250 mL) and loaded
onto 50 mm ꢃ 300 mm ion-exchange column (Amberlite CG50, H+
form). The column was first washed extensively (about 10 column
volumes) with a mixture MeOH/H2O (50:50), followed by elution
with the mixture of MeOH/H2O/NH4OH (25% solution in water)
(20:3:2) to afford the desired N-1-Boc derivative in 60% yield
(5.5 g). 1H NMR (500 MHz, D2O): ‘Ring I’: dH 1.12 (d, 3H,
J = 6.5 Hz, CH3-6), 2.73 (dd, 1H, J1 = 4.0, J2 = 10.5 Hz, H-2), 3.25
(dd, 1H, J1 = J2 = 9.0 Hz, H-4), 3.49 (dd, 1H, J1 = J2 = 9.0 Hz, H-3),
3.76 (dd, 1H, J1 = 2.0, J2 = 10.0 Hz, H-5), 4.03–4.08 (m, 1H, H-6),
5.14 (d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH 1.20 (ddd, 1H,

3744
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
5.2.5. 20,3-Diamino-1-N-[(S)-4-amino-2-hydroxybutanoyl]-60-
(R)-methyl-paromamine (6)
(m, 2H), 2.06 (s, 3H, Ac), 2.08 (s, 3H, Ac), 2.09 (s, 3H, Ac), 2.15 (s,
3H, Ac), 2.19 (s, 3H, Ac), 3.36–3.40 (m, 2H), 5.14 (dd, 1H, J1 = 5.0,
J2 = 6.5 Hz). 13C NMR (125 MHz, CDCl3) dC 14.0 (CH3-60), 20.6–
21.0 (Ac, 5C), 30.5, 32.6 (C-2), 47.0, 48.3, 58.3, 61.7, 68.8, 69.0,
70.9, 70.9, 71.4, 74.0, 74.1, 83.6, 98.5 (C-10), 169.1, 169.8, 170.0,
The titled compound was prepared as was described for the
preparation of compound 5 with the following quantities: com-
pound 11 (70.0 mg, 0.14 mmol), THF (3 mL), NaOH (1 mM,
3.0 mL), PMe3 (1 M solution in THF, 2.00 mL, 2.00 mmol) to yield
170.0, 170.1, 172.5. MALDI TOFMS calcd for C27H38N10O14
K
compound
6
as
a
free amine form (39 mg, 65%). 1H NMR
([M+K]+) m/e 765.2; measured m/e 765.2).
(500 MHz, D2O, pD 6.5): ‘Ring I’: dH 1.14 (d, 3H, J = 7.0 Hz, CH3-6),
3.18 (dd, 1H, J1 = 4.0, J2 = 10.5 Hz, H-2), 3.33 (dd, 1H,
J1 = J2 = 9.5 Hz, H-4), 3.71 (dd, 1H, J1 = 10.5, J2 = 10.0 Hz, H-3),
3.82–3.85 (m, 1H, H-5), 4.11–4.15 (m, 1H, H-6), 5.39 (d, 1H,
J = 4.0 Hz, H-1); ‘Ring II’: dH 1.73 (ddd, 1H, J1 = J2 = J3 = 12.5 Hz, H-
2ax), 2.09–2.13 (m, 1H, H-2eq), 3.25–3.30 (m, 1H, H-3), 3.48 (dd,
1H, J1 = J2 = 9.5 Hz, H-6), 3.58 (dd, 1H, J1 = J2 = 9.5 Hz, H-5), 3.61
(dd, 1H, J1 = J2 = 9.5 Hz, H-4), 3.78–3.84 (m, 1H, H-1); the additional
peaks in the spectrum were identified as follows: d 1.90–1.97 (m,
1H), 2.07–2.10 (m, 1H), 3.06–3.14 (m, 2H), 4.25 (dd, 1H, J1 = 4.0,
J2 = 7.5 Hz). 13C (NMR 125 MHz, D2O): dC 14.5 (CH3-60), 30.8 (C-
2), 31.2, 36.5, 49.4, 50.9, 54.5, 65.5, 69.5, 70.4, 70.9, 73.5, 75.0,
75.3, 83.6, 98.5 (C-10), 175.5 (CO). MALDI TOFMS calcd for
C17H34N4O9K ([M+K]+) m/e 477.2; measured m/e 477.1).
