Biodegradable broad-spectrum antimicrobial polycarbonates: Investigating the role of chemical structure on activity and selectivity

Article abstract of DOI:10.1021/ma4019685

A series of biodegradable polycarbonate polymers was designed and synthesized via organocatalytic ring-opening polymerization of functional cyclic carbonate monomer (MTC-OCH₂BnCl). By adopting a facile postpolymerization functionalization strategy, the polycarbonates were quaternized to yield cationic polymers with quaternary ammonium groups of various pendant structures (e.g., alkyl, aromatic, imidazolinium). The biological properties of these polymers were investigated by microbial growth inhibition assays against clinically relevant Gram-positive and Gram-negative bacteria, fungus as well as hemolysis assays using rat red blood cells. A judicious choice in the structure of the cationic appendages elucidated that the amphiphilic balance of the polymers is a pertinent determinant to render substantial antimicrobial potency and low hemolysis, consequently affording the polymer pButyl-20 (degree of polymerization, 20; quaternary ammonium group, N,N-dimethylbutylammonium) as a highly efficacious and nonhemolytic antimicrobial agent with a remarkable selectivity of more than 1026. To ameliorate the selectivity against a wider spectrum of microbes including the difficult-to-kill Pseudomonas aeruginosa, it was shown that polymers containing N,N-dimethylbutylammonium and N,N-dimethylbenzylammonium groups in 1:1 molar ratio exerted considerable antimicrobial potency while remaining relatively nonhemolytic. Biophysical studies encompassing the determination of water-octanol partition coefficients (log P) and dye leakage studies from model liposomes provided useful insights which delineate the pivotal role of cationic group structure in the antimicrobial activity and mechanism of these polymers. Through field emission scanning electron microscopy (FE-SEM), a physical lysis of microbial cell membranes was deemed operative in the antimicrobial action of these macromolecular agents, which would consequently reduce the propensity toward resistance development. These polymers are envisaged to be promising antimicrobial agents for the prevention and treatment of multidrug-resistant pathogenic infections.

Full text of DOI:10.1021/ma4019685

Article
pubs.acs.org/Macromolecules
Biodegradable Broad-Spectrum Antimicrobial Polycarbonates:
Investigating the Role of Chemical Structure on Activity and
Selectivity
Willy Chin,†,⊥ Chuan Yang,† Victor Wee Lin Ng,† Yuan Huang,‡ Junchi Cheng,‡ Yen Wah Tong,§
Daniel J. Coady,∥ Weimin Fan,‡ James L. Hedrick,∥,* and Yi Yan Yang†,*
†Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, Singapore 138669, Singapore
‡State Key Laboratory for Diagnosis and Treatment of Infectious Diseases, First Affiliated Hospital, College of Medicine, Zhejiang
University, 79 Qingchun Road, Hangzhou 310003, China
§Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576,
Singapore
⊥NUS Graduate School for Integrative Sciences and Engineering (NGS), 28 Medical Drive, Singapore 117456, Singapore
∥IBM Almaden Research Center, 650 Harry Road, San Jose, CA 95120, United States
S
* Supporting Information
ABSTRACT: A series of biodegradable polycarbonate poly-
mers was designed and synthesized via organocatalytic ring-
opening polymerization of functional cyclic carbonate
monomer (MTC−OCH2BnCl). By adopting a facile post-
polymerization functionalization strategy, the polycarbonates
were quaternized to yield cationic polymers with quaternary
ammonium groups of various pendant structures (e.g., alkyl,
aromatic, imidazolinium). The biological properties of these
polymers were investigated by microbial growth inhibition
assays against clinically relevant Gram-positive and Gram-
negative bacteria, fungus as well as hemolysis assays using rat
red blood cells. A judicious choice in the structure of the
cationic appendages elucidated that the amphiphilic balance of the polymers is a pertinent determinant to render substantial
antimicrobial potency and low hemolysis, consequently affording the polymer pButyl_20 (degree of polymerization, 20;
quaternary ammonium group, N,N-dimethylbutylammonium) as a highly efficacious and nonhemolytic antimicrobial agent with
a remarkable selectivity of more than 1026. To ameliorate the selectivity against a wider spectrum of microbes including the
difficult-to-kill Pseudomonas aeruginosa, it was shown that polymers containing N,N-dimethylbutylammonium and N,N-
dimethylbenzylammonium groups in 1:1 molar ratio exerted considerable antimicrobial potency while remaining relatively
nonhemolytic. Biophysical studies encompassing the determination of water-octanol partition coefficients (log P) and dye
leakage studies from model liposomes provided useful insights which delineate the pivotal role of cationic group structure in the
antimicrobial activity and mechanism of these polymers. Through field emission scanning electron microscopy (FE-SEM), a
physical lysis of microbial cell membranes was deemed operative in the antimicrobial action of these macromolecular agents,
which would consequently reduce the propensity toward resistance development. These polymers are envisaged to be promising
antimicrobial agents for the prevention and treatment of multidrug-resistant pathogenic infections.
macromolecules that are capable of exerting their antimicrobial
activity by disrupting the microbial cell membranes.3 Owing to
the physical nature of its mode of activity, it is postulated to be
less susceptible than metabolic targeting of conventional
antibiotics to the development of resistance. Despite being
highly efficacious, their applicability is still beset with a number
of limitations including: high cytotoxicity (e.g., hemolysis),
1. INTRODUCTION
The ongoing prevalence of drug-resistant bacterial infections
represents a challenging and eminent health issue facing society
today.1 Exacerbated by the diminishing number of novel and
effective antibiotics,2 this precarious situation has since
prompted unremitting efforts for developing alternative
compounds that not only show high efficacy toward pathogenic
microbes, but also exhibit reduced propensity toward resistance
development. To that end, an unconventional class of
antimicrobial agents known as natural antimicrobial peptides
(AMPs) has emerged in the past decade as effective
Received: September 23, 2013
Revised: November 5, 2013
Published: November 15, 2013
© 2013 American Chemical Society
8797
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Scheme 1. Synthetic Scheme and Structure of (a) MTC−OCH2BnCl Monomer and (b) Amphiphilic Cationic Polycarbonates
by Varying the Cationic Group Structures and (c) Gel-Permeation Chromatograph of Precursor Polymer Prior to Post-
Polymerization Quaternization
poor proteolytic stability and pharmacokinetics (i.e., short half-
life in vivo), and high manufacturing cost.4
have also recently demonstrated their ability to target
pathogens via a membrane disruption mechanism. Notwith-
standing the numerous synthetic efforts in developing potent
macromolecules for targeting a wide spectrum of microbes, the
relatively low selectivity of many polymers resulting in
cytotoxicity to mammalian cells (in particular, hemolysis)
remains an unsolved issue.9 Furthermore, the nonbiodegradable
polymeric backbones observed in the majority of these
synthetic materials would present significant problems during
in vivo administration, thus hampering their potential for
clinical applications.
