Rapid synthesis, RNA binding, and antibacterial screening of a peptidic-aminosugar (PA) library

Article abstract of DOI:10.1021/cb5010367

A 215-member mono- and diamino acid peptidic-aminosugar (PA) library, with neomycin as the model aminosugar, was systematically and rapidly synthesized via solid phase synthesis. Antibacterial activities of the PA library, on 13 bacterial strains (seven G

Full text of DOI:10.1021/cb5010367

Articles
pubs.acs.org/acschemicalbiology
Rapid Synthesis, RNA Binding, and Antibacterial Screening of a
Peptidic-Aminosugar (PA) Library
Liuwei Jiang,† Derrick Watkins,‡ Yi Jin,† Changjun Gong,† Ada King,‡ Arren Z. Washington,§
Keith D. Green,∥ Sylvie Garneau-Tsodikova,∥ Adegboyega K. Oyelere,§ and Dev P. Arya*,†,‡
†Laboratory of Medicinal Chemistry, Department of Chemistry, Clemson University, Clemson, South Carolina 29634, United States
‡NUBAD, LLC, Greenville, South Carolina 29605, United States
§School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332, United States
∥College of Pharmacy, Department of Pharmaceutical Sciences, University of Kentucky, Lexington, Kentucky 40536-0596, United
States
S
* Supporting Information
ABSTRACT: A 215-member mono- and diamino acid peptidic-aminosugar (PA) library, with neomycin as the model
aminosugar, was systematically and rapidly synthesized via solid phase synthesis. Antibacterial activities of the PA library, on 13
bacterial strains (seven Gram-positive and six Gram-negative bacterial strains), and binding affinities of the PA library for a 27-
base model of the bacterial 16S ribosomal A-site RNA were evaluated using high-throughput screening. The results of the two
assays were correlated using Ribosomal Binding-Bacterial Inhibition Plot (RB-BIP) analysis to provide structure−activity
relationship (SAR) information. From this work, we have identified PAs that can discriminate the E. coli A-site from the human
A-site by up to a 28-fold difference in binding affinity. Aminoglycoside-modifying enzyme activity studies indicate that APH(2″)-
Ia showed nearly complete removal of activity with a number of PAs. The synthesis of the compound library and screening can
both be performed rapidly, allowing for an iterative process of aminoglycoside synthesis and screening of PA libraries for optimal
binding and antibacterial activity for lead identification.
minoglycoside antibiotics, e.g., amikacin, gentamicin,
kanamycin, and neomycin, are a class of clinically significant
activity: maintaining the backbone of aminoglycosides but
adding additional recognition/binding elements. Schweizer et
al. showed that amphiphilic neomycin-peptide/aromatic ring
conjugates with triazole or amide linkages on the 5″ position
exhibit better activity than neomycin against neomycin-resistant
strains.8−10 Substitution of the 5″-OH group of neomycin with
mono- or disaccharides also has shown improved activity
compared to that of neomycin against neomycin-resistant
strains.11−13 These previous studies produced compounds with
improved binding affinity to rRNA and more potent antibacterial
activity than the original aminoglycosides.13
A
chemotherapeutics, which have broad-spectrum activities against
both Gram-negative and Gram-positive bacteria.1,2 Aminoglyco-
sides inhibit bacterial growth by binding to the A-site decoding
region of the bacterial 16S rRNA within the 30S ribosomal
subunit. A-site binding by aminoglycosides leads to mistransla-
tion of mRNA or premature termination of protein synthesis.
The loss of protein translational fidelity results in the death of
bacteria.3−6
Emergence of drug resistance has become a public problem in
applying aminoglycoside antibiotics in the therapy of bacterial
infections.1,7 In response to the requisite but still deficient rRNA
binding affinity research, an approach for a comprehensive study
of the binding and antibacterial activity of aminoglycoside-small
molecule conjugates is urgently needed. A common strategy is
used in the aforementioned studies to improve antibacterial
Received: September 15, 2014
Accepted: January 28, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a
Scheme 1. Synthesis of Neomycin-Amino-Acid Conjugates
a
Reagents and conditions: (a) glutaric anhydride, Et3N, DMF, r.t., 95%.
Amino acids (natural and unnatural) that can comprise
peptides have various chemical structures, polarity, electrical
charges, and hydrophobicity/hydrophilcity resulting in diverse
biological activity. Some peptides are known for their ability to
interact with microbial lipid membranes, resulting in destabiliza-
tion, translocation, pore formation, or lysis of microorgan-
isms.14,15 A number of cationic amphiphilic peptides are found to
have the ability to traverse intact cytoplasmic membranes to
interact with internal targets.16 The chemical diversity of amino
acids, their ability to form contacts with ribosome RNA and
protein constituents, and their ability to regulate uptake across
the bacterial membrane makes them perfect moieties for
conjugation to enhance the A-site binding affinity and/or
antibacterial activity of aminoglycosides. The ribosome (RNA
and peptides) offers a multitude of sites for interaction with the
peptide-linked aminosugars. Moreover, newer targets for
aminoglycosides are being discovered, further enhancing the
need for efficient and rapid methods for their synthesis and
evaluation.17−21
describing the ligand−target interaction (aminoglycoside−
rRNA).8,9,13,22−24 Recently, a single concentration assay was
adapted to screen compound libraries for antibacterial activity.
The results of the assay were well correlated with antibacterial
activity of bacterial strains determined by MIC calculations.18
Antibacterial activity was correlated with data reporting on
binding to the ribosomal A-site to provide information about the
possible mechanism of action. The A-site binding data were
determined from a fluorescent probe displacement assay that can
be performed rapidly to screen large libraries of compounds in a
high-throughput format.18,19 The A-site binding affinity and
antibacterial activity from the compound library can be rapidly
analyzed by Ribosomal Binding-Bacterial Inhibition Plots (RB-
BIP). Using the RB-BIP here, we correlate the antibacterial
activity to the A-site binding affinity of compound libraries. We
have successfully employed RB-BIP to analyze a neomycin dimer
library in a previous study.18 With RB-BIP, the amino-acid-
neomycin library will provide important structure−activity
relationship (SAR) information with which aminoglycoside-
based lead compounds may be designed.
Recent studies have shown increased antibacterial activities
upon conjugation of aminoglycosides with small molecules, as
demonstrated by improvement in MICs; however, the majority
of these studies have not provided binding information
The solution phase synthesis of aminoglycoside conjugates,
utilized in studies to date,25−34 have limited the number of
compounds that can be synthesized and studied. A library
B
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 1. (A) Structure of F-neo reporter molecule. (B) Schematic of the displacement assay of F-neo (cyan) by a competing ligand (neomycin; green)
from the model ribosomal A-site (red and blue). The yellow star indicates fluorescence. (C) Ribosomal RNA A-site binding screening results of 215
neomycin-amino acid conjugates. The percentage displacement of neomycin was determined as described in eq 1. Neomycin (Neo) is shown in a red
box and was set to 100%. The bindings of all other compounds are relative to the binding of neomycin. Each experiment was performed in duplicate, and
standard deviation in the measurement is shown as error bars for each compound. The assay was performed with each compound at a 0.3 μM
concentration to 0.1 μM F-neo:A-site complex using 100 reads per well at an excitation/emission of 485/535 nm. All experiments were performed in
duplicate in 10 mM MOPSO (pH 7.0), 50 mM NaCl, and 0.4 mM EDTA.
approach to screening requires a rapid method for conjugation of
required moieties to the aminoglycoside. Here, we show that 215
peptidic-aminosugars (PAs) with mono- and diamino acids were
rapidly synthesized by using neomycin and 14 amino acids via
solid phase synthesis. We also report a SAR analysis of A-site
binding and antibacterial activity for amino acid-neomycin
conjugated PAs. Because the synthesis, screening, and analysis
can be performed rapidly, we have employed RB-BIP analysis for
every compound in the library. The synthesis and screening
approach described here is amenable to the development of
larger and more diverse aminoglycoside modifications using solid
phase synthesis. Coupled with rapid screening approaches
presented here, the identification of useful hits can be rapidly
performed.
thoxycarbonyl (Fmoc)-based solid-phase peptide synthesis
protocol (Scheme 1).35 Fourteen L-amino acids (β-alanine,
arginine, asparagine, aspartic acid, cysteine, histidine, leucine,
phenylalanine, proline, serine, threonine, tryptophan, tyrosine,
and valine) were studied in the full library. In addition to the full
library, five lysine-neomycin conjugates were synthesized for
comparison with arginine-neomycin compounds. Protected 5″-
amino-neomycin (1) was synthesized by modifying previous
methods.36−39 A glutaric anhydride reaction with 5″-amino-
neomycin yielded the amide bond linked monomer 2 required
for solid-phase synthesis (Scheme 1). Due to presence of the
large tert-butoxycarbonyl (Boc) group, as observed via character-
istic resonances in proton NMR spectra, compound 2 was
deprotected to remove all Boc groups before NMR character-
ization. This modified method is capable of synthesizing
hundreds of PAs in a short time, on small scales with low costs
and minimal labor. The ribosomal A-site binding affinity and
bacterial growth inhibition were then examined for all 215
compounds in the library.
