Introduction

Reactive oxygen species (ROS) such as hydrogen peroxide (H2O2), hydroxyl radical, and superoxide anion are produced constantly in the human body as a consequence of oxygen metabolism [1]. Although ROS at basal levels are important in cell proliferation and survival, elevated ROS levels have been implicated in diseases like atherosclerosis, diabetes, neurodegeneration, and cancer [2]. Therefore, there is an interest to control ROS levels in different diseases by using antioxidant compounds.

Among antioxidants, polyphenols, major antioxidant compounds in plant, are known to show preventive effects against some diseases including cancer [3]. For example, it has been reported that resveratrol, a polyphenol contained in red wine, inhibited tumor growth, attenuated angiogenesis, and prevented tumorigenesis after carcinogen treatment in a mouse model [4, 5]. However, a drawback to use these low molecular weight polyphenols for therapeutic applications is their rapid diffusion throughout the body and short circulation time. Furthermore, these compounds can readily be oxidized and metabolized, thereby losing their antioxidant properties.

The use of polymeric nanoparticles is one of the approaches to solve some of the problems associated with small antioxidants for therapeutic application. It is well explored in the field of drug delivery that polymeric nanoparticles in the range of 20–200 nm show increased circulation times in the blood compared to low molecular weight drugs. Furthermore, polymeric nanoparticles of sub-hundred nanometer in dimension are known to selectively accumulate in cancer tissues [6, 7]. In view of these characteristics, the group of Nagasaki reported redox-active polymeric micelles containing nitroxyl radicals as ROS scavengers. It was shown that these micelles were bioactive in an ischemia-reperfusion model, enhanced cancer drug toxicity, reduced gastric mucosal inflammation as well as offered renal protection [8]. Inspired by the antioxidant properties of polyphenols, we recently reported a polymeric micelle having catechol moieties located inside the core. We showed that the micelles interfered with ROS-mediated vascular endothelial growth factor (VEGF) signaling by showing inhibitory effects in HUVEC tube formation and angiogenesis in the chicken chorioallantoic membrane (CAM) assay [9].

Poly(γ-glutamic acid) (γ-PGA) is a natural polymer produced by Bacillus anthracis and Bacillus subtilis. γ-PGA differs from normal PGA because the polymer backbone is formed by linking the γ-COOH group of glutamic acid instead of the α-COOH group. This difference in structure makes γ-PGA more resistant to proteases [10, 11]. Because of its hydrophilic nature, γ-PGA has found application in both regenerative medicine [12] and drug delivery [13]. For example, the group of Akashi previously reported that γ-PGA randomly grafted with hydrophobic amino acid self-assembled to form nanoparticles about 200 nm in diameter. These nanoparticles were used to deliver proteins [14], siRNA [15], and other drugs [16].

Since our previous work had shown that the catechol groups were hydrophobic enough to drive micellization, we reasoned that randomly grafting of these groups to a hydrophilic γ-PGA would potentially induce self-assembly to form nanoparticles. In this study, we synthesized amphiphilic catechol-grafted γ-PGA (catechol-PGA) and prepared nanoparticles by in situ self-assembly. The nanoparticles were characterized by dynamic light scattering (DLS), atomic force microscopy (AFM), cyclic voltammetry (CV), and their ROS-scavenging activity was evaluated using H2O2 as a model.

Experimental

Instrumentation

Dynamic light scattering

(DLS) measurements were done in 4.5-mL disposable polystyrene cuvettes using an Otsuka Electronics ELSZ-2 dynamic light scattering instrument. The diameter of nanoparticles (Z-average in nm) and polydispersity index (PDI = μ 2 / Γ 2) were obtained by fitting the correlation function using the cumulant method.

Absorbance and fluorescence

measurements were done with a Tecan infinite M200 pro microplate reader using transparent or black 96 polystyrene plates respectively.

UV-vis spectra

were measured with a Hitachi U-2810 spectrophotometer using a 10-mm path-length quartz cuvette.

Atomic force microscope

(AFM) images were acquired on a Seiko SPI300 in dynamic mode using a Si probe (SI-DF20, Seiko). Nanoparticles after dialysis were adsorbed onto a fresh mica surface and air-dried.

Cyclic voltammetry

measurements were done with an ALS 600E electrochemical analyzer and a CS-3A Cell Standard. Cyclic voltammograms were measured using a three-electrode set up: reference electrode: Ag/AgCl, working electrode: glassy carbon (0.79 cm2), and counter electrode: platinum coil.

pH measurements

were done on Laqua F-74 Horiba pH meter.

