Sugar-grafted cyclodextrin nanocarrier as a "trojan Horse" for potentiating antibiotic activity

Article abstract of DOI:10.1007/s11095-016-1861-0

Purpose: The use of "Trojan Horse" nanocarriers for antibiotics to enhance the activity of antibiotics against susceptible and resistant bacteria is investigated. Methods: Antibiotic carriers (CD-MAN and CD-GLU) are prepared from β-cyclodextrin grafted with sugar molecules (D-mannose and D-glucose, respectively) via azide-alkyne click reaction. The sugar molecules serve as a chemoattractant enticing the bacteria to take in higher amounts of the antibiotic, resulting in rapid killing of the bacteria. Results: Three types of hydrophobic antibiotics, erythromycin, rifampicin and ciprofloxacin, are used as model drugs and loaded into the carriers. The minimum inhibitory concentration of the antibiotics in the CD-MAN-antibiotic and CD-GLU-antibiotic complexes for Gram-negative Escherichia coli, Pseudomonas aeruginosa and Acinetobacter baumannii strains, and a number of Gram-positive Staphylococcus aureus strains, including the methicillin-resistant strains (MRSA), are reduced by a factor ranging from 3 to >100. The CD-MAN-antibiotic complex is also able to prolong the stability of the loaded antibiotic and inhibit development of intrinsic antibiotic resistance in the bacteria. Conclusions: These non-cytotoxic sugar-modfied nanocarriers can potentiate the activity of existing antibiotics, especially against multidrug-resistant bacteria, which is highly advantageous in view of the paucity of new antibiotics in the pipeline.

Full text of DOI:10.1007/s11095-016-1861-0

Pharm Res
DOI 10.1007/s11095-016-1861-0
RESEARCH PAPER
Sugar-Grafted Cyclodextrin Nanocarrier as a BTrojan Horse^
for Potentiating Antibiotic Activity
Min Li1 & Koon Gee Neoh1 & Liqun Xu1 & Liang Yuan1 & David Tai Leong1 & En-Tang Kang1 & Kim Lee Chua2 & Li Yang Hsu3
Received: 12 August 2015 /Accepted: 13 January 2016
#
Springer Science+Business Media New York 2016
Conclusions These non-cytotoxic sugar-modfied
nanocarriers can potentiate the activity of existing antibiotics,
especially against multidrug-resistant bacteria, which is highly
advantageous in view of the paucity of new antibiotics in the
pipeline.
ABSTRACT
Purpose The use of BTrojan Horse^ nanocarriers for antibi-
otics to enhance the activity of antibiotics against susceptible
and resistant bacteria is investigated.
Methods Antibiotic carriers (CD-MAN and CD-GLU) are
prepared from β-cyclodextrin grafted with sugar molecules
(D-mannose and D-glucose, respectively) via azide-alkyne
click reaction. The sugar molecules serve as a chemoattractant
enticing the bacteria to take in higher amounts of the antibi-
otic, resulting in rapid killing of the bacteria.
.
.
.
KEY WORDS antibacterial antibioticnanocarrier glucose
.
mannose β-cyclodextrin
Results Three types of hydrophobic antibiotics, erythromy-
cin, rifampicin and ciprofloxacin, are used as model drugs
and loaded into the carriers. The minimum inhibitory con-
centration of the antibiotics in the CD-MAN-antibiotic and
CD-GLU-antibiotic complexes for Gram-negative Escherichia
coli, Pseudomonas aeruginosa and Acinetobacter baumannii strains,
and a number of Gram-positive Staphylococcus aureus strains,
including the methicillin-resistant strains (MRSA), are re-
duced by a factor ranging from 3 to >100. The CD-MAN-
antibiotic complex is also able to prolong the stability of the
loaded antibiotic and inhibit development of intrinsic antibi-
otic resistance in the bacteria.
ABBREVIATIONS
A. baumannii Acinetobacter baumannii
CD-GLU
CD-MAN
CIP
D-glucose-grafted cyclodextrin (CD)
D-mannose-grafted cyclodextrin (CD)
Ciprofloxacin
E. coli
ERY
Escherichia coli
Erythromycin
MIC
MTT
Minimum inhibitory concentration
3-[4,5-Dimethyl-thiazol-2-yl]-2,
5-diphenyltetrazolium bromide
P. aeruginosa Pseudomonas aeruginosa
PBI
3,4,9,10-perylenetetracarboxylic
3,4:9,10-dianhydride
Rifampicin
Staphylococcus aureus
RIF
S. aureus
Electronic supplementary material The online version of this article
(doi:10.1007/s11095-016-1861-0) contains supplementary material, which is
available to authorized users.
INTRODUCTION
* Koon Gee Neoh
chenkg@nus.edu.sg
Antibiotics have been of critical importance in the fight
against infectious diseases caused by bacteria since 1920s.
However, the extensive use of antibiotics has given rise to
increasing frequency of antibiotic resistance due to bacterial
evolution from mutation and horizontal transfer of genetic
materials from other resistant bacteria (1). Currently, about
70% of bacteria that cause infections in hospitals are resistant
to at least one of the widely used antibiotics (2). These resistant
1
Department of Chemical and Biomolecular Engineering, National
University of Singapore, Kent Ridge, Singapore 117585, Singapore
2
Department of Biochemistry, National University of Singapore, Kent
Ridge, Singapore 117543, Singapore
3
Department of Medicine, National University of Singapore, Kent Ridge
Singapore 119228, Singapore

Li et al.
bacteria are a growing serious public health concern and have
resulted in increased economic burden of infection control.
For example, methicillin-resistant Staphylococcus aureus
(MRSA) strains-induced infections resulted in~$9.7 billion
extra medical cost in United States in 2005 (3). A recent re-
view has bleakly predicted that the antibiotic resistance prob-
lem will continue into the foreseeable future, and even if an-
tibiotic use could be reduced, resistant clones would remain
persistent (4). In addition, the antibiotic pipeline is running
almost dry, but the number of new synthetic antibiotics annu-
ally approved by Infectious Diseases Society of America
(IDSA) for marketing continues to decrease since 1983, and
there were only two new approved antibiotics from 2009 to
2013 (5). Thus, it would be highly advantageous if existing
antibiotics can be modified to be more effective against resis-
tant bacteria.
The use of antibiotics in combination with nanoparticulate
systems for combating resistant bacteria has gained increasing
interest in past decades. This strategy capitalizes on the unique
physiochemical properties of nanoparticles and their interac-
tions with biological systems. Metal nanoparticles with intrin-
sic bactericidal properties, such as gold, silver, iron oxide,
magnesium oxide and zinc oxide, have been used as antibiotic
delivery systems (6). A typical example is gold nanoparticles
combined with erythromycin, which reduced the minimum
inhibitory concentration (MIC) of erythromycin against
Yersinia enterocolitica and Escherichia coli (E. coli) by about a factor
of 8 as compared to erythromycin alone (7). The gold nano-
particles were shown to breach the bacterial cell wall thereby
increasing the uptake of the antibiotic. Organic nanoparticles,
including polymeric nanoparticles, DNA based nanostruc-
tures, liposomes and dendrimers, have also been used for load-
ing antimicrobial agents through physical adsorption, encap-
sulation, or chemical conjugation (8,9). For example,
carboxymethyl chitosan (CMCS) nanoparticles were used to
deliver ciprofloxacin into bacterial cells by altering bacterial
cell membrane and the MIC of CMCS nanoparticle-loaded
antibiotic on E. coli was reduced by a factor of 2 as compared
to the free form (10). A number of prevailing strategies using
nanoparticles in combination with antibiotics to enhance the
efficacy of the antibiotics depend on the ability of the nano-
particles to disrupt the bacterial cell membrane. However, a
recent article has shown that nanoparticles which damage
bacterial cell membranes promote the horizontal conjugative
transfer of multidrug-resistance genes and may increase anti-
biotic resistance (11). Therapeutic strategies based on
nanoparticle-antibiotic complexes that can inhibit bacterial
infection as well as prevent development of bacterial resistance
are still lacking.
hydrophobic inner cavity, which can be used for loading a
variety of hydrophobic agents via hydrophobic association to
form inclusion complexes (host-guest complexes) (13). This
inclusion complex leads to advantageous changes to the phys-
iochemical properties of guest molecules, such as improve-
ment of solubility, modification of chemical activity and pro-
tection against degradation by enzymatic substances. As a
result of these unique physiochemical properties, CDs have
been widely used in many applications, including cosmetics,
personal care, food, pharmaceutics and chemical industries.
In addition, CDs are also appealing as cores for the design of
multivalent glycoconjugates due to the selective functionality
of their hydroxy moieties. For example, n-heptyl α-D-
mannoside, a nanomolar FimH antagonist, was tethered on
β-CD to develop heptavalent β-CD to capture and aggregate
living bacteria and reduce bacterial adhesion in mouse blad-
der (14). In this work, β-CD, which is the most widely-used
CD in its family because of its availability, lowest price and
acceptance by USFDA as generally regarded as safe (GRAS)
for use in food (15), was selected as an antibiotic carrier. The
β-CD was grafted with sugar molecules, D-mannose or D-
glucose, to form the CD-MAN and CD-GLU carriers, respec-
tively. We hypothesize that since mannose and glucose are
carbon sources for bacteria and can readily permeate into
bacterial cells through sugar transporters on the cell mem-
brane (16), the grafted sugar moieties on the CD carrier will
serve as chemoattractant for the bacteria. The subsequent
interactions between the bacteria and the CD carrier may
enhance the uptake of the antibiotic and rapidly increases its
cytoplasmic concentration and efficacy. Furthermore, the car-
rier may also protect the antibiotic from hydrolysis and deg-
radation and preserve its activity for a longer period than its
free form. The antibacterial efficacy of CD-MAN and CD-
GLU loaded antibiotics (CD-MAN-antibiotic and CD-GLU-
antibiotic complexes) was evaluated with Staphylococcus aureus
(S. aureus), E. coli, Pseudomonas aeruginosa (P. aeruginosa) and
Acinetobacter baumannii (A. baumannii) bacteria. The possibility
of the bacteria developing resistance to the sugar-modified
CD-antibiotic complexes and cytotoxicity of the carrier-
complexes were also evaluated.
MATERIALS AND METHODS
Materials
Beta-cyclodextrin (β-CD, 97%), triphenylphosphine (Ph3P,
98.5%), iodine (99.8%), sodium azide (99.5%), copper(I) bro-
m i d e ( C u B r , 9 8 % ) , N , N , N ’ , N ^, N ^-
pentamethyldiethylenetriamine (PMDETA, 99%), D-
mannose (99%) and D-glucose (99.5%) were purchased from
Sigma–Aldrich (St. Louis, MO, USA). 3-[4,5-Dimethyl-
thiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) was
Cyclodextrins (CDs) are cyclic oligosaccharides containing
six or more glucopyranose units linked via α-(1,4) bonds (12).
A CD molecule is considered as a nanoparticle (outer diame-
ter from 1.46 nm onwards) with hydrophilic outer wall and

