Modulating bioactivities of primary ammonium-tagged antimicrobial aliphatic polycarbonates by varying length, sequence and hydrophobic side chain structure

Kazuki Fukushima *a, Kohei Kishi a, Keita Saito a, Kazuki Takakuwa a, Shunta Hakozaki a and Shigekazu Yano b
aGraduate School of Organic Materials Science, Yamagata University, 4-3-16 Jonan, Yonezawa, Yamagata 992-8510, Japan
bGraduate School of Science and Engineering, Yamagata University, Japan

Received 19th March 2019 , Accepted 9th April 2019

First published on 10th April 2019

Cationic aliphatic polycarbonates bearing primary ammonium side chains have been developed with relatively high molecular weights and controlled macromolecular architectures. These polycarbonates exhibit reasonable antimicrobial activity against Gram-negative and Gram-positive bacteria. The prepared homopolymers could be effective against Gram-negative bacteria whose growth is usually inhibited by copolymers with hydrophobic comonomer units when quaternary ammonium salts (QAS) are used at the cationic side chains. A methoxyethyl (ME) side chain was explored as a comonomer unit for modulating biological activities, besides conventional hydrophobic side chains including ethyl and benzyl groups. In contrast to the ethyl side chain that increases both antimicrobial and hemolytic activities, the ME side chain serves to enhance the antimicrobial activity, but suppresses the hemolytic activity. This could be attributed to the unique characteristics of an aliphatic polycarbonate bearing a ME side chain: hemocompatibility, cell adhesion property, and selective interactions with proteins. The benefits of blood compatibility of the cationic aliphatic polycarbonates with the use of the primary ammonium side chains have been reported for the first time. The polycarbonate main chain is subjected to hydrolysis, which reduces the inherent cytotoxicity of polycations. This hydrolytic property is specific to these primary ammonium-tagged polycarbonates and could be an advantage over previously reported QAS-tagged antimicrobial polycarbonates.


Antibiotics are used for the prevention and treatment of infections in immune-compromised patients. The widespread use and abuse of life-saving antibiotics has led to the emergence of drug-resistant bacteria over the last decade. To treat these life-threatening infections, there is a pressing need for the development of new types of drugs with broad spectrum activity to target drug-resistant infections.

Amphiphilic polycations mimicking host defense peptides (HDPs)1 physically break down negatively charged bacterial cell membrane by electrostatic interaction and do not lead to drug-resistance.2 Nederberg and Fukushima et al. reported first the use of biodegradable polycarbonates with quaternary ammonium salts (QAS) as effective therapeutics against the drug-resistant Gram-positive bacteria, methicillin-resistant Staphylococcus aureus (MRSA).3 Hedrick, Yang, and co-workers have subsequently explored QAS-bearing antimicrobial polycarbonates extensively by modifying comonomers and N-substituents around the cationic nitrogen center and reported several efficient derivatives.4–11 Ong and Fukushima et al. also have found that QAS-bearing polycarbonates containing tertiary amines on the same side chain could serve as a gene delivery carrier.12 Duan, Chan-Park, and co-workers have recently revealed the efficacy of the polycarbonates as antimicrobials.13 One of the key priorities in the discovery of novel antimicrobial polymers is broad-spectrum activity. In particular, Gram-negative bacteria and fungi are sometimes more challenging targets for antimicrobial polymers.14 In general, hydrophobicity and amphiphilic balance of the polymer play a key role in the treatment efficacy against Gram-negative bacteria.15 Cationic polymers with a high hydrophobic content tend to cause lysis of mammalian cells, which results in serious side reactions.16 High selectivity between Gram-negative bacteria and mammalian cells has been achieved in a few QAS-bearing polycarbonates by balancing amphiphilicity using copolymerization and N-substituent modification,4,10 among which only one example was found to be highly selective as a single polymer component.10

The use and regulation of self-assembled structures formed by some cationic block copolymers could be another option to alter antimicrobial activity.17 Rod/fiber-shaped cationic assemblies have demonstrated high antifungal activity.18,19 More fundamentally, cationic micelles have performed better than their homopolymer counterparts in the case of QAS-bearing polycarbonates.3

Most antimicrobial aliphatic polycarbonates employ QAS, pyridinium, and imidazolium salts as cationic side chains, probably due to the relative ease of synthesis with these moieties.13 On the contrary, reversibly N-protonated amines, such as primary and tertiary ammonium salts, have been more generally used for other polymer platforms, including poly(meth)acrylates, poly(vinyl ether)s, and polynorbornenes, in antimicrobial applications.20–23 Extensive studies by Kuroda et al. suggest that N-protonated amines are more appropriate than QAS to balance the high antimicrobial activity and minimum hemolytic activity.23,24 The synthetic routes to aliphatic polycarbonates with primary ammonium side chains have already been established.25,26 However, an application as an antimicrobial has never been reported. In this context, we began the evaluation and use of primary ammonium-tagged polycarbonates for antimicrobial applications.27 Recently, Cai and co-workers demonstrated that the cationic polycarbonates with primary ammonium side chains exhibit high antimicrobial activity against Gram-positive bacteria including MRSA and display minimal hemolytic properties.28 This opens another exciting possibility for cationic polycarbonate antimicrobials. Nevertheless, the molecular weights of the polymers used in their study were quite low (less than 10-mers),28 and their efficacy against Gram-negative bacteria has never been evaluated. Therefore, precise synthesis of primary ammonium-tagged polycarbonates and their substantial evaluation as antimicrobials have not been sufficiently concluded.