Materials

Poly(γ-glutamic acid) (γ-PGA) in the sodium form with a M w = 50 kDa was provided by BioLeaders Corp. (Daejeon, Korea). Dopamine hydrochloride and 0.1 M NaOH (aq) solution were purchased from Wako Pure Chemical Industries (Osaka, Japan). 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC⋅HCl), 4-aminoantipyrine (4-AAP), and carbonate buffer (pH 10) were purchased from Tokyo Chemical Industry (Tokyo, Japan). Acetate buffer solution (0.1 M, pH 5.0) and phosphate buffer solution (0.1 M, pH 7.0) were purchased from Nacalai Tesque, Inc. (Kyoto, Japan). 2-Morpholinoethanesulfonic acid monohydrate (MES) was purchased from Dojindo Molecular Technologies, Inc. (Kumamoto, Japan). Amicon Ultra 0.5 mL (MWCO 3,000 Da) centrifugal filters were purchased from Millipore (Billerica, MA, USA). Dialysis tubing (MWCO 2,000 Da) was purchased from Spectrum Laboratories (Rancho Dominguez, CA, USA). The Amplex red assay kit was purchased from Life Technologies Corporation (Tokyo, Japan).

Synthesis of catechol-grafted γ-PGA (catechol-PGA) nanoparticles

The conjugation of dopamine to γ-PGA via an amide bond was carried out under argon atmosphere using EDC as the coupling agent. The polymer concentration is stated as the concentration of sodium glutamate repeating units and abbreviated as COONa. For the conjugation reaction, the COONa concentration was fixed at 33, 66, and 83 mM and catechol/COONa molar ratio was varied while keeping the EDC concentration at 130 mM. A typical procedure for nanoparticles prepared from PGA = 83 mM and the catechol/COONa = 0.17 is as follows: 25.0 mg (0.17 mmol COONa groups) of γ-PGA and 5.5 mg (0.029 mmol) of dopamine hydrochloride were dissolved in 1.5 mL of 0.1 M MES buffer solution (pH 5.0). To the clear solution was added 50.0 mg (0.26 mmol) of EDC in 0.5 mL of 0.1 M MES buffer solution (pH 5.0) and the reaction mixture was stirred for 12 h at RT. The reaction mixture was then transferred to a cuvette for measuring DLS.

Synthesis of cross-linked catechol-grafted PGA (catechol-PGA) nanoparticles

After EDC coupling, the solution was first diluted to 20.0 mL by the addition of 0.1 M MES buffer solution (pH 5.0). The pH was then raised to pH 8.9 by the addition of 0.1 M NaOH (aq) to induce cross-linking of the catechol groups. After stirring for an additional 30 min, the pH was adjusted to pH 6.8 by the addition 0.1 M MES (aq) and the cross-linked nanoparticles were dialyzed against milliQ water for 24 h (5 × 3 L). Nanoparticles obtained in this way could be lyophilized and resuspended in aqueous solution.

Catechol quantification

The amount of catechol groups conjugated to γ-PGA was quantified by the colorimetric 4-aminoantipyrine (4-AAP) assay. In this assay, 4-AAP reacts with catechol in a 1:1 stoichiometry at alkaline pH [17]. Contrary to phenol, catechol is oxidized by air in this assay and an external oxidant is not required. The reaction product has a strong absorbance at 510 nm that can be used to quantify the amount of catechol. After EDC coupling, the non-cross-linked nanoparticles solutions were freed from small reaction products and buffer components by ultrafiltration with water using Amicon filter (4000 g for 90 min). To 100 μL of the nanoparticle solution obtained after ultrafiltration was added 100 μL of 4-AAP solution (50 mM, pH 10.0) in a 96-well plate and incubated for 90 min at 25 °C. The catechol concentrations were then calculated from a standard curve of poly(ethylene glycol)-dopamine conjugate (PEG-dopamine) [18] by measuring the absorbance at 510 nm. A known volume of the solution after Amicon filtration was lyophilized and from the weight of the solid obtained and the 4-AAP data, the catechol/COONa molar ratios were calculated which are shown in Table 1.

Table 1 Catechol/COONa molar ratios as measured by the 4-AAP assay and nanoparticle size before and after chemical cross-linking. [COONa] = 83 mM
Full size table

Time course chemical cross-linking reaction

The pH of the nanoparticle (PGA 83 mM, catechol/COONa = 0.16) solution in 0.1 M MES buffer after EDC coupling was adjusted from 5 to 8.9 by the addition of 0.1 M NaOH (aq). The solution was quickly transferred to a quartz cuvette for measuring UV spectra at different time points. The final concentration of catechol groups in solution was 0.4 μM.

Cyclic voltammetry measurements

Lyophilized nanoparticles (PGA 83 mM, catechol/COONa = 0.16) were dissolved in 0.1 M acetate buffer solution (pH 5.0) with a final catechol concentration of 100 μM. The solution was then measured at a scan rate of 0.2 V/s.