Trojan Horse Antibiotic Nanocarriers
obtained from Alfa Aesar Co. (Ward Hill, MA, USA).
Spectra-Por dialysis membranes (molecular weight cut-off:
1000) were purchased from Spectrum Laboratories, Inc.
(Rancho Dominguez, CA, USA). E. coli DH5α, S. aureus
ATCC 25923, Human HEK 293T embryonic kidney cells,
murine RAW 264.7 macrophage and murine 3T3 fibroblast
cells were purchased from American Type Culture Collection
(Manassas, VA, USA). Human NCM460 colonic epithelial
cells were purchased from INCELL Corporation (San
Antonio, TX, USA). P. aeruginosa PAO1 was purchased from
National Collection of Industrial Food and Marine Bacteria
(NCIMB, Bucksburn, Aberdeen, Scotland). P. aeruginosa
PAO397 was obtained by blocking the multidrug efflux
pumps of P. aeruginosa PAO1 as reported in an earlier publi-
cation (17). The MRSA strains, DM23605 and DR9369, were
clinical isolates provided by the Singapore General Hospital
(SGH, Singapore), while the MRSA strain ATCC BAA-44
was purchased from American Type Culture Collection.
A. baumannii S1, a clinical isolate, was provided by the
Network for Antimicrobial Resistance Surveillance
(Singapore). 1-(2′-Propargyl)-D-mannose (mannose-alkyne)
and 1-(2′-propargyl)-D-glucose (glucose-alkyne) were synthe-
sized using procedures similar to that described in the litera-
ture (18).
1.0 mmol), mannose-alkyne (1.64 g, 7.5 mmol) and PMDETA
(156.6 μL, 130 mg, 0.75 mmol) were dissolved in DMF
(15 mL) in a 50 mL flask. The solution was degassed by purg-
ing with argon for 30 min and CuBr (107.6 mg, 0.75 mmol)
were then added. The reaction mixture was further purged
with argon for 10 min. The flask was then sealed tightly and
the reaction was allowed to proceed under continuous stirring
at 60°C for 24 h. After reaction, the solution was poured into
excess ethyl ether. The precipitate was collected after filtration
and re-dissolved in doubly distilled water. The aqueous solu-
tion was then subjected to dialysis against doubly distilled wa-
ter for 3 days followed by freeze-drying to obtain the final CD-
MAN product. CD-GLU was prepared using similar proce-
dures with glucose-alkyne instead of mannose-alkyne.
Preparation of CD-MAN or CD-GLU Loaded Antibiotic
(CD-MAN-antibiotic or CD-GLU-antibiotic) Complex
Three different antibiotics, erythromycin (ERY, inhibitor of
protein synthesis), rifampicin (RIF, inhibitor of transcription)
and ciprofloxacin (CIP, inhibitor of DNA replication), were
used to prepare CD-MAN-antibiotic complexes. Typically,
the antibiotic (0.01 mmol) and CD-MAN (0.01 mmol) were
mixed and triturated with ethanol-water mixture (200 μL,
ethanol/water=100 μL/100 μL). The mixture was kneaded
for 45 min and then dissolved in doubly distilled water
(20 mL). The solution was passed through a 0.2 μm Nylon
membrane filter followed by freeze-drying to obtain the CD-
MAN-antibiotic complex. CD-antibiotic and CD-GLU-
antibiotic complexes were prepared using similar procedures
with β-CD and CD-GLU, respectively, instead of CD-MAN.
The content of antibiotics in CD-MAN-antibiotic com-
plexes was determined as follows: CD-MAN-antibiotic
(1 mg) was suspended in methanol (10 mL) and subjected to
ultrasonication for 30 min. The suspension was kept overnight
for the loaded antibiotic to be released from the complex.
After the precipitation of the undissolved CD-MAN by cen-
trifugation at 8000 rpm for 10 min, the concentration of the
antibiotic in the supernatant was analyzed to calculate the
content of antibiotics in the complexes. The concentration of
ERY was analyzed by high-performance liquid chromatogra-
phy (HPLC). HPLC assay was performed with an Agilent
Technologies 1200 series liquid chromatographic system at a
45°C (oven temperature). The mobile phase was 15 wt%
methanol, 45 wt% acetonitrile and 40 wt% ammonium phos-
phate 0.01 M solution, and the flow rate was 1 mL/min. The
injection volume was 50 μL and the effluent was monitored at
215 nm. For the HPLC analysis of CIP, the mobile phase was
5 vol% methanol, 5 vol% acetonitrile and 90 vol% acetic acid
aqueous solution (50 mL/L). The injection volume was 50 μL
and the effluent was monitored at 280 nm. The oven temper-
ature was set at 50°C and flow rate was 1 mL/min. The
concentration of RIF was analyzed by the colorimetric
Synthesis of D-mannose- or D-glucose-grafted β-CD
(CD-MAN or CD-GLU)
Heptakis(6-deoxy-6-azido)-β-cyclodextrin (CD-N3) was pre-
pared using procedures described in an earlier publication
(19). Briefly, Ph3P (40.1 g, 153 mmol) was dissolved in dry
N,N-dimethylformamide (DMF, 160 mL) and iodine (40.5 g,
160 mmol) was added to this solution over 10 min. The solu-
tion was then heated to 70°C and dry β-CD (11.6 g,
10.2 mmol) was added. The reaction was allowed to proceed
at 70°C for 24 h under N2 protection. After completion of the
reaction, the solution was partially concentrated under re-
duced pressure followed by addition of sodium methoxide
(3 M, 60 mL). The reaction mixture was then poured into
excess methanol. The precipitate was collected by filtration
and purified by Soxhlet extraction with methanol for 72 h
followed by drying under reduced pressure to obtain
heptakis(6-deoxy-6-iodo)-β-cyclodextrin (CD-I). Then, CD-I
(2.99 g, 1.57 mmol) was dissolved in dry DMF (50 mL) and
sodium azide (1 g, 15.4 mmol) was added. The reaction sus-
pension was heated to 70°C and kept at this temperature for
36 h under N2 atmosphere. After reaction, DMF was partially
removed and the reaction suspension was poured into excess
doubly distilled water. The resultant precipitate was collected
by filtration, washed with water and freeze-dried to obtain the
CD-N3 product.
CD-MAN was synthesized via azide-alkyne click reaction
between CD-N3 and mannose-alkyne. Briefly, CD-N3 (1.33 g,