5.2.8. 5-O-(5-Azido-2,3-O-dibenzoyl-5-deoxy-b-D-ribofuranose)-
30,40,60,6-tetra-O-actate-1,20,3-triazido-60-(R)-methyl-parom-
amine (15)
To powdered, flame-dried 4 Å molecular sieves (1.5 g) was
added anhydrous CH2Cl2 (12 mL), followed by the addition of
acceptor 12 (600 mg, 1.03 mmol) and donor 1425 (2.2 g, 4.06
mmol). The mixture was stirred for 10 min at room temperature
and was then cooled to ꢂ30 °C. A catalytic amount of BF3–Et2O
(100 lL) was added and the mixture was continued to stir at
ꢂ30 °C. The reaction progress was monitored by TLC (EtOAc/hex-
ane 35:65), which indicated completion after 15 min. The reaction
was diluted with EtOAc and filtered through Celite. After thorough
washing of the Celite with EtOAc, the washes were combined and
extracted with saturated aqueous NaHCO3, brine, dried over MgSO4
and concentrated. The crude product was purified by flash chroma-
tography to yield 15 (750 mg, 77%). 1H NMR (500 MHz, CDCl3):
‘Ring I’: dH 1.27 (d, 3H, J = 6.0 Hz, CH3-6), 3.50 (dd, 1H, J1 = 4.0,
J2 = 11.0 Hz, H-2), 4.45 (dd, 1H, J1 = 0.7, J2 = 10.5 Hz, H-5), 4.97
(dd, 1H, J1 = 9.5, J2 = 10.0 Hz, H-4), 4.98–5.02 (m, 1H, H-6), 5.44
(dd, 1H, J = 10.0 Hz, H-3), 5.94 (d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH
1.51 (ddd, 1H, J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.39 (dt, 1H, J1 = 4.0,
J2 = 13.0 Hz, H-2eq), 3.48–3.58 (m, 2H, H-1, H-3), 3.76 (dd, 1H,
J1 = J2 = 10.0 Hz, H-4), 3.91 (dd, 1H, J = 10.0 Hz, H-5), 5.03 (dd, 1H,
J1 = J2 = 10.0 Hz, H-6); ‘Ring III’: dH 3.58 (dd, 1H, J1 = 6.0,
J2 = 13.0 Hz, H-50), 3.64 (dd, 1H, J1 = 3.0, J2 = 13.0 Hz, H-5), 4.53–
4.56 (m, 1H, H-4), 5.42 (dd, 1H, J1 = 5.0, J2 = 10.0 Hz, H-3), 5.62 (d,
1H, J = 5.0 Hz, H-2), 5.64 (s, 1H, H-1); the additional peaks in the
spectrum were identified as follow: 2.05–2.09 (m, 12H, Ac), 7.36
(t, 2H, J = 8.0, Bz), 7.42 (t, 2H, J = 8.0 Hz, Bz), 7.53–7.55 (m, 1H,
Bz), 7.57–7.60 (m, 1H, Bz), 7.88 (dd, 2H, J1 = 1.0, J2 = 7.5 Hz, Bz),
7.95 (dd, 2H, J1 = 1.0, J2 = 7.5 Hz, Bz). 13C NMR (125 MHz, CDCl3):
dC 13.3 (CH3-60), 20.6–21.2 (Ac, 4C), 32.1 (C-2), 53.1 (C-500), 58.3,
58.9, 61.6, 68.5, 69.0, 70.2, 70.6, 72.0, 74.0, 74.5, 76.8, 80.0, 80.8,
96.2 (C-10), 107.6 (C-100), 128.5 (Bz, 2C), 128.6 (Bz, 2C), 129.6 (Bz,
2C), 129.7 (Bz, 2C), 133.7 (Bz, 2C), 165.2 (Bz, CO), 165.4 (Bz, CO),
5.2.6. 30,40,60,6-Tetra-O-actate-20,3-diazido-1-N-[(S)-4-azido-2-O-
acetate-butanoyl]-60-(R)-methyl-paromamine (12)
Compound 9 (900 mg, 2.20 mmol) was dissolved in dry pyridine
(10 mL), cooled at ꢂ12 °C and then acetic anhydride (4.5 equiv,
1.0 mL, 10.0 mmol) was added. The temperature was kept at
ꢂ12 °C and the reaction progress was monitored by TLC (EtOAc/
