Di-alkylated paromomycin derivatives: Targeting the membranes of Gram positive pathogens that cause skin infections

Article abstract of DOI:10.1016/j.bmc.2013.03.046

A collection of paromomycin-based di-alkylated cationic amphiphiles differing in the lengths of their aliphatic chain residues were designed, synthesized, and evaluated against 14 Gram positive pathogens that are known to cause skin infections. Paromomyci

Full text of DOI:10.1016/j.bmc.2013.03.046

Bioorganic & Medicinal Chemistry 21 (2013) 3624–3631
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Di-alkylated paromomycin derivatives: Targeting the membranes
of Gram positive pathogens that cause skin infections
⇑
Yifat Berkov-Zrihen, Ido M. Herzog, Mark Feldman, Adar Sonn-Segev, Yael Roichman, Micha Fridman
School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 2 April 2013
A collection of paromomycin-based di-alkylated cationic amphiphiles differing in the lengths of their ali-
phatic chain residues were designed, synthesized, and evaluated against 14 Gram positive pathogens that
are known to cause skin infections. Paromomycin derivatives that were di-alkylated with C7 and C8 linear
aliphatic chains had improved antimicrobial activities relative to the parent aminoglycoside as well as to
the clinically used membrane-targeting antibiotic gramicidin D; several novel derivatives were at least
16-fold more potent than the parent aminoglycoside paromomycin. Comparison between a di-alkylated
and a mono-alkylated paromomycin indicated that the di-alkylation strategy leads to both an improve-
ment in antimicrobial activity and to a dramatic reduction in undesired red blood cell hemolysis caused
by many aminoglycoside-based cationic amphiphiles. Scanning electron microscopy provided evidence
for cell surface damage by the reported di-alkylated paromomycins.
Keywords:
Amphiphilic aminoglycosides
Cationic amphiphiles
Membrane-targeting antibiotics
Skin infection causing bacteria
Hemolysis
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
positive bacteria and are also highly toxic when used internally.9,10
The clinical efficacy of the aminoglycoside neomycin B is continu-
Bacterial skin infections, including chronic infections related to
diabetes, venous stasis, or arterial insufficiency account for a sig-
nificant percentage of infectious diseases.1–3 Patients with infected
wounds are frequently treated with systemic antibiotics and, in
addition, topical antibiotic treatments.4,5 There is a large repertoire
of potent antimicrobial agents that are unfit for internal use due to
their toxicity but that are tolerated topically.6 These can be used in
treatment of skin infections caused by multidrug-resistant organ-
isms that are unaffected by systemic antibiotic treatment.
Amongst the frequently used topical antimicrobial agents are
gramicidins, polymyxins, and aminoglycosides such as neomycin
B.4–7 The heterogeneous oligo-peptide mixture of gramicidins is
effective mainly against Gram positive bacteria but not against most
Gram positive bacilli. Gramicidins are also highly hemolytic, making
them too toxic for internal use.6,8,9 The membrane targeting cyclic
lipopeptide antibiotics polymyxins are active against several Gram
negative pathogens, but polymyxins are not active against Gram
ously reduced as an ever increasing number of bacterial strains ac-
quire resistance to this aminoglycoside antibiotic.11,12 Topical use of
neomycin B may be accompanied by undesired contact dermatitis
side effects.13 Aminoglycosides are also highly nephrotoxic and oto-
toxic when used internally.14,15 Although topical antibiotic treat-
ment is tolerated with gramicidins, polymyxins and neomycin B,
side effect occur when these antibiotics are used internally, or if
high doses of these toxic antimicrobial agents can make their way
into the blood system through open wounds or highly damaged
external tissue. Therefore, there is a constant need for topical anti-
biotics that are effective against a wide spectrum of bacteria and ex-
hibit minimal toxic side effects.
In recent years, we and others have demonstrated that the
attachment of hydrophobic residues to aminoglycoside antibiotics
results in cationic amphiphiles that possess potent antimicrobial
activities.16–21 Several lines of evidence support the hypothesis
that these cationic amphiphiles act by disrupting bacterial mem-
branes.16,20 We previously developed and studied a library of
tobramycin-based cationic amphiphiles generated by the alkyl-
ation of the 600 position of this penta-amino aminoglycoside with
series of aliphatic chains (Fig. 1). Of the various hydrophobic resi-
dues that were tested, tobramycin-based cationic amphiphiles
with C12, C14, and C16 linear aliphatic chains demonstrated the
most potent antimicrobial activity against a wide selection of both
Gram positive and Gram negative bacteria.16,17 The selectivity of
the most potent tobramycin-based cationic amphiphiles for bacte-
rial membranes was tested by measuring the damage that these
Abbreviations: BHI, brain heart infusion; BOC, tert-butoxycarbonyl; 1-D-TOCSY,
total correlation spectroscopy; DMF, dimethylformamide; EtOAc, ethyl acetate; HR-
ESI, high resolution electron spray ionization; LR-ESI, low resolution electron spray
ionization; MIC, minimum inhibitory concentration; MTT, thiazolyl blue tetrazo-
lium bromide; PBS, phosphate buffered saline; RBC, red blood cells; rt, room
temperature;
SEM,
scanning
electron
microscope;
TIBSCl,
2,4,6-tri-
isopropylbenzenesulfonyl chloride; TFA, trifluoroacetic acid; TLC, thin layer
chromatography.
⇑
Corresponding author. Tel.: +972 3 640 8687.
E-mail address: mfridman@post.tau.ac.il (M. Fridman).
0968-0896/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2013.03.046

Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
3625
physiological conditions. Second, unlike tobramycin, paromomycin
(1) has two primary alcohols therefore making it possible to readily
and chemo-selectively alkylate these two alcohols in the presence
of the six secondary alcohols of this aminoglycoside. In our previ-
ous studies with tobramycin-based cationic amphiphiles, the opti-
mal antimicrobial activity was obtained by attaching C12, C14, and
NH2
6'
HO
O
1'
H2N
2
H2N
O
NH2
HO
6
O
6'' R
OH
1''
O
HO
H2N
C16 alkyl chains to the aminoglycoside. A dramatic drop in antimi-
crobial activity was observed for tobramycin-based amphiphiles
with shorter or longer alkyl chains. To maintain the hydrophobic-
ity/hydrophilicity ratio that was optimal in the case of the potent
tobramycin cationic amphiphiles, we used C6, C7, and C8 alkyl
chains for the preparation of the di-alkylated paromomycins.
The five amine groups of commercially available paromomycin
(1) were protected with Boc groups to afford the penta-NHBoc pro-
tected paromomycin derivative 1a in 85% yield (Scheme 1A).25,26
Selective conversion of the two primary alcohols at positions C-60
and C-500 of 1a to the corresponding O-trisyl leaving groups using
30 equiv of 2,4,6-triisopropylbenzene-sulfonyl chloride gave com-
pound 1b in 86%. Compound 1b was reacted with 1-n-hexanethiol,
1-n-heptanethiol, or 1-n-octanethiol resulting in the NHBoc-pro-
tected di-alkylated paromomycin derivatives 2a, 3a, and 4a in
yields ranging from 76% to 80%. Removal of the NHBoc protecting
groups in neat TFA gave the penta TFA salts of the 60, 500 dithioether
paromomycin derivatives 2, 3, and 4 with no need for further puri-
fication. When the penta-NHBoc paromomycin 1a was reacted
with 20 equiv of 2,4,6-triisopropylbenzene-sulfonyl chloride, a
mixture of the di-trisylated product 1b and of the mono-trisylated
product 5a was obtained (43% and 33% isolated yields, respectively,
Scheme 1B). NMR characterization of 5a confirmed that the O-
trisylation took place on the C-60 primary alcohol of paromomycin
(confirmed by 1D-TOCSY NMR performed on a sample of 5a after
the removal of the NHBoc groups). The 60-O-trisyl of 5a was dis-
placed by 1-n-hexadecanethiol resulting in compound 5b; 5b
was treated with TFA to yield the C16 aliphatic chain thioether 5.