Li et al.
method described in the literature (20). The content of anti-
biotics in CD-antibiotic and CD-GLU-antibiotic complexes
was analyzed using similar methods as described above.
P. aeruginosa and S. aureus was qualitatively evaluated using fluo-
rescence microscopy. Overnight bacterial culture broth was di-
luted to a concentration of 1 ×108 CFU/mL with PBS. A piece
of medical grade silicone film (1× 1 cm2) was placed in a 24-well
plate and covered by 1 mL of bacterial suspension at 37°C for
4 h to allow bacterial attachment. After the bacterial attachment
process, the film was washed thrice with PBS to remove the non-
adherent bacteria. Then, 1 mL of MHB medium containing
either PBI (10 μg) or CD-MAN-PBI (99 μg, corresponding to
10 μg of PBI) or CD-PBI (49.3 μg, corresponding to 10 μg of
PBI) was added to cover the silicone film with adherent bacterial
cells and incubated at 37°C for 4 h. After incubation, the film
was washed thrice with PBS followed by observation under a
fluorescence microscope.
The accumulation of CD-MAN-PBI and CD-PBI within
the bacterial cells was evaluated in a similar manner as that
described in an earlier report with a few modifications (23).
One mL of bacterial suspension containing 1×105 CFU in
MHB was incubated with PBI (10 μg) or CD-MAN-PBI
(99 μg) or CD-PBI (49.3 μg) for 0.5 and 4 h, respectively.
After incubation, the bacterial cells were collected by
centrifuging at 2700 rpm for 10 min and washed thrice with
PBS. The bacteria were then lysed using 1 mL of bacterial lysis
solution to release the PBI which had accumulated in the
bacterial cells. After bacterial lysis, the suspension (200 μL)
was transferred to the wells of a black microtitre tray
(Corning, Amsterdam, The Netherlands) and fluorescence in-
tensity was measured by a fluorescence microplate reader.
Antimicrobial Susceptibility Assay
In this work, the antibacterial activity of the antibacterial agents
(free antibiotic, CD-MAN-antibiotic, CD-GLU-antibiotic and
CD-antibiotic complexes) was evaluated using MIC assay.
MIC is defined as the lowest concentration of an antimicrobial
agent that will prevent the visible growth of a microbe after
overnight culture (21). MIC is commonly used as a research
tool to evaluate the antimicrobial activity of a new agent in vitro
and used by diagnostic laboratories to determine resistance.
The MIC was determined by a standard turbidimetric method.
Bacteria were cultured overnight in growth medium (nutrient
broth (NB) for E. coli, tryptic soy broth (TSB) for S. aureus and
lysogeny broth (LB) for P. aeruginosa and A. baumannii). The
bacteria-containing growth medium was then centrifuged for
10 min at 2700 rpm to remove the supernatant. The bacterial
cells were washed with phosphate buffer saline (PBS, pH 7.4)
and resuspended in Mueller-Hinton broth (MHB) at a concen-
tration of 2 × 105 colony forming unit (CFU)/mL, as estimated
from the optical density of the suspension at 600 nm (OD600).
OD600 of 0.1 is equivalent to ~108 CFU/mL based on calibra-
tion from spread plate counting. One hundred μL of two-fold
serially diluted antibacterial agent (free antibiotic, CD-antibiot-
ic, CD-MAN-antibiotic or CD-GLU-antibiotic complex) solu-
tion in MHB and 100 μL of the bacterial-containing MHB
were added in each well of a 96-well plate. Initial screening
was carried out with 1024 to 0.125 mg/L of the antibacterial
agent in MHB. If the MIC was determined to be out of this
range, further screening was carried out with the concentration
range adjusted accordingly. The plate was then incubated in an
orbital shaker at 37°C for 20 h with a shaking speed at
200 rpm/min. The MIC of the antibacterial agent was deter-
mined as the minimum concentration at which there was no
visible change in the turbidity of the medium in the well. The
corresponding MIC of the antibiotic in the complex was calcu-
lated from the antibiotic content in the complex. All the tests
were carried out at least in triplicate.
Antibiotic Resistance Assay
Since bacteria can develop resistance to antibiotics, an impor-
tant consideration is whether bacteria will develop resistance
to the sugar-modified CD-antibiotic complexes. An antibiotic
resistance assay using S. aureus ATCC 25923 was carried out
to assess its potential to develop resistance to free ERY, CD-
ERY and CD-MAN-ERY. To encourage S. aureus to develop
resistance to free ERY, TSB medium (1 mL) containing 0.25
times MIC of free ERY was inoculated with overnight culture
(10 μL) of S. aureus. The bacterial culture was passaged every
2 days by inoculating the culture (10 μL) into fresh TSB me-
dium containing 0.25 times MIC of free ERY. At the end of 8
and 16 days of incubation, the bacterial suspension after ap-
propriate dilution was spread onto TSB agar to determine
total bacterial counts. The bacterial suspension was also
spread onto TSB agar containing 2 times MIC of free ERY
to determine the portion of bacteria that had developed resis-
tance to free ERY. Bacterial cultures in TSB medium alone
were also tested to confirm no spontaneous mutations that
might give rise to ERY-resistant bacteria. The portion of bac-
teria that developed resistance to CD-ERY or CD-MAN-
ERY complex was determined using similar procedures with
CD-ERY and CD-MAN-ERY, respectively, instead of free
Bacterial Uptake and Accumulation Assays
A h y d r o p h o b i c f l u o r e s c e n t d y e , 3 , 4 , 9 , 1 0 -
perylenetetracarboxylic 3,4:9,10-dianhydride (PBI), was loaded
in place of antibiotic into CD-MAN and β-CD to form CD-
MAN-PBI and CD-PBI complexes, respectively, using similar
procedures as those described above for CD-MAN-antibiotic or
CD-antibiotic complexes. The resultant CD-MAN-PBI and
CD-PBI contained 10.1% and 20.3% (wt%) of PBI, respective-
ly, as measured by a fluorometric method reported earlier (22).
The uptake of CD-MAN-PBI and CD-PBI by E. coli,

