Structure-activity relationships of bivalent aminoglycosides and evaluation of their microbiological activities

Article abstract of DOI:10.1016/j.bmcl.2005.02.029

A library of 4,5- and 4,6-linked bivalent aminoglycoside (AMG) antibiotics consisting of neamine and nebramine pharmacophores have been synthesized. We probed the effect of the linker on antibiotic activity with a series of selected synthetic analogues wi

Full text of DOI:10.1016/j.bmcl.2005.02.029

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2123–2128
Structure–activity relationships of bivalent aminoglycosides
and evaluation of their microbiological activities
Chang-Hsing Liang,a Alex Romero,a David Rabuka,a Paulo W. M. Sgarbi,a
Kenneth A. Marby,a Jonathan Duffield,a Sulan Yao,a Mayling L. Cheng,a
Yoshi Ichikawa,a Pamela Sears,b Changyong Hu,c San-Bao Hwang,c
Youe-Kong Shuea and Steven J. Suchecka,*
aDepartment of Chemistry, Optimer Pharmaceuticals, Inc., 10110 Sorrento Valley Rd., Suite C, San Diego, CA 92121, USA
bDepartment of Biology, Optimer Pharmaceuticals, Inc., 10110 Sorrento Valley Rd., Suite C, San Diego, CA 92121, USA
cDepartment of Biology, Optimer PTE. Ltd, 41 Science Park Road, #01-18/20 The Gemini,
Singapore Science Park II, Singapore 117610
Received 1 December 2004; revised 7 February 2005; accepted 9 February 2005
Available online 16 March 2005
Abstract—A library of 4,5- and 4,6-linked bivalent aminoglycoside (AMG) antibiotics consisting of neamine and nebramine phar-
macophores have been synthesized. We probed the effect of the linker on antibiotic activity with a series of selected synthetic
analogues with varied length and substituents. A number of compounds demonstrated in vitro activity against several bacterial
strains and showed activity against drug resistant strains of Pseudomonas aeruginosa. Among the compounds prepared, analogues
12a–d were novel 4,6-linked AMGs containing the nebramine pharmacophore. In addition the lead compound OPT-11 possessed an
ED50 of 65 mg/kg in a Staphylococcus aureus ATCC 29213 mouse protection model.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
through a diamine linker and are believed to interact
with the target 16S rRNA in a bivalent fashion, as
shown in Figure 1A. Furthermore, these novel AMGs
were found to be nanomolar inhibitors of the APH(200)
activity of the AMG modifying-enzyme AAC(60)–
APH(200) making the linked AMGs ÔbifunctionalÕ as
well.
Nosocomial infections of P. aeruginosa are associated
with high mortality and morbidity; in addition, P. aeru-
ginosa is the primary pathogen associated with the ge-
netic disease cystic fibrosis (CF).1,2 Treatment options
for infections caused by P. aeruginosa are limited in
comparison to many Gram-positive infections due to
the organismÕs intrinsic resistance to antibiotics. Amino-
glycosides (AMGs) are particularly active against sev-
eral Gram-negative pathogens and are commonly used
to treat severe P. aeruginosa infections. However, resis-
tance to this important class of antibiotics as well as
other antipseudomonals is growing. Recently, a new
class of linked AMGs was reported to have good antimi-
crobial activity against AMG sensitive strains of S. aur-
eus and P. aeruginosa.3 The linked AMGs consistof the
2-deoxystreptamine (2-DOS) pharmacophore connected
OPT-11, shown in Figure 1B, is a bivalentAMG wiht
potent antibacterial activity.3 The neamine motif present
in OPT-11 has been shown to complex the 16S rRNA
model AS–wt with a unique 2:1 binding stoichiometry
in contrast to other naturally occurring AMGs, which
bind AS–wt in a 1:1 fashion. Tethering neamine with
an appropriate linker resulted in OPT-11, which re-
verted to 1:1 binding and showed a 250-fold improve-
mentin Kd over monomeric neamine and showed a
significant increase in antimicrobial activity compared
to neamine. Figure 2 illustrates a theoretical model of
OPT-11 bound to the AS–wt. The model was con-
structed using the reported structure of AS–wt bound
to paromomycin.4 In our model one molecule of ne-
amine is docked to the A1408–A1493 base pair in the bulge
region of the RNA while the tether lies in the major
Keywords: Aminoglycosides; RNA; Antibiotics; Ribosome; 2-Deoxy-
streptamine.
*
Corresponding author. Tel.: + 1 858 909 0736; fax: + 1 858 909
0737; e-mail: ssucheck@optimerpharma.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.02.029