hexane 6:4), which indicated completion after 16 h. The reaction
mixture was diluted with EtOAc (100 mL) and extracted with
aqueous solutions of NaHCO3, HCl (2%), saturated aqueous NaH-
CO3, and brine. The combined organic layer was dried over MgSO4
and concentrated. The crude product was purified by flash chroma-
tography (EtOAc/hexane 3:7) to afford 12 (840 mg, 60% yield). 1H
NMR (500 MHz, CDCl3): ‘Ring I’: dH 1.24 (d, 3H, J = 6.0 Hz, CH3-6),
3.56 (dd, 1H, J1 = 3.5, J2 = 10.5 Hz, H-2), 4.31 (dd, 1H, J1 = 2.0,
J2 = 10.5 Hz, H-5), 4.92–4.98 (m, 1H, H-6), 4.96 (dd, 1H, J1 = 10.0,
J2 = 9.0 Hz, H-4), 5.43 (d, 1H, J = 4.0 Hz, H-1), 5.45 (dd, 1H,
J1 = 9.0, J2 = 10.0 Hz, H-3); ‘Ring II’: dH 1.56 (ddd, 1H,
J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.36 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-
2eq), 3.34–3.40 (m, 1H, H-1), 3.50–3.55 (m, 1H, H-3), 3.47 (dd,
1H, J1 = 9.0, J2 = 10.0 Hz, H-4), 3.67 (dd, 1H, J1 = 9.0, J2 = 9.5 Hz, H-
5), 4.89 (dd, 1H, J1 = J2 = 10.0 Hz, H-6); the additional peaks in the
spectrum were identified as follow: 2.07 (s, 3H, Ac), 2.10 (s, 3H,
Ac), 2.12 (s, 3H, Ac), 2.15 (s, 3H, Ac). 13C NMR (125 MHz, CDCl3):
dC 14.0 (CH3-60), 20.6–21.0 (Ac, 4C), 32.0 (C-2), 57.9, 58.4, 61.5,
68.7, 69.1, 70.8, 71.5, 74.5, 75.1, 82.4, 98.4 (C-10), 169.8, 170.0,
170.2, 170.5. MALDI TOFMS calcd for C21H29N9O11 Na ([M+Na]+)
m/e 606.2; measured m/e 606.3)
169.9–170.2 (4Ac, CO). MALDI TOFMS calcd for C40H44N12O16
K
([M+K]+) m/e 971.3; measured m/e 971.3).
5.2.9. 5-O-(5-Azido-2,3-O-dibenzoyl-5-deoxy-b-D-ribofuranose)-
30,40,60,6-tetra-O-actate-20,3-diazido-1-N-[(S)-4-azido-2-O-ace-
tate-butanoyl]-60-(R)-methyl-paromamine (16)
The titled compound was prepared as was described for the
preparation of compound 15 with the following quantities: com-
pound 13 (400 mg, 0.55 mmol), donor 14 (1.50 g, 2.84 mmol).
Yield: 550 mg (91%). 1H NMR (500 MHz, CDCl3): ‘Ring I’: dH 1.26
(d, 3H, J = 6.0 Hz, CH3-6), 3.46 (dd, 1H, J1 = 4.0, J2 = 10.5 Hz, H-2),
4.46 (dd, 1H, J1 = 2.0, J2 = 10.5 Hz, H-5), 4.96–5.01 (m, 1H, H-6),
5.00 (dd, 1H, J1 = J2 = 10.0 Hz, H-4), 5.44 (dd, 1H, J1 = 9.5,
J2 = 10.0 Hz, H-3), 5.92 (d, 1H, J = 3.5 Hz, H-1); ‘Ring II’: dH 1.58
(ddd, 1H, J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.50 (dt, 1H, J1 = 4.0,
J2 = 13.0 Hz, H-2eq), 3.59–3.64 (m, 1H, H-3), 3.74 (dd, 1H, J1 = 9.0,
J2 = 8.0 Hz, H-4), 3.98–4.03 (m, 1H, H-1), 4.00 (dd, 1H, J1 = 8.5,
J2 = 9.0 Hz, H-5), 4.92 (dd, 1H, J1 = 9.5, J2 = 10.0 Hz, H-6), 6.64 (d,
1H, J = 7.5 Hz, NH); ‘Ring III’: dH 3.60–3.68 (m, 2H, H-5), 4.52–
4.57 (m, 1H, H-4), 5.50 (dd, 1H, J1 = 5.0, J2 = 7.0 Hz, H-3), 5.63 (dd,
1H, J = 5.0 Hz, H-2), 5.67 (s, 1H, H-1); the additional peaks in the