RESULTS AND DISCUSSION
■
Synthesis of a Neomycin PA Library. A 215 PA library
containing mono- and diamino acid conjugated neomycin was
rapidly synthesized by modifying the standard 9-fluorenylme-
C
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a
Table 1. A-site Binding Analysis
a
The compounds were analyzed based on the charge on individual amino acids. The numbers are the percentage displacement of F-neo from an A-
site/F-neo complex by peptidic neomycin as shown in Figure 1C for each DPA numbered compound as a function of each amino acid(s) conjugated.
The third row gives the one-letter code of the amino acid on the conjugate position 1 (firstly conjugated to neomycin). The second column gives the
one-letter code of the amino acid on conjugate position 2 (secondly conjugated to neomycin). Results are color coded as total charges of the amino
acids on peptidic neomycin. *Total charges of the amino acids on peptidic neomycin under physiological pH. Zwitter indicates that the peptidic
neomycin contains an amino acid with a positively charged side chain and an amino acid with a negatively charged side chain.
Screening of the Peptide-Neomycin Conjugate Library
for A-site Binding. The peptide-neomycin compound library
was screened for A-site binding affinity using a high-throughput
assay previously developed in our laboratory.19 This method is a
fluorescence-based competition assay, which uses a 27-base RNA
model of the ribosomal A-site and a fluorescent reporter
molecule, F-neo (Figure 1). The fluorescence of F-neo increases
when competitively displaced by an A-site binding compound.
This screening technique is used for measuring the relative
binding affinity of a compound library compared to neomycin
(Figure 1C). The assay was validated for A-site binding by
determining the relative displacement of known aminoglycosides
with previously reported results;40 additionally, the assay has
been successfully employed in screening of a neomycin dimer
compound library.18
Data collected for the single amino acid-neomycin conjugates
demonstrated that the A-site has a strong preference for neomycin
conjugated to positively charged amino acids, such as arginine and
lysine. These conjugates have binding affinities for the A-site that
is greater than that of neomycin alone (Table 1). Polar residues,
overall, reduce the affinity of neomycin less than nonpolar
residues, with cysteine maintaining a moderate binding affinity
for the A-site while serine and asparagine have half the original
affinity of neomycin. With the exception of β-alanine, all
nonpolar amino acids resulted in a reduction of neomycin
binding below 50% when conjugated to neomycin. The
hydrophobic aromatics of phenylalanine and tyrosine reduce
the binding of the conjugated neomycin to 33% of the
nonconjugated neomycin, while the more polar tryptophan
only reduces the binding to 62%. The negatively charged
aspartate is the most detrimental residue to binding, reducing the
conjugate to 14% binding of the nonconjugated neomycin.
The docking of compounds into the ribosomal A-site suggests that the
amino acid conjugated to neomycin may alter the orientation of the
rings on the aminoglycoside. Docking of neomycin alone results in
an orientation of the aminoglycoside rings that is similar to that
of the crystal structure of neomycin bound by the E. coli
ribosomal A-site (Figure 2).41 However, the addition of the
amino acid side chain to neomycin results in an alternate
conformation of ring I and ring II as well as the positioning of the
amino acid side chain between rings III and IV and the floor of
the RNA groove. Future structural work will determine the
relevance of the calculations for the binding of the amino acid
conjugates, but these results indicate that the introduction of the
amino acid side chain to neomycin introduces alternate binding
modes to the aminoglycoside.
The diamino acid conjugates prefer only a single positive charge,
cysteine or aromatic residues. An arginine at position 1 results in
three diamino acid conjugates with higher affinity than that of the
nonconjugated neomycin, with NeoRW binding 20% higher.
Three of the four lysine diamino acid conjugates have affinity
almost equivalent to or greater than that of neomycin. However,
arginine at postion 2 has negative effects on the binding of most
compounds, indicating that consecutive positive charges
diminish binding affinity. Cysteine also appears to be beneficial
for binding in the diamino acid conjugates, with three
compounds binding better than neomycin when cysteine is
D
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
sites indicates that the conjugation of amino acids has the
potential to produce compounds that are highly specific for the
rRNA target (Figure 3).
Figure 2. Docking of neomycin and NeoR (DPA1102) onto the E. coli
ribosomal A-site. Neomycin (A: cyan sticks) and NeoR (B: green and
blue sticks) were docked onto a model of the ribosomal A-site (partially
transparent surface overlaid onto a cartoon representation) using
Autodock Vina. Similar to the crystal structure, ring II of neomycin in
panel A is below ring I and closer to the unstacked adenines. The
neomycin moiety (green sticks) in panel B has ring II below ring I and
closer to the unstacked adenines. The amino acid side chain (blue sticks)
is between the floor of the groove and rings III and IV. The E. coli
ribosomal A-site was modeled from the crystal structure of neomycin in
complex with the A-site (PDB ID: 2A04).41
present at position 1, including the highest affinity compound,
NeoCY.
A tryptophan at position 1 has the most consistent positive
effect on the binding of the diamino acid conjugates. Despite the
fact that the monoamino acid of tryptophan has diminished
binding compared to the nonconjugated neomycin, four of the
conjugates have greater affinity than the nonconjugated
neomycin, four of the conjugates have reduced affinity of 15%
or less, and only three of the conjugates with tryptophan at
position 1 have the affinity reduced by more than 50%. Although
the aromatic nature of tryptophan would indicate that stacking
interactions are important, the presence of the other aromatic
compounds, phenylalanine and tyrosine, reduces the affinity of
all the diamino acid conjugates with most having over 50%
reduction in affinity compared to neomycin. This observation
indicates that the size and/or presence of the amine of
tryptophan contribute to the affinity of the conjugates.
The presence of the negatively charged aspartate is detrimental
to binding when present in the monoamino acid conjugate, or
when present at position 1 or position 2 in the diamino acid
conjugates. The electrostatic repulsion between aspartate with
the phosphate backbone likely precludes the presence of any
negatively charged residues for RNA binding in the small chain of
amino acids.
The conjugation of amino acids to neomycin increases the difference
in the affinity of the aminoglycoside between the E. coli and the
human ribosomal A-site models. Previous work has been shown
that there is approximately a 5-fold difference in the affinity of
neomycin for the human A-site model and the E. coli model.42
The screening of the compound library for F-neo displacement
from the human A-site identified compounds that discriminate
between the two models. The results of the screen were
confirmed for two of the identified compounds, NeoWL and
NeoYW, by determining the IC50 for the compound displace-
ment of F-neo from each site. Each of these compounds has a
greater difference in binding affinity for the E. coli A-site and the
human A-site than that of neomycin, with NeoWL having almost
a 15-fold greater affinity of the E. coli A-site and NeoYW having
almost a 30 fold greater affinity for the E. coli A-site. The
identification of a compound that can discriminate between A-
Figure 3. PA compounds that discriminate between the E. coli and
human model A-site. The secondary structure of the model A-site used
in the development of the high throughput screen is shown at the top.
The red box indicates the residues that are present in the E. coli or human
rRNA and are numbered according to the E. coli ribosome. Compounds
DPA1199 (NeoWL) and compound DPA1308 (NeoYW) bind to the E.
coli A-site (left) 14−28 times stronger than to human A-site (right).