Trojan Horse Antibiotic Nanocarriers
ERY. A schematic illustration of the procedures of the antibi-
otic resistance assay is shown in Fig. S1 (Supplementary
Material). The antibiotic resistance assay was also carried
out using E. coli DH5α to assess its potential to develop resis-
tance to free RIF, CD-RIF and CD-MAN-RIF. The proce-
dures used were similar to those using S. aureus ATCC 25923
but with NB as the medium instead of TSB.
containing the antibacterial agent (free ERY, CD-MAN-
ERY complex or CD-MAN carrier) at different concentra-
tions, and the cells were incubated at 37°C for 24 and 72 h.
Control experiments were carried out without the antibacte-
rial agents. The culture medium in each well was then re-
moved and the medium (90 μL) and MTT solution (10 μL,
5 mg/mL in PBS) were added to each well. After 4 h of
incubation at 37°C, the medium was removed and the
formazan crystals were dissolved with dimethylsulfoxide
(DMSO, 100 μL) for 15 min. The optical absorbance of the
DMSO solution was measured at 560 nm on a BIO-TEK
microplate reader. The cell viability is expressed as a percent-
age relative to that obtained in the control experiment.
For the Tali™ scan analysis, NCM460 and HEK 293T cells
were both seeded on 24 wells plates at a seeding density of
1×105 cells per well. After overnight incubation, the original
media solutions were replaced with CD-MAN-ERY or free
ERY suspension in DMEM at different concentrations. The
treated cells were further incubated for 24 h, washed once
with PBS containing 0.05% trypsin and then incubated for
5 min. Fresh medium was added to neutralize the trypsin.
After flushing the wells several times with fresh medium, the
cell suspensions were pelleted down by centrifugation at
2000 g for 5 min. The supernatant was then removed and the
remaining cell pellet was re-suspended in medium.
Subsequently, chilled propidium iodide (PI, 1 μL, 50 μM)
was added as a dye to the suspension, and the suspension
was vortexed lightly and incubated for 5 min. The mixture
(25 μL) was loaded onto a Tali™ Cellular Analysis Slide (Life
Technologies, Invitrogen, USA) and analyzed with the Tali™
Image Based Cytometer (Invitrogen) for cell viability.
Bacterial Lysis Assay
The possibilty of lysis arising from the interactions of the CD-
MAN nanocarrier with bacterial cell membrane was assessed
from detection of bacterial nucleic acids after incubation with
CD-MAN. Ten mL of PBS suspension of bacteria
(2 × 108 CFU/mL) was mixed with 10 mL of PBS solution
of CD-MAN (0.6 g/L). The bacterial suspension was incubat-
ed at 37°C with 150 rpm of constant shaking. At hourly inter-
vals, 2 mL of the suspension was then collected and filtered
through a 0.2 μm Nylon membrane filter to remove the bac-
terial cells. The absorbance of the filtrate at 260 nm (OD260
)
was measured on a Shimadzu UV-1601 spectrophotometer to
detect any nucleic acids leached from the bacteria. Bacterial
lysis buffer (sodium dodecyl sulfate in PBS, 2.0 wt%) and PBS
were used as positive and negative control, respectively.
Stability Assay
The stability of the antibacterial agents was assessed by
conducting the bacterial growth inhibition assay over a period
of time. Two hundred μL of MHB medium containing
1 × 105 CFU/mL of bacterial cells (E. coli DH5α,
P. aeruginosa PAO1 and S. aureus ATCC 25923) and the anti-
bacterial agent (free ERY or CD-MAN-ERY complex) at
MIC or 2 times MIC were added in a well of a 96-well plate.
The plate was incubated at 37°C for 21 days. The bacterial
culture was replenished by adding fresh MHB every 2 days
and the volume of culture medium in each well was main-
tained at 200 μL. The OD600 of the bacterial culture was
measured on a BIO-TEK microplate reader (Model
Powerwave XS) at regular intervals.
Characterization
Fourier transform infrared (FT-IR) spectra were obtained in
the transmission mode on a Bio-Rad FT-IR spectrophotom-
eter (Model FTS135). UV-visible (UV–vis) absorption spectra
were recorded on a Shimadzu UV-1601 spectrophotometer.
Hydrodynamic diameter (Dh) of CD-MAN, CD-GLU and
their antibiotic complexes in aqueous solution, and zeta po-
tential of these samples in PBS buffer were measured using a
Zetasizer Nano ZS (Malvern Instruments, Southborough,
MA, USA) with a 633 nm He-Ne laser at a scattering angle
of 173°. PBS buffer was used to stabilize the pH and ionic
strength of sample solution during zeta potential measure-
ment. Electrospray ionization mass spectra (ESI-MS) were
obtained on a Bruker MicroTOF-Q system (Bruker
Corporation, Karlsruhe, Germany).
Cytotoxicity Assay
The cytotoxicity of CD-MAN-ERY complex was investigated
using two methods, the MTT assay and Tali™ scan analysis,
which are based on different principles. The MTT assay was
carried out with murine 3T3 fibroblasts and RAW 264.7
macrophages, while the Tali™ scan analysis was carried out
with human NCM460 colonic epithelial and HEK 293T
embryonic kidney cells. 3T3 fibroblasts were seeded in
DMEM growth culture medium at a density of 1×104 cells
per well in a 96-well plate and incubated at 37°C for 24 h.
The growth culture medium was then replaced with one
Statistical Analysis
The results were reported as mean standard deviation (SD).
Statistical assessment was made with Tukey post hoc test using

Li et al.
one-way analysis of variance (ANOVA). Statistical signifi-
cance was accepted at P<0.05.
attributed to the 1,2,3-triazole rings (27) of CD-MAN and
CD-GLU generated by the alkyne-azide click reaction. The
molecular weight of CD-MAN and CD-GLU were analyzed
by ESI-MS. The molecular weight of the as-prepared CD-
MAN and CD-GLU was found to be 2860.1 for both com-
pounds, which is consistent with the theoretical value of 2860
for C105H161O70N21 (M + Na+), confirming the successful
preparation of CD-MAN and CD-GLU.
RESULTS AND DISCUSSION
Characterization of Antibiotic Nanocarriers
The preparation of D-mannose-modified β-CD (CD-MAN)
and D-glucose-modified β-CD (CD-GLU) is illustrated sche-
matically in Scheme 1. CD-MAN (or CD-GLU) was synthe-
sized via the azide-alkyne click reaction between CD-N3 and
mannose-alkyne (or glucose-alkyne). Azide-alkyne click reac-
tion has gained much attention as a promising approach to
synthesize new materials with almost complete conversion of
reagents to a single product since it was developed in 1940s
(24). In recent years, azide-alkyne click reaction has been ex-
tensively used for the synthesis of biomaterials with high con-
version efficiency under mild conditions (25). The chemical
structure of CD-MAN and CD-GLU were analyzed by FT-
IR and UV–vis spectroscopy. In the FT-IR spectrum of β-CD
(Fig. 1a), absorption bands at 3370, 1156 and 1037 cm−1
corresponding to the stretching vibrations of –OH, C-O and
C–O-C groups, respectively, of the β-CD are observed. In the
spectrum of CD-N3 (Fig. 1b), the band at 2108 cm−1 is attrib-
uted to the stretching vibrations of the azide group (26). The
successful conjugation of CD-N3 and mannose-alkyne was
confirmed by the FT-IR spectrum of CD-MAN (Fig. 1c), in
which the characteristic band of the azide group at 2108 cm−1
disappeared after the azide-alkyne click reaction. The success-
ful synthesis of CD-GLU was similarly confirmed by the dis-
appearance of the stretching vibration absorption band of the
azide group in the FT-IR spectrum of CD-GLU (Fig. 1d). In
the UV–vis spectra of CD-MAN, β-CD and D-mannose
(Fig. 2a), no absorption was observed for β-CD and D-
mannose between 200 and 600 nm, while a strong absorption
peak at around 216 nm was found in the spectrum of CD-
MAN. A similar absorption was found in the UV–vis spec-
trum of CD-GLU, which was not present in that of D-glucose
(Fig. 2b). The strong absorption at around 216 nm can be
The most well-known feature of CD is its ability to load
hydrophobic agents into its inner cavity to form solid inclusion
complexes. The effect of inclusion of antibiotics (erythromycin
(ERY), rifampicin (RIF) and ciprofloxacin (CIP)) in the sugar-
grafted CD carriers on the hydrodynamic diameter (Dh) of the
carriers was investigated. The Dh and zeta potential of CD-
MAN and CD-GLU before and after inclusion with antibiotics
are shown in Fig. 3 and Table S1. There was no significant
difference between the zeta potential of CD-MAN, CD-GLU
and their antibiotic complexes. On the other hand, changes
were observed in the Dh. The mean Dh of CD-MAN and
CD-GLU is 3.9 and 3.8 nm, respectively, which is slightly big-
ger than that of unmodified β-CD (1.54 nm) (13). After antibi-
otic inclusion, the mean Dh of CD-MAN-antibiotic (CD-
MAN-CIP, CD-MAN-RIF and CD-MAN-ERY) and CD-
GLU-antibiotic (CD-GLU-CIP and CD-GLU-ERY) com-
plexes increased to 9.8-24.4 nm, depending on the carrier
and antibiotic. The increase in Dh may have resulted from
some degree of aggregation of CD-MAN-antibiotic and CD-
GLU-antibiotic in aqueous medium due to the hydrophobic
nature of the antibiotics (ERY, RIF and CIP), similar to the
aggregation of CD-loaded ERY complex (CD-ERY) reported
in an earlier work (28). While the free antibiotics have low
solubility in aqueous medium (≤1 mg/mL), the CD-MAN-
antibiotic and CD-GLU-antibiotic complexes dissolve readily
in such medium. The observed changes in the Dh of the carriers
before and after antibiotic inclusion, as well as the change in the
solubility of the antibiotic before and after loading into the
carriers, indicate that CD-MAN-antibiotic and CD-GLU-
antibiotic complexes were successfully prepared. The content
of antibiotics in the complexes (wt%) is: ERY/CD-MAN-
ERY =15.02%. ERY/CD-ERY= 23.09%. ERY/CD-GLU-
Scheme 1 Schematic illustration of the preparation procedures for CD-MAN (or CD-GLU) and CD-MAN-antibiotic (or CD-GLU-antibiotic) complex.