spectrum were identified as follow: 2.03–2.15 (m, 2H), 2.08 (s,
3H, Ac), 2.09 (s, 6H, 2Ac), 2.19 (s, 3H, Ac), 2.23 (s, 3H, Ac), 3.35
5.2.7. 30,40,60,6-Tetra-O-actate-20,3-diazido-1-N-[(S)-4-azido-2-O-
acetate-butanoyl]-60-(R)-methyl-paromamine (13)
The titled compound was prepared as was described for the
preparation of compound 12 with the following quantities: com-
pound 11 (1.0 g, 1.94 mmol), pyridine (10 mL) and acetic anhy-
dride (5.5 equiv, 1.0 mL, 10.60 mmol). Yield: 1.1 g (70%). 1H NMR
(500 MHz, CDCl3): ‘Ring I’: dH 1.25 (d, 3H, J = 7.0 Hz, CH3-6), 3.62
(dd, 1H, J1 = 3.5, J2 = 10.5 Hz, H-2), 4.34 (dd, 1H, J1 = 2.0,
J2 = 10.5 Hz, H-5), 4.96–5.00 (m, 1H, H-6), 4.98 (dd, 1H, J1 = 10.5,
J2 = 9.0 Hz, H-4), 5.35 (d, 1H, J = 4.5 Hz, H-1), 5.50 (dd, 1H,
J1 = 9.0, J2 = 10.5 Hz, H-3); ‘Ring II’: dH 1.47 (ddd, 1H,
J1 = J2 = J3 = 12.5 Hz, H-2ax), 2.50 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-
2eq), 3.39–3.45 (m, 1H, H-3), 3.38–3.43 (m, 1H, H-4), 3.77 (dd,
1H, J1 = 9.0, J2 = 9.5 Hz, H-5), 3.95–4.03 (m, 1H, H-1), 4.80 (dd,
1H, J1 = J2 = 10.0 Hz, H-6), 6.62 (d, 1H, J = 7.5 Hz, NH); the addi-
tional peaks in the spectrum were identified as follow: 2.05–2.13

I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
3745
(dd, 2H J1 = 7.0, J2 = 6.5 Hz), 5.16 (dd, 1H, J1 = 5.0, J2 = 7.0 Hz), 7.34
5.3. Dual luciferase readthrough assays
(dd, 2H, J1 = J2= 7.5 Hz, Bz), 7.43 (dd, 2H, J1 = J2= 7.5 Hz, Bz), 7.54
(dd, 1H, J1 = J2 = 7.5 Hz, Bz), 7.59 (dd, 1H, J1 = J2 = 7.5 Hz, Bz), 7.86
(dd, 2H, J1 = 1.0, J2 = 8.5 Hz, Bz), 7.95 (dd, 2H, J1 = 1.0, J2 = 8.5 Hz,
Bz). 13C NMR (125 MHz, CDCl3): dC 15.4 (CH3-60), 22.5–23.0 (Ac,
5C), 26.5, 32.3 (C-2), 48.9, 50.3, 54.5 (C-500), 60.5, 63.3, 70.4, 70.8,
72.1, 72.5 72.7, 73.3 75.3, 76.5, 79.0, 81.9, 81.9, 98.2 (C-10), 109.2
(C-100), 130.3 (Bz, 4C), 130.4 (Bz, 2C), 131.5 (Bz, 4C), 135.5 (Bz,
2C), 167.0 (Bz, CO), 167.1 (Bz, CO), 170.8 (AHB, CO), 171.6–174.2
(5Ac, CO). MALDI TOFMS calcd for C46H53N13O19 Na ([M+Na]+)
m/e 1114.3; measured m/e 1114.5).
DNA fragments derived from PCDH15, CFTR, Dystrophin, and
IDUA cDNAs, including the tested nonsense mutation or the corre-
sponding wild type codon and four to six upstream and down-
stream flanking codons, were created and inserted into the
polylinker of the p2luc plasmid, as previously described by us.26
For the in vitro readthrough assays, the obtained plasmids or the
original p2luc plasmid, with addition of the tested aminoglyco-
sides, were transcribed and translated using the TNT reticulocyte
lysate quick coupled transcription/translation system (Promega™).
Luciferase activity was determined after 90 min of incubation at
30 °C, using the Dual luciferase reporter assay system (Promega™).