2124
C.-H. Liang et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2123–2128
Figure 1. (A) Model of a bivalentAMG binding an AS–wtrRNA model sequence. The pharmacophore is neamine or nebramine (shaded). (B) The
structure of OPT-11, a bivalent AMG.
achieve high levels of activity against P. aeruginosa we
wanted to evaluate the effect of replacing the neamine
motif with the nebramine motif which is found in the
AMG tobramycin, the AMG with the highest potency
against P. aeruginosa.
2. Chemistry
The synthesis of 4,5-linked neamine, 4,5-linked nebr-
amine and 4,6-linked nebramine analogues is outlined
in Schemes 1 and 2, respectively. 5-free OH neamine
13 and 5-free OH nebramine 25 were obtained following
previously reported procedures. The alcohols 1 and 2
were reacted with (R)-glycidol tosylate to afford the
epoxides 3 and 4 which were subsequently linked with
Figure 2. A model of OPT-11 bound to the AS–wt rRNA, a model for
the 16S rRNA domain.
a variety of amines using 2 equiv of epoxide heated at
70 ꢁC in ethanol. After 24–48 h of heating, the resulting
linked AMGs were purified by silica gel and de-pro-
tected in a single step by hydrogenation over 20%
Pd(OH)2/C in 1:1 acetic acid–water. The resulting
AMGs were efficiently purified by mass-directed pre-
parative RP-HPLC. Alternatively, the compounds could
be purified by AmberliteTM CG50 cation exchange resin
using the ammonium form of the resin. Elution of the
resin with a linear gradient 0–3% NH4 OH was generally
sufficient to purify the compounds. The latter method was
more convenient than RP-HPLC when multigram quanti-
ties of linked AMG were needed for in vivo studies.
Table 1 shows compounds 7a–n, which contain a
representative selection of flexible and rigid linkers in
the 4,5-linked neamine series, while Table 2 shows 8a,b
the nebramine homologues of OPT-11. In order to
determine if the 4,5-linking pattern was critical,
groove of the RNA and the distal neamine motif
interacts near the loop domain.
Based on this model we aimed to determine whether
modification of the linker domain and the linkage posi-
tion could further enhance the antimicrobial activity of
bivalent AMGs because optimized binding or selectivity
for 16S rRNA would result in the improvement in anti-
microbial activity (vide infra).
To accomplish this aim we selected a series of substi-
tuted and unsubstituted linkers with lengths between 4
and 10 atoms that were either flexible or rigid. In addi-
tion we also wanted to assess linking the 2-DOS phar-
macophore through the 6-position because of the
abundance of naturally occurring AMGs linked through
that position. And finally, since our ultimate goal was to
protected nebramine
9 was prepared as recently