The Screening of the PA Library for Bacterial Growth
Inhibition. The bacterial growth inhibition of the neomycin PA
library was determined by using a single concentration high-
throughput screen (HTS) that was developed by De La Fuente et
al.43 This screening method has been used in our laboratory to
identify compounds with antibacterial activity and was validated
by determining MIC of the compounds.44 Antibacterial HTS was
performed on seven Gram-positive bacterial strains (Staph-
ylococcus aureus ATCC 25923, S. aureus NorA,45 methicillin-
resistant S. aureus (MRSA) ATCC 33591, Listeria monocytogenes
ATCC 19115, Streptococcus pyogenes ATCC 12384, Enterococcus
faecalis ATCC 49533, E. faecalis ATCC 700802) and six Gram-
negative bacterial strains (Escherichia coli ATCC 25922, E. coli
TolC,45 Enterobacter cloacae ATCC 13047, Serratia marcescens
ATCC 13880, Pseudomonas aeruginosa ATCC 27853, Acineto-
bacter baumannii ATCC 19606) for all PA at 5 μM. The resulting
data are presented in the Supporting Information in detail with
errors associated with each measurment (Table S1). The amino
acid preferences on bacterial growth inhibition are summarized
E
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Table 2. Favored Amino Acids from Antibacterial Screening at 5 μM
*
The average inhibition and standard deviation from the mean (σ) was taken for all compounds in the library. The σ for the inhibition was
determined for each compound. The average σ for all compounds with the indicated amino acid at postion 1 (2) was determined for all compound
variations (X) at position 2 (1).
Table 3. Minimal Inhibitory Concentration (MIC) in μM
bacterial strains (μM)
name
Neo
DPA no.
A
B
C
D
E
F
G
H
I
J
0.625
1.25
5
>320
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
>80
>80
>80
>80
>80
>80
>80
>80
>80
>80
>80
>80
>80
10
10
40
40
40
20
20
20
40
20
20
20
20
2.5
2.5
160
80
2.5
1.25
5
>320
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
>160
80
80
NeoR
1102
1113
1115
1142
1145
1173
1229
1243
1257
1271
1285
1299
5
20
20
80
5
20
80
NeoC
>160
>160
>160
>160
>160
80
>80
>80
>80
>80
>80
80
NeoK
5
80
5
NeoCR
NeoRN
NeoRH
NeoRS
NeoRT
NeoRY
NeoRV
NeoRC
NeoRW
10
>80
20
5
5
2.5
5
5
20
5
40
1.25
5
20
2.5
5
5
20
>160
>160
160
>80
>80
80
5
10
20
20
>80
20
2.5
5
5
20
5
40
160
5
>80
80
10
>80
160
10
A = S. aureus ATCC 25923, B = MRSA ATCC 33591, C = P. aeruginosa ATCC 27853, D = E. coli ATCC 25922, E = E. cloacae ATCC 13047, F = S.
marcescens ATCC 13880, G = S. pyogenes ATCC 12384, H = S. aureus NorA, I = E. coli TolC, J = L. monocytogenes ATCC 19115.
in Table 2. Results of S. pyogenes ATCC 12384, E. faecalis ATCC
29533, E. faecalis ATCC 700802, and A. baumannii ATCC 19606
are omitted from the table due to negligible bacterial growth
inhibition.
above or below the mean for each class of compounds that
inhibited the growth of each strain. This treatment allows us to
determine a quantitative measurement of the preferred amino
acid at each position.
In order to define amino acid preferences at position 1 and at position
2 in the neomycin conjugates, we analyzed the inhibition data based
on the average standard deviation (σ) for all compounds with specific
residue at position 1 or position 2 above the mean of inhibition of all
compounds for a specific bacterial strain. The average inhibition
and standard deviation from the mean for all compounds was first
determined for each strain. The average inhibition was found for
each class of compounds NeoYX or NeoXY, where Y is a specific
amino acid and X is varied across all amino acids used in the
conjugation. The average for each class was divided by the
standard deviation of all conjugates to determine the number of σ
The inhibition of amino acid conjugates on bacterial growth shows a
strong preference for the positive amino acids arginine and lysine. The
NeoR and NeoK conjugates show the broadest impact of all
amino acid conjugates at position 1, with the average over 1σ
above the mean for six of the 10 strains tested for NeoRX and
three of the 10 strains tested for NeoKX. NeoRX compounds
were favored approximately 2σ above the mean for Gram-
positive bacterial strains S. aureus ATCC 25923, S. aureus NorA,
and L. monocytogenes ATCC 19115. NeoRX was also favored 2σ
above the mean for Gram-negative bacterial strains E. coli ATCC
25922, E. cloacae ATCC 13047, and S. marcescens ATCC 13880
F
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 4. RB-BIP of neomycin amino acid compounds with arginine on conjugate position 1.
and 0.78σ above the mean for E. coli TolC. NeoKX compounds
were favored over 1σ above the mean for Gram-positive bacterial
strains S. aureus NorA and L. monocytogenes ATCC 19115 and
Gram-negative bacterial strains E. coli TolC and E. cloacae ATCC
13047 and 2.13σ above the mean for E. coli ATCC 25922.
Both cysteine and tryptophan amino-acid-neomycin deriva-
tives, NeoCX and NeoWX, were also broadly effective in
inhibiting the strains, with both classes of compounds being the
only derivatives that were significantly above the mean for
inhibiting the Gram-negative P. aeruginosa ATCC 27853. The
NeoCX conjugates appear more selective toward the Gram-
positive strains, inhibiting S. aureus ATCC 25923, S. aureus
NorA, and L. monocytogenes ATCC 19115 by over 0.5σ. In
addition to P. aeruginosa ATCC 27853, NeoWX conjugates also
inhibit the Gram-negative strain E. coli TolC. Most significantly,
NeoWX compounds were favored 1σ above the mean for
inhibition of neomycin resistant MRSA ATCC 33591.
The amino acid at position 2 has less influence on inhibition
compared to that of the amino acid at position 1. Only the
NeoXW class of compounds appears to have a consistent
influence on inhibition. NeoXW is favored over 1σ for the Gram-
positive neomycin resistant MRSA ATCC 33591 and the Gram-
negative P. aeruginosa ATCC 27853. NeoXW and NeoWX are
the only class of the conjugates that are favored for both of these
strains, indicating that the presence of a tryptophan may
influence mechanisms of resistance that arise in bacteria.
previously shown that HTS at a single concentration for these
strains is an excellent predictor of ligand antibacterial effects, as
verified by MIC determinations. To verify the extrapolation of
single point screening to MIC, we randomly picked a few ligands
and determined their MIC values for select strains. MIC results
show excellent correlation with single point determination
(Table 3). For example, the MIC of NeoR was determined for S.
aureus ATCC 25923 and MRSA ATCC 33591. In both cases, the
MIC agreed with the results of the single point screen for NeoR,
with NeoR having a MIC of 1.25 μM in S. aureus and inhibiting
95% growth in the single point screen. The MIC of NeoR in
MRSA was determined to be >160 μM, which is also in
agreement with the 2% inhibition observed in the single point
screen.
Furthermore, the MIC was also determined for NeoRW for S.
aureus ATCC 25923 and MRSA ATCC 33591. Consistent with
the single point screen, the MIC for NeoRW was greater than
that of NeoR in S. aureus at 10 μM and inhibited growth by 85%
in the single point screen. The MIC for NeoRW was also found
to be between 40 and 80 μM in MRSA. This compound inhibited
the growth of MRSA by 45% in the single point screen and
indicated that the single point screen is representative of the MIC
of the compound for the neomycin resistant strain. Neomycin
alone required >320 μM for complete inhibition of MRSA
growth and showed negligible inhibition in the single point
screen for this strain.
The results from the high-throughput antibacterial screen were
Correlating Ribosomal Binding Data and Bacterial
further verified by MIC calculation of select compounds. We have
Inhibition Data with RB-BIP Analysis. In order to rapidly
G
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Table 4. IC50 of PA Compounds for Cell Free Luciferase Inhibition Assay
compound
Neo
24
NeoK
262
NeoR
34
NeoRH
226
NeoRC
104
NeoRS
34
NeoRT
56
NeoRV
73
NeoRW
58
NeoRY
56
NeoC
118
NeoCR
246
a
IC50 (nM)
a
IC50 is the concentration of compound that inhibits the translation of luciferase as determined by the luciferase assay as described in the Materials
and Methods. Fits and errors of measurements can be found in the Table S13 of Supporting Information.
draw structure activity relationships for the compounds tested,
the data for both binding and inhibition assays of the PAs were
plotted on RB-BIP. A similar assay has been used to correlate
bacterial growth inhibition with translational inhibition.46 RB-
BIP analysis was performed on neomycin position 1 conjugates
of arginine, cysteine, and tryptophan. A representative plot with
arginine on conjugate position 1 from the two screens is shown in
Figure 4, and the plots for the cysteine and tryptophan
conjugates are given in the Supporting Information (Figures
S1, S2). RB-BIPs outline the percent ribosomal binding of the
compound as compared to neomycin on the y axis and the
percent bacterial inhibition as compared to untreated cultures on
the x axis. The RB-BIP can then be divided into quadrants I, II,
III, and IV. Compounds plotted in quadrant II display both high
binding to the ribosomal A-site and inhibition of bacterial
growth. The correlation using the RB-BIP analysis for the PA
compound library is similar to that observed by Wong et al. in
correlating naturally occurring and bifunctional aminoglycosides,
in which a linear relationship is observed between the MIC and
translation inhibition. This correlation between binding and
inhibition is similar to what we observe in quadrant II.46
Compounds plotted in quadrant I exhibit strong binding to the
model ribosomal A-site but weakly inhibit bacterial growth.