Compound 5 was prepared as the single aliphatic chain anchor
analogue of the di-C8 aliphatic chain paromomycin analogue 4.
R = OH, (ribosome targeting Tobramycin)
R = linear aliphatic chains
(membrane targeting amphiphilic tobramycins)
OH
6'
O
HO
HO
1'
H2N
H2N
2
O
NH2
5''
O
HO
6
OH
O
1''
NH2
2'''
OH
O
6'''
O
OH
H2N
1'''
OH
Paromomycin (1)
Figure 1. Structures of tobramycin-based cationic amphiphiles and of the amino-
glycoside paromomycin (1).
compounds inflict on red blood cell (RBC) membranes. Hemolysis
tests revealed that there was not a linear correlation between the
antimicrobial potency and undesired hemolytic activity of these
tobramycin-based cationic amphiphiles and that the minimal
inhibitory concentrations (MICs) of the most potent compounds
were in some cases well over an order of magnitude lower than
the concentrations in which these compounds caused detectible
hemolysis.
Although the MIC and hemolysis experiments revealed that
tobramycin-based cationic amphiphiles have a degree of selectiv-
ity for membranes of some bacterial strains over red blood cells,
the therapeutic window of these compounds is still not wide en-
ough to be considered for the development of membrane-targeting
antibiotics for internal use.5,22 We were therefore interested in
exploring strategies to increase the selectivity of aminoglycoside-
based cationic amphiphiles for bacterial membranes and to widen
their potential therapeutic window.
Compared to mammalian cell membranes, bacterial mem-
branes contain a high percentage of anionic lipids.23,24 We there-
fore hypothesized that optimization of the interactions between
the positively charged amines on the aminoglycoside segment of
the cationic amphiphile and the negatively charged bacterial mem-
brane surface may improve the antimicrobial activity of these com-
pounds and their selectivity for bacterial membranes. Our strategy
was to enhance anchoring of the positively-charged aminoglyco-
side to the bacterial membrane by attaching aliphatic chains on
two different positions on the aminoglycoside scaffold.
2.2. Antimicrobial activity tests against skin infection causing
bacteria
Compounds 2–5 were screened against 14 bacterial strains
known to cause skin infections, and their minimum inhibitory con-
centrations (MICs) were determined using the double dilution pro-
tocol (Table 1).
We focused on strains belonging to two major families of Gram
positive bacteria: Staphylococci and Streptococci. Of the strains
belonging to the Staphylococci genus were pathogens such as
methicillin-resistant Staphylococcus aureus (MRSA, strain B)27
,
Staphylococcus aureus (Cowan, Strain E),28 which causes skin infec-
tions in patients with compromised immune systems such as HIV
carriers, and two strains of Staphylococcus epidermidis (strains C
and D)29 that were once regarded as harmless human skin coloniz-
ing bacteria but are now recognized as major opportunistic patho-
gens. We also tested nine strains of Streptococcus pyogenes. These
bacteria colonize mainly the throat and skin of humans and are of-
ten the cause of skin and soft-tissue infections.30
Of the tested di-alkylated paromomycin derivatives, the C6 di-
alkyl chain derivative 2 was the least active against all of the
tested strains. However, compound 2 had superior activity com-
pared to paromomycin 1 against all nine Streptococcus pyogenes
2. Results and discussion
2.1. Design and synthesis of di-alkylated paromomycin-based
cationic amphiphiles
strains F–N (MIC range from 4 to 16
and 16 to >64 g/mL for paromomycin 1). The di-C7 alkyl chain
derivative 3 was superior to compound 2 against most of the 14
tested strains (MIC range from 2 to 16 g/mL). The most potent
antimicrobial activity was observed for the di-C8 alkyl chain
derivative 4 with MICs ranging from 2 to 8 g/mL against all
lg/mL for compound 2,
l
For the preparation of di-alkylated aminoglycosides we chose
the pseudo-tetrasaccharide paromomycin (1, Fig. 1) for two rea-
sons: first, like tobramycin (Fig. 1), this aminoglycoside scaffold
has five amine functionalities that are positively charged under
l
l

3626
Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
OR
O
SR''
O
6'
A.
HO
HO
R'HN
HO
HO
R'HN
R''S
R'HN
R'HN
O
R''SH(14 eq)
Boc2O, NEt3
O
O
O
NHR'
NHR'
1
O
RO
5''
Cs2CO3, DMF
OH
OH
dioxane / H2O
O
NHR'
OH
OH
O
NHR'
O
OH
O
OH
O
R'HN
R'HN
OH
OH
R' = Boc, R'' = (CH2)5CH3; 2a, (78%)
R' = H, R'' = (CH2)5CH3; 2, (quantitative)
R' = H, R'' = (CH2)6CH3; 3, (quantitative)
R' = H, R'' = (CH2)7CH3; 4, (quantitative)
1a
R = H, R' = Boc; , (85%)
TIBSCl (30 eq.),
pyridine rt.
TFA
R' = Boc, R'' = (CH2)6CH3; 3a, (76%)
R' = Boc, R'' = (CH2)7CH3; 4a, (80%)
1b
R = TIBS, R' = Boc; , (86%)
OTIBS
S(CH2)15CH3
6'
B.
O
HO
HO
R'HN
O
HO
HO
CH3(CH2)15SH(7 eq)
Cs2CO3, DMF
R'HN
R'HN
R'HN
TIBSCl (20 eq)
O
O
NHR'
O
5''
1b
(43%)
NHR'
1a
O
HO
O
HO
pyridine rt.
OH
OH
O
OH NHR'
NHR'
OH
O
O
OH
(33%)
O
OH
O
R'HN
R'HN
OH
OH
R' = Boc; 5b, (77%)
5a
TFA
5
R' = H; , (quantitative)
Scheme 1. Synthesis of mono- and di-alkylated paromomycin-based cationic amphiphiles.
Table 1
Antibacterial activity: MIC values (
lg/mL) of mono- and di-alkylated amphiphiles (2–5), the parent drug paromomycin (1), and the membrane targeting antibiotic gramicidin D
Bacterial strainsa
Antibiotics
A
B
C
D
E
F
G
H
I
J
K
L
M
N
Gramicidin D
32
>64
32
8
4
64
64
>64
>64
16
8
8
2
8
4
2
32
2
>64
16
64
16
4
>64
>64
8
4
4
<0.5
64
16
4
2
>64
<0.5
64
8
2
2
<0.5
32
4
2
2
<0.5
32
8
4
2
<0.5
64
8
4
2
<0.5
64
8
4
2
16
64
8
4
2
2
16
4
4
2
1
2
3
4
5
>64
16
4
2
64
64
64
32
16
32
64
32
8
32
8
a
MIC values were determined against: A, S. aureus oxford NCTC6571; B, MRSA; C, S. epidermidis ATCC12228 (biofilm negative); D, S. epidermidis ATCC35984/RP62A (biofilm
positive); E, S. aureus Cowan ATCC12598; F, S. pyogenes serotype M12 (strain MGAS9429); G, S. pyogenes serotype M1T1; H, S. pyogenes serotype M2; I, S. pyogenes serotype
M3; J, S. pyogenes serotype M5; K, S. pyogenes serotype M24; L, S. pyogenes JRS75; M, S. pyogenes glossy; and N, S. pyogenes serotype T5.
tested strains. The 60-C16 linear aliphatic chain paromomycin
derivative 5, designed as the mono-alkyl chain anchor analogue
of the di-C8 alkyl chain paromomycin derivative 4, had poor
antimicrobial activity against all of the tested bacterial strains.
