Synthesis of biologically derived poly(pyrogallol) nanofibers for antibacterial applications

Abstract

Herein, we present the facile synthesis of poly(pyrogallol) biopolymers and their application as antibacterial agents. Pyrogallol is a class of phenolic compounds that can be found in various plants. Polymerization was performed by the auto-oxidation of pyrogallol under a hydrated condition. The microscopic image of poly(pyrogallol) shows a highly homogenous nanofibrous structure with a diameter of 100.3 ± 16.3 nm. Spectroscopic analysis by FT-IR spectroscopy, Raman spectroscopy, and XPS corroborated the formation of ether (C-O-C) bonds between the hydroxyl group and adjacent carbons of pyrogallol during polymerization. The FT-IR and XPS spectra also revealed redox-active gallol functional groups on poly(pyrogallol) nanofibers, which can be used to release free electrons and protons during oxidation followed by the generation of reactive oxygen species (ROS). The generated ROS from poly(pyrogallol) was used to inhibit the growth of bacteria, Escherichia coli, at a inhibition rates of 56.3 ± 9.7% and 95.5 ± 2.0% within 0.5 and 2 h, respectively. This finding suggests that poly(pyrogallol) can be used as a naturally occurring antibacterial agent for various biomedical and environmental applications.

Graphical Abstract

Synthesis of biologically derived polyphenols, polypyrogallol nanofibers, and their applications as antibacterial agents are shown. Polypyrogallol contains a huge potential for wound healing, blood filtration, and applications in biomedical electronics.

Bacterial aggregation and their rapid proliferation have been a primary challenge in a variety of areas including food industry, biomedical devices, wastewater treatment, and wound healing process.¹ There has been a high risk of bacterial infection when using biomedical devices like central venous and urinary catheters, which leads to acute hematogenous infections.² Bacterial infections also lead to major challenges in the wound healing process. In the United States, 2.5% of the total population suffered from chronic wounds and related infections.³ Numerous efforts have been made to develop effective strategies to minimize bacterial growth. It is well known that the bacterial populations attach to solid substrates for survival, forming biofilms.⁴ Biofilms are dense microbial communities adhering to surfaces, which secrete an extracellular matrix mainly composed of water, polysaccharides, DNA, or proteins. Several strategies on adjusting the surface properties have been studied during the past few decades to prevent the biofilm formation.⁵

Two universal approaches have been widely used for treating bacterial infections. The first strategy is to impede bacterial attachment via morphological, physical, or chemical interactions, including electrostatic repulsion and steric hindrance with the super hydrophobic surface.^6–8 Various surface modification techniques have been studied for decreasing the surface energy; however, the poor stability and low persistence in the hydrated environment by surface hydrolysis have been technical challenges for long-term applications.⁹ The other approach is to actively kill the bacteria using antimicrobial compounds by either direct contact or exogenous release of biocides.

For example, metal nanoparticles, such as silver (Ag), copper (Cu), titanium (Ti), and zinc (Zn), have been used for suppressing bacterial growth in many applications.^10–13 Oxidative stress generated by reactive oxygen species (ROS) and metal ion penetration into the bacteria are two main mechanisms that damage bacterial cells and lead to cell death.¹⁴ However, the tedious preparation steps of nanoparticles and the risk of depleting unwanted substances, such as metallic ions, to the targets have become other technical challenges to address. Ag nanoparticles are the most widely used material for antibacterial applications. However, recent studies have indicated that Ag⁺ ions released from sliver nanoparticles could be toxic to cell lines, damaging DNA, and breaking the DNA strands.^15–17 The cytotoxicity of metal nanoparticles is challenging to investigate since various parameters, such as size, charge, composition, and shape of nanomaterials can adversely affect.¹⁸ In order to overcome these challenges, there has been extensive research on the use of naturally occurring materials as antibacterial agents.¹⁹ For instance, chitosan, a natural cationic biopolymer synthesized by the deacetylation of chitin, has been used as an antibacterial carrier.^20–23 Qi et al. evaluated the in vitro antibacterial activities of chitosan nanoparticles with and without copper (Cu), showing the significant inhibition of various microorganisms by both chitosan and copper-loaded chitosan nanoparticles. One proposed mechanism is that chitosan and copper-loaded chitosan nanoparticles exhibit high surface charge density that can provide a strong affinity toward the membrane of bacterial cells, thereby altering membrane permeability and leading to cell death.²³

