Design of bifunctional antibiotics that target bacterial rRNA and inhibit resistance-causing enzymes [11]

Full text of DOI:10.1021/ja000575w

5230
J. Am. Chem. Soc. 2000, 122, 5230-5231
Design of Bifunctional Antibiotics that Target
Bacterial rRNA and Inhibit Resistance-Causing
Enzymes
Steven J. Sucheck,† Andrew L. Wong,† Kathryn M. Koeller,†
David D. Boehr,‡ Kari-ann Draker,‡ Pamela Sears,*,†
Gerard D. Wright,‡ and Chi-Huey Wong*,†
Department of Chemistry and The Skaggs Institute for
Chemical Biology, The Scripps Research Institute
10550 North Torrey Pines Road, La Jolla, California 92037
Antimicrobial Research Centre, Department of Biochemistry
McMaster UniVersity, 1200 Main Street
West Hamilton, Ontario L8N 3Z5, Canada
ReceiVed February 14, 2000
Deoxystreptamine-based aminoglycosides are a clinically im-
portant class of antibiotics that are effective against a broad range
of microorganisms.1 It is believed that aminoglycosides exert their
therapeutic effect by interfering with translational fidelity during
protein synthesis via interaction with the A-site rRNA on the 16S
domain of the ribosome.2 Unfortunately, the high toxicity and
rapid emergence of high-level aminoglycoside resistance have
severely limited the usefulness of this class of antibiotics. Num-
erous aminoglycoside resistance mechanisms have been identified,
and enzymatic acetylation, phosphorylation and ribosylation are
the primary causes of high level resistance in most clinical
isolates.3 Of the modifying enzymes, the acetyl- and phospho-
transferases (AAC and APH) have been extensively studied with
respect to their specificity.3,4 To tackle the problem of antibiotic
resistance, we are developing novel bifunctional aminoglycosides
that can resist or inhibit aminoglycoside-modifying enzymes while
simultaneously targeting ribosomal RNA.
The dissociation constant (Kd) and binding stoichiometry were
determined using surface plasmon resonance (SPR) against an
immobilized rRNA sequence modeling the A-site of prokaryotic
rRNA (Figure 1).5c,d The dissociation constants were obtained
from equilibrium binding curves through nonlinear curve-fitting
and were comparable to those obtained using Scatchard analysis.
We focused on neamine as it represents the simplest effective
aminoglycoside antibiotic and contains the key â-hydroxyamine
motif for interaction with the phosphodiester group and the
Hoogsteen face of guanine residues in RNA (Figure 1b).6 Neamine
was found to bind biotinylated AS-wt in a 2:1 complex with a
Kd of 10 µM for each binding site (Figure 1c). Various dimers of
neamine were therefore constructed to identify a bivalent ami-
noglycoside that would bind AS-wt with high affinity (Figure
1d), and at the same time resist or inhibit the modifying enzymes
due to its unnatural structure.7
Figure 1. (A) Biotinylated E. coli 16S rRNA A-site (AS-wt) rRNA
sequence, (B) The mode of action of â-hydroxyamine commonly found
in aminoglycosides antibiotics, (C) Binding isotherm and Scatchard plot
(inset) of neamine binding to AS-wt (circles) and control mutants
(U1406A, squares; U1485A, diamonds) for determination of dissociation
constants (Kd ) inverse slope) and binding stoichiometry (x-intercept).
The binding is sequence selective. (D) Energetic analysis of a bivalent
neamine.
Neamine dimers were prepared starting from perbenzyl per-
azido 5-O-carboxyethylneamine8 (see Scheme 1), which was
prepared from the 5-O-allyl precursor.5e Carboxyethylneamine was
distributed into a Quest 210 parallel synthesizer and was activated
using a cyclohexylcarbodiimide bound to macroporous polysty-
rene resin. Resin (2 equiv), acid (1 equiv), and various diamines
(0.4 equiv each) were utilized to synthesize a library of neamine
dimers of variable linker length. The intermediate amides were
isolated by filtration and were >95% pure, as determined by
NMR. The resulting dimers were first reduced under Staudinger
conditions to convert the azides to amines, which were captured
from solution using the resin-bound sulfonic acid scavenger MP-
TsOH (Argonaut). The resin was washed, and the free amine was
released from the resin by elution with 2 M NH3 in methanol.
The resulting amines were debenzylated by hydrogenolysis in the
presence of 2 equiv of acetic acid per amine. The reaction mixture
was filtered, concentrated, and purified by silica gel chromatog-
raphy using 8:2:4:5 NH4OH-CHCl3-n-BuOH-EtOH, followed
by cation exchange chromatography to give the pure aminogly-
coside dimers 4-13. The amide-linked dimers could also be
prepared via Ugi reactions, for example, dimer 14, starting from
the same perbenzyl perazido 5-O-carboxyethylneamine. This
procedure is also directly applicable to parallel synthesis and could
be used to increase the molecular diversity of the library.