C.-H. Liang et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2123–2128
2125
Scheme 1. Synthesis of 4,5-linked neamine and nebramine congeners. Reagents and conditions: (a) (R)-glycidyl tosylate, NaH, DMF; (b) 0.5 equiv
diamine, EtOH, 70 ꢁC, 48 h; (c) 20% Pd(OH)2/C, H2, H2O, AcOH.
described.5 The linking chemistry was repeated to afford
the 4,6-linked compounds 12a–d, Scheme 2 and Table 3,
respectively.
The lead compound OPT-11 was further evaluated for
efficacy in a S. aureus mouse protection model (Table
6). A fresh inoculum of S. aureus ATCC 29213 was
adjusted to 100 · LD50 dose and 500 mL was injected
intraperitoneally to five female Balb/c mice. The test
compound was dissolved in a sterile normal saline
(0.85% NaCl) solution. The pH was between 7.0 and
7.5 and required no adjustment. The drug (60 mg) was
dissolved in 3 mL of sterile saline, the drug was rapidly
solubilized and this solution was then used as the stock
solution for all lower concentrations used in this study.
OPT-11 was administered over a five-dose range by
intravenous injection into the tail vein 1 h and 5 h post-
infection and observed for the following five days. An
3. Biology
As a screen for broad spectrum activity, minimum inhibi-
tory concentrations (MICs) in microgram per milliliter
values were determined after 16–20 h incubation6 for
compounds 7a–n, 8a,b and 12a–d in comparison with
tobramycin against reference strains S. aureus ATCC
29213; Enterococcus faecalis ATCC 29212; Escherichia
coli ATCC 25922; P. aeruginosa ATCC 27853, and
PAO-1 and S. aureus MRSA 33591, Table 4. As a sec-
ondary panel, 16 clinical isolates of P. aeruginosa
(PAE_NUH01-7, 17-25) were obtained which include
both sensitive and resistant strains (National University
Hospital, National University of Singapore, Depart-
mentof Laboraotry Medicine, Clinical Microbiology
Lab, Singapore). The disk diffusion method was used
to obtain the antibiotic susceptibility profiles against a
variety of empirically used antipseudomonal antibiotics
(Table 5).7 Also shown are the MICs for tobramycin,
OPT-11 and the des-methyl analogue 7h.
exactED
was not calculated because there were not
50
sufficient mortalities in the lower dosing groups, but
was less than 5 mg/kg. The results show protection of
1/2 of the mice at concentrations below 5 mg/kg, a result
that compares favourably with currently used AMGs.
4. Discussion
Derivatives 7a–n, 8a,b and 12a–d were synthesized utili-
zing two pharmacophore domains, neamine and

2126
C.-H. Liang et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2123–2128
Scheme 2. Synthesis of 4,6-linked nebramine congeners. Reagents and conditions: (a) (R)-glycidyl tosylate, NaH, DMF; (b) 0.5 equiv diamine,
EtOH, 70 ꢁC, 48 h; (c) 20% Pd(OH)2/C, H2, H2O, AcOH.
Table 1. 4,5-Linked neamine structures
No.
Linker
No.
Linker
H
N
7a
7h
N
N
H
N
N
H
N
H
7b
7c
7i
7j
N
N
N
N
N
H
N
N
H
N
H
N
N
N
7d
7k
H
N
H
N
N
H
N
H
7e
7l
N
H
N
N
7f (OPT-11)
7m
N
N
HN
NH
O
N
7g
7n
N
O