Recently, redox active biomaterials have shown a promising result as antibacterial agents since diverse forms of ROS can be generated via free electrons and protons by the redox reaction of the biomaterials within the aqueous environment. Catechol-bearing biomaterials have been one of the well-known redox active materials that exhibits reversible oxidation to ortho (o)-quinones in aqueous media.^24,25 The catechol (1,2-dihydroxyphenyl) is a type of polyphenol that contains two or more hydroxyl groups attached to benzene backbones. Antimicrobial activities based on the redox chemistry of catechols have been reported for melanins that are naturally sourced from animal or fungi.^26–28 Eumelanin biopigments extracted from Equus ferus hair exhibited a promising antibacterial activity by achieving complete inhibition of bacterial growth within 4 h of incubation. This suggests that the naturally occurring melanin may serve as a promising antibacterial agent with unique redox chemistry and ROS generation capability.²⁸ However, poor adhesion with native substrates and uncontrollable topography are other challenges especially for coating or fabricating the composite. Therefore, synthesizing biologically derived redox active materials would be advantageous for developing more functional antibacterial agents.

Polyphenols are biologically active compounds derived from plant-based natural products such as fruits, vegetables, coffee, and tea. Polyphenols are the most abundant antioxidants in human diets, which are known to prevent degenerative diseases, cardiovascular diseases, cancers, and osteoporosis.²⁹ Catechol and/or gallol (1,2,3-trihydroxyphenyl) groups are the characteristic chemical functionalities of the monomer of polyphenols such as tannic acid, epigallocatechin gallate, and epicatechin gallate.^30,31 Moreover, polyphenols are redox active under a hydrated condition that can be reversibly oxidized and reduced by free electrons and protons.³² Due to their various physical, chemical, and electrochemical properties, polyphenols have gained significant interest in various research fields including energy storage, heavy metal ion removal, drug delivery, and hydrogel preparation.^33–36 Most approaches of producing the polyphenols are associated with the synthetic procedure using templates or metallic ions such as Fe³⁺ or Cu²⁺.^37,38 Template-free and metal ion-free synthesis could be advantageous for biomedical and environmental applications as it reduces the use of potentially toxic metallic ions and the synthetic steps. Herein, we report the synthesis of a biopolymer, poly(pyrogallol), via auto oxidative polymerization of pyrogallol (1,2,3-trihydroxybenzene) and its antibacterial activity. The microstructures, redox chemistry, and ROS generation potential of poly(pyrogallol) were evaluated by microscopic and spectroscopic techniques as compared to the counterpart, natural melanin pigments.

The chemical structure of pyrogallol consists of a 3,4,5-trihydroxylbenzoyl (gallol) group, which is capable of forming noncovalent bonds with neighboring molecules including proteins, peptides, or polysaccharides via hydrogen bonds or hydrophobic interactions.³⁹ The localized hydroxyl groups of the subunits are capable of forming a coordination bonding with multivalent metallic cations such as iron (Fe²⁺, Fe³⁺), magnesium (Mg²⁺), or calcium (Ca²⁺).^38,40 Furthermore, the oxidation of the gallol functional group in an aqueous solution can reproduce free electrons and protons, which can further form ROS including hydroxyl radicals, superoxide anions, or hydrogen peroxide. This suggests that the amount of generated ROS is based on the number of hydroxyl groups existing in the benzene ring. Eumelanins (Mel) extracted from Sepia officinalis were selected as a counterpart to corroborate our hypothesis. Eumelanins are a subset of naturally occurring melanins that consist of randomly polymerized tetramer units of 5,6-dihydroxyindole (DHI) and 5,6-dihydroxyindole-2-carboxylic acid (DHICA) subunits, as shown in Figure 1(d).⁴¹ The protomolecules change the self-assembly into a spherical nanostructure via strong π-π stacking and hydrogen bonding (Figure 1(c)).⁴² Eumelanins extracted from the horsehair have been previously studied to generate ROS by reversible oxidation from catechols to ortho-quinone, which were applied to completely reject both Gram-negative and Gram-positive bacteria within 4 hours of incubation under the aqueous media.²⁸