H2O2 scavenging capacity

of the cross-linked nanoparticles (γ-PGA 83 mM, catechol/COONa = 0.16) was evaluated by the Amplex Red assay. The catechol concentration of the nanoparticle solution was varied between 0.6 and 10 μM. The nanoparticle solution in PBS (pH 7.4, 25 μL) was incubated with 25 μL of H2O2 solution (final concentration, 10 μM) for 30 min. After that, 100 μM Amplex Red reagent and 0.2 U/mL horseradish peroxidase were added to the solution. After incubation for another 30 min, the H2O2 concentration was quantified by measuring the fluorescence intensity (λ excitation = 560 nm, λ emission = 590 nm).

Results and discussion

Synthesis and characterization of catechol-grafted γ-PGA (catechol-PGA) nanoparticles

γ-PGA used for preparing the amphiphilic polymer was in the sodium form (COONa) and had a molecular weight of 50 kDa. For introducing the catechol groups, we used dopamine and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) as the coupling agent as shown in Fig. 1. To see the effect of polymer concentration, expressed as the concentration of sodium glutamate repeat units COONa, we used γ-PGA at 33, 66, and 83 mM and varied the catechol/COONa molar ratio. The polymer and dopamine hydrochloride were dissolved in 2-(N-morpholino)ethanesulfonic acid (MES) buffer (0.1 M, pH 5.0) and to the clear solution was added EDC to start the coupling reaction. After stirring the mixture at RT for 12 h, the solutions were analyzed by dynamic light scattering (DLS) to see whether nanostructures were formed. As shown in Fig. 2, at 33 mM of γ-PGA, nanostructures about 25 nm in diameter were formed and increasing the catechol/COONa feed ratio showed only a slight increase in size. The reactions at 66 and 83 mM of γ-PGA, however, resulted in formation of nanoparticles in the range of 25–130 nm in diameter depending on γ-PGA concentration and the catechol/COONa feed ratio. To obtain quantitative data of how much of the added dopamine was conjugated to the polymer, we measured the catechol content using the 4-aminoantipyrine (4-AAP) assay. The 4-AAP assay showed near quantitative incorporation of dopamine into γ-PGA (Table 1).

Fig. 1
figure 1

Synthesis scheme for catechol-grafted γ-PGA by coupling of dopamine to γ-PGA in the sodium form using EDC as the coupling agent

Full size image
Fig. 2
figure 2

Diameter of the catechol-PGA nanoparticles obtained at different γ-PGA concentrations (33, 66, and 83 mM of sodium glutamate repeat units) and the catechol/COONa feed ratio as measured by DLS

Full size image

Unfortunately, the nanoparticles after dialysis were not stable and precipitated in water over the course of several days. This may relate to the increased chain-chain repulsion caused by the higher number of negatively charged carboxylate groups. The pKa of γ-PGA has been reported to be 4.8 which means that during conjugation at pH 5.0, about 61 % of the COOH groups will be present as carboxylate groups [19]. On the other hand, in deionized water after dialysis (pH ~5.6), the amount of carboxylate groups will increase to about 86 %. Furthermore, in the presence of salts, the negative charge of carboxylate groups was shielded, thereby further reducing charge repulsion between γ-PGA polymer chains within the particles. It should be noted that conversion of both phenol groups into phenolate groups are not responsible for this phenomenon due to the fact that both pKa values for dopamine are above 10 [20]. It therefore seems that during the coupling reaction, the interaction between catechol groups was strong enough to hold the γ-PGA polymer chains together. However, at the higher pH and in the absence of salts, the catechol-catechol interaction is not strong enough to prevent polymer dissociation. As a consequence, the nanoparticles may slowly disassemble and expose hydrophobic catechol groups resulting in formation of inter-particle aggregates. It should be noted that the precipitate after dialysis could not be resuspended in any buffer solutions. It is known that catechol undergoes cross-linking reactions in the presence of an oxidant like O2 as shown in Fig. 3a [21, 22]. As the first step, the catechol is oxidized to the quinone that can undergo a reverse dismutation reaction with a close by catechol to form aryloxy radicals that then dimerizes to the dicatechol with the formation of a covalent bond (Fig. 3b). Since the oxidation of the catechol to a quinone is favored at basic pH, this cross-linking will be more pronounced at higher pH.