In addition to comparing the antimicrobial activity of the alkyl-
ated paromomycins to that of the ribosome targeting parent
antibiotic paromomycin 1, we evaluated their activity against
that of the membrane targeting oligopeptide mixture gramicidin
D.31 With the exceptions of the two Staphylococcus epidermidis
strains (C and D), gramicidin D was ineffective against the tested
Staphylococci strains (Table 1). Compounds 3 and 4 were signifi-
cantly more potent than gramicidin D against the tested Staphy-
lococci strains A, B, and E with MIC values comparable to or
better than those of gramicidin D against strains C and D. In
contrast to the lack of activity against Staphylococci strains,
more potent than compounds 2, 3, and 4 against seven of the
tested Streptococci strains, gramicidin D demonstrated very poor
activity against S. pyogenes serotype M12 (strain F) and only
moderate activity against S. pyogenes glossy (strain M; MIC
>64 lg/mL, and MIC = 16 lg/mL, respectively); compounds 2–4
were more potent than gramicidin D against these Streptococci
strains. However, gramicidin D was at least eight times more po-
tent than compounds 3 and 4 against 6 of the tested streptococci
strains (Table 1). The antimicrobial activity tests demonstrated
that the di-alkylated paromomycin amphiphiles have a broader
spectrum of antimicrobial activity than the parent aminoglyco-
side 1 or the membrane targeting antibiotic gramicidin D. The
significantly more potent antimicrobial activity of the di-C8 alkyl
chain paromomycin derivative 4 compared to that of the mono
C16 linear aliphatic chain paromomycin derivative 5 demon-
strates the favorable effect that may result from the attachment
of more than one linear aliphatic chain to the aminoglycoside
scaffold.
gramicidin
against seven of the nine tested Streptococci strains (for strains
G–L, MICs <0.5 g/mL, and for strain N, MIC = 2 g/mL). Although
D demonstrated very good antimicrobial activity
l
l

Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
3627
2.3. Red blood cells hemolysis test
derivative with the di-C7 alkyl chain, 3, was either as active or
one double dilution less active than compound 4 against the tested
bacterial strains and caused significantly less RBCs hemolysis than
compound 4. At 256 lg/mL, compound 3 caused 30.4 8.9% hemo-
lysis; compound 4 caused 62.5 7.9% hemolysis at the same con-
centration. The high level of hemolysis caused by compound 4 at
256 lg/mL was similar to that caused by gramicidin D. Hence, of
The selectivity of all of the amphiphilic paromomycin deriva-
tives 2–5 for bacterial membranes was studied by testing the
hemolytic activity of these compounds on red blood cells (RBCs)
isolated from laboratory rats. The percentage of hemolysis was
determined after 1 h of incubation with increasing concentrations
of the tested compounds (up to 256
brane targeting gramicidin D caused hemolysis even at low con-
centrations close to the MIC range of this antimicrobial agent
l
g/mL) at 37 °C. The mem-
the three di-alkylated paromomycins, the di-C7 alkyl chain deriva-
tive 3 demonstrated potent antimicrobial activities against all 14
bacterial strains and was dramatically less hemolytic then both
gramicidin D and di-C8 alkyl chain paromomycin derivative 4. Up
(2.4 1.4% at 2
lg/mL, Fig. 2). In contrast, even at concentrations
of 128 g/mL, compounds 2, 3, and 4 were significantly less hemo-
l
to a concentration of 32
less hemolysis than gramicidin D and as of a concentration of
64 g/mL and up to 256 g/mL this amphiphilic paromomycin
derivative was approximately twofold less hemolytic than gramici-
din D. Therefore, in terms of the ratio of hemolysis to antimicrobial
activity, compound 3 is the most potent of the di-alkylated parom-
omycins that were studied.
lg/mL compound 3 caused considerably
lytic then gramicidin D (Fig. 2). The aliphatic chain length affected
the percentage of hemolysis with the lowest hemolysis caused by
the di-C6 alkyl chain paromomycin derivative 2 (3.4 1.2% hemo-
lysis at a concentration of 256 lg/mL) and the highest by the di-
C8 alkyl chains paromomycin derivative 4 (62.5 7.9% at 256 lg/
l
l
mL).
Interestingly, whereas the mono C16 linear aliphatic chain par-
omomycin derivative 5 demonstrated poor antimicrobial activity
against the tested bacterial strains, this compound caused drastic
RBC hemolysis at concentrations that were lower than its MIC
2.4. Scanning electron microscopy (SEM) evidence for bacterial
cell damage
range. At a concentration of 16
l
g/mL, compound 5 already caused
In order to visualize bacterial cell morphological changes that
result from exposure to the synthetic di-alkylated paromomycin-
based cationic amphiphiles, we incubated a culture of S. epidermi-
dis ATCC12228, which was susceptible to both the parent amino-
glycoside paromomycin (1) and to the most potent of the di-
33.8 5.2% hemolysis of rat RBCs; at the same concentration com-
pounds 2, 3, and 4 caused almost no measurable hemolysis (0% for
2 and 3, and 0.8 0.3% for 4). Moreover, at a concentration of
32 lg/mL compound 5 caused almost a 100% hemolysis; the di-
alkylayed paromomycin derivative 4 and the highly hemolytic
alkylated paromomycins, compound 4 at a concentration of 1 lg/
gramicidin D caused no more than 60% hemolysis even at a con-
mL; this concentration is half of the MIC of the two selected anti-
centration of 256 lg/mL. Hemolysis experiments demonstrated
microbial agents against this bacterial strain.
that compound 5 had no selectivity for bacterial membranes and
acted in a manner similar to that of non-specific membrane dis-
rupting detergents.
Under these experimental conditions, untreated bacterial cells
had a smooth surface morphology (Fig. 3a). After 1 h of incubation
with paromomycin 1, moderate wrinkling of the bacterial cell sur-
face was observed (Fig. 3b). As paromomycin 1 is a protein synthe-
sis inhibitor, the accumulation of damaged and non-functional
membrane proteins may explain the observed wrinkled cell sur-
faces. Bacterial cells that were treated with the di-C8 alkyl chain
paromomycin derivative 4 were severely damaged (Fig. 3c). These
cells had severe surface wrinkling and roughening that resulted
from irregularly-shaped cell walls and damaged membranes. At
higher concentrations of compound 4, mainly severely damaged
cells and cellular debris were observed. The percentage of genes
in the bacterial genome that encode bacterial membrane proteins
averages between 20% and 30% depending on the strain.32 Unlike
the minor cell surface damage caused by paromomycin 1 that
At a concentration of 128 lg/mL, the di-C6 alkyl chain paromo-
mycin derivative 2 did not cause any measurable hemolysis; this
compound was also the least potent antimicrobial agent against
the 14 tested bacterial strains. The high end of the MIC range of
derivative 3 against the 14 tested bacterial strains was 16
for 11 of the tested strains the MIC of compound 3 was not higher
than 4 g/mL. However, at 32 g/mL which is two to eight times
lg/mL;
l
l
higher than the MICs, this compound caused almost no measurable
hemolysis (3.6 1.9%).
The di-C8 alkyl chain paromomycin derivative 4 demonstrated
the most potent antimicrobial activity against all of the tested bac-
terial strains (MIC range from 2 to 8 lg/mL). The paromomycin
Figure 2. Rat RBCs were incubated with tested compounds for 1 h at 37 °C: 2 (X), 3 (N), 4 (d), 5 (–ꢀ–), gramicidin D (j). All experiments were performed in triplicate, and
results are the average of at least two different sets of experiments using blood samples from different laboratory rats.

3628
Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
a
b
c
Figure 3. Scanning electron microscopic (SEM) images of S. epidermidis ATCC12228 with and without drug: (a) Untreated control bacteria cells. (b) Cells after 1 h of
incubation at 37 °C with 1 g/mL paromomycin 1. (c) Cells after 1 h of incubation at 37 °C with 1 g/mL of compound 4.
l
l
presumably results from impaired membrane protein synthesis,
the severe cell surface damage caused by compound 4 can be ratio-
nalized by direct and rapid membrane disrupting effects of this
compound.
enhancing the antimicrobial activity and in reducing the undesired
hemolytic effect.
4. Experimental
3. Summary and conclusions
4.1. General methods and instrumentation
In conclusion, three 500,60-di-alkylated paromomycin-based cat-
ionic amphiphiles differing in the length of the linear hydrophobic
chain (C6, C7, and C8 chains) were synthesized and evaluated for
their antimicrobial activity against 14 bacterial pathogens known
to cause skin infections.