Figure 1.

Scanning electron microscope (SEM) images are shown for (a) poly(pyrogallol) (PPG) and (c) eumelanin (Mel) extracted from Sepia officinalis (scale bar = 1 μm). Proposed molecular structures of (b) PPG and (d) Mel are shown.

Pyrogallol was oxidatively polymerized within the aqueous phosphate buffer solution (pH = 7.4). The microstructure of polymerized pyrogallol (PPG) is nanofibrous with a diameter of 100.3 ± 16.3 nm (Figure 1(a)). In general, it is common that the synthetic procedure without the templates leads to non-homogenous microstructures. Numerous research efforts have been made to synthesize homogeneous microstructures using the natural or synthetic templates such as surfactants, proteins, and peptides.^43–47 Unlike the heterogeneous microstructure of synthetic melanins synthesized via template-free oxidative polymerization,⁴⁸ it was found that polymerized pyrogallol exhibits homogenous microstructures that are largely uniform in diameter and length. We speculate that the hydroxyl group at 3, 4, or 5 position could act as a nucleophile that donates the electron pair to the neighboring sp2-hybridized carbon that could further be self-assembled into a nanofiber.

The variation in the chemical functionality during PG polymerization was examined by spectroscopic techniques, namely, Fourier transform infrared (FT-IR) spectroscopy, Raman spectroscopy (Figure 2 (a) and (b)) and XPS. The peak assignment of the FT-IR spectra is summarized in Table S1. Several peaks between 1615 and 1361 cm⁻¹ (grey band) are considered as sp3 C-C or sp2 C=C bonds from the backbone of PG, which were shifted during polymerization to PPG. Broad O-H stretching vibration peaks (blue band) were found at 3365 and 3223 cm⁻¹, which were shifted to 3446 and 3300 cm⁻¹ after polymerization.^49,50 This is due to the valence electron changes from the adjacent carbons that form covalent bonding with the PG subunits. Five distinct peaks (red band) were observed only from PPG. Two peaks located at 1590 cm⁻¹ and 1562 cm⁻¹ could be assigned to the aromatic skeletal vibration of PPG.⁵¹ The characteristic C-O-C stretching vibration peaks of phenyl ether can be found at 1278 cm⁻¹ and 1038 cm⁻¹.^51,52 The peaks located at 1012 cm⁻¹ are characteristic of vinyl ethers.⁵² Taken together, the FT-IR spectra suggest that ether bonding was formed during polymerization to PPG, as shown in Figure 2 (c).

Figure 2.

(a) FT-IR spectra are measured for pyrogallol (PG) and poly(pyrogallol) (PPG). Four distinct peaks at 1590, 1562, 1278, and 1038 cm⁻¹ suggests that ether (C-O-C) bond is formed during the oxidative polymerization. (b) Raman spectra are shown before and after the polymerization of PG. Polymerization mechanism is proposed in (c).