C.-H. Liang et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2123–2128
Table 2. 4,5-Linked nebramine structures
2127
No.
Linker
No.
Linker
H
N
N
8a
8b
N
N
H
Table 3. 4,6-Linked nebramine structures
No.
Linker
No.
Linker
H
N
N
H
N
H
12a
12c
N
H
N
12b
12d
HN
NH
N
Table 4. In vitro antimicrobial activities of (7a–n), (8a,b) and (12a–d)
No.
MIC (lg/mL)
Sa
Ef
>50
Ec
Pa
>50
12.5
Paa
Sab
7a
7b
>50
>50
>50
25
>50
50
>50
>50
>50
>50
50
12.5
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
32
7c
7d
12.5
12.5
3.1
1
25
12.5
6.3
8
25
50
50
12.5
7e
7f (OPT-11)
6.3
8
4
25
16
50
7g
7h
50
1
1.6
4
—
4
2
12.5
16
50
7i
1.6
6.3
12.5
12.5
12.5
25
6.3
3.1
25
7j
—
12.5
7k
7l
50
>50
50
>50
>50
>50
>50
32
>50
6.3
>50
>50
>50
12.5
>50
7m
7n
25
>50
16
>50
16
8a
8b
2
16
4
2
16
20.9
4
1.3
16
10.4
12a
12b
12c
12d
1.3
0.6
0.5
0.5
>20.9
20.9
32
2.6
5.2
4
10.4
2
2.6
2
20.9
16
32
4
2
2
16
Tobramycin range
0.12–1
8–32
0.25–1
0.25–1
0.25–1
>64
Sa, Staphylococcus aureus ATCC 29213; Ef, Enterococcus faecalis 29212; Ec, Escherichia coli ATCC 25922; Pa, Pseudomonas aeruginosa ATCC
27853.
a Pa, Pseudomonas aeruginosa PAO-1.
b Sa, Staphylococcus aureus MRSA 3359.
nebramine. A number of compounds demonstrated
MICs below 6 lg/mL against several bacterial strains.
In the 7a–n series the conformationally restrained linkers
7b–d and g all showed reduced activities as did linkers
shorter or longer than 4Cs, 7a,e and 7i–n, respectively
(Table 4). Thus, the C4 linker remained the optimal lin-
ker length. The des-methyl analogue 7h was slightly
more potent against P. aeruginosa whilstlinkers of any
length with substitution larger in size than a methyl
group lost activity (data not shown). These data are