Trojan Horse Antibiotic Nanocarriers
Fig. 3 Distribution of hydrodynamic diameter of CD-MAN, CD-GLU and
their antibiotic complexes in aqueous medium.
Fig. 1 FT-IR spectra of (a) β-CD, (b) CD-N3 (c) CD-MAN and (d)
CD-GLU.
and susceptibility breakpoints for the bacteria were based on
data published by The European Committee on Antimicrobial
Susceptibility Testing (29). ERY was chosen as a model antibi-
otic because it is the most important macrolide antibiotic and
has been widely used to treat infectious diseases caused by both
Gram-positive and Gram-negative bacteria for more than
60 years. However, it has been reported that common bacteria
have generally been resistant to ERY since 1970s (30). As
shown in Table Ii, ERY in CD-MAN-ERY complex has the
lowest MIC for all the test bacteria as compared with that in
CD-ERY and free ERY. S. aureus was selected as a model
Gram-positive bacterium because it is one of the most prevail-
ing causes of serious infections, and methicillin-resistant S. aureus
(MRSA) is responsible for several most difficult-to-treat infec-
tions in humans (31). The MIC of ERY in CD-MAN-ERY for
the ERY-susceptible strains (29), MRSA DM23605 and
ATCC 25923, (MIC of free ERY= 0.5 mg/L for both) was
reduced by 3.3 and 6.7 times, respectively. For the highly ERY-
resistant strains, ATCC BAA-44 and MRSA DR9369, (MIC
of free ERY >1024 mg/L for both), the ERY in CD-MAN-
ERY shows a reduction of MIC by a factor of >10 and >100,
to 76.8 and 4.8 mg/L, respectively. These results indicate that
the antibacterial activity of ERY was potentiated by forming a
complex with CD-MAN even though the carrier itself (CD or
CD-MAN) has no antibacterial property, as indicated by its
MIC of > 1024 mg/L for all the bacteria tested.
ERY = 15.10%. RIF/CD-MAN-RIF = 15.38%. RIF/CD-
RIF = 20.30%. CIP/CD-MAN-CIP = 8.10%. CIP/CD-
GLU-CIP=7.90%. The loading efficiency (defined as antibi-
otic loaded into carrier/initial amount of antibiotic) ranged
from 48 to 78%, depending on the carrier and antibiotic.
Antibacterial Activity of CD-MAN Loaded Antibiotics
In this study, MIC assay was used to evaluate the antibacterial
activity of CD-MAN-antibiotic complex. Four clinically impor-
tant bacteria were selected for our investigations, and both
susceptible and resistant strains were tested. The resistance
Some of the Gram-negative bacteria, such as E. coli,
P. aeruginosa and A. baumannii, are among the most challenging
multidrug-resistant (MDR) microorganisms (32). For the
Gram-negative bacteria, the MIC of ERY in CD-MAN-
ERY for E. coli DH5α and P. aeruginosa PAO1 decreased to
9.6 and 19.2 mg/L, respectively, which is 3.3 times lower than
that of free ERY (32 and 64 mg/L, respectively). In addition to
ERY, rifampicin (RIF), a widely-used hydrophobic antibiotic
of the rifamycin group, was also loaded into CD-MAN to test
its antibacterial efficacy. As shown in Table Iii, the RIF in CD-
MAN-RIF complex also shows a lower MIC for S. aureus, E. coli
and P. aeruginosa as compared to free RIF. These results
Fig. 2 UV–vis spectra of 1 mM aqueous solution of (a) CD-MAN, β-CD
and D-mannose, and (b) CD-GLU and D-glucose.

Li et al.
indicate that as an antibiotic carrier, CD-MAN is able to im-
prove the antibacterial activity of the loaded hydrophobic an-
tibiotics against both Gram-positive and Gram-negative bac-
teria, possibly by rapidly increasing their intracellular
concentrations.
kinetics since an antibiotic can have different degree of ioniza-
tion at different pH, and unionized drugs usually form more
stable complexes with CD carrier than their ionic counterparts
(33). The three antibiotics used in this work have different
ionization constant (pKa): ERY 8.8, RIF 1.7 and 7.9, CIP
6.1 and 8.7. Depending on the antibiotic, the effect of pH in
the range of physiological interest on antibiotic release may or
may not be significant.
It has been reported that there are several mechanisms
involved in the development of bacterial resistance to antibi-
otics, namely 1) changes in the antibiotic-binding sites on
RNA/proteins to prevent the antibiotic action, 2) hydrolysis
of antibiotics by enzymes, 3) active efflux to decrease the in-
tracellular concentration of antibiotics, and 4) alteration in the
permeability of the outer membrane of Gram-negative bacte-
ria to antibiotics (34). Among these mechanisms, the activity of
the efflux system can confer bacterial resistance to macrolides.
To investigate whether the CD-MAN-ERY complex will po-
tentiate the antibiotic activity in efflux-proficient bacteria, we
compared the antimicrobial activity of ERY in an
efflux-proficient strain of P. aeruginosa (PAO1) and its efflux-
deficient derivative, PAO397 (Table I). The MIC of free ERY
for P. aeruginosa PAO397 (8 mg/L) was eight times lower than
that for P. aeruginosa PAO1 (64 mg/L), indicating that the
efflux system has a significant contribution to ERY resistance
in P. aeruginosa PAO1. In the efflux-proficient PAO1 strain, the
In addition to CD-MAN-antibiotic complexes, CD-GLU-
antibiotic complexes were also prepared and investigated
(Scheme 1). The MIC of ERY in CD-GLU-ERY for S. aureus
ATCC 25923 and P. aeruginosa PAO1 is similar to that of ERY
in CD-MAN-ERY (Table Ii), which is not surprising since both
glucose and mannose are carbon sources for S. aureus and
P. aeruginosa. A similar result was obtained for CIP, a fluoro-
quinolone antibiotic, when loaded in CD-MAN or CD-GLU
against P. aeruginosa PAO1 (Table Iiii). Further investigation
with CIP was carried out against A. baumannii. Both CD-
GLU-CIP and CD-MAN-CIP complexes was able to reduce
its MIC for A. baumannii S1 (Table Iiii) by a factor of ~6. This
result is consistent with the postulate that the sugar-modified
carriers deliver the antibiotic more effectively into the bacterial
cells. It should be noted that the above-mentioned results were
obtained in simple media. The efficacy of this carrier system
may be different when applied in different biological environ-
ments (such as via intravenous injections or eye drops) since
biomolecules present in a particular environment may be ca-
pable of displacing the loaded antibiotic. Furthermore, the pH
of biological environments may also affect the antibiotic release
Tab le I MIC of antibiotic in CD-MAN-antibiotic, CD-GLU-antibiotic, CD-antibiotic complex, and free antibiotic for different bacterial strains
(i) Bacterium
Gram
Number of strains
Strain name
MIC (mg/L)
ERY in CD-MAN-ERY
(CD-GLU-ERY)a
0.15
ERY in CD-ERY
Free ERY
S. aureusb
Positive
4
MRSA DM23605
ATCC 25923
MRSA DR9369
ATCC BAA-44
0.23
0.12
7.36
>1024
0.5
0.5
>1024
>1024
0.075 (0.075)a
4.8
76.8
E. coli
Negative
Negative
1
2
DH5α
9.6
19.2 (19.2)a
4.8
29.4
32
P. aeruginosa
PAO1
29.4
7.36
64
8
PAO397 (PAO1 without efflux pumps)
(ii) Bacterium
Gram
Number of strains
Strain name
MIC (mg/L)
RIF in CD-MAN-RIF
0.0012
RIF in CD-RIF
0.0033
1.62
Free RIF
0.016
4
S. aureusb
E. coli
Positive
Negative
Negative
Gram
1
ATCC 25923
DH5α
1
1.23
P. aeruginosa
(iii) Bacterium
1
PAO1
4.92
12.99
32
Number of strains
Strain name
MIC (mg/L)
CIP in CD-MAN-CIP
0.088
CIP in CD-GLU-CIP
0.088
Free CIP
0.25
P. aeruginosab
A. baumanniib
Negative
Negative
1
1
PAO1
S1
0.044
0.044
0.25
a The value in brackets means the MIC of ERY in CD-GLU-ERY complex for the corresponding bacterial strain
b Values indicating susceptibility to the antibiotic are highlighted in bold. For S. aureus, MIC of >2 mg/L of ERY is considered resistant and MIC of ≤1 mg/L is
considered susceptible to ERY; MIC of ≤0.06 mg/L of RIF is considered susceptible and >0.5 mg/L is considered resistant to RIF. For P. a e r u g in o s a , MIC
≤0.5 mg/L of CIP is considered susceptible and MIC >1 mg/L is considered resistant to CIP. For A. baumannii, MIC ≤1 mg/L of CIP is considered susceptible and
MIC >1 mg/L is considered resistant to CIP. There are no corresponding resistance and susceptibility breakpoints for ERY for E. coli and P. aeruginosa, or RIF for
E. coli and P. aeruginosa (29)