Similar to previous work, the compounds in quadrant I could be
correlated to transport limitations,46 or other modes of resistance
associated with ineffective binding. Quadrant III of the RB-BIP
contains compounds with low ribosomal binding and inhibition
of bacterial growth. Finally, quadrant IV shows compounds
exhibiting inhibition of bacterial growth with weak binding to the
model A-site, which could have a different mode of action
associated with the bacterial inhibition.46
unlikely to be the target of NeoRN and NeoRR for inhibition of
bacteria growth.
The NeoRW conjugate was the most effective compound
against the neomycin resistant strains MRSA ATCC 33591 and
P. aeruginosa ATCC 27853 as well as one of the strongest A-site
binding compounds. Additionally, in the MRSA strain, it very
closely approaches quadrant II. Its broad spectrum of activity and
improved inhibition against neomycin-resistant strains makes
NeoRW an attractive compound to use for expanding and further
optimizing the PA library.
The translation inhibition ability of the 10 compounds active in the
antimicrobial assays was tested in a cell-free translation inhibition
assay. A number of these compounds (Table 4 and Table S13)
are potent nanomolar inhibitors of protein synthesis, suggesting
that these amino acid modifications do not adversely affect the
ability of aminoglycosides to inhibit bacterial protein synthesis.
The 12 most promising PA conjugates were tested against a variety of
aminoglycoside-modifying enzymes. The chosen enzymes included
three regiospecific acetyltransferases,47,48 a phosphotransfer-
ase,49 and an acetyltransferase enzyme known to multiply
acetylate aminoglycosides.47 We can use the data obtained from
the AME reactions to predict how the peptide-aminoglycoside
conjugates might be modified in bacterial cells. The data
presented in Figure 5 show that in almost all cases, the reactivity
As shown in Figure 4, the position 1 arginine conjugate that
showed high bacterial growth inhibition (inhibition >50% at 5
μM) on neomycin susceptible Gram-positive bacterial strains (S.
aureus ATCC 25923, S. aureus NorA, and L. monocytogenes
ATCC 19115) are generally good A-site binding ligands, namely,
NeoR, NeoRS, NeoRT, NeoRY, NeoRV, NeoRC, and NeoRW.
The presence of a compound in or approaching quadrant II gives
a strong SAR correlation between antibacterial activity and A-site
affinity, indicating that the primary target of these conjugate
molecules is likely the A-site of 16S rRNA. The same general
trend is true for compounds against the neomycin susceptible
Gram-negative strains E. coli ATCC 25922, E. coli TolC, E.
cloacae ATCC 13047, and S. marcescus ATCC 13880 with the
majority of the compounds falling in quadrant II. However, more
of the compounds tend toward quadrant I for the Gram-negative
strains than for the Gram-positive strains, indicating a preference
for inhibition of Gram-negative strains.
Compounds NeoRN and NeoRR show up in quadrant IV of
many of the bacterial strains. NeoRN has low A-site affinity but
good bacterial growth inhibition on three out of seven neomycin
susceptible bacterial strains. NeoRR showed very low A-site
affinity (0% displacement on A-site screen) but good bacterial
growth inhibition on S. aureus NorA, L. monocytogenes ATCC
19115, and E. coli TolC. From the RB-BIP, our model A-site is
Figure 5. PA conjugates tested against a variety of aminoglycoside-
modifying enzymes.
of the peptide-neomycin conjugate is reduced when compared to
neomycin alone. The three regiospecific AACs, AAC(2′)-Ic,
AAC(3)-IV, and AAC(6′)-Ie/APH(2″)-Ia (used solely for its
acetylation activity), targeting the I and II rings show a reduction
in rate of 50% or less with the exception of AAC(6′)-Ie/
APH(2″)-Ia, with DPA1115 and DPA1229. However, this is the
most substrate- and cosubstrate-promiscuous of the three
regiospecific AACs.49 Eis was unaffected by the addition of the
peptides to the neomycin structure; this is not all that surprising
considering the large active site of the enzyme and its ability to
H
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Miroflex MALDI-TOF spectrometer. E. cloacae ATCC 13047, E. coli
ATCC 25922, P. aeruginosa ATCC 27853, S. marcescens ATCC 13880,
and MRSA ATCC 33591 were purchased from the American Type
Culture Collection (ATCC; Manassas, VA). S. aureus ATCC 25923 and
S. pyogenes (ATCC 12384) were a gift from Dr. Tamara McNealy
(Clemson University, Clemson, South Carolina). L. monocytogenes
ATCC 19115, E. faecalis ATCC 700802, E. faecalis ATCC 49533, A.
baumannii ATCC 19606, S. aureus NorA, and E. coli TolC were from Dr.
Garneau-Tsodikova’s laboratory (University of Kentucky). All culture
supplies were purchased from Thermo Fisher Scientific (Pittsburgh,
PA) and Sigma-Aldrich (St. Louis, MO).
acetylate neomycin three times. There was a reduction in activity
with DPA1102, DPA1113, DPA1142, and DPA1285 with Eis.
However, it should be noted that three of these compounds have
a free thiol group (cysteine) that reacted with the indicators used
for the acetyltransferase assay, and this decrease could be
indicative of the consumption of indicator through this
mechanism.
The most affected enzyme was APH(2″)-Ia, which showed
nearly complete removal of activity with 8 of the 12 compounds.
Since this enzyme is noted to modify the 3′ or 3‴ position of
neomycin and other 4,5-DOS substituted aminoglycosides50
rather than the actual 2″ position, the data indicate that
modifying the 5″ position of neomycin with amino acids and
peptides reduces the ability of neomycin to be phosphorylated by
APH(2″)-Ia. While there was no singular compound that
completely removed all activity of the modifying enzymes tested,
it is possible that the reduction in the rate of modification could
increase the time the unmodified PA conjugate is available to
target the ribosome and perform its original function.
Conclusion. Large libraries of amino acid and aminosugar
conjugated PAs can be rapidly synthesized via solid phase as an
approach to find new effective broad-spectrum antibiotics. The
analysis of the library identifies the residues that are preferred in
the first two positions of an amino acid chain. The work
presented here provides the first building block in the expansion
of the library and identifies the optimal residues to further expand
and optimize the peptide chain. Through rapid synthesis and
screening, we identified PAs that bind strongly and selectively to
the E. coli A-site model with up to 28 times greater affinity than to
the model human A-site rRNA. To the best of our knowledge,
this is one of the most selective E. coli rRNA aminoglycoside ever
reported, particularly for a neomycin class aminoglycoside, which
is known to have very poor selectivity for bacterial over
mammalian rRNA.
The approach described in this work can identify compounds
that correlate well between binding and growth inhibition, as well
as those that bind well but do not inhibit the growth of bacteria.
The resistance of bacteria to compounds that bind to the A-site
may be a product of limited transport or a promiscuous enzyme.
It has been previously shown that the RNA of the A-site is
dynamic, and compounds that induce the destacking of A1492
and A1493 of the E. coli 16S rRNA result in mistranslation by the
ribosome.51−54 Further analysis of the high affinity compounds
that do not inhibit bacterial growth can be followed by dynamic
assays using 2-amino purine labeled RNA and an in vitro
translation inhibition assay to assess the mode of resistance and
help direct future modifications of these compounds.