Of the three chain lengths tested, the di-C8 aliphatic chain par-
omomycin derivative 4 was the most potent antimicrobial agent
against all of the tested strains and was at least 16 times more po-
tent than the membrane targeting antibiotic gramicidin D and the
parent ribosome-targeting aminoglycoside paromomycin 1 against
most of the tested Staphylococci strains and some of the tested
Streptococci strains.
However, the di-C7 aliphatic chain paromomycin analogue 3 is
of particular interest. This compound was either as effective or
one double dilution less potent than the di-C8 aliphatic chain par-
omomycin analogue 4 against all of the tested bacteria, yet this
compound caused considerably less hemolysis of laboratory rat
RBCs compared to both compound 4 and the clinically used gram-
icidin D. Moreover, the mono C16 linear aliphatic chain paromomy-
cin derivative 5, which was synthesized as the mono aliphatic
chain analogue of the di-C8 chain paromomycin analogue 4, dem-
onstrated poor antimicrobial activity against all tested bacterial
strains and caused almost 100% hemolysis of RBCs even at a con-
centrations close to the MIC.
Scanning electron microscopy (SEM) experiments revealed that
bacterial cells that were incubated with the di-C8 aliphatic chain
paromomycin analogue 4 had extensive cell surface damage com-
pared to that caused by the parent ribosome-targeting aminogly-
coside 1. This observed extensive damage caused by analogue 4
further supports the suggestion that these aminoglycoside-based
cationic amphiphiles exert their biological activity through disrup-
tion of the bacterial cell membrane.
1H NMR spectra (including 1D-TOCSY) were recorded on Bruker
Avance™ 400 or 500 spectrometers, and chemical shifts (reported
in ppm) were calibrated to CD3OD (d = 3.31) when CD3OD was the
solvent. 13C NMR spectra were recorded on Bruker Avance™ 400
or spectrometers at 100.6 or 125 MHz. Multiplicities are reported
using the following abbreviations: b = broad, s = singlet, d = doublet,
t = triplet, hep = heptet, m = multiplet. Coupling constants (J) are gi-
ven in Hertz. Low-resolution electron spray ionization (LR-ESI) mass
spectra were measured on a Waters 3100 mass detector. High-reso-
lution electron spray ionization (HR-ESI) mass spectra were mea-
sured on a Waters Synapt instrument. Chemical reactions were
monitored by TLC (Merck, Silica gel 60 F254). Visualization was
achieved using a cerium-molybdate stain ((NH4)2Ce(NO3)6 (5 g),
(NH4)6Mo7O24ꢀ4H2O (120 g), H2SO4 (80 mL), H2O (720 mL)). All
reactions were carried out under an argon atmosphere with anhy-
drous solvents, unless otherwise noted. All chemicals, unless other-
wise stated, were obtained from commercial sources. Compounds
were purified flash chromatography (SiO2, Merck, Kieselgel 60).
4.2. Synthetic procedures
4.2.1. Compound 1a
Paromomycin sulfate (1.0 g, 1.2 mmol) was dissolved in a mix-
ture of dioxane/water (2:1, 15 ml), and triethylamine (1.7 ml,
12 mmol) was added. The solution was sonicated for 10 min and
then Boc2O (2.6 g, 12 mmol) was added, and the mixture was stir-
red at ambient temperature for 12 h. Reaction progress was mon-
itored by TLC (9% methanol in dichloromethane). Upon
completion, the reaction mixture was concentrated, diluted with
ethyl acetate (150 mL), washed with brine, dried over MgSO4,
and concentrated under reduced pressure. The crude was purified
by flash column chromatography over silica gel (7.5% methanol/
dichloromethane) to afford compound 1a as a white solid (1.15 g,
85%).
The results of this study demonstrate that di-alkylation of par-
omomycin results in potent antimicrobial agents that are effective
against a broad spectrum of Gram positive pathogens. The newly
synthesized compounds caused significantly less hemolysis com-
pared to membrane targeting antibiotics such as gramicidin D. This
study also demonstrates that the di-alkylation approach may be
more favorable than the mono-alkylation approach in designing
aminoglycoside-based cationic amphiphiles both in terms of
4.2.2. Compound 1b
A solution of 1a (1.5 g, 1.34 mmol) in pyridine (20 mL) was trea-
ted with 2,4,6-triisopropylbenzenesulfonyl chloride (12.22 g,
40.35 mmol, 30.0 equiv). The reaction mixture was stirred at 23 °C
for 12 h. Reaction progress was monitored by TLC (3% methanol in

Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
3629
dichloromethane). Upon completion, the reaction mixture was con-
centrated, diluted with ethyl acetate (100 mL), washed with brine,
dried over MgSO4, and concentrated under reduced pressure. The
concentrated crude was purified by flash column chromatography
(2.5% methanol in dichloromethane) to afford the desired product
1b as a white solid (1.9 g, 86%): 1H NMR (500 MHz, methanol-d4) d
7.32 (s, 2H, SO3C6H2(CH)3(CH3)6), 7.31 (s, 2H, SO3C6H2(CH)3(CH3)6),
5.47 (s, 1H, H-10), 5.19 (s, 1H, H-100), 4.85 (m, 1H, H-1000), 4.40 (m, 1H,
H-60), 4.38–4.30 (m, 2H, H-200, H-60), 4.25-4.14 (m, 4H, H-300, H-500, H-
400), 4.01 (m, 1H, H-50), 3.91 (t, J = 3.1 Hz, 1H, H-4000), 3.81 (t, J = 7.2 Hz,
1H, H-5000), 3.77 (m, 1H, H-3000), 3.62 (dd, J = 10.3, 4.0 Hz, 1H, H-40),
3.56–3.26 (m, 12H, H-1, H-3, H-4, H-5, H-6, H-20, H-30, H-2000,
SO3C6H2(CH)3(CH3)6), 3.20 (m, 1H, H-6000), 2.99 (m, 4H,
SO3C6H2(CH)3(CH3)6, H-6000), 1.97 (m, 1H, H-2eq), 1.47–1.43 (m,
2H, H-500), 2.73 (m, 1H, H-60), 2.70-2.62 (m, 4H,
2ꢁSCH2CH2(CH2)4CH3), 2.02 (dd, J = 13.5, 4.8 Hz, 1H, H-2eq),
1.66-1.59 (m, 4H, 2ꢁSCH2CH2(CH2)4CH3), 1.57–1.15 (m, 62H,
5ꢁCO2C(CH3)3, 2ꢁSCH2CH2(CH2)4CH3, H-2ax), 0.94- 0.91(m, 6H,
2ꢁSCH2CH2(CH2)5CH3). 13C NMR (125 MHz, methanol-d4) d 157.4,
157.1, 156.7, 156.3, 109.9, 98.8, 97.5, 86.0, 81.2, 79.6, 79.4, 79.2,
79.1, 78.8, 74.5, 73.9, 73.0, 72.9, 72.3, 70.2, 67.4, 55.3, 52.1, 50.9,
50.1, 48.4, 40.4, 34.2, 33.6, 33.3, 32.3, 31.6, 31.6, 29.6, 29.5, 28.8,
28.7, 28.6, 28.5, 27.6, 27.4, 27.4, 22.3, 22.3, 13.1. LR-ESI m/z calcd
for C62H114N5O22S2 1344.73, found 1344.67 [M+H]+.
4.2.5. Compound 4a
To a solution of compound 1b (250 mg, 0.15 mmol) and Cs2CO3
(198 mg, 0.6 mmol) in dry DMF (4 mL) was added 1-octanethiol
(0.36 mL, 2.12 mmol), and the mixture was stirred at 23 °C over-
night. Reaction progress was monitored by TLC (5% methanol in
dichloromethane). Upon completion, the reaction mixture was
concentrated, diluted with ethyl acetate (100 mL), washed with
brine, dried over MgSO4, and concentrated under reduced pressure.
