Towards Catalytic Antibiotics: Redesign of Aminoglycosides To Catalytically Disable Bacterial Ribosomes

Article abstract of DOI:10.1002/cbic.201800549

The emergence of multidrug-resistant pathogens that are resistant to the majority of currently available antibiotics is a significant clinical problem. The development of new antibacterial agents and novel approaches is therefore extremely important. We set out to explore the potential of catalytic antibiotics as a new paradigm in antibiotics research. Herein, we describe our pilot study on the design, synthesis, and biological testing of a series of new derivatives of the natural aminoglycoside antibiotic neomycin B for their potential action as catalytic antibiotics. The new derivatives showed significant antibacterial activity against wild-type bacteria and were especially potent against resistant and pathogenic strains including Pseudomonas aeruginosa and methicillin-resistant Staphylococcus aureus. Selected compounds displayed RNase activity even though the activity was not as high and specific as we would have expected. On the basis of the observed chemical and biochemical data, along with the comparative molecular dynamics simulations of the prokaryotic rRNA decoding site, we postulate that the rational design of catalytic antibiotics should involve not only their structure but also a comprehensive analysis of the rRNA A-site dynamics.

Full text of DOI:10.1002/cbic.201800549

Accepted Article
Title: Towards Catalytic Antibiotics: Redesign of Aminoglycosides to
Catalytically Disable Bacterial Ribosome
Authors: Boris Smolkin, Alina Vilensky, Tomasz Pieńko, Michal Shavit,
Valery Belakhov, Joanna Trylska, and Timor Baasov
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201800549
Link to VoR: http://dx.doi.org/10.1002/cbic.201800549
A Journal of
www.chembiochem.org

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Towards Catalytic Antibiotics: Redesign of Aminoglycosides to
Catalytically Disable Bacterial Ribosome
Boris Smolkin,[a] Alina Khononov,[a] Tomasz Pieńko,[b,c] Michal Shavit,[a] Valery Belakhov,[a] Joanna
Trylska[b] and Timor Baasov*[a]
Dedicated to Professor Chi-Huey Wong on the occasion of his 70th birthday.
Abstract: The emergence of multidrug-resistant pathogens that are
resistant to the majority of currently available antibiotics is a significant
clinical problem. The development of new antibacterial agents and
novel approaches is therefore extremely important. We set out to
explore the potential of catalytic antibiotics as a new paradigm in
antibiotics research. Here we describe our pilot study on the design,
synthesis, and biological testing of a series of new derivatives of the
natural aminoglycoside antibiotic neomycin B for their potential action
as catalytic antibiotics. The new derivatives showed significant
antibacterial activity against wild-type bacteria and were especially
potent against resistant and pathogenic strains including P.
aeruginosa and MRSA. Selected compounds displayed RNase
activity even though the activity was not as high and specific as we
would have expected. Based on the observed chemical and
biochemical data, along with the comparative molecular dynamics
simulations of the prokaryotic rRNA decoding site, we postulate that
the rational design of the catalytic antibiotics should involve not only
their structure but also a comprehensive analysis of the rRNA A-site
dynamics.
bacterial target or a semisynthetic derivative that counters the
resistance to its parent drug, within only a short matter of time new
resistance will emerge and create a serious public health
problem[3]. Some bacterial strains have developed multi-drug
resistance that covers the majority of currently available
antibiotics. The significance of this health problem has re-
energized the search for new antibacterial agents and novel
approaches.
One innovative approach is the development of catalytic
antibiotics: the pharmacophore of an existing antibiotic is modified
to include a catalytic warhead that disables the target in a catalytic
manner. Unlike conventional antibiotics that act on their targets
either in a reversible (non-covalent interaction) or irreversible
manner (covalent interaction), the antibiotics acting in a catalytic
manner promote multiple turnovers of a catalytic cycle. The
possible benefits include: (i) activity at lower dosages and
consequently reduced side effects, (ii) activity against drug-
resistant bacteria, and (iii) reduced potential for generating new
resistance.
In general, the idea of catalytic drugs is not novel and
several studies towards the development of such agents have
been reported previously [4],[5]. These include numerous peptide-
cleaving agents based on small molecule metal complexes as
artificial proteases[4], site-specific RNA-cleaving agents that
combine a reactive moiety (phosphodiester cleavage directed
non-metallic, warhead) with a recognition element (sequence-
specific hybridization to target RNA)[6], and non-metallic small
1. Introduction
The ongoing emergence of multidrug-resistant pathogens
requires continuous intensive search for novel antibiotics.
Unfortunately, only two new classes of antibiotics,
oxazolidinones[1] and lipoproteins[2] have been introduced into
clinical practice during the last decades. Furthermore, it has been
well documented that once a new antibiotic is introduced into the
clinic, whether it is a novel chemical entity acting at a distinct
organic molecules as artificial ribonucleases[7],[8]
.
Inspired by these findings, we set out to explore the potential
of catalytic antibiotics as a new paradigm in antibiotics research.
Herein we focus on redesigning aminoglycosides, which
represent a particularly well-studied and broad-spectrum class of
antibiotics. These molecules exert their therapeutic (bactericidal)
effect by selectively binding to the aminoacyl-tRNA binding site
(A-site) of the bacterial 16S rRNA, thereby interfering with
translational fidelity during protein synthesis[9]. Previous reports
on the ability of copper-aminoglycoside complexes to promote
hydrolytic and oxidative cleavage of RNA[10][11][12][13] have
prompted the potential use of these complexes as metallodrugs
with potent antibacterial activity. However, antibacterial tests
showed no significant enhancement in the activity of the copper-
[a]
Dr. B. Smolkin, A. Khononov, Dr. M. Shavit, Dr. V. Belakhov, Prof.
Dr. T. Baasov
The Edith and Joseph Fischer Enzyme Inhibitors Laboratory,
Schulich Faculty of Chemistry Department
Technion - Israel Institute of Technology
Haifa 3200003, Israel
E-mail: chtimor@technion.ac.il
[b]
[c]
Dr. T. Pieńko, Prof. Dr. J. Trylska
Centre of New Technologies
University of Warsaw
Banacha 2c, 02-097 Warsaw, Poland
Dr. T. Pieńko
aminoglycoside
complex
compared
to
the
parent
Department of Drug Chemistry, Faculty of Pharmacy with the
Laboratory Medicine Division
Medical University of Warsaw
aminoglycoside[14], suggesting that a more sophisticated design
process is needed to obtain a biologically functional catalytic
antibiotic.
Banacha 1a, 02-097 Warsaw, Poland
Supporting information for this article is given via a link at the end of
the document.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Particularly improved activities were observed against Methicillin-
resistant Staphylococcus aureus (MRSA) and Pseudomonas
aeruginosa strains, which are both highly resistant to conventional
aminoglycosides. Interestingly, three of the 4’-amide derivatives
showed a 2-fold stronger inhibition of protein synthesis in
comparison to NeoB and other clinically used aminoglycosides.
However, all attempts to demonstrate the cleavage of the scissile
phosphodiester bond of rRNA by these derivatives have been
unsuccessful. Based on the observed chemical and biochemical
data, along with the comparative molecular dynamics simulations
of the prokaryotic rRNA decoding site, we postulate that the
rational design of the catalytic antibiotics should involve not only
their structure but also a comprehensive analysis of the rRNA A
site dynamics.