Chemical cues of PG and PPG were investigated by Raman spectroscopy. Four significant peaks were observed from the spectra of PG (Figure 3(a)). The doublet peaks in the range of 240–350 cm⁻¹ and the peak at 1059 cm⁻¹ (green regions) were assigned to the benzene deformation out-of-plane and in-plane bending vibrations, respectively,^49,50 which were absent in the spectra after the polymerization. This is because the benzene rings of the PG subunit become less mobile and free during polymerization into PPG. Both PG and PPG exhibited a peak at 710 cm⁻¹ (yellow region) that was assigned to the O-H out-of-plane bending vibration. The attenuated peak of PPG compared to PG suggests that the hydroxyl groups of PG were consumed with the adjacent monomers during polymerization.⁴⁹ In addition, PPG exhibits similar broad spectra to Mel (Figure S1) between 1000 and 1800 cm⁻¹, demonstrating the characteristic sp² carbon-rich material peaks,⁵³ whereas Mel showed two broad peaks at about 1380 and 1580 cm⁻¹, similar to D and G bands observed from the Raman spectra of disordered graphite. These peaks are attributed to the stretching of the hexagonal carbon rings and linear stretching of three of the six C-C bonds.⁵⁴ PPG exhibited much lower intensities in Raman than Mel. We cautiously posit that this variation resulted from the low degree of polymerization of PPG with a variety of configurations in meso-scale.

Figure 3.

XPS survey peaks and their atomic percentages are shown for (a) PG, (c) PPG and (e) Mel. Atomic percentage suggests that PPG contains higher oxygen content than Mel. (b) High-resolution O 1s XPS of (b) PG and (d) PPG shows the formation of ether bonding during the polymerization of PG. Comparison between (b) PPG and (d) Mel indicates the higher presence of C–OH/C=O functionalities in PPG. High-resolution peaks are deconvolved by CasaXPS and shown as color lines.

XPS was further used to examine the change in the functional groups during PG polymerization and the disparity between PPG and Mel (Figure 3). No significant changes in carbon and oxygen amounts were found from the survey peaks and atomic weight percentage after polymerization (Figure 3(a) and (c)). PPG exhibited more oxygen and less carbon than Mel, which is associated with the fact that PG contains more hydroxyl groups than Mel in their subunits.

The chemical functional group regarding oxygen was further investigated by high-resolution O 1s peak analysis. PPG exhibited three individual peaks at binding energies of 532.1, 533.2, and 536.3 eV, while PG displayed two distinctive peaks at binding energies of 533.4 and 539.8 eV (Figure 3(b) and (d)). It is noted that the C-O bond at 532.1 eV was decreased and the C-OH and C-O-C bonds at 533.2 and 536.3 eV emerged during polymerization. This corroborates that the polymerization of PG is associated with the formation of ether bonding between OH and the carbon of benzene.⁵⁵

In addition, the deconvolution of the O 1s peak of Mel disclosed three peaks at 531.6, 532.9, and 535.1 eV. These peaks could be assigned based on the subunit of melanins, DHICA: C-O, C-OH/C=O bond of catechols or ortho-quinones, and O=C-O bond of carboxylic acid. Compared to the C-OH/C=O bond of PPG at 533.2 eV, a higher density of C-OH/C=O groups is considered to be present from PPG (45.99%) in comparison to Mel (40.10%).^28,48

Furthermore, the association of the high-resolution O 1s peak at ~ 533 eV and the atomic weight percentage from the survey peak can provide a deeper understanding of the potential density of redox-active C-OH groups. The polymerization mechanism indicates that the C-OH groups are much more than C=O groups. Catechols should be dominant in the monomer of Mel due to the acidic condition of the extraction process. PPG contains 27.05% oxygen from the survey spectra, of which 45.99% is present in the C-OH chemical form, indicating that the total C-OH population was calculated to be 12.44%. A similar assessment leads to 7.29% (= 18.18% O from survey × 40.10% from O 1s high-resolution spectra) of the C-OH composition present in Mel. Considering the molecular weight of the subunits of PPG and Mel (PG = 126.11 g mol⁻¹, DHICA = 193.16 g mol⁻¹), we could speculate that approximately 2.6 times higher density of redox-active C-OH groups are present in PPG than Mel.