The concentrated crude was purified by flash column chromatogra-
phy over silica gel (4% methanol/dichloromethane) to afford com-
pound 4a (162 mg, 78%) as a white solid: 1H NMR (500 MHz,
methanol-d4) d 5.48 (bs, 1H, H-10), 5.15 (d, J = 2.8 Hz, 1H, H-100),
4.94 (bs, 1H, H-1000), 4.46–4.22 (m, 2H, H-200, H-400), 4.11 (q,
J = 5.6 Hz, 1H, H-300), 3.93 (m, 3H, H-50, H-3000, H-5000), 3.79 (m, 1H,
H-4000), 3.70–3.57 (m, 6H, H-4, H-5, H-6, H-20, H-30, H-2000), 3.39–
3.25 (m, 5H, H-1, H-3, H-40, H-6000), 3.01 (dd, J = 13.9, 2.7 Hz, 1H,
H-60), 2.87 (m, 2H, H-500), 2.72 (dd, J = 13.8, 7.3 Hz, 1H, H-60),
2.70–2.61 (m, 4H, 2ꢁSCH2CH2(CH2)5CH3), 2.02 (m, 1H, H-2eq),
1.62 (m, 4H, 2ꢁSCH2CH2(CH2)5CH3), 1.53–1.41 (m, 66H,
5ꢁCO2C(CH3)3, 2ꢁSCH2CH2(CH2)5CH3, H-2ax), 0.94–0.91 (m, 6H,
2ꢁSCH2CH2(CH2)5CH3). 13C NMR (125 MHz, methanol-d4) d 157.4,
157.0, 156.8, 156.7, 156.3, 109.9, 98.9, 97.5, 86.0, 81.2, 79.7, 79.4,
79.2, 79.1, 78.8, 74.5, 73.9, 73.0, 72.9, 72.3, 70.1, 67.4, 55.2, 52.1,
50.9, 50.1, 48.4, 40.3, 34.4, 34.2, 33.6, 33.3, 32.3, 31.7, 31.6, 29.6,
29.5, 29.1, 29.1, 29.0, 28.7, 28.6, 27.6, 27.4, 27.3, 22.3, 13.11. LR-
ESI m/z calcd for C64H118N5O22S2 1372.76, found 1372.54 [M+H]+.
46H,
5ꢁCO2C(CH3)3,
H-2ax),
1.33–1.23
(m,
36H,
SO3C6H2(CH)3(CH3)6). 13C NMR (100 MHz, methanol-d4) d 157.5,
157.1, 156.8, 156.7, 156.3, 154.0, 150.8, 149.4, 148.0, 129.3, 129.1,
123.6, 122.0, 109.2, 98.8, 97.3, 85.0, 79.5, 79.2, 79.1, 78.8, 77.9,
77.1, 74.0, 73.2, 73.0, 72.5, 70.0, 69.9, 68.5, 67.1, 54.6, 52.0, 50.9,
49.5, 39.8, 34.1, 34.0, 33.9, 29.4, 29.4, 29.3, 29.0, 27.5, 27.4, 27.3,
24.0, 23.8, 22.8, 22.5. LR-ESI m/z calcd for C78H128N5O28S2 1646.83,
found 1646.01 [MꢂH]ꢂ.
4.2.3. Compound 2a
To a solution of compound 1b (250 mg, 0.15 mmol) and Cs2CO3
(198 mg, 0.6 mmol) in dry DMF (4 mL) was added 1-hexanethiol
(0.30 mL, 2.12 mmol), and the mixture was stirred at 23 °C over-
night. Reaction progress was monitored by TLC (5% methanol in
dichloromethane). Upon completion, the reaction mixture was
concentrated, diluted with ethyl acetate (100 mL), washed with
brine, dried over MgSO4, and concentrated under reduced pressure.
The concentrated crude was purified by flash column chromatogra-
phy (4% methanol/dichloromethane) to afford compound 2a
(159 mg, 80%) as a white solid: 1H NMR (400 MHz, methanol-d4)
d 5.44 (bs, 1H, H-10), 5.13 (d, J = 2.7 Hz, 1H, H-100), 4.92 (d,
J = 1.8 Hz, H-1000), 4.25 (m, 2H, H-200, H-400), 4.09 (m, 1H, H-300),
3.91 (m, 3H, H-50, H-3000, H-5000), 3.76 (m, 1H, H-4000), 3.66–3.42 (m,
6H, H-4, H-5, H-6, H-20, H-30, H-2000), 3.33-3.30 (m, 5H, H-1, H-3,
H-40, H-6000), 2.99 (dd, 1H, J = 13.8, 2.7 Hz, H-60), 2.85 (m, 2H, H-
500), 2.71(m, 1H, H-60), 2.69–2.57 (m, 4H, 2ꢁSCH2CH2(CH2)3CH3),
4.2.6. Compound 2
Compound 2a (20 mg, 0.014 mmol) was treated with 95% TFA
(0.7 mL) at ambient temperature for 3 min. The TFA was removed
under reduced pressure, and the residue was dissolved in a minimal
volume of H2O and freeze-dried to afford compound 2 (21 mg,
quantitative yield) as a white foam: 1H NMR (400 MHz, metha-
nol-d4) d 5.73 (d, J = 3.9 Hz, 1H, H-10), 5.36 (d, J = 3.7 Hz, 1H, H-100),
5.34 (d, J = 1.7 Hz, 1H, H-1000), 4.45 (t, J = 4.7 Hz, 1H, H-300), 4.40 (m,
1H, H-200), 4.33 (dt, J = 8.4, 4.2 Hz, 1H, H-400), 4.29 (ddd, J = 7.2, 3.9,
1.5 Hz, 1H, H-5000), 4.15 (t, J = 3.2 Hz, 1H, H-3000), 3.99 (t, J = 9.6 Hz,
1H, H-4), 3.96–3.85 (m, 2H, H-30, H-50), 3.81 (t, J = 9.0 Hz, 1H, H-
5), 3.69 (m, 1H, H-4000), 3.63 (m, 1H, H-6), 3.57–3.33 (m, 5H, H-3,
H-20, H-40, H-2000, H-6000), 3.29-3.22 (m, 2H, H-1, H-6000), 3.10-3.04
(m, 2H, H-500, H-60), 2.83 (dd, J = 13.0, 8.0 Hz, 1H, H-500), 2.75 (dd,
J = 14.3, 6.7 Hz, 1H, H-60), 2.66–2.56 (m, 4H, 2ꢁSCH2CH2(CH2)3CH3),
2.48 (dt, J = 12.4, 4.2 Hz, 1H, H-2eq), 1.92 (q, J = 12.6 Hz, 1H, H-2ax),
1.67–1.54 (m, 4H, 2ꢁSCH2CH2(CH2)3CH3),1.50–1.24 (m, 12H,
2ꢁSCH2CH2(CH2)3CH3), 0.90 (m, 6H, 2ꢁSCH2CH2(CH2)3CH3). 13C
NMR (100 MHz, methanol-d4) d 161.7 (q, J = 34.6 Hz, CF3CO2H),
116.7 (q, J = 292.8 Hz, CF3CO2H), 110.7, 96.2, 95.5, 85.2, 80.7, 78.4,
78.0, 74.3, 73.9, 72.4, 71.82, 70.6, 68.8, 67.7, 54.3, 51.5, 49.6, 49.2,
40.2, 35.1, 32.8, 32.6, 32.2, 31.2, 31.1, 29.4, 29.3, 28.3, 28.1, 22.2,
22.2, 13.0, 12.9. HRESI-MS m/z calcd for C35H70N5O12S2 816.4462,
found 816.4460 [M+H]+.