2. Results and Discussion
2.1. Design hypothesis
Initially, we considered the following three key aspects in our
design of a potential aminoglycoside catalyst: (1) the choice of the
phosphodiester bond in the A site that should be the most
susceptible to catalytic cleavage, (2) the potential “catalytic
warhead” structures and (3) the attachment site of a “catalytic
warhead” on the aminoglycoside structure.
(1) The choice of the phosphodiester bond. Compelling evidence
is now available that the successful cleavage of an RNA
phosphodiester bond requires substantial motion in the HO-C2'-
C3'-O-P bonds of the ribose-3'-phosphate region to reach the
necessary low energy transition state wherein the C2'-OH group
is orientated for in-line nucleophilic attack on the scissile bond [7]
.
Such flexibility is usually achieved by the enzyme-induced flipping
of the base attached to the RNA scissile bond. The mechanisms
suggested for the RNase T1[15], RNase α-sarcin[16], and for
several ribozymes[17], are only a few of the examples that support
this notion. Of particular relevance is the proposed mechanism for
colicin E3 (ColE3), a natural enzymatic toxin produced in several
E. coli strains, that selectively cleaves a phosphodiester bond
between A1493 and G1494 of 16S rRNA[18]. This cleavage
impairs the protein translation process and consequently leads to
cell death. The proposed mechanism of ColE3 also explains why
this natural ribonuclease cleaves the specific position in the A site
of rRNA, between A1493 and G1494. This region of the A site is
very important functionally (for correct proofreading) and is also
one of the most flexible and accessible regions in the whole
ribosome because it needs to accommodate the incoming
aminoacyl-tRNA.
Figure 1. Structures of neamine, neomycin
aminoglycosides 1-10 that were investigated in this study.
B (NeoB) and synthetic
We hypothesized that by employing the available structural
and mechanistic library of data on the aminoglycoside target,
natural ribonuclease systems, and recently reported artificial,
small molecule systems capable of cleaving phosphodiester
bonds, we could create a new variant of aminoglycoside antibiotic
that would selectively and catalytically act on the bacterial
ribosome and irreversibly deactivate it. To test this hypothesis,
herein we describe our pilot studies on the design, synthesis and
biological evaluation of a series of new Neomycin B (NeoB)
derivatives (compounds 1-10, Figure 1) substituted at the 4'
position (via ether and amide linkages) or at the 6’ position (via
amide linkage) with various diamine moieties as potential catalytic
warheads for the hydrolysis of rRNA. The observed data
demonstrates that most of the new derivatives retain the
antibacterial potency of the parent NeoB against wild-type strains
of both Gram-negative and Gram-positive bacteria and display
significantly better activity against the tested resistant strains.
Consequently, we postulated that the target phosphodiester
bond must be within the region of rRNA that upon binding of an
aminoglycoside undergoes the most extensive conformational
change. Moreover, this region is virtually the same as that of
ColE3 binding: G1491-A1492-A1493-G1494. Since the binding of
most aminoglycosides induces extensive flipping of A1492 and
A1493 base residues from the bulged-in (ligand unbound
ribosome) to the bulged-out conformation[19], similar to that of the
ColE3 binding[20]
, it is most likely that the best three
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
phosphodiester bond candidates within the A-site are between
G1491-A1492, A1492-A1493, and A1493-G1494.
(2) The choice of the catalytic warheads. Previous studies with
simple diamines have demonstrated their ability to accelerate the
cleavage of adenylyl(3'-5')-adenosine (ApA), from one to three
orders of magnitude more efficiently than the corresponding
monoamines[21],[22]. Furthermore, it has been shown that the order
of reactivity for the simple diamine series is as follows: N-2-N >
N-3-N > N-1-N > N-4-N > N-5-N. The strong activities of N-2-N
and N-3-N were primarily ascribed to the abundance of
catalytically active monocations (61% for N-2-N and 7.4% for N-
3-N) that exist at pH 7 because the second protonation is
suppressed due to the electrostatic repulsion of the positively
charged ammonium ions (the corresponding pKa’s are 6.8 and 9.4
for N-2-N, and 8.1 and 9.8 for N-3-N). Based on these
Figure 2. (A) Ball and stick representation of compound 2-induced cleavage site
in the bacterial rRNA A-site. Modelling has been performed by superimposition
of 2 with NeoB structure in the crystal structure of NeoB bound to the rRNA
olignonucleotide model (PDB ID 2ET4)[19] by using PyMol. (B) Proposed
catalytic action of compound 2 on the hydrolysis of the phosphodiester bond
between G1491 and A1492. Modelling has been performed by superimposition
of 2 structure with Neomycin A-site crystal structure (PDB ID 2ET4) by using
PyMol.
observations,
we
selected
ethylenediamine,
methyl
ethylenediamine, diethylenetriamine, N-(2-aminoethyl)pyrrolidine
and guanidine-ethyleneamine as potential “catalytic warheads”
and prepared the new NeoB derivatives 1-10 (Figure 1).
(3) The choice of the attachment site of a catalytic warhead. We
selected the 4'-hydroxyl group (ring I) of NeoB (Figure 2) as the
attachment site for the following reasons. Firstly, the available
structural data on the interaction of NeoB with its ribosomal
target[19] indicate that the 4'-hydroxyl is positioned in front of the
scissile phosphodiester bond of G1491-A1492, and is near a large
cavity formed in the A-site that makes its modification feasible.
Secondly, our preliminary molecular modelling studies of the
proposed warheads linked at the 4’ position suggested that the
phosphodiester bond between G1491 and A1492 is the closest
one and its cleavage is feasible through acid-base catalysis
(Figure 2): the terminal amino group in its ammonium form can
activate the phosphate between G1491 and A1492 as a general-
acid (3.9 Å distance), and the next nearest amine can activate the
2'-hydroxyl of G1491 as a general-base (2.6 Å distance). Finally,
Ye and co-workers have recently reported a series of new
derivatives of kanamycin B, modified at the 4'-OH position, which
show excellent antibacterial activity against both wild-type and
resistant bacteria[23]. The following findings of this study are of
particular importance: (1) the side-chain free amine is best
tolerated by the ribosome; (2) the A-site of the ribosome can
accommodate bulky substituents linked at the 4'-position. Based
upon these collective data, we selected the 4'-OH group (ring I) of
NeoB as an attachment site for the catalytic warhead and the
G1491-A1492 as the cleavage site, as schematically illustrated in
Figure 2.
free base form was then converted to the corresponding perazido
derivative 11 by the diazo transfer reaction in the presence of TfN3,
CuSO4 ·5H2O and Et3N.
Treatment of 11 with benzaldehyde dimethylacetal in dry
DMF in the presence of camphor sulfonic acid (CSA) afforded the
corresponding benzylidene acetal 12, which was then O-
benzylated with benzyl bromide (BnBr) in the presence of NaH in
DMF to yield the tribenzyl ether 13. Removal of the benzylidene
group (acetic acid, 60oC) gave the corresponding diol 14, which
was then selectively tosylated at the 6'-hydroxyl group followed by
nucleophilic substitution with sodium azide to yield compound 15.