For the ex vivo readthrough assays, the constructs harboring the
R3X, Q70X and W1282 mutations were transfected to HEK-293
cells with Lipofectamine 2000 (Invitrogen) and addition of the
tested compounds was performed 6 h post transfection. The cells
were harvested following 16 h incubation with the aminoglyco-
sides. Stop codon readthrough was calculated as previously de-
scribed.35 The experimental details for all the readthrough
experiments were identical to our recently reported procedures
for the in vitro and ex vivo suppression tests of aminoglycosides
variants.25,26
5.2.10. 60-(R)-Methyl-5-O-(5-amino-5-deoxy-b-
D-ribofuranose)-
20,1,3-triamino-paromamine (7)
Compound 15 (750 mg, 0.79 mmol) was treated with a solution
of MeNH2 (33% solution in EtOH, 30 mL) at room temperature and
the reaction progress was monitored by TLC (EtOAc/MeOH 8:2),
which indicated completion after 8 h. The reaction mixture was
evaporated to dryness and was subjected to a Staudinger reaction
as described above for the preparation of compound 5 under fol-
lowing conditions: THF (6.0 mL), NaOH (1 mM, 5.0 mL), PMe3
(7.0 mL, 7.0 mmol) to yield 7 as a free amine form (295 mg, 80%
yield). 1H NMR (500 MHz, D2O, pD 3.5) ‘Ring I’: dH 1.09 (d, 3H,
J = 6.5 Hz, CH3-6), 3.32–3.36 (m, 1H, H-4), 3.36 (dd, 1H, J1 = 4.0,
J2 = 9.0 Hz, H-2), 3.70 (dd, 1H, J1 = 2.5, J2 = 10.0 Hz, 1H, H-5), 3.83
(dd, 1H, J1 = 9.0, J2 = 10.0 Hz, H-3), 4.09–4.15 (m, 1H, H-6), 5.64
(d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH 1.78 (ddd, 1H, J1 = J2 =
J3 = 12.5 Hz, H-2ax), 2.37 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-2eq),
3.22–3.28 (m, 1H, H-1), 3.45–3.51 (m, 1H, H-3), 3.65 (dd, 1H,
J1 = J2 = 9.5 Hz, H-6), 3.88 (dd, 1H, J1 = J2 = 9.5 Hz, H-5), 3.99 (dd,
1H, J1 = J2 = 9.5 Hz, H-4); ‘Ring III’: dH 3.11 (dd, 1H, J1 = 8.0,
J2 = 13.5 Hz, H-5), 3.25–3.29 (m, 1H, H-50), 3.99–4.05 (m, 1H, H-
4), 4.08–4.11 (m, 1H, H-3), 4.14 (d, 1H, J = 5.0, H-2), 5.26 (d, 1H,
J = 1.0, H-1). 13C NMR (125 MHz, D2O): dC 14.6 (CH3-60), 27.7 (C-
2), 41.5 (C-500), 49.2, 49.7, 53.6, 65.4, 69.2, 69.9, 70.8, 71.4, 74.4,
75.4, 77.5, 78.2, 82.0, 95.0 (C-10), 108.5 (C-100). MALDI TOFMS calcd
for C18H37N4O10 ([M+H]+) m/e 469.2; measured m/e 469.2).
5.4. Protein translation inhibition tests
Prokaryotic in vitro translation inhibition by the different ami-
noglycosides was quantified in coupled transcription/translation
assays36 by using E. coli S30 extract for circular DNA with the pBES-
Tluc plasmid (Promega), according to manufacturer’s protocol.
Translation reactions (25 lL) that contained variable concentra-
tions of the tested aminoglycoside were incubated at 37 °C for
60 min, cooled on ice for 5 min, and diluted with a dilution reagent
(tris-phosphate buffer (25 mM, pH 7.8), DTT (2 mM), 1,2-diam-
inocyclohexanetetraacetate (2 mM), glycerol (10%), Triton X-100
(1%) and BSA (1 mg mLꢂ1)) into 96-well plates. Eukaryotic
in vitro translation inhibition was quantified by use of TNTÒ T7
Quick Coupled Transcription/Translation System with a luciferase
T7 control DNA plasmid (Promega), according to the manufacturer
5.2.11. 60-(R)-Methyl-5-O-(5-amino-5-deoxy-b-
D-ribofuranose)-
20,3-diamino-1-N-[(S)-4-amino-2-hydroxybutanoyl]paromamine
Compound 16 (520 mg, 0.48 mmol) was treated with a solution
of MeNH2 (33% solution in EtOH, 20 mL) and the reaction progress
was monitored by TLC (EtOAc/MeOH 8:2), which indicated comple-
tion after 12 h. The reaction mixture was evaporated to dryness and
was directly subjected to a Staudinger reaction as was described
above for the preparation of compound 3 with the following
quantities: THF (3.0 mL), NaOH (1 mM, 3.0 mL), PMe3 (4.00 mL,
4.00 mmol) to afford 8 as a free amine form (195 mg, 72% yield).