2.05–1.94
2ꢁSCH2CH2(CH2)3CH3), 1.53–1.37 (m, 58H, 5ꢁCO2C(CH3)3,
H-2eq), 0.99–0.82 (m, 6H,
(m,
1H,
H-2eq),
1.67–1.53
(m,
4H,
2ꢁSCH2CH2(CH2)3CH3,
2ꢁSCH2CH2(CH2)3CH3). 13C NMR (125 MHz, methanol-d4) d 157.4,
157.0, 156.8, 156.7, 156.3, 109.9, 98.9, 97.5, 85.9, 81.2, 79.7, 79.4,
79.2, 79.1, 78.8, 74.5, 73.9, 73.0, 72.8, 72.3, 72.2, 70.1, 67.4, 55.2,
52.1, 40.3, 34.2, 33.5, 33.3, 32.3, 31.3, 31.2, 29.6, 29.4, 28.3, 28.2,
27.6, 22.3, 22.3, 13.1, 13.0. LR-ESI m/z calcd for C60H110N5O22S2
1316.70, found 1316.58 [M+H]+.
4.2.4. Compound 3a
To a solution of compound 1b (250 mg, 0.15 mmol) and Cs2CO3
(198 mg, 0.6 mmol) in dry DMF (4 mL) was added 1-heptanethiol
(0.33 mL, 2.12 mmol), and the reaction was stirred at 23 °C over-
night. Reaction progress was monitored by TLC (5% methanol in
dichloromethane). Upon completion, the reaction mixture was
concentrated, diluted with ethyl acetate (100 mL), washed with
brine, dried over MgSO4, and concentrated under reduced pressure.
The concentrated crude was purified by flash column chromatogra-
phy (4% methanol/dichloromethane) to afford compound 3a
(155 mg, 76%) as a white solid: 1H NMR (500 MHz, methanol-d4)
d 5.47 (bs, 1H, H-10), 5.15 (d, J = 3.6 Hz, 1H, H-100), 4.93 (bs, 1H,
H-1000), 4.39–4.22 (m, 2H, H-400, H-200), 4.10 (q, J = 5.2 Hz, 1H, H-
300), 4.02–3.87 (m, 3H, H-50, H-3000, H-5000), 3.79 (bs, 1H, H-4000),
3.66–3.42 (m, 6H, H-4, H-5, H-6, H-20, H-30, H-2000), 3.38–3.27 (m,
5H, H-1, H-3, H-40, H-6000), 3.09–2.96 (m, 1H, H-60), 2.98–2.84 (m,
4.2.7. Compound 3
Compound 3a (15 mg, 0.011 mmol) was treated with 95% TFA
(0.7 mL) at ambient temperature for 3 min. The TFA was removed

3630
Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
under reduced pressure, and the residue was dissolved in a mini-
mal volume of H2O and freeze-dried to afford compound 3
(15.6 mg, quantitative yield) as a white foam: 1H NMR (400 MHz,
methanol-d4) d 5.70 (d, J = 3.9 Hz, 1H, H-10), 5.35 (d, J = 3.8 Hz,
1H, H-100), 5.34 (d, J = 2.0 Hz, 1H, H-1000), 4.45 (t, J = 4.6 Hz, 1H, H-
300), 4.41 (t, J = 4.2 Hz, 1H, H-200), 4.33 (m, 1H, H-400), 4.29 (ddd,
J = 7.1, 3.8, 1.5 Hz, 1H, H-5000), 4.14 (t, J = 3.2 Hz, 1H, H-3000), 3.98–
3.84 (m, 3H, H-4, H-30, H-50), 3.80 (t, J = 9.0 Hz, 1H, H-5), 3.70 (m,
1H, H-4000), 3.61 (t, J = 9.7 Hz, 1H, H-6), 3.50–3.34 (m, 5H, H-3, H-
20, H-40, H-200, H-6000), 3.29–3.19 (m, 2H, H-6000, H-1), 3.07 (m, 2H,
H-500, H-60), 2.83 (dd, J = 13.0, 8.1 Hz, 1H, H-500), 2.75 (dd, J = 14.2,
6.9 Hz, 1H, H-60), 2.69–2.55 (m, 4H, 2ꢁSCH2CH2(CH2)4CH3), 2.44
(dt, J = 12.5, 4.2 Hz, 1H, H-2eq), 1.87 (q, J = 12.6 Hz, 1H, H-2ax),
1.62 (m, 4H, 2ꢁSCH2CH2(CH2)4CH3), 1.50–1.19 (m, 16H,
2ꢁSCH2CH2(CH2)4CH3), 0.98–0.78 (m, 6H, 2ꢁSCH2CH2(CH2)3CH3).
13C NMR (125 MHz, methanol-d4) d 161.7 (q, J = 34.6 Hz, CF3CO2H),
116.7 (q, J = 292.7 Hz, CF3CO2H), 110.7, 96.2, 95.5, 85.4, 80.7, 78.4,
74.3, 73.9, 72.5, 71.8, 70.7, 68.8, 67.8, 67.7, 54.4, 51.5, 49.7, 49.2,
40.2, 35.1, 32.8, 32.7, 32.2, 31.58, 31.5, 29.5, 29.4, 28.7, 28.6,
28.4, 22.3, 22.2, 13.0. HRESI-MS m/z calcd for C37H74N5O12S2
844.4775, found 844.4770 [M+H]+.
SO3C6H2(CH)3(CH3)6). 13C NMR (100 MHz, methanol-d4) d 159.7,
159.4, 159.0, 159.3, 158.7, 156.2, 153.0, 131.6, 125.8, 111.9, 100.9,
100.0, 88.9, 84.1, 81.6, 81.5, 81.3, 81.0, 79.7, 79.3, 76.8, 76.4, 75.22,
73.9, 72.4, 72.3, 71.9, 70.3, 69.6, 65.1, 57.4, 54.4, 53.3, 52.21, 42.5,
36.4, 31.7, 29.8, 29.7, 29.6, 26.0, 24.8. LR-ESI m/z calcd for
C
63H106N5O26S 1380.69, found 1380.58 [MꢂH]ꢂ.
4.2.10. Compound 5b
To a solution of compound 5a (80 mg, 0.057 mmol) and Cs2CO3
(mg, 0.115 mmol) in dry DMF (1 mL) was added 1-hexadecanethiol
(0.12 mL, 0.4 mmol), and the mixture was stirred at 23 °C over-
night. Reaction progress was monitored by TLC (5% methanol in
dichloromethane). Upon completion, the reaction mixture was
concentrated, diluted with ethyl acetate (100 mL), washed with
brine, dried over MgSO4, and concentrated under reduced pressure.