Allylation of the 4'-hydroxyl with allyl bromide in the presence of
NaH in DMF gave the 4’-allyl derivative 16. Attempts to convert
16 to the corresponding aldehyde by ozonolysis resulted in a
mixture of products due to the partial oxidation of the benzyl
groups. To solve this problem, the double bond in 16 was first
converted to the corresponding diol 17 by using the procedure of
Nicolaou[24]. Oxidative cleavage of the diol 17 [PhI(OAc)2, DCM]
was followed by in-situ reductive amination with 2-
azidoethanamine[25] to yield the corresponding 4’-azido amine 18
in 66% isolated yield. Finally, after several unsuccessful attempts
to remove the benzyl and azide protections in 18, we found the
sequential operation of the Staudinger and Birch reactions to be
the best protocol. Thus, the Staudinger reaction (PMe3, NaOH)
followed by the Birch reduction (Na/NH3, THF) gave the target
compound 1 in 65% isolated yield.
The 4’-O-substituted derivatives of NeoB, compounds 2-5
(Figure 1), were synthesized using the same strategy described
for the synthesis of compound 1 with some modifications as
illustrated in Scheme 2. The modifications used were as follows.
Firstly, unlike the azidation of paromamine with TfN3 to yield the
corresponding perazido derivative 11 (Scheme 1), the same
reaction on paromomycin gave a very low yield of the desired
perazido derivative 19.[26][27] In an attempt to improve the yield of
the desired perazido product, instead of TfN3 we used the
imidazole-1-sulfonyl azide hydrochloride (ImSO2N3·HCl)[28] and
2.2. Synthesis of 4’-O-linked compounds
To selectively modify NeoB at the desired 4’ position, we
initially developed the required synthetic pathway for its simplest
fragment neamine, which consists of rings I and II of NeoB, and
prepared the derivative 1 as illustrated in Scheme 1. The
synthesis started from the commercial paromomycin sulfate,
which was treated with anhydrous HCl (AcCl in MeOH) at reflux
to give a highly regioselective hydrolysis between the rings II and
III, and afforded paromamine as its hydrochloride salt. The
observed salt was converted to the free base form by passing it
through a column of Dowex 50W (H+ form). Paromamine in its
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Scheme 1. Chemical transformation of paromomycin into pseudo-disaccharide 1.
replacement of tosyl chloride with the more bulky tri-isopropyl-
sulfonyl chloride (trisyl chloride), which was more selective for the
protection of the 6'-hydroxyl (conversion of compound 22 to 23)
and gave 60% yield over two steps (trisylation and azidation).
The common intermediate diol 25 was separately subjected
to in-situ oxidation and reductive amination steps with four
axial hydroxyl at the 4' position. Compound 31 was treated with
triflic anhydride (Tf2O, pyridine, CH2Cl2) to form the corresponding
4’-triflate, which was then reacted with ammonia in acetone to
yield 32 with an equatorial amine at 4’ position. Next, 32 was
treated with chloroacetyl chloride to give the 4’-chloride 33, which
was then separately reacted with three different amines,
compounds A, B, and diethylene triamine, to afford the
corresponding 4’-amide derivatives of NeoB in their protected
forms (compounds 34, 35 and 36, respectively). These products
were then deprotected using the two-step procedure described
above (Staudinger and Birch), to afford the corresponding 4’-
amide derivatives of NeoB, compounds 6, 7 and 8 in 64%, 68%
and 20% isolated yields, respectively. Interestingly, during the last
deprotection step (the Birch reduction), we discovered that if this
step proceeds in the presence of excess sodium, transamidation
rearrangement of the warhead, from the 4' position to the 6'
position, takes place. The structure of the rearrangement product
(6’-amide) was confirmed by its isolation and subsequent spectral
assignment by using a combination of various 1D and 2D NMR
techniques (see Figure S1 in Supporting Information). This
rearrangement probably occurs due to the strong basic conditions
generated after quenching of the reaction, which results in the
different
amine
linkers,
compounds
A,
B,
1-(2-
aminoethyl)pyrrolidine and C (see Chart 1)[25], to afford the
corresponding protected 4’-O-derivatives of NeoB, compounds
26-29 (Scheme 2). The Staudinger reaction (PMe3, NaOH)
followed by the Birch reduction (Na/NH3, THF) gave the target
compounds 2-5 in average to modest isolated yields. The
structures of all new compounds (1–5) were confirmed by
combining 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.
2.3. Synthesis of 4’- and 6’-amide linked compounds
For the synthesis of the 4’-amide derivatives, the alcohol 23
(Scheme 2) was first oxidized with Dess–Martin periodinane
(DMP) to form the corresponding 4'-ketone 30, which was then
reduced with sodium borohydride to afford compound 31 with an
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Scheme 2. Chemical transformation of paromomycin into synthetic derivatives 2-5.
formation of sodium hydroxide. We exploited this transformation
by performing the Birch reaction step under excess sodium, and
synthesized the corresponding 6'-amide linked compounds 9 and
10 in 74% and 36% isolated yields, respectively.
values for the new designer structures 1-10 were determined
against wild-type (WT) Gram-negative and Gram-positive
bacteria. Although this simple test cannot confirm or disprove
catalytic activity, it is nonetheless of the utmost importance as it
reveals how the compounds act in vivo. Furthermore, we
anticipated that unusually low MIC values would be indicative of
a high catalytic rate and turnover.
2.4. Antibacterial activity and protein translation inhibition
tests
In order to probe the influence of the attached warheads on
antibacterial activity, the minimal inhibitory concentration (MIC)
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Scheme 3. Chemical transformation of intermediate 23 in to the synthetic derivatives 6-10.
Table 1 shows the comparative MIC values of NeoB and of
compounds 1-10 against both Gram-negative and Gram-positive
bacteria, including pathogenic and resistant strains. The bacterial
strains that were included in these tests were as follows: two WT
E. coli strains (R477-100 and 25922) as representatives of Gram-
negative bacteria with unknown resistance to aminoglycosides[29]
and two WT Staphylococcus epidermidis and Bacillus subtilis
strains as representatives of Gram-positive bacteria (the clinically
used aminoglycosides have significant antibacterial activity
against these strains).[30]The resistant strains included Methicillin-
included several strains of Pseudomonas aeruginosa that have
an inherent resistance to aminoglycosides[33],[34]
.
The comparative data collected in Table 1 show that all the
new derivatives of NeoB, compounds 2-10, exhibit significant
antibacterial activity against both the WT and aminoglycoside-
resistant strains including Gram-negative and Gram-positive
bacteria. In general, the activity of the novel NeoB derivatives
against the WT Gram-negative bacteria is similar to or slightly
lower than that of the parent NeoB. The activities against the WT
Gram-positive bacteria were diverse across the different strains
tested. The activity of most of the compounds against S.
epidermidis is similar to or better than that of NeoB, while the
activity against B. subtilis is generally lower than that of NeoB.