Trojan Horse Antibiotic Nanocarriers
MIC for ERY was reduced by 3.3 times using CD-MAN-
ERY as compared to free ERY. By comparison, the corre-
sponding MIC reduction for the efflux-deficient P. aeruginosa
PAO397 strain was only 1.7 times. This suggests that the CD-
MAN carrier can potentiate the activity of the loaded antibi-
otic even in efflux-proficient bacterial strains. The mechanism
by which the carrier is able to accomplish this is not clear at
present, but it is plausible that effective delivery of the antibi-
otic into the bacterial cytoplasm rapidly increases its intracel-
lular concentrations and thus optimizes its activity.
hydrophobic fluorescent dye probe PBI was loaded into CD-
MAN to evaluate the ability of CD-MAN loaded drugs to
permeate into the bacterial cells. The uptake of hydrophobic
fluorescent dyes by bacteria is commonly applied as a method
to gauge bacterial resistance to hydrophobic antibiotics (23).
The fluorescent images of bacteria after incubation with free
PBI, and PBI loaded into CD (CD-PBI) and CD-MAN (CD-
MAN-PBI) are illustrated in Fig. 4. As shown in Fig. 4a^’,
E. coli DH5α incubated with CD-MAN-PBI exhibited the
strongest fluorescence intensity as compared to those incubat-
ed with the other two agents (Fig. 4a, a’ and a^). Similar results
for P. aeruginosa PAO1 and S. aureus ATCC 25923 were also
observed in Fig. 4b-b^’ and c-c^’, respectively. In addition, the
accumulation assay results (Fig. S2) show that the accumula-
tion of PBI from CD-MAN-PBI in these bacteria after 4 h is
significantly higher than that from CD-PBI and free PBI,
which is consistent with the results in Fig. 4. These results
suggest that antibiotic loaded in CD-MAN carriers is deliv-
ered into the bacterial cells more effectively than either the
antibiotic loaded in the CD carrier or the free form antibiotic.
A recent study has reported that the strongest resistance of
S. aureus to free ERY (MIC > 1024 mg/L, as shown as by
MRSA DR9369 and ATCC BAA-44 strains in Table Ii) is
the result of decreased uptake of the antibiotic into the bacte-
rial cells (38). Thus, the ability of CD-MAN to effectively
deliver the loaded antibiotic into bacterial cells is of crucial
importance in potentiating its efficacy against these highly
resistant bacteria.
Previous studies have reported that antibiotics in CD-
antibiotic complexes have a lower MIC (usually by a factor of
2 to 8) than the free form for some bacterial strains because of
the inhibition of enzymatic hydrolysis and improved intestinal
permeability of the antibiotic after inclusion in β-CD or β-CD
derivatives (methylated β-cyclodextrin and hydroxypropyl-β-cy-
clodextrin). However, the MIC of these CD-loaded antibiotics
are still similar to their free form for other bacterial strains (35).
Similarly, in the present study, it was found that although the
ERY loaded in the CD carrier exhibited lower MIC for some of
the tested bacterial strains, its efficacy against S. aureus ATCC
BAA-44, E. coli DH5α and P. aeruginosa PAO397 is not different
from that of free ERY (Table Ii). On the othr hand, with the
grafting of mannose and glucose moieties on the CD carrier,
ERY in CD-MAN-ERY and CD-GLU-ERY complexes was
able to reduce the MIC for the bacteria indicated in Table I by
a factor of 3 to >100, with the exception of efflux-deficient
P. aeruginosa PAO397 for reasons mentioned above. While it is
not entirely clear how the CD carrier increases the intracellular
concentration of the antibiotic, it is known that CDs can pass
through the bacterial outer membrane via a specific transport
system (porins) (36). In addition, D-mannose, like D-glucose, is a
carbon source for many bacteria which have transporters and
active uptake systems for mannose on the cell membrane. Thus,
bacteria, through their sugar sensing mechanisms, will be
attracted to the sugar on the CD carriers. We postulate that
in the subsequent interactions between the bacterial cell and
CD carrier, the antibiotic will be released from the carrier,
perhaps by actions of bacterial enzymes, resulting in high local
concentration of the antibiotic which enhances its entry into the
cell. Thus, in the presence of β-CD-degradable enzymes like α-
amylase (37), a faster antibiotic release from CD-MAN-
antibiotic and CD-GLU-antibiotic complexes can be expected.
The presence of free D-mannose on the efficacy of the CD-
MAN carrier was investigated by determining the MIC of
CD-MAN-ERY on S. aureus ATCC25923 in MHB medium
containing different concentrations of free D-mannose (1 g/L
and 10 g/L, respectively). It was found that the MIC values are
the same as that without D-mannose. Thus, an excess amount
of free D-mannose did not competitively affect the interactions
of the bacterial cells with CD-MAN-ERY.
Antibiotic Resistance Assay
Antibiotic resistance assays using S. aureus ATCC 25923 and
E. coli DH5α were carried out to assess their potential to de-
velop resistance to free antibiotic, CD-antibiotic and CD-
MAN-antibiotic complexes. Table IIi shows the resistant
S. aureus ATCC 25923 count after the antibiotic resistance
assay. After incubation with CD-MAN-ERY, CD-ERY and
free ERY, respectively, for 8 days, no resistant bacteria were
found for these agents. When the incubation period was in-
creased to 16 days, about 1 in 26700 and 1 in 25300 bacterial
cells developed resistance to free ERY and CD-ERY, respec-
tively. However, no resistant bacteria against CD-MAN-ERY
were found after incubation for 16 days. Table IIii shows the
corresponding antibiotic resistance assay results for E. coli
DH5α. After incubation with CD-MAN-RIF, CD-RIF and
free RIF, respectively, for 8 days, no resistant bacteria for all
the agents was found. When the incubation period was in-
creased to 16 days, E. coli developed resistance to CD-RIF
and free RIF. The corresponding number of resistant bacteria
exposed to CD-RIF and free RIF is 1 in 6210 and 1 in 6250,
respectively, indicating that antibiotic loaded into CD is not
able to inhibit the evolution of bacterial resistance. However,
no resistant bacteria against CD-MAN-RIF were found after
To further investigate the mechanism for the enhanced
antibacterial activity of the CD-MAN-antibiotic complex, a

Li et al.
incubation for 16 days. Detailed results of the antibiotic resis-
tance assay are given in Table S2 and S3. We postulate that
CD-MAN-ERY and CD-MAN-RIF did not promote the de-
velopment of intrinsic resistance in S. aureus ATCC 25923 and
E. coli DH5α, unlike CD-ERY, CD-RIF and the free form
antibtiotics, because enhanced delivery of the antibiotics from
the sugar-modified carriers into the bacterial cytosol resulted
in rapid attainment of its effective intracellular concentration.
The bacterial lysis assay results are shown in Fig. 5. The
higher OD260 of the bacterial suspension incubated with CD-
MAN in PBS as compared to incubation with only PBS can be
attributed to the absorbance of CD-MAN carrier, as shown in
Fig. S3. However, the OD260 of the bacterial suspension
(S. aureus, E. coli and P. aeruginosa) changed very little with time
when incubated with either a high-dose of CD-MAN carrier
in PBS (0.3 g/L) or just PBS (negative control). On the other
hand, when the bacterial suspension was incubated with bac-
terial lysis buffer (positive control), a significant increase in
OD260 was observed within the first hour (Fig. 5). These re-
sults indicate that the CD-MAN carrier does not cause bacte-
rial cell lysis and the principle of enhancing the delivery of
antibiotics into bacteria using such carriers does not rely on
damaged cell membrane. It has been reported that nanopar-
ticles which damage bacterial cell membranes promote the
horizontal conjugative transfer of multidrug-resistance genes
and may increase antibiotic resistance (11). Hence, this prob-
lem may not arise with carriers such as CD-MAN which do
not cause cell lysis.
Fig. 4 Fluorescence microscopy images showing accumulation of a hydrophobic fluorescent dye probe (3,4,9,10-perylenetetracarboxylic 3,4:9,10-dianhydride,
PBI) in E. coli DH5α (a-a^’), P. aeruginosa PAO1 (b-b^’) and S. aureus ATCC 25923 (c-c^’) after incubation for 4 h in culture medium only (a-c), culture medium
containing free PBI (a’-c’), culture medium containing CD-PBI (a^-c^) and culture medium containing CD-MAN-PBI (a^’-c^’). Scale bar is 100 μm.