We have been acquainted with cyclic carbonate monomer synthesis and controlled polymerization using organocatalysts to obtain high molecular weight aliphatic polycarbonates through several relevant studies.29–35 In this study, we synthesized the primary ammonium-tagged cationic polycarbonates, whose molecular weights were around 10[thin space (1/6-em)]000 g mol−1. Furthermore, several copolycarbonates were also prepared by incorporating different hydrophobic side chain units (Fig. 1). In particular, the introduction of a methoxyethyl (ME) side chain, which has been recently found highly blood compatible, appears promising to mitigate hemolytic activity.34 The antimicrobial activity of the prepared polymers was evaluated to elucidate effects of the hydrophobic side chains on blood compatibility and antimicrobial activity, concerning the sequence and structure. In addition, we first evaluated the hydrolytic property of these primary ammonium-tagged aliphatic polycarbonates that could exhibit different degradation behavior from the QAS-tagged polycarbonates that are stable in a physiological condition.3

image file: c9bm00440h-f1.tif
Fig. 1 Synthesis of cationic-functionalized aliphatic polycarbonates bearing primary ammonium salts.

Materials and methods


2,2-Bis(hydroxymethyl)propionic acid (bisMPA) was purchased from Sigma-Aldrich Japan (Tokyo, Japan). (+)-Sparteine (SP) and bis(pentafluorophenyl carbonate) (PFC) were purchased from Kanto Chemical (Tokyo, Japan). Triphosgene, trifluoroacetic acid, and benzyl alcohol (BnOH) were purchased from Tokyo Chemical Industry (TCI, Tokyo, Japan). Methylene chloride, THF, toluene, and DMF were obtained from a dry solvent supply system (GlassContour), unless specified otherwise. 1-(3,5-Bis(trifluoromethyl)phenyl)-3-cyclohexyl-2-thiourea (TU) was prepared as reported elsewhere.36 BnOH and SP were vacuum-distilled over CaH2 and stored in a nitrogen-filled glove box. TU was dehydrated over CaH2 and stored in the glove box. Other chemicals were purchased from any of the distributors above and were used as received unless stated otherwise.

Phosphate-buffered saline (PBS, pH 7.4) was prepared by dissolving ten tablets of phosphate buffer salts (Takara Bio, Tokyo, Japan) in 1 L of ultrapure water.

Luria–Bertani (LB) broth was formulated by dissolving 2.5 g of yeast extract, 5 g of tryptone, and 2.5 g of sodium chloride in 500 mL of ultrapure water and the subsequent sterilization by an autoclave. All the ingredients were purchased from Nacalai Tesque (Kyoto, Japan). Muller–Hinton (MH) broth was purchased from Sigma-Aldrich (Tokyo, Japan).


1H-NMR spectra were acquired on a JEOL JNM-ECA400 and a JNM-EC500 operated at 400 and 500 MHz, respectively. The residual solvent peaks were used as a reference (CDCl3: 7.27 ppm, DMSO-d6: 2.50 ppm). The number-averaged molecular weight (Mn) and molar-mass dispersity (ĐM) were estimated by size-exclusion chromatography (SEC). The SEC using THF as the eluent was performed at 40 °C using an integrated SEC unit of Malvern Viscotek TDAmax equipped with three TSK-gel columns connected in series (G2000HHR and two GMHHR-H) and a refractive index (RI) detector. The Mn and ĐM were calibrated with polystyrene (PS) standards (ranging from 580 to 2.2 × 106 g mol−1). The SEC eluting DMF containing 0.1 M LiBr was operated by a Tosoh HPLC HLC-8220 system equipped with four TSK-gel columns connected consecutively (α-M, α-4000, α-3000, and α-2500) and RI and ultraviolet (UV) detectors at 40 °C. PS standards ranging from 1050 to 1.09 × 106 g mol−1 were employed for calibration. Hydrodynamic diameters and zeta potentials of polymers were measured by Malvern Zetasizer Nano Series Nano-ZS (Malvern Instrument Ltd) with a predetermined concentration (1.0 mg mL−1) of polymers dissolved in water, PBS, or broth. The measurement was performed at least five times, and the mean diameters and PDIs were calculated.

Synthesis of cationic functionalized polycarbonates

Synthesis of carbonate monomers. All cyclic carbonates were prepared according to previously reported procedures.25,29 MTC-BAE (Fig. 1) was obtained by following a method developed by Sanders et al.26
General ring-opening polymerization (ROP) of cyclic carbonates. All polymerization reactions were performed in CH2Cl2 at room temperature using organocatalysts, TU and SP, in a nitrogen-filled glove box.35 For both homo- and co-polymerizations, 1 or 2 mol% of benzyl alcohol (BA) and 1-pyrenebutanol (PB) relative to the total amount of monomers was used as an initiator. The total concentration of the cyclic carbonates was adjusted as 1.0 or 2.0 M. Acetic anhydride was used for quenching the reaction as well as end capping of hydroxyl end groups of polymers to increase the tolerance of hydrolysis that starts from backbiting.
General procedure for formation of polycarbonate cationic side chains. tert-Butoxycarbonyl (Boc) groups at the side chains were treated with either of CF3CO2H or conc. HCl, aq. (35–37 wt%). The Boc-protected precursor polymers were dissolved in acetonitrile and cooled down to −5 °C by an ice-salt cooling bath. The acid (15 eq. relative to the Boc group) was then gradually added, and the solution was allowed to warm to room temperature and stirred for up to 6 h. After the completion or plateauing of the reaction was confirmed by 1H-NMR, the polymers were precipitated in a mixed solvent of hexane and diethyl ether (1[thin space (1/6-em)]:[thin space (1/6-em)]1) and isolated by centrifugation.