The generation of ROS by PPG and Mel was quantitatively analyzed by the H₂O₂ generation using a fluorometric assay. ROS may lead to cell death by damaging their DNA, RNA, proteins, or lipids when ROS levels exceed an organism’s detoxification and repair capabilities.⁵⁶ ROS mainly includes superoxide anions (O₂^•–), hydrogen peroxide (H₂O₂), singlet oxygen (¹O₂), and hydroxyl radicals (^•OH), which are chemical species formed upon incomplete reduction of oxygen.⁵⁷ In this study, photooxidation of PPG or Mel induces free electrons (e⁻) and protons (H⁺). O₂ can be reduced by one electron to generate O₂^•– (primary ROS), which can be converted into more stable H₂O₂ (secondary ROS) by reacting with H⁺.^28,58,59 PPG and Mel were incubated in ddH₂O under ambient light, and H₂O₂ concentrations were monitored (Figure 4(a)).

Figure 4.

(a) Time-course generation of H₂O₂ are measured for PPG and Mel using colorimetry assay of H₂O₂ (n = 4). Samples were collected after the incubation of PPG and Mel in aqueous solution under the ambient light. Proposed mechanisms of ROS generation by (b) PPG and (c) Mel are shown. Superoxides and free protons can incorporate to reproduce H₂O₂. Photo-oxidation of PPG induces free electrons and protons leading to the production of three H₂O₂ while Mel generates two H₂O₂. Stoichiometric contrast of the number of hydroxyl groups could result in the variance of H₂O₂ generation shown in (a).

The concentrations of H₂O₂ were largely increased with the time of measurement from both PPG and Mel. H₂O₂ concentrations generated by PPG and Mel were 10.40 ± 1.20, and 8.0 ± 0.43 at 30 min of incubation period, which were comparable to each other. However, PPG could generate 1.76, 2.11, and 3.08 times more H₂O₂ than Mel at 1, 2, and 4 h of incubation. Primary variation in the H₂O₂ concentrations of PPG and Mel is due to the ROS generation mechanisms, as shown in Figure 4 (b) and (c). The gallol group of PPG consists of three hydroxyl groups in its subunit, which theoretically is capable of producing higher levels of ROS than the catechol groups of Mel that only contain two hydroxyl groups. The deviation of generated H₂O₂ between PPG and Mel is speculated to be higher with the incubation time due to the larger number of free electrons and protons propagated within the media. This is in line with XPS analysis, which indicates that PPG contains broadly 2.6 times higher C-OH redox-active groups than Mel.

To investigate the antibacterial activity of PPG and Mel, we chose Gram-negative anaerobic bacteria, Escherichia coli (E. coli). E. coli is one of the most common pathogens in a living organism, and is known to be the major cause of a broad spectrum of diseases.^60,61 The antibacterial activity of both PPG and Mel was quantitatively evaluated by a colony count method after incubation with E. coli. Then, 20 mg mL⁻¹ PPG and Mel were incubated with 500 μL of E. coli suspensions (c = 1.5 × 10⁵ CFU mL⁻¹; CFU = colony forming units) in a 24-well plate for different lengths of time. Relative E. coli activities after incubation are shown in Figure 5(a).

Figure 5.

In vitro antibacterial activities are tested using E. coli. Bacterial colony count was performed after exposure to PPG and Mel with various time courses (n = 3). Both PPG and Mel exhibited significant antibacterial activities after 2 h and 4 h of incubations. Statistically significant differences are indicated by *p < 0.05 compared to the control. (b) Representative images of agar plates are shown after incubating E. coli at 37 °C for 0.5 h and 4 h, respectively in comparison to the controls.