The design principles of most AMPs have hitherto been
governed by two common characteristics, which are a facially
amphiphilic nature and cationic charge.5 Such attributes allow
the AMPs to electrostatically adhere to the negatively charged
microbial cell surface while the hydrophobic moieties disrupt
the membrane by insertion into the bilayer core. By drawing
“molecular inspiration” from AMPs so as to mimic their
cationic and amphiphilic structure, synthetic antimicrobial
polymers6,7 (e.g., polymethacrylates8 and polynorbornenes9)
8798
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Recently, we reported biodegradable antimicrobial polycar-
bonates, which were designed via the “segregated monomer”
approach, wherein a relatively hydrophobic monomer is
randomly copolymerized with a cationic monomer to afford
either statistical or block copolymers.10,11 It is widely
established that the hydrophobic/hydrophilic balance (i.e.,
amphiphilicity) of antimicrobial polymers represents a pivotal
determinant in influencing how they interact with cellular
membranes, and as a consequence imparts selectivity toward
bacteria over mammalian cells.12 The amphiphilic balance of
polycarbonates could thus be fine-tuned by varying the
monomeric composition across the polymer chain. These
materials demonstrated antimicrobial activities toward a wide
spectrum of microbes with good hemolytic profiles; however,
the observed minimum inhibitory concentrations (MICs) were
still considerably high when compared to most AMPs, which
consequently limits their usage in biomedical applications.
As highlighted in our recent review,13 many classes of
antimicrobial polymers have primarily focused on adopting the
“segregated monomer” and “facially amphiphilic monomer”
(i.e., each repeat unit is comprised of a separate hydrophobic
and cationic section) approaches for tuning the hydrophobic/
hydrophilic balance. In contrast, relatively fewer studies have
been dedicated toward polymers employing the “same
centered” approach for developing macromolecular antimicro-
bial agents. In the “same centered” approach, a hydrophobic
moiety (usually an alkyl chain) is directly conjugated to a
cationic center, allowing the facile optimization of its
amphiphilicity by varying various structural parameters
pertaining to the hydrophobic component. A notable work
described by Sens and co-workers14 have shown that
polymethacrylates with cationic pyridinium and hydrophobic
tails on the “same center” led to improved selectivities while a
spatial separation of the cationic and hydrophobic components
(i.e., “segregated monomer”) rendered the polymers to become
significantly biocidal (hemolytic and antimicrobial). Taking into
account the findings gathered from these studies, it therefore
remains desirable to design and synthesize biodegradable
antimicrobial polycarbonates via the less explored “same
centered” approach which we postulate to be able to lead to
a pronounced membrane-lytic activity toward microbes while
retaining or even augmenting their previously observed
biocompatibility, consequently giving rise to promising
materials with potent antimicrobial activity and high selectivity.
The study presented herein describes a series of amphiphilic
polycarbonates with the “same centered” structure, which was
synthesized via metal-free organocatalytic ring-opening poly-
merization of functional cyclic carbonate monomer to
investigate the structure−activity relationship. By adopting a
facile and efficient postpolymerization quaternization reaction,
the cationic group structure and their respective compositions
were systematically modulated to study their effect on
antimicrobial activity against clinically relevant microbes and
toxicity against mammalian erythrocytes. Results revealed that
by carefully controlling the hydrophobic/hydrophilic balance of
the polymers, it is possible to dramatically enhance the
selectivity through subtle structural modifications in the
cationic groups, giving rise to biodegradable polymers with
remarkably high and unprecedented selectivity toward a broad
range of pathogenic microbes over mammalian cells. To
elucidate the role of the cationic groups in the observed
biological activity, a series of biophysical studies such as water-
octanol partition coefficients and polymer-induced dye leakage
from model liposomes was carried out. In addition, the
antimicrobial mechanism of action was explored by visualizing
the morphological changes incited by the polymers on the
bacteria cell surfaces.
2. EXPERIMENTAL SECTION
2.1. Materials. 2,2-Bis(hydroxymethyl)propionic acid (bis-MPA)
and N-(3,5-trifluoromethyl) phenyl-N′-cyclohexylthiourea (TU) were
prepared according to our previous protocol.15,16 TU was dissolved in
dry tetrahydrofuran (THF), stirred with CaH2, filtered, and freed of
solvent in vacuo. Prior to use, 1,8-diazabicyclo [5,4,0]undec-7-ene
(DBU) were stirred over CaH2 and vacuum distilled before being
transferred to a glovebox. All other chemical reagents were purchased
from Sigma-Aldrich and used as received unless specified. Ultra pure
(HPLC grade) water was obtained from J.T. Baker (U.S.A.).
Phosphate-buffered saline (PBS) at 10× concentration was purchased
from 1st BASE (Singapore) and diluted to the intended concen-
trations before use. Tryptic soy broth (TSB) powder was bought from
BD Diagnostics (Singapore) and used to prepare the microbial broths
according to the manufacturer’s instructions. Methanol, ethanol,
formalin solution (10% neutral buffered) and calcein were purchased
from Sigma-Aldrich (Singapore) and used as received. The
phospholipids 1,2-dioleoyl-snglycero-3-phospho-(10-rac-glycerol)
(DOPG), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE)
and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) were obtained
as dry powder from Avanti Polar Lipids, Inc. Staphylococcus aureus
(ATCC No. 6538), Escherichia coli (ATCC No. 25922) and
Pseudomonas aeruginosa (ATCC No. 9027) were obtained from
ATCC (U.S.A) and reconstituted according to the suggested
protocols. Methicillin-resistant S. aureus (MRSA), vancomycin-
resistant Enterococcus (VRE), carbapenem-resistant Acinetobacter
baumannii and fluconazole-resistant Cryptococcus neoformans were
extracted from patients’ blood (MRSA, VRE and carbapenem-resistant
A. baumannii) and cerebrospinal fluid (fluconazole-resistant C.
neoformans) samples, and kindly provided by Z. Q. Wei, Department
of Infectious Diseases, The First Affiliated Hospital, College of
Medicine, Zhejiang University, P. R. China.
2.2. Synthesis of MTC−OCH2BnCl Monomer (Scheme 1a).
MTC−OCH2BnCl was synthesized with reference to the protocol
reported in the previous work.15 Briefly, in a dry three-neck round-
bottom flask equipped with a stir bar, MTC−OH (3.08 g, 19.3 mmol)
was dissolved in dry THF (50 mL) with a 3−4 drops of
dimethylformamide (DMF). A solution of oxalyl chloride (2.45 mL,
28.5 mmol) in THF (50 mL) was subsequently added from a dropping
funnel. Under an inert atmosphere, the solution was stirred for 1 h,
after which volatiles were removed under vacuum, yielding an off-
white solid (i.e., 5-chlorocarboxy-5-methyl-1,3-dioxan-2-one inter-
mediate). The solid was heated to 60 °C for a brief 2−3 min to
remove any residual solvent, and then redissolved in dry CH2Cl2 (50
mL) and cooled down to 0 °C via an ice bath. A mixture of p-
chloromethyl benzyl alcohol (2.79 g, 17.8 mmol) and pyridine (1.55
mL, 19.3 mmol) dissolved in dry CH2Cl2 (50 mL) was then added
dropwise over a duration of 30 min, and allowed to stir at 0 °C for an
additional 30 min before leaving it at ambient temperature for further
stirring overnight. After removal of solvent, the crude product was
subjected to purification by flash column chromatography using silica
gel and a hexane-ethyl acetate solvent system as the eluent (gradient
elution up to 80% vol. ethyl acetate) to yield MTC−OCH2BnCl as a
white solid (4.1 g, 13.6 mmol, 70% yield). 1H NMR (400 MHz,
CDCl3, 22 °C): δ 7.36 (dd, J = 30.3, 8.0 Hz, 4H, Ph-H), 5.20 (s, 2H,
OCH2), 4.69 (d, J = 10.8 Hz, 2H, CHaHb), 4.58 (s, 2H, CH2Cl), 4.20
(d, J = 10.8 Hz, 2H, CHaHb), 1.31 (s, 3H, CCH3).