Interestingly, all new derivatives (compounds 2-10) have shown
significantly improved activity against the Gram-negative strains
of pathogenic P. aeruginosa in comparison with NeoB.
resistant Staphylococcus aureus (MRSA),
a Gram-positive
bacterium, whose treatment represents a great challenge in the
clinic, the MRSA 252, which is known for its high resistance to
aminoglycosides[31] and the MRSA CI 15877, which is resistant to
natural aminoglycosides[32]. Other pathogens that were tested
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Table 1. Comparative antibacterial activity (MIC values in µg/mL) and inhibition of protein translation (IC50 values) in prokaryotic system of NeoB and
synthetic compounds 1-10
IC50
Bacteria/
a
b
c
d
e
f
g
h
i
j
k
compound
Gram-negative
Gram-positive
P. aeruginosa
MRSA
Geobacillus
(μM)
NeoB
12
12
6
0.75-1.5
>192
48-96
192
>192
48
0.2
12
0.01±0.002
1
2
384
48
96
48
192
24
24
48
48
24
>384
48
192
6
48
6
48-96
24-48
48
48-96
6-12
12
48-96
24
192
6-12
24
48
0.75
3
6
12
0.8
0.8
0.8
-
2.03±0.3
0.02±0.001
0.03±0.005
0.03±0.007
0.08±0.005
0.005±0.0005
0.07±0.004
0.07±0.007
0.006±0.0009
0.006±0.0009
0.2
0.2
0.2
-
3
96
6
6-12
3-6
12
2
24
4
48-96
192
48
6
96
48
48
24
6
5
24
6
-
-
-
-
-
6
24
6
48
24-48
24
1.5
1.5
0.75
1.5
1.5
0.2
0.2
0.4
0.2
0.2
0.4
0.4
1.5
0.8
0.4
7
24
6
3-6
6
24
24
48
8
48
3
24
24-48
48-96
24
48
12-24
12
9
48
6
3-6
3
48
24
10
48
6
24
48-96
12-24
The shadowed rows in the table highlight the most potent compounds.
[a] Escherichia coli R477-100; [b] Escherichia coli 25922; [c] Staphylococcus Epidermidis; [d] Bacillus Subtilis; [e] Pseudomonas aeruginosa 1275; [f] Pseudomonas
aeruginosa 27853; [g] Pseudomonas aeruginosa O1; [h] MRSA 252; [i] MRSA 15877; [j] Geobacillus stearothermophilus T-1 60oC; [k] Geobacillus
stearothermophilus KanR 60oC;
P. aeruginosa is a nosocomial human pathogen known to
be inherently resistant to aminoglycosides due to the presence of
the chromosomally encoded APH(3’)-IIb enzyme. This enzyme
catalyzes the transfer of the ATP γ-phosphoryl group to the 3'-
hydroxyl of many aminoglycosides, rendering them inactive as
the hydrolysis of the scissile phosphodiester bond of rRNA. To
test this hypothesis, we decided to test the antibacterial activity
against the thermophilic strain Geobacillus T1 with the optimal
growth temperature of 60°C (Table 1). We expected that if the
new compound had catalytic activity, it would show better activity
than NeoB at 60°C. However, all the new compounds, 1-10,
showed antibacterial activity similar to that of NeoB against the
WT Geobacillus T1. Against the Geobacillus T1 harboring the
resistance to kanamycin, as expected, most of the new
compounds maintained their high antibacterial activity, while
NeoB almost lost its activity.
In summary, even though we cannot conclude from the
observed comparative MIC data whether or not compounds 1-10
have catalytic activity, we can definitely claim that the
modifications we introduced have not hindered the binding to the
A-site and most of the derivatives have retained significant
antibacterial activity. Moreover, the new compounds have
overcome the existing resistance of P. aeruginosa and MRSA
pathogens to aminoglycosides, which has been one of the
greatest challenges in today's antibacterial research!
antibiotics[34]
. The observed improved activity of the new
derivatives in comparison to NeoB, against the tested strains of P.
aeruginosa can be explained by the steric hindrance of the
cationic warhead, which introduces unfavorable interactions with
the APH(3’)-IIb enzyme active site.
A very similar improvement in antibacterial performance of
the new designers versus that of NeoB was also observed against
the Gram-positive pathogenic MRSA strains. Especially large
improvements were observed for compounds 2 and 8, exhibiting
MIC values 64 times lower that of NeoB. No significant difference
in antibacterial activity was observed between the 4’-ether
(compounds 1-5) and 4’-amide (compounds 6-8) derivatives,
even though we had expected the 4’-amide derivatives to be more
active because of the potential for additional attractive
interactions between the amide bond and the rRNA. Compound
1, a neamine-based derivative, has a substantially lower
antibacterial activity in comparison with the other compounds
Next, we tested the protein translation inhibition by
determining half-maximum inhibition levels (IC50 values, Table 1).
tested, indicating that its binding affinity to the A-site is much lower. While most of the new compounds show activity in the same order
Since the successful cleavage of an RNA phosphodiester
bond requires substantial conformational flexibility, it was
important to investigate the possibility that the standard
incubation temperature of 37°C is not high enough to allow the
“installed warheads” to reach the activation energy required for
of magnitude as the parent NeoB, compounds 6, 9 and 10 with a
4’-nitrogen have a 2-fold inhibitory potency compared to NeoB
(IC50 values of 0.006, 0.005, 0.006 and 0.01 for 6, 9, 10, and
NeoB respectively). This could be explained by the additional
interactions of the 4’-amide (compound 6) and 4’-amine
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
Figure 3. Cleavage experiments of E.coli ribosomes in the presence of ethylenediamine, colicin E3, NeoB and compound 3. A) Lane 1, E.coli ribosomes (control).
Lanes 2-5, ribosomes treated with increasing concentrations of ethylenediamine. B) Lane 1, (control). Lanes 2-6, ribosomes treated with decreased concentrations
of colicin E3. C) Lane 1, (control); Lane 2, ribosomes treated with 7.3 M colicin E3 (ColE3); Lanes 3-6, ribosomes treated with increasing concentrations of NeoB.
D) Lane 1, (control). Lanes 2-4, ribosomes treated with increasing concentrations of compound 3. rRNA fragments were analyzed on 6% acrylamide TBE/urea gel,
stained with syber gold, and analyzed by fluorescence
(compounds 9 and 10) of these compounds with the ribosomal A-
site. However, the increased binding affinity of the compound for
the A site may result in its slow dissociation from the ribosome
(very small koff value), which would then adversely affect the
catalytic performance. To test this possibility, we synthesized
compound 1, which is a smaller analog of compound 2, as it lacks
rings III and IV of NeoB. We found that the inhibition potency of
compound 1 (IC50 = 2.03 µM, a neamine derivative) is two orders
of magnitude lower than the other compounds tested and is one
order of magnitude lower than that of the parent neamine scaffold
(IC50 = 0.28 µM). Thus, it would appear that the neamine
pharmacophore is not a suitable choice for the aminoglycoside
scaffold unless the catalytic efficiency of the warhead is greatly
improved, since the attachment of the warheads has a deleterious
effect on the binding affinity at the target.
(~1540 nucleic bases) in a dose-dependent manner. Figure 3 also
shows the 5S and tRNA fragments. However, the experiments
with NeoB and compound 3 (Figure 3, panels C and D) did not
show any signs of the cleaved product at concentrations up to 400
M. At higher concentrations, we encountered solubility problems,
which prevented us from detecting RNA cleavage. At the same
concentration range (up to 400 M), ethylenediamine (a negative
control) did not cleave the full ribosome (Figure 3, panel A),
suggesting that it is unable to bind to rRNA effectively.