1H NMR (500 MHz, D2O, pD 3.3): ‘Ring I’: dH 1.16 (d, 3H, J = 6.5 Hz,
CH3-6), 3.40 (dd, 1H, J1 = 9.0, J2 = 10.0 Hz, H-4), 3.42 (dd, 1H,
J1 = 4.0, J2 = 10.5 Hz, H-2), 3.76 (dd, 1H, J1 = 2.5, J2 = 10.0 Hz, H-5),
3.90 (dd, 1H, J1 = J2 = 10.0 Hz, H-3), 4.17–4.19 (m, 1H, H-6), 5.70
(d, 1H, J = 4.0 Hz, H-1); ‘Ring II’: dH 1.72 (ddd, 1H, J1 = J2 =
J3 = 12.5 Hz, H-2ax), 2.15 (dt, 1H, J1 = 4.0, J2 = 12.5 Hz, H-2eq),
3.46–3.52 (m, 1H, H-3), 3.64 (dd, 1H, J1 = 9.5, J2 = 10.0 Hz, H-6),
3.83–3.88 (m, 1H, H-1), 3.89 (dd, 1H, J1 = J2 = 9.0 Hz, H-5), 3.99
(dd, 1H, J1 = J2 = 9.5 Hz, H-4); ‘Ring III’: dH 3.16 (dd, 1H, J1 = 7.5,
J2 = 13.5 Hz, H-5), 3.32 (dd, 1H, J1 = 3.5, J2 = 13.5 Hz, H-50), 4.02–
4.08 (m, 1H, H-4), 4.13 (dd, 1H, J1 = 5.0, J2 = 7.5 Hz, H-3), 4.16–4.19
(m, 1H, H-2), 5.31 (s, 1H, H-1); the additional peaks in the spectrum
were identified as follow: 1.88–1.97 (m, 1H), 2.05–2.12 (m, 1H),
3.02–3.12 (m, 2H), 4.25 (dd, 1H, J1 = 4.0, J2 = 8.0 Hz). 13C NMR
(125 MHz, D2O) : dC 16.5 (CH3-60), 31.5 (C-2), 32.6, 38.4, 43.3 (C-
500), 50.4, 51.6, 55.6, 67.3, 71.1, 71.4, 71.8, 72.7, 74.4, 76.3, 77.3,
79.7, 79.9, 84.8, 96.8 (C-10), 110.4 (C-100), 177.4. MALDI TOFMS calcd
for C22H43N5O12K ([M+K]+) m/e 608.3; measured m/e 608.3).
protocol. Translation reactions (25 lL) containing variable concen-
trations of the tested aminoglycoside were incubated at 30 °C for
60 min, cooled on ice for 5 min, diluted with the dilution reagent
and transferred into 96-well plates. In both prokaryotic and
eukaryotic systems the luminescence was measured immediately
after the addition of the luciferase assay reagent (50 lL; Promega),
and the light emission was recorded with an FLx800 Fluorescence
Microplate Reader (Biotek). The half-maximal inhibition concen-
tration (IC50) values were obtained from fitting concentration–re-
sponse curves to the data of at least two independent
experiments by using Grafit 5 software,45 and are presented in
Table 1 and Figure 1S (see Supplementary data).
5.5. Antibacterial activity and human cell toxicity tests
Comparative antibacterial activities were determined in two
representative strains of Gram-negative (E. coli R477–100) and
Gram-positive (B. subtilis ATCC-6633) bacteria, by measuring the
MIC values using the double-microdilution method according to
the National Committee for Clinical Laboratory Standards
(NCCLS)46 with starting concentration of 384
lg/mL of the tested
compound. All the experiments were performed in triplicates and
analogous results were obtained in three different experiments.
For the cytotoxicity assays, HEK-293 cells were grown over
night in 96-well plates (5000 cells/well) in DMEM medium con-
taining 10% FBS, 1% penicillin/streptomycin and 1% glutamine
(90 lL; Biological Industries) at 37 °C and 5% CO2. The medium

3746
I. Nudelman et al. / Bioorg. Med. Chem. 18 (2010) 3735–3746
14. Sangkuhl, K.; Schulz, A.; Rompler, H.; Yun, J.; Wess, J.; Schoneberg, T. Hum. Mol.
Genet. 2004, 13, 893.
15. Helip-Wooley, A.; Park, M. A.; Lemons, R. M.; Thoene, J. G. Mol. Genet. Metab.
2002, 75, 128.
was then changed with the medium without penicillin/streptomy-
cin (90 L), different concentrations of the tested aminoglycosides
were added (10 L) and the cells were incubated for 48 h. A cell
l
l
16. Grayson, C.; Chapple, J. P.; Willison, K. R.; Webster, A. R.; Hardcastle, A. J.;
Cheetham, M. E. J. Med. Genet. 2002, 39, 62.
17. Xi, B.; Guan, F.; Lawrence, D. S. J. Am. Chem. Soc. 2004, 126, 5660.
18. Holbrook, J. A.; Neu-Yilik, G.; Hentze, M. W.; Kulozik, A. E. Nat. Genet. 2004, 36,
801.
19. (a) Aguiari, G.; Banzi, M.; Gessi, S.; Cai, Y.; Zeggio, E.; Manzati, E.; Piva, R.;
Lambertini, E.; Ferrari, L.; Peters, D. J.; Lanza, F.; Harris, P. C.; Borea, P. A.;
Somlo, S.; Del Senno, L. FASEB J. 2004, 18, 884; (b) Kellermayer, R.; Szigeti, R.;
Keeling, K. M.; Bedekovics, T.; Bedwell, D. M. J. Invest. Dermatol. 2006, 126, 229;
(c) Olson, T. M.; Alekseev, A. E.; Liu, X. K.; Park, S.; Zingman, L. V.;
Bienengraeber, M.; Sattiraju, S.; Ballew, J. D.; Jahangir, A.; Terzic, A. Hum.