2.3. Synthesis of Representative Cationic Polymers (Scheme
1b). The detailed procedures for the ring-opening polymerization
(ROP) of MTC−OCH2BnCl with 4-methyl benzyl alcohol as initiator
are given as a representative example. Using a glovebox, MTC−
OCH2BnCl (448 mg, 1.2 mmol) was added to a reaction vial
containing 1-(3,5-bis(trifluoromethyl)-phenyl)-3-cyclohexyl-2-thiourea
(TU) (22.2 mg, 0.05 mmol) dissolved in dry DCM (2 mL). The
8799
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
mixture was subsequently charged with 4-methyl benzyl alcohol (7.32
mg, 0.06 mmol), before adding 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) (8.96 μL, 0.05 mmol) and left to stir at room temperature for
approximately 30 min. At the end of the reaction, an excess of benzoic
acid (10 mg, 0.08 mmol) was added to quench the catalyst. The crude
polymer was then precipitated twice into cold methanol and the
killing-efficiency studies, the samples were taken after 18 h incubation
and plated using the same protocol for viable counts.
2.6. Hemolysis Assay. The toxicity of the polymers against
mammalian erythrocytes was tested using fresh rat red blood cells
(rRBCs). Briefly, rRBCs were diluted 25-fold in PBS to achieve 4% v/v
of blood content. The polymers were dissolved in PBS at
concentrations ranging from 0 to 4000 μg mL−1 by serial dilutions.
Equal volumes of polymer solutions (100 μL) were then mixed with
the diluted blood suspension (100 μL). The mixtures were then
incubated at 37 °C for 1 h to allow for the interactions between rRBC
and the polymers to take place. After that, the mixture was subjected to
centrifugation (1000 g for 5 min, 4 °C), and 100 μL aliquots of the
supernatant was pipetted into a 96-well microplate. The hemoglobin
release was measured spectrophotometrically by measuring the
absorbance of the samples at 576 nm using the microplate reader
(TECAN, Switzerland). Two control groups were employed for this
assay: untreated rRBC suspension (negative control), and rRBC
suspension treated with 0.1% Triton-X (positive control). Each assay
was performed in 4 replicates. The percentage of hemolysis was
defined as follows:
1
supernatant decanted to obtain a white solid (91% yield). H NMR
(400 MHz, CDCl3, 22 °C): δ 7.41−7.26 (m, 89H, Ph-H), 5.18−5.06
(m, 48H, −OCOCH2-), 4.58−4.52 (m, 45H, −CH2Cl), 4.34−4.21
(m, 83H, −OCOOCH2- and −OCH2CCH3−), 2.34 (s, 3H, initiator
CH3), 1.28−1.15 (m, 69H, −CH3).
For postpolymerization quaternization, the aforementioned poly-
mer (0.35 g, 0.05 mmol) was initially added to a reaction vial and
dissolved in acetonitrile (10 mL). The quaternizing agent was added in
excess (5 equiv of benzyl chloride groups) and the reaction mixture
left to stir overnight at ambient temperature. Quaternization reactions
involving trimethylamine (TMA) were carried out in a pressure safe
Schlenk tube owing to the gaseous nature of TMA. The addition of
TMA gas was carried out by initially cooling the reaction mixture to
−78 °C using dry ice and the condensed TMA gas (4 mL, 42.6 mmol)
was subsequently added and sealed before leaving it to stir overnight at
ambient temperature. For the quaternization of polymers with
pyridine, the reaction was carried out at 40 °C. The synthesis of
polymers containing randomly distributed quaternary ammonium
centers was carried out in two consecutive steps, whereby the
precursor polymer was first quaternized with N,N-dimethylbutylamine
to the desired molar composition, as verified by in situ 1H NMR. After
that, the partially quaternized polymer was reacted with an excess of
the second quaternizing agent.
Following quaternization, the crude product was then subjected to
purification via dialysis against a 1:1 acetonitrile-isopropanol solvent
system. After removal of solvent, the resultant polymer was freeze-
dried to yield a white solid. The yields and analytical data for all
quaternized polymers are provided in the Supporting Information.
2.4. MIC Measurements. All bacteria strains obtained from
ATCC were reconstituted from its lyophilized form according to the
manufacturer’s protocol. Bacteria from ATCC and clinical samples
were cultured in tryptic soy broth (TSB) and Mueller−Hinton broth
(MHB) solutions, respectively, at 37 °C under constant shaking of 100
rpm. The MICs of the polymers were measured using the broth
microdilution method. Briefly, 100 μL of TSB or MHB solution
containing the polymer (with a fixed deionized (DI) water
concentration of 20% v/v) at various concentrations (0−500 μg
mL−1) was placed into each well of a 96-well microplate. An equal
volume of microbial suspension (3 × 105 CFU mL−1) was added into
each well. Prior to mixing, the microbial sample was first inoculated
overnight to enter its log growth phase. The concentration of bacterial
solution was adjusted to give an initial optical density (OD) reading of
approximately 0.07 at 600 nm wavelength on a microplate reader
(TECAN, Switzerland), which corresponds to the concentration of
McFarland 1 solution (3 × 108 CFU mL−1). The bacterial solution was
then further diluted 1000-fold to achieve an initial inoculum of 3 × 105
CFU mL−1. The 96-well plate was kept in an incubator at 37 °C under
constant shaking of 100 rpm for 18 h. The MIC was taken as the
concentration of the antimicrobial polymer at which no microbial
growth was observed with unaided eyes and the microplate reader at
the end of 18 h incubation. Broth containing microbial cells alone was
used as the negative control, and each test was carried out in 6
replicates.
Hemolysis (%) = [(OD576 nm of the treated sample − OD576 nm of
the negative control)/(OD576 nm of positive control − OD576 nm of
negative control)] × 100%.
2.7. FE-SEM. Bacteria cells grown in TSB with or without polymer
treatment were performed using a similar protocol as MIC
measurements but with a 2 h incubation time. All the samples were
collected into a microfuge tube and pelleted at 4000 rpm for 5 min,
and then washed twice with PBS buffer. Subsequently, fixing the
samples with formalin solution (10% neutral buffered) for 60 min was
conducted, followed by washing with DI water twice. Dehydration of
the samples was performed by using a series of ethanol/water solution
(35%, 50%, 75%, 90%, 95%, and 100%). The dehydrated samples were
dried at room temperature for 2 days before being mounted on carbon
tape and coated with platinum for imaging using a JEOL JSM-7400F
(Japan) field emission scanning electron microscope.
2.8. Water−Octanol Partition Coefficient (log P). Solutions
containing representative dansyl-conjugated polymers (500 μL, 63 μg
mL−1) were prepared in PBS buffer in a microfuge tube and octanol
(500 μL) was added. The tube was vortexed for 5 min and then
allowed to sit overnight in the dark. After 5 min of centrifugation at
3000 rpm, an aliquot of each phase was diluted 10-fold into methanol
and the fluorescence spectra were recorded. The concentration of
polymer in each phase was determined with reference to calibration
curves obtained by measuring the fluorescence spectra of the polymers
at various concentrations in methanol. The partition coefficient was
defined as log P = log ([P]oct/[P]aq) where [P]oct and [P]aq are the
concentration of the polymer in the octanol and aqueous phases,
respectively. Measurements were performed in triplicates.