To avoid precipitation and allow the use of higher
concentrations of aminoglycosides, instead of ribosomal particles,
we decided to use an A-site oligonucleotide model. We selected
an oligonucleotide model similar to that used by Westhof and co-
workers[36],[37] for crystallographic studies. To improve the RNA
detection, we added a fluorescent tag Cy3 at the 3' end (and not
at 5’ end) in order to ensure that there is a significant difference
between the size of the full length RNA and the cleaved RNA (see
Figure S2 in Supporting Information). The cleavage experiments
indicated that with the ethylenediamine (N-2-N) we observe
nonspecific cleavages only at high concentrations, 100 and 200
M of N-2-N (Figure 4, panel A).
In the presence of compound 6, we detected some RNA
cleavage at substantially lower concentrations, 10 M (Figure 4,
panel B). We observed the double-strand RNA (DS band in panel
B) suggesting that the aminoglycoside binding stabilizes the
double-stranded RNA even though the gel was under denaturing
conditions. In addition, only nonspecific cleavage bands were
observed at the concentrations tested which are far longer
fragments than expected for a specific and selective cleavage
(less than 8 bases).
2.5. RNase activity tests
Since compounds 1-10 did not show substantially higher
antibacterial activity than NeoB against WT bacteria, we decided
to directly assess the potential RNase activity of these
compounds by using gel electrophoresis experiments as
previously reported for ColE3[20]
.
We envisioned that by
incubating the designer structures with either the complete 70S
ribosomes, the isolated 30S ribosomal particles, and/or the
synthetic A-site oligonucleotide model structures, we could
determine whether the new compounds have catalytic activity.
Initially, the experiments were performed on full-size
ribosomes isolated from E. coli as previously reported[35]. As a
positive control, we used the RNase domain of the natural toxin
ColE3 obtained from Prof. Colin Kleanthous from Oxford
University (Figure 3, panel B). As expected, with ColE3 we
observed the cleavage of ~40 bases from the 16S rRNA fragment
All in all, we conclude that the initial NeoB derivatives
(compounds 1-10) that we designed and tested, exhibit significant
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
antibacterial activity and overcome existing resistance to
aminoglycosides, however they lack any significant catalytic
activity as we would have expected according to our initial design
strategy.
2.6. MD simulations
2.6.1. Conformational dynamics of the warheads and the
possibility of RNA cleavage
In an attempt to explain the experimental data of the newly
designed compounds 1-10 at the molecular level, we performed
full-atom molecular dynamics (MD) followed by Gaussian
accelerated MD (GaMD)[38],[39]. The crystal structure of NeoB
bound to the oligonucleotide model of the A-site rRNA (PDB code:
2ET4)[19] was used as a template for building the systems used in
the simulations (for details see Simulation Methods in Supporting
Information, Figure S6). Four representative derivatives of NeoB,
compounds 2, 5, 8 and 10, were simulated and NeoB was used
as a control. Total MD and GaMD simulation time was about 5.5
µs.
For compounds 2 and 5, we found two and three different
conformations of the warheads, respectively (Figures S3 and S4).
For both, the dominant conformation of the warhead (82.7% of the
population in 2, and 76.4% of the population in 5) is characterized
by a common intra-molecular hydrogen bond between the N1
amine of the warhead), and the N6' ammonium of the
aminoglycoside ring I. Unfortunately, these intra-molecular
hydrogen bonds prevent the N1 amine of the warhead from acting
as the general base to activate 2’-OH of the G1491 ribose as a
nucleophile (see Figure 2 for the proposed mechanism).
Figure 5. The normalized occurrence of the three binding modes of compound
8 warhead to A site as a function of the two intermolecular distances N3-O2'
and N4-(OP1, OP2). The representative structures are presented. For clarity,
only ring I of the aminoglycoside (in green) and hydrogen atoms crucial for
interactions of the warhead are shown. Black dashed lines denote donor-
acceptor short-range interactions.
In contrast to 2 and 5, the warhead of 8 and 10, which
represent the 4’- and 6’-amide derivatives of NeoB, is longer and
does not form any similar intra-molecular interactions with the rest
of the molecule (Figure 5 and Figure S5). For compound 10
(Figure S5), we observed that the largest conformational
variability of the warhead is associated with its rotation around the
dihedral angle N2-C3-C4-N3. Therefore, we used this coordinate
in the clustering analysis and found two major conformations of
the warhead (72.2 % and 27.8% of the population). The short-
range contacts of the N4' ammonium with the phosphates of
A1492 and A1493 conformationally restrict not only the position
of ring I in the A site but also the A1492 and A1493 backbone
atoms.
For compound 8, we identified three principal modes of
binding of the warhead to the rRNA (Figure 5). In the most
abundant binding mode (58.6% of the population), we observed
two short-range interactions between the 2’-hydroxyl of the ribose
(G1491) and the N3 amine of compound 8 (proposed general
base) and between the A1492 phosphate and the N4 ammonium
of compound 8 (general acid). Importantly, this conformational
Figure 4. Cleavage experiments of the A-site oligonucleotide model rRNA (23
bases labeled at 3’ with the fluorescent Cy3 tag, incubated for 24 hr, pH 8, 37 °C;
for the sequence of rRNA see Figure S2 in Supporting Information) in the
presence of ethylenediamine and compound 6. A) Lane 1, RNA markers; Lane
2, blank lane; Lane 3, not treated (control); Lanes 4-7, rRNA oligonucleotide
treated with increased concentrations of ethylenediamine. B) Lane 1, RNA
markers; Lanes 2-7, rRNA oligonucleotide treated with increased
concentrations of compound 6; Lanes 8 and 9, rRNA oligonucleotide treated
with 500 M 1,2-cyclohexane diamine (Cyclo) and ethylenediamine (N2N),
respectively; Lane 10, not treated (comtrol). rRNA fragments were analyzed on
20% TBE/urea gel and visualized by fluorescence. DS, double stranded rRNA.
SS, single stranded rRNA.
state of compound
8 is consistent with the hypothetical
mechanism of the A-site rRNA cleavage between G1491 and
A1492 (Figure 2). In the second most abundant binding mode of
compound 8 (21.6% of the population), the N2 amine hydrogen
bonds with the 2’-hydroxyl of the ribose of G1491, which actually
now serves as the general base. However, the concomitant
stabilization of the transition state via the interaction of the
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
warhead amines with the phosphate of A1492 is lacking, which
would significantly limit the efficiency of hydrolysis.
interactions necessary to cleave the rRNA. Therefore, together
with compound 8, we consider 10 as a lead compound for the next
generation of new aminoglycoside derivatives that are designed
to achieve rRNA catalysis.
Considering the formation of the interactions required for the
A-site rRNA cleavage between G1491 and A1492, compound 8
seems the best candidate among the selected derivatives for the
development of a catalytic antibiotic. The N3 amine of the
warhead activates the 2’-OH of the G1491 ribose for the
nucleophilic attack and the N4 ammonium of the warhead
favorably binds to the OP2 and O3' atoms of the A1492 phosphate,
facilitating the nucleophilic attack. However, the warhead hardly
interacts with the O5' atom and consequently the activation of the
leaving group would be minimal. We therefore predict that
elongating the warhead might be beneficial so that the N4
ammonium could interact with the A1492 phosphate-leaving
group more directly and frequently.
In general, the efficiency of rRNA hydrolysis is highly
dependent on the ability of the catalyst to induce the correct
positioning of the nucleophile for in-line attack on the scissile bond.