Compound 1. Compound 1 was synthesized by modifying
previously reported methods.36−39 A mixture of neomycin sulfate
(9.09 g, 10.0 mmol), Na2CO3 (13.78 g, 130.0 mmol), di-tert-butyl
dicarbonate (28.37 g, 130.0 mmol), H2O (480 mL), and CH3OH (480
mL) was stirred at RT for 3 h. After removal of 500 mL of the solvent
under reduced pressure, the solid was collected by filtration and washed
with H2O. The solid was then dried and purified by column
chromatography using EtOAc/hexane/isopropanol (v/v, 90/30/3) as
an eluent. Intermediate 1,3,2′,6′,2‴,6‴-hexa-N-(tert-butoxycarbonyl)-
neomycin was obtained as a white solid in 61% yield (7.41 g).
Intermediate 1,3,2′,6′,2‴,6‴-hexa-N-(tert-butoxycarbonyl)-neomycin
(4.86 g, 4.0 mmol) was then dissolved in dry pyridine (40 mL). To
the above solution, 2,4,6-triisopropylbenzenesulfonyl chloride (TPSCl;
9.69 g, 32.0 mmol) was added. The reaction mixture was stirred at RT
for 41 h and was then neutralized with saturated NaHCO3. The turbid
liquid was extracted with EtOAc, and the organic layer was washed with
saturated NaCl, aqueous HCl (0.5 M; two times), saturated NaHCO3,
and H2O. After solvent evaporation, the solid was then purified by
column chromatography using CH2Cl2/CH3OH/EtOAc (v/v, 100/3/
3) as an eluent. Intermediate 1,3,2′,6′,2‴,6‴-hexa-N-(tert-butoxycar-
bonyl)-5″-O-(2,4,6-triisopropylbenzenesulfonyl)-neomycin was ob-
tained as a white solid (3.08 g, 51%). A mixture of 1,3,2′,6′,2‴,6‴-
hexa-N-(tert-butoxycarbonyl)-5″-O-(2,4,6-triisopropylbenzenesulfon-
yl)-neomycin (1.46 g, 0.98 mmol), NaN3 (0.96 g, 14.7 mmol), and N,N-
dimethylformamide (DMF; 15 mL) was stirred for 12 h at 80 °C. The
mixture was then extracted with EtOAc and washed with H2O. After
solvent removal, 1,3,2′,6′,2‴,6‴-hexa-N-(tert-butoxycarbonyl)-5″-
azido-neomycin was obtained as a white solid in 96% yield (1.17 g).
1,3,2′,6′,2‴,6‴-Hexa-N-(tert-butoxycarbonyl)-5″-azido-neomycin (500
mg, 0.4 mmol) was dissolved in ethanol (5 mL) and mixed with Pd/C
(10 wt % on activated carbon; 250 mg). The mixture was stirred under a
hydrogen atmosphere for 12 h at RT. Pd/C was removed by filtration,
and the filtrate was concentrated by vacuum. The crude product was
then purified by column chromatography using CH2Cl2/CH3OH (v/v,
100/5) as an eluent. Compound 1 was obtained as white solid in 90%
yield (440 mg).
Compound 2. To a DMF solution (10 mL) of compound 1 (1.00 g,
0.82 mmol), glutaric anhydride (86.1 mg, 0.86 mmol) and triethylamine
(TEA) (120 μL, 0.86 mmol) were added dropwise. The resulting
solution was stirred at RT for 30 min before dilution with 50 mL of
EtOAc. The EtOAc layer was washed with 0.01 M HCl solution (50 mL)
followed by washing with water (100 mL × 3) and brine (50 mL) and
then was dried with anhydrous Na2SO4. The crude product was stirred
in ethanol (10 mL) and a K2CO3 (1.0 g) mixture for 4 h and diluted with
50 mL of EtOAc. The EtOAc layer was washed with 0.01 M HCl
solution (50 mL) followed by washing with water (100 mL × 3) and
brine (50 mL) and then dried over anhydrous Na2SO4. The solution was
concentrated in a vacuum affording compound 2 as a light yellow solid in
95% yield (1.1 g).
Deprotection of Compound 2. To 500 μL of TFA/water (v/v 90/
10) solution, 30 mg (22.6 mmol) of compound 2 was added. The
resulting solution was decanted into 10 mL of ice-cold diethyl ether after
stirring at RT for 2 h. White precipitate formed in ether was collected by
centrifugation and dried by a vacuum to afford 30 mg of a light yellow
solid. Yield = 98%. 1H NMR (500 MHz, D2O): δ 5.88 (d, J = 3.8 Hz,
1H), 5.35 (d, J = 4.2 Hz, 1H), 5.26 (s, 1H), 4.35 (t, J = 4.8 Hz, 1H),
4.29−4.26 (m, 1H), 4.25−4.21 (m, 1H), 4.21−4.17 (m, 2H), 4.07 (t, J =
9.6 Hz, 1H), 3.98 (t, J = 9.7 Hz, 1H), 3.95−3.91 (m, 1H), 3.91−3.87 (m,
1H), 3.79 (d, J = 1.1 Hz, 1H), 3.72−3.65 (m, 2H), 3.56 (s, 1H), 3.54−
The methodology for synthesizing, screening for both
ribosomal binding and bacterial growth inhibition, and RB-BIP
analysis of the data is a rapid, optimal approach that provides a
systematic method for the rapid synthesis and screening of large
aminoglycoside antibiotic libraries. The approach described here
is expected to impact the analysis of modified aminoglycoside
function against established and newly discovered targets, in
addition to assisting in the development of new leads for
targeting bacterial, viral, and fungal pathogens that contain
potential nucleic acid binding sites for these PAs.
METHODS
■
Unless otherwise specified, chemicals were purchased from commer-
cially available sources and used without further purification. NMR
spectra were recorded on a Bruker Avance-500 spectrometer. MS
(MALDI-TOF) spectra were recorded using a Bruker Daltonics/
I
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
3.47 (m, 3H), 3.45−3.28 (m, 6H), 2.48−2.44 (m, 1H), 2.38 (t, J = 7.4
Hz, 2H), 2.29 (t, J = 7.5 Hz, 2H), 1.93−1.80 (m, 3H). 13C NMR (126
MHz, D2O): δ 177.97, 176.60, 163.35, 163.07, 162.79, 162.51, 119.95,
117.63, 115.31, 112.99, 109.32, 95.93, 95.06, 85.08, 81.07, 77.56, 75.25,
73.72, 72.38, 70.43, 70.26, 69.98, 68.18, 67.66, 67.50, 53.32, 50.95, 49.64,
48.65, 40.94, 40.58, 40.09, 34.92, 33.14, 27.97, 20.78. MS (MALDI-
TOF) m/z calcd for C28H54N7O15 [M + H]+: 728.4. Found: 728.3.
General Procedure for the Solid Phase Synthesis of PA
Library. Coupling of Amino Acids. Rink amide MBHA resin (5.08 mg)
with 0.59 mmol/g loading was swollen in DMF (2 mL, 2 × 1 min) with a
constant stream of nitrogen bubbling before Fmoc deprotection. Resin
was deprotected by mixing with 25% piperidine in DMF (2 mL, 2 × 5
min). N-Fmoc protected amino acids (5 equiv) were then coupled to
resin in DMF in the presence of HCTU (5 equiv) and DIPEA (10 equiv)
for 1 h. Resin was washed extensively with DMF (4 × 1 mL) between
reactions, and the reaction was tested for completeness via Kaiser test.
Coupling of Compound 2. After amino acid couplings were
completed, resin was deprotected with 25% piperidine in DMF (2
mL, 2 × 5 min) and then washed extensively with DMF (4 × 1 mL).
Compound 2 (3 equiv) was coupled in NMP in the presence of HCTU
(3 equiv) and DIPEA (5 equiv) for 24 h. After coupling, resin was
washed with DMF (4 × 1 mL) and CH2Cl2 (4 × 1 mL) and then was
dried under room atmosphere for 1 h. Resin was stored under 0 °C
before cleavage.
Side Chain Deprotection and Cleavage of PA from Resin. The
PAs were cleaved from resin, and all protecting groups were removed by
a 2 h treatment with 88:5:5:2 TFA/phenol/H2O/TIPS (v/v; 150 μL).
After the treatment, TFA was evaporated under nitrogen pressure. To
the resulting resin, 600 μL of deionized water was then added to dissolve
the cleaved PAs. The water layer was washed extensively with CH2Cl2
(500 μL) and EtOAc (3 × 500 μL) then lyophilized to afford final
products as white powders. Final products were stored at −20 °C and
protected from light and moisture.
General Procedure for Large-Scale Synthesis of Selected PA.
Coupling of Amino Acids. Rink amide MBHA resin (19.9 mg) with 0.51
mmol/g loading was swollen in DMF (2 mL, 2 × 1 min) with stirring.