2.9. Liposome Dye-Leakage Assay. To prepare the liposomes,
the following buffers were prepared: buffer 1 consisted of 10 mM
Na2HPO4 in H2O at pH = 7.0, buffer 2 consisted of 10 mM Na2HPO4
and 90 mM NaCl. Calcein dye was dissolved in buffer 1 to achieve a
concentration of 40 mM. To a clean round-bottom flask, appropriate
volumes of lipid stocks (25 mg/mL CHCl3) were added to make up 1
mL of CHCl3 (For 4:1 PE/PG vesicles, 238 μL of PE and 63.6 μL of
PG were used; For PC vesicles, 314.4 μL of PC was used). The solvent
was removed by a stream of nitrogen gas to obtain a thin lipid film,
which was then hydrated by 1 mL of calcein solution. The mixture was
left to stir for 1 h, after which it was subjected to 10 freeze−thaw cycles
(using dry ice/acetone to freeze and warm water to thaw). The
suspension was extruded 20 times through a polycarbonate membrane
with 400 nm pore diameter. The excess dye was removed using
Sephadex G-50 column and buffer 2 as the eluent. The dye-filled
vesicle fractions were diluted 2000 times with buffer 2 (final lipid
concentration: ∼5.0 mM). This suspension (90 μL) was subsequently
mixed with polymer stock solutions (10 μL) on a 96-well black
microplate (Greiner, flat bottom). Buffer 2 (10 μL) and Triton-X
(0.1% v/v, 10 μL) were employed as the negative and positive
controls, respectively. After 1 h, the fluorescence intensity in each well
was recorded using the microplate reader (TECAN, Switzerland) with
2.5. Time-Kill and Killing Efficiency Tests. The bacteria were
inoculated and prepared according to the same procedure in the MIC
measurement described above. The samples were treated with the
polymer at various concentrations (0, 1/2 × MIC, MIC, and 2 ×
MIC), and were incubated at 37 °C under constant shaking of 100
rpm. At regular time intervals (0, 0.5, 1, 2, 4, 6, 8, and 18 h), the
bacteria samples were taken out from each well for a series of 10-fold
dilutions. Twenty μL of the diluted bacterial solution was streaked
onto an agar plate (LB Agar from 1st Base). The plate was incubated
for 24 h at 37 °C and counted for colony-forming units (CFU). For
8800
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Table 1. Antimicrobial (MIC, μg mL−1) and Haemolytic (HC50, μg mL−1) Activities of Homopolymers with Varying Cationic
Group Structures
MIC (μg mL−1
)
selectivity (HC50/MIC)
polymer
R1 or R2
DP
E. coli
S. aureus
P. aeruginosa
HC50 (μg mL−1
)
E. coli
S. aureus
P. aeruginosa
pMethyl_20
pEthyl_20
pButyl_20
pHexyl_20
pCyhexyl_20
pOctyl_20
pBenzyl_20
pPyr_20
methyl
ethyl
22.5
23.6
23.8
23.5
23.9
20.7
22.7
21.2
20.1
23.4
125
125
15.6
7.8
62.5
31.3
3.9
>500
>500
>500
62.5
>4000
>4000
>4000
15.6
>32
>32
>256
2
>64
>128
>1026
4
<8
<8
<8
butyl
hexyl
3.9
0.25
<0.5
0.06
cyclohexyl
octyl
15.6
7.8
>500
250
16
32
a
a
a
62.5
15.6
15.6
15.6
7.8
31.3
125
7.8
0.12
0.25
32
benzyl
7.8
7.8
3.9
3.9
125
125
250
16
64
2
8
pyridine
1000
>1000
250
128
>256
64
pMeImi_20
pBuImi_20
1-methyl imidazole
1-butyl imidazole
>500
62.5
>64
32
<2
4
a
Poorly soluble in nutrient broth; measured as finely ultrasonic-distributed suspensions.
excitation and emission wavelengths of 490 and 515 nm, respectively.
The percentage of leaked calcein dye in each well was determined as
follows: leakage (%) = [(F − F0)/(FTX − F0)] × 100% where F is the
fluorescence intensity recorded in the well, F0 is the intensity in the
negative control well, and FTX is the intensity in the positive control
well.
virtue of its chemical structure (e.g., pyridine). As a
consequence of the reactive benzylic center, the reaction was
found to be remarkably efficient in all the cases where near-
quantitative (>99%) degree of quaternization was typically
achieved, as determined by 1H NMR spectroscopy (Figure S1).
Such high extent of quaternization would indubitably be an
attractive synthetic approach for the facile tuning of the
polymers’ composition and amphiphilicity, so as to effectively
investigate their influence on antimicrobial and hemolytic
activities.
As a means to characterize the hydrophobic/hydrophilic
balance of polycarbonates, water-octanol partition coefficients
were investigated. Representative polymers conjugated with a
highly fluorescent dansyl group at the chain end were prepared
in a similar fashion to that described above using a modified
hydroxyl-terminated dansyl as the initiator (Scheme S1). The
wavelengths of maximum absorbance and fluorescence
emission of the polymers in methanol ranged from 332 to
336 nm and 519−520 nm respectively (Table S1), and are
comparable to those of free dansyl and dansyl-conjugated
proteins reported in literature.17,18 These data indicated that
the conjugation to these polymers did not affect or diminish the
absorbance and emission properties of the dansyl fluorophore.
For clarity, the polymers are labeled according to the
structure of their respective quaternary ammonium centers and
their respective composition, as well as the degree of
polymerization. As a representative example, pButyl_20
denotes a homopolymer containing 20 repeating units
quaternized with N,N-dimethylbutylamine, while pBu-
tyl0.5Benzyl0.5_20 denotes a polymer of similar chain length
containing two dissimilar quaternary ammonium centers (i.e.,
N,N-dimethylbutylammonium and N,N-dimethylbenzylammo-
nium) randomly distributed along the polymeric backbone in a
final molar ratio of 1:1.
3. RESULTS AND DISCUSSION
3.1. Monomer and Polymer Synthesis. In light of
designing antimicrobial polymers with hydrophobic pendent
groups bearing synthetically accessible and reactive moieties
toward various tertiary, cyclic as well as heterocyclic amines for
facile postpolymerization quaternization, a carbonate monomer
containing a benzyl chloride functional group was synthesized
using a previously established route,15 as shown in Scheme 1.
Metal-free organocatalytic ring-opening polymerization (ROP)
of the functional carbonate was subsequently employed to yield
homopolymers with varying degrees of polymerization (20−
60). Owing to the living nature and highly controlled ability of
organocatalytic ROP, the homopolymers were obtained with
respective lengths that were dictated by initial monomer:-
initiator feed ratios, as ascertained by comparing the integrated
intensities of the relevant 1H NMR resonances from the
terminal initiator (4-methyl benzyl alcohol) relative to the side
chains (Figure S1 in the Supporting Information). All polymers
exhibited narrow molecular weight distribution with a
polydispersity index ranging between 1.2 and 1.3, which was
determined by gel permeation chromatography prior to
postpolymerization quaternization. The metal-free nature of
the catalysts and low catalyst loadings (5 mol %) are also
anticipated to circumvent potential cytotoxic issues typically
associated with catalyst residues.