Enzymes, being large, can mechanically achieve this step ‘easily’
by distorting the substrate to reach the conformation necessary
for efficient catalysis. For example, ColE3[20] and α-sarcin[40], the
two bacterial toxins that cleave a single phosphodiester bond of
rRNA (in the small and in the large ribosomal subunit,
respectively), both use RNA base flipping in order to dock the
substrate into the active site in such a manner as to facilitate the
crucial in-line attack. Whether the aminoglycoside-warhead
combination can induce a similar conformational change of the
rRNA A-site is one of the most important questions of this work.
To address this question, we measured the angle created
between the 2’-OH (the G1491 ribose, the nucleophile), the
phosphorus of the phosphate between G1491 and A1492, and
the 5’O (the leaving group) – O—P-O angle.
Figure 6. The distribution of the O—P-O angle for NeoB and compounds 2, 5,
8 and 10 in GaMD simulations. The inset shows the O—P-O angle in the
representative structures of NeoB and compound 10. The donor-acceptor short-
range interactions important for stabilization of the O—P-O angle are marked
by black dashed lines. For clarity, only ring I of the aminoglycosides (in green)
and selected hydrogen atoms are shown.
In the crystal structure of the Westhof model that we used
for the simulations, the O—P-O angle does not exceed 90
degrees. Figure 6 compares the distribution of the O-P-O angle
as obtained from GaMD simulations of NeoB, and its derivatives
2, 5, 8 and 10. The smallest values of this angle are found for the
NeoB complex, distributed in the range of 45-105 degrees. The
interactions formed between O3’ of NeoB (ring I) and OP2 of the
A1492 phosphate, and between O4’ of NeoB (ring I) and OP2 of
the A1493 phosphate seem to be the most important for the O-P-
O angle orientation. The modifications introduced into the NeoB
ring I to make compounds 8 and 10, clearly led to the increase of
this angle for both derivatives, reaching as high as 170 degrees
for 10. We found that for the nearly linear orientation of the O-P-
O angle, a strong and stable interaction with the OP1 atom of the
A1492 phosphate is essential. For compound 8, this requirement
is fulfilled thanks to the persistent hydrogen bond formed by the
3’-hydroxyl of ring I and concomitant stabilization of the A1492
phosphate by the N4 ammonium of the warhead. For compound
10, the crucial short-range interaction with the OP1 atom of the
A1492 phosphate is made by the N4' ammonium in ring I. Even
though the warhead of compound 10 does not reach the putative
rRNA cleavage site, the stabilization of the – O—P-O angle in the
nearly in-line orientation is remarkable. Thus, we believe that
using flexible docking algorithms, the structure of the warhead of
compound 10 can be remodeled to optimize its ability to form the
2.7 Summary and Conclusions
We adopted the idea of catalytic antibiotics from nature:
bacteria produce bacteriocins that exert their lethal action by an
enzymatic nuclease digestion mechanism. Based upon the most
recent structural and mechanistic data on the bacteriocin ColE3
and on the aminoglycoside antibiotics, we have designed and
synthesized a series of aminoglycosides bearing a potential
catalytic moiety, with the expectation that these molecules could
mimic ColE3 activity. Our design principles included a careful
analysis of the choice of the ‘target’ phosphodiester bond, the
‘catalytic warhead’ structures, and the attachment site on the
aminoglycoside scaffold. We selected the phosphodiester bond
between the rRNA bases G1491 and A1492 as the potential
cleavage site, the 4’-OH group (ring I) of the natural
aminoglycoside NeoB as an attachment site for the catalytic
warheads, and by attaching a series of different 1,2-diamines as
potential catalytic warheads we prepared the new NeoB
derivatives (compounds 1-10, Figure 1).
To probe the influence of the attached warheads on
antibacterial activity, the derivatives 1-10 were tested against WT,
pathogenic and aminoglycoside resistant strains (Table 1). In
general, the compounds 2-10 showed significant antibacterial
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
activity against WT bacteria, which was similar to or slightly lower
than that of the parent NeoB. The observed antibacterial data was
corroborated by the in vitro protein translation inhibition data (IC50
aminoglycoside-induced rRNA cleavage; this progress may offer
promise for the development of catalytic antibiotics as a new
paradigm in antibiotics research. Thus, although ‘the catalytic
values, Table 1) showing activity similar to that of the parent NeoB. aminoglycoside’ is yet to be found, the results introduced in this
Of particular note is the potent antibacterial activity of the new
derivatives (versus that of NeoB) against resistant and pathogenic
strains like P. aeruginosa and MRSA, with the lead compounds
exhibiting MIC values 8-16 times lower in P. aeruginosa
(compound 6) and 64 times lower in MRSA (compounds 2 and 8).
We anticipated that unusually low MIC and IC50 values
against WT bacteria would be indicative of significant catalytic
activity on the ribosomal RNA, but in this regard the observed data
was not encouraging. However, these data alone cannot
conclusively disprove the desired hydrolytic activity, since they
could also be explained by a low catalytic turnover resulting from
strong binding affinity of the derivative to the cleaved rRNA.
To address whether the new derivatives exhibit the
anticipated hydrolytic activity, we tested the comparative RNase
activity using full-size ribosomes isolated from E. coli (Fig. 3) and
the bacterial A-site oligonucleotide model RNA (Fig. 4). In the E.
coli ribosomes, we confirmed the reported activity of the ColE3 as
the positive control. However, neither NeoB nor compound 3
showed any activity at concentrations of up to 400 µM; at higher
concentrations, we encountered solubility problems. The
cleavage experiments with the A-site oligonucleotide model RNA
indicated weak, dose dependent but nonspecific hydrolytic
cleavage activity for compound 6. Although other RNA models
could be employed that would more precisely monitor strand
cleavage (e.g. the use of 32P end-labeled A-site RNA construct[41]),
it is clear that if the new derivatives would exhibit the desired
function at an appreciable level, the experiments we performed
would certainly detect them. In view of these discouraging results,
it is prudent to consider whether or not there is sufficient
justification to encourage further research towards catalytic
aminoglycosides and if so then what should be the next step?
The observed similar MIC and IC50 values of 2-10 to those
of NeoB against WT bacteria indicated that the modifications on
NeoB have not hindered the bacterial cell permeability or the
binding affinity of the aminoglycoside scaffold to the target site.
Furthermore, full-atom GaMD simulations (Fig. 6) on the crystal
structure of NeoB bound to the oligonucleotide model of the A-site
rRNA revealed that the O—P-O angle for compounds 8 and 10 is
significantly greater than that of NeoB (between 45 and 105
degrees), reaching as high as 170 degrees for 10, which is
remarkably close to the ideal, in-line orientation of nucleophilic
attack. These data support the notion that appropriately designed
aminoglycoside-warheads such as 8 or 10, have the capacity 1)
to bind selectively to the regular aminoglycoside binding site and
2) to induce the conformational changes that are necessary to
lower the activation barrier of the transition state for hydrolytic
cleavage. Taken together, these results suggest that using
flexible docking algorithms, the structures of the warheads in
compounds 8 and 10 could be re-modelled to optimize their ability
to catalytically cleave the rRNA.
study indicate that this is an achievable goal. More
comprehensive design strategies are now being employed, which
incorporate advanced molecular dynamics techniques and more
powerful metal-free and metal-based catalytic warheads, to
optimize the catalytic activity of the new designer structures.