Mol. Genet. 2006, 15, 2185; (d) Porgpermdee, S.; Yu, X.; Takagi, A.; Mayuzumi,
N.; Ogawa, H.; Ikeda, S. J. Dermatol. Sci. 2005, 40, 137.
20. Karpati, G.; Lochmuller, H. Ann. Neurol. 2001, 49, 693.
21. Carter, A. P.; Clemons, W. M.; Brodersen, D. E.; Morgan-Warren, R. J.; Wimberly,
B. T.; Ramakrishnan, V. Nature 2000, 407, 340.
22. Borovinskaya, M. A.; Pai, R. D.; Zhang, W.; Schuwirth, B. S.; Holton, J. M.;
Hirokawa, G.; Kaji, H.; Kaji, A.; Cate, J. H. Nat. Struct. Mol. Biol. 2007, 14,
727.
23. Kondo, J.; Hainrichson, M.; Nudelman, I.; Shallom-Shezifi, D.; Barbieri, C. M.;
Pilch, D. S.; Westhof, E.; Baasov, T. ChemBioChem 2007, 8, 1700.
24. Hainrichson, M.; Nudelman, I.; Baasov, T. Org. Biomol. Chem. 2008, 6, 227.
25. Nudelman, I.; Rebibo-Sabbah, A.; Shallom-Shezifi, D.; Hainrichson, M.; Stahl, I.;
Ben-Yosef, T.; Baasov, T. Bioorg. Med. Chem. Lett. 2006, 16, 6310.
26. Nudelman, I.; Rebibo-Sabbah, A.; Cherniavsky, M.; Belakhov, V.; Hainrichson,
M.; Chen, F.; Schacht, J.; Pilch, D. S.; Ben-Yosef, T.; Baasov, T. J. Med. Chem. 2009,
52, 2836.
proliferation assay (XTT based colorimetric assay, Biological
Industries) was performed by using the 3-h incubation protocol,
according to manufacturer’s instructions. O.D. (optical density)
was then read by using an Elisa plate reader. Cell viability was cal-
culated as the ratio between the numbers of living cells in cultures
grown in the presence of the tested compounds, versus cultures
grown under the identical protocol but without the tested com-
pound. Primary human dermal fibroblasts (HDF) were isolated
from adult skin and cultivated as previously described.47 Cytotox-
icity assays in HDF cells were performed as described above for
HEK-293 cells, except the following modification: the HDK cells
were incubated with aminoglycosides for 72 h. The half-maximal
lethal concentration (LC50) values were obtained from fitting
concentration–response curves to the data of at least two indepen-
dent experiments, using GraFit 5 software45 and are presented in
Table 1 and Figure 2S (see Supplementary data).
Acknowledgments
We thank John F. Atkins (University of Utah) for providing the
p2luc plasmid. This work was supported by research grants from
the US-Israel Binational Science Foundation (Grant No. 2006/301)
and by Mitchel Fund (Grant No. 2012386). I.N. acknowledges the
Irwin and Joan Jacobs Fellowship; D.G. acknowledges the Miriam
and Aaron Gutwirth Memorial Fellowship; V.B. acknowledges the
financial support by the Center of Absorption in Science, the Min-
istry of Immigration Absorption and the Ministry of Science and
Technology, Israel (Kamea Program).
27. Rebibo-Sabbah, A.; Nudelman, I.; Ahmed, Z. M.; Baasov, T.; Ben-Yosef, T. Hum.
Genet. 2007, 122, 373.
28. Recht, M. I.; Fourmy, D.; Blanchard, S. C.; Dahlquist, K. D.; Puglisi, J. D. J. Mol.
Biol. 1996, 262, 421.
29. Recht, M. I.; Douthwaite, S.; Puglisi, J. D. EMBO J. 1999, 18, 3133.
30. Pfister, P.; Hobbie, S.; Vicens, Q.; Bottger, E. C.; Westhof, E. ChemBioChem 2003,
4, 1078.
31. Manuvakhova, M.; Keeling, K.; Bedwell, D. M. RNA (New York, NY) 2000, 6,
1044.
32. Howard, M. T.; Shirts, B. H.; Petros, L. M.; Flanigan, K. M.; Gesteland, R. F.;
Atkins, J. F. Ann. Neurol. 2000, 48, 164.
33. Chernikov, V. G.; Terekhov, S. M.; Krokhina, T. B.; Shishkin, S. S.; Smirnova, T.
D.; Kalashnikova, E. A.; Adnoral, N. V.; Rebrov, L. B.; Denisov-Nikol’skii, Y. I.;
Bykov, V. A. Bull. Exp. Biol. Med. 2003, 135, 103.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.03.060.