As amphiphilic balance has been shown to play a pivotal role
in the antimicrobial and hemolytic properties, a series of
polymers was synthesized by reacting the precursor polymer
with a variety of quaternizing agents in a simple and facile
quaternization reaction performed under ambient conditions
(Scheme 1). An exception was made for the reaction with
pyridine as the quaternizing agent most likely owing to the
slightly reduced nucleophilicity of the nitrogen center, which
thus led to the need for slight heating (40 °C) during
quaternization. It was anticipated that the amphiphilicity of the
resultant polymers could be tuned by simple modifications of
the cationic group structures either by varying the length of the
alkyl chains of homologous tertiary (i.e., N,N-dimethylalkyl-
amines) and heterocyclic amines (i.e., 1-alkylimidazoles) or by
3.2. Antimicrobial and Hemolytic Activities of
Homopolymers. The antimicrobial activity of all polymers
in this study was probed against three representative clinically
relevant bacterial strains, namely S. aureus (Gram-positive), E.
coli and P. aeruginosa (Gram-negative). By employing a broth
microdilution assay, minimal inhibitory concentrations (MICs)
were taken at the lowest polymer concentrations required for
the inhibition of bacterial growth after 18 h of incubation with
an initial inoculum size of 3 × 105 CFU/mL. Relevant polymer
toxicity against erythrocytes was also evaluated via hemolysis
assays using rat red blood cells to determine the HC50, i.e.,
8801
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Figure 1. Antimicrobial and hemolytic activities for polymer series containing homologous linear alkyl chain (pMethyl_20 to pOctyl_20),
pCyhexyl_20, pBenzyl_20, and pPyr_20. Bars labeled with a “>” symbol indicate that the MIC or HC50 was greater than the highest polymer
concentration tested (500 μg mL−1 for MIC and 4000 μg mL−1 for HC50). Through an extensive structure−activity study, a highly efficacious and
nonhemolytic polymer, pButyl_20, was achieved. Remarkably, it exhibited a high selectivity ratio of an unprecedented value (>1026 against S.
aureus) for biodegradable synthetic polymers.
concentration at which 50% of red blood cells are lysed upon
exposure to the polymer.
homopolymers bearing a variety of quaternary ammonium
centers (Table 1) and at a fixed uniform chain length of about
20 repeating units. By comparing the MICs and HC50 for
different polymers, it appeared evident that the length of the
alkyl substituent in homologous quaternary centers had a
significant bearing on the antimicrobial and hemolytic activities.
For the series of polymers quaternized with N,N-dimethyl
alkylamines of varying alkyl chain lengths, the potency of the
polymers ameliorated with increasing alkyl chain length while
maintaining its nonhemolytic properties, until it reached a
threshold in terms of hydrophobicity imparted by the alkyl
chain wherein the polymers started to lose its initial selectivity
and became gradually biocidal (i.e., nonselective) in nature. To
elaborate the observed trend, the findings are collectively
summarized in Figure 1. As the number of carbon atoms in the
alkyl substituent increased from pMethyl_20 to pButyl_20, the
antimicrobial activities of the polymers, particularly toward E.
coli and S. aureus, increased as much as 8-fold (125 μg mL−1 to
15.6 μg mL−1) and 16-fold (62.5 μg mL−1 to 3.9 μg mL−1)
respectively, while concomitantly retaining its nonhemolytic
properties (HC50 > 4000 mg/L). This optimum in MIC and
HC50 values consequently gave rise to a remarkably efficacious
antimicrobial polymer with a measured selectivity (HC50/MIC)
of more than 256 against E. coli and 1026 against S. aureus
respectively. To the best of our knowledge, these are the
highest selectivity values reported in the literature for
biodegradable synthetic polymers. Furthermore, the MIC
values for E. coli and S. aureus are comparable to and in
some cases lower than many highly effective naturally occurring
antimicrobial peptides (e.g., cecropin A and B, magainin 1 and
3.2.1. Effect of Molecular Weight. As a preliminary study,
the influence of molecular weight on antimicrobial and
hemolytic activities was evaluated with a series of polycar-
bonates bearing a common quaternary ammonium center
where R1 = methyl (i.e., pMethyl_20, pMethyl_30, pMeth-
yl_40, and pMethyl_60) (Table S2). Changes in molecular
weight over the investigated range did not result in significant
changes in hemolytic activities (HC50 > 1000 mg/L for all
polymers), whereas they did influence antimicrobial activities
depending upon the type of bacteria strains. While the
polymers were seen to be relatively inactive (nonbiocidal and
nonhemolytic) against P. aeruginosa across the range of
molecular weights tested, a modest extent of antimicrobial
activity and selectivity was seen for the polymers against E. coli
and S. aureus. Interestingly, the polymers demonstrated an
apparent correlation between molecular weight and antimicro-
bial activity against the Gram-positive S. aureus, wherein the
antimicrobial activity was observed to increase as the molecular
weight decreased from 23 800 g mol−1 (pMethyl_60) to 8170 g
mol−1 (pMethyl_20). This may be ascribed to the “sieving
effect” as previously reported by Lienkamp et al.,5 which
accounts for such a loss in antimicrobial activity as higher
molecular weight polymers are trapped by the dense outermost
peptidoglycan layer of S. aureus.
3.2.2. Effect of Cationic Group Structure. Having
established the optimal polymer chain length, the role of the
cationic group structure on the antimicrobial and hemolytic
activities was subsequently evaluated using a series of
8802
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
2, and defensing,3,19−22) which also typically exhibit pro-
nounced hemolytic activities. As the alkyl substituent length
further increased to six carbons (pHexyl_20), the MIC against
E. coli and S. aureus remained relatively constant while for P.
aeruginosa, it decreased significantly to 62.5 μg mL−1.
Notwithstanding the lowest MIC values observed with the
hexyl substituent, the dramatic increase in its hemolytic activity
(HC50 = 15.6 μg mL−1) and resultant biocidal properties
precluded its usefulness for therapeutic applications. A similar
extent of nonselective biocidal activity was also seen for the
octyl-substituted polymer pOctyl_20, albeit its poor solubility
in the nutrient broth.
From Figure 1, additional intriguing findings pertaining to
the influence of cationic group structure on the polymers’
antimicrobial and hemolytic activities are revealed. While the
polymers pHexyl_20 and pCyhexyl_20 contain the same
number of pendent carbon atoms in their respective alkyl
substituent, the cyclic homologue was apparently much less
hemolytic (HC50 = 250 μg mL−1) relative to pHexyl_20 despite
having a comparably similar extent of potent antimicrobial
activity, particularly toward E. coli and S. aureus. Interestingly,
polymers containing aromatic substituents in their quaternary
ammonium centers exhibited an analogous trend as well. The
polymer pBenzyl_20 showed similar extent of antimicrobial
and hemolytic activities relative to pCyhexyl_20, with the
notable exception of its stronger activity against P. aeruginosa
(MIC = 125 μg mL−1 vs >500 μg mL−1). Similarly, pPyr_20
exerted considerable antimicrobial potency against the
microbes tested while being less hemolytic than linear alkyl
chains of comparable number of carbons (pHexyl_20). These
results thus led us to posit that the properties of fine structures
and conformations (aliphatic cyclic and aromatic vs linear alkyl)
does have a profound influence on the observed biological
activities. Such observations could be putatively attributed to
the cyclohexyl, benzyl and pyridinium groups being fixed in an
extended conformation, consequently limiting the insertion of
the respective substituents into the hydrophobic lipid region of
the cellular membrane, relative to the polymers bearing flexible
and linear alkyl groups which therefore render them more
hemolytic in nature.
ultimately causes cell death. For similar reasons in each polymer
series, the HC50 decreased accordingly with increasing
hydrophobicity imparted by the alkyl chain length. Much akin
to native antimicrobial peptides,3,23 a general comparison of
polymers across each series in Table 1 highlights the delicate
balance of hydrophobicity and hydrophilicity required to
develop effective macromolecular antimicrobials with favorable
biocompatibility toward mammalian cells.