Antimicrobial test

Bacillus subtilis 168 (B. subtilis) and Escherichia coli DH5α (E. coli) were grown in LB or MH broth at 37 °C. Polymers were directly dissolved in ultrapure water, which was previously sterilized using an autoclave, to formulate polymer solutions with concentrations of 8, 16, 32, 64, 128, 250, 500, and 1000 μg mL−1. Poly(ethylene glycol) (PEG: Mn 5000; Sigma-Aldrich) and poly(ethylene imine) (PEI: Mn 70[thin space (1/6-em)]000; TCI) were used as the controls. Bacteria solution was adjusted to an optical density of around 0.2 at 600 nm (O.D.600) by dilution with a broth before the test. In each well of a 96-well plate, 100 μL of the polymer solutions was added followed by 100 μL of the bacteria solution. The mixed samples were then incubated at 37 °C, and the readings of O.D.600 of the mixed samples were measured by a microplate reader (Multiscan GO, Thermo Fischer Scientific) at a predetermined time (0, 2, 4, 6, 8, and 24 h). The tests were repeated at least three times, and the average readings of O.D.600 and standard deviations (SD) were plotted with error bars as a function of time (Fig. 3, S2, and S5). The minimum inhibitory concentrations (MICs) of the polymers were defined as the concentration at which no bacterial growth is observed after 24 h of incubation, which was verified by the absence of increase in the reading of O.D.600 in the graphs.

Hemolysis test

Sheep red blood cells were obtained by centrifugation of sheep whole blood that was purchased from Cosmo Bio (Tokyo, Japan). The red blood cell (RBC) suspension was prepared by dilution with PBS to adjust the concentration to 4% in volume. Aqueous solutions of the polymers were prepared using procedures similar to those used for the antimicrobial test, and the concentrations ranged from 100 up to 4000 μg mL−1. The polymer solutions (100 μL) were placed in each well of a 96-well plate, and the same volume of the RBC suspension (100 μL) was added to each well. The mixing did not lead to hemolysis, which was confirmed using a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 mixture of RBC and PBS as a blank sample. The plates after incubation for 1 hour at 37 °C were then centrifuged at 1000g for 5 minutes. Aliquots (100 μL) of the supernatant were transferred to another 96-well plate, and the hemoglobin release was monitored at 570 nm using a microplate reader. The RBC suspension in PBS was used as negative control. The absorbance of the RBC suspension treated with 1% Triton X-100 (100 μL) was considered 100% hemolysis. PEG and PEI were also used as controls.

Percentage of hemolysis was calculated using the following formula: Hemolysis (%) = [(O.D. in the mixture of polymer and RBC − O.D. in the mixture of PBS and RBC)/(O.D. in RBC treated with 1% Triton X-100 – O.D. in the mixture of PBS and RBC)] × 100, where O.D. is an optical density at 570 nm. The test was at least triplicated, and the average hemolysis and SDs were plotted with error bars as a function of polymer concentration (Fig. 4).

In vitro degradation test

The polymer was dissolved in PBS prepared with D2O (10 mg mL−1). The polymer solution was incubated at 37 °C, and the structure change along with hydrolysis was monitored by 1H-NMR at several time points.

Results and discussion

Polymer synthesis and characteristics

Three types of cationic polycarbonates were synthesized in this study; homopolymers (H1-a, H1-b, and H2), random copolymers (R1–R4), and a diblock copolymer (B1). The characteristics of the cationic polycarbonates and their precursors are summarized in Table 1. Each polymer was successfully formed in a relatively controlled manner, indicated by the relatively low ĐM values of the polymers determined by SEC. For the most ROP, PB or BA was used as an initiator. Since the ROP of these cyclic carbonates often involves concomitant initiation by moisture and impurities in monomers,33,34,37 no initiator was used to obtain a higher molecular weight polymer (H1-b in Table 1). The molecular weights and degrees of polymerization (DPs) of each monomer unit in the polymers were calculated from 1H-NMR. For the block copolymer B1, m and n (Fig. 1) denote the DPs of NH3+ and R in Table 1, respectively. Table 1 shows that the changes in the molecular weights and DPs were minimal before and after deprotection. Thus, the deprotection reaction hardly appears to affect the aliphatic polycarbonate main chain. More than 90% of the Boc groups were removed by deprotection to form primary ammonium salts for all the polymers studied. Most of the cationic polycarbonates were recovered in high yields (>70%), except for R1 and R2 containing the ME side chain. This is attributed to the high miscibility of the ME side chain in a wide range of solvents including the precipitation solvent (Et2O/hexane 1[thin space (1/6-em)]:[thin space (1/6-em)]1). As described later, these polycarbonates are easily subjected to hydrolysis and solvolysis. Thus, all polycarbonates in this study were treated with acetic anhydride to modify hydroxyl end groups that facilitate hydrolysis by a back-biting reaction.29 The characteristic peak of the acetyl end of the polymers appears at ∼2 ppm (h′ and h′′ in Fig. 2). In addition, the deprotection was conducted in an aprotic solvent such as acetonitrile.
image file: c9bm00440h-f2.tif
Fig. 2 1H-NMR spectra of MTC-BAE (A), Boc-H1-a (B), H1-a (C), and R1 (D). (A, B): 400 MHz in CDCl3. (C, D): 500 MHz in DMSO-d6.
Table 1 Characteristics of cationic polycarbonates used in this study
Polymers Precursor (Boc-)b After deprotection
Ia M n (SEC) Đ M M n (NMR)e DPe,f M n (NMR)e DPe,f N+[thin space (1/6-em)]e,g Yield
kg mol−1 kg mol−1 Boc R kg mol−1 NH3+ R % %
a I: initiator, PB: 1-pyrenebutanol, BA: benzyl alcohol. b Polymers with a Boc group. c Determined by SEC (DMF). d Determined by SEC (THF). e Determined by 1H-NMR. f Degree of polymerization. g Conversion for deprotection of the Boc group.
H1-a PB 20c 1.21c 11 36 0 12 36 0 >98 99
H1-b None 31c 1.39c 20 65 0 21 66 0 94 82
H2 PB 18c 1.22c 10 33 0 8.5 34 0 >98 99
R1 BA 7.5d 1.34d 10 19 19 10 19 18 95 51
R2 PB 8.1c 1.32c 9.5 10 28 11 12 30 95 47
R3 BA 8.2d 1.26d 9.0 18 18 8.4 16 17 >98 70
R4 BA 9.5d 1.19d 9.3 18 18 10 18 18 98 79
B1 BA 7.1d 1.30d 10 21 18 10 20 16 >98 82