3. Experimental Section
Experimental Details. 3.1. General Techniques: NMR spectra (including
1H, 13C, DEPT, 2D-COSY, 1D TOCSY, HMQC, HMBC) were routinely
recorded on a Bruker AvanceTM 500 spectrometer, and chemical shifts
reported (in ppm) are relative to internal Me4Si (δ=0.0) with CDCl3 as the
solvent, and to MeOD (δ=3.35) as the solvent. 13C NMR spectra were
recorded on a Bruker AvanceTM 500 spectrometer at 125.8 MHz, and the
chemical shifts reported (in ppm) relative to the solvent signal for CDCl3 (δ
=77.00), or to the solvent signal for MeOD (δ=49.0). Mass spectra
analyses 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)Mo7O24.4H2O (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 under an argon atmosphere with anhydrous
solvents, unless otherwise noted. Neomycin B and Paromomycin as
analytical samples for comparative biochemical assays were purchased
from Sigma. For the chemical synthesis, large scale Paromomycin (used
as a starting material) was purchased from Apollo Scientific LTD (SK6 2QR
United Kingdom). All other chemicals and biochemicals, unless otherwise
stated, were obtained from commercial sources. In all biological tests, all
the tested aminoglycosides were in their sulfate salt forms except
compound 5, which was used as its trifluoroacetate salt.
3.2. Biochemical assays: Prokaryotic in-vitro translation inhibition by the
different standard and synthetic aminoglycosides was quantified in
coupled transcription/translation assays by use of E. coli S30 extract for
circular DNA with the pBESTlucTM plasmid (Promega), according to the
manufacturer protocol. Translation reactions (25 µL) containing variable
concentrations of the tested aminoglycoside were incubated at 37 ºC 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-
diaminocyclohexanetetraacetate (2 mM), glycerol (10%), triton x100 (1%)
and BSA (1 mg/mL)) into 96-well plates. The luminescence was measured
immediately after the addition of Luciferase Assay Reagent (Promega) (50
µL), and light emission was recorded with Victor3TM Plate Reader (Perkin-
Elmer). The concentration of half-maximal inhibition (IC50) was obtained
from fitting concentration-response curves to the data of at least three
independent experiments, using Grafit 5 software[42]
.
Comparative antibacterial activities were determined by measuring the
MIC values using the double-microdilution method according to the
National Committee for Clinical Laboratory Standards (NCCLS)[43]. All the
experiments were performed in triplicates and analogous results were
obtained in three different experiments.
In summary, this pilot study provides a new direction for the
development of novel aminoglycoside-based small molecules that
target bacterial rRNA by means of optimizing the efficacy of
For the rRNA cleavage experiments, the ribosomes were isolated from E.
coli cells (R477-100) by following the reported protocol.[35] Ribosomes are
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
pelleted from pooled fractions (35 K for 15 h at 4°C) and resuspended in
buffer for snap freezing in liquid nitrogen and storage at -80 °C. The resin
is rinsed with water after use and stored in 20% ethanol at 4°C. The
catalytic domain of ColE3 (provided by Prof. Klenathous) was purified from
its immunity protein as previously described[44]. Briefly, after elution from
the Ni-affinity column with 6M GnHCl, the ColE3 RNase becomes unfolded.
It refolds upon dialysis into 50 mM potassium phosphate or 20 mM Tris pH
7.5 buffer. All the purification procedure can be carried out at room
temperature and the product was analyzed on 16% SDS-PAGE.
The authors thank Prof. Colin Kleanthous from Oxford
University for providing ColE3 toxin, Dr. Eli Shulman for
performing the initial RNase tests by gel electrophoresis, and Dr.
Moran Shalev for helping in the initial design and modelling
experiments. This work was supported by research grants from
the Israel Science Foundation founded by the Israel Academy of
Sciences and Humanities (grant no. 1845/14 for TB), from the
Ministry of Science, Technology and Space, State of Israel (grant
no. 2022675 for TB), by the Interdisciplinary Centre for
Mathematical and Computational Modelling University of Warsaw
(grants G31-4, GA65-16, GA73-21 for JT and TP) and by the
National Science Centre, Poland (UMO-2017/26/M/NZ1/00827).
J. T. acknowledges the Polish-U.S. Fulbright Commission. V.B.
acknowledges the financial support by the Ministry of Immigration
Absorption and the Ministry of Science and Technology, Israel
(Kamea Program).
The cleavage experiments of rRNA with E. coli ribosomes were performed
by incubation of freshly isolated ribosomes for 24 h (5 minutes in case of
ColE3) (37°C, pH 7.0) in the presence of ethylenediamine, NeoB
compound 3 or ColE3. After incubation, RNA was phenol/chloroform
extracted from samples and electrophoresed on a 6% acrylamide
TBE/urea gel for 100 min at 180 V, stained with syber gold, and analyzed
by fluorescence. Short RNA oligomer that represents the bacterial A-site
sequence labeled with fluorescent tag (23 bases, for sequence see Figure
S2) was also used for rRNA cleavage experiments. This RNA sequence
was purchased from Dharmacon and was used without further purification.
The cleavage experiments were performed by using gel electrophoresis;
the rRNA fragments were analyzed on 20% TBE/urea gel and visualized
by fluorescence.
Conflicts of interest
The authors declare no conflict of interest.
3.3. Molecular dynamics simulations. MD simulations were carried out on
the model of the A site containing two symmetric aminoglycoside binding
sites using the crystal structure of the A site with neomycin B bound (PDB
Keywords: Aminoglycosides • Catalytic Antibiotic • RNA
Cleavage • Bacterial Ribosome • Decoding A-site
code: 2ET4)[19]
. The MD simulation protocol consisted of energy
minimization, thermalization, equilibration, and production phases. In the
first two phases, harmonic constraints with the force constant of 10
kcal/mol/Å2 were imposed on heavy atoms of the solute. First, all systems
were energy minimized with the above restraints undergoing 5000 steps
of steepest descent followed by 4000 steps of conjugate gradient
minimization using sander (Amber 12). The next phases were carried out
with NAMD[45]. Second, during thermalization (in the NVT ensemble), each
system was heated from 10 to 310 K increasing the temperature by 10 K
every 100 ps. Then 2 ns simulations at 310 K were carried out. Third,
equilibration was performed in the NpT ensemble with a constant pressure
of 1 atm controlled using Langevin Piston method and at constant
temperature of 310 K regulated by Langevin dynamics with a damping
factor of 1 ps-1. During 5 ns equilibration the restraints were exponentially
decreased in 50 time windows (scaled from 1 to 0.0065). Further, the 120
ns production runs were performed without any restraints. Periodic
boundary conditions and Particle Mesh Ewald method with grid spacing of
1 Å were used. The SHAKE algorithm and the integration time step of 2 fs
were applied. For non-bonded interactions a short-range cutoff of 12 Å was
used.
[1]
B. Bozdogan, L. Ednie, K. Credito, K. Kosowska, P. C. Appelbaum,
Antimicrob. Agents Chemother. 2004, 48, 4762–4765.
B. I. Eisenstein, Expert Opin. Investig. Drugs 2004, 13, 1159–1169.
V. Pokrovskaya, I. Nudelman, J. Kandasamy, T. Baasov,
Aminoglycosides. Redesign Strategies for Improved Antibiotics and
Compounds for Treatment of Human Genetic Diseases, 2010.