3.3. Antimicrobial and Hemolytic Activities of
Homopolymers with Two Different Quaternary Ammo-
nium Centers. To broaden the extent of the polymers’
antimicrobial activity toward the highly opportunistic and
difficult-to-kill Gram-negative P. aeruginosa without compro-
mising much of its nonhemolytic properties, a series of
homopolymers bearing a mixture of two different quaternary
ammonium centers randomly distributed along the polymer
chain was designed and synthesized as shown in Scheme 1. It
was envisaged that the low hemolytic activity of pButyl_20 and
potent antimicrobial activities of pBenzyl_20, pPyr_20 and
pBuImi_20 across the range of microbes tested, would allow for
the concomitant installation of dissimilar quaternary ammo-
nium centers on a single polymeric chain as a facile synthetic
strategy for optimizing activity and selectivity. As the precursor
polymer remains essentially the same as that employed for the
homopolymerization studies, such a synthetic route rendered
the various polymeric compositions to be easily explored
without compromising narrow polydispersities, in contrast to
statistical copolymerization and polycondensation ap-
proaches.24 From Figure 2, Table 1 and Table S3, it was
shown that the polymer pButyl0.5Benzyl0.5_20 with N,N-
dimethylbutylammonium and N,N-dimethylbenzylammonium
groups randomly distributed in a final molar ratio of 1:1
effectively alleviated the hemolytic activity previously observed
in the comparatively more hydrophobic pBenzyl_20, con-
To shed further light on investigating the role of cationic
group structure, two different polymers containing homologous
1-alkylimidazolinium moieties were evaluated for their
corresponding antimicrobial and hemolytic activities. Interest-
ingly, the polymer pMeImi_20 which contains 1-methylimida-
zolinium as the cationic group, was found to exhibit high
selectivity toward E. coli and S. aureus (Table 1). As the alkyl
substituent chain length increased to four carbon atoms, the
polymer became more potent toward the pathogens in general,
even against the elusive P. aeruginosa. However, this apparent
improvement in antimicrobial activity was concomitantly
accompanied by an undesirable increase in its hemolytic
activity as well, thus causing pBuImi_20 to exhibit a weak
extent of selectivity against the microbes.
The results from these studies have effectively demonstrated
the cogent influence of nuanced structural changes on the
biological activities of these amphiphilic polymers. The
observed increase in activity with increasing alkyl substituent
chain length suggested that hydrophobicity is a pivotal
determinant in modulating the antimicrobial potency of the
polymers. Increasing the hydrophobicity augments the
propensity of the polymers to bind onto the lipid membrane,
which in turn leads to pronounced membrane disruption and
Figure 2. Effect of cationic group composition on the antimicrobial
and hemolytic activities of polymers with varying molar compositions
of N,N-dimethylbutylammonium and N,N-dimethylbenzylammonium
groups. Bars labeled with a “>” symbol indicate that the MIC or HC50
was greater than the highest polymer concentration tested (500 μg
mL−1 for MIC and 4000 μg mL−1 for HC50). The results revealed that
50% of N,N-dimethylbutylammonium group content was sufficient to
alleviate hemolytic activity by 3 times, consequently enhancing its
potency against a wide spectrum of microbes including the highly
opportunistic and elusively hard-to-kill Gram-negative P. aeruginosa.
8803
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Table 2. Antimicrobial (MIC, μg mL−1) Activity and Selectivity of Polymers against Clinically Isolated Nosocomial Microbes
MIC (μg mL−1
)
selectivity (HC50/MIC)
polymer
MRSA
VRE
A. baumannii
C. neoformans
HC50 (μg mL−1
)
MRSA
VRE
A. baumannii
C. neoformans
pMethyl_20
pButyl_20
250
7.8
7.8
3.9
31.3
3.9
500
31.3
31.3
7.8
>4000
>4000
250
>16
>513
32
>128
>1026
128
>8
>64
16
>128
>128
32
62.5
15.6
31.3
pBenzyl_20
1.95
3.9
pButyl0.5 Benzyl0.5_20
7.8
750
192
192
24
96
Figure 3. Killing efficiency and bactericidal kinetics of polymers. (a) Fractional cell survival of S. aureus, E. coli and P. aeruginosa after 18h incubation
with pButyl_20 and pButyl0.5Benzyl0.5_20 at various concentrations (0, 1/2 × MIC, MIC and 2 × MIC) (b) Bactericidal kinetics of S. aureus and E.
coli after 0.5, 1, 2, 4, 6, 8, and 18 h incubation with pButyl_20. Taken together, the results evidently demonstrate that the mode of antimicrobial
activity incited by the polymers is bactericidal, rather than just bacteriostatic.
sequently augmenting the selectivity against E. coli, S. aureus,
and, most importantly, P. aeruginosa. Notably, 50% of N,N-
dimethylbutylammonium group content was sufficient to
increase the HC50 by 3 times, which consequently afforded a
broad-spectrum antimicrobial polymer with favorable anti-
microbial potency and selectivity (Table 1). Notwithstanding
the successful attempt to ameliorate its biocompatibility (HC50
> 1000 μg mL−1), a further increase in the molar ratio of N,N-
dimethylbutylammonium group content for pBu-
tyl0.75Benzyl0.25_20 however resulted in an apparent loss of
activity toward P. aeruginosa. From Table S3 and Table 1 a
similar observation was made for the polymers pBu-
tyl0.5Pyr0.5_20 and pButyl0.75Pyr0.25_20, wherein the introduc-
tion of N,N-dimethylbutylammonium groups significantly
improved their hemolytic activities but was not adequate to
exhibit observable potency toward P. aeruginosa. In contrast,
while the polymers bearing N,N-dimethylbutylammonium and
1-butylimidazolium groups (pButyl0.5BuImi0.5_20 and pBu-
tyl0.75BuImi0.25_20) retained their broad-spectrum and potent
activities against the tested microbes, they were still generally
hemolytic in nature, with minimal differences relative to
pBuImi_20. The polymers from this study are excellent
examples for demonstrating the dramatic effect of fine-tuning
the hydrophibic/hydrophilic balance (i.e., amphiphilicity) via a
synthetically facile method, which in turn allows for the
development of efficacious antimicrobial polymers (pBu-
tyl0.5Benzyl0.5_20) with broad-spectrum activity and favorable
selectivity.
clinically isolated nosocomial microbes (Table 2). Similar to
the in vitro results against E. coli, S. aureus, and P. aeruginosa,
pButyl_20 showed superior extent of selectivities against a wide
spectrum of pathogens in comparison to pMethyl_20 despite
both polymers being nonhemolytic. Notably, pButyl_20
exhibited excellent antimicrobial potency and selectivity toward
methicillin-resistant S. aureus (MRSA, Gram-positive, HC50/
MIC > 513), vancomycin-resistant Enterococcus (VRE, Gram-
positive, HC50/MIC > 1026) as well as carbapenem-resistant
Acinetobacter baumannii (Gram-negative, HC50/MIC > 64).
Furthermore, the polymers were also shown to possess
efficacious antifungal activities against fluconazole-resistant
Cryptococcus neoformans, which are generally comparable to
their antibacterial activity. A further survey of the results also
revealed that pButyl0.5Benzyl0.5_20 retained most of its
antimicrobial activity (relative to pBenzyl_20) toward the
range of nosocomial microbes tested while having its hemolytic
activity ameliorated by about 2-fold. This further substantiates
the synthetic feasibility of tuning the amphiphilic balance so as
to allow the preparation of potent and selective antimicrobial
polymers. These promising results using clinical isolates of
drug-resistant microbes thus suggest excellent potential of
pButyl_20 and pButyl0.5Benzyl0.5_20 for use as antimicrobial
agents in future clinical applications.