As found in the 1H-NMR spectra of Boc-H1 and H1 (Fig. 2b and c, respectively), the integral ratios of the signals a′ + d′ (a′′+ d′′), e′ (e′′) and h′ (h′′) were unchanged, confirming that the polycarbonate main chain remained intact after the deprotection of the Boc group with CF3CO2H. These results indicate that the prepared cationic polycarbonates used in this study possess sufficient DPs to be recognized as polymer antimicrobials. The primary ammonium-tagged aliphatic polycarbonates with molecular weights over 10[thin space (1/6-em)]000 g mol−1 have been obtained for the first time in this study.

Antimicrobial activity of cationic polycarbonates against E. coli

The growth curves of E. coli in the presence of different concentrations of polymers as a function of time are shown in Fig. 3. PEI is used as a representative control of conventional cationic polymers. Most of the cationic polycarbonates exhibited reasonable antimicrobial activity against E. coli that were physically disrupted as observed by SEM (Fig. S1). The high O.D. values at t = 0 seem to be attributed to the electrostatic interaction between the polymers and ingredients in a broth. We determined the MIC as the concentration at which the O.D. did not increase after 8 h of the treatment. For instance, the MIC of PEI against E. coli is 250 μg mL−1. The bacterial growth curve also provides information regarding the immediate effect of the polymer antimicrobials. All polymers exhibited the lowest O.D. within 4 h after the treatment with the polymers of concentrations higher than the MIC. In turn, most of the bacteria could be destroyed or at least the bacterial growth is inhibited in such a short time.
image file: c9bm00440h-f3.tif
Fig. 3 Bacterial growth of E. coli treated with different concentrations of polymers as a function of time.

The MIC values of the cationic polycarbonates are summarized in Table 2. The MIC of H1-b was at the same level as that of H1-a (Fig. S2), implying little influence of molecular weight on the antimicrobial activity at the range of more than 10[thin space (1/6-em)]000 g mol−1 in this study. According to the previous studies, the molecular weight dependence on antimicrobial activities was often observed at ranges below 10[thin space (1/6-em)]000 or over 50[thin space (1/6-em)]000 g mol−1.22,38,39H2, which has chloride counter anions, was quite less active against E. coli, as the MIC was determined as more than 500 μg mL−1. This result, along with a few reported results, demonstrates the effect of counter anions on antimicrobial activities.40 Nevertheless, H1 in this study is recognized as a rare case that aliphatic polycarbonate-based cationic polymers present a pragmatic antimicrobial activity against Gram-negative bacteria by a single component.41

Table 2 Biological activities of cationic polycarbonates and PEI control
Polymers MICa/μg mL−1 (μM) Hemolysisb (%)
E. coli B. subtilis
a Minimum inhibitory concentration. b The value at 2000 μg mL−1. c The value at 1000 μg mL−1. d The value at 500 μg mL−1.
H1-a 64 (5.3) 64 (5.3) 16 ± 2.5
B1 64 (6.4) 250 (25.1) 4 ± 0.7
R1 32 (3.2) 32 (3.2) 26 ± 3.3
R2 250 (23.8) 125 (11.9)
R3 16 (1.9) 96 ± 4.3
R4 64 (6.8) 99 ± 5.7c
PEI 250 32 51 ± 7.8d

Effect of non-cationic side chain: sequence and distribution of charged units

Bactericidal mechanisms of cationic polymers include toroidal-pore, barrel stave, and carpet models.42 In either case, hydrophobic moieties play a role in interacting with or inserting in a lipid layer of the cell membrane for the membrane disruption-based bacteriolysis. We designed the cationic copolycarbonates R1, R2, and B1 introducing a hydrophobic comonomer unit with a ME side chain. In our previous study, a high blood compatibility and a high affinity to cells by the ME side chain have been demonstrated.34 Thus, we estimated that the ME side chain could also enhance the affinity to bacterial cells (antimicrobial activity) as well as blood compatibility. Multi-ether side chains such as a methoxyethoxy group and oligo(ethylene glycol), which are ordinarily used as a biocompatible moiety,41,43 are not appropriate for this purpose due to their hydrophilicity and water solubility. R1 and R2 are random copolymers, and B1 is a block copolymer. For a simple comparison, molecular weights of the polymers were uniform to around 10[thin space (1/6-em)]000 g mol−1, and the comonomer ratios were almost 1[thin space (1/6-em)]:[thin space (1/6-em)]1, except for R2 (Table 1). The MIC values against E. coli became the smallest for R1 and the largest for R2. H1-a and B1 showed the same MIC values.