Resin was deprotected by stirring with 25% piperidine in DMF (2 mL, 2
× 3 min). N-Fmoc protected amino acids (5 equiv) were then coupled
to resin in DMF in the presence of HCTU (5 equiv) and DIPEA (10
equiv) for 1 h. Resin was washed extensively with and DMF (4 × 1 mL)
between reactions and was tested for completeness via Kaiser test.
Coupling of Compound 2. After amino acid couplings were
completed, resin was deprotected with 25% piperidine in DMF (2 mL, 2
× 5 min) and then washed extensively with DMF (4 × 1 mL). Resin was
mixed with compound 2 (3 equiv) in DMF in the presence of HCTU (3
equiv) and DIPEA (5 equiv) for 24 h. After coupling, resin was washed
with DMF (4 × 1 mL) and CH2Cl2 (4 × 1 mL) and then was dried under
room atmosphere for 1 h. Resin was stored under 0 °C before side chain
deprotection and cleavage.
Deprotection and Cleavage of PA from Resin. The PAs were
cleaved from resin, and all protecting groups were removed by a 2 h
treatment with 88:5:5:2 TFA/phenol/H2O/TIPS (v/v; 500 μL). After
deprotection, resin was removed from reaction mixture by filtration. Ice-
cold diethyl ether (10 mL) was added into the filtrate to afford a white
precipitate. The precipitated product was collected by centrifugation
and washed with diethyl ether (3 × 1 mL) three times. The product was
dried under a rough vacuum and lyophilized for 48 h to completely
remove diethyl ether. The resulting final products were stored at −20 °C
protected from light and moisture.
High-throughput Screening (HTS) of Antibacterial Activity
and Determination of the Minimal Inhibitory Concentrations
(MIC). Antibacterial HTS and the determination of MIC were
performed following a previously reported method.43,55 S. pyogenes, E.
faecalis, and L. monocytogenes were cultured on Mueller Hinton (MH)
agar with 5% lysed horse blood at 35 °C for 20 h and were then collected
with a sterile cotton swab. The other bacteria were cultured in MH broth
in a shaker−incubator at 37 °C for 20 h. The cultures were then diluted
to an absorbance at 595 nm of 0.006 using sterile MH broth (lysed horse
blood was added to S. pyogenes, E. faecalis, and L. monocytogenes strains
after dilution, and the resulting blood concentration is 5%). The diluted
bacteria aliquots were placed (90 μL per well) into 96-well clear flat
bottom plates. For HTS, compounds (10 μL, 50 μM, in H2O solution)
were then added, and the final concentration of the compounds was 5.0
μM. For the determination of MIC, compounds were diluted with water
to give the following concentrations (800, 400, 200, 100, 50.0, 25.0, 12.5,
6.25 μM). The resulting solutions (10 μL) were then added to the 96-
well plates, and the final concentrations of the compounds were 80, 40,
20, 10, 5.0, 2.5, 1.25, and 0.625 μM, respectively. Background controls
(10 μL water and 90 μL MH broth solution), positive controls (a
mixture of water (10 μL) and the diluted bacteria solution (90 μL)), and
known drug controls (kanamycin, neomycin, gentamicin, and
ampicillin) were run in each plate. In all plates, the absorbance at 595
nm was measured after 22 h incubations at 35 °C. The percentage
bacterial growth inhibition was calculated as percentage inhibition =
100−100 × (Atestcompound − Abackgroundcontrol)/(Anegativecontrol
−
Abackgroundcontrol). In cases where treatment appeared to enhance the
growth of the culture, inhibition was scored as zero rather than negative
inhibition. The MIC is determined to the lowest (minimum)
concentration at which no visible growth is present (percentage
inhibition ≥99%).
Ribosomal RNA A-site Binding. The described compound library
was screened using an F-neo competitive binding assay as described in
detail by Watkins et al.40 In short, the displacement assay was performed
in duplicate with each compound at a 0.3 μM concentration to 0.1 μM F-
neo:A-site complex using 100 reads per well at an excitation/emission of
485/535 nm. All experiments were performed in 10 mM MOPSO (pH
7.0), 50 mM NaCl, and 0.4 mM EDTA. The displacements of F-neo by
test compounds were determined by the increase of fluorescence
measured after the addition of the test compound to wells of a 96-well
plate containing F-neo:A-site complex compared to the fluorescence of
the F-neo:A-site complex alone (ΔF). Percent binding was calculated
from the ratio of the change in fluorescence by the addition of the test
compound (ΔFDrug) with the change in fluorescence by the addition of
neomycin (ΔFNeo), using eq 1:
%binding = (ΔFdrug/ΔFneomycin) × 100%
(1)
All dockings were performed as blind dockings using Autodock Vina
1.0.56 Docking was performed using an “exhaustiveness” value of 12. All
other parameters were used as a default. All rotatable bonds within the
ligand were allowed to rotate freely, and the A-site was kept rigid.
Autodock Tools version 1.5.457 was used to convert the ligand and
receptor molecules to the proper file formats for AutoDock Vina
docking. All structures were visualized and images made using Pymol.58
The concentration of compound that displaces 50% of the F-neo
(DC50) of DPA1199 and DPA1308 was determined by the titration of
each compound into 0.1 μM F-neo:A-site complex containing either the
E. coli or human model of the A-site, using 50 reads per well at an
excitation/emission of 485/535 nm. All experiments were performed in
duplicate in 10 mM MOPSO (pH 7.0), 50 mM NaCl, and 0.4 mM
EDTA. The initial concentration range was 0.05 nM to 1 μM of each
compound. The concentration was extended to 20 μM of each
compound for the human A-site to obtain the DC50 due to a lack of
saturation at the lower range. The DC50 was determined from the
sigmoidal fit using Origin 5.0.
Cell Free Translation Assays. Working solutions were made of
each compound using filtered nanoPure H2O by serially diluting with
final dose ranges of 1.25 μM to 2 nM for prokaryotic systems. Escherichia
coli (E. coli S30 Extract System for Circular DNA, Promega, L1020)
assays were performed as recommended by the manufacturer.59,60
Briefly, varying concentrations of the compounds of interest were
allowed to incubate in a solution of cellular extract and all amino acids
for 20 min at RT. Following brief centrifugation, 0.40 μL of luciferase
control template (Promega, L492A-C) was added to each tube. After
gentle mixing and spin down, tubes were incubated at 37 °C for 60 min.
Translation was terminated by inactivating on ice for 5 min. Upon
returning to ambient temperature, 5 μL per tube was delivered to a
LUMITRAC 200 96-well plate. Luminescence was immediately read
following the addition of the luciferin solution using a Molecular Devices
SpectraMax M2. IC50 values were determined by nonlinear fit using
J
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
GraphPad Prism 6. All compounds were done in triplicate and
standardized against an internal vehicle control.
(4) Purohit, P., and Stern, S. (1994) Interactions of a small RNA with
antibiotic and RNA ligands of the 30S subunit. Nature 370, 659−662.
(5) Fourmy, D., Recht, M. I., Blanchard, S. C., and Puglisi, J. D. (1996)
Structure of the A site of Escherichia coli 16S ribosomal RNA complexed
with an aminoglycoside antibiotic. Science 274, 1367−1371.
(6) Tor, Y. (2006) The ribosomal A-site as an inspiration for the design
of RNA binders. Biochimie 88, 1045−1051.
(7) Chambers, H. F., and DeLeo, F. R. (2009) Waves of resistance:
Staphylococcus aureus in the antibiotic era. Nat. Rev. Microbiol. 7, 629−
641.
Aminoglycoside-modifying Enzyme Activity. To determine if
the PA compounds are enzymatically modified by aminglycoside-
modifying enzymes, previously developed assays were employed to
monitor the possible inactivation of these compounds.47−49 All
reactions were monitored at 25 °C (AAC(2′)-Ic, AAC(3)-IV, Eis, and
APH(2″)) or 37 °C (AAC(6′)-Ie/APH(2″)-Ia) on a SpectraMax M5
microplate reader and performed in triplicate. All rates were normalized
to neomycin (NEO).