2128
C.-H. Liang et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2123–2128
Table 5. Susceptibility testing of P. aeruginosa clinical isolates and in vitro antimicrobial activities of OPT-11 and 7h
Strain
Antibiotic
CIP
MIC (lg/mL)
AMK
GM
TOB
IMP
CEF
AZT
PTZ
CO-T
CHL
TET
TOB
OPT-11
7h
ATCC 27853
PAE_NUH01
PAE_NUH02
PAE_NUH03
PAE_NUH04
PAE_NUH05
PAE_NUH06
PAE_NUH07
PAE_NUH17
PAE_NUH18
PAE_NUH19
PAE_NUH20
PAE_NUH21
PAE_NUH22
PAE_NUH23
PAE_NUH24
PAE_NUH25
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
R
R
S
S
S
S
R
S
S
S
S
S
S
S
S
S
R
R
R
R
R
R
R
R
S
R
R
R
R
R
R
R
R
S
S
1
>200
50
8
12.5
6.3
6.3
6.3
6.3
12.5
6.3
>32
16
1
6.3
R
R
R
R
R
R
S
R
R
R
R
R
R
R
S
R
R
R
R
R
R
S
R
R
R
R
R
R
I
R
R
R
R
R
R
R
S
R
R
R
R
R
R
R
S
R
R
R
R
R
R
R
R
R
R
R
I
6.3
6.3
3.1
3.1
6.3
3.1
50
100
100
100
6.3
1
S
S
S
32
S
I
S
S
S
R
R
R
R
R
R
R
R
R
R
S
R
R
S
0.5
0.5
2
8
8
S
S
S
S
S
16
16
S
S
S
S
S
16
16
16
16
16
8
S
S
S
S
S
R
R
R
R
S
I
1
16
S
S
S
S
S
I
R
R
R
R
0.25
>32
8
>32
32
S
I
R
S
S
S
I
S
I
S
S
I
1
0.5
S
S
S
S
S
R
8
AMK = amikacin; GM = gentamicin, TOB = tobramycin; IMP = imipenem; CEF = ceftazidime; CIP = ciprofloxacin; AZT = aztreonam;
PTZ = piperacillin/tazobactam, CO-T = co-trimoxazole; CHL = chloramphenicol; TET = tetracycline.
Table 6. In vivo efficacy of OPT-11 in a S. aureus mouse protection
model
of Allergy and Infectious Diseases (1 R43 AI056617-01
to S.J.S.).
Dose
Survivors
Mortalities
% Survivors
Control
5 mg/kg
10 mg/kg
25 mg/kg
1 of 5
4 of 5
5 of 5
4 of 5
4 of 5
1 of 5
0 of 5
1 of 5
20
References and notes
66.6
90
1. Centers for Disease Control and Prevention National
Nosocomial Infections Surveillance (NNIS) Report, Octo-
ber 1986–April 1996 . Am. J. Infect. Control 1996, 24, 380–
388.
92.9
2. Lyczak, J. B.; Cannon, C. L.; Pier, G. B. Clin. Microbiol.
Rev. 2002, 15, 194.
3. Sucheck, S. J.; Wong, A. L.; Koeller, K. M.; Boehr, D. D.;
Draker, K.-A.; Sears, P.; Wright, G. D.; Wong, C.-H. J.
Am. Chem. Soc. 2000, 122, 5230.
4. Fourmy, D.; Recht, M. I.; Blanchard, S. C.; Puglisi, J. D.
Science 1996, 274, 1367.
5. Agnelli, F.; Sucheck, S. J.; Marby, K. A.; Rabuka, D.; Yao,
S.-L.; Sears, P. S.; Liang, F.-S.; Wong, C.-H. Angew.
Chem., Int. Ed. 2004, 1562–1566.
6. Individual compounds are tested for growth inhibition and
minimum inhibitory concentrations (MIC) values calcu-
lated according to the protocols outlined by the NCCLS in
their publication M7-A5, ÔMethods for Dilution Antimi-
crobial Susceptibility Tests for Bacteria that Grow Aero-
bically,Õ 5th ed., 2000; with reference to the standards in
their publication M100-S12, ÔPerformance Standards for
Antimicrobial Susceptibility Testing, Twelfth Informa-
tional Supplement,Õ 2002. For most(nonfasitdious) organ-
isms, the protocol is the same: compounds are dissolved to
10 mg/mL in water and diluted to 128 lg/mL in Mueller–
Hinton Broth, then successively diluted (2-fold dilutions) in
microtiter plates. Bacteria are suspended in Mueller–Hin-
ton Broth (cation adjusted) and mixed 1:1 with the diluted
antibiotic solution. Plates are incubated at 35 ꢁC for 16–
20 h. The MIC is read as the lowest antibiotic concentration
atwhich no growth is observed. All MICs are performed in
duplicate, with a no-antibiotic and a no-inoculum (sterility)
control.
consistent with a model where the linker lies in the groove
of the RNA. The 4,5-linked nebramine compounds 8a,b
were less active than their neamine homologues while
the 4,6-linked nebramine dimers were consistently more
active than their 4,5-linked homologues 12a–d. Also
intriguing were the MIC results obtained with the ex-
tended panel of P. aeruginosa (PAE_NUH01-7, 17-25),
Table 5. The bivalent AMGs were demonstrated to re-
tain activity against P. aeruginosa including strains resis-
tant to a variety of antipseudomonal agents. In
conclusion, these studies have established that the length
of the linker and its substituents were shown to be
important determinants of antibacterial activity in biva-
lent AMGs and the results appear to be consistent with
the model proposed in Figure 2. Furthermore, bivalent
AMGs showed potent activity against drug resistant
strains of P. aeruginosa and the lead compound OPT-
11 demonstrated the ability to protect mice in an infec-
tion model. Taken together, these data suggest that
bivalent AMGs may lead to the development of AMGs
active against bacteria that are resistant to a wide variety
of antipseudomonal drugs.
Acknowledgements
We thank Dr. Chi-Huey Wong for helpful discussions
and insightful suggestions. This work was supported,
in part, by a SBIR grant from the National Institute
7. NCCLS, Methods for antimicrobial susceptibility tests for
bacteria that grow aerobically, M100-S8 and M7-A5,
NCCLS, Wayne, PA, 2000.