3.5. Mechanistic Studies. 3.5.1. Bactericidal Mechanism
and Kinetics. To study the antimicrobial mechanism of the
polymers, colony counting assays via a traditional surface
plating method were performed. Each bacterial strain was
treated using the most potential candidates, which were
identified as the most potent and selective (i.e., pButyl_20
for E. coli and S. aureus, pButyl0.5Benzyl0.5_20 for P. aeruginosa).
As shown in Figure 3a, at both the MIC and 2xMIC
3.4. Antimicrobial Activity against Clinically Isolated
Microbes. To assess the potential of these polymers toward
clinical applications as broad-spectrum therapeutics, the
antimicrobial activity was investigated against a series of
8804
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Figure 4. Field emission scanning electron microscopy (FE-SEM) images of E. coli (a, b) and S. aureus (c, d) before (a, c) and after (b, d) 2 h
treatment with pButyl_20 at 2 × MIC. Compared with the intact controls, the treated E. coli cells exhibited distorted and corrugated surfaces.
Significant fusion of bacterial membrane and debris were observed for S. aureus cells.
Figure 5. Extent of calcein efflux in neutral vesicles (DOPC) and negatively charged vesicles (4:1 DOPE/DOPG) after treatment with representative
polymers for 1 h. In general, a good correlation between leakage and selectivity was observed, wherein nonhemolytic polymers such as pMethyl_20
and pButyl_20 induced little or negligible extent of dye leakage for neutral vesicles while relatively more hydrophobic polymers such as pHexyl_20
and pOctyl_20 demonstrated a nonselective behavior toward both type of vesicles.
concentrations of the respective polymers, more than 99.9%
killing efficiency (i.e., 3-log reduction of the initial inoculum)
was achieved, strongly suggesting its definitive bactericidal
mechanism. Further substantiation of this antimicrobial
mechanism was evidently shown from time-kill experiments
using pButyl_20 as a model polymer against E. coli and S.
aureus over an 18-h period. From Figure 3b, pButyl_20
displayed expeditious bactericidal effect against the pathogens:
more than 3-log reduction in terms of viable colonies was
observed within 4 h of treatment, at the respective lethal
concentrations for each bacterial strain. Taken together, these
results conferred definitive evidence for indicating the polymers
to be bactericidal, rather than being merely bacteriostatic (i.e.,
growth inhibiting) in nature.
3.5.2. Membrane-Lytic Studies with FE-SEM. Using field
emission scanning electron microscopy (FE-SEM), the
antimicrobial mechanism of the polymers was further probed
through observation of the pathogen cell morphologies upon
polymer treatment. With reference to Figure 4, the microbial
cells (E. coli and S. aureus) treated with pButyl_20 as a model
polymer (at a lethal dose well above its MIC) exhibited distinct
and stark changes in their morphology when compared to their
corresponding intact controls. In contrast to the smooth
surfaces of the control cells, distorted and wrinkled cell walls
8805
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Figure 6. Correlation studies between water-octanol partition coefficient (log P) of representative polymers with their respective antimicrobial
(MIC) and hemolytic (HC50) activities. Bars labeled with a “>” symbol indicate that the HC50 was greater than the highest polymer concentration
tested (i.e., 4000 μg mL−1).
were clearly observed upon polymer treatment, particularly for
E. coli cells (Figure 4a,b). Some S. aureus cells were also seen to
be damaged to such an extent that the cell structure wholly
collapsed and consequently had their cell contents leaked out
and membranes fused together to form distinctive groups of
cell debris (Figure 4c,d). These results gathered from FE-SEM
images collectively indicate that the polymer caused severe cell
wall and membrane lysis via physical means, analogous to
antimicrobial peptides.3,23
3.5.3. Membrane-Lytic Mechanism Studied Through Dye
Leakage Assay. In order to assess the ability of the polymers to
disrupt the integrity of model membranes, we monitored the
leakage of an encapsulated fluorescent dye (calcein) from
within negatively charged (4:1 DOPE/DOPG) and neutral
(DOPC) large unilamellar vesicles, which mimic Gram-negative
bacteria and mammalian cell membranes respectively, upon
polymer treatment. Although these assays are known to
underestimate certain factors such as cell walls and lip-
opolysaccharides in bacterial cell membranes, they are none-
theless deemed to be useful in providing biophysical insights
underlying the membrane disrupting abilities of macro-
molecular antimicrobial materials.19,24 These assays were
therefore employed to investigate the overall membrane
disruption activities of polymers.
As shown in Figure 5, pMethyl_20 and pButyl_20 were
relatively inactive or showed little disruption against neutral
vesicles while exhibiting observable activity against negatively
charged vesicles, demonstrating a good correlation between
leakage and selectivity. A strong agreement between leakage
and in vitro results was further seen for the nonselective and
biocidal polymers pHexyl_20 and pOctyl_20, which was found
to induce almost similar extent of leakage toward negatively
charged and neutral vesicles. Similarly, the extent of dye leakage
demonstrated by pCyhexyl_20 and pBenzyl_20 was higher
from the negatively charged vesicles as compared to the neutral
vesicles, which is in agreement with the good selectivity
determined from in vitro studies. The general trend gathered
from these dye leakage assays provides further substantiation to
our postulation wherein amphiphilic balance as well as nuanced
structural factors (e.g., alkyl chain length and cationic group
structure) are equally important determinants of membrane-
lytic potency and selectivity.
3.5.4. Water−Octanol Partition Coefficients (log P). It has
been widely established that the hydrophobic/hydrophilic
balance serves as a pivotal structural determinant for the
design of membrane active and biocompatible polymers.
Despite its significance, a majority of studies concerning the
overall hydrophobicity of amphiphilic antimicrobial polymers
has been typically based on qualitative determinants, such as
molar ratio of monomers in copolymers and length of alkyl side
chains. To gain a better understanding of how the overall
hydrophobicity of these polymers affect their biological activity,
we sought to determine the water-octanol partition coefficient
(log P), where P is defined as the ratio of polymer
concentrations in the octanol and water phases at equilibrium.
Polymer concentration was measured based on the fluorescence
properties of the dansyl-conjugated polymers. When log P is
lower than 0, a polymer is considered to be relatively
hydrophilic, while the polymer is hydrophobic if log P is
above 0. While log P values are conventionally useful in
structure−activity relationship studies for small molecules and
drugs,25 it is expected that it would provide an equally germane
approach as a quantitative measure of the polymers’ overall
hydrophobicity or amphiphilic balance.
From Figure 6, the log P values of the representative dansyl-
labeled polymers appeared to correlate well with its
corresponding biological activity. In general, it was shown
that as the alkyl substituent of the homopolymers increased in
chain length (i.e., from methyl to hexyl groups), their respective
log P gradually increased from −1.46 to +0.05 as the polymers
became concomitantly more potent toward the microbes. It was
interesting to note that the log P of pButyl_20 (−1.20)
increased only slightly from that of pMethyl_20 (−1.46),
suggesting the predominantly hydrophilic nature of the
polymers albeit the relatively significant increase in the alkyl
chain length. As the alkyl chain increased by two carbon atoms
to pHexyl_20, the log P dramatically increased 24-fold to 0.05.