The concentrated crude was purified by flash column chromatogra-
phy over silica gel (4% methanol/dichloromethane) to afford com-
pound 5b (60 mg, 77%) as a white solid: 1H NMR (400 MHz,
methanol-d4) d 5.32 (bs, 1H, H-10), 5.14 (d, J = 2.6 Hz, 1H, H-100),
4.89 (s, 1H, H-1000), 4.22 (m, 1H, H-200), 4.17 (m,1H, H-300), 3.97–
3.79 (m, 5H, H-50, H-400, H-500, H-3000, H-5000), 3.76 (m, 1H, H-4000),
3.68 (dd, J = 12.2, 6.4 Hz, 1H, H-500), 3.63–3.19 (m, 11H, H-1, H-3,
H-4, H-5, H-6, H-20, H-30, H-40, H-2000, H-6000), 2.99 (dd, J = 13.8,
2.7 Hz, 1H, H-60), 2.72 (dd, J = 13.8, 7.0 Hz, 1H, H-60), 2.64 (t,
J = 7.2 Hz, 2H, SCH2CH2(CH2)12CH3), 1.98 (dt, J = 13.0, 4.0 Hz, 1H,
H-2eq), 1.59 (m, 2H, SCH2CH2(CH2)12CH3), 1.53–1.39 (m, 46H,
5ꢁCO2C(CH3)3, H-2ax), 1.30- 1.29 (m, 24H, SCH2CH2(CH2)12CH3),
0.97–0.70 (m, 3H, SCH2CH2(CH2)12CH3). 13C NMR (100 MHz, meth-
anol-d4) d 157.5, 157.1, 156.8, 156.3, 109.6, 98.9, 97.8, 86.4, 82.0,
79.3, 79.2, 79.0, 76.9, 74.6, 74.1, 73.0, 72.6, 72.3, 71.6, 70.2, 67.5,
62.4, 55.5, 52.1, 40.4, 34.2, 33.7, 33.2, 31.6, 29.6, 29.3, 29.0, 29.0,
4.2.8. Compound 4
Compound 4a (30 mg, 0.021 mmol) was treated with 95% TFA
(0.7 mL) at ambient temperature for 3 min. The TFA was removed
under reduced pressure, and the residue was dissolved in a minimal
volume of H2O and freeze-dried to afford compound 4 (31.5 mg,
quantitative yield) as a white foam: 1H NMR (400 MHz, methanol-
d4) d 5.73 (d, J = 3.9 Hz, 1H, H-10), 5.36 (d, J = 3.7 Hz, 1H, H-100), 5.34
(d, J = 1.7 Hz, 1H, H-1000), 4.45 (t, J = 4.7 Hz, 1H, H-300), 4.40 (m, 1H,
H-200), 4.33 (dt, J = 8.4, 4.2 Hz, 1H, H-400), 4.29 (ddd, J = 7.2, 3.9,
1.5 Hz, 1H, H-5000), 4.15 (t, J = 3.2 Hz, 1H, H-3000), 3.99 (t, J = 9.5 Hz,
1H, H-4), 3.88 (m, 2H, H-30, H-50), 3.81 (t, J = 9.0 Hz, 1H, H-5), 3.69
(m, 1H, H-4000), 3.63 (dd, J = 10.4, 9.0 Hz, 1H, H-6), 3.56–3.33 (m,
5H, H-3, H-20, H-40, H-200, H-6000), 3.30–3.19 (m, 2H, H-1, H-6000), 3.07
(m, 2H, H-60, H-500), 2.83 (dd, J = 13.1, 8.0 Hz, 1H, H-500), 2.76 (dd,
J = 14.3, 6.7 Hz, 1H, H-60), 2.67–2.56 (m, 4H, 2ꢁSCH2CH2(CH2)5CH3),
2.47 (dt, J = 12.5, 4.3 Hz, 1H, H-2eq), 1.92 (q, J = 12.5 Hz, 1H, H-2ax),
1.72–1.51 (m, 4H, 2ꢁSCH2CH2(CH2)5CH3), 1.48–1.17 (m, 20H,
2ꢁSCH2CH2(CH2)5CH3), 0.97–0.67 (m, 6H, 2ꢁSCH2CH2(CH2)5CH3).
13C NMR (100 MHz, methanol-d4) d 161.7 (q, J = 34.7 Hz, CF3CO2H),
116.7 (q, J = 292.7 Hz, CF3CO2H), 110.7, 96.1, 95.5, 85.3, 80.7, 78.4,
78.0, 74.4, 73.9, 72.4, 71.7, 70.6, 68.8, 67.8, 54.3, 51.5, 49.6, 49.2,
40.2, 35.7, 32.8, 32.6, 32.2, 31.6, 31.5, 29.5, 29.4, 28.7, 28.4, 28.1,
22.3, 13.0. HRESI-MS m/z calcd for C39H78N5O12S2 872.5088, found
872.5087 [M+H]+.
28.5, 27.5, 27.4, 27.3, 22.3, 13.0. LR-ESI m/z calcd for C64H118N5O23
S
1356.79, found 1356.34 [M+H]+.
4.2.11. Compound 5
Compound 5b (20 mg, 0.014 mmol) was treated with 95% TFA
(0.7 mL) at ambient temperature for 3 min. The TFA was removed
under reduced pressure, and the residue was dissolved in a mini-
mal volume of H2O and freeze-dried to afford compound 5
(21 mg, quantitative yield) as a white foam: 1H NMR (400 MHz,
methanol-d4) d 5.57 (d, J = 3.8 Hz, 1H, H-10), 5.35 (d, J = 3.1 Hz,
1H, H-100), 5.30 (d, J = 1.7 Hz, 1H, H-1000), 4.52 (t, J = 5.5 Hz, 1H, H-
300), 4.36–4.24 (m, 2H, H-5000, H-200), 4.17 (m, 1H, H-400), 4.14 (m,
1H, H-3000), 3.95 (t, J = 9.8 Hz, 1H, H-4), 3.92–3.73 (m, 5H, H-5, H-
30, H-50, H-500), 3.68 (m, 1H, H-4000), 3.61 (m, 1H, H-6), 3.54 (td,
J = 12.6, 10.2, 4.3 Hz, 1H, H-3), 3.44 (m, 1H, H-2000), 3.42–3.21 (m,
5H, H-1, H-20, H-40, H-6000), 3.09 (dd, J = 14.1, 2.2 Hz, 1H, H-60),
2.68 (dd, J = 14.3, 8.0 Hz, 1H, H-60), 2.59 (m, 2H,
SCH2CH2(CH2)12CH3), 2.48 (dt, J = 12.6, 4.5 Hz, 1H, H-2eq), 1.90
(q, J = 12.6 Hz, 1H, H-2ax), 1.71–1.52 (m, 2H, SCH2CH2(CH2)12CH3),
1.41-1.25 (m, 24H, SCH2CH2(CH2)12CH3), 0.90-0.85 (m, 3H,
SCH2CH2(CH2)12CH3). 13C NMR (100 MHz, methanol-d4) d 161.7
(q, J = 34.8 Hz, CF3CO2H), 116.7 (q, J = 292.4, CF3CO2H), 110.3,
97.3, 95.6, 84.2, 82.0, 79.6, 75.5, 74.6, 73.9, 72.5, 72.2, 70.6, 68.9,
67.8, 67.7, 59.6, 54.2, 51.5, 49.7, 49.4, 40.2, 32.7, 32.6, 31.6, 29.4,
4.2.9. Compound 5a
A solution of 1a (0.5 g, 0.44 mmol) in pyridine (10 mL) was treated
with 2,4,6-triisopropylbenzenesulfonyl chloride (2.71 g, 8.96 mmol,
20.0 equiv). The reaction mixture was stirred at 23 °C for 12 h. Reac-
tion progress was monitored by TLC (3% methanol in dichlorometh-
ane). Upon completion, the reaction mixture was concentrated,
diluted with ethyl acetate (100 mL), washed with brine, dried over
MgSO4, and concentrated under reduced pressure. The concentrated
crude was purified by flash column chromatography over silica gel
(3%methanolindichloromethane) to afford1b(0.31, 43%)andthe de-
sired product 5a as a white solid (0.2 g, 33%): 1H NMR (400 MHz,
methanol-d4) d 7.29 (s, 2H, SO3C6H2(CH)3(CH3)6), 5.43 (bs, 1H, H-10),
5.12 (d, 1H, J = 2.2 Hz, H-100), 4.89 (bs, 1H, H-1000), 4.35 (m, 3H, H-60,
H-50), 4.18 (m, 2H, H-200, H-300), 4.01–3.86 (m, 4H, H-400, H-40, H-3000,
H-5000), 3.83–3.73 (m, 2H, H-2000, H-500), 3.64 (dd, J = 12.2, 6.7 Hz, 1H,
H-500), 3.61–3.21 (m, 12H, H-1, H-3, H-4, H-5, H-6, H-20, H-30, H-4000,
29.3, 29.0, 28.9, 22.3, 13.01. HRESI-MS m/z calcd for C39H78N5O13
S
856.5317, found 856.5318 [M+H]+.
4.3. Biological assays
4.3.1. Minimal inhibitory concentration protocol
Starter cultures were incubated for 24 h (37 °C, 5% CO2, aerobic
conditions) and then diluted in fresh medium to obtain an optical
density of 0.004 (OD600). All strains were tested using the double-
H-6000,
SO3C6H2(CH)3(CH3)6), 1.93 (dt, J = 12.1, 4.3 Hz, 1H, H-2eq), 1.51-1.38
(m, H-2ax), 1.32-1.22 (m, 18H,
46H, 5ꢁCO2C(CH3)3,
SO3C6H2(CH)3(CH3)6),
2.95
(hep,
J = 6.9 Hz,
1H,
dilution method starting at 64
lg/mL in 96-well plates (Sarstedt).