Both PPG and Mel exhibited significantly efficient antibacterial activities within 30 min of incubation. Compared to the control, PPG and Mel achieved 56.3 ± 9.7 % and 36.9 ± 13.4% antibacterial rates, respectively. In 1 h of incubation, PPG showed approximately 30% higher antibacterial rate than that of Mel. This is due to the higher number of hydroxyl groups present in PPG that can be oxidized to generate ROS in aqueous media. The antibacterial activities of PPG and Mel could lead to the saturated levels of 95.5 ± 2.0% and 93.6 ± 1.5% within 2 h of incubation.

This is comparable to the antibacterial activities of other bioinspired polyphenols. Direct comparison can be made with the melanins extracted from horsehair and the synthetic melanins, which exhibited less than 10% of antibacterial activities in 1 h of incubation.²⁸ Taleb et al. reported that the date syrup polyphenols (32 mg mL⁻¹) extracted from the date fruit cultivar Khadrawi achieved 99.9% reduction in E. coli growth after incubation overnight.⁶² Among the previously introduced polyphenols, it was found that PPG outperforms the antibacterial activity within a short period of incubation.

Conclusions

The auto-oxidation of PG under a hydrated condition provided a facile synthetic procedure to form a highly homogeneous nanofibrous microstructure of PPG without adding any template. The formation of the new bond, C-O-C, was corroborated by FT-IR and XPS spectra, which suggests that PPG was polymerized between the hydroxyl group and adjacent carbons of the PG backbone. We found that the polymerized polyphenolic PPG contained a high density of redox-active gallol functional groups that could be oxidized by releasing three electrons and protons. These free electrons and protons were used to react with oxygen in water to generate ROS. Gallols from PPG and catechols from Mel are considered to be the primary functional groups of polyphenols that allow reversible oxidation and reduction under the hydrated condition. Compared to the eumelanins (Mel) extracted from Sepia officinalis, PPG exhibited at least 30% (1 h incubation) improved E. coli. rejection kinetics by releasing more ROS from the higher density of hydroxyl groups within the polymers.

The results herein indicate that PPG exhibits a huge potential as a biologically derived antibacterial agent for a variety of applications. The synthetic procedure shown here is advantageous as (1) the nanofibrous structure can be formed in a single step, (2) the polymerization procedure is cost-effective, and (3) the as-prepared polymer is bioinert without releasing any metallic ions. Antibacterial performance including kinetics would be further improved by manipulating the redox state of gallols. Reduction of gallols by a chemical reductant such as ascorbate can assist in enhancing the generation of ROS, resulting in improved antibacterial performance.^32,63 Finally, these classes of biopolymers can be further applied in a variety of engineering fields. The fabrication of scaffolds with PPG would be advantageous to prevent infections and, therefore, promote the wound healing process accordingly.^64,65 In addition, PPG can be utilized as an antibacterial agent in many biomedical applications including blood filtration membrane, biomedical electronics, or clinical equipment.^66–68

Experimental

Materials

Pyrogallol (PG), phosphate buffered saline (PBS), and eumelanins (Mel; extracted from Sepia officinalis ink sac) were purchased from Sigma-Aldrich (St. Louis, MO USA) and used as received unless otherwise stated. Amplex Red hydrogen peroxide/peroxidase assay kit was obtained from ThermoFisher Scientific (Waltham, MA USA). Luria-Bertani (LB) broth and agar were purchased from VWR (Radnor, PA USA).

Synthesis of poly(pyrogallol) (PPG) nanofibers

PPG was synthesized by auto-oxidation and polymerization under ambient conditions. Specifically, 2 mg mL⁻¹ PG was dissolved into 50 mL 0.01 M PBS (pH = 7.4) in the darkness with oscillation. After 24 h, the solution was centrifuged at 4000 rpm for 5 min, and the precipitates were washed three times with double-distilled water (ddH₂O). After discarding the supernatant, the sediment was dried in a vacuum oven overnight. Synthesized PPG was kept in a closed container in the darkness under ambient conditions until further processing.