J. Suh, W. S. Chei, Curr. Opin. Chem. Biol. 2008, 12, 207–213.
Z. Yu, J. A. Cowan, Chem. - A Eur. J. 2017, 23, 14113–14127.
T. Niittymaki, H. Lonnberg, Org. Biomol. Chem. 2006, 4, 15–25.
T. Lönnberg, K. M. Kero, Org. Biomol. Chem. 2012, 10, 569–574.
R. Salvio, R. Cacciapaglia, L. Mandolini, F. Sansone, A. Casnati,
RSC Adv. 2014, 4, 34412–34416.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
S. Magnet, J. S. Blanchard, Chem. Rev. 2005, 105, 477–498.
A. Sreedhara, A. Patwardhan, J. A. Cowan, Chem. Commun. 1999,
2, 1147–1148.
[10]
After the above MD production phase, the GaMD simulation[38],[39] started
with another 2 ns MD simulation to gather the statistics of the potential
energy of the system in order to calculate the GaMD acceleration
parameters. After adding the boost potential, the simulation was continued
for 30 ns to equilibrate the system. Subsequently, ten independent GaMD
production runs were conducted for 100 ns each starting with randomized
initial atomic velocities. The GaMD simulations were performed in the dual-
boost mode in which the boost potential is applied to the dihedral and total
potential energy terms. The threshold energy was set to the lower bound,
i.e., E = Vmax. The upper limit of the boost potential standard deviation, σ0,
was set to 10 kcal/mol for the dihedral and total potential energetic terms.
[11]
[12]
A. Patwardhan, J. A. Cowan, Dalt. Trans. 2011, 40, 1795–1801.
W. Szczepanik, A. Krezel, M. Brzezowska, E. Dworniczek, M.
Jezowska-Bojczuk, Inorganica Chim. Acta 2008, 361, 2659–2666.
W. Szczepanik, J. Ciesiołka, J. Wrzesiński, J. Skała, M. Jezowska-
Bojczuk, Dalt. Trans. 2003, 1488–1494.
[13]
[14]
W. Szczepanik, E. Dworniczek, J. Ciesiołka, J. Wrzesiński, J. Skała,
M. Jezowska-Bojczuk, J. Inorg. Biochem. 2003, 94, 355–364.
S. M. K. Takahashi, Acad. Press. New York 1982, 435–468.
X. J. Yang, T. Gerczei, L. Glover, C. C. Correll, Nat. Struct. Biol.
2001, 8, 968–973.
[15]
[16]
[17]
[18]
P. B. Rupert, A. R. Ferre-D’Amare, Nature 2001, 410, 780–786.
C. L. Ng, K. Lang, N. A. G. Meenan, A. Sharma, A. C. Kelley, C.
Kleanthous, V. Ramakrishnan, Nat. Struct. Mol. Biol. 2010, 17,
Acknowledgements
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
1241–1246.
Agents Chemother. 2005, 49, 3734–3742.
[19]
B. François, R. J. M. Russell, J. B. Murray, F. Aboul-ela, B.
[34]
M. Hainrichson, O. Yaniv, M. Cherniavsky, I. Nudelman, D. Shallom-
Shezifi, S. Yaron, T. Baasov, Antimicrob Agents Chemother 2007,
51, 774–776.
Masquida, Q. Vicens, E. Westhof, Nucleic Acids Res. 2005, 33,
5677–5690.
[20]
C. L. Ng, K. Lang, N. A. G. Meenan, A. Sharma, Nat. Struct. &
2010, DOI 10.1038/nsmb.1896.
[35]
[36]
B. a Maguire, L. M. Wondrack, L. G. Contillo, Z. Xu, RNA 2008, 14,
188–195.
[21]
[22]
[23]
M. Komiyama, K. Yoshinari, J. Org. Chem. 1997, 62, 2155–2160.
K. Yoshinari, M. Komiyama, Chem. Lett. 1990, 519–522.
R.-B. B. Yan, M. Yuan, Y. F. Wu, X. F. You, X.-S. S. Ye, Bioorg.
Med. Chem. 2011, 19, 30–40.
P. Pfister, S. Hobbie, Q. Vicens, E. C. Böttger, E. Westhof,
ChemBioChem 2003, 4, 1078–1088.
[37]
[38]
Q. Vicens, E. Westhof, Chem. Biol. 2002, 9, 747–755.
Y. Miao, V. A. Feher, J. A. McCammon, J. Chem. Theory Comput.
2015, 11, 3584–3595.
[24]
K. C. Nicolaou, V. a. Adsool, C. R. H. Hale, Org. Lett. 2010, 12,
1552–1555.
[39]
[40]
[41]
[42]
[43]
Y. T. Pang, Y. Miao, Y. Wang, J. A. McCammon, J. Chem. Theory
Comput. 2017, 13, 9–19.
[25]
[26]
N. S. Chindarkar, A. H. Franz, Arkivoc 2008, 2008, 21.
R. Pathak, D. Perez-Fernandez, R. Nandurdikar, S. K. Kalapala, E.
C. Bottger, A. Vasella, Helv. Chim. Acta 2008, 91, 1533–1552.
R. Pathak, E. C. C. Böttger, A. Vasella, Helv. Chim. Acta 2005, 88,
2967–2985.
C. C. Correll, X. Yang, T. Gerczei, J. Beneken, M. J. Plantinga, J.
Synchrotron Radiat. 2004, 11, 93–96.
[27]
M. J. Belousoff, B. Graham, L. Spiccia, Y. Tor, Org. Biomol. Chem.
2009, 7, 30–33.
[28]
[29]
E. D. Goddard-borger, R. V Stick, 2007, 7515–7517.
V. Pokrovskaya, V. Belakhov, M. Hainrichson, S. Yaron, T. Baasov,
J. Med. Chem. 2009, 52, 2243–2254.
Leatherbarrow, R. J. GraFit 5; Erithacus Software Ltd.: Horley, U.K.,
2001.
NCCLS, National Committee for Clinical Laboratory Standards,
Performance Standards for Antimicrobial Susceptibility Testing. Fifth
Information Supplement: Approved Standard M100-S5, NCCLS,
Villanova, Pa., 1994.
[30]
J. Kondo, M. Hainrichson, I. Nudelman, D. Shallom-Shezifi, C. M. C.
M. Barbieri, D. S. D. S. Pilch, E. Westhof, T. Baasov, Chembiochem
2007, 8, 1700–1709.
[31]
[32]
[33]
M. T. G. Holden, E. J. Feil, E. Al., Proc. Natl. Acad. Sci. 2004, 101,
9786–9791.
[44]
[45]
S. Carr, D. Walker, R. James, C. Kleanthous, A. M. Hemmings,
Structure 2000, 8, 949–960.
G. Kaneti, H. Sarig, I. Marjieh, Z. Fadia, A. Mor, FASEB J. 2013, 27,
4834–43.
J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E.
Villa, C. Chipot, R. D. Skeel, L. Kalé, K. Schulten, J. Comput. Chem.
2005, 26, 1781–1802.
J. I. Sekiguchi, T. Asagi, T. Miyoshi-Akiyama, T. Fujino, I.
Kobayashi, K. Morita, Y. Kikuchi, T. Kuratsuji, T. Kirikae, Antimicrob.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800549
ChemBioChem
FULL PAPER
FULL PAPER
Text for Table of Contents
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.