Trojan Horse Antibiotic Nanocarriers
Stability
viability of 3T3 fibroblasts incubated with free ERY of con-
centration ranging from 8 to 512 mg/L (i.e. 1/8 to 8 times
MIC for P. aeruginosa PAO1) for 24 and 72 h was higher than
85% in all cases (Fig. 7a). The viability of the cells incubated
with CD-MAN-ERY complex of concentration ranging from
8 to 2048 mg/L (containing 1.2 to 307.2 mg/L of ERY, cor-
responding to 1/16 to 16 times MIC for P. aeruginosa PAO1)
was also higher than 85% (Fig. 7b). Similarly, the viability of
the cells incubated with CD-MAN carrier of concentration
ranging from 125 to 2000 mg/L was higher than 85%
(Fig. S4). There is no significant difference in cell viability
when the cells were incubated with either CD-MAN-ERY
complex or CD-MAN carrier, compared to that with free
ERY and the control experiment (cells incubated in growth
medium only), indicating that CD-MAN-ERY possesses no
significant cytotoxicity to 3T3 fibroblasts and the complexa-
tion with CD-MAN does not increase the cytotoxicity of ERY.
The stability of the antibiotic in the CD-MAN-antibiotic com-
plex was evaluated by bacterial growth inhibition assay over
an extended period. As shown in Fig. 6a, E. coli when cultured
in growth medium without antibiotic grew well and the bac-
terial concentration increased (as indicated by the increasing
OD600) with increasing time. When culture medium with free
ERY and CD-MAN-ERY at their respective MIC was used,
the growth of E. coli can be inhibited for only 1 day (Fig. 6a
and a’). The loss of ERY’s antibacterial activity can be attrib-
uted to its decomposition by hydrolysis in the medium (39).
On the other hand, when culture medium with 2 times the
respective MIC of free ERY and CD-MAN-ERY was used,
the growth of E. coli was inhibited for 3 and 18 days,
respectively (Fig. 6a and a’). Similarly, with P. aeruginosa
(Fig. 6b and b’) and S. aureus (Fig. 6c and c’), bacterial
growth was surpressed even after 18 days with 2 times
the respective MIC of CD-MAN-ERY whereas free ERY at
the corresponding MIC lost its efficacy much earlier. These
results indicate that ERY in CD-MAN-ERY is more stable
and its antibacterial activity is preserved for a much longer
period than the free form.
Cytotoxicity
The cytotoxicity of the free ERY, CD-MAN-ERY complex
and CD-MAN carrier was evaluated on a variety of mamma-
lian cell lines. Murine 3T3 fibroblasts and RAW 264.7 mu-
rine macrophages were evaluated using the MTT assay. 3T3
fibroblasts were used as model cells to study cytotoxicity of
CD-MAN-ERY complex in vitro as fibroblasts play a critical
role in fibrous encapsulation and wound healing due to their
ability tosynthesizeproteins of extracellular matrix (40). The
Table II Resistant bacterial count after antibiotic resistance assay
(i) S. aureus ATCC 25923
Culture medium
After 8 days
After 16 days
0
TSB + 0.25 × MICa CD-MAN-ERY
TSB + 0.25 × MICa CD-ERY
TSB + 0.25 × MICa free ERY
(ii) E. coli DH5α
0
0
0
1 in 26700
1 in 25300
Culture medium
After 8 days
After 16 days
0
NB + 0.25 × MICb CD-MAN-RIF
NB + 0.25 × MICb CD-RIF
NB + 0.25 × MICb free RIF
0
0
0
1 in 6210
1 in 6250
a Refer to Table I for MIC value of free ERY and ERY in CD-MAN-ERY and
CD-ERY. ERY in CD-MAN-ERY and CD-ERY constitutes 15 wt% and
23 wt%, respectively
b Refer to Table I for MIC value of free RIF and RIF in CD-MAN-RIF and CD-
RIF. RIF in CD-MAN-RIF and CD-RIF constitutes 15 wt% and 20 wt%,
respectively
Fig. 5 Comparison of OD260 of (a) S. aureus ATCC 25923, (b)
E. coli DH5α and (c) P. a e r u g in o s a PAO1 suspension after incubation
with CD-MAN in PBS. Bacterial lysis buffer and PBS were used as positive and
negative control, respectively.

Li et al.
Fig. 6 Comparison of antibacterial
activities of free ERY versus
CD-MAN-ERYusing growth curves
for (a,a’) E. coli DH5α, (b,b’)
P. aeruginosa PAO1 and (c,c’)
S. aureus ATCC 25923. Growth of
bacteria (as indicated by OD600
)
was monitored over 21 days when
cultured in medium containing free
ERY (a-c) or CD-MAN-ERY (a’-c’)
at MIC or 2 times MIC. Control
experiments were carried out in the
culture medium without ERYor
CD-MAN-ERY.
MTT assay was also carried out with RAW 264.7 macro-
phages, which express mannose receptors on their cell surface.
The results showed that after incubation with CD-MAN-ERY
complex at a concentration 2048 mg/L for 24 h, the viability
Fig. 7 Effect of (a) free ERY
and (b) CD-MAN-ERYon the
viability of 3T3 fibroblasts after 24
and 72 h. Viability is expressed as a
percentage relative to the result
obtained with the control
(3T3 fibroblast cells incubated in
growth medium only).

Trojan Horse Antibiotic Nanocarriers
of macrophages was ~88%, and there was no significant dif-
ference in cell viability as compared to the control experiment
(cells incubated in growth medium only), indicating that CD-
MAN-ERY possesses no significant cytotoxicity to RAW
264.7 macrophages. In addition to murine 3T3 fibroblasts
and RAW 264.7 macrophages, human NCM460 colonic ep-
ithelial and HEK 293T embryonic kidney cells were also used
for the evaluation of cytotoxicity of CD-MAN-ERY with
Tali™ scan analysis. As shown in Fig. S5, more than 85%
of NCM460 and HEK 293T cells remained viable after
incubation with 5 to 400 mg/L of free ERY for 24 h.
Similarly, more than 80% of NCM460 and HEK 293T cells
remained viable after incubation with CD-MAN-ERY, even
at a very high dosage (2000 mg/L, containing 300 mg/L of
ERY, corresponding to ~16 times of MIC for P. aeruginosa
PAO1) for 24 h. These findings are very similar to those ob-
tained from the MTT assay, and confirm that the CD-MAN-
ERY has no significant cytotoxicity to mammalian cells over
the concentration range tested.
authors would also like to thank Dr. Rong Wang from the
Department of Chemical & Biomolecular Engineering,
National University of Singapore for his assistance in the an-
tibiotic resistance assay.
REFERENCES
1. Chang HH, Cohen T, Grad YH, Hanage WP, O’Brien TF,
Lipsitch M. Origin and proliferation of multiple-drug resis-
tance in bacterial pathogens. Microbiol Mol Biol Rev.
2015;79(1):101–16.
2. Klevens RM, Edwards JR, Richards CL, Horan TC, Gaynes RP,
Pollock DA, et al. Estimating health care-associated infections
and deaths in US hospitals, 2002. Public Health Rep.
2007;122(2):160–6.
3. Klein E, Smith DL, Laxminarayan R. Hospitalizations and deaths
caused by methicillin-resistant Staphylococcus aureus, United States,
1999–2005. Emerg Infect Dis. 2007;13(12):1840–6.
4. Andersson DI, Hughes D. Persistence of antibiotic resistance in
bacterial populations. FEMS Microbiol Rev. 2011;35(5):901–11.
5. Boucher HW, Talbot GH, Benjamin DK, Bradley J, Guidos RJ,
Jones RN, et al. 10 x ′20 Progress-development of new drugs active
against Gram-negative bacilli: an update from the infectious dis-
eases society of america. Clin Infect Dis. 2013;56(12):1685–94.
6. Pelgrift RY, Friedman AJ. Nanotechnology as a therapeutic tool to
combat microbial resistance. Adv Drug Deliv Rev. 2013;65(13–14):
1803–15.
7. Adhikari MD, Goswami S, Panda BR, Chattopadhyay A, Ramesh
A. Membrane-directed high bactericidal activity of (gold
nanoparticle)polythiophene composite for niche applications
against pathogenic bacteria. Adv Healthc Mater. 2013;2(4):
599–606.
8. Quinn JF, Whittaker MR, Davis TP. Delivering nitric oxide with
nanoparticles. J Control Release. 2015;205:190–205.
9. Seleem MN, Jain N, Pothayee N, Ranjan A, Riffle JS,
Sriranganathan N. Targeting Brucella melitensis with polymeric
nanoparticles containing streptomycin and doxycycline. FEMS
Microbiol Lett. 2009;294(1):24–31.
10. Zhao L, Zhu BY, Jia YH, Hou WJ, Su C. Preparation of biocom-
patible carboxymethyl chitosan nanoparticles for delivery of
antibiotic drug. Biomed Res Int. 2013;13:Article ID 236469,
236467 pages.
11. Qiu ZG, Yu YM, Chen ZL, Jin M, Yang D, Zhao ZG, et al.
Nanoalumina promotes the horizontal transfer of multiresistance
genes mediated by plasmids across genera. Proc Natl Acad Sci U S
A. 2012;109(13):4944–9.
12. Chilajwar SV, Pednekar PP, Jadhav KR, Gupta GJC, Kadam VJ.
Cyclodextrin-based nanosponges: a propitious platform for enhanc-
ing drug delivery. Expert Opin Drug Deliv. 2014;11(1):111–20.
13. Del Valle EMM. Cyclodextrins and their uses: a review. Process
Biochem. 2004;39(9):1033–46.
14. Bouckaert J, Li Z, Xavier C, Almant M, Caveliers V, Lahoutte T,
et al. Heptyl α-D-mannosides grafted on a β-cyclodextrin core to
interfere with Escherichia coli adhesion: an in vivo multivalent effect.
Chem Eur J. 2013;19(24):7847–55.
CONCLUSIONS
Mannose and glucose molecules were grafted on modified β-
cyclodextrin via alkyne-azide click reaction. The resultant CD-
MAN and CD-GLU were used as nanocarriers for loading
hydrophobic antibiotics (ERY, RIF and CIP) and delivering
these antibiotics into Gram-positive (S. aureus including MRSA)
and Gram-negative (E. coli, P. aeruginosa and A. baumannii) bac-
teria. These sugar-modified nanocarriers enhanced the deliv-
ery and accumulation of the loaded antibiotic in the bacterial
cells, and as a result, the minimum inhibitory concentration of
the loaded antibiotics against the tested bacteria were reduced
by a factor ranging from 3 to >100 compared to the free
antibiotics. In addition, the CD-MAN-antibiotic complex is
able to inhibit the development of antibiotic resistance of
S. aureus and E. coli, and prolong the stability of the loaded
antibiotic and its activity to inhibit bacterial growth. In view
of the paucity of new antibiotics in the pipeline, it is highly
advantageous that these non-cytotoxic sugar-modfied
nanocarriers can potentiate the activity of existing antibiotics,
especially against multidrug-resistant bacteria. However, in vivo
evaluation of the performance of these sugar-modified
nanocarriers is necessary since the in vitro assays cannot dupli-
cate the complicated biological environment, and the perfor-
mance of these carriers in oral or intravenous applications may
be very different from that in topical application.
15. Rulis AM. USFDA agency response letter GRAS notice no. GRN
000074. Available from: http://www.fda.gov/Food/
IngredientsPackagingLabeling/GRAS/NoticeInventory/
ucm154182.htm.
ACKNOWLEDGMENTS AND DISCLOSURES
16. Postma PW, Lengeler JW. Phosphoenolpyruvate - carbohy-
drate phosphotransferase system of bacteria. Microbiol Rev.
1985;49(3):232–69.
This work was financially supported by the National
University of Singapore Grant R-279-000-359-731. The