Hydrophobicity and amphiphilic balance are believed to be a key factor for eradicating Gram-negative bacteria such as E. coli.2,15 PEI is a control cationic polymer that contains cationic groups in the polymer backbone, and is therefore quite hydrophilic. The MIC of PEI against E. coli was relatively high (250 μg mL−1), compared to those of the cationic polycarbonates in which the polymer main chain contributes supplementary to hydrophobicity. The further lower MIC of R1 is attributed to the incorporated hydrophobic ME side chains, while the antimicrobial activity of R2 deteriorated due to the decreased composition of the cationic repeating units (∼29 mol%).

Hydrodynamic diameters of the cationic polymers were measured by DLS and are summarized in Table 3 with the zeta potential data. As found in the size distribution charts (Fig. S3), most polymers showed the multimodal size distribution in water, which reflects relatively large numbers of PDI values. There seems to be no correlation between antimicrobial activity and surface potential. Comparing z-average diameters, B1 appears to form a more stable assembling structure than any other polymers, supposedly core–shell micelles. Previous studies also support the fact that amphiphilic or cationic block copolymers using these types of functionalized polycarbonates form core–shell micelles with diameters below 50 nm.4,18,29,44 Due to these reasons, hydrophobic moieties of B1 are sequestered in the assembling core, which could lessen the chance for the ME side chains to encounter the bacterial cell surface. On the other hand, R1 could form loose aggregates comprising several polymer chains where some ME side chains can be exposed outward, which may enhance interaction with the bacterial surface of E. coli. In this way, R1 demonstrates a higher efficacy against E. coli than both H1 and B1.

Table 3 Size and surface potential of cationic polycarbonatesa
Polymers z-Average (d, nm) PDI z-Potential (mV)
a Measured at 1 mg mL−1 in water at 25 °C.
H1-a 374 0.38 +40
B1 21 0.42 +14
R1 303 0.57 +23
R2 94 0.21 +33
R3 156 0.42 +18
R4 125 0.79 +19

Effect of non-cationic side chain structure

R3 and R4 are random copolymers with more hydrophobic side chains compared to the ME side chains. Indeed, mean diameters of R3 and R4 were quite lower than that of R1 (Table 3), suggesting a stronger contribution by the ethyl and benzyl groups to hydrophobic interaction to form smaller aggregates. Contrary to expectations from the hydrophobicity of these side chains, R4 exhibited the same MIC value as H1 and B1 and higher than R1 and R3. This motivated us to consider another factor that could affect antimicrobial activity of the polymers other than the hydrophobicity of the side chains such as the assembling structure and conformation in the test environments.

We studied next the size distribution of polymers in different media by DLS. The mean diameters and PDIs of R1, R3, and R4 are summarized in Table 4. All three random copolymers were found to form larger aggregates in a broth than in water (see also Fig. S4). In particular, R3 and R4 form the aggregates over 1 μm, whereas the increase in the size of R1 was not as large as those of R3 and R4. This difference could be caused by interaction between the hydrophobic side chains of the charged copolymers and ingredients in the broth such as proteins. Overall, this proves that the ME side chains of R1 contribute to the suppression of protein adsorption, considering the previous study where the ME side chains were found to alleviate the interaction with proteins.34 In contrast, the benzyl side chains of R4 strongly interact with the ingredients, which facilitates further aggregation. Such large aggregates probably impair the antimicrobial property due to the large size leading to cancellation of charges within the aggregate, limiting the interaction of the benzyl groups with the bacterial cell membrane. Although R3 also forms large aggregates in the broth, the aggregates are sufficiently flexible and likely dissociated to expose the ethyl groups by the interaction with the bacterial cell surface.

Table 4 Size and size distribution of cationic copolycarbonates in different mediaa
Polymers z-Average (d, nm)/PDI
Water PBS Broth
a Measured at 1 mg mL−1 at 25 °C.
R1 303/0.57 592/0.40 514/0.18
R3 156/0.42 207/0.40 1817/0.53
R4 125/0.79 477/0.36 1728/0.31

Antimicrobial activity of cationic polycarbonates against B. subtilis

The primary ammonium-tagged polycarbonates used in this study were also effective against B. subtilis, except for the block copolymer B1 (Fig. S5). Even though small size aggregates are usually favored to permeate a thick peptidoglycan layer, a characteristic of Gram-positive bacteria,2 the MIC of B1 against B. subtilis was quite higher than those of any other polymers (250 μg mL−1, Table 2). Gram-positive bacteria are usually more negatively charged than Gram-negative bacteria. Thus, the cationic property of the polymer should influence increased interaction with the bacterial cell surface of Gram-positive bacteria compared to Gram-negative bacteria. Although the high MIC could be due to the lowest zeta potential for B1, the details are still unclear. For membrane-disruption-based bacteriolysis, the hydrophobic part of the polymer is required to invade the lipid bilayer. The self-assembled structure of B1 may approach and attach to the bacterial cell surface, but may not be disassembled. In this manner, the ME side chains in the hydrophobic core of the assembly are not sufficiently revealed, decreasing the activity of B1 against B. subtilis. The difference with E. coli (64 μg mL−1) could be explained by the difference in the initial interaction between the least positively charged surfaces of the B1 assembly and the less negatively charged bacterial surface.

On the contrary, R1 is effective against B. subtilis, showing the MICs as low as 32 μg mL−1, which is lower than that of H1. This indicates that the ME side chains partially exposed to the surface of the loose assemblies are effective for the bacteriolysis. Similarly, R2 displayed a lower MIC against B. subtilis (125 μg mL−1) than B1, despite possessing a lower composition of the cationic repeating units. These results indicate that the ME side chain is also beneficial in increasing the antimicrobial activity against Gram-positive bacteria.