Acetylation. The activity of several aminoglycoside acetyltransferases
(AAC(6′)-Ie/APH(2″)-Ia, AAC(3)-IV, AAC(2′)-Ic, and Eis) was
monitored using Ellman’s method, coupling the release of CoASH
with DTNB (ε = 14150 M−1cm−1) and monitored at 412 nm. Briefly,
reactions (200 μL) contained PAs (100 μM), AcCoA (500 μM for Eis
and 150 μM for all other acetyltransferases) and DTNB (2 mM) and
were inititated with enzymes (0.125 μM AAC(3)-IV and AAC(2′)-Ic,
0.5 μM for all remaining acetyltransferases) in a corresponding buffer
(100 mM sodium phosphate pH 7.4 for AAC(2′)-Ic, 50 mM MES pH
6.6 for AAC(6′)-Ie/APH(2″)-Ia and AAC(3)-IV, and 50 mM Tris pH
8.0 for Eis). Reactions were monitored taking measurements every 30 s
for 30 min. Initial rates were calculated using the first 2−5 min of the
reaction.
(8) Bera, S., Zhanel, G. G., and Schweizer, F. (2010) Evaluation of
amphiphilic aminoglycoside-peptide triazole conjugates as antibacterial
agents. Bioorg. Med. Chem. Lett. 20, 3031−3035.
(9) Bera, S., Zhanel, G. G., and Schweizer, F. (2010) Antibacterial
activities of aminoglycoside antibiotic-derived cationic amphiphiles:
Polyol-modified neomycin B-, kanamycin A-, amikacin-, and neamine-
based amphiphiles with potent broad spectrum antibacterial activity. J.
Med. Chem. 53, 3626−3631.
(10) Findlay, B., Zhanel, G. G., and Schweizer, F. (2012) Neomycin-
phenolic conjugates: polycationic amphiphiles with broad-spectrum
antibacterial activity, low hemolytic activity and weak serum protein
binding. Bioorg. Med. Chem. Lett. 22, 1499−1503.
Phosphorylation. The phosphorylation of PAs by APH(2″)-Ia was
monitored using the absorbance of NADH at 340 nM (ε = 6220 M−1
cm−1) by the enzyme-coupled response to the production of GDP.
Reactions (200 μL) contained PA (100 μM), Tris-HCl (50 mM, pH
8.0), MgCl2 (10 mM), KCl (40 mM), NADH (0.5 mg mL−1), PEP (2.5
mM), GTP (2 mM), PK/LDH (4 μL), and reactions were initiated by
the addition of APH(2″) (1 μM). Reactions were monitored every 30 s
for 30 min, and initial rates were determined using the first 2−5 min of
reaction time.
(11) Zhang, J., Chiang, F. I., Wu, L., Czyryca, P. G., Li, D., and Chang,
C. T. (2008) Surprising Alteration of Antibacterial Activity of 5″-
Modified Neomycin against Resistant Bacteria. J. Med. Chem. 51, 7563−
7573.
(12) Fridman, M., Belakhov, V., Yaron, S., and Baasov, T. (2003) A
new class of branched aminoglycosides: pseudo-pentasaccharide
derivatives of neomycin B. Org. Lett. 5, 3575−3578.
(13) Hainrichson, M., Pokrovskaya, V., Shallom-Shezifi, D., Fridman,
M., Belakhov, V., Shachar, D., Yaron, S., and Baasov, T. (2005) Branched
aminoglycosides: biochemical studies and antibacterial activity of
neomycin B derivatives. Biorg. Med. Chem. 13, 5797−5807.
(14) Sitaram, N., and Nagaraj, R. (1999) Interaction of antimicrobial
peptides with biological and model membranes: structural and charge
requirements for activity. Biochim. Biophys. Acta 1462, 29−54.
(15) Dathe, M., and Wieprecht, T. (1999) Structural features of helical
antimicrobial peptides: their potential to modulate activity on model
membranes and biological cells. Biochim. Biophys. Acta 1462, 71−87.
(16) Hancock, R. E. W., and Rozek, A. (2002) Role of membranes in
the activities of antimicrobial cationic peptides. FEMS Microbiol. Lett.
206, 143−149.
(17) Fosso, M. Y., Li, Y., and Garneau-Tsodikova, S. (2014) New
trends in aminoglycosides use. MedChemComm 5, 1075−1091.
(18) Kumar, S., Spano, M. N., and Arya, D. P. (2013) Shape readout of
AT rich DNA by carbohydrates. Biopolymers 101, 720−732.
(19) Hamilton, P. L., and Arya, D. P. (2012) Natural product DNA
major groove binders. Nat. Prod. Rep. 29, 134−143.
ASSOCIATED CONTENT
* Supporting Information
■
S
High-throughput bacterial screening results, selected com-
pounds NMR characterizations, NMR spectrum, and mass
spectrometry results are available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Telephone: 1-864-656-1106. Fax: 1-864-656-6613. E-mail:
dparya@clemson.edu.
■
Notes
The authors declare the following competing financial
interest(s): DPA has ownership interest in NUBAD LLC.
(20) Xi, H., Davis, E., Ranjan, N., Xue, L., Hyde-Volpe, D., and Arya, D.
P. (2011) Thermodynamics of nucleic Acid ″shape readout″ by an
aminosugar. Biochemistry 50, 9088−9113.
(21) Willis, B., and Arya, D. P. (2006) An expanding view of
aminoglycoside-nucleic acid recognition. Adv. Carbohydr. Chem.
Biochem. 60, 251−302.
(22) Bera, S., Zhanel, G. G., and Schweizer, F. (2008) Design,
synthesis, and antibacterial activities of neomycin-lipid conjugates:
polycationic lipids with potent Gram-positive activity. J. Med. Chem. 51,
6160−6164.
(23) Bera, S., Zhanel, G. G., and Schweizer, F. (2010) Antibacterial
activity of guanidinylated neomycin B- and kanamycin A-derived
amphiphilic lipid conjugates. J. Antimicrob. Chemother. 65, 1224−1227.
(24) Bera, S., Zhanel, G. G., and Schweizer, F. (2011) Synthesis and
antibacterial activity of amphiphilic lysine-ligated neomycin B
conjugates. Carbohydr. Res. 346, 560−568.
(25) Kumar, S., and Arya, D. P. (2011) Recognition of HIV TAR RNA
by triazole linked neomycin dimers. Bioorg. Med. Chem. Lett. 21, 4788−
4792.
ACKNOWLEDGMENTS
■
This research was supported by National Institute of Health
grants (1R41GM100607, 2R42GM100607) to D. P. Arya and
(1R01AI090048) to S. Garneau-Tsodikova, and support from
NASA astrobiology institute NNA09DA78A to A. K. Oyelere.
We thank T. McNealy (Clemson University; Clemson, South
Carolina) for the S. aureus (ATCC 25923) and S. pyogenes
(ATCC 12384) strains.
REFERENCES
■
(1) Houghton, J. L., Green, K. D., Chen, W., and Garneau-Tsodikova,
S. (2010) The Future of Aminoglycosides: the End or Renaissance?
ChemBioChem. 11, 880−902.
(2) Yang, L., and Ye, X. (2010) Development of aminoglycoside
antibiotics effective against resistant bacterial strains. Curr. Top. Med.
Chem. 10, 1898−1826.
(3) Moazed, D., and Noller, H. F. (1987) Interaction of antibiotics with
functional sites in 16S ribosomal RNA. Nature 327, 389−394.
K
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(26) Xue, L., Ranjan, N., and Arya, D. P. (2011) Synthesis and
spectroscopic studies of the (neomycin)-perylene conjugate binding to
human telomeric DNA. Biochemistry 50, 2838−2849.
(27) Willis, B., and Arya, D. P. (2010) Triple recognition of B-DNA by
a neomycin-Hoechst 33258-pyrene conjugate. Biochemistry 49, 452−
469.
cause of drug resistance in XDR-TB. Proc. Natl. Acad. Sci. U. S. A., Early
Ed. 1−5, 5.
(48) Green, K. D., Chen, W., Houghton, J. L., Fridman, M., and
Garneau-Tsodikova, S. (2010) Exploring the substrate promiscuity of
drug-modifying enzymes for the chemoenzymatic generation of N-
acylated aminoglycosides. ChemBioChem 11, 119−126.
(49) Green, K. D., Chen, W., and Garneau-Tsodikova, S. (2011)
Effects of altering aminoglycoside structures on bacterial resistance
enzyme activities. Antimicrob. Agents Chemother. 55, 3207−3213.
(50) Daigle, D. M., Hughes, D. W., and Wright, G. D. (1999)
Prodigious substrate specificity of AAC(6′)-APH(2″), an aminoglyco-
side antibiotic resistance determinant in enterococci and staphylococci.
Chem. Biol. 6, 99−110.