Li et al.
17. Kumar A, Chua KL, Schweizer HP. Method for regulated expres-
sion of single-copy efflux pump genes in a surrogate Pseudomonas
aeruginosa strain: identification of the BpeEF-OprC chloramphenicol
and trimethoprim efflux pump of Burkholderia pseudomallei 1026b.
Antimicrob Agents Chemother. 2006;50(10):3460–3.
18. Zhang Q, Slavin S, Jones MW, Haddleton AJ, Haddleton DM.
Terminal functional glycopolymers via a combination of catalytic
chain transfer polymerisation (CCTP) followed by three consecu-
tive click reactions. Polym Chem. 2012;3(4):1016–23.
29. Breakpoint tables for interpretation of MICs and zone diameters.
Version 5.0, 2015. The European Committee on Antimicrobial
Susceptibility Testing.; Available from: http://www.eucast.org/
clinical_breakpoints/.
30. Roberts MC. Update on macrolide-lincosamide-streptogramin,
ketolide, and oxazolidinone resistance genes. FEMS Microbiol
Lett. 2008;282(2):147–59.
31. Fowler VG, Proctor RA. Where does a Staphylococcus aureus vaccine
stand? Clin Microbiol Infect. 2014;20(S5):66–75.
19. Ashton PR, Koniger R, Stoddart JF, Alker D, Harding VD. Amino
acid derivatives of beta-cyclodextrin. J Org Chem. 1996;61(3):903–8.
20. Mariappan TT, Jindal KC, Singh S. Overestimation of rifampicin
during colorimetric analysis of anti-tuberculosis products contain-
ing isoniazid due to formation of isonicotinyl hydrazone. J Pharm
Biomed Anal. 2004;36(4):905–8.
32. Singh R, Smitha MS, Singh SP. The role of nanotechnology in
combating multi-drug resistant bacteria. J Nanosci Nanotechnol.
2014;14(7):4745–56.
33. Loftsson T, Brewster ME. Pharmaceutical applications of cyclodex-
trins.1. Drug solubilization and stabilization. J Pharm Sci.
1996;85(10):1017–25.
21. Andrews JM. Determination of minimum inhibitory concentra-
tions. J Antimicrob Chemother. 2001;48:5–16.
34. Oldfield E, Feng XX. Resistance-resistant antibiotics. Trends
Pharmacol Sci. 2014;35(12):664–74.
22. Xu LQ, Wang L, Zhang B, Lim CH, Chen Y, Neoh KG, et al.
Functionalization of reduced graphene oxide nanosheets via stack-
ing interactions with the fluorescent and water-soluble perylene
bisimide-containing polymers. Polymer. 2011;52(11):2376–83.
23. Richmond GE, Chua KL, Piddock LJV. Efflux in Acinetobacter
baumannii can be determined by measuring accumulation of
H33342 (bis-benzamide). J Antimicrob Chemother. 2013;68(7):
1594–600.
24. Hüttel R. Über einige aldehyde der pyrazol- und der 1.2.3-triazol-
reihe. Chem Ber. 1941;74(10):1680–7.
25. Tang W, Becker ML. BClick^ reactions: a versatile toolbox
for the synthesis of peptide-conjugates. Chem Soc Rev.
2014;43(20):7013–39.
26. Xu J, Liu SY. Synthesis of well-defined 7-arm and 21-arm poly (n-
isopropylacrylamide) star polymers with beta-cyclodextrin cores via
click chemistry and their thermal phase transition behavior in aque-
ous solution. J Polym Sci Part A: Polym Chem. 2009;47(2):404–19.
27. Theocharis AB, Alexandrou NE. Synthesis and spectral data of 4,5-
bis[5-aryl-1,3,4-oxadiazol-2-yl]-1-benzyl-1,2,3-triazoles. J
Heterocycl Chem. 1990;27(6):1685–8.
35. Athanassiou G, Michaleas S, Lada-Chitiroglou E, Tsitsa T,
Antoniadou-Vyza E. Antimicrobial activity of beta-lactam antibi-
otics against clinical pathogens after molecular inclusion in several
cyclodextrins. A novel approach to bacterial resistance. J Pharm
Pharmacol. 2003;55(3):291–300.
36. Pajatsch M, Andersen C, Mathes A, Bock A, Benz R, Engelhardt
H. Properties of a cyclodextrin-specific, unusual porin from
Klebsiella oxytoca. J Biol Chem. 1999;274(35):25159–66.
37. Fetzner A, Bohm S, Schreder S, Schubert R. Degradation of raw or
film-incorporated beta-cyclodextrin by enzymes and colonic bacte-
ria. Eur J Pharm Biopharm. 2004;58(1):91–7.
38. Piatkowska E, Piatkowski J, Przondo-Mordarska A. The strongest
resistance of Staphylococcus aureus to erythromycin is caused by de-
creasing uptake of the antibiotic into the cells. Cell Mol Biol Lett.
2012;17(4):633–45.
39. Steffansen B, Bundgaard H. Erythromycin prodrugs - kinetics of
hydrolysis of erythromycin and various erythromycin 2′-esters in
aqueous-solution and human-plasma. Int J Pharm. 1989;56(2):
159–68.
28. Song W, Yu XW, Wang SX, Blasier R, Markel DC, Mao GZ, et al.
Cyclodextrin-erythromycin complexes as a drug delivery device for
orthopedic application. Int J Nanomedicine. 2011;6:3173–86.
40. Schultz GS, Wysocki A. Interactions between extracellular matrix
and growth factors in wound healing. Wound Repair Regen.
2009;17(2):153–62.