Hemolytic activity

Hemolytic activity is often studied for estimating cytotoxicity of the antimicrobial agents evaluated for administration via blood. Since the mammalian cell membranes are comprised of neutral phospholipids, cationic polymers capable of a strong hydrophobic interaction are liable to induce lysis of the healthy mammalian cells.14,15 As already described, hydrophobicity is necessary to eradicate Gram-negative bacteria. Many efforts directed at exploring the optimum balance of the hydrophobicity to maximize the antimicrobial activity and minimize the hemolytic property have been reported.7,21,22

The cationic copolycarbonates with typical hydrophobic side chain units R3 and R4 indicated a high hemolytic property (Table 2 and Fig. 4), with a high sensitivity to polymer concentration. The hemolysis reaches almost 50% at 50 μg mL−1 of the polymer concentrations for R3 and R4. The HC50 of the polymers, which is the concentration of the drug needed to cause 50% of hemolysis, was 500 μg mL−1 for PEI, which is better than those of R3 and R4 but is still low. H1-a and polycarbonates with the ME side chain units (B1 and R1) exhibited a low hemolytic property with more than 2000 μg mL−1 of HC50. By comparison with H1-a, the high hemolytic property of R3 and R4 is proved to be associated with the incorporated hydrophobic side chains. Ethyl and benzyl groups have a high affinity to a lipid moiety of phospholipids in the mammalian cell membrane.

image file: c9bm00440h-f4.tif
Fig. 4 Hemolytic property of cationic polycarbonates.

By contrast, an increase of hemolysis as a function of the concentration of R1 is quite moderate, and the hemolysis at 2000 μg mL−1 of the polymer concentration was 26%. Although the value is somewhat higher than that of H1-a, the ME side chains can be recognized to be sufficiently blood compatible, compared to ethyl and benzyl groups. Furthermore, the MICs of R1 are one level lower than those of H1-a (Table 2 and Fig. 3). Thus, R1 can be regarded as a more efficient antimicrobial polymer than H1-a in terms of the selectivity (HC50/MIC: >63 (R1) vs. >31 (H1-a)). The ME side chains could improve antimicrobial activity while suppressing elevation of hemolysis. B1 exhibited among the lowest hemolytic activities of the tested polymers. The difference in the hemolytic property of H1-a, B1, and R1 can be rationalized by the affinity of the polymer side chains with phospholipids in the cell membrane of RBCs, as described for antimicrobial activity against E. coli. The sequestered ME side chains of B1 are hardly exposed to the surface of the assembly, and therefore their interaction with RBCs becomes minimum. Loosely assembled R1 could actively interact with RBCs, resulting in the slightly higher hemolysis. According to Kuroda's report for the poly(methacrylate) antimicrobials, concentration dependence on hemolysis is characteristic of N-protonated amines as the cationic moiety.24 Unlike the QAS-tagged polycarbonates,3,18 our polycarbonate platforms also indicated the concentration dependence, but with a very small increment. Nevertheless, this study proves the “mono-ether” ME side chain to be a blood compatible hydrophobic unit which plays a key role in improving the cell selectivity of antimicrobial cationic polymers.

Hydrolytic property

Biodegradability is a key feature of these aliphatic polycarbonate-based biomaterials, and understanding their degradation rate and behavior is indispensable for designing their application styles and predicting the materials lifetime. According to Nederberg's report, cationic polycarbonates with QAS are not sensitive to hydrolysis in PBS at 37 °C,3 and as previously reported by us, these polycarbonates could be eventually degraded in the presence of enzymes.34

In contrast, the cationic polycarbonates with primary ammonium groups reported in this study were readily subjected to hydrolysis. Fig. 5 represents the time course of 1H-NMR of H1-a treated in PBS formulated with D2O at 37 °C. Surprisingly, hydrolysis proceeds fast, and several signals (a′ to d′) appeared in the NMR chart after 24 h. By reference to the previous reports regarding the cyclic carbonate synthesis,25,26,29 these signals likely originate from the formed substituted 1,3-diols and oligocarbonates and are observed after 1 h of treatment. These results indicate the fact that the polymers degrade during antimicrobial tests. In addition, the NMR charts in Cai's report28 were found to be rather similar to our data after hydrolysis, indicating that the molecular lengths of their polymers are quite short. This suggests two possible modes of action for the bacterial growth inhibition by these cationic polycarbonates. One possibility is the eradication of the bacteria by the polymers in the initial stages of the test, such as when t = 2 and 4 h, and the other is that quite short oligomers, such as a hydrolysate, could still retain high antimicrobial activity. Either way, the quick hydrolytic property affords an additional benefit of reducing the toxicity of the polycations by degradation,45 due to which the slightly high hemolytic property of the cationic copolycarbonate R1 could be within an acceptable range.

image file: c9bm00440h-f5.tif
Fig. 5 Time course of H1-a in PBS at 37 °C monitored by 1H-NMR (400 MHz, D2O).


Cationic aliphatic polycarbonates with primary ammonium side chains and molecular weights of around 10[thin space (1/6-em)]000 g mol−1 have been synthesized by exploiting engineered monomer synthesis and controlled polymerization. In addition, this study has demonstrated for the first time the use of a cationic polycarbonate comprised of a single component, the primary ammonium-tagged polycarbonate, for inhibition of Gram-negative bacteria. We further found that a comonomer unit with methoxyethyl side chains can serve as a hydrophobic moiety enhancing antimicrobial activity and blood compatibility. Unlike cationic polycarbonates with quaternary ammonium salt side chains, the primary ammonium-tagged polycarbonates are degraded in a quite short period, which could not only mitigate the toxicity inherent to polycationic compounds but also improve bioabsorbability. While the developed polymers display excellent activity, the relation between molecular weights and biological activity needs to be further explored. In summary, we report the development of a novel and highly promising class of cationic polycarbonates for use as biodegradable antimicrobials, to improve therapeutic development in this area by compensating the conventional quaternary ammonium system.