(51) Ogle, J. M., Brodersen, D. E., Clemons, W. M., Jr., Tarry, M. J.,
Carter, A. P., and Ramakrishnan, V. (2001) Recognition of cognate
transfer RNA by the 30s ribosomal subunit. Science 292, 897−902.
(52) Lynch, S. R., and Puglisi, J. D. (2001) Structural Origins of
Aminoglycoside Specificity for Prokaryotic Ribosomes. J. Mol. Biol. 306,
1037−1058.
(53) Barbieri, C. M., Kaul, M., Bozza-Hingos, M., Zhao, F., Tor, Y.,
Hermann, T., and Pilch, D. S. (2007) Defining the molecular forces that
determine the impact of neomycin on bacterial protein synthesis:
importance of the 2′-amino functionality. Antimicrob. Agents Chemother.
51, 1760−1769.
(54) Barbieri, C. M., Kaul, M., and Pilch, D. S. (2007) Use of 2-
aminopurine as a fluorescent tool for characterizing antibiotic
recognition of the bacterial rRNA A-site. Tetrahedron 63, 3567−6574.
(55) Cudic, M., Lockatell, C. V., Johnson, D. E., and L, O., Jr (2003) In
vitro and in vivo activity of an antibacterial peptide analog against
uropathogens. Peptides 24, 807−820.
(56) Trott, O., and Olson, A. J. (2010) AutoDock Vina: Improving the
Speed and Accuracy of Docking with a New Scoring Function, Efficient
Optimization, and Multithreading. J. Comput. Chem. 31, 455−461.
(57) Sanner, M. F. (1999) Python: a programming language for
software integration and development. J. Mol. Graphics & Modelling 17,
57−61.
(58) Brunger, A., Adams, P., Clore, G., DeLano, W., Gros, P., Grosse-
Kunstleve, R., Jiang, J., Kuszewski, J., Nilges, M., Pannu, N., Read, R.,
Rice, L., Simonson, T., and Warren, G. (1998) Crystallography & NMR
system: A new software suite for macromolecular structure determi-
nation. Acta Crystallogr., Sect. D: Biol. Crystallogr. 54, 905−921.
(59) Pratt, S.-D., David, C.-A., Black-Schaefer, C., Dandliker, P.-J.,
Xuei, X., Warrior, U., Burns, D.-J., Zhong, P., Cao, Z., Saiki, A.-Y., Lerner,
C.-G., Chovan, L.-E., Soni, N.-B., Nilius, A.-M., Wagenaar, F.-L., Merta,
P.-J., Traphagen, L.-M., and Beutel, B.-A. (2004) A strategy for discovery
of novel broad-spectrum antibacterials using a highthroughput
Streptococcus pneumoniae transcription/translation screen. J. Biomol
Screen. 9, 3−11.
(28) Xue, L., Xi, H., Kumar, S., Gray, D., Davis, E., Hamilton, P., Skriba,
M., and Arya, D. P. (2010) Probing the recognition surface of a DNA
triplex: binding studies with intercalator-neomycin conjugates. Bio-
chemistry 49, 5540−5552.
(29) Shaw, N. N., Xi, H., and Arya, D. P. (2008) Molecular recognition
of a DNA:RNA hybrid: sub-nanomolar binding by a neomycin-
methidium conjugate. Bioorg. Med. Chem. Lett. 18, 4142−4145.
(30) Willis, B., and Arya, D. P. (2006) Recognition of B-DNA by
neomycin-Hoechst 33258 conjugates. Biochemistry 45, 10217−10232.
(31) Charles, I., and Arya, D. P. (2005) Synthesis of neomycin-DNA/
peptide nucleic acid conjugates. J. Carbohydr. Chem. 24, 145−160.
(32) Arya, D. P., Coffee, R. L., Jr., and Xue, L. (2004) From triplex to B-
form duplex stabilization: reversal of target selectivity by aminoglycoside
dimers. Bioorg. Med. Chem. Lett. 14, 4643−4646.
(33) Arya, D. P., and Willis, B. (2003) Reaching into the major groove
of B-DNA: synthesis and nucleic acid binding of a neomycin-hoechst
33258 conjugate. J. Am. Chem. Soc. 125, 12398−12399.
(34) Arya, D. P., Xue, L., and Tennant, P. (2003) Combining the best
in triplex recognition: synthesis and nucleic acid binding of a BQQ-
neomycin conjugate. J. Am. Chem. Soc. 125, 8070−8071.
(35) Coin, I., Beyermann, M., and Bienert, M. (2007) Solid-phase
peptide synthesis: from standard procedures to the synthesis of difficult
sequences. Nat. Protoc. 2, 3247−3256.
(36) Michael, K., Wang, H., and Tor, Y. (1999) Enhanced RNA
binding of dimerized aminoglycosides. Biorg. Med. Chem. 7, 1361−1371.
(37) Boer, J., Blount, K. F., Luedtke, N. W., Elson-Schwab, L., and Tor,
Y. (2005) RNA-selective modification by a platinum(II) complex
conjugated to amino- and guanidinoglycosides. Angew. Chem., Int. Ed.
44, 927−932.
(38) Quader, S., Boyd, S. E., Jenkins, I. D., and Houston, T. A. (2007)
Multi-site modification of neomycin B: combined mitsunobu and click
chemistry approach. J. Org. Chem. 72, 1962−1979.
(39) Kumar, S., Kellish, P., Robinson, W. E., Wang, D., Appella, D. H.,
and Arya, D. P. (2012) Click dimers to target HIV TAR RNA
conformation. Biochemistry 51, 2331−2347.
(40) Watkins, D., Norris, F. A., Kumar, S., and Arya, D. P. (2013) A
fluorescence-based screen for ribosome binding antibiotics. Anal.
Biochem. 434, 300−307.
(41) Zhao, F., Zhao, Q., Blount, K., Han, Q., Tor, Y., and Hermann, T.
(2005) Molecular recognition of RNA by neomycin and a restricted
neomycin derivative. Angew. Chem., Int. Ed. 44, 5329−5334.
(42) Ryu, D., and Rando, R. (2001) Aminoglycoside binding to human
and bacterial A-site rRNA decoding region constructs. Bioorg. Med.
Chem. 9, 2601−2608.
(43) De La Fuente, R., Sonawane, N. D., Arumainayagam, D., and
Verkman, A. S. (2006) Small molecules with antimicrobial activity
against E. coli and P. aeruginosa identified by high-throughput
screening. Br. J. Pharmacol. 149, 551−559.
(60) Thorne, C.-A., Lafleur, B., Lewis, M., Hanson, A.-J., Jernigan, K.-
K., Weaver, D.-C., Huppert, K.-A., Chen, T.-W., Wichaidit, C., Cselenyi,
C.-S., Tahinci, E., Meyers, K.-C., Waskow, E., Orton, D., Salic, A., Lee,
L.-A., Robbins, D.-J., Huppert, S.-S., and Lee, E. (2011) A biochemical
screen for identification of small-molecule regulators of the Wnt
pathway using Xenopus egg extracts. J. Biomol. Scr. 16, 995−1006.
(44) King, A., Watkins, D., Kumar, S., Ranjan, N., Gong, C., Whitlock,
J., and Arya, D. P. (2013) Characterization of ribosomal binding and
antibacterial activities using two orthogonal high-throughput screens.
Antimicrob. Agents Chemother. 57, 4717−4726.
(45) Shaul, P., Green, K. D., Rutenberg, R., Kramer, M., Berkov-Zrihen,
Y., Breiner-Goldstein, E., Garneau-Tsodikova, S., and Fridman, M.
(2011) Assessment of 6′- and 6‴-N-acylation of aminoglycosides as a
strategy to overcome bacterial resistance. Org. Biomol. Chem. 9, 4057−
4063.
(46) Sucheck, S. J., Wong, A. L., Koeller, K. M., Boehr, D. D., Draker, K.
-., Sears, P., Wright, G. D., and Wong, C. (2000) Design of Bifunctional
Antibiotics that Target Bacterial rRNA and Inhibit Resistance-Causing
Enzymes. J. Am. Chem. Soc. 122, 5230−5231.
(47) Chen, W., Biswas, T., Porter, V. R., Tsodikov, O. V., and Garneau-
Tsodikova, S. (2011) Unusual regioversatility of acetyltransferase Eis, a
L
DOI: 10.1021/cb5010367
ACS Chem. Biol. XXXX, XXX, XXX−XXX