After 24 h of incubation, MTT (50
l
L of a 1 mg/mL solution in

Y. Berkov-Zrihen et al. / Bioorg. Med. Chem. 21 (2013) 3624–3631
3631
H2O) was added to each well followed by additional incubation at
37 °C for 2 h. MIC values ( g/mL) were determined as the lowest
Supplementary data
l
concentration at which no bacterial growth was observed. Results
were obtained from two independent experiments, and each
experiment was done in triplicate.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2013.03.046.
References and notes
4.3.2. Red blood cell hemolysis protocol
A sample of rat RBCs (2% w/w) were incubated with each of the
tested compounds for 1 h at 37 °C, 5% CO2 using the double dilu-
1. Landis, S. J. Adv. Skin Wound Care 2008, 21, 531.
2. Lipsky, B. A.; Berendt, A. R.; Deery, H. G.; Embil, J. M.; Joseph, W. S.; Karchmer,
A. W., et al Clin. Infect. Dis. 2004, 39, 885.
3. Abdulrazak, A.; Bitar, Z. I.; Al-Shamali, A. A.; Mobasher, L. A. J. Diabet.
Complications 2005, 19, 138.
4. Schwartz, R. A.; Al-mutairi, N. Gulf J. Dermatol. Venereol. 2010, 17, 1.
5. Lipsky, B. a.; Hoey, C. Clin. Infect. Dis. 2009, 49, 1541.
6. Gelmetti, C. Dermatol. Ther. 2008, 21, 187.
7. Hancock, R. E. W.; Chapple, D. S. Antimicrob. Agents Chemother. 1999, 43, 1317.
8. Hartmann, M.; Berditsch, M.; Hawecker, J.; Ardakani, M. F.; Gerthsen, D.; Ulrich,
A. S. Antimicrob. Agents Chemother. 2010, 54, 3132.
9. Mogi, T.; Kita, K. Cell. Mol. Life Sci. 2009, 66, 3821.
10. Zavascki, A. P.; Goldani, L. Z.; Li, J.; Nation, R. L. Antimicrob. Agents Chemother.
2007, 60, 1206.
tion method starting at concentration of 256 lg/mL. Negative con-
trol was PBS and positive control was 1% w/v solution of Triton
X100 (100% hemolysis). Following centrifugation (2000 rpm,
10 min, ambient temperature), the supernatant was removed and
its absorbance measured at 550 nm using a microplate reader
(Genios, TECAN). The results are expressed as percentage of hemo-
globin released relative to the positive control (Triton X100).
Experiments were performed in triplicate, and the results are an
average of experiments in blood samples taken from at least two
rats.
11. Glupczynski, Y. Antimicrob. Agents Chemother. 1999, 43, 727.
12. Houghton, J. L.; Green, K. D.; Chen, W.; Garneau-Tsodikova, S. ChemBioChem
2010, 11, 880.
13. Gehrig, K. a.; Warshaw, E. M. J. Am. Acad. Dermatol. 2008, 58, 1.
14. Huth, M. E.; Ricci, A. J.; Cheng, A. G. Int. J. Otolaryngol. 2011, 1.
15. Mingeot-Leclercq, M. P.; Tulkens, P. M. Antimicrob. Agents Chemother. 1999, 43,
1003.
16. Herzog, I. M.; Green, K. D.; Berkov-Zrihen, Y.; Feldman, M.; Vidavski, R. R.;
Eldar-Boock, A.; Satchi-Fainaro, R.; Eldar, A.; Garneau-Tsodikova, S.; Fridman,
M. Angew. Chem., Int. Ed. 2012, 51, 5652.
17. Herzog, I. M.; Feldman, M.; Eldar-Boock, A.; Satchi-Fainaro, R.; Fridman, M.
MedChemCommun 2013, 4, 120.
18. Bera, S.; Zhanel, G. G.; Schweizer, F. Antimicrob. Agents Chemother. 2010, 65,
1224.
19. Bera, S.; Zhanel, G. G.; Schweizer, F. J. Med. Chem. 2010, 53, 3626.
20. Ouberai, M.; El Garch, F.; Bussiere, A.; Riou, M.; Alsteens, D.; Lins, L.; Baussanne,
I.; Dufrêne, Y. F.; Brasseur, R.; Decout, J.-L.; Mingeot-Leclercq, M.-P. Biochim.
Biophys. Acta 1808, 2011, 1716.
21. Baussanne, I.; Bussière, A.; Halder, S.; Ganem-Elbaz, C.; Ouberai, M.; Riou, M.;
Paris, J.-M.; Ennifar, E.; Mingeot-Leclercq, M.-P.; Décout, J.-L. J. Med. Chem.
2010, 53, 119.
22. Howell-Jones, R. S.; Wilson, M. J.; Hill, K. E.; Howard, A. J.; Price, P. E.; Thomas,
D. W. Antimicrob. Agents Chemother. 2005, 55, 143.
23. Weghuber, J.; Aichinger, M. C.; Brameshuber, M.; Wieser, S.; Ruprecht, V.;
Plochberger, B.; Madl, J.; Horner, A.; Reipert, S.; Lohner, K.; Henics, T.; Schütz, G.
J. Biochim. Biophys. Acta 1808, 2011, 2581.
24. Epand, R. F.; Savage, P. B.; Epand, R. M. Biochim. Biophys. Acta 2007, 1768, 2500.
25. Michael, K.; Wang, H.; Tor, Y. Bioorg. Med. Chem. 1999, 7, 1361.
26. Pathak, R.; Böttger, E. C.; Vasella, A. Helv. Chim. Acta 2005, 88, 2967.
27. Bearden, D. T.; Allen, G. P.; Christensen, J. M. Antimicrob. Agents Chemother.
2008, 62, 769.
28. Löffler, B.; Hussain, M.; Grundmeier, M.; Brück, M.; Holzinger, D.; Varga, G.;
Roth, J.; Kahl, B. C.; Proctor, R. a.; Peters, G. PLoS Pathog. 2010, 6, 1.
29. Otto, M. Nat. Rev. Microbiol. 2009, 7, 555.
30. Johansson, L.; Thulin, P.; Low, D. E.; Norrby-Teglund, A. Clin. Infect. Dis. 2010, 51,
58.
31. Koo, S.; Bayer, A. S.; Yeaman, M. R. Infect. Immun. 2001, 69, 4916.
32. Wallin, E.; Von Heijne, G. Protein Sci. 1998, 7, 1029.
4.3.3. Scanning electron microscopy protocol
Cultures of S. epidermidis ATCC12228 were incubated for 4 h at
37 °C under aerobic conditions to obtain an optical density of 0.2
(OD600) and then harvested by centrifugation at 4000 rpm for
10 min at 4 °C. The bacterial pellet was washed twice with PBS
(pH 7.4) and resuspended in PBS (pH 7.4) to obtain an optical den-
sity of 1.0 (OD600). The bacterial suspension was diluted twofold
after treatment with the tested compound at a concentration of
1 lg/mL at 37 °C for 1 h. The cells were then spun down at
6000 rpm for 4 min at 4 °C, washed with PBS (pH 7.4) three times,
and fixed with 2.5% glutaraldehyde/PBS buffer overnight at 4 °C.
The cells were then washed three times in 0.1 M PB (pH 7.4), dehy-
drated in series of graded ethanol solutions (30–100%), and dried in
vacuum desiccation. Finally, the samples were coated with palla-
dium-gold and viewed via scanning electron microscopy (Quanta
200FEG ESEM).
Acknowledgments
This work was supported by the FP7-PEOPLE-2009-RG Marie
Curie Action: Reintegration Grants (Grant 246673). We thank Pro-
fessors Itzhak Ofek, Dani Cohen (Tel Aviv University), and Doron
Steinberg (The Hebrew University of Jerusalem) for the gift of bac-
terial strains. We thank Anat Eldar-Boock from the group of Profes-
sor Ronit Satchi-Fainaro (Tel Aviv University) for her help with the
hemolysis assays.