Conflicts of interest

There are no conflicts to declare.


This work has been supported by funding from Takeda Science Foundation and Ito Science Foundation. Haruka Tsuchiya and Hyunmi Choi are thanked for technical support.

Notes and references

  1. R. E. W. Hancock and H.-G. Sahl, Nat. Biotechnol., 2006, 24, 1551–1557 CrossRef CAS PubMed.
  2. C. Ergene, K. Yasuhara and E. F. Palermo, Polym. Chem., 2018, 9, 2407–2427 RSC.
  3. F. Nederberg, Y. Zhang, J. P. K. Tan, K. Xu, H. Wang, C. Yang, S. Gao, X. D. Guo, K. Fukushima, L. Li, J. L. Hedrick and Y.-Y. Yang, Nat. Chem., 2011, 3, 409–414 CrossRef CAS PubMed.
  4. A. C. Engler, J. P. K. Tan, Z. Y. Ong, D. J. Coady, V. W. L. Ng, Y. Y. Yang and J. L. Hedrick, Biomacromolecules, 2013, 14, 4331–4339 CrossRef CAS PubMed.
  5. V. W. L. Ng, J. P. K. Tan, J. Leong, Z. X. Voo, J. L. Hedrick and Y. Y. Yang, Macromolecules, 2014, 47, 1285–1291 CrossRef CAS.
  6. R. J. Ono, A. L. Z. Lee, W. Chin, W. S. Goh, A. Y. L. Lee, Y. Y. Yang and J. L. Hedrick, ACS Macro Lett., 2015, 4, 886–891 CrossRef CAS.
  7. Y. Qiao, C. Yang, D. J. Coady, Z. Y. Ong, J. L. Hedrick and Y.-Y. Yang, Biomaterials, 2012, 33, 1146–1153 CrossRef CAS PubMed.
  8. J. P. K. Tan, D. J. Coady, H. Sardon, A. Yuen, S. Gao, S. W. Lim, Z. C. Liang, E. W. Tan, S. Venkataraman, A. C. Engler, M. Fevre, R. Ono, Y. Y. Yang and J. L. Hedrick, Adv. Healthcare Mater., 2017, 6, 1601420 CrossRef PubMed.
  9. C. Yang, S. Krishnamurthy, J. Liu, S. Liu, X. Lu, D. J. Coady, W. Cheng, G. De Libero, A. Singhal, J. L. Hedrick and Y. Y. Yang, Adv. Healthcare Mater., 2016, 5, 1272–1281 CrossRef CAS PubMed.
  10. W. Chin, C. Yang, V. W. L. Ng, Y. Huang, J. Cheng, Y. W. Tong, D. J. Coady, W. Fan, J. L. Hedrick and Y. Y. Yang, Macromolecules, 2013, 46, 8797–8807 CrossRef CAS.
  11. J. M. W. Chan, X. Ke, H. Sardon, A. C. Engler, Y. Y. Yang and J. L. Hedrick, Chem. Sci., 2014, 5, 3294–3300 RSC.
  12. Z. Y. Ong, K. Fukushima, D. J. Coady, Y.-Y. Yang, P. L. R. Ee and J. L. Hedrick, J. Controlled Release, 2011, 152, 120–126 CrossRef CAS PubMed.
  13. L. Pu, J. Xu, Y. Sun, Z. Fang, M. B. Chan-Park and H. Duan, Biomater. Sci., 2016, 4, 871–879 RSC.
  14. L. Brown, J. M. Wolf, R. Prados-Rosales and A. Casadevall, Nat. Rev. Microbiol., 2015, 13, 620–630 CrossRef CAS PubMed.
  15. A. C. Engler, N. Wiradharma, Z. Y. Ong, D. J. Coady, J. L. Hedrick and Y.-Y. Yang, Nano Today, 2012, 7, 201–222 CrossRef CAS.
  16. H. Takahashi, G. A. Caputo, S. Vemparala and K. Kuroda, Bioconjugate Chem., 2017, 28, 1340–1350 CrossRef CAS PubMed.
  17. Y. Li, K. Fukushima, D. J. Coady, A. C. Engler, S. Liu, Y. Huang, J. S. Cho, Y. Guo, L. S. Miller, J. P. K. Tan, P. L. R. Ee, W. Fan, Y. Y. Yang and J. L. Hedrick, Angew. Chem., Int. Ed., 2013, 52, 674–678 CrossRef CAS PubMed.
  18. K. Fukushima, J. P. K. Tan, P. A. Korevaar, Y. Y. Yang, J. Pitera, A. Nelson, H. Maune, D. J. Coady, J. E. Frommer, A. C. Engler, Y. Huang, K. Xu, Z. Ji, Y. Qiao, W. Fan, L. Li, N. Wiradharma, E. W. Meijer and J. L. Hedrick, ACS Nano, 2012, 6, 9191–9199 CrossRef CAS PubMed.
  19. K. Fukushima, S. Liu, H. Wu, A. C. Engler, D. J. Coady, H. Maune, J. Pitera, A. Nelson, N. Wiradharma, S. Venkataraman, Y. Huang, W. Fan, J. Y. Ying, Y. Y. Yang and J. L. Hedrick, Nat. Commun., 2013, 4, 2861 CrossRef PubMed.
  20. R. Liu, X. Chen, S. Chakraborty, J. J. Lemke, Z. Hayouka, C. Chow, R. A. Welch, B. Weisblum, K. S. Masters and S. H. Gellman, J. Am. Chem. Soc., 2014, 136, 4410–4418 CrossRef CAS PubMed.
  21. L. M. Thoma, B. R. Boles and K. Kuroda, Biomacromolecules, 2014, 15, 2933–2943 CrossRef CAS PubMed.