34. Sandoval, R. M.; Reilly, J. P.; Running, W.; Campos, S. B.; Santos, J. R.; Phillips, C.
L.; Molitoris, B. A. J. Am. Soc. Nephrol. 2006, 17, 2697.
35. Grentzmann, G.; Ingram, J. A.; Kelly, P. J.; Gesteland, R. F.; Atkins, J. F. RNA (New
York, NY) 1998, 4, 479.
References and notes
36. Greenberg, W. A.; Priestley, E. S.; Sears, P. S.; Alper, P. B.; Rosenbohm,
C.; Hendrix, M.; Hung, S. C.; Wong, C. H. J. Am. Chem. Soc. 1999, 121,
6527.
37. Howard, M. T.; Anderson, C. B.; Fass, U.; Khatri, S.; Gesteland, R. F.; Atkins, J. F.;
Flanigan, K. M. Ann. Neurol. 2004, 55, 422.
1. Keeling, K. M.; Bedwell, D. M. Curr. Pharmacogenomics 2005, 3, 259.
2. Kellermayer, R. Eur. J. Med. Genet. 2006, 49, 445.
3. Zingman, L. V.; Park, S.; Olson, T. M.; Alekseev, A. E.; Terzic, A. Clin. Pharmacol.
Ther. 2007, 81, 99.
38. Vicens, Q.; Westhof, E. Biopolymers 2003, 70, 42.
39. Recht, M. I.; Puglisi, J. D. Antimicrob. Agents Chemother. 2001, 45, 2414.
40. This hypothesis was further supported by the data in the same reports showing
that gentamicin C1 (closely resembles G418 structure but contains 60-NH2) and
neomycin (same structure to that of paromomycin except the NH2 group at 60
position) conferred >200–1000-fold resistance in A1408G mutant.
41. Kondo, J.; Westhof, E. Nucleic Acids Res. 2008, 36, 2654.
42. Bottger, E. C.; Springer, B.; Prammananan, T.; Kidan, Y.; Sander, P. EMBO Rep.
2001, 2, 318.
43. Hobbie, S. N.; Bruell, C. M.; Akshay, S.; Kalapala, S. K.; Shcherbakov, D.; Bottger,
E. C. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 3244.
44. Alper, P. B.; Hendrix, M.; Sears, P. S.; Wong, C. H. J. Am. Chem. Soc. 1998, 120,
1965.
4. Burke, J. F.; Mogg, A. E. Nucleic Acids Res. 1985, 13, 6265.
5. Bedwell, D. M.; Kaenjak, A.; Benos, D. J.; Bebok, Z.; Bubien, J. K.; Hong, J.;
Tousson, A.; Clancy, J. P.; Sorscher, E. J. Nat. Med. 1997, 3, 1280.
6. Howard, M.; Frizzell, R. A.; Bedwell, D. M. Nat. Med. 1996, 2, 467.
7. Du, M.; Jones, J. R.; Lanier, J.; Keeling, K. M.; Lindsey, J. R.; Tousson, A.; Bebok, Z.;
Whitsett, J. A.; Dey, C. R.; Colledge, W. H.; Evans, M. J.; Sorscher, E. J.; Bedwell,
D. M. J. Mol. Med. 2002, 80, 595.
8. Barton-Davis, E. R.; Cordier, L.; Shoturma, D. I.; Leland, S. E.; Sweeney, H. L. J.
Clin. Invest. 1999, 104, 375.
9. 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. Eng. J. Med.
2003, 349, 1433.
45. Leatherbarrow, R. J. GraFit 5; Horley, U.K.: Erithacus Software Ltd, 2001.
46. NCCLS. National Committee for Clinical Laboratory Standards, Performance
standards for antimicrobial susceptibility testing. Fifth information supplement:
Approved Standard M100-S5; Villanova, Pa.: NCCLS, 1994.
47. Stark, H. J.; Baur, M.; Breitkreutz, D.; Mirancea, N.; Fusenig, N. E. J. Invest.
Dermatol. 1999, 112, 681.
10. Politano, L.; Nigro, G.; Nigro, V.; Piluso, G.; Papparella, S.; Paciello, O.; Comi, L. I.
Acta Myol. 2003, 22, 15.
11. Zilberberg, A.; Lahav, L.; Rosin-Arbesfeld, R. Gut 2010, 59, 496.
12. Brendel, C.; Klahold, E.; Gartner, J.; Huppke, P. Pediatr. Res. 2009, 65, 520.
13. Keeling, K. M.; Brooks, D. A.; Hopwood, J. J.; Li, P.; Thompson, J. N.; Bedwell, D.
M. Hum. Mol. Genet. 2001, 10, 291.