The dimers with the highest affinity for AS-wt determined by
SPR were also the most potent antibiotics, as determined by the
† The Scripps Research Institute.
‡ McMaster University.
(1) Edson, R. S.; Terrel, C. L. Mayo Clin. Proc. 1991, 66, 1158.
(2) (a) Moazed, D.; Noller, H. F. Nature 1987, 327, 389. (b) Purohit, P.;
Stem, S. Nature 1994, 370, 659. (c) Formy, D.; Recht, M. I.; Blanchard, S.
C.; Puglisi, J. D. Science 1996, 274, 1367.
(3) (a) Wright, G. D.; Berghuis, A. M.; Mobashery, S. AdV. Exp. Med.
Biol. 1998, 456, 27. (b) Kondo, S.; Hotta, K. J. Infect. Chemother. 1999, 5,
1. (c) Mingeot-Leclerco, M.-P.; Glupczynski, Y.; Tulkens, P. M. Antimicrob.
Agents Chemother. 1999, 43, 727.
(4) (a) Daigle, D. M.; Hughes, D. W.; Wright, G. D. Chem. Biol. 1999, 6,
99. (b) Azucena, E.; Grapsas, I.; Mobashery, S. J. Am. Chem. Soc. 1997,
119, 2317. (c) Patterson, J.-E.; Zervos, M. J. ReV. Infect. Dis. 1990, 12, 644.
(5) (a) Recht, M. I.; Fourmy, D.; Blanchard, S. C.; Dahlquist, K. D.; Puglisi,
J. D. J. Mol. Biol. 1996, 262, 421. (b) Miyaguchi, H.; Narita, H.; Sakamoto,
K.; Yokoyama, S. Nucleic Acids Res. 1996, 24, 3700. (c) Hendrix, M.;
Priestley, E. S.; Joyce, G. F.; Wong, C.-H. J. Am. Chem. Soc. 1997, 119,
3641. (d) Wong, C.-H.; Hendrix, M.; Priestley, E. S.; Greenberg, W. A. Chem.
Biol. 1998, 5, 397. (e) Greenberg, W. A.; Priestley, E. S.; Sears, P. S.; Alper,
P. B.; Rosenbohm, C.; Hendrix, M.; Hung, S.-C.; Wong, C.-H. J. Am. Chem.
Soc. 1999, 121, 6527.
(6) Hendrix, M.; Alper, P. B.; Priestley, E. S.; Wong, C.-H. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 95.
(7) Some aminoglycoside dimers were prepared previously; however, the
monomoers bind the A-site stoichiometrically, see: Michael, K.; Wang, H.;
Tor, Y. Bioorg. Med. Chem. 1999, 7, 1361. For vancomycin dimers, see: Rao,
J.; Whitesides, G. H. J. Am. Chem. Soc. 1997, 119, 10286. Sundram, U. N.;
Griffin, J. H.; Nicas, T. I. J. Am. Chem. Soc. 1996, 118, 13107.
(8) Sucheck, S. J.; Greenberg, W. A.; Tolbert, T. J.; Wong, C.-H. Angew.
Chem., Int. Ed. 2000, 39, 1080.
10.1021/ja000575w CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/16/2000

Communications to the Editor
J. Am. Chem. Soc., Vol. 122, No. 21, 2000 5231
Scheme 1. Neamine Dimers Linked via Amides and
1,2-Hydroxyaminesa
a (a) CH3NH2; (b) 15, EtOH, 95 °C, 16 h; (c) i. P(CH3)3, THF, H2O;
ii. 20% Pd(OH)2/C (Degussa), H2, H2O, AcOH; (d) 16, EtOH, 95 °C, 16
h; (e) diamine, EtOH, 95 °C, 16 h.
Figure 2. The relationship between antibiotic activity (MIC, minimum
inhibitory concentration) and translation inhibition (IC50). The compounds
above the line do not target RNA and have different modes of antibiotic
action, while those to the right of the line exhibit transport limitations.
antimicrobial assays5e,9 and by IC50 of in vitro translation.5e Of
this series, the dimers with the highest antibiotic activity, 4 and
6, showed a Kd of 1.1 and 0.8 µM on AS-wt, respectively, 10-
fold greater than neamine. Dimers with longer linker lengths had
weaker affinities for AS-wt, a trend that correlated with antibiotic
activity. Interestingly, all of the dimers continued to display a
2:1 binding stoichiometry, indicating that the increase in affinity
is most likely due to an additional favorable (not dimeric) yet
weak interaction with AS-wt. Antibiotic activities of dimers 4
and 6 were comparable to neamine, MIC ) 31 and 125 µM,
respectively, against the Escherichia coli reference strain (See
Supporting Information for antibiotic testing data).