Fig. 3
figure 3

Cross-linking of the catechol-PGA nanoparticles. a Reaction mechanism of the oxidative cross-linking of the catechol groups. b Time course of the cross-linking reaction as measured by UV-vis. c DLS of the catechol-PGA nanoparticles before and after cross-linking. The catechol-PGA nanoparticles were prepared at 83 mM of sodium glutamate repeat units and catechol/COONa = 0.16

Full size image

Time course of cross-linking of catechol-PGA nanoparticles

In order to stabilize the nanoparticles, we considered chemical cross-linking of neighboring catechol groups within a nanoparticle using the oxidative coupling reaction. To favor intra- but not inter-particle cross-linking, the reaction mixture after EDC coupling was diluted prior to cross-linking. Then, cross-linking was initiated by raising the pH quickly to pH 8.9 by the addition of 0.1 M NaOH (aq). The reaction was followed by UV-vis spectrometry as shown in Fig. 3b. As can be seen, the catechol-PGA nanoparticles showed a broad absorption at around 270 nm that can be assigned to the catechol groups. Within 1 min, the absorption peak starts shifting to about 279 nm along with a strong increase in absorbance. Similar behavior has been observed in NaIO4-induced oxidation of mussel-adhesive protein Mefp-1, and the broad band around 279 nm has been assigned to the coupling product between catechol groups [23].

Characterization of the cross-linked catechol-PGA nanoparticles

The size of the cross-linked catechol-PGA nanoparticles was measured by DLS. After 30 min of cross-linking, the diameter of the particles was 98 nm, comparable to the particle size before cross-linking (Fig. 3c). However, after 4 h, the size increased to 448 nm which may be due to inter-particle cross-linking. Therefore, even though the time course experiment showed an increase of the absorbance assigned to cross-linked catechol groups for up to 4 h, the reaction was stopped after 30 min by quickly adjusting the pH to near neutral (pH 6.8) (Table 1). The cross-linked nanoparticles did not precipitate when stored in water and could be lyophilized and redissolved in aqueous solutions showing their structural integrity. The cross-linked nanoparticles were stable at pH 7.4 for 2 days, and DLS showed a slight shift of the size distribution to smaller sizes (Figure S1, Online Resource 1).

The nanoparticles were also characterized by atomic force microscopy (AFM). As shown in Fig. 4a, b, whereas non-cross-linked particles were present as large spherical structures are about 60–100 nm in size, the diameter of the cross-linked nanoparticles was about 40 nm. As shown in Fig. 4c, d, the cross-sectional images of these particles showed that the nanoparticles without cross-links collapsed upon drying in air and were present as flat structures on the mica surface. Compared to the non-cross-linked nanoparticles, the cross-linked nanoparticles did not lose their structure and shrank upon drying, which may be due to the chemical cross-links that hold the structure together. This also explains why these particles appeared smaller in AFM than DLS.

Fig. 4
figure 4

Morphology of the catechol-PGA nanoparticles before and after cross-linking. AFM images of the nanoparticle solutions a before and b after cross-linking. Cross-sectional images of the nanoparticles c before and d after cross-linking. The catechol-PGA nanoparticles were prepared at 83 mM of sodium glutamate repeat units and catechol/COONa = 0.16

Full size image

Redox and H2O2 scavenging activity of the cross-linked catechol-PGA nanoparticles

Because the chemical structure of the formed dicatechol is different from that of the monocatechol, it was important to know whether the formed dicatechol would be still redox active and capable of scavenging ROS. To answer this question, the redox activity of the cross-linked nanoparticles was studied by cyclic voltammetry (CV). As shown in Fig. 5a, the CV showed a clear but broad oxidation peak with a maximum at around 0.34 V. This result shows that the nanoparticles remained redox active.

Fig. 5
figure 5

Redox and ROS-scavenging activity of the cross-linked catechol-PGA nanoparticles. a Cyclic voltammogram of the cross-linked nanoparticles at pH 5.0. The arrow indicates oxidation peak. Scan rate, 0.2 V/min. b Concentration-dependent H2O2 scavenging activity as measured by the Amplex Red assay. The catechol-PGA nanoparticles were prepared at 83 mM of sodium glutamate repeat units and catechol/COONa = 0.16

Full size image

We further evaluated the H2O2 scavenging activity of the cross-linked nanoparticles using the Amplex Red assay. Nanoparticles at different concentrations were incubated with 10 μM of H2O2 solution for 30 min before measuring the remaining H2O2 concentrations [24]. As can be seen from Fig. 5b, the nanoparticles scavenged H2O2 in a concentration-dependent manner. The nanoparticles slowly lost their H2O2 scavenging activity by air oxidation with about 50 % of the original activity remaining after 2 days at pH 7.4 (Figure S2, Online Resource 1).

Conclusions

To conclude, we prepared nanoparticles from amphiphilic catechol-grafted γ-PGA via in situ self-assembly. The size of the nanoparticles varied in the range of 25–130 nm in diameter depending on the γ-PGA concentration and the catechol/COONa feed ratio. Although the formed nanoparticles were not stable in water, subsequent oxidative cross-linking of catechol groups stabilized the structure without impairing redox and H2O2 scavenging activity. These nanoparticles might find application in antioxidant therapies.