  22. K. Lienkamp, K.-N. Kumar, A. Som, K. Nüsslein and G. N. Tew, Chem. – Eur. J., 2009, 15, 11710–11714 CrossRef CAS PubMed.
  23. E. F. Palermo, D.-K. Lee, A. Ramamoorthy and K. Kuroda, J. Phys. Chem. B, 2011, 115, 366–375 CrossRef CAS PubMed.
  24. E. F. Palermo and K. Kuroda, Biomacromolecules, 2009, 10, 1416–1428 CrossRef CAS PubMed.
  25. R. C. Pratt, F. Nederberg, R. M. Waymouth and J. L. Hedrick, Chem. Commun., 2008, 114–116,  10.1039/B713925J.
  26. D. P. Sanders, K. Fukushima, D. J. Coady, A. Nelson, M. Fujiwara, M. Yasumoto and J. L. Hedrick, J. Am. Chem. Soc., 2010, 132, 14724–14726 CrossRef CAS PubMed.
  27. Yamagata University, WO/2015/170769, 2015.
  28. A. Nimmagadda, X. Liu, P. Teng, M. Su, Y. Li, Q. Qiao, N. K. Khadka, X. Sun, J. Pan, H. Xu, Q. Li and J. Cai, Biomacromolecules, 2017, 18, 87–95 CrossRef CAS PubMed.
  29. K. Fukushima, R. C. Pratt, F. Nederberg, J. P. K. Tan, Y. Y. Yang, R. M. Waymouth and J. L. Hedrick, Biomacromolecules, 2008, 9, 3051–3056 CrossRef CAS PubMed.
  30. D. J. Coady, K. Fukushima, H. W. Horn, J. E. Rice and J. L. Hedrick, Chem. Commun., 2011, 47, 3105–3107 RSC.
  31. S. H. Kim, J. P. K. Tan, K. Fukushima, F. Nederberg, Y. Y. Yang, R. M. Waymouth and J. L. Hedrick, Biomaterials, 2011, 32, 5505–5514 CrossRef CAS PubMed.
  32. K. Fukushima, Biomater. Sci., 2016, 4, 9–24 RSC.
  33. K. Fukushima, K. Honda, Y. Inoue and M. Tanaka, Eur. Polym. J., 2017, 95, 728–736 CrossRef CAS.
  34. K. Fukushima, Y. Inoue, Y. Haga, T. Ota, K. Honda, C. Sato and M. Tanaka, Biomacromolecules, 2017, 18, 3834–3843 CrossRef CAS PubMed.
  35. K. Fukushima, Polym. J., 2016, 48, 1103–1114 CrossRef CAS.
  36. B. G. G. Lohmeijer, R. C. Pratt, F. Leibfarth, J. W. Logan, D. A. Long, A. P. Dove, F. Nederberg, J. Choi, C. Wade, R. M. Waymouth and J. L. Hedrick, Macromolecules, 2006, 39, 8574–8583 CrossRef CAS.
  37. The impurities could be ring-opened monomer and a residual alcohol used when the side chain functionality is introduced. Since they are alcohols, they would also serve as an initiator.
  38. K. Lienkamp, A. E. Madkour, A. Musante, C. F. Nelson, K. Nüsslein and G. N. Tew, J. Am. Chem. Soc., 2008, 130, 9836–9843 CrossRef CAS PubMed.
  39. A. King, S. Chakrabarty, W. Zhang, X. Zeng, D. E. Ohman, L. F. Wood, S. Abraham, R. Rao and K. J. Wynne, Biomacromolecules, 2014, 15, 456–467 CrossRef CAS PubMed.
  40. M. Isik, J. P. K. Tan, R. J. Ono, A. Sanchez-Sanchez, D. Mecerreyes, Y. Y. Yang, J. L. Hedrick and H. Sardon, Macromol. Biosci., 2016, 16, 1360–1367 CrossRef CAS PubMed.
  41. C. A. Hae Cho, C. Liang, J. Perera, J. Liu, K. G. Varnava, V. Sarojini, R. P. Cooney, D. J. McGillivray, M. A. Brimble, S. Swift and J. Jin, Biomacromolecules, 2018, 19, 1389–1401 CrossRef CAS PubMed.
  42. L. T. Nguyen, E. F. Haney and H. J. Vogel, Trends Biotechnol., 2011, 29, 464–472 CrossRef CAS PubMed.
  43. C. Wang, O. Y. Zolotarskaya, S. S. Nair, C. J. Ehrhardt, D. E. Ohman, K. J. Wynne and V. K. Yadavalli, Langmuir, 2016, 32, 2975–2984 CrossRef CAS PubMed.
  44. S. H. Kim, J. P. K. Tan, F. Nederberg, K. Fukushima, J. Colson, C. Yang, A. Nelson, Y.-Y. Yang and J. L. Hedrick, Biomaterials, 2010, 31, 8063–8071 CrossRef CAS PubMed.
  45. I. Ullah, K. Muhammad, M. Akpanyung, A. Nejjari, A. L. Neve, J. Guo, Y. Feng and C. Shi, J. Mater. Chem. B, 2017, 5, 3253–3276 RSC.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9bm00440h
Present address: Department of Chemistry and Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. E-mail:; Fax: +81-3-5841-7306; Tel.: +81-3-5841-8661.

This journal is © The Royal Society of Chemistry 2019