Microscopic and spectroscopic characterization

Scanning electron microscopic (SEM) images were acquired using a Lyra3 GMU FIB SEM (Tescan, Brno, Czech Republic). The Mel and as-prepared PPG were fixed on aluminum stubs with double-sided carbon adhesive tape followed by gold/palladium coating.

Fourier transform infrared (FT-IR) spectra were recorded with attenuated total reflection (ATR) using a Nicolet iS10 FTIR Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Spectra were recorded in a wavenumber range of 400–4000 cm⁻¹ with a resolution of 4 cm⁻¹ and 30 scans.

UV-vis absorption spectra were recorded using a SHIMADZU UV-1800 UV spectrophotometer. Spectra were recorded in the wavelength range of 340–900 nm.

Raman spectra were recorded using an AFM-Raman microscope (NTEGRA Spectra, NT-MDT Spectrum Instruments, Moscow, Russia) with a 10 × objective and a 500 nm wavelength laser over a Raman shift range of 200–2000 cm⁻¹. Data from ten separate scans using 1 mW of laser power and 1 min exposure time were averaged to maximizing the signal-to-noise ratio.

X-ray photoelectron spectroscopy was performed using a Kratos Axis Supra XPS. Survey and high-resolution spectra of 1s orbitals of oxygen (O) were obtained using an Al source. The XPS analysis was conducted using the Casa XPS software.

Pro-oxidant activity assay

The level of H₂O₂ production was used as the metric for reactive oxygen species (ROS) generation. First, 5 mg PPG or Mel were incubated with 0.5 mL ddH₂O for different lengths of time (30 min, 1 h, 2 h, and 4 h) under ambient conditions. The supernatant solutions were assayed for the generation of H₂O₂ using an Amplex red hydrogen peroxide/peroxidase assay kit. Specifically, in the presence of peroxidase, the Amplex Red reagent (10-acetyl-3,7-dihydroxyphenoxazine) reacts with H₂O₂ in a 1 : 1 stoichiometry to produce red-fluorescent oxidation product, resorufin. Resorufin has excitation and emission maxima of approximately 571 nm and 585 nm. The amount of H₂O₂ can be determined by the fluorescence intensity. Data were measured using a fluorescence plate reader (SpectraMax M2e/EA, Molecular Devices, San Jose, CA USA). Each experiment was repeated four times.

Antibacterial activity

The antibacterial activity of PPG and Mel was evaluated using Gram-negative Escherichia coli (E. coli) DH5α strains. PPG and Mel were stored in the darkness before incubation with E. coli. Prior to each antibacterial test, E. coli was streaked from a frozen glycerol stock onto a LB agar plate. A single E. coli colony was collected from the plate and inoculated in 5 mL LB broth. The cultures were incubated at 37°C for 16 h in a platform shaker.

Bacterial growth concentrations were determined by means of optical density (OD) (Cell density meter, Ultraspec 10, Amersham Biosciences, Cambridge, UK) at 600 nm wavelength. Then, 10 mg PPG and Mel were inoculated with 1.5 × 10⁵ CFU mL⁻¹ E. coli suspensions (V = 500 μL) in a 24-well plate. The plates were incubated at 37°C in a platform shaker for different lengths of time (30 min, 1 h, 2 h, and 4 h). Aliquots of the samples were serially diluted and plated on agar media for incubation at 37°C overnight. Visible colonies were counted and compared with the negative controls, which were grown without PPG or Mel. Each experiment was repeated three times. The relative E. coli activity rate (R, %) was calculated using the following equation:

$$R=\frac{N_a}{N_c}\times 100\mathrm{\%}$$

where ${\text{N}}_{\text{c}}$ represents the average concentration of E. coli in control and ${\text{N}}_{\text{a}}$ represents the average concentration of E. coli when treated with PPG or Mel.

Statistical analysis

All data are presented as mean ± standard deviation. The ROS generation and antibacterial test are the average of four and three parallel experiments, respectively. The statistical significance was analyzed by one-way ANOVA using the Origin software.