The relatively weak antibiotic activity of these dimers led us
to design a flexible and hydrophilic linker by opening the 1,2-
propyloxirane with an amine as shown in Scheme 1. The triflate
of (S)-(-)- and (R)-(+)- glycidol10 was used to alkylate perbenzyl
perazido neamine to form epoxides 15 and 16, respectively.
Epoxides 15 and 16 were heated for 16 h in a sealed tube with
excess methylamine to form 1,2-hydroxy amines 17 and 18,
respectively. These hydroxy amines could then be used in an
addition reaction with another equivalent of epoxide 15 or 16 to
form dimers 19 and 20, respectively, after deprotection. Epoxides
15 and 16 were also opened with 0.5 equiv of a N,N′-
methyldiamines to afford protected dimers 21-28. N,N′-meth-
yldiamines that were not commercially available were readily
prepared by a one-pot synthesis via imine formation with a
primary diamine and benzaldehyde, alkylation of the intermediate
imime with dimethyl sulfate followed by hydrolysis of the
alkylimine afforded N,N′-methyldiamines in high yield.11 The
resulting dimers were deprotected as previously described to
afford dimers 21-28. These dimers possessed significantly
increased antibiotic activity compared to the amide-linked dimers.
Antibiotic activity was greatest with the diaminobutane linker in
dimer 27, which showed a MIC ) 6.25, µM against E. coli and
Kd ) 40 nM (AS-wt) with 1-to-1 stoichiometry.12
In an attempt to better understand the relationship between
RNA binding and antibiotic activity, inhibition of in vitro
translation of luciferase gene5e was measured as a function of
MIC, Figure 2. This analysis was used to validate the target and
characterize potential transport limitations for the aminoglyco-
sides, and in vitro translation inhibition is expected to be a better
indicator of aminoglycoside selectivity for 16S rRNA compared
to binding affinity measurements with the A-site sequences.5c-e
A nearly linear relationship between the IC50 of translation
inhibition and the MIC was observed. This analysis should be
useful for analyzing structure activity relationships within a similar
series of compounds. Compounds falling below the line in Figure
2 may suffer from transport limitation while compounds above
the line may act via a fundamentally different mode of action
than compounds at or near the line.
Further study of neamime dimers 4, 6, and 27 using several
aminoglycoside-modifying enzymes revealed that the dimers were
poor substrates for AAC(6′)-Ii and APH(3′)-IIIa, responsible for
6′- and 3′-N-acetylation and O-phosphorylation, respectively.3 In
addition, dimers 4, 6, and 27 were poor substrates for the AAC-
(6′) activity of the bifunctional aminoglycoside modifying-enzyme
AAC(6′)-APH(2′′),3,4 and not substrates for the APH(2′′) activity
of AAC(6′)-APH(2′′). They were in fact potent competitive
inhibitors of the APH(2′′) activity, Kis ) 0.8 µM for dimer 4, 0.1
µM for 6, and 0.7 µM for 27.
In summary, the neamine dimers prepared in this study
represent a new class of aminoglycoside antibiotics that are
functionally simpler than previously known aminoglycosides. In
addition, they are potent inhibitors of the APH(2′′) activity of
the bifunctional AAC(6′)-APH(2′′) enzyme, one of the most
clinically significant of the aminoglycoside-modifying enzymes.
This research thus provides a new direction for the development
of novel antibiotics that target bacterial RNA and resistance-
causing enzymes.
Acknowledgment. This work was supported by the NIH (GM58439,
to CHW). G.D.W. was funded by the Medical Research Council of
Canada (MTI13536). We thank Pathogenesis Co. for the antibiotic assay
of resistant strains.
(9) Phillips, I.; Williams, D. In Laboratory Methods in Antimicrobial
Chemotherapy; Gerrod, L., Ed.; Churchill Livingstone Press: Edinburg, 1978;
pp 3-30.
(10) (a) Baldwin, J. J.; McClure, D. E.; Gross, D. M.; Williams, M. J.
Med. Chem. 1982, 25, 931. (b) Schlecker, R.; Thieme, P. C. Tetrahedron
1988, 44, 13289.
Supporting Information Available: Detailed procedures for the
synthesis of compounds 4-28 as well as their characterization and
functional analysis (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
(11) Devinsky, F.; Lacko, I.; Krasnec, L. Synthesis 1980, 4, 303.
(12) Compound 27 is also effective against other strains, including
Pseudomonas aeruginosa ATCC 27853, P. aeruginosa, PAO-1, S. aureus
ATCC 29213 and ATCC 33591-MRSA, and Enterococcus faecalis ATCC
29212 and is 3 times more effective than tobramycin against the tobramycin-
resistant strain of P. aeruginosa from cystic fibrosis patients.
JA000575W