This similar behavior in hemolytic activity as observed earlier
8806
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807

Macromolecules
Article
Author Contributions
for the homopolymers in Table 1 led us to ascribe the
pronounced toxicity of pHexyl_20 as a highly plausible
consequence of the increase in its overall hydrophobicity as
reflected in its log P value. Another intriguing finding stems
from the polymers pBenzyl_20, wherein the log P value of
−0.85 is seen to be relatively hydrophilic in comparison to
pHexyl_20, which might in turn attribute to its lower hemolytic
activity.
The observed log P values have effectively demonstrated that
the overall hydrophobic/hydrophilic balance is not a simple
function of the individual monomers’ alkyl substituent chain
length or the cationic group structure, but also depends on the
polymeric structure as a whole (“global amphiphilicity”). This is
in accord with our previous hypothesis as evidenced through
the dramatic effects these subtle yet pertinent determinants
have on the biological activities of the polymers.
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was financially supported by the Institute of
Bioengineering and Nanotechnology (Biomedical Research
Council, Agency for Science, Technology and Research,
Singapore), and IBM Almaden Research Center (USA). W.C.
acknowledges the support of National University of Singapore
(NUS) Graduate School for Integrative Sciences & Engineering
through a Ph.D. scholarship.
REFERENCES
■
4. CONCLUSION
(1) Taubes, G. Science 2008, 321, 356−361.
(2) Boucher, H. W.; Talbot, G. H.; Bradley, J. S.; Edwards, J. E.;
Glibert, D.; Rice, L. B.; Scheld, M.; Spelberg, B.; Bartlett, J. Clin. Infect.
Dis. 2009, 48, 1−12.
A series of biodegradable and amphiphilic antimicrobial
polycarbonates was synthesized via metal-free organocatalytic
ring-opening polymerization. By means of a facile and efficient
post-polymerization quaternization reaction, the hydrophobic/
hydrophilic balance was easily controlled by varying the cationic
group structure as well as their respective compositions along
the polymeric chain. Nuanced modifications to the hydro-
phobic character of the cationic amphiphilic polymers were
demonstrated to dramatically influence the antimicrobial and
hemolytic activities, which consequently afforded the polymer
pButyl_20 as a highly efficacious (MICs ranging from 3.9 to
62.5 μg mL−1 against clinically isolated drug-resistant Gram-
positive and Gram-negative bacteria as well as fungi) and
nonhemolytic (HC50 > 4000 μg mL−1) antimicrobial polymer.
In an effort to ameliorate its antimicrobial activity while
retaining its nonhemolytic character, polymers bearing a
mixture of two dissimilar quaternary ammonium centers were
systematically prepared so as to fine-tune their amphiphilic
balance. It was successfully demonstrated that 50% of N,N-
dimethylbutylammonium group content was sufficient to
alleviate hemolytic activity by 3 times, consequently enhancing
its potency against a wide spectrum of microbes including the
highly opportunistic and elusively hard-to-kill Gram-negative P.
aeruginosa. Through a series of mechanistic studies, the
polymers are demonstrated to be expeditiously bactericidal in
nature, with a membrane-lytic mechanism deemed to be
operative. It is envisaged that these excellent broad-spectrum
and highly selective antimicrobial polymers hold great potential
for use as promising antimicrobial agents in combating many
challenging pathogenic infection.
(3) Zasloff, M. Nature 2002, 415, 389−395.
(4) Marr, A. K.; Gooderham, W. J.; Hancock, R. E. W. Curr. Opin.
Pharmacol. 2006, 6, 468−472.
(5) Lienkamp, K.; Tew, G. N. Chem.Eur. J. 2009, 15, 11784−
11800.
(6) Kenawy, E.-R.; Worley, S. D.; Broughton, R. Biomacromolecules
2007, 8, 1359−1384.
(7) Munoz-Bonilla, A.; Fernan
́
dez-García, M. Prog. Polym. Sci. 2012,
̃
37, 281−339.
(8) Sovadinova, I.; Palermo, E. F.; Huang, R.; Thoma, L. M.; Kuroda,
K. Biomacromolecules 2011, 12, 260−268.
(9) Lienkamp, K.; Madkour, A. E.; Musante, A.; Nelson, C. F.;
Nesslein, K.; Tew, G. N. J. Am. Chem. Soc. 2008, 130, 9836−9843.
(10) Nederberg, F.; Zhang, Y.; Tan, J. P.; Xu, K.; Wang, H.; Yang, C.;
Gao, S.; Guo, X. D.; Fukushima, K.; Li, L.; Hedrick, J. L.; Yang., Y. Y.
Nat. Chem. 2011, 3, 409−414.
(11) Qiao, Y.; Yang, C.; Coady, D. J.; Ong, Z. Y.; Hedrick, J. L.; Yang,
Y. Y. Biomaterials 2012, 33, 1146−1153.
(12) Palermo, E. F.; Sovadinova, I.; Kuroda, K. Biomacromolecules
2009, 10, 3098−3107.
(13) Engler, A. C.; Wiradharma, N.; Ong, Z. Y.; Coady, D. J.;
Hedrick, J. L.; Yang, Y. Y. Nano Today 2012, 7, 201−222.
(14) Sambhy, V.; Peterson, B. R.; Sen, A. Angew.Chem. Int. Ed. 2008,
47, 1250−1254.
(15) Pratt, R. C.; Nederberg, F.; Waymouth, R. M.; Hedrick, J. L.
Chem. Commun. 2008, 114−116.
(16) Pratt, R. C.; Lohmeijer, B. G. G.; Long, D. A.; Pontus Lundberg,
P. N.; Dove, A. P.; Lo, H.; Wade, C.; Waymouth, R. M.; Hedrick, J. L.
Macromolecules 2006, 39, 7863−7871.
(17) Chen, R. F. Arch. Biochem 1968, 25, 412−416.
(18) Chen, R. F. Arch. Biochem. Biophys. 1968, 128, 163−175.
(19) Oren, Z.; Shai, Y. Biochemistry 1997, 36, 1826−1835.
(20) Liu, D. H.; DeGrado, W. F. J. Am. Chem. Soc. 2001, 123, 7553−
7559.
(21) Schmitt, M. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc.
2004, 126, 6848−6849.
(22) Liu, D. H.; Choi, S.; Chen, B.; Doerksen, R. J.; Clements, D. J.;
Wrinkler, J. D.; Klein, M. L.; DeGrado, W. F. Angew. Chem., Int. Ed.
2004, 43, 1158−1162.
ASSOCIATED CONTENT
■
S
* Supporting Information
General sysnthesis scheme for dansyl-conjugated polycarbon-
1
ates, H NMR spectrum and data of representative polymers,
characterization data of dansyl-conjugated polycarbonates, and
additional data on biological activity of polycarbonates. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(23) Andreu, D.; Rivas, L. Biopolymers 1998, 47, 415−433.
(24) Tew, G. N.; Liu, D. H.; Chen, B.; Doerksen, R. J.; Kaplan, J.;
Carroll, P. J.; Klein, M. L.; DeGrado, W. F. Proc. Natil. Acad. Sci. 2002,
99, 5110−5114.
(25) Leo, A. J.; Hansch, C. Perspect. Drug. Discovery Des. 1999, 17, 1−
25.
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: yyyang@ibn.a-star.edu.sg (Y.Y.Y.).
*E-mail: hedrick@us.ibm.com (J.L.H.).
8807
dx.doi.org/10.1021/ma4019685 | Macromolecules 2013